text stringlengths 1.23k 293k | tokens float64 290 66.5k | created stringdate 1-01-01 00:00:00 2024-12-01 00:00:00 | fields listlengths 1 6 |
|---|---|---|---|
A method based on Dempster-Shafer theory and support vector regression-particle filter for remaining useful life prediction of crusher roller sleeve
In order to solve the problem of accurately predicting the remaining useful life (RUL) of crusher roller sleeve under the partially observable and nonlinear nonstationary running state, a new method of RUL prediction based on Dempster-Shafer (D-S) data fusion and support vector regression-particle filter (SVR-PF) is proposed. First, it adopts the correlation analysis to select the features of temperature and vibration signal, and subsequently utilize wavelet to denoising the features. Lastly, comparing the prediction performance of the proposed method integrates temperature and vibration signal sources to predict the RUL with the prediction performance of single source and other prediction methods. The experiment results indicate that the proposed prediction method is capable of fusing different data sources to predict the RUL and the prediction accuracy of RUL can be improved when data are less available.
Introduction
Roller sleeve as an important component is widely used in crusher, the status of its running performance directly affects the health of the whole equipment [1]. However, due to the complex working load, dust and other harsh working conditions, the service life of crusher roller sleeve is not long, and the accurate life of high speed spindle in all kinds of crushers is only thousands of hours. Once the working hours exceed the service life limit, the operation precision of roller will drop sharply and further cause that the machine cannot work properly. So it is very important to improve the reliability, safety and work efficiency of the crusher roller sleeve by means of prognostics and health management (PHM). It is an important part of PHM to predict the RUL of equipment and evaluate the performance of devices [2].
It is critical to establish an appropriate model in the life prediction process. In brief, condition-based monitoring is becoming more and more significant, especially in the RUL prediction. Vibration signal on-line monitoring is one of the most effective methods to monitor the state of crusher roller sleeve health condition (SOH) [3]. The RUL prediction of crusher roller sleeve based on vibration monitoring data is divided into two steps: firstly, construct an indicator to accurately assess the performance degradation of crusher roll sleeve; subsequently, establish an effective model to predict the RUL of crusher roller sleeve.
How to establish an appropriate model under partially observable state is the key to predict the RUL accurately, and it is also the urgent demand for the industrial production. SVR-PF is such a machine learning algorithm to make classification and prediction under small samples [4]. This method based on the statistics theory has been successfully applied to the prediction in the financial, electric and other systems [5,6].
In this paper, a method using acceleration and temperature data is proposed firstly to solve the challenge of low RUL prediction precision based on single data source. However, the noise and vibration interferences caused by other mechanical and systems may severely obscure the roller sleeve signal collected from sensors and make it very challenging to reliably detect the effective components. For the above reasons, a variety of signal analysis methods have been proposed by researchers, such as time-domain, frequency-domain and time-frequency technique. Wavelet analysis is such a widely accepted approach. Then it selects the sensitive features of two signals as input and constructs the SVR-PF model to solve the problem that is difficult to predict with finite state data. The proposed method is evaluated using experimental data respectively. Finally, after assessing the prediction result errors, the conclusion is given in Table 5.
2 Theory introduction of D-S data fusion and SVR-PF 2.1 Theory introduction of D-S data fusion D-S theory fuses data from different sources through the basic probability assignment (BPA) function, and analyses the belief of all the possible propositions in the identification framework, so as to achieve the goal of data fusion [7][8][9][10].
To set the identification framework consists of evidence B and C, m 1 and m 2 are two BPA functions in the identification framework, m 1,2 is the fused BPA function, then the D-S data fusion can be expressed as: where K is the degree of conflict between the evidences B and C K ¼ D-S data fusion theory combines with the same view for different sources of the same problem and eliminates the all conflicting views at once, so that a more reliable fused posterior BPA function can be obtained.
The RUL prediction of crusher roll sleeve based on D-S data fusion has two data sources: (1) the RUL prediction based on temperature data; (2) the RUL prediction based on acceleration data. In this paper, the proposed prediction method based on D-S data fusion and SVR-PF fuses the results of two prediction methods to gain the fused RUL prediction result.
The whole identified framework is set as V Because there's no intersection between T and a, prediction by acceleration data and prediction by temperature data are independent events, then the power set can be expressed as 2 V .
The meaning of all the propositions in the power set 2 V is explained as follows.
-{T} represents the RUL prediction credibility obtained by temperature data; -{a} represents the RUL prediction credibility obtained by acceleration data; -{T ∪ a } represents the RUL prediction credibility obtained by acceleration or temperature data.
Meanwhile, the BPA functions m 1 and m 2 defined in the power set 2 V mean that: m 1 represents the prediction credibility distribution obtained by temperature data in the power set 2 V ; m 2 represents the prediction credibility distribution obtained by acceleration data in the power set 2 V .
The combination of the BPA function based on data fusion is shown in Table 1.
From Table 1, the posterior BPA function based on data fusion can be expressed as: That's the proposed RUL prediction method for crusher roller sleeve, it uses the posterior fusion BPA function and combines the prediction results of two prediction methods to gain a more accurate prediction result.
2.2 Basic theory of SVR-PF 2.2.1 Basic theory of particle filter On the basis of the recursive Bayesian estimation [11][12][13], particle filter becomes a universal algorithm drawing samples from posterior distributions and assigns weights to all the particles by using the Monte Carlo method [13][14][15][16]. Particle filter has more excellent performance on nonlinear and non-Gaussian system than Kalman filter which only has good performance on liner and Gaussian system [17]. The particle filter system state space model can be described as: where x k is the system state, z k is either the system output or the measurement, v kÀ1 is the system noise, and n k is the measurement noise.
We assume that the prior distribution pðx i 0:kÀ1 jz 1:kÀ1 Þ of system is known and N samples from the posterior distribution of system (11) have drawn. The posterior distribution can be approximately described as: The higher is the weight, the higher is the sample probability. d(⋅) represents the Dirac-Delta function.
In order to solve the problem that is very difficult to sample directly from a posterior distribution, a good deal of the problem is the importance sampling technique. It can draw samples directly from the importance distribution. The importance distribution can be described as: Plugging the importance distribution (13) into (12), then the weight can be updated: where pðz k jx i k Þ is the likelihood function, pðx i k jx i kÀ1 Þ is the state transfer distribution. If system (11) subjects to the Markov process, the weight update equation (14) can be reduced to: We set state transfer distribution as the importance distribution: If the likelihood function pðz k jx i k Þ and the prior weights are used to update the new weights [15], the weight renew equation can be reduced to equation (17): There is a wider problem of PF, which is known as degeneracy phenomenon. In order to avoid the problem, resampling is a suitable method. If the system iterates without resampling, the weight of some particles will tend to zero, and all efforts for the weights calculation become meaningless.
The standard method to avoid the degeneracy phenomenon is to renormalize the distribution by removing the small weight particles and duplicating the large weight particles. The weights of all the particles are set to 1/N (N is the number of particles). The resampling algorithm of the standard PF is shown above.
where N eff is the threshold of resampling.
Support vector regression-particle filter
The standard PF algorithm eliminates the small weight particles and duplicates the large weight particles to avoid the degeneracy phenomenon that would cause the loss of particle diversity. Which would make most particles aggregate around the larger weighted ones, so the degeneracy phenomenon still exists. In view of this problem, a new resampling algorithm known as SVR is introduced to rebuild a posterior distribution [18], which has an extremely fast learning speed and advantageous generalization capability. Moreover, SVR has a commendable performance in both classification and regression with a simple structure. Compared with other methods, SVR can avoid the degeneracy phenomenon and keep the diversity of particles in the case of limited samples. What's more, the training speed of SVR is much faster while obtaining better generalization. In view of these advantages, the SVR is selected in this paper to establish RUL prediction model. The application of the SVR is detailed in some studies [19,20]. The fundamental principle of SVR is known as an optimization problem expressed by a regularized functional with constraints [21], the form can be described as: where the regularized functional defined in Hilbert space and generated by s l is represented by V = (f, f) H . The error between the distribution functions F(x) and their estimation F l (x) is represented by s l . The constraint is represented by e. The estimated probability density function (PDF) of distribution F l (x) is F(x). Only the points x i (i = 1, 2, . . . , m) in the particle set should be considered, so equation (19) can be reduced to: If the PDF f(x) is described by kernel functions: Then the regularized functional can be described as: The posterior distribution prediction can be described as an optimization problem with constraints: Set i are non-negative slack variables, then equation (23) can be reduced to a quadratic programming problem: where C is the penalty coefficient. By introducing Lagrange coefficients a i, a à i to equation (24), we get: . . . ; m: Now the solution of equation (25) can be described as: In equation (26), x i is the support vector and the corresponding parameter of non-zero coefficients a à i , ai. Substituting equation (26) into (21), the solution can be transformed into a posterior distribution estimation of an optimization problem.
As discussed above, the PF algorithm can be modified into a new PF algorithm by integrating SVR, which can be described as follows.
Resampling of the posterior distribution starts once the effective sample N eff below the threshold. The two training groups are particle x i k and corresponding weight The resampling posterior distribution is rebuild by these groups. The flow chart of the SVR-PF algorithm is shown in Figure 1.
Proposed prediction method
The proposed method mainly consists of three parts, feature construction, feature signal processing and RUL prediction, see Table 2 for details.
Feature construction
Feature signals are extracted from respective original vibration and temperature signal of crusher roller sleeves.
Since the definition of the original signal in different stage is relatively vague, it is crucial to select a significant sensitive feature that can fully reflect the degradation of roller sleeve. The proposed method evaluates the degradation of roller sleeves by calculating the tendency degree between each feature and running time, which is defined as the Karl Pearson coefficient of the feature. The Karl Pearson coefficient uses the rank to evaluate the tendency degree of a feature. It cannot only evaluate the nonlinear relationship but the monotonicity of the features where x i and y i are the ranks of the time t i , and the ith feature, respectively. N i is the length of the time sequence.
x and y are the means of x i and y i , respectively. The sensitive feature is chosen from the features which has the highest tendency value, i.e., the original feature has the most obvious monotonic trends.
Feature signal processing
In the signal processing part, original vibration and temperature signals are usually formed by the superposition of the characteristic signal and the noise signal, and the random disturbance signal in the noise affects the precision of the prediction result deeply. So, it is important to process the signal for a more accurate prediction result. Signal processing contains the removal of outliers, eliminates the trend item and denoising.
In general, the detection of outliers in signal is based on the previous normal monitoring data. The least squares polynomial is established to estimate the value of the observation data at the next moment, the absolute value of the estimated value subtracts the actual data at current time and the difference is further determined whether the difference is more than a given threshold. If the difference is more than the threshold, it is considered that the observation data are outlier, otherwise they are considered normal data.
In measurement process b xðnÞ ¼ xðn À 1Þ þ 1 2 or signal processing, set x (n À 4) , x (n À 3), x (n À 2) , x (n À 1) are four consecutive data of signal x(n) before time point n. The estimated value of the current time b xðnÞ can be obtained by the linear extrapolation of the least square estimation [22][23][24][25].
Calculating the absolute value of b xðnÞ subtracts the measured value, and compares it with the threshold value d, i.e.
where x(n) are current data, b xðnÞ are the estimated value of the current data obtained by the linear extrapolation through the least square estimation; s is the standard deviation of measured data residuals. If the equation (29) is established, x(n) is the normal value, otherwise, it is the outlier.
The trend term in the measurement signal is the frequency component of the signal, which is larger than the sampling length of the signal. It is generally the result of a slow change of the time sequence in the measurement system. Except the working frequency of the original signal collected by the sensor, there are some random interference signals. The existence of these trends, will cause great error in the correlation analysis or power spectrum analysis in space domain, even distort the low frequency completely. If the measurement signal without removing the trend term is directly used to predict the RUL of roller sleeve, it will directly affect the forecast results, make inappropriate judgments and conclusions. So the extraction and elimination of the measurement signal trend term is an important part of tested data processing.
The original signal is x(t), which can get a discrete time series x(n) by uniformly-spaced sampling. The least square method is used to construct a pth-order polynomial [26][27][28].
where p is a positive integer, means the order of polynomial, and the selection of p value is based on the estimation of the signal trend. If the trend of the signal is linear, choose p = 1.
With y(t), we subtract the original signal x(t) by the polynomial trend term y(t), i.e.
b yðtÞ À xðtÞ À yðtÞ ð 31Þ where b yðtÞ is the signal that removes the trend item. Because of the existence of so lot of noise, the field monitoring signal may submerge in other vibration signals and random noise, which can further causes great impact on the online monitoring. In this paper, wavelet is used to Table 2. BPA function combination based on D-S data fusion.
Step 1: Collect the vibration and temperature signal Step 2: Calculate the absolute correlation matrix of the selected features Step 3: Find the feature which has the highest correlation coefficient among these features.
Step 4: Remove the interference terms of the feature signals Step 5: Calculate each BPA function based on acceleration and temperature data Step 6: Calculate the RUL prediction based on acceleration and temperature fusion data and repeat step 5 to step 6 until a k equals the threshold value Step 7: Calculate the RUL prediction L ¼ ðk þ 1Þ À N denoise the signals. The basic idea of wavelet denoising is to decompose and reconstruct the signal. Because signal and noise at different wavelet spectrum scales have different expressions, we remove the spectral components especially the dominant portions generated by noise at different scales. The wavelet spectrum preserved in this way is the wavelet spectrum of the original signal, basically. We reconstruct the original signal using the reconstruction algorithm of wavelet transform at last [29][30][31][32].
Signal x(t) can be expressed as: In the multi-scale decomposition process, x(t) is always progressively decomposed to two subspaces V j and W j from the space V jÀ1. According to the two-scale equation, we can get the fast recursive algorithm about projection coefficient from c jÀ1,k of x(t) in V jÀ1 to c j,k and d j,k of x(t) in V j and W j.
On the contrary, c jÀ1,k also can be reconstructed by c j,k and d j,k , and the reconstruction formula is as follows.
RUL prediction 3.3.1 Initial state of fusion prediction
Set N as the initial time of prediction, a N is the acceleration in the Nth time point, a à T ;N is the acceleration prediction obtained by temperature data in the Nth time point,a à a;N is the acceleration prediction obtained by acceleration data in the Nth time point.
The first step is calculating the initial value of m 1,i (T) BPA function. From the central limit theorem, a large number of temperature data measurement errors obeys normal distribution, i.e.
where m T T ;N is the mean value of temperature, s 2 T is the variance.
From the experimental result we can see that there is a strong linear relationship between acceleration signal and temperature signal. So, the a T,N estimation also obeys normal distribution, i.e. a T ;N ∼ Nðm a T ;N ; Thus the initial value of the BPA function m 1,i (T) can be described as m 1,N+1 (T) The second step is calculating the initial value of the BPA function m 2,i (a). From the central limit theorem, a large number of acceleration data measurement errors obeys the normal distribution, i.e. a a;N ∼ Nðm a a ;N ; s 2 a Þ ð 41Þ where m a,N is the mean value of acceleration; s 2 a is the variance. If m a,N meets: The initial value of the BPA function m 2,i (a) can be described as m 2,N+1 (a) After getting the value of BPA function m 1,N+1 (a) and m 2,N+1 (a), then we calculate the value of the other BPA functions. Because there is no correlation between the two kinds of prediction methods, it's easy to know that: According to the properties of BPA function: After determining all the values of the BPA function, the third step is calculating the posterior fusion BPA function. The fusion BPA function can be described as m, from formulas (6) and (7), it's easy to know that: With the fusion posterior BPA function, the acceleration can be predicted in the N+1th time point. Set a à T ;Nþ1 is the acceleration prediction obtained by temperature data in the N+1th time point, a à a;Nþ1 is the acceleration prediction obtained by acceleration data in the N+1 th time point, a à Nþ1 is the acceleration prediction obtained by data fusion method in the N+1th time point. Based on formulas (47)-(49), a à Nþ1 meets: Thus we get the initial state of RUL prediction based on D-S data fusion and SVR-PF:
Fusion prediction process
Step 1: calculating the BPA functions. Set N EOL is the end of roller sleeve working life, a à k is the acceleration prediction obtained by data fusion in kth time point, a à T ;k is the acceleration prediction obtained by the temperature data, a à a;k is the acceleration prediction obtained by the acceleration data. Where N + 1 k N EOL , the BPA function can be described as m 1,k+1 (T) and m 2,k+1 (a).
Because there is no correlation between the two prediction models, it's easy to know that: Based on the properties of BPA functions, there are: Step 2: calculating the acceleration prediction in the k +1th time point. From the formulas (6) and (7), formulas (52)-(56), the posterior fusion BPA function m k+1 (T) and m k+1 (a) can be described as: Thus we obtain the acceleration prediction a à kþ1 in the K+1th time point through D-S data fusion.
Step 3: determining whether the acceleration reaches the threshold. If not, then go back to step 1 and continue the prediction; otherwise, calculating the RUL prediction LÃ: In summary, the proposed RUL prediction method based on D-S data fusion and SVR-PF can be described as flow chart in Figure 2.
A prediction model based on D-S data fusion and SVR-PF is established as: where X k is the prediction state, X k , V k are noise.
where X T T ;k is the state obtained by the analysis of temperature data; X T a;k is the state obtained by the analysis of acceleration data; X T DS;k is the state obtained by the analysis of D-S data fusion, and there are: In formulas (63)-(65): l T,k is the degradation parameter of temperature in the kth time point [33]; T T,k , l T T ;k are the temperature degradation parameters in the kth time point [34]; * is the prediction of each corresponding variable.
Thus, in combination with the RUL prediction model of literature [33,34] we can obtain the state equation of prediction model (61), the partial state equation of prediction by temperature data can be expressed as follows.
where a N , b N is the degradation parameter prediction of acceleration in the Nth time point through the analysis of the temperature data. The partial state equation of prediction by acceleration data can be described as follows.
The partial state equation of prediction by data fusion can be described as follows.
Combining formulas (66)-(68) can obtain the state equation of prediction model (61), the measurement equation (61) can be described as follows.
where b T T ;k is the temperature prediction obtained by temperature data, and T a,k is the degradation parameter prediction of acceleration obtained by temperature data.
Experimental demonstration 4.1 Introduction to the data acquisition platform
The test platform layout is shown in Figure 3, named PRONOSTIA [35]. The testing platform is designed by the AS2M Department of the FEMTO-ST Association, the full life test of the roller sleeve is carried out on the data acquisition platform of the rolling bearing, the vibration signal is collected by the 3035B Dytran acceleration sensor (the maximum acquisition range is 50 g), temperature signal acquisition using JCJ100TLB temperature sensor (maximum acquisition range is 200°C). Because the acceleration signal is more severe than the temperature signal, so the full-life test process stops if the acceleration signal amplitude is found to exceed 20 g. Even if the roller sleeve does not out of work, in order to avoid the test platform damage caused by the roller sleeve, we determine the failure of the roller sleeve, stop testing. The acceleration test sampling frequency is 25.6 kHz, each 10 s stores a set of data, each group of data 2560 points, the temperature test sampling frequency is 10 Hz, and each 10 s stores a set of data, each group of data 100 points.
In the test, the testing roller sleeve is 22,324 tapered roller bearing, the roller life-test is carried out 4 times, each time one test roll is damaged. The 1st test and the 2nd test are worked under the condition of radial load 4000 kN, speed 1800 rpm/min; the 3rd test and the 4th test are worked under the condition of radial load 4200 kN, speed 1650 rpm/min. The test results are shown in Table 3.
Due to the different working state of the 4 roller sleeves and the different structure of the roller sleeves, the experimental results are different, which is conform to the actual engineering facts. Then we use the measured experimental data to verify the performance of the proposed RUL prediction method. Because the 4 groups tests are carried out on the same platform, and the analysis methods of each roller sleeve is same, below takes the 1st roller sleeve as the research object to make explanation. The measured temperature and vibration data are shown in Figures 4 and 5.
Feature construction
In order to compare the prediction performance of proposed D-S data fusion and SVR-PF prediction method with the prediction method using the single acceleration data and the prediction method with the temperature data with finite data available, the first key step is to select a good feature signal.
In this paper, the features are selected by calculating the Karl Pearson correlation coefficient between the timedomain characteristics of temperature and acceleration and RUL. As an indicator the features whose correlation coefficient is highest are selected. As a result, the root mean square (RMS) feature of vibration and the absolute mean value feature of temperature are selected. The Karl Pearson correlation coefficient result is shown in Table 4. The feature signal of acceleration and temperature are shown in Figures 6 and 7.
Feature signal processing 4.3.1 Removal of outliers
The first step of signal processing is the removal of outliers. For the nonlinear and non-stationary signal, the existence of outliers can produce spurious harmonic components, further can influence the prediction accuracy. According to the statistical properties of the original data, 3s criterion is used to remove the outliers here. If the residuals in equation (29) exceed 3s the outliers can be eliminated. The temperature and vibration signals after removing the outliers are shown in Figures 8 and 9.
Remove the trend of a smooth
Because of the zero drift of the amplifier caused by temperature variation, the performance of low frequency which exceeds the frequency range of the sensor is not stable with ambient interference around the sensor etc., which caused the collected data of vibration signal and temperature signal in life-test will often deviate from the baseline, and even the degree of the deviation from the baseline will vary over time. The whole process of the deviation from the baseline directly affects the correctness of the signal and should be removed as the trend term. This paper from the perspective of engineering application adopts a simple and practical method to remove the trend items À the modified function method. The temperature and vibration signals after removing the trend term are shown in Figures 10 and 11.
Denoising
Wavelet analysis is known as the microscope of signal processing. The key of wavelets analysis is the selection of wavelet basis and the decomposition level. Decomposition layer has great influence on the effect of denoising. The more is the decomposition layer, the lower is the noisesignal ratio. Meanwhile, when the layer increases, the processing becomes slow. Although few decomposition layer has high noise-signal ratio, the signal is decomposed to very small frequency bandwidths. Only the high frequency coefficients can be processed to remove the corresponding noise, while the corresponding low frequency noise is all reserved. Therefore, the choice of wavelet decomposition layer should be neither too large as consider-ing the improvement of the noise-signal ratio nor too small as considering the suppression of low frequency noise. The purpose of denoising is to get the useful feature signal, so wavelet coefficients can reflect the minimum frequency components in the useful signal. The wavelet decomposition is to decompose the signal into various independent bands, high detail coefficient reflects the low-frequency part of the signal. So this paper is based on the minimum frequency signal to determine the maximum level of wavelet decomposition. In this paper, sym8 wavelet is chosen as the wavelet base and soft thresholds are used to denoise. The temperature and vibration signal denoising processes are shown in Figures 12-17.
RUL prediction
We compare the prediction performance of proposed prediction method which based on D-S data fusion and SVR-PF with prediction method which uses single data source and other prediction methods. Roller sleeve 1 was used to predict in three cases, predicted by acceleration data, predicted by temperature data and predicted by fused data. From Figures 18-20, it's easy to know that the results predicted by the proposed prediction method are more accurate than other prediction methods.
Conclusion
In view of the engineering problem that is difficult to accurately predict the remaining useful life under partially observed state, a new method based on D-S data fusion and SVR-PF is proposed.
From Table 5, we can see that compared to other prediction methods such as obtained by the temperature or acceleration-based data-driven, the prediction accuracy of proposed method is significantly improved. Meanwhile, it provides a basis to make maintenance | 6,843.2 | 2019-09-01T00:00:00.000 | [
"Engineering",
"Computer Science"
] |
WHEAT EAR DETECTION IN PLOTS BY SEGMENTING MOBILE LASER SCANNER DATA
The use of Light Detection and Ranging (LiDAR) to study agricultural crop traits is becoming popular. Wheat plant traits such as crop height, biomass fractions and plant population are of interest to agronomists and biologists for the assessment of a genotype's performance in the environment. Among these performance indicators, plant population in the field is still widely estimated through manual counting which is a tedious and labour intensive task. The goal of this study is to explore the suitability of LiDAR observations to automate the counting process by the individual detection of wheat ears in the agricultural field. However, this is a challenging task owing to the random cropping pattern and noisy returns present in the point cloud. The goal is achieved by first segmenting the 3D point cloud followed by the classification of segments into ears and non-ears. In this study, two segmentation techniques: a) voxelbased segmentation and b) mean shift segmentation were adapted to suit the segmentation of plant point clouds. An ear classification strategy was developed to distinguish the ear segments from leaves and stems. Finally, the ears extracted by the automatic methods were compared with reference ear segments prepared by manual segmentation. Both the methods had an average detection rate of 85%, aggregated over different flowering stages. The voxel-based approach performed well for late flowering stages (wheat crops aged 210 days or more) with a mean percentage accuracy of 94% and takes less than 20 seconds to process 50,000 points with an average point density of 16 points/cm. Meanwhile, the mean shift approach showed comparatively better counting accuracy of 95% for early flowering stage (crops aged below 225 days) and takes approximately 4 minutes to process 50,000 points.
INTRODUCTION i. Background
LiDAR (Light Detection And Ranging) or laser scanning technology has been identified to hold high potential for meeting the demands of next generation phenotyping (Lin, 2015).This is attributed to the availability of LiDAR systems with small footprint and high pulse emission frequency and their capability to provide high-throughput plant traits.These systems provide robust data in varied illumination conditions and effective reconstruction of the in-field 3D crop architecture.Understandably, the use of laser scanners, both stationary and robot-mounted scanners, for field crop monitoring is becoming common for crop-height measurement and biomass estimation (Garrido et al., 2015;Hofle, 2014;Koenig et al., 2015).However, these applications have not yet exploited the full potential of LiDAR data i.e. 3D point clouds.It could be extended to plant population estimation.
Figure 1.A sample 3D point cloud acquired over a wheat plot.
The random spacing between plants, irregular orientation of the ears and noisy air returns make individual ear detection a challenge.
The number of wheat ears per unit area, which is an indication of the plant population, can be obtained by manual, semi-automated and automated techniques.Even though it is favorable to incorporate automatic counting of wheat ears, it is a challenging task due to the complex crop architecture with close plant spacing and high extent of overlap (LemnaTec, 2015) as can be seen in Figure 1.Moreover, the automatic technique should be able to handle the changing size and orientation of the wheat ears with the developmental stage of the plant as shown in Figure 2.
ii. Related Work
Few automated image processing techniques have been proposed to count wheat ears using 2D images from Charge-Coupled Device (CCD) cameras by applying texture-based classification in hybrid space (Cointault et al., 2008) or by high pass Fourier filtering (Journaux et al., 2010).The counting accuracy from 2D cameras is constrained by the illumination conditions at the time of image acquisition and undetected wheat ears due to overlap and obstruction by other plant organs.Deery et al. (2014) showcase the use of rasterized elevation images to count the number of wheat ears by applying a simple particle count algorithm on the segmented image.This approach does not perform well for fields with high crop density and overlapping ears.
The use of terrestrial laser scanners to estimate wheat crop density was demonstrated by Lumme et al. (2008).A detailed methodology and validation with reference dataset was not presented and the authors conclude that a laser scanner mounted on a mobile platform could be used as a tool in precision farming.Saeys et al. (2009) developed a method to estimate wheat crop density by 3D reconstruction of an artificial canopy set-up.They extracted the ear layer by fitting a "thin-plate smoothing spline" and generated a point density image from which the approximate location of the ears was identified.This method was found to be computationally intensive due to the fitting of the spline.Also, the method was demonstrated on an artificial canopy, where overlap among the adjacent plants is minimal.
Thus, manual counting is still widely used in practice.However, it is a tedious and labour intensive method and obviously cannot meet the needs of high-throughput phenotyping.Hence, 3D point clouds from LiDAR sensors could be used as an alternative for the automatic counting of wheat ears as they help to overcome the limitations of the existing methods owing to the availability of the depth information and reliability in all illumination conditions.This study presents two methods (segmentation followed by classification) developed to extract wheat ears from the LiDAR 3D point clouds.
Point cloud segmentation may be defined as the clustering of points based on their properties and spatial distribution to form homogenous regions.In most scenarios, segmentation is an integral step to recognize objects in the point cloud which determines the amount of useful information retrieved (Wang & Shan, 2009).Hence, the application and the objects present in the point cloud should determine the choice of segmentation technique.
There are two basic design mechanism to segment point clouds.
The first approach is based on methods that have mathematical model assumptions or geometric reasoning such as model fitting, probability density estimators or region growing.Though these methods achieve quick results for simple scenarios, they are sensitive to noise and exhibit poor performance in complex scenarios.The other approach is based on machine learning algorithms trained to classify the different object types in the scene (Nguyen & Le, 2013).A review of various segmentation techniques has been presented by Vosselman et al. (2004) and they are categorized based on the surface being extracted.For the extraction of smooth surfaces, they have suggested region growing, scan line segmentation and connected components in voxel space.
For the segmentation of wheat plant organs, literature reporting methods for the detection of individual trees from airborne laser scanner and crop detection were reviewed.Two methods that do not require labeled training samples and can identify irregular structures from dense point clouds were chosen as follows: • Voxel-based connected component: Voxel-based tree detection (Hosoi & Omasa, 2006) has been demonstrated for the estimation of forest parameters.Another popular application of voxelization of 3D point cloud is to extract the skeleton of trees (Bucksch et al., 2009) and plants (Ramamurthy et al., 2015).These works demonstrate that the voxel approach is adaptable to suit the object of interest.
•
Mean shift segmentation: The mean shift approach (Fukunaga & Hostetler, 1975) uses a non-parametric density estimator function to look for modes in the data.This was first used in computer vision by Comaniciu & Meer (2002) where they demonstrated the applicability of mean shift in image segmentation, to look for clusters in the feature space of images.Since then, this method has gained popularity and has also been demonstrated efficient on 3D point clouds.
Thus, these two segmentation methods were incorporated in the ear detection methods and their performances were compared in this paper.The following section describes the steps involved in the extraction of the wheat ears.The third section presents the datasets and design of the experiments used for the comparison of the two methods.Section 4 contains a short discussion on the results and the final section presents the conclusion and future works.
METHODOLOGY
The procedure to extract the wheat ears from the point cloud involves three main steps as follows: i. Filtering of noisy points ii.Segmentation of the point cloud iii.Classification of the segments
i. Filtering of noisy points
An initial filtering step was designed to remove the outliers placed either high above the plant canopy or too low below the ground level.The next step was designed to deal with the points floating around the objects in the scene that are "air returns" or "ghost returns" owing to the mixed edge effect.The mixed-edge effect occurs when a laser beam is intercepted by the edge of an object (Van Genchten et al., 2008).The beam splits and thus two signals reflected by two different objects are sent back.The receiver records an averaged distance between the two received signals and stores it as a point, which does not actually exist.In our case, the presence of "ghost returns" is expected due to the mixed-edge effect owing to the beam divergence varying with range and the densely occurring leaves, ears and soil particles.
These "ghost return" points surrounding the objects were removed in the first step of the segmentation process.This removal was carried out based on the neighboring point density, and was handled in different ways for the two short-listed segmentation methods.
ii. Segmentation
The following two segmentation techniques were adapted to identify and remove the air returns due to multi-edge effect.The ear-classification procedure that was developed as a part of this study was applied on the segmentation output.
A. Voxel-based connected component
In this approach, the point cloud is split into equal sized voxelcubes and the number of points within each voxel-cube is calculated.Based on the point density and the objects being Table 1 User-defined parameters used in voxel-based segmentation studied, only the voxels with number of points above a threshold value are considered for further processing.This approach is suitable to delineate individual plants in the canopy layer by removing overlapping points from adjacent plants and segment the ears by using a connected component analysis.The following steps were carried out in Matlab to segment the input point cloud as shown in Figure 3 and the user-defined parameters are explained in Table 1.
a. Voxel definition: A voxel coordinate system is defined for the point cloud with equal sided voxels and origin at (Xvox, Yvox, Zvox) = (0,0,0).The position of each point in the voxel space was calculated.b.Voxel-based thinning: For each voxel, the number of points inside the voxel is counted.For further processing, only the voxels which contain more than one point and which have at least one direct neighbor that contain more than a user-defined minimum number of points were selected.This thinning process ensures only voxels that belong to plant organs with high point density are selected for further processing.c.Connected components analysis: On the voxels selected from the thinning process, a connected components analysis was performed.This clusters points belonging to a single object and assign a unique segment number to it.
d. Second iteration of thinning and segmentation: In areas of high overlap, the objects appear connected and result in two or more ears being clustered together as a single segment.In order to address these cases of undersegmentation, a second iteration of point cloud thinning followed by connected components segmentation is carried out with half of the original voxel size thinning threshold.
Figure 3 Flow-chart depicting the steps involved in the voxelbased connected components segmentation.
B. Mean shift segmentation
Mean shift segmentation method, which is point based uses a probability density estimator function to search for modes in the data and cluster the points that fall within function for the kernel bandwidth.The only user-defined parameter is the kernel bandwidth that is defined based on the characteristics of the object being segmented (Melzer, 2007).The following steps as shown in Figure 4 were carried out: a. Filtering: The points hovering over the canopy and belonging to the stem region were removed based on point density.For each point, the number of neighbors within an imaginary 3D cylindrical neighborhood of (diameter = 3 cm, height = 6 cm) is calculated.Only points with more than 30 neighbors in this cylindrical neighborhood are used in further processing.
The dimension of the cylindrical neighborhood was decided based on a trial and error method which helped to identify the ideal dimensions and the range of neighboring density within which wheat ears can be approximately identified.An analysis of manually extracted ear segments showed that a wheat ear contains at least 30 points and thus, the threshold of 30 neighbors was used.b.Connected components: Next, a connected components analysis was performed on the remaining point cloud to identify under-segmented blobs of connected points.This rough connected component step based on proximity and neighborhood definitions reduces the processing time for mean shift segmentation.
Userdefined Parameter
Definition voxel-side • Length of the side of voxel cubes (1 cm was used in this study); • Should be fixed according to the average spacing between the plants in the field.
thinningthreshold
• Threshold for points removal for voxelbased thinning (threshold of 2 points for voxel-side = 1cm) • Should be fixed depending on the size of the voxels; a larger threshold should be used for bigger voxel sizes. smallsegmentthreshold • Threshold for removing small segments (the minimum number of points per segment to be considered as ear segment) The average number of points per ear was always found to be higher than 30.Hence, segments with less than 30 points could be considered as non-ear segments.However, a threshold of 15 was used so as to consider the thinning of the point cloud.
c. Segmentation: For each component, mean shift segmentation was carried out using the (X, Y, Z) coordinates with the help of existing software from ITC, Enschede.The (X, Y, Z) coordinates of the points were used as input parameters in the algorithm.The kernel bandwidths were selected as X = Y < Z following a prolate spheroid, to approximate the average size and shape of a wheat ear.
Table 2 User-defined parameters used in mean shift segmentation
iii. Classification of Segments
After segmenting the point cloud using either the mean shift method or the voxel-based method, we are left with unlabeled segments that could be ear, leaf or stem.Hence, the following strategy was developed to distinguish the ear segments and the parameters used are listed in Table 3.
a. Top-most segment selection: Since the wheat ear is present at the top of the canopy, the first step was to extract the topmost segments in the segment-space.These top segments were labeled as ears.For the segments below, the extent of overlap with top ear segment is calculated.If the overlap percentage is above a certain value, say 50%, then the segment is assumed to be directly below an ear segment and permanently labeled as a non-ear segment.
In some scenarios where the ears are bent and overlapping, as highlighted in the green box in Figure 5 b.Height Thresholding: After the above steps, some of the non-ear segments might still be wrongly labeled as ear (an example as shown in the orange box in Figure 5 (b)).On analyzing the height histogram (Figure 6) of the segments labeled as ears, two distinct peaks could be identified; one peak corresponding to the ear segments and the other one to the non-ear segments.This is because there are always more points describing the ear segments since they are present at the top of the canopy.Hence, in most cases, they are not overshadowed by other plant parts.
Hence, identifying the trough between the two peaks of the height histogram will aid in correct labelling of the ear segments.This height threshold is selected using Otsu's threshold selection method (Otsu, 1979) for histograms.The segments that lie below this height threshold but were classified as ear in the previous steps are reclassified as non-ear segments.No changes are made to the labels of the segments above this height threshold.The orange boxes in Figure 5 (b, c) highlight a mislabeled ear segment being corrected in the thresholding step.
i. Datasets
The laser scanning survey was conducted on three micro-plots, roughly of size 10 m x 2 m, in Gréoux, France, by Arvalis (A technical institute specialized in cereals) and INRA (National Agricultural Research Institute).Three plots sown with three different varieties of wheat on 29 October 2015 and subject to different irrigation treatments were selected for the survey.In order to have a time series, the survey was conducted on three different dates with the crops aged 194 days, 209 days and 225 days (May 10; May 25; June 10, 2016 respectively), each 15 days apart, on the same micro-plots.
The datasets were acquired with a SICK LMS 400 LiDAR (SICK Germany, 2007) sensor mounted on an unmanned ground vehicle (UGV), with nadir looking.This time of flight laser scanner has a scanning frequency of up to 500 Hz with an operating range of 0.7 m to 3 m and a systematic error of ± 4 mm.The scanner uses visible red light at 650 nm as its light source.Along with the LiDAR observations, wheat ear density estimates from manual counting in the field were also provided.To check the segmentation accuracy, reference wheat ear segments were extracted and labeled manually using CloudCompare.The datasets provided were acquired from a height of 2.1 m above the ground and had a point density of 16 pts/cm 2 .
ii. Design of the accuracy assessment
From the point clouds acquired on the three crop developmental stages, i.e. over crops of age 194, 209 and 225 days, a subset corresponding to the same area was taken from the three plots.From these 9 subsets, ear segments were manually segmented and labeled by 6 operators using CloudCompare.These manual segments were used to evaluate the correctness and completeness measures of the two methods (shown in Table 6).In order to check the variability between the operators, one common dataset (crops of age 225 days from Plot A) was provided to them.Table 4 shows the number of ears identified by different operators for the same subset.In addition to these manual labels, reference ear density for each plot was made available from manual counting in the field.This was used to calculate the Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) shown in Table 5.
iii.Ear detection results
a. Voxel-based ear detection
The thinning of point clouds in the voxel-based segmentation results in segments with much lower number of points as compared to the mean shift segmentation.From Figure 7 (b), it is seen that most of the noisy points surrounding the ears in the original point cloud have been removed.This results however in the removal of few ear points as well.This ensures that the ear segments are distinctly identified even in regions of overlap.In Figure 7 (c), we see the finally extracted ear segments in red, while the non-ear segments are displayed in blue.It is noticed that the voxel-based method does not retain the size of the original ear and hence if used to estimate the size of the wheat ear, it will lead to underestimation.Thus, the thinning threshold should be reconsidered if the size of the wheat ear is needed for the study.However, for plant population estimation the size of the ear segments is not a concern.
b. Mean shift ear detection
The segmentation results from the mean shift segmentation are representative of the raw point cloud i.e. the segments are not thinned and the size of the segments is retained.From Figure 8 (b), we notice that the segments retain the majority of the points from the raw point cloud and hence are mostly noisy.Thus, the mean shift based ear detection method can be directly extended to applications such as ear size estimation, without severe underestimation.
iv. Accuracy
The fitness of the two ear-detection methods was evaluated by comparing the number of ears observed by the automated methods with the reference dataset prepared by manual extraction.The following error metrics were used to evaluate the error associated with the detection of number of wheat ears per plot.
RMSE:
The ear detection method being evaluated might produce good estimates for a particular plot while returning poor estimates for another one.Therefore, RMSE was chosen to represent the collective error associated with the ear count over different plots and plant developmental stages as it is sensitive to outliers (Chai & Draxler, 2014).
MAPE:
This gives a measure of the overall performance of the ear detection algorithm by providing an averaged absolute precision in terms of percentage.The MAPE provides an estimate that is independent of the spatial extent of the plot.Hence, using the MAPE in combination with RMSE gives an overall understanding of the performance of the algorithm across different plots.
Table 5 Error in ear counts from the automatic detection methods calculated using the manual counting from the field as reference.In addition to the RMSE and MAPE, we used two more metrics, Correctness and Completeness, that have been used in literature (Y.Wang et al., 2016;Yao et al., 2014) to evaluate the performance of object detection algorithms.These were calculated using the manually extracted segments as reference.The voxel-based detection method had a comparatively higher Correctness value of 80.15% while compared to the 60.57% of the mean shift method as seen in Table 6.A similar trend is noticed for the Completeness value for which the voxel-based method has a 90.27% while the mean shift has only 83.33%.This indicates that the voxel-based approach is more successful than mean shift in identifying ear segments present in the scene with a relatively lower number of false ear segments.
v. Analysis of Detection Results
The following three criteria were used to understand the robustness of the two ear detection methods developed in this study.The accuracies stated were calculated using the manual counts from the field.
• Crop developmental stage: The size and orientation of the wheat ear changes throughout its flowering stage (refer to Figure 2).Thus, using the datasets available for three different developmental stages of the same plot, the effect of crop growth on the ear detection rate was evaluated.The voxel-based method performs better for the crops aged 209 days and more with an average percentage error of 5.2% and 8.03%.Whereas, the mean shift-based approach shows relatively better results for the crops of age 209 days and more, with an average percentage error of 3.8% and 6.68% calculated by using the manual counts from the field as reference.Thus, it is concluded that the crop developmental stage influences the performance of the ear detection methods.If the age of the plants is known, then the input parameters should be set accordingly for optimal performance.
•
Wheat variety and irrigation treatment: The variety of wheat and water stress are found to influence the size of the wheat grains (Sionit, Hellmers, & Strain, 1980).Hence, the ear detection methods developed were tested on three different varieties of wheat subject to different irrigation treatments.This helped to understand how the methods perform for plots with varying ear density.From this analysis, it was concluded that adaptive selection of segmentation parameters is advisable.
•
Effect of point density: Point density is one of the important factors that influence the choice of input parameter values and ear detection rate.Hence, to understand how the ear detection methods perform for different point densities three degraded versions of the datasets was created by retaining 75%, 50% and 25% of the original points in a random manner.It was found that the ear detection rates for both the methods were consistent for up to 75% of the original point density.However, modifying the parameters according to the point density improves the performance of the ear detection methods.
DISCUSSION
After aggregation of the MAPE values in Table 5 over different plots and developmental stages, it was observed that the voxelbased ear detection method had a lower value of 14.90% whereas the mean shift-based ear detection had a relatively higher value of 41.00%.It should be noted that before aggregation over different developmental stages, the voxel-based method had an average MAPE of 6% for crops aged 209 days and above.Whereas, the mean shift based method had an average MAPE of a very low 5% for crops of age 209 days and below.
This detection rate is an improvement when compared to the 0.9 correlation with 40 training points reported for image-based ear detection by Cointault et al. (2008).In a different study, Saeys et al. (2009) reported a 94% ear detection rate while using LiDAR observations with optimal acquisition settings.However, this was demonstrated on artificially constructed canopy and the proposed method was computationally very expensive.
CONCLUSIONS AND FUTURE WORK
In this study, we designed, analyzed and evaluated two approaches for the detection of wheat ears from laser scanned 3D point clouds.The task of ear detection was split into two stepssegmentation and classification of segments.For segmentation of the point cloud, voxel-based segmentation and mean shift segmentation methods were chosen and implemented.For classification of the segments, an ear classification methodology was developed taking into consideration the position of wheat ears in the canopy, overlap between plants and presence of stunted plants within the canopy.
The performance of these two ear detection methods was then assessed based on developmental stage of the crop, wheat variety and point density.
•
The voxel-based approach performs well for late developmental stages with a comparatively very short computational time.
•
Meanwhile, the mean shift method performs well for different point densities and gives reliable results for early developmental stages and has a higher processing time.
Since both the ear-detection approaches are in some way influenced by the point density of the dataset, the input detection parameters need to be adapted to achieve good count results.
Another interesting observation was the variation in the number of ears extracted by different operators by manual segmentation which focuses on the fact that ear detection is a difficult task even for a human operator.
In the following works, the use of adaptive kernel bandwidth and adaptive voxel sizes for segmentation of the point cloud should be investigated.Experiments should be done to utilize RGB images to improve the ear classification performance.Depending on the available prior knowledge regarding the crops studied and accuracy required for the application, one of the two ear detection approaches may be selected.Thus, through this study, two ear detection techniques that depend solely on the 3D coordinates of the points describing wheat ears have been developed and compared.
Figure 4
Figure 4 Work-flow involved in mean shift segmentation illustrated with a sample point cloud.
(a), the ear segments from the shorter plant initially tend to get mislabeled as non-ears.However, reconsidering these segments based on the extent of their overlap with the topmost segments ensures that they are correctly relabeled as ear segments as shown in the green box in Figure5(b).
Figure 5
Figure 5 Example to illustrate the steps involved in ear classification.Ear segments are displayed in red while the non-ear segments are displayed in blue (a) Ear segments identified by searching for the top-most segment in segment space.The green box highlights a case of mislabeling due to overlap.(b) Relabeling nonear segments depending on extent of overlap with the segments above.The orange box focuses on a case of mislabeled non-ear segments.(c) Final segment labels assigned based on height thresholding selected by Otsu's method which identifies the trough between two distinct peaks in a histogram.
Figure 6
Figure 6 Height histogram showing the two distinct peaks between the ear and non-ear segments.The line at 0.727 m denotes the threshold chosen from Otsu's thresholding.
Figure 7
Figure 7 Voxel-based ear detection results.The size of the ears is reduced due to point cloud thinning (a) Raw point cloud, color coded based on height value.(b) Voxelbased segmentation results with each segment displayed using a random color (c) Ear classification where the red segments are the extracted ears (crops of age 225 days).
Figure 8
Figure 8 Mean shift-based ear detection results.The size of the ears is retained (a) Raw point cloud, color coded based on height value (b) Mean shift-based segmentation results with each segment displayed with a random color (c) Ear classification where the red segments are the extracted ears (crops of age 225 days).
= Number of detected by the automatic method = Number of ears identified by manual counting N = Number of datasets used in the evaluation
Table 3
Parameters used for the ear classification
Table 4
Variability among different operators in manually extracting the ear segments.The subjectiveness of manual segmentation and difficulty to distinctly identify ear segments for humans is observed.
Table 6
Error in ear counts from the automatic detection methods calculated using the manually extracted segments as reference (aggregated over different plots and developmental stages). | 6,626.2 | 2017-09-13T00:00:00.000 | [
"Agricultural and Food Sciences",
"Engineering"
] |
Measuring Young’s modulus with a tensile tester
We use a tensile testing machine to create stress-strain plots and determine Young’s modulus for some ductile and plastic materials. Supplementary videos are also provided.
Young's modulus is a key concept in elasticity theory and mechanics of solids [1,2]. Measurements of Young's modulus in a classroom setting have been studied many times in the literature. These studies typically focus on simple or inexpensive ways of measuring Young's modulus [3,4]. Related studies describe alternative indirect ways of measuring the modulus [5,6]. Nunn, for example, developed a technique for classroom use by which one can determine the Young's modulus of a solid using the speed of sound in that material [7]. In this article, we will perform accurate measurements with a tensile testing machine to create stress-strain plots and determine Young's modulus for several materials. We also discuss some possible limitations of our method of measurement.
Young's modulus is defined to be the ratio: Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
where σ is the tensile strength (defined as force applied per unit area), ϵ is the tensile strain (defined as displacement per unit length), F is the applied lengthwise force, A is the cross sectional area of the sample, L is the total displacement, and l is the test length which was initially between the two grippers in the device. This ratio defines how a sample of material deforms in a response to a force which is applied lengthwise. The stress analysis device which we use to measure this modulus in our demonstration is the Tinius Olsen 25ST with a 25 KN load cell. We will mention here that the device and analysis software which we use are somewhat sophisticated and will likely only be available at a metrology laboratory at a university. High school students may be able to contact a technician at an Engineering department at a local university and ask if they can view a demonstration. The novelty of our result is to present complete stress-strain plots which are hard to obtain in a classroom environment with limited equipment. The technical method which we have used in carrying out our experiments can be compared with measurements which students can obtain using alternative means. We also provide several videos of the tests which may be useful to students. Note that we have chosen everyday materials for our tests to aid with comparisons in a classroom environment.
A pedagogical point here is that different materials can have very different relationships between stress and strain. For example, students might intuitively think that a plastic straw would not have much resistance to stress because the straw is easy to cut or snap. However, in figure 1, we show a plastic straw (similar to a drinking straw) being used in our apparatus, along with the raw force-displacement plot. This plot can be converted to a stress-strain plot by dividing the force by the cross sectional area and the displacement by the original test length. This is done in figure 2 (left-hand image) to produce a stress-strain plot for the plastic straw. We can see from the plot that the plastic straw has an initial elastic region where the relationship between the stress and the strain is close to linear.
In the mathematical modelling of elasticity phenomena, this is known as linear elasticity. Since E is defined as the ratio of tensile strength to tensile strain, students should know that Young's modulus can be found by taking the gradient of the curve in figure 2. This only holds when the graph is linear because Young's modulus is strictly speaking not valid outside of the linear elastic region. This turning point where the graph stops being linear is called the elastic limit and at this point, the material becomes stretched and deformed far enough that non-linear effects are introduced and the material cannot return to its original length [2]. This can be checked afterwards by seeing that the top of the straw has been stretched into a position from which it cannot recover.
In figure 2 (right-hand image), we zoom in on the region of the stress-strain plot which is approximately a straight line. Young's modulus can now be found from the gradient of this line. In this case, E = 1.75 × 10 5 Pa, or 1.75 GPa. The type of irreversible deformation which occurs beyond the elastic limit due to dislocations of the material on the atomic scale is typically known as plastic deformation in the literature. This is to be contrasted with an elastic deformation, which is reversible and does not cause a permanent change to the structure of the material.
In figure 3, we figure 3, we repeat the experiment with flexible plastic wire cord. Students might guess that since the material is flexible, it will have a lot of resistance to tensile strain and that it will stretch out by a large displacement before suddenly breaking. This approximate behaviour is indeed observed in figure 3. The applied force is lower than the force for figure 1, but the corresponding displacement (and hence resistance to strain) is lower. The sudden drop corresponds to the point at which the cord breaks so that a force is no longer being applied. For completeness, we plot the stress-strain plot for the plastic cord in figure 4. In this case, the tangent of the linear region tells us that Young's modulus can be estimated at around 1.25 GPa.
We will now attempt to test a material which is neither ductile nor plastic. In figure 5, we show a stress test with a piece of nylon rope. It might intuitively be guessed that a piece of rope could undergo a huge lengthwise applied force but that it will not stretch very much and will break at a very high tensile strain. This type of behaviour is typical of materials which are strong but not ductile. In figure 5, we show that this greater stress resistance is indeed observed when the rope sample is placed in the tensile tester. The sudden drop in the graph occurs because the grip used in a tensile tester device causes the rope to fray as it moves upwards such that rope strands start to be worn away by shearing forces. This then causes the rope to slip and the force drops off before the full forcedisplacement plot can be obtained. In figure 5, we attempted to avoid this problem by tying knots at the end points of the rope to allow for more grip but slippage still occurs. This is a limitation of our method, but nevertheless one can see that the behaviour which occurs for the rope is somewhere between that of a brittle and a ductile material. The rope is clearly not ductile, since as stated earlier, ductile materials like steel typically yield under an applied force, before hardening and then finally breaking (see figure 1 for an example).
We have also included videos in the supplementary material so that students can see the tests being performed on the apparatus. In video 1, we use the rope sample and demonstrate how the rope sample is pulled until it is tense. As the name of the device suggests, it is necessary to have this tension before we can perform any measurements. In video 2 (sped up 4×), we show the stress test being performed on the plastic cord. Note how after some time the cord starts to turn white (corresponding to hardening and passing beyond the elastic region) before the cord finally breaks. In video 3 (also sped up 4×), we show the stress test for the plastic straw. It can be clearly seen the straw is stretched by a substantial amount March 2022 and that it also turns white as it is deformed, starting to form a neckpinch in the middle under tension.
Data availability statement
No new data were created or analysed in this study.
Hollis Williams is an Engineering
PhD student at the University of Warwick. He is interested in various aspects of physics education and theoretical physics and has published articles on fluid dynamics, quantum mechanics and particle physics. | 1,950.2 | 2022-01-14T00:00:00.000 | [
"Physics"
] |
A Hybrid Path-Oriented Code Assignment CDMA-Based MAC Protocol for Underwater Acoustic Sensor Networks
Due to the characteristics of underwater acoustic channel, media access control (MAC) protocols designed for underwater acoustic sensor networks (UWASNs) are quite different from those for terrestrial wireless sensor networks. Moreover, in a sink-oriented network with event information generation in a sensor field and message forwarding to the sink hop-by-hop, the sensors near the sink have to transmit more packets than those far from the sink, and then a funneling effect occurs, which leads to packet congestion, collisions and losses, especially in UWASNs with long propagation delays. An improved CDMA-based MAC protocol, named path-oriented code assignment (POCA) CDMA MAC (POCA-CDMA-MAC), is proposed for UWASNs in this paper. In the proposed MAC protocol, both the round-robin method and CDMA technology are adopted to make the sink receive packets from multiple paths simultaneously. Since the number of paths for information gathering is much less than that of nodes, the length of the spreading code used in the POCA-CDMA-MAC protocol is shorter greatly than that used in the CDMA-based protocols with transmitter-oriented code assignment (TOCA) or receiver-oriented code assignment (ROCA). Simulation results show that the proposed POCA-CDMA-MAC protocol achieves a higher network throughput and a lower end-to-end delay compared to other CDMA-based MAC protocols.
with the three-way handshake mechanism. According to the long propagation delay, DACAP makes a sender wait for a mandatory waiting time before it sends the data packet after receiving CTS; in addition, it allows the destination to send a warning packet to the source to cancel the transmission if it receives an RTS packet from another node. An adaptive propagation delay tolerant collision avoidance protocol (APCAP) is proposed in [11]. In this protocol, a sender is allowed to respond to other senders while waiting for the CTS packet from an intended receiver, and this modified mechanism improves the network throughput. In order to alleviate the funneling effect [12] happening in a localized and sink-oriented network, a funneling MAC for UWASNs (FMAC-U) is proposed in [13]. With the three-way handshake mechanism, this protocol makes the sink receive data packets from multiple neighboring nodes in a fixed order during each round of handshakes.
COPE-MAC [14] is another protocol based on a three-way handshake mechanism. It uses parallel reservation and carrier sensing methods to avoid the packet collision when a node has received more than one request to send data. Focusing on the case in which two nodes can transmit to each other at around the same time without collision, a bidirectional-concurrent MAC protocol (BiC-MAC) is proposed in [15]. Since the sender-receiver pair is allowed to transmit data packets to each other in a successful handshake, the BiC-MAC protocol greatly improves the data transmission efficiency and channel utilization.
The triple hidden terminal problem in single-transceiver multi-channel long propagation delay underwater networks, the multi-hop, multi-channel and long-delay hidden terminal problem are defined in [16]. Then, a cooperative underwater multi-channel MAC protocol based on a three-way handshake mechanism and a cooperative collision detection mechanism is proposed to solve the triple hidden terminal problem.
Slotted FAMA [17] is a contention-based protocol based on floor acquisition multiple accesses (FAMA) [18] for UWASNs. In this protocol, all nodes share the common slot synchronization, and initiate the RTS-CTS handshake at the beginning of a slot. Compared to TDMA, Slotted FAMA has no idle slots.
On the other hand, receiver-reservation-based MAC protocols have been investigated to avoid the hidden terminal problem. In [19], authors proposed the receiver-initiated packet train (RIPT) protocol for multi-hop UWASNs. In the RIPT protocol, an intended receiver invites senders to transmit the data packets, and coordinates data packets from multiple neighboring nodes in a packet chain manner.
In order to take advantage of the low delay of contention-based MAC protocols and the high throughput of schedule-based MAC protocols, designing a hybrid MAC protocol has also been investigated for UWASNs. CDMA is the most promising technique used in a hybrid MAC protocol since it is robust to frequency-selective fading and it can easily compensate for the effect of multi-path transmission at the receiver. In [20], authors proposed a distributed protocol for long latency access networks (PLAN), in which a node uses a unique spreading code to encode its signals (such as RTS, CTS and DATA) before transmitting. Then, the intended receivers broadcast a CTS packet for several accumulated RTS packets and receive data packets from multiple senders at the same time. In [21], by combining ALOHA and CDMA, a transmitter-based CDMA MAC protocol is proposed for UWASNs. Since a closed-loop distributed algorithm is used to set the optimal transmit power and the code length to minimize the near-far effect [22], the protocol achieves a low channel access delay, low energy consumption and high network throughput. Inspired by the theory of compressed sensing, a distributed energy-efficient sensor network scheme, random access compressed sensing (RACS), is proposed in [23]. This protocol is suitable for long-term deployment underwater networks in which energy saving is of crucial importance. It also prolongs network lifetime since a simple and distributed scheme is used to eliminate the need of scheduling. In [24], a hybrid spatial reuse TDMA (HSR-TDMA) protocol based on time division technology and direct sequence spread spectrum technology is proposed for broadcasting UWASNs. This protocol improves the network performance since the near-far problem is resolved.
The MAC protocols mentioned above are proposed to reduce the packet collision or deal with the long propagation delay problems in UWASNs, and the network throughput is improved to some extent with a random topology. However, many applications of UWASNs have some special characteristics which should be considered in designing the MAC protocol. For example, in a UWASN for oceanic environment monitoring and oceanic data collection, the network generally consists of a sink and many sensor nodes deployed surrounding the sink. That is, the network topology is sink-oriented, in which sink is the destination of the information generated at sensor nodes all over the network. When data packets are transmitted hop-by-hop from the sensor nodes to the sink, the funneling effect occurs [12], but the existing MAC protocols mentioned above fail to deal with the funneling effect in UWASNs. In [13], FMAC-U is proposed to improve three-way handshake mechanism between the sink and one hop neighboring nodes in order to alleviate the funneling effect. However, this protocol did not consider the funneling effect problem in multi-hop networks.
In order to resolve the funneling effect in multi-hop UWASNs, an improved CDMA-based MAC protocol, named path-oriented code assignment (POCA)-CDMA-MAC protocol, is proposed in this paper. In the proposed protocol, sensor nodes perform a round-robin scheduling in the same route and CDMA technology for different routes. That is, the nodes in the same path are assigned the same spreading sequence, and then they transmit data packets pre-processed with the spreading sequence in a round-robin method. With POCA-CDMA-MAC protocol, a sink can receive data packets from multiple paths simultaneously, and the packet collision, therefore, is reduced. Since the number of paths is much less than that of nodes, the length of the spreading sequences required in POCA-CDMA-MAC protocol is shortened. Hence, the proposed POCA-CDMA-MAC protocol performs well, in term of a high network throughput, a low packet loss rate and a small end-to-end delay.
The remainder of the paper is organized as follows: in Section 2, we introduce the system model and formulate the problem we aim to resolve. In Section 3, the POCA-CDMA-MAC protocol for UWASNs is presented. In Section 4, the proposed MAC protocol is evaluated by simulations. In Section 5 further insights into the proposed MAC protocol are provided. Section 6 gives the conclusions and suggestions for future work.
System Model and Problem Formulation
A typical architecture of a UWASN is shown in Figure 1. A sink is located at the center of the monitored area, many sensor nodes are deployed surrounding the sink, and a surface station is deployed to act as a gateway between the on-shore control center and the sink. The information gathered by sensor nodes in the monitored field transmits to the sink hop-by-hop. The data packets at the sink are forwarded to the surface station via cable. Finally, the data packets are transmitted to the on-shore control center by radio signals [25,26].
In UWASNs, the sink is assumed to be a sufficient energy supply and capable of handling parallel communications with senor nodes, all sensor nodes are homogenous and quasi-stationary. In Figure 1, since the data packets generated at sensor nodes are transmitted to the sink hop-by-hop in a many-to-one pattern, the funneling effect happens, as shown in Figure 2. From the Figure 2, we observe that the number of sensor nodes near the sink is much less than that of sensor nodes far from the sink. Therefore, the sensor nodes near the sink need to transmit more data packets than those far from the sink and the quantity of data transmitted by the neighboring nodes of the sink is the most, and the area around the sink becomes the choke point for the whole network. As shown in Figure 3, the sensor nodes are randomly deployed in a circular area. When the distance from the sink to the network edge is larger than the maximum transmission range of sensor nodes, the sensor nodes near the sink should act as the intermediate nodes to forward the data from sensor nodes located far away from the sink.
The average data packets transmitted by each sensor node in the i th circle, including the data generated by itself and the forwarded data, are: . From Figure 4, we observe that the data transmitted by the sensor nodes near the sink are much more that of the sensor nodes far from the sink. The reason for this phenomenon is that the sensor nodes near the sink need to forward the data from the sensor nodes far from the sink.
The fairness index of sensor nodes transmitting data packets, the Jain's fairness index, is defined as: Figure 5 shows the fairness of sensor nodes in each circle transmitting data with different network radius. From Figure 5, we observed that when the radius of the network equals the maximum transmission range of the sensor nodes, data packets transmitted by sensor nodes are almost the same, and then the fairness is 1. This is the case of a single-hop network, as sensor nodes only need to transmit their own data to the sink. When the radius of the network increases, the fairness decreases. The reason for this phenomenon is that the data packets forwarded by sensor nodes near the sink increases as the network radius increases. When M = 9, the fairness is 0.308, the worst.
As analyzed above, in UWASNs the traffic intensity increases as data packets move more closely toward the sink. Hence, the funneling effect leads to not only the increase of packet collision, network congestion and packet loss, but also the increase of the energy consumption of nodes near the sink. Because of the long propagation delay and low available bandwidth, it is necessary to alleviate the funneling effect in designing the MAC protocol for UWASNs. Although the throughput is improved by mitigating the hidden terminal problem and reducing the packet collision to some extent, existing MAC protocols for UWASNs do not consider the funneling effect in a multi-hop UWASN. The RTS-CTS handshaking and the CDMA-based method in PLAN make receivers concurrently receive packets from multiple sources, which is a good choice for alleviating the funneling effect. However, since the protocol considered that each node is allocated a quite long spreading sequence using distributed TOCA, its efficiency is low. The T-Lohi and the Ordered CSMA protocols without the handshake mechanism work well in single-hop UWASNs, but they cannot achieve good performance in multi-hop networks. Furthermore, the funneling-MAC is proposed to resolve the funneling effect problem with a hybrid TDMA/CSMA approach in terrestrial wireless sensor networks. However, due to the long propagation delay in UWASNs, a slot time in TDMA protocol, the sum of the transmission delay of data packets and the maximum propagation delay of the network, is also quite long, and the performance of networks, therefore, is reduced.
Proposed POCA-CDMA-MAC Protocol
In this section, an improved CDMA-based protocol for UWASNs, named POCA-CDMA-MAC, is presented. A round-robin method and CDMA technology are utilized to reduce the packet collisions. It allows the sink to receive packets from multiple neighbors belonging different paths at the same time. It also works well in a multi-hop network with the hidden/exposed terminal problems.
Overview of POCA-CDMA-MAC Protocol
In the proposed POCA-CDMA-MAC protocol, the sink determines the routes and spreading sequences for nodes. When sensor nodes are deployed, routes will be built by the sink in the initialization phase, the detail of the determination of the routes is shown in Section 3.2. Each node is ordered with its position in the path. That is, the first node in the path is the first sender, and the next node in the path is the second sender, and so on. The sink is the destination of all paths. Each node follows the order to transmit in a round-robin manner. When a node except the first sender wants to transmit a data packet, it has to wait for data packets from its previous node. The start transmitting time of the first node at each path is determined by the sink, which is described in the Section 3.3.
The packet transmission via an established path is demonstrated in Figure 6. At the start transmitting time, t 0 , the first sender, node D, sends its data packet (packet 1) to its next-hop node, node C. Receiving packet 1 from node D, node C first forwards packet 1, and then transmits its own packet (packet 2). After transmitted packet 1, node D does not delete the packet until it receives the forwarded packet from node C. When node B received and forwarded packets from node C (packets 1 and 2), it only sends a token (packet 3) since it has no generated data. The token is used to trigger next transmission. Receiving packets from node B, node A forwards packets (packets 1-3) and transmits its own packet (packet 4) to the sink. Then, node D starts another transmission at t 0 + T.
Path-Oriented Code Assignment (POCA)
When several packets from one-hop neighbors arrive at the sink at the same time, packet collision happens. To avoid the collision, each one-hop neighbor of the sink pre-processes packets with a spreading sequence. Unlike the TOCA and the ROCA, the sink uses the path-oriented code assignment (POCA) mechanism. Nodes in the same path are assigned with a same spreading sequence, whereas nodes in different paths are assigned with different spreading sequences. In the initialization phase, the sink assigns a pseudo-random binary spreading sequence to each established path in the network. It is assumed that the spreading sequence is orthogonal or quasi-orthogonal. In general, since the number of the established paths is much less than the number of nodes, the length of the spreading code used in POCA is much shorter than that of TOCA or ROCA.
As shown in Figure 7, there are a sink and some sensor nodes belonging to seven paths around the sink in the network. There are two nodes, A P1 and B P1, in path 1. There are three nodes, A P2 , B P2 and C P2, in path 2. There are two nodes, A P2 and B P3, in path 3. There are two nodes, A P2 and B P4, in path 4. There are three nodes in paths 5-7. A P2 is shared by three paths, paths 2-4. A P6 and B P6 are shared by two paths, paths 6 and 7. In the initial phase, the sink generates and sends spreading sequences to its neighbor nodes according to the number of established paths in the network. When A P1 received its spreading sequence (c 1 ) from the sink, it transfers c 1 to B P1 . The sink sends three spreading sequences (c 2 , c 3 , c 4 ) 3 Passing token to A P2 , and then A P2 assigns these three spreading sequences to B P2 , B P3 , and B P4 . When B P2 has received c 2 , it transfers c 2 to C P2 . Other nodes in the paths 5-7 are assigned the spreading sequences in the same way. Finally, seven spreading sequences are assigned to seven paths separately. When B P1 has packets to send, it pre-processes its packets with c 1 and transmits the processed packets to A P1 . When A P1 received the packets from B P1 , it forwards them to the sink, and pre-processes its own packets with c 1 , and transmits them to the sink. When A P2 received the packets from B P2 , B P3 and B P4 , it identifies them with different spreading sequences and forwards them to the sink. Then it pre-processes its own packets with c 2 and transmits them to the sink.
As mentioned above, there are two cases about the transmission time of each node in a given path. In one case, for a node, such as node A P1 as shown in Figure 6, there is only one previous node. In this case, A P1 forwards the packets received from its previous node, B P1 , and pre-processes its own packets with its spreading sequence, c 1 , and transmits them. In another case, for a node, such as A P2 as shown in Figure 6, there is more than one previous node, three previous nodes, nodes B P2 , B P3 and B P4 , for A P2 . In this case, A P2 forwards the data packets from its three previous nodes after received and identified these packets with their spreading sequences. Then, A P2 pre-processes its own packets with its own spreading sequence, c 2 , and transmits them.
When the sink has received the data packets from different paths with different spreading sequences, it decodes the packets with corresponding spreading sequences. For example, B P1 pre-processes packet In a real UWASN, there are some nodes except the sink are shared by more than one path, and the number of the path is more than the number of one-hop neighbors of the sink. Hence, the spreading sequences can be also reused in two paths when the distance between two paths is more than two times of the maximum transmission range of nodes. From Figure 7, there are seven paths in the network, but only four one-hop neighbors of the sink. Hence, code c 7 can be the same as one of the codes c 1 -c 4 .
Timing of the First Node Sending Packets
The timing of the first transmission node in each path is an important issue. If the first node starts transmitting at the proper time, nodes may share the channel well, and this favors reducing the packet collisions and improving the network throughput.
The time for the first node to send packets depends on the traffic load of the network, the deployment of nodes and the hops of the path, and so on. For simplicity, the first node sends its packets periodically, and the period interval of each path is assigned by the sink according to the hops of the path, the maximum data packets transmitted by a node each time, and the maximum propagation delay between two neighboring nodes. A long periodic interval may be set for a path with many nodes and high traffic load, while a short periodic interval may be set for a path with few nodes and low traffic load.
In the POCA-CDMA-MAC protocol with a round-robin method, the first node at a path starts the transmission during a periodic interval, and the sink has to wait for long time to receive DATA packets from its neighbors. In a path with more than two hops from the first node to the sink, the first node can start the transmission when the sink is receiving DATA packets from its neighbors. As shown in Figure 6, node A can start another transmission at t 2 , and t 2 < t 0 + T. Hence, the sink waits for a short time to receive DATA packets, which increases the network throughput.
Analysis of the Maximum Network Throughput
The throughput is an essential metric to evaluate the performance of MAC protocols. In sink-oriented networks, the throughput is defined as the ratio of the total received packets by the sink to the packets it can receive during a given time [19]. When the spreading sequence technology is not used in the system, the sink can only receive packets from a neighboring node. For simplicity, a network with only one path is considered, and the network throughput is the same as the throughput of the path. Hence, the network throughput can be written as: where, S is the network throughput, L is the length of each DATA packet, r is the data rate of the system. And n DATA,0 is the number of DATA packets received by the sink during T. From Figure 6, there are a sink and four nodes at the path. When each node transmits its DATA packets on its order, and the first node transmits its DATA packets within T, the maximum throughput of network with only one path is: where, S max is the maximum network throughput. n DATA,1 is the number of DATA packets transmitted by a one-hop neighbor (node A) each time, including forwarded DATA packets. n DATA,i is the number of DATA packets transmitted by the ith hop neighbor each time, Hop is the number of hops at a path, and Hop = 4 in Figure 6. And D p,i is the propagation delay of packets transmitted from the ith hop neighbor to the (i-1)th hop neighbor. In schedule-based networks without packet collision, the sink can receive all of the DATA packets transmitted by its one-hop neighbor (node A), then n DATA,0 = n DATA,1 .
Since the CDMA technology is used in the system, the data packets are pre-processed with the spreading sequence before transmission, and the sink can receive from multiple neighboring nodes simultaneously. On the other hand, the transmission delay of the packets is expanded. Hence, the maximum throughput in a path can be re-written as: (8) where, L c is the length of the spreading sequence, , the factor of the CDMA technology, is relevant to the parameters of the technology, such as the signal-to-interference-plus-noise ratio (SINR), the spreading gain, the power of the receiving signal and the length of the spreading sequence. Since the paths are set up and the spreading sequences are assigned statically by the sink in the initialization phase, each node has relative static neighboring nodes. The interference signals from other senders can be neglected if the sink assigns a long spreading sequence to nodes.
When many paths are set up around the sink, the nodes at each path will transmit the DATA packets pre-processed by different spreading sequences. Then, the maximum network throughput is: where S max,k is the maximum throughput of the k th path, and can be obtained in Equation (8), N path is the number of the paths in the network.
In a sink-oriented UWASN, the sink can build the paths with the same hops. When each node at a path transmits the same number of data packets, the number of one-hop neighbors, n DATA,1 , of these paths is the same, the maximum network throughput, therefore, can be achieved. (10) When there are eight one-hop neighbors in the network, the data rate is 1 kbps, and the length of the data packets is 4,000 bits, the propagation delay is 4 s. It is also assumed that each node in a path transmits packets with the same size. The impact of the length of the spreading sequences and the hops of the path on the maximum network throughput of the proposed protocol is shown in Figure 8. decreases fast, and then the maximum network throughput decreases. From Figure 8b, we observe that the maximum network throughput decreases quickly when the number of hops at a path increases. The reason is that the time of the channel used by one-hop neighbors decreases when the number of hops increases, and then the efficiency of the channel and the maximum network throughput decrease. Therefore, two useful conclusions can be drawn. First, the network throughput increases when the number of DATA packets transmitted by one-hop neighbors during T increases, and it increases along with the ratio of the number of DATA packets transmitted by one-hop neighbors of the sink to the number of DATA packets transmitted by other nodes. Second, the performance of the network throughput is improved by alleviating the effect of the long propagation.
Simulation Results
To evaluate the performance of our proposed POCA-CDMA-MAC protocol, simulations were performed under different traffic loads. The performance of the proposed MAC protocol, in terms of the network throughput, the packet drop rate and the end-to-end delay, is compared to two MAC protocols for UWASNs, the Slotted FAMA in [17] and the RIPT protocol in [19]. The proposed MAC protocol with different spreading sequences for each path and the proposed MAC protocol with spreading sequences reuse for different paths, are labeled as -proposed MAC without spatial reuse‖ and -proposed MAC with spatial reuse‖.
The packet drop rate is the ratio of the number of packets dropped by all nodes to the number of packets generated by them. The end-to-end delay is the average time of the packets received from generation to reception. That is: where, n d is the number of packets dropped by all sensor nodes, n G is the number of packets generated by all sensor nodes. t R,k is the time of the kth packet received by the sink, t G,k is the time of the kth packet generated by one of sensor nodes. The network topology used for simulations is shown in Figure 9. There are 24 sensor nodes deployed in a grid topology, and a sink is located at the center. Instead of precisely placing each node at the grid intersection point, some randomness is introduced by allowing each node to deviate from the grid intersection point within 10% of the grid spacing in both x and y directions. In the network, the grid spacing is 5 km, and the transmission range is set to be 1.75 times of the grid spacing so that each sensor node has exactly eight neighbors in its transmission range.
All sensor nodes are equipped with a half-duplex, omni-directional transceiver with a data rate of 1 kbps. The length of control packet is 12.5 bytes. The length of data packet is 500 bytes in the simulations in Figures 10-12. The acoustic propagation speed is 1,500 mps. In the RIPT, The average time interval between initiating ready-to-receive (RTRs) at a node (T avg in [18]) is 100 s, the number of DATA slots reserved at a receiver (M train in [18]) is initialized to 1 and the maximum allowable value for M train (M train,max in [18]) is 50. In the Slotted FAMA, the M train,max is 2 considering that there is only one sender to transmit DATA packet through a handshake. The size of time slot is 4.1 s. In the proposed POCA-CDMA-MAC protocol, the period of the first node starting transmission is the sum of the propagation time and the time for a node to transmit its DATA packet. The length of the spreading sequences for proposed MAC without spatial reuse is 64 considering that there are 16 paths in the networks. The length of the spreading sequences for proposed MAC with spatial reuse is 32 considering that there are only eight one-hop neighbors around the sink. The value of M train,max is 2. Figure 10, we observe that the throughput of Slotted FAMA and RIPT MAC protocols increases along with the increase of the traffic load when the traffic load is small, and it decreases slightly along with the increase of the traffic load after reaching peak values. However, the throughput of proposed POCA-CDMA-MAC protocol increases along with the increase of the traffic load when the traffic load is small, and it becomes a relatively stable value when the traffic load increases. The maximum network throughput of Slotted FAMA, the proposed protocol without spatial reuse, RIPT, and the proposed protocol with spatial reuse is 0.004, 0.006, 0.010, and 0.0142 when the traffic load is 0.005, 0.007, 0.013 and 0.018 packet/s, respectively. The achieved maximum network throughput in the network is only about 0.0142, partly because of the long propagation delay and the collision of the data from multiple senders.
From Figure 10, we also observe that the maximum network throughput of the proposed protocol with spatial reuse is the highest among all the protocols.
The reason for this phenomenon is that the packets are transmitted from several paths to the sink without handshaking, while the packets are transmitted with four-way handshaking in the other two protocols. The maximum network throughput of Slotted FAMA protocol is the lowest because that the sink can receive packets from only one neighbor for each round of handshakes in this protocol, while the sink can receive packets chain in the other two MAC protocols. Moreover, we also observe that the maximum network throughput of the proposed protocol without spatial reuse is higher than that of the Slotted FAMA, but it is lower than that of the RIPT protocol. This is because the spreading gain is too large to reduce the channel efficiency. When the spreading gain is reduced in the proposed protocol with spatial reuse, the network throughput is improved. Sink Sensor Figure 10. The network throughput.
The comparison of the packet drop rate of the three MAC protocols is shown in Figure 11. From Figure 11, we observe that the packet drop rate of the MAC protocols increases along with the increase of the traffic load. The packet drop rate of the proposed protocol with spatial reuse is much smaller than that of other two protocols when the traffic load is small. This is because the CDMA technology used in the proposed protocol can effectively reduce packet collisions from different paths. However, the packet drop rate of the proposed protocol without spatial reuse is larger than that of RIPT when the traffic load is larger than 0.05 packet/s. The reason is that the proposed protocol without spatial reuse uses too much time to forward the packets when long spreading sequences are used. In the proposed protocol, the first node is the initiator of the packet transmission, and other nodes must wait for packets from their previous nodes before transmitting their own packets. However, the packet drop rate in RIPT protocol is the largest among the protocols when the traffic load is less than 0.04 packet/s because the receiver-initiated approach used in this protocol makes senders wait too long to send packets. Figure 12 shows the performance of the end-to-end delay of the three MAC protocols. From Figure 12, we observe that the end-to-end delay increases when the traffic load increases for all three MAC protocols. Among these protocols, the end-to-end delay of Slotted FAMA is the smallest. The reason is that Slotted FAMA is a sender-initiated protocol, and a few packets are forwarded during a handshake round. The end-to-end delay performance of the proposed protocol without spatial reuse is the worst because the packets pre-processed with long spreading sequences need much time to be forwarded. Moreover, the end-to-end delay of the proposed protocol with spatial reuse is almost as same as that of the RIPT protocol. Hence, the end-to-end delay becomes small when the spreading sequences with relatively short length are used in the proposed protocol with spatial reuse. However, the end-to-end delay of RIPT protocol increases faster than that of the proposed protocol with spatial reuse when the traffic load increases. The reason for this phenomenon is that a sender cannot transmit its DATA packets until a handshake is initiated by the receiver in this protocol. Figure 13 shows the impact of the data packet size on the network throughput in three MAC protocols, where the traffic load is 0.018 packet/s. From Figure 13, we observe that the network throughput increases when the data packet size increases. The reason is that the data transmitted in each successful period increases along with the increase of the data packet size. The network throughput of the proposed protocol increases slightly when the data packet size increases since the time for transmitting data packets increases when the data packet size increases.
The network throughput of Slotted FAMA and RIPT increases fast when the data packet size increases. This is because that the rate of the data transmission in each time slot increases as the data packet size increases in these two protocols. End-to-end delay (s)
Traffic load
SlottedFAMA RIPT proposed MAC without spatial reuse proposed MAC with spatial reuse Figure 13. The impact of the data packet size on the network throughput.
Discussion
The primary aim of the proposed POCA-CDMA-MAC protocol is to alleviate the funneling effect around the sink in a sink-oriented UWASN. Different from the CDMA technology used in the PLAN protocol [20] and the CDMA-based protocol [21], a round-robin method is used in the proposed protocol, each node, therefore, can transmit its data packets without a handshake mechanism, which reduces the packet collision. Moreover, the POCA is used to reduce the length of the spreading sequences, which improves the channel efficiency.
The round-robin method is used first in the Ordered CSMA MAC protocol [5]. The authors also briefly claimed that the Ordered CSMA works well in a cluster and nodes in different clusters could transmit packets without interference with CDMA technology.
However, there are three significant differences between two protocols. First, the round-robin method is used among nodes within a cluster in the Ordered CSMA protocol, and a node senses the channel and transmits its own data packets in its order, and it does not need receiving and forwarding the data packets from its front node. In the POCA-CDMA-MAC protocol, the round-robin method is used in multi-hop networks, and a node does not transmit its own packets until it has received and forwarded the data packets from its front node. Second, the CDMA technology is suggested to be used between different clusters in the Ordered CSMA protocol, while the CDMA technology is used to different established paths in proposed POCA-CDMA-MAC protocol. Third, a simple, sink-oriented packet transmission is used for avoiding the hidden terminal problem in the multi-hop networks in the POCA-CDMA-MAC protocol, while the Ordered CSMA protocol is designed for a single-hop network.
The acknowledgement of the successful DATA reception is important for the protocol, especially in UWASNs with unreliable channels. In the Slotted FAMA protocol [17], a receiver replies an ACK packet when it has received DATA packet correctly. In the Ordered CSMA MAC protocol [5], a receiver just combines the ACK and DATA packet in one transmission, so the sender waits for a long time to receive the ACK packet. In the proposed protocol, a receiver forwards the DATA packet when it has received it correctly from the sender. Hence, the sender can just wait for to sense its DATA packet forwarded by the receiver. If the sender does not sense its DATA packet forwarded by the receiver, it triggers packet retransmission at the next round. The packet forward mechanism in our proposed MAC protocol saves the energy consumption of sensor nodes without acknowledgements, and makes the protocol work well in multi-hop networks. However, in order to prevent the sender from listening to a large packet, the receiver can also transmit a small ACK before forwarding the packet from the sender. Then the sender can listen to the small ACK to confirm whether its data packet is received by next-hop node or not. Moreover, when a node near the sink becomes invalid, the sink will assign another node to replace the invalid node. The paths will be re-built and the spreading sequences will be re-assigned. However, the hops of a path around the sink will not be long and the sink will monitor and control the paths, and the nodes far from the sink can random access the channel since the DATA packets transmitted by nodes far from the sink are much less than that transmitted by nodes near the sink.
Conclusions
In this paper, an improved CDMA-based MAC protocol with a round-robin method and CDMA technology is proposed for UWASNs. In the proposed MAC protocol, each node pre-processes its data packets with a spreading sequence so that the sink can receive data packets from multiple paths at the same time. Moreover, the nodes in the same path are assigned with the same spreading sequence, and they transmit their data packets pre-processed with the assigned spreading sequence in a round-robin method. Hence, the packet collision is reduced and the length of the spreading sequence codes is shortened. Simulation results show that the maximum network throughput of proposed MAC protocol with spatial reuse is almost two times of that of other two protocols, Slotted FAMA protocol and RIPT MAC protocol. Moreover, the proposed MAC protocol also achieves a lower packet drop rate and a smaller end-to-end delay compared to other two MAC protocols.
Since the primary aim of proposed MAC protocol is to alleviate the funneling effect in multi-hop UWASNs, the principle of the protocol is paid attention to rather than the theoretical algorithm. Our future work will concentrate on the development of a proper spread-spectrum code assignment algorithm, the time selection for the first node to send packets, and methods dealing with nodes without data in a round-robin cycle. | 8,853.6 | 2013-11-01T00:00:00.000 | [
"Computer Science",
"Engineering",
"Environmental Science"
] |
Airborne hyperspectral observations of surface and cloud directional reflectivity using a commercial digital camera
Spectral radiance measurements by a digital single-lens reflex camera were used to derive the directional reflectivity of clouds and different surfaces in the Arctic. The camera has been calibrated radiometrically and spectrally to provide accurate radiance measurements with high angular resolution. A comparison with spectral radiance measurements with the Spectral Modular Airborne Radiation measurement sysTem (SMART-Albedometer) showed an agreement within the uncertainties of both instruments (6 % for both). The directional reflectivity in terms of the hemispherical directional reflectance factor (HDRF) was obtained for sea ice, ice-free ocean and clouds. The sea ice, with an albedo ofρ = 0.96 (at 530 nm wavelength), showed an almost isotropic HDRF, while sun glint was observed for the ocean HDRF ( ρ = 0.12). For the cloud observations with ρ = 0.62, the cloudbow – a backscatter feature typically for scattering by liquid water droplets – was covered by the camera. For measurements above heterogeneous stratocumulus clouds, the required number of images to obtain a mean HDRF that clearly exhibits the cloudbow has been estimated at about 50 images (10 min flight time). A representation of the HDRF as a function of the scattering angle only reduces the image number to about 10 (2 min flight time). The measured cloud and ocean HDRF have been compared to radiative transfer simulations. The ocean HDRF simulated with the observed surface wind speed of 9 m s −1 agreed best with the measurements. For the cloud HDRF, the best agreement was obtained by a broad and weak cloudbow simulated with a cloud droplet effective radius of Reff = 4 μm. This value agrees with the particle sizes derived from in situ measurements and retrieved from the spectral radiance of the SMART-Albedometer.
Introduction
Surface reflectivity is a key parameter to estimate the Earth's atmosphere energy budget.As a lower boundary condition it is a parameter controlling the solar radiative transfer in the atmosphere.Considering the directional nature of radiometric quantities, such as radiance, the bidirectional reflectance distribution function (BRDF) fully describes the surface characteristics (e.g.Nicodemus et al., 1977;Schaepman-Strub et al., 2006).For the application of spaceborne instruments based on measurements of solar radiation, the BRDF is critical to retrieve aerosol or cloud properties.Hyer et al. (2011) found that correcting the surface albedo in the aerosol retrieval of the Moderate Resolution Imaging Spectroradiometer (MODIS) significantly reduces the variability of the bias between MODIS and ground based AOD measurements.
To estimate the impact of clouds on the Earth's energy budget from spaceborne measurements, the BRDF of clouds is required.Satellite instruments primarily measure spectral radiance and do not cover the entire hemisphere.However, the energy budget is calculated by hemispheric irradiance/top-of-atmosphere (TOA) albedo.To convert the satellite observations of reflectivity into TOA albedo, the cloud BRDF has to be known in terms of an angular distribution model (ADM, Loeb et al., 2000Loeb et al., , 2005)).From multiangular instruments such as the Clouds and the Earths Radiant Energy System (CERES, Loeb et al., 2005) and the Polarization and Directionality of the Earth's Reflectances instrument (POLDER, Loeb et al., 2000), empirical ADMs were derived from 24 and 5 months of observations, respectively.A different approach utilizing radiative transfer simulations was applied by Buriez et al. (2005) to measurements by POLDER.Plane-parallel radiative transfer calculations of the cloud BRDF for different cloud properties were used to convert the observations into TOA albedo.
However, for inhomogeneous clouds, plane-parallel radiative transfer calculations are not sufficient to simulate the angular reflectivity above clouds (e.g.Loeb and Davies, 1997;Varnai and Marshak, 2007).Analyzing observations of the Earth Radiation Budget Satellite (ERBS), Loeb and Davies (1997) found that plane-parallel simulations underestimate the reflectivity in the backscattering direction.Varnai and Marshak (2007) observed a bias in the cloud optical thickness retrieved by MODIS, which depends on the viewing angle of the sensor and cloud inhomogeneity.Both effects are significant for viewing angles of about 60 • and larger.Three-dimensional models may improve cloud BRDF simulations.However, given the diversity and complexity of clouds and the computational time of three-dimensional calculations, plane-parallel models are used for operative products, such as optical thickness and TOA albedo.These problems show that there is a need for measurements of the directional reflectivity above clouds.
Several ground-based and airborne retrieval techniques have been developed to derive the BRDF of different surfaces and clouds.While local ground-based measurements provide the BRDF of characteristic homogeneous surfaces (e.g.von Schönermark et al., 2004;Dumont et al., 2010), airborne data cover a larger measurement area and average over a mixture of different surface types, which is more suitable to the pixel size of spaceborne observations.However, for atmospheric measurements it has to be considered that the surface is illuminated by both the direct solar and the diffuse sky radiation.In this case, the measurements provide the hemispherical directional reflectance factor (HDRF) instead of BRDF.Based on radiative transfer calculations, the BRDF is derived afterwards by applying an atmospheric correction to the measured HDRF data.
State-of-the-art airborne BRDF instruments are mostly based on a scanning system measuring spectral radiance in different viewing angles.The Cloud Absorption Radiometer (CAR) presented by Gatebe et al. (2005) utilizes an optical system with a 1 • field of view.The mirror of the optical system is rotated at 100 r min −1 .An entire BRDF measurement of the lower hemisphere is obtained within 2-3 min.BRDF measurements with CAR are reported above ocean, savanna, salt pans, snow and clouds (Gatebe et al., 2003;Lyapustin et al., 2010).A similar instrument including polarimetric data, the Research Scanning Polarimeter (RSP), is used by Litvinov et al. (2011) to validate BRDF models of vegetation and soil surfaces.The RSP employs a telescope with 0.8 • field of view and a double mirror with a scan rate of about 70 r min −1 to cover zenith angles of ±60 • from the nadir direction.
Spaceborne multi-angular observations are obtained by instruments such as the Polarization and Directionality of the Earth's Reflectances instrument (POLDER, Descloitres et al., 1998) and the Multiangle Imaging SpectroRadiometer (MISR, Ovtchinnikov and Marchand, 2007).While POLDER provides a full image in ±43 • along track and ±51 • across track, MISR uses nine separate line cameras to cover nine different viewing angles.Using the airborne version of POLDER, Descloitres et al. (1998) compared the measured cloud HDRF (without atmospheric correction) to plane-parallel radiative transfer simulations, assuming spherical cloud particles.Differences ranged between 2 % for liquid water clouds and 9 % for ice clouds, which indicates that the scattering phase function of the cloud particles is essential for calculating HDRF.Assuming nonspherical ice crystals for the simulations, the differences are reduced to 2 %.With a similar approach, Ovtchinnikov and Marchand (2007) compared the radiance of different view angles measured by the airborne version of MISR and three-dimensional radiative transfer simulations.Differences appeared mainly in the nadir direction and are suggested to result from differences in the three-dimensional structure between observed and simulated clouds.
Here we present airborne measurements of HDRF using a commercial, single-lens reflex digital camera.Compared to scanning instruments, digital cameras instantly obtain a full scene of measurements without the need of high-precision movable components.The camera is easy to mount on an aircraft and relatively cheap.The high spatial resolution of the camera allows measurements with an angular resolution of about 0.1 • .However, due to the imaging system including lens and sensor, a careful calibration of the camera is required to quantify the angular dependence of the camera sensitivity, which might be affected by dark noise, saturation, distortion, or polarization effects.
Such a type of camera is still rarely applied in atmospheric sciences, even though there is an increasing use in vegetation and soil monitoring (Lebourgeois et al., 2008).Only a few studies have used such camera measurements quantitatively.From radiance-calibrated conventional photographs, Cox and Munk (1954) derived a parametrization of ocean BRDF.Digital cameras were introduced in the last century for ground-based cloud-cover detection (e.g.Long et al., 2006;Schade et al., 2009).However, instead of calibrated radiance, Long et al. (2006) and Schade et al. (2009) used the radiance-uncalibrated signals of the camera sensor to detect clouds by analyzing the three spectral channels (red, green, blue; RGB) of the CCD (charged coupled device) sensor.
We analyze radiance-calibrated digital camera images obtained from airborne measurements performed during a campaign in the Arctic.They are introduced in Sect. 2. The processing of the digital camera images, including radiometric and spectral calibration, is shown in Sect.3. HDRF measurements for different surfaces are presented in Sect. 4. The results for cloud and ocean HDRF are discussed and compared to radiative transfer simulations in Sect. 5. Section 6 presents the conclusions of this paper.
Instrumentation and measurements
We report on data collected during the Solar Radiation and Phase discrimination of Arctic Clouds (SORPIC) campaign in May 2010.During SORPIC the Polar 5 aircraft, owned by the Alfred Wegener Institute for Polar and Marine Research (AWI), Bremerhaven, Germany, was deployed to investigate Arctic clouds with a set of remote sensing and in situ instruments.With the Polar 5 based in Longyearbyen on Svalbard (78 • 13 N, 15 • 38 E), in total 13 flights were conducted covering the area of the Greenland Sea west of Svalbard.
The major purpose of the flights was to quantify the horizontal and vertical distribution of ice and liquid water in mixed-phase clouds by different independent approaches, including remote sensing and in situ measurements.The airborne instrumentation for remote sensing included the Spectral Modular Airborne Radiation measurement sysTem (SMART-Albedometer), the hyperspectral camera system AISA Eagle, the Airborne Mobile Aerosol Lidar (AMALi), and a commercial CANON EOS-1D Mark III digital camera.Additionally, an airborne sun photometer was operated to characterize aerosol properties.For in situ measurements, a Nevzorov probe, the Polar Nephelometer, a Cloud Particle Imager (CPI), and the Particle Measuring System (PMS) Forward Scattering Spectrometer Probe (FSSP-100) were installed on Polar 5. A detailed description of the instrumentation is given by Lampert et al. (2009) and Gayet et al. (2009).The SMART-Albedometer was described by Wendisch et al. (2001) and Ehrlich et al. (2008).For sea ice measurements, the electromagnetic-induction (EM) system EM-bird was operated in a towed sonde during the flights on 13 and 14 May 2010 (Haas et al., 2009).
To demonstrate the potential of HDRF measurements with the CANON camera, we present three selected cases of measurements above clouds, sea ice, and open water.For the cloud case, we focus on observations of pure liquid water clouds observed on 17 May 2010 south of Svalbard over icefree sea.A strong advection of warm air produced a persistent cloud layer in the lower boundary layer, with cloud top rising from 200 m in the south to 700 m in the north.Measurements above sea ice and open water were obtained during a flight with clear sky conditions on 14 May 2010.The sea ice was observed at about 82 • N, 2 • W; the open water at about 79 • 20 N, 10 • E. In all cases, clear sky was reported above the aircraft.
General characteristics
The CANON EOS-1D Mark III is a digital single-lens reflex (DSLR) camera, which incorporates a CMOS (Complementary Metal Oxide Semiconductor) image sensor providing the three spectral channels (RGB).The advantage of the CMOS image sensor compared to CCD sensors is the possibility of using larger sensors with low power consumption.This allows pixels with larger surface area, which increases the dynamic range of the sensor.With the new sensor generation, the noise and dark current level of CMOS sensors have been reduced to typical values of CCD sensors (Kaufmann, 2010).
The CMOS sensor applied in the CANON EOS-1D Mark III has the Advanced Photo System APS-H format with a 28.1 × 18.7 mm sensor area (crop factor of 1.3).The sensor has a 3908 × 2600 pixel grid and covers a total of about 10.2 × 10 6 pixels (10 megapixels).
To cover a large area, the camera was configured with the wide-angle lens Canon EF 14 mm f/2.8LII USM.Compared to a fisheye lens, this lens provides distortion-free images all the way across the frame.Thus, the camera field of view is calculated from the lens focal length of f = 14 mm and the sensor chip size d: . (1) For the horizontal (d = 28.1 mm) and vertical (d = 18.7 mm) direction, the angle of view is = 90.2• and = 67.5 • , respectively.The image diagonal has an angle of view of = 100.6• .The corresponding angular resolution of each pixel is about 0.025 • .
The camera was installed on Polar 5 close to a low definition digital video camera, as shown in Fig. 1.To protect the camera lens from damage by stone chipping and rain water, a A. Ehrlich et al.: Bidirectional reflectivity observations using a digital camera glass window was integrated in the aircraft frame in front of the lens.The camera was fixed to the aircraft frame, which made a correction to the aircraft attitude necessary.To guarantee the overlap of at least two subsequent images, the camera was aligned with its long image side along the aircraft axis.
To obtain the full dynamic range of the camera sensor chip, only raw data (RAW) were analyzed.Compared to the standard JPG format (8 bit), the RAW format provides 16 bit dynamic range.
To read the camera manufacturer-specific RAW format (Canon RAW version 2, CR2), we employed the open source tool DCRAW (http://www.cybercom.net/∼ dcoffin/dcraw/).With DCRAW, the CR2 images were converted into portable pixmap format (PPM) files using the command: To avoid any manipulation of the original measurements, no white balance was applied by setting the multipliers of all channels to 1.The darkness level was set to 0 and the saturation level to 16 384, respectively, with linear interpolation in between.Finally, the dark current of the images was determined in the laboratory for different camera settings and environmental conditions.Images without illumination were taken for temperatures between 10 • C and 25 • C. All data were taken with the same exposure as used during the airborne measurements (1/2656 s), which showed that the dark current does not exceed one digital unit.Thus the dark current is negligible, which agrees with Kaufmann (2010) who found that the dark current is no issue for exposure times below 33 ms.
Spectral calibration
To compare the camera measurements with spectral measurements of the SMART-Albedometer and radiative transfer simulations, the spectral sensitivity of each RGB channel was determined in the laboratory.The camera was mounted in front of a grating monochromator (Zolix Omni-λ300).A 200 W halogen lamp was used as radiation source.The spectral irradiance emitted by the lamp was determined by cross calibration of a 1000 W halogen lamp traceable to the National Institute of Standards and Technology (NIST) standard.Measurements with the camera were made between 300 nm and 700 nm wavelength for steps of 5 nm.For the monochromator, a grating with a blaze wavelength of 500 nm and a groove density of 1200 mm −1 was chosen, providing a spectral resolution of 0.1 nm.The wavelength accuracy of the monochromator is specified as 0.2 nm.The bandwidth was set to 5 nm, providing a sufficiently high radiance to be detected by the camera.The relative spectral response function RSR λ is defined by the normalization: (2) It is calculated from the measured camera signal S λ , and the irradiance F l,λ emitted by the 200 W halogen lamp by: The RSR λ function measured in the laboratory is shown in Fig. 2 for all three camera channels.The RSR λ of all channels is non-Gaussian, with their full-width of half-maximum (FWHM) ranging between 76 nm for the blue channel and 89 nm for the green channel.The center wavelength λ c of each channel (median value of RSR λ ) was determined as 591 nm (red, channel 1), 530 nm (green, channel 2), and 446 nm (blue, channel 3).
Radiometric calibration
The exposure time of the camera, aperture (f-number), and film speed were fixed during the measurements.The settings with an exposure time of 1/2656 s, an f-number of F/9.1, and a film speed of ISO-400 were chosen for cloud and sea ice observations with high reflectivities, but these settings worked as well for measurements above the open ocean.The short exposure time was chosen to avoid distortion due to the aircraft movement.
The radiometric calibration was obtained in the laboratory with the use of a NIST traceable radiance source (integrating sphere).The camera was mounted in the laboratory together with the protective glass window required for the aircraft installation (Fig. 1) in front of the aperture of the integrating sphere at 5 cm and 15 cm distances.For both distances the exit port of the integrating sphere with 6 cm diameter did not cover the whole image.Therefore, a series of images was taken while the camera was moved horizontally and vertically.No differences between measurement at both distances were observed.Therefore, all images were merged into a single calibration.
The calibration coefficients k λ were calculated for each camera pixel (x,y) and each camera channel using the camera signal S λ,C (x,y) (digital counts) and the NIST traceable radiance emitted by the integrating sphere I λ,IS by: Figure 3a shows the original calibration coefficients k λ of the merged images for channel 1 (591 nm).The plot indicates that the raw data of the camera is noisy.The noise is typical for CMOS image sensors and randomly distributed independent of the pixel position, as shown by laboratory tests (not shown here).Compared to channel 1, the noise of channel 3 (446 nm) is of similar magnitude, while channel 2 (530 nm) shows a reduced noise level.This is probably caused by the doubled number of channel 2 pixels of the Bayer filter used in the CMOS sensor.The data analysis is not seriously effected by the noise, as it is counterbalanced by the high number of pixels.To remove the noise in the calibration, a two-dimensional polynomial fit of 4th degree was applied to smooth the data.The final calibration coefficients kλ used to process the data are shown in Fig. 3b for channel 1.It shows that the sensitivity of the CMOS sensor is maximal in the center and decreases towards the edges of the sensor.The difference between maximum and minimum is about 40 %.This vignetting effect is well known for digital cameras (see Lebourgeois et al., 2008;Olsen et al., 2010).This pattern has been observed in all the three channels, indicating that the pattern results from lens effects.Lebourgeois et al. (2008) corrected the vignetting effect by fitting a polynomial function onto an average image of about 500 images.This method does not work if the observed surface is a non-isotropic reflector itself, e.g.sea ice, clouds, or open water.For such surfaces, the vignetting effect will be superimposed by the BRDF of the surface.In this case the vignetting effect has to be eliminated by a radiometric calibration, as presented above for the CANON camera.
Polarized radiation (e.g.sun glint) might increase the uncertainty of the camera measurements if the camera lens acts like a polarization filter.The sensitivity to linear polarized radiation of different orientation was tested in the laboratory using a source of 100 % linear polarized radiation.Differences between measurements of parallel and perpendicular polarized radiation were found to be negligible for the center of the images.Toward the edge of the image, this polarization effect slightly increased.Maximum effects were estimated to be 3 %.It has to be taken into account that for radiation, which is not 100 % polarized, this effect will be reduced by the degree of polarization.
The sensitivity of the CMOS image sensor was additionally tested for linearity.The results (not shown here) agree with the study reported by Kaufmann (2010), who showed an almost perfect linear response of the CMOS image sensor to the intensity of the incoming radiation.Thus, the uncertainty in the radiometric calibration results mainly from the uncertainty given for the certified radiance source.For the camera setup used in this study, an uncertainty in the radiance measurements of about 7 % was considered for each camera channel.
Geometry
As the camera is fixed to the aircraft frame, a correction for the aircraft attitude has to be applied before averaging different images.The definition of the coordinate systems is shown in Fig. 4, where the position of the Sun is defined by the solar zenith angle θ 0 and the solar azimuth angle ϕ 0 .The pixel coordinates are given by the viewing zenith angle θ v and the viewing azimuth angle ϕ v .The viewing zenith angle is derived from Eq. ( 1) by replacing the diameter of the sensor with the corresponding distance of each pixel to the center of the sensor.The viewing azimuth angle is defined clockwise, with 0 • showing into flight direction.
θ v and ϕ v have been corrected for the aircraft roll and pitch angle.Therefore, Euler rotations of the pixel coordinates with roll and pitch angles were applied.The rotation of θ v and ϕ v gives θ r and ϕ r , the zenith and azimuth angles of the reflected radiation in Earth fixed coordinates.Finally, the images have been rotated into the azimuthal direction of the Sun ϕ 0 .
Assuming single scattering, the scattering angle ϑ of direct solar radiation has been calculated for each image pixel (e.g.Wendisch and Yang, 2012).ϑ is defined as the angle between the direction of the Sun (θ 0 , ϕ 0 ) and the viewing direction (θ r , ϕ r ), and is calculated by: (5)
Hemispherical-directional reflectance factor HDRF
The reflectivity of surfaces is generally described by the bidirectional reflectance distribution function BRDF (Nicodemus et al., 1977;Schaepman-Strub et al., 2006).The BRDF describes how the incident irradiance F i from one direction (θ i ,ϕ i ) is reflected by a surface or a cloud into the direction (θ r ,ϕ r ).Here, F i = cosθ i • F 0,i refers to a horizontal surface.
With the reflected radiation being the radiance I r (θ r ,ϕ r ), the BRDF in units of sr −1 is defined by: In the literature, the dimensionless bidirectional reflectance factor BRF is often used instead of BRDF.It is defined as the ratio of the radiance I r actually reflected by a surface to the radiance I r,L reflected by an ideal (non-absorbing) and diffuse (Lambertian) standard surface for identical irradiation and beam-geometry.An ideal Lambertian surface reflects the radiation isotropically, and it holds that BRDF L = (π sr) −1 .This results in the definition of the BRF: However, both BRDF and BRF can be measured directly only when an artificial radiation source is applied.We present measurements in atmospheric conditions where the surface is illuminated by the Sun ( and by diffuse radiation (F diff ).Both components give the global irradiance F glob = F dir +F diff .In this case, the hemispherical-directional reflectance factor HDRF is measured (Schaepman-Strub et al., 2006): Using the definition of F glob and introducing the fraction of direct incident radiation f dir = F dir /(F dir +F diff ), Eq. ( 8) can be transformed to: The measured HDRF can be split into the BRF(θ 0 ,ϕ 0 ;θ r ,ϕ r ) for illumination of the surface by the Sun and the BRF(2π;θ r ,ϕ r ) for pure diffuse illumination of the surface.Both components are weighted with f dir , the fraction of direct incident radiation.
From HDRF measurements at a certain altitude, the BRDF, BRF, and the HDRF at surface level can be derived by applying an atmospheric correction using radiative transfer simulations, as shown by Gatebe et al. (2003) and Lyapustin et al. (2010).With the intention to validate the radiance and HDRF measurements of the camera as they are (at flight altitude), we do not apply an atmospheric correction for the comparison of measurements and simulations and present HDRF measurements instead of atmospherically corrected BRDF or BRF.
Spectral radiance
By applying the radiometric calibration, the camera provides spectral radiances for each pixel and camera channel.The accuracy of the calibration was verified by comparing the nadir radiance of the camera to spectral measurements of the SMART-Albedometer, which has an uncertainty of 6 % for radiance measurements (Ehrlich et al., 2008).The radiance optical inlet of the SMART-Albedometer is horizontally stabilized into nadir direction and has a field of view of 2.1 • .This spot corresponds to about 16 000 pixels of each camera image.These nadir pixels were averaged for each image.Furthermore, the spectral data of the SMART-Albedometer was adapted to the camera measurements by convolving the relative spectral response functions of the three camera channels (see Sect. 3.2).
In Fig. 5a, measurements of channel 1 are compared for an exemplary time interval on 17 May 2010, which were chosen to cover different surfaces such as sea ice, open ocean, and clouds.Despite the lower temporal resolution of the camera measurements (one image within 12 s), the time series of radiances obtained from the camera images agree with the SMART-Albedometer measurements (temporal resolution of about 1 s).The mean value of the SMART-Albedometer measurements between 09:28 and 10:28 UTC is Ī = 0.108 W m −2 nm −1 sr −1 , while the camera observed a mean nadir radiance of Ī = 0.104 W m −2 nm −1 sr −1 .This difference of 4 % ranges in the uncertainties range of the radiometric calibration of both instruments.As illustrated by the ratio of both measurements in the lower panel of Fig. 5a, the single data points differ more due to a non-perfect temporal allocation (integration times and sampling frequency) which makes averaging necessary.The standard deviation between both data sets is 0.006 W m −2 nm −1 sr −1 with a correlation coefficient of 0.99 (see Fig. 5b).For the other spectral channels (not shown here), a similar behavior was observed, with differences in the mean values of 1 % for channel 2 and 2 % for channel 3. Standard deviation and correlation coefficient are almost identical for all channels.The agreement between both instruments shows that the CANON camera is capable of quantitatively measuring the distribution of reflected radiances, which can be used to derive the HDRF.In the following, results will be shown for channel 2 (530 nm) only.Channel 2 was chosen because it shows the smallest differences (1 %) to the SMART-Albedometer data and has the lowest electronic noise, as discussed in Sect.3.3.
HDRF examples
Images of the CANON camera were analyzed for three cases: sea ice, open water, and clouds.The HDRF was calculated using Eq. ( 8).The downward irradiance F glob (θ 0 ,ϕ 0 ) was obtained from measurements of the SMART-Albedometer.
The time and position of the observations, the corresponding position of the Sun, and the number of images used to build the averaged HDRFs are given in Table 1.The measurements above sea ice and open water were conducted on 14 May.The cloud scene was part of the measurements on 17 May, which were analyzed in Sect.4.1.Using the irradiance measurements of the SMART-Albedometer, we additionally calculated the spectral albedo for each case.The measured albedo corresponding to the 530 nm channel of the camera is given in Table 1.For sea ice the albedo reaches a mean value of ρ = 0.96.Above open water and clouds, ρ = 0.12 and ρ = 0.62, respectively, were observed.The mean HDRF of each case is shown in Fig. 6 for camera channel 2. Additionally, a single characteristic image of the observed surface is shown.
Sea ice
Similarly to the albedo, the highest HDRF with values exceeding 1.0 was observed for sea ice, which was almost completely covered by snow.The measurements were conducted during the release of the towed EM-bird sonde, which observed a mean sea ice thickness of 2.5 m.Therefore, the rope of the sonde was present in all images slightly affecting the HDRF measurements.Furthermore, ice ridges, as shown in Fig. 6a, have been frequently observed on the sea ice, showing a high contrast in the reflected radiance between shadow and illuminated areas.These horizontal inhomogeneities remain partly present in the mean HDRF calculated from 46 single images.However, the HDRF shows an almost Lambertian-like pattern, with only slight variability between 0.95 and 1.10 in the magnitude of HDRF.The lowest values are observed for nadir direction.The weak anisotropy with increasing HDRF along the principal plane is slightly stronger in the direction of the Sun (up to 1.1) than in the opposite direction (up to 1.0).For a similar solar zenith angle of θ 0 = 67 • , Lyapustin et al. (2010) showed that the hot spot of the Sun influences the measured HDRF for zenith angles higher than 30 • along the principal plane.This is in agree-ment with the camera measurements covering zenith angles up to about 60 • .
Sea water
Due to the high altitude at which the measurements above open water have been conducted, an atmospheric correction has been applied to extract the contribution of radiation reflected by the atmosphere below the aircraft.We adapted the iterative correction method by Wendisch et al. (2004) for the radiance measurements of the camera assuming a Lambertian-reflecting surface in the radiative transfer simulations.
Compared to the sea ice, the camera measurements above open water show a non-Lambertian pattern dominated by sun glint.In general, the HDRF of sea water (with minimum values of about 0.02) is significantly lower than for the sea ice and cloud case, which agrees with the low albedo.The sun glint area, which was only partly covered by the camera, shows values of up to 0.4.The maximum of the sun glint (specular reflection for 61 • ) ranges outside the camera angle of view and might have even higher values.As discussed by Cox and Munk (1954), sun glint is caused by specular reflection at the surface waves, which is visible in the individual camera image of Fig. 6d.The surface wind measured by a drop sonde during the observations had a speed of about 9 m s −1 with a northerly direction (360 • ).
Compared to the sea ice and cloud measurements, the open water measurements required fewer images (11) for averaging because the sea surface is more homogeneous (as seen from about 3000 m altitude) than for the sea ice and cloud observations.
Clouds
The HDRF of a representative cloud was derived from measurements (50 individual images) above a low-level stratus cloud layer.For the area covered by the camera, the cloud HDRF ranges between 0.45 and 0.8 for the area covered by the camera.The anisotropy of the cloud HDRF mainly reflects the anisotropy of the scattering phase function of the cloud particles.For liquid water droplets, the scattering phase function has a maximum in forward scattering direction which explains the increasing HDRF in the direction of the Sun.The minimum values in both HDRF and scattering phase function are observed for the broad range
Averaging
Due to inhomogeneities of the observed scene, the camera images had to be averaged to obtain a representative HDRF measurement.In the above examples, all available images (46, 11 and 50) were averaged for the sea ice, open water and cloud case, respectively.Especially for clouds, the narrow patterns of the glory and the cloudbow are visible in a single image only if the clouds are highly homogeneous.In most cases, even cloud layers such as stratocumulus have small-scale inhomogeneities which disturb the view of the glory and cloudbow.Therefore, averaging of several images was required to remove cloud inhomogeneities in the HDRF measurements.
For airborne measurement with POLDER, Descloitres et al. (1998) showed that after averaging a sequence of cloud observations, the scene acts like a plane-parallel cloud.The averaging approach assumes that the temporal cloud variability observed by each pixel in a sequence of images is similar to the spatial variability of one single image.Descloitres et al. (1998) found that about 10 images are required to sufficiently reduce the spatial variability for the observed cloud cases.
In a similar way, we investigated how many images are needed for sufficient averaging for the cloud observed during SORPIC.Data were analyzed for a typical stratocumulus observed on 17 May, 09:32 to 09:42 UTC.A single image Using more images (20 and 50), the cloud structure begins to vanish but the cloudbow becomes more pronounced in the mean HDRF.This implies that for the stratocumulus case investigated here, averaging of about 50 images or more is necessary to obtain a HDRF in which the scattering phase function of the cloud droplets dominates the mean HDRF compared to cloud inhomogeneities.The number of 50 images is limited to this single case study only and may significantly differ for clouds with stronger inhomogeneity and observations at different altitudes.A stronger inhomogeneity would require more images to be averaged.On the other hand, images taken close to cloud top (not shown here) indicated the glory and cloudbow already in one single image.
To quantify the inhomogeneity of the HDRF, the standard deviation σ 15 • of the HDRF was calculated for a circle of zenith angles lower than 15 • (about 70 000 camera pixels).This narrow area was chosen to ensure that the standard deviation is not affected by the cloudbow at zenith angles larger than 15 • .Before calculating σ 15 • , the HDRF was filtered by a 2-D low pass filter using an averaging window of 50 × 50 pixels.The filter removes the electronic noise in the images, which would also have been reduced by the averaging of images and thus biased σ 15 • .The filter window of 50 × 50 pixels is smaller than the natural cloud homogeneities and thus separates the effects of the electronic noise and natural cloud inhomogeneities.The values of σ 15 • calculated for the mean HDRF of 5, 10, 20, and 50 images are given in Table 2.The values decrease with increasing number of images -from σ 15 • = 0.014 for 5 images to σ 15 • = 0.009 for averaging 50 images.In general, σ 15 • does not converge to zero with increasing number of images, which is due to the anisotropy of the theoretical HDRF in the 15 • circle.To estimate the range of σ 15 • for a perfectly homogeneous cloud, radiative transfer simulations were performed.For a cloud of optical thickness of τ = 12 and particle effective radius of R eff = 10 µm, the simulations give a σ 15 • of 0.01.This ideal value ranges above σ 15 • = 0.009 obtained for the mean HDRF using 50 images.This contradiction can only be explained by general A. Ehrlich et al.: Bidirectional reflectivity observations using a digital camera differences of the measured and simulated HDRF, but shows that an average of 50 images is sufficient to eliminate cloud inhomogeneities for the case presented here.
A way to reduce the required number of images is to present the HDRF as a function of the scattering angle ϑ.Assuming that the scattering at homogeneous surfaces is rotationally symmetric with respect to the solar zenith angle, each image can be translated from the HDRF(θ 0 ,ϕ 0 ;θ r ,ϕ r ) defined by the solar and viewing zenith and azimuth angles into a HDRF(ϑ) defined by the scattering angle ϑ.This transformation allows the averaging of several image pixels into one HDRF value for the corresponding ϑ.We calculated HDRF(ϑ) with a resolution of 0.1 • .For each incremental scattering angle ϑ = 0.1 • , about 10 000 pixels were averaged.In this way, the electronic noise of the camera sensor and the horizontal cloud inhomogeneities are smoothed more efficiently.The corresponding mean HDRF(ϑ) for averaging 5, 10, 20, and 50 images are shown in the right panels of Fig. 8 for channel 2. The HDRF(ϑ) shows much less variability due to cloud inhomogeneities compared to the mean HDRF shown in the left panels of Fig. 8.The cloudbow can be already identified in the mean of 5 images.Averaging 10 images or more, the horizontal cloud inhomogeneities have been removed almost completely.
Although the average of 50 images indicates the backscatter glory for scattering angles larger than 176 • , the glory was not perfectly covered on most images, being situated at the edge of the images.Therefore, the glory was not analyzed in the following.
Simulated HDRF
For the measurements above open water and above clouds, the HDRF was simulated by one-dimensional plane-parallel radiative transfer calculations.The simulations were run with the library for radiative transfer libRadtran by Mayer and Kylling (2005) using the discrete ordinate radiative transfer solver DISORT version 2.0 by Stamnes et al. (1988).The meteorological input (profiles of static air temperature, relative humidity, and static air pressure) was obtained from the drop sound released from Polar 5 at 10:25 UTC, 14 May for the open water case and 09:36 UTC, 17 May for the cloud case.
Radiances were calculated for the entire lower hemisphere.For one half of the cloud case with viewing direction into the Sun and for the entire open water case (where the HDRF is more homogeneous), the angular resolution was 5 • for the azimuth angle and 3 • for the zenith angle.The second half of the cloud case, including the glory and cloudbow in the backscattering region, was simulated with a higher angular resolution of 0.5 • for both angles.The results have been interpolated to the same grid as obtained by the camera measurements to allow a direct comparison.
Open water
The BRDF of sea water calculated by libRadtran is based on the parametrization of Cox and Munk (1954) and Nakajima and Tanaka (1983).The magnitude of the sun glint and the shape of the BRDF are mainly determined by the surface wind speed.Therefore, the parametrization was adjusted to the surface wind speed measured by the drop sonde.To analyze the sensitivity of the simulations with respect to the wind speed, three simulations with 5 m s −1 , 9 m s −1 , and 15 m s −1 were performed, with 9 m s −1 being the value measured during the observations by a drop sonde.The wind direction was set to a northerly direction (360 • ,) corresponding to the observations.For the pigment concentration and the salinity, default values (0.01 mg m −3 for pigment concentration and 0.1 ppt for salinity) were used.The simulations were performed for both surface and flight altitude to allow a direct comparison of the uncorrected measurement.
The result of the HDRF simulations with 9 m s −1 wind speed is shown in Fig. 9a for the flight altitude.The low values and the position of the sun glint agree with the measurements presented in Fig. 6e.In Fig. 9b, the absolute differences between measurements and simulations at flight altitude are given.For most of the areas covered by the camera measurements, the differences range below 0.01, indicated by the turquoise color.Only for the sun glint area at zenith angles larger than 45 • did the differences increase significantly and exceed values of −0.2.The negative values show that in this area the simulations overestimate the HDRF compared to the measurements.Unfortunately, the sun glint is located at the outer edge of the image, where measurement uncertainties may increase due to a decreasing sensitivity of the camera sensor towards the sensor edges.However, an improper radiometric calibration of the camera can be excluded as reason for the deviations, as the differences occur only in the sun glint, while other boundary areas of the image agree well with the simulations.
Radiation reflected at angles similar to the sun glint is partially polarized, which may have affected the measurements for these scattering angles (Takashima, 1985).For the solar zenith angle (61 • ) and a scattering angle of about 60 • , where maximum differences show up between simulated and measured HDRF(ϑ), the degree of polarization may reach maximum values up to 0.9 (A. Hollstein, personal communication, 2011).However, the uncertainty of the camera due to polarization was estimated to be 3 % at maximum and cannot completely explain the differences between measurements and simulations.The images were also checked for saturation.With a low exposure time (1/2656 s) adjusted to the bright scenes of clouds and sea ice, no saturation was evident in the data.The raw data of the images showed maximum digital counts of about 12 000 in the sun glint and about 25 000 counts for sea ice, with a saturation value of 65 536 (16 bit).The angular distribution HDRF(ϑ) can be used to analyze the differences between simulations and measurements, as shown in Fig. 9c-d.From the simulations, results for the en-tire hemisphere are obtained.However, we only calculated HDRF(ϑ) from the area which was covered by the camera images.The HDRF(ϑ) of the entire hemisphere (not shown here) differs significantly for most of the scattering angles because they include the viewing directions close to the horizon, where multiple scattering leads to enhanced reflection.Figure 9c-d additionally shows the measurements corrected for the atmosphere and the simulations at surface altitude (red lines).
Comparing simulations and measurements of the selected area (Fig. 9d), the simulated HDRF(ϑ) differs from the measurement for scattering angles lower than 80 • , while for larger scattering angles they fit into the uncertainty range of the measurements.The position of the local sun glint maximum within the image is well covered by the measurements, but the magnitude differs by up to 0.25.Similar differences are observed for the HDRF at surface altitude.
The HDRF of simulations carried out with surface wind speeds of 5 m s −1 and 15 m s −1 are shown in Fig. 9c and e.Compared to the HDRF using 9 m s −1 wind speed, it stands out that the maximum HDRF values, which are located in the sun glint area, decrease with increasing wind speed.The closest agreement with the measurements is obtained from the simulations with 15 m s −1 wind speed.However, while the simulations with 9 m s −1 wind speed fit to the measurements at flight altitude for all scattering angles larger than 80 • , the simulated HDRF at 5 m s −1 and 15 m s −1 differ for these scattering angles.The HDRF simulated using a wind speed of 5 m s −1 ranges significantly below the measurements for scattering angles between 80 • and 120 • , while the 15 m s −1 HDRF ranges above the measurements for all scattering angles.The higher HDRF values simulated for 15 m s −1 wind speed may result from an increase of white caps.The amount of white caps, which is correlated to the albedo of open water (Gordon and Jacobs, 1977), increases nearly linearly between 5 m s −1 and 15 m s −1 , as shown by Stramska and Petelski (2003).A higher albedo is directly linked to a higher HDRF.Therefore, we argue that the measured HDRF correspond best to the observed wind speed of 9 m s −1 , despite the differences in the sun glint area.
Clouds
To analyze the HDRF measured above clouds, radiative transfer calculations were used to simulate the cloud case observed on 17 May.The cloud optical properties required for the model input have been retrieved from SMART-Albedometer measurements using the method introduced by Nakajima and King (1990).For the flight leg between 09:49 UTC and 09:59 UTC (see Table 1), the mean optical thickness was about τ = 11.5, with the cloud droplet effective radius R eff ranging between 4 µm and 10 µm.The effective radius obtained from the in situ instrumentation about one hour after the remote sensing measurements was about R eff = 9 µm.
Considering the variation of R eff , simulations for two clouds with R eff = 4 µm and R eff = 10 µm were performed.As τ and R eff are linked with each other, we adjusted τ to fit the simulated HDRF in nadir direction to the measurements of the SMART-Albedometer.For the cases of R eff = 4 µm and R eff = 10 µm, the cloud optical thickness was scaled to τ = 10.5 and τ = 12.0, respectively.The spectral surface albedo is represented by SMART-Albedometer measurements above sea water obtained for similar conditions during the ASTAR 2007 campaign (Ehrlich et al., 2008).For this cloud case with a moderate cloud optical thickness, the albedo is sufficient to describe the surface reflectivity.The BRDF derived using the parametrization of Cox and Munk (1954) has not been applied, because additional simulations have shown no differences between simulations using the albedo or the BRDF.
For both simulations with R eff = 4 µm and R eff = 10 µm, the HDRF is shown in Fig. 10 (upper panels).In both cases, the HDRF is characterized by the sun glint for zenith angles larger than 75 • in the direction of the Sun and the glory and cloudbow in the backscattering region.The lowest HDRF is simulated in the nadir direction.The comparison of both simulations indicates that the size of the glory decreases with increasing cloud droplet effective radius.The first order maximum is at 176.3 • scattering angle for R eff = 4 µm and ϑ = 178.4• scattering angle for R eff = 10 µm, respectively.Similarly, the characteristics of the cloudbow change with cloud droplet size.For simulations with larger droplet size, the cloudbow is more pronounced and the maximum is at a smaller scattering angle compared to a cloud with small droplets.
The simulations are compared to the HDRF derived from the camera measurements in the middle panels of Fig. 10, where the differences between measurements and simulations are shown.Positive differences (green and orange color) correspond to cases where the measurements showed higher values than calculated by the simulations.The blue color indicates negative differences where the measured values range below the simulations.
For both simulations the lowest differences are observed in nadir direction, which results from scaling the optical thickness with regard to the measured nadir radiance.Higher differences up to values of −0.2 are obtained for larger zenith angles.In the direction of the Sun, these differences are related to the sun glint where the HDRF is enhanced.Here, the simulation did calculate significantly higher HDRF values than observed by the camera.In the backscatter region, the differences are related to the glory and cloudbow, with maximum differences close to the 180 • point at 56 • zenith angle.Most striking are the differences corresponding to the cloudbow in the simulations for R eff = 10 µm.While the cloudbow pattern is visible for large droplets, the simulations for the smaller cloud droplets (R eff = 4 µm) do not significantly differ from the measurement.This indicates that the HDRF measurements can be used to characterize the cloud particle size.
A similar picture is obtained by comparing the angular distribution HDRF(ϑ) of the limited area to the camera measurements, as illustrated in the lower panels of Fig. 10.Again, the simulations for R eff = 4 µm fit better to the observations than the simulations using R eff = 10 µm.The signature of the cloudbow with a broader and lower maximum is reproduced best if smaller cloud droplets are assumed in the simulations.The narrow and intense cloudbow simulated for the larger cloud droplets significantly exceeds the measured HDRF(ϑ) at scattering angles around 142 • .
The differences at small and large scattering angles below 80 • and above 150
Conclusions
Images measured with a commercial digital single-lens reflex camera have been analyzed to produce the HDRF of different surfaces and clouds.For this purpose, the camera was calibrated spectrally and radiometrically.The central wavelengths of the three spectral channels are 591 nm (red), 530 nm (green), and 446 nm (blue) with a FWHM of about 80 nm.The radiometric calibration showed a decreasing sensitivity towards the boundaries of the camera sensor, which is a typical vignetting effect of digital photo cameras (Lebourgeois et al., 2008).Dark current, polarization effects, and sensor saturation were found to be negligible for the measurement uncertainty.A comparison with spectral radiance measurements provided by the SMART-Albedometer shows differences below the uncertainty range of both instruments (6 %).This agreement shows that the CANON camera is capable of measuring calibrated radiances.HDRF measurements were obtained for sea ice, open water, and clouds.In general, the results agree with known literature.Compared to traditional measurements, the high spatial resolution of the camera provides a detailed view on the angular pattern of the HDRF, including the sun glint of open water and the cloudbow for the cloud HDRF.However, to obtain a representative HDRF, averaging was necessary.Due to the high spatial resolution of the camera, small-scale inhomogeneities of the surface (sea ice or open water) or of the cloud were resolved by the observations and then averaged.For the inhomogeneous stratocumulus clouds analyzed here, the required number of images was estimated to be 50.When the HDRF was translated into an angular distribution HDRF(ϑ), the required number of images was reduced to 10.With a sampling frequency of one image per 12 s, these required numbers of 50 and 10 images correspond to sampling times of 10 min and 2 min, respectively.These could be reduced if the maximum sampling frequency provided by the camera (one image in 6 s) was applied.These numbers hold only for the clouds investigated here and may change for clouds with different inhomogeneity.Also, the flight altitude will alter the details resolved by the camera and thus the sampling time for one HDRF measurement.
For the measurements above open water and clouds, radiative transfer simulations providing HDRF were applied and compared to the measurements.Except for the sun glint region, the open-water case agreed well with the HDRF based on the BRDF parametrization of Cox and Munk (1954).The magnitude of the sun glint was simulated with higher values compared to the measurements.Simulations with a higher surface wind speed reduced the sun glint but also increased the HDRF outside the sun glint.Known measurement uncertainties like polarization effects, lens distortion and sensor saturation were estimated and ruled out as reasons for the differences.Further measurements with different surface wind conditions and solar zenith angles have to be analyzed to determine whether these differences are model or measurement based.
The measurements above clouds showed that the cloudbow can be extracted from the images.The position, magnitude, and width of the cloudbow agreed with simulations, assuming cloud droplets with an effective diameter R eff = 4 µm.Simulations assuming R eff = 10 µm failed to reproduce the observed cloudbow.This indicates that the analysis of the cloudbow could possibly be used to retrieve the cloud effective diameter.A similar approach was successfully applied by Mayer et al. (2004) who derived the particle size from analyzing the width of the backscatter glory.For ice clouds, multi-angle satellite measurements have been utilized by Chepfer et al. (2002) to retrieve the ice crystal shape.This method, based on differences in scattering phase functions of ice crystals, might be applied to our camera measurements in future.However, detailed analysis of the images and further observation of different clouds are necessary to obtain a reliable retrieval.Uncertainties in the aircraft attitude may broaden the cloudbow when the images are not perfectly corrected.Additionally, stereo effects for inhomogeneous clouds with varying cloud top height may broaden the cloudbow.This would lead to an underestimation of the cloud droplet size.However, HDRF above different clouds (not shown here) did show a narrower cloudbow, indicating larger cloud droplets.As it was not our intention to provide a retrieval method for the cloud effective radius, no detailed studies on radii are shown here.
The HDRF measurements presented here are limited to the field of view of the camera lens with a maximum of 100 • in the image diagonal.A circular flight pattern might be helpful to increase the angular coverage of the images.However, considering the required averaging, several circles would have to be flown, which would increase the sampling time of one HDRF measurement significantly.An alternative to improve the camera measurements would be the application of a 180 • field of view lens, which would enable us to cover the entire lower hemisphere within one single image.In this way, the full sun glint pattern and the entire backscatter glory would be covered and provide more detailed information on the surface and cloud microphysical properties.
Fig. 1 .
Fig. 1.Photographs of the installation of the CANON camera on board of Polar 5 (arrows indicate the flight direction).In both photographs (interior and bottom view) the CANON camera is labeled with (a) and the digital video camera with (b).
Fig. 2 .
Fig. 2.Relative spectral response function RSR λ of the three camera channels (red, green, blue).For each channel, the center wavelength (median value) λ c and the FWHM are given.
Fig. 3 .
Fig. 3. Radiometric calibration coefficients k and k of camera channel 1 (λ = 591 nm).Panel (a) shows the noisy raw data (k).In Panel (b) a two-dimensional fit was applied to smooth the data ( k).The calibration is valid for an exposure time of 1/2656 s, an f-number of F/9.1, and a film speed of ISO-400.
Fig. 4 .
Fig. 4. Illustration of the airborne fixed(left, θv, ϕv) and Earth fixed coordinates (right, θr, φr) of one single camera pixel.Additionally, the scattering angle ϑ is indicated with the position of the Sun defined by θ0, φ0.
Fig. 4 .
Fig. 4. Illustration of the airborne fixed (left, θ v , φ v ) and Earth fixed coordinates (right, θ r , ϕ r ) of one single camera pixel.Additionally, the scattering angle ϑ is indicated with the position of the Sun defined by θ 0 , ϕ 0 .
Fig. 5 .
Fig. 5. Time series (17 May 2010) comparison (a) of a spectral radiance measured by the SMART-Albedometer (red) and by the CANON camera (black).Data are shown for the red camera channel (λ = 591 nm).The correlation between both measurements is illustrated in second panel (b).
Table 2 .
Standard deviation σ 15 • of the mean HDRF using 5, 10, 20 or 50 images calculated for a circle of zenith angles lower than 15 • .Additionally, σ 15 • is given for a plane-parallel cloud of optical thickness τ = 12 and particle effective radius R eff = 10 µm.horizontal cloud structure of the stratocumulus is shown in Fig.7.The mean HDRFs for averaging 5, 10, 20 and 50 images are shown in Fig.8for channel 2. With a sampling frequency of one image for each 12 s, this corresponds to flight times of 1 min, 2 min, 4 min and 10 min.The plots show that for averaging 5 and 10 images, the cloud structure is still visible in the mean HDRF.
Fig. 9 .
Fig. 9. Simulated HDRF of open water.The left panels show (a) the HDRF simulated with 9 m s −1 wind speed and (b) the difference to the measurements.The angular distribution of the HDRF is compared between measurements (solid line) and simulations (dashed line) with different wind speed in panels (c) 5 m s −1 , (d) 9 m s −1 and (e) 15 m s −1 .The uncertainty of the measurements is indicated by the gray area.The red solid and dashed lines represent the measured HDRF(ϑ) after atmospheric correction and the HDRF(ϑ) simulated at the surface, respectively.
Fig. 10 .
Fig. 10.Simulated HDRF for clouds with optical thickness and effective diameter of τ = 12 and R eff = 10 µm (left) and of τ = 10, R eff = 4 µm (right).The upper panels show the HDRF of the entire lower hemisphere.Differences between measured and simulated HDRF are given in the middle panels(c, d).The lower panels (e, f) give the HDRF(ϑ) as function of the scattering angle for measurements and simulations.The uncertainty of the measurements is indicated by the gray area.
Table 1 .
HDRF measurements above sea ice, open water, and clouds.
www.atmos-chem-phys.net/12/3493/2012/ Atmos. Chem. Phys., 12, 3493-3510, 2012 the
• correspond to the margins of the camera images where the statistics are bad compared to the center of images.Furthermore, 3-D effects may reduce the measured HDRF(ϑ) compared to the simulations.As Loeb and Coakley Jr. (1998) andLoeb et al. (1998)have shown, the 3-D structure of clouds leads to a decreasing cloud reflectivity towards the horizon compared to one-dimensional planeparallel simulations which we applied here. | 12,968.2 | 2012-04-11T00:00:00.000 | [
"Environmental Science",
"Physics",
"Mathematics"
] |
Experimental and numerical analysis of solar cell temperature transients
Many factors determine the efficient operation of a photovoltaic cell. These factors can be the intensity and spectral composition of illumination, the surface temperature, the ambient temperature, and the amount contaminations in the air and on the surface of the cells. The aim of the present study is to describe the effect of temperature gradient on the voltage and amperage changes, as well as the power output of a commercial solar cell through experimental methods and numerical simulations performed in MATLAB. The transient temperature investigations have allowed better understanding the time-dependent behavior of a solar cell under constant intensity illumination. Measurements prove that an increase in the surface temperature of the solar cell significantly reduces its performance. Measurements performed with the solar simulator show good conformity with simulated results.
INTRODUCTION
The 21st Century is considered by many to be the golden age of solar power utilization. Their efficiency is increasing steadily, but it should not be overlooked that their operation is affected by several environmental factors throughout the present study, the effects of the change of the surface temperature of the solar cell on the cell's electrical parameters are investigated.
Experimental results were obtained by providing artificial illumination using an ASTM E972 (IEC 60904-9) [1], standard solar simulator. Correlations were then obtained with the help of the measurement results and the results acquired from MATLAB simulations [2]. The future development goal is to refine the established theoretical model based on the measurement results. There is a large body of literature devoted to the experimental and numerical investigation of solar cells at constant temperatures [3][4][5], but very few researchers [6,7] investigate the phenomenon of transient temperature and its effects. The present study primarily contains results measured and simulated in the transient state, which also represents the novelty of the research work.
Literature research, overview of solar cell models
To simulate the operation of a solar cell, the first step is to establish its electronic model. Several models of equivalent circuits of a solar cell can be found in the related literature [8][9][10][11][12][13][14], this study is started by reviewing them. In this chapter, without being exhaustive, the most commonly used models will be briefly described. Figure 1 shows the described models and Fig. 2 presents the experimental arrangements.
Model a) in Fig. 1 is an ideal equivalent circuit of a solar cell, consisting of a current source and a diode [9,15]. Compared to the ideal circuit, model b) contains a seriesconnected resistor, which is intended to incorporate the resistance of the constructed solar cell [9]. In model c), a further extension is the resistor connected in parallel with the shunt diode [9]. Model d) is the most complex equivalent circuit of a solar cell. In this case, a double shunt diode is incorporated into the model [8,9,15]. This variant is considered to be the most accurate model to simulate the operation of a solar cell [8,9,12,15]. The other described equivalent circuits can be derived from this one as well [16].
Some literature discusses how to compare the accuracy of different models, which can be helpful to choose the model to be applied. In this case, the model of Fig. 1c was taken as the basis of the simulation model of the solar cell. The reason for this is that, according to the literature, there is no significant difference between the accuracy of models c) and d), the calculations when using the c) equivalent circuit are, however, much simpler [8,9,12,15].
Model construction
Accordingly, the photo-current (I ph ) provided by the current source of this model, describes the charge carrier separation occurring because of the sunlight in the p-n junction of the solar cell well; the diode of the model adequately models the processes occurring within the p-n junction [10]. The serial and parallel resistors describe the deviation from the ideal model and the individual losses of a solar cell. The series resistance (R s ) is given by the distance between the p-n junction and the metallic conductors on the surface of the semiconductor layer, and to a small extent by the resistivity of the conductors [10]. The parallel resistance (R p ) mainly occurs at the edge of the cells, and it indicates the effect of currents caused by the recombination of charges that are bypassing the p-n junction [10]. This leakage current can be minimized with proper insulation; therefore, it has a negligible impact on the operation of today's modern solar cells [3,10,11]. Based on the equivalent circuit is shown in Fig. 1, the following equation can be written [11,16]: where ; g is a cell-specific factor [-]; k is the Boltzmann-constant [8,11,17]. In this equation, the value of the parallel resistance R p , based on previously discussed considerations, is chosen to be infinitely large [8]. Normally R p would be rather difficult to determine, choosing it to represent a break does, however, not result in a significant change in the accuracy of the model [18]. Therefore, the last term of Eq. (1) is zero, so there are four variables: I ph , I 0 , g, R s [16,18]. These variables can be determined using the solar cell characteristics given in Table 1.
When determining the variables of Eq. (1), it must be considered that the main goal of the research is to model the operation of a solar cell as a function of temperature and the irradiation [3]. To determine the four unknown parameters taking the changes in the intensity of irradiation and temperature into accountthe following equations can be devised [3,[8][9][10][11]19]: (2) (4) and (5) merely simplify the expression of the equations and can be derived from the following equations [2,3,8,[18][19][20][21][22]: In order to solve the equation system, it is also necessary to determine the reference value of the photo current I F ref and the reference value of the saturation current of the diode I 0 ref [2,8,21,22]. These two parameters can be expressed by substituting the open circuit and short circuit cases into Eq. (1). The values of the two parameters, after further sorting, can be expressed as follows [2,8,18,21,22]: With the help of the described Eqs (2)-(9), the main Eq.
(1) of the equivalent circuit of the solar cell becomes implicitly solvable, hence the curve of the solar cell can be determined as a function of temperature and irradiation by the model [11,12,20]. So far, the effect of the temperature of the solar cell was only considered, but in practical applications it is useful to determine the relationship between the temperature of the solar cell and the ambient temperature [20,22]. The basis is the following solar energy balance (Electric energy is equal to the difference of adsorbed and dissipated energy) [23]. Energy balance can be expressed reduced to a unit surface of the solar cell as follows: (10), the cell temperature can be expressed as a function of ambient temperature [20,22]: There are several unknown variables in Eq. (11), of which the product of τa is chosen to be 0.9 as recommended by the literature. As it can be seen, the efficiency of a solar cell depends on the temperature, so accurate determination can only be achieved using an iterative approach. By executing the calculations with efficiency valid for Maximum Power Point (MPP), the equation can be written as follows [24]: The last missing parameter, namely the heat transfer coefficient of the solar cell (U L ), is determined by the help of the so-called Nominal Operating Cell Temperature (NOCT), which is found among the solar cell data. The required equation can be written as follows: where E NOCT is the solar irradiation in case of NOCT, usually 800 W/m 2 -1,000 W/m 2 ; T a NOCT is the ambient temperature in case of NOCT, usually 20 8C; T c NOCT is the cell temperature in case of NOCT, usually 40-50 8C; and in case of NOCT a wind speed of 1 m/s on the solar cell surface is also assumed [9,10,24]. Therefore, based on the equations and considerations described above, a correlation between the cell temperature and the ambient temperature can be established [9,10,15,19,24]:
Implementing the simulation program
The program is structured into several blocks, which are [ The first four blocks of the program are unambiguous and merely involve substitution into the equations that were described along with the model. To solve the implicit equation of the current, the following are required: create a target function from Eq. (1) (Eq. (15)) and look for the zero value (or root) of this function where I is the variable. Using MATLAB's 'fzero' command the root of a target function can be found rather easily [2,12,22,24]: To determine the voltage-current curve of a solar cell, the output current (I) needs to be determined for the entire voltage range 0-U oc . To solve this problem numerically, it is sufficient to define a cycle which repeatedly searches for the root of Eq. (16) at a given resolution (U step), and registers the amperage for that given voltage [2,3,8,10,15,17,22,24,27,28].
Since the behavior of the solar cell's electronic parameters is also investigated during the transient temperature stage, it becomes necessary to use the simulation model in this way as well. During the transient temperature measurements, the illumination is constant [3]. The U oc and the I sc of the unloaded solar cell are recorded with varying temperatures. With the help of the mathematical correlations stated above this phenomenon can easily be described.
To be able to do that Eq. (1) just needs to be rearranged for the short-circuit and the open-circuit cases. The equations for the U oc and the I sc can be written as [8,24]: The previously stated correlations can be used to solve the described equations, with which the photoelectric current I ph , and the saturation current I 0 of the diode can be calculated while considering the effect of temperature [2]. The transient temperature calculation method is also built in MATLAB environment. Along with the already requested solar cell properties and irradiation values, the program also requests temperature values, with which it solves Eqs (16) and (17) for each temperature value. This task is feasible by implementing a 'for loop' to the program code. The transient temperate simulating program did not receive a unique graphical interface [2,8].
As a result of computer simulations, in addition to voltage and current values, theoretical and real power values are also determined. The theoretical performance of a solar cell is calculated from Eq. (18), and the real power of a solar cell is given by Eq. (19) [8].
During the investigation of the transient phenomenon, the correlation between theoretical power and temperature is determined from Eq. (18). In addition to the voltage-current characteristic of a loaded solar cell, the voltage-power characteristic can also be plotted by the cyclic solution of Eq. (19).
THE EXPERIMENTAL COMPOSITION
As a precursor to this research, a standard solar simulator was developed. Requirements for solar simulators are managed by American Standard for Testing and Materials (ASTMs) E972 (IEC 60904-9) [1]. The solar simulator implemented in the current research is a standard Class C, so both spatial non-uniformity and temporal non-uniformity are below 10%. The light intensity distribution of our sun simulator has a 9.96% inhomogeneity, which means the device complies with the standard.
The temperature of the solar cell is controlled by a cooling module made using Peltier modules [1]. The temperature of the solar cell is measured by a Voltcraft PL-125-T4 four-channel digital thermometer, furthermore current and voltage measurements are performed by two METEIX MX 59H digital multimeters. Figure 2 shows the experimental arrangements. This is reasonable as the cell in the investigation area is illuminated by 36 LED units in addition to the 8 halogen lamps. With an average irradiation of 1,000 W/m 2 , the cell temperature steadies at 88 8C. The whole investigation is carried out over a period of roughly 20 minutes, since steady-state temperature values are obtained at each of the three measurement points by then.
COMPARISON BETWEEN EXPERIMENTAL AND NUMERICAL RESULTS
The results of the simulations are plotted against the experimental results so that the difference between the measured and the simulated values is shown, thus showing the correctness of the simulation. The temperatures recorded during the measurements are used to calculate the temperature transient. Parameters used during the simulations are obtained from the solar cell's product data sheet. Open-circuit voltage, short-circuit current, and theoretical performance are plotted against time/temperature in case of three different heating curves (no cooling, half cooling, full cooling). The graphs show measured and simulated data simultaneously under STCs. Figure 3 shows that under STC conditions and without cooling the cell surface temperature reached steady state at 70 8C [20]. In case of half cooling, the maximum steady-state temperature of the cell was reduced by 10 8C, and by 18 8C under full cooling. The experiments and simulations were also performed under Non-Standard Test Conditions (NSTCs) conditions, in which case similar results were obtained. Many other researchers received similar results, for example Singh et al. [29], Wood et al. [30] and Malik et al. [31].
Observing graphs in Fig. 4, it can be concluded that the results of the transient investigations are in good agreement with experimental results [30]. It can be observed in both the simulation and the measurement results, the curves of the chilled and non-cooled solar cell cross each other, just like in case of other researches: Chantana et al. [23], Singh et al. [29] and Malik et al. [31].
CONCLUSION
In summary it can be stated that the activity in matter of solar cell simulation and measurement results in a mathematical model based on the study of the relevant literature that can describe the operation of the solar cell. The correct operation of the model implemented in MATLAB was based on the results of our measurements. The model validation was performed by comparing the measured and simulated results. Validation can be said to be successful, but it should be mentioned, that while transient examinations showed excellent agreement, the simulations of the loaded solar cell worked with greater error compared to the measurement. There may be two reasons for this, on the one hand, the measurements have errors as well, and on the other hand the calculations of the loaded solar cell required more complicated solutions and influenced the upshot with larger errors. The main goal of the cooling is to improve the solar cell's energetic efficiency and to increase its lifetime.
The results of the experimental and simulation examinations clearly reflect that the cooling changes the solar cell power in a positive direction, so the basic assumption is correct. | 3,434.2 | 2021-04-24T00:00:00.000 | [
"Physics"
] |
Real-time reorientation and cognitive load adjustment allow for broad application of virtual reality in a pediatric hospital
Background: With a new generation of affordable portable virtual reality (VR), clinicians are discovering more utility for VR, while also identifying opportunities for improvement, such as the inability to reorient the horizon line during repositioning or transport, or modulate cognitive load in real time. Aim: At our institution, this lack of functionality prohibited or decreased VR usage in some clinical scenarios such as dressing changes with dynamic positioning. The purpose of this brief report is to describe the development and use of a VR application that is optimized for the healthcare setting and report historical effects of patients who utilized VR as supplement to Child Life procedures. Eligible affects per chart review included Happy, Relaxed, Anxious, Distressed, Unable to Assess. Materials and Methods: Given the need for real-time reorientation and cognitive load modulation, we created the Space Pups™ VR application. The experience was launched as part of the Stanford Chariot Program in the summer of 2017, and its usage was tracked through the electronic medical record and a VR application dashboard. Chart review was queried from 3 January 2018 to 9 August 2021 for pediatric patients who used VR with real-time reorientation and cognitive load modulation as a supplement to their Child Life interventions. Results: The Space Pups™ experience has been successfully used in a variety of settings, including perioperative care, vascular access, wound care, and ENT clinic, a total of 1696 times. Patients ranged from 6 years to 18-year old, with no reports of side effects. Significant results (P<0.001) were observed pre- and post-VR use for affect improvements in Happy, Relaxed, and Anxious, but not for Distressed. Conclusions: The ability to reorient VR experiences in real time has increased functionality where other applications have failed. Relevance for Patients: While more studies are needed to quantify the anxiolytic and pain-reducing effect of Space Pups™, our report demonstrates the feasibility of this VR experience as a non-pharmacological modality to safely increase patient cooperation in a wide variety of clinical settings.
Introduction
Pediatric patients with untreated pain and anxiety during medical procedures may experience short-and long-term consequences, including post-traumatic stress disorder, needle phobia, and lack of trust of healthcare providers [1]. Virtual reality (VR) has emerged as a promising non-pharmacologic tool for reducing pain and anxiety in some adults and children [2,3]. Children, in particular, can benefit greatly from VR as a distractive tool to minimize attention to aversive stimuli by focusing on a 3-dimensional interaction of a computer-generated environment. Recent studies have citied large effect sizes in overall pain and anxiety reduction with VR use in pediatric patients undergoing a wide range of medical procedures such as venous access, burn, and oncological care [4,5]. VR has also been reported as an effective anxiolytic adjunct in a variety of settings, including phlebotomy, wound care, and chronic pain rehabilitation with additional potential to reduce the need for opioid-based analgesia [2,3,5,6]. Given the availability of low cost, portable, head-mounted VR units, clinicians have novel opportunities to integrate VR therapy into a wide range of clinical modalities. However, given the paucity of studies focusing on the clinical relevance of VR therapy within pediatric populations, it is imperative to further elucidate and address any shortcomings in a clinical context [7].
Most VR experiences are designed for users to play in a fixed viewing direction, usually seated or standing. This feature limits its application in healthcare settings given that many patients are supine. Furthermore, many patients experience fluctuations of noxious stimuli during encounters that include dressing changes, minor procedures, and phlebotomy. A VR application that provides healthcare workers the ability to modulate gameplay as stimuli change allows for increased distractibility at opportune moments. We introduce a novel VR experience that can reorient the horizon line in real time to allow for gameplay in any position with cognitive load modulation.
Development of VR application
The Space Pups™ VR experience was developed by a team of physicians, research fellows, and a software engineer through the Chariot Program at Lucile Packard Children's Hospital Stanford. Initial design considerations were developed in consultation with patients, parents, and pediatric psychologists. The process included five iterative revisions based on their feedback. During each round of feedback, adjustments were made to the software to (1) increase cognitive load demand during provider-initiated accelerated gameplay; (2) improve the gaming motivation to progress through levels; (3) optimize the audio accompaniment to the application; and (4) insert additional rewards for users based on game progression. Optimization of gameplay was completed when consensus was reached between the developers and providers that the cost of further iterations did not dramatically increase perceived patient benefit.
During the Space Pups™ experience, patients choose from one of five different pups that they steer down a highway in outer space, collecting treats from one of three highway lanes that synchronize with the beat of music ( Figure 1). The patient or clinician can change the horizon line orientation of the highway at any point during the game by triple swiping the side finger pad on a Samsung Gear VR or by holding down the App button (center button) on the Lenovo Mirage remote controller ( Figure 2). Patients were excluded from VR use if they had any of the following-significant cognitive impairment, history of severe motion sickness, current nausea, prone to seizures, visual problems, clinically unstable, or required urgent/emergent intervention. The experience is played either by gaze (lane changes are accomplished by rotating their head 5° left or right) or controller for patients with limited head mobility or whose heads are required to be still for clinical procedures, such as nasal endoscopy. Intuitive gameplay and minimization of menus allow the patient to be fully immersed in the game during minor procedures. The scoring system (point accumulation based on the number of treats retrieved) encourages patients to remain focused on the game to reach a high score. There is no fail state, allowing users to remain in the game continuously. In addition, a mechanism to transiently increase cognitive load for 12 s initiates a visual vortex and increases the frequency targets during times of increased pain or stress [6]. This is important to the game's mechanics because it allows for further distraction and cognitive load modulation during periods of potentially increased pain or anxiety.
Case
An 11-year-old boy was admitted to the patient care unit with a broken right leg, requiring multiple procedures and dressing changes. Written informed consent was obtained from the patient's guardians, as well as an assent from the patient. Space Pups™ was loaded in the VR headset. Pain and anxiety scores were monitored through self-reported questionnaires before and after VR use. The VR headset and game were initially presented to the patient as a source of entertainment to increase comfort during his recovery. After the patient became comfortable with Space Pups™, it was used to alleviate pain and anxiety during wound dressing changes. The horizon reorientation allowed the patient to continuously play while he was being moved from supine and fowler position during his dressing changes. During periods of potentially increased pain and anxiety, the providers would initiate the visual vortex within the gameplay to increase cognitive load and further immerse the patient. The VR headset was used daily by the patient for 4 weeks during his recovery.
Effectiveness
Through partnerships with Child Life Specialists, nurse specialists, and physicians, the Chariot Program provided a combination of Gear VR and Lenovo headsets to multiple healthcare environments within Lucile Packard Children's Hospital Stanford. Through electronic medical record (EMR) integration with the Child Life note and the headset interface alongside self-reported questionnaires, side effects and usage were analyzed.
After obtaining IRB approval, the EMR was queried from 3 January 2018 to 9 August 2021 for patients who used VR as a supplement to their Child Life interventions. Patient affect was measured pre-and post-VR utilization by Child Life Specialists. Chi-square tests for equivalent proportions (P<0.0001) were performed to determine affect count differences between both groups (pre-vs post-VR).
Results
Before using the Lenovo VR headset, the patient had never experienced VR. Despite its novelty, he learned how to choose a character and play the game within minutes, with limited instruction. By intentionally developing the game with limited head movement, nurses were able to successfully change his dressings while he remained in game play. The patient explained that pain and anxiety during dressing changes were significantly lower when using the headset. The patient's parents stated that they would recommend VR for other anxious children and expressed how thankful they were that their son was able to use the VR headset.
Space Pups™ has been used in eight different clinical settings a total of 1696 times with an average in-application duration of 4.56 min ( Table 1). There have been no reports of nausea, motion sickness, or dizziness on review of Child Life notes and selfreported questionnaires. Significant differences (P<0.0001) in affect pre-versus post-VR use were observed between Happy versus Not Happy, Relaxed versus Not Relaxed, and Anxious versus Not Anxious responses with no significant differences between Distressed versus Not Distressed and Unable to Assess versus Able to Assess (Figure 3).
Discussion
This report highlights the successful application of VR in the clinical setting, utilizing horizon line reorientation and cognitive load modulation. By increasing cognitive load and reducing attention to aversive stimuli, the patient reported reduction of subjective pain intensity, and in-application reorientation allowed for uninterrupted care. The application has been widely and successfully used in a variety of settings.
Although VR and other immersive technologies have previously been studied as non-pharmacologic adjuncts for the modulation of pain and anxiety in clinical contexts, the inherent limitations of many VR games -specifically, the inability to reorient the horizon for patients in non-standard gaming positions -limit the patient population that can safely and easily utilize this technology. Cognitive load modulation has previously been shown to subjectively reduce one patient's anxiety during a vascular access procedure [8], but its use has not yet been extended into broader clinical contexts. To the best of our knowledge, this is the first report showing the broad clinical applicability of horizon line reorientation and cognitive load modulation in VR with chart review usage to support the clinical efficacy of VR as a means to increase positive affect in various clinical settings.
Future studies will quantify the anxiolytic and analgesic effect of Space Pups TM . This case study demonstrates the use of this VR game as a non-pharmacological treatment to distract patients during wound care. The game can accommodate different clinical contexts due to horizon line reorientation and provides increased distraction during increased cognitive loads. | 2,486.4 | 2021-11-06T00:00:00.000 | [
"Medicine",
"Engineering"
] |
Analyzing Methods to Achieve Successful Development
While it is easy to set worldwide development goals, achieving and meeting those benchmarks is not so simple. The goal of many nongovernmental organizations (NGOs) is to spread development to communities. Two months were spent in Arusha, Tanzania interviewing NGOs which complete projects in many disciplines ranging from water and sanitation to energy and agriculture. The NGOs were asked how they define the success of a project, about challenges encountered and lessons learned, and to discuss any monitoring strategies they may have. The interviews revealed that many organizations have not established metrics of success nor do they have well-defined procedures for evaluating their projects. This leads to a lack of focus on monitoring following the implementation of a project. However, many donor organizations and agencies have recently added monitoring and quantitative metrics of project success. The most frequently identified challenge was regarding community interaction. Community participation is a challenge for various NGOs; however, many NGOs explained that through increased community interaction success of a project amplifies. These survey results can be used as a guide for engineers interested in entering the development sector, or by those who are already involved in it. Index Terms – international sustainable development, reflection, service learning, qualitative survey
INTRODUCTION
Development remains a concept and not a reality for ninety percent of the world's population.While most businesses, schools, and technological advances focus on the richest ten percent of people, there are some organizations that exist to serve the poorest ninety percent.These organizations are governmental, nongovernmental, and for-profit groups. 1 Many of the projects these groups work on are successful, but others fail.Past studies have demonstrated that
INTERVIEW METHODOLOGY
The interview questions were created to allow participants to provide detailed qualitative responses using personal experiences.The goal of qualitative interviewing is to prevent guiding the subject of the interview to a specific answer.Therefore, the flow of interviews was flexible based on the conversations that arose. 6The interview questions and procedures were reviewed by the Michigan Tech Internal Review Board (IRB) to ensure confidentiality and safety of participants.
Interviews were conducted in a variety of settings, with many interviews conducted at the NGO office or project location in the Arusha area.The interviews were recorded using voice recording equipment, and afterwards the interviews were transcribed and the results were discussed and analyzed.The phrasing of interview questions and order was revised following the second interview.These adjustments took place based on responses from the initial NGOs interviewed and an increased understanding of the Tanzanian culture and interviewing methods.
SUMMARY OF NONGOVERNMENTAL ORGANIZATIONS INTERVIEWED
A variety of NGOs were interviewed.Their focuses varied but included engineering technologies, agriculture, education, vocational training, and health awareness.A description of the ten NGOs interviewed is included in Table 1, along with the title in which each NGO will be identified throughout the report.
Successes
Success is often difficult to define and can vary between projects or NGOs.However, it is vital that organizations set end goals in order to eventually determine their successes.The World Association of Non-Governmental Organizations suggests in their recommendations for starting an NGO that the first necessary step is to establish a purpose, vision, and short-and long-term goals for the organization. 7It is also important to set realistic goals, since no organization could ever solve all of the world's problems; it will take millions of small efforts. 8The survey asked participants to explain how they define success, in order to help the research team gain a better understanding of the expectations and areas of emphasis of the groups actively working on projects.This often led into discussing specific goals set by the NGO.NGO 1 has defined success for implemented projects and also for the overall organization.Success is defined by each individual person that participates in a project.The NGO focuses on the values and benefits of impacting one person's life positively, and success is determined by observing that change.For instance, after implementing a project focused on education of women on healthy eating habits, the NGO observed a change in groceries purchased by many participants.Furthermore, while creating goals for a project the NGO actively involves the community so that participants can understand the final impacts of a project.However, NGO 1 warns that there must be a balance between the NGO's community expectations and the willingness of the community to participate in the project.Setting realistic goals for the community and thoroughly explaining project expectations, results and benefits will aid in active participation in a project.This will also prevent any "negative attitudes" that the community may eventually form.
Another way in which NGO 1 has ascertains success is through acknowledgement from governmental agencies and peers.Evidence that this NGO has completed successful projects includes multiple awards from municipal governments.Additionally, the organization has been invited to numerous conferences to share their story.This demonstrates that success can also be defined by influencing other groups to allow the work of one NGO to impact people beyond its original scope.NGO 2 explained that they are better able to determine community needs and identify potential projects by conducting a baseline surveys prior to project implementation.The information obtained is then compared to a similar survey that is conducted after the conclusion of the project.The organization believes they are then able to determine whether or not the project actually met the needs of the community.NGO 2 also partners with other NGOs during projects to better utilize the strengths of each organization.The support and effectiveness of this partnership is also used to define the success of the project.
NGO 3 only recently began to measure the success of their projects.Although the organization has been established for over 15 years, monitoring is a new concept and the meaning of success had not been previously analyzed.However, a recent demand from funders has lead to a focus on quantitative results including health statistics and crop yields.NGO 3 acknowledged the benefits of setting goals, so that success can be identified, instead of following their previous generic goal: "let's help as many as we can." While establishing quantitative goals is a new concept to NGO 3, not every NGO has embraced this as the true definition of success.NGO 4 explained that their organization is not "exactly measuring success."The NGO conducts life skills education classes, which discuss reproductive health, communication, decision making, goal setting, and other important skills.The course is based in the U.S. Peace Corps life skills curriculum. 9Even though the NGO was reluctant to define success, they did explain that at the end of each term the students complete an evaluation form.Within that form students are asked to discuss lessons that were of particular interest and lessons they will never forget.Success then can be identified by the responses of the students.NGO 4 considers a project successful "[i]f every kid can at least list one or two things that they'll never forget." In summary, both NGOs 1 and 4 analyze the overall success of a project, and their definitions of success are driven by the impacts to the individual participants.Additionally, NGO 2 evaluates success in a similar manner, though at a community level, with success equated to fulfilling a need that the community had previously identified.For NGO 2 success is measured by conducting community surveys and evaluating the strength of partnerships.A summary table of selected interview NGO responses regarding success follows (Table 2).The vision desired is realized.An example includes observing changes in eating habits to include vegetables and other nutrients NGO 2 People eat more and that the village's food insecurities diminish NGO 3 Depends on the issues being addressed but a real need must be identified, not a precooked idea NGO 4 Success is achieved when students retain lessons taught during class and can list the gained knowledge NGO 5 Adoption of the project NGO 6 Community integration and education NGO 7 Establishing financial independence and the street children holding jobs after graduation NGO 8 The women's ability to use the life skills taught after leaving the center
Challenges
Numerous challenges arise throughout the implementation process for any project.However, these challenges are rarely discussed in detail. 10To expand on the existing understanding of challenges faced by NGOs, the survey included a question regarding obstacles encountered.Many lessons can be learned from analyzing challenges, and the knowledge gained can be used to avoid similar obstacles in the future.NGO 4 explained numerous challenges that they have come across, most of them centered on cultural differences and employees not completing their assigned duties.More specifically, NGO 1 identified finding the correct participants and explaining individuals' roles as a challenge.Similarly, NGO 6 stated, "You have to find the right people, the right people who will accept the message."The project participants must understand the "motifs" and goals of the implemented project from the beginning.NGO 1 has been able to overcome this challenge by thoroughly explaining the project and integrating it into the community.
NGO 3 expressed a similar sentiment-finding the right people to take on varying roles of responsibility is crucial-and provided an example.For their chicken vaccination program they train one vaccinator in each village.There are some perks to this job, including a small stipend and access to a bicycle to complete the work.The vaccinator is selected -according to local custom -by the village leadership.There have been a few instances where the correct individual for the job was not selected, instead a relative of one of the leaders is chosen.However, they have had some extremely successful vaccinators selected.NGO 3 attributed this success to establishing their organization in the community and meeting the villagers early in the process.Then, they can offer more suggestions for which person would be the best candidate for the job.
Conducting work remotely has been a challenge for NGO 5.The communities that they serve are located outside of Arusha, which is the location of their headquarters.They frequently travel to complete field work and realize that the day they selected does not work for the village due to a holiday, funeral, or other event.This costs them money and time, delaying the success of the projects.Through establishing closer relationships and becoming more integrated with the communities, this challenge could be overcome.NGO 6 also explained that this has historically been a challenge for their organization.Consequently, they have decided to relocate to be able to have constant communication with their participants.NGO 5 explained that implementing a project that takes three years to reach completion has presented numerous challenges.They stated that it is difficult to convince people that the project will benefit them unless there are immediate benefits to participants.Working within the agricultural sector, long time frames for projects has been particularly challenging for their organization.To overcome this obstacle, they conduct education sessions.However, they have been able to observe greater adoption of jatropha plants, which have a three-year maturity period, through introducing additional crops, such as sweet potatoes, that produce significant yields within one season.
NGO 7 explained that whenever a challenge arises they move quickly to solve and eliminate it.Although this process may work for this NGO, they do not keep any record of these challenges and methods to overcome them.This could perpetuate repeated problems if there is significant employee turnover.
The challenge that was most frequently identified by the interviewed NGOs was a lack of cultural understanding and community integration.These are two factors that are commonly identified as requirements to complete a successful project. 11Although these needs are well known, the challenge of assimilation and understanding perpetuates in active projects in Arusha, TZ.By carefully selecting the communities in which to complete work, and establishing close community contacts, this challenge could be greatly reduced.A summary of challenges encountered by selected interviewed NGOs is shown below (Table 3).
Monitoring
The number of NGOs involved in development aid has increased greatly in recent years.With this growth, identifying the achievements of NGOs has been a developing concern. 12Monitoring of projects should be a regular process in development work in order to keep track of project progress and adjust the plans if necessary to improve efficiency.Effective monitoring practices can also lead to improvement in program management, strengthening of capacity building, and better alignment with the expectations of donors. 13Even though continued follow-up and monitoring of projects is stressed frequently in literature regarding development 14 , it has generally been difficult for organizations to reach persuasive conclusions about their work.This may be due to ambitious expectations, complexity caused by scale, activity diversity, vague objectives, or the absence of baseline data. 15The interviews contained a question regarding monitoring after a project is completed to determine if and how existing NGOs complete this stage of the development process.A few NGOs interviewed had some sort of monitoring plan in place, but many did not.It is argued that without an established presence in a community, projects will have a decreased chance of success.
NGO 8 provided a good example of monitoring.The NGO maintains contact with participants for one year after they leave the vocational center.They are required to report where they are living, their employment status, and their earnings.This allows the NGO to measure the effectiveness of their training and ensure that the graduates of their program are able to maintain healthy and successful lives.The information gained is then used to make modifications to their program.The NGO has served 167 young women in three years.
NGO 5 completes monitoring in two stages.Working in agricultural sector, the NGO establishes a three-year commitment to the community.After the three years of implementation, initial monitoring takes place.Then, three months later they return to the community to conduct the second stage of monitoring.Their focus is on adoption of the plants they introduce and use of the crops to create a means of business for the village.Expectations from donors have dictated NGO 4 stressed that flexibility is key while completing development work, and they noted that the majority of the challenges that they had encountered were overcome by adapting and being flexible.NGO 4 also explained, in the context of a life skills class taught to secondary school children, that repetition of concepts during educational programs needs to take place.This class covers a range of topics, many of which are health-related.Through repetition, each time in new innovative ways, the students have been able to better understand and remember the material covered.
NGO 1 cited that lots of energy is crucial for a successful project.Since there is a significant amount of work associated with every project, and often a large amount of information that needs to be refined and explained to the community, energetic individuals are needed to drive the process.The NGO also cited the importance of acknowledging the value of making a difference in individual lives.However, it was observed through talking with NGO 1 that much of the work is completed by one individual.This could be problematic in the long term sustainability of projects since one person holds all of the knowledge of the projects and operations of the NGO.Additional lessons learned expressed by selected NGOs follows (Table 4).If a project is long term, short term projects are also necessary in order to encourage community members and earn their trust 7 Strict discipline is important and making sure everything is locked so that stealing is not an issue (speaking about a boy's school)
IMPLICATIONS FOR SERVICE LEARNING PROGRAMS
Many organizations can benefit from the knowledge gained from these interviews with ten NGOs.The organizations include Engineers Without Borders-USA, Engineers for a Sustainable World, and other student and professional groups.This information should be used to promote more effective communication and partnerships between these groups and NGOs.First, it is vital that the partnering NGO has a strong presence in the community in which the project will take place.Since time in-country for service learning groups is often short, the NGO's responsibilities should include understanding the community and establishing a strong relationship with community leaders.Since the most valuable lessons learned and responses to challenges faced involve community interaction, this point is essential.If a service learning group is not able to form an appropriate partnership, with an NGO or other community organization, a different project should probably be sought, unless the group is prepared to devote substantial time towards gaining community understanding and respect.
Next, if a project needs to take place over a long time span, such as with the Jatropha project, other smaller projects should be implemented in the same communities that can produce results more quickly, such as within a few months or within a single growing season.The community will have more faith in the long-term project, and they will be more likely to respect the knowledge of the service learning program participants, thus increasing the prospects of success.Setting measurable goals is another important aspect of achieving successful development.These goals should be both quantitative and qualitative.Additionally, in order to evaluate whether the goals were achieved, monitoring must be completed over an appropriate amount of time (varies dependent on the project).Measuring the influences of a project, such as by evaluating design performance and health impacts, will allow future projects to be completed more efficiently.
Overall, two of the most important skills to utilize in a project are communication and flexibility.Service learning programs should be able to communicate effectively with an NGO and the community they are working in.They should also have the ability to be flexible in every part of the project, including assessment, implementation, and monitoring.By taking into consideration communication, flexibility, and the other lessons learned, the prospect of a project's success will greatly increase.
CONCLUSION
After conducting ten interviews with development NGOS, a greater understanding of methods to achieve successful development has been gained.While some organizations were hesitant to openly discuss sensitive issues such as challenges encountered, others candidly explained the specific challenges that have plagued their organizations and ways they have attempted to overcome them.
The most significant observation following the interviews is that many NGOs do not adequately reflect on the projects that they conduct.Those that do complete monitoring frequently have added this step in their projects only due to pressures from donors.This lack of monitoring was also linked to a lack of goal setting at the initial stages of a project.By establishing a defined direction for a project, communities may be better engaged, and more people will likely be served.Additionally, by documenting challenges and successes, organizations will e able to troubleshoot common problems and have a greater chance of achieving success in future projects.
Success is an elusive concept that can be defined in many ways.However, NGOs frequently used common words, including "health", "community", and "individual", which indicates that success must be assessed quantitatively as well as qualitatively since these attributes require an in depth understanding of community impacts.This has presented many challenges to the NGOs interviewed, which frequently do not set measurable goals or complete follow-up monitoring of projects.However, by establishing goals and a strong presence in a community, by being flexible and adapting to change, and by including a strong monitoring program in every project, many of the challenges faced by the NGOs may be overcome.It is hoped that by reflecting on the experiences of practicing NGOs working in the development field, successful development projects can be more easily attained.
ACKNOWLEDGMENT
This material is based upon work supported by the National Science Foundation under Grant No. OISE-0854050.The authors are grateful to the NGOs that participated in this study.Dr. Kari
TABLE 1 DESCRIPTION
OF NGOS Identifier Description of Work NGO 1 Works with environmental conservation, nutrition encouragement, and water quality in the Arusha area.NGO 2 Monitors the implementation of rural agricultural projects.Works with NGO 3 and 5. NGO 3 Works with projects in rural communities including a chicken vaccination project, HIV/AIDS awareness project, and sustainable agriculture projects.NGO 4 School for orphans and disadvantaged youth.NGO 5 Works with rural farmers to promote planting and use of jatropha seeds to obtain oil for soap, cooking and lighting.NGO 6 Provides health education to rural communities, especially focused on HIV/AIDS.NGO 7 Works with street children and conservation efforts.NGO 8 Community based organization that provides housing and vocational training for young women, focuses on single women with one child.NGO 9 Works on projects with land conservation, HIV/AIDS awareness, animal research, and entrepreneurial training.NGO 10 Provide individual farmers and communities with livestock, trees, and training.They also provide HIV/AIDS education to communities.
TABLE 2 :
SELECTED DEFINITIONS OF SUCCESS BY NGOS
TABLE 3 :
KEY CHALLENGES IDENTIFIED BY NGOS Realizing that the village leaders have a huge role in the projects and working with them to help them make the best decisions for their communities NGO 4 Reliability of coworkers and setting fair and official policies NGO 5 Integration with community and gaining trust of community when working on multi-year projects NGO 6 Obtaining funding and working within cultural traditions NGO 7 Discipline NGO 8 Funding, finding volunteers and insufficient space to meet the needs of the surrounding communities NGO 9 Working with policymakers
TABLE 4 :
SUMMARY OF NGO LESSONS LEARNED | 4,851 | 2011-05-07T00:00:00.000 | [
"Environmental Science",
"Economics"
] |
Physical Education Teaching Based on Human Model Design and Computer Simulation of Human Two-Dimensional Moving Image
The traditional gait analysis is mainly used in the research of joint dynamics and electromyography, which is helpful to develop a medical feedback system. The system judges the leg injury by analyzing the patient’s gait, so as to provide effective treatment. In order to study how computer technology can be more appropriately introduced into physical education, this paper puts forward a new design method of manikin. The application of some key implementation technologies of computer simulation of two-dimensional motion images of human body is conducive to the scientization of physical education teaching and training. The experiment analyzes the daily movement of human body. MEMS technology is a technology to obtain human inertial data by simulating human motion with one or more computers. By integrating the collected data, different characteristics can be obtained. Then, classify these feature processing algorithms, and finally, identify the actions performed by the human body. Finally, the advantages of this recognition method in recognition accuracy are verified.
Introduction
e core goal of computer graphics is to create e ective visual communication. In the eld of science, graphics can display scienti c achievements to the public through visualization. In the eld of entertainment, such as PC games, mobile games, 3D movies, and movie special e ects, computer graphics is playing a more and more important role. Graphics also plays an important basic role in creative or artistic creation, commercial advertising, product design, and other industries. In the eld of science, this point was highlighted in the 1987 visual report on scienti c computing. e report cites Richard Hamming's classic assertion in 1962: "the purpose of calculation is to gain insight into the nature of things, not to obtain numbers." e report mentioned the important role of computer graphics in helping the human brain understand the essence of things from the perspective of graphics and images, because graphics and images have stronger insight than simple numbers. "Computer graphics" is a related theory, method, and technique for generating object graphic output based on a model. A computer is used to build, store, and process a model of an object [1]. Computer graphics has a wide range of applications. One important area is the use of computers to simulate the process of systems or phenomena, for system analysis and process control [2]. ere are three geometric models commonly used in computer graphics: wireframe, surface model, and volume model. e single-line model of human body evolved from the wireframe model [3]. Its description method is similar, and the principle is similar. e surface of polygonal surface model is very complex, which can generate various views and realistic graphics. Model information is an ideal model for the two-dimensional spatial wire frame of a simulated object, which can be used to evolve into a single reasoning model of human body [4]. It is scienti c and feasible to use computer graphics model to build a suitable model for transfer motion. However, scholars have not done much research on this issue. e design technology of human model simulating two-dimensional human motion on computer has become the focus of research [5]. Taking gymnastics as the research object and computer simulation as the design method, the computer simulation gymnastics graphics movement teaching system was developed in one and a half year.
The Research Status of Human Body Model
According to Hanna's body model theory and textbook schematic diagram, the human body is understood [6]. Similarly, the human model can be divided into 41 parts. In computer graphics, data structure is essentially a data file, which is used to generate all the graphics that the object describes. ese descriptions include geometric information and topology information that defines the shape and size of objects for all components and is used to describe information about objects that are related to the object, such as color and skin quality [7]. e common data structure is a common data file, and its description information is universally applicable. For example, the head consists of four polygons, while the polygon is composed of dozens of vertices [8]. e description of the head link model should include the following information: the general data structure of the joint center (part of the head of the shoulder center ring) coordinates, the link angle joint center (the attachment of the centroid and the active X-axis) coordinates, the center vertex of the distance (distance) of the vertex, the vertex angle (the vertex distance and the positive angle of the axis), the number of vertices, the color of the filling, and the color of the line [9].
Bilateral Filtering Algorithm Based on Hidden Markov
Model. Two-dimensional human models have broad application prospects in modern medicine, virtual reality, and other fields. Bilateral filtering technology has become a hot research topic. As a nonlinear filter, it plays an important role in solving the pixel values of spatial similarity. is technology also needs to consider the influence of spatial noise to remove gray similarity [10]. In general, the detail protection that can clearly penetrate the blurred edge is not obvious. Research needs models with simple, noniterative, and local characteristics. Second, the bilateral filter can just meet the above characteristics [11]. General Gaussian blur mainly considers the spatial distance relationship between pixels, but does not consider the similarity between pixel values. erefore, the blur result we get is usually the blur of the whole picture. e improvement of bilateral blur is that when sampling, we not only consider the relationship between pixels in spatial distance, but also consider the similarity between pixels, so that we can maintain the general block of the original image and then maintain the edge. Due to the high frequency of high-frequency human motion signal, the high-frequency noise in color image cannot be completely filtered out. In digital image processing, only low-frequency human motion signals can be filtered better [12]. In digital image processing, bilateral filtering algorithm is often used for image denoising. It is a nonlinear filtering method, which considers the adjacency between pixels and human gray motion signals. e bilateral filters used in digital image processing are where q is a pixel, k is an adjacent pixel of q, N(q) is a set of q adjacent pixels, ‖p − k‖ is the Euclidean distance between q and k, and L(q) is the gray value of q. |L(q) − L(k)|. It is the grayscale similarity between q and k, and both W c and W s are weight functions. In order to introduce the bilateral filtering algorithm from the digital image processing to the scattered point cloud denoising, a visual plane is first defined: for the neighborhood point set N(P), the three-dimensional space R3 is described as R3 � N ++ S2. N is a one-dimensional space along the normal direction at the P point of the neighborhood point. S2 is a two-dimensional tangent plane over P points. + is straight sum operation. In the local scope, S2 is defined as the visual plane, while the neighborhood point is defined as the projection point on S2. e distance from the neighborhood point to the pixel is defined as the gray value of the pixel. e bilateral filtering algorithm is defined as where qʹ is the filtered point, q is the point of the original point cloud, n is the normal vector of point q, and α is the bilateral filtering factor. e formula for calculating α is as follows: e definition of the smoothing filter function W c (x) and the feature-preserving weight function W s (y) is σ c and σ s are the Gauss filtering coefficients of the plane of view, namely, the weight of the spatial domain and the weight of the feature domain. σ c describes the effect of distance from point q i to neighborhood point on this point. σ s describes the influence of projection from point q i to neighborhood points on the normal n i of q i to point q i . e larger the σ c , the better the filtering effect. e larger the σ c , the better the cloud motion characteristics of the human body. e two influence each other. K-nearest neighboring point culling algorithm and K-nearest neighbor-based outlier culling algorithm are selected. e algorithm actually takes the idea: for any point in the point cloud, that is, the average distance between the point and the corresponding nearest neighbor points, if the average distance from the whole point to the K-nearest neighbor is in accordance with the Gauss distribution, the average distance to the K-nearest neighbor point at the point of M + D is treated as the outlier point and is removed. e rst step is to properly initialize the network structure so that each weight w ij and threshold θ j of the neural network are initialized between 0 and 1 ( Figure 1). en, the square error SSE of the network error and the maximum number of iterations are set to 0, and the number of iterations is calculated one by one. e second step is to extract an input vector X and its corresponding output vector T from the training set randomly. e third step is the forward propagation input process. e net input vector I j of each neuron is calculated relative to the previous layer, and a continuous derivable function is selected as a transfer function (such as logistic or sigmoid function). en, the output vector O j of neuron J is mapped to 0 and 1 according to e fourth step is to calculate the error squared sum SSE of the neural network, as shown in e fth step is the reverse propagation process. e error vector ERR j of each neuron is calculated according to the expected output value O j , as shown in e sixth step is to dynamically adjust the weights and thresholds of the vectors in the neural network according to the formulas (9) and (10), in which α is the learning rate.
e seventh step is that if SSE is less than or equal to the target error, it means that the network is convergent. If t M, it indicates that the network is not convergent. When these cases occur, the iteration process ends. Otherwise, it will continue to be executed from the second step. BP network adopts a guided learning method, and its learning includes the following four processes. (1) e input mode consists of the "mode forward propagation" process from the input layer through the hidden layer to the output layer. (2) e error signal of the di erence between the expected output and the actual output of the network is the "error back propagation" process of rounding the connection weight layer by layer from the output layer through the hidden layer. (Reverse calculation process) (3) network "memory training" process is repeated by "mode forward propagation" and "error inverse propagation." (4) e network tends to converge, that is, the "learning convergence" process in which the overall error of the network tends to the minimum. In the training phase, the training instance repeatedly passes through the network, modi es each weight, and changes the connection. e initial set of corresponding points is predicted. Reasonable calculation of the characterization factor is implemented correspondingly to the important points obtained. Its position in point cloud is discussed in detail. Based on their characteristics and location similarity, the relationship between them is predicted. A preliminary estimate of the response point pair is implemented, and the initial corresponding point set is clari ed. e error point must be reasonably removed.
Model Correction Method.
e commonly used data mining techniques include statistical analysis class and knowledge discovery class. One is statistical analysis. e data mining models used by statistical analysis technology include linear analysis and nonlinear analysis, regression and logistic regression analysis, single variable and multivariable analysis, cluster analysis, time series analysis, and so on [13]. First of all, these technologies can be used to check which human model information is abnormal human model information.
en, a statistical model or a mathematical model can be established to analyze the information of these human models so that the rules and information hidden between the information of the human body model are found. Photos and videos based on or containing real human bodies are used as objects of data input and postprocessing. We choose the existing models in the human model library of the existing human modeling commercial software PR as our analysis. e reconstructed and adjusted object is used to replace the digital model obtained by scanning the real human body. e main purpose of this is to make a compromise between the cost of building the system and the authenticity of the system results. For example, in order to increase market share and pro t, general statistical analysis tools can be used to seek the best business opportunities. A comprehensive quality management system is utilized to improve the quality of products or services, and customer satisfaction and user experience are improved. Or by restructuring business processes or reshaping the pipeline structure, the pro ts of enterprises can be increased accordingly. e data mining technology of statistical analysis is one of the most mature data mining technologies at present, which has been widely applied [14]. e two is knowledge discovery technology. Knowledge discovery class technology is based on human model information driven, which can discover business patterns from human model information in human model information warehouse. is kind of data mining technology does not need to be like the statistical analysis data mining technology in the analysis of human body model information. What is the target variable before starting the analysis, and what information needs to be mined must be clearly known; otherwise, it is di cult to successfully mine. Knowledge discovery technology mainly includes neural network, decision tree, association rules, and so on [15]. In order to apply the knowledge of data mining in practice to improve production, the model generation management subsystem in the human model building system has been optimized. New modules have been added, and the corresponding functions are realized. e production management system adopts B/S (browser/server) mode. Based on the MVC three-tier architecture developed in ASP.NET environment, SQL SERVER 2008 is used as the background human model information base management system. e functional implementation of the system is implemented on the server side. e client can get the result of server-side operation through browser, so as to provide convenient and quick interaction. MVC (model view control) divides the software system into three parts: model, view, and control, as shown in Figure 2. e view layer is the interface between the user and the whole system to complete the interaction between the system and the user, for example, data entry, deletion, modi cation, or query result display and summary of various kinds of data.
Action Signal Acquisition.
In this paper, the analog circuit is designed to collect the surface EMG signals on human arms and legs, and lter the signals, including low pass and high pass. After ampli cation, the analog-to-digital conversion is realized through the A/D of American Delsys EMG processor, and then simple digital ltering and processing, drawing surface EMG (sEMG), feature extraction, pattern recognition, and so on. Identify the simple actions of a part of the human body. As shown in Figure 3, di erent testers have di erent actions and di erent test results. After data normalization is processed, Figure 4 is obtained. e ordinate of Figure 4 is the number of times, and the abscissa is the time of each action, which lists several time curves to get the action, such as the BPF curve and the M curve. Figure 5 shows the relationship between the residual energy and the number of principal components in the principal component analysis.
In the actual system, the signal extracted by means of signal preprocessing often has small amplitude and large noise, which is easy to be disturbed. erefore, it is necessary to isolate and lter impedance transform and other means to extract and amplify the signal because the human motion signals collected by motion capture system are doped with some noise during the experiment. e number of extracted features is large. erefore, in order to ensure the accuracy of the research results, it is necessary to process the original signal accurately, feature extraction and selection, so as to Mathematical Problems in Engineering reduce the noise and reduce the recognition rate on the basis of reducing the number of features. First, the noise contained in the original action data is analyzed. By using principal component analysis, one-dimensional features are selected and merged into high-dimensional features. In the speci c operation, the classi er's feature vectors are trained according to the formula stochastic gradient descent algorithm. e loss function is calculated according to the formula, and the classi er is trained if the stopping condition is reached; otherwise, the procedure is transferred. In traditional Taiji movements, there are many repetitions of movements, complicated routes, and emphasis on the right frame. e practice time is about one minute for a routine, which is not conducive to popularization and other drawbacks. e Tai Chi movement is simpli ed, the arrangement is re ned, and nearly half of the repetitive movements are reduced. e route is two back and forth on the two point one line. ere are four ways: straightforward, backward, sideway, and turning.
Hidden Markov model (HMM) is a statistical model, which is used to describe a Markov process with hidden unknown parameters. e di culty is to determine the hidden parameters of the process from the observable parameters.
en, these parameters are used for further Halt 71 5 3 200 93 Bend 70 6 3 200 89 Sit 86 6 8 200 91 Left turn 87 7 2 200 91 Right turn 87 9 3 200 92 Sideslip 85 8 3 200 87 Bow step 83 6 7 200 89 Normal walking 87 7 3 200 90 Up the steps 71 3 6 200 88 analysis, such as pattern recognition. In the modeled system, it is considered as a Markov process and unobserved (hidden) state statistical Markov model. In Figure 6, q i represents a possible gesture. b i (k) is the probability of a specific gesture. e number a ij marked on the path of the state transfer is the probability of the occurrence of a possible movement attitude q i to another possible action attitude q j . When the pose number is N, the gesture transfer probability can be expressed by the matrix A � a ij of N × N.
e Analysis and Application of the Data Set of the Action.
In order to ensure the accuracy of the recognition results, the research needs to test everyone's physical education teaching mode behavior. e study used 20 different teaching behaviors and collected 10 groups of different data. e collected teaching behavior of each person is cross verified, and the state parameters are distinguished by principal component analysis and vectorization. Due to the limited number of training, the TV value in the experiment is too large. When the item value corresponding to parameter a is 0 or 1, the complexity state coefficient is fixed to 3 to 8. Tables 1 and 2 show the identification results when the number of states is 4 and 7, respectively. e effects of different state numbers on the recognition rate of physical education teaching were compared. Figure 7 shows the recognition rate of the number of states when the number of states is taken. It can be seen that with the increase of the number of states, the recognition rate is on the rise as a whole. When the number of states increased from 3 to 7, the recognition rate increased significantly, and after 7, there was little change. At the same time, the system runs as shown in Figure 8: en, based on the index system of algorithm and construction, the situation of university sports teaching informatization is evaluated. First, the algorithm is analyzed by factor. After processing the original data of 15 people's physical education teaching actions, they are programmed by MATLAB. e results are as shown in Table 3.
As can be seen from Tables 1-3, HMM can achieve good recognition results. e algorithm design is reasonable and feasible. At the same time, it shows that the pretreatment method and feature selection method are effective. It shows that the information collection of these physical education movements is better, and the extracted feature parameters can better reflect the teaching results. Although the Markov model has obtained good test results, there are still many problems to be further improved. After extracting the parameters of physical education teaching model, dimension reduction will result in the loss of some original information.
In order to train a good hidden Markov model, a lot of data training is needed. Because the training data are not enough, the hidden Markov model cannot accurately represent the characteristics of a certain sport teaching action. In the training and testing of the hidden Markov model, many parameters are selected by manual setting, and the results are not very good.
Conclusion
Computer technology is increasingly being introduced into physical education. On this basis, this paper puts forward a new design method of human model and some key implementation technologies of computer simulation of human two-dimensional motion image, which is helpful to the scientization of physical education teaching and training. e experiment analyzes the daily movement of human body. According to the number of vulnerable parts with the best identification effect, 10 different actions are summarized. Different characteristics can be obtained by integrating the collected data. en, these feature processing algorithms are classified, and finally, the actions performed by the human body are recognized. Finally, the advantages of the recognition method in recognition accuracy are verified. However, due to the lack of training data in this paper, hidden Markov model cannot accurately reflect the characteristics of a physical education teaching behavior. In the training and testing of hidden Markov model, many parameters are selected manually, and the results are not very good. erefore, after extracting the parameters of physical education teaching model, dimensionality reduction will lead to the loss of some original information. In order to train a good hidden Markov model, a lot of data training is needed for further research.
Data Availability
e data used to support the findings of this study are included within the article.
Conflicts of Interest
e authors declare that they have no conflicts of interest.
Acknowledgments
e Scientific Research Project of Hunan Education Department in 2020 "E-sports training method in the training of high-level football team practice research" (20C1013). | 5,314.4 | 2022-05-30T00:00:00.000 | [
"Computer Science"
] |
The concept of organizing educational literary tourism by means of information technologies as a region development tool
The article discusses the development of the cultural potential of the region through the development of tourism. Educational tourism was chosen as one of the potentially attractive types of tourism, as well as its private direction - literary tourism, which has specificity in terms of methodological, organizational and communicational components. In the presented concept of organizing literary tourism, the development of an information system is described that allows one to fill the database with all the necessary information to implement the educational goals of literary tourism, taking into account the needs of users. Due to the ability to automatically retrieve the necessary information the designed and implemented information system significantly saves user’s time and financial resources for organizing tourism, as well as automatically collects and stores data about the regional attractions found, which confirms the relevance of the study and the need to develop the described system, which has successfully passed testing and approbation. Prospects for further development imply a complete transition to web technologies in order to further popularize this information system among a large user audience of students and teachers both in megacities and in the regions.
Introduction
The most important area of the socio-economic development of the region is to involve society in this issue, including the popularization of the problems of cultural heritage in the educational sphere. This will solve several problems simultaneously: it will allow preserving old cultural sites by increasing the demand for them, expanding the list of objects visited by tourists including little-known and unknown previously, will form a new generation of people interested in preserving the culture of their region; partial commercialization will provide an inflow of funds to cultural and recreational facilities, contributing to their preservation and development.
Recently, tourism in Russia has been developing significantly due to the popularization of domestic direction and the significant economic effect of the industry both in megalopolises and in the regions. A significant number of local tourist sites became known to the general population and attracted attention to their preservation and dismantling thanks to informing about the socio-economic potential of Russian regions, as well as through the development of various strategies for organizing tourist decentralization and additional attraction of the flow of people to local cultural centers.
Among the various approaches to the classification of tourism according to the purpose of visit, one can single out the following categories of tourism activities related to sustainable development of territories: 1 Thus, it can be concluded that the use of tourism in education is one of the areas of attracting society to the problems of region's socio-economic development. Nevertheless, the tourism implementation in the specifics of educational activities requires significant methodological study, which actualizes new research in this area.
The main implementation directions of educational tourism as a field of knowledge at the intersection of psychology, pedagogy and management are such functional components as: 1. Learning process management. 2. Organizational and communicative component management, based, inter alia, on the use of information technology in the field of tourism.
Thus, for the competent implementation of educational tourism, it is crucial to pay attention to the methodological, organizational and managerial support of this process, the main problems of the implementation of which in the field of education are: 1. Insufficient development of the theoretical base. 2. Insufficient degree of empirical achievements generalization. 3. Insufficient methodological study of the tourism and local history implementation. 4. Labor intensity of search work and research results fixation in a material format. 5. Weak plans for tourism activities development aimed at eliminating cultural and educational barriers.
6. Insufficient integration of the educational organizations' activities of with specialized travel agencies.
7. Lack of developed approaches to the tourist routes design, taking into account the educational specifics.
One of the directions of modern educational tourism development is one, that is specialized for the study of specific academic disciplines or related ones. An example of such an activity is literary tourism -the direction of cultural and educational tourism associated with the study of works of world classics, as well as the traditions of familiarization with regional literature.
The implementation of literary tourism has the following advantages in the framework of educational technology: 1. A deeper understanding of the relevance of the literary places' preservation [1].
2. An empirical assessment of the tourist's awareness of a literary work and the satisfaction degree with the location [2].
3. Rereading literary works based on personal experience, impressions and memories from trips. 4. Popularization of literary works as a source material before the start of tourist activities [3].
Within the framework of literary tourism, the user has the opportunity not only to get acquainted with the places of life and work of prominent literary figures, but also to pay attention to a wide range of related historical and cultural sites, such as museums, monuments, theaters, buildings, gardens and parks associated with the study of literary heritage, thereby immersing students in the socio-cultural environment in which the classics were born. As studies show [4], an increase in tourist attractiveness allows preserving archaic cultural objects due to an increase in demand for them, expands the list of objects by including little-known and previously unknown regional objects, forms a new generation of people interested in preserving the culture of their region; has a significant economic effect, allows marketing strategies [5], affects the minimization of a services package cost.
Most of the requirements currently associated with the development of information technology in tourism are associated with the search for methodological, algorithmic and software for the development of personalized tourist routes [6][7][8], which are diverse and realistic. However, tourists still have difficulties with the formation of a route affecting several destinations when integrating them in one trip, taking into account financial, timely and other constraints [9]. Significant predicaments exist with the calculation of time and distance for individual tourists and travel in groups, when the preferences and goals of group members may differ and conflict [10]. The business tourism is also interested in data analytics to distinguish the types of tourists based on the choice of places of their visit and traffic patterns [11]. In the works [3,5], the problems of planning literary tourism are considered, and it is also noted that keen readers prefer to plan their trips and avoid organized attractions, criticizing the methods of mass literary tourism.
Among the possible solutions to the above problems, mention should be made of the development of intelligent information systems that allow the large flow information analysis [9][10][11][12]. The use of intelligent systems makes it possible to combine user preferences [13] with effective automated methods for solving complex computational problems and provides a better understanding of the tourists' behavior for making better and faster decisions [6].
One of the solutions to the above problems is the development of a problem-oriented automated information system for organizing literary tourism in St. Petersburg, which would allow to quickly compose thematic routes based on data mining, as well as comprehensively study the heritage and biography of literary figures.
Materials and methods
The presented work relies on empirical research methods based on observation and research in the field of the literary educational tourism implementation by means of information technology. Studies [14] present a list of basic informational entities that are associated with literary tourism. For the system being developed, they can be used as entities stored in a database, with the ability to expand and add new categories: 1. Information about the socio-cultural environment with a description of the historical, cultural, every day, educational value.
2. Information about typology according to educational value. 3. Information about the diversity of traditions and customs of the sociocultural environment.
4. Information about the priority of visiting the sight, the possibility of its dynamic nature due to the change in the relevance of its visit by the society.
5. Information about the components of the natural and built environment in terms of their historical and aesthetic appeal.
6. Information about literary works, as well as audio, video information and photographs of sites.
7. Information about guides and maps, distances between points of routes. 8. Expert information from experienced teachers in the field of organizing educational tours regarding the methodological and logistic areas.
Based on the analysis of the subject area and a review of similar solutions, the general functionality of the developed information system was formed: 1. Availability of a unified database containing the necessary information about authors, films, performances, works and locations; 2. Search for objects by author with the ability to filter; 3. Viewing the list of authors with the ability to search; 4. Viewing the biography of the author; 5. Watching movies about the author; 6. Viewing works of authors with allocation of places in St. Petersburg, mentioned in them; 7. Viewing the list of performances based on the works of the author or his biography 8. User registration; 9. User authorization; 10. Ability to comment and evaluate database objects; 11. Ability to search for places by authors and topics; 12. Route development; 13. Route editing; 14. Downloading the route for offline access. Figure 1 shows the approbation results of the developed information system on specific examples of literary tourism:
Results
At the current stage, a desktop application has been developed that allows filling the database with the entire list of information necessary for the implementation of literary tourism, as well as taking into account the needs of users. The developed software was tested in the framework of the educational work organization with students of the Faculty of Secondary Professional Education of ITMO University in the category of educational tourism. Empirical studies have proven the pedagogical effectiveness of the developed information technology for both the student and the teacher.
Students were able to automatically build individual routes; carry out an automatic search for works with navigation by authors of interest, literary directions; achieved greater immersion in the educational material through visualization and direct contact with the objects of literary study. Teachers noted the ability to automatically track both the literary interests of students and logistic preferences (travel options, infrastructure requirements), the most interesting tourist sites for students (museums, parks, monuments, places from the author's biography or heritage, etc.). From the point of view of organizing the didactics of the educational process, the teachers appreciated the opportunity to design tourism activities both in relation to the work program of the discipline, and within the framework of extracurricular reading and educational work; to organize a complex of tourist events for logically related or competing literary movements that need to be studied in combination (for example, symbolism and acmeism, formed as a reaction to the extremes of symbolism; baroque and its antipode classicism). From the point of view of the organizational and communicative component, the teachers noted a significant reduction in time and economic costs for the preparation and organization of tourist events due to the centralized storage of all content in various literary directions and the construction of the shortest routes to reach tourist sites.
Conclusions
Thus, the automation of the routes compilation of literary tourism allows one to significantly save time and makes it a possible means for solving the problem of attracting students of educational organizations to the cultural and recreational facilities of the region, and hence the socio-economic development of the region as a whole. | 2,661.4 | 2020-01-01T00:00:00.000 | [
"Computer Science"
] |
Fast Trajectory Tracking Control Algorithm for Autonomous Vehicles Based on the Alternating Direction Multiplier Method (ADMM) to the Receding Optimization of Model Predictive Control (MPC)
In order to improve the real-time performance of the trajectory tracking of autonomous vehicles, this paper applies the alternating direction multiplier method (ADMM) to the receding optimization of model predictive control (MPC), which improves the computational speed of the algorithm. Based on the vehicle dynamics model, the output equation of the autonomous vehicle trajectory tracking control system is constructed, and the auxiliary variable and the dual variable are introduced. The quadratic programming problem transformed from the MPC and the vehicle dynamics constraints are rewritten into the solution of the ADMM form, and a decreasing penalty factor is used during the solution process. The simulation verification is carried out through the joint simulation platform of Simulink and Carsim. The results show that, compared with the active set method (ASM) and the interior point method (IPM), the algorithm proposed in this paper can not only improve the accuracy of trajectory tracking, but also exhibits good real-time performance in different prediction time domains and control time domains. When the prediction time domain increases, the calculation time shows no significant difference. This verifies the effectiveness of the ADMM in improving the real-time performance of MPC.
Introduction
Autonomous driving is the main direction of development in the future automobile field.Its advantages have been demonstrated in many aspects and attracted wide attention and recognition from society.The main content of autonomous driving technology includes perception, decision-making, planning, control and other aspects [1].Among them, trajectory tracking is the main content of the control level and one of the core technologies of autonomous vehicles, which determines the performance indicators such as the safety and comfort of vehicle operation [2].
Trajectory tracking is a fundamental function of autonomous vehicles that ensures that the vehicle travels along a preset path.Proportional integral derivative (PID) control, preview control theory, sliding mode control, deep learning, model predictive control (MPC), and other methods have provided numerous trajectory tracking strategies for autonomous vehicles [3][4][5][6][7][8].MPC is widely used by researchers to address the trajectory tracking control problems for intelligent vehicles.MPC is characterized by predictive modeling, receding optimization, feedback correction, and the ability to handle constrained optimization problems [9][10][11][12][13].Aiming at the trajectory tracking problem of distributed drive vehicles, based on the MPC and vehicle dynamics model, a method combining the trajectory tracking preview time control and differential torque control based on the reference angle has been proposed, which improved the trajectory tracking performance [14].In order to improve the tracking ability of autonomous vehicles, fuzzy control is used in combination with MPC to improve the weight of the cost function in MPC through the fuzzy adaptive algorithm.This not only improves the accuracy of trajectory tracking, but also improves the steering smoothness of the vehicle [15].Using the Kalman filter algorithm combined with the MPC controller, the online estimation of the vehicle's nonlinear curvature response can reduce the wear and tear of vehicle components while ensuring comfort [16].Feedforward compensation is integrated with MPC.The error from the expected path is calculated using a pure tracking algorithm as the input for feedback tracking control.The control output is then determined through MPC calculations, resulting in improved accuracy and robustness in trajectory tracking [17].The weights of MPC are dynamically adjusted using a PSO-BP neural network, which improves the tracking performance of the autonomous vehicle under different speeds and road curvatures [18].A vehicle model with steering dynamics is proposed.The cascaded MPC structure is used to separate the steering system of the vehicle dynamics model from the trajectory tracking controller.The simulation and test results show better performance than only considering the vehicle dynamics.However, integrating the steering system into the dynamics model can achieve optimal performance, but leads to higher computational requirements [19].The above methods improve the tracking accuracy, but do not consider the complexity of MPC calculation.
MPC has significant advantages in solving trajectory tracking problems with various constraints.Extensive research has also been conducted to study the accuracy, stability, and robustness of vehicle trajectory tracking based on MPC [20,21].However, MPC has a large online computation workload and low real-time performance.In addition, if the controlled model is too complex, it will significantly increase the MPC online iterative calculation, which will hinder the application of MPC in practice.
Therefore, improving the real-time performance of MPC has received increasing attention, and many research results have emerged [22,23].A dynamic optimization toolkit can improve the computational speed and reduce the computation time of MPC [24,25].A genetic algorithm is used to compute the optimal time domain parameters for realtime vehicle speed and road conditions, which improves the real-time performance of the controller [26].But, it uses a relatively simple kinematic model.To improve the realtime performance of vehicle longitudinal speed planning, a nonlinear space-domain MPC (SMPC) is proposed to accelerate the nonlinear SMPC computation by generating thermal initialization and subsequently forming SMPC-RTI.However, it considers the longitudinal motion of the vehicle as well as the energy it saves [27].In [28], the real-time performance and accuracy of vehicle trajectory tracking are improved by reducing the number of control steps or reducing the control frequency domain.Reducing the control frequency domain can meet the real-time requirements but may result in a slightly higher error compared to reducing the number of control steps.Using the linear parameter time-varying MPC and designing a linear quadratic regulator can reduce the computational cost of the dynamic control of the vehicle [29].Alternating the direction multiplier method (ADMM) is favored by many researchers because of its good scalability.It has been successfully applied to machine learning, distributed computing, and other fields [30][31][32].ADMM can divide the optimization variables into two parts and solve them in a separate framework, which can reduce the computational burden caused by the large scale of the system.In [33], a class of ADMM with nonlinear equality constraints is empirically studied and its convergence is analyzed.To enhance the real-time performance of the iterative linear quadratic regulator (iLQR), an optimization is conducted using the ADMM.When employing a logarithmic barrier function, this algorithm can circumvent the feasibility requirements of the initial trajectory during the first iteration, thereby expediting the optimization process [34].It shows good real-time performance in different driving scenarios.
In summary, many methods have been proposed to improve the real-time performance of MPC, and some research results have been obtained.However, there are still some issues regarding improving the real-time performance of MPC for autonomous vehicle trajectory tracking.The combination of optimization algorithm and MPC can improve the real-time performance of MPC to a certain extent, but it will take up additional computational resources when solving the optimal parameters of MPC [34].Simplifying the vehicle model can also improve the real-time performance of the controller, but it may not achieve ideal control effects in complex road conditions [28].When using optimization algorithms or simplified models in conjunction with MPC controllers, the MPC problem is usually transformed into a quadratic programming (QP) problem [35].The interior point method (IPM) or the active set method (ASM) are commonly used to solve the QP problem, but the ASM cannot utilize the sparsity of the MPC problem, and has to perform the active set operation at each iteration, which is suitable for the case of a small number of controls and constraints [36].The IPM can utilize the sparsity of the MPC problem, but each iteration needs to solve the Karush-Kuhn-Tucker (KKT) system, and when the KKT system changes during the iteration, the solution process needs to decompose or inverse the matrix.When the number of constraints is large or the prediction time domain increases, the matrix dimension becomes larger and the computation time becomes longer [37].This paper focuses on the problem of a long MPC solution time.Starting from the MPC problemsolving process, based on the framework of trajectory tracking control for autonomous vehicles, an auxiliary variable method is introduced to transform the quadratic problems in MPC into a separable structured optimization problem.The ADMM is used to solve it, and a decreasing penalty factor is used to ensure the convergence of the ADMM algorithm during the solution.
The organization of this article is as follows.The vehicle dynamic model and the MPC controller are presented in Section 2. The combination of the ADMM algorithm and MPC is introduced in Section 3. The simulation results of the controller are shown in Section 4. Finally, a brief conclusion is given in Section 5.
Vehicle Dynamics Model
To ensure that the vehicle follows its desired trajectory, the bicycle model in [38] is adopted to represent the vehicle dynamics.In this model, we do not consider the effects of suspension and aerodynamics, the state variables are defined as ξ(t) = [ ẋ, ẏ, ψ, ψ, Y, X] T , and the control input vector is the front wheel angle, defined as u(t) = δ f .The threedegrees-of-freedom dynamics model of the vehicle is shown in Figure 1.The x axis represents the longitudinal axis in the vehicle coordinate system, the y axis represents the lateral axis in the vehicle coordinate system, the xoy is the body coordinate system, and XOY is the ground coordinate system.
where m is the mass of the vehicle, I z is the moment of inertia about the z axis, a and b are the distances between the center of mass and front/rear axles of the vehicle, ψ is the yaw angle of the vehicle, F x f and F c f represent the tire force on the longitudinal axis of the vehicle in the body coordinate system, and F d f and F lr represent the tire force on the transverse axis of the vehicle in the body coordinate system, respectively.
When the lateral acceleration is not greater than 0.4 g, the longitudinal force and lateral force exerted on the tire can approximate a linear relationship as given below: In Equation ( 2), C l f and C lr represent the longitudinal stiffness of the front and rear tires of the vehicle, C c f and C cr represent the cornering stiffness of the front and rear tires of the vehicle, s f and s r represent the slip ratio of the front and rear tires of the vehicle, α f and α r represent the cornering angle of the front and rear tires of the vehicle, respectively.Assuming that both front wheel angles of the vehicle are equally small, the lateral acceleration satisfies the small angle assumption, and in this case, the following approximate relationship can be used: Finally, consider the transformed relationship between the vehicle body coordinate system and the earth inertial coordinate system as follows: From Equations ( 1)-( 4), we can obtain the following nonlinear dynamic model of intelligent vehicles: It can be represented by the following differential equation: The state variables of the system are ξ(t) = [ ẋ, ẏ, ψ, ψ, Y, X] T , and the control variable is selected as u(t) = δ f .
Trajectory Tracking Based on Model Predictive Control
The nonlinear vehicle dynamics model developed in Equation ( 6) is complex and takes a long computational time to solve, so it is linearized using Taylor's formula, omitting the higher-order terms except the first order.Assuming the Taylor expansion of Equation ( 6) at [ξ 0 , u 0 ], the resulting linear time-varying equation is given as follows: where are the Jacobi matrices.
Discretizing the above state space equations using the forward Euler's method, the following equation can be obtained: From the above equations, the discrete linearized system can be shown: where state error of the system at k-time, and T is the sampling time.
Taking the increment of the control variable, i.e., ∆u(k), as the input to the system reduces the effect of the sudden change in the control variable on the system, and therefore the original state vector needs to be augmented.Let the new state vector be ξ(k) = [ξ(k), u(k − 1)] T , we can obtain the new state space form as follows: where The predicted output expression can be obtained as follows: where . . .
The primary objective of the MPC controller is to minimize the disparity between the output and the reference value.This ensures precise vehicle tracking along the desired path while maintaining lateral stability.However, due to the intricacies of the vehicle dynamics model and associated constraints, situations may arise where a numerical solution cannot be obtained within a single control cycle.To address this, a relaxation factor, denoted by , is introduced.This factor guarantees that a viable solution can be obtained in each control cycle.The objective function is expressed as follows: where N p is the prediction time domain, N c is the control time domain, is the relaxation factor, ρ is the weight of the relaxation factor, η(t + i) represents the actual output of the system, η re f (t + i) represents the reference output, and ∆U(t + i) represents the increment of the forward turn angle.
According to Equation (12), the total cost of the objective function consists of three main parts.The first term relates to the cost associated with the deviation between the output trajectory and the reference trajectory.A greater deviation leads to a higher cost.The second term considers the cost associated with the forward corner increment.A larger forward corner increment during the tracking process results in a higher cost.Achieving a smoother control process is of the utmost importance when minimizing the control amplitude.The third term encompasses the relaxation factor and weight, ensuring the attainability of an executable solution for the objective function.When the vehicle is performing trajectory tracking, it is necessary to consider both the dynamic constraints of the vehicle itself and the limitations of the actuating mechanism.The constraints of vehicle dynamics are as follows: where y hc is the hard constraint output and y sc is the soft constraint output.
Quadratic programming can solve the optimization problem of the MPC objective function; therefore, Equation ( 12) can be rewritten to the standard quadratic programming form, and the state variables and control variables constraint are introduced: Solving Equation ( 14) in each control cycle yields the optimal sequence of control increments: If the first item in the control sequence is added to the system, the current control variable is given as follows: After the whole system enters the next cycle, the system repeats the above process and updates the control sequence to complete the trajectory tracking of the intelligent vehicle.MPC obtains the optimal control sequence through iterative optimization search, and the actual application of the algorithm is limited when the iteration process is long.The ADMM algorithm has the factorization of dual ascent and the global convergence of a multiplier method, which has received widespread attention in recent years due to its low computational complexity and simple algorithm structure.In this paper, it is used to solve the optimization problem in the trajectory tracking MPC of autonomous vehicles.
Alternating Direction Method of Multipliers
The alternating direction multiplier method is generally used to solve constraint programming problems with the following equation: where f and g are convex functions, z and v are the variables to be optimized, and are the linear equality constraints that the problem needs to satisfy.By introducing the dual variable ω, the augmented Lagrangian function of the above equation is constructed as follows: where ρ > 0 is the penalty parameter.
The iterative process of the ADMM algorithm includes three parts: iteratively updating the original variable z, iteratively updating the original variable v; and updating the process of the dual variable ω.The update strategy is given as follows [39]: In general, it is convenient to use the scaling dual variable µ = ω ρ to represent the iteration process: One advantage of the ADMM method is that it has only one parameter ρ, and under general conditions, the method can demonstrate convergence to all penalty factors.During the iterative process, the solution for the primal variables z and v are performed alternately, which reduces the scale of the problem and improves computational efficiency.
Model Predictive Controller Based on ADMM Improvement
Based on the previous section, this section provides a detailed explanation of the combination of the ADMM algorithm and trajectory tracking MPC problem.As shown in Figure 2, the ADMM algorithm is applied to the optimization and solving process in MPC.Based on the dynamic model of intelligent vehicles, the future outputs of the system are predicted using the state information of the vehicle and the control input, and then the feedback correction is performed using the actual output of the detection object.Finally, the ADMM algorithm is used to solve the optimization target online, and the current optimal control input is obtained.To apply ADMM to the receding optimization process of MPC, the QP problem described by Equation ( 14) can be abbreviated as follows: where After formulating the quadratic programming in standard form for addressing the trajectory tracking challenge in autonomous vehicles, it is imperative to incorporate auxiliary variables and subsequently reformulate the problem into a structure conducive to the ADMM algorithmic resolution: where According to the multiplier method, the augmented Lagrangian function for the optimization problem can be obtained as follows: By introducing the dual scaling variable µ = y ρ, according to Equation ( 14), the iterative update process of the ADMM algorithm can be obtained as follows: Compute the gradients of the original variable x and the auxiliary variable z, respectively.According to the first-order optimality condition, the iterative process formula is given as follows: To accelerate convergence, a relaxation factor of α ∈ [1, 2] is added, and the above iterative process can be obtained as follows: According to the first-order optimality condition of the QP problem, define the original residuals s prim and dual residuals s dual of ( 21), so we can obtain Equation ( 27): The algorithm convergence criterion is given according to the two residuals: where In general, ε rel = 10 −3 , ε abs can be selected according to the required accuracy.
The update of the ADMM algorithm variables requires the use of the results of the previous moment, and the optimal solution of the previous moment (d − 1) is selected as the initial value of the algorithm at the current moment (d).The algorithm flow of the MPC solution control input ∆u under trajectory tracking can be obtained.
(1) Initialize the MPC parameter to obtain the system status information at the dth moment.
(2) According to the equation of the state variables of the system and the input and output variables, the objective function is converted into a quadratic programming problem in the form of Equation ( 14).(3) Rewrite Equation ( 14) to form as Equation ( 22) conforms to the ADMM solution.(4) The optimal solution obtained at time d − 1 is used as the initial value of the solution to the time problem.(5) The variables are updated according to the iterative process of the ADMM algorithm, as shown in Equation ( 26). ( 6) According to Equations ( 27)- (30), to determine whether the iteration process meets the termination conditions, if it is met, stop the iteration, send the first term in the calculated optimal solution sequence to the control system as input, and enter step (7); if not, continue to iterate until the maximum number of iterations is reached.(7) Go to the next sampling moment d + 1, and repeat step (1).
Simulation
To verify the effectiveness of the proposed algorithm, a joint simulation platform of Simulink and CarSim was built to simulate and validate the designed controller.The ASM and the IPM are commonly used methods for solving traditional MPC problems, and this paper compares and analyzes the proposed algorithm with ASM and IPM.The processor parameters of the laptop used in the simulation are AMD Ryzen 7 5700U with Radeon Graphics 1.80 GHz.The vehicle parameters used in simulation are shown in Table 1.In the ordinary vehicle driving test, the double-shift condition is the test section with high frequency.Many scholars have also used it to test the trajectory tracking capabilities of autonomous vehicles.The desired trajectory is given by: where In this section, the traditional MPC is solved by ASM and IPM, and the simulation results of the proposed controller were analyzed under the same prediction and control time domains.The simulation results are shown in Figures 4-6 From Figure 4, it can be seen that, during trajectory tracking, both the MPC with ADMM algorithm improvement and the traditional MPC controller can achieve good tracking effect.The tracking accuracy of the MPC with the ADMM algorithm improvement is higher during the control time.We can also obtain an improved MPC that shows a better tracking accuracy, the proposed controller has a tracking error of 0.266 m at the maximum lateral displacement, while the other two controllers have a tracking error of 0.443 m.During the whole control time, we used the trajectory data driven by the vehicle and the expected trajectory data to obtain the root mean square error (RMSE) of each controller.The RMSE for improved MPC is 0.193, and it is 0.231 for both other controllers.The RMSE formula is shown below: where N represents the total simulation time divided by the sampling time, Y represents the actual trajectory of the vehicle, and Y re f represents the expected trajectory.Figure 5 reflects the tracking of the three controllers on the yaw angle.It can be seen that all controllers show a good tracking effect on the change of yaw angle.At the longitudinal position of 80 m, the MPC with the ADMM algorithm improvement has some overshoot in tracking the desired yaw angle, and there are some oscillations at the longitudinal position of 90 m.It is worth noting that the driving comfort may be reduced due to the change in the heading angle, but it can quickly track the reference yaw angle afterwards.
Figure 6 reflects the calculation performance of three controllers.The initial iteration of ADMM takes a longer computational time, this is because, in the first iteration, due to the presence of constraints, the computation of the variables needs to solve a large system of linear equations, which will take a long time, and the next step is carried out alternatively to greatly reduce the computational time.The average computation time is shown in Table 2.As shown in Table 2, the average computation time of the MPC improved by the ADMM algorithm is 0.0013 s, the average computation time of the ASM is 0.0020 s, and the average computation time of the IPM is 0.0035 s.Compared with the ASM, the average computation time of the MPC improved by the ADMM algorithm is reduced by 35%.The real-time performance of the MPC improved by the ADMM algorithm is higher than that of the IPM, with an average computation time reduction of 62.8%.
Comparison of Controllers under Different Control Horizons
In this section of the simulation, the control time domain of the MPC based on ADMM improvement is set to 10, and the control time domain of the traditional MPC is 4. The simulation results are shown in Figures 7-9.By choosing a prediction time domain of N p = 14, it can be seen from Figures 7 and 8 that the trajectory tracking accuracy of the improved MPC is significantly better than that of the IPM and ASM.The error of the improved MPC at the maximum lateral displacement is 0.273 m, while that of the IPM and ASM is 0.665 m.The tracking precision achieved by both the IPM and the ASM falls below the desired level, primarily due to the limited control time domain.Notably, IPM performs less effectively in tracking.The RMSE of improved MPC is 0.95, ASM is 0.317, and IPM is 0.319.The IPM causes the vehicle's front wheel angle to vary more in the 0-20 m range, which affects the stability, whereas the other two controllers show a smoother control performance.
Figure 9 reflects the computation time of the controller.It can be seen that the improved MPC algorithm takes a longer computation time in the first iteration, which is because it has a larger control time domain, which means that more dimensions of the control volume need to be solved in the iteration.Depending on the good performance of the ADMM algorithm, the average computation time of the improved MPC is still smaller than IPM and ASM.As can be seen from Table 3, the average computation time (0.0015 s) of the MPC improved by the ADMM algorithm is slightly lower than that of the ASM (0.0018 s).Compared with the IPM (0.0038 s), the average computation time of the MPC based on the ADMM algorithm is significantly reduced.
Comparison of Controller Computation Time as the Prediction Horizon Increases
This section of simulation discusses the advantages and disadvantages of the improved MPC based on ADMM and traditional MPC in real-time performance under different prediction time domains.When the prediction time domains are set to 8, 10, 12, 14, 16, 20 and 22, the average computation time of the three controllers is compared as shown in Table 4, and the average solution time of the control sequence by the model based on ADMM improvement is basically unchanged in different prediction time domains, as shown in Figure 10.From Figure 10, it can be seen that as the prediction horizon increases, the computation time of the MPC based on ADMM improvement is basically stable at 0.0015 s.Compared with the active set method, the maximum average computation time is reduced by 51.6%.Compared with the interior point method, as the prediction horizon increases, the computation time of the interior point method also increases.The maximum average computation time of the MPC based on ADMM improvement is reduced by 64.3%.
The simulation results indicate that the MPC improved by the ADMM algorithm can effectively reduce the online solving time of the traditional MPC.Under the same prediction time domain and control time domain, the average computation time is reduced by 35% compared to the ASM, and reduced by 62.8% compared to the IPM.When the same control time domain is used and the prediction time domain is changed, the average computation time of the MPC based on the improved ADMM method is smaller than that of the traditional MPC.Compared with the ASM, the computation time is reduced by up to 51.6%, and compared with the IPM, the computation time is reduced by up to 64.3%.Additionally, the proposed controller has higher tracking accuracy for double lane changes with some improvement in changing lanes.
Conclusions
This paper aims to address the problem of slow online solving and the low real-time performance of traditional MPC.A combination of ADMM and MPC is used to improve the real-time performance of trajectory tracking for autonomous vehicles.The ADMM algorithm is used to solve the QP problem transformed by MPC, and the ADMM algorithm is incorporated into the receding optimization of MPC.A relaxation factor α is added to accelerate the convergence.
To validate the effectiveness of the proposed method, we established a joint simulation platform within the Carsim-Simulink environment and conducted system simulations and tests in a double-shifted lane scenario.Regarding tracking accuracy, the algorithm demonstrates significant improvements.Numerical results indicate a reduction in the tracking error by 0.217 m at the maximum transverse displacement under identical prediction and control time domains, and by 0.392 m, at the maximum transverse displacement under varying control time domains.In terms of computation time, the proposed algorithm notably enhances the real-time performance of MPC when compared to previous methods.Numerical results reveal an average computation time reduction of 64.3% compared to IPM and 51.6% compared to ASM under different control time domains.Furthermore, the ADMM algorithm exhibits stability in the computation time compared to the IPM and ASM.As the prediction time domain increases, the computation time for MPC improved by the ADMM algorithm remains nearly constant, while the computation time for the ASM fluctuates significantly and the computation time for the IPM increases.
Future research may consider integrating the proposed algorithm with an optimization toolkit to further enhance its real-time performance.Additionally, this paper does not address longitudinal control for intelligent vehicles.Therefore, the proposed algorithm could potentially be applied to the combined horizontal and longitudinal control of autonomous vehicles in the future.
Figure 1 .
Figure 1.Vehicle dynamics model.Differential equations for vehicle motion in lateral, longitudinal, and yawing directions are established based on Newton's second law, as given below:
2, d x1 = 25, d x2 = 21.95,d y1 = 4.05, d y2 = 5.7.The initial parameters of the algorithm are set as the road adhesion coefficient µ = 0.85, v = 20 m/s, relaxation factor α = 1.7, and the penalty factor ρ is a decreasing sequence with respect to time.The calculation time of the three algorithms is compared in the same and different prediction time domain and control time domain.The joint simulation diagram of the MPC trajectory tracking improved by ADMM is shown in Figure 3.
4. 1 .
Comparison of Controllers under the Same Prediction and Control Horizon (N p = 11, N c = 6) .
Figure 6 .
Figure 6.Computation time of different control methods.
Figure 9 .
Figure 9. Computation time of different control methods.
Figure 10 .
Figure 10.Comparison of the computation time of the three controllers in different prediction time domains.
Table 1 .
Vehicle parameters in the simulation.
Table 2 .
Computation time of different control methods.
Table 3 .
Computation time of different control methods.
Table 4 .
Comparison of controller computation time in different prediction time domains. | 6,826.4 | 2023-10-01T00:00:00.000 | [
"Engineering",
"Computer Science"
] |
Origins of a Low-Sulfur Superalloy Al 2 O 3 Scale Adhesion Map
: Low-sulfur single-crystal Ni-base superalloys have demonstrated excellent cyclic oxidation resistance due to improved Al 2 O 3 scale adhesion. This derives from preventing deleterious interfacial sulfur segregation that occurs at common ppm levels of S impurity. Multiple hydrogen-annealing desulfurization treatments were employed to produce a continuum of levels demonstrating this oxidative transition, using 1 h cyclic oxidation at 1100 ◦ C for 500 h to 1000 h. The sulfur content was determined by glow discharge mass spectrometry. The complete gravimetric database of 25 samples is revealed and correlated with sulfur content. Maximum adhesion (i.e., no weight loss) was achieved at ≤ 0.3 ppmw S, significant spallation (20–30 mg/cm 2 ) above 2 ppmw, with transitional behavior between 0.3 and 2 ppmw S. A map suggested that adhesion was enabled when the total sulfur reservoir was less than one S atom per Ni interface atom. Equilibrium models further suggest that segregation may be minimized (~1% at 0.2 ppmw bulk), regardless of section thickness. 1st order adhesion effects have thus been demonstrated for PWA 1480 having no Y, Zr, or Hf reactive element dopants and no possibility of confounding reactive element effects. The results are compared with 2nd generation PWA 1484, Rene’N5, N6, and CMSX-4 ® SLS, all having Hf dopants.
Introduction
Al 2 O 3 -forming coatings and superalloys are widely accepted as among the most oxidation-resistant high-temperature materials used commercially. In addition to low scale growth rates, the materials must also possess good scale adhesion to prevent spallation and wastage caused by thermal cycling. Generally, this was spectacularly accomplished through reactive element dopants, Y, Zr, Hf, etc. Among numerous other theories, it was postulated that these elements served to getter ppm levels of a sulfur impurity, thus preventing deleterious interfacial segregation and bond weakening. Upon cooldown, significant compressive thermal stress produces spallation of the scale, with concurrent loss of alloy Al content. This may proceed initially with some gradual weight loss in repeated thermal cycling exposures. However, more seriously, continued cycling allows a transition to other oxides that grow much faster and eventually produce breakaway oxidative degradation.
In order to support this mechanism, various methods were explored to produce low sulfur materials with improved cyclic oxidation behavior. The significant effect of low sulfur content on the cyclic oxidation and spalling behavior of high-temperature alloys has been well established. Initially, studies were on cyclic polish-purged NiCrAl alloys, where improved adhesion was initiated at low sulfur contents on the order of 1 ppmw. These transitioned to single-crystal alloys and the use of 1 atm hydrogen annealing as a more efficient, controlled desulfurization technique [1,2]. As this process was refined, multiple PWA 1480 (Pratt and Whitney, East Hartford, CT, USA) single-crystal samples were desulfurized by adjusting the thickness of the sample and time/temperature of the anneal. Correlation of the desulfurization conditions with cyclic oxidation [3] provided the source data for this paper. At that time, the only sulfur content available was that predicted by the thin slab diffusion solution. The corresponding measured sulfur contents
Materials and Methods
A polycrystalline ingot of PWA 1480 melt stock was sectioned by electro-discharge machining (EDM) into rectangular samples of various thicknesses. The overall dimensions were~12.7 ×25.4 mm ( 1 2 x 1 ). The machined thicknesses were nominally 0.25, 0.51, 1.27, 2.54, and 5.08 mm (labeled "10, 20, 50, 100, and 200 mils"). They were further hand polished on SiC emery paper down to 600 grit. Hydrogen annealing took place in a horizontal alumina tube under slowly flowing 5% H 2 /Ar, using Zr foil as an oxygen getter. Four temperatures, 1000 • C, 1100 • C, 1200 • C, and 1300 • C, were used. Times were 8, 20, 50, or 100 h. Annealing weight changes were less than 0.03 mg/cm 2 as samples were clean and metallic in appearance (no surface scale). Samples had a 1.5 mm slice taken from the end for sulfur analyses by glow discharge mass spectroscopy (GDMS, Shiva Technologies, Syracuse; now Evans Analytical Group, EAG, Liverpool, NY, USA). Samples were repolished as before, re-measured, and tested by cyclic oxidation in a vertical alumina tube furnace at 1100 • C (1 h heating and 10 min cooling). Weight change was monitored on a Sartorius analytical balance, sensitive to 0.01 mg. Balance drift was periodically checked using standard 1.00000 and 5.00000 g weights. Scales were examined by XRD and SEM with results highlighted previously [1-5].
Cyclic Oxidation vs. Sulfur Database
The cyclic oxidation results were tabulated for the series of sample thicknesses, as delineated by annealing temperature/time. These are presented in Tables 1-4 for nominally "10 mil", "20 mil", "50 mil", "100 mil" and "200 mil" samples (0.15, 0.42, 1.22, 2.53, and 5.02 mm actual thickness). The sulfur content of the starting material was~6.15 ppmw; that of the annealed samples is also included. These complete data sets allow for multiple analysis schemes, as opposed to simply presenting representative plots. It can be seen in Table 1 that significant weight losses occurred after 500 h oxidation for the "10 mil" samples (polished to~0.15 mm) hydrogen-annealed at 1000 • C for 20 h or less. Oxidation resistance improved for 100 h anneals and achieved small gains for samples annealed at 1200 • C, indicating adherent behavior. Table 2 presents results for the "20 mil" samples (polished to~0.42 mm), indicating a substantial −33 mg/cm 2 loss for the control (no anneal). It improved slightly with 1000 • C, and 1100 • C hydrogen anneals. They exhibited small gains for samples annealed at 1200 • C and only a modest loss for the sample annealed at 1300 • C (samples annealed at 1300 • C may exhibit anomalous behavior because they suffered incipient melting during the anneal and may not represent master ingot microstructure). Duplicate samples annealed for 20 h both showed small 0.6 mg/cm 2 gains after 500 h, with one decreasing to near 0 mg/cm 2 after 1000 h. The "50 mil" samples (~1.22 mm), Table 3, showed the highest 500 h weight losses (−66 mg/cm 2 ) for the control sample, i.e., with no hydrogen anneal. Annealing at 1200 • C produced progressively less oxidative weight loss with annealing time, achieving a small weight gain for the 100 h annealed sample. The "100 mil" samples (~2.53 mm), Table 4, exhibited weight loss for samples hydrogenannealed at 1100 • C and 1200 • C, the latter with some improvement for 100 h annealing. The 1300 • C annealing appeared to produce an improved performance here. Finally, the "200 mil" samples (~5.09 mm) exhibited various degrees of non-adherent behavior for both the 1200 • C and 1300 • C-annealed samples. Table 1. Cyclic-oxidation specific weight change data for "10 mil" PWA 1480 samples (1100 • C for 500 h in air). Effect of hydrogen annealing time and temperature. Measured sulfur content (C S ) after hydrogen annealing; thickness (L) measured after polishing and before oxidation testing.
Cyclic Oxidation Plots
The cyclic oxidation weight change vs. time was plotted for many conditions [4]. They show variations of a classic cyclic oxidation form: initial parabolic weight gain, a small maximum less than 1 mg/cm 2 , a gradual decrease to cross zero weight change, and a final, perhaps linear, rate of loss. An example is presented in Figure 1 for a full range of sulfur contents. This family of curves is for samples of various nominal thicknesses ("10 mil", "20 mil", and "50 mil"), annealed at various T, t, plus the un-annealed 6.7 ppmw S control sample baseline. Other figures illustrated the basic trend toward greater adhesion for thinner samples, higher annealing temperatures and longer annealing times [4]. These are keyed to the diffusion factor containing thickness (L) and time (t) as Dt/L 2 and in the temperature control of D s as D s = D 0 exp (−Q/RT). Although difficult to resolve because of overlapping curves, it is noted that early growth rates, before measurable spallation occurred, fall within a small range. The average 5-h data from 23 samples in Tables 1-4 (less two outliers) was 0.272 mg/cm 2 , with a standard deviation of 0.065 (i.e., 24%). These low values suggest no significant change in growth mechanism occurred in the initial Al 2 O 3 films. This is contrasted with~5000% differential in final weight change, e.g., −30 mg/cm 2 vs. +0.6 mg/cm 2 . Table 2. Cyclic-oxidation specific weight change data for "20 mil" PWA 1480 samples (1100 • C for 500 or 1000 h in air). Effect of hydrogen annealing time and temperature. Measured sulfur content (C S ) after hydrogen annealing; thickness (L) measured after polishing and before oxidation testing. Table 3. Cyclic-oxidation specific weight change data for "50 mil" PWA 1480 samples (1100 • C for 500 h in air). Effect of hydrogen annealing time and temperature. Measured sulfur content (C S ) after hydrogen annealing; thickness (L) measured after polishing and before oxidation testing.
Cyclic Oxidation Plots
The cyclic oxidation weight change vs. time was plotted for many conditions [4]. They show variations of a classic cyclic oxidation form: initial parabolic weight gain, a small maximum less than 1 mg/cm 2 , a gradual decrease to cross zero weight change, and a final, perhaps linear, rate of loss. An example is presented in Figure 1 for a full range of sulfur contents. This family of curves is for samples of various nominal thicknesses ("10 mil", "20 mil", and "50 mil"), annealed at various T, t, plus the un-annealed 6.7 ppmw S control sample baseline. Other figures illustrated the basic trend toward greater adhesion for thinner samples, higher annealing temperatures and longer annealing times [4]. These are keyed to the diffusion factor containing thickness (L) and time (t) as Dt/L 2 and in the temperature control of Ds as Ds = D0 exp (−Q/RT). Although difficult to resolve because of overlapping curves, it is noted that early growth rates, before measurable spallation occurred, fall within a small range. The average 5-h data from 23 samples in Tables 1-4 (less two outliers) was 0.272 mg/cm 2 , with a standard deviation of 0.065 (i.e., 24%). These low values suggest no significant change in growth mechanism occurred in the initial Al2O3 films. This is contrasted with ~5000% differential in final weight change, e.g., −30 mg/cm 2 vs +0.6 mg/cm 2 .
Correlation with Sulfur Content
For theoretical and practical purposes, it is instructive to catalog cyclic oxidation behavior as a function of sulfur content. The first attempt at a concise formulation may be to correlate a figure of merit with sulfur content. This had been done by plotting 500 h weight change vs. sulfur content [4]. While illustrating the overall trend, it was seen that plots for various sample thicknesses did not collapse on a single-valued, monotonic, universal curve. However, the data strongly suggested fully adherent behavior for sulfur contents at or below about 0.1 ppmw S for most thicknesses. It also suggested that thicker samples required lower sulfur contents to become adherent, and, conversely, thinner samples maintained adhesion at higher sulfur contents. Thus, a unique value of critical sulfur content could not immediately be claimed for all section thicknesses.
Since breakaway weight change may follow nonuniform, less repeatable behavior, a second attempt was made to correlate performance with sulfur content by using the time to cross zero weight change, i.e., t 0 , as compiled in Table 5 for convenient reference. This construction is shown in Figure 2. Here each symbol shape refers to one section thickness, and fill color refers to one anneal temperature. It is seen that high sulfur contents are inevitably associated with short times to cross zero, t 0 , as low as 20 h (lower right quadrant). At lower sulfur contents, below about 0.3 ppmw, most data points indicate nearly full adhesion (upper left quadrant), some not exhibiting any decrease in weight change over the entire test duration (i.e., t 0 > 500 h or 1000 h) (Red symbols, 1300 • C anneals, may indicate anomalous behavior as mentioned previously). The advantage of this construction is that a sharp edge (≤ 0.3 ppmw S) is illustrated, separating excellent performance from the downward trend toward that exhibited by unannealed samples at 6 ppmw S. While this representation provides a useful graphical tool, a more mechanistic approach might involve the rate of weight loss or a spall constant extracted from cyclic oxidation models, such as COSP [7].
Scale-Adhesion Map
These results were influenced by sample thickness, first because desulfurization was limited for thick samples with a large diffusion distance. However, there is a secondary thickness consideration. It derives from the larger reservoir of residual sulfur contained in thick samples, even at equivalent sulfur content to thin samples. During cycling, this allows for more re-segregation after repeated spallation events that remove sulfur from the damaged interface. To accommodate this factor, an adhesion map was constructed that included sample thickness as a variable affecting the total sulfur content. The other variable was the GDMS measured sulfur content of each individual sample, with various hydrogen annealing pretreatments. The data were color-coded by regions of 500 h cyclic weight change behavior, from very adherent (± 1 mg/cm 2 ) to poorest performance (< −30 mg/cm 2 ) and is shown in Figure 3. Furthermore, approximate positions of 0 and −10 mg/cm 2 were interpolated from data corresponding to each individual set of nominal thickness. This allowed a locus of initially adherent and non-adherent boundaries to be delineated, as indicated by the solid lines. These show a downward trend of lower sulfur content needed for thicker sections to produce adherent behavior. A simple example interpolates a critical sulfur content of ~0.2 ppmw for a 1 mm thick section.
Scale-Adhesion Map
These results were influenced by sample thickness, first because desulfurization was limited for thick samples with a large diffusion distance. However, there is a secondary thickness consideration. It derives from the larger reservoir of residual sulfur contained in thick samples, even at equivalent sulfur content to thin samples. During cycling, this allows for more re-segregation after repeated spallation events that remove sulfur from the damaged interface. To accommodate this factor, an adhesion map was constructed that included sample thickness as a variable affecting the total sulfur content. The other variable was the GDMS measured sulfur content of each individual sample, with various hydrogen annealing pretreatments. The data were color-coded by regions of 500 h cyclic weight change behavior, from very adherent (± 1 mg/cm 2 ) to poorest performance (<−30 mg/cm 2 ) and is shown in Figure 3. Furthermore, approximate positions of 0 and −10 mg/cm 2 were interpolated from data corresponding to each individual set of nominal thickness. This allowed a locus of initially adherent and non-adherent boundaries to be delineated, as indicated by the solid lines. These show a downward trend of lower sulfur content needed for thicker sections to produce adherent behavior. A simple example interpolates a critical sulfur content of~0.2 ppmw for a 1 mm thick section. Lastly, a mass balance approach was applied to the total amount of sulfur contained in the sample as compared to the amount required to provide one full monolayer coverage due to interface segregation [4,6]: where CS = bulk sulfur content in weight fraction; Nm = number of S atoms per metal surface atom; A = surface area of sample (cm 2 ); and W = sample weight (gm). These are shown as dashed lines for 1 full monolayer (blue) and 10 monolayers (orange). It is interesting to see that the slopes correspond roughly to the interpolated experimental adhesion boundaries (solid). Furthermore, a segregation level accruing to about 1 monolayer can be taken as a rough criterion to define adherent behavior. This is not to say that actual interfacial segregation must reach a full monolayer at one time, but just that there is a sufficient reservoir to replenish the interface only up to one monolayer total (It is generally believed and shown that surface and interfacial segregation saturates at low monolayer levels at any one time [8,9]. The above discussion thus provides a perspective or guideline for low sulfur oxidation-resistant single-crystal superalloys. Air-cooled blades in aero-turbines typically have very thin sections, ~1-2 mm. Ground power or industrial turbines can have thicker sections, which may require more stringent sulfur limits. Nevertheless, at very low levels, ≤ 0.2 ppmw, thermodynamic segregation laws may limit segregation, according to the Langmuir-McLean isotherm, even for larger section thickness [8]. This behavior may be seen for the 2.5 and 5 mm ("100 mil", "200 mil") samples in the adhesion map, Figure 3. Lastly, a mass balance approach was applied to the total amount of sulfur contained in the sample as compared to the amount required to provide one full monolayer coverage due to interface segregation [4,6]: where C S = bulk sulfur content in weight fraction; N m = number of S atoms per metal surface atom; A = surface area of sample (cm 2 ); and W = sample weight (gm). These are shown as dashed lines for 1 full monolayer (blue) and 10 monolayers (orange). It is interesting to see that the slopes correspond roughly to the interpolated experimental adhesion boundaries (solid). Furthermore, a segregation level accruing to about 1 monolayer can be taken as a rough criterion to define adherent behavior. This is not to say that actual interfacial segregation must reach a full monolayer at one time, but just that there is a sufficient reservoir to replenish the interface only up to one monolayer total (It is generally believed and shown that surface and interfacial segregation saturates at low monolayer levels at any one time [8,9]. The above discussion thus provides a perspective or guideline for low sulfur oxidationresistant single-crystal superalloys. Air-cooled blades in aero-turbines typically have very thin sections,~1-2 mm. Ground power or industrial turbines can have thicker sections, which may require more stringent sulfur limits. Nevertheless, at very low levels, ≤ 0.2 ppmw, thermodynamic segregation laws may limit segregation, according to the Langmuir-McLean isotherm, even for larger section thickness [8]. This behavior may be seen for the 2.5 and 5 mm ("100 mil", "200 mil") samples in the adhesion map, Figure 3.
Segregation Projections
The relation for S surface segregation in 200 ppm S Ni was obtained by Miyahara et al. [10] and generalized to all S levels as Equation (2) [8]: where θ is the surface segregation level (relative to a maximum saturation value assumed to be 0.5), and C S is the bulk sulfur content (atomic fractions). It was shown that between 0.1 and 10 ppma (0.06 and 6 ppmw) S, large changes in equilibrium segregation are predicted for the 900-1200 • C temperature range, covering most operational turbine airfoils. This dependency is shown in Figure 4 for 1100 • C. One can estimate the bulk sulfur content that yields a low θ = 1% surface saturation (relative to 0.5 monolayers). At 1100 • C, for example, Equation (2) yields the low bulk sulfur level of 0.33 ppma (~0.20 ppmw) S to predict low 1% segregation. Sulfur levels in this lower range should, therefore, result in minimal surface segregation, regardless of sample thickness, as indicated by the dashed horizontal line in the adhesion map of Figure 3. That is, adhesion will not be as adversely affected by larger reservoirs for thick sections at these low bulk levels. At higher amounts,~5 ppmw, it is seen that 20% segregation is projected. Furthermore, higher temperatures retain more sulfur in solution and segregate less at the same bulk level (low temperatures project as 0.5 monolayers, i.e., full saturation, but can only occur if diffusion is sufficient). Experimentally, low surface segregation had been demonstrated below 900 • C for standard PWA 1480 at 7 ppmw S). Moreover, lower segregation had been demonstrated for PWA 1480 hydrogen desulfurized to 0.1 ppmw [8]. Interface segregation is more appropriate than surface segregation for scale adhesion issues, as demonstrated experimentally [9]. Again, lower bulk S levels were shown to produce lower S interfacial levels.
Segregation Projections
The relation for S surface segregation in 200 ppm S Ni was obtained by Miyahara et al. [10] and generalized to all S levels as Equation (2) [8]: where Ɵ is the surface segregation level (relative to a maximum saturation value assumed to be 0.5), and CS is the bulk sulfur content (atomic fractions). It was shown that between 0.1 and 10 ppma (0.06 and 6 ppmw) S, large changes in equilibrium segregation are predicted for the 900-1200 °C temperature range, covering most operational turbine airfoils. This dependency is shown in Figure 4 for 1100 °C. One can estimate the bulk sulfur content that yields a low Ɵ = 1% surface saturation (relative to 0.5 monolayers). At 1100 °C, for example, Equation (2) yields the low bulk sulfur level of 0.33 ppma (~0.20 ppmw) S to predict low 1% segregation. Sulfur levels in this lower range should, therefore, result in minimal surface segregation, regardless of sample thickness, as indicated by the dashed horizontal line in the adhesion map of Figure 3. That is, adhesion will not be as adversely affected by larger reservoirs for thick sections at these low bulk levels. At higher amounts, ~5 ppmw, it is seen that 20% segregation is projected. Furthermore, higher temperatures retain more sulfur in solution and segregate less at the same bulk level (low temperatures project as 0.5 monolayers, i.e., full saturation, but can only occur if diffusion is sufficient). Experimentally, low surface segregation had been demonstrated below 900 °C for standard PWA 1480 at 7 ppmw S). Moreover, lower segregation had been demonstrated for PWA 1480 hydrogen desulfurized to 0.1 ppmw [8]. Interface segregation is more appropriate than surface segregation for scale adhesion issues, as demonstrated experimentally [9]. Again, lower bulk S levels were shown to produce lower S interfacial levels.
High Performing 2nd Generation Single Crystals
It is noted that some samples exhibited very adherent behavior even up to 1000 1 h cycles at 1100 °C (Table 3). While it is important here to display and map the full range of behavior over a significant range of sulfur contents, it is recognized that improved performance is observed for low sulfur Gen 2, and beyond, single-crystal superalloys. One primary reason may be the widespread use of 0.1-0.2 wt% Hf in most formulations, as well
High Performing 2nd Generation Single Crystals
It is noted that some samples exhibited very adherent behavior even up to 1000 1 h cycles at 1100 • C (Table 3). While it is important here to display and map the full range of behavior over a significant range of sulfur contents, it is recognized that improved performance is observed for low sulfur Gen 2, and beyond, single-crystal superalloys. One primary reason may be the widespread use of 0.1-0.2 wt% Hf in most formulations, as well as the absence of Ti in Rene'N5 or PWA 1484. Here, the gettering potential of Hf for S and C is expected to play an additional positive role compared to PWA 1480 having no Hf.
In that regard, we have shown very adherent behavior (+0.80 mg/cm 2 ) for Rene'N5 hydrogen-annealed to 0.01 ppmw S after 1000 1 h cycles at 1150 • C [11], Table 6 and Figure 5. Furthermore, PWA 1484, hydrogen-annealed to 0.01 ppmw S or melt desulfurized to 0.25 ppmw S, exhibited adherent behavior after 2000 1 h cycles at 1100 • C (+0.93 mg/cm 2 or +0.77 mg/cm 2 ), and after 1000 1 h cycles at 1150 • C (0.52 mg/cm 2 ) [12,13]. In addition, Rene'N6, hydrogen-annealed to 0.08 ppmw S, lost 0.8 mg/cm 2 after 200 h at 1200 • C [14]. These small weight changes (< 1 mg/cm 2 ) are among the lowest produced and recorded for these oxidation times and temperatures for single-crystal superalloys without Y-doping (They may appear exaggerated here on the highly expanded scale of Figure 5). Here sporadic maxima and abrupt drops due to moisture or water immersion (delayed spallation) are demonstrated. The best 1000 h results from PWA 1480 in this study are shown as an average of three samples, with 0.05, 0.08, and 0.28 ppmw S, samples 20-3,6,9, Table 1. The hydrogen-annealed samples are compared and similar to, for example, melt desulfurized 0.2 ppmw S PWA 1484 after 500 h oxidation at 1177 • C (−0.56 mg/cm 2 ) [15]. | 5,698 | 2021-01-13T00:00:00.000 | [
"Materials Science"
] |
Human Uterine Decidual NK Cells in Women with a History of Early Pregnancy Enhance Angiogenesis and Trophoblast Invasion
Objective The present study aimed to identify changes in decidual natural killer (dNK) cells and related cytokines in women who have undergone induced abortions (IAs). The effects of dNK cells on subsequent pregnancies remain unknown. Accordingly, we sought to investigate whether a history of early pregnancy can change dNK cells and facilitate their role in the regulation of angiogenesis and trophoblast invasion. Materials and Methods. dNK cells were obtained from primiparous women who had undergone IA(s) prior to this study and primiparous women who had never been pregnant before this IA (control). Real-time polymerase chain reaction (PCR) was used to measure the mRNA levels of IFN-γ, IP-10, VEGF, and PLGF in dNK cells. The levels of these cytokines were quantified using the enzyme-linked immunosorbent assay. HUVEC and HTR-8/SVneo cells were used to evaluate the angiogenesis, migration, and invasion activities influenced by dNK cells. Results In dNK cells, the mRNA level of IFN-γ, IP-10, VEGF, and PLGF in dNK cells. The levels of these cytokines were quantified using the enzyme-linked immunosorbent assay. HUVEC and HTR-8/SVneo cells were used to evaluate the angiogenesis, migration, and invasion activities influenced by dNK cells. Conclusion The findings of this study suggest that a history of early pregnancy has an impact on dNK cells. These trained dNK cells can regulate angiogenesis and trophoblast invasion and migration by promoting the production of certain cytokines.
Introduction
Placentation in the first trimester substantially affects reproductive success [1]. Research has shown that normal pregnancy and delivery will protect the subsequent pregnancies. First pregnancies are linked to lower birth weights and increased risk of pregnancy disorders such as preeclampsia (PE) [2]. It has been reported that repeated pregnancies can train the memory of decidual natural killer (dNK) cells leading to their enhanced function in subsequent pregnancies [3]. In a previous study, single-cell reconstruction of the early maternal-fetal interface was performed and it was verified that the initiation of dNK1 cells during the first pregnancy responds more effectively in subsequent pregnancies [4]. e dNK cells account for 70% of the decidual immune cells with the capacity of producing cytokines, but limited cytotoxicity. Previous reports suggested that dNK cells might be involved in decidualization, angiogenesis [5], regulation of trophoblast invasion [6], and spiral artery remodeling [7].
IA is a remedial method to deal with contraceptive failure. Although IAs are generally safe and effective, the side effects and influence on subsequent pregnancies should not be ignored. According to previous studies, repeated IAs may cause uterine adhesion, placenta previa, and pelvic inflammation, which impact subsequent pregnancies [8]. However, studies have shown a lower risk of PE associated with IAs in primiparous women [9][10][11][12][13]. A history of light endometrium injuries, such as biopsy and curettage may increase the success rate of implantation in assisted reproductive techniques [14]. e underlying mechanism of these findings has not been elucidated so far. However, it could possibly be attributed to the inflammatory response that contributes to decidualization and implantation. Furthermore, surgical abortions may be associated with endometrium injuries and subsequent spiral artery remodeling, which may be associated with the pathogenic pathway of PE. It was observed that the invasiveness of a trophoblastic cell can be increased by decidual injury [15]. Additionally, longer interpregnancy intervals can lead to a higher risk of PE [11,16]. It seems that among women with a prior IA or shorter interpregnancy interval, maternal immunological recognition of trophoblast may be improved, which may contribute toward placentation and uteroplacental perfusion. Parker et al. [12] speculated that the inflammatory response to IA improves placentation and reduces the risk of PE. However, the pregnancy itself may change the uterine immunological environment, which may contribute to the maternal-fetal crosstalk in the subsequent pregnancy.
To date, there is no evidence showing a relationship between IA or early pregnancy and changes of maternal-fetal interface. Gamliel et al. [3] found that repeated pregnancies were associated with improved placentation and pregnancytrained dNK cells might contribute to improved placentation. erefore, this study aimed to identify changes in the function of dNK cells and the related cytokines in women who have undergone IA. We also sought to investigate whether the dNK cells are trained during the first trimester of gestation. Additionally, we intended to elucidate the role of dNK cells in the regulation of trophoblast invasion and angiogenesis.
Study Population.
All tissue samples were collected with informed consent according to the requirements of the Research Ethics Committee of Beijing Obstetrics and Gynecology Hospital, Capital Medical University (Beijing, China). e inclusion criteria were as follows: (1) pregnancy with a gestational age of 6-9 weeks; (2) no prior births; (3) fetal heartbeat detectable through ultrasound. e exclusion criteria were as follows: (1) women with a history of missed abortion, medical abortion, or delivery; (2) severe inflammation (pelvic inflammatory disease or endometriosis); (3) severe liver or kidney disease; (4) immunological disease; (5) ultrasound confirmed that there was no fetal heartbeat. Decidual samples (n � 33) were obtained from healthy women whose pregnancies were terminated for nonmedical reasons at 6-9 weeks of gestational age. Seventeen of these individuals had never been pregnant before this IA (age, 25 ± 2.57 years; gestational age, 7.41 ± 0.78 weeks). e other 16 participants had no history of delivery but had undergone IA(s) prior to the study (age, 27.44 ± 5.45 years; gestational age, 7.44 ± 0.80 weeks; number of IAs, 1.25 ± 0.58). e tissue fragments derived from the decidua after surgically induced abortions were placed into a container with 25 mL of sterile Roswell Park Memorial Institute (RPMI, Gibco, Grand Island, NY, USA) 1640 medium.
Uterine Flush Sample Collection.
After sterilization, the cervix was cleansed with 0.9% bacteriostatic sodium chloride to remove mucus from the cervical canal. e uterus was then flushed with 5 mL of sterile saline solution using an artificial insemination catheter. Ultrasound was used to supervise the whole process. Uterine flush samples (n � 10) were collected and then centrifuged to confirm cytokine expression in vivo.
Isolation of Decidual Cells.
Decidual tissues derived from the first trimester surgical abortions were washed in sterile phosphate-buffered saline (PBS, Gibco, Grand Island, NY, USA). After the clearance of blood, the tissue was minced and digested with 0.1% collagenase and 100 IU/ml DNase I by shaking in a 37°C water bath for 30 min. Fetal bovine serum (FBS, Gibco, Grand Island, NY, USA) was added to stop the digestion followed by centrifugation. e resulting pellet was resuspended in complete RPMI containing 20 μg/ ml penicillin/streptomycin and then seeded onto Petri dishes followed by overnight incubation at 37°C in a humidified 95%O 2 : 5% CO 2 incubator to enrich decidual cells.
Isolation and Culture of dNK Cells.
e decidual NK cells were filtered and resuspended in RMPI 1640. e leukocyteenriched supernatants were centrifuged and cell pellets resuspended in complete RPMI, layered over Ficoll-Hypaque and centrifuged. Decidual cells were centrifuged (1800 rpm, room temperature) for 30 min in Ficoll gradients (MP Biomedical, Santa Ana, CA, USA) to collect the mononuclear cells. Ficoll layers were collected and resuspended in RPMI 1640. e supernatant was then filtered. Aliquots of 10 7 cells were resuspended in 40 μl MACS buffer followed by the incubation of cells with a CD56+ microbead cocktail at 4°C for 15 min. Microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) were used for the selection of dNK cells. e cells were centrifuged and resuspended in MACS buffer solution, and then loaded onto a column positioned in a magnetic separator (MACs; Miltenyi Biotec, Bergisch Gladbach, Germany). Cells were separated with MACS cell separation columns. Flow cytometry was used to examine the purity of dNK cells.
Real-Time Polymerase Chain Reaction (PCR).
Real-time PCR was used to measure the mRNA levels of IP-10, VEGF, PLGF, and IFN-c in dNK cells. TRIzol reagent (Tiangen Biotech, Beijing, China) was used to extract the total RNA from 1 × 10 6 dNK cells. e RNA samples were assessed for their purity and concentration by spectrophotometry. e cDNA was synthesized from 500 ng of total RNA using PrimeScript RT Master Mix (TaKaRa, Japan) in a total volume of 10 μl at 37°C for 15 min, 85°C for 5 s, and stored at − 20°C before use. Real-time PCR was performed based on the detected fluorescence with SYBR Premix Ex Taq (TaKaRa, Japan). e RNA levels were used in the mRNA levels of the standard. PCR was performed using specific primers synthesized by Invitrogen (Table 1). Relative mRNA expression levels were calculated using REL � 2 − ΔΔCt , where ΔCt � Ct target gene-Ct 18SrRNA. All experiments were performed in triplicate.
Enzyme-Linked Immunosorbent Assay (ELISA).
e culture medium from the dNK cells was collected and stored at − 80°C for cytokine measurements. e levels of cytokines (IFN-c, IP-10, VEGF, and PLGF) in the medium were quantified by ELISA based on kit instructions (R&D Systems, Minneapolis, MN, USA). All measurements were performed in triplicates to minimize the influence of technical error and intra-assay variation.
Transwell Invasion and Migration Assays.
After coculturing with dNK cells, HTR-8/SVneo and HUVEC cell lines were used to evaluate invasion and migration. Transwell assays were performed to assess the influence of dNK cells from each group on the invasion and migration potential of HTR-8/SVneo cells. Approximately 200 μl of HTR-8/SVneo cells (at a concentration of 5 × 10 5 cells/ml) suspension was plated in invasion chambers, which were immersed in 24well cell culture plates containing the supernatant of dNK cells from each group. e plates were cultured at 37°C with 5% CO 2 . Noninvading cells were removed from the top of the Matrigel (BD Biosciences, San Jose, CA, USA) by cotton swabs after 24 h. e cells that had invaded the Matrigel were fixed using 4% paraformaldehyde and dyed with 0.1% crystal violet. e average number of invading cells in five random fields at 100x magnification was calculated for each insert.
We assessed the migratory behavior of HTR-8/SVneo cells and HUVECs incubated with dNK using Transwell assays. Cells were plated on the well (5 × 10 5 cells/ml) and incubated with the supernatant of dNK cells from each group. e migrated cells were then counted using a light microscope (Olympus, IX51, Japan).
Tube Formation.
After coculturing with dNK cells, HUVEC suspension was added to each well of a 96-well BD BioCoat angiogenesis plate. Approximately 20 μl of HUVECs (5 × 10 5 cells/ml) suspension was placed in 96-well plate with 20 μl of supernatant from dNK cells and incubated at 37°C with 5% CO 2 . We evaluated the extent of tube formation microscopically after 6 h.
Statistical Analysis.
e means and standard deviations or medians and interquartile ranges for variables were calculated. Since the data did not meet the assumptions for normality, nonparametric tests (Mann-Whitney U test) were used for comparisons. p value less than 0.05 was considered to be significant. SPSS software (SPSS for windows version 23.0; SPSS Inc., Chicago, IL,USA) and GraphPad Prism 8.0 (GraphPad Software, La Jolla, CA, USA) were used to perform statistical analyses.
Differences in the Secretion Function and Level of mRNA in dNK Cells of the Control and IA Groups.
e dNK cells can secrete cytokines to affect pregnancy. rough ELISA and Real-time PCR, we detected the secretion of four cytokines in the supernatant and their mRNA levels in dNK cells. e results demonstrated that more IP-10 and VEGF were secreted in the supernatants of dNK cells isolated from women with a history of IAs compared to those in the control group (p < 0.01).
e mRNA level of IFN-c in dNK cells was higher in the control group than that in the IA group (p < 0.05). However, the relative levels of PlGF, IP-10, and VEGF mRNA in dNK cells of both groups were not significantly different between the two groups ( Figure 1).
Differences in Cytokine Expression Levels between Uterine
Flush Samples from Control and IA Groups. Because cytokine expression differs between in vivo and in vitro environments, we used an ELISA to detect the expression of the four cytokines in the uterine flush samples. Results showed higher IP-10 and VEGF levels in uterine flush samples from women with a history of IA(s) than in samples from the control group (p � 0.037 and p � 0.032, respectively; Figure 1).
Differences in the Regulation of HUVEC Migration and Angiogenesis by dNK Cells from Different Groups.
Tube formation and Transwell migration assays were performed to evaluate the influence of dNK cells on angiogenesis. e dNK cells isolated from IA and control groups were cocultured with HTR-8/SVneo to imitate the uterine environment. Next, the supernatants were cocultured with HUVEC to facilitate tube formation. e dNK cells from women with a history of early pregnancy were observed to enhance the angiogenesis of HUVECs. Tube formation was significantly enhanced in the IA group compared to the control (p < 0.01; Figures 2(a) and 2(b)). e migration activity of HUVECs from the IA group was significantly enhanced compared to that in the control group (p < 0.01; Figures 2(c) and 2(d)).
Difference in the Regulation of Invasion and Migration of HTR-8/SVneo Cells by dNK Cells from Different Groups.
Transwell migration and invasion assays were performed to evaluate the influence of dNK cells on the invasion and Figures 3(a) and 3(b)). Transwell invasion assays showed a significant increase in
Discussion
e maternal-fetal interface consists of decidual immune cells, decidual stromal cells, and trophoblast cells. It has been reported that the crosstalk among these cells may contribute to the success of pregnancy [17]. Dysregulation of maternalfetal interface can lead to many pregnancy disorders. Moreover, several studies have demonstrated the role of cytokines in the maintenance of pregnancy. dNK cells participate in spiral artery remodeling with trophoblasts [18]. It has been demonstrated that dNK cells regulate trophoblast invasion by producing IP-10. In addition, dNK cells secrete PLGF and VEGF to induce vascular growth [19,20]. IFN-c may have a positive effect on endothelial integrity. It was confirmed that in early pregnancy, dNK cells are in close contact with the trophoblast and promote invasion and angiogenesis. Many cytokines are secreted due to this cellular interaction [21]. e coculture of dNK and trophoblast cells leads to the production of cytokines such as VEGF and IP-10 [22]. IP-10 has a positive impact on trophoblast invasion, but the mechanism underlying this phenomenon has not yet been elucidated. VEGF is an angiogenesis factor, which is secreted by both decidual and dNK cells [19,23].
e primary finding of present study is that dNK cells may secrete cytokines (IP-10, VEGF) related to angiogenesis and trophoblast invasion, which contribute to successful pregnancy. e experiments on cell morphology also verified the promotive effect of dNK cells on pregnancy. Besides, we would like to explore as to when the training of dNK cells begins. Our findings provide a further understanding of the role of dNK cells in pregnancy. e history of an early pregnancy might train uterine NK cells to exert a protective effect in a subsequent pregnancy. However, due to the limited sample size, further studies are needed to support these conclusions.
Complete gestation and delivery processes were reported to have a positive influence on the success of subsequent pregnancy. erefore, in order to eliminate the bias from the history of gestation, women with a history of delivery were excluded.
is criterion allowed us to better evaluate the influence of early pregnancy by analyzing decidual tissues that were obtained from IA in first trimester participants. Besides, considering the potential side effects of IA, in the observation group, we only enrolled women with a history of one or two IAs.
One may question that in vitro culture can possibly alter cytokine expression of dNK cells. Hence we collected and analyzed uterine flushing samples by ELISA.
e results showed that the four target cytokines (IFN-c, IP-10, VEGF, and PLGF) exhibited the same expression trends between flush samples and the supernatants of dNK cells. Many other types of cells exist at the maternal-fetal interface, the influence of which should not be ignored. However, all things considered, results from the ELISA of supernatants and flush samples confirmed that the in vivo and in vitro environments are comparable. e present study observed that the level of mRNA and the protein level of some cytokines were not correlated. We hypothesized that there might be some noncoding-RNAs that regulate the expression of these cytokines. Further studies need to be conducted to explain this phenomenon. BioMed Research International e limitation of our study is that we only performed in vitro analysis. We used the HTR-8/SVneo cell line to verify the contribution of dNK cells to placental progress. e HTR-8/SVneo cell line used in our research was derived from the first trimester of gestation, which is somewhat different from actual in the human maternal-fetal interface.
Considering the side effects of IA, we chose the decidual tissue from patients with a history of only one or two IAs to avoid inflammation or endometrial damage. Moreover, we only explored the function of dNK cells. e potential role of decidual stromal cells and the possible interaction between cells in the maternal-fetal interface were not explored in this study. Our results demonstrated that the dNK cell may affect the HUVEC angiogenesis. However, the mechanism remains unknown. Further investigations should be conducted to validate the relationship between the number of early pregnancies and the effect on subsequent pregnancies.
Conclusion
is study suggests that a history of early pregnancy has a training effect on dNK cells. ese trained dNK cells can promote angiogenesis and trophoblast invasion and migration by promoting the production of certain cytokines.
Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
e authors declare that they have no conflicts of interest. | 4,148.2 | 2020-02-18T00:00:00.000 | [
"Biology",
"Medicine"
] |
A reproducing kernel approach to Lebesgue decomposition
We show that properties of pairs of finite, positive and regular Borel measures on the complex unit circle such as domination, absolute continuity and singularity can be completely described in terms of containment and intersection of their reproducing kernel Hilbert spaces of `Cauchy transforms' in the complex unit disk. This leads to a new construction of the classical Lebesgue decomposition and proof of the Radon--Nikodym theorem using reproducing kernel theory and functional analysis.
is then called the reproducing kernel of H.In this paper, all inner products and sesquilinear forms are conjugate linear in their first argument and linear in their second argument.Any reproducing kernel function is a positive kernel function on X × X, i.e. for any finite set {x 1 , is positive semi-definite.Conversely, by a theorem of Aronszajn and Moore, given any positive kernel function, k, on X × X, one can construct a RKHS of functions on X with reproducing kernel k [3], see Subsection 1.2.Given this bijective correspondence between positive kernel functions on X and RKHS of functions on X, one writes H = H(k) if H is a RKHS with reproducing kernel k.
Equipping the vector space of µ−Cauchy transforms with the H 2 (µ)−inner product, C µ g, C µ h µ := g, h L 2 (µ) ; g, h ∈ H 2 (µ), yields a reproducing kernel Hilbert space (RKHS) of analytic functions in D, H + (µ), with reproducing kernel, Using the above formula (2), it is easy to check that domination of measures implies domination of the reproducing kernels for their spaces of Cauchy transforms, see Theorem 3, where we write k ≤ K for positive kernel functions k, K on X, if K − k is a positive kernel function on X.We will say that λ dominates µ in the reproducing kernel sense (by t 2 > 0) and write µ ≤ RK t 2 λ to denote that k µ ≤ t 2 k λ .By results of Aronszajn, domination of kernels, k ≤ t 2 K, is equivalent to bounded containment of their RKHS, i.e. k ≤ t 2 K if and only if H(k) ⊆ H(K) and the norm of the linear embedding e : H(k) ֒→ H(K) is at most t > 0 [3].See Subsection 1.2 for a review of RKHS theory and these results.In summary, domination of measures implies bounded containment of their spaces of Cauchy transforms: µ ≤ t 2 λ ⇒ H + (µ) ⊆ H + (λ), e µ,λ : H + (µ) ֒→ H + (λ), e µ,λ ≤ t, i.e. µ ≤ t 2 λ ⇒ µ ≤ RK t 2 λ.
Building on this observation, we show that domination and, more generally, absolute continuity, as well as mutual singularity of measures can be completely characterized in terms of their spaces of Cauchy transforms.Moreover, we develop an independent construction of the Lebesgue decomposition and new proof of the Radon-Nikodym theorem using reproducing kernel methods and operator theory.
Outline
The following Background section, Section 1, provides an introduction to (i) the bijective correspondence between positive, finite and regular Borel measures on the circle and contractive analytic functions in the disk, (ii) reproducing kernel theory and (iii) the theory of densely-defined and positive semi-definite quadratic forms in a separable, complex Hilbert space.
Section 2 introduces the reproducing kernel Hilbert spaces, H + (µ), of µ−Cauchy transforms associated to any positive, finite and regular Borel measure, µ, on the complex unit circle.These are Hilbert spaces of holomorphic functions in the complex unit disk.
Our first main results appear in Section 3. Theorem 3 proves that domination of positive measures in the reproducing kernel sense is equivalent to domination in the classical sense: Theorem (Theorem 3).Given positive, finite and regular Borel measures, µ and λ on the unit circle, µ ≤ RK t 2 λ for some t > 0 if and only if µ ≤ t 2 λ.This result is extended to general absolute continuity, written µ ≪ λ, in Theorem 6. Namely, we say that µ is absolutely continuous in the reproducing kernel sense with respect to λ, written µ ≪ RK λ, if the intersection of the space of µ−Cauchy transforms with the space of λ−Cauchy transforms, int(µ, λ), is norm-dense in H + (µ).
These are satisfying results, however, actual construction of the Lebesgue decomposition of µ with respect to λ using reproducing kernel methods is more subtle and bifurcates into the two cases, where: The intersection space, int(µ, λ) = H + (µ) ∩ H + (λ), of the spaces of µ and λ−Cauchy transforms is (i) invariant, or, (ii) not invariant, for the image, V µ , of Z µ = M µ ζ | H 2 (µ) under Cauchy transform.Some necessary and sufficient conditions for this to hold are obtained in Lemma 9 and Proposition 4. Namely, as described in Subsection 1.1, there is a bijection between contractive analytic functions in the complex unit disk and positive, finite and regular Borel measures on the circle.If a positive measure, µ, corresponds to an extreme point of this compact, convex set of contractive analytic functions, we say that µ is extreme, otherwise µ is non-extreme.As established in Lemma 9 and Proposition 4, the intersection space, int(µ, λ) will be V µ −reducing if (i) λ is non-extreme or if (ii) µ + λ is extreme, and the intersection space will be non-trivial and not V µ −invariant if µ, λ are both extreme but µ + λ is non-extreme.
Given two positive, finite and regular Borel measures, µ and λ, on the complex unit circle, ∂D, one can associate to µ a densely-defined and positive semi-definite sesquilinear or quadratic form in H 2 (λ).Namely, we define the form domain, Dom q µ ⊆ H 2 (λ), as the disk algebra, Dom q µ := A(D), the unital Banach algebra of all uniformly bounded analytic functions in the unit disk which extend continuously to the boundary, equipped with the supremum norm.The disk algebra embeds isometrically into the continuous functions on the circle, C (∂D) and A(D) can be viewed as a dense subspace of H 2 (λ).The quadratic form, q µ : Dom q µ × Dom q µ → C is then defined in the obvious way by integration against µ, As described in Section 3 and Theorem 4, there is a theory of Lebesgue decomposition of densely-defined and positive semi-definite quadratic forms in a Hilbert space, H. Namely, given any such form, there is a unique Simon-Lebesgue form decomposition, q = q ac + q s , where 0 ≤ q ac , q s ≤ q, q ac is absolutely continuous in the sense that it is closeable and it is maximal in the sense that q ac is the largest closeable quadratic form bounded above by q.The form q s is singular in the sense that the only closeable positive semi-definite form it dominates is the identically 0 form.Here, a positive semi-definite quadratic form, q, with dense form domain Dom q in H, is closed, if Dom q is a Hilbert space, i.e. complete, with respect to the norm induced by the inner product q(•, •) + •, • H .A form is then closeable if it has a closed extension.See Subsection 1.3 for an introduction to the theory of densely-defined and positive semi-definite quadratic forms.An immediate question is whether the Simon-Lebesgue decomposition of the form, q µ , in H 2 (λ) coincides with the Lebesgue decomposition of µ with respect to λ. Namely, if µ = µ ac + µ s and q µ = q ac + q s , then is it true that q ac = q µac and q s = q µs ?A complete answer, summarized in the theorem below, is provided in Theorem 9, Theorem 10, Corollary 5 and Corollary 6.
Theorem.If q µ = q ac + q s is the Simon-Lebesgue form decomposition of q µ in H 2 (λ), then If µ = µ ac + µ s is the Lebesgue decomposition of µ with respect to λ, then is a complementary space decomposition in the sense of de Branges and Rovnyak, with H + (µ ac ), H + (µ s ) contractively contained in H + (µ).Moreover, H + (µ ac ) is the largest RKHS, H(k), contractively contained in H + (q ac ) ⊆ H + (µ) so that the closed embedding, e : , is such that τ := ee * is Toeplitz for the image, V λ , of Z λ under Cauchy transform, i.e.V λ * τ V λ = τ .In particular, the Simon-Lebesgue decomposition of the quadratic form, q µ , in H 2 (λ) coincides with the Lebesgue decomposition of µ with respect to λ if and only if int(µ, λ) is V µ −invariant.
In the above, the spaces of q ac and q s −Cauchy transforms are defined in an analogous way to the space of µ−Cauchy transforms, see Subsection 4.1.By Proposition 4, the intersection space, int(µ, λ), is not always V µ −invariant.Example 2 (continued in Example 3) provides a concrete example, where µ = m + and λ = m − are the mutually singular restrictions of normalized Lebesgue measure, m, to the upper and lower half-circles, so that the Lebesgue decomposition of m + with respect to m − has m +;ac = 0 but int(m + , m − ) = {0}, so that q m+;ac = 0.
Remark 1.This 'reproducing kernel approach' to measure theory on the circle and Lebesgue decomposition of a positive measure with respect to Lebesgue measure was first considered and studied in [14,15], in a more general and non-commutative context.
Background
1.1 Function theory in the disk, measure theory on the circle Classical analytic function theory in the complex unit disk and measure theory on the complex unit circle are fundamentally intertwined.There are bijective correspondences between (i) contractive analytic functions in the disk, (ii) analytic functions in the disk with positive semi-definite real part, i.e.Herglotz functions and (iii) positive, finite and regular Borel measures on the complex unit circle.Namely, starting with such a positive measure, µ, its Herglotz-Riesz transform is the Herglotz function,
It is easily verified that
(By the maximum modulus principle, b µ is strictly contractive in D unless it is constant.)Each of these transformations is essentially reversible.Namely, given any contractive analytic function, b, the Cayley transform, H b := 1+b 1−b , is a Herglotz function and the Herglotz representation theorem states that if H is any Herglotz function in the disk, then there is a unique finite, positive and regular Borel measure, µ on the circle, so that see [13,Boundary Values,Chapter 3].To be precise, two Herglotz functions correspond to the same positive measure, µ, if and only if they differ by an imaginary constant.If H 1 , H 2 are two Herglotz functions so that H 2 = H 1 + it for some t ∈ R, then their corresponding inverse Cayley transforms obey so that b 2 is, up to multiplication by the unimodular constant z(t) z(t) , a Möbius transformation, m z(t) , of b 1 , where m z(t) defines an automorphism of the disk interchanging 0 with z(t).
If a contractive analytic function, b, corresponds, essentially uniquely, to a positive measure, µ, in this way, we write µ := µ b , and µ b is called the Clark or Aleksandrov-Clark measure of b [5].Many properties of contractive analytic functions in the disk can be described in terms of corresponding properties of their Clark measures and vice versa [1,2].For example, by Fatou's theorem, the Radon-Nikodym derivative of any Clark measure, µ b , with respect to normalized Lebesgue measure, m, on the circle is given by the radial, or more generally non-tangential, limits of the real part of its Herglotz function, This follows from the characterization of extreme points in the set of contractive analytic functions given in [13, Extreme Points, Chapter 9] and Fatou's Radon-Nikodym formula as described above.Here, equipping the set of all bounded analytic functions in the disk with the supremum norm, we obtain the unital Banach algebra, H ∞ , the Hardy algebra, whose closed unit ball, [H ∞ ] 1 , is the compact and convex set of contractive analytic functions in the disk.It further follows from a well-known theorem of Szegö (later strengthened by Kolmogoroff and Kreȋn), that H 2 (µ) = L 2 (µ) if and only if µ = µ b for an extreme point b ∈ [H ∞ ] 1 [13, Szegö's Theorem, Chapter 4], [23].Namely, Szegö's theorem gives a formula for the distance from the constant function 1 to the closure of the analytic polynomials with zero constant term in L 2 (µ): It follows, in particular, that b is an extreme point so that dµ dm is not log-integrable if and only if 1 belongs to the closure, H 2 0 (µ), in L 2 (µ) of the analytic polynomials obeying p(0) = 0.That is, if and only if is the Clark measure of an extreme point, b, we will say that µ is extreme, and that µ is non-extreme if b is not an extreme point.
The results of this paper reinforce the close relationship between function theory in the disk and measure theory on the circle by establishing the Lebesgue decomposition and Radon-Nikodym theorem for positive measures using functional analysis and reproducing kernel theory applied to spaces of Cauchy transforms of positive measures.We will see that the reproducing kernel construction of the Lebesgue decomposition of a positive measure µ, with respect to another, λ, bifurcates into the two cases, where: the intersection of the spaces of µ and λ−Cauchy transforms, is (i) invariant, or (ii), not invariant for the image of Z µ under Cauchy transform.Moreover, whether or not this intersection space is invariant is largely dependent on whether λ, or µ + λ are non-extreme or extreme.
Reproducing kernel Hilbert spaces
As described in the introduction, a reproducing kernel Hilbert space (RKHS) is any complex, separable Hilbert space of functions, H, on a set X, with the property that the linear functional of point evaluation at any x ∈ X is bounded on H.Further recall, as described above, that for any x ∈ X, there is then a unique kernel vector or point evaluation vector, k x ∈ H so that k x , h H = h(x) for any h ∈ H and we write H = H(k), where k : X × X → C is a positive kernel function on X in the sense of Equation (1).Much of elementary reproducing kernel Hilbert space theory was developed by N. Aronszajn in his seminal paper, [3].In particular, there is a bijective correspondence between RKHS on a set X and positive kernel functions on X given by the Aronszajn-Moore theorem, [3, Part I], [20, Proposition 2.13, Theorem 2.14] and this motivates the notation H = H(k).
Theorem (Aronszajn-Moore).If H = H(k) is a RKHS of functions on a set, X, then k is a positive kernel function on X.Conversely, if k is a positive kernel function on X, then there is a (necessarily unique) RKHS of functions on X with reproducing kernel, k.
Any RKHS, H(k), of functions on a set X, is naturally equipped with a multiplier algebra, Mult(k), the unital algebra of all functions on X which 'multiply' H(k) into itself.That is, g ∈ Mult(k) if and only if g • h ∈ H(k) for any h ∈ H(k).Any h ∈ Mult(k) can be identified with the linear multiplication operator M h : H(k) → H(k).More generally, one can consider the set of multipliers, Mult(k, K), between two RKHS on the same set.If h ∈ Mult(k, K), then M h is always bounded, by the closed graph theorem.Adjoints of multiplication operators have a natural action on kernel vectors: If h ∈ Mult(k, K), then All RKHS in this paper will be RKHS, H(k), of analytic functions in the complex unit disk, D = (C) 1 , with the additional property that evaluation of the Taylor coefficients of any h ∈ H(k) (at 0) defines a bounded linear functional on H(k).Again, by the Riesz representation lemma, for any j ∈ N ∪ {0}, there is then a unique Taylor coefficient kernel vector, k j ∈ H(k), so that if h ∈ H(k) has Taylor series at 0, defines a positive kernel function, the coefficient reproducing kernel of H(k), on the set N ∪ {0}.It is easily checked that for any such Taylor coefficient reproducing kernel Hilbert spaces, H(k) and H(K), of analytic functions in D, The reproducing and coefficient reproducing kernels of a Taylor coefficient RKHS in D are related by the formulas: k(j, ℓ)z j w ℓ , and Adjoints of multipliers also have a natural convolution action on coefficient kernels, if h ∈ Mult(K, k), then, We will say that a RKHS, H(k), of analytic functions in X = D is a coefficient RKHS in D, if Taylor coefficient evaluations define bounded linear functionals on H(k).In this case the positive coefficient kernel function k on N ∪ {0} is an example of a discrete or formal reproducing kernel in the sense of [4].
In this paper it will be useful to consider densely-defined multipliers between RKHS H(K), H(k) on X, which are not necessarily bounded.
Proposition 1 (Multipliers are closeable).Let k, K be positive kernel functions on X, and let h be a function on X so that the linear operator M h : Dom M h ⊆ H(k) → H(K) has dense domain, Dom M h .Then M h is closeable, and closed on its maximal domain, , and x∈X K x is a core for M * h , if M h is defined on its maximal domain.Recall that a linear operator with dense domain in a Hilbert space, H, is said to be closed if its graph is a closed subspace of H ⊕ H.Further recall that a dense set, D ⊆ Dom A, contained in the domain of closed operator, A, is called a core for A if A is equal to the closure (minimal closed extension) of its restriction to D. In general, given any two linear transformations A, B, we say that B is an extension of A or that A is a restriction of B, written A ⊆ B, if Dom A ⊆ Dom B and B| Dom A = A. Equivalently, the set of all pairs (x, Ax), for x ∈ D, is dense in the graph of A. Finally, A is closeable if it has a closed extension.
Proof.Define Dom max M h to be the linear space of all g ∈ H(k) so that h • g ∈ H(K).This is the largest domain on which M h makes sense.If g n ∈ Dom max M h is such that g n → g and M h g n → f , then since H(k), H(K) are RKHS, it necessarily follows that This proves that f = h • g, so that g ∈ Dom max M h and M h is closed on Dom max M h .If M h is denselydefined on some other domain, Dom M h , then Dom M h ⊆ Dom max M h by maximality, so that M h has a closed extension, and is hence closeable.
The fact that K x is a core for M * h follows from the assumption that M h is defined (and closed) on its maximal domain.By maximality, M h , with domain Dom max M h , has no non-trivial closed extensions which act as multiplication by h.Let T * be the closure of the restriction of M * h to K x .Then T * ⊆ M * h is densely-defined and closed so that M h ⊆ T := T * * , where T * * , the adjoint of T * is necessarily closed so that T * = T * .However, , so that T necessarily acts as multiplication by h on its domain.By maximality, Dom T = Dom max M h and M h = T .Remark 2. If H(k) and H(K) are Taylor coefficient RKHS in D, then one can further show that the adjoint of any closed multiplication operator, M h : H(k) → H(K) acts as a convolution operator on coefficient kernels, as in Equation (4), and the linear span of all Taylor coefficient kernels is also a core for M * h .One can define a natural partial order on positive kernel functions on a fixed set, X. Namely, if k and K are two positive kernel functions on the same set, X, we write k ≤ K, if K − k is a positive kernel function on X.Notice that the identically zero kernel function is a positive kernel on X, so that k ≤ K can be equivalently written as K − k ≥ 0. The following theorem of Aronszajn describes when one RKHS of functions on X is boundedly contained in another in terms of this partial order, [3, Section 7] [20, Theorem 5.1].
Theorem (Aronszajn's inclusion theorem).Let k, K be positive kernel functions on a set, X.Then H(k) ⊆ H(K) and the norm of the embedding e : If k and K are both positive kernel functions on a set, X, it is immediate that k + K is also a positive kernel function on X.The following 'sums of kernels' theorem of Aronszajn describes the norm of H(k + K) and the decomposition of this space in terms of H(k) and H(K) [3], [20, Theorem 5.4, Corollary 5.5].Notice, in particular that k, K ≤ k + K as kernel functions so that H(k) and H(K) are contractively contained in H(k + K), by the inclusion theorem.
Theorem (Aronszajn's sums of kernels theorem).Let k, K be positive kernel functions on a set, X.Then, Observe that the sums of kernels theorem asserts that the algebraic sum H(k + K) = H(k) + H(K) is a direct sum if and only if it is an orthogonal direct sum.More can be said about this decomposition and the structure of H(k + K) using the theory of operator-range spaces of contractions and their complementary spaces in the sense of de Branges and Rovnyak [6], [8,Chapter 16].Let A ∈ L (H, J) be a bounded linear operator.The operator-range space of A, R(A), is the Hilbert space obtained by equipping the range of A with the inner product that makes A a co-isometry onto its range.That is, R(A) = Ran A ⊆ J, with inner product, Ax, Ay A := x, P ⊥ Ker A y H .
One can generally show that R(A) = R( √ AA * ), [8,Corollary 16.8].If A is a contraction, A ≤ 1, then R(A) ⊆ J is contractively contained in J in the sense that the embedding, e : R(A) ֒→ J is a linear contraction.In this case, one can define the complementary space of A, R c (A) := R( √ I − AA * ).The notion of complementary space was originally introduced in a more geometric way by de Branges and Rovnyak [6].Namely, if H is any Hilbert space and R ⊆ H is a Hilbert space which is contractively contained in H, then R = R(j), where j : R ֒→ H is the contractive embedding.L. de Branges and J. Rovnyak defined the complementary space, R c of R as the set of all y ∈ H so that sup One can prove that R c = R c (j) and that the above formula is equal to the norm of y in R c (j), so that these two definitions coincide [8,Chapter 16].The following theorem summarizes several results in the theory of operator-range spaces, see [8,Chapter 16].
Theorem 1 (Operator-range spaces of contractions).Let A ∈ L (H, J) be a contraction.If e : R(A) ֒→ J and j : R c (A) ֒→ J are the contractive embeddings, then For any x = y + z ∈ J so that y ∈ R(A) and z ∈ R c (A), the Pythagorean equality, holds if and only if y = ee * x and z = jj * x, so that, in particular, I J = ee * + jj * .As a vector space, the overlapping space is and A : R c (A * ) → R c (A) acts as a linear contraction.
Moreover, the following are equivalent: Observe that, as in Aronszajn's sums of kernels theorem, the algebraic sum Theorem 2. Let H(K) be a RKHS on a set, X.If H(k) is another RKHS on X which embeds, contractively, in H(K), and e : H(k) ֒→ H(K) is the contractive embedding, then H(k) = R(e) and the complementary space, R c (e), is the RKHS on X with reproducing kernel K − k.
Proof.Let e : H(k) ֒→ H(K) be the contractive embedding and consider the operator-range space of e.Given any g, h ∈ H(k), we have that eg, eh e = g, h H(k) , since e is injective.Hence, for any x ∈ X, and it follows that R(e) = H(k).Indeed, equation (6) shows that R(e) is a reproducing kernel Hilbert space on X with point evaluation vectors k x := ek x and that for any x, y ∈ X, Since this complementary space is contractively contained in H(K), for any proving that R c (e) is also a RKHS on X with point evaluation vectors k ′ If j : R c (e) ֒→ H + (µ) is the contractive embedding, then observe that jj * + ee * = I H(K) , so that The previous theorem and Theorem 1 provide additional information on the structure and decomposition of H(k + K) in Aronszajn's sums of kernels theorem.
Corollary 1.Let k, K be positive kernel functions on a set, X and let e : H(k) ֒→ H(k + K) and j : H(K) ֒→ H(k + K) be the contractive embeddings.Then we can identify H(k) and H(K) with the operator range spaces R(e) and R(j), respectively.Moreover, I H(k+K) = ee * + jj * so that H(K) = R c (e) is the complementary space of R(e) = H(k), and given any h ∈ H(k + K), ) and jR c (j * ), and e : R c (e * ) → R c (e) = H(K), j : R c (j * ) → R c (j) are contractions.
Finally, as described in [18] and [17, Section 5], we can define a pair of natural 'lattice operations', ∨ and ∧ on the set of all positive kernel functions on a fixed set, X.Given two positive kernel functions, k and K, on X, let k ∨ K := k + K, a positive kernel function on X.We can also construct a second RKHS on X by defining int(k, equipped with the inner product g, h k∧K := g, h k + g, h K . It is not difficult to verify that int(k, K), equipped with this inner product is complete, and that point evaluation at any x ∈ X defines a bounded linear functional on int(k, K), so that this is a RKHS, H(k ∧ K), of functions on X.The following theorem describes a useful relationship between H(k + K) and H(k ∧ K) [17, Theorem 5.2], [18].
Theorem (Sums and intersections of RKHS).Let k, K ≥ 0 be positive kernel functions on a set, X. Define two linear maps, U ∨ and The point evaluation vectors for H(k ∧ K) = H(k) ∩ H(K) are given by the formulas
Positive quadratic forms
A quadratic or sesquilinear form, q : Dom q × Dom q → C, with dense form domain, Dom q in a separable Hilbert space, H, is said to be positive semi-definite if q(x, x) ≥ 0 for all x ∈ Dom q.Such a quadratic form is said to be closed, if Dom q is complete with respect to the norm induced by the inner product and q is closeable if it has a closed extension.We will let Ĥ(q) denote the Hilbert space completion of Dom q with respect to this q + id−inner product.Hence, q is closed if and only if Ĥ(q) = Dom q.If q ≥ 0 is closeable, then its closure, q, is the minimal closed extension of q.By Kato, closed positive semi-definite forms obey an 'unbounded version' of the Riesz Lemma [16, Chapter VI, Theorem 2.1 and Theorem 2.23].Namely, q ≥ 0 is closed if and only if there is a unique self-adjoint, densely-defined and positive semi-definite operator, A, so that Dom q = Dom √ A and Any self-adjoint operator is necessarily closed.Following Kato and Simon, we can define a partial order on densely-defined and positive semi-definite forms by q 1 ≤ q 2 if (i) Dom q 2 ⊆ Dom q 1 , and In particular, if q A and q B are the closed forms of the self-adjoint and positive semi-definite operators A and B, we say that A ≤ B in the form sense if q A ≤ q B as forms.
for any t > 0.
Recall that if A is a closed operator with dense domain, Dom A ⊆ H, that D ⊆ Dom A is called a core for A, if A is equal to the closure of its restriction to D. Similarly, if q is a closed, densely-defined and positive semi-definite form, a (necessarily dense) set D ⊆ Dom q is called a form-core for q, if D is dense in Ĥ(q).It is not difficult to verify that if q = q A is closed, then D is a form-core for q if and only if D is a core for √ A.
Toeplitz forms.The classical Hardy space, H 2 = H 2 (D), is the Hilbert space of square-summable Taylor series in the complex unit disk, equipped with the ℓ 2 −inner product of these coefficients.By results of Fatou, any element of H 2 has non-tangential boundary limits almost everywhere on the unit circle, ∂D, with respect to normalized Lebesgue measure, m [13].Identifying any h ∈ H 2 with its boundary limits defines an isometric inclusion of H 2 into L 2 = L 2 (m).Classically, Toeplitz operators, T , on H 2 , are defined as the compression of bounded multiplication operators on L 2 to H 2 .Namely, T = T g := P H 2 M g | H 2 , and A theorem of Brown and Halmos, [10, Theorem 6], characterizes the Toeplitz operators as the set of all bounded operators, T , on H 2 which obey the simple algebraic condition, where S = M z , is the shift on H 2 , the isometry of multiplication by z.Under the boundary value identification of H 2 with the subspace H 2 (m) ⊆ L 2 (m, ∂D), the shift is identified with the isometry . Recall, as described in the Outline, given a positive, finite and regular Borel measure, µ, on ∂D, we can associate to µ the densely-defined and positive semi-definite quadratic form, q µ , with dense form domain, Dom , where m is normalized Lebesgue measure.This positive form, q µ , is an example of a Toeplitz form, as studied by Grenander and Szegö in [9].Namely, Dom q µ = A(D) obeys SDom q µ ⊆ Dom q µ , and If q µ is closeable so that q µ = q T for a closed, self-adjoint T ≥ 0, then by Kato's unbounded Riesz lemma we have that S * T S = T , and our results will show that a closed, potentially unbounded Toeplitz operator with symbol f , in this 'quadratic form sense'.In particular, if T ≥ 0 is bounded, which happens if and only if µ ≤ t 2 m for some t > 0, then by the Riesz representation lemma, S * T S = T , so that T is a bounded Toeplitz operator by Brown-Halmos, in which case T = T f for f = dµ dm ∈ L ∞ and f ∞ ≤ t, by Theorem 3 and Corollary 2.
Spaces of Cauchy transforms
Let µ be a positive, finite and regular Borel measure on the complex unit circle.Recall that given any We will call the functions k z , z ∈ D, Szegö kernel vectors.
Recall that A(D) denotes the disk algebra, the unital Banach algebra of analytic functions in D which extend continuously to the boundary, ∂D.Since the analytic polynomials are supremum-norm dense in A(D), viewed as a subspace of the continuous functions, C (∂D), on the circle and (Here, recall that our inner products are conjugate linear in their first argument.)This limit will exist and C µ h will be holomorphic, if and only if the limit of ǫ −1 (k z+ǫ − k z ) exists in H 2 (µ).This limit exists in supremum norm on the circle (and so belongs to A(D)), and so it certainly exists in the L 2 (µ)−norm by Cauchy-Schwarz.Indeed, and this limit is continuous on ∂D for any fixed z ∈ D.
⊆ O(D) be the complex vector space of µ−Cauchy transforms equipped with the inner product, Lemma 3. The space of µ−Cauchy transforms, H + (µ), is a RKHS of analytic functions in D with reproducing kernel where Proof.To show that this inner product is well-defined, we need to check that C µ h ≡ 0 in the disk implies that h = 0 in H 2 (µ).Indeed, and since k z is dense in A(D), the linear span of the Szegö kernels is also dense in H 2 (µ) as described above.Hence this vanishes for all z ∈ D if and only if h = 0.By definition, for any z ∈ D, Finally, establishing the second formula.
Example 1 (Hardy space).If µ = m is normalized Lebesgue measure, then, It follows that b m := Hm−1 Hm+1 ≡ 0, so that m = µ 0 is the Clark measure of the identically 0 function.Moreover, is the Szegö kernel.This is the reproducing kernel for the classical Hardy space H 2 = H 2 (D), of squaresummable Taylor series in the complex unit disk, equipped with the ℓ 2 −inner product of the Taylor coefficients.That is, Since any h := C µ g ∈ H + (µ) is holomorphic in the open unit disk, its Taylor series at 0 has radius of convergence at least one, and it follows that the Taylor coefficients are given by That is, for any j ∈ N∪{0}, the linear functionals ℓ j (h) = ĥj are bounded on H + (µ) and are implemented by inner products against the Taylor coefficient kernel vectors k µ j := C µ ζ j .Hence H + (µ) is a Taylor coefficient RKHS in D with coefficient reproducing kernel, kµ (i, j), on the set N ∪ {0}, and kµ is then a positive kernel function on N ∪ {0}.
Given a positive measure µ, let This is an isometry on H + (µ) that will play a central role in our analysis.This operator has a natural action on kernel vectors: In particular, z∈D where, here, denotes closed linear span.It is easy to check that a function, h ∈ H + (µ), is orthogonal to Ran V µ if and only if h = c1, c ∈ C, is constant in the disk.Hence the following statements are equivalent: and hence V µ is unitary, (iv) H + (µ) does not contain the constant functions.
Lemma 4. Given any finite, positive and regular Borel measure, µ, on ∂D, the co-isometry V * µ , acts as a backward shift on Given any h in the classical Hardy space, H 2 = H + (m), one can check that S := V m = M z is the isometry of multiplication by the independent variable, z, on H 2 , the shift.In this case, adjoint of S is called the backward shift and acts as It is straightforward to verify that if h ∈ H 2 has Taylor series h(z) = ĥj z j , then (S * h)(z) = ∞ j=0 ĥj+1 z j .This motivates the terminology 'backward shift' in the above lemma statement.This lemma is easily verified and we omit the proof.
Absolute continuity in the reproducing kernel sense
Recall that given positive measures µ and λ, we say that µ is dominated by λ if there is a t > 0 so that µ ≤ t 2 λ, and we say that µ is reproducing kernel or RK-dominated by λ, if H + (µ) ⊆ H + (λ) and there is a t > 0 so that the norm of the embedding e µ,λ : H + (µ) ֒→ H + (λ) is at most t, written µ ≤ RK t 2 λ.We will begin this section by showing that these two definitions of domination are equivalent.
Theorem 3. Given positive, finite and regular Borel measures µ, λ on the unit circle, µ ≤ t 2 λ for some t > 0 if and only if µ ≤ RK t 2 λ.
Proof.(Necessity.)If µ ≤ t 2 λ, then γ := t 2 λ − µ is a positive measure and First proof of sufficiency.Conversely, suppose that where the k µ j , j ∈ N ∪ {0} are the Taylor coefficient evaluation vectors, and since similar formulas hold for λ, we obtain that where the K i , i ∈ N ∪ {0} are the Taylor coefficient evaluation vectors in H(K).Namely, H(K) is also a Taylor coefficient RKHS in D so that K(i, j) := K i , K j H(K) defines a positive kernel function on the set in the positive cone of the continuous functions, it follows that ℓ K is a bounded, positive linear functional on C (∂D), with norm ℓ K = ℓ K (1) = t 2 λ(∂D) − µ(∂D) = K(0, 0) ≥ 0. By the Riesz-Markov theorem, there is then a unique, positive, finite and regular Borel measure, γ, on ∂D, so that for any f ∈ C (∂D), i.e. ℓ K = γ, and we conclude that γ = t 2 λ − µ ≥ 0 so that t 2 λ ≥ µ.
Second proof of sufficiency.If t 2 k λ ≥ k µ , then by Aronszajn's inclusion theorem, H + (µ) ⊆ H + (λ) and the norm of the embedding e µ,λ : Observe that e := e µ,λ acts trivially as a multiplier by the constant function 1, so that e * k λ z = k µ z , and e * k λ j = k µ j , for any z ∈ D and j ∈ N ∪ {0}.Hence, so that e * intertwines V λ with V µ , e * V λ = V µ e * .Setting E := C * µ e * C λ , we see that for any monomial, At this point one could argue using the Riesz-Markov theorem as above, however, here is an alternative argument.Since Z λ and Z µ are contractions (they are isometries), we can apply the intertwining version of the commutant lifting theorem [19, Corollary 5.9] to conclude that E can be 'lifted' to a bounded operator Ê : L 2 (λ) → L 2 (µ), with norm Ê = E , so that ÊM λ ζ = M µ ζ Ê, and ζ is a dense set in L 2 (λ), a simple argument shows that T acts as multiplication by f := T 1 ∈ L 2 (λ).However, since T = M f is a bounded and positive semi-definite operator, it is easy to check that t 2 ≥ T = f ∞ , T = Ê 2 = E 2 = T , and f ≥ 0 λ−a.e.Finally, one can also check that This formula extends to elements of the form Again, by Weierstraß approximation, since , which is in turn dense in L 2 (λ) and L 2 (µ), it follows that for any g, h ∈ L 2 (λ), where f ≥ 0, λ−a.e. and f ∞ ≤ t 2 .We conclude that µ ≤ t 2 λ and that is the (bounded) Radon-Nikodym derivative of µ with respect to λ.
Definition 1.Let T ∈ L (H) be a bounded operator and let V be an isometry on H.We say that T is If q ≥ 0 is a positive semi-definite quadratic form with dense form domain, Dom q ⊆ H, we say that q is V −Toeplitz if Dom q is V −invariant and q(V g, V h) = q(g, h); g, h ∈ Dom q.
In particular, if T ≥ 0 is a positive semi-definite, self-adjoint and densely-defined operator in H, we say that T is V −Toeplitz if the closed, positive semi-definite form it generates, If T ≥ 0 is bounded, this latter definition reduces to the definition of a bounded, positive semi-definite V −Toeplitz operator.
Corollary 2. Let µ, λ be positive, finite and regular Borel measures on ∂D so that µ ≤ t 2 λ.In this case, , where f ≥ 0 λ− a.e., f ∞ ≤ t 2 , and f = µ(dζ) λ(dζ) is the Radon-Nikodym derivative of µ with respect to λ. Remark 3.While the co-embedding, E µ,λ : H 2 (λ) ֒→ H 2 (µ) always has dense range, it may have nontrivial kernel.For example, if λ is the sum of two Dirac point masses at distinct points ζ, ξ ∈ ∂D, µ is the point mass at ζ, then µ ≤ λ and if p is any polynomial vanishing at ζ, then E µ,λ p = 0 ∈ H 2 (µ).To be precise, H 2 (µ) is the closure of the disk algebra, A(D), or the polynomials, More generally, absolute continuity of positive measures can also be described in terms of their spaces of Cauchy transforms.It is a straightforward exercise, using the Radon-Nikodym formula, to show that µ is absolutely continuous with respect to λ, if and only if one can construct a monotonically increasing sequence of positive measures, µ n ≥ 0, so that µ n ≤ µ for all n, the µ n ↑ µ increase monotonically to µ, and there is a sequence of positive constants, t n > 0 so that µ n ≤ t 2 n λ.Indeed, this can be readily established by taking the 'join' or point-wise maxima of dµ dλ and the constant functions t 2 n • 1.Since µ n ≤ µ for all n, Aronszajn's inclusion theorem implies that H + (µ n ) ⊆ H + (µ) and that the embeddings e n : H + (µ n ) ֒→ H + (µ) are all contractive.Moreover, and again by Aronszajn's inclusion theorem, each H + (µ n ) ⊆ H + (λ) is boundedly contained in H + (λ) so that Proof.We have that for all n, 0 ≤ µ n ≤ µ and µ n ↑ µ.Moreover, H + (µ n ) ⊆ int(µ, λ) for all n.If 1 ≥ g n ≥ 0, µ − a.e. are the Radon-Nikodym derivatives of the µ n with respect to µ, and p ∈ C[ζ], let C n := C µn and let e n : H + (µ n ) ֒→ H + (µ).Then, by the Lebesgue monotone convergence theorem.In conclusion, where, here, denotes closed linear span.
This motivates the following definitions: Definition 2. Let µ, λ be finite, positive and regular Borel measures on ∂D.We say that µ is absolutely continuous with respect to λ in the reproducing kernel sense, µ ≪ RK λ, if the intersection space, We say that µ is reproducing kernel singular with respect to λ, written µ ⊥ RK λ, if the intesection space is trivial, int(µ, λ) = {0}.
By the previous proposition, µ ≪ λ implies that µ ≪ RK λ.The main result of this section will be to show that this new 'reproducing kernel' definition of absolute continuity is equivalent to the classical one.To check that .
As the polynomials are a core for E = E µ,λ , this calculation holds on Dom E = Dom √ T .Moreover, since Dom E = Dom √ T , by polar decomposition of closed operators, and since C[ζ] and K D are Z λ −invariant cores for E, they are also cores for √ T , and it follows that Dom In this case, the self-adjoint λ−Toeplitz operator T µ ≥ 0 is Toeplitz with respect to the unitary M λ ζ .That is, , or, equivalently, This shows that T µ , and hence √ T µ are affiliated to the commutant of the unitary operator Tµ1 acts as multiplication by T µ 1 =: f ∈ L 2 (λ).Since T µ ≥ 0, we necessarily have that f ≥ 0, λ − a.e., and we conclude that for any polynomials p, q, p, q H 2 (µ) = T µ p, T µ q As in the proof of sufficiency in Theorem 3, we conclude that the above formula holds for any g, h ∈ which is dense in C (∂D) and L ∞ (µ).In particular, the formula holds for all simple functions and characteristic functions of Borel sets.Since f ∈ L 2 (λ), f 2 ∈ L 1 (λ) and it follows that is the Radon-Nikodym derivative of µ with respect to λ.
To prove that absolute continuity in the reproducing kernel sense is equivalent to absolute continuity in general, we will appeal to B. Simon's Lebesgue decomposition theory for positive quadratic forms in Hilbert space [22], [21,Supplement to VIII.7].Let Ĥ(q) be the Hilbert space completion of Dom q with respect to the inner product •, • H + q(•, •), and let j q : Dom q ֒→ Ĥ(q) denote the formal embedding.Further define the co-embedding E q : Ĥ(q) ֒→ H by E q (j q (x)) := x, x ∈ Dom q.
By construction, j q is densely-defined, has dense range, and E q is contractive with dense range in H. Hence E q extends by continuity to a contraction, also denoted by E q , E q : Ĥ(q) ֒→ H. Lemma 6.A densely-defined and positive semi-definite quadratic form, q, in H, is closeable if and only if j q is closeable, or equivalently, if and only if E q is injective.
This lemma is a straightforward consequence of the definitions, see also [22].
Theorem 4 (Simon-Lebesgue decomposition of positive forms).Let q ≥ 0 be a positive semi-definite quadratic form with dense form domain, Dom q, in a separable, complex Hilbert space, H. Then q has a unique Lebesgue decomposition, q = q ac + q s , where 0 ≤ q ac , q s ≤ q in the quadratic form sense, q ac is the maximal absolutely continuous form less than or equal to q and q s is a singular form.
If P s denotes the projection onto Ker E q , and P ac = I − P s , then q ac is given by the formula, q ac (x, y) = j q (x), (P ac − E * q E q )j q (y) Ĥ(q) = j q (x), P ac j q (y) Ĥ(q) − x, y H .
In the above theorem statement, recall that we defined the notions of an absolutely continuous or singular positive quadratic form in the introduction.Namely, a positive semi-definite and densely-defined quadratic form, q : Dom q × Dom q → H, Dom q ⊆ H, is called absolutely continuous if it is closeable, and singular if the only absolutely continuous and positive semi-definite form it dominates is the identically zero form.Remark 4. If, now, µ, λ ≥ 0 are measures on the circle, we can take H := L 2 (λ) or H 2 (λ), and define q µ ≥ 0 on a dense form domain in H by the formula For example, if H = L 2 (λ), one can take Dom q µ = C (∂D), the continuous functions.In this case, by the remark on [22, p. 381], the quadratic form Lebesgue decomposition of q µ coincides with the classical Lebesgue decomposition of µ with respect to λ. Namely, in this case, the absolutely continuous part of q µ , q µ;ac is equal to q µac , the positive form of the absolutely continuous part of µ with respect to λ, µ ac , and q µ;s = q µs .In particular, q T := q ac is the form of the positive semi-definite, self-adjoint operator T = M f ≥ 0, where f ∈ L 1 (λ) is the Radon-Nikodym derivative of µ with respect to λ.This follows because, as observed by Simon, in this case his construction of the absolutely continuous and singular parts of q µ essentially reduces to von Neumann's functional analytic proof of the Lebesgue decomposition and Radon-Nikodym theorem in [24,Lemma 3.2.3].See also [12,Section 5], which arrives at the same conclusion with the choice of form domain, Dom q µ ⊆ L 2 (λ), equal to the simple functions, i.e. linear combinations of characteristic functions of Borel sets.
Theorem 5. Let q ≥ 0 be a densely-defined and positive semi-definite quadratic form in a separable complex Hilbert space, H.If q T = q ac is the closure of q ac , then (I + T ) −1 = E q E * q , where E q : Ĥ(q) ֒→ H is the contractive co-embedding.Lemma 7. Let A : Dom A ⊆ H → H be a densely-defined linear operator.Then A is closeable if and only if the positive semi-definite quadratic form q A * A (x, y) := Ax, Ay H , with form domain Dom q A * A := Dom A, is closeable.
In the above statement, note that A * A is not defined if A is not closeable.
Proof of Theorem 5. Let (x j ) ∞ j=1 ⊆ Dom q be a sequence with dense linear span.Apply Gram-Schimdt orthogonalization to (x j ) with respect to the q + id−inner product of Ĥ(q).This yields a countable basis (y j ) ∞ j=1 ⊆ Dom q, so that the sequence (j q (y j )) is an orthonormal basis of Ĥ(q).Hence, By [22, Theorem 2.1, Corollary 2.3], see Theorem 4 and Equation ( 10) above, q ac (x, y) + x, y H = q I+T (x, y) = √ I + T x, √ I + T y H = j q (x), P ac j q (y) Ĥ(q) , for any x, y ∈ Dom q ⊆ Dom q ac ⊆ Dom √ I + T .Hence, for any x, y ∈ Dom T , ), j q (y j ) q+id j q (y j ), P ac j q (y) q+id = P ac j q (x), P ac j q (y) q+id = q I+T (x, y) That is, the (closeable) quadratic forms of I + T and (I + T )E q E * q (I + T ) agree on Dom T , which is a core for √ I + T , and a form-core for q I+T .Moreover, That is, the (bounded) positive quadratic form of the identity, I, agrees with the quadratic form of √ I + T E q E * q √ I + T on the dense subspace Dom √ I + T .Here, if V := E * q √ I + T , this is a closeable operator by Lemma 7. In fact, V extends by continuity to an isometry, since q V * V = q I | Dom √ I+T .Moreover, E q , and hence E q E * q have dense range, so that E q E * q √ I + T extends to an isometry with dense range, i.e. a unitary.For any x, y ∈ Dom √ I + T , we have that Hence, by definition of the adjoint, for any y ∈ Dom Hence for any and we conclude, by the Riesz lemma for bounded sesquilinear forms, that Theorem 6.Let µ, λ be positive, finite and regular Borel measures on ∂D.Then µ ≪ λ if and only if µ ≪ RK λ.
This calculation shows that the 'compression' of I + T to the intersection of its domain with the subspace H 2 (λ) is equal to T , in this quadratic form sense.In particular T − I ≥ 0 is the compression of T = M f to H 2 (λ), where f ∈ L 1 (λ) is the Radon-Nikodym derivative of µ with respect to λ.In conclusion, for any polynomials p, q, As in the second proof of sufficiency of Theorem 3, this equality can be extended to arbitrary g, h
Lebesgue decomposition via reproducing kernels
By Theorem 6, our definition of reproducing kernel absolute continuity is equivalent to the classical definition of absolute continuity for finite, positive and regular Borel measures on the complex unit circle.In particular, if µ ≪ λ, it follows that the intersection space of µ and λ−Cauchy transforms is dense in the space of µ−Cauchy transforms.Hence, if µ = µ ac + µ s is the Lebesgue decomposition of µ with respect to λ, then int(µ ac , λ) is dense in H + (µ ac ), and since µ ≥ µ ac , int(µ ac , λ) ⊆ int(µ, λ).That is, if µ ac = 0, it follows that int(µ, λ) = {0} is not trivial.This raises several natural questions: How can we identify the space of µ ac −Cauchy transforms?Is int(µ, λ) −µ := int(µ, λ) − • µ equal to the space of µ ac −Cauchy transforms?We will see that the answer to the second question is positive if λ is non-extreme, but that in general, int(µ, λ) −µ is not the space of Cauchy transforms of any positive measure, see Corollary 6 and Example 3.
In this case, M = H + (γ) = R(e), and the complementary space of is trivially a (contractive) multiplier so that, as before, e * k µ z = k γ z , and e * V µ = V γ e * .
Conversely, if τ = ee * is V µ −Toeplitz and contractive, then, as in the proof of Theorem 3, T := C * µ ee * C µ is a contractive Z µ −Toeplitz operator and we can appeal to the Riesz-Markov theorem to show that there is a γ ≥ 0, so that , where f ≥ 0, f ∞ ≤ 1 is the Radon-Nikodym derivative of γ with respect to µ. Namely, one can define a linear functional, μT , on It is easy to check that μT is bounded and positive using the Fejér-Riesz theorem, as in the proof of Theorem 3. The fact that T is a positive semi-definite Z µ −Toeplitz contraction ensures that μT extends to a bounded, positive linear functional on C (∂D), and that μT ≤ μ, so that μT = γ for some finite, regular and positive Borel measure, γ ≤ µ, by the Riesz-Markov theorem.
By Theorem 2, the complementary space, R c (e), of R(e) = H + (γ) is a RKHS in D with reproducing kernel k ′ (z, w) = k µ (z, w) − k γ (z, w) and it is contractively contained in H + (µ), by the inclusion theorem.Moreover, if j : R c (e) ֒→ H + (µ) is the contractive embedding, then it follows that jj * = I − ee * ≥ 0 is also a positive semi-definite V µ −Toeplitz contraction.Hence, by the first part of the proof, H = H + (ν) for a positive measure, ν.Finally, since k µ = k γ + k ν , we obtain that µ = γ + ν.Lemma 8. Given any µ, λ, the intersection space int(µ, λ), is both V λ and V µ −co-invariant, and Proof.This is immediate, by Lemma 4, since both V * µ and V * λ act as 'backward shifts' on power series.
Theorem 8. Let µ, λ ≥ 0 be finite, positive and regular Borel measures on ∂D.If the intersection space, int(µ, λ), is V µ −reducing and µ = µ ac + µ s is the Lebesgue decomposition of µ with respect to λ, then
In this case,
H + (µ ac ) = int(µ, λ) −µ , and That is, µ ac is the largest positive measure ≤ µ which is RK-ac with respect to λ, and µ s is RK-singular with respect to λ.
In particular, int(µ, λ) will be V µ −reducing if λ is non-extreme by Lemma 9. Hence, ∨ is also reducing.The previous lemma now implies that int(µ, λ) is V µ −reducing.
If, on the other hand, µ, λ are both extreme but µ + λ is not, then V µ , V λ are both unitary but V µ+λ is not.Hence, since 1 ⊥ Ran V µ+λ , 1 ∈ H + (µ + λ), we have that We can calculate some vectors in int(m + , m − ) more explicitly.By the proof of Proposition 4, we have that Here, 1 = k m 0 , where m = m + + m − , so that Since the unitaries V * ± both act as backward shifts on power series, we can compute these elements of the intersection space explicitly.First, the kernel vectors of H + (m ± ) at 0 are: where log is the branch of the logarithm fixed by the choice of the argument function taking values in [0, 2π).
Here, the branch cut is along the positive real axis, and is strictly negative for any z ∈ D so that this formula defines a holomorphic function in D. (We know, of course, that k + 0 must be holomorphic in D.) Since Also note that ± act as backward shifts on power series, it follows that as required.
Lebesgue decomposition of measures and their forms
As described in Remark 4 and Subsection 1.3, if µ, λ ≥ 0 are positive, finite and regular Borel measures on the unit circle, ∂D, then one can construct the Lebesgue decomposition of µ with respect to λ by considering the densely-defined positive quadratic form, q µ : C (∂D) × C (∂D) → 0, with dense form domain C (∂D) ⊆ L 2 (λ), the continuous functions on the unit circle.Namely, applying the Simon-Lebesgue decomposition to q µ , viewed as a positive, densely-defined form in L 2 (λ), one obtains, q µ = q µ;ac + q µ;s , where q µ;ac is an absolutely continuous (closeable) form and q s is a singular form and moreover, q µ;ac = q µac , q s = q µs , where µ = µ ac + µ s is the Lebesgue decomposition.However, in this paper, since we wish to apply analytic and function theoretic methods, we instead consider the positive quadratic Z λ −Toeplitz form, q µ , associated to µ ≥ 0, with dense form domain Dom q µ = C[ζ] or Dom q µ = A(D), in H 2 (λ) ⊆ L 2 (λ).As we will show, if q µ = q ac + q s is the Simon-Lebesgue form decomposition of q µ in H 2 (λ), then one can define reproducing kernel Hilbert spaces of q ac and q s −Cauchy transforms, H + (q ac ) and H + (q s ).The goal of this subsection is to compare the Lebesgue decomposition of µ with respect to λ with the Simon-Lebesgue decomposition of q µ in H 2 (λ).
Let µ, λ ≥ 0 be finite and regular Borel measures on ∂D.Consider the positive quadratic form, q µ , with dense form domain, A(D) ⊆ H 2 (λ).Observe that Ĥ(q µ ) = H 2 (µ + λ) so that C[ζ] and K D are both dense sets in this space.Consider the Simon-Lebesgue decomposition, q µ = q ac + q s , of q µ in H 2 (λ).By Theorem 4, q ac ≥ 0, is the largest closeable quadratic form bounded above by q µ .Since q ac ≤ q µ , this implies that Dom q µ = A(D) ⊆ Dom q ac , and if q D = q ac denotes the closure of q ac , then A(D) must be a form-core for the closed form q D by the maximality statement in Theorem 4. We define H 2 (q ac ), H 2 (q s ) as the Hilbert space completion of the disk algebra, A(D), modulo vectors of zero length, with respect to the pre-inner products, q ac , q s , respectively.Since 0 ≤ q ac , q s ≤ q µ , we can define the contractive co-embeddings E ac : H 2 (µ) ֒→ H 2 (q ac ) and E s : H 2 (µ) ֒→ H 2 (q s ) by E ac a = a ∈ H 2 (q ac ) and E s a = a ∈ H 2 (µ s ).(Here, an element a ∈ A(D) could be equal to 0 as an element of H 2 (µ), or as an element of the spaces H 2 (q ac ), H 2 (q s ).
However, the inequality 0 ≤ q ac , q s ≤ q µ , ensures that if a ∈ A(D) is zero as an element of H 2 (µ), i.e. it vanishes µ − a.e., then a = 0 as element of both H 2 (q ac ) and H 2 (q s ).A more precise notation would be to let N ac denote the subspace of all elements of A(D) of zero-length with respect to the q ac −pre-inner product so that equivalence classes of the form a + N ac , a ∈ A(D), are dense in H 2 (q ac ). ) Observe that if D ⊆ A(D) is any supremum-norm dense set, such as then D is dense in H 2 (µ), and since the co-embedding E ac : H 2 (µ) ֒→ H 2 (q ac ) is a contraction with dense range, D will be dense in H 2 (q ac ) and it will be similarly dense in H 2 (q s ).
Lemma 11.If q D = q ac is the closure of q ac , and D ⊆ A(D) is supremum-norm dense, then D is a core for √ D.
Proof.Since Dom q µ = A(D), and q ac ≤ q µ is the largest closeable and positive semi-definite quadratic form, A(D) is a form-core for q D , and hence a core for √ D. Hence, A(D) is dense in Ĥ(q D ) = Ĥ(q ac ).Given any a ∈ A(D), let x n ∈ D be a sequence which converges to a in supremum-norm.Then This proves that D is dense in the dense subspace A(D) ⊆ Ĥ(q D ), and hence D is a form-core for q D and a core for √ D.
Given any h ∈ H 2 (q ac ) or in H 2 (q s ), we can now define the q ac or q s −Cauchy transform of h as before: and similarly for q s .As in Lemma 2 and Lemma 3, Cauchy transforms of elements of H 2 (q ac ), H 2 (q s ) are holomorphic in the unit disk, and if we equip the vector space of q ac −Cauchy transforms with the inner product C ac x, C ac y ac := q ac (x, y), we obtain a reproducing kernel Hilbert space of analytic functions in the disk, H + (q ac ) with reproducing kernel: k (ac) (z, w) := q ac (k z , k w ).
Finally, since q µ = q ac + q s , q µ ≥ q ac , q s ≥ 0, we obtain the following.
Proposition 5.The RKHS of q ac and q s −Cauchy transforms are contractively contained in H + (µ) = H + (q µ ) and k µ = k (ac) + k s so that Moreover, if e ac : H + (q ac ) ֒→ H + (µ) and e s are the contractive embeddings, then Proof.To check the decomposition of the identity, it suffices to calculate so that k µ∩λ ≤ k ac , and by Aronszajn's inclusion theorem, int(µ, λ) −µ is contractively contained in H + (q ac ) which is in turn contractively contained in H + (µ).Hence, if e 1 is the first embedding into H + (q ac ) and e 2 is the second embedding into H + (µ), the composite embedding, e = e 2 e 1 : int(µ, λ) −µ ֒→ H + (µ) is again a contractive embedding and it must be isometric since int(µ, λ) −µ is a closed subspace of H + (µ).It follows that e 1 must be an isometric embedding.Indeed, if there is a unit vector x so that e 1 x < 1 then Similarly e 2 must be isometric on the range of e 1 .On the other hand, since int(q ac , λ) := H + (q ac ) ∩ H + (λ) is dense in H + (q ac ) and H + (q ac ) is contractively contained in H + (µ), we must have that int(q ac , λ) ⊆ int(µ, λ) ⊆ Ran e 1 .Hence, by the previous argument, since int(q ac , λ) ⊆ Ran e 1 is dense in H + (q ac ) and e 2 is isometric on the range of e 1 , e 2 : H + (q ac ) ֒→ H + (µ) is also an isometric inclusion.In conclusion, int(µ, λ) −µ and H + (q ac ) are both closed subspaces of H + (µ), int(µ, λ) −µ is a closed subspace of H + (q ac ) and int(q ac , λ) ⊆ int(µ, λ) is dense in H + (q ac ) so that int(µ, λ) −µ = H + (q ac ).It follows that q B = q D on K D so that by Lemma 12 and the uniqueness of representation of closed forms, D = B.
Corollary 5.If µ, λ ≥ 0 are finite, positive and regular Borel measures on ∂D and q µ is the densely-defined positive quadratic form associated to µ with form domain A(D) ⊆ H 2 (λ), then the space of µ−Cauchy transforms decomposes as the orthogonal direct sum, In particular, H + (q s ) ∩ int(µ, λ) = {0}.
Proof.By Proposition 5 and Theorem 4, we have that the identity operator on H + (µ) decomposes as and H + (q ac ) = int(µ, λ) −µ is a closed subspace of H + (µ) so that the contractive embedding, e ac : H + (q ac ) ֒→ H + (µ) is an isometry.Hence, P ac := e ac e * ac is an orthogonal projection onto the range of e ac and hence P s = I − P ac = e s e * s is the projection onto the orthgonal complement of Ran e ac in H + (µ).It follows that e s is also an isometric embedding and that we can identify H + (q ac ), H + (q s ) with the ranges of these isometric embeddings so that H + (µ) = H + (q ac ) ⊕ H + (q s ).Corollary 6.Let µ, λ be positive, finite and regular Borel measures on the unit circle.The Lebesgue decomposition of µ with respect to λ, µ = µ ac + µ s , coincides with the Simon-Lebesgue decomposition of q µ with form domain Dom q µ = A(D) in H 2 (λ), q µ = q ac + q s , in the sense that q ac = q µac and q s = q µs if and only if int(µ, λ) is V µ −reducing.
Remark 5.More generally, one can apply the methods of this section to construct a Lebesgue decomposition for pairs of positive kernel functions k, K on the same set, X, see Appendix A.
Example 3 (Lebesgue measure on the half-circles).As before, let m ± denote normalized Lebesgue measure restricted to the upper and lower half-circles.These are mutually singular measures so that m + = m +;s is the singular part of m + with respect to m − , and yet by Example 2, int(m + , m − ) = {0}, so that q + = q m+ has a Simon-Lebesgue decomposition q + = q ac + q s in H 2 (m − ), where q ac is non-trivial, by Theorem 9.Moreover, in this example, m − is extreme, so that H 2 (m − ) = L 2 (m − ).This means that while the quadratic form, q µ , associated to µ, with dense form domain, A(D) ⊆ L 2 (m − ) = H 2 (m − ) has non-zero absolutely continuous part, if we instead define the form domain of q µ to be Dom q µ = C (∂D), then, with this form domain, q µ has vanishing absolutely continuous part (since the decompositions of q µ and µ always coincide in this case, see Remark 4).This shows, that in dealing with these unbounded positive quadratic Toeplitz forms, the choice of form domain is crucial!
Lebesgue decomposition for arbitrary measures
The question remains: If µ, λ ≥ 0 are arbitrary, how can we construct the Lebesgue decomposition of µ with respect of λ using reproducing kernel theory and their spaces of Cauchy transforms?If λ is non-extreme, or more generally if int(µ, λ) is V µ −reducing, Theorem 8 provides a satisfying answer.However, as Proposition 4, Example 2 and Theorem 9 show, the intersection of the spaces of µ and λ Cauchy transforms cannot be reducing in general, and that there are examples of pairs of positive measures µ, λ, for which int(µ, λ) cannot be equal to, or even contain, the space of Cauchy transforms of any non-zero positive measure.
By Theorem 6, we do know that if µ = µ ac + µ s is the Legbesgue decomposition of µ with respect to λ, that µ ac ≪ RK λ so that int(µ ac , λ) ⊆ int(µ, λ) ⊆ int(µ, λ) −µ = H + (q ac ).The final result below provides an abstract characterization of the Lebesgue decomposition for arbitrary pairs of positive measures.
Proof.This follows from the definition of q ac , Theorem 2 and Theorem 7. Remark 6.In the case where the complementary space decomposition of H + (µ) = H + (µ ac ) + H + (µ s ), appearing in the above theorem statement, is not an orthogonal direct sum, this yields a corresponding decomposition of the quadratic form q µ , q µ = q µac + q µs , ( where q µac < q ac and q µ = q ac + q s is the Simon-Lebesgue decomposition of q µ .In this case, the decomposition of Equation ( 13) is an example of a 'psuedo-orthogonal' Lebesgue decomposition of q µ as recently defined and studied in [11].
The previous theorem is, while interesting, admittedly not very practical for construction of the Lebesgue decomposition of µ with respect to λ.A simpler, albeit somewhat ad hoc, approach using our reproducing kernel methods is simply to 'add Lebesgue measure'.Namely, if µ ac;λ is the absolutely continuous part of µ with respect to λ, then µ ac;λ = µ ac;λ+m − µ ac;m and both λ + m and m are non-extreme so that Theorem 8 applies.
A Lebesgue decomposition of positive kernels
Let K be a fixed positive kernel function on a set, X.Given any other positive kernel, k, on X, we can associate to it the densely-defined and positive semi-definite quadratic form, q k : Dom q k × Dom q k → C, with dense form domain Dom q k := x∈X K x in H(K), q k (K x , K y ) := k(x, y).
One can then apply B. Simon's Lebesgue decomposition of positive quadratic forms to q k .Such a Lebesgue decomposition of positive kernels was first considered in [12,Section 7,Theorem 7.2].The theorem below provides some more details about this decomposition.
Theorem 11.Let k, K be positive kernel functions on a set, X.If q k is the densely-defined positive quadratic form of k in H(K), as defined above, with Simon-Lebesgue form decomposition q k = q ac + q s , then there are positive kernels, k ac and k s on X, so that q ac = q k (ac) , q s = q k s , k = k ac + k s , and Moreover, H(k ac ) = int(k, K) −k := (H(k) ∩ H(K)) − • k , and if e : int(k, K) ֒→ H(K) is the (closed) embedding, then q ac = q ee * .
Proof.Let h := n i=1 c i K xi be any finite linear combination of the kernels K xi , {x i } n i=1 ⊆ X.Then, since q ac ≤ q k , we obtain that c i c j k(x i , x j ) = c i c j q k (x i , x j ) = q k (h, h) ≥ q ac (h, h) ≥ 0, where 0 ≤ q ac (h, h) = c i c j q ac (K xi , K xj ).
It follows that k ac (x, y) := q ac (K x , K y ), defines a positive kernel function on X so that 0 ≤ k ac ≤ k.Similarly, k s (x, y) := q s (K x , K y ) defines a positive kernel function on X so that 0 ≤ k s ≤ k, and since q k = q ac + q s , we obtain that k ac + k s = k.By definition, q ac is the largest closeable quadratic form bounded above by q k .In particular q ac = q D is the positive form of some densely-defined, self-adjoint and positive semi-definite operator D, so that K x := x∈X K x is a core for √ D. (Here, denotes non-closed linear span.)If e : int(k, K) ⊆ int(k, K) −k ֒→ H(K) is the densely-defined and closed embedding, let A := ee * .We claim that A = D. First, A ≥ 0 is self-adjoint, hence closed, and since e is trivially a multiplier, we obtain that q A (K x , K y ) = e * K x , e * K y k = k ∩ x , k ∩ y k = k ∩ (x, y), where k ∩ denotes the reproducing kernel of the subspace int(k, K) −k ⊆ H(k), the closure of the intersection space, int(k, K) in H(k).In particular, since k ∩ x = P ∩ k x , where P ∩ : H(k) → int(k, K) −k is the orthogonal projection, it follows that k ∩ ≤ k, and hence that q A ≤ q k .Since q A | KX is closeable, it follows, by maximality of the Simon-Lebesgue decomposition, that q A ≤ q D .This inequality implies that k ∩ ≤ k ac as positive kernels on X.
Re H µ (z) ≥ 0, is a positive harmonic function.Applying the inverse Cayley transform to any Herglotz function, i.e. the Möbius transformation sending the open right half-plane onto the open unit disk, D, which interchanges the points 1 and 0, yields a contractive analytic function, b µ , in the disk, a positive kernel.View the analytic polynomials, C[ζ], as a dense subspace of the disk algebra A(D), embedded isometrically in the Banach space C (∂D).For any finite, positive and regular Borel measure on the complex unit circle, µ, we define the positive linear functional, μ on C (∂D) by μ(f ) := ∂D f dµ.(The map µ → μ is a bijection, by the Riesz-Markov theorem.)We then define a bounded linear functional, ℓ K on C (∂D) by ℓ K := t 2 λ − μ.By Weierstraß approximation, C[ζ] + C[ζ] is supremum-norm dense in the continuous functions, C (∂D).Since the Fejér kernel is positive semi-definite, the partial Cesàro sums of any positive semi-definite f ∈ C (∂D) will be a positive trigonometric polynomial, i.e. a positive semi-definite element of C[ζ] + C[ζ] and, by Fourier theory, it follows that the positive cone of C[ζ] + C[ζ] is supremum norm-dense in the positive cone of C (∂D).Moreover, by the Fejér-Riesz theorem, any positive trigonometric polynomial, p + q ≥ 0, on ∂D factors as |g| 2 for an analytic g ∈ C[ζ] (and necessarily, p = q, deg(p) = deg(g)).Hence, to check that ℓ K is a positive linear functional on C (∂D), it suffices to check that ℓ K (p + p) ≥ 0 for any p + p = |g| 2 ≥ 0, p, g ∈ C[ζ].If p = n j=0 pj ζ j and g = n j=0 ĝj ζ j , then by construction for any g ∈ C[ζ], or, equivalently, ℓ K (p + p) = ∂D (p + p)(t 2 dλ − dµ) ≥ 0, for any positive semi-definite p + p ∈ C[ζ] + C[ζ].By density of the positive cone of C[ζ] + C[ζ]
all h ∈ Dom T µ .Hence the quadratic forms for (M λ ζ ) * T µ M λ ζ and T µ are the same.By uniqueness of the unbounded Riesz representation, (M λ ζ ) * T µ M λ ζ = T µ , so that, by Lemma 1, and hence int(µ, λ) is not V µ −reducing by the previous lemma.Example 2 (Lebesgue measure on the half circles).Let m ± be normalized Lebesgue measure restricted to the upper and lower half-circles.Then m = m + + m − , and m + ⊥ m − .Note that both m ± are extreme since dm± dm = χ ∂D± , where χ Ω denotes the characteristic function of a Borel set, Ω, is not log-integrable (with respect to m).On the other hand, m is non-extreme.By the previous proposition, int(m+ , m − ) = {0} is non-trivial, and yet m + ⊥ m − .If int(m + , m − ) contained a non-trivial V + := V m+ or V − := V m− −reducing subspace, M , then the closure, M + or M − in the norms of H + (m ± ) would be a closed V + or V − −reducing subspace.In the first case, Corollary 3 would then imply that M + = H + (γ) for some 0 ≤ γ ≤ m + .On the other hand, int(γ, m − ) ⊇ M is dense in M + = H + (γ) so that γ ≪ RK m − .Since RK-absolute continuity is equivalent to absolute continuity by Theorem 6, this contradicts the mutual singularity of m + and m − .A symmetric argument shows that int(m + , m − ) cannot contain a non-trivial V − −reducing subspace either.Similarly, m = m + + m − can be viewed as the Lebesgue decomposition of m with respect to m + .In this case, int(m, m + ) = H + (m + ) = {0} since m + ≤ m.However, int(m, m + ) cannot be S = V m −reducing as then its closure, int(m, m + ) −m would be a closed, S−reducing subspace of H 2 = H + (m) and the shift has no non-trivial reducing subspaces.(Hence this intersection space cannot contain any non-trivial S−reducing subspace.)In fact, int(m + , m − ) cannot (contractively) contain the space of γ−Cauchy transforms of any non-zero positive measure, γ, as then γ ≪ RK m + and γ ≪ RK m − , so that γ ≪ m + , m − by Theorem 6 and γ ≡ 0 since m + and m − are mutually singular.Finally, we cannot have int(m, m + ) dense in H 2 either as this would imply that m ≪ RK m + which would imply that m ≪ m + by Theorem 6.
z , (e ac e * ac + e s e * s )k µ w µ .
Now suppose that h ∈ Dom D ⊆ H(K) and choose h n ∈ K X = x∈X K x so that h n → h and √ Dh n → √ Dh. (This can be done since Dom D is a core for √ D.) If h n = mn j=1 c j (n)K xj (n) , a finite linear combination, then note that (Dh)(x) = lim n↑∞ c j (n) √ DK x , √ DK xj(n) K = lim c j (n)k ac (x, x j (n)) = lim g n (x),whereg n = c j (n)k ac xj (n) ∈ H(k ac ) ⊆ H(k).c i (n)c j (n)k ac (x i (n), x j (n)) 13, Fatou's Theorem, Chapter 3].As a corollary of this formula, we see that b is inner, i.e. it has unimodular radial boundary limits m−a.e. on the circle, if and only if its Radon-Nikodym derivative vanishes almost everywhere, i.e. if and and only if its Clark measure is singular with respect to Lebesgue measure.As a second example which will be relevant for our investigations here, b is an extreme point of the closed convex set of contractive analytic functions in the disk if and only if its Radon-Nikodym derivative with respect to Lebesgue measure is not log-integrable.That is, b is an extreme point if and only if log | 19,282.2 | 2023-12-04T00:00:00.000 | [
"Mathematics"
] |
Effect of Sn content on the mechanical properties and corrosion behavior of Mg-3Al-xSn alloys
The effect of Sn content on the mechanical properties and corrosion behavior of Mg-3Al-xSn alloys was investigated by SEM-EDXS, XRD, electrochemical measurements, and scanning Kelvin probe force microscopy (SKPFM). The results showed that when the Sn content was 1.4 wt%, Sn dissolved in the α-Mg matrix and then precipitated as an intermetallic compound (Mg2Sn). The combined results of mass loss, hydrogen evolution, and electrochemical measurements indicated that Mg-3Al-1Sn had a low corrosion rate. The SKPFM results showed that the Volta potential of Mg2Sn particles, Al-Mn, and β-Mg17Al12 phases were 100, 80, and 50 mV higher than the matrix, respectively. Therefore, the Mg2Sn phase that formed in Mg-3Al-xSn served as a local cathode due to its high potential, which accelerated microgalvanic corrosion along with the secondary local cathode (Al-Mn). The Sn solution strengthening and secondary phase strengthening (fine Mg2Sn particles) improved the mechanical properties of the Mg-3Al-xSn alloys.
Introduction
Over the past decade, Mg-Al series alloys have been widely used due to their low density, high specific strength, good electrical conductivity, and easy mechanical processing [1][2][3]. However, their poor mechanical properties and corrosion resistance have restricted their applications in automobiles, 3C products, and aeronautical fields [4][5][6]. Therefore, minor alloying elements are often added to modify the alloy microstructure and form new intermetallic compounds, which improves their tensile strength and corrosion resistance.
Sn is relatively cheap and can improve the corrosion resistance, mechanical properties, and castability of Mg alloys [7][8][9][10][11]. Park et al [12] revealed that when Sn was added into Mg-5Al-1Zn alloys, it stabilized the Mg(OH) 2 layers on the alloy surface, which dramatically improved the corrosion resistance. Zeng et al [13] reported that Sn decreased the cathodic reaction and the hydrogen evolution rate of alloys due to its high hydrogen overpotential. Moreover, Polina et al [14] indicated that Sn dissolved in the α-Mg matrix of AT53 and AT54 alloys and facilitated the formation of a protective layer that reduced the anodic dissolution rate of the matrix [15]. However, excessive Sn addition to Mg alloys can result in the formation of Mg 2 Sn intermetallic compounds, which increase the hydrogen evolution rate and deteriorate the corrosion resistance of as-cast AM70 magnesium alloys [16,17]. Mg 2 Sn intermetallic compounds in magnesium alloys were assumed to be the cathode phase responsible for severe microgalvanic corrosion [18,19]. However, there remains an incomplete understanding of how Mg 2 Sn phases influence the microgalvanic corrosion of Mg-Al alloys.
Recently, scanning Kelvin probe force microscopy (SKPFM) has been used to determine the Volta potential difference between the α-Mg matrix and secondary phases, such as Al 2 Y, Al 2 Nd, Al 2 Gd, and Al-Mn [20][21][22][23]. This has increased the understanding of the local microgalvanic corrosion behavior of such alloys. Therefore, the Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. effect of Sn content on the mechanical properties and corrosion behavior of Mg-3Al-xSn alloys were investigated by SEM-EDXS, electrochemical measurements, and SKPFM. The main focus was on the effect of Mg 2 Sn on the mechanical properties of Mg-3Al-xSn alloys and their corrosion mechanism.
Experimental procedure
Mg-3Al-xSn (x=0, 1.0, 1.5, 2.0) alloys were prepared using pure Mg (>99.9 wt%), Al (>99.9 wt%), Sn (>99.9 wt%), and Al-10Mn master alloy ingots. First, pure Mg and Al were heated to 720°C in a steel crucible inside a resistance furnace, and the melt was protected by a mixture of CO 2 and SF 6 gases in a volume ratio of 6:1. When the temperature of the melt decreased to 680°C, Sn wrapped in aluminum foil was added. After holding for 15 min at 700°C, the melt was poured into a metal mold (180 mm in length and 40 mm in diameter) preheated to 200°C. Subsequently, these experimental alloys were processed into Φ38×40 mm by a lathe. Samples were first homogenized at 400°C for 12 h in a box-type resistance furnace and then hot extruded at 300°C into bars with 16 mm diameters at an extrusion speed of 2.5 mm s −1 . The chemical compositions of the specimens were analyzed by inductively coupled plasma atomic emission spectroscopy (ICP-AES), and the results are shown in table 1. The extruded bar was processed into a standard tensile specimen, and the remainder was used for microscopic observations and electrochemical tests. Tensile tests were carried out at a strain rate of 1 mm min −1 using an electronic tensile testing machine (CMT5105) at room temperature. To obtain microstructural characteristics, all samples were cut from as-cast and extruded bars, then polished by 600#-2000# silicon carbide (SiC) paper and 5-0.5 μm diamond polishing paste on an MP-2A metallographic polishing machine. Then, sample surfaces were corroded by a 4% nitric acid/ethanol solution. Microstructural characteristics were observed by optical microscopy (OM, Nican M200) and scanning electron microscopy (SEM, TESCAN VEGA 3) equipped with energy-dispersive x-ray spectroscopy (EDXS).
Surface potential maps of the phases in the Mg-3Al-xSn alloys were obtained by atomic force microscopy (AFM, Agilent Technologies 5500 Scanning Probe Microscope) working in the KFM-AM mode. A probe with a gold-plated tip was used to simultaneously acquire topographic and surface potential images. The images were scanned with a pixel resolution of 512×512 and a scan rate of 1 Hz. The tip force constant was 1 N m −1 , and all measurements were conducted at room temperature.
Specimens for corrosion tests, including potentiodynamic polarisation, H 2 gas collection, and mass loss evaluation, were ground with a 2000 grit SiC paper, cleaned in alcohol, and then dried in air. The weight of each sample was measured using an analytical balance with an accuracy of 0.0001 g. Corrosion experiments were performed by placing samples in a 3.5 wt% NaCl solution for 24 h at room temperature. H 2 was collected into an inverted acid dropper above the corroded samples. The volume of H 2 was obtained by measuring the change in the liquid level. Then, samples were soaked in a mixed solution of 20% CrO 3 and 1% AgNO 3 for 15 min to remove corrosion products. Finally, samples were dried in the air and weighed. All tests were repeated five times to assess their reproducibility. The hydrogen evolution and weight loss rates of samples were calculated by the following formulas [24-27]: where M 1 is the original weight, and M 2 is the weight after corrosion for 24 h; H 1 is the original liquid level of the acid dropper, and H 2 is the liquid level after 24 h; A is the total surface area of the samples; t is the corrosion time.
All electrochemical measurements were carried out on an electrochemical workstation (Princeton P4000). The three-electrode system was composed of a working electrode with an exposed area of 1 cm 2 , a platinum counter electrode, and a saturated calomel electrode as the reference electrode. Polarization scans of the potentiodynamic polarization tests were in the range of ±250 mV SCE with a scan rate of 1 mV SCE s −1 after samples were immersed for 15 min to stabilize the open circuit potential. Electrochemical impedance spectra were obtained at frequencies from 100 kHz to 10 MHz. The disturbance amplitude of the open circuit potential was 5 mV. The EIS results were fitted using ZView-Impedance software, and the fitting errors of the parameters were less than 10%. All tests were repeated five times to reduce errors. Figure 1 presents the SEM images of the Mg-3Al alloys with different Sn contents. Two different phases were observed, and the corresponding EDS results are listed in figure 2. As shown in figure 1(a), the microstructure of the matrix and Mg-3Al-1Sn alloys consisted of an α-Mg matrix and β-Mg 17 Al 12 and Al-Mn particles, according to the EDS results ( figure 2(a)). When 1.5 wt.% Sn was added into the alloy, some bright particles with sizes of ∼2 μm were formed ( figure 1(c)). The corresponding EDS results (figure 2(b)) indicated that these particles were an Sn-rich intermetallic compound. As the Sn content increased to 2.0 wt.%, the Sn-rich intermetallic compound particles coarsened, as shown in figures 1(d) and 2(c).
Microstructure
To further analyze the Sn distribution in the alloy, EDXS elemental maps of Mg-3Al-2Sn are shown in figure 3. Significant amounts of Sn were evenly distributed throughout the matrix alloy, but it tended to accumulate on the white particles. It can be inferred that a small amount of Sn was present in the matrix alloy in the form of a solid solution. When the content of Sn exceeded the solid solubility of the Mg matrix, an Sn-rich intermetallic compound formed, as shown in figure 3(d). The electronegativities of Mg, Al, Mn, and Sn are 1.31, 1.61, 1.55, and 1.96, respectively. The electronegativity difference between Mg and Sn was higher than that of the other two elements, which suggests that the Mg-Sn phase would form more easily than others. The Sn-rich intermetallic compounds in the as-cast Mg-3Al-2Sn alloy were analyzed by XRD, as shown in figure 4. Many peaks corresponding to α-Mg and β-Mg 17 Al 12 were observed in the patterns of the Mg-3Al-xSn alloys, while peaks associated with Mg 2 Sn were found in the patterns of the Mg-3Al-1.5Sn and Mg-3Al-2Sn alloys. Therefore, it was concluded that the Sn-rich intermetallic compound was Mg 2 Sn.
Optical microstructures of the hot-extruded Mg-3Al-xSn (x=0, 1.0, 1.5, 2.0) alloys are shown in figure 5. Many coarse α-Mg grains were observed in the Mg-3Al alloy after hot extrusion, and the average grain size of the alloy matrix was measured via the linear intercept method at 103 μm, as shown in figure 5(a). Upon increasing the Sn content, the grain size of the Mg-3Al alloys containing Sn decreased. At 1.5 wt.% Sn, the grain size decreased to 66 μm, as shown in figure 5(c), and the average grain size of the Mg-3Al-1Sn and Mg-3Al-2Sn alloys were 81 and 78 μm, respectively.
Recent studies have used the different Volta potentials between the secondary phases and the alloy matrix to understand local microgalvanic corrosion behavior [28]. In this study, the Volta potential of the intermetallic compounds in Mg-3Al and Mg-3Al-1.5Sn alloys were determined using SKPFM. Figures 6 and 7 show the topographic images, surface potential maps, and surface potential profiles of the α-Mg matrix, β-Mg 17 Al 12 , Al-Mn, and Mg 2 Sn phases in the Mg-3Al and Mg-3Al-2Sn alloys. As shown in figures 6 and 7, strip-shaped Al-Mn particles exhibited cathodic behavior with a potential 80 mV higher than the surrounding matrix alloy. In contrast, the potentials of the submicron Mg 2 Sn particles and β-Mg 17 Al 12 were 100 mV and 50 mV higher than the matrix, respectively. These values are similar to those observed for the Mg-Al alloys [29], which suggests that a more intense galvanic couple formed between Mg 2 Sn and the α-Mg phase in the Mg-3Al-2Sn alloy. Figure 8 shows the hydrogen evolution and weight loss rates of Mg-3Al-xSn alloys after immersion in a 3.5% NaCl solution for 24 h. The H 2 volume and weight loss rate first decreased with the Sn content and then increased significantly. Mg-3Al-1Sn exhibited low hydrogen evolution and weight loss rates of 0.18×10 −2 ml·cm −2 ·h −1 and 0.21×10 −2 mg·cm −2 ·h −1 , respectively. It also had a maximum corrosion rate at 2.0 wt.% Sn, i.e. an H 2 volume of 1.69×10 −2 ml·cm −2 ·h −1 and a weight loss rate of 2.07×10 −2 mg·cm −2 ·h −1 . The immersion test results indicated that the lowest corrosion rate and best corrosion resistance were obtained at 1.0 wt.% Sn.
Corrosion and electrochemical analysis
The polarization curves of Mg-3Al-xSn in 3.5 wt.% NaCl solution are displayed in figure 9. The corrosion potential (E corr ) of Mg-3Al-1Sn was higher than other alloys, and E corr of the matrix alloy was −1.41 V SCE . It first increased to −1.35 V SCE for Mg-3Al-1Sn, then decreased to −1.426 V SCE for Mg-3Al-1.5Sn, and finally decreased to −1.451 V SCE for Mg-3Al-2Sn. The fitting data of the polarization curves in table 2 show that the alloy prepared with 1.0 wt.% Sn exhibited the lowest corrosion current density (I corr ) of 34.56 μA·cm −2 . The Nyquist plots and Bode plots of Mg-3Al-xSn alloys immersed in 3.5 wt.% NaCl solution are shown in figure 10. The Nyquist plot of Mg-3Al consists of two capacitive loops and a short low-frequency inductive loop, while the other alloys all displayed a capacitive loop and an inductive loop. The capacitive loop diameters were ordered as follows: Mg-3Al-1Sn>Mg-3Al>Mg-3Al-1.5Sn>Mg-2Sn. It has been reported that a larger capacitive loop diameter corresponds to a higher corrosion resistance [30]; therefore, the addition of 1.0 wt.% Sn enhanced the anticorrosion performance of Mg-3Al. In figure 10(b), the value of |Z| for Mg-3Al-1Sn was also the highest, and most of the phase angles in figure 10(c) exceeded 50°.
The corresponding equivalent circuit of each alloy is shown in figure 11. Figure 11(a) represents Mg-3Al, and figure 11(b) represents Mg-3Al-xSn (x=1.0, 1.5, 2.0). In figures 11(a) and (b), R s is the solution resistance, R f and CPE f are the resistance and capacitance of the corrosion product layer, R ct is the resistance to transfer charge during corrosion, L and R L represent inductance resistance, and CPE dl is the capacitance of the electric doublelayer at the interface of the metal surface and the corrosive medium [31][32][33]. Table 3 lists the fitting parameters obtained from the EIS spectra of the studied alloys. The polarization resistance (R p ) can be calculated as follows [34]: The R p value of Mg-3Al-1Sn was the highest (227.54 Ω·cm 2 ), while Mg-3Al-2Sn had the lowest value (133.37 Ω·cm 2 ). Metalnikov et al [14] reported that a high R p value indicates a better corrosion resistance, indicating that Mg-3Al-1Sn has the highest corrosion resistance.
Mechanical properties
The stress-strain curves and tensile properties of the extruded Mg-3Al-xSn alloys at room temperature are shown in figure 12. The ultimate tensile strength of the Mg-3Al-xSn alloys increased with the Sn content, reaching a maximum of 271 MPa at 1.5 wt.% Sn; however, as the Sn content further increased, the ultimate tensile strength decreased to 250 MPa. The elongation of the extruded Mg-3Al-xSn alloys decreased from 13.8% to 9.4% as the Sn content increased, as shown in table 4.
Effect of Sn content on the corrosion resistance of alloys
Electrochemical potential differences between secondary-phase particles and the α-Mg matrix influence the corrosion behavior, which is usually quantitatively examined to evaluate the microgalvanic corrosion behavior of alloys [20,35]. In this study, the Volta potential of the Al-Mn phase showed a higher corrosion potential (30 and 80 mV) than that of the Mg 17 Al 12 and α-Mg phases, indicating that Al-Mn served as a local cathode in Mg-3Al. It was also revealed that Mg 2 Sn served as a local cathode during the microgalvanic corrosion of Mg-3Al-Sn alloys. Therefore, the effect of microgalvanic couple corrosion between Mg 2 Sn/Al-Mn phases and the matrix increased the corrosion rate and decreased the polarization resistance of Mg-3Al-2Sn alloys.
To further understand the corrosion process of Sn-containing Mg-Al alloys, a schematic of the corrosion mechanism is shown in figure 13. According to the microstructural characteristics and SKPFM results, welldefined anodic α-Mg and cathodic Al-Mn phases within the Mg-3Al alloy were confirmed ( figure 13(a)). When 1.0 wt.% Sn was added into the Mg-3Al alloy, most Sn dissolved in α-Mg (figures 3 and 13(b)), which decreased the weight loss rate of Mg-3Al-1Sn to 0.21×10 −2 mg·cm −2 ·h −1 . Similar results have been previously reported, in which Sn dissolved in the α-Mg matrix increased the corrosion resistance by stabilizing the oxide layer of corrosion products and reducing the anodic dissolution rate of the matrix [15][16][17]19]. As the Sn content further increased, Mg 2 Sn formed in Mg-3Al-1.5Sn, which served as a local cathode due to its high potential and accelerated microgalvanic corrosion along with the secondary local cathode, Al-Mn ( figure 13(c)). Significant amounts of cathodic Mg 2 Sn in Mg-3Al-2Sn further deteriorated its corrosion resistance, as shown in figure 13(d).
Effect of Sn content on the mechanical properties of alloys
The strengthening mechanism of Mg-3Al-xSn alloys may be attributed to grain refinement strengthening, solution strengthening, and secondary-phase strengthening. According to the Hall-Petch relationship, grain size has a significant effect on mechanical properties. In this work, the average grain size of the Mg-3Al-xSn alloys decreased from 103 to 66 μm during dynamic recrystallization due to hot extrusion. The lower average grain size increased the surface tension and the interaction of neighboring grains, which resulted in the formation of a hard deformation area near grain boundaries [36]. Therefore, the formation of fine grains increased the deformation resistance, which greatly improved the tensile properties of the extruded Mg-3Al-xSn alloys. Moreover, the atomic radius of Sn (R Sn =0.14 nm) is larger than Mg (R Mg =0.136 nm), which allowed Sn to dissolve and evenly distribute throughout the Mg matrix, as shown in figure 3. These solid solution Sn atoms caused lattice distortion, increased the dislocation motion resistance, remarkably impeded slip, and improved the strength of the magnesium alloy. For example, a Δσs increase of 29.8 MPa was caused by the dissolution of Sn in Mg [37]. As shown in figures 1(c) and 7, many fine Mg 2 Sn particles were uniformly distributed in the α-Mg grain in Figure 11. Equivalent circuit of the EIS spectra for the Mg-3Al (a) and Mg-3Al-xSn (x=1, 1.5, 2) alloys (b). Table 3. Electrochemical parameters of studied alloys attained from the EIS data. Mg-3Al-1.5Sn that facilitated the formation of pinning points and dislocations. During tensile deformation, an increasing number of pinning points and dislocations in the grain interior due to Mg 2 Sn hindered deformation and increased the deformation resistance and ultimate tensile strength. However, the coarser Mg 2 Sn particles in Mg-3Al-2.0Sn separated the matrix alloy under loading, which reduced the tensile strength. Meanwhile, the solution strengthening and secondary phase strengthening also weakened the plasticity of the alloys, as shown in figure 12 and table 2.
Conclusions
(1) Increasing the Sn content in Mg-3Al-xSn alloys first caused Sn to dissolve in the α-Mg matrix, but Mg 2 Sn particles precipitated when the Sn content exceeded 1.0 wt.%.
(2) As the Sn addition increased, the H 2 volume and weight loss rates first decreased, followed by a significant increase. Mg-3Al-1Sn had the best corrosion potential, and its larger polarization resistance (determined by EIS) indicated a better corrosion resistance.
(3) The Volta potentials of Mg 2 Sn particles, Al-Mn, and β-Mg 17 Al 12 phases were 100, 80, and 50 mV higher than the matrix, respectively. The results showed that the Mg 2 Sn phase served as a local cathode and accelerated microgalvanic corrosion, along with the secondary local cathode, Al-Mn.
(4) The addition of Sn improved the mechanical properties of the Mg-3Al-xSn alloys by solution strengthening and secondary phase strengthening. | 4,446.4 | 2020-07-08T00:00:00.000 | [
"Materials Science"
] |
Phosphorylation of Tyr-398 and Tyr-402 in Occludin Prevents Its Interaction with ZO-1 and Destabilizes Its Assembly at the Tight Junctions*
Occludin is phosphorylated on tyrosine residues during the oxidative stress-induced disruption of tight junction, and in vitro phosphorylation of occludin by c-Src attenuates its binding to ZO-1. In the present study mass spectrometric analyses of C-terminal domain of occludin identified Tyr-379 and Tyr-383 in chicken occludin as the phosphorylation sites, which are located in a highly conserved sequence of occludin, YETDYTT; Tyr-398 and Tyr-402 are the corresponding residues in human occludin. Deletion of YETDYTT motif abolished the c-Src-mediated phosphorylation of occludin and the regulation of ZO-1 binding. Y398A and Y402A mutations in human occludin also abolished the c-Src-mediated phosphorylation and regulation of ZO-1 binding. Y398D/Y402D mutation resulted in a dramatic reduction in ZO-1 binding even in the absence of c-Src. Similar to wild type occludin, its Y398A/Y402A mutant was localized at the plasma membrane and cell-cell contact sites in Rat-1 cells. However, Y398D/Y402D mutants of occludin failed to localize at the cell-cell contacts. Calcium-induced reassembly of Y398D/Y402D mutant occludin in Madin-Darby canine kidney cells was significantly delayed compared with that of wild type occludin or its T398A/T402A mutant. Furthermore, expression of Y398D/Y402D mutant of occludin sensitized MDCK cells for hydrogen peroxide-induced barrier disruption. This study reveals a unique motif in the occludin sequence that is involved in the regulation of ZO-1 binding by reversible phosphorylation of specific Tyr residues.
Epithelial tight junctions (TJs) 2 form a selective barrier to the diffusion of toxins, allergens, and pathogens from the external environment into the tissues in the gastrointestinal tract, lung, liver, and kidney (1). Disruption of TJs is associated with the gastrointestinal diseases such as inflammatory bowel disease, celiac disease, infectious enterocolitis, and colon cancer (2)(3)(4) as well as in diseases of lung and kidney (5,6). Numerous inflammatory mediators such as tumor necrosis factor ␣, interferon ␥, and oxidative stress (7)(8)(9)(10)(11)(12) are known to disrupt the epithelial TJs and the barrier function. Several studies have indicated that hydrogen peroxide disrupts the TJs in intestinal epithelium by a tyrosine kinasedependent mechanism (11,12).
Four types of integral proteins, occludin, claudins, junctional adhesion molecules, and tricellulin are associated with TJs. Occludin, claudins, and tricellulin are tetraspan proteins, and their extracellular domains interact with homotypic domains of the adjacent cells (1,2,13). The intracellular domains of these proteins interact with a variety of soluble proteins such as ZO-1, ZO-2, ZO-3, 7H6, cingulin, and symplekin (14 -23); this protein complex interacts with the perijunctional actomyosin ring. The interactions among TJ proteins are essential for the assembly and the maintenance of TJs. Therefore, regulation of the interactions among TJ proteins may regulate the TJ integrity. A significant body of evidence indicates that numerous signaling molecules are associated with the TJs. Protein kinases and protein phosphatases such as protein kinase C (PKC), PKC/ (24), c-Src (25), c-Yes (26,27), mitogenactivated protein kinase (28), PP2A, and PP1 (29) interact with TJs, indicating that TJs are dynamically regulated by intracellular signal transduction involving protein phosphorylation. Additionally, other signaling molecules such as calcium (30), phosphatidylinositol 3-kinase (31), Rho (32), and Rac (33) are involved in the regulation of TJs.
Occludin, a ϳ65-kDa protein, has been well characterized to be assembled into the TJs. Although occludin knock-out mice showed the formation of intact TJs in different epithelia (34), numerous studies have emphasized that it plays an important role in the regulation of TJ integrity. Occludin spans the membrane four times to form two extracellular loops and one intracellular loop, and the N-terminal and C-terminal domains hang into the intracellular compartment (35)(36)(37). In epithelium with intact TJs, occludin is highly phosphorylated on Ser and Thr residues (38), whereas Tyr phosphorylation is undetectable. However, the disruption of TJs in Caco-2 cell monolayers by oxidative stress and acetaldehyde leads to Tyr phosphorylation of occludin; the tyrosine kinase inhibitors attenuate the disruption of TJs (39,40). Furthermore, a previous in vitro study dem-onstrated that Tyr phosphorylation of the C-terminal domain of occludin leads to the loss of its interaction with ZO-1 and ZO-3 (25).
In the present study we identified the Tyr residues in occludin that are phosphorylated by c-Src and determined their role in regulated interaction between occludin and ZO-1 and its assembly into the TJs. Results show that 1) Tyr-379 and Tyr-383 in chicken occludin and Tyr-398 and Tyr-402 in human occludin are the exclusive sites of phosphorylation by c-Src, and these Tyr residues are located in a highly conserved sequence of occludin, YET-DYTT, 2) deletion of YEDTYTT or point mutation of Tyr-398 and Tyr-402 in human occludin attenuates the phosphorylation-dependent regulation of ZO-1 binding, 3) Y398D/Y402D mutation of human occludin leads to loss of ZO-1 binding and prevents its translocation to the plasma membrane and cell-cell contact sites in Rat-1 cells, 4) Y398D/Y402D mutation of occludin delays its assembly into the intercellular junctions during the calcium-induced assembly of TJs, and 5) expression of Y398D/Y402D mutant occludin sensitizes cell monolayers for hydrogen peroxide-induced disruption of barrier function.
Chemicals
Cell culture reagents and supplies, G418, Lipofectamine-R, and Plus reagent were purchased from Invitrogen. FuGENE was purchased from Roche Diagnostics, and glutathione (GSH), leupeptin, aprotinin, pepstatin A, phenylmethylsulfonyl fluoride, protease inhibitor mixture, GSH-agarose, Triton X-100, and vanadate were purchased from Sigma. The QuikChange XL site-directed mutagenesis kit was from Stratagene, La Jolla, CA. Active c-Src (recombinant protein) was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). All other chemicals were of analytical grade and were purchased either from Sigma or Fisher.
Plasmids and Recombinant Proteins
cDNA for the C-terminal tail of chicken occludin (amino acids 358 -504) was a kind gift from Dr. James Anderson (University of North Carolina, Chapel Hill, NC); this was used to amplify and insert into pGEX2T vector. The C-terminal tail of human occludin 378 -522 was amplified from a full-length human occludin in pEGFP and then shuttled into pGEX2T vector. Site directed mutations were induced in both chicken and human occludin (C-terminal domain as well as full-length occludin). The sequences of the primers used for this are provided in Table 1. The mutations were confirmed by sequencing. pGEX2T constructs containing wild type occludin C-terminal domain (GST-cOcl-C and GST-hOcl-C) were transformed into BL21DE3 cells, and recombinant proteins were purified. The full-length human occludin (wild type and mutants) in pEGFP vector was used for transfection into Rat-1 and MDCK cells.
Mass Spectrometric Analysis
Trypsin Digestion-GST-cOcl-C WT , tyrosine-phosphorylated by c-Src, was suspended in 10% acetonitrile in 10 mM ammonium bicarbonate, pH 8.5, and incubated at 37°C overnight with TPCKtreated trypsin (enzyme to substrate molar ratio of 1:10). The digest was passed through 0.45-m filters. Clear supernatant was lyophilized to dryness. Air-dried samples were equilibrated in an aqueous solution containing 0.1% trifluoroacetic acid and desalted by passing through C-18 ZipTip (Millipore, Bedford, MA) using the manufacturer's protocol. Peptides extracted from the ZipTip were subjected to phosphopeptide extraction.
Extraction of Phosphopeptides-The phosphopeptides from trypsin digestion of Tyr-phosphorylated GST-cOcl-C were isolated using a phosphopeptide isolation kit (Pierce). Phosphopeptides were bound to immobilized gallium matrix at acidic pH (Ͻ3.5) and eluted in 50 mM ammonium bicarbonate at pH 10. Phosphopeptide extracts were then subjected to MALDI and LC/MS/MS analysis.
MALDI-TOF-Phosphopeptide extracts were dried under vacuum and reconstituted in 2 l of matrix (␣-cyano-4-hydroxycinnamic acid) and spotted for crystallization. Crystals were analyzed for mass by MALDI-TOF using Voyager Biospectrometry work station DE (delayed extraction technology) (Perseptive Biosystems Inc., Framingham, MA) and Data Explorer (Perseptive Biosystems). A Prescriptive Biosystems MALDI time-of-flight instrument incorporating a nitrogen laser (Laser Science, Newton, MA) was used to obtain MALDI mass spectra. Samples solubilized in 85% acetic acid and mixed (1:3 v/v) with ␣-cyano-4-hydroxycinnamic acid matrix were spotted in 1-l aliquots and air-dried. Typically, 100 -250 laser shots were used to obtain one mass spectrum. Mass scale was calibrated with peptide internal standards.
LC/MS/MS Analysis-Sequence analysis of tryptic peptides was performed by injecting 3 l of the ZipTip-purified sample onto a capillary C-18 LC column on-line with a Finnigan LCQ DECA (Thermoquest, San Jose, CA) ion-trap mass analyzer that is equipped with a nanoelectrospray ionization source. The capillary C-18 column was prepared in-house using New Objective Pico Frit (360-m outer diameter, 75-m inner diameter, 15-m tip, 10.4-cm length) and Magic C18AQ packing material (5-m beads, 200 A°pores). The peptides were fractionated using 0.1% formic acid in water as solvent A and 90% acetonitrile as solvent B. The acquired spectra were visualized using Qual-browser in the X-Calibur software suite. Raw
Sequences of primers used to generate various mutations
Sequences in bold substitute tyrosine residues in the wild type occludin.
Tight Junction Regulation by Tyr Phosphorylation of Occludin
data thus obtained was analyzed against a protein data base generated from Swissprot using the Sequest software suite (Sequest Technologies Inc., Lisle, IL).
Cell Culture and Transfection
Caco-2, Rat-1, and MDCK cells were cultured in Dulbecco's modified Eagle's medium from Invitrogen and supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, and 2 mM glutamine as per the ATCC guidelines. MDCK cells were seeded on 6-well plates a day before transfection to achieve 50 -60% confluency. The cells were transfected using 1 ml of antibiotic-free Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 1 g of DNA plasmid (empty vector pEGFP or vector carrying hOcl WT or its mutants), 1 l of Plus reagent, and 3 l of Lipofectamine-R for each well. After 20 h, the cell monolayers were trypsinized and seeded onto 100-mm plates. The cells were subjected to G418 selection (0.7 mg/ml) for 2 weeks. Resistant cells were sorted to obtain only GFPexpressing cells by fluorescence-activated cell sorter. Cells were maintained in the medium that was supplemented with 0.3 mg/ml G418. Rat-1 cells were transfected using FuGENE reagent as per the manufacturer's protocol, and the GFP-positive cells were sorted by fluorescence-activated cell sorter. Stably transfected cells were selected using G418 as described above.
Immunofluorescence Microscopy
Cell monolayers (12-mm transwells) were washed with phosphate-buffered saline and fixed in acetone-methanol (1:1) at 0°C for 5 min. Cell monolayers were blocked in 3% nonfat milk in TBST (20 mM Tris, pH 8.0, containing 150 mM NaCl and 0.5% Tween 20) and incubated for 1 h with primary antibodies (rabbit polyclonal anti-ZO-1 and mouse monoclonal anti-GFP) followed by incubation for 1 h with secondary antibodies (Cy3conjugated anti-rabbit IgG and AlexaFluor 488-conjugated anti-mouse IgG). The fluorescence was visualized using a Zeiss LSM 5 laser scanning confocal microscope, and images from Z-series sections (1 m) were collected by using Zeiss LSM 5 Pascal Confocal Microscopy Software (Release 3.2). Images were stacked using the software, Image J (NIH), and processed by Adobe Photoshop (Adobe Systems Inc., San Jose, CA).
Occludin Phosphorylation in Vitro
Recombinant GST-cOcl-C WT or GST-hOcl-C WT (5 g) was incubated with 500 ng of active c-Src in 250 l of kinase buffer (50 mM Hepes, pH 7.4, 1 mM EDTA, 0.2% -mercaptoethanol, 3 mM MgCl 2 ) containing 100 M ATP at 30°C for 3 h on a shaking incubator. Control reactions were done in the absence of ATP.
GST Pulldown Assay
To determine the interaction of occludin with ZO-1 and ZO-3, GST-hOcl-C GST-cOcl-C (2.5-10 g) was incubated with Caco-2 whole cell extract made in phosphate-buffered saline containing 0.2% Triton X-100, 1 mM sodium vanadate, and 10 mM sodium fluoride for 16 h at 4°C on an inverter. GST-occludin-C (GST-conjugated C-terminal tail of occludin) was pulled down with 20 l of 50% GSH-agarose slurry at 4°C for 1 h. The amounts of ZO-1 and ZO-3 bound to GSH-agarose were determined by immunoblot analysis. Nonspecific binding was determined by carrying out the binding with GST.
Immunoblot and Densitometric Analysis
Proteins were separated by 7% SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. Membranes were blotted for ZO-1, ZO-3, and p-Tyr by using specific antibodies in combination with HRPconjugated anti-mouse IgG or HRP-conjugated anti-rabbit IgG antibodies. HRP-conjugated anti-GST antibody was used for immunoblot analysis of GST or GST-occludin. The blot was developed using ECL chemiluminescence method (Amersham Biosciences). Quantitation was performed by densitometric analysis of specific bands on immunoblots by using the software, Image J.
Hydrogen Peroxide Treatment and Paracellular Permeability
MDCK cell monolayers that stably express GFP-hOcl WT , GFP-hOcl Y398A/Y402A , or GFP-hOcl Y398D/Y402D were exposed to varying concentrations (20 -2500 M) of hydrogen peroxide for 2 h, and paracellular permeability was evaluated by measuring the unidirectional flux of inulin as described before (11).
TJ Assembly by Calcium Switch
MDCK cell monolayers that stably express GFP-hOcl WT , GFP-hOcl Y398A/Y402A , or GFP-hOcl Y398D/Y402D were incubated overnight with low calcium medium followed by calcium replacement as described before (29). TJ assembly was evaluated by measuring transepithelial electrical resistance, inulin permeability, and confocal microscopy.
Immunoprecipitation
GFP was immunoprecipitated from cells under native or denatured conditions as described before (29). Anti-GFP immunocomplexes at native conditions were immunoblotted for ZO-1, whereas complexes under denatured conditions were immunoblotted for p-Tyr.
Statistics
Comparison between two groups was made by Student's t tests for grouped data. Significance in all tests was set at 95% or greater confidence level.
Tyr-379 and Tyr-383 in Chicken Occludin Are the Sites of
Phosphorylation by c-Src-A previous study showed that Tyr phosphorylation of occludin C-terminal domain by c-Src resulted in the loss of its interaction with ZO-1 (25). In the present study we identified the phosphorylation sites in occlu-
Tight Junction Regulation by Tyr Phosphorylation of Occludin
din C-terminal domain by mass spectrometric analysis. GST-fused C-terminal region (150 amino acids) of chicken occludin (GST-cOcl) was prepared and phosphorylated by incubation with c-Src and ATP. Tyr-phosphorylated GST-cOcl-C was digested with trypsin. Generation of five different tryptic peptides containing Tyr residues was predicted (Fig. 1A); the mass of these peptides was expected to increase by 80 Da with phosphorylation of each Tyr residue. MALDI mass spectrometric analysis of the phosphopeptide extracts from the tryptic digest detected several phosphopeptides with masses slightly deviated from the predicted mass analysis (Fig. 1B). Fig. 1D. Three different Tyr-phosphorylated peptides were identified. All three peptides were identified as the derivatives of tryptic peptide, P1 with single or double phosphorylation of Tyr residues. These results determine that two Tyr residues in occludin C-terminal region corresponding to the sequence of P1 were singly or doubly phosphorylated. These two Tyr residues correspond to Tyr-379 and Tyr-383 in chicken occludin.
Sequence alignment of occludin from different species (Fig. 2A) demonstrated that Tyr-379 and Tyr-383 are located in a highly conserved sequence of occludin (YETDYTT) and that Tyr-398 and Tyr-402 are the corresponding Tyr residues in human occludin. Therefore, we induced mutations in chicken and human occludin C-terminal region (cOcl-C and hOcl-C). The YETDYTT was deleted or the tyrosine residues in this region were subjected to point mutation in cOcl-C and hOcl-C (Fig. 2B) and inserted into pGEX2T vector to generate GST-fused mutant proteins. Tyr-398 and Tyr-402 in full-length human occludin in pEGFP vector were mutated to phenylalanine or aspartic acid and expressed in Rat-1 or MDCK cells as GFP fusion proteins.
Y379F and Y383F Mutation of Chicken Occludin Attenuates Its Phosphorylation and Regulation of ZO-1 Binding-GST pulldown assay for ZO-1 binding showed that GST-cOcl-C WT , GST-cOcl-C Y379F , GST-cOcl-C Y383F , and GST-cOcl-C Y379F/Y383F bind ZO-1 and ZO-3 in a dose-dependent manner (Fig. 4A). Incubation with c-Src in the presence of ATP showed a partial phosphorylation of single mutants, GST-cOcl-C Y379F and GST-cOcl-C Y383F , whereas phosphorylation was undetectable in the double mutant, GST-cOcl-C Y379F/Y383F (Fig. 4B). ZO-1 binding was not significantly different among unphosphorylated occludin (Fig. 4, B and C), except that ZO-1 binding of GST-cOcl-C Y379F was slightly greater than that of GST-cOcl-C WT . Incubation in the presence of ATP and c-Src resulted in a reduced ZO-1 binding by GST-cOcl-C WT and GST-cOcl-C Y379F (Fig. 4, B and C). However, incubation with c-Src in the presence of ATP did not alter the ZO-1 binding by GST-cOcl-C Y383F and GST-cOcl-C Y379F/Y383F (Fig. 4, B and C).
Mutation of Tyr-398 and Tyr-402 in Human Occludin Prevents Phosphorylation and Alters ZO-1 Binding-The sequence analysis indicated that Tyr-398 and Tyr-402 are the residues in human occludin that correspond to Tyr-379 and Tyr-383 in chicken occludin. Therefore, we mutated Tyr-398 and Tyr-402 in hOcl-C. Similar to GST-cOcl-C WT , incubation with c-Src in the presence of ATP induced Tyr phosphorylation of GST-hOcl-C WT , whereas phosphorylation was undetectable in GST-hOcl-C Y398F/Y402F and GST-hOcl-C Y398D/Y402D mutants (Fig. 6A). GST pulldown assay showed that GST-hOcl-C WT binds to ZO-1, and this binding was reduced by incubation with c-Src in the presence of ATP. GST-hOcl-C Y398D/Y402D showed only a trace amount of ZO-1 binding. ZO-1 binding to GST-hOcl-C Y398F/Y402F was lower than GST-hOcl-C WT ; however, the ZO-1 binding was not further reduced by incubation with c-Src in the presence of ATP (Fig. 6, A and C). Unlike reduced binding to GST-hOcl-C Y398F/Y402F , the ZO-1 binding to GST-hOcl-C Y398A/Y402A was similar to that of GST-hOcl-C WT (Fig. 6B). Once again, GST-hOcl-C Y398A/Y402A showed no Tyr phosphorylation or regulation of ZO-1 binding when incubated with c-Src in the presence of ATP (Fig. 6B). Densitometric analysis (Fig. 6C) confirmed that ZO-1 binding to GST-hOcl-C Y398A/Y402A and GST-hOcl-C Y398D/Y402D is significantly lower than that of GST-hOcl-C WT , whereas the binding to GST-hOcl-C Y398A/Y402A was similar to that of GST-hOcl-C WT . Furthermore, incubation with c-Src in the presence of ATP significantly reduced ZO-1 binding to GST-hOcl-C WT but not to GST-hOcl-C Y398F/Y402F , GST-hOcl-C Y398A/Y402A , or GST-hOcl-C Y398D/Y402D when compared with corresponding ZO-1 binding in the absence of ATP.
Y398D and Y402D Mutation in Human Occludin Prevents Its Localization at Plasma Membrane and Cell-Cell Contact
Sites in Rat-1 Cells-The regulation of ZO-1 binding by Tyr-398 and Tyr-402 raised the question of whether phosphorylation of Tyr-398 and/or Tyr-402 of occludin affects its localization at the TJs. To determine the effect of mutation of Tyr-398 and Tyr-402 on the distribution of occludin at the plasma membrane and the intercellular junctions, we transfected Rat-1 cells (occludin null) with GFP-hOcl, GFP-hOcl Y398A/Y402A , and GFP-hOcl Y398D/Y402D and visualized the cells by confocal microscopy. GFP-hOcl WT and GFP-hOcl Y398A/Y402A were localized to both the plasma membrane and the intracellular compartment (Fig. 7). A greater level of occludin was found at the cell-cell contact sites, which was associated with the redistribution of ZO-1 at the cell-cell contact and the plasma membrane. In contrast, GFP-hOcl Y398D/Y402D was localized exclusively at the intracellular compartment with no trace of distribution at the plasma membrane or cell-cell contact sites (Fig. 7).
Y398D/Y402D Mutation of Occludin Delays Its Assembly at the TJs and Sensitizes MDCK Cells for TJ Disruption by Hydrogen Peroxide-Unlike
Rat-1 cells, in MDCK cell monolayers, GFP-hOcl Y398D/Y402D appeared at the intercellular junctions. However, during the calcium switch-induced assembly of TJs, GFP-hOcl Y398D/Y402D localized predominantly at the intracellular compartment, whereas GFP-hOcl WT and GFP-hOcl Y398A/Y402A appeared at the intercellular junctions 1 h after calcium replacement (Fig. 8A). The inulin permeability in cell monolayers that express GFP-hOcl Y398D/Y402D was significantly greater than those in cell monolayers expressing GFP-hOcl WT or GFP-hOcl Y398A/Y402A (Fig. 8B). As reported before (11), hydrogen peroxide induced a dose-dependent increase in inulin permeability in MDCK cell monolayers that stably express GFP-hOcl WT (Fig. 9A). A hydrogen peroxide-induced increase in inulin permeability was significantly lower in cell monolayers that express GFP-hOcl Y398A/Y402A , whereas it was significantly higher in cells expressing GFP-hOcl Y398D/Y402D . Incubation of cell monolayers that express GFP-hOcl WT , GFP-hOcl Y398A/Y402A , and GFP-hOcl Y398D/Y402D with 500 M hydrogen peroxide for 1 h increased inulin permeability (% flux/h/cm 2 ) from 0.025 Ϯ 0.005 to 0.035 Ϯ 0.01, 0.02 Ϯ 0.01 to 0.025 Ϯ 0.006, and 0.07 Ϯ 0.007 to 0.25 Ϯ 0.03, respectively. These observations were confirmed by analyzing the junctional distribution of GFP and ZO-1 after hydrogen peroxide treatment. Hydrogen peroxide induced a slight redistribution of GFP in cells expressing GFP-hOcl WT , whereas the hydrogen peroxideinduced redistribution of GFP from the junctions was much more dramatic in cells that express GFP-hOcl Y398D/Y402D
FIGURE 3. Deletion mutation of occludin prevents phosphorylation and attenuates regulation of ZO-1 binding.
A, varying amounts of GST-cOcl-C WT or GST-cOcl-C (⌬378 -385) was analyzed for ZO-1 binding by GST pulldown assay using Caco-2 cell extract. GST pulldown was immunoblotted (IB) for ZO-1, ZO-3, and GST. Binding to GST (5 g) was performed as a control. The labels p44, p42, and p22 correspond to the molecular weight of GST-cOcl-C WT , GST-cOcl-C (⌬378 -385) , and GST, respectively. B, densitometric analysis of ZO-1 and ZO-3 binding to 5 g of wild type (WT) or mutant (⌬378 -385) occludin. Values are the mean Ϯ S.E. (n ϭ 3). Asterisks indicate the values that are significantly (p Ͻ 0.05) different from corresponding value for WT group. C, GST-cOcl-C and GST-cOcl-C (⌬378 -385) , 2.5 g, were incubated with c-Src in the absence or presence of ATP and analyzed for ZO-1 binding by GST pulldown assay. GST pulldown assays were immunoblotted for ZO-1, p-Tyr, and GST. D, densitometric analysis of ZO-1 bands from three different experiments described in panel B. In each experiment, ZO-1 band density for GST-cOcl-C WT incubated without ATP was designated to 100, and corresponding bands in other groups were normalized as a percent of that number. Values are the mean Ϯ S.E. (n ϭ 3). The asterisk indicates the value that is significantly (p Ͻ 0.05) different from corresponding value for ϪATP group. (Fig. 9B). ZO-1 distribution in hydrogen peroxide-treated cell monolayers paralleled the distribution of GFP.
DISCUSSION
A significant body of evidence suggests that Tyr phosphorylation of TJ proteins may play an important role in the regulation of epithelial TJs. Occludin is highly phosphorylated on Ser and Thr residues in the intact epithelium, and Tyr phosphorylation is undetectable (38). Previous studies, however, demonstrated that occludin undergoes Tyr phosphorylation during the disruption of TJs by hydrogen peroxide (11,39). Furthermore, a recent in vitro study demonstrated that Tyr phosphorylation of the C-terminal region of occludin reduces its ability to interact with ZO-1, a TJ plaque protein (25). In the present study we identified the Tyr phosphorylation sites in the C-terminal region of occludin and demonstrated their role in regulation of ZO-1 binding.
The Tyr phosphorylation sites in the C-terminal region of chicken occludin were determined by mass spectrometric analysis of phosphooccludin. Mass analysis of phosphopeptide extracts from tryptic digests detected the presence of one phosphopeptide corresponding to the predicted peptide fragment of chicken occludin (amino acids 371-393) with phospho-Tyr residues. There are two Tyr residues within this sequence (Tyr-379 and 383), which were found to be phosphorylated as the mass of this peptide was 160 daltons greater than the predicted mass value. Another phosphopeptide detected by MALDI showed a molecular mass of 2915.7, which is 156 daltons greater than the predicted mass of the monophosphate peptide fragment. LC/MS/MS analysis determined that this peptide corresponds to the sequence 370 - A, GST-cOcl-C WT , GST-cOcl-C Y379F , GST-cOcl-C Y383F , and GST-cOcl-C Y379/383F were analyzed for ZO-1 binding by GST pulldown assay. GST pulldown assays were immunoblotted (IB) for ZO-1, ZO-3, and GST. B, GST-cOcl-C, GST-cOcl-C Y379F , GST-cOcl-C Y383F , and GST-cOcl-C Y379/383F were incubated with c-Src in the absence or presence of ATP. Five g of phosphorylated and non-phosphorylated occludins were analyzed for ZO-1 binding by GST pulldown assay. GST pulldown assays were immunoblotted for ZO-1, p-Tyr, and GST. Control binding was performed by using 5 g of GST. C, densitometric analysis of ZO-1 bands from three different experiments described in panel B. In each experiment, ZO-1 band density for GST-cOcl-C WT incubated without ATP was designated to 100, and corresponding bands in other groups were normalized as a percent of that number. Values are the mean Ϯ S. E. (n ϭ 3). Asterisks indicate the values that are significantly (p Ͻ 0.05) different from corresponding value for ϪATP group. 393 of chicken occludin with an extra Arg (Arg-370) at the N terminus compared with the predicted 371-393 fragment. This is possibly caused by a misdigestion by trypsin due to the presence of sequential Arg residues in this region of the occludin sequence. LC/MS/MS analysis also detected two types of 2915.2-dalton peptides; one in which Tyr-379 was phosphorylated and another in which Tyr-383 was phosphorylated. Therefore, the sequences of all three phosphopeptides identified in the study demonstrate that Tyr-379 and Tyr-383 are the phosphorylation sites in chicken occludin; Tyr-398 and Tyr-402 are the corresponding tyrosines in human occludin. These two tyrosines are located in a highly conserved sequence of occludin HYETDYTT. BLAST analysis of this sequence demonstrated that this is a unique motif that is not present in other proteins, including claudins.
Deletion of HYETDYTT (378 -385) from chicken occludin abrogated c-Src-induced Tyr-phosphorylation of the occludin C-terminal region, confirming the mass spectrometric data that Tyr-379 and Tyr-383 are the phosphorylation sites in chicken occludin. Point mutation of Tyr-379 or Tyr-383 to phenylalanine resulted in a partial decrease in c-Src-induced Tyrphosphorylation. Decrease in Tyr phosphorylation was greater in Y383F mutants compared with that in Y379F mutants, sug-FIGURE 5. Y379D and Y383D in chicken occludin attenuates its binding to ZO-1. A, GST-cOcl-C WT , GST-cOcl-C Y379D , GST-cOcl-C Y383D , and GST-cOcl-C Y379/383D were analyzed for ZO-1 binding by GST pulldown assay. GST pulldown assays were immunoblotted (IB) for ZO-1 and GST. B, GST-cOcl-C, GST-cOcl-C Y379D , GST-cOcl-C Y383D , and GST-cOcl-C Y379/383D were incubated with c-Src in the absence or presence of ATP. Ten g each of phosphorylated and non-phosphorylated occludins were analyzed for ZO-1 binding by GST pulldown assay. GST pulldown assays were immunoblotted for ZO-1, p-Tyr, and GST. C, densitometric analysis of ZO-1 bands from three different experiments described as in panel B and selected ZO-3 bands in panel A. In each experiment, ZO-1 (or ZO-3) band density for GST-cOcl-C incubated without ATP was designated to 100, and corresponding bands in other groups were normalized as a percent of that number. Values are the mean Ϯ S.E. (n ϭ 3). Asterisks indicate the values that are significantly (p Ͻ 0.05) different from corresponding value for ϪATP group, and the symbols # indicate the values that are significantly different (Ͻ0.05) from corresponding value for WT group. FIGURE 6. Mutation of Tyr-398 and Tyr-402 in human occludin attenuates its phosphorylation and altered regulation of ZO-1 binding. A, GST-hOcl-C WT , GST-hOcl-C Y398F/Y402F , and GST-hOcl-C Y398D/Y402D were incubated with c-Src in the absence or presence of ATP. Five g of phosphorylated or nonphosphorylated occludin was analyzed for ZO-1 binding by GST pulldown assay. GST pulldown assays were immunoblotted (IB) for ZO-1, p-Tyr, and GST. B, five g of phosphorylated or non-phosphorylated GST-hOcl-C and GST-hOcl-C Y398A/Y402A were incubated with c-Src in the absence or presence of ATP and analyzed for ZO-1 binding by GST pulldown assay. GST pulldown assays were immunoblotted for ZO-1, p-Tyr, and GST. C, densitometric analysis of ZO-1 bands from three different experiments described in panel B. In each experiment, ZO-1 band density for GST-cOcl-C incubated without ATP was designated to 100, and corresponding bands in other groups were normalized as a percent of that number. Values are the mean Ϯ S.E. (n ϭ 3). The asterisk indicates the value that is significantly (p Ͻ 0.05) different from corresponding value for -ATP group. # indicates the values that are significantly (p Ͻ 0.05) different from value for non-phosphorylated wild type occludin.
gesting that Tyr-383 is a preferred phosphorylation site. Double mutation of Tyr-379 and Tyr-383 abolished c-Src-induced Tyr phosphorylation. Similarly, mutation of Tyr-398 and Tyr-402 in the C-terminal region of human occludin also abrogated the c-Src-induced Tyr phosphorylation. Previous studies showed that c-Src plays an important role in hydrogen peroxide-induced disruption of TJs and barrier dysfunction in Caco-2 cell monolayers (11). Expression of inactive c-Src significantly reduced hydrogen peroxide-induced Tyr-phosphorylation of occludin and TJ disruption, suggesting that c-Src-induced Tyrphosphorylation may be involved in this process. The present study suggests that hydrogen peroxide may induce phosphorylation of Tyr-379 and Tyr-383 in Caco-2 cells. A previous study demonstrated that Tyr phosphorylation of the C-terminal region of occludin results in the loss of its interaction with ZO-1 (25). ZO-1, a major adaptor protein of TJs, interacts with C-terminal region of occludin on one hand and with actin cytoskeleton on the other (12,14).
The interaction between the C-terminal region of occludin and ZO-1 is crucial for the assembly and the maintenance of occludin at the TJs (14). Truncation of the C-terminal region of occludin resulted in a loss of its interaction with ZO-1 and prevented its assembly into TJs. In the present study we determined the role of phosphorylation of specific Tyr residues in the C-terminal region of occludin in the regulation of its interaction with ZO-1. GST pulldown assays demonstrated that C-terminal regions of both chicken and human occludin bind to ZO-1. Deletion of the sequence HYET-DYTT in chicken occludin resulted in a significant reduction in binding to ZO-1 at higher concentrations of occludin; however, at low concentrations the deletion mutant bound to ZO-1 at a similar level to that of the wild type occludin. When wild type occludin C-terminal domain was incubated with c-Src and ATP, there was a significant reduction in ZO-1 binding; however, this was not observed with the deletion mutant, indicating that the phosphorylation of Tyr-379 and Tyr-383 is important in the regulation of interaction between occludin and ZO-1. This was confirmed by point mutations of Tyr-379 and -383. Y379F mutation partially reduced c-Src-induced regulation of ZO-1 binding, whereas Y383F or Y379F/Y383F mutation completely attenuated c-Src-induced regulation of ZO-1 binding. Similarly, Y398F/Y402F mutation in human occludin attenuated c-Src-induced regulation of ZO-1 binding. However, Y398F/Y402F mutation by itself resulted in significant reduction in ZO-1 binding. On the other hand, Y398A/Y402A mutation did not affect ZO-1 binding in the absence or presence of c-Src. Therefore, the results of this study demonstrate that Tyr-398 and Tyr-402 are important in regulation of ZO-1 binding by human occludin.
The crystal structure of the occludin C-terminal region (383-522) has been determined recently (15). This coiled-coil region C-terminal to the Tyr phosphorylation sites does bind to ZO-1 quite well. This indicates that the main function of the phosphorylation is not ZO-1 binding; rather, it plays a role in the regulation of ZO-1 binding. Similar to ZO-1 binding, ZO-3 binding to GST-hOcl-C was also altered by mutation of Tyr-398 and -402. Interestingly, ZO-3 binding to Y398F mutant was greater than its binding to GST-hOcl-C WT . However, at present, the reason for this enhanced binding is not clear and needs further studies.
Tight Junction Regulation by Tyr Phosphorylation of Occludin
To determine the effect of Tyr-398 or Tyr-402 phosphorylation on the assembly of occludin into intercellular junctions, we induced point mutations to GFP-tagged full-length human occludin. The GFP-hOcl WT and its mutants, GFP-hOcl Y398/Y402A , GFP-hOcl Y398/Y402D , and the corresponding single mutants were transfected into Rat-1 fibroblasts or MDCK cells. Confocal immunofluorescence microscopy demonstrated that the GFP-hOcl WT and its Y398/Y402A mutant were localized at the plasma membrane of Rat-1 cells, forming intercellular contact sites. Rat-1 cells express high levels of ZO-1, but it is predominantly localized in the intracellular compartment. However, transfection of GFP-hOcl WT or GFP-hOcl Y398/Y402A induced a recruitment of ZO-1 at the plasma membrane and the intercellular contact sites. Y398A/Y402A mutation did not alter the distribution of occludin at the plasma membranes. However, Y398D, Y402D, and Y398/Y402D mutants of occludin failed to localize at the plasma membrane or cell-cell contact sites; rather, they were distributed in the intracellular compartment. This indicates that mimicking the phosphorylation of Tyr-398 and Tyr-402 by mutation to aspartic acid results in the loss of its ability to assemble at the plasma membranes and cell-cell contact sites. This may be due to the loss of its ability to bind to ZO-1.
A significant portion of GFP-occludin in vesicular structure did appear near the plasma membrane, suggesting that the mutation did not result in a defect in its ability to integrate into the plasma membrane; rather, these occludin mutants are internalized into the cells due to the lack of their interaction with ZO-1 and inability to integrate into TJ structure. Both single mutants, Y398D and Y402D, similarly failed to localize at the intercellular junctions, indicating that phosphorylation of either Tyr residues is enough to alter its ability to bind ZO-1 and integrate into TJs. In vitro phosphorylation and ZO-1 binding studies indicated that Tyr-383 is more important than Tyr-379 in chicken occludin. However, this is in contrast to cell data, which shows that both Tyr-398 and Tyr-402 in human occludin are important for its localization at the cellcell contact sites. This may be explained by the lower preference of c-Src for Tyr-379 compared with Tyr-383. In the cell, Tyr-379 may be phosphorylated by some other tyrosine kinase.
In MDCK cells, wild type and Y398A/Y402A mutant occludin were localized at the intercellular junctions. However, during the early stages of TJ assembly by calcium replacement, we saw a delay in the organization of GFP-hOcl Y398D/Y402D at the intercellular junctions. One hour after calcium replacement, Y398D/Y402D mutant occludin was distributed predominantly in the intracellular compartment, whereas WT and Y398A/ Y402A mutant occludin were organized at the intercellular junctions. This confirmed that phosphorylation of Tyr-398 and Tyr-402 of occludin does prevent its ability to integrate into the TJs. The organization of Y398D/Y402D mutant at the intercellular junctions may be mediated by dimerization of mutant with the endogenous occludin. Expression of Y398D/Y402D mutant of occludin did disrupt the junctional distribution of ZO-1. This dominant negative effect is evident only at higher levels of mutant expression. The expression of mutant in relation to endogenous occludin is difficult to assess. However, ZO-1 redistribution was seen only in cells with higher level of mutant expression. Therefore, it is not clear whether such an effect can be seen at endogenous levels of phospho-occludin.
Furthermore, MDCK cell monolayers that express Y398D/ Y402D mutant of occludin were dramatically more sensitive to hydrogen peroxide-induced disruption of barrier function, whereas cell monolayers expressing Y398A/Y402A mutant occludin showed significant resistance to hydrogen peroxide compared with cell monolayers expressing wild type occludin. The present study also shows that hydrogen peroxide failed to induce Tyr phosphorylation in MDCK cell monolayers, demonstrating that Tyr-398 and Tyr-402 are the phosphorylation sites in hydrogen peroxide-treated cells. The loss of co-immunoprecipitation of ZO-1 with GFP-hOcl Y398D/Y402D in MDCK cells confirms our observation made in in vitro studies that FIGURE 9. Y398D/Y402D mutation of occludin sensitizes MDCK cell monolayers for hydrogen peroxideinduced disruption of TJs. A and B, MDCK cell monolayers that express GFP-hOcl WT , GFP-hOcl Y398A/Y402A , or GFP-hOcl Y398D/Y402D were exposed to varying concentrations of hydrogen peroxide. Inulin permeability was measured 1 h after hydrogen peroxide (A), and fixed cell monolayers were double-stained for GFP and ZO-1 by the immunofluorescence method (B). C, MDCK cell monolayers that express GFP-hOcl WT , GFP-hOcl Y398A/Y402A , or GFP-hOcl Y398D/Y402D were incubated with or without (control) hydrogen peroxide for 1 h. Anti-GFP immunocomplexes prepared under denatured conditions were immunoblotted (IB) for p-Tyr and GFP. The arrow with p92 label corresponds to GFP-occludin. IP, immunoprecipitates. D, anti-GFP immunocomplexes prepared under non-denaturing conditions from MDCK cells expressing wild type or mutant occludins were immunoblotted for ZO-1 and GFP. Density of ZO-1 bands from three different experiments was measured. The arrow with p92 label corresponds to GFP-occludin. phosphorylation of Tyr-398 and Tyr-402 does prevent its interaction with ZO-1.
As shown in Fig. 9, the GFP-hOcl Y398D/Y402D eventually tends to organize at the intercellular junctions, although not as discretely as GFP-hOcl WT or GFP-hOcl Y398A/Y402A . This is possibly due to oligomerization of GFP-hOcl Y398D/Y402D with the endogenous occludin. The inulin flux in hydrogen peroxide-treated cell monolayers that express GFP-hOcl Y398D/Y402D is significantly higher than that in cells that express GFP-hOcl WT or GFP-hOcl Y398A/Y402A . The mechanism for greater sensitivity to hydrogen peroxide or delayed assembly is not clear at this point. However, we speculate that mixed oligomerization of GFP-Ocl Y398D/Y402D and endogenous occludin facilitates the dissociation of ZO-1 by hydrogen peroxide due to weak interaction between GFP-Ocl Y398D/Y402D with ZO-1; this may work synergistically with the hydrogen peroxide-induced Tyr phosphorylation of endogenous occludin and loss of its interaction with ZO-1. In summary, this study identifies Tyr-398 and Tyr-402 as the phosphorylation sites in human occludin and demonstrates that phosphorylation of these Tyr residues results in the loss of interaction between occludin and ZO-1 and attenuation of its integration into the epithelial TJs. | 8,467.2 | 2009-01-16T00:00:00.000 | [
"Biology",
"Chemistry"
] |
Research on slicing chip forming and wear analysis of arc-tooth slice tool with equal-rake angle
: As a new technology of gear processing, the gear slicing technology has distinct characteristics and broad application prospects. However, the lack in wear and life of the slice tool hinders the further widespread application of the gear slicing technology. Considering the chip forming mechanism is an important basis affecting the wear and service life of the cutter, this paper takes the arc-tooth slice tool with equal-rake angle (ATST-ERA) as the carrier and studies the chip forming and tool wear. Aiming at the particularity of the slicing process, the cutting process is unified and abstracted, and the cutting edge model, the cutting edge race model and the workpiece tooth surface model of the ATST-ERA are established, and then the undeformed chip geometry model is obtained. The correctness of the parametric model of the undeformed chip was verified by comparative analysis. Based on this model, the load of the cutting edge is calculated using the Kienzle empirical formula of the cutting force. By analyzing the relationship with the wear of the rake face, the positive correlation between the cutting edge load and the rake face wear of the ATST-ERA is found. The chip forming modeling method proposed in this paper provides a theoretical basis for the control of the slicing chip forming and the improvement of tool wear.
As a totally new gear machining technology, the gear slicing technology is not realized until the 21st century. The gear slicing technology can not only complete the machining of special structural gears such as thin-walled and non-penetrating internal helical teeth, but also has the characteristics of high efficiency, high precision and environmentally friendly. Therefore, many scholars have studied the gear slicing technology [1][2][3][4][5][6][7][8][9][10][11][12] . Though scholars have achieved breakthrough, the gear slicing technology has not been widely used. The reason is mainly because the slice tool is easy to wear and has a short service life. According to the metal cutting principle, the chip forming law is an important factor affecting tool wear. Because of this, in order to improve the wear performance of the slice tool and improve its service life, it is necessary to conduct an in-depth study on the chip forming mechanism during the slicing process.
Schulze et al. analyzed the chip forming mechanism and its effect on machining reliability by modeling the motion process of the gear skiving [13] . Mazak et al. used the simulation method to study the machining process of the intermittent cutting of the bevel gear. It was found that the peak point of the chip thickness could not be explained from the physical and kinematics, by analyzing the chip shape. Therefore, the method of calculating the chip thickness was proposed to describe the thickness of each part of the chip [14] . Tapoglou established a chip simulation model for power skiving in the CAD simulation environment. Based on this model, the geometry and size of nondeformed chips can be obtained [15] . Pierce et al. used a solid modeling software to create a virtual model of the power skiving process, extracting the chip geometry [16].
The above research is mainly based on simulation techniques, and the relevant research results have not been experimentally verified. In order to carry out in-depth research on the chip forming mechanism during slicing process, we can learn from the related research of chip forming of other machining process, such as turning, milling and hobbing.
Uysa, Su and Zang used the cutting process of stainless steel, graphite cast iron and titanium alloy as the carrier, and studied the chip forming mechanism and threedimensional shape of the chip by means of finite element simulation and cutting experiments. Then they analyzed the effects of cutting speed, feed rate and cutting temperature on the chip shape, and calculated the chip thickness, chip curl radius, shear zone thickness, etc. according to the chip model [17][18][19] . Ahmed et al. considered the influence of tool wear and established the chip model of AISI 304 turning in high pressure cooling environment. It was verified that the model prediction value of chip curl radius was basically consistent with the measured value [20] . Rakesh et al. studied the influence of cutting speed on chip shape by using dry-cutting chrome-nickel alloy as a carrier, and verified it by experiments [21] . Molnar et al. used orthogonal cutting as the research object. By comparing non-stationary cutting and stationary cutting experiments, the influence of cutting thickness on cutting force was analyzed, and the theoretical model related to chip breaking was evaluated [22] . [23] . Nikawa et al. used CAM simulation software to control the chip thickness by modifying the cutting parameters of the ball-end milling cutter in the NC code, and finally improved the machining ability of the machining method [24] .
Harry et al. proposed a theoretical modeling method for the machining process of turning and milling to predict the geometry of the uncut chip, and then calculated the cutting force, and verified the accuracy of the cutting force calculation result [25] .
Berenji et al. considered the turning and milling process under different machining conditions, studied the influence of machining process parameters on the chip forming mechanism. It was verified that good process parameters improved the machining quality [26] . Bergmann et al. considered the limitations of specific machining methods and specific workpiece materials, and studied the chip forming mechanism during the cutting process of circular cutting edges with different microscopic shapes, and verified them by planning and milling experiments [27] Chen et al. used the gear hobbing as the carrier to establish the space forming model of cylindrical gear of multi-edge interrupted cutting. The chip geometry was obtained by numerical method, and the relationship between the load distribution and wear of the cutting edge was further analyzed [28,29] . Likely, the gear hobbing process is the object. Li et al. used the finite element software to establish the simulation model of gear hobbing of high-speed and dry-cutting, and obtained the deformation law of the chip during the gear hobbing process and the wear state of the hob. The model was validated by the actual gear hobbing experiment [30] . Zhang used the solid modeling method to realize the material removal simulation of the gear hobbing process, and obtained the geometric entity of the undeformed chip. But the research result was not verified by experiments [31] .
As can be seen from the above literature, the motion relationship involved in the gear hobbing process is more complicated than in turning and milling process. In order to study the chips in the gear hobbing process, it is necessary to analyze and simplify the geometry of the gear hobbing process, and to study the chips of the gear hobbing process based on the orthogonal or bevel cutting principle. Similarly, this paper draws on the research method of chips in cutting process such as turning, milling and hobbing, analyzes the machining process of the ATST-ERA, and establishes the parametric model of the gear slicing process and the geometry model of the undeformed chip. The geometry model of the undeformed chip is verified by comparison with the chips obtained from the gear slicing experiment. On this basis, the cutting edge load is calculated based on the empirical formula. By comparing the load distribution of the cutting edge with the wear of the rake face, the positive correlation between the load distribution of the cutting edge and the wear of the rake face is obtained, which further verifies the correctness of the theoretical model established in this paper. The gear slicing technology can realize the machining of various gears such as internal teeth, external teeth, straight teeth and helical teeth. Regardless of the type of gear machining, the slicing process is based on the conjugate rotary cutting principle.
The gear slicing process is similar to the gear meshing process. The tool and the workpiece rotate in a certain speed ratio. The cutting edge that meets the conjugate condition of the line and surface continuously removes excess material during the continuous sweeping process, and finally forms the tooth surface of the workpiece.
According to the principle of conjugate rotary cutting, there are three kinds of motions involved in the gear slicing process, including the rotation of the workpiece, the rotation of the slice tool and the feed of the workpiece or slice tool along the axial direction. In addition, a shaft angle is always maintained between the tool axis and the workpiece axis during the movement of the tool and the workpiece. For the machining of different types of gears, the biggest difference lies in the relative position of the tool and the workpiece, and the cutting amount.
Looking at the tool axis as a reference, if the tool rotates counterclockwise with respect to the workpiece axis, it is considered to be "left pendulum"; if the tool rotates clockwise with respect to the workpiece axis It is considered to be "right pendulum".
When the workpiece rotates counterclockwise with respect to the workpiece axis, the angular velocity is positive, and vice versa; when the tool rotates counterclockwise with respect to the tool axis, the angular velocity is positive, and vice versa.
As shown in Figure 1, when slicing a spur and internal gear, if the cutter shaft swings to the left, the cutter rotates counterclockwise around the tool axis, and the workpiece rotates counterclockwise around the workpiece axis. If the cutter shaft swings to the right, the cutter rotates clockwise around the tool axis, and the workpiece rotates clockwise around the workpiece axis. As shown in Figure 2, when cutting a spur and external gear, if the cutter shaft swings to the left, the cutter rotates counterclockwise around the tool axis, and the workpiece rotates clockwise around the workpiece axis. If the cutter shaft swings to the right, the cutter rotates clockwise around the tool axis, and the workpiece rotates counterclockwise about the workpiece axis. Considering the shaft angle between the tool axis and the workpiece axis, the helix angle of the slice tool must be equal to shaft angle to ensure that the tooth direction of the cutter is consistent with the direction of the tooth groove of the workpiece, so that the tooth groove of the spur gear can be machined.
For the cutting of the spur gear, the difference between the machining of the internal gear workpiece and the external gear workpiece is the center distance value and the tool axis direction, as shown in Figure 3. According to the basic principle of the above-mentioned caries processing, the caries processing process of different types of gears is simplified into a unified motion abstract model, as shown in Figure 4. According to the position and motion relationship between the tool and the workpiece, the coordinate system of slicing is established, as shown in Figure 5. In the figure, the coordinate system S1(o1, x1, y1, z1) is the workpiece coordinate system, o1 represents the coordinate origin of the S1 coordinate system, the unit vectors of x1、y1、 z1 are i1、 j1、 k1, respectively; and the coordinate system S2(o2, x2, y2, z2) is the tool coordinate system, o2 is the coordinate origin of the S2 coordinate system, and the unit vectors of x2、y2、z2 are i2、 j2、 k2, respectively; the coordinate system Sp(op, xp, yp, zp) is The auxiliary coordinate system of the coordinate system S1, op represents the coordinate origin of the Sp coordinate system, the unit vectors of xp、yp、zp are ip、 jp、 kp, respectively; the coordinate system S0(o, x, y, z) is the auxiliary coordinate system of S2, o represents the coordinate origin of the S0 coordinate system, and the unit vectors of x、y、z are respectively i、 j、 k. The coordinate system S1 is used to establish the workpiece tooth surface model, and the coordinate system S2 is used to establish the tool model. The spatial positions of the auxiliary coordinate systems Sp and S0 are fixed and do not change with time. The center distance a represents the vertical distance between the tool axis and the workpiece axis, and the axis angle γ represents the angle between the tool axis and the workpiece axis. φ1 represents the angle of the workpiece coordinate system S1 as the workpiece rotates relative to the initial position, and l represents the distance that the workpiece coordinate system S1 moves relative to the auxiliary coordinate Sp along the workpiece axis (the positive direction of the zp axis), and φ2 represents the angle of the tool coordinate system S2 as the tool rotates relative to the initial position.
Figure 5 Coordinate system
According to the gear machining coordinate system shown in Figure 5, the total transformation matrix of the tool coordinate system S2 to the workpiece coordinate system S1 is: (1) In this formula 11 11 20 cos sin 0 0 sin cos 0 0 0 (4) In the above formula, T20 represents a transformation matrix of the tool coordinate system S2 to the auxiliary coordinate system S0, T0p represents a transformation matrix of the auxiliary coordinate system S0 to the auxiliary coordinate system Sp, and Tp1 represents a transformation matrix of the auxiliary coordinate system Sp to the workpiece coordinate system S1.
Model of edge-sweeping surface
According to the conjugate rotary cutting principle, the cutting edge and the theoretical tooth surface of the workpiece satisfy the conjugate relationship. Since many curves on the conjugate surface of the tooth surface of the workpiece satisfy the conjugate relationship with the tooth surface of the workpiece, a curve that satisfies the conjugate relationship with the tooth surface of the workpiece from tooth root to tooth crest is selected as the cutting edge of the slice tool. Firstly, the conjugate surface of the theoretical tooth surface of the workpiece is obtained, and then a plane or other regular curved surface is used to intersect with the conjugate surface, and the intersection line is the cutting edge curve. The method of deriving the cutting edge curve is basically the same as that used by Wang et al. [12] . It will not be repeated here. The cutting edge curve of the slice tool is shown in Figure 6.
图 6 The cutting edge of the ATST-ERA
The final formed tooth surface of the workpiece is made by multiple cuts of the cutting edge of the cutter, and the motion trace of the cutting edge relative to the tooth surface of the workpiece is the edge-sweeping surface. The cutting edge of the slice tool is usually indicated on the tool coordinate system S2. The tooth surface of the workpiece is usually indicated on the workpiece coordinate system S1. To accurately calculate the position relationship between the edge-sweeping surface and the tooth surface of the workpiece, the coordinate transformation is necessary to be done between coordinate system S2 and S1. In order to facilitate the expression later, the cutting edge on the tool coordinate system S2 is selected to be transformed to the workpiece coordinate system S1. On the workpiece coordinate system S1, the tooth surface of the workpiece is fixed. The cutting edge is fixed on the tool coordinate system S2. After the cutting edge on the tool coordinate system S2 is transformed to the workpiece coordinate system S1, the position of the cutting edge changes with time, and a motion trajectory is formed on the workpiece coordinate system S1, which is the edge-sweeping surface.
The spatial curved surface formed by the cutting edge of any one of the arc tooth cutters swept on the workpiece coordinate system S1 is shown in Figure 7. In the figure, where, j r represents the vector of the point on the j-th edge-sweeping surface on the workpiece coordinate system S1 , r r represents the vector of the point on the cutting edge on the tool coordinate system S2, j indicating the distance between the point on the cutting edge of the j-th edge-sweeping surface on the tool coordinate system S2 and the tool axis, T represents the total transformation matrix from the tool coordinate system S2 to the workpiece coordinate system S1. Let The simultaneous equations (1), (5), (6), (7) can obtain the parametric equation of the edge-sweeping surface as cos sin sin sin cos sin cos 11 where, . x2、 y2、 z2 respectively represent components of the vector of the points on the cutting edge in the tool coordinate system S2 on the x2 axis, the y2 axis, and the z2 axis, all of which are functions of r. Analysis of equations (5) and (8) shows that the mathematical model of the edge-sweeping surface can be seen as a parametric equation for the variables r and . As shown in Figure 8, the edge-sweeping surface of the arc-tooth slice tool is obtained according to the above formula.
The edge-sweeping surfaces (Σ1, Σ2, ⋯, Σn, n≥3) formed on the tooth groove Ei is the edge-sweeping surface family. Assume that the center angle of the tooth Fi on the slice tool is θ at the time t1. The center angle of the tooth Fk on the slice tool is k after the workpiece is rotated for n weeks. Where The simultaneous formulae (6) and (11) (12) According to the formulas (1), (11) and (12) . The edge-sweeping surface family of the arc-tooth slice tool is shown in Figure 9. The cutting amount of the caries processing is very small, and it takes many times to complete the machining of the tooth surface. In order to make the cutting amount as reasonable as possible, the cutting layer in the radial direction is divided into multiple parts, and the layer-by-layer cutting of the workpiece in the radial direction is realized by sequentially adjusting the vertical distance (center distance) between the tool axis and the workpiece axis, until the workpiece tooth surface meets the accuracy requirements.
The slicing of the tooth groove Ei is taken as an example. Figure 10 is a schematic diagram of the sweeping of the tooth slot Ei by two adjacent edge-sweeping surfaces.
It is assumed that the center distance between the arc-tooth slice cutter and the workpiece is a1. Assuming that the cutting amount is δ1, the total cutting amount δ needs to be completed to form the tooth surface of the workpiece that meets the accuracy requirements. Adjust the center distance between the slice tool and the workpiece so that the center distance is a2. Repeat the previous process and the cutting amount that can be completed is δ2. Cycling and reciprocating, when the cumulative amount of cutting reaches δ, the arc-tooth slice cutter completes the machining of the tooth surface of the workpiece to form the final gear. The method of deriving the tooth surface equation is basically the same as that used by Wang et al. [12] .
It will not be repeated here. The tooth surface of the arc-tooth gear is shown in Figure 11. It can be seen from the above analysis that in the cut-in state and the cut-out state, the upper and lower end faces of the workpiece tooth grooves serve as geometric boundaries of the undeformed chips, so that the geometry of the undeformed chips is incomplete and complicated. In addition, most of the cutting conditions during the gear slicing process are completely cut. Therefore, the main research on the undeformed chips in the full cutting state is studied here.
The arc gear is machined by the ATST-ERA as the carrier, and the tool, workpiece and machining parameters are shown in Table 1. Taking a tooth groove Ei of the gear workpiece as an example, during the machining process of the tooth, the p-1 tooth cutting groove Ei-1, the p-th tooth cutting tooth groove Ei, ,the p+1th tooth cutting tooth groove Ei+1., and after the time t, the q-th tooth cutting the tooth groove Ei. With the pth tooth contact slot Ei as the initial time, the q-th tooth starts to contact the slot Ei after the workpiece rotates one revolution, and the required time is t.
According to the geometric relationship, the chip length along the tooth width direction is: In the formula, n1 represents the workpiece rotation speed, v1 represents the workpiece feed speed, and f represents the feed rate. (16) In the formula, t1 indicates the moment when the j+1th blade sweep surface is just in contact with the tooth surface obtained by the previous cutting process, and t2 indicates the moment when the j-th blade sweep surface is just in contact with the tooth face obtained by the previous cutting process. , t3 represents the moment when the j-th edge sweeping surface is in contact with the finally formed tooth surface.
According to the basic theory of differential geometry, the chip volume corresponding to the spatial path of the cutting edge can be obtained: (17) In this formula, According to the above formula, the shape of the undeformed chips can be obtained by means of the three-dimensional modeling software, as shown in Figure14.
Experimental verification
Using the geometry model of undeformed chip, the undeformed chips with different cutting parameters can be obtained. For the different cutting parameters in Table 2, the undeformed chips got by simulation calculation are shown in Figure 15. In order to verify the geometric model of the undeformed chips, the cutting parameters in Table 2 were selected, and a trial cutting experiment was carried out on the cutting machine tool to collect chips. As mentioned above, for the complete cutting state, the shape of the chips obtained by cutting the workpiece material with the cutting tool teeth is basically the same. Therefore, the chips with common shapes are collected in the experiment as a typical example. The chips produced by different cutting parameters are shown in Figure 16. As can be seen from Figure 15, the contour of the undeformed chip is a special free curve, and its projection on the plane is approximately crescent-shaped. As for the chip thickness, when the cutting parameters are unchanged, the chip thickness in the cutting-in and cutting-out stages is the smallest, and the chip thickness is the largest in the part corresponding to the boundary line of the blade-sweeping surface. In addition, when the cutting parameters change, the chip thickness increases as the depth of cut increases. Figure 16 shows that the chips collected in the experiment are all deformed chips, and the main deformation form of the chips is curling. The chips after curling are also approximately crescent-shaped. As the cutting depth increases, the chip thickness also shows an increasing trend. Through the above qualitative comparative analysis, it can be seen that the shape of the chip obtained by the simulation calculation is basically the same as the shape of the chip collected by the experiment. The chip shape obtained by the experiment is partially bent or broken which is caused by the extrusion or the friction during the chip-forming process. In addition, the thickness changes of the chips got by the simulation calculation and the experiment are basically the same, and both have a positive correlation with the depth of cut.
In order to quantitatively analyze and compare the simulated and experimental For the undeformed chips obtained by the simulation calculation, the undeformed chips are sectioned in Solidworks, and then the maximum chip thickness is measured; for the deformed chips obtained by the experiment, the curled chips are first spread out, and then with the help of the optical microscope, the thickness value of the deformed chip is obtained by measuring the distance between the upper and lower surfaces of the chip, as shown in Figure 18. For three groups of different cutting parameters, the measured simulation and experimental chip shape parameter values (including: contour radius, arc length, chord length and chip thickness) are shown in Table 3. According to the empirical formula of cutting load proposed by scholar Kienzle [32] , the cutting thickness and the area of the cutting layer directly affect the cutting load during the cutting process. During the cutting process, the force of the cutting edge against the chip and the force of the chip against the cutting edge are equal and opposite, so the calculation of the load of the cutting edge can be converted into the calculation of the chip load. The chip load can be calculated by the following formula.
Where kc represents the cutting force per unit area when the nominal section thickness and the nominal width of the chip are 1 mm, A represents the area of the cutting layer, h represents the cutting thickness, and u is a constant value coefficient, indicating the influence of the cutting thickness h on the cutting load.
According to the chip forming principle of the gear slicing, the chip formed by the edge-sweeping surface sweeping the workpiece tooth groove reflects the shape of the cutting layer of the cutting edge of the slice tool at different times, as shown in Figure 19. According to the formula 16, the chip thickness s(t,h) of the cutting edge of the slice tool can be extracted at any time, as shown in Figure 20. The contour of the cutting edge of the cutting edge of the slice tool is mainly crescent-shaped. According to the shape of the cutting layer, the area of the cutting layer micro-element at any point on the cutting edge determined by the time t can be calculated as The area of the micro-element at a certain point of the cutting edge is shown in Figure 21. According to the calculation formula of the Kienzle cutting load, the micro-area cutting load at any point on the cutting edge can be derived: Then, the cutting load per unit length of the cutting edge is: In the formula, The linear value of the cutting edge load of the different teeth of the arc tooth cutter can more accurately reflect the comprehensive loading condition of the different cutting edges of the arc tooth cutter. The integrated load of the blade is: Select the parameters in Table 1 to calculate the cutting edge load of the ATST-ERA. The calculated coefficients Kc and u in the Kienzle formula are related to the workpiece material and are usually obtained experimentally. The workpiece material selected here is QT450-10, the parameter value in reference [33] , taking Kc=1451N/mm2, u=0.17. During the machining process of the caries, there are differences in the cutting process of the different teeth of the arc tooth cleaver. In order to get closer to the real situation, the teeth of No. 3, No. 6 and No. 15 of the arc knives are randomly selected for calculation. The calculated load spectrum of the three teeth is shown in Figure 22.
It can be seen that the load distribution has a tendency to gradually decrease toward the both sides in the middle, and the position where the load is the largest is at the left-toleft position in the middle. Considering that the cutting load is the main cause of tool wear, the parameters in Table 2 are selected for the cutting experiment. After testing 100 gear workpieces, the optical microscope of KEYENCE VX-1000 is used to photograph the rake faces of arc cutters No. 3, No. 6 and No. 15, and the wear conditions are shown in Figure 23. It can be seen from the figure that the left part of the cutting edge wears the most severely and gradually weakens to the two sides, and the wear on the left part is more serious than the wear on the right part. Although the number of gear workpieces cut by the cutting tool is small and the wear of the rake face of the cutting teeth is not too serious, there is still crater wear near the cutting edge on the No. 15 cutting tooth, as shown in Figure 24. In order to further analyze the relationship between the cutting edge load distribution and the tool wear, the load spectrum of the No. 15 cutter cutting edge is compared with the rake face wear, as shown in Figure 25. It can be seen from the figure that the wear of the cutting edge in the middle left position is the most serious, and the degree of wear to the left side of the cutting edge is gradually lower, which is also consistent with the average load distribution of the cutting edge. Therefore, the wear amount of the arc tooth cutter is positively correlated with the load distribution of the cutting edge. | 6,668.2 | 2021-05-27T00:00:00.000 | [
"Materials Science"
] |
'Defiant', A Today Unique Helicopter in the World
You do not see helicopters like this every day. Sikorsky-Boeing SB1 Defiant in his first demonstration flight showed that it is much more than just a fancy design on paper as he could initially think, with a first relevant flight demonstration. The initial flight lasted less than 30 min, enough time for Defiant to prove he can climb and descend easily, move back and forth and turn left and right with extraordinary ease and maneuverability. Companies have described this action as a low-speed flight maneuver and a short film has revealed that the helicopter can also operate in the area of about 10 to 20 m of land successfully. His first start in the inaugural flight test was held at a Sikorsky airport in West Palm Beach, Florida. SB1 Defiant has an unusual design that instantly conquers you. If most helicopters have one main rotor for vertical lifting, "Defiant" has a pair of coaxial rotors, one rotating clockwise and the other rotating trigonometrically (counterclockwise) to achieve a balancing the dynamics of the couple - that is, to prevent it from overturning one side or another. This means that the aircraft no longer needs the standard rotor at the helicopter tail. Instead, it has a new, distinctive, rear-wheel drive, which has the role of getting Defiant's zoom along the high-speed horizontal flight. "Defiant" is designed to fly at almost twice the speed of classic helicopters while maintaining the best low speed and hover performance of conventional helicopters. Coaxial design can also be found in some Russian military helicopters but without a propeller. Sikorsky used the project earlier (with the propeller) in the experimental X2 helicopter and the next S-97 Raider. The Defiant Project is considered by the United States Army for the Future Lift Vertical program, which intends to find a replacement for many of the US military helicopters used today (Finally, as in the early 2030s, where Defiant's rival in that area is the Bell V-280 Valor, which is not a helicopter, but rather a V-22 Osprey-like tiller-like aircraft used by US marines. The V-280 has been able to improve flight quality since its first flight in December 2017). The Sikorsky-Boeing team will then analyze the flight data recorder and film on this first demonstration flight with the aim of establishing a plan for several "Defiant" test flights in the coming weeks and months to confirm other important features of the new helicopter, before making a complete presentation, for massive future production orders. Sikorsky and Boeing have long been working on this high-speed military helicopter project and have released the first images of what is expected to become a very efficient utility helicopter of the United States Army sometime in 2030. SB1 Defiant is a technology demonstrator with Future Lift technology, built around the Sikorsky X2 platform, using two counter-rotating blades at the top of the plane to eliminate the problem of a 150 km/h traction rotor (~ 240 km/h).
Introduction
Sikorsky and Boeing have reunited this time to build a new helicopter, able to quickly take off vertically, thus achieving a true ultra-fast lift, but also with great maneuverability, capabilities that will give it some advantages special in front of classical helicopters. It can move easily in any direction, simultaneously with tilt maneuvers and with very fast attack and defense, regardless of the height it is found, being able to move permanently including the zigzag as we have shown already at even low altitudes, to which other helicopters do not have the courage to handle either side or the other, in order not to become unbalanced and fall out of the lack of altitude. In other words, it is a beehive or a beehive bee that may, while struggling to carry out various operations including military, obviously of recognition, espionage, or attack even during confusing flight including low altitude, acceleration and large decelerations, with higher speeds than classic helicopters, which can double from them.
Sikorsky and Boeing have long been working on this high-speed military helicopter project and have released the first images of what is expected to become a very efficient utility helicopter of the United States Army sometime in 2030. SB1 Defiant is a technology demonstrator with Future Lift technology, built around the Sikorsky X2 platform, using two counter-rotating blades at the top of the plane to eliminate the problem of a 150 km/h traction rotor (~ 240 km/h).
Simply placing it in a stable hover without wind, a helicopter tip blade generates an equal amount of lifting of its entire rotation around the central axis.
But when you begin to move through the air, the blade starts to develop an extra height on the side where the blades hurry forward in the wind and rise less on the opposite side, where the blades turn with the wind.
This can become so unbalanced that it threatens to defy the helicopter altogether if you do not stay at a defined maximum air speed.
Sikorsky technology places two main rotors one above the other, rotating in opposite directions, balancing the lift profile on both sides and allowing the ship to fly much faster -up to twice the speed of a regular helicopter -agility.
It also eliminates the need for a rotor because the rotation can only be managed with top rotors. SB1 Defiant uses a propeller to the rear, along with active hooks and lifts.
At take-off, this means that the aircraft can rise and move quickly without having to lean forward. At higher speeds, the rear carrier provides an extra mechanism and elevators and lifts help with quick handling.
Defiant will have a retractable landing system, reducing traction to allow higher speeds at greater efficiency, resulting in a greater distance. Its double rotor system will reduce leakage and stop leakage and Sikorsky also claims a "dramatically reduced acoustic signature". On the ground, the top rotors can be folded back, allowing for easy storage and transport.
When he enters the service, Defiant will wear a team of four and a cabin equipped for up to 12 troops ready for battle or eight medevac rifles.
Materials and Methods
You do not see helicopters like this every day. Sikorsky and Boeing have reunited this time to build a new helicopter, able to quickly take off vertically, thus achieving a true ultra-fast lift, but also with great maneuverability, capabilities that will give it some advantages special in front of classical helicopters.
It can move easily in any direction, simultaneously with tilt maneuvers and with very fast attack and defense, regardless of the height it is found, being able to move permanently including the zigzag as we have shown already at even low altitudes, to which other helicopters do not have the courage to handle either side or the other, in order not to become unbalanced and fall out of the lack of altitude.
In other words, it is a beehive or a beehive bee that may, while struggling to carry out various operations including military, obviously of recognition, espionage, or attack even during confusing flight including low altitude, acceleration and large decelerations, with higher speeds than classic helicopters, which can double from them. Electronic copy available at: https://ssrn.com/abstract=3417345 Sikorsky-Boeing SB1 Defiant in his first demonstration flight showed that it is much more than just a fancy design on paper as he could initially think, with a first relevant flight demonstration.
The initial flight lasted less than 30 minutes, enough time for Defiant to prove he can climb and descend easily, move back and forth and turn left and right with extraordinary ease and maneuverability (Fig. 1).
Companies have described this action as a low-speed flight maneuver and a short film has revealed that the helicopter can also operate in the area of about 10 to 20 meters of land successfully. His first start in the inaugural flight test was held at a Sikorsky airport in West Palm Beach, Florida. SB1 Defiant has an unusual design that instantly conquers you. If most helicopters have one main rotor for vertical lifting, "Defiant" has a pair of coaxial rotors, one rotating clockwise and the other rotating trigonometrically (counterclockwise) to achieve a balancing the dynamics of the couple -that is, to prevent it from overturning one side or another. This means that the aircraft no longer needs the standard rotor at the helicopter tail. Instead, it has a new, distinctive, rearwheel drive, which has the role of getting "Defiant's zoom" along the high-speed horizontal flight.
"Defiant" is designed to fly at almost twice the speed of classic helicopters while maintaining the best low speed and hover performance of conventional helicopters.
Coaxial design can also be found in some Russian military helicopters but without a propeller. Sikorsky used the project earlier (with the propeller) in the experimental X2 helicopter and the next S-97 Raider.
The Defiant Project is considered by the United States Army for the Future Lift Vertical program, which intends to find a replacement for many of the US military helicopters used today (Finally, as in the early 2030s, where Defiant's rival in that area is the Bell V-280 Valor, which is not a helicopter, but rather a V-22 Osprey-like tiller-like aircraft used by US marines. The V-280 has been able to improve flight quality since its first flight in December 2017).
The Sikorsky-Boeing team will then analyze the flight data recorder and film on this first demonstration flight with the aim of establishing a plan for several "Defiant" test flights in the coming weeks and months to confirm other important features of the new helicopter, before making a complete presentation, for massive future production orders.
Results
Defiant is designed to fly at nearly twice the speed and has twice the range of conventional helicopters while retaining the very best, if not better low-speed and hover performance of conventional helicopters... This design provides for exceptional performance in the objective area, where potential enemy activity places a premium on maneuverability, survivability and flexibility. We are thrilled with the results of today's flight and look forward to an exciting flight test program.
Defiant is a further extrapolation of Sikorsky's X2 technology demonstrator, which pioneered its unique configuration.
The S-97 Raider, an armed reconnaissance coaxial-rotor compound helicopter that the company also derived from the X2 concept and has helped further inform the development of the SB1, has been flying for years now and two prototypes are currently in testing. Lockheed Martin has since purchased Sikorsky, but it continues to operate under that name and as a separate business unit.
The Defiant and the Valor are set to go head to head as part of the Army's Joint Multi-Role (JMR) technology demonstration program. The service had hoped to begin these flight tests in December 2017 but pushed its timeline back due to the delays with the SB1.
The Army plans to use the results from JMR to help better understand and define the Future Vertical Lift (FVL) program requirements. The SB>1 and V-280, or refined versions of these designs, will be heading for a brawl over the "medium" portion of that project, which aims to replace hundreds of UH-60 Black Hawks and AH-64 Apaches. The Sikorsky-Boeing team plans to pitch the Defiant as a successor to the Black Hawks and has shown a companion gunship design as the Apache replacement.
Discussion
Sikorsky and Boeing have long been working on this high-speed military helicopter project and have released the first images of what is expected to become a very efficient utility helicopter of the United States Army sometime in 2030. SB1 Defiant is a technology demonstrator with Future Lift technology, built around the Sikorsky X2 platform, using two counter-rotating blades at the top of the plane to eliminate the problem of a 150 km/h traction rotor (~ 240 km/h).
Simply placing it in a stable hover without wind, a helicopter tip blade generates an equal amount of lifting of its entire rotation around the central axis. But when you begin to move through the air, the blade starts to develop an extra height on the side where the blades hurry forward in the wind and rise less on the opposite side, where the blades turn with the wind. This can become so unbalanced that it threatens to defy the helicopter altogether if you do not stay at a defined maximum air speed.
Sikorsky technology places two main rotors one above the other, rotating in opposite directions, balancing the lift profile on both sides and allowing the ship to fly much faster -up to twice the speed of a regular helicopter -agility.
It also eliminates the need for a rotor because the rotation can only be managed with top rotors. SB1 Defiant uses a propeller to the rear, along with active hooks and lifts. At take-off, this means that the aircraft can rise and move quickly without having to lean forward. At higher speeds, the rear carrier provides an extra mechanism and elevators and lifts help with quick handling.
Defiant will have a retractable landing system, reducing traction to allow higher speeds at greater efficiency, resulting in a greater distance. Its double rotor system will reduce leakage and stop leakage and Sikorsky also claims a "dramatically reduced acoustic signature". On the ground, the top rotors can be folded back, allowing for easy storage and transport.
When he enters the service, Defiant will wear a team of four and a cabin equipped for up to 12 troops ready for battle or eight medevac rifles. It is in an assault configuration; there will be a variant of attack that shares a common transmission and many other systems, but it has a different fuselage and is more armed.
Conclusion
Sikorsky and Boeing have reunited this time to build a new helicopter, able to quickly take off vertically, thus achieving a true ultra-fast lift, but also with great maneuverability, capabilities that will give it some advantages special in front of classical helicopters.
It can move easily in any direction, simultaneously with tilt maneuvers and with very fast attack and defense, regardless of the height it is found, being able to move permanently including the zigzag as we have shown already at even low altitudes, to which other helicopters do not have the courage to handle either side or the other, in order not to become unbalanced and fall out of the lack of altitude. In other words, it is a beehive or a beehive bee that may, while struggling to carry out various operations including military, obviously of recognition, espionage, or attack even during confusing flight including low altitude, acceleration and large decelerations, with higher speeds than classic helicopters, which can double from them.
Sikorsky and Boeing have long been working on this high-speed military helicopter project and have released the first images of what is expected to become a very efficient utility helicopter of the United States Army sometime in 2030. SB1 Defiant is a technology demonstrator with Future Lift technology, built around the Sikorsky X2 platform, using two counter-rotating blades at the top of the plane to eliminate the problem of a 150 km/h traction rotor (~ 240 km/h).
Acknowledgement
The work was appreciated by teams of professors from the departments of automobiles from several universities in Romania and Italy. This text was acknowledged and appreciated by Associate Professor Aniello Riccio SECONDA UNIVERSITA' DEGLI STUDI DI NAPOLI Italy, whom we thanks and in this way.
Ethics
Author declares that are not ethical issues that may arise after the publication of this manuscript. This article is original and contains unpublished material. Optimization of energy dispersive x-ray fluorescence spectrometer to analyze heavy metals in moss samples. | 3,652.6 | 2019-01-01T00:00:00.000 | [
"Physics"
] |
Ancient Delta Deposits in the Ivie Creek Area, Ferron Sandstone Member of the Mancos Shale, Western San Rafael Swell, East-Central Utah
In contrast to the beautiful array of colorful layers and spectacular cliffs of the Triassic and Jurassic (251 to 148 million years ago [Ma]) sections in the San Rafael Swell of east-central Utah, most of the Upper Cretaceous (96 to 86 Ma) Mancos Shale produces a drab, barren landscape. However, lying within the Mancos, the Ferron Sandstone, is the most studied unit in the San Rafael Swell. The Ferron has world-class outcrops of rock layers deposited near the shorelines of a sinking, fluvial- (stream) dominated delta system. Along the west flank of the San Rafael Swell, the 80-mile-long (130 km) Ferron outcrop belt of cliffs and side canyons (e.g., the Coal Cliffs, Molen Reef, and Limestone Cliffs [not actually limestone, just misnamed]) provides a three-dimensional view of vertical and lateral changes in the Ferron’s rock layers (facies and sequence stratigraphy), and, as such, is an excellent model for fluvial-deltaic oil and gas reservoirs worldwide (e.g., Chidsey and others, 2004).
INTRODUCTION
In contrast to the beautiful array of colorful layers and spectacular cliffs of the Triassic and Jurassic (251 to 148 million years ago [Ma]) sections in the San Rafael Swell of east-central Utah, most of the Upper Cretaceous (96 to 86 Ma) Mancos Shale produces a drab, barren landscape. However, lying within the Mancos, the Ferron Sandstone, is the most studied unit in the San Rafael Swell. The Ferron has world-class outcrops of rock layers deposited near the shorelines of a sinking, fluvial-(stream) dominated delta system. Along the west flank of the San Rafael Swell (figure 1), the 80-mile-long (130 km) Ferron outcrop belt of cliffs and side canyons (e.g., the Coal Cliffs, Molen Reef, and Limestone Cliffs [not actually limestone, just misnamed]) provides a three-dimensional view of vertical and lateral changes in the Ferron's rock layers (facies and sequence stratigraphy), and, as such, is an excellent model for fluvial-deltaic oil and gas reservoirs worldwide (e.g., Chidsey and others, 2004).
The Ferron Sandstone consists of a stacked series of sandstone-dominated rock layers deposited as local sea level rose and fell (depositional transgressive-regressive cycles). Collectively these sandstone layers form an eastward-thinning wedge into the Mancos Shale. The Ivie Creek area along the north side of Interstate 70 (I-70) (figure 2), displays spectacular and abrupt changes in two of these regional-scale depositional cycles ("parasequence sets" referred to as Kf-1 and Kf-2 [figure 3]). These cyclic deposits represent the those typically found in a deltaic oil and gas reservoir. The Kf-1 parasequence set displays spectacular clinoforms. The term "clinoform" is used to identify a group of beds which are inclined seaward in an en echelon pattern and generally separated from one another by a distinctive bounding surface. As a reservoir model, the Ferron in the Ivie Creek area displays variations that influence the reservoir (both its compartmentalization and permeability). These outcrops are often a standard stop for geology field trips. Thus, the Ferron Sandstone in the Ivie Creek area was selected as a geosite (also see Anderson's Ferron geosite in Emery County, this volume) providing a more complete picture of this classic outcrop.
HOW TO GET THERE
The Ferron Sandstone Ivie Creek area is about 170 miles (270 km) or a 2 hour and 45-minute drive from Salt Lake City, Utah, via I-15 and U.S. Highway 6 to State Highway 10 to I-70. Proceed eastbound on I-70 for 6 miles (10 km) to the junction with County Road 912 (Miller Canyon Road), cross the overpass and re-enter I-70 heading westbound for about 2.3 miles (3.7 km) to a well-used but difficult to spot dirt pull-off area for the best view (to the north) of the Ferron clinoforms, parasequence sets, and fluvial-deltaic facies in the "Ivie Creek amphitheater, " an informal name applied to a broad, curving area of cliffs north of I-70 (38°48'35" N., 111°15'10" W., elevation 5813 feet [1772 m]) (figures 1 and 3). The Ferron Ivie Creek geosite is about 61 miles (98 km) west of the junction of U.S. Highway 6 and I-70 near the town of Green River, Utah (figure 1).
Upon approaching the Ivie Creek stop carefully slow down with flashers on and pull off the interstate down a slight incline to a well-used parking area along the right-of-way fence. There are no well-marked trails to examine the Ferron outcrops up close; the terrain is rugged with vertical cliffs and requires first negotiating a fence designed to keep livestock and wildlife from the interstate. Anderson and others (1997) provided a geologic field guide to several key locations within the Ferron section in the Ivie Creek amphitheater (figures 2 and 3).
San Rafael Swell
The Ferron Sandstone is exposed in an outcrop belt along the west flank of the San Rafael Swell, a broad, asymmetric, north-south-to southwest-northeast-trending anticline about 75 miles (120 km) long and 35 miles (56 km) wide. The structure formed in response to compressional forces of the Laramide orogeny between latest Cretaceous time (about 70 Ma) and the Eocene (about 40 Ma) (Hintze and Kowallis, 2009 and references therein) (figure 1). Uplift and erosion has made it a showcase of Colorado Plateau geology with a colorful array of sedimentary rocks over 7000 feet (2100 m) thick, ranging in age from Permian (276 Ma) to Cretaceous (86 Ma), exposed in spectacular cliffs along cuestas, mesas, and deep canyons. The rocks in the San Rafael Swell are folded, faulted, jointed, fractured, and uplifted, with deformation likely controlled by a large, blind, basement-involved reverse fault (up on the west side) bounding the east flank of the structure. Beds in the Ferron outcrop belt dip west from 2° to 12° with only a few minor faults present (Gloyn and others, 2003;Quick and others, 2004). Small to large subsidiary anticlines and synclines are found north to south along the uplift. Ferron field produces natural gas from the Ferron Sandstone in one of these subsidiary structures northeast of the Ivie Creek area; farther north coalbed methane is produced from Ferron coal seams in a series of major fields (Wood and Chidsey, 2015).
Stratigraphy
The Mancos Shale is divided into four members along the west flank of the San Rafael Swell, which in ascending order are: Tununk, Ferron Sandstone, lower Blue Gate, and Emery Sandstone (Hintze and Kowallis, 2009 and references therein). It epitomizes deposits from the Cretaceous Interior Seaway on its western margin and the influence of the Sevier orogenic belt in western Utah.
The Ferron Sandstone ranges in thickness from 150 to 400 feet (45-230 m) on the west flank of the San Rafael Swell (Witkind, 1979(Witkind, , 1980(Witkind, , 1988Weiss and others, 1990;others, 2009, 2015;Hintze and Kowallis, 2009 and references therein;Kuehne, 2012, 2016). The Ferron is informally divided into lower and upper sections. The lower Ferron consists of two thin-bedded silty intervals called, in ascending order, the Clawson figure 9A); note that Kf-1 bedsets are also annotated (see figure 9C). The Kf-2 represents a significant change to wave-modified deposition (subtle left-to-right dips) as seen by shoreface, distributary channel, bay, and coastal plain/swamp facies. Modified from Anderson and others (2004). Photograph by Michael Chidsey, Sqwak Productions Inc. and Washboard units, that are within the Tununk Member ( figure 6). The upper Ferron is a series of major cliff-forming sandstone units described by Ryer (1981Ryer ( , 1991, Gardner (1993), Barton (1994), Barton and others (2004), Garrison and van den Bergh (2004), , and many other workers. They have been referred to as "delta-front units" (Ryer, 1981), "genetic units" (Gardner, 1993), and "stratigraphic cycles" (Barton, 1994). defined these units as parasequence sets based on genetically related parasequences within each set, bounded by major flooding surfaces, and they paired the parasequence sets to major coal zones (note: a parasequence is a small-scale, genetically related succession of bedsets bounded by marine flooding surfaces or their correlative surfaces on top and at the bottom [ Van Wagoner and others, 1988]). The Ferron parasequence sets create stacks of seaward-stepping, vertically stacked, and landward-stepping packages of rocks, each generally composed of sandstone, siltstone, mudstone, and coal intervals (Gardner and others, 2004). There are nine parasequence sets in the Ferron, which referred to as Kf-Last Chance through Kf-8: "Kf " for Cretaceous Ferron Sandstone, followed by the first, second, third, etc. (Kf-LC, Kf-1, Kf-2…) (figures 6 and 7); we use this nomenclature. The parasequences in each set include an abbreviation for the location followed by the letter "a" for the lowermost unit (e.g., Kf-2-Iv-a, Kf-2-Iv-b, and Kf-2-Iv-c parasequences, where Iv stands for Ivie Creek [figure 7]) . Bedsets are recognizable on outcrop as distinct and mappable genetic units but unrelated to flooding surfaces ( Van Wagoner and others, 1990;. The nomenclature used for the bedsets in the Ivie Creek area are the Kf-1-Iv[a] and Kf-1-Iv[c] .
Lithology
The Ferron Sandstone consists of yellow-gray, light-brown, to white sandstone and siltstone, gray sandy to black carbonaceous shale, and coal. Sandstone (quartz arenites to quartzo-feldspathic arenites) is very fine to coarse grained, poor to moderately sorted, subrounded to angular, and cemented with calcite, dolomite, or iron oxide (Jarrard and others, 2004). Sedimentary structures in the sandstone that determine facies include ripples, channel scours, soft-sediment deformation, cross-stratification (trough, swaley, and hummocky), and planar beds. Many beds contain rooted zones, plant fossils, and a large variety of trace fossils from burrows to dinosaur tracks. Fauna such as bivalves and gastropods are common; some are so enriched with shells that the rocks are considered coquinas. Bedding in sandstone is thin or lenticular to massive, forming vertical cliffs, whereas siltstone, shale, and coal create recesses and slopes. Coal beds are persistent but lenticular and commonly burned (due to spontaneous combustion or lightning strikes), creating "clinker" zones with red oxidation staining.
Paleogeography and Facies
The Ferron Sandstone was deposited in fluvial-dominated and wave-modified deltaic environments that prograded into the Cretaceous Interior Seaway (figure 8). Sediment was transported by east-to northeast-flowing rivers and streams that originated in the nearby Sevier orogenic belt. The alluvial to lower delta or coastal plain of the Ferron contained meandering streams (creating single-and multi-storied complexes), swamps and peat bogs, distributary channels, levees, crevasse spays/overbanks, and bays. Deltaic facies varied and were primarily controlled by sediment input and accommodation space. Wave energy at the coastline influenced the redistribution of these sediments and fluctuations in sea level had profound effects on the accumulating sediment package. Sediment supply was high during early Ferron deposition resulting in seaward-prograding fluvial-dominated conditions that Lupton (1916) is still followed. The green area represents coastal-and alluvial-plain deposits. After Ryer (1991), Anderson and others (1997), and .
Figure 7. Ivie
Creek area Ferron stratigraphy including parasequence sets, parasequences, and named bedsets. Gray = marine shale, blue and gray-blue = marine siltstone, light orange and yellow = shoreface sandstone, green = bay-fill mudstone and siltstone, black = coal, grass green = coastal and alluvial plain siltstone and mudstone, tan = fluvial sandstone. After Anderson and others (2004). created lobate deltas consisting of delta-front deposits, distributary channels, splay complexes, and interdistributary bays. Later, conditions changed at Ivie Creek to wave-dominated or wave-modified deltaic deposition forming cuspate deltas that grade laterally into strandplains and barrier islands. Facies associated with wave-dominated or wave-modified deltas consist of prodelta deposits; lower, middle, and upper shoreface; foreshore; distributary channels; and distributary-mouth bars. The strandplains and barrier island facies include washover fans, lagoons, bays, and tidal inlets and associated flood-tidal deltas (Ryer, 1991;Ryer and Anderson, 2004). Elongate silty sand bodies, such as the Clawson and Washboard units (figure 6), represent offshore bars produced by longshore drift or sand plumes/turbidites that accumulated on the shallow-marine shelf to the east.
Landslides and Slumps
Landslides and slumps can occur when erosion oversteepens slopes and undercuts resistant bedrock. The Ferron Sandstone is highly jointed and underlain by shale beds of the Tununk Member of the Mancos Shale and thus very susceptible to rockfalls (figures 2 and 3). The Tununk contains significant amounts of swelling clay, particularly bentonite/smectite in volcanic ash beds. These unstable soft units cause joints in the overlying Ferron to further enlarge. Large Pleistocene-(?)-age (1,800,000 to 12,000 years ago) landslides can be observed along the south-facing slopes of the Ferron escarpment, east of Quitchupah Creek along the I-70 route (figure 2).
Emery Coalfield
The Emery coalfield is located along the west flank of the San Rafael Swell from the Limestone and Coal Cliffs continuing into the subsurface under Castle Valley and Wasatch Plateau (figure 1). The coalfield consists of 13 coal beds contained in the Ferron Sandstone that are given letter designations from A to M in ascending order based on Lupton's (1916) Ferron description system (figure 6); this system is still used by geologists working the Ferron over 100 years later.
Ferron coal beds are lenticular, which has limited their commercial development. Seven seams exceed 4 feet (1.2 m) in thickness; the maximum coal seam thickness is 30 feet (9 m) (Doelling, 1972). However, several seams coalesce to form a single, thicker coal with at least one at 25 feet (8 m) thick (Doelling, 1972(Doelling, , 1976. Coal beds A (exposed in the Ivie Creek area), C, I, and M are the most important (figure 6), and several are stratigraphically close to each other. Analyses of these beds indicate an overall high-volatile C bituminous coal (Doelling, 1972(Doelling, , 1976. Estimated available, inplace coal resources for the southern Emery coalfield are 2.2 billion short tons, based on coal beds greater than 4 feet (1.2 m) thick and less than 3000 feet (900 m) deep, of which about 500 million short tons are recoverable (Quick and others, 2004). Presently, one coal mine is extracting coal from the coalfield from the I seam.
Kf-1 Parasequence Set
The Kf-1 parasequence set consists of river-dominated deltaic deposits that prograded from the south-southeast to the north-northwest, proximal to distal, as delta lobes across the Ivie Creek area. Progradation was parallel or onshore to the northwest-southeast regional shoreline trend. The Ivie Creek area is best known for its distinctive, steeply inclined bedsets or clinoforms dipping west-northwest 10° to 15° (figures 3 and 9). These unique sedimentary features have been the subject of major academic and industry studies and are incorporated in hydrocarbon reservoir simulation models others, 1997, 2004;Mattson, 1997;Forster and others, 2004;Jarrard and others, 2004;Mattson and Chan, 2004;Enge and Howell, 2010;Deveugle and others, 2011;Graham and others, 2013Graham and others, , 2015aGraham and others, , 2015b. 7), that represent the two delta lobes sourced from a common point .
Kf-1-Iv[a] Bedset
The Kf-1-Iv[a] bedset can be traced on outcrop from landward to seaward pinchout for about 1.5 miles (2.4 km). Delta-front sandstones of a modified Gilbert delta make up these deposits, which change from proximal to distal as they pass from east to west across the Ivie Creek area. This bedset is a fan-like deposit that thickens to the east and was deposited into an area having minimal wave influence; therefore, the primary beds are preserved as clinoforms (figures 3 and 9). The clinoforms change characteristics from the shallower water, more proximal locations, to the deeper water, more distal locations. They thin rapidly to the west, as well as thin to the more typical seaward direction of north. This is anomalous for Ferron delta lobes.
• Clinoform proximal facies is sandstone, mostly fine to medium grained. The chief sedimentary structure is low-angle cross-stratification with minor horizontal and trough cross-stratification and rare hummocky bedding ( figure 11). This facies is dominantly thick-to medium-bedded, well to moderately indurated.
Clinoform proximal facies is interpreted to be the highest energy and most proximal to the sediment supply. The steep inclinations are interpreted to represent deposition into a relatively localized deep portion of an open bay environment.
• Clinoform medial facies is dominantly sandstone with about 5% shale ( figure 12). The sandstone is primarily fine grained with slightly more fine-to very fine grains than fine to medium grains. Horizontal beds dominate with some rippled, trough, and low-angle cross-stratified beds (figure 13). Bed thickness ranges from laminated to very thick, but most beds are medium. Clinoform medial facies is generally transitional between end members clinoform proximal and distal.
• Clinoform distal facies is sandstone (sometimes silty) and about 10% shale ( figure 14). The sandstone grain size is dominantly fine to very fine grained, with considerable variation. Sedimentary structures in clinoform distal facies are chiefly horizontal laminations and ripples in medium to thin beds ( figure 13). Clinoform distal facies is gradational with clinoform medial facies and represents the deepest water and lowest energy deposits within the clinoform. It can be traced distally into prodelta to offshore facies.
• Clinoform cap facies is sandstone, generally very fine to fine grained. The beds are horizontal, with some trough and low-angle cross-stratification in thick to medium beds (figure 11). The clinoform cap facies is interpreted to represent an eroded and reworked delta top.
The deposits of the Kf-1-Iv[a] bedset accumulated on an arcuate delta lobe that prograded into a deeper-water, fully marine bay. The main delta, located to the east and northeast, created a protected embayment in the northwest part of the Ivie Creek area (Anderson and others, 1997). The Kf-1 clinoforms represent deposition into the embayment fed by river channels from the southeast ( figure 15A) (Anderson and others, 2004). The distributary complexes/delta front, shallow marine, and deep marine environments produced the clinoform proximal, medial, and distal facies, respectively.
Kf-1-Iv[c] Bedset
The Kf-1-Iv[c] bedset is sand rich and varies in thickness within the Ivie Creek area. In contrast to the Kf-1-Iv[a], delta-front, subtle clinoforms of the lower part of the bedset dip less than 5°. This sand-rich bedset thins in both up-depositional-dip and down-depositional-dip directions due to lateral facies changes. A major flooding surface is at the top of the Sub-A coal zone (figure 3 and 16) and corresponds to the boundary with the overlying Kf-2 parasequence set (figures 6 and 7).
The lower section of the Kf-1-Iv[c] laps onto the more distal parts of the Kf-1-Iv[a] in the western part of the area and represents the distal part of another delta lobe, probably sourced from the southwest. This lower section may be completely absent in some locations as a result of erosion and/or non-deposition. It also includes a distributary channel sandstone body, lenticular in cross section, deposited by a northwesterly flowing stream (figures 9B and 9C). Mattson, 1997; for data and detailed descriptions see Anderson and others, 2004). Figure 3, Kf-1 captures approximately the same area as the cross section. The artificially abrupt end of the medial facies is representative of most of the contacts portrayed between proximal, medial, and distal clinoform facies in this study because polygonal packages of rock are necessary for reservoir modeling and simulation. In reality, the transition from medial to distal (as well as medial to proximal) is gradational. In the Ivie Creek amphitheater, above the basal cross-bedded sandstone, are 10 to 15 feet (3-5 m) of coarsening-and bed-thickening-upward, brackish-water/bay-fill deposits consisting of carbonaceous mudstone; thin, rippled to horizontally laminated, sparsely to intensely burrowed sandstone and siltstone; fossiliferous mudstone to sandstone; oyster coquina; and ash-rich coal (figures 16 and 17). The presence of a brackish-water fauna (Crassostrea, Corbula securis, Lucinid, Caryo corbulais, aff. Varia, and Serrthid) is the distinguishing characteristic of this facies. This fauna is present within "dirty" sandstone, siltstone, mudstone, and carbonaceous mudstone. Oyster coquinas commonly are found within the brackish-water/bay facies ( figure 17B). Typically, the rocks are rich in carbonaceous debris. The uppermost, carbonaceous mudstone or ash-rich coal is the Sub-A coal zone (figures 3 and 16). The coal zone (swamp [paludal] facies) shows considerable variation in thickness (0 to 1 foot [0-0.3 m]) and intertongues with underlying bay deposits. The paludal facies is commonly underlain and/or overlain by coastal plain facies (the top is frequently rooted) and may grade into shoreface deposits below.
At the mouth of Ivie Creek Canyon, the Kf-1-Iv[c] bedset is anomalously thick due to large slump features or rotated blocks that formed shortly after deposition. Failure of the rotated blocks is consistently toward the north to northwest, the direction that the delta lobe appears to have prograded.
KF-2 PARASEQUENCE SET
The Ferron sediment source switched from the east-southeast to the west going from Kf-1 to Kf-2 deposition. The Kf-2 parasequence set represents wave-modified deltaic deposits that generally coarsen east to west, consisting of shoreface and distributary complex facies gently inclined (< 3°) to the east (figure 3). These relatively clean, sand-rich deposits accumulated along a local north-south shoreline trend defined by a landward pinchout of marine shoreface facies (observed in the I-70 roadcut just to the west of the Ivie Creek area and farther east in Ivie Creek), as opposed to the more common regional northwest-southeast shoreline trend recognized in other Ferron cycles above and below Kf-2.
The Kf-2 parasequence set contains three parasequences: the Kf-2-Iv-a, Kf-2-Iv-b, and Kf-2-Iv-c (figures 7, 9B, and 9C). These parasequences show less lateral variation in lithofacies than the Kf-1 bedsets due to greater wave influence (reworking). Within the Ivie Creek area, there is also little lateral variation in thickness of sand-rich Kf-2 facies, even when lateral change occurs from one depositional subfacies to another. Kf-2 cycles accumulated in sheet-like bodies that pinchout laterally, forming a wedge due to wave-action along the delta front. The contact between Kf-1 and Kf-2 is generally drawn at the top of the Sub-A coal zone in the Ivie Creek area ( figure 3). The top of Kf-2 lies near the top of the C coal and includes all of the A coal zone and delta-plain strata which separate the A and C coal zones (figure 6).
In general, the western part of the Ivie Creek area, east-to northeast-flowing distributary channels deposited large amounts of sand in north-south-trending distributary-mouth bars. Shallow-to moderate-depth marine conditions existed in the eastern part of the area. An uncommon transition from shoreface, to bay, to coastal plain/ swamp occurred during the late stage of Kf-2 deposition (Anderson and others, 2004). Above the Kf-2 parasequence set are coastal-plain and alluvial-plain facies which represent the landward equivalents of the marine portions of Kf-3 through Kf-7 parasequence sets.
The wave-modified parasequences of the Kf-2 parasequence set consist predominantly of shoreface, distributary complex, and coastal plain/swamp facies.
Shoreface
The shoreface facies consists of a relatively steeply dipping zone (compared to the shelf/slope) from the subaerial beach to a poorly defined point where the slope flattens on the sea floor. Wave energy is sufficient to move sand-size grains in this zone. At the seaward end of the shoreface is the prodelta facies. This mud-dominated facies represents the area just seaward of the dominant influence of wave energy and typically interfingers with the lower shoreface defined below. Anderson and others (2004).
A B C D
The lower shoreface facies consists of thinly interbedded shale to siltstone and very fine to fine-grained sandstone having wave ripples to horizontal laminations; hummocky stratification is common ( figure 16). Burrowing is generally found on the top of thin sandstone beds and the shale is often bioturbated.
Middle shoreface facies consists of very fine to fine-grained sandstone composed of hummocky, swaley, and planar laminations with minor ripple laminations ( figure 18). This facies is generally thick, representing 50% or more of the shoreface sequence. The most common burrow types are Thalassinoides and Ophiomorpha with numerous other shallow marine traces and the amount of burrowing varies from moderately to intensely bioturbated.
Upper shoreface facies is characterized by fine-to medium-grained, multidirectional to bimodal, cross-stratified sandstone, in sets that are occasionally separated by planar laminations, and that is generally moderately to well sorted ( figure 19). This facies is about 10 feet (3 m) thick, but greater in the vicinities of the landward pinchouts of the parasequences where the upper shoreface may reach 20 feet (6 m) in thickness. The upper shoreface facies is slightly to moderately burrowed and Ophiomorpha is the most common trace fossil.
The foreshore facies consists of fine-to medium-grained sandstone with planar to inclined bedding, slightly to intensely burrowing, and is sometimes rooted ( figure 19). This facies is sometimes absent at the top of the shoreface sequence due to erosion, but when present ranges up to a few feet in thickness.
Distributary Complex
The distributary complex facies is subdivided into distributary channel and distributary mouth-bar in the Ivie Creek area (figures 9B and 9C shows an example in the Kf-1). This facies is commonly characterized by the complicated geometry of cross-bedding and large-scale bounding surfaces that are related to cut-and-fill processes, in contrast to the flat to very gently inclined surfaces of the lower delta-front.
A B
The distributary channel facies is common in river-dominated delta-front deposits and is also found within the wave-modified shoreline deposits of the Ivie Creek area (figures 9B and 9C). It is characterized by channels with high height-to-width ratios and unidirectional trough cross-stratified and current ripple cross-laminated deposits. Channel fills are sandstone-dominated, but heterolithic channel fills are also found. In sand-filled channels, the grain size is commonly coarser than the surrounding delta-front or shoreface and fines upward. Troughs in the channel base generally contain mud rip-up clasts, woody fragments, and rare shark teeth. This facies grades seaward into the distributary-mouth bar facies.
The distributary-mouth bar facies is found in the upper parts of delta-front sequences of Kf-2 and is associated with distributary channel deposits. This facies is characterized by fine-grained, or coarser, trough-cross-stratified sandstone, with moderate to intense burrowing in areas that had lower flow velocities and decreased sedimentation rate; some intervals between trough sets are completely bioturbated; Ophiomorpha is common. Paleoflow directions show a strong offshore component with the amount of scatter increasing with increased wave influence and distance from the distributary channel. Traced laterally and seaward, the facies commonly grades into middle shoreface or lower delta-front facies (fluvial-dominated shoreline).
Coastal Plain/Swamp
The coastal plain/swamp facies is represented by a sequence of non-marine rocks dominated by mudstone, carbonaceous mudstone, and siltstone with minor sandstone. Coal is commonly interstratified with the other rock types. These rocks are interpreted as inter-fluvial environments along a low-gradient coast.
Kf-2-Iv-a Parasequence
The oldest parasequence of the Kf-2 parasequence set in the Ivie Creek area is the Kf-2-Iv-a. The Kf-2-Iv-a thickens to 50 feet (15 m) westward across the area. The base of the Kf-2-Iv-a parasequence consists of interbedded sand and minor shale deposited in prodelta to lower-shoreface environments (figures 9C and 16). These deposits are thin, typically less than 10 feet (3 m) thick. The Figure 18. Middle shoreface facies in parasequence Kf-2-Iv-a usually forms a vertical cliff and consists of very fine to fine-grained sandstone that is horizontally bedded and often moderately burrowed to bioturbated. Note the horizontal bedding indicated by the light horizontal lineations in the photograph. Figure 19. Upper shoreface and foreshore facies in parasequence Kf-2-Iv-C, exposed in Ivie Creek Canyon. Note the cross-beds in the base of the upper shoreface, typical of this facies. Ophiomorpha is common in both facies. The foreshore bedding is horizontal to sometimes slightly inclined. Inset shows vertical rootlets penetrating the facies and originating from the overlying A coal.
prodelta and lower-shoreface deposits are overlain by a 0.5-to 1-foot-thick (0.2-0.3 m) zone of highly carbonaceous to coaly sandstone which grades into 20 to 30 feet (6-9 m) of very fine grained, silty, and slightly carbonaceous sandstone representing a middle shoreface environment ( figure 18). The middle-shoreface unit is moderately to intensely bioturbated.
Near the mouth of Ivie Creek Canyon, the Kf-2-Iv-a parasequence consists of a thin sequence of lower shoreface heterolithics overlain by about 28 feet (8 m) of middle-shoreface deposits ( figure 18). The westernmost exposure of Kf-2-Iv-a is in Ivie Creek Canyon where the facies is a distributary-mouth bar. This indicates that the shoreline was probably not too far to the west. Utah Geological Survey (UGS) drill hole Ivie Creek no. 11 contained distributary-mouth bar facies of the unit ( figure 15B), but interpretation of drill holes farther west indicate the shoreline has been crossed and only non-marine deposits are present. The shoreline orientation of the Kf-2-Iv-a parasequence was primarily northsouth ( figure 15B). Possible foreshore deposits are near the last outcrop of the Kf-2-Iv-a in the bottom of Ivie Creek.
Kf-2-Iv-b Parasequence
In the Ivie Creek amphitheater, the Kf-2-Iv-b parasequence was deposited in a middle-shoreface environment. Kf-2-Iv-b exhibits gently seaward-inclined beds that are very conspicuous when viewed from west to east along the outcrop. The Kf-2-Iv-b parasequence consists of horizontally bedded, silty sandstone at the base (middle shoreface) (figure 9C) and unidirectional, trough cross-bedded sandstone at the top (distributary channel to mouthbar deposits). These units are moderately to intensely bioturbated.
Along Ivie Creek Canyon, Kf-2-Iv-b is dominantly unidirectional, trough cross-bedded sandstone of a mouth-bar complex which continues westward up the canyon and is present in the subsurface at UGS drill hole Ivie Creek no. 10 to the northwest ( figure 15C). The unit thickens slightly from west to east, in contrast to the underlying parasequence.
Like the preceding unit, the shoreline orientation during deposition of the Kf-2-Iv-b parasequence was generally north-south. In the landward exposures at Ivie Creek Canyon, the unit is dominated by distributary-mouth bar and distributary channels associated with wave-modified delta deposits ( figure 15C). Farther seaward, the distributary-mouth bar deposits give way to middle-shoreface deposits. Seaward (east) of the Ivie Creek area, a well-developed lower shoreface is present at the base of the unit. Paleoflow of several of the distributary channels indicates a general northerly trend in the local area.
At the I-70 road cut, a highway drainage channel was cut through solid rock to accommodate storm runoff. This cut provides superb views of Kf-2-Iv-b mouth-bar deposits. From the downstream end of the drainage channel, cross-bedded, mouth-bar deposits can be visually followed into distal bar or middle shoreface deposits to the east.
Kf-2-Iv-c Parasequence
The Kf-2-Iv-c parasequence is separated from the underlying Kf-2-Iv-b parasequence by a siltstone to shale interval that varies in thickness across the east part of the Ivie Creek area. Generally, the entire parasequence fines from west to east. In the Ivie Creek amphitheater, Kf-2-Iv-c is interpreted as a bay-fill deposit (although it is devoid of body fossils). There is evidence for bay-head deltas and tidal channels feeding the bay to the northeast in Quitchupah Canyon (figures 2 and 15D). At the top of this sequence is a thin, medium-grained carbonaceous sandstone, which may represent the migration of a low-energy beach (foreshore deposits) across the bay fill prior to capping by coastal-plain deposits and deposition of the overlying A-coal zone (which is locally burned). In Ivie Creek Canyon, Kf-2-Iv-c forms a 10-foot (3 m) cliff having excellent upper shoreface facies exposed ( figure 19). The top of the unit is rooted by the overlying coastal-plain vegetation.
The landward pinchout of the marine facies of Kf-2-Iv-c trends just slightly east of south toward I-70. At the mouth of Ivie Creek Canyon, shoreline sandstone changes over a distance of about 300 feet (90 m) from a strongly wave-modified shoreface unit to a much lower energy unit that contains mud interbeds and finer sand and has a silvery gray color in outcrop, but is carbonaceous on a fresh surface. This suggests a change from a coast directly facing the sea to one that was sheltered from wave energy. The environment of this shoreline unit is a wave-modified coast, probably shoreface facies in the proximal part, transforming laterally to low-wave-energy bay facies.
A large meanderbelt channel system cuts into the top of the Kf-2-Iv-c parasequence just south of the immediate Ivie Creek area.
The channel system flowed to the northeast, but much of this Ferron channel system has been removed by erosion. A late-stage episode of lateral channel migration across the top of the Kf-2-Iv-c delta-plain deposits is recorded by scours into the meanderbelt deposits. The best example of this channel system is exposed on the south side of I-70 across from the Ivie Creek amphitheater, just east of the road cut through the Ferron. This same channel is exposed in the road cut on the north side of I-70, but the channel orientation and the nature of its exposure in the cut face do not present the classic channel shape exhibited in the southern exposure. The meanderbelt and younger channel systems fed the continued progradation of Kf-2-Iv-c and stratigraphic equivalents to the east and northeast. | 7,529.2 | 2019-12-01T00:00:00.000 | [
"Geology"
] |
Frequent mutation of the PI 3 K pathway in head and neck cancer defines predictive biomarkers
Vivian Wai Yan Lui, Matthew L. Hedberg, Hua Li, Bhavana S. Vangara, Kelsey Pendleton, Yan Zeng, Yiling Lu, Qiuhong Zhang, Yu Du, Breean Gilbert, Maria Freilino, Sam Sauerwein, Noah Peyser, Dong Xiao, Brenda Diergaarde, Lin Wang, Simion Chiosea, Raja Seethala, Jonas T. Johnson, Seungwon Kim, Umamaheswar Duvvuri, Robert L. Ferris, Marjorie Romkes, Tomoko Nukui, Patrick Kwok-Shing Ng, Levi A. Garraway, Peter S. Hammerman, Gordon B. Mills, Jennifer R. Grandis
Sentence Statement of Significance:
Treatment options for HNSCC are limited by an incomplete understanding of the targetable mutations that "drive" tumor growth.Here, we define a subgroup of HNSCC harboring activating mutations of genes in the PI3K pathway where targeting the pathway demonstrates antitumor efficacy.These results suggest that PI3K pathway mutation assessment may be used to guide HNSCC therapy.
Introduction
Head and neck squamous cell carcinoma (HNSCC) is a frequently lethal cancer with few effective therapeutic options.Recent genomic findings in head and neck cancer revealed a wide spectrum of unexpected genetic aberrations (1,2).This genomic heterogeneity of HNSCC tumors underscores an obstacle to the identification of effective molecular targeting agents likely to benefit the majority of HNSCC patients.To date, there is a translational gap between genomics and treatment selection for HNSCC patients.TP53 mutation is the single most common mutational event.Yet the loss of function of this tumor suppressor gene has remained challenging to exploit therapeutically.Mitogenic pathways are crucial for cancer development and progression.In other malignancies, mutations of growth pathway genes have been shown to result in pathway activation, enhanced tumor growth, and increased sensitivity to agents targeting the mutated pathway.However, the potential of genomics-based therapy selection has not been widely investigated in HNSCC.
We and others recently reported genomic mutational profiles of over 100 HNSCC tumors (1,2).
Here, we analyzed an additional 45 HNSCC tumors by whole exome sequencing using the Illumina platform.In an effort to identify mutationally altered, targetable mitogenic pathways in HNSCC, we combined all currently available mutational data (from whole-exome sequencing) of 151 HNSCC primary tumors and evaluated the mutational events of genes in three major mitogenic pathways that have been previously implicated in HNSCC pathophysiology, namely the MAPK(3), JAK/STAT(4), and the PI3K pathways (3).These key mitogenic pathways are targetable in human cancers with a variety of agents currently in various stages of clinical development. Research.
on July 4, 2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Nearly one third of HNSCC tumors harbor PI3K pathway mutations
To date, whole-exome sequencing data of 106 HNSCC tumors are available.Here, we reported the whole-exome sequencing data of additional 45 HNSCC tumors collected at the University of Pittsburgh (Supplementary Table 1).Our results, similar to previous reports, showed a high degree of inter-tumor mutation heterogeneity, further confirming the complexity of HNSCC biology and the associated challenges in defining subsets of patients likely to respond to specific targeted therapies.With the aim of identifying mutationally altered, targetable mitogenic pathways in HNSCC, we assessed the mutational events of genes comprising three major mitogenic and targetable pathways in HNSCC; the JAK/STAT, MAPK and PI3K pathways.
A detailed analysis of the PI3K pathway mutational events showed that 19 of the 151 tumors (12.6%) harbor a PIK3CA mutation (Figure 1B).This mutation rate is similar to that detected in prior reports of HNSCC tumors (7.4% and 10.8% rate (6,7)).Further, we found PIK3CG and PTEN mutations in 4.0% (6/151) of HNSCC tumors, while PIK3R1 (also known as p85), PIK3R5 and PIK3AP1 were mutated in 2.7% tumors (4/151).Other components of the PI3K pathway were mutated in <2% cases (Figure 1B).Major downstream effectors of the PI3K pathway, including PDK1, AKT1 were not mutated, while AKT2 and mTOR were only mutated in 1.3% (2 mutations) of HNSCC tumors.Although PIK3CA gene amplification data is not available for the previously sequenced tumors, in the newly added cohort, PIK3CA was amplified in 24.4% (11/45) of the cases.
Previous reports noted loss of heterozygosity (LOH) of the tumor suppressor PTEN in HNSCC, by PCR-based microsatellite analysis primarily using D10S215 and/ D10S541 probes in relatively small HNSCC cohorts (e.g. 17 and 39 tumors, respectively) (8,9).Although comprehensive analysis of PTEN copy number was not available in the published genomic HNSCC studies (1,2), PTEN gene copy number change was analyzed using a highly sensitive Affymetrix Genome-Wide Human SNP Array 6.0 platform in our 45 newly sequenced HNSCC tumors.Our results showed that PTEN gene copy loss (≥1 copy loss) was only found in 8.16% of cases (4/45), indicating that PTEN loss is not likely to be the primary mediator of PI3K pathway alteration in HNSCC (unlike other cancers such as glioblastoma, where PTEN loss can be as high as 20-60%) (10,11).However, all 4 tumors with PTEN gene copy loss expressed relatively low levels of PTEN protein when compared to HNSCC tumors without PTEN gene copy alteration (P<0.001,Supplementary Figure 1).
PI3K pathway-mutated HNSCC tumors demonstrate an increased rate of cancer gene mutations
To determine if HNSCC tumors harboring mutations in PI3K pathway genes contained a higher number of mutations in other cancer-associated genes, we compared the mutation rates of PI3K pathway-mutated tumors to non-PI3K pathway-mutated tumors.We found that tumors harboring PI3K pathway mutations have higher rates of mutation than non-PI3K-mutated HNSCC tumors.
Further, cancer gene filtering analysis showed that these PI3K pathway-mutated HNSCC tumors harbored twice as many cancer gene mutations than those without PI3K pathway mutations (Figure 1D, 7.2±0.8vs 3.6±0.3,P<0.0001: defined by the Cancer Gene Census, COSMIC Database) (12).These results suggest that PI3K pathway mutations in HNSCC may facilitate the expansion or selection of tumor cells that are already genetically unstable and thus harbor more genomic aberrations including aberrations in known cancer genes.This contention is supported by further analysis demonstrating that DNA damage/repair genes (based on the DNA damage gene list in the cBio portal database (13), which includes ATM, ATR, CHEK1/2, BRCA1/2, FANCF, MLH1, MSH2, MDC1, PARP1 and RAD51) were found to be mutated at a significantly higher frequency in the PI3K-mutated tumors (average mutation rate of 37.0%; 17 mutations in 46 tumors) compared to tumors without PI3K pathway mutations (average mutation rate of 15.2%; 16 mutations in 105 tumors) (P=0.033).The association between PI3K pathway mutations and genomic instability is observed in HNSCC derived from all anatomic sites in our cohort (e.g.oral cavity, pharynx and larynx, data not shown).Mutation rates in laryngeal tumors (186.3±27.1,n=32, data not shown) are significantly higher than the rates of mutation in tumors from the other anatomical locations (78.6±8.3,n=116, P=0.0005, data not shown).Additionally, the prevalence of PI3K pathway mutations is higher in laryngeal tumors (53.1%±9.0%,n=32, data not shown) compared with tumors from the other anatomical locations (25.0%±4.0%,n=116, P=0.0005, data not shown).
Co-mutation analysis showed that tumors with PI3K pathway mutation(s) are also associated with mutations of several known tumor suppressor genes including ARID1A, MLL and MLL3 (P<0.05;Supplementary Table 3), which contribute to chromatin remodeling and transcriptional regulation in cancers (14)(15)(16)(17).Intriguingly, ARID1A has been shown to influence signaling through the PI3K pathway, suggesting that ARID1A may regulate the PI3K pathway and expand the number of tumors susceptible to targeting the PI3K pathway (18).Of note, PI3K pathwaymutated tumors are not associated with TP53 (P=1.0) or NOTCH1 mutations (P=0.34,data not shown).
Interestingly, we also identified 3 tumors where PIK3CA or PIK3R1 was the only known mutated cancer gene [HN_00361, HN_63027 and HN41PT with the respective PI3K-mutations of PIK3R1 (453_454insN), PIK3CA(E542K) and PIK3CA(H1047L)].Strikingly, all three tumors were associated with infection by the human papillomavirus (HPV).Although the number of HPV-positive HNSCC tumors in this cohort is relatively small (15 cases, see Supplementary Table 4), 5 of these tumors harbored PI3K pathway mutations (33.3) suggesting that a subset of HPV-positive HNSCC tumors (20%; 3/15 cases) may be driven by PI3K-pathway mutation(s) alone, without an associated increase in underlying genomic instability.
Only advanced stage HNSCC tumors harbor multiple PI3K-pathway mutations
In HNSCC cancers containing PI3K pathway mutations, 21.7% (10/46) harbored mutations in more than one PI3K pathway member genes, indicating that genetic alterations at multiple levels in the PI3K pathway are relatively common in HNSCC (Table 1).In contrast, HNSCC tumors rarely, if ever, harbored multiple mutations in the MAPK pathway (0 tumors), or the JAK/STAT pathway (only one contained both JAK3 and STAT1 mutations; HN_63080) (Figure 1C).Strikingly, all of these HNSCC tumors (100%; 10/10 cases) with concurrent PI3K mutational events were advanced (Stage IV) (Table 1).None of these tumors was associated with HPV infection.These findings suggest that concerted PI3K pathway aberrations may contribute to HNSCC progression.This finding appears to be unique to HNSCC.Examination of recently published tumor datasets including breast, colon and lung SCC showed that only 1/25 breast tumors, 1/27 colon carcinoma, and 0/31 lung SCC tumors harbored multiple PI3K pathway mutations were stage IV (data not shown; cBio portal (13)).While all 10 tumors with multiple PI3K pathway mutations were advanced (Stage IV), there is no significant association between advanced disease and individual PI3K pathway mutations (data not shown).
Additionally, mutation rates do not vary significantly between Stage IV and earlier Stage (I-III)
disease (data not shown).In the absence of models assessing the specific contribution of each mutation to cell growth or survival, it is not possible to determine the precise effect of individual mutations in tumors that harbor more than one mutation of genes in the PI3K pathway.
PIK3CA canonical and novel mutations increase survival and pathway activation in HNSCC tumors
PIK3CA is a critical gene in the PI3K signaling pathway.In HNSCC tumors, the most common sites of PIK3CA mutations included H1047R/L (8 mutations total), E545K/G (4 mutations) and E542K (3 mutations) (Figure 2A).All of which represent previously reported canonical ("hotspot") mutation sites.This HNSCC-PIK3CA mutation pattern (~90% of mutations found in the helical/kinase domains), is similar to that observed in cervical, breast and lung SCC cancers, but is distinct from other tumors such as endometrial cancer, lung adenocarcinoma, glioblastoma multiforme, and prostate carcinoma (Supplementary Table 5).In addition, we detected 4 previously unreported, novel PIK3CA mutations (R115L, G363A, C971R, R975S).To determine the functional consequences of these mutations we stably expressed, by retroviral infection, each of the novel mutations, and a hotspot mutation (H1047R), in a representative HNSCC cell line that is WT for all PI3K pathway components.Overexpression of WT PIK3CA (mimicking PIK3CA gene amplification), and expression of all the engineered PIK3CA mutants individually, resulted in enhanced growth compared to infection with EGFP control.Furthermore, the canonical hotspot mutation showed significantly enhanced growth compared to overexpression of WT PIK3CA (p = 0.0001).The novel mutations were found to confer moderate growth advantage compared to simulated WT amplification (R115L; p = 0.1174, G363A; p = 0.9637, C971R; p = 0.6503, R975S; p = 0.0958).Immunoblotting of cell lysates revealed that enhanced HNSCC growth conferred by introduction of the novel mutations was associated with increased PI3K pathway activation as reflected by elevated expression of phosphorylated AKT (Figures 2B & C).In the absence of complete functional characterization of these novel mutations, these findings should be considered supportive but not definitive evidence of oncogenic function.
HNSCC patient tumorgrafts with PIK3CA mutations are exquisitely sensitive to BEZ-235
Reports in other cancers suggest that tumors with PI3K pathway activation may be more sensitive to PI3K pathway inhibitors (19).To determine the predictive value of PIK3CA mutational status in HNSCC, we examined the sensitivity of HNSCC cell lines that did and did not harbor intrinsic activating driver PIK3CA(H1047R) hotspot mutations to PI3K pathway inhibitors.As shown in Figure 3A 3E).Another HNSCC patient-derived tumorgraft model (HPVnegative) harboring a PIK3CA mutation (E110K) was also found to be sensitive to BEZ-235 treatment (Supplementary Figure 2).In contrast, patient tumorgrafts with WT PIK3CA and low baseline p-AKT levels were not sensitive to the growth inhibitory effects of BEZ-235 (Figure 3F and Supplementary Figure 3).agent in HNSCC) compared with cetuximab alone (Supplemental Figure 4), suggesting that targeting PI3K in the setting of PIK3CA mutant tumors can enhance treatment responses to cetuximab.
Discussion:
The increase in targeted agents for cancer treatment results in an unprecedented opportunity for personalized cancer medicine.Selection of therapies based on mutation status of molecular targets has transformed clinical management and survival of several human malignancies.The EGFR monoclonal antibody cetuximab is the only targeted therapy that is FDA-approved to date for HNSCC treatment, yet there are no biomarkers that can be assessed in the primary tumor to predict clinical responses to this agent.The recent elucidation of HNSCC genomics offers an opportunity to identify genetic subgroups of HNSCC tumors to guide treatment decisions.
In this report, we employed a bioinformatic approach to identifying mutationally altered, targetable mitogenic pathways in HNSCC.Analyses of all currently available HNSCC wholeexome sequencing data (a total of 151 primary HNSCC tumors) revealed several key findings with important implications for HNSCC pathobiology and treatment.The PI3K pathway is the most frequently mutated oncogenic pathway in HNSCC, with the relative number of PI3Kmutated tumors compared to RAS/MAPK and JAK/STAT-mutated tumors being approximately 3-fold greater.Similar ratios of PI3K pathway mutations (relative to RAS/MAPK or JAK/STAT) are seen in squamous cell carcinoma of the lung and in cervical cancer; both of which share common risk factors with HNSCC, including tobacco and HPV infection, respectively.In contrast, the RAS/MAPK pathway is more frequently mutated than the PI3K pathway in colon and thyroid cancers, and both the PI3K and RAS/MAPK pathways are mutated at comparable rates in lung adenocarcinomas (13).The percentage of HNSCC tumors harboring multiple mutations in the PI3K pathway is similar to that observed in breast cancers (4.9%, 25/507 tumors) and glioblastomas (9.1%, 25/276 tumors) , higher than in thyroid cancer (0.3%, 1/323 tumors) and much lower than most other cancers, including uterine carcinoma (65.7%, 163/248 tumors), melanoma (24.9%, 63/253 tumors), and interestingly lung squamous cell carcinoma (17.4%, 31/178 tumors), which otherwise shares common risk factors and similar relative rates of pathway mutations with HNSCC.(13) Using novel patient-derived tumorgraft models with an oncogenic PIK3CA(E542K) mutation we demonstrated that these tumors are exquisitely sensitivity to a PI3K pathway inhibitor (BEZ-235).Similar results were demonstrated in another HNSCC patient-derived tumorgraft model with a PIK3CA (E110K) mutation, previously reported in breast cancer (21).In contrast, treatment of human-derived heterotopic tumorgrafts with wildtype PIK3CA and low basal expression levels of phospho-AKT, with a PI3K pathway inhibitor, was ineffective.These findings suggest that: 1) PI3K-pathway inhibitors can be effective for treating HNSCC tumors with PI3K mutations; and 2) mutation-guided treatment responses can be evaluated/monitored using patient-derived HNSCC tumorgraft models in vivo.In fact, early-phase clinical trial results showed that patients with solid tumors harboring a PIK3CA hotspot mutation (H1047R) were found to be responsive to PI3K pathway inhibitors (22).However, the effects of other PIK3CA mutations on mediating drug sensitivity in HNSCC preclinical models or clinical trials has not been previously reported.Findings from our study implicate that PIK3CA(E542K) mutation, as well as other non-hotspot mutations (such as E110K) may also identify an HNSCC subgroup potentially responsive to PI3K-pathway inhibitors.In particular, our results using HNSCC patient-derived tumorgrafts suggest that HNSCC tumors with activating PIK3CA mutations may be more sensitive to a dual PI3K/mTOR inhibitor (such as BEZ-235) compared to tumors with wildtype PIK3CA (Figures 3E and Supplementary Figure 3), as indicated by significant inhibition of p-S6 expression in the PIK3CA mutated, but not in the wildtype tumorgrafts.In fact, a recent report of five HNSCC cases found that mTOR-based targeted therapy may be more effective in HNSCC tumors harboring PIK3CA mutation and/or PTEN loss (23).PI3K pathway-mutated HNSCC tumors were found to have a higher rate of non-synonymous mutations, including an increased number of defined cancer gene mutations, compared to tumors without PI3K pathway mutations.This observation implies that the PI3K pathwaymutated HNSCC tumors have "oncogenic" advantage even with genomic instability, and/or that PI3K-mutated HNSCC tumors intrinsically harbor a "mutator" phenotype rendering them more prone to mutation.The oncogenic advantage of PI3K pathway-mutated tumors can be partly explained by PIK3CA "driver" mutations' growth-promoting activity(24) (Figure 2C), while the "mutator" phenotype of these tumors is supported by the our finding that PI3K pathway-mutated tumors are associated with ARID1A and MLL3 mutations, which are important tumor suppressor genes (15,16,25).It is possible that both the "oncogenic/growth" advantage and "mutator" phenotype associated with PI3K pathway-mutated HNSCC tumors are necessary for HNSCC progression; especially since PI3K pathway mutations in these tumors are not associated with TP53 mutation, a previously recognized tumor suppressor alteration that contributes to HNSCC carcinogenesis.Although the relationship of PI3K pathway mutations and TP53 mutation has not been carefully examined in most cancers, a recent study in bladder cancer showed that PIK3CA mutations were significantly more common in TP53 WT tumors (26).Hence, PI3K pathway mutations may mediate tumor progression in the absence of TP53 genetic alteration.
Our finding that all 10 HNSCC tumors with concurrent mutations of multiple PI3K-pathway genes were advanced stage cancers (Stage IV) suggests the potential involvement of concurrent alterations of multiple nodes of the PI3K pathway in HNSCC progression.This agrees with the recent report that in addition to PIK3CA mutation, other pathway components such as PIK3R1 and PIK3R2, when mutated, can also serve to drive cell growth/survival (27).
Although the effects of multiple PI3K pathway mutations on cancer cell growth or progression has not been previously investigated, our results support the possibility that genetic alterations at multiple nodes in this oncogenic pathway, a common feature of many solid tumors, may identify a subgroup of cancer patients most likely to respond to PI3K pathway inhibitors.These cumulative findings identify the PI3K pathway as the most frequently mutated mitogenic pathway in HNSCC tumors.Prospective identification of patients whose tumors harbor these mutations is likely to identify a subgroup of individuals who may benefit from treatment with PI3K pathway inhibitors.
Materials and Methods
Additional methods are detailed as Supplementary Information.
Cancer Gene Census Comparison and Co-mutation Analysis
A mutation comparison program was written in Visual Basic for Microsoft Excel to compare the existence of HNSCC mutations vs. a reference list of mutations of interest (in this case, cancer genes).The program allows side-by-side comparison between multiple groups (2 or more) to find out common mutational events, as well as the number of common events in multiple groups.A cancer gene list was generated in each subgroup of tumors by comparing the Cancer Gene Census list (COSMIC Database) with non-synoynomous mutation gene list of each tumor subgroup (the PI3K-mutated tumors, tumors without PI3K-mutation, PIK3CA-mutated tumors, PIK3CA WT tumors) using this comparison program.This analysis allows us to find out the number of cancer genes mutated in each subgroup.
Mutation Validation by Sanger Sequencing
Sanger sequencing was performed on patient tumors that were grafted for tumorgraft studies.
2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103 , HNSCC cell lines containing endogenous PIK3CA(H1047R) mutations (CAL-33 and Detroit 562)(20) demonstrated increased sensitivity to PI3K pathway inhibition by the mTOR/PI3K inhibitor BEZ-235 compared to representative HNSCC cells with WT PIK3CA [SCC-9 and PE/CA-PJ34(clone C12)].Next, mice bearing CAL-33 xenografts were found to be sensitive to BEZ-235 treatment in vivo when compared to vehicle control (Figure 3B).Due to the lack of HPV-HNSCC cell line models that contain PIK3CA mutations, we developed an HPV-positive PIK3CA-mutated HNSCC patient tumorgraft model (E542K) (Figure 3C) to determine the sensitivity of HPV-positive PIK3CA-mutated HNSCC tumors to PI3K pathway targeting.As shown in Figure 3D, BEZ-235 treatment (at 25mg/kg/day by oral gavage) significantly inhibited the growth of HPV-positive PIK3CA-mutated patient tumorgraft in vivo (P<0.0001).Inhibition of tumor growth was accompanied by decreased PI3K signaling as evident by down regulation of p-AKT(S473) (P=0.0124), and p-S6(S235/236) (P<0.0001) in the BEZ-235-treated tumors (Figure These results indicate that activating mutations of PI3K pathway have the potential to serve as biomarkers for treatment selection in HNSCC.Xenografts developed from a HNSCC cell line harboring a PIK3CA mutation (H1047R) were more sensitive to the combination of BEZ-235 plus cetuximab (the only FDA-approved molecular targeting Research.on July 4, 2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103 2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103 2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103 2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103 2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103
About 25 -
50mg of tumor tissues (pathologically confirmed HNSCC with >70% tumor cell contents) were used for extraction of DNA by QIAamp DNA Mini Kit (Qiagen, Inc, Valencia, CA).Sequencing primers for HNSCC-associated PIK3CA hotspot mutations were synthesized (Sigma-Aldrich, St. Louis, MO) and used for Sanger sequencing.The primer sequences for E542 site mutation are: 5'-cacgagatcctctctctaaaatcactgagcaggag-3' (forward) and 5'ctcctgctcagtgattttagagagaggatctcgtg-3' (reverse).Sanger sequencing was performed at the Genomics and Proteomics Core Laboratories at the University of Pittsburgh.HNSCC Tumorgraft Model and Drug Treatment BEZ-235 was obtained as a kind gift from Novartis, USA.HPV-positive HNSCC patient tumorgrafts were derived under the auspices of an IRB-approved protocol, with PIK3CA WT or PIK3CA(E542K) mutation were implanted into the flanks of NOD SCIDγ mice and treatment was started when tumors became palpable.BEZ-235 (25mg/kg) or vehicle control was given daily by oral gavage.Tumor volumes were measured every two days.Research.on July 4, 2017.© 2013 American Association for Cancer cancerdiscovery.aacrjournals.orgDownloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103
Figure 1 .
Figure 1.Mutations in oncogenic signaling pathways in HNSCC.(A) Mutation rates of the major mitogenic pathways (the PI3K pathway, the MAPK pathway and the JAK/STAT pathway) in 151 HNSCC patient tumors determined by whole exome sequencing.Components of each pathway examined are displayed underneath each pie chart.(B) Bar graph detailing the number of mutations (dark bars) of each particular component of the PI3K pathway as well as the number of HNSCC tumors harboring these mutations (grey bars).(C) PI3K pathway-mutated HNSCC tumors have higher of non-synonymous mutations and (D) cancer gene mutations when compared to HNSCC tumors without any PI3K pathway mutations.Bar graph representing the average number of non-synonymous mutations per tumor (C) and the average number of cancer gene mutations (D) in 151 HNSCC tumors.Statistical significance was calculated by Fisher's Exact test, P<0.0001 (N=151).(E) Graphical representation of the number of HNSCC tumors with mutation of multiple components of the PI3K, MAPK and the JAK/STAT pathways, respectively.
Figure 2 .
Figure 2. PIK3CA mutations in HNSCC tumors.(A) Schematic of all PIK3CA mutations found in 151 HNSCC tumors by whole exome sequencing.The amino acid (a.a.) positions of each domain is shown in grey below each domain.The number of mutational events at each site is indicated by a filled triangle (▲) in the graph above.Blue triangles indicate mutations found in HPV/HNSCC tumors.Keys: ABD: p85 binding domain; RBD: Ras binding domain; C2 Superfamily; Helical: PIK domain; Kinase: Kinase domain of PIK3CA.(B) Effects of PIK3CA mutations on PI3K signaling in HNSCC cells.WT PIK3CA, hotspot mutant H1047R, and novel mutants: R115L, G363A, C971R, and R975S were stably expressed in an HNSCC cell line harboring no endogenous mutations in the PI3K pathway, PE/CA-PJ34 (clone C12)
Table 1 : HNSCC Tumors with Multiple Mutations in a Single Mitogenic Pathway. Table 1: Describes all tumors in our cohort harboring mutations in more than one gene in a defined mitogenic pathway by mutation type and pathological stage.
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.Author Manuscript Published OnlineFirst on April 25, 2013; DOI: 10.1158/2159-8290.CD-13-0103 | 5,556.6 | 2013-07-01T00:00:00.000 | [
"Biology",
"Medicine"
] |
Ultra-wideband filtering of spoof surface plasmon polaritons using deep subwavelength planar structures
Novel ultra-wideband filtering of spoof surface plasmon polaritons (SPPs) is proposed in the microwave frequency using deep subwavelength planar structures printed on thin and flexible dielectric substrate. The proposed planar SPPs waveguide is composed of two mirror-oriented metallic corrugated strips, which are further decorated with parallel-arranged slots in the main corrugated strips. This compound structure provides deep subwavelength field confinement as well as flexible parameters when employed as a plasmonic waveguide, which is potential to construct miniaturization. Using momentum and impedance matching technology, we achieve a smooth conversion between the proposed SPPs waveguide and the conventional transmission line. To verify the validity of the design, we fabricate a spoof SPPs filter, and the measured results illustrate excellent performance, in which the reflection coefficient is less than −10 dB within the −3 dB passband from 1.21 GHz to 7.21 GHz with the smallest insertion loss of 1.23 dB at 2.21 GHz, having very good agreements with numerical simulations. The ultra-wideband filter with low insertion loss and high transmission efficiency possesses great potential in modern communication systems.
In this paper, we firstly propose a novel structure for spoof SPPs at microwave frequencies. Based on the structure, we design and fabricate an ultra-wideband filter with low reflection and high transmission coefficient for SPPs waves. The spoof SPPs waveguide is composed of mirror-oriented corrugated metallic strips, in which compound slot geometry is further designed. The compound slot structure is composed of mother slots and son slots, where the son slots are parallel and symmetrically arranged on the two sides of the mother slots, as can be referred to the inset of Fig. 1. Using the presented artificial plasmonic waveguide, we can confine the microwave energy tightly with little propagation loss. Also, in order to reach a perfect momentum matching between the spoof SPPs waveguide and the signal input port, where a traditional co-planar waveguide (CPW) working in quasi TEM mode is employed, a transition section with gradient slots and flaring ground is designed for high-efficiency conversion. Numerical simulations and experimental results show that the presented plasmonic waveguide owns high efficiency filtering of spoof SPPs in ultra-wide frequency band, which builds a solid avenue for large-scale plasmonic integrated circuits in microwave and terahertz devices.
Results
The designed compound slots structure for plasmonic waveguide. The corrugated metal with compound slots structure is printed on a 0.5 mm thick dielectric substrate F4B with dielectric constant ε r = 2.65 and loss tangent tanδ = 0.003 at microwave frequencies. The metal is chosen as annealed copper with film thickness of 0.018 mm. The thickness of dielectric substrate and copper film correspond to 167/10000 and 6/10000 of the EM wavelength in free space at 10 GHz respectively. Figure 1 depicts the dispersion and E-filed distribution characteristics of the proposed artificial plasmonic structure. Figure 1(a) shows the corrugated metallic unit cell is composed of two sets of periodic slots, where the depth, width and period of the main slot (here we call it mother slot) are denoted as hm, b, and p, respectively. On the two sides of the mother slots, there distributes parallel, periodical and mirror symmetric son slots. The son slots are designed to enhance the equivalent capacitance and inductance of the mother slot. The geometrical parameters of son slots are denoted as periodic ps, depth hs and width ds, respectively, in which, the periodic ps varies with the son slot number (denoted as ssn) so as to guarantee the even distribution of son slots in the mother slot. According to R. F. Harrington 18 , for one dimensional groove array Figure 1. (a) The schematic unit cell of the spoof SPPs waveguide, which is composed of compound slot structure (mother slot and son slot) and in this sample the son slot number is 12. (b) The schematic of one dimensional groove array with the depth of hm, the width of b and the lattice constant of p on a perfect metal surface with an infinite thickness along the z direction. (c) The dispersion diagrams of the spoof SPPs structure with different son slot depths hs (with the son slot number being 20, the son slot width ds being 0.1 mm, the son slot period ps being 0.35 mm, the mother slot depth hm being 4.5 mm, the mother slot width b being 2 mm and the mother slot period p being 5 mm). (d) The dispersion diagrams of the spoof SPPs structure with different son slot numbers ssn (with the son slot depth hs being 1.0 mm, ds being 0.1 mm and the parameters of the mother slot being the same as that in (c)). (e) Represents the EM confinement (simulated) in the H-shaped mother slot (hs = 0, left) and in the compound slot with hs = 1.0 and ssn = 20 (right).
Scientific RepoRts | 6:37605 | DOI: 10.1038/srep37605 with the geometry parameters of the above mother slot, but being infinitely thick in z direction, as illustrated in Fig. 1(b), when the incident EM wave is p-polarized, the wave impedance looking into the corrugated surface in the negative y direction is: in which, k x is the wave number in the x direction of the surface of F4B substrate and k 0 is the wave umber in free space. On the other side, the parallel-plate transmission-line mode exists in the short-circuited slots of the metal corrugation, thus, the input wave impedance from the negative y direction can be described as ref. 19: where, η µ ε = / 0 0 0 is the intrinsic impedance of free space. When EM wave resonates at the interface of metal corrugate, we can deduce Eq. (4) by equating Eqs (1)~(3) as following: From Eq. (4), one can clearly infer that when the slot width b increases, the wave number k x will be enhanced as well. Inspired by this, we design the corrugated son slots on the two sides of mother slots with finite film thickness to expect the analogous physical effects. Meanwhile, for symmetrical transmission, the above groove array has been designed mirror-oriented along it propagation direction. It is anticipated that the increase of the son slot depth hs will augment the wave number k x as well as the input wave impedance Z −y of the mother slots (refer to Eq. (3)) in the present model. To verify the above designing theory, all the numerical simulations of the compound slots have been carried out by the commercial software, CST Microwave Studio. Results show that the dispersion curve deviates more quickly and the equivalent cutoff plasmonic frequency reduces with the increase of son slot depth hs from 0 to 1.0 mm (whilst the son slot width ds is fixed at 0.1 mm during the numerical simulation), as illustrated in Fig. 1(c), where hs = 0 corresponds to the dispersion curve of the previously proposed metallic corrugated strips by Ma et al. 20 . These results correspond quite well with the above designing theory and this unique feature much resemble the behavior that originally happen in optical frequencies of the natural SPPs, which will result in a smaller propagating wavelength and a tighter electric field confinement of the EM wave at gigahertz frequency, ensuring reliable applications in miniaturized plasmonic devices and low-interference SPPs circuits 21 . Also, as the son slot number increases from 0 to 30 (with the son slot depth hs being 1.0 mm), we find a similar dispersion curve with lower asymptotical cutoff frequencies as that in the increase of son slot depth, as depicted in Fig. 1(d). These dispersion curves with lower cutoff frequencies will lead to larger propagating vectors in the plasmonic waveguides, which indicates the proposed spoof SPPs provides more tailoring parameters than the previously reported corrugated metallic structure 20 and allows easier tuning of the wave momentum and propagating vector in plasmonic circuits. Also, the field confinement effect is much more significant in compound slots structure than that in ref. 20 as indicated by Fig. 1(e). In this regard, the proposed structure allows more flexible and convenient application in plasmonic circuit industry without increasing the size or cost of devices.
Plasmonic wide bandwidth filter with transition section. By employing the proposed spoof SPPs waveguide, we designed an ultra wideband microwave filter, whose structure is formed by three sections printed on the top surface of the substrate, whose lengths are l 1 = 8 mm, l 2 = 80 mm and l 3 = 58 mm respectively, as depicted in Fig. 2. In the input and output sections, conventional CPWs are employed to feed microwave signal or to receive the transmitted signal. The middle section of the filter is composed of the proposed mirrorly oriented corrugated metallic strip with the compound slot structure. It has been analyzed above that the compound slot structure is a good plasmonic waveguide with tight EM field confinement and small transmission loss. However, this plasmonic waveguide cannot be directly integrated with the first CPW section because the quasi TEM mode in the latter is mismatched in wave momentum with the spoof SPPs mode in the former.
Previously reported methods such as dipole antenna method 21,22 , prism method 23 and gratings method 24,25 have been put forward to connect SPPs modes with quasi TEM modes. However, the relatively low conversion efficiency due to momentum mismatch between them have prevented these methods from industrial applications. Later, a high efficiency transition waveguide between the quasi TEM transmission line and the spoof SPPs line was proposed in ref. 26, which is composed of a linearly gradient corrugated metallic strip and an exponential flaring curve ground. The mixed modes of quasi-TEM and spoof SPPs are simultaneously supported on the transition section, where quasi-TEM mode dominates in the beginning and is gradually converted to the designed plasmonic mode as the groove depth increases step by step until it turns into the pure spoof SPPs mode at last. Although the transition efficiency is significantly increased through this design, it can still be improved through modifying the flaring ground curve, which is actually a crucial point to provide gradient momentum compensation and impedance matching between quasi-TEM modes and plasmonic TM modes. Actually, we have modified the flaring ground curve using several mathematical functions in the present work as will be discussed later in this paper, and we found a better transition effect than that given in ref. 27.
After designing the structure of the particular filter, we simulated the scattering parameters (S 11 and S 21 ) from 0 to 15 GHz and the results are illustrated in Fig. 3. The geometrical parameters for the designed filter are as follows: the width of the ground w = 25.0 mm, the gap in CPW section g = 0.4 mm and the width of the transmission It is observed that the spoof SPPs waveguide is capable of EM wave transmission from almost DC frequency to the cutoff gigahertz frequency with high transmission and low reflection, as illustrated in Fig. 3. The upper stop band of the spoof SPPs filter can be artificially adjusted by the geometrical dimensions of the son slot, such as the son slot number and the son slot depth. As can be seen, with the increase of son slot number or son slot depth, the upper stop band frequency decreases and the stop band becomes steeper, which means the filter possesses better out-of-band properties. Especially, the simulated upper stop band frequency decreases almost linearly from 9.03 GHz to 7.25 GHz with the increase of son slot depth from 0 to 1.0 mm, which provides a convenient and precise control of the stop frequency for the filter. Note that these adjustment is very flexible and without any increase of filter size or cost.
At lower stop band frequency, the S-parameter curves are rarely adjusted by the geometrical parameters of the son slot. In the whole 3 dB passband the reflection is less than − 10 dB and the transmission curve is flat with ripples of less than 1.5 dB. However, near the upper and lower stop frequencies, the reflection ripple becomes higher, resulted from the momentum mismatch at these frequency ranges. The mismatch can actually be improved by tailoring the shape and length of the transition section. In terms of this, we deduce that the transition section is a key part to optimize the operation properties of the spoof SPPs filter. To more accurately and quantitatively evaluate the transmission properties of the presented spoof SPPs structure, we calculate the transmission loss in the spoof SPPs mode section, where the influence of the transition sections at both ends are extracted. Result shows that the transmission loss is only 0.02 dB/cm at 7 GHz and it decreases along with the frequency. For comparison, we have also calculated the transmission loss in a conventional microstrip line with the same length and strip width as the spoof SPPs waveguide in Fig. 2. It is found that although the propagation loss of the presented spoof SPPs structure is twice larger that of conventional microstrip, which is about 0.01 dB/cm at 7 GHz, the former provides an approach for the convenience of integrated circuits with its single-side conductor feature and low crosstalk property. In addition, the simulated relative bandwidth of the filter with son slot number being 20 and son slot depth being 1.0 mm can reach 117%, manifesting itself an ultra-wideband filter that can be utilized in wideband high speed data communication.
The influence of the flaring ground curve. As has been mentioned above, it is worthy to note that the flaring ground curve in the transition section is very important for high efficient transmission and minimum reflection loss of the spoof SPPs filter. Thus, we have designed several kinds of flaring ground using different mathematical functions, including circle function, parabola function and exponential function for comparison. Results show that the exponential one is the optimal function for highly efficient EM modes conversion. Also, the shape of the exponential curve is critical for the highly efficient signal transmission. Here, we define the exponential curve as: and where n is the shape parameter that can control the flaring speed of the exponential curve. As can be seen in Fig. 4, with the variation of n (n ≥ 1), the reflection and transmission characteristics of the filter can be obviously improved, especially at the frequency close to the lower stop band of the S 11 curve, which means that the exponential curve shape contributes greatly to the momentum conversion in the pass band of the filter. However, as the parameter n increases to 4, the reflection energy increases again, particularly at the frequency close to the upper stop band of the S 11 curve. Therefore, n = 3 is the optimal shape curve parameter for the presented transition section. Fabrication and measurement of the spoof SPPs filter. In order to verify the design and simulation, we fabricate the above spoof SPPs filter by traditional PCB printing method, as illustrated in Fig. 5. The production geometry parameters are the same as the optimal parameters in simulation. Two samples with the exponential curve parameter n = 1 and n = 3 were synthesized for comparison. The S-parameter of the synthesized filter was measured by Agilent vector network analyzer (VNA, N5230C). The measured and simulated results of the S-parameters, including the reflection coefficients S 11 and transmission coefficients S 21 , are illustrated in Fig. 6. It is obvious that there is a good agreement between the measured and simulated S-parameters, especially for the two S 21 curves. Moreover, the measured S 21 parameter is even better than the simulated one at lower frequencies.
The measured S 11 parameter, however, is worse than the simulated one, which can be ascribed to the impedance mismatch at two welding end points. The transmission coefficient S 21 indicates good frequency-selective property of the proposed structure with the transmission loss being low and companied by a transmission zero at 0.21 GHz, which helps to suppress the low-frequency interference. The reflection coefficient S 11 of the n = 3 sample is much better than that of the n = 1 sample, which is in good agreement with the simulated result in Fig. 4.
In the whole pass band from 1.21 GHz to 7.21 GHz, the S 11 parameter of the n = 3 sample is less than − 11.2 dB, which manifests the good impedance and momentum matching behavior from the CPW waveguide to the SPPs waveguide through the presented exponential curve ground.
Field confinement effects.
To get a direct physical insight into the mode matching transition, as well as the properties of field propagation and confinement on the spoof SPPs waveguide with compound slot structure, we performed full-wave simulations using commercial CST Microwave Studio. Figure 7 demonstrates the energy flows (on a dB scale) toward x direction on the xoy plane that is 1.5 mm above the plasmonic surface of the waveguide. We monitored and measured four different frequencies at 1 GHz, 3 GHz, 7 GHz and 10 GHz for the sample of the son slot number being 20 and the exponential shape coefficient n being 1. It is also verified that the measured and simulated EM fields correspond quite well with each other, both in magnitude and distribution size. It is obviously evidenced the quasi TEM mode in CPW waveguide is smoothly converted to the SSPPs mode with low reflection. The EM energy is tightly confined in deep subwavelength scale around the plasmonic waveguide and it propagates with little reflection and low absorption and radiation loss in the whole pass band from 1.21 GHz to 7.21 GHz. Moreover, the significant Ex component is detected on the plasmonic waveguide due to the transverse magnetic behavior of its eigenmodes.
Discussion
In the present paper, we have proposed a compact spoof SPPs structure with compound corrugated metallic slots to produce frequency selective microwave filter. The compound slots structure is composed of the so called mother slot and son slot, orthogonally arranged to each other, to obtain better spoof SPPs properties. Besides the spoof SPPs waveguide using the compound slot structure, we also employ a transition waveguide to smoothly covert the quasi TEM mode to spoof SPPs mode, so as to feed and receive electromagnetic signal with low energy reflection. Plentiful numerical simulations and experiments have been employed to validate the designing theory and the transmission properties of the proposed filter. And both the simulated and measured S-parameters demonstrate that the proposed spoof SPPs structure has high efficiency transmissions (with transmission loss being only 0.02 dB/cm at 7 GHz) and low reflections loss (measured below − 11.2 dB) in the designed frequency band from 1.21 GHz to 7.21 GHz. The simulated and measured near field distributions indicate that the EM energy is tightly confined in deep subwavelength scale around the spoof SPPs waveguide. Such unique performance endows the proposed spoof SPPs band-pass filter a very encouraging future in high compact microwave or even terahertz wave integrated circuits and plasmonic functional devices.
Methods
All numerical simulations including SPPs filters and E-field distributions are conducted by the commercial software, CST Microwave Studio. The experimental structure is fabricated using a 0.5 mm thickness dielectric film, F4B, which is a kind of Teflon woven, composed of polytetra-fluoroethylene and glass fiber with its relative permittivity being 2.65 and tangent loss being 0.003 at microwave frequency. The patterned copper conductor film is printed onto the F4B substrate for the SPPs filtering with its thickness of 0.018 mm. We employ Agilent vector network analyzer (VNA, N5230C) to measure the S parameters of the SPPs filters, including the reflection coefficients S 11 and transmission coefficients S 21 . The near E-field distributions along the z-direction of the SPPs filter are examined by a home-made near-field scanning system, where the testing antenna probe linearly scans in the xoy plane, 1.5 mm above the surface of the fabricated SPPs filter. | 4,649.4 | 2016-11-24T00:00:00.000 | [
"Physics"
] |
In vivo microsampling to capture the elusive exposome
Loss and/or degradation of small molecules during sampling, sample transportation and storage can adversely impact biological interpretation of metabolomics data. In this study, we performed in vivo sampling using solid-phase microextraction (SPME) in combination with non-targeted liquid chromatography and high-resolution tandem mass spectrometry (LC-MS/MS) to capture the fish tissue exposome using molecular networking analysis, and the results were contrasted with molecular differences obtained with ex vivo SPME sampling. Based on 494 MS/MS spectra comparisons, we demonstrated that in vivo SPME sampling provided better extraction and stabilization of highly reactive molecules, such as 1-oleoyl-sn-glycero-3-phosphocholine and 1-palmitoleoyl-glycero-3-phosphocholine, from fish tissue samples. This sampling approach, that minimizes sample handling and preparation, offers the opportunity to perform longitudinal monitoring of the exposome in biological systems and improve the reliability of exposure-measurement in exposome-wide association studies.
Materials and Methods
Study design. Adult white sucker (Catostomus commersonii) (40.8 ± 3.6 cm, 969.7 ± 303.3 g, n = 60) were collected by boat electrofishing from the Athabasca River in the Alberta oil sands region (Northern Alberta, Canada). As part of a larger sampling effort, twelve fish were collected in September 2013 from: 2 sites outside of the deposit, M0 (Athabasca) and M1, which are both downstream of a pulp and paper mill discharge; 1 site upstream of the oil sands development but within the deposit around Northlands Sawmill (Downstream of M3); 1 site adjacent to the oil sands development (Upstream of M4); and 1 site downstream of the Muskeg River within the deposit and downstream of the development (Downstream M4) (Fig. 1). In total, a subset of 6 males and 6 females were selected at each site and held briefly (< 1 h) in cages in the river until sampling. SPME blade coating preparation. PAN-C18 SPME blades were prepared as previously described 12 by immobilization of particles on the surface of stainless steel blades. Briefly, 5 μ m particles (Supelco, PA) were immobilized (60 μ m coating thickness) using a polyacrylonitrile (PAN) solution which acted as a bonding agent.
In vivo and ex vivo SPME sampling procedure of fish tissue. All experimental protocols were in accordance with and as approved by the University of Waterloo Animal Care Committee (AUPP #10-17). In vivo extraction of metabolites from fish tissue was conducted by inserting a PAN-C18 SPME coated blade into the dorsal-epaxial muscle (near the dorsal fin) of fish immobilized using a large foam bed 13,14 . The blade remained in place for 20 min while fish were held in an aerated, 28 L covered bucket. After 20 min, the blade was removed, rinsed with nanopure water to remove matrix components, and frozen in liquid nitrogen. Fish were then sacrificed and a small part of the dorsal-epaxial muscle was cut out, placed in aluminum foil and frozen in liquid nitrogen on-site and shipped to our laboratory. In the laboratory, ex-vivo SPME sampling was performed by inserting a PAN-C18 coated blade into a thawed, non-homogenized tissue sample for 20 min without agitation. Desorption of metabolites from the SPME coating of the blades was done by immersing them for 60 min in 1 mL of acetonitrile/water (80/20, v/v) with vortex agitation at 1000 rpm. Extracts were stored at − 80 °C until analysis. UPLC-Q-Exactive Orbitrap HRMS analysis. Metabolite profiling was conducted using an LC-MS system consisting of a ThermoAccela autosampler, pumps and a Q-Exactive Orbitrap System (Thermo Fisher Scientific, CA, USA). Metabolites were separated by a reversed-phase method using a pentafluorophenyl column (Kinetex Phenomenex, 2.1 mm × 100 mm, 1.7 μ m particle size) at a flow rate of 300 μ L/min. Mobile phase A consisted of water/formic acid (99.9/0.1, v/v) and mobile phase B consisted of acetonitrile/formic acid (99.9/0.1, v/v). The starting mobile phase conditions were 90% A from 0 to 1.0 min, followed by a linear gradient to 10% A from 1.0 to 9.0 min and an isocratic hold at 10% A until 12.0 min. The total run time was 18 min per sample, including a 6 min re-equilibration time. The injection volume was 10 μ L. Analyses were performed in positive ionization mode in the mass range of m/z 50-750. To maintain a mass accuracy better than 5 ppm, we used the following lock mass: m/z 391.2843. Instrument parameters were set as follows: sheath gas (Nitrogen) flow rate, 35 arbitrary units; capillary voltage, 3.1 kV; ion source temperature 280 °C; full MS automatic gain control (AGC), 1.10 6 ; spectra rate acquisition, 3.7 spectra/s; full MS resolution, 70,000. MS/MS fragmentation of the ten most intense ions per spectrum was performed using a normalized collision energy (NCE) of 50; MS/MS resolution, 17,500; MS/MS AGC, 2.10 5 ; precursor ion mass isolation window, 1 ppm. MS/MS exclusion list was set after 3 analyses of blank SPME samples.
UPLC-MS/MS data processing and molecular networking. Molecular networking of LC-MS/MS
data was performed using the Global Natural Products Social Molecular Networking (GNPS) software 15 . For MS and MS/MS spectral library search and molecular networking, we used an ion mass tolerance of 1 and 0.01 Da for precursor ions and fragment ions, respectively. A minimum cosine similarity score of 0.7 was used for MS/ MS spectral library matching. A minimum cosine similarity score of 0.7 and a minimum number of 6 matched fragment ions were used to form a network of two consensus MS/MS spectra. Resulting molecular networks were built and visualized using Cytoscape 3.2.1 16 . Precursor ion m/z was used as node attribute and cosine similarity score was used as edge attribute. . Green and blue nodes represent small molecules observed only after in vivo SPME sampling and ex vivo SPME sampling, respectively. Grey nodes represent small molecules detected using both sampling methods. Nodes with red border indicates annotated molecule by matching LC-MS/MS libraries with spectral similarity ≥ 0.7 Edge represents similarity between MS/MS spectra. Thickness of the edges indicates the level of similarity (the thicker is an edge, the more similar are MS/MS spectra).
Results and Discussion
After molecular networking of fish tissue LC-MS/MS data, we found 494 nodes representing consensus of at least two or more MS/MS spectra (Fig. 2). The majority of the nodes matched chemicals detected in fish tissue using both sampling approaches. Only 7% of them could be identified by matching LC-MS/MS libraries with spectral similarity ≥ 0.7 ( Table 1), suggesting that the vast majority of chemicals in fish tissue are unknown. However, mapping the chemical similarity between unknown and identified molecules can help uncover chemical classes of the uncharacterized molecules. Although the majority of identified chemicals were endogenous molecules, four molecules including 4-methoxycinnamic acid, 1-hydroxybenzotriazole, diethylphthalate, and phenoxybenzamine used in the formulation of sunscreen products 17 dishwasher cleaning products 18 and plasticizers 19 or as active ingredients in pharmaceutical products 20 respectively, were detected in fish tissues. These chemicals probably originated from external exposures (i.e. water contamination), confirming that fish tissue offers opportunity to reconstruct past environmental exposures due to bioaccumulation of persistent organic toxicants, and can be used to monitor aquatic ecosystem health.
We then evaluated the ability of in vivo SPME to capture unstable molecules. We observed that 16% of the nodes were only detected after in vivo SPME sampling, while 21% only found after ex vivo SPME sampling. Molecules only observed with in vivo SPME include cinnamic acids, glycerophosphocholines, phenylmethylamines, phenoxybenzamines and pyridincarboxylic acids originating from both exogenous and endogenous exposures (Table 1). For example, 1-oleoyl-sn-glycero-3-phosphocholine is a lipid-signaling molecule generated by phospholipase enzymes 21 . This compound contains an unsaturated acyl chain where the hydrogen atom on methylene groups adjacent to the double bounds has low carbon-hydrogen (C-H) bond energies, and is therefore a major target for modification under oxidative conditions after sample collection 22 . Five other glycerophosphocholines ( Fig. 1B; m/z 542.322, m/z 524.376, m/z 568.342, m/z 544.340, and m/z 510.359) were only detected using in vivo SPME, but was not successfully identified due to the lack of adequate similarity (cosine ≥ 0.7) between their MS/MS spectra and those from public LC-MS/MS libraries. This result possibly indicates that these molecules have not been previously identified due to their chemical instability or experimental MS/MS spectra produced were a composite of two or more molecules. Similarly, cinnamic acids, phenylmethylamines, phenoxybenzamines and pyridincarboxylic acids only detected after in-vivo SPME are also highly reactive compounds prone to auto-oxidation during sample transportation and storage. However, glycerophosphocholines containing saturated acyl chains, such as 1-palmitoyl-sn-glycero-3-phosphocholine, which was detected after both in vivo and ex vivo SPME, are more resistant to oxidation 22 . Some molecules were only observed after ex-vivo SPME, including 6-hydroxynicotinate, hydroxyproline and 2-heptyl-3-hydroxy 4-quinolone. Previous studies have demonstrated that 6-hydroxynicotinate and 2-heptyl-3-hydroxy 4-quinolone are essentially produced from bacterial metabolism 23,24 and possibly originated from tissue degradation during sample transportation and storage. Hydroxyproline, a major constituent of collagen, is also used as an indicator of tissue damage or degradation 25 .
Conclusion
In summary, we demonstrated that in vivo SPME sampling provided extraction and stabilization of highly reactive chemicals, not detected by ex vivo sample preparation techniques due to their auto-oxidation during sample collection, transportation and storage. In vivo SPME sampling provides extraction of molecules with a wide range of chemical and physical properties 9 (balanced coverage), which originate from exogenous sources via environmental exposures or from endogenous processes including host and microbial metabolism. In vivo SPME sampling as a minimally invasive technique also offers the opportunity to perform repeated sampling over time on the same subject, limiting inter and intra-individual variability in levels of circulating molecules arising from changes in environmental factors (e.g. diet or sources of pollutants). | 2,231.6 | 2017-03-07T00:00:00.000 | [
"Chemistry",
"Environmental Science"
] |
Effect of Annealing Process on the Properties of Ni(55%)Cr(40%)Si(5%) Thin-Film Resistors
Resistors in integrated circuits (ICs) are implemented using diffused methods fabricated in the base and emitter regions of bipolar transistor or in source/drain regions of CMOS. Deposition of thin films on the wafer surface is another choice to fabricate the thin-film resistors in ICs’ applications. In this study, Ni(55%)Cr(40%)Si(5%) (abbreviated as NiCrSi) in wt % was used as the target and the sputtering method was used to deposit the thin-film resistors on Al2O3 substrates. NiCrSi thin-film resistors with different thicknesses of 30.8 nm~334.7 nm were obtained by controlling deposition time. After deposition, the thin-film resistors were annealed at 400 °C under different durations in N2 atmosphere using the rapid thermal annealing (RTA) process. The sheet resistance of NiCrSi thin-film resistors was measured using the four-point-probe method from 25 °C to 125 °C, then the temperature coefficient of resistance could be obtained. We aim to show that resistivity of NiCrSi thin-film resistors decreased with increasing deposition time (thickness) and the annealing process had apparent effect on the sheet resistance and temperature coefficient of resistance. We also aim to show that the annealed NiCrSi thin-film resistors had a low temperature coefficient of resistance (TCR) between 0 ppm/°C and +50 ppm/°C.
Introduction
BiCMOS can provide superior performance for the applications in different circuits. Thus, the analog functions in BiCMOS require passive circuit components having small temperature coefficients, and the process sequence must be able to accommodate their fabrication. Resistors in integrated circuits (ICs) are implemented using diffused methods fabricated in the base and emitter regions of bipolar transistor or in source/drain regions of CMOS [1,2]. To deposit thin films on the wafer surface is another way to fabricate the passive resistors. Thin-film resistors are used extensively in electronic circuits due to their high accuracy and excellent long term stability. So far, there are various specific materials currently used as thin-film resistors in ICs' applications, such as Cr-Si-Ta-Al [3], Ti-Si-V-N [4], and CuAlMo [5], respectively. One of the most widely used thin-film resistors is nickel-hromium (NiCr) which has a low temperature coefficient of resistance (abbreviated as TCR) between´50 ppm/˝C and +50 ppm/˝C [6,7] and has a wide sheet resistance range of 10 to 500 Ω/square [8]. Ni-Cr thin-film-based resistors are extensively used as discrete loads or potentiometers in hybrid circuits. Excellent wear and corrosion resistance makes Ni-Cr thin films an attractive material for fusible links in programmable read only memories [6]. Recently, there has been an increase in demand for thin-film resistors with lower value of 0.1~10 Ω, especially in portable electronic devices for the purpose of saving battery power [9]. This requirement is difficult to meet with the properties of NiCr-based thin-film resistors, because as the thickness of NiCr-based thin-film resistors increases, the cost of deposition process increases and problems at the subsequent laser trimming stage will occur, because they become difficult to ablate.
There are different materials currently used as thin-film resistors in IC applications and Cr-Si thin films are very interesting materials. The Cr-Si thin films can be deposited as the thin-film-resistors because of existing certain advantages, including high sheet resistance, low TCR, high thermal stability, good long-term reliability, and chemical stability, respectively [10]. Dong et al. previous investigation proved that Cr-Si (Cr:Si = 1:3) thin films doping with 3-6 at % Ni had the resistivity of 1.31-1.49 times higher than that of Cr-Si thin films without Ni [11,12]. They also found that the TCR of thin films could be adjusted to close zero by using annealing process. However, the resistivity of Cr-Si-Ni films was decreased when Ni content was higher than 6 at %. The results suggested that a moderate Ni addition, for example, equal and higher than 40 wt % of Ni in Cr films [7] or equal and higher than 6 at % of Ni in Cr-Si films was favorable to the electrical stability [11]. In this study, Ni(55%)Cr(40%)Si(5%) (abbreviated as NiCrSi) in wt % was used as the target because the NiCrSi thin films had the merit of long-term stability [13] and the sputtering method was used to deposit the thin-film resistors on glass substrates. After deposition, the thin-film resistors were also annealed at 400˝C under different duration using the rapid thermal annealing (RTA) process. The sheet resistance of thin-film resistors was measured using the four-point-probe method. We would show that resistivity of NiCrSi thin-film resistors decreased with increasing deposition time and the annealing process had no apparent effect on the value of sheet resistance and the TCR.
Experimental Section
The structure of deposited NiCrSi-based thin-films using for the measurements of physical and electrical properties was shown in Figure 1. Figure 1a shows the side view of the structure. The green glass paste on Al 2 O 3 ceramic shown in Figure 1b was used to protect the substrate, and it could be removed during the NiCrSi thin-film resistors' annealing process and following ultrasonic clean in deionized water. The length between two electrodes (or called the length of thin-film resistors) was 4.0 mm, the widths of thin-film resistors and electrodes were 2.8 mm, and the length of electrodes was 1.2 mm, respectively, as Figure 1a,b show. Commercial composition material Ni(55%)Cr(40%)Si(5%) (abbreviated as NiCrSi) in wt % was used as the target and the sputtering method was used to deposit NiCrSi thin films on Al 2 O 3 substrates. The deposition parameters were power of 150 W at 5 mTorr and at room temperature (25˝C), and the deposition time was changed from 30 min to 150 min, respectively. After deposition, the deposited NiCrSi thin films were annealed by using the rapid thermal annealing (RTA) at 400˝C in N 2 under different durations of 30 s~10 min. After NiCrSi-based thin-films were annealed, the glaze layer was removed by using ultrasonic method in deionized water. The surface morphology of deposited NiCrSi thin-film resistors was shown in Figure 1c. The crystalline structures of deposited and annealed NiCrSi-based thin-films were determined by means of X-ray diffraction (XRD) (Cu-Kα, Bruker D8, Billerica, MA, USA). The thickness measurement of deposited and annealed Ni-Cr-Si-based thin-films was obtained by using α-step (Kosaka ET 4000, Tokyo, Japan). rystalline structures of deposited and annealed NiCrSi-based thin-films were determined by means of -ray diffraction (XRD) (Cu-Kα, Bruker D8, Billerica, MA, USA). The thickness measurement of eposited and annealed Ni-Cr-Si-based thin-films was obtained by using α-step (Kosaka ET 4000, okyo, Japan). The surface morphology of NiCrSi thin-film resistors was observed by field emission canning electron microscopy (FESEM) (JEOL JSM-6700, Tokyo, Japan). As Figure 1d shows, only ano-crystalline grains were observed and surface morphology was almost unchanged as the different The surface morphology of NiCrSi thin-film resistors was observed by field emission scanning electron microscopy (FESEM) (JEOL JSM-6700, Tokyo, Japan). As Figure 1d shows, only nano-crystalline grains were observed and surface morphology was almost unchanged as the different deposition time and the annealing process were used. The resistance was measured by using the four-point-probe method and the resistivity was calculated by the measured resistance and the thickness of Ni-Cr-Si-based thin films. Resistance values for materials at any temperature other than the standard temperature (usually specified at 20˝C) on the specific resistance table can be determined through the following formula: where R is the material resistance at temperature T, R ref is the material resistance at temperature T ref , usually 20˝C or 0˝C, α is temperature coefficient of resistance (TCR) for the material symbolizing the resistance change factor per degree of temperature change, T is the material temperature in degrees Celcius, and T ref is the reference temperature that α is specified at for the material, respectively. In this study, the measured temperatures were 25˝C, 50˝C, 75˝C, 100˝C, and 125˝C, respectively, the resistivity measured at the two temperatures were used to find the temperature coefficient of resistance of NiCrSi thin-film resistors.
Results and Discussion
The effect of deposition time on the thickness of as-deposited and annealed NiCrSi thin films was investigated, and the results are shown in Figure 2. As the deposition temperature was 25˝C, the thicknesses of the 15 min-, 30 min-, 60 min-, and 150 min-deposited NiCrSi thin films measured by using Ellipsometer and the thickness were around 30.8 nm, 90.7 nm, 140.1 nm, and 334.7 nm, respectively. As the deposition time increased, the increase in thickness of NiCrSi thin films is expectable. However, as NiCrSi thin films were annealed at 400˝C in N 2 atmosphere, the thicknesses of the 15 min-, 30 min-, 60 min-, and 150 min-deposited NiCrSi thin films decreased slightly. thicknesses of the 15 min-, 30 min-, 60 min-, and 150 min-deposited NiCrSi thin films measured by using Ellipsometer and the thickness were around 30.8 nm, 90.7 nm, 140.1 nm, and 334.7 nm, respectively. As the deposition time increased, the increase in thickness of NiCrSi thin films is expectable. However, as NiCrSi thin films were annealed at 400 °C in N2 atmosphere, the thicknesses of the 15 min-, 30 min-, 60 min-, and 150 min-deposited NiCrSi thin films decreased slightly. XRD was used to investigate the structural properties of NiCrSi thin films deposition at room temperature in a pure Ar atmosphere. Figure 3 shows the XRD patterns of Al2O3 substrate, Al2O3 substrate with Ag paste. Figure 3 also shows the XRD patterns of the 15 min-, 30 min-, 60 min-, and 150 min-deposited NiCrSi thin films with the different thicknesses. As Figure 3 shows, the XRD patterns of all deposited NiCrSi thin films revealed an amorphous structure and no apparent crystalline phases were observed. Only the Ag and Al2O3 phases were observed in Figure 3. Those results suggest XRD was used to investigate the structural properties of NiCrSi thin films deposition at room temperature in a pure Ar atmosphere. Figure 3 shows the XRD patterns of Al 2 O 3 substrate, Al 2 O 3 substrate with Ag paste. Figure 3 also shows the XRD patterns of the 15 min-, 30 min-, 60 min-, and 150 min-deposited NiCrSi thin films with the different thicknesses. As Figure 3 shows, the XRD patterns of all deposited NiCrSi thin films revealed an amorphous structure and no apparent crystalline phases were observed. Only the Ag and Al 2 O 3 phases were observed in Figure 3. Those results suggest that the thickness (or deposition time) will not affect the crystalline structure of deposited NiCrSi thin films.
Materials 2015, 8 5
that the thickness (or deposition time) will not affect the crystalline structure of deposited NiCrSi thin films. Figure 4 shows the effect of thickness on the variations of resistance and resistivity for NiCrSi thin-film resistors measured at 25 °C and 125 °C. The NiCrSi thin-film resistors' resistance was recorded by the four-point measurement and the resistivity was derived from the resistance using a measurement of film thickness. As Figure 4 shows, as the measured temperatures were increased from 25 °C to 125 °C (only measured at 25 °C to 125 °C were shown); however, the resistance of NiCrSi thin-film resistors increased slightly and they had similar results. Figure 4 also shows that the thinner NiCrSi thin-film resistors showed higher resistivity and the resistivity reached a saturation value as the thickness was equal and more than 140.1 nm. If we suppose the thickness of NiCrSi thin-film resistors was independent of measured temperature, the resistivity of the 25 °C-and 125 °C-measured NiCrSi thin-film resistors had almost the same values and the variations in resistivity were not apparently observed. Figure 4 shows that the resistivity linearly decreased with increasing NiCrSi thin-films' thickness. In the free-electron model of a metallic thin film with hard-wall boundary conditions, the discretization of energy levels makes it impossible to treat both the Fermi energy and the electron Figure 4 shows the effect of thickness on the variations of resistance and resistivity for NiCrSi thin-film resistors measured at 25˝C and 125˝C. The NiCrSi thin-film resistors' resistance was recorded by the four-point measurement and the resistivity was derived from the resistance using a measurement of film thickness. As Figure 4 shows, as the measured temperatures were increased from 25˝C to 125˝C (only measured at 25˝C to 125˝C were shown); however, the resistance of NiCrSi thin-film resistors increased slightly and they had similar results. Figure 4 also shows that the thinner NiCrSi thin-film resistors showed higher resistivity and the resistivity reached a saturation value as the thickness was equal and more than 140.1 nm. If we suppose the thickness of NiCrSi thin-film resistors was independent of measured temperature, the resistivity of the 25˝C-and 125˝C-measured NiCrSi thin-film resistors had almost the same values and the variations in resistivity were not apparently observed. Figure 4 shows that the resistivity linearly decreased with increasing NiCrSi thin-films' thickness. In the free-electron model of a metallic thin film with hard-wall boundary conditions, the discretization of energy levels makes it impossible to treat both the Fermi energy and the electron density as independent of the thickness [14].
NiCrSi thin-film resistors showed higher resistivity and the resistivity reached a saturation value as the thickness was equal and more than 140.1 nm. If we suppose the thickness of NiCrSi thin-film resistors was independent of measured temperature, the resistivity of the 25 °C-and 125 °C-measured NiCrSi thin-film resistors had almost the same values and the variations in resistivity were not apparently observed. Figure 4 shows that the resistivity linearly decreased with increasing NiCrSi thin-films' thickness. In the free-electron model of a metallic thin film with hard-wall boundary conditions, the discretization of energy levels makes it impossible to treat both the Fermi energy and the electron density as independent of the thickness [14]. Nevertheless, many scattering effects are believed to affect the resistivity of NiCrSi thin-film resistors, including surface scattering effect, grain boundaries scattering effect, uneven or rough surfaces scattering effect, and impurities scattering effect, respectively [15]. Surface scattering effect is dependent on the thickness of thin-film resistors and other effects are dependent on the procedures and conditions used to fabricate the thin films, and thus, it is very difficult to quantify each of these effects without measurement [15]. From the Lacy's propose, the electrical resistivity as a function of film thickness can be expressed as: and k " t{2 l where k is constant and 0 < k ď 1, ρ is resistivity of thin film resistors, ρ o is the bulk resistivity of the material, l is the average traveling distance of electrons an, t is the thickness of thin film conductors and it is assumed to have smooth or even surfaces.
The TCR values of deposited NiCrSi thin-film resistors are shown in Figure 5 as a function of film's thickness, using the measured results shown in Figure 4. All of the TCR values of deposited NiCrSi thin-film resistors has a negative number, meaning that resistance decreases with increasing measured temperature. For pure metals, this coefficient is a positive number, meaning that resistance increases with increasing measured temperature. The TCR value of Ni metal is 0.00017 and the TCR value of Cr metal is 13ˆ10´8, respectively. Dhere et al. found that Ni-Cr thin films of low positive TCR value (less than 100 ppm/˝C) were obtained at all thicknesses studied when the sum of the total atomic contents of chromium, oxygen and carbon reached 50%-55% [16]. For the elements silicon in single crystalline type, the TCR value is a negative number of about´0.04 (depending strongly on the presence of impurities in the material). Because of this, we believe the negative TCR value of NiCrSi thin-film resistors is caused by the addition of Si in Ni-Cr alloy, and Si can be added in the NiCr composition for shifting the TCR values to close 0 ppm/˝C. The TCR shown in Figure 5 first shifted to small negative value as the thin films' thickness increased from 30.8 nm to 140.1 nm and then shifted to large negative value as the thin films' thickness increased to 334.7 nm. The reason for this result is not really known, the formations of unknown alloy or compound during the annealing process is the possible reason for this result. composition for shifting the TCR values to close 0 ppm/°C. The TCR shown in Figure 5 first shifted to small negative value as the thin films' thickness increased from 30.8 nm to 140.1 nm and then shifted to large negative value as the thin films' thickness increased to 334.7 nm. The reason for this result is not really known, the formations of unknown alloy or compound during the annealing process is the possible reason for this result. Lacy's proposal shows that as the thickness of thin-film resistors is thinner than 20 nm, the resistivity of thin-film resistors will exponentially decrease. As the thickness of thin-film resistors is thicker than 20 nm, the resistivity of thin-film resistors will linearly decrease. In a thin film material, if, as proposed, the thin films have smooth or even surfaces, the surface scattering (rather than other scatterings) is believed to be the main reason that will affect the resistivity of thin-film materials. In this study, the thickness of deposited thin films is thicker than 30 nm. We believe the thin films are thick enough and the bulk mean free path of the electrons in NiCrSi thin-film resistors is less than t/2. Therefore, only the partial electrons located at in the t/2 region of upper face and down face will be scattered by the thin films' surfaces, and only partial electrons located at t/2 the mean free path of the average conduction electrons will be altered or the electrons will be scattered by the surface. The ratio of conduction electrons will be altered or be scattered decreased with increasing the thickness of NiCrSi thin films. We believe this is the reason that the resistivity of NiCrSi thin-film resistors linearly decreases with increasing thickness. Figure 6 shows the XRD patterns of 60 min-deposited NiCrSi thin films as a function of annealing time. As Figure 6 shows, only the Ag and Al 2 O 3 phases were observed and no other crystalline phases were observed. Because the NiCrSi thin films are annealed in N 2 atmosphere, we believe that the oxidation will not happen during the annealing process. Those results suggest the annealing process has no effect on the crystallization of NiCrSi thin films but it will have the chance to densify the NiCrSi thin films.
Materials 2015, 8 7
Lacy's proposal shows that as the thickness of thin-film resistors is thinner than 20 nm, the resistivity of thin-film resistors will exponentially decrease. As the thickness of thin-film resistors is thicker than 20 nm, the resistivity of thin-film resistors will linearly decrease. In a thin film material, if, as proposed, the thin films have smooth or even surfaces, the surface scattering (rather than other scatterings) is believed to be the main reason that will affect the resistivity of thin-film materials. In this study, the thickness of deposited thin films is thicker than 30 nm. We believe the thin films are thick enough and the bulk mean free path of the electrons in NiCrSi thin-film resistors is less than t/2. Therefore, only the partial electrons located at in the t/2 region of upper face and down face will be scattered by the thin films' surfaces, and only partial electrons located at t/2 the mean free path of the average conduction electrons will be altered or the electrons will be scattered by the surface. The ratio of conduction electrons will be altered or be scattered decreased with increasing the thickness of NiCrSi thin films. We believe this is the reason that the resistivity of NiCrSi thin-film resistors linearly decreases with increasing thickness. Figure 6 shows the XRD patterns of 60 min-deposited NiCrSi thin films as a function of annealing time. As Figure 6 shows, only the Ag and Al2O3 phases were observed and no other crystalline phases were observed. Because the NiCrSi thin films are annealed in N2 atmosphere, we believe that the oxidation will not happen during the annealing process. Those results suggest the annealing process has no effect on the crystallization of NiCrSi thin films but it will have the chance to densify the NiCrSi thin films. The variations of resistance and resistivity of NiCrSi thin-film resistors after annealing process are shown in Figure 7 for the deposited thin films measured at 25 °C and in Figure 8 for deposited thin films measured at 125 °C, respectively. Figures 7 and 8 show that as the thickness of NiCrSi thin-film resistors was equal and more than 91 nm and as the annealing time was equal and more than 60 s, the The variations of resistance and resistivity of NiCrSi thin-film resistors after annealing process are shown in Figure 7 for the deposited thin films measured at 25˝C and in Figure 8 for deposited thin films measured at 125˝C, respectively. Figures 7 and 8 show that as the thickness of NiCrSi thin-film resistors was equal and more than 91 nm and as the annealing time was equal and more than 60 s, the resistance and resistivity would reach a stable value, independent of deposition time. Nocerinot and Singer find that the variations of resistance values were caused by two effects [17]. The first effect was adsorption of residual O 2 gas during the deposition process. Because we deposited the thin films in the pure argon atmosphere, the effect of residual O 2 gas on the properties of NiCrSi thin films can be neglected. The second effect was annealing of thin films, which would change the crystalline structure of thin-film materials or form chemical compound with thin-film materials. Thus, we anneal the NiCrSi thin films in the pure N 2 atmosphere, the effect of chemical compound formed can also be neglected. However, the annealing process can stabilize the crystalline structure of deposition on thin films, because the resistivity and resistance of NiCrSi thin-film resistors will have a stable property after a period of annealing process.
Materials 2015, 8 8 of thin-film materials or form chemical compound with thin-film materials. Thus, we anneal the NiCrSi thin films in the pure N2 atmosphere, the effect of chemical compound formed can also be neglected. However, the annealing process can stabilize the crystalline structure of deposition on thin films, because the resistivity and resistance of NiCrSi thin-film resistors will have a stable property after a period of annealing process. For applications of thin-film resistors, the temperature coefficient of resistance is hoped to be very close to zero, meaning that the resistance hardly changes at all with variations in temperature. In the past, the TCR values could be kept near zero for a wide variety of hybrid substrate materials with varied surface finishes by changing dopant concentration in the sputtering target. During the deposition process, the resistance and TCR value of thin-film resistors are also found to vary with both deposition and annealing parameters. Variations of TCR values of deposited NiCrSi thin-film resistors are shown in Figure 9 as a function of annealing time. Figure 9 shows that, as the annealing time was increased from 0 s to 60 s, the TCR value of the 33 nm-NiCrSi thin-film resistors changed from negative to positive; as the annealing time was increased from 60 s to 600 s, the TCR value changed from of thin-film materials or form chemical compound with thin-film materials. Thus, we anneal the NiCrSi thin films in the pure N2 atmosphere, the effect of chemical compound formed can also be neglected. However, the annealing process can stabilize the crystalline structure of deposition on thin films, because the resistivity and resistance of NiCrSi thin-film resistors will have a stable property after a period of annealing process. For applications of thin-film resistors, the temperature coefficient of resistance is hoped to be very close to zero, meaning that the resistance hardly changes at all with variations in temperature. In the past, the TCR values could be kept near zero for a wide variety of hybrid substrate materials with varied surface finishes by changing dopant concentration in the sputtering target. During the deposition process, the resistance and TCR value of thin-film resistors are also found to vary with both deposition and annealing parameters. Variations of TCR values of deposited NiCrSi thin-film resistors are shown in Figure 9 as a function of annealing time. Figure 9 shows that, as the annealing time was increased from 0 s to 60 s, the TCR value of the 33 nm-NiCrSi thin-film resistors changed from negative to positive; as the annealing time was increased from 60 s to 600 s, the TCR value changed from For applications of thin-film resistors, the temperature coefficient of resistance is hoped to be very close to zero, meaning that the resistance hardly changes at all with variations in temperature. In the past, the TCR values could be kept near zero for a wide variety of hybrid substrate materials with varied surface finishes by changing dopant concentration in the sputtering target. During the deposition process, the resistance and TCR value of thin-film resistors are also found to vary with both deposition and annealing parameters. Variations of TCR values of deposited NiCrSi thin-film resistors are shown in Figure 9 as a function of annealing time. Figure 9 shows that, as the annealing time was increased from 0 s to 60 s, the TCR value of the 33 nm-NiCrSi thin-film resistors changed from negative to positive; as the annealing time was increased from 60 s to 600 s, the TCR value changed from +31.7 ppm/˝C to 94.2 ppm/˝C. As the thicknesses of NiCrSi thin-film resistors were 91 nm, 140 nm, and 330 nm and as the annealing time was increased from 0 s to 60 s, the TCR values changed from´106.4~´153.3 ppm/˝C to´41.9~0 ppm/˝C; as the annealing time was increased from 60 s to 600 s, the TCR values were stable in the range of´14.5 ppm/˝C to´37.7 ppm/˝C. Nocerinot and Singer found that the effects of introducing oxygen into the Ni-Cr thin films during deposition process would decrease the TCR value [17]. However, we deposited and annealed the NiCrSi thin-film resistors in Ar and N 2 atmosphere, respectively, the substrate used in this study is the high resistivity Al 2 O 3 , which has the resistivity higher than 10 8 Ω-cm. Therefore, the effect of substrates will be neglected. Bayne found that the TCR value of standard 60/40 Ni-Cr thin-film resistors was around~100 ppm/˝C for constant sheet resistance films [18]. Those results suggest that before the annealing process the Si will dominate the characteristic of TCR, and after annealing process the effect of Ni-Cr Si will affect the characteristic of TCR, because the TCR value of NiCrSi thin-film resistors changes from negative to positive or shifts to near 0 ppm/˝C. [17]. However, we deposited and annealed the NiCrSi thin-film resistors in Ar and N2 atmosphere, respectively, the substrate used in this study is the high resistivity Al2O3, which has the resistivity higher than 10 8 Ω-cm. Therefore, the effect of substrates will be neglected. Bayne found that the TCR value of standard 60/40 Ni-Cr thin-film resistors was around ~100 ppm/°C for constant sheet resistance films [18]. Those results suggest that before the annealing process the Si will dominate the characteristic of TCR, and after annealing process the effect of Ni-Cr Si will affect the characteristic of TCR, because the TCR value of NiCrSi thin-film resistors changes from negative to positive or shifts to near 0 ppm/°C.
Figure 9.
Variations of temperature coefficient of resistance of deposited NiCrSi thin-film resistors as a function of annealing time.
Conclusions
As the deposition time was 15 min, 30 min, 60 min, and 150 min, the thickness of NiCrSi thin films was around 30.8 nm, 90.7 nm, 140.1 nm, and 334.7 nm, respectively. All of the NiCrSi thin films showed amorphous phase even the deposition time was 150 min, and even thin films annealed at 400 °C for 300 s NiCrSi also showed amorphous phase. The resistivity of deposited NiCrSi thin-film resistors first increased with thickness and reached a saturation value as the thickness was equal and more than 140.1 nm, and the resistivity linearly decreased with increasing NiCrSi thin-films' thickness. When the annealing process was used, as the thickness of NiCrSi thin-film resistors was equal and more than 91 nm and as the annealing time was equal and more than 60 s, the resistance and resistivity would reach a stable value, independent of deposition time. As the thicknesses of NiCrSi thin-film resistors were 91 nm, 140 nm, and 330 nm, and the annealing time was more than 60 s, the TCR values were stable in the range of −14.5 ppm/°C to −37.7 ppm/°C.
Conclusions
As the deposition time was 15 min, 30 min, 60 min, and 150 min, the thickness of NiCrSi thin films was around 30.8 nm, 90.7 nm, 140.1 nm, and 334.7 nm, respectively. All of the NiCrSi thin films showed amorphous phase even the deposition time was 150 min, and even thin films annealed at 400˝C for 300 s NiCrSi also showed amorphous phase. The resistivity of deposited NiCrSi thin-film resistors first increased with thickness and reached a saturation value as the thickness was equal and more than 140.1 nm, and the resistivity linearly decreased with increasing NiCrSi thin-films' thickness. When the annealing process was used, as the thickness of NiCrSi thin-film resistors was equal and more than 91 nm and as the annealing time was equal and more than 60 s, the resistance and resistivity would reach a stable value, independent of deposition time. As the thicknesses of NiCrSi thin-film resistors were 91 nm, 140 nm, and 330 nm, and the annealing time was more than 60 s, the TCR values were stable in the range of´14.5 ppm/˝C to´37.7 ppm/˝C. | 6,907.6 | 2015-10-01T00:00:00.000 | [
"Engineering",
"Materials Science"
] |
A Larger Membrane Area Increases Cytokine Removal in Polymethyl Methacrylate Hemofilters
Blood purification is performed to control cytokines in critically ill patients. The relationship between the clearance (CL) and the membrane area during adsorption is not clear. We hypothesized that the CL increases with the hydrophobic area when hydrophobic binding contributes to cytokine adsorption. We investigated the relationship between the hemofilter membrane area and the CL of the high mobility group box 1 protein (HMGB-1) and interleukin-6 (IL-6). We performed experimental hemofiltration in vitro using polymethyl methacrylate membranes CH-1.8W (1.8 m2) and CH-1.0N (1.0 m2), as well as polysulfone membrane NV-18X (1.8 m2). After adding 100 mg of HMGB1 or 10 μg of IL-6 into the test solution, experimental hemofiltration was conducted for 360 min in a closed-loop circulation system, and the same amount of HMGB1 and IL-6 was added after 180 min. With CH-1.8W and CH-1.0N, both HMGB-1 and IL-6 showed a rapid concentration decrease of more than 70% at 180 min and 360 min after the re-addition. At 15 min, the CL of HMGB-1 was CH-1.8W: 28.4 and CH-1.0N: 19.8, and that of IL-6 was CH-1.8W: 41.1 and CH-1.0N: 25.4. CH-1.8W and CH-1.0N removed HMGB1 and IL-6 by adsorption and CH-1.8W was superior in CL, which increased with a greater membrane area.
Introduction
Blood purification plays an important role in the treatment of patients who are severely ill, including those with sepsis, and is often performed not only as renal replacement therapy, but also for the modulation of cytokines and other humoral mediators [1][2][3]. Sepsis occurs when there is organ damage caused by an uncontrolled host immune response that has been triggered by infection [4] and its pathogenesis, hypercytokinemia, includes interleukin-1 (IL-1), interleukin-6 (IL-6), tumor necrosis factor, and high mobility group box 1 protein (HMGB-1), which is a major inflammatory cytokine [5]. IL-6 is a cytokine released by the immune cells and plays a role in the systemic inflammatory changes caused by infection or tissue injury [6]. Recently, IL-6 produced by the marginal zone B cells in animal experiments was found to be a factor that promoted the development of sepsis. In these experiments, an improvement in mortality from sepsis by administering IL-6 antibodies has been reported [7]. Several studies have also reported that the serum IL-6 concentration is associated with disease severity [8], organ dysfunction [9], and overall mortality among patients with sepsis, burn and trauma injuries, and cardiovascular diseases, as well as those undergoing hemodialysis [10][11][12][13]. HMGB-1 is one of the DAMPs (damage associated molecular patterns) released from injured tissues into the bloodstream that metastasizes and amplifies inflammation. Wang et al. reported that HMGB-1 was a late mediator of endotoxin lethality and that the serum levels of HMGB-1 were increased significantly in the septic patients who did not survive compared with the survivors [14]. Hence, HMGB-1 has been studied as a key mediator of sepsis. In clinical settings, the plasma levels of HMGB-1 were associated with the severity and mortality attributed to sepsis and correlated with RIPK3 (receptor interacting protein kinase 3) and MLKL (mixed lineage kinase domain-like protein), suggesting an association of HMGB-1 with necroptosis [15].
In our previous investigation of HMGB-1, we found that the efficiency of the removal differed based on the membrane material of the hemofilter [16]; although the upper limit of the filtration clearance (CL) was the filtrate flow rate, HMGB-1 with a molecular weight of 30 kD was removed with a several times greater efficiency than the filtration clearance of the small molecular weight creatinine. This may be explained by the principle of adsorption [17]. In cases of continuous kidney replacement therapy, where the intensity of blood purification is low, the effect of the membrane area on the CL due to filtration or dialysis is negligible [1]. However, the relationship between the CL and the membrane area in cases of adsorption remains unknown.
Hydrophilic binding, such as ionic binding, and hydrophobic binding are the typical adsorption principles of hemofilters. Hydrophilic bonds involve polar molecules with negatively charged oxygen and positively charged hydrogen ions. Typical hydrophobic substances in vivo are membrane proteins, while hydrophilic substances include proteins dissolved in blood, which are determined by the number and steric configuration of the hydrophilic and hydrophobic groups. On the one hand, the ionic binding is affected by the charge of the target substance (isoelectric point in the case of proteins), and it has been described as an adsorption mechanism for acrylonitrile-co-methallyl sulfonate surfacetreated (AN69ST) materials [18]. On the other hand, hydrophobic binding is affected by the molecular weight of the target substance and the hydrophobic area (adsorption site) of the membrane. However, the effect of the membrane area of the hydrophobic binding membrane on cytokines has not been evaluated. While p-methyl methacrylate (PMMA) acrylic resin is hydrophobic [19], it is not clear whether the PMMA membrane hemofilter adsorbs cytokines via hydrophobic binding [1].
We hypothesized that, if hydrophobic binding contributes to cytokine adsorption, the CL will increase in parallel with an increase in the hydrophobic area. We investigated the relationship between the hemofilter membrane area and the CL of HMGB-1 and IL-6 in vitro using PMMA membranes with different membrane areas using a polysulfone membrane as a non-adsorptive reference.
Materials and Methods
The in vitro, closed, experimental system consisted of a solution reservoir, hemofilter, and fully automated blood purification machine (TR525, Toray Medical Co., Tokyo, Japan), as shown in Figure 1. The hollow fiber hemofilters investigated were CH-1.8W (Toray Ind., Tokyo, Japan; polymethylmethacrylate: PMMA) with a large membrane area of 1.8 m 2 , CH-1.0N (Toray, Tokyo, Japan; PMMA) with a normal membrane area of 1.0 m 2 , and NV-18X (Toray, Tokyo, Japan; polysulfone: PS) with a large membrane area of 1.8 m 2 . In the evaluation of CH-1.8W, CH-1.0N was used for the membrane area comparison and NV-18X was used as the control membrane. The test solution was prepared by dissolving 35 g of bovine serum albumin (MW 66 kDa; Fuji Film Wako Pure Chemical Co., Osaka, Japan) in 1000 mL of the substitution fluid for hemofiltration Sublood BSG ® (Fuso Pharmaceutical Ind., Osaka, Japan), adding 100 µg of HMGB-1 (Shino-Test Co., Sagamihara, Japan) or 10 µg of IL-6 (Kamakura Techno Science, Kamakura, Japan), and stirring the mixture thoroughly with a magnetic stirrer. The circuit containing the hemofiltration membrane was primed with a saline solution and extruded slowly with the above test solution, and then the priming volume was discarded to make a closed circuit. The test solution was pumped from the solution reservoir to a hemofilter at a solution flow rate (Qb) of 100 mL/min and returned to the same reservoir. The ultrafiltrate was pumped in the post-dilution mode at a filtrate flow rate (Qf) of 1000 mL/h (16.7 mL/min) and returned to the reservoir in a closedloop circulation system. The in-flask concentration decreased only when cytokines were removed by adsorption in this system. The experimental hemofiltration was conducted for 360 min at 37 • C. To confirm the adsorption saturation limit of the membranes, 100 µg of HMGB-1 or 10 µg of IL-6 was added again after 180 min; the mixture was circulated for 5 min to make the solution concentration uniform and then evaluated for a period of 185 min to 360 min. The samples for laboratory analyses were taken from both the inlet and outlet of the filter and from the filtrate port. The in-flask concentration was the same as the inlet concentration. The point at which the test solution was circulated for 3 min to equalize the concentration in the circuit was set at 0 min. The samples were collected at 0, 5, 15, 30, 45, 60, 90, 120, and 180 min. After adding HMGB-1 or IL-6 again after 180 min, the samples were collected at 185, 195, 210, 225, 240, 270, 300, and 360 min. Over time, the clearances were calculated using the following formula: This is a recirculation model for testing the adsorption mechanism. Only when the adsorption occurred, did the concentration of HMGB-1 or IL-6 at the inlet side decrease.
Results
The changes in the IL-6 and HMGB-1 concentrations at the inlet of each hemofilter are shown in Figures 2 and 3, respectively. In CH-1.8W and CH-1.0N, both IL-6 and HMGB-1 showed a rapid decrease in concentration of more than 70% at 180 min. Even after re-addition, there was a decrease in the concentration of approximately 70% at 360 min. To evaluate the effects of time from the addition (or re-addition), the number of times added, interaction between time and number of times added, type of hemofilter, and inlet concentration on the CL, we used a linear mixed effects model with the intercepts being random effects. The modifying effect of the inlet concentration on the association between the type of hemofilter and the CL was assessed using separate models after the initial addition and re-addition, each including a non-linear term and an interaction term between the type of hemofilter and the inlet concentration. The association of the CL was confirmed using an F-test on the regression coefficients obtained from the mixed-effect model. The statistical significance was set at a two-sided p-value of <0.05. All of the statistical analyses were performed using R version 4.1.1 patched (http://www.r-project.org, accessed on 1 August 2022).
Results
The changes in the IL-6 and HMGB-1 concentrations at the inlet of each hemofilter are shown in Figures 2 and 3, respectively. In CH-1.8W and CH-1.0N, both IL-6 and HMGB-1 showed a rapid decrease in concentration of more than 70% at 180 min. Even after re-addition, there was a decrease in the concentration of approximately 70% at 360 min. Figure 1. Diagram of the in vitro experimental setup. This is a recirculation model for testing the adsorption mechanism. Only when the adsorption occurred, did the concentration of HMGB-1 or IL-6 at the inlet side decrease.
Results
The changes in the IL-6 and HMGB-1 concentrations at the inlet of each hemofilter are shown in Figures 2 and 3, respectively. In CH-1.8W and CH-1.0N, both IL-6 and HMGB-1 showed a rapid decrease in concentration of more than 70% at 180 min. Even after re-addition, there was a decrease in the concentration of approximately 70% at 360 min. In NV-18X, the inlet concentration decreased up to 15 min after addition and then remained almost unchanged until 180 min. After the re-addition, the concentration remained the same after the added amount was increased.
At 15 min, the IL-6 CL was CH-1.8W: 4 1.1 mL/min, CH-1.0N: 2 5.4 mL/min, and NV-18X: 25.4 mL/min and the HMGB-1 CL was CH-1.8W: 28.4 mL/min, CH-1.0N: 19.8 mL/min, and NV-18X: 6.6 mL/min ( Figure 4). With CH-1.8W and CH-1.0N, the CL of both IL-6 and HMGB-1 consisted almost completely of adsorption and increased with the increasing membrane area. In NV-18X, the inlet concentration decreased up to 15 min after addition and then remained almost unchanged until 180 min. After the re-addition, the concentration remained the same after the added amount was increased.
At 15 min, the IL-6 CL was CH-1.8W: 4 1.1 mL/min, CH-1.0N: 2 5.4 mL/min, and NV-18X: 25.4 mL/min and the HMGB-1 CL was CH-1.8W: 28.4 mL/min, CH-1.0N: 19.8 mL/min, and NV-18X: 6.6 mL/min ( Figure 4). With CH-1.8W and CH-1.0N, the CL of both IL-6 and HMGB-1 consisted almost completely of adsorption and increased with the increasing membrane area. The results of the mixed-effects model analysis and the predicted values of change in the IL-6 CL over time for each hemofilter from the initial addition and re-addition are shown in Figure 5 and Table 1. After the re-addition, the IL-6 CL was, on average, 8.879 mL/min lower than after the initial addition (p = 0.005). The IL-6 CL decreased by an average of 0.08 mL/min per minute (p = 0.001). Furthermore, the IL-6 CL in CH-1.8W was significantly higher than the IL-6 CL in CH-1.0N (3.443) (p = 0.016). For every 1 pg/mL The results of the mixed-effects model analysis and the predicted values of change in the IL-6 CL over time for each hemofilter from the initial addition and re-addition are shown in Figure 5 and Table 1. After the re-addition, the IL-6 CL was, on average, 8.879 mL/min lower than after the initial addition (p = 0.005). The IL-6 CL decreased by an average of 0.08 mL/min per minute (p = 0.001). Furthermore, the IL-6 CL in CH-1.8W was significantly higher than the IL-6 CL in CH-1.0N (3.443) (p = 0.016). For every 1 pg/mL increase in concentration, the IL-6 CL increased by an average of 0.002 mL/min (p = 0.023). A significant interaction was observed between the number of additions and time (p = 0.023). The results of the mixed-effects model analysis and the predicted values of change in the HMGB-1 CL over time for each hemofilter from the initial addition and re-addition are shown in Figure 6 and Table 2. After the re-addition, the HMGB-1 CL was, on average, 13.309 mL/min lower than after the initial addition (p < 0.001). The HMGB-1 CL decreased by an average of 0.075 mL/min per minute (p < 0.001). The HMGB-1 CL in CH-1.8W was significantly higher than the HMGB-1 CL in CH-1.0N (2.923) (p = 0.006). For every 1 ng/mL increase in concentration, the HMGB-1 CL increased by an average of 0.287 mL/min (p < 0.001). A significant interaction was observed between the number of additions and time (p < 0.001). The relationship between the inlet concentration and the IL-6 CL is shown separately after the initial addition and re-addition for each hemofilter (Figure 7a,b). The interaction between the inlet concentration and hemofilter type was significant, after both the initial addition and the following re-addition (p = 0.007 and p = 0.010, respectively); however, the respective associations showed the opposite results. The CL was higher for CH-1.8W at higher inlet concentrations after the initial addition, whereas after the re-addition, the CL decreased at higher inlet concentrations and remained flat at approximately 20 mL/min. Table 2. The relationship between each parameter, the number of additions, time, hemofilter type, and HMGB-1 concentration at the inlet side (CBi), and interaction between the number of additions and CBi, and with the HMGB-1 CL as an objective variable, using a linear mixed model. The results of the linear mixed model evaluating the association between the IL-6 CL and the inlet concentrations are shown after the initial addition and the following re-addition (Table 3). After the initial addition, for every 1 pg/mL increase in concentration, the IL-6 CL increased by an average of 0.005 mL/min (p = 0.001). After the re-addition, the IL-6 CL decreased, on average, by 0.002 mL/min for each 1 pg/mL increase in concentration (p = 0.049). After re-administration, the IL-6 CL decreased by an average of 0.121 mL/min (p = 0.001). Table 3. The relationship between each parameter, the IL-6 concentration at the inlet side (CBi), hemofilter type, and time, and the interaction between the CBi and hemofilter type, with the IL-6 CL as an objective variable, using a linear mixed model. *: p-value of <0.05 was considered as significant. The results of the linear mixed model evaluating the association between the IL-6 CL and the inlet concentrations are shown after the initial addition and the following re-addition (Table 3). After the initial addition, for every 1 pg/mL increase in concentration, the IL-6 CL increased by an average of 0.005 mL/min (p = 0.001). After the re-addition, the IL-6 CL decreased, on average, by 0.002 mL/min for each 1 pg/mL increase in concentration (p = 0.049). After re-administration, the IL-6 CL decreased by an average of 0.121 mL/min (p = 0.001).
Coefficient
The relationship between the inlet concentration and the HMGB-1 CL is shown separately after the initial addition and re-addition for each hemofilter (Figure 8a,b). The interaction between the inlet concentration and hemofilter type was significant after the initial addition (p = 0.003). Table 3. The relationship between each parameter, the IL-6 concentration at the inlet side (CBi), hemofilter type, and time, and the interaction between the CBi and hemofilter type, with the IL-6 CL as an objective variable, using a linear mixed model. *: p-value of <0.05 was considered as significant. The results of the evaluation of the association between the HMGB-1 CL and the inlet concentrations using a linear mixed model are shown in Table 4, after the initial addition and the following re-addition. After the initial addition, the HMGB-1 CL decreased by an average of 0.069 mL/min (p = 0.001). The results of the evaluation of the association between the HMGB-1 CL and the inlet concentrations using a linear mixed model are shown in Table 4, after the initial addition and the following re-addition. After the initial addition, the HMGB-1 CL decreased by an average of 0.069 mL/min (p = 0.001).
Discussion
In recent years, the control of humoral mediators such as cytokines, using the principle of adsorption, has attracted attention in blood purification therapy for sepsis [20]. Membranes with a high adsorption capacity include AN69ST, oXiris, and PMMA [21]. PMMA and AN69ST membranes have been used extensively in Japan, and many treatment experiences, especially for sepsis, have been reported [22][23][24]. However, there have been no direct reports indicating that the removal of IL-6 or HMGB-1 improved sepsis mortality.
PMMA membranes have a homogenous symmetric structure and may exhibit adsorption capacity owing to hydrophobic binding, not only on the hollow fiber inner surface, but also on the entire membrane, including the bulk layer [25]. The diameter of the pores is designed to be 6.6-7.0 nm, and IL-6, with a radius of 2 nm, is also expected to be adsorbed efficiently on the inner surface of the pore.
During filtration, water and other substances move from the blood to the filtrate side through the pores. If the pore size is sufficiently large to allow the substances to pass through, removal by filtration is expected according to the filtration flow rate (1000 mL/h = 17 mL/min). However, in the present study, there was little removal of both IL-6 and HMGB1 with filtration through the CH-1.0N and CH-1.8W PMMA membranes (Figure 4). This is because most of the substance is adsorbed when passing through the hollow fiber inner surface and bulk layer and, therefore, does not enter the filtrate. This difference in removal capacity according to the membrane area was also observed in the comparison of 15-minute values: the IL-6 CL was approximately 1.6-fold larger and the HMGB-1 CL was approximately 1.4-fold larger when the membrane area was 1.8-fold larger. Although the CL did not increase at the same ratio as the increase in membrane area, it increased with increasing membrane area, indicating that a larger membrane area increases the cytokine removal capacity.
Matsumura et al. devised a method using two PMMA membranes with a large membrane area of 2.1 mm 2 as enhanced-intensity PMMA-CHDF and used them clinically. They reported that the IL-6 removal ability was superior, and that the mortality rate was also reduced compared with the normal CHDF group that used a 1.0 m 2 PMMA membrane [26]. As the procedure was performed using two consoles, the efficiency of dialysis and filtration was doubled; however, the increased membrane area may also have had a significant impact.
The increase in the concentration before and after the re-addition of IL-6 was smaller in CH-1.8W and CH-1.0N than in NV-18X. As mentioned above, this is because NV-18X has a low level of removal, whereas CH-1.8W and CH-1.0N continue to be adsorbed and removed for 5 min until the next measurement after the re-addition.
The CL of CH-1.8W and CH-1.0N decreased with time for both IL-6 and HMGB1, but showed a significant increase in CL after the re-addition. This indicates that the adsorption capacity of the membrane did not reach saturation. Compared with the CL at the time of the initial addition, the CL after the re-addition was low, and while the CL decreased with time, it still maintained a sufficient CL. This may have been because of the main adsorption site during the initial addition being the hollow fiber lumen, whereas the adsorption site after the re-addition was the bulk layer. Figure 6 shows the relationship between the inlet concentration and the IL-6 CL for CH-1.8W and CH-1.0N, without the time factor. In CH-1.8W, after the re-addition, the IL-6 CL appears to be lower at higher inlet concentrations. However, in the actual data, the IL-6 CL in CH-1.8W was higher than in CH-1.0N, while the inlet concentration was high (180-225 min) even after the re-addition (Supplement Table S1). After the re-addition, the inlet concentration was lower in CH-1.8W at all the measurement points, that is, a greater amount of IL-6 was removed. As the substance concentration was lower, the removal efficiency also decreased because the frequency of contact with the adsorption surface was lower. In CH-1.8W, the inlet concentration dropped sharply after 300 min. On the other hand, the CL is calculated from the ratio of the inlet and outlet concentrations, and while an accurate clearance is obtained when the target substance concentration is sufficiently high, a small change is reflected in the clearance, to a large extent, when the target substance concentration is low. The fact that the IL-6 CL appears to be lower at higher inlet concentrations in CH-1.8W after the re-addition may have been because of these errors.
The IL-6 CL after the re-addition was approximately 20 mL/min without the time factor. Although little filtration removal was observed in this study (Figure 4), both showed IL-6 CLs with a filtration flow rate of approximately 17 mL/min, suggesting that adsorption occurs mainly in the bulk layer rather than in the inner surface of the hollow fiber. As adsorption in the bulk layer is considered to occur as the filtration passes through the pores, the efficiency of adsorption in the bulk layer approximates the filtration efficiency. The reason CH-1.8W showed a high IL-6 CL according to the inlet concentration after the initial addition may have been because of CH-1.8W having a sufficiently large membrane area; thus, a high percentage of adsorption was also observed in the inner surface of the hollow fiber at the initial addition. After the adsorption sites on the hollow fiber inner surface reached saturation, adsorption would occur in the bulk layer, so the IL-6 CL was considered to be equivalent to that of CH-1.0N according to a filtration flow rate of 17 mL/min. The nature and capacity of adsorption differs between the inner surface of the hollow fiber and the bulk layer, suggesting that they act as if they are different membranes. The hollow fiber lumen surface is 1.8 m 2 or 1.0 m 2 , as described in the manufacturer's instructions. On the other hand, just as the intestinal tract has a huge surface area because of its numerous microvilli, the surface area of the bulk layer is considered to have a huge surface area compared with the hollow fiber lumen surface.
The added amount of 10 µg of IL-6 used in this study corresponds to the amount of IL-6 in the total circulating blood when a patient, weighing 70 kg and with a circulating blood volume of 4.6 L, develops an IL-6 hypercytokinemia of approximately 2170 pg/mL. At this amount, the adsorption sites on the hollow fiber inner surface reached saturation even at CH-1.8W; however, the adsorption capacity in the bulk layer did not reach saturation even at CH-1.0N and still had sufficient adsorption capacity. Although not considered in this study, a larger membrane area may have a higher adsorption capacity, as the bulk layer volume is inevitably larger, and the surface area of the hydrophobic material in the pores is larger. Protein adsorption on PMMA membranes is reported to be via an occlusion into pores of the membrane, which is dependent on the filtration and, with a larger membrane area, a larger surface area is adsorbed [27]. When dialysis is used, an even higher adsorption capacity in the bulk layer may be expected because of internal filtration. The present experiment was a simple experimental system comprising additional filtration. However, in a clinical continuous hemodiafiltration practice, there is dialysate flowing in the opposite direction to the blood on the outside of the hollow fiber, and even though the flow rate is lower than that in conventional hemodialysis, a significant internal filtration effect is expected. Therefore, the diffusion, filtration, and adsorption clearances may be even greater. Usually, the circuit exchange for blood purification is performed every day and, considering the high Il-6 CL at the time of the initial addition, a larger membrane area may be advantageous when selecting PMMA membranes.
There were several limitations to this study. First, this was an in vitro study in which bovine albumin was added to a phosphate buffer solution, which has different properties from blood, which contains many other components, and may differ from the adsorption CL in actual clinical practice. However, because it was a simple experimental system, the adsorption characteristics of HMGB-1 and IL-6 on the membrane itself could be evaluated more accurately. Second, both CH-1.0N and CH-1.8W showed a CL reuptake after a single re-addition, indicating that the CL had not reached saturation. However, we could not indicate how many re-additions were required to reach saturation. In our previous multiple addition experiment on an AN69ST membrane, saturation was not reached even after seven re-additions of 100 mcg of HMGB-1 [17]. The adsorption in the bulk layer of the PMMA membrane was significant, and it may have a comparably large adsorption capacity. Third, only HMGB-1 and IL-6 were examined in this study. When the adsorption principle is ionic bonding, the adsorption capacity depends on the charge of the substance to be adsorbed. Meanwhile, the CL increased as the membrane area increased, suggesting that the PMMA membrane hydrophobicity was involved. However, this experiment did not directly demonstrate the presence of any hydrophobic binding between the PMMA membranes and cytokines. Fourth, the hollow fiber inner diameters were different; that is, CH-1.8W: 240 µm and CH-1.0N: 200 µm. For a more accurate comparison of the effect of an increased membrane area, CH-0.6W or CH-1.3W, which have the same hollow fiber inner diameter as CH-1.8W, should be considered and, in addition, further studies are needed.
Conclusions
The removal characteristics of different membrane materials, membrane areas, and substances to be removed were investigated, and we found that IL-6 and HMGB-1 were removed mainly according to the principle of adsorption in hemofiltration using PMMA membranes, and their removal capacity increased in parallel with an increase in the membrane area. The CL increased as the membrane area increased, suggesting that hydrophobic bonding may be involved in cytokine adsorption, which may allow for the removal of many other cytokines in a non-specific manner. Larger membrane areas may be effective in controlling pathological conditions and in disease control when PMMA membranes are used for blood purification with the intention of controlling mediators.
Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/membranes12080811/s1, Supplemental Table S1. The concentration at the inlet side (CBi) and the change in IL-6 CL over time when using CH-1.0N and CH-1.8W. IL-6 was added again after 180 min to confirm the adsorption saturation limit of the membranes. | 6,332.2 | 2022-08-01T00:00:00.000 | [
"Medicine",
"Materials Science"
] |
Quantifying mRNA and MicroRNA with qPCR in Cervical Carcinogenesis: A Validation of Reference Genes to Ensure Accurate Data
A number of recent studies have catalogued global gene expression patterns in a panel of normal, tumoral cervical tissues so that potential biomarkers can be identified. The qPCR has been one of the most widely used technologies for detecting these potential biomarkers. However, few studies have investigated a correct strategy for the normalization of data in qPCR assays for cervical tissues. The aim of this study was to validate reference genes in cervical tissues to ensure accurate quantification of mRNA and miRNA levels in cervical carcinogenesis. For this purpose, some issues for obtaining reliable qPCR data were evaluated such as the following: geNorm analysis with a set of samples which meet all of the cervical tissue conditions (Normal + CIN1 + CIN2 + CIN3 + Cancer); the use of individual Ct values versus pooled Ct values; and the use of a single (or multiple) reference genes to quantify mRNA and miRNA expression levels. Two different data sets were put on the geNorm to assess the expression stability of the candidate reference genes: the first dataset comprised the quantities of the individual Ct values; and the second dataset comprised the quantities of the pooled Ct values. Moreover, in this study, all the candidate reference genes were analyzed as a single “normalizer”. The normalization strategies were assessed by measuring p16INK4a and miR-203 transcripts in qPCR assays. We found that the use of pooled Ct values, can lead to a misinterpretation of the results, which suggests that the maintenance of inter-individual variability is a key factor in ensuring the reliability of the qPCR data. In addition, it should be stressed that a proper validation of the suitability of the reference genes is required for each experimental setting, since the indiscriminate use of a reference gene can also lead to discrepant results.
Introduction
Cervical cancer is one of the most common cancers affecting women worldwide and is linked to human papillomavirus (HPV) infection [1,2]. This type of cancer is preceded by preventable precancerous lesions; however, conventional screening tests lack both sensitivity (in the Pap test) and specificity (in the HPV test) [3][4][5][6]. Hence, there is an urgent need for new effective biomarkers to improve the triage tests and determine how affected women can be treated in an appropriate way [7].
Some studies have catalogued global gene expression patterns in a panel of normal, tumoral cervical tissues so that potential biomarkers can be identified [8,9]. The real-time quantitative PCR (qPCR) has been one of the most widely used technologies for detecting these potential biomarkers. However, reliable results can only be achieved with this technology by evaluating some crucial parameters [10,11] such as the validation of reference genes, which must be as stable as possible in the investigated samples [12,13]. Additionally, some studies have included a pool of samples in the qPCR assays [14][15][16][17]. This strategy is usually employed to reduce biological variability and also to reduce the costs of the experiments. However searching for variations in gene expression should take account of the endogenous variations of the biological individuals to avoid an erroneous interpretation of the data [18].
Until now, few studies have investigated a correct strategy for data normalization in qPCR assays for cervical tissues; one of them recommended the use of reference genes for mRNA expression studies [19], and another for microRNA (miRNA) expression studies [20]. Nevertheless, several research groups have stressed the importance of evaluating normalization targets as has been demonstrated in the way the results varied in accordance with the choice of reference gene in both the mRNA [21] and miRNA [22] qPCR assays. Additionally, even though it has been well established that the use of a single or unvalidated reference gene is not suitable to obtain reliable qPCR data [12], studies in cervical cancer continue to use the most well-known reference genes such as GAPDH and RNU-6, as single reference gene to measure mRNA and miRNA expression levels, respectively [23][24][25][26][27].
In the light of this, the aim of this study was to validate reference genes in cervical tissues to ensure accurate quantification of mRNA and miRNA levels in cervical carcinogenesis. For this purpose, some issues for obtaining reliable qPCR data were evaluated such as the following: geNorm analysis with a set of samples which meet all of the cervical tissue conditions (Normal + CIN1 + CIN2 + CIN3 + Cancer); the use of individual samples (or individual Ct values) versus a pool of samples (or pooled Ct values); and the use of a single (or multiple) reference genes to quantify mRNA and miRNA expression levels.
Ethics statement
This study was approved by the ''Research Ethics Committee of the Federal University of Pernambuco'', Brazil, (number: 03606212.7.0000.5208) and also by the Institutional Review Board of the Clinical Hospital of UFPE, and the Prof. Fernando Figueira Institute of Integral Medicine -IMIP. All the patients signed a written consent form prior to the collection of the samples.
Patients and samples
The experiments were planned and carried out in accordance with the MIQE guidelines [10]. The biopsies of patients were collected at the Clinical Hospital of UFPE and Institute of Integral Medicine Prof. Fernando Figueira Institute of Integral medicine (IMIP). Biopsies were obtained from women undergoing colposcopy, with different degrees of cervical intraepithelial neoplasia-CIN (CIN 1, 2, 3), and cancer (Ca). Normal cervical tissue samples (negative for neoplasia) were included as controls. Written consent forms were obtained from all the patients, prior to the sample collection. Women with the Human Immunodeficiency Virus (HIV) and/or during pregnancy were excluded from this study. Fresh cervical biopsies were immediately preserved in RNAlater (Qiagen) and stored at 280uC. The biopsies were used in their entirety. HPV detection in samples was performed by PCR [28], after extraction and purification of total DNA with Trizol (Invitrogen) and DNeasy Blood & Tissue Kit (Qiagen), respectively. A total of 65 samples were used that consisted of five groups: CIN1 (12), CIN2 (6), CIN3 (14), cancer (14) and normal cervical tissue (19). All the samples from cancer and CIN were found to be positive for HPV, and all the normal cervical tissue samples were found to be HPV negative.
Isolation of total RNA and cDNA synthesis
The preserved samples (25-100 mg) were ground while still nitrogen-frozen and homogenized with 1 ml of Trizol (Invitrogen) for isolation of total RNA (including miRNAs and mRNAs), in accordance with the manufacturer's instructions. Total RNA was purified in a subsequent stage by means of the miRNA Absolutely RNA Kit (Agilent Technologies). The quantity and purity of total RNA were estimated by NanoDrop 2000 Spectrophotometer (ThermoScientific), and the criterion for the inclusion of the RNA cDNA was synthesized from 1mg of total RNA using the miScript II RT kit (Qiagen) in a 20 ml reaction volume. The cDNA generated with the aid of the miScript II RT Kit was used as a template for quantification of miRNA and mRNA. An RT-minus negative control reaction with all the components for the RT reaction (except the Reverse Transcriptase enzyme) was carried out for each sample to control genomic DNA contamination.
Selection of gene sequences and primer design
Four protein coding genes (mRNA genes) were selected for expression analyses (GAPDH, ACTB, EEF1A1 and RPLPO) based on previous qPCR studies in cervical cancer [19,23,29,30]. The characteristics of each gene, such as the accession number, genomic location, function, and amplicon size are summarized in Table 1. Primers were designed on the basis of the sequence data obtained from GenBank (http://www.ncbi.nlm.nih.gov/) using the CLCBioMain Workbench 5.7.1 software ( Table 2).
In the case of the analysis involving miRNA expression studies, three non-protein coding genes (npcRNA genes) were selected for the evaluation of stability: RNU6-2, miR-191 and miR-23a. These genes are small npcRNAs and correspond to the family of snoRNAs and microRNAs, respectively, which are commonly used as reference genes, not only in cervical tissues, but also in other types of tissues [15,20,22,31]. The primers were purchased from miScript primer assay (Qiagen) and from miScript PCR Starter Kit (Qiagen), which contains the miScript Universal Primer. The characteristics of each npcRNA gene are summarized in Table 3.
Real-time quantitative polymerase chain reaction (qPCR)
The Rotor Gene 6000 thermocycler (Qiagen) was used to run the qPCR reactions. The reactions were in duplicate and the final volume for each reaction was 20 ml, containing 10 ml of 2X QuantiTect SYBR Green PCR kit (Qiagen), 1 ml of forward primer, 1 ml of reverse primer, 6 ml of RNAse-free water and 2 ml of cDNA. The final concentration of each primer in the PCR reaction was 0.5 mM. The final concentration of cDNA was 20 ng per qPCR reaction for the mRNA measurement, and 2 ng of cDNA per qPCR reaction for microRNA measurement. Negative controls without cDNA for each primer pair were added to detect contamination. Negative controls of cDNA synthesis (not submitted to the reverse transcriptase action) were also added to detect possible contamination with genomic DNA. The reaction conditions for the quantification of mRNA were as follows: 15 min at 95uC (initial activation of HotStarTaq DNA Polymerase), followed by 30 cycles of 95uC for 25 s, 60uC for 25 s, and 72uC for 25 s, with a final extension at 72uC for 2 min. For miRNA qPCR, the conditions were: 15 min for 95uC (initial activation of HotStart-Taq DNA Polymerase), followed by 40 cycles of 94uC for 15 s, 55uC for 30 s and 70uC for 30 s.
The amplification efficiency for each primer pair was determined by a qPCR assay using triplicates of a 10-fold dilution series (1:10, 1:100, 1:1000, 1:10.000, 1:100.000) of normal cervical tissue cDNA as a template. The mean Ct values for each serial dilution were plotted against the logarithm of the cDNA dilution factor. The amplification efficiency for each primer pair was calculated by standard curve methods using the Efficiency = (10 (21/slope) -1)6100 formula. The melting curve was obtained to confirm the specificity of the primers.
Analysis of gene expression stability
The software program used to calculate the expression stability of reference candidate genes was geNorm [12]. The geNorm calculates the average expression stability value (M) with a standard deviation between the logarithmically transformed expression rates. This M value is the average pairwise variation of one particular gene compared to all the other tested genes. This program recommends using an M below the threshold of 1.5 to identify the reference genes. The geNorm also estimates the pairwise variation value (Vn/n +1), by allowing the identification of the optimal number of reference genes to be used. Pair-wise variation values with a threshold #0.15 are considered sufficient for normalization, although this limit should not be seen as a very narrow cut-off point. Thus, the geNorm indicates the number of genes necessary for normalization through normalization factors (NF n ) or providing the geometric means of combining the most stable reference genes: the two most stable genes (NF 2 ), the three most stable genes (NF 3 ), etc. Two different data sets were put on the geNorm to assess the expression stability of the candidate reference genes: the first dataset comprised quantities from individual Ct values; and the second dataset comprised quantities from pooled Ct values. In the first geNorm analysis, the large number Ct values from independent replicates in each cervical tissue condition were taken into account. Thus, assuming that 65 samples -CIN1 (12), CIN2 (6), CIN3 (14), cancer (14), normal (19) (1962). It should be noted that these two types of data (individual Ct values and pooled Ct values) were applied to the mRNA genes (GAPDH, ACTB, EEF1A1, RPLPO) and npcRNA genes (RNU-6, miR-23a, miR-191), in two independent analyzes conducted by means of geNorm.
Validation of reference genes in cervical tissues
In order to validate the most stable genes recommended by geNorm as suitable reference genes for normalization of qPCR data in cervical carcinogenesis, two targets were evaluated. Thus, the two most stable genes, the three most stable genes and the two least stable genes were used to normalize the expression levels of the chosen targets (p16 INK4a and miR-203) for each sample (in the same batch of cDNA). The p16 primer pair (F_ACATCCCC-GATTGAAAGAACC; R_ATGAAAACTA CGAAAGCGGGG) was designed on the basis of the GenBank data (ID: 1029) with the aid of CLCBioMain Workbench 5.7.1 software. The primers for miR-203 amplification was purchased from the miScript primer assay (Qiagen) and the miScript PCR Starter Kit (Qiagen), which contains the miScript Universal Primer. Moreover, with the purpose to demonstrate the effect of using a single reference gene on the target expression, we assessed the relative expression of p16 INK4 obtained by each of the four single reference genes (GAPDH, ACTB, EEF1A1, RPLPO), and the relative expression of miR-203 obtained by each of the three single reference genes (RNU-6, miR-23a, miR-191). In employing this normalization strategy, the linear scale expression quantities of the reference genes obtained from the individual Ct values, as well as from the pooled Ct values, were directly used to calculate the relative quantification of the targets. The Ct values of each target were not pooled and each biological replicate was kept independent (65 biological samples 62 PCR repeats = 130 Ct values per target) since our objective was only to assess the effect of the pooled Ct values for the reference genes.
Statistical analysis
A statistical analysis was conducted by making use of two kinds of software: R (version 3.1.0) and GraphPad Prism (version 6.0). Before the geNorm analysis the D9Agostino-Pearson normality test was carried out to determine the distribution of the data. A one-way analysis of variance (ANOVA) was employed to compare the relative quantities of the p16 INK4 and miR-203 targets across all the cervical tissue conditions. The Bonferroni correction was used to correct P values. The P value ,0.05 was considered as statistically significant.
Determination of RNA quality and qPCR efficiency
The RNA concentrations from cervical tissues were suitable and ranged from 200-3000 ng/ml, depending on the size of the biopsy (25-100 mg). All these RNA samples were checked for purity and integrity. The value of the purity ranged from 1.8 to 2.0, in accordance with the absorbance ratio at 260/280 nm. The integrity was visualized by the presence of intact 28 S and 18 S ribosomal subunits on electrophoresis gel. Thus, all the RNA samples included in this study were reliable and were representative of the evaluated tissues.
The qPCR efficiency was determined for each primer pair by using the slope of a linear regression model ( Figure S1). All the PCR primer pairs showed correlation coefficients of R2 = 0.99 and primer efficiency values (E) ranging from 0.99 to 1.00 (Table 4). The specificity of all the primer pairs was confirmed by a single peak in the melting curve ( Figure S2).
Expression ranges of candidate normalizers in cervical tissues
In this study we evaluated the expression pattern of the most commonly used reference genes (protein coding genes and nonprotein coding genes) for qPCR assays in cancer research. As can be seen in Figure 1
Determination of the most stable reference genes
The stability ranking and the best combination of genes that could be used as a normalizer, were provided by geNorm after conducting an analysis involving individual Ct values (Table 5) as well as the pooled Ct values ( Table 6). All the candidate reference genes showed average expression stability values (M) below the threshold of 1.5 intragroup, as recommended by geNorm. GAPDH and ACTB were recommended as the most stable genes followed by EEF1A1 and RPLPO from an analysis involving individual Ct values (Table 5). Of all the npcRNA genes, miR-191 was found to be the most stable, followed by miR-23a. RNU6 was revealed to be the least stable gene ( Table 5). The ranking of the mRNA genes (as well as the best combination of genes) was altered in this second analysis which involved pooled Ct values (Table 6). GAPDH remained the most stable gene, but EEFA1 became the second most stable. However, no alteration was observed in the stability ranking of the npcRNA genes ( Table 6).
Validation of reference genes for measuring mRNA expression in cervical tissues
For validation purpose, the relative quantification of the p16 INK4a target was assessed by using a combination of the two most stable genes, the three most stable genes and the two least stable genes. Target gene expression was normalized through a stability ranking of the genes based on an analysis of individual Ct values (Figure 3a), as well as an analysis involving pooled Ct values (Figure 3c). The overexpression of p16 INK4a has been linked to the Quantifying mRNA and MicroRNA with qPCR in Cervical Carcinogenesis PLOS ONE | www.plosone.org severity of premalignant lesions, i.e. there is a greater expression of this protein in CIN2 and CIN3 (which corresponds to a highgrade squamous intraepithelial lesions-HSIL) than in CIN1 which corresponds to a low-grade squamous intraepithelial lesions-LSIL [32,33]. In view of this, p16 INK4a expression levels across cervical tissues were reproduced more effectively by normalizations with a combination of the two or three most stable genes (GAPDH and ACTB; GAPDH, ACTB and EEF1A1) obtained from an analysis of individual Ct values ( Figure 3a). Conversely, the p16 INK4a profiling based on the normalization factors (NF) from the pooled Ct values varied, and showed similar expression levels between CIN1 (LSIL) and CIN2 (HSIL) (Figure 3c). In addition, as can be seen in Figure 3a and Figure 3c, the use of the two least stable genes resulted in a discrepant expression profile of p16 INK4a . The relative quantification of p16 INK4a was also assessed; this used a single reference gene that took account of individual Ct values (Figure 3b), as well as pooled Ct values (Figure 3d). In both types of analysis, the results showed that the use of a single reference gene can lead to discrepancies. In Figure 3b, it can be observed that the use of GAPDH and ACTB resulted in a similar expression pattern to p16 INK4a when this is compared with the combination of both genes (GAPDH and ACTB) from an analysis of individual Ct values (Figure 3a). However, the expression levels of p16 INK4a were higher in all instances (Figure 3b and Figure 3d) than the normalization carried out by the combined genes provided by the geNorm (Figure 3a and Figure 3c).
The relative quantification of the p16 INK4 transcript in all the cervical tissue conditions was better represented by using the two most stable genes (GAPDH and ACTB) based on the analysis with individual Ct values (Figure 3a). Thus, in Figure 4a significant differences are shown between this normalization strategy and a strategy involving each single gene, which also takes account of the individual Ct values (Figure 3b). As previously demonstrated, the combination of the two or three most stable genes from the individual Ct values did not change the expression profile of p16 INK4a . Hence, the target gene expression does not significantly differ if two reference genes are used rather than three (Figure 4a). Conversely, the use of a single reference gene significantly differs when compared to the normalization based on the two and three most stable genes (Figure 4a). In view of this, the combined use of GAPDH and ACTB for the normalization of the target expression significantly reduced the magnitude of error when compared with the use of a single gene.
Additionally, in Figure 4b the normalized expression levels of p16 INK4a are demonstrated through a combination of GAPDH and ACTB (based on the analysis involving individual Ct values). It should be noted that p16 INK4a has shown a significant overexpression in all the cervical tissue conditions, except between CIN2 and CIN3.
Validation of reference genes for measuring microRNA expression in cervical tissues
The validation of the best normalizers for measuring miRNA expression across cervical tissues was based on a combination of the most stable npcRNA genes obtained by the geNorm analysis, as well as each single gene. The two stability rankings provided by the geNorm (with individual Ct values and pooled Ct values) were identical. However, the use of combined genes (from both analyses) to normalize the relative quantification of miR-203, has reflected variations in its expression profile throughout all the cervical carcinogenesis (Figure 5a and Figure 5c). In a similar way to the results obtained for the measurement of the p16 INK4a expression levels; miR-203 demonstrated an expected pattern of 'downregulated' expression across cervical tissues [15,25,34,35], when the two most stable genes (miR-191 and miR-23a) were used that originated from the geNorm analysis with individual Ct values (Figure 5a). In Figure 5c a suggested 'upregulation' can be observed in the target expression from the normal tissue to CIN1 when the two or three most stable genes are used (based on an analysis involving pooled Ct values). An analysis was also conducted with each candidate gene as a single normalizer using individual Ct values (Figure 5b) as well as pooled Ct values (Figure 5d). Wider discrepancies in the miR-203 expression profile were observed when each candidate gene was used as a single reference. For example, the use of RNU6 as a normalizer resulted in a discrepant expression profile with elevated expression of miR-203 in normal tissue and CIN2 when individual Ct values were employed (Figure 5b), as well as in normal and CIN1, when pooled Ct values were employed (Figure 5d).
Given the factors outlined above, the combination of miR-191 and miR-23a based on the analysis with individual Ct values, was suggested as the best normalizer for measuring miRNA expression in cervical tissues because this provided a better representation of the miR-203 expression profile across cervical carcinogenesis. Nevertheless, no significant difference was observed in the effects on normalization between the use of two or three genes ( Figure 6a). However, target gene expression differs significantly when a single reference gene is used (that takes account of individual Ct values) when compared with the normalization obtained by the two or three most stable genes (Figure 6a). Figure 6b shows normalized expression levels of miR-203 across cervical tissue conditions by the combination of miR-191 and miR-23a obtained from the individual Ct analysis. Interestingly, no significant differences were detected in the expression levels of the target between all the tissues, despite the fact that miR-203 has been reported to be downregulated across cervical carcinogenesis.
Discussion
Studies of gene expression profile in cervical tissues (normal and neoplastic) have been performed to find biomarkers for cervical cancer [23,36]. Some of these studies showed discrepant results, e.g. in more recent studies of miRNA expression in cervical carcinogenesis [15,25,37]. Some authors suggest that these discrepancies can be attributed to the different platforms and methods employed and the diversity of the samples of the control groups, perhaps due to ethnic variability [15,37]. However, the use of unsuitable reference genes seems to be one of the reasons for the differences in the results obtained in qPCR studies [16,38]. The importance of choosing a correct standardization strategy has already been emphasized, both in the qPCR analysis of mRNA [21], and in the miRNA profiles [22].
To the best of our knowledge, the present study is the first to perform a geNorm analysis with a set of samples which meet all of the cervical tissue conditions: Normal + CIN1 + CIN2 + CIN3 + Cancer. This strategy enables reference genes to be used for the identification of potential biomarkers, not only in normal and cancer, but also in premalignant lesions. In view of the examples of failure in the current tests for screening premalignant cervical lesions [3][4][5][6], these biomarkers could be useful to distinguish between CIN1 and CIN2, CIN2 and CIN3, and cancer as well as in providing more information about the severity and progression of these lesions. Additionally, we have chosen commonly used reference genes for qPCR studies of mRNA [19,23,29,30] and miRNA expression levels in cervical cancer [15,20,22,31] with the aim of validating them in our specific experimental design, as recommended by the MIQE guidelines [10]. The determination of gene expression stabilities was performed with the aid of geNorm software, first developed by Vandesompele, et al. in 2002 [12], and since then widely adopted to evaluate the expression stability of the candidate reference genes [39,40].
In addition, to our knowledge, this is the first time that a study has evaluated the effects of pooling Ct values across replicates (by simulating a pool of samples) on an expression stability analysis, as well as on the qPCR results with regard to cervical tissues. We have proposed to make a comparison between the use of pooled Ct values and individual Ct values (by treating samples as independent) in a qPCR assay, since pooling samples (or RNA) is an alternative method that reduces the costs incurred by a qPCR assay. In this regard, we provide evidence that the use of pooled Ct values is not a strategy to obtain valid qPCR data that is as reliable as a strategy that employs individual Ct values. This evidence was obtained by validating reference genes for measuring mRNA and miRNA expression in cervical tissues. The p16 INK4 expression profile across all the cervical tissue conditions was represented better by means of the two most stable genes (GAPDH and ACTB) obtained from the geNorm analysis with the individual Ct values. This strategy has provided an expression profile of p16 INK4a that corresponds to that of other studies, which have found a greater expression of p16 protein in HSIL than in LSIL [9,32,33]. Similarly, miR-203 displayed a suggestive pattern of downregulated expression [15,25,34,35], with the use of the two most stable genes (miR-191 and miR-23a) derived from the geNorm analysis with the individual Ct values. The use of the two or three most stable genes from an analysis that involve pooled Ct values, is able to cause discrepancies in the expression profiles of both p16 INK4a and miR-203. Thus, the results clearly suggest that pooled Ct values can lead to a misinterpretation of the qPCR data. Some studies have included a pool of samples (or RNA pool) in the qPCR assays to reduce biological variability and also to reduce the costs of the experiments [14][15][16][17]. However, we suggest that the effect of including pooled samples should be evaluated for each experiment which involves qPCR assays, since our study shows that the maintenance of inter-individual variability is a key factor which can ensure the reliability of the qPCR data in cervical tissues.
To date, only two studies have investigated a correct strategy for data normalization in qPCR assays for cervical tissues; one of these studies recommended reference genes for mRNA expression studies [19], and the other for microRNA expression studies [20]. According to Shen et al. [19] EEF1A1 was the most stable gene (followed by GAPDH and RPLP0) for mRNA quantification in human cervical tissues. Interestingly, in the same work, ACTB was found to be the least stable gene in cervical tissues, in contrast with our study where ACTB was the second most stable gene. This variation may be linked to the geNorm analysis carried out in our study, which includes all the cervical tissue conditions, as well as the use of different platforms by the laboratories, or else it may be due to ethnic variability. In this way, this data strengthens the need to validate reference genes in a specific experimental setting. With regard to non-protein coding genes, the combined use of miR-191 and miR-23a, based on the analysis of individual Ct values, was the best normalizer for measuring the miRNA target expression in accordance with Shen et al. [20]. The use of these two most stable genes reflected the profile of miR-203 across cervical carcinogenesis that was most expected, even though expression levels between cervical tissues have no significance, perhaps due to the small size of the sample [41,42]. Thus, larger sample sizes are required to obtain valid conclusions.
Even though it has been well established that a normalization strategy is an essential component in ensuring the reliability of the qPCR data, and that reference genes must be validated for each particular experimental setting [10,11,21,22]; a large number of studies of cervical cancer continue to use the most well-known reference genes, on the basis of previous studies and without proper validation, or without mentioning whether this stage has been carried out accurately [23][24][25][26][27]. For this reason, in evaluating the effects of using a single reference gene to normalize the target expression, we have assessed the relative expression of p16 INK4 obtained by each of the four single reference genes (GAPDH, ACTB, EEF1A1, RPLPO), and the relative expression of miR-203 obtained by each of the three single reference genes (RNU6, miR-23a, miR-191). The relative expression of both the targets (p16 INK4 and miR203), which were normalized by each single reference gene, did not show exactly the same pattern as the expression profile obtained by the most suitable normalizer observed in this study: the combination of the two most stable genes provided by the geNorm analysis, involved individual Ct values. The results suggested that the use of a single reference gene can lead to discrepancies in the qPCR data, which is in agreement with the findings of other studies [12,43,44]. Apart from this, Shen et al. 2010 [19] recommends EEF1A1 as the most stable gene that can be used as a single reference gene for normalization in gene Figure 6. Effect of normalization options on miR-203 expression in cervical tissues. In a), it is shown that there is no significant effect on normalization between the use of the two most stable genes and the three genes. The differences in the expression levels of miR-203 were statistically significant with ANOVA when the use of the two and three most stable genes were compared with the use of each gene as a single normalizer. Graph b), shows the relative quantification of miR-203 in cervical tissues using a combination of miR-191 and miR-23a (from an analysis involving individual Ct values) as the normalizer. The error bars indicate a 95% confidence interval; **, p,0.01; ****, p,0.0001. doi:10.1371/journal.pone.0111021.g006 profiling studies involving human cervical tissues. In our study, the use of EEF1A1 as a single normalizer led to an erroneous normalization up to 3.0-fold when compared with the normalization where two or three of the reference genes are used together (as shown in Figure 4a). Our findings are in agreement with those of Vandesompele et al. [12] where it is stated that ''a conventional normalization strategy based on a single housekeeping gene leads to erroneous normalization up to 3.0-and 6.4-fold in 25% and 10% of the cases, respectively, with sporadic cases showing error values above 20''. In this way, it should be stressed that the suitability of reference genes in some studies does not necessarily apply to others. Other authors also recommend the use of at least two reference genes for human tissues [43,44]. However, even though it has been widely accepted that one of the best ways to normalize the qPCR data is to use at least 2 to 3 reference genes, several studies of cervical cancer continue to use the most wellknown reference genes such as GAPDH [23,45], ACTB [30], EEF1A1 [46] and RNU6 [15,[24][25][26][27][46][47][48][49][50][51], as a single reference gene and without mentioning whether this stage has been performed accurately. Furthermore, even though it has been established that a normalization standard must reflect the quantity and size of the target of interest to obtain comparable samples [12,22]; some studies have used an mRNA as a reference gene to normalize the miRNA expression levels in cervical cancer [25,51]. The use of an unvalidated or single reference gene in qPCR remains a recurring problem that has been addressed and critically discussed in recent papers [38,40,52].
Conclusion
An increasing number of publications have used qPCR to identify differentially expressed messenger RNAs as well as microRNAs between several types of tissues and cells, in various biological conditions or experimental situations. qPCR has become the most widely used technique in these studies due to its simplicity and that fact that it can provide results quickly. However, careful standardization of each stage is of crucial importance to obtain accurate data, such as the inclusion of validated reference genes. In this study, we performed the validation of reference genes for mRNA and miRNA quantification in cervical carcinogenesis. We have made a serious attempt to evaluate the effects of important issues in qPCR assay to ensure accurate data, since the main purpose of our line of research is to identify changes in mRNA and miRNA expression which have a real significance in cervical carcinogenesis. It should be underlined that the suitability of reference genes in some studies does not necessarily apply to others and that the use of a single reference gene is not sufficient to obtain reliable qPCR data; even though several studies continue to employ this methodology in qPCR studies on cancer research. It is worth noting that the best combination of reference genes which can be used for the measurement of targets in this study was selected after a comparison had been made between the use of individual Ct values and pooled Ct values. The results clearly showed that pooled Ct values can lead to an unreliable results, which suggests that studies on cancer research by means of a qPCR assay, should take into account the individuality of each biological sample. Finally, we believe that this study raises important issues and points to the need for further research that is not confined to the area of cervical cancer, but also leads to the question of the qPCR assay. Figure S1 Real-time PCR standard curve of all primer pairs. The slope of the standard curves indicates the efficiency of qPCR. (DOCX) Figure S2 Melting peaks of all primer pairs. The specificity of all the primer pairs was confirmed by a single peak in the melting curve. (DOC) | 7,898.8 | 2014-11-03T00:00:00.000 | [
"Biology"
] |
WAYFINDING AND AUGMENTED REALITY: APP FOR OUTDOOR EXPERIMENTS IN THE PERUGIA STATION AREA
: This research aims to improve wayfinding in the Fontivegge district of Perugia, a chaotic and disorienting area due to the numerous redevelopment projects. The goal is the development of an Augmented Reality application to improve the urban orientation experience. The app will provide users with indications on how to reach pre-selected places of interest, through the visualisation of directional arrows placed horizontally along the route and vertically at major turning points. In addition, it will provide a number of thematically categorised infopoints, which will accompany the user along the route, enriching it with information. The aim is, therefore, to create a more positive and engaging orientation experience for users and to promote a sense of belonging and social cohesion in the place.
INTRODUCTION
The purpose of this research is to improve the wayfinding strategies through the use of an Augmented Reality smartphone application designed for the railway station area of the city of Perugia, the urban hub of modal interchange between public transport services.The district of Fontivegge, an historic and cardinal place of the city of Perugia, due to its degradation, has been affected by various redevelopment and rehabilitation interventions (Figure 1).The main one is represented by the "periphery plan" (Piano, 2015) that was developed from the collaboration between the Department of Civil and Environmental Engineering and the Municipality of Perugia (Bianconi et al., 2020b;Bianconi and Filippucci, 2018a).Through altering perceptions (Bianconi et al., 2021;Bianconi and Filippucci, 2019;Maffei, 2007) of the area covered by the plan, the aim is to mend the relationship between the community and its places (Appadurai, 1996;Bauman, 2000), associating public space with a community wellbeing service (Burry and Burry, 2012;Gehl and Gemzøe, 2003;Kim and Kaplan, 2004;Steg et al., 2013).The choice of the field of investigation related to the problems of the area, identified in the neighborhood can be traced to the difficulty of orientation encountered by users, proven in research previously conducted (Bianconi et al., 2020b(Bianconi et al., , 2018;;F. Bianconi et al., 2023;Fabio Bianconi et al., 2023;Bianconi and Filippucci, 2017); the chaotic nature of the area emerges and there appear to be numerous nodes of urban flows such as the railway, bus stop and minimetro, which are poorly connected and difficult to identify.
In this path of urban regeneration, the research of wayfinding is included, which can be likened to the process of orientation in a place, as it literally means "finding one's way" (Barnard, 1998).This term is historically attributed to Kevin Lynch, who uses it in his famous work "The image of the city" (Lynch, 1960) to indicate that then very innovative approach of mixing architecture, urbanism, semiotics and psychology, issues that are connected to the cognition of space, to the transformation of the images experienced in a scheme correlated to the urban form and ascribable to the two-dimensionality of drawing (Alexander, 1964;Bridgman, 1959;Venturi, 1967;Wolbers et al., 2008).During a period of great transformation of western cities, such as the years after the Second World War, the reading of urban space and, in particular, the legibility of the complexity experienced as the foundation of contemporaneity (Alexander, 1964;Bridgman, 1959;Empire, 1955;Venturi, 1967), opened up a new research theme, namely the revitalisation of the increasingly structural relations between the built environment and images (Jencks and Baird, 1969).The totality of sensations and what is perceived is, in fact, re-elaborated in a design where orientation and identification are two essential aspects of the abstract processes of our mind, ascribable in any case to representative, immaterial, virtual acts (Mirzoeff, 1999;Picazo et al., 2020).At the centre is the value of images, because "it should be borne in mind that man is a predominantly visual animal.More than 50% of the neurons in his brain respond to this sensory input" (Maffei, 2007).
Emphasis is on the interpretation that wayfinding is a section of "navigation", which should be placed in parallel with "locomotion", the act of moving along a path.Wayfinding represents that process of actions implemented in order to solve the problem of identifying the route to be followed.The information anxiety of our culture (Wurman, 1989), exalted by the disruptive digitalisation, and in particular, the continuous accessibility guaranteed by smartphones, activating an addiction to their use (Matar Boumosleh and Jaalouk, 2017), poses this issue as a purely contemporary theme: spatial neophobia is a common feeling, more or less felt, that is connected to the survival instinct, to the necessary attention that comes from discovering what one does not already know.The study of the cognitive aspects of wayfinding must premise the evaluation of the impacts of an environment on the people who experience it and the mechanisms by which they move in spatiality, since loss of orientation creates anxiety and increases stress levels in users (Chías and Fernández-trapa, 2022).The sense of disorientation and bewilderment experienced by users lead to an alteration in the physiological and psychological state of the user who experiences the place, involuntarily associating it with a negative experience and leading, consequently, to the belittling of the area and its subsequent emptying (Goldhagen, 2017).
Wayfinding is based on the relationship between dots and lines, stasis and movement, focussing on spatial memory, to be understood (Siegel and White, 1975) as the set of remembrances of images, paths and relationships, which are recognised and reconstructed in a figurative process (Bechtel and Churchman, 2002).The idea behind the project is the development of an Augmented Reality application, capable of explicating the digital language of wayfinding, so that the user can be informed, first of all, of points of interest in the proximity and, subsequently, assisted in reaching them.Outdoor wayfinding is a little-covered topic, while there are numerous case studies of implementations on indoor locations, such as airports (Lampazz et al., 2020), schools (Cibilić et al., 2020) and hospitals (Basri and Sulaiman, 2013;Drewlow et al., 2022;Prodi and Stocchetti, 1990).To date, GPS data in augmented reality has been used for on-screen positioning of information popups that indicate to the user the direction they should follow to reach the destination, as well as the distance to be traveled, while, in this case, the approach aims to accompany the user throughout the journey by providing continuous information.A step forward from the studies conducted so far is to be performed, proposing the accompaniment of the user along the entire path and additionally providing timely and valuable information on the various critical situations of the route to be taken.
METODOLOGY
The goal of the research consists in reversing this paradigm, thereby creating favorable experiences that can generate positive emotional memories; these promote a sense of belonging and, secondarily, a social cohesion associated with experiencing the place.A cross-media strategy is chosen to define the wayfinding structure, which offers integration between different communicative approaches: physical, currently being designed, and digital; a transversality is sought that extends communication to the user and allows a clearer and broader reading of the place.
The research aims to use the new devices in the service of wayfinding, setting a standardized methodological approach through the implementation of an augmented reality application.
The smartphone and Augmented Reality become the tools for a new approach to urban exploration.
The application was developed using the Unity graphics engine, preferred over the competition for its easy integration with Google services and, in particular, with the ARCore development kit.
The American computer company recently announced a new tool for developers who want to create and launch Augmented Reality (AR) experiences in real-world locations.Once the development environment was configured, through the installation of the necessary plug-ins within the graphics engine, the user experience was programmed.First of all, a number of information points were inserted inside the identified area, which, once reached by the user, allow the display of in-depth information tabs.To do this, the new anchoring functionality made available by the ARCore kit was used, associating the object to be displayed during the augmented experience with a suitably configured script with real geographical coordinates.
The infopoints were subdivided according to three themes (historical-cultural, mobility, commercial) through the insertion of a specially created script that allows the category to be associated with the object, allowing it to be displayed in the correct real position only if the category is activated in the created interface.
The same script manages the colouring of the infopoint's graphic indicator, which varies according to the selected category, and also controls the assignment of the information pop-up that can be displayed via that specific information point (Figure 2, Figure 3).Another remarkable feature in the development of the application is the navigation to the remarkable places identified.Through a specially programmed menu, it is possible to select an item, which is linked to coordinates (Figure 4).Arrows have been used to represent the graphical indications useful for reaching the selected place.These will be positioned on the ground along the route and vertically at notable turning points.For positioning purposes, the most convenient routes for reaching the places made available were previously created on Unity and, by means of a script, when navigation is activated, the route considered closest to the user's position at that time is displayed (Figure 5).The methodology applied is easily replicable to other case studies and also scalable to more complex and extended situations.The aspects to focus on are the battery consumption of the devices and the quality of the GPS signal, which is not always optimal.
RESULTS
The results of the research are inherent in the implementation of the app for orientation within the Perugia Station area.The application, when first opened, displays what is visible through the rear camera of the smartphone in use, with the overlay of the user interface intentionally minimal in order to emphasise the augmented view.
On the screen, in fact, there are only two buttons that allow the opening of two menus for selecting the information to be shown.At the bottom right is a button that activates the visualisation of the menu of places of interest, the selection of which activates the display of directions to reach them.At the bottom left is the menu for choosing the infopoints that the user wishes to activate or deactivate according to the categories of interest.The infopoints are represented in the augmented scene through the classic 'pin' graphic element, reproduced in 2D with automatic rotation to be always perpendicular to the user's gaze.
Once reached, the application exits AR mode and displays an information popup across the screen, with a graphic composition containing data of interest to the user and appropriately decorated with chromatic elements coordinated with the category of the infopoint.For the historical-cultural category red was associated as the representative colour of the city of Perugia (Figure 6), for mobility blue was chosen because the graphics of the city's main transport companies use these shades (Figure 7).For the last category, that is commercial activities, purple was chosen, simply to have as much contrast as possible with the others, excluding green from the possible choices, which is not very visible in urban contexts with natural elements.The deactivation of the augmented mode was a design choice aimed at avoiding distractions when reading the information texts provided and can be resumed by closing the panel using the X-shaped button at the top right of the graphic.As for the 2D elements indicating the path to the user, a graphic composed of a double arrow coloured with two different colours, dark red and orange, was selected to guarantee the visibility of these elements in overlapping with every possible scenario.Thus, the goal of the app is not only to take the user to their destination but also to engage and allow them to explore and investigate the potential of the place they are in.During the journey, the user will be accompanied by the app through the display of specific remarkable points to allow the user to deepen their knowledge of the place.
CONCLUSION
The research conducted focuses on the representational issues inherent in the interaction between urban users and places.It is a proposal that supports seeing, intended not only as a functional process but as a cultural act, in the aim of requalifying the place by identifying the reasons for the development of the area (Bianconi et al., 2022) and searching for a juxtaposition of signs to improve identification and orientation (Filippucci, 2012), thus legibility (Lynch, 1960;Sancar, 1986;Schultz, 1987;Smardon, 1988) and consequently the accessibility of places (Calori, Chris et al., 2015;Devlin, 2014;Passini, 1981;Symonds, 2017).The great challenge to be addressed goes beyond the boundaries of transport objectives inherent in orientation research, but proposes processes of re-appropriation of places through the rediscovery of the qualities of place, which is the subject of an important process of territorial regeneration (Carr, 1992;Castells, 2008;Gehl, 2007;Molinari, 2021;Purini, 2021).This is the context for the studies and experiments carried out over the last five years, aimed at creating added value through the rediscovery of places.The process of reconstructing historical evolution through the realisation of immersive models of spaces erased from collective memory becomes the key to understanding the current urban form (Bianconi et al., 2022;F. Bianconi et al., 2023), transforming the immaterial into the material (Bianconi et al., 2020a;Bianconi and Filippucci, 2018b).In the hypothesis of enhancing the image culture of our era, this path was supported by digital representation, as the history of the original architecture planned for the station was rediscovered and subsequently reconstructed three-dimensionally to ensure its exploration through Virtual Reality (Bianconi et al., 2022).The same Virtual Reality has been made usable through smartphones, with which greater interactivity is guaranteed through a 360degree view.This path highlights the desire to create wayfinding in a process that operatively wants to place the relationship between places and those who live them at the centre, exploiting the logic of serious games as tools for creating empathy relationships.(F.Bianconi et al., 2023;Dominici et al., n.d.;Ioannides et al., 2016;Kuliga et al., 2015;Larson, 2020;Meng and Zhang, 2012;Mortara et al., 2014;Theodoropoulos and Antoniou, 2022;Wilson and Soranzo, 2015).
The studies described allow us to constestualise the value of this research in the cultural proposal system, setting as a goal the development of an application that detects the gps position in real time and, through augmented reality, projects orientation information and additional data on the architecture present, thus improving the wayfinding of the study area.In fact, the app not only wants to take the user to their destination but also to engage and allow them to explore and investigate the potential of the place they are in.The usability of the app suggested some potential future implementations, such as exploiting user profiling to suggest routes to follow in relation to physical abilities, personal tastes or needs of the moment, thus improving urban navigation, and thus wayfinding.
Figure 1 .
Figure 1.Current state of the Fontivegge station affected by requalification works.
Figure 3 .
Figure 3. Screenshot of the infopoint pop-up menu.
Figure 4 .
Figure 4. Screenshot of the informational route pop-up menu.
Figure 5 .
Figure 5. Screenshot of the directional arrow.
Figure 6 .
Figure 6.Development of the historical-cultural pop-up inside Unity platform.
Figure 7 .
Figure 7. Screenshot of the pop-up about minimetrò. | 3,504.2 | 2023-12-13T00:00:00.000 | [
"Computer Science",
"Geography",
"Environmental Science",
"Engineering"
] |
A Component Architecture for the Internet of Things
In this paper, we describe a component-based software architecture for the Internet of Things in which proxies for Things and services that we call “accessors” interact with one another under a concurrent, time-stamped, discrete-event (DE) semantics. These proxies are analogous to web pages, which proxy a cloud-based service such as a bank, but instead of being designed to interface those services with humans, accessors are designed to interface services and Things with other services and Things. A deterministic DE semantics is combined with a widely used pattern for handling network interactions that we call asynchronous atomic callbacks (AACs). AAC enables many concurrent pending requests to be active at once without blocking and without the treacherous concurrency pitfalls of threads. In effect, our architecture combines AAC with actors where the actor model has been endowed with a temporal semantics. We show how this architecture can leverage the previously reported secure swarm toolkit (SST) to achieve stateof- the-art authentication, authorization, and encryption of interactions across networks.
I. INTRODUCTION
The Internet of Things (IoT) is the class of cyberphysical systems (CPS) that leverage Internet technology for interactions between the physical world and the cyber world.The vision embodied by IoT appeals to the imagination of many-our environment and virtually anything in it will turn "smart" by having otherwise ordinary things be furnished with sensors, actuators, and networking capability, so that we can patch these things together and have them be orchestrated by sophisticated feedback and control mechanisms.Supported by Wegner's argument that interaction is more powerful than algorithms [45], Lohstroh and Lee [30] point out that interaction indeed opens up limitless possibilities for Things to harness their environment and compensate for a lack of self-sufficient cleverness; sensors aside, a connection to the Internet alone allows a Thing to tap into an exceedingly rich environmentunleashing a real potential for making things smarter.
Ensuring safety, reliability, privacy, and security of systems that rely on open networks is extremely challenging.There is precedent, however, for high confidence systems that use open networks.Today, the world's financial system operates almost entirely electronically and with heavy use of the open Internet.No engineered system is perfect, but the benefits appear to outweigh the risks, and losses due to technical failures and malicious actors are simply factored into the cost of operation.Can cyberphysical systems achieve the same balance, where the benefits of open networks outweigh the costs?
The web focuses on the interaction between people and information or services hosted on servers or in the cloud.The IoT, on the other hand, will emphasize interactions between Things and Things, Things and cloud services, and Things and people, rather than people with cloud services.In this paper, we describe a design pattern that we call accessors 1 , where an accessor is like a web page for a Thing or service, but instead of being designed for humans to interact with it, it is designed for other Things and services to interact with it.
Accessors are based on an adapted actor model, where an accessor is a parameterized actor with input ports and output ports through which timed events stream.The accessor serves as a proxy for a Thing or service that may be local or remote.This proxy is analogous to the proxying of a service that occurs in your web browser when you download HTML and JavaScript from a web server.The browser instantiates a proxy for a remote service, such as your bank.The proxy runs on the local host, your computer, but interacts with a remote service using mechanisms that are largely invisible and irrelevant to you, the user of the proxy.
The input ports and parameters of an accessor are analogous to form boxes on web pages, and output ports are analogous to rendered pages.But form boxes are designed for human input, and rendered pages for human consumption.The input and output ports of accessors are designed for interaction with other Things and services.The accessor itself provides functionality analogous to the scripts that a web page runs in the browser, which communicate in proprietary ways with the (possibly remote) Thing or service.A critical part of our model is that the host environment that executes the accessor must have standardized capabilities, just as browsers today (mostly) support common languages (HTML and JavaScript) with a common set of bindings that the program can use (the Window object and the XMLHttpRequest object, for example).This enables web designers to design a proxy that will work in any browser.Analogously, accessors are designed to execute in a variety of hosts, ranging from deeply embedded processors to cloud servers.
We also describe a design environment called CapeCode 2 that can be used to compose accessors to create services which can then further be proxied by accessors.CapeCode is, in effect, a computer-aided design tool for IoT applications.It facilitates debugging and design-space exploration by providing a friendly graphical environment that includes all of the analytical tools of Ptolemy II, on which it is based.
Our focus in this paper will be to show how the particular actor model by which accessors interact with one another, a timed discrete-event model, matches well the requirements of IoT applications.It provides a measure of determinism that helps to counter the chaos of unpredictable latencies and unreliable networks that is intrinsic to applications that are distributed on the open Internet.In particular, its deterministic semantics enables well-defined test cases, rigorous specifications, and reliable error checking.Deterministic semantics means that there is a well-defined notion of "correct behavior," and that behavior is repeatable.Our semantics also enables a more deterministic use of timing by replacing best-effort timeouts with a model of time that has a semantic notion of simultaneity and well-defined ordering of events.Finally, we show how accessors can leverage edge computing to improve security, privacy, predictability, and robustness to network outages.
II. MOTIVATING EXAMPLE
Consider a device, such as a tablet, smart phone, or augmented reality (AR) goggles, that has a camera, an interactive screen, an audio speaker, and a microphone.Consider an app on this device that invokes an image processing service to recognize Things seen by its camera and then overlays the Things in the display with current sensor data and any interactive controls provided by the Thing.Fig. 1 shows such an overlay in what might be a mechanical room of the future.Consider further that the device can respond to voice commands to scroll through a suite of recognized Things in the field of view or turn on and off the overlay display.Such an app would be useful, for example, for factory floor inspection, equipment maintenance, configuring smart conference rooms, and myriad other applications.
All of the technology exists today to build such an app, and indeed similar systems are familiar to those working in the field of augmented reality.But anyone familiar with the technologies involved will realize that the complexity is considerable and that the result is likely to be brittle.Very likely, a realization today will be a stovepiped solution, where every component is entirely under the control of a single vendor.But making something that is open, for example something that is able to discover devices from new vendors in the local environment or to leverage machine learning that integrates data from outside sources will be extremely challenging.
Fig. 2 shows a prototype of such a system built using accessors in the CapeCode design environment.Our prototype constructs the overlay display shown in Fig. 1, a user interface for a Thing detected in the local environment.The app tracks movement as the camera pans over the scene.As motivation for subsequent discussion, we walk you through what the components of this prototype do.Hopefully, the reader will see readily that this is, in fact, a prime example of the use of reusable components in a platform-based design.It will also lend insight into the reasons for our choice of using a discreteevent semantics to govern the interactions between accessors.
First, the three boxes in the figure whose icons contain "JS" encapsulate small, simple scripts that are specific to this app.These are the only components in the design that are not intended to be reusable.They are specific to this app.In our prototype, they are written in a few lines of JavaScript using the same framework used to create the reusable components, which we call "accessors."All remaining components, those with icons not containing "JS," are accessors and are intended to be reusable.
Beginning at the left of the top row of icons, the Camera accessor provides access to a hardware device: a camera, connected to the host computer that runs this app.That accessor outputs a stream of images and has parameters for controlling the frame rate, resolution, and selection of camera, in case there is more than one camera on the host.It can alternatively provide access to a network-connected camera, in which case the structure of the app does not change at all.Only the parameters change.
To the upper right, receiving the stream of images, is an ObjectRecognizer accessor.This can use any of a variety of technologies to recognize Things in images.In our prototype, we simply assume that Things are labeled with AR tags, which are like simplified QR codes that are easier for cameras to recognize at a distance.Three AR tags can be seen in Fig. 1.More elaborate technologies could identify objects by their visual appearance, with the help of a challengeresponse interaction, leveraging indoor localization and device telemetry, or with the help of a discovery service such as the Summon app [46].
The script labeled "TagToAccessor" receives from the Ob-jectRecognizer an array of zero or more IDs for AR tags found Fig. 2. Augmented reality application for interacting with sensors and actuators.This composition of accessors generates the interface shown in Fig. 1 in each image frame along with the X-Y position of the tag in the field of view.Based on the "index" input (which comes from the lower loop of icons, discussed below), it selects one of these tags and looks up an accessor for a Thing associated with the tag's ID.In the simplest case, TagToAccessor could perform a table lookup to match a tag with an accessor for a Thing.A more advanced TagToAccessor implementation would use a location-based discovery service to dynamically obtain the tag to accessor matching for the user's current environment and update the matching when new AR tags are deployed.The accessor itself could come from a web service or from a local edge computer that is aware of devices in the local environment.TagToAccessor feeds that accessor to an accessor labeled "Mutable." Mutable is perhaps the most interesting component here.It accepts as input an accessor, which it checks for compatibility with its ports, and if it is compatible, reifies the accessor and delegates handling of streaming inputs and outputs to it.The input to Mutable is the source code for an accessor that it instantiates and begins executing.Note that since this source code can be downloaded at the time of instantiation, it can be assumed to be up-to-date and compatible with the current version of the Thing's API, which itself may be periodically updated, for example, to patch for security vulnerabilities.
The Mutable accessor can be seen as an abstract interface specification for candidate accessors.The reified accessor effectively replaces the Mutable accessor, taking its place in the block diagram.If later a new accessor appears, it will be reified and will replace the previously reified accessor, which will be shut down.In this case, Mutable expects the provided accessor to have an input port named "control" and two output ports, "data" and "schema."All three ports are typed to handle JSON formatted data.It can also accept accessors that partially match, for example omitting the control input port, as we explain below.The schema output provides the app with a specification of what is expected on the control input port.
That specification is used by the ConstructUI script to build an HTML table with input boxes as shown in Fig. 1.The UserInterface accessor uses any resource on the local host that can display HTML5.On a laptop computer this could be a browser, whereas on a mobile device it could be a service provided by the operating system or app development framework such as Apache Cordova.
The reified accessor in Mutable is a proxy running on the local host that represents the Thing identified in the camera's field of view.Depending on the capabilities of the host, the reified accessor may communicate with the Thing via Bluetooth, Wi-Fi, ZigBee, or any other technology supported by both the host and the Thing.The actual communication mechanism and protocols can be proprietary to the Thing and its accessor, just as the details of the communication between a web page and a bank are specific to the bank.
The design pattern here is similar to what has proven so effective in the web.To make browsers talk to banks, for example, the world's banks could have gotten together and established a standard messaging protocol that would include specifications for messages to view balances, make payments, etc.But this is not what happened.Instead, the designers of browsers standardized an execution environment within which a proxy for the bank, in the form of downloaded JavaScript code, could execute.Analogously, the makers of Things could get together to establish standard messaging protocols, for example to turn on lights or establish a temperature setpoint.But that strategy is not likely to work very well; it will be foiled by the very richness of the IoT ecosystem and its dynamically evolving nature.Our design pattern is inspired by what has worked so well in the web.We standardize on the execution environment for proxies for Things.
Once the Thing's accessor has been reified, it begins producing output data.Notice here a subtle but important point.The diagram in Fig. 2 is not a simple flowchart nor a simple dataflow diagram.Were it either of these, the Mutable accessor block would produce one datum on its output in reaction to each accessor and/or control input.But this is not what it does.Each time it receives a new unique accessor input, it reifies the accessor, and that accessor begins to produce outputs that can be either spontaneous or reactions to inputs.If the outputs are spontaneous, they occur at some rate determined by the device and will continue to emerge from the Mutable accessor until a new accessor input is received or the device itself stops providing data.
The sensor data emerging from the Mutable accessor streams right to the ConstructUI script actor.In our prototype, this script produces HTML and CSS code to be overlaid on the image, as shown in Fig. 1.This HTML code includes input elements that can be used to send control data back to the accessor that reifies Mutable.In Fig. 1, the "step size" and "sampling period" controls are fed back to the Thing's accessor whenever the user taps or clicks on the Submit button.
The loop at the bottom illustrates the integration of entirely disjoint technologies into the app in order to get voice control.This could be used, for example, to scroll between tags in the event that multiple AR tags are detected in the image.Beginning at the bottom left, the "VoicePromptGenerator and Tag Selector" script starts off by outputting text which then gets converted to an audio signal using a text-to-speech service.This could be a cloud-based service such as those provided by Google or Amazon or a locally realized service.The resulting audio data is fed to an AudioPlayer accessor, which provides access to the local audio hardware.Having produced a voice prompt, the AudioCapture accessor is triggered, which listens for a response.The resulting audio signal is sent to a SpeechToText accessor, which again could use a cloud-based service or a local one.In our prototype, we then further process the resulting text using a natural language engine, in our case the one provided by Google at API.AI, which can convert natural language into specified fulfillment commands.This is fed back to the leftmost script, which can produce a new prompt and/or an index to select a new tag in the image.This part of the app could, for example, state, "I found three devices.Is this the device you want?"If the user says "no" or "not really" (the natural language processor would handle such variability) it could scroll to the next one and state "How about this one?"
III. THE ACCESSOR PATTERN
An early version of the accessor pattern is given by Latronico et al. [24], who explain them using a diagram similar to Fig. 3.The upper part of the figure shows a "swarmlet," which is a composition of components called "actors" connected by streams.The middle actor is an accessor and serves as a proxy for a "swarm service or thing."The proxy runs on the accessor host and communicates with the service or Thing by some proprietary mechanism.Lohstroh and Lee [30] distinguish the interaction between the proxy and its service or Thing, which conforms to a "vertical contract," from the interaction between the accessor and the other actors in the swarmlet, which conforms to a "horizontal contract."The horizontal contract enables composition of multi-vendor services and Things, whereas the vertical contract enables device and hostspecific interaction mechanisms, using for example any of a variety of radio or networking technologies and protocols such as HTTP, CoAP, TCP sockets, or WebSockets.
The host that runs the accessor may be a microcontroller, mobile device, edge computer, or server, whereas a Thing it interacts with is typically a separate piece of hardware, not necessarily proximate to the host.Or the accessor can access a cloud-based service.
A simple accessor definition is shown in Fig. 4. The "setup" function defines two ports, an untyped "trigger" input and JSON-typed "data" output.The accessor then declares that it requires an "http-client" module, which must be provided by the host if the host is to be compatible with this accessor.The host, moreover, can restrict the sorts of HTTP requests that the module supports, as done for example by the "sameorigin" policy in browsers, which enhances security of web pages.The "initialize" function adds an input handler for the "trigger" input.Whenever an event arrives on this input, this handler function will be invoked.On line 13, the input handler function uses the required module to make an HTTP request.It provides to that module a callback function that will parse the response and send it to the "data" output port.Of course, a better designed accessor should check for errors and handle timeouts, but we hope this is enough to get a feel for how an accessor is specified.
In this example, inputs at the "trigger" port cause an output to be produced sometime later.These inputs and outputs interact with other accessors or actors according to the horizontal contract, which is based on actors [17], [1], but with a more deterministic temporal semantics similar to that used in discrete-event simulators.We explain this in the next section.
IV. COORDINATION
Coordination can be thought of as constrained interaction [44].Much like type systems, which can prevent programs from "going wrong" [35], coordination can make programs satisfy certain desirable properties, such as determinism, liveness, and fairness, by construction.Importantly, disciplined coordination reduces the need for burdensome validation and testing.A coordination model or language [38] implements coordination rules that endow a coordinated ensemble with a semantics, often called a "model of computation" (MoC).For accessors, the host realizes the MoC.
The MoC is common for all accessors, but the mechanism by which the accessor interacts with a Thing or service is not.That mechanism can be built on top of established standards, such as HTTP, MQTT, datagrams, WebSockets, etc., but there is too much diversity among devices and services to reasonably expect common details to emerge.Hence, an accessor can be thought of as an adapter that translates between two incompatible protocols.Such adapters have been used in other design frameworks for heterogeneous systems such as Metropolis [10].Of course, within specific application domains, such as Internet-controlled lighting, manufacturers could benefit from establishing local standards.This would enable, for example, the same accessor to work with products from multiple vendors.But nothing about our approach requires the establishment of such standards.Our approach is therefore friendlier to innovation, multi-vendor compositions, and new entrants into markets.As illustrated by the example in Fig. 2, IoT applications tend to be highly concurrent, so concurrency needs to figure prominently in the MoC.In that application, concurrency appears in two ways.First, the actors in the model are concurrent in that they can (conceptually or physically) execute at the same time modulo data dependencies.The ObjectRecognizer and the AudioPlayer, for example, are concurrent with no data dependencies between them.Second, several of these actors spawn remote actions on Things or cloud-based services and there can be several pending actions awaiting responses all active at the same time.The ObjectRecognizer, TextToSpeech, Speech-ToText, and NaturalLanguage accessors, for example, may all use cloud-based services with RESTful APIs.Notice that neither of these forms of concurrency is about performance, about speeding the execution of software.Instead, concurrency is intrinsic in the distributed nature of IoT applications and the interaction of software with Things.If the application designer attempts to manage this concurrency using threads, chaos is likely to ensue, with unexpected nondeterminism and deadlocks a constant threat.The accessors framework instead provides a much more structured concurrent MoC, which we describe in this section.The key idea is to combine a discrete-event model of computation with asynchronous atomic callbacks.
A. Discrete-Event Systems
The diagram in Fig. 2 is an executable model given in a graphical syntax.Each icon represents a reactive piece of software, either an accessor from a reusable library or an application-specific script.We call these pieces of software "actors."Each actor has parameters and input and output ports.The "wires" connecting ports convey messages from one actor to another.In our semantics, these messages are events occurring at a logical time.An event has a time stamp, and the execution semantics ensures that an actor sees input events in time-stamp order.Moreover, if events with the same time stamp are received on multiple ports, our semantics ensures that the actor can see all such events in the same reaction.We thereby avoid a form of nondeterminism in which simultaneous events (those with identical time stamps) may be processed in nondeterministic order, as they would be in a more classical actor model.
Such discrete event (DE) systems have a long history and have been used primarily for simulation [31], [8], [15], [25], [7], [47].In our case, we are using DE semantics not for simulation, but for run-time deployment, as has been done in Ptides [48], [13] and Spanner [9].Ptides and Spanner extend the time-stamp semantics across networks so that time stamps have a global meaning in a distributed system.Accessors are compatible with Ptides, in principle, and therefore the semantics of DE can extend across a network to coordinate actions on several distributed hosts.Such an extension, however, is beyond the scope of this paper.We focus instead on how DE is combined with the highly asynchronous actions of Internet interactions.
As with any framework, there is overhead associated with abstraction.A DE scheduler needs to maintain a list of pending events to be processed sorted in time-stamp order.The core part of our implementation, which is shared among all hosts, supports defining, instantiating, connecting, and executing accessors and comprises only approximately 3000 lines of heavily-commented JavaScript code.
B. Asynchronous Atomic Callbacks
The Asynchronous atomic callbacks (AAC) concurrency pattern is used extensively in web programming, both on the server side (using for example Node.js 3 and Vert.x 4 ) and on the client side, in browsers.On the server side, it has proven scalable to very large numbers of clients and servers.It has also been used in some other (non-web) applications such as parallel computing (e.g.Active Messages [43]) and embedded systems (e.g.TinyOS [28]).
A central feature of this pattern, which drives its scalability, is its dependence on a functional style of programming, where functions are first-class objects in the language.Functions are invoked asynchronously, typically when some request that has long and/or variable latency has been satisfied.For example, the callback function on line 15 in Fig. 4 will be invoked when a response from an HTTP server has been received.This callback function is passed as an argument to the module's get() function.This non-blocking behavior prevents programs from becoming unresponsive while waiting for responses from remote servers.
Importantly, every such asynchronously invoked function invocation is atomic with respect to every other function invocation; that is, a callback function invocation waits until no other function is being executed before beginning, and the callback function executes to completion before any other function can begin executing.This atomicity distinguishes the AAC concurrency model from interrupt-driven I/O, threads, and many asynchronous remote procedure call mechanisms.The same benefits can, in principle, be accomplished with threads, but the resulting programs are much less scalable, more difficult to understand, and vulnerable to the many nefarious bugs that multithreaded programs inevitably have [26], including unexpected nondeterminism and deadlocks.Properly written AAC applications do not use locks explicitly and cannot deadlock.
AAC comes with costs, however.First, it becomes essential to write code carefully to consist only of quick, small function invocations.A long-running function will block all callback functions, reducing the responsiveness of applications and compromising real-time performance.Second, AAC accentuates the chaos of asynchrony, where achieving coordinated action can become challenging.For example, if you make multiple requests in sequence to a service, each time passing a callback function, there is no assurance that the callbacks will be invoked in the same order as the requests.Both problems are important for IoT, where heavy computation may be required to analyze sensor data, and coordinated physical actions may be dependent on the order in which things occur.
Because of these limitations, several alternatives mix AAC with other concurrency models.Many JavaScript implementations realize a thread-like mechanism called a Web Worker, which runs tasks in the background concurrently with the main AAC function invocations.Unlike threads, these Web Workers cannot share data with the main application.Instead, they send messages to the main application, which, if it is listening, will invoke a callback to handle the message.ECMAScript 6, a recent version of JavaScript, enriches AAC with a cooperative multitasking model, which allows a function to suspend execution at well-defined points, allowing other functions to be invoked while it waits for some event.The Vert.x framework enriches AAC with so-called "verticles" (think "particles"), which can execute in parallel while preserving atomicity.Verticles can interact with one another through a publish-and-subscribe bus or through shared but immutable data structures.All of these extensions can, in principle, be used with accessors.But the combination of AAC with DE makes them much less necessary.
C. Combining DE with AAC
DE provides a streaming data model, suitable for many IoT applications, augmented with time stamps for improved determinism.AAC provides a mechanism for handling highly asynchronous delayed events.Both models are concurrent, are more disciplined than threads, and have achieved widespread acceptance.Accessors combine the two, getting the best of both worlds.
Consider again the simple accessor in Fig. 4. A trigger input that it receives occurs at a logical time and the output data occurs distinctly later than that logical time.Line 5 declares the output to be "spontaneous," which means that it does not have an immediate direct dependency on the input.The accessor host uses this information when analyzing data dependencies between actors to come up with a schedule that reacts deterministically to any particular set of time-stamped inputs.
On line 16, a callback function produces an output.This occurs asynchronously, but the AAC model ensures that it occurs atomically.No other actor can be in the middle of executing anything.This event can therefore be assigned a time stamp without risk of violating DE semantics by having events appear "in the past" or by having simultaneous events (those with identical time stamps) appear at some destination in two distinct reactions.The actual time stamp that is assigned to this "data" output is nondeterministic because it depends on the reaction time of some remote server and on network latencies.But once the time stamp is assigned, how the swarmlet reacts to this event is completely deterministic.This semantics makes behaviors more repeatable and testable.For example, we can write regression tests that check behavior given particular time-stamp assignments.The determinism of the model ensures that there is exactly one "correct" reaction to a set of time-stamped inputs.
The combination of DE with AAC was first realized by the authors by embedding a JavaScript interpreter in an actor and coordinating it with the DE director of Ptolemy II [40], which is implemented in Java.Lohstroh and Lee formalized this combination using interface automata [30].
D. Timing
Time stamps in a DE semantics are a logical concept, not a physical one.But the use of time stamps suggests a connection with the physical world.Indeed, in the IoT, physical timing of events can be quite important.
The most straightforward connection we can make between time stamps and physical time is to attempt to align them as much as possible.For example, the time stamp assigned to the asynchronous, spontaneous output on line 16 of Fig. 4 may be taken from a physical clock on the host, and that clock may be synchronized with other clocks on the network.For this to be valid, the host needs to maintain a correspondence between the logical notion of "current time" and the time recorded on the physical clock.A simple way to do that is to delay handling of time-stamped events until the physical clock of the host reaches or exceeds the time of the time stamp.In a distributed system, clock drift will have to be taken into account as done for example in Ptides [48], [13] and Spanner [9].
Accessors will also need mechanisms to invoke actions in the future at specified logical times.These are similar conceptually to the callbacks in Fig. 4, but these future events will be assigned time stamps deterministically.
To get timed behavior, most AAC frameworks support delayed callbacks.For example, most JavaScript environments provide a setInterval(F, T ) function, where function F is to be invoked after T milliseconds and then again periodically with intervals of T milliseconds.Of course, the actual time of the function invocations cannot be exactly every T milliseconds, since that would require a perfect timekeeper, which does not exist, and it would require that the JavaScript engine be idle at every multiple of T milliseconds, since the AAC model requires atomicity.We expect (and get) some jitter in the actual timing of the function invocations.Such jitter is unavoidable in any software platform.
But the situation is worse because the time T actually has very little meaning at all.It is interpreted in the JavaScript language as a suggestive guideline to invoke the function at some time near the multiples of T milliseconds.When there are multiple such delayed callbacks, there are no guarantees on the order of invocation of the callbacks even if the time intervals are identical or related by integer multiples.
In [18], two of the authors (Jerad and Lee) show that these mechanisms can be given a stronger temporal semantics.For example, it is possible to ensure that if two calls to setInterval(F, T ) and setInterval(G, T ) are made with the same T , then the host can ensure that F and G are invoked together atomically and hence will appear to any observer as being simultaneous.Moreover, Jerad and Lee define labeled logical clock domains (LLCDs), within which islands of synchrony can be created asynchronously and coexist with a clean semantics.
These timing mechanisms have been integrated into our framework.They can be used, for example, to extend the AR application in Fig. 2 with timed behavior, for example, to synchronize video and audio feedback to the user.Multiple clock domains could become useful in more complex applications where several concurrent islands of synchrony coexist.
Many callbacks, however, are untimed, like the ones in Fig. 4. Since callbacks are all atomic, during the invocation of the callback, each logical clock will have a specific "current time" that is frozen during the invocation of the callback.This makes it possible for such an asynchronous callback to then make a request for a timed callback that will be invoked logically simultaneously with other actions on the same logical clock domain.Referring again to the example in Fig. 2, this could enable audio stimuli to be synchronized with visual stimulus even if the onset of events that trigger these stimuli is asynchronous.
Ultimately, no strategy can guarantee that a timing goal is met.If a Thing fails, it fails!Our contribution is enabling detectability.A predictable, composable timing semantics is necessary to detect abnormal timing variability.The order of events is well-defined by time stamps, and any anomalous order that emerges at runtime is an indicator of a fault.
In addition, accessors are designed to be able to be run locally, close to the Things they interact with, in contrast with cloud-based services.This makes latencies more predictable, repeatable, and controllable.Some functions, however, are much easier to provide in the cloud than locally.Fig. 2 uses cloud services for speech synthesis and recognition and natural-language processing.For fault tolerance, we can use a "local first" architecture to treat cloud services as an enhancement instead of a requirement.It would be easy, for example, to augment Fig. 2 with a fallback user-interaction mechanism, such as pushbuttons on the screen.This means that the application can reliably deliver real-time behavior even if Internet connectivity is lost altogether.
E. Context Sensitivity
The Mutable accessor in Fig. 2 provides an example of a context-sensitive swarmlet, where the actual behavior depends on the Things that are locally available.Such mutation, if done in an ad hoc way, would amount to self-modifying code, which is usually not a good idea.Self-modifying code is notoriously difficult to understand and has even been found suitable as a code obfuscation technique [32].The mechanism in the accessors framework is much more disciplined.
The Mutable accessor is a placeholder for an accessor that can reify it.Before being reified, the Mutable accessor has no functionality.It ignores input events and produces no output events, except for input events on the "accessor" input.In our implementation, when the "accessor" input receives an event, the Mutable accessor interprets this input event as a request to reify an accessor that is specified by the value of that event.The value of the event could be text similar to what is shown in Fig. 4. If the Mutable accessor has already reified an accessor, then it unreifies it and then reifies the new one.In the application in Fig. 2, this can be used to provide entirely different visual interfaces to Things in the field of view.
Note that, in principle, this dynamic substitution mechanism could be leveraged to optimize for locality or availability, as done in the example of Fig. 2, but also to cope with unexpected network outages, or for keeping up with firmware updates or unpredictable changes to remote APIs.A new accessor could be downloaded and instantiated at run-time to replace one that (for whatever reason) no longer optimally performs its function.
Discipline is important, however.The Mutable accessor has a specific role in this swarmlet, and not all accessors can satisfy this role.The role is specified by its input and output ports, "control" and "data."These ports have types, and the reified accessor must have matching ports that conform to those types.A perfect match, however, is not required.We follow the type refinement schema for actors similar to that of Lee and Seshia [27, chapter 14].An output data type of a reified accessor, for example, can be a subtype of the type of the corresponding output of the Mutable accessor.Conversely, an input type of the reified accessor can be a supertype of the corresponding input port of the Mutable accessor.In addition, the reified accessor need not match all the input and output ports present in the Mutable accessor.Any output port that is present in the reified accessor but not in the Mutable accessor will have its events ignored, and any input present in the Mutable accessor but not in the reified accessor will not be receiving events.Of course, a useful reification will have at least some input ports that match.
Dynamically reified accessors may be downloaded from the Internet as part of a discovery process.Hence, as we will discuss in Section VI, the accessor to reify is likely to not be completely trusted.Much as a browser controls the local resources that a web page can access, our hosts control the resources that the reified accessor can access.All access to resources is meditated by modules, like the http-client module that is required in Fig. 4. The module is implemented by the host and hence can be constrained in any way appropriate.
F. Hierarchy
The model in Fig. 2 is an instance of what we call a composite accessor.In that example, the composite accessor itself has no input and output ports, so it cannot be directly embedded in another swarmlet.But our accessor framework supports composite accessors with input and output ports, so models can be constructed hierarchically.
Even more interestingly, the swarmlet in Fig. 2 interacts with outside services through the network, for example by making HTTP requests.Those outside services could themselves be swarmlets, and they may have embedded within them an accessor designed for accessing the services of the swarmlet in Fig. 2.
This schema is illustrated in Fig. 5.In that figure, two networked hosts have each instantiated a swarmlet containing an accessor for the other swarmlet.When the accessor on host A receives an input event, it sends a message to the accessor on host B, which then produces an output event.The swarmlet on host B constructs a response and provides that response as input event for the accessor, which sends a message back to host A. Finally, the accessor on host A produces an output event with the response.This mechanism can be used to construct services that can then be easily instantiated remotely; the service (a swarmlet) provides an accessor that another swarmlet can instantiate.
Of course, once such peer-to-peer interactions exist, a new form of brittleness appears.One piece of a distributed application may be updated, for example, without being able to simultaneously update the other pieces.Some sort of coordinated deployment and update will have to be developed.
V. A PLATFORM FOR COMPOSING THINGS
Accessors are generic reusable components that can be composed in a common semantic domain with an actor-based discrete-event semantics furnished by a host implementation.As such, the host can be thought of as a platform in the sense of Platform-Based Design (PBD) [41].The key goal of PBD is to separate functionality (the what) from architecture (the how) and be able to map a design (or parts of it) onto different architectures without having to change the design.
Platforms abound in IoT.A typical philosophy is to offer an application programming ecosystem deployed on a certain type of host, such as Node.js or a centralized cloud service.An application facilitates communication among Things using information streams, which can be acted on directly or scanned for events.Generally, the focus has been on supporting diverse host-to-Thing connections, with some success.An application developed on a particular platform is usually not transferable to another platform.This paradigm works well for a set of Things owned by the same entity and a community substantial enough to afford its own application designers.
Looking to the future, it is desirable to write an application once and deploy it on any host connected to the right Things.How should this application be written?JavaScript is an attractive candidate due to its widespread usage and compatibility with heterogeneous hosts, such as web browsers and Node.js.
One underlying problem is that JavaScript host environments differ, particularly in their mechanisms for allocating compute resources for large computations, providing permanent data storage, and global variable management.Pure JavaScript provides no such mechanisms, and hence, when such mechanisms are provided by a host, they are often provided in a host-specific way.Typically, these mechanisms are implemented in the host's native language, such as C, C++, or Java, and then provided to the JavaScript programs through modules that must be explicitly "required" by the program.
To solve this, the Accessor approach leverages an information hiding strategy.Accessor code has no direct access to platform-specific primitives.Instead, an accessor may declare dependencies on functionality contained in a host-provided module, like the http-client module in Fig. 4. A module ideally has a common API for all hosts but may have a host-specific implementation.Thus, the host implementation details are hidden from the application developer.
An accessor can execute on any host that meets all its module dependencies.The binding between an accessor and the module(s) it relies on takes place upon instantiation of the accessor on a host.For instance, the Camera accessor in Fig. 1 can run on various hosts without assuming anything about the mechanisms used to access the camera other than what is defined in the camera module's API.
The ability to determine whether a host supports an accessor at run time provides advantages for IoT applications.Some Things execute on energy and cost-constrained leaf nodes, and therefore it may be desirable to reject accessors that push the system over budget.Accessor rejection may also be a security strategy.
Ideally, since accessors may contain untrusted code, it is preferable to execute them in a contained environment.The most straightforward way to do that is with a language interpreter or a virtual machine that provides a sandbox.Browsers, for example, already execute JavaScript in such a sandbox, and our browser host, described below, takes advantage of this extra measure of security.
Browsers and server-side infrastructure such as Node.js and Vert.x provide powerful JavaScript interpreters, but they are not lightweight enough for installation in many leaf devices.The situation may improve in the future, as a number of small JavaScript kernels have appeared recently in open-source form, such as Duktape.Some of these can execute without much operating system support, and hence may be suitable for deployment in quite constrained environments.In the remainder of this section, we outline the hosts we have prototyped, thereby demonstrating that the accessor architecture is deployable on a large variety of platforms.
A. CapeCode
The CapeCode host is a Ptolemy II configuration that provides a GUI for composing, executing, and deploying composite accessors.Ptolemy II is a Java-based open-source software laboratory that supports experimenting with actororiented design [14].CapeCode provides a graphical user interface for building swarmlets.The name of this host is derived from Cape Cod, Massachusetts, where much of the development has occurred.
Models that consist solely of accessors may be code generated into composite accessors that may be executed by any accessor host that implements the modules used by the accessors.CapeCode can execute the generated composite accessors either locally or deploy them remotely on the Node host, which we discuss in the next section.These composite accessors can also be loaded and executed in the Browser host by embedding a reference in a web page.Developing composite accessors using the CapeCode GUI and then deploying them to possibly less powerful remote machines is an effective development strategy.Further, CapeCode can execute models combining accessors with other Ptolemy II actors, thereby providing a rich library of predefined actors.
A drawback of using a graphical syntax to express a more complex or elaborate design is that it tends to result in exceeding the Deutsch limit [12] of fifty graphical elements, which can make the model unwieldy and difficult to understand.Ptolemy II uses hierarchy to decompose a design into comprehensible pieces.To scale up to very large numbers of actors that are widely distributed in networks, we can consider application builders like Chisel [5], which provides programmed construction of digital circuits.It leverages higherorder functions featured in the Scala programming language.In Chisel, rather than directly instantiating and connecting components, a designer writes a program that instantiates and connects components when it executes.
B. Node
The Node host is meant for use with Node.js.Node.js is a popular cross-platform server-side JavaScript runtime environment.Because of the popularity of Node.js, we anticipate that the Node host will have the greatest impact of all of the hosts.In particular, we feel that providing well-defined temporal semantics to the Node environment could increase robustness and reliability.
The Node host is available via the Node Package Manager (npm) as the @terraswarm/accessors module.
C. Browser
The Browser host allows users to inspect and execute accessors in a web browser.Any web page may load an accessor by including a <script> tag pointing to the Browser host script plus a reference to the accessor(s).It is assumed that some web server is available to provide these files.A demonstration server is available publicly 5 , and it features an interactive tutorial 6 for writing accessors that executes in the browser host.
The JavaScript engine in a modern browser is designed to safely execute third-party code in the local environment.Therefore, the Browser host does not typically have direct access to the local file system or to machine hardware, leaving it slightly underpowered compared to other hosts.One advantage of the Browser host is ease of deployability.Any device with a web browser can download and execute accessors simply by pointing to a URL.
An example of the Browser host is an accessor that uses a JavaScript implementation of OpenCV [42] to recognize faces in an image 7 .
D. Cordova
Smartphones provide access to a wealth of sensors, offer an uplink to a LAN or the Internet, and are mobile, which means that their environment is subject to change as they are carried around by their owners.These three aspects combined make smartphones a very appealing deployment platform.
Apache Cordova8 provides a development toolchain amenable to a variety of smartphone operating systems.Apps are constructed much like an ordinary web page, using HTML, CSS, and JavaScript, and can be deployed to different targets, including Apple iOS and Google Android.Superior to a browser-hosted web application, Cordova's plugins expose platform-dependent functionality such as geolocation and a file system to the application's JavaScript environment.Hence, unlike the browser host, which can also be deployed on mobile platforms, the Cordova host can selectively bypass the browser's security restrictions, giving access to platformspecific functionality.
Apache Cordova is itself a platform-based design tool, where Cordova plugins offer a uniform API while hiding Android-specific or iOS-specific implementation details.This points to the scalability of PBD, as the PBD approach supports a series of platform mappings; here, from accessors to Cordova, then Cordova to Android or iOS.
E. Duktape
The Duktape accessor host uses the Duktape JavaScript engine 9 to deploy accessors on small embedded systems.
As a proof of concept, we deployed a composite accessor to a Maxim Integrated MAX32630, which is a Cortex-M4 with 512K RAM and 2Mb flash.Our simple accessor was an accessor that produces integers connected to a display accessor.The executable had the following sizes in bytes: • text: 291,848 -program code in flash • data: 2,964 -initialized data in RAM • bss: 8400 -uninitialized data in RAM This shows that accessors can be deployed on deeply embedded platforms.
To use the Duktape host on a composite accessor would require implementing in C/C++ the modules used by the accessors.For example, to support the accessor in Fig. 4 requires implementing the http-client module, which would most likely be implemented in C or C++ using low-level networking primitives.
VI. SECURITY
Accessors are untrusted code that serve as proxies for sensors, actuators, and services.Inspired by the web, accessors are therefore executed in a virtualized environment that controls access to resources and data.Such encapsulation provides a starting point for ensuring security and privacy, but it is not sufficient by itself.In particular, the execution environment will have to grant access to physical resources such as sensors and actuators in order to realize IoT applications.How should it authenticate the IoT applications (e.g., whether an application running remotely is modified from the original program components)?Moreover, how can we make sure we only grant permission to legitimate IoT applications to access certain resources (authorization or access control)?
While it is crucial to provide appropriate authentication and authorization for the IoT applications, it is also challenging to do so.Many IoT applications can run on a variety of software/hardware platforms, possibly in a distributed fashion.Some of them run under a constrained energy budget or with restricted communication capability.It would be unreasonable to expect those types of applications to incorporate traditional security measures for the web, such as SSL/TLS (Secure Socket Layer/Transport Layer Security), based on a Public Key Infrastructure (PKI), because it requires a frequent use of power-hungry public-key cryptography.Leveraging the Kerberos authentication system [37] is also difficult because it requires a constant stable connection to the authentication server and an interactive prompt for users to enter passwords.Passwords are not appropriate for Thing-to-Thing interaction.
In addition, IoT applications tend to operate in open (or even hostile) environments and thus are at higher risk of being compromised or subverted.As an example of this, Ghena et al. [16] demonstrated an attack on traffic controllers on the streets of Ann Arbor, Michigan, by leveraging access to wireless communication used by traffic controllers.Therefore, it is sorely important to monitor the behavior of IoT applications and revoke access to safety-critical resources as soon as a security breach has been detected.Due to the scale of IoT applications, in terms of both the number of applications and the volume of data traffic, it is not a feasible strategy to solely rely on digital certificates because a PKI with tens of billions of certificates will quickly become unmanageable.
Our open-source toolkit, SST (Secure Swarm Toolkit) [20], provides a set of accessors for bringing authentication and authorization to the IoT while addressing above-mentioned challenges.SST uses a local authorization entity called Auth [21] deployed on edge computing devices that act as local gateways to the Internet for IoT applications.SST employs a locally centralized, globally distributed [19] approach, which has two key benefits: (1) dependency on a reliable connection to Cloud servers is limited, which improves robustness to network failures, and (2) better scalability can be achieved by distributing the workload for authentication and authorization among Auths.Various security configuration alternatives supported by SST also embrace heterogeneous security requirements and resource availability in the platforms for IoT applications.
Fig. 6 illustrates a part of an extended version of the augmented reality example in Fig. 2, secured by one of the accessors provided by SST, SecureCommClient.A stream of output data from the Mutable accessor is encrypted and sent to a cloud server via the SecureCommClient accessor.Let us assume there is another IoT application, namely SensorAnoma-lyDetector, running on a remote cloud server and programmed using another accessor in SST, called SecureCommServer.SensorAnomalyDetector takes streams of data from the distributed augmented reality applications reporting their sensor data, executes a machine learning algorithm on collected data, and sends feedback to the applications when any sensor data anomaly is detected.When a client application receives feedback on a detected anomaly from the server, the feedback is sent to the graphic overlay's additional input port, metadata, to indicate the anomaly as part of the overlay.
In this extended example, the main function of these security accessors is to access the authentication and authorization services required to establish a secure channel between two swarmlets, each of which are associated with their own (possibly different) Auth.It is assumed that a trust relationship between the involved Auths exists.After being authenticated and authorized by their respective Auths, SecureCommClient and SecureCommServer establish a secure communication channel between each other similar to that of a client-server connection via SSL/TLS.But with SST, we have the option to choose the underlying network protocols (e.g., TCP or UDP) or the cryptographic protocol.We also do not have to maintain a centralized certificate authority (CA).Adding SST accessors provides additional security guarantees including confidentiality and message authenticity, preventing network-based attackers from eavesdropping or staging a man-in-the-middle attack.In addition to the two aforementioned accessors, SST provides accessors for constructing IoT applications based on a publish-subscribe communication style using accessors called SecurePublisher and SecureSubscriber.
Another benefit of using SST accessors comes from encapsulation of cryptography operations and cryptographic key management.As Myers and Stylos [36] point out, design of APIs is critical for software security, especially in the sense that misuse of APIs can lead to serious security problems.With SST accessors, all software developers need to do is to specify configuration parameters and set up initial credentials (e.g., generate a public-private key pair and register the public key with an Auth).Even developers with moderate knowledge in security need not worry about internal cryptographic operations and encryption key management for accessors once the accessor design is correct.Specifying security configurations can be further simplified by using the given profiles as suggested in [20].
Combined with actor-oriented modeling semantics where actors communicate only through input and output ports, isolation of cryptographic keys and operations in SST accessors can enhance security when supported by OS-level or architecturelevel security mechanisms.By sandboxing the execution of SST accessors, a swarmlet host can restrict the privilege of accessors to read from or write to arbitrary files or network ports, preventing credentials from being leaked or being used maliciously (e.g., by an attacker to spoof the device).If the host is equipped with an architectural security mechanism such as Intel's SGX (Software Guard Extensions) [33], the credentials can be protected even when other processes on the host or the host's middleware or hardware components are compromised.Like other accessors, SST accessors also require some modules to be available on the host.These modules include the crypto module and the TCP or UDP modules.
VII. PRIVACY
Fundamentally, accessors are a disciplined form of mobile code.An accessor can be dynamically downloaded and instantiated on a wide variety of platforms, including deeply embedded ones.This has a potential advantage for preserving privacy because computation can be easily moved to a data source rather than the more common scenario where the data is fed into a cloud service for processing.
As Jaron Lanier explains in his book "Who Owns the Future?" [23], a key aspect of the business models of Silicon Valley tech companies that provide such cloud services (and typically do so free of charge) is the extraction and accumulation of information from and about its users.Furthermore, cloud services that make use of virtual resources such as those provided by Amazon AWS 10 or Microsoft Azure Cloud11 may not be designed to spy on its users, but can still be vulnerable to side channel attacks such as Meltdown [29] and Spectre [22] and potentially leak sensitive data.
A shift away from a centralized cloud-based paradigm towards mobile code and local computation could thus be an effective strategy to foster better privacy on the IoT.Accessors can reduce the need to transport data over the open Internet and allow for designs that carry out analysis locally and therefore rely less on the cloud, or they can anonymize data before sending it into the cloud for further processing.
VIII. RELATED WORK
A number of projects adopt an actor-oriented approach similar to ours for IoT system development.The closest are probably Calvin [39], Node-RED 12 , and NoFlo 13 , which use a dataflow concurrency model for interactions between services that are using AAC.Also reasonably close, although very different in syntax, is Rx (Reactive Extensions), which combine callbacks with streams [34].We believe that our use of discrete-event semantics is unique and offers a solid foundation for deploying timing-sensitive IoT applications.
Also related are publish-and-subscribe services such as MQTT, DDS, and ROS.MQTT, for Message Queue Telemetry Transport, is an ISO standard (ISO/IEC PRF 20922) intended for embedded applications where small footprint code is required and/or network bandwidth is limited.The Data Distribution Service (DDS) is an Object Management Group (OMG) machine-to-machine standard for real-time communication using publish-and-subscribe pattern.ROS, the Robot Operating System, an open-source software framework originated by Willow Garage, is widely used for building robot applications.These can be used by accessors for vertical communication with Things and services, and hence are complementary to our work.We have used MQTT in accessors for wireless sensornetwork devices and ROS in accessors that control robots.But similar to accessors, these services provide a communication fabric that can stitch together components.Unlike accessors, they use a publish-and-subscribe pattern, where a data producer generates messages for a "topic" and data consumers that are subscribed to the topic will be notified, typically by a centralized broker that mediates the communication.None of these use time stamps to provide a deterministic interaction semantics, however, so applications are less testable and harder to deploy in timing-sensitive scenarios.Time stamps help by controlling the order in which events are handled, thereby ensuring that given the same inputs, the application always delivers the same behavior.
IFTTT (IF This Then That) is a free web-based service originally created by Linden Tibbets and Jesse Tane that enables composing Things and services using a simple imperative style with chains of conditional statements.It is fundamentally a cloud-based approach, and it excels at integrating with other cloud-based services such as email, Facebook, and Pinterest.Accessors, in contrast, need not run in the cloud.An application built with accessors would usually run in a host that is much closer to the Things it is interacting with.The AR application in Fig. 2, for example, is designed to run on a phone, tablet, or head-mounted AR display that is in the same room as the Things it is interacting with.When feedback control is involved, local execution can be critical because latency can strongly affect behavior.
Accessors can, in fact, leverage the considerable ecosystem around IFTTT.Like the accessors in Fig. 2 that use cloudbased services for language and speech, accessors could be easily designed to interact with IFTTT.
In IFTTT, for the most part, each Thing requires its own "channel," the name for the mechanism that IFTTT uses to interact with the Thing.Accessors, similarly, can be created independently for each Thing.An application that is built with such "channels" or Thing-specific accessors will not work if the Thing is replaced by a variant from another manufacturer.However, as illustrated by the Mutable accessor in Fig. 2, accessors offer the intriguing possibility of more vendor-independent applications.The application in Fig. 2 will work with any device for which there is an accessor that provides JSON-formatted data and (optionally) accepts JSONformatted commands.There are many such devices.We need no prior agreement on what particular JSON structures are used nor even on what communication mechanism is used to talk to the Thing.The accessor may access the Thing via the Internet, Bluetooth, or ZigBee, for example, and it may use protocols such as HTTP, WebSockets, MQTT, TCP sockets, or datagrams, as long as the host provides modules supporting these technologies.IFTTT has no such mechanism for this level of vendor neutrality.
One approach to achieving vendor neutrality is represented by Home Connect, a protocol and a corresponding app from BSH Bosch and Siemens Hausgerate that controls multiple brands of home appliances.This is superficially similar to accessors in that it integrates Things from diverse vendors, but it is really quite different.First, it does not directly support Thing-to-Thing interaction, although its protocol is supported by IFTTT, and hence, using IFTTT, it is possible to build Thing-to-Thing interaction with Home Connect appliances.But more importantly, Home Connect standardizes the communication protocol between the Thing and the app.It requires the Thing to use a specific communication protocol, such as HTTP over the Internet, and it dictates the format that commands and data must have.Accessors, in contrast, standardize the interface between an accessor and its host, like web browsers, which standardize the interface between a downloaded JavaScript program and the computer on which the browser runs.Accessors do not standardize the mechanism that is used to communicate with the Thing.
Accessors, fundamentally, are stitched together by a coordination language, which has a syntax and a semantics.Recent years have seen an explosion of innovation in programming languages and programming models.New languages, such as Rust, Scala, Hack, Clojure, Julia, F#, Go, and Dart, and frameworks, such as Apache Spark, Microsoft Orleans [6], and Akka, codify programming models that manage parallel computing resources, scalable workloads, and/or long network latencies.A common thread in the new languages is to embrace elements of functional programming, particularly to make functions first-class objects in the language.Functional programming can be used to realize design patterns such as asynchronous atomic callbacks and structured parallelism such as map-reduce [11].A common thread in the frameworks (Spark, Orleans, Akka) is support for stream computation based on actors [17], [1].Our work overlaps with these by embracing functional programming and stream-based computation, but our work appears to be unique in its adoption of discrete-event semantics.
Another recent trend that pays more attention to timing is the focus on real-time data analytics.The IoT promises a flood of sensor data.Many organizations already are collecting but not effectively using vast amounts of data.The research and consulting firm Forrester defines "perishable insights" as "urgent business situations (risks and opportunities) that firms can only detect and act on at a moment's notice."Fraud detection for credit cards is one example of such perishable insights.This has a real-time constraint in the sense that once a fraudulent transaction is allowed, the damage is done.In CPS, a perishable insight may be, for example, a determination of whether to apply the brakes on a car, where a wrong or late decision can be quite destructive.
Real-time data analytics implies both timing constraints and streaming data.Computing on streaming data fundamentally means that you don't have all the data, but you have to deliver results.It differs from standard computation in that the data sets are unbounded, not just big, and you can't do random access on input data, which constrains the types of algorithms you can use.Major research efforts, such as the industryfunded RISELab launched at Berkeley in 2016 (Real-time Intelligent Secure Execution 14 , are getting a lot of attention.Examples of algorithmic innovations for real-time streaming data include adaptations of machine learning and optimization algorithms [2], [3] and adaptations of formal methods [4] to operate on streams.We believe that our component architecture could contribute quite a bit to such projects by integrating Things.
IX. CONCLUSION
IoT services are intrinsically an amalgam of heterogeneous and distributed components.It is naive to assume that any single standard will emerge for interaction between Things, services, and users.The accessors framework described in this paper provides a number of key properties not found (at least not all together) in any IoT framework that we know of today: • The use of proxies for Things, where proxies execute in a host-controlled environment, similar to web pages with scripts; • Embracing heterogeneity by not standardizing the means by which Things and services communicate with applications; • An actor-oriented streaming model of computation for interactions between components; • Time-stamped messages with deterministic interleaving semantics; • Use of functional programming concepts, particularly asynchronous atomic callbacks, to hide network latencies; • Deterministic timed interactions between components; and • Integration of state-of-the-art security including encrypted communication, authentication, and authorization.A great deal of work remains to be done.Most interesting to us is the possibility of leveraging the time-stamped interactions between components to build more deterministic distributed real-time applications.These could be based on the semantics of Ptides [48], [13] and Spanner [9], but also will require more support for dynamically changing component interactions and for large numbers of components.The mechanisms used in Orleans [6] look particularly promising, where a distributed registry of actor instantiations together with multiplexing of streams through host-to-host communication channels facilitates scalability to very large numbers of actors and hosts.
Fig. 6 .
Fig.6.Modified part of the example in Fig.2with a SecureCommClient accessor for additional security. | 14,711.2 | 2018-04-20T00:00:00.000 | [
"Computer Science",
"Engineering"
] |
Oxysterols in the Immune Response to Bacterial and Viral Infections
Oxidized cholesterols, the so-called oxysterols, are widely known to regulate cholesterol homeostasis. However, more recently oxysterols have emerged as important lipid mediators in the response to both bacterial and viral infections. This review summarizes our current knowledge of selected oxysterols and their receptors in the control of intracellular bacterial growth as well as viral entry into the host cell and viral replication. Lastly, we briefly discuss the potential of oxysterols and their receptors as drug targets for infectious and inflammatory diseases.
Introduction
Cholesterol is a ubiquitous sterol that is synthesized by mammalian cells and forms an essential component of the cell membrane [1]. It is involved in a variety of cellular processes, ranging from controlling membrane permeability and integrity to cellular signaling. In addition, cholesterol also serves as an important precursor for several molecules, such as steroid hormones, bile acid, vitamin D, and oxysterols [2]. Oxysterols are formed by the hydroxylation of cholesterol by enzymes or pro-oxidants and are important for regulating cholesterol homeostasis. In recent years, oxysterols have also emerged as important bioactive molecules in immunity and inflammation, with multiple immunoregulatory functions across several different cell types and organs. However, the actions and implications of oxysterols in the immune response to bacterial and viral pathogens are largely underexplored. Here, we review the literature and summarize the currently available knowledge on the role of oxysterols in the immune response to infectious pathogens of both bacterial and viral origin. Clear evidence exists that some oxysterols possess anti-microbial and anti-viral activities, mostly through the modulation of the host immune response. Lastly, we propose and provide evidence that this group of lipid mediators and their respective receptors may be targeted pharmacologically to improve the treatment outcomes of infectious diseases.
Types of Oxysterols and Their Receptors
Structurally similar to cholesterol but with additional oxygen-containing functional groups, oxysterols are oxidized cholesterols that can either be derived from the diet or produced endogenously [3]. Dietary sources of oxysterols include cholesterol-rich foods such as meats, eggs, and dairy products [4]. Endogenously, oxysterols are derived through the oxidation of cholesterol by enzymes or reactive oxygen species (ROS) (Figure 1) [5]. The majority of enzymes involved in the synthesis of oxysterols belong to the family of cytochrome P450 enzymes [3,5], with the exception of cholesterol 25 hydroxylase (CH25H), which belongs to the family of enzymes that utilize oxygen and diiron cofactors for hydroxylation [6].
(CH25H), which belongs to the family of enzymes that utilize oxygen and diiron cofactors for hydroxylation [6].
Figure 1.
Overview of the generation of selected oxysterols from cholesterol. Oxysterols are generated from cholesterol through oxidation by pro-oxidants or enzymes. The majority of these enzymes involved in the synthesis of oxysterols are from the family of cytochrome P450 enzymes [3,5], with the exception of cholesterol 25 hydroxylase (CH25H).
Oxysterols are present at low concentrations, up to 5-fold lower compared to cholesterol, in circulation [7], and were initially thought to only play a role in regulating cholesterol homeostasis. Oxidization makes cholesterol more hydrophilic, thereby facilitating its elimination. In addition, oxysterols act as a negative feedback control in response to high cholesterol concentrations [8] and can restore homeostasis [9]. Oxysterols are known to activate the transcriptional regulators of cholesterol synthesis-namely, the sterol regulatory element binding protein (SREBP) and the Liver X receptors (LXRs) [9,10]. In response to low cholesterol concentrations, SREBP is translated in the ER membranes and transported to the Golgi complex with the help of the SREBP cleavage activating protein (SCAP) for cleavage. The cleaved SREBP is subsequently able to enter the nucleus, triggering the transcription of genes involved in cholesterol synthesis [11,12]. In contrast, Figure 1. Overview of the generation of selected oxysterols from cholesterol. Oxysterols are generated from cholesterol through oxidation by pro-oxidants or enzymes. The majority of these enzymes involved in the synthesis of oxysterols are from the family of cytochrome P450 enzymes [3,5], with the exception of cholesterol 25 hydroxylase (CH25H).
Oxysterols are present at low concentrations, up to 5-fold lower compared to cholesterol, in circulation [7], and were initially thought to only play a role in regulating cholesterol homeostasis. Oxidization makes cholesterol more hydrophilic, thereby facilitating its elimination. In addition, oxysterols act as a negative feedback control in response to high cholesterol concentrations [8] and can restore homeostasis [9]. Oxysterols are known to activate the transcriptional regulators of cholesterol synthesis-namely, the sterol regulatory element binding protein (SREBP) and the Liver X receptors (LXRs) [9,10]. In response to low cholesterol concentrations, SREBP is translated in the ER membranes and transported to the Golgi complex with the help of the SREBP cleavage activating protein (SCAP) for cleavage. The cleaved SREBP is subsequently able to enter the nucleus, triggering the transcription of genes involved in cholesterol synthesis [11,12]. In contrast, when cholesterol levels are high, cholesterol binds to SCAP, which in turns binds to insulin-induced gene 1 protein (INSIG1) and INSIG2, which are ER retention proteins, thereby preventing SREBP activation [12]. Oxysterols have shown to directly bind to and activate INSIGs, thereby preventing the activation and cleavage of SREBP, leading to the downregulation of cholesterol synthesis [8,11,12]. In addition, oxysterols are also ligands for LXRs that exist in two isoforms, LXRα and LXRβ [8]. In response to high cholesterol concentrations, the activation of LXR by oxysterols leads to the upregulation of the genes involved in lipid metabolism, notably genes from the ATP-binding cassette family of membrane transporters, which regulate cholesterol efflux and excretion, resulting in a reduction in intracellular cholesterol accumulation [8,10]. Additionally, oxysterols have been shown to acutely regulate cholesterol concentrations through several post-transcriptional regulatory mechanisms [9]. For instance, several oxysterols, including 25-hydroxycholesterol (25-OHC), alter the enzymatic activity of 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR), an important rate-limiting enzyme involved in the synthesis of mevalonate in the cholesterol synthesis pathway [13]. Some oxysterols can also induce cholesterol esterification through the activation of acyl-CoA cholesterol acyl transferase (ACAT) within the cell, resulting in a rapid reduction in cholesterol [14].
The essential role of oxysterols in governing cholesterol homeostasis has long been established [8]. More recently, several studies have shown the numerous immunological functions of oxysterols, ranging from their involvement in chemotaxis and the development of immune cell niches [15][16][17][18] to skewing immune cell phenotypes and coordinating inflammatory responses [19] (Table 1). In addition, studies have also demonstrated that oxysterols bind to a wide range of receptors, from nuclear receptors such as retinoic acid receptor-related orphan receptors (RORs) to LXRs, estrogen receptors (ERs), and transmembrane G-protein coupled receptors (GPCRs) to carry out their diverse immunological functions [20]. This review focuses on the role oxysterols play in the host response to bacterial and viral infections. LXRs [21] RORα [22] RORγt [23] INSIGs [11] GPR183 [18,24] ERα [25] Produced by macrophages upon viral infection to mediate antiviral functions; broad antiviral activity against enveloped and non-enveloped viruses.
[ [26][27][28][29][30] Triggers cholesterol remodeling on the plasma membrane, restricting the intracellular dissemination of Listeria monocytogenes and Shigella flexneri; prevents CDC-induced pore damage. [31,32] Produced upon lipopolysaccharide (LPS) stimulation in the lungs. CH25H was found to be upregulated up to 24 h post-infection. Pulmonary administration of 25-OHC resulted in reduced immune cell infiltration and inflammation in the lung. [33,34] Downregulated upon exposure to house dust mites. CH25H was found to be upregulated in contrast. Pulmonary administration of 25-OHC resulted in a more severe onset of inflammation and airway remodeling.
Oxysterols in Bacterial Infections
Over the past decade, various studies have elucidated the link between oxysterols and the innate immune response to intracellular bacteria. An increased susceptibility to infection by Listeria monocytogenes and Mycobacterium tuberculosis was observed in LXR knock-out (KO) models, which provided potential evidence of the involvement of oxysterols in the immune response to these infections, given that oxysterols are ligands for LXR receptors [59,60]. Indeed, two recent studies have demonstrated an immunomodulatory role of the oxysterol 25-OHC against intracellular bacteria and secreted bacterial toxins [31,32]. The first study revealed a role for 25-OHC in the immunity against several intracellular bacteria. The authors demonstrated that in Listeria monocytogenes infection, CH25H KO mice had increased bacterial dissemination compared to wild-type (WT) -infected mice [31]. Additionally, the in vivo administration of 25-OHC reduced the bacterial burden, supporting the immunomodulatory role of 25-OHC. Furthermore, using both L. monocytogenes and Shigella flexneri, the authors demonstrated that the in vitro administration of 25-OHC resulted in the downstream activation of ACAT, which in turn triggers cholesterol remodeling on the plasma membrane and thereby restricts the cell-to-cell dissemination of these pathogens [31]. The second study elucidated a protective role of 25-OHC against cholesterol-dependent cytolysins (CDCs), a pore-forming toxin that is secreted by a variety of pathogenic bacteria [32]. The authors demonstrated in bone marrow-derived macrophages (BMDMs) that interferon (IFN) signaling mediates cholesterol remodeling on plasma membranes through the CH25H/25-OHC axis [32]. The reduced availability of cholesterol on the plasma membrane induced by 25-OHC resulted in a reduced binding of CDCs, thereby conferring resistance to CDC-induced pore damage. These findings were further explored in in vivo models, whereby the authors demonstrated that CH25H deficiency resulted in ulcerative lesions and larger lesion sizes. Additionally, pre-treatment with 25-OHC was found to be protective against CDC-mediated tissue damage [32]. Collectively, both studies showed that 25-OHC modifies the cholesterol content on the plasma membrane, thereby conferring resistance to CDCs as well as bacterial pathogens.
The role of oxysterols in the innate immune response to M. tuberculosis has also been investigated [35,61]. IL-36, a newer family of the IL-1 family of cytokines, is produced in macrophages upon M. tuberculosis infection to regulate the synthesis of oxysterols such as 25-OHC and 27-hydroxycholesterol (27-OHC) [61]. The production of these oxysterols subsequently led to downstream LXR activation, suppressing cholesterol metabolism and reducing mycobacterial growth. More recently, our laboratory discovered an important role for 7α,25-dihydroxycholesterol (7α,25-OHC) and its receptor GPR183 in M. tuberculosis infection [35]. GPR183, also known as Epstein-Barr virus-induced G protein coupled-receptor 2, was discovered in the 1990s as one of the genes that were upregulated upon infection with Epstein-Barr virus in Burkitt's Lymphoma cell lines [62]. GPR183 is expressed across several types of innate and adaptive immune cells such as dendritic cells, innate lymphoid cells 3 (ILC3s), macrophages, and T and B lymphocytes [5]. Oxysterols are known ligands for GPR183, with 7α,25-OHC being the most potent endogenous agonist [24]. The intracellular growth of both M. tuberculosis and M. bovis BCG in primary human monocytes was significantly restricted in the presence of 7α,25-OHC, an effect that was abrogated by a specific GPR183 antagonist [35]. This growth inhibitory effect, which was associated with induction of autophagy and the negative regulation of type I IFNs, was specific to primary human monocytes [35] and was not observed by others in a murine macrophage cell line [36]. In patients with pulmonary tuberculosis, lower GPR183 expression in the blood correlated significantly with more severe disease on chest X-ray, and GPR183 KO mice exhibited a higher lung mycobacterial burden and dysregulated type I IFNs in early infection [35], highlighting the important roles of oxysterols and GPR183 in the immune response to mycobacterial infections.
In addition to the immunomodulatory effects, a recent study provided evidence that the enzymes produced by M. tuberculosis regulate the oxysterol metabolism to limit the effective induction of the immune response [63]. The mycobacterial enzyme 3β-hydroxysteroid dehydrogenase (3β-HSD) can metabolize 25-OHC and 7α,25-OHC, among others, to render them inactive. As these oxysterols have protective roles in immunity against M. tuberculosis, it is possible that 3β-HSD, along with the other identified Mtb enzymes CYP124, CYP125, and CYP142, targets these oxysterols to interfere with and evade the host immune response, thereby allowing its continual persistence in host macrophages [63]. It is also possible that M. tuberculosis produces itself antagonists against GPR183, as has been demonstrated for other bacteria. Eubacterium rectale, for instance, produces lauroyl tryptamine, which is able to bind to and antagonize GPR183 against its endogenous agonist 7α,25-OHC at a 0.98 µM half-maximal inhibitory concentration [64]. Whether M. tuberculosis is able to produce antagonists for GPR183 or other oxysterol receptors remains to be elucidated.
Oxysterol gradients can also facilitate the migration and recruitment of cells. For instance, 7α,25-OHC attracts GPR183-expressing immune cells to secondary lymphoid organs [15][16][17][18]. In the host immune response to Citrobacter rodentium infection, 7α,25-OHC production attracts GPR183-expressing ILC3s and GPR183 KO mice have a reduced abundance of IL-22-producing intestinal ILC3s, which is associated with greater disease severity and mortality rates as compared to WT mice [37].
In addition, the immunomodulatory function of 25-OHC against a broad range of bacterial pathogens has been proposed in studies conducted with lipopolysaccharide (LPS), a major component of the outer membrane of Gram-negative bacteria. In the lung, the anti-inflammatory role of 25-OHC against acute lung inflammation through LPS stimulation has been studied [33,34]. 25-OHC was observed to be produced upon LPS stimulation in the lungs and bronchoalveolar lavage fluid. CH25H KO mice displayed a delayed resolution of inflammation [33]. Additionally, alveolar macrophages from CH25H KO mice displayed increased cholesterol accumulation and defective efferocytosis [33]. The in vivo administration of 25-OHC led to a reduction in immune cell infiltration and inflammation [34] and accelerated the resolution of inflammation in CH25H KO mice [33]. Another study also demonstrated the anti-inflammatory role of the CH25H/25-OHC axis in vivo in LPS-stimulated BMDMs [65]. Upon LPS stimulation, 25-OHC is produced to prevent the cholesterol-dependent DNA sensor protein absent in melanoma 2 (AIM2) activation and the subsequent downstream activation of IL-1β. The CH25H/25-OHC axis prevents the translocation of SREBP2 to the nucleus for the onset of cholesterol synthesis. In addition, CH25H KO BMDMs resulted in an increased accumulation of sterols (desmosterol, lanosterol, and 7-dehydrocholesterol) involved in cholesterol biosynthesis. The increased cholesterol load in CH25H KO BMDMs was associated with impaired mitochondrial metabolism and mitochondrial dysfunction, releasing mitochondria DNA into the cytosol for the downstream activation of AIM2 and inflammatory responses [65]. A schematic overview of the known mechanisms of oxysterol action in the immune response to bacterial infections is shown in Figure 2.
Oxysterols and Viral Entry
Cellular membranes contain microdomains that are rich in cholesterols and sphingolipids called lipid rafts that are often exploited by viruses, both enveloped and nonenveloped viruses, for their viral life cycle [77]. Many viruses utilize these cholesterol-rich regions for internalization [78] and entry into the host cell through a variety of mechanisms ranging from curvature formation to receptor clustering and binding to viral fusion proteins to gain entry. This, hence, highlights the need for cholesterol for efficient entry into cells, although it is noted that not all viruses depend on lipid rafts for entry [77]. In this regard, the CH25H/25-OHC axis has shown to inhibit viral entry through modifying the cholesterol composition on the plasma membrane, preventing membrane fusion for enveloped viruses [27,46,49,[66][67][68][69][70]73,79] and altering endosomal dynamics and its cholesterol composition, thereby preventing cytosolic entry for non-enveloped viruses [46,74]. The majority of the research conducted so far has demonstrated the potent antiviral activity of 25-OHC in inhibiting viral entry. Although other oxysterols such as 22(S)-OHC, 20α-OHC, and 7β-hydroxycholesterol (7β-OHC) can also inhibit viral entry, their molecular mechanisms of action have not been fully characterized [49].
Several mechanisms of action for 25-OHC regarding inhibiting the viral entry of many enveloped viruses-for instance, Porcine reproductive and respiratory syndrome virus [79]; hepatitis B virus [49]; ZIKV [69]; and, more recently, SARS-CoV-2 [30,68]-have been suggested. The majority of these studies demonstrate that the alteration of cholesterol on the plasma membrane is the key to restriction for viral entry, although various studies have demonstrated that 25-OHC is able to be localized within the plasma membrane [27], directly affecting membrane properties [31,80]. Mechanistically, infection studies with SARS-CoV-2 have elucidated that 25-OHC induces the depletion of cholesterol on the plasma membrane through ACAT activation [68]. This observation is consistent with work on intracellular pathogens, whereby cholesterol remodeling on the plasma membrane upon infection is caused by the activation of ACAT [31]. Therefore, this mode of action appears to be conserved across viral and bacterial infections [31,68]. In addition, the benefits of utilizing 25-OHC in combination with other viral inhibitors have been studied recently on human coronaviruses [30]. The authors conjugated 25-OHC with a peptide-based viral inhibitor with a different mode of action and tested its inhibitory efficacy against a broad range of coronaviruses. The resulting 25-OHC-conjugated lipopeptide (EK1P4HC) demonstrated a synergistic antiviral effect on inhibiting viral entry against SARS-CoV-2 and its variants as well as other human coronaviruses [30]. Apart from inducing cholesterol remodeling in plasma membranes, 25-OHC also localizes within late endosomes, where it prevents SARS-CoV-2-mediated entry to the cytosol through inhibiting cholesterol export [67].
The antiviral effects of 25-OHC have also been studied in vivo, with studies demonstrating that the administration of 25-OHC is protective against HIV [27], ZIKV [69], and SARS-CoV-2 [66], and transgenic KO studies showing that CH25H deficiency results in an increased susceptibility to murine gammaherpesvirus 68 (MHV68) [27].
A recent study highlighted the importance of 27-OHC in SARS-CoV-2 infection. Consistent with previous experiments conducted on other viruses, the in vitro administration of 27-OHC prior to infection reduced the intracellular accumulation of SARS-CoV-2 as well as human coronavirus OC-43. Furthermore, using mass spectrometry analysis, the authors demonstrated that the serum 27-OHC levels of patients were inversely correlated with the disease severity of COVID-19 [47]. Among non-enveloped viruses, a study conducted on human rotavirus demonstrated that 25-OHC and 27-OHC alter the endosomal dynamics in MA104 cells through preventing the interaction between oxysterol binding protein (OSBP) and vesicle-associated membrane protein-associated protein-A (VAP-A) [46]. This study demonstrated that these oxysterols are able to displace OSBP from the ER to the Golgi, preventing its interaction with VAP-A. The OSBP-VAP-A complex regulates cholesterol transport from the ER to intracellular organelles such as the endosomes. The disruption of the complex by these oxysterols thereby prevents cholesterol recycling between the ER and the late endosomes. This, in turn, results in the accumulation of cholesterol within the late endosomes and inhibits rotavirus entry into the cytoplasm [46]. Similarly, in reovirus infections 25-OHC alters endosomal dynamics upon infection. However, the authors suggested an alternative mechanism of restriction induced by 25-OHC on reovirus-infected cells. They demonstrated that, in HeLa cells, 25-OHC reduces the co-localization of the viral particles with the late endosomal marker Rab7. As reovirus entry into the cell is dependent on the timely uncoating of the virus in the late endosomes, the authors suggested that the main mechanism of restriction is due to this delayed trafficking of the viral particles to the late endosomes, which alters the uncoating of the virus and its subsequent penetration efficiency into the cell cytoplasm [74]. In Seneca valley virus infection, a study highlighted that CH25H activity is inversely correlated with viral replication and that 25-OHC inhibits Seneca valley virus replication in a dose-dependent manner in HEK-293T and BHK-21 cells [29]. The authors further elucidated that 25-OHC specifically inhibited the viral absorption process of the viral life cycle, with no effect observed on the later stages of the viral replication cycle [29].
Oxysterols and Restriction of Viral Replication
In addition to the preventing viral entry, the antiviral effects of 25-OHC extend further to modifying cholesterol content intracellularly and restricting viral replication [75]. In vitro, both the pre-and post-treatment of Poliovirus pseudovirus-infected HEK293 cells with 25-OHC was found to reduce viral replication. The authors further demonstrated that 25-OHC interacts with OSBP, leading to the reduced accumulation of Phosphatidylinositol 4-phosphate (PI4P) at the golgi apparatus [75,81]. PI4P has been implicated in supporting poliovirus replication partially through the recruitment of unesterified cholesterol to PVinduced membrane structures [81]. Hence, the reduction in PI4P induced through the 25-OHC/OSBP axis might reduce the cholesterol availability on PV-induced membranes that is required for replication.
Another mechanism of the antiviral activity of 25-OHC has been suggested to be the upregulation of the integrated stress response pathway. Authors found that the endogenous production of 25-OHC in BMDMs during MCMV infection leads to the induction of stress response genes independently of LXRs [82]. Furthermore, the addition of 25-OHC in BMDMs leads to the activation of general control nonderepressible 2 (GCN2), one of the eIF2α kinases that senses and activates the integrated stress response pathway [82]. Thus, the activation of the integrated stress response could lead to the suppression of protein synthesis, which viruses depend on for viral replication [82].
A cell-based screening of oxysterols in SARS-CoV-2-infected TMPRSS2-overexpressed VeroE6 cells identified 7-KC, 22R-hydroxycholesterol (22(R)-OHC), 24S-hydroxycholesterol (24S-OHC), and 27-OHC as potent inhibitors of SARS-CoV-2 replication in vitro [44]. In addition to these natural oxysterols, the authors further demonstrated that the semi-synthetic oxysterol derivatives Oxy210 and Oxy232 have a higher antiviral potency than the natural oxysterols. They further elucidated that Oxy210 inhibits the in vitro replication of SARS-CoV-2 and HCV through reducing double membrane vesicles (DMVs)-dependent replication [44]. These existing findings place oxysterols and their synthetic analogues in the spotlight as novel therapeutics for infectious diseases.
Oxysterols in Viral Assembly and Release
Another antiviral role of 25-OHC has been suggested in Lassa virus (LASV) infections, whereby 25-OHC was able to interfere with the late stages of the LASV life cycle through interfering with the viral glycosylation, affecting the production of infectious viruses [83]. In vitro, 25-OHC alters the glycosylation of the LASV glycoprotein 1 (GP1) in Huh-7 cells, causing an increased presence of immature forms of N-glycans on GP1 and thereby leading to the production of less infectious virus progeny which have defective entry. Furthermore, the overexpression of CH25H affected GP1 glycosylation, and the infectious viral production and knockdown of CH25H led to an increased production of infectious LASV, further supporting the importance of the CH25H/25-OHC axis in LASV infection [83].
Through an autophagy compound screening in ZIKV-infected Vero and C6/36 cells, one group identified an antiviral role of 7-KC in ZIKV infections [55]. The in vitro administration of 7-KC interfered with the later stages of the viral life cycle, as characterized by a reduction in the viral budding efficiency and infectious virion production, with no impact on viral entry or intracellular viral replication. Although the exact mechanism has yet to be elucidated, the authors suggest that 7-KC could influence the intracellular lipid environment in the organelles involved in ZIKV trafficking [55]. A summary overview of the known mechanisms of oxysterol action in the immune response to viral infections is shown in Figure 3.
Oxysterols in Viral Assembly and Release
Another antiviral role of 25-OHC has been suggested in Lassa virus (LASV) infections, whereby 25-OHC was able to interfere with the late stages of the LASV life cycle through interfering with the viral glycosylation, affecting the production of infectious viruses [83]. In vitro, 25-OHC alters the glycosylation of the LASV glycoprotein 1 (GP1) in Huh-7 cells, causing an increased presence of immature forms of N-glycans on GP1 and thereby leading to the production of less infectious virus progeny which have defective entry. Furthermore, the overexpression of CH25H affected GP1 glycosylation, and the infectious viral production and knockdown of CH25H led to an increased production of infectious LASV, further supporting the importance of the CH25H/25-OHC axis in LASV infection [83].
Through an autophagy compound screening in ZIKV-infected Vero and C6/36 cells, one group identified an antiviral role of 7-KC in ZIKV infections [55]. The in vitro administration of 7-KC interfered with the later stages of the viral life cycle, as characterized by a reduction in the viral budding efficiency and infectious virion production, with no impact on viral entry or intracellular viral replication. Although the exact mechanism has yet to be elucidated, the authors suggest that 7-KC could influence the intracellular lipid environment in the organelles involved in ZIKV trafficking [55]. A summary overview of the known mechanisms of oxysterol action in the immune response to viral infections is shown in Figure 3.
Oxysterols in Disease Pathogenesis and as Potential Biomarkers
Apart from their antiviral activities, several studies have also implicated oxysterols such as 7-KC as being detrimental in severe viral infections. While 7-KC has been demonstrated to have antiviral activities across several different viruses in vitro, in patients, however, a pathological role of 7-KC has been suggested with several viral pathogens due to its known cytotoxic properties at high concentrations [56]. 7-KC promotes a proinflammatory phenotype in human macrophages [53] and may thus, together with other oxysterols, contribute to excessive inflammation. Several studies have demonstrated the cytotoxic effects of these oxysterols among several non-immune cells, such as endothelial cells [84], neuronal cells [85], and mesenchymal stem cells [86,87]. Mechanistically, a study conducted on human umbilical vein endothelial cells (HUVECs) found that 7-KC and 7β-OH drive endothelial dysfunction by inducing early lipid accumulation and lysosomal permeabilization [84]. This results in increased oxidative stress, leading to apoptosis in these cells [84]. Other mechanisms, such as mitochondrial hyperpolarization [85] and changes in actin polarization, caspase activation, and autophagy, have also been described [86,87]. In viral infections, the plasma concentrations of 7-KC were elevated following human herpesvirus 8 infection [57] in diabetes patients and patients with influenza [51], as well as patients with COVID-19 [50]. In COVID-19, an increase in 7-KC is observed in patients with moderate and severe COVID-19, but this rises gradually along with disease severity. It has been proposed that the pro-inflammatory and cytotoxic effects of 7-KC contribute to the cytokine storm and acute respiratory distress and that 7-KC is thus involved in disease progression and poor outcomes [50]. Consequently, 7-KC may therefore serve as a biomarker for COVID-19 severity [50]. 7-KC has also been implicated in driving disease pathogenesis in cardiovascular diseases [56], and oxysterols have also been proposed as potential biomarkers for chronic and neurodegenerative diseases [88]. The translational potential of oxysterol research and the therapeutic application of oxysterols for viral infections and other chronic diseases are gaining momentum [88].
Conclusions and Future Perspectives
In this review, we highlighted an emerging field of research: oxysterols as important immunomodulators of infectious diseases. Initially thought to only be involved in cholesterol homeostasis, many studies conducted throughout the past decade have described the important role of oxysterols in both physiological and pathological conditions. Oxysterols are produced as part of the host immune response towards several bacterial and viral infections. In addition, the mechanisms of action induced by these immune oxysterols are broad, ranging from chemotaxis and regulating inflammatory responses to restricting intracellular cholesterol content, some of which are conserved across viral and bacterial pathogens. Given the increasing prevalence of antimicrobial resistance and the emergence of novel viruses, there is an increasing need to find new and innovative strategies to combat these pathogens. The growing evidence of the role of oxysterols in contributing to the immune response to infections hence presents them as novel biomarkers of disease severity and suggests oxysterol receptors to be attractive targets for host-directed therapy for improving bacterial and viral infectious disease outcomes. | 6,218.8 | 2022-01-01T00:00:00.000 | [
"Biology"
] |
Frequency-Selective MHz Power Amplifier for Dielectric Barrier Discharge Plasma Generation
Plasma-assisted nitrogen fixation at atmospheric pressure is a clean and decentralized method for fertilizer production. Among many different plasma discharge types, dielectric barrier discharge (DBD) is one of the few that can generate high output of nitrogen compounds at atmospheric conditions for fertilizer usage. Most DBD generators operate at kHz switching frequencies, however, previous research has found that discharge activities occur at the intervals of the greatest voltage slew rate. The finding implies that operating at higher frequencies can lead to more discharge activities and higher plasma generation. This paper presents a MHz Class E power amplifier with frequency selection to generate DBD plasma at two distinct frequencies. The power amplifier achieves a peak efficiency of 91.5% and outputs 600 W at frequencies of 12.4 MHz and 15.5 MHz.
I. INTRODUCTION
The Haber-Bosch process is one of the most impactful chemical reactions in history. This process improves agricultural yield by increasing ammonia fertilizer production, helping in feeding the world's growing population. In 2020, global ammonia production was around 144 million metric tons [1], and more than 95% of ammonia is produced through the Haber-Bosch process [2]. Between 75% to 90% of the ammonia produced is used in making fertilizer, benefiting 50% of the world's food production [3]. The Haber-Bosch process fixes nitrogen with hydrogen to produce ammonia with the feedstock of natural gas (50%), oil (31%) or coal (19%) [2]. Burning these fossil fuels during the process leads to more than 1% of the world's total CO 2 emission. The process requires a centralized plant for high-temperature and highpressure reactions. After the production, the ammonia then needs to be transported to local farms. The transportation not only has environmental and cost concerns, it is also dangerous since ammonia is highly toxic and flammable when exposed to high temperatures [4].
Recently, Non-Thermal Atmospheric Pressure Plasma (NTAPP) has gained attention as a sustainable replacement for conventional nitrogen fixation process [5]. With the development of low-temperature plasma science, the minimum theoretical energy consumption of non-thermal plasmaassisted fixation is lower than that of the Haber-Bosch process [6]. The fixation process applies plasma at the surface of air and water to form soluble reactive oxygen and nitrogen species (RONS). The species include nitrate (NO − 3 ) and other forms of nitrogen compounds, which can all be used as fertilizers [5], [7]. Plasma-assisted nitrogen fixation works under atmospheric conditions without emission of green house gases. Localized production due to the reduced system size and simplified process also eliminates potential cost and hazards during transportation.
The literature has discussed different discharge types for nitrogen fixation, including microwave plasma, spark discharge, gliding arc discharge, and dielectric barrier discharge (DBD) [6]. Among these alternatives, both the gliding arc discharge and DBD have high output of NO x compounds and can operate under atmospheric pressure. This work focuses on designing a power amplifier to produce DBD plasma. Most DBD systems use power supplies that generate a low-frequency (in the range of 50 Hz to a few tens of kHz) multi-kV output at the plasma load [8]. Previous research has shown that the discharge activities occur during the zero crossings of the applied voltage, corresponding to the intervals of the greatest voltage slew rate and maximum current [9]. The finding implies that increasing the frequency can lead to more discharge activities and higher plasma production rate, motivating the design of a MHz power amplifier to drive a DBD plasma load. Conventionally, RF system manufacturers use linear power amplifiers in high-frequency applications, including plasma generation. The linear power amplifiers have the advantages of linearity but low efficiencies. The switch-mode power amplifiers such as Class D, Class E, and Class 2 can achieve much higher efficiencies under zero-voltage-switching (ZVS) operations. However, due to their resonant nature, these switch-mode power amplifiers maintain high efficiencies only at a fixed frequency and constant load condition. Previous research has introduced different methods to extend the frequency and load range, including reactance compensation [10], impedance compression [11], phase-switched impedance modulation [12], and load-independent power amplifier design [13]. The power delivered to a plasma load at MHz can be much higher than that at kHz due to higher plasma generation rate. As a result, it is necessary to operate the power amplifier at a low burst rate to reduce the average power delivered to the DBD reactor. This paper presents a wideband Class E power amplifier using the reactance compensation technique and combined with frequency selection networks to drive multiple DBD loads sequentially. The frequency selection networks allow the power amplifier to operate for longer time within each burst period while each DBD reactor only receives a fraction of the total average power.
One challenge before designing the power amplifier is to measure the plasma load impedance. Before the plasma breakdown, the load is almost entirely reactive. To start the plasma, one requirement is to carefully tune the matching network to minimize the reflective power. After the plasma breakdown, the load has an additional resistive part, and the resistance varies significantly with the operating conditions. Depending on the reactor design and operating conditions, the plasma load after the breakdown can still be very reactive, making it hard to measure precisely. Section II describes the DBD electrode design and plasma impedance measurement in details. The rest of the paper is organized as follows. Section III explains the design of the frequency-selective power amplifier; Section IV shows the experimental results; and Section V concludes the paper.
II. ELECTRODE DESIGN AND PLASMA LOAD MEASUREMENT
A DBD reactor has two electrodes located on the opposite sides of a dielectric barrier [8]. The electrode on the front side is exposed to air, and the electrode on the back side is covered by a layer of dielectric material. Generating DBD plasma at MHz frequencies requires a careful selection of the dielectric material for the board to produce a certain capacitance with low loss. Previous research has used mica at 13.56 MHz [14]. In this work, Rogers 4003C [15] is used because of its low dissipation factor at high frequencies and the possibility of it being integrated into a PCB to keep the costs low. Figure 1(a) shows the electrode on the front side of the Rogers board made by a thin line of copper tape, and Figure 1(b) shows the grounded electrode on the back made by a copper plane covered by a layer of Kapton tape. After experimenting with several different widths of the front electrode and thicknesses of the Rogers board, the designs with low capacitance (narrow copper lines on the front side) and small board thickness are found to break down at lower voltages. The electrodes used in this work has a board thickness of 0.2 mm, a front electrode width of 1 mm, and a length of 100 mm. The plasma strikes when the voltage across the electrodes of this DBD reactor exceeds 700 V rms . After the plasma strikes, a plasma sheath forms only around the top electrode without a breakdown of the dielectric. The detailed measurement procedure will be described later in this section.
Plasma load impedances are sensitive to the operating conditions, such as temperature, pressure, and power delivered [9]. Before the plasma generation starts, the load is nearly purely reactive; after the plasma breakdown, the load impedance changes with the increased applied voltage, especially for the resistive part. A commercial linear power amplifier is used to drive the plasma load for impedance measurement. When the power amplifier drives a highly reactive load, most of the power gets reflected, which makes it hard to reach the voltage for plasma breakdown. Therefore, it is necessary to match the plasma load to 50 using a tunable matching network and actively adjust the network as the plasma impedance changes. Calibration of the probes is also essential for accurate measurements, especially the phase angle between the voltage and current. Even after the breakdown, the plasma load can still have a relatively high quality factor (Q), and a small skew can result in a large error in the measured impedance. Knowing the caveats above, the measurement of the load impedance in a DBD plasma system includes the following steps:
1) VOLTAGE AND CURRENT PROBE CALIBRATION
The voltage measurement requires a capacitor divider by connecting a high-voltage C0G capacitor between the load and the probe because of the high breakdown voltage of the plasma load. The current probe used in the measurement setup is Pearson current monitor 2878. Before connecting to the DBD load, it is important to calibrate the probes. The calibration includes measuring the voltage and current across a 50 radio-frequency (RF) load driven by a linear power amplifier (ENI A1000) at the desired frequency and power level. During this measurement, it is important to adjust the skew time between the two probes to make the phase angle between them 0 • . The measured V r I r is a reference for 50 .
2) SMALL SIGNAL IMPEDANCE MATCHING OF THE DBD LOAD TO 50
After calibrating the probes, the next step is to connect them to the DBD load and measure the small signal impedance with a tunable matching network connected to the load. The tunable matching network can be adjusted until the measured impedance is 50 . Taking the measurements with the probes connected allows for better matching, since the capacitance of the DBD load can be small (on the order of a few pF to hundreds of pF depending on the electrode design). The small signal measurement sets the initial matching conditions before the system is powered on, which accelerates the impedance matching in the next step.
3) MEASUREMENT OF THE DBD PLASMA LOAD
The DBD load and the matching network are then connected to the linear power amplifier and a power meter (Keysight N1914 A). Figure 2 illustrates the overall measurement setup. Figure 3 shows the setup of the voltage and current probes (V & I Probes) in Figure 2. The voltage probe connects to a capacitor divider to measure the high output voltage. The Pearson current monitor 2878 is used to measure the current. The input terminal connects to the tunable matching network, and the output terminal connects to the DBD load. As the applied voltage increases, the matching network can be adjusted when the reflected power becomes significant. Finally at the desired output power after the plasma breakdown, the voltage and current measurements are taken along with the phase angle between them (V m , I m ).
4) CALCULATE THE PLASMA LOAD IMPEDANCE
Using the V r I r reference for 50 in step 1) and the voltage and current measurements in step 3), the following equation calculates the plasma impedance: (1) Figure 4 shows the measured plasma impedance at 13.56 MHz in terms of the parallel resistance and capacitance as modeled in Figure 5. The plasma breakdown starts at an input power of around 200 W and an output voltage of 700 V rms . As the power increases after the breakdown, the output voltage stays relatively unchanged, but the length of the plasma strip generated at the front electrode gradually extends, causing a decrease in the resistance. The plasma strip reaches the full length of the front electrode at about 600 W. The measured load impedance has a parallel capacitance of 33 pF (including 13 pF of the connectors and the board) and a resistance of 758 with a phase angle of approximately -70 • . Because of its significant real part, the measured impedance provides a relatively reasonable estimate of the plasma load impedance. However, depending on the plasma type, electrode design, and the operating conditions, the load impedance can have a close-to-±90 • phase angle. In those cases, it can be challenging to measure the plasma impedance accurately using voltage and current probes. In step 4), it is also possible to use the input power and output voltage measurements to calculate the plasma impedance. However, the input power measurement includes the power lost in the matching network, which can make the impedance measurement less accurate.
III. DESIGN OF THE FREQUENCY-SELECTIVE POWER AMPLIFIER
Most of the DBD plasma generators operate at low switching frequencies. However, [9] has shown that most of the discharge activities occur during the intervals of the greatest voltage slew rate and maximum current, implying a higher plasma generation rate at a higher frequency. As the frequency increases into MHz, the amount of the plasma generated can result in a much higher power delivery to the load. Even with the high-frequency dielectric materials, it is still hard for the DBD reactor to dissipate all the power. Therefore, the need to operate the power amplifier at a low burst rate arises to avoid such high power dissipation in the load, as illustrated in Figure 6. However, operating at a low burst rate is not a good utilization of the power amplifier's high power delivering capability. A better operation of the power amplifier is to drive multiple loads sequentially. As illustrated in Figure 7, within each pulsing period, the power amplifier is delivering power selectively to only one of the DBD reactors, which is labeled by the same color as the input signal. As a result, the power amplifier is able to operate longer, while each DBD reactor only receives a fraction of the total average power. The literature has introduced a few techniques for selective power delivery, including using switch selection [16], and frequency selection [17]. Both techniques require active switches to either selectively connect to the load or be used in the switched capacitor banks to change the resonant frequency.
Depending on the applications, these switches may require high voltage blocking and fast switching capabilities, as well as auxiliary gate drive circuitry, which could be complicated for the non-ground-referenced switches. In contrast, this paper proposes to use the similar idea of frequency selection but without any additional switches. Table 1 compares the selective power delivery technique presented in this work with the two mentioned above. Using only passive elements makes the design less expensive and less complicated. Similar to the frequency selection technique using switched capacitor banks, the design presented in this work also requires a wide bandwidth so that selective power delivery can be achieved by setting different switching frequencies. Figure 8 shows the proposed frequency-selective power delivery system. The LC resonant networks connecting to the loads are resonating at different frequencies, and all of these resonant frequencies are within the bandwidth of the designed wideband power amplifier. The prototype design of a frequency-selective Class E power amplifier to drive two DBD loads follows the design procedure below. Figure 9 shows the Class E power amplifier with finite DC-feed inductance that is able to maintain efficient ZVS operations when driving a variable load [13].
1) DESIGN THE CLASS E POWER AMPLIFIER
L F and C F are the main design components to meet ZVS, while the output network of L S -C S -L P -C P filters out higher harmonics. Following the design steps in [13], L F and C F can be calculated: where f IN = 1.4× f S ( f S is the switching frequency, 13.56 MHz in this case) and k f = 1. The input voltage is 200 V and the power is 600 W based on the measurements from Figure 4.
2) EXTEND THE BANDWIDTH OF THE POWER AMPLIFIER
The conventional switch-mode power amplifiers only operate at a single frequency. When shifting away from the designed frequency, both the efficiency and output power degrade significantly. Extending the bandwidth for the frequencyselective power amplifier is necessary to ensure efficient operations at different frequencies and same power delivery to all of the loads. The literature proposes different methods to extend the bandwidth of these power amplifiers, including tunable and switchable matching networks and reactance compensation techniques [10], [18], [19], [20], [21]. This work uses reactance compensation because it only requires two sets of LC resonant tanks (L S -C S and L P -C P ), which are already in the design. The first step is to design L S and C S based on the output filtering requirement and the targeted bandwidth. A larger Q S makes the output more sinusoidal with less harmonics but the trade-off is a smaller bandwidth. In this design, Q S is selected to be 2.5 to provide a bandwidth that covers the targeted operating frequencies. Next, L P and C P are calculated based on the reactance compensation requirement following the equations below: [22] C F + 1 where Q P = ωRC P , Q S = ωL S R , ω is the the center frequency, and R is the output resistance of the Class E power amplifier determined in the previous step. Substituting the C F and L F calculated from the previous step, Q S = 2.5, and ω = (2π )13.56 MHz into the equation results in a Q P of 1.48.
3) DESIGN THE OUTPUT RESONANT NETWORKS FOR THE TWO LOADS
Selective power delivery to one of the two loads requires two output resonant networks, resonating at different frequencies. Figure 10 shows the schematic of the output networks.
The first design requirement is to deliver the same amount of power to the active plasma load at the corresponding frequency. Specifically, which leads to where Q 2 = 2π f 2 (C 2 + C load )R load , where f 1 is the operating frequency to deliver power to the plasma load Z 1 , f 2 is the frequency to deliver power to Z 2 , and C load and R load are defined in Figure 5. At f 1 , and at f 2 , The second requirement is to make the voltage across the active plasma load to be significantly larger than the other load voltage: so that only one load at a time is receiving power and generating plasma. Designing for a minimum voltage ratio of 2 A high Q allows the two designed frequencies to be close together; however, it also makes the network hard to tune and sensitive to component variations, including the variation in the plasma capacitance. A low Q requires the two frequencies to be farther apart, but at the same time, both f 1 and f 2 need to be within the bandwidth of the designed power amplifier. For a plasma load of C load = 20 pF and R load = 758 , Q = 5 is selected for both output networks with f 1 = 12 MHz and f 2 = 15.5 MHz. The third requirement is to make sure the output impedance (Z out ) at f 1 and f 2 are equal. If Z out at the two frequencies are not the same due to the interaction of the two close resonances, additional tuning of the component values as well as the designed frequencies can help to achieve equal Z out .
4) MATCH THE OUTPUT RESISTANCE TO THE PLASMA LOAD
The last step is to match the output resistance of the designed Class E power amplifier to the total impedance of the load (Z out ). Figure 11 shows the impedance transformation of the power amplifier's output resistance by a factor of 1 m 2 using Norton transform [23], [24], where m = L S L Fa + 1 (11) and L Fa is part of L F in Figure 9, where The reactive part of Z out can be combined to L P or C P , so that
TABLE 2. Design Parameters of the Frequency-Selective Class E Power Amplifier
One of the advantages of this technique is that the impedance waveform across frequency is exactly the same after the transformation. Therefore, the transformation step does not change the bandwidth of the designed power amplifier. Figure 12 shows the overall schematic of the design, and Table 2 lists the design parameters.
IV. EXPERIMENTAL RESULTS
This section discusses the experimental results of the design presented, including the small-signal impedance, power, and efficiency measurements. First, it is necessary to measure the small-signal output impedances (Z out ) of the designed power amplifier (Figure 13) at f 1 and f 2 to ensure that they match at the two frequencies. Figure 14 shows the modified front side of the electrode design to include two DBD loads, where each top electrode has the same dimensions as the one in Figure 1(a). Figure 15 shows the dummy loads to measure the impedance and efficiency. The dummy loads are designed using high-voltage resistors and capacitors to emulate the DBD loads with and without plasma generated. The measured output impedances shown in Figure 16 confirm that Z out at f 1 = 12.4 MHz roughly equals Z out at f 2 = 15.5 MHz.
The next step is to test the design with the dummy loads and calculate the system efficiency using the voltage measurements V 1 and V 2 and the dummy load resistance. Figure 17 shows the waveforms of V 1 , V 2 , and V drain at 12.4 MHz and 15.5 MHz. The drain voltage waveforms show ZVS switching behaviors at the two frequencies. V 1 at f 1 = 12.4 MHz and V 2 at f 2 =15.5 MHz are significantly larger than the voltage across the no-plasma load in each case, ensuring the power delivery and plasma generation only at the selected load. The ratios of V 1 V 2 at f 1 and V 2 V 1 at f 2 are not the same, because otherwise, which contradicts Equation 5 for f 1 = f 2 . Ensuring the same power delivery to each load is more critical than keeping V 1 V 2 at f 1 and V 2 V 1 at f 2 equal as long as the voltages across the no-plasma loads are much lower than the plasma breakdown voltage. Figure 18 shows the power and efficiency plots at both of the designed frequencies. The design achieves a peak efficiency of 91.5% at 15.5 MHz. Finally, it is ready to test the design with the DBD loads as shown in Figure 19. The function generator is programmed to output a series of 12.4 MHz gate pulses followed by an idle period (setting gate to low), and then output a series of 15.5 MHz gate pulses followed by another idle period. The 12.4 MHz and 15.5 MHz operations are both set to 1000 cycles with a burst frequency of 100 Hz. At 12.4 MHz, plasma is generated on the left electrode (Figure 20(a)), and at 15.5 MHz, plasma is generated on the right electrode ( Figure 20(b)). Figure 21 shows the measured output voltages and input power when the designed power amplifier drives the DBD loads. At both frequencies, the plasma starts at an input voltage of around 60 V and an input power of around 100 W. After the breakdown, the measured voltage across the plasma load remains at about 720 V rms regardless of the input power. Since the plasma load is variable and difficult to measure precisely, the measurements of the output power and efficiency when directly driving the DBD loads are neglected. Table 3 compares this design with the other high-frequency wideband power amplifiers in the literature. By using the reactance compensation technique, the design presented in this work does not require any additional switches, which simplifies the control and reduces cost. The design is able to maintain its high efficiency at two distinct frequencies, 12.4 MHz and 15.5 MHz, with an output power of 600 W. In addition to bandwidth extension, the frequency selection network also allows the designed power amplifier to deliver power to multiple loads selectively, which is an additional feature to all of the previous work in this table.
V. CONCLUSION
As an alternative for the Haber Bosch process, the plasmaassisted nitrogen fixation using the dielectric barrier discharge is able to achieve much cleaner and decentralized fertilizer production. This paper presents a frequency-selective Class E power amplifier that is able to drive DBD plasma loads sequentially at MHz frequencies by using the reactance compensation and frequency-selective resonant networks. The designed power amplifier is able to output 600 W at frequencies of 12.4 MHz and 15.5 MHz with a peak efficiency of 91.5%. | 5,684 | 2022-01-01T00:00:00.000 | [
"Physics"
] |
Distributed Signal Processing Algorithms for Wireless Sensor Networks
- Wireless Body Area Networks (WBAN), in particular in the field of wearable health monitoring system (WMB), such as electromagnetic cardiograms (ECG) data collecting system via WBANs in e-health applications, is becoming increasingly important for future communication systems. Compressive sensing (CS), on the other hand, has been shown to consume less power compared classic transform-coding-based approaches. We propose a new low-rank sparse deep signal recovery algorithm for recovering ECG data in the context of CS (Compressive sensing) because the spatial and temporal data collected by a WBAN have some closely correlated structures in certain wavelet domains e.g., the discrete wavelet transform (DWT) domain.
INTRODUCTION
For statistical inference in wireless connections, distributed signal processing techniques are used. Utilizing distributed analysis methods, data is extracted from data gathered at nodes dispersed along a geographic region. A group of surrounding nodes collects nearby data for each node, together including their local estimations for a better estimate, and sends it to each node [1] [2].
Figure 1. Distributed wireless network
Distributed signal processing (DSP) is a networking design and evaluation approach that may solve optimization and adaptation issues in a distributed and productive way [3].
Distributed Wireless Networks
A WLAN network allows a WLAN hotspot to be extended along a greater geographical region without any need for cables to be attached to each access point (AP). It is comprised of two or even more Wi-Fi (access point) base stations that work together as one platform. The backbone will be utilised as a backbone for computers, cell telephones, monitors, controllers, and processors in upcoming data transmission in distributed wireless networks. In this scenario, the dispersed networks has no central controller, causing the entire network to collapse. Furthermore, in a dispersed network, many data transmission paths are possible.
Distributed Estimation -In the distributed parameters estimation, we concentrate on processing an unknown vector 0 of size M x 1. The intended signal for each node at the point in a given time i is d k (i) = ω 0 H x k (i) + n k (i), i = 1, 2, . . . , I The result of the adaptive algorithm for every node is provided by y k (i) = ω k H x k (i)), i = 1, 2, . . . , I A problem described in the form of a minimizing cost in the distributing format at each node k is used to predict unknown vector the global network cost function could be described as
Compressive Sensing Techniques
5G wireless communication technologies are intended to power the next generation Internet of Things (IoT). In view of the growing number of Internet-of-things [1] devices, wireless IoT will become dense in the near future, taking place a tremendous variety of complicated exchanges and turning into self-organized systems [4]- [10]. For instance, contemporary advances in industrial automation approaching Industry 4.0 models drive factory transformations into extremely flexible and recognizable manufacturing processes. For instance, contemporary advances in industrial automation approaching Industry 4.0 models drive factory transformations into extremely flexible and recognizable manufacturing processes. Radio technologies also can only play an important part in this context when combined with modern systems to allow ultrareliable Low Latency (URLL), decentralized computing capabilities and massive machine communication (mMTC), for self-organizing essential applications [11]. Current technology is confronted with an expanding amount of information in the era of the Internet of Things, and the increasing amount of information of all types, as well as the growing capabilities of modern gadgets in the processing of high-quality information. These massive amounts of data must be sensed like signal acquisition transmitted like telecommunication technology processed. All three functions, notably sensing, storage memory, high transition bandwidth, necessitate data processing, and of course, greater power. Compression was proposed and applied in the field of signals and images, in which a signal/image was altered and saved in a considerably smaller size. When the compressed signal/image was needed, it was decompressed to its original size. Compressive sensing was proposed to prevent unnecessarily sampling redundant data and bearing the aforementioned adverse effect [12]. The signal is sampled through compressive sensing with a substantially fewer amount of trials than the Nyquist-Shannon theorem requires. It is predicated on the premise that the signal is sparse. Naturally, this approach assumed true for the vast majority of data types found in nature. Compressive sensing mainly is a challenge to compressively measure a signal while its information content is kept preserved to recover the original signal after compressive sensing. Compressive methods have a wide range of applications, including signal processing, smart environments, telecommunications, acoustic OFDM. In its most basic form, compressive sensing is modeled as follows: Consider x to be a signal vector with length n in the Rn vector space. Then x can be compressed felt, or displayed as a vector y of length m in the vector space Rm, as follows: Y m×1 = A m×n * X n×1 The name "compressive sensing" comes from the fact that m is substantially shorter than n, m n. A compressive sensing matrix is what it's called. A long signal vector is sensed as a much shorter signal vector in compressive sensing, as shown in Fig 2.
LITERATURE REVIEW
Some research has been done into the usability and structure of CS when utilized for signal acquisition in WBAN. Mamaghanian et al. [2] evaluated the efficiency of CS for ECG capture in detail, as previously stated. Lastly, the energy savings were enhanced to 52.04 percent. The CS acquisition of non invasive baby ECG, which is a key branch in healthcare systems and can be employed for the identification of embryonic development and behavior, was also studied by the researchers. The downsides of fetal ECG, such as strong noise and non-sparsity, which are incompatible with standard CS frameworks, have been well remedied by sparse Bayesian learning; raw foetal ECG recordings are recreated with acceptable quality while synchronously maintaining interconnectedness relationships of multi-channel signals. Brunelli and Caione [3] investigated the energy usage issue of both digital and analog CS, conducting a valid assessment utilizing real resource-constrained hardware architecture to investigate the influence of CS variable on signal recovery performance and sensor longevity. Majumdar and Ward [4] integrate state-of-the-art blind CS and low-rank approaches, and then create a Split Bregman strategy to address the EEG signal restoration issue in WBAN. In a rakeness-based CS structure with a zeroing approach, Mangia et al. [5] examined tradeoffs among data compression and signal reconstructing. The potential of CS for signal acquisition in IoT has been widely investigated in terms of sensor, transmitting, and reconstructing to relax energy usage and extend network capacity by establishing revolutionary cluster-sparse signal reconstruction methods. Furthermore, to the consumption-efficient signal sampling technique, Peng et al. [6] considered secure transmission of the received signals in WBAN. In order to construct a joint signal capture, compressing, and encrypting system, chaos theory was added into the CS network to obtain a secret measurement matrix. Such a concept demonstrates chaos-based CS's huge potential for secure CS in WBAN and other IoT applications. The sliding window treatment method was also abandoned in order to make significant advances in signal recovery while lowering communication needs.
METHODOLOGY
By employing, the concept of sparsely, CS develops as a novel framework for signal collection. Rather than using the usual sampling and compression data-gathering method, CS shows how the number of data points necessary can be drastically decreased if the signal is sparse or compressing in some way. In this case, CS collects the information directly, compressive and at a lesser pace. Figure 4.9 depicts the sampling method, which is discussed below. Suppose that the signal to be sampled is k-sparse and column vector = ( 1 , 2 , … … . . , ) , under the basis and that it is represented as (1). In comparison to the Nyquist sampling theorem, it is evident that M N denotes a decrease in dimension, and so CS denotes a reduced sampling frequency. Where, = ( • log ( ) ) predetermined projecting vectors make up the measurement matrix ( ). In comparison to the Nyquist sampling theorem, it is evident that < denotes a decrease in dimension, and so CS denotes a reduced sampling frequency. To summarize, in the CS procurement process, the old method of high-frequency sampling using Nyquist theory and then data compression using the sparse transformation has been substituted by a simple low-rate linear projection. As a result, system complexity reduction can be achieved and energy efficiency can be enhanced.In the Nyquist sampling theorem, signal restoration is done via linear sinc interpolation, that does not match the CS architecture, which is defined by an easily interpretable linear system. The inherent nonlinear technique solves the following optimization issue by exploiting the sparsity constraint.
The actual signal x is then retrieved using = . Addressing equation demands thoroughly searching all k sub-columns from N ones, and it is determined to be a non-deterministic polynomial-time (NP)-hard task for normal-sized signals. The convex relaxed expression of equation is referred to as the alternate reconstructive approach.
To ensure that this optimization issue has a distinctive and consistent solutions, the sensing matrix ⨀ must satisfy the restrictive isometric property (RIP) of order 2k, which means that there must be a constant 2 ∈ [0,1]such that The following theorem dominates the precision of restoration for any 2k-sparse signals x.
Figure 3: Proposed Compressive Sensing Modules
In the suggested technique, the sensor initially detects the × 1 vector ( ) at every node, and then estimates 0 in the compressive region with the support of the × measurement matrix Γ. Data Exchange: Only the localized compression estimator Ψ ( ) will be communicated among node k and all its neighbour nodes based on the network topology structure. Locally, the measuring matrix Γ will be stored.
Merger: After the data sharing is completed at every time instant I = 1, 2,..., I, the combining phase started. Through a given equation, every node will aggregate the locally compression estimation method from its adjacent nodes and itself, to calculate the compression estimator that has been modified as ( + 1).
In conclusion, only the localized compression estimator Ψ ( ) will be broadcast over the network during the deep compressive estimator operation, cumulating in a reducing the number of variables to be sent from M to d.
Deep Learning Network Description
The robust system is developed by application of convolution neural network (CNN). In this work, residual learning is taken as proposed approach. For residual learning loss function is:
Sparse and Low Rank Reconstruction
The compressive multichannel ECG signal will be corrupted when noise penetrate the multichannel ECG signal CS system at the sender side. In this work, we examine the multichannel ECG signal CS in the existence of noise. = +
Decompression Module
Testing Process In terms of maximum probability, the matrix Frobenius norm is best for modeling Gaussian noise. As a result, the related optimization model is specified: , , ‖Ω ‖ 0 + ‖ ‖ 2 .
= +
Where represents two regularization parameters , and ||. || f signifies the matrix Frobenius norm. The optimization issue employing the ℓ 0 norm, on the other hand, is strongly non-convex and generally NP-hard. We use the ℓ 1 norm instead of the initial ℓ 0 norm for convenience. As a result, can be rephrased as follows: The suggested optimum framework based on sparse and low rank depiction can be written as: , , ‖Ω ‖ 1 + * rank‖ ‖ * + * rank‖ ‖ 2 .
= +
Here ‖ ‖ * denotes x's nuclear norm, which is the same as the total of singular values of x and λ denotes the regularization parameter.
RESULT ANALYSIS
The result analysis is performed by simulating the WBAN scenario on MATLAB platform. Three conditions are considered and discussed further: With respect to variable nodes, With respect to variable SNR, With respect to variable compression ratio. The ECG samples in our experiments are obtained from MIT-BIH Arrhythmia Database is used.
Analysis with respect to variable number of Nodes
In this section, the simulation is performed with variable number of WBAN nodes as well as with different CR. The simulation results are analyzed on three parameters such as PRD, MSE and Time. From the result it is analyzed that with increasing number of nodes the PDR decreases and time increases. The MSE was evaluated between -30 to -40 db. Analysis with respect to SNR In this section, the simulation is performed with variable SNR for WBAN nodes as well as with different CR. The simulation results are analyzed on three parameters such as PRD, MSE and Time. From the result it is analyzed that with increasing SNR the PDR increases. The MSE was evaluated between -30 to -40 db and it seems to be increasing. The performance of proposed deepCS is compared with existing research work presented by Zhang et al. The author presented a sparse signal recovery algorithm for recovering ECG data in the framework of CS. The model presented in this work, shows the MSE that ranges from -7 to -18. The MSE evaluation of proposed work ranges between -30 to -40. The comparative analysis shows and it can be concluded that proposed deepCS algorithm is more efficient as compared to existing work.
CONCLUSION
Compressive sensing (CS) has recently proven to be an excellent data compression approach for wireless body-area network tele monitoring of multichannel ECG signals. Most present EEG CS multi-channel algorithms disregard the noise.
Other noise kinds, such as heavy impulsive tailed noise, are also accessible. This study proposes a deep learning-based sparse and low rank representation in the presence of noise to address the aforementioned difficulties. The simulations findings demonstrate the advantage of the deep CS offered over advanced noise strategies. The simulation is performed under different scenarios: In first scenario, the simulation was performed with variable number of nodes and with variable compression ratios. It was observed that with increasing number of nodes the PRD decreases and the MSE was evaluated between -30 to -40 db.
In second scenario, the simulation was performed with variable SNR for WBAN nodes as well as with different CR. From the result it is analyzed that with increasing SNR the PDR increases. The MSE was evaluated between -30 to -40 db and it seems to be increasing. In last, the performance of proposed deepCS is compared with existing research work presented by Zhang et al. [1]. The MSE evaluation of proposed work ranges between -30 to -40 that is more efficient as compared to existing work. From result analysis of proposed work, it can be concluded that the proposed deepCS algorithm is more efficient as compared to existing work. The MSE evaluation of the proposed approach is good enough. But it is seen from the result that the proposed algorithm shows good MSE evaluation at lower SNR value and it gets increased with increased SNR. This needs to be focused for future research direction. | 3,518.2 | 2021-08-28T00:00:00.000 | [
"Engineering",
"Computer Science",
"Medicine"
] |
Fabrication of low loss dispersion engineered chalcogenide photonic crystals
We demonstrate low loss photonic crystal waveguides in chalcogenide (Ge33As12Se55) glasses. The measured losses are as low as 21dB/cm. We experimentally determine the refractive index of the thin film chalcogenide glass to be n = 2.6 and demonstrate that dispersion engineering can be performed up to a group index of ng = 40 in this relatively low refractive index contrast system. © 2011 Optical Society of America OCIS codes: (230.5298) Photonic crystals; (220.4000) Microstructure fabrication; (190.4360) Nonlinear optics, devices; (130.5296) Photonic crystal waveguides. References and links 1. T. K. Liang, and H. K. Tsang, “Optical Limiting and Raman Amplification in Silicon Waveguides”, in Optical Fiber Communication Conference and Exposition and The National Fiber Optics Engineers Conference, Technical Digest (CD) (Optical Society of America, 2005), paper JWA15. 2. M. Dinu, F. Quochi, and H. Garcia, “Third-order nonlinearities in silicon at telecom wavelengths,” Appl. Phys. Lett. 82(18), 2954–2956 (2003). 3. V. Mizrahi, K. W. Delong, G. I. Stegeman, M. A. Saifi, and M. J. Andrejco, “Two-photon absorption as a limitation to all-optical switching,” Opt. Lett. 14(20), 1140–1142 (1989). 4. D. Lezal, J. Pedlikova, and J. Zavadil, “Chalcogenide Glasses for optical and photonics applications,” Chalcogenide Lett. 1, 11–15 (2004). 5. K. A. Cerqua-Richardson, J. M. McKinley, B. Lawrence, S. Joshi, and A. Villeneuve, “Comparison of nonlinear optical properties of sulfide glasses in bulk and thin film form,” Opt. Mater. 10(2), 155–159 (1998). 6. Y. Ruan, M. Kim, Y. Lee, B. Luther-Davies, and A. Rode, “Fabrication of high-Q chalcogenide photonic crystal resonators by e-beam lithography,” Appl. Phys. Lett. 90(7), 071102 (2007). 7. K. Suzuki, Y. Hamachi, and T. Baba, “Fabrication and characterization of chalcogenide glass photonic crystal waveguides,” Opt. Express 17(25), 22393–22400 (2009). 8. S. A. Schulz, L. O’Faolain, D. M. Beggs, T. P. White, A. Melloni, and T. F. Krauss, “Dispersion engineered slow light in photonic crystals: a comparison,” J. Opt. 12(10), 104004 (2010). 9. M. Bass, ed., Handbook of Optics II, 2nd ed. (McGraw-Hill, 1994). 10. T. P. White, L. O’Faolain, J. Li, L. C. Andreani, and T. F. Krauss, “Silica-embedded silicon photonic crystal waveguides,” Opt. Express 16(21), 17076–17081 (2008). 11. J. Li, T. P. White, L. O’Faolain, A. Gomez-Iglesias, and T. F. Krauss, “Systematic design of flat band slow light in photonic crystal waveguides,” Opt. Express 16(9), 6227–6232 (2008). 12. A. Gomez-Iglesias, D. O’Brien, L. O’Faolain, A. Miller, and T. F. Krauss, “Direct measurement of the group index of photonic crystal waveguides via Fourier transform spectral interferometry,” Appl. Phys. Lett. 90(26), 261107 (2007).
Introduction
The increasing demand for bandwidth in telecommunication systems suggests that all-optical signal processing based on nonlinear effects will become essential in the long term.In silicon, the obvious material for photonic-electronic integration, nonlinear processes are limited by nonlinear loss mechanisms, such as two-photon absorption (TPA) and subsequent free carrier absorption (FCA) [1].At the telecom wavelength of λ = 1540nm, the high TPA coefficient of β 2 0.79cm/GW results in a nonlinear figure of merit of FOM = n 2 /β 2 λ0.37 despite the sizeable nonlinear refractive index of n 2 0.45•10 13 cm 2 /W [2].The nonlinear figure of merit is very useful for comparing the suitability of materials for all-optical signal processing [3] irrespective of the waveguide geometry employed.Chalcogenide glasses, for example, offer a much higher nonlinear figure of merit of typically FOM>>1.These glasses contain one or more of the chalcogen elements (sulphur, selenium or tellurium) from group 16 of the periodic table as a major constituent combined with glass-forming elements such as As, Ge, Sb, Ga, etc. Being composed of covalently bonded heavy elements leads to a window of transparency ranging from the near-well into the middle infrared [4].Their high nonlinear refractive index in the range of n 2 1-30•10 14 cm 2 /W combined with a low two-photon absorption coefficient, typically β 2 <1•10 11 cm/GW, makes them ideal candidates for nonlinear optical applications [5].Previous work on chalcogenide photonic crystals has focused on cavities [6] and more recently, nonlinear optical effects have also been demonstrated in chalcogenide photonic crystal waveguides [7] that used silver doping to increase n 2 (but also β).Clear evidence for the benefits of using photonic crystals in the slow light regime is still missing, however, and a key stepping stone towards this important demonstration is the development of dispersionengineered slow light photonic crystal waveguides in highly nonlinear materials.Here, we report the successful fabrication of dispersion engineered chalcogenide waveguides in the slow light regime and demonstrate low loss operation, both of which are important steppingstones towards efficient nonlinear operation.
Slow light is an important phenomenon for enhancing nonlinear operations, however, it has only been realized so far in high contrast silicon waveguides using various techniques [8].It is not obvious that the same techniques can be applied to lower refractive index systems.We highlight the issue in Fig. 1, which displays the dispersion curves of the even mode of a photonic crystal W1 waveguide for materials with a refractive index of n2.6 (e.g.chalcogenide, dashed), and n3.48 (i.e.silicon, solid) with respect to the light line (straight dashed).It is clear that the useful k-space between the light line and the cut-off point for AMTIR-1 is much smaller than that for silicon.The corresponding wavelength ranges are Δλ87nm for silicon and Δλ28nm for chalcogenide, which clearly limits the performance of the chalcogenides.Firstly, the operation is closer to the band-edge, which can increase linear losses, and secondly, the window for dispersion engineering is much smaller.Despite these considerable limitations, we demonstrate here that dispersion engineering methods can indeed be applied very successfully.Fig. 1.Band diagram for the fundamental mode of a W1 photonic crystal waveguide for silicon (solid curve, n = 3.48) and AMTIR-1 (Ge33As12Se55, n = 2.6, dashed curve).The light cone is outlined as the light grey area which is delimited with the straight dashed line.The dark grey areas outline the useful frequency range that is accessible below the light line and above the cut-off frequency.The corresponding accessible k-range lies between the crossing point of the dispersion curve of the respective material (denoted by the vertical solid lines) and the band edge at k = π/a.It is obvious that the operating range for the lower index contrast chalcogenide system is much smaller than that for the higher contrast silicon system.
Fabrication and loss measurements
Our devices were fabricated from a 300nm layer of AMTIR-1 that had been deposited by thermal evaporation onto 1.5μm of silica on a silicon substrate.The photonic crystal pattern was exposed in ZEON ZEP 520A electron-beam resist using a ZEISS GEMINI 1530/ RAITH ELPHY system.In order to achieve high resolution, an acceleration voltage of 30kV was chosen with a write field size of 100μm and pixel size of 2nm.The electron beam resist was developed using Xylene at 23C and the pattern was transferred into the chalcogenide layer using chemically assisted ion beam etching (CAIBE) with a chlorine/argon gas combination.The sample was heated to T = 115C and a chlorine flow of 2.5sccm was used.For the argon beam, a flow rate of 2.5sccm was used with a beam current of 7.5mA and a beam voltage of 300V.These conditions required an etching time of 9 min and resulted in vertical sidewalls.The remaining electron beam resist was then removed by dimethylformamide.Due to the lower mechanical stability of AMTIR-1 compared to silicon, the use of liquid hydrofluoric (HF) acid to create the photonic crystal membrane had a very low success rate.Figure 2(a) shows a collapsed waveguide taper from the access waveguide to the photonic crystal waveguide.The stresses related to the liquid HF etching step tend to break away the mechanically sensitive taper from the photonic crystal membrane.To address this issue, we developed a vapour HF technique which avoids the disruptive surface tension.Hydrofluoric acid vapour reacts with silicon dioxide according to: The resulting water can condensate on the sample and react with the surrounding HF vapour to form liquid phase HF.In order to avoid such condensation, it is essential to keep the sample at elevated temperature.In fact, controlling the temperature gives a means of finely controlling the etch rate down to a few nm/minute.In order to implement the technique, we use a standard HDPE bottle that was partially filled with a 48% HF solution.The lid of the bottle incorporated a heater.The sample was then glued to the inside of the lid with standard heat sink paste in order to increase the thermal conductivity.The sample was heated up to T = 50 C in order to start the etch process.The temperature was then increased to T = 70C over 3-4h.Controlling both, the temperature and the atmosphere gave fine control over the etch rate and time.Figure 2(b) shows the resulting successful application of HF vapour phase etching to chalcogenide photonic crystal waveguides.The access taper is partially suspended but clearly still connected to the photonic crystal waveguide, in contrast to the device shown in Fig. 2a that was etched using the liquid phase etching process.To avoid sagging of the under-etched access taper, a new design for the access waveguides was used.We widened the access waveguides from 3μm to 6μm and added a 50μm long taper region.Figure 3(a) shows a sketch of this taper layout and Fig. 3(b) shows a corresponding cleaved facet.The remaining 2μm wide silica under the waveguide acts as a supporting pedestal.The advantage of this new design is that no etch mask is required for the HF step, thus reducing the number of processing steps and the exposure of the chemically sensitive chalcogenides to resist developer and remover.To determine the refractive index of the AMTIR-1 layer experimentally, we fabricated a set of 80μm long photonic crystal waveguides with a range of lattice constants, whilst keeping the air hole fill-factor constant.The measured spectra were then matched to 3D MPB simulations, using the refractive index as a fitting parameter.The best match was achieved for a refractive index of n = 2.6 which agrees reasonably well with the literature bulk value of n = 2.54 [9].We then fabricated W1s with a lattice constant of a = 525nm and lengths varying from 95μm to 1095μm with a typical hole diameter of d = 315nm (r/a = 0.3).The high quality of the fabrication process could be assessed from the sharpness of the transitions near cut-off as well as from cut-back measurements which determined the propagation loss (see Fig. 4).The transmission was found to drop by 35dB over a range of 3 nm near the band-edge suggesting our samples were of high quality (Fig. 4(a)).Cut-back measurements confirmed this and showed the loss, determined from plots of transmission versus length (Fig. 4(b)), was 21dB/cm.Given the relative immaturity of chalcogenide photonic crystals, this compares well with losses of 12dB/cm obtained for a silicon membrane W1 waveguide fabricated with the same lithography system [10].
Dispersion engineering in low-index photonic crystal waveguides and experiments
We used the method based on shifting rows of holes to engineer the dispersion curve to produce a slow-light regime as was pioneered in silicon [11] using the group-index-bandwidth-product (GIBP) n g (Δω/ω) as a figure of merit (Fig. 5(a)).As expected for a lower refractive index material, the GIBP reaches a somewhat lower value of about 0.25, compared to 0.3 that can be achieved in silicon.Another difference is that the highest achievable groupindex is around n g 40 rather than >100.Both of these observations highlight the reduced degrees of freedom given by the lower refractive index contrast, which also agrees with the smaller operating window observed in Fig. 1.The GIBP map was used to confirm the experimental design and Fig. 5(b) shows the group-index curves including the four different designs that were used for the fabrication of chalcogenide slow light samples.Figure 6 shows the measured group-index curves for four different designs with a target n g of 20, 30, 35 and 40.The measurements were carried out using Fourier transform spectral interferometry [12].It is clear that despite the lower refractive index contrast, sizeable group index values can be achieved, e.g.n g 40 over a 5 nm bandwidth (Fig. 6d), thus highlighting the suitability of the method.
Conclusion
Chalcogenide photonic crystals are a favorable platform for nonlinear optics due to their high nonlinear figure of merit.Due to their lower refractive index and corresponding weaker confinement, it was not obvious whether the same dispersion engineering techniques previously explored in silicon can be used, and whether similar low losses can be achieved.To investigate these issues, we have fabricated dispersion engineered chalcogenide photonic crystal waveguides and demonstrated losses as low as 21dB/cm.In addition, we have shown that the dispersion engineering toolkit can be applied to the chalcogenide system and have demonstrated slow light waveguides with a group index of n g 40.Given the lower phase index n φ of the waveguide mode, this corresponds to a slowdown factor (S = n g /n φ ) of S20, which is considerable and highlights the potential of the system for nonlinear applications.To our knowledge, this is the first demonstration of systematic dispersion engineering in relatively low refractive index photonic crystal waveguides, in particular in chalcogenides.Furthermore, we have shown the benefits of using vapour phase HF etching for the fabrication of photonic crystal membranes.
Fig. 2 .
Fig. 2. Scanning electron micrographs of AMTIR-1 photonic crystal waveguides at the coupling section to the taper for the access waveguides.The smooth and vertical sidewalls of the holes are shown.(a) The access taper is collapsed due to stress acting on the waveguide and membrane during the liquid HF process.(b) An intact photonic crystal coupling section due to the use of vapour HF.
Fig. 3 .
Fig. 3. (a) Sketch of the taper which allows performing the HF undercut without an additional masking step.The 6μm wide waveguide is tapered down to the waveguide width of 909nm within 50μm.The supporting, not-etched pedestal can be seen.(b) Cleaved facet of the access waveguide after underetching also showing the remaining silica which acts as a supporting pedestal under the access waveguide.
Fig. 4 .
Fig. 4. a) Transmission measurements of the chalcogenide W1 photonic crystal waveguides for the determination of the refractive index.b) Losses extracted from cut-back measurements.
Fig. 5 .
Fig. 5. (a) Map of 3D calculations for the group-index-bandwidth-product (as a function of s1 and s2).The color map displays the calculated GIBP whilst the contours map the achievable group index.(b) Calculated group index curves for modified W1 modes for certain values for s1 and s2.The red areas indicate the constant (ng ± 10%) group index regions. | 3,350.2 | 2011-01-31T00:00:00.000 | [
"Engineering",
"Materials Science",
"Physics"
] |
Barotropic response in a lake to wind-forcing
. We report results gained with a three-dimensional, semi-implicit, semi-spectral model of the shallow water equations on the rotating Earth that allowed one to compute the wind-induced motion in lakes. The barotropic response to unidirectional, uniform winds, Heaviside in time, is de-termined in a rectangular basin with constant depth, and in Lake Constance, for different values and vertical distributions of the vertical eddy viscosities. It is computationally demonstrated that both the transitory oscillating, as well as the steady state current distribution, depends strongly upon the absolute value and vertical shape of the vertical eddy viscosity. In particular, the excitation and attenuation in time of the inertial waves, the structure of the Ekman spiral, the thickness of the Ekman layer, and the exact distribution and magnitude of the upwelling and downwelling zones are all significantly affected by the eddy viscosities. Observations indicate that the eddy viscosities must be sufficiently small so that the oscillatory behaviour can be adequately modelled. Comparison of the measured current-time series at depth in one position of Lake Constance with those computed on the basis of the measured wind demonstrates fair agreement, including the rotation-induced inertial oscillation.
Introduction
Knowledge of water movements is a prerequisite for the study of a multitude of water quality problems of natural and artificial lakes. The computation of the current distribution in homogeneous lakes is commonly performed with the shallow water equations on the rotating Earth. In the present study, density variations are ignored (that is the case for most of the Alpine lakes in winter), so that the wind-induced motion is Correspondence to: Y. Wang<EMAIL_ADDRESS>restricted to the barotropic component for which the internal hydrostatic pressure gradient is exclusively due to the variation of the free surface over its equilibrium value. It is common knowledge that numerical codes integrating these spatially three-dimensional shallow water equations, explicitly in time, are conditionally stable; due to the explicit treatment of the friction (or diffusion) terms, the time step is limited by the smallest mesh size. For most problems of ocean dynamics, since space scales are usually large, the maximum usable time steps can equally be relatively large, and thus, time integration over physically relevant intervals is often possible. Since lakes are small, sufficient spatial resolution requires small mesh sizes limiting the economically justifiable integration time to values below the physically relevant intervals. Thus, implicit or semi-implicit integration routines are needed. For the latter, time steps are still limited by the value of the eddy viscosity and the space steps in the horizontal direction; but this restriction is not severe because the horizontal mesh sizes are usually fairly large. That is why, often, only semi-implicit schemes are applied instead of fully implicit or ADI schemes. There are also many ocean models which include implicit schemes for good reasons, especially in coastal ocean models or when a time-and space-dependent eddy viscosity is computed. Various implicit models have been developed and used by many researchers (e.g. Backhaus, 1983Backhaus, , 1985Backhaus, , 1987Blumberg and Mellor, 1987;Davies, 1987;Bleck et al., 1992;Davies and Lawrence, 1994;Ip and Lynch, 1994;Song and Haidvogel, 1994).
On the other hand, with regard to non-linear advection terms, if the second-order centered finite difference scheme is used (or other traditional high-order schemes, e.g. the spectral method used here), numerical oscillations always occur. The smaller the eddy viscosity is, the smaller the mesh size and time step for stable integration must be. Thus, less numerical oscillations are provoked, so that the small eddy viscosity can still assure stable numerical simulations. The attenuation with time of any oscillating motion increases with growing eddy viscosity. It follows that amplitude and persistence of oscillations are directly tied to the eddy viscosities. 368 Y. Wang et al.: Barotropic response in a lake to wind-forcing Values needed for stable integration are often larger than is physically justified, i.e. computed physical oscillations are faster attenuated than in reality. A study how the vertical eddy viscosity affects the time dependent and steady state wind-induced velocities is, therefore, urgent.
Wind-induced barotropic responses in lakes or the ocean have been studied by many researchers (e.g. Mortimer, 1974Mortimer, , 1980Simons, 1980;Hutter, 1982;Heaps, 1984;Sündermann, 1984). However, for most of these works, if the threedimensional field equations including the nonlinear advection terms are numerically solved, the inertial waves seem to be hardly observable in the numerical results, or they remain only for a fairly short time and then are rapidly damped out. This fact is due to the much larger eddy viscosities used in the models in order to restrain the numerical (not physical) oscillations and thus, assure numerical stability.
We employ the three-dimensional, hydrodynamic, semiimplicit, semi-spectral model SPEM, as originally developed by Haidvogel et al. (1991) and extended for semi-implicit integration in time by Wang and Hutter (1998). This software is able to cope with the above questions in a reasonable way. We propose three functional relations for the vertical distribution of the vertical eddy viscosity: (i) constant, (ii) sinusoidal half wave in an upper turbulently active layer extended to a constant value below and (iii) a linear increase with depth in the upper layer to a maximum continued by a constant value at this maximum below it. All these distributions are based on well justified arguments; here we use these distributions and vary their intensity.
We show that the offset of the surface current to the right of the wind, the Ekman spiral and the thickness of the Ekman layer depend strongly upon both the absolute value and the distribution of the vertical eddy viscosity, as do the windinduced inertial oscillations, the stored kinetic energy and the details in the up-and downwelling distributions of the motion. It is further demonstrated that the numerical values of the eddy viscosities need to be sufficiently small, i.e., at levels of their physical counterparts if the observed dynamics is to be reproduced by the computations. This demonstrates that robust software is needed if the barotropic circulation dynamics is to be adequately modelled. This paper, together with the previous work relating to the baroclinic circulation dynamics in lakes , makes up a complete pattern of forced motion response in enclosed lakes.
The paper is structured as follows: In Sect. 2, the numerical method is briefly introduced. Results are discussed in Sect. 3 for a rectangular basin with a constant depth, and in Sect. 4, for Lake Constance. Comparison of a series of measured results with the numerical simulation of that case is performed in Sect. 5. We conclude with a summary and some remarks in Sect. 6.
Numerical method
The balance law of mass and the balances of linear momentum form the hydrodynamic field equations for the consid-ered fluid system. We apply these hydrodynamic equations in the shallow water approximation, with the Coriolis term and the hydrostatic pressure equation implemented. Thus, the field equations read Here a Cartesian coordinate system (x, y, z) has been used; (x, y) are horizontal, and z is vertically upwards, against the direction of gravity. The field variables (u, v, w) are the components of the velocity vector v in the (x, y, z) directions, φ is the dynamic pressure φ = p/ρ with pressure p and density ρ, f is the Coriolis parameter and g is the gravity acceleration. Eddy viscosities are taken into account by differentiating between the horizontal and vertical directions, ν h and ν v , respectively; thus, the anisotropy effects of the turbulent intensity are considered accordingly.
Wind stress forcing at the water surface, free-slip lateral boundary conditions, and linear bottom friction are used. Along all boundaries the normal component of the current is set equal to zero, which, at the free surface, corresponds to the rigid-lid approximation. With this method, however, a Poisson equation must be solved at every time step to obtain a pressure field which ensures ∇ · (hv) = 0, where h indicates water depth andv is the depth-averaged velocity (see Haidvogel et al. (1991) and Wang (1996) for details).
A semi-spectral model was designed with semi-implicit integration in time to solve the system of differential equations numerically. The model is based on the semi-spectral model SPEM developed by Haidvogel et al. (1991), in which the variation of the field variables in the vertical direction is accounted for by a superposition of Chebyshev polynomials, whereas a centered finite difference discretization on a staggered Arakawa grid is employed in the horizontal direction; it was extended by Wang and Hutter (1998) to account for implicit temporal integration. Due to the small water depths of lakes, in comparison to the ocean, the original SPEM model had to be altered to permit economically justifiable time steps in the computation of the circulation of a lake. In Wang and Hutter (1998) several finite difference schemes, implicit in time, were introduced; that scheme which uses implicit integration in time for the viscous terms in the vertical direction was the most successful one.
For a realistic irregular lake, a vertical topography following the σ -coordinate, and a horizontal shore following the curvilinear coordinate system is employed. By using the so- FIG. 1: Rotation angle Φ between wind and surface current (a) and absolute fluid velocity V surf at the free surface (b). Here Φ and V surf are plotted against ν v (solid curves), A (broken curves), A 0 (dotted curves), respectively, corresponding to the cases (i), (ii) and ( Here and V surf are plotted against ν v (solid curves), A (broken curves), A 0 (dotted curves), respectively, corresponding to the cases (i), (ii) and (iii).
called σ -transformation, a lake domain with varying topography is transformed to a new domain with constant depth, and this region is once again transformed in the horizontal coordinates by using conformal mapping, which maps the shore, as far as possible, onto a rectangle. Uniformity in grid size distribution is intended because numerical oscillations (instabilities) occur preferably on the small scales; however, it is difficult to achieve it in complex geometries. In such cases, to attain an uniform grid as far as possible, a bounding line, which deviates in some segments from the actual lake boundaries, is used for the conformal mapping. In these segments the actual boundaries can only be approximated by a step function, and the land areas must be masked with a special technique (Wilkin et al., 1995). For the computation in a rectangular basin, such a conformal mapping in the horizontal direction is not needed; however, for the Lake Constance, it is required, but not unique. In this case, a conformal mapping is performed for a bordering line which deviates in some segments from the actual shore line, so that a more uniform grid system is obtained. Then, in the numerical computations, the land areas within the grid system must be excluded by the aforementioned masking technique. This numerical code has proved its suitability in several lake applications, including diffusion problems , substructuring procedures and wave dynamics in baroclinic wind-driven circulation dynamics . Here, the application of this code to the barotropic motions is performed. It will be seen in the following numerical results that almost all typical features in a homogeneous lake, e.g. inertial waves, Ekman spirals, can be reproduced by this model.
Barotropic motions in a rectangular basin
A rectangular basin of 65×17 km 2 extent with a 100 m depth corresponds to the space scale of Lake Constance; we assume homogeneous water, initially at rest, and subject to external wind forcing. This wind blows in the long direction of the rectangle, uniformly in space, Heaviside in time, and with strength 4.5 m s −1 (10 m above the water surface), corresponding to a wind stress of approximately 0.0447 Pa at the water surface. This is the wind-forcing we apply throughout the paper, except for Sect. 5. Integration starts at rest until a steady state is reached. We shall implement the discretization with 10-30 Chebyshev polynomials; the exact number depends on the magnitude of the vertical eddy viscosities (the smaller the vertical eddy viscosities, the more Chebyshev polynomials are needed to achieve stable and convergent numerical computations). We choose x = y = 1 km and let the numerical value of the horizontal eddy viscosity of momentum be constant ν h = 1 m 2 s −1 , because it turned out that the numerical values of ν h are not very crucial (Wang, 1996), and this value is generally reasonable (Csanady, 1978;Hutter, 1984). The insensitivity of the horizontal eddy viscosity can be expected and easily recognized because the horizontal variation of the velocity field is relatively small, in comparison with its vertical variation, except in the vicinity of shores and when tracer diffusion is considered.
3.1 Surface steady current in relation to distributions and amplitudes of the vertical eddy viscosity Three chosen vertical eddy viscosities are prescribed as follows: -case (i): ν v ∈ [0.005, 1] m 2 s −1 , -case (ii): FIG. 1: Rotation angle Φ between wind and surface current (a) and absolute fluid velocity V surf at the free surface (b). Here Φ and V surf are plotted against ν v (solid curves), A (broken curves), A 0 (dotted curves), respectively, corresponding to the cases (i), (ii) and (iii) -case (iii): where A 0 ∈ [0.01, 1] m 2 s −1 .
In case (i), the vertical eddy viscosities are spatially constant, but their values are varied. Case (ii) corresponds to a z-dependence of ν v , where ν v is large in the upper layer, growing first with depth until it reaches a maximum in 30 m and then rapidly decreases to the value 0.01 m 2 s −1 at 45 m depth and below. This form corresponds with the vertical eddy viscosity distribution obtained by Svensson employing an one-dimensional k − ε model (Svensson, 1979) and by Güting, using a three-dimensional k − ε closure condition (Güting, 1998). Case (iii) assumes a linear variation of ν v from the free surface to 10 m depth below, in which ν v is kept constant at the value ν v = 0.03 m 2 s −1 (Csanady, 1980;Madsen, 1977). The maximum values of ν v can be varied by varying the free parameters A and A 0 . Choices like these are considered realistic and have, to some extent, been analysed with a linear equation system, in which the nonlinear advection terms are not included in finite depth oceans (see Heaps, 1984).
One typical result of the Ekman problem is the angle between the direction of the wind and the surface current in steady state. In the original problem, an infinite ocean with constant vertical eddy viscosity exists, and the surface current v surf is 45 • to the right of the wind. Results of our nonlinear calculations are summarized in Fig. 1. In case (i), the maximum value of is 42.5 • for ν v = 0.005 m 2 s −1 , and decreases rapidly with increasing vertical eddy viscosity to a value as low as 4 • for ν v = 1 m 2 s −1 . It can never reach 45 • as it does (independent of the value of ν v ) in the infinite ocean, because of the finiteness of the rectangular basin that induces a geostrophic flow which has the tendency to reduce . This simply indicates how significantly the circulating motion in a finite basin can depend on the absolute values of the eddy viscosities. Even more surprising are the results of case (ii). Here varies non-monotonously with A: For A ≤ 0.12 it decreases with growing A, while it grows with A when A ≥ 0.12. Case (iii) presumes, for most values of A 0 , a very active turbulent near-surface layer and illustrates that for such a stiff situation, the typical dependence of can even be reversed. In Fig. 1b the moduli of the surface velocities v surf are shown as functions of the eddy viscosities. As expected, they rapidly decrease with increasing eddy viscosities. These results demonstrate how critical the distribution and absolute values of the eddy viscosities are; reducing the numerical diffusion to levels below the physical values is, therefore, compelling, if results are not to be physically falsified. Figure 2 shows the time series of the vertical velocity component w in 30 m depth at the four nearshore midpoints, as sketched in panel (b) of Fig. 2, subject to three different magnitudes of the constant vertical eddy viscosity, respectively. Immediately after the west wind sets in, the motion starts rapidly with an upwelling at the western end (Fig. 2a) and a downwelling at the eastern end (Fig. 2b), due to the direct effect of wind stress at the surface. In the northern midboundary point, there occurs an upwelling (Fig. 2c), while in the southern shore, a downwelling emerges (Fig. 2d). This behaviour is, clearly, due to the Coriolis force that causes a horizontal velocity drift to the right and therefore, is responsible for the downwelling (upwelling) at the southern (northern) shore on the northern hemisphere. The motion is characterized by oscillation, with a period of approximately 16.3 hours, which can obviously be identified as the inertial period using a Coriolis parameter approach for Lake Constance, f = 1.07 × 10 −4 s −1 . The smaller the vertical eddy viscosity is, the stronger the superimposing oscillations will be, and the longer the oscillations will persist. For the larger value ν v = 0.02 m 2 s −1 , only one day after the wind starts, the inertial oscillations are damped out, whereas the oscillations for ν v = 0.001 m 2 s −1 can still be clearly identified after six days. For this small eddy viscosity, due to the emerging strong oscillation, even a short-time downwelling (upwelling) at the western (eastern) shore can be observed.
Time series of the velocity components
In Fig. 3, the time series of the horizontal velocity components u (a) and v (b) for a medium value of ν v = 0.005 m 2 s −1 in the center of the basin, at different depths, are displayed. Steady motion is reached approximately after five days, but initially, an oscillating motion can be seen at all water depths, however, with decreasing amplitude as the depth increases. It is also interesting that the horizontal velocity that oscillates above 30-40 m, with the opposite phase from below this depth, corresponds approximately to the Ekman depth. This can be explained by the use of a linear equation system. It is known that for a linearized equation system, the horizontal motion consists of a drift current directly due to windstress, and a geostrophic current due to the surface slope (the surface pressure gradient) that can be described as follows: Neglecting the horizontal friction terms, the linearized horizontal momentum equations (2) and (3) then take the form with ν v = const. For the homogeneous case the dynamic pressure is independent of z. The local solution u(z, t), v(z, t) of (5) (depending only parametrically on x, y) may be written as where u 1 (t), v 1 (t) is the geostrophic current which is a solution of due to the surface pressure gradient, whilst u 2 (z, t), v 2 (z, t) is the depth-dependent or frictionally induced velocity components With the suddenly imposed windstress τ 0 at the surface in the positive x-direction, and with a surface pressure gradient (a surface slope) ∂φ/∂x in the positive x direction and bottom stresses neglected, (7) and (8) have the local solutions of the geostrophic current and of the drift motion, respectively. The solution (10) was first given by Fredholm (Ekman, 1905). It is obvious that the two motions have opposite phase. It is of interest to note that such a combination of the currents can give a satisfactory approximation for the vertical current profile in regions far away from the lateral boundaries, even in those cases where nonlinear effects cannot be neglected in solving the horizontal circulation problem (Nihoul and Ronday, 1976). Above the Ekman depth, the drift motion (u 2 , v 2 ), induced directly by windstress, is dominant, whilst below this depth, the Ekman drift motion dies out and only the geostrophic current (u 1 , v 1 ) remains, which is independent of depth under homogeneous, hydrostatic conditions. From the solutions of such a simplified linear, horizontal momentum equation system, one can easily see that these two contributions to the current oscillate with opposite phase; at 30 to 40 m depth, the two oscillations possess comparably large amplitudes, and hence, the superimposing oscillation is hardly visible at this depth (see the 40m-depth curves in Fig. 3). Such oscillations have been discussed by Krauss (1979) and Csanady (1984). The same time series of the velocity components u and v, as shown in Fig. 3, are displayed again in the form of hodographs in Fig. 4 for 0, 5, 10, 20 and 60 m depths. The small circles along the curves mark time intervals of 1/4 of the inertial oscillations, which is approximately 4 hours. The arrows represent the velocity vectors in the steady state. This graph is very similar to the result obtained by Krauss (1979) in an infinite channel of rectangular cross-section. In the upper layer (0, 5, 10, 20 m), the wind-induced drift current, due to sudden wind, approximates the steady state in the form of inertial oscillations. The spiral in 60 m depth reflects the adaptation of the current field to the geostrophic current. The reflection of the water surface, due to the sudden wind, produces a sudden change of the pressure field, in the entire water column. The adaptation of the current field, which was zero, to the new condition, occurs in the form of inertial waves. Similar to those in the surface layer, these waves start together with the wind because the pressure gradient is felt immediately. In the hodograph for 20 m depth, one can see, at first, (perhaps during the first two hours) the same northwestward current as in 60 m depth; at this initial time, only the geostrophic current exists in 20 m, but when the surface drift current reaches this depth, an abrupt change in direction occurs. It can be explicitly detected that the wind-induced current in the upper layer (in 0, 5, 10 and 20 m depth) oscillates with opposite phase of the geostrophic current in the lower layer.
Ekman spirals
In Fig. 5 the vertical velocity is negligible, for example, in the middle of a lake, the motion under homogeneous conditions can be considered a linear combination of (i) the geostrophic current, constant in the vertical, (ii) a bottom current deviation indicating the bottom Ekman layer, decaying exponentially away from the bottom which, together with the geostrophic current, satisfies the bottom boundary condition, and (iii) a surface-drift current (the Ekman layer) decreasing rapidly with depth and satisfying the surface dynamic boundary condition. In an infinite ocean with infinite depth, there exists only the surface-drift current. As is evident from Fig. 5a, for a constant vertical eddy viscosity, the drift current at the surface, in steady state, is directed 45 • to the right of the windstress (on the northern hemisphere). In a bounded rectangular basin, due to the surface slope under windstress, a geostrophic current shares the wind-induced motion. Subject to a western wind, a surface pressure gradient towards the east is built up, and hence, the caused geostrophic current points towards the north, as displayed in Fig. 5b; the current below the Ekman depth is approximately 60 m for ν v = 0.02 m 2 s −1 . If the northward geostrophic current is eliminated from the motion in Fig. 5b, a velocity field is obtained which is almost identical with that of the infinitely large ocean (Fig. 5a). If bottom friction cannot be ignored, the bottom Ekman layer occurs (Fig. 5c). The surface drift current, as well as the bottom current deviation, decreases rapidly with distance from the boundaries and a decay rate depending on the magnitude of the vertical eddy viscosity, as can be seen from a comparison of Fig. 5c,e. We also note that the northward geostrophic current automatically reduces the off-set angle of the surface current relative to that of the Ekman drift, a fact visible when comparing Figs. 5a and 5b. Finally, Figs. 5d,f display the Ekman spirals for a position near the midpoint of the southern shore (the distance from shore is 500 m) that is computed for two values of the verti-cal eddy viscosities. Here too, the value of the vertical eddy viscosity considerably affects the distribution of the velocity.
Total kinetic energy in relation to the vertical eddy viscosity and the wind-fetch
In Fig. 6, the time series of the total kinetic energy in the homogeneous rectangular basin of constant depth, subject to constant western (a) and southern (b) wind, respectively, are displayed, as previously computed with three vertical eddy viscosities, as indicated. As has been seen in the time series for the horizontal velocity, the inertial motions of the transient energy persist much longer, when ν v = 0.005 m 2 s −1 than when ν v = 0.02 m 2 s −1 , and this difference is even more distinct between ν v = 0.005 m 2 s −1 and ν v = 0.001 m 2 s −1 . The total kinetic energy stored in the basin for ν v = 0.001 m 2 s −1 is much larger than for ν v = 0.005 m 2 s −1 or ν v = 0.02 m 2 s −1 , due to the much larger energy input from windstress because of the much larger water velocity at the free surface (or more precisely, its component in the wind direction), and the likely smaller dissipation for the smaller vertical eddy viscosity. Comparison of Figs. 6a,b shows that the kinetic energy subject to the longitudinal, western wind is only slightly larger than that which is subject to the transverse, southern wind. The dependency of the total kinetic energy on the wind direction is worked out in terms of a polar diagram, which plots the variation of the total kinetic energy stored in the stationary circulation of the rectangular basin as a function of the direction of the wind. We compute the water motion subject to a constant wind from different directions (in intervals of 10 • ) for a fixed vertical eddy viscosity ν v = 0.02 m 2 s −1 . The polar diagram of the total kinetic energy is displayed in Fig. 7a. It is seen that the kinetic energy in the rectangular basin with constant depth depends only marginally on the wind direc- Fig. 7. The dashed circles show the maximum value of the total kinetic energy, which is 9.71 × 10 8 N m for the case with consideration of the Coriolis force (a), but 8.57 × 10 9 N m for the case without the Coriolis effect (b). Fig. 8. Same as Fig. 7, but for a rectangular basin with larger aspect ratio. Here, the basin dimension is 60 km × 5 km × 100 m instead of the dimension of 65 km × 17 km × 100 m in Fig. 7. The dashed circles show the maximum value of the total kinetic energy, which is 9.71 × 10 8 N m for the case with consideration of the Coriolis force (a), but 8.57 × 10 9 N m for the case without the Coriolis effect (b).
tion. The kinetic energy attains its maximum for longitudinal, western wind, while for a transverse southern wind, the kinetic energy is minimum, although the maximum and the minimum differ from each other only by approximately 15%. As expected, the dependence of the kinetic energy on the wind direction is anti-symmetric about the E-W or S-N axis. It is also seen that the kinetic energy subject to southwest or northeast wind is larger than that for northwest or southeast wind. This loss of symmetry is due to the effect of the Coriolis force, which can be demonstrated by means of repeating the computations, but under the neglect of the Coriolis force. The analogous polar diagram of the total kinetic energy, without consideration of the Coriolis force, is displayed in Fig. 7b. In this case, the dependence of the kinetic energy on the wind direction is symmetric about the E-W or S-N axes. When the Coriolis force is ignored, the kinetic energy is almost an order of magnitude larger. One likely reason may be that the water motion decays exponentially with increasing depth when the Coriolis force is accounted for, while the velocity decreases only linearly with depth, if the Coriolis force is neglected. The other reason may be that in the absence of the Coriolis force, the water motion at the free surface is mainly along the wind direction, hence, there is more energy input from the windstress. When the Coriolis force is neglected, the maximum kinetic energy occurs for a wind blowing in the diagonal direction of the basin. In fact, the difference in the total kinetic energy depends primarily on the wind-fetch, i.e. the integrated distance over which windaction takes place, taken along the wind direction. For diagonal winds, the wind-fetch is at maximum. In the presence of rotation, the integral for the fetch would probably have to be taken over the Ekman layer. This may be why the maximum of the kinetic energy is not reached for diagonal winds.
The stronger dependence of the total kinetic energy on the wind direction can be expected for a basin with a larger aspect ratio (e.g. a larger variation of the wind-fetch). In order to demonstrate this point, we repeat the computations for a basin with a larger aspect ratio. A basin dimension of 60 km × 5 km × 100 m is used instead of 65 km × 17 km × 100 m. The results are displayed in Fig. 8. In this case, the total kinetic energy depends on the wind direction much more intensively than in Fig. 7. Obviously, the basin must be very long and narrow before a directional dependence of the total kinetic energy is appreciable.
In the next section we will see that the response of Lake Constance, subject to a wind, is much more conspicuously dependent on the wind direction with regard to the total kinetic energy, even though its geometry is close to the rectangular basin in Fig. 7. This means that such a dependence on the wind direction depends not only on the wind-fetch, but also on the bathymetry of the lake basin.
Barotropic circulation in Lake Constance
Lake Constance consists of three basins: Obersee,Überlinger See and Untersee, but the Untersee is separated from the other two basins by a 5 km long channel; we shall be concerned here with the ensemble Obersee+Überlinger See, for brevity, also referred to as Lake Constance. It is approximately 64 km long and 16 km wide with has a maximum depth of 253 m and an approximate mean depth of 100 m, as shown in Fig. 9.
A study analogous to that above was also performed for Lake Constance with the same number of grid points (65 × 17) as for the rectangle. Computations were also done for ν h = 1.0 m 2 s −1 , ν v = 0.02 m 2 s −1 and ν v = 0.005 m 2 s −1 , respectively. In the second case, the larger number of polynomials, namely N = 25 instead of N = 10, was needed to achieve stable and convergent numerical integration. Let a spatially uniform, temporally constant wind with a wind speed of 4.5 m s −1 blow from Northwest (305 • (NW) in the longitudinal direction of the lake) until steady state conditions are established.
Horizontal distribution of the steady currents
The horizontal variation of the currents at the surface and at depth levels 10, 20 and 40 m is depicted in Fig. 10. The horizontal motion is displayed by arrow-diagrams. In the central part of the lake, the surface current is deflected by about 50-60 • to the right of the downwind direction and amounts to 5 cm s −1 , on average. The boundary currents somewhat off the northern and southern shore increase to a magnitude of 8-10 cm s −1 and are directed parallel to the boundary.
As is typical for drift currents, the direction of the motion in the open lake continuously turns to the right on with increasing depth and decreasing velocities. At 10 m depth and far away from the shores, the current deflection to the right of the wind exceeds 90 • , and the average velocity decreases to 3-4 cm s −1 . At the 40 m depth, the motion in the middle of the lake is nearly reflected to the upwind direction. However, above the 40 m depth, the shore currents are still basically parallel to the shores with the exception of the northeastern shore, where the motion below the 20 m depth is almost in the upwind direction. Near the northeastern corner, a cyclonic gyre is visible.
In order to show the general influence of the depth configuration on the wind-driven circulation, the vertically integrated volume transport is displayed in Fig. 11 ume transport streamlines. The circulation consists, in principle, of two cells which rotate in such a way that the transport is downwind along the shores, orientated roughly in the direction of the wind. The central part of the circulation shows a net transport with an upwind component which results from the geographic current (caused and bounded by the bottom topography); with the bottom current prevailing over the surface drift current. In contrast, the drift current dominates the other currents in the shallow, nearshore region. Due to the Coriolis force, subject to a western wind forcing, the sloping bottom causes an intensification of the northward flow component at the northwestern shore as well as at the southeastern shore (Serruya et al., 1984;Wang, 1996). If there would be no Coriolis effect, the circulation would consist of two gyres, which are located almost exactly to the north and south of the Talweg, rotating in the clockwise and anticlockwise directions, respectively.
Ekman spirals and time series of the horizontal velocity
The vertical variation of the current depends strongly on the eddy viscosity. We display in Fig. 12 two steady Ekman spirals at the midlake positions of theÜberlinger See (above) and the Obersee (below), as they form, for an impulsively applied uniform wind from 305 • (NW) (approximately in the long direction), and as obtained with the two different indicated eddy viscosities. Those Ekman spirals are considerably affected by the ν v -values. The turning of the arrows which make up the spirals, also indicates that the surface Ekman boundary layer is thinner for the smaller value of the eddy viscosities (right panel) than for the larger value (left panel).
Equally interesting is the comparison of the time series of the horizontal velocity components u and v for various depths at the midlake positions of theÜberlinger See and the Obersee, as displayed by the hodographs in Fig. 13 and obtained with ν v = 0.005 m 2 s −1 . At both positions, transient oscillations can be discerned with the inertial period of approximately 16 h. The oscillations can be seen at all water depths, however, with decreasing amplitude as the depth increases. Furthermore, they die out before two inertial periods in theÜberlinger See, but only after 5-6 inertial periods in the Obersee. The reason is the smaller size of thë Uberlinger See and, therefore, the enhanced frictional resistance due to the lake bottom and the side shores. In the Obersee, the transient velocities oscillate around a slowly FIG. 13: Hodographs, i.e., time series of the horizontal velocities in the middles of theÜberlinger See (a) and the Obersee (b) in 0, 5, 10 and 60 m depths. The motion is set up from rest. The small circles mark time intervals of approximately four hours. In order not to confuse the spirals at 20 m and 60 m depth the 20 m hodographs are shown separately in the panels (c) and (d) for the two positions.
Total kinetic energy [×10 9 Nm] FIG. 13: Hodographs, i.e., time series of the horizontal velocities in the middles of theÜberlinger See (a) and the Obersee (b) in 0, 5, 10 and 60 m depths. The motion is set up from rest. The small circles mark time intervals of approximately four hours. In order not to confuse the spirals at 20 m and 60 m depth the 20 m hodographs are shown separately in the panels (c) and (d) for the two positions.
Total kinetic energy [×10 9 Nm] varying mid-velocity at each depth; this is different from the corresponding behaviour in the rectangular basin where the motion oscillates approximately around the steady state (compare with Fig. 4). This difference is due to the complex topography of Lake Constance. From the spirals at larger depth (e.g. 20 m), during an initial time interval after the wind started, a sudden change in the direction of the motion can be seen when the wind-induced surface drift flow reaches this depth. Before this time, only the geostrophic current exists at this depth which, due to the restriction of the bottom topography, is along the Talweg. This is different from the behaviour of the flow in a basin with constant depth, in which the geostrophic motion is basically perpendicular to the wind direction or surface pressure gradient (90 • to the left on the northern hemisphere).
Total kinetic energy in relation to the vertical eddy viscosity and the wind direction
The time series of the stored, total kinetic energy in Lake Constance for NW and SW wind, ν v = 0.02 m 2 s −1 and ν v = 0.005 m 2 s −1 (Fig. 14) show that the inertial oscillations persist longer, and the value of the total kinetic energy is much larger for the smaller ν v -value (as shown before for the rectangular basin). More interesting is the comparison of the kinetic energies for the case of longitudinal NW wind and transverse SW wind (Fig. 14a,b). For the transverse wind, the temporal inertial oscillations persist much longer than those associated with the longitudinal wind, whereas the magnitude of the kinetic energy subject to the longitudinal wind is much larger. The variable response of the lake to constant wind forcing from different directions at 10 • interval is represented in the polar diagram of the kinetic energy uptake in Fig. 15a, computed for ν v = 0.02 m 2 s −1 . The stored, total energy depends very strongly on the wind direction. The geographical direction of maximum exposure coincides approximately with the longitudinal direction of the lake. The kinetic energy under a transverse wind is only one fifth of thekinetic energy for a longitudinal wind. The strong dependence of the total kinetic energy on the direction of the wind should be mainly due to the topography, and less to the dimension of the lake, since Lake Constance has a comparable aspect ratio to the rectangular basin in Fig. 7, in which the dependence on the direction of the wind is much weaker. This dependence of the total kinetic energy on the direction of the wind is similar, to some extent, to that obtained by Serruya et al. (1984), where only the kinetic energy of the two-dimensional vertically integrated net transport was calculated. The dependence of the kinetic energy on the wind direction, without consideration of the Earth's rotation, is also displayed in Fig. 15, in panel b. In this case, the total kinetic energy is much larger than the kinetic energy with the Coriolis force, while the dependence of the kinetic energy on the direction of the wind is somewhat weaker; here, the kinetic energy under a transverse wind is nearly half as large as the kinetic energy for a longitudinal wind.
Vertical distributions of the steady currents
The isotachs of the three velocity components in a steady state on the cross-section through the center of theÜberlinger See are presented in Fig. 16. The wind is blowing in the direction of the longitudinal axis of the lake, perpendicular to the plane of the graphs upwards. The most distinguished feature of the motion is a nearshore coastal jet in the direction of wind (Fig. 16a). The change of the horizontal velocity with depth happens primarly in the upper layer and in the middle of the cross-section (Fig. 16a,b). At lower depths, where the geostrophic current plays an important part in the motion, the velocity components do not significantly change with depth. Due to the effect of the Coriolis force, for a west wind, there occurs a downwelling along the southern shore (left side in Fig. 16c) and an upwelling along the northern shore (right in Fig. 16c), however, the vertical velocity component is much smaller than the horizontal components. In the zone far away from shore, the vertical velocity component is practically negligible. Figure 18 displays the transverse variations of the longshore transports. These are the depth integrals of the positive and negative velocities in the x-direction, as functions of the y-coordinates in the cross-section, with the wind direction (positive, broken lines) and against the wind (negative, dotted lines), in cross-sections through the centers of thë Uberlinger See and the Obersee, respectively. From Fig. 18, the nearshore coastal jet in the direction of the wind can be seen more clearly, not only in theÜberlinger See, but also in the Obersee. Close to the shore, the transport is almost only in the direction of the wind; far away from shore, where the motion is dominantly in the upwind direction, the transport is against it. This feature is due to the approximate parabolic shape of the cross direction.
If the positive and the negative transports are added, the net volume fluxes per unit width (the solid lines in Fig. 18) are obtained. These lines would be obtained with a depth integrated model. Obviously, the horizontal integration of the net volume flux along the cross-section must vanish in steady state.
The vertical distributions of the horizontally integrated transport at the centers of theÜberlinger See and the Obersee can be extracted from Fig. 19. It can be seen that the transport occurs primarly in the top 80 m, especially in the top 40 m. The transport in the wind direction (positive, broken lines) assumes its maximum at the surface, while the maximum of the transport against the wind (negative, dotted lines) occurs at approximately 20-30 m depth. Below 100 m there exists only the transport against the wind with very small values. The sum of the two fluxes for a fixed depth yields the net volume flux per unit depth, which is also displayed as solid lines in Fig. 19. Obviously, its vertically integrated total flux through a cross-section must vanish in a steady state. It is also obvious that these results cannot be obtained with the use of a vertically integrated model.
In Figs. 17, 20 and 21, graphs are displayed for the isotachs of the three velocity components in a cross-section of theÜberlinger See (Fig. 17) as well as the horizontal and vertical distributions of the longshore transport in the two indicated cross-sections of theÜberlinger See (Fig. 20) and the Obersee (Fig. 21), respectively. This is also the case in Figs. 16, 18 and 20, but now with a smaller vertical eddy viscosity ν v = 0.005 m 2 s −1 . Principally, they are very similar to those subject to ν v = 0.02 m 2 s −1 . The most distinguishing differences is that for the smaller vertical eddy viscosity, the surface layers of positive x-velocity are much smaller (Figs. 17 and 21) than in Figs. 16 and 19, while the absolute values of the velocity are much larger, and the two-dimensional depth integrated model is less justified (compare Figs. 18 and 20).
Comparison of measured values and computational results in Lake Constance
Unfortunately, there are insufficient data available which would allow validation of the model. Nevertheless, a partial check of the reliability of the numerical method was possible with data from 16 February 1993 ( weak. On 19 February, a strong wind was initiated; its amplitude reached 8 ms −1 and it persisted for the remainder of the period. For the same period, the water velocity and its direction at the "Mainauschwelle" (near the island Mainau) at 80 m depth were also measured; they are displayed as solid curves in Fig. 23a,b. In the first 80 hours, the water velocity was less than 1.1 cm s −1 , which was below the threshold of the current meter. At the fourth day, strong currents started with superimposed oscillations of a period of approximately 16 hours, which can be interpreted as inertial oscillations. The flow direction superimposed by oscillations is approximately 300-340 • (NW).
Here we simulate, numerically, the corresponding water motion with the measured wind input and check if the measured water velocity can be reproduced by the computed velocity field. In winter (here, in February), the water density in Lake Constance can be considered to be homogenous (Bäuerle et al., 1998). For this simulation the measured wind at the station "Boje Mitte" is used and extrapolated to the entire lake. This uniformity is certainly unlikely to be realistic, but we apply it due to lack of better knowledge. The shear stress at the water surface can be calculated with the aid of the classical drag formulas where V wind is the wind speed 10 m above the water surface, ρ a the air density (ρ a = 1.225 kg m −3 ) and c 0 a dimensionless friction coefficient. The specification of c 0 is not unique and varies from author to author. A typical value for a weak or medium wind strength (V wind < 10 m s −1 ) is c 0 = 1.8×10 −3 (Lehmann, 1992). The wind speed measured at 4.4 m above the lake surface (Fig. 22) must be converted to the value at 10 m. Assuming a logarithmic wind profile, yields where z 0 is a measure for the corrugation of the water sur-face (roughness length). In the computations, we choose z 0 = 1.0 × 10 −4 m. The vertical eddy viscosity is assumed constant with value ν v = 0.02 m 2 s −1 and the simulation is started from rest, although at the initial time, a small motion must exist in nature.
The computed water velocity and its direction in 80 meters depth, at the position "Mainauschwelle" where the measured time series are plotted in Fig. 23, are also displayed in Fig. 23 as dashed lines. During the first three days, when a weak wind prevailed, the measured and computed current speeds could not be compared, but their directions could (Fig. 23b), and, they deviated considerably from one another. This is most likely due to the fact that the computations started from a state of rest, but there was some (small) motion in nature (however, not caught by the current meter) which was not considered in the simulation. A relatively strong wind is needed to bring the computed and the measured velocities together. This strong wind occurs after 80 hours when both measured and computed current speed and direction coincide quite well with one another. Only at the seventh day do the current speeds of the computations differ from the respective measured values. The computed current orientation at this depth lies between 300-340 • (towards Northwest), which is in fair agreement with the measured values (compare the solid and dashed lines of Fig. 23b), which are basically around 320 • , superimposed by oscillations. Given the measuring technique and the bold extrapolation of the wind over the entire basin, a better agreement can hardly be expected.
In Fig. 24a,b we have plotted time series of the two horizontal velocity components at several depths in the midlake position in the Obersee, as indicated in the inset. Evidently, the strong wind commencing after 80 hours generates equally strong oscillations; with a period of almost exactly 16.3 hours, they are most likely, inertial oscillations. On the other hand, comparison of the water velocity at 80 m depth at the Mainauschwelle (Fig. 23) with the driving wind suggests that these oscillations might simply be the direct response to the wind forcing. In order to confirm that the oscillations exhibited by the time series of water velocity are indeed inertial oscillations due to the Coriolis force, the same computation but without consideration of the Coriolis force was repeated. For this case, the computed water velocities at the Mainauschwelle, at 80 m depth, and at the center of the Obersee, at several depths, are shown in Figs. 25 and 26. Comparison of Figs. 23 and 25 shows that without the Coriolis term, the computed velocity at the Mainauschwelle does not exhibit the oscillations shown by the measurements. Moreover, the direction of the current is toward 260 • , which is substantially different when one accounts for the Coriolis effects. Similar differences are also seen when Figs. 24 and 26 are compared. The computed current oscillations at the midpoint position of the Obersee are not established without the Coriolis force. These facts should be sufficient demonstration that the measured oscillations are indeed rotational effects due to the motion of the Earth.
Concluding remarks
In this paper, wind induced barotropic circulation in lakes was studied, first, from a more fundamental point of view, using a rectangular basin with constant depth, but later, with Lake Constance, a medium-size Alpine lake. viscosities were varied in three different ways, as suggested by other studies, and the steady Ekman problem was solved. The following results were obtained, each demonstrating the significance of the eddy viscosity: -The direction of the surface water velocity at the midpoint of a rectangular basin relative to that of the wind, the so-called wind set-off, depends strongly on the absolute value and vertical distribution of the vertical eddy viscosity. It cannot be concluded that the current set-off (to the right on the northern hemisphere) decreases with increasing eddy viscosity. Depending upon the vertical distribution of the eddy viscosity, opposite or even nonmonotonic behaviour may occur (Fig. 1a).
-The surface current speed, however, decreases monotonically with an increase in the eddy viscosity (Fig. 1b).
-The amount of kinetic energy stored in the water depends equally upon the absolute value of the eddy vis-cosity (but less on its vertical distribution). Its transient behaviour from a state of rest to a steady state is characterized by oscillations that are attenuated in time, but equally, the smaller the vertical eddy viscosity is, the larger the oscillations. This is true for the rectangle (Fig. 6), as well as Lake Constance (Fig. 14).
-The wind-directional dependence of the total kinetic energy depends strongly on the wind-fetch and the lake bathymetry. For a rectangle with constant depth and a width to length ratio of 0.25, this dependence is less than approximately 10% (Fig. 7); for Lake Constance, it is substantial (Fig. 15).
-The Coriolis force implies a significant reduction of the total kinetic energy that can be established in a basin due to wind forces, when compared to the case when rotational effects are neglected (Figs. 7, 8 and 15). -The vertical distribution of the water velocity, i.e., the Ekman spiral, depends considerably upon the absolute values of the eddy viscosities. Not only the Ekman depth is affected by these values, but equally so, the orientational distribution of the horizontal current with depth (Figs. 5, 12).
-Inertial oscillations are more easily established in open water than bounded channels; this demonstrates the significance of the frictional effects due to boundaries (Fig. 13).
All these results support the conjecture that it is important for adequate reproduction or prediction of observed current fields in lakes to use the correct orders of magnitudes of the numerical eddy viscosities. On the other hand, the eddy viscosity should be chosen according to physical considerations, and simultaneously, ensure numerical stability of a simulation. On the other hand, if the eddy viscosity desired by numerical stability is much larger than the values suggested by measurements, the numerical code needs to be adjusted, especially with some modern numerical treatments of the nonlinear advection terms. Inappropriate numerical schemes of these advection terms are often the reason for the production of numerical oscillations, which require large eddy viscosities in order to prevent their development and hence, assure numerical stability. At present, we are making an effort in this aspect. It is exactly the problem in many existing threedimensional lake circulation models, that in order to reach numerical stability, the eddy viscosities must be chosen to be much larger than physically acceptable so that the physical oscillations are also rapidly damped away or are even indiscernible. It has been clearly demonstrated by this model that observed inertial oscillations due to the rotation of the Earth persist long and are slowly attenuated. That these inertial waves are indeed a dominant effect is shown in a restricted comparison of the measured current at 80 m depth for Lake Constance. Neglecting the Coriolis effects does not yield re-sults that can be compared with the measured ones. This wave dynamic, in turn, determines the advective properties and thus, the transports of nutrients and pollutants both horizontally and vertically. | 12,481 | 2001-03-31T00:00:00.000 | [
"Environmental Science",
"Physics"
] |
NOISE CONTROL OF THE BEGINNING AND DEVELOPMENT DYNAMICS OF FAULTS IN THE RUNNING GEAR OF THE ROLLING STOCK
In contrast to traditional systems for monitoring fault of the running gear of the rolling stock, this paper proposes a technology of noise control at the onset of defects. The authors consider the possibility of creating an intelligent system that can perform noise diagnostics with the indication of the beginning of the latent period of the initiation of typical defects preceding faults. To this end, using the noise technology, sets of reference informative attributes are created in the training process. The reference sets, in turn, are used to determine the condition of the object at the beginning of the development of defects by comparing them with current noise estimates. It also allows controlling the dynamics of the development of defects.
INTRODUCTION
The main condition for ensuring the safety of train movement in railway transport is the reliable and fail-safe operation of rolling stock. To ensure the required reliability of the rolling stock, it is necessary to constantly control the technical condition of its running gear. Reliable information about its technical condition is provided by technical diagnostic systems. Various diagnostic systems are currently used to assess the technical condition of the running gear of rolling stock in motion (based on the principle of their use: stationary, airborne, portable and incorporated directly into the controlled object, etc.). The main goal of technical diagnostics is to determine the type and location of defects. Vibration parameters, pressure, force, voltage, resistance, pulses, time intervals, etc. are used as diagnostic indicators. Receiving the information on deviations from the nominal readings of the controlled parameters (temperature, vibration, noise, etc.) during movement, the driver of a high-speed passenger train informs the dispatcher, who, in turn, informs the relevant units [1][2][3][4][5].
Traditional technologies for the analysis of noisy signals used in control systems do not allow extracting sufficient diagnostic information to identify the beginning of the latent period of the development of defects in the core components of the running gear of the rolling stock. This affects the reliability of the control results, which sometimes leads to errors that inevitably cause accidents with undesirable consequences. Therefore, to enhance the reliability of a fail-safe operation by early detection of the onset of faults and organization of timely maintenance of the rolling stock, it is essential to create new, more effective technologies for analyzing noisy signals. As follows from [6,7], the development of effective technologies for calculating noise parameters is relevant for the quality operation of modern real systems, which takes into account the signal to noise ratio and the effect of the noise on the operation of the system.
PROBLEM STATEMENT
It is known that to ensure the safety of trains, it is necessary to ensure diagnostics of such faults of the running gear of the rolling stock as bearing defects, lacking and insufficient lubrication, malfunctions of wheel-and-motor units, mounting defects, imbalance of rotating parts, gearbox defects, leakage of the feed and brake lines, break valve malfunctions, brake cylinder malfunctions, compressor malfunctions, etc.
As an example, let us consider one of the critical components of the running gear of the rolling stock -axle boxes of wheelsets, which mainly consist of roller bearings. Currently, automatic temperature and noise control systems are used to monitor the technical condition of axle box bearings. Such systems can be either stationary (for any type of rolling stock) or have built-in sensors. These systems make it possible to monitor the parameters of the technical condition of axle boxes and generate information on deviations from nominal readings. This allows taking timely measures, thereby preventing accidents. For instance, for passenger cars, the system can obtain information about the temperature of the axle box using sensors built into the casing of the axle box of wheelset bearings. The control system ensures the processing and storage of the obtained information and signals about dangerous heating of bearings. However, this system only monitors the temperature of the unit and notifies the train staff of its significance, which does not always allow one to control the beginning and dynamics of the development of bearing damage. As a result, it becomes difficult to eliminate the fault.
Our analysis shows that the diagnostics of the technical condition in other components of the running gear of the rolling stock also has similar shortcomings. Therefore, it is necessary to control the beginning of the latent period of a fault by early detection of the initiation of defects preceding it in the initial stage, when negative effects on the reliability or operability of the rolling stock do not yet manifest themselves. Therefore, we need to develop more effective alternative options for solving the problem of controlling the beginning of initiation and development dynamics of faults in the running gear of the rolling stock.
POSSIBILITIES OF USING THE THEORY OF FUZZY SETS FOR FAULT DIAGNOSTICS
Studies have shown that to solve the problems of fault diagnostics in case of axle boxes and other units of rolling stock, fuzzy sets can be used, as they take into account such difficult-to-formalize factors as the experience and intuition of a highly qualified expert specialist. For instance, the apparatus of the theory of fuzzy sets for the diagnosis of axle boxes of the rolling stock (ARS) in a fuzzy expert system (FES) allows one to arrive at operational conclusions about the technical diagnosis of faults by abandoning the traditional requirements for the accuracy of its functional description.
To enter knowledge into the knowledge base (KB) of a diagnostic FES, a knowledge representation language is used that takes into account the specific features of the object. A set of heuristics used by highly qualified specialists serves as the algorithm for solving the problem. The knowledge formulated by experts is entered into the system knowledge base [8,9] An analysis of the possibility of diagnosing faults of the running gear of the rolling stock with the use of this technology demonstrates that the use of these systems makes it possible to detect the initiation of a defect with sufficient reliability in its pronounced stage. Unfortunately, sometimes, it can be delayed, which in some cases can cause accidents with catastrophic consequences.
In view of the above, to enhance the validity and reliability of control results, it is advisable to use in diagnostic systems a technology for monitoring the beginning of the transition of the main running gear of the rolling stock to the latent period of an emergency state using [1,10].
Noise control of the beginning and development dynamics… 85.
POSSIBILITIES OF NOISE MONITORING OF THE ONSET OF DEFECTS PRECEDING RUNNING GEAR FAULTS
It is known that the running gear of a rolling stock breaks down due to the initiation of various defects, such as wear, crack, fatigue deformation, etc. [1][2][3][4][5]8]. In some cases, they lead to disastrous consequences. As was stated earlier, to prevent this, it is necessary to control the initiation of defects preceding such accidents. The solution to the problem of controlling the onset of the initiation of a defect that leads to a violation of the integrity and operability of the structure requires, first, creating appropriate technologies and software for analyzing the signals received at the outputs of the corresponding sensors. Here, it is important to obtain the necessary information to control the onset of all kinds of defects. To this end, based on the statistics of the most dangerous accidents that have occurred, it is necessary to determine the type of sensors, the location of their installation ("vulnerable spots"), which ensures obtaining sufficient information from the object, and making early detection of defect initiation reliable enough [1].
The sensors that receive signals reflecting the beginning of the initiation of the most common defects have the largest information capacity. Such technological parameters as temperature, pressure, vibration, acoustic and thermal radiation, etc. contain sufficient data to control the initiation of the corresponding defects. For instance, the types of data required for controlling the condition of most running gear of the rolling stock include the vibration spectrum of rolling stock elements, the spectrum of acoustic vibrations and other parameters that characterize the functioning of the system. Moreover, at the beginning of fault initiation, not only are the values of these parameters important but also the dynamics of changes in their noises at a given time. For instance, axle boxes of wheelsets are typical systems characterized by the variation of the vibration parameters of both the useful signal and the noise due to changes in the technical condition during operation (1) In the control of the beginning of the latent period of faults, it is natural that, in addition to the vibration signals, it is also possible to analyze acoustic and other signals received at the outputs of appropriate sensors [11][12][13][14][15][16][17].
Studies and analysis of failures of running gear of the rolling stock demonstrate that the onset of faults and the dynamics of their development are accompanied by the appearance of the noise correlated with the useful signal. The noise forms from the noise caused by the influence of external factors and by the noises that emerge as a result of the initiation of the corresponding defects. As a result, the sum noise (2) forms, which, in the latent period of accidents, correlates with the useful signal.
Therefore, when solving the problem of controlling the beginning and development dynamics of this process, it is advisable to use the estimates of the static characteristics of the sum noise as informative attributes.
In the following paragraphs, we consider one of the possible versions of the noise control of the onset of a fault in the running gear of the rolling stock. In this version, vibration sensors are installed, for instance, on the axle boxes of the wheelsets. When the rolling stock moves, vibration signals from these sensors are transmitted wirelessly to the input of the controller of that particular car. Any change in the condition (deviation from the norm) of the controlled components is reflected in the signals of the vibration sensors, which are analyzed on the noise analysis controllers. For instance, they are analyzed based on the technology of relay correlation noise analysis, using the expressions is the estimate of the relay cross-correlation function between the useful signal and the noise ; is the variance of the noise ; and is the noisy vibration signal.
Due to this, in the beginning of the initiation of faults, the estimates and Dε of the noise characteristics of the vibration signal differ from the normal (reference) estimates, which makes it possible to register the beginning of the latent period of changes in the technical condition of the corresponding component. Similarly, using the appropriate formulas for other noise characteristics of the vibration signal given in the following paragraphs, it is possible to control the technical condition of all controlled nodes of the train's running gear. To this end, during the system operation, as a result of certain amount of training, the maximum threshold values of the reference estimates of all noise characteristics are determined, at which the technical condition of the controlled component is considered normal. A reference set of informative attributes is formed from them in the form [1] .
(5) In a similar manner, the reference set of other noise characteristics is formed. Due to this, during the train movement, as a result of application of all kinds of noise technologies for analyzing vibration signals, a set of informative attributes are formed on the noise controller, reflecting the current technical condition of all the controlled components of the train's running gear along the entire route of its movement. In case of a defect, e.g. in the axle boxes of wheelsets at the current moment of the train car's movement, some current estimates of the noise characteristics will be greater than the corresponding reference threshold value formed from expression (5) of reference informative attributes. In other words, if there is a fault in the current state of the controlled component, the current estimates of some noise characteristics will be greater than the corresponding maximum reference estimates. Due to this, the information can be compiled for wireless transmission from the noise controllers. Therefore, the information that will reflect the technical condition of the corresponding components of the running gear of the rolling stock can be registered and displayed on the driver's monitor screen.
Thus, along the entire route, the version under discussion will provide noise control of both the technical condition of all the controlled components of individual cars and the rolling stock as a whole. It is clear that duplication of the correlation noise analysis of vibration signals using other noise analysis technologies can enhance the reliability and validity of the results of the proposed system. Some algorithms of spectral analysis of the noise of vibration signals that are also advisable for use in the noise control of the running gear of the rolling stock are given in the following paragraphs.
SPECTRAL TECHNOLOGY FOR THE NOISE CONTROL OF THE BEGINNING OF FAULTS IN THE RUNNING GEAR OF THE ROLLING STOCK
As mentioned earlier, the initiation of faults and the dynamics of their development are accompanied by the emergence of the noise correlated with the useful signal . The noise is added to the noise , forming the sum noise that, in the latent period of accidents, correlates with the useful signal.
Therefore, when solving the problem of controlling the beginning and dynamics development of faults, it is advisable to also use estimates of the spectral characteristics of the sum noise as informative attributes. An analysis of possible solutions to this problem showed [1,10] that for this purpose, it is advisable to replace non-measurable samples of the noise with their approximate which can also be represented as .
(7) Due to this, assuming the notation the formula for calculating the equivalent values of the samples of the noise can be represented as .
(10) Here, assuming that the expression (11) holds true, the formula for calculating the mean value of samples of the noise can be reduced to calculating the mean value of equivalent samples of the noise , i.e.
. (12) Due to this, the expression for calculating the estimates of the spectral characteristics of the noise can be represented in the following form: (13) .
(14) It is easy to see that, taking into account notations (8) and (9), expressions (13) and (14), i.e. formulas for calculating the estimates of the spectral characteristics of the noise can be represented in the following form: (16) Thus, the use of algorithms (15) and (16) opens the possibility for registering the beginning of the latent period of faults, since the estimates and will differ from the reference informative attributes only at the beginning of an emergency state. Because of this, the use of these expressions will make it possible to enhance the reliability of the control of the onset of the latent period of initiation of faults in the running gear of the rolling stock. Studies have shown that the dynamics of development of running gear faults affects the degree of correlation between the samples of the noise , as well as the correlation between samples of the equivalent noise . In this case, the formula for forming the equivalent noise at can be written as .
(17) At , this expression will take the following form: .
(18) At , the expression can be written in a generalized form: .
(19) Due to this, based on the results of a spectral analysis of the equivalent of the noise at , i.e.
, it is possible to control the dynamics of an accident using the following expressions: . (20) If the fault is stable, then these estimates will be equal. However, in the presence of fault development dynamics, the estimates , ; , ; … ; , will differ from each other, and in the case of high dynamics of the development of the defect degree, these differences will be significant.
SPECTRAL TECHNOLOGY FOR NOISE SIGNALING OF THE BEGINNING OF THE LATENT PERIOD OF FAULTS
Our analysis of the spectral noise control technologies has demonstrated that during the operation of the rolling stock, the signaling of the beginning of the latent period of faults is also important. For this purpose, in addition to the aforementioned estimates, it is also advisable to use the estimates and and the relay spectral characteristics of the noise of the noisy signals , which can be calculated from expressions [1] (21) .
(22) These studies also demonstrated that the technology of sign spectral noise analysis can also be used for the signaling of the beginning of the onset of faults by means of the following expressions: .
(24) The advisability of applying the technologies of relay and sign spectral analysis for the signaling of the beginning of the latent period of accidents is due to the fact that its hardware implementation is easy. At the same time, it is advisable to ensure the reliability of the signaling by duplicating these technologies with the technology of relay correlation analysis calculated from the formula: is the estimate of the relay cross-correlation function between the useful signal and the noise .
If the obtained estimates are greater than the reference ones at , the dynamics of the development of faults can be considered slow. If the estimates are greater than the reference ones at , then, the dynamics is moderate. In case of a difference from the reference informative attributes at and above, the development dynamics can be considered accelerated.
CONCLUSION
Traditional technologies do not allow extracting sufficient diagnostic information to identify the beginning of the latent period of the initiation of defects in the core components of the running gear of the rolling stock. This affects the time of registration of the onset of faults, which sometimes leads to inevitable accidents with undesirable consequences. Therefore, to enhance the reliability of fail-safe operation and timely maintenance of the rolling stock, it is necessary to create new, more effective technologies for analyzing noisy signals that allow early detection of the onset of faults.
Our analysis of running gear of the rolling stock has demonstrated that during the initiation of corresponding defects, the noisy signals at the outputs of the sensors carry the information about it in the form of the noise of a random function. This is because during the onset of an accident due to the imposition of a large number of various dynamic effects in the controlled components, noises appear.
T. Aliev, T. Babayev, T. Alizada, N. Rzayeva Therefore, noise components of noisy signals, being of a chaotic random nature, contain enough information about the beginning of changes in the technical condition of an object. For instance, vibration signals of axle boxes of wheelsets contain a large number of different noises. They make it difficult to detect the onset of a defect when traditional signal analysis technologies are used. At the same time, in some cases, noises are the carriers of diagnostic information about the onset of a fault. Therefore, to control the beginning of the initiation of faults, it is necessary to create technologies that allow calculating informative attributes by using not only useful signals but also noise [1,10]. Here, to ensure the control of a defect at the beginning of its initiation, the first and foremost task is to choose the type and place of installation of the appropriate sensors that ensure the object's controllability. To analyze both the sum signal and the noise , it is advisable to employ the technologies that allow calculating the appropriate informative attributes.
Due to the extreme importance of ensuring a fail-safe operation of the rolling stock, it is advisable to control the beginning and development dynamics of faults by duplicating several noise control and noise signaling technologies proposed in [1]. In this case, the reference set of the estimates of the noise characteristics of noisy signals will take the form , (26) which, combined with current informative attributes, will constitute the basis of the dataware for the solution of the control problem. As a result, the reliability and validity of the results of the control of the beginning and development dynamics of faults will increase.
In conclusion, it should be noted that, despite the influence of various factors that make it difficult to ensure a fail-safe operation of the running gear of the rolling stock, currently used technologies and systems provide satisfactory control of their functioning. However, due to the extreme importance of this issue, to control the current state of the running gear of the rolling stock, it is advisable to duplicate traditional control algorithms with the proposed algorithms for the noise control of the onset and development dynamics of faults. This will ensure early diagnostics of such faults of the running gear of the rolling stock as bearing defects, lacking and insufficient lubrication, malfunctions of wheel-and-motor units, mounting defects, imbalance of rotating parts, gear defects, leakage in the feed and brake lines, break valve malfunctions, brake cylinder malfunctions, compressor malfunctions, etc. Thus, the use of algorithms and technology of noise control in combination with traditional algorithms and technologies can significantly enhance the effectiveness and reliability of ensuring a fail-safe operation of the rolling stock. | 4,963 | 2020-01-01T00:00:00.000 | [
"Engineering"
] |
Effects of musical ear training on lexical tone perception
The effect of short term musical experience on lexical tone perception was examined by administering four hours of daily musical ear training to non-tone language speakers. After training, participants showed some improvement in a tone labeling task, but not a tone discrimination task; however, this improvement did not differ reliably from controls indicating that short-term musical training is thus far not able to replicate language effects observed among lifelong musicians, but some linguistic differences between musicians and nonmusicians may likely be due to experience, rather than individual differences or other factors.
Introduction.
Musicians perceive pitch differently from others, which is reflected in their perception, production, and learning of lexical tones (Bradley, 2013).Musicians also possess more robust phonetic pitch representations, which are prerequisite for acquiring tonal categories and words (Wong & Perrachione, 2007).Musicianship has therefore been argued to change linguistic pitch perception, but cross-sectional evidence remains circumstantial, because musiclanguage studies typically compare musicians with many years of training to those with little to no formal training.It remains poorly established how what music skill level or degree of training is required to see differences in linguistic pitch perception, but differences likely arise from a combination of explicit training, implicit learning, and individual aptitudes and personalities (Corrigall, Schellenberg, & Misura, 2013).Neural differences associated with musical experience suggest that musical training could be harnessed for non-musical gains (Kraus & Chandrasekaran, 2010;Patel, 2011).
To determine to what degree explicit perceptual musical training shapes linguistic perception (and whether such shaping can still occur in adulthood), we administered computerized musical ear training to non-tone language speakers who were not musicians in order to determine: 1. Whether music-induced changes in lexical tone perception experience be observed among adult non-tone language speakers.2. What musical skill level or degree of training is sufficient to produce changes in lexical tone perception.
Because musicians use ear training to increase sensitivity to musical pitch, if nonmusicians improve their pitch perception, we hypothesized that they will also benefit from the same linguistic "side effects," namely: 1. Improved perception of unfamiliar lexical tones; 2. Improvement primarily in phonetic (vs.phonological) tasks.
Procedure.
Participants were 32 non-tone language speakers (L1 English = 24) with musical backgrounds ranging from no musical participation to moderate amateur participation-however, none were professional musicians, and none had studied music theory or aural skills (m=3.9 years instrument study).
Participants completed 4 hours of computerized aural skills (ear) training in daily one-hour sessions over one week, as follows.2. Tone Comparison Task: Participants decided whether two words are the "same" or "different."Both words in a trial had the same syllable, and their tone differed on twothirds of trials ("same" trials had two different tokens of the same word).The test contained nine blocks of 20 trials (180 total); six blocks had a single speaker (lowvariability) and three had two speakers per trial (high-variability).
Results & Discussion
. Results were analyzed using separate analyses of variance for the two tone tests, with accuracy as the dependent variable, and session and group as independent variables.Neither group improved their performance in Tone Comparison.For Tone Labeling (see Figure 2), there was a marginally significant main effect of session, F(1, 30) = 3.70, p = 0.059; both the Interval and Rhythm groups improved slightly from pre-test to post-test (7.1% vs. 6.1%, respectively), although the size of this improvement did not differ significantly between the groups.Unexpectedly, there was a main effect of group, with the Interval Training group performing better than the Rhythmic Training group at both pre-test and post-test, F(1, 30) = 6.08, p = 0.017; this was surprising, given that participants were randomly assigned to training conditions, and could possibly confound the results by flattening any expected greater gains by the interval group if they were already outperforming the control group.Post-hoc, Bonferroni-adjusted t-tests were used to compare performance in low-and highvariability blocks of the Tone Labeling Task (see Figure 3).In low-variability (1-speaker) blocks, the Interval group improved in Labeling more than the Rhythm group (8.0%vs 5.5%), but this was not significant (t(29) = −0.57,p = 0.29).Labeling improvement in the high-variability (2speaker) blocks was similar for the Interval and Rhythm groups (6.1% vs. 6.7%). .This would be consistent with progression along a phonetic-phonological continuity (Wong & Perrachione, 2007), if training affects phonetic (pitch shape, in its low-variability, more phonetic condition) but not phonological (category decision) tasks.Further studies with increased training with additional contexts (e.g., pitch production) are necessary.
Day 1: Lexical Tone Pre-tests + 30 minutes Ear Training Days 2-4: 1 hour Ear Training Day 5: 30-minute Ear Training + Lexical Tone Post-tests 2.1.EAR-TRAINING.Ear training is perceptual training to identify elements of music (e.g., scales, rhythms) through active listening and feedback.Training was administered through software (EarMaster) which controls training parameters and tracks progress.Participants were randomly assigned to two training conditions: 1. Interval Training (n=16), in which participants heard synthesized piano intervals (melodic note pairs defined by the distance/ratio between their pitches).During training, participants were asked to compare (say which pitch distance was larger) and name (select the appropriate label) intervals.The difficulty increased as training progressed through the addition of more similar interval types.Interval training was expected to boost lexical tone performance due to its basis in pitch.2. Rhythmic Training (n=16), in which participants repeated computer generated drum rhythms by tapping on the space bar.The difficulty increased as training progressed through the addition of longer and more complex rhythms.Unlike the interval training, this training involved production, as well as perception, but it was the only rhythmic task available in EarMaster which does not require the participant to read musical notation.Rhythmic training was not expected to affect lexical tone perception, because it does not involve pitch.
Figure 1.Intervals used during ear training.2.2.LEXICAL TONE TESTS.Before and after training, tone perception was assessed through two tests based on Mandarin lexical tones.Stimuli both tasks consisted of the syllables [ma], [ku], and [di] spoken with each of the four Mandarin lexical tones by two native speakers (1 female, 1 male) in citation form.Presented with PsychoPy software, and responses were collected via keyboard.1. Tone Labeling Task: participants matched a Mandarin word with a visual symbol depicting its pitch pattern (→ ↘ ↗ U).The test contained eight blocks of 24 trials (192 total); four blocks had a single speaker (low-variability) and four had two speakers (highvariability).
Figure 2 .
Figure 2. Changes in Mandarin Tone Identification after ear training.
Figure 3 .
Figure 3. Changes in high-and low-variability Mandarin Tone Identification after ear training.4.Conclusions.Four hours of melodic interval ear training does not seem sufficient to improve tone perception relative to rhythmic training.There is some evidence that those achieving more in interval training improved more in tone, but additional data are needed.If interval training affects tone perception, it affects Tone Labeling more than it does Tone Comparison (category matching).This would be consistent with progression along a phonetic-phonological continuity(Wong & Perrachione, 2007), if training affects phonetic (pitch shape, in its low-variability, more phonetic condition) but not phonological (category decision) tasks.Further studies with increased training with additional contexts (e.g., pitch production) are necessary. | 1,642.2 | 2016-06-12T00:00:00.000 | [
"Linguistics",
"Psychology"
] |
Dynamics of Mandelbrot Set with Transcendental Function
These days Mandelbrot set with transcendental function is an interesting area for mathematicians. New equations have been created for Mandelbrot set using trigonometric, logarithmic and exponential functions. Earlier, Ishikawa iteration has been applied to these equations and generate new fractals named as Relative Superior Mandelbrot Set with transcendental function. In this paper, the Mann iteration is being applied on Mandelbrot set with sine function i.e. sin(zn)+c and new fractals with the concept of Superior Transcendental Mandelbrot Set will be shown. Our goal is to focus on the less number of iterations which are required to obtain fixed point of function sin(zn)+c.
I. INTRODUCTION
Complex dynamics have a new significant development with the explosion of popular interest in the beautiful fractal objects that form the subject matter of the theory [5,6].The computer-generated images of Julia and Mandelbrot sets bombard mathematicians to investigate the nature of both the Fatou and Julia sets of a given complex function [4].
Though a moment's reflection confirms the origin is a critical point for the cosine and for the sine, the transcendental Julia sets fill major regions of the complex plane [3].
Based on this concept, authors found the region M of nonescape corresponds to a principal central bulb set with a fractal series of black hearts, including a series lining the x-axis, and that any c value in M corresponds precisely to Julia set kernels of the corresponding quadratic type [7].
Ereneko [1] studied that for every transcendental functions, the set of escaping points is always non-empty.The set of parameter values of c for which the Julia set of Q is connected forms the well-known Mandelbrot set.
Mandelbrot set serves as a lexicon for the Julia Set.The location of the parameter c with in the Mandelbrot set furnishes information on properties of the corresponding Julia set.There are similarities between magnified positions of the Mandelbrot set and the corresponding filled Julia set holds only near certain c values such as the central junctions of the antenna.These are c values, for which 0 is eventually periodic; such c values are called Misiurcwicz points [10].In 2004, Rani jointly with Kumar applied Mann iteration to functions and introduces superior iteration in non linear sciences and gave new escape criterions for complex polynomials.Thus authors computed superior Julia set [8] and superior Mandelbrot set [9] for complex polynomials.
Fixed point can be obtained by repeated function iteration or Picard iteration.There are some other iteration processes like Ishikawa iteration and Mann iteration, which is required to obtain weak or strong convergence to a fixed point in case of non-expansive maps, pseudocontrative maps etc [6].The purpose of this paper is to generate superior Julia set and Mandelbrot set using transcendental sine function.We are generating fractals for sin(z n )+c using Mann iteration and calculating the fixed points for the same.
II. DEFINITIONS & PRELIMINARIES
The generation of fractal for sin(z n )+c is much similar to standard quadratic equation of Mandelbrot set but it consists of repeated iterations upto n times with respect to sine function.Following are some basic definitions require for subsequent analysis.
A. Definition
(Superior iterates) Let X be a non-empty set of real numbers and f : X->X.For x 0 belongs to X, construct a sequence {x n } in the following manner [11,12]: The sequence {x n } constructed this way is called a superior sequence of iterates, denoted by
B. Definition
(Superior Orbit) The sequences x n constructed above is called Mann sequence of iteration or superior sequences of iterates.We denote it by SO(x 0 , s, t) [11,12].This procedure was essentially given by Mann, was the first to study it for β n in 1955.www.ijacsa.thesai.orgSince the results obtained in fractal modelling via Mann iterates are the super set of their corresponding fractal models in the Picard orbit.Researchers have since developed superior fractal models for β n = β, n = 1, 2,..., for various values of β.
C. Definition
(Superior Julia sets) The set of complex points SK whose orbits are bounded under superior iteration of a function Q is called the filled superior Julia set.A superior Julia set SJ of Q is the boundary of the filled superior Julia set SK [8].
D. Definition
(Superior Mandelbrot sets) A Superior Mandelbrot set SM for a function of the form Q c (z) = z n + c, n = 1, 2,..., is defined as the collection of c C for which the superior orbit of the point 0 is bounded,
E. Definition
Suppose x 0 is a fixed point for F. Then x 0 is an attracting fixed point if |F'(x 0 )|<1.The point x 0 is a repelling fixed point if |F'(x 0 )|>1.Finally if |F'(x 0 )|=1, the fixed point is neutral [10].
III. GENERATING THE FRACTAL
The Mandelbrot set is the collection of C points for which the orbit is bounded.The set of those points are known as prisoner set and remaining points comes under escape set.The escape criteria for the function sin (z n ) +c is given as follows: A. Escape Criteria for Quadratic Function: For n = 2, the escape criteria is depends on a constant value or (z>=1/ ω).
C. Escape Criteria for General Polynomial:
The escape criteria for the general polynomial equations using Mann iteration procedure for n is Note that the initial value z 0 should be infinity, since infinity is the critical point of z for sin (z n )+c .However instead of starting with z 0 = ∞, it is simpler to start with z 1 = c, which yields the same result.A critical point of z →f(z)+c is a point where f'(z)=0.The point z in Mandelbrot set for sine function has an orbit that satisfies imag(z) > 50, then the orbit of z escapes [2].
IV. GEOMETRY OF SUPERIOR TRANSCENDENTAL MANDELBROT SETS AND SUPERIOR TRANSCENDENTAL JULIA SETS:
The fractal generated by this iteration process possesses symmetry about x axis in case of all polynomials.
A. Description of Superior Transcendental Mandelbrot Set:
The fractal is symmetrical about x axis for all values of s.Initially for a quadratic polynomial the value of s =1 has been taken and got the fractal with two bulbs.As the value of s has been changed from 1 to 0.3, both bulbs merged together and resultant fractal is not very much sharp.There are very few tiny bulbs attached to primary bulbs for s=0.5.Subsequently we change the value of s to 0.7; the primary bulb is showing too much tiny bulbs with attached decorations.
For a cubic polynomial, fractal shows its beautiful images.Starts with the value of s=0.1, and move towards 0.3, we got an image of lord Vishnu according to Hindu mythology.With the s= 0.5, there is an image of sparkling earthen lamp (diya) with its own image has been shown.There is a symmetrical image about x axis for s=0.7.
For a biquadrate polynomial, the fractal is having three primary bulbs and large number of secondary bulbs attached to it.For s=0.7, shape of primary bulbs are approximately same as with the value of s=1 but the number and shapes of secondary bulbs reduced.In case of s=0.3 and 0.5, the shapes of primary bulbs are same but the axis of secondary bulb varies based on above mentioned values of s.
B. Description of Superior Transcendental Julia Set:
Transcendental function sin(z) with superior Julia set, which follows the law of having 2n wings has analyzed, where n is the power of z.The images for all polynomials possesses symmetry about both x and y axis.Here extremely beautiful images of superior Julia set have been generated for different power of polynomial.
An image of superior Julia set has a resemblance with hippocampus for a quadratic polynomial.In case of cubic polynomial, a star fish shape has formed with large central body having rotational and reflection symmetry along with axes symmetry.Finally a biquadrate polynomial is having same structure as cubic polynomial with eight wings.
C. Generation of Superior Transcendental Mandelbrot Sets
For Quadratic function: For Cubic function: For Biquadratic function: For Quadratic Function:
E. Fixed Points
Fixed points of Quadratic function: V. CONCLUSION In this paper, we have analyzed sine function in Mandelbrot equation with Mann iteration.Superior Julia set possess 2n wings with central black region.Our study shown the striking properties and escape criteria for transcendental function and generated the corresponding fractals using Mann iterates in which most of the images are having symmetry along x axis and y axis.The images revealed their own identity.As on a particular value of constant s, an image resembled to Lord Vishnu icon according to Hindu Mythology see Fig. [6].Another image shows earthen lamp (diya) with its own reflection along with real axis see Fig. [7].We obtained fixed point for quadratic function after 7 iterations, for cubic function after 3 iterations and for biquadratic function after 9 iterations.
The surrounding region of superior Mandelbrot set shown to be invariant cantor set in the form of curve or hair that tends to infinity under iteration in all figures.
TABLE III .
Orbit | 2,108.4 | 2012-01-01T00:00:00.000 | [
"Mathematics"
] |
SN1987A neutrino burst: limits on flavor conversion
In this paper, we revisit the SN1987A neutrino data to see its constraints on flavor conversion. We are motivated by the fact that most works that analyze this data consider a specific conversion mechanism, such as the MSW (Mikheyev-Smirnov-Wolfenstein) effect, although flavor conversion is still an open question in supernovae due to the presence of neutrino-neutrino interactions. In our analysis, instead of considering a specific conversion mechanism, we let the electron antineutrino survival probability $P_{\overline{e}\overline{e}}$ be a free parameter. We fit the data from Kamiokande-II, Baksan, and IMB detected spectrum with two classes of models: time-integrated and time-dependent. For the time-integrated model, it is not possible to put limits above $1\sigma$ (68% confidence level) on the survival probability. The same happens for the time-dependent model when cooling is the only mechanism of antineutrino emission. However, for models considering an accretion phase, $P_{\overline{e}\overline{e}}\sim0$ is strongly rejected, showing a preference for the existence of an accretion component in the detected antineutrino flux, and a preference for normal mass ordering when only the MSW is present.
One of the main questions regarding supernova neutrinos today is the flavor conversion mechanism.It is expected for the supernova neutrinos to suffer MSW conversion [12][13][14] and a substantial number of works were done considering this as the only conversion mechanism in action, including the ones that analyze the SN1987A data [6,7].However, today it is expected that neutrino-neutrino interactions (forward scattering) become relevant in a supernova environment leading the neutrinos to a non-linear collective evolution [15].Due to the complications that emerge from this type of evolution, there is not a conclusive picture of neutrino conversion in the supernova environment.
Nevertheless, given the equal amount of non-electron antineutrinos ν x = (ν µ , ν τ ) emitted from the supernova, it is possible to write the flavor conversion in terms of only the electron antineutrino survival probability P ee .Therefore, we treat this probability as a free parameter to see how SN1987A data can constrain it.Something similar was done by F. Vissani in [16].However, it seems that the influence of the survival probability is analyzed only for the MSW normal hierarchy scenario (P ee = 0.64) against the no oscillation one (P ee = 0).Here we take a more complete analysis for P ee , allowing it to range from 0 to 1.
In section 2 we describe our model for the detected event rate in each detector (KII,IMB, Baksan) based on two different neutrino emission models, the flavor conversion mechanism, and the detection properties.In section 3 we describe our statistical analysis of the SN1987A data.In section 4 we show our results and discuss them, and finally, in section 5 we present our conclusions.
Model for the neutrino signal
In this section, we describe the model for the expected neutrino event rate in each of the detectors, which is used to fit the SN1987A data.First, we describe the two neutrino emission models considered in this paper: a time-dependent and a time-integrated.In sequence, we describe the flavor conversion in the flux, which depends only on P ee , and, in the end, we discuss the detection features of this analysis.Given that the most relevant cross-section for the considered detectors is the IBD, we will restrict our model to the antineutrino sector ( νe , νµ , ντ )
Time-dependent Given that the neutrino emission evolves in time, a time-dependent model should be at least considered in data analysis.This approach can be found in the famous paper of Lamb and Loredo [6] and some other works [7].In this approach, the antineutrino emission can be divided into two phases: the accretion and cooling phases.Here we will follow the path of [6,7] and model each phase by its most relevant mechanism of emission.
In this case, the accretion phase can be modeled as a positron thermal flux with temperature T a incident in a neutron target, that composes the mass in accretion in the protoneutron star.Therefore, as in [6,7], we consider that only electron antineutrinos are emitted in this phase and the flux is given by: with , where N n (t) is the number of neutrons as a function of the time, σ e + n (E ν ) the positron-neutron cross-section, and g e + (E e+ , T a ) the thermal distribution of positrons with energy E e+ in a temperature T a .The number of neutrons is given by the initial accreting mass M a with a fraction of neutrons Y n , and its time behavior is given by the factor j k (t) = exp − (t/τ a ) k , with τ a being the characteristic time of the accretion phase and the parameter k = 2 following the parametrization in [7] 1 .The denominator 1 + t/0.5s, as in [6,7], is used to mimic the behavior from supernova simulations, where we have a constant flux within the first 0.5 s followed by a fast decrease.The cooling phase, which is dominated by neutrinos and antineutrinos of all flavors emitted by the cooling neutron star, is modeled by a thermal distribution of fermions with temperature T c (t), with characteristic time τ c , emitted from a sphere with fixed radius R c and is given by with the cooling temperature being a function of time As already pointed out, different from the accretion component, the cooling one is composed of antineutrinos of all flavors.However, the non-electron antineutrinos ν x are emitted from deeper regions in the supernova, which can be effectively implemented by considering that they are emitted with higher initial temperatures T c, νx .In fact, during the rest of the paper, we will talk about the ratio between the flavors temperatures τ = T νx /T νe .
To combine the fluxes of both phases of emission, we follow [7] where the cooling phase starts after the accretion one.As argued in the cited work, if the accretion and cooling phases were contemporaneous the first seconds would be composed of two different spectra, given the different temperatures of each of these phases.As numerical simulations of supernovae do not show this feature, we assume that the different emission phases are separated in time.We do this using the following parameterization: where the accretion flux is only composed of electrons antineutrinos φ 0 a, νe , while the cooling flux contains an electronic φ 0 c, νe and non-electronic component φ 0 c, νx .
Time-integrated
In this model, we consider that the timeintegrated flux can be described by the following pinched spectrum [17]: where, for a specific neutrino flavor β , L β is the total energy (time-integrated luminosity), E 0β the mean energy, and α β the pinching parameter.We are mainly motivated to use this model due to a collection of works that only use the energy information from the SN1987A [8][9][10].Although the time data could bring new information, it is interesting to check if the energy alone can say something about the flavor conversion.
Flavor Conversion
From emission until detection, the neutrino may suffer flavor conversion.It is still an open question for supernova neutrinos which is the complete mechanism of flavor conversion, given the complications that arise with neutrinoneutrino interactions.However, due to unitarity and the equal initial flux of non-electron antineutrinos φ 0 ν µ = φ 0 ν τ = φ 0 ν x , the equations for flavor conversion can be simplified so that it will only depend on the electron antineutrino survival probability P ee and initial fluxes [18], such that Therefore, we can explore the survival probability P ee as a free parameter representing the flavor conversion occurring during the neutrino propagation.In this paper, we want to see how strong the SN1987A data can constrain P ee in the fitted models, given that the flavor conversion mechanism is still an open question in a supernova environment.Although this probability may be time and/or energy-dependent, we will consider it independent of these variables, given that we do not want to use a specific model.
We will also consider the MSW-only conversion scenario in order to compare it to our free P ee model.In this scenario, the electron antineutrino is created as a ν1 for normal mass hierarchy (NH) and ν3 for inverted mass hierarchy (IH).Therefore, the survival probability for each mass ordering can be written as follows [19] where we have considered an adiabatic evolution, with a flipping probability equal to zero at the high and low-density resonances.The vacuum mixing parameters are taken from the update values published for the global fit analysis in [20].Although this energy dependence of P ee is negligible in the standard MSW effect, other possible effects associated with collective effects, such as spectral split among different neutrino flavors lead to a strong energy dependency, changing drastically this scenario [15].However, given the unknowns associated with such collective effects nowadays, we limit our analysis to consider a P ee that is uniform in energy, leaving the spectral split analysis for a future work.
Detection
In the case of the SN1987A, we have data from three detectors: Kamiokande-II, IMB, and Baksan.In all of them, the dominant channel for electron antineutrino detection is the Inverse Beta-decay (IBD), which is the only one that we will consider.Therefore, the event rate R IBD νe as a function of the positron measured energy E e + , the angle between the incoming neutrino and the scattered positron θ and time (for the time-dependent model) can be calculated as follows where N p is the number of free protons, φ νe (E ν ,t) the electron antineutrino flux at the detector, dσ IBD νe (E ν )/d cos θ the differential cross-section for IBD, and η d (E e + ) the detector intrinsic efficiency.For the IBD, the incoming neutrino energy E ν is related to the created positron energy by E e + ≈ E ν − 1.293MeV , due to the mass difference between the initial proton and the final neutron.The energy threshold for the IBD is E th ν = 1.806MeV [21].
Efficiency
As pointed out by [16], when calculating the differential event rate in equation 9, one should use the detector intrinsic efficiency η d (E e + ).However, when integrating the event rate to get the total number of detected events, one should account for the threshold energy considered when selecting the events.This is achieved by multiplying the intrinsic efficiency by a function g(E e + , E min ) resulting in a total efficiency g(E e + , E min ) = in which the error function Erf accounts for the threshold energy E min and the uncertainty σ (E e + ) on the energy.This distinction between intrinsic and total efficiency is relevant when talking about the ones reported by the experiments, which are total efficiencies accounting for the threshold energies used during the events selections.This distinction becomes even more relevant in the case of the Kamiokande-II when using the low-energy events (numbers 13-16 nad 6 in table 5) added a posteriori and which are below the energy threshold of 7.5 MeV used in the first published data.To incorporate these events in our analysis, we need to infer the intrinsic efficiency from the published total efficiency and extrapolate the last to lower energies in the case of Kamiokande-II.Following this reasoning, we adopt the same parametrization for the intrinsic efficiency as reported in [16], with E min = 4.5 MeV for Kamiokande-II.Both total and intrinsic efficiencies used in this work are shown in figure 10.
Uncertainties
The uncertainties used in this work are experimental ones shown in tables 5, 6, and 7.Although we have the angle uncertainty, we will not consider it in our analysis, due to its non-significant impact on the likelihood, given that the considered cross-section (IBD) has a weak angular dependency.Also, as pointed out in [7], the relative time between the events is measured with good precision so that we also ignore the time uncertainty.As for the energy uncertainty, in addition to reported values for the energy of the events, to implement it in the efficiency expressions, such as equation in 10b, we need to estimate the uncertainty for other values of energy.For this purpose, we adopt an uncertainty parametrization with a statistical component that goes with the square root of the measured energy E e + and a systematic one that grows linear with the energy.as done in [16]: The values that we used for the coefficients are shown in table 4 corresponding to the ones that best adjust the function to the reported uncertainties.
Cross-section
The exclusive interaction considered in the analysis was the inverse beta decay, given the high cross-section compared to other possible channels of KII, IMB, and Baksan.We adopted the differential cross section (in the scattering angle) calculated by Vogel and Beacom in [22].
Off-set time
Another thing that we have to be careful of is to not confuse the time of the first detected neutrino t 1 with the time t 0 = t = 0 which indicates the time that the first neutrino arrives at the detector, even if it was not detected.Not considering this may force that the first detected neutrino is originated from the initial accretion phase, which may not be the case.As we will discuss later, for the MSW conversion in the inverted mass hierarchy scenario (IH), the initial νe flux contributes only to 2% of the detected flux, which makes it probable that the first detected neutrino came from the cooling phase and then t 1 ̸ = t 0 .To get around this problem, it is usual to introduce an offset time t d off = t 1 − t 0 between the first detected neutrino and the time of arrival of the first neutrino, which may be different for each detector given that they do not have an equal absolute time.
Background Modeling
In a realistic approach, we have to consider that detected events may come from background sources.The background rate is considered to be constant over the time of exposure, and also uniform over space, i.e., it depends only on the positron energy of the event B = B(E i ) = d2 N B /dtdE.The independence regarding the spatial position is an approximation, given that there is more background at the wall of the detector, due to the surrounding material.
The background can be measured and it is published by the collaborations.As argued in [23], there is no need to do a convolution of these measured background rates with a Gaussian uncertainty in the energy, as done in [6], given that the background curve adjusted to the data already accounts for the uncertainty in the measurement.Therefore, one only needs to take the background rate from the experimental curve without doing a posteriori uncertainty convolution, which would double count the uncertainty effect.In our case, we use the background rate from [16] for both Kamiokande-II and Baksan, whereas the background is irrelevant for the IMB detector.In the case of the Time-Integrated analysis, we have to integrate the background rate in time to get the event rate per energy B = B(E i ) = dN B /dE.The integration has to be done on the time of exposure to the supernova signal, i.e., the data-taking duration (∼ 30s).
Statistical Analysis
For the statistical analysis, we use the method of maximum unbinned likelihood, due to the low number of events.Our expression for the likelihood is similar to the one adopted in [7] Here we made implicitly the dependency of L in the parameters of our models.In this equation, i is the index of each event, R(t, E, cos θ ) is the expected event rate from equation (9), R(t) the event rate integrated in the angle and energy, and B the background rate 2 discussed in section 2.8.
Here we differ from [7] in the definition of R(t), in which we consider the total efficiency to calculate the event rate integrated in the energy, as discussed in section 2.4.The integration in the positron energy E e is made considering a Gaussian distribution L i (E e ) around the measured value E e,i with standard deviation given by the measurement uncertainty.As already discussed, we consider that the time and angle uncertainties are irrelevant.We also consider the dead time τ d for each detector (d = K, B, I), where f d is the live-time fraction [7].In the case of the time-independent model, we only have to consider a time integration in the event rate for the signal R(t i , E e,i , cos θ i ) and for the background B(E i ).
To find the set of parameters that best adjusts our model to the data, we only have to maximize the likelihood L or minimize −2 log(L ).The last one is useful because it transforms multiplication into a sum and has a straightforward connection to confidence intervals.Given that we have a set of parameters ⃗ θ , taking their the best-fit ⃗ θ we can define the likelihood ratio as follows.
so that −2 log λ ( ⃗ θ ) follows a χ 2 distribution in the asymptotic limit of large samples N → ∞, with m degrees of freedom representing the number of parameters not constrained to be in its best-fit value.With this procedure, we can estimate the best-fit values for the parameters and their confidence interval, given a confidence level.However, we have to note that our data is not a large sample so our confidence level is an approximation.In any case, in this paper, we consider that it is an acceptable approximation given the allowed region for the astrophysical parameters to be comparable to previous works [6] that use other approaches to set the confidence levels, as we discuss in Appendix A.
Time-dependent model
For the time-dependent model, following the references [6,7], we consider two possible cases, one with just cooling emission and the other with an initial accretion phase.For the cooling component, we have four astrophysical parameters, the initial cooling temperature T c , the time constant of the phase τ c , the radius of the neutrinosphere R c , and the ratio between the initial temperatures of the electronic and non-electronic antineutrinos τ = T νx /T νe .Previous works [7] fix this temperature ratio based on supernova simulations.Here, we check the impact of changing this ratio given that it has strong implications in how similar the initial spectra are, which reflects how well we can identify flavor conversion in the detected spectrum.Nevertheless, we limit ourselves to the range of temperature ratio expected from supernova simulations [17].When considering the accretion phase, we introduce three new astrophysical parameters: the initial accretion temperature T a , the time constant of the phase τ a , and the accretion mass M a .In addition to the astrophysical parameters, there is the offset time for each detector and the survival probability, resulting in a total of 8 parameters for the cooling model and 11 for the cooling plus accretion.
To analyze how the SN1987A data can put limits on P ee , we can do a marginal analysis, as described in section 3. Figures 1 and 2 show the marginal plot of P ee for the models with only cooling component and for the one with cooling and accretion, respectively.For the model with just cooling, we can see that it is not possible to put limits on P ee up to the 1σ for τ values considered.This probably happens because both initial fluxes φ 0 ν e and φ 0 ν x come from the same mechanism, resulting in almost indistinguishable spectra, even allowing the temperatures to be different.
When we consider the accretion phase, we have a different scenario, where P ee ∼ 0 is strongly rejected, as we can see in Figure 2.This stronger constraint in P ee happens because in the accretion mechanism only electrons antineutrinos are emitted, making their initial flux φ 0 ν e more distinguishable from the non-electronic one φ 0 ν x , which in turns facilitates the identification of flavor conversion.Given that, the excluded region of P ee ∼ 0 corresponds to the case where the detected flux is composed only by the initial φ 0 ν x , i.e., a flux with no accretion component.This shows us that the detected electron antineutrinos are better described by a flux with an accretion component coming from φ 0 ν e , as already found by [6].However, in [6] they do not consider the role of flavor conversion, while here we can see that the existence of an accretion component has strong implications on the conversion mechanism.If we consider only the MSW effect with adiabatic propagation, this implies that the normal hierarchy scenario is favored over the inverted.Comparing them with the best-fit of free P ee , the normal hierarchy scenario is not significantly rejected, while the inverted one is rejected by ∼ 3σ of significance.
It is also possible to see in figure 2 some kind of discrete transition to a lower ∆ χ 2 at P ee ∼ 0.5.This happens because there is a preference for a non-zero off-set time in the IMB data, as can be seen in the best-fit value of t I o f f in table 2 if the accretion component is strong enough (MSW-NH or free P ee ).However, if we go to lower values of P ee , such as in the MSW-IH, it becomes preferable to describe some of the first events of IMB as coming from the cooling, i.e. t I o f f = 0.This transition can be seen in figure 3 in which we plot the ∆ χ 2 profile for t I o f f = 0.5 and 0 s.We have also tested the implications of considering the cooling and accretion components as contemporaneous.As argued by [7], there is no evidence of a composed spectrum in supernova simulations, so the two mechanisms with dif- ferent mean energies should occur at different times.However, from supernovae physics, we may expect that the PNS starts to cool down by neutrino emission soon after its formation, simultaneously with the accretion mechanism [24].Therefore, we decide to test the implications of that hypothesis in our analysis.As we can see in Figure 4 there is no significant modification on P ee limits.The only modification appears on the best-fit of t IMB off , which can be seen in Appendix A.
Time-integrated model
For the time-integrated model, we considered a Fermi-Dirac emission (α ν e = α ν x = 2.3), a choice that does not have big impact in the fitting for 2.3 < α < 4 3 .We also con- 3 By letting α νe and α νx run free in this interval, the variation of the likelihood ratio L /L max was not above 1σ (C.L. ≈ 68%).sider a hierarchy for the mean energy E ν x > E ν e , which is physically motivated given that non-electron neutrinos interact less (lack of τ and µ leptons in the environment) and then escape from deeper regions in the supernova with higher temperatures.The best-fit values for the astrophysical parameters are shown in Table 3 considering the 3 different conversion scenarios.As we can see, there is a preference for a detected spectrum φ ν e to be composed mostly by the initial non-electron neutrino spectrum φ 0 ν x , given that there is basically no constraint for the total energy ε ν e , the same behavior was also found in [10].Even in the MSW mechanism with inverted mass hierarchy, where the composition of φ 0 ν x in the final flux is small (P ee ≈ 2.18%, the flavor conversion is compensated by a higher total energy ε ν x .This preference is a combination of the imposed energy hierarchy E ν x > E ν e and the low detection efficiency for lower energies, where the low energy events can be as well described as coming from the background.However, we did not inves- tigate this preference deeply 4 .As we are interested in the flavor conversion parameter P ee , we leave the Appendix A to compare our marginal and contour plots with previous analyses to show the consistency of our method, at least regarding the astrophysical parameters. For the flavor conversion analysis, we again fix the initial temperature ratio (more precisely the mean energy ratio τ = E ν x /E ν e = T ν x /T ν e ) and let the other parameters run freely over the allowed range (Table 3).Figure 5 shows the marginal plot of P ee minimizing over the other model parameters.Again, there is no constraint on the survival probability above 68% of confidence, even for spectra with higher mean energy differences such as τ = 1.4.
Problems with fitting the data with some models
In our numerical implementation, we found some difficulties in working with the two-component model (accretion + cooling).The main one is the existence of different local minima, which make the minimizer algorithm give different best fits depending on the initial conditions.To get around this problem, we used two methods to find the global minimum.In the first method we fit this model multiple times (≈ 1000) fluctuating the initial conditions of parameters uniformly in the ranges shown in Table 2, and taking the minimum value of −2 log L as the initial condition to find the global best-fit.The second method was based on using different minimizers (MINOS, scipy, simplex) 5 to see if this dependency on the initial conditions was algorithm dependent.In the end, we found that all the different minimizers obtained the same best fit given initial conditions around it, and in agreement with the first method.Given the concordance between the two methods and algorithms, we have confidence that the best fit obtained is the most probable one inside the allowed parameter space.
Conclusion
In this paper, we have explored the role of flavor conversion in the SN1987A neutrino data, and how it can impose limits on the flavor conversion mechanism.We found that the time-integrated model, which uses only the energy information, could not put any limit on the electron antineutrino survival probability P ee .The same happens for the timedependent models that consider antineutrino emission only from the cooling mechanism.However, with the existence of an accretion emission of electron antineutrinos, strong limits are imposed on low values of P ee .This is impressive given the low statistics of the SN1987A neutrino data and it is in agreement with the previous work of Lamb and Loredo [6] in which the data shows a strong preference for the existence of an accretion component.
In previous works, such as [19], it was already pointed out that the inverted mass hierarchy was disfavored in MSW adiabatic scenario with a significance of 3σ for some values of θ 13 , which was unknown at that time.Here we confirm this statement, as it can be seen from the figures 2 and 4. Our improvement to their analysis was to use the current well-known neutrino vacuum mixing angles [20] and extend the analysis to the whole spectrum of possible values for the survival probability P ee .
As we discussed, our analysis does not consider any time or energy dependency on P ee , which may happen when we consider collective effects due to neutrino-neutrino forward scattering.We leave the study of time and energy dependency for a future paper.In any case, our results can still be used to constrain conversion models that result in a fixed value for P ee .Fig. 6 T c,0 vs R c contour plots comparing our results with previous ones [6,7].
parameter agrees with the obtained contour plots.Also, the conversion model with free P ee encompasses the MSW-IH and MSW-NH scenarios, as one would expect given that the latter are specific cases from the former, with P ee ≈ 2.18% and P ee ≈ 67.8% respectively.7 Marginal and contour plots for the astrophysical parameters T c , τ c , R c for the only cooling model, keeping the detection off-set times t KII off ,t IMB off ,t Bak off in their best-fit value.For the contour plots, we use color bands for the MSW-IH and MSW-NH scenarios and lines for the free P ē ē, corresponding to confidence levels of 68% (dashed) and 99.7% (solid).Note that minimum χ 2 min = −2 log L max is the absolute one among all the curves/conversion scenarios.
2 |U e1 | 2 |U e3 | 2 MSWτFig. 1 P 2 τFig. 2
Fig. 1 P ee likelihood ratio (∆ χ 2 = −2 log L /L max ) for the SN1987A data considering the time-dependent model with only the cooling component.The horizontal dashed lines correspond to 1, 2 and 3σ of C.L. Note that minimum χ 2 min = −2 log L max is the one absolute regarding all the curves.
Fig. 3 2 τFig. 4
Fig. 3 Same as Fig. 2 but fixing τ 1.2 for two different values of off-set time for the IMB data t I o f f
2 τFig. 5 P
Fig. 5 P ee likelihood ratio for the SN1987A data considering the timeintegrated model.
Fig.
Fig.7Marginal and contour plots for the astrophysical parameters T c , τ c , R c for the only cooling model, keeping the detection off-set times t KII off ,t IMB off ,t Bak off in their best-fit value.For the contour plots, we use color bands for the MSW-IH and MSW-NH scenarios and lines for the free P ē ē, corresponding to confidence levels of 68% (dashed) and 99.7% (solid).Note that minimum χ 2 min = −2 log L max is the absolute one among all the curves/conversion scenarios.
Fig. 8
Fig. 8 Same as figure 7, but including the accretion component with parameters new parameters T c , τ c , R c , T a , τ a , M a .
Fig. 10
Fig.10Intrinsic (solid curves) and total efficiencies (dashed and dotted curves) for each detector.
Table 1
Range and best-fit (BF) for all parameters in the timedependent model Only Cooling.We show the best-fit for three flavor conversion scenarios: MSW with NH, MSW with IH, and a modelindependent free P ee .
Table 4
Characteristics of each detector | 7,088.2 | 2023-01-26T00:00:00.000 | [
"Physics"
] |
Observation of centimetre-scale argon diffusion in alkali feldspars: implications for 40Ar/39Ar thermochronology
Abstract New data from a gem-quality feldspar from Itrongay, Madagascar, record naturally occurring 40Ar/39Ar age profiles which can be numerically modelled by invoking a single diffusion mechanism and show that microtexturally simple crystals are capable of recording complex thermal histories. We present the longest directly measured, naturally produced 40Ar*-closure profiles from a single, homogeneous orthoclase feldspar. These data appear to confirm the assumption that laboratory derived diffusion parameters are valid in nature and over geological timescales. Diffusion domains are defined by crystal faces and ancient cracks, thus in gem-quality feldspars the diffusion domain size equates to the physical grain size. The data also illustrate the potential of large, gem-quality feldspars to record detailed thermal histories over tens of millions of years and such samples should be considered for future studies on the slow cooling of continental crust. Supplementary material: Ar-isotope data, standards and constants used in calculations and irradiation parameters are available at http://www.geolsoc.org.uk/SUP18720.
Understanding the diffusive transport of argon in minerals, both within nature and the laboratory, is critical to our ability to apply 40 Ar/ 39 Ar data to thermochronological studies. Advances in laser spot and depth profile extraction techniques, and in stepped and cycle heating, have largely confirmed the Arrhenius relationship between diffusion rates and temperature (e.g. Foland 1974;Arnaud & Kelley 1997;Lovera et al. 1997;Wartho et al. 1999;Lovera et al. 2002), although recent analyses of argon diffusion at submicron spatial resolution (Watson & Cherniak 2003;Thomas et al. 2008;Baxter 2010;Clay et al. 2010) appear to indicate more than one diffusion mechanism in silicates. In contrast to the general consensus over the relationship between diffusion rates and temperature, there has been considerable debate and uncertainty concerning argon-loss systematics resulting from complexity in sample microstructure, including perthite boundaries, subgrains, and fast diffusion pathways. It has been hypothesized that multipath diffusion combining lattice diffusion and diffusion via defect pathways might act to limit the effective domain size to less than the visible grains (e.g. Lee 1995). Parsons et al. (1988) and Burgess et al. (1992) demonstrated that alkali feldspar 'patch' perthite domain boundaries define diffusion domains while 'tweed' perthite domain boundaries do not. Further, Reddy et al. (2001) showed that apparent ages within a detrital K-feldspar did not vary across coarse perthite boundaries, indicating that such boundaries do not always act as diffusion domain boundaries, but in the same sample deformationinduced microstructures did define diffusion domains. More recently, Mark et al. (2008) showed that the smallest observed subgrains do not always define diffusion domains in authigenic K-feldspar. In summary, there have been several datasets on natural argon diffusion in complex samples but not in individual K-feldspar domains. We note that the debate continues in articles in the present special issue. Our motivation in analysing large individual grains of gem-quality K-feldspar was to test the coherence of argon diffusion in large grains from the micron scale at the grain surface to bulk diffusion over many millimetres within the same grain.
The Itrongay feldspar from Madagascar is an ideal sample on which to test the coherence of diffusion and the potential for multiple diffusion mechanisms for argon in K-feldspar. It is a clear, yellow-coloured, pegmatitic low-sanidine/orthoclase feldspar and detailed descriptions of various samples have been previously published (Coombs 1954;Wartho et al. 1999). Pegmatites generally form rapidly from supersaturated, undercooled fluids so are unlikely to be maintained at high temperatures for long periods of time (London 1996). Thus, the simple microtexture of Itrongay feldspar was thought to reflect a combination of both the simple geological history and the purity of the orthoclase (Or) composition lattice. Given the apparent simple crystal structure and rapid formation, these crystals represent an excellent natural laboratory to test the models for bulk argon diffusion which cannot be imaged in smaller crystals, and thus have to be inferred from low resolution laser profiles or from cycle or stepped heating experiments. Published radiometric ages for Itrongay feldspar range from 435 Ma (Arnaud & Kelley 1997) to 461 Ma (Wartho et al. 1999;Nägler & Villa 2000) using the 40 Ar/ 39 Ar dating technique, and up to 477 Ma for K/Ca ages (Nägler & Villa 2000). While the presence of this age discrepancy has been noted (Wartho et al. 1999) only Nägler & Villa (2000) have discussed possible causes; they have invoked either a discrete thermal disturbance or maintenance of temperatures above 350 8C as mechanisms that would have resulted in resetting of the 40 Ar/ 39 Ar system.
We have investigated two samples of Itrongay K-feldspar: one which was cut from a fragment of a larger grain (Crystal A, Fig. 1b); and a second sample which was a complete crystal exhibiting an untwinned euhedral shape and the characteristic monoclinic crystal habit (Crystal B, Fig. 1c). The perfect shape of Crystal B indicates that the crystal surfaces are original and formed as the pegmatite crystallized. Their origin and the planar form of these crystal surfaces make them ideal targets for the same depth profiling that has been used for the analysis of laboratory induced argon diffusion (e.g. Arnaud & Kelley 1997;Wartho et al. 1999). Arnaud & Kelley (1997) investigated diffusion in the Itrongay feldspar by combined cycle heating, laser depth profiling and crushing on millimetresized fragments. Their data showed age variations related to domains that differed in size by a factor of around five, but the sizes were unspecified. The conclusions of combining laser depth profile measurements of 39 Ar appeared to indicate grains acting as single diffusion domains with some evidence for fast-path diffusion, an observation backed up by the better fits achieved in Multi-Domain Diffusion (MDD) modelling by the use of variable activation energies. In addition, sample crushing appeared to indicate traps in the structure with the potential to retain argon relative to the mineral lattice. Arnaud and Kelley concluded that volume diffusion took place over scales of at least 100 mm and that defects in the crystal acted as a reservoir for an apparent excess 40 Ar component. These traps were also considered to influence argon loss from the outer 10 mm of the sample and to act as fast diffusion pathways once they were opened to the surface during sample preparation. Subsequent UV laser depth profiling provided high spatial resolution analysis of laboratory induced argon diffusion. Samples were heated in a high pressure argon atmosphere by Wartho et al. (1999), who concluded that argon diffusion occurred by volume diffusion to the visible grain boundary. The diffusion parameters from these studies corroborated the accepted laboratory values of activation energy (E) and pre-exponential factor (D 0 ) for alkali feldspars and values determined by step heating and cycle heating of natural samples. While it has been implied that argon diffusion domains in Itrongay feldspar should equate to physical grain sizes, Arnaud & Kelley (1997) hypothesized fast-track diffusion in the outermost few microns, and other studies have measured varying laboratory diffusion and solubility in the outermost micron (Watson & Cherniak 2003;Thomas et al. 2008;Baxter 2010;Clay et al. 2010).
Sample preparation and 40 Ar/ 39 Ar methodology
Two samples of Itrongay alkali feldspar were used for this study: the first (Crystal A) was a 100 mm thick polished slice (final polish using 0.3 mm aluminium oxide paste) cut from a larger broken crystal more than 1 cm in diameter (Fig 1b); the second sample (Crystal B) was a whole crystal exhibiting a characteristic feldspar monoclinic shape and thus ancient crystal boundaries (Fig. 1c). From Crystal B we sampled three crystal surfaces to a depth of c. 1 mm using a slow saw (Fig. 1d were very near planar (although scratched) and we interpret these as representing the original crystal faces, while the [110] sample was considerably more undulating, possibly representing loss of material from the surface. Selected slices of the two samples were ultrasonically washed in acetone and distilled water, dried, wrapped in aluminium foil and irradiated at the McMaster Reactor (Canada) using cadmium shielding. The neutron flux was monitored using biotite standard GA1550, which has an age of 99.8 + 0.2 Ma (Renne et al. 1998(Renne et al. , 2010 and gave J-values of 0.01041 + 0.000052 and 0.01265 + 0.000063. Standards were analysed by melting in vacuo using an infrared 1090 nm fibre laser to melt individual biotite grains. Gases were cleaned by getters in a similar manner to the Itrongay samples, except that an in-line cold finger was used to reduce water and carbon dioxide prior to gettering. The cleaned gases from the standards were analysed using an MAP 215-50 noble gas mass spectrometer. A New Wave Research Ltd LUP 213 nm pulsed quintrupled Nd-YAG laser was used to ablate the K-feldspar samples producing two different types of profiles. Pits c.100 mm in diameter were ablated producing a traverse across the polished and cleaved surfaces, and depth profiles from the natural crystal surfaces were produced by repeated rastering over 200 mm squares (see Wartho et al. 1999). Three transect profiles were analysed on the polished surface of Crystal A, creating age profiles nearly 1 cm long. Each of the three ancient surfaces of Crystal B was analysed twice, creating six depth profiles of around 20 mm. In addition, since the laser ablation process is inefficient at great depths, we cleaved two of the three ancient crystal surfaces to reveal fresh surfaces orthogonal to the crystal face and created ablation traverses up to 1 mm depth into the original crystal using laser spots. Once the gas released had equilibrated in the extraction system, two SAES getters removed unwanted gas species over a period of less than 5 min, one operating at room temperature and the other at 450 8C, for a minimum of 30 s prior to automatic inlet into the Nu Instruments Noblesse noble gas mass spectrometer. Interfering hydrocarbon and chlorine peaks were measured at masses 35, 36 and 39 in addition to the argon isotopes to monitor for hydrocarbon or chlorine interferences, none of which were observed at levels greater than system blanks. Hydrocarbon peaks at masses 36 and 39 were measured as shoulders on the side of the argon peaks, always at the high mass side. All analyses were corrected for blanks, 37 Ar decay and neutron-induced 40 Ar (using the correction factor ( 40 Ar/ 39 Ar) K ¼ 0.0085); Ca-derived 39 Ar and 36 Ar were not corrected for due to the very low abundance of Ca in Itrongay feldspar (10 ppm : Nägler & Villa 2000) resulting in measured levels of 37 Ar Ca being indistinguishable from the blank. Yields of 36 Ar were low in all analyses, although it was always measurable in the analysis of the natural crystal surfaces. In many analyses, however, 36 Ar measurements were indistinguishable from the blank even when averaged over a day. Regular monitoring of the 36 Ar peak position, instrument background and hydrocarbon peaks confirmed that the low 36 Ar yields are genuine and not due to measurement error or reduced detection limits from an elevated instrument background. In addition, detailed analysis of both young sanidines and mineral standards, using a similar approach to Renne et al. (2012) on the same instrument, confirmed that this procedure ensures the correct measurement of 36 Ar. Therefore, these low 36 Ar yields are considered to reflect very low 36 Ar levels within the sample. Where the 36 Ar yields are indistinguishable from or lower than the blank (i.e. the blankcorrected 36 Ar value is zero or negative), it is assumed that the atmospheric argon content of the analysis is effectively zero.
Results
Crystal A argon-isotope and age data are given in the supplementary data table S1 and the ages are shown in Figure 2. The ages range from 450.3 + 2.4 Ma to 473.8 + 2.3 Ma, showing over 20 Ma of age variation. Most of this age range is included in an 8 mm long profile (A -B) across the slab (Fig. 2). This profile shows a steep increase in age from c. 450 -458 Ma within the first millimetre, followed by a gentler increase to c. 469 Ma at 4 mm where the profile plateaus. This profile begins within the crystal, at the edge of a fragment of the polished slab, the edge being formed by a cleavage plane next to a pre-existing crack in the crystal (Fig. 1b).
Profile C-D, which runs orthogonal to the other profiles, shows a smooth but less pronounced increase in age (c. 464 to c. 473 Ma) from the top to the bottom of the slab. The final profile (E-F) runs horizontally across the bottom of the slab and shows a slight, but insignificant age increase from left (471 Ma) to right (473 Ma). As a whole, the slab shows an increase in age from the top left corner to the bottom right.
The 40 Ar/ 39 Ar isotope and age data for Crystal B are given in the supplementary data table S2 and shown in Figure 3. Age variations are greater than for Crystal A, from as low as 415.7 + 3 Ma to 471.2 + 3.4 Ma, a range of 55 Ma (cf. 20 Ma range in Crystal A). Note, however, that the range only extends to lower values; the highest value for both crystals are within experimental errors and also agree closely with the determined K/Ca age for samples from the same locality (477 + 2 Ma: Nägler & Villa 2000). When all age analyses of Crystal B are plotted against the distance from the crystal boundary (Fig. 3) they display a very coherent trend from young ages at the surface, increasing rapidly with distance at first to around 440 Ma by 20 mm depth, but rising at a decreasing rate at greater distances to reach around 470 Ma at 1 mm distance from the ancient crystal surface.
There is one departure from this pattern -the data we obtained from the third surface [110], which we noted was less regular in form than the two planar surfaces (see supplementary data table S2). Individual 40 Ar/ 39 Ar ages in this depth profile did not show any consistent variation with depth, and appeared to correspond with the highest ages in the other depth profiles. Ages ranged from 457.5 + 3.6 Ma to 471.7 + 3.1 Ma. It seems likely that the present surface was not a grain boundary during the time at which the other boundaries lost radiogenic argon. Given its more pitted appearance it seems likely that dissolution has taken place, and, since ages from other surfaces were consistently around 415-425 Ma, that up to several hundred microns have been lost from the surface. This observation also raises the possibility that mass has been lost at other crystal faces and that ages younger than 415-425 Ma may have been recorded (and subsequently lost) on completely undisturbed surfaces, but the planar nature of the [010] and [101] surfaces suggests that this is unlikely. We also cannot rule out the possibility that this is the original crystal surface, but it did not act as a boundary connected to an infinite reservoir -that is, it was not a grain boundary for diffusional argon loss. Mark et al. (2008) have shown that such effects can occur on a micron scale in feldspar overgrowths, and there is evidence that mica grain boundaries have variable permeability (Smith et al. 2005) leading to variations in edge ages across one grain.
One feature of the Crystal B age profiles is that there is little variation between ages in different profiles at the same distance from the grain boundary. The only zone which deviates from this pattern is that of the outermost 2 mm where we recorded ages varying by over 10 Ma. However, we hesitate to place too much of our interpretation on this outer zone since we note that the crystal is visibly scratched on the two flat surfaces. Such variations from a perfect surface may lead to enhanced argon loss via fast-track diffusion pathways (i.e. loss via deformed lattice beneath the scratches), thus it may not be appropriate to interpret these ages in terms of volume diffusion and we place less emphasis on the outermost points while noting that the thermal history recorded lasted beyond the c. 425 Ma ages recorded by many points close to the grain edge. We also note that it is possible that the age profiles continue to steepen towards the grain boundary up to the final few nanometres such that we were unable to measure them, and that there is a pattern reflecting a younger thermal history in the outermost micron of the grain.
The coherence of all the profiles in the Crystal B data allows us to observe that the age profile appears to deviate from those that might be Fig. 1b) where ages are displayed rounded to the nearest Ma; errors are not shown but are +2-3 Ma at 2s. Dashed line represents the DIFFARG model of Scenario 3 (see text) and the arrows show the direction this curve is expected to move if an age profile was measured at an angle not perpendicular to the grain edge. produced from rapid 40 Ar* loss or gain (e.g. Pickles et al. 1997) in that it appears to steepen close to the grain edge, resembling more closely a slow-cooling profile (Dodson 1986).
Argon diffusion modelling and discussion
All the age profiles described and illustrated in Figures 2 and 3 resemble age profiles resulting from diffusional loss of 40 Ar* across the whole crystal and thus centimetre-scale distances. Indeed to our knowledge these are the longest natural diffusion profiles yet recorded and are around an order of magnitude longer than previously known examples (e.g. Reddy et al. 2001). The implication of such long and coherent age profiles in a gem-quality alkali feldspar is that the thermal history reflects either long slow cooling, or reheating and loss of radiogenic argon at some point after the initial cooling. Pegmatites form rapidly from supersaturated undercooled liquids (e.g. London 1996), so the presence of these age profiles must represent diffusional loss of 40 Ar* through the crystal lattice during a prolonged thermal history. Since the thermal history of the region is already constrained (see below) we will be able to compare diffusion profiles produced by combined loss and radiogenic build up, as observed from natural diffusion, with models based on laboratory determined diffusion rates with natural diffusion.
Measurements at high spatial resolution of 40 Ar/ 39 Ar age profiles can be modelled, using the numerical model DIFFARG (Wheeler 1996), to reconstruct possible thermochronological histories for the sample. Conversely, where the thermal model is already known to some extent, such models can be used to test the assumptions underlying the model including the diffusion parameters E and D 0 , the boundaries and grain size assumed, and also the relationship between diffusion rates measured in a laboratory and those extant in natural samples cooled on geological timescales. DIFFARG calculates radial age profiles that are perpendicular to the grain boundary. Comparison of modelled age profiles with our data provides insights into both the thermal history and diffusion of argon within the sample.
While it is not possible for modelled age profiles to be precisely fitted to the data for Crystal A because the crystal orientation and the precise location of original crystal surfaces are unknown, the measured age profiles suggest coherent diffusive loss of 40 Ar* over distances of at least 1 cm. Fitting the precise shape using spherical or even one dimensional diffusion is open to some interpretation since we do not know the orientation relative to the ancient grain boundaries. However, we note that the minimum and maximum ages are comparable to the profiles from Crystal B, and while the youngest ages in Crystal A occur within the crystal, rather than at its edge, they are found close to two prominent cracks, one of which forms the edge of the fragment and is likely to be a cleavage plane. The most likely explanation of this age distribution is that the crystal acted as a single domain bounded by crystal edges and some early fractures, which resulted in diffusional mal history of the area based on the known thermal history constraints and the argon-loss model which assumes the use of laboratory determined diffusion parameters (Wartho et al. 1999). The thermal history of Madagascar has been extensively studied in the context of the amalgamation and subsequent break-up of Gondwana, and thermochronological modelling of the 40 Ar/ 39 Ar data is likely to reflect these events. The Pan-African Orogeny began 650-700 Ma, with the later stages of continental collision occurring c. 550-570 Ma and final suturing completed by c. 535 Ma, with localized thermal perturbations relating to shear zones (Paquette & Nédélec 1998;Emmel et al. 2006;Grégoire et al. 2009; and references therein). Despite the localized thermal and structural complexities, various thermochronological studies from different terrains show similar cooling histories after c. 535 Ma (Fig. 4). We have used these models to provide constraints on the post-amalgamation cooling history of Madagascar for the area around Itrongay from where the feldspars derive. The final event occurred around 88 Ma when widespread magmatism and volcanism occurred in southern Madagascar over a 6 myr period, related to the Marion Hotspot and the separation of Madagascar from India (Storey et al. 1995). This may have caused crustal heating in southern Madagascar, although this was not identified by Seward et al. (2004). We have assumed the thermal history started at 477 Ma based on the K/Ca age of Nägler & Villa (2000). The oldest ages recorded in either Crystal A or B are around 471 Ma in the core of the grains, and thus K-feldspar began to retain argon quantitatively c. 6 million years after the initial intrusion. Nägler & Villa (2000) suggested two thermal history scenarios to account for the difference between K/Ca and 40 Ar/ 39 Ar ages, based on the different diffusivities of calcium and argon, and on the monoclinic structure of the feldspar. However, both scenarios were based on the assumption that the oldest 40 Ar/ 39 Ar age is 461 Ma (previously measured on bulk samples), whereas our experiments have demonstrated that the oldest ages recorded are c. 471 Ma. In order to further constrain the thermal history and model the argon loss, we determined the youngest ages in the thermal history based on ages at the edge of Crystal B profiles, ignoring the youngest ages which appear to steepen suddenly in the outer 2 mm. We thus assume a youngest age of 423 Ma since it produces the best fits to the lower age data.
Three thermal history scenarios were modelled. In the first (Scenario 1) we attempted to model the age profiles based on a simple thermal history of pegmatite emplacement followed by rapid cooling to 471 Ma and subsequent reheating at 423 Ma to test whether this scenario reproduced the form of the age profiles in Crystal B. The two grey dashed lines in Figure 3 illustrate model curves for the scenario of reheating at 423 Ma for 1 myr to temperatures of 320 8C and 270 8C. The lower temperature produced an age profile that fits the ages in the outer 20 mm but not the rest of the profile. The higher temperature model fits the data between 400-800 mm from the crystal surface but not the data closer to the crystal surface. Thus neither these two end members nor any intermediate temperature reheating event produces a good fit to all the data and we conclude that the simple reheating model is not a valid explanation for the age profiles in Crystal B. We hypothesized above that the steepening of the profiles towards the grain boundary resembled slow cooling, but a linear cooling history (Scenario 2, various cooling rates modelled) produced a similarly poor fit because the age difference (50 myr) and difference in temperatures recorded are insufficient to produce the hypothetical 'Dodson slow cooling' patterns.
We noted that the thermal histories for Madagascar (Fig. 4) have a pattern that appears to indicate rapid initial cooling but then an extended history of much lower cooling rates, a process resembling 'cratonization' of the continental crust. Thus, in Scenario 3 (Fig. 4), we modelled a similar rapid cooling followed by a slower cooling pattern for the Itrongay sample. We used a simple two-step model to test this hypothesis although we note that this is not a unique solution. The model started with rapid cooling to 350 8C at the emplacent age of 471 Ma. This temperature was maintained for 20 myr and was followed by a second cooling to either 250 8C, 260 8C or 270 8C, which was maintained until 423 Ma (Fig. 4). This simple two-stage model produced an excellent fit to the data (Fig. 3). The best fit was for the model which combined 350 8C and 260 8C steps (solid line on Fig. 3); we (5) Meert et al. (2001), based on U-Pb zircon ages and 40 Ar/ 39 Ar hornblende, biotite and K-feldspar ages. While the published thermal histories are quite variable, owing to their regional distribution over hundreds of kilometres and their varying tectonic origins (basement rocks, upper crustal rocks, metamorphism, plutonism), they nevertheless record a general cooling trend, which is consistent with our modelled thermal history. noted that all data points (other than the three points within 2 mm of the surface) fell within two sigma errors of the model curve and the majority less than one sigma from the curve. The two-step model reproduced the inflection in the age profile seen in Figure 3 and also conformed very closely to the known thermal history. This is not presented as a unique thermal history solution, but it does demonstrate that these crystals are capable of recording variations in thermal history and that a detailed thermal history could be recovered by more analytical work on a range of grains.
Our model thermal history for Itrongay (Fig. 4), which is based on a single diffusion pathway for K-feldspar, compares very well with the thermal histories for southern and central Madagascar derived from other sources, and potentially refines them since the lower bound on many of the histories is defined by either a single published apatite fission track age (Emmel et al. 2006) or an 40 Ar/ 39 Ar multi-domain modelled age-temperature pair for a K-feldspar in the Carion granite, north-central Madagascar (Meert et al. 2001). However, we note that it may also be possible to reproduce the observed profile with a simpler thermal history if two (or more) diffusion pathways for argon in K-feldspar are invoked (see Lee 1995;Baxter 2010), but we have not attempted to numerically model this due to a lack of model constraints. Nevertheless, an important conclusion is that the model thermal history, based on the age profiles in a single gem-quality alkali feldspar grain and laboratory diffusion parameters, reproduces the thermal history derived from other sources. This corroborates the assumption that laboratory derived diffusion parameters can be used to model natural diffusion over geological timescales. Furthermore, the coherency of the Crystal B age profiles confirms that argon diffusion in homogeneous K-feldspar is isotropic and does not occur at different rates depending on the crystallographic orientation.
Our analysis of a gem-quality grain appears to confirm the capability of the lattice to retain coherent argon closure profiles (a combination of diffusive loss and radiogenic build-up) over distances of many millimetres, confirming the capability of smaller subgrains or domains (that form above their closure temperature and remain unchanged) to retain genuine thermal histories in the form of closure profiles as hypothesized by the MDD model (Lovera et al. 1989(Lovera et al. , 1997(Lovera et al. , 2002. Conversely, laser profiles of the type illustrated above are only possible on the largest gem-quality grains, but there is considerable scope for similar studies in other areas that may be able to reveal slow cooling of continental crust.
Finally, we caution that, when using pegmatitic gem-quality feldspars for diffusion experiments, a simple geological and thermochronological history should not be assumed and that chemically homogeneous feldspars do not necessarily have a homogeneous distribution of argon isotopes; any simple argon-loss experiments should check the samples for pre-existing age gradients.
Conclusions
We conclude that the Itrongay alkali feldspar records closure age profiles formed over 50 million years, and thus natural argon diffusion in alkali feldspars can be explained by the same diffusion mechanism as that observed in short-term laboratory experiments. While we did detect an apparent variation from the trend in the outer 2 mm of the grain this is likely to be the result of surface imperfections and defects deviating from a planar surface behaviour in this sample. Our modelling, which used a thermal history consistent with published constraints, shows that this naturally occurring closure profile can be produced by invoking a single diffusion mechanism and that additional argon diffusion mechanisms, pathways or defect-enhanced fast-track diffusion are not necessarily required to reproduce the steepening in the outer few tens of microns of the age profile.
The thermal history and coherence of the closure profiles also explain variations in measured ages of different samples of Itrongay feldspar over the years, since the sampling location within the grain and the original grain size will have a profound effect upon the bulk age. Large gem-quality grains such as from the Itrongay sample have the capability to reveal long and slowly changing thermal histories for areas of continental crust such as Madagascar. | 6,881 | 2013-01-10T00:00:00.000 | [
"Geology"
] |
Impact of wormlike micelles on nano and macroscopic structure of TEMPO-oxidized cellulose nanofibril hydrogels
In this work, we investigated the effect of adding surfactant mixtures on the rheological properties of TEMPO-oxidized cellulose nanofibril (OCNF) saline dispersions. Three surfactant mixtures were studied: cocamidopropyl betaine (CAPB)/sodium dodecyl sulfate (SDS), which forms wormlike micelles (WLMs); cocamidopropylamine oxide (CAPOx)/SDS, which forms long rods; and CAPB/sodium lauroyl sarcosinate (SLS), which forms spherical micelles. The presence of micelles in these surfactant mixtures, independent of their morphology, leads to an increase of tan d , making the gels less solid-like, therefore acting as a plasticizer. WLMs were able to suppress strain stiffening normally observed in OCNF gels at large strains. OCNF/WLM gels have lower G 0 values than OCNF gels while the other micellar morphologies have a reduced impact on G 0 . The presence of unconnected micelles leads to increased dissipative deformation in OCNF gels without affecting the connectivity of the fibrils, while the presence of entangled micelles interferes with the OCNF network.
Introduction
Soft materials and complex fluids are ubiquitous materials in modern life. Ever present in food and health-care products, additives are employed as tools to tailor the right rheological response according to the end applications. The rheological behaviour of these additives is not simply determined by the intrinsic properties of their components, but how their 3D structure interacts and is shaped at different length-scales. Physical hydrogels are of particular interest, as their selfassembled 3D-structures are maintained by a fine balance of transient interactions. By playing with this balance, the physical properties can be tuned. One way to influence the gel network self-assembly is to combine different types of networks in order to guide or restrict their self-assembly. 1 Cellulose nanofibrils, a type of nanocellulose, can form colloidal networks. 2,3 In plants, cellulose is found as a tightly bound pack of nanosized fibrils, which, once individualized, provide a renewable source of nanoparticles, also called nanocelluloses. 4,5 These nanocelluloses can be roughly divided into two main groups: cellulose nanofibrils (CNFs), obtained via mechanical disintegration, and cellulose nanocrystals (CNCs), obtained from acid hydrolysis of plant-based cellulose. CNFs have lengths of a few hundred nanometres and cross-sections of up to tens of nanometres, leading to particles with very large aspect ratios, while CNCs are generally shorter cylinders. 5 Nanocelluloses can be readily surface modified, allowing an exceptional level of tailoring to specific applications. 6-10 A common modification is TEMPOmediated oxidation, 11 a chemo-selective oxidation of the glucosyl C6 primary hydroxyl groups by NaOCl mediated by (2,2,6,6tetramethyl-piperidin-1-yl)oxyl (TEMPO)/NaBr in water. The resulting oxidized cellulose nanofibrils (OCNFs) are anionic nanoparticles capable of forming stable dispersions of individualized nanofibrils in water. OCNF dispersions can undergo gelation due to changes in the aqueous environment such as addition of alcohols, 12 surfactants, 13,14 salts, [15][16][17][18] or block copolymers, 19 and changes in pH [20][21][22] or temperature. 23 Surfactants are capable of selfassembling in solution to form supra-molecular aggregates that can adopt a myriad of forms. 24 Spherical micelles are the most common morphology, but cylindrical, lamellar and vesicle morphologies can form given the right environment. 25 Of particular interest are the long and flexible cylindrical micelles, commonly named wormlike micelles (WLMs), as their rheological behaviour is similar to that of polymers and they are able to form entangled networks, imparting strong viscoelastic properties to the solution. 25,26 A convenient way to obtain WLMs is through surfactant mixtures, as the range of mixed micellar aggregates accessible can be controlled by tuning the mixture composition. Also, the system can benefit from the surfactants' individual properties. 26 Combinations of surfactants are also more relevant to commercial and technological products, which usually contain a multitude of components. 27 To obtain the WLMs, we chose to work with the mixture of cocamidopropyl betaine and sodium dodecyl sulphate under saline conditions. 28 Cocamidopropyl betaine is one of the most common foam boosters used in shampoos, mainly due to its mildness and ability to form WLMs. 29 In this work, we combine both contributions: colloidal networks from OCNFs and WLM entangled networks from the surfactant mixtures. We studied the influence of surfactant mixtures on the gelation behaviour of OCNFs under saline conditions. Three mixtures, offering three different micellar morphologies were studied: cocamidopropyl betaine (CAPB)/sodium dodecyl sulphate (SDS), cocamidopropylamine oxide 30 (CAPOx)/SDS and CAPB/sodium lauroyl sarcosinate 31 (SLS). Under saline conditions, CAPB/SDS mixtures will form long rodlike micelles and, under the right conditions, 28 generate an entangled network of wormlike micelles (WLMs), while CAPOx/ SDS formed shorter cylinders and CAPB/SLS formed spherical micelles. OCNF alone also gels under saline conditions. 18 Salt is a common component in formulations where these surfactants and their mixtures are employed. The use of sodium chloride as the salt of choice both serves to induce the different micellar morphologies observed as well as imitate a common environment in applied uses. We explored the impact of the WLM entangled network on the OCNF network in contrast with other micellar aggregates also obtained from these surfactant mixtures. The gels obtained were studied via rheology and small-angle neutron scattering, providing insights on the benefits of combining these two networks and the presence or absence of cross-interactions between the micellar and nanoparticle networks.
Materials
TEMPO oxidized cellulose nanofibrils, OCNFs, with an B25% degree of oxidation, produced from purified softwood fibre processed via high pressure homogenization, were kindly provided by Croda Europe Ltd. These were further purified by dialysis against ultra-pure water (DI water), 18.2 MO cm, for 24 h. Then, the dispersion was acidified to pH 3 using HCl solution and dialysed (cellulose dialysis tubing MWCO 12400) against DI water for 24 h. The dialysed OCNF was processed via mechanical shear (ULTRA TURRAX, IKA T25 digital, for 30 minutes at 6500 rpm) and the pH was adjusted to 7 using NaOH solution and further dialysed against DI water for 3 days. The DI water was replaced twice a day. This leads to the formation of a sodium-salt, as all -COOH groups on the OCNF are now converted to -COONa. After a second dialysis step, the dispersion was diluted to ca. 2 wt% and dispersed using a sonication probe (Ultrasonic Processor, FB-505, Fisher). 40 mL of the 2 wt% dispersion was sonicated via a series of 1 s on 1 s off pulses for a total time of 60 min at 30% amplitude in an ice bath. Sodium dodecyl sulfate, SDS, (Z99.0%), N-lauroylsarcosine sodium salt, SLS, (Z99.0%) and sodium chloride, NaCl, (Z99%) were obtained from Sigma-Aldrich and used without further treatment. Commercial grade cocamidopropyl betaine (CAPB, Crodateric CAB 30-LQ-(MH), 30% aqueous solution, batch No. 1189504) and cocamidopropylamine oxide (CAPOx, Incromine Oxide C-LQ-(MH), 25% aqueous solution, batch No. 838616) were kindly donated by Croda Europe Ltd. CAPB and CAPOx were freeze-dried and redispersed before use. All samples were prepared by dilution of the aqueous stock dispersions and concentrations are given in weight/weight. Samples for smallangle neutron scattering experiments were redispersed in D 2 O (Sigma-Aldrich, 99.9 atom% D) from freeze-dried stock. All samples are in 1 wt% (ca. 173 mM) NaCl solutions. Samples were measured within 4 days of preparation. This time window allows the samples to reach a steady-state while minimizing chances of microbial contamination.
Methods
Rheological measurements were conducted using a stresscontrolled Discovery Hybrid Rheometer, Model HR-3 (TA Instruments, USA) equipped with a sand-blasted 40 mm parallel plate geometry over a sand-blasted lower plate. Temperature was controlled via a Peltier unit (AE0.1 1C) and kept at 25 1C. A thin layer of low viscosity mineral-oil was added to the edge of the geometry to prevent sample evaporation. Oscillatory amplitude sweeps were done at a fixed angular frequency (o) of 6.28 rad s À1 and amplitude strain (g) from 0.01 to 100%. Frequency sweeps were conducted at an g of 0.1%, within the linear viscoelastic range, covering the o of 0.01 to 50 rad s À1 . All samples were measured between 24 and 48 h after preparation. Small-angle neutron scattering (SANS) measurements were conducted using the time-of-flight diffractometer instrument SANS2d at the STFC ISIS Neutron and Muon Source (Didcot, UK). 32 Incident wavelengths from 1.75 to 16.5 Å were used with a sample-to-detector distance of 4 m, corresponding to a total scattering vector range q from 4.5 Â 10 À3 to 0.75 Å À1 . The sample temperature was controlled using an external circulating thermal bath (Julabo, DE). The scattering intensity was converted to the differential scattering cross-section in absolute units using ISIS standard procedures. 33 Samples were loaded in 1 mm path length, 1 cm wide optical quartz cells. 34,35 Contrast match experiments were done at 15 wt% D 2 O, which is the contrast match point for the surfactant mixtures. All other SANS experiments were done in 100 wt% D 2 O. SANS data were fitted using SASView 36 or internally developed routines written in FORTRAN. 18 The intensity I(q) can be written as follows: with P(q) being the form factor of the objects studied, giving information about their shape and S(q) being the structure factor associated with the interactions between the objects probed. Concerning OCNF, a detailed description of the data
View Article Online
This journal is © The Royal Society of Chemistry 2020 Soft Matter, 2020, 16, 4887--4896 | 4889 treatment for oxidized cellulose nanofibrils under water or saline conditions has been published by Schmitt et al. 18 Briefly, the nanofibrils are modelled as rigid cylinders with an elliptical cross-section (eqn (2)), 37,38 wherein R maj (Å) is the major radius of the fibrils, e is the ellipticity of the cross-section (e = R min /R maj ), and L (Å) is the length of the fibrils.
qR maj sin 2 y þ e 2 cos 2 y À Á 1=2 dy is the contribution associated with the elliptical cross-section and P rod ðq; LÞ ¼ L 2 2SiðqLÞ qL À 4 sin 2 qL 2 is the contribution from the length of the fibrils, with J 1 is a first order Bessel function and s represents the polydispersity in size of the cross-section, following the Schulz-Zimm distribution. 18 This decoupling between length and cross-section can be achieved due to the random distribution of orientation of the fibrils, and because R maj o L/5.
Fibril-fibril attraction was modelled using the PRISM model, 39 with n RPA o 0 being the strength of the attraction.
Depending on the surfactant used, micelles in saline solution can be shaped as spheres, (CAPB/SLS), rodlike particles (CAPOx/ SDS) or even WLMs (CAPB/SDS). Spherical micelles are defined by their radius R (Å), and the polydispersity in size s. 40 The form factor used to describe spheres (eqn (3)) is: where F sphere ðq; RÞ ¼ VðRÞ 3 sinðqRÞ À qR cosðqRÞ ½ ðqRÞ 3 where V(R) is the volume of a sphere of radius R.
Interactions between spherical micelles are modelled using the hard-sphere repulsion model, with f HS being the volume fraction of micelles interacting (in %) and R HS 4 R being the hard-sphere radius of interaction. For SLS, the Hayter-Penfold mean spherical approximation was used instead of hard-spheres to model the electrostatic interactions between micelles. This model uses the parameter Z, the value of the charge (in electrons), and f int , the volume fraction of micelles interacting (in %), to describe the interactions. 41,42 For rodlike micelles with a cross-section radius R, length L and polydispersity in radius s, the intensity can be described similarly to that for OCNFs without interaction, hence using the form factor given in eqn (2), with e = 1 for a spherical cross section. Finally, wormlike micelles can be modelled as flexible cylinders. This can be obtained by replacing the term P rod in eqn (2) by P flex = (q,L,b Kuhn ), which depends on the length, L, of the flexible cylinder and b Kuhn , the Kuhn length of the wormlike micelle. A complete description of P flex can be found in ref. 37 and is not repeated here.
Similar to OCNFs, rodlike micelles can experience repulsion, modelled via the PRISM model, 39 where n RPA 4 0 is the strength of the repulsion and R c Z R is the cross-section radius of interaction. WLMs are modelled using the same parameters, but adding the The Kuhn length is defined as twice the persistence length of the WLMs, i.e., the length scale where WLMs can be considered as rigid. Details of modelling of spherical, 41,42 rodlike 43 or wormlike micelles 37 can be found elsewhere. Radial polydispersity improved the fittings for the CAPB/SLS mixtures. For the other systems, polydispersity was not considered as it did not result in better fitting of the data. In the case of mixtures of OCNFs and micelles, the signal is modelled as a sum of the two contributions, without adding any extra contributions from OCNF-micelle interactions. This is a simplified calculation of the intensity but one that is sufficient to describe the data accurately.
Raw data is available on University of Bath Research Data Archive. 44 3 Results and discussion
Surfactant behaviour
In this work, we studied the impact of the micellar morphology on the gelation properties of TEMPO-oxidized cellulose nanofibrils (OCNFs). We focused on three surfactant mixtures, CAPB/SDS, CAPOx/SDS and CAPB/SLS, as each of these forms a different type of micelle. Before covering the OCNF/surfactant hydrogel systems, we present a brief characterisation of the surfactant solutions on their own. Here, we establish the baseline behaviour of the individual surfactants first, as a prerequisite to determining how they interact with the OCNF network. The solution behaviour of SDS micelles is well established. 13,43,45,46 They are expected to be ellipsoidal objects with half axes of 13 Â 23 Â 23 Å, in solutions with salt concentrations smaller than 0.5 M NaCl. 43 For the three other surfactants studied in this work, CAPB, 28,31,[47][48][49] CAPOx, and SLS, 47,50 the information available is more limited. We conducted SANS experiments at three different concentrations above the critical micellar concentration (CMC) for each of these surfactants. The CMC value for SLS at 25 1C in water is 12.7 mM; 47 for CAPB, at 25 1C in 0.2 M NaCl, it is 5.6 mM; 51 and for CAPOx at room temperature in 1 wt% NaCl, it is 0.06 mM (measured as part of this work via surface tension - Fig. S1, ESI †). The SANS data for the CAPB and CAPOx solutions at all concentrations can be adequately fitted using a spherical model with a hard-sphere structure factor 40,52 (Fig. 1A and B). They form spherical micelles of similar dimensions, radius R = 23 AE 1 and 24 AE 1 Å for CAPB and CAPOx, respectively. This journal is © The Royal Society of Chemistry 2020 No interaction was needed to fit the lowest concentration while a small increase in the repulsion is observed with concentration; nonetheless, this is found to be very weak (fitted parameters are in Table S1A and B, ESI †). The anionic surfactant, SLS, forms spheres of radius R = 19 AE 1 Å and evidenced stronger interactions, which are better described using a structure factor for charged spheres, in this case the Hayter-Penfold mean spherical approximation (MSA) 41,42 ( Fig. 1C and Table S1C, ESI †). Overall, the shape and dimensions of the micelles did not show any significant change over the concentration range studied: 20 to 60 mM ( Fig. 1A-C). All surfactants studied have a hydrocarbon tail containing twelve carbon atoms, thus the difference between them lies in the headgroups. CAPB and CAPOx micelles, within the error, have the same shape and size. The headgroups are of a similar nature, betaine for CAPB and an amine oxide for CAPOx. The main difference between these surfactants lies in the charge distribution in the headgroup, CAPOx has a N-O dative bond, which has higher polarity than a covalent bond but not a formal charge separation, making CAPOx a non-ionic surfactant. For CAPB, charge separation is present in the betaine headgroup, but the net charge is zero. However, the differences between them are small enough to not affect the morphology of the single surfactant micelles. SLS micelles are significantly smaller, having a shorter and more compact headgroup than the other surfactants. A more in-depth discussion of the pure surfactant phase diagrams is outside the scope of this work, but this topic has been studied in depth elsewhere. 53 The mixtures of surfactants, however, showed a wider range of micellar morphology. In Fig. 2, the SANS data and fittings are presented for the mixtures CAPB/SDS, CAPOx/SDS and CAPB/SLS as a function of the concentration. We kept an equimolar ratio while changing the overall concentrations to give solutions at 20/20, 40/40 and 60/60 mM surfactants. At first glance, the mixtures can be split into two groups depending on the cosurfactant, with either SDS ( Fig. 2A and B) or with SLS (Fig. 2C). For CAPB/SDS and CAPOx/SDS, the data were fitted using a combination of a semi-flexible cylinder 37 form factor and a random-phase approximation structure factor. 18,54,55 The fittings suggest long, rodlike, strongly interacting micelles for both CAPB/SDS and CAPOx/SDS mixtures. The length of the micelles, however, lies outside of the q-range accessible by the measurement and cannot be determined from the SANS data. It was arbitrarily fixed as 1000 Å. From the fittings (Table S2A and B, ESI †), we obtained the cross-sectional radius for the rodlike micelles, R = 18 Å, for both CAPB/SDS and CAPOx/SDS, and the n RPA coefficient, which provides insights on the magnitude and type of interactions (n RPA 4 0 repulsive, n RPA o 0 attractive). However, the magnitude of n RPA depends on the micelle length, which was arbitrarily fixed. Therefore, only relative changes within a given series will be discussed. At 20/20 mM, n RPA = 16.2 was obtained for CAPB/SDS and n RPA = 64.2 was obtained for CAPOx/SDS, showing that strong interactions are present in both cases but CAPOx/SDS micelles show apparently stronger repulsive interactions. As the actual value of n RPA will depend on the length of the micelle, the stronger repulsive interactions for CAPOx/SDS compared to CAPB/SDS at the same concentration can either be described by a stronger interaction or longer micelles, or even both. While the SANS data cannot provide information about the length of the micelles, the rheological behaviour of the mixtures can add further clarification. Systems formed by long flexible cylinders such as polymers and WLMs will form entangled networks above an overlap concentration. These entangled networks have a characteristic rheological profile. They follow the behaviour expected for a Maxwell fluid. 25,26 In Fig. 3, we present the frequency sweep for both CAPB/SDS and CAPOx/SDS at 60/60 mM in 1 wt% NaCl. All of the single surfactants and the mixture CAPB/SLS showed Newtonian behaviour, however the CAPB/SDS solution shows the characteristic behaviour of a Maxwell fluid, associated with wormlike micelles. 25,26 In this case, a two-element Maxwell model was used to satisfactorily fit the data. CAPOx/SDS solutions showed very weak viscoelasticity close to the operational limit of the rheometer, which could not be satisfactorily fitted with a Maxwell model. A likely explanation is that CAPOx/SDS micelles are short rods, not long enough to generate an entangled network at the concentration assayed. The high repulsion observed via SANS also suggests they are not flexible cylinders. One cause for the lack of flexibility is a highly charged surface, since repulsive electrostatic interactions would lead to a loss of flexibility. As discussed earlier, the amine oxide group from the CAPOx headgroup has a N-O dative bond, which has no formal charge separation, which could result in weaker electrostatic interactions with SDS than the betaine headgroup in CAPB, where formal charges are present. Even though CAPB and CAPOx have no net charge at the experimental pH (neutral), the local charge distribution of the CAPB headgroup should be more effective at reducing electrostatic repulsion than the CAPOx headgroup. For the CAPB/SLS mixture, as shown in Fig. 2C, the data were fitted with a combination of a sphere form factor 40 and a Hayter-Penfold MSA (charged spheres) structure factor. 41,42 The fittings resulted in spherical micelles of R = 24 AE 1 Å in radius, essentially identical to CAPB micelles except for the enhanced interactions, due to the charge contribution of SLS.
OCNF/surfactant systems
The effect of anionic surfactants on OCNF gelation has been previously studied by our group. 13,14 In this work, we focus on the contribution of CAPB and CAPOx, both alone and in mixtures with either SDS or SLS. In Fig. 4, the SANS curves for CAPB, CAPOx and SLS in 1 wt% OCNF/1 wt% NaCl are presented for three surfactant concentrations, 20, 40 and 60 mM. The surfactant contribution is strong in the SANS pattern dominating the overall signal, however the OCNF signal is still clearly present. The curves can be described by the addition of spherical micelles, using the previously established behaviour for CAPB, CAPOx and SLS (Fig. 1), plus the contribution from OCNF (without polydispersity for the cross-section), R maj = 53 AE 1 Å, fixed length = 1600 Å and e = 0.24 AE 0.02, as established in a previous study, 18 adjusting the scale factor and n RPA . These results show that the OCNFs and the surfactant micelles in these systems are not disturbed by the other's presence and that no strong interactions (either attractive or repulsive) between them are present at this length scale. The extracted n RPA parameter associated with OCNF is negative, À1.9, in all of the scattering patterns from OCNF/CAPB, OCNF/CAPOx and OCNF/SLS, which is in line with the expected attractive interactions of OCNFs in saline solutions. 18 The OCNF does not affect the morphology of the single surfactant micelles, nor is the OCNF network sensitive to the type of surfactant used. The rheological behaviour, however, is strongly affected by the presence of the surfactants (Fig. 5A-C). OCNF in saline solutions is known to form gels due to strong attractive interactions. 18 The addition of either CAPB or CAPOx leads to a significant reduction of G 0 and G 00 but without affecting the overall gel behaviour, with G 0 c G 00 for all concentrations of surfactant, suggesting that the micelles disrupt the OCNF network. SLS (Fig. 5C) has a smaller effect, increasing G 0 slightly at 20 mM, with a further concentration increase leading to a reduction of G 0 . A likely scenario is that the micelles populate the interstices of the OCNF network and disrupt the network connectivity by imposing a steric barrier to OCNF interactions.
The combination of OCNF and the surfactant mixtures (CAPB/SDS, CAPOx/SDS and CAPB/SLS), however, leads to more visible changes in the OCNF gel behaviour. In Fig. 6A-C, oscillatory frequency behaviour for the three mixtures as a function of the surfactant concentration is shown. Based on the previously discussed SANS and rheological data ( Fig. 2A-C and 3), WLMs, long rods and spherical micelles are formed in CAPB/SDS, CAPOx/SDS and CAPB/SLS mixtures, respectively. For the OCNF/CAPB/SDS mixtures (Fig. 6A), containing WLMs, we observe a significant drop in G 0 and G 00 when compared to OCNF gels (Fig. 5A). An inversion of the G 00 frequency dependence is also observed. In the presence of WLMs, an exponential increase, which decreases with the surfactant concentration, is observed, while without WLMs, G 00 values decayed exponentially with increasing frequency. Moreover, G 00 is also more sensitive to the surfactant concentration than G 0 . For both CAPOx/SDS and CAPB/SLS ( Fig. 6B and C), the impact of the micellar aggregates on the OCNF gel network is smaller than that with the WLMs (Fig. 6A) or single surfactant micelles ( Fig. 5A and B). WLMs, under the right conditions, form entangled networks, adding a new relaxation process, reptation, which is not present in the OCNF network. Based on the rheological behaviour of the CAPB/SDS solution (Fig. 3), one can assume that the entangled network has a shorter relaxation time than the OCNF network ( Fig. 5 -black line). OCNF forms physical gels, and their frequency sweeps show a weak frequency dependency that translates to finite but very long relaxation times for the OCNF network. Meanwhile, as can be seen from Fig. 3, CAPB/SDS WLMs have a G 0 , G 00 cross-over at 1 rad s À1 , therefore a relaxation time of ca. 150 ms. Hence, in OCNF/CAPB/SDS hydrogels, the WLM entangled network is capable of undergoing mechanical relaxation in a time frame much shorter than that of the OCNF network, adding extra dissipative contributions, in the form of reptation and breaking/reforming micellar dynamics, to the system. This leads to the observed increase of G 00 in addition to the drop in G 0 due to the steric constraints imposed on the OCNF network by the micelles. Only a small drop in G 0 is observed for CAPOx/SDS (rodlike micelles) in the presence of OCNF and a slight increase is observed for CAPB/SLS (spherical micelles). G 00 is more sensitive to the surfactant presence, as, in both cases, an increase in the frequency dependence of G 00 is observed. Curiously, when comparing spherical micelles of the single surfactant and 56 suggests that synergism between CAPB and anionic surfactants results in lower free surfactant concentrations in their mixtures than alone, so this is unlikely to be due to free surfactant binding to the OCNFs and changing inter-fibrillar interactions. Another visible change in OCNF gel behaviour in the presence of WLMs can be seen in the deformation behaviour. In Fig. 7A-C, oscillatory strain sweeps are presented for OCNF dispersions and OCNF/surfactant mixtures. In the nonlinear region of the strain sweep, G 00 shows strain stiffening, i.e., an increase in magnitude of G 00 or overshooting 57 in OCNF gels ( Fig. 7 -black line), which is suppressed when WLMs are present (Fig. 7A), but the gels are unaffected by rodlike or spherical micelles ( Fig. 7B and C), apart from a decrease of the strain value at which the stiffening of the gel occurs. This strain stiffening, during strain deformation, reflects an additional resistance to the applied strain from the system. 58 For the OCNF case, or generally speaking, for charged anisotropic particles, the applied deformation leads to an increase in the orientation of the particles; below the linear viscoelastic range (LVR), this shear induced orientation is within the normal range of movement of the particles at rest. Past the LVR, the shear induced orientation is hindered by the repulsive interactions of the charged OCNFs, leading to the strain stiffening. The WLMs could be suppressing the strain stiffening in two ways. The added relaxation through reptation could help to offset the stiffening as observed in interpenetrating networks, 59 or the physical presence of the WLMs could affect the OCNF shear orientation. We note that both systems are likely to be similar in terms of changes in ionic strength arising from the surfactant counterions; as CAPB/SDS and CAPB/SLS are sodium salts, they do not appear to play a role in suppressing the strain stiffening. However, rheological data do not provide molecular insights into these systems and since the SANS data were collected in a quiescent state, it is therefore not possible to provide an unambiguous clear pictorial view of the actual mechanism or mechanisms taking place. The SANS data for OCNF/surfactant mixtures, however, show that the OCNF and micelle contributions to the scattering patterns are still additive. In Fig. 8A and B, SANS curves for OCNF/ CAPB/SDS and OCNF/CAPOx/SDS systems are presented. In both cases, the curves can be fitted using the previous surfactant mixture scattering patterns ( Fig. 2A and B) and the previously established OCNF profile. 18 To clarify the effect of the WLMs on the OCNF network in these mixtures, SANS data were also collected in 15/85 wt% D 2 O/H 2 O mixtures (Fig. 8C). At this D 2 O/ H 2 O ratio, the scattering length density of the surfactant is matched to that of the solvent and the surfactant is no longer visible. Therefore, the SANS signal originates only, or in majority, from the OCNFs. The OCNF scattering is not affected by the concentration of the surfactant mixtures, from 20/20 to 60/60 mM, and it shows a weaker attractive interaction than expected at 1 wt% NaCl, n RPA = À1.1 vs. n RPA ca. À1.8, respectively. 18 This suggests that the WLMs, once in place, slightly reduce the attractive forces between the OCNFs, but do not change the OCNF dispersion behaviour, which would be in line with the WLMs just creating an excluded-volume barrier to the OCNF-OCNF interactions without specific WLM-OCNF interactions.
Conclusions
In this work, we studied the effect of micellar morphology on the gel behaviour of OCNFs under saline conditions. We observed that both micellar morphology and the presence of single surfactants or surfactant mixtures can have an impact on OCNF gel behaviour. Small-angle neutron scattering was used to probe the structure of the micellar aggregates within the OCNF network, showing both that the micelles are unperturbed by the presence of the OCNF network and that they are not interacting with the fibrils at the length scale probed in SANS. This shows that the rheological modulation provided by the micelles is likely due to excluded-volume effects rather than molecular interactions. Wormlike micelles (CAPB/SDS mixtures) had the largest impact, both reducing G 0 and G 00 as well as changing the moduli frequency dependence, making the gel less stiff and more plastic. WLMs also suppressed strain stiffening normally observed in OCNF gels at large oscillatory strains. For spherical micelles, the single surfactants studied (CAPB and CAPOx) lowered both G 0 and G 00 , while the mixtures CAPB/SLS and CAPOx/ SDS lead to small variations in G 0 and a large effect on G 00 frequency dependency. These results suggest that unconnected spherical and rodlike micelles present within the OCNF network only add to the dissipative deformation (increase in G 00 ) without major impacts on the structure of the OCNF network (G 0 ), while the WLMs affect the connectivity of the OCNF network, leading to the drop of G 0 and G 00 observed. In summary, small, unconnected micellar aggregates occupy space within the OCNF network, weakening it, likely by reducing the connectivity of the fibril network. For long, interconnected micellar aggregates, in our case WLMs, the system behaves as a sum of the fibril and entangled network contributions, both reducing the OCNF network connectivity and adding reptation relaxation to the hybrid hydrogel.
Conflicts of interest
There are no conflicts to declare. | 7,124.6 | 2020-05-19T00:00:00.000 | [
"Materials Science"
] |
Semi-tightly coupled integration of multi-GNSS PPP and S-VINS for precise positioning in GNSS-challenged environments
Because of its high-precision, low-cost and easy-operation, Precise Point Positioning (PPP) becomes a potential and attractive positioning technique that can be applied to self-driving cars and drones. However, the reliability and availability of PPP will be significantly degraded in the extremely difficult conditions where Global Navigation Satellite System (GNSS) signals are blocked frequently. Inertial Navigation System (INS) has been integrated with GNSS to ameliorate such situations in the last decades. Recently, the Visual-Inertial Navigation Systems (VINS) with favorable complementary characteristics is demonstrated to realize a more stable and accurate local position estimation than the INS-only. Nevertheless, the system still must rely on the global positions to eliminate the accumulated errors. In this contribution, we present a semi-tight coupling framework of multi-GNSS PPP and Stereo VINS (S-VINS), which achieves the bidirectional location transfer and sharing in two separate navigation systems. In our approach, the local positions, produced by S-VINS are integrated with multi-GNSS PPP through a graph-optimization based method. Furthermore, the accurate forecast positions with S-VINS are fed back to assist PPP in GNSS-challenged environments. The statistical analysis of a GNSS outage simulation test shows that the S-VINS mode can effectively suppress the degradation of positioning accuracy compared with the INS-only mode. We also carried out a vehicle-borne experiment collecting multi-sensor data in a GNSS-challenged environment. For the complex driving environment, the PPP positioning capability is significantly improved with the aiding of S-VINS. The 3D positioning accuracy is improved by 49.0% for Global Positioning System (GPS), 40.3% for GPS + GLOANSS (Global Navigation Satellite System), 45.6% for GPS + BDS (BeiDou navigation satellite System), and 51.2% for GPS + GLONASS + BDS. On this basis, the solution with the semi-tight coupling scheme of multi-GNSS PPP/S-VINS achieves the improvements of 41.8–60.6% in 3D positioning accuracy compared with the multi-GNSS PPP/INS solutions.
Introduction
Precise Point Positioning (PPP) has been demonstrated as an effective tool in high-precision positioning and shows the advantages of efficiency and flexibility compared to the baseline network approach (Zumberge et al. 1997;Bisnath and Gao 2009). In recent years, the rapid development of Chinese BeiDou navigation satellite System (BDS) and European Galileo navigation satellite system (Galileo) brings new opportunities for PPP. A four-system PPP model was proposed by Li et al. (2015) to fully use the Global Positioning System (GPS), Global Navigation Satellite System (GLONASS), Galileo, and BDS observations. In their study, the multi-constellation Global Navigation Satellite System (GNSS) PPP presented faster solution convergence and higher positioning accuracy than single-system PPP. Recently, the investigation of multi-GNSS PPP data processing is not only about the dual-frequency models (Cai et al. 2015), but also focusing on the multi-frequency observations (Li et al. 2019b(Li et al. , 2020a. Briefly, the multi-frequency and multi-GNSS based PPP is becoming increasingly fashionable for precise positioning services (Alkan and Öcalan 2013;Guo et al. 2018), particularly in some new applications such as self-driving cars and unmanned aerial vehicles (Nie et al. 2019;Geng and Guo 2020).
However, PPP fails in the cases of observation outages or harsh signal environments (Zhang and Li 2012). Consequently, the Inertial Navigation System (INS) has been utilized to assist PPP in GNSS-challenged environments in the last decades (Roesler and Martell 2009;Gao et al. 2017). Shin and Scherzinger (2009) demonstrated that PPP/INS integration could realize a better accuracy and reliability of positioning in both open sky and GNSS blocked areas. Rabbou and El-Rabbany (2015) presented a tightly coupled multi-GNSS PPP/INS solution and achieved the positioning accuracy at decimeter to centimeter-level when the measurement updates from GNSS are available. Nevertheless, the performance of the GNSS/INS integration is degraded due to the rapid INS drift errors for the case of the long-term GNSS outages.
Favorable complementary properties of visual and inertial measurements make them suitable for fusion. Thus, extensive applications based on a visual-inertial integration were found in drones (Weiss et al. 2012) and selfdriving vehicles (Li and Mourikis 2012). Generally, the existing visual-inertial fusion methods can be classified into the optimization-based (Yang and Shen 2017;Usenko et al. 2016) and the filter-based approaches (Bloesch et al. 2015;Tsotsos et al. 2015). A popular filter-based Visual-Inertial Odometry (VIO) algorithm was proposed by Mourikis and Roumeliotis (2007). In their approach, a versatile measurement model was presented to express the geometric constrains among multiple-camera poses with a common view. In practice, the optimizationbased approaches can provide higher accuracy than the filter-based approaches given adequate computational resources (Delmerico and Scaramuzza 2018). The property of the re-linearization at each iteration contributes to the high accuracy of the optimization-based methods. Leutenegger et al. (2015) presented a keyframe-based Visual-Inertial Navigation System (VINS) and used Google's Ceres solver to perform the nonlinear optimization (Agarwal et al. 2012). Besides, the sliding window strategy was adopted in their study to reduce the computation complexity of optimization. Qin et al. (2018) proposed a complete and versatile monocular VINS, which can realize the indoor positioning of drones with accuracy at a decimeter-level. Additionally, the translation error of the stereo VINS (S-VINS) is about 1% of the driving distance in an outdoor vehicular experiment (Qin et al. 2019). Although VINS can achieve a robust and accurate local pose estimation, the errors still accumulate over the time.
To eliminate the accumulated errors of VINS, many researchers integrate the GNSS and VINS for realizing a local accurate and global drift-free localization. Lynen et al. (2013) proposed a basic multi-sensor fusion framework to process delayed, relative, and absolute measurements from different sensors. Mascaro et al. (2018) proposed a decoupled optimization-based multi-sensor fusion method, which is demonstrated to be more accurate than other decoupled fusion strategies. Although some progress has been made with these methods in multi-sensor fusion navigation, they adopt the decoupled way to integrate the GPS and VINS. In addition, only the GPS derived positions are utilized in their framework rather than the GNSS raw observations with more available information. Vu et al. (2012) developed a multisensor fusion framework with differential GPS (DGPS), vision, and INS, which can provide a lane-level vehicle navigation in GNSS open-sky conditions. Moreover, Li et al. (2019a) proposed a tightly coupled fusion solution of multi-GNSS Real-Time Kinematic (RTK)/INS/vision, which can achieve centimeter-level positioning accuracy in GNSS degraded conditions. In the above two studies, the relative positioning methods were used to provide the global locations, which requires additional GNSS infrastructures such as reference stations and receivers in comparison to PPP. Zhu (2019) proposed a new structure named Semi-Tightly Coupled (STC) integration, which realized multi-sensor information fusion by the bidirectional location transfer and sharing in two separate navigation systems. The STC not only combines the advantages of the Loosely Coupled (LC) integration and Tightly Coupled (TC) integration, but also overcomes their main deficiencies.
In this contribution, we present a graph-optimization based and semi-tight coupling framework of multi-GNSS PPP and S-VINS for improving the PPP performance in a GNSS-challenged environment and realizing a stable and accurate global positioning outputs in a complex driving environment. In addition to a GNSS outage simulation test to verify the positioning capacity of S-VINS, the vehicle-borne experiment was also carried out in the campus of Wuhan University to assess the positioning performances of the S-VINS aided PPP solution and the multi-GNSS PPP/S-VINS solution. The contribution of the proposed method to precise positioning is presented and analyzed. In the following parts of this paper, we first describe the methods used in this study and then explain the algorithm implementation for the triple integrated system. Subsequently, the experimental situation is introduced, and the results are analyzed. Finally, the conclusions are summarized.
Methodology
In this section, we firstly introduce the PPP observation model. Then, a tightly coupled stereo VIO algorithm is described. Subsequently, the semi-tightly coupled multi-GNSS PPP/S-VINS fusion method is presented. Finally, the algorithm implementation of the developed multisensor fusion framework is explained.
PPP observation model
The GNSS observation equations for raw pseudorange and carrier phase are formulated as (Li et al. 2015): where the symbols s , r , and j represent the satellite, receiver and carrier frequency, respectively; ρ is the geometric distance between the satellite and receiver; c is the speed of light in vacuum; t r and t s denote the receiver and satellite clock offsets, respectively; d r,j and d s j are the code hardware delays for the receiver and the satellite, respectively; I s r,1 is the ionospheric delay at the first carrier frequency, and µ j = f 2 j f 2 1 is the ionospheric coefficient associated to a frequency f j ; T s r is the tropospheric delay; j and N s r,j denote the wavelength and the integer ambiguity; b r,j and b s j are the phase delays in receiver and satellite sides (Ge et al. 2008;Li et al. 2011); �ρ denotes the other corrections which should be considered in the PPP model, such as phase wind-up effect, antenna Phase Center Offset (PCO) and Phase Center Variation (PCV), relativity effect, and earth rotation effect (Wu et al. 1993;Schmid et al. 2007); ε P s r,j and ε L s r,j represent the sum of measurement noises and multipath errors for code and phase, respectively.
The Ionospheric-Free (IF) combinations are usually applied to eliminate the ionospheric delay in the PPP model. The dual-frequency IF combinations can be written as: where γ = f 2 1 (f 2 1 − f 2 2 ) , f 1 and f 2 are the frequencies of two carriers; IF N IF = γ ( 1 (N s r,1 + b r,1 − b s 1 )) + (1 − γ ) ( 2 (N s r,2 + b r,2 − b s 2 )) is the IF ambiguity in meters. The measurements noises of IF pseudorange and phase can (1) be denoted by ε P IF = γ ε P s r,1 + (1 − γ )ε P s r,2 and ε L IF = γ ε L s r,1 + (1 − γ )ε L s r,2 . Additionally, d r,j is absorbed in receiver clock offset, and d s j is corrected in the IF combinations when applying the precise clock products. The tropospheric delay T in Eqs. (1) and (2) is made up of the dry and wet components which can be expressed by the zenith delays ( T dry , T wet ) and the corresponding mapping functions ( m dry , m wet ). An empirical model can be used to correct the dry delay part ( m dry · T dry ) (Saastamoinen 1972), while the wet component delay ( m wet · T wet ) can be estimated from the observations. When multi-GNSS observations are involved, the different signal structures and different hardware delays for each GNSS system will result in different code biases in one multi-GNSS receiver (Li et al. 2015). The differences between these biases are usually called Inter-System Biases (ISB) or Inter-Frequency Biases (IFB) for GLONASS satellites. ISB/IFB parameters must be introduced into the multi-GNSS estimator. The IF combinations of the multi-GNSS code and phase observations can be written as: where t r denotes the receiver clock offset of the reference GNSS system, namely GPS; ISB sys represents the ISB of the non-reference GNSS system. As for GLONASS, the ISB sys parameter will be set for each frequency. ρ T represents the sum of the geometric distance and the dry tropospheric delay. The linearized equations of the IF combination can be expressed as: where p IF and l IF signify observed-minus-computed pseudorange and phase IF measurement residuals; u represents the unit vector of the direction from the receiver to the satellite; δp is the position correction vector. In this paper, the GNSS raw measurements are processed by the individual multi-GNSS PPP module of the multi-sensor fusion system. The detailed information on the multi-GNSS data processing in PPP is listed in Table 1.
Stereo visual-inertial odometry formulation
The visual front-end processes the stereo pairs from the stereo camera. For each new stereo pair, the Kanade-Lucas-Tomasi (KLT) sparse optical flow algorithm is applied to perform feature tracking of existing features (Lucas and Kanade 1981). In addition, the forward (5) (previous frame to current frame) and backward (current frame to previous frame) feature tracking are both implemented to acquire high quality tracking results. Meanwhile, new corner features are detected to maintain a certain number of features (e.g., 100-300) in each image (Shi and Tomasi 1994). The stereo matches are also obtained by the KLT sparse optical flow algorithm between left and right images. As for the raw Inertial Measurement Unit (IMU) measurements, the IMU preintegration technique is used to generate relative IMU measurements between two consecutive states in VIO sliding window (Lupton and Sukkarieh 2012). For the IMU state propagation in pre-integration, the mid-point integration is used for the discrete-time implementation.
To propagate the uncertainty of the state, the covariance of the IMU state can be computed recursively, referring to Qin et al. (2018). An initialization procedure is required for the stereo VIO. For each frame in the sliding window, we triangulate all features observed in the stereo pairs. Based on these triangulated features, a Perspective-n-Point (PnP) method is used to estimate the poses of all other frames in the window (Lepetit et al. 2009). Additionally, the pre-integration factor is constructed between each frame in the window. When the window size reaches 10, a visual-inertial bundle adjustment is performed to obtain the optimized states in the window.
After the initialization of estimator, a tightly coupled sliding window-based VIO is carried out to achieve accurate and robust state estimation, serving as local constraints in the global fusion. The definition of state vector in the sliding window can be written as (Qin et al. 2018): where χ l vio denotes the complete state vector including the IMU state vector x k , the extrinsic parameter x b c of IMU-camera, and the inverse depth l of the l th feature from its first observation; c and b represent the camera frame and IMU frame, respectively. n and m are the quantities of keyframes and features in the sliding window, respectively; the x k consists of the IMU states at the time when the k th image is captured; the position p l vio b k , velocity v l vio b k , and orientation q l vio b k of the IMU center is with respect to the local reference frame l vio which is defined by the first IMU pose; b a and b g represent the accelerometer bias and gyroscope bias, respectively.
A maximum posteriori estimation of the VIO system states can be acquired by minimizing the sum of a priori and the Mahalanobis norm of all measurement residuals: , χ) and r C (ẑ c j l , χ ) denote the inertial and visual residuals, respectively; r p − H p χ represents the a priori information obtained by the process of marginalization in the sliding window; ρ(·) is the Huber function used for reducing the weight of the outliers in the least squares problems (Huber 1964). In addition, a strict outlier rejection mechanism is performed after each optimization by checking the average reprojection errors of (13), (14), and (15). When the window size is full, the oldest IMU state and corresponding features in the sliding window will be marginalized to bound the computational complexity of VIO.
There are two additional types of reprojection equations for the stereo VIO compared to the mono-VIO presented in Qin et al. (2018). Supposed that the l th feature is observed by the i th stereo images and the j th stereo images. Additionally, the first observation of the feature happens in the former. Three types of reprojection equations are used in our method, which can be expressed as: l ] is the first observation of the lth feature, and c i,1 denotes the left image of the ith stereo images; π −1 c is the back projection function which turns a pixel location into a unit vector using camera intrinsic parameters; R b c 1 , p b c 1 and R b c 2 , p b c 2 are the extrinsic parameters of left IMU-camera and right IMU-camera, respectively; P
Multi-GNSS PPP/S-VINS fusion
The multi-sensor fusion problem is depicted by constructing a graph structure displayed in Fig. 1. The graph structure consists of a series of nodes and edges. Each node denotes the vehicle state in the global frame.
The edge between two consecutive nodes is a local constraint formed by S-VINS. Another type of edge is the global constraint provided by the multi-GNSS PPP solution. Because of the low satellite availability in complex driving conditions, the positioning results from the PPP are selectively used as the global constraint. A Quality Number (QN) is adopted to indicate the accuracy of PPP solution, referring to (NovAtel Corporation 2018a). The quality of the positioning results from PPP solution are labeled with an integer 1-6 based on their covariances. In this paper, the QN within 4 will be maintained in the pose graph; the QN equal to 5 will be used only once and removed after the global optimization; and the QN more than 5 will be rejected. The growth rate of the node is dependent on the GNSS outputs. The mathematical model of the fusion method can be expressed as a Maximum Likelihood Estimation (MLE) problem as described in Qin et al. (2019). For the completeness, we briefly introduce the theory. The state estimation of the global fusion can be converted to a nonlinear least squares problem, which can be written as: where χ= [x 0 , x 1 , . . . Fig. 1 The graph structure for global fusion represents the vector of the measurement residual, and is the corresponding covariance. The error function r = z k t − h k t (χ ) consists of two parts in the fusion model. Part one is the local measurement residual, which is formulated as: the upper equation describes the relative pose error between time t − 1 and t . The first row denotes the relative position errors, and the second row denotes the relative rotation error. ⊖ is the minus operation on the error state of quaternion. The unified covariance is applied for all local measurements in our framework. Part two is the global measurement residual, which can be written as: where p ppp t is the position measurement from the multi-GNSS PPP. The global location is directly used as the position constraint for every node. It should be noted that the local-level frame (ENU, East-North-Up) is adopted to represent the global reference frame G , and the origin point is located at the first global location from the multi-GNSS PPP solution during the global fusion. Furthermore, the subsequent global positioning results are converted from the Earth-Centered Earth-Fixed (ECEF) frame to the ENU frame with respect to the first global location. The proposed triple integrated system can provide the covariances of the global locations, which contributes to a better use of the position information from the GNSS. In comparison, the original work in Qin et al. (2019) determines the covariance only by the number of the visible satellites.
The nature of the fusion method is a rigid base frame alignment problem between a local reference frame and a global reference frame. The multi-sensor-fusion positioning in the global frame can be realized by carrying out this alignment process. The transformation between the local and the global reference frame will be updated after each global optimization. The subsequent positioning results from S-VINS can be converted from the local frame to the global frame by this transformation. Moreover, the predicted positions maintain a high accuracy in a short term, which can be utilized in the multi-GNSS PPP data processing. Thus, we transmit the global forecast position to the multi-GNSS PPP processor as the a priori information.
The a priori information is used for the following purposes in the multi-GNSS PPP processing. Firstly, the (17) predicted position is used as an initial value for the PPP data processing to replace the Standard Point Positioning (SPP) result. On the one hand, SPP produces a position with low accuracy in GNSS-challenged conditions (Angrisano et al. 2013). On the other hand, the priori location has a comparable positioning accuracy with PPP in a short term, which is verified in the following experimental part. Secondly, when the number of available satellites is less than six, the forecast position will be used as the position constraint in the PPP processing. This criterion is used mainly to cope with the extremely poor observation conditions. The variance of the predicted position can be determined by: where σ 0 = σ 2 x 0 + σ 2 y 0 + σ 2 z 0 is the standard deviation of the global location used in last graph optimization; (σ 2 x 0 , σ 2 y 0 , σ 2 z 0 ) is the variances of position in ECEF; D denotes the diving distance from the vehicle state of last graph optimization to current vehicle state in meters; 1% is the degradation rate of the local positioning accuracy (Qin et al. 2019). The unified variance σ 2 is applied for different axes of the position vector p = [p e x , p e y , p e z ] in our algorithm for the degradation rate of the local positioning accuracy is hard to be decomposed to different axes. The position feedback mechanism in our solution is bootable when the number of the global locations maintained in the global fusion processor exceeds a certain threshold.
Implementation of multi-GNSS PPP/S-VINS algorithm
The architecture of the proposed semi-tightly coupled multi-GNSS PPP/S-VINS integration is shown in Fig. 2. A sliding window-based nonlinear optimization is operated for state updates after finishing the visual-inertial initialization. The newest local state is converted to the corresponding global state by the transformation between the local frame and the global frame. In addition, the transformation matrix is initially set to the identity matrix and gets updated after each global optimization. The IF combinations of GNSS raw pseudorange and phase measurements are applied to the PPP data processing. Once the feedback mechanism is activated, the predicted positions from S-VINS can be utilized in the PPP processing. When the PPP solution is completed, the global position with its uncertainty will be transferred to the global fusion processor. Nevertheless, only the positioning result that passes the quality check will be used in the global fusion. Practically, the measurements from the local (S-VINS) and global (PPP) processor have different sampling rates. If the timestamp of the newest local state is synchronized with the current GNSS epoch, the global optimization will proceed. Otherwise, only the forecast positions can be acquired. The Ceres Solver is used in the triple integrated system for state optimization (Agarwal et al. 2012). The optimal positioning results are obtained after carrying out the global graph optimization. Meanwhile, the transformation from the local frame to the global frame is updated.
Experimental description
The vehicular road test was carried out in the campus of Wuhan University where the trees with dense forest canopies are on both sides of the roads. Figure 3 displays the top view of the trajectory and the typical surroundings in the road. The total distance of the trajectory is about 2670 m, and it takes about 12 min in our experiment.
The equipment used for collecting the multi-sensor fusion data is displayed in Fig. 4. In our vehicular road test, only single GNSS antenna was used. As shown in the top panel of Fig. 4, a GNSS receiver and an IMU device are connected to the GNSS antenna through a signal power divider. Two cameras are tightly mounted on the front of the platform with a 505 mm baseline. The detailed information on the devices is listed in Table 2, and the specification of the IMU sensor is provided in Table 3. We achieved the time synchronization at the hardware level. More specifically, the Pulses Per Second (PPS) generated by the GNSS receiver is used to trigger the IMU to work and the stereo cameras to exposure at different frequencies. By this means, the time stamps of different sensors will synchronize to GPS time. The offset between GNSS antenna reference point and IMU center was measured precisely to compensate the lever-arm effect. The extrinsic parameters for stereo cameras and IMU-camera were calibrated offline (Furgale et al. 2013). Moreover, the extrinsic parameters of IMU-camera are also estimated in S-VINS based on the pre-calibrated values to compensate the small variations caused by the vehicle motion. We also calibrated the intrinsic parameters of the stereo cameras before and after the test.
Additionally, the multi-sensor fusion data were collected under the normal driving conditions including most common ground vehicle dynamics, such as acceleration, deceleration, and cornering. The bidirectional
Result analysis
In this part, the number of available GNSS satellites and the corresponding position Dilution of Precision (PDOP) are firstly presented. Then, a simulation test of complete GNSS outage is proceeded to validate the positioning capacity of S-VINS. Subsequently, the positioning capacity of the S-VINS aided multi-GNSS PPP solution is discussed. Finally, we assess the positioning performance of the multi-GNSS PPP/S-VINS solution.
Satellite availability
The top panel of Fig. 5 depicts the evolutions of the number of available satellites for GPS (G), GLONASS (R), BDS (C), and GPS + GLONASS + BDS (G + R+C) during the test at a cutoff elevation angle of 7 • . The Galileo system is absent for only the single-frequency signals of Galileo can be received by our receiver during the test.
The mean values of the number of visible satellites for different GNSS constellations are 4.8 (G), 3.2 (R), 4.1 (C), and 12.1 (G + R + C). There are frequent decreases in the satellite numbers as shown in Fig. 5, and the number of available GLONASS or BDS satellites sometimes becomes zero. The PDOP variations of different GNSS constellations are presented in the bottom of Fig. 5. The average PDOP values for different GNSS constellations are 3.1 (G), 4.7 (R),4.3 (C), and 1.2 (G + R + C). It is obvious that the value of the PDOP increases as the number of observed satellites decreases. On account of such GNSS partly blocked conditions, the number of observed satellites drops frequently, and the signal tracking is discontinued, which is a challenge to precise positioning.
S-VINS positioning performance during GNSS outage
In this section, we simulated the complete GNSS outage conditions to investigate the positioning performance of S-VINS compared with the INS-only solution. A complete dynamic trajectory (about 2670 m) in the real driving environment was divided into ten segments with the driving time of 100 s each. Meanwhile, the complete GNSS outage for 50 s was simulated in each segment. The average root mean square (RMS) values of the position drifts for the two solutions are shown in Fig. 6. During the GNSS outage time from 5 s to 50 s, the position RMS values of the INS mode are degraded from 0.05, 0.02, and 0.01 m to 3.12, 3.04, and 0.15 m in the east, north, and vertical directions, respectively. By contrast, the RMS values of S-VINS drop from 0.05, 0.06, and 0.01 m to 0.80, 1.16, and 0.12 m in east, north, and vertical directions, respectively. It can be seen that the S-VINS mode has a slower degradation in positioning accuracy than the INS-only mode. This indicates that redundant visual observation from the tracked features can help the S-VINS maintain an accurate local position. As described above, the GNSS is in normal operation in the remaining 50 s of each segment. Thus, the accumulated positioning errors of S-VINS in the triple integrated system can be corrected after each global optimization. To have a comprehensive assessment of the positioning performance of S-VINS, the predicted position accuracy of S-VINS before each global optimization is calculated. The distribution of the predicted position differences is shown in Fig. 7. The results show that the percentage of position differences less than 5 cm is 71.9%, 63.8%, and 98.5% for east, north, and up components, respectively. The corresponding percentage is 22.6%, 33.2%, and 0.5% in the range of 5 cm to 10 cm. Given the above, it can be found that more than 90% of the predicted position differences are at centimeter level when GNSS is in normal operation, despite of the outliers caused by the visual instability resulting from the feature mismatches in the complex driving conditions. Additionally, the S-VINS-only positioning performance is also evaluated in the same dynamic driving environment. We aligned the S-VINS trajectory (local coordinate) with the ground truth (ECEF coordinate) using a rigid body transformation (Horn 1987) and calculated the position differences of each matched positions. The RMS of position differences of S-VINS in the local coordinate system is given in Table 5.
Positioning capacity of the S-VINS aided multi-GNSS PPP solution
In our triple integrated system, the forecast position from S-VINS is used as an initial value or a position constraint to assist multi-GNSS PPP in GNSS-challenged conditions. The position differences of the IF PPP solution for the GPS, GPS + GLONASS, GPS + BDS, and GPS + BDS + GLONASS modes are shown in Fig. 8. The corresponding position differences for the S-VINS aided IF PPP solution are shown in Fig. 9. The results of both modes are listed in Table 4 still largely impacted by the PPP performance due to the location-based information fusion. In addition, the major improvement of PPP is in vertical component while the horizontal components obtain a modest improvement with the aiding of S-VINS. In conclusion, the multi-GNSS PPP/S-VINS solution achieves a higher positioning accuracy and availability compared with multi-GNSS/ INS solutions in such GNSS-challenged environment.
Conclusion
To improve the positioning performance in GNSSchallenged environments, an optimization-based semi-tightly coupled multi-sensor fusion framework of multi-GNSS PPP/S-VINS was developed and validated in this study. Based on the GNSS outage simulation test and the vehicle-borne experiment, the positioning performances of the multi-GNSS PPP/S-VINS solution were comprehensively evaluated with respect to the stand-alone S-VINS positioning, the S-VINS aided multi-GNSS PPP positioning, and the triple integrated system positioning.
The GNSS outage simulation test demonstrates that the S-VINS can achieve a slower degradation in positioning accuracy than the INS-only. The statistical analysis of the complete GNSS outages for 50 s shows that the average RMS of position drifts for S-VINS is 0.80, 1.16, and 0.12 m with an improvement of 74.4%, 61.8%, 20.0% in north, east, and up components, respectively, compared with the INSonly mode. Furthermore, more than 90% of the predicted position differences is at centimeter level during one-second GNSS outages. According to the results of the vehicle-borne experiment, the accurate predicted positions from S-VINS can assist PPP to improve the overall positioning performance. The maximum position error of the stand-alone PPP (GPS + GLONASS + BDS) solution is reduced from (5.36, − 19.14, − 44.99) m to (2.00, − 2.78, − 3.13) m compared with the results of the aiding of S-VINS in east, north, and up components, respectively. Besides, the improvements of 3D positioning accuracy for the unaided PPP solution are 49.0% for GPS, 40.3% for GPS + GLONASS, 45.6% for GPS + BDS, and 51.2% for GPS + GLONASS + BDS. Due to the improvement in the positioning accuracy of the S-VINS aided PPP solution, better positioning results can participate in the graph optimization for global fusion. The statistics shows that that the RMSs of position errors of the multi-GNSS PPP/S-VINS solution are 0.88, 1.47, and 0.96 m with an improvement of 7.4%, 6.4%, and 27.3% in east, north, and up components, respectively, compared with the S-VINS aided PPP (GPS + GLO-NASS + BDS) solution. Moreover, the multi-GNSS PPP/S-VINS solution improves 3D positioning accuracy by 60.6% and 41.8% compared with the LC multi-GNSS PPP/INS solution and the TC multi-GNSS PPP/ INS solution, respectively.
In conclusion, the positioning performance of the PPP solution can be significantly improved with the aiding of S-VINS. Meanwhile, the multi-GNSS PPP/S-VINS solution realizes a higher positioning accuracy and availability compared with the multi-GNSS PPP/ INS solutions in GNSS-challenged environments, which shows a great potential of the multi-sensor fusion system for precise positioning. | 7,530.6 | 2021-01-04T00:00:00.000 | [
"Computer Science"
] |
New data‐driven method for estimation of net ecosystem carbon exchange at meteorological stations effectively increases the global carbon flux data
The eddy covariance (EC) flux stations have great limitations in the evaluation of the global net ecosystem carbon exchange (NEE) and in the uncertainty reduction due to their sparse and uneven distribution and spatial representation. If the EC stations are linked with widely distributed meteorological stations using machine learning (ML) and remote sensing, it will play a big role in effectively improving the accuracy of the global NEE assessment and reducing uncertainty. In this study, we developed a framework for estimating NEE at meteorological stations. We first optimized the hyperparameters and input variables of the ML model based on the optimization method called an adaptive genetic algorithm. Then, we developed 566 random forest (RF)‐based NEE estimation models by the strategy of spatial leave‐out‐one cross‐validation. We innovatively established the Euclidean distance‐based accuracy projection algorithm of the R square (R2), which could test the accuracy of each model to estimate the NEE of the specific flux at the weather station. Only the model with the highest R2 was selected from the models with a prediction accuracy of R2 > 0.5 for the specific meteorological stations to estimate its NEE. 4674 out of 10,289 weather stations around the world might match at least one of the 566 NEE estimation models with a projected accuracy of R2 > 0.5. The NEE estimation models we screened for the meteorological stations showed a reliable performance and a higher accuracy than the former studies. The NEE values of the most (96.9%) screened meteorological stations around the world are negative (carbon sink) and most (65.3%) of those showed an increasing trend in the mean annual NEE (carbon sink). The NEE dataset produced at the meteorological stations could be used as a supplement to the EC observations and quasi‐observation data to assess the NEE products of the global grid. The NEE dataset is publicly available via the figshare with https://doi.org/10.6084/m9.figshare.20485563.v1.
| INTRODUC TI ON
The net ecosystem carbon exchange (NEE) is the key carbon flux component within terrestrial ecosystems and plays an essential role in a better understanding of the global carbon cycle and landatmosphere interaction (Shiri et al., 2022). Accurate estimation and validation of NEE of the terrestrial ecosystems in regions or globally are of great significance in evaluating the function of the regional carbon source and sink. However, the NEE estimation has several common issues, such as the poorer modelled performance than the gross primary production (GPP). The main reason for this difference is that NEE is associated with both GPP and ecosystem respiration (RE) and remote sensing data cannot readily capture the parameters related to RE (Ichii et al., 2017;Tramontana et al., 2016).
The eddy covariance (EC) flux measurements have been providing detailed time series of the carbon fluxes, energy fluxes and atmospheric conditions across a large range of biomes and climate types. However, the EC measurements are site-scale observations which only represent the fluxes from the tower footprint up to several square kilometres (Gockede et al., 2008). Moreover, the observational constraints, such as a low number and unevenly distributed observation stations, limit the validation and extrapolation of the NEE estimation at large scales (Jung et al., 2009).
The ongoing efforts of the FLUXNET community and continuous improvement of the spatiotemporal resolution of remote sensing data have encouraged the application of the data-driven machine learning (ML) method such as the random forest (RF, Shiri et al., 2022), artificial neural networks (ANNs, Evrendilek, 2014), support vector regression (SVR, Ichii et al., 2017), cubist (Xiao et al., 2008;Xiao et al., 2011) or model trees ensemble (MTE, Jung et al., 2009 to estimate the terrestrial ecosystems' carbon dioxide, water and energy fluxes from a site scale to the regional and global scale (Xiao et al., 2019). The accuracy of the ML model is generally better than linear regression, ecosystem model, remote sensing inversion and other model methods, which has been proved in the application research of related geosciences . However, various uncertainties still exist in the ML upscaled output, for example the hyperparameters' setting of the ML, data quality, uneven spatial distribution of the EC stations and the representativeness of the training examples.
As with other NEE estimation models at a global or regional scale (e.g. process-based biophysical models and atmospheric inverse models), the evaluation and validation of the NEE estimation were only conducted at limited EC stations. When applied in regional or global extrapolation, it lacks a validation of the simulation results at regions without flux stations. More widely distributed meteorological stations have the potential to deliver more reliable NEE datasets to offset the limitation of the NEE validation in regions without flux stations .
The input variables and hyperparameter settings seem two factors that substantially improve the performance and reduce the computation time of the ML and should also be carefully considered . These two factors are critical because the choice of the feature subsets affects the appropriate hyperparameters and vice versa (Huang & Wang, 2006). Previous studies only individually optimized these two, either feature subset selection or hyperparameters, which greatly limits the potential for enhancing the ML performance (Ichii et al., 2017;Tramontana et al., 2016). A synergistic optimization of these two factors to search for every possible combination is a computationally expensive task. The genetic algorithm (GA) has been widely used in previous studies to find the most efficient and accurate model combination automatically and has been proven to be an effective method so as to solve this problem (Tao et al., 2019). In addition, due to the dual effects of natural and human activities, the observation timeseries of NEE and the environmental variables (e.g. the precipitation, the NDVI and EVI) often have complex time series characteristics such as the non-linear, nonstationary, lagged response and multi-time scale Tramontana et al., 2016) and thus the prediction of the long-term NEE dynamics appears to be a difficult modelling task (Friedlingstein et al., 2020;Jung et al., 2019). By decomposing the time series data into separate time series components representing the long-term trend variation, seasonal variation and residual variation (i.e. remaining information in the time series), the complexity of the time series might be effectively reduced while protecting a large part of the small-scale information (Horemans et al., 2020) (Table S1.1), which were collected from the FLUXNET2015 Dataset (Pastorello et al., 2020). The daily data were aggregated to 8-day values according to the following criteria: (1) The corresponding quality flags of NEE (i.e. NEE_VUT_REF_QC) should be >0.75 (Yang et al., 2020).
| Global meteorological station observations
The global meteorological station data used in this study were derived from the Global Surface Summary of the Day, which provides daily weather observations beginning in 1929. These daily observations were screened and aggregated to 8-day values according to the same criteria (2), (3), (4) and (5) with the above-mentioned EC stations. Finally, the observation data from 10,289 global meteorological stations were available, which could be used for a subsequent analysis.
| Explanatoryvariables
We obtained four types of explanatory variables (
| ME THODOLOGY
To improve the accuracy of the NEE prediction and to effectively fill the NEE data in the flux tower observation gap area, a three-step methodology framework with validation and extrapolation experiments was implemented in this study (Figure 1
| Time series decomposition
The NEE variations may be affected by the different time series components of the driver variables. Therefore, we decomposed each time series to explain the variables (i.e. 8-day continuous variables of the meteorological and remote sensing data over time in Table 1) into the major components: long-term trend variation, seasonal variation and residual variations, by using the Prophet model performed in Python. Then, these components are used as part of the explanatory variables to train the ML model in our study.
The Prophet model, which was recently developed by Facebook (Taylor & Letham, 2018), was designed for the analysis and forecasting of the time series data based on an additive model. Compared with the traditional time series decomposition methods, the Prophet model has no requirements regarding the regularity of the measurements' spacing and excellent adaptability to the change points in the data. It is extremely robust to the missing values, trend shifts and a large number of outliers and could achieve better results than the other traditional methods. The Prophet model has been widely evaluated in various research including the atmosphere and air quality assessment (Belikov et al., 2019).
| Adaptive genetic algorithm
The GA is a random global search optimization algorithm that simulates the biological evolution raw of the natural selection and genetic processes such as selection, cross-over and mutation to identify the optimal solutions (Whitley, 1994). In this paper, we first encoded the ML hyperparameters and the subset of the explanatory variables in a chromosome (also known as individual) using the binary of '0' and '1' that are analogous to the genes, in which '1' indicates that the hyperparameters or variables were selected. value (to produce a group of individuals more suitable for the environment, which makes the population evolving to a better area in TA B L E 1 List of the explanatory variables used for the ML training. The 'Type of Variability' indicates how the values of various variables change for a given pixel. '8-day' is the time step of this study. 'Static' variables mean never changing over time but can be used to illustrate some specific characteristics of NEE. The 'Monthly' and 'yearly' variables refer to a change in month and year, respectively.
F I G U R E 1
The framework for the global net ecosystem carbon exchange estimation at the weather stations.
the search space). In this way, the population continues to reproduce and evolve and finally converges into a group of individuals who are most suitable for the environment, thus obtaining a global optimal solution for the given problem ( Figure 2e). The probability of cross-over and mutation in the standard GA is a constant, while the improvement of adaptive genetic algorithms (AGA) includes the adaptive adjustment of the genetic parameters according to the fitness function, which will maintain the population diversity, improve the computational efficiency and speed up the convergence of the algorithm. In addition, we also adopted the Elitist Preservation (Leung & Liang, 2003) strategy so as to directly copy the best individual in the population evolution process to the next generation without a genetic operation, which could effectively prevent the loss of the best individuals in the next generation and improve the global convergence ability of the standard GA. (Table S3.2). Moreover, the Taylor diagram (Taylor, 2001) was used to compare the performance of the different algorithms.
| Algorithm training
In the training process, the spatial leave-out-one cross-validation (SLOOCV) was applied so as to develop a series of spatially extrapo-
| TransferabilityevaluationandNEEprediction for the meteorological stations
The grid-based NEE data did not evaluate the accuracy of each grid due to the sparse flux stations. Therefore, the NEE estimation at the weather stations will greatly expand the number of the global EC stations to evaluate the grid products. However, not all meteorological stations can be used to predict the NEE, and it is necessary to preevaluate the precision of the weather stations. The magnitude of R 2 has been selected as a metric to measure the precision of the meteorological stations in this study. We established the R 2 prediction model of the meteorological stations according to the Euclidean distance (ED), and we selected the model corresponding to the maximum R 2 value for each meteorological station to predict the NEE of the station. The framework was constructed using the following five steps.
| Step 1-Calculation of ED at the flux stations
After applying the SLOOCV to all flux stations, each IGBP classification and climate zone, we developed a series of prediction models based on 189 flux stations. The selection of the matched model for each weather station from the prediction models (i.e. transferability evaluation) depends on the relation of the geographic similarity and R 2 between the test sets and the training sets. Here, the geographic similarity was quantified using the ED between each test set and training set in the attributed space (Yang et al., 2008).
The ED is defined as: where d is the ED, x shows the explanatory variables in the training set, and y represents the corresponding variables in the test set. N represents the sample size of the variables for each station. For a specific flux station, the R 2 value (obtained from the NEE prediction model when it is a test set in the SLOOCV) and the ED together construct the MLR database (Figure 3a).
|
Step 2-R 2 estimation model (M ~ R 2 ) According to the EC between the training set and the test set, an estimation model of R 2 was constructed using a multivariate statistical model (MLR), which is expressed as: where M ∼ R 2 demonstrates the R 2 values of the meteorological stations, the a 0 , a 1 , … , a n stand for the regression coefficients and d 1 , d 2 , … , d n illustrate the ED of the factors influencing the NEE flux factors between the training set and test set.
|
Step 3-ED database between the explanatory variables of the flux stations and the same variables at the meteorological stations The method in step 1 has been migrated to the meteorological stations so as to calculate the ED between an influencing factor in the training set and the same influencing factor at the meteorological stations. This produces a large database (Figure 3b).
| Step 4
The R 2 migrated to the meteorological stations was obtained by means of the MLR model (Figure 3b), based on the ED database.
Here, the R 2 thresholds (low: R 2 < 0.5; moderate: 0.5 ≤ R 2 < 0.75; high: R 2 ≥ 0.75) were used to assess the transferability of each meteorological station, that is which NEE prediction models could be migrated to the current station or how many available prediction models could be matched at the current station.
| Step 5
With respect to the meteorological stations that could be connected with an applicable RF model, the NEE dataset of the meteorological stations is constructed in order to analyse the mechanisms behind the carbon dynamics.
| 1Modelperformance
The Taylor diagram summarized the performance of six models through a combination of the correlation coefficient (R), root-meansquare error deviation and standard deviation between the NEE estimation and observations (Figure 4). The RF model showed a better performance than the other models and was selected for further analysis as a result.
In this study, a total of 566 NEE inversion models were constructed based on the RF model with the SLOOCV strategy. For the 8-day NEE estimation (Table 3) (189). The NEE prediction model indicated a variation in performance for each IGBP type and climate zone ( Table 3). Among the 10,829 meteorological stations, 4674 stations met the migration conditions (≥0.5). Overall, the proportion of (2) M ∼ R 2 = a 0 + a 1 d 1 + a 2 d 2 + … +a n d n ,
F I G U R E 4
Taylor diagram of the simulated net ecosystem carbon exchange computed from six independent models with the observed data. The red line indicates the reference observations. the meteorological stations satisfying R 2 ≥ 0.5 was 45.5%, which is distributed over the world, while the remaining 54.5% of the meteorological stations could not find any matching RF model to predict the NEE (i.e. R 2 < 0.5), which is mainly spread over Central America, most parts of Africa, western and eastern Europe and southern Asia (Figure 5a). 42.8% of the meteorological stations could detect more than three RF models for the NEE prediction
| Spatiotemporalpatternsofthe meteorological site NEE
The spatial pattern of the averaged annual NEE at the meteorological stations was comparable to the results of the NOAA CarbonTracker, version CT2019B (Jacobson et al., 2020), which denotes that our methods and results are reliable for the NEE estimation at the meteorological station. The NEE values of most meteorological stations around the world are negative, indicating that these stations are carbon sinks. There are four areas with significant carbon sinks including Eastern America, Central Europe, and East and South Asia.
The long-term trend of NEE was determined by means of a linear regression method (Figure 6b). In general, most of the meteorological stations showed a decreasing trend in the mean annual NEE located in North America, Europe and Northern Asia. According to the results of the spatial pattern and trend analysis, the eastern part of America is a region where carbon dynamics change dramatically.
TA B L E 3
The model performance.
In the second column, N refers to the number of flux stations. In the RMSE, R 2 and MAE columns, the evaluation and standard deviation (between brackets) of the random forest output per IGBP and climate zone were shown. The last column illustrates the number of flux stations that satisfy R 2 ≥ 0.5 and the proportion (in parentheses). The list of acronyms refers to Table S3.1.
types, NEE showed better outcomes in DBF, MF, boreal and temperate continental, which also have a high proportion of meteorological stations that could be migrated. In contrast, the relatively poor performance in EBF and tropical forest is explained by difficulties in capturing the small seasonal dynamics of NEE in the EBF and tropical regions. Moreover, the limitations of the availability of the explanatory variables caused by the large amount of cloud coverage may also lead to low performance in these ecosystems. Difficulties F I G U R E 5 Global distribution of the transferability (a) and number of the matched models (b) at the meteorological stations. In (a), the metrics for transferability are represented by the magnitude of R 2 , where R 2 ≥ 0.5 refers to the fact that the stations meet the prediction model migration criteria and could be used to predict NEE. In (b) N represents the number of the matched models. N = 0 indicates that the meteorological stations could not be linked with any type of model (i.e. the prediction accuracy of the meteorological stations is R 2 < 0.5). N = 3 indicates that the meteorological stations could be connected with three types of models (i.e. the prediction accuracy of this meteorological station is R 2 ≥ 0.5).
in data-driven estimation of CO 2 fluxes in EBF and tropical regions are also reported in previous studies (Ichii et al., 2017;Tramontana et al., 2016). An effective way to improve the prediction accuracy of NEE in these areas is to increase the number of the EC stations in these areas.
Our results showed that the NEE performance is slightly better than for previous studies, and these studies have also pointed out more challenges in reproducing the NEE magnitude and seasonality than the other carbon components (e.g. GPP; Ichii et al., 2017;Tramontana et al., 2016;Xiao et al., 2011). The main reason for these differences in the model performance is likely as follows: Firstly, the NEE in some stations experienced substantial disturbances and fire. Ueyama et al. (2013) added the fire history at the fire-disturbed stations into the SVR-based modelling and found that the disturbance history was important to estimate NEE. Other disturbance information, including both natural and anthropogenic factors, such as land use change, turnover of the soil organic matter, hurricanes, afforestation, and extreme water stress, might improve the performance of the NEE estimations.
These disturbances may reduce the carbon exchange rate, while the MODIS data are not sensitive to capturing these disturbances, which could lead to a serious overestimation of the carbon exchange rate. Secondly, NEE is difficult to estimate because it is connected to the GPP and RE. It is particularly the RE that is difficult to estimate at a global scale due to the limited understanding of the complex interactions of the physical, chemical and biological processes and the remote sensing-based data unable to capture the parameters related to the RE (Ichii et al., 2017;Jägermeyr et al., 2014).
Furthermore, our migration model has an overall accuracy of 80.4% and is reliable. We thus assessed the transferability of 10,829 meteorological stations using the migration model, of which 4674 stations met the migration conditions. The DBF, MF and boreal have a relatively high proportion of stations to be migrated. We believe that the NEE accuracy of the meteorological stations is higher than the corresponding pixels in the global grid results that are generated by other upscale methods. Firstly, the meteorological data directly observed by the meteorological stations are used as input variables TA B L E 4 Transferability evaluation by the multivariate statistical model (MLR) at all meteorological stations over PFT and climate zone. N refers to the number of meteorological stations in different groups, R 2 < 0.5 represents the low transferability, 0.5 ≤ R 2 < 0.75 refers to the medium transferability, and R 2 ≥ 0.75 refers to the high transferability. The coloured metrics showed the percentage of different groups of the coefficient of determination (R 2 ) value (the darker the blue, the more stations in the group). The acronyms in the first columns are the same as for Table 3. in the data-driven model to estimate NEE instead of using the grid data based on reanalysis. The grid data such as reanalysis data are integrated data which assimilate the meteorological prediction data generated by atmospheric dynamics process, various ground-based observation data measured by meteorological stations, and satellite remote sensing data. It is the optimal reflection of atmospheric conditions but not the true reflection. Grid data has inherent uncertainty in the production process. In contrast, the meteorological variables obtained from meteorological stations are direct measured by related instruments and represent the actual situation of the near-surface.
| CON CLUS IONS
In this study, we estimated the NEE at the meteorological stations with 8-day intervals by combining the remote sensing data and FLUXNET2015 data with ML algorithms. The main results of the study were as follows: (1) The RF-based NEE estimations showed reliable outcomes and performed better than other similar studies.
F I G U R E 6 (a) The spatial pattern of the averaged annual NEE (g C m −2 year −1 ) at the meteorological stations. Negative fluxes (blue colours) represent the CO 2 uptake by the land biosphere, whereas the positive fluxes (red colours) indicate the CO 2 release from the land biosphere to the atmosphere. (b) The trend of the average annual NEE.
(2) Among the 10,829 meteorological stations, a total of 4674
CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare that they have no competing interests.
PEER R E V I E W
The peer review history for this article is available at https:// Supporting Information S1. Eddy covariance study sites used for this study. were optimized synchronously using the AGA algorithm.
Supporting Information S4. IGBP plant functional types and
Köppen-Geiger climate zones. | 5,401.2 | 2023-07-29T00:00:00.000 | [
"Environmental Science",
"Computer Science"
] |
Hydrodynamic Radii of Intrinsically Disordered Proteins: Fast Prediction by Minimum Dissipation Approximation and Experimental Validation
The diffusion coefficients of globular and fully unfolded proteins can be predicted with high accuracy solely from their mass or chain length. However, this approach fails for intrinsically disordered proteins (IDPs) containing structural domains. We propose a rapid predictive methodology for estimating the diffusion coefficients of IDPs. The methodology uses accelerated conformational sampling based on self-avoiding random walks and includes hydrodynamic interactions between coarse-grained protein subunits, modeled using the generalized Rotne−Prager−Yamakawa approximation. To estimate the hydrodynamic radius, we rely on the minimum dissipation approximation recently introduced by Cichocki et al. Using a large set of experimentally measured hydrodynamic radii of IDPs over a wide range of chain lengths and domain contributions, we demonstrate that our predictions are more accurate than the Kirkwood approximation and phenomenological approaches. Our technique may prove to be valuable in predicting the hydrodynamic properties of both fully unstructured and multidomain disordered proteins.
Recursive algorithm for generating Self-Avoiding Random Walks of Spheres (SARWS)
To efficiently generate GLM protein conformations, we use a recursive approach.The recursive implementation relies on the observation that for the whole chain to be free of self intersection each sub-chain within it has to be free of self intersections as well.Based on that we can generate conformations of a given length recursively by randomizing two chains of length /2 separately and then gluing them together.
For each half-chain, we add spheres subsequently, starting from one end of the protein in such a way that each added sphere has one point of contact with the previous one.The position of the point of contact is selected randomly from a uniform probability distribution on the surface of the previous sphere.Then, after the whole chain is assembled, the final construct is checked for intersections between different spheres, and self-intersecting chains are discarded.
We note that an alternative approach of simply re-randomizing the location of the last attached sphere if an intersection is detected leads to biased distributions and therefore cannot be used to generate conformations.This recursive strategy is captured by the pseudocode below: We implemented this algorithm as part of the SARWS package on which the GLM-MDA method is based.
This strategy leads to a significant performance benefit.Consider a situation where we try to generate a chain of length 4 and in our first round of randomization only beads 3 and 4 intersect.This would be detected when recursion depth is equal to 3 (combining two chains of length 1) and only two beads would have to be re-randomized rather than four in the iterative approach.Further performance gains can be achieved by implementing the algorithm above with no memory allocations as it requires only N memory cells for locations of bead centers at any moment (in our case we chose std::span to pass locations and radii in an elegant way without performance drawbacks).
The recursive approach involves a time complexity of O(N 1+γ ), and provides a satisfactory and unbiased ensemble for the largest of the proteins considered here in under a minute using only a personal computer (a single thread at 1.8 GHz).The speed of the recursive approach should be contrasted with an iterated one where steps are simply added one by one, and intersecting chains are discarded.This easier-to-implement method is characterized by a time complexity of ( ) which becomes prohibitively slow for chains with > 20.
Fast convergence of the MDA-GLM algorith for computation of R h values.
A B
Figure S1.
Computed R h value (blue) and computational time (orange) as a function of ensemble size for two cases, A) a small SAP 1A protein (n = 149, id = 13, Table S1) and B) a large H 6 -SUMO-GW182 SD-mCherry protein (n = 809, id = 42), presented with 2 standard deviations error bars estimated using 10 rounds of bootstrap, included in the computation time.Even for moderate ensemble sizes (N=20), Monte Carlo errors are smaller than hydrodynamic approximation errors.
Chemicals
The chemicals for protein expression and purification were purchased from Merck (Sigma-Aldrich) and were analytically pure, grade A, or specified for molecular biology.The AF488 NHS ester was purchased from Lumiprobe GmbH.Alexa Fluor 546 NHS ester was purchased from Invitrogen.
Proteins were labelled by using the AF488 NHS ester according to the manufacturer's protocol (Lumiprobe GmbH ) and purified from the excess of the unreacted dye by Zeba spin columns (Thermo Scientific), multi-step dialysis with use of Pur-A-Lyzers (Sigma-Aldrich), or by another SEC run on Superdex 200 Increase 10/300 GL(Cytiva), depending on the protein properties.The residual presence of the unreacted dye was taken into account in the FCS data analysis as a second component.
Fluorescence correlation spectroscopy measurements
The FCS experiments were performed essentially as described previously 6 , at Zeiss LSM 780 with ConfoCor 3, in 50 mM Tris/HCl buffer pH 8.0 (at 25 C), 150 mM NaCl, 0.5 mM EDTA, and 1 mM TCEP or DTT, in droplets of 25-30 µl.The buffer and the samples were filtered through the membrane of 0.22 m pore sizes immediately before the experiment.The protein concentrations were in the range of 10-20 nM after the filtration.The temperature inside the droplet, 25 0.5 C, was checked after the FCS measurements by means of a certified calibrated micro-thermocouple.A single measurement time was 3 to 6 s, repeated 10 to 100 times in a set.The set of measurements was repeated 3 times in 5 independent droplets.
The structural parameter (s) was determined every time with use of AF488 (DAF488 = 435 μm 2 s −1 ) or Alexa Fluor 546 (D = 341 μm 2 s −1 ) in pure water 7 , individually for each microscopic slide previously passivated with BSA in the working buffer.The actual solution viscosity was taken into account by comparison of the diffusion time for AF488 or Alexa Fluor 546 in pure water and in the buffer at the same equipment calibration.
The experiments for proteins labelled by AF488, SUMO-mαEGFP-H 6 , and mαEGFP-H 6 were performed at the 488 nm excitation wavelength with a relative Argon multiline laser power of 3 %, MBS 488 nm, BP 495-555 nm.For the mCherry-fused proteins and Alexa Fluor 546 calibration, the excitation wavelength was 561 nm at 2 % relative DPSS laser power, MBS 488/561 nm, LP 580 nm.A dampening factor of 10 % and a dust filter of 10 % were applied.
Photophysical processes of AF488 and fluorescent proteins, mCherry and mαEGFP, were investigated in independent sets of experiments.A relative laser power ranging from 3 to 20 % at 488 nm was used for the AF488 triplet state lifetime measurements.The average lifetime was determined to be about 4 µs.In the case of mCherry and mαEGFP, the measurements were performed in 30 % glycerol to slow down the protein diffusion and extract the blinking 8 .The fraction of mCherry population that undergoes blinking was found to be about 24-28 % both for the fluorescent protein alone and in the fusion constructs, and about 15 % for mαEGFP.
FCS data analysis
The FCS data were analysed by using the Zen2010 software (Zeiss).The raw measurements were closely inspected and refined to exclude possible oligomerization or aggregation of the protein sample in the confocal volume during the experiment.Global fitting of the autocorrelation curve was performed to data sets containing 10 to 50 single measurements.The autocorrelation function for 3D diffusion, including photophysical processes (triplet state for chemical dyes or blinking for fluorescent proteins) was fitted according to the equations 9 : (eq. 5) where: () is the fitted autocorrelation function; (), normalized autocorrelation function for photophysical processes; (), normalized autocorrelation function for the diffusion of n components; PT, triplet state or blinking fraction; , lifetime of the photophysical process; , , diffusion time for the i-th component; s, structural parameter of the confocal volume; Φ , fraction of the i-th diffusing component.
A one-component model (n = 1) providing for the fluorescent protein blinking was fitted for the fusion proteins, and a two-component model (n = 2), taking into account the AF488 triplet state and the presence of a residual freely diffusing dye, was used for the chemically labelled proteins.The mCherry and mαEGFP blinking fraction, as well as the AF488 triplet state lifetime determined from the independent experiments were fixed during the global analysis.
The Rh values were determined from the diffusion times, , providing for the actual buffer viscosity, as follows: where 0 is the viscosity of pure water 10 at the temperature T and _ and is the diffusion time of AF488 or Alexa Fluor 546 in the buffer at the same calibration.
The numerical regressions were performed by Prism 6 (GraphPad Software).
The total experimental uncertainty was determined according to the propagation rules for small errors 11 , taking into account both numerical uncertainty of the fitting, statistical dispersion of the results, and uncertainties of other experimental values used for calculation of the results.
A power function of the number of the polymer units (N) was fitted to the experimental R h values of folded proteins, determined by FCS (Table S1) according to the equation: The critical exponent value, , was calculated as 0.33 ± 0.02, in agreement with the value of 1/3 for a polymer chain packed into a spherical shape, and the R 0 was determined as 3.9 ± 0.6 Å, which corresponds to an average R h value for free amino acids, 3.2 ± 0.4 Å 12 .
Bioinformatics
Example conformations of the IDPs were generated by AlphaFold 2.0 notebook 13,14 .Protein structures were drawn by using Discovery Studio v3.5 (Accelrys Software).
Identification of the protein sequence fragments to be treated as ordered regions and mimicked by larger balls in the globule-linker model (GLM) was done by using Disopred3 15 .
The fragment was assumed to be ordered if the disorder probability P was less than 50 % for at least three subsequent amino acid residues, including loops linking such fragments not exceeding 14 residues 16 .
Selection of R h from literature data
The experimental benchmark set was complemented by the Rh values selected from literature.
FCS
This work 1) the R h value from FCS is slightly underestimated due to the residual presence of the freely diffusing dye impossible to be completely separated from the protein by SEC and the short diffusion time of lysozyme.
2) shown in Figures 2 and 3A (main text) for comparison with other proteins; not included in the analysis of the theoretical model S2) and experimental results (Table S1) for the benchmark set.IDPs (full green squares) and folded proteins (full black circles) from this work; IDPs from literature (blank squares); two largest outliers are marked in red (fesselin, Id. 43, N = 996, SEC) and magenta (OMM-64, Id. 39, N = 608, AUC).Error bars reflect both theoretical (Table S2, column F) and experimental uncertainties (Table S1) calculated according to small errors propagation rules.S2, columns D, F) vs. experimental results (Table S1, the benchmark set excluding globular proteins); 1:1 relationship (thin black line); linear fit to the data points without free y-intercept (green broken line, except F); (A) all R h values (full green squares, this work; blank green squares, literature); (B-F) subsets of results obtained using different experimental approaches, i.e.PFG-NMR, FCS (this work), SEC, DLS, and AUC, respectively.
Figure S8. Correlation analysis of R h values predicted for IDPs by MDA+GLM (Table
The Snedecor's F-test for the linear functions with and without the y-intercept as a free parameter fitted to the data points from the IDPs benchmark set showed that the y-intercept is insignificantly different from zero, -0.26 ± 3.6.The fit (Figure S8 A) yielded the slope of 0.96 ± 0.03 (with 90% confidence interval, CI, of 0.905 to 1.006).This means that MDA+GLM provides good 1:1 correlation with the experimental results for IDPs even at the level of 90% CI.
The R 2 of the linear correlation between the predicted and experimental results for all IDPs from Table S1 is 0.7534 (Figure S8 A), which means that our model explains ~75% of the R h variability within the IDP benchmark set.The remaining part of the variability as well as the slightly underestimated slope value can have several sources.Among the main reasons for the discrepancies are the intrinsic properties of individual experimental methods, which suffer from typical errors or limitations and are usually not taken into account when reporting the final experimental results.
The root mean square of the relative uncertainty for all experimental data (Table S1), when given, is 5.8%.Even for a perfect model that accurately predicts the diffusion coefficient, assuming the measurement uncertainty is only random (not systematic), achieving R 2 = 1 is impossible due to the inherent random noise in the data.The median R 2 values under these conditions, determined theoretically, are gathered in Table S5.
Relative error % Median R 2 of a perfect model However, the value of 5.8% seems underestimated.This is because it relies on undervalued figures provided in literature, where only some parts of the uncertainty are included in the error estimates, and in some cases, no error analysis is provided.Assuming a more realistic overall measurement error of 10%, which may still be considered small for certain measurements, the best possible model should give a typical R 2 of ~0.9.
Considering that our GLM-MDA approach involves approximated hydrodynamics, the predictions result in ~5% error of the theoretical R h values.Therefore, one should expect results only up to an R 2 of 0.85, even with exceptionally precise modeling of conformers, hydration layers, and other complex factors.
Intrinsically disordered benchmark proteins gathered in Table S1.
Sequence numbering according to Table S1.
Table S1 .
Experimental values of hydrodynamic radii, R h , for the benchmark proteins.Most of them are intrinsically disordered proteins (otherwise noticed in the Remarks column).N, number of amino acid residues in the protein chain. | 3,217.2 | 2024-02-08T00:00:00.000 | [
"Chemistry",
"Physics"
] |
Derived equivalences of gentle algebras via Fukaya categories
Following the approach of Haiden-Katzarkov-Kontsevich arXiv:1409.8611, to any homologically smooth graded gentle algebra $A$ we associate a triple $(\Sigma_A, \Lambda_A; \eta_A)$, where $\Sigma_A$ is an oriented smooth surface with non-empty boundary, $\Lambda_A$ is a set of stops on $\partial \Sigma_A$ and $\eta_A$ is a line field on $\Sigma_A$, such that the derived category of perfect dg-modules of $A$ is equivalent to the partially wrapped Fukaya category of $(\Sigma_A, \Lambda_A ;\eta_A)$. Modifying arguments of Johnson and Kawazumi, we classify the orbit decomposition of the action of the (symplectic) mapping class group of $\Sigma_A$ on the homotopy classes of line fields. As a result we obtain a sufficient criterion for homologically smooth graded gentle algebras to be derived equivalent. Our criterion uses numerical invariants generalizing those given by Avella-Alaminos-Geiss in math/0607348, as well as some other numerical invariants. As an application, we find many new cases when the AAG-invariants determine the derived Morita class. As another application, we establish some derived equivalences between the stacky nodal curves considered in arXiv:1705.06023.
Introduction
Given a Liouville manifold (M, ω = dλ), a rigorous definition of the compact Fukaya category, F (M), appears in the monograph [20]. This is a triangulated A ∞ -category linear over some base ring K. Roughly speaking, the objects of F (M) are compact, exact, oriented Lagrangian submanifolds in M, equipped with spin structures (if charK = 2). The orientations on each Lagrangian determine a Z 2 -grading on F (M), and the spin structures enter in orienting the moduli spaces of holomorphic polygons that enter into the definition of structure constants of the A ∞ operations. It is often convenient to upgrade the Z 2grading on F (M) to a Z-grading, which can be done under the additional assumption that 2c 1 (M) = 0 (see [15], [19]). Under this assumption, one defines a notion of a grading structure on M, and correspondingly considers only graded Lagrangians as objects of F (M), which now becomes a Z-graded category. We refer to [19] for these general notions. In this paper, we focus our attention to the case where M = Σ is punctured (real) 2-dimensional surface, equipped with an area form. A grading structure on Σ can be concretely described as a homotopy class of a section η of the projectivized tangent bundle of P(T Σ). Note that there is an effective H 1 (Σ)'s worth of choices (see Sec. 1). A Lagrangian can be graded if the winding number of η along L vanishes, and in such a situation a grading is a choice of a homotopy from the tangent lift L → T L ⊂ T Σ to η |L along L. These gradings extend in a straightforward manner to the wrapped Fukaya category W(Σ) which contains F (Σ) as a full subcategory, but also allows non-compact Lagrangians in Σ and more generally, partially wrapped category W(Σ, Λ), as studied in [10,Sec. 2.1], where Σ is a surface with boundary and Λ is a collection of stops (i.e., marked points) on ∂Σ.
Given two graded surfaces with stops, (Σ i , Λ i ; η i ) for i = 1, 2, a homeomorphism φ : Σ 1 → Σ 2 , which restricts to a bijection Λ 1 → Λ 2 , and a homotopy between φ * (η 1 ) and η 2 , one gets an equivalence between the partially wrapped Fukaya categories W(Σ 1 , Λ 1 ; η 1 ) and W(Σ 2 , Λ 2 ; η 2 ). Thus, it is important to have a set of explicit computable invariants of a line field η on a surface with boundary that determine the orbit of η under the action of the mapping class group of Σ. Our first result (see Theorem 1.2.5) gives such invariants in terms of winding numbers of η. In the most interesting case when genus is ≥ 2, the invariants consist of the winding numbers along all the boundary components, plus two more invariants, each taking values 0 and 1. The first of them is a Z 2 valued invariant which decides whether the line field η is induced by a non-vanishing vector field, while the second is the Arf-invariant of a certain quadratic form over Z 2 . Note that from the numerical invariants of Theorem 1.2.5 one can also recover the genus of the surface and the numbers of stops on the boundary components, so if these invariants match then then the corresponding partially wrapped Fukaya categories are equivalent.
Next, we use this result to construct derived equivalences between gentle algebras, introduced by Assem and Skowrónski in [3]. This is a remarkable class algebras with monomial quadratic relations of special kind with a well understood structure of indecomposable modules. Furthermore, their derived categories of modules also enjoy many nice properties (see [7] and references therein). Avella-Alaminos and Geiss [5] gave a combinatorial definition of derived invariants of finite-dimensional gentle algebras, which form a collection of pairs of non-negative integers (m, n) with multiplicities. We refer to these as AAG-invariants. It is known that these invariants do not completely determine the derived Morita class of a gentle algebra in general (for example, see [1]).
We consider Z-graded gentle algebras and their perfect derived categories (the classical case corresponds to algebras concentrated in degree 0). For such an algebra A, we denote by D(A) the perfect derived category of dg-modules over A viewed as a dg-algebra with zero differential. The category D(A) has a natural dg-enhancement which we take into account when talking about equivalences involving D(A).
The connection between graded gentle algebras and Fukaya categories was established by Haiden, Katzarkov and Kontsevich in [10] (cf. [6]): they constructed collections of formal generators in some partially wrapped Fukaya categories whose endomorphism algebras are graded gentle algebras. In Theorem 3.2.2 we give an inverse construction 1 : starting from a homologically smooth graded gentle algebra A we construct a graded surface with stops (Σ A , Λ A ; η A ) together with a set formal generators whose endomorphism algebra is isomorphic to A. This leads to an equivalence of the partially wrapped Fukaya category W(Σ, Λ) with the derived category D(A). In addition, we generalize the combinatorial definition of AAG-invariants to possibly infinite-dimensional graded gentle algebras and show that they can be recovered from the winding numbers of η A along all boundary components. Now recalling our numerical invariants of graded surfaces with stops from Theorem 1.2.5 we obtain a sufficient criterion for derived equivalence between homologically smooth graded gentle algebras. Namely, if we start with two such algebras A and A ′ and find that the corresponding invariants from Theorem 1.2.5, determined by winding numbers of η A and η A ′ , coincide then we get a derived equivalence between A and A ′ . Note that this involves checking that A and A ′ have the same AAG-invariants, and in addition that two more invariants with values in {0, 1} match.
As an application, using the above approach we obtain a sufficient criterion for derived equivalence of homologically smooth graded gentle algebras given purely in terms of AAGinvariants (see Corollary 3.2.5). Using Koszul duality, we also get a sufficient criterion for derived equivalence of finite-dimensional gentle algebras with grading in degree 0 (see Corollary 3.2.6).
In a different direction, we construct derived equivalences between stacky nodal curves studied in [16], Namely, these are either chains or rings of weighted projective lines glued to form stacky nodes, locally modelled by quotients (xy = 0)/(x, y) ∼ (ζ k x, ζy), where ζ r = 1 and k ∈ (Z/r) * . In [16,Thm. B] we constructed an equivalence of the derived category of coherent sheaves on such a stacky curve with the partially wrapped Fukaya category of some graded surface with stops (this can be viewed as an instance of homological mirror symmetry). Thus, using Theorem 1.2.5 we get many nontrivial derived equivalences between our stacky curves. In the case of balanced nodes (those with k = −1) we recover the equivalences between tcnc curves from [21].
Acknowledgments. Y.L. is partially supported by the Royal Society (URF) and the NSF grant DMS-1509141, and would like to thank Martin Kalck for pointing out the reference [5]. A.P. is supported in part by the NSF grant DMS-1700642 and by the Russian Academic Excellence Project '5-100'. While working on this project, A.P. was visiting King's College London, Institut des Hautes Etudes Scientifiques, and Korea Institute for Advanced Study. He would like to thank these institutions for hospitality and excellent working conditions. 1. Line fields on surfaces 1.1. Basics on line fields. Let Σ be an oriented smooth surface of genus g(Σ) with nonempty boundary with connected components ∂Σ = b i=1 ∂ i Σ. The mapping class group of Σ is M(Σ) = π 0 (Homeo + (Σ, ∂Σ)), where Homeo + (Σ, ∂Σ) is the space of orientation preserving homeomorphism of Σ which are the identity pointwise on ∂Σ. Definition 1.1.1. An (unoriented) line field η on Σ is a section of the projectivized tangent bundle P(T Σ). We denote by G(Σ) = π 0 (Γ(Σ, P(T Σ))) the set of homotopy classes of unoriented line fields.
A non-vanishing vector field i.e. a section of the unit tangent bundle SΣ induces a line field via the bundle map SΣ → P(T Σ) which is a fibrewise double covering. However, not all line fields come from non-vanishing vector fields: a section of P(T Σ) may not lift to a section of SΣ.
The trivial circle fibration S 1 ι − → P(T Σ) p − → Σ induces an exact sequence A line field η determines a class [η] ∈ H 1 (P(T Σ)) such that ι * [η]([S 1 ]) = 1 by taking the Poincaré-Lefschetz dual of the class of the image [η(Σ)] ⊂ H 2 (P(T Σ), ∂P(T Σ)). Via this construction, we get an identification where ζ ∈ H 1 (S 1 ) is the generator which integrates to 1 along S 1 . Thus, the set G(Σ) is a torsor over H 1 (Σ). We denote the corresponding action of c ∈ H 1 (Σ) on G(Σ) by The mapping class group M(Σ) acts on G(Σ) on the right. Our goal in this section is to understand the orbit decomposition of G(Σ) with respect to this action.
The winding number w η (γ) with respect to η only depends on the homotopy class of η and the regular homotopy class of γ. From the definition we immediately get the following compatibility with the action of H 1 (Σ): Throughout, ∂Σ is oriented with respect to the natural orientation as the boundary of Σ. In particular, w η (∂D 2 ) = 2 for the unique homotopy class of line fields on D 2 . For a boundary component B ⊂ ∂Σ with the opposite orientation, we write −B. Then, we have w η (−B) = −w η (B).
1.2.
Invariants under the action of the mapping class group. The winding numbers along boundary components of Σ gives the first set of invariants of elements of G(Σ). To go further, we need to study the winding numbers along non-separating curves on Σ. As is well-known, the winding number invariants do not descend to a map from H 1 (Σ). Indeed, if S ⊂ Σ is a compact subsurface with boundary ∂S = n i=1 ∂ i S, by Poincaré-Hopf index theorem (see [11,Ch. 3]), we have: However, considering the reduction modulo 2 we still get a well-defined homomorphism: i.e an element H 1 (Σ; Z 2 ).
We have a natural inclusion induced map The cokernel of i is isomorphic to Z 2g 2 and comes equipped with a non-degenerate intersection pairing.
Note that the numbers r i (η) mod 2 are precisely the values of [w η ] (2) on the boundary cycles. Thus, if r i (η) is odd for at least one i then σ(η) = 1. If all r i (η) are even then we can check whether σ(η) = 0 by looking at the winding numbers of a collection of cycles projecting to a basis of the cokernel of i. given by where α i are simple closed curves. It satisfies where a, b ∈ H 1 (Σ; Z 4 ), and a · b denotes the intersection pairing on H 1 (Σ; Z 4 ).
Proof. In the case when η comes from a vector field v, we have w η (a) = 2w v (a), where w v (·) is the winding number of the vector field. Hence, the assertion in this case follows from [12, Thm 1A, Thm 1B]. In general we have [η] = η 0 + c, for some c ∈ H 1 (Σ). Thus, the function q η (a) := q η 0 (a) + c, a has the claimed properties. Proof.
Thus, the study of the M(Σ)-orbits on G(Σ) reduces to the study of M(Σ)-orbits on the set of functions q : Let us denote by Quad 4 = Quad 4 (Σ) the set of all such functions (it is an H 1 (Σ, Z 4 )-torsor).
Recall that given a symplectic vector space V, (− · −) over Z 2 , one can consider the set Quad(V ) of quadratic forms q : V → Z 2 satisfying For every q ∈ Quad(V ), the Arf-invariant ( [2], [8]) is the element of Z 2 given by where (a i , b i ) is a symplectic basis of V . The Arf invariant is the value that q attains on the majority of vectors in V . In the case when r i (η) = w η (∂ i Σ) + 2 ∈ 4Z for every i = 1, . . . , d, and the quadratic function q = q η takes values in 2Z 4 , we can associate to q a certain quadratic form on a Z 2 -vector space, and its Arf-invariant will give us an additional invariant of η modulo the mapping class group action.
Let us set H := H 1 (Σ, Z 4 ), K = im(i * : H 1 (∂Σ, Z 4 ) → H 1 (Σ, Z 4 )), H = H/2H, K = K/2K. Since K lies in the kernel of the intersection pairing, for any q ∈ Quad 4 the restriction q| K is a homomorphism K → Z 4 . Note that for q = q η the value of this homomorphism on [∂ i Σ] is r i (η) mod 4. Now let q ∈ Quad 4 be such that q| K is zero. Then it is easy to see that q descends to a well defined function q H/K on H/K. Assume in addition that σ(η) = 0, i.e., q takes values in 2Z 4 . In this case we have q H/K = 2q, where q is a function H/K → Z 2 satisfying (1.3). It is easy to see that q(x + 2y) = q(x), so q can be viewed as a Z 2 -valued quadratic form on H/K ≃ Z 2g 2 . Thus, q is an element of Quad(H/K) and we define A(η) as the Arf-invariant of q.
Here α, β are simple curves such that [α] and [β] project to a basis of Then two line fields η and θ are in the same M(Σ) orbit if and only if the following conditions are satisfied: Proof. (i) This follows immediately from the fact that G(Σ) is an H 1 (Σ)-torsor and the boundary curves ∂ i Σ generate the group H 1 (Σ). (ii) This is proved in the same way as Theorem 2.8 in [13]. (iii) We need to prove that if the invariants match then η and θ are in the same M(Σ)orbit. Note that σ(η) is determined by whether the quadratic function q η is trivial modulo 2 or not. By Lemma 1.2.4, it is enough to prove that the quadratic functions q η and q θ are in the same M(Σ)-orbit.
First, let us analyze the result of the action of a transvection T a (x) = x + (a · x)a on quadratic functions in Quad 4 . We have Assume now that q ∈ Quad 4 is such that q| K is surjective, i.e., the reduction of q| K modulo 2 is nonzero. Then we claim that any q ′ ∈ Quad 4 with q ′ | K = q| K lies in the M(Σ)-orbit of q. Indeed, we have q ′ − q = (a·?) for some a ∈ H. By surjectivity of q| K we can find k ∈ K such that q(k) = −1 − q(a), i.e., q(a + k) = −1. Then from (1.4) we get Next, let us consider q ∈ Quad 4 such that q| K takes values in 2Z 4 . Assume also that q mod 2 = 0. We claim that in this case the M(Σ)-orbit of q is determined by q| K . Note that q mod 2 is a homomorphism H → Z 2 trivial on K, so it is an element of Hom(H/K, Z 2 ). Since M(Σ) acts transitively on nonzero elements in Hom(H/K, Z 2 ), it is enough to prove that if q ′ ≡ q mod 2 and q ′ | K = q| K then q ′ and q are in the same M(Σ)-orbit. As before we deduce that q ′ − q = 2(a·?) for some a ∈ H. If q(a) ≡ 1 mod 2 then this immediately gives q ′ = qT 2 a . On the other hand, if q ′ (a) ≡ q(a) ≡ 0 mod 2 then for any element b with q(b) ≡ 1 mod 2 we have (1.5) 2. In the case g(Σ) = 1, let α, β be the standard non-separating curves in Σ. Then, it can be shown as in [13, Lemma 2.6] that We also note that in the case d = 1, w η (∂Σ) = −2, hence this invariant reduces to gcd(w η (α), w η (β)) considered in [1].
3. In the case σ(η) = 0, the line field η is induced by a non-vanishing vector field v. This induces a spin structure on the surface Σ (by considering its mod 2 reduction). The condition that w η (∂ i Σ) ∈ 2 + 4Z means that this spin structure extends to a spin structure on the compact surface obtained from Σ by capping off the boundaries with a disk. Now, it is a theorem of Atiyah [4] (see also [12]) that the action of the mapping class group on the spin structures on a compact Riemann surface has exactly 2 orbits distinguished by the Arf invariant.
Partially wrapped Fukaya categories
The partially wrapped Fukaya category W(Σ, Λ; η) (with coefficients in a field K) is associated to a graded surface (Σ, Λ; η), where Σ is a connected compact surface with non-empty boundary ∂Σ, Λ ⊂ ∂Σ is a collection of marked points called stops, and η is a line field on Σ. There is a combinatorial description of W(Σ, Λ; η) provided in [10]. A set of pairwise disjoint and non-isotopic Lagrangians {L i } in Σ\Λ generates the partially wrapped Fukaya category W(Σ, Λ; η) as a triangulated category if the complement of the Lagrangians is a union of disks D f each of which has exactly one stop on its boundary. Figure 1 illustrates how each D f may look like, where the blue arcs are in i L i while the black arcs lie in ∂Σ.
Furthermore, in this case, the associative K-algebra is formal, and it can be described by a graded gentle algebra (see Def. 3.1.1). The generators of this quiver can easily be described following the flow lines corresponding to rotation around the boundary components of Σ connecting the Lagrangians. Note that each boundary component of Σ is an oriented circle (where the boundary orientation is induced by the area form on Σ). Specifically, a flowline that goes from L j to L i gives a generator for hom(L i , L j ) (note the reversal of indices). The data of Λ enters by disallowing flows that pass through a marked point. The algebra structure is given by concatenation of flow lines. Given α i ∈ hom(L i , L i+1 ) for i = 1, . . . , n, we write for their product, read from right-to-left, and if non-zero, this expression corresponds to a flow from L n to L 1 . Finally, the line field η is used to grade the morphism spaces. A convenient way to determine the line field η is by describing its restrictions along each of the disks D f . Each such disk is as in Figure 1. Different disks are glued along the curves L i (the blue parts in their boundary). Changing a line field by homotopy, we can arrange that it is tangent to L i (as L i are contractible). Thus, every line field on Σ (up to homotopy) can be glued out of such line fields on the disks D f .
Note that if we have an embedded segment c ⊂ Σ and a line field η, which is transversal to c at the ends p 1 , p 2 of c, then we can define the winding number w η (c) (first, one can trivialize T Σ along c in such a way that the tangent line to c is constant, then count the number of times (with sign) η coincides with the tangent line to c along c. An equivalent definition is given in [10,Sec. 3.2]). Now a line field on a disk D f , tangent to {L i }, is determined (up to homotopy) by the integers θ i , for i = 1, . . . , m, given by its winding numbers along the boundary parts on ∂Σ (the black parts in Figure 1). By definition, these numbers are the degrees of the corresponding morphisms in the wrapped Fukaya category.
The numbers θ i can be chosen arbitrarily subject to the constraint This last constraint is the topological condition that needs to be satisfied in order for the line field to extend to the interior of the disk. (Note that the stops do not play a role in this discussion.) The gentle algebra A L• is always homologically smooth since so is W(Σ, Λ; η). In what follows, it will be convenient to consider A op L• as a quiver algebra KQ/I, so that flow lines from L i to L j correspond to arrows from the i th vertex to j th vertex. Note that the collection {L i } generates the partially wrapped Fukaya category W(Σ, Λ; η). Therefore, we have an equivalence , where the category on the left denotes the bounded derived category of perfect (left) dgmodules over A op L• .
Gentle algebras and Fukaya categories
3.1. Graded gentle algebras and AAG-invariants. A quiver is a tuple Q = (Q 0 , Q 1 , s, t) where Q 0 is the set of vertices, Q 1 is the set of arrows, s, t : Q 1 → Q 0 is the functions that determine the source and target of the arrows. We always assume Q to be finite. A path in Q is a sequence of arrows α n . . . α 2 α 1 such that s(α i+1 ) = t(α i ) for i = 1, . . . , (n − 1). A cycle in Q is a path of length ≥ 1 in which the beginning and the end vertices coincide but otherwise the vertices are distinct. For K a field, let KQ be the path algebra, with paths in Q as a basis and multiplication induced by concatenation. Note that the source s and target t maps have obvious extensions to paths in Q.
Definition 3.1.1. A gentle algebra 2 A = KQ/I is given by a quiver Q with relations I such that (1) Each vertex has at most two incoming and at most two outgoing edges.
(2) The ideal I is generated by composable paths of length 2.
(3) For each arrow α, there is at most one arrow β such that αβ ∈ I and there is at most one arrow β such that βα ∈ I. (4) For each arrow α, there is at most one arrow β such that αβ / ∈ I and there is at most one arrow β such that βα / ∈ I. In addition, we always assume Q to be connected.
We will consider Z-graded gentle algebras, i.e., every arrow in Q should have a degree assigned to it. For a Z-graded gentle algebra A we denote by D(A) the derived category of perfect dg-modules over A, where A is viewed as a dg-algebra with its natural grading and zero differential. . But the latter space can be computed using the standard Koszul complex, and the presence of forbidden cycles would mean that for some S the space Ext * A (S, S) is infinite-dimensional.
We will use the following notions from [5]. such that all (α i ) are distinct and for all i = 1, . . . , (n − 2), α i+1 α i ∈ I. It is a forbidden thread if for all β ∈ Q 1 neither βα n . . . α 2 α 1 nor α n . . . α 2 α 1 β is a forbidden path. In addition, if v ∈ Q 0 with #{α ∈ Q 1 |s(α) = v} ≤ 1, #{α ∈ Q 1 |t(α) = v} ≤ 1, then we consider the idempotent e v as a (trivial) forbidden thread in the following cases: • either there are no α with s(α) = v or there are no α with t(α) = v; • we have β, γ ∈ Q 1 with s(γ) = v = t(β) and γβ ∈ I. The grading of a forbidden thread is defined by such that all (α i ) are distinct and for all i = 1, . . . , (n − 1), α i+1 α i / ∈ I, and it is a permitted thread if for all β ∈ Q 1 neither βα n . . . α 2 α 1 nor α n . . . α 2 α 1 β is a permitted path. In addition, if v ∈ Q 0 with #{α ∈ Q 1 |s(α) = v} ≤ 1, #{α ∈ Q 1 |t(α) = v} ≤ 1, then we consider the idempotent e v as a (trivial) permitted thread in the following cases: • either there are no α with s(α) = v or there are no α with t(α) = v; • we have β, γ ∈ Q 1 with s(γ) = v = t(β) and γβ / ∈ I. The grading of a permitted thread is defined by Remark 3.1.6. Inclusion of the idempotents as forbidden and permitted threads ensures that every vertex appears in exactly two forbidden threads/cycles and exactly two permitted threads/cycles. such that s(f i ) = s(p i ) for i ∈ Z/n, and t(p i ) = t(f i+1 ) for i ∈ Z/n with the following condition: The winding number associated to a combinatorial boundary component b of type I is defined to be We also denote the number n of forbidden threads in b as n(b).
A combinatorial boundary component of type II (that can appear only if A is not proper) is simply a permitted cycle pc = α m . . . α 1 . The winding number associated to such a cycle is A combinatorial boundary component of type II' (that can appear only if A is not homologically smooth) is simply a forbidden cycle The winding number associated to such a cycle is For combinatorial boundary components of types II and II' we set n(b) = 0. Proof. This follows directly from the description of the AAG-invariants in [5,Sec. 3]. Note that the pair (0, m) in Step (3)
3.2.
Relation to Fukaya categories. The definition of the combinatorial boundary component for a gentle algebra is motivated by the following proposition: Proof. Figure 2 shows an example of the way the surface Σ looks around a boundary component B. Assume first that there is at least one stop on B. Let q 1 (1), . . . , q 1 (k 1 ), q 2 (1), . . . , q 2 (k 2 ), . . . , q n (1), . . . , q n (k n ) be the endpoints of the Lagrangians ending on B, ordered compatibly with the orientation of B. Here we assume that there are no stops between q i (j) and q i (j + 1) and there is exactly one stop s i between q i (k i ) and q i+1 (1), for i ∈ Z/n. Then for every i ∈ Z/n we have a permitted thread p i = β i (k i − 1) . . . β i (1), where β i (j) is the generator of A corresponding to the flow on B from q i (j) to q i (j + 1). On the other hand, each stop s i lies on a unique disk D, and by looking at the pieces of ∂D formed by other boundary components of Σ, we obtain a forbidden thread f i = α m i . . . α 1 starting at the Lagrangian corresponding to q i (1) and ending at the one corresponding to q i−1 (k i−1 ). Thus, we get a combinatorial boundary component of type I, b = p n f n . . . p 1 f 1 .
The winding number of η along the arc passing through the stop, oriented in the opposite direction to the boundary direction, is determined using the constraint (2.1) to be On the other hand, the winding number of η along the arc corresponding to the permitted thread p is simply |p|. Thus, we get the equality w η (B) = w(b).
In the case of a boundary component B ⊂ ∂Σ with no stops, the sequence of flows between the corresponding ends of Lagrangians on B gives a permitted cycle, i.e., a combinatorial boundary component of type II. Again, the winding numbers match.
It is easy to see that in this way we get a bijection between the boundary components B and the combinatorial boundary components of A. Figure 2. The boundary component is given by the cyclic sequence p 2 f 2 p 1 f 1 where f 1 = α 3 α 2 α 1 , p 1 = β 2 β 1 , f 2 = γ 2 γ 1 and p 2 = δ 1 . Note that if instead of f 1 , we considered the forbidden threadf 1 =β 2 β 1 , the condition (⋆) is violated.
Theorem 3.2.2. (i) Given a homologically smooth graded gentle algebra A over a field K (with |Q 1 | > 0), there exists a graded (connected) surface with stops (Σ A , Λ A , η A ), with non-empty boundary and a derived equivalence Furthermore, the AAG-invariants of A are given by the collection of pairs Proof. (i) We define a ribbon graph R A whose vertices are in bijection with the collection of forbidden threads in Q, and whose edges are in bijection with vertices of Q.
Recall that there are precisely two forbidden threads that pass through a vertex of Q. The corresponding edge on R A is defined to connect the two forbidden threads. Furthermore, we can equip the set of edges in R A incident to a vertex f with a total ordering. Namely, the set of edges incident to a vertex f of R is in bijection with the set of vertices of Q which appear in the forbidden thread f . Hence, we can use the order in which these vertices appear in the forbidden thread. This linear order of edges incident to vertices of R A induces a ribbon structure on R A , i.e., a cyclic order of edges incident to each vertex. Therefore, we can consider the associated thickened surface Σ A such that R A is embedded as a deformation retract of Σ A . (A graph with the additional data of a linear ordering on the edges incident to a vertex is called a ciliated fat graph [9].) Thus, to construct Σ A we replace each vertex of R A with a 2-disk D 2 and each edge with a strip, a thin oriented rectangle [−ǫ, ǫ] × [0, 1], where the rectangles are attached to the boundary of the disks according to the given cyclic orders at the vertices. On the boundary of each disk associated to the vertex of R A we also mark a point, called a stop as follows. If the linear order on edges incident to this vertex is given by e 1 < e 2 < . . . < e k , the stop e 0 appears in the circular order such that e k < e 0 < e 1 . We define Λ A by taking the union of all such points. In particular, the cardinality of Λ A , is equal to the number of forbidden threads in A.
We claim that the ribbon graph R A and hence the associated surface Σ A is connected. Indeed, for every vertex v of Q let e(v) be the corresponding edge in R A , viewed as a subgraph in R A . Since Q is connected, it is enough to check that if v and v ′ are connected by an edge α in Q then e(v) and e(v ′ ) intersect in R A . Indeed, let f be a forbidden thread containing α (it always exists). Then f is a vertex of both e(v) and e(v ′ ). This proves our claim that R A is connected.
Dual to the edges of R A we obtain a disjoint collection of non-compact arcs L v indexed by vertices of Q. Thus, Σ A is a surface with non-empty oriented boundary, Λ A is a set of marked points in its boundary, and {L v : v ∈ Q 0 } is a collection pair-wise disjoint and non-isotopic Lagrangian arcs in Σ A \ Λ A . Furthermore, the complement is a union of disks D f indexed by forbidden threads f in Q, with exactly one marked point on its boundary (see Examples 3.2.7, 3.2.8 below). In particular, the collection {L v } gives a generating set.
By construction, there is a bijection between arrows in the quiver Q and the generators of the endomorphism algebra A L := v,w hom(L v , L w ) since each edge α in Q is in exactly one forbidden thread f , and the corresponding D f has a flow associated to α. Furthermore, two flows α 1 : L v 2 → L v 1 and α 2 : L v 3 → L v 2 can be composed in A L if and only if α i is in a forbidden thread f i , for i = 1, 2, such that the disks D f 1 and D f 2 are glued along the edge corresponding to v 2 . But this means that the corresponding elements of A satisfy α 2 α 1 / ∈ I, as otherwise condition (3) of Definition 3.1.1 would be violated. This imples that A is naturally identified with A op L as an ungraded algebra. We define the line field η A on Σ A as follows. We require that the line field is tangent to each L v . Then it suffices to describe its restrictions to the disks D f . Each D f is a 2m-gon as in Figure 1. The homotopy class of a line field on D f is determined by the winding numbers θ i along the boundary arcs of D f , α i , for i = 1, . . . , (m − 1), avoiding the unique stop (black in Figure 1) between the Lagrangians (blue in Figure 1). Indeed, the remaining winding number θ m along the boundary arc that passes through the stop is determined by the condition m i=1 θ i = m − 2, and we can define η A | D f as the unique line field with these winding numbers. Now we set θ i , for i = 1, . . . , m − 1, to be the degree of the generator of A corresponding to α i .
With this definition A and A op L are identified as graded algebras. Since we also know that the collection Finally, the last statement follows from Proposition 3. Combining this with the previous formula we get v(R A ) = 2|Q 0 | − |Q 1 |, so we deduce that χ(R A ) = χ(Q).
Using formula (1.5) we derive the following property of the AAG-invariants.
..,d be the AAG-invariants of a homologically smooth graded gentle algebra A. Then where g ≥ 0 is the genus of the corresponding surface Σ A . As a particular case of the last Corollary, we can describe some cases when already looking at the AAG-invariants gives the derived equivalence. We can use Koszul duality to convert our results about homologically smooth graded gentle algebras into those about finite-dimensional gentle algebras. Namely let A be a finite-dimensional gentle algebra with grading in degree 0. Let A ! be the Koszul dual gentle algebra (with respect to the generators given by the edges). We equip A ! with the grading for which all edges have degree 1 (i.e., path-length grading). Then the result of Keller in [14, Sec. 10.5] ("exterior" case) gives an equivalence where D f (A) is the bounded derived category of finite-dimensional A-modules (and D(A ! ) is the perfect derived category of A ! viewed as a dg-algebra, as before).
Furthermore, it is easy to check that the AAG-invariants of A and A ! are the same. Thus, Corollary 3.2.5 leads to the following result. The associated ribbon graph is given in Figure 4, where the cyclic order at vertices are given by counter-clockwise rotation. Figure 5 depicts the corresponding surface, together with the dual arcs L 1 , L 2 , L 3 , L 4 . The corresponding surface is given in Figure 7.
Remark 3.2.9. An optimist's conjecture would be that conversely if A and B are homologically smooth graded gentle algebras which are derived equivalent, then there exists a homeomorphism φ : Σ A → Σ B inducing a bijection Λ A → Λ B and such that φ * (η A ) is homotopic to η B . Note that to prove this, one needs to show that the topological type of (Σ A , Λ A ; η A ) is a derived invariant of A. This is encoded by the numerical invariants of η A introduced in Theorem 1.2.5 (from which one can recover the topological type of the surface), together with the numbers of marked points on each boundary component.
Remark 3.2.10. In Theorem 3.2.2, it is possible to drop the assumption that A is smooth. Assume for simplicity that A is proper. In this case, the surface Σ would be glued together from the disks D f associated to forbidden threads as before, and also disks D c with an interior hole, associated with forbidden cycles. In other words, D c is an annulus whose inner boundary component has no marked points and is not glued to anything, while its outer boundary component is connected by strips, corresponding to the vertices in c, to other disks (this boundary component of D c still has no stops). In the presence of unmarked boundary components, there is a dual construction to the construction of partially wrapped Fukaya categories, W(Σ, Λ; η), namely, the infinitesimal wrapped Fukaya categories F (Σ, Λ; η), studied in [16]. Its objects are graded Lagrangians which do not end on the unmarked components of the boundary. Thus, for non-smooth proper gentle algebras, a version of Theorem 3.2.2 should state the equivalence However, we have not checked that the collection of Lagrangians {L v } given by the construction in Theorem 3.2.2 (and modified as above) generates F (Σ A , Λ A ; η A ).
Derived equivalences between stacky curves
4.1. Chains. Recall that in [16] we considered stacky curves C(r 0 , . . . , r n ; k 1 , . . . , k n−1 ) obtained by gluing weighted projective lines B(r 0 , r 1 ), B(r 1 , r 2 ), . . . , B(r n−1 , r n ) into a chain, where k i ∈ (Z/r i ) * are used to determine the stacky structure of the nodes in this chain. We showed in [16,Thm. B] that the bounded derived category of coherent sheaves on such a stacky curve is equivalent to the partially wrapped Fukaya category of a surface obtained by a certain linear gluing of the annuli. Namely, let A(r, r ′ ) denote the annulus with ordered boundary components that has r marked points p − 1 , . . . , p − r on the first component and r ′ marked points p + 1 , . . . , p + r ′ on the second boundary component (the points are ordered cyclically compatibly with the orientation of the boundary). Given a collection of permutations σ i ∈ S r i , i = 1, . . . , n − 1, we consider the surface Σ lin (r 0 , . . . , r n ; σ 1 , . . . , σ n−1 ) obtained by gluing the annuli A(r 0 , r 1 ), A(r 1 , r 2 ), . . . , A(r n−1 , r n ) in the following way. For each i = 1, . . . , n − 1, j = 1, . . . , r i , we glue a small segment of the boundary around the marked point p + j in A(r i−1 , r i ) with a small segment of the boundary around the point p − σ i (j) in A(r i , r i+1 ) by attaching a strip, as in Figure 8. Note the resulting surface has two special boundary components equipped with r 0 and r n marked points, respectively (there are no other marked points on the other boundary components). The boundary components that arise because of the gluing are from n − 1 groups, so that components in the ith group are in bijection with cycles in the cycle decomposition of the commutator [σ i , τ ] ∈ S r i , where τ is the cyclic permutation j → j −1.
We equip each annulus with the standard line field that has zero winding numbers on both boundary components. These line fields glue into a line field η on the surface Σ lin (r 0 , . . . , r n ; σ 1 , . . . , σ n−1 ). More precisely, we take η that corresponds to the horizontal direction in Figure 8.
It is easy to see that the boundary invariants of η are given as follows. For the two special boundary components the winding numbers are equal to zero, so the corresponding invariant is 2. For a boundary component corresponding to a k-cycle in the cycle decomposition of [σ i , τ ] the winding number is −2k, so the invariant is 2 − 2k.
Note that for each i the commutator [σ i , τ ] is given by x → x + k i + 1 mod(r i ), so its cycle decomposition has p i = gcd(k i + 1, r i ) cycles of length r i /p i . Thus, the corresponding boundary invariants are 2, 2 (for the special boundary components) and for each i = 1, . . . , n − 1, the number 2 − 2r i /p i repeated p i times.
4.1.2.
Trade-off for balanced nodes. More generally, let I ⊂ [1, n − 1] be the subset of indices i such that k i = −1, and let r I = i∈I r i . Then, we have a homeomorphism Σ lin (r 0 , . . . , r n ; k 1 , . . . , k n−1 ) ≃ Σ lin (r 0 , r I , (r i ) i ∈I , r n ; −1, (k i ) i ∈I ) preserving the line fields. This can be either derived from Theorem 1.2.5 as above or constructed directly. As before, this leads to a derived equivalence of the corresponding stacky curves. 4.1.4. Genus ≥ 2. Because of the two special components with the boundary invariant 2, the Arf-invariant never appears. Thus, two surfaces Σ lin (r • ; k • ) and Σ lin (r ′ • ; k ′ • ) of genus g ≥ 2 are homeomorphic as surfaces with a line field, whenever we have r 0 = r ′ 0 , r n = r ′ n and the sequence ((r 1 /p 1 ) p 1 , . . . , (r n−1 /p n−1 ) p n−1 ) differs from the corresponding sequence for (r ′ • , k ′ • ) by a permutation (here (r i /p i ) p i means the number r i /p i repeated p i times). For example, we can specialize to the case n = 2, r 0 = r 2 = 0, r 1 = r. Note that the corresponding stacky curve C(0, r, 0; k) is the global quotient of the affine coordinate cross xy = 0 by the µ r -action ζ · (x, y) = (ζ k x, ζy). We obtain that for k, k ′ ∈ (Z/r) * , such that gcd(k + 1, r) = gcd(k ′ + 1, r), there exists an equivalence D b Coh(C(0, r, 0; k)) ≃ D b Coh(C(0, r, 0; k ′ )).
4.2.
Rings. Now let us consider another class of stacky curves considered in [16], denoted by R(r 1 , . . . , r n ; k 1 , . . . , k n ). They are defined by gluing the weighted projective lines B(r 1 , r 2 ), B(r 2 , r 3 ), . . . , B(r n , r 1 ) into a ring, where as before k i ∈ (Z/r i ) * are used to determine the stacky structure of the nodes.
On the symplectic side we can modify our definition of the surfaces Σ lin (r 0 , . . . , r n ; σ 1 , . . . , σ n−1 ) as follows. Starting with the annuli A(r 1 , r 2 ), A(r 2 , r 3 ), . . . , A(r n , r 1 ) we can glue them circularly using permutations σ 1 , . . . , σ n . Thus, the corresponding surface could be represented similarly to Figure 8 but with the right and left ends identified (so that the corresponding boundary components disappear). We denote the resulting surface by Σ cir (r 1 , . . . , r n ; σ 1 , . . . , σ n ). Similarly to the case of a linear gluing it is equipped with a natural line field η that corresponds to the horizontal direction when the surface is depicted as on Figure 8.
4.3.
Case of irreducible stacky curve. This is the case n = 1. Let r = r 1 . Let us consider the case of k ∈ Z r such that gcd(k + 1, r) = 1 (note that this is possible only when r is odd). Then the surface Σ cir (r; k) has genus g = (r + 1)/2 and one boundary component with the winding number −2r, i.e., the invariant 2 − 2r. Note that 2 − 2r is divisible by 4, so to determine the orbit of the line field under the mapping class group we have to calculate the corresponding Arf-invariant. This invariant will depend on k.
First, let us consider the case k = 1. Let us look at the simple curves α i , i = 1, . . . , r − 1, depicted on Figure 9. In addition, we have two simple curves α and β corresponding to a vertical and horizontal line on Figure 9.
Hence, the Arf-invariant is given in this case by 1 + (r−1)/2 Next, assume in addition that r is not divisible by 3 and consider the case k = 2 (then gcd(k + 1, r) = 1). Then we claim that the classes [α i ], together with (α, β) still project to a basis of H 1 (Σ, Z 2 ), however, their intersection numbers are now given by where i < j (we still have α i · α = α i · β = 0 and α · β = 1. It is easy to see that by renumbering the classes (α i ) as follows: we get the quadratic form of Example 4.5.2. Hence, the Arf invariant is given by 1 + (r − 1)/2 mod 2. Thus, we deduce the following derived equivalence. Proposition 4.3.1. Assume that r ≥ 7 is not divisible by 3 and r ≡ ±1 mod (8). Then the stacky curves C ring (r; 1) and C ring (r; 2) are derived equivalent.
4.4.
Merging two nodes into one. Let us fix an odd r. Then the surfaces Σ cir (r, r; 1, 1) and Σ cir (2r; 1) are homeomorphic: they both have genus r and 2 boundary components. One can ask whether they are homeomorphic as surfaces with line fields. The boundary invariant on each component is equal to 2 − 2r, so we need to look at the Arf-invariant.
Proposition 4.4.1. The Arf-invariant of the form associated to the line field on Σ cir (r, r; 1, 1) is equal to 1. The Arf-invariant of the form associated to the line field on Σ cir (2r; 1) is equal to (r + 1)/2 mod 2. Hence, if r ≡ 1 mod(4) then the stacky curves C ring (r, r; 1, 1) and C ring (2r; 1) are derived equivalent.
Proof. In the case of the surface Σ cir (r, r; 1, 1) we have two collections of simple curves (α 1 , . . . , α r−1 ), (α ′ 1 , . . . , α ′ r−1 ) associated with each of the two segments where the gluing happens. In addition we have two standard curves α and β as before. So the corresponding quadratic space will be a direct sum of two copies of V r−1 together with the 2-dimensional space spanned by (α, β). Thus, the Arf-invariant is equal to 1.
Computation of Arf-invariants.
Example 4.5.1. Let V n be a Z 2 -vector space with the basis α 1 , . . . , α n , and the even pairing given by α i · α j = 1 for i = j. Let q be the unique quadratic form in Quad(V n ) such that q(α i ) = 0 for all i. First, assume that n is even. Then we claim that this pairing is nondegenerate and the Arf-invariant of q is given by A(q) = n/2 2 mod 2.
Indeed, it is enough to prove that the Gauss sum is equal to ±2 n/2 . Then the sign will determine the Arf-invariant. It is easy to see that q(x) = (−1) ( k 2 ) , where k is the number of nonzero coordinates of x. Thus, we have G(q) = n k=0 n k (−1) ( k 2 ) .
Now, let us assume that n is odd. Then the vector v 0 = n k=1 α k lies in the kernel of the pairing and q(v 0 ) = n 2 mod 2. Thus, if we assume in addition that n ≡ 1 mod 4 then we have q(v 0 ) = 0 and so the form q descends to a well-defined quadratic form q on V n−1 = V n / v 0 . We claim that its Arf-invariant is A(q) = n − 1 4 mod 2.
We have Example 4.5.2. Now let V be a Z 2 -vector space with the basis α 1 , . . . , α n , where n ≥ 4 is even, the even pairing given by the rule α i · α j = 1, i < j < i + n/2, 0, j ≥ i + n/2, and the quadratic form q in Quad(V ) such that q(α i ) = 0 for all i. Assume also that n ≡ 2 mod(3). Then we claim that the pairing is nondegenerate and A(q) = n/2 mod 2.
We will prove this by relating (V, q) with another quadratic form. For every k ≥ 0, such that k ≡ 2 mod(3), let us consider a Z 2 -vector space W k with the basis β 1 , γ 1 , . . . , β k , γ k , the even pairing given by the rule β i · β j = 1 for i = j; γ i · γ j = 1 for i = j; β i · γ j = 1 for i ≤ j; β i · γ j = 0 for i > j, and the quadratic form q k in Quad(W k ) such that q k (β i ) = q(γ i ) = 1 for all i.
First, we will prove that A(q) = A(q n/2−2 ) and then we will prove that A(q k ) = k mod 2 (4.1) To relate (V, q) with (W n/2−2 , q n/2−2 ) let us consider the 2-dimensional isotropic subspace I ⊂ V spanned by α 1 and α n . Then we have q| I ≡ 0, so the Arf-invariant of q is equal to that of the induced quadratic form on I ⊥ /I. Now setting γ i = α 2 + α 2+i , β i = α n/2+1 + α n/2+1+i , for i = 1, . . . , n/2 − 2, we get an identification of I ⊥ /I with W n/2−2 , compatible with the quadratic forms.
To prove (4.1) we use induction on k. It is easy to check that A(q 1 ) = 1 (and A(q 0 ) = 0 for trivial reasons), so it is enough to establish the formula A(q k ) = A(q k−3 ) + 1.
To this end we consider the 2-dimensional isotropic subspace J ⊂ W k spanned by β k + γ 1 and β 1 + β k + γ k . We have q k | J = 0, and our formula follows from the identification where the standard basis of W k−3 corresponds to the elements (β 2 + β 2+i mod J, γ 2 + γ 2+i mod J) 1≤i≤k−3 while a copy of W 1 spanned by β k mod J and γ k mod J.
King's College London
University of Oregon, National Research University Higher School of Economics, and Korea Institute for Advanced Study | 13,192.2 | 2018-01-19T00:00:00.000 | [
"Mathematics"
] |
On backward problem for fractional spherically symmetric diffusion equation with observation data of nonlocal type
The main target of this paper is to study a problem of recovering a spherically symmetric domain with fractional derivative from observed data of nonlocal type. This problem can be established as a new boundary value problem where a Cauchy condition is replaced with a prescribed time average of the solution. In this work, we set some of the results above existence and regularity of the mild solutions of the proposed problem in some suitable space. Next, we also show the ill-posedness of our problem in the sense of Hadamard. The regularized solution is given by the fractional Tikhonov method and convergence rate between the regularized solution and the exact solution under a priori parameter choice rule and under a posteriori parameter choice rule.
Introduction
In recent decades, the study of noninteger diffusion equations has received great attention from mathematicians around the world. These models have many applications in various types of research fields, for example, thermal diffusion in fractal domains [1] and protein dynamics [2], finance [3], systems biology [4], physics [5] and medicine [6], and besides, there are also some references as follows [7][8][9][10][11], and [12]. In this work, we consider the following problem: (1.1) Here Caputo fractional derivative D β t is defined as follows: t 0 u s (r, s) (ts) β ds, 0<β < 1, (1.2) and the source function G(r, t) ∈ L ∞ ([0, R]; r 2 ), the final data f (r) ∈ L 2 ([0, R], r 2 ) are given. Note that when the fractional order β is equal to 1, the fractional derivative D β t u(r, t) is equal to the first-order derivative du dt (see in [13]), and thus problem (1.1) reproduces the classical diffusion problem. In practice, the input data (f , G) is noisy by the observed data f ε , G ε which satisfy ff L 2 ([0.R];r 2 ) + G -G L ∞ (0,T;L 2 ([0,R];r 2 )) ≤ . (1.3) Our problem is called inverse problem and its solution is not stable. This property is called ill-posed in the sense of Hadamard. In other words, easier to understand, if is small, it will lead to large errors for the corresponding solution if using an unapproximate model for observed data f ε , G ε . The question mentioned in this paper is: Find an approximation method for the solution of the problem with noisy input data f ε , G ε . Before discussing the main results, we would like to outline a few previous papers that mentioned problem (1.1).
Our novel point in this paper is to replace the final condition (1.4) with the nonlocal condition ξ 1 u(r, T) + ξ 2 T 0 u(r, t) dt = f (r) as introduced in the last condition of our problem. This condition is proposed in the paper by Dokuchaev [29]. Very recently, Tuan and coauthors used this condition to solve some nonlocal problem, for example [30][31][32], and [33]. Motivated by this above reason, in this paper, we apply the fractional Tikhonov method to solve problem (1.1). To the best of authors' knowledge, there are not any results concerning problem (1.1). Our paper investigates problem (1.1), and the main results of this work are as follows: • We give the stability and the regularity of the mild solution.
• We show the ill-posedness and the conditional stability of solution in L 2 ([0, R]; r 2 ).
• We propose a regularized method and prove the convergence rate under a priori parameter choice rule and a posteriori parameter choice rule. Let us say that in an analytical sense, our problem seems to be more complicated than the models studied before.
This paper is organized as follows. Section 2 gives some preliminaries that are needed throughout the paper. In Sect. 3, we show the sought solution of problem (1.1), and an example describes the ill-posedness of the problem. In Sect. 4, we study the fractional Tikhonov method to solve problem (1.1) and show the convergence rate under a priori parameter choice rule and a posteriori parameter choice rule. Finally, we add the conclusion for this paper.
Proof This lemma provision can be found in [20].
Lemma 2.4
For any j ≥ 1, we have the following estimate: Proof From Lemma 2.2, we need to show that This leads to Next, due to the fact, we also get which implies that In this section, we need the solution of the direct problem of (1.1) (2.14) From [20], by using the Fourier expansion, we know that , and j 0 (y) denotes the 0th order spherical Bessel functions of the first kind. Besides, we know that {ψ j (r)} ∞ j=1 from an orthonormal basis in L 2 ([0, T], r 2 ).
The mild solution of problem (1.1)
Theorem 3.1 Let f ∈ L 2 ([0, R]; r 2 ) and G ∈ L ∞ (0, T; L 2 ([0, R]; r)). Let us further assume that Then problem (1.1) has a unique solution u given as follows: Then we get the following regularity: ) and the regularity result holds Proof From (2.15) and using the nonlocal condition in problem (1.1), we obtain that By integrating both sides from 0 to T for equation (2.15), we immediately have the following equality: From two observations (3.5) and (3.6), we get the following equality: Our next aim is to express the formula of the function in terms of two input data f and G.
In view of the nonlocal condition as in the last condition of problem (1.1) we find the following identity for the Fourier coefficient of the function : , Therefore, by taking Fourier series for the term u j (t), the formula of the mild solution to problem (1.1) can be given by . (3.9) Using the inequality (a Now, we give the regularity result of a mild solution. First of all, from Lemma 2.4, it gives (3.11) Next, let us evaluate for A i (·, t) 2 L 2 ([0,T];r 2 ) , i = 1, 4, one by one.
The fractional Tikhonov method
In this section, we apply the fractional Tikhonov method given by Morigi [35]. From now on, we denote we propose the following regularized solution with exact data (f , G): However, if the measured data {f , G} are noised by {f , G }, then we get where δ ∈ ( 1 2 ; 1] and α > 0. Noting that when δ = 1, the fractional Tikhonov method becomes a standard Tikhonov regularization. β,j (π, R, T) with Z depending on δ.
Proof The proof of lemma can be found in [36].
An a priori parameter choice rule
Let us consider the operator for ν ∈ L 2 ([0, R]; r 2 ) and 0 ≤ t ≤ T. By applying the fractional Tikhonov method, we can see that (Ts) β G(·, s), ψ j ds ψ j (r). (5.7) By choosing the regularization parameter α, the following theorem gives that the choice α is valid by using suitable assumptions. In order to give error estimate, let us assume that H s ([0,R];r 2 ) ≤ C for any s > 0, where C is a positive constant. Before going to the main theorem, we have auxiliary lemmas as follows.
An a posteriori parameter choice rule
In this subsection, considering the choice of the a posteriori regularization parameter in Morozov's discrepancy principle, [37] we choose the regularization parameter α such that
Lemma 5.5 From (5.23), if we can find that ζ is satisfied, then we have the estimate of α as follows: (5.26) We have the estimate ζ through two steps as follows, one by one.
Step 1: Estimate of X 1 , to do this, we recall , from expressions (5.2) and (5.4), and we have Step 2: Estimate of X 2 , using again the a priori bound condition of , we obtain α, i fs ≥ 1.
(5.29)
From the analytics assessment on the side, we get This yields This ends the proof of this theorem.
Conclusion
In this paper, we focus on the spherically symmetric backward time-fractional diffusion equation with the nonlocal integral condition. By using some properties of the Mittag-Leffler function, we show two results as follows. First of all, we show the properties of the well-posedness and regularity of the mild solution to this problem. Next, we present that our problem is ill-posed. In addition, we construct a regularized solution and present the convergence rate between the regularized and exact solutions by the fractional Tikhonov method under a priori parameter choice rule and under a posteriori parameter choice rule. | 2,051 | 2021-10-09T00:00:00.000 | [
"Mathematics"
] |
Performance of Concrete Mixes Containing TBM Muck as Partial Coarse Aggregate Replacements
This study investigated the potential utilization of the TBM muck obtained from the Gold Line of the Doha Metro Project as a partial replacement of coarse aggregates in concrete mixes. First, the TBM muck particles were screened to coarse aggregate standard sizes. Then, concrete mixes were prepared using 0%, 25%, 50%, and 75% TBM muck replacement of coarse aggregates. The compressive and flexural strengths were determined for all mixes at 28 and 56 days. Moreover, the results obtained were validated using EDX analysis and SEM images. A t-statistical analysis did not show a significant impact of TBM muck usage on the compressive strength results of the concrete mixes. However, another t-statistical analysis showed that TBM muck replacement of coarse aggregates had adversely affected the flexural strength results. The EDX analysis indicated the presence of Na+ ions, which can replace the Ca2+ ions in the C-S-H gel, cause discontinuities of it, and hence reduce the strength at later ages. Finally, the SEM images showed that the ettringite and carbon hydroxide (C-H) contents in the mixes with TBM muck were higher than that of the control mix, while the C-S-H gel was less in such mixes.
General
With the population and economy boom witnessed globally, the demand for new buildings, structures, and effective infrastructural systems has considerably increased. Constructing underground highway networks and railway tunnels above all is the most suitable method to adapt to such development. Instead of conventional blast and drill excavation methods, Tunnel Boring Machines (TBMs) have been recently used in many countries to excavate tunnels in hard rocks. It has been also demonstrated that the TBM technology is a feasible method for the excavation of underground research laboratories (URLs), which play important roles in the safe disposal of high-level radioactive waste (HLW) [1]. The TBM machines produce millions of tons of muck every year, which are mostly stockpiled as waste disposal.
The consumption of non-renewable natural raw materials has recently become a challenge. TBM muck use presents a viable utilization for sustainable economic and environmental developments. However, TBM muck must achieve certain physical and chemical properties to be used as a building material. Several studies investigated the potential use of TBM muck in construction applications. Few of these studies concluded that TBM could be used in many specialized building applications as an aggregate substitute with less processing durations [2,3]. Few others indicated that rock formation muck, which possesses low-to-medium brittleness and medium-to-high hardness, had the highest potential to be used in road and building construction [4][5][6][7][8][9]. Grünner et al. [6] evaluated the muck extracted from a Mismove tunnel in northern Slovakia using a set of laboratory tests including adsorption and short frost resistance tests. They concluded that muck's use must be evaluated based on its features, extraction area geology, and driving methodology. Thus, a final decision was not possible to be made on the use of this muck in different construction applications. Bellopede and Marini [7] investigated the influence of treatment on the muck that was extracted using TBM as well as conventional drilling and blast excavation methods. They observed that raw TBM muck particles are flaky in shape and therefore not suitable to replace concrete aggregate in construction applications. They observed that mobile and fixed plants-treated TBM muck by a hammer crusher and a secondary jaw crusher can reduce the muck flakiness index to conform to concrete aggregate specifications. Berdal et al. [9] investigated the use of TBM fines, with sizes ranging between 0 and 0.125 microns from two Norwegian projects, in concrete production. They concluded that TBM fines had a better effect on concrete rheology as well as better flow properties compared to natural coarse aggregates. Thalmann-Suter [4] conducted several laboratory tests such as point load, breakability, and petrographic tests to investigate the properties of the TBM muck that was extracted from the AlpTransit tunnelling project in Switzerland. They reported that the TBM muck is suitable for use as concrete aggregates after adequate preparation and special processing. Berdal [8] observed that the properties of TBM muck are largely affected by the area geology and extraction technology used. Therefore, extracted muck from each project must be studied separately to determine its appropriate processing requirements. Gertsch et al. [5] analyzed a laboratory boring sample from the Yucca Mountain project site in the USA, which was obtained by fitting five boulders into a laboratory tunnel boring machine (LTBM) and crushing them. The sample was composed of well-graded hard rocks with flaky chips and little fines. The authors concluded that TBM muck could be successfully used as a fill or sub-grade material in special construction projects, with little processing of the TBM muck. Moreover, they reported on a few tunneling projects, where their extracted TBM muck was used in construction applications, such as for fill material, road base and/or sub-base, and landscaping. However, the muck that was extracted from the Jostedal hydropower project in Norway was used in Portland cement concrete along with 30% natural gravel. The produced concrete mixes exhibited large variations in the compressive strength and concrete quality because the TBM muck lacked quality control as it was taken blindly from stockpiles without consideration of its properties. Moreover, the authors also reported that differences in the untreated and uncharacterized muck particle-grading, shape, and mechanical strength contributed to the large performance variations. Hence, the muck should be carefully characterized before using it in concrete mixes [5].
Recently, in the 14th Baltic Sea Region Geotechnical Conference, Alnuaim [10] presented several potential areas to utilize the TBM material excavated from the Riyadh Metro project. After investigating the grain size distribution, Atterberg limits, compaction test, and direct shear and permeability tests, the author concluded that utilizing the TBM excavated material in the asphalt mix designs or as a subbase for road construction requires some processing. He also mentioned that the TBM excavated material should be treated on site using mobile crusher equipment in order to be used as a non-structure concrete. However, using the TBM excavated material for backfilling was not demonstrated due to its high fine content and low hydraulic conductivity. Finally, the potential use of such material as a landfill liner was proposed with the requirement of additional sample crushing and the addition of some fine material such as bentonite. Voit and Kuschel [11], as well as and Voit et al. [12] examined the quality of the rock excavated from the longest underground railway line in the world, namely the Brenner Base Tunnel, and the treatment processes required to accomplish a high-quality level of recycled aggregates from such excavated material. After extensive research on the rock quality and concrete mix design, as well as on the appropriate implementation of the processing techniques and concrete mixing plants, they concluded that calcareous schists rock could be successfully recycled and processed as aggregate for shotcrete, structural, and inner-lining concrete, and as filter gravel. However, the quartz phyllite rock was found unsuitable as a concrete aggregate substitute due to intense foliation and mica-rich mineral composition.
Doha Metro Project: Background and Previous Studies
The Doha Metro project may represent the latest achievement in the field of TBM tunneling. It was entered into the Guinness Book of Records [13] due to its rapid construction, utmost safety considerations, and excellent quality standards. The Qatar Railway company constructed, using TBMs, 111 km of metro tunnels in less than two years. The Doha Metro tunnels included the construction, in several phases, of four main lines, namely Gold, Red, Green, and Blue. In the first phase, twenty-one TBMs were designed and their operations started in 2020. The second phase, which includes the excavation and construction of the Blue Line, is estimated to be completed in 2026. The excavated TBM materials of the first phase tunnels, which were about five million tons, were stored in four logistical areas in Doha. These areas were designated as MLPA South, MLPA West, MLPA North, and MLPA Al-Rifaa to differentiate between the characteristics of the TBM materials extracted from each area. Abu-Taqa et al. [14] studied the potential use of the TBM muck generated from Doha's Metro Gold Line that was stockpiled at MLPA AL-Riffa in various construction applications. Laboratory tests were conducted on raw TBM material and the results obtained were compared with the specification limits established in the Qatari Construction Standards (QCS 2014) [15]. The raw TBM muck gradation complied with the QCS 2014 requirements for a fill material under buildings or road subgrades. However, the liquid limit and plasticity index of TBM muck were higher than the permissible limits. Conversely, TBM muck was not suitable as a concrete coarse aggregate substitute because its gradation did not meet the QCS 2014 requirements. Hence, physical processing is needed for TBM muck before using it as a concrete coarse aggregate substitute. The results obtained also showed that the raw TBM complies with the requirements of coarse aggregates, except for the water absorption and acid-soluble sulphate. Moreover, the results showed the presence of silica (Si) as a predominant component in TBM muck, which may cause higher Coefficient of Thermal Expansion (CTE) values. Consequently, this leads to a reduction in the compressive strength of the concrete mixes prepared using muck.
Research Objectives and Significance
As the properties of TBM muck are highly dependent on the geology of the tunneled area and the chemicals used in the TBMs facilitate the excavation, this study aims specifically at investigating the potential use of the Doha Metro Gold line TBM muck stockpiled at MLPA Al-Riffa as a partial or full replacement for concrete aggregates in Portland cement concrete (PCC). The properties of such TBM material have been investigated and reported by the authors in a previous work [14]. The results of this study may be applied for concrete mixtures prepared with TBM material that are extracted from different areas but have similar properties of Doha Metro's TBM. It is worth noting that the raw TBM muck was subjected to physical modification (sieving) only to obtain sizes between 10 mm and 20 mm, and between 4 mm and 10 mm as per the QCS 2014 coarse aggregate requirements. The gabbro coarse aggregates used in the control mix were replaced with the raw TBM muck at percentages of 0%, 25%, 50%, and 75%. The fresh and hardened properties of the mixes were tested and compared to those of the control mix. It is worth noting that the hardened properties investigated in this study were limited to the compressive and flexure strengths. The elasticity and ductility should be also investigated by defining the stress-strain plots and future work should be directed towards this end. The results were used to decide the optimum TBM muck replacement percentage that may be utilized without a significant loss of the mechanical properties. The demonstration of the possibility of utilizing the TBM muck from the Doha Metro project excavations as well as using it as an alternative local aggregate, even partially, will lead to significant cost savings involved in importing aggregates from other countries, alongside environmental benefits related to the disposal of the stockpiled muck.
Experimental Program
Four mixes were prepared by replacing 0%, 25%, 50%, and 75% of gabbro coarse aggregates with TBM muck. The workability was determined for all mixes using the slump test. The mechanical properties were determined for all the mixes using flexural and compressive strength tests. The test results for TBM muck mixes were compared to those of the control mix (without TBM muck). Moreover, the fractured surfaces of the strongest samples of each mix were examined using SEM and EDX to understand their microstructures. Finally, the mechanical properties results were statically analyzed using t-tests.
Materials
Generally, Portland cement (CEM I) Class 42·5 R complying with EN 197-1 [16], natural-washed sand conforming to BS EN 12620 [17] (0-4 mm diameter), and tap water were used to prepare all concrete mixes. Gabbro aggregates and TBM muck of 4-10 mm and 10-20 mm diameters were used as coarse aggregates. The TBM muck, which was used in this study, was extracted from the Gold Line of the Doha Metro project. It was thoroughly investigated by the authors in an earlier study [14]. The TBM muck gradation was found to comply with the QCS 2014 [15] requirements for concrete coarse aggregates. Hence, the muck was classified into the acceptable sizes for coarse aggregates in concrete mixtures (10-20 mm and 4-10 mm). Table 1 summarizes the physical properties of TBM muck as concrete coarse aggregates [14]. Table 1 shows that many physical properties of TBM muck did not meet aggregate acceptable limits. Nevertheless, TBM muck was used herein to partially replace gabbro coarse aggregates to investigate the effect of such replacement on concrete's strength. Its high-water absorption value may be the limiting factor for its use in concrete mixes. The high-water absorption and loss of magnesium sulphate soundness values of TBM muck may affect the frost and chemical resistance of the concrete mixture. However, this study was limited to the investigation of the strength of the TBM concrete mixture and future studies should be directed towards investigating other mixture properties, such as the freeze-thaw behavior, etc. To avoid concrete mix segregation and other associated concerns, a high-range water-reducer (HRWR), specifically CHRYSO Delta CQ 25, which conforms to ASTM C 494 Type-A, F and G [28], and BS EN 934-2 [29] standards, was incorporated into the mixes in various dosages based on the TBM muck replacement ratio. It is worth noting that cost of the HRWR used in this study is responsible compared to other similar products. As per the supplier, one liter of CHRYSO Delta CQ 25 costs USD 0.75, which is competitive among others. As per the manufacturer's recommendations, the HRWR's dosage shall not exceed the permissible one stated in the data sheet to avoid its adverse effect on fresh or hardened concrete properties. Table 2 summarizes the super-plasticizer's properties and its recommended dosage.
Mixture Composition, Mixing, and Sample Preparation Procedures
Four different concrete mixes were designed to achieve a 28-day target cubic compressive strength of 40 MPa (approximately 32 MPa for cylindrical strength) with 0%, 25%, 50%, and 75% replacement of gabbro aggregates with TBM muck. As per Al-Ansary and Iyengar [30], the relative density of the imported gabbro (which is normally used for concrete production in Qatar) is about 2.8-2.9 and as per Table 1, the relative density of the TBM muck is about 2.75-2.8. Hence, the TBM replacement percentages based on the aggregate weight rather than volume could be feasible. The control mix (without TBM muck) had a water/cement ratio of 0.43 and a HRWR's content of 0.82 L/100 kg of cementitious material. As previously noted and due to its high-water absorption, the water demand of the TBM muck mixes was higher than the control ones. However, QCS 2014 [15] limited the water/cement ratio to 0.5 for such a concrete grade. Hence, this value was not exceeded in all TBM muck mixes and the HRWR's dosage was selected accordingly. The slump flow was checked immediately after mixing according to BS EN 12350-8 to achieve a target value of 160 ± 40 mm [31]. The HRWR's dosage was adjusted gradually for the mixes with a slump value lower than the target one and the mix was repeated until achieving the target's slump value. Table 3 presents the mix design and slump value for each mix. It should be noted that the HRWR's dosages in the mixes containing 50% and 75% TBM muck replacement exceeded the manufacturer's recommended values (i.e., 0.5 to 2.5 L per 100 kg of total cementitious material) to achieve the workability of the mix. The effect of this over-dosage on the hardened concrete properties should be noted and investigated. Immediately after mixing, the slump flow value was measured according to BS EN 12350-8 [31]. Then, the specimens needed for the strength testing were cast. For each mix, 10 standard cylinders, according to ASTM C470 [32] (i.e.,152 mm in diameter and 304 mm in length), and 10 plain concrete beams with dimensions of 100 mm × 100 mm × 500 mm were casted. The specimens were demolded after 1 day and kept in the curing tanks until the day of testing.
Compressive and Flexure Strength Testing
Half of the samples were tested after 28 days of moist curing and the other half were tested after 56 days. The compressive and flexural strength testing of the samples was carried out according to ASTM C39 [33] and ASTM C78 [34], respectively. Then, t-statistical tests were used to determine whether there were any statistical differences between the strengths of the control mix and those containing the TBM muck.
Microstructural Analysis
From each mix, the fracture surface samples were collected from the strongest tested beam specimens. Scanning electron microscopy (SEM) and energy dispersive X-ray analysis (EDX) were used to investigate the microstructure and element composition, respectively, which may infer the mineralogy of the samples. It is worth noting that the EDX analysis was used herein to investigate the minerals present in the sample, which helps in justifying the strength behavior and other mixture properties (for example, thermal expansion). The SEM imaging started with vacuum drying of the samples followed by gold palladium coating in order to enhance the image resolution. A low-energy secondary and back-scattered electron were used to capture the SEM images at an accelerating voltage of 20 kV. The SEM images and EDX results were compared between the mixes.
Compressive Strength
The cylindrical compressive strength results at 28 and 56 days are presented in Table 4. Table 4 shows that the compressive strength of the control mix was slightly enhanced by incorporating 25% of TBM muck. Strength increases of 4.8% and 6.1% were determined at 28 and 56 days, respectively. However, the compressive strengths for the mixes containing 50% TBM muck decreased by 10.4% and 11.5% at 28 and 56 days, respectively. Moreover, the compressive strengths for the mixes containing 75% TBM muck decreased by 16.5% and 14.4% at 28 and 56 days, respectively.
Additionally, t-statistical tests were used to draw a more accurate conclusion about the measured compressive strength results. The test null hypothesis assumed that the average strength values of both mixes are equal (i.e., µ control mix = µ s (TBM mix) ) and a two-tailed significance level of 0.05 (α = 0.05) was considered. The null hypothesis should be rejected if the t-statistic is larger than or equal to the critical t-test value. Table 5 shows the t-statistical analysis and the critical t-test values for the compressive strengths of all mixes at 28 and 56 days. It is worth noting that in the t-statistical test, the degree of freedom is dependent on the standard deviation and hence different degrees of freedom could be observed for the same sample size depending on the standard deviation of the results. Table 5 shows that the enhancement of the compressive strength presented in Table 4 for Mix B (containing 25% TBM muck replacement) is statically insignificant. It shows that the reduction of the compressive strength in Mix C containing 50% TBM muck is also not significant. However, the reduction of the compressive strength due to the replacement of 75% of the gabbro aggregates with TBM muck is statistically significant.
The failure modes of the strongest specimen from each mix are presented in Figure 1a-d. As reported by Xu and Cai [35], the crack pattern of the specimen under uniaxial compression depends on its microstructure and stress state. The fracture patterns shown in Figure 1a-d are similar to those specified in ASTM C39 [33] for the cylinders under uniaxial compressive loads. The figures show that the samples from Mix D (containing 75% TBM muck replacement) split vertically after failure, which indicates a brittle failure under uniaxial compression. However, the cylinders of the control Mix A, Mix B, and Mix C (which contained 0%, 25%, and 50% TBM muck replacement, respectively) failed along inclined shearing planes, which is considered more of a ductile shear failure rather than a brittle splitting one. It could be also concluded that as the TBM muck content increased above 50% of the weight of the coarse aggregates in the mix, it tended to have a brittle failure pattern under uniaxial compression loads and hence no apparent deformation may be witnessed before fracture. This may indicate a degradation in the mixture's ductility. Table 6 presents the flexural strength results for the samples at 28 and 56 days. The results showed that the flexural strength decreased due to the partial replacement of coarse aggregates with TBM muck, even for small replacement percentages. This finding does not correspond with the increased compressive strength results for Mix B containing 25% TBM muck as a coarse aggregate replacement. Table 6 shows that the flexural strength decreased by 9% and 21.2% at 28 days and by 15% and 24.2% at 56 days for Mixes B and C, respectively. However, the flexural strength decreased by only 3.4% at 28 days and 17.3% at 56 days for Mix D containing 75% TBM muck replacement. The muck material's heterogeneity may be one reason for these inconsistent results. However, the analysis of the fractured samples' microstructures will be analyzed in the following section to draw a better explanation.
Flexural Strength
A t-statistical analysis was also carried out to confirm the extent of the significance of the flexural strength reduction. Table 7 summarizes the t-statistical analysis for all mixes at 28 and 56 days. Table 7 shows that the reduction in the 28th day of the flexural strength may be considered significant for Mix C containing 50% TBM muck. However, the reduction in the flexural strength becomes significant for all mixes at 56 days. Thus, replacing concrete coarse aggregates with untreated muck, even with small percentages, may adversely affect the flexural strength, especially in the long-term.
Microstructural Analysis
The EDX analysis of the fractured surfaces for the strongest beam samples at 28 days defines their element composition, which may infer the mineralogy as shown in Figures 2-5. Moreover, the SEM images were taken from the fractured surfaces (scale of 2 µm) to investigate their microstructures (Figure 6a-d). The EDX graphs show that Mixes B, C, and D contained sodium ions (Na + ) in contrast with the control Mix A. Moreover, the calcium (Ca ++ ) ions' content in Mixes B, C, and D were less than the Ca ++ content in Mix A. The muck Na + ions' source may be the rock, which is being excavated, or the chemical additives used to facilitate the excavation process. The presence of sodium (Na + ) ions in the mix may justify the strength's loss as witnessed in the samples containing TBM muck, especially in the long-term. Wang et al. [36] studied the effects of sodium bicarbonate and sodium carbonate as concrete accelerators on the hydration process and on the Portland cement paste hardened properties. They reported that both accelerators, with an optimum concentration of 1% by weight of the Portland cement, may shorten the paste's setting time and increase the early strength without having a detrimental effect on its long-term strength. However, an increase of both accelerators above 1% by weight could decrease the paste's long-term strength. The strength reduction may be attributed to the fact that Na + ions may replace the Ca ++ in the C-S-H gel at a later age, causing discontinuities of the gel and hence reducing the strength. This conclusion may be also applicable for the mixes containing TBM muck due to the presence of the Na + in the muck material and justifies the significant strength reduction at 56 days.
The SEM results for all mixes at 28 days are shown in Figure 6a-d. The images show that the needle-shaped ettringite and carbon hydroxide (C-H) contents in the TBM muck mixes were higher than those in the control mix. It is worth mentioning that the C-S-H gel in the TBM muck mixes was less than that in the control mix. While the early age strength is mainly influenced by both ettringite and the C-S-H gel, the later age strength is mainly influenced by the C-S-H gel's microstructure. These results agree with the flexural strength results at a later age, as presented in Table 5.
To reduce their adverse effects on the C-S-H microstructure and mixture strength, the Na + content must be reduced in TBM muck when used in concrete. The use of pozzolanic materials such as fly ash, silica fume, etc., may be useful. Further research work will be needed using a stabilization technique.
Conclusions
This study investigated the potential use of TBM muck obtained from the Doha Metro project's Gold Line as a partial replacement for gabbro aggregates. In addition to the control mix (without TBM muck), TBM muck mixes with 25%, 50%, and 75% replacement of coarse aggregates were prepared. The slump test was performed on all mixes to ensure good workability. The compressive and flexural strengths were determined for all mixes at 28 and 56 days, and were compared to those of the control mix. Moreover, EDX analysis and SEM images were taken in order to study the microstructure and element composition, which may infer the mineralogy of the mixes and analyze the results. The following conclusions can be drawn:
•
Due to TBM muck's high-water absorption, the amount of water needed to achieve a sufficient mixture's workability was increased. Hence, a high-range water-reducer (HRWR) was added with different dosages to achieve standard slump values. Increasing the HRWR dosages above the recommended limits could affect the strain gain.
•
The compressive strength was slightly better for the 25% TBM muck mix in comparison with the control at 28 and 56 days. However, the t-statistical analysis showed that this improvement was statically insignificant.
•
Increasing the TBM muck's content beyond 25% reduced the compressive strength of the mixes at 28 and 56 days. This reduction was significant for the 75% TBM muck mix.
•
The failure modes under uniaxial loads showed that by increasing TBM muck content above 50%, the samples tend to exhibit columnar brittle failure and hence no apparent deformation may be witnessed before fracture. Accordingly, the mixture's ductility may be decreased by increasing the TBM content above 50% of coarse aggregates.
•
The use of TBM muck in concrete mixtures had an adverse effect on the flexural strength, especially in later ages. The t-statistical analysis showed that the flexural strength reductions in TBM muck concrete mixes were statically significant and should be taken into consideration.
•
The EDX analysis showed that the mixes prepared with TBM muck contained Na + ions, which may be the reason for the strength decay observed, especially at later ages. Thus, it is recommended to investigate the use of fly ash, silica fume, etc., in TBM muck-concrete mixes.
•
The SEM images showed that the ettringite and carbon hydroxide (C-H) contents in TBM muck mixes were higher than those in the control mix, while the C-S-H gel in the TBM muck mixes was less. This may justify the reduction of the flexural strength in TBM muck-concrete mixes.
According to the aforementioned observations, it can be concluded that the TBM muck could be utilized into concrete mixtures without a significant loss of strength; however, physical processing and chemical stabilization may be needed to enhance the properties of the TBM muck and to avoid the adverse effect of its composition on the mechanical properties of concrete mixtures. This should be the subject for future study to demonstrate the potential utilization of TBM muck for construction applications.
Data Availability Statement:
The data used to support the findings of this study are available upon request from the corresponding author. | 6,454 | 2021-10-21T00:00:00.000 | [
"Engineering",
"Materials Science"
] |
Investigating the Place of Imagination in Farabi ’ s Epistemological Theory
The issue of the role of imagination and its importance and position in Islamic epistemological structure, especially for Farabi and advocates of Gnostic approach, has established the chance to achieve a theoretical framework in understanding Islamic epistemology. For Gnostic approach advocates, wisdom and truth have appearance, essence, perception, and intuition in a way that imagination as a means of abstraction of objects and shapes from their material reality has an active and decisive role in creation of wisdom and truth, it can also interpret human conscience and inner being; therefore, as a theoretical means for thought and analysis, the basis for imaginative reaction is to symbolize and interpret. From this perspective, every sensory experience and awareness necessarily contains a certain level of imaginative nature and quality. There is a special type of dualism in Farabi’s approach and imagination which is related to his interpretation of religion and religious reality. He believes that religious truth is a metaphorical, artistic, literary, and imaginative truth. For him, human possess another unique value which is the value he generally possesses towards art and aesthetic work. Another important point according to Farabi is the difference between philosophy and religion. According to him, philosophy inclines towards sense and experience while religion deals with intuition, predictions, and speculation. In intuition, philosophy supports necessity whereas philosophy supports essence, and this is what causes to think that these two are of different types. Due to their differences, specific transverses have appeared in them. This division cannot be solved except for Farabi’s philosophy and his epistemology theory which is based on the unity between nature and essence (from the beginning) and totally deals with maintenance of balance. Farabi’s epistemology theory in imagination is a philosophical-rational theory that is based on logical credits and rational preliminaries and bold dependence on the relevant results. From this perspective, Farabi was creative in philosophical speech and has exceeded the words of Plato and Aristotle, and through imagination his theories proposed a new critical epistemology in associations like religion and philosophy, art and science, and wisdom and essence, which was effective in the history of philosophy after him.
Introduction and Background
Islamic scholars have seldom paid attention to imagination and its position in the system of Islamic epistemology.Position of imagination in Islamic philosophy and especially in Islamic epistemology and specifically for Aristotle and the strong position of rational reason, imagination is an issue that scientific communities of Islamic countries have never paid attention to.Henry Carbon, Nasrullah Hekmat, William Chitik, Maryam Sane'pour, Rashideh Kalla', and Hazim al-Ghartanji are among scholars who have paid limited attention to the role of imagination in mystics, rhetoric, and Islamic critic.An important point in these studies is the relationship between imagination and theoretical mysticism and illuminative approach and transcendental philosophy and the relationship between Sahrvardi and Ibn Arabi and Sadr al-Din Shirazi.
The present mainly study deals with Farabi's effect on Islamic philosophers' approach and Islamic epistemology.Farabi's epistemology in imagination is a philosophical-rational theory and relies on rational credits, logical preliminaries, and bold submission to its relevant results.In fact, the position of logic for him and his position in logic are noteworthy.Another important point is that Farabi did not consider logic as a mere entry; however, he thought of it as a basis for philosophical speech in metaphysical framework.In this regard, he is like the American philosopher Josiah Royce who preceded him centuries.Royce (1855Royce ( -1916) ) is neo-Hegelian and tried to unite idealism and pragmatism and philosophy and religion.He considered logic as a necessary means to achieve this significant goal (Royce, 2009, p. 215).
Investigation into this issue and the approach of queries adopts mechanisms of analysis and interpretation, and through them and from the perspective of epistemology it tries to examine Farabi's philosophical and epistemology in a framework in order to reveal the importance of imagination in Farabi's epistemology and his role in creation of the new concept of epistemology (imaginative epistemology).The present study was conducted in order to explore this theory in the structures of epistemology and ontology which have a metaphysically tight relationship.
Ultimate Division as Unity
In regard with the issue of essence and existence in Farabi's definition, there are some questions such as whether he believes that essence and origin have priority over existence or vice versa?Some believe that Farabi is the first philosopher that paid attention to this issue and tried to solve it.Sadraldin Shirazi later stated this issue first in regard with the contradiction and second tendency toward existentialism (Khayyereh, 2013, p. 48).Now, the question is that whether Farabi believes in existentialism or essentialism (al-Farabi, 1988, p. 74) or believes in something other than these two.
Farabi is the first one that has considered the theory of unity of existence and essence self-existent.However, stating for the first one that there is no essence other than existential essence is the same as claiming that existence is truth if it possesses the necessary quality which is existentialism (al-Farabi, 1988, p. 42).
This short statement refers to the point that essence is existence and there is no essence other than existence; therefore, the two are the same.Therefore, the first creature is the same essence (self) of existence.And since the real existence for every creature is not anything other than the specific existence that is outside the soul, all real creatures are in fact a grant from Him and a sign for His presence; therefore, their existence is from Him, which is and existence for the way of imagination because it is actually nothing outside His soul.So whatever is originated from self-existence is merely because of their belonging to Him, and the rational images are originated from His existence (al-Farabi, 1988, p. 44).This self-existence is reason, rational, and reasonable, and only His existence (He Himself) thinks because there is nothing but Him such that there is no creating existence other than reasonability for His own essence (al-Farabi, 1988, p. 45).
Therefore, the belief that one of them is prior to the other one is not compatible with Farabi's philosophy that is mainly based on unity between the two, and what is important is the essence of imagination.
According to Farabi, thought in response to "What is this?" is called essence since it is reasonable, and in regard with its existence outside, it is called reality (Jahami, 1998, pp. 862-863).Correctness of this reality depends on the compatibility and creative power of imagination and negation of realization barriers, among which one can brefer to contradiction between imaginations and contradiction of imagination companying with movement (al-Farabi, 1988, pp. 57-58).
The fundamental problem that Farabi highlights clearly out of the statement, existence unity and essence in necessity, is that this issue can hardly be abstracted because His reasonability is the same necessity and essence.
According to Farabi, what is noteworthy is that active intellect gets involved with creation by affecting the initial or the common material and gives images to this material, and by imposing and assuming reasons for it, the intellect domination over them will be possible.However, the second ones have the same appearance or due to their relative simplicity, they are all given one appearance.An instance is celestial bodies whose common essence or quality is their initial common material for all creatures existing in the universe (al-Badavi, 2006, p. 106, Vol. 2).
The appearance of Farabi's statement indicates that celestial intellects, souls, and active intellect do not turn into imagination neither into sense; however, they inspire it.Although imagination and thinking are different, they are one thing; however, they are different from one intellect to another and from one soul to another.Therefore, where there is imagination, there exists creation, and this imagination is unique to intellect which is interpreted as blessing and is merely an abstraction from internal senses of cognitive activity.Farabi states, "The stars are imagined in the form of slight movement, and movement does not reach them, and these bodies are bound to movement.So you think of what is created out of this movement and understand it, and you do not understand what is created other than the movement.And if one could imagine other than the movement, it would be necessary that two movements be created; however, it is impossible and those bodies do not imagine impossibility although it is not fake.
On the other hand, Farabi has founded a thought that intuition or sensory observation is reflection or intuition, and imagination intellect is the abode for and surrounds both and is like the initial or abstracted material.Farabi has really affected these two a lot and there is no doubt that he mastered over them and took different ways through them because environment is a container for them.In fact, this activity is close to Farabi's understanding of gender tie as male and female.Female is the same female or a power that material provides and male is the same male or material that gives it an appearance that provides the power (Farabi, 1986, p. 97).
In the 19 th chapter of his book, "Views of utopia residents", Farabi focuses on sequence of images on the main material.He states that existence is in fact a total unity and soul, powers and talents are a stage of existence, the most inferior of the lowest material for the highest level, and the highest or most valuable for the most inferior.Aristotle's statements and approaches have sometimes affected Muslims.Therefore, he is the first one that has stated, "Soul is the initial completion of the body with living powers (Badavi, 1980, p. 30), whereby he has joined the soul with the body firmly, and if it is separated from soul, the body will not exist.Therefore, if we imagine a gradual ladder for creatures, all of them are created from the main material and appearance.
On Soul and Intellect
The second or average level, understanding and awareness, is the firmest and most reasonable one.The researcher tries to name this level best which is speech on imagination, orders, and secrets, and belongs to all troublesome philosophical issues that Farabi encountered with and attempted to solve and imagined the essence of his philosophy and is based on his imaginative logic.Therefore, despite of many similar aspects, his logic is Aristotelian imaginative and not Platonic dialectic one, and in fact, it is a new type of both.In this regard, Farabi is completely like Hegel.His uniting philosophy does not bring about unity, to an extent that the solver and the creation activity are new, phenomenology is specific, in which all philosophical and religious heritages affect one another.His philosophy is not the collection of these to philosophers, neither the unity between religion and philosophy or philosophy and literature and art and music and immateriality of mathematics, but it is gnostic, in which divine is tied to human, intellectual to material, social to individual, subjective to objective, and imaginary to reality.In his philosophy, Gnosticism is tied with a type of trust, reasonable with unreasonable, real empirical materialism with idealism, and mysticism with intuition and traditionalism, in a way that it later included all structures and studies especially schools of Islamic and human thought.In his philosophy, his personal surrounding and understanding of myths form, and there are studies on his methods, resources, and cognition in oral speech.This philosopher has considered it as an acquired reality like human as an incomplete form.Farabi is named "the second teacher" because of his superiority in logic and rank of knowledge writings (Nasr, 2003, p. 431).Due to his superiority in writing issues and their relationship, existence and creature and their ranks, his ponder in meaning of words and the boundary of objects and group relationship of them with each other, and the properties of studies, and types of speech, serious issues are brought forth in this type of research in the field of science philosophy and epistemology.All of these levels of research have a direct relationship with logic in accordance with Farabi's concept of these sciences which are bound to obtaining prosperity and goodwill.
The present study focuses on reason and imagination or soul and power in order not to digress from the issue and deal with different schools and ways.Therefore, these are the main issues that the present study deals with.Muslims have been affected by the definitions and approaches about soul proposed Greeks, especially Aristotle and Plato.For these two philosophers, the issue of soul has a close relationship with wisdom.Plato's theory about reminder is clear, and some scholars believe that in Aristotelian religion, there is a close relationship between soul and rejection of this theory by him and the world of forms.He understood the origin of the theory of forms by Plato.Therefore, he decided not to accept the foundation on which Plato's theory was based, which is presented i the theory of immaterial or spirituality of the soul (Ghasem, 2002, pp. 70-71).
The present study is aimed at resolving the severe ambiguity in how some Muslims deal with these two philosophers, which has caused religious sensitivity and confusion of temporary Islamic philosophers and jurists (Badavi, 1980, p. 12).This confusion causes misunderstanding their ideas such that Farabi is accused of making contradicting comments about soul and intellect, and that he is doubtful about Aristotle and Plato and has made many strange comments (Ghasem, 2002, pp. 74-75).
Farabi states, "Soul is the first perfection for the body and possesses power and life" (Ghasem, 2002, p. 73).Soul is the created form within the body and completely belongs to it and to the material of the initial material (Ghasem, 2002, p. 75).Therefore, soul is a necessity for individuals while it disappears as the body dies.The initial gems that individuals hold inside their bodies do not need anything except for soul; however, secondary gems include all generalities that individuals need inside themselves.Therefore, individuals have priority over the so-called general gems (al-Farabi, 1987, p. 89).However, generalities are constant.Moreover, abstract intellectual concepts after destruction of individuals and bodies, generalities will sustain; therefore, generalities under the label of individual gems have priority (Farabi, 1987, p. 89) because reasonability in the same abstract existence of that thing (Farabi, 1988, p. 46).About human and intellect, Farabi states, "It is true that a child has a soul that is aware of its power and has senses that are tools to understanding details out of generalities which in turn are experiences, and when soul forms out of these experiences, the soul becomes wise because wisdom is not other than experiences, and with more experiences, the soul will becomes wiser (Farabi, 1986, p. 98).Therefore, what actualizes the child's world wisdom and soul is specifically the same works accomplished by cognitive tools that are senses and experiences.Therefore, reason has no appearance and cannot be defined or specified and does not have nature; however, it turns into forms.Therefore, reason does not naturally have any form, but if a form appears, there will be a barrier between it and external forms (Badavi, 1980, p. 20).Farabi also emphasizes that for reason and soul to be acquirable is not anything other than experiences.Is there any other activity that meddles in this activity?Does anything else other than experiences interfere with mental and intellectual growth of a child?Farabi believes that there are two types of experiences; those obtained on purpose and those acquired without any intention.Scholars name experiences obtained without intention the preliminary knowledge, reason basis, and things like that, which are kinds of knowledge that most people doubt about they have always existed in the soul (Farabi, 1986, p. 99).So when these aimless experiences are acquired, most people are confused and doubt that they have always existed inside the soul and are the beginning of knowledge and basis for reason.The soul becomes wise through action because wisdom is not other than experiences (Farabi, 1986, p. 99).So what make wisdom become actualized is merely the experiences that are acquired without any intentions, and most people think that they have always existed in the soul; therefore, the thought that there are initial knowledge or forms or concepts or principles before experiences is not more than a delusion.
Under another topic, Farabi talks about active intellect which causes forms and negates and orders things that exists in the universe, a wisdom that is originated from the forms of creatures, including solids, plants, and animals, or from human soul.He is fame for his theory about the mediation between active intellect and its relationship with prophet and philosopher does not need to be mentioned.
Although the first hypothesis is in line with the form of some statements by Farabi, especially those made in "Fosus al-Hikam" and with the appearance of religious thought; however, it is in contradiction with Farabi's logic and thinking system which puts emphasis on a kind of phenomenalism and believes that human does not know the reality of the object because his basis for his knowledge is his sense, then he distinguishes between similarities and dissimilarities with his reason, whereby he knows some of his equipment, internalities, and qualities through which he gradually approaches a limited knowledge about his achievements (Jahami, 1998, p. 124).
On Transcendental Intellects
The concept of transcendental realities is the fundamental problem and challenge that sometimes Farabi is faced with.Islamic philosophers know these realities as intellects each of which is active compared to its subordinate and inactive compared to its superior.An example is the active intellect dominating the things that exists in the universe, and this is the same challenge that interprets what Farabi states about human soul or intellect.In this regard, Eskandar Afroudisi, the first interpreter of Aristotle, believes that active determining intellect that Aristotle talks about its sustainability is holy and pure God.However, Ibn Roshd, his greatest interpreter, believes in eternality of human soul and what Aristotle believed about intellect is because of this eternality, he also believes that active intellect is the only essence of human soul (Badavi, 1980, pp. 5-10).
However, based on his theory of blessing, Farabi believes that the foundation of existence and the comprehensive essential reality in which existence and essence unite is reason.This theory or interpretation that brings the universe and corruption under its 10 disciplines and systems according to a numerical system, and this is the interpretation that has more agreement with Farabi's logic and his thinking system.Therefore, intellect or reason, existence on metaphysical level is the same rules or initial categories and necessary eternal principles in which corruption and change cannot occur; therefore, these are its origin, and human does not achieve them intentionally, but he obtains them through unintentional experience along with sensory understanding.And, here Farabi interprets them as unintentional or urgent experiences.
Farabi's attempts in this regard caused the two philosophers and philosophy and religion to unite, and his pondering on reason caused a third theory to be created, in which there is a kind of similarity and closeness between Farabi's concept of form and Hegel's temper and synthesis.For Farabi, form is basically derived from the relationship and dialect between the two; necessity and refusal, existence and absence.In the reason, there is no special reaction to sense except for understanding of all objects and opposition and imagination of creatures (Farabi, 1986, p. 99).Moreover, Farabi believes that sense understands an object; the external form of the object; however, the reason understands what sense perceives and its opposite (Farabi, 1986, p. 99).This dialectic essence accompanies awareness and imagination of an object and its opposite, which is both a reaction and a fantasy and imagination.It prepares the situation for thinking and reasoning.However, what makes this activity certain is confirmation of rejection of the form, which is the first principle of existence and is called the perceived necessity through intuition and sensory experience, which cannot be separated from existence.The same present existence principle in this creature and with its existence will gain the credit to pass for its abstracted form and will be sensible after it has separated, which is accomplished with active intellect, i.e. the imagination that causes this form.In this conditions, the form will be abstracted from its imaginary world, and all imagination and abstractions and in fact the reason will be the same principle for prediction of abstract motivators.
Farabi is the first believer in the theory of existence unity in the system of Islamic thinking.After him, no Muslim philosopher or mystic, in spite of their intellectual features, have not freed from this theory.This religion is in the unity of intellectual schools and the conclusion is that intellect is the origin.Therefore, the principles of transcendent single abstraction and then the abstraction principle or the principle of systematic abstraction is based on imagination and whatever is reasonable, to the highest level, is imaginary, on which revealed imagination has influenced, which means the structure of existence on the level of awareness is an imaginary one and it is impossible to neglect imagination in the presence of the reasonable.
Human and His Understanding Structure from Farabi's Perspective
Human is one of the creatures whose perfection was not bestowed to him from the beginning.He is one of the creatures that have given the most deficient perfection.Meanwhile, he has given some principles whereby he tries and progresses in order to reach perfection through instinct and nature or on his own intention (Jahami, 1998, p. 124).
With these definitions, human is a combination of instinct and intention and he specifically approaches perfection, and he uses the principle to reach perfection (Ibid,124).In this regard, there are two types of wisdom; the first type and the created necessary rules for him without intention and along with tendency toward generalization, and this is what can be emphasized in tastes an cannot be separated from it.In knowledge and intention and beliefs and legal and Sharia rules and civic interactions, the emphasis is on the whole time of induction of details like the principle that any stone sinks in the water and some may float (Farabi, 1986, p. 82).Therefore, any intention of understanding objects a human bears in mind is caused by awareness of the situations and strictness of that situation, and this is only the demand of the object that exists inside his soul (Farabi, 1986, p. 99).It seems that he seeks a kind of agreement with his soul within the unity between what exists inside him and what he finds outside and by combining the condition he possesses and what is there in his soul, which means the cognition activity is the activity of compromising with soul within compromising with reality.This is a kind of activity is generalization, comparison, analogy, and a kind of real remembrance; however, it is different from Platonic ideal remembrance because observation and experience are the basis for knowledge.Therefore, the apparent five senses and their perception belong to the power of senses, which is the same place where intellect and material meet and sense is only choosing the existing form, and this is the simplest level of understanding which forms a type of understanding or sense, through which human understands, sees, hears, and touches himself and this is a type of internal perception.
According to Farabi, when human was created, the first thing that forms inside him is his sensory part along which there is a tendency toward something that he feels, then another power maintains the opposite of the felt things, which is called imagination power that combines sensible things together and separates from one another.It accompanies with different combinations and separations, truth and fakeness; however, it keeps a tendency toward things he imagines (Farabi, 1986, p. 87).And there is a distinction between sensible things for the common sense, like distinction between white and black in visionary understanding.Therefore, the distinction power is used for different senses and coordination among them by use of law of association, and also understanding the form, continuum, place, amount, motion, volume, and distance play an important role in evolution of sensory knowledge and acquisition of experiences in order to understand common sensible things.Afterwards, the role of imagination will be put forth, and imagination power catches the effect of common sense and welcomes it, and after it hides, imagination will understand it, and sometimes it is an invention in agreement or disagreement with the truth, and in this case imagination power is submissive to intention and willing, and sometimes it occurs without intention as happens in sleep and daydreams.Afterwards, the memory (reminder) comes, which is based on imagination power.And after it, illusion and intuition power will come, which understands incomprehensible concepts in detailed sensible things, like sheep's understanding of wolf.Illusion power understands through inspirations, opacity, and symmetries and distinguishes benefits and losses and acts like a judge because it is the origin of principles and beliefs that reason trust in their correctness and certainty, like belief in the fact that every confused creature in in a place; therefore, this power is the origin of intellectual principles.Afterwards, memory comes, which stores whatever illusion understands (al-Kordi, 2003, pp. 57-112).
In Farabi's perspective, active intellect which is the same extra category of intellectual form, as Gabriel is the religious form for Him and this is the same category that orders and influences on the world of human understanding through these categories.In other words, active intellect is the same fundamental form that collects all categories, and this intellect is one of them.However, the speaking power, which is like active intellect and on the same level, is an imagination that alters between sensible issues and the principles, and it seems that active intellect is divine imagination and speaking power is human imagination.
The theory of blessing goes into the details of this fact and lightens an interpretation on the level of understanding and awareness.So he ponders on his essence, specifies and defines and delimits it and surrounds it completely, i.e. he imagines and depicts it as if he separates necessity from existence, and this is a kind of internal feeling, awareness of the soul, which is established in the whole essence, then it becomes full which cannot be controlled on the level of human imagination because He is the greatest entity and has the complete form and its necessity is consideration and heed.What does the intellect do?He imagines His essence within His essence, i.e.He imagines and understands His essences as a form or imagination, or through this another depiction for His abstracted reality will be issued, so a form with more visibility will be created.Therefore, trend and orientation of this thinking is blessing and He thinks of His essence, and this thinking is returnable and reversible; therefore, secondary reason and mental tendency and motion will be issued.The first thinking does more like the internal understand "I think" and He is the necessary condition for understanding the external reality and truth.These three are the essence of the theory of blessing on the level of single or reasonable creature or the simple cognitive unity whether it is imagination or affirmation, and it depicts awareness nature and understanding and structure of the form or thinking in an apparent expression.Therefore, human logic for awareness is necessarily apparent and imaginary.Thus, if thinking of the soul were not the beginning and if its role were not like the role of light in the body, it would be necessary to interpret it as revelation and blessing.
The issue for Farabi is not existence and essence because they are the most obvious and at the same time the most hidden issues.However, the problem is with the structure of awareness and human perception and the mediation of imagination in every perception and thinking because sometimes it is believed that the appearance of objects is obtained in the reason, the time for straightness of sensible entities without mediation; however, the issue is not like this and there are actually mediators among them and the reason is the form of all these mediators.Therefore, sense directly deals with sensible things and the forms of the sensible things and results in mutual sense.Thus, its common sense brings the forms to imagination and imagination to distinction power, so it can performs with purity within it and brings it to reason and guess in a purified way (Farabi, 1987, p. 104).Therefore, the basis for knowledge is sense unless the reasonable belongs to an abstract object (Jahami, 1998, p. 814).And imagination is carried out with reason, and human feels issues that are outside the soul and their forms are obtained in the senses combined with relationships.And the imaginer performs in a very abstracted way and imagines and feels his soul within it in a way that the one who is outside is the same sensible issue which is not in fact in agreement with the thing that exists in his soul such that he understands the difference between two forms, which is the peak of abstraction.Therefore, the reason is the most delicate thing and what it imagines is the most delicate imagination (Farabi, 1987, p. 103).Therefore, although the initial sense is reason that is not anything other than experiences and since all understandings and thinking are combined with imagination unless the sensible issue is clear rather the high samples are for knowledge (Farabi, 1987, p. 105).
According to Farabi, the right thing that imagination creates is selecting and extracting in the mind, which is the thing that gives experiences meaning.Therefore, when the imagination power succeeds in understanding these issues or reasons, miracle will occur and this essence will be expressed among other powers of the soul, and "if fantasy utilizes it, it is called imagination" (Farabi, 1979, p. 79), and in its holy complete level, it is the inspiring source for forms, samples, ideals, and imaginations in a way that an inspiring source in theory, reason, and selection in theoretical and practical philosophy will be created.In a general form, it is concluded from philosophy and the Prophet that philosophy inspires abstraction and likewise the Prophet inspires analogy, imagination, and symbol.The philosopher inspires high imagination or the performance of the reason within the sensible and ideal samples and forms with their abstraction while the Prophet inspires future imaginations, approaches, and symbols and their forms, whereby he imagines unity of religion and reason and that of art and science, and prepares theories for the community and provides the opportunity for human's perfection and real prosperity, which in turn creates the real balance inside the human who is equipped with the two wings of reasonable and unreasonable or imaginary ones.This power is talents and readiness that intensifies among some people until there is no need for that big thing.Its perfection is depicted by philosopher, prophet, or scholar poet, someone who deserves being called human because it is not impossible that human's imagination power reaches its ultimate perfection and is affected by active intellect and reasonable and transcendent imitations are accepted, through which human has a responsibility toward divine things and this is the complete level of imagination power (Farabi, 1986, p. 115).
Conclusion
Discussing the role and importance of imagination in the structure of epistemology creates new horizons in a way that our ideas about Islamic philosophical and in short the gnostic scholars in Islamic philosophy are influenced by the second teacher.In this regard, knowledge and reality have appearance, inside, understanding, and intuition, and imagination plays an active and decisive role in its structure, and any awareness and sense necessarily has a certain amount of imagination.
There is a special type of dualism in Farabi's approach and imagination which is related to his interpretation of religion and religious reality.He believes that religious truth is a metaphorical, artistic, literary, and imaginative truth.For him, human possess another unique value which is the value he generally possesses towards art and aesthetic work.Another important point according to Farabi is the difference between philosophy and religion.According to him, philosophy inclines towards sense and experience while religion deals with intuition, predictions, and speculation.In intuition, philosophy supports necessity whereas philosophy supports essence, and this is what causes to think that these two are of different types.Therefore, a type of symbolism and affection and brevity ad interest accompanying with the form dominates him.And finally, it takes the extremist form of formalism, in a way that sometimes philosophy becomes an extremist form of idealism.This derivation and division cannot be solved except for through Farabi's philosophy which is based on unity of existentialism (from the beginning) and ultimately deals with maintenance of the balance and places the world of awareness as the world of darkness and is based on essence and its principles are among the appearance, the inside part, the reason, and the form.
Shortly, the present study was aimed at highlighting the fact that Farabi has caused innovation in speech and has exceeded from Aristotle and Plato's words, and his ideas are a combination of religion and philosophy, art and science, and reason and existence.He established a critical epistemology that has a remarkable effect on the whole history of philosophy, and this effect is apparent in the tradition of Islamic and Western philosophy. | 7,655.2 | 2015-08-18T00:00:00.000 | [
"Philosophy"
] |
Electronically decoupled stacking fault tetrahedra embedded in Au(111) films
Stacking faults are known as defective structures in crystalline materials that typically lower the structural quality of the material. Here, we show that a particular type of defect, that is, stacking fault tetrahedra (SFTs), exhibits pronounced quantized electronic behaviour, revealing a potential synthetic route to decoupled nanoparticles in metal films. We report on the electronic properties of SFTs that exist in Au(111) films, as evidenced by scanning tunnelling microscopy and confirmed by transmission electron microscopy. We find that the SFTs reveal a remarkable decoupling from their metal surroundings, leading to pronounced energy level quantization effects within the SFTs. The electronic behaviour of the SFTs can be described well by the particle-in-a-box model. Our findings demonstrate that controlled preparation of SFTs may offer an alternative way to achieve well-decoupled nanoparticles of high crystalline quality in metal thin films without the need of thin insulating layers.
S tacking faults (SFs) are defect-type structures that occur in crystalline materials, for example, due to a local mismatch of the atomic stacking within the crystallographic planes or due to a deviation in the stacking sequence of the planes. These defects can occur more frequently in crystalline films that are grown on substrates with different lattice. SFs are typically considered to be undesired defects that lower the film structural properties and therefore much effort is done to avoid their formation 1,2 . However, they also show intriguing electronic properties 3,4 that may be exploited if they can be created in a controlled manner. The formation of SFs in crystals is promoted by quenching from high temperature, by high-energy particle irradiation and by doping with, for example, Mg, Cd or Zn 1,2,5-7 . Alternatively, their formation can also be promoted by growth of thin films on selected substrates with an appropriate lattice mismatch. 3 One very particular type of defect is the so-called stacking fault tetrahedron (SFT), which consists of four different triangularshaped SF planes that together demarcate a three-dimensional quasi-perfect nanocrystal. Previously, we reported on lateral quantization effects in SFTs in pristine Ag(111) surfaces grown on mica 4 . These SFTs appear spontaneously during the Ag film growth and they are known to exist in various metals [5][6][7] , including Ag(111) films 3,8 . Thus far, studies have focused on the growth and annihilation of SFs, yet the electronic properties of SFTs have remained largely unexplored.
Here, we report on the electronic properties of SFTs that are retrieved in Au(111) films grown on mica and that we investigate by scanning tunnelling microscopy (STM), scanning tunnelling spectroscopy (STS) and high-resolution transmission electron microscopy (HRTEM). We find that the embedded Au SFTs reveal a remarkable decoupling from their metal surrounding, which can be attributed to their stacking-fault-type origin. This implies that the SFTs may be considered as metallic quantum dots that are embedded in a metallic film. The Au SFTs accommodate an electronic state at their exposed surface that differs strongly from that of the surrounding Au(111) surface state. Au SFTs therefore reveal a very different electronic behaviour than previously investigated Ag SFTs, which showed a lateral quantization effect of the Ag(111) surface state without a clear decoupling from the surrounding Ag(111) substrate 4 . This remarkable difference highlights the rich and diverse electronic properties of SFTs, which appear to strongly depend on the material.
Results
Identification of SFTs. The Au(111) surface is well known for its remarkable surface reconstruction that is commonly referred to as a herringbone reconstruction 9,10 . It consists of a periodic modulation of the surface topography, in which surface atoms rearrange in either face centred cubic (fcc) or hexagonal close packed (hcp) stacking. Hcp and fcc regions are separated by discommensuration lines in which the atoms are slightly squeezed out of the otherwise atomically flat (111) surface. These ridges are running along three directions following the (111) surface and typically switch their orientation in a periodic, herringbone-type manner. At the elbows of the reconstruction ridges, a single atomic point dislocation exists 11,12 .
Locally, the herringbone ridges can show a more disordered appearance. In these regions, it is occasionally observed that three pairs of herringbone ridges seem to merge together as illustrated in Fig. 1a. Within our experiments, we find that at 1 to 10% of such crossroads, a larger defect-like feature can exist . These defect-like features have a very regular shape and they either appear as depressions or as protrusions, as is the case for the two highlighted defects in Fig. 1a. At the used tunnelling voltage, the defect enclosed by the dotted circle has a depth of 25±5 pm, while the defect enclosed by the dashed circle has a height of 35±5 pm. Owing to their exclusive appearance at these herringbone crossroads and their crystalline shapes, these defects are considered to have a similar origin as the SFTs that exist in Ag(111) films 8 . Apart from the regular-shaped defects at the crossroads, also other defects can be observed in Fig. 1a. These defects are commonly observed in STM images of our Au(111) films and are interpreted as (sub)surface atomic-size defects (an impurity atom or a Au vacancy) in the top atomic layers of the Au(111) film. These defects act as effective scattering centres for both surface and bulk electrons, which we have previously reported in ref. 13 15 . For our experiments, the observed lateral sizes of the Au SFTs are in the 1 to 5 nm range, that is, about two to three times smaller than the Ag(111) SFTs reported in ref. 4. The Au SFTs were observed on nine different samples (each prepared as described in the 'Methods' Section), irrespective of the amount of cleaning cycles (ranging from one up to eight cycles of ion bombardment and annealing). In total, more than 100 SFTs were retrieved, all of them at herringbone crossroads similar to the SFT presented in Fig. 1a. The amount (density) of SFTs can vary considerably from one sample to another (ranging from only one SFT within 1 mm 2 to several SFTs within 0.01 mm 2 ). We did not find a clear relation between the annealing time (ranging from 1 h to more than 12 h) and the annealing temperature (ranging from 330 to 430°C) on one hand, and the density of SFTs on the other hand. The observed variation of the SFT densities among our samples may be related to a variation of the density of impurities and vacancies in our gold films, which can act as nucleation centres for SFT formation 7 . The SFT size and density may be tuned by varying the growth parameters (for example, film thickness, involved temperatures, rate of deposition, and so on) and by selecting a different substrate for the Au(111) film growth 3 .
The electronic behaviour of SFTs. Next, we performed a detailed STM investigation of the SFTs. Remarkably, the height of the SFTs (with respect to the surrounding Au(111) surface) in the STM topographies depends strongly on the tunnelling voltage. This is illustrated in Fig. 2a,b. At voltages below 1 V, the SFT appears as a depression with a voltage-independent depth ( Fig. 2a). At voltages above 1V, the depth is strongly voltagedependent and the SFT can even appear higher than the surrounding Au(111) surface (Fig. 2b). The dependence of the height of the SFT on the tunnelling voltage is demonstrated in more detail in Fig. 2c, which shows the height of the SFT in Fig. 2a,b as inferred from STM topographies (blue dots) and from distance-voltage z(V) curves (red dotted line) that are recorded on the SFT (green solid line) and the surrounding Au(111) surface (black dashed line). Figure 2c demonstrates that the height of the SFT oscillates with increasing voltage. This peculiar behaviour is observed for all thus investigated SFTs and can be accounted for by the specific electronic structure of the SFTs. Figure 2d presents dI/dV spectra that are recorded together with the (black dashed and green solid) z(V) spectra in Fig. 2c. These spectra reflect the local density of states (LDOS) of the SFT and the Au(111) surface. The Au(111) spectrum is more or less featureless within the surface band gap that ranges up to about 3.7 V (ref. 16). Around this voltage, the bottom of the bulk conduction band appears as a pronounced step in the dI/dV signal (indicated by vertical black dotted line). At higher voltages, four image-potential states are revealed (labelled 1 to 4). These electronic resonances in Fig. 2d are observed as a step in the corresponding z(V) spectrum in Fig. 2c. Image-potential states exist below the vacuum level, yet they are shifted to higher voltages due to the applied electric field in the STM experiments 17 . Electrons within an image-potential state act as a two-dimensional free-electron-like gas that can move freely parallel to the surface. In contrast to the Au(111) spectrum, the SFT spectrum reveals the presence of pronounced electronic resonances in the same voltage region, that is, around 1.6, 2.3 and 3.2 V (labelled A to C), as well as three image-potential states (labelled 1 0 to 3 0 ). Remarkably, the step-like onset of the bulk conduction band is completely absent in the Height from topography images SFT spectrum. This absence indicates that scattering from SFT electrons to Au bulk states is (nearly) absent and hence points to a strong decoupling of the Au SFT from the surrounding Au(111) surface. The image-potential states of the investigated Au SFTs always appear at higher voltages compared with the corresponding voltages of the surrounding Au(111) surface, as illustrated in Fig. 2d (also see Supplementary Fig. 1). We note here that it can be excluded that the states labelled A to C are imagepotential states, as will be demonstrated below. The voltage difference between the image-potential states of the Au(111) and SFT (1 0 -1, 2 0 -2, 3 0 -3, y) is typically several hundreds of meV (see Supplementary Fig. 1). We can then conclude that the work function of the SFTs is higher than that of the bare Au(111) surface 18 . Following theoretical work reported in ref. 19, this implies that the two-dimensional image-potential states should not be confined within the contours of the Au SFTs, in contrast to Ag SFTs in Ag(111) (ref. 4). The increased work function may be attributed to a decreased distance between the successive atomic layers in the Au SFT compared with the surrounding Au(111) atomic layers 20 .
The above-described observations hold for all SFTs that we investigated with STS: (1) absence of Au(111) bulk conduction band, (2) pronounced resonances between 1 and 4 V and (3) image-potential states occur at higher voltages compared with the surrounding Au(111). These common properties of the SFTs, in addition to their similar appearance, imply that the SFTs have a similar structure and hence a similar origin.
Quantization effects in SFTs. Next, we focus on the electronic resonances of the SFTs that are resolved at lower voltages, such as those labelled A to C in Fig. 2d. Figure 3a presents an STM topography of another, larger Au SFT. Corresponding dI/dV spectra are presented in Fig. 3b and reveal maxima around 1.6, 2.1 and 2.6 V. The spectra are recorded with the same tunnelling voltage setpoint yet with different tunnelling current setpoints, implying different electric fields between the STM tip and the sample. It can be seen in Fig. 3b that the resonances do not exhibit a detectable Stark shift for the used settings. This excludes interpretation of the resonances in terms of imagepotential states, which are strongly dependent on the electric field 17 . Figure 3c displays a selection of LDOS maps recorded on the Au SFT in Fig. 3a. It can be seen that pronounced wave patterns start to develop within the contours of the Au SFT for voltages exceeding about 1.5 V. The patterns within the SFT have a very high intensity compared with that of the surrounding Au(111) surface up to the bottom of the bulk conduction band around 3.7 V (see Fig. 2d). Above this value, the signal on the Au(111) surface increases drastically and wave patterns of the SFT become more and more difficult to discern. Standing waves cannot be observed on the surrounding Au(111) surface above 3.7 V due to the strong coupling of the Au(111) surface state to the bulk states. On the Au SFT, wave patterns can be resolved up to about 4.5 V (more dI/dV maps of the Au SFT in Fig. 3a are presented in Supplementary Fig. 2). This further confirms the strong decoupling of the Au SFTs from the surrounding Au(111).
It is clear that the resonances and the wave patterns in Fig. 3c do not exhibit a repeating periodic behaviour with applied tunnelling voltage (also see Supplementary Fig. 2). This excludes interpretation of the Au SFTs in terms of subsurface Ar bubbles that may remain after cleaning of the sample. Scattering of bulk electrons between the subsurface Ar bubble and the metal film surface leads to quantum-well-type resonances and electron standing waves that show a periodic behaviour with applied tunnelling voltage 21,22 . Given the applied high annealing temperature, the presence of remaining Ar bubbles in our Au(111) films is unlikely. Moreover, in the case of subsurface Ar bubbles, one expects to still probe the Au(111) bulk conduction band in dI/dV spectra recorded above the bubble. The Au SFTs have an appearance that is similar to that of Au vacancy islands that can be controllably created by mild ion bombardment and annealing 23 . However, while vacancy islands have a depth of one atomic layer that is independent of the applied tunnelling voltage, Au SFTs have a sub-monolayer depth/height that depends strongly on the applied tunnelling voltage as discussed above. Moreover, the electronic behaviour of Au SFTs differs strongly from that of Au vacancy islands 23,24 . Au vacancy islands confine the surface state of the Au(111) surface within their step boundaries. The surface state electrons within the vacancy islands experience a considerable coupling to the bare Au(111) substrate and their confinement persists only up to the onset of the bulk conduction band around 3.7 eV. We therefore interpret the observed wave patterns and energy resonances of the Au SFTs as a new electronic state that exists within the Au SFTs and that can be probed at the SFT facet that is exposed at the surface. This electronic state of the SFT exhibits a behaviour reflecting that of a surface state. The known surface state of the Au(111) surface is characterized by a parabolic-like dispersion with an onset energy E 0 ¼ À 460 meV and an effective electron mass m* ¼ 0.23 m e (refs 13,24). To learn more about the electronic state of the Au SFTs, we performed simulations using the particle-in-a-box software (available via ref. 25) developed by K.-F. Braun 26 . For the Au SFT in Fig. 3, we achieve good agreement between the experimental LDOS maps and the 2D particle-in-a-box model when using an onset energy ARTICLE are presented in Supplementary Fig. 2. These used values for E 0 and m* differ strongly from those of the Au(111) surface state, which may be accounted for by a different stacking of the Au atoms within the SFT when compared with the Au(111) film, as indicated above.
Comparison with the particle-in-a-box model. To further verify our interpretation in terms of quantum confinement within the Au SFTs, we now focus on a Au SFT that is significantly smaller. The lateral size (exposed facet) of the SFT in Fig. 4a is only about 1.5-2.0 nm. Corresponding dz/dV maps that reflect the LDOS are presented in Fig. 4b-f. The dz/dV spectrum in Fig. 4g reveals the existence of three electronic resonances below the voltage at which the first image-potential state occurs, that is, at 2,050, 3,100 and 4,100 mV. Here as well, the imagepotential states above the SFT occur at higher voltages compared with the surrounding Au(111) surface. The dz/dV maps at 2,050, 3,100 and 4,100 mV are similar to the maps for the larger SFT in Fig. 3 at 1,600, 1,900 and 2,100 mV, respectively.
The three electronic resonances of the small SFT exist at higher energies compared with the larger SFT in Fig. 3, in agreement with our interpretation in terms of the particle-in-a-box model. Considering the same onset energy E 0,SFT and effective electron mass m à , the electronic resonances E n,SFT (eigenstates) and the corresponding wave patterns of SFTs with different size O can be linked to each other using the particle-in-a-box equation E n;SFT ¼ E 0;SFT þ l n;SFT Âðm ?  OÞ À 1 ; n ¼ 1; 2; 3; . . . ð1Þ In equation (1), the ;eigenvalues l n,SFT depend solely on the shape of the confining box. Assuming the same (truncated triangular) shape for the SFT in Fig. 4 as that used to model the data in Fig. 3, we find good agreement between the simulated images and the experimental dz/dV maps if O is taken 25% the size of that used for Fig. 3. This is in very good agreement with the size determined based on the STM topography in Fig. 4a.
The dI/dV maps of the Au SFT in Fig. 2a,b are presented in Supplementary Fig. 3. Maps of yet another Au SFT are presented in Supplementary Fig. 4. This SFT has a very similar lateral size and shape as that in Supplementary Fig. 3 and shows quasi-identical wave patterns and voltage-dependent behaviour. The SFT in Fig. 3 and Supplementary Fig. 2 is slightly larger than the two SFTs in Supplementary Fig. 3 and Supplementary Fig. 4 and similar wave patterns are formed at somewhat lower voltages (for example, the wave pattern at 3,200 mV in Supplementary Fig. 3 and Supplementary Fig. 4 occurs at 2,900-3,000 mV for the NC in Supplementary Fig. 2), again in agreement with the particle-in-a-box model. Figure 5 presents an overview of the energy values of the electronic resonances of all investigated Au SFTs as a function of their surface area. It can be seen that there exists a strong correlation between the energies and the SFT surface area. This is again in line with interpretation of the electronic resonances in terms of the particle-in-a-box model, that is, the electronic behaviour can be described by the same onset energy E 0 ¼ 1,490 meV and the same effective electron mass m* ¼ 0.33 m e for all SFTs. In turn, this additionally indicates that all investigated SFTs have the same atomic structure, that is, they are all (truncated) Au SFTs that occur spontaneously in the Au(111) film and of which one facet is exposed at the Au(111) surface.
Deviations of the experimental data in Fig. 5 from the theoretical model may be attributed at least partially to deviations from the assumed idealized hexagonal (truncated triangular) shape. In particular, deviations in Fig. 5 are most pronounced for the smaller SFTs such as the one in Fig. 4a. For these smaller SFTs, the precise shape can be observed less clearly in STM images and it may be more close to that of a triangle rather than a hexagon. In addition, electron scattering at the subsurface SF planes of the SFT may affect the ideal particle-in-abox type confinement of electrons.
Finally, as indicated above, the exposed surfaces of the Au SFTs all have a regular crystalline shape, that is, the shape of a hexagon or truncated triangle. Exceptionally, an equilateral, triangularly shaped defect-like feature is found (only one observation, see Supplementary Fig. 5). The arrangement of the herringbone ridges at the triangular shaped feature differs from that of the other SFTs (Supplementary Fig. 5a). Moreover, the triangular feature exhibits an electronic behaviour that differs from the other SFTs. It shows the Au(111) bulk conduction band similar to the surrounding Au(111) surface ( Supplementary Fig. 5k). In addition, wave patterns can already be observed in LDOS maps at voltages close to the Fermi level ( Supplementary Fig. 5f,g), while the SFTs discussed above reveal wave patterns only above 1,490 mV. The resolved wave patterns ( Supplementary Fig. 5e-i) and resonance in the STS spectrum ( Supplementary Fig. 5j) can be interpreted as confinement of the bare Au(111) surface state ( Supplementary Fig. 5j) than the other Au SFTs. The triangular feature may be interpreted as a so-called Frank loop, consisting of a single stacking fault. The Frank loop is very similar to the SFT in terms of the spatial coordinates of atoms.
In conclusion, we performed a detailed STM investigation of Au SFTs that exist in Au(111) films, as confirmed by TEM experiments. The SFTs exhibit a set of discrete electronic resonances and reveal pronounced voltage-dependent wave patterns in maps of the density of states. The wave patterns exist up to energies well above the bottom of the bulk conduction band of the Au(111) film, indicative of a strong decoupling of the Au SFT from its surroundings. This behaviour is found to correlate with the size of the Au SFTs. From analysis using a two-dimensional particle-in-a-box model, we find that the electronic behaviour can be described well by an electronic state with parabolic dispersion having an onset energy of about 1,490 meV above the Fermi level and an effective electron mass of about 0.33 m e .
Our findings demonstrate that controlled introduction of SFTs may offer an alternative way to obtain well-decoupled quantum dots of high crystalline quality in metal thin films without the need of thin insulating layers, which often make sample preparation more cumbersome 27 . In addition, SFTs can be expected to have an enhanced stability at room temperature when compared with deposited nanoclusters of similar size 27,28 . Obviously, their controlled preparation will be a crucial issue for further developments in this direction. A potential route to overcome this issue could be by creating regular patterns of defects in the substrate, at which SFTs may preferentially start to form during the metal film growth on the support. However, it is already evident from our present findings that investigating the electronic properties of SFTs provides a new playground for in-depth studies of quantum mechanical finite size effects in surfaces.
Methods
Sample preparation. Epitaxially grown, 140 nm thick Au(111) films on freshly cleaved mica were prepared by molecular beam epitaxy at elevated temperatures as described in ref. 29. Sample transfer from the molecular beam epitaxy set-up to the low-temperature ultra-high vacuum STM set-up was performed under ambient conditions. The Au(111) surfaces were cleaned in the preparation chamber of the STM set-up by repeated cycles of Ar ion bombardment (at about 4 keV and 10 À 6 mbar Ar partial pressure) and annealing (at about 720 K). The resulting film surfaces consist of atomically flat islands with dimensions up to 500 Â 500 nm 2 (ref. 13).
STM experiments. All the experiments were conducted in a ultra-high vacuum system (for sample preparation, base pressure in the 10 À 9 mbar range) that is connected to a low-temperature STM (Omicron Nanotechnology) operated at 4.5 K (for sample measurement, base pressure in the 10 À 11 mbar range). (dI/dV)(V) spectra and dI/dV maps (commonly referred to as LDOS maps) were acquired by lock-in detection with closed feedback loop (amplitude is typically about 20 to 50 mV) at 800 Hz. (dz/dV)(V) spectra and dz/dV maps that also reflect the LDOS are obtained numerically from recorded z(V) spectra. STM data in this work were obtained with mechanically cut PtIr (10% Ir) STM tips, and with polycrystalline W tips that were electrochemically etched and cleaned in situ by thermal treatment. All bias voltages mentioned are with respect to the sample, and the STM tip was virtually grounded. The STM images were analysed using the Nanotec WSxM software 30 .
TEM experiments. The cross-sectional TEM thin foils were fabricated in a dual beam Helios NanoLab 660 (FEI) setup using the lift-out procedure. To protect the surface of the Au film from damage caused by the incident Ga þ ions of the focused ion beam (FIB) 31 , a protective Pt layer was first deposited using electronbeam assisted deposition (5 kV, 0.8 nA) followed by an ion-beam assisted deposited Pt layer (30 kV, 0.23 nA). To minimize any damage on the sample during thinning, final cleaning on both sides of the thin lamella was performed using a low energy ion beam of 1 kV and 95 pA. The HRTEM characterizations of the Au films were carried out using a FEI Tecnai G2 (FEG, 200 kV). To achieve clear visualization of single dislocations and SFs, local strain mapping was performed using the Geometric Phase Analysis, which is an image processing technique that is sensitive to small displacements of the lattice fringes in HRTEM images 32 . Energydispersive X-ray analysis was performed in TEM and did not reveal any trace of Ga þ ions in the TEM sample. However, with a detection limit of around 1 at.%, this does not exclude the existence of some Ga in the film and thus the production of extra vacancies, so in order to confirm that SFTs are intrinsic to the present Au films and not FIB induced artifacts, a bulk pure Au (99.99%) reference sample was annealed at 973 K for 24 h. Next, a cross-sectional FIB sample was produced using the same conditions as for the Au film. HRTEM investigation on this sample only revealed dislocation loops and individual SFs, while no SFTs were found (also see Supplementary Fig. 6).
Calculations. Simulations were performed using particle-in-a-box software (available via ref. 25) developed by K.-F. Braun 26 (the Schrödinger equation was solved by treating scattering centres at the box boundaries as zero-range potentials).
Data availability. All relevant data related to this manuscript are available from the authors. | 6,010.2 | 2016-12-23T00:00:00.000 | [
"Materials Science",
"Physics"
] |
Crystal Growth and Stoichiometry of Strongly Correlated Intermetallic Cerium Compounds
Strongly correlated electron systems are among the most active research topics in modern condensed matter physics. In strongly correlated materials the electron interaction energies dominate the electron kinetic energy which leads to unconventional properties. Heavy fermion compounds form one of the classes of such materials. In heavy fermion compounds the interaction of itinerant electrons with local magnetic moments generates quasiparticles with masses up to several 1000 electron masses. This may be accompanied by exciting properties, such as unconventional superconductivity in a magnetic environment, non-Fermi liquid behavior and quantum criticality. Strong electronic correlations are responsible for physical phenomena on a low energy scale. Consequently, these phenomena have to be studied at low temperatures. This, in turn, requires ultimate quality of single crystals to avoid that the low temperature intrinsic properties are covered by extrinsic effects due to off-stoichiometry, impurities or other crystal imperfections. The overwhelming majority of heavy fermion systems are cerium and ytterbium intermetallic compounds. In the present paper we discuss the crystal growth of three cerium compounds, Ce3Pd20Si6, CeRu4Sn6 and CeAuGe. Ce3Pd20Si6 undergoes an antiferromagnetic phase transition at low temperatures and shows a magnetic field induced quantum critical point [Takeda et al (1995), Strydom et al (2006)]. CeRu4Sn6 [Das & Sampathkumaran (1992)] appears to be a Kondo insulator [Aepli & Fisk (1992)] with a highly anisotropic energy gap. CeAuGe is one of a the few cerium compounds showing a ferromagnetic phase transition at low temperatures [Pöttgen et al (1998), Mhlungu & Strydom (2008)]. This is of special interest in the context of quantum criticality, since the occurrence of quantum criticality on the verge of a ferromagnetic ground state is a subject of current debate. Much attention in this paper is paid to the problem of stoichiometry. Single crystals of intermetallic compounds are grown at high temperatures, which facilitates the formation of thermal defects realized often as deviation from the stoichiometric composition. Thermal instabilities of some intermetallic phases require the use of flux techniques, i.e., growth from off-stoichiometric melts, which is another source of non-stoichiometry. Sizeable non-stoichiometries can be detected by measuring the composition by chemical and physical analytical techniques, e.g. energy dispersive X-ray spectroscopy analysis (EDX). Tiny deviations from the stoichiometry, on the other hand, can be found only from an analysis of the physical properties of single crystals. Thus physical property measurements are not only the 11
Introduction
Strongly correlated electron systems are among the most active research topics in modern condensed matter physics.In strongly correlated materials the electron interaction energies dominate the electron kinetic energy which leads to unconventional properties.Heavy fermion compounds form one of the classes of such materials.In heavy fermion compounds the interaction of itinerant electrons with local magnetic moments generates quasiparticles with masses up to several 1000 electron masses.This may be accompanied by exciting properties, such as unconventional superconductivity in a magnetic environment, non-Fermi liquid behavior and quantum criticality.Strong electronic correlations are responsible for physical phenomena on a low energy scale.Consequently, these phenomena have to be studied at low temperatures.This, in turn, requires ultimate quality of single crystals to avoid that the low temperature intrinsic properties are covered by extrinsic effects due to off-stoichiometry, impurities or other crystal imperfections.The overwhelming majority of heavy fermion systems are cerium and ytterbium intermetallic compounds.In the present paper we discuss the crystal growth of three cerium compounds, Ce 3 Pd 20 Si 6 ,C e R u 4 Sn 6 and CeAuGe.Ce 3 Pd 20 Si 6 undergoes an antiferromagnetic phase transition at low temperatures and shows a magnetic field induced quantum critical point [Takeda et al (1995), Strydom et al (2006)].CeRu 4 Sn 6 [Das & Sampathkumaran (1992)] appears to be a Kondo insulator [Aepli & Fisk (1992)] with a highly anisotropic energy gap.CeAuGe is one of a the few cerium compounds showing a ferromagnetic phase transition at low temperatures [Pöttgen et al (1998), Mhlungu & Strydom (2008)].This is of special interest in the context of quantum criticality, since the occurrence of quantum criticality on the verge of a ferromagnetic ground state is a subject of current debate.Much attention in this paper is paid to the problem of stoichiometry.Single crystals of intermetallic compounds are grown at high temperatures, which facilitates the formation of thermal defects realized often as deviation from the stoichiometric composition.Thermal instabilities of some intermetallic phases require the use of flux techniques, i.e., growth from off-stoichiometric melts, which is another source of non-stoichiometry.Sizeable non-stoichiometries can be detected by measuring the composition by chemical and physical analytical techniques, e.g.energy dispersive X-ray spectroscopy analysis (EDX).Tiny deviations from the stoichiometry, on the other hand, can be found only from an analysis of the physical properties of single crystals.Thus physical property measurements are not only the final purpose of a crystal growth but also a valuable diagnostic tool for further improvement of their quality.Therefore, in the paper the consideration of crystal growth is accompanied by the discussion of their physical properties.Physical property measurements on Ce 3 Pd 20 Si 6 single crystals grown by the floating zone technique have been reported earlier [Goto et al (2009), Prokofiev et al (2009), Mitamura et al (2010)].In Ref. Prokofiev et al (2009) a systematic study of the relationship between the growth technique, stoichiometry and physical properties of single crystals has been done.The crystal growth and stoichiometry of CeRu 4 Sn 6 and CeAuGe are reported for the first time.The physical properties of the CeRu 4 Sn 6 single crystals were published partially in an author's earlier paper [Paschen et al (2010)].
Ce 3 Pd 20 Si 6
The recent observation of magnetic-field induced quantum criticality [Paschen et al (2007); Strydom et al (2006)] in the cubic heavy fermion compound Ce 3 Pd 20 Si 6 [Takeda et al (1995)] has attracted considerable attention.Ce 3 Pd 20 Si 6 crystallizes in a cubic Cr 23 C 6 -type structure with the space group Fm3m [Gribanov et al (1994)].The cubic cell with a= 12.161 Å [Gribanov et al (1994)]; 12.280 Å [Takeda et al (1995)] contains 116 atoms.The Ce atoms occupy two distinct sites of cubic point symmetry.At the octahedral 4a (tetrahedral 8c) site Ce atoms are situated in cages formed by 12 Pd and 6 Si atoms (16 Pd atoms).These high coordination numbers allow to classify Ce 3 Pd 20 Si 6 as a cage compound.Strongly correlated cage compounds are of much interest as potential candidates for thermoelectric applications [Paschen (2006)].Similar to the isostructural germanide compound Ce 3 Pd 20 Ge 6 , two phase transitions -a presumably antiferromagnetic one at T L of 0.15 K [Takeda et al (1995)], 0.17 K [Goto et al (2009)], or 0.31 K [Strydom et al (2006)] and a possibly quadrupolar one at T U of 0.5 K [Strydom et al (2006)] -have been found in the silicide compound Ce 3 Pd 20 Si 6 .Similar to the effect of magnetic field applied along [100] or [110] in Ce 3 Pd 20 Ge 6 [Kitagawa (1998)], a magnetic field shifts the two transitions of polycrystalline Ce 3 Pd 20 Si 6 in opposite directions: At the critical field of about 1 T the transition at T L is suppressed to zero whereas the transition at T U shifts to 0.67 K [Strydom et al (2006)].The non-Fermi liquid behavior of the electrical resistivity observed at the critical field in the temperature range 0.1-0.6K is an indication for the existence of a field-induced quantum critical point [Paschen et al (2007)].
Neutron scattering experiments on polycrystalline Ce 3 Pd 20 Si 6 have to date failed to detect any kind of magnetic order [Paschen et al (2008)].Thus, to clarify the nature of the phases below T L and T U large single crystals of high quality are needed.In fact, in the first neutron scattering study on Ce 3 Pd 20 Si 6 single crystals [Paschen et al (2008)] the absence of signatures of magnetic order was attributed to a too low T L value of the investigated specimen.Since both phase transitions take place at rather low temperatures, disorder may influence them significantly.The discrepancy in the reported ordering temperatures (e.g.[Goto et al (2009); Strydom et al (2006); Takeda et al (1995)]) demonstrates this delicate dependence.Also the non-negligible difference in the reported unit cell constants [Gribanov et al (1994), Takeda et al (1995)] needs a clarification.This has motivated us to undertake a systematic investigation of the relation between crystal growth techniques/conditions, sample quality, and the resulting physical properties down to dilution refrigerator temperatures.We show here that both phase transitions are extremely sensitive to small stoichiometry variations that result from different growth procedures.
To elucidate the discrepancies in the low-temperature data reported on the quantum critical heavy fermion compound Ce 3 Pd 20 Si 6 and reveal the compound's intrinsic properties, single crystals of varying stoichiometry were grown by various techniques -from the stoichiometric and slightly off-stoichiometric melts as well as from high temperature solutions using fluxes of various compositions.The results of this work on Ce 3 Pd 20 Si 6 have been partially reported earlier [Prokofiev et al (2009)].Here a more detailed analysis including also information on new crystal growth experiments as well as the physical property measurements on new single crystals are reported.
Growth from the stoichiometric melt
To investigate the melting character of Ce 3 Pd 20 Si 6 a differential thermal analysis (DTA) run up to 1350 • C was carried out (Fig. 1).There is only a single peak both on the heating and on the cooling curve, indicating congruent melting as previously reported [Gribanov et al (1994)].
A closer inspection of the shape of the melting peak (Fig. 1, inset) might, however, suggest merely a quasi-congruent melting character.The onset of melting occurs at T M ≈ 1250 • C. Because of undercooling the crystallization begins about 100 • C lower than the melting.The inset shows a magnification of the melting peak.From Ref. Prokofiev et al (2009).
The floating zone technique with optical heating was used for crystal growth from the melt, with a pulling rate of 10 mm per hour and an upper rod rotation speed of 5 rpm.Due to the high density and the low surface tension of the Ce 3 Pd 20 Si 6 melt the floating zone was rather unstable, and the melt dropped down repeatedly (Fig. 2, right).However, one growth could be kept running long enough to grow an ingot of 25 mm length (Fig. 2, left).Over the growth run the originally shiny and clear surface of the melt became more and more opaque, and a crust on the surface could be seen after some time.This is a sign of incongruent melting supposedly due to a peritectic reaction.The crust is the higher temperature phase, therefore it forms on the optically heated surface which is the hottest place of the melt.This observation can be explained by close proximity of the peritectic point to the temperature of the complete melting of the system with the total composition Ce 3 Pd 20 Si 6 , as indicated already by the peculiar shape of the DTA curve (Fig. 1 inset).Therefore the formation of a small amount of foreign phase on the hot surface results in a slight shift of the melt composition, which returns the crystallization process in the melt deep into the primary crystallization field of the Ce 3 Pd 20 Si 6 phase.Due to the thin foreign phase film the surface of the recrystallized (lower) rod was yellowish.
The EDX measurement detected Ce (∼ 70 at.%),Pd (∼ 22 at.%), and Si (∼ 8 at.%) in this film.However no inclusions of foreign phases were found by scanning electron microscopy (SEM) inside the crystal.A single crystal grown in this way is specified here as sc1(seeT able1forthe nomenclature of all crystals).In order to trace segregation effects we differentiate additionally between the part of this crystal grown at the beginning of the growth process (bottom part of the ingot, sc1b) and that grown at the end (top part, sc1t).For the growth of sc1thelower purity starting materials (Ce 99.99 at.%,Pd 99.95 at.%) were used.Further growth runs with rotation of the upper rod led to a permanent breaking of the surface crust.The melt leaked out through the cracks of the crust.Hence, the crust may serve as a quasi-crucible if it remains intact during the whole growth time.Based on this observation a growth run without rotation of the upper rod was carried out with the same growth parameters.The melt zone was quite stable in this experiment.A crystal grown by this procedure will be specified as sc2.For the growth of this crystal higher purity starting materials (Ce 99.99 at.%,Pd 99.998 at.%) were used.Laue investigations show very good crystallographic perfection of both types of crystals grown from the melt.Before annealing an SEM/EDX investigation of both crystals was carried out.The polished cross-section of the sc1 rod is more homogeneous in composition than that of sc2.The surface of the latter had a well distinguishable 300 µm thick outer shell where the concentration of Ce was about 5% higher and that of Si somewhat lower than in the core region (Fig. 3, left).This shell may result from a partial dissolution and diffusion of the crust into the core of the rod.In the core region of sc2d i f f u s e1 0µmi n c l u s i o n so f supposedly the same shell phase, however with lower Ce concentration, were found (Fig. 3, right).These inclusions were not detected by X-ray powder diffraction and they disappeared after annealing, according to SEM analysis.After annealing of both crystals for 3 weeks at 900 • C a second EDX investigation was carried out.At first we did the analysis without any reference sample.This yielded a stoichiometric Ce content (10.3 at.%) but an over-stoichiometric Si content (22.0 instead of 20.7 at.%) and an under-stoichiometric Pd content (67.7 instead of 69.0 at.%) for both sc1andsc2, corresponding to a partial substitution of Pd by Si on their sites.These results motivated our efforts to grow crystals using the flux technique, where the stoichiometry of the crystals can be tuned through the variation of the flux composition.The results of further EDX investigations with a reference sample will be discussed in Sect.2.4.
Flux growth 2.2.1 Tin flux
We tried at first a crystal growth with standard fluxes.The related compound Ce 2 Pd 3 Si 5 can be grown with Sn as flux at below 1100 • C [Dung et al (2007)].To check for the compatibility of Sn flux with Ce 3 Pd 20 Si 6 ,amixtureofCe 3 Pd 20 Si 6 and Sn was heated up to 1100 • Cinaboron nitride crucible and then cooled slowly (1 • C/h) down to 700 • C. The solute-to-solvent ratio was 2:1.After crystallization the ingot was cut, polished, and investigated by SEM/EDX and X-ray powder diffraction (XRD).The crystallization yielded relatively large single crystals of the non-stoichiometric phase CePd 2−x Si 2+x with x ≈ 0.25, incorporated in a matrix of Sn 4 Pd.In addition, small inclusions 267 Crystal Growth and Stoichiometry of Strongly Correlated Intermetallic Cerium Compounds www.intechopen.com of other phases were found.The experiment thus shows the inapplicability of Sn as a flux because its affinity to Pd leads to a destruction of the Ce 3 Pd 20 Si 6 phase.However, other Sn-based flux compositions with a lower affinity to Pd can be found in the Sn-Pd binary phase diagram [Chandrasekharaiah (1990)].A series of Sn-Pd compounds -Sn 4 Pd, Sn 3 Pd, Sn 2 Pd -with low melting points (below 600 • C) exists.Similar experiments as the one with pure Sn flux were carried out using the above Sn-Pd compositions.In all cases the primarily crystallized phase was CePd 2−x Si 2+x according to SEM/EDX.The x-value diminished with increasing Pd content in the flux, reaching about 0.05 for Sn 2 Pd flux.The stoichiometric CePd 2 Si 2 phase has the same Ce/Si ratio as the Ce 3 Pd 20 Si 6 phase but a strongly reduced Pd content.However, further increasing of the Pd concentration in the flux for tuning of Pd content in crystals was impossible: with the higher melting compound SnPd, only partial melting of the crucible content was observed at 1100 • C. Thus, foreign flux growth was not successful.
Self-flux Pd 5 Si
The reason for using the flux method was, on one hand, to obtain single crystals with exact stoichiometric composition and, on the other hand, the expectation that the off-stoichiometric melt would have a higher surface tension and hence the floating zone would be more stable than without flux.To avoid a contamination by foreign atoms we first opted for self flux growth.Since the concentration of thermal defects (e.g., Si -Pd substitutions) in the crystal is expected to decrease with decreasing growth temperature, we searched in the Ce-Pd-Si phase diagram (Fig. 4a) for low-melting (at first binary) compositions with an over-stoichiometric Pd content.The phase Pd 5 Si [Seropegin (2001)] which, according to a later study [Gribanov et al (2006)], appears to consists or two scarcely distinguished phases Pd 14 Si 3 and Pd 84 Si 16 , fulfills all requirements: it melts at a relatively low temperature of 835 • C [ Chandrasekharaiah (1990)], has an over-stoichiometric (> 20:6) Pd/Si ratio, and there are no stable Ce-containing intermediate phases between Pd 5 Si and Ce 3 Pd 20 Si 6 in the Ce-Pd-Si ternary phase diagram (cross-section at 600 • C [Gribanov et al (2006); Seropegin (2001)]).As above, the floating zone technique with optical heating was used.The feed and the seed rods had the stoichiometric starting composition Ce 3 Pd 20 Si 6 , and the zone was a molten mixture of Ce 3 Pd 20 Si 6 and Pd 5 Si.Contrary to the melt growth, the floating zone was very stable, and its surface remained clear during the entire growth run.The latter means that the growth occurred within the Ce 3 Pd 20 Si 6 primary crystallization field.The stability of the melt zone allowed the rotation of the upper rod.The pulling rate was 0.6 mm/h.Two growth runs with different flux compositions (molar ratios Ce 3 Pd 20 Si 6 to Pd 5 Si of 2:1 and 1:2) were carried out (Fig. 4b).The crystals were annealed for 3 weeks at 900 • C. The corresponding samples are referred to as sc3andsc4, respectively.
Growth from slightly off-stoichiometric melt
As the analysis of the composition and the properties of the crystals sc1-sc4grownfromthe melt and from flux has shown (see Sections 2.4, 2.5) the stoichiometry was strongly sensitive to the starting composition of the melt.For a fine correction of non-stoichiometry a growth from a slightly off-stoichiometric melt was carried out, too (sc5).For reasons discussed later, the feed rod for sc5 contained 0.3 at.% excess of Ce and 0.2 at.% deficiency of Pd.To avoid the floating zone instability the growth was carried out without rod rotation, as in case of sc2.The pulling rate was 4 mm/h.Similar to sc1 we differentiate additionally between the bottom and top parts of the ingot, sc5b and sc5t.
Composition of the grown crystals
The usual EDX technique without standards cannot determine absolute atomic concentrations with sufficient accuracy.This problem can, in principle, be solved by using a standard of exactly known composition.However, as such a sample is not readily available, we used, instead, the polycrystalline sample (pc) which will be shown below to be of best quality, according to the physical property measurements.Irrespective of whether its composition may be identified with the exact stoichiometry 3:20:6 or not it served as a practical guideline to establish the crystal composition-property relationship.With additional improvements of our EDX setup (improved measurement statistics, counting time, beam current control) we can measure, with a high accuracy, deviations from the reference sample stoichiometry.The results are summarized in Table 2 where the measured lattice parameters are given, too.Because crystals grown from the Sn-containing fluxes are not the title phase and their compositions vary sizably they are simply represented by CePd 2−x Si 2+x in Table 2.Even though, in the absence of a real standard, there remains some uncertainty about the absolute values measured by EDX (which even for the reference sample pc differ from the calculated stoichiometry 3:20:6), trends in the composition of the investigated series of crystals can be discussed.While the Si content varies only weakly a stronger variation of the Ce and Pd content is observed, the Ce concentration nearly anticorrelating with the one of Pd (Fig. 5).2. From Ref. Prokofiev et al (2009).
Ce over-stoichiometry can be realized by a substitution of Pd or Si atoms on their crystallographic sites by excess Ce-atoms or, alternatively, by vacancies on the Pd and/or Si sites.Since the lattice parameter of sc1t and sc5 is smaller than that of the pc,t h e latter option is more probable, but a combination of both options cannot be excluded.If only Pd/Si vacancies were present, the Ce sublattice would remain occupied and fully ordered.Ce under-stoichiometry, on the other hand, can be associated either with Ce vacancies or, which is more probable taking into account the approximate anti-correlation between the Ce and Pd contents, with a partial Pd substitution on Ce sites.The broad dark grey line in Fig. 5 represents the Ce-Pd concentration relation for a crystallochemical model which, as an example, assumes a half-filling of the Ce vacancies V by excess Pd atoms: [Ce 1−x V 0.5x Pd 0.5x ] 3 Pd 20 Si 6 .The same model was used to describe the ∆Si vs ∆Ce relation (light grey line).The agreement with the data is excellent.
Influence of the growth technique on the physical properties
The resulting stoichiometry of the crystals as well as their physical properties show sizable dependence on the different growth techniques.Neutron diffraction experiments were carried out on an oriented sample cut from sc1b.T h e crystal was confirmed to have excellent crystallinity but, contrary to our expectations from the investigations on polycrystalline samples, showed no phase transition down to the lowest accessed temperature of about 0.15 K [Paschen et al (2008)].The present investigation shows that, for this very single crystal, this temperature was still too high.The specific heat C p was measured for sc1-sc3a n dsc5.Figure 6a shows the temperature dependence of C p /T of these crystals, together with published data of a polycrystalline sample [Strydom et al (2006)].C p /T(T) of the polycrystalline sample pc has a sharp peak with a maximum at T L = 0.31 K and a shoulder at T U ∼ 0.5 K [Strydom et al (2006)].These features get successively broadened and suppressed to lower temperatures in the crystals sc1t, sc5b, sc1b,a n dsc2, respectively.In sc3 no clear signature of T L can be identified, suggesting that it has shifted further to lower temperatures and was further broadened or, alternatively, has vanished altogether.Due to the suppression of this lower temperature feature the anomaly at T U , identified in all other samples as shoulder, now appears as a maximum.From the specific heat measurements the best single crystals thus appear to be sc1t and sc5.With their rather sharp peaks at 0.22 K and 0.20 K, respectively, and a shoulder at about 0.4 K they demonstrate all features associated with the intrinsic behavior of Ce 3 Pd 20 Si 6 .polycrystalline sample drops to very low values at low temperatures, the residual resistivities of the single crystals are considerably higher.They increase in the sequence pc → sc5t → sc1t → sc5b→ sc1b → sc2 → sc3.This is about the same sequence in which the temperature and the sharpness of the low-temperature phase transition of the C p /T curves decreases (Fig. 6a).Thus, it is natural to assume that in the sequence pc → (sc1t and sc5)→ sc1b→ sc2 → sc3 the lattice disorder increases.Since the starting material purity of sc1t and sc1b was lower than that of sc2, we conclude that the main reason of the disorder is a deviation of the sample composition from the exact stoichiometry 3:20:6 and not the concentration of foreign atoms.For single crystal sc4 the high temperature minimum occurs at lower temperatures than expected, leading to a slightly lower residual resistivity than for sc3.
In sc1 a remarkable increasing of the quality from the bottom to the top part of the crystal is observed.Only a small (top) part of the total crystal has an excellent quality and can be used for physical property investigations.sc5 is indeed ranked after sc1t according to the C p data but it is more homogeneous throughout the whole volume of the batch.A scenario which can explain the difference between the bottom (sc1b)andthetop(sc1t) parts of the crystal sc1isas follows.At first (bottom part of the ingot) the crystal phase captures less Ce and more Pd from the stoichiometric melt (Table 1).While the crystal grows the melt gets enriched by Ce and depleted by Pd.At the end of the crystallisation (top part) this change in the melt composition results in a shift of the crystal composition to a more stoichiometric one, in accordance with the law of mass action.It was this observation that motivated us to perform the off-stoichiometric growth with a little Ce excess and Pd deficiency that resulted in sc5.As Fig. 7 shows the low temperature (down to 2 K) relative resistivity of sc5 is comparable with the best sample sc1t and the spatial (top-bottom) difference in the resistivity is much smaller for sc5thanforsc1.The highest residual resistance ratio (RRR, defined here as ρ(200 K)/ρ(50 mK))a n d the sharpest and most pronounced phase transition features in C p /T are found for the polycrystalline sample (pc) which therefore appears to be the most stoichiometric one.This can be easily understood by the specifics of preparation.The accuracy of the stoichiometric total composition of a polycrystalline sample is limited only by the accuracy of the weighing process of starting materials and by their purity.A possible high temperature non-stoichiometry of the main phase of an as-cast polycrystal is compensated by the presence of minor impurity phases, the phase separation being heterogeneous on a microscopic scale.This heterogeneity can be lifted by annealing at lower temperatures due to the short diffusion path.During single crystal growth, however, a macroscopic phase separation can occur, making annealing very inefficient.Actually the resistivity curves of sc1b before and after annealing were practically identical.T h eC ec o n t e n t∆Ce varies by more than 3 at.%among all grown single crystals.We have revealed a systematic dependence of the residual resistance ratio, the lattice parameter, the (lower) phase transition temperature T L , and the maximum in the temperature dependent electrical resistivity T max with ∆Ce.This clarifies the sizable variation in the values of T L reported in the literature.We discuss the physical origin of the observed composition-property relationship in terms of a Kondo lattice picture.We predict that a modest pressure can suppress T L to zero and thus induce a quantum critical point.
While no clear correlation between the physical properties and the Si content can be demonstrated, a pronounced dependence on the Ce content (or anti-correlated Pd content) is observed.Figure 8a shows the dependencies of the lattice parameter a and the residual resistance ratio RRR as function of the deviation from stoichiometry ∆Ce.The largest deviations from the "intrinsic behavior" (which is most closely met by the polycrystal) are seen for sc4.It has the lowest Ce content of 7.4 at.% (∆Ce = 2.9 at.%).The polycrystalline sample (pc) which demonstrates the most pronounced phase transition features and sc1t with the second-sharpest features have the highest Ce contents, -and the lowest deviation from the exact Ce stoichiometry.The deviations ∆Ce of sc1t and sc5 lie on the other side of the stoichiometry line.One may argue that only the absolute value of ∆Ce is relevant with respect to the composition-disorder-property relationship since any non-stoichiometry is usually associated with lattice imperfection.To test this conjecture we plot, in Fig. 8a, the data point for sc1t also mirrored through the ∆Ce = 0 axis (open symbol).Indeed this point fits nicely into the overall RRR(∆Ce) behavior.
As it was pointed out in Sec.2.4, the Ce-understoichiometry may be realized by two ways: either by Ce-vacancies or by Ce-vacancies partially (to 50%) filled by Pd atoms.Both cases of Ce under-stoichiometry correspond, in a Kondo lattice description, to Kondo holes.In the Kondo coherent state at low temperatures Kondo holes act as strong scattering centers, decreasing the RRR.This effect is seen in Fig. 8a.Since also sc1t has a reduced RRR, over-stoichiometry appears to be indeed realized by the combination of excess Ce and Pd/Si vacancies.Excess Ce, just as Ce holes, creates Kondo disorder.In addition to reducing RRR, Kondo disorder is expected to successively suppress the temperature T max where the high-temperature incoherent Kondo scattering with an approximate ρ ∝ − ln T behavior crosses over to coherent Kondo scattering at low temperatures.In samples sc3andsc4Kondo disorder is so strong that ρ ∝ − ln T is observed in a wide temperature range, and ρ(T) shows only a tiny drop below 2 K. Figures 6b and 7 show that a sizable suppression of T max occurs in our sample series pc − (sc1t and sc5) -sc1b -sc2-sc3-sc4o n l yf r o msc2o n .T h u s , in addition to Kondo disorder, there appears to be a second effect influencing T max in the opposite direction.This can be identified as a volume effect: the Ce vacancies or (smaller) Pd atoms on Ce sites in Ce under-stoichiometric samples as well as the Pd or Si vacancies in the over-stoichiometric sample sc1t lead to a shrinkage of the crystal lattice, which is seen in Table 2 and in On the upper axis the corresponding pressure as estimated via the bulk modulus of Ce 3 Pd 20 Ge 6 [Nemoto et al (2003)] is given.The line represents a linear fit to the data and its extrapolation to T L = 0. From Ref. Prokofiev et al (2009).
Figure 9 shows T max (full dots) vs the relative volume shrinkage −∆V/V of our single crystals with respect to our polycrystal (pc) (bottom axis) and vs a hypothetical external pressure calculated via the bulk modulus [Nemoto et al (2003)] (top axis).The step-like, as opposed to continuous, decrease of T max with −∆V/V, addressed already above, is clearly seen.In order to understand this behavior we compare our results to a pressure study [Hashiguchi et al (2000)].We include the T max vs p data from this investigation on a polycrystalline sample under quasi-hydrostatic pressure in Fig. 9 (crosses and grey fit to these).From a value T max ≈ 20 K at p = 0, similar to the value for our polycrystal, T max increases continuously with increasing pressure.This behavior is typical for Ce-based heavy fermion compounds: with increasing pressure the hybridization between the 4 f electron of Ce and the conduction electrons increases and thus the Kondo temperature which is proportional to T max increases.Our data follow this trend only at low non-stoichiometry (low −∆V/V values) while at higher non-stoichiometry T max decreases quickly.Thus we conclude that the volume effect on T max dominates in our sample series pc -sc1t -sc1b -sc2-sc3-sc4u pt osc1b while the Kondo disorder effect dominates from sc2on.
Next we analyze the evolution of the ordering temperature T L along our sample series.
For this purpose T L , extracted from the specific heat data in Fig. 6a, is plotted vs RRR in Fig. 9b and vs −∆V/V (bottom axis) or p (top axis) in Fig. 8b.T L decreases continuously with decreasing RRR (Fig. 9b).This might be taken as indication for the strong influence of disorder on T L .However, in our sample series, the variation of RRR is intimately linked to a variation of ∆Ce (Fig. 8a) and thus to a variation of the lattice parameter a,t h e relative volume change −∆V/V, and the corresponding pressure p.Thus, an alternative Prokofiev et al (2009).
view of the situation is that not increasing disorder but decreasing volume is responsible for suppressing T L (Fig. 8b).To decide which effect is dominant we come back to our above discussion on Fig. 9a which revealed that the volume effect on T max dominates in the series pc − sc1t − sc1b − sc2 − sc3 − sc4uptosc1b while the disorder effect dominates from sc2on.
If this holds true also for the effect on T L it is Fig. 8b (without sc2) that captures the essential physics while Fig. 9b only displays implicit dependencies (except for sc2).
In a Kondo lattice picture the physical origin of the suppression of T L with p is that pressure increases the hybridization between conduction electrons and 4 f states, thus strengthening Kondo compensation, diminishing Ce magnetic moments, and suppressing the (most probably magnetic) ordering temperature T L .Extrapolating the T L vs p dependence of Fig. 8b to higher p suggests that a pressure induced quantum critical point is at reach for Ce 3 Pd 20 Si 6 .A linear extrapolation of the fit shown in Fig. 8b yields p c ≈ 0.5 GPa.Of course, actual low-temperature pressure measurements are needed to verify this possibility.Finally, we comment on a related study on the germanide compound Ce 3 Pd 20 Ge 6 where the influence of the starting composition on the physical properties was investigated [Kitagawa et al (1999)].A strong composition effect on C p and ρ was found to occur inside a narrow homogeneity range where a volume contraction of up to 0.7% takes place with increasing Pd content.In that work the concentration of Pd and not of Ce was concluded to be variable and responsible for the changing physical properties.It should be noted that the real composition of the resulting phases was not investigated.Thus, it is plausible that the increasing Pd content is accompanied by a decreasing Ce content.Under this assumption the composition effect on the properties of Ce 3 Pd 20 Ge 6 and Ce 3 Pd 20 Si 6 are indeed comparable.
CeRu 4 Sn 6
In Kondo insulators (frequently also referred to as heavy fermion semiconductors) a narrow gap develops at low temperatures in the electronic density of states at the Fermi level [Aepli & Fisk (1992)].While most Kondo insulators known to date adopt a cubic crystal structure (e.g., YbB 12 ,S m B 6 ,F e S i ,C e 3 Bi 4 Pt 3 ) a few compounds (e.g., CeNiSn, CeRhSb) are orthorhombic.These latter show anisotropic properties which suggest that the energy gap vanishes along certain directions in k-space.CeRu 4 Sn 6 , first synthesized by Das and Sampathkumaran [Das & Sampathkumaran (1992)], crystallizes in a tetragonal structure of space group I42m (a = 6.8810Å, c = 9.7520 Å, c/a = 1.4172) [Venturini et al (1990)].A peculiarity of this compound is that, in addition to the tetragonal (body-centered) cell with lattice parameters a and c there exists a quasi cubic (face-centered) cell with lattice parameters c' and c,w h e r ec' is the diagonal of the tetragonal plane which differs by only 0.2% from c.This makes it very difficult to orient single crystals in an unambiguous way.On the other hand it allows us to study a "tetragonal" and a "quasi-cubic" Kondo insulator within the same material which is very appealing.
Choice of the growth method and the growth procedure
The compound CeRu 4 Sn 6 melts incongruently.Melting followed by rapid quenching yields CeRuSn 3 ,R u 3 Sn 7 and a tiny amount of the title phase in the solidified material.The incongruent melting was confirmed by a DTA experiment which showed multiple peaks on the heating and on the cooling curves (Fig. 10).Since crystal growth from the melt is impossible for an incongruently melting compound we searched for an appropriate flux.Usually low melting metals are used as high temperature solvents for intermetallic compounds.For CeRu 4 Sn 6 this might be tin.CeRu 4 Sn 6 was shown to react with tin with the formation of other phases.To avoid possible contaminations by foreign elements we searched thereafter for a flux composition in the ternary Ce-Ru-Sn system (self-flux).
Unfortunately, the Ce-Ru-Sn phase diagram which would be helpful for the choice of the best flux composition has not been reported.A Ce-rich flux for the Ce-poor CeRu 4 Sn 6 is expected to lead to the formation of the highly stable phase CeRuSn 3 .Therefore, a Ce-free binary Ru-Sn mixture is more appropriate.Taking into account the Ru:Sn ratio of 2:3 in CeRu 4 Sn 6 the flux composition "Ru 2 Sn 3 " is optimal to maintain the Ru:Sn stoichiometry in the crystals.Moreover, the lowest melting (1160 • C) composition in the Ru-Sn binary diagram is the eutectic with the composition 42.5 at.%Ru and 57.5 at.%Sn [ Massalski (1990)] which is close to the element ratio 2:3.According to Ananthasivan (2002) the system is a (partially immiscible) liquid above 1200 • C in the composition range of 37-75 at.%Sn.Finally, the large excess of Ru and Sn in the melt is expected to suppress the formation of the stable Ce-rich 3 phase.The crystals were grown by the floating-solution-zone traveling heater method (THM) in a mirror furnace.The seed and the feed rods had the stoichiometric composition whereas the melt zone was a solution of CeRu 4 Sn 6 in Ru 2 Sn 3 .The growth rate was 0.3-1.0mm/h.The composition of the crystals along their length in the growth direction was investigated by SEM/EDX analysis.No noticeable deviation from the stoichiometric ratio 1:4:6 can be observed (Fig. 11).
Physical properties of CeRu 4 Sn 6 single crystals
Figure 12 shows the temperature dependence of the electrical resistivity, ρ(T),o fC e R u 4 Sn 6 on a semi-logarithmic scale.With decreasing temperature ρ first increases steeply (range 1), then passes over a maximum at about 10 K, increases again, albeit less steeply (range 2), and finally tends to saturate at the lowest temperatures.A possible explanation of this behaviour is a double-gap structure frequently encountered in simple semiconductors: a larger intrinsic gap visible at high temperatures (range 1) and a smaller extrinsic gap between impurity states and the band edge that dominates the low-temperature behaviour (range 2).Approximating ρ(T) between 120 and 300 K with exponential behaviour (ρ = ρ 0 exp(∆ 1 /2k B T), Arrhenius law) yields an energy gap ∆ 1 /k B = 125 K, sizeably larger than previously reported for polycrystalline samples [Brünig et al (2006); Das & Sampathkumaran (1992)].Fitting the data between 0.8 and 1.8 K with the same function yields ∆ 2 /k B = 0.1 K (see full red lines in Fig. 12 for both fits).While this gap value may seem incongruous with respect to the fit range, it has to be bourn in mind that the influence of the low-T gap on ρ(T) is expected to diminish by eventual thermal depopulation of the upper states towards T = 0 and hence the observed Now, having single crystals an investigation of the anisotropy of CeRu 4 Sn 6 became feasible.
The magnetic susceptibility was measured on an oriented single crystal in two mutually perpendicular directions (Fig. 13a).One of these directions is the crystallographic c axis, the other one is situated within the tetragonal plane.A pronounced difference is seen.For both directions Curie Weiss-type behaviour is observed at high temperatures, with an effective magnetic moment that is roughly consistent with the full effective moment of Ce 3+ ,a n d with the paramagnetic Weiss temperatures Θ c ≈ 395 K and Θ ⊥c ≈ 155 K for H c and H⊥c, respectively.In order to test whether anisotropy also exists within the quasi-cubic cell we prepared small single crystalline platelets (with geometries which allowed for specific heat measurements only) cut from one piece in such a way that three mutually perpendicular directions were obtained: for two samples a c' axis is perpendicular to the platelet plane, for one sample it is c.As explained above, our Laue diffractograms cannot identify which sample is which.Our specific heat measurements, however, allow to clearly identify the c-a n dc'-oriented samples: while the zero field data are very similar for all three samples a magnetic field applied perpendicular to the platelet planes induces sizable anisotropy.The difference is best seen by plotting the relative difference in specific heat induced by a magnetic field, (c p (B) − c p (0))/c p (0), which reaches a maximum of more than 70% at 3.5 K for B||c ′ but is below 15% for B||c at this temperature (Fig. 14).Two samples (sc1andsc2) show very similar behavior and must thus be c'-oriented while sc3 shows distinctly different behavior and is thus identified as the c-oriented sample.A possible interpretation of these data is that a narrow energy gap which is present along c' in zero field is suppressed/diminished by a field of 5 T in this direction.Since the same field applied along c leads to a much weaker increase of c p we believe that no or a much less field sensitive gap is present along c [Paschen et al (2010)].
CeAuGe
A sizable number of Ce-and Yb-based intermetallic compounds demonstrate quantum critical behaviour.Most of the compounds are antiferromagnets, whereas only a handful of Ce and Yb compounds with a ferromagnetic phase transition at low temperatures is known.The occurrence of quantum criticality in a ferromagnetic ground state is a subject of current debate.CeAuGe orders ferromagnetically at a relatively low Curie temperature of T C = 10.9K [Sondezi- Mhlungu et al (2009), Mhlungu & Strydom (2008)].Thus it can be expected that the magnetic order can be tuned or fully suppressed by modest variations in magnetic field or pressure.
279
Crystal Growth and Stoichiometry of Strongly Correlated Intermetallic Cerium Compounds www.intechopen.com CeAuGe is a phase with a homogeneity range.Stoichiometric and nearly stoichiometric CeAuGe adopts the NdPdSb structure, an ordered variant of the AlB 2 structure type.The unit cell is hexagonal (space group P6 3 mc) with the lattice parameters a= 4.4569 Å and c= 7.9105 Å [Pöttgen et al (1998)].At larger deviations from the elemental 1:1:1 ratio a phase with a slightly different structure forms.The non-stoichiometric CeAu 1−x Ge 1+x crystallizes inthetrueAlB 2 structure type and has about twice smaller unit cell.Unlike the ordered 1:1:1 phase, Au and Ge atoms are distributed statistically and the Au/Ge layers are not puckered but planar [Jones et al (1997)].Since CeAuGe is not cubic it is especially important to investigate its physical properties on oriented single crystalline samples.We have investigated the crystal growth of this phase of various stoichiometries by the floating zone technique.In the course of the growth experiments we encountered a severe non-stoichiometry problem.We report here on our efforts to diminish the deviation from the 1:1:1 stoichiometry and the segregation effects resulting from it.
Crystal growth using stoichiometric feed rods
As a starting point, growth from the stoichiometric melt was tried.We studied the evolution of the crystal composition during the crystallization by measuring the composition at the starting, middle and final part of the crystallized ingot by the EDX technique.The concentration profiles are represented in Figs. 15 and 16.At the left border of each panel the initial compositions of the respective polycrystalline feed rod is shown.The beginning of the crystallization corresponds thus to the length coordinate z = 0.As Fig. 15a shows, the crystals primarily crystallized from the stoichiometric melt (i.e. the first portion of the crystalline phase) have a non-stoichiometric composition with a reduced Au content and increased Ce and Ge contents.This leads to a change of the melt composition with an accumulation of Au and a depletion of Ce and Ge.As a consequence, the Au content increases the crystals and the Ce and Ge contents decrease in the course of further crystallization (Fig. 15a).Due to the composition change the lattice parameters change too (Fig. 15b).For the first solidified crystals the a-parameter is lower and the c-parameter is higher than the stoichiometric values (Fig. 15b).Further crystallization leads to a decrease of c and to an increase of a. Thus, the growth using a stoichiometric feed rod results in a non-stoichiometric single crystalline ingot which, along its length, is macro-inhomogeneous with respect to all three constituting elements.
Crystal growth using off-stoichiometric feed rods
In order to suppress the above discussed starting deviation in the element concentration in the crystals with respect to the melt we used a feed rod enriched in Au and depleted in Ge content, with the off-stoichiometric composition CeAu 0.96 Ge 1.04 .T h eprimary crystals obtained from this run appear to be nearly stoichiometric (Fig. 16a, z=0 mm), but in the course of further crystallization the composition again drifts away from the 1:1:1 stoichiometry: the Au content increases and the Ge content decreases, both summing up to a constant value.However, the Ce content remains constant along the whole solidified ingot (Fig. 16a).This fact is favorable for physical investigations on large crystals because non-stoichiometry of Ce is usually the most disturbing factor in heavy fermion systems.The crucial role of the Ce stoichiometry was demonstrated in Section 2 on Ce 3 Pd 20 Si 6 .Crystallization from the off-stoichiometric melt CeAu 0.96 Ge 1.04 seems to be most promising for the growth of stoichiometric homogeneous CeAuGe crystals, provided that the growth technique should be modified.In the floating zone method the melt-to-crystal volume ratio is very small.Therefore the segregation phenomena have very strong impact on the composition of the melt zone, and the crystal composition varies strongly during the growth.In other techniques, e.g. the Czochralski method, the melt-to-crystal volume ratio can be rather large, and the growth of homogeneous (also with respect to the Au/Ge ratio) crystals from the melt with a practically constant composition appears feasible.Crystallization from a more strongly off-stoichiometric melt CeAu 0.88 Ge 1.12 yielded strongly non-stoichiometric crystals with a large excess of Ge and a lack of Au, the Ce content remaining at nearly stoichiometric level (Fig. 16b.)While the primarily crystallized material is, according to SEM, single phased, the finally solidified ingot consists of two phases (Fig. 17, right).These two phases are the ordered and the disordered variants of the AlB 2 s t r u c t u r e( t h eN d P d S ba n dt h et r u eA l B 2 types).The material from the middle part of the ingot seems to be single-phased at first glance (Fig. 17, upper left), but a closer inspection under higher magnification reveals a two-phase pattern (Fig. 17
Conclusions
Single crystals of the Ce 3 Pd 20 Si 6 phase were grown from the melt (stoichiometric and slightly off-stoichiometric) under various growth conditions and from high-temperature solutions using Pd 5 Si as a flux.Ce 3 Pd 20 Si 6 melts quasi-congruently, i.e., the peritectic temperature is very close to the temperature of the complete melting.This fact follows from our DTA experiments and the observation of the melting zone during the growth process.The floating zone with stoichiometric composition was very unstable because of the low surface tension, which made the melt growth problematic.In contrast, the off-stoichiometric flux growth ran stably but resulted in non-stoichiometric single crystals.While the Si content varies only slightly for different crystals, the Ce and Pd contents do so sizably, the decrease of Ce being partially compensated by an increase in Pd.The existence of a homogeneity range in Ce 3 Pd 20 Si 6 is the reason for the strong variation of the properties of single crystals grown by different techniques.The sharpness of the lower (presumably antiferromagnetic) phase transition, its transition temperature T L , the residual resistivity, and the temperature T max of the (local) maximum in ρ(T) were shown to be measures of the crystal quality.Based on all these properties the upper part of the crystal grown from the stoichiometric melt (sc1t) and the whole volume of the crystal sc5 grown from a slightly off-stoichiometric melt can be ranked as having the highest perfection among all the grown single crystals.Their lattice parameters together with their compositions indicate that the slight off-stoichiometry is not dominated by Ce on the Pd sites but by Pd vacancies, which do not directly disturb the 4 f lattice.
282
Modern Aspects of Bulk Crystal and Thin Film Preparation www.intechopen.com CeRu 4 Sn 6 melts incongruently.Single crystals of CeRu 4 Sn 6 can be grown from Ru 2 Sn 3 flux.The grown single crystals show no marked deviation from the ideal stoichiometry, which indicates a very narrow homogeneity range of the phase.On the grown crystals anisotropies of the magnetic properties of two types were demonstrated: along the tetragonal unit cell axes and along the axes of a quasi-cubic unit cell.Single crystal growth of CeAuGe is complicated by a wide homogeneity range of the phase.Growth from the stoichiometric melt yields single crystals with an essential deviation from the stoichiometry.The composition of single crystals varies strongly along the growth direction.The crystal composition depends complexly on the melt composition and on the crystallisation temperature (the latter is a function of the former).Growth from the off-stoichiometric melt CeAu 0.96 Ge 1.04 is optimal for nearly stoichiometric crystals provided that the melt-to-crystal volume ratio is large enough for keeping the melt composition quasi-constant during the entire growth run.
Fig. 2 .
Fig. 2. Samples resulting from crystal growth experiments from the stoichiometric melt with upper and lower rod rotation.Left -Successfully grown single crystal (sc1); Right -The crystallized ingot experienced repeated dropping down of the melt because of the unfavorable combination of the high density and of the low surface tension of the melt.
Fig. 3 .
Fig. 3. Microstructure of sc2 before annealing.Left -Surface shell (darker top part) with a 5% higher Ce concentration; Right -An inclusion with a lower Ce concentration in the core (shown by arrow).
Fig. 4
Fig. 4. a) Phase diagram of the Ce-Pd-Sn ternary system (from Gribanov et al (2006)).Here Ce 3 Pd 20 Si 6 is denoted as τ 9 -phase.b) Magnified part of the phase diagram with the compositions of the solutions from that sc3andsc4weregrown.
Fig. 5. Deviation of the Pd and Si content, ∆Pd and ∆Si, from the exact 3:20:6 stoichiometry as function of the Ce non-stoichiometry ∆Ce.The dashed lines indicate the exact stoichiometry.The meaning of the broad grey lines is discussed in the text.Error bars indicate the standard deviations given in Table2.From Ref.Prokofiev et al (2009).
Fig.6.a) Specific heat divided by temperature C p /T plotted for all single crystals prepared here, and for a polycrystal[Strydom et al (2006)] for comparison, as function of temperature T on a logarithmic scale.The lower transition temperatures T L are taken here as position of the maxima.The maximum in the data for sc3 can be identified with T U (see text).b) Electrical resistivity of the Ce 3 Pd 20 Si 6 single crystals grown here normalized to the respective values ρ 200 K vs temperature T. The arrows indicate the positions T max of the (local) maxima.Data of a polycrystal[Strydom et al (2006)] are shown for comparison.From Ref.Prokofiev et al (2009).The electrical resistivities of all single crystalline samples and of one polycrystalline sample are shown in Figs.6b and 7 as function of temperature.While the resistivity of the
Fig. 7 .
Fig. 7. Relative resistivities of the bottom and top parts of sc1andsc5.
Fig. 8. a) Dependence of the lattice parameter a and the residual resistance ratio RRR on the deviation ∆Ce from the stoichiometric Ce content.The open symbol in the lower panel represents the data point of sc1t, mirrored through the ∆Ce = 0 line.The thick grey lines are guides to the eye.b) Lower transition temperature T L vs relative volume shrinkage −∆V/V of the different Ce 3 Pd 20 Si 6 single crystals with respect to the volume of the polycrystal pc.On the upper axis the corresponding pressure as estimated via the bulk modulus of Ce 3 Pd 20 Ge 6[Nemoto et al (2003)] is given.The line represents a linear fit to the data and its extrapolation to T L = 0. From Ref.Prokofiev et al (2009).
Fig.9.a) Temperature T max of the maximum in the electrical resistivity vs relative volume shrinkage − ∆V/V of our Ce 3 Pd 20 Si 6 single crystals with respect to the polycrystal (dots) and, for comparison, T max vs pressure p for the polycrystal ofHashiguchi et al (2000) (crosses and grey line, which is fit to the data).∆V/V and p are related to each other via the bulk modulus of Ce 3 Pd 20 Ge 6 .b) Lower transition temperature T L vs residual resistance ratio RRR for the different Ce 3 Pd 20 Si 6 samples.The line represents a quadratic fit to the data.From Ref.Prokofiev et al (2009).
along the crystal, mm
Fig. 11 .
Fig. 11.Element concentration profiles of the CeRu 4 Sn 6 single crystal in the growth direction.
Fig. 13.a) Temperature dependence of the inverse magnetic susceptibility, χ −1 (T),of CeRu 4 Sn 6 for the magnetic field µ 0 H = 0.4 T applied along the crystallographic c axis and within the tetragonal plane.b) Magnetic field dependence of the magnetization, M(µ 0 H), taken at 3 K for fields applied along the crystallographic c and c' axes.Fig. 13b from Ref. Paschen et al (2010).
278Modern
Aspects of Bulk Crystal and Thin Film Preparation www.intechopen.com
Fig. 15.a) Element concentration profiles along the growth direction at the crystallization from a stoichiometric feed rod.b) Profiles of the unit cell parameters a and c.
Fig. 16 .
Fig. 16.Element concentration profiles along the growth direction for crystal growth from off-stoichiometric feed rods with the compositions (a) CeAu 0.96 Ge 1.04 and (b) CeAu 0.88 Ge 1.12 .
Fig. 17.Microstructure of the ingot crystallized from the feed rod with the composition CeAu 0.88 Ge 1.12 .Left -middle part, right -upper part.The magnified (×25) image of the middle part shows a fine inhomogeneity of the material.
269
Crystal Growth and Stoichiometry of Strongly Correlated Intermetallic Cerium Compounds www.intechopen.com
Table 2 .
Composition and lattice parameter a of the investigated Ce 3 Pd 20 Si 6 samples. | 12,504.8 | 2012-01-13T00:00:00.000 | [
"Materials Science",
"Physics"
] |
General Power Flow Calculation for Multi-Terminal HVDC System Based on Sensitivity Analysis and Extended AC Grid
The power flow calculation of the multiterminal high voltage direct current (MTDC) system is essential for planning sustainable energy sources and power flow analysis for the MTDC system. However, the traditional unified methods require a large system of non-linear equations leading to low calculational efficiency. Also, for a large DC grid, there is a concern about the convergence of sequential methods. This paper proposes a general power flow calculation method for voltage source converter (VSC) based MTDC systems. Based on an extended topology of an AC grid, a generalized calculation model of the MTDC power flow is proposed. Then, a novel sensitivity analysis-based power flow (SAPF) method is proposed, in which the state variables of the extended AC grid are calculated via sensitivity analysis. With a smaller system of equations, the proposed SAPF method has less computational burden than the traditional unified methods and causes no convergence problem compared to sequential methods. To further improve the calculational efficiency, the sensitivity-based variable updating is adopted to accelerate the iterative process. By comparing with the existing methods in calculating power flow for different MTDC systems, including large systems with multiple AC/DC grids, the effectiveness and scalability of the proposed methods are verified.
I. INTRODUCTION
W ith the advantages of large transmission capacity and low power loss [1], the high voltage direct current (HVDC) technology has been applied in offshore wind power systems [2] and power transmission between asynchronous grids [3]. However, the HVDC transmission is usually with the back-to-back and point-to-point configuration [4], which limits the application of the HVDC system. With the development of the direct current (DC) circuit breaker and voltage source converter (VSC), the interest of the multiterminal high voltage direct current (MTDC) system increases. As a promising power transmission network, the VSC based MTDC system can realize the connection of multiple asynchronous power grids, DC sources and loads, which shows flexible power control ability and wide applications [5].
The power flow calculation of the MTDC system is the foundation of the steady-state studies of the MTDC system, such as adaptive power control [6] and static security analysis [7]. To calculate the power flow of the MTDC system, the outputs of DC sources, the power control of multi-VSCs and the power loss of VSCs need to be considered. Generally, the existing methods for power flow calculation of the MTDC system can be mainly divided into two categories: the sequential methods [8]- [13] and the unified methods [14]- [19].
In sequential methods, the power flow for the alternating current (AC) and DC grids is calculated separately and repeatedly, where the state variables of each VSC are set as coupling variables for determining the consistency of power flow at the AC side and DC side of the MTDC system. The main advantage of sequential methods is that the existing AC power flow algorithms can be adopted without any modifications. As presented in [8], the traditional Newton-Raphson (NR) method is used to solve the AC power flow efficiently, and the DC state variables of the MTDC system are iterated until convergence. In [9] and [10], the sequential method is applied for solving the power flow of the MTDC system under droop control. A double-layer loop iterative method is proposed in [11] to improve the convergence of the sequential method where the approximation is introduced into the calculation. In [12], the MTDC system is divided into smaller sub-systems and analyzed separately to improve computational efficiency. Also, to further improve the computational speed, the iteration of the sequential method is simplified by approximating the VSC current in [13]. In unified methods, solving the power flow of the MTDC system is to solve a large system of nonlinear equations which can be built by combining all the power flow equations of AC grids and DC grids. The results of the unified methods are generally with high accuracy as no assumptions are included in the unified method. In [14], the system of equations of the power flow calculation for a common MTDC system is built. Also, considering the power loss of the VSC, a general unified method for MTDC is presented in [15] where the mismatch equations are built with a general model of the VSC station. The NR method is adopted in [16] and [17] to solve the large system of equations built in the unified method where the DC slack control and droop control are considered. To further accelerate the calculation, the smooth This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ approximation of the Fischer-Burmeister is used in [18], but the computational accuracy is reduced by the approximation. In [19], the performance of the typical unified method is tested based on simulations where the precision of the unified method is verified.
However, in certain situations, sequential methods may lead to the convergence problem [14]. For example, when the MTDC system contains multiple VSCs with the droop control or the system is with multiple AC and DC grids, multiple coupling variables need to be determined by the iterative process which does not converge until all coupling variables converge [11]. Therefore, with the increase of the number of VSCs, it would be a challenge to make a large number of coupling variables converge at the same time [12]. Although the approximation is adopted in [11] to improve the convergence of the sequential method, approximations will reduce the calculation accuracy. In unified methods, although there is no need to set coupling variables, it is computationally inefficient to solve a large system of equations, which limits the scalability of unified methods for optimal power flow calculation [20] and dynamic analysis [21] of a large system. Also, when the number of buses and VSCs increases, the computational burden of the unified method will increase nonlinearly and dramatically [10]. Besides, the modification of power flow equations in unified methods introduce more variables into the NR iteration [18], which could lead to the difficulty of finding suitable initial values for the convergence [22]. Therefore, to improve the previous methods, this paper proposes a novel power flow calculation method for the MTDC system. The main contributions of this paper are as the following, 1) Based on an extended topology of the AC grid, a general calculation model of the MTDC power flow is presented. One advantage of this topology is that the calculation equations of the AC grid do not need to be modified. Also, the power flow calculation model is built by combining the sub-model of each DC bus, which can be generalized to multiple large DC grids. 2) A novel sensitivity analysis-based power flow (SAPF) method is proposed to calculate the power flow for MTDC systems, in which the state variables of the extended AC grid are calculated via sensitivity analysis. Thus, the power flow calculation model can be represented by a small-scale system of equations, contributing to less computational burden and higher calculational efficiency compared to unified methods. In addition, the NR method is used in the SAPF avoiding the convergence problems occurred in the sequential method. 3) To further improve the calculation speed of the SAPF method, a fast SAPF (F-SAPF) method is proposed, in which the AC power flow is obtained by the sensitivity analysis instead of the traditional AC power flow calculation. The results obtained by the F-SAPF method are as accurate as that of the SAPF method because of the unchanged mismatch equations. 4) Different MTDC systems, including the scenarios of multiple DC grids and large AC networks, are studied to verify the effectiveness of the proposed method whose performance is compared with other existing methods. Also, the power flow variations and contingencies in an offshore wind power system are investigated by the proposed method. The rest of the paper is organized as follows. Section II presents the extended AC grid and the power flow calculation model of the MTDC system. Then, the SAPF and F-SAPF methods are proposed in Section III. Case studies are implemented in Section IV, and Section V concludes the paper.
A. The MTDC System and the VSC Station
The structure of the generalized MTDC system is presented in Fig. 1, where the DC grid includes DC transmission lines, VSCconnected DC buses (namely VSC buses) and pure DC buses (the DC buses without any VSCs connected). The DC sources denote DC power supplies such as photovoltaic (PV) power stations and battery storage systems. A complex MTDC system normally consists of multiple AC and DC grids, as shown in Fig. 1. Fig. 2 shows the steady state model of the VSC station connecting the k-th AC bus in the AC grid and i-th DC bus in the DC grid, where R T,i and X T,i are the equivalent resistance and reactance of the transformer, respectively. R c,i and X c,i are the equivalent resistance and reactance of the phase reactor respectively. Also, P k and Q k denote the active and reactive power injected to the gird at the k-th AC bus respectively. The magnitude and angle of the voltage at the k-th AC bus are V k and θ k respectively, and V f,i , θ f,i denote the magnitude and angle of the voltage at the node f i respectively.
According to [8], [23] and [24], the total loss of the VSC can be described as where a 0,i , a 1,i , a 2,i are three loss coefficients of the VSC connected to the i-th DC bus. I c,i is the magnitude of the converter current which can be calculated as where P f,i , Q f,i , S f,i are the active power, the reactive power and the magnitude of the apparent power, respectively, from the node f i to c i . And V f,i is the magnitude of the voltage at node f Fig. 2.
B. The Extended AC grid
The nodal balance for each AC bus in an N-bus AC grid can be described as where P z and Q z denote the active power injection and reactive power injection at the z-th bus (z = 1, 2, · · · , N) respectively. V z and θ z denote the magnitude and angle of the voltage at the z-th bus respectively. In (3) and (4), G zj and B zj are the real part and imaginary part of the admittance of the branch between the z-th and j-th AC bus respectively. So, by applying (3) and (4) to all buses in the N-bus AC grid, a system of 2N equations can be built to calculate the AC power flow. Then, we suppose this AC grid is connected to an MTDC system and one of the original AC buses, namely the k-th AC bus, is connected to the i-th DC bus via a VSC station, as shown in Fig. 2. In this paper, an extended AC grid is built by combining the transformer, filter, node f i and the N-bus AC grid. This extended model of the AC grid is different from the model presented in [11] and [19] in which both node f i and c i are considered as the additional AC buses. Also, in [11] and [19], Equations (1)-(4) are all needed to build the nodal balance model for node c i in each VSC station, which causes complexity to build the AC power flow model and reduces the computational efficiency of the AC power flow calculation. In the extended model proposed in this paper, only the node f i is regarded as an additional AC bus to the original AC grid, and the filter is regarded as an additional susceptance of the transformer branch between the node f i and the k-th AC bus. In this way, the node f i can be regarded as the (N+1)-th AC bus of this extended AC grid (which has N original AC buses). Without adding (1) and (2) to the AC power flow model, (3) and (4) are adequate for building the nodal balance model at the (N+1)-th AC bus. Note that, considering the direction of P f,i and Q f,i shown in Fig. 2, the active power injection at the (N+1)-th AC bus is P N+1 = −P f,i and the reactive power injection is Q N+1 = −Q f,i .
Normally, for each AC bus (including additional AC buses), the four variables P z , Q z , V z , θ z shown in (3) and (4) can be classified as input variables and state variables. Table I shows the input and state variables of the z-th AC bus according to its bus type, where two input variables are referred to as u 1,z , u 2,z and two state variables are referred to as y 1,z , y 2,z . Table I can be applied to each AC bus in the extended AC grid including additional AC buses. Note that the bus type of each additional AC bus depends on the reactive power control of the connected VSC, which will be presented in Part C of this Section.
C. The DC Power Flow Calculation and Power controls
The active power P dc,i from the VSC to the i-th DC bus in an n-DC-bus MTDC system can be presented as where P G,i is the power injection from the DC sources to the i-th DC bus. Also, P L,i is the DC load and U dc,i is the DC voltage at the i-th DC bus. Y dc ij is the admittance between the i-th and j-th DC bus in this MTDC system.
The control of a VSC is with the corporation of an inner d-q current control and an outer power control, in which the active power and reactive power are controlled independently [25]. According to [3], [13], [25] and [26], the control objectives of the outer power control are the variables of the AC side, including P f,i , Q f,i , V f,i of the VSC station as shown in Fig. 2. From a steady state point-of-view, the outer power control (reactive and active power control) is the key to calculating the power flow of the MTDC system. 1) Reactive power control.
One of the following strategies is generally used to control the reactive power of a VSC: f,i and the node f i is regarded as an additional PQ bus of the extended AC grid); f,i and the node f i is regarded as an additional PV bus of the extended AC grid.
The superscript * denotes the reference value of control objectives. According to the model of the extended AC grid, the node f i is numbered as the (N+i)-th AC bus in the extended AC grid. As shown in Table I, with both the P-Q mode and P-V mode, the controlled variable (Q f,i or V f,i ) can be referred to as u 2,N+i . So, the reactive power control can be described as 1) Active power control (also known as DC voltage control). In this paper, the two most commonly used modes of DC voltage control are introduced: a) DC slack control: one of VSC buses is selected as the slack bus whose DC voltage is controlled to be constant, as shown in Fig. 3(a), and the other VSC buses are the non-slack DC buses whose P f,i is controlled to be constant, b) Droop control: for each VSC, either P f,i or I dc,i is controlled to be constant, or where K i is the droop coefficient of the VSC connected the i-th DC bus, and I dc,i is the DC current from the VSC to this DC bus shown in Fig. 2. The first type of droop control is called droop voltage-power (U-P) control, as shown in (9) and Fig. 3(b). The other one is called droop voltage-current (U-I) control, as shown in (10) and Fig. 3(c). According to (10), P dc,i at the i-th DC bus can be calculated by
D. Power Flow Calculation Model of the MTDC System
The power flow calculation model for an n-DC-bus MTDC system should be built on the basis of different control modes. Suppose the i-th DC bus is 1) a pure DC bus: P dc,i = 0 as the pure DC bus is not connected to the AC grid, so that an equation, namely f 1,i , for this pure DC bus can be obtained by substituting P dc,i = 0 into (5) as So, the following simultaneous equations F s,i can be built: where
2) a VSC bus with the DC slack control:
The active power balance at node f i can be described as By substituting (1) and (5) into (14), the f 1,i of the i-th DC bus can be built as Then, another equation, namely f 2,i , of the i-th DC bus can be obtained from (2): As the node f i is considered as the (N+i)-th bus (PQ or PV bus) of the extended AC grid, the set {V f,i , P f,i , Q f,i } can be referred to as the set {u 2,N+i , u 1,N+i , y 2,N+i } according to Table I. Also, according to (6), u 2,N+i is controlled to be the pre-set value. Therefore, the independent variables of (17) can be summarized as {I c,i , u 1,N+i , y 2,N+i }.
By combing the (15) and (17), the following simultaneous equations can be built for the i-th DC bus: 3) a VSC bus with the droop control: Two equations f 1,i and f 2,i for the i-th DC bus are as the same as (15) and (17) respectively. Then, another equation, namely f 3,i , is added. The f 3,i can be built according to the type of the droop control: if the connected VSC is under droop U-P control, by (9), or under droop U-I control, by substituting (11) into (5), Similarly, the following simultaneous equations can be built for the i-th DC bus with droop control: According to (13), (18) and (21) The sensitivity analysis is an effective tool to analyze the state of power systems, which has been used in the power flow coordination [27], line losses determination [28], convergency analysis [29], optimal power flow calculation [30], etc. In this Section, the sensitivity analysis is adopted to calculate the power flow for the MTDC system.
A. Sensitivity Analysis for the Extended AC Grid
Suppose an N-bus original AC grid is connected to an n-DC-bus MTDC system which contains m VSC buses. Thus, an extended AC grid containing N+m AC buses can be built by adding m additional AC buses (the node f i in each VSC station) to the original AC grid. By applying (3) and (4) to all buses in the extended AC grid, the following simultaneous equations can be built to solve the AC power flow: where u and y are the vectors consisted of input variables and state variables of all AC bus respectively. According to Table I, u = [u 1,1 , u 2,1 , u 1,2 , u 2,2 , · · · , u 1,N +m , u 2,N +m ] T and y = [y 1,1 , y 2,1 , y 1,2 , y 2,2 , · · · , y 1,N +m , y 2,N +m ] T . Normally, the AC power flow calculation is to solve y by inputting u into (22), which can be described as where G( · ) denotes the process of the AC power flow calculation. Suppose the extended AC grid is on a state with [u (k) , y (k) ], by the first-order Taylor polynomial, where Δu and Δy are the variation of input variables and corresponding state variables. By substituting (22) into (24), where S ac = Δy/Δu is called the sensitivity matrix whose entries describe the relationship between the variation of each state variable and input variable, which can be expressed as where B( · ) denotes the process to calculate S ac . Based on the data of the N-bus original AC grid, the input variables (u 1,1 , u 2,1 , u 1,2 , u 2,2 , · · · , u 1,N , u 2,N ) of N original AC buses are normally pre-known. Instead, for the m additional AC buses, the input variables (u 1,N +1 , u 2,N +1 , u 1,N +2 , u 2,N +2 , · · · , u 1,N +m , u 2,N +m ) are unknown. According to the reactive power control of each VSC as shown in (6), u 2,N +i = u * 2,N +i (i = 1, 2, 3, · · · , m) for these additional AC buses. Thus, the unknown variables in u can be summarized as u 1 = [u 1,N +1 , u 1,N +2 , · · · , u 1,N +m ] T , which means u can be determined once u 1 is given, namely u = g(u 1 ). By rewriting (23), the AC power flow calculation for the extended AC grid can be presented as Similarly, (27) can be rewritten as
B. The SAPF Method
The basic idea of the SAPF method is to build a small-scale system of equations based on the equations of DC buses, and the mismatch equations of the NR method are modified by using the sensitivity matrix. Taking the n-DC-bus MTDC system mentioned in Section III Part A as an example, suppose the m VSC buses are with the DC slack control and the DC-1 (the 1st DC bus) is set as the DC slack bus. For each DC bus, simultaneous equations F s,i can be built based on its type, as shown in Section II Part D. Thus, by combining F s,i of each DC bus, the power flow calculation model of the MTDC system can be built, namely F dc , which is a system of equations shown as where I c = [I c,1 , I c,1 , · · · , I c,m ] T and y 2 = [y 2,N +1 , y 2,N +2 , · · · , y 2,N +m ] T . Then, a new matrix S MT DC =Δy 2 /Δu1 is set, whose entries can be extracted from S ac as shown in (31). (30), the NR method is adopted, and the mismatch equations of (30) can be presented as Then, by substituting S MT DC =Δy 2 /Δu1 into (32), Fig. 4. The sensitivity analysis implemented in the SAPF method.
where X = [I T c , U T dc , u T 1 ] T , and J denotes the Jacobi matrix. According to (33), the sensitivity analysis is implemented between NR iterations for updating y 2 , as shown in Fig. 4. Note that, to reduce the computational burden, y obtained by each AC power flow calculation (represented by (28)) is adopted to be the initial value of the next AC power flow calculation in the sensitivity analysis, as the process represented by the green line in Fig. 4. The detailed steps of the SAPF method are presented in Fig. 5 where ε is the convergence tolerance.
Similarly, the calculation model presented in (30) can also be built for the MTDC system with the droop control, where the simultaneous equations F s,i of the VSC buses can be obtained by (21). Then, the SAPF method can also be implemented by following the steps shown in Fig. 5. For the MTDC system consisting of multiple DC grids, the calculation model F dc shown in (30) can be built by combining the simultaneous equations F s,i for all DC buses from different DC grids. To solve F dc , the SAPF method are also applicable by following the steps shown in Fig. 5. The multiple DC grids should be seen as a whole for building (5), where the disconnection of multiple DC grids can be represented by setting the admittance between disconnected DC buses to zero. For the MTDC system consisted by multiple AC grids (namely AC-1, 2, · · · , M), multiple S ac (S AC−1 ac , S AC−2 ac , · · · , S AC−M ac ) and corresponding S MTDC (S AC−1 MT DC , S AC−2 MT DC , · · · , S AC−M MT DC ) can be generated. Therefore, the SAPF method can be implemented as Δy 2 in mismatch (33) is calculated by are the corresponding u 1 in AC-1, 2, · · · , M. Thus, there are multiple sensitivity analysis processes for multiple AC grids in the SAPF method. Once the number of AC buses in one of the AC grids increases, it will only increase the computational burden in the sensitivity analysis for this AC grid, where the sensitivity analysis for other AC grids is not affected. While, in the unified method, the increased number of the AC buses will directly increase the scale of the large simultaneous equations for the whole system, which will significantly increase the computational burden of the unified method.
For the MTDC system with the DC slack control, the number of equations in (30) is m+n, and this number will be 2m+n for the MTDC system with the droop control. By introducing the control equations into the calculation model represented by (30), the power flow of the MTDC system with other control methods such as voltage margin control [31] and droop control with dead-band [32] can also be solved by the SAPF method. In addition, the model of the extended AC grid can also be built for the line-commuted converter (LCC) stations which is similar to the VSC stations except for the inner structure of the converter [33]. The sensitivity analysis shown in Fig. 4 can be used to determine the state variables of the extended AC models for both the LCC and VSC side. Therefore, the SAPF method can potentially solve the power flow problem for the hybrid LCC-VSC HVDC system, where the corresponding control strategies [33] of LCC stations will be applied to build the calculation model in (30).
C. The F-SAPF Method
In the iterative process of the SAPF method, the AC power flow calculation represented by (28) is implemented repeatedly to update y. To further improve the calculational efficiency, a fast SAPF (F-SAPF) method is proposed in this part. In the F-SAPF method, y is updated by sensitivity analysis instead of the AC power flow calculation, which can be described as y (k+1) = y (k) + Δy (k) = y (k) + S y · Δu (k) 1 (35) where S y = Δy/Δu1, whose entries can also be extracted from S ac : The detailed steps of the F-SAPF method are presented in Fig. 6. The AC power flow calculation, by (28), is implemented only twice in the F-SAPF method (once before the iteration and once after the iteration) and is not required in the iterative process, which greatly shortens the calculation time. Although y updated in the F-SAPF and SAPF methods will be a little different in each iteration, the accuracy of the F-SAPF method is high because the value of Δu1 in each iteration is small and the mismatch equations are the same as that of the SAPF method. Also, when the iteration converges, an AC power flow calculation will be implemented for correcting y as shown in Fig. 6. The number of iterations of the F-SAPF method might be different from the SAPF method, and this specific difference will be presented in case studies.
A. Case A: 3-Bus Small Test System
In this case, a small MTDC system with 3 DC buses (the total number of DC buses: n = 3) is used to illuminate the detailed calculation process of the SAPF and F-SAPF methods, in which the AC side is based on the CIGRÉ NORDIC 32-A grid [34]. This 32-bus AC grid (the total number of original AC buses: N = 32) can be considered as a scaled down version of the Swedish grid with three areas (North, Central, South), and the AC buses of this grid are renumbered as shown in Fig. 7. Three buses (No. 2,16,18) from different areas are connected to a 3-bus ±320kV MTDC system with DC slack control, and each DC bus is a VSC bus (the total number of VSC buses: m = 3). The control modes and corresponding references are presented in Table II. The resistance of DC cables is 0.012 Ω/km, and the parameters of VSC stations are shown in Table III [14] where B f,i denotes the susceptance of the filter. Based on the extended AC model presented in Fig. 2, the node f 1 , f 2 , f 3 of VSC-1, 2, 3 are considered as the additional AC buses and numbered as the No. 33, 34, 35 bus of the extended AC grid respectively. Thus, the original 32-bus AC grid is extended to a 35-bus grid where the three additional AC buses are marked as green in Fig. 7. Accordingly, the sensitivity matrix S ac of the extended AC grid can be built by (26). The equations of each bus for building the calculation model represented by (30) are summarized in Table IV. Some independent variables in (30) are pre-set based on the power control of VSCs, which can be summarized as 1) For the DC-1 bus (i = 1), as the slack bus, U dc,1 = U * dc,1 by (7). Also, with the P-Q mode of the reactive power control, u 2,33 = u * 2,33 by (6), where u 2,33 denotes Q 33 (Q f,1 , as the AC bus 33 is renamed from the node f 1 of VSC-1) of the AC grid according to Table I. 2) For the DC-2 and DC-3 bus (i = 2 and 3), as the non-slack bus, P f,2 = P * f,2 and P f, 3 = P * f,3 by (8). Similarly, with the P-V mode, u 2,34 = u * 2,34 and u 2,35 = u * 2,35 by (6), where u 2,34 and u 2,35 denotes V 34 and V 35 (V f,2 and V f,3 ) respectively. So, considering the pre-set variables, (30) can be built and then simplified as where Both S MT DC and S y can be extracted from S ac which can be calculated by inputting u and y into (29) in each iteration. The results are obtained after 5 NR iterations of both methods. Table V shows numerical results of the key independent variables in (33) with iterations. And the final result of y and y 2 of the F-SAPF method is obtained by inputting converged (28), as the steps shown in Fig. 6. From Table V, the results of the F-SAPF method are the same as that of the SAPF method except for the difference of 0.004% on P f,1 .
B. Case B: 6-Bus MTDC System
In Case B, the SAPF and F-SAPF methods are used to calculate the power flow of a 6-bus MTDC test system (n = 6). The results obtained by a typical unified method [19] are taken as the benchmark, and another sequential method proposed in [13] is adopted as the comparison. This MTDC system contains four VSC buses (m = 4), two pure DC buses (the DC buses without any VSCs connected), a PV power station, and a DC load of electric vehicle (EV) charging station. As shown in Fig. 8, the four VSC buses (DC-1, 2, 3, 4) are connected to the No. 8, 10, 30, 81 bus of an IEEE 118-bus system (N = 118) [35]. According to the extend AC model presented in Section II, the node f i in each VSC bus is considered as an additional AC bus, and these four additional AC buses are named as additional AC-119, 120, 121, 122 bus respectively, as shown in Fig. 8. Table VI shows the control references of VSCs. Also, two types VSCs, the half-bridge and full-bridge modular multilevel converter (MMC), are used in this case, of which the parameters and loss coefficients [13] are given in Table VII.
The calculation results of the four methods are presented in Table VIII. With DC slack control (DC-1 as the slack bus), the DC bus voltages obtained by the SAPF and F-SAPF methods are the same as the benchmark obtained by the method used in [19]. Instead, the sequential method in [13] has computational errors, which is because the converter current of each VSC is approximated. With droop control, the SAPF method and the method used in [19] also obtain the same results of DC bus voltages, while the results obtained by the F-SAPF method are very close to that of the SAPF method and the method used in [19] (only with an acceptable error of 0.001% on U dc,2 ). Meanwhile, the method used in [13] shows a relatively low accuracy. In Table VIII, the CPU time of the F-SAPF method is 63%, 70%, 72% less than that of the other three methods with the DC slack control, respectively, and 67%, 74%, 56% less with the droop control. This is because the calculation of the AC power flow in each iteration is replaced by the faster sensitivity analysis in the F-SAPF method. In addition, the numbers of iterations in the SAPF method, F-SAPF method and the method used in [19] are similar, which is because these three methods all iterate in a second-order speed based on the NR method with the same initial values. Although these three methods are with a similar number of iterations, the CPU time of each iteration of the method used in [19] is longer due to a larger system of nonlinear equations that needs to be solved [10]. Besides, the calculation speed of the method in [13] is faster than that of the SAPF method with droop control, but its calculation error is also larger, and more iterations are required. To summarize, in Case B, the SAPF and F-SAPF methods proposed in this paper are reliable in term of accuracy, and the F-SAPF method has a significantly higher computational efficiency than other methods.
C. Case C: MTDC System With 2 AC Grids
In Case C, a New England 39-bus AC grid [36] is connected to the IEEE 118-bus system by extending the existed 6-bus MTDC system in Case B, where three scenarios with different topologies of the MTDC grid are presented in Fig. 9. In scenario (a), a VSC bus (DC-7) is added to build a 7-DC-bus system (n = 7, m = 5) with 2 AC grids (N = 118+39). In scenario (b), two VSC buses (DC-8, 9) and a pure DC bus (DC-10) with another PV power station are added to build another MTDC gird which is connected to the New England 39-bus AC grid. Thus, this power network contains two AC grids (N = 118+39) and two separated DC grids (n = 7+3, m = 5+2). In scenario (c), another two lines (yellow lines shown in Fig. 9) are added to connect the two separated DC grids into one DC grid (N = 118+39, n = 10, m = 7). In Case C, the newly added VSC-5, 6 and 7 are connected to bus No. 39, 14, 23 of the New England 39-bus AC grid respectively, and all VSCs are with the droop control. Detailed control modes and reference values of VSC-5, 6, 7 are Table IX, and the parameters of them are consistent with those of the half-bridge MMC shown in Table VII. Table X shows the power flow calculation results of the test system in these three scenarios, where U dc , V f and P loss dc denote the average voltage of DC buses, average voltage of additional AC buses, and power loss of the MTDC grid (including the power loss of VSCs) respectively. Compared to scenario (a), the CPU time of the SAPF, F-SAPF and the method used in [19] in scenario (b)/(c) increases by 9%/12%, 8%/10% and 16%/17% respectively. In terms of calculation accuracy, for all three scenarios, the calculation results of the SAPF method and the method used in [19] are still consistent except for the 0.1% differences in P loss dc . Also, the difference between the results obtained by the F-SAPF method and the method used in [19] is less than 0.3%.
In order to study the influence of the initial values X (0) of the power flow calculation on the CPU time and convergence of given methods, the initial values of the power flow calculation are selected in two ways and tested based on scenario (b). The first way to select initial values is by where X * is the accurate result of the power flow obtained by the previous simulation, and c is the proportion coefficient. Fig. 10(a) shows the CPU time cost by the SAPF, F-SAPF and the method used in [19] with c = 1, 2, 3, · · · , 20, where the calculation of the method used in [19] does not converge with c = 1 and c = 2. Besides, the CPU time of the three methods is less when the initial values are closer to the accurate results X * . Also, the CPU time of the F-SAPF method is the least, as shown in Fig. 10(a). The initial values selected by another way are the random number in the interval of (0.5, 1.5). Fig. 10(b) shows the CPU time with the random initial values. Similarly, the F-SAPF method costs the least time to calculate the power flow in the 10-bus DC system compared to other two methods. In Fig. 10, it can be concluded that the small initial values will increase the calculation time and may cause the calculation of the NR method to be non-convergent.
D. Case D: Offshore Wind Power System
In this case, a fraction of a developing offshore wind power grid called the North Seas Countries Offshore Grid [37] is introduced, in which offshore wind farms of three countries (the United Kingdom: U.K., The Netherlands: NL and Germany: DE) are connected to onshore grids by DC submarine power cables (with resistance 0.0195 Ω/km) and VSC stations. Fig. 11 shows the topology of this system (n = 9, m = 6). The proposed methods are implemented to study the impact of the varying wind farm outputs and presumptive contingencies on the power flow of the offshore wind power system. Three independent IEEE 300-bus systems [35] are used to represent the U.K., NL Table XI. Fig. 12 shows the variations of power flow indicators with the changing output of the U.K. offshore wind farm, while the other two wind farms are with the rated power output (1000MW). In Fig. 12(a), the average voltage of DC buses drops when the output of the U.K. wind farm decreases. Compared to the droop control, the DC slack control maintains the average DC voltage at a higher level. Also, a higher average voltage of DC buses can be observed under droop control with a higher droop coefficient. The opposite trend can be observed in Fig. 12(d) that the maximum current of DC branches rises as the decreasing output of the U.K. wind farm. Besides, the DC slack control leads to more power losses in the offshore wind power system than the droop control as shown in Fig. 12(b). In Fig. 12(c), the inverter power of DE VSC with DC slack control shows a more is the maximum current among the DC cables. As shown in Table XII, the CPU time of the F-SAPF method is over 70% less than that of the other three methods, showing the high efficiency of the F-SAPF method. The power flow results obtained by the SAPF, F-SAPF and the method used in [19] are almost the same while a relatively large difference can be seen between the results obtained by the method in [13] with other methods, which is similar to the observations in Case B and C. In addition, the CPU time of all methods for the calculation with droop control is larger than that of the DC slack control. This is because the number of equations for the droop control is more than that of the DC slack control.
According to the power flow results shown in Table XII, when the U.K. wind farm is forced to an outage, the total power loss of the DC grid drops by 21% with droop control but increases by 77% with DC slack control. Also, with DC slack control, the inverter mode of the DE VSC needs to be adjusted to the rectifier mode after the outage of the U.K. wind farm, which is difficult to implement in a short time. On the contrary, with the droop control, three onshore VSCs in this MTDC system are maintained in stable operation status. When the NL or DE onshore VSC is offline, the offshore wind power system can keep stability only by the droop control.
E. Case E: the CIGRÉ B4 DC Grid With Large AC Systems
To evaluate the performance of the proposed methods for large networks, the CIGRÉ B4 DC grid [38] is adopted in this case, which contains 15 DC buses (n = 15), including 11 VSC buses (m = 11) and 4 pure DC buses, as shown in Fig. 13. This DC grid shows the ability of the MTDC system to integrate renewable generation from offshore grids into different synchronous networks. In this case, the DC-DC converters in the MTDC system are considered as DC sources with constant power flow to the connected DC buses. Also, the five AC offshore VSCs (Cm-C1, Cb-C2, Cb-D1, Cm-E1, Cm-F1) are set to be with DC slack control, and corresponding DC buses are non-slack DC buses. The other VSCs are with the droop U-P control where K i = 40 and U * dc,i = 1 p.u. The VSC Cm-A1 and Cm-B2 are set to be working in PQ mode with Q * dc,i = 0, and the other VSCs are working in PV mode where V f,i = 1 p.u. The setpoint of P * dc,i and parameters of the transmission line can be found in [38]. The VSCs in this case are all the half-bridge MMC. Note that, there is a situation where two VSCs are connected to the same AC bus, where two additional AC buses will be added and connected to this AC bus. In this case, the CIGRÉ NORDIC 32-A system [34], a 3012-bus AC test grid [35], and a scaled down version of the European system model [39] containing 2869 buses are used for representing the real AC grids. Different AC grids are connected to the bus Ba-A0 and Ba-B0 shown in Fig. 13, in which the connection buses of these large AC grids are shown in Table XIII. The power injections from AC bus Ba-A1, Ba-B1, Ba-B2, Ba-B3 to the DC grid are denoted as P AC Ba−A1 , P AC Ba−B1 , P AC Ba−B2 , P AC Ba−B3 respectively, which are obtained by three methods as shown in Table XIII. Based on the results shown in Table XIII, it can be observed that the impact of the scale of the AC grid on the accuracy of the proposed SAPF and F-SAPF methods is limited, as the difference between the results by the SAPF method and the method used in [19] is less than 0.15%. Besides, with the significantly larger AC systems, three methods cost a longer CPU time to calculate the power flow than previous cases. In the calculation with N = (32+2869), the CPU time of the F-SAPF method is less by 69% and 74% than that of the SAPF method and method used in [19], respectively, which shows the high efficiency of the F-SAPF to calculate the power flow for the MTDC system with large AC grids. In the calculation with N = (3012+2869), compared with the results with N = (32+2869), the CPU time of the three methods is increased by 91%, 78%, 208% respectively, which shows the calculational burden caused by introducing the very large AC system into the calculation. Because the proposed SAPF and F-SAPF methods can separately generate the sensitivity matrix for each AC grid instead of combing the two AC grids into a large system as the method used in [19], the increase in the CPU time of these two methods is significantly less than that of the method used in [19].
V. CONCLUSION
This paper proposes a novel power flow calculation method for the MTDC system based on the sensitivity analysis and the extended AC grid. In the proposed power flow calculation model, the mismatch equations of the DC side have been built according to the control mode of the VSCs, while the AC grid has been extended without any change to the AC power flow calculation. The sensitivity analysis has been adopted in the proposed SAPF method which has higher calculation efficiency than the unified method and causes no convergence problem compared to the sequential method. To further improve the calculation efficiency, the F-SAPF method has been proposed, where the time-consuming AC power flow calculation in each iteration of the SAPF method has been replaced by the faster sensitivity-based calculation. A small test system has shown in Case A for presented the detailed calculation process with numerical results. In Case B, the accuracy and higher computational speed of the proposed SAPF and F-SAPF methods have been shown by comparing with the other two existing methods. Case C has presented the performance of the proposed methods with different topologies of the MTDC system. With high convergence, the wide applicability and high robustness of the proposed method have also been illuminated in Case D, where the proposed methods have been applied to determine the power flow variations of an offshore wind power system with the changing output of offshore wind farms. It also shows the advantage of droop control over DC slack control in maintaining power flow stability in given contingencies. The accuracy of the proposed methods to calculate power flow for large power networks has been presented in Case E, where the comparison of the results obtained by different methods has also shown the advantage of the F-SAPF method in computational efficiency with the scenarios of large power networks. | 10,858.4 | 2022-10-01T00:00:00.000 | [
"Physics"
] |
Space Syntax in Mixed Reality Gaming Applications
With the rapidly growing interest in AR, grows the motivation to overcome some of the problems facing the implementations of the technology. The main challenge encountered in the building of large-scale mixed reality AR games is the uniqueness of the spatial settings in which the game will be experienced by the user. Game designers will require data of the spatial settings to determine game object placement, events and narrative flow. The problem arises because the designers are not aware of the physical environment in which the game will be played. In our research, we address this problem and take an approach to solving it by using Space Syntax techniques. We demonstrate the use of this technique, using a proof-of-concept game called Adventure AR.
Introduction
With no prior knowledge of the physical spaces in which the game will be experienced, a mechanism is needed, to allow level design to adapt to different physical environments. To do this, our research seeks to use Space Syntax analysis to define the spatial structure of the area in which the game will be played and use the results to place game assets according to the design of the environment. In this paper we will focus on acquiring and encoding the Space Syntax data that includes the connectivity, visual complexity, and openness graphs. Using this data and a 2D/3D model of the floor or region the game will be played in, we will develop a RGBA bitmap representing Space Syntax values. A calibration step is performed for scaling the real world to the virtual space. Finally, the retrieved data for the spatial attributes (Visual Complexity, Openness, connectivity) can be displayed on the Heads-up-display (HUD).
Furthermore, we developed a proof-of-concept game called Adventure AR to explain how the Space Syntax data is implemented to determine the placement of spawning game objects. This game is developed in Unity for the Microsoft Holo-Lens. This paper discusses the methodologies to retrieve the Space Syntax data for a given space, and then apply that to dynamically place assets in that region.
Space syntax [1] involves the study of the relationship between the geometric and relational structure of built environments (buildings, cities) and experience of (and behaviour within) the environment. Actively developed over the past decades, Space syntax is not a single method, but a family of graph-theoretic approaches that describe the structure of a space from the perspective of experience [2].
Methods
We are using attributes of Space Syntax such as connectivity, visual complexity, and openness, where connectivity represents how well a region in space is connected to another region [3], visual complexity represents how much area can be seen from one point in space and openness is defined by how open the area is [4]. Values of these attributes are then calculated for a given space and are scaled to the ratio of 255. These individual attribute values are then mapped to an RGBA data structure where blue represents connectivity, green represents visual complexity, and red represents openness (as shown in Figure 1). At the start of the gameplay, the spaces are calibrated by scanning a static object. We use Vuforia, an object recognition API to retrieve coordinates of this static object. The algorithm will then drift the virtual space and map it with the real environment. Finally, the Space Syntax data of any given point in the calibrated space can be retrieved through this RGBA data structure (as shown in Figure 2).
Results
Our research is focused on the fundamental issue facing the development of large-scale mixed reality, immersive AR games. In this paper, so far, we have discussed about the methods we have followed to retrieve the Space Syntax data. But how is this data used to distribute spawning game objects around the gaming area?
A proof-of concept game Adventure AR, based on the classic Atari console game Adventure, is used to test the results of the research that we have carried out. The rules of the game are simple and concise. It requires the player to navigate around the space looking for keys which can open chests that are distributed among the region, while avoiding contact with roaming Artificial Intelligence Monsters. The game is over when the player collides with a monster or when all the chests are opened. Now, the placement of these game assets (Keys, Chests and monsters) will vary according to the spatial settings of the gaming environment. This is the part where the attributes of SS (openness, visual complexity and connectivity) are used to determine the placement of the spawning game objects.
As soon as the game loads and the spaces get calibrated, the RGBA bitmap is loaded. Alongside this, a configuration file is loaded that identifies which room should hold the game objects, and rules for placing the chest and key assets. Once the game ends, the configuration is randomized to generate a new set of location for the game assets keeping the rules of placement constant. For instance, the keys are likely to be placed in areas with low values of openness, visual complexity and connectivity. Similarly, the chests will be placed in spaces with a higher value of these attributes. The game predefines ranges for "low", "medium" and "high" for the spatial attributes. If the location has enough spaces for game object placement, the objects are randomly placed within those spaces. In case, the candidate locations are not enough according to the placement rules, the game won't play, and the rules will need to be changed, or a new location found.
Conclusions
This paper defines a way to derive and encode Space Syntax data to dynamically place game objects in a given space. In the proof-of concept game Adventure AR we have | 1,342.6 | 2019-11-12T00:00:00.000 | [
"Computer Science"
] |
Pneumasis/pneumafication Based on Romans 8:1–17: Highlighting the Spirit’s Role in Deification
In view of the two key themes found in Romans: pneumatology and deification, some pressing questions can be asked. One of these is, what is the role of the Holy Spirit in deification? This essay identifies one area of the work of the Holy Spirit presented in Romans that is often neglected in New Testament (NT) pneumatology, soteriology, and anthropology. This paper argues that, in Romans 8:1–17, the crucial role of the Spirit, as an active person in the triune Godhead, in possessing and being possessed by believers and facilitating the mutual indwelling of Christ and his co-sufferers, is best captured by a new term, namely, pneumasis or pneumafication. In other words, theosis/deification and Christosis/Christification are made possible by pneumasis/pneumafication.
Introduction
The role of the Holy Spirit in Romans has been widely recognized in biblical studies.As Gordon Fee observes, "It is fair to say that Paul's entire theology without the supporting pinion of the Spirit would crumble into ruins" (Fee 1996, p. 7).A parallel development in recent Pauline studies is the increased attention given to the Pauline theology of deification (or theosis) as revealed in Romans. 1 For example, Michael Gorman detects a renewal of interest in participation, 2 with a family of words including theosis, deification, divinization, Christosis, and Christification, within various theological subdisciplines, including biblical studies, theological ethics, spirituality, and others.Furthermore, he observes that "participation has been proposed as an essential aspect . . . of Pauline theology and spirituality in particular" (Gorman 2019, p. xv).Among the many key themes that scholars such as Gorman and M. David Litwa find in Romans are these two: pneumatology and theosis/deification. 3 Consequently, some stimulating questions can be asked: What is the relationship between the Spirit and deification?What is the role of the Holy Spirit in deification?
In his magisterial God's Empowering Presence, Fee laments: "By and large the crucial role of the Spirit in Paul's life and thought-as the dynamic, experienced reality of Christian life-is often either overlooked or given mere lip service" (Fee 1994, p. xxi).For Susan Eastman, the Holy Spirit as God's presence is elusive-at least for Pauline scholars, with only a few exceptions.She observes "the relative paucity of scholarly work on the Spirit" regarding the importance of the Spirit at certain key junctions in Paul's letters (Eastman 2018, p. 103).She suspects that the experiential aspect of Paul's language regarding the Spirit may be partly to blame.Moreover, modern tendencies in the Global North have exacerbated the difficulty of talking about experience due to their inclination to think of "spiritual experience" as individual, private, and esoteric (ibid., pp. 103-4).
This essay identifies one area of the work of the Holy Spirit presented in Romans that is often neglected in NT pneumatology, soteriology, and anthropology. 4Constantine Campbell summarizes the sixteen scholars 5 who have made "significant academic contributions concerning union with Christ through the twentieth century to the present day" (Campbell 2012, p. 59).However, the Holy Spirit is hardly even mentioned in his synthesis of these scholarly works. 6In Pauline scholarship, Grant Macaskill recognizes "the broad awareness of the role of Christ as the focus of union" (Macaskill 2013, p. 41).After reviewing the twelve scholars 7 whose "key contributions . . .rightly or wrongly, have shaped the discussions during the last century or so, including the recent resurgence of interest in participatory accounts of atonement" (Campbell 2012, p. 17), Macaskill concludes that "most scholars have recognized the distinctive Christocentrism of Pauline mysticism" 8 .Bernie A. Van De Walle asserts that "the doctrine of theosis is essentially Christocentric.While theosis occurs through the agency of the Spirit, it is complementary to the work of the Incarnation and results from union with Christ" (Van De Walle 2008, p. 140).Clearly, in the contemporary scholarly work on theosis, the role of the Holy Spirit as a person in the triune Godhead has, by and large, been marginalized. 9 To reverse that trend to some degree, I propose to work with Rom 8.1-17, which is notable for its high incidence of Spirit language. 10I argue that, in 8.1-17, the crucial role that the Spirit plays through possessing believers and being possessed by them, and facilitating the mutual indwelling of Christ and his co-sufferers, is best captured by a new term, namely, pneumasis or pneumafication.In other words, Christosis/Christification, a synonym of theosis/deification, is made possible by pneumasis/pneumafication.If, by theosis, we mean participation in God's divine nature 11 and, by Christosis, conformation to the image of Christ in his death and resurrection, 12 then by pneumasis/pneumafication, we mean being in the Spirit and indwelt by the Spirit, which entails walking according to the Spirit, setting our minds on the Spirit, and being led by the Spirit as children of God.Thus, pneumasis highlights Paul's recognition of the spiritual reality that the Spirit is not merely the means or the power through which we participate in Christ, but is an active person in whose work "the stress [is] upon the immediacy of the divine, and the direct encounter of man with the Holy Spirit" (McGrath 1986, pp. 1:3-4).Theosis, Christosis, pneumasis-or deification, Christification, pneumafication-form a trinitarian doctrine that does justice to Paul's proto-trinitarian thought, which penetrates Rom 8. Through the use of this neologism-pneumasis/pneumafication, this paper hopes to underscore the preeminent role of the Holy Spirit, to match Paul's high and elevated view of the Holy Spirit without undermining the salvific power of Christ.For Paul, the Spirit serves as God's solidarity with Christ's co-sufferers.Due to some problematic readings of Rom 8, much of Western Christianity has been plagued by a type of individualized and arrogant morality that divides the world into the spiritual and the fleshly.Such a dualistic, black-and-white divide results in spiritual pride, which not only is detrimental to Christian spirituality and humanity, but also loses sight of God's solidarity with those suffering with Christ.The significance of this paper lies in its contribution to the NT studies on deification; it brings to light the crucial role of the Holy Spirit not only in soteriology, but also in anthropology, highlighting human participation in the Spirit as Christ's co-sufferers for the sake of the eschatological and cosmological freedom from corruption that creation eagerly awaits.Romans 8.1-17 invites believers into such a pneumatic/pneumaficational reading and transforms them into Christ's co-sufferers, who no longer remain captive to spiritual pride, but genuinely care about the suffering of humanity and creation.
Brief Survey of Recent Exegetical Works on Paul's Pneumatology and Deification in Romans 8:1-17
The sheer volume of Pauline scholarship on Romans makes any attempt at even a brief survey of the history of exegesis a gargantuan task.Even if we narrow it down to the Pauline theme of union with Christ, we have in our hands the scholarly works of more than a century to cover. 13However, the research carried out by Campbell and Macaskill provides excellent limitations for our work.Among the works surveyed, only those of Käsemann, Sanders, Gaffin, Dunn, Wright, Gorman, Campbell, and Macaskill have provided in-depth exegetical work on Rom 8.1-17. 14First, Käsemann initiated a line of scholarship that concentrates on "apocalyptic" readings of Paul (Macaskill 2013, p. 34).Käsemann divides his exegesis of Rom 8.1-17 into two subsections, entitled "the Christian life as being in the Spirit" (vv. 1-11) and "the state of sonship" (vv. 12-17).He is correct to refute the view that one cannot start with the parallels "in the Spirit" and "in the flesh," because the former usually designates inspiration.However, his opinion that "in the Spirit" is interchangeable with "in Christ" betrays his reductionistic view (Käsemann 1980, pp. 212-29).Moreover, his insistence that "in Christ" interprets "in the Spirit", but not vice versa (ibid., pp. 222-23), suggests his subordination of the role of the Spirit.
Second, Sanders refocused attention on "participation" as the more important dimension of Paul's soteriology (Dunn 1998, p. 393).In the context of Rom 8, he observes that "having the Spirit as guarantee and salvation by participation in the Spirit or in Christ (or participation with the Spirit or Christ by having them in one) are not separate themes.. . .Having the Spirit results in (or is) real participation in the Spirit and the resurrected Lord, which participation provides the best guarantee of all: Christians are sons of God" (Sanders 1977, p. 460).Sanders also notices that "the reference in [Phil 3.10] to suffering with Christ is to be connected with other passages in which Paul says that Christians share Christ's sufferings so as to share his life: Rom.8.17" (Ibid., p. 467).On one hand, Campbell rightly applauds Sanders, who "revitalized the concept of participation with Christ, describing Paul's pattern of religion as 'participationist eschatology'" (Campbell 2012, p. 53).Sanders also sees the intrinsic link between suffering with Christ and sharing his life.On the other hand, he does not notice the connection between participation and the mutual indwelling of believers and the Spirit/Christ.
Third, in his work The Centrality of the Resurrection, Gaffin notes that: The assumption expressed in Rom 8.9a ("if the Spirit of God dwells in you") is basic to the reasoning in the sentences immediately following.Essential also is the intimate bond between Christ and the Spirit.The Spirit is "the Spirit of Christ" (v.9b).In the experience of believers, "in the Spirit" (v.9a), "the Spirit in you" (vv. 9a, 11a, c) and "Christ in you" (v.10a) are all used correlatively, and the remaining possibility on this combination, the more usual "in Christ" is certainly present by implication (cf. the apodosis of v. 9b; v. 1).The idea of solidarity, then, has an important place in these verses.(Gaffin 1978, p. 66) Another of Gaffin's contributions is his argument (against Käsemann and Bultmann) that "Paul considers the Spirit a (divine) person in the same sense as the Father and Christ" (Ibid., p. 71).However, in considering Paul's use of "in Christ" and "in the Spirit" as equivalent, Gaffin fails to discern their inherent distinctions.Fourth, Dunn carefully analyzes Paul's theme of participation in Christ (Dunn 1998, pp. 390-412).Commenting on Rom 8.9-10, Dunn argues that "where 'in Spirit,' 'have Spirit,' and 'Christ in you' all serve as complementary identifying descriptions, the dividing line between experience of Spirit and experience of Christ has become impossible to define in clear-cut terms.At best we may speak of Christ as the context and the Spirit as the power" (Ibid.,p. 408).Realizing the mutual indwelling between the Spirit and believers, Dunn concludes that "the Spirit is the medium of Christ's union with his own" (Ibid., p. 264).However, by merely viewing the Spirit as the "power" and "medium," Dunn fails to meet his own classification of Rom 8.1-27 as "the high point of Paul's theology of the Spirit" (Ibid.,p. 423).
Fifth, in his classic work on Paul, The Climax of the Covenant, Wright captures the covenantal dimension of Paul's thought with a narrative substructure shaped by the story of Israel, which leads to the new reality of Christ.Wright studies prepositional phrases that include Xριστóς (Wright 1991, pp. 44-49).He lists as a red herring, though entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid.,p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid.,p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid.,p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8. [9][10][11]p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid.,p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid.,p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid.,p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid.,p. 448,italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid.,p. 240). However,in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
ν Xριστ entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8. [9][10][11]p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid.,p. 240). However,in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
, ε entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8. [9][10][11]p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰ Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid.,p. 448,italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid.,p. 240). However,in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
ς Xριστóν, σ entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8. [9][10][11]p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid.,p. 448,italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid.,p. 240). However,in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
ν Xριστ 4 of 13 be mistaken to suggest that Paul's language about ss the same meaning as "in Christ".He further notices er to Paul's language about the Spirit, as is clear from nately, due to his emphasis on the covenant and the not tackle the crucial issues related to the terms "in xegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Theological and Pastoral Commentary, Gorman notices the Spirit within the framework of a trinitarian unders at least five key elements of Christian life from Rom 8, .1-17: first, the mutual indwelling of Christ and the lievers' adoption as God's children; and third, the followed by glory (Gorman 2022, p. 192).Gorman afte aspects of participation in and with Christ (Ibid., p. e identification of the Holy Spirit as "the Spirit of crution in Christ" in Rom 8 (Ibid., p. 41), with the role of body (8.13) and empowering believers to suffer with 7) (Ibid.,p. 193).Particularly germane to this paper is tual indwelling of Christ and the faithful "takes place that "Paul can use the language of mutual indwelling irit in the same breath" (Ibid., p. 40).In interpreting (8.13), Gorman clarifies that, for Paul, "This is not a iting glory but a claim about the nature of full particist's story is a narrative of suffering before full and final , of being humbled before being exalted" (Ibid.,p. 202).e argument presented by this paper, but still falls short role of the Spirit in the mutual indwelling of Christ and aul's concept of union with Christ primarily in his exses associated with Christ, for example, ἐν Χριστῷ, εἰς ῦ, σύν-compounds, and their variations.He concludes ntial ingredient that binds all other elements together; e ideas of Paul's web-shaped theological framework" ll opines that "in the life of the believer, the Spirit beunion with Christ is lived out".He suggests that "sufatio Christi and not imitatio Christi only.Believers share eath and the power of his resurrection, and one conseill undergo suffering" (Ibid., p. 448, italics original).ical work on union with Christ, Campbell's recognition means of the work of Christ downplays the personal marginalizes her role.15 His understanding would have the same exegetical rigor to the study of the Spirit.a descriptive task on participation in the New Testaology and, to a lesser extent, systematic theology.In rgues that the covenantal framework must serve as the articipation or union with Christ (Macaskill 2013, pp.ovenant is the covenant of the Spirit.The Spirit "is the t, who conforms our being to its terms by writing those our conformity to Christ" (Ibid., p. 300).Commenting es the distinctive partnership of the Holy Spirit and the this treatment of Romans (Ibid.,, he fails to namely, paying attention to "the distinctive place of n particular, he misses the different roles of "in Christ" heme of participation. , δι entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid.,p. 448,italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
Xριστo entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8. [9][10][11]p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid.,p. 448,italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
, σ entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8. [9][10][11]p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid.,p. 448,italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
ν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
In addition, I propose reviewing Eastman's work for its relevance to this research.In her recent study of Rom 8, Eastman notices the Spirit's role in mediating the experience of union with Christ.Eastman argues that the Spirit generates and sustains a mutually participatory bond of love between believers and God, as well as between people "in Christ."For Eastman, "The central motif therefore is the indwelling of Christ through the Spirit" (Eastman 2018, p. 111).In her consideration of the entire chapter, she is insightful in realizing the Spirit's role in generating a community bonded with love.However, the motif of love is only visible in the second half (vv. 28, 35, 39), suggesting that her exegetical "center of gravity" leans more toward vv.18-39.Therefore, it is necessary to conduct a recalibrated analysis of the first half of the chapter, without losing sight of Eastman's contribution.
So far, I have reviewed scholars surveyed by Campbell and Macaskill and the works of Fee and Eastman.These studies show a trend toward an increased appreciation for Paul's motif of theosis.However, comparatively speaking, what is lacking is the due recognition of an elevated role of the Spirit in theosis.Moreover, there has been a lack of appreciation of the Spirit's work in (trans)forming believers into Christ's co-sufferers so that eschatological freedom can be enjoyed by the suffering creation.With that in mind, I proceed to my analysis of Rom 8.1-17.
Exegesis of Romans 8:1-17
Romans 8.1-17 is situated in a chapter that "contains one of the thickest clusters of Spirit language in Paul's letters" (Ibid., p. 103).Commenting on this chapter, Dunn vividly portrays the highly elevated role of the Spirit: "when the reader reaches Rom. 8, not least after the agonizing testimony of 7.7-25, it is almost as though a pent-up flood has been released, and out pour Paul's convictions about the decisive role of the Spirit in determining and shaping the believer's life.Rom.8.1-27 is unquestionably the high point of Paul's theology of the Spirit" (Dunn 1998, p. 423).Moreover, 8.1-17 is populated with the densest references to the Spirit.The seventeen instances of πνε "Christ in me/you" has more or less the same meaning as "in Christ".He furth that "Christ in you" is much closer to Paul's language about the Spirit, as is c Rom 8. [9][10][11]p. 45).Unfortunately, due to his emphasis on the covenan vindication of the law, he does not tackle the crucial issues related to the Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorma that Rom 8 focuses on the life in the Spirit within the framework of a trinitaria standing of salvation.He identifies at least five key elements of Christian life fro three of which originate from vv. 1-17: first, the mutual indwelling of Chris Spirit with believers; second, believers' adoption as God's children; and Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).G firms the individual and corporate aspects of participation in and with Christ 41).His unique contribution is the identification of the Holy Spirit as "the Spi ciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with t putting to death the deeds of the body (8.13) and empowering believers to su Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to thi Gorman's affirmation that the mutual indwelling of Christ and the faithful "ta by the means of the Spirit," and that "Paul can use the language of mutual in with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In int Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "Thi statement about suffering as meriting glory but a claim about the nature of fu pation in the messianic story.Christ's story is a narrative of suffering before full glory, of death before resurrection, of being humbled before being exalted" (Ibid Gorman's insight comes close to the argument presented by this paper, but still in understanding the preeminent role of the Spirit in the mutual indwelling of C his co-sufferers. Seventh, Campbell studies Paul's concept of union with Christ primarily egesis of a few prepositional phrases associated with Christ, for example, ἐν Χρ Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He c that union with Christ is the "essential ingredient that binds all other elements it is the webbing that connects the ideas of Paul's web-shaped theological fra (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the comes the means through whom union with Christ is lived out".He suggests fering is to be viewed as a participatio Christi and not imitatio Christi only.Believ in the ongoing force of Christ's death and the power of his resurrection, and o quence of this is that believers will undergo suffering" (Ibid., p. 448, italics Despite his comprehensive exegetical work on union with Christ, Campbell's re of the Spirit's role as merely the means of the work of Christ downplays the nature of the Spirit and, therefore, marginalizes her role. 15His understanding wo been much thicker had he applied the same exegetical rigor to the study of the S Finally, Macaskill performs a descriptive task on participation in the N ment, informed by historical theology and, to a lesser extent, systematic the synthesizing his conclusions, he argues that the covenantal framework must se starting point for reflection on participation or union with Christ .For Macaskill, the new covenant is the covenant of the Spirit.The Spi gift given within the new covenant, who conforms our being to its terms by wri terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Com on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spir human (Ibid., p. 240).However, in this treatment of Romans (Ibid., pp.237-44), execute his own prescribed task, namely, paying attention to "the distinctive Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of " and "in the Spirit" in the Pauline theme of participation.
µα mentioned in 8.1-17 occupy half of all the references in the epistle, dwarfing the remaining chapter and all other parts of the epistle.In comparison, θεóς appears nine times, and Xριστóς six times.Ironically, the passage has been chiefly read christologically, in terms of justification. 16
Of all these references to πνε
Religions 2023, 14, x FOR PEER REVIEW 4 of 13 entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role.15 His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
ν πνε
Religions 2023, 14, x FOR PEER REVIEW entirely correctly, that we would be mistaken to sug "Christ in me/you" has more or less the same meaning that "Christ in you" is much closer to Paul's language Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his e vindication of the law, he does not tackle the crucia Christ" and "in the Spirit" in his exegesis of Rom 8.1-1 Sixth, in his work Romans: A Theological and Past that Rom 8 focuses on the life in the Spirit within the standing of salvation.He identifies at least five key elem three of which originate from vv. 1-17: first, the mut Spirit with believers; second, believers' adoption as Christ-shaped pattern of suffering followed by glory ( firms the individual and corporate aspects of particip 41).His unique contribution is the identification of the ciformity-cross-shaped participation in Christ" in Ro putting to death the deeds of the body (8.13) and em Christ as the prelude to glory (8.17) (Ibid.,p. 193).Par Gorman's affirmation that the mutual indwelling of C by the means of the Spirit," and that "Paul can use th with respect to Christ and the Spirit in the same bre Christians' suffering with Christ (8.13), Gorman clari statement about suffering as meriting glory but a claim pation in the messianic story.Christ's story is a narrativ glory, of death before resurrection, of being humbled b Gorman's insight comes close to the argument presente in understanding the preeminent role of the Spirit in th his co-sufferers.
Seventh, Campbell studies Paul's concept of unio egesis of a few prepositional phrases associated with C Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, a that union with Christ is the "essential ingredient that it is the webbing that connects the ideas of Paul's we (Campbell 2012, p. 442).Campbell opines that "in the comes the means through whom union with Christ is fering is to be viewed as a participatio Christi and not im in the ongoing force of Christ's death and the power o quence of this is that believers will undergo sufferin Despite his comprehensive exegetical work on union w of the Spirit's role as merely the means of the work o nature of the Spirit and, therefore, marginalizes her role been much thicker had he applied the same exegetical r Finally, Macaskill performs a descriptive task on ment, informed by historical theology and, to a lesse synthesizing his conclusions, he argues that the covena starting point for reflection on participation or union 297-98).For Macaskill, the new covenant is the covena gift given within the new covenant, who conforms our terms on our hearts and realizing our conformity to C on Rom 8.14-17, Macaskill observes the distinctive part human (Ibid., p. 240).However, in this treatment of Rom execute his own prescribed task, namely, paying atte Jesus and the Spirit" (Ibid., 145).In particular, he misse and "in the Spirit" in the Pauline theme of participation µατι in v. 9, through which Paul speaks of the believers' new sphere.Namely, they are no longer in the flesh but in the Spirit.Functionally similar to R PEER REVIEW 4 of 13 entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role.15 His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
ν πνε
Religions 2023, 14, x FOR PEER REVIEW entirely correctly, that we would be mistaken to suggest that Paul's langua "Christ in me/you" has more or less the same meaning as "in Christ".He furthe that "Christ in you" is much closer to Paul's language about the Spirit, as is c Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenan vindication of the law, he does not tackle the crucial issues related to the t Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorma that Rom 8 focuses on the life in the Spirit within the framework of a trinitaria standing of salvation.He identifies at least five key elements of Christian life fro three of which originate from vv. 1-17: first, the mutual indwelling of Chris Spirit with believers; second, believers' adoption as God's children; and t Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Go firms the individual and corporate aspects of participation in and with Christ 41).His unique contribution is the identification of the Holy Spirit as "the Spir ciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with th putting to death the deeds of the body (8.13) and empowering believers to su Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this Gorman's affirmation that the mutual indwelling of Christ and the faithful "ta by the means of the Spirit," and that "Paul can use the language of mutual in with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In int Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This statement about suffering as meriting glory but a claim about the nature of fu pation in the messianic story.Christ's story is a narrative of suffering before full glory, of death before resurrection, of being humbled before being exalted" (Ibid Gorman's insight comes close to the argument presented by this paper, but still f in understanding the preeminent role of the Spirit in the mutual indwelling of C his co-sufferers. Seventh, Campbell studies Paul's concept of union with Christ primarily egesis of a few prepositional phrases associated with Christ, for example, ἐν Χρ Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He c that union with Christ is the "essential ingredient that binds all other elements it is the webbing that connects the ideas of Paul's web-shaped theological fra (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the comes the means through whom union with Christ is lived out".He suggests fering is to be viewed as a participatio Christi and not imitatio Christi only.Believ in the ongoing force of Christ's death and the power of his resurrection, and o quence of this is that believers will undergo suffering" (Ibid., p. 448, italics Despite his comprehensive exegetical work on union with Christ, Campbell's re of the Spirit's role as merely the means of the work of Christ downplays the nature of the Spirit and, therefore, marginalizes her role.15 His understanding wo been much thicker had he applied the same exegetical rigor to the study of the S Finally, Macaskill performs a descriptive task on participation in the Ne ment, informed by historical theology and, to a lesser extent, systematic the synthesizing his conclusions, he argues that the covenantal framework must ser starting point for reflection on participation or union with Christ .For Macaskill, the new covenant is the covenant of the Spirit.The Spi gift given within the new covenant, who conforms our being to its terms by writ terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Com on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spir human (Ibid., p. 240).However, in this treatment of Romans (Ibid., pp.237-44), h execute his own prescribed task, namely, paying attention to "the distinctive Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "i and "in the Spirit" in the Pauline theme of participation.
µατι is πνε
Religions 2023, 14, x FOR PEER REVIEW entirely correctly, that we would be mistaken to suggest that P "Christ in me/you" has more or less the same meaning as "in Chri that "Christ in you" is much closer to Paul's language about the Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on vindication of the law, he does not tackle the crucial issues rel Christ" and "in the Spirit" in his exegesis of Rom 8.1-11 (Ibid., pp. Sixth, in his work Romans: A Theological and Pastoral Commen that Rom 8 focuses on the life in the Spirit within the framework standing of salvation.He identifies at least five key elements of Chr three of which originate from vv. 1-17: first, the mutual indwel Spirit with believers; second, believers' adoption as God's chi Christ-shaped pattern of suffering followed by glory (Gorman 202 firms the individual and corporate aspects of participation in and 41).His unique contribution is the identification of the Holy Spiri ciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p putting to death the deeds of the body (8.13) and empowering b Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly ge Gorman's affirmation that the mutual indwelling of Christ and th by the means of the Spirit," and that "Paul can use the language with respect to Christ and the Spirit in the same breath" (Ibid., Christians' suffering with Christ (8.13), Gorman clarifies that, fo statement about suffering as meriting glory but a claim about the pation in the messianic story.Christ's story is a narrative of sufferin glory, of death before resurrection, of being humbled before being Gorman's insight comes close to the argument presented by this pa in understanding the preeminent role of the Spirit in the mutual in his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Chr egesis of a few prepositional phrases associated with Christ, for ex Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their va that union with Christ is the "essential ingredient that binds all ot it is the webbing that connects the ideas of Paul's web-shaped th (Campbell 2012, p. 442).Campbell opines that "in the life of the comes the means through whom union with Christ is lived out".fering is to be viewed as a participatio Christi and not imitatio Chris in the ongoing force of Christ's death and the power of his resurr quence of this is that believers will undergo suffering" (Ibid., p Despite his comprehensive exegetical work on union with Christ, C of the Spirit's role as merely the means of the work of Christ do nature of the Spirit and, therefore, marginalizes her role.15 His unde been much thicker had he applied the same exegetical rigor to the Finally, Macaskill performs a descriptive task on participati ment, informed by historical theology and, to a lesser extent, sy synthesizing his conclusions, he argues that the covenantal framew starting point for reflection on participation or union with Chris 297-98).For Macaskill, the new covenant is the covenant of the Sp gift given within the new covenant, who conforms our being to its terms on our hearts and realizing our conformity to Christ" (Ibid.on Rom 8.14-17, Macaskill observes the distinctive partnership of t human (Ibid., p. 240).However, in this treatment of Romans (Ibid., execute his own prescribed task, namely, paying attention to "th Jesus and the Spirit" (Ibid., 145).In particular, he misses the differe and "in the Spirit" in the Pauline theme of participation.
µατι, used without the preceding preposition.Paul exhorts the Roman believers to put to death the deeds of the body in/by the Spirit (v.13) and to be led by/in the Spirit (v.14).At the other end of the spectrum of the relationship between the Spirit and believers is the fact that "the Spirit of God dwells in you" (9, 11 [2x]).Therefore, the concept of mutual indwelling between the Spirit and the believers is one of the hidden "jewels" in 8.1-17, recognized by Gorman as "the center of Paul's spirituality of participation and transformation" (Gorman 2019, p. 16).
The phrase 4 of 13 hat we would be mistaken to suggest that Paul's language about has more or less the same meaning as "in Christ".He further notices is much closer to Paul's language about the Spirit, as is clear from .45).Unfortunately, due to his emphasis on the covenant and the aw, he does not tackle the crucial issues related to the terms "in pirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].ork Romans: A Theological and Pastoral Commentary, Gorman notices on the life in the Spirit within the framework of a trinitarian under-.He identifies at least five key elements of Christian life from Rom 8, inate from vv. 1-17: first, the mutual indwelling of Christ and the s; second, believers' adoption as God's children; and third, the rn of suffering followed by glory (Gorman 2022, p. 192).Gorman afl and corporate aspects of participation in and with Christ (Ibid., p. tribution is the identification of the Holy Spirit as "the Spirit of cruped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of deeds of the body (8.13) and empowering believers to suffer with e to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is n that the mutual indwelling of Christ and the faithful "takes place Spirit," and that "Paul can use the language of mutual indwelling ist and the Spirit in the same breath" (Ibid., p. 40).In interpreting with Christ (8.13), Gorman clarifies that, for Paul, "This is not a fering as meriting glory but a claim about the nature of full particinic story.Christ's story is a narrative of suffering before full and final e resurrection, of being humbled before being exalted" (Ibid., p. 202).mes close to the argument presented by this paper, but still falls short e preeminent role of the Spirit in the mutual indwelling of Christ and bell studies Paul's concept of union with Christ primarily in his exositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς ῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes ist is the "essential ingredient that binds all other elements together; at connects the ideas of Paul's web-shaped theological framework" 442).Campbell opines that "in the life of the believer, the Spirit berough whom union with Christ is lived out".He suggests that "sufd as a participatio Christi and not imitatio Christi only.Believers share of Christ's death and the power of his resurrection, and one conseat believers will undergo suffering" (Ibid., p. 448, italics original).ensive exegetical work on union with Christ, Campbell's recognition s merely the means of the work of Christ downplays the personal nd, therefore, marginalizes her role.15 His understanding would have ad he applied the same exegetical rigor to the study of the Spirit.ill performs a descriptive task on participation in the New Testahistorical theology and, to a lesser extent, systematic theology.In clusions, he argues that the covenantal framework must serve as the flection on participation or union with Christ (Macaskill 2013, pp.ill, the new covenant is the covenant of the Spirit.The Spirit "is the new covenant, who conforms our being to its terms by writing those and realizing our conformity to Christ" (Ibid., p. 300).Commenting caskill observes the distinctive partnership of the Holy Spirit and the ).However, in this treatment of Romans (Ibid.,, he fails to entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role.15 His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid., pp.237-44), he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of µατι appears eleven times in the undisputed Pauline epistles (Rom 2.29; 8.9; 9.1; 14.17; 15.16; 1 Cor 12.3; 14.16; 2 Cor 6.6; Gal 6.1; 1 Thess 1.5), of which at least four (Rom 8.9; 17 1 Cor 12.3; 14.16; Gal 6.1) refer to the believers' new relationship with the Holy Spirit. 18The semantic analysis of prepositions is a complex enterprise (Campbell 2012, p. 74).Without question, the preposition OR PEER REVIEW 4 of 13 entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the ν is by far the most commonly used preposition in the NT (Wallace 1996, pp. 357, 372); it is also the most significant and the most perplexing of the relevant prepositions (Campbell 2012, p. 75).According to Albrecht Oepke, "The spatial sense is always the starting-point, but we have to ask how far there is an intermingling of other sense, esp. the instrumental" (Oepke 1964, p. 2:538).He regards the use of 4 of 13 would be mistaken to suggest that Paul's language about re or less the same meaning as "in Christ".He further notices ch closer to Paul's language about the Spirit, as is clear from nfortunately, due to his emphasis on the covenant and the does not tackle the crucial issues related to the terms "in in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
mans: A Theological and Pastoral Commentary, Gorman notices life in the Spirit within the framework of a trinitarian underentifies at least five key elements of Christian life from Rom 8, om vv.1-17: first, the mutual indwelling of Christ and the nd, believers' adoption as God's children; and third, the ffering followed by glory (Gorman 2022, p. 192).Gorman aforporate aspects of participation in and with Christ (Ibid., p. n is the identification of the Holy Spirit as "the Spirit of crurticipation in Christ" in Rom 8 (Ibid., p. 41), with the role of of the body (8.13) and empowering believers to suffer with ory (8.17) (Ibid.,p. 193).Particularly germane to this paper is the mutual indwelling of Christ and the faithful "takes place ," and that "Paul can use the language of mutual indwelling the Spirit in the same breath" (Ibid., p. 40).In interpreting Christ (8.13), Gorman clarifies that, for Paul, "This is not a as meriting glory but a claim about the nature of full particiy.Christ's story is a narrative of suffering before full and final rection, of being humbled before being exalted" (Ibid., p. 202).se to the argument presented by this paper, but still falls short inent role of the Spirit in the mutual indwelling of Christ and dies Paul's concept of union with Christ primarily in his exal phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστοῦ, σύν-compounds, and their variations.He concludes e "essential ingredient that binds all other elements together; ects the ideas of Paul's web-shaped theological framework" ampbell opines that "in the life of the believer, the Spirit bewhom union with Christ is lived out".He suggests that "sufparticipatio Christi and not imitatio Christi only.Believers share rist's death and the power of his resurrection, and one conseevers will undergo suffering" (Ibid., p. 448, italics original).exegetical work on union with Christ, Campbell's recognition ly the means of the work of Christ downplays the personal refore, marginalizes her role. 15His understanding would have pplied the same exegetical rigor to the study of the Spirit.entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role.15 His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testa- entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal ν as "a marker of close personal association-'in, one with, in union with, joined closely to'" (Louw and Nida 1989, p. 1:793).Therefore, "in the Spirit" denotes believers' intimate relationship with the Spirit, in which "all believers individually are constantly enveloped and possessed by Christ's Spirit, like the air around and within them" (Gorman 2022, p. 199).Such a close, personal association between the Spirit and believers is also similarly expressed by the standalone word πνε statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid., pp.237-44), he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
µατι, without the preceding Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid., pp.237-44), he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.ν, such as in "I serve God in/with my spirit" (1.9; cf.9.1; 12.11; 15.16).The same is true in other undisputed Pauline epistles: "I come with love in the Spirit of gentleness" (1 Cor 4.21;cf. 5.3;7.34;14.2,15 [2x]; 2 Cor 2. 13;12.18;Gal 3.3;5.5,16,18,25 [2x]; Phil 3.3).
Paul's use of e to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is on that the mutual indwelling of Christ and the faithful "takes place e Spirit," and that "Paul can use the language of mutual indwelling rist and the Spirit in the same breath" (Ibid., p. 40).In interpreting g with Christ (8.13), Gorman clarifies that, for Paul, "This is not a ffering as meriting glory but a claim about the nature of full partici-nic story.Christ's story is a narrative of suffering before full and final re resurrection, of being humbled before being exalted" (Ibid.,p. 202).mes close to the argument presented by this paper, but still falls short e preeminent role of the Spirit in the mutual indwelling of Christ and bell studies Paul's concept of union with Christ primarily in his ex-ositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς τῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes rist is the "essential ingredient that binds all other elements together; at connects the ideas of Paul's web-shaped theological framework" 442).Campbell opines that "in the life of the believer, the Spirit be-rough whom union with Christ is lived out".He suggests that "suf-ed as a participatio Christi and not imitatio Christi only.Believers share e of Christ's death and the power of his resurrection, and one conse-at believers will undergo suffering" (Ibid., p. 448, italics original).hensive exegetical work on union with Christ, Campbell's recognition as merely the means of the work of Christ downplays the personal and, therefore, marginalizes her role. 15His understanding would have had he applied the same exegetical rigor to the study of the Spirit.kill performs a descriptive task on participation in the New Testa-historical theology and, to a lesser extent, systematic theology.In nclusions, he argues that the covenantal framework must serve as the eflection on participation or union with Christ (Macaskill 2013, pp.kill, the new covenant is the covenant of the Spirit.The Spirit "is the e new covenant, who conforms our being to its terms by writing those s and realizing our conformity to Christ" (Ibid., p. 300).Commenting acaskill observes the distinctive partnership of the Holy Spirit and the ).However, in this treatment of Romans (Ibid.,, he fails to rescribed task, namely, paying attention to "the distinctive place of " (Ibid., 145).In particular, he misses the different roles of "in Christ" in the Pauline theme of participation.
ν πνε Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid.,p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).
Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role.15 His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.
µατι in Rom 8.9 differs from that in other non-disputed Pauline epistles.For example, in 1 Cor 12.3, 14.16, and Gal 6.1, Paul focuses on the believers' new lifestyle in speaking and behaving, whereas in Rom 8.9, Paul's attention is on the believers' new identity, demonstrated through his sharp antithesis between "in the flesh" and "in the Spirit" (8.8, 9).Now, let us turn to the other side of the coin by looking at how Paul describes the Spirit's dwelling in believers.In 8.9, Paul speaks of the Spirit of God dwelling in his believers.The word o m vv.1-17: first, the mutual indwelling of Christ and the d, believers' adoption as God's children; and third, the fering followed by glory (Gorman 2022, p. 192).Gorman afporate aspects of participation in and with Christ (Ibid., p. is the identification of the Holy Spirit as "the Spirit of cruticipation in Christ" in Rom 8 (Ibid., p. 41), with the role of f the body (8.13) and empowering believers to suffer with y (8.17) (Ibid.,p. 193).Particularly germane to this paper is e mutual indwelling of Christ and the faithful "takes place and that "Paul can use the language of mutual indwelling he Spirit in the same breath" (Ibid., p. 40).In interpreting hrist (8.13), Gorman clarifies that, for Paul, "This is not a meriting glory but a claim about the nature of full partici-.Christ's story is a narrative of suffering before full and final ction, of being humbled before being exalted" (Ibid.,p. 202).e to the argument presented by this paper, but still falls short nent role of the Spirit in the mutual indwelling of Christ and ies Paul's concept of union with Christ primarily in his exl phrases associated with Christ, for example, ἐν Χριστῷ, εἰ ριστοῦ, σύν-compounds, and their variations.He concludes "essential ingredient that binds all other elements together; cts the ideas of Paul's web-shaped theological framework" pbell opines that "in the life of the believer, the Spirit behom union with Christ is lived out".He suggests that "sufrticipatio Christi and not imitatio Christi only.Believers share st's death and the power of his resurrection, and one conseers will undergo suffering" (Ibid.,p. 448,italics original).xegetical work on union with Christ, Campbell's recognition the means of the work of Christ downplays the personal fore, marginalizes her role. 15His understanding would have plied the same exegetical rigor to the study of the Spirit.rms a descriptive task on participation in the New Testal theology and, to a lesser extent, systematic theology.In , he argues that the covenantal framework must serve as the on participation or union with Christ (Macaskill 2013, pp.ew covenant is the covenant of the Spirit.The Spirit "is the enant, who conforms our being to its terms by writing those lizing our conformity to Christ" (Ibid., p. 300).Commenting bserves the distinctive partnership of the Holy Spirit and the er, in this treatment of Romans (Ibid.,, he fails to task, namely, paying attention to "the distinctive place of 45).In particular, he misses the different roles of "in Christ" line theme of participation.
κέω appears four times in Romans (7.18,20;8.9,11),portraying a stark contrast between the believers' old identity and their new one.Their old identity is characterized by statements such as "nothing good dwells in me" (7.18) and "sin dwells within me" (7.20), whereas with their new status, the indwelling sin is replaced by the indwelling Spirit (8.9, 11).In 1 Cor 3.16 and 6.19, Paul uses the vivid metaphor of believers as the temple of God indwelt by the Spirit.In Rom 8.11, Paul employs another verb, tely, due to his emphasis on the covenant and the tackle the crucial issues related to the terms "in esis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
heological and Pastoral Commentary, Gorman notices Spirit within the framework of a trinitarian under-least five key elements of Christian life from Rom 8, -17: first, the mutual indwelling of Christ and the vers' adoption as God's children; and third, the llowed by glory (Gorman 2022, p. 192) 15 His understanding would have same exegetical rigor to the study of the Spirit.escriptive task on participation in the New Testa-gy and, to a lesser extent, systematic theology.In es that the covenantal framework must serve as the cipation or union with Christ (Macaskill 2013, pp.nant is the covenant of the Spirit.The Spirit "is the ho conforms our being to its terms by writing those r conformity to Christ" (Ibid., p. 300).Commenting he distinctive partnership of the Holy Spirit and the is treatment of Romans (Ibid.,, he fails to mely, paying attention to "the distinctive place of articular, he misses the different roles of "in Christ" e of participation. νoικέω, to express the similar notion of the Spirit's dwelling in the believers.This verb also appears in 2 Cor 6.16, in which God is said to indwell his people who are the temple of the living God. 19 There are various allusions to the indwelling Spirit in Israel's Scriptures.Paul could have in mind something akin to God's promise in Exod 29.45-46 to dwell among the people of Israel (Jewett 2007, p. 490).Both Ezek 36.26-27 and Jer 31.33 contain prophetic promises of the indwelling divine Spirit, which can now be experienced by those in the Spirit.In Ezek 43.5, the prophet was lifted by the Spirit and then brought into the inner court, where he saw the glory of the Lord fill the temple.If it is true that Paul borrows this imagery from Ezekiel, then he reappropriates both the temple as God's justified people, and the glory of the Lord as the Holy Spirit.For Paul, however, the glory is eschatological and cosmological (Rom 8.18,21), whereas the Spirit is presently "set out as the Spirit of cruciformity-cross-shaped participation in Christ" (Gorman 2022, p. 192).
In summary, Paul speaks of the mutual indwelling of the Holy Spirit and believers as their new identity, contrary to their old identity, which is marked by their being in the flesh and indwelt by sin.
Mutual Indwelling of Christ and the Believers
Paul begins Rom 8 by announcing the good news to a particular group of people, namely those who receive no more condemnation as they are now "in Christ Jesus" (v. 1).The reciprocal relationship between Christ and believers is expressed in v. 10: "Christ is in you."Hence, similar to the mutual indwelling of the Spirit and the believers, Paul speaks of an intimate relationship between Christ and the faithful.
There are thirteen occurrences of the precise phrase , 14, x FOR PEER REVIEW 4 of 13 entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian under-standing of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman af-firms the individual and corporate aspects of participation in and with Christ (Ibid.,p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cru-ciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full partici-pation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his ex-egesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit be-comes the means through whom union with Christ is lived out".He suggests that "suf-fering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one conse-quence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testa-ment, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the ν Xριστ 3, 14, x FOR PEER REVIEW 4 of 13 entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the in Romans (Rom 3.24;6.11,23;8.1,2,39;9.1;12.5;15.17;16.3,7,9,10),of which nine (6.11,8.1;9.1;12.5;15.17;16.3,7,9,10)refer to the believers' new relationship with Christ.Their new identity is understood as being alive to God in Christ Jesus (6.11), many members being "as one body in Christ" (12.5), and so on (cf. 16.7, 9, 10).This new lifestyle is demonstrated through their "speaking the truth in Christ" (9.1; cf.15.17; 16.3).Thirty-nine occurrences of the phrase 4 of 13 e mistaken to suggest that Paul's language about the same meaning as "in Christ".He further notices o Paul's language about the Spirit, as is clear from tely, due to his emphasis on the covenant and the tackle the crucial issues related to the terms "in esis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].
heological and Pastoral Commentary, Gorman notices Spirit within the framework of a trinitarian under-least five key elements of Christian life from Rom 8, -17: first, the mutual indwelling of Christ and the vers' adoption as God's children; and third, the llowed by glory (Gorman 2022, p. 192).Gorman af-spects of participation in and with Christ (Ibid., p. entification of the Holy Spirit as "the Spirit of cru-n in Christ" in Rom 8 (Ibid., p. 41), with the role of dy (8.13) and empowering believers to suffer with (Ibid.,p. 193).Particularly germane to this paper is l indwelling of Christ and the faithful "takes place t "Paul can use the language of mutual indwelling t in the same breath" (Ibid., p. 40).In interpreting 13), Gorman clarifies that, for Paul, "This is not a g glory but a claim about the nature of full partici-story is a narrative of suffering before full and final being humbled before being exalted" (Ibid., p. 202).rgument presented by this paper, but still falls short of the Spirit in the mutual indwelling of Christ and 's concept of union with Christ primarily in his ex-associated with Christ, for example, ἐν Χριστῷ, εἰς ύν-compounds, and their variations.He concludes al ingredient that binds all other elements together; deas of Paul's web-shaped theological framework" pines that "in the life of the believer, the Spirit be-on with Christ is lived out".He suggests that "suf- e mistaken to suggest that Paul's language about the same meaning as "in Christ".He further notices to Paul's language about the Spirit, as is clear from tely, due to his emphasis on the covenant and the t tackle the crucial issues related to the terms "in gesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].heological and Pastoral Commentary, Gorman notices Spirit within the framework of a trinitarian undert least five key elements of Christian life from Rom 8, -17: first, the mutual indwelling of Christ and the vers' adoption as God's children; and third, the llowed by glory (Gorman 2022, p. 192).Gorman afaspects of participation in and with Christ (Ibid., p. dentification of the Holy Spirit as "the Spirit of crun in Christ" in Rom 8 (Ibid., p. 41), with the role of dy (8.13) and empowering believers to suffer with (Ibid.,p. 193).Particularly germane to this paper is al indwelling of Christ and the faithful "takes place t "Paul can use the language of mutual indwelling it in the same breath" (Ibid., p. 40).In interpreting .13),Gorman clarifies that, for Paul, "This is not a g glory but a claim about the nature of full particis story is a narrative of suffering before full and final f being humbled before being exalted" (Ibid., p. 202).argument presented by this paper, but still falls short e of the Spirit in the mutual indwelling of Christ and l's concept of union with Christ primarily in his exassociated with Christ, for example, ἐν Χριστῷ, εἰς σύν-compounds, and their variations.He concludes ial ingredient that binds all other elements together; ideas of Paul's web-shaped theological framework" pines that "in the life of the believer, the Spirit beion with Christ is lived out".He suggests that "sufo Christi and not imitatio Christi only.Believers share h and the power of his resurrection, and one consel undergo suffering" (Ibid., p. 448, italics original).l work on union with Christ, Campbell's recognition eans of the work of Christ downplays the personal arginalizes her role.15 His understanding would have are recorded in the other undisputed Pauline epistles, among which nineteen refer to the believers' new identity in Christ (1 Cor 1.2,4,30;3.1;4.10,15a;2 Cor 5.17;12.2;Gal 1.22;2.4,17;3.26,28;Phil 1.1,4.21;1 Thess 2.14;4.16;Phlm 8,23), and twelve to the new lifestyle in Christ, personally and corporately (1 Cor 4.15b,17;15.18,19,31;2 Cor 2.17;12.19;Phil 1.13,26;3.3,14;Phlm 20).
Gorman helpfully identifies in Rom 8 two groups of phrases and words: the "in" group dominating the first half of the chapter, and the "with" group in the second half, such as "bearing witness with" (v.16), "joint heirs," "suffer with," "be glorified with" (v.17), "groans and suffers together" (v.22), and so on (vv. 26, 28, 29).In addition, there is the phrase "with him [Christ]" (σ with respect to Christ and the Spirit in the same breath" (Ibid., p. 40 Christians' suffering with Christ (8.13), Gorman clarifies that, for Pau statement about suffering as meriting glory but a claim about the natu pation in the messianic story.Christ's story is a narrative of suffering be glory, of death before resurrection, of being humbled before being exalte Gorman's insight comes close to the argument presented by this paper, b in understanding the preeminent role of the Spirit in the mutual indwel his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ pr egesis of a few prepositional phrases associated with Christ, for examp Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variatio that union with Christ is the "essential ingredient that binds all other e it is the webbing that connects the ideas of Paul's web-shaped theolog (Campbell 2012, p. 442).Campbell opines that "in the life of the believ comes the means through whom union with Christ is lived out".He su fering is to be viewed as a participatio Christi and not imitatio Christi onl in the ongoing force of Christ's death and the power of his resurrection quence of this is that believers will undergo suffering" (Ibid., p. 448 Despite his comprehensive exegetical work on union with Christ, Camp of the Spirit's role as merely the means of the work of Christ downp nature of the Spirit and, therefore, marginalizes her role. 15His understan been much thicker had he applied the same exegetical rigor to the study Finally, Macaskill performs a descriptive task on participation in ment, informed by historical theology and, to a lesser extent, system synthesizing his conclusions, he argues that the covenantal framework starting point for reflection on participation or union with Christ (Ma 297-98).For Macaskill, the new covenant is the covenant of the Spirit.gift given within the new covenant, who conforms our being to its terms terms on our hearts and realizing our conformity to Christ" (Ibid., p. 3 on Rom 8.14-17, Macaskill observes the distinctive partnership of the H human (Ibid., p. 240).However, in this treatment of Romans (Ibid., pp. 2 execute his own prescribed task, namely, paying attention to "the di Jesus and the Spirit" (Ibid., 145).In particular, he misses the different ro and "in the Spirit" in the Pauline theme of participation.ν α of the phrase ἐν Χριστῷ are recorded in the other undisputed Pauline epistles, among which nineteen refer to the believers' new identity in Christ (1 Cor 1.2,4,30;3.1;4.10,15a;2 Cor 5.17;12.2;Gal 1.22;2.4,17;3.26,28;Phil 1.1,4.21;1 Thess 2.14;4.16;Phlm 8,23), and twelve to the new lifestyle in Christ, personally and corporately (1 Cor 4.15b,17;15.18,19,31;2 Cor 2.17;12.19;Phil 1.13,26;3.3,14;Phlm 20).
Gorman helpfully identifies in Rom 8 two groups of phrases and words: the "in" group dominating the first half of the chapter, and the "with" group in the second half, such as "bearing witness with" (v.16), "joint heirs," "suffer with," "be glorified with" (v.17), "groans and suffers together" (v.22), and so on (vv. 26, 28, 29).In addition, there is the phrase "with him [Christ]" (σὺν αὐτῷ) (8.32).All these "exhibit Paul's profound spirituality of participation,… echo and further develop the 'in/into' and 'with' language associated with baptism and new life in Christ in 6.1-7.6"(Gorman 2022, pp. 193-94).
To flip the coin once again, the phrase "Christ in you" (Χριστὸς ἐν ὑμῖν) occurs three times in undisputed Pauline epistles (Rom 8.10;2 Cor 13.5;Gal 4.19). 20For Paul, the believers are not only the temple for the Holy Spirit, but are also a habitation for Christ.The Christians' taking up residence in Christ is realized though faith and baptism, when "people are moved into Christ, into the sphere of his Spirit; simultaneously, those who move into Christ find that Christ has moved into them" (Gorman 2022, p. 199).
Having identified the mutual indwelling of Christ/Spirit and the believers in Paul, I will now investigate the role of the Spirit in Rom 8.1-17, and the significance of pneumasis to Christosis.
Pneumasis: A New Terminology Highlighting the Primacy of the Spirit in Deification
According to Gorman, terms such as theosis/deification and Christosis/Christification have summarized a Christian understanding of salvation since at least the second century: Christ became what we are so that we could become what he is (Ibid.,p. 197).These terms, however, do not fully capture the role of the Holy Spirit in the Pauline understanding of participation.To do justice to Paul's prevailing references to the Spirit (compared to God and Christ) and his proto-trinitarian motif, this paper proposes the use of the word pneumasis or pneumafication.To coin a new theological term is, understandably, a risky business.However, I am persuaded, if not compelled, by the text to propose this new term as a protest against the downplaying or marginalizing of the role of the Holy Spirit in recent biblical and theological works on deification.As shown above, Rom 8.1-17 is so deeply saturated by Spirit language that it calls for the naming of the preeminent role of the Spirit.The occurrences of πνεύμα exceed the total number of references to God and Christ combined (πνεύμα: 17; θεός: 9; Χριστός: 6).Therefore, pneumasis deserves a seat at the table of participation as an equal partner with theosis and Christosis.If, using Pauline terminology and following Eastern Orthodox tradition, Christosis may be defined in terms of "in Christ" and "Christ in you" 21 , namely, the mutual indwelling of Christ and the believers, then pneumasis refers to believers' experience of being in the Spirit and being indwelt by the Spirit, or the mutual indwelling of the Holy Spirit and believers as a temple of the Spirit.
Based on his study of Rom 8, Blackwell suggests that Christosis is a better term than theosis/deification to describe Paul's specific soteriological emphasis, for two reasons.The first substantive reason is the particularly Christo-form nature of the experience.The more pragmatic reason relates to the "meaning" of the term theosis, which "can be ambiguous with regard to its referent because of its varied use in ancient and modern contexts" (Blackwell 2016, pp. 264-66).However, Blackwell acknowledges that the term Christosis may not do justice to the fact that "the distinctive role of the Spirit permeates our passages, especially in Rom 8, Gal 3-4, and 2 Cor 3".While I am appreciative of Blackwell's genuine struggle with how to balance the roles of God, Christ, and the Spirit in Pauline deification, I propose that the newly coined term pneumasis, or pneumafication, can better avoid the "possible overemphasis on Christ and underemphasis on the necessary and unique roles of … the Holy Spirit" (Ibid., p. 265).First, in Rom 8, when τ y the means of the Spirit," and that "Paul can use the language of mutual indwelling ith respect to Christ and the Spirit in the same breath" (Ibid.,p. 40).In interpreting hristians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a atement about suffering as meriting glory but a claim about the nature of full particiation in the messianic story.Christ's story is a narrative of suffering before full and final lory, of death before resurrection, of being humbled before being exalted" (Ibid.,p. 202).orman's insight comes close to the argument presented by this paper, but still falls short understanding the preeminent role of the Spirit in the mutual indwelling of Christ and is co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς ριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes at union with Christ is the "essential ingredient that binds all other elements together; is the webbing that connects the ideas of Paul's web-shaped theological framework" ampbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit bemes the means through whom union with Christ is lived out".He suggests that "sufring is to be viewed as a participatio Christi and not imitatio Christi only.Believers share the ongoing force of Christ's death and the power of his resurrection, and one conseuence of this is that believers will undergo suffering" (Ibid.,p. 448,italics original).espite his comprehensive exegetical work on union with Christ, Campbell's recognition f the Spirit's role as merely the means of the work of Christ downplays the personal ature of the Spirit and, therefore, marginalizes her role. 15His understanding would have een much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testaent, informed by historical theology and, to a lesser extent, systematic theology.In nthesizing his conclusions, he argues that the covenantal framework must serve as the arting point for reflection on participation or union with Christ (Macaskill 2013, pp. 7-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the ift given within the new covenant, who conforms our being to its terms by writing those rms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting n Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the uman (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to ecute his own prescribed task, namely, paying attention to "the distinctive place of sus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" d "in the Spirit" in the Pauline theme of participation.
To flip the coin once again, the phrase "Christ in you" (Xριστòς putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid.,p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).
Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.ν of the phrase ἐν Χριστῷ are recorded in the other undisputed Pauline epistles, among which nineteen refer to the believers' new identity in Christ (1 Cor 1.2,4,30;3.1;4.10,15a;2 Cor 5.17;12.2;Gal 1.22;2.4,17;3.26,28;Phil 1.1,4.21;1 Thess 2.14;4.16;Phlm 8,23), and twelve to the new lifestyle in Christ, personally and corporately (1 Cor 4.15b,17;15.18,19,31;2 Cor 2.17;12.19;Phil 1.13,26;3.3,14;Phlm 20).
Gorman helpfully identifies in Rom 8 two groups of phrases and words: the "in" group dominating the first half of the chapter, and the "with" group in the second half, such as "bearing witness with" (v.16), "joint heirs," "suffer with," "be glorified with" (v.17), "groans and suffers together" (v. 22), and so on (vv. 26, 28, 29).In addition, there is the phrase "with him [Christ]" (σὺν αὐτῷ) (8.32).All these "exhibit Paul's profound spirituality of participation,… echo and further develop the 'in/into' and 'with' language associated with baptism and new life in Christ in 6.1-7.6"(Gorman 2022, pp. 193-94).
To flip the coin once again, the phrase "Christ in you" (Χριστὸς ἐν ὑμῖν) occurs three times in undisputed Pauline epistles (Rom 8.10;2 Cor 13.5;Gal 4.19). 20For Paul, the believers are not only the temple for the Holy Spirit, but are also a habitation for Christ.The Christians' taking up residence in Christ is realized though faith and baptism, when "people are moved into Christ, into the sphere of his Spirit; simultaneously, those who move into Christ find that Christ has moved into them" (Gorman 2022, p. 199).
Having identified the mutual indwelling of Christ/Spirit and the believers in Paul, I will now investigate the role of the Spirit in Rom 8.1-17, and the significance of pneumasis to Christosis.
Pneumasis: A New Terminology Highlighting the Primacy of the Spirit in Deification
According to Gorman, terms such as theosis/deification and Christosis/Christification have summarized a Christian understanding of salvation since at least the second century: Christ became what we are so that we could become what he is (Ibid.,p. 197).These terms, however, do not fully capture the role of the Holy Spirit in the Pauline understanding of participation.To do justice to Paul's prevailing references to the Spirit (compared to God and Christ) and his proto-trinitarian motif, this paper proposes the use of the word pneumasis or pneumafication.To coin a new theological term is, understandably, a risky business.However, I am persuaded, if not compelled, by the text to propose this new term as a protest against the downplaying or marginalizing of the role of the Holy Spirit in recent biblical and theological works on deification.As shown above, Rom 8.1-17 is so deeply saturated by Spirit language that it calls for the naming of the preeminent role of the Spirit.The occurrences of πνεύμα exceed the total number of references to God and Christ combined (πνεύμα: 17; θεός: 9; Χριστός: 6).Therefore, pneumasis deserves a seat at the table of participation as an equal partner with theosis and Christosis.If, using Pauline terminology and following Eastern Orthodox tradition, Christosis may be defined in terms of "in Christ" and "Christ in you" 21 , namely, the mutual indwelling of Christ and the believers, then pneumasis refers to believers' experience of being in the Spirit and being indwelt by the Spirit, or the mutual indwelling of the Holy Spirit and believers as a temple of the Spirit.
Based on his study of Rom 8, Blackwell suggests that Christosis is a better term than theosis/deification to describe Paul's specific soteriological emphasis, for two reasons.The first substantive reason is the particularly Christo-form nature of the experience.The more pragmatic reason relates to the "meaning" of the term theosis, which "can be ambiguous with regard to its referent because of its varied use in ancient and modern contexts" (Blackwell 2016, pp. 264-66).However, Blackwell acknowledges that the term Christosis may not do justice to the fact that "the distinctive role of the Spirit permeates our passages, especially in Rom 8, Gal 3-4, and 2 Cor 3".While I am appreciative of Blackwell's genuine struggle with how to balance the roles of God, Christ, and the Spirit in Pauline deification, I propose that the newly coined term pneumasis, or pneumafication, can better avoid the "possible overemphasis on Christ and underemphasis on the necessary and unique roles of … the Holy Spirit" (Ibid., p. 265).First, in Rom 8, when µ of the phrase ἐν Χριστῷ are recorded in the other undisputed Pauline epistles, among which nineteen refer to the believers' new identity in Christ (1 Cor 1.2,4,30;3.1;4.10,15a;2 Cor 5.17;12.2;Gal 1.22;2.4,17;3.26,28;Phil 1.1,4.21;1 Thess 2.14;4.16;Phlm 8,23), and twelve to the new lifestyle in Christ, personally and corporately (1 Cor 4.15b,17;15.18,19,31;2 Cor 2.17;12.19;Phil 1.13,26;3.3,14;Phlm 20).
Gorman helpfully identifies in Rom 8 two groups of phrases and words: the "in" group dominating the first half of the chapter, and the "with" group in the second half, such as "bearing witness with" (v.16), "joint heirs," "suffer with," "be glorified with" (v.17), "groans and suffers together" (v. 22), and so on (vv. 26, 28, 29).In addition, there is the phrase "with him [Christ]" (σὺν αὐτῷ) (8.32).All these "exhibit Paul's profound spirituality of participation,… echo and further develop the 'in/into' and 'with' language associated with baptism and new life in Christ in 6.1-7.6"(Gorman 2022, pp. 193-94).
To flip the coin once again, the phrase "Christ in you" (Χριστὸς ἐν ὑμῖ) occurs three times in undisputed Pauline epistles (Rom 8.10;2 Cor 13.5;Gal 4.19). 20For Paul, the believers are not only the temple for the Holy Spirit, but are also a habitation for Christ.The Christians' taking up residence in Christ is realized though faith and baptism, when "people are moved into Christ, into the sphere of his Spirit; simultaneously, those who move into Christ find that Christ has moved into them" (Gorman 2022, p. 199).
Having identified the mutual indwelling of Christ/Spirit and the believers in Paul, I will now investigate the role of the Spirit in Rom 8.1-17, and the significance of pneumasis to Christosis.
Pneumasis: A New Terminology Highlighting the Primacy of the Spirit in Deification
According to Gorman, terms such as theosis/deification and Christosis/Christification have summarized a Christian understanding of salvation since at least the second century: Christ became what we are so that we could become what he is (Ibid.,p. 197).These terms, however, do not fully capture the role of the Holy Spirit in the Pauline understanding of participation.To do justice to Paul's prevailing references to the Spirit (compared to God and Christ) and his proto-trinitarian motif, this paper proposes the use of the word pneumasis or pneumafication.To coin a new theological term is, understandably, a risky business.However, I am persuaded, if not compelled, by the text to propose this new term as a protest against the downplaying or marginalizing of the role of the Holy Spirit in recent biblical and theological works on deification.As shown above, Rom 8.1-17 is so deeply saturated by Spirit language that it calls for the naming of the preeminent role of the Spirit.The occurrences of πνεύμα exceed the total number of references to God and Christ combined (πνεύμα: 17; θεός: 9; Χριστός: 6).Therefore, pneumasis deserves a seat at the table of participation as an equal partner with theosis and Christosis.If, using Pauline terminology and following Eastern Orthodox tradition, Christosis may be defined in terms of "in Christ" and "Christ in you" 21 , namely, the mutual indwelling of Christ and the believers, then pneumasis refers to believers' experience of being in the Spirit and being indwelt by the Spirit, or the mutual indwelling of the Holy Spirit and believers as a temple of the Spirit.
Based on his study of Rom 8, Blackwell suggests that Christosis is a better term than theosis/deification to describe Paul's specific soteriological emphasis, for two reasons.The first substantive reason is the particularly Christo-form nature of the experience.The more pragmatic reason relates to the "meaning" of the term theosis, which "can be ambiguous with regard to its referent because of its varied use in ancient and modern contexts" (Blackwell 2016, pp. 264-66).However, Blackwell acknowledges that the term Christosis may not do justice to the fact that "the distinctive role of the Spirit permeates our passages, especially in Rom 8, Gal 3-4, and 2 Cor 3".While I am appreciative of Blackwell's genuine struggle with how to balance the roles of God, Christ, and the Spirit in Pauline deification, I propose that the newly coined term pneumasis, or pneumafication, can better avoid the "possible overemphasis on Christ and underemphasis on the necessary and unique roles of … the Holy Spirit" (Ibid., p. 265).First, in Rom 8, when ν) occurs three times in undisputed Pauline epistles (Rom 8.10;2 Cor 13.5;Gal 4.19). 20For Paul, the believers are not only the temple for the Holy Spirit, but are also a habitation for Christ.The Christians' taking up residence in Christ is realized though faith and baptism, when "people are moved into Christ, into the sphere of his Spirit; simultaneously, those who move into Christ find that Christ has moved into them" (Gorman 2022, p. 199).
Having identified the mutual indwelling of Christ/Spirit and the believers in Paul, I will now investigate the role of the Spirit in Rom 8.1-17, and the significance of pneumasis to Christosis.
Pneumasis: A New Terminology Highlighting the Primacy of the Spirit in Deification
According to Gorman, terms such as theosis/deification and Christosis/Christification have summarized a Christian understanding of salvation since at least the second century: Christ became what we are so that we could become what he is (Ibid.,p. 197).These terms, however, do not fully capture the role of the Holy Spirit in the Pauline understanding of participation.To do justice to Paul's prevailing references to the Spirit (compared to God and Christ) and his proto-trinitarian motif, this paper proposes the use of the word pneumasis or pneumafication.To coin a new theological term is, understandably, a risky business.However, I am persuaded, if not compelled, by the text to propose this new term as a protest against the downplaying or marginalizing of the role of the Holy Spirit in recent biblical and theological works on deification.As shown above, Rom 8.1-17 is so deeply saturated by Spirit language that it calls for the naming of the preeminent role of the Spirit.The occurrences of πνε Religions 2023, 14, x FOR PEER REVIEW 4 of 13 entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role. 15His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.entirely correctly, that we would be mistaken to suggest that Paul's language about "Christ in me/you" has more or less the same meaning as "in Christ".He further notices that "Christ in you" is much closer to Paul's language about the Spirit, as is clear from Rom 8.9-11 (Ibid., p. 45).Unfortunately, due to his emphasis on the covenant and the vindication of the law, he does not tackle the crucial issues related to the terms "in Christ" and "in the Spirit" in his exegesis of Rom 8. [1][2][3][4][5][6][7][8][9][10][11].Sixth, in his work Romans: A Theological and Pastoral Commentary, Gorman notices that Rom 8 focuses on the life in the Spirit within the framework of a trinitarian understanding of salvation.He identifies at least five key elements of Christian life from Rom 8, three of which originate from vv. 1-17: first, the mutual indwelling of Christ and the Spirit with believers; second, believers' adoption as God's children; and third, the Christ-shaped pattern of suffering followed by glory (Gorman 2022, p. 192).Gorman affirms the individual and corporate aspects of participation in and with Christ (Ibid., p. 41).His unique contribution is the identification of the Holy Spirit as "the Spirit of cruciformity-cross-shaped participation in Christ" in Rom 8 (Ibid., p. 41), with the role of putting to death the deeds of the body (8.13) and empowering believers to suffer with Christ as the prelude to glory (8.17) (Ibid.,p. 193).Particularly germane to this paper is Gorman's affirmation that the mutual indwelling of Christ and the faithful "takes place by the means of the Spirit," and that "Paul can use the language of mutual indwelling with respect to Christ and the Spirit in the same breath" (Ibid., p. 40).In interpreting Christians' suffering with Christ (8.13), Gorman clarifies that, for Paul, "This is not a statement about suffering as meriting glory but a claim about the nature of full participation in the messianic story.Christ's story is a narrative of suffering before full and final glory, of death before resurrection, of being humbled before being exalted" (Ibid., p. 202).Gorman's insight comes close to the argument presented by this paper, but still falls short in understanding the preeminent role of the Spirit in the mutual indwelling of Christ and his co-sufferers.
Seventh, Campbell studies Paul's concept of union with Christ primarily in his exegesis of a few prepositional phrases associated with Christ, for example, ἐν Χριστῷ, εἰς Χριστόν, σὺν Χριστῷ, διὰ Χριστοῦ, σύν-compounds, and their variations.He concludes that union with Christ is the "essential ingredient that binds all other elements together; it is the webbing that connects the ideas of Paul's web-shaped theological framework" (Campbell 2012, p. 442).Campbell opines that "in the life of the believer, the Spirit becomes the means through whom union with Christ is lived out".He suggests that "suffering is to be viewed as a participatio Christi and not imitatio Christi only.Believers share in the ongoing force of Christ's death and the power of his resurrection, and one consequence of this is that believers will undergo suffering" (Ibid., p. 448, italics original).Despite his comprehensive exegetical work on union with Christ, Campbell's recognition of the Spirit's role as merely the means of the work of Christ downplays the personal nature of the Spirit and, therefore, marginalizes her role.15 His understanding would have been much thicker had he applied the same exegetical rigor to the study of the Spirit.
Finally, Macaskill performs a descriptive task on participation in the New Testament, informed by historical theology and, to a lesser extent, systematic theology.In synthesizing his conclusions, he argues that the covenantal framework must serve as the starting point for reflection on participation or union with Christ (Macaskill 2013, pp. 297-98).For Macaskill, the new covenant is the covenant of the Spirit.The Spirit "is the gift given within the new covenant, who conforms our being to its terms by writing those terms on our hearts and realizing our conformity to Christ" (Ibid., p. 300).Commenting on Rom 8.14-17, Macaskill observes the distinctive partnership of the Holy Spirit and the human (Ibid., p. 240).However, in this treatment of Romans (Ibid.,, he fails to execute his own prescribed task, namely, paying attention to "the distinctive place of Jesus and the Spirit" (Ibid., 145).In particular, he misses the different roles of "in Christ" and "in the Spirit" in the Pauline theme of participation.µα: 17; θεóς: 9; Xριστóς: 6).Therefore, pneumasis deserves a seat at the table of participation as an equal partner with theosis and Christosis.If, using Pauline terminology and following Eastern Orthodox tradition, Christosis may be defined in terms of "in Christ" and "Christ in you" 21 , namely, the mutual indwelling of Christ and the believers, then pneumasis refers to believers' experience of being in the Spirit and being indwelt by the Spirit, or the mutual indwelling of the Holy Spirit and believers as a temple of the Spirit.
Based on his study of Rom 8, Blackwell suggests that Christosis is a better term than theosis/deification to describe Paul's specific soteriological emphasis, for two reasons.The first substantive reason is the particularly Christo-form nature of the experience.The more pragmatic reason relates to the "meaning" of the term theosis, which "can be ambiguous with regard to its referent because of its varied use in ancient and modern contexts" (Blackwell 2016, pp. 264-66).However, Blackwell acknowledges that the term Christosis may not do justice to the fact that "the distinctive role of the Spirit permeates our passages, especially in Rom 8, Gal 3-4, and 2 Cor 3".While I am appreciative of Blackwell's genuine struggle with how to balance the roles of God, Christ, and the Spirit in Pauline deification, I propose that the newly coined term pneumasis, or pneumafication, can better avoid the "possible overemphasis on Christ and underemphasis on the necessary and unique roles of . . . the Holy Spirit" (Ibid., p. 265).First, in Rom 8, when speaking of believers' experiences, Paul's attention is always on the Spirit.The Spirit's work sustains a type of "dual agency" inaugurated through God's action in Christ (Eastman 2018, p. 123).Blackwell himself concurs that "the Spirit is central to Paul's portrayal of the believer's experience of the divine".He quickly adds that "such experience is christo-telic in nature, such that believers embody the Christ-narrative in death and life through the Spirit" (Blackwell 2016, p. 265).To overcome his circular reasoning, it makes more sense to use pneumasis to capture Paul's incidence of references to the Spirit in Rom 8. Second, Blackwell proposes the term Christosis because of the diversity of meanings associated with theosis/deification, a fact that "makes it difficult in some discussions to distinguish theosis as a modern term from theosis as a historical term" (Ibid., p. 266).Of course, by introducing the term Christosis, Blackwell does not want to do away with theosis.Instead, he suggests that "no single term will be sufficient for fully encompassing Pauline soteriology because he uses diverse terminology and metaphors to describe his theology.Accordingly, Christosis cannot be the only term we use to describe Pauline soteriology any more than justification can, but at the same time deification, theosis, and Christosis would not be inadequate terminology for describing Pauline soteriology" (Ibid., p. 267).The concept of pneumasis proposed in this paper fits well into his insightful observation.Paul's specific contribution in Rom 8 is a Spirit-filled soteriology; thus, pneumasis helps sharpen the analysis of this dimension of Paul's theology.
Having demonstrated the necessity and significance of pneumasis, the paper proceeds to define the role of the Spirit as illustrated in 8.1-17.
Pneumasis Effectuates Christosis
In 8.1-17, the role of the Holy Spirit is to effectuate, facilitate, and enable Christosis, namely the mutual indwelling of Christ and his co-sufferers.The reasoning is as follows.First, the role of the Spirit is highlighted through the following inverted parallelism.
A 8.1 Christosis: "in Christ" B 8.2-8 The Spirit liberates the believers and invokes their obligations C 8.9 Pneumasis: the mutual indwelling of the Spirit and the believers B' 8.10-15 The effect of pneumasis and the believers' further obligations A' 8.16-17 Christosis: "co-heirs, co-sufferers, being co-glorified" with/in Christ What is crucial to our purpose is that pneumasis lies at the center of the Pauline idea of Christosis.In particular, the γάρ (v. 2) explains that the believers' new status "in Christ" is made possible by the law of the Spirit, which sets them free from the bondage of the law of sin and death.Subsequently, the Spirit invokes the believers' participation to "walk according to the Spirit" (v.4), so that they can "set their minds on the things of the Spirit" to harvest life and peace (vv.5-6).For Lancaster, walking according to the Spirit entails "responding appropriately to God, living in that relationship and thus fulfilling the requirement of the law (8.4)" 22.As a result of pneumasis (v.9), God will give life to their mortal bodies through the Spirit (v.11) who is life (v.10).In v. 12, "thinking about debt Paul mentions as social obligation can open up the text to a reading about the community", in which the communal relationship determines our responsibilities to one another (Ibid., p. 139).Believers are "obligated" to put to death the deeds of the body by/in the Spirit, so that they will live (v.13). 23They manifest their status as children of God if they are led by the Spirit (v. 14).Paul concludes this pericope with the idea of Christosis, namely, of believers as Christ's co-heirs, co-sufferers, and co-sharers of his glorification (v.17), which is the result of the Spirit bearing witness with our spirit (v.16).
Third, the mutual indwelling of Christ and his believers can be shown to be effectuated by the mutual indwelling of the Spirit and the believers.First, for the believers to be "in Christ" (8.1), they need to be released from the bondage of the law of sin and death (v.2).Paul makes it crystal clear that such is the work of the Spirit, and that the law of the Spirit sets them free from the tyranny of sin and death (v.2).Second, the only possible way for Christ to indwell believers is through .Therefore, it is only reasonable to appreciate Paul's high pneumatology (at least in this pericope) and to recognize that the Spirit facilitates, empowers, enables, and effectuates the mutual indwelling of Christ and believers.Gorman identifies two dimensions of the Spirit-filled life as the effects of the Spirit's presence: (1) putting to death the deeds of the body, or dying to the flesh in order to truly live (8.13); and (2) suffering with Christ as the prelude to glory (8.17) (Gorman 2022, p. 193).Therefore, a conclusion can be drawn that it is the Spirit who facilitates believers' participation in Christ's life and suffering.
Fourth, the leading role of the Spirit in Christosis can be seen in 8.14-15.Here, Lancaster is insightful in pointing out Paul's wordplay in Greek: "The word translated as 'adoption' conveys the idea of 'son-making' in Greek (huiothesias), so Paul refers to the community members as 'sons' (huioi) in 8.14.The Greek wordplay between 'sons' and 'son-making' also calls to mind that in 8.3 Paul says God sent the Son into the world to deal with sin.Through Jesus Christ, God's own Son, we become adopted sons and therefore join heirs with Christ" (Lancaster 2015, p. 141).Paul's wordplay around Christ as God's Son and believers as "adopted sons" captures the essence of Christosis.The primacy of the Spirit can be shown by the fact that, first, "God's sons" are those who are led by the Spirit (8.14) and, second, we can call "Abba!Father!" because we have received the Spirit of "son-making" (8.15).God sends the Spirit into the community of believers so that they are made "sons" of God, who can cry "Abba!Father!" just like Jesus, the Son of God, did (Mark 14.36;cf. Gal 4.6).Without possessing and being possessed by the Spirit, such a "son-making" process cannot be achieved.
Pneumaficational Anthropology
The neglected role of the Holy Spirit is also reflected in the one-sided emphasis on terms such as "the indwelling Christ," and much less on "the indwelling Spirit," and the neglect of the Spirit-enabled and experiential "indwelling humanity": "in Christ" and "in the Spirit."Combatting an individualistic, hidden, and inward mindset concerning spiritual experiences, Eastman is insightful in stating that "the basis of the Spirit's claim on every aspect of life is not simply our given reliance on God, but Paul's participatory Christology and pneumatology; because Christ entered fully into human existence, there is no unclaimed or 'secular' territory of the self or society; because the Spirit continues that divine participation, there is no 'unspiritual' experience beyond the reach of transformation" (Eastman 2018, p. 123).Part and parcel of Paul's participatory anthropology is the interpersonal experience that exists as the relational bonds between members of Christ's body (Ibid.,.Building on Eastman's insight, this paper proposes another aspect of such a participatory or, better, pneumaficational anthropology: humanity dwells in the Spirit and, consequently, in Christ, which is qualified by the fact that, in contrast to Christ's pervasive entrance into human existence, and the Spirit's ubiquitous penetration into the self and society, human participation in the divine is always limited by human weakness (8.26).In Rom 8, Paul portrays ongoing warfare, a constant struggle between the Spirit and the flesh (vv. 4, 5, 6, 9, 13).Human failures due to weakness (v.26) in this warfare cause creation to be subjected to futility (v. 20), to suffer the bondage to decay (v.21), and to groan in labor pains, until now (v.22).However, pneumasis entails first, the Spirit helping us in our weakness by interceding for us (v.26), and second, that the new life in Christ "is not a spiritual free-for-all" (Gorman 2022, p. 194).Life in the Spirit is not automatic, but requires active participation by believers (Ibid.,p. 200).To live in the Spirit "is to live according to, or in sync with . . . the Spirit rather than the flesh (8.4, 5, 12, 13).. . .In other words, the Spirit enables the children of God to resemble their Father by resembling his Son, who is their elder brother" (Ibid.,p. 194).As a result, Paul's pneumaficational anthropology brings hope through the eschatological tension between the "already" ("in hope we were saved" [v.24]) and the "not yet" (by active participation in the Spirit "we hope for what we do not see" [v. 25]).
The result of such a participatory anthropology is, first, "the freedom of the glory of the children of God," contingent on which is the cosmological freedom shared between creation and humanity (v.21).Gorman summarizes Rom 8 with the heading, "Life in the Spirit: Resurrectional Cruciformity" (Ibid.,p. 191).Life in the Spirit "is a life of joyful, resurrection-infused cruciformity" (Ibid.,p. 192,italics original).I would further add that, between such a present joyful life in the Spirit (vv.1-17a) and future glory and freedom (vv.18b-30) lies a life of co-suffering with .Such Spirit-enabled co-suffering is, paradoxically, integral to the joyful life affirmed in Rom 8.For Paul, the key to realizing eschatological and cosmological hope is in Christ's co-sufferers staying actively "in the Spirit" by allowing the Spirit to indwell them and direct them; hence, pneumasis.
Summary of the Spirit's Role in Deification
In this section, I have presented an exegetical analysis of Rom 8.1-17 from two angles: the mutual indwelling of the Spirit and believers, and the mutual indwelling of Christ and the faithful.I have offered a theological rationale for coining the new theological term pneumasis/pneumafication.I have further demonstrated that the Holy Spirit is God's facilitator of the mutual indwelling of Christ and believers.Finally, based on these observations, I have proposed a pneumaficational anthropology: although limited by human weakness, believers are enabled by the Spirit to participate in the divine as Christ's co-sufferers.
Conclusions
In this paper, I have followed the general contours established by Campbell and Macaskill and surveyed the works on Romans by a number of scholars, ranging from the early twentieth century to the present.What I have noticed is the increasing scholarly attention on Pauline pneumatology and deification.However, a lacuna has been identified concerning the neglect of the preeminent role of the Holy Spirit in deification and as God's solidarity with Christ's co-sufferers.In order to fill such a lacuna, I have argued for the crucial role of the Holy Spirit as one who facilitates, enables, and empowers the mutual indwelling of Christ and his co-sufferers in Rom 8.1-17.To avert the misconception of a subordinated role of the Spirit in deification, I have coined the term pneumasis, or pneumafication, which goes hand in hand with theosis and Christosis.Such a term is necessary to nuance and supplement the language of theosis and Christosis and constitute a genuinely trinitarian doctrine of participation.Theosis, Christosis, and pneumasis-or deification, Christification, and pneumafication-work together to reflect, and do full justice to, Paul's proto-trinitarian thoughts on participation.The newly coined term also makes it possible to capture the distinct but inseparable functions of God, Christ, and the Spirit in participation.One cannot separate pneumasis from Christosis (or the Spirit from Christ); neither can one separate it from theosis (or the Spirit from God), because participation is, indeed, a triune divine encounter.Another benefit of this new term is an increased awareness of a participatory or, better, pneumaficational anthropology, in which human participation into the divine, though often hindered by human weakness in the struggle between the Spirit and flesh, calls for the Spirit-enabled Christian obligation to walk in the Spirit, while being led by the Spirit in every daily mundane affair.
Pauline pneumasis is pivotal to constructing contemporary Christian ethics through questioning the "private", "inner", and "subjective" Cartesian understandings of Christian spiritual experiences.Moreover, the pneumaficational experiences highlighted in Rom 8.1-17 focus on the Spirit as God's solidarity with Christ's co-sufferers in the present suffering creation.This may serve as a healthy critique of some parts of Western Christianity, which have led to a black-and-white worldview that divides humanity into the spiritual and the fleshly, resulting in an arrogant mindset reflected in civil and international politics.On the contrary, based on a close reading of Paul's passage, pneumaficational ethics invite readerly formation to be identified with the Holy Spirit.In this way, we can be rescued from an introspective and judgmental spirituality, participate in Christ's suffering, and stand in solidarity with Christ's co-sufferers and the suffering creation.By doing this, the voice of the Holy Spirit's groaning, joined by Paul in his pneumaficational experiences, will not be left unheard.
Blackwell coined the term Christosis to nuance and supplement the language of theosis.Such a term speaks to his reading of Paul's Christocentric doctrine and spirituality.For Blackwell, "the term 'christosis' (or christopoiesis) serves to capture the embodiment of Christ's death and life that is so fundamental for Paul's spirituality."See (Blackwell 2016, p. xix).
13 Constantine Campbell argues that, "while union with Christ has been discussed and explored at various times in the history of the Christian church, the volume and intensity of such discussions and explorations became significantly heightened through the twentieth century."See (Campbell 2012, p. 32). 14 This does not mean that the works that do not contain exegetical works are unimportant.Many of these twenty-three authors contributed significantly to the topic of union with Christ and participation.However, without detailed exegesis, many nuanced observations could be missed.15 I follow Johannes Van Oort, who argues that, "in the image of the Holy Spirit as woman and mother, one may attain a better appreciation of the fulness of the Divine."See (Van Oort 2016, p. 1).16 Fee's analysis of the inadequacy of the traditional view of Paul's theology applies here.The traditional view, fostered by the Reformers and perpetuated by generations of Protestants, is that "justification by faith" is the key to that theology.The inadequacy of the view is that "it focuses on one metaphor of salvation, justification, to the exclusion of others."See (Fee 1996, p. 5).
of 13
en to suggest that Paul's language about meaning as "in Christ".He further notices language about the Spirit, as is clear from to his emphasis on the covenant and the he crucial issues related to the terms "in om 8. [1][2][3][4][5][6][7][8][9][10][11].and Pastoral Commentary, Gorman notices ithin the framework of a trinitarian undere key elements of Christian life from Rom 8, , the mutual indwelling of Christ and the option as God's children; and third, the y glory (Gorman 2022, p. 192).Gorman aff participation in and with Christ (Ibid., p. ion of the Holy Spirit as "the Spirit of crust" in Rom 8 (Ibid., p. 41), with the role of ) and empowering believers to suffer with 193).Particularly germane to this paper is ling of Christ and the faithful "takes place an use the language of mutual indwelling same breath" (Ibid.,p. 40).In interpreting an clarifies that, for Paul, "This is not a ut a claim about the nature of full particia narrative of suffering before full and final mbled before being exalted" (Ibid.,p. 202).presented by this paper, but still falls short pirit in the mutual indwelling of Christ and t of union with Christ primarily in his exed with Christ, for example, ἐν Χριστῷ, εἰς pounds, and their variations.He concludes ient that binds all other elements together; aul's web-shaped theological framework" t "in the life of the believer, the Spirit be-Christ is lived out".He suggests that "sufnd not imitatio Christi only.Believers share power of his resurrection, and one consesuffering" (Ibid., p. 448, italics original).union with Christ, Campbell's recognition e work of Christ downplays the personal s her role.15 His understanding would have egetical rigor to the study of the Spirit.e task on participation in the New Testato a lesser extent, systematic theology.In e covenantal framework must serve as the or union with Christ (Macaskill 2013, pp. he covenant of the Spirit.The Spirit "is the µα as local: "The thought of the Spirit in [humanity] is local.. ..The converse that [humanity] is in the Spirit . . . is also based on a spatial sense" (Ibid., p. 2:540).Oepke's definition is complemented byLouw and Nida, who interpret . Gorman af-spects of participation in and with Christ (Ibid., p. entification of the Holy Spirit as "the Spirit of cru-n in Christ" in Rom 8 (Ibid., p. 41), with the role of dy (8.13) and empowering believers to suffer with(Ibid., p. 193).Particularly germane to this paper is l indwelling of Christ and the faithful "takes place t "Paul can use the language of mutual indwelling t in the same breath" (Ibid., p. 40).In interpreting 13), Gorman clarifies that, for Paul, "This is not a g glory but a claim about the nature of full partici-story is a narrative of suffering before full and final being humbled before being exalted" (Ibid., p. 202).rgument presented by this paper, but still falls short of the Spirit in the mutual indwelling of Christ and 's concept of union with Christ primarily in his ex-associated with Christ, for example, ἐν Χριστῷ, εἰς ύν-compounds, and their variations.He concludes al ingredient that binds all other elements together; deas of Paul's web-shaped theological framework" pines that "in the life of the believer, the Spirit be-on with Christ is lived out".He suggests that "suf-Christi and not imitatio Christi only.Believers share and the power of his resurrection, and one conse-undergo suffering" (Ibid., p. 448, italics original).work on union with Christ, Campbell's recognition ans of the work of Christ downplays the personal rginalizes her role.
Christi and not imitatio Christi only.Believers share and the power of his resurrection, and one conse-undergo suffering" (Ibid., p. 448, italics original).workonunion with Christ, Campbell's recognition ans of the work of Christ downplays the personal rginalizes her role.15Hisunderstanding would have | 41,777.4 | 2023-09-20T00:00:00.000 | [
"Philosophy"
] |
Human Rhinovirus Selectively Modulates Membranous and Soluble Forms of Its Intercellular Adhesion Molecule–1 (ICAM-1) Receptor to Promote Epithelial Cell Infectivity*
Human rhinoviruses are responsible for many upper respiratory tract infections. 90% of rhinoviruses utilize intercellular adhesion molecule-1 (ICAM-1) as their cellular receptor, which also plays a critical role in recruitment of immune effector cells. Two forms of this receptor exist; membrane-bound (mICAM-1) and soluble ICAM-1 (sICAM-1). The soluble receptor may be produced independently from the membrane-bound form or it may be the product of proteolytic cleavage of mICAM-1. The ratio of airway epithelial cell expression of mICAM-1 to the sICAM-1 form may influence cell infectivity and outcome of rhinovirus infection. We therefore investigated the effect of rhinovirus on expression of both ICAM-1 receptors in normal human bronchial epithelial cells. We observed separate distinct messenger RNA transcripts coding for mICAM-1 and sICAM-1 in these cells, which were modulated by virus. Rhinovirus induced mICAM-1 expression on epithelial cells while simultaneously down-regulating sICAM-1 release, with consequent increase in target cell infectivity. The role of protein tyrosine kinases was investigated as a potential mechanistic pathway. Rhinovirus infection induced rapid phosphorylation of intracellular tyrosine kinase, which may be critical in up-regulation of mICAM-1. Elucidation of the underlying molecular mechanisms involved in differential modulation of both ICAM-1 receptors may lead to novel therapeutic strategies.
Human rhinoviruses (HRV) 1 are the most frequent cause of upper respiratory tract infections known as the "common cold." Although these infections are generally mild and self-limiting, they inflict a heavy economical burden due to high loss of productivity and medical costs (1). Currently, there is no effective treatment for HRV infections; over the counter cold rem-edies only alleviate the symptoms but do not eradicate the virus.
Primarily, HRV target epithelial cells for attachment and entry. These cells express intercellular adhesion molecule 1 (ICAM-1), the receptor for 90% of HRV serotypes (2). Both this major group of HRV and the 10% of HRV that use alternative receptors for cell attachment enhance cell surface ICAM-1 expression (3). This glycoprotein, belonging to the immunoglobulin supergene family, consists of five Ig-like domains (4); domains 1 and 2 have been shown to fit snugly in a key-lock relationship into reciprocal canyons on the HRV shell (5). In addition, to this critical role as a docking molecule during HRV infection, ICAM-1 through separate domains with its cognate ligand LFA-1 (CD18/CD11a) drives the migration of immuneeffector cells to sites of inflammation (6). While most studies refer to the membranous form of ICAM-1, a soluble form (sICAM-1) has also been described (7). The molecular mass of sICAM-1 is similar to the molecular mass of the extracellular domain of ICAM-1 (80 -114 kDa) depending on the level of glycosylation, suggesting that this soluble circulating form of ICAM-1 consists of most of the extracellular domain of membranous ICAM-1 (7). Several circulating isoforms of sICAM-1 have also been detected of 240, 430, and Ͼ500 kDa in size, indicating that sICAM-1 may circulate in a complexed form either with itself or with other proteins (7)(8). While the exact origin of sICAM-1 is unclear, sICAM-1 may be produced directly from the membrane-bound form by proteolytic cleavage (9 -10) or produced independently by an alternative splicing mechanism (11). What role is played by soluble ICAM-1 in disease pathogenesis remains to be elucidated.
Previous studies have demonstrated that pro-inflammatory cytokines can alter the expression of mICAM-1 (3,(12)(13)(14) and sICAM-1 (15)(16). In addition, HRV infection has been shown to significantly up-regulate the expression of its membranebound receptor ICAM-1 on the surface of epithelial cells (3,(12)(13)(14) leading to an increase in epithelial cell infectivity (12). However, there is evidence indicating that sICAM-1 may have the opposite effect because it possesses antiviral properties both in vitro (17) and in vivo (18 -19). Therefore, the dynamic inter-relationship between mICAM-1 and sICAM-1 forms may have a critical bearing on the pathogenesis as well as course of HRV infection. Thus, there is a need for a better understanding of the interaction between mICAM-1 and sICAM-1, leading to potential targets for therapeutic modulation of the course of HRV infection.
To investigate this hypothesis, a series of studies were undertaken to establish the presence of distinct mRNA transcripts coding for mICAM-1 and sICAM-1 in an in vitro airway epithelial cell model and to determine whether HRV has the ability to modulate the two forms of ICAM-1 and how this affects epithelial cell infectivity. Having found that epithelial cells express the two ICAM-1 forms and that HRV could selectively modulate membrane and soluble ICAM-1 expression in an inverse fashion to promote/propagate infection, we explored the potential intracellular mechanisms influencing this differential modulation to identify potential targets for antiviral therapy.
ICAM-1 is a rapid response gene with a complex pattern of regulation (20). Numerous second messenger signaling pathways involved in the activation of the ICAM-1 gene have been identified (20 -21). A recent study demonstrated that HRV up-regulates membrane-bound ICAM-1 expression in airway epithelial cells via an NFB-dependent mechanism (3). We investigated this intracellular signaling pathway further and examined the effect of HRV on tyrosine kinase phosphorylation in airway epithelial cells. In addition, we utilized inhibitors of gene transcription and protein synthesis to investigate potential mechanisms responsible for the modulation of sICAM-1 production/release during HRV infection.
Our results demonstrate that HRV selectively induces mICAM-1 expression on epithelial cells, at least in part, through a tyrosine kinase-dependent pathway, while, HRV influence on sICAM-1 release may involve the down-regulation/inhibition of proteolytic enzymes associated with the cleavage of mICAM-1 from the epithelial cell surface. The interaction of HRV with these intracellular molecular pathways controlling mICAM-1/sICAM-1 ratios will need to be further dissected, specifically for the development of new anti-HRV strategies aimed at either halting progression or reversing host cell infectivity.
Viral Stocks-The main rhinovirus seed (HRV-14) was kindly donated by J. Kent (University of Leicester). A stock solution of HRV-14 was generated by infecting confluent monolayers of HeLa Ohio cell line as described previously (Sethi et al.,Ref. 12). Briefly, confluent monolayers of Hela cells were inoculated with a known dilution (10 2.5 , TCID 50 /ml) of HRV-14 and incubated for 90 min at 34°C in humidified air containing 5% CO 2 , after which, cells were cultured until the cytopathic effect (CPE) was Ͼ80%. Medium containing virus was centrifuged at 600 ϫ g for 10 min, after which the viral suspension was stored at Ϫ80°C until required.
Viral Purification-Prior to use viral stocks were purified using a sucrose gradient. 20 g/ml RNase A (Sigma) was added to the viral suspension and incubated at 35°C for 20 min. 1% sodium sarkosyl (Sigma) and 2-mercaptoethanol (1 g/ml) were added to the RNasetreated viral suspension. This was then overlaid on 1 ml of purification solution (20 mM Tris acetate, 1 M NaCl, 30% w/v sucrose) and centrifuged at 200,000 ϫ g for 5 h at 16°C. The supernatant was discarded, and the resulting virus pellet was resuspended in medium and stored at Ϫ80°C until required.
HRV-14 Infection of Epithelial Cells-Once cell monolayers were 70 -80% confluent the culture medium was removed, and the cells were inoculated with HRV-14 10 2.5 TCID 50 /ml for 90 min at 34°C, 5% CO 2 / air. The cells were washed, and maintenance medium was added to sustain cell growth. infection. At each time point, cells from 6-well plates were collected via trypsinization and centrifugation; 6 cytospins for each experimental condition at each time point were prepared for immunostaining. The remaining cells were utilized for RNA extraction, cDNA synthesis, and reverse transcription (RT)-PCR. Internal controls consisting of unstimulated and uninfected cells were set up at each time point to allow comparisons between controls and treated cells. A cytokeratin immunoglobulin IgG-specific monoclonal antibody (Sigma) was used to confirm the epithelial origin of NHBE cell lines. Surface ICAM-1 was semiquantified using a 3-step indirect immunoenzymatic labeling method (22) and modified as described previously (12)(13). Briefly, exogenous peroxidase staining was blocked using 2% bovine serum albumin/phosphate-buffered saline solution. NHBE cells were incubated with ICAM-1 monoclonal antibody at a concentration of 5 g/ml, (R1/1.1, IgG, Boehringer Ingelheim) at room temperature for 30 min. Cells were then washed using a washing buffer (Tris stock solution, 0.05 mol/liter, pH 7.4, NaCl, 0.9% saline solution), and incubated with rabbit anti-mouse IgG conjuagated to peroxidase at a concentration of 1 mg/ml (Dako) for 30 min; the cells were then washed again. A third antibody, swine anti-rabbit IgG also conjugated to peroxidase, (DAKO), was then added at a concentration of 0.8 mg/ml to amplify the staining intensity. The cells were then incubated with the substrate 3,3-diaminobenzidine tetrahydrochroride (0.6 mg/ml, Sigma) and stained with Mayers Hemalum solution (BDH). The cells were incubated with an anti-mouse IgG antibody (Coulter clone) at a concentration of 10 g/ml, which acted as a negative control.
To avoid observer bias, the cytospins were scored by two independent observers (A. Bianco and S. Whiteman); a mean of three readings of each slide was performed. Two cytospins for each experimental condition per time point were assessed at ϫ400 magnification with a light microscope (Olympus CH-2 microscope, Olympus Optical Co., Ltd., Tokyo). 300 cells per microscopic field were counted and surface ICAM-1 on epithelial cells was assessed using a 5 point scoring scale based on the intensity of staining and appearance of the nucleus: 0, gray/brown; 1, light brown; 2, medium brown; 3, medium/dark brown; 4, dark brown; in grades 0 -2 the nucleus appears well defined, in 3-4 the nucleus is partially or fully obliterated. The number of cells scored in each grade was then multiplied with the respective grade index, and the resulting values summed. The final result was expressed as the Pox score (12)(13)22), defined as the difference between the sum of the specific and background staining: [(a ϫ 0) ϩ (b ϫ 1) ϩ (c ϫ 2) ϩ (d ϫ 3) ϩ (e ϫ 4)] Ϫ value for the control slide ϭ POX score, where each letter represents the number of cells scored in the respective grade. The coefficient of variability of the differences between the counts obtained from all slides by both observers was between 4 and 12%; and that between the two observers for each time point was less than 5%.
Measurement of sICAM-1 Protein in Cell Culture Supernatants-Soluble ICAM-1 expression was also evaluated at 0, 8, 24, and 96 h post-HRV-14 infection. Internal controls consisting of unstimulated and uninfected cells were setup at each time point to allow comparisons between controls and treated cells. Cell culture supernatants retrieved from each experimental condition at each time point were assayed for soluble ICAM-1 using a commercially available ELISA kit (BioSource International, CA). The minimum detectable level of human soluble ICAM-1 (hsICAM-1) was Ͻ0.04 ng/ml. 100 l of undiluted cell culture supernatant or standard were utilized in the assay, which was performed in accordance with the manufacturer's guidelines.
RNA Extraction and cDNA Synthesis-Total RNA was extracted from NHBE cells at 0, 8, 24, and 96-h postinfection using Trizol (Invitrogen) according to the manufacturer's guidelines, and cDNA was synthesized from 2 g of RNA. cDNA synthesis was conducted in a reaction mixture containing 20 pmol oligo(dT) primer, 5ϫ buffer (50 mM, pH 8.3, 75 mM KCl, 3 mM MgCl 2 ), 0.5 mMdNTP mixture, 0.5 units of RNase inhibitors, and 200 units of MMLV reverse transcriptase; the total reaction volume was 20 l. All cDNA synthesis reagents were obtained from Clontech. This was then incubated at 42°C for 1 h after which the reverse transcriptase and DNase were heat-inactivated at 94°C for 5 min. The cDNA was then diluted to a final volume of 100 l and stored at Ϫ80°C for RT-PCR.
Detection of Membrane and Soluble ICAM-1 Gene Expression using RT-PCR-Glyceraldehyde-3-phosphate dehydrogenase (G3PDH) was used as a control for cDNA synthesis and RT-PCR. Primers used to detect G3PDH were 5Ј-TGA AGG TCG GAG TCA GA-3Ј (sense) and CAT GTG GGC CAT GAG GTC CAC CAC (antisense). Primers for the detection of mICAM-1 and sICAM-1 were based on those described previously by Wakatsuki et al. (11). The sequence of the forward primer used to detect mICAM-1 was 5Ј-CAA GGG GAG GTC ACC CGC GAG GTG-3Ј and 5Ј-CAA GGG AGG TCA CCC GCG AGC C-3Ј. Both primers were used in combination with a common reverse primer with the following sequence 5Ј-TGC AGT GCC CAT TAT GAC TG-3Ј. These primer pairs encompass the transmembrane domain of ICAM-1. The primers used to detect HRV were CGG ACA CCC AAA GTA G (sense) and GCA CTT CTG TTT CCC C (antisense). The RT-PCR consisted of 25 pmol of primers, 200 M dNTPs, 1.5 mM MgCl 2 , 5 l 10ϫ PCR buffer, and 2.5 units of Amplitaq Gold (PerkinElmer Life Sciences) in a 50-l reaction mixture in a thermal cycler (PTC 200 Pielter Thermal cycler) under the following conditions: 95°C for 12 min, 94°C for 1 min and 15 s (denaturation step), 60°C (G3PDH) or 65°C (ICAM-1) for 1 min and 15 s (annealing step) and 72°C for 1 min (extension step) for a total of 30 cycles (G3PDH) or 35 cycles (ICAM-1), after which a final extension step was performed at 72°C for 10 min. RT-PCR products were resolved using 3% metaphor agarose (Flowgen) gel in TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA, Sigma). Gels were visualized using ethidium bromide and UV light and analyzed densitometrically (Model GS-670, BioRad) using Molecular Analyst (version 1.5). Restriction endonucleases were used to confirm the size of the RT-PCR products. Membrane-bound ICAM-1 RT-PCR products were digested to give product sizes of 45 and 57 base pairs, and soluble ICAM-1 RT-PCR products were digested at the site of the deletion to give 63 and 20 bp products. In addition, to confirm the presence of the 19-bp deletion, RT-PCR amplicons were sequenced using an ABI PRISM automated sequencer model 310.
Viral Titer Assay-The TCID 50 method was used to calculate the concentration of the virus in cell culture supernatants at 0, 8, 24, and 96 h postinfection. Serial dilutions of cell culture supernatants were incubated in cell monolayers in 96-well plates for 5 days at 34°C in humidified air containing 5%CO 2 . The presence of cytopathic effect (CPE) in the wells was used to calculate the TCID 50 using the Karber formula (12)(13)(14). Furthermore, cell infectivity was confirmed by detecting the presence of HRV RNA within the cell using RT-PCR (23).
Prevention of Virus-Receptor Binding-To confirm the changes observed in ICAM-1 expression were true effects of HRV and not due to soluble factors within the viral inoculum, separate studies were designed using anti-ICAM-1 monoclonal antibodies to block HRV attachment and subsequent infection. NHBE cell monolayers were washed and incubated with anti-ICAM-1 monoclonal antibodies (mAb) (R1/1.1 Boerhinger Ingelheim) at separate concentrations of 4, 8, and 16 g/ml for 1 h at 37°C under 5% CO 2 -humidified air (24). After which the anti-ICAM-1 mAb solution was removed, and cell monolayers were washed and inoculated with HRV-14 at a concentration of 10 2.5 TCID 50 /ml for 90 min at 34°C under 5% CO 2 -humidified air. After 90 min the viral inoculum was removed, cell monolayer washed, and maintenance medium was replaced. Gene expression of both ICAM-1 forms was evaluated using RT-PCR at 24 h, because this time point was previously shown to reflect an optimum response in gene expression. Corresponding viral titers were also measured.
Inhibition of de Novo Protein Synthesis-To assess the level at which HRV-14 regulates the expression of both ICAM-1 receptors, separate cell cultures were preincubated with cycloheximide, an inhibitor of de novo protein synthesis at a concentration of 10 g/ml for 2 h at 37°C in humidified air containing 5% CO 2 . Cycloheximide is widely used as an inhibitor of protein synthesis and has an effect at 10 g/ml (25). The cell monolayers were then washed and infected with HRV-14 as described above. Membrane and soluble ICAM-1 protein levels were assessed as described above.
Inhibition of Gene Transcription-Actinomycin D, an inhibitor of gene transcription was used to assess the effect of HRV-14 on ICAM-1 gene transcription. NHBE cells were incubated with actinomycin D at a concentration of 10 g/ml for 2 h at 37°C in humidified air containing 5% CO 2 . 10 g/ml was identified as the optimum dose in previous dose response experiments (data not shown). The cell monolayers were then washed and infected with HRV-14 as described above. Membranebound and soluble ICAM-1 were assessed using RT-PCR and semiquantified using densitometry (Model GS-670, BioRad) and Molecular Analyst Software (version 1.5).
Investigation of the Role of Tyrosine Kinases in HRV Induction of ICAM-1 Expression-The involvement of tyrosine kinase in the HRVdriven up-regulation of mICAM-1 gene expression was investigated using Western blot. NHBE cells were cultured in SABM (detailed under "Materials and Methods") and infected with HRV-14 (TCID 50 10 2.5 ) for the indicated times (0 -30 min, 0 min represents the viral inoculum placed on the cell monolayer and then immediately removed). Cell were also pretreated with genistein (50 and 100 M), an inhibitor of tyrosine kinase for 1 h at 37°C in 5% CO 2 -humidified air. NHBE cells were lysed using a lysis buffer containing 1% Triton X-100, 20 mM Tris-HCl, pH 8.0, 137 mM sodium chloride, 10% glycerol, 1 mM sodium orthovanadate, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 20 M leupeptin, and 0.15 units/ml aprotinin. All reagents were molecular biology grade and obtained from Sigma. The cells were placed on ice for 20 min, and total protein was collected by centrifugation. Total protein was assayed using a commercially available kit based on the Lowry assay (Bio-Rad). 25 g of reduced protein samples were electrophoresed on 12.5% SDS-PAGE and transferred to nitrocellulose membranes (sandwiches, 0.45-m pore size Novex, San Diego). Molecular weight markers and epidermal growth factor receptor (control) were run with the samples. Membranes were blocked with 10% (w/v) low fat milk for 1 h in TBS-T and probed for 2 h with mouse anti-human phosphotyrosine kinase (clone 4G10, Upstate Technology) diluted 1:3000 in TBS-T. Membranes were incubated with a horseradish peroxidase-conjugated rabbit anti-mouse antibody (Dako) diluted 1:40,000 in TBS-T for 1 h. The ECL system was used for detection (Amersham Biosciences). The membranes were reprobed with a control mouse IgG as a negative control. The membranes were analyzed densitometrically (Model GS-670, BioRad) and Molecular Analyst Software (version 1.5).
Statistical Analysis-Each experiment was performed three times. Data were expressed as means Ϯ S.E., and comparisons between experimental conditions and controls were performed by paired Student's t test. Probability values Ͻ 0.05 were considered significant.
Epithelial Cell Expression of Two Distinct Forms of mRNA
Coding for ICAM-1-Two distinct mRNA transcripts were observed in NHBE cells after RT-PCR (Fig. 1). These RT-PCR products corresponded with those observed by Wakatsuki et al. (11) with RT-PCR products of 102 and 83 bp for mICAM-1 and sICAM-1, respectively. Product sizes were confirmed by restriction enzyme digestion (data not shown).
Influence of HRV-14 on sICAM-1 Expression-Basal sICAM-1 protein release in cell culture supernatants increased at 8 h ICAM-1 mAb also resulted in a decrease in mICAM-1 expression. In addition, these higher concentrations also produced small but significant increases in sICAM-1 expression (*, p Ͻ 0.02, Fig. 4A), suggesting the effects of HRV were inhibited. ICAM-1 mAb decreased viral titers in cell culture supernatants in a dose-dependent manner (Fig. 4B, *, p Ͻ 0.02). Inhibition of de Novo Protein Synthesis-Treatment of NHBE cells with cycloheximide (10 g/ml) resulted in an inhibition of HRV-induced mICAM-1 protein expression at 0 and 8 h, suggesting that HRV-14 induces de novo protein synthesis of mICAM-1. However cycloheximide had no significant effect on mICAM-1 protein expression at 24 and 96 h (data not shown). This may be caused by the depletion of cycloheximide in the culture medium as it has a short half-life. In contrast, sICAM-1 release from the same HRV-infected NHBE cells was not affected by cycloheximide, implying that the regulation of sICAM-1 release from infected cells is not dependent on de novo protein synthesis.
Inhibition of Gene Transcription-Treatment of NHBE cells with actinomycin D (10 g/ml) resulted in a complete inhibition of the expected HRV-induced mICAM-1 protein expression at 8, 24, and 96 h (*, p Ͻ 0.001, Fig. 5A). In contrast, actinomycin D enhanced sICAM-1 protein release from HRV-infected cells at all experimental time points post-HRV inoculation. These observations indicate that HRV may down-regulate sICAM-1 release by inhibiting gene transcription of a suppressor/inhibitor of an enzyme responsible for the cleavage of mICAM-1 (*, p Ͻ 0.001, Fig. 5B).
Investigation of the Role of Tyrosine Kinases in HRV Induction of ICAM-1 Expression-Previous studies have demonstrated that the malarial parasite Plasmodium falciparum utilizes and up-regulates ICAM-1 expression via a tyrosine kinasedependent mechanism (26). We therefore sought to determine whether HRV-14 also acts on ICAM-1 expression via this mechanism. Phosphorylation of tyrosine kinase was assessed using Western analysis. HRV-14 infection of NHBE cells induced the de novo phosphorylation of a number of cellular substrates (Fig. 6A). Densitometry of three separate blots was performed, and the data are expressed as a percentage increase in tyrosine kinase phosphorylation in HRV-infected cells compared with Ͻ 0.001, Fig. 7). DISCUSSION In this present study, we have investigated the mechanisms driving the regulation of the major rhinovirus group receptor, ICAM-1 at the epithelial cell level during infection. As ICAM-1 can exist in two distinct forms, which appear to have opposing influences on host cell infectivity, we postulate that the dynamic inverse relationship between the membranous and soluble ICAM-1 receptor types is a critical catalyst in the pathogenesis and outcome of HRV infection.
We have demonstrated the presence of distinct mRNA transcripts coding for membrane-bound and soluble ICAM-1 in airway epithelial cells. We have also shown that these two receptors may be regulated independently. Consistent with previous studies (3,(12)(13)(14), our experiments demonstrated that rhinovirus infection of NHBE cells, increased the expression of mICAM-1 at both the protein and gene level peaking in this study at 8 and 24 h after virus inoculation, respectively, and thereafter, remained elevated above comparative uninfected control cells for up to 96 h after inoculation.
In addition, we explored effects of HRV on sICAM-1 gene and protein expression on the above cells. Rhinovirus appeared to down-regulate the gene expression of sICAM-1 throughout the study period reaching half-basal levels of expression by 24 h. This observed down-regulation of sICAM-1 mRNA is supported by the absence of detectable sICAM-1 in cell culture supernatants retrieved from HRV-infected cells. As virus and the antibody applied in the sICAM-1 ELISA utilize the same binding site on sICAM-1 molecules, the possibility of interference by HRV with the ELISA was eliminated by assaying controls "spiked" with recombinant sICAM-1 (data not shown).
To our knowledge, we are the first to report the simultaneous differential effect of HRV on both membrane and soluble forms of ICAM-1 receptor in an epithelial cell model, such that the virus appears to induce the membrane-bound form, while dramatically decreasing the soluble component; thereby facilitating and promoting cell infectivity. Of particular interest, we have conducted other studies, which show this same pattern of inverse relationship between mICAM-1 and sICAM-1 pattern of regulation following HRV inoculation of other epithelial cell lines, BEAS-2B and H292 (data not shown) (27). However, these studies also suggest that differences can occur in the time kinetics and magnitude of responses dependent on cell type used. Indeed, this is supported by the observations of Papi and Johnston (3), who observed significant magnitude differences in viral-induced ICAM-1 cell surface expression between primary bronchial epithelial cells and A549 (bronchial carcinomaderived) epithelial cells. Thus direct comparison of data between studies needs to take into account differences in cell origin, culture techniques as well as assay conditions used, as all these factors may account for observed differences in responsiveness to HRV infection.
Taken together, these results suggest that sICAM-1 may be detrimental to the virus as its competitive binding to available virus particles would facilitate a defensive role in limiting viral infection of target cells. Indeed, a recombinant form of sICAM-1 has been shown to have inhibitory effects on HRV infection in vitro (17). Studies in chimpanzees (18) and humans (19) have shown recombinant sICAM-1 to have some prophylatic effects with reduced severity of experimental rhinoviral colds. A recent in vivo study investigated mICAM-1 expression on nasal scrape biopsies and sICAM-1 levels in nasal lavage fluid from separate volunteer groups inoculated with experimental rhinovirus (28). mICAM-1 expression was increased in 87% of the volunteers following infection with rhinovirus; however, sICAM-1 levels in the nasal lavage fluid were only increased in 47% of volunteers. This differs from our current study; we found no detectable sICAM-1 in cell culture supernatants from HRV-infected cells. This apparent discrepancy in results may be caused by fundamental differences in the study design as our study is an in vitro study using cell lines. In addition, as less than half the volunteers exhibited an increase in sICAM-1 in the nasal lavage fluid, it is equally possible that the sICAM-1 measured may be due to mICAM-1 dislodged from the cell surface as a result of the sampling process and not through direct release. Time kinetics of ICAM-1 induction was consistent with our study in that up-regulation of ICAM-1 occurred within 24 h of infection and declined by day 5 (28).
Levels of infectious virus in cell culture supernatants from infected cells increased 8-h postinoculation and remained elevated for up to 96 h, suggesting an increase in viral replication over time. These data were confirmed by RT-PCR, in which a time-dependent increase in HRV RNA within NHBE cells was observed.
Blocking viral binding and subsequent viral internalization using monoclonal antibodies against ICAM-1 resulted in the inhibition of both the HRV-induced increase in mICAM-1 and down-regulation of sICAM-1 expression. In addition, viral titers in retrieved cell culture supernatants were significantly lower in cells pretreated with ICAM-1 monoclonal antibodies, suggesting a reduction of initial viral binding, entry, and subsequent infection. These data suggest that the observed HRV effects on ICAM-1 expression are due to virus specific-epithelial cell receptor interactions. This information may facilitate the design of potential small molecule therapeutic inhibitors targeting HRV-cell receptor interactions or subsequent intracellular events following viral binding and release of genetic material into the cell.
The HRV-induced increase in mICAM-1 expression was inhibited by cycloheximide indicating that HRV initiates de novo protein synthesis of ICAM-1. However, cycloheximide had no effect on sICAM-1 release from HRV-inoculated cells. Treatment of NHBE cells with actinomycin D also inhibited the HRV-induced increase in mICAM-1 suggesting HRV initiates transcription of the ICAM-1 gene. In contrast, actinomycin D increased the release of sICAM-1 from HRV-inoculated NHBE cells, suggesting HRV may increase transcription of a suppressor/accessory protein preventing the activity of a proteolytic enzyme involved in the cleavage of mICAM-1 from the cell surface. Possible candidate enzymes involved in the cleavage of mICAM-1 include the metalloproteinases (MMPs), which are tightly regulated by tissue inhibitors of metalloproteinases (TIMPs). In human gastric adenocarcinoma cells, Helicobateur pylori has been demonstrated to modulate MMP and TIMP secretion and that host MMP-3 and a TIMP-3 homolog expressed by H. pylori mediate at least in part of the host cell response to infection (29). A similar mechanism may operate in HRV-infected bronchial epithelial cells. In addition, a study utilizing astrocytes demonstrated that the mechanism involved in sICAM-1 release was sensitive to metalloproteinase inhibitors (30). Proteolytic cleavage of mICAM-1 has also been observed in keratinocytes, where the addition of protease inhibitors resulted in a dose-dependent inhibition of sICAM-1 production (31). It is therefore plausible that HRV may modulate the release of sICAM-1 by both down-regulating the gene expression of sICAM-1 and manipulating the potential enzymatic reactions involved in the cleavage of mICAM-1 from bronchial epithelial cells. Further studies need to identify enzymatic pathways responsible for the cleavage of mICAM-1 in bronchial epithelial cells.
HRV has been shown to inhibit nuclear import, thus preventing signal transduction into the nucleus (32). This could serve as one pathway utilized by HRV during the down-regulation of sICAM-1 expression. Alternatively, expression of sICAM-1 protein may be blocked at the level of translation. Previous studies have demonstrated that certain viruses may block translation of mRNA at the initiation step (33). Indeed, a study conducted by Svitkin et al. (34) demonstrated that HRV inhibits host cell protein synthesis by cleaving the eukaryotic initiation factors eIF4G11 and eIF4G1 resulting in a 60% decline in host protein synthesis by 6 h. This mechanism may contribute to the HRVinduced inhibition of sICAM-1 release as no sICAM-1 protein was detected at 0 h (8 h after inoculation with HRV-14). Translation termination factors may serve as a target for the virus. It has been demonstrated that certain isolated RNAs have an affinity for eukaryotic translation termination factors, eRF1, and eRF1⅐eRF3 complexes; to which they not only bind but also inhibit eFR1-mediated release of protein precursor chains from ribosomes (35).
Furthermore, HRV could modulate sICAM-1 secretion. Studies have demonstrated that poliovirus, also a member of the picornaviridae family, inhibits the transport of both plasma and secretory proteins from the endoplasmic reticulum to the Golgi apparatus early in the infection cycle (36 -37). It is therefore plausible that the above pathways, either solely or partly in combination could drive the observed HRV-induced downregulation in sICAM-1 release.
Since our studies have shown an HRV-induced increase in mICAM-1 expression at the transcriptional level, we proceeded to investigate the potential molecular mechanisms involved. Previous studies have shown that HRV induced up-regulation of mICAM-1 gene promoter activity involves initiation of NFB proteins binding to the NFB binding site on the ICAM-1 gene promoter region (3). Other studies have also indicated that tyrosine kinases may play a role in the regulation of the ICAM-1 gene (26, 38 -39). We have examined this intracellular pathway further and demonstrated that HRV initiates rapid onset of tyrosine phosphorylation of multiple substrates. There was strong phosphorylation of two substrates, 85 and 200 kDa, 5 and 15 min post-HRV inoculation. This response was totally inhibited by genistein at concentrations of 50 and 100 M. Kelley and Drumm (40) also demonstrated that ICAM-1 expression was mediated through tyrosine kinases of 85 and 154 kDa in endotheial cells and showed that ICAM-1 expression was completely inhibited with genistein. Studies investigating tyrosine phosphorylation events during Coxsackie virus, also a member of the Picornaviridae family have also demonstrated an increase in tyrosine phosphorylation of a 200-kDa protein (41). Huber et al. (41) concluded that this protein was of cellular origin and may play a critical role in effective viral replication. In addition, tumor necrosis factor (TNF)-induced ICAM-1 expression has been demonstrated to involve the tyrosine phosphorylation of an 85-kDa protein, which was thought to be a cytoskeletal protein (41). Further studies utilizing more specific inhibitors for example herbimycin A, a selective inhibitor of Src-like kinases or tyrphostin, an inhibitor of Janus kinase (JAK) are required to identify these proteins. In addition, specific monoclonal antibodies to proteins of a similar molecular weight may be utilized to identify these substrates. In this study, genistein (100 M) significantly inhibited the HRV-induced up-regulation in mICAM-1 expression at 8 h with levels decreasing to basal levels of expression by 96 h. These data support the hypothesis that HRV may modulate mICAM-1 expression through a tyrosine kinase-dependent signaling pathway.
In conclusion, we have demonstrated that HRV manipulates the expression of both ICAM-1 receptors in airway epithelial cells to promote and sustain infection. We have attempted to elucidate the complex molecular mechanisms of ICAM-1 regulation and identified protein tyrosine kinases as critical components. It is plausible that detailed dissection of the molecular driving forces involved in coordinating the inverse relationship between membranous and soluble ICAM-1 receptors in the context of HRV-epithelial cell membrane interaction may lead to the development of novel anti-HRV therapeutic strategies. | 7,389.2 | 2003-04-04T00:00:00.000 | [
"Biology",
"Medicine"
] |
Integrated profiling identifies DXS253E as a potential prognostic marker in colorectal cancer
Background Increasing evidence suggests that DXS253E is critical for cancer development and progression, but the function and potential mechanism of DXS253E in colorectal cancer (CRC) remain largely unknown. In this study, we evaluated the clinical significance and explored the underlying mechanism of DXS253E in CRC. Methods DXS253E expression in cancer tissues was investigated using the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. The Kaplan-Meier plot was used to assess the prognosis of DXS253E. The cBioPortal, MethSurv, and Tumor Immune Estimation Resource (TIMER) databases were employed to analyze the mutation profile, methylation, and immune infiltration associated with DXS253E. The biological functions of DXS253E in CRC cells were determined by CCK-8 assay, plate cloning assay, Transwell assay, flow cytometry, lactate assay, western blot, and qRT-PCR. Results DXS253E was upregulated in CRC tissues and high DXS253E expression levels were correlated with poor survival in CRC patients. Our bioinformatics analyses showed that high DXS253E gene methylation levels were associated with the favorable prognosis of CRC patients. Furthermore, DXS253E levels were linked to the expression levels of several immunomodulatory genes and an abundance of immune cells. Mechanistically, the overexpression of DXS253E enhanced proliferation, migration, invasion, and the aerobic glycolysis of CRC cells through the AKT/mTOR pathway. Conclusions We demonstrated that DXS253E functions as a potential role in CRC progression and may serve as an indicator of outcomes and a therapeutic target for regulating the AKT/mTOR pathway in CRC. Supplementary Information The online version contains supplementary material available at 10.1186/s12935-024-03403-4.
Introduction
Colorectal cancer (CRC) remains the major cause of tumor-related mortality in the world [1,2].Although significant progress in treatment has reduced recurrence and improved patient survival rates, the prognosis of CRC patients remains largely unknown.Therefore, it is necessary to gain a better understanding of oncogenic genes involved in CRC to develop therapeutic strategies for improving the long-term outcomes of CRC patients.
The TCGA and GEO databases are currently the most widely used public resource centers for oncology-related research [3][4][5].These databases not only cover gene expression, methylation status, noncoding RNA, and other data, but, more importantly, contain clinical data and dynamically updated survival data.Bioinformatics analyses of TCGA and GEO data in cancer basic and clinical studies have facilitated the identification of numerous tumor markers, with some of these markers already being tested in clinical trials.Based on TCGA, Jin et al. identified a previously uncharacterized immunosuppressive tumor necrosis factor ligand molecule named CD70 that may serve as a possible chimeric antigen receptor target for immunotherapy in gliomas [6].Additionally, another study listed in the TCGA's database suggested that amplification of mouse double minute 4 (MDM4) is an important genetic change in the development of hepatocellular carcinoma (HCC), and a drug targeting this amplification has been undergoing clinical trials [7].
Moreover, previous studies have revealed the pivotal role of DXS253E in a series of cancers.Wang et al. found that DXS253E influences the chemotherapy resistance in polyploid cancer cells [21].Recent studies have indicated that DXS253E is associated with the infiltration of immune cells in liver cancer [22].Similarly, Wang et al. analyzed public databases and identified that DXS253E influences the immune microenvironment and is related to poor prognosis in patients with CRC [23].However, the exact role of DXS253E in CRC remains to be further determined through the specific experiments.
In this study we evaluated the clinical significance and effects of DXS253E in CRC through TCGA and GEO, and with an experimental cohort.We found that high levels of DXS253E correlate with poor prognosis in CRC patients.Our bioinformatics analyses also show that the methylation of DXS253E is associated with the prognosis of CRC patients and that DXS253E levels correlate with immune cell infiltration.Moreover, we, for the first time, explored the function of DXS253E in vitro using cell culture experiments.We found that DXS253E enhances the malignant biological behavior and aerobic glycolysis through the AKT/mTOR pathway in CRC.Systematically investigating the role of DXS253E may facilitate the identification of novel therapeutic targets for CRC.
DXS253E expression analysis
DXS253E mRNA expression levels were determined for 33 human cancers including CRC tissues based on TCGA and the GEO databases.The R programming language (version 3.6.3)was used to download, clean, and visualize the data.In detail, the 'TCGAbiolinks' (version 2.28.4) and 'GEOquery' (version 2.72.0)R packages were used to download RNA-sequencing and clinical data from TCGA and the GEO databases, respectively [24,25].Subsequently, all of the cancer RNA-sequencing data were merged into a single meta-cohort via ComBat algorithm from the 'sva' R package (version 3.48.0) to eliminate batch effects [26].Following the combination of 33 datasets, we normalized the raw data using a log2(FPKM + 1) transformation.Probes were converted into corresponding gene symbols based on a GPL570 annotation file.Finally, DXS253E mRNA expression levels were filtered using the R package 'tidyverse' (version 2.0.0) and plotted using the 'ggplot2' R package (version 3.4.4)[27].
Patients and tissue specimens
The Department of Gastrointestinal Surgery IV, Peking University Cancer Hospital & Institute, provided eight matched sets of CRC and normal tissues from patients who had surgical resection between September 2009 and October 2011.
Immunohistochemistry
Samples of CRC patients and nearby healthy tissues were preserved in 4% paraformaldehyde (PFA), embedded in paraffin, cut into 4 μm slices, and then heated at 65 °C for two hours.After heat-induced epitope retrieval, the slides were incubated with DXS253E antibody (1:100, Cat #19,909-1-AP, Proteintech) overnight at 4 °C, and subsequently exposed to an anti-rabbit antibody for 40 min and stained with fresh 3,3′-Diaminobenzidine substrate within a controlled reaction time of 1-2 min.Next, sections were counterstained with hematoxylin, rinsed to blue, dehydrated, and sealed.
Survival analysis
The association between DXS253E expression and the prognosis of CRC patients was evaluated by Kaplan-Meier analysis with the log-rank test.The results were visualized using the R survminer package.The effect of clinical variables on patient outcome was assessed using univariate and multivariate Cox regression analyses.Tumor stage correlation was analyzed using the R ggplot2 and stats packages.
DNA methylation analysis
The status of DXS253E methylation in CRC was investigated using the MethSurv web-based tool (https://biit.cs.ut.ee/methsurv/).Moreover, the prognostic value of DXS253E methylation levels in CRC patients was evaluated using MethSurv and its built-in database.
Differentially expressed gene analysis
The Differentially expressed gene analysis (DEGs) between DXS253E high-and DXS253E low-expression groups were identified using the R Bioconductor package DESeq2 (https://www.bioconductor.org/).Then, the DEGs were drawn as volcano maps using the R ggplot2 package.
Gene function enrichment analysis
The R Bioconductor clusterProfiler package was used to evaluate our Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses and GSEA with the identified DEGs.An adjusted P value < 0.05 was regarded as statistically significant.
Immune correlation analysis
The ssGSEA method provided by the GSVA package was used to determine the infiltration degree of immune cells with notable DXS253E expression levels [28].Given that the expression matrix is in the format of FPKM (Fragments Per Kilobase of transcript per Million mapped reads), the kcdf parameter was set to "Gaussian" [28].We downloaded markers for 24 types of immune cells following the Bindea et al. protocol [29].Subsequently, a Pearson's correlation analysis was conducted to investigate the relationship between DXS253E expression levels and the infiltration of immune cells based on a previous study [30].An absolute correlation coefficient value greater than 0.3, accompanied by a P value less than 0.05, was considered significant for these correlations [31].Additionally, the association between DXS253E and immune-related genes was examined using the Sangerbox toolkit (http://www.sangerbox.com/)[32].
Cell lines and cell culture
All human CRC cell lines (LoVo, HCT116, RKO, SW480, and SW620) and a healthy human intestinal epithelial cell line (NCM460) were obtained from American Type Culture Collection.Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) high glucose medium (HyClone, Logan, UT, USA), which was supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C with 5% CO 2 .
Quantitative real-time PCR
Total RNA from CRC tissues and cells was isolated using Trizol (Invitrogen, Waltham, MA, USA) and then reversely transcribed into cDNA with a reverse transcription kit (Promega, Madison, WI, USA).Thereafter, qRT-PCR was carried out using a SYBR-Green Master kit (Cat# 147,100, TOYOBO, Japan).The primers used for amplification are presented in Table S2.
CCK-8 and colony formation assay
CRC cell proliferation was assessed using CCK-8 assays (Dojindo, Japan).The cells were inoculated into 96-well plates at a density of 5,000 cells per well with complete medium.After being cultured for the indicated time, 10 µL of CCK-8 solution was added to each well and incubated for an additional 2 h at 37 °C.Spectrometric absorbance values were then measured at 450 nm.To assess colony formation ability, the cells were plated at a density of 500 cells per well and incubated in 6-well plates for 12 days.Colonies were stabilized with 4% PFA, and stained with 0.1% crystal violet.
Transwell assays
Cell migration and invasion assays were performed using 24-well Corning® Costar® Transwell chambers (NY, USA).The cells were resuspended in serum-free medium and inoculated into the Transwell chambers with or without Corning® Matrigel and then we conducted migration or invasion assays, respectively.Complete medium was added to the lower chambers.After incubation for 24 h, migrated cells were treated with 4% PFA and stained with 0.1% crystal violet.Images were taken under a microscope and the number of migrating cells was counted.
Lactate production and detection of reactive oxygen species (ROS)
Cells were transfected with plasmids containing DXS253E and cultured for 48 h in complete medium.The culture medium was harvested to determine lactate production with a lactic acid assay kit (Nanjing Jiancheng, China).Cells were incubated with 2 µM diacetyldichlorofluorescein (Solarbio, China) at 37 °C for 30 min.After incubation, cells were washed with 1 ml of phosphate buffered saline three times.Fluorescence intensity of the cells was then recorded using a flow cytometer.
Statistical analysis
All statistical analyses were conducted using R (version 3.6.3).DXS253E mRNA expression levels across various cancers were calculated using a Wilcoxon rank-sum test based on TCGA and GEO databases.Student's t tests were conducted to determine mRNA expression differences between the CRC and normal samples in our cohort.The correlation between clinicopathologic features and the level of DXS253E was analyzed with a Wilcoxon rank-sum test and logistic regression.For in vitro experiments, comparisons between groups were detected using Student's t tests.
DXS253E is highly expressed in a series of cancers including CRC
A flowchart containing the detailed procedures of this study is presented in Fig. 1.First, our pan-cancer analyses based on TCGA's database suggest that high DXS253E expression is involved in a series of cancers, such as breast invasive carcinoma (BRCA) and cholangiocarcinoma (CHOL) (Fig. 2A, B).The expression level of DXS253E is notably higher in colon adenocarcinoma (COAD) and rectum adenocarcinoma (READ) samples than in normal tissues (Fig. 2C, D).Moreover, we identified high DXS253E expression in CRC tissues from the GEO datasets GSE9348 and GSE23878 compared with corresponding controls (Fig. 2E).To verify our database results, we determined DXS253E expression in CRC and adjacent normal tissues using quantitative real-time PCR (qRT-PCR) and immunohistochemistry from our CRC cohort.Our results show a markedly increased expression of DXS253E in CRC tissues compared with adjacent normal tissues (Fig. 2F, G).
High DXS253E expression indicates worse prognosis for CRC patients
We evaluated the mRNA levels of DXS253E in different clinical categories to determine the correlation between DXS253E expression and clinicopathological features in CRC patients within TCGA's database.Our findings indicated that the expression of DXS253E is significantly associated with N stage (P = 0.003), M stage (P = 0.002), pathologic stage (P = 0.001), perineural invasion (P = 0.039), lymphatic invasion (P < 0.001), neoplasm type (P = 0.038), overall survival (OS) (P = 0.010), and disease-specific survival (DSS) events (P = 0.040) in CRC patients (Table S3, Fig. 3A).However, no significant correlations were observed between DXS253E and other clinicopathological variables, such as T stage, gender, age, and race (Table S3).Collectively, these data suggest that high DXS253E expression is related to more lymph node metastases, more perineural invasion, higher pathologic, N, and M stage occurrences, and poor prognosis in CRC patients.
Kaplan-Meier plots were used to analyze the relationship between DXS253E expression and prognosis of patients with CRC in the TCGA's database.We found that higher DXS253E expression has a significant association with poor OS in COAD (P = 0.004) and READ (P = 0.022) (Fig. 3B).Similarly, DSS analysis data showed that higher DXS253E expression also correlated with poor prognosis in COAD (P = 0.009) and READ (P = 0.024) (Fig. 3C).
Finally, we conducted a multivariate Cox regression analysis using the significant factors identified in our univariate analysis (Fig. 3E).Our results demonstrate that advanced pathologic stage (P = 0.005) and advanced M stage (P < 0.001) are independent prognostic variables of poor prognosis for OS and DSS, respectively.
Genetic and epigenetic alterations of the DXS253E gene in CRC
To explore mutations in the DXS253E gene in a series of cancers, we analyzed the mutation status of SCL10A3 using the Web cBioPortal platform (https://www.cbioportal.org/)with TCGA's pan-cancer database.The highest alteration frequency of the DXS253E gene (10.42%) was identified in diffuse large B cell lymphoma patients with amplification and deep deletion as the primary alterations (Fig. 4A).Our analysis results showed a DXS253E gene mutation frequency of 0.84%, and an amplification frequency of 0.51% for DXS253E gene in CRC (Fig. 4A).
Additionally, we performed exon missense mutation analysis in high and low DXS253E expression groups using CRC TCGA data.The top 15 notably different somatic mutations occurred in the following proteins: TP53, ZFHX4, FAT3, DNAH11, RYR1, MUC5B, BRAF, ADGRV1, KMT2B, KMT2D, HMCN1, DCHS2, UNC80, DMD, and COL6A3 (Fig. 4B).Our results uncovered a remarkably high mutation frequency for the TP53 gene in the DXS253E high-expression group.These mutated genes may prove valuable for determining tumor progression and therapeutic response in CRC patients.
DNA methylation was investigated to determine epigenetic effects, according to TCGA's database.DNA methylation levels of the gene encoding DXS253E and the prognostic value of the CpG islands in this gene were determined using the web's MetSurv tool.Twelve methylated CpG islands were identified as being correlated with levels of DXS253E gene methylation, in particular cg09418475 and cg05424879 (Fig. 4C).Furthermore, Kaplan-Meier survival analysis suggested that two CpG sites (cg09418475, P = 0.039 and cg06266461, P = 0.048) exhibited high methylation levels that were significantly correlated with favorable OS of CRC.We observed that following four other DXS253E CpG sites were borderline significantly related to the OS of CRC patients: cg05424879 (P = 0.082), cg06616857 (P = 0.084), cg11667509 (P = 0.057), and cg24546622 (P = 0.07) (Fig. 4D).
We next used the LinkedOmics web portal (https:// www.linkedomics.org/)to identify genes correlated with DXS253E in CRC.Pearson correlation analysis revealed that a total of 1,919 co-expressed genes were remarkably associated with DXS253E in CRC (FDR < 0.05, P < 0.05, and |cor.|≥ 0.2) (Fig. 5F, Table S8).The top 50 positively (r > 0) and negatively (r < 0) correlated genes that we found were included in the heat maps of this cohort (Fig. 5G).
Next, the 5,761 DXS253E-associated DEGs from TCGA's database that we identified and the 1,919 significantly co-expressed genes found with LinkedOmics were selected to determine how many occurred in both dataset results.Among them, 53 genes overlapped and these were selected for further functional analyses (Fig. 5H, Table S9).We then performed a combined GO/KEGG analysis through the Metascape tool (https://metascape.org/) to explore biological functions of the 53 overlapping genes.This enrichment analysis suggested that the following three functions were most significantly enriched in the 53 genes: vitamin metabolic processes, response to salt, and calcium-independent cell-cell adhesion via plasma membrane cell-adhesion molecules (Fig. 5I, Table S10).
DXS253E expression is associated with immune-related genes and immune cell infiltration
Given the known essential influence of the tumor microenvironment (TME) on patient prognosis and treatment decisions [33,34], we investigated the association between DXS253E and the TME.First, in pan-cancer dataset, we determined the correlation of DXS253E with immunomodulatory genes which encoded chemokines, receptors, MHC, immuno-inhibitors, and immunostimulators.Our results showed that DXS253E was significantly correlated with various immune-associated genes in multiple cancers, including COAD and READ (Fig. 6A).
Immune cells play an important role in orchestrating the TME.Increasing evidence shows that single-cell RNA sequencing (scRNA-seq) is very useful for analyzing the TME and immune cell infiltration [35,36].In this study, we found five independent CRC scRNA-seq entries in the Tumor Immune Single-cell Hub database (http://tisch.comp-genomics.org/home/)at the time of our research (CRC_GSE108989, CRC_GSE136394, CRC_ GSE139555, CRC_GSE166555, and CRC_GSE179784).We then explored DXS253E expression levels at the single-cell level in various cells of the TME.Our findings indicated that DXS253E was highly expressed in immune and endothelial cells (Fig. 6B).Specifically, in the CRC_ GSE108989, higher levels of DXS253E expression were found in Tprolif cells.In the CRC_GSE136394 dataset, DXS253E was mainly distributed in conventional CD4Tconv and Tprolif cells (Fig. 6C).
Tumors can modulate the immune cells infiltration and immune response through specific molecules in the TME [37].Therefore, we aimed to investigate whether the DXS253E level in tumors could impact the recruitment of immune cells.Our analysis using TCGA datasets demonstrated that DXS253E was significantly linked to the infiltration of multiple immune cells, including NK, NK CD56bright, eosinophil, Tcm (a subgroup of memory CD8 + T cells), T helper, and Th2 cells (a subgroup of CD4 + T cells) (all P < 0.05, |R| ≥ 0.3, Fig. 6D) cells.Our scatter plot illustrates that the expression of DXS253E was positively connected with NK, NK CD56bright, and eosinophil cells (P < 0.001), but was negatively connected with Tcm, T helper, and Th2 cells (P < 0.001) (Fig. 6E).Similarly, NK, NK CD56bright, eosinophil, Tcm, T helper, and Th2 cells showed differential infiltration when samples were subdivided according to DXS253E expression levels in CRC (Fig. 6F).Thus, DXS253E expression levels are significantly associated with immune-related genes and immune cell infiltration in CRC.
DXS253E overexpression promotes malignant biological behavior of and aerobic glycolysis in CRC cells
Because DXS253E expression was elevated in the tissue of CRC patient and was linked to the poor prognosis, we explored the biological role of DXS253E in CRC.First, we examined DXS253E expression levels in different cell lines.Our western blot analyses of a normal human colon mucosal epithelial cell line (NCM460) and CRC cell lines (LoVo, RKO, HCT116, SW480, and SW620) suggested that DXS253E is highly expressed in CRC cell lines (Fig. 7A).Subsequently, the mRNA levels encoding DXS253E were obtained through qRT-PCR (Fig. 7B).We found that high DXS253E mRNA and protein levels were both positively associated with NCM460, RKO, HCT116, SW480, and SW620 cells, but not with LoVo cells.The results of a CCK-8 assay and plate cloning assay showed that DXS253E overexpression markedly enhances the proliferative ability of CRC cells in RKO and HCT116 compared with the control group (Fig. 7C, D).In addition, we used a Transwell assay to assess migration and invasion ability in RKO and HCT116.These results suggested that the migration and invasion ability of CRC cells was significantly increased when DXS253E was overexpressed (Fig. 7E).Collectively, these findings show that DXS253E can significantly enhance the malignant progression of CRC cells.
Our GSEA findings suggested that DXS253E could regulate the processes of oxidative stress in CRC.To determine whether DXS253E does regulate oxidative stress processes in CRC cells, we detected intracellular reactive oxygen species (ROS) levels using dichlorodihydrofluorescein diacetate probes.Flow cytometry results from this experiment suggested that overexpression of DXS253E decreases ROS production in CRC cells (Fig. 7F).Moreover, due to associations between the processes of aerobic glycolysis and oxidative stress, lactate production assays were performed to determine the role of DXS253E in aerobic glycolysis.Our results indicated that DXS253E overexpression significantly increases lactate production in CRC cells (Fig. 7G).Furthermore, qRT-PCR and western blot procedures were performed to assess the expression of glycolytic genes.These results revealed that the overexpression of DXS253E remarkably increases the expression of several glycolytic genes, in particular HK2, PKM2, GLUT1, and LDHA, at both the mRNA and protein level in CRC cells (Fig. 7H, I).
Increasing evidence has demonstrated that aerobic glycolysis in tumor cells is associated with the AKT/mTOR pathway [38].Therefore, we conducted western blot analysis to evaluate the role of DXS253E in the AKT/mTOR pathway.Our results showed that DXS253E overexpression promoted the phosphorylation of AKT and mTOR in RKO and HCT116 cells (Fig. 7J).These findings raise the possibility that DXS253E increases aerobic glycolysis through the activation of the AKT/mTOR pathway in CRC cells.
Discussion
CRC is a heterogeneous malignant tumor that exhibits a complex microenvironment [39].This complicates efforts to understand how different genes and cell types promote disease progression.In this study, based primarily on TCGA and the GEO database, we identified that high DXS253E levels predict poor prognosis in patients with CRC.Bioinformatics analyses indicated that DXS253E is associated with genetic change and immune infiltration.Functionally, DXS253E enhances malignant phenotypes and aerobic glycolysis in CRC cells through the AKT/mTOR pathway, according to our research.Previous studies have demonstrated the clinical and prognostic significance of DXS253E in various tumors.In liver cancer, high DXS253E protein levels were observed to be significantly associated with immune cell infiltration and poor prognosis [22].In terms of LGG, higher DXS253E expression levels contribute to programmed cell death and immune infiltration, resulting in poor prognosis of the patients with LGG [40].Consistently, Ma et al. also found that the upregulation of DXS253E is associated with poor OS in LGG and glioblastoma [41].Moreover, they suggested that SLC10A3 is a possible therapeutic target for LGG [41].Furthermore, Wang et al. 's study that analyzed public databases indicates that a high DXS253E expression predicts poor prognosis in CRC patients, and they propose that it could be used as an immunotherapy target in CRC given its significant influence on the immune microenvironment [23].Similarly, we found that CRC patients with higher DXS253E tumor expression exhibit worse outcomes in respect to both OS and DSS, suggesting that DXS253E can serve as a predictor of CRC prognosis.Therefore, DXS253E may be considered a cancer-promoting gene and could be used as a novel prognostic biomarker as well as a potential therapeutic target for CRC.
Previous studies have implicated that somatic mutations and high methylation levels within a gene are correlated with the regulation of gene expression [42,43].
Local DNA hypermethylation analyses have potential as a diagnostic and outcome prediction tool for some cancers [44,45].Thus, we analyzed genetic and epigenetic alterations of the DXS253E gene in CRC.Our exploratory findings demonstrated that DXS253E genetic alterations, including mutation and amplification, can be observed in a variety of cancers.Furthermore, significant differences existed in the DXS253E gene DNA methylation levels between tumor tissues and normal tissues.Although the role of gene body methylation and the methylation of CpG sites in both introns and exons remains less characterized, we identified six CpG sites in the DXS253E gene that were correlated with the prognosis of CRC patients.Additionally, due to the reversibility of DNA methylation, hypermethylated genes could prove to be targets for the treatment of cancers.
We also investigated the DEGs between DXS253E low-and DXS253E high-expression groups in CRC to elucidate the potential biological functions and regulatory pathways of DXS253E.A previous study using GO/ KEGG enrichment analysis demonstrated that DXS253Ecorrelated genes were related to substance transport in LGG [41].However, we identified that DXS253E affected the transcriptome and oxidative phosphorylation metabolic pathway in CRC by conducting enrichment analyses of the GO and KEGG databases and GSEA.The difference in enrichment analyses between LGG and CRC may be due to the heterogeneity of the tumors.Furthermore, we explored genes encoding DXS253Erelated proteins and co-expressed genes in CRC tissues to identify DXS253E molecular mechanisms contributing to CRC prognosis.Consequently, we defined 53 overlapping genes between TCGA and LinkedOmics as CRC coexpressed hub genes and DEGs.Our enrichment pathway analysis showed that these hub genes were involved in vitamin metabolic processes and response to salt.Overall, these data raise the possibility that DXS253E may participate in cellular metabolism and intracellular transport and may facilitate the proliferation and metastasis of CRC by adjusting these biological processes.
Cancer progression is related to the TME, which is composed of tumor cells, mesenchymal cells, the extracellular matrix, immune cells.The infiltration of immune cells into the TME influences tumor prognosis [46].Tian et al. confirmed significant associations between DXS253E expression and CD4 + Tconv and tumor-infiltrating CD20 + B cells via bioinformatics analysis and multispectral imaging techniques in HCC [22].Similarly, a LGG study showed that DXS253E expression was associated with macrophage, CD4 + Tconv cell, and B cell levels [40].We consistently found that DXS253E levels were associated with the level of multiple immune-related genes and immune cell infiltration in CRC.DXS253E was significantly connected with the infiltration of NK, NK CD56bright, eosinophil, Tcm, T helper, and Th2 cells, any of which could be used as a clinical outcome indicator in CRC patients.DXS253E has dramatic effect on the partial infiltrations of immune cells in CRC tissues.
We also explored the impact of DXS253E on cell phenotype and mechanism, and identified that DXS253E overexpression can remarkably enhance the proliferation, migration, and invasion capacity of CRC cells.Our results indicate that the specific influence of DXS253E on the malignant proliferation of tumor cells may occur to some degree through effects on aerobic glycolysis via the AKT/mTOR pathway.
Although our current study provides new understanding about the relations between DXS253E expression and prognostic value in CRC patients, some limitations need further consideration.Primarily, this study used opensource datasets, which can lead to selection bias.Other independent large-scale CRC cohorts with detailed clinicopathologic characteristics will be required to further validate our findings and the application of DXS253E as a biomarker in CRC should be evaluated.Additionally, although DXS253E may offer an alternative opportunity to delay the progression of CRC, further investigation of its functions in vitro and in vivo is necessary to identify the molecular mechanism underlying DXS253E regulation of the tumor microenvironment, as well as the potential of AKT/mTOR-targeted treatment for CRC patients with high DXS253E expression.
Conclusions
This study demonstrates that high levels of DXS253E are an independent unfavorable prognostic indicator in CRC and are notably correlated with aggressive clinical characteristics.High DXS253E levels facilitate malignant biological behavior and aerobic glycolysis in CRC cells though AKT/mTOR signaling.Our study identifies DXS253E as a potentially cancer-promoting gene and suggests its potential application as a therapeutic target for regulating the AKT/mTOR pathway in CRC.
Fig. 1
Fig.1The analysis flowchart of this study
Fig. 4
Fig. 4 Mutation and methylation analysis of DXS253E in CRC.A Alteration frequency of the DXS253E gene across various cancers analyzed using the cBioPortal web resource.B Differential somatic mutations identified in CRC between low and high DXS253E expression groups.C Correlation between DXS253E mRNA expression level and methylation level.D Kaplan-Meier survival curves showing six methylation sites in the DXS253E gene
Fig. 5
Fig. 5 Functional enrichment analysis of differentially expressed genes (DEGs) according to DXS253E expression level in CRC.A Volcano plot for DEGs between low DXS253E and high DXS253E expression groups.B Heatmap showing the top 10 DEGs between low and high DXS253E-expression groups.C GO enrichment analysis of DXS253E-associated DEGs.D KEGG enrichment analysis of DXS253E-associated DEGs.E GSEA of relevant signaling pathways in CRC tissues based on DXS253E-related DEGs.F Volcano plot of co-expressed genes correlated with DXS253E expression using the LinkedOmics web resource.G Heatmaps of the top 50 genes that are positively or negatively associated with DXS253E.H Venn diagram of the number of intersections between DXS253E DEGs and co-expressed genes in CRC.I Enrichment analysis of overlapping genes analyzed by the Metascape web resource
Fig. 6
Fig. 6 DXS253E expression is associated with multiple immune-related genes, and immune cell infiltration occurs in CRC.A Correlation of DXS253E with immunomodulatory genes in pan-cancer.B Heatmap of DXS253E expression with various tumor microenvironment cells in five independent datasets from the Tumor Immune Single-cell Hub (TISCH) database.C DXS253E expression in immune cells according to the Gene Expression Omnibus (GEO) GSE108989 and GSE136394 datasets.D Correlation between DXS253E and immune cell infiltration.E, F Relationship of DXS253E expression with the infiltration level of NK, NK CD56bright, eosinophil, Tcm, T helper, and Th2 cells using scatter plots (E) and box plots (F).*P < 0.05, **P < 0.01, ***P < 0.001
Fig. 7
Fig. 7 DXS253E facilitates CRC cell malignant phenotype and aerobic glycolysis via the AKT/mTOR pathway.A Western blot assay of DXS253E protein expression levels in NCM460, LoVo, HCT116, RKO, SW480, and SW620 cell lines.B qRT-PCR assay of DXS253E mRNA expression levels in normal epithelial colon cell line NCM460 and CRC cell lines.C DXS253E overexpression accelerates the proliferation of RKO and HCT116 cells.D High levels of DXS253E expression enhance colony formation in CRC cells.E Overexpression of DXS253E enhances the malignant progression of RKO and HCT116 cells.Bar graphs show the number of migrated or invaded cells.F High levels of DXS253E decrease the generation of reactive oxygen species (ROS).G DXS253E overexpression elevates lactate production.H, I DXS253E regulates the level of glycolytic genes in CRC cells, including HK2, PKM2, GLUT1, and LDHA with qRT-PCR and western blot.J DXS253E overexpression in CRC cells mediates the activation of the AKT/mTOR pathway.*P < 0.05, **P < 0.01, ***P < 0.001 Table S1 contains detailed clinical information.Written informed consent was taken from each patient before sample collection.The Research Ethics Committee of Peking University Cancer Hospital & Institute approved and supervised this study (2021KT134).
Table 1
Univariable and multivariable analysis for OS in CRC patients with TCGA cohort HR: hazard ratio; CI: confidence interval.P-values in bold were statistically significant
Table 2
Univariable and multivariable analysis for DSS in CRC patients with TCGA cohort HR: hazard ratio; CI: confidence interval.P-values in bold were statistically significant | 6,767.6 | 2024-06-18T00:00:00.000 | [
"Medicine",
"Biology"
] |
Modified Naiver-Stokes equation for conceptual tests of pure field physics
Cartesian relativistic physics has its own nondual analog of the 1915 Einstein Equation for pure field physics in nonempty space. This tensor field analog leads to the vector geodesic equations for relativistic accelerations of Ricci material densities. Extended states of inertial energy densities modify the Navier-Stokes equation by the kinematic ‘living forces’ for the slow motion of material media. The new inertial feedback enables a conceptual choice between Newtonian and Cartesian alternatives (with localized or extended, respectively, elementary masses in the Universe) because of different pressure and temperature gradients across laboratory flows of liquids and gases.
Introduction
Cartesian physics [1,2] of material space is currently described only in qualitative terms and without quantitative predictions of specific phenomena for the theory verification/falsification procedure. The concept of void space regions without matter was unclear to Descartes, who after Aristotle also maintained the extension of matter or 'matter-extension' as the continuous material plenum. The Cartesian vortex mechanics for all types of observable spatial displacements in this material plenum was published in 1629. Later Newton's successful dynamics of point-like masses shook the need in sophisticated matter-extensions with vortex states. Nowadays, Newton's mechanics of localized masses and continuous gravitation fields in empty space forms the dual core of contemporary space theories in the Solar system and beyond. Not Cartesian physics, but Newtonian empty space modeling became mandatory for 1916 Einstein's gravitation in the weak field limit. No one looked at mechanical theories if they did not fit Newton's mass transport at low speeds.
Dual classical physics for spatially separated matter and fields is traditionally assigned to the macroscopic world, while physicists tend to assign nondual material densities (or quantum fields) only to the microscopic world. But the unique physical reality is either dual or nondual regardless of the spatial scaling and mathematical formalisms in available theories. Similarly, the matter is either local or non-local regardless of suitable approaches to describe it. The celebrated Einstein-Podolsky-Rosen paradox initiated the long-term discussion and many tests of the material world nonlocality. Now nonlocality of matter is reasonably considered beyond quantum physics [3]. The world holism [4] and the global direct overlap of all material elements motivates our attempts to develop non-Newtonian mechanics for mass densities of the extended particle. There is no in reality a mesoscopic a e-mail<EMAIL_ADDRESS>EPJ Web of Conferences 182, 02022 (2018) https://doi.org/10.1051/epjconf/201818202022 ICNFP 2017 scale for possible transitions from nondual microscopic physics to the dual macroscopic model of Newton. Material states should initially be considered in nondual field terms on micro-, macro-, and mega-scales in qualitative, if not quantitative, approaches to any energy flows in the Universe.
The purpose of this paper is to trace the balanced energy origin of extended sources in Einstein's metric gravitation, to study the strong field compensation for positive kinetic and negative potential self-energies of equal inertial and gravitational masses, to propose a nondual field analogue of the Einstein Equation, and to predict new laboratory phenomena for conceptual tests of extended matter physics. Based on propositions of Aristotle, Descartes, Mie [5] and Einstein [6], we shall try to study how to redesign classical relativistic gravitation in pure field terms -heavy material fields without localized particles. Our initial idea is that 1938 Einstein-Infeld physics may coherently replace the supposed empty space (between the localized charges) with the Aristotle space plenum of moving material fields. This may assist to give up the unnecessary concept of Newton's point-like particles and to modify the old building of classical gravitation by strong field solutions without energy divergence and metric singularities in the center of gravitational sources.
The Einstein Equation in the dual (field + matter) Newtonian paradigm has been known since 1915. The similar tensor equation in the Cartesian paradigm of the united space-matter operates with a nondual field continuum of extended mass-energies. This continuum is filled by overlapping radial vortexes with chaotic auto-rotations of elementary metric densities. Such permanent metric motions within the elementary Cartesian vortex reveal the kinematic origin of the rest-mass energy mc 2 . This active, kinetic energy may be called the internal relativistic heat. The positive (kinetic, active) energy is always accompanied by equal, but negative (potential, passive) self-energy according to the principle of energy self-compensation (introduced below). There is no elementary inertial mass without vortex auto-rotations of metric space. And circular material densities with positive kinetic energy generate self-gravitation with negative (passive) elementary energies of Cartesian vortex states that results in the global compensation of active and passive energies in the non-empty space Universe.
We will return to the Einstein idea that the geodesic equations of motion should follow from the strong-filed gravitation equation. This was difficult to prove from iterations in the dual post-Newtonian physics. But Cartesian geodesic relations for Ricci material densities can be derived exactly from the nondual field analogue of the Einstein Equation. In order to demonstrate some practical benefits of non-empty space physics, we will modify the Navier-Stokes equation in terms of Ricci energy flows with the inertial feedback of 'living forces', −µ∂ i V 2 /2. This dynamical pressure averts asymptotically divergent energy flows and can initiate conceptual tests of Newton vs Cartesian mechanics in the laboratory.
1938 Einstein's material fields for 1629 Cartesian vortex mechanics
The predominant majority of people believe that electric charges and inertial masses are located within visible frames of macroscopic bodies. Such bodies can be divided into smaller parts in accordance with everyday observations. Thus, the smallest part of substance was introduced as a so-called corpuscle (the point particle in the indivisible limit) that is a carrier of the elementary mass and charge. Classical fields between localized particles are considered massless and chargeless. Nevertheless, the experiment is only a criterion of truth, but not the truth itself. The materialistic pragmatism of highly educated and well-equipped researchers of the supposed dual 'reality' (particle and field are different notions) was rejected by many philosophers. Indeed, by dividing an infinitely extended field-energy object into parts one should again obtain only infinite objects rather than corpuscles of limited size.
In contrast to available observations, nonempty space of continuous material flows has been recognized by philosophers not only in the Ancient East and in Ancient Greece, but also by many contemporary thinkers in the West ( et al. [7]). "A coherent field theory requires that all elements be continuous... And from this requirement arises the fact that the material particle has no place as a basic concept in a field theory. Thus, even apart from the fact that it does not include gravitation, Maxwell's theory cannot be considered as a complete theory"as was stated by Einstein and Infeld [6] in 1938. Indeed, the postulated pointparticle paradigm results in Coulomb energy divergence, which terminates Classical Electrodynamics as a self-consistent theory. A point source in the Maxwell-Lorentz Equations may be considered as "an attempt which we have called intellectually unsatisfying" according to De Broglie [8]. Einstein also criticized his 1915 field equation for the point gravitational source: "it resembles a building with one wing built of resplendent marble and the other built of cheap wood" (translation [9]). The Dirac delta operator for the density of point matter in the void seems to have pushed physicists away from physical reality and the Eastern approaches to superimposed energy flows.
The evolution of Einstein's theory of relativity has already passed three milestones. They are the 1905 postulates and the internal energy mc 2 of mechanical bodies, the 1915 geometrization of massless metric fields in the Einstein Equation under the Newton's empty space dogma, and the 1938 proposition to distribute particle's mass-energy mc 2 continuously over all spatial points of the material metric field in the non-Newton nondual approach to physical reality. Recall that the integration of particles into spatial structures of their fields was suggested by Einstein together with Infeld [6] for the further evolution of all natural disciplines: "We would regard matter as being made up of regions of space in which the field is extremely intense... There would be no room in this new physics for both field and matter, for the field would be the only reality." However, the extended mass has not been yet adopted by modern relativists. Their Newtonian references traditionally associate General Relativity (GR) only with the 1916 empty space metric of Schwarzschild, later denied by Einstein for physical reality in the 1939 thought experiment [10]. And they use the Dirac delta-function for the formal presentation of the Lagrange material density µ( 4 x for the Newton empty space x i with material peculiarities along the path ξ(x o ).
The elementary mass density is to be a continuous spatial function in the Cartesian world and in the Einstein-Infeld pure field approach to matter [6]. According to Einstein, there are equal inertial (active µ a ) and gravitational (passive µ p ) masses. Today the best mathematical candidate in nonempty pseudo-Riemannian world to match the sum of active and passive mass-energy densities is the Ricci scalar R ≡ g µν R µν = (µ a + µ p )c 2 /ζϕ 2 o [11,12]. Here ϕ o = c 2 / √ G = 1.04 × 10 27 V is the universal potential for the inertial/gravitational mass-energy charge g oo was already found for the static metric of nonempty space, when This geometrical formalism for static gravitational fields reveals the Ricci scalar meaning as a sum of passive and active mass densities, Principle of Equivalence for active and passive masses and with the mutual compensation of active (kinetic, positive) and passive (gravitational, negative) self-energies under such a mechanical motion. The relativistic invariant for the active (positive) mass-energy density of vortex mechanical space is (−ϕ 2 o f µν f µν /8π), which is together with its passive (yin-yang) partner is distributed continuously over all elementary 4-volumes √ −gd 4 x ≡ √ g oo dx o √ γd 3 x in the action integral for the 'void nothing': Here we introduced the kinematic scalar field B(x) = ds(x)/ √ g oo dx o next to the active and passive mass densities in the space-time elementary volume √ −gd 4 x, This inertial field depends on the GR metric tensor g µν = g νµ , the physical three-velocity The inertial field B contributes to the four-current of mass-energy, This Maxwell-type equation is the equality due to the Einstein equality of active and passive mass densities. The variational equation (4) supports the complex charge unification [13] of extended massenergies with imaginary electric self-energies, where E = ( √ Gm + ie)ϕ o = 511KeV − i104 × 10 22 KeV for the electron.
By applying the Einstein Principle of Equivalence to moving densities of active and passive masses in (2), one can say from (4) that the mass density µ(x) of extended matter in Cartesian mechanics originates from vortex velocities of material space: Relations (4)-(5) for the rest frame of the extended elementary particle without net rotations lead µ(r) = mr o /4πr 2 (r + r o ) 2 due to the following metric components g oo = r 2 /(r + r o ) 2 , r o ≡ √ Gm/ϕ o = Gm/c 2 , γ i j = δ i j , and g i j = −δ i j [11].
Despite inertial (kinetic) and gravitational (potential) mass densities are equal in (4), the positive kinetic (+µ a c 2 ) and the negative potential self-energies have opposite signs. The gravitational self-action always arises due to the intense interaction potential [11][12][13] Here the energy density is negative only in the very core of the radial carrier and is positive for r > r o /(e − 1). The integral energy balance (6) means the exact mutual compensation of kinetic and potential self-energies of any rest-mass body due to the Einstein Principle of Equivalence for inertial (m in ) and heavy (m gr ) masses. Contrary to measurable exchanges of internal and translation kinetic energies, the negative potential energy (−m gr c 2 ) is not relevant to observations in practice. But negative energy tensions within the continuous carrier (of the elementary relativistic heat Q = mc 2 ) is relevant to stability of the elementary extended mass-energy by the internal negative pressure, once assumed by Poincare for the extended particle.
Nondual Einstein-type equation for moving material space
The internal heat, as well as an internal kinetic degree of freedom under the spatial transport of elementary energy, cannot be reasonably assigned to the Newton point mass. Therefore, the variable internal heat is the principle difference between Newton and Einstein transport of energy, even at the low speed motion of mechanical bodies. Indeed, the Newton point particle possesses only one degree for the summary (internal and translation) kinetic energy, while the Cartesian distributed vortex for the elementary heat-energy possesses both kinetic internal and kinetic translation energies. The Special Relativity kinematic cooling Q(β 2 ) = Q o 1 − β 2 ≈ (mc 2 − mv 2 /2 − mv 4 /8c 2 − ...) of the elementary rest-mass energy mc 2 ≡ Q o , which is the elementary relativistic heat, and the pure translation (mechanical) energy reitirate together the Newton-type summary changes of internal (heat) and external (mechanical, translation) kinetic energies, ∆[Q(β 2 ) + p i v i ] = ∆(mv 2 /2 + 3mv 4 /8c 2 + ...) for the low speed (β 2 ≪ 1) transport of the elementary energy carrier. There is no mass transpart and conservation in relativistic energy physics.
The slow spatial motion of the Cartesian mechanical body is first of all the transport of the variable heat integral Q o (1 − β 2 /2) of vortex densities. The accompanying fraction of the kinetic energy Q o β 2 due to the translation ordering is very small compared to the kinetic energy of internal chaos. Cartesian transport of two variable energies just formally correspond to the Newton transport of the constant mass Q o /c 2 plus one variable energy Q o β 2 /2. Newton had no idea about the internal degree of kinematic heat-energy and modeled transport of net energy changes, It is clear from Cartesian physics that Newton's model works properly only for small probe bodies in empty space where internal (heat) and external (translation) kinematic degrees of freedom obey the collinear transport.
In condensed media, where heat energy gradients and mechanical energy flows may have different directions, Cartesian dynamics with internal degrees of freedom may not reiterate Newtonian one. Moreover, the Newton model of motionless point masses without internal heat does not coincide with the Cartesian system of static vortices which initially have internal kinetic energy or heat. One can expect some principle differences between the dual and nondual theories of matter for its spatial acceleration and its heat transfer in liquids and gases. Here, different exchange mechanisms can control non-Newtonian energy gains, mv 2 , and thermal losses, −mv 2 /2, of the relativistic kinematic cooling. Early or later, the elementary relativistic energy of internal vortex motions with variable heat should replace the Newtonian constant mass as a basic notion for description of warm material space in General Relativity. The latter can not rely on Newton's cold masses in order to incorporate thermodynamics at the low speed limit. Equally, metric gravitation of cold masses cannot be considered as a true limit for gravitation of Cartesian energy-charges, which depend on variable heat.
The last integral in the 'void nothing' action (2) can be replaced by integration of 4-forces u ν f µν B over the hyper-surface that provide the internal Poincare pressure for stabilization of the extended mass. The other part of the action integral is related to active and passive mass densities, which we Now we perform the Hilbert metric variations over δg µν in (2), 4 xT µν δg µν /2c = 0, in order to derive a nondual analog of the Einstein Equation for pure field physics of moving material flows, The Ricci material flows of active and passive mass densities are balanced in (8) by the Poincare negative stress P µν , which originates from the Hilbert variations of local force densities δ( √ −gBu ν f µν )/δ µν . Einstein was the first who tried to derive the motion of particles directly from his gravitational field equation. He was engaged into this conceptual problem for more than 20 years. Empty space regions in Newtonian dual physics are described by field solutions with zero Ricci curvatures, R µν = 0. The motion of point peculiarities in such massless fields can be approximated only through successive iterations [14][15][16]. Cartesian physics of material fields is free from these complicated iteration procedures. One can find the exact geodesic equations for any non-equilibrium motion of inertial Ricci densities in the world metric space-time as the vector flow conservation from the general tensor balance (8), ∇ ν T ν µ = 0. Here it is sufficient to make use of the known geometrical equalities This general equation of motion can be equally read for the 4-acceleration u ν ∇ ν u µ along a normal axis to the 4-velocity u µ of the scalar mass density µ = ϕ 2 o R/16πc 2 , The vector equation (10) for Cartesian mechanics claims that the inertial field B(x) const contributes to geodesic dynamics of material space densities, while dual theories with the Newton empty space alternative are free from such kinematic feedbacks. ICNFP 2017 Recall again that Newton's low speed dynamics works satisfactorily for small probe bodies where transports of the internal kinetic energy or relativistic heat Q = mc 2 (1 − β 2 /2) and the kinetic energy of spatial translations mc 2 β 2 have collinear directions. Newtonian proponents have to consider the mechanical motion as a transport of the constant elementary mass m = const plus the transport of one kinetic energy mv 2 /2. They derive dynamical laws for acceleration of such constant masses and, basing on these laws, compute accompanying energy flows. Cartesian physicists have to consider first of all the transport of elementary vortex energies with the kinematic cooling under motion, mc 2 (1 − β 2 /2) const, and they have to derive dynamical laws for vector energy flows. There is no mass transport and mass conservation law in cartesian physics. Basing on dynamical laws for variable heat and translation energies, one can also model the accompanying transport of elementary objects with the rest-frame energy mc 2 . The competing force and energy approaches lead to the same low speed dynamics of small probe energies mc 2 in external fields but not of material densities µ within continuous media with inhomogeneous 'life forces' µc 2 β 2 /2. Streams of liquids and gases with heat gradients do not obey Newton's dynamic for cold masses and appropriate conceptual tests can distinguish Newton and Cartesian world organizations in the regular Earth laboratory.
The Cartesian acceleration law (10) for mass-energy flows within the Einstein-Infeld material space can be considered in the limit of vanishing gravitational fields, g µν → η µν ≡ {1, −1, −1, −1}, and low speeds, Here the post Euler force density V i ∂ t µ is the 1903 Tsiolkovsky force for the reactive rocket motion. The new kinematic feedback µ∂ i V 2 /2 originates from local gradients of the scalar inertial field (3), which is not relevant to the 1916 GR dynamics in Newtonian empty space. The new kinematic potential V 2 /2, which modifies Euler's fluid dynamics and the Navier -Stokes equation, appears in (11) exclusively in Cartesian physics due to the nondual tensor equation (8) for the nonempty inertial space. The right-hand side of the equation (11) is associated with net forces from the stress tensor P ν µ in (10). Here we underlined the Euler's equation force with the pressure p gradient and the simplest drag force with the Stokes kinematic viscosity ν next to all other possible forces f i . The pressure p = µ(sT + µ ch − ε) within fluids can be bound quantitatively, as is known [18,19], with the temperature T , the specific entropy s, the chemical potential µ ch , and the specific internal energy ε, with dε = T ds − pd(1/µ) and dµ ch = (dp/µ) − sdT . The pressure differential, dp = (sT + µ ch − ε)dµ + µ(sdT + dµ ch + pdµ −1 ) = µ(sdT + dµ ch ), is related only to changes of the local temperature and the chemical potential. The net dynamical balance, dV i /dt = −s∂ i T − ∂ i (µ ch + V 2 /2) − V i ∂ t lnµ + ν∂ j ∂ j V i , of the Euler material derivative dV i /dt in the equation (11) depends on speed gradients next to thermal ones. Such a modification can suggest a lot of conceptual tests to compare Newtonian and Cartesian physics even for steady flows of liquids and gases with dV i /dt = 0. For example, contrary to Newtonian physics for laminar fluids with homogeneous pressure and temperature across the radial section of tubes, Cartesian physics predicts inhomogeneous p(r) and T (r) across the same steady flows. This can shed new light on 'unexpected' heat and pressure on boundaries of many dynamical systems like rocket engines, Ranque-Hilsch vortex tubes [20], plasma beams, etc. There are also various applications of the modified Navier-Stokes equation (11) to turbulent flows of energy in industrial hydro generators of electric power systems.
The kinematic self-deceleration −∂ i V 2 /2 due to the feedback of Leibniz -de Coriolis 'living forces' of a moving material space can be understood only through the Cartesian approach to physical reality. This inertial self-deceleration prevents asymptotic flows with diverging en-ergy densities. The unphysical energy divergence in the original Euler and Navier-Stokes equations was questioned by many researchers, including experts of the Clay Mathematical Institute (http://www.claymath.org/millennium-problems/navier-stokes-equation). Newtonian based mechanics results in the incomplete description of Navier-Stokes flows, while the Cartesian replica (11) with the inertial damping by 'living forces' is a more realistic approximation of real fluid dynamics.
In general, the new inertial force with its kinematic potential (3) for moving material spaces can justify the nondual field reality of Einstein and Infeld for the macroscopic Cartesian world. Nondual extended matter refutes applications of Newtonian dynamics to continuous media and considers the Navier -Stokes equation with diverging asymptotic solutions as conceptually insolvent. By closing, Cartesian mechanics does deserve very careful investigation for low speed energy flows and for further technological developments in a line of the Einstein and Infeld physics of the pure field. | 5,299 | 2018-08-01T00:00:00.000 | [
"Physics"
] |
REPRESENTATION OF COMPLEX OBJECT AND PROCESS STRUCTURES IN GRAPH DATA BASE ON THE EXAMPLE OF AGRICULTURAL TRACTOR
The paper presents an application of Graph Data Base technology in knowledge engineering task concerned with knowledge representation of complex structures. The first example is graph representation of a composition of tractor gearbox. It is an example of application of static partonomy relation. The adequate codding of that relation in Graph Data Base is easily expressed in Cypher language. There were presented some applications of Graph Data Base useful in the management of the technical object existence in different “life” phases, as design, production, maintenance and recycling. The second example concerns the Graph representation of activity structure. Such structures were built based on an agricultural tractor’s technical revision system. Created Data Base allows for processing the queries referred to the questions of what activities in what period should be performed to maintain the agricultural tractor ready to work in good technical condition. The Graph Data Base technology is the first step to create semantic systems for data storing and processing in order to extract knowledge and information useful in precise physical process management.
Introduction
The structure of complex facilities is defined by the term BOM (Bill of Material) [2].The precise specification of a technical object structure is used in all stages of its existence -from construction through production, operation and maintenance phases to recycling.Data describing the structure is an important element of the material flow logistics.There are many approaches for analytical representation BOM structures [7].Representation of structure in classical, relational databases is not very effective, because the number 1 Transport and Computer Science, University of Economics and Innovation in Lublin, Projektowa 4,Lublin,Polska, of parts of a complex object can reach thousands as well as the number of relations between them.Apart from the labour-intensive nature of implementing such a database, it is difficult to update it.Each extension or reduction of the number of elements requires changes in the entire database.The most important advantage of the graph databases is the reduction of search and upgrade operations to the detection and evaluation of paths in the graph [6,11].
In addition, the semantics of structural relationships represented explicitly by graph arcs enables data-driven machine learning and algorithmic inference.
Complex object BOM representation in graph databases GDB
Theoretical basis of graph databases is the theory of graphs, [1,16,18].The graph is a mathematical entity (abstraction) defined as a pair of sets (N, A), where N is a set of nodes and A is a set of relationships between nodes.Mathematical definition of the graph has a well-defined semantics and in terms of logic it corresponds to the predicate's logic.This makes graph language an expressive system of knowledge representation about complex structure objects and processes.
Graph nodes represent objects of the physical world and relationships determine how they constitute a complex system.These relationships may be expressed with predicates such as is_a, has_a and their extensions.This allows to represent symbolically the knowledge of what exists in the subject domain, and what properties it has.Extensions to these relationships, such as part_of and cause/effect_of are used in various logic inference schemes (induction, deduction, abduction) to describe and reason about an object's structure and processes.Specified relations are important in problems of ontological [9] knowledge representation, particularly that involves meronimic approach [13,14].From the pragmatic point of view, the language of graphs makes it possible to construct symbolic, highly modular and recursive representation of knowledge.
In a particular case, the nodes of the graph represent components (parts) of a complex object.A node may have multiple labels that are used to group nodes into sets.Labels are indexed to accelerate finding nodes in the graph.Nodes can have one or more properties (attributes) stored as key/value pairs).Nodes are connected to other nodes via relationships.Relationships are directional, connect two nodes and can have one or more properties (i.e., attributes stored as key/value pairs).Properties are named values where the name (or key) is a string.Properties can be indexed, and constrained, composite indexes can be created from multiple properties, [25].
Properties determine the role of nodes and, in the case of relationships, specify the flow between nodes.The term "flow" is a good metaphor when a process view of complex structures is needed.
In BOM structure, relationships represent predicate part_of or has_a_part.In a simple case, the graph representing the complex object is a Tree, but when a higher-order structural levels have common downstream components, it becomes a Network.
Technology of graph databases can be applied not only to the semantic representation of the complex objects structure but also to the structure of processes.Particularly useful is a representation of complex processes comprising all phases of a technical object existence [2].Technical object existence starts from its conceptualization phase, through design, production, operation and maintenance to recycling.The GDB can be used to record the facts that occur in the process.That register of events is required by standards for tracing and tracking operation, maintenance and repair services [4], emission processes [3], the state of road infrastructure [15], measurements of Loyalty [8], documentation and modelling of all transport's risks [10].Graph representation of process participants and relationships between them provides not only factual data but also meta data needed to extract from data their semantic content.
BOM as an example of GDB application can be generalized to BOP (Bill Of Process).Then, for example, the graph nodes represent activities/ operations that create different levels of process description and each node can be described with any number of properties, e.g. the purpose of an action, resources needed to perform it, criteria for correctness of its execution, etc. Relationships may be temporal interdependencies, necessary to organize and track the process over time.Each operation then has a specific start time, duration, end time, determined with accuracy to the probability distribution.Due to temporal relations, we know the temporal structure of operations in and the possibility of replacing one operation with several others, e.g. the operation of the whole assembly replacement is substituted by the exchange of specific parts, which may result with different cost and time.
GDB Design with NEO4J
A convenient tool for building graph database is Neo4J, [5,12,17,21,22,26].In this system, all operations on the GDB from creation through searching and updating are performed using queries coded in the CYPHER language, [24].
The correct sentence in that graphs language has a form: (node)-[relationship]->(node).
We use the word sentence here, because this structure is exactly the sentence structure: (subject) -[predicate] -(object), used in many natural languages.The graph is the equivalent of sentences in which relations represent predicates of sentences and the nodes of the graph are predicate arguments.In order to put a graph segment: (node) -[relation] -> (node) in the database, it should be preceded by the word CREATE or MERGE.We suggest using the MERGE command, because it does not duplicate the sentences previously introduced in contrast to CREATE.The definition of a node has the following scheme: (variable: label {property1, property2,..., propertyN }) where property is a couple {Variable: value}.The value of a variable can be a number or a string, and must be enclosed in quotation marks, for example {Color: 'red'}.A record of a relationship has a similar pattern: [:relationship_name {property1, property2,..., propertyN }] where properties, as of nodes, they are pairs of {Variable: value} eg.{Distance: '10 km', Flow: '100 car/hour '}.The sentence in the CYPHER language takes the form: MERGE (variable1: label1 {List of node1 properties) MERGE (variable2: label2 {List of node2 properties}) MERGE [:relationship_name { List of relationship propeties}]; CYPHER language is so expressive as SQL.Specification of CYPHER language is shown in [23,14].
Example of building GDB
As an example, the BOM of the URSUS 3514, models 102, 304, 315 the unit gearbox was selected.The data from the spare parts catalogue [20] were used.The gearboxes in each model differ from each other and for that we construct a graph whose input node is represented by the 3514 tractor and three of its models: 102, 304 and 315.The query Q1 (described precisely in supplement to the paper) creates the subsequent structural levels of the tractor URSUS 3514, produced in three models: 102, 304, 315 built from hundreds components and parts, from which only gearbox will be the subject of next consideration created in query Q2 (see in supplement).
Query Q3 displays graphically (Fig. 1.) the result of queries Q1 and Q2.Thanks to Neo4J interactives of each graph it is possible to see immediately all properties of the clicked node or relations in between.
Each gearbox consists of many parts, most of them are common in all models of tractors.Some parts appear repeatedly in each model.Cypher query Q4 (see in supplement) adds to GDB parts of gearbox model 102 and corresponding graph as a result of query Q5 (see in supplement) is shown in Fig. 2. Graph in Fig. 3. is too big to display readable names of gearbox parts and components.
When the graph is displayed on the screen in Neo4J, then after clicking on chosen node, thanks to graph reactiveness, all descriptions of each node like node name and other properties can be found out.Such possibility is shown in Fig. 4., It is possible to create more questions if we enter additional attributes (properties) to the nodes and relations; for example: parts price, parts weight, material that the part is made of, spare parts producer's name, climatic condition in which the parts are properly working, substitute of given parts, the year of tractor's or parts production; then the questions could be: 1. What is the material's price of gearbox assembled to given model or price of chosen components and parts (it is an important topic for logistic specialists who are working on price optimization of final product)?2. What is the weight of a given model's gearbox or its components?3. What type of materials are implemented in the gearbox of given model (important question in recycling activity)?4. What is the name of the certain spare part's producer? 5. Does the pointed part suitable to work in tropic climate condition (important topic for the designers, technologist and foreign traders planning the export of final product to different destinations all over the world)?6.What substitutes of original parts are allowed to use in certain circumstances of a tractor working (this question can concern the foreign markets)?7. What spare part is suitable to be installed in the gearbox of the tractor's model made in certain years?
Expanding BOM approach to the whole tractor, the questions presented above will relate to all components, subcomponents, parts and operating liquids.
In analogy to the construction of the graph that represents a BOM, it is possible to create graphical representations of Processes.This is illustrated on the example of technical inspections of the Ursus MF255 tractor, according to data specifying in the repair instructions [19].According to this manual, there are 5 types of technical revisions P-1 ... P-5.
Technical revision includes 8 inspection objects (engine, fuel system and air filter, ..., cabin) and in each of them there are 3 to 6 activities such as check and fill, clean, replace (Table 1.).Graph database representing technical revisions was created as well in Neo4j system.Some example queries Q7, Q8, Q9 (description see in supplement) are shown below.
Resulting answer drawings are self-explanatory.The relation in Fig. 6 represent actions of technical revisions operations listed in Table 1.For example: relation 49 means "check the oil level in engine's oil sump and fill it up if necessary".That information is available due to the graph interactive or a query.
The same explanation is valid also to the relations represented on Fig. 7., on which the index points the actions from Table 1.Designed GDB can be used for documentation (tracing) revision's history of tractors and then tracking it with appropriate queries.It seems that tracing and tracking process is very important application of GDB
Summary
The semantic character of GDB allows to formulate questions regarding internal regularities inherent in specific structures of complex objects.Detection of such regularities using machine learning and inference algorithms provides knowledge enabling effective evolution of existing structures.Having GDB of the entire evolutionary sequence of objects connected to the functional structure, it would be possible to automatically generate future design solutions depending on external scenarios considering the anticipated changes, for example the type of energy sources and operational infrastructure.
Another application of GDB is the possibility of constructive technological development of assemblies and subassemblies of a complex object, considering technical and technological progress based, for example, on replacing the force interaction between elements of the structure with the information effect in order to achieve the desired functionality with a lower energy and cost.
Graph databases can be used in many industries, improving the work of designers, technologists, logisticians, service staff, traders or marketing specialists.Once established and systematically updated, GDB would collect not only the construction and operation data, but also their change in time."CABIN" "check and fill up the window spray liquid" "CABIN" "check and tighten the cabin fastening screws" "CABIN" "clean the cabin filter" "CLUTCH and BRAKES" "check the brakes and adjust" "CLUTCH and BRAKES" "check the clutch pedal idle and adjust" "CLUTCH and BRAKES" "check the work of pneumatic installation and brake valves" "COOLING SYSTEM" "check the cooling liquid level in radiator and fill it up" "COOLING SYSTEM" "clean the radiator ribs" "COOLING SYSTEM" "drain, clean and fill up the cooling system" "ELECTRIC SYSTEM" "check the alternator" "ELECTRIC SYSTEM" "check the electrolyte level in battery and fill it up with distilled water" "ELECTRIC SYSTEM" "check the fan's belt tension and adjust properly" "ELECTRIC SYSTEM" "clean the upper battery surface and grease the clamps with technical Vaseline" "ENGINE" "change the engine oil" "ENGINE" "change the oil in oil sump" "ENGINE" "check and adjust the clearance of engine valves" "ENGINE" "check the oil level in engine's oil sump and fill it up if necessary" "ENGINE" "clean the engine ventilation host" "FUEL SYSTEM and AIR FILTER" "change the cartridge of fuel filter" "FUEL SYSTEM and AIR FILTER" "check the condition of glass fuel filter settler and drain the water" "FUEL SYSTEM and AIR FILTER" "check the liquid level in settler and oil level in air filter" "FUEL SYSTEM and AIR FILTER" "clean the cartridge and change the oil in wet air filter" "FUEL SYSTEM and AIR FILTER" "drain and clean the fuel tank" "FUEL SYSTEM and AIR FILTER" "verify the condition of injectors" "POWERTRAIN SYSTEM AND HYDRAULIC SYSTEM" "change the oil in powertrain system and planetary gears" "POWERTRAIN SYSTEM AND HYDRAULIC SYSTEM" "check and adjust the differential lock pedal" "POWERTRAIN SYSTEM AND HYDRAULIC SYSTEM" "check oil level in gearbox and fill it up" "POWERTRAIN SYSTEM AND HYDRAULIC SYSTEM" "check the oil level in powertrain system" "POWERTRAIN SYSTEM AND HYDRAULIC SYSTEM" "clean the transmission oil filter" "STEERING SYSTEM" "check the bearing pressure of the front wheel hubs and adjust" "STEERING SYSTEM" "check the geometry of the front wheels and adjust" e.Name p.ID c.Name "POWERTRAIN SYSTEM AND HYDRAULIC SYSTEM" "P-2" "check oil level in gearbox and fill it up" "FUEL SYSTEM and AIR FILTER" "P-2" "check the condition of glass fuel filter settler and drain the water" "FUEL SYSTEM and AIR FILTER" "P-2" "check the liquid level in settler and oil level in air filter" "ELECTRIC SYSTEM" "P-2" "check the electrolyte level in battery and fill it up with distilled water" "ELECTRIC SYSTEM" "P-2" "check the fan's belt tension and adjust properly" "COOLING SYSTEM" "P-2" "check the cooling liquid level in radiator and fill it up" "CLUTCH and BRAKES" "P-2" "check the clutch pedal idle and adjust" "ENGINE" "P-2" "check the oil level in engine's oil sump and fill it up if necessary" "CABIN" "P-2" "check and fill up the window spray liquid" "CABIN" "P-2" "clean the cabin filter"
Fig. 1 .
Fig. 1.Result of query Q3 (displaced as a screen-print from Noe4J, showing graph interactives, red colour represents note of URSUS tractor 3514, green colour -models of tractor, pink colour -three types of gearbox, blue colour represents relation has a part and red colour has a model; properties of the clicked note 7007 977 M91 is listed as: subassembly <id> Symbol: 7007 977 M91 name: Eight speed gearbox).
Fig. 2 .
Fig. 2. Graph of gearbox parts, model 102 as a result of query Q5
Fig. 4 .
Fig. 4. Graph representing additional explanation of Fig. 3., (by clicking the node with running number 115, the identification code: 1869 712 M91(from BOM) is appeared as well as the part name: PTO clutch shaft).
Table 1 .
Technical Revisions of Ursus MF255 | 3,824.8 | 2019-03-29T00:00:00.000 | [
"Agricultural and Food Sciences",
"Engineering",
"Computer Science"
] |
Evaluating the value of individualized 3D printed models for examination, diagnosis and treatment planning of cervical cancer
Background 3D printing holds great potential of improving examination, diagnosis and treatment planning as well as interprofessional communication in the field of gynecological oncology. In the current manuscript we evaluated five individualized, patient-specific models of cervical cancer FIGO Stage I-III, created with 3D printing, concerning their value for translational oncology. Methods Magnetic resonance imaging (MRI) of the pelvis was performed on a 3.0 Tesla MRI, including a T2-weighted isotropic 3D sequence. The MRI images were segmented and transferred to virtual 3D models via a custom-built 3D-model generation pipeline and printed by material extrusion. The 3D models were evaluated by all medical specialties involved in patient care of cervical cancer, namely surgeons, radiologists, pathologists and radiation oncologists. Information was obtained from evaluated profession-specific questionnaires which were filled out after inspecting all five models. The questionnaires included multiple-select questions, questions based on Likert scales (1 = „strongly disagree “ or „not at all useful “ up to 5 = „strongly agree “ or „extremely useful “) and dichotomous questions (“Yes” or “No”). Results Surgeons rated the models as useful during surgery (4.0 out of 5) and for patient communication (4.7 out of 5). Furthermore, they believed that the models had the potential to revise the patients’ treatment plan (3.7 out of 5). Pathologists evaluated with mean ratings of 3.0 out of 5 for the usefulness of the models in diagnostic reporting and macroscopic evaluation. Radiologist acknowledged the possibility of providing additional information compared to imaging alone (3.7 out of 5). Radiation oncologists strongly supported the concept by rating the models highly for understanding patient-specific pathological characteristics (4.3 out of 5), assisting interprofessional communication (mean 4.3 out of 5) and communication with patients (4.7 out of 5). They also found the models useful for improving radiotherapy treatment planning (4.3 out of 5). Conclusion The study revealed that the 3D printed models were generally well-received by all medical disciplines, with radiation oncologists showing particularly strong support. Addressing the concerns and tailoring the use of 3D models to the specific needs of each medical speciality will be essential for realizing their full potential in clinical practice. Supplementary Information The online version contains supplementary material available at 10.1186/s41205-024-00229-8.
Background
Cervical cancer represents the fourth leading cancerentity among women worldwide, with 604,127 new cases in 2020 [1].Treatment options include surgery, chemoradiotherapy, chemotherapy and radiotherapy (Cervical Cancer: ESMO Clinical Practice Guidelines).Treatment decision mainly depends on tumor size and tumor spread, but also age, general state of health and family planning are considered.Selection of the appropriate therapy is made by a multidisciplinary team of surgeons, radiologists and radiation oncologists and requires an accurate understanding of the patient's uterine anatomy, tumor location and disease progression.
So far, gynecologic pelvic examination and medical imaging are the only means to provide the above-mentioned information.Conventional diagnostic methods rely on 2D imaging modalities, presenting the complexity of cervical tumors only to a limited extent.Moreover, the interpretation of a large number of 2D slices is dependent on the experience of the examining physician and is timeconsuming [2].In contrast, the utilization of 3D printed models, based on MRI segmented scans, is expected to provide high-resolution anatomical insights allowing a more comprehensive visualization of the patients' individual cervical cancer disease [3].Until now, there is only limited data on the use of 3D printed models in daily clinical practice, mainly in the fields of dentistry [4,5].Yet, there are studies showing that 3D printed teaching models of cervical cancer can be helpful for postgraduate gynecological training [2,[6][7][8].The present work aims to evaluate the effectiveness and value of 3D printed models in improving the understanding of cervical cancer progression, aiding in diagnosis and facilitating more tailored and efficient treatment planning strategies.It was also investigated whether the 3D visualization could improve the interprofessional communication between surgeons, radiologists, pathologists and radiation oncologists.Moreover, the study evaluated whether these models could serve as a tool for patient education enabling more informed discussions and shared decision-making.
Subjects and acquisition conditions
This study was registered with the Institutional Review Board (Ref.No. 20200910 02).Five patients (mean age 38 years; age range 27-47 years) with histologically confirmed cervical cancer were included in the study.MRI of the pelvis as indicated by the national guidelines for locoregional staging [9] was performed on a 3.0 Tesla MRI System (Magnetom Skyra, Siemens Healthcare, Erlangen, Germany) in supine patient position.Following the recommendations of the European Society of Urogenital Radiology [10] a multiparametric approach, including but not limited to two-dimensional T2-weighted sequences, diffusions-weighted imaging and contrast-enhanced imaging (Gadovist, Bayer Vital GmbH, Germany; 0.1 mmol/ kg body weight) was used.Additionally, a coronar T2-weighted isotropic 3D fast turbo spin echo sequence was acquired (in plane resolution 0.9 mm × 0.9 mm, slice thickness 1 mm, field-of-view 300 mm × 300 mm) and reconstructed in transversal and sagittal orientation.This sequence was exported anonymized and used for segmentation.
Segmentation
The procedure from MRI scan to final print is depicted in Fig. 1.Based on the MRI scans, one board-certified radiologist with more than 5 years of experience in gynecologic pelvic imaging segmented uterus, bladder, rectum and cervical cancer.The segmented data was then processed by our own software (written in Python, Van Rossum, G., & Drake Jr, F. L. (1995), Python reference manual, Centrum voor Wiskunde en Informatica Amsterdam).Our software read in the dataset using the PyDicom library, extracted the individual slices and passed them on to a Marching Cubes algorithm.This algorithm was provided by the SKImage.measurelibrary which generated a list of vertices and facets (i.e.triangles).Due to the underlying voxel grid used in the marching cubes algorithm, the surface is usually not continuously smooth.Therefore we first processed the mesh in Blender (Community, B. O. 2018.Blender-a 3D modelling and rendering package.Stichting Blender Foundation, Amsterdam) by decimating the surface by a constant factor.We then applied Laplacian smoothing to the surface.The resulting virtual model was then exported from Blender as a.stl file for printing.
3D Printing
The resulting virtual 3D models, created from the segmented MRI scans, were subsequently 3D printed to create haptic and real-life representations.The overarching aim in the 3D printing process was to give the final products a desktop-model-like character.Therefore, a weighted stand was designed to mount the organs with the correct orientation and its base inscribed with the respective Study ID.The organs were printed at life-size scale.
One challenge which needed to be resolved was an organ overlap in the 3D-model resulting from imperfect scanning and extraction.This organ overlap was removed with Boolean operations, following a set logic sequence.The carcinoma was kept complete and no modifications to its shape were allowed, so as not to compromise operation planning.In case of an overlap with the uterus, the overlapping volume was removed from the uterus.In case of overlaps between uterus and bladder or uterus and rectum, the overlap volume was attributed to the uterus and cut out of the respective other organ.
The different model elements are mechanically fastened to each other but can be taken apart as well.The model base inserts into the lowest-lying organ (uterus or rectum) via a press fit.The rectum and bladder connect to the uterus via custom round pegs.The connecting pegs have a notch, so that the respective organs can only fit in one correct position, thereby ensuring an unambiguous orientation.Because one goal was to leave the carcinoma shape unaltered, this implies that no cut-outs for mechanical fastening are allowed in the carcinoma, presenting a challenge to the model design.This was solved by printing the entire carcinoma from ferromagnetic material and inserting magnets into the uterus at the uterus-carcinoma area of contact.An overview of the mechanical model design is given in Fig. 2.
The model elements were 3D printed via fused deposition modeling (FDM) on Prusa MK3s and Prusa Mini machines (Prusa, Praha, Czech Republic).The same filament was used to color-code the different elements across all models.This helps to quickly identify the organs, as the shapes and orientation vary considerably between patients.The filaments used were Polymaker PolyTerra PLA, type "Lava Red" for the uterus; Polymaker PolyTerra PLA, type "Sakura Pink" for the recta; Polymaker PolyTerra PLA, type "Forrest Green" for the bladder (Polymaker, Changshu, China); Filamentworld PLA "Grey" for the stand and connecting pegs (Filamentworld, Neu-Ulm, Germany); Proto-Pasta, type "Magnetic Iron", composite PLA as the ferromagnetic filament for the carcinomas (Protopasta, Vancouver, Washington, USA).All organs were printed with 15% gyroid infill, while the stand was printed at 30% gyroid infill for stability.The carcinomas were printed with gyroid infill depending on their size.Further, stainless steel washers (50 mm outer diameter) were glued into a slot at the base of the stand as weights.Different sizes of Neodym magnets (10 mm, 4 mm and 3 mm diameter, all with 2 mm height) were used to hold each carcinoma.A depiction of one 3D printed model is given in Fig. 3. anonymized.All groups assessed the five models with an evaluation-form which was individualized for every profession.The questionnaires were divided into three sections.The first section included questions concerning the general analysis of the five models.The second section evaluated the possibilities of application of the 3D printed models in the respective fields (surgery, pathology, radiology and radiotherapy).The third section asked about the physicians' general experience.Section one and two involved questions with a five-point Likert scale, dichotomous questions ("Yes" or "No") and multipleselect questions.Section three only included questions with a five-point rating scale.
Statistics
Statistical analysis was performed using by SPSS Statistics Version 25.0 (SPSS Inc., Chicago, IL.USA) and Microsoft Excel Version 2023 (Microsoft Corp., Washington, USA).
Use of large language models
ChatGPT Version 4.0 (OpenAI Inc., San Francisco) was used for language quality check.
Results
The 3D printed individualized models can be of aid for supporting patient care in case of cervical cancer.
Common questions for all disciplines
Twelve questions were answered by all disciplines: surgeons, radiologists, pathologists and radiation oncologists (Fig. 4).The first part of the questions was answered by choosing on a scale of one to five (1 = Strongly disagree, 2 = Disagree, 3 = Neither agree, nor disagree, 4 = Agree, 5 = Strongly agree, or 1 = Not at all useful, 2 = Not so useful, 3 = Somewhat useful, 4 = Very useful, 5 = Extremely useful).The three oncologic surgeons, who evaluated the five models, rated the models' resolution with a mean of 3.7 of 5 points and the helpfulness of the different coloring of the models' single parts for identifying single structures with a mean of 4.7 of 5 points.The three pathologists also partly agreed to these two aspects (3.0 and 4.0 of 5 points).Radiologists and radiation oncologists rated the detail of the single parts with 4.3 and 3.7 points and the coloring with 5.0 and 4.7 points.The models' size was assessed as accurate: 4.7 of 5 points from the pathological and surgical point of view Fig. 4 Mean scores of all common questions. 1 = "strongly disagree " or "not at all useful " up to 5 = "strongly agree " or"extremely useful ".Yellow = Radiation therapists, Grey = Radiologists, Orange = Pathologists, Blue = Surgeons and 4.3 of 5 points from the radiological and radiooncological perspective.The surgeons perceived the support generated by the 3D printed models for interprofessional communication in treatment planning for patients with locally advanced cervical cancer (LACC) and early cervical cancer (ECC) as somewhat useful and useful (3.3 and 3.7 of 5 points, respectively).The pathologists rated these aspects with 3.3 and 2.0 points and the radiologists with 4.3 and 2.7 points.The radiation oncologists evaluated the support for interprofessional communication with 4.7 points for LACC and 4.3 points for ECC.The option of utilizing the models for educational purposes for junior colleagues received 4.0 from 5 points from the surgical oncologists, 3.0 points by their pathological colleagues, 4.7 points by radiological specialists and 5.0 points by the radiation oncologists.
Two surgeons agreed and one strongly agreed that the models could save them time, leading to a mean point score of 4.3 of 5 points.Two pathologists disagreed and one neither agreed nor disagreed concerning this matter, making up for 2.3 points.Two radiologists neither agreed nor disagreed and one strongly disagreed, resulting in a mean of 2.3 points.One radiation oncologist neither agreed nor disagreed and two agreed.A mean of 3.7 points was given.All specialties were unsure, if the use of the 3D printed models could decrease overall supplies related to their diagnostic and therapeutical procedures (3.0 of 5 points by surgeons and pathologists, 3.3 of 5 points by radiation therapists and 2.7 points by radiologists).All participants of the study agreed that the models were easy to use, resulting in 3.7 points from the surgeryand pathology-department, 4.0 points from the radiology and 4.3 points from the radiation oncology department.The surgeons answered with mean point score of 3.7 to the question whether the models met their needs.The pathologists answered this question with 3.0 and the radiologists together with the radiation oncologists with a mean of 4.0 points.The surgeons and the radiation oncologists both agreed to using and recommending the use of patient-specific 3D models in the future (mean points of 4.0 each).The pathologists agreed to this in an equal manner (mean 4.0 points for "I would use a patientspecific 3D model in the future" and 3.7 points for "I would recommend use of a patient-specific 3D model in the future").The radiologists also gave a mean of 4.0 points for using and 4.3 points for recommending the use of patient-specific 3D models in the future.
Department-specific questions
Every discipline also answered questions tailored to their everyday clinical work with one part of the questions to be answered by using the rating scale from one to five points as stated above (Figs.5 and 6).The second part of the questions was answered by using "Yes " or "No " statements.
Department-specific questions: surgeons
The oncologic surgeons rated the question "When compared to the MRI-/CT-/ultrasound-images and to the description of palpation findings, is the additionally provided 3D printed model useful to improve your understanding of patient-specific pathological characteristics?" with a mean point score of 3.7 of 5.The question whether the 3D printed models could assist the surgeon to improve surgical planning by e.g.allowing the identification of a safer surgical pathway received a rating of 3.3 points.The question "Would the physical 3D printed model be useful during the surgical procedure?For example, the model may be used as a guide to orientate or navigate the cutting plane, by being physically positioned in the same way as the patient itself." scored a mean of 4.0 of 5 points.Two surgeons strongly agreed and one agreed that the models could be useful for communication with patients (4.7 of 5 points).The surgeons rated the question "The use of the model could revise the patient's treatment plan" with a mean of 3.7 of 5 points.All surgeons agreed that the 3D printed models provided additional information compared to the MRI-/CT-/ ultrasound-images and to the description of palpation findings.Moreover, all confirmed, that the models gave a better perception of the depth of local invasion (three times "Yes").The models also gave a better perception of the spatial relationships between surrounding structures (e.g.bladder, rectum), with all surgeons answering this question with "Yes".In addition, all surgeons were of the opinion, that understanding the depth of local invasion and the spatial relationship between surrounding structures is important for surgical planning for cervical cancer.The question "Do you believe that 3D printed cervical models could have the potential to reduce the chance of intraoperative complications?" was answered with "Yes" twice and with "Yes, in complex cases only".Furthermore, every questioned surgeon believed 3D printed cervical models could reduce the duration of surgery.
Department-specific questions: pathologists
The questioned pathologists rated the question "How useful do you perceive the 3D printed model to be during the diagnostic reporting process for the case provided in this study?"with a mean of 3.0 of 5 points.The same score was given to the question whether the 3D printed models were useful during the macroscopic evaluation of the specimen.The question "How useful do you perceive the 3D printed model to be during the evaluation of the resection margins?"scored 2.0 points.All pathologists believed the models served as a useful supplementary tool in the diagnostic reporting process for patients with cervical cancer in some cases (less than 50%).
Department-specific questions: radiologists
The department of radiology rated the question "When compared to the MRI-/CT-/ultrasound-images and to the description of palpation findings, is the additionally provided 3D printed model useful to improve your understanding of patient-specific pathological characteristics?" with a mean point score of 3.3 of 5.The question "When comparing the MRI images to the additionally provided 3D printed tumor model, is the 3D printed model useful to provide any additional information?"scored a mean of 3.7 points.The radiologists rated the question of how useful they perceive the 3D printed model to be during the diagnostic reporting process with a mean point score of 3.3.Two radiologists answered the question "When comparing the MRI-/CT-/ ultrasound-images to the 3D printed cervix model, does the 3D printed model give you a better perception of the depth of local invasion?"with "Yes" and one radiologist with "They provide the same information".Two radiologists agreed that the 3D printed cervix model gave a better perception of the spatial relationships between surrounding structures (e.g.bladder, rectum) then MRI-/ CT-/ultrasound-images.One radiologist was of the opinion, that they provided the same information.All radiologists stated that understanding the depth of invasion and the spatial relationship between surrounding structures is important in their diagnostic reporting process for cervical cancer.The question "Do you perceive the 3D printed model to be a useful supplementary tool in the diagnostic reporting process for patients with cervical cancer?" was answered once with "Yes, in some cases (less than 50%), once with "No" and once with "Yes, in most cases (more than 50%)".
Department-specific questions: radiation oncologists
The physicians of the department for radiation oncologists rated the question "When compared to the MRI-/ CT-/ultrasound-images and to the description of palpation findings, is the additionally provided 3D printed model useful to improve your understanding of patientspecific pathological characteristics?" with a mean of 4.3 points of 5 points.The same score was reached by the question "Could the 3D printed model be useful in assisting interprofessional communication between radiation oncologists and medical physicist?"and by the question Fig. 5 Mean scores of speciality-specific questions. 1 = "strongly disagree " or "not at all useful " up to 5 = "strongly agree " or"extremely useful ".Orange = Radiation therapists, Blue = Radiologist "Could the 3D printed model assist you to improve radiotherapy treatment planning?".The question "Could the 3D printed models be useful for communication with patients?"scored a mean of 4.7.The radiation oncologists rated the question of how useful they perceive the 3D printed model to be during radiotherapy treatment with a mean point score of 4.0.All radiation oncologists agreed that the 3D printed model provided additional information compared to the MRI-/CT-/ultrasound-images and to the description of palpation findings.The question "When comparing the MRI-/CT-/ultrasound-images to the 3D printed cervix model, does the 3D printed model give you a better perception of the depth of local invasion?"was answered twice with "No" and once with "They provide the same information".All radiation oncologists agreed that the 3D printed cervix model, gave a better perception of the spatial relationships between surrounding structures (e.g.bladder, rectum) compared to MRI-/CT-/ultrasound-images.The same confirmation was found for the question "Is understanding the depth of local invasion and the spatial relationship between surrounding structures important for surgical planning for cervical cancer?".The question "Do you believe that 3D printed cervical models could have the potential to reduce the chance of complications?" was answered with "Yes, the enhanced perception that the printed models provide could further reduce the chance of complications during radiation" twice and once with "Yes, in all cases (complex and relatively simple cases)".
Discussion
This work focuses on the application of 3D printed models in the clinical field, particularly emphasizing the use of individualized 3D printed models for the diagnostic process and treatment planning of cervical cancer by surgeons, pathologists, radiologists and radiation oncologists.
Accurate anatomical representation
An important condition for further integration of 3D printed models into clinical practice is an accurate and easily comprehensible representation of the patients' anatomy.All participants including surgeons, radiologists, pathologists and radiation oncologists evaluated the model resolution (mean scores between 3.0 and 4.3 points out of 5) and its usefulness in identifying specific structures positively (mean scores between 4.0 and 5.0 Fig. 6 Mean scores of speciality-specific questions. 1 = "strongly disagree " or "not at all useful " up to 5 = "strongly agree " or"extremely useful ".Orange = Pathologists, Blue = Surgeons out of 5).Additionally, the models were considered to be of accurate size, with ratings ranging from 4.3 to 4.7 out of 5, and easy to use, with ratings from 3.7 to 4.3 out of 5, depending on the discipline.Therefore, the condition of accurate and comprehensible anatomy representation was met.
Interprofessional communication
Another aspect explored in this study was the potential for 3D printed models to enhance translational oncology by supporting interprofessional communication.Interestingly, opinions varied among radiation oncologists, radiologists, surgeons and pathologists.Radiation oncologists and radiologists highly valued the models for facilitating the exchange between medical professions concerning LACC, rating them with 4.7 and 4.3 points out of 5.However, both surgeons and pathologists saw limited usefulness in these models, with mean scores of 3.3 out of 5.The discrepancy may be due to the fact that the two disciplines, surgery and pathology, are less accustomed to working with replicas and models and typically involve hands-on practice.Tailored training and strategies for integrating 3D models into their workflow could potentially bridge this gap.
Usefulness during treatment/diagnostic procedure
The potential for utilizing the models during treatment/ diagnostic procedure was also evaluated differently by the four specialty fields: Surgeons and radiation oncologists expressed a high interest with mean scores of 4.0 out of 5, whereas pathologists and radiologist were more critical with mean score of 3.0 and 3.3 out of 5.This is not surprising, considering that both disciplines, radiologists and pathologists, are typically not directly involved in treatment planning.The surgeons' and radiation therapists' high interest, in contrast, reflects the potential for enhanced treatment precision and reduced side effects.
Educational value
Despite the lack of consensus among the four disciplines regarding the benefits during treatment and diagnostic reporting, there was agreement that 3D printed models could be used and recommended in future, particularly for educational purpose.Remarkably, radiation oncologists expressed the highest enthusiasm for educational value, rating it at 5 points.Surgeons, radiologists, and pathologists also recognized the educational benefits of the models rating with mean scores between 3.0 and 4.7 out of 5.This aligns with the fact that 3D printed models are already used in medicine for educational purposes in various ways, as demonstrated by previous studies (Marconi et al., 2017 and Wake et al., 2019).
Evaluation of the 3D printed models by department specific questions Surgeons
When examining the utility of 3D printed models in the field of oncologic surgery, the findings reveal a generally positive perception among surgeons.They found the 3D printed models very useful during surgical procedure, such as for guiding surgical orientation and navigation (mean score of 4.0 out of 5).This could be attributed to the surgeons' discovery that the models offered additional information beyond traditional imaging and palpation findings enhancing the perception of the depth of local invasion and spatial relationships between surrounding structures.Surgeons concluded that 3D printed models could reduce the chance of intraoperative complications and that they could reduce the duration of surgery which has implications for improving patient outcome and resource efficiency.Overall, these findings suggest that, in the surgeons' opinion, 3D printed cervical models hold promise in enhancing their clinical practice.
Pathologist
When examining the utility of 3D printed models in the field of pathology the findings reveal a more critical opinion.Pathologist rated the overall usefulness during the macroscopic evaluation and the diagnostic reporting process with a moderate mean score of 3.0 out of 5.These results underline the fact that 3D printing does not facilitate the examination of the resection margin, which is one of the main questions posed to the pathologists (mean score of 2.0 out of 5).Despite these moderate ratings, all pathologists agreed that the 3D printed models-in a limited number of cases-could be useful as supplementary tools in the diagnostic reporting process.Additional research and refinement of the technology, especially concerning the tumor margins, could lead to a more widespread and effective adoption in pathological practice.
Radiologists
When examining the utility of 3D printed models in the field of radiology the findings reveal a mixed opinion.The 3D printed models were considered only moderately useful for improving the understanding of patient-specific pathological characteristics and during the diagnostic reporting process with mean scores of 3.3 out of 5.This is rather surprising as most radiologists agreed that the 3D printed model enhanced their understanding of the spatial relationships between surrounding structures.Moreover, when compared to MRI images the 3D printed tumor model was generally seen as providing additional information, receiving a mean score of 3.7 points out of 5. When asked about the overall usefulness as a supplementary tool, responses were mixed with one radiologist agreeing in some cases, one disagreeing and one agreeing in most cases.This underlines once more that further investigation is needed to optimize the 3D printed models' role for radiologists.
Radiation oncologists
Among radiation oncologists the assessment of 3D printed models for cervical cancer demonstrates highly positive findings.Experts in radiotherapy rated the 3D printed models with 4.3 out of 5 when compared to traditional imaging and palpation findings.Regarding treatment planning and radiotherapy treatment the models received a strong score of 4.3 and 4.0 out of 5, suggesting they are beneficial in the phase of treatment process.This might be due to the fact that the 3D printed models were seen as particularly effective in improving the perception of spatial relationships between surrounding structures and reducing the chance of complications.These findings strongly underline the value of 3D printed models in radiotherapy for cervical cancer.They are perceived as highly effective tools for improving understanding, communication and treatment planning.
Existing data and limitations
The advancement of 3D printing technology has led to an increasing number of publications assessing the use of 3D printing in medicine.Systematic reviews have outlined its advantages in surgery, such as enhanced pre-operative planning, improved operative outcomes and reduced surgical time [5,11].Initially focused on specific surgical specialties, recent reviews, including one from 2023, highlight a growing interest in 3D printing in gynecology, particularly in gynecologic oncology [12][13][14].Cooke et al. searched medical databases and systematically reported on the clinical applications of individualized 3D printing in gynecology [12].63% of the studies investigated printed five or less models which is equivalent to the number of 3D printed models in our study.Notably, patient specific 3D printed brachytherapy devices emerged as a widely studied application [12,15].Studies reported increased radiation doses to the target volume and decreased dose to organs at risk, emphasizing the potential benefits of individualized 3D printing in improving treatment outcomes [15,16].This high interest in 3D printing by radiation oncologists is also confirmed by our study results: Physicians in the field of radiooncology rated our 3D printed models with the highest possible score of 5.0 out of 5 when compared to traditional imaging and palpation findings.Regarding treatment planning and radiotherapy, the models received a strong score of 4.0 out of 5, suggesting they are beneficial in the phase of treatment process.
Even though 3D printed models in gynecologic oncology have mainly been utilized for production of patient specific 3D printed brachytherapy guides/ applicators, studies indicate its additional potential in surgical planning and education [6,12,17,18].To validate this concept initially, Ajao et al. and Mackey et al. generated highly accurate individualized models of endometriotic nodules or fibroids by applying additive manufacturing.These models closely aligned with surrounding tissue and.
and adequately depicted patient anatomy [19,20].Studies by Baek et al. and Sayed Aluwee et al. further substantiated this approach, revealing that gynecologic oncologists experienced enhanced comprehension of patient anatomy and pathology, including factors such as tumor size, shape, and borders.The use of individualized 3D printed models notably increased their confidence in determining the optimal route of excision, particularly in the preparatory phase of oncologic surgeries [2,6].Overall, these studies highlight the accuracy of 3D printed models in representing gynecologic pathology, aiding surgical comprehension, and increasing confidence in surgical routes.This observation was confirmed by our study results: Surgeons rated the possibility of additional information with a mean point score of 3.7 of 5 and the usefulness during surgical procedure with a mean of 4.0 of 5 points.
In contrast to existing studies, our patient-specific 3D printed models of cervical cancer were evaluated by all medical specialties involved in diagnosis and treatment planning.Thereby we could highlight the differences of opinions and identify the topics which surgeons, radiologists, radiation therapist and pathologists agreed on.Interestingly the benefit for interprofessional communication for instance was rated differently: Radiation oncologists and radiologist highly valued the models for facilitating the exchange between medical professions.Surgeons and pathologist were more reserved.Nevertheless, all specialties agreed that 3D printed models could be used and recommended in future, particularly for educational purpose.
Given the pivotal role of interprofessional collaboration in oncology, we recommend further studies integrating all specialties to comprehensively assess the impact of 3D printing in medical practice.
Conclusion
All specialty fields agreed on the usefulness of the individualized 3D printed models as supplementary tools.There were mixed responses concerning the utilization for diagnostic processes and treatment planning of cervical cancer by surgeons, pathologists, radiologists and radiation oncologists.As imaging techniques and data processing, particularly supported by artificial intelligence, continue to advance, the integration of 3D printing in medicine is likely to become increasingly prevalent in future.Novel frameworks, which automate the digitization of patient anatomies, facilitate precise and swift three-dimensional reconstructions while ensuring a faithful representation of their physical characteristics.The ability to create accurate, patient-specific models holds considerable potential for personalized medicine, further extending the quality of clinical diagnostics und treatment planning.Furthermore, 3D printed models can be produced at low costs by anyone with access to a 3D printer, which makes it a cost-effective tool for every clinic.The nearly worldwide and quick accessibility combined with minimal resource demands facilitates a more efficient dissemination of gynecologic knowledge and skills among health care providers worldwide.This underlines the importance of this field of research and shows that further investigation should be conducted.Addressing the concerns and tailoring the use of additive manufacturing to the specific needs of each medical specialty will be essential for realizing its full potential in clinical practice.
Appendix
In the following, the questions are stated, which were used to evaluate the 3D-printed model.
QUESTIONNAIRE 1 (Radiology)
☐ I have received information regarding this research and had an opportunity to ask questions.I believe I understand the purpose, extent and possible risks of my involvement in this project and I voluntarily consent to take part.
SECTION ONE -Applications of the 3D printed model.
( (5) On a scale from 1 to 5, when compared to the MRI images, is the additionally provided 3D printed model useful to improve your understanding of patient-specific pathological characteristics?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.
(6) On a scale from 1 to 5, when comparing the MRI images to the additionally provided 3D printed tumor model, is the the 3D printed model useful to provide any additional information?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.☐ YES ☐ NO ☐ They provide the same information (c) Is understanding the depth of invasion and the spatial relationship between surrounding structures important in your diagnostic reporting process for cervical cancer?
☐ YES, information related to both depth and spatial relationships is important.☐ YES, only the depth of invasion.☐ YES, only the spatial relationships between structures.☐ NO, neither is important in surgical planning.
SECTION TWO -Applications of the 3D printed model (8) How useful do you perceive the 3D printed model to be during the diagnostic reporting process for the case provided in this study?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.
(9) Do you perceive the 3D printed model to be a useful supplementary tool in the diagnostic reporting process for patients with cervical cancer?☐ YES, in most cases (more than 50%) ☐ YES, in some cases (less than 50%) ☐ YES, in simple cases ☐ YES, in complex cases ☐ NO (10) How useful do you perceive the 3D printed model to be in assisting interprofessional communication between radiologists and associated health professionals involved in the treatment planning for patients with locally advanced cervical cancer?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.
(11) How useful do you perceive the 3D printed model to be in assisting interprofessional communication between radiologists and associated health professionals involved in the treatment planning for patients with early cervical cancer?
QUESTIONNAIRE 2 (Surgeons)
☐ I have received information regarding this research and had an opportunity to ask questions.I believe I understand the purpose, extent and possible risks of my involvement in this project and I voluntarily consent to take part.SECTION ONE -Analysis of the 3D printed model.
(1) The resolution of the model (i.e. the detail of the single parts) was adequate.
(2) The model coloring helped identify relevant structures.
(3) The model accurately reflected the patient's anatomy in size.
(4) When compared to the MRI-/ CT-/ultrasound images and to the description of palpation findings, is the additionally provided 3D printed model useful to improve your understanding of patient-specific pathological characteristics?
(5) When compared to the MRI-/ CT-/ultrasound images and to the description of palpation findings, does the 3D printed model provide any additional information?☐ They provide the same information (c) Is understanding the depth of invasion and the spatial relationship between surrounding structures important in surgical planning for cervical cancer?
☐ YES, information related to both depth and spatial relationships is important.☐ YES, only the depth of invasion.☐ YES, only the spatial relationships between structures.☐ NO, neither is important in surgical planning.
SECTION TWO -Applications of the 3D printed model.(7) Could the 3D printed model assist you to improve surgical planning by allowing you to identify a safer surgical pathway?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.
(8) Would the physical 3D printed model be useful during the surgical procedure?For example, the model may be used as a guide to orientate or navigate the cutting plane, by being physically positioned in the same way as the patient itself.
(9) Do you believe that 3D printed cervical models could have the potential to reduce the chance of intraoperative complications?Note: you may tick more than one box to answer this question ☐ YES, in all cases (complex and relatively simple cases) ☐ YES, in complex cases only.☐ YES, in simple cases only.☐ YES, the enhanced perception that the printed models provide could further reduce the chance of intraoperative complications.☐ NO, the 3D reconstructed images give enough information to reduce the chance of intraoperative complications, the printed model would give no extra information to further reduce chances.☐ Unsure.
(10) Do you believe that 3D printed cervical models could have the potential to reduce the duration of surgery?Note: you may tick more than one box to answer this question.
☐ YES, in all cases (complex and relatively simple cases).☐ YES, in complex cases only.☐ YES, in simple cases only.☐ YES, the enhanced perception that the printed models provide could further reduce the chance of intraoperative complications.☐ NO, the MRI images give enough information to reduce the chance of intraoperative complications, the printed model would give no extra information to further reduce chances.☐ Unsure.
(11) How useful do you perceive the 3D printed model to be in assisting interprofessional communication between radiologists and associated health professionals involved in the treatment planning for patients with locally advanced cervical cancer?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.(12) How useful do you perceive the 3D printed model to be in assisting interprofessional communication between radiologists and associated health professionals involved in the treatment planning for patients with early cervical cancer?
( End of questionnaire.Thank you.
QUESTIONNAIRE 3 (Radiation therapist)
☐ I have received information regarding this research and had an opportunity to ask questions.I believe I understand the purpose, extent and possible risks of my involvement in this project and I voluntarily consent to take part.SECTION ONE -Applications of the 3D printed model.
( (5) When compared to the MRI images, is the additionally provided 3D printed model useful to improve your understanding of patient-specific pathological characteristics?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.
(6) When compared to the MRI images, does the 3D printed model provide any additional information?
☐ YES ☐ NO (7) When comparing the MRI images to the 3D printed cervix model... (c) Is understanding the depth of invasion and the spatial relationship between surrounding structures important in surgical planning for cervical cancer?☐ YES, information related to both depth and spatial relationships is important.☐ YES, only the depth of invasion.☐ YES, only the spatial relationships between structures.☐ NO, neither is important in surgical planning.
SECTION TWO -Applications of the 3D printed model.
( (5) When compared to the MRI images, is the additionally provided 3D printed model useful to improve your understanding of patient-specific pathological characteristics?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.☐ YES ☐NO ☐ They provide the same information (c) Is understanding the depth of invasion and the spatial relationship between surrounding structures important in surgical planning for cervical cancer?☐ YES, information related to both depth and spatial relationships is important.☐ YES, only the depth of invasion.☐ YES, only the spatial relationships between structures.
☐ NO, neither is important in surgical planning.
SECTION TWO -Applications of the 3D printed model.(8) How useful do you perceive the 3D printed model to be during the diagnostic reporting process for the case provided in this study?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.
(9) Do you perceive the 3D printed model to be a useful supplementary tool in the diagnostic reporting process for patients with cervical cancer?☐ YES, in most cases (more than 50%) ☐ YES, in some cases (less than 50%) ☐ YES, in simple cases ☐ YES, in complex cases ☐ NO (10) How useful do you perceive the 3D printed model to be in assisting interprofessional communication between pathologists and associated health professionals involved in the treatment planning for patients with locally advanced cervical cancer?1=Not at all useful.2=Not so useful.3=Somewhat useful.4=Very useful.5=Extremely useful.(11) How useful do you perceive the 3D printed model to be in assisting interprofessional communication between pathologists and associated health professionals involved in the treatment planning for patients with early cervical cancer?
1=not at all useful 2=not so useful 3=somewhat useful 4=very useful.5=extremely useful (12) Could the 3D printed models be useful in education and training of junior pathologists or students?
Fig. 1
Fig. 1 Our proposed pipeline of creating individualized 3D printed models.Orange: manual steps.Blue: Automated steps
Fig. 2 Fig. 3
Fig. 2 Depiction of the anatomical structures which were segmented from MRI scans, 3D printed and attached by connecting pegs and a magnetic pin.Green: Bladder.Red: Uterus.Grey: Cervical cancer.Pink: Rectum
( 7 )
When comparing the MRI images to the 3D printed cervix model... (a) Does the 3D printed model give you a better perception of the depth of invasion?☐ YES ☐ NO ☐ They provide the same information (b) Does the 3D printed model give you a better perception of the spatial relationship between surrounding structures (e.g.bladder, rectum)?
( 6 )
When comparing the MRI images to the 3D printed cervix model... (a) Does the 3D printed model give you a better perception of the depth of invasion?☐ YES ☐ NO ☐ They provide the same information (b) Does the 3D printed model give you a better perception of the spatial relationships between surrounding structures (e.g.bladder, rectum)?☐ YES ☐ NO
( 6 )( 7 )
When compared to the MRI images, does the 3D printed model provide any additional information?When comparing the MRI images to the 3D printed cervix model... (a) Does the 3D printed model give you a better perception of the depth of invasion?☐ YES ☐ NO ☐ They provide the same information (b) Does the 3D printed model give you a better perception of the spatial relationships between surrounding structures (e.g.bladder, rectum)? | 9,361.8 | 2024-07-27T00:00:00.000 | [
"Medicine",
"Engineering"
] |
Effects of Mutual Interaction between Constituent Elements on Phase Formation of High Entropy Alloys
Introduction High entropy alloys (HEAs) are alloys that contains multiple elements, often five or more principal elements in equiatomic or near equiatomic ratio with or without minor elements. As felt from the name, high mixing entropy is the sole parameter to decide the solid solution alloy formation in HEAs but, the mutual interaction between elements are key parameters to predict the phase formation in HEAs as proposed by some researchers. In this review the guideline for alloy design in HEAs for the formation of solid solution phases are proposed on basis of atomic size difference (δ), mixing entropy (ΔSmix), mixing enthalpy (ΔHmix), valence electron concentration (VEC) and electronegativity difference (Δχ). Effect of one more parameter Ω is also discussed. Ω is a parameter of the entropy of mixing timing and the average melting temperature of the elements over the enthalpy of mixing. Solid solution phase forms only when the requirement of these parameters are all met. These parametric constraints will lend a valuable motivation for future developments in HEAs.
Introduction
High entropy alloys (HEAs) are alloys that contains multiple elements, often five or more principal elements in equiatomic or near equiatomic ratio with or without minor elements.As felt from the name, high mixing entropy is the sole parameter to decide the solid solution alloy formation in HEAs but, the mutual interaction between elements are key parameters to predict the phase formation in HEAs as proposed by some researchers.In this review the guideline for alloy design in HEAs for the formation of solid solution phases are proposed on basis of atomic size difference (δ), mixing entropy (ΔSmix), mixing enthalpy (ΔHmix), valence electron concentration (VEC) and electronegativity difference (Δχ).Effect of one more parameter Ω is also discussed.Ω is a parameter of the entropy of mixing timing and the average melting temperature of the elements over the enthalpy of mixing.Solid solution phase forms only when the requirement of these parameters are all met.These parametric constraints will lend a valuable motivation for future developments in HEAs.
Keywords: High entropy alloys; Atomic size difference solid solution; Crystal structure; Solid solution ISSN: 2348-9812 After the development of bulk metallic glasses (BMGs), high entropy alloys (HEAs) emerged as an advanced metallic material.The concept of HEAs was introduced by Achard in the late of eighteen century, then in the mid of 1900s Yeh and his colleagues explored the world of multicomponent alloys [1,2].HEAs are often in single solid solution form has been attracting extensive research of materials community for their excellent properties like wear resistance, high hardness, softening resistance at higher temperature, corrosion and oxidation resistance.The alloy design concept in HEAs is beyond the realm of traditional powder metallurgy concept.The phase formation in HEAs are typically solid solution FCC and/or BCC structured which is far from intermetallic compounds formation in conventional physical metallurgy process.Thus, HEAs be given notable attention in last one decade.State of the art gives a significant idea that most studies in HEAs are focused on phase identification, microstructures, mechanical and physical properties [3][4][5][6][7][8].Although to a small extent the attention was paid to the effect of atomic size, valence electrons and electronegativity of constituent elements on HEAs, they are as the truth also quite encouraging [9,10].
The effect of vanadium (V) addition on the microstructure and mechanical properties of as cast AlCoCrFeNiV x , Al 0.5 CoCrFeCuFeNiV x and CoCrFeMnNiV x HEAs have been studies and remarked that vanadium greatly increases the hardness of the material and decreases the ductility due to formation of BCC phase with increase in V percentage [15,16,17].The phase analysis report for Notably, when the amounts of Al, Cu, Mo, Nb, Si, Ti, and V increases beyond a certain limit then, phase change in HEAs occurs for the combination of certain elements.In AlCoCrFeNiTi x the addition of Ti changes BCC phase to BCC1 (AlNi rich) and BCC2 (Fe-Cr rich) due to lattice expansion shifting of phase towards lower 2θ value [11].Similarly, in Al x CoCrCuFeNi HEAs without Al (x=0) the FCC structure obtained but, by adding Al in this combination tunes the FCC structure to FCC+BCC structure (x=1-1.3).Further addition of Al (x=2.3-3)converts the FCC+BCC crystal structure to BCC structure [12,13].He JY, et al. synthesized a series of (FeCoNiCrMn)100x Al x (x=0-20 at.%) to investigate the effect of Al addition on the tensile properties [14].They concluded that, tensile strength increases and ductility decreases with the Al addition and hence brittleness of HEAs increases due to formation of BCC phase.The aim of this review is hence to get idea of physical parameters that strongly dominance the stability of FCC and BCC phases in HEAs.
Analysis
CoCrFeMnNiV x clearly indicates that without Al and Cu the phases are single solution FCC phases.In AlCoCrFeNiV x the phases obtained are totally BCC, whereas Al 0.5 CoCrFeCuFeNiV x shows both FCC and BCC phases (x=0.2-1).The conversion of FCC phase to BCC phase in AlCoCrFeNiV x is due to large atomic radius of Al compared to other compositional elements.Similarly, FCC+BCC phase obtained in Al 0.5 CoCrFeCuFeNiV x is due to Al is a BCC stabilizer and large positive enthalpy of Cu than other elements [9,18].In CoCrFeNiNb x alloys, the FCC phase formed for the range of x=0.103-0.412whereas for AlCoCrFeNb x Ni alloys the phases obtained are BCC for x=0.1-0.75 due to large atomic radius of Nb [19,20].Here also Al acts as a BCC stabilizer.The phase analysis of Al 0.5 CoCrCuFeNiSi x was done by Xiaotao Liu, et al., and they reported that for x=0.4 the peaks of BCC phase appears corresponding to Si content and the intensity goes on increasing by increasing the Si content.Hence by Si addition the single solid solution of FCC phase transforms into FCC+BCC phase due to positive enthalpy of Cu.This results into segregation of Cu [21].
Zhu, et al. did intense work on AlCoCrCuFeNiMo x and AlCoCrFeNiMo x to investigate the effect of Mo as well as of Cu.Upto x=0-0.2, the material exhibits combined BCC and FCC structure in earlier one after that only BCC phase formation occurs (x=0.2-1),whereas in latter composition only BCC phase was seen for x=0-0.1 and further addition of Mo (x=0.1-0.5)resulted into the formation of an unidentified α phase [22,23].In HEAs, transition materials (TM) are the most studied elements and they have spin d-electrons.Therefore, it is also important to review the characteristic of d-electrons for structural stability in HEAs.Valence electron concentration and electronegativity are the two parameters that depends on the d-electrons.VEC is basically defined as the total electrons together with the d-electrons accommodated within the valence band [9,24].Electronegativity is the property of an atom to attract electrons.One more very important factor as stated by hume-rothery in alloy design is atomic mismatch, also called atomic size difference.Atomic mismatch is also a critical issue within the new category of HEAs by taking into account the lattice distortion [25].In present study we are considering these parameters to predict the stability of solid solution phases.[30] ΔH Mix , ΔS Mix , VEC, δ, Δχ, and Ω for a series of multicomponent alloys which are in equiatomic or near to equiatomic were statistically analysed and listed in Table 1.The calculations of different parameters were done by Meidema's approach.All the listed alloys were prepared by casting technique and phase analysis was carried out in as cast condition.Table 1 shows the mixing enthalpy calculated by Miedema's approach.
Result and Discussion
According to the Hume-Rothery (H-R) rule, the atomic size difference is a potent parameter in controlling or determining the solubility of alloys.H-R rule for binary alloys, the atomic size difference is defined by |r A -r B |/r B where r A and r B are atomic radii of solute and solvent respectively.For simplicity, Zhang, et al. studied the collective behavior of constituent elements and gives the relation for atomic mismatch (δ), mixing enthalpy (ΔH Mix ), and mixing entropy (ΔS Mix ) on the solid solution phases of multicomponent alloy, which are: Where , n is the number of alloying elements, C i and r i are the atomic percentage and atomic radius of the i th element respectively.1 , is the Pauling Electronegativity for the i th element [28]. Where
the mixing enthalpy of binary liquid AB alloys and
Where R is the gas constant (8.314J/mol.K) [26,27].
Fang, et al. studied the relationship between the phase stability and electronegative difference in a multicomponent alloy system, which is given as Gue and Thaddeus introduced the effect of valence electron concentration (VEC), which is the number of total electrons accommodated in valence band, given by Where (VEC) i is the VEC for the i th element [24,29].
In addition of these three critical parameters, another one more parameter, Ω is used to predict the phase formation behavior of HEAs, given as From Figure 1 we can easily say that for light transition metal (Cr, Mn, Fe, Co, Ni, Cu,) based HEAs the FCC phase shows higher mixing enthalpy but as the large size elements like Al, Ti, V, and Nb is added, then the enthalpy decreases.For most of the materials which are studied in this paper shows the range of mixing enthalpy -9.7 ≤ ∆H mix ≤ 3.2 for FCC phase as shown in Figure 1.The ranges of VEC for all combinations, for VEC ≥ 7.95, single BCC phase is not stable and most of the FCC phases are obtained in the region.Whereas for 7.2 ≤ VEC ≥ 7.95 all three combinations have been seen i.e., FCC, BCC and FCC+BCC but stability of BCC phase is high for the VEC ≤ 7.2.Atomic mismatch (δ) of FCC phase ranges from 1.03 ≤ δ ≤ 3.7 whereas for BCC phase the range is 5.33 ≤ δ ≤ 7.01 and for FCC and FCC+BCC phases obtained in the range 3.7 ≤ δ ≤ 7.01.
Where T m is the melting temperature of n-elements alloy and calculated as 1 ( ) here, (T m ) i is the melting point temperature of the i th element [21].
However, these empirical relationship criterion works for some combinations of the constituent elements to predict the stability of FCC and BCC solid solutions.Hence, the objective of this paper is to focus on the investigation of those physical parameters that would strongly predicts the stability of FCC and BCC phases in HEAs.
Tong, et al. discussed the effect of Al addition on microstructure and mechanical properties of Al x CoCrCuFeNi multi-component alloy.They observed that when x<0.5, the alloys showed a simple FCC structure.As x=0.8, mixed FCC+BCC phases were observed.When Al contents x>1.0-2.8 a single BCC structure was obtained due to spinodal decomposition.Hence, the structural transmission from FCC→FCC+BCC→BCC and also increased hardness value obtained.This is explained by higher atomic size of Al as compared to others and solid solution hardening and strong binding forces of Al with other metallic atoms [12].Chen, et al. reported that the hardness value increased with titanium content in Al 0.5 CoCrCuFeNiTi x .Upto 0.6 of Ti content, there was not much increase in wear resistance but, it increases rapidly with increasing the Ti content from 0.6 to 1.0 and reached a maximum at x=1.0 [9].
The relation between mixing enthalpy (∆H mix ) and atomic mismatch (δ) is clearly seen that, the value of δ increases by increasing the amount of element in composition whereas for the same compositions enthalpy decreases.Electronegativity (∆χ) variation is shown in Figure 1, and this indicates that for the value of ∆χ ≥ 0.138 FCC phase was not stable whereas for ∆χ ≤ 0.11 FCC and FCC+BCC phases are stable.Figure 2 confirms the relation between ∆H mix and Ω as in mathematical expression that ∆H mix and Ω are inversely proportional.
The critical value Ω=1 is proposed to form the solid solution.If Ω>1, the solid solution formation will be more and if Ω≤1 intermetallic compound formation and also segregation will exceed that of solid solution to form [13,31].The analysis done in this work of Ω also satisfies the condition given above.However, the exactness of these criterion requires more experimental analysis to make sure.Figure 2 shows the variation of different parameters with respect to valence electron concentration.This clearly indicates that, the beyond the 7.25 value of VEC, mixing enthalpy goes on increase.But, changes in cases of mixing entropy, atomic size mismatch and electronegativity variation in values with VEC value does not follows any specific pattern.Phase selection analysis for as cast high entropy alloys prepared by different researchers have been studied based on thermodynamic approach as well as geometric effect.Notably, the high mixing entropy is not only criteria to predict the BCC, BCC and FCC+BCC solid solution phases in equiatomic or non equiatomic multi component system.Some empirical conditions have been summarized here for the prediction of phases in HEAs.The mixing enthalpy, valence electron concentration, atomic size mismatch and
Figure 1 :
Figure 1: Variation of different parameters with Valence Electron Concentration (VEC).Note on the Legend: Fully Closed Symbols for Sole FCC, Fully Open Symbols for Sole BCC Phase; Right-Half Closed Symbols for Mixes FCC and BCC Phases
Figure 2 :
Figure 2: Variation of Ω with respect to mixing enthalpy (∆H mix ) for variety of high entropy alloys
Submit your next manuscript to Annex Publishers and benefit from: Submit your manuscript at http://www.annexpublishers.com/paper-submission.php→ Easy online submission process → Rapid peer review process → Open access: articles available free online → Online article availability soon after acceptance for Publication → Better discount on subsequent article submission → More accessibility of the articles to the readers/researchers within the field | 3,143 | 2017-12-01T00:00:00.000 | [
"Materials Science"
] |
The ultimate theoretical error on γ from B → DK decays
The angle γ of the standard CKM unitarity triangle can be determined from B → DK decays with a very small irreducible theoretical error, which is only due to second order electroweak corrections. We study these contributions and estimate that their impact on the γ determination is to introduce a shift |δγ| ≲ \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$ \mathcal{O}\left( {1{0^{-7 }}} \right) $\end{document}, well below any present or planned future experiment.
Introduction
The determination of the standard CKM unitarity triangle angle γ ≡ arg(−V ud V * ub /V cd V * cb ) from B → DK and B →DK decays is theoretically extremely clean. The reason is that the B → DK transitions receive contributions only from tree operators, and none from penguin operators. Furthermore, all the relevant matrix elements can be obtained from data if enough D-decay channels are measured. The sensitivity to γ comes from the interference of b → cūs and b → ucs decay amplitudes, which have a relative weak phase γ, cf. figure 1. These quark-level transitions mediate B − → D 0 K − and B − →D 0 K − decays, respectively. The D 0 andD 0 subsequently decay into a common final state f , which allows the two decay channels to interfere. Several variants of the method have been proposed, distinguished by the final state f : i) f can be a CP eigenstate such as K S π 0 and K S φ [1,2], ii) a flavor state such as K + π − and K * + ρ − [3,4], or iii) a multibody state such as K S π + π − , π + π − π 0 [5][6][7]. Other possibilities include the decays of neutral B mesons, B 0 and B s , [8][9][10][11][12][13], multibody B decays [14][15][16][17][18][19] and D * or D * * decays [20,21] (see also the reviews in [22][23][24] and the current combination of LHCb measurements in [25]).
The above set of methods has several sources of theoretical errors. Most of them can be reduced once more statistics becomes available. For instance, in the past the D → K S π + π − Dalitz plot needed to be modeled using a sum of Breit-Wigner resonances or using the Kmatrix formalism. Utilizing the data from entangled ψ(3770) → DD decays measured at CLEO-c [26] and BES-III, this uncertainty can in principle be completely avoided [6]. The related error is now statistics-dominated [27,28].
Other sources of reducible uncertainties are D −D mixing and K −K mixing (for final states with K S ). Both of these effects can be included trivially by modifying the expressions for the decay amplitudes, taking meson mixing into account, and then using experimentally measured mixing parameters [29]. The effect of D −D mixing is most significant if the D decay information comes from entangled ψ(3770) → DD decays. The shift in γ is then linear in x D , y D , giving ∆γ 2.9 • [30] (see also [31]). For flavor-tagged D JHEP01(2014)051 decays (i.e. from D * → Dπ) the effect is quadratic in x D , y D and thus much smaller [32]. The effect of K −K mixing in D decay modes involving a neutral kaon was discussed recently in [33]; it introduces a shift in the extraction of γ of order 10 −2 which can be systematically incorporated into the analysis. Similarly, for γ extraction from untagged B s → Dφ decays the inclusion of ∆Γ s can be important and can be achieved once ∆Γ s is well measured [34].
In the extraction of γ from B → DK, CP violation in the D system was usually neglected. Even if this assumption is relaxed, it is still possible to extract γ by appropriately modifying the expressions for the decay amplitudes (and using the fact that in Cabibboallowed D decays there is no direct CP violation 1 ) [25,[35][36][37][38][39].
Yet another source of reducible theory error are QED radiative corrections to the decay widths. The uncertainties from this source are expected to be below present experimental sensitivity on γ so that not much work has been done on them. Since the corrections are CP conserving they can be reabsorbed in the CP-even measured hadronic quantities and would not affect γ, as long as in the measurements the radiative corrections are treated consistently between different decay modes.
The first irreducible theory error on γ thus comes from higher-order electroweak corrections. This error cannot be eliminated using just experimental information and may well represent the ultimate precision of the γ determination from B → DK decays. The resulting uncertainty was estimated using scaling arguments in ref. [40] and found to be of the order of δγ/γ ∼ O(10 −6 ). In this paper we perform a more careful analysis, and find that the induced uncertainty is in fact most probably even an order of magnitude smaller. The one-loop electroweak corrections give rise to local and nonlocal contributions. We estimate the size of the local contributions using naive factorization and obtain δγ/γ O(10 −7 ). The nonlocal contributions are more difficult to estimate, but naively one expects that they are not significantly larger than the local ones.
The paper is organized as follows. In section 2 we give a brief discussion of electroweak corrections for B → DK decays with a focus on the γ extraction. We also give numerical estimates for the shift, δγ, utilizing the analytic results of section 3, where further details of the calculation are given. Finally, we conclude in section 4.
2 The shift in γ from B → DK due to electroweak corrections The measurement of γ in B → DK decays is based on the interference between the treelevel b → cūs and b → ucs mediated processes, cf. figure 1. The sensitivity to the weak phase γ enters through the amplitude ratio where the four-fermion operators are Above we have used the short-hand notation (cb) V −A (su) V −A ≡ cγ µ (1 − γ 5 )b sγ µ (1 − γ 5 )u , and similarly for the other quark flavors. The scale at which the Wilson coefficients are evaluated is close to the b quark mass, µ ∼ m b , with C 1 (m b ) = 1.10, and C 2 (m b ) = −0.24 at leading-log order [42], for m b (m b ) = 4.163 GeV [43] and α S (M Z ) = 0.1184 [44]. The decay amplitudes in eq. (2.1) are then given at leading order in the electroweak expansion by there are corrections to (2.1) and (2.6) from W box diagrams, and from vertex corrections, shown in figure 2, and from double penguin diagrams. In addition there are also self-energy diagrams for the W -propagator and wave function renormalization diagrams for external legs, which however have exactly the same CKM structure as the leading order contributions and thus do not affect the γ extraction. The same is true of the vertex corrections due to a Z or W loop, shown in figure 2 (right), which correct the CKM matrix at one-loop. The double penguin insertions are two-loop and are thus subleading, as can be easily checked from the small sizes of the respective Wilson coefficients. They are safely neglected in the following. The leading effect on extracted γ at O(G 2 F ) then comes from the box diagram in figure 2 (left). The dominant contribution is effectively due to the top and bottom quark running in the loop, as we show in the next section. The CKM structure of the box diagram is different from that of the O(G F ) tree contribution and is given, Since the weak phases of the two contributions are different, this results in a shift δγ in the extracted value of γ.
A similar higher-order electroweak diagram contributes also to the b → ucs transition, which is given by exchanging the external u and c quarks in figure 2 (left). Again, the dominant contribution is effectively due to the top and bottom quark running in the loop, so that the CKM factors are In this case the weak phases of the LO and NLO contributions are the same to a very good approximation, so that the electroweak contributions do not induce a shift in γ.
Keeping only the local part of the box diagram, the relevant change to the effective weak Hamiltonian is very simple. The structure of the CKM coefficients in (2.7) and (2.8) is such that all the corrections relevant for the γ extraction are in the Hc u effective weak Hamiltonian eq. (2.2), which at O(G 2 F ) takes the form The Wilson coefficients C 1,2 (µ) are the same Wilson coefficients as in eqs. (2.2) and (2.3), while ∆C 1,2 (µ) are calculable corrections. They depend on the CKM elements and carry a weak phase γ. They therefore have a different weak phase than C 1,2 (µ), which in our phase convention are real. This introduces a shift in δγ in the extraction of the weak phase γ from B → DK decays. This shift represent the ultimate theory error on the measurement of γ.
Defining the ratio of matrix elements for the two relevant operators where we expanded in the small corrections ∆C 1 , ∆C 2 to linear order. The resulting shift in the extracted value of γ is The size of the corrections ∆C 1,2 will be calculated in the next section, while here we only quote the numerical results. The unresummed result for Im(∆C 2 ), cf. eq. (3.5) below, is where the error only reflects the experimental errors due to the input parameters. The results with log(m b /M W ) resummed, cf. eq. (3.22) below, are (2.14) In order to obtain δγ we also need to estimate the ratio of the matrix elements, r A , in (2.10).
In naive factorization this ratio is where we used f D = 0.214 GeV [45], F B→K 0 (0) = 0.34 [46], f K = 0.16 GeV, F B→D 0 (0) = 1.12 [47]. In eq. (2.15) we only quote the central value, since the error on this estimate is bigger than the errors on the form factors themselves. However, we do not expect the error on the estimate of r A in (2.15) to be bigger than a factor of a few.
Using this and setting γ = 68 • for definiteness, we obtain the estimate for the shift δγ, δγ 2.0 · 10 −8 (2.16) where to this accuracy the resummed expressions for ∆C 1,2 (with nonlocal contributions neglected) and unresummed results coincide. An uncertainty of at most an additional factor of a few can be expected on the above estimate, so that we can conclude that the ultimate theoretical error on γ measurement is safely below |δγ| 10 −7 . (2.17) In the next section we derive the analytic expressions for ∆C 1,2 (µ), and then draw our conclusions in section 4.
Corrections to the electroweak Hamiltonian
In this section we consider the b → cūs box diagram, figure 2 (left), in detail. The results can be readily adapted to the b → ucs case by exchanging the external quarks and adjusting the CKM factors. The diagram in figure 2 (left) is superficially similar to the box diagrams contributing toK 0 − K 0 andB 0 (s) − B 0 (s) mixing [42], and to b → ssd, dds decays [48]. The difference is that the box diagram in figure 2 (left) has both up-and down-quarks running in the loop, in contrast to the case ofK 0 − K 0 andB 0 (s) − B 0 (s) mixing where both quarks in the loop are of up-type.
JHEP01(2014)051
We will calculate the shift δγ in two ways -first by keeping only the log(m b /M W ) enhanced local contribution, but without resumming it. Subsequently we will resum this log. In the first case we will take b, t and W in the loop to be heavy and integrate them out at µ ∼ M W . In this way one obtains the local operator part of the effective field theory (EFT) with only the light quarks, u, d, s, c, and an external non-dynamical b-quark field. Keeping only the local operators in EFT is a crude approximation that does, however, suffice for our purposes -to show that the induced corrections on the γ extraction are exceedingly small. The obtained result will also give us better understanding of the correct EFT results with resummed log(m b /M W ), which we will perform next. The resummation is achieved by first integrating out t and W at µ ∼ M W and matching onto the effective theory with b, and c, s, d, u quarks. We will then evolve the Wilson coefficients down to the scale µ ∼ m b using the renormalization-group (RG).
The result without resummations
We first evaluate the box diagram at µ ∼ M W , treating t and b quarks as massive and u, c and d, s quarks as massless, and set all external momenta to zero (including the external b-quark momentum). This will give us the local part of the EFT contributions with unresummed Wilson coefficients. Because of the double GIM mechanism, acting on both the internal up-quark and down-quark lines, the leading contribution is proportional to This is easy to see by expanding the matrix element for the box-diagram correction to the B → DK decay in terms of the quark masses, and readŝ Note that the loop functionĈ full (x, y) vanishes if either x → 0 or y → 0. This proves that the only nonzero contribution in (3.1) is A 4 ∝ x t y b . In fact, it is a very good approximation to keep in this result only the log y b enhanced contribution, where the finite terms amount to an O(10%) correction. Using the values for the CKM matrix elements from the CKMfitter collaboration [41] and further input from [44], we find 5) where the error shown is only due to the CKM elements. The Wilson coefficientĈ(x t , y b ) contains the unresummed large logarithm log y b . The logarithm is multiplied by 2y b and would vanish in the limit of zero b quark masses. However, since the Wilson coefficientĈ(x, y) starts only at O(y b ), the term with log y b represents a large correction. In the next subsection we therefore perform a resummation of this logarithm.
The resummed result
In order to resum log(m b /M W ) we need to explicitly keep the hierarchy of scales, m b M W , in the construction of the effective theories. For µ > M W one has the full SM, for JHEP01(2014)051 m b < µ < M W one has an effective theory with massless b and c, s, d, u quarks but no top quark, while below m b there is an effective theory with only the light quarks, c, s, d, u.
In the matching at µ ∼ M W the top quark and the W, Z bosons are integrated out, while the massless bottom quark is still a dynamical degree of freedom also in the effective theory -this is the main difference to the previous subsection. Integrating out the W at tree level in electroweak counting generates the effective Hamiltonians (2.2), (2.3) and its variants containing also the dynamical d-quark field. The contribution proportional to y b now vanishes at the electroweak scale to the order considered. However, this contribution will be generated by mixing of two insertions of dimension-six operators below the electroweak scale. It is therefore useful to write the Hamiltonian describing the five-flavor effective theory in the following way, where we used V tb V * ts = −V cb V * cs + O(λ 2 ), with λ = |V us | 0.23 (numerically, this replacement is valid up to a three-permil correction). Moreover, we denoted the usual four-quark operators by and definedQ The last two operators denoted by a tilde are formally of dimension eight because of the m 2 b factor. They have the same four-quark structure as the leading power operators Q 1,2 so that their contributions could be absorbed by redefining the Wilson coefficients C 1,2 allowing them to be complex. It is more practical, however, to keep the Wilson coefficients real and split-off explicitly the contributions to the effective Hamiltonian that carry the extra weak phase as we did in (3.6). Note that in the second line in eq. (3.6) we neglect all the O(G 2 F ) terms with the same weak phase as the O(G F ) terms in eqs. (2.2), (2.3), since these are not relevant for calculating δγ. We also neglect the six-quark operators which arise from integrating out the W boson and the top quark, as they are suppressed by an additional factor of 1/M 2 W . The dimension-eight Wilson coefficients at the electroweak scale vanish to leading order. The mixing of double insertions of dimension-six operators intoQ 1,2 will generate non-vanishing Wilson coefficientsC 1,2 (µ) below the electroweak scale. The inverse powers of g s in the definition ofQ 1,2 in (3.8) take into account that we will sum the leading logarithms proportional to the strong coupling constant.
JHEP01(2014)051
Let us now look at some of the contributing terms in more detail. The sum of the two diagrams denoted by 2) in figure 3 yields where Q 2 Q 2 div is the common divergence of the two diagrams, which is independent of the light-quark masses. In the last step we kept only the term proportional to the factor with a weak phase, which is the only contribution entering the shift δγ. The Lorentz and color structure of Q 2 Q 2 div is the same as ofQ 2 , so that this gives the anomalous dimension of the double insertion mixing intoQ 2 . The sum of the two diagrams denoted by 1) in figure 3 is similar to the first case, eq. (3.9), but with the replacement C 2 → C 1 , Q 2 → Q 1 . The sum of the two diagrams denoted by 3) in figure 3 yields and does not carry a weak phase. As such it does not contribute to δγ and can be discarded. There are also four additional diagrams, shown in figure 4, which lead to the mixing of double insertions into the Fierz-transformed operatorQ 1 .
To obtain the contributions of double H f =5 eff insertions to the running ofQ 1,2 we thus only need to compute the diagrams denoted by 1) and 2) in figure 3, with a double insertion of Q 1 and Q 2 , respectively, plus two additional diagrams with an insertion of Q 1 and then Q 2 at each of the two weak vertices, cf. figure 4. We expandγ i,j;k = αs 4πγ (0) i,j;k + . . ., where i, j denote the Q 1,2 insertions, and k is the labeling of theQ k operators. Extractingγ with all the remaining entries either vanishing or not contributing. The initial conditions for the dimension-six Wilson coefficients are given by C 1 (µ W ) = 1, C 2 (µ W ) = 0 to leading order [42]. ExpandingC k =C k (µ W ) = 0 at leading order. A nonvanishing value will be induced by RG running for µ < µ W , which we compute by solving where γ lk is the well-known anomalous dimension for the mixing of the Q 1,2 operators, (3.13)
JHEP01(2014)051
It is advantageous to go to the diagonal basis of the current-current operators, by defining (3.14) In this way eq. (3.12) gets rewritten as a homogeneous equation [51], for which the standard techniques of obtaining closed expressions for the RG evolution apply. The transformed LO anomalous dimensions and the Wilson coefficients are [50] where R = 1 2 By explicit calculation we findγ while the remaining entries are zero. Defining D + ≡ (C − ,C + /C − ) T , the renormalizationgroup equations forC + and C − can be combined into Here γ + = 4 and γ − = −8 are the eigenvalues of the matrix (3.13). We obtain the corresponding solution forC − and C + by exchanging the subscripts + ↔ −. Note that we have also included the running of the mass and the coupling constant related to the factor m 2 b /g 2 s in the definition of the operatorsQ k , given by the anomalous dimension of the quark mass γ m and the QCD beta function β. Transforming back to the original basis, we find numerically where we used α s (M Z ) = 0.1184 [44] and m b (m b ) = 4.163 GeV [43]. Note that the RG running has now also induced a nonzero correction to C 1 in (2.9), in contrast to the unresummed result. We used the mathematica package "RunDec" [52] for the numerical running of the strong coupling constant. Finally, at the bottom-quark scale we need to calculate the B → DK matrix elements using our EFT Hamiltonian (3.6) in order to obtain the shift δγ. This will give the leading y b behavior with resummed logarithms. We write the matrix elements suggestively as (3.20)
JHEP01(2014)051
Here we expand ∆C k = 4π αs ∆C (0) k +O(1); note that in this way the artificially inserted factor of 1/g 2 s in the definition ofQ k (3.8) is canceled. At LO it is not necessary to compute the double insertions Q i Q j since these are loop suppressed, and therefore we effectively obtain the matching condition for the Wilson coefficients of the local operators (2.9) Numerically, we find
Conclusions
The determination of the SM weak phase γ from the B → DK decays has a very small irreducible theoretical error which is due to one-loop electroweak corrections. In this paper we have estimated the resulting shift in γ. Treating m b ∼ M W or resumming logs of m b /M W gives in both cases an estimated shift δγ ∼ 2·10 −8 , keeping only the local operator contributions at the scale µ ∼ m b . It is unlikely that the neglected non-local contributions, which come with the same CKM suppression as the local contributions, would differ from the above estimate by more than a factor of a few. For instance, in K 0 −K 0 mixing the long distance contributions to ImM 12 are even an order of magnitude smaller than the short distance ones [53]. We can thus safely conclude that the irreducible theoretical error on the extraction of γ from B → DK is |δγ| O(10 −7 ). | 5,383 | 2014-01-01T00:00:00.000 | [
"Physics"
] |
Bidirectional CLLLC Resonant Converter Based on Frequency-Conversion and Phase-Shift Hybrid Control
: Due to the symmetrical structure, consistent working characteristics in the forward and reverse operations, and good soft-switching characteristics, bidirectional CLLLC resonant converters are widely used in electric vehicles and other fields. To meet the requirements of the on-board charger, this paper uses a bidirectional CLLLC resonant converter as the rear-stage of the on-board charger. The converter uses a constant voltage and constant current control in the forward operation and voltage and current double-closed-loop control in the reverse operation. Aiming at the problem whereby the voltage gain range of the bidirectional CLLLC resonant converter is relatively narrow under frequency-conversion control, the paper proposed a hybrid control method of frequency conversion and phase shift. Frequency-conversion control is used when the voltage gain is high, and phase-shift control is used when the voltage gain is low. The output voltage range of the converter is effectively broadened, and zero voltage switching and zero current switching can be realized in the full load range to improve operating efficiency. A 3.3 kW bidirectional CLLLC resonant converter simulation system is built in the simulation software. The simulation results verify the correctness and feasibility of the designed circuit and control method.
Introduction
With the development of the economy, the problems of environmental pollution and energy shortage are becoming more and more serious.Electric vehicles that are low-pollution and low-cost have become the main development direction of new energy vehicles, and have been vigorously supported and promoted by the state [1,2].As one of the core technologies of electric vehicles, the on-board charger (OBC) has also become a research hotspot.Its main part is generally composed of the front-end AC/DC converter and the rear-end DC/DC converter [3].The bidirectional DC/DC converter is the key part to realizing the power conversion and energy exchange [4,5].
The topology of a bidirectional DC/DC converter is mainly divided into the nonisolated bidirectional DC/DC converter and the isolated bidirectional DC/DC converter [6].The non-isolated bidirectional DC/DC converter topology mainly includes buck/boost, buck-boost, Ćuk, and Sepic-Zeta, etc.It has the advantages of a simple structure, small size, and low cost, but it has no electrical isolation, poor equipment stability, and low transmission efficiency [7].The isolated bidirectional DC/DC converter is generally divided into the non-resonant and resonant modes.The most typical non-resonant bidirectional DC/DC converter is the dual-active-bridge DC/DC converter (DAB) [8,9].Its structure is simple: the primary and secondary sides are electrically isolated, and zero-voltage soft-switching (ZVS) characteristics can be achieved, but there are problems such as power backflow, high current stress, and failure to achieve soft-switching at a light load [10,11].Compared with the DAB, the resonant DC/DC converter can achieve zero-voltage switching of the primary switches and zero-current switching (ZCS) of the secondary switches in a wide voltage range and a wide load range at the same time, which can reduce losses and achieve a higher voltage gain and efficiency [12][13][14].Among them, the LLC bidirectional resonant converter has been widely studied because of its simple circuit structure and easy realization of soft switching [15].However, the voltage gain adjustment range of the traditional LLC bidirectional resonant converter is not wide, and it is limited in the application of a wide input voltage or wide output voltage.It is equivalent to an LC series resonant converter when working in the reverse.The resonant state is inconsistent with the forward working mode.The voltage gain is always less than 1, and the boost cannot be achieved [16,17].
The L-LLC bidirectional resonant converter proposed in Reference [18] adds an auxiliary inductor between the arms of the H-bridge on the input side of the traditional LLC bidirectional resonant converter so that the state of the converter is consistent in the forward and reverse operations, but the increase of the auxiliary inductor reduces the overall efficiency of the converter.Reference [19] proposed a boost-LCL resonant converter, which improves the voltage gain and efficiency by multiplexing the LCL resonant inductor, the flyback primary inductor, and the boost inductor.However, the structure of the converter is complex and it is not easy to achieve an optimal design.In [20], a novel bidirectional isolated LLC resonant converter is proposed.The inverter and rectifier parts of the converter have two modes by reconfiguring different modulation schemes, which can achieve an ultra-wide voltage gain range in a narrow modulation frequency, but there are many switching elements.In [21], a symmetrical CLLC resonant converter is proposed.The converter adds a resonant cavity on the secondary side of the traditional LLC resonant converter, which has a completely symmetrical structure.It has the characteristics of the LLC's natural soft switching and high power density, but the increase of resonant elements will make the gain curve non-monotonic and lose the ability of ZVS in the capacitive interval.In [22], a new topology based on the LLC and LCL-T resonant tank is proposed, because the output voltage characteristics of the LLC resonant tank and the fixed output current characteristics of the LCL-T resonant tank have high accuracy, which can greatly reduce the operating switching-frequency range.Reference [23] proposed a variable frequency-phaseshift control method for the LLC resonant converter applied to electric vehicle charging, so that the LLC converter can always operate within the range of EV charging requirements and realize the functions of step-down and step-up, and soft switching.Reference [24] proposed a wide-output dual-full-bridge LLC resonant converter with a hybrid control strategy.However, it can only work in the forward direction.Reference [25] proposes a bidirectional LLC resonant converter with a pulse frequency modulation and phase-shift combined control that can work in the forward and reverse direction.However, its forward and reverse mode are inconsistent, and the control mode only uses the voltage single loop.
In this paper, a hybrid control strategy of frequency conversion and phase shift for a CLLLC resonant converter is proposed.Frequency-conversion control is used when the boost is needed, and the voltage gain is determined by the switching frequency.Phase-shift control is used when the buck is needed, and the voltage gain is determined by the phaseshift angle.For the CLLLC resonant converter, the fundamental wave analysis method is used to optimize the parameter design.Based on satisfying the voltage gain characteristics, the constant voltage and constant current control are used in the forward operation, and the voltage and current double-closed-loop control are used in the reverse operation [26].The CLLLC resonant converter can achieve soft switching in the full load range during forward charging and reverse discharging, thereby improving the working efficiency.
Birirectional CLLLC Resonant Converter
The topology of the bidirectional CLLLC resonant converter is shown in Figure 1.The switching tubes S 1 ~S4 constitute the primary-side full bridge, and S 5 ~S8 constitute the secondary-side full bridge.D 1 ~D8 are the body diodes of S 1 ~S8 , and C coss1 ~Ccoss8 are the output capacitors of S 1 ~S8 .The ratio of the transformer is n:1; L m is the excitation inductance of the transformer; L r1 , L r2 , C r1 , and C r2 are the resonant inductance and resonant capacitance of the primary side and the secondary side, respectively, and the high-frequency transformer is used for electrical isolation.U AB is the primary voltage, U CD is the secondary voltage, and C in and C o are the filter capacitors.When the converter is working forward, the power is transmitted from U AB to U CD .Since the converter operates in the same state when it is working in the forward and reverse operations, only the positive working time is taken as an example for the analysis.The topology of the bidirectional CLLLC resonant converter is shown in Figure 1.The switching tubes S1~S4 constitute the primary-side full bridge, and S5~S8 constitute the secondary-side full bridge.D1~D8 are the body diodes of S1~S8, and Ccoss1~Ccoss8 are the output capacitors of S1~S8.The ratio of the transformer is n:1; Lm is the excitation inductance of the transformer; Lr1, Lr2, Cr1, and Cr2 are the resonant inductance and resonant capacitance of the primary side and the secondary side, respectively, and the high-frequency transformer is used for electrical isolation.UAB is the primary voltage, UCD is the secondary voltage, and Cin and Co are the filter capacitors.When the converter is working forward, the power is transmitted from UAB to UCD.Since the converter operates in the same state when it is working in the forward and reverse operations, only the positive working time is taken as an example for the analysis.The left and right sides are full-bridge structures, and the structures are symmetrical to each other.In both the forward and reverse modes, energy can be transferred to achieve a bidirectional flow of energy, and the converter acts as an LLC resonant converter in both the forward and reverse operations.When the converter is working in the forward operation, switches S1~S4 obtain the driving signal to realize the inverter and S5~S8 realize the rectification.
Working Characteristics of Frequency-Conversion Control
The fundamental analysis method is used to analyze the operating characteristics of the converter, and the fundamental equivalent circuit is shown in Figure 2.Among them, VAB_FHA(t) is the square-wave fundamental component of the input DC-bus voltage after the full-bridge inverter.Lr2′ and Cr2′ are the values of the capacitance and inductance converted from Lr2 and Cr2 to the primary side of the transformer.Req is the equivalent load converted to the primary side.The expression is: where n is the transformer ratio; Vout, Iout, and Pout are the output voltage, output current, and output power, respectively.The left and right sides are full-bridge structures, and the structures are symmetrical to each other.In both the forward and reverse modes, energy can be transferred to achieve a bidirectional flow of energy, and the converter acts as an LLC resonant converter in both the forward and reverse operations.When the converter is working in the forward operation, switches S 1 ~S4 obtain the driving signal to realize the inverter and S 5 ~S8 realize the rectification.
Working Characteristics of Frequency-Conversion Control
The fundamental analysis method is used to analyze the operating characteristics of the converter, and the fundamental equivalent circuit is shown in Figure 2.Among them, V AB_FHA(t) is the square-wave fundamental component of the input DC-bus voltage after the full-bridge inverter.L r2 and C r2 are the values of the capacitance and inductance converted from L r2 and C r2 to the primary side of the transformer.R eq is the equivalent load converted to the primary side.The expression is: where n is the transformer ratio; V out , I out , and P out are the output voltage, output current, and output power, respectively.
Electronics 2023, 12, x FOR PEER REVIEW 3 of 16 The topology of the bidirectional CLLLC resonant converter is shown in Figure 1.The switching tubes S1~S4 constitute the primary-side full bridge, and S5~S8 constitute the secondary-side full bridge.D1~D8 are the body diodes of S1~S8, and Ccoss1~Ccoss8 are the output capacitors of S1~S8.The ratio of the transformer is n:1; Lm is the excitation inductance of the transformer; Lr1, Lr2, Cr1, and Cr2 are the resonant inductance and resonant capacitance of the primary side and the secondary side, respectively, and the high-frequency transformer is used for electrical isolation.UAB is the primary voltage, UCD is the secondary voltage, and Cin and Co are the filter capacitors.When the converter is working forward, the power is transmitted from UAB to UCD.Since the converter operates in the same state when it is working in the forward and reverse operations, only the positive working time is taken as an example for the analysis.The left and right sides are full-bridge structures, and the structures are symmetrical to each other.In both the forward and reverse modes, energy can be transferred to achieve a bidirectional flow of energy, and the converter acts as an LLC resonant converter in both the forward and reverse operations.When the converter is working in the forward operation, switches S1~S4 obtain the driving signal to realize the inverter and S5~S8 realize the rectification.
Working Characteristics of Frequency-Conversion Control
The fundamental analysis method is used to analyze the operating characteristics of the converter, and the fundamental equivalent circuit is shown in Figure 2.Among them, VAB_FHA(t) is the square-wave fundamental component of the input DC-bus voltage after the full-bridge inverter.Lr2′ and Cr2′ are the values of the capacitance and inductance converted from Lr2 and Cr2 to the primary side of the transformer.Req is the equivalent load converted to the primary side.The expression is: where n is the transformer ratio; Vout, Iout, and Pout are the output voltage, output current, and output power, respectively.When the voltage at both ends of L m is clamped by the output voltage, the corresponding resonant frequency is the first resonant frequency: When L m is no longer clamped by the output, the corresponding resonant frequency is the second resonant frequency: We define impedance as follows: where Z 1 is the primary-side series-resonant impedance; Z 2 is the equivalent secondary series-resonant impedance; Z m is the excitation inductance impedance; and w s is the switching angular frequency, w s = 2π f s .
From Figure 2, the transfer function of the converter is as follows: By substituting Equation (4) into Equation ( 5), the voltage gain of the converter can be obtained: where k is the inductance coefficient, k = L m /L r1 ; Q is the quality factor, Q = √ L r1 /C r1 /R eq ; f n is the normalized frequency; f n = f s / f r1 ; and f s is the actual switching frequency.
It can be seen from Equation (6) that the voltage gain of the CLLLC resonant converter includes three parameters: the inductance coefficient k, the quality factor Q, and the normalized switching frequency f n .The main parameters affecting the voltage gain are the inductance coefficient k and quality factor Q, which are analyzed in detail below: (1) Influence of inductance factor k on voltage gain Figure 3 is the curve of the voltage gain and normalized frequency under different inductance coefficients k when Q = 0.3.When f n = 1, the resonant frequency is equal to the switching frequency, the system works at the resonant point, and no matter how k changes, the voltage gain is always 1.In the under-resonant region, the smaller the k value, the greater the maximum voltage gain, and the narrower the frequency modulation range, which is conducive to the adjustment of the wide gain range.As the k value increases, the peak value of the converter voltage gain gradually decreases, the normalized frequency corresponding to the maximum gain of the curve gradually decreases, and the frequency modulation range becomes wider.In the over-resonant region, the voltage gain curve becomes smooth and the effect of adjusting the voltage gain by adjusting the switching frequency is very limited.If the gain is expected to remain unchanged in a wide range around the resonance point, then k should take a larger value.However, when the k value is greater than a certain value, the peak voltage gain of the converter will be less than the maximum value required by the design, and the minimum switching frequency is too low.(2) Influence of quality factor Q on voltage gain Figure 4 shows the curves of the voltage gain and normalized frequency under different quality factors Q when k = 5.When fn = 1, the resonant frequency is equal to the switching frequency, and the system works at the resonant point; no matter how Q changes, the voltage gain is always 1.When the converter is at a light load, that is, the Q value is small, the gain of the converter can be greatly improved when the switching frequency is reduced in the under-resonant region.In the over-resonant region, when the switching frequency increases, the converter gain will only decrease slightly, and the frequency adjustment effect is low.To achieve the effect of voltage regulation, a high frequency is required, and the switching loss increases.Therefore, the selection of the Q value needs to ensure that the converter can meet the requirements of a minimum voltage gain at a light load and that the maximum switching frequency cannot be too high.As the Q value increases, the load becomes heavier and heavier.In the over-resonant region, the gain of the converter changes with the switching frequency.However, when the load is too heavy, the peak value of the maximum gain of the converter decreases in the underresonant region.To obtain the maximum voltage gain, the switching frequency needs to be reduced to broaden the frequency modulation.The voltage gain curve has two peak points, which are not conducive to the stability of the converter and the design of the control system.Therefore, the selection of the Q value should also be compromised.(2) Influence of quality factor Q on voltage gain Figure 4 shows the curves of the voltage gain and normalized frequency under different quality factors Q when k = 5.When f n = 1, the resonant frequency is equal to the switching frequency, and the system works at the resonant point; no matter how Q changes, the voltage gain is always 1.When the converter is at a light load, that is, the Q value is small, the gain of the converter can be greatly improved when the switching frequency is reduced in the under-resonant region.In the over-resonant region, when the switching frequency increases, the converter gain will only decrease slightly, and the frequency adjustment effect is low.To achieve the effect of voltage regulation, a high frequency is required, and the switching loss increases.Therefore, the selection of the Q value needs to ensure that the converter can meet the requirements of a minimum voltage gain at a light load and that the maximum switching frequency cannot be too high.As the Q value increases, the load becomes heavier and heavier.In the over-resonant region, the gain of the converter changes with the switching frequency.However, when the load is too heavy, the peak value of the maximum gain of the converter decreases in the under-resonant region.To obtain the maximum voltage gain, the switching frequency needs to be reduced to broaden the frequency modulation.The voltage gain curve has two peak points, which are not conducive to the stability of the converter and the design of the control system.Therefore, the selection of the Q value should also be compromised.(2) Influence of quality factor Q on voltage gain Figure 4 shows the curves of the voltage gain and normalized frequency under different quality factors Q when k = 5.When fn = 1, the resonant frequency is equal to the switching frequency, and the system works at the resonant point; no matter how Q changes, the voltage gain is always 1.When the converter is at a light load, that is, the Q value is small, the gain of the converter can be greatly improved when the switching frequency is reduced in the under-resonant region.In the over-resonant region, when the switching frequency increases, the converter gain will only decrease slightly, and the frequency adjustment effect is low.To achieve the effect of voltage regulation, a high frequency is required, and the switching loss increases.Therefore, the selection of the Q value needs to ensure that the converter can meet the requirements of a minimum voltage gain at a light load and that the maximum switching frequency cannot be too high.As the Q value increases, the load becomes heavier and heavier.In the over-resonant region, the gain of the converter changes with the switching frequency.However, when the load is too heavy, the peak value of the maximum gain of the converter decreases in the underresonant region.To obtain the maximum voltage gain, the switching frequency needs to be reduced to broaden the frequency modulation.The voltage gain curve has two peak points, which are not conducive to the stability of the converter and the design of the control system.Therefore, the selection of the Q value should also be compromised.For the primary-side switching tube of the converter, after the end of the half-cycle, there is still an inductor current i Lr flowing through the resonant circuit, which can discharge and charge the parasitic capacitance C oss of the switching tube.It is assumed that the current flowing through the switching tube remains unchanged as i Lr within the dead time t d , and the parasitic capacitance C oss of the switching tube is equal.To achieve ZVS conditions, i Lr must be greater than the minimum current for charging and discharging C oss during the dead time, that is:
Working Characteristics of Phase-Shift Control
When the converter adopts phase-shift control, due to the introduction of the phaseshift angle ϕ, the voltage of the input resonant cavity U AB changes from a square wave with a duty cycle of 50% to a square wave with a duty cycle of less than 50% and a zero level.From the fundamental equivalent model and Fourier analysis shown in Figure 2, the fundamental component of the input voltage square wave under phase-shift control is expressed as: where D is the phase shift, D = ϕ/π, 0 < D < 1; and ϕ is the phase-shift angle.
Combining Equations ( 5)-( 8), the voltage gain expression under phase-shift control is: It can be seen that the voltage gain is approximately cosine with D, and the variation range is between 0 and 1.
In the 0.5(1 − D)T s period, the L m2 two-terminal voltage is U 2 .From U Lm2 = L m2 (di Lm2 /dt) and Equation (7), it is necessary to ensure that the voltage at both ends of C oss meets U Coss > U 2 in the dead time td to achieve soft switching.Further, the conditions for soft switching under phase-shift control are:
Resonant Network Parameter Design
The design specifications of the bidirectional CLLLC resonant converter are shown in Table 1.(1) Turn ratio of transformer n At the resonant frequency point of the bidirectional resonant converter, the voltage gain of the converter is always 1, which is independent of the change of the circuit parameters and the load.The rated operating point is usually set here, so the turn ratio n of the transformer is: (2) Transformer normalized voltage gain M The normalized maximum voltage gain of this converter is: The normalized minimum voltage gain of this converter is: (3) Parameter design of inductance coefficient k and quality factor Q Because the phase-shift control mode is adopted in the output low voltage, the limit condition of k is no longer applicable to the no-load condition of the minimum output voltage corresponding to the highest input voltage in the traditional design method.Considering that the maximum output voltage corresponding to the lowest input voltage is fully loaded as the limit condition of the k value, the range of the inductance coefficient is: Although increasing the k value can reduce the loss of the loop, it will increase the range of the switching frequency.In combination with Equation ( 14) and Figure 3, we finally choose k = 5.
In combination with Equation (15) and Figure 4, Q = 0.3 is selected.At this time, the requirements of maximum gain and minimum gain can be met by smaller frequency adjustments, and the monotonicity can be maintained.
(4) Design of resonant inductor and resonant capacitor
The equivalent impedance of the converter is: According to Equation (8) through Equation ( 16), the parameters of the primary and secondary resonant inductors and resonant capacitors can be obtained, as shown in Table 2.
Frequency-Conversion and Phase-Shift Hybrid Control Strategy
Combining Equations ( 6), (9), and (10), the relationship curve of the voltage gain with the normalized frequency f n (under frequency-conversion control) and the phase-shift ratio D (under phase-shift control) can be obtained, as shown in Figure 5.After the quality factor is selected as 0.3, when the bidirectional CLLLC resonant converter is controlled by frequency conversion alone, the voltage gain range is narrow in the step-down mode, and the switching frequency needs to change greatly to change the voltage gain.When phase-shift control is used alone, the maximum voltage gain is only 1, which can only be realized in the step-down mode.The minimum voltage gain in the phase-shift mode is much smaller than the minimum voltage gain under frequency-conversion control.Therefore, this paper proposes a method of frequency-conversion and phase-shift hybrid control applied to a bidirectional CLLLC resonant converter to broaden the voltage gain range.At the resonance point, the working conditions of frequency-conversion control and phase-shift control are the same.Setting this as a switching point can achieve a seamless connection of the voltage gain; that is, frequency-conversion control is adopted when M > 1, and phase-shift control is adopted when M < 1.
Frequency-Conversion and Phase-Shift Hybrid Control Strategy
Combining Equations ( 6), (9), and (10), the relationship curve of the voltage gain with the normalized frequency fn (under frequency-conversion control) and the phase-shift ratio D (under phase-shift control) can be obtained, as shown in Figure 5.After the quality factor is selected as 0.3, when the bidirectional CLLLC resonant converter is controlled by frequency conversion alone, the voltage gain range is narrow in the step-down mode, and the switching frequency needs to change greatly to change the voltage gain.When phaseshift control is used alone, the maximum voltage gain is only 1, which can only be realized in the step-down mode.The minimum voltage gain in the phase-shift mode is much smaller than the minimum voltage gain under frequency-conversion control.Therefore, this paper proposes a method of frequency-conversion and phase-shift hybrid control applied to a bidirectional CLLLC resonant converter to broaden the voltage gain range.At the resonance point, the working conditions of frequency-conversion control and phaseshift control are the same.Setting this as a switching point can achieve a seamless connection of the voltage gain; that is, frequency-conversion control is adopted when M > 1, and phase-shift control is adopted when M < 1.The bidirectional CLLLC resonant converter designed in this paper is applied to the rear stage of the vehicle charger.The charging process of the charging vehicle generally has two stages: constant current charging and constant voltage charging.The charging curve is shown in Figure 6.The first stage is the constant-current-charging stage.At this stage, a certain constant current is used for constant current charging.The voltage values of the battery are relatively small at the beginning, far less than the set value.The voltage loop is always saturated.At this time, only the current loop plays a regulatory role.The voltage value continues to rise during the charging process.When the voltage rises to the set value, it enters the second stage.The second stage is the constant-voltage-charging stage.At this stage, the voltage loop exits the saturation state and the output value begins to decrease.Until it is lower than the output value of the current loop regulator, it begins to be regulated by the voltage loop and enters the constant-voltage-charging stage.At this time, the current loop is saturated, and the charging current begins to decrease gradually until the charging current of the battery is equal to the self-discharge current, and the whole charging process ends.The bidirectional CLLLC resonant converter designed in this paper is applied to the rear stage of the vehicle charger.The charging process of the charging vehicle generally has two stages: constant current charging and constant voltage charging.The charging curve is shown in Figure 6.The first stage is the constant-current-charging stage.At this stage, a certain constant current is used for constant current charging.The voltage values of the battery are relatively small at the beginning, far less than the set value.The voltage loop is always saturated.At this time, only the current loop plays a regulatory role.The voltage value continues to rise during the charging process.When the voltage rises to the set value, it enters the second stage.The second stage is the constant-voltage-charging stage.At this stage, the voltage loop exits the saturation state and the output value begins to decrease.Until it is lower than the output value of the current loop regulator, it begins to be regulated by the voltage loop and enters the constant-voltage-charging stage.At this time, the current loop is saturated, and the charging current begins to decrease gradually until the charging current of the battery is equal to the self-discharge current, and the whole charging process ends.Figure 7 is the control block diagram of using the frequency-conversion and phaseshift hybrid control mode.In the whole control process, the load uses the resistance load, and the output voltage and current are sampled as the judgment basis.To realize the au- Figure 7 is the control block diagram of using the frequency-conversion and phaseshift hybrid control mode.In the whole control process, the load uses the resistance load, and the output voltage and current are sampled as the judgment basis.To realize the automatic switching of the constant current and constant voltage loop, the voltage loop and the current loop are controlled in parallel when working forward.Comparing the actual battery voltage value obtained by sampling with the set value U ref , the constant-currentcontrol mode is selected when the battery voltage is less than the set value U ref , and the constant-voltage-control mode is selected when the battery voltage reaches the set value U ref .When working in the reverse direction, the voltage and current double-closed-loop control method are adopted to output the electric energy in the battery to the primary side through the resonant cavity to obtain a stable 400 V voltage.Figure 7 is the control block diagram of using the frequency-conversion and phaseshift hybrid control mode.In the whole control process, the load uses the resistance load, and the output voltage and current are sampled as the judgment basis.To realize the automatic switching of the constant current and constant voltage loop, the voltage loop and the current loop are controlled in parallel when working forward.Comparing the actual battery voltage value obtained by sampling with the set value Uref, the constant-currentcontrol mode is selected when the battery voltage is less than the set value Uref, and the constant-voltage-control mode is selected when the battery voltage reaches the set value Uref.When working in the reverse direction, the voltage and current double-closed-loop control method are adopted to output the electric energy in the battery to the primary side through the resonant cavity to obtain a stable 400 V voltage.Figure 8 is the hysteresis-control-switching diagram of the bidirectional CLLLC resonant converter.This paper uses the two control modes of frequency-conversion control and phase-shift control.In order to avoid unstable oscillation, the switching of these two modes should be set to hysteresis control, rather than switching at a single point.When Figure 8 is the hysteresis-control-switching diagram of the bidirectional CLLLC resonant converter.This paper uses the two control modes of frequency-conversion control and phase-shift control.In order to avoid unstable oscillation, the switching of these two modes should be set to hysteresis control, rather than switching at a single point.When the voltage rises to the maximum set value of phase-shift control, it is switched to frequency-conversion control.When the voltage drops to the minimum set value of frequency-conversion control, phase-shift control is switched.When setting the value of the mode-switching point, the maximum setting value of phase-shift control is higher than the minimum setting value of frequency-conversion control.The existence of the buffer region can make the converter switch reliably between the two control modes.
the voltage rises to the maximum set value of phase-shift control, it is switched to frequency-conversion control.When the voltage drops to the minimum set value of frequency-conversion control, phase-shift control is switched.When setting the value of the mode-switching point, the maximum setting value of phase-shift control is higher than the minimum setting value of frequency-conversion control.The existence of the buffer region can make the converter switch reliably between the two control modes.
Forward Simulation
When the converter is working in the forward direction, the input voltage remains unchanged at 400 V, and the output voltage is gradually increased from 250 V to 430 V by adjusting the phase-shift angle between S1 and S2, and S4 and S3, or adjusting the switching frequency of S1, S2, S3, and S4.In
Forward Simulation
When the converter is working in the forward direction, the input voltage remains unchanged at 400 V, and the output voltage is gradually increased from 250 V to 430 V by adjusting the phase-shift angle between S 1 and S 2 , and S 4 and S 3 , or adjusting the switching frequency of S 1 , S 2 , S 3 , and S 4 .In Figures 9-11, V gs and V ds represent the driving signal and drain-source voltage of the corresponding switching tube, respectively.i D represents the forward current of the rectifier diode D, V AB is the resonant slot input voltage, i Lr is the resonant current, and i Lm is the excitation inductance current.
Figure 9 shows the simulation results when the output voltage V o and the output current I o are 250 V/7.67A in the constant current mode.At this time, the output power P o = 1917.5W, which is the lightest working point of the converter during the constant current operation.It can be seen from Figure 9a that the converter operates in phase-shift mode, and the output voltage is controlled by adjusting the phase-shift angle between S 1 and S 4 , and S 2 and S 3 .It can be seen from Figure 9b that the drain voltage V ds4 of S 4 has been reduced to zero before the arrival of the driving signal V gs4 of the switching tube S 4 .It can be seen that the primary-side switch can achieve ZVS.Before the arrival of i D5 and i D8 , i D6 and i D7 have been reduced to zero.Similarly, before the arrival of i D6 and i D7 , i D5 and i D8 have also been reduced to zero, so the secondary rectifier diode can achieve ZCS. the voltage rises to the maximum set value of phase-shift control, it is switched to frequency-conversion control.When the voltage drops to the minimum set value of frequency-conversion control, phase-shift control is switched.When setting the value of the mode-switching point, the maximum setting value of phase-shift control is higher than the minimum setting value of frequency-conversion control.The existence of the buffer region can make the converter switch reliably between the two control modes.
Forward Simulation
When the converter is working in the forward direction, the input voltage remains unchanged at 400 V, and the output voltage is gradually increased from 250 V to 430 V by adjusting the phase-shift angle between S1 and S2, and S4 and S3, or adjusting the switching frequency of S1, S2, S3, and S4.In Figure 9 shows the simulation results when the output voltage Vo and the output current Io are 250 V/7.67A in the constant current mode.At this time, the output power Po = 1917.5W, which is the lightest working point of the converter during the constant current operation.It can be seen from Figure 9a that the converter operates in phase-shift mode, and the output voltage is controlled by adjusting the phase-shift angle between S1 and S4, and S2 and S3.It can be seen from Figure 9b that the drain voltage Vds4 of S4 has been reduced to zero before the arrival of the driving signal Vgs4 of the switching tube S4.It can be seen that the primary-side switch can achieve ZVS.Before the arrival of iD5 and iD8, iD6 and iD7 have been reduced to zero.Similarly, before the arrival of iD6 and iD7, iD5 and iD8 have also been reduced to zero, so the secondary rectifier diode can achieve ZCS.
Figure 10 shows the simulation results of the converter in the constant current mode when the output voltage Vo and the output current Io are 320 V/7.67 A. At this time, the output power Po = 2454.4W. As can be seen from Figure 10a, the converter works in the frequency-conversion mode, the switching frequency is 100 kHz, and the primary-side resonant current changes in the form of a sine wave.At this time, the converter works at the rated operating point and is in a quasi-resonant state.It can be seen from Figure 10b Figure 9 shows the simulation results when the output voltage Vo and the output current Io are 250 V/7.67A in the constant current mode.At this time, the output power Po = 1917.5W, which is the lightest working point of the converter during the constant current operation.It can be seen from Figure 9a that the converter operates in phase-shift mode, and the output voltage is controlled by adjusting the phase-shift angle between S1 and S4, and S2 and S3.It can be seen from Figure 9b that the drain voltage Vds4 of S4 has been reduced to zero before the arrival of the driving signal Vgs4 of the switching tube S4.It can be seen that the primary-side switch can achieve ZVS.Before the arrival of iD5 and iD8, iD6 and iD7 have been reduced to zero.Similarly, before the arrival of iD6 and iD7, iD5 and iD8 have also been reduced to zero, so the secondary rectifier diode can achieve ZCS.
Figure 10 shows the simulation results of the converter in the constant current mode when the output voltage Vo and the output current Io are 320 V/7.67 A. At this time, the output power Po = 2454.4W. As can be seen from Figure 10a, the converter works in the frequency-conversion mode, the switching frequency is 100 kHz, and the primary-side resonant current changes in the form of a sine wave.At this time, the converter works at the rated operating point and is in a quasi-resonant state.It can be seen from Figure Figure 10 shows the simulation results of the converter in the constant current mode when the output voltage V o and the output current I o are 320 V/7.67 A. At this time, the output power P o = 2454.4W. As can be seen from Figure 10a, the converter works in the frequency-conversion mode, the switching frequency is 100 kHz, and the primary-side resonant current changes in the form of a sine wave.At this time, the converter works at the rated operating point and is in a quasi-resonant state.It can be seen from Figure 10b that before the arrival of the driving signal V gs4 of the switching tube S 4 , the drain voltage V ds4 of S 4 has dropped to zero, which shows that the primary-side switching tube can achieve ZVS.Since the resonant current and the excitation current are equal only for a moment, i D6 and i D7 are generated at the moment i D5 and i D8 are reduced to zero.Similarly, i D5 and i D8 are generated at the moment when i D6 and i D7 are reduced to zero, so the secondary rectifier diode can achieve ZCS.
Figure 11 shows the simulation results when the output voltage V o and the output current I o are 430 V/7.67 A. At this time, the output power P o = 3300 W. It can be seen from Figure 11a that the converter works in the frequency-conversion mode, and the switching frequency is 67 kHz.At this time, the converter works in an under-resonant state.It can be seen from Figure 11b that the primary-side switch can achieve ZVS, and the secondary-side rectifier diode can achieve ZCS.
To verify the performance of the proposed bidirectional CLLLC resonant converter under constant current and constant voltage charging, the closed-loop simulation of the converter is carried out.The simulation results are shown in Figures 12 and 13.
achieve ZVS.Since the resonant current and the excitation current are equal only for a moment, iD6 and iD7 are generated at the moment iD5 and iD8 are reduced to zero.Similarly, iD5 and iD8 are generated at the moment when iD6 and iD7 are reduced to zero, so the secondary rectifier diode can achieve ZCS.
Figure 11 shows the simulation results when the output voltage Vo and the output current Io are 430 V/7.67 A. At this time, the output power Po = 3300 W. It can be seen from Figure 11a that the converter works in the frequency-conversion mode, and the switching frequency is 67 kHz.At this time, the converter works in an under-resonant state.It can be seen from Figure 11b that the primary-side switch can achieve ZVS, and the secondaryside rectifier diode can achieve ZCS.
To verify the performance of the proposed bidirectional CLLLC resonant converter under constant current and constant voltage charging, the closed-loop simulation of the converter is carried out.The simulation results are shown in Figures 12 and 13 Figure 12 is the closed-loop simulation waveform of the converter in the constant current stage.When the converter is in the constant current stage, the voltage loop is saturated.At this time, the current loop works.The output current remains constant at 7.67 A throughout the constant current stage, and the output voltage rises from 250 V to 430 achieve ZVS.Since the resonant current and the excitation current are equal only for a moment, iD6 and iD7 are generated at the moment iD5 and iD8 are reduced to zero.Similarly, iD5 and iD8 are generated at the moment when iD6 and iD7 are reduced to zero, so the secondary rectifier diode can achieve ZCS.
Figure 11 shows the simulation results when the output voltage Vo and the output current Io are 430 V/7.67 A. At this time, the output power Po = 3300 W. It can be seen from Figure 11a that the converter works in the frequency-conversion mode, and the switching frequency is 67 kHz.At this time, the converter works in an under-resonant state.It can be seen from Figure 11b that the primary-side switch can achieve ZVS, and the secondaryside rectifier diode can achieve ZCS.
To verify the performance of the proposed bidirectional CLLLC resonant converter under constant current and constant voltage charging, the closed-loop simulation of the converter is carried out.The simulation results are shown in Figures 12 and 13 Figure 12 is the closed-loop simulation waveform of the converter in the constant current stage.When the converter is in the constant current stage, the voltage loop is saturated.At this time, the current loop works.The output current remains constant at 7.67 A throughout the constant current stage, and the output voltage rises from 250 V to 430 Figure 12 is the closed-loop simulation waveform of the converter in the constant current stage.When the converter is in the constant current stage, the voltage loop is saturated.At this time, the current loop works.The output current remains constant at 7.67 A throughout the constant current stage, and the output voltage rises from 250 V to 430 V.At the beginning of the simulation, the load is set to 32.59 Ω.Before the simulation time reaches 0.06 s, the output current remains 7.67 A, and the output voltage quickly reaches and can be stabilized at 250 V.At 0.06 s, the load changes from 32.59 Ω to 41.72 Ω. Due to the existence of the filter capacitor, the output voltage does not change immediately, and the output current drops sharply.After about 13 ms, the output current is stable at 7.67 A, and the output voltage reaches 320 V.At 0.12 s, the load changes from 41.72 Ω to 56.03 Ω at full load.After a sharp drop, the output current is stabilized at 7.67 A after about 8 ms, and the output voltage finally reaches and stabilizes at 430 V.The above process proves that the converter can work normally in the constant current stage.
Figure 13 is the closed-loop simulation waveform of the converter in the constant voltage stage.In the constant voltage stage, the current loop of the converter is saturated, and the voltage loop works at this time.The output voltage remains constant at 430 V throughout the constant voltage stage, and the output current gradually decreases from 7.67 A to equal to the self-discharge current.At the beginning of the simulation, the load is set to 56.03 Ω.Before the simulation time reaches 0.06 s, the output voltage remains constant at 430 V and the output current remains at 7.67 A. At 0.06 s, the load changes from 56.03 Ω to 81.03 Ω. Due to the existence of the voltage closed-loop regulation, the output voltage fluctuates slightly, and the output current drops sharply to 5.31 A. At 0.12 s, the load changes from 81.03 Ω to 106.03 Ω, and the output current drops sharply to 4.05 A. The two output voltages are quickly restored to stability after the oscillation, and the ripple is within 0.6%.The above process proves that the converter can work normally in the constant voltage stage.
Reverse Simulation
When the bidirectional CLLLC resonant converter is in the reverse discharge mode, the input voltage range is 250 V~430 V, and the voltage and current double-closed-loop control mode is adopted to control the output voltage to be stable at 400 V DC.The magnetizing inductance is equivalent to the secondary side, and the magnetizing inductance is L m /n 2 = 60.03 µH.
Figure 14 is the simulation result when the reverse input voltage is 250 V.As can be seen from Figure 14a, at this time in the variable-frequency-boost mode, the operating frequency is 69 kHz.Before the arrival of the driving signal V gs5 of the switching tube S 5 , the drain voltage V ds5 of the S 5 has been reduced to zero, which shows that the switching tube S 5 can achieve ZVS.Before the arrival of i D1 and i D4 , i D2 and i D3 have been reduced to zero.Similarly, before the arrival of i D2 and i D3 , i D1 and i D4 have also been reduced to zero, so the rectifier diode on the primary side can achieve ZCS.It can be seen from Figure 14b that the output voltage is stable at 400 V after 8 ms, indicating that the reverse frequency-conversion boost of the bidirectional CLLLC resonant converter can work normally.Figure 15 is the simulation result when the reverse input voltage is 450 V.As can be seen from Figure 15a, it is in the phase-shift buck mode, and the operating frequency is 100 kHz.Before the arrival of the driving signal V gs5 of the switching tube S 5 , the drain voltage V ds5 of the S 5 has been reduced to zero, which shows that the switching tube S 5 can achieve ZVS.Before the arrival of i D1 and i D4 , i D2 and i D3 have been reduced to zero.Similarly, before the arrival of i D2 and i D3 , i D1 and i D4 have also been reduced to zero, so the rectifier diode on the primary side can achieve ZCS.From Figure 15b, it can be seen that the output voltage is stable at 400 V after 2.9 ms, indicating that the reverse phase-shift buck of the bidirectional CLLLC resonant converter can work normally.
Conclusions
To meet the requirements of the on-board charging equipment of electric vehicles, this paper adopts the bidirectional CLLLC resonant converter as the rear stage of the onboard charger and analyzes the characteristics of this converter under frequency-conversion control and phase-shift control, and proposes a hybrid method of frequency-conversion and phase-shift control.Through the above analysis, the following conclusions can be drawn: (1) Due to its complete symmetrical and good soft-switching characteristics, the bidirectional CLLLC resonant converter solves the problem where the traditional LLC resonant converter has different resonant states during the forward and reverse operations, and cannot achieve soft switching at the same time and is difficult to control.(2) The hybrid control method of frequency conversion and phase shift adopts frequency-conversion control at a higher voltage gain and phase-shift control at a lower voltage gain.It solves the problem of low voltage gain of the single frequency-conversion control method in the step-down mode, effectively broadens the output voltage range of the bidirectional CLLLC converter, and is suitable for wide-range output occasions.At the same time, it achieves zero-voltage switching and zero-current
Conclusions
To meet the requirements of the on-board charging equipment of electric vehicles, this paper adopts the bidirectional CLLLC resonant converter as the rear stage of the on-board charger and analyzes the characteristics of this converter under frequency-conversion control and phase-shift control, and proposes a hybrid method of frequency-conversion and phase-shift control.Through the above analysis, the following conclusions can be drawn: (1) Due to its complete symmetrical and good soft-switching characteristics, the bidirectional CLLLC resonant converter solves the problem where the traditional LLC resonant converter has different resonant states during the forward and reverse operations, and cannot achieve soft switching at the same time and is difficult to control.(2) The hybrid control method of frequency conversion and phase shift adopts frequencyconversion control at a higher voltage gain and phase-shift control at a lower voltage gain.It solves the problem of low voltage gain of the single frequency-conversion control method in the step-down mode, effectively broadens the output voltage range of the bidirectional CLLLC converter, and is suitable for wide-range output occasions.At the same time, it achieves zero-voltage switching and zero-current switching in the full load range, and has a high operating efficiency.This control method is convenient and easy to implement.(3) The control method proposed in this paper is superior to the traditional control method in reducing the switching frequency-conversion range and improving the efficiency of the converter, which is helpful in the popularization and application of the high-efficiency and high-power bidirectional DC/DC converter in the distributed new energy generation.(4) In the actual debugging process, it is difficult to establish an accurate model of a bidirectional CLLLC circuit.In the later stage, it is necessary to further analyze the circuit model to design more appropriate closed-loop control parameters and test how well it actually works.
Figure 2 .
Figure 2. Forward equivalent circuit model of bidirectional CLLLC resonant converter.
Figure 2 .
Figure 2. Forward equivalent circuit model of bidirectional CLLLC resonant converter.Figure 2. Forward equivalent circuit model of bidirectional CLLLC resonant converter.
Figure 2 .
Figure 2. Forward equivalent circuit model of bidirectional CLLLC resonant converter.Figure 2. Forward equivalent circuit model of bidirectional CLLLC resonant converter.
Figure 3 .
Figure 3. Voltage gain and normalized frequency at different k values when Q = 0.3.
Figure 4 .
Figure 4. Voltage gain and normalized frequency at different Q values when k = 5.
Figure 3 .
Figure 3. Voltage gain and normalized frequency at different k values when Q = 0.3.
Figure 3 .
Figure 3. Voltage gain and normalized frequency at different k values when Q = 0.3.
Figure 4 .
Figure 4. Voltage gain and normalized frequency at different Q values when k = 5.Figure 4. Voltage gain and normalized frequency at different Q values when k = 5.
Figure 4 .
Figure 4. Voltage gain and normalized frequency at different Q values when k = 5.Figure 4. Voltage gain and normalized frequency at different Q values when k = 5.
Figures 9 - 11 ,Figure 9 .
Figure 9.The simulated waveform of Vo and Io is 250 V/7.67 A: (a) main working waveform; and (b) ZVS and ZCS.Figure 9.The simulated waveform of V o and I o is 250 V/7.67 A: (a) main working waveform; and (b) ZVS and ZCS.
Figure 9 .Figure 10 .Figure 11 .
Figure 9.The simulated waveform of Vo and Io is 250 V/7.67 A: (a) main working waveform; and (b) ZVS and ZCS.Figure 9.The simulated waveform of V o and I o is 250 V/7.67 A: (a) main working waveform; and (b) ZVS and ZCS.
Figure 10 .Figure 10 .Figure 11 .
Figure 10.The simulated waveform of V o and I o is 320 V/7.67 A: (a) main working waveform; and (b) ZVS and ZCS.
Figure 11 .
Figure 11.The simulated waveform of V o and I o is 430 V/7.67 A: (a) main working waveform; and (b) ZVS and ZCS. .
Figure 14 .Figure 14 .
Figure 14.Simulation results when the reverse input voltage is 250 V: (a) main working waveform; and (b) output voltage.
Figure 14 .Figure 15 .
Figure 14.Simulation results when the reverse input voltage is 250 V: (a) main working waveform; and (b) output voltage.
Figure 15 .
Figure 15.Simulation results when the reverse input voltage is 430 V: (a) main working waveform; and (b) output voltage.
Table 1 .
Design specifications of bidirectional CLLLC resonant converter.
Table 2 .
The parameters of primary and secondary resonant inductance and resonant capacitance.
Figure 12.Simulation waveform of forward constant-current stage.
Figure 12.Simulation waveform of forward constant-current stage.Figure 13.Simulation waveform of forward constant-voltage stage. | 12,157.6 | 2023-03-29T00:00:00.000 | [
"Engineering"
] |
Dual Role of Beam Polarization and Power in Nematic Liquid Crystals: A Comprehensive Study of TE- and TM-Beam Interactions
Optical spatial solitons are self-guided wave packets that maintain their transverse profile due to the self-focusing effect of light. In nematic liquid crystals (NLC), such light beams, called nematicons, can be induced by two principal mechanisms: light-induced reorientation of the elongated molecules and thermal changes in the refractive index caused by partial light absorption. This paper presents a detailed investigation of the propagation dynamics of light beams in nematic liquid crystals (NLCs) doped with Sudan Blue dye. Building on the foundational understanding of reorientational and thermal solitons in NLCs and the effective breaking of the action–reaction principle in spatial solitons, this study examines the interaction of infrared (IR) and visible beams in a [-4-(trans-4′-exylcyclohexyl)isothiocyanatobenzene] (6CHBT) NLC. Our experimental results highlight the intricate interplay of beam polarizations, power levels, and the nonlinear properties of NLCs, offering new insights into photonics and nonlinear optics in liquid crystals.
Introduction
Exploring light-matter interactions in nematic liquid crystals (NLCs) has become a cornerstone of contemporary photonics research, unveiling phenomena with profound theoretical and practical implications.NLCs, characterized by their unique anisotropic optical properties, serve as a versatile medium for studying various nonlinear optical effects, including the formation and dynamics of spatial solitons called nematicons [1,2].This research area bridges fundamental physics and technological innovations, leading to advancements in optical communication, signal processing, and material science.
The intriguing aspect of NLCs lies in their ability to manipulate light propagation through their molecular orientation, which can be controlled by external fields, temperature gradients, and light itself [3][4][5][6].By inducing dipoles in the NLC molecules, the extraordinary component of the optical field is responsible for the rotation of the molecular optical axis, an effect called reorientational nonlinearity [3].In addition, the NLCs exhibit additional substantial nonlinear mechanisms, i.e., thermal nonlinearity, associated with wavelength-dependent partial absorption.This responsiveness enables the study of various nonlinear optical effects, most notably the formation and propagation of solitons.Solitons in NLCs, specifically reorientational and thermal solitons, represent a crucial area of research due to their distinct characteristics and potential for practical applications [7][8][9][10].
Reorientational solitons, or "nematicons", arise from the interplay between the lightinduced reorientation of NLC molecules and the elastic restoring forces, creating selftrapped light beams [1,2].On the other hand, thermal solitons emerge from localized temperature changes induced by light absorption, altering the refractive-index profile of the NLC [11][12][13][14].These solitons offer a different perspective on light-matter interactions, where thermal effects play a pivotal role.Understanding thermal solitons and their impact on molecular reorientation is crucial for developing temperature-sensitive photonic devices and exploring the thermal dynamics in anisotropic materials.The coexistence and interplay between reorientational and thermal solitons in NLCs present a complex and affluent area of study.The response of NLCs to different light polarizations, such as transverse electric (TE) and transverse magnetic modes (TM), adds another layer of complexity.The coexistence of two competing nonlinear mechanisms, one of which (molecular reorientation) is always positive in sign and the second (i.e., the thermal effects) either negative or positive in sign, depending on the specific properties of the NLC, opens a wide range of different possibilities related to a non-trivial type of light-matter interaction in nonlinear reorientational soft matter [12,13,15,16].A more detailed analysis of the nonlinearity coefficient versus temperature for the [-4-(trans-4 ′ -exylcyclohexyl)isothiocyanatobenzene] (6CHBT) nematic liquid crystal [17] shows, among other things, the temperature ranges in which the two mechanisms leading to nonlinear refractive-index changes compete, coexist, or enhance each other [18].This interplay, influenced by light intensity, beam polarization, and external fields, leads to a diverse range of nonlinear phenomena [7,18,19].The differential interaction of these polarization states with NLCs has significant implications for the design of optical components, like beam steerers, modulators, and adaptive lenses.It also provides a deeper understanding of the anisotropic nature of these materials and their potential for creating tunable and adaptive optical systems.
Harnessing the competing and coexisting reorientational and thermal nonlinearities in suitably doped nematic liquid crystals (NLCs) presents significant advantages.These nonlinearities have been utilized to control soliton trajectories and positions within NLC cells.A particularly intriguing aspect is the mutual interaction of light beams in NLCs, leading to aligned accelerations, a phenomenon known as "diametric drive" [16,[20][21][22].This interaction, akin to two massive objects influencing each other in a manner that defies action-reaction symmetry, is an impossibility in classical mechanics but becomes feasible in the realm of NLCs, offering new frontiers in optical control and manipulation [23].
This paper aims to contribute to this growing field by presenting new insights into the dynamics of solitons in NLCs and exploring the implications of these findings for the future of photonics and advanced materials.We seek to expand the understanding of light propagation in anisotropic media and its potential applications in next-generation optical technologies through a combination of different experimental approaches.
Materials and Methods
In our study of nonlinear optics in NLCs, we introduce a guest dopant to the host LC mixture to heighten light absorption within a specific wavelength range.This allows for a detailed investigation into the interaction between reorientation, which typically dominates in pure NLCs, and thermo-optical effects.We explore these interactions by employing wavelengths inside and outside the dopant's absorption band.Both nonlinearity mechanisms, i.e., reorientation and thermal ones, allow for a straightforward modification of the magnitude of the NLC nonlinear response.NLCs, as uniaxial materials, consist of elongated molecules aligned along the molecular director (optic axis) n.The ordinary refractive index corresponds to eigenwaves with an electric field perpendicular to the n; the extraordinary refractive index depends on the angle θ between the optical wave vector and n.The value and spatial distribution of θ in an NLC defines the phase velocity and diffraction of a propagating light beam.The ordinary electric field (o-wave) is always perpendicular to the director n, and its phase velocity is c/n o .In contrast, the extraordinary electric field (e-wave) lies in the same plane as the wave vector k and the director n, with a phase velocity of c/n e , which depends on the orientation angle θ between k and n.The refractive indices of NLCs are temperature-dependent, with n e typically decreasing as the temperature increases and n o either increasing or decreasing depending on the material's properties [3].The exact relationship between temperature and refractive index depends on the specific NLC material and its temperature range.
The extraordinary wave polarization propagates in a θ-dependent refractive index: , with energy flux (Poynting vector) angularly displaced by the walk-off angle δ(θ, T) from the wave vector k.This causes the beam trajectory to vary through changes in walk-off, the latter conveniently expressed as [24] δ(θ, T) = arctan ϵ a (T)sin2θ In the case of the mechanism based on reorientation, the magnitude of the nonlinear response, expressed by the n 2R coefficient, depends both on the angle between the director and the electric component of the optical wave field and temperature.The nonlinear index change ∆n 2 e (θ, T) = n 2 e (θ, T) − n 2 e (θ 0 , T) depends on the director distribution, where the optic axis distribution is given by θ = θ 0 + ψ, i.e., the superposition of the reorientation at rest (θ 0 ) and the nonlinear reorientation ψ due to the beam.The nonlinearity, hence the degree of confinement of the solitary beam, can be expressed as [7] n 2R (θ 0 , T) = 2ϵ 0 ϵ a (T) where ϵ a (T) = n 2 e (T) − n 2 o (T) is the optical anisotropy and K(T) the Frank elastic constant, quantifying the temperature-dependent strength of the intermolecular links.
In the tailored design of an NLC cell, the molecular alignment direction is strategically determined to achieve a specific level of nonlinear optical response.The absolute value of the nonlinear coefficient n 2T , which is related to the thermal nonlinearity mechanism, depends on the temperature characteristics of the refractive indices for a given NLC compound.It can be expressed by n 2T = σ•dn/dT, where σ and n are the absorption coefficient and the ordinary or extraordinary refractive index, respectively.By increasing the linear absorption of light, for example, by doping absorbent dyes, the magnitude of the n 2T coefficient can be adjusted to the desired level.
In experimental investigation, we deployed dual-wavelength optics to probe the intricate nonlinearity in nematic liquid crystals (NLCs) doped with Sudan Blue II [12,25], which purely enhances the thermal nonlinear response.Employing a transverse electric (TE) infrared Nd:YAG laser at λ = 1064 nm-chosen for its minimal absorption by the dye -and a transverse magnetic (TM) visible beam at λ = 642 nm or λ = 532 nm, aligned with the dopant's peak absorption, we dissected the complex dynamics of nematicon propagation.
The experimental investigation was performed in planar NLC cells consisting of two glass plates bonded together with a gap of d = 80 µm and the homogeneous orientation of the director at θ 0 = 30 • with respect to the z-axis (the z-axis coincides with the wave vector k).The cell was filled with a 6CHBT nematic liquid crystal doped with a 0.05% weight concentration of Sudan Blue II dye, with an absorption peak localized at λ = 604 nm and significantly enhanced absorption at λ = 532 nm [25].
The initial orientation angle was chosen for two purposes: firstly, to induce a negative refractive gradient due to the temperature increase (manifesting as negative thermal nonlinearity for the e-wave) and, secondly, to balance the minimization of temperature sensitivity in the walk-off angle against the maximization of the reorientation nonlinearity coefficient n 2R , as depicted in Figure 1.
Both input beams (orthogonally polarized, continuous wave (CW) characterized by a Gaussian electric-field profile) were collimated by a microscope objective to comparable widths in their waist (about 3 microns) and simultaneously coupled into the NLC medium.The beams remained close to each other and were aligned so that their Poynting vectors were kept parallel within the NLC cell.The light propagation was observed by the CCD camera combined with the microscope objective, which collected the out-of-plane scattered light in the yz-plane.Both input beams (orthogonally polarized, continuous wave (CW) characterized by a Gaussian electric-field profile) were collimated by a microscope objective to comparable widths in their waist (about 3 microns) and simultaneously coupled into the NLC medium.The beams remained close to each other and were aligned so that their Poynting vectors were kept parallel within the NLC cell.The light propagation was observed by the CCD camera combined with the microscope objective, which collected the out-of-plane scattered light in the yz-plane.
Experimental Results and Discussion
At this stage, our investigation centers on the TE-infrared-beam propagation, which operates through molecular reorientation and gives rise to a reorientational nematicon.This reorientation is contrasted against the effects of a temperature gradient produced by the highly absorbed visible TM-polarized wave on the NLC medium.Under the influence of the visible TM-polarized beam, the NLC s ordinary refractive index increases, forming a red thermal nematicon.Simultaneously, this thermally induced soliton diminishes the extraordinary refractive index perceived by the TE-induced reorientation nematicon.According to (1) and Figure 1a,b, this temperature gradient introduces dual perturbations to the infrared beam s trajectory: (1) a thermally induced modification of the walk-off angle due to birefringence changes and (2) a drift towards regions of higher refractive index.This, in turn, substantially alters the trajectory of the extraordinary beam.Please note that in this configuration, the TM-polarized beam does not perturb the optical parameters of the medium in any way that could affect the nature or the propagation direction of a TM-polarized beam.Despite the launch angle of the red transverse magnetic (TM) beam being slightly angled away from the z-axis direction, owing to the infraredtransverse-electric (TE)-beam walk-off angle, it encounters an ordinary refractive index when coupling, regardless of the coupling direction.d-f) the same as in (a-c) for the infrared wavelength spectral range.
Experimental Results and Discussion
At this stage, our investigation centers on the TE-infrared-beam propagation, which operates through molecular reorientation and gives rise to a reorientational nematicon.This reorientation is contrasted against the effects of a temperature gradient produced by the highly absorbed visible TM-polarized wave on the NLC medium.Under the influence of the visible TM-polarized beam, the NLC's ordinary refractive index increases, forming a red thermal nematicon.Simultaneously, this thermally induced soliton diminishes the extraordinary refractive index perceived by the TE-induced reorientation nematicon.According to (1) and Figure 1a,b, this temperature gradient introduces dual perturbations to the infrared beam's trajectory: (1) a thermally induced modification of the walk-off angle due to birefringence changes and (2) a drift towards regions of higher refractive index.This, in turn, substantially alters the trajectory of the extraordinary beam.Please note that in this configuration, the TM-polarized beam does not perturb the optical parameters of the medium in any way that could affect the nature or the propagation direction of a TM-polarized beam.Despite the launch angle of the red transverse magnetic (TM) beam being slightly angled away from the z-axis direction, owing to the infrared-transverseelectric (TE)-beam walk-off angle, it encounters an ordinary refractive index when coupling, regardless of the coupling direction.In this way, for a TM-polarized beam, the NLC behaves as an isotropic medium with an index of refraction equal to the ordinary refractive index of the liquid crystal.
The nematicon trajectory's alteration, consequent to the walk-off angle decreasing as a function of temperature, results in the propagation direction shifting toward the system's origin.Furthermore, the increased temperature caused by the red beam reduces the refractive index encountered by the infrared beam, causing it to move away along the temperature gradient created.Depending on the specific geometry, i.e., whether the red beam is launched above or below the infrared beam, the cumulative or counteracting effect of these induced refractive-index gradients on the overall beam trajectory is critically essential, i.e., it can either complement or compete with the temperature-induced walkoff shift.
Figure 2 provides a comprehensive illustration of these phenomena, depicting the NLC sample and the coupling geometry of the light beams and presenting experimental results.In Figure 2a, the top-view capture realized in the yz-plane shows a red beam launched above the infrared soliton.When a low-power red beam is used (P RED = 0.4 mW), as presented in Figure 2b,c, the infrared soliton's trajectory alteration is minimal, reducing the walk-off angle by about 0.4 degrees.However, increasing the power of the red beam up to 4 mW causes a further alteration in the infrared nematicon trajectory and a reduction in the effective walk-off angle by approximately 1.2 degrees, corresponding to an effective change in the y-coordinate of the soliton of 20 microns at a distance of z = 1000 µm.The exact trajectories of an infrared beam are plotted in Figure 2d.In the presented configuration, the thermal effects decrease the transverse shift in the nematicon, which resulted from the walkoff and light-induced molecular reorientation.The light-to-dark grey solid lines represent the trajectories of the TE beam along with the increased power of a highly absorbed red light beam.Simultaneously, we can see that the trajectory of the nematicon induced by thermal nonlinearity remains unaltered.system s origin.Furthermore, the increased temperature caused by the red beam reduces the refractive index encountered by the infrared beam, causing it to move away along the temperature gradient created.Depending on the specific geometry, i.e., whether the red beam is launched above or below the infrared beam, the cumulative or counteracting effect of these induced refractive-index gradients on the overall beam trajectory is critically essential, i.e., it can either complement or compete with the temperature-induced walk-off shift.
Figure 2 provides a comprehensive illustration of these phenomena, depicting the NLC sample and the coupling geometry of the light beams and presenting experimental results.In Figure 2a, the top-view capture realized in the yz-plane shows a red beam launched above the infrared soliton.When a low-power red beam is used (PRED = 0.4 mW), as presented in Figure 2b,c, the infrared soliton s trajectory alteration is minimal, reducing the walk-off angle by about 0.4 degrees.However, increasing the power of the red beam up to 4 mW causes a further alteration in the infrared nematicon trajectory and a reduction in the effective walk-off angle by approximately 1.2 degrees, corresponding to an effective change in the y-coordinate of the soliton of 20 microns at a distance of z = 1000 µm.The exact trajectories of an infrared beam are plotted in Figure 2d.In the presented configuration, the thermal effects decrease the transverse shift in the nematicon, which resulted from the walk-off and light-induced molecular reorientation.The light-to-dark grey solid lines represent the trajectories of the TE beam along with the increased power of a highly absorbed red light beam.Simultaneously, we can see that the trajectory of the nematicon induced by thermal nonlinearity remains unaltered.In the second configuration presented in Figure 2e, we indicate the red beam that is launched below the infrared nematicon.The experimental results are presented in Figure 2f,g.As in the low power configuration, the modification of the infrared soliton trajectory is small.The walk-off angle increases by 0.25 degrees, and the trajectory shifts towards a In the second configuration presented in Figure 2e, we indicate the red beam that is launched below the infrared nematicon.The experimental results are presented in Figure 2f,g.As in the low power configuration, the modification of the infrared soliton trajectory is small.The walk-off angle increases by 0.25 degrees, and the trajectory shifts towards a positive value on the y-axis.If the power of the red beam is increased, similar to the configuration in Figure 2a, the trajectory is further modified due to the repulsion of the infrared beam by the red beam.For a red beam power of 4 mW, the infrared soliton experiences a change in the walk-off angle of about one degree.This represents an increase in the walk-off angle relative to the propagation undisturbed by the red beam.The exact trajectories of an infrared beam are plotted in Figure 2h.Similar to Figure 2d, the lightto-dark grey solid lines represent the trajectories of the TE beam along with the increased power of a highly absorbed red light beam.The interaction changes the TE beam's walk-off angle and propagation direction, highlighting the complexity of light-matter interactions within NLCs and the influence of polarization and temperature on soliton response.The detailed trajectory plots underscore the significance of beam placement and the refractiveindex gradient's role in steering the propagation within the medium.
The above results considered the effect of the thermal soliton on the propagation, particularly the direction of the reorientation soliton, and concerned configurations where the beams are introduced at some distance apart.The results presented in Figure 3 consider a configuration where both beams, the TE infrared and the visible TM, are launched along the same path.Similarly, as explained above, the TM beam with a power of P = 6.25 mW forms the thermal soliton, and the infrared TE beam forms the reorientational soliton.We show here that the interplay between both nonlinearities leads to extraordinary beam splitting and its propagation in the form of two solitons propagating at an angle dependent on the power of the ordinary beam (thermal soliton).
red beam by the red beam.For a red beam power of 4 mW, the infrared soliton experiences a change in the walk-off angle of about one degree.This represents an increase in the walkoff angle relative to the propagation undisturbed by the red beam.The exact trajectories of an infrared beam are plotted in Figure 2h.Similar to Figure 2d, the light-to-dark grey solid lines represent the trajectories of the TE beam along with the increased power of a highly absorbed red light beam.The interaction changes the TE beam s walk-off angle and propagation direction, highlighting the complexity of light-matter interactions within NLCs and the influence of polarization and temperature on soliton response.The detailed trajectory plots underscore the significance of beam placement and the refractive-index gradient s role in steering the propagation within the medium.
The above results considered the effect of the thermal soliton on the propagation, particularly the direction of the reorientation soliton, and concerned configurations where the beams are introduced at some distance apart.The results presented in Figure 3 consider a configuration where both beams, the TE infrared and the visible TM, are launched along the same path.Similarly, as explained above, the TM beam with a power of P = 6.25 mW forms the thermal soliton, and the infrared TE beam forms the reorientational soliton.We show here that the interplay between both nonlinearities leads to extraordinary beam splitting and its propagation in the form of two solitons propagating at an angle dependent on the power of the ordinary beam (thermal soliton).The power of the reorientational soliton is P = 6.5 mW.At this power, two breaths [19,26,27] were observed along the propagation path, indicating strong beam localization.Virtually no change in soliton propagation is observed when a green beam of initially low power is introduced along the same path.Both beams propagate without mutual influence.Increasing the power of the green beam generates a thermal soliton and thus reduces the extraordinary refractive index observed in the TE beam.The mechanism of the effect ,c) experimentally recorded mutual propagation of beams for P IR = 6.5 mW, P GREEN = 6.25 mW and P IR = 6.5 mW, and P GREEN = 6.5 mW: filtered view of a visible wavelength beam; (d) independent propagation of nematicon induced by an IR beam of a power P IR = 6.5 mW and (e) co-propagation with a thermal nematicon induced by a visible beam of a power P GREEN = 6.25 mW and (f) P GREEN = 6.5 mW: filtered view of an IR spectral range.The right panels in (b-f) show the beam profile at the output of an NLC cell, at z = 1500 µm, filtered at particular wavelengths; (g) intensity distribution along the y-axis direction, at the propagation distance of z = 1000 µm, corresponding to (f), for increasing powers of the green beam: P GREEN = 0 mW, 6.25 mW, 6.50 mW, 7.00 mW, and 7.50 mW (light-grey-to-dark-red lines, respectively).
The power of the reorientational soliton is P = 6.5 mW.At this power, two breaths [19,26,27] were observed along the propagation path, indicating strong beam localization.Virtually no change in soliton propagation is observed when a green beam of initially low power is introduced along the same path.Both beams propagate without mutual influence.Increasing the power of the green beam generates a thermal soliton and thus reduces the extraordinary refractive index observed in the TE beam.The mechanism of the effect of the thermal soliton on the infrared beam is analogous to that described above: the temperature increase induced by the green beam reduces the extraordinary refractive index seen by the TE beam and modifies the value of the walk-off angle.This temperature increase causes the repulsion of the orientation soliton.In this case, the beams have been guided along a single path.This leads to the separation of the infrared beam and its further propagation in the form of two solitons, each with a lower power.The soliton is split into two for a thermal soliton power of P = 6.25 mW.A further increase in the thermal soliton power leads to the repulsion of these solitons, in agreement with the results presented earlier.
The results presented above detail the propagation of two orthogonally polarized beams with different wavelengths, where one wavelength is highly absorbed by the medium, causing a notable temperature increase.In contrast, the other is unaffected by the dye.Additionally, investigations were conducted with two beams of the same wavelength, specifically λ = 532 nm, within the dye's enhanced absorption spectrum.Both TEand TM-polarized beams at this wavelength lead to a rise in temperature during propagation.The study involved introducing TM and TE beams into a liquid-crystal cell separately and analyzing the nematicon formation and propagation.TM beams generate solitons through thermal nonlinearity, increasing the ordinary refractive index, whereas TE beams, through reorientation nonlinearity, cause an increase in the extraordinary refractive index and molecular rotation, leading to beam localization.The TE beam's induced localization, accompanied by medium heating, results in a reduced effective refractive index and an altered walk-off angle, impacting the beam's guidance direction.This effect on localization is observed primarily in the change in the guidance direction due to the walk-off angle shift.
In the next step, we study the independent propagation of beams at the wavelength of enhanced absorption of Sudan Blue II dye (λ = 532 nm).At this optical frequency, the thermal nonlinearity supports beam self-focusing for TM polarization in the form of a thermal soliton.On the other hand, the negative contribution of thermal effects (defocusing nonlinearity) for TE-polarized light is much weaker than the local increase in refractive index imposed by molecular reorientation.Thus, the stable propagation of a solitary wave is possible due to the reorientation nonlinearity mechanism for this particular polarization of light.Still, as the light absorption caused the temperature increase in the NLC, we observed a significant variation in the beam trajectory related to decreased birefringence.The defocusing character of the thermal nonlinearity for the extraordinary component is compensated by the reorientational nonlinearity, thus ensuring the formation of selfcollimated beams in both polarizations.
Figure 4 illustrates the evolution of a TM-and-TE green light beam in a homogeneously aligned NLC cell.The TM-ordinary-wave propagation was observed in both low-power (P GREEN = 0.5 mW) and high-power (P GREEN = 6.0 mW) regimes, as shown in Figure 4b,c.An increase in optical power leads to a significant decrease in beam width without altering its trajectory, as depicted in Figure 4d,e, across various power levels from 1.0 mW to 6.0 mW.
The TE-polarized beam of the same wavelength was analyzed in the same NLC cell with a slightly tilted coupling direction to maintain propagation along the z-axis (Figure 4f).Low-and high-power regimes (P GREEN = 0.5 mW and 5.0 mW, respectively) are presented in Figure 4g,h.The TE beam, supporting molecular reorientation, starts forming a solitary wave at P GREEN = 4.0 mW in Figure 4i.The thermal nonlinearity becomes more significant for higher powers, leading to notable walk-off variations.Comparing beam trajectory at P GREEN = 0.5 mW and P GREEN = 5.0 mW over z = 1000 µm reveals a transverse shift of approximately 10 µm, as demonstrated in Figure 4j.
The last part of our analysis explores the interaction of TE-and TM-polarized beams of the same wavelength, λ = 532 nm.The experimental setup, organized as schematically shown in Figure 5a, examines the propagation of TE and TM beams through an NLC cell, tracking the effects of polarization-dependent nonlinearities on beam behavior.The critical point here is that the reorientational nonlinearity is dominant over the TE beam's thermal effect, thus ensuring spatial soliton's existence.Conversely, thermal nonlinearity needed to be sufficiently strong to generate a thermal soliton for the TM beam, requiring a significant positive gradient in the ordinary refractive index.Similar to the above results, the TE beam, through its interaction with the optical axis, created a reorientational nematicon.The substantial absorption and resultant temperature gradient altered the TE beam's trajectory by reducing the walk-off of the extraordinary beam.This modification in the walk-off angle corresponds to a displacement along the y-axis at a specific distance.The last part of our analysis explores the interaction of TE-and TM-polarized beams of the same wavelength, λ = 532 nm.The experimental setup, organized as schematically shown in Figure 5a, examines the propagation of TE and TM beams through an NLC cell, tracking the effects of polarization-dependent nonlinearities on beam behavior.The critical point here is that the reorientational nonlinearity is dominant over the TE beam s thermal effect, thus ensuring spatial soliton s existence.Conversely, thermal nonlinearity needed to be sufficiently strong to generate a thermal soliton for the TM beam, requiring a significant positive gradient in the ordinary refractive index.Similar to the above results, the TE beam, through its interaction with the optical axis, created a reorientational nematicon.The substantial absorption and resultant temperature gradient altered the TE beam s trajectory by reducing the walk-off of the extraordinary beam.This modification in the walk-off angle corresponds to a displacement along the y-axis at a specific distance.
Interestingly, we observed that the extraordinary beam is repelled from the ordinary beam, yet the ordinary beam is simultaneously attracted to the extraordinary one.In this setup, the TE beam s dual role of reorientation and induced positive temperature gradient (resulting in a negative refractive-index gradient for the TE beam and a positive one for the TM beam) was crucial.This negative refractive-index gradient affected the TE beam s propagation direction, countering the TM beam s repulsion.However, the TM beam s repulsion proved to be stronger.At the same time, the TM beam experienced attraction towards the TE beam due to the positive temperature gradient relevant to the ordinary wave.The attractive and repulsive force ratio can be effectively controlled by varying the two input powers.In the Figure 5b,c, we can observe propagation of extraordinary beam of a power PTE = 2 mW co-propagated with an ordinary wave of a power PTM = 2 mW and PTM = 6 mW, respectively.In such conditions, we can observe significant change in trajectory of a TE beam as a result of mutual interaction.For a higher power of TE beam, equal to PTE = 4 mW, we can also observe a shift in the TM-polarized beam trajectory, towards TE beam, as shown in Figure 5d,e, for ordinarily polarized beams of a powers PTM = 2 mW and PTM = 6 mW.
The exact trajectories of the propagating beam that correspond to the cases presented in Figure 5b-e are plotted in Figure 5f,g.As can be clearly seen, the NLC medium, due to its peculiar polarization-dependent nonlinearity, both reorientation and thermal, via nonlinear interaction, can support the so-called diametric drive.Diametric drive means that the first beam must be drawn towards the second, while conversely, the second beam is repulsed by the first.
Conclusions
In conclusion, this intricate interdependence of thermal and reorientational effects, Interestingly, we observed that the extraordinary beam is repelled from the ordinary beam, yet the ordinary beam is simultaneously attracted to the extraordinary one.In this setup, the TE beam's dual role of reorientation and induced positive temperature gradient (resulting in a negative refractive-index gradient for the TE beam and a positive one for the TM beam) was crucial.This negative refractive-index gradient affected the TE beam's propagation direction, countering the TM beam's repulsion.However, the TM beam's repulsion proved to be stronger.At the same time, the TM beam experienced attraction towards the TE beam due to the positive temperature gradient relevant to the ordinary wave.The attractive and repulsive force ratio can be effectively controlled by varying the two input powers.In the Figure 5b,c, we can observe propagation of extraordinary beam of a power P TE = 2 mW co-propagated with an ordinary wave of a power P TM = 2 mW and P TM = 6 mW, respectively.In such conditions, we can observe significant change in trajectory of a TE beam as a result of mutual interaction.For a higher power of TE beam, equal to P TE = 4 mW, we can also observe a shift in the TM-polarized beam trajectory, towards TE beam, as shown in Figure 5d,e, for ordinarily polarized beams of a powers P TM = 2 mW and P TM = 6 mW.
The exact trajectories of the propagating beam that correspond to the cases presented in Figure 5b-e are plotted in Figure 5f,g.As can be clearly seen, the NLC medium, due to its peculiar polarization-dependent nonlinearity, both reorientation and thermal, via nonlinear interaction, can support the so-called diametric drive.Diametric drive means that the first beam must be drawn towards the second, while conversely, the second beam is repulsed by the first.
Conclusions
In conclusion, this intricate interdependence of thermal and reorientational effects, as governed by beam polarization and geometry, underscores the complex dynamics within doped NLCs, with profound implications for designing advanced photonic systems for controlling and manipulating light.The thermal and reorientational effects offer a deeper understanding of light-matter interactions in anisotropic media, paving the way for developing novel optical devices leveraging these solitonic properties.
The intricate dynamics discovered in our study, such as the diametric drive and interaction between TE-and TM-polarized beams, suggest several promising avenues for future research.Integration with photonic structures: The integration of NLC-based nonlinear phenomena with photonic crystals, waveguides, and microresonators can open new possibilities for on-chip optical signal processing, light routing, and switching.Such integration efforts can leverage the unique properties of NLCs for the development of reconfigurable photonic devices and circuits.Dynamic Control and Reconfigurability: Another promising direction involves developing methods for dynamic control of nonlinear optical phenomena in NLCs using external fields, temperature gradients, or light itself.This research could result in the creation of adaptive optical elements capable of real-time adjustment of their optical properties in response to changing environmental conditions or operational requirements.
The potential applications span a wide range of advanced photonic systems that exploit the unique properties of NLCs for the development of the next generation of optical devices.The intricate interplay between beam polarization, power, and the nonlinear properties of NLCs opens new avenues for designing advanced photonic devices.In particular, the ability to manipulate light propagation through reorientation and thermal nonlinearities in NLCs can be exploited for the design of tunable optical components, such as lenses, modulators and beam steering devices.These components have important applications in optical communications.They enhance signal processing capabilities and improve the efficiency of photonic circuits.In addition, the observed phenomena, such as the "diametric drive" effect, suggest innovative approaches to light control and manipulation in optical computing and information processing.Here, light beams can interact in unconventional ways to perform computational tasks or control signal paths.In addition, the sensitivity of NLCs to different wavelengths and polarization states can be exploited in sensors and imaging devices.This offers improved sensitivity tailored to specific environmental conditions.These applications highlight the potential of NLC-based photonic systems in signal processing, control, and switching.
Figure 2 .
Figure 2. (a) Sketch of a liquid-crystal sample (6CHBT NLC doped with 0.05% of Sudan Blue dye, θ0 = 60°) with infrared (TE-polarized) and visible light (TM-polarized) beams coupling geometry and trajectories during propagation in NLC for a visible beam of low (left panel) and high (right panel) powers; (b,c) photographs of a simultaneous propagation in configuration form (a), presented for optical powers PIR = 5.6 mW and PRED = 0.4 mW, respectively; (d) trajectories of the infrared (PIR = 5.6 mW) corresponding to increasing powers of a red beam, shown from light-to-dark grey solid lines, respectively: PRED = 0.4 mW, PRED = 2.7 mW, and PRED = 4.0 mW.The dashed black line denotes the trajectory of the infrared beam in the case of independent propagation.Red dashed and solid lines refer to the visible beam of low and high powers; (e-h) refer to the same as in (a-d) for inverted-IR and visible-beam sequences.
Figure 2 .
Figure 2. (a) Sketch of a liquid-crystal sample (6CHBT NLC doped with 0.05% of Sudan Blue dye, θ 0 = 60 • ) with infrared (TE-polarized) and visible light (TM-polarized) beams coupling geometry and trajectories during propagation in NLC for a visible beam of low (left panel) and high (right panel) powers; (b,c) photographs of a simultaneous propagation in configuration form (a), presented for optical powers P IR = 5.6 mW and P RED = 0.4 mW, respectively; (d) trajectories of the infrared (P IR = 5.6 mW) corresponding to increasing powers of a red beam, shown from light-to-dark grey solid lines, respectively: P RED = 0.4 mW, P RED = 2.7 mW, and P RED = 4.0 mW.The dashed black line denotes the trajectory of the infrared beam in the case of independent propagation.Red dashed and solid lines refer to the visible beam of low and high powers; (e-h) refer to the same as in (a-d) for inverted-IR and visible-beam sequences.
Figure 3 .
Figure 3. (a) Sketch of a liquid-crystal sample (6CHBT NLC doped with 0.05% of Sudan Blue dye, θ0 = 30°) with indicated coupling geometry of IR (TE-polarized) and visible light (TM-polarized) beams and their trajectories during propagation in the NLC: reported for low and high powers of TM-polarized light (left and right panels, respectively); (b,c) experimentally recorded mutual propagation of beams for PIR = 6.5 mW, PGREEN = 6.25 mW and PIR = 6.5 mW, and PGREEN = 6.5 mW: filtered view of a visible wavelength beam; (d) independent propagation of nematicon induced by an IR beam of a power PIR = 6.5 mW and (e) co-propagation with a thermal nematicon induced by a visible beam of a power PGREEN = 6.25 mW and (f) PGREEN = 6.5 mW: filtered view of an IR spectral range.The right panels in (b-f) show the beam profile at the output of an NLC cell, at z = 1500 µm, filtered at particular wavelengths; (g) intensity distribution along the y-axis direction, at the propagation distance of z = 1000 µm, corresponding to (f), for increasing powers of the green beam: PGREEN = 0 mW, 6.25 mW, 6.50 mW, 7.00 mW, and 7.50 mW (light-grey-to-dark-red lines, respectively).
Figure 3 .
Figure 3. (a) Sketch of a liquid-crystal sample (6CHBT NLC doped with 0.05% of Sudan Blue dye, θ 0 = 30 • ) with indicated coupling geometry of IR (TE-polarized) and visible light (TM-polarized) beams and their trajectories during propagation in the NLC: reported for low and high powers of TMpolarized light (left and right panels, respectively); (b,c) experimentally recorded mutual propagation of beams for P IR = 6.5 mW, P GREEN = 6.25 mW and P IR = 6.5 mW, and P GREEN = 6.5 mW: filtered view of a visible wavelength beam; (d) independent propagation of nematicon induced by an IR beam of a power P IR = 6.5 mW and (e) co-propagation with a thermal nematicon induced by a visible beam of a power P GREEN = 6.25 mW and (f) P GREEN = 6.5 mW: filtered view of an IR spectral range.The right panels in (b-f) show the beam profile at the output of an NLC cell, at z = 1500 µm, filtered at particular wavelengths; (g) intensity distribution along the y-axis direction, at the propagation distance of z = 1000 µm, corresponding to (f), for increasing powers of the green beam: P GREEN = 0 mW, 6.25 mW, 6.50 mW, 7.00 mW, and 7.50 mW (light-grey-to-dark-red lines, respectively).
Figure 4 .
Figure 4. Analysis of unperturbed Gaussian beam propagation (λ = 532 nm) in 6CHBT NLC doped with 0.05% of Sudan Blue dye: (a) sketch of the NLC cell (θ0 = 30°) and the geometry of the TMpolarized-light-beam coupling; (b,c) the yz-plane view photographs of the TM-beam propagation for the optical powers of PTM = 1.0 mW and PTM = 6.0 mW, respectively; (d,e) evolution of the beam width and trajectory versus propagation distance, plotted for optical powers PTM = 1.0 mW (black line) and 4.0 mW, 5.0 mW, and 6.0 mW (light-to-dark green lines, respectively); (f) the sketch of the NLC cell (θ0 = 30°) and the geometry of the TE-polarized-light-beam coupling; (g-j) the same as in (b-e) regarding the TE-polarized-light-beam propagation in the configuration shown in (f).
Figure 4 .
Figure 4. Analysis of unperturbed Gaussian beam propagation (λ = 532 nm) in 6CHBT NLC doped with 0.05% of Sudan Blue dye: (a) sketch of the NLC cell (θ 0 = 30 • ) and the geometry of the TMpolarized-light-beam coupling; (b,c) the yz-plane view photographs of the TM-beam propagation for the optical powers of P TM = 1.0 mW and P TM = 6.0 mW, respectively; (d,e) evolution of the beam width and trajectory versus propagation distance, plotted for optical powers P TM = 1.0 mW (black line) and 4.0 mW, 5.0 mW, and 6.0 mW (light-to-dark green lines, respectively); (f) the sketch of the NLC cell (θ 0 = 30 • ) and the geometry of the TE-polarized-light-beam coupling; (g-j) the same as in (b-e) regarding the TE-polarized-light-beam propagation in the configuration shown in (f).
Figure 5 .
Figure 5. Co-propagation of TM-and TE-polarized Gaussian beams (λ = 532 nm) in 6CHBT NLC doped with 0.05% of Sudan Blue dye: (a) sketch of the NLC cell (θ0 = 30°) and optical fields coupling geometry with initial separation equal to 15 µm; (b) experimental verification of the co-propagation two orthogonally polarized light beams with optical powers corresponding to PTM = 2 mW, PTE = 2 mW, (c) PTM = 6 mW, PTE = 2 mW, (d) PTM = 2 mW, PTE = 4 mW, and (e) PTM = 6 mW, PTE = 4 mW.The dotted and dashed lines in (b-e) indicate the trajectories of the unperturbed and low-power TM and TE beams, respectively.(f) The trajectories of the co-propagated light beams correspond to (b,c): the TE beam in the unperturbed case (PTE = 4 mW, black dashed line) and influenced by the TM beam of powers PTM = 2 mW and PTM = 6 mW (light and dark green solid lines, respectively), the TM beam in the unperturbed case (PTM = 2 mW, black dotted line), and PTM = 2 mW simultaneously propagated with PTE = 4 mW (dark brown solid line); (g) the same as in (f) plotted for the TE beam in the unperturbed case (PTE = 2 mW, black dashed line) and influenced by the TM beam of powers PTM = 2 mW and PTM = 6 mW (light and dark green solid lines), the TM beam in the unperturbed case (PTM = 2 mW, black dotted line), and PTM = 2 mW and PTM = 6 mW simultaneously propagated with PTE = 4 mW (light and dark brown solid lines, respectively).
Figure 5 .
Figure 5. Co-propagation of TM-and TE-polarized Gaussian beams (λ = 532 nm) in 6CHBT NLC doped with 0.05% of Sudan Blue dye: (a) sketch of the NLC cell (θ 0 = 30 • ) and optical fields coupling geometry with initial separation equal to 15 µm; (b) experimental verification of the co-propagation two orthogonally polarized light beams with optical powers corresponding to P TM = 2 mW, P TE = 2 mW, (c) P TM = 6 mW, P TE = 2 mW, (d) P TM = 2 mW, P TE = 4 mW, and (e) P TM = 6 mW, P TE = 4 mW.The dotted and dashed lines in (b-e) indicate the trajectories of the unperturbed and low-power TM and TE beams, respectively.(f) The trajectories of the co-propagated light beams correspond to (b,c): the TE beam in the unperturbed case (P TE = 4 mW, black dashed line) and influenced by the TM beam of powers P TM = 2 mW and P TM = 6 mW (light and dark green solid lines, respectively), the TM beam in the unperturbed case (P TM = 2 mW, black dotted line), and | 9,909.2 | 2024-02-22T00:00:00.000 | [
"Physics"
] |
An Instantaneous Recombination Rate Method for the Analysis of Interband Recombination Processes in ZnO Crystals
Time-resolved photoluminescence (TRPL) analysis is often performed to assess the qualitative features of semiconductor crystals using predetermined functions (e.g., double- or multi-exponentials) to fit the decays of PL intensity. However, in many cases—including the notable case of interband PL in direct gap semiconductors—this approach just provides phenomenological parameters and not fundamental physical quantities. In the present work, we highlight that within a properly chosen range of laser excitation, the TRPL of zinc oxide (ZnO) bulk crystals can be described with excellent precision with second-order kinetics for the total recombination rate. We show that this allows us to define an original method for data analysis, based on evaluating the “instantaneous” recombination rate that drives the initial slope of the decay curves, acquired as a function of the excitation laser fluence. The method is used to fit experimental data, determining useful information on fundamental quantities that appear in the second-order recombination rate, namely the PL (unimolecular) lifetime, the bimolecular recombination coefficient, the non-radiative lifetime and the equilibrium free-carrier concentration. Results reasonably close to those typically obtained in direct gap semiconductors are extracted. The method may represent a useful tool for gaining insight into the recombination processes of a charge carrier in ZnO, and for obtaining quantitative information on ZnO excitonic dynamics.
Introduction
The phenomenon of band-to-band (or "interband") photoluminescence (PL) in semiconductor materials is caused by the radiative recombination of free charge carriers (i.e., valence holes and conduction electrons), generated via the absorption of optical radiation, whose photon energy is larger than the bandgap energy of the material.
Surface traps, bulk defects and band offsets at interfaces significantly affect the rates of charge recombination [1][2][3][4] and thus of PL intensity, by introducing efficient non-radiative recombination pathways and/or by affecting the average spatial separation between electrons and holes. As a consequence, the optical PL spectroscopy of semiconductors is often employed to assess crystal quality and surface/interface characteristics [5][6][7][8][9][10]. Very often, these assessments do not involve quantitative analysis and phenomenological approaches are sufficient for extracting the desired information. A typical example of that is the use of a predetermined decaying function to fit the time-resolved photoluminescence (TRPL) decays. For example, lifetimes are often extracted by fitting the data with a sum of exponentially-decaying functions. However, such lifetimes are just phenomenological parameters and not fundamental physical quantities. In fact, a multi-exponential function would be the correct description of a PL decay only if the total recombination was caused by different populations of excited charges all decaying simultaneously via a first-order kinetic process. In reality, this does not occur: it is instead well established in the theory of interband recombination [11,12] that the electron-hole decay does not follow first-order kinetics, except for very low charge densities. Moreover, there is just one population of excited charges involved in a direct-bandgap semiconductor, namely the free carrier population (electrons energetically close to the minimum of the conduction band and holes close to the maximum of the valence band).
Performing quantitative analyses that involve the actual physical parameters can be especially relevant for gaining insights on the charge-carrier recombination mechanisms. This is particularly valid in regard to semiconductors relevant for their light-emitting and/or photocatalytic and/or gas-sensing properties. The validity of this statement is quite evident in regard to the light-emitting properties, as the efficiency of interband PL is a prerequisite for achieving efficient amplified spontaneous emission and lasing [13].
In regard to photocatalytic properties, the key point is that a photocatalyst provides free charges on the semiconductor surface for chemical redox reactions of adsorbed molecules. Once used for a reduction or an oxidation reaction, the free charge is no longer available for PL emission: therefore, PL and photocatalytic processes are in competition. For this reason, monitoring of the PL intensity is often employed to probe photocatalytic efficiency and to monitor surface reactions [14][15][16][17].
Additionally, the photo-generation of excited charges always occurs in a region close to the sample surface. Hence, changes in the gaseous environment surrounding the sample typically determine changes in the charge recombination probability, i.e., in the PL intensity and lifetime. For this reason, PL spectroscopy is a sensitive technique for analyzing surfaces [18], and optically-based gas sensing can be achieved by monitoring the PL lifetime and/or intensity modulation [19][20][21].
Zinc oxide (ZnO) is a metal oxide semiconductor that exhibits all the three mentioned functionalities. It is an efficient light emitter [22,23] thanks to its direct bandgap. It is also a well-known photocatalyst, relevant in the field of visible light-activated photocatalysis (also known as "solar" photocatalysis) and of photo-induced antibacterial surface functionalization [24][25][26]. Finally, ZnO is a gas-sensitive semiconductor employable as both a conductometric [27][28][29] and optical [30][31][32] sensor.
Based on these considerations, we focused the present work on developing a simple method for the data analysis of TRPL in ZnO, aiming to determine the actual physical quantities that are defined in the second-order rate equation for the electron-hole recombination, namely, the unimolecular lifetime, τ, the bimolecular coefficient, B and the equilibrium density of free carriers, N. In order to simplify the exposure, the definitions and physical meaning of these quantities are discussed in more detail in Appendix A.
In order to achieve the aim of this work, we firstly highlighted that the interband recombination in ZnO subjected to pulsed laser excitation at moderate values of optical fluences, is described with good precision by a total electron-hole rate of recombination of the second order (i.e., a second-order kinetic process). This is here clearly evidenced experimentally by a transition from an almost single-exponential decay at very low optical fluences, to a bimolecular recombination decay at moderate optical fluences. These different decay kinetics affect mainly the initial stage of the PL decay; this allowed us to focus on the initial slope of the PL decay curves and to define a procedure, based on the initial "instantaneous lifetime", that can provide a quantitative determination of the above-mentioned characterizing parameters.
The procedure is described in the next section, followed by the discussion and conclusions.
Materials and Methods
Measurements of continuous-wave, time-resolved and time-integrated photoluminescence-labelled as CWPL, TRPL and TIPL, respectively-were carried out at room temperature on (0001) oriented single crystal ZnO substrates purchased from CrysTec GmbH (hydrothermal growth, 10 × 5 mm 2 surface area, 0.5 mm thickness). Two as-received crystals belonging to the same batch were characterized by means of TRPL, TIPL and continuous-wave photoluminescence (CWPL). Laser excitation for CWPL measurements was provided by the 325 nm wavelength of an He-Cd laser (10 mW, laser spot size of approximately 4*10 −2 cm 2 ). Pulsed laser excitation for TRPL was provided by the third harmonic beam of a mode-locked Nd:YAG laser, characterized by the following parameters: 355 nm wavelength, 25 picoseconds pulse duration (average full width at half the maximum of the temporal profile of the laser pulses) and 10 Hz repetition rate. TIPL spectra Φ(λ) were determined by the time integration of the TRPL intensities φ(t), i.e.,: where ∆t = 2 ns is the time range of over which the TRPL spectra were acquired. The excitation laser beam was slightly focused on the sample surface, impinging with about a 50 • angle of incidence and a 10 −1 cm 2 laser spot area. The sample's reflectivity (R) for different values of laser excitation fluency was obtained by measuring the energy of impinging and reflected laser pulses by means of a pyroelectric joulemeter. A wedge window placed along the laser path was used to extract a fraction of excitation energy and to direct it towards an energy meter in order to monitor the laser energy.
The photoluminescence emission emerging from the excited surface was collected by means of an achromatic confocal lens system. The time-resolved PL spectral components were dispersed by a spectrometer (25 cm equivalent focal length) and detected via a streak camera optically coupled with the exit slit of the spectrometer. The acquisition electronics of the streak camera were externally triggered by the laser, allowing shot-byshot detection. The overall system provided simultaneous time-resolved and wavelengthresolved detection (overall experimental time and spectral resolution: 30 ps and 1 nm, respectively). The residual signal at excitation wavelength (355 nm) was eliminated by means of a sharp cut-off optical filter.
For each sample, multiple TRPL data were acquired using different values of the optical excitation fluency, F, defined as the laser pulse energy absorbed by the crystal, divided by the laser spot area at sample surface. It is worth underlining that the large thickness of the crystals (d = 500 µm) assures that all the samples were optically thick, i.e., αd 1 where α is the absorption coefficient of the ZnO crystal at 355 nm wavelength. Given a typical value of α ≈ 10 5 cm −1 , which corresponds to an optical penetration length of 100 nm, the Lamber-Beer low assures that 99.7% of the optical energy is absorbed within a thickness of about 0.3 µm. Considering that our crystals were 500 µm thick, the complete absorption of the optical energy which is not reflected away by the sample surface is assured.
Results
The results are presented in two separate subheadings, where the first describes the theoretical model used to set the analytical procedure of data analysis, while the second shows and discusses the experimental results obtained for ZnO bulk crystals.
Theoretical Model
The model presented in this work considers a n-type direct gap semiconductor in which an equilibrium density of free carriers, indicated by N, is present in the conduction band even in absence of photoexcitation, due to the partial ionization of shallow donors. In addition, photoexcitation by laser pulses induces the presence of non-equilibrium (i.e., time dependent) excess concentration, indicated by ρ(t). After pulsed laser excitation, the electronic fundamental state (i.e., equilibrium) is restored through the recombination of the excess carriers, whose time evolution is determined by the rate equation: where the quantity W represents the total recombination rate, defined as the total number of recombination events (including both radiative and non-radiative ones) per unit time and volume. It is worth noting that the total recombination rate is itself a function of the excess charge density, as it has been explicitly indicated.
We further indicate by W R and W NR the radiative and non-radiative recombination rates, respectively, so that W(ρ(t)) = W R (ρ(t)) + W NR (ρ(t)). Again, our nomenclature explicitly expresses the fact that both recombination rates depend on ρ.
As an interband electron-hole recombination occurs as a two-body scattering event, its probability is proportional to the product between the hole density p = ρ and the electron density n = N + ρ, leading to the following general expression of a radiative recombination rate [11]: where the coefficient B is the radiative bimolecular coefficient, and where the second equality defines the radiative unimolecular lifetime 1/τ R def = BN. The quantity W R (t) gives, by definition, the number of emitted (i.e., photoluminescent) photons per unit volume and time. Therefore, the experimentally measured time-resolved photoluminescence intensity φ(t) at each wavelength is proportional to the expression of W R (t), which is obtained by using in Equation (3) the function ρ(t) that solves the rate equation Equation (2).
It is immediately seen that a recombination rate W that is linearly proportional to ρ, would be trivial and unrealistic, as it would lead to a single-exponential decay for both the excess charge density and the photoluminescence intensity. Such a simple case is almost never encountered in the actual experimental results for photoluminescence decay in oxides and, even more relevantly, does not describe the experimental results discussed later.
Therefore, a less simple expression is required. The core of the present model is to consider the expression of W(ρ(t)) up to the second order in terms of the power of the ρ(t) variable, that is to use a generalized second-order recombination rate W (2) = aρ + bρ 2 and to solve the following rate equation: where the unimolecular lifetime τ appears in the definition of the linear term (i.e., a = 1/τ). The rate Equation (4) thus neglects the eventual contributions of the third order ρ 3 and is successive to the total recombination rate. The radiative recombination rate in interband transition is intrinsically of second order vs. charge density, as shown in Equation (2). Hence, to discuss the approximations implied by Equation (3), we have to consider the non-radiative recombination.
The main non-radiative recombination processes in semiconductors are the Shockley-Read-Hall (SRH) recombination and the Auger recombination. The SRH processes consist of the simultaneous capture of an electron and a hole at deep defect levels, and various kind of defects (extrinsic as well as intrinsic) can give place to SRH recombination, which thus significantly limit the radiative recombination probability, unless the material has high crystalline quality. On another hand, the Auger processes are three-body processes, in which the kinetic energy of a pair is transferred to a third charge carrier. As such, an Auger process involving three free carriers is proportional to the third power (ρ 3 ) of the injected carrier density and can hence become important at large densities of photogenerated carriers.
To summarize, the use of a second-order effective recombination rate is reasonable, as far as exceedingly high optical injection regimes are not probed experimentally. As shown in the next section, this hypothesis allows us to significantly simplify the analysis, while maintaining the main experimental features.
The solution of the second-order differential Equation (4) is: where ρ 0 = ρ(t = 0) is the initial density of photoinduced excess charge carriers, as it can be easily verified by using t = 0 and hence exp(−t/τ) = 1 in the expression. Inserting then the expression (5) into Equation (2), we get the following expression for time-resolved photoluminescence decay: The initial density of the excess charge ρ 0 is linked to the excitation fluence F by the equation ρ 0 = αF/ ω, which expresses the fact that one pair of free carriers (i.e., an electron and a hole) is generated for each photon of energy ω belonging to an excitation pulse of fluence F absorbed by the material. The 1:1 correspondence between photon density and charge carrier density assumes that multi-photon absorption processes are negligible in comparison with linear absorption. This is indeed correct in our case, in which supra-gap radiation is used to excite the material.
The peculiar characteristic of the function (6) is that for a long time (t τ) it tends to a single-exponential function, and thus shows a constant and ρ 0 -independent (i.e., fluence independent) logarithmic slope, while instead the initial temporal behavior significantly depends on the value of ρ 0 = αF/ ω, i.e., it depends on the laser fluence.
Extracting physically relevant information from the fitting of experimental data through fittings of experimental data using Equation (6), is complicated, due to the number of unknown quantities that appear in it. In this work, we propose as a useful procedure for data analysis, to extract an "instantaneous recombination rate" W e f f (t), by means of the logarithmic derivative of φ(t), i.e., : By integrating (6) we obtain φ(t) = φ(0) exp − t 0 w e f f (t )dt , which shows again that the simple case of a time-independent instantaneous recombination rate corresponds to simple single-exponential (first order) kinetics.
At a generic instant, t, the instantaneous recombination rate describes (by definition) the logarithmic slope of the φ(t) curve, which is related to the laser fluence, as mentioned before. Therefore, we have a link between the instantaneous recombination rate at instant t = 0 and the laser fluence; by simple algebra we can calculate the derivative in (6) at t = 0 and obtain the following relationship: By using the expression (8), it is possible to determine the initial recombination rate via a numerical derivative of the TRPL data. Next, the extraction of best-fit values of b, τ and N becomes possible by determining the ρ 0 quantities (i.e., the densities of photoinjected excess carrier) at different optical fluencies. This procedure avoids performing more complicated fittings of Equation (6).
Experimental Results and Discussion
Representative examples of a low-fluence CWPL spectrum and of TIPL spectra acquired for different values of laser fluence, are reported in Figure 1. The spectra refer to the ultraviolet range of the ZnO photoluminescence emission (excitonic emission).
Experimental Results and Discussion
Representative examples of a low-fluence CWPL spectrum and of TIPL spectra acquired for different values of laser fluence, are reported in Figure 1. The spectra refer to the ultraviolet range of the ZnO photoluminescence emission (excitonic emission). As expected for room temperature measurements, spectral features related to bound excitons and donor-acceptor pair recombination [33] are not evidenced, while an unstructured emission spectrum that peaked at about 3.28 eV was observed in all samples.
The curve (a) was obtained using the HeCd laser as the excitation source. As the HeCd is a continuous-wave source, it distributes the energy over long time periods and leaves stationary electron-hole density much lower (about four orders of magnitude lesser) than the instantaneous charge carrier density, determined by pulsed laser excitation. The PL spectra measured via picosecond pulsed excitation for different values of F (laser fluencies) are reported in curves (b)-(f).
The TIPL spectra have been measured to assure that the optical fluences involved are weak enough to avoid the activation of the "p-line" (caused by exciton-exciton scattering), observable as a PL contribution peaking at 3.18 eV [34]. It is also possible to infer that the experimental conditions do not cause the dissociation of excitons due to the screening of the Coulomb potential by electron-hole plasma, which would lead to an emission peak at even lower photon energies [34,35], which are not observed here.
No significant shift of the PL peak was noticed for excitation fluencies of a few μJ/cm 2 (see curve (b) of Figure 1), while a red shift of PL peak emission vs. laser fluency is evidenced, with values compatible with the bandgap narrowing caused by electronelectron and hole-hole exchange interactions [36]. The maximum measured PL shift was about 50 meV for the employed values of laser fluences. This value is small enough to ensure that that neither the activation of the p-line nor the plasma-induced exciton dissociation occurs. Hence, we can still model the measured PL in the frame of interband recombination.
The PL spectra of Figure 1 exhibit a small energy shift toward lower energies, but retain the same spectral widths and shapes at increasing excitation fluences. These facts suggest that the excitons retain their individual character and binding energy at investigated fluencies, while the shift of the emission peak results from the bandgap narrowing. This is supported by the evidence shown in Figure 2, where the shift Δ of emission peak energy is plotted vs. the excitation laser fluency, F. Notably, a power law provides a good fitting: As expected for room temperature measurements, spectral features related to bound excitons and donor-acceptor pair recombination [33] are not evidenced, while an unstructured emission spectrum that peaked at about 3.28 eV was observed in all samples.
The curve (a) was obtained using the HeCd laser as the excitation source. As the HeCd is a continuous-wave source, it distributes the energy over long time periods and leaves stationary electron-hole density much lower (about four orders of magnitude lesser) than the instantaneous charge carrier density, determined by pulsed laser excitation. The PL spectra measured via picosecond pulsed excitation for different values of F (laser fluencies) are reported in curves (b)-(f).
The TIPL spectra have been measured to assure that the optical fluences involved are weak enough to avoid the activation of the "p-line" (caused by exciton-exciton scattering), observable as a PL contribution peaking at 3.18 eV [34]. It is also possible to infer that the experimental conditions do not cause the dissociation of excitons due to the screening of the Coulomb potential by electron-hole plasma, which would lead to an emission peak at even lower photon energies [34,35], which are not observed here.
No significant shift of the PL peak was noticed for excitation fluencies of a few µJ/cm 2 (see curve (b) of Figure 1), while a red shift of PL peak emission vs. laser fluency is evidenced, with values compatible with the bandgap narrowing caused by electron-electron and hole-hole exchange interactions [36]. The maximum measured PL shift was about 50 meV for the employed values of laser fluences. This value is small enough to ensure that that neither the activation of the p-line nor the plasma-induced exciton dissociation occurs. Hence, we can still model the measured PL in the frame of interband recombination.
The PL spectra of Figure 1 exhibit a small energy shift toward lower energies, but retain the same spectral widths and shapes at increasing excitation fluences. These facts suggest that the excitons retain their individual character and binding energy at investigated fluencies, while the shift of the emission peak results from the bandgap narrowing. This is supported by the evidence shown in Figure 2, where the shift ∆E p of emission peak energy is plotted vs. the excitation laser fluency, F. Notably, a power law provides a good fitting: where F and ∆E p are reported in µJ/cm 2 and in eV units, respectively. The same function as in Equation (9) has been employed by Schmid et al. [37] to describe the gap shrinkage in heavily n-doped and p-doped silicon versus doping concentration, while Lu and co-workers [36] showed that the same empirical power law behavior also fits with the experimental bandgap narrowing in aluminum-doped ZnO films. Notably, the value γ = 0.43 ± 0.03 found for the data in Figure 2 is very close to the one (γ = 0.40) obtained and reported in the mentioned work [36].
where F and Δ are reported in μJ/cm 2 and in eV units, respectively. The same function as in Equation (9) has been employed by Schmid et al. [37] to describe the gap shrinkage in heavily n-doped and p-doped silicon versus doping concentration, while Lu and coworkers [36] showed that the same empirical power law behavior also fits with the experimental bandgap narrowing in aluminum-doped ZnO films. Notably, the value = 0.43 ± 0.03 found for the data in Figure 2 is very close to the one ( = 0.40) obtained and reported in the mentioned work [36]. Figure 1 and the "zero value" of the shift refers to the low fluence measurement obtained via the use of an HeCd laser. The blue curve is a best fit of the data obtained through the function in Equation (9).
As mentioned before, the linear absorption of the excitation laser pulses dictates the relationship = /ℏ between the initial density of photogenerated charges and the laser fluence F, where α is the absorption coefficient at the photon energy ℏ = 3.49 eV, corresponding to the laser wavelength of 355 nm. As mentioned before, the proportionality relationship between and might be modified by nonlinear contributions to the optical absorption. In order to check for eventual nonlinear effects, measurements of crystal reflectivity at 355 nm were performed at various excitation fluencies. As shown in Figure 3, no fluence-dependent variation in reflectivity was observed in the investigated range. This indicates that it is indeed reasonable to neglect nonlinear optical effects and use the relationship = /ℏ to describe the experimental data. Figure 1 and the "zero value" of the shift refers to the low fluence measurement obtained via the use of an HeCd laser. The blue curve is a best fit of the data obtained through the function in Equation (9).
As mentioned before, the linear absorption of the excitation laser pulses dictates the relationship ρ 0 = αF/ ω between the initial density of photogenerated charges and the laser fluence F, where α is the absorption coefficient at the photon energy ω = 3.49 ev, corresponding to the laser wavelength of 355 nm.
As mentioned before, the proportionality relationship between ρ 0 and F might be modified by nonlinear contributions to the optical absorption. In order to check for eventual nonlinear effects, measurements of crystal reflectivity at 355 nm were performed at various excitation fluencies. As shown in Figure 3, no fluence-dependent variation in reflectivity was observed in the investigated range. This indicates that it is indeed reasonable to neglect nonlinear optical effects and use the relationship ρ 0 = αF/ ω to describe the experimental data.
where F and Δ are reported in μJ/cm 2 and in eV units, respectively. The same function as in Equation (9) has been employed by Schmid et al. [37] to describe the gap shrinkage in heavily n-doped and p-doped silicon versus doping concentration, while Lu and coworkers [36] showed that the same empirical power law behavior also fits with the experimental bandgap narrowing in aluminum-doped ZnO films. Notably, the value = 0.43 ± 0.03 found for the data in Figure 2 is very close to the one ( = 0.40) obtained and reported in the mentioned work [36]. Figure 1 and the "zero value" of the shift refers to the low fluence measurement obtained via the use of an HeCd laser. The blue curve is a best fit of the data obtained through the function in Equation (9). As mentioned before, the linear absorption of the excitation laser pulses dictates the relationship = /ℏ between the initial density of photogenerated charges and the laser fluence F, where α is the absorption coefficient at the photon energy ℏ = 3.49 eV, corresponding to the laser wavelength of 355 nm. As mentioned before, the proportionality relationship between and might be modified by nonlinear contributions to the optical absorption. In order to check for eventual nonlinear effects, measurements of crystal reflectivity at 355 nm were performed at various excitation fluencies. As shown in Figure 3, no fluence-dependent variation in reflectivity was observed in the investigated range. This indicates that it is indeed reasonable to neglect nonlinear optical effects and use the relationship = /ℏ to describe the experimental data. A reference value of α(λ = 355 nm) = 1.5 · 10 5 cm −1 was taken from optical transmittance analysis of ZnO single-crystal films, as reported by Muth et al. [38]. Consistent values for absorption coefficient were also reported in earlier optical transmission analysis of ZnO platelets [39] and in ellipsometry analysis of ZnO single crystals [40]. Using this reference value, densities of photogenerated charge carriers of 2 × 10 18 cm −3 , 5.5 × 10 19 cm −3 , 2.4 × 10 20 cm −3 and 3.9 × 10 20 cm −3 were obtained for curves from (b) to (f). We would like to underline that, regardless of the value of the absorption coefficient, the red shift of the photoluminescence spectra becomes evident at an excitation fluency of 100 µJ/cm 2 . This suggests probing optical excitation densities at such a value and above for the TRPL experiments. Figure 4 shows some of the peak-normalized TRPL curves measured for the ZnO samples at different values of optical fluence. The reported PL signal is obtained by integration over the UV peak. The curves shown on Figure 4 are just a representative excerpt of the set of measurements performed. It is also to be underlined that the data reported in Figure 4 are different from the data used to determine the TIPL spectra of Figure 1, so that there is no correspondence between the optical fluencies in Figures 1 and 4. transmittance analysis of ZnO single-crystal films, as reported by Muth et al. [38]. Consistent values for absorption coefficient were also reported in earlier optical transmission analysis of ZnO platelets [39] and in ellipsometry analysis of ZnO single crystals [40]. Using this reference value, densities of photogenerated charge carriers of 2 × 10 cm , 5.5 × 10 cm , 2.4 × 10 cm and 3.9 × 10 cm were obtained for curves from (b) to (f).
We would like to underline that, regardless of the value of the absorption coefficient, the red shift of the photoluminescence spectra becomes evident at an excitation fluency of ~100 μJ/cm 2 . This suggests probing optical excitation densities at such a value and above for the TRPL experiments. Figure 4 shows some of the peak-normalized TRPL curves measured for the ZnO samples at different values of optical fluence. The reported PL signal is obtained by integration over the UV peak. The curves shown on Figure 4 are just a representative excerpt of the set of measurements performed. It is also to be underlined that the data reported in Figure 4 are different from the data used to determine the TIPL spectra of Figure 1, so that there is no correspondence between the optical fluencies in Figure 1 and Figure 4. The data in Figure 4 evidence that different decay kinetics occur for different values of excitation fluence, and thus for different initial values of the charge densities. At larger ρ 0 values, the TRPL signal starts with a fast non-exponential decay, whose slope is fluencedependent and which dominates for about the first 200 picoseconds. Successively, a slower decay is observed, characterized by an almost constant and laser fluency-independent logarithmic slope of the signal. Once the laser fluency is decreased, the fast component of TRPL becomes less evident, until the decay curve tends to a simple single-exponential.
As discussed previously, this behavior shows the characteristic features for a secondorder recombination rate and agrees with the PL decay reported in Equation (6). Therefore, from decay curves acquired at different values of the laser fluence F, the instantaneous lifetime was determined via numerical derivative and the initial values were fitted as a function of F by using Equation (8). The results are reported in Figure 5.
The analysis results are consistent with the predictions, as they give a lower value for (0), which is close to the fluence-independent slope (i.e.,1/ ) of the TRPL curves.
On the other hand, (0) increases at larger values of the laser fluence. The black curves represent the best fit of the data obtained through Equation (8) and using = /ℏ . The values for the quantities N, b and τ obtained from best fits are reported in Additional considerations can be elaborated by considering the second order expression of the non-radiative rate of Shockley-Read-Hall recombination. For recombination centers whose energy levels are close to the center of the bandgap, this rate is given by [12]: The analysis results are consistent with the predictions, as they give a lower value for W e f f (0), which is close to the fluence-independent slope (i.e., 1/τ) of the TRPL curves. On the other hand, W e f f (0) increases at larger values of the laser fluence. The black curves represent the best fit of the data obtained through Equation (8) and using ρ 0 = αF/ ω. The values for the quantities N, b and τ obtained from best fits are reported in Table 1. Additional considerations can be elaborated by considering the second order expression of the non-radiative rate of Shockley-Read-Hall recombination. For recombination centers whose energy levels are close to the center of the bandgap, this rate is given by [12]: where τ min and τ maj are the SRH lifetimes for the minority carrier and majority carrier, respectively [12]. In Equation (10), we also show the expansion of the SRH rate up to the second order in the density of photogenerated carriers, so that it is possible to estimate the radiative bimolecular coefficient B and the SRH non-radiative minority lifetime τ min . In fact, by neglecting third-order terms (i.e., Auger rate) in the total recombination rate W (2) = W R + W SRH , and using Equations (3), (4) and (10), we get the relations 1/τ = BN + 1/τ SRH min and b = B − τ SRH maj /N(τ SRH min ) 2 .
Rough but reasonable estimations can be made by considering similar values for the non-radiative lifetimes of the majority and minority carriers (i.e., τ SRH min ∼ = τ SRH maj ) [12], so that we obtain values of 0.8 ns for τ min (for both samples) and values of B ≈ 1 × 10 −10 cm 3 s −1 (sample 1) and B ≈ 0.7 × 10 −10 cm 3 s −1 (sample 2) for the radiative bimolecular coefficient B, which are reasonably close to room temperature B values in epitaxial GaAs [41,42] and GaAs/AlGaAs quantum wells [43].
Conclusions
In conclusion, we proposed here a relatively simple method for the extraction of the parameters defining the second-order kinetics that rule the interband recombination in ZnO crystals. The goal of this work was to highlight the possibility of extracting well-defined (i.e., physically meaningful) parameters from TRPL data without resorting to phenomenological fittings. This aim is, to our opinion, especially important for semiconductors that exhibit significant light-emitting, photocatalytic or gas-sensing properties. In this sense, ZnO is a particularly suitable material, as it exhibits all these three functionalities.
The key point allowing the method described here is that, within a suitable range of laser excitation fluences, the interband recombination of ZnO crystals is very well described using second-order kinetics for the total recombination rate, characterized by an onset of non-exponential (bimolecular) decay observed for carrier densities of about 10 19 cm −3 . This allowed us to determine and fit the initial values of the "instantaneous" recombination rate (Equations (7) and (8)) measured at different excitation fluencies. The procedure allowed us to extract the parameters b and τ by simple data fittings and may represent a useful tool for quantitatively characterizing the charge recombination in ZnO.
Furthermore, we also proved that the procedure provides information on radiative vs. non-radiative recombination contributions by estimating the bimolecular radiative coefficient B and the non-radiative Shockley-Read-Hall lifetime, without the need to perform further experiments (e.g., measuring the quantum efficiency of PL emission).
We propose the procedure developed here as a tool for sketching a quantitative outline of the ZnO exciton recombination dynamics, going beyond the use of TRPL analysis as a mere characterization technique.
Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A
To improve the readability of the manuscript, some definitions and explanations are reported here in the Appendix.
Unimolecular lifetime, τ: This is the quantity often referred to as simply "lifetime". It defines the linear (i.e., first order) term of a decay rate in a kinetic equation of the form −dρ/dt = ρ/τ. It is clear to see that the solution of this equation is a single-exponential. This kinetic can describe the recombination for very low optical injection levels, i.e., when the density of photogenerated carriers is low compared to the density of equilibrium carriers due to the doping. However, it cannot describe the recombination between two photogenerated carriers, which is a bimolecular recombination. Bimolecular coefficient, b: This describes the probability of a recombination (either radiative or non-radiative) between a photogenerated hole and a photogenerated electron, according to a quadratic (two-body) rate equation of the form −dρ/dt = bρ 2 , whose solution is clearly non-exponential.
Radiative bimolecular coefficient, B: This is the analogue of the b coefficient for the radiative recombination processes only, i.e., Bnp = Bρ(ρ + N) is the number of photons produced via electron-hole radiative recombination per unit time and volume.
Once terms of third order and higher (i.e., ρ 3 , ρ 4 etc.) are neglected, the number of recombination events (either radiative or non-radiative) per unit time and volume shall include a linear ("unimolecular") term and a quadratic ("bimolecular") term, which lead to the general expression of Equation (4). It is clear that for densities of optically-excited carriers that are negligible with respect to the equilibrium charge density N (the one caused by doping), the overall kinetic will be dominated only by the first term, i.e., a single-exponential PL decay shall be expected. For higher values of ρ, the deviation from single-exponential is expected to become more and more relevant. | 8,407 | 2022-02-01T00:00:00.000 | [
"Physics"
] |
The COsmic-ray Soil Moisture Interaction Code ( COSMIC ) for use in data assimilation
Soil moisture status in land surface models (LSMs) can be updated by assimilating cosmic-ray neutron intensity measured in air above the surface. This requires a fast and accurate model to calculate the neutron intensity from the profiles of soil moisture modeled by the LSM. The existing Monte Carlo N-Particle eXtended (MCNPX) model is sufficiently accurate but too slow to be practical in the context of data assimilation. Consequently an alternative and efficient model is needed which can be calibrated accurately to reproduce the calculations made by MCNPX and used to substitute for MCNPX during data assimilation. This paper describes the construction and calibration of such a model, COsmic-ray Soil Moisture Interaction Code (COSMIC), which is simple, physically based and analytic, and which, because it runs at least 50 000 times faster than MCNPX, is appropriate in data assimilation applications. The model includes simple descriptions of (a) degradation of the incoming high-energy neutron flux with soil depth, (b) creation of fast neutrons at each depth in the soil, and (c) scattering of the resulting fast neutrons before they reach the soil surface, all of which processes may have parameterized dependency on the chemistry and moisture content of the soil. The site-to-site variability in the parameters used in COSMIC is explored for 42 sample sites in the COsmic-ray Soil Moisture Observing System (COSMOS), and the comparative performance of COSMIC relative to MCNPX when applied to represent interactions between cosmic-ray neutrons and moist soil is explored. At an example site in Arizona, fast-neutron counts calculated by COSMIC from the average soil moisture profile given by an independent network of point measurements in the COSMOS probe footprint are similar to the fast-neutron intensity measured by the COSMOS probe. It was demonstrated that, when used within a data assimilation framework to assimilate COSMOS probe counts into the Noah land surface model at the Santa Rita Experimental Range field site, the calibrated COSMIC model provided an effective mechanism for translating model-calculated soil moisture profiles into aboveground fast-neutron count when applied with two radically different approaches used to remove the bias between data and model.
Introduction
Until recently area-average soil moisture at the hectometer horizontal scale has been difficult and costly to measure because of the need to take many point samples, but with the advent of the cosmic-ray method (Zreda et al., 2008(Zreda et al., , 2012Desilets et al., 2010) it is now feasible with a single instrument. However, a complicating aspect of measuring soil moisture using this method is that the volume of soil measured in the vertical varies with soil moisture content (Franz et al., 2012a).
One potentially important use of area-average soil moisture measured with the cosmic-ray method is through data assimilation methods to update the value of soil moisture states represented in the LSMs which are used to describe surface-atmosphere exchanges in meteorological and hydrological models. Typically such LSMs calculate (among many other things) time-varying estimates of soil moisture content in discrete layers of soil defined within the vertical soil profile. In order to make use of the area-average soil moisture provided by the cosmic-ray method, it is necessary to a. diagnose if there is a discrepancy in the modeled soil moisture status from the aboveground measured fastneutron count; and b. interpret knowledge of the extent of any discrepancy back into the LSM with weighting between layers reflecting their relative influence on the aboveground measured fast-neutron count.
This requires the availability and use of an accurate model to interpret the modeled soil moisture profiles in terms of the aboveground fast-neutron count.
In principle the required model needed to make such interpretation exists, specifically the Monte Carlo N-Particle eXtended (MCNPX: Pelowitz, 2005) neutron transport code, which was much used in establishing the cosmic-ray method (Zreda et al., 2008(Zreda et al., , 2012Desilets et al., 2010) currently being deployed in the COsmic-ray Soil Moisture Observing System (COSMOS: http://cosmos.hwr.arizona.edu/; Shuttleworth et al., 2010;Zreda et al., 2011Zreda et al., , 2012. Given the specified chemistry of the atmosphere and soil (including the amount of hydrogen present as water in the system), the MC-NPX code uses knowledge of nuclear collisions and libraries of nuclear properties for these constituents to track the life history of individual, randomly generated, incoming cosmic rays and their collision products through the atmosphere and in the soil. The code then counts the resulting fast neutrons (we use those in the range 10 eV to 100 eV) that enter a defined detector volume above the ground. In principle the MC-NPX code could be used in data assimilation applications to define (a) and (b) in the last paragraph. However, although accurate, the MCNPX code uses the time-consuming Monte Carlo computational method, and this means its use in data assimilation applications is impractical. Therefore an alternative model is needed which can efficiently reproduce the belowground physics, the resulting aboveground count rate and the belowground vertical source distribution of fast neutrons simulated by MCNPX. This paper describes the construction and calibration of such a model, the COsmic-ray Soil Moisture Interaction Code (COSMIC), which is simple, physically based and analytic, and which runs much faster than MCNPX because the nuclear processes and collision cross sections that are explicitly represented in MCNPX are re-captured in parameters that have dependency on the site-specific soil properties. These parameters are calibrated using multi-parameter optimization techniques against MCNPX calculations for a suite of hypothetical soil moisture profiles.
Physical processes represented in COSMIC
The COSMIC model assumes there are three dominant processes involved in generating the fast neutrons detected above moist soil (see Fig. 1). It is first assumed that there is an exponential reduction with depth in the number of the high-energy neutrons that are available to create fast neutrons at any level in the soil. Calculations made with MC-NPX indicate that assuming such an exponential reduction in neutron flux is appropriate. There is reduction due to interaction both with the (dry) soil and with the water that is present in the soil. The exponential reduction therefore depends on two length constants L 1 and L 2 , in units of g per cm 2 , corresponding to interaction with the soil and the water (hydrogen), respectively. The mass of water includes both lattice water, i.e., that which is in the mineral grains and bound chemically with soil and considered fixed in time, and the pore water which is available to support transpiration or drainage and which consequently changes with time. Thus, the number of high-energy neutrons available at depth z in the soil is given by where N 0 he is the number of high-energy neutrons at the soil surface, m s (z) and m w (z) are respectively the integrated mass per unit area of dry soil and water (in g cm −2 ) between the depth z and the soil surface, and L 1 and L 2 (in g per unit area) are respectively determined by the chemistry of the soil and its total water content, including any chemically bound lattice water.
Second, it is assumed that at each depth z the number of fast neutrons created in the soil is proportional to the product of the number of high-energy neutrons available at that depth with the local density of dry soil per unit soil volume and the local density of soil water per unit soil volume at that depth, assuming the relative efficiency of creation of fast neutrons by soil is a factor α of the efficiency of their creation by water. Consequently, the number of fast neutrons created in the soil in the plane at level z is given by Hydrol. Earth Syst. Sci., 17, 3205-3217, 2013 www.hydrol-earth-syst-sci.net/17/3205/2013/ where C is a (unitless) "fast-neutron creation" constant for pure water, ρ s (z) is the local bulk density of dry soil and ρ w (z) the total soil water density, including lattice water. It is assumed that the direction in which the fast neutrons are generated at level z is isotropic, i.e., that they leave with equal probability in all directions.
Finally it is assumed that the fraction of fast neutrons originating in the soil in the plane at level z that are detected above the ground are reduced exponentially by an amount related to the distance traveled between the point of origin in this plane and the detector at the surface. There is then little reduction in the neutron count in the air between the soil surface and the fast-neutron detector mounted just a few meters above the surface. The reduction in fast neutrons in the moist soil is assumed to follow a functional form similar to that in Eq. (1), i.e., an exponential reduction, as for highenergy neutrons, but with different length constants L 3 and L 4 , in units of g cm −2 , corresponding to attenuation by soil and by (total) soil water, respectively. However, because the direction in which fast neutrons are generated at level z is assumed to be isotropic, fast neutrons reaching the surface will travel further if they do not originate directly below the detector, rather from a point that is more distant in the horizontal plane at level z. To allow for this it is necessary to calculate the integrated average of the attenuation for all points in this plane to the detector, with the attenuation distance being inversely proportional to cos (θ), where θ is the angle between the vertical below the detector and the line between the detector and each point in the plane; see Fig. 2. Consequently, the integrated average attenuation of the fast neutrons generated at level z before they reach the detector is given by the function A(z): which, because there is assumed symmetry around the vertical through the detector, reduces to The value of A(z) can be found numerically, but for efficiency it could also be adequately calculated using the approach described in Appendix A.
Combining the representations of the three physical processes considered in COSMIC described above, the analytic function describing N COSMOS , the number of fast neutrons reaching the COSMOS probe at a near-surface measurement point is The source volume element of fast neutrons created in the plane at depth z in the soil which may reach the measurement point P, but whose number is attenuated by an exponential factor with length constants L3 and L4 (in g per unit area), these being respectively determined by the chemistry of the soil and by the total water content of the soil, including lattice water.
Fig. 2.
The source volume element of fast neutrons created in the plane at depth z in the soil which may reach the measurement point P but whose number is attenuated by an exponential factor with length constants L 3 and L 4 (in g per unit area) -these being respectively determined by the chemistry of the soil and by the total water content of the soil, including lattice water.
Note that in Eq. (5), the product of the two constants (CN 0 ) that appears in Eq.
(2) has been replaced by a single constant, N, because the values of C and N 0 cannot be separately determined from a comparison between calculations made using COSMIC and MCNPX.
Determining the parameters to be used in COSMIC
To determine the values of the (in some cases site-specific) parameters to be used in COSMIC, at 42 selected sites in the COSMOS network (see Fig. 3) for which the required data were available at the time of this analysis, simulations using COSMIC were calibrated against equivalent calculations made with the MCNPX model. The MCNPX calculations were made using the site-specific COSMOS probe calibration based on gravimetric samples (see, for example, Franz et al., 2013a, b), corrected for the effect of atmospheric humidity (see Rosolem et al., 2013), and with site-specific bulk density of the soil, soil chemistry and lattice water content (see Table 2 in Zreda et al., 2012, for values).
Because L 2 and L 4 relate to attenuation by water alone, their values are independent of the soil chemistry of the site and they can be determined by substituting pure water for dry soil in MCNPX and COSMIC calculations. A simulation with MCNPX was made with pure water substituting for soil, and an exponential function then fitted to the calculated reduction in high-energy neutrons with depth calculated by MCNPX for pure water to determine L 2 . The original San Pedro site was then selected for determining L 4 and the required value of the parameter N first defined at this site. This was accomplished by first optimizing the values of all remaining four COSMIC parameters (N, α, L 3 , L 4 ) at this site, with L 2 given as previously discussed and L 1 computed directly from MCNPX, in a similar manner to that described below. Once N is determined, COSMIC is configured to simulate pure water, and the parameter L 4 is fine-tuned to match the same neutron count obtained directly from MCNPX at the San Pedro site (after appropriate scaling using the F term described in the last paragraph of this section and shown in Table 1). Notice that for pure water simulations, the terms associated with parameters α, L 1 , and L 3 no longer appear in Eq. (5). Based on these pure water simulation comparisons, the values of L 2 and L 4 were set to 129.1 and 3.16 g cm −2 at all COSMOS sites.
The value of L 1 is easily determined for each site by running MCNPX with dry soil that has the site-specific soil chemistry and then fitting an exponential function to the calculated exponential reduction in high-energy neutrons with depth simulated by MCNPX (analogous to the method used to determined L 2 described above). Although the value of L 1 may depend on the soil chemistry present, our simulations with MCNPX at the 42 COSMOS sites considered in this study suggest that L 1 is only weakly related to soil chemistry, with site-to-site variability around the mean value for all sites being just ∼ 1 %. On this basis, adopting a fixed value equal to 162.0 g cm −2 irrespective of site is a reasonable assumption.
Data from individual sites in the COSMOS network are corrected for site to site differences in elevation and cutoff rigidity but local variability remains, likely associated with site-to-site differences in soil chemistry or vegetation cover. Individual site calibration of sensors is therefore required to allow for the fact that the observed neutron flux intensity at calibration does not necessarily equal the neutron flux intensity calculated by MCNPX when run with the soil chemistry and water content observed at calibration; see the final paragraph in Sect. 4. The values of the site-specific constants N, α and L 3 at all sites were then determined using multiparameter optimization techniques against calculations made using MCNPX. At each site calculations of the aboveground fast-neutron count are made using MCNPX for the 22 hypothetical profiles of volumetric water content illustrated in Fig. 4, i.e., for 10 profiles with different uniform volumetric water content, and 12 with different linear gradients of volumetric water content to a depth of 1 m and with uniform volumetric water content below 1 m. One criterion used in parameter optimization to define the preferred values of N, α, and L 3 is the weighted mean absolute error (MAE) between the aboveground fast-neutron counts calculated using the COSMIC model and the equivalent counts calculated by MCNPX with the same profiles. In each case, the weighted MAE is calculated based on the individual differences between the COSMIC neutron flux and MCNPX neutron flux for each profile, in absolute terms, and weighted by the probability density function of soil moisture historically observed at each site, with the most such commonly observed soil moisture values weighted to be twice as important as the least commonly observed value. The second criterion used in the optimization was that the cumulative contribution to aboveground fast neutrons as a function of depth given by the COS-MIC model matches that calculated by MCNPX as reported Table 1. Site-specific values of latitude and longitude; ρ s (g cm −3 ), θ lattice (m 3 m −3 ) and F ; and the parameters N, α (cm 3 g −1 ) and L 3 (g cm −2 ) obtained by calibrating the COSMIC model against MCNPX at the 42 COSMOS sites shown in Fig. 3 with L 1 = 162.0 g cm −2 , L 2 = 129.1 g cm −2 , and L 4 = 3.16 g cm −2 .
Latitude
Longitude by Zreda et al. (2008); i.e., at the site the cumulative contribution has a 2-e folding depth of around 0.76 m for a prescribed uniform volumetric water content of 0 %, and around 0.12 m for a prescribed uniform volumetric water content of 40 %, with zero lattice water content in both cases.
The multi-algorithm genetically adaptive multi-objective (AMALGAM) method (Vrugt and Robinson, 2007) was used to solve this multi-criteria minimization problem. AMAL-GAM contains highly desirable features for model optimization which facilitate parameter convergence, such as the use of multi-operator search and self-adaptive offspring creation, as well as the implementation of population-based elitism search. The initial parent population of size n is generated using Latin hypercube sampling (McKay et al. 1979). The fast non-dominated sorting algorithm approach (Deb et al., 2002) is used to assign the Pareto-rank for multiple criteria. Subsequent generation of the offspring (with the same size n) occurs with the use of k operators. The approach adopted in this study, which is similar to that presented by Rosolem et al. (2012), uses a population of size n = 100, and number of operators (search strategies) k = 4, and set the maximum number of generations, s = 1000, so that the total number of simulations (s × n) is 100 000.
This multi-parameter optimization was made at all 42 sites considered in this study to obtain the site-specific preferred values of N, α, and L 3 when the values of L 1 , L 2 and L 4 are specified to be 162.0, 129.1 and 3.16 g cm −2 , respectively. The resulting optimal parameters are given in Table 1 (the factor F given in column four of this table is discussed and used later in Sect. 5). Figure 5 summarizes the overall results of the multi-parameter optimization procedure, given the value of the difference between the simulated neutron count given by COSMIC (with optimized parameters) and the equivalent neutron count scaled from MCNPX, normalized by the MCNPX count (represented by colors for each site and each hypothetical soil moisture profile). Because MCNPX is a Monte Carlo model, the neutron count given by MCNPX is subject to random sampling errors of the order 1 %, and this contributes to some of the normalized differences illustrated in Fig. 5. For a substantial majority of the sites and hypothetical soil moisture profiles the normalized difference between the COSMIC-and MCNPX-simulated neutron counts is within the range 2-3 %, and when averaged over all sites the normalized difference is much less than this (Fig. 5, bottom row). This range in normalized difference is comparable to the measurement uncertainty in the COSMOS probe and the sampling error in the soil moisture field at probe calibration, including for the drier soil profiles for which the differences are greatest.
Correlations and dependencies of optimized parameters
It is of interest to investigate the extent to which the sitespecific optimized values of N, α and L 3 are correlated with each other and with the site-specific values of ρ s , the average bulk density for the soil in g cm −3 , and θ lattice , the lattice water content of the soil in m 3 m −3 . In practice, there is no evidence of correlation between the site-specific value of the parameter N and the site-specific values of ρ s , α and L 3 : linear correlation of these three parameters with N gives R 2 values of 0.01, 0.19, and 0.01, respectively. There is also no evidence of correlation between the site-specific optimized values of α and N with θ lattice at each site (R 2 = 0.04 and 0.06, respectively), and little evidence of correlation of L 3 with θ lattice (R 2 = 0.30). However as Fig. 6 shows, the sitespecific values of L 3 and α both exhibit evidence of correlation with ρ s , the bulk density for the soil at each site, and the site-specific values of L 3 and α are also mutually correlated. Arguably L 3 and α are indeed both independently correlated with ρ s , but the possibility exists that one of the parameters (likely L 3 ) is correlated, and the apparent correlation of the second parameter (α) is because the process of optimization is not able to clearly separate these two variables because their influence on N COSMOS calculated by Eqs. (3) and (5) is to change its value in opposite directions.
It is worth noting that, in physical terms, a strong correlation between L 3 and ρ s implies the attenuation of fast neutron by (dry) soil is not well described as an exponential decay with a simple single length constant that is independent of the density of soil as assumed in COSMIC. Instead the effective value of the length constant appears to be a near-linear function of soil density. Similarly a (true) correlation between α and ρ s implies that the creation of fast neutron from high neutrons is not perfectly described as a linear function of the local density of dry soil; i.e., in Eq. (2) the product [αρ s ] becomes [0.404(ρ s ) − 0.101(ρ s ) 2 ]. It is possible that the observed correlations of L 3 and α with ρ s may be useful for COSMOS sites where a multi-parameter optimization against MCNPX is not feasible because approximate estimates of L 3 and α might then be made from measured value of ρ s using the following equations: The marked variability in the site-specific optimized values of the parameter N must reflect substantial variability in one or both of the component constants C or N 0 he . However, there should be limited variability in N 0 he because the site-specific neutron calculations given by MCNPX against which calibration was made were corrected for local station effects using a scaling factor to account for differences in cosmic ray intensity as a result of the elevation/cutoff rigidity of the site where the probe is located (for details see Desilets and Zreda, 2003). The contributing variability is therefore presumably primarily associated with the effective value of C. This siteto-site variability is intrinsic to the COSMOS array (rather than a feature associated with the COSMIC model) and is present in the site-specific factor F (given in column 4 of Table 1). F is the ratio between the number of counts observed during COSMOS probe calibration at a specific site and the calculated neutron flux intensity given by MCNPX when run with the soil chemistry and water content (including lattice water) observed at each probe site during calibration. (Note: the factor 10 14 in F arises because MCNPX actually calculates neutron fluence, the time integration of neutron flux, rather than neutron count rate directly.) Figure 7 shows the strong interrelationship between the COSMIC parameter N found by multi-parameter optimization and the factor F : N = −24.46 + 63.16 × 10 −14 F.
The origin of the real site-to-site variability in F across the COSMOS array is currently under investigation. It is possible there is some remnant contribution to variability in F associated with the location and altitude of the probe although the neutron count rates were corrected for these (Desilets and Zreda, 2003). It is also possible that differences in the ambient water vapor content of the air during probe calibration may make some contribution to the variability in F at the level of a few percent (for details, see Rosolem et al., 2013). Otherwise the variability in F is presumably associated with site-to-site differences in soil chemistry or more likely vegetation cover (Franz et al., 2013a, b).
Application of the COSMOS probe at the Santa Rita study site
We tested COSMIC using soil moisture data from a COS-MOS probe and from a distributed sensors network at the Santa Rita Experimental Range field site in southern Arizona. A total of 180 time domain transmissivity (TDT) sensors (Fig. 8a) were installed (Franz et al., 2012b) in 18-paired profiles at 10, 20, 30, 50 and 70 cm within the footprint of the COSMOS probe (Fig. 8b). Figure 8c shows a comparison between the fast-neutron count observed by the COS-MOS probe and that calculated from the area-average soil moisture as measured with TDT sensors using MCNPX and COSMIC. Overall the COSMIC-derived fast-neutron intensity compares quite well with measurements from the COS-MOS probe, and (as should be expected) it compares extremely well with the fast-neutron intensity computed using MCNPX. In some cases, the after-rainfall response is slower than the COSMOS probe because the area-average . Relationship between the COSMIC parameter N found by multi-parameter ization and the factor F, this being the ratio between the number of counts observed the COSMOS probe calibration at a specific site and the calculated neutron flux ity given by MCNPX when run with the soil chemistry and water content (including water) observed at each probe site during calibration. Fig. 7. Relationship between the COSMIC parameter N found by multi-parameter optimization and the factor F , this being the ratio between the number of counts observed during the COSMOS probe calibration at a specific site and the calculated neutron flux intensity given by MCNPX when run with the soil chemistry and water content (including lattice water) observed at each probe site during calibration. soil moisture calculated from TDT point sensors does not sample the near-surface soil moisture above 10 cm depth and, as a result, does not recognize the faster rate of drying of surface soil moisture. Consequently, when the area-average profile measured by the TDT probes is used in the COSMIC model to calculate the COSMOS probe count, the estimated COSMOS count is underestimated.
As previously stated, the primary purpose of the COSMIC model is to facilitate use of observed COSMOS probe counts into LSMs through ensemble data assimilation methods. We foresee two broad data assimilation applications using COS-MIC, specifically to provide i. the best estimate of the rate of change in the areaaverage soil moisture profile when this is being calculated by a prescribed (but perhaps imperfect, e.g., biased) LSM, to obtain improvement in the calculated moisture loss from the surface to the atmosphere, in a Numerical Weather Prediction model for example. Arguably in this application the data assimilation process primarily needs to correct for weaknesses in the highfrequency dynamics of the soil moisture profile calculated by the model rather than its absolute value; and ii. the best estimate of the (albeit LSM-calculated) areaaverage profile of soil moisture at a COSMOS probe site, this as a basis for investigating and building models of the relationship between area-average soil moisture and area-average hydro-ecological behavior at the site for example. In this application the data assimilation process primarily needs to correct for weaknesses in the absolute value of the model-calculated profile.
It is not the purpose of this paper to consider detailed aspects of the assimilation of COSMOS probe counts into LSMs at many sites and to investigate the validity of particular LSMs; these details will form the subject of future papers. Meanwhile we illustrate the fact that the COSMIC model can be used in the two applications described above by providing an overview of studies in which COSMOS probe data was assimilated into the Noah Land Surface Model (see Koren et al., 1999;Chen and Dudhia, 2001;Ek et al., 2003) at the Santa Rita Range field site (see Kurc and Benton, 2010;and Cavanaugh et al., 2011) for a period during the North American Monsoon when there were rainstorms that generated rapid changes in soil moisture. Ancillary near-surface hourly measurements of meteorological variables available at this site were used to provide the Noah forcing. Noah represents soil moisture in four layers (0.0-0.1 m; 0.1-0.4 m, 0.4-1.0 m, and 1.0-2.0 m) by calculating the input of water at the surface and the movement of water between layers and loss by transpiration from the upper three layers. The data assimilation used only the COSMOS data (i.e., hourly neutron counts) to update the values of soil moisture in each layer. The observational uncertainty in the COSMOS counts is well defined by Poisson statistics and equal to the square root of the sensor hourly count (Zreda et al., 2008), but, given the typical number of counts from an individual COS-MOS probe, this Poisson distribution of the errors can be adequately approximated by a Gaussian distribution. In each of the example cases discussed below the data assimilation is carried out within the National Center for Atmospheric Research (NCAR) Data Assimilation Research Testbed (DART) framework (Anderson et al., 2009), this being a community facility for ensemble data assimilation. The Bayesian framework employed in DART combines the probability distribution of the prior ensemble with the observation likelihood (data distribution) to compute an updated ensemble estimate (posterior distribution) and increments to the prior ensemble. Increments for each component of the prior state vector are computed by linear regression from the increments calculated in observation space. We use the ensemble adjustment Kalman filter (EAKF) discussed in Anderson (2001) applied hourly. The updated ensemble is obtained by shifting the prior ensemble to have the same mean as the continuous posterior distribution, and the posterior ensemble standard deviation is kept the same as the continuous posterior by linearly contracting the ensemble members around the mean. In this application we used 40 ensemble members with both the meteorological forcing and soil moisture initial conditions perturbed following standard procedures as described in the literature (see Table 2). The soil moisture initial conditions are perturbed around a reference value determined by the COSMOS sensor with an initial assumed uniform profile (the conversion from neutron counts to integrated soil moisture is achieved by applying Eq. A1 in Desilets et al., 2010). Sequential data assimilation was applied via the EAKF to neutron counts, and the soil moisture state variables in Noah updated appropriately every time a new (hourly) observation was available. Draper et al. (2011) state that when applying data assimilation methods, a primary goal is to address the cause of bias between the data and model rather than to rely on data assimilation to correct it, while Yilmaz and Crow (2013) also emphasize that biases should be removed prior to assimilating data. There are several ways to remove such bias (through a priori scaling approaches or through a bias estimation module, for example); in the context of a paper whose primary purpose is to describe the formulation and calibration of the COSMIC model, we follow Kumar et al. (2012) and choose to demonstrate application of COSMIC when using two radically different alternate approaches for removing relative bias, i.e. first by assuming the bias is solely in the data and "modifying the data to match the model", and second by assuming the bias is in the model and "recalibrating the model to match the data".
In fact there is a large systematic bias between soil moisture calculated by the Noah LSM and the value deduced from COSMOS observations at the Santa Rita field site. This Table 2. List of meteorological forcing variables applied to the Noah model and perturbed during ensemble data assimilation together with the nature of the perturbation applied to them. The perturbation distribution was either log-normal (i.e., multiplying the reference variable) or normal (i.e., adding to or subtracting from a reference value). The magnitude of perturbations used in the DART framework is based on a literature review of several studies including Zhou et al. (2006), Zhang et al. (2010), Reichle et al. (2002Reichle et al. ( , 2007Reichle et al. ( , 2008, Walker and Houser (2004), Sabater et al. (2007), Kumar et al. (2012), Dunne and Entekhabi (2005) is clearly apparent in the inset graph in the top panel of Fig. 9, which shows that the cumulative distribution function (CDF) of neutron counts computed by COSMIC using soil moisture profiles from an offline simulation of Noah LSM (NOAH-COSMIC, shown in black) has systematically lower values than those observed by both the COSMOS sensor (COSMOS-real, shown in blue) and the counts computed with the average soil moisture profile from the TDT network (TDT-derived, shown in purple). Although it is clear that in this particular case the source of bias originates from the inability of the model to accurately represent reality, nonetheless we proceed to demonstrate use of COSMIC when used in the "modifying the data to match the model" approach and apply CDF-matching (Reichle and Koster, 2004;Drusch et al., 2005) to scale the COSMOS observations (COSMOSscaled, in green) to match the CDF obtained from Noah LSM offline simulation. Figure 9a shows the time series of the resulting scaled version of the observed neutron count (green) together with the neutron count (calculated by COSMIC) from the soil moisture profiles simulated by the Noah model when running open loop (black) and with the assimilation of COSMOS data (red). Similarly, Fig. 9b shows the depthaverage soil moisture for the Noah model when running open loop (black) and with the assimilation of COSMOS data (red), together with the area-average soil moisture measured by the TDT network (purple). To enhance consistency between these three depth averages they are all weighted by the relative contribution to the aboveground fast-neutron flux for each level (calculated by COSMIC).
To demonstrate use of COSMIC when used in the "recalibrating the model to match the data" approach, we next Panel (b) shows the depthaverage soil moisture for the Noah model when running open loop (black) and with the assimilation of COSMOS data (red) together with the area-average soil moisture measured by the TDT network (purple). To enhance consistency between the three depth averages in (b) they are all weighted by the relative contribution to the aboveground fast-neutron flux for each level (calculated by COSMIC). sought to eliminate the systematic bias by improving the performance of Noah LSM via a priori parameter calibration. When doing this we again employed the AMALGAM method (see Sect. 3) with n = 100, k = 4, and s = 200 to constrain 10 parameters used in Noah (and each individual layer soil moisture initial condition) which were selected based on a preliminary sensitivity analysis. We found that the values of all ten parameters were changed by calibration to some extent, but four model parameters changed significantly, namely FXEXP, REFKDT, SMCREF, and DKSAT, which control bare-soil evaporation, surface infiltration, onset of transpiration stress due to soil water content, and soil hydraulic conductivity, respectively. This multi-objective optimization was performed on the individual components of the mean squared error (Gupta et al., 2009;Rosolem et al., 2012) between observed neutron counts and neutron counts computed via COSMIC from model-derived soil moisture profiles. The recalibrated version of the Noah model was then used in an experiment in which the observed (unscaled) neutron counts were assimilated. Figure 10a shows the time series of the observed neutron count (green) together with the neutron count (calculated by COSMIC) from the soil (black) and with the assimilation of COSMOS data (red), together with the area-average soil moisture measured by the TDT network (purple). To enhance consistency between these three depth averages they are all weighted by the relative contribution to the above ground fast neutron flux for each level (calculated by COSMIC). (black) and with the assimilation of COSMOS data (red), together with the area-average soil moisture measured by the TDT network (purple). To enhance consistency between these three depth averages they are all weighted by the relative contribution to the aboveground fast-neutron flux for each level (calculated by COSMIC). moisture profiles simulated by the recalibrated Noah model when running open loop (black) and with the assimilation of COSMOS data (red). Figure 10b shows the depth-average soil moisture for the Noah model when running open loop (black) and with the assimilation of COSMOS data (red), together with the area-average soil moisture measured by the TDT network (purple). Again, to enhance consistency between these three depth averages they are all weighted by the relative contribution to the aboveground fast-neutron flux for each level (calculated by COSMIC).
In both of the very different data assimilation demonstrations just described, the COSMIC model provided an effective mechanism for translating model-calculated soil moisture profiles into aboveground fast-neutron count when applied using EAKF-based assimilation within the DART framework. The resulting improvements in model performance are illustrated in Figs. 9 and 10 and documented in Table 3. Arguably the two different approaches for removing bias between data and model just demonstrated (i.e., "modifying the data to match the model" and "recalibrating the model to match the data") might respectively be considered appropriate for use in the COSMOS probe data assimilation applications (i) and (ii) (see above). The results in Table 3 clearly demonstrate that there is an improvement in the statistical metrics when neutron counts are assimilated relative to the open-loop case either when the observations are scaled or the Noah model calibrated. However, Table 3 also suggests that the calibration of Noah successfully removed most Table 3 when other metrics are analyzed and when the integrated soil moisture is compared with the average soil moisture measured by the TDT network.
Summary and conclusions
This study showed that COSMIC, a simple, physically based analytic model, can substitute for the time-consuming MC-NPX model in data assimilation applications, and that COS-MIC can be calibrated by multi-parameter optimization at 42 COSMOS sites to provide calculated neutron fluxes which are within a few percent of those given by the MCNPX model. The parameters α and L 3 are correlated with ρ s , the bulk density for the soil at each site, and consequently are mutually correlated. This correlation with ρ s might provide an approximate estimate of their value if parameter optimization against MCNPX model is not feasible. The value of N , the third optimized parameter in COSMIC, is very strongly related to F , i.e., to the ratio between the number of counts observed during COSMOS probe calibration at a specific site and the calculated neutron fluence given by MCNPX when run with the soil chemistry and water content (including lattice water) observed at each probe site during calibration. The origin of this real site-to-site variability in F across the COSMOS sensor array, which is presumably mainly associated with site-to-site differences in soil chemistry or more likely vegetation cover, is currently under investigation.
It was demonstrated at the Santa Rita Experimental Range field site that the aboveground neutron count rates calculated by COSMIC from an area-average soil moisture profile independently measured using TDT sensors agreed well with observed neutron count rates measured by the COSMOS probe at this site. It was further demonstrated that when the calibrated COSMIC model was applied at this site, it provided an effective mechanism for translating model-calculated soil moisture profiles into aboveground fast-neutron count when used within a data assimilation framework to assimilate COSMOS probe counts into the Noah model with two radically different approaches used to remove the bias between observations and model. The COSMIC model is freely available for download at the COSMOS website (http://cosmos. hwr.arizona.edu). | 9,240 | 2013-08-14T00:00:00.000 | [
"Environmental Science",
"Physics"
] |
IMPROVEMENT OF NAIVE BAYES ALGORITHM IN SENTIMENT ANALYSIS OF SHOPEE APPLICATION REVIEWS ON GOOGLE PLAY STORE
Reviews of the shopee application on the google play store are included in useful information if processed properly. Old or new users can analyze app reviews to get information that can be used to evaluate services. The activity of analyzing application reviews is not enough just to see the number of stars, it is necessary to see the entire contents of the review comments to be able to know the intent of the review. A sentiment analysis system is a system used to automatically analyze a review to obtain information including sentiment information that is part of an online review. The data is classified using Naive Bayes. A total of 1,000 shopee app user reviews were collected to form the sample dataset. The purpose of this study is to determine the sentiment analysis of shopee application reviews in the Google Play Store using the Naive Bayes algorithm. The stages of this research include, data collection, labeling, pre-processing, sentiment classification, and evaluation. In the pre-processing stage there are 6 stages, namely Cleaning, Case folding, Word Normalizer, Tokenizing, Stopword Removal and Stemming. TF-IDF (Term Frequency - Inverse Document Frequency) method is used for word weighting. The data will be grouped into two categories, namely negative and positive. The data will then be evaluated using accuracy parameter testing. The test results show an accuracy value of 81%, this result shows that shopee application reviews tend to be negative.
INTRODUCTION
The rapid development of E-commerce in Indonesia has given rise to many marketplaces that offer various products and services to the public.This phenomenon not only reflects the rapid growth of technology, but also creates various opportunities for consumers or businesses.With the increasing number of marketplace options, consumers can easily search and compare products and prices [1].It also creates healthy competition among marketplaces, encouraging innovation in terms of user experience, services, and promotions.
Currently, the Google Play Store offers a plethora of marketplace applications.Google play store is one of the digital content services owned by Google which contains digital products such as applications, music or songs, books, games and cloud-based media players [2].In the play store there are many features such as rating and review features where users of products from the play store can provide their opinions on the products they have used.
One of the e-commerce applications found on the play store is the shopee application launched in 2015, based on data from (SimilarWeb, May 2023) shopee users in Indonesia are around 161 million people.Because everyone has different opinions and thoughts, there are many opinions that are pro and contra to this application.User reviews are very diverse ranging from good and bad talk, user criticism and there are also those who provide suggestions regarding features.Examining reviews can be simplified by observing the star ratings provided by users.However, these star ratings alone do not provide a comprehensive understanding of the complete review content [3].Review analysis can be done manually by looking at reviews one by one, but if there are too many reviews then we can't do it manually there is a faster and more efficient way to use automated systems such as sentiment analysis [4].
Sentiment analysis is the process of identifying, extracting, and evaluating sentiments or opinions expressed in text or data, particularly in the context of reviews or opinions.The goal of sentiment analysis is to understand people's viewpoints, feelings, or attitudes towards a specific topic or entity.Sentiment analysis can be done manually by humans or automatically using computational algorithms and techniques, such as natural language processing and machine learning.In practical applications, sentiment analysis is often used in customer surveys, social media monitoring, product review analysis, and public feedback analysis to gain insights into perceptions and opinions.This research was conducted because there are so many reviews given by shopee application users but these reviews are not utilized as well as possible.Even the review is only limited to a review without any action, this causes the shopee application rating to drop on PlayStore.
In this research, we will conduct sentiment analysis on reviews of the Shopee application accessible on the Google Play Store.The goal is to evaluate Shopee app users by utilizing the Naive Bayes technique with TF-IDF weighting to classify them into negative and positive categories.Furthermore, this study seeks to assess the precision of the Naive Bayes algorithm in performing sentiment analysis of Shopee application reviews.
Furthermore, as for previous research related to this research, it will be used as the author's research material and used as a reference.In the following, research that discusses sentiment analysis in reviews.
First research was carried out by arif rahman, Ema Utami, Sudarmawan come Amikom University Yogyakarta in 2021 entitled "Analysis of Sentiment on Google Play Store Apps utilizing Naive Bayes Algorithm and Genetic Algorithm" aims for determining level accuracy between the two algorithms then for the dataset used comes from several applications such as shopee, ruang guru, pedia Shopee and gojek.In this study the pedia store dataset have the highest accuracy rate of 96.87 percent, then for Shopee obtained an overall accuracy 96.53 percent, Ruangguru obtained the 95.54%, and Gojek obtained an accuracy of 96.54%.The research collected datasets using the web scrapper application from Google Chrome then the labeling was done manually in the Ms Excel application [5].
Second research was conducted by dedi darwis, nery siskawati, zaenal abidin from teknokrat "In 2021, the University of Indonesia conducted research titled 'Utilization of the Naive Bayes Algorithm for Analyzing Sentiments in National BMKG Twitter Data".Python 3.74 was employed as the primary programming tool, and a dataset comprising 1179 tweets was used for testing.The data was divided into two categories: training data and testing data, with a ratio of 70:825.It's worth noting that the dataset was imbalanced, with more positive reviews than negative ones.Notably, the author opted not to perform feature normalization, resulting in relatively low accuracy test results when applying the Naive Bayes method for classification namely 69.97% [6] .
Third research was conducted by styawati, nirwana hendrastuty, auliya Rahman isnain, ari yanti rahmadhani from teknokrat university of Indonesia in 2021 entitled "Analysis of Public Sentiment Towards the Pre-Employment Card Program on Twitter with the Support Vector Machine Method".Aims to analyze public opinion with data obtained on twitter social media using Support Vector Machine (SVM).Then to measure the performance of SVM classification, the Confusion Matrix method is used.
In this study, two kernels were compared, namely linear and RBF kernels.The evaluation results show that the precision of the linear kernel is 98.67%, the precision is 98%, the recall is 99%, and the F1 score reaches 98%.Meanwhile, the precision of the RBF kernel is 97%, the recall is 98.67%, and the F1 score is also 98%.The accuracy of the RBF kernel reached 98.34%.From these results, it can be concluded that the public opinion of Twitter users towards the preemployment card program during the pandemic tends to be neutral, with a neutrality rate of 98.34%.When viewed from the accuracy aspect, the linear kernel is more accurate than the RBF kernel with an accuracy of 98.67% and 98.34% respectively [7].
Fourth studies was carried out with the aid of Billy, Helen, Enda Esyudha retrieved tanjungpura university in 2018, discussing Indonesian online product opinions to achieve records such as sentiment records that's a part of the web evaluation.The information is run through a classification process the use of the Naive Bayes approach.The sentiment analysis gadget is going via 5 levels, namely records series, pre-processing, phrase weighting, model building, and sentiments class.within the phrase weighting degree, the TF-IDF (term Frequency-Inverse document Frequency) approach usage.Facts is assessed into five lessons: strongly bad, terrible, neutral, tremendous, and strongly advantageous.moreover, the statistics might be evaluated the usage of a confusion matrix by using measuring accuracy, bear in mind, and precision.The take a look at outcomes show that in the 3-magnificence take a look at (bad, neutral, and superb) the first-rate performance is completed with ninety% education facts and 10% test data, ensuing in an accuracy of 78%, bear in mind of 93.33%, and precision of 77%.inside the 5-class take a look at, the high-quality performance turned into also finished with 90% schooling statistics and 10% take a look at information, with accuracy reaching 59.33%, remember 58.33%, and precision 59.33% [8].
The fifth study conducted by Apriani & Gustian in 2019 aims to evaluate the performance of accuracy, class recall, and AUC of sentiment analysis using the Tokopedia application by applying the Naïve Bayes algorithm.The results of this study show that the precision value for class 1 is 1.On the other hand, in class recall, the result is 95.49% (positive class: negative class) and the AUC value is 0.980.From the analyzed reviews, it can be seen that comments with negative sentiments have a percentage of 63.53%, while comments with positive sentiments are only about 36.37% [9].
The cryptocurrency.They proposed a CGWELSTM (Convolutional Gated Word Embedding Long Short-Term Memory) architecture that combines a convolutional layer, bidirectional layer, group enhancement mechanism, attention mechanism, and fully connected layer.The accuracy results of this study show that the CGWELSTM architecture achieves the highest level of accuracy in cryptocurrency-related sentiment analysis.In the dataset used, CGWELSTM achieved a classification accuracy of 85.4001% [10].
Based on this background, author does not conduct the same research as others.This research uses review data from the Shopee application on the Google Play Store, then the method used is the naive bayes algorithm.Later the review data that has been collected will be grouped into negative and positive reviews / comments before a process called sentiment analysis is carried out.
RESEARCH METHODS
Then for the research strategy that will be carried out, explain in Figure 1.The following is an explanation of figure 1 concept study.
Data Selection
In this preliminary stage, the writer conducts information collection and statistics labeling.records taken from the shopee utility on the google play shop using the scraping technique by using using python programming.
Pre-Processing
A pre-processing step in sentiment analysis is the initial stage of a sentiment analysis process.In processing unstructured input data into structured data before the main process such as classification or sentiment analysis is carried out [11].At this stage, several stages will be carried out including the following: 1. Cleaning is the method of cleansing emoticons and emblems executed in this observe.Emoticons and emblems in the evaluation are unnoticed due to the fact These studies most simply make a specialization of the text embodied inside the evaluation.The characters that were ignored included "~", " `", " !", " $", " %", " ^", " &", " *", " (", " )", " _", "-", "+", "=", ":", "'", "comma", "duration", "?".reviews included characters similar to " tolong min, untuk jasa kirim diperbaiki " after processing grow to be " tolong min untuk jasa kirim diperbaiki ". 2. Case folding is a degree that involves changing the text in the file to a standardized shape, that's lowercase.for instance, the comment "Tolong perbaiki fitur pembayaran" will become "tolong perbaiki fitur pembayaran ", the uppercase "T" is converted to lowercase "t".
Word Normalizer is a technique used to rectify
words in reviews in order to generate sentences that adhere to proper and accurate Indonesian grammar.This enhancement is essential to facilitate comprehension for readers regarding the sentence's intended meaning.As an illustration, if there is a comment "knpa lama proses update", after going through the Word Normalization process, it will become "kenapa lama proses update".The change of the word "knpa" to "kenapa" is done to make the message easier to understand. 4. Stopwoard removal is a level of putting off phrases based at the phrases contained inside the conjunction list.phrases inclusive of "di", "dan", "yang", might be removed.5. Tokenizing is a technique used to break the text into phrases, deliberating punctuation and space obstacles.as an instance, the sentence "shopee sekarang lemot" is then transformed into "shopee", "sekarang", "lemot".6. Stemming is the manner of reducing phrases to their simple form or "root" phrase.as an example, the word "batalkan" might be changed to "batal".Stemming is utilized in textual content processing and textual content analysis to help reduce phrase versions and improve consistency in subsequentanalysis or processing.
Text Transformation
This stage of text transformation or attribute formation refers to the process of obtaining the required representation.At this stage the author performs feature extraction with TF-IDF.Here is the formula of TF -IDF.
Data Mining
Data mining is the step-by-step process of extracting relevant, hidden, and potentially valuable information from large and complex data sets [12].The main purpose of data mining is to uncover new knowledge and information that can provide added value or competitive advantage in a domain or industry.This research uses one of the existing algorithms in data mining, namely naive baiyes.
Naive baiyes is a working method for the most popular classifiers with a good average accurization rate.Naive Baiyes classification technique is based on simple probabilities and is designed to be used with the assumption of independence between explanatory variables.This algorithm emphasizes probability learning, here is the formula of naive baiyes [13].
This formula explains that X is evidence, H the hypothesis, then P(H | X) is part of the probability that hypothesis H is true if accompanied by evidence X, or vice versa, P(H | X) enters the probability of hypothesis H in the presence of condition X. P(X | H) can be interpreted as the probability of evidence X in the attendance of hypothesis H, then P(H) is the initial ability of hypothesis H, while P(X) is the initial probability of evidence X..
Data Selection
1. Data Scraping refers to the process of extracting information or data from a source or website.This is done by using algorithms or specialized software that automatically extracts data from web pages and stores it in a format that can be further used and analyzed.Data scraping methods can be done manually, where a person manually copies and pastes information from a website into a file or database, or automatically using specialized software or scripts designed to automatically retrieve data from web pages [14], In this research, scraping shopee application review data is done automatically using the API retrieved from Google Play Store.The results of this data collection are 1000 data, then the data that has been taken is stored in a csv format file.2. Data labeling, also known as data tagging, involves assigning a specific category or label to each data point or sample in a data set.The purpose of data labeling is to give context or meaning to the data so that it can be used to train or test machine learning models [15].In this research, we divided them into negative and positive labels and the subsequent labelin process is done automatically using the python programming language.
Pre-Processing
Derived from the completed labeling process, the subsequent section presents the outcomes of the pre-processing phase.
Text Transformation
The text transformation or attribute generation stage is a step in the process of obtaining the required document representation [16].In this stage, the author performs feature extraction using the TF-IDF method, which is a technique for generating attributes from documents.This process aims to transform the document text into a numerical representation that can be used for further analysis.
Data Mining
After all steps completed, calculations will be made to see accuracy rate using naive baiyes algorithmic approach.The accuracy rate when performing sentiment analysis using this method is quite good with an accuracy value of 81%, then there are also results of precision 80%, recall 97%, f1-score 88% of negative comment data and precision 86%, recall 45%, f1-score 60% of positive comment data.
Frequency of Word Appearance
Frequently occurring words in user reviews can provide information about the app.This time the author will provide an overview of user review based on positive and negative reviews.Visible in Figure 8 and Figure 9.
DISCUSSION
Of the findings obtained from the research that has been carried out, namely sentiment analysis using the naive baiyes algorithm, several things can be used as evaluation material, among others, In this study, the Naive Baiyes algorithm proved to be an accurate algorithm because it produced an accuracy value of 81%.The advantage of this research is that it has good accuracy, precision and recall values.Then the calculation of precision, recall and f1-score is also carried out based on positive and negative reviews.Then there is previous research conducting sentiment analysis with several algorithms where the result is the naive baiyes algorithm which has the highest algorithm until the accuracy rate reaches 65,26 % [17].Therefore, it can be concluded that naive baiyes is accurate enough to be used.The drawback in this study is that it uses a fairly small dataset so that the results are less satisfying, for further researchers can add the number of dataset.
CONCLUSION
The research findings indicate that an accuracy rate of 81% was attained.Additionally, there were precision results of 80% and recall of 97%, resulting in an f1-score of 88% for negative comments.For positive comments, the precision was 86%, recall was 45 percent, and f1-score was 60%.The survey employed The TF-IDF combined with the Naive Bayes algorithm weighting technique to categorize user reviews into either negative or positive sentiments.Furthermore, the sentiment analysis of Shopee app users reveals a prevailing negative sentiment towards Shopee's services.This suggests that there are areas where Shopee could improve, particularly in terms of features and other aspects.Based on the analysis, it can be concluded that the data sourced from the Google Play Store platform for the Shopee application indicates a higher frequency of negative reviews compared to positive ones.The negative reviews often revolve around issues with the application's features, shipping process, and frequent slow updates.Conversely, positive feedback frequently highlights various promotions offered by Shopee, such as free shipping, product discounts, and cashback.
Figure 2 .
Figure 2. Script for scraping data
Figure 4 .
Figure 4. Results of labeling negative and positive reviews.
Figure 7 .
Figure 7. Calculation results with the naive baiyes method. | 4,180.4 | 2023-12-23T00:00:00.000 | [
"Computer Science"
] |
Multi-casting approach for vascular networks in cellularized hydrogels
Vascularization is essential for living tissue and remains a major challenge in the field of tissue engineering. A lack of a perfusable channel network within a large and densely populated tissue engineered construct leads to necrotic core formation, preventing fabrication of functional tissues and organs. We report a new method for producing a hierarchical, three-dimensional (3D) and perfusable vasculature in a large, cellularized fibrin hydrogel. Bifurcating channels, varying in size from 1 mm to 200–250 µm, are formed using a novel process in which we convert a 3D printed thermoplastic material into a gelatin network template, by way of an intermediate alginate hydrogel. This enables a CAD-based model design, which is highly customizable, reproducible, and which can yield highly complex architectures, to be made into a removable material, which can be used in cellular environments. Our approach yields constructs with a uniform and high density of cells in the bulk, made from bioactive collagen and fibrin hydrogels. Using standard cell staining and immuno-histochemistry techniques, we showed good cell seeding and the presence of tight junctions between channel endothelial cells, and high cell viability and cell spreading in the bulk hydrogel.
Introduction
A perfusable vascular network is necessary to support the mass transport requirements of a metabolically active and highly populated tissue. The lack of a three-dimensional (3D) system has held back progress into fabrication of complex tissues and organs, as diffusion from the surface of a tissue construct becomes unfeasible for rapid delivery of oxygen and nutrients [1]. A functioning blood vessel network, which is 3D, hierarchical and perfusable, is thus a necessary requirement of most functional tissues; the use of channel hierarchy enabling both extensive coverage of the tissue while simultaneously providing the low pressure heads required for efficient diffusion in and out of the bulk. Porous materials, while improving the permeability of tissue engineered constructs, are still limited in the capacity for this exchange in thick samples especially in scaffolds containing biologically relevant cell densities [2]. Additionally, blood vessel infiltration by the host's vasculature tends to be a slow process, meaning limited functionality of implanted scaffolds [3]. Thus tissue engineering, which seeks to fabricate new tissues and organs, requires approaches for the formation of 3D vascular networks in materials suitable for supporting tissue development, of which hydrogels are the most important group.
Current vascularization methods are limited in their capacity for supporting a large volume and achieving full three-dimensionality; a variety of techniques have been suggested for forming vascular systems, each with specific limitations. Layer-by-layer approaches, such as those based on lithography, have been reported [4][5][6][7][8][9][10][11] but are limited by the lack of three-dimensionality, which is achievable from a single plane of channels. However, such approaches do show that a perfusable vascular system greatly improves the functionality of a tissue construct and can display characteristics of an active vascular system, such as angiogenic sprouting from the channels into a bulk hydrogel [4]. Gelatin and alginate have previously been used as sacrificial materials for the formation of planar vascular networks, in which polydimethylsiloxane (PDMS) moulded casts of hydrogel templates yielded perfusable networks at very fine length-scales [9,12].
A range of direct 3D printing-based approaches have created vascular networks in three-dimensions [13]. However, they often require the casting material to be chemically or photo-cross-linked, requiring cells to be seeded after casting which leads to a lack of cell uniformity in the bulk of the tissue construct, or they use materials that are synthetic in origin and thus often have properties that limit cellular activities [14 -16]. One promising sacrificial 3D printing technique uses 3D filament networks of carbohydrate glass, allowing casting in ideal tissue engineering materials, and which generates open channels upon immersion in water [17]. However, there are issues with the cytotoxicity from the dissolution of carbohydrate glass templates (osmotic damage to cells) and to prevent early degradation, the template requires a synthetic coating (PLGA) [3]. Another promising technique uses printing of Pluronic F127 as the sacrificial material [18][19][20][21], which can be printed alongside cell-loaded hydrogel materials.
The direct printing of these sacrificial materials limits the complexity and structure of the vascular network that is producible, due to certain properties of the materials (e.g. hydrophilicity, gelling speed, removal method). For instance, in the case of agarose [3] or carbohydrate glass printing [17], a sequence of one-dimensional (1D) filaments is extruded out of the printhead which touch at various locations to produce network junctions. In the case of the carbohydrate glass, the thickness of the channel is determined by the speed of the printhead, limiting the structure of the finer features that can be fabricated [17].
Other work has used thermoplastic 3D printed models to produce internal channels in porous scaffolds, such as with freeze-dried collagen [22,23] and PCL (collagen coating on the channels) [24]. However, they require freezing and harsh solvents as a means of formation of the scaffold and removal of the channel template material. Thus, they must remain acellular until after the process is completed and both approaches require the collagen to be chemically cross-linked for stability. Finally, 3D printed materials have recently been cast in PDMS and removed to reveal a 1D helical channel (500 mm in diameter) [25]. The use of PDMS as the bulk material limits the use of cells to the channels only.
We have developed a novel process in which standard 3D printed thermoplastic models, which are highly customizable and reproducible, are converted into a gelatin hydrogel material. This can then be used as a vascular network template in highly bioactive hydrogel materials, which are pre-loaded with cells. The ability to use 3D printed materials in this way is achieved by casting an alginate intermediary, which is ionically cross-linked with calcium, around the 3D printed model. Alginate hydrogels are stable at high temperatures (approx. 1008C), and do not decay in acetone, both of which are required to remove the thermoplastic ( polyesterbased) model in its entirety. By making use of the different physical and chemical properties of the 3D printed material, and of alginate and gelatin, we are able to selectively add and remove positive and negative casts of the original network design, resulting in a channelled construct made from thermo-gelled collagen or enzymatically cross-linked fibrin hydrogel. The use of gelatin gel as a vascular template has the added advantage of enabling prompt perfusion once the final construct has formed.
Recently, a 3D printed material has been converted into an alginate template material and used in agarose and genipin-cross-linked gelatin constructs, with trifurcating channel features down to 400 mm, supporting liver cancer cells [26]. In our study, via a different set of materials, we produced a bifurcating template, which was hierarchical in nature with features down to 200 mm. The template material was removable within 20 min of casting, meaning minimal delay prior to perfusion, which can be performed at physiologically comparable rates. A flow rate of approximately 2 ml h -1 through the 200 mm channels can be achieved by applying 32 ml h -1 at the 1 mm inlet [17]. The implantable, bioactive hydrogel materials collagen [27], as the major component of extracellular matrix, and fibrin [28], as the major protein precursor for wound healing, were used as our final constructs. These materials supported primary cells; here, we used endothelial cells in the channels and connective tissue cells, which showed spreading in the hydrogel constructs.
Cell culture
Human umbilical vein endothelial cells (HUVEC, Public Health England) and human dermal adult fibroblasts (HDFs, Public Health England) were used for the experiments. The cells were trypsinized using TrypLE Express (Life Technologies), centrifuged at 250 g for 5 min, and counted using a Scepter Cell Counter (Millipore). HUVECs were cultured in a T75 flask to 80% confluence, using endothelial growth medium (no vascular endothelial growth factor or platelet-derived growth factor) (EGM Plus, Lonza), and HDFs were cultured in a T75 flask to 80% confluence using high-glucose Dulbecco's modified Eagle's medium (Life Technologies) containing 10% fetal bovine serum (FBS, Life Technologies).
HUVECs were imaged using Green Cell Tracker dye (Life Technologies), allowing for immediate imaging and tracking of cell seeding and distribution in the channels. Lyophilized fluorescent dye was resuspended in dimethyl sulfoxide (DMSO, Sigma) at 10 mM and stored at 2208C until use. A working solution was prepared in DMEM at 10 mM. Plated cells were washed with phosphate-buffered saline (PBS, Life Technologies) and the working solution added. Flasks were incubated at 378C for 45 min prior to trypsinization.
Three-dimensionally printed design and printing
The CAD model vasculature was designed using Autodesk Inventor 2014 (figure 1a); the structure consists of a 1 mm inlet which symmetrically bifurcates into a series of 200 -250 mm channels, which then converge into a single outlet of 1 mm. Thermoplastic models in support wax were a gift from Solidscape UK. Models were fabricated using 12 mm layer thickness and 10 base support layers, on a 3Z Studio printer. Following printing, models in support wax were removed from the ceramic base (8 mm) by placement on a hot plate at 1208C for 7 min. Models in support wax could then be gently removed using stiff card. To remove the support wax from the models (figure 1b), they were put in a bath of selective solvent (BioAct VSO, Solidscape) and placed in an oven at 558C for 4 h. The model was subsequently stored in fresh solvent at room temperature, until use (figure 1c).
Silicone chamber design and model mounting
The vascularized hydrogel was produced in a custom-made silicone chamber. For the base, a 1 mm silicone sheet (Silex) was adhered to a glass imaging slide (Fisher Scientific) and a section of 12 mm internal diameter silicone tubing (VWR) attached to this sheet, both using silicone sealant (Dow Corning). Two holes were drilled on opposite sides of the chamber and flexible silicone tubing with 1 mm internal diameter (VWR) fitted. A silicone lid was made using a CNC milling machine from a 3 mm silicone sheet (Silex), into which two 2 mm filling holes were drilled (chamber design shown in figure 1d ). The models were briefly dried on tissue paper and then very carefully placed between the filled silicone inlets using tweezers. The inlets were rotated so as to angle the model such that all channel features are visible under an optical microscope at the same time (figure 1e). The silicone inlets were then super-glued in place.
Alginate preparation
Calcium alginate gels were prepared using an 'internal setting' method. This involved adding 300 mM CaHPO 4 (Sigma) and 600 mM gluconolactone (Sigma) to 7.5% (w/v) sodium alginate (Sigma). In order to prevent the trapping of bubbles between the fine features of the 3D printed template, the chamber was first filled with DI water. The viscous alginate precursor solution was then slowly injected into the silicone chamber, displacing the water and fully encapsulating the thermoplastic model. The gluconolactone functions as an acidifier and causes the slow release of calcium, enabling a controlled gelling mechanism and a uniformly cross-linked hydrogel. The alginate gel was left overnight at 48C for maturation (figure 1f ).
Model removal and gelatin template preparation
The model material was removed by placing the chamber in a bath of near-boiling DI water. After 5 min, the liquid model material was removed from the channels by gentle suction with a 1 ml syringe. Following 10 min on ice, the channel structure was cleaned by injecting 2 ml of acetone into the channels to purge any remaining model material. Subsequently, 6 ml of 250 mM CaCl 2 (Sigma) was injected around the network which cross-links the surface of the alginate channels more thoroughly, without causing significant syneresis.
Liquid gelatin (Sigma, porcine skin, low bloom, 15% (w/v), pH 7.4) was infiltrated into the channel structure via suction. The chamber lid was removed and the chamber placed in a bath containing 200 mM sodium citrate (Sigma) and 200 mM glycine (Sigma) to chelate the calcium cross-linking ions, and was then left for 24 h at 48C, at which point the alginate was fully dissolved into solution. Once the alginate gel had dissolved, the chamber was placed in a bath of sterile, calcium free PBS to remove any residual alginate in the chamber, and a lid attached under liquid, which was then sealed using super-glue. The chambers were finally purged of any remaining citrate or alginate by flowing 5 ml of PBS through the lid holes (figure 1g).
Extracellular matrix hydrogel preparation
Collagen hydrogels were used at 7.5 mg ml -1 final concentration. Soluble collagen solution was prepared using rat-tail tendon. We solubilized the tendon using 0.1% (v/v) acetic acid on a magnetic stirrer (250 ml g -1 tendon) for 60 h at 48C, and centrifuged the solution at 9000g for 90 min. The collagen was subsequently lyophilized, weighed and resuspended at the required stock concentration in 0.1% (v/v) acetic acid. Collagen was gelled using a standard approach [6], employing the addition of 10Â M199 (Sigma), 1 M NaOH, and 0.2% (w/v) NaHCO 3 (Sigma), in order to raise the ionic content and pH of the collagen solution, thereby inducing a gelling response. HDFs were carefully mixed into the neutral precursor solution at 1 Â 10 6 cells ml 21 , just prior to casting. Owing to the high viscosity of the collagen solution, it was injected slowly on ice into the chamber using a syringe pump at 15 ml h -1 , displacing the PBS supporting the gelatin template (figure 1h). Following 20 min at room temperature, the chamber was placed for an additional 20 min in a sterile 378C bath, which rapidly gelled the collagen and melted the gelatin.
Both were stored at 2208C until use. In total, 0.1 M NaOH was added to the fibrinogen solution to raise the pH to 7.4. HDFs, in cell medium, were carefully mixed into the precursor solution at 1 Â 10 6 cells ml 21 and thrombin was added just prior to casting. The solution was subsequently loaded into a 1 ml syringe and injected into the chamber via the lid holes, displacing the PBS supporting the gelatin template. Following 5 min at room temperature, the chamber was placed in a sterile 378C bath for an additional 15 min, which melted the gelatin.
Gelatin removal and channel seeding
Gelatin was removed from the channel network by gentle suction of PBS at 378C through the inlets. Channels were checked for patency using fluorescent 1 mm red tracer beads (Life Technologies). HUVECs were subsequently suspended in 100 ml of endothelial growth medium and injected into the channels at 2 Â 10 7 cells ml 21 . The chamber was left for 5 h-overnight prior to commencing perfusion. Non-adherent HUVECs were flushed from the channels and perfusion continued with endothelial growth medium at 0.1 ml h -1 for 7 days, using a syringe pump. For HDF experiments, high-glucose DMEM containing 10% FBS was perfused at a range of flow rates (1-30 ml h -1 at the 1 mm inlet) for 7 days using a peristaltic pump, and the medium reservoir was replaced every 2 days.
Characterization of microchannel formation
After 7 days of perfusion, the development of tight cell-cell junctions between the channel HUVECs was determined using CD31 immunocytochemistry. Cells were fixed by injecting 4% (w/v) paraformaldehyde (PFA) into the channels and leaving the chamber at 48C for 1 h. Following three 5 min washes with PBS, a permeabilizing solution containing 0.25% (v/v) Triton X-100 (Sigma) was injected into the channels for 10 min at room temperature, followed by a blocking solution of 1% bovine serum albumin (BSA, Sigma) for 30 min at room temperature. The primary antibody, mouse monoclonal anti-CD31 (HEC7, Abcam), was added at 1 : 100 in 1% BSA, injected through the inlet and incubated overnight at 48C. Unbound antibody was flushed out using three 5 min washes of PBS, and a secondary antibody, goat anti-mouse IgG H&L conjugated to AlexaFluor 568 (Abcam), was added at 1 : 200 in 1% BSA, injected into the channels, and incubated for 2 h in the dark at room temperature. The unbound antibody was subsequently flushed out using three 5 min washes of PBS. Fluoroshield with 4 0 ,6-diamidino-2-phenylindole (DAPI, Life Technologies) was subsequently injected as a nuclear counterstain. Cell viability was performed using a Live-Dead kit (Life Technologies). Fibrin gels containing 1 Â 10 6 cells ml 21 were cross-sectioned through the centre of the construct, and the surface stained with 1 ml ml -1 calcein AM and 4 ml ml -1 ethidium homodimer-1. After 20 min, slices were washed in PBS and imaged.
Imaging
Imaging was performed using phase contrast and epi-fluorescent microscopy (Zeiss Observer.Z1 with ORCA-Flash4.0). Images were processed using Zen software (Zeiss) and ImageJ. Other images were taken using an overhead microscope (Olympus SZX16 with PixeLINK camera and software), and also with an optical camera (Canon Digital IXUS70).
Statistical analysis
Statistical analysis of live-dead data was performed using IBM SPSS Statistics 23. Cells were manually counted using ImageJ by taking representative sample areas of 1 mm 2 (at least four were analysed per flow rate). A Levene's test showed variances were not homogeneous and so a comparison of values was carried out using a Kruskal -Wallis and a Games -Howell post hoc test. Differences were considered statistically significant for p , 0.05. Data are presented as the mean + s.d.
Results and discussion
The channel network produced for this study has a hierarchical structure, with channels bifurcating symmetrically from 1 mm to 200-250 mm, producing a single inlet and outlet, and 16 channels to support a large 3D volume. 3D printed models are prepared using a commercially available printer, mounted into a custom-made silicone chamber, which is filled with alginate gel. Following evacuation of the 3D printed model material from the alginate cast, gelatin is subsequently infiltrated into the channels and the alginate is removed by chelation of the cross-linking calcium ions. This template is then cast in a collagen or fibrin gel, which can be pre-loaded with cells, and the gelatin liquefied at 378C. The use of standard CAD and commercial 3D printing to generate the first stages of the process conveys a high degree of precision and reproducibility to the gelatin ( figure 2a,b), and thus the vascular channel structure. In particular, Solidscape inkjet technology enables very precise features to be fabricated at significant volumes, and the use of support wax enables a wide range of designs, including overhanging features. Additionally, there is some choice in the shape of the channel cross-section and the morphology of the channel junctions. Thus, our approach enables gelatin templates to be fabricated with a highly complex and fine network architecture, which is fully scalable.
The CAD design in this study was influenced by a number of considerations, including a correspondence to physiological channel structures and fluidic conditions, but also to an ease of handling and imaging. Vascular systems are constrained physiologically by a number of relationships, most notably Murray's law relating parent and daughter branch diameters, usually quoted as the sum of cubic terms which, among other results, yields a constant shear force around the whole network [29,30]. In this study, we have designed a network that uses the sum of squared terms. If we were to follow the sum of cubic terms, a higher number of bifurcations would be needed and the same volume would have a much higher density of channels. Mounting of the 3D printed models requires an inlet size that is not difficult to handle; hence, we used inlets with a diameter of 1 mm. For ease of imaging, a template was desired with fewer channels such that there is minimal overlap between them. The chamber is constrained in this study to a volume of 1 ml in order to maximize cell concentration, though we postulate that much larger volumes are possible as this is a casting-based technique, and thus a closer proximity to Murray's law is possible in the future.
The angles between parent and daughter branches are chosen as 708, to minimize turbulent flow at junctions while also maximizing the volume supported by the finest channels of the network. The finest channels are 2 mm in length and are separated by 1.5 mm in plane and 1 mm between planes. Thus, these thoroughfare channels enable diffusion of nutrients and oxygen into a significant proportion of the bulk hydrogel.
The choice of the gelatin and alginate used is essential for this process to be successful. A low bloom (80 g) gelatin is used that can withstand alginate removal and subsequent rsif.royalsocietypublishing.org J. R. Soc. Interface 13: 20160768 casting with collagen or fibrin, but has a viscosity which is low enough to be easily removed from the final hydrogel. It is type A gelatin, which has a better gel strength to viscosity ratio than type B [31]. A concentration of 15% (w/v) is sufficient for the chelation and casting process, and has a melting point such that a low viscosity liquid is produced upon melting at 378C. However, the gelatin template still requires mechanical support from the surrounding liquid in order to preserve its structure, prior to casting.
To enable the alginate to be cast and removed effectively, a low viscosity sodium alginate is used at a concentration of 7.5% (w/v), which provides sufficient rigidity for the fabrication process; in particular resisting syneresis from high temperatures, solvent exchange and CaCl 2 addition. The alginate gels are fabricated using an internal setting process using CaHPO 4 and gluconolactone as an acidifier, which slowly releases the calcium producing a uniform alginate hydrogel. Overloading the alginate gel with calcium at gelation causes significant syneresis and thus poor reproduction of the original CAD design. However, if extra calcium is not added, the pore size of the alginate gel is sufficiently large to allow the liquid gelatin to penetrate the surface of the alginate gel. Following chelation, this generates a 'halo' of gelatin around the templated gelatin channels, greatly increasing the diameter of the finest channels, as illustrated in figure 2c. Additionally, the molecular weight of gelatin reduces with increasing temperature [18] and thus using gelatin at higher temperatures produces a thicker halo.
To overcome these issues, one must further cross-link the channel surfaces of the alginate hydrogel by injecting CaCl 2 via the inlets, and use gelatin at 378C. Sufficient cross-linking of the alginate at the channel surface prevents the halo effect and produces channels close to the 3D printed model design. In order that the injected calcium penetrates to the alginate surface, the 3D printed model material must be completely removed; a layer of model material prevents the calcium from cross-linking the alginate sufficiently. To this end, the remaining ( polyester-based) model material is dissolved using acetone once the channels have become patent. If acetone is not used, the gelatin template is turbid and some model material can be left on the surface (figure 2a). In total, 250 mM CaCl 2 is then injected around the channel network, cross-linking the alginate channel surface but without causing significant syneresis of the whole gel.
The choice of alginate as an intermediary casting medium enables fine features of the network to be translated into the gelatin network, though there is some variation between the CAD design and the 3D printed model, between the model and the gelatin template, and between the gelatin template and hydrogel channels, as shown in figure 2d. As the 3D printed models are produced with features at the lower limit of the Solidscape machines, there is some variation in their size; a CAD feature with a diameter of 170 mm yielded a model with features in the range of 200-250 mm.
Using negative pressure, liquid gelatin can infiltrate the channel network. If positive pressure is used instead, gelatin infiltrates the interface between the alginate gel and silicone chamber, and also coats the inlets, which prevents the fibrin hydrogel from adhering to the (hydrophobic) silicone chamber. When the gelatin is liquefied, this opens leak paths around the fibrin hydrogel. The alginate gel was chosen to withstand the applied negative pressure without collapsing.
The thermoplastic model has in its surface 12 mm thick terraces due to the layer-by-layer nature of the 3D printing process. This fine stepping is visible in the gelatin template, as shown in figure 2e, which demonstrates a high casting fidelity of surface features. On a larger scale however, the overall channel structure is less well preserved due to the weak mechanical properties of gelatin and the suction employed for its infiltration into the alginate channels. Additionally, there is a notable change in the diameter of the gelatin features on variation of the volumes of acetone and CaCl 2 injected into the alginate channels, both of which produce at least some shrinkage of the alginate gel, acting to reduce the volume of the channels within. The result is channels finer than those modelled in the CAD (figure 2d). However, as shown in figure 2f-m, there is good correspondence between the 3D printed model and gelatin template.
In this study, both fibrin and collagen hydrogels were used to test the use of the gelatin template with ideal tissue construct materials. Fibrin was easier to handle because at 20 mg ml -1 , the precursor solution has low viscosity and once thrombin has been added, can be cast quickly using a 1 ml syringe. Conversely, at concentrations of 7.5 mg ml -1 , collagen gel is much more viscous and requires slow casting using a syringe pump. Collagen precursor solutions gel rapidly at room temperature and so must be kept on ice during this process.
The choice of bulk material was influenced by a need for perfusion and thus a requirement for a tight interface to an external pumping system; thus, silicone inlets were used. It is well known that hydrophobic materials, such as silicone and Teflon, have some protein-binding capacity [32]. In this study, we rely on this adsorption of protein to the surface of silicone rather than the use of other agents to directly bond the gel at the interface. This interface, as shown in figure 2n, was tested using fluorescent red 1 mm beads, which showed minimal leaking around the silicone inlet surface when used with fibrin gels. Further, perfusion was tested up to 30 ml h -1 (at the inlet) with 10 mg ml -1 fibrin gels, which showed no signs of leaking from the chamber.
The use of gelatin as the sacrificial template material has the advantage that it is thermally removed at physiological temperatures. This allows the cell-loaded constructs to be immediately placed in incubator conditions and puts minimal stress on the cells themselves. The use of gelatin enables a wide range of materials to be employed as the final cast, beyond those of thermally cross-linked collagen and enzymatically cross-linked fibrin hydrogels. The template is capable of withstanding a viscous casting material, as shown by 7.5% (w/v) collagen (figure 1h), which can be infiltrated without significant displacement of the gelatin template, by way of slow casting with a syringe pump. The release of the template constituents upon thermal dissolution impacts minimally upon the surrounding cellular environment due to the relatively benign nature of gelatin. We postulate that the gelatin template could be used with other hydrogels of relevance to the tissue engineering field, such as alginate, hyaluronic acid, Matrigel, synthetic materials such as PDMS, and also interpenetrating polymeric networks.
As a preliminary evaluation of the vascular system, a series of pilot cell experiments was performed. HUVECs were seeded into the channels via the inlets, as shown in figure 3a. HUVECs spread evenly around the symmetrically bifurcating channels and, after several hours, had adhered to the hydrogel channel walls. After 8 days of perfusion, a higher number of HUVECs was observed in all four channel levels of the template (figure 3b-e). These cells were perfused at relatively low flow rates (0.1 ml h -1 ) so as to not dislodge the cell layer in the initial stages of the endothelium formation. Further, the observation of high numbers of cells over all layers shows that flow is not confined to a single layer and that nutrients and oxygen would reach a fully 3D volume when cells are added to the bulk. Immunocytochemistry for the cell adhesion marker CD31 showed the formation of tight junctions between HUVECs in the channels (figure 3f ).
To evaluate the ability of the channel network to support cells in the bulk of the hydrogel, HDFs were mixed into the fibrin precursor solution prior to gelation and cell-loaded fibrin gels (10 mg ml -1 ) were perfused at a range of flow rates (0.5-30 ml h -1 ). After 7 days, a live-dead viability assay was performed by cross-sectioning the gel and staining the central surface with calcein AM and ethidium homodimer-1. There was a notable difference in the viability and cell spreading in the constructs at different flow rates; good viability of HDFs was observed in the bulk for flow rates above 10 ml h -1 and further showed increased cell elongation with increasing flow rate, as shown in figure 4a-c. To quantify any improvement in viability, live and dead cells were counted in representative sample regions and plotted against flow rate, as shown in figure 4d. Static conditions (non-perfused) or low flow rates were incapable of supporting the metabolic It is conceivable that with improvements in commercial 3D printing technologies, models could be produced with higher precision, smaller feature sizes, and in new materials with improved mechanical, physical, and chemical properties. The process described here, of converting these 3D printed models into gelatin vascular templates, and subsequently cellularized hydrogel materials, would utilize these advances. Further, as we make use of standard 3D printing technology, the fabrication method is relatively inexpensive in comparison with other approaches, such as those requiring custom-made printers, and so could be widely used for 3D network formation in cellularized constructs.
Conclusion
Current vascularization methods are limited in their capacity for supporting a large construct volume and achieving full three-dimensionality. We report a new method for producing vasculature, which is hierarchical, 3D, and perfusable, in a fully cellularized and bioactive hydrogel. Vascular networks are created by converting 3D printed thermoplastic material into gelatin which, importantly, can be used in cellularized environments cast from extracellular matrix materials and can be removed efficiently for prompt perfusion. We were able to fabricate channels with diameters of 200 mm and apply flow rates comparable with physiological conditions. Further, the ability to use bioactive hydrogel materials with our process enables the support of primary cells, displaying tight cell junctions in the channels, and a high viability and cell spreading in the bulk. These thick constructs are suitable as an inexpensive platform for 3D cell studies and engineering tissue in vitro. Figure 4. To determine the ability to support cells in thick constructs, HDFs were encapsulated in 10 mg ml -1 fibrin gels at 1 Â 10 6 cells ml 21 and perfused at a range of flow rates. After 7 days, the gels were cross-sectioned and stained with calcein AM (green) and ethidium homodimer-1 (red). (a-c) Live-dead staining for flow rates at which the majority of cells are viable. At 10 ml h -1 , most cells are round while at 30 ml h -1 , most cells are spread out. Shown is a cross-section through the middle of the gel normal to the long axis of the channels. (d ) In order to quantify cell viability at different flow rates, the number of live and dead cells were counted in sample regions. Static conditions (no perfusion) and low flow rates produced constructs with high numbers of dead cells and flow rates above 10 ml h -1 produced constructs with high numbers of live cells (*p , 0.05, ***p , 0.001). Comparisons show a statistically significant difference between live cell numbers for flow rates of 1 ml h -1 and 10 -30 ml h -1 , and for those between 10 ml h -1 and 30 ml h -1 . | 7,480.2 | 2016-12-01T00:00:00.000 | [
"Biology",
"Materials Science",
"Engineering"
] |
Ten simple rules for helping newcomers become contributors to open projects
To survive and thrive, a community must attract new members, retain them, and help them be productive [1]. As openness becomes the norm in research, software development, and education, knowing how to do this has become an essential skill for principal investigators and community managers alike. A growing body of knowledge in sociology, anthropology, education, and software engineering can guide decisions about how to facilitate this.
What exactly do we mean by "community"? In the case of open source and open science, the most usual meaning is a "community of practice." As defined by Lave and Wenger [2, 3], groups as diverse as knitting circles, oncology researchers, and web designers share three key characteristics:
Participants have a common product or purpose that they work on or toward.
They are mutually engaged, i.e., they assist and mentor each another.
They develop shared resources and domain knowledge.
Brown [4] specializes this to define a "community of effort" as
…a community formed in pursuit of a common goal. The goal can be definite or indefinite in time, and may not be clearly defined, but it is something that (generally speaking) the community is aligned on.
People working to preserve coral reefs in the face of global climate change are an example of such a community. No central organization coordinates their work, but the scientists who study coral reefs, the environmentalists who work to protect them, and the citizens who support them financially and politically are aware of each other’s efforts, collaborate in ad hoc ways, and are conscious of contributing toward a shared purpose.
Open-source software projects are also communities of effort. E.g., the Mozilla Firefox [5] community includes a mix of paid professionals, highly involved volunteers, and occasional contributors who not only create software, documentation, and tutorials but also organize events, answer questions in online forums, mentor newcomers, and advocate for open standards.
Every community of effort has unique features, but they have enough in common to profit from one another’s experience. The 10 rules laid out below are based on studies of such communities and on the authors’ experience as members, leaders, and observers. Our focus is on small and medium-sized projects, i.e., ones that have a handful of to a few hundred participants and are a few months to a few years old but may not (yet) have any formal legal standing, such as incorporation as a nonprofit.
Introduction
To survive and thrive, a community must attract new members, retain them, and help them be productive [1]. As openness becomes the norm in research, software development, and education, knowing how to do this has become an essential skill for principal investigators and community managers alike. A growing body of knowledge in sociology, anthropology, education, and software engineering can guide decisions about how to facilitate this.
What exactly do we mean by "community"? In the case of open source and open science, the most usual meaning is a "community of practice." As defined by Lave and Wenger [2,3], groups as diverse as knitting circles, oncology researchers, and web designers share three key characteristics: 1. Participants have a common product or purpose that they work on or toward.
2. They are mutually engaged, i.e., they assist and mentor each another.
They develop shared resources and domain knowledge.
Brown [4] specializes this to define a "community of effort" as . . .a community formed in pursuit of a common goal. The goal can be definite or indefinite in time, and may not be clearly defined, but it is something that (generally speaking) the community is aligned on.
People working to preserve coral reefs in the face of global climate change are an example of such a community. No central organization coordinates their work, but the scientists who study coral reefs, the environmentalists who work to protect them, and the citizens who support them financially and politically are aware of each other's efforts, collaborate in ad hoc ways, and are conscious of contributing toward a shared purpose.
Open-source software projects are also communities of effort. E.g., the Mozilla Firefox [5] community includes a mix of paid professionals, highly involved volunteers, and occasional contributors who not only create software, documentation, and tutorials but also organize events, answer questions in online forums, mentor newcomers, and advocate for open standards.
Every community of effort has unique features, but they have enough in common to profit from one another's experience. The 10 rules laid out below are based on studies of such communities and on the authors' experience as members, leaders, and observers. Our focus is on small and medium-sized projects, i.e., ones that have a handful of to a few hundred participants and are a few months to a few years old but may not (yet) have any formal legal standing, such as incorporation as a nonprofit.
Rule 1: Be welcoming
Karl Fogel wrote [6], "If a project doesn't make a good first impression, newcomers may wait a long time before giving it a second chance". Other authors have empirically confirmed the importance of kind and polite social environments in open-source projects [7][8][9]. Therefore, projects should not just say that they welcome new members: they should make a proactive effort to foster positive feelings in them. One way to do this is to post a welcome message on the project's social media pages, Slack channels, forums, or email lists. Projects might also consider maintaining a dedicated "Welcome" channel or list, where a project lead or community manager writes a short post asking newcomers to introduce themselves.
Other ways to be welcoming include offering assistance in finding ways to make an initial contribution, directing the newcomer to project members who have a similar background or skillset so as to demonstrate fit to the newcomer, and pointing the newcomer to essential project resources (e.g., the contribution guidelines). It also helps to clearly identify work items they can start with; a growing number of projects explicitly tag bugs or issues as "suitable for newcomers" and ask established members not to fix them in order to ensure there are suitable places for new arrivals to start work.
Projects can further designate one or two members to serve as a point of contact for each newcomer. Doing this may reduce the newcomer's hesitancy to ask questions, particularly when they are told from the outset that there are no dumb questions in the community.
Rule 2: Help potential contributors evaluate if the project is a good fit
People could contribute to many different projects; the first and most important step in being welcoming is to help them determine whether your project is a good fit for their interests and abilities. Their decision to contribute can be related to reputation or external needs but also to a desire to learn or give back to the community. In all of these cases, the more you help newcomers understand whether this is the right project for them, the more quickly they will either start contributing or look elsewhere.
To do this, the project should explicitly state what the different types of skills required are. This information should be easily accessible and guide new members to the tasks they may handle. LibreOffice, e.g., provides a way for developers to filter available tasks by required skills and difficulty [10].
The project should also help developers evaluate their skills, since "basic Python skills" means very different things to different people. Tools like My GitHub Resume [11] and Visual Resume [12] that aggregate information from previous contributions can help with this assessment, while the now-defunct OpenHatch project [13] aggregated entry-level issues from a variety of open-source projects and classified them according to language and other required skills to provide a one-stop portal for finding appropriate projects.
Rule 3: Make governance explicit
Raymond's "The Cathedral and the Bazaar" [14] described an egalitarian world in which everyone could contribute equally to open projects. Two decades later, we can see how unequal and unwelcoming the supposedly egalitarian "bazaar" of open source can be if authority lies with those willing to shout loudest and longest. As Bezroukov pointed out [15], Raymond ignored the realities of how power arises, becomes concentrated in a few hands, and is then used to perpetuate itself.
Bezroukov's criticism drew on Freeman's influential essay "The Tyranny of Structurelessness" [16], which explained how an apparent lack of structure in organizations " . . .too often disguised an informal, unacknowledged and unaccountable leadership that was all the more pernicious because its very existence was denied". The solution is to make a project's governance explicit so that people know who makes which decisions.
Large, well-established projects that incorporate as nonprofits are required to promulgate bylaws, such as those for the Python Software Foundation [17]. What smaller projects should do is less well-documented but generally falls under one of three headings [6]. The first is a "benevolent dictator" (often the project founder), who the community agrees has final say on important issues. This model is common in young or small projects but is brittle and inevitably fosters the emergence of unofficial (and hence unaccountable) de facto leaders in specific areas.
The second model formalizes a consensus-building process in which the whole community can take part. One example is Martha's Rules [18], under which anyone can put forward proposals, but those proposals are only adopted once it is clear that most people are not strongly opposed. The third model is based on elected representation. In the Carpentries [19], e.g., the electorate includes anyone who has • completed instructor certification in the preceding year; • completed certification in the last two years and taught at least one workshop; • been certified for more than two years and has taught at least twice in that time; or • made a significant contribution to lesson development, infrastructure, or other activities as determined by the Executive Council.
Decisions are then made by those elected, though they may decide or be required to take some matters to a referendum vote.
More complex models are possible [20], but the most important thing is to decide on the rules well in advance of contentious issues emerging, since tempers may already be running hot by the time this point is reached.
Rule 4: Keep knowledge up to date and findable
When starting to contribute to a project, newcomers must orient themselves in an unfamiliar landscape [21]. It is therefore important to make sure that all necessary information is both accessible and findable. A single project may use wikis, files in GitHub, shared Google Docs, old tweets or Slack messages, and email archives; keeping information about a specific topic in a single place and clearly defining the purpose of each communication medium saves newcomers from having to navigate multiple unfamiliar data sources to find what they need. Doing this makes newcomers more confident and oriented [22].
At the same time, outdated documentation may lead newcomers to a wrong understanding of the project, which is also demotivating. While it may be hard to keep material up to date, community members should at least remove or clearly mark outdated information. Signaling the absence or staleness of material can save newcomers time and also suggest opportunities for them to make contributions that they themselves would find useful.
One special case of this rule is to provide "how to contribute" guidelines in easy-to-find, readily available places. Many projects follow GitHub's recommendation for placing such information in a CONTRIBUTING.md file [23]. Other projects, such as the Apache Open Office Suite and rOpenSci, provide newcomer manuals and learning modules accessed through a web interface [24,25]. Still others take a more interactive approach; e.g., the GNOME project's Newcomers' Guide [26] walks potential contributors through the contribution pipeline: choosing a project, acquiring and installing the necessary computing tools, finding problems or choosing issues to work on, submitting changes, and following up on feedback.
Such guidelines do more than just describe how to contribute. First, their mere existence can ease newcomers' hesitation about whether or not their work is sufficient and suitable for the project. Second, they provide a centralized, well-organized description of resources that a newcomer can consult while learning to navigate the project's technical and social environments [27]. Guidelines also acclimate newcomers to the norms of work and communication, particularly when items such as necessary computing tools and codes of conduct are foregrounded.
Rule 5: Have and enforce a code of conduct
Community leaders should model the behaviors they want to encourage, but that by itself is not enough: experience shows that communities must also make norms about acceptable behavior explicit. This helps ensure that everyone, not just newcomers, will find the environment healthy and welcoming. It also sends a clear signal that the community actually has standards: many potential contributors will be painfully familiar with communities that don't and are more likely to give yours a try if they believe it is not just another troll-infested chat room. Being explicit also makes the project more accessible to people from differing cultural backgrounds because it helps them understand how expectations may differ from what they are used to.
A popular way to make norms explicit is to adopt a code of conduct. Research on these is still in its infancy [28], but many projects such as rOpenSci [29], NumPy [30], and Project Jupyter [31] have adopted the Contributor Covenant [32] or used other frameworks such as SciPy's Code of Conduct [33].
A code of conduct is only useful if there is a clear reporting mechanism that community members trust and if it is enforced [34]. Projects should designate an independent party (i.e., an individual not employed by or otherwise closely connected to the project) to receive and review reports. An independent party offers a degree of objectivity and can help protect reporters from hesitating to raise issues concerning project leaders out of fear of retribution or damage to their reputation. When possible, the independent party should be part of a more extensive code of conduct committee made up of several people with varied characteristics (e.g., gender identity, race, ethnicity, roles in the community). Any member of the committee implicated in the incident should recuse themselves from reviewing the violations.
Project leaders should also develop and publicize enforcement mechanisms, which may range from verbal or written warnings, limits on access to project communication avenues (e.g., Slack channels or mailing lists), or suspension or expulsion from contributing to the project. When safe for the reporter, project leaders should also publicize enforcement decisions: if this is not done, the community may come to believe that the code is meaningless.
Rule 6: Develop forms of legitimate peripheral participation
A core concept in the theory of communities of practice is that of legitimate peripheral participation (LPP) [2,3]. Newcomers become members of a community by participating in simple, low-risk tasks that further the goals of the community. Through these peripheral activities, newcomers become acquainted with the community's tasks, vocabulary, and governance so that they can ease into the project.
In communities such as GitHub, core activities such as committing code and submitting pull requests can be socially daunting for newcomers [35]. One way to encourage LPP in this case is to encourage newcomers to submit issues to a repository when they notice a bug or to join the dialog on recently submitted pull requests or issues. Another way is to have newcomers help with documentation, particularly with translation and localization, and a third (mentioned in Rule 3) is to mark some issues as suitable for newcomers.
Building multiple ways of participating in a community demonstrates the variety of approaches newcomers can take to join the community. This further demonstrates that there is not just one way to make technical contributions. E.g., the main form of interaction in the community on Stack Overflow is to ask a question and post an answer, but engaging in that type of interaction can present barriers for some users, including an intimidating community size and fear of negative feedback [36]. Thus, it is important to provide additional forms of participation. On Stack Overflow, this is demonstrated through the ability to edit questions and answer without the restriction of reputation points. Developing a pathway to participation can decrease the presence of barriers. In studying the evolution of how content is formed in these communities [37], newcomers can better understand the norms of a community and the best way to contribute [38].
Rule 7: Make it easy for newcomers to get started
One way to facilitate LPP is to make it easy for newcomers to get set up so that they can start work on contributions. Getting set up to work on a project-going from "I want to help" to "I'm able to help" to "I'm helping"-is often someone's first experience as a community participant. Any complexity or confusion at this point is therefore a significant barrier to participation [39]. By treating the process of getting involved with the same care and attention you give to the product itself, you're making it clear that you value those contributors' time and effort and forestalling reactions like this [40]: I am still trying to build, because many errors occurred. . . I was expecting to move forward, because so far I did not have time to look at the source code. . . It is frustrating.
This work does not just benefit newcomers; it also helps retention of existing intermittent contributors, and the same work that makes your project more accessible to new contributors today will do the same for future you. Wheelchair ramps and the buttons that open heavy doors are not just used by those in wheelchairs: they are just as helpful to people with strollers or one too many bags of groceries. None of us are ever more than a sprained ankle away from desperately wanting that wheelchair ramp to be there. In that same vein, a drive failure will someday force you to download a gigabyte of data and reinstall some software, inevitably at the least convenient moment imaginable. There is therefore a lot to be gained from automating as much of your setup process you can and thoroughly documenting whatever you cannot.
Rule 8: Use opportunities for in-person interaction-With care
Open-source software projects often rely heavily on remote workers communicating via text, audio, and video. Research on face-to-face and audio/video-mediated communication is mixed with regard to their comparative effectiveness [41][42][43] but demonstrates that each form has benefits and drawbacks. In-person interaction is valuable for uninterrupted, synchronous dialog and helps to establish mutual understanding in a streamlined way [44]. Projects can therefore benefit from engaging newcomers in in-person interaction from time to time.
According to Huppenkothen and colleagues [45], newcomers may particularly benefit from events that ". . .combine structured periods focused on pedagogy (often with an emphasis on statistical and computational techniques) and less structured periods devoted to hacks and creative projects, with the goal of encouraging collaboration and learning among people at various stages of their career." Combining newcomer-friendly events and activities with larger gatherings such as conferences also amortizes participants' financial costs and travel time.
However, potential contributors might shy away from the project if they are introverted, suffer from social anxiety, or have had bad experiences in the past in face-to-face settings. A code of conduct helps allay these concerns, but some newcomers may still feel uncomfortable in group settings. In this case, not going to a meetup may leave them feeling less a part of the community.
Face-to-face communication also involves forms of information exchange that are not easily captured and archived for all project members to see. E.g., collocated project members might hash out ideas on whiteboards, by scribbling notes, or through informal chats. Even when transcribing and/or taking photos of these is possible, important contextual information may be lost [46]. Decisions and changes may seem to come out of nowhere when evaluated by a nonattendee, so project leads should develop universally accessible ways to communicate and explain the results of in-person activities.
Rule 9: Acknowledge all contributions
People in open source sometimes joke that a programmer is someone who will do something for a laptop sticker that they would not do for a hundred dollars. The kernel of truth in this joke is that gratitude and recognition are the most powerful tools community builders have. It is therefore crucial to acknowledge newcomers' contributions and thank them for their work. Every hour that someone has given your project may be an hour taken away from their personal life or their official employment; recognize that fact and make it clear that while more hours would be welcome, you do not expect them to make unsustainable sacrifices.
To ensure completeness and fairness, every project should adopt and publicize guidelines describing what constitutes a contribution, how contributions will be acknowledged, and how they will be used. Who can use the data collected by the project for what purposes, and what attribution do they have to give? How must they acknowledge the project and/or its contributors? Who holds the copyright on contributed material? Most projects now place this information in files called LICENSE.md and CITATION.md and place a brief, readable summary in plain language in onboarding materials.
Rule 10: Follow up on both success and failure
Once someone has carried their first contribution over the line, you and they are likely to have a better sense of what they have to offer and how the project can help them. Helping newcomers find the next problem they might want to work on or pointing them at the next thing they might enjoy reading is both helpful and supportive. In particular, encouraging them to help the next wave of newcomers is both a good way to recognize what they have learned and an effective way to pass it on.
Mentoring programs are a popular way to do this. However, their effectiveness appears mixed. [47] found that ". . .developers receiving deliberate onboarding support through mentoring were more active at an earlier stage than developers entering projects through conventional means". In contrast, [48] found that ". . .developers who join an organization through these programs are half as likely to transition into long-term community members than developers who do not use these programs. . . although developers who do succeed through these programs find them valuable." One explanation for this disparity is that people become members of open projects for different reasons and hence respond to things like mentoring programs in different ways. E.g., Barcomb and colleagues identified four types of episodic or intermittent contributors to opensource projects [49], while Mäenpää and colleagues looked at how to reconcile the competing yet complementary needs of stakeholders in hybrid open/commercial projects [50]. More research is needed, but as openness becomes the norm in research, doing it well becomes a core skill for every researcher.
When they can, projects should also try to follow up on their failures. Why did potential contributors not become community members? Did they realize that the project wasn't a good fit (in which case, the overview may need an overhaul)? Was it too difficult to find a starting point or to get set up to start work (in which case information may need to be consolidated, tagged, filled in, or updated)? Or did they feel uncomfortable or undervalued (in which case the community may need to have a more difficult conversation)? The conversations with individuals should in most cases be confidential, but making the conclusions and corrective actions public is the best possible way to signal that you are serious about building the best community you can.
Martha's rules
1. Anyone may put forward a proposal up to 24 hours before a meeting. Proposals must include a one-line summary, a brief description, any required background information, and a discussion of pros and cons (including alternatives).
2. Once a person has sponsored a proposal, they are responsible for it; the group may not discuss or vote on the issue unless the sponsor is present.
3. After the sponsor presents the proposal, a "sense" vote is taken prior to any discussion in which people indicate whether they like the proposal, can live with it, or are uncomfortable with it.
4. If all or most of the group likes or can live with the proposal, it moves to a formal vote with no further discussion.
5. If most of the group is uncomfortable with the proposal, it is postponed for further rework by the sponsor.
6. If some members are uncomfortable, they can briefly state their objections. After 10 minutes of moderated discussion, the facilitator calls for a yes-or-no vote on adoption. If a majority vote "yes", the proposal is implemented. Otherwise, the proposal is returned to the sponsor for further work. | 5,780.6 | 2019-09-01T00:00:00.000 | [
"Computer Science",
"Sociology",
"Education"
] |
Chaotic Convection in a Viscoelastic Fluid Saturated Porous Medium with a Heat Source
Chaotic convection in a viscoelastic fluid saturated porous layer, heated from below, is studied by using Oldroyd’s type constituting relation and in the presence of an internal heat source. A modified Darcy law is used in the momentum equation, and a heat source term has been considered in energy equation. An autonomous system of fourth-order differential equations has been deduced by using a truncated Fourier series. Effect of internal heat generation on chaotic convection has been investigated. The asymptotic behavior can be stationary, periodic, or chaotic, depending upon the flow parameters. Construction of four-scroll, or “two-butterfly,” and chaotic attractor has been examined.
Introduction
Transport phenomenon in a fluid saturated porous medium is of great practical importance in many areas such as geothermal energy utilization, oil reservoirs, solar energy storage systems, passive cooling of nuclear reactors, pollutant transport in ground water, and storage of chemical and agricultural products, to mention a few.The problem of convection in a fluid saturated porous medium has been studied during the past few decades due to its applications in thermal and engineering sciences.An interesting problem was studied by Horton and Rogers [1] and independently by Lapwood [2], who addressed the Rayleigh-Bénard convection in porous media.Katto and Masuoka [3] employed Darcy's law to express the fluid characteristics in porous layer and experimentally showed the effect of Darcy's number on the onset conditions of buoyancy-driven convection.Important reviews of most of the findings on convection in porous medium are given by Ingham and Pop [4], Vafai [5], and Nield and Bejan [6].
The concept of chaos was first introduced by Poincaré [7,8], who investigated orbits in celestial mechanics and realized that the dynamical system generated by the three-body problem is quite sensitive to the initial conditions exhibiting chaotic behavior.Since the introduction of the chaotic attractors by Lorenz [9] to study atmospheric convection, many chaotic systems have been introduced, such as Rössler [10], Chen and Ueta [11] systems.Because of their potential applications in engineering, the study of chaotic systems has attracted the interest of many researchers.Recently Vadász et al. [12] have investigated the effect of vertical vibrations on chaotic convection in porous medium employing Darcy model.Their results show that periodic solutions and chaotic solutions alternate as the value of the scaled Rayleigh number varies, when forced vibrations are present.Very recently Kiran and Bhadauria [13] have studied chaotic convection in a Newtonian fluid saturated porous medium under temperature modulation at the boundaries.They found that the effect of temperature modulation is to enhance the behavior of chaotic motion.Very recently Bhadauria et al. [14] and Bhadauria and Kiran [15] have studied the chaotic convection using different models.
Although viscoelastic fluid in porous media has been considered many years before by Marshall and Metzner [16] and James and McLaren [17], it is only recently that attention has been given to convection in viscoelastic fluid saturated porous media.Kim et al. [18] studied thermal instability of viscoelastic fluid in porous media by making linear and nonlinear stability analyses and obtained the stability criteria The motive for the present work is to study the effect of internal heat source on dynamics of convection in a viscoelastic fluid saturated porous medium.In the momentum equation, the viscoelastic model of the Oldroyd type is considered by a modified Darcy law.To deduce fourdimensional system, the Fourier series expansion has been applied to the governing equations of the thermal convection in a viscoelastic fluid saturated porous medium.The effects of internal heat source, relaxation, and retardation parameters on the dynamics of the system have been examined in detail.Further, the present system has been reduced to some famous systems provided in the literature.
Mathematical Formulation
An infinitely extended horizontal viscoelastic fluid saturated porous layer, confined between two impermeable boundaries at = 0 and = , heated from below and cooled from above, has been considered.A Cartesian frame of reference is chosen in such a way that the origin lies on the lower plane and the -axis as vertical upward.Adverse temperature gradient is applied across the porous layer and the lower and upper planes are kept at temperatures 0 + Δ and 0 , respectively.Oberbeck-Boussinesq approximation is applied to account the effect of density variations.Under these postulates the governing equations for thermal convection in a viscoelastic fluid saturated porous medium are given by where q is velocity (, V, ), is the dynamic viscosity, is permeability, is the thermal diffusivity, is temperature, is thermal expansion coefficient, and is the density, while 0 and 0 are the reference density and temperature, respectively.The externally imposed thermal boundary conditions are given by = 0 + Δ, at = 0, = 0 , at = , (5) where Δ is the temperature difference across the porous medium.
Basic State
The basic state is assumed to be quiescent, and the quantities in this state are given by Substituting ( 6) in ( 1)-( 4), we get the following relations, which helps us to define basic state pressure and temperature: The solution of ( 8), subject to the boundary conditions (5), is given by The finite amplitude perturbations on the basic state are superposed in the following form: We introduce (11) and the basic state temperature field given by ( 10) in ( 1)-( 4).The resulting equations are then nondimensionalized using the following transformations: Introducing the stream function such that * = −/ * and * = / * in the resulting momentum equation (1) and then eliminating the pressure by operating curl on it, we get the following momentum and energy equations in nondimensionalized form (dropping the asterisks) as where Va = Pr/Da is the Vadasz number, Ra = (Δ)/] is the Rayleigh number, = 2 / is the internal Rayleigh number, Da = / 2 is the Darcy number, and Pr = ]/ is the Prandtl number.
The nondimensional basic temperature field , which appears in (14), can be obtained from the expression (10) as
Mathematical Solution
To obtain the solution of the nonlinear coupled system of partial differential equations ( 13)-( 14), we represent the stream function and temperature in the following Fourier series expressions [12,13,15]: Substituting expressions ( 16) in ( 13)-( 14), using the orthogonality condition with the eigenfunctions associated with expressions (16), and integrating over the domain, we get a set of ordinary differential equations for the time evolution of the amplitudes, in the form where Γ is nondimensional relaxation time, Λ is ratio of retardation time to relaxation time, and 2 = ( 2 + 2 ).Also In ( 17) above, time has been rescaled as = ( 2 + 2 ).The above low-order spectral model may qualitatively reproduce convective phenomenon observed in the full system.Further, the solution of the system can be used as initial values in studying the fully nonlinear convection problem.Now, for our convenience, we use the following notations: Then after rescaling the amplitudes in the form we get the following set of equations: where "⋅" denotes the derivative with respect to the scaled time .
Equilibrium Points.
Setting the time derivatives of system (21) to vanish, we obtain the equilibrium points for velocity and temperature fields as Solving the above algebraic equations, we get the trivial solution which corresponds to pure heat conduction solution.This is known to be a possible solution though it is unstable when (Ra) is sufficiently large.The other two equilibrium points are the solutions ( 2,3 , 2,3 , 2,3 , 2,3 ) characterize the onset of finite amplitude steady motions, where > /.
Stability of the Equilibrium Points.
The Jacobian matrix of system (21) may be written as The stability of the fixed point corresponding to pure conduction solution ( 1 = 1 = 1 = 1 = 0) is governed by the roots of the following characteristic polynomial equation for the eigenvalues, ( = 1-4): Stability depends upon the value of Γ; for Γ < ( − )/[(Λ − 1)] there is an exchange of stability, and for other two steady state solutions origin loses its stability.When Γ > ( − )/[(Λ − 1)], there is a pair of pure imaginary roots of (32).The oscillatory or overstable solutions arise at a critical value of Rayleigh number given by The stability of the fixed point corresponding to convection solution ( = = ±√( − ), = ( − ), = 0) is governed by the roots of the following characteristic polynomial equation for the eigenvalues, ( = 1-4): The steady state solutions are useful because they predict that a finite amplitude solution to the system is possible for subcritical values of the Rayleigh number and that the minimum values of Ra for which a steady solution is possible lie below the critical values for instability to either a marginal state or an overstable infinitesimal perturbation.
Results and Discussions
In the previous section, the set of results were obtained for supercritical values of with the effect of internal heat generation.All the calculations were taken over using Mathematica's inbuilt forth-order Runge-Kutta method.Solutions were obtained using the same initial conditions which were selected to be in the neighbourhood of the positive convective fixed point.The common initial conditions, (0) = (0) = (0) = 0.9 and (0) = 0.1, have been chosen.The time domain is taken as 0 ≤ ≤ 100 with a constant time step, Δ = 0.001.In the paper, we demonstrate the effects of internal heat source and viscoelastic parameters on the system, in the form of space projections of trajectories onto the -, -, and - planes, as the value of increases.
In Figure 1, the initial supercritical solutions are presented at fixed value of = 0.1[1( 1 - 4 )], = 2[1( 5 - 8 )], and = 5[1( 9 - 12 )] with Λ = 0.7 and have been projected onto - planes.From Figure 1, we observed that for weak internal heating, that is, = 0.1 at = 0.987361, the motionless solution loses its stability and the convection solution takes over.For = 2, trajectories move towards steady convection stability point on a straight line for a Rayleigh number slightly above the loss of stability of the motionless solution.From Figure 1, it is evident that there is phase (spiral) trajectory for = 8, as the flow exhibits an oscillatory decay.In Figure 1, spiralling approach of the trajectories towards the steady state fixed point is quite pronounced.This behavior is also inferred from the roots of (32).The flow undergoes a homoclinic bifurcation, as shown in Figure 1 for = 9.294, similar to that predicted by the Lorenz [9] equations.This is known as a global bifurcation which cannot be detected through local stability analysis around the fixed point.On further increasing the value of , the flow becomes completely chaotic as can be depicted from the spatiotemporal structure of flow.The transition to chaos in this case is similar to that leading to the Lorenz attractor.It does not follow any of (a 1 ) the well known routes to chaos, such as through period doubling, quasiperiodicity, or intermittency [43].The behavior of the system at = 28 is chaotic which is confirmed by the broadening of the base in the power spectrum.For the moderate heating, that is, = 2 at = 0.756963, the motionless solution loses its stability and the convection solution takes over.For = 2, trajectories move towards steady convection stability point on a straight line for a Rayleigh number slightly above the loss of stability of the motionless solution.At = 8, it shows the spiral approach before it is attracted towards steady state fixed point.As the value of increased to = 9.294, the flow becomes homoclinic in nature, and, on further increasing to = 28, a transition to chaos occurs.In the case of strong heating, for = 5, at = 0.430905, the motionless solution loses its stability and the convection solution takes over.For = 2, trajectories move towards steady convection stability point on a straight line for a Rayleigh number slightly above the loss of stability of the motionless solution.In this case for = 8 spiralling approach is even more pronounced, a fact which is consistent with the eigenvalues.From Figure 1 ( 11 ), for = 9.294, a homoclinic bifurcation occurs which is more pronounced in this case.Finally at = 28, transition to chaos occurs.Figure 2 ( 1 - 12 ) is the plots of - plane at = 0.1[2( 1 - 4 )], = 2[2( 5 - 8 )], and = 5[2( 9 - 12 )] with Λ = 0.9.In Figure 2 ( 1 - 12 ), we find qualitatively similar results as shown in Figure 1 ( 1 - 12 ), except that amplitudes take a little bit larger value than for Λ = 0.7.
Figure 3 shows the initial supercritical solutions at fixed value of = 0.1[1( 1 - 4 )], = 2[1( 5 - 8 )], and = 5[1( 9 - 12 )] with Λ = 0.7, projected onto - planes.From Figure 3 (c 1 -c 12 ) we observed that for weak heating, that is, = 0.1 at = 0.987361, the motionless solution loses stability and the convection solution takes over.For = 2 trajectories move towards steady convection stability point through a spiral for a Rayleigh number slightly above the loss of stability of the motionless solution.From Figure 3 ( 2 , 6 , 10 ) it is clear that there is spiral trajectory for = 8 as the flow exhibits an oscillatory decay.In Figure 3 (c 2 , c 6 , c 10 ), spiralling approach of the trajectories towards the steady state fixed point is more pronounced.For = 9.19, there is a homoclinic pattern of flow, as shown in Figure 3 (c 3 , c 7 , c 11 ).Further increasing the value of , there occurs a transition to chaotic convection as can be seen from Figure 3 (c 4 , c 8 , c 12 ).The transition to chaos in this case is similar to that leading to the Lorenz attractor; one can easily see "two-butterfly" Figure 3 ( 4 ) for = 28.For the case of moderate heating, that is, = 2 at = 0.756963, the motionless solution loses stability and the convection solution takes over.Trajectories move towards steady convection stability point through a spiral for a Rayleigh number slightly above the loss of stability of the motionless solution for = 2.At = 8 trajectories move via a spiral, before they are attracted to their steady state fixed points.At = 9.19, the flow becomes homoclinic in nature and attracted towards a steady state fixed point.On increasing the value of at = 28, a transition to chaos occurs and "two-butterfly" Figure 3 ( 4 ) are obtained.For the case of strong heating, namely, = 5, the motionless solution loses its stability and the convection solution takes over at = 0.430905.For = 2 trajectories move towards steady convection stability point through a spiral for a Rayleigh number slightly above the loss of stability of the motionless solution.In this case for = 8, spiralling approach is even more pronounced, a fact which is consistent with the eigenvalues.From Figure 3 ( 11 ) for = 9.19, a homoclinic bifurcation occurs which is more pronounced in this case.Finally at = 28 transition to chaos occurs and "twobutterfly" nature occurs.In Figure 4 ( 1 - 12 ), which are for Λ = 0.9, we found qualitatively similar results to those of Figure 3 ( 1 - 12 ) with slightly large amplitudes.
The initial supercritical solutions, calculated at = 0.1, 2, and 5 with Λ = 0.7, projected onto - planes have been presented in Figure 5. From Figure 5 ( 1 - 4 ), we observe that for the case of weak heating, that is, at = 0.1, the motionless solution loses its stability at = 0.987361, and the convection solution takes over.For = 2 trajectories move towards steady convection stability point through a spiral for a Rayleigh number slightly above the loss of stability of the motionless solution.From Figure 5 ( 6 ) it is clear that there are spiral trajectory for = 8 as the flow exhibits an oscillatory decay.In Figure 5 ( 5 ), spiralical approach of the trajectories towards the steady state fixed point is more pronounced.For = 9.45, there is a homoclinic bifurcation, as shown in Figure 5 ( 7 ).Further increasing the value of , there occurs a transition to chaotic convection as can be seen from Figure 5 ( 8 ).The transition to chaos in this case is similar to that leading to the Lorenz attractor, and one can easily see the transition to chaos for = 28.For the case of moderate heating, that is, = 2 at = 0.756963, the motionless solution loses its stability and the convection solution takes over; the trajectories move towards steady convection stability point through a spiral for a Rayleigh number slightly above the loss of stability of the motionless solution at = 2.At = 8 trajectories move via a spiral, before it is attracted to its steady state fixed point.At = 9.45, the flow becomes homoclinic in nature and attracted towards a steady state fixed point.On increasing the value of at = 28, a transition to chaos occurs.For the case of strong heating, namely, = 5 at = 0.430905, the motionless solution loses its stability and the convection solution takes over; for = 2 trajectories move towards steady convection stability point through a spiral for a Rayleigh number slightly above the loss of stability of the motionless solution.In this case for = 8, spiralling approach is even more pronounced, a fact which is consistent with the eigenvalues.From Figure 5 ( 11 ) for = 9.45, a homoclinic bifurcation occurs which is more pronounced in this case.Finally at = 28 there occurs transition to chaos.Figure 6 ( 1 - 12 ) is the plots of initial supercritical solutions, and it is found that the results are qualitatively similar to those obtained in Figure 5 ( 1 - 12 ).
Figure 1 :
Figure 1: Projections and evolution of trajectories over the planes -.
Figure 7 :
Figure 7: Projections and evolution of trajectories. | 4,219.6 | 2016-03-10T00:00:00.000 | [
"Engineering",
"Environmental Science",
"Physics"
] |
X-domatic Partition of Bipartite Graphs
In this study, we define X-domatic partition of a bipartite graph. X-domatic number of a unicyclic graph is less than or equal to 3 is proved. Nordhaus Gaddum type of results involving X-domatic number are obtained and the graphs for which d x (G) + d x (G) = 2P, d x (G) + d x (G) = 2P-1, are characterized.
INTRODUCTION
Graph colouring and domination in graphs are two areas within graph theory which have been extensively studied. Graph coloring partitions the set of objects into classes according to certain rules. Similar to the concept of chromatic partition, (Cockayne and Hedetniemi, 1977) introduced the concept of domatic partition of a graph. The idea of k-rainbow domatic number was introduced by Sheikholeslami and Volkmann (2012). The concept X-domatic partition of a bipartite graph is an extension of the concept of domatic partition of a graph.
Let G be a graph. A subset S of V is a dominating set of G if for every v ∈ V-S there exists a vertex u∈ S such that u and v are adjacent. The minimum cardinality of a dominating set is called the domination number of G and is denoted by γ(G). A partition {V 1 , V 2 ,⋯, V n }of V is a domatic partition of G if every V i is a dominating set. The maximum order of a domatic partition of G is called the domatic number of G and is denoted by d(G). For further details on domatic number, the reader is referred to the paper on domatic number of graphs by Zelinka (1981).
We shall consider only connected bipartite graphs with bipartition (X, Y) where |X| = P; |Y| = q and have no loops and multiple edges. Bipartite theory of graphs was proposed by Stephen and Renu (1986 a, b) in which concepts in graph theory have equivalent formulations as concepts for bipartite graphs. One such reformulation is the concept of X-dominating and Ydominating set.
Two vertices u, v ∈ X are X-adjacent if they are adjacent to a common vertex in Y. A subset D of X is an X-dominating set if every vertex in X -D is Xadjacent to at least one vertex in D. The minimum cardinality of a X-dominating is called the Xdomination number of G and is denoted by γ X (G). If x ∈X, then the X-degree of x is defined to be the number of vertices in X which are X-adjacent to x and is Proposition 1: γ x (C l ) = [ ] for every l 4, l ≡ 0, 2, 4, (mod 6).
A subset S ⊆ X which dominates all vertices in Y is called a Y-dominating set of G. The Y-domination number denoted by γ Y (G) is the minimum cardinality of a Y-dominating set of G.
Vertices u, v, w are adjacent to a vertex y ∈ Y, we say that the vertices u, v, w are X-adjacent through same y ∈ Y. If the vertices u, w are adjacent to y 1 ∈ Y and u, v are adjacent to y 2 ∈ Y, we say that the vertices u, v, w are X-adjacent through different Y y . A bipartite graph G is said to be X-complete if every vertex in X(G) is X-adjacent to every other vertex in X (G).
X-domatic partition of a graph:
An X-domatic partition of G is a partition of X, all of whose elements are X-dominating sets in G. The X-domatic number of G is the maximum number of classes of an X-domatic partition of G. The X-domatic number of a graph G is denoted by d x (G).
Remark 1:
Since {X} itself is a X-domatic partition of G , the existence of a X-domatic partition is guaranteed for any bipartite graph and so the parameter d x (G) is well defined. Further, obviously 1 d x (G) P for any bipartite graph.
The X-domatic number of a bipartite graph is analogous to the concept of domatic number of an arbitrary graph. For more information on the domination number and on the domatic number and their invariants, the reader is referred to survey book by Haynes et al. (1998) and Zelinka (1988).
Theorem 2: If G is a connected unicyclic graph then d x (G) 3.
Proof: Assume dx(G) = 3. When |X| 2, G will have a vertex of full X-degree implying d x (G) = 2, a contradiction. Hence, |X| 3. G Cannot have a pendent X-vertex since in that case d x (G) 2. Hence, G = C 2n .
Corollary 2:
If G is a connected unicyclic graph with |X| 3, then d x (G) = 2 if and only if G C 2n where |X| ≡ 0 (mod 3).
Proof:
Let d x (G) = 2 and |X| 3. By the above corollary G C 2n where |X| ≡ 0 (mod 3). Conversely, let G C 2n where |X| ≡ 0 (mod 3). Since G is unicyclic, d x (G) 3. d x (G) cannot be three by the above corollary and d x (G) cannot be 1. Therefore, d x (G) = 2. It is easy to prove the following.
Proposition 2: In any graph G, d x (G) = P if and only if G is X-complete.
Now we define the complement of a bipartite as defined by Sampathkumar and Pusphalatha (1989). Let G = (X, Y, E) be a bipartite graph. We define the complement of G denoted by G = (X, Y, F) as follows: No two vertices in X are adjacent No two vertices in Y are adjacent x ∈ X and y ∈ Y are adjacent in G if and only if x ∈ X and y ∈ Y are not adjacent in G Proposition 3: In any graph G, d x (G ) = P if and only if for every u, v ∈ X, there exists y ∈ Y such that y is not adjacent to u, v in G.
Proof:
Let d x (G ) = P. Every vertex in G is X-adjacent to all other vertices. There exists y ∈ Y such that u and v are adjacent to y ∈ Y in G . Then, y ∈ Y is not adjacent to u and v are in G. Conversely, suppose that for every u, v ∈ X, there exists y ∈ Y such that y is not adjacent to u, v in G. Then y is adjacent to u, v in G . Hence, u and v are Xadjacent in G . Hence, d x (G ) = P.
NORDHAUS-GADDUM TYPE RESULTS
Nordhaus-Gaddum type theorem establish bounds on θ(G) + θ(G ) for some parameter θ, where G is the complement of G. Several Nordhaus-Gaddum type theorems were characterized for various domination parameters, some are rainbow vertex-connection number (Chen and Memngmeng, 2011) of a graph, rainbow connection number (Chen et al., 2010) of graphs and k-rainbow domatic number (Meierling et al., 2011 ) of a graph. Here we characterize graphs for which d x (G) + d x (G ) equal to 2P and 2P-1. To characterize the graphs attaining the bounds, we define the following family of graphs.
Λ is the family of graphs such that every vertex in X is of X-degree P-1 and for any two vertices u and v in X, there exists y ∈ Y such that u and v are not adjacent to y.
Λ 1 is the family of graphs with a vertex in Y of degree |X| and a unique γ Y -set of cardinality 2.
Λ 2 is the family of graph with |X| = 2 and at least one vertex in X and Y is of full degree. Proof: If d x (G) + d x (G ) = 2P then d x (G) = P and d x (G ) = P. By proposition 2, we get every vertex in X is of X-degree (P-1). Conversely, suppose that for every u, v ∈ X, there exists y ∈ Y such that y is not adjacent to u, v in G. Then y is adjacent to u, v in G . Hence, u and v are X-adjacent in G . Hence, d x (G ) = P. Converse is obvious. If d x (G) = P, every vertex in X is of X-degree P-1. If every vertex in X is of X-adjacent to other vertices through different y ∈ Y, then d x (G ) = P, a contradiction. Therefore, there exists a vertex y ∈ Y of full degree.
Claim: γ Y (G) 2. Suppose, γ Y (G) 3. Then any subset of X of cardinality 2 is not a Y-dominating set of G. Therefore, if u, v ∈ X, then there exists y ∈ Y such that u and v are not adjacent with y. Hence, u and v are X-adjacent in G and so d x (G ) = P P-1, a contradiction. Hence, γ Y (G) 2.
Let S 1 = {u 1 , u 2 } be a γ Y -set of G.
Claim: u 1 and u 2 are not X-adjacent in G . Suppose u 1 and u 2 are X-adjacent in G . Then there exists y ∈ Y (G ) such that u 1 and u 2 are adjacent with y ∈ Y (G ). Therefore, u 1 and u 2 are not adjacent with y in G. Hence, y is not dominated by u 1 and u 2 in G, a contradiction.
Claim:
The existence of γ Y -set is unique in G.
Let S 1 and S 2 be two minimum Y-dominating sets of G. Let S 1 = {u 1 , u 2 } and S 2 = {u 3 , u 4 }. Therefore, u 1 and u 2 are not X-adjacent in G . Similarly u 3 and u4 are not X-adjacent in G . If u 1 , u 2 , u 3 and u 4 are all distinct, then {u 1 }, {u 2 }, {u 3 }, {u 4 } are all non X-dominating set in G . Therefore, d x (G ) P-2, a contradiction. Let S 1 and S 2 have a common vertex. Let u 2 = u 4 . Then {u 1 }, {u 2 }, {u 3 } are all non X-dominating set in G . Therefore, d x (G ) P-2, a contradiction. Therefore, γ Y -sets of cardinality 2 in G is unique. Therefore, G ∈ Λ 1 .
Case 2: Let d x (G) = P-1 and d x (G ) = P. Using the above argument, we get G is in {Λ 1 , Λ2}. | 2,538.2 | 2013-01-11T00:00:00.000 | [
"Mathematics"
] |
Drought Tolerance of Legumes: Physiology and the Role of the Microbiome
Water scarcity and global warming make drought-tolerant plant species more in-demand than ever. The most drastic damage exerted by drought occurs during the critical growth stages of seed development and reproduction. In the course of their evolution, plants form a variety of drought-tolerance mechanisms, including recruiting beneficial microorganisms. Legumes (one of the three largest groups of higher plants) have unique features and the potential to adapt to abiotic stress. The available literature discusses the genetic (breeding) and physiological aspects of drought tolerance in legumes, neglecting the role of the microbiome. Our review aims to fill this gap: starting with the physiological mechanisms of legume drought adaptation, we describe the symbiotic relationship of the plant host with the microbial community and its role in facing drought. We consider two types of studies related to microbiomes in low-water conditions: comparisons and microbiome engineering (modulation). The first type of research includes diversity shifts and the isolation of microorganisms from the various plant niches to which they belong. The second type focuses on manipulating the plant holobiont through microbiome engineering—a promising biotech strategy to improve the yield and stress-resistance of legumes.
Introduction
Drought is one of most important stress factors that negatively affects plant development.Climate change and population growth expand the territories affected by water shortages.Drought is the main cause of reductions in crop yield and hinders the progress of agriculture in many countries [1].Soil aridity and the spread of deserts have become a global challenge that cannot be solved using irrigation and technical measures.Water shortages deteriorate morphological, physiological and biochemical processes such as the development of leaves and roots, oxygen absorption and photosynthesis.The most drastic damage exerted by drought takes place during the critical phases of growth: seed development and reproduction.The main symptoms of drought occurring in legumes (as for other crops) are twisting, burning, wilting and premature shedding [2].
During their evolution, plants, including legumes, developed a variety of droughtresistance mechanisms to regulate their morphological and physiological characteristics such as root and leaf structure, stomatal and photosynthetic regimes, and to accumulate water.Plants are called "drought resistant" if they can grow in medium and severe drought conditions [3].Legumes are one of the three largest higher plant groups.Many legume species are well-known food crops that can acquire soil nitrogen via symbiotic bacteria and resist metals in soil.It should be underlined that Fabaceae members have unique features and the potential to be adapted to abiotic stress [4].To overcome the negative effect of drought, legumes attract beneficial microbes that can form a symbiosis with the host.Experimental studies show that the composition and diversity of the microbial community varies between the plant parts (leaves, roots and seeds).Drought tolerance in legumes can be promoted by mediating the microbial community using a variety of approaches.Although drought is a well-known type of abiotic stress, existing reviews mostly cover the genetic (breeding) and physiological aspects of drought resistance in legumes [5][6][7].Even Special Issues devoted to the role of the microbiome in climate change mitigation are missing reviews of recent studies on the topic [6,8].
The available literature discusses the genetic (breeding) and physiological aspects of drought resistance in legumes, neglecting the role of the microbiome.Filling this gap, in this review we consider the mechanisms used by legumes to endure droughts and droughtinduced changes in the microbiome.These adaptive mechanisms are diverse and include the synthesis of phytohormones and osmoprotectants, the recruitment of microorganisms and changes in plant metabolism.In addition to legumes' own ability to attract beneficial microbes, there are a number of approaches to modifying the microbial community to make the plant more drought-tolerant, such as microbial engineering.The aim of our review is to highlight the "best practices" in producing drought-tolerant legumes and to suggest directions for further research.
Common Drought-Adaptation Mechanisms in Plants
Drought stress leads to a high concentration of reactive oxygen species (ROS), which can be extremely harmful, especially hydroxyl radicals and singlet oxygen.ROS causes cell damage, membrane and protein degradation, and lipid oxidation DNA fragmentation; eventually, these processes lead to cell death [9].Drought stress can also change the carbon and nitrogen biogeochemical cycle, which reduces the absorption of water and nutrients by the roots and lowers the cations' conductivity (e.g., Ca 2+ , K + and Mg 2+ ).Legumes have diverse mechanisms to reduce the amount they are affected by drought, which is common for the majority of crops.The first response to drought stress involves receiving a membrane receptor-mediated signal.The signal is then transduced to express the appropriate genes.Mitogen-activated protein kinases (MAPK) and Ca 2+ -dependent protein kinases transmit the signals to the nucleus, which activates various regulons, controlling the expression of drought-resistant genes, DREB, MYB/MYC, NAC, ABRE and WRKY.
Role of Phytohormones
Phytohormones are crucial substances, which activate a lot of cell pathways during stress.One of the main stress phytohormones is abscisic acid (ABA).It takes part in signal acquisition from the environment as well as in the regulation of physiological and biochemical features [10].Despite the numerous works on crops devoted to ABA pathways, the number of studies covering legumes is limited [11,12].
Salicylic acid (SA) is another important phytohormone that regulates plant growth and abiotic stress reactions [13].Exogenous SA promotes drought resistance in plants by regulating the protein kinases' activity, as well as the chlorophyll and rubisco concentration [14].Melatonin is an evolutionally conserved molecule, which is contained in most living organisms and possesses biologically essential properties.The role of melatonin in drought-resistant crops is poorly understood, but a number of studies show that the exogenous treatment of several legumes promotes drought resistance through inhibiting membrane injury [15,16].The physiological and molecular activity of melatonin in plants shows that it is an important substance to stimulate Fabaceae plants, especially under the action of abiotic stress [17].
Osmoprotection System
Many osmoprotectants (sugars and sugar alcohols), such as mannitol, sorbitol, inositol, trehalose, proline, ectoine, glycine and betaine, play a key role in cells' drought resistance.They inactivate ROS and stabilize proteins and membranes [18].To regulate the osmotic pressure in cells and protect the membrane from damage, some oligosaccharides (trehalose, raffinose, fructose and saccharose) are employed.They can also act as signaling molecules.A high content of carbohydrates in plants may evidence their drought resistance, while wa-ter shortages increase the expression of genes related to the synthesis of carbohydrates [19].The metabolism of amino acids during abiotic stress, in which proline plays an important role, should also be borne in mind.Polyamines are low-molecular aliphatic compounds.Among them, putrescine, spermidine and spermine are most frequently met in plants.The level of endogenous polyamine is induced by abiotic stress (including by drought), but exogenous spermidine treatment overexpresses the polyamine biosynthetic genes [20].
Reactive Oxygen Species (ROS)
As mentioned above, prolonged drought leads to the excessive accumulation of ROS in plant tissues.To protect plants against ROS, bacteria with antioxidant potential are used.These bacteria support biochemical changes (content of proline, proteins, and antioxidant enzymes) and promote plant growth [21].Elevated ROS levels are also controlled by ROS efflux systems, which include both non-enzymatic antioxidants and antioxidant enzymes.Furthermore, the levels of various ROS are regulated during N 2 accumulation in legume roots using symbiotic bacteria [9].
The Role of the Beneficial Microbes to Face Drought
In nature, plants co-exist with various microorganisms, such as viruses, bacteria, archaea, oomycetes, and fungi.Plants usually attract various kinds of microbes to promote their growth, and all of them interact with each other in a complex manner.Some experts even call these microorganisms "second genome of the plant".The soil microbiome protects plants against drought and improves the yield and soil fertility.Many studies highlight the importance of rhizosphere microbiome in improving drought resistance.It seems obvious that attracting beneficial microbes is a common evolution strategy for plants under water shortages [22,23].The plant microbiome has attracted the interest of the agricultural research community in terms its uses in sustainable crop production and food security [24].Soil microbes take part in a variety of processes that are crucial for plant productivity: nutrients circulation, soil mineralization, resistance to diseases and overcoming abiotic stresses (high salinity and drought).Many legumes are known to be involved in such plant-microbe interactions [25].
Generally, a plant can be considered as a holobiont and unified biological object of evolution (Figure 1).During mutual adaptation, many plant species (including legumes) formed close symbiotic relationships with bacteria and fungi.Such symbioses allow for the acquisition of a rich spectrum of nutrients that would be unavailable without the symbionts-plants themselves lack the necessary enzymatic systems (such as nitrogen fixation).Symbiotic relationships bring advantages to both sides: microsymbionts obtain access to the host's resources, while the macrosymbiont (host) uses microbial metabolites to enrich nutrition and resist abiotic stresses.From an evolutionary perspective, the symbiose leads to the pooling of hereditary information towards more diverse, top-specie heredity systems, which accelerate the evolution processes [26].Drought stress is a major selection factor, shaping the rhizospheric drought-resistant microbiome over many seasons.This trait may be inherited from the entire hologenome of the plant and its biota, i.e., it indicates how the plant can memorize multiple past drought stress events [27].A prolonged drought period may irreversibly change the microbiome.Recently, Santos-Medellín et al. showed that, even after removing drought stress, the endophytic microbiome could not be regenerated to the initial state.During long periods of water shortage, the microbiome configuration changes and plant health is damaged [28].Let us take a closer look at each of the hologenome niches.A number of high-throughput sequencing studies show that the composition of the microbial community differs significantly between plant parts (leaves, roots, seeds, and rhizosphere).In general, plant host characteristics, species, age, crone type, genotype and sterility can significantly change the microbial composition.Moreover, the plant host genotype determines the profile of microbial community members, as reported for soybean [29].Usually, the population of microbes is highest in the soil and decreased in the rhizosphere, phyllosphere and endosphere (Figure 1), showing the selection gradient [30].
Phyllosphere has a more dynamic environment than the endosphere and rhizosphere (Figure 1) because its microbiome has fewer common taxa than the endosphere and rhizosphere.To colonize various parts of a plant, microbes use a vast arsenal of tools: biofilms, biosurfactants, quorum sensing, pilis, flagella, adhesion molecules, etc.To protect plants from sunlight (UV), epiphytic bacteria synthesize the pigments.Ascomycota and Acidobacteria are dominating phyla in the phyllosphere and rhizosphere, as reported for the soybean [31].Compared to the leaf surface, the endosphere is likely to be richer in nutrients and has a more stable environment (protected from fluctuations in the atmosphere, including UV radiation, temperature and moisture).Otherwise, endophytes have a closer interaction with the plant host immunity system that restrains bacteria reproduction.
To attract the beneficial bacteria in soil and leaves, plant uses exudates.These substances can contain amino and fatty acids, sugars, growth factors, vitamins, etc.With the help of secondary metabolites (flavonoids, coumarins, citrates oxalates), plants can recruit specific types of microbes in the rhizosphere, phyllosphere and endosphere [30,32].Some researchers call this action a "cry for help" [33,34].This phenomenon is well-known for A number of high-throughput sequencing studies show that the composition of the microbial community differs significantly between plant parts (leaves, roots, seeds, and rhizosphere).In general, plant host characteristics, species, age, crone type, genotype and sterility can significantly change the microbial composition.Moreover, the plant host genotype determines the profile of microbial community members, as reported for soybean [29].Usually, the population of microbes is highest in the soil and decreased in the rhizosphere, phyllosphere and endosphere (Figure 1), showing the selection gradient [30].
Phyllosphere has a more dynamic environment than the endosphere and rhizosphere (Figure 1) because its microbiome has fewer common taxa than the endosphere and rhizosphere.To colonize various parts of a plant, microbes use a vast arsenal of tools: biofilms, biosurfactants, quorum sensing, pilis, flagella, adhesion molecules, etc.To protect plants from sunlight (UV), epiphytic bacteria synthesize the pigments.Ascomycota and Acidobacteria are dominating phyla in the phyllosphere and rhizosphere, as reported for the soybean [31].Compared to the leaf surface, the endosphere is likely to be richer in nutrients and has a more stable environment (protected from fluctuations in the atmosphere, including UV radiation, temperature and moisture).Otherwise, endophytes have a closer interaction with the plant host immunity system that restrains bacteria reproduction.
To attract the beneficial bacteria in soil and leaves, plant uses exudates.These substances can contain amino and fatty acids, sugars, growth factors, vitamins, etc.With the help of secondary metabolites (flavonoids, coumarins, citrates oxalates), plants can recruit specific types of microbes in the rhizosphere, phyllosphere and endosphere [30,32].Some researchers call this action a "cry for help" [33,34].This phenomenon is well-known for the nitrogen fixation with Rhizobia and plant-growth-promoting fungi in cases where there are low levels of nitrogen and phosphates in soil [35].Their high ability to acquire atmospheric nitrogen is the key feature of legumes.The symbiosis of Fabaceae plants with soil bacteria (rhizobia) represents one of the best-known mutualistic plant-microbe interactions because of its contribution to the sustainability of agricultural systems and human nutrition [36].This interaction is so close that it even affects the flowering pathways, as was shown for soybean [37].During symbiosis, rhizobia fixate on atmospheric nitrogen, which becomes available to the plant.The amount of acquired nitrogen often meets most of the plant's needs, and the nitrogen retained in the soil becomes available for the following season's crops [38].
We should note that the interaction with the root exudates of cereals affects rhizobia activity during intercropping.In [39], the greatest effect was found for maize (compared with wheat, or barley).The involvement of nitrogen-fixing symbionts has an important advantage: it allows for the use of synthetic nitrogen fertilizers, the long-term use of which seems to lead to the predominance of less effective strains of rhizobia in the agroecosystem, to be reduced.Rhizobia are not the only root symbionts: the existence of various bacterial endophytes within nodules was reported for many legumes [40,41], and experiments involving co-inoculation with rhizobia suggest that a number of endophytes associated with nodules may stimulate growth and be safe and effective partners.Rhizobial inoculants are available and easy to use, so they have been developed and employed worldwide.However, the effect of the application of outdoor rhizobia on legume productivity varies widely and seems to depend on both environmental constraints and cultivation history.Recently, the advantages of legume selection for N 2 fixation in parallel with crop rotation (versus the intercropping of grain legumes) in small-scale agriculture in Africa have been highlighted [42].
Although the legumes can attract the beneficial microbes themselves, a number of approaches can mediate the microbial community to make the plant more tolerant toward drought.These approaches will be considered in the following sections.
Microbiome Engineering
The mechanisms of plant-bacterial communication are still poorly understood, but this knowledge is crucial for the further engineering of microbiomes with particular features [43,44].This kind of host-mediated microbiome engineering was described for the soil microbiome of Arabidopsis, when plants selected microbes that would help to change the leaf biomass and flowering time [45], and may develop seeds before drought, which would cause plant disease or death.The endophytic microbiome is sensitive to drought and quickly responds to drought stress.As a result, diversity rises and shifts, while the interaction between plants and endophytes intensifies [46].There are two main microbiome-engineering approaches to overcome drought stress: "synthetic communities" (SynComs) [47,48] and "host-mediated microbiome engineering" (HMME) [49].Both approaches have recently been applied to legumes, but the number of studies where the authors clearly define the approach used as HMME or SynComs is limited: SynComms for Medicago sativa [50], and for Crotalaria juncea, and Canavalia ensiformis [51].Nevertheless, new experiments and methods for Fabaceae species can be suggested based on the results for other crops, some of which are briefly described below.
The first approach (SynComs) deals with the design of inoculants using microbial ecology and genetics approaches, as well as functions, which could improve plant characteristics and promote plant-microbe and microbe-microbe interaction [52].For example, Rolli et.al. developed SynCom using Bacillus, Acinetobacter, Sphingobacterium, Delftia and Enterobacter for grapes, which not only protect plants during drought, but promote growth and yield [53].A similar approach was applied to blue maize when combined inoculation with several bacterial strains (P.putida KT2440, A. brasilense Sp7, Acinetobacter sp.EMM02, and Sphingomonas sp.OF178) promoted growth better than a monoculture.This bacterial consortium possesses desirable features for application in sustainable agriculture, even for different maize varieties [54].Using synthetic biology approaches, we can construct specialized SynComs, selecting members of the community to evaluate the impact of each bacterial strain.
Host-mediated microbiome engineering is an innovative approach to developing longterm beneficial microbiome features when the host phenotype is employed for the indirect iterative selection of microbiomes.Its main advantage over the SynComs approach is in the fact that most selected microbes are adapted to stress conditions and have a strong relation to the plant host.Despite the elegant concept, this approach usually has modest efficiency and the selection process can be unsuccessful [55].The number of studies using the HMME approach has been limited to date, but some of the results are encouraging.For example, to protect Brachypodium distachyon from salinity stress, Muller et al. applied the HMME and defined the beneficial microbiome.Some microbial communities increased the seed yield to 55-205% in comparison with the control (in addition to salinity resistance) [56].An important question is how to maximize the impact of the genetic effects encoded by the microbial community to the host traits.A promising strategy is to limit or infer the host's genetic contribution to the phenotype of the whole system (plant-microbiome).The inbred or cloned plant populations could be used to minimize genetic variation.Under such conditions, genetic effects encoded by the microbial community become the major factor [57].
The typical scenario of root microbiome changes during drought was documented in a number of works.Diversity shifts towards Gram-positive bacteria, especially Actinobacteria, while the Gram-negative ones (the usual population of the rhizosphere) lose their niche.This Gram-positive enrichment is proportional to drought duration and severity [58].When water returns to the soil, the microbiome quickly returns to its original state.Several hypotheses were proposed to explain this conservative pattern [59].In general, these hypotheses are based on metatrancriptome data of microbial communities suffering from drought, but a detailed analysis is hard to perform because information about the functional and genetic features of rhizosphere community members is lacking [60].Manipulating the plant holobiont through microbiome engineering is a promising biotechnology strategy to improve the yield and stress resistance of legumes.In the next section, we review two types of studies: inoculation of Fabaceae plants with microbes (with sole and mixtures), and isolation of microbial cultures for the further inoculation of non-legumes.
Microbiome Modulation of Fabaceae Plants
A lot of work was devoted to plant-growth-promoting rhizobacteria (PGPR), but most of it was not focused on specific plant species or families [61].An important approach to overcome the negative impact of drought is various types of inoculation (seeds or soil treatment with microbial mixtures).Here, we briefly describe some recent successful efforts in this direction for legumes.
It seems that the Pseudomonas species is a very common beneficial component in drought-resistant bacterial mixture or in sole action.In many studies, these bacteria were isolated from soil roots or used as biofertilizers.Pseudomonas bacteria can synthesize indole acetic acid (IAA), 1-aminocyclopropane-1-carboxylate-deaminase (ACC), siderophores and successfully colonize roots by forming biofilms.The ability of Pseudomonas to alleviate drought stress in Vigna radiata was evaluated by Uzma et al.Five Pseudomonas were isolated and used as bioinoculants [62].Conversely, the drought-tolerant species can be a source of specific bacterial isolates.An effort to move microbiome components (bacteria and fungi) to improve drought resistance was made using Alhagi sparsifolia, a known desert plant.Lei Zhang et al. extracted microbes from the plant host rhizosphere and isolated the Pseudomonas strain LTGT-11-2Z cell culture.When introduced to the wheat soil, this cell culture improved wheat drought resistance [63].The isolated microbes can also help plants adapt to the non-natural environment in the case of transplantation.The composition of the microbiome is an important factor for the growth of wild plant species in the field or greenhouse.Zuo et al. tried to transplant the natural-habitat soil fungal community to the pot experiment.These fungi (species A. chlamydospora, S. kiliense, and Monosporascus sp.) showed high survivability under drought stress, which appeared to be developed during their long-term adaptation to low water conditions [64].
In addition to microbiome manipulation, the genetics are also studied, but such works are rare for legumes.Most studies are focused on the microbiome composition, neglecting the genetic features of particular beneficial bacteria, the expression levels of genes related to plant growth promotion and drought resistance.To provide further insight into plant-bacterial interactions under stress conditions, Nishu et al. isolated the Pseudomonas fluorescens DR397 and performed in vitro polyethylene glycol-based screening experiments [65].As a result, the versatile strain Pseudomonas fluorescens DR397 could be used as a promising biofertilizer, improving plant drought tolerance.We previously mentioned the common beneficial factor, the expression of ACC deaminase, which reduces the concentration of ethylene in plants.Andrey Belimov and coauthors evaluated the role of the acdS gene using a knockout mutant.The experiment showed that the ACC deaminase of rhizosphere bacteria promoted the successful nodulation of pea (Pisum sativum) [66].Further, the same team performed a pot experiment with pea line SGE and its Cd-tolerant mutant SGECd t , which were cultivated under optimal and limited water conditions.They reported that water stress affected the rhizosphere microbiome far more significantly than plant genotype (in terms of alpha and beta diversity indices) [67].
The application of the microbial mixture instead of a monoculture seems to be a more promising approach due to the higher stability and versatility of the obtained community.To extend the biochemical activity of sole Pseudomonas, Mora et al. added Bacillus bacteria to an organic biofertilizer [68].The authors note that many species among the Bacillus and Pseudomonas genera have plant-growth-promoting activity.This mixture allows for microbes to hydrolyze and transform complex organic molecules into simpler ones that are accessible for root adsorption.This nutrient biotransformation has a second positive effect, because biomolecules hold better and remain available to the plant's root system.The growth of fava beans (Vicia faba), with a mixture of Rhizobium leguminosarum (Rl) and Pseudomonas putida (Pp) added to the soil, improved water absorption and increased the expression of photosynthetic pigments [69].A very similar study was performed with soya beans (Glycine max) and Azotobacter chroococcum (Az) and Piriformospora indica (Pi) bacterial species.It was reported that water deficiency reduced the growth and yield of soya bean, but the application of Az and Pi decreased the negative effect of water shortages, with no dependence on the irrigation regime being detected [70].
To extend the metabolic potential of rhizosphere organisms, a fungi-bacteria mixture was used in several studies (Table 1).Laranjeira et.al. mixed the prokaryotes (Mesorhizobium sp.UTADM31, Burkholderia sp.UTADB34 and Pseudomonas sp.UTAD11.3) and mycorrhizal fungi (Rhizophagus irregularis, Funneliformis geosporum and Claroideoglomus claroideum) in chickpea soil.As a result, crop yield increased by 6% compared with single inoculation, and by 24% compared to the control plants [71].Long-term drought caused root degradation since the acquisition of water and nutrients stopped.To help plants to restore the root system, Yue et al. grew licorice with Bacillus amyloliquefaciens in near-root soil.This measure promoted root development and changed its structure [72].
Generally, modulated microbiome studies use different compositions of organisms to select the most productive solution.He et.al. described two septate endophytes of A. mongholicus that can colonize the rhizosphere of roots.The combined inoculation of this legume with various fungal and bacterial species showed that dual inoculation with Paraboeremia putaminum and Trichoderma viride had a stronger effect than inoculation with Trichoderma viride and Acrocalymma vagum [73].Inoculating plants with beneficial bacteria and fungi could also help in well-watered environments.The planting of soybean, one of the most important legumes, often takes place in poor soils with an unfavorable water regime.Sheteiwy and coauthors co-inoculated soybean with mycorrhizal spores (inoculum was added to 5 g of trapped soil) and endophytic bacterium Bacillus amyloliquefaciens MN592674B (soybean seeds were soaked in the bacterial culture).Biofertilizers contributed to an obvious reduction in cell size and granularity, which may improve soybean tolerance to drought stress conditions [74].The main source of this promising PGPR for droughttolerant species is arid soils.Verma et al. isolated 50 bacterial strains from rhizosphere samples of lobia (Vigna unguiculata) and performed a set of treatments with different strain combinations (all in pot experiments).The authors reported that treatment with No. T27 (Pseudomonas sp.IESDJP-V1 + Ochrobactrum sp.IESDJP-V5 + A. brasilense) led to more significant results (comparing the plant development to control samples and other treatments).They also note that these findings require further "in the field" validation [75].Thus, the results may provide a platform for further understanding of the molecular mechanisms of bacterially mediated drought resistance in plants.Another important aspect of drought resistance is the microbiome's reaction to the drought stress of various severities.Legumes are known for their ability to survive in mid-arid environments, but how the diversity and shape of the microbial community depends on the irrigation regime remains poorly understood.In the next section, we cover the known studies on this topic.
Legume Microbiome for Different Watering Regimes
It is particularly interesting to compare the microbiome changes that occur under drought stress when other environment conditions (plant site, salinity, nutrients, etc.) remain the same.Unfortunately, the design of the experiments in the considered studies varies significantly so they cannot be directly compared; in addition, some studies lack information on the microbial community composition (16S or ITS amplicon sequencing).However, we can highlight some common aspects.Most studies use two (drought stress and control) or three (well-watering, medium drought, severe drought) irrigation regimes for the soil or simulate drought stress with polyethylene glycol treatment.In addition [76], monocropping was used in all studies.Table 2 shows the common taxa present in both "drought" and "control" samples; a brief summary of each study is given below.
One of the negative factors during low water periods is an excessive concentration of ethylene and its precursor (1-aminocyclopropane-1-carboxylic acid, ACC).Some rhizobacteria can hydrolyze these harmful compounds and reduce their negative impact on the plant host (Cicer arietinum) [77].Other bacteria have the same ACC hydrolyzation activity, particularly the Bacillus species.Andy et al. showed that Bacillus strains (B.cereus and B. haynesii) had deaminase activity, which is sufficient to overcome abiotic stress for two plant hosts (Vigna mungo and Phaseolus vulgaris) [78].The formation of a microbiome during seedling is the initial stage of plant growth.Bintarti et al. focused on the endophytic community of the dormant seeds of Phaseolus vulgaris and its changes during drought.They hypothesize that seed microbiome likely characterizes taxa that have been transferred from parent to seed.Growing common beans in pot experiments with different irrigation regimes showed that, under stress conditions, diversity shifts were much higher in the bacterial community than in the fungal community [79].
A common source of both oil and protein, peanut (Arachis hypogaea L.), has important advantages: self-pollination and aerial flowering.Its main disadvantage is sensitivity to monocropping (and this sensitivity increases during planting years).We found four microbiome studies of peanut related to drought tolerance; common taxa are highlighted with bold font (Table 2).In greenhouse experiments, Dai et al. compared the diversity and composition of the peanut microbial community in control and drought conditions.Six major phyla were dominant in all samples (Actinobacteria, Proteobacteria, Saccharibacteria, Chloroflexi, Acidobacteria, and Cyanobacteria), which are common for many plant species.However, three of them, Actinobacteria, Acidobacteria and Proteobacteria, seem to have mutualistic relationships in peanut soil [80].Although the PGPR has wide metabolic potential, the role of the fungal community cannot be neglected.The arbuscular mycorrhizal (AM) fungi are known to enhance plant growth in many species.Xu and coworkers compared the microbial community in natural, drought and fungal-mediated soil conditions [81].The authors showed that peanut plant drought resistance inreases when the rhizosphere community contains AM fungi.The application of biofertilizers can be combined with the intercropping approach.Intercropping with Mulberry (Morus alba L.) is popular for planting peanuts in China.Li et al. reported that, in an experiment with three plant configurations, pure mulberry planting, pure peanut planting, and mulberry and peanut intercropping, there were significant differences in the bacterial and fungal communities [76].As with most legumes, peanut plants have a close relationship with soil organisms, and in monocropping practices they become more sensitive to fungal pathogens from the soil.To compare the transcriptional response to monocropping combined with drought, Luo et al. performed both field trial and pot experiments with peanut.The authors revealed that long-term monocropping altered the soil structure, raising the percentage of small aggregates and lowering water availability.Monocropping practices increase the severity of drought stress [82].
Studies of microbiome phylogeography are of particular interest: they allow for the determination of the core species of microbiome and match the environmental factors at each plant site with the specialized bacteria.The symbiotic efficiency of fenugreek (Trigonella foenumgraecum) rhizobia depends on the bacterial strain and environmental conditions.Khairnar et al. performed a phylogenetic analysis of housekeeping genes, which revealed unique genotypes of fenugreek rhizobia, such as Ensifer (Sinorhizobium) meliloti.These strains are characteristic of agroclimatic regions of India and differ from other known genotypes [83].Understanding the mechanisms of the mutual interaction of plant-associated microbes in different niches is key to the promotion of plant growth.Shirley Evangilene and Sivakumar Uthandi compared the diversity of the bacterial community in four niches-soil, rhizosphere, root nodules and seeds-of the horse gram (Macrotyloma uniflorum).They reported that the ammonium-oxidizing metabolism (amoA), nitrite-reducing metabolism (nirK) and nitrogenfixing metabolism (nifH) were common and prominent in all niches, but the alpha diversity showed no significant difference.The obtained microbial cultures can substitute the synthetic fertilizers and maintain soil fertility for sustainable agricultural practices [84].
Although most attention is paid to cultured edible Fabaceae representatives, some wild types and varieties are understudied.There are known species growing in arid soils.Bambara groundnut is one of them.It can survive in marginal soils and become tolerant to drought.Ajilogba et al. reported that Bambara groundnut could selectively modulate the composition and potential functions of its microbiome during all developmental stages [85].
We should note that some studies do not contain metagenome or amplicon sequencing data to allow for reproductions of the results.However, the isolated bacterial and fungal strains could be used as a PGPR for other species [77,78].Although the number of studies describing the above certainly is not exhaustive, it allows for us to define the future research directions. 1 for the seedling and flowering stages, respectively. 2no amplicon sequencing for microbiome analysis was performed in this study.
Conclusions
Several approaches can be used to improve drought tolerance in legumes: selection (breeding), genotype modification, agronomic methods and microbiome modulation.Although the reviewed studies are methodologically quite different, we can conclude that a single approach is not sufficient to obtain a stable and productive Fabaceae crop under drought conditions.In works on microbiome modulation, it has been shown that mixtures of bacteria and fungi are more promising biofertilizers than monocultures.
In order to achieve better reproducibility and an easy comparison of results, it seems important to develop a protocol for research on the microbiome of drought-tolerant plant species, particularly legumes.In this protocol, the irrigation regime, sample preparation and DNA extraction technique could be standardized.Unfortunately, many studies do not pay attention to the physical and chemical properties of the soil, which undoubtedly influences the composition of the microbial community.To demonstrate the high potential of the plant microbiome in agriculture, future studies will use a complex approach combining the methods of microbiology, metagenomics, metatranscriptomics and metabolomics.
Figure 1 .
Figure 1.Fabaceae plant hologenome.Legumes interact with microorganisms in the soil and air.The composition of the microbiome depends on the part of the plant.Usually, three compartments, the rhizosphere (near root soil), endosphere (inner tissue of the plant) and phyllosphere (leaves and stems surface), are considered.The beneficial bacteria can promote plant growth by direct and indirect mechanisms.
Figure 1 .
Figure 1.Fabaceae plant hologenome.Legumes interact with microorganisms in the soil and air.The composition of the microbiome depends on the part of the plant.Usually, three compartments, the rhizosphere (near root soil), endosphere (inner tissue of the plant) and phyllosphere (leaves and stems surface), are considered.The beneficial bacteria can promote plant growth by direct and indirect mechanisms.
Table 1 .
Studies devoted to promoting drought resistance by inoculation with the microbe mixture.
Table 2 .
Studies of legume microbiomes under drought stress. | 7,428.4 | 2023-07-28T00:00:00.000 | [
"Biology"
] |
Autophagy of Candida albicans cells after the action of earthworm Venetin-1 nanoparticle with protease inhibitor activity
The present studies show the effect of the Venetin-1 protein-polysaccharide complex obtained from the coelomic fluid of the earthworm Dendrobaena veneta on Candida albicans cells. They are a continuation of research on the mechanisms of action, cellular targets, and modes of cell death. After the action of Venetin-1, a reduced survival rate of the yeast cells was noted. The cells were observed to be enlarged compared to the controls and deformed. In addition, an increase in the number of cells with clearly enlarged vacuoles was noted. The detected autophagy process was confirmed using differential interference contrast, fluorescence microscopy, and transmission electron microscopy. Autophagic vesicles were best visible after incubation of fungus cells with the Venetin-1 complex at a concentration of 50 and 100 µg mL−1. The changes in the vacuoles were accompanied by changes in the size of mitochondria, which is probably related to the previously documented oxidative stress. The aggregation properties of Venetin-1 were characterized. Based on the results of the zeta potential at the Venetin-1/KCl interface, the pHiep = 4 point was determined, i.e. the zeta potential becomes positive above pH = 4 and is negative below this value, which may affect the electrostatic interactions with other particles surrounding Venetin-1.
Scanning electron microscopy (SEM).
The yeast cultures were centrifuged and the supernatant was suspended in a fixative pH = 7 (phosphate buffer, glutaraldehyde, saccharose) and incubated for 2 h at room temperature.Then, the fixative was discarded and 0.1 M phosphate buffer was added.Next, 1.5% OsO 4 was added to the pelleted cells and centrifuged for 30 min at 2500×g.Then, OsO 4 was removed and the cells were resuspended in 0.1 M phosphate buffer and centrifuged for 30 min at 2500×g.After that, the cells were dehydrated in acetone solutions with increasing concentrations: at 30%, 50%, 70%, and twice at 100%.The cell cultures were transferred onto SEM stages, stored in a desiccator with silica gel for 24 h, and sputtered with gold (K550X sputter coater, Quorum Technologies).The cells were imaged with a Vega 3 scanning electron microscope (Tescan, Czech Republic) 49 .
Cryo-scanning electron microscopy (Cryo-SEM). Samples with control C. albicans cells and cells after
incubation with Venetin-1 were centrifuged and the supernatant was discarded.The pellet was suspended in 200 µL of a GH solution, centrifuged (10 min, 6000×g), and almost all of the supernatant was withdrawn.The fungal cells in the GH solution were placed in a sublimation chamber for 12 min at -92 °C.Then, the samples were cut with a special blade in the preparation chamber and observed with the use of an EM ZEISS Ultra Plus SEM microscope (Carl Zeiss, Germany) at 5 kV.
Fluorescence microscopy.
Candida albicans cells for staining with fluorochromes were prepared as described previously in "Microorganisms".Cell suspensions were then incubated with different fluorochromes in proper conditions in order to visualize cell organelles.Stained yeast cells were observed using a confocal microscope (Carl Zeiss, Germany) with immersion.
Quinacrine dihydrochloride.Quinacrine dihydrochloride (Sigma-Aldrich) stains acidic organelles like autophagic bodies (yellowish fluorescence) and autophagosomes (blue fluorescence).Fungal cells for Quinacrine dihydrochloride staining were mixed in a proportion of 10 µL of the cell suspension with 10 µL of fluorochrome www.nature.com/scientificreports/ (1 mg mL −1 water solution) and incubated for 10 min at 37 °C in the dark. 2 µL of cell suspension was transferred onto glass slide, covered with a cover slip, and observed at a 436-nm excitation wavelength for Quinacrine.
Yeast LIVE/DEAD test.The LIVE/DEAD test is used to distinguish between metabolically active, inactive, and dead cells.However, such cellular organelles as mitochondria are stained as well.For the experiment, C. krusei cells were prepared in the same way as the C. albicans cells.The yeast cell suspensions were centrifuged, the supernatant was removed, and the pellets were suspended in GH buffer.Next, the yeast cell cultures were mixed with 10% FUN-1 in GH buffer in a 1:1 ratio and incubated for 30 min at 30 °C. 2 µL of stained cells were transferred onto microscopic slides and fluorescence was observed at a 480-nm emission wavelength 49 .Cells with red inserts in the cytoplasm were considered active, and green and yellow cells were inactive.Mitochondria fluoresced yellow-green at the periphery of the cells.
Acridine Orange.Acridine Orange (AO) is a fluorochrome indicating affinity for nucleic acids and acidic cellular compartments.An AO water solution (0.1 mg mL −1 ) was mixed with the yeast cell suspension in a 1:1 ratio and incubated for 10 min in the dark at room temperature.Samples in a volume of 2 µL were placed on a cover slip and observed at the excitation wavelength λ = 502 nm and emission wavelength λ = 526nm using a Zeiss/ LEO 912AB microscope at 1000× magnification 51 .The cells undergoing the autophagy process were counted after acridine orange staining with the ImageJ program using the multitool.Approximately 500 cells in each sample were counted, and the experiment was repeated three times.
Transmission electron microscopy (TEM).
Yeast cell cultures (preparation described in "Microorganisms") were fixed in GA (4% glutaraldehyde in 0.1 M cacodylate buffer pH = 7.2).Next, the C. albicans cultures were centrifuged (2500×g, 12 min) and rinsed twice with 0.1 M cacodylate buffer with centrifugation (2500×g, 12 min).Then, the pellet was fixed in 1.5% KMnO 4 for 1 h 15 min at 5 °C.After this time, the cells were rinsed several times with distilled water until discolored.Embedded fungal cells were then contrasted with uranium acetate 1%, dehydrated with subsequent concentrations of ethanol, infiltered, and embedded in resin (LR White).The resin blocks were then cut into ultrathin sections and observed using a transmission electron microscope JEM-1400Flash (JEOL, Japan).
Flow cytometry analysis.
The flow cytometry analysis of live/dead yeast cells was performed using a mixture of Propidium iodide (10 µg mL −1 water solution) and Hoechst 33342 (5 µg mL −1 water solution) mixed in a proportion of 2:1.The cell cultures were incubated with a staining mixture at room temperature for 20 min in the dark; then, distilled water was added to a final volume of 800 µL 58 .The analysis was performed using a Guava easyCyte Flow Cytometer (Luminex, USA) with a 695/50 nm laser (Red-B-HLog), a 785/70 nm laser (NIR-R-HLog), Forward Scatter (FSC-HLog), Side Scatter (SSC-HLog), Threshold set on FSC at 68, and the total number of counted cells 5000.The obtained dot-plots were gated with the Quad Stat Marker for viable (upper left square) and dead (lower left square) cell regions.
Cryo-TEM analysis of Venetin-1.
Venetin-1 at a concentration of 1 mg mL −1 was ultrasonicated for 10 min at 40 °C in an ultrasonic chamber (Pol-Sonic, Poland).Then, the sample was vitrificated in a water solution on the TEM grid covered with Quantifoil R 2/2 carbon film (Quantifoil Micro Tools GmbH, Großlöbichau, Germany).Before observations, the grids were activated with oxygen plasma for 15 s in a Femto plasma cleaner (Diener Electronic, Germany).Next, 3 μL of the Venetin-1 suspension was transferred onto the grid, blotted with filter paper and immersed in liquid ethane for instant freezing by Vitrobot Mark IV (FEI Company, USA).Before observations, the samples were stored in liquid nitrogen.To transfer the specimens to the TEM microscope, they were loaded into the Gatan 626 Cryo-TEM holder (Gatan Inc., USA) 59 .The samples were observed using a Tecnai F20 X TWIN microscope (FEI Company, USA) with 200-kV acceleration voltage of the emission gun.An Eagle 4k HS camera (FEI Company, USA) was used to record the images.
Dynamic light scattering domain-specific characterization of Venetin-1. The DLS technique
with the Prometheus Panta was used to characterize the Venetin-1 nanoparticle.It provides the highest quality data on the biophysical characteristics of the analyzed material.Venetin-1 was dissolved in water at a concentration of 1.3 mg mL −1 .The samples were loaded into eight capillaries acting as replicas, where a parallel analysis allowed us to analyze the homogeneity of the nanoparticles and their aggregation abilities under different temperature conditions.
Zeta potential determination.
The electrophoretic mobility, conductivity, and zeta potential measurements of the NaCl solution with the concentration of 10 -4 , 10 -3 M and 10 -3 M KCl were carried out using the Zetasizer Nano ZS90 by Malvern.Smoluchowsky's equation was applied due to the value of κa ~ 150.The measurements were performed at a 100 ppm solid concentration of Venetin-1.The compound was added to the solution and subjected to dispersion using the ultrasound probe Sonicator XL 2020 produced by Misonix.Then, the suspension was poured into 125 mL flasks, and pH was established to be in the range of 3-11 using 0.1 M HCl and NaOH solutions.Five measurements of electrophoretic mobility, conductivity, and zeta potential were made for each solution.
The concentration of the bovine trypsin stock solution was determined spectrophotometrically at 410 nm by titration with the chromogenic burst substrate 4-nitrophenyl 4-guanidinobenzoate (NPGB, Sigma-Aldrich, USA).Later, the standardized trypsin solution was used to titrate turkey ovomucoid third domain OMTKY-3 (Sigma-Aldrich, USA) used as a mutual inhibitor of trypsin and chymotrypsin) with BAPNA as a substrate.Then, OMTKY-3 was used to determine the concentration of the α-chymotrypsin stock solution in the presence of its substrate Suc-Ala-Ala-Pro-Leu-pNA.In the case of MT1 and MT2, the concentrations were calculated in accordance with the information provided by the suppliers.
The tested compound was added to the appropriate protease in different concentrations and incubated in assay buffer at 37 °C for 30 min.After this time, the solution of an appropriate substrate was added.Reactions were monitored for at least 35 min at 37 °C.The final volume in each well was 200 µL.Measurements were carried out using excitation and emission wavelengths of 380 nm and 450 nm (for the substrates of MT1, MT2, and furin), and absorbance for the trypsin and chymotrypsin substrates was monitored at 410 nm.Percentage inhibition of an enzyme was calculated relative to the control sample without the inhibitor.Determination of the inhibitory activities were done according to Gitlin et al. 61 and Gitlin-Domagalska et al. 62 .
The IC 50 values (inhibitor concentrations giving 50% inhibition of enzyme activity) were calculated from plots of enzyme activity (% of the control sample) versus the inhibitor concentration using a four-parameter fit model (GraFit 5.0.12Erithacus Software Ltd.).The IC 50 values were determined from triplicate measurements with at least ten different inhibitor concentrations.
Statistical analysis.
Statistical analyses for assessment of changes in the number of C. albicans cells with autophagic bodies and changes in the cell size after the treatment with Venetin-1 were performed with the Statistica program (Tibco Software Inc., USA; serial number: JPZ009K288211FAACD-Q).The Shapiro-Wilk test was used to check the type of data distribution, and the Levene test was employed to determine the homogeneity of variance.The post-hoc Tukey HSD test and one-way ANOVA were used to examine the level of significance of differences in the number of cells with autophagic bodies and changes in the cell size.
Flow cytometry analysis of C. albicans cells after treatment with Venetin-1.
The flow cytometry analysis of C. albicans cells treated with Venetin-1 performed using a mixture of Hoechst 33342 and Propidium iodide showed significant changes in the level of dead cells in cultures treated with Venetin-1 in comparison to the control culture (Fig. 2).The upper left side of the dot-plot shows that 86% of cells were viable in the control culture.After the incubation with Venetin-1, the level decreased to 62.34%, 56.34%, and 30.02% in cultures after treatment with active compound at 25 µg mL −1 , 50 µg mL −1 , and 100 µg mL −1 , respectively.
SEM analysis of C. albicans cells after treatment with Venetin-1. Control C. albicans cells and
those treated with Venetin-1 were imaged by SEM.The morphology analysis showed that the cells of the control culture had an oval shape and a rough but regular cell wall (Fig. 3A1-A2).After the incubation with Venetin-1 at a protein concentration of 50 µg mL −1 , the yeast cells underwent visible changes.The cells exposed to the active compound clearly enlarged compared to the control cells (Fig. 3B1-B2).In addition, Fig. 3B1 shows cell with exfoliated of the cell wall outer layer, and cell shape deformation is visible in Fig. 3B2.After the treatment with Venetin-1 at 100 µg mL −1 (Fig. 3C1-C2), enlarged cells (Fig. 3C1-C2), as well as cells with numerous division scars and shrunken forms were observed (Fig. 3C2).The average cell size was 5.67 µm in the control cell culture, 7.28 µm in the culture treated with Venetin-1 at the concentration of 25 µg mL −1 , 7.1 µm in the variant with the concentration of 50 µg mL −1 , and 7.37 µm in the treatment with the concentration of 100 µg mL −1 (Supplementary Information Table S1, Fig. S1).Parametrical data distribution was determined with the Shapiro-Wilk test (results in Supplementary Table S1).The Levene test results (F(3, 795) = 12.3181; p < 0.001) indicated nonhomogenous variances.Statistical significance was analyzed with the post-hoc Tukey HSD test and one-way ANOVA: F(3, 795) = 219.12;p < 0.001.
DIC and Cryo-SEM analysis of C. albicans autophagy. The C. albicans control culture cells and the
Venetin-1-treated cells were imaged by DIC.The control culture cells had an oval shape with a vacuole constituting no more than half of the cell lumen (Fig. 4A).After the exposure to the preparation at a concentration of 50 µg mL −1 , enlarged cells with clearly enlarged vacuoles occupying most of the cell lumen were observed (Fig. 4B).Small vesicles (autophagosomes) (marked with red arrows in Fig. 4B) were noticeable at the outer edge of the vacuole.After the incubation of the yeast cells with Venetin-1 at a concentration of 100 µg mL −1 , in addition to enlarged vacuoles and autophagosomes (marked by red arrows), there were vesicles penetrating the vacuole or already present inside (marked with green arrows in Fig. 4C,D; autophagic bodies).After the exposure to the preparation at a higher concentration (100 µg mL −1 ), there were definitely more vesicles around the vacuole than in the treatment with the complex at a concentration of 50 µg mL −1 .When absorbed by vacuoles, they lost their regular round shape.
Images E and F in Fig. 4 show a cross section of a C. albicans cell after the treatment with Venetin-1 at a concentration of 100 µg mL −1 captured by Cryo-SEM.Images E and F present vacuoles and autophagosomes (marked with arrows).Large vacuoles are visible in Fig. 4E and an enlarged cell nucleus is shown by image Fig. 4F.
Detection of autophagy using fluorescence microscopy.
The Quinacrine dihydrochloride staining revealed the presence of blue and blue-green vesicles located both inside and outside the vacuoles (Fig. 5I).The staining of cellular structures in the presented images is non-specific.No vesicle fluorescence was observed in the control C. albicans culture cells.In the microscopic image, the outline of vacuoles with a normal size only was visible (Fig. 5I A1-A2).The C. albicans cells treated with Venetin-1 at a concentration of 50 µg mL −1 were characterized by larger cell and vacuole sizes than in the untreated culture.Both autophagosomes, i.e. vesicles located outside the vacuolar lumen (indicated by red arrows), and autophagic bodies, i.e. vesicles located inside the vacuoles (indicated by yellow arrows), were visible (Fig. 5I B1, B2).In turn, the C. albicans cells treated with Venetin-1 at a concentration of 100 µg mL −1 exhibited a greater amount of both types of autophagic vesicles: autophagosomes (red arrows) and autophagic bodies (yellow arrows).The size of the vacuoles did not differ from the normal size (Fig. 5I C1-C4).
The above observations were confirmed by the analysis of C. albicans cells performed using acridine orange (AO), which stains acidic cell compartments orange or red.This fluorochrome is used to stain autophagic cells with characteristic red fluorescent vesicles (Fig. 5II).The microscopic image of the control culture showed green cells with a normal size (Fig. 5II A1-A2).In turn, small red spots were visible in the yeast cells incubated with Venetin-1 at a concentration of 50 µg mL −1 , indicating the presence of acidic vesicles (pointed by white arrows) (Fig. 5II B1-B2).The C. albicans cells treated with Venetin-1 at a concentration of 100 µg mL −1 had a significantly greater number of red vesicles (indicated by white arrows).In addition, they were larger and more intensely stained than in cells treated with a lower concentration of Venetin-1 (Fig. 5II C1-C4).The average percentage of cells with visible red autophagic bodies was 9.44% in the control cell culture and was higher in samples treated with Venetin-1: 16.25% in the culture treated with the concentration of 25 µg mL −1 , 33.85% in the treatment with 50 µg mL −1 , and 62.70% in the variant with 100 µg mL −1 (Supplementary Information Table S2, Fig. S2).The parametrical data distribution was confirmed with the Shapiro-Wilk test (results in Supplementary Table S2).
TEM analysis of C. albicans cell autophagy.
The ultrastructure of C. albicans cells after the application of Venetin-1 was analyzed with the use of transmission electron microscopy.The control cells were characterized by a regular cell wall and normal intracellular structure with visible organelles, i.e. the nucleus (N), a single vacuole (V), and mitochondria (M) with no signs of autophagy (Fig. 6A).In turn, numerous autophagic vesicles of two types were visible in the C. albicans cells incubated with Venetin-1 at the concentrations of 50 µg mL −1 and 100 µg mL −1 .Moreover, the complex-treated cells had larger vacuoles and more mitochondria than the control culture cells.
The TEM technique allowed observation of several steps of the autophagy process.Mature and completely closed autophagosomes (indicated by red arrows) transporting their contents to the vacuole were visible in the cell cytoplasm (Fig. 6B1-B2,C1-C6).Subsequently, the stages of autophagosome fusion with vacuoles, release of the internal vesicle into the vacuole lumen, and formation of autophagic bodies were captured (Fig. 6B1,C3,C6; autophagic bodies are pointed by yellow arrows).
The vesicles visible in the cytoplasm varied in size, which may indicate the occurrence of different types of macroautophagy.The larger autophagosomes are typical of non-selective macroautophagy, where the vesicle contains bulk cytoplasm, and the smaller ones are typical of the selective process, where specific structures of the yeast cell constitute the contents.However, in the C. albicans cells treated with Venetin-1 at the concentration of 50 µg mL −1 , all autophagy vesicles were noticeably larger than in the cells incubated with Venetin-1 at the concentration of 100 µg mL −1 .The presented pictures are representative of the 30 images obtained.
Visualization of mitochondria using fluorescence microscopy.The Live/Dead Yeast Viability Kit is used as a standard to determine the metabolic activity of fungal cells.The analysis is based on the ability of cells www.nature.com/scientificreports/ to metabolize the FUN-1 dye, as a result of which red structures are visible in cell vacuoles.It has been used in our previous studies 49 .The use of this kit facilitated additional visualization of mitochondria (Fig. 7).The labeling of these organelles was a non-specific effect.The C. albicans control culture was characterized by metabolically active cells with red structures inside the vacuole.In turn, no structures other than vacuoles were visible in the control cells (Fig. 7A1-A2).After the incubation of the C. albicans cells with Venetin-1 at the protein concentrations of 25, 50, 100 µg mL −1 , spot yellow fluorescence of numerous mitochondria was observed.These organelles were arranged in a characteristic way forming a ring under the surface of the membrane (Fig. 7B1-B2,C1-C2,D1-D2; pointed by arrows).
To confirm our observations of mitochondria, another strain of Candida, C. krusei, was incubated with Venetin-1 as well.The effect of mitochondrial fluorescence after the treatment with Venetin-1 in this case was stronger.The control cells showed no stained mitochondria (Fig. 7E1-E2).In turn, in the cells incubated with the active compound, yellow-green fluorescing mitochondria were clearly visible.After the treatment of C. krusei cells with Venetin-1 at a concentration of 25 µg mL −1 , their mitochondria were small and visible as spots arranged on the periphery of the cells (Fig. 7F1-F2; pointed by arrows).In the case of cells treated with Venetin-1 at the concentrations of 50 and 100 µg mL −1 , elongated forms of brightly fluorescent mitochondria were observed (Fig. 7G1-G2,H1-H2; pointed by arrows).
Cryo-TEM analysis of Venetin-1.
The analysis of the Venetin-1 complex by Cryo-TEM confirmed the nanoparticle size determined by the DLS analysis 25 .The nanoparticles were visible under the microscope as spherical forms, which sometimes coalesced to form double forms indicated by red arrows in Fig. 8A,B.Interestingly, the DLS analysis on the fly clearly showed an increase in the overall nanoparticle size resulting in formation of larger 100 nm nanoparticles at a temperature above the melting point (Fig. 9, Table 1).
Determination of Venetin-1 zeta potential.
Five measurements of electrophoretic mobility, conductivity, and zeta potential were made for each solution (Table 2) (Fig. 10A,B).The zeta potential of the tested samples, in the tested pH ranges from 3 to 12 and different electrolyte concentrations, ranged from 10 to − 40 mV.
In the majority of the ranges tested, the selected systems are unstable in the colloidal form and may be delaminated.Stable systems at the absolute potential lower than − 30 mV were obtained only for the following samples: V-1/0.0001M NaCl at pH = 7, 9, 11; V-1/0.001M KCl at pH < 8 (Fig. 10B).The comparison of the results for 0.001 mol/dm 3 NaCl and 0.001 M KCl showed very different absolute values for pH > 4 at the same electrolyte concentrations (Fig. 10A,B).Therefore, the surface potential of Venetin-1 was found to exert the main impact on the zeta potential.
As shown in Fig. 10A,B, the studied system had the highest absolute values of the zeta potential at the lowest electrolyte concentration and the lowest absolute values at the highest electrolyte concentration.It was found that the zeta potential decreased with the increasing electrolyte concentration, resulting from the dissociation of surfaces subjected to ionization on Venetin-1.In the examined systems, the zeta potential decreased with increasing pH.The state of the solid surface in which the amounts of positive and negative charges in the diffuse layer of the electric double layer (edl) are equal to each other is called the isoelectric point of the solid surface (IEP-IsoElectric Point) pH IEP .Then, the resultant charge of the diffuse layer of the double electrical layer is zero.Since the concentration of potential-forming ions (H + and OH − ) depends on the pH of the solution, the pH IEP point corresponds to a precisely determined pH value.The pH IEP values of Venetin-1 are below pH 6 for NaCl (Fig. 10A) and pH IEP = 4 for KCl (Fig. 10B).This means that in a KCl solution above pH = 4, the zeta potential becomes positive and negative below, which may affect the electrostatic interactions with other particles surrounding Venetin-1.
Inhibitory potency against selected serine proteases.The inhibitory activity of Venetin-1 was examined against selected serine proteases, i.e. two digestive enzymes: trypsin and chymotrypsin, and three transmembrane proteases: matriptase-1 (MT1), matriptase-2 (MT2), and furin.In this study, Venetin-1 exhibited moderate inhibitory potency towards MT1 (IC 50 value of 176.608 ± 91.164 μg mL −1 ; Fig. 11A), while the inhibition of MT2 was much more remarkable with IC 50 0.06 ± 0.01 μg mL −1 (Fig. 11B).The inhibition of furin was marginal at the highest concentration of 250 μg mL −1 ,and this enzyme retained over 70% of its initial proteolytic activity.Due to the close structural homology but distinct physiological functions, selective inhibitors of matriptases are of great interest However, further more detailed research is essential to clearly determine the selectivity of Venetin-1 towards MT1 and MT2, which is beyond the scope of this paper.Additionally, it was shown that the tested compound inhibited chymotrypsin (IC 50 1.26 ± 0.14 µg mL −1 ) and was inactive towards trypsin (Fig. 11C,D).
Discussion
Annelids inhabit soil rich in microorganisms and have developed a unique ability to survive in this environment; therefore they represent a group of animals that is intensively studied by biologists in order to find new molecules with potential therapeutic applications.These model invertebrates are crucial for elucidation of the mechanisms of important biological and developmental processes in organisms.Earthworms as a model organism are inexpensive and, importantly, ethically uncontroversial 63 .Currently, bearing in mind the lack of sufficient effectiveness and selective action of synthetic preparations, the pharmaceutical industry is returning to natural bioactive substances that may prove effective where modern medicine fails 64 .Venetin-1, previously characterized as a protein-polysaccharide fraction, also showed inhibitory activity against C. albicans ATCC 10231 and C. krusei ATCC 6258.Venetin-1 disrupted cell division and led to fungal cell death 49 .The preparation induced changes in the fungal cell wall structure and in the nanomechanical properties of the cell wall, as evidenced by atomic force microscopy (AFM) 50 .Upon incubation of C. albicans with Venetin-1, mitochondrial DNA migrated towards nuclear DNA.Both genetic materials combined into one nuclear structure, which was the beginning of the apoptosis process.It was observed that the analyzed compound affected the expression of oxidative stress proteins 51 .The protein part of the nanoparticle was shown to consist mainly of two lysenin proteins 50 .The presence of carbohydrate compounds in the preparation was confirmed as well.The analyzed compound did not show endotoxicity and cytotoxicity in relation to normal human skin fibroblasts 49 .
In previous analyses, Venetin-1 was shown to cause C. albicans cell death via necrosis and apoptosis 49,51 .Generally, there are three main types of cell death: apoptosis, necrosis, and autophagy.Apoptosis, or programmed cell death, is an active process that requires activation of many genes and energy expenditure.Necrosis, in contrast to apoptosis, is a passive and pathological process.This type of death occurs under the influence of physical, chemical, and biological factors after exceeding the threshold value of cell immunity.Autophagy is a mechanism referred to as type II programmed cell death 65 .In this process, proteins with a long half-life and other components of organelles are directed to vacuoles, where they are degraded 66,67 .In the present study, autophagy observed during the analysis of the action of Venetin-1.Autophagy, which is a process specific for eucaryotic cells, has three main forms: chaperone-mediated autophagy, microautophagy, and macroautophagy.Macroautophagy is the most common type and is activated to degrade damaged organelles and proteins.A small amount of cytoplasm containing damaged structures is enveloped by a double-layered vesicle creating an autophagosome and transported to the vacuole.The outer layer of the autophagosome fuses with the vacuolar membrane, and the inner layer with its contents passes inside the vacuole as an autophagosomal body, where it is degraded by vacuolar enzymes [68][69][70] .
The presented fluorescent microscopy and transmission electron microscopy analyses of vacuoles convincingly showed the process of autophagy, i.e. engulfment of vesicles after the incubation with Venetin-1.The C. albicans cells treated with Venetin-1 had significantly bigger vacuoles than the control cells.Those changes may have a source in the autophagy process, where every autophagosome adds its membrane to the vacuolar membrane.This additional membrane intake is not balanced by outflow and leads to vacuole enlargement and dilution of intramembrane contents [71][72][73][74] .
The process of autophagy usually involves leading the cell through unfavorable living conditions.Autophagy may be a strategy for cell survival in stress conditions such as starvation, hypoxia, or the use of chemotherapeutics.The mechanism of autophagy is activated to obtain an additional source of energy.Autophagy is an essential process for cell survival, as it removes damaged or abnormal organelles and molecules that can disrupt cell homeostasis.After the action of Venetin-1, oxidative stress was observed, which led to damage to intracellular structures.There are studies that prove that ROS (reactive oxygen species) play a regulatory role in the process of autophagy-the appearance of ROS activates this process.Activation of the autophagy process by ROS aims to minimize ROS-induced damage 75 .
Mitochondrial ROS can be beneficial or harmful to cells depending on their concentration and location 76 .Mitophagy, where inefficient mitochondria are degraded, is a type of selective autophagy.Mitochondria provide most of the energy required by the cell.Energy production by mitochondria is accompanied by generation of ROS.An increased amount of mitochondrial ROS damages cell components, including the mitochondria.This leads to various pathological changes and results in cell death 76 .An important process for the cell is the elimination of overactive or damaged mitochondria in order to maintain mitochondrial homeostasis.This removal must take place in a timely manner so as not to cause further cell damage 77,78 .
After the treatment of the C. albicans cells with Venetin-1, these overactive mitochondria were enlarged and showed stronger fluorescence in the microscopic image.In addition, the C. krusei strain was analyzed in this respect.It turned out that the effect was repeated and was additionally stronger than in the case of C. albicans, and the mitochondria were visible as large, elongated structures located at the edge of the cell.Enlarged mitochondria and increased production of ROS were observed in our previous studies 51,52 .There are data reporting that mitochondrial outer membrane proteins have been identified as mitophagy receptors in yeast 77 .The patch of vacuoles and mitochondria is an interesting and increasingly often discussed topic in the characterization of yeast biology.In C. albicans cells, the Mcp1 vacuole and mitochondrial patch protein (vCLAMP) is involved in maintaining mitochondrial function and mitophagy 77 .The observed morphological and proteomic changes in the C. albicans cells after the treatment with Venetin-1 51 suggest occurrence the mitophagy, which will be analyzed in further research.
SEM micrographs of C. albicans cells after treatment with the test preparation were presented in an earlier publication 51 .In those studies, we showed cell deformation and changes in the surface of the cell wall, which was irregular after the action of the analyzed compound.In the current studies, we have shown significant cell enlargement after the incubation with Venetin-1 in comparison to the control culture cells, which is related to autophagy.After the action of Venetin-1, numerous post-division scars on the surface of cells were observed.They were similar to those observed in SEM micrographs after exposure to fluconazole, a known drug that disrupts the structure of the cell membrane of C. albicans 79 .In our opinion, the cell dividing many times in a short time does not keep up with the production of building compounds required to synthesize the structures of the cell wall in place of scars.In addition, there are disturbances inside the cell, the inhibition of which is associated with the expenditure of cell energy in order to keep the cell alive.
Continuing the multidirectional characterization of the Venetin-1 nanoparticle, we decided to analyze the zeta potential for this complex.An important quantum of particles dispersed in the liquid is its stability, i.e. the ability of the particles to remain in the form of a colloidal dispersion.Dispersions that are unstable may undergo coagulation or sedimentation processes, which result in the delamination of the sample.The zeta potential is the potential occurring in the double layer at the surface of dispersed particles.It is a potential between the dispersant and the fluid layer attached to the surface of the particle.This parameter is used to determine the stability of colloidal systems.It is assumed that the dispersion is stable for an absolute value of zeta potentials >|± 30| V.These conditions were met by some of the analyzed samples.The zeta potential of Venetin-1 depends on the surface potential.The zeta potential was measured in the pH range from 3 to 7 in the 0.001 M NaCl electrolyte and in the pH range from 3 to 8 in 0.001 M KCl.The tested systems were unstable and aggregated, which was confirmed by the particle size measurements made using the Dynamic Light Scattering method.
The present study showed that Venetin-1 additionally exhibits inhibitory activity towards certain proteolytic enzymes, namely matriptase-1 (MT1), matriptase-2 (MT2), and chymotrypsin.Both matriptases are type II transmembrane serine proteases, have trypsin-like specificity, and share high structural similarity but differ in biological activity 80 .MT1 affects the formation and integrity of epithelial tissues and is implicated in various epithelial-derived cancers, including breast, prostate, and ovary tumors 81 .Noteworthy, its increased activity is often regarded as a predictive factor of poor cancer prognosis 82 .MT2 acts as the proteolytic regulator in human iron homeostasis.Mutations in its gene, TMPRSS6, correlate with iron-refractory iron deficiency anemia, while its up-regulated activity may lead to iron overload disorders 83 .The inhibitory activity of Venetin-1 against furin, i.e. a type I transmembrane serine protease involved in the cleavage of many inactive protein precursors in the constitutive secretory pathway, was insignificant 84 .
In addition, the analyses showed that Venetin-1 inhibited strongly chymotrypsin and was inactive towards trypsin.Both proteolytic enzymes are pancreatic serine proteases involved in food digestion displaying different substrate preferences.Trypsin specifically hydrolyzes peptide bonds after basic amino acid residues, such as Lys and Arg, while chymotrypsin cleaves peptide bonds formed by the carboxyl groups of either aromatic Tyr, Phe, and Trp or aliphatic Leu residues.
Inhibition of proteolytic activity is a very important molecular mechanism by which organisms prevent self-injury 85 and protect themselves against pathogens [86][87][88][89] or predators [90][91][92] .Of all types of protease inhibitors, serine protease inhibitors (SPIs) are the most common 93 .They are widespread in invertebrates, e.g. in kuruma shrimp (Marsupenaeus japonicas), black tiger shrimp (Penaeus monodon) 92 , sea anemones 94,95 , scorpions 93,95,96 , hookworms 97 , and numerous poisonous animals-spiders and snails 93,98,99 .SPIs with antifungal activity have been described in the silkworm (Bombyx mori) 100 .Studies reported by Lee 92,101 also showed SPIs produced by the oyster Crassostrea gigas having activity against the HIV-1 virus.On the other hand, it has been reported that HIV protease inhibitors have been shown to reduce C. albicans cell adherence by inhibiting yeast-secreted aspartic proteases 102,103 .The incidence of candidiasis in HIV-infected patients has been proven to be significantly reduced 104 .
The pharmaceutical industry offers several protease inhibitors, which are used in the treatment of human diseases, such as dabigatran used in the treatment of pulmonary embolism and angiotensin converting enzyme inhibitors (ACEI) used in the treatment of hypertension 92,105,106 .The American Association for the Study of Liver Diseases (AASLD) recommends therapy for genotype-1 chronic hepatitis C virus (HCV) based on serine protease inhibitors together with pegylated interferon α and ribavirin 107 .The proteasome inhibitor bortezomib having also antifungal activity 108,109 has been approved for clinical treatment of multiple myeloma 110 .It is similar to the Venetin-1 nanoparticle, which inhibits the 20S proteasome and shows antifungal activity at the same time.
Discovering protease inhibitors from the natural environment may have advantages over development of synthetic compounds with such properties There are known natural inhibitors whose mechanism of action is similar to that of Venetin-1, i.e. they are targeted at such intracellular organelles as mitochondria and nucleus and cause oxidative stress and cell death via apoptosis [111][112][113] .Natural protease inhibitors more easily overcome resistance mechanisms and are more stable and less toxic.These arguments support the advisability of searching for these compounds in nature 92 .
The antifungal effect of earthworm CF has been described mainly in relation to fungi attacking plants 8,22,114 .When CF is used against plant pathogens, the cytotoxicity of the fluid is irrelevant and it can be used in its raw form.However, to be used against fungal infections in humans, the CF preparation must be properly prepared so that it does not show cytotoxicity or endotoxicity.Because our research aims to use Venetin-1 as a chemotherapeutic drug, it had to undergo many laboratory processes to meet the required conditions.Taking into account the properties and multidirectional antifungal and anticancer effects of Venetin-1 and its complex structure, the structural connection of polysaccharides and lysenins seems to be an interesting research topic to be explored using techniques from various scientific disciplines.
In conclusion, the conducted analyses showed that the decrease in the survival of C. albicans cells corresponded to changes in the morphology of the cells and cellular organelles, such as vacuoles and mitochondria.Vacuole changes typical of autophagy and enlargement of mitochondria were observed.The zeta potential of the tested preparation was characterized.The aggregation ability and properties of the protease inhibitor type of the Venetin-1 nanoparticle have been proved.The compound, exhibiting inhibitory activities against certain
Figure 2 .
Figure 2. Flow cytometry analysis of viable and dead cells using mixture fluorochromes of Hoechst 33342 and Propidium iodide.The upper left quarter gates viable cells; the lower left quarter gates dead cells.
Figure 10 .
Figure 10.(A) Dependence of ζ potential of Venetin-1 in the pH function in 0.001 M and 0.0001 M NaCl solutions.(B) Dependence of ζ potential of Venetin-1 in the pH function in 0.001 M NaCl and 0.001 M KCl solutions. | 8,449 | 2023-08-30T00:00:00.000 | [
"Biology"
] |
Design and Implementation of the Detection Software of Wireless Microseismic Acquisition Station Based on Android Platform
New energy acquisition devices are urgently required to address the increasing global energy consumption and 10 increasing difficulty of energy exploitation. Devices for seismic exploration appear to be small in size, wireless and rapidly becoming more intelligent; hence, a traditional operating platform can no longer satisfy the demand of portable exploration device usage. This study investigates and develops hardware for a wireless microseismic acquisition station, then uses this hardware as a platform to address the distribution of wireless microseismic acquisition stations and deliver monitoring software based on the Android platform, which is portable, popular and has a large number of users. In large-scale field constructions, 15 software can provide operators with visualised station layouts throughout the process, including positioning, ranging, angle measuring and network monitoring. It also offers a real-time network for monitoring smalland medium-sized microseismic acquisition station arrays under construction as well as other functions, such as intelligent control and real-time data monitoring of the status of the acquisition station. A drainage blast monitoring test is conducted on the system, showing positively monitored data and accurate results in the inverse operation. Moreover, the software and hardware are proven to be highly 20 stable and portable through a post-construction test, which can help enhance the field construction efficiency.
such as fractures in mines. A microseism is defined as the strain energy released in the form of elastic waves in the stress redistribution process caused by structural break and elongation (Mulargia, F et al.,2016;Kurzon I et al.,2018;Pavlova A et al.,2014). Locating the seismic source with the microseismic data is key to the geometric characteristic analysis of the underground primary and induced fractures, as well as techniques related to effective stimulated reservoir volume (ESRV) and prediction of trend for future production (Thorne P W et al.,2017). In 1912, Geiger (Germany) proposed a positioning 5 method within the actual domain based on the ideal Earth (Geiger L et al.,1912) , which was later applied to the study of Meles et al.,2010). As the main source for seismic exploration, the seismic data acquisition stations play a critical role in resource exploration and energy acquisition (Mazza, S et al.,2012). Similarly, the microseismic data generation also relies 10 on the microseismic data acquisition stations. With the rapid development of science and technology, the evolution of seismic exploration sources tends to focus more on miniaturisation, intelligence and portability whilst maintaining high resolution, high precision, synchronous data acquisition and high-speed data transmission with low power consumption (Qisheng Zhang et al.,2013). From a broader perspective, in the future, seismic data will be collected by a network made up of many miniature seismic instruments in most areas (Willett K M et al.,2014). At a local scale, data will be acquired from a local area network 15 (LAN) consisting of small seismic instrument arrays in local construction areas. Nevertheless, the current development of seismic exploration sources has seen two bottlenecks in terms of the front-end sensor and the software management platform of a corresponding exploration device.
Computers are used as terminal devices in most of the existing software management platforms. Despite their high computing 20 speed and stable performance, computers are not suitable for many operators in terms of working in the field with a large area and a large distance, considering their poor portability, short life and complicated operations. In addition, most of the current Android software have been applied to seismic data display and rapid earthquake reporting(LIU Jun et al.,2014), whilst only a few have been directly used in the large-scale field construction sites.
25
Based on the research group has developed the hardware of the acquisition station. transmission. Furthermore, the position information of the Android device user can be quickly inputted to WLAN, thanks to its built-in GPS unit and the GSM/GPRS access module, providing more valid data for construction.
The Android software was designed for the hardware structure of the acquisition system, which monitored and controlled the operation and status of the acquisition system. Designed in the published article "Development of high-precision distributed wirelessmicroseismic acquisition stations", then a brief introduction to its structure.
Architecture of the acquisition system 5
The main tasks completed by the acquisition system included signal synchronisation, data acquisition, AD conversion and data storage. Fig. 1 shows that the vibrations generated from the operation, production and geological movements are firstly received by a geophone. Next, mechanical vibrations were converted into electrical signals, then inputted into the front-end processing circuit for front-end processing. After the analogue signals are processed by primary filtering and impedance matching in the front-end processing circuit, an amplifier will automatically adjust the amplification factor based on the input 10 signal amplitude. Amplified analogue signals will be further filtered by the filter circuit. Filtered single-ended signals are then converted into differential signals through a differential amplifier, which will then be inputted into an A/D (analogue-to-digital) circuit for conversion. Finally, analogue signals are converted into digital signals, and digital filtering is completed.
The main control unit of the acquisition system consisted of FPGA and CPU (integrated ARM and DSP dual-core processor). 15 FPGA is mainly used for completing multiple tasks that cannot be performed by general-purpose processors: firstly, it is used for allocating and executing any task related to time (e.g. generating the sampling clock, scheduling related strategies and realtime clock); and secondly, it collects the voltage, current, temperature and other information from sensors of the acquisition system and controls the data collection in the acquisition system and data buffering. Finally, it performs wired data communication, GPS information decoding and time synchronisation. The communication between the CPU core controller 20 and Wi-Fi is achieved through the extended USB of the system. The Android platform also uses that module to communicate with the acquisition system and control the upload and local storage backup (SD card) of data. Additionally, the interaction between the system and the external devices was also completed through core controllers, including LED indicators, on/off switch buttons and battery indications.
Construction of the operation network for the acquisition system 25
A poor outdoor environment together with remote station layouts in large areas can cause data transmitting problems.
Accordingly, constructing suitable supported wireless networks is vital in solving the problems and decreasing the difficulty and complexity of collecting field seismic data. Given that most of the construction sites for seismic acquisition are located in field areas with unstable networks, setting up a miniature WLAN in the array is more practical than applying the internet considering the actual construction. Setting up a miniature wireless WLAN in the array is a critical step. The WLAN should 30 have a high and stable transmission efficiency to minimise the negative influence of topography, and should be easily maintained and expanded. This study proposed the following WLAN construction plans based on the transmission layer interfaces provided by the existing distributed wireless microseismic acquisition stations as well as actual applications: The 2.4 GHz WLAN to provide access points and the 5.8 GHz WLAN for remote data transmission were adopted in this plan .
A single acquisition station was connected to a single 2.4 GHz omni-directional AP through the TCP/IP protocol by binding SSID, connecting the MAC address of the device and limiting the number of connected devices. The data were transmitted 5 from the network cable to the 5.8 GHz directional antenna, then to the hub station AP through directional transmission. Whilst the 2.4 GHz wireless signal was largely affected by atmospheric attenuation, the 5.8 GHz WLAN proved to have a higher data transmission rate and a greater signal radiation distance under the same transmitting power of the directional antenna. The WLAN constructed in this manner had a clear network topology, and can easily be maintained.
Android platform
Android is a free and open-source Linux-based operating system mainly used for mobile devices, such as smartphones and tablets. The Android system architecture has employed a layered architecture composed of four layers. The layers from top to bottom are the application layer, Android framework, Android runtime and Linux kernel. Android 9.0 is the latest version.
Software demand analysis 15
(1) The software should have visualised layouts (i.e. custom maps) to display the specific location of the arranged wireless microseismic acquisition stations on the map based on the field construction requirements. In addition, offline maps are required to ensure the normal use of the software when network connection is unavailable outdoors.
(2) During the station distribution, the relative distance and the angle between the acquisition and host stations as well as the communication status of the WLAN must be checked. 20 (3) The parameters and status of all acquisition stations in the current network should be configured and controlled through an Android mobile device upon the completion of the station distribution.
(4) After the acquisition stations start collecting data, the software should be able to monitor the status of the acquisition station, including the temperature, power, remaining storage space and the number of satellites connected to the GPS to help operators monitor the acquisition stations. 25 (5) The software should achieve real-time data collection from the acquisition stations, which can be displayed in multi-channel waveforms to observe the real-time acquisition.
Software architecture design
According to the demand analyses in Section 3.2, the Android software should be equipped with WLAN access and communication, control, monitoring, real-time position, waveform display, data storage and other functions of the multiple 30 distributed wireless microseismic acquisition stations in the WLAN. Additionally, it should be utilised to obtain the location and the network status of the Android mobile device where the software is installed. Fig.2 shows the designed software architecture interface.
The software design follows the MVP software design pattern, where MVP stands for model, view and presenter. The MVP 5 has evolved from the MVC pattern, which is short for model, view and controller. The presentation logic and the business logic are separated in the MVP to reduce coupling. In a traditional MVP architecture, the model layer is used for business logic and solid modelling. The view layer is an interface displaying an updated view to the user and receiving the data inputted by the user. The presenter layer is used to separate the model and view layers and reference through the interfaces of the view and the model. It also loads data and updates UI. Business logic is performed by the interfaces in the presenter layer. In the 10 traditional pattern, the system is packed with implementation classes and interfaces when processing simple requests. The view layer is rarely modified in the actual project. Many bugs of business logic must be fixed during programme upgrade and iteration. Therefore, in the MVP officially defined by ©Google, logic processing has been added to the presenter layer such that the model layer has less functions and only provides data models. In the MVP pattern, the implementation classes of the view layer, such as activity and fragment, are responsible for processing the life cycle of the presenter layer, avoiding the 15 storage leakage of activity. Activity does not process specific business logic. The interfaces in the presenter make it easier to conduct unit tests. The MVP pattern can generally offer a lower coupling in programming, a clearer software structure and improvement in code flexibility and maintainability.
A framework with a single activity and multiple fragments in the MVP pattern was adopted. Whilst the framework has a clear 20 logic with one activity on one page, the activity may put pressure on the loading of the mobile device. The system is also more likely to slow down and freeze because of frequent jumps, resulting in a poor immersive experience for users. On the contrary, the framework with a single activity and multiple fragments can reduce the amount of codes whilst ensuring a clear logic to decrease the difficulty in interacting among different interfaces. Moreover, the transition of pages when swiping smoothly can enhance the fluency of the software and immersive experience for users. 25 Fig. 3 shows the main process designed for the Android software. After connecting into the established WLAN, the connection and the communication of the Android mobile device will be conducted by accessing the static IP addresses of a single or multiple acquisition station(s) in the WLAN. Station-searching broadcast packets will be sent to the broadcast address in the WLAN through the UDP protocol in the software. Alternatively, the packets can be sent to a series of consecutive static IP 30 addresses in loops. In the defined protocol, the acquisition station(s) will send addresses to the broadcast packets to respond to the status packets and obtain the position and status of all acquisition stations in the WLAN. Finally, real-time data interaction and work status control will be completed by establishing the TCP/IP connections with all the acquisition stations.
Development environment for the Android programme
The Android software was written in Java with Android Studio. The API level for the Android SDK packages was 26, and the version strings in JDK 8 were 1.8.0. Meanwhile, BlackBerry Priv (Android 6.0) and Samsung S3 (Android 7.1) with a 3.0.1 base kernel were used for testing and running the software. 5 We solve these problems by replacing unsigned data types with high-precision data types and data types with the same 25 byte or shifting byte array in the process of converting the structure and the class. (4) Real-time data imaging technology: this software uses MPAndroidChart, which is a third-party open source charting library, to receive real-time imaging data. MPAndroidChart can display different types of images, such as polylines, scatter and histograms, with X-/Y-axis scaling and a customised outlook. The X-and Y-axis data sets will be generated once the real-5 time data are received from the blocking array. They will then be transferred to the line chart control unit in the MPAndroidChart to achieve data imaging in real time.
Specific functions of the Android software
The software can be installed on mobile phones with an Android 4.0 or above OS and can self-adapt to the screen size.
WIFIFragment will be loaded first and displayed on the first page. The main functions include online/offline custom maps, 10 layouts, control, status monitoring, data imaging and parameter settings of the acquisition stations. (4) Setting of the acquisition parameters: each channel gain of the target acquisition station must be set. The acquisition station will be controlled to start collecting data after setting the relevant parameters. The set parameters and command packet will be sent to the acquisition stations. Assuming that the acquisition parameters are set after the start of the data acquisition, 25 the setting will be dispatched before the next start of data acquisition.
(5) Real-time data imaging: the software can analyse the multi-channel data returned by real-time transmission from the target acquisition station and visualise the data in waveforms (Fig. 5). The channel can be selected on the waveform page. This test aimed to measure the blasting location and the range affected by the blasting in the previously designed acquisition stations. Sixteen microseismic acquisition systems were used in the test to conduct real-time data acquisition, data recording and post-monitoring analysis. The acquisition system was a 24-bit AD acquisition system with a sampling rate of 1 kHz and a dynamic range of 137 dB. After measurement and calculation, the acquisition accuracy of the acquisition station can reach 5 0.5μV.
According to the data collected by the collection station, the processor is used to process and detect the collected data, and the picked up vibration events are counted in a histogram. The horizontal axis of the microseismic time-frequency diagram is the monitoring time and the vertical axis is The number of micro earthquakes per unit time. As can be seen from the above figure, 10 in the time course, the microseismic distribution can be divided into three clusters. The frequency of microseismic rupture is low, indicating that there is a rupture process; the frequency of microseismic events during the process of fracture extension, sanding and shut-in Lower, fracturing construction is smooth. According to the data processing and calculation of the central station, the geophysical data can be used to invert the original 20 fracture trend map and three-dimensional fracture map of underground fractures (as shown in Fig. 9).
According to the experimental data analysis, the acquisition system can complete the exploration task very well under the Android software monitoring. The acquisition system can properly collect data and run stably. The positions of the seismic sources and the induced seisms were obtained through the experimental data inversion. Ultimately, the degree and the trend of 25 the fracture fission were calculated.
Conclusion
The distribution of stations and the status of data monitoring in the large-scale construction of field seismic data collection were studied herein. A microseismic acquisition system was designed for application to complex outdoor environments. Based on that, this study introduced software for the distribution and monitoring of distributed wireless microseismic acquisition 30 stations based on the Android platform. The software combined UDP and TCP/IP to achieve data communication and monitoring with the acquisition stations. This study also provided solutions to the incompatibility between different programming languages in the current communication protocol. Visualised station layout and real-time data imaging were accomplished using third-party libraries. All construction processes, including station layouts, data acquisition and monitoring of the acquisition stations, were precisely designed. After many outdoor experiments, the software was proven to be easy to use with a simple operation. It had a stable performance and exhibited smooth runs. The software can make the operators' lives easier and can also reduce the workload of data observers at the host station. Construction workers can collect construction 5 data in any area covered with WLAN by using only an Android mobile device, rather than monitoring the status of the acquisition station at a fixed position with a computer/laptop. In this way, the outdoor workload of an operator can be significantly reduced. These results indicate that the software has a great value for future applications and marketing promotions. The outdoor tests obtained positive data and accurate test results, as expected. This study has laid a foundation for the implementation of the acquisition system in future mining explorations. 10 | 4,243.4 | 2020-11-04T00:00:00.000 | [
"Computer Science"
] |
Performance Evaluation of 3D Descriptors Paired with Learned Keypoint Detectors †
: Matching surfaces is a challenging 3D Computer Vision problem typically addressed by local features. Although a plethora of 3D feature detectors and descriptors have been proposed in literature, it is quite difficult to identify the most effective detector-descriptor pair in a certain application. Yet, it has been shown in recent works that machine learning algorithms can be used to learn an effective 3D detector for any given 3D descriptor. In this paper, we present a performance evaluation of the detector-descriptor pairs obtained by learning a 3D detector for the most popular 3D descriptors. Purposely, we address experimental settings dealing with object recognition and surface registration. Our results show how pairing a learned detector to a learned descriptors like CGF leads to effective local features when pursuing object recognition (e.g., 0.45 recall at 0.8 precision on the UWA dataset), while there is not a clear performance gap between CGF and effective hand-crafted features like SHOT for surface registration (0.18 average precision for the former versus 0.16 for the latter).
Introduction
Surface matching is a fundamental task in 3D Computer Vision, key to the solution of major applications such as object recognition and surface registration. Nowadays, most surface matching methods follow a local paradigm based on establishing correspondences between 3D patches referred to as features. The typical feature-matching pipeline consists of: feature-detection to identify surface points surrounded by patches amenable to finding correspondences, usually referred to as keypoints or feature points; feature-description to encode the distinctive geometrical traits of a patch around a keypoint into a compact representation, referred to as descriptor, while filtering out nuisances like noise, viewpoint changes and point density variations; feature-matching to estblish feature correspondences by comparing descriptors computed around keypoints, usually by means of the Euclidean distance in the descriptor space.
Although over the last decades many 3D detectors and descriptors have been proposed in literature, it turns out rather unclear how to effectively combine these proposals. Indeed, unlike the related field of local image features, methods to either detect or describe 3D features have been designed and proposed separately, alongside with specific application settings and related datasets. This is also vouched by the main performance evaluation papers in the field, which address either repeatability of 3D detectors designed to highlight geometrically salient surface patches [1] or distinctiveness and robustness of popular 3D descriptors [2]. An attempt to investigate on the affinity between 3D detectors and descriptors can be found in [3], where the authors highlight how, depending on the considered application scenario, it may be possible to identify certain preferred detector-descriptor pairs.
More recently, however, [4,5] have proposed a machine learning approach that allows for learning an optimal 3D keypoint detector for any given 3D descriptor so as to maximize the end-to-end performance of the overall feature-matching pipeline. The authors show that this approach can handle effectively the diversity issues related to applications and datasets. Moreover, their object recognition experiments show that, with the considered descriptors (SHOT [6], SI [7], FPFH [8]), learning to detect specific keypoints leads to better performance than relying on existing general-purpose handcrafted detectors (ISS [9], Harris3D [10], NARF [11]).
By enabling an optimal detector to be learned for any descriptor, [5] sets forth a novel paradigm to bridge the gap between 3D detectors and descriptors. This opens up the question of which learned detector-descriptor pair may turn out most effective in the main application areas. This paper tries to answer this question by proposing an experimental evaluation of learned 3D features. In particular, we address object recognition and surface registration, and compare the performance attained by learning a paired feature detector for the most popular handcrafted 3D descriptors (SHOT [6], SI [7], FPFH [8], USC [12], RoPS [13]) as well as for a recently proposed descriptor based on a deep learning approach (CGF-32 [14]).
3D Local Feature Detectors and Descriptors
This section reviews popular methods for detection and description of 3D local features. Both tasks have been pursued through hand-crafted and learned approaches.
Hand-Crafted Feature Detectors
Keypoint detectors have traditionally been conceived to identify points that maximize a saliency function computed on a surrounding patch. The purpose of this function is to highlight those local geometries that turn out repeatedly identifiable in presence of nuisances such as noise, viewpoint changes, point density variations and clutter. Stateof-the-art proposals mainly differ for the adopted saliency function. Detectors operate in two steps: first, the saliency function is computed at each point on the surface, then non-maxima suppression allows for sifting out saliency peaks. Intrinsic Shape Signature (ISS) [9] computes the Eigenvalue Decomposition of the scatter-matrix of the points within the supporting patch in order to highlight local geometries exhibiting a prominent principal direction, Harris3D [10] extends the idea of [15] by deploying surface normals rather than image gradients to calculate the saliency (i.e., Cornerness) function. Normal Aligned Radial Feature (NARF) [11] was originally designed to operate on the range images but its usage has been extended also to point clouds. The algorithm first selects stable surface points, then highlights those stable points showing sufficient local variations. This leads to locate keypoints close to boundaries.
Learned Feature Detectors
Unlike previous work in the field, Salti et al. [4] proposed to learn a keypoint detector amenable to identify points likely to generate correct matches when encoded by the SHOT descriptor. In particular, the authors cast keypoint detection as a binary classification problem tackled by a Random Forest and show how to generate the training set as well as the feature representation deployed by the classifier. Later, Tonioni et al. [5] have demonstrated that this approach can be applied seamlessly and very effectively to other popular descriptors such as SI [7] and FPFH [8].
Hand-Crafted Feature Descriptors
Many hand-crafted feature descriptors represent the local surface by computing geometric measurements within the supporting patch and then accumulating values into histograms. Spin Images (SI) [7] relies on two coordinates to represent each point in the support: the radial coordinate, defined as the perpendicular distance to the line trough the surface normal, and the elevation coordinate, defined as the signed perpendicular distance to the tangent plane at the keypoint. The space formed by this two values is then discretized into a 2D histogram. In 3D Shape Context (3DSC) [16] the support is partitioned by a 3D spherical grid centered at the keypoint with the north pole aligned to the surface normal. A 3D histogram is built by counting up the weighted number of points falling into each spatial bin along the radial, azimuth and elevation dimensions. Unique Shape Context (USC) [12] extends 3DSC with the introduction of a unique and repeatable canonical reference frame borrowed from [6]. SHOT [6], alike, deploys both a unique and repeatable canonical reference frame as well as a 3D spherical grid to discretize the supporting patch into bins along the radial, azimuth and elevation axes. Then, the angles between the normal at the keypoint and those at the neighboring points within each bins are accumulated into local histograms. Rotational Projection Statistics (RoPS) [13] uses a canonical reference frame to rotate the neighboring points on the local surface. The descriptor is then constructed by rotationally projecting the 3D points onto 2D planes to generate three distribution matrices. Finally, a histogram encoding five statistics of distribution matrices is calculated. Fast Point Feature Histograms (FPFH) [8] operates in two steps. In the first, akin to [17], four features, refereed to as SPFH, are calculated using the Darboux frame and the surface normals between the keypoint and its neighbors. In the second step, the descriptor is obtained as the weighted sum between the SPFH of the keypoint and the SPFHs of the neighboring points.
Learned Feature Descriptors
The success of deep neural networks in so many challenging image recognition tasks has motivated research on learning representations from 3D data. One of the pioneering works toward this direction is 3D Match [18], where the authors deploy a siamese network trained on local volumetric patches to learn a local descriptor. The input to the network consists of a Truncated Signed Distance Function (TSDF) defined on a voxel grid. A similar approach is pursued in [19], where smoothed density values in a voxel grid are used instead of the TSDF. Along the lines of the end-to-end computational paradigm advocated by deep learning, PointNet [20] directly consumes point clouds as input. The authors show that the network can learn a global descriptor suited to tasks like part segmentation, classification and scene parsing. However, so far it has not been demonstrated whether and how the PointNet architecture may be deployed to learn a local descriptor amenable to the typical feature-matching pipeline. This is, in fact, exactly the goal pursued in [14], where the authors deploy a fully-connected deep neural network together with a state-of-the-art feature learning approach based on the triplet ranking loss [21,22] in order to learn a very compact 3D descriptor, referred to as CGF-32. Unlike PointNet, their approach does not rely on raw data but on the hand-crafted representation proposed in [16] canonicalized by the local reference frame presented in [6]. A multi-view approach can also be adopted: in [23] Li et al. integrate a differentiable renderer into a 2D neural network so to optimize a multi-view representation in order to learn a local feature descriptor. A parallel line of research, instead, tries to learn descriptors with unsupervised approaches and proposes to employ the latent codeword of an encoder-decoder architecture as a 3D feature descriptor. PPFFoldNet [24] and 3D-PointCapsNet [25] learn to reconstruct the 4-dimensional Point Pair Feature [26,27] of a local patch by a FoldingNet [28] decoder. Differently, Spezialetti et al. [29] reconstruct the 3D points of the input patch by means of a AtlasNet decoder [30] from an equivariant embedding computed by a Spherical CNN encoder [31].
Materials and Methods
In this section, we first briefly summarize the methodology proposed in [5] to learn a descriptor-specific keypoint detector, which is used as the keypoint detector in this evaluation due to its ability to learn effective keypoints for each decriptor, and then we review the evaluation methodology used to test descriptors for object recognition and surface registration.
Keypoint Learning
In order to carry out the performance evaluation proposed in this paper, for most local descriptors reviewed in Section 1.1 we did learn the corresponding optimal detector according to the keypoint learning methodology [5]. We provide here a brief overview of this methodology and refer the reader to [4,5] for a detailed description.
The idea behind keypoint learning is to learn to detect keypoints that can yield good correspondences when coupled with a given descriptor. To this end, keypoint detection is cast into binary classification, i.e., a point can either be good or not as concerns undergoing matching by the given descriptor, and a Random Forest [32] is used as classifier. The reasons to select Random Forests as classifier to perform keypoint detection as classification includes: they have been used to successfully solve several computer vision problems, and 3D keypoint detection in particular [33]; they are among the fastest classifiers as regards run-time prediction, especially when dealing with complex classification functions; Random Forest can be seamlessly extended to perform multi-class classification, which is exploited in [5] to generalize the KPL framework to handle multiple support sizes. The second trait is particularly relevant when using a classifier as a keypoint detector, as the prediction must be carried out at every point of the input cloud.
Learning the classifier requires defining the training set, i.e., both positive (good-tomatch) and negative (not-good-to-match) samples, as well as the feature representation. As for positive samples, the method tries to sift out those points that, when described by a chosen descriptor, can be matched correctly across different 2.5D views of a 3D object. Thus, starting from a set of 2.5D views {V i }, i = 1, . . . , N of an object from a 3D dataset, each point p ∈ V i in each view V i is embedded by the chosen descriptor, d = φ(p), where φ represents the descriptor algorithm. Then, for each view V i , a subset of overlapping views V i is selected based on an overlap threshold τ, A two-step positive samples selection is performed on V i and each overlapping view V j ∈ V i . In the first step, for each view V j ∈ V i a list of correspondences between descriptors D j i is created by searching for each descriptors d k = φ(p k ), ∀p k ∈ V i the nearest neighbor in the descriptor space between all descriptors d m = φ(p m ), ∀p m ∈ V j . Formally, A preliminary list of positive samples P j i for view V i is created by taking only those points that have been correctly matched in V j , i.e., the points belonging to the matched descriptors in the two views correspond to the same 3D point of the object according to threshold : where T i and T j denote the ground-truth transformations that, respectively, bring V i and V j into a canonical reference frame. The list is then filtered keeping only points corresponding to local minima of the descriptor distance within a radius nms . In the second step, the list of positive samples is further refined by keeping only the points in V i that can be matched correctly also in the others overlapping views V k,k =j ∈ V i that have not been used in the first step. Negative samples are extracted on each view, picking random point between those points which are not included in the positive set. A distance threshold neg is used to avoid a negative being too close to a positive and to balance the positive and negative sets. As far as the representation input to the classifier is concerned, the method relies on histograms of normal orientations inspired by SHOT [6]. However, to avoid computation of the local Reference Frame while still achieving rotation invariance, the spherical support is divided by considering only N r subdivisions along the radial dimension so as to compute a histogram for each spherical shell thus obtained. Hence, similarly to SHOT, the feature vector for a given point p is obtained by quantizing and accumulating into a histogram with N b bins the angle between the normal at p and those at the points within the spherical shells resulting from the N r radial subdivisions.
Evaluation Methodology
The performance evaluation proposed in this paper is aimed at comparing different learned detector-descriptor pairs while addressing two main application settings, namely object recognition and surface registration. In this section, we highlight the key traits and nuisances which characterize the two tasks, present performance evaluation metrics used in the experiments and, finally, provide the relevant implementation details.
Object Recognition
In typical object recognition settings one wishes to recognize a set of given 3D models into scenes acquired from an unknown vantage point and featuring an unknown arrangement of such models. Peculiar nuisances in this scenario are occlusions and, possibly, clutter, as objects not belonging to the model gallery may be present in the scenes.
According to the protocol adopted in [4,5], to evaluate the effectiveness of the considered learned detector-descriptor pairs we rely on descriptor matching experiments. An overview of how descriptor matching is performed in our object recognition experiments is depicted in Figure 1. Specifically, for both datasets, we run keypoint detection on synthetically rendered views of all models. Then, we compute and store into a single kd-tree all the corresponding descriptors. Keypoints are detected and described also in the set of scenes provided with the dataset, {S j }, j = 1, . . . , S. Eventually, a correspondence is established for each scene descriptor by finding the nearest neighbor descriptor within the models kd-tree and thresholding the distance between descriptors to accept a match as valid. Correct correspondences can be identified based on knowledge of the ground-truth transformations which bring views and scenes into the canonical reference frames linked to models by checking whether the matched keypoints lay within a certain distance . Indeed, denoted as (k j , k n,m ) a correspondence between a keypoint k j detected in scene S j and a keypoint k n,m detected in the n-th view of model m, as T j,m the transformation from S j to model m, as T n,m the transformation from the n-th view and the canonical reference frame of model m, the set of correct correspondences for scene S j is given by: This allows for calculating True Positive and False Positive matches for each scene and, by averaging across scenes, for each of the considered datasets. Like in [4,5], the final results for each dataset are provided as Recall vs. 1-Precision curves, with curves obtained by varying the threshold on the distance between descriptors.
Surface Registration
The goal of surface registration is to align into a common 3D reference frame several partial views (usually referred to as scans) of a 3D object obtained by a certain optical sensor. This is achieved through rather complex procedures that, however, typically rely on a key initial step, referred to as Pairwise Registration, aimed at estimating the rigid motion between any two views by a feature-matching pipeline. Thus, in surface registration, 3D feature detection, description and matching are instrumental to attain an as good as possible set of pairwise alignments between the views which then undergoes further processing to get the final global alignment. Differently from object recognition scenarios, the main nuisances deal with missing regions, self-occlusions, limited overlap area between views and point density variations. An overview of how pair-wise correspondences are attained in a typical surface registration pipeline is depicted in Figure 2. For each pair, we run keypoint detection on both views. Due the partial overlap between the views, a keypoint belonging to V i may have no correspondence in V j . Hence, denoted as T i and T j the ground-truth transformations that, respectively, bring V i and V j into a canonical reference frame, we can compute the set O i,j that contains the keypoints in V i that have a corresponding point in V j . In particular, given a keypoint k i ∈ V i : where N (T i k i , T j V j ) denotes the nearest neighbor of T i k i in the transformed view T j V j . If the number of points in O i,j is less than 20% of the keypoints in V i , the pair (V i , V j ) is not considered in the evaluation experiments due to insufficient overlap. Conversely, for all the view pairs (V i , V j ) exhibiting sufficient overlap, a list of correspondences between all the keypoints detected in V i and all the keypoints extracted from V j is established by finding the nearest neighbor in the descriptor space via kd-tree matching. Then, given a pair of matched keypoints (k i , k j ), k i ∈ V i , k j ∈ V j , the set of correct correspondences, C i,j , can be identified based on the available ground-truth transformations by checking whether the matched keypoints lay within a certain distance in the canonical reference frame: Then, the precision of the matching process can be computed as a function of the distance threshold [14]: The precision score associated with any given model is obtained by averaging across all view pairs. We also average across all test models so as to get the final score associated the Laser Scanner dataset.
Eventually, we point out that rather than showing the precision as a function of the distance threshold , we provide the score for a fixed and tight distance threshold value, the same value used to establish upon correctness of matches in object recognition experiments. Indeed, as highlighted in [14], the truly meaningful precision score is that dealing with a distance threshold tight enough to sift out useful correspondences in the addressed scenario. According to our experience, the adopted common threshold for the object recognition and surface registration experiments fulfills this requirement in both settings. Moreover, we think that having the same threshold value in both experiments renders the results more easily comparable.
Datasets
Akin to [5], in our experiments on object recognition we rely on the following popular datasets (see also Figure 3): • UWA dataset, introduced by Mian et al. [34]. This dataset consists of 4 full 3D models and 50 scenes wherein models significantly occlude each other. To create some clutter, scenes contain also an object which is not included in the model gallery. As scenes are scanned by a Minolta Vivid 910 scanner, they are corrupted by real sensor noise. • Random Views dataset, based on the Stanford 3D scanning repository (3 http:// graphics.stanford.edu/data/3Dscanrep/ accessed on 14 November 2020) and originally proposed in [1]. This dataset comprises 6 full 3D models and 36 scenes obtained by synthetic renderings of random model arrangements. Scenes feature occlusions but no clutter. Moreover, scenes are corrupted by different levels of synthetic noise.
In the experiments we consider scenes with Gaussian noise equal to σ = 0.1 mesh resolution units. As far as surface registration is concerned, we use the Laser Scanner dataset, proposed in [14]. This dataset includes 8 public-domain 3D models, i.e., 3 taken from the AIM@SHAPE repository (Bimba, Dancing Children and Chinese Dragon), 4 from the Stanford 3D Scanning Repository (Armadillo, Buddha, Bunny, Stanford Dragon) and Berkeley Angel [35]. According to the protocol described in [14], training should be carried out based on synthetic views generated from Berkeley Angel, Bimba, Bunny and Chinese Dragon, while the test data consists of the the real scans available for the remaining 3 models (Armadillo, Buddha and Stanford Dragon). Figure 4 reports exemplar scans from the Laser Scanner test set. We follow the split into train and test objects proposed by the authors.
Implementation
For all handcrafted descriptors considered in our evaluation, i.e., FPFH, ROPS, SHOT, SI, and USC, we use the implementation available in PCL [36], which relies on the original parameter settings proposed by the authors. As for the learned descriptor, i.e., CGF-32, we use the public implementation made available by the authors [14]. As for the Keypoint Learning (KPL) framework described in Section 2.1, we use the publicly available original code for the generation of the training set (http: //github.com/CVLAB-Unibo/Keypoint-Learning accessed on 14 November 2020.) together with the OpenCV (opencv.org) implementation of Random Forest classifier. We also keep all the KPL hyperparameters to the values proposed by the authors in [5]. Accordingly, each Forest consists of 100 trees of maximum depth equal to 25 while the minimum number of samples to stop splitting a node is 1. During the detection phase, each point of an unseen point cloud is passed through the Random Forest classifier which produces a score. A point is identified as a keypoint if it exhibits a local maximum of the scores in a neighborhood of radius r nms and the score is higher than a threshold s min .
For each descriptor considered in our evaluation, we train its paired detector according to the KPL framework. As a result, we obtain six detector-descriptor pairs, referred to from now on as KPL-CGF32, KPL-FPFH, KPL-ROPS, KPL-SHOT, KPL-SI, KPL-USC.
In object recognition experiments, the training data for all detectors are generated from the 4 full 3D models present in the UWA dataset. Purposely, according to the KPL methodology [4,5], for each model we render views from the nodes of an icosahedron centered at the centroid. Some of the generated views are presented in the Figure 5. Then, the detectors are used in the scenes of the UWA dataset as well as in those of the Random Views dataset. Thus, similarly to [4,5], we do not retrain the detectors on Random Views in order to test the ability of the considered detector-descriptor pairs to generalize well to unseen models in object recognition settings. A coherent approach was pursued for the CGF-32 descriptor. As the authors do not provide a model trained on the UWA dataset, we trained the descriptor on the synthetically rendered views of the 4 UWA models using the code provided by the authors and following the protocol described in the paper in order to generate the data needed by their learning framework based on the triplet ranking loss. Thus, KPL-CGF32 was trained on UWA models and, like all other detector-descriptor pairs, tested on both UWA and Random Views scenes.
In surface registration experiments we proceed according to the protocol proposed in [14]. Hence, detectors are trained with rendered views of the train models provided within the Laser Scanner dataset (Angel, Bimba, Bunny, Chinese Dragon) and tested on the real scans of the test models (Armadillo, Buddha, Stanford Dragon). As CGF-32 was trained exactly on the above mentioned train models [14], to carry out surface registration experiments we did not retrain the descriptor but used the trained network published by the authors (https://github.com/marckhoury/CGF accessed on 14 November 2020).
The values of the main parameters of the detector-descriptor pairs used in the experiments are summarized in Tables 1 and 2. Accordingly to [5], we used the same support size for all descriptors r desc and all detectors r det across all the datasets. As it can be observed from Table 1, train parameters for Random Views dataset are not specified as we did not train a new KPL detector on this dataset.
As regards the parameters used in surface registration, since models belong to different repository, we report parameters grouped by model. Test parameters for Angel, Bimba, Bunny and Chinese Dragon are not reported as they are only used in train. Similarly, we omit train parameters for Armadillo, Buddha and Stanford Dragon. Due to the different shape of the models in the dataset, τ is tuned during the train stage so that the number of overlapping views remains constant across all models.
Results and Discussion
In this section, we report experimental results and discuss them. Compared to the most recent evaluation in the field [2], our results allow to assess the performance of the recently proposed KPL methodology based on machine learning as a keypoint detector (which is not considered in [2]) when paired with descriptors or when tested on datasets not considered in [5]. Moreover, we also include a learned descriptor (CGF) in the pool of compared methods, while [2] only consider hand-crafted methods. The main limitation of our study is the smaller number of datasets and acquisition modalities with respect to [2], although on the other hand we rely on real partial scans for the surface registration experiment, while only one synthetic dataset, obtained by creating virtual views of the UWA dataset model gallery, is used in [2].
Object Recognition
The results on the UWA dataset are shown in Figure 6. First, we wish to highlight how the features based on descriptors which encode just the spatial densities of points around a keypoint outperform those relying on higher order geometrical attributes (such as, e.g., normals). Indeed, KPL-CGF32, KPL-USC and KPL-SI yield significantly better results than KPL-SHOT and KPL-FPFH. These results are coherent with the findings and analysis reported in the performance evaluation by Guo et al. [2], which pointed out the former feature category being more robust to clutter and sensor noise. It is also worth observing how the representation based on the spatial tessellation and point density measurements proposed in [16] together with the local reference frame proposed in [6] turn out particularly amenable to object recognition, as it is actually deployed by both features yielding neatly the best performance, namely KPL-CGF32 and KPL-USC. Yet, learning a dataset-specific non-linear mapping by a deep neural network on top of this good representation does improve performance quite a lot, as vouched by KPL-CGF32 outperforming KPL-USC by a large margin. Indeed, the results obtained in this paper by learning both a dataset-specific descriptor as well as its paired optional detector, i.e., the features referred to as KPL-CGF32, turn out significantly superior to those previously published on UWA object recognition dataset (see [4,5]). In [5], the results achieved on Random Views by the detectors trained on UWA prove the ability of the KPL methodology to learn to detect general rather than dataset-specific local shapes amenable to provide good matches alongside with the paired descriptor, and even more effectively, in fact, than the shapes found by handcrafted detectors. Thus, when comparing the different descriptors, we can assume here that descriptors are computed on the best available patches and results highlight the ability to handle the specific nuisances of the Random Views dataset. As shown in Figure 7, KPL-FPFH and KPL-SHOT outperform KPL-USC, KPL-CGF32 and KPL-SI. Again, this is coherent with previous findings reported in literature (see [2,5]), which show how descriptors based on higher order geometrical attributes turn out more effective on Random Views due to the lack of clutter and real sensor noise. As for KPL-CGF32, although its performances are still overall better than those of the other descriptors based on point densities, we observe quite a remarkable performance drop compared to the UWA dataset, much larger, indeed, than the feature sharing the very same input representation, i.e., KPL-USC. This suggests that the non-linear mapping learnt by KPL-CGF32 is highly optimized to tell apart the features belonging to the objects present in the training dataset (i.e., UWA) but turn out quite less effective when applied to unseen features, like those found on the objects belonging to Random Views. This domain shift issue is quite common to learned representations and may yield to less stable performance across different datasets than handcrafted representations.
Surface Registration
First, we deem it worth pointing out how, unlike object recognition settings, in surface registration it is never possible to train any supervised machine learning operator, either detector or descriptor, on the very same objects that would then be processed at test time. Indeed, should one be given either a full 3D model or a set of scans endowed with groundtruth transformations, as required to train 3D feature detectors (i.e., KPL) or descriptors (e.g., CGF-32), there would be no need to carry out any registration for that object. Surface registration is about stitching together several scans of a new object than one wishes to acquire as a full 3D model. As such, any supervised learning machinery is inherently prone to the domain shift issue.
As shown in Figure 8, when averaging across all test objects, the detector-descriptor pair based on the learned descriptor CGF-32 provides the best performance, which confirms the findings reported in [14] where the authors introduce CGF-32 and prove its good registration performance on Laser Scanner. It is worth highlighting here that, as we evaluate detector-descriptor pairs, our results are determined by both the repeatability of the detector as well as the effectiveness of the descriptor, whilst the registration experiments in [14], according to the established descriptor evaluation methodology (e.g., [2,6]) assume an ideal feature detector, i.e., the feature points are mapped from one view into another by the available ground-truth transformation. This different experimental settings explain the difference between the precision values in Figure 8 and those reported in [14]. Figure 8 shows also that, compared to the features based on handcrafted descriptors, and in particular KPL-SHOT, the gain provided by KPL-CGF32 is not as substantial as we did found in object recognition experiments (see Figure 6). Moreover, looking at the results on the individual test objects, reported in Figures 9-11, one may observe how KPL-CGF32 outperforms the feature pairs based on handcrafted descriptors only on Buddha, with KPL-SHOT yielding better results on Stanford Dragon and both KPL-SHOT as well as KPL-USC comparing favourably to KPL-CGF32 on Armadillo. This last experiment is particularly interesting due to CGF-32 consisting in a non-linear mapping learnt on top of the input representation deployed by USC. Unlike in object recognition, though, this mapping can only be learnt on different objects than those seen at test time and, hence, because of the domain shift, may not always turn out beneficial.
Conclusions and Future Work
Object recognition settings turn out quite amenable to deploy learned 3D features. Indeed, one can train upon a set of 3D objects available beforehand, e.g., due to scanning by some sensor or as CAD models, and then seek to recognize them into scenes featuring occlusions and clutter. These settings allow for learning a highly specialized descriptor alongside its optimal paired detector so to achieve excellent performance. In particular, the learned pair referred to in this paper as KPL-CGF32 sets the new state of the art on the UWA benchmark dataset, being able to achieve 0.45 recall at 0.8 precision. Although the learned representation may not exhibit comparable performance when transferred to unseen objects, in object recognition it is always possible to retrain on the objects at hand to improve performance. An open question left to future work concerns whether the input parametrization deployed by CGF-32 may enable to learn an highly effective non-linear mapping also in datasets characterized by different nuisances (e.g., Laser Scanner) or one should better try to learn 3D representations directly from raw data, as vouched by the success of deep learning from image recognition. Features based on learned representations, such as KPL-CGF32, are quite effective also in surface registration, where it delivers 0.18 precision on the Laser Scanner dataset, the highest value attained in our study, although this scenario is inherently more prone to the domain shift issue such that features based on handcrafted descriptors, like in particular KPL-SHOT and KPL-USC, turn out still very competitive and delivers similar performance (0.16 precision for the former on the same dataset).
We believe that these findings may pave the way for further research on the recent field of learned 3D representations, in particular in order to foster addressing domain adaptation issues, a topic investigated more and more intensively in nowadays deep learning literature concerned with image recognition (see e.g., [37]). Indeed, 3D data are remarkably diverse in nature due to the variety of sensing principles and related technologies and we witness a lack of large training datasets, e.g., at a scale somehow comparable to ImageNet, that may allow learning representations from a rich and varied corpus of 3D models. Therefore, how to effectively transfer learned representations to new scenarios seems a key issue to the success of machine/deep learning in the most challenging 3D Computer Vision tasks.
Finally, KPL has established a new framework whereby one can learn an optimal detector for any given descriptor. In this paper we have shown how applying KPL to a learned representation (CGF-32) leads to particularly effective features (KPL-CGF32), in particular when pursuing object recognition. Yet, according to the KPL training protocol, the descriptor (e.g., CGF-32) has to be learnt before its paired detector: one might be willing to investigate on whether and how a single end-to-end paradigm may allow learning both components jointly so as to further improve performance. | 8,201.2 | 2021-05-11T00:00:00.000 | [
"Computer Science"
] |
Probabilistic Prognostic Estimates of Survival in Metastatic Cancer Patients (PPES-Met) Utilizing Free-Text Clinical Narratives
We propose a deep learning model - Probabilistic Prognostic Estimates of Survival in Metastatic Cancer Patients (PPES-Met) for estimating short-term life expectancy (>3 months) of the patients by analyzing free-text clinical notes in the electronic medical record, while maintaining the temporal visit sequence. In a single framework, we integrated semantic data mapping and neural embedding technique to produce a text processing method that extracts relevant information from heterogeneous types of clinical notes in an unsupervised manner, and we designed a recurrent neural network to model the temporal dependency of the patient visits. The model was trained on a large dataset (10,293 patients) and validated on a separated dataset (1818 patients). Our method achieved an area under the ROC curve (AUC) of 0.89. To provide explain-ability, we developed an interactive graphical tool that may improve physician understanding of the basis for the model’s predictions. The high accuracy and explain-ability of the PPES-Met model may enable our model to be used as a decision support tool to personalize metastatic cancer treatment and provide valuable assistance to the physicians.
Introduction
Scientific review (1) and clinical studies (2) have indicated that physicians are overly-optimistic during survival prediction of patients with terminal cancer.This leads to overutilization of aggressive medical interventions and protracted radiation treatment, which increase side effects and health care bills, while other patients who may benefit for continued therapy are likely under-treated.The electronic medical records systems serving major medical centers store a wealth of potentially valuable clinical notes acquired over time ("longitudinal data") that contains rich information on performance status, imaging findings, and tolerance to systemic therapy.These data contain key covariates for informing diagnostic decisions, and may also serve as critical resources for survival estimation.However, the massive explosion of the medical data outstrips the manual ability to comprehend the entire source of information.On the contrary, a machine learning model can be trained on the large amount of heterogenous data which allows seamless integration between multi-source information, and, as a result, it may outperform physicians in estimating life expectancy.Using the large Stanford Cancer Institute Research Database (SCIRDB), we created a dynamic sequence-dependent deep learning model -Probabilistic Prognostic Estimates of Survival in Metastatic Cancer Patients (PPES-Met).The PPES-Met model takes as input a sequence of longitudinal free-text clinical visit narratives ordered according to the date of visits, and computes as output a probability of short-term life expectancy (> 3 months) for each visit.The complexity of the model mainly lies in extracting relevant information from the heterogeneous types of free-text clinical notes along with modeling temporal irregularity of the visits.
Methods
Patient population: With the approval of Institutional Review Board (IRB), we created a database that includes adult cancer patients (13,523) seen at the Stanford Cancer Center from 2008-2017 and diagnosed with distant metastases.This database contains various types of free-text patient visit notes (e.g.radiology reports, oncologist notes, discharge summary) from date of metastatic cancer diagnosis to death.A separate "Palliative radiation dataset" was created using patients (899) enrolled from 2015-2016 in a prospective survey study conducted in our institution's Radiation Oncology department.The overall group of patients were seen for 471,005 daily encounters/visits, including outpatient and inpatient contact.For these visits, median follow-up was 12.7 months.Median overall survival was 22.4 months.Patients were hospitalized for 115,716 (24.6%) visits.There were 1,403,544 provider notes.The training set contains 10,239 patients with 380,080 visits, validation set of 1,785 patients and test set of 1,818 patients (15%) with 90,925 visits.Proposed System -PPES-Met: The model is composed of two core processing blocks: i. Semantic Word Embedding (SWE): We adopted a completely unsupervised hybrid methodan updated version of Intelligent Word Embedding (IWE) method(3) that combines semantic-dictionary mapping, neural embedding, and context-based windowing technique for creating dense vector representation of free-text clinical narratives.The method leverages the benefits of unsupervised learning along with expert-knowledge to tackle the major challenges of information extraction of informative information from clinical texts, while accounting for ambiguity of free text narrative style, lexical variations, arbitrary ordering of words, and frequent appearance of abbreviations and acronyms.
ii. Stacked RNN model: On top of the word embeddings obtained from patient visit notes, we designed a many-tomany RNN model using two-layer one directional stacked stateful Long short-term memory (LSTM) units for learning survival across the sequence of clinical narratives.The model takes as input a series of vectorized patient visit notes ordered according to the timestamp of visits, and it predicts probability of survival at each patient visit.In the stacked RNN layers, the first layer's one directional LSTM block receives the input and previous hidden state ℎ −1 , and pass the current state ℎ to the successive LSTM block and to the corresponding block in the upper layer.LSTM blocks also maintains an H-dimensional cell state −1 ∈ and the second layer units have modeled to maintain the recurrent connections in multiple dimensions.The final output of double layer stacked-RNN is modeled as: ̂() = ℎ (ℎ , ̂−1 () ) where ℎ is the hidden state of the first level and ̂() is the predicted survival at time − 1.The time distributed weighted cross entropy loss is used as optimization function during model training.By using stateful LSTM, the model includes long-term dependencies that exist in the longitudinal data, which is generally very informative for the prediction task.
Results
Quantitative evaluation: We used 3-month survival record defined at each time point as categorical class labels for probabilistic prediction.The overall prognosis estimation accuracy on the test dataset was validated as Precision-Recall curve and AUC score was 0.97.The high AUC scores show that the model is predicting survival accurately with high precision, as well as returning most of positive results (high recall).To check the calibration with the ground truth, we also measure the Brier score and the low brier-score (0.069) shows the prediction was highly calibrated with the ground truth.
Figure 1 Intelligible longitudinal survival curve of a patient
Qualitative evaluation: To provide a formalized mechanism to reason about computerized model's predictions at a specific timepoint, we implemented an interactive graphical interface that generates a longitudinal probabilistic summary (Figure 1) for each patient.Clicking on a time point, the system will retrieve not only the visit type, but it will also exploit the controlled-terms and will extract the core findings of the visit by highlighting the context of the controlled-terms (see Figure 1).This intuitive illustration may help the clinician to reason on the PPE-Met prediction and perform a qualitative error-analysis.
Discussion
The objective of our work is to improve physicians' knowledge of their patients' prognosis to help tailor treatment strategy, improve quality of life, as well as reduce costs.Early stage results of an ongoing prospective study conducted by Stanford Radiation Oncology department which includes 899 patients enrolled in the palliative radiation study found that the physicians are not able to accurately estimate life expectancy.We tested our PPES-Met model on a combination of a general group of metastatic patients and data from a palliative radiation study, and the probabilistic prediction accuracy was 0.97 AUC-PR-curve.Initial evaluation suggests that the PPES-Met prediction model produces good accuracy.This is probably due to the PPES-Met model's capability of integrating a large amount of patient-specific facts while preserving long-term dependencies in the data, which is not a trivial task for human experts.The high accuracy and the ability of our visualization the PPES-Met model output to show the critical data driving its predictions suggests that the model might ultimately be clinically useful in the future as a decision support tool to personalize metastatic cancer treatment and to aid physicians.The core limitations of the current work are: (i) many patients in the training dataset lost follow-up which may create an inaccurate assumption about survival during the model training.However, we consulted the central cancer registry to validate the survival; (ii) the time points are not equally spaced, and a single day may contribute multiple data points which may affect the temporal dependency and introduce fluctuation in the survival curve.In future, an NLP technique can be applied to combine the visit data from the same day for date-based survival analysis. | 1,876 | 2018-01-09T00:00:00.000 | [
"Medicine",
"Computer Science"
] |
Electrical Stimulation Modulates High γ Activity and Human Memory Performance
Visual Abstract
Introduction
Studies of direct electrical stimulation of the human brain were pioneered in epilepsy patients undergoing surgery to treat drug resistant focal epilepsy. During the surgery when patients were awake and stimulated in specific areas of the neocortex, they reported conscious experience of past events (Penfield, 1958). This phenomenal effect of invoking declarative memory representations was more likely to occur when stimulating in a discrete range of spectral and temporal parameters, which led to a hypothesis that the electrical current that was passed through the neural tissue activated specific neurophysiological activity supporting memory (Bickford et al., 1958;Penfield and Perot, 1963). In the current study, free recall tasks were used to investigate how stimulation in specific brain regions modulated the electrophysiological activities induced by word presentation and their subsequent recall.
Recent attempts at human memory enhancement have primarily focused on the hippocampus (HP) and the associated mesial temporal lobe structures, with reports of positive outcomes described in small studies of individual brain regions (Suthana and Fried, 2014;Kim et al., 2016). In general, however, studies have shown inconsistent results for stimulation in mesial temporal lobe structures, including: HP (Coleshill et al., 2004;Suthana et al., 2012;Fell et al., 2013;Jacobs et al., 2016), entorhinal cortex (Suthana et al., 2012;Fell et al., 2013;Jacobs et al., 2016), and fornix (Hamani et al., 2008;Miller et al., 2015). The effect of stimulation on the neurophysiological activity associated with memory tasks was largely unexplored. The positive effects of stimulation on memory reported in some of these studies were observed either in a single case (Hamani et al., 2008) or at the level of a group of patients (Suthana et al., 2012;Miller et al., 2015) without a detailed analysis of the electrophysiological signals, which is often challenging because of the stimulation artifacts (Johnson et al., 2013). In summary, limitations in the sample size, number of brain regions tested, and analysis have impeded our understanding of the impact of direct human brain stimulation on memory processes.
␥ activities in the local field potential present one plausible target for exploring the neurophysiology of memory processes and the effect of stimulation. These activities have been associated with cognitive functions, including perception, attention and memory (Singer, 1993;Tallon-Baudry and Bertrand, 1999;Fries, 2009;Düzel et al., 2010). ␥ activities in the high frequency ranges were proposed to be generated by local neuronal assemblies underlying cognitive processing during task performance (Crone et al., 2006;Lachaux et al., 2012), and thus provide a potential biomarker for mapping brain functions. Recent studies of ␥ activity in humans and nonhuman primates showed discrete bursts of ␥ power induced by memorized stimuli (Kucewicz et al., 2014;Lundqvist et al., 2016). In these studies, the rate of high ␥ burst events was associated with memory performance and proposed to underlie the differences in average power induced between trials with remembered and forgotten items, i.e., the subsequent memory effect Sederberg et al., 2007). Although the physiologic source of ␥ activities, local field oscillations or firing of neuronal assemblies, and their role in cognitive function are actively debated (Crone et al., 2006;Waldert et al., 2013;Kucewicz et al., 2017), they may still be useful as a measure of neuronal processing and modulation.
There is growing evidence that ␥ activities can be modulated by external interventions. Optogenetic stimulation of distinct neuron types was shown to increase ␥ power in the local field potential and enhance neuronal network performance in rodents (Sohal, 2016). ␥ power can also be increased through neurofeedback training in specific brain regions, as reported in nonhuman primate recordings that showed synchronous neuronal firing and enhanced behavioral performance (Engelhard et al., 2013). Transcranial current stimulation is another approach used to modulate ␥ activities and, for instance, was shown to induce dream self-awareness in the human subjects (Voss et al., 2014). The effect of direct stimulation of the human brain on ␥ activities linked to memory performance has been largely unexplored. However, the reports of a positive effect on memory performance in humans were all stimulating at frequencies in the ␥ range (40/50/200 Hz; for review, see Kim et al., 2016), suggesting that the applied current presumably modulated similar frequencies of neuronal oscillations. Here, we tested the effect of 50-Hz electrical stimulation on ␥ activity and task performance in four brain regions supporting declarative memory.
Study participants
Patients undergoing intracranial electroencephalographic monitoring as part of their clinical treatment for drugresistant epilepsy were recruited to participate in this multi-center collaborative study. Data were collected from the following clinical centers: (Mayo Clinic, Thomas Jefferson University Hospital, Hospital of the University of Pennsylvania, Dartmouth-Hitchcock Medical Center, Emory University Hospital, University of Texas Southwestern Medical Center). The research protocol was approved by the respective IRB at each clinical center and informed consent was obtained from each participant. Electrophysiological data were collected from standard clinical subdural and penetrating depth electrodes (AdTech Inc., PMT Inc.) implanted on the cortical surface and into the brain parenchyma, respectively. The subdural electrode contacts were arranged either in a grid or a strip configuration with contacts separated by 10mm. The depth electrode contacts were separated by 1.5-10 mm spacing. In each case, the placement of the electrodes was determined by a clinical team whose sole purpose was to localize seizures for possible epilepsy surgery. In this study, we identified 22 patients (nine males) with subdural or depth electrodes implanted in at least one of the four brain regions of the cortical-hippocampal declarative memory system (Eichenbaum, 2000), who completed at least two stimulation sessions in any of these regions (Tables 1, 2).
Anatomic localization and brain surface mapping
Cortical surface parcellations were generated for each participant from preimplant magnetic resonance imaging (MRI) scans (volumetric T1-weighted sequences) using Freesurfer software (RRID: SCR_001847;Fischl et al., 2004). The HP and surrounding cortical regions were delineated separately based on an additional 2-mm-thick coronal T2-weighted scan using the Automatic Segmentation of Hippocampal Subfields (ASHS) multi-atlas segmentation method (Yushkevich et al., 2015). Electrode contact coordinates derived from registered postimplant CT scans were then mapped to the preimplant MRI scans to determine their anatomic locations. For subdural strips and grids the electrode contacts were additionally projected to the cortical surface using an energy minimization algorithm to account for postoperative brain shift (Dykstra et al., 2012). For comparisons across subjects, coordinates were transformed to the MNI brain space, in which distance between bipolar electrode pairs was estimated using the shortest path from the stimulating electrode pair. Contact locations were reviewed and confirmed on surfaces and cross-sectional images by a neuroradiologist. For further visualization and presentation purposes, surfaces and contact coordinates were rendered using Blender (http://blender.org) and Blend4web (http://blend4web. org) open source software in a customized interactive web application.
Electrophysiological recordings
Intracranial data were recorded using one of the following clinical electrophysiological acquisition systems spe-New Research cific to a given site of data collection: Nihon Kohden EEG-1200, Natus XLTek EMU 128, or Grass Aura-LTM64. Depending on the acquisition system and the preference of the clinical team, the signals were sampled at either 500, 1000, or 1600 Hz and were referenced to a common contact placed either intracranially, on the scalp, or on the mastoid process. For analysis all recordings using higher sampling rates were down-sampled to 500 Hz. A bipolar montage was calculated post hoc for each subject by subtracting measured voltage time series on all pairs of spatially adjacent contacts. This resulted in N -1 bipolar signals in case of the penetrating and the strip electrodes, and N ϩ x bipolar signals for the grid electrodes, where N is the number of electrode contacts and x is the number of extra combinations of bipolar contacts that resulted from the montage.
Memory tasks with brain stimulation
The tasks were based on classic paradigms for probing verbal memory , in which subjects learned lists of words for subsequent recall (Fig. 1A). Subjects were instructed to study lists of individual words presented sequentially on a laptop computer screen for a later memory test. Lists were composed of 12 words chosen at random and without replacement from a pool of high frequency nouns (either English or Spanish, depending on the participant's native language; http://memory. psych.upenn.edu/WordPools). Each session had a set of 25 specific lists using words from the same general pool. The words on each list were either sampled from specific categories like vehicles, music instruments and vegetables, or they were sampled randomly. Each word re- Patient demographic data are presented together with clinical observations from structural MRI, clinically identified seizure onset zones (SOZs), pathology for those subjects who underwent respective surgery, hemispheric laterality of language functions together with the method of determination ("aphasia" means that the determination was done based on an identified lesion/pathology in a specific hemisphere), overlap of the stimulating electrodes with the language areas for patients who have undergone cortical stimulation mapping ("-" means that the stimulation mapping was not performed or the report was not available), verbal IQ (vIQ), and the clinical qualitative description of verbal memory deficits as concluded in the neuropsychological assessment. FC, frontal cortex; PC, parietal cortex; OC, occipital cortex; IC, insular cortex; aTC, anterior TC; MTL, mesial temporal lobe; TPC, temporo-parietal cortex; FPC, fronto-parietal cortex; OPC, occipito-parietal cortex; CD, cortical dysplasia; HS, hippocampal sclerosis; MCD, malformation of cortical development; MTS, mesial temporal sclerosis; PMG, polymicrogyria; DNET, dysembryoplastic neuroepithelial tumor.
mained on the screen for 1600 ms, followed by a random jitter of 750-to 1000-ms blank interval between stimuli. Immediately following the final word in each list, participants performed a distractor task (20 s) consisting of a series of arithmetic problems of the form "A ϩ B ϩ C ϭ ??", where A, B, and C were randomly chosen integers ranging from 1 to 9. Following the distractor task subjects were given 30 s to verbally recall as many words as possible from the list in any order. Vocal responses were digitally recorded by the laptop computer and later manually scored for analysis. Each session consisted of 25 lists of this encoding-distractor-recall procedure. Stimulation was applied by passing electrical current between two adjacent electrode contacts using parameters from the study (Suthana et al. 2012) showing a positive effect of stimulation on memory performance (bipolar symmetric, charge-balanced, square-wave stimulation at a frequency of 50 Hz and 300-s pulse width). Safe amplitude for stimulation was determined at the start of each session using a mapping procedure in which stimulation was applied at 0.5 mA while a neurologist monitored for after-discharges. This procedure was repeated, incrementing the amplitude in steps of 0.5 mA, up to a maximum of 1.5 mA for depth contacts and 3.5 mA for cortical surface contacts. These maximum amplitudes were chosen to be below the after-discharge threshold and below accepted safety limits for charge density (Mc-Creery et al., 1990). The stimulation was delivered for 4600 ms during the presentation of two subsequent words (from 200 ms before the first word onset to 200 -450 ms after second word offset due to a random jitter in inter-stimulus interval) on every other word pair (three pairs on every list with first pair pseudorandomized across all lists in a given session). Stimulation was applied on 20 out of 25 randomly assigned lists of a full session. There were no more than two sessions a day of a given task separated by at least three hours. The target electrode pair for stimulation was selected based on the anatomic coverage of brain regions associated with declarative memory functions (Eichenbaum, 2000), including hippocampus (HP), parahippocampal region (PH), temporal cortex (TC), and prefrontal cortex (PF). Within these regions specific target electrode pairs for stimulation were selected based on anatomic localization in one the studied brain regions and based on mapping of active areas showing a subsequent memory effect Sederberg et al., 2007). Electrodes had to be localized outside the seizure onset zone, as defined by the local clinical team. Additional clinical data were collected about the localization of language functions relative to the stimulation sites and neuropsychological assessment of verbal memory (Table 1). Stimulation amplitude was determined using conservative limits for safe charge density (Gordon et al., 1990;McCreery et al., 1990) for subdural or depth electrode contact, not higher than 3.0 and 1.5 mA, respectively.
Electrophysiological analysis
Brain activity induced by word presentation was analyzed in this study, and comprised 1600 ms of word display on the screen and 200-ms blank interval before and after each word (total 2000 ms segments). Stimulated word pair epochs were excluded from analysis to prevent potential contamination of spectral analysis with the stimulus artifact. Hence, one complete session yielded electrophysiological signal from 60 nonstimulated list epochs (five lists ϫ 12 words) and 120 stimulated list epochs (20 lists ϫ six words). Every signal epoch was spectrally decomposed in 50-ms time bins using multi-taper Fast Fourier Transform (Chronux toolbox, RRID:SCR_005547; Bokil et al., 2010); taper parameters: 4-Hz bandwidth, 250-ms timewidth, one taper). To estimate power in distinct frequency bands (high ␥: 62-118 Hz, low ␥: 30 -58 Hz, : 14 -26 Hz, /␣: 6 -14 Hz) signals were bandpass filtered between the corresponding cutoff frequencies (Barlett-Hanning, 1000 order) before spectral decomposition to reduce any possible influence of lower frequencies on the power estimate. The cutoff frequencies for the high ␥ band were chosen to minimize contamination of the 60-Hz line noise and its first harmonic at 120 Hz. The decomposed spectral power values in a given frequency band were log and z-score transformed in each frequency bin to account for the power law effect and obtain values that can be compared in the same scale across sessions and subjects. Frequency bands in the low and ␦ ranges between 1 and 5 Hz were not included in this study due to different high-pass filters applied in signal acquisition across the data collection centers. Average power estimates were calculated from all epochs of the studied words from nonstimulated lists. Exact time of memory encoding during the stimulus presentation is difficult to determine and can vary between subjects. We used the maximum peak value of the average power estimates as proxy for the brain response related to the memory encoding. This maximum value of the average power estimate was defined as peak power, and the difference between peak power values from the stimulated (P stim ) and nonstimulated (P non ) list condition was defined as the "neuromodulation (NM) index": is the nth power estimate.
Surface plots were created using the peak power and the NM index values interpolated between all bipolar pairs on an electrode grid. Active electrodes were selected by identifying outliers of the peak power value distributions above the upper adjacency value (UAV; Ͼ third quartile ϩ 1.5 ϫ interquartile range), which were calculated from all nonstimulated list epochs for every electrode in a given patient. The identified active electrodes were used to determine mean value of the NM index across all electrodes in a given subject or brain region, which had active electrodes from at least two subjects.
Behavioral analysis
Memory performance was quantified as count of words recalled per list (with or without stimulation). To compare the effect of stimulation on performance across subjects the raw counts from all sessions in a given subject were normalized into z-scores. Difference between means of the scores on the stimulated and nonstimulated lists was defined as a measure of stimulation's effect on memory performance (⌬ behavioral score). At least two sessions in a given stimulation target were required to be included in data analysis to ensure an accurate estimate of the mean for the nonstimulated lists, i.e., more than five scores were required to estimate the mean. ). B, Example of an 8 ϫ 8 electrode grid implanted over the lateral TC highlights in red two adjacent contacts used for brain stimulation (connected red dots) in subject 1050. C, Broadband spectrogram (left column) shows trial-averaged power changes aligned to the time of word presentation for encoding, in contrast to the power changes in the signal prefiltered in the four studied frequency bands (middle column), as recorded from a representative electrode example from subject 1111. Line plots on the right summarize the mean power change response independently for the four bands (rows) and separately for the good and poor encoding trials (columns) in the two conditions of list stimulation, color-coded as in A. Notice the difference in peaks of the response (NM index) caused by stimulation in the poor encoding trials specifically in the high ␥ frequency band.
Statistical analysis
All statistical tests were performed in Matlab (Math-Works Inc., RRID:SCR_001622) using built-in and custom written codes. One-way ANOVA test compared NM index calculated from the same set of electrodes from one subject in different frequency bands (Fig. 3C). The test was followed by Tukey-Kramer post hoc group comparison of the 95% confidence intervals of the means. Pearson's correlation was chosen to test dependence between NM index and: peak power value (Fig. 3D), distance from the stimulating electrode (Fig. 3D), and the behavioral effect of stimulation on memory performance (Fig. 5D). For the former two the correlation was additionally confirmed on the level of electrodes from individual patients. The correlation plots were complemented with least-squares lines to aid visual interpretation. ANOVA test was used to compare the effect of stimulating in the four studied regions on the NM index and on behavioral performance. The test was followed by Tukey-Kramer post hoc group comparison of the 95% confidence intervals of the means. Data are shown as mean Ϯ SEM. ANOVA tables are summarized in Table 3. All data collected in this project are available at: http://memory. psych.upenn.edu/RAM_Public_Data.
Results
We investigated the effect of direct brain stimulation on electrophysiological activity and memory performance in epilepsy patients undergoing evaluation for surgery to treat refractory seizures. Each patient was implanted with intracranial subdural, depth, or subdural and depth electrode arrays in multiple cortical and subcortical brain regions selected based solely on the clinical considerations. We identified 22 patients who were implanted in one of the four brain regions of the declarative memory system (Eichenbaum, 2000) and completed at least two sessions of free recall tasks with stimulation (Tables 1, 2). The tasks were based on a classic paradigm for probing verbal short-term memory , in which subjects learned lists of twelve words to be freely recalled in any order following a distractor (Fig. 1A). Electrical stimulation was applied between a pair of adjacent electrode contacts during encoding of words for subsequent recall (Fig. 1B). Low amplitude stimulation (Ͻ1.5 mA, 50-Hz frequency, pulse width 300 µs; Table 2) was applied for 4.6 s during presentation of two consecutive words, followed by presentation of two other words without any stimulation to enable electrophysiological analysis without stimulus artifact (Fig. 1A).
We found that stimulation in the lateral TC modulated the spectral power specifically in the high ␥ band (62-118 Hz) on electrodes showing induced responses to word presentation (Fig. 1C), which was associated with enhanced memory performance (for behavioral analysis, see Fig. 5). The high ␥ response on trials with words that are subsequently not recalled ("poor" encoding) is known to be decreased relative to trials with the subsequently recalled words ("good" encoding), as previously described , Sederberg et al., 2007. Stimulation on the poor encoding trials increased this high ␥ response and restored it to the magnitude observed on the good encoding trials with words that were subsequently recalled (Fig. 1C). Thus, the subsequently forgotten words from the stimulated "STIM" lists had increased high ␥ response relative to the words from the "NON-STIM" lists that were not stimulated. Each experimental session comprised of both the STIM and the NON-STIM lists, which were randomly assigned in a double-blind fashion. The modulatory effect of stimulation was quantified as a difference between peaks of the power response in the STIM minus the NON-STIM condition, which we called the NM index (Fig. 1C). The peak response was thus used as proxy for brain activity related to memory encoding.
This NM effect was localized to "activated" areas of the brain showing the induced high ␥ response in the tasks. Figure 2 presents three exemplar cases of stimulation from subdural surface grid electrodes in the TC, which modulated the peak power responses. The top case depicts a single discrete area of the peak activation. The magnitude of this discrete high ␥ response is greater on the good than the poor encoding trials in the NON-STIM control condition. This disparity between the remembered and the forgotten word trials is not present in the STIM condition with similar peaks on the two trial types ( Fig. 2A). Stimulation therefore increased the high ␥ response on the poor encoding trials to the levels seen during good encoding, selectively in the area of the induced task activity. The middle case reveals that this effect was also observed in an activated area of the occipital cortex, which was distant from the site of stimulation located in the TC (Fig. 2B). We did not observe this neuromodulatory effect (quantified as the NM index) in the bottom case, where no area in the TC was activated in the tasks (Fig. 2C). Cortical stimulation mapping of language functions was performed as part of the clinical evaluation in patients 1050 and 1177, which showed no overlap with the target stimulation electrodes (Table 1). We quantified these observations for all electrodes in the activated brain areas ("active" electrodes) in the only stimulated subject who had more than ten such active electrodes (n ϭ 22). The active electrodes were selected based on the distribution of the peak values of the high ␥ response from all available electrodes in a given patient (Fig. 3A,B). To test whether the observed modulation was specific to the high ␥ band we compared the NM index Figure 1B, are interpolated and visualized as surface plots overlaid on this subject's brain surface (left side). The first two columns present peaks of the high ␥ power in the STIM (first) and the NON-STIM (second) conditions, the third column presents the NM index, i.e., the effective difference between the first two columns. Arrows point to a discrete area of peak power modulated by stimulation particularly in the poor encoding trials. B, C, Analogous plots from two other cases of subject 1111 (brain surface rendering was turned upside down to aid visualization) and 1177, respectively. Notice that the high ␥ modulation is observed also at a distance from the stimulation site in subject 1111 and is not observed in subject 1177.
New Research values in four nonoverlapping frequency bands (/␣, , low ␥ and high ␥). A significant difference was found between the studied bands in the condition of poor memory encoding (p Ͻ 0.0001, ANOVA, F ϭ 14.8, degrees of freedom ϭ 3, 84) but not in the good memory encoding (p ϭ 0.171, ANOVA, F ϭ 1.71, df ϭ 3, 84) in this subject. NM index values for the high ␥ band in the poor encoding condition were significantly more positive (Tukey-Kramer post hoc test, p Ͻ 0.05) than for any of the other bands (Fig. 3C). We further investigated whether these significantly more positive values of NM index were correlated with the amplitude of the high ␥ response and with the distance from the source of stimulation (Fig. 3D). The mean NM index was positively correlated with the mean amplitude of the high ␥ response (Pearson's correlation, R ϭ 0.627, p ϭ 0.0018) and negatively correlated with the distance from stimulation site (Pearson's correlation, R ϭ -0.429, p ϭ 0.0461). These correlations suggest that the strength of modulation was dependent on the electrode activity in the tasks and its proximity to the site of stimulation.
In the final part of this study, we asked whether this positive modulation of the high ␥ activities induced in the free recall memory tasks is specific to stimulation in the lateral TC. We observed an inverse pattern of modulation when the other studied brain regions were stimulated. Figure 4 shows two example electrodes showing a positive NM index with TC stimulation (top rows) and two negative index values with stimulation in the HP (bottom rows). The latter came from subject 1024, who noted decreased memory performance on the STIM relative to the NON-STIM lists.
To test this observed relationship between the behavioral performance and magnitude of the modulation in different brain regions, we compared the effect of stimulation in the four regions involved in the declarative memory system: PH (entorhinal/perirhinal and PH gyrus), HP (subiculum and HP proper), lateral TC (middle and superior temporal gyrus), and PF (middle and inferior frontal gyrus). Precise localization of all stimulation targets used in every subject (N ϭ 23) is shown on a unified brain surface (Fig. 5A) and can be viewed online (to be identified if the article is published). We summarized the behavioral effect of stimulation across the studied brain regions to find that all four subjects stimulated in the lateral TC showed a positive effect on memory performance (Fig. 5B). There was a significant effect of the brain region (p ϭ 0.0019, ANOVA test, F ϭ 7.31, df ϭ 3, 19) revealing a stronger positive modulation of memory performance in the TC stimulation group than any other brain region (Tukey-Kramer test, p Ͻ 0.05). Stimulating in the four regions also exerted different effects on the high ␥ modulation (p Ͻ 0.001, ANOVA test, F ϭ 23.27, df ϭ 3, 194). We found that the NM index, averaged over active electrodes from stimulation in a given region (n ϭ 198), followed the same pattern (Fig. 5C) with a stronger positive NM in the TC group compared to any other group (Tukey-Kramer test, p Ͻ 0.05). Plotting the behavioral modulation score as a function of the mean NM index for every subject (Fig. 5D) confirmed that the electrophysiological effect of stimulation and memory performance were correlated (Pearson's correlation, R ϭ 0.50, p ϭ 0.016). Subjects 1050 and 1111, who noted the highest NM index values, demonstrated the Stimulation selectively modulates task responses in the high ␥ frequency band. A, Spectrogram of trial-averaged high ␥ response to word presentations recorded on an electrode in the brain area activated in the tasks. B, Active electrodes showing this response were identified as positive outliers of the peak value distribution of this response (red data points above the solid line of UAV). C, Mean NM index of all active electrodes in one stimulated patient (n ϭ 36) is compared among four frequency bands in the poor and good memory encoding conditions. Subplots on the right show post hoc comparison of the group means, dashed lines mark the 95% CI intervals (error bars) for the high ␥ group, and red indicates significant group with the intervals that do not overlap with any other group. D, Scatterplots with least-square lines show correlations of the NM index values in the poor encoding condition plotted against peak value of the task response (left) and against the distance from the stimulation site (right) for the active electrodes from C. Notice that NM index was proportional to the induced power response and inversely proportional to the distance from the stimulation site. greatest memory enhancement (Fig. 5D). Conversely, subject 1024 with the lowest mean NM index, noted the greatest memory impairment.
Discussion
In this work, we found evidence that electrical stimulation in specific regions of the human brain modulates high ␥ activities induced during encoding of words for subsequent recall. Positive high ␥ modulation, as observed with stimulation in the lateral TC, was associated with the brain region showing enhanced memory performance with stimulation, whereas negative modulation was seen in the HP, a region where stimulation had the opposite effect on memory recall. Both structures have been proposed to play differential roles in the declarative memory. HP and the medial temporal lobe structures are thought to be critical for binding episodic memory representations from distributed regions in the neocortex, which process and store memory (Squire and Zola-Morgan, 1991;Eichenbaum, 2000). Previous studies using electrical stimulation in the medial temporal lobe during memory performance in human subjects showed mixed results . Our results corroborate a recent report of stimulation-induced impairment in a range of tasks, including the free recall of word lists, applied in a large number of patients stimulated in the HP and the entorhinal cortex (Jacobs et al., 2016). Much less is known about the effect of stimulation in the lateral TC. Since the original reports of eliciting memory experience in individual epilepsy patients (Penfield and Perot, 1963), stimulation in this region of the human brain has been predominantly used for mapping language functions (Ojemann, 1991). Noninvasive stimulation (Tune and Asaridou, 2016) and imaging studies (Binder et al., 2009) support the role of brain regions in the lateral TC in processing semantic information. Another study with large number of epilepsy patients implanted with electrodes in various regions of the brain found that epileptiform discharges were impairing memory encoding of word lists specifically if they occurred in the lateral TC (Horak et al., 2017). Our results show that stimulation applied in the lateral TC enhanced the high ␥ activities in response to word encoding. In summary, there is a growing body of literature implicating the lateral TC in verbal memory functions.
Stimulation-related enhancement of the induced high ␥ activities was observed on trials with poor memory encoding and not on the good encoding trials. In fact, the average NM index for the high ␥ band during good encoding trials turned out to be negative (Fig. 3C). In a recent study of electrical stimulation applied during word encoding, the induced high ␥ activity was used to classify brain states into good and poor encoding states and predict that stimulating in the good state decreased the probability of recall and vice versa increased probability of recall when stimulating in the poor encoding state (Ezzyat Figure 1C. Arrows mark the positive and negative NM index changes in the three patients who showed the greatest positive (upper rows) and negative (lower rows) behavioral effects of stimulation (Fig. 5). et al., 2017). This interesting finding of good and poor encoding state-dependency is consistent with our observation of a positive stimulation-induced NM index during the poor encoding trials and a negative index during the good trials. Still, the positive effect of stimulation on the high ␥ activity was restricted to trials with words that were ultimately forgotten, making it challenging to explain the overall enhancement observed in the increased number of recalled words.
The outcome of stimulation was not only determined by the encoding brain state, but also by anatomic location. Our results show that both the neurophysiology and the behavior (recall performance) were differentially modulated depending on the brain region tested. The same stimulation pattern applied in the lateral TC versus the HP had opposite effects on the high ␥ responses and the associated recall performance (Fig. 5). The exact factors causing these differential effects on the neurophysiology and behavior remain unclear. The difference could be related to the qualitative differences in the electrode contacts used for stimulation, i.e., penetrating depth electrodes in the HP and subdural electrodes on lateral TC, but the surface area of the different electrodes is similar. In addition, five out of six subjects undergoing stimulation in the PF group were stimulated using subdural electrodes and did not show the same neurophysiological or behavioral effect as in the TC group. Further, the difference could be attributed to the range of stimulation parameters used. The original studies with epilepsy patients found that only a given set of amplitude and frequency parameters elicited the memory experience (Bickford et al., 1958;Penfield and Perot, 1963). Stimulating the same regions of the brain with higher amplitudes is known to disrupt cognitive processing of, e.g., verbal information mapped in these patients (Ojemann, 1991) as applied in clinical language mapping. Therefore, our reported results may not necessarily generalize to other tasks or be replicable with different set of parameters, which could not be tested within the scope of this study. Nevertheless, the results hold promise for using high ␥ activities as a biomarker of NM to target optimal parameters, phases and sites for stimulation and support that the stimulated region in the posterior half of the middle and superior temporal gyrus is specifically important for modulating memory processes engaged in these tasks.
Regarding the possible target sites, within the lateral TC there were distinct focal areas where word encoding induced the high ␥ activity (Fig. 2). These "islands" of high frequency power have been reported in the intracranial recordings during tasks (Kucewicz et al., 2014, which may indicate local processing of neuronal assemblies (Crone et al., 2006;Lachaux et al., 2012) and be used to map target sites for stimulation. Interestingly, the precise localization of the foci of high ␥ activities was not exactly the same in the studied subjects even within the lateral TC, possibly due to different strategies employed by subjects in these tasks (e.g., remembering more semantic or visual representations). At this point we can only speculate about the effects of stimulating in the focus or perimeter of these islands, over a gyrus or a sulcus, or at various scales of neuronal organization. Successful stimulation sites were localized on the middle temporal gyrus adjacent to the high ␥ island in two out of four subjects, who showed the strongest positive effect on neurophysiology and behavior (Fig. 5). Our study as well as others in the field (Suthana and Fried, 2014;Kim et al., 2016;Jacobs et al., 2016;Ezzyat et al., 2017) were performed with standard clinical electrodes with contacts of diameters ranging from 1 to 10mm 2 and separated by 5-10 mm. We speculate that future studies using combined macroand micro-electrode arrays could provide additional information of the spatial scale of the neuronal networks underlying memory function (Le Van Quyen et al., 2010;Viventi et al., 2011;Worrell et al., 2012;Kucewicz et al., 2016).
With regard to the target phases and parameters for stimulation, there are many other possible approaches to enhance memory processing and task performance. We have focused on modulating the encoding of memorized stimuli during their presentation, which induces high frequency activities. Another approach is to modulate maintenance, consolidation or retrieval of memory for the encoded items, which are thought to engage oscillatory activities in the lower frequency bands, including the rhythm (Buzsaki, 2006;Düzel et al., 2010). These lower frequency oscillations were shown to be more widely spread than the focal ␥ responses Kucewicz et al., 2014), thus possibly providing a viable target for noninvasive stimulation techniques. For instance, transcranial magnetic stimulation was employed to modulate oscillations mapped in parietal cortex to enhance retention of nonverbal memory (Albouy et al., 2017). Memory performance was increased in 13 out of 17 subjects and attributed to entrainment of the oscillations during the maintenance phase of the task. Other studies using noninvasive stimulation in similar tasks to probe active maintenance of memory in the PF showed mixed effects on reaction time and accuracy (Brunoni and Vanderhasselt, 2014). Although these studies are limited in terms of elucidating the ongoing neurophysiological activity, they complement the invasive intracranial recordings with insight into other measures of neural excitability and plasticity (Kincses et al., 2004;Fregni et al., 2005).
The precise memory processes that were modulated in our study are elusive. The observed NM did not directly enhance memory encoding per se since the high ␥ modulation was observed on the poor encoding trials with words that were subsequently forgotten. It could rather enhance memory performance through an associated process. Selective attention, perception and computation of sensorimotor information were all proposed as functions of ␥ oscillations (Singer, 1993;Tallon-Baudry and Bertrand, 1999;Jensen et al., 2007;Fries, 2009), which are essential to memory performance. If stimulation worked by increasing the level of attention and/or sensory processing of words, it would aid their encoding but not necessarily improve the retention and recall of all of them. In this scenario, the likelihood of successful memory en-coding would be increased specifically on the trials with words that were not adequately attended and processed. As a result, more of these words would end up being recalled due to this enhanced attention or perception to the verbal stimuli, which is what we observed on the behavioral level. There would still be words that did not end up being recalled despite the stimulation-induced enhancement of these associated processes. In summary, stimulation would restore processing of these allegedly "less attended" words, increasing their subsequent recall probability that would lead some, but not all, to transition and add to the number of the recalled words (the good encoding group). Disentangling this challenging relationship between memory and the associated processes requires additional experiments that can track attention and sensory processing through other behavioral or autonomic measures, e.g., the eye movements or pupil dilation.
Another way to identify the cognitive processes modulated by electrical stimulation is to test the existing computational models of memory. One can look for example at the classic primacy and recency effect in remembering lists of stimuli (Murdock, 1962) or the temporal contiguity effect (Sederberg et al., 2010). The former model incorporates serial position of a word on the presented lists with a prior knowledge that the ones in the beginning and in the end of the list tend to be more attended, and thus better recalled, than the middle-list words. The latter is a model of the probability of recall based on temporal proximity of the presented words; words presented next to each other are more likely to be recalled together. In the current paper, we explored these possibilities and did not find compelling evidence for either; however, the current study is limited by a small number of trials to compare. Both of these models may prove useful in future for elucidating the effect of stimulation on memory processing with larger number of subjects.
Finally, physiologic mechanisms of the high ␥ modulation and how it is linked with the associated behavioral effect remain to be explored. Direct brain stimulation is thought to primarily activate neuronal axons rather than cell bodies (Perlmutter and Mink, 2006), which would provide one explanation for why the electrophysiological effects were observed and not only in the region of stimulation but also in more distant islands of high ␥ activity, presumably connected with each other. It could also account for the disparity between the frequency of stimulation (50 Hz) and the higher frequencies of the modulated high ␥ response. Axons of the stimulated white matter tracts may be depolarized and trigger a response of neuronal assemblies oscillating at other frequencies in the distant brain regions they connect. Supporting evidence for the role of axonal stimulation comes from micro-electrode stimulation combined with calcium imaging that shows wide-spread activation of sparsely distributed neurons instead of local depolarization of neurons surrounding the stimulating electrode (Histed et al., 2009).
We observed that the modulation was stronger on electrodes closer to the stimulation site and more active in the tasks. This may possibly reflect a small-world network organization of the brain (Bassett and Bullmore, 2006), which proposes higher number of local and fewer longdistance connections. Therefore, more of the short-range local connections would be depolarized by the electric current and activate more proximal neuronal assemblies, relative to the longer-distance assemblies. In this network view of brain modulation, stimulation would also exert the strongest effect when applied to brain regions, which were critical nodes, i.e., hubs, with many connections to other active nodes in a given network. The lateral TC and the HP, where we observed the strongest positive and negative modulation of high ␥ activities respectively, are both considered critical hubs for declarative memory networks. Therefore, finding and targeting these critical connection hubs to modulate the whole network instead of a single brain region may be the most efficient strategy for enhancing memory processes . In our study, stimulation in the lateral TC could work by activating a network hub for verbal declarative memory. These network hubs can potentially be more effectively identified using various measures of connectivity and temporal interactions like spectral coherence or cross-frequency coupling. Future investigations of the brain connectomics data and modeling tools combined with high-density electrophysiological recordings promise to shed light on the mechanisms of electrical modulation for memory and cognitive enhancement. | 9,364.6 | 2018-01-24T00:00:00.000 | [
"Biology",
"Psychology"
] |
Inhibiting acid‐sensing ion channel exerts neuroprotective effects in experimental epilepsy via suppressing ferroptosis
Abstract Background Epilepsy is a chronic neurological disease characterized by repeated and unprovoked epileptic seizures. Developing disease‐modifying therapies (DMTs) has become important in epilepsy studies. Notably, focusing on iron metabolism and ferroptosis might be a strategy of DMTs for epilepsy. Blocking the acid‐sensing ion channel 1a (ASIC1a) has been reported to protect the brain from ischemic injury by reducing the toxicity of [Ca2+]i. However, whether inhibiting ASIC1a could exert neuroprotective effects and become a novel target for DMTs, such as rescuing the ferroptosis following epilepsy, remains unknown. Methods In our study, we explored the changes in ferroptosis‐related indices, including glutathione peroxidase (GPx) enzyme activity and levels of glutathione (GSH), iron accumulation, lipid degradation products‐malonaldehyde (MDA) and 4‐hydroxynonenal (4‐HNE) by collecting peripheral blood samples from adult patients with epilepsy. Meanwhile, we observed alterations in ASIC1a protein expression and mitochondrial microstructure in the epileptogenic foci of patients with drug‐resistant epilepsy. Next, we accessed the expression and function changes of ASIC1a and measured the ferroptosis‐related indices in the in vitro 0‐Mg2+ model of epilepsy with primary cultured neurons. Subsequently, we examined whether blocking ASIC1a could play a neuroprotective role by inhibiting ferroptosis in epileptic neurons. Results Our study first reported significant changes in ferroptosis‐related indices, including reduced GPx enzyme activity, decreased levels of GSH, iron accumulation, elevated MDA and 4‐HNE, and representative mitochondrial crinkling in adult patients with epilepsy, especially in epileptogenic foci. Furthermore, we found that inhibiting ASIC1a could produce an inhibitory effect similar to ferroptosis inhibitor Fer‐1, alleviate oxidative stress response, and decrease [Ca2+]i overload by inhibiting the overexpressed ASIC1a in the in vitro epilepsy model induced by 0‐Mg2+. Conclusion Inhibiting ASIC1a has potent neuroprotective effects via alleviating [Ca2+]i overload and regulating ferroptosis on the models of epilepsy and may act as a promising intervention in DMTs.
reducing the toxicity of [Ca 2+ ] i .However, whether inhibiting ASIC1a could exert neuroprotective effects and become a novel target for DMTs, such as rescuing the ferroptosis following epilepsy, remains unknown.
Methods:
In our study, we explored the changes in ferroptosis-related indices, including glutathione peroxidase (GPx) enzyme activity and levels of glutathione (GSH), iron accumulation, lipid degradation products-malonaldehyde (MDA) and 4-hydroxynonenal (4-HNE) by collecting peripheral blood samples from adult patients with epilepsy.Meanwhile, we observed alterations in ASIC1a protein expression and mitochondrial microstructure in the epileptogenic foci of patients with drug-resistant epilepsy.Next, we accessed the expression and function changes of ASIC1a and measured the ferroptosis-related indices in the in vitro 0-Mg 2+ model of epilepsy with primary cultured neurons.Subsequently, we examined whether blocking ASIC1a could play a neuroprotective role by inhibiting ferroptosis in epileptic neurons.
Results: Our study first reported significant changes in ferroptosis-related indices, including reduced GPx enzyme activity, decreased levels of GSH, iron accumulation, elevated MDA and 4-HNE, and representative mitochondrial crinkling in adult patients with epilepsy, especially in epileptogenic foci.Furthermore, we found that inhibiting ASIC1a could produce an inhibitory effect similar to ferroptosis inhibitor Fer-1, alleviate oxidative stress response, and decrease [Ca 2+ ] i overload by inhibiting the overexpressed ASIC1a in the in vitro epilepsy model induced by 0-Mg 2+ .
| INTRODUC TI ON
Epilepsy is one of the most prevalent central nervous system disorders, which features recurrent seizures caused by sudden hypersynchronous neuron discharges. 1 Although the first-line treatments for epilepsy are anti-seizure medications (ASMs), about 30% of patients with epilepsy fail to benefit from seizure control. 2,3The traditional ASMs mainly target remodeling the balance of excitation and inhibition, including regulating ligand-gated glutamate receptors, enhancing γ-aminobutyric acid (GABA) function, etc. 4 However, epilepsy is a chronic progressive disease with cell damage, triggering the inflammation response and recapitulation of development. 5In 2002, Löscher et al. 6 proposed that the development of diseasemodifying therapies (DMTs) is one of the important future goals for epilepsy treatment.However, there is currently no drug available that can ameliorate the course of epilepsies and related comorbidities.Therefore, developing effective and safe DMTs is a high priority in epilepsy research and care. 7][10] Notably, acid-sensing ion channels (ASICs) are a family of ion channels mainly activated by H + .They are expressed throughout the central and peripheral nervous systems, including the brain, spinal cord, and sensory ganglia. 115][16] In a previous study, we detected six tag single-nucleotide polymorphisms of the ASIC1a encoding gene in 560 patients with temporal lobe epilepsy (TLE) and 401 healthy controls.Notably, we found that an ASC1a variant allele (rs844347: A>C) was significantly associated with TLE. 17 However, this study is limited to direct and more powerful evidence based on brain tissue.
Interestingly, a recent study found high levels of ASIC1a in reactive astrocytes in the hippocampi of patients with TLE and epileptic mice.Moreover, selectively inhibiting the expression of ASIC1a on astrocytes by injecting rAAV-ASIC1a-shRNA into the dentate gyrus reduced the spontaneous seizures following pilocarpine injection.This finding points to the possible trafficking of ASIC1a in astrocytes during chronic epilepsy pathology. 18However, this study did not explore the role of ASIC1a on neurons in the development of epilepsy.Notably, psalmotoxin 1 (PcTX1) effectively and specifically inhibits the ASIC1a current without affecting the currents mediated by other configurations of ASICs, indicating that PcTX1 could be considered an indispensable pharmacological tool for the studies of ASIC1a. 19Xiong et al. reported that PcTX1 targeting ASIC1a on neurons protected the brain from ischemic injury by reducing the toxicity of [Ca 2+ ] i . 20However, whether PcTX1 could exert neuroprotective effects and be a potential direction for DMTs in epilepsy remains unknown.
2][23] When intracellular iron is overloaded, the excess iron generates reactive oxygen species (ROS) via the Fenton reaction, which promotes lipid peroxidation. 24In doing so, the glutathione antioxidant system is weakened, contributing to subsequent cell death. 257][28] Notably, abnormal iron metabolism is associated with epilepsy.Cortical iron injections can induce recurrent seizures, which are used as a model of post-traumatic epilepsy. 29Moreover, a previous study verified that inhibiting ferroptosis could mitigate pentylenetetrazol kindling, and pilocarpine-induced seizures in mice, suggesting that ferroptosis-mediated pathological alterations play important roles in the initiation and progression of epilepsy. 30,31wever, few studies evaluate the neuroprotective effects of suppressing ferroptosis in epilepsy.[Ca 2+ ] i plays an important role in regulating neuron function under physiological and pathophysiological conditions, such as cell death.Importantly, when ASIC1a opens, Ca 2+ enters in. 32However, whether blocking ASIC1a could exert neuroprotective effects through rescuing the ferroptosis associated with [Ca 2+ ] i following epilepsy remains unclear.
In this study, we first aimed to confirm the ferroptosis phenomenon on both brain tissues and blood samples in adult patients with epilepsy.Then, we investigated whether ASIC1a inhibitors-PcTX1 could exert neuroprotective effects using an in vitro epilepsy model Conclusion: Inhibiting ASIC1a has potent neuroprotective effects via alleviating [Ca 2+ ] i overload and regulating ferroptosis on the models of epilepsy and may act as a promising intervention in DMTs.
K E Y W O R D S
acid-sensing ion channel, disease-modifying therapies, drug-resistant epilepsy, ferroptosis, neuroprotection and its associated underlying mechanisms to probe its potential value as a DMT for epilepsy.
| Brain samples collection
Brain tissues used in this study were acutely resected from 5 patients with DRE who underwent surgical treatment.All patients were diagnosed with DRE based on the ILAE updated definition and were hospitalized for preoperative evaluation at the Epilepsy Center in the Beijing Tiantan Hospital of Capital Medical University. 3We selected the neocortical regions in the epileptogenic zone and corresponding distant non-epileptic temporal neocortex based on preoperative evaluation, including clinical history, neurological examination, scalp video-electroencephalogram (EEG) monitoring, magnetic resonance imaging (MRI), and 18-fluoro-deoxyglucose positron emission tomography, as epilepsy and control groups, respectively. 33,34All resected tissues were presented by using photographs taken after surgery and classified into two groups, namely, epilepsy and control groups (Figure 1C).Compared with the preoperative MRI, the postoperative CT clearly showed the location and extent of the resected zone (Figure 1A,B).Electroencephalography showed the typical epileptic pattern of activities in the frontotemporal region, which represented the resected epileptogenic zone for the epilepsy group (Figure 1D).Here, to safeguard brain tissue function and tissue activity, we collected only two tissues in the resected regions as epilepsy and control group from the same patient, which was approved by the surgery scheme.All human tissues were obtained with patient consent.This study was approved by the Medical Ethics Committee of Tiantan Hospital, Capital Medical University (Beijing, China).
| Blood sample collection
We enrolled 13 unrelated patients diagnosed with epilepsy and 15 age-matched healthy controls (Table 1).The age-matched healthy adults were collected at the health examination center without a diagnosis of any neurological diseases in the Beijing Tiantan Hospital of Capital Medical University for routine peripheral blood tests.
Considering the effect of multiple factors on ferroptosis, there were exclusion criteria to reduce bias, such as intake of antioxidants in 6 months and intake of iron products in 1 year.All participants signed an informed consent, and the study was approved by the Medical Ethics Committee of Tiantan Hospital, Capital Medical University (Beijing, China).
Peripheral blood samples were collected into EDTA Vacutainer Tubes (Shenzhen, China).After 45 minutes (mins) at room temperature, the plasma was obtained by centrifuging the whole blood for 3 min at 450 g, and it was stored at 80°C for measurements of ferroptosis-related indices.Take at least 500 μL of whole blood and centrifuge it at 4°C for 5 min.After discarding the supernatant, the precipitate was resuspended in 10-fold volumes of ice-cold homogenate, then centrifuged as before and discarded the supernatant.
Erythrocytes were lysed by approximately 4 times the volume of icecold Milli-Q water.Lysate was pelleted by centrifugation at 12,000 g for 5 min.The supernatant was taken for the determination of GPx enzyme activity.
| Primary cortical neuron culture
The protocol of primary neuronal cells was performed using the methods described previously. 35Cortices from embryonic rat (E17) were dissected in cold phosphate-buffered saline.Neural forceps, large scissors, and eye scissors were used to strip embryonic rats, and fine forceps, iris scissors, curved scissors, and eye scissors were used to rotate the brain tissue.The experimental devices were all sterilized in an autoclave.All subsequent operations were performed in a super-clean table.The tissue was then cut into pieces and digested with 0.05% trypsin and DNAase for 10-20 min.Digestion was terminated by adding DMEM supplemented with 10% fetal bovine serum, 10% horse serum, and 50× penicillin/streptomycin.After filtering the cell suspension with 70μm mesh, it was centrifuged at 200 g, dispersed, and resuspended with DMEM.Next, cells were mechanically counted and added at densities of 1 × 10 4 , 8 × 10 4 , and 5 × 10 5 cells/ well to poly-D-lysine pre-coated 96-well, 24-well, or 6-well plates, respectively.After 4 h of incubation to allow for cell attachment, DMEM was replaced with NBA Plus [Neurobasal-A medium supplemented with 1× GlutaMAX, 1× B27 supplement, and 100× penicillin/streptomycin].Cells were treated with cytarabine (5 μM) and cultured for 3-5 days to inhibit the proliferation of gliocytes.Every 2 days, 50% of the media was replaced with fresh NBA Plus.
| Establishment of epilepsy cell model
The in vitro 0-Mg 2+ model of epilepsy was established according to previous studies. 36
| Cell treatment
There are four groups: (1) Control group: The cells were pre-treated with isodose neurobasal, and then the cells were cultured in normal extracellular culture medium for 3 h.(2) 0-Mg 2+ group: The cells were pre-treated with isodose neurobasal, and then the cells were cultured in 0-Mg 2+ extracellular medium for 3 h.(3) PcTX1+ 0-Mg 2+ group: The cells were pre-treated with the concentration gradient of PcTX1 for 1 h, and then the cells were cultured in 0-Mg 2+ extracellular medium for 3 h.(4) Fer-1+ 0-Mg 2+ group: The cells were pretreated with Fer-1 for 1 h, and then the cells were cultured in 0-Mg 2+ extracellular medium for 3 h.The experimental design was summarized in Figure 4A.
| Cell viability assay
Cell viability assays were performed using Cell Counting Kit-8 (CCK8) (C0038; Beyotime Institute of Biotechnology, China) following the manufacturer's instructions.Briefly, cells were seeded in 96well plates in a medium.After drug treatment at the indicated time points, 10 μL CCK8 solution was added to each well and incubated for 3 h.Cell viability was finally measured using a microplate reader at a wavelength of 450 nm.
| Electrophysiology
Whole-cell recordings were performed on primary cortical neurons in parallel on the same day (days 13-16 in vitro), and viewed with an infrared differential interference contrast microscope with an
| Immunofluorescence staining
Immunofluorescence staining was performed as previously described. 37Briefly, frozen sections or cells were incubated with anti-ASIC1a rabbit polyclonal (1:500, Bioss) and anti-NeuN mouse monoclonal (1:300, Sigma) antibodies overnight at 4°C, followed by the appropriate secondary antibodies (1:500; Cell Signaling Technology) for 1.5 h at room temperature.Nuclei were counterstained with DAPI for 5 min at room temperature.Confocal images were captured using a laser-scanning microscope (A1R; Nikon, Tokyo, Japan).
| Mitochondria observation by transmission electron microscopy (TEM)
Transmission electron microscopy (TEM) was performed for mitochondrial observation of brain tissue.The prepared tissues were dehydrated and fixed with 2% paraformaldehyde and 2.5% glutaraldehyde.
Ultrathin sections of 70 nm thickness were cut and stained with lead citrate and uranyl acetate.Ultrathin sections of the samples were observed and pictured using an electron microscope (H-7650 system, Hitachi, Chiyoda-ku, Tokyo, Japan).The ultrastructural changes were assessed empirically by two blinded pathologists from our institute.
Mitochondrial ultrastructure from each sample was observed in five random visual fields according to the criteria.
| Ferroptosis-related indicator determination
All samples' proteins were determined by the BCA method (Thermo Fisher, USA).
| GPx enzyme activity assay
The GPx enzyme activity in leukocytes, tissue homogenate, and cell extracts was measured by a glutathione peroxidase assay kit (S0056, Beyotime, China).GPx could catalyze GSH to produce glutathione disulfide (GSSG), while glutathione reductase can catalyze GSSG to produce GSH using NADPH, and the level of GPx activity can be calculated by detecting the reduction of NADPH at A340.
| Glutathione assay
The GSH levels in the plasma, tissue homogenate, and cell extracts were performed according to the manufacturer's protocols (A006-1-1; Nanjing Jiancheng, China).Glutathione levels were detected by using the DTNB-GSSG reductase recycling methods.A405 was determined by enzyme marker.
| Iron assay
The iron concentration in the plasma, tissue homogenate, and cell extracts was measured by using Iron Assay Kit according to the manufacturer's instructions (A039-2-1; Nanjing Jiancheng, China).Iron levels were measured by the α, α'-dipyridyl method and determined by A520.
| MDA assay
The MDA concentration in the plasma, tissue homogenate, and cell extracts was determined according to the manufacturer's protocols (A003-1-2; Nanjing Jiancheng, China).MDA levels were measured using 2-thiobarbituric acid methods and followed by determination of A532.
| Mitochondrial membrane potential (ΔΨm)
The mitochondrial membrane potential (ΔΨm) was detected by using the mitochondrial membrane potential assay kit (ab113850; Abcam,
| Measurement of intracellular calcium concentration
Fluorescence imaging and qualification of Ca 2+ in primary cortical neurons were performed using the indicator dye Fluo-3AM (S1056; Beyotime Institute of Biotechnology, China).Cells were incubated in Fluo-3AM (2 μM) for 45 min at 37°C.The Fluo-3AM-loaded cells can then be used for qualitative fluorescence imaging and quantitative flow cytometry measurement.
| Quantification of ROS in primary cortical neurons
The ROS formation in primary cortical neurons was detected by a cellular ROS assay kit (ab113851; Abcam, Cambridge, MA).According to the manufacturer's instructions, the cellular ROS was detected by utilizing the cell-permeable reagent 2′,7′-dichlorofluorescein (DCFDA).After drug treatment, incubate cells with the diluted DCFDA/H2DCFDA and DAPI (100 ng/mL) for 45 min at 37°C in the dark.Finally, the plate was measured immediately on a fluorescence plate reader at Ex/Em = 485/535 nm and observed under a fluorescence microscope.
| Increased expression of ASIC1a and ferroptosis occurrence in epileptogenic foci from patients with DRE
To determine the expression of ASIC1a protein in the DRE, we colabeled the ASIC1a with NeuN which is specific to nuclei and perinuclear cytoplasm of most of the neurons in the central nervous system.We found that ASIC1a was expressed on the neurons in the epileptogenic foci (Figure 2A).In addition, we quantified the total expression of ASIC1a and found the level of ASIC1a was higher in the epilepsy group than in the control group (control: 1, epilepsy: 1.30 ± 0.02, p < 0.001, Figure 2B).We observed mitochondrial microstructure using electron microscopic to further confirm the occurrence of ferroptosis in DRE.Likewise, we explored the level of GPx4, a crucial mediator of ferroptosis, using Western blotting on brain tissues from epilepsy and control groups.Our results revealed significantly crinkled mitochondria in the epileptogenic foci compared to the relative non-epileptic neocortex (Figure 2C).In contrast to the control group, there was a significant decrease in the expression of GPx4 in the epilepsy group (control: 1, epilepsy: 0.62 ± 0.09, p < 0.05, Figure 2D).To sum up, as far as we know, these results provided direct evidence of the increased level of ASIC1a and ferroptosis in epileptogenic foci from patients with DRE for the first time.137.3 ± 20.16 mU/mg, p < 0.001, Figure 3B).However, there was no significant difference in the level of GSH between epilepsy and control groups (Figure 3E).These results powerfully articulated the presence of ferroptosis in adult patients with epilepsy.
Next, we measured the patient's clinical characteristics, including seizure frequency (Sz frequency), the interval time between the most recent episode and blood draw and course, to check the relationship between clinical symptoms and the ferroptosis-related indices.We listed the absolute and relative values of enrolled patients with epilepsy (Figure 3H).Interestingly, it showed a negative correlation between GPx enzyme activity and course, which meant that the longer course, the significantly lower GPx enzyme activity was found in epilepsy (R: −0.69, 95% CI: −0.89 to −0.15, p = 0.01, Figure 3F).In addition, a negative correlation was found between the MDA level and the time interval since the most recent episode, meaning that the closer the last seizure attack, the higher level of MDA was found in the plasma sample (R: −0.64, 95% CI: −0.88 to −0.12, p = 0.02, Figure 3G).Therefore, these suggested that ferroptosis was an irreplaceable part of the pathological alterations in epilepsy.
| PcTX1 recured the cell death induced by 0-Mg 2+
To further explore the potential value of PcTX1, a specific blocker of ASIC1a in DMTs, we established a 0-Mg 2+ -induced epilepsy model on primary cortical neuronal to mimic the epileptic state reported in a previous study. 36For identification of the primary cortical neurons in culture, the primary cultured cortical neurons were immunostained for DAPI and NeuN (Figure 4B).The purification of primary cortical neurons is 93.75%.Electrophysiology studies revealed that the 0-Mg 2+ extracellular medium-induced recurrent spontaneous seizure-like activities on the cultured primary cortical neurons (Figure 4C).Next, we evaluated the safety of the PcTX1 and examined its effects on rescuing cellular survival rate using CCK8 analyses (Figure 4D).As expected, it only showed significant inhibition on the cellular survival rate with the relatively high concentration of PcTX1 (control: 1.06 ± 0.04; vs. PcTX1 with 0.1 μM: 0.87 ± 0.06, p < 0.05; PcTX1 with 0.2 μM: 0.79 ± 0.05, p < 0.01, Figure 4D, the upper row), which indicated the significant cytotoxicity on primary cortical neurons.However, the results from the CCK8 assay showed that treatment with 0-Mg 2+ extracellular medium for 3 hours significantly reduced the survival rate of primary cortical neurons (control: 1.00 ± 0.13, vs. 0-Mg 2+ : 0.44 ± 0.06, p < 0.01, Figure 4D, the under row).To explore the potential effects of pretreatment with PcTX1 on 0-Mg 2+ -induced cellular injury, primary cortical neurons were exposed to different concentrations of PcTX1 (0.01, 0.02, 0.04, and 0.08 μM) for 1 h.Notably, they demonstrated that PcTX1 exerted the most apparent protective effects with a concentration of 0.02 μM (PcTX1 with 0.02 μM: 0.87 ± 0.11, p < 0.5, Figure 4D, the under row).
| PcTX1 inhibited the increased ASIC1a on the in vitro 0-Mg 2+ -induced epilepsy model
We next confirmed the increased ASIC1a in the 0-Mg 2+ group, and that PcTX1 could recover the changes of ASIC1a following 0-Mg 2+ induction in the subsequent experiments.As expected, we found an increased fluorescence intensity of ASIC1a in the 0-Mg 2+ group.
Moreover, we observed the reduced fluorescence intensity of
ASIC1a in neurons pre-treated and treated with PcTX1 compared with the 0-Mg 2+ group (Figure 5A).Consistent with immunohistochemistry staining, Western blot showed significantly increased expression of ASIC1a in the 0-Mg 2+ group, while PcTX1 pretreatment significantly reversed the increase in ASIC1a (control: 1, 0-Mg 2+ group: 1.22 ± 0.08, PcTX1 + 0-Mg 2+ group: 0.92 ± 0.04, p < 0.01, Figure 5B).As reported in the previous study, ASIC1a responded to the pH for half-maximal activation (pH 50 ) at 6.2, mediating fast decaying and transient currents. 11Thus, extracellular solution adjusted to pH = 6.0 was applied by a hand-made drug delivery device which could achieve rapid perfusion within 10 s to the cultured cortical neurons to activate ASIC1a and record the corresponding currents.To know whether 0-Mg 2+ could influence acid-evoked currents, we determined whether pretreatment of cells with 0-Mg 2+ / PcTX1 affects the amplitude of ASIC currents.Electrophysiological data then revealed that ASIC1a currents (I ASIC ) were significantly elevated after 0-Mg 2+ extracellular medium treatment compared to the control group, while pretreatment with the PcTX1 significantly reversed the elevation of I ASIC (control: 0.36 ± 0.13 nA; vs. 0-Mg 2+ group: 1.46 ± 0.40, p < 0.01; vs. PcTX1 + 0-Mg 2+ group: 0.10 ± 0.04, p < 0.001, Figure 5C).However, there were no significant differences in action potential frequency and rheobase (Figure 5D), as well as resting potential (Data S2).Thus, these results indicated that both of expression and function of ASIC1a were enhanced following 0-Mg 2+ induction and PcTX1 inhibited the increased ASIC1a in the in vitro 0-Mg 2+ -induced epilepsy model.
PcTX1 mediates the mitochondrial membrane function using the in vitro 0-Mg 2+ -induced epilepsy model (Figure 6H).The fluorescence transition from red to green indicated the loss of ΔΨm and hence significant mitochondrial damage.FCCP (a potent mitochondrial membrane disruptor) is used as the positive control group.
In summary, these results suggest that the ferroptosis pathway is
| PcTX1 rebalanced the disputed [Ca 2+ ] i homeostasis and reduced oxidative stress reaction induced by 0-Mg 2+
As an indicator of neuronal hyperexcitability, the disputed Ca 2+ homeostasis, represented as the elevation of Ca 2+ has been found in numerous studies based on epilepsy models and patients with epilepsy. 41Here, we found that intracellular Ca 2+ was significantly increased after treatment with 0-Mg 2+ extracellular medium using
| DISCUSS ION
In this study, we first provided direct evidence of increased ASIC1a and the occurrence of ferroptosis in the brain tissue of patients of DRE.In addition, we found the corresponding changes in ferroptosisrelated indices using blood samples from adult patients with epilepsy compared with age-matched healthy control.As far as we know, it Epilepsy is a chronic neurological disorder, afflicting approximately 50 million people worldwide. 1Currently, the first-line treatments for epilepsy are anti-seizure medications (ASMs), which are mainly designed for acute seizures control.However, epilepsy is a chronic progressive disease with complicated pathological changes. 5though more than 20 ASMs have been approved for treating epilepsy, no available drug can ameliorate the course of epilepsies and related comorbidities. 6In 2002, Löscher et al. first proposed the conception of DMTs and emphasized that the development of DMTs is one of the essential goals for epilepsy in the future. 6Our study mainly targeted finding a novel direction for developing DMTs for epilepsy.
6][47] High expressions of polyunsaturated fatty acids on neuronal membranes make the neurons more sensitive to ROS products. 48In addition, releasing excitotoxic substances, such as excessive glutamate and acidic synaptic vesicles, causes tissue acidification, acidotic reactions, and neuron death. 49,50erefore, suppression of excessive oxidative stress and acidosis reactions induced by seizures deserves more attention for protecting neuron function.
2][53] There existed a difference between rodent and human ASIC1a.It is worth noting that resected cortical tissue from humans exhibits a higher membrane/total ratio of ASIC1a than that from mice. 545][56] Acidotoxicity mediated by hA-SIC1a appears to be more severe than that by mASIC1a. 54These findings suggest that hASIC1a exhibits a greater response to acid signaling and has a stronger impact on the related biological effects than mASIC1a, so we selected clinically excised samples as experimental subjects.Current studies resolving the correlation between ASIC1a and epilepsy focus on alterations of ASICs protein and the direct effect of its antagonists on acute seizures. 18,57 our preliminary study, we found that I ASIC was significantly attenuated when exposed to ketone bodies which have been shown as an effective long-term treatment for epilepsy, suggesting that ASIC1a may be an essential part of the development and progression of epilepsy. 35To further confirm the changes of ASIC1a in epileptogenic foci from patients with chronic epilepsy, we selected inpatients diagnosed with DRE for surgical evaluation as subjects.
According to a previous study, we chose the temporal neocortex tissue, which is distant from the epileptogenic zone, from patients with mesial temporal epilepsy as the control group. 34en, we performed immunostaining with anti-ASIC1a and anti-NeuN antibodies and observed a high intensity of ASIC1a on the NeuN-positive cells.We found that the expression of ASIC1a was significantly increased in excised epileptogenic foci compared to the control group.PcTX1, a specific inhibitor of ASIC1a, has been reported to protect the brain from ischemic injury by reducing the toxicity of [Ca 2+ ] i by targeting ASIC1a expressed in neurons.influx, promoting the release of related neurotransmitters or even causing neuronal damage. 32,490-Mg 2+ -induced primary neuron is a classical epilepsy neuronal damage model: the absence of magnesium removes the block from the NMDA receptor channels and leads to their uncontrolled activation and excessive calcium influx, which can trigger a cascade of events that ultimately affect the activity of AMPA receptors and release excitatory glutamate.The activation of AMPA receptors by glutamate leads to an influx of Na + into the neuron, which further depolarizes the postsynaptic membrane.This depolarization can trigger action potentials and increase the firing rate of the neuron, contributing to neuronal hyperexcitability.9][60] In our study, PcTX1 was found to rescue 0-Mg 2+ -induced neuron death, rebalance the Ca 2+ homeostasis, and alleviate ROS production induced by 0-Mg 2+ .However, the detailed mechanism of PcTX1 exerting neuroprotective effects remains unclear.
Ferroptosis is a type of cell death mainly caused by lipid peroxidation of unsaturated fatty acids in the cell membrane. 2161,62 Ferroptosis in neurons was also observed in various epileptic animal models, causing neuron death increase, mitochondrial volume reduction, and ferroptosis-related index changes.After treatment with ferroptosis inhibitor Fer-1, all the above phenomena could be reversed, and the cognitive function of animals could be improved, suggesting that reversing ferroptosis was an essential component of the neuroprotective effect. 30,31multaneously ASIC1a is the only member of ASICs with specific permeability to Ca 2+ . 20,62,63It has been well documented that [Ca 2+ ] i overload can provoke cytotoxicity and induce cell death. 64,65In a recent study, overload in epileptic models.Here, we found that PcTX1 significantly recovered the changes of ferroptosis indexes on primary cortical neurons induced by 0-Mg 2+ , which showed a similar effect as the ferroptosis-specific inhibitor Fer-1.In addition, we found that PcTX1 had a stronger inhibition on the Ca 2+ than Fer-1.There are two potential reasons to explain it: firstly, ferroptosis might be partially due to Ca 2+ influx by ASIC1a; second, we proposed the possibility involved in the relationship of the ASIC1a-Ca 2+ -ferroptosis axis for epilepsy (Figure 8).In summary, the results suggested that remodeling [Ca 2+ ] i might be associated with regulating ASIC1a on ferroptosis.
It is worth mentioning that there still exist some limitations in our research, which are worthy of further research.
F I G U R E 1 | 5 of 19 SHI
Selection of the brain tissue for epilepsy and control groups.Representative presurgical evaluation of a patient was presented (A, D). (A) The preoperative magnetic resonance imaging (MRI): abnormal parasagittal T2-weighted MRI of the left frontotemporal regions, which is considered to be the epileptogenic zone; (B) the postoperative computed tomography (CT): CT after brain tissue resection; (C) all resected tissues were classified into two groups, namely, epilepsy and control groups; (D) the scalp video-EEG monitoring of ictal states of focal seizures evolves to generalized seizures: 3-5 s spike and wave discharges in left anterior regions displayed prior to 3-5 s diffuse low voltage and then spike and wave discharges in the bilateral frontotemporal areas rapidly generalized to all leads.et al.
The 4 -
hydroxynonenal (4-HNE) concentration in the plasma was determined according to the manufacturer's protocols (MM-0796R1; Mlbio, China).The concentration of 4-HNE in the samples was determined by comparing the O.D.(A450) of the samples to the standard curve.
(
median: 27 years old), while that of the control group ranged from 18 to 53 years old (median: 31 years old).As shown in Figure3E, the mean disease duration was 17.07 years (range 3-34 years), the mean generalized tonic-clonic seizure frequency was 1.06 times/ month (range 0.08-4.0times/month), and the mean interval time between the most recent episode and blood draw was 54.46 days (range 7-180 days).Although a gender bias was present in our study population, it showed no statistical difference in gender distribution between the epilepsy group and the control group.
3. 3 |
Figure3D) in the blood increased in adult patients with epilepsy.The GPx enzyme activity in leukocytes was significantly decreased in the epilepsy group (control: 355.40 ± 20.72 mU/mg, epilepsy:
F I G U R E 2
Increased expression of ASIC1a and ferroptosis occurrence in epileptogenic foci from patients with drug-resistant epilepsy (DRE).(A) Representative confocal photomicrographs of DRE patients with NeuN (red), ASIC1a+ (green), and DAPI (blue) triple-labeled (scale bar, 10 μm).(B) Representative Western blot images showing brain tissue of ASIC1a (B 1 ).Western blot analysis of ASIC1a protein expression in brain tissues (B 2 , n = 3 per group).(C) Western blot analysis of GPx4 protein expression in brain tissues (n = 3 per group).GAPDH served as a loading control.(D) Mitochondria observation by TEM analysis (scale bar, 200 nm).Yellow arrows, normal mitochondria; Red arrows, crinkled mitochondria.All data are presented as the mean ± SEM.Independent samples t-test was applied for statistical analysis.*p < 0.05, **p < 0.01.
F I G U R E 3 | 11 of 19 SHI
Correlation of clinical information and ferroptosis-related indices in adult patients with epilepsy.The ferroptosis-related indices in adult patients with epilepsy (n = 13) were analyzed and compared with healthy controls (n = 15).(A) Iron accumulation level; (B) GPx activity; (C) MDA concentration; (D) 4-HNE level; (E) GSH level; the absolute (H, the left side) and relative (H, the right side) values of enrolled patients with epilepsy were listed (n = 13).(F) The correlation between GPx enzyme activity and course.(G) The correlation between MDA level and the time interval since the most recent episode.All data are presented as the mean ± SEM.Independent samples t-test was applied for statistical analysis.A correlation was determined using Spearman's correlation.*p < 0.05, **p < 0.01, ***p < 0.001.et al.
F I G U R E 4
Experimental design and PcTX1 recured the cell death induced by 0-Mg 2+ .(A) Schematic of experimental timeline in vitro.(B) Representative immunofluorescence staining of NeuN in primary cortical neurons (scale bar, 125 μm, n = 3 fields per slide).(C)
is the first time to confirm the presence of ferroptosis in the epileptogenic foci from adult patients with DRE and peripheral blood samples from adult patients with epilepsy.The significant findings of our present work demonstrated that PcTX1, a specific blocker of ASIC1a, could rescue neuronal death and exert neuroprotective effects, including reversing the ferroptosis-related indices, ROS and [Ca 2+ ] i alterations on the in vitro 0-Mg 2+ -induced epilepsy model.We suggested that inhibiting ASIC1a could play a neuroprotective role on epileptic neurons through attenuating ferroptosis, which implies the potential value of ASIC1a in DMTs for preclinical and clinical studies for epilepsy.
F I G U R E 7
PcTX1 reduced [Ca 2+ ] i and oxidative stress reaction induced by 0-Mg 2+ .(A) The intracellular Ca 2+ was determined with Fluo-3 AM.Representative images of the changes in Ca 2+ were captured (scale bar, 125 μm).(B) Statistical analyses of intracellular Ca 2+ in different groups were shown (n = 6-8 per group).(C) The intracellular ROS were determined with DCFH-DA.Representative images of the changes of ROS were captured (scale bar, 125 μm).(D) Statistical analyses of intracellular ROS in different groups were shown (n = 7-8 per group).All data are presented as the mean ± SEM.One-way analysis of variance was performed for statistical analysis.ns, no significance; *p < 0.05, **p < 0.01, ***p < 0.001 versus control group; # p < 0.05, ## p < 0.01 versus 0-Mg 2+ group.
,
Petrillo et al. found the abnormal expression of ferroptosis indicators in the blood samples of children with epilepsy, which suggests that the determination of ferroptosis indicators may be a biomarker that plays a particular role in evaluating the ferroptosis in the brain tissues and the prognosis of epilepsy.38However, this study was limited to the specific cohort.Therefore, we collected blood from adult patients with epilepsy on the day of hospitalism before the clinical intervention to restore the patient's daily state to the greatest extent.Moreover, we observed the mitochondria and measured the expression of key enzymes of ferroptosis in the epileptogenic foci of patients with DRE.It is the first study to provide direct and indirect evidence to verify ferroptosis in patients with epilepsy, including representative crinkled mitochondria, increased iron deposition, depletion of glutathione, decreased GPx enzyme activity, and increased MDA content.Furthermore, we performed a deeper correlation analysis on changes in ferroptosis indicators in blood with clinical data.We surprisingly found that the activity of GPx, the key enzyme of ferroptosis, was negatively correlated with the duration of the disease.In addition, the content of MDA was negatively correlated with the interval between seizures, suggesting that F I G U R E 8 Diagram of the possible mechanism by which PcTX1 regulates the ferroptosis in epilepsy.Left: During the pathogenic process of epilepsy, cellular acidosis developed.Right: ASIC1a opens due to a rapid decrease in pH.The development of epilepsy is accompanied by the occurrence of ferroptosis.PcTX1 exerted neuroprotective effects by inhibiting the ferroptosis pathway.ferroptosis indicators in blood might become potential biomarkers in DRE.Given that our blood sample collection is limited due to the gender-biased population and statistical parameters only based on generalized tonic seizures, further research on female patients and different types of DRE needs further improvement.Consistent with the above-mentioned results, we found similar changes in ferroptosis indicators on the in vitro 0-Mg 2+ induced epilepsy model.
( 1 ) 5 |
The mechanism for the change of ASIC1a activity after 0-Mg 2+ ; (2) the difference between human ASIC1a and rat ASIC1a in expression and channel function; and (3) the specific mechanism of [Ca 2+ ] i in the ASIC1a-Ca 2+ -ferroptosis axis for epilepsy.CON CLUS ION In conclusion, our study provided evidence of the increased ASIC1a in epileptogenic foci and ferroptosis in patients with epilepsy.Meanwhile, we suggested inhibiting excessive ASIC1a with PcTX1 could exert neuroprotective effects via mediating neuronal ferroptosis by alleviating [Ca 2+ ] i overload.On the one hand, our work reveals a novel strategy of disease-modifying therapy for epilepsy to prevent ferroptosis-induced neuronal death, including the regulation of the Ca 2+ signaling pathway.On the other hand, it further expands the possibility of ASIC1a-[Ca 2+ ] i overload-ferroptosis in epilepsy, which supports a potential value of ASIC1a as a DMT for epilepsy, providing a theoretical basis for clinical translation.
Demographic and clinical data of adults with epilepsy.
TA B L E 1 All of the uncropped blots of these western blot experiments were presented in the Data S1.
UK), and anti-GAPDH (1:1000, Sigma Aldrich) as a loading control.Horseradish peroxidase-conjugated goat anti-mouse or goat antirabbit IgG (1:10,000, Applygen) was used as the secondary antibody, and the signal was visualized using Super ECL Plus substrate (P1050, Applygen).The indicated proteins were quantified with Image J software.2.14 | Statistical analyses 3 | RE SULTS3.1 | Patient cohorts and clinical informationA total of 13 unrelated patients with epilepsy (12 males, 1 female) and 15 age-matched healthy controls (13 males, 2 females) were enrolled.Detailed clinical information and pathological characteristics of all patients with epilepsy are summarized in Table1.The patients in the epilepsy group ranged from 19 to 49 years old 69,67.found that ferroptosis had been linked to the activation of protein kinase C isoforms, which are Ca 2+ dependent.62Apointworthmentioning is that the crosstalk between Ca 2+ , iron, and ferroptosis is bidirectional via ROS signaling.Likewise, iron-induced ROS generation elicits RyR-mediated Ca 2+ signals that promote ERK1/2 phosphorylation in primary hippocampal cultures kept in a Ca 2+ -free medium.66,67Besides,Bostancietal.found that blocking Ltype VGCCs may reduce the neurotoxic effects of iron by inhibiting the cellular influx of excessive Ca 2+ and/or iron ions.68Notably,excessironcausesmitochondrial fragmentation, elevated Ca 2+ levels that subsequently stimulate the Ca 2+ -dependent phosphatase calcineurin, and neuronal cell death in the HT-22.69Besides,ferroptosisalso generates excessive ROS. Coequently, excessive ROS production leads to oxidative stress and is detrimental to neurons.These combined observations support a mechanistic link between iron, Ca 2+ , and ferroptosis in the central nervous system.The balance of the triangle breaks down, the neuronal damage occurs.Therefore, we hypothesize that inhibiting ASIC1a can rescue neuronal death by regulating neuronal ferroptosis via alleviating [Ca 2+ ] i | 8,653.4 | 2024-02-01T00:00:00.000 | [
"Medicine",
"Biology"
] |
A novel approach with a fuzzy sliding mode proportional integral control algorithm tuned by fuzzy method (FSMPIF)
An automobile's vibration can be caused by stimulation from the road's surface. The change in displacement and acceleration values of the sprung mass is used to evaluate the automobile's vibration. Utilizing an active suspension system is recommended in order to attain an increased level of ride comfort. This article presents a novel strategy for regulating the operation of an active suspension system that has been put up for consideration. The PI (Proportional Integral) algorithm, the SMC (Sliding Mode Control) algorithm, and the Fuzzy algorithm served as the basis for developing the FSMPIF algorithm. The signal generated by the SMC algorithm is what is used as the input for the Fuzzy algorithm. In addition, the settings of the PI controller are modified with the help of yet another Fuzzy algorithm. These two Fuzzy methods operate independently from one another and in a setting that is wholly distinct from one another. This algorithm was created in a wholly original and novel way. Using a numerical modelling technique, the vibration of automobiles is investigated with a particular emphasis on two distinct usage situations. In each case, a comparison is made between four different circumstances. Once the FSMPIF method is implemented, the results of the simulation process have demonstrated that the values of displacement and acceleration of the sprung mass are significantly decreased. This was determined by looking at the values before and after implementing the new algorithm. In the first case, these figures do not surpass a difference of 2.55% compared to automobiles that use passive suspension systems. The second case sees these figures falling short of 12.59% in total. As a direct result, the automobile's steadiness and level of comfort have been significantly improved.
The automobile's comfort and steadiness are crucial factors. It can impact the comfort of the vehicle's passengers. The suspension system guarantees the proper level of ride comfort 1 . Typically, the suspension system is between the vehicle's body and the wheel. The components above a suspension system are known as the sprung mass (vehicle body). The components underneath a suspension system are referred to as unsprung mass 2 . A suspension system's primary components are a shock damper, lever arms (upper or lower lever arm), and springs (coil spring, leaf spring) 3 . According to certain studies, the anti-roll bar is also a suspension system component 4,5 . Compared to other systems, the suspension system's construction is relatively complex.
Uneven road surfaces are the primary source of automobile vibration, according to Zuraulis et al. 6 . Several more variables can also contribute to variations. However, the impact of these variables is negligible. Wheel vibrations are transferred to the car body via the suspension system. The suspension system will regulate these vibrations. In addition, the suspension system will decrease the vibration energy. When analyzing the vibration of a vehicle, several factors are considered, but displacement and acceleration values of the sprung mass are vital factors. These two markers have been utilized in much earlier research 7,8 . The displacement and acceleration of a vehicle body can be determined by simulation or experiment. Only the highest vehicle body displacement and acceleration values should be considered for discontinuous vibrations. The average and maximum values of the two parameters above may be employed for continuous vibrations. RMS critical allows for calculating mean values [9][10][11] .
The performance of the passive suspension system (mechanical suspension system) is poor. It does not meet the requirements for smoothness for substantial frequencies and continuous volume excitations. Instead of this, mechatronics suspension system solutions should be utilized. Zhang www.nature.com/scientificreports/ suspension 12 . This system utilizes balloons with completely automated control systems. These pneumatic balloons are variable-stiffness pneumatic springs 13 . The hardness of pneumatic springs may be altered by adjusting the pressure within pneumatic balloons. This was emphasized by Geng et al. 14 . When a vehicle is equipped with a pneumatic suspension system, its ride quality is good. However, this type is rather expensive. The use of electromagnetic absorbers to replace traditional absorbers often described as a "semi-active suspension system, " is another technology presented by Oh et al. 15 . According to Basargan et al., the current within the damper will alter the arrangement of the metal particles in its vicinity. Consequently, the damping stiffness is continually variable 16 . This kind is simpler and less expensive. Their efficacy, however, is typical. To better manage the automobile's vibrations, an extra actuator is required to upgrade the suspension system. Based on this approach, an active suspension system was implemented 17 . The active suspension system incorporates a hydraulic actuator. This actuator may apply force on the vehicle's mass from two sides. Consequently, its performance will improve. Nevertheless, the suspension system's construction will become more complex. Additionally, active suspension is more expensive than a semi-active suspension system. Recently, several publications concerning control for suspension systems have been published. Nguyen proposed i 18 employing the double combined PID (Proportional Integral Derivative) controller for the vehicle's quarter-dynamics model. This integrated double controller comprises two separate controllers. Each component controller regulates a distinct parameter. The PID controller's parameters K P , K I , and K D , must be chosen suitably. If a FOPID (Fractional Order Proportional Integral Derivative) controller is used in place of a PID controller, the number of variables will double 19 . Han et al. 20 developed a Fuzzy method to modify these settings. According to Mahmoodabadi and Nejadkourki's demonstration 21 , the value of these three factors may be altered continually. In addition, intelligence algorithms have been employed to optimize the PID controller settings [22][23][24] . For systems with multiple objects, either LQR (Linear Quadratic Regulator) or LQG (Linear Quadratic Gaussian) control algorithms are preferable 25 . By reducing the cost function, this approach will assist in optimizing automobile vibration 26 . Frequently, the preceding techniques are used to operate linear systems. The SMC method must be utilized for nonlinear systems. According to Azizi and Mobki, the objects will slide over the surface. The object then advances toward the location of equilibrium 27 . According to Nguyen, a sliding surface is a complicated function dependent on the controller's error signal 28 . The error signal is evaluated using the derivative of a high order. In order to ease the issue, it is essential to linearize a hydraulic actuator. This information was provided by Nguyen et al. 11 . Combining the SMC and Fuzzy techniques will improve its performance 29 . This has been demonstrated by Chen et al. in 30 when they used a combination of SMC and Fuzzy algorithms for the nonlinear system. Besides, the Fuzzy adaptive algorithms also help better observe the system's error state 31 . Many other intelligent control algorithms have also been applied to the suspension controller. In 32 , Liu et al. introduced the ANN (Adaptive Neural Network) algorithm for active suspension with constraints related to vehicle speed and displacement. The parameters of the controllers for the suspension can be optimally selected through methods such as GA (Genetic algorithm) 33 or PSO (Particle Swarm Optimization) 34 . Some techniques that use artificial intelligence to design the suspension controller have also been applied to heavy trucks 35 . In addition, several suspension system control methods are highly efficient 36,37 .
In order to meet the specifications for the automobile's ride comfort, it is crucial to regulate the suspension system's operation. The authors offer an original control algorithm, FSMPIF, in this work, based on four distinct perspectives. Besides, the controller design procedure is described in the article's content. In addition, a numerical simulation approach is used to analyze the vibration of the automobile. This article consists of four sections. In the Introduction section, some concepts and literature reviews are pointed out. In the Models section, the authors will explain the process of establishing a vehicle dynamics model and a control algorithm. The calculation and simulation process are done in the Results and Discussions section next. Finally, some comments will be indicated in the Conclusions section. In the following sections of the article, specific details are offered.
Models
Initially, developing a dynamics model of the vehicle's vibrations is necessary. This research used a quarterdynamics model with two masses; m s will produce the vertical displacement z s , whereas m u will do the vertical displacement z u (Fig. 1).
The differential equations describing vehicle vibrations are listed as follows: where: (1) and (2) produces: The control signal of the actuator, u(t), is determined by its controller. A completely innovative control algorithm named FSMPIF is proposed. This algorithm is developed with the following perspectives in mind: Firstly, the PI algorithm provides a more stable response, whereas the Fuzzy algorithm is more adaptable. Both of these components must concurrently exist in the control signal. Therefore, these two algorithms are required to combine an ultimate control signal.
Secondly, the PI algorithm settings must be adjusted appropriately. These values must be modified to accommodate the pavement's excitation signals. Consequently, they must be controlled by a fuzzy system. The vibration of the vehicle body is the input signal for the first fuzzy controller.
Thirdly, because the vehicle vibration is nonlinear, it is essential to design a nonlinear control algorithm to fulfill the system's stability requirements. The SMC algorithm is appropriate for this function. The output signal of the SMC technique will serve as the input signal for the second fuzzy controller described in the first point.
Fourthly, the second fuzzy controller is a crucial component of the integrated controller. Consequently, the signal of the second Fuzzy controller will consist of three components: the output signal of the previously described SMC algorithm, the error signal of the PI algorithm, and the vibration signal of the vehicle body.
On the basis of the considerations above, the Fuzzy Sliding Mode Proportional Integral tuned by Fuzzy (FSMPIF) technique was suggested. This algorithm satisfies all system stability criteria. Figure 2 displays the system's schematic.
Vehicle comfort can be measured through values related to car oscillations, such as displacement and acceleration of the sprung mass. These values are measured directly by the sensors fitted on the car. The result obtained from the sensor is the feedback signal of the system (Fig. 2). When evaluating ride comfort, we often consider the average, RMS, or maximum value.
Synthesis of the final control signals u(t) from the two-component signals u 1 (t) and u 2 (t).
The first element signal, u 1 (t), is the PI controller's output signal. Including the first viewpoint, the PI controller settings must be continually adjusted to fulfill the system's requirements. Therefore, tuning these settings with a Fuzzy system is a viable option. This is the first controller for Fuzzy. This controller's input is the sprung mass displacement value. Figure 3 depicts the membership function of this controller. This function is developed from the perspective of the author. A control signal will be transmitted as soon as the vehicle's body vibrates. Equation (12) can also be expressed as: The second Fuzzy controller's output signal is the integrated controller's second component signal, u 2 (t), also known as the central controller. The input signals for this controller are u 21 (t), u 22
(t), and u 23 (t).
The initial input signal, u 21 (t), represents the vehicle body displacement. The PI controller error signal is the second input signal, u 22 (t). It is multiplying a gain factor (k gf ) by this signal.
The SMC controller's output signal is the last input signal u 23 (t). An SMC controller is an integral part of the integrated controller.
Consider a nonlinear control object with u(t) as an input signal and y(t) as an output signal. A function determined by the component derivation signals and the input signals is referred to as the nth derivative of the output signal.
In this scenario, it is assumed that the function f(y) is limited and subject to uncertainty, i.e. Let the following be the values of the model's state variables: The object's model is returned as a system of state equations as follows: x 1 = y x 2 =ẏ ... www.nature.com/scientificreports/ The slip surface's bi coefficients must be set correctly, so that (26) is a Hurwitz polynomial. When this condition is fulfilled, state variables return to zero after a specific time (27).
Due to:
As a result, the Eq. (27) may be written as follows: where: T is a finite time point.
If the equation s(e) = 0 includes coefficients bi that fulfill the Hurwitz polynomial (26) condition, the sliding surface s(e) tends to zero, i.e.
The sliding condition (sliding surface) of the controller is defined by Eq. (30). We get the following from (22), (24), and (30): If s(e) is less than zero, the value of (31) is positive; otherwise, it is negative. The control signal u(t) may be rewritten as follows by combining (20) and (31): The control signal u(t), as given by Eq. (32), is independent of (22). As a result, it is regarded as a reliable controller. If condition (20) is not met, an upper limit of the function f(y(x)) must be defined, i.e.
Then the condition becomes: www.nature.com/scientificreports/ However, traditional SMC control algorithms still often cause the "chattering" phenomenon mentioned in 38 by Slotine and Li.
The procedure for designing an SMC controller is described in 28 . According to 9 , the SMC controller's output signal may be written as follows: The second component signal is a complicated function of the vehicle body vibration signal when Eqs. (16), (17), (18), and (19) are combined.
where: χ is the ratio coefficient between the inertial forces. This should be referenced in 28 . Figure 4 displays the membership function of this approach. The defuzzification procedure is carried out following the fuzzy rules specified in Table 1 and Fig. 5. www.nature.com/scientificreports/
Results and discussions
Condition of the simulation process. This study uses numerical simulation as its approach. This approach utilizes the MATLAB-Simulink ecosystem. The specs of the vehicle are listed in Table 2. These parameters are taken from the CARSIM ® application and slightly modified. Two case studies were conducted corresponding to two forms of road surface excitation (Fig. 6). In each scenario, the vibration of a vehicle will be evaluated under four conditions: passive suspension, PID, SMC, and FSMPIF. With road roughness as the input www.nature.com/scientificreports/ excitation signal, the system's output signal is the vehicle body displacement and acceleration. Maximum and average (RMS-calculated) outcomes for each condition will be compared.
Results of the simulation process. Case 1. In the initial instance, a road surface excitation of the sine cyclic form is utilized. According to this rule, the displacement and acceleration of the vehicle body will cycli- www.nature.com/scientificreports/ cally vary. Figure 7 depicts the change in sprung mass displacement over time. If an automobile has a mechanical suspension system, its maximum displacement can reach 100.83 (mm). This value can be decreased by employing an active suspension system. This result indicates that the vehicle body displacement is only 61.72 (mm) and 39.84 (mm), respectively when the PID and SMC algorithms handle the active suspension system. In particular, once the FSMPIF algorithm is used to control an active hydraulic suspension system, the maximum displacement value may decrease drastically, reaching about 1.55 (mm). Compared to the initial circumstance, this is merely 1.54%. This is a highly positive outcome. When evaluating vehicle stability, a mean vibration value must also be considered. This value may be determined using the RMS standard. According to simulated data, the sprung mass average displacement achieved 65.60 (mm), 43.60 (mm), 28.16 (mm), and 0.95 (mm) for the four examination scenarios. Using the value of the first scenario as a reference, the following numbers may be transformed equivalently to 100%, 66.46%, 42.93%, and 1.45%, respectively.
The acceleration of a sprung mass may be used to evaluate its vibrations. The value of vertical acceleration could be examined in this work. Figure 8's graph reveals that the greatest vertical acceleration for four simulated conditions is 1.96 (m/s 2 ), 1.91 (m/s 2 ), 1.64 (m/s 2 ), and 0.05 (m/s 2 ), in that order. Due to the continuous character of this vibration, the average value may also be determined using the RMS criteria. An average vertical acceleration of an automobile with passive suspension can reach 0.67 (m/s 2 ). This number may be dramatically lowered to as low as 0.01 (m/s 2 ) when the FSMPIF algorithm-controlled active suspension is utilized. This discrepancy is huge. Thus, the vehicle's comfort and stability may be significantly enhanced.
Considering the change in the acceleration value in percent, it can be clearly seen that the average value of the acceleration when using the FSMPIF algorithm is only 1.49% compared to the situation of the car without the controller for the suspension system. In terms of the SMC situation and the PID scenario, these numbers reach 43.28% and 64.18%, respectively. Regarding using the maximum value in comparison, if the value of the Passive situation is fixed at 100%, the values of the other three scenarios are only 2.55%, 83.67%, and 97.45%. The difference between the FSMPIF and the Passive situation is very large, while the difference between the SMC and the PID with the Passive is not much. This further helps demonstrate the efficiency of the new algorithm proposed in this article.
The control signal for the system is shown in Fig. 9. According to this result, the voltage value in the FSMPIF situation is highest, but there is a decrease over time to return to a stable threshold. This is consistent with the car body acceleration result shown in Fig. 8. Meanwhile, the output signal of the conventional SMC controller is unstable, also known as the "chattering" phenomenon. The control signal of the PID algorithm is more stable, but its response is not good (it causes the car body to fluctuate more than SMC and FSMPIF).
Case 2.
Random pavement stimulation is utilized in the second scenario. This is the actual variety of pavement. In this case, the amplitude and frequency of the vibrations are significantly greater than in the previous instance. Two results, including the vehicle body displacement and acceleration, are comparable to those in (Fig. 10). This number may be decreased by almost half to 48.29 (mm) if the PID algorithm handles the active suspension. This number can be decreased further by substituting the PID algorithm with the SMC method, which requires just 31.27 (mm). As soon as the new method FSMPIF presented in this article is used, the maximum displacement value may drastically decrease to 2.15 (mm). The Figure 11. Acceleration of the vehicle body (Case 2). In this situation, the acceleration value of the vehicle's body is relatively substantial. This might influence the vehicle's ride quality while in motion. These values regularly change throughout simulation time (Fig. 11). The maximum acceleration for an automobile with passive suspension is 13.45 (m/s 2 ). If active suspension using SMC or PID algorithms is employed, the acceleration value can be more extensive. This affects the vehicle's comfort. Only after the FSMPIF algorithm is implemented will the vertical acceleration value drop. This decrease is significant, just about 1.30 (m/s 2 ). For the two algorithms, SMC and PID, the percentage of maximum acceleration value are even more significant than that of Passive (102.30% and 106.17%, respectively). Meanwhile, the value belonging to FSMPIF is only 9.67%. In addition, the average values obtained from the calculation are 12.59%, 102.29%, and 108.24%, respectively, compared with Passive. Consequently, utilizing this innovative technique can increase vehicle stability.
The simulation's findings are presented in Table 3. The percentage differences between values are depicted in Table 4.
In the second case, the control signal changes continuously. The amplitude and frequency of the control signal are larger than in the first case (Fig. 12). The "chattering" phenomenon still occurs even when using only the traditional SMC algorithm. Meanwhile, the FSMPIF algorithm helps to limit this phenomenon more effectively.
Conclusions
The roughness of the road surface can cause the vehicle's body to vibrate. This vibration will damage the passengers' riding comfort. Consequently, an active suspension system is utilized to address this issue. The controller of the active suspension system will have a significant impact on its performance. In this article, the FSMPIF active suspension control algorithm is described. The proposed algorithm by the author is strict. This method is a combination between intelligent control, linear control, and nonlinear control.
The displacement and acceleration data of the vehicle's body are used to determine vibration levels. Through numerical simulation, these values are determined. Simulation findings indicate that when the FSMPIF algorithm is employed to regulate the active suspension system, the car body's displacement and acceleration values are significantly decreased. In both instances under examination, the maximum and mean values of displacement and acceleration are small compared to other circumstances. As a result, the vehicle's smoothness and comfort have been improved. This new method yields positive results. This method, however, is rather complicated. So, it should be simplified in the future to be applied to automobile mechatronic systems. Further, vehicle vibration testing must be done to confirm the effectiveness of this new control mechanism.
Data availability
The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.
Author contributions
All content belonging to this article is prepared by T.A.N. | 5,005 | 2023-05-05T00:00:00.000 | [
"Computer Science"
] |
A new paradox for well-being subjectivism
Abstract Subjectivists think that our well-being is grounded in our subjective attitudes. Many such views are vulnerable to variations on the ‘paradox of desire’, where theories cannot make determinate judgements about the well-being of agents who take a positive valuing attitude towards their life going badly. However, this paradox does not affect all subjectivist theories; theories grounded on agents’ prudential values can avoid it. This paper suggests a new paradox for subjectivist theories which has a wider scope, and includes such prudential judgement theories. I outline the new paradox and show how two plausible idealisztions (coherence and consideration) will not help. Subjectivists about well-being must either add an additional idealization that can solve the paradox of judgement or explain why such paradoxes do not constitute serious objections to a theory of well-being.
Introduction
Subjectivists think a subject S's well-being is determined by S's subjective attitudes.A common candidate is desire (Barrett 2022, Heathwood 2006, 2022, Lin 2016, Murphy 1999).The simplest desire view is that S's well-being depends on S getting what she desires.
Many subjectivist views face a paradox: the 'paradox of desire'.However, views which ground well-being in what S prudentially values avoid it (e.g.Dorsey 2012, 2021, Tiberius 2018).Thus, avoidance of paradox is one point in favour of subjectivism based on judgements of prudential value over rival subjectivist views: call such views 'judgement subjectivism' or JS.
This paper outlines a new paradox for subjectivism.Like the paradox of desire, it applies to a wide range of subjectivist views; unlike the paradox of desire, this includes subjectivist views grounded in what agents judge prudentially valuable.After outlining the paradox of desire and showing how JS avoids it, I outline the new paradox which affects JS along with a range of other subjectivist views.I show that two idealizations adopted by leading JS theorist Dale Dorsey, and which might plausibly be adopted by a range of subjectivist views, will not help to avoid this new paradox.
The paradox of desire is outlined by Bradley 2007(see also 2009: 30-32, Feldman 2004: 17, Heathwood 2005).DS stands for 'desire-satisfactionism': Suppose DS is true, and suppose Epimenides has just two desires.His first desire, D a , is a desire of intensity +5 for an apple.He does not get the apple, so his life includes a desire frustration of value -5.His second desire, D b , is a desire of intensity +10 that his life goes badly for him.Is D b satisfied?If it is, then Epimenides' life contains a desire-satisfaction of value +10, in which case his life has an overall value of +5 (it goes well for him), in which case D b is not satisfied after all.If D b is not satisfied, then his life contains a desire frustration of value -10, in which case his life has an overall value of -15 (it goes badly for him), in which case D b is satisfied.Thus if DS is true, D b is satisfied if and only if it is not satisfied, and Epimenides' life goes well if and only if it does not go well.(Bradley 2007: 46) Desire theorists cannot avoid the paradox; nor can many other potential attitude-based theories of well-being.But so long as the prudential judgements grounding well-being are required to be coherent, JS can avoid an equivalent challenge (Dorsey 2012: 422-24).
To generate a judgement-translated equivalent to the paradox of desire, I swap references to desires for references to Epimenides' coherent judgements about what is intrinsically good for him.I translate 'desires of intensity N' to 'judgements of value N'.
The translated scenario is: Suppose JS is true.Epimenides forms just two considered judgements about what is intrinsically good for him.His first judgement, J a , is a judgement of value +5 that eating an apple would be intrinsically good for him.He does not get the apple, so his life includes a judgement frustration of value -5.His second judgement, J b , is a judgement of value +10 that it would be intrinsically good for him if his life goes badly for him.
Whereas this scenario led to a paradox concerning desires, JS makes this scenario non-paradoxical because it is incoherent to judge that your life going badly for you is intrinsically good for you.It is not coherent to judge that something being intrinsically bad for you is intrinsically good for you. 1he paradox does not arise.Proponents of JS can offer a determinate judgement about how well Epimenides' life went for him: it was net negative, since his sole relevant prudential value judgement was frustrated.
The new paradox
JS can avoid the judgement-translated paradox of desire.However, a related paradox is not far away.Given the facts about his previous judgements, Ten may initially seem to be satisfied: Pythagoras has had nine relevant judgements satisfied, fewer than ten.But then Ten itself would be satisfied, taking his total to ten, and so on.
The paradox occurs at two levels.First, it is unclear whether Ten has been satisfied.Second, if Pythagoras's lifetime welfare is neutral apart from Ten, JS cannot say whether Pythagoras's lifetime well-being is positive, negative or neutral.Pythagoras's attitude is odd, but not unrealistic.Many people ascribe value to particular numbers. 3Perhaps Pythagoras thinks that there is something bad about the number ten, and it is best to avoid it.The judgement is also stable across his life: he avoids going over ten wherever he can (he keeps his stamp collection meagre, refuses to rent homes numbered above ten etc.).Importantly, Pythagoras need not be overly unrealistic to make the new paradox work.He may be just like any other individual, forming a range of 'ordinary' subjective attitudes alongside Ten: the paradox arises if those other attitudes balance out to make his well-being neutral.So, we cannot dismiss him as irrelevantly unrealistic.
I will shortly consider whether certain idealizations can rescue JS from paradox.However, I first show that the paradox applies more widely.
Consider desire again.If Pythagoras has had nine desires satisfied, and forms a tenth desire that fewer than ten of his desires are satisfied, an analogous problem arises.So too for the claim that Pythagoras 'prefers' Ten to be satisfied over its not being satisfied (Barrett 2019), and for several of the 'positive attitudes' mentioned by Heathwood (2014: 202), including 'caring about it … having it as a goal, being fond of it, being for it'.Insofar as these attitudes can be taken at a higher order (e.g.having goals about your goals), the paradox can arise.Finally, it is also worth noting that some idealizations suggested in the literature will not help.For instance, Ten is a judgement about Pythagoras's whole life, which he holds robustly; and thus restrictions to 'global' (Griffin 1988: 105;see discussion in Heathwood 2014: 213, Raibley 2012) or 'stable' attitudes (Raibley 2010, Tiberius 2018, Tiberius and Plakias 2010) will not help.Additionally, assume Pythagoras gets pleasure from satisfying Ten in different domains, e.g. he feels pleased when he thinks about the fact that he has read fewer than ten books in his life.Thus, more complex accounts of what it means to value something which includes an affective component (e.g.Tiberius 2018, Tiberius and Plakias 2010) are also vulnerable.
Possible solutions to the new paradox
I now consider two possible idealizations that subjectivists might adopt to avoid the new paradox.These two idealizations are adopted by Dorsey, and it is his formulations that I engage.However, they are also two obvious routes for other subjectivists, and I consider deviations from Dorsey's theory where this makes a difference.The two idealizations are that value judgements must be 'coherent' and 'considered'.
Take coherence first.Ten is not internally incoherent.There is nothing self-contradictory about it; we can understand what Pythagoras means when he makes the judgement, and it is satisfiable in many cases.For a coherence requirement to help, we need a more complex understanding.Dorsey (2021: 144) ' (2021: 144).Thus, judgements should not just be internally coherent, but also mutually coherent.
Mutual coherence is not joint satisfiability -I can coherently judge it good that I get two different cakes from the baker but be unable to afford both -but rather coherence in judgement.A person's coherent set of value judgements is determined through 'minimal mutilation', eliminating weaker or more peripheral judgements first (2021: 145).Dorsey imagines someone (call her the Gourmand) who judges it intrinsically prudentially good to eat food from Julia Child recipes and intrinsically prudentially bad to eat French cooking.Since Child's recipes are French, these judgements are mutually, though not internally, incoherent.The Gourmand values and disvalues the same thing at the same time, under different descriptions.If we tell the Gourmand that Julia Child's recipes are all French, she should adjust her judgements.
If coherence is to rescue JS from the new paradox, the tension involved should be more like the Gourmand tension than the bakery tension.In other words, it should be a genuine incoherence, not simply a case where mutually coherent judgements cannot be jointly satisfied.
To fail this more demanding coherence test, Ten must face tension with some other value judgement or with itself under another description.And indeed, the new paradox does seem to involve considering a particular event under different descriptions.If we ask Pythagoras whether he judges it prudentially good that Ten is satisfied, he will say yes.If we ask him whether he wants his tenth judgement to be satisfied, he will say no.We can present him with the fact that Ten is his tenth judgement, similarly to presenting the Gourmand with the fact that Child's cooking is French.
But these tensions are importantly different.Remember that in some cases, such as wanting two cakes but only having money for one, we cannot get all the things we judge to be good for us due to circumstance.There is no inherent tension between having cakes A and B; but if I get cake A I cannot get cake B, and vice versa.If things were different -if the cakes were cheaper or I had more money -I could satisfy both judgements.For the Gourmand, there are no circumstantial changes that would enable them to satisfy both judgements.
In this respect, Pythagoras's judgement is more like the bakery example than the Gourmand.There are circumstances under which Pythagoras can satisfy Ten.If he had only satisfied eight prudential judgements, he would be able to coherently satisfy Ten.It is bad luck that his circumstances mean he cannot satisfy it.And the reason he cannot satisfy it is because of its relation to his nine other satisfied judgements.Ten is satisfiable under the right conditions; but it is not jointly satisfiable with the nine other prudential judgements Pythagoras made previously.
Moreover, for the Gourmand, the solution is obvious: she must either change her mind about French food ('Julia Child cooks French food?Turns out I do like it!'),or Child ('Julia Child's been cooking French food this whole time?Disgusting!').But there is no straightforward resolution to the paradox of judgement.In the circumstances, there is no coherent way for Pythagoras to abandon just one of two evaluative attitudes that have come into tension, since the problem comes not from an incoherence among different value attitudes, but from the same event both satisfying and frustrating the same value attitude.
A defender of judgement subjectivism might object here that an important difference between Pythagoras's situation and the bakery customer's is that although the customer cannot jointly satisfy her judgements, they do still fit together.The problem with Ten is different: although not essentially incoherent, it turns out to be self-defeating in some contingent circumstances.Some subjectivists might insist that this contingent self-defeat is a kind of incoherence which also rules an attitude out of grounding well-being.One observation that potentially supports this is that while the case of Epimenides is not paradoxical for non-subjectivist views of well-being, the case of Pythagoras retains its paradoxical nature independently of its implications for well-being.For instance, a hedonist about well-being will still see that there is a puzzle about whether Ten is satisfied or not, even if they regard this puzzle as irrelevant to well-being.
It is worth thinking about what this means for Pythagoras's attitudes.One option is that Pythagoras can abandon the judgement Ten.If he did this, there would be no tension.However, this is unsatisfying, at least if we accept Dorsey's theory in full.Dorsey's concern with coherence is not a post hoc attempt to tidy up our evaluative sets so they fit neatly into theory; rather, it is an attempt to get at what we really value.It seems reasonable to say that the Gourmand either did really value some French cooking, or did not really value Julia Child's recipes; dropping a judgement is a way of getting at what she really values.But for Pythagoras, this is not true.If we say to him, 'Look, your judgement is paradoxical.You must have made a mistake about what you value', it would be reasonable for Pythagoras to insist there has been no mistake -he really does value avoiding going above ten -he has just been unlucky.Dropping the value judgement Ten might avoid the paradox; but it is not a good way of getting at what he really values.
Still, I have claimed that this new paradox is a problem for a wide variety of subjectivist views.Even if Dorsey's theory does not sit well with stipulating that the judgement Ten (or some equivalent attitude) is to be excluded in problematic cases, a different subjectivist could make this stipulation.One such example is Tiberius's (2018) value theory, which may endorse the stronger version of mutual coherence as joint satisfiability that Dorsey rejects.Tiberius says that our well-being is determined by how well we satisfy our 'appropriate' values, where appropriate values are, inter alia, 'capable of being fulfilled together over time'.This is amenable to weaker and stronger readings.On a weaker reading, it is enough that two values are in principle mutually fulfillable, even if the real world makes doing so impossible.On this reading, Tiberius's view of mutual coherence is close to Dorsey's.But on a stronger reading, values cease to be appropriate if they cannot in fact be fulfilled together.For instance, if someone values becoming a professional chef but also values becoming an accountant then, to the extent that it is not possible to be both, at least one ceases to be an 'appropriate' value.Thus, it may be that Ten becomes an inappropriate value simply by virtue of being de facto unfulfillable.
I suspect this reading is too strong.Assume that values-based views also hold that not getting what we value is bad for us (this is not strictly required by a view that holds that getting what we value is good for us, but it is a natural extension of it).If we adopted the stronger reading of joint satisfiability, this would mean that whenever one value clashes with another this is not really bad for our well-being because whichever value is not fulfilled is 'inappropriate', and thus irrelevant to well-being.That would imply -implausibly, I think -that there are never well-being trade-offs in making difficult choices, such as a choice over what career to pursue.For instance, if you choose to become a chef while still also valuing becoming an accountant, the stronger reading implies that it is not at all intrinsically bad for you to have this latter value go unsatisfied.But precisely what makes such choices difficult is that they do require giving up things we value; even if this is net positive for our well-being, there is some cost to doing so.
Thus, although on this reading Tiberius's theory may escape the paradox, I suggest that it does so at the cost of problems elsewhere.Note that this is different from the claim that the best life will be one where all one's values are jointly satisfiable: at other points Tiberius (2018: 57, 68; see also Raibley 2012: 252) seems to have this question in mind.Ten is not conducive to a good life; it makes it harder for the other things Pythagoras values to promote his well-being.But this is a different question than whether the judgement Ten, given Pythagoras actually does make it, is admissible as a welfare-grounding judgement.
Finally, a subjectivist might exclude judgements simply when they risk the kind of paradoxical result I have outlined.On this view, even if Pythagoras had only fulfilled eight other value judgements, Ten could not contribute to his welfare even though it is non-paradoxically satisfiable.Again, I acknowledge this as a possibility.But it is important to have an external motivation for such exclusions beyond the avoidance of paradox, and I am unsure why the risk is enough to exclude Ten in cases where it does not result in paradox.
Thus, the situation Pythagoras finds himself in seems to me to be a temporally extended, and numerically expanded, version of the situation where I can only buy one of two cakes.Pythagoras's situation is not like the tension involved in valuing Julia Child's food but disvaluing French food.
Turn now to a second idealization.Dorsey also says that the judgements which ultimately ground facts about well-being must be 'considered'.The reason for the consideration requirement relates to the role of what a person values in JS; Dorsey suggests that if S takes a valuing attitude towards something, that is not enough to say they really value that outcome.For instance, he imagines someone who claims to value becoming US President, but only because they misunderstand what being President would be like.Were they to learn what the Presidency really involves they would no longer value it.Thus, says Dorsey, they do not actually value becoming President even before learning the truth (see also Raibley 2010: 607-8).Rather, they value becoming President given the conditions they imagine it to involve.Dorsey suggests that S does not really value a particular thing 'in and of itself' if they 'take the relevant valuing attitude toward [it] under a particular description, but don't take that same valuing attitude under some other description ' (2021: 150).
What does this mean in practice?Dorsey considers Sobel's (2009: 337) suggestion that the pro-attitudes relevant to well-being must be those we would make on the basis of full information, rejecting this as insufficient because it is too tethered to the actual world.Dorsey's (2021: 150) 'full consideration' condition is that 'a necessary condition for x to value is that x would take the relevant attitude toward given full consideration of the ways might be'.If I value becoming President in a wide range of conditions, but not under conditions in which I have children, Dorsey's analysis suggests that I do not value becoming President, but rather value something else, such as 'becoming President while childless' (see also Tiberius and Plakias 2010: 422-23).
Let us return to Pythagoras.There are various reasons he might have for forming the judgement Ten.For instance, he might think the number ten is unlucky, and going above it will make bad things happen.Since that is untrue, full consideration would rule out Pythagoras's attitude as grounding well-being.However, this explanation of Pythagoras's aversion to the number ten is irrelevant to the paradox, since the relevant judgement concerns instrumental rather than intrinsic prudential badness.
Here is a more relevant story.Imagine that Pythagoras developed an instrumental 'bad luck' judgement such as the one mentioned just above early on in his life, and believed that exceeding the number ten would cause bad things to happen.However, the attitudes and habits he developed around this instrumental belief became so central to his life that he now holds an intrinsic prudential attitude towards the number ten.Ten is a special number, he now thinks, to be avoided for its own sake.He has forgotten the childish origins of his belief.And that new belief is now so deeply rooted that revealing the truth -that failing to avoid the number ten will not cause any harm -will not shake him from his judgement that it is to be avoided.Nor will reminding him of the origins of his judgements.
Such an attitude seems irrational.But it is not, I think, unconsidered in Dorsey's sense.Indeed, the sense that it is irrational is most obviously explained as a judgement that exceeding the number ten is not intrinsically bad for anyone.However, Dorsey cannot appeal to this as one of the facts which full consideration would reveal to Pythagoras, since it is precisely this sort of claim that is decided by each individual's values on JS.
I noted earlier that in the circumstances in which Pythagoras finds himself, there is a tension in satisfying the judgement Ten.I suggested that although there was such a tension, it was not of the sort needed for Dorsey's coherence condition to rescue JS from the paradox of judgement.One might think that this tension is, however, susceptible to the consideration condition.The line of thought may run as follows: Pythagoras thinks he values 'avoiding having ten or more of my prudential judgements satisfied'.But under conditions of full consideration, he would have to consider the following circumstance: 'avoiding having ten or more of my prudential judgements satisfied where doing so satisfies my tenth satisfied judgement'.And, one might think, Pythagoras will reject satisfying his tenth judgement where this frustrates that very judgement.
However, I think such a response misunderstands the tension at the heart of the paradox.The problem raised by the paradox is unlike Dorsey's case of becoming President, where things are not the way the valuing agent expects.If Pythagoras is thinking straight, when we ask him whether he values satisfying Ten in circumstances where he has already satisfied nine such judgements, he should not reply 'yes' or 'no'.He should rather reply that it is simply not clear whether it is possible to satisfy Ten in this case.The basic problem with the paradox of judgement is not that the 'less-than-ten' judgement is satisfied in circumstances that make it unattractive, but that there is a paradox in the very question of whether it is satisfied.And it is this paradox that leads, for JS, to a paradox in determining Pythagoras's overall lifetime well-being.
I suggest, then, that it is possible for Pythagoras to value avoiding the number ten as far as possible even following full consideration, and that appeals to full consideration will not help resolve the tension at the heart of the paradox of judgement.
I have argued that a wide range of subjectivist theories are vulnerable to paradoxical implications in unusual but conceivable circumstances, including views based on prudential value judgements which avoid the paradox of provides one.According to his view, S's well-being is grounded by what S 'values' (see also Raibley 2010, Tiberius 2018 and Tiberius and Plakias 2010 though these are each more complex than Dorsey's account); and it is in determining what S values that we consult S's prudential judgements.Coherence requires that 'one's evaluative judgements should not offer inconsistent evaluative verdicts concerning individual bearers of intrinsic prudential value | 5,254.2 | 2023-08-25T00:00:00.000 | [
"Philosophy"
] |
FACTORS INFLUENCING THE ADOPTION OF E-TILANG; EMPIRICAL EVIDENCE FROM THE UTAUT MODEL
: Mid-year 2017 The National Police of the Republic of Indonesia publishes e-tilang technology innovation. Traffic police use e-tilang in handling vehicles that violate traffic on the highway. This is to improve service to the public. This research factor influences acceptance and use of e-tilang by using UTAUT model. This research was conducted in Bengkulu area with 152 traffic policemen. The findings of this study indicate that effort expectancy, performance expectancy, and social influences positively affect the use of e-tilang. Furthermore, no positive effect on the intention of using e-tilang is the Facilitation Conditions. The results of this study are important steps to improve e-tilang services.
Introduction
Currently, the development of information technology in the government is growing rapidly.This is particularly important given the potential to improve services, reduce costs and accessibility to citizens (Carter & Bélanger 2005).E-government provides certain benefits of society transparency in government processes, efficient services will reduce costs and time.E-government for government helps simplify procedures, improve office management and create effective government regulations (Kayani et al. 2011).
In mid-2017, the State Police of the Republic of Indonesia issued e-tilang innovation technology.E-tilang is an online application used by traffic police.E-tilang aims to improve service to the public and reduce the misuse of payment of fines to traffic police on the highway.Traffic policemen are responsible for ensuring that traffic rules are adhered to by vehicle drivers (Bates et al. 2017;Bates et al. 2014).E-tilang reflects the strategy of the traffic police in delivering information and communications to the public.This is in line with Lindsay et al. (2011) study, which states that 72% of technology can help police work and be able to solve problems.
This study uses a model reference Venkatesh et al. (2003).The model used in evaluating the acceptance and application of technology.Unified Theory Acceptance and Use of the Technology (UTAUT) is able to explain behavioral intentions in using system information.UTAUT explains the intent and behavior of users in using system information.
Performance
Expectations, Business Expectations, Social Influence, and facilitating conditions have a direct effect on the intentions of using the system.This model describes the intent of the user in using the system information and behavior of its users.
The Performance Expectations, Effort Expectations, Social Influence, and Facilitation Conditions have a direct effect on behavioral intentions in UTAUT.These four constructs are used to measure from e-Government services at present.The linkage between the main constructs at UTAUT shows a high significance of technology acceptance.
www.japmnt.comBased on the above explanation, the researcher will identify the factors influencing the use and acceptance of e-tilang at traffic police by using the UTAUT method.This research will be useful for developing additional literature based on current conceptual data and future research in Indonesia and its territory.The findings of this research can be a consideration to improve information technology services in the police.
Literature Review
According to Alshehri, his research investigates the feasibility of UTAUT for the use of technology received in government (Alshehri 2012).The findings indicate UTAUT is eligible for use on e-government acceptance.Increasing use and acceptance of e-government especially in the service to the community makes the service easier and faster.Information technology affects the ability of police to solve problems.Legohérel (2013) said the use of new technologies can improve performance.Lau (2016), explains that performance expectancy greatly influences behavioral intentions to use technology.Effort expectancy explains the ease of service in e-government.Effort expectancy has the effect on behavioral intentions in using technology.Social influences can convince each other that it is easy to accept the use of technology.Facilitating conditions is a picture of infrastructure and technical support for the system used within the organization, this is like previous studies (Akhtar Shareef et al. 2014;Lin et al. 2010).
Research Model and Hypothesis
This method UTAUT is composite of Theory Reasoned Action, Planned Behavior Theory (Ajzen 1991), Technology Acceptance Model, Diffusion of Innovation Model, (Mustonen-Ollila & Lyytinen 2003) and TAM2 (Madden et al. 1992).The strength of this model is widely used in various studies and applies it extensively to various technologies (Williams et al. 2015).The UTAUT model consists of 4 variables, that is performance expectancy, effort expectancy, social influence, facilitation conditions.According to Lin et al. (2010), the variable will be an important role as a direct determinant of usage behavior.In this study, performance expectations, social influence, effort expectations, and facilitation conditions are the main constructs that will be the effect on Behavioral Intention.Based on the UTAUT model, it can be estimated that this key factor influences the adoption of e-tilang.The linkages between major constructs at UTAUT have demonstrated a high significance of technology acceptance and have been shown to be consistent with many studies (Azam 2015, Raja Yusof et al. 2017).The UTAUT model as shown in figure 1. Empirically this study is the most widely used and applied adoption and acceptance model, using e-government services can provide useful insights and implications for understanding someone's intentions (Barua 2012).
Performance expectations are one of UTAUT constructions, performance expectations explain how much one believes that information systems can help the work to get the desired results (Venkatesh et al. 2003).This performance expectancy is derived from a combination of the concept benefit perception, extrinsic motivation, occupational conformance, relative profit, and expected outcomes.Performance expectancy makes it possible to access information quickly and conveniently.performance expectancy greatly affects the user's intentions (Lau 2016).The hypotheses are summarized as follows: H1: Performance Expectancy positively influences behavioral intentions in using etilang.
Effort expectancy describes the level of ease of using the system.Effort expectancy is a combination of TAM2 and MPCU methods.Effort Expectancy explains the ease of service e-tilang, how users interact with the interface.This construction is like that of Williams et al. (2015) that influenced users' attitudes toward usage.The hypotheses are summarized as follows: H2: Effort Expectancy positively influences behavioral intentions in using e-tilang.
The Social Effect is the UTAUT construct that explains how much one be sure lest employ a new system can minimize effort in work.This construct describes the person's ease or environment to influence each other.The hypotheses are summarized as follows: H3: Social Influence positively influences behavioral intentions in using e-tilang.
Facilitating conditions are how much one believes that the existence of a good organizational and technic infrastructure could support the employ from the system.Facilitating conditions is a very important service to the organization.The hypotheses are summarized as follows: H4: Facilitating Conditions positively affects behavioral intentions in using e-tilang.
Research Methods
Case studies in this research are traffic police in Bengkulu area.The number of traffic police respondents is 165 personnel.The survey was conducted using a questionnaire given to all traffic police.Based on Kumar (2016), the questionnaire was distributed with probability sampling technique.Determination of the number of respondents refers to previous studies (Chang, 2013) and is based on MacCallum et al (MacCallum et al. 1996).Another consideration is that in SEM, a sample size of 100-200 is required for sample size determination.
SEM techniques allow researchers to evaluate the construction model and to estimate the structural relationship between latent variables simultaneously (Hair et al., 2006).The data were tested using Structural Equation Modeling (SEM) with the help of AMOS 22 devices.Researchers collected data through 165 questionnaires and 152 returned, with a response rate of 87.36%.The questionnaire used was 165 and returned as many as 152 questionnaires, with a response rate of 87.36%.Respondent's demographics the results are shown in table 1.From the data processing, it is known that the value of CR shows the value of 1.96 and the value of P below 0.05 so it can be said that 3 hypotheses have an influence, and one hypothesis has no significant effect because the P value is more than 0.05.The results show that performance expectancy, effort expectancy, and social influence have a positive effect on behavioral intentions.Effort expectancy (p = 0.002) has a greater effect than performance expectation (p = 0.035).The social influence (p = 0,000) also has a positive effect.Facilitation conditions have a value of p = 0.850 which shows a P value above 0.05, which means that the facilitation conditions have no positive effect on behavioral intent.Venkatesh (2003) explains that facilitating conditions are cost and system availability.According to Lin et al, facilitating conditions and good infrastructure can facilitate the use of the system.
Discussions
The results show that performance expectation has a positive effect using e-tilang.E-tilang is useful in work because it can improve performance.The results of this study support the research of Lin et al (2010), which says using online systems can help improve performance.Usefulness is useful for traffic police because e-tilang can be used anywhere and anytime.The perceived benefits have described a trust to use the system in improving performance.The results showed that e-tilang use can improve service to the public.
The results showed effort expectancy positively influences behavioral intentions in using e-tilang.Traffic cops say e-tilang is easy to learn.The level of ease is very influential in using the system.Construct This means an important factor in e-tilang adoption.E-tilang also provides an easy user interface.These results indicate that ease of use in e-tilang can improve performance and expected effort.Etilang can reduce illegal levies with good results.The results of this study support some previous research (Bagozzi & Yi 1988;Alshehri 2012;Rehman et al. 2012;Verdegem & Verleye 2009).
Social influences also have a positive effect in using e-tilang.People who are important to traffic police have a big influence on the use of e-tilang.Director of traffic directs traffic police to use e-tilang.The use of e-tilang is also influenced by other co-workers.This supports Barua (2012) research, which says users are likely to comply with applicable regulations.Liu et al. (2014) also revealed that a significant social influence on the continuation of the intention of using mobile services. www.japmnt.com The study states that the facility does not affect the intention to use e-tilang.The condition of Police Facilitation in Bengkulu is still inadequate.organizations do not have special personnel if electronic tilang have problems.According to Ajzen (1991), facilitating conditions can act as a proxy for controlling the direct behavior in using the system.traffic police still use manual tilang, so it is less effective.From the survey results are also known to the Internet network is sometimes a constraint.A favorable facilitation condition can increase the user to use e-tilang.Lindsay et al (2011) demonstrates easy and fast technology services that can rapidly improve the police ability to solve problems.
Conclusions
This study examines the factors affecting acceptance and use of e-tilang issued by the Indonesian National Police.Using the UTAUT model, this study shows that almost all the variables used in the method positively affects behavioral intentions in using e-tilang, such as work expectation, performance expectation, and social influence positively affects behavioral intentions in using e-tilang.For the construct of facilitation conditions, the results do not positively affect behavioral intentions in using e-tilang.This is because there are still inadequate facilities within the organization.Therefore, facilitation conditions should be improved to improve more effective services.
This study is also important steps to improve e-tilang services in the future, especially improving facilities conditions by the Police of the Republic of Indonesia in support of such e-tilang, such as facilities handphone and other internet devices.The more complete facilities and facilities provided to traffic police, the intention to use e-tilang will also be higher.
Limitations
This study has limitations because it only uses the main construction of the UTAUT model and has not included moderating effects (age, gender, and experience).Furthermore, the researchers only tested the acceptance of etilang at the traffic police organization.
Figure 2 .
Figure 2. Model of Research
Table 4 .
Criteria model fit | 2,679.6 | 2018-05-23T00:00:00.000 | [
"Computer Science",
"Business"
] |
Existentialism- The Struggle Remains in Mulk Raj Anand’s Major Novels
Existentialism is a philosophy that emphasizes individual existence, freedom and choice. It is centered upon the analysis of existence and of the way humans find themselves existing in the world. The perception is that, humans exist first and then each individual spends a lifetime changing their essence or nature. Existentialism is a philosophy concerned with finding self and the meaning of life through free will, choice, and personal responsibility. Existentialism is a quest for authentic existence. Jean-Paul Sartre says, ‘Man is nothing else but what he makes of himself. Such is the first principle of existentialism.’ Man’s sufferings and humiliations comes under the aspect of existentialism, which is found in the novels of Anand. Anand is a humanist and his humanism manifests itself in a realistic representation of the inhumanity of the situation of the oppressed masses, suffering, various types of disability, discrimination and alienation. Existentialism is an aspect of humanism and Anand has portrayed it through human beings pathetic sufferings and miseries. Anand’s humanism dwells into the survival of human love through existentialism. The humanism of Anand showcases the concerns of existentialism, exposing the reality of life and its tragic condition of suffering and misery. The pathetic condition of suffering and misery is existential since it has the elements of chance, absurdity and nothingness in them. Their alienated conditions are shaped by fear and loneliness. Though Anand denies of being an existentialist, his most of the works reveal existential ideologies of Sartre and Heidegger.
Introduction:
Existentialism is a philosophy that emphasizes individual existence, freedom and choice. It is the view that humans define their own meaning in life, and try to make rational decisions despite existing in an irrational universe. It focuses on the question of human existence, and the feeling that there is no purpose or explanation at the core of existence. It holds that, as there is no God or any other transcendent force, the only way to counter this nothingness is by embracing existence.
Existentialism believes that individuals are entirely free and must take personal responsibility for themselves. It therefore emphasizes action, freedom and decision as fundamental, and holds that the only way to rise above the essentially absurd condition of humanity is by exercising our personal freedom and choice.
Existentialism originated with the 19th Century philosophers Soren Kierkegaard and Friedrich Nietzsche. In the 1940s and 1950s, French existentialists such as Jean-Paul Sartre, Albert Camus and Simone de Beauvoir wrote scholarly and fictional works that popularized existential themes, such as dread, boredom, alienation, the absurd, freedom, commitment and nothingness.
Existentialism in the broader sense is a 20th century philosophy that is centered upon the analysis of existence and of the way humans find themselves existing in the world. The notion is that, humans exist first and then each individual spends a lifetime changing their essence or nature. In simpler terms, existentialism is a philosophy concerned with finding self and the meaning of life through free will, choice, and personal responsibility. The belief is that people are searching to find out who and what they are throughout life as they make choices based on their experiences, beliefs and outlook. And personal choices become unique without the necessity of an objective form of truth. An existentialist believes that a person should be forced to choose and be responsible without the help of laws, ethnic rules, or traditions.
Existentialism is a quest for authentic existence. Man must decide who he would be.
Each Individual must decide the question for himself. Each one's existence is his own. There is no universal pattern that can be imposed on all. Each must invent his values and he exists authentically in so far as he strives to realize values that really are his own.
Man is exercising freedom, will, decision, creativity, setting goals and striving for the attainment of selfhood. He appears as being possessed. Man is described as 'a being with others' capable of love and community. Many of the great existentialist thinkers have stressed the individualist's need to extricate himself from the crowd in order to be fully himself.
Sartre claims that a central proposition of Existentialism is, that existence precedes essence, which means that the most important consideration for individuals is that they are individuals, who act independently and are responsible conscious beings ("existence")rather than what labels, roles, stereotypes, definitions, or other preconceived categories the Existentialism deals with choices, decisions and personal commitments. The existentialists have a great concern for human existence, especially the plightful situation of the present human as being or an individual in the society, since an individual is pushed against all odds. Jean-Paul Sartre says, 'Man is nothing else but what he makes of himself.
Such is the first principle of existentialism.' Anand is a humanist and his humanism manifests itself in a realistic representation of the inhumanity of the situation of the oppressed masses, suffering, various types of disability, discrimination and alienation. Man's sufferings and humiliations comes under the aspect of existentialism, which is found in the novels of Anand. Existentialism is an aspect of humanism and Anand has portrayed it through human beings pathetic sufferings and miseries. Anand's humanism dwells into the survival of human love through existentialism.
The humanism of Anand showcases the concerns of existentialism, exposing the reality of life and its tragic condition of suffering and misery. The pathetic condition of suffering and misery is existential since it has the elements of chance, absurdity and nothingness in them. Anand slog like anything, it is viewed not only as a social activity but also as an instrument of self-realization. If man is alienated from the products of his labour, it is a moment of despair. The act of production brings for his hero suffering. D.Riemenschneider rightly remarks: "If man is alienated from his own nature, he is also alienated from the human nature of his fellow beings, a fact most obvious in the existence of antagonistic classes within a society." 2 In Anand's Untouchable, Bakha is very much aware of the discord between the world he is condemned to inhabit and the new world of his undying aspirations. He tried hard in vain to be in harmony with himself but soon realizes that he is an alien, an outsider who "Munoo did not laugh and talk even as much as he used to at Babu's house. He was possessed by moods of extreme melancholy in the mornings, dark of self-distrust and brooding sinking feeling which oppressed his heart and expressed itself in his nervous, agitated manner" 4 Anand's description of the working place and the analysis of Ganpat's attitude are the two important factors in the treatment of alienated labour as a motif in the novel .
The intimidating presence of Ganpath , who believes that the success of an entrepreneur lies in extracting the maximum out of the labourers, tyrannizes the poor workers to silence. But in his absence they sing together. As they chant the hill songs aloud, as if by a spell, their subdued spirits get charged with vigour and a consciousness of their identity that has almost been forgotten gets revived. It was 'as if he regained the wild freedom of his childhood' and he would take out his mirror and comb his hair with' the desire to be a man, to flourish the true dignity of manhood' 5 Munoo's experiences in Bombay and Daulatpur depicts his savage struggle for survival. In that struggle, life seemed to be a threat and death was a release. Life in Bombay was a dreadful pattern of garish opulence and rampant filth. The novel shows death through alienation.
Anand in his Two Leaves and a Bud, has taken up the problem of the appropriation and alienation of labour as one of the prominent themes. In the Economic Philosophical Manuscripts of 1844, Marx says,"for the worker who appropriates nature through his work, this appropriation appears as alienation, his own activity as activity for and of someone else, his vitality as sacrifice of his life, production of objects as their loss to an alien power" 6
This novel delineates Gangu's despair and anxiety as an alienated peasant in
MacPherson Tea Estate in Assam Hills. Life of Gangu is marked with fear, anxiety and frustration. He is ruthlessly exploited by the colonists who own the tea estate. Gangu is aware that the real cause of his tragedy is his poverty, but he puts up with his suffering and humiliation very abjectly. In the new environment, Gangu has to face the crisis of existence.
The wages given to Gangu's whole family was very meagre. It was only a hand to mouth existence to them. In case of emergency, they had to borrow money from the money-lenders, who charged them with heavy interest and they were unable to return the money in their whole life. This way the coolies in the tea plantation had to toil their whole soul and heart for their whole life, for the colonists to lead a luxurious life.
Gangu is a victim of man, god and civilization. He faces the storm which ruins his harvest with a feeling of resignation. Anand describes his feelings thus: "Gangu watched the violent play of God, the storm with an almost imperturbable calm, as if in the moment of his uttermost anguish, in the very moment of his despair, at the loss of his harvest, he had been purged of his fear of the inevitable." 7 The pattern of despair and delight has many dimensions. Binod Mishra rightly remarks here: "The doctrinal despair and delight is only a fragment in the total pattern. Anand's stress on hope and harmony is as explicit as his stress on despair and disillusionment." 8 According to Ralph Fox, for Anand, novel was the epic of struggle and the struggle for Anand was to be directed towards recreating the individual and his community. The optimistic view of human development stressed by Anand in his fiction is not without delight.
Despair and delight had been a continuous theme of his fiction. These two terms can be explained with sun and slum, happiness and sorrow, energetic and dispirited etc. Life moves on the two wheels of despair and delight, which gives a slice to our existence. Anand's envy "But I do not apologize for this because it is not easy in the face of such wretchedness and misery as I had seen in India to believe that material happiness and wellbeing had no connection with real happiness and the desire for beauty." 9 The theme of self-realization through labour is developed more meaningfully in Anand's another slim novel, 'The Road'. Reimenschender aptly remarks: "It is in this short novel that Anand more than in any other work has found a more profound insight into a possible solution of overcoming alienation." 10 In the novel, 'The Road', the government hires Bhikhu and the other untouchables to build a road so that milk can be easily transported from the village of Govardhan to the city.
The road would mean prosperity for the village, but the construction is opposed by the caste Hindus who refuse even to touch the stones quarried by the untouchables. There are two | 2,619.4 | 2019-11-28T00:00:00.000 | [
"Philosophy"
] |
Discovering Spotify – A Thematic Introduction i
With a user base now officially reaching more than 100 million, which includes 60 million paying subscribers, the music streaming platform Spotify is today widely recognized as the solution to problems caused by recent decades of digital disruption within the music and media industries. Spotify resembles Netflix, YouTube, and Apple Music as an epitome of streaming’s digital Zeitgeist that is shaping our future. Industry interviews, trade papers, academic books, and the daily press reiterate numerous versions of this “technological solutionism” (Morozov 2013) in almost as many variations. This thematic section of Culture Unbound is broadly concerned with the music service Spotify, and novel ways to situate and do academic research around streaming media. Approached through various forms of digital methods, Spotify serves as the object of study. The four articles presented here—three full length research articles and a shorter reflection—emanates from the cross-disciplinary research project “Streaming Heritage: Following Files in Digital Music Distribution”. It was initially conceived at the National Library of Sweden (hence the heritage connection), but the project has predominantly been located at the Umeå University’s digital humanities hub, Humlab, where the research group has continuously worked with the lab’s programmers. The project involves four researchers and one PhD student and is funded by the Swedish Research Council between 2014 and 2018.ii While most previous scholarship on Spotify has primarily focused on its service role within the music industry, its alterations to the digital music economy, or its influence on ending music piracy (Wikström 2013, Wikström & DeFilippi 2016, Allen Anderson 2015, Galuszka 2015, Andersson Schwarz 2013), our
Culture Unbound
Journal of Current Cultural Research project mainly takes a software studies and digital humanities approach towards streaming media.The project "Streaming Heritage" broadly engages in reverse engineering Spotify's algorithms, aggregation procedures, metadata, and valuation strategies to study platform logics, including underlying norms and structures.Reverse engineering starts with the final product (in our case the music service Spotify) and tries to take it apart, backwards, step-by-step.Basically, we draw a more holistic picture by using Spotify as a lens to explore social, technical, and economic processes associated with digital media distribution.The key research idea within our project is to follow files (rather than the people making, using, or collecting them) on their distributive journey through the streaming ecosystem, taking empirical advantage of inherent data flows at media platforms (such as Spotify).
Over the last ten years, the extensive field of media and Internet studies have used several digital methods to develop pioneering ways to analyse and understand the digital, the Internet, as well as digital media production, distribution, and consumption.Following the catchphrase "the system is the method" (Bruhn Jensen 2011), digital methodologies are increasingly deployed to perform social science or humanistic inquiries on, for example, big data and black-boxed media platforms (such as Spotify) (Ruppert, Law & Savage 2013).As a research practice, digital methods "strive to follow the evolving methods of the medium" (Rogers 2013:1).The issue of data of, about, and around the Internet, as Klaus Bruhn Jensen has eloquently stated, "highlights the common distinction between research evidence that is either 'found' or 'made'".If one disregards various complexities, basically all evidence needed for Internet or digital studies is already at hand.When interacting, searching, and listening to music at Spotify, for example, user data are constantly being produced.Such data are "documented in and of the system" and "with a little help from network administrators and service providers" it can be used as the empirical base for research (Bruhn Jensen 2011:52).
For researchers seeking to take empirical advantage of data flows at contemporary media platforms, it quickly becomes apparent "that such platforms do not present us with raw data, but rather with specially formatted information" (Marres & Gerlitz 2015).Data, in short, are often biased.Twitter, for example, determines what data are available and how the data can be accessed, and researchers often have a hard time knowing what relevant data might be missing.Hence, the major academic problem confronting media scholars working with digital methods is the lack of access to data.In our project, the main difficulty in doing research on and around Spotify is the reluctance of the company to share data.
Consequently, user data must be acquired and compiled through other means such as by deploying bots as research informants or by recording and aggregating self-produced music and sounds.Building on the tradition of breaching ex-
Culture Unbound
Journal of Current Cultural Research periments in ethnomethodology (Garfinkel 1967), where reactions are caused by disturbing or even violating commonly accepted rules or norms, our project has tried via repeated and modified so-called "interventions" to break into the hidden infrastructures of digital music distribution.On the one hand, we have been interested in broadly studying different data patterns and media processes at Spotify.On the other hand, we have also been keen on producing and obtaining research data, for example, by using bots as virtual listeners, by documenting (and tracing) Spotify's history through constantly changing interfaces, or by tracking and archiving advertisement flows.Using debugging software such as Fiddler or Ghostery, we have also tracked traffic between a computer and the Internet.
Although this thematic section of Culture Unbound is concerned with Spotify, basically any other streaming media services could be studied in similar ways.The various digital methods we present, use, and critically discuss can be used to analyse a range of different online services or platforms that today serve as key delivery mechanisms for works of culture, including YouTube, Netflix as well as various platforms for e-books or academic articles.Although our analysis is specific, the methods we propose are of more general relevance and concern.For example, using bots as research informants can be deployed for many different types of digital scholarship.Due to the transformation of media into data, digital methods can easily be used in research (albeit with some coding skills).When media at online services (such as Spotify) are coded and redefined as a purely data-driven communication form-with, on the one hand, content (e.g., media files and metadata) being aggregated through external intermediaries, and, on the other hand, user-generated data being extracted from listening habits-the singularity of the media experience is transformed and blended into what Jeremy Wade Morris has termed "a multimediated computing experience" (Wade Morris 2015: 191).
For a regular user, today's multimediated and exceedingly computational experience of online media takes on different and sometimes personalised forms.To understand the logic and rationale of contemporary media services and platforms, one should not shy away from but rather ask what exactly happens when data are turned into media and vice versa.What occurs and takes place beneath the black shiny surface of, say, the Spotify desktop client, with its green and greyish interface details and whited fonts and textures?It goes without saying, that research on the cultural implications of software-whether in the form of software studies, digital humanities, platform studies, or media archaeology-has repeatedly stressed the need for in-depth investigations on how computing technologies work combined with (more or less) meticulous descriptions of technical specificities (Kirschenbaum 2008, Chun 2011, Sterne, 2012, Ernst 2013).
Localising Spotify
Departing from the interventionist and experimental approaches we have used in our research project, which both metaphorically and practically try to track and follow the transformation of audio files into streamed experiences in the simple way a postman would follow the route of a parcel from packaging to delivery, the notion of localisation has become salient.Following files is a technical impossibility in a streaming media context, yet approaching, encircling, and circumscribing Spotify, both as a company and a service, has also proven to be hard.In our research project, we have repeatedly asked insidiously simple questions: Where is Spotify?When is Spotify used?What is Spotify?It might seem naive, but during the research process it has become increasingly difficult for us to understand and grasp our object of study.
Asking Google the search question "What is a Spotify?" returns a snippet from Wikipedia: "Spotify is a music, podcast, and video streaming service, officially launched on 7 October 2008.It is developed by start-up Spotify AB in Stockholm, Sweden" (Wikipedia 2017).But such an answer hides more than it shows and can easily be problematized.Is Spotify, for example, a content platform, a distribution service, or a media company?Furthermore, music naturally lies at the heart of Spotify (even if podcasts and videos seem increasingly important), but what kind of content is accepted-i.e., how is music defined?And what about the Swedishness of Spotify?Where is the company located?Headquarters are still to be found in central Stockholm on Birger Jarlsgatan 61, but the service is now available in some 60 countries, not to mention the digital variety of desktop and mobile versions (which all differ slightly).In addition, how does one situate Spotify commercially and financially (i.e., how much money is Spotify making (or losing) and how can one measure its economic impact?
As is apparent from the four issues above-and one could easily have included yet another-localising Spotify is easier said than done.Starting, however, by determining whether Spotify is a tech or a media company, it was obvious that Spotify for several years foremost offered a technological solution for record companies struggling with piracy.In a private conversation in 2012, one of the authors of this introduction (Snickars) asked Sophia Bendz (at the time Head of Marketing at Spotify) what kind of company Spotify actually was.Without hesitating, Bendz stated that Spotify was a tech company, only distributing content produced by others.The tech identity, however, was somewhat dubious even in 2012 and has become increasingly harder to sustain.Advertisement serves an illustrative case in point.In endless discussions with record labels (around rights management), Spotify took the stance that the continuous offering of a zero-price version with recurrent advertisement (Spotify Free) would in the long run be the best solution, as this strategy would serve as an incentive to scale businesses and attract glo-
Culture Unbound
Journal of Current Cultural Research bal listeners.Spotify's classification as strictly a tech company misses the fact that a core part of its business has been to provide content to audiences and selling those audiences to advertisers.Other music streaming services used a different strategy and Spotify has consequently struggled, and increasingly become more of a media company, all in order to keep to its business model with two versions of the product: the Free version (with embedded advertising) and the Premium version (without advertising).
Arguably, the music industry still sees Spotify as the top streaming service around, yet Spotify "has done little to address the lack of new music from a large collection of major artists when their albums are released" (Singleton 2016).That is, in a digital environment where streaming music becomes default, a focus on tech and distribution will only result in missed business opportunities.Indeed, Spotify has not really entered into content production (e.g., like Netflix), although some self-made videos are provided such as interviews with artists as well as other content (e.g., pop-ups that explain lyrics).Hence, stating that Spotify is only a tech company (in the form of a streaming service) fails to see other defining characteristics of the enterprise.
Secondly, "Music for everyone" is the company catch phrase, displayed, for example, when entering spotify.com.To localise Spotify, one might ask what kind of music does the service offer?In fact, one fundamental question we have struggled with in our research project is determining what sounds are perceived as music according to Spotify.It should be stressed that uploading music onto the service is outsourced to several so-called aggregation services.In short, these (and not Spotify) regulate content appearing on different music streaming platforms.In one of our interventions, we experimented with uploading self-produced music via different aggregators.These explorations with artificial sounds and music resulted in different responses.The same music (or sounds) passed some aggregators, but others did not define these "sounds" as music content at all.In short, rejection criteria of music aggregators turned out to be arbitrary.Hence, when principles as to what is considered music vary at the aggregation level, and consequently on streaming platforms such as Spotify, usually depending on whether users pay an aggregation fee or not, the line between music and non-music, artist and machine, becomes increasingly blurred.
A third way to use the notion of localisation to pinpoint Spotify is to look closer at geography and the hype around the "Swedishness" of Spotify.On the one hand, the company is still often associated with Sweden: "Swedish music-streaming service provider Spotify is in advanced talks to acquire German rival SoundCloud" (The Guardian, 2016).Yet, on the other hand, geographical localisation strategies also make it apparent that Spotify tries hard to transform itself into a global media company: "Spotify is tailoring its service for local tastes, from topical playlists to
Journal of Current Cultural Research
tiered pricing, as it prepares to expand its music streaming in Asia" (Bloomberg 2016).Spotify, in fact, increasingly acts as a global media company, and as a result, Patrick Vonderau (one of the researchers in our project) has recently claimed that "Spotify is neither particularly Swedish nor about music".While invocations of the company's Swedishness have been needed to sustain venture capital, and a "vision of 'European unicorns' . . . to position Spotify at the sexy, cool end of digital innovation", Vonderau argues that in financial terms Spotify now acts more "as a digital broker whose history of equity rounds, market and debt capitalization, and board of directors firmly ties brokerage strategies to U.S.-based financial interests" (Vonderau 2017).Spotify, in short, operates increasingly like a traditional American media company.
A fourth way to try to frame and localise Spotify is to follow the money and look at the company's evasive finances.Some figures estimate that the company makes more than two billion dollars a year from subscription fees and advertising, yet approximately 80 percent of that income is (all likely) paid to record labels and artists.In general, the financial situation and status of Spotify remains concealed, yet the same basically goes for the commodity that is being sold.As Rasmus Fleischer argues in his article in this thematic section, a crucial issue when dealing with the political economy of digital media is understanding what kind of commodity is being sold and to whom.
Lately, it has even been claimed that Spotify is "causing a major problem for economists" (Edwards 2016).Within mainstream economics, it is now commonly acknowledged that GDP is just an empirical construct that is becoming ever more misleading (Coyle 2014, Economist 2016).One main problem is how to measure inflation: to establish a price index, it is necessary to quantify differences in quality between last year's products and this year's products.It is difficult to compare the price of music sold as discrete units and music bundled as a monthly subscription (Spotify Premium) or offered with advertisements (Spotify Free).Is it meaningful to calculate a hypothetical "price per track listened to" in any of these cases?And how should we measure, in monetary terms, the value of music recommendations?Because of such quandaries, economists like Erik Brynjolfsson and Andrew McAfee have pointed to Spotify as an example of how national accounts fail to capture the "consumer surplus" resulting from rapid technological progress (Brynjolfsson & McAfee 2014: 174-189).Even a more traditional calculation of national accounts, which only includes those transactions where money is changing hands, poses delicate problems when locating Spotify.Thus, recent government inquiries from Sweden and the U.K. have singled out Spotify as the epitome of problems with measuring an economy increasingly built on digital services (Felländer 2015; Bean 2016).It seems that Spotify has not only disrupted the music and media industries but also has disrupted the ways in which the economic sta-
Culture Unbound
Journal of Current Cultural Research tistics surrounding user data need to be measured and interpreted.
Historicising Spotify
The story of Spotify is commonly told as an extraordinary success story: over 100 million users and over $8 billion valuation and growing.However, Spotify has yet to show a profit.So far, its losses have tended to grow faster than its turnover, so the survival of the service depends on ever larger injections of venture capital.This situation, typical for today's technology start-ups, tends to limit the opportunities for independent research.To attract investment and to secure deals with partner companies, it is necessary for Spotify to maintain a certain level of buzz in the news media, confirming the image of a company always expanding, always innovating, and always headed on a straight path towards a future monopoly position.No information will be let out if it does not play a predefined role in this public relations strategy.
One might argue that the buzz and hype, including problems in localising the company, makes it difficult for researchers to approach Spotify, at least compared to more established companies that have already gone public.Throughout most of Spotify's lifetime, there have been speculations about an imminent stock market launch, an IPO (Initial Public Offering), or a possible acquisition in which Spotify would be bought up by Google, Apple, or Facebook.Certain commentators have also questioned whether Spotify's business model is sustainable.These discussions and speculations have not lead anywhere and often remain obscure as vital details are kept secret via nondisclosure agreements between Spotify and the music industry.Another impossible (but lively) discussion has been concerned with whether Spotify is good for artists, as if artists exist as a homogenous group to which Spotify can be either good or bad.
From our research perspective, it is more relevant to ask how Spotify takes part in a redefinition of what it means to be a successful artist or a record company by changing the ways in which music is presented, commodified, and valuated.In other words, the producer of musical recordings cannot be thought of as existing independently of the distributor.As researchers, we must simply acknowledge that Spotify is a moving object and that the results from our digital experiments and interventions must be situated within a historical context (even though the company is not much older than ten years).One important source material for the historiography of Spotify, which has been essential for our research, is a major archive of news reports, including trade journals focusing on tech (e.g., Wired and Techcrunch), music (e.g., Billboard and Music Week) and advertising (e.g., Advertising Age and Marketing Week), all sources we have constantly been collecting.
Going through this archive, one is confronted by an immense level of buzz,
Culture Unbound
Journal of Current Cultural Research speculations, rumours, and empty promises.Localising and historicising Spotify is in many ways a task of how one approaches this constant murmur.One possibility is to regard this buzz simply as a kind of noise that ought to be filtered out, leaving a smaller selection of verified stories, useful for producing a historiography over what Spotify has really done.We propose the opposite approach, however: Just as we follow the files using digital methods, we follow the buzz using archives (i.e., our historiography).This means working through a tremendous source material looking not only for what happened, but also after what Richard Barbrook has described as "the beta version of a science fiction dream: the imaginary future" ( 2007).The history of Spotify is, in fact, full of false predictions and visions.Taking these shortcomings into account provides an important corrective to the conventional narrative about the gradual realisation of a grandiose entrepreneurial vision.It may surely be true that Spotify CEO, Daniel Ek, has a deep passion for music and that he enjoys playing the guitar, but when he and Martin Lorentzon founded Spotify in 2006, it was certainly not an attempt to disrupt the music industry to save it from piracy, as the official story now goes (Bertoni 2012) The original idea behind Spotify was purely technological: to create a platform for media distribution based on a peer-to-peer network.The first news reports in Sweden, in fact, presented Spotify as a company building a new infrastructure for film distribution.However, because video demanded too much bandwidth, Spotify's first set up and trials used music files as distribution content (Åkesson 2007(Åkesson , Johansson 2015)).To be more precise, the beta version of Spotify was loaded with pirated music files, downloaded by its employees through file-sharing services like The Pirate Bay (Andersson Schwarz 2013: 149).Music streaming proved attractive, and soon enough Ek and Lorentzon had conceived a business model for music, clearly inspired by the popularity of illicit file-sharing in Sweden.Spotify was to make music free but legal, available to consumers at no cost, while advertising provided all revenues.
Spotify's launch, thus, coincided perfectly with the broader hype around the idea that "$0.00Is the Future of Business" (Anderson 2008, Fleischer 2017), but also with the onset of a global financial crisis, which was soon to decimate the advertising market, making it hard to sustain ad-funded "free" services.The business of selling subscriptions for media services, however, tended to do remarkably well in the recession (Economist 2009).Spotify hence gradually changed its mind, now declaring that both advertising and subscriptions were to be equally important sides of their business model, while also dabbling with ideas of making money on sales of merchandise and concert tickets.In retrospect, it is striking how long the founders of Spotify resisted the idea of building a business fully dependent on subscription revenues.
Journal of Current Cultural Research
Historiography cannot do without an element of periodisation.With respect to Spotify's financial uncertainty and its dependence on venture capital, the company history can thus be understood over a timeline of investments.These have come in a series of funding rounds, from the first round (Series A) of about $20 million to the most recent round of $1 billion in convertible debt.Each time, the value of existing stocks has been diluted, the balance of ownership displaced in a new direction.The identity of the investors is usually public information, aggregated on websites like Crunchbase ( 2016), but the conditions detailed in each deal is always a secret.However, if one follows the buzz and maps it over the investment timeline, some of it becomes evident.Investments have, for example, been used mostly for international expansion (Series D, Series F) or for developing the streaming service in a specific direction (Series E).
Daniel Ek has been dubbed "the most important man in music" by Forbes (Bertoni 2012) and one of the ten most powerful people in the music industry (Billboard 2016), yet he is not in control of Spotify.The company's founders most certainly lost their majority share by 2009.In addition, Spotify's existence remains dependent on the willingness of the Big Three record labels (Universal Music Group, Sony Music Entertainment, and Warner Music Group) to renew their licensing deals.Hence there are several reasons why Spotify is not like Facebook: it is not profitable, it is not publicly traded, and it cannot dictate the terms in dealing with content providers.It would be silly to deny that Spotify is not dominant and mighty, but the power of Spotify is not easily located.Rather than being a single forceful actor trying to shape the future of music, Spotify indeed exists at the intersection of competing industries (tech, content, advertising, and finance).
One way to historicise Spotify in a more concrete manner is to look at altered strategies for music discovery.In the earlier period before its U.S. launch, Spotify's interface was centred around the search box (Fleischer 2015).Not much effort was put into assisting users who did not immediately know what music they wanted to hear.In other words, Spotify's ideal user was an individual with strong musical preferences (as part of his or her identity).When asked about the lack of social features in 2009, a Spotify director simply answered: "We're coming at it from the on-demand side" (Music Week 2009).This was also Spotify's real strength, according to influential magazines like Billboard and Wired; the service was considered fast, clean, and easy to use, and importantly so because it did not push music recommendations to its users (Bruno 2009, Peoples 2010, Pollack 2011).
This partly began to change in 2010-11, when Spotify established a strategic partnership with Facebook, following a Series C investment by Sean Parker (co-founder of Facebook and, before that, of Napster) who also joined Spotify's board of directors.The interface was gradually redesigned, moving away from the individualism of the search box and towards more social approaches of friction-
Culture Unbound
Journal of Current Cultural Research less sharing: all music listening would be automatically shared with friends.This was met, however, by an outcry from many users, forcing Spotify to introduce new options for protecting the privacy of musical preferences (Spotify 2011a, Spotify 2011b, Financial Times 2011).In short, the social turn provided a new direction for Spotify's developers, moving away from the poverty of the empty search box and towards a third way, different from both algorithmic and expert-curated music recommendations (Fleischer 2017).By integrating with Facebook, Spotify hoped to create the ultimate discovery engine.Spotify's approach was to recommend music based on what the user's friends had put in their playlists.Friends, however, can have bad taste.Ultimately, social discovery turned out to be a failure in the light of Spotify's experience on the U.S. market.Spotify had emphasised the freedom to choose, but many Americans seemed to prefer the freedom from choice.By the end of 2012, Daniel Ek admitted that "Spotify is great when you know what music you want to listen to, but not so great when you don't" (Bercovici 2012).
Spotify's social turn was followed, just a couple of years later, by a curatorial turn.The development of this type of new music discovery approach (throughout 2013) was financed by a $100 million investment round (series E) led by Goldman Sachs.Spotify was indeed not a vanguard in this movement.During 2012, industry observers began establishing as a fact that people love to simply lean back and listen.The future of streaming music was now more commonly sought in radio-like lean-back services such as Pandora, while the lean-forward approach of Spotify was seen as its Achilles' heel (Peoples 2012, Warren 2012).Trying to remedy this, Spotify first acquired Tunigo, a company specialised in building expert-curated music playlists.At the same time, Spotify discarded its old, individualist slogan: "Whatever you want, whenever you want it" (Spotify 2011c).New slogans were put in use: "Music for every moment" (Spotify 2013a) and "Soundtrack your life" (Spotify 2013b).In every country where Spotify was active, the local office began to recruit playlist curators with knowledge of local culture, but not specialised in any specific genres.The standard job description used was typical of Spotify's new approach: playlist curators should identify "songs to fit different situations" and create "playlist listening experiences for a multitude of moods, moments, and genres" (Spotify 2014).Here, it seems that Spotify had opted for a more human approach of expert curation, but Spotify was simultaneously working on algorithmic recommendation systems in close cooperation with the music intelligence company The Echo Nest, which it acquired in 2014.Neither a purely human nor a purely algorithmic curation system would be conceivable, but a combination of the two could work.In any case, it is finally interesting to note how this dichotomy was reinforced in 2015 by Apple when it presented its new streaming music service.Apple Music was then framed as the more warm and human alternative to the allegedly cold and all-too-algorithmic Spotify (Apple 2015, Dredge 2015).
About the Articles
As is evident from the discussions above, analysing Spotify is not an easy task.If localising Spotify is hard, historicizing the company's whereabouts doesn't result in a particularly straight trajectory either.On the contrary, users, competitors, and investors have all influenced the different directions that Spotify has taken and will all likely continue to do so.Hence, if music discovery today is important for Spotify to both satisfy and create a desire to consume and listen to more music, discovering Spotify is another matter.This thematic section of Culture Unbound, however, tries to locate the streaming service from several different perspectives.It brings together ongoing and differentiated research within the project "Streaming Heritage: Following Files in Digital Music Distribution".The four articles presented are, in short, all concerned with uncovering and finding out more about Spotify via different research strategies and methods.Three of the articles use digital methods in their approach, trying to get closer to Spotify through inventive experiments.Two of the longer articles (Eriksson & Johansson and Snickars) also explicitly use bots as research informants.A bot is a small software application that runs automated tasks (or scripts), and within interventions at Humlab we have repeatedly used massive set-ups of bots, sometimes working with up to 500 virtual listeners.
In the first article in the thematic section, "If the song has no price, is it still a commodity?",Rasmus Fleischer reviews some of the recent literature on how music is marketed.Over the last century, music has been subject to different regimes of commodification, sold as a published score, as a live performance, or as recorded sound.Streaming services like Spotify, however, represent a different commodification regime, Fleischer argues.Therefore, it is necessary to identify and define the commodity Spotify sells.Fleischer criticises prevalent conceptions of the digital music commodity that often assume that each song (whether downloaded or streamed) is a commodity, which is indeed correct in the case of downloading services like the iTunes Store.But the user of Spotify will (currently) never see a price tag on a song.In fact, Spotify is not selling discrete pieces of recorded sound and is not offering consumers millions of commodities; Spotify offer only one commodity: the subscription.This product is a bundle that includes not only access to all songs in the catalogue, but also the maintenance of a personalised profile connected to a variety of playlists tailored for pre-defined activities.Music is still commodified by Spotify, Fleischer argues, but as a commodity, music can mean different things.Spotify is, for example, buying music through various aggregation services in the form av copyright licenses, bundling it, adding new features, and then selling music as a personalised experience.When analysing commodification, it is always necessary to ask what kind of object is the commodity.
Journal of Current Cultural Research
In their article, "Tracking Gendered Streams", Maria Eriksson and Anna Johansson investigate whether music recommendations at Spotify are gendered.As is well known, one of the most prominent features on contemporary music services is the provision of personalised music recommendations that come about through the profiling of users and audiences.Based on a range of bot experiments, their article explores patterns in music recommendations provided by Spotify in its Discover feature.The article specifically focuses on issues around gender and explores whether the Spotify client and its music recommendation algorithms are performative of gendered user identities and taste constellations.Exploring the tension between gendered publics and Spotify's promise to deliver personalised music recommendations to everyone, Eriksson and Johansson's research ties into broader questions about the workings and effects of algorithmic knowledge production.They argue that issues around gender are important in this context, since Spotify's music recommendations can be considered as one of the venues where gendered norms and ideals are reproduced and manifested.Eriksson and Johansson's results for example reveal that male artists were highly overrepresented in Spotify's music recommendations; an issue which they argue prompts users to reproduce hegemonic masculine norms within the music industries.Although the results should be approached as highly historically and contextually contingent, Eriksson and Johansson argue that they do give some evidence of the ways in which gender becomes tied to issues of taste and identity formation in algorithmic knowledge-making processes.
In his article, "More of the Same -On Spotify Radio", Pelle Snickars takes a similar approach as Eriksson and Johansson, working extensively with bots as research informants.Snickars main interest is the so-called radio function at streaming services, and Spotify Radio in particular.It is a service that "lets you sit back and listen to music you love.The more you personalise the stations to match your tastes the better they get", at least according to the company slogan.Basically, the radio functionality allows users (via various unknown algorithms) to find new music within Spotify's vast back-catalogue, offering a potential infinite avenue of discovery.Nevertheless, the radio service has also been disliked and blamed for playing the same artists over and over.Together with the Humlab programmers, Snickars set up an experiment to explore the possible limitations and restrains found within "infinite archives" of music streaming services.The hypothesis was that the radio function of Spotify does not consist of an infinite series of songs although it may appear so to the listener; it is actually a finite loop.Spotify Radio claims to be personalised and never-ending, yet music seems to be delivered in limited loop patterns.What would such loop patterns look like?The intervention used 160 bot listeners programmed to listen to different Swedish music from the 1970s.Snickars is not primarily interested in personalised recommendations,
Culture Unbound
Journal of Current Cultural Research but rather how Spotify Radio functions generically.The first (and major) round of bots started Spotify Radio based on the highly popular Abba song "Dancing Queen" (with some 65 million streams).The second (and minor) round of bots used the less well-known Swedish progressive rock band Råg i Ryggen's "Queen of Darkness" (with some 10,000 streams).Snickars article describes different research strategies when dealing with proprietary data as well as the background and the establishment of the radio functionality at streaming services like Spotify.Essentially, his article empirically recounts, discusses, and analyses the radio looping interventions set up at Humlab.
Finally, in their co-written article, "Studying Ad Targeting with Digital Methods: The Case of Spotify", Patrick Vonderau and Roger Mähler provide a brief description of digital methods used in studying digital advertising technologies.To study ad targeting, researchers have an inventory of tested methods at their disposal but a problem of access to verifiable data persists.In order to understand which types of key stakeholders are involved in ad targeting processes, the authors experimented with digital tools to complement data collection.In doing so, they followed the well-established idea of taking up methods that are already embedded in digital infrastructures and practices.
This thematic section of Culture Unbound goes under the hood of Spotify and looks critically at its tech stack.It is important to remember that Spotify's data infrastructure resembles other services.The analyses put forth in the different articles (sometimes) approximates media specific readings of the computational base; that is, the mathematical structures underlying various interfaces and surfaces resonate with media scholarly interests in technically rigorous ways of understanding the operations of material technologies.Then again, it is also important to stress that the Spotify infrastructure is hardly a uniform platform.Rather it is downright traversed by unseen data flows, file transfers, and information retrieval in all kinds of directions, be they metadata traffic identifying music, aggregation of audio content, playout of streaming audio formats (in different quality ratings), programmatic advertising (modelled on finance's stock exchanges), or interactions with other services (notably social media platforms).This thematic section tries to uncover and make visible some of these streams.
Rasmus Fleischer is a postdoctoral researcher based at the Department of Economic History, Stockholm University.His research interests are located in the intersections between culture and economy, as well as technology and politics.Most of all, he has explored 20th century media history, transformations of copyright and the commodification of music.E-mail<EMAIL_ADDRESS>
Journal of Current Cultural Research
Pelle Snickars is Professor of Media and Communication Studies, specialising in digital humanities at Umeå university and is affiliated with the Humlab research centre.His research focuses on the relationship between old and new media, media economy, digitisation of cultural heritage, media history as well as the importance of new technical infrastructures for the humanities.E-mail<EMAIL_ADDRESS>
Notes
ii The research project, "Streaming Heritage.Following Files in Digital Music Distribution" involves system developers Roger Mähler and Johan von Boer (at Humlab, Umeå University), as well as researchers Pelle Snickars, Maria Eriksson, Anna Johansson (at Umeå University), Rasmus Fleischer (at Stockholm and Umeå University), and Patrick Vonderau (at Stockholm University).For more information: http://streamingheritage.se/. | 8,290.8 | 2017-10-31T00:00:00.000 | [
"Computer Science",
"Business"
] |
Detecting Violent Radical Accounts on Twitter
In the past few years and as a result of the enormous spreading of social media platforms worldwide, many radical groups tried to invade social media cyber space in order to disseminate their ideologies and destructive plans. This brutal invasion to society daily life style must be resisted as social media networks are interacted with on daily basis. As some violent radical groups such as ISIS has developed well designed propaganda strategies that enables them to recruit more members and supporters all over the world using social media facilities. So it is crucial to find an efficient way to detect the violent-radical accounts in social media networks. In this paper, an intelligent system that autonomously detects ISIS online community in Twitter social media platform is proposed. The proposed system analyzes both linguistic features and behavioral features such as hashtags, mentions and who they follow. The system consists of two main sub-systems, namely the crawling and the inquiring subsystems. The crawling subsystem uses the initially known ISIS-related accounts to establish an ISISaccount detector. The inquiring subsystem aims to detect Pro
I. INTRODUCTION
Social media become an essential part of everyone's life style nowadays. Everyone now can easily express his thoughts, feelings and even his emotions through internet and these ideas will spread within seconds among the world. These posts may be viewed by millions who will interact with them. Online Social Network (OSN) has grown significantly over the past decade. Within a population of 7.676 billion human being which is the entire population on earth, there are 4.388 billion Internet users, 3.256 billion smart phones and 3.484 billion active social media users in 2019 [1]. Its existence gave humanity the gift of spreading civilization, literature, science, arts, and others which are means of fulfilling prosperity and welfare all over the world. On the other hand, if this power was misused, many unpleasant sequences may occur. Hatred may be spread instead of constructive ideas, violence instead of literature and science, war instead of welfare.
Many extremist radical hate groups and violent associations are consistently trying to spread their ideologies and hate speech through various social media platforms. In other words, these groups are using the social media facilities to recruit new members and to distribute destructive ideas and plans. It is very important to identify the members of these groups to prevent them from spreading their harmful ideologies, disseminating violence and hatred on social media platforms that may result war and conflicts in peaceful societies. A living example on these radical groups is Islamic State of Iraq and al-Sham "ISIS", also is known as "Daesh" in Middle East. It is recognized by its adherence to the fundamentalist Salafi faith of Sunni Islam [2]. It gained international fame in early 2014 when it expelled Iraqi government forces from the main cities of Western Iraq then seizing Mosul and committing the massacre of Sinjar. Since its appearance, "ISIS" is continuously trying to leverage its ideology through social media platforms.
Twitter social media is one of the most popular online social media networks in the world with 330 million monthly active users and 500 million tweets per day. A tweet is a message composed of 140 characters that any user can easily share among millions of accounts. Tweets may contain hashtags that highlight the tweet's main topic. Although there is no accurate statistics shows the existence of ISIS members but in 2014, it was an expected that from 46,000 to 90,000 Twitter accounts that upheld for ISIS or they were controlled by its supporters. Over 1.2 million accounts suspended for terrorist content since August 2015 [3]. Suspension criteria are mainly based on other accounts reporting an account that generates violent radical tweets. However, inspecting the posts of a single user in a social media application such as Twitter is a tedious work. The need for an artificial intelligent approach that mimics the human inspection to solve this problem is a must.
In this paper, a new architecture that can autonomously detect ISIS members' community in the Twitter social media network is proposed. The proposed system analyzes both linguistic features and behavioral features such as hashtags, mentions and the following lists. These will enable us to continuously have a live image of how ISIS members interact with online social media. The rest of the paper is organized as follows: Section 2 reviews the previous studies and related work. Section 3 clarifies dataset gathering and analysis. Section 4 introduces the proposed architecture. Section 5 discusses the experimental results. Finally, Section 6 includes the conclusions and future work.
II. RELATED WORK ISIS existence in twitter has been expanded enormously that catch the concerns of the international society. That resulted the appearing of some online groups volunteering their time and resources to report those violent accounts. One of these groups was Ctrl-sec [4] that was responsible of deactivation of about 25,000 twitter accounts in three years by identifying radical accounts manually [5]. Others could suspend about 25,000 Pro-ISIS accounts through crowd-sourced reporting [3]. Many researchers were attracted to this topic. They are trying to find new means to discover and limit the vast spread of these accounts. Ashcroft et al. [6] adopted machine learning techniques based on a list of English predefined hashtags related to ISIS to detect ISIS related tweets using three feature classes, namely sentiment, temporal and stylometric features. He found that his approach is highly dependent on data and the approach of detecting radical content should always be a helpful tool to support humans to asset accounts not to replace them. Choudhary et al. [7] tried to detect behavior patterns and key player features to identify terrorist community.
In a trial to understand what happens through the phase of an account being converted from an ordinary account to a Pro-ISIS one sharing Pro-ISIS content, Rowe and Saif [8] defining the Pro-ISIS account by the account that shares more radical Pro-ISIS content than the Anti-ISIS one. Although this approach seems effective but couldn't deal with the lexical diversity.
Klausen et al. [9] studied the communication flow among ISIS members on twitter using 59 manually assessed Pro-ISIS accounts and found that female members has a key role in ISIS propaganda. Carter et al. [10] found that newly bounded ISIS members seek guidance through online spiritual ISIS figures on twitter. Chatfield et al. [11] investigated how ISIS recruits new members through disseminating their ideology on social media platforms. Also Vergani et al. [12] studied how ISIS uses emotional speech and religious quotes to recruit online supporters. Berger et al. [13] found that twitter users who follow ISIS members are highly affected by their ideology. Saif et al. [14] found that semantic features based models out performs other lexical, topic and sentiment based models in detecting Pro-ISIS accounts. Berger et al. [15] found that Pro-ISIS accounts can be identified through their profile description. Agarwal et al. [16] expressed the presence of offensive, war and hate speech terms in ISIS propaganda.
A. Data Gathering
As the research objective is to automatically detect Arabic speaking violent radical accounts on twitter, dataset should be found and properly cleaned to be analyzed, studied and to apply various machine learning algorithms on it. Although native ISIS members are Arabic speakers as they mainly located in Syria and Iraq, no Arabic ISIS related dataset was found due to the lack of proper Arabic resources in science society especially in machine learning field. In order to prepare a suitable Arabic dataset that can be used in this research, two approaches for collecting data were adopted. The first approach is to collect proper dataset from Arabic speaking twitter accounts that represent Pro-ISIS, Anti-ISIS and non-ISIS. This dataset is referred to hereafter as the collected dataset. The second approach is to collect the published available non-Arabic ISIS related datasets that can be found in online datascience communities and translate them. This dataset is referred to hereafter as the translated dataset.
1) The collected dataset By studying and examining extremist accounts in Twitter, the most frequent hashtags in ISIS propaganda were manually identified such as اﻟﺪوﻟﺔ وأﻋﺪوا, ﺑﺎﻗﯿﺔ, _اﻹﺳﻼﻣﯿﺔ, ﺗﺘﻤﺪد . Using these hashtags, 42 accounts were collected and annotated as the most violent and ISIS influential accounts. Using twitter API [17] the tweets feed of the annotated accounts were collected which resulted to downloading of 21,000 tweets. These accounts will be referenced as "collected Pro-ISIS accounts". Similarly and in order to collect balanced dataset 21,000 Anti-ISIS tweets were gathered in the same way using the following hashtags such as ﺑﻠﻎ داﻋﺶ, ﻓﻀﺎﺋﺢ داﻋﺶ, ﺿﺪ ﻣﺴﻠﻤﻮن اﻟﺪم, ﺗﺠﺎر داﻋﺶ داﻋﺶ, ﺟﺮاﺋﻢ داﻋﺶ ﻋﻦ as they were the most used hashtags in the most active Anti-ISIS accounts namely "collected Anti-ISIS accounts". 21,000 random non-ISIS related tweets were collected as-well from different domains: "News -Religion -Sports -Art" to represent "collected non-ISIS related accounts". Data preprocessing such as URL and mentions removal, discarding non alpha characters such as (@, #, $, %, _), characters normalization such as ئ( ؤ, آ, إ, ),أ, and stop words removal such as إﻟﻰ( ﻣﻦ, أى, ﻋﻠﻰ, )ﻓﻰ, , tashkeel removal such as ( َ◌, ً◌, ّ◌, ٌ◌, ٍ◌ ) and prefix/suffix removal such as و( ﻧﺎ, ﻧﻰ, )اﻟـ, was applied to prepare the collected dataset for training stage.
2) The translated dataset By searching online data-science communities such as "Kaggle", three non-Arabic ISIS related datasets were found: a) Fifth-tribe "How ISIS Uses Twitter" dataset [18]. It is consisted of 17,000 tweets was collected from 112 Pro-ISIS twitter accounts from all over the world that supported 2015 terrorist Paris attacks [19]. These tweets are mostly written in English.
b) "Religious Texts Used By ISIS" [20]. 2,685 religious texts dataset which was collected by scrapping 24 issues of Dabiq and Rumiyah English-based ISIS magazines that ISIS uses to spread their ideology in Europe and western world.
c) "Tweets Targeting ISIS" dataset which contained 122,000 ISIS related tweets was collected all over the world in many languages, mostly in English [18]. These tweets were collected by following ISIS related terms such as (ISIS, Daesh, Islamic State, Raqqa, Mosul) 13,000 tweets that were against ISIS ideology and terrorism were translated into Arabic.
Translating into Arabic language was carried out by custom python scripts using Google translating service [21]. Translated dataset was manually reviewed to correct mistranslated words/expressions. Finally, data pre-processing steps were carried out including data cleaning, normalization, stop words removal and stemming as mentioned in the collected dataset subsection.
B. Text Features Vectoring
Dataset was collected from 2-main different sources collected dataset and translated one. These two datasets may suffer from the domain divergence problem [22]. To make sure of suitability of applying any recognition and detection technique for both of these datasets, visual testing was adopted. In the first step to carry out this test, the datasets are first represented as vectors. Then visualization techniques are used to represent the converted words. "Mazajak" word embedding 517 | P a g e www.ijacsa.thesai.org was used [23] to convert dataset corpus into vector domain. Mazajak is considered to be the largest Arabic word embedding models based on a corpus of 250 million tweets converted using skip-gram architecture [24]. In Mazajak, each word token is converted to a 300-D vector. Tweets can be represented by the mean of its contained word embedding's. Similarly the whole user's tweets thread can be represented by the mean of its vector tweets embedding.
The second step in the proposed visual testing is dimensionality reduction of the resulting vectors because it is not proper to visualize 300-D vectors. So embedding vectors dimensions have to be reduced in order to visualize and study the dataset. TSNE [25] is a machine learning technique for dimensionality reduction. This technique is applied here to reduce the vector dimensions from 300-D to just 2-D.
C. Text Features Analytics
After converting the dataset's corpus into a vector form using "Mazajak" word embedding model [23] where the distance between any two vectors is proportional to the difference in the meaning of the words they represent. The vectors should be plotted graphically so they can be studied and analyzed to get better understanding for dataset. That is how the consistency of the dataset can be checked as it was collected from multiple sources and to make sure that vectors that represent the same class can be clustered in spite of the lexical diversity between the collected and translated data. Fig. 1 illustrates both the collected and the translated tweets. The "collected Pro-ISIS accounts" tweets and both the translated "How ISIS Uses Twitter" tweets and the translated "Religious Texts Used By ISIS." These tweets were labeled as Pro-ISIS in "Red", Anti-ISIS in "Green" and Random in "Blue" colors. From Fig. 1, it can be noticed that some tweets from both Anti-ISIS and Pro-ISIS classes are overlapped with random tweets. This can be easily interpreted because ISIS related tweeters may have other interests in different topics such as sports and news. Also it is noticed that non-ISIS related tweets are distributed over a wide range of fields that belong to different domains such as "news -religion -sports -art". Non-ISIS related dataset should be collected in order to avoid detection accuracy degradation if the system is not trained to detect non-ISIS related tweets. Fig. 2 represents in "Red" ISIS related tweets from both classes Anti-ISIS and Pro-ISIS tweets, where non-ISIS related tweets are represented in "Blue". It is obvious that ISIS related class is clearly separated from non-ISIS related.
Finally the relation between Pro-ISIS and Anti-ISIS tweets has to be expressed. They may share the same vocabularies, as twitters from both sides (Pro and Anti ISIS) regularly discuss similar subjects in their tweets what will make it harder for us to separate between them. To gather appropriate tweets that represent these two classes, the translated 13,000 tweets from "Tweets Targeting ISIS" dataset and 18,000 manually labeled tweets from collected Anti-ISIS tweets were used to represent the Anti-ISIS data class. On the other side, the translated 12,000 tweets from "How ISIS Uses Twitter" plus 16,200 tweets from collected Pro-ISIS and 2,600 translated texts from "Religious Texts Used By ISIS" were used to represent Pro-ISIS data class. Fig. 3 shows that the two classes can't be separated linearly.
D. Behavioral Features Organization
In addition to the usage of the lexical features of the dataset, it is important to make use of other behavioral features got from the collected data such as mentions, hashtags, likes, retweets and follow lists. These features express the behavior of the user, as the user's used hashtags, retweets and likes defines the topics he is interested in. On the other hand, his mentions and follow list express his spiritual leaders. With the aid of unstructured Mongo database management system [26], the collected mentions, likes, retweets, hashtags and follow lists can be aggregated. Each of these features will have a score that will be identified with its commitment in ISIS community. If a certain hashtag or mention is found more often in ISIS propaganda, it will have higher score which will reflect how much it is related to ISIS. This facilitates capturing the topics they are keen on, how they influence their beliefs and who they follow. Finally a list of the most followed accounts, the most liked and retweeted tweets in addition to the most tweeted hashtags by ISIS supporters will be obtained. 518 | P a g e www.ijacsa.thesai.org
IV. PROPOSED ARCHITECTURE
Although Pro-ISIS propaganda is distinguished by the presence of offensive, war and hate speech terms [16], the lexicon itself varies according to the undergoing events [8]. They always use the world trending topics to ensure the wide spread of their propaganda. Other problem is that ISIS propaganda evolves during the process of recruitment itself within many stages that includes religious, emotional and hate speech [7]. These problems cause degradation in the detection accuracy on the long term for non-maintainable detection systems, as they rely on old detection models or outdated lexical dictionaries in the detection process.
The main challenge of the proposed system is to maintain it up to date autonomously. This includes tracking the changes in the online-behavior of ISIS members as well as tracking the evolution of their ideologies and propaganda plans. Moreover, it is essential to identify the fundamental key-players or profound pioneers that originate hate speech content. In a trial to overcome these challenges, the proposed architecture is designed to continuously crawl on ISIS online community updating its corpora and other semantic features such as used hashtags, mentions, who they follow and what they share. That will give us a live image of how ISIS behaves on Twitter social media platform. In order to invade ISIS online community on Twitter, an initial seed of ISIS related accounts will be needed which can be found in the collected dataset. The proposed architecture consists of two main subsystems, namely the crawling subsystem and the inquiring subsystem. The details of these subsystems are explained in the next two subsections. They are depicted separately in Fig. 4 and Fig. 5.
A. The Crawling Subsystem
The Crawling subsystem enables us to invade ISIS online Twitter community. Fig. 4 depicts the subsystem components. The crawling subsystem is started by targeting the most followed accounts from the follow lists in the collected dataset. Using Twitter's REST API [17], the account info, followers list and the last posted 3,000 tweets for each targeted account can be downloaded. The downloaded tweets will undergo data preprocessing steps. The steps include data cleaning, normalization, stop words removal and stemming prior inputting to ISIS-Content detector. ISIS-Content detector will detect ISIS related tweets in order to define for how far the user is involved into ISIS community. As the ISIS-Content detector will assess all of the downloaded tweets for each user, it will distinguish whether each tweet is Pro, Anti or non-related to ISIS in addition to updating hashtags and mention lists in the system's database if the tweet is predicted as Pro-ISIS. If the anticipated Pro-ISIS tweets ratio for the tested account surpassed certain ratio (Pro-ISIS threshold) the account is considered to be Pro-ISIS. Finally all the collected and analyzed data for each account will be stored in the unstructured MongoDB [26]. The collected database can be effortlessly used to investigate and aggregate most followed/active accounts, hashtags and mentions. Updated most followed ISIS accounts will be sent back to be cushioned as a contribution to the following cycle to keep updating the database. As a result of this subsystem, ISIS members, supporters and leaders can be easily tracked in addition to addressing hot topics undergoing in ISIS. Thus the gain of this subsystem is the expansion of the user's knowledge. Likewise an up to date dataset of ISIS-related Arabic labeled tweets will be continuously available that can be used in further studies.
B. The Inquiring Subsystem
Inquiring subsystem enables the user to inspect specific twitter account by twitter ID as an input. Using Twitter API [17], the system will download account data along with the latest posted 3,000 tweets. The downloaded tweets will undergo the same cleaning and data pre-processing steps before usage as in the crawling subsystem. Downloaded hashtags, mentions and following list will be calculated and correlated with the data stored in the system DB which was collected from crawling subsystem to calculate its behavioral features. Behavioral scores will be determined by contrasting the calculated behavioral features by the blacklisted data from crawling subsystem. Behavioral scores along with the pre-processed tweets will be the input to ISIS Accounts Detector which will detect whether the account is Pro-ISIS or not.
C. System Maintenance
In order to maintain the system up to date with highperformance and acceptable detection rates, it must be periodically updated and retrained. It should be introduced to undergoing ISIS related events and its members reactions on them, track their spiritual leaders and key players, learn about their ideologies and propaganda methods. As a result of continues pursue of ISIS community, a live image of how ISIS members spread their news, ideology and even guidance on social media will be obtained. Helpful reports of tracking ISIS members and leaders can be easily developed. Also a proper up to date dataset will be developed which can be used to periodically retrain our detectors on up to date data to avoid deprecation that causes degradation in detection accuracy.
V. EXPERIMENTAL RESULTS
The proposed system includes two pre-trained supervised classification detectors which are key nodes in the system. They detect whether the tweet is Pro-ISIS or not and whether the account is Pro-ISIS or not. Their accuracies define the overall performance of the system as they categorize and define the quality of the crawled data which is important in system evolution. So their accuracy should be boosted and lower down the possible False Positive or False Negative rates.
A. ISIS-Content Detector
ISIS-Content Detector should process on all of the downloaded tweets downloaded from crawling subsystem that targets ISIS community on Twitter. Although tracking ISIS members in the proposed system, also their tweets should be checked before labeling them where some of ISIS members may have other interests such as news, sports or religion so Pro-ISIS tweets obtained from stalked ISIS members should be inspected in order to increase the quality of the collected dataset.
Other task is to determine how far the stalked account owner is involved into ISIS community, as the ratio of his Pro-ISIS tweets to all of its tweets is calculated. If it exceeds a certain value, the account will be labeled as Pro-ISIS account in the collected database. So a supervised pre-trained detector on labeled data classes that represent Pro-ISIS, Anti-ISIS and non-ISIS related tweets should be prepared. In order to collect these three classes, the translated 13,000 tweets from "Tweets Targeting ISIS" dataset and 18,000 manually labeled tweets from collected Anti-ISIS tweets were used for representing the Anti-ISIS class. For Pro-ISIS data class 12,000 translated tweets from "How ISIS Uses Twitter" plus 16,200 tweets from collected Pro-ISIS and 2,600 translated texts from "Religious Texts Used By ISIS" were used. Finally, the collected 21,000 random tweets were used to represent the non-ISIS related class. Table I shows the results of the testing process. Table I shows that TF-IDF embedding outperforms skipgram embedding in all experiments. Linear SVC achieved accuracy 89% followed by Bernulli NB in "UG" TF-IDF [27] and Logistic Regression [27] in "TG" TF-IDF achieved accuracy 88% regarding F1-score [27]. Detector's accuracy can be boosted by using all the three detectors as a voting detector as an ensemble learning system [29]. So each tweet will be labeled as the majority of the three detectors decide which will avoid any weaknesses found in any individual one of them in addition to avoiding over fitting.
B. ISIS-Account Detector
ISIS-Account Detector is a key node in the proposed system. It should check if the account inquired by the system user is Pro-ISIS or not. Unlike ISIS-Content detector that operates on discrete tweets, Pro-ISIS accounts detector run on all account data including tweets as textual features, in addition to hashtags, mentions, likes, retweets and following list as behavioral features. Thus beside vectoring dataset corpus by word embedding techniques, other behavioral features such as used hashtags, mentions and following list were vectored by comparing them by the collected data from the crawling subsystem and assigning scores to them. In order to train this detector, two data classes that properly represent Pro-ISIS accounts and non-Pro-ISIS were created. The translated "How ISIS Uses Twitter" dataset in addition to the collected Pro-ISIS accounts were used to represent Pro-ISIS data class. On the other side the collected Anti-ISIS accounts plus collected non-ISIS related accounts were used to represent non-ISIS data class. Non-ISIS related accounts were introduced to the detector to avoid detection accuracy degradation if a non-ISIS related account was inquired in the system. Fig. 6 expresses the scattering of the accounts in the dataset. However Pro-ISIS and Anti-ISIS accounts may share the same lexicons as they are both involved into the same topics, Fig. 6 shows that Pro-ISIS accounts can easily be separated from both Anti-ISIS and non-ISIS-related accounts. Table II shows that this time skip-gram embedding [24] outperforms TF-IDF embedding. Both Linear SVC and Logistic Regression [27] achieved F1-score 94% followed by Decision Tree [27] by score 89%. Such as "ISIS-Content Detector", accuracy can be boosted and avoid over fitting by using all three in a voting ensemble learning system [29].
VI. CONCLUSIONS AND FUTURE WORK
A new intelligent system architecture that autonomously detects Pro ISIS-accounts in twitter social media platform is introduced. The system consists of two main sub-systems, namely the crawling and the inquiring subsystems. The kernels of the two subsystems are intelligent detectors. The attributes of these detectors are both linguistic features and behavioral features.
For proper testing the proposed system, a new collected dataset of Pro-ISIS, Anti-ISIS and non-ISIS-related accounts were gathered. Three English datasets about ISIS were translated to Arabic. All datasets were represented as vectors using "Mazajak" word embedding "skip-gram" scheme to 300-D vectors. Vectors dimensions were reduced into 2-D and plotted to check their consistency.
The intelligent detectors kernels for the crawling and the inquiring subsystems were developed using supervised machine learning techniques. The results show that ISIS-Content Detector gave best accuracy 89% according to f1score by linear SVM algorithm using TF-IDF embedding. They show also that ISIS-Account Detector gave pest accuracy 94% according to f1-score by linear SVM algorithm Skip-gram embedding.
As next steps, extension of the proposed architecture to other social media applications such as Facebook and Instagram is planned. Enlarge the used dataset to include different Arabic delegates. | 5,979.4 | 2020-01-01T00:00:00.000 | [
"Political Science",
"Computer Science"
] |
Numerical Solution of Second-Orders Fuzzy Linear Differential Equation
In this paper, the numerical solution of the boundary value problem that is two-order fuzzy linear differential equations is discussed. Based on the generalized Hukuhara difference, the fuzzy differential equation is converted into a fuzzy difference equation by means of decentralization. The numerical solution of the boundary value problem is obtained by calculating the fuzzy differential equation. Finally, an example is given to verify the effectiveness of the proposed method.
Introduction
Many engineering system problems are too complex to be directly converted into system equations to solve, and they often involve parameter uncertainties, which often appear as the fuzzy numbers. Therefore, when solving such problems, it is often necessary to model it to a fuzzy differential equation to consider, which always makes the solution of fuzzy system equations very important.
Prior to discussing fuzzy differential equations and their associated numerical algorithms, it is necessary to present an appropriate and brief introduction to the derivative of the fuzzy-valued function. The concept of a fuzzy derivative was first introduced by Zadeh [1], followed up by Dubois and Prade [2] who used the extension principle in their methods. Other fuzzy derivative concepts have been proposed by Puri and Ralescu [3] as an extension of the Hukuhara derivative of multivalued functions.
In recent years, many scholars have made profound research on two-order differential equations given the fuzzy boundary conditions. Wu, Q. [4] discussed the uncertainty of the two-point boundary value of the two-order differential equation. Using the fuzzy simulation principle and different methods, the numerical solution of the boundary value problem is obtained. Regan, O. et al. [5] proved a prior result on the solvability of fuzzy boundary value problems based on the generalized Schauder theorem. Wu, C.X. et al. [6] proved that the existence of analytical solutions of fuzzy boundary value problems completely depends on the definition, structure and properties of fuzzy numbers. Guo et al. [7] [8] studied the approximate solutions of two-order linear differential equations with several fuzzy boundary conditions. In this paper, we investigate the boundary value problem of two-order fuzzy linear differential equations. . We say that f is Hukuhara differential at 0 t , if there exists an element
Preliminaries
such that for all 0 h > sufficiently small, and the limits: Definition 2.6 [11]: The second-order fuzzy differential equation: is called the two-order fuzzy boundary value problems, where In this paper, we mainly study the boundary value problem of two-order fuzzy linear differential equations, under the condition of parameter numbering.
, , are fuzzy numbers, ( ) ( ) , p t q t are coefficient function. And parameter form is: y t r p t y t r q t y t r g t r y a r r y b r r y t r p t y t r q t y t r g t r y a r r y b r r
The Establishment of the Difference Method
For the boundary value problem of two-order fuzzy linear differential equations, we are going to discuss the establishment of the difference method and study its solvability in this section.
First, we use the following first-order difference quotient to approximate the first derivative ( ) f t ′ , which is: Then the second derivative can be approximated by the first-order difference quotient of the first-order difference quotient, that is, Finishing the following the ( ) ( ) where: Finishing the following the ( ) ( ) where: , , Proof First, we can eliminate the first-order difference in the fuzzy difference equation by appropriate transformation of the independent variables.
According to the results of the Negoita-Ralescu Characterization Theorem [12], Just prove that the corresponding homogeneous linear equations as:
Numerical Example
Example 4.1 Consider the boundary value problem of fuzzy differential equations as follows: Then, the solution of the boundary value problem of fuzzy differential equations is:
Conclusion
In this paper, an approximate method based on a positive basis for the undetermined coefficients second-orders fuzzy linear boundary value problems was discussed. A class of boundary condition and the general case were considered.
According to the sign of coefficient functions of the fuzzy linear differential equation, the corresponding function systems of linear equations were composed.
And then, fuzzy approximate solutions were obtained by solving a crisp function extended system of linear equations. For next investigation, we can consider the other class of boundary conditions by the same method. | 1,038.8 | 2021-01-01T00:00:00.000 | [
"Computer Science",
"Mathematics"
] |
Conditionally unbiased estimation in phase II/III clinical trials with early stopping for futility
Seamless phase II/III clinical trials combine traditional phases II and III into a single trial that is conducted in two stages, with stage 1 used to answer phase II objectives such as treatment selection and stage 2 used for the confirmatory analysis, which is a phase III objective. Although seamless phase II/III clinical trials are efficient because the confirmatory analysis includes phase II data from stage 1, inference can pose statistical challenges. In this paper, we consider point estimation following seamless phase II/III clinical trials in which stage 1 is used to select the most effective experimental treatment and to decide if, compared with a control, the trial should stop at stage 1 for futility. If the trial is not stopped, then the phase III confirmatory part of the trial involves evaluation of the selected most effective experimental treatment and the control. We have developed two new estimators for the treatment difference between these two treatments with the aim of reducing bias conditional on the treatment selection made and on the fact that the trial continues to stage 2. We have demonstrated the properties of these estimators using simulations. Copyright © 2013 John Wiley & Sons, Ltd.
Introduction
Modern innovations in clinical trial design have led to the availability of new approaches referred to as adaptive seamless designs (ASDs). Using an ASD, a clinical trial is conducted in 2 or more stages with interim analyses performed before the final stage to make adaptations. In this paper, we will consider two-stage ASDs where several doses or formulations of a drug, or several different treatments, are simultaneously compared with a standard/control with the poorly performing treatments dropped at stage 1 on the basis of interim analysis results. Such a trial is often termed a seamless phase II/III clinical trial. Unlike the traditional approach in which promising treatments are selected in a phase II trial separate to a confirmatory phase III trial, seamless phase II/III clinical trials combine aspects of both phases into a single trial with two or more stages. At the end of stage 1 of a two-stage seamless phase II/III clinical trial, an interim analysis is conducted to select the most promising treatment so that stage 1 resembles a phase II trial. The selected treatment together with the control treatment continues to stage 2 after which a confirmatory analysis is performed so that stage 2 resembles a fixed-sample-size phase III trial. The confirmatory analysis includes data from stages 1 and 2.
An ASD such as that described earlier poses a number of statistical challenges in both hypothesis testing and estimation of treatment effects because at the end of the trial, they use the data used in treatment selection to make inferences. An appropriate hypothesis testing method must be used to ensure that the overall type I error rate of the trial is not inflated. The evidence from the two stages can be combined using sufficient statistics from the accumulated data after each of the two stages or using the p-values from the two stages. Several authors [1][2][3] have proposed ASDs for which hypothesis testing is based on the sufficient statistics for the selected treatment effect, where the selected treatment is that which is seen to be most effective in the trial. If the selected treatment is not the most effective but testing is performed as if the most effective treatment has been selected using the preceding methods, the test is conservative [4]. Hypothesis testing following an ASD can be carried out by combining the p-values from stages 1 and 2 [5,6]. This method of testing is very flexible with regard to the choice of the selection rule. The flexibility of this testing has been exploited to propose ASDs that use Bayesian techniques to make the selection but use frequentist methods for hypothesis testing [7][8][9]. A third technique of testing hypotheses after an adaptive trial is by using the conditional error principle [10] as in the adaptive Dunnett test [11].
The focus in this paper is estimation following an ASD. Estimation in this context is challenging because experimental treatments are retained in the trial precisely because they appear to be the most promising. Data suggesting that one treatment is superior may arise by chance even if the treatment is not truly superior to the other experimental treatments. Although the estimates may be biased, the bias can be quantified only if the rule for selecting the most promising treatment is specified in advance [12]. This is because bias is defined as an expectation and expectations are taken over all possible outcomes, requiring specification of the selection rule used. The most promising treatment may be chosen on the basis of effectiveness and other factors such as safety. In this paper, we will focus on selection where the most promising treatment is that which has the highest apparent effectiveness at the end of stage 1. For such a selection, the effectiveness of treatments chosen to remain in the clinical trial is likely to be overestimated.
Regulatory guidance [13,14] indicates that the bias of estimates obtained following an ASD should be considered. Cohen and Sackrowitz [15] and Shen [16] have proposed methods for estimating the mean of the selected treatment. The Cohen and Sackrowitz estimator is unbiased, whereas the Shen estimator reduces the bias relative to the naive estimator that ignores selecting the most effective treatment to continue to stage 2 based on the observed stage 1 data. Stallard and Todd [17] have proposed a method for estimating the mean of the selected treatment and also the means of the treatments that are dropped at stage 1. Cohen and Sackrowitz [15] and Shen [16] assume that the trial always continues to stage 2, whereas Stallard and Todd [17] assume that the trial may stop either for futility (when none of the experimental treatments are sufficiently effective on the basis of stage 1 data) or for efficacy. In this paper, we extend these methods to the setting where the trial can stop at stage 1 for futility, but not efficacy. This setting is common in practice. We will derive new estimators for the treatment difference for the selected treatment when estimation is unbiased conditional on continuing to stage 2. This differs from the Stallard and Todd estimator because the Stallard and Todd estimator is derived to be approximately unbiased conditional on the selected treatment whereas the estimators we will derive in this paper are obtained conditional on the selected treatment and the fact that the trial continues to stage 2. The two new conditional estimators (the word conditional is used to emphasize that estimation is unbiased conditional on continuing to stage 2) that we will derive extend the Cohen and Sackrowitz estimator and the Stallard and Todd estimator.
Like Koopmeiners et al. [18], we believe estimation unbiased conditional on continuing to stage 2 is of practical importance because when the trial cannot stop for efficacy at stage 1, it is reasonable to be interested in making a claim only when the trial continues to stage 2. Also, unconditionally unbiased estimators, that is estimators that do not condition on the stage at which the trial stops, may be conditionally biased [19]. Because of this, in [19], the authors proposed to obtain estimators unbiased conditional on the stage at which the trial stops.
We organized the remainder of the paper as follows. In Section 2.1, we describe the setting of interest while giving the notation. In Section 2.2, we derive an estimator that extends the Cohen and Sackrowitz estimator, and in Section 2.3, we derive expressions used to obtain an estimator that uses the Stallard and Todd principle. In Section 3, we present a worked example. We compare the various estimators using a simulation study in Section 4. The paper ends with a discussion in Section 5.
Setting and notation
As already mentioned, we will consider ASDs with two stages where stage 1 is used to select the most effective treatment and stage 2 is used for confirmatory analysis. Let k .>2/ denote the number of experimental treatments available at stage 1 for comparison with the control treatment, with the experimental treatment showing the highest effectiveness based on stage 1 data selected to continue to stage 2 together with the control. Let the number of subjects allocated to each treatment at stage 1 be denoted by n 1 .
We assume outcomes from treatment i (i D 0; 1; :::; k), with i D 0 corresponding to the control treatment, are normally distributed with mean i and a known common variance 2 , so that the stage 1 sample mean for treatment i follows a normal distribution N i ; 2 1 , where 2 1 D 2 =n 1 . We denote the stage 1 sample mean for treatment i by X i and the observed sample mean by x i . Let the selected treatment be denoted by S .S 2 f1; :::; kg/, noting that S is a random variable, and the sample mean from the stage 2 data for treatment i (i D 0; S ), with i D 0 corresponding to the control treatment, be denoted by Y i with observed sample mean denoted by y i so that Y i N i ; 2 2 , where 2 2 D 2 =n 2 , with n 2 the number of subjects allocated to each treatment at stage 2. We suppose that the trial continues to stage 2 if x S x 0 > b. We will refer to b as the futility boundary.
We define the selection time as the proportion n 1 =.n 1 C n 2 /. This is the proportion of stage 1 data for the control and the selected treatment. We denote the selection time by t so that the sample mean from the two stages for the control treatment is given by Z 0;MLE D tX 0 C .1 t/Y 0 and the sample mean for the selected treatment is given by Z S;MLE D tX S C .1 t/Y S . After completion of the trial, the objective is to estimate the treatment difference  S D S 0 . We can base the inference on the naive maximum likelihood estimator (MLE) for the difference between the selected and control treatments given by We will refer to this as the naive estimator. When there is no opportunity to stop at stage 1, the naive estimator is positively biased [12,20,21]. This is because the chosen experimental treatment is selected on the basis of having the maximum observed treatment difference compared with the control treatment.
In this paper, the objective is to seek estimators, which are unbiased conditional on continuing to stage 2, for the setting where a trial can stop at stage 1 for futility. For this setting, the naive MLE is also positively biased because of selecting the highest effective treatment and also requiring x S x 0 , the observed difference between the selected and control treatments at stage 1, to exceed the critical value b. If estimation is conditional on continuing to stage 2, Y S and Y 0 are respectively unbiased estimators for S and 0 so that the stage 2 sample difference is an unbiased estimator for  S . However, this estimator, which we will henceforth refer to as the stage 2 estimator, is likely to be inefficient compared with estimators that use both stage 1 and 2 data. In Sections 2.2 and 2.3, we will derive two new estimators that use both stages 1 and 2 data .
A new unbiased estimator for the treatment difference
Cohen and Sackrowitz [15], although not considering the control treatment, derived a uniformly minimum variance unbiased estimator (UMVUE) for S when the trial always continues to stage 2. When the trial always continues to stage 2, the bias of the naive estimator of  S arises from using Z S;MLE as an estimator for S [12]. Thus, replacing Z S;MLE with the Cohen and Sackrowitz UMVUE for S in Equation (1) gives an unbiased estimator for  S in the case where the trial always continues to stage 2. In this paper, we are interested in a setting where a trial can stop for futility and estimation is conditional on continuing to stage 2. For this setting, Z 0;MLE is biased for 0 and also the estimator for S derived by Cohen and Sackrowitz is no longer unbiased because it does not condition on continuing to stage 2. In the rest of this section, we will derive the UMVUE for S and the UMVUE for 0 when estimation is conditional on continuing to stage 2, and hence an unbiased estimator for  S .
The UMVUEs are based on the Rao-Blackwell theorem (for example [22]). If estimation is conditional on continuing to stage 2, Y S is an unbiased estimator of S . In the Rao-Blackwell theorem, a new estimator defined as the expected value of Y S given a sufficient statistic for S is the UMVUE for S . Similarly, Y 0 is an unbiased estimator of 0 so that the expected value of Y 0 given a sufficient statistic for 0 is the UMVUE for 0 . Let X .1/ > X .2/ > ::: > X .k/ be the order statistics of stage 1 sample means so that X S D X .1/ . For the selected treatment S , we show in Appendix A that the UMVUE for S , which we denote by Z S;CHN with the notation chosen such that it reflects the fact that the estimator Copyright where . / andˆ. / respectively denote the density and distribution functions of a standard normal, and B D X 0 C b. For the control treatment, we show in Appendix B that the UMVUE for 0 , which we denote by Z 0;CHN , is given by Because Z S;CHN and Z 0;CHN are unbiased estimators for S and 0 , then is an unbiased estimator for  S . We will refer to this estimator as the (new) unbiased estimator. If we set the futility boundary b D 1 so that B D 1, B 1 D 1 and maxfB; X .2/ g D X .2/ , then Equation (5) simplifies to Z 0;CHN D Z 0;MLE and Equation (4) simplifies to The simplification of Z 0;CHN to Z 0;MLE supports the finding in [12] that when there is no opportunity to stop at stage 1, the bias when the naive estimator is used as an estimator of the treatment difference is only contributed to by using Z S;MLE as an estimator for the selected treatment in Equation (1). If further we have 2 D 1 and n 1 D n 2 D 1 so that 2 1 D 2 2 D 1, formulae (3) and (7) reduce to the formulae given by Cohen and Sackrowitz.
Koopmeiners et al. [18] considered the setting with k D 1. Note that for a trial with no control arm, the UMVUE for this setting is given by Equation (3), but with W B expressed as where B is the futility boundary and Z S;MLE is the sample mean for the experimental treatment. The same formula applies for the case with a control, replacing Z S;MLE with D S;MLE , the sample difference between the experimental treatment and the control, and appropriately defining 2 1 and 2 2 as the variances of stages 1 and 2 sample differences. Koopmeiners et al. also derived the UMVUE for the setting with k D 1. The formula given by Koopmeiners et al. has a typological error. Defining 2 all D 2 =n, where n D n 1 C n 1 , using our notation, they give the formula for UMVUE as where B is the futility boundary, instead of This formula can be shown to be equivalent to our formula. Thus, the Koopmeiners et al. estimator is a special case of our estimator.
A new bias-adjusted estimator for the treatment difference
Stallard and Todd [17] proposed a bias-adjusted estimator that involves estimating the bias of the naive estimator. A bias-adjusted estimate is then obtained by subtracting the estimate of the bias from the naive estimate. The bias-adjusted estimator is obtained as follows. Let O  i (i D 1; :::; k) denote the naive maximum likelihood estimate for the treatment difference The initial value of Q  in the iteration procedure could be set to be O Â. If the solution is achieved at iteration r, then the bias-adjusted estimator for  S is given by where the notation is chosen to reflect the fact that this estimator is obtained using the principles described by Stallard and Todd. Stallard and Todd derived the bias vector conditional on the selected treatment. Let stage 1 treatment differences X i X 0 (i D 1; :::; k) be denoted by D i and the observed differences x i x 0 by d i . One of the densities Stallard and Todd need while deriving the bias vector, which we also need in this paper, is the joint density of S D i and d i given by In the rest of this section, we will derive the bias vector when estimation is conditional on continuing to stage 2, where the trial continues to stage 2 if d S > b. If the trial continues to stage 2, the expected value of the treatment difference for the selected treatment i is given by where f .d i ; S D i/ is the density given by Equation (10). The numerator and the denominator in Equation (11) are simplified to expressions with single integrals in Appendix C.1. The expected value of the naive estimator given by Equation (1) can be expressed as t.EOEX S X 0  S / C  S so that the bias of the treatment difference for the selected treatment i, given that the trial continues to stage 2, may be written as Conditional on the trial continuing to stage 2, the expected value of the treatment difference between a dropped treatment i 0 and the control treatment is expressed by The expression for pr .S D i; D i > b/ is given earlier. The expression for EOED i 0 ; S D i; D i > b while using D i 0 directly involves multidimensional integrals that cannot be simplified to fewer integrals. To overcome this, we define a new variable W 0 that has a normal distribution N 0; 2 1 and its covariance with D i .i D 1; :::; k/, Cov.
and Cov.W i ; W j / D 0 for i ¤ j D 0; 1; :::; k. Note that The expressions for EOEW i 0 ; S D i; D i > b and EOEW 0 ; S D i; D i > b are simplified to single integrals in Appendix C.2. The bias of the treatment difference for a dropped treatment i 0 given that the trial continues to stage 2 may be written as To obtain b S . Q  r / to substitute in Equation (9) and obtain a bias-adjusted estimate when estimation is conditional on continuing to stage 2, expressions (12) and (13) are used in the iteration procedure but with  i and  i 0 replaced by Q  i and Q  i 0 , respectively. We will refer to this estimator for  S as the (new) bias-adjusted estimator. Koopmeiners et al. [18] derived a similar bias-adjusted estimator for the setting with k D 1 so that their bias-adjusted estimator is a special case of our bias-adjusted estimator.
Example
In this section, using the two new estimators described in Sections 2.2 and 2.3, we compute estimates for an example constructed from the case study described in [21]. The case study is based on a comparison of three doses of an experimental drug for generalized anxiety disorder with a placebo. The primary endpoint is the change from baseline at 8 weeks of treatment in the total score on the Hamilton Rating Scale for Anxiety. The primary endpoint is taken to be normally distributed with a common standard deviation across the four treatment arms assumed to be 6 points. As in [21], we consider a two-stage ASD for the case study with n 1 D n 2 D 71 so that t D 0:5.
Suppose that the true treatment means are the stage 1 estimates from [21], which we give in Table I (column 1), and that the observed stage 1 means from an adaptive trial are as given in column 2. We suppose the trial continues to stage 2 if the highest effective experimental dose is at least as effective as the placebo, that is, the observed difference between the highest effective experimental dose and the placebo is at least 0. On the basis of the observed stage 1 data, dose 2 and placebo would be tested further in stage 2. We suppose the results from stage 2 are as given in column 3.
With the results in Table I (3) and (4),´2 ;CHN D 1:261. For´0 ;CHN , the only component we have not calculated is B 1 , which is given by 1:766 0. By substituting the appropriate values in Equation (5),´0 ;CHN D 0:017. Therefore, the unbiased estimate for the difference between dose 2 and placebo is 1.278.
For the new bias-adjusted estimator, we note that the naive maximum likelihood estimate for (Â 1 ; Â 2 ; Â 3 ) is (0:495; 1:626; 1:649). The bias function for doses 1 and 3 is given by expression (13), and the bias function for dose 2 is given by expression (12). Using a program written in the R statistical package, we obtain the value of b. Q Â/ and hence Q Â D O Â b. Q Â/ at each iteration. The iteration procedure stops at iteration r if the Euclidean distance between Q Â r 1 and Q Â r is less than or equal to 0.0005. The program is available at https://files.warwick.ac.uk/nstallard/browse/adaptive. We set Q Â 1 D .0:495; 1:626; 1:649/. The iteration procedure stopped at iteration 15, and the bias-adjusted estimate for the difference between dose 2 and placebo is 1.135.
Thus, the naive, stage 2, unbiased, and bias-adjusted estimates for the difference between dose 2 and the placebo are 1.626, 1.402, 1.278, and 1.135, respectively. The estimates are different with the naive estimate, as expected, having the highest value. The unbiased and bias-adjusted estimates correct for the bias, and their values are below both the stage 1 and 2 differences. The unbiased and bias-adjusted estimates are closer to the stage 2 difference, which is an unbiased estimate of the treatment difference. We explore the properties of the four estimators in the next section.
Simulation study settings
In this section, we describe a simulation study that was used to assess the bias and the mean squared error (MSE) of the estimators described in Section 2. Following expressions (10), (11), and (12), the bias of the naive estimator depends on the number of experimental treatments k, the selection time t , the value of the futility boundary, and the true parameter values. Therefore, we will consider several scenarios in the simulation study. We will consider scenarios where k is between 2 and 5. We believe this encompasses the majority of practical scenarios with k > 1. We will also consider different true parameter values for the means. In all simulations, we will take the variance of the outcomes 2 to be 1. Hence, we will only consider small differences in true treatment means corresponding to the small standardized effect sizes that we might anticipate in clinical trials.
We will assess three different values for the futility boundary. In most simulations that we will describe, we will take the treatment difference between the most effective treatment(s) and the control treatment to be 0.05. The first futility boundary value is 0, so because it is below the highest treatment difference(s), this boundary will be used to assess the bias when some of the experimental treatments are more effective than minimally required. The second futility boundary value is 0.05, so it will be used to assess the bias when the highest treatment difference is on the futility boundary. The third futility boundary value is 0.10, so it will be used to assess the bias when none of the experimental treatments are as effective as is minimally desired. We will also describe simulation results for some scenarios where the treatment difference between the most effective treatments(s) is 0.1, 0.2, and 0.5 while using the same futility boundary values (0, 0.05, and 0.1). These simulations will be used to assess the bias and MSE when most bias is contributed by the selection of the most effective treatment and not because of the futility boundary.
We perform simulations for 14 values of t , the selection time point, in the interval (0, 1). Because of the computations required, at each time point, we run 10,000 simulations that would continue to stage 2, that is, 10,000 simulations for which the simulated stage 1 treatment difference of the selected treatment is equal to or greater than the futility boundary value. For the treatment difference of the selected treatment S , in each simulation, we obtain the naive MLE d S;MLE using Equation (1), the stage 2 estimate d S;2 using Equation (2), the unbiased estimate d S;CHN using Equation (6), and the bias-adjusted estimate d S;STL using Equation (9). We then calculate the differences .d S;MLE Â S /, .d S;2 Â S /, .d S;CHN Â S /, and .d S;STL Â S / and the respective squares .d S;MLE Â S / 2 , .d S;2 Â S / 2 , .d S;CHN Â S / 2 , and .d S;STL Â S / 2 . Then at each selection time point, for each estimator, the mean bias is obtained by taking the average of its corresponding 10,000 differences and the MSE by taking the average of its corresponding 10,000 square differences. Copyright We will present the bias and the p MSE of the various estimators in units of the standard error (SE), the standard deviation for the estimator of the difference of a single experimental treatment-control comparison given by p 2=.n 1 C n 2 /. This makes the results invariant to changes in the sample sizes.
4.2.
Simulation results for k D 2 with  1 D  2 Figure 1 shows the bias and p MSE when two experimental treatments and a control are included in stage 1 with  1 D  2 D 0:05. Columns 1, 2, and 3 correspond to futility boundary values 0, 0.05, and 0.1, respectively. The dashed, dotted, continuous, and dash-dotted lines correspond to the naive, stage 2, unbiased, and bias-adjusted estimators, respectively. The naive estimator is biased, and the bias increases with selection time but not linearly and also as the futility boundary value increases. The stage 2 estimator, as expected, is mean unbiased for all selection time points and all futility boundary values. Because of the theoretical derivation of the unbiased estimator, this is also mean unbiased for all scenarios. The bias-adjusted estimator overcorrects for bias, and the overcorrection increases with selection time but decreases as the value of the futility boundary increases. The naive estimator has the lowest MSE at all selection times for all scenarios. The stage 2 estimator has the highest MSE. In all scenarios, up to selection time 0.7, the unbiased estimator and the bias-adjusted estimator have approximately equal MSE. Tables giving more details of the results in Figure 1 and of additional simulations mentioned in the following are available from the authors.
We also assessed the characteristics of the four estimators when  1 D  2 D 0:1,  1 D  2 D 0:2, and and the stage 1 sample difference for experimental treatment 2 beat the futility boundary values in most simulation runs so that the bias arises mostly because of the treatment selection and hence the similar bias of the naive estimator. As expected, the unbiased estimator and the stage 2 estimator are mean unbiased for all values of t , b,  1 , and  2 , whereas the bias-adjusted estimator is negatively biased, but the bias decreases as the values of  1 and  2 increase. For the three futility boundary values, as in the case where  1 D  2 D 0:05, for the cases where  1 D  2 D 0:1,  1 D  2 D 0:2, and  1 D  2 D 0:5, the naive estimator has the lowest MSE at all selection times, the stage 2 estimator has the highest MSE at all selection times, and up to selection time 0.7, the unbiased estimator and the bias-adjusted estimator have approximately equal MSE.
Simulation results for
To assess the bias and the MSE when Table II. This is unlike the setting in which the trial always continue to stage 2, where bias decreases as one of the experimental treatments becomes distinctly superior to the competing treatment [12]. To assess what may be causing this difference, in Table II, we observe that as the selection is made later in the trial, it is more likely that a right decision of whether to continue to stage 2 or not will be made. However, we note that for the case where  1 D 0:025 and  2 D 0:005, treatment 1 is still selected with relatively high probability (the minimum probability is 0.34 Table II when t D 0:8 and boundary value b D 0:1). Also, when treatment 1 is selected to continue to stage 2, the treatment difference for treatment 2 is usually below the boundary (Pr.S D 1; d 2 > b/ is small). We use Figure 2 to assess whether the instances where treatment 1 is selected are the ones that make bias higher when  1 D 0:025 and  2 D 0:05 than when  1 D  2 D 0:05. Figure 2(a) shows the bias of the naive estimator when  1 D  2 D 0:05 (dashed line) and when  1 D 0:025 and  2 D 0:05 (continuous line) in the case where the trial always continues to stage 2, and as expected, following [12], bias is higher when  1 D  2 D 0:05. The proof that for the case where the trial always continues to stage 2, the naive estimator is maximally biased when all experimental treatments are equally effective is given in [24]. Figure 2(b) shows the bias of the naive estimator when k D 1 and the futility boundary value b D 0:05. The continuous and dashed lines correspond to  1 D 0:025 and  1 D 0:05, respectively. The bias is higher when  1 D 0:025 than when  1 D 0:05. Comparing Figure 2(a and b), we see that the futility boundary seems to contribute more to the bias. This may explain why in the case where there is a futility boundary and k D 2, the bias of the naive estimator is higher when  1 D 0:025 and  2 D 0:05 than when  1 D  2 D 0:05. Although the selected treatment may be the most promising because the treatment effects are distinct and hence reduce the selection bias, whenever the least effective treatment is selected, the bias is higher because we have a futility boundary. We also performed simulations when . 1 ;  2 / D .0:075; 0:1/, . 1 ;  2 / D .0:175; 0:2/, and . 1 ;  2 / D .0:475; 0:5/ using the futility boundary values b D 0, b D 0:05, and b D 0:1. Note that for these parameter vectors, as for the case considered earlier where  1 D 0:025 and  2 D 0:05,  2  1 D 0:025. We describe the findings from these scenarios without giving the figures. For the three futility boundary values, the bias of the naive estimator decreases as the values of  1 and  2 increase. For . 1 ;  2 / D .0:475; 0:5/, the biases of the naive estimator for b D 0, b D 0:05, and b D 0:1 are identical. This is because for this case, for the three futility boundary values, the stage 1 sample difference for experimental treatment 1 and the stage 1 sample difference for experimental treatment 2 beat the futility boundary values in most simulation runs so that the bias arises mostly because of the treatment selection and hence the similar bias of the naive estimator. For the three futility boundary values, compared with the case where . 1 ;  2 / D .0:5; 0:5/, the biases of the naive estimator when . 1 ;  2 / D .0:475; 0:5/ are lower. This is because, for futility boundary values 0, 0.05 and 0.1, stage 1 sample differences for treatments 1 and 2 beat the futility boundary values in most simulations for the cases where . 1 ;  2 / D .0:5; 0:5/ and . 1 ;  2 / D .0:475; 0:5/ so that most bias arises from treatment selection and selection bias is maximal when experimental treatments are equally effective [12,24].
Simulation results for k > 3
When three or more experimental treatments are tested in stage 1, there are several possible configurations of the treatment differences, and this leads to several scenarios. Therefore, we will first describe general findings for such scenarios without presenting figures and then describe results of a few specific scenarios using a figure. On the basis of results that are not presented here, as in the case when two treatments are tested in stage 1, when three or more experimental treatments are tested in stage 1, the bias of the naive estimator increases with the futility boundary value, and estimation using the bias-adjusted estimator improves with higher futility boundary value whereas the stage 2 and unbiased estimators, as expected, provide unbiased estimators for all futility boundary values. Figure 3 shows results when treatment differences are all equal to 0.05 and the futility boundary is 0.05. Columns 1 to 3 give results when three, four, and five experimental treatments, respectively, are tested in stage 1. For the naive estimator, we observe that the bias increases slightly as the number of treatments increases. The stage 2 and unbiased estimators, as expected, are mean unbiased at all selection times and when three, four, or five experimental treatments are tested in stage 1. The bias-adjusted estimator again overcorrects for bias, and the overcorrection increases with selection time. Also, as the number of treatments increases, the overcorrection of the bias-adjusted estimator increases slightly. The naive estimator has the least MSE whereas the unbiased and bias-adjusted estimators have similar MSE for selection times up to 0.6. The stage 2 estimator has the highest MSE, and the difference between the MSE for the stage 2 estimator and the other estimators increases with selection time.
Summary of findings from the simulation study
From the simulation study, we have observed that the bias of the naive estimator increases with the selection time, the number of experimental treatments, and the futility boundary value. The treatment differences affect the bias of the naive estimator, but this also depends on the futility boundary value so that it is not possible to generalize the bias on the basis of treatment differences only. The stage 2 and unbiased estimators, as expected, provide mean unbiased estimates. The bias-adjusted estimator overcorrects for bias, but under some configurations of treatment differences, if selection is carried out up to selection time 0.4, it performs fairly well. For MSE, the unbiased and bias-adjusted estimators perform similarly up to time 0.6, whereas unsurprisingly, the stage 2 estimator performs worst. Regulation guidelines [13] suggest that methods for estimating treatment effect and confidence intervals with appropriate coverage should be provided as well as for controlling the prespecified type I error, whereas in [14], the importance of controlling the bias of the point estimate is emphasized. Hence, from the simulation findings and the importance of not overestimating treatment effect as described in [13,14], we recommend the unbiased estimator. Copyright
Discussion
In drug development, the need to reduce the cost and time taken to test new treatments has led to the use of ASDs. ASDs combine several phases of a clinical development program into a single trial. However, compared with traditional testing strategies, ASDs pose additional challenges in statistical analysis. In this paper, we have considered point estimation following an ASD where, on the basis of observed data at stage 1, the experimental treatment that is superior to the competing experimental treatments at stage 1 continues to stage 2 together with the control. Cohen and Sackrowitz [15] and Shen [16] have considered this setting in the case where the trial always continues to stage 2 and proposed estimators for the treatment difference. Stallard and Todd [17] have also proposed an estimator that can be applied in this setting.
In this paper, we have considered the setting where the trial can stop for futility and estimation is unbiased conditional on continuing to stage 2. We have extended the Cohen and Sackrowitz method to construct an unbiased estimator for this setting. We have referred to this estimator as the (new) unbiased estimator. Carreras and Brannath [24] compared the Cohen and Sackrowitz estimator and the Stallard and Todd estimator when the trial always continues to stage 2. Their findings show that although the Cohen and Sackrowitz estimator is unbiased, it is not the best in terms of MSE. Thus, although the estimator we derive by extending the Cohen and Sackrowitz estimator to the setting where the trial can stop at stage 1 for futility is unbiased by construction, it is of interest to compare it with other estimators in terms of bias and MSE. Therefore, we have also developed a new bias-adjusted estimator that extends the Stallard and Todd estimator to our setting.
We also considered extending the Shen [16] estimator. The Shen estimator was proposed when the trial always continues to stage 2 and adjusts for bias by proposing a step function. When the trial always continues to stage 2, the step function depends on the absolute differences between the experimental treatment means and a tuning parameter. The best value for the tuning parameter depends on the unknown true values of the treatment means. With the possibility of early stopping, the bias depends not only on the absolute differences between the means of the experimental treatments but also on the values of observed differences between these and the mean of the control because of the futility boundary. This makes it challenging to propose a step function, and because we also know it will depend on a tuning parameter whose best value depends on the unknown true treatment means, we did not pursue this estimator further.
In terms of MSE, if treatment selection and the decision whether to continue to stage 2 are made at a selection time t < 0:6, the unbiased and bias-adjusted estimators perform similarly. The stage 2 estimator performs worst in terms of MSE, and the naive estimator (unadjusted for the possibility of stopping and for selection) performs the best. In terms of bias, the unbiased and stage 2 estimators are unbiased, and the naive estimator is positively biased whereas the bias-adjusted estimator is negatively biased. From this finding, we propose using the new unbiased estimator we have derived in this paper by extending the Cohen and Sackrowitz estimator [15] when a trial can stop for futility and estimation is performed conditional on continuing to stage 2. We emphasize that, although in the simulation study, we averaged over all simulations and the selected treatments, by derivation, the new unbiased estimator fulfills a stronger condition of unbiasedness in that it is unbiased with respect to each treatment whenever it is selected.
In this paper, we have considered point estimation following a two-stage adaptive seamless trial in which at stage 1, there is treatment selection and the possibility of early stopping for futility and estimation is conditional on the trial continuing to stage 2. As mentioned in Section 4.5, methods for interval estimation (confidence intervals) that adjust for the adaptation so that the right coverage is achieved are also important. There exist methods for constructing confidence intervals that can be used for the setting considered in this paper [17,20,25]. However, the confidence intervals following these methods are not based on the principle used to develop the estimators in this paper. For further research, we are considering confidence intervals based on the principle used to derive the unbiased estimator.
Appendix A. Deriving the uniformly minimum variance unbiased estimator for S For ease of notation in the derivation of UMVUEs for S and 0 , without loss of generality, we let X 1 > ::: > X k so that X .i / D X i (i D 1; :::; k) and X S D X 1 and Y S D Y 1 . For the mean of the selected treatment, we are seeking the UMVUE for 1 . We will skip details of steps that are similar to steps given in [15] and [26]. Denote by Q B .X / the event fX 0 ; X 1 > ::: > X k ; X 1 > Bg, where B D X 0 C b. We will first write and re-express the density of .Y 0 ; Y 1 ; X/, where X D .X 0 ; X 1 ; :::; X k / given Q B in order to deduce the sufficient statistics for estimating 1 that combine stage 1 and 2 means for treatment 1 into a single quantity. The density of .Y 0 ; Y 1 ; X/ given Q B , denoted by f .y 1 ; y 0 ; xjQ B /, is given by where 1 and 2 are as defined in Section 2.1, 1 OEQ B is the indicator for Q B .x/, K. / D Prob 1 OEQ B .x/ D 1 , and The preceding density can be re-expressed as Let´1 D . 2 = 1 /x 1 C . 1 = 2 /y 1 ; then from the preceding density, .X 0 ; X 2 ; :::; X k ; Y 0 ; Z 1 / is sufficient and complete for the problem of seeking an estimate for 1 given Q B . Therefore, conditional on Q B , the UMVUE for 1 is given by EOEY 1 jX 0 ; X 2 ; :::; X k ; Y 0 ; Z 1 ; Q B . We obtain the expression for this by deriving the density f .y 1 jx 0 ; x 2 ; :::; x k ; y 0 ;´1; Q B / and using it to get the expected value we are seeking. | 10,108 | 2013-02-15T00:00:00.000 | [
"Mathematics",
"Medicine"
] |
Classifying Code Comments in Java Open-Source Software Systems
Code comments are a key software component containing information about the underlying implementation. Several studies have shown that code comments enhance the readability of the code. Nevertheless, not all the comments have the same goal and target audience. In this paper, we investigate how six diverse Java OSS projects use code comments, with the aim of understanding their purpose. Through our analysis, we produce a taxonomy of source code comments, subsequently, we investigate how often each category occur by manually classifying more than 2,000 code comments from the aforementioned projects. In addition, we conduct an initial evaluation on how to automatically classify code comments at line level into our taxonomy using machine learning, initial results are promising and suggest that an accurate classification is within reach.
I. INTRODUCTION
While writing and reading source code, software engineers routinely introduce code comments [6]. Several researchers investigated the usefulness of these comments, showing that thoroughly commented code is more readable and maintainable. For example, Woodfield et al. conducted one of the first experiments demonstrating that code comments improve program readability [35]; Tenny et al. confirmed these results with more experiments [31], [32]. Hartzman et al. investigated the economical maintenance of large software products showing that comments are crucial for maintenance [12]. Jiang et al. found that comments that are misaligned to the annotated functions confuse authors of future code changes [13]. Overall, given these results, having abundant comments in the source code is a recognized good practice [4]. Accordingly, researchers proposed to evaluate code quality with a new metric based on code/comment ratio [21], [9].
Nevertheless, not all the comments are the same. This is evident, for example, by glancing through the comments in a source code file 1 from the Java Apache Hadoop Framework [1]. In fact, we see that some comments target enduser programmers (e.g., Javadoc), while others target internal developers (e.g., inline comments); moreover, each comment is used for a different purpose, such as providing the implementation rationale, separating logical blocks, and adding reminders; finally, the interpretation of a comment also depends on its position with respect to the source code.
Defining a taxonomy of the source code comments that developers produce is an open research problem. 1 https://tinyurl.com/zqeqgpq Haouari et al. [11] and Steidl et al. [28] presented the earliest and most significant results in comments' classification. Haouari et al. investigated developers' commenting habits, focusing on the position of comments with respect to source code and proposing an initial taxonomy that includes four highlevel categories [11]; Steidl et al. proposed a semi-automated approach for the quantitative and qualitative evaluation of comment quality, based on classifying comments in seven high-level categories [28]. In spite of the innovative techniques they proposed to both understanding developers' commenting habits and assessing comments' quality, the classification of comments was not in their primary focus.
In this paper, we focus on increasing our empirical understanding of the types of comments that developers write in source code files. This is a key step to guide future research on the topic. Moreover, this increased understanding has the potential to (1) improve current quality analysis approaches that are restricted to the comment ratio metric only [21], [9] and to (2) strengthen the reliability of other mining approaches that use source code comments as input (e.g., [30], [23]).
To this aim, we conducted an in-depth analysis of the comments in the source code files of six major OSS systems in Java. We set up our study as an exploratory investigation. We started without hypotheses regarding the content of source code comments, with the aim of discovering their purposes and roles, their format, and their frequency. To this end, we (1) conducted three iterative content analysis sessions (involving four researchers) over 50 source files including about 250 comment blocks to define an initial taxonomy of code comments, (2) validated the taxonomy externally with 3 developers, (3) inspected 2, 000 source code files and manually classified (using a new application we devised for this purpose) over 15, 000 comment blocks comprising more than 28, 000 lines, and (4) used the resulting dataset to evaluate how effectively comments can be automatically classified.
Our results show that developers write comments with a large variety of different meanings and that this should be taken into account by analyses and techniques that rely on code comments. The most prominent category of comments summarizes the purpose of the code, confirming the importance of research related to automatically creating this type of comments. Finally, our automated classification approach reaches promising initial results. Listing 1 shows a Java source file example containing both code and comments. In a well-documented file, comments help the reader with a number of tasks, such as understanding the code, knowing the choices and rationale of authors, and finding additional references. When developers perform software maintenance, the aforementioned tasks become mandatory steps that practitioners should attend. The fluency in performing maintenance tasks depends on the quality of both code and comments. When comments are omitted, much depends on developers' ability and code complexity; when well-written comments are present, the maintenance could be simplified.
A. Code/comment ratio to measure software maintainability When developers want to estimate the maintainability of code, one of the easiest solutions consists in the evaluation of the code/comment ratio proposed by Garcia et al. [9]. By evaluating the aforementioned metric in the snippet in Listing 1, we find an overall indicator of good quality. However, the evaluated measure is inaccurate. The limitation arises from the fact that this metric considers only one kind of comment. More precisely, Garcia et al. focus only on the presence or absence of comments, omitting the possibility of use comments with different benefits for different end-users. Unfortunately, the previous sample of code represents a case where the author used comments for different purposes. The comment on line 31 represents a note that developers use to remember an activity, an improvement, or a fix. On line 20 the author marks his contribution on the file. Both these two comments represent real cases where the presence of comments increases the code/comment ratio without any real effect on code readability. This situation hinders the validity of this kind of metric and indicates the need for a more accurate approach to tackle the problem.
B. An existing taxonomy of source code comments A great source of inspiration for our work comes from Steidl et al. who presented a first detailed approach for evaluating comment quality [28]. One of the key steps of their approach is to first automatically categorize the comments to differentiate between different comment types. They define a preliminary taxonomy of comments that comprises 7 high-level categories: COPYRIGHT, HEADER, MEMBER, INLINE, SECTION, CODE, and TASK. They provide evidence that their quality model, based on this taxonomy, provides important insights on documentation quality and can reveal quality defects in practice.
The study of Steidl et al. demonstrates the importance of treating comments in a way that suits their different categories. However, the creation of the taxonomy was not the focus of their work, as also witnessed by the few details given about the process that led to its creation. In fact, we found a number of cases in which the categories did not provide adequate information or did not differentiate the type of comments enough to obtain a clear understanding. Noise. Line 36 represents a case of a comment that should be disregarded from any further analysis. Since it does not separate parts, the SECTION would not apply and an automated classification approach would try to wrongly assign it to one of the other categories. No sort of noise category is considered.
With our work, we specifically focus on devising an empirically grounded, fine-grained classification of comments that expands on previous initial efforts. Our aim is to get a comprehensive view of the comments, by focusing on the purpose of the comments written by developers. Besides improving our scientific understanding of this type of artifacts, we expect this work to be also beneficial, for example, to the effectiveness of the quality model proposed by Steidl et al. and other approaches relying on mining and analyzing code comments (e.g., [21], [30], [23]). Pascarella, and Bacchelli -Classifying Code Comments in Java Open-Source Software Systems
III. METHODOLOGY
This section defines the overall goal of our study, motivates our research questions, and outlines our research method.
A. Research Questions
The ultimate goal of this study is to understand and classify the primary purpose of code comments written by software developers. In fact, past research showed evidence that comments provide practitioners with a great assistance during maintenance and future development, but not all the comments are the same or bring the same value.
We started analyzing past literature looking for similar efforts on analysis of code comments. We observed that only a few studies define a rudimentary taxonomy of comments and none of them provides an exhaustive categorization of all kinds of comments. Most of past work focuses on the impact of comments on software development processes such as code understanding, maintenance, or code review and the classification of comments is only treated as a side outcome (e.g., [31], [32]). Therefore, we set our first research question:
RQ1. How can code comments be categorized?
Given the importance of comments in software development, the natural next step is to apply the resulting taxonomy and investigate on the primary use of comments. Therefore, we investigate whether some classes of comments are predominant and whether there is a pattern across different projects. This investigation is reflected in our second research question: RQ2. How often does each category occur?
Finally, we investigate to what extent an automated approach can classify unseen code comments according to the taxonomy defined in RQ1. An accurate automated classification mechanism is the first essential step in using the taxonomy to mine information from large-scale projects and to improve existing approaches that rely on code comments. This leads to our last research question: RQ3. How effective is an automated approach, based on machine learning, in classifying code comments?
B. Selection of subject systems
To conduct our analysis, we focused on a single programming language (i.e., Java, one of the most popular programming languages [5]) and on projects whose source code is publicly available, i.e., open-source software (OSS) projects. Particularly, we selected six heterogeneous software systems: Apache Spark [2], Eclipse CDT, Google Guava, Apache Hadoop, Google Guice, and Vaadin. They are all open source projects and the history of the changes are controlled with GIT version control system. Table I details the selected systems. We select unrelated projects emerging from the context of different four software ecosystems (i.e., Apache, Google, Eclipse, and Vaadin); the development environment, the number of contributors, and the project size are different, thus mitigating some threats to the external validity.
C. Categorization of code comments
To answer our first research question about categorizing code comments, we conducted three iterative content analysis sessions [15] involving 4 software engineering researchers (3 Ph.D. candidates and 1 faculty member) with at least 3 years of programming experience. Two of these researchers are authors of this paper. In the first iteration, we started choosing 6 appropriate projects (reported in Table I) and sampling 35 files with a large variety of code comments. Subsequently, together we analyzed all source code and comments. During this analysis we could define some obvious categories and left undecided some comments; this resulted in the first draft taxonomy defining temporary category names. In the course of the second phase, we first conducted an individual work analyzing 10 new files, in order to check or suggest improvements to the previous taxonomy, then we gathered together to discuss the findings. The second phase resulted in a validation of some clusters in our draft and the redefinitions of others. The third phase was conducted in team and we analyzed 5 files that were previously unseen. During this session we completed the final draft of our taxonomy verifying that each kind of comments we encountered was covered by our definitions and those overlapping categories were absent.
Through this iterative process, we defined a taxonomy having a hierarchy with two layers. The top layer consists of 6 categories and the inner layer consists of 16 subcategories.
Validation. We externally validated the resulting taxonomy with 3 professional developers having 3 to 5 years of Java programming experience. We conducted one session with each developer. At the beginning of the session, the developer received a printed copy of the description of the comment categories in our taxonomy (similar to the explanation we provide in Section IV-A) and was allowed to read through it and ask questions to the researcher guiding the session. Afterwards, the developer was required to login into COM-MEAN (a web application, described in Section III-D, that we devised for this task and to facilitate the large-scale manual classification necessary to answer RQ2 and RQ3) and classify each comment in 3 Java source code files (the same files have been used for all the developers), according to the provided taxonomy. During the classification, the researcher was not in the experiment room, but the printed taxonomy could be consulted. At the end of the session, the guiding researcher came back to the experiment room and asked the participant to comment on the taxonomy and the classification task. At the end of all three sessions, we compared the differences (if any) among the classifications that the developers produced.
All the participants found the categories clear and the task feasible; however, they also reported the need for consulting the printed taxonomy several times during the session to make sure that their choice was in line with the description of the category. The analysis of the three sets of answers showed a few minor differences with an agreement above 92%. The differences were all within the same top category and mostly regarding where the developers split certain code blocks into two sub-categories.
D. A dataset of categorized code comments, publicly available
To answer the second research question about the frequencies of each category, we needed a statistically significant set of code comments classified accordingly to the taxonomy produced as an answer to RQ1.
Sampling approach. Since the classification had to be done manually, we relied on random sampling to produce a statistically significant set of code comments from each one of the six OSS projects we considered in our study. To establish the size of such sample sets, we used as a unit the number of files, rather than number of comments: This results in sample sets that give a more realistic overview of how comments are distributed in a system. In particular, we established the size (n) of such set with the following formula [33]: The size has been chosen to allow the simple random sampling without replacement. In the formulap is a value between 0 and 1 that represents the proportion of files containing a specific block of code comment, whileq is the proportion of files not containing such kind of comment. Since the apriori proportion ofp is not known, we consider the worst case scenario wherep ·q = 0.25. In addition, considering we are dealing with a small population (i.e., 557 Java files for Google Guice project) we use the finite population correction factor to take into account their size (N ). We sample to reach a confidence level of 95% and error (E) of 5% (i.e., if a specific comment entity is present in f % of the files in the sample set, we are 95% confident it will be in f % ± 5% files of our population). The suggested value for the sample set is 1, 925 files. In addition, since we split the sample sets in two parts with an overlapped chunk for validation, we finally sampled 2, 000 files. This value does not change significantly the error level that remains close to 5%. This choice only validates the quality of our dataset as a representation of the overall population: It is not related to the precision and recall values presented later, which are actual values based on manually analyzed elements.
Manual classification. Once the sample of files with comments was selected, each of them had to be manually classified according to our taxonomy. To facilitate this error-prone and time-consuming task, we build a web application, named COMMEAN. Figure 1 shows the main page of COMMEAN, which comprises the following components: • The Actions panel (1) handles the authentication of the users and several actions such as 'start', 'suspend', or 'send classification'. In addition, the panel keeps the user updated on the status of the classification showing the path of the resource loaded in the application and the progress with the following syntax: I-P /T . Where I represents the current index, P is the progress, and T is the total number of files to be processed.
• The Annotation panel (2) allows the user to append a pre-defined label to the selected text or define a new label. It enables the possibility to append a free text comment, create a link between comments and code, or categorize text composed of multiple parts. In addition, two keyboard shortcuts help the user to append the current label to selected text and create a link between source code and comments.
• The Source view panel (3) is the main view of the application. It contains the Java source file with highlighted syntax to help users during the classification and increase the quality of the analysis. In addition, the processed parts of the file are marked with different colors.
• The Status panel (4) shows the progress of the current file. A dynamic table is created when a new comment is added. A row of the table contains the initial position, the final position, the label used in the categorization, a summary of how many parts compose it, and a summary of linked code (if any). Clicking on rows, the correspondent text is highlighted and using the delete button the user is able to cancel a wrong classification.
• The Selection panel (5) shows details such as selected test, initial position, final position, and length of the text.
The two authors of this paper manually inspected the sample set composed of 2, 000 files. One author analyzed 100% of these files, while another analyzed a random, overlapping subset comprising 10% of the files. These overlapped files were used to verify their agreement, which, similarly to the external validation of the taxonomy with professional developers (Section III-C), highlighted only negligible differences. Moreover, this large-scale categorization also confirmed the exhaustiveness of the taxonomy created in RQ1: None of the annotators felt that comments, or parts of the comments, should have been classified by creating a new category.
Finally, the two researchers annotated, when present, any link between comments and the code they are referring to. This allows the use of our dataset for future approaches that attempt to recover the traceability links between code and comments. We make our dataset publicly available [24]. Pascarella, and Bacchelli -Classifying Code Comments in Java Open-Source Software Systems
E. Automated classification of source code comments
In the third research question we set to investigate to what extent and with which accuracy source code comments can be automatically categorized according to the taxonomy resulting from the answer to RQ1. Employing sophisticated classification techniques (e.g., based on deep learning approaches [10]) to accomplish this task goes beyond the scope of the current work. Our aim is to two-fold: (1) Verifying whether it is feasible to create an automatic classification approach that provides fair accuracy and (2) defining a reasonable baseline against which future methods that aim at a more accurate, project-specific classification can be tested.
Classification granularity. We set the automated classification to work at line level. In fact, from our manual classification, we found several blocks of comments that had to be split and classified into different categories (similarly to the block defined in the lines 5-8 in Listing 1) and in the vast majority of the cases (96%), the split was at line level. In only less than 4% of the cases, one line had to be classified into more than one category. In these cases, we replicated the line in our dataset for each of the assigned categories, to get a lower bound on the effectiveness in these cases.
Classification technique. Having created a reasonably large dataset to answer RQ2 (it comprises more than 15,000 comment blocks totaling over 30,000 lines), we employ supervised machine learning [8] to build the automated classification approach. This kind of machine learning uses a pre-classified set of samples to infer the classification function. In particular, we tested two different classes of supervised classifiers: (1) probabilistic classifiers, such as naive Bayes or naive Bayes Multinominal, and (2) decision tree algorithms, such as J48 and Random Forest. These classes make different assumptions on the underlying data, as well as have different advantages and drawbacks in terms of execution speed and overfitting.
Classification evaluation. To evaluate the effectiveness of our automated technique to classification code comments into our taxonomy, we measured two well known Information Retrieval (IR) metrics for the quality of results [18], named precision and recall: The union of T P and F N constitutes the set of correct classifications for a given category (or overall) present in the benchmark, while the union of T P and F P constitutes the set of comments as classified by the used approach. In other words, precision represents the fraction of the comments that are correctly classified into a given category, while recall represents the fraction of correct comments in that category.
F. Threats to validity
Sample validity. One potential criticism of a scientific study conducted on a small sample of projects is that it could deliver little knowledge. In addition, the study highlights the characteristics and distributions of 6 open source frameworks mainly focusing on developers practices rather than end-users patterns. Historical evidence shows otherwise: Flyvbjerg gave many examples of individual cases contributing to discoveries in physics, economics, and social science [7]. To answer to our research questions, we read and inspected more than 28, 000 lines of comments belonging to 2, 000 Java files (see Section III-D) written by more than 3, 000 contributors in 6 different projects (in accord to Table I). We also chose projects belonging to four open-source software ecosystems and with different development environments, number of contributors, and size of the project.
Taxonomy validity. To ensure that the comments categories emerged from our content analysis sessions were clear and accurate, and to evaluate whether our taxonomy provides an exhaustive and effective way to organize source code comments, we conduced a validation session that involved three experienced developers (see Section III-C) external to content analysis sessions. These software engineers conducted an individual session on 3 unrelated Java source files. They observed that categories were clear and the task feasible, and the analysis of the three sets of answers showed a few minor differences with an agreement above 92%. In addition, we reduce the impact of human errors during the creation of the dataset by developing COMMEAN, a web application to assist the annotation process.
External validity. Threats come with the generalization of our results. The proposed approach may show different result on different target systems. To reduce this limitation we selected 6 projects with unrelated characteristics and with different size in term of contributors and number of lines.
To judge the generalizability we tested our results simulating this circumstance using the project cross validation. Similarly, another threat concerning the generalizability is that our taxonomy refers only to a single object-oriented programming language i.e., Java. However, since many objectoriented languages descend to common ancestor languages, many functionalities across object-oriented programming are similar and it is reasonable to expect the same to happen for their corresponding comments. Further research can be designed to investigate whether our results hold in other programming paradigms.
IV. RESULTS AND ANALYSIS
In this section, we present and analyze the results of our research questions aimed at understanding what developers write in comments and with which frequency, as well as at evaluating the results of an automated classification approach.
A. RQ1. How can code comments be categorized?
Our manual analysis led to the creation of a taxonomy of comments having a hierarchy with two layers. The top level categories gather comments with similar overall purpose, the internal levels provide a fine-grained definition using explanatory names. The top level categories are composed of 6 distinct groups and the second level categories are composed of 16 definitions. We now describe each category with the corresponding subcategories.
A. PURPOSE
The PURPOSE category contains the code comments used to describe the functionality of linked source code either in a shorter way than the code itself or in a more exhaustive manner. Moreover, these comments are often written in a pure natural language and are used to describe the purpose or the behavior of the referenced source code. The keywords 'what', 'how' and 'why' describe the actions that take place in the source code in SUMMARY, EXPAND, and RATIONALE groups, respectively, which are the subcategories of PURPOSE: A.1 SUMMARY: This type of comments contains a brief description of the behavior of the source code referenced. To highlight this type of comments the question word 'what' is used. Intuitively, this category incorporates comments that represent a sharp description of what the code does. Often, this kind of comments is used by developers to provide a summary that helps understanding the behavior of the code without reading it. A.2 EXPAND: As with the previous category, the main purpose of reading this type of comment is to obtain a description of the associated code. In this case, the goal is to provide more details on the code itself. The question word 'how' can be used to easily recognize the comments belonging to this category. Usually, these comments explain in detail the purpose of short parts of the code, such as details about a variable declaration. A.3 RATIONALE: This type of comment is used to explain the rationale behind some choices, patterns, or options.
The comments that answer the question 'why' belong to that category (e.g., "Why does the code use that implementation?" or "Why did the developer use this specific option?").
B. NOTICE
The NOTICE category contains the comments related to the description of warning, alerts, messages, or in general, functionalities that should be used with care. It also includes the reasons and the explanation of some developers' choices. In addition, it covers the description of the adopted strategies to D.1 DIRECTIVE: This is an additional text used to communicate with the IDE. It is in form of comments to be easily skipped by the compiler and it contains text of limited meaning to human readers. These comments are often added automatically by the IDE or used by developers to change the default behavior of the IDE or compiler. D.2 FORMATTER: This type of comments represents a simple solution adopted by the developers to separate the source code in logical sections. The occurrence of patterns or the repetition of symbols is a good hint at the presence of a comment in the formatter category.
E. METADATA
The METADATA category aims to classify comments that define meta information about the code, such as authors, license, and external references. Usually, some specific tags (e.g., "@author") are used to mark the developer name and its ownership. The license section provides the legal information about the source code rights or the intellectual property.
E.1 LICENSE: Generally placed on top of the file, this types of comments describes the end-user license agreement, the terms of use, the possibility to study, share and modify the related resource. Commonly, it contains only a preliminary description and some external references to the complete policy agreement. E.2 OWNERSHIP: These comments describe the authors and the ownership with different granularity. They may address methods, classes or files. In addition, this type of comments includes credentials or external references about the developers. A special tag is often used e.g., "@author". E.3 POINTER: This types of comments contains references to linked resources. The common tags are: "@see", "@link" and "@url". Other times developers use custom references such as "FIX #2611" or "BUG #82100" that are examples of traditional external resources.
F. DISCARDED
This category groups the comments that do not fit into the categories previously defined; they have two flavors: F.1 AUTOMATICALLY GENERATED: This category defines auto-generated notes (e.g., "Auto-generated method stub"). In most case, the comment represents the skeleton with a comment's placeholder provided by the IDE and left untouched by the developers. F.2 NOISE: This category contains all remaining comments that are not covered by the previous categories. In addition, it contains the comments whose meaning is hard to understand due to their poor content (e.g., meaningless because out of context).
B. RQ2. How often does each category occur?
The second research question investigates the occurrence of each category of comments in the 2, 000 source files that we manually classified from our 6 OSS subject projects. Figure 2 shows the distribution of the comments across the categories; it reports the cumulative value for the top level categories (e.g., NOTICE) and the absolute value for the inner categories (e.g., EXCEPTION). For each category, the top red bar indicates the number of blocks of comments in the category, while the bottom blue bar indicates the number of non-blank lines of comments in the category.
Comparing blocks and lines, we see that, unsurprisingly, the longest type of comments is LICENSE, with more than 11 lines on average per block. The EXPAND category follows with a similar average length. The SUMMARY category has only an average length of 1.4 lines, which is surprising, since it is used to describe the purpose of possibly very long methods, variables, or blocks of code. The remaining categories show negligible differences between number of blocks and lines.
We consider the quality metric code/comment ratio, which was proposed at line granularity [21], [9], in the light of our results. We see that 59% of lines of comments should not be considered (i.e., categories from C to F), as they do not reflect any aspect of the readability and maintainability of the code they pertain to; this would significantly change the results. On the other hand, if one considers blocks of comments, the result would be closer to the aspect that is set to measure with the code/comment metric. In this case, a simple solution would be to only filter out the METADATA category, because the other categories seem to have a more negligible impact.
Considering the distribution of the comments, we see that the SUMMARY subcategory is the most prominent one. This confirms the value of research efforts that attempt to generate summaries for functions and methods automatically, by analyzing the source code [26]. In fact, these methods would alleviate developers from the burden of writing a significant amount of the comments we found in source code files. On the other hand, the SUMMARY accounts for only 24% of the overall lines of comments, thus suggesting that they only give a partial picture on the variety and role of this type of documentation. The second most prominent category is USAGE. Together with the prominence of SUMMARY, this suggests that the comments in the systems we analyzed are targeting end-user developers more frequently than internal developers. This is also confirmed by the low occurrence of the UNDER DEVELOPMENT category. Concerning UNDER DEVELOPMENT, the low number of comments in this category may also indicate that developers favor other channels to keep track of tasks to be done in the code.
Finally, the variety of categories of comments and their distribution underlines once more the importance of a classification effort before applying any analysis technique on the content and value of code comments. The low number of discarded cases corroborates the completeness of the taxonomy proposed in RQ1.
C. RQ3. How effective is an automated approach, based on machine learning, in classifying code comments?
To evaluate the effectiveness of machine learning algorithm in classifying code comments we employed a supervised learning method. Supervised machine learning bases the decision evaluating on a pre-defined set of features. Since we set to classify lines of code comments, we computed the features at line granularity.
Text preprocessing. We preprocessed the comments by doing the following actions in this order: (1) tokenizing the words on spaces and punctuation (except for words such as '@usage' that would remain compounded), (2) splitting identifiers based on camel-casing (e.g., 'ModelTree' became 'Model Tree'), (3) lowercasing the resulting terms, (4) removing numbers and rare symbols, and (5) creating one instance per line.
Feature creation. Table II shows some of the features we devised and all that appear in the final model. Due to the optimal set of features is not known a priori, we started with some simple, traditional features and iteratively experimented with others more sophisticated, in order to improve precision and recall for all the projects we analyzed. A set of features commonly used in text recognition [25] consists in measuring the occurrence of the words; in fact, words are the fundamental tokens of all languages we want to classify. To avoid overfitting to words too specific to a project, such as code identifiers, we considered only words above a certain threshold t. This value has been found experimentally, Pascarella, and Bacchelli -Classifying Code Comments in Java Open-Source Software Systems we started with a minimum of 3 increasing up to 10. Since the values around 7 do not change the precision and recall quality, we chose that threshold. In addition, other features consider the information about the context of the line, such as the text length, the comment position in the whole file, the number of rows, the nature of the adjacent rows, etc.
SERG
The last set of features is category specific. We defined regular expressions to recognize specific patterns. We report three detailed examples: • This regular expression is used to match comments in single line or multiple lines with empty body.
([ˆ*\s])(\1\1)|ˆ\s * \/\/\/\s * \S * |\$\S * \s * \S * \$ Machine learning validation with 10-fold. We tested both probabilistic classifiers and decision tree algorithms. When using probabilistic classifiers, the average values of precision and recall were usually lower than values obtained using decision tree algorithms, thus a minor number of comments are correctly classified. Conversely, using decision tree algorithm the percentage value associated with the correctly classified instances is better, with Random Forest we obtain up to 98.4% and the effect is that more comments are correctly classified. Nevertheless, in the latter case, many comments belonging to classes with a low occurrence were wrongly classified. Since the purpose of our tool is to best fit the aforementioned taxonomy we discovered that the best classifier is based on a probabilistic approach. In Table III we report only the results (precision, recall, and weighted average TP rate) for the naive Bayes Multinominal classifier that on average, considering whole categories, achieves a better result accordingly to the aforementioned considerations. In Table III we intentionally leave empty cells that correspond to categories of comments that are not present in related projects. For the evaluation, we started with a standard 10-fold cross validation. Table III shows the results in the column '10-fold'.
Cross-project validation. Different systems have comments describing different code artifacts and are likely to use different words and jargons. Thus, term-features working for the comments in one system may not work for others. To better test the generalizability of the results achieved by the classifier, we conduct a cross-project validation, as also previously proposed and tested by Bacchelli et al. [3]. In practice, cross-project validation consists in a 6-folds cross validation, in which folds are neither stratified nor randomly taken, but correspond exactly to the different systems: We train the classifiers on 5 systems and we try to predict the classification of the comments in the remaining system. We do this six times rotating the test system. The right-most columns (i.e., 'cross-project') in Table III show the results by tested system.
Summary. The values for 10-fold cross validation reported in Table III show accurate results (mostly above 0.95%) achieved for top-level categories. This means that the classifier could be used as an input for tools that analyze source code comments of the considered systems. For inner-categories, the results are lower; nevertheless, the weighted average TP rate remains 0.85. Furthermore, we do not see large effects due to the prominent class imbalance. This suggests that the amount of training data is enough for each class. As expected, testing with cross-project validation, the classifier performance drops. However, this is a more reliable test for what to expect with JAVA comments from unseen projects. The weighted average TP rate that goes as low as 0.74. This indicates that project-specific terms are key for the classification and either an approach should start with some supervised data or more sophisticated features must be devised.
A. Information Retrieval Technique
Lawrie et al. [14] use information retrieval techniques based on cosine similarity in vector space models to assess program quality under the hypothesis that "if the code is high quality, then the comments give a good description of the code". Marcus et al. propose a novel information retrieval techniques to automatically identify traceability links between code and documentation [19]. Similarly, de Lucia et al. focus on the problem of recovering traceability links between the source code and connected free text documentation. They propose a comparison between a probabilistic information retrieval model and a vector space information retrieval [16]. Even though comments are part of software documentation, previous studies on information retrieval focus generally on the relation between code and free text documentation.
B. Comments Classification
Several studies regarding code comments in the 80's and 90's concern the benefit of using comments for program comprehension [35], [31], [32]. Stamelos et al. suggest a simple ratio metric between code and comments, with the weak hypothesis that software quality grows if the code is more commented [27]. Similarly, other two authors define metrics for measuring the maintainability of a software system and discuss how those metrics can be combined to control quality characteristics in an efficient manner [21], [9].
New recent studies add more emphasis to the code comments in a software project. Fluri et al. present a heuristic approach to associate comments with code investigating whether developers comment their code. Marcus and Maletic propose an approach based on information retrieval technique [20]. Maalej and Robillard investigate API reference documentation (such as javadoc) in Java SDK 6 and .NET 4.0 proposing a taxonomy of knowledge types. They use a combination of grounded and analytical approaches to create such taxonomy. [17]. Instead Witte et al. used Semantic Web Technologies to connect software code and documentation artifacts [34]. However, both approaches focus on external documentation and do not investigate evolutionary aspects or quality relationship between code and comments, i.e., they do not track how documentation and source code changes together over time or the combined quality factor. More in focus is the work of Steidl et al. where they investigate over the quality of the source code comments [29]. They proposed model for comment quality based on different comment categories and use a classification based on machine learning technique tested on Java and C/C++ programs. Despite the quality of the work, they found only 7 high-level categories of comments based mostly on comment syntax, i.e., inline comments, section separator comments, task comments, etc. A different approach is adopted by Padioleau et al. [22]. The innovative idea is to create a taxonomy based on the comment's meaning. Even if it is more difficult to extract the content from human sentences, their proposal is a more suitable technique for defining a taxonomy. We follow this path in our work.
VI. CONCLUSION
Code comments contain valuable information to support software development especially during code reading and code maintenance. Nevertheless, not all the comments are the same, for accurate investigations, analyses, usages, and mining of code comments, this has to be taken into account. In this work we investigated how comments can be categorized, also proposing an approach for their automatic classification.
The contributions of our work are: • A novel, empirically validated, hierarchical taxonomy of code comments for Java projects, comprising 16 inner categories and 6 top categories.
• An assessment of the relative frequency of comment categories in 6 OSS Java software systems.
• A publicly available dataset of more than 2,000 source code files with manually classified comments, also linked to the source code entities they refer to.
• An empirical evaluation of a machine learning approach to automatically classify code comments according to the aforementioned taxonomy.
Pascarella, and Bacchelli -Classifying Code Comments in Java Open-Source Software Systems SERG | 9,559.2 | 2017-05-20T00:00:00.000 | [
"Computer Science"
] |
Composition Under Distributive Natural Transformations: Or, When Predicate Abstraction is Impossible
Natural language semanticists have often found it useful to assume that all expressions denote sets of values. The approach is most prominent in the study of questions and prosodic focus, but also common in work on indefinites, disjunction, negative polarity, and scalar implicature. However, the most popular compositional implementation of this idea is known to face technical obstacles in the presence of object-language binding constructs, including, chiefly, lambda abstraction. The problem has been well-described on several occasions in the literature, and in fact several solutions have been explored. This paper seeks to formalize the challenge of defining an indeterminate semantics for binding operators, and to formally establish the intuition that the challenge is in fact insurmountable. The primary benefit to this exercise is that it offers an abstract characterization of what it means to lift an operation from one semantic space to another, a notion which may be applied to domains having nothing to do with sets of alternatives.
Introduction
Compositional theories of natural language semantics often take the form of inductive interpretations of simply-typed lambda terms. In much early work following Montague (1973) these terms comprise the metalanguage in which denotations are expressed. Semantic analysis consists in translating natural language constituents into formal terms with variable-binding constructs, and then assigning meanings to constituents by interpreting their formal translations. In much contemporary work since the publication of Heim and Kratzer's (1998) influential textbook, object languages are themselves B Dylan Bumford<EMAIL_ADDRESS>1 UCLA, Los Angeles, USA assumed to be structured as lambda terms, with variables and variable-abstractions appearing directly in natural language abstract syntax.
Canonically, these expressions are interpreted in the familiar manner of Henkin (1950). Basic types τ are associated with domains D τ , and functional types (στ) with functions f : D σ D τ between domains. Terms are evaluated relative to assignments that map typed variables v τ to objects in the relevant domain D τ . Abstractions denote functions, modifying the assignments at which their bodies are evaluated.
For some purposes, however, these object-or meta-language lambda terms may be interpreted differently. Perhaps most prominently, Rooth (1985) defined a semantics for Montague's higher-order logic in which terms of type τ are interpreted not as functions from assignments to objects in D τ , but rather as sets of such functions. This, he argued, provides a formal strategy for comparing the meaning of an expression to the alternative meanings it evokes, crucial to understanding the effects of prosodic focus. Later work in the spirit and shadow of Rooth has extended this idea to explicate the semantic force of question particles (Hagstrom, 1998), indeterminate pronouns (Kratzer & Shimoyama, 2002), disjunctions and indefinites (Alonso-Ovalle, 2006), among other things. But in many of these sequels, syntactic terms of type τ are not interpreted as sets of functions from assignments to D τ , as in Rooth (1985), but as functions from assignments to sets of D τ values. Somewhat unfairly, but in keeping with what terminological consistency there is in the literature, I will call interpretations in this latter type signature Hamblin denotations, after Hamblin's (1973) pioneering analysis of questions. In particular, where Rooth interprets abstraction terms λv σ . φ τ as sets of functions from assignments to elements of [ D σ D τ ], 1 the Hamblin semantics defined in, e.g., Hagstrom (1998) interprets them as functions from assignments to elements of ℘ [ D σ D τ ], the powerset of [ D σ D τ ]. Shan (2004) points out that these respective semantic spaces are far from equivalent, and moreover that the denotations assigned to abstraction terms by Hagstrom and the others are inadequate to the empirical tasks they are designed for. Worse, Shan contends that it is in fact impossible to define a suitable Hamblin denotation for abstraction. Charlow (2019a) observes that the same difficulties arise in trying to define other binding constructs, including existential closure, if the denotations of their prejacents are functions from assignments into sets instead of sets of functions from assignments.
The demonstrations in Shan (2004) and Charlow (2019a;2019b) leave no doubt that the definition of abstraction in Hagstrom (1998) and Kratzer and Shimoyama (2002) is empirically defective; the reader is referred to these works for concrete illustrations of its failure, and illuminating discussion of the general nature of the problem. In this note, I wish to contribute two small things to the discussion. First, I hope to offer a somewhat more abstract criterion of adequacy for a Hamblin semantics of abstraction, or any other syntactic construct that has an ordinary interpretation of the sort discussed in the second paragraph above. The goal here is to lay out a few algebraic laws that we should expect to hold between the intended ordinary denotation of an operator and its Hamblin denotation in the more structured semantic space. Second, I will prove that 1 I will use [ D σ D τ ] to pick out the set of functions from D σ to D τ , in lieu of the more traditional D D σ τ , as the arrow notation is better suited to higher-order signatures.
for a class of such operators, including lambda abstraction, such a definition is indeed impossible, as Shan ordained. 2 In short, I will propose that an adequate semantics for abstraction depends on the existence of a natural transformation between Roothian denotations and Hamblin denotations satisfying two monad distributivity laws. These are notions from Category Theory, though I will attempt to justify them on conceptually and linguistically intuitive grounds (see Asudeh & Giorgolo (2020) for a more general linguistic introduction to the mathematics). The strategy is general enough that it should extend to analyses that locate denotations in other sorts of regimented structures, including values paired with propositions (Potts, 2005;Koev 2017), domains augmented with undefined elements (Beaver & Coppock, 2015;Grove, 2019), and values paired with continuations (Krifka, 1991;Barker, 2016). I discuss these extensions briefly in Sect. 6.
Denotations in Hamblin Space
I assume the simplest canonical definition of natural language logical forms consists in at least the following syntax-driven interpretive clauses: 3 Expressions of type σ denote functions from assignments to values in D σ . Lexical items c are looked up in a lexicon I. Variables t n correspond to projection functions on assignments. And appropriately typed binary constituents are combined by function application.
To this bare-bones formalism, any number of unary variable-binding and typeshifting constructs and/or alternative binary modes of combination may be added. Most fragments since Montague (1973) include a "lambda" of some sort that denotes a function parameterized on the value assigned to a particular variable, as in (4). Less selective binders have also been proposed (Lewis, 1975;Heim, 1982), along the lines of the universal closure operator in (5). Linguistic type-shifters have come in all shapes and sizes; Partee's (1986) Lift and Be, in (6) and (7), are paradigmatic examples of the genre. Alternative binary modes of combination oriented around Boolean operations and function composition are commonly entertained, as in (8) and (9).
Early in the Montagovian era, Hamblin (1973) sketched an analysis of questions that sought to maintain as much of the canonical fragment above as possible while assuming that 'wh'-words like 'who' and 'what' ought to behave for all compositional purposes like names. Yet of course such words cannot be assumed to refer to specific individuals. Hamblin's solution was to reanalyze all argument terms as denoting sets of individuals. Names then played the special role of terms whose denotations were determinate, or singleton. In fact all non-'wh'-words were taken to denote determinate sets of values.
Let ⦃E⦄ signify the Hamblin denotation of E. At any assignment, 'wh'-phrases denote sets of entities, suitably sortably restricted. Variables and non-'wh'-constants of type σ are interpreted, relative to assignments, as elements of ℘ D σ , that is, as sets of D σ objects.
The question then is how to restore compositional, recursive operations like those in (4)-(9), keeping in mind that their inventory may be open-ended and depend on analyses of phenomena completely independent of questions. Hamblin himself suggested that function application ought to be executed pointwise; the combination of an expression φ of type (στ) with an expression ψ of type σ should be the set of all possible applications of a function in ⦃φ⦄ to an argument in ⦃ψ⦄.
There is an obvious sense in which this definition generalizes the ordinary applicative mode of combination. Let f • a := f a be the semantic operation of function application. Then the relevant clauses of both · and ⦃·⦄ can be written in terms of •: This much is trivial, but it provides a clear schema for other modes of combination, as in (15) and (16).
Likewise, various unary syntactic operators can be associated with semantic functions. Hamblin denotations may then be generated by mapping these functions over the set of values returned by ⦃·⦄.
However, this general strategy is no help in defining Hamblin denotations for the binding constructs. What the rules above have in common is that the operations •, •, , B, and L are all assignment-independent. They are extensional. Binders, on the other hand, are definitionally assignment-modifying. In other words, there is no function A such that n σ φ = λg. A ( φ g), and so no function to stick into the translational template in (17c) and (18c).
In light of this, a Hamblin denotation for abstraction terms must be ad-hoc. Given the typing regime, the goal for any expression E of type τ is to find for ⦃n σ E⦄ a function from assignments to sets of D σ D τ functions. For instance, Kratzer and Shimoyama (2002), following Hagstrom (1998), define Hamblin abstraction as in (19).
This definition succeeds in picking out, relative to an assignment, a set of functions in the desired signature in terms of the Hamblin denotation of its prejacent ⦃E⦄. It is therefore well-typed and compositional. However, Shan (2004), Romero and Novel (2013), and Charlow (2019a;2019b) argue that this definition yields sets of alternatives that include many conceptually suspect functions, and demonstrate that it leads directly to a variety of empirical inadequacies. Could we do better? There are certainly other ways to assemble a set of functions with the right domain and codomain from ⦃φ⦄. It is hard to say a priori whether any of them would be adequate to empirical purposes, that is, whether they would "do the job" that abstractions are supposed to do. The same goes for any other binding construct. Given the lack of a systematic mapping procedure from ordinary space to Hamblin space for such constructs, how should we determine which possible function from assignments to sets of the relevant shape is appropriate for an operation, given its ordinary meaning, or whether any such appropriate denotation even exists?
In the next Section, I formalize this question and offer a criterion of adequacy for any Hamblin denotation of a compositional operation. I then prove that no denotation for an intensional operation, including variable binding, can meet this criterion.
The Formal Problem
Let υ be a unary syntactic operator, so that for some class of expressions E, if E ∈ E, then [υ E] is a well-formed constituent. Assume also that the interpretation of υ is compositional, so that for any E, υ E is a function of E . As usual, we identify υ with this function. Finally, assume that the expressions in E take denotations (relative to assignments) in some set A, and that the expressions [υ E] take denotations (relative to assignments) in a set R.
where G is the domain of assignments.
Call any such υ extensional if υ is such that for any g ∈ G and α : G A: That is, at any assignment g, υ cares only about the value that its argument α takes at g. For instance, the Partee type-shifters L and B are extensional, given the definitions in (6) and (7): Call any υ intensional if it doesn't satisfy the equation in (20). In general, any sort of binding or context-shifting operator will be intensional. For instance, the abstraction and closure constructs in (4) and (5) are intensional, as seen in (26) and (28).
Similarly, when μ is a compositional binary mode of combination φ ψ μ , we write · μ for that function which combines any appropriately typed φ and ψ to produce In this case, we say μ is extensional if · μ is such that for any g ∈ G, α : G A, and β : G B: And otherwise, we say μ is intensional. The modes of combination in (3), (8) and (9) are all extensional. Finally, as discussed in Sect. 2, for any expression E of type σ, assume that the Hamblin denotation ⦃E⦄ is (relative to an assignment) a subset of D σ . That is, ⦃E⦄ : Here then is the question: For an arbitrary operator υ, is it possible to define a sensible Hamblin denotation ⦃υ E⦄ in terms of υ and ⦃E⦄? Or, likewise, for an arbitrary mode of combination μ, is it possible to define a sensible Hamblin mode ⦃φ ψ⦄ μ in terms of · μ , ⦃φ⦄, and ⦃ψ⦄?
Say for concreteness that E has type τ and [υ E] has type ρ. Then ⦃υ E⦄ should be a function from G to ℘ D ρ , assembled somehow from ⦃E⦄ : G ℘ D τ and There are two basic obstacles here. One is that υ expects a function from G to D τ , but ⦃E⦄ is a function from G to ℘ D τ . The other is that υ returns a function from G to D ρ , but the Hamblin denotation ⦃υ E⦄ should be a function from G to ℘ D ρ .
The same issue arises with binary combination. Say φ has type σ, ψ has type τ and [φ ψ] has type ρ. Then ⦃φ ψ⦄ μ ought to be a function from G to ℘ D ρ . But the mismatches now are doubled. On the one hand, · μ expects a pair of functions G D σ and G D τ , though ⦃φ⦄ and ⦃ψ⦄ deliver a pair of functions G ℘ D σ and G ℘ D τ . On the other hand, ⦃φ ψ⦄ is supposed to deliver a function G ℘ D ρ , but · μ provides only a function G D ρ .
What is needed then is a map ϒ : [G ℘ D τ ] ℘ [G D τ ] to transform a function into sets into a set of functions. With that, the Hamblin denotations could be defined as follows: For extensional rules, it's easy enough to concoct such a map. The following would work (note the similarity with (19)), as would others.
Since an extensional υ only ever evaluates its argument at its own input g, the set determined by (31) and (33) depends only on the image of ⦃E⦄ at g. For example, The set described in (34c) is determined by the collection of functions ϕ which have the property that at any assignment h, they pick out an element of ⦃E⦄ h. There are a great many such functions. But since for each one, λk. k (ϕ g) simply projects out the value that it takes at g, almost all of the variation among the possible ϕs is irrelevant. All that matters is what value they choose at g. And as (34c) guarantees, the values available at g are all and only the values in ⦃E⦄ g. Thus the intimidating expression in (34c) is reduced to the familiar (34d), which is exactly in accordance with the schema in (17c). But for intensional operators, this ϒ is clearly inappropriate. Consider for instance the closure operator † . Perhaps the simplest cases to inspect are those in which † 's prejacent contains no free variables at all. The ordinary denotation of such an expression is equivalent to the ordinary denotation of the prejacent, since the quantification over assignments is vacuous: But things are quite different for the Hamblin denotation of such a vacuously closed expression. Assume, following the discussion in Sect. 2, that ⦃who left⦄ = λg. {left z | z ∈ D e }. Then (31)-(33) predict: Any ϕ that meets the comprehension condition in (36d) is a function that sends each h ∈ G to a proposition of the form (left z), where z is an entity. Some such functions map every assignment to left j, others to left m; others may map some assignments to left j and other assignments to left m; etc. For any of these ϕs, to say that ϕ h is true at every h is to say that every proposition in the range of ϕ is true. In other words, at any input g, the alternatives in ⦃
Proposal: Distributive Natural Transformations
I would like to suggest that any transformation ϒ worth considering ought to satisfy the following three constraints. For any types σ and τ, and any P ⊆ D σ , f : D σ D τ , ϕ : G D σ , and : G ℘ D σ : These are the natural transformation and distributivity laws when assignmentdependence and indeterminacy are considered as monads. More on this in Sect. 5. But first, let's see what they amount to for practical linguistic purposes.
Imagine that an operator's prejacent contains no variables, traces, pronouns, etc. The Hamblin denotation of such a constituent will be a constant function from assignments to some set P ⊆ D σ . What set of assignment-dependent values should ϒ return? Left guarantees that in this case, the set we get back is just the set of constant functions into the elements of P.
Assume, as is common, that focus alternatives are calculated using Hamblin denotations. That is, prosodically focused expressions contribute ordinary denotations to ordinary truth conditions, but at the same time raise the specter of alternative denotations from the same domain. These alternative values contribute to calculations of alternative propositions, ideas evoked but not uttered. For instance, the alternatives elicited by placing stress on 'John' in 'John smiled' are just propositions about other people smiling (Mary smiled, Bill smiled, etc.). Left ensures the result in (37). Upon repackaging, the alternatives of 'John F smiled' are just the ordinary denotations 'z smiled' for other type-e expressions z.
More generally, if [· · · x F · · · ] is variable-free, then Left guarantees the identity in (38), where a ∼ b means that a has the same type as b. In other words, in purely extensional contexts where assignments are irrelevant, focus alternatives can be computed from ordinary denotations by simply replacing focused constituents with expressions of the same type.
Now imagine that an operator's prejacent contains no interesting alternativegenerating language; no 'wh'-words, focused constituents, etc. Its Hamblin denotation ϕ, at any assignment, is just the singleton set of its ordinary denotation at that assign-ment. Then Right guarantees that when this denotation is passed to ϒ, the result contains exactly the one function that maps any g to the single element in ϕ g. 5 For instance, assume that the Hamblin denotation of a vanilla declarative sentence, like 'he smiled', is just the function λg. t 1 smiled g , mapping assignments to singleton ordinary denotations. Then Right ensures: More generally, if E contains no alternative-generating expressions (is focus-free, etc.), then Right guarantees: At the very least, this ensures that abstraction satisfies η-reduction in the absence of focus. Assume that E contains no focus-marking and no free occurrence of t n (the latter is only important to justifying the step from (42f) to (42g)). ⦃n Nat is more complicated, but guarantees that ϒ is not ad-hoc, in the sense that it should perform "the same" operation no matter what size or type of denotation it is passed. To see this, let R and H be defined as in (43) and (44).
This rule, incidentally, is in line with Charlow's (2019a) interpretation of the abstraction rule proposed in Kotek (2017): ⦃n E⦄ := λg. λa. ιp. p ∈ ⦃E⦄ g n →a [Charlow 2019a: (16)] In other words, Kotek can be seen as adopting the special case of Right that applies to lambda abstraction, though she also argues, in effect, that when Right is inapplicable, abstraction should be undefined. In any case, the point is that Right does the only reasonable thing when alternatives are trivial (i.e., when Hamblin denotations are isomorphic ordinary denotations).
Both of these operations have a claim to the name Pointwise Function Application. The former is just Hamblin's version of application, introduced in (13), reified as a combinator. It combines an assignment-dependent set of functions from A to B with an assignment-dependent set of values in A to produce an assignment-dependent set of results in B. The latter is the Roothian analog of this. It combines a set of assignmentdependent functions from A to B with a set of assignment-dependent values in A to produce a set of assignment-dependent results in B. In other words, (43) would combine two denotations before ϒ has gone to work on them, (44) after. What Nat guarantees is that when is deeply uninteresting, it won't make any difference which of (43) or (44) you pick. Let = λg. { f }, for some function f : A B. If this is the Hamblin denotation of an expression, then that expression must be free of any alternative-generating constituents (declarative, unfocused, etc.), since the value at any assignment is a singleton set of alternatives. Likewise, it must be free of any assignment-sensitive constituents (traces, pronouns, etc.), since the value at every assignment is the same. And in fact, every such focus-and pronoun-free expression will have as its Hamblin denotation some function in the shape of . In these circumstances, Nat is exactly the requirement that ϒ is a homomorphism with respect to the two notions of Pointwise Function Application.
ϒ (
This, together with Left and Right, ensures that the alternatives generated by ϒ depend only on the anaphoric and indeterminate components of its prejacent. Ordinary content rides free. For instance, consider the sentence 'her mom F called John', with the referent of 'her' free and prosodic focus on mom. This sentence contains a mixture of anaphoric, indeterminate, and plain language. By Nat, we are assured that Hamblin application H of the predicate 'called John' to the subject 'her mom' satisfies the following equations: Combining this with the general recipe in (31), we are guaranteed that abstracting over n in the constituent [t n 's mom F called John] will generate the set of alternative properties in (47).
⦃n [t n 's mom F called John]⦄ = λg. λx. call j (ϕ g n →x ) | ϕ ∈ ϒ ⦃t n 's mom F ⦄ For each assignment-dependent entity ϕ that ϒ transforms ⦃t n 's mom F ⦄ into, we get a function that is true of x whenever ϕ g n →x called John. The important thing to see here is that the property of John-calling is what each alternative ϕ is tested for, but by Nat, this property can play no role in determining what the alternatives actually are. Those are determined entirely by the constituents containing pronouns and focus.
The Impossibility of Hamblin Abstraction
Here is the main result: the three requirements Left, Right, and Nat are together inconsistent. There cannot be any such ϒ that is well-behaved on focus-free and pronoun-free language. And consequently there cannot be any means of interpreting an intensional operator [υ E] in terms of its ordinary behavior υ , and its prejacent's Hamblin denotation ⦃E⦄.
The result actually has nothing to do with assignments, variables, or "alternatives" per se, as others have deduced (e.g., Rooth 1985: ch. II, pt. 3; Heim 2011: sec. 3; Charlow 2019a: pg. 10). So what is proved here is a more general fact about functions with certain signatures. We start by situating Hamblin's and Rooth's typing assumptions in a more abstract setting. Let S, 1 S , S and R, 1 R , R be the following functors, where R is any fixed set.
S and R are functions from domains to domains. For any set A, S A returns the powerset of A, and R A returns the set of functions from R to A. These constructors are both equipped with two polymorphic operations, a unit and a map. For any domain A, the unital operation lifts objects in A to objects in the relevant domain image. The mapping operation lifts functions between domains f : A B to functions between their images, e.g., S f : S A S B. To say that R and S are functors is to say that these latter operations are homomorphisms for function composition, as the reader may verify: From these, define the two composite functors RS, 1 RS , RS and SR, 1 SR , SR . These are simply the compositions of the constructors R and S, together with the compositions of their associated maps.
When we fix R to be the domain of assignments G, the first composition RS is the set of possible Hamblin denotations. For any expression E of type σ, the set RS D σ contains ⦃E⦄. The latter composition SR is the set of possible denotations à la Rooth.
The question put to us in (30) turns on what kinds of transformations between RS and SR are possible. For our purposes, let us say that a transformation ϒ : RS SR is a polymorphic function from the space of Hamblin denotations to those of Rooth. That is, for any domain A, ϒ determines a function from RS A to SR A.
Then we say ϒ : RS SR is natural iff the following diagram commutes for all sets A and B.
That is, for any f : A B and : RS A, we have the following equivalence. When R = G, this is the eponymous law proposed in Sect. 4.
We say a transformation ϒ : RS SR is distributive iff the following diagrams commute.
That is, for any A ∈ S A and ϕ : R A, we have the following equivalences. Again, when R = G, these are the laws of Sect. 4.
Claim No transformation ϒ : RS SR is both natural and distributive.
Proof Consider the following model. The functions f 1 , f 2 , and f 3 map elements of A to truth values. The first of these f 1 is the characteristic function of even numbers in A. The second is the characteristic function of the numbers in A that are factors of 10. And the third the characteristic function of numbers less than 4. As we'll see, the crucial configuration in this model is that the three functions cross-cut A in three different ways, relative to : : Let ϒ be a transformation RS SR, and assume for the purposes of contradiction that ϒ satisfies Nat, Left, and Right. Since ϒ maps any object of type RS A to an object of type SR A, it must in particular map to some set of functions ϒ in Thus there is no natural, distributive ϒ : RS SR, since any such function would have to make all three diagrams commute for at f 1 , f 2 , and f 3 , which is impossible.
It should be clear that I have used numbers and truth values here just to make the functions simple, familiar, and describable. The reasoning is entirely abstract as regards the elements of R, A, and B. In particular, when R is G, we are assured that there is no natural, distributive transformation between the two denotational spaces for composing alternatives at issue in this paper. Moreover, it follows immediately that any parameter to the ordinary denotation function · will suffer the same fate. This includes worlds, times, indexical contexts, judges, etc. And consequently, we cannot expect to define abstractions or any other modal operators.
Outlook
I have argued that the Hamblin denotation of an intensional rule of composition depends on the existence of a transformation ϒ : (31) and (32), repeated here.
The narrow formal result established in the previous section is that no such transformation can be both natural and distributive in the sense of Left, Right, and Nat. I've also suggested that these are requirements we should expect any decent Hamblin definition to meet. They guarantee very basic things about the correspondence between a rule's ordinary behavior and its Hamblin behavior. But strictly speaking, the logic here is very weak. One is free to reject the entire framing in terms of (66) and (67) and/or reject any of the distributivity and naturality laws. Ultimately the question is about when a definition counts as a reasonable generalization of an operation. Of course, in analyzing a particular construct, we may assign it an arbitrary Hamblin denotation to see if it makes predictions we like. What the proof here shows is that the hopes of doing this in any systematic way, any way which is neutral about the ordinary semantics of the construct, are dashed.
Before closing, let me also point out that the laws here are highly abstract. They depend only on the functorial nature of S and R. And sets are not the only functors that have been exploited to structure natural language denotations. To give two examples: First, Heim & Kratzer (1998) interpret expressions of type σ as denoting partial functions from assignments to D σ . The denotation of an expression like 'her 1 guitar' is only defined for assignments that map 1 to an individual with a guitar. Let M A = A ∪ {⊥}, where ⊥ is an object not in the domain of any type. Then we may identify the partial denotation of an expression E σ as a function from G to M D σ . If E σ is defined at g, then the partial denotation and the ordinary denotation coincide; otherwise E σ g = ⊥.
Second, for various purposes ordinary denotations can be paired with supplemental content (e.g., Potts, 2005;Martin, 2013;Koev 2017). For concreteness, say that an expression like 'John, a linguist' denotes an entity paired with the proposition that John is a linguist. Let W A = A × D π , where π is the type of propositions. Under such assumptions, the bidimensional denotation of an expression E σ lies in G W D σ ; at any assignment, it denotes a pair whose left component is in D σ and whose right component is in D π .
Both M and W are functors, under the associated maps below (where is the denotation of a tautology): Consequently, their compositions MR, RM, WR and RW are all functors as well. Unsurprisingly, when attempting to extend ordinary compositional rules to rules in these enriched spaces, we run into precisely the same obstacles as in Sect. 2. Namely, the only way to utilize υ is to provide it with a function ϕ : Outfitted with such transformations, the following general templates would allow operations to be lifted as follows: where ϕ, p = ϒ W E (75a) These transformations have corresponding naturality and distributivity laws, derived by replacing S with M or W in the diagrams above. For any types σ and τ, and any x ∈ D σ , p ∈ D π , m ∈ M D σ , f : D σ D τ , ϕ : G D σ , and : G M D σ (or : G W D σ , as appropriate): I will mostly leave to future research the investigation of what exactly these constraints impose on transformations. It is clear that there are solutions in both cases. For instance, the ϒs defined in (76) and (77) satisfy their three respective laws.
In (76) any total function is returned as is; any (non-total) partial function is sent to ⊥. This yields a kind of Weak Kleene semantics for binding operators, such that if any assignment would lead the prejacent to failure, then the entire operation is undefined.
(77) is considerably more arbitrary. For any : G W D σ , the transformation ϒ W returns a pair whose left component is the function that at any g projects the left component of g, but whose right component is whatever supplemental content takes at some fixed assignment g * . I take it to be obvious that at least the second of these would yield a completely inadequate semantics for, say, lambda abstraction. Clearly if these laws are relevant to the program of lifting denotations from one setting to another, they are merely necessary conditions. But then it is perhaps all the more surprising that no transition from Hamblin to Rooth can meet even this minimal muster.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. | 7,910 | 2022-04-21T00:00:00.000 | [
"Linguistics",
"Computer Science"
] |
Dust observations with antenna measurements and its prospects for observations with Parker Solar Probe and Solar Orbiter
. The electric and magnetic field instrument suite FIELDS on board the NASA Parker Solar Probe and the radio and 15 plasma waves instrument RPWS on the ESA Solar Orbiter mission that explore the inner heliosphere are sensitive to signals generated by dust impacts. Dust impacts were observed using electric field antennas on spacecraft since the 1980s and the method was recently used with a number of space missions to derive dust fluxes. Here, we consider the details of dust impacts, subsequent development of the impact generated plasma and how it produces the measured signals. We describe empirical approaches to characterise the signals and compare to a qualitative discussion of laboratory simulations to predict signal shapes 20 for spacecraft measurements in the inner solar system. While the amount of charge production from a dust impact will be higher near the sun than observed in the interplanetary medium before, the amplitude of pulses is determined by the recovery behaviour that is different near the Sun since it varies with the plasma environment
Introduction
The space missions Parker Solar Probe and Solar Orbiter to explore the inner heliosphere and close vicinity of the Sun carry antennas experiments that respond to dust impacts onto the spacecraft.Parker Solar Probe (Parker Probe) is a NASA mission that was launched in August 2018 and that studies the vicinity of the Sun at closest distance ~ 10 solar radii near the solar equator.The mission payload includes the electric and magnetic fields instrument suite FIELDS (Bale et al. 2016).Solar Orbiter is an ESA mission with a launch planned in 2020 (Mueller and al., 2019).
It will study the vicinity of the Sun as close as 0.3 AU and at maximum 35° inclination from the Solar equatorial plane includes the Radio and Plasma Waves (RPW) experiment (Maksimovic et al., 2019).Dust impacts are observed with electric antennas for field measurements since the 1980s beginning with the Voyager mission (cf.Gurnett et al., 1997, Meyer-Vernet, 2001).The method was recently used with a number of space missions to derive dust fluxes.While antenna measurements do not replace those of dedicated dust detectors, they are interesting because many space missions carry electric field instruments than carry dedicated dust detectors.In addition, antenna measurements can observe lower dust fluxes because of their large collecting area of the whole spacecraft in comparison to the small collecting area of dust detectors.A limitation of the antenna measurements is that they do not provide information on dust composition and only limited information, if at all, on impact direction and dust mass.These derived values are highly uncertain (Meyer-Vernet et al. 2009, Zaslavski et al. 2012, Malaspina et al. 2015).The relationship between dust impacts and the signals they produce in electric field instruments has also been considered in new instrument development and laboratory measurements.
Cosmic dust particles are one of the major constituents of the interplanetary medium in the inner heliosphere and knowledge on dust near the Sun is so far based on remote observations and model assumptions.An exception are the measurements of the HELIOS mission with two spacecraft that reached a minimum distance 0.31 AU from the Sun and each carried a dust detector (Grün et al. 1980).Our basic understanding (see, e.g.Mann et al. 2004) is that large (> micrometre) dust particles that are fragments of comets and asteroids are in Keplerian orbits around the Sun.Their velocities and number densities increase with decreasing distance from the Sun.Fragments are produced in dust-dust collisions for which the rates increase with decreasing distance from the Sun.The majority of fragments smaller than micrometre are pushed outward by radiation pressure (cf.Wehry and Mann 1999) and deflected by electro-magnetic forces (cf.Mann et al. 2004, Czechowski andMann 2010).In addition to the interplanetary dust, interstellar dust particles stream into the inner heliosphere from interstellar upstream direction and move ~parallel to the ecliptic plane (Mann 2010).Because of repulsion by the radiation pressure force, only the large interstellar dust reaches the inner heliosphere.A large fraction of dust is destroyed in the inner heliosphere, in sublimation and other destruction processes and this generates a dust-free zone.Sun-grazing comets (cf.Jones et al. 2018) are a local highly variable source for dust particles (cf.Fig. 1).Dust material is released in the ambient solar wind, a process which is not well quantified.The solid fragments that are not fully destroyed are pushed outward and produce the small size portion of the interplanetary dust flux observed near Earth orbit.Parker Probe and Solar Orbiter will for the first time explore the inner heliosphere in-situ.The dust impacts on the spacecraft will influence electric field measurements on these spacecraft and provide an opportunity to study the dust environments of the inner heliosphere.and atomic and molecular ions are produced by thermal ionisation of the impact vapour.The initial ionisation is followed by recombination and thermalisation and a residual ionisation remains in the impact vapour.The amount of the residual ionisation can be derived from laboratory measurements of the charge production.An empirical relation describes the charge production , as a function of the dust mass, and speed, according to = , where is given in Coulombs, the in kilograms and in km/s.The exponents and are dimensionless and determined from experimental data.The constant gives the proportionality and, as parameters and , it is dependent on both impactor and target composition.This impact ionisation model (cf.Drapatz and Michel, 1974) is in a good agreement with laboratory experiments in relatively thin targets, for speeds between the supersonic limit and some tens of km/s.The model bases on the assumption that neutral vapour forms in the impact process and that it subsequently ionises because of its high temperature.Ionisation can also occur directly in the target where the propagating shock wave leads to a high-pressure high temperature state so that the ejected target material is already ionized (see Hornung and Kissel 1994).The properties of the vapour cloud are also depending on material composition and on the impact angle (cf. Sugita et al. 1998).Moreover, the shock wave propagation in the target (spacecraft) and dust material produces not only vapour, but also solid fragments and the fragment formation is predominant at smaller impact velocities (Jones et al. 1996).In summary, there is no theory that fully describes the charge production, sometimes also denoted as plasma production, that is caused by dust impacts onto spacecraft.The following discussion uses the term impact cloud for the impact generated electrons, ions, and neutrals to avoid confusion with the surrounding plasma.
Charge production is often determined empirically in dust accelerator facilities, in recent years at those of the University of Stuttgart (Mocker et al., 2011) and the University of Colorado (Shu et al., 2012) for the range of impact velocities shown in Figure 2. The parameters to describe impact charge production derived from observations vary strongly for different impact materials; has reported values between 0.7 and 1 while has been measured between 2.5 and 6.2 (see e.g.Dietzel et al. 1973, Auer 2001, Collette et al. 2014, and references therein).
An often-used relationship for dust impacts on spacecraft is ≈ 0.7 3.5 , which was reported for aluminium targets (McBride and McDonnell, 1999).It should be noted that the exponents also change for low impact energies (speeds below ~1 km/s and sizes below ~10 nm).For low energy collisions, where fragments of significant sizes compared to the initial impactor survive, there may be surface effects such as capacitive contact charging (see e.g.John et al. 1980).The exponent also changes at high impact energies for speeds above ~50 km/s and dust sizes above ~1 (Auer 2001;Göller and Grün, 1989).Collette et al. (2014) measured the charge impact yield as a function of impact velocity for common materials used on spacecraft, they point out the need for dedicated studies for > 50 km/s impacts.For very large impact energies, where the impactor gets completely vaporised, surface effects are negligible and the charge generation can be modelled through hydrodynamic theory (Hornung and Kissel, 1994).Moreover, there is a dependence of impact angle on the charge generation (Schulz and Sugita, 2006;Collette et al. 2014).Based on spectroscopic analysis of 4.7 -5.6 km/s impact flashes Sugita et al. (1998) find temperatures of about 4000 K to 5000 K for the impact vapour cloud.Subsequent observations yield temperatures of 0.9 to 3 eV for impact speeds varying from 10 to 40 km/s (Miyachi et al. 2008).Laboratory measurements find for the impact vapour ion temperatures of about 5 eV at 4 km/s impact speed and > 10 eV at 20 km/s (Collette et al. 2016).
The Impact Process
Formation of the dust impact signal involves the dust impact process, the interaction of the impact cloud with the surrounding plasma and finally the detection by electric field measurement.At the most basic level, dust impacts on the spacecraft body generate clouds of free electrons and positive or negative ions.These charged particles are attracted to, or repulsed from, the spacecraft surface according to its electric potential relative to the surrounding ambient plasma.Charged particles from the impact cloud can be re-collected by the spacecraft or escape to free space and generate a transient deviation from the equilibrium spacecraft surface potential.The potential change can be positive or negative, e.g., escaping electrons generate a positive signal.Electrons are significantly faster than ions, thus the signal generated by escaping electrons appears before the signal generated by escaping ions for the case when the spacecraft potential is not too large.The amplitude of the spacecraft potential deviation is given by the amount of escaped charges and by spacecraft capacitance.The spacecraft potential relaxes back to the equilibrium value via interaction with ambient plasma according to ~−/ , where is a characteristic relaxation time (Meyer-Vernet, 1985).
The different phases of the impact process for various spacecraft potentials (slightly positive, zero, and slightly negative) are illustrated in Fig. 3.In the first phase (T1), at which the spacecraft is assumed to be in equilibrium potential, the impact occurs, and an impact cloud is generated (green).Some of the cloud particles may be recollected.The second phase (T2) is characterised by electron escape and partial recollection depending on the target's potential, yielding an initial rise in signal strength (blue).The third phase (T3) is characterised by the ion escape, decreasing the spacecraft potential (red).The final phase (T4) is the relaxation phase when the spacecraft potential returns to the equilibrium value (orange).Individual time steps are summarised in Table 1 and sketched in Fig. 3.The ratio between escaping electrons and ions in phases T2 and T3 depend on the spacecraft potential.
For example, more electrons than ions leave for a negatively charged spacecraft (left part of Fig. 3).
Impact cloud generation and expansion -T1
Charged particles at a small distance from the spacecraft body still influence its potential.The change in the spacecraft potential can thus not be observed directly after impact cloud formation but when charged particles are recollected or escape far enough and/or are sufficiently shielded by the ambient plasma or photoelectrons that their influence on the spacecraft potential is reduced (Meyer-Vernet et al. 2017).The number of escaping particles depends on initial impactor energy and velocity after the initial cloud expansion.It is possible to assume that impact cloud electrons move in random directions due to collisions with ions.This implies that half of the electrons move towards the spacecraft before they are influenced by spacecraft potential, whereas the other half moving initially outwards is recollected if the target potential is positive and higher than their temperature (in eV).An alternative model assumes approximately half of both the electrons and the ions move towards the spacecraft.This model is in good agreement with recent laboratory measurements (Nouzák et al., 2018, see below).It bases on the assumption that the impact vapour is initially neutral and that the free charges form as a result of thermal ionisation in the impact vapour.
Electron escape -T2
The first part of the signal shown in Fig. 3 is generated by electrons escaping from the spacecraft body.The amplitude of the electron signal is reduced when the spacecraft is charged positively because some electrons are attracted back to the spacecraft.All electrons are re-collected when the positive spacecraft potential is significantly higher than the temperature of electrons (no electron part in the signal).This is a very fast process (µs) and the characteristic time depends on a number of parameters.Independent from the ambient plasma this process is influenced by the geometry of the system and specifics of the antenna and parts of the spacecraft body as well as the energy (velocity) of the electrons.The cloud expansion and internal shielding depends on the size of cloud formed by the impact.In space, the ambient plasma Debye length and magnitude of the photoemission current from the spacecraft determine the length scale of the spacecraft potential influence on the expansion.
Ion escape -T3
The ion escape works in a similar manner as electron escape, but happens at a lower rate if electron and ion temperature are comparable.The potential induced by ions on the spacecraft and antennas is progressively shielded by the ambient plasma electrons and photoelectrons.Moreover, escaping electrons can drag some of the ions behind them.This can result in double population of escaping ions: fast and slow.All ions are recollected when the negative spacecraft potential is significantly higher than the temperature of ions.
Relaxation -T4
The spacecraft potential returns back to the equilibrium value due to interaction with ambient plasma.The relaxation time is determined by the ambient environment (plasma density, temperature, photoemission), and by the capacitance of the spacecraft and antennas.On the other hand, higher plasma density and stronger photoemission result in stronger currents from ambient plasma and thus in a significantly shorter relaxation time.
The typical relaxation time measured by various spacecraft is in the range from ≈ 100 µs up to several ms.The relaxation time could be comparable or shorter than the ion escape or shielding time in dense plasma environments (or under strong photoemission), and by the capacitance of the spacecraft and antennas.This will result in a reduction of the detected signal and lowers the sensitivity of dust detection via electric field antennas.Relaxation time can also be reduced by active experiments, for example by ASPOC (Active Spacecraft POtential Control) (Vaverka et al., 2017b).
Antenna Signal Shapes
Electric field antennas can be operated as a dipole, where the voltage difference between two antenna booms is measured, or a monopole, where the voltage difference between an antenna boom and the spacecraft body is measured.It has been noted that the power spectral density of dust impact signals measured by monopole antenna is significantly larger than that measured by the dipole antennas (Meyer-Vernet 1985, Tsintikidis et al., 1994;Meyer-Vernet et al., 2014), and this difference is attributed to the low sensitivity of a symmetric dipole antenna to dust impacts on spacecraft body.It is important to note that dust impact on a spacecraft body described by this model, can be detected by the monopole electric field antenna as a potential drop between the spacecraft body and one antenna.A dipole configuration measuring electric field as a potential difference between two antennas can be utilized to detect a signal only when escaping electrons or ions influence the potential of one of the dipole antennas to be used for calculating the density of dust populations encountered by the spacecraft (Ye et al., 2016).
A few cases of impact events are shown in Fig. 4 derived from laboratory studies on scaled down Cassini model (Nouzák et al., 2018).Although the signals are measured in dipole configuration, since the dust impacts one of the dipole antennas, this configuration corresponds to monopole measurement when dust impacts the spacecraft body as described above.The laboratory experiment is performed in the vacuum chamber without ambient plasma.
Relaxation process is simulated by discharge of electronics circuit inside of the Cassini model.
Figure 4 shows different signal shapes measured in the Cassini laboratory simulation and the signal development of the different stages are described for each case in Table 2: • The signal shown in panel (A) is for a strongly negatively biased target potential.All electrons are repulsed from the spacecraft and all ions are recollected back to a strongly negatively biased target.The ion escape part (red) is not apparent in this case.The electron part (blue) is followed directly by relaxation (orange).
• Panel (B) describes the signal shape for a reduced negative target potential.The number of escaping ions increases with the reduction of the negative potential (panel B).The electron part (blue) is followed by the smaller ion part (red) and relaxation (orange).A kink appears between the ion part (red) and relaxation (orange) see a left panel in Fig. 3.
• Panel (C) describing the signal measured at an unbiased target shows that similar numbers of electrons and ions escape.The amplitude of the electron part (blue) is similar to the ion part (red).
• Panel (D) shows the case of positively charged target.The number of escaping electrons is reduced and the ion part of the signal exceeds the electron one.This results in a bipolar signal where the first part is typically called "pre-spike" (Collette et al. 2015, Thayer et al. 2016).A larger number of escaping ions change the polarity of the signal.
• Panel (DD) shows the signal for a higher positive target bias potential-The first (electron) part of the bipolar pulse is reduced with increasing positive target potential.
• Panel (E) shows a case of even higher positive bias.All electrons are re-collected in this case.The signal has no electron (blue) part and it has no "pre-spikes" in this case.
The shapes of all pulses measured in the laboratory for various biases can be explained by the model described above.It must be noted, however, that since electron and ion escapes are very fast processes (~µs), detection of a detailed structure of initial parts of pulses including "pre-spikes" thus requires fast electronics (sampling in the order of 100 kHz).Therefore, not all spacecraft are able to detect them, and a thorough inquiry into signal shapes using in-situ data is difficult. .The antenna signal can be also affected by the response of instrument electronics.
For example, transfer function of electronics can modify a shape and duration of the measured signal (Ye et al., 2019).
Antenna signals observed in previous space missions
Detection of dust impacts with antenna measurements has recently been done in several space missions.In the following, we discuss the major findings related to dust detection from the respective missions.(Pantellini et al., 2012), producing a ratio between antenna voltages in agreement with the mechanism producing the pulses (Zaslavsky et al. 2012).The formation of the signal involves a transient local perturbation of the photoelectron equilibrium current on the antenna being close to the impact.The steps that lead to the antenna signals have been studied with plasma simulations and semi-empirically (see e.g.Pantellini et al., 2012;Meyer-Vernet et al., 2014;Zaslavsky, 2015).Kellogg et al. 2018 suggested that STEREO does not observe nanodust, but did not propose an alternative mechanism able to explain the observations.The larger dust impacts observed with STEREO/WAVES are observed with similar amplitudes at all three antennas.
Based on STEREO/WAVES Zaslavsky (2015) proposed a model accounting for electric pulses generation by electron collection after an impact, linking the shape and amplitude of the electric signals to the dust and local plasma parameters.Figure 5 shows the model applied to typical impact clouds.
Cluster
The Cluster mission launched in 2000 consists of four identical spacecraft orbiting the Earth in close formation.
The highly elliptical orbit (4-20 Earth radii) crosses various parts of the Earth's magnetosphere.Each spacecraft is equipped with two pairs of dipole electric field sensors (on 88 m booms tip-to-tip) (Gustafsson et al., 2001).Wide Band Data (WBD) instrument provide data of single electric or magnetic field component with a high sampling frequency in three modes (27.4 kHz, 54.9 kHz, and 219.5 kHz) (Gurnett et al., 1997).This resolution is sufficient to detect signals triggered by dust impacts.The dipole configuration is not sensitive to dust impacts on the spacecraft body.Some signal can be detected only after a direct dust impact on the one of the antennas or when the expanding impact cloud influences the potential of the antenna.On the other hand, Cluster 1 operates with the only one remaining probe in the monopole configuration since 2009 (three probes have been lost during time).This situation makes the detection of dust impacts by the Cluster 1 spacecraft possible (Vaverka et al., 2017a).On the other hand, a presence of a large number of natural waves including electrostatic solitary waves in the Earth's magnetosphere significantly complicates such detection (Vaverka et al., 2018).The fact that solitary waves are much more numerous than the expected amount of detected dust grains makes a reliable detection of dust impacts by the Cluster spacecraft very challenging.For this reason, the Cluster spacecraft are not optimal for dust studies.
MMS
The MMS mission consists of four Earth-orbiting spacecraft lunched in 2015 (Burch et al., 2016).While the missions are similar, the MMS electric field instruments just slightly differ from the Cluster ones.Each of the spacecraft is equipped with three pairs of electric field probes, two in the spin plane (120 m tip-to-tip) and one in the axial plane (29 m, Torbert et al., 2016).The electric field is measured in the dipole configuration in all three directions with sampling frequency up to 8 kHz (burst mode) and up to 256 kHz in wave burst mode.The main difference is that the instrument operates simultaneously also in the monopole configuration.The combination of dipole and monopole measurements provides a complex information about the ambient electric field and the spacecraft potential which is possible to use for the reliable identification of dust impacts.Solitary waves and other structures in the ambient plasma or electric field generate simultaneously pulses both in monopole and dipole configuration.On the other hand, changes in the spacecraft potential triggered by the dust impact generate identical pulses on all monopole antennas and no signal in the dipole configuration (electric field data).This allows us reliably distinguish changes in the spacecraft potential from the other pulses as solitary waves (see Vaverka et al., 2018).A measurement with MMS, shown in Fig. 6, illustrates the different detections in monopole and dipole configuration.No evidence for moon-related dust rings or dust lifted from the surface (e.g.Sanchez-Lavega et al. 2015) was found with MAVEN LPW.
Wind
The NASA Wind spacecraft launched in November 1994 with the goal of studying the solar wind upstream of (https://cdaweb.sci.gsfc.nasa.gov/index.html/).
Cassini
The Cassini Radio and Plasma Wave Science (RPWS) instrument measures oscillating electric fields over the frequency range 1 Hz to 16 MHz and magnetic fields in the range 1 Hz to 12 kHz (Gurnett et al., 2004).The instrument uses three nearly orthogonal electric field antennas (Eu, Ev, Ew, each 10 m long and 2.86 cm in diameter) and three orthogonal magnetic search coil antennas.The Eu and Ev antennas are often used together as a dipole antenna and Ew and the spacecraft body as a monopole antenna (Gurnett, 1998) (WBR) of the RPWS instrument was switched from monopole mode to dipole mode at a ring plane crossing, so that the responses of these two antenna modes to dust impacts were compared, assuming the dust density and size distribution did not change across the ring plane (Ye et al., 2016).Figure 8 shows an RPWS wave power spectrogram, which covers a one-hour period around a ring plane crossing on DOY 001, 2016.As the antenna mode switched from monopole to dipole at the ring plane at ~10:30, the spectral power decrease was accompanied by a significant decrease in the negative impact rates (blue) and the polarity ratio jumping to ~1.The spectral power is proportional to the product of impact rate and average voltage jump size squared (Meyer-Vernet, 1985).So, the difference in spectral power at the antenna switch could be due to either lower impact rate or smaller average voltage pulse size, or both.
In figure 9, we show a comparison of the vertical dust density profiles measured by RPWS Wideband Receiver (WBR) and the Cassini Dust Analyzer (CDA) High Rate Detector (HRD) during the ring plane crossing on DOY 361, 2016.HRD uses polarized foils for dust detection and can measure high impact rates of particles bigger than a size threshold that depends on the impact speed (Srama et al.,2004).Discontinuities in the RPWS dust density profile are due to gain changes of WBR.The CDA data showed consistent peak densities around 0.04 m -3 (threshold ~ 0.8 micron) during the Ring Grazing orbits, less than one order of magnitude higher than the RPWS dust density, which is within the uncertainty limit of the method (Ye et al., 2014).The density peak measured by RPWS (FWHM 600 to 1000 km) is wider than that by CDA (averaged profile shows a FWHM of 475 km).This difference is discussed in detail in Ye et al. (2018a).The E ring density structure based on RPWS measurements has been shown to be consistent as well with that revealed by optical observation (Ye et al., 2016a).Ye et al. (2016b) compared the data collected with these two antenna setups and found that the wave power spectral density observed by the monopole antenna is ≈ 10 dB higher than that observed by the dipole antenna.This does not necessarily mean that the monopole antenna is more sensitive to individual dust impacts, because direct comparison of the waveforms observed by these two antennas showed that the sizes of the voltage jumps induced by dust impacts are comparable.Comparison of the impact rates showed that the monopole antenna detects ≈ 10 times more dust impacts than the dipole antenna.This difference in impact rates is roughly in line with the difference in the effective impact areas of the spacecraft body and the dipole electric antenna.Detailed analysis showed that the polarity ratio of the impacts detected by the dipole antenna changes with the projected area ratio of the dipole antenna elements (Eu and Ev) as the spacecraft rotates, providing strong evidence that the dipole mode detects primarily impacts on the antenna booms.
Cassini cruise measurements between 1 and 5 AU also enabled us to study the rise time of the impact ionisation pulses as a function of dust mass and of heliocentric distance (Meyer-Vernet et al. 2017), a quantity of great importance for future missions since it determines the frequency range and voltage amplitude for dust detection.
Dust in the inner heliosphere
Many dust observations describe the dust flux near Earth orbit, it can be estimated from meteor observations, crater statistics and measurements from spacecraft.Based on these sources, an empirical polynomial mass distribution was found (Grün et al., 1985;Ceplecha et al., 1998).Observations by the STEREO spacecraft allowed extending this distribution to smaller masses (Zaslavsky et al., 2012, Meyer-Vernet et al., 2009, Malaspina et al. 2015).There is still an uncertainty, however on the absolute flux values.Estimates of cosmic dust fluxes near 1 AU and onto Earth range over several orders of magnitude and are based on a number of different assumptions (cf.e.g.Nesvorný et al. 2011, Mann et al., 2011;Plane, 2012).There is even less known on the dust flux inside 1 AU and estimates are often based on extrapolation of the flux curve obtained near Earth and considering the major forces acting on dust particles (Mann et al. 2004): Large dust particles ("micron dust" m > 10 -14 kg) are mainly influenced by gravity force and move in Keplerian orbits superimposed by a slow migration inward caused by the Poynting-Robertson effect.For dust with masses 10 -19 kg < m < 10 -14 kg ("beta meteoroids") the radiation pressure force is comparable to the gravitation; when these small particles form, typically by collisions of larger dust, they move outward in hyperbolic orbits (Czechowski and Mann 2007).For even smaller dust with m < 10 -19 kg (nanodust) electromagnetic forces prevail, they are deflected in a way that is similar to the pick-up process of ions that newly form in the solar wind (Mann et al., 2010).
The observational studies of scattered light and thermal emission from the dust are constrained by large contributions from dust near Earth to the brightness because the observed brightness is an integrated signal along the line of sight (Mann et al., 2004).Since early infrared eclipse observations (MacQueen, 1968) showed irregularities in the slope of the corona brightness with varying distance from the solar limb, the possibility of dust rings existing around the Sun is discussed.Indeed, model calculations were made to show that dust rings can form for specific dust properties in the initial stage of dust sublimation when the dust size is reduced and the radiation pressure force increases with reduced dust size (Mukai and Yamamoto, 1979).The solar eclipse observations made over the years suggest however that the observed features can be explained without the existence of pronounced rings (Mann 1992) and that the average dust properties in the inner heliosphere change over time scales of years (Kimura et al., 1997;Ohgaito et al., 2002).The spatial distribution of the dust that can be derived from the scattered light and thermal emission observations suggests that the dust number density increases with distance from the Sun and in combination with the increasing orbital velocities this leads to increasing dust flux inversely proportional to the distance, r from the Sun within or close to the ecliptic plane (Mann et al. 2004).The amount of dust above the solar poles and in orbits with high inclination is even less known (Mann et al. 2004).Recent white light observations from STEREO A (Stenborg andHoward 2017a, 2017b) provide the shape of the F-corona and inner Zodiacal light from 5 to 24 degree line-of-sight elongation and show its flattening to larger elongation.Closer analysis also showed that the flattening varied with spacecraft position indicating an influence of the dust brightness near the spacecraft (Stenborg, et al.2018, Stauffer et al. 2018), again showing the importance of considering the line-of sight effects when analysing brightness observations.
While we can expect that the flux of large particles in the inner solar system increases proportional to 1/r, where r ist the distance from the Sun, it is difficult to predict the flux of smaller dust.Detailed trajectory calculations show that nanodust can be trapped in orbits with perihelia very close to the Sun instead of being ejected (Czechowski and Mann, 2010).Trapping conditions depend on a number of different parameters so that the nanodust flux outward can vary in time (Czechowski andMann, 2010, 2012).The majority of nanoparticles and those that form at distances 0.2 AU from the Sun or larger are usually ejected outward.Figure 10 shows the velocity of nanoparticles that are ejected from circular orbit at 0.2 AU.While the particles gradually gain speed, one can see that in the inner solar system they still have a velocity close to that of the parent object.The trajectory of the largest particle shown in the figure with approximate radius 100 nm corresponds to a beta-meteoroid that is mainly influenced by radiation pressure force (Wehry and Mann, 1999) because it has a smaller value of surface charge to mass than the smaller particles.Figure 11 shows the velocity as function of distance for the same range of parameters for the nanodust but when particles are released from initially highly elongated orbits.The orbital eccentricity and perihelion assumed for the model calculations presented correspond to the orbits of the Aquarids meteoroids and the nanoparticles are released at different locations of the orbit.One can see that their trajectories in the inner solar system strongly depend on the initial condition.
In addition to uncertainty of the dust trajectories, the dust production rate is hard to predict.The majority of dust smaller micron-size inside 1 AU are produced by fragmentation during collisions of larger dust.The dust formation by mutual collisions depends on the dust material compositions (Ishimoto and Mann, 1999;Mann and Czechowski, 2005), it varies with the dust velocities and it is for instance enhanced when coronal mass ejections push out nanoparticles (Czechowski and Kleimann 2017).Sun grazing comets (cf.Jones et al. 2018) are another source of time variable dust flux.The nanodust flux can vary also due to other effects, like the variation of the source, e.g.
when the dust flux is enhanced by a single collision event in the inner heliosphere, or due to the influence of the solar magnetic field structure (see Czechowski andMann 2012, Juhasz andHoranyi, 2013).While estimates are made for time-stationary conditions, current sheet crossings occur along the trajectories and in addition, the magnetic field is time-variable.Also coronal mass ejections change the conditions pushing outward large fractions of nanodust and with speed reaching 1000km/s (Czechowski and Kleimann 2017).
From ten years of STEREO A observations attributed to nanodust impacts several important properties can be obtained.The signal explained as nanodust is 10-100 times more frequent during Stream Interaction Region (SIR) or Interplanetary Coronal Mass Ejections (ICMEs).The observed signals exhibited a periodicity due to the crossing by STEREO of the solar magnetic equator.A correlation with solar wind perturbations, and periodicities corresponding to those of Mercury and Venus were also detected (Le Chat et al. 2015).These signals nearly disappeared on STEREO A around 2012 (Le Chat et al. 2013, Malaspina et al. 2015) when the heliosphere entered a defocusing configuration in which the nanodust coming from the inner heliosphere are pushed away from the solar magnetic equator, therefore possibly preventing their observation.Observations with the RPWS instrument on board Cassini between 1 and 5 AU have produced two further important properties of interplanetary nanodust.
Firstly, the average nanodust flux measured at 1 AU was similar in order of magnitude to the average of the highly variable flux measured by STEREO when the heliosphere was in a focusing configuration (Schippers et al., 2014), and decreased roughly as the inverse squared heliocentric distance (Schippers et al. 2015).Secondly, the nanodust fluxes were found to follow the variation in solar wind drift speed closely (Meyer-Vernet et al. 2017) as predicted by nanodust dynamics (Mann and Czechowski 2012).
Finally, aside from providing information on dust in the inner heliosphere, it is also quite possible that the Parker Probe and Solar Orbiter missions will find more effects that dust particles have on the solar wind.Cosmic dust particles interact with the surrounding plasma through electric charge collection, the photoelectric effect (Mann et al. 2014), and destruction processes (sputtering, fragmentation, sublimation).Photoionisation, electron-impact ionisation, and charge exchange quickly ionize the atoms and molecules in the solar wind (Mann and Czechowski, 2005) so that dust destruction generates pick-up ions.While typically near 1 AU, those interactions little affect solar wind measurable parameters (Mann et al., 2010), conditions are possibly different near the Sun.Dust particles sublimate at bulk temperatures ≈ 1000-2000 K inside ≈ 10 solar radii (Mukai and Mukai, 1973;Mann et al., 2004;Mann and Murad, 2005).A fraction of dust material vaporizes during collision (Mann and Czechowski, 2005).The effect of dust on the solar wind is also time-variable, as for instance the dust destruction rates due to sputtering increase during coronal mass ejections (Ragot and Kahler, 2003).The solar wind particles also change charge state by interaction with the dust surface or passing through the particles (Mann et al., 2010;Minato et al., 2004).And some authors suggest that newly formed charged dust fragments generate features in the solar wind magnetic field (Connors et al., 2014;Lai et al., 2013Lai et al., , 2015)).
7 Discussion of Implications for Observations with Parker Probe and Solar Orbiter The design of the radio and plasma waves instrument (RPW) on the ESA Solar Orbiter (Mueller and al., 2019) is similar to STEREO/WAVES instrumentation.The electric Antenna system (ANT) on RPW consists of a set of three identical antennas deployed from +Z axis and from the opposite corners of the spacecraft and can operate in dipole and monopole modes.RPW antennas consist each of a 1 metre rigid deployable boom and a 6. and a Radio Frequency Spectrometer (RFS) for signal processing and digitization.The DFB and TDS make rapid samples of waveforms with a highest sampling rate of 150,000 samples per second (DFB) and 2,000,000 samples per second (TDS), with an on-board selection of events to reduce bit-rate.The low frequency (LF) part of the RFS is a dual channel digital spectrometer receiving inputs from the four first antennas, either in dipole or monopole mode, with a frequency range of 10 kHz to 2.4 MHz, allowing a relative frequency spacing of about 4.5 %.
Parker Solar Probe orbit the Sun in the ecliptic plane, making seven Venus gravity assist manoeuvres during the seven-year nominal mission duration, which will lessen its perihelia to less than 10 RS, the closest any spacecraft has come to the Sun.In this way, the spacecraft will spend a total of 937 hours inside 20 RS, 440 hours inside 15 RS, and 14 hours inside 10 RS (Fox et al. 2015).The surrounding plasma changes considerably along the spacecraft orbits (Bale et al., 2016).The orbital trajectory for the first orbit around the Sun is shown in Fig. 12.
Our considerations suggest that both RPW and FIELDS measurements in monopole mode will be able to detect signals generated by dust impacts.Distinction between dust and other wave features needs to be considered based on the observational data.At present, we do not know the mass range of dust particles that will be detectable.As heliocentric distance decreases, the pulse's decay time will decrease faster than its rise time (Meyer-Vernet et al. 2017), eventually becoming smaller than the rise time, which will decrease the dust signal for large grains.The spacecraft charging and charged particles dynamics close to the Sun are expected to be considerably complicated by presence of a potential barrier (sheath structure) due to strong photoemission (Ergun et al. 2010, Campanell M. D., 2013) as well as by presence of the thermal shield and non-conducting solar panels.
Vaverka et al, (2017b) simulated pulses generated by dust impacts in various plasma environments using a simple numerical model.The spacecraft potential is calculated using orbital-motion-limited theory and the current generated by the dust impact is represented by Gaussian function.We used this approach to simulate the spacecraft charging in the inner solar system.The rise time of the pulse was estimated according to Meyer-Vernet et al. (2017) for variable photoelectron sheath.The rise time of the measured pulse is also affected by response of the antenna electronics.These electronic effects were not taken into account in the model.This model also does not describe the detailed structure of the pulses including "pre-spikes" but only their general shapes, and can be appropriate to estimate the conditions for dust impacts.Figure 13 shows estimated signals for impacts of 0.1 micrometre particles with speeds of 100 km/s for spacecraft at different distances from the Sun.The top panel represents temporal evolution of the spacecraft potential and bottom panel shows changes in the equilibrium potential.The charge production of the impact is assumed Q = 30 pC according the equation presented in chapter 2. The amplitude of the pulse is then proportional to the mass of the impinging grain and to its velocity with power between 2.5 and 6.2 (depends on materials).It is possible to see that the amplitude and duration of the pulses are reduced with decreasing distance from the sun.This fact means that the sensitivity of dust impact detection is smaller close to the sun.It is necessary to mention that the conditions for the orbital-motion-limited theory are not satisfied near the sun.The presence of the potential barrier created due to strong photoemission, described by Ergun et al. (2010) and Campanell (2013), strongly influence the spacecraft charging and charge dynamics.An interesting result is that the shape of the detected signal depends only weakly on the solar UV illumination, which leads to the photocurrent as shown in Fig. 14 for 1 AU from the sun.The increase or decrease of the solar activity which varies the UV flux, influence the spacecraft potential but not so much the profile of the pulse generated by dust impacts.1.
asymmetrically.The described model shows a strong dependence on the spacecraft potential.This can be compared with laboratory experiments for various polarity and sizes of bias voltage.A series of such measurement campaigns have been performed at the dust accelerator facility at the University of Colorado in order to aid the interpretation of signals collected in space.Collette et al. (2015) successfully identified different mechanisms of voltage signals generation on the antennas.The experiments performed by Nouzák et al. (2018) have used a scale model of the Cassini spacecraft to investigate the differences between antennas operated in monopole vs. dipole modes.The results show that in the dipole mode the antennas are greatly insensitive to dust impacts occurring on the spacecraft and only impacts on the antennas generate clear signals.This study helped clarifying the appropriate cross section
5. 4
MavenMAVEN is a NASA mission to Mars.It launched on November 18, 2013 and arrived at Mars on September 22, 2014.MAVEN is designed to study the escape of Mars's atmosphere, including the contribution of plasma processes associated with the interaction between the solar wind and the planet(Jakosky et al. 2015).Voltage spikes consistent with the impact of micron dust on the spacecraft were detected by the MAVEN LPW (Langmuir Probe and Waves) experiment at orbital altitudes between 200 km and 1500 km(Andersson et al. 2015a).Andrews et al. 2015 found large variations in plasma density and spacecraft surface charging encountered by MAVEN as it dipped into the Martian ionosphere.This resulted in strong variation in the detectability of dust impact voltage spikes.Once these effects were taken into consideration, the estimated near-Mars micron dust flux observed by MAVEN was found to be consistent with the interplanetary dust flux expected at Mars(Andersson et al. 2015b).
Earth.From 1994 to 2004, Wind executed a series of high apogee (100 Re) orbits about Earth and several lunar flybys before being stationed in an orbit about the first Lagrange point (L1) ~250 Re Sunward of Earth, where it remains operational to the present day (2019).The Wind WAVES experiment(Bougeret et al. 1995) detects voltage spikes consistent with the impact of micron-sized dust on the spacecraft body(Malaspina et al. 2014).These dust spikes are observable even though Wind WAVES makes only dipole electric field measurements, likely due to strong asymmetries of the dust impact signal on oppositely mounted antennas.Further, the rapid spin of the Wind spacecraft (one rotation every 3s) and asymmetry of dust impact voltage signals on the electric field wire antennas allows a crude directionality of the dust to be determined(Malaspina et al. 2014, Malaspina andWilson 2016).The observed amplitude and polarity of such signals are consistent with voltage induced on the antennas by positive ions produced by impacts on the spacecraft, after it has recollected the electrons(Meyer-Vernet et al. 2014); this new mechanism explained the previously unexplained voltage sign and amplitude for interstellar dust impacts on Wind, and also the absence of nanodust detection on this spacecraft.The yearly modulation of Wind-observed impacts was found to be consistent with the yearly variation in interplanetary micron dust(Malaspina et al. 2014, Wood et al. 2015).Further supporting this conclusion was the observation that both Wind and STEREO observe the same yearly modulation of interstellar dust flux(Kellogg et al. 2016).The long duration of the Wind mission (> 25 years, over two full solar cycles) presents a unique opportunity to study how the solar magnetic field modulates the entry of interstellar dust into the solar system and its arrival at 1 AU.To facilitate such studies, a database cataloguing all dust impacts observed by Wind was created(Malaspina and Wilson 2016) and made publicly available through the NASA Space Physics Data Facility Coordinated Data Analysis Web(CDAWeb) , both sensitive to dust impacts.The south-polar plume of Enceladus was one of the top discoveries made Cassini mission.During the Enceladus plume crossing, besides dust impact signals, RPWS detected plasma oscillations induced by dust impacts, the frequencies of which are equal to the local plasma frequencies (Ye et al. 2014a), which can be explained by a beam-plasma instability induced by the impact-produced electrons when their speed exceeds the thermal speed of the ambient plasma (Meyer-Vernet et al. 2017).Comparison of observations (Ye et al., 2014b), showed that the dust density profile measured by RPWS is consistent with that measured by the dedicated dust detector on board.Cassini allowed for a comparison of measurements in dipole and monopole configuration.The difference is clearly seen on Fig. 7 which shows the electric power spectrum measured by the Cassini Radio and Plasma Wave Science (RPWS) HFR receiver simultaneously in dipole (top) and monopole (bottom) mode in Saturn's E-ring at the first close approach of Enceladus (Meyer-Vernet et al. 2014).During the subsequent mission, the Wideband receiver 5 meters stacer deployable monopole, which has a 1.5 cm radius.The Time Domain Sampler (TDS) subsystem of the RPW instrument (Maksimovic et al., 2019) is designed to capture electromagnetic waveform snapshots at high cadence from 200 Hz to 200 kHz, resolving in particular voltage spikes associated with interplanetary dust impacts.Solar Orbiter will make observations of the Sun and in-situ measurements from elliptic orbits coming as close as ~60 solar radii (~0.285AU) to the Sun.The aphelia lie outside 0.8 AU for large part of the 7 years nominal mission time during which orbital latitude reaches 25 degree.The long cruise phase of Solar Orbiter and the elongated spacecraft orbits with aphelia close to 1 AU provide the opportunity to study in detail the dust flux near 1AU and to estimate the flux of sub-μm dust onto Earth, its time variation and variation during part of a solar cycle.The FIELDS instrument on Parker Solar Probe (Bale et al. 2016) combines magnetic and electric field measurements into a single, coordinated experiment.Four electric field antennas (2 m long, 3.18 mm diameter Niobium C-103 thin-walled tubes) are mounted at the base of the heat shield, and deploy in full sunlight out of the spacecraft wake, whereas a fifth antenna is mounted on the magnetometer boom in the umbra of the spacecraft.The sensor electric field signals are transferred to a Digital Fields Board (DFB), a Time Domain Sampler (TDS)
Figure 1 :Figure 2 :
Figure 1: Sketch of different dust components and dust interactions in the vicinity from the Sun as given in an overview (adapted from Mann et al., 2014).Recent results are presented in section 6.875
Figure 3 :
Figure 3: This figure sketches the impact process for a spacecraft that is slightly negatively charged (left panel), zero biased (middle panel), and slightly positively charged (right panel).It is further described in the text and parameters given inTable 1.
Figure 4 :
Figure 4: Laboratory simulation of dust impacts on Cassini model showing the impact signal detected by the antenna (EU boom was bombarded) for different polarity and size of bias voltage.Different phases of dust impact signal are distinguished by colours (green -cloud generation, blue -electron escape, red -ion escape, orange -relaxation).The inserts show details of the pre-spikes (modified from Nouzák et al., 2018).The conditions in these laboratory measurements are comparable to a measurement in monopole 890
Figure 5 : 895 Figure 6 :
Figure 5: Dust impact signals recorded by the STEREO/WAVES TDS on STEREO A shown with black crosses show in comparison to fit with semi-empirical model shown with red solid lines (from Zaslavsky et al. 2015).895
Figure 7 :
Figure 7: From Meyer-Vernet et al. (2014): Time-frequency electric power spectral density measured by Cassini/ RPWS on 9 March 2005 in Saturn's E ring, in dipole (top) and monopole mode (bottom).The increase due to micron-sized dust impacts on the spacecraft only appears in monopole mode, whereas the dipole only measures the weaker plasma quasi-thermal and impact noise.
Figure 8 :
Figure 8: Adapted from Fig. 4 and 5 of Ye et al. 2016b.RPWS wave power spectrogram around a ring plane crossing on DOY 001, 2016.The top panel shows the positive (red) and negative (blue) impact rates.The middle panel shows the impact signal polarity ratios with the moving averages (teal).At ~10:30, the antenna used was switched from monopole to dipole, which was accompanied 910
Figure 9 :
Figure 9: Adapted from Fig. 4 of Ye et al. 2018a.Comparison of vertical dust density profiles of the Janus-Epimetheus ring measured by RPWS and CDA during the ring plane crossing on DOY 361, 2016.There is one order of magnitude difference between the two results, which is within the uncertainty limit estimated for the RPWS measurement (Ye et al. 2014).
Figure 10 : 920 Figure 11 :Figure 12 : 930 Figure 13 :
Figure 10: Velocity as function of distance from the Sun for particles with Q/m = 10 −4 , 10 −5 , 10 −6 , and 10 −7 e/mp released from a circular orbit with the radius 0.2 AU near the ecliptic.Solid lines correspond to the focusing, and dashed to defocusing, magnetic field orientation (adapted from Mann et al., 2014).
Figure 14 :
Figure 14: The spacecraft potential change during dust impact for different values of the photocurrent.A typical value of 935 STEREO is a NASA mission , Belheouane et al. 2012ober 26, 2006, with the study of coronal mass ejections as primary science goal.The mission consists of two twin spacecraft that orbit the Sun at around 1 AU, one trailing the Earth (STEREO B) while the other leads (STEREO A).The study of the STEREO/WAVES radio receiver data proved to be of great interest for dust studies.STEREO/WAVES measured the flux of submicrometre dust near 1 AU(Meyer-Vernet et al. 2009, Belheouane et al. 2012, Zaslavsky et al. 2012)and discovered a highly time-variable flux of dust with size few nm(Meyer-Vernet et al., 2009).The nanodust impacts were observed frequently on both STEREO spacecraft as radio pulses on single monopole antennas.The physical mechanism that leads to their generation is not yet fully understood.The voltage was much higher on the antenna that was adequately located to be sensitive to impacts of prograde nanodust on each spacecraft(Meyer-Vernet et al. 2009), which destabilized the photoelectron sheath of that antenna | 11,510.8 | 2019-07-15T00:00:00.000 | [
"Physics"
] |
Electron-positron vacuum instability in strong electric fields. Relativistic semiclassical approach
Instability of electron-positron vacuum in strong electric fields is studied. First, falling to the Coulomb center is discussed at $Z>137/2$ for a spinless boson and at $Z>137$ for electron. Then, focus is concentrated on description of deep electron levels and spontaneous positron production in the field of a finite-size nucleus with the charge $Z>Z_{\rm cr}\simeq 170$. Next, these effects are studied in application to the low-energy heavy-ion collisions. Then, we consider phenomenon of"electron condensation"on levels of upper continuum crossed the boundary of the lower continuum $\epsilon =-m$ in the field of a supercharged nucleus with $Z\gg Z_{\rm cr}$. Finally, attention is focused on many-particle problems of polarization of the QED vacuum and electron condensation at ultra-short distances from a source of charge. We argue for a principal difference of cases, when the size of the source is larger than the pole size $r_{\rm pole}$, at which the dielectric permittivity of the vacuum reaches zero, and smaller $r_{\rm pole}$. Some arguments are presented in favor of the logical consistency of QED. All problems are considered within the same relativistic semiclassical approach.
Introduction
I dedicate this review to the blessed memory of Vladimir Stepanovich Popov, who recently left us as the result of a many-year hard illness, which prevented him working actively in his last years. The problem of the electron-positron pair production when the ground-state electron level dives below the energy −mc 2 (m is the electron mass, c is the speed of light) was of his interest starting from the end of 1960-th. Especially he contributed to this problem during the 1970s. V. S. Popov was awarded the I. Y. Pomeranchuk Prize in 2019 for his outstanding contributions to the theory of ionization of atoms and ions in the field of intense laser radiation and the theory of the creation of electron-positron pairs in the presence of superstrong external fields.
We worked together with Vladimir Stepanovich on problems of supercritical atoms with the charge Z > Z cr = 170 − 173 during 1976-1978 when we developed semiclassical treatment of this problem. These works, cf. [1][2][3][4][5][6][7] became a part of my PhD thesis [8] that was defended in 1977 under the guidance of Arkadi Benediktovich Migdal. As follows from the Dirac equation in the Coulomb field of a point-like nucleus with Z > 1/e 2 (in unitsh = c = 1, which will be used in this paper, e 2 1/137), the electron that occupied the ground-state level should fall to the center. Following the idea of I. Pomeranchuk and Ya. Smorodinsky [9], the solution of the problem of the falling of the electron to the center can be found while taking into account the fact that the real nuclei have a finite radius. With increasing Z, the energy of the ground state level decreases and, at Z > Z cr , crosses the boundary of the lower continuum = −m. The problem received a new push in the end of the 1960sThe important role of the Pauli principle was emphasized in [10]. However the authors erroneously assumed delocalization of the electron state with −m. Independently, W. Pieper and W. Greiner [11] (in numerical analysis) and V. S. Popov [12][13][14][15][16] (in analytical and numerical studies) correctly evaluated the value of the critical charge to be Z cr 169 − 173, depending on assumptions regarding the charge distribution inside the nucleus and the ratio Z/A. It was argued that two positrons with the energies > m go off to infinity and electrons with < −m screen the field of the nucleus by the charge −2e. The typical distance characterizing electrons of the vacuum K shell is ∼ 1/(3m) R nucl , cf. [7]. Subsequently, there appeared an idea to observe positron production in heavy-ion collisions, where the supercritical atom is formed for a short time [17,18]. As the reviews of these problems, I can recommend [19][20][21].
In 1976, with the inauguration of the UNI-LAC accelerator in GSI, Darmstadt, it became possible to accelerate heavy ions up to uranium below and above the Coulomb barrier. Instead of a positron line that is associated with the spontaneous decay of the electron-positron vacuum, mysterious line structures were observed, which, in spite of many attempts, did not get a reasonable theoretical interpretation. The experimental results on the mentioned positron lines proved to be erroneous. New experiments were conducted during 1993-1995, cf. [22][23][24]. The presence of the line structures was not observed. Events, which could be interpreted as the effect of the decay of the QED vacuum with the spontaneous production of the electron-positron pair, were not selected. In spite of the effect of the spontaneous production of positrons in the electric field of the supercharged nucleus being predicted many decades ago, it has not yet been observed experimentally in heavy-ion collisions.
One also studied a possibility of a nuclear sticking in the process of the heavy-ion collisions [25,26]. Although these expectations did not find a support in further investigations, extra arguments were given for a possibility of the observation of the spontaneous positron production in the heavy-ion collisions, cf. [27]. Especially, the usage of transuranium ions looks very promising [28]. Besides a spontaneous production of positrons, a more intensive induced production of pairs occurs due to an excitation of nuclear levels, cf. [20]. Therefore, the key question is how to distinguish spontaneous production of positrons that originated in the decay of the electron-positron vacuum from the induced production and other competing processes.
New studies of low-energy heavy-ion collisions at the supercritical regime are anticipated at the upcoming accelerator facilities in Germany, Russia, and China [29][30][31]. This possibility renewed theoretical interest to the problem [27,[32][33][34]. As one can see from the numerical results reported in [34], these results support those that were obtained in earlier works, although a comparison with the analytical results derived in [1][2][3][4][5][6][7] was not performed. Additionally, it should be noted that there recently appeared statements that the spontaneous production of positrons should not occur in the problem under consideration. I see no serious grounds for these revisions and, thereby, will not review these works.
A General Picture
States with | | < m correspond to the energy E = ( 2 − m 2 )/2m < 0 and effective potential U, see Figure 1. In terms of the Schrödinger equation these are ordinary bound states. Let the ground state level be empty and we are able to adiabatically increase the charge of the nucleus Z. The latter means that the time τ Z characterizing the increase of Z is much larger when compared to 1/| 0 − njm |, where njm are the energies of other bound states in the potential well, and τ Z > 1/m for the case of transitions from the ground-state level, 0 , to the continues spectrum. The empty level with < −m becomes quasistationary, see Figure 1. When penetrating the barrier between continua, see Figure 2 below, two electrons (with opposite spins) are produced, which occupy this level, whereas two positrons of the opposite energy go off through the barrier to infinity. In the standard interpretation, cf. [16], the electron states, ψ ∝ e −i t , with = 0 + iΓ( 0 )/2 for 0 < −m, Γ > 0, cf. Equations (3.5) and (3.6) in [35], are occupied due to the redistribution of the charge of the vacuum. The vacuum gets the charge 2e < 0 distributed in the region of the supercritical ion. Two positrons with e + = − 0 − iΓ( 0 )/2 go off to infinity after passage of a time ∼ τ 0 e Γt , τ 0 ∼ R, where R is the size of the potential well for R > ∼ 1/m, as it occurs for any decaying quasistationary state, producing a diverging spherical wave ψ ∝ e ikr , k = 2 e + − m 2 for the positron. For far-distant potentials, the situation is similar to that for the charged bosons, cf. [36]. For the case V = −Ze 2 /r for r > R nucl , one obtains Γ(−m) = 0. Typical dependence of effective Schrödinger potential U on r for a charged particle in an electric central-symmetric potential well, r ± are turning points, and r 0 corresponds to maximum of the effective potential U. The dashed line describes the quasistationary level with < −m.
For Z < Z cr electrons of the lower continuum (with < −m), fill all energy levels according to the Dirac picture of the electron-positron vacuum. They are spatially distributed at large distances. For Z > Z cr the process of the tunneling of the electron of the lower continuum to the empty (localized) state that was prepared in the upper continuum with < −m can be treated as the tunneling of the virtual positron (electron hole) with e + = − 0 − iΓ/2 from the region of the potential well to infinity, where it already can be observed. If one scatters an external real positron with a resonance energy e + − 0 > m on such a potential, this positron, for a short time, forms a resonance quasistationary state in the effective potential, which, after passage of a time ∼ 1/Γ, is decayed. As the result, the positron goes back to infinity. After that, during a time of the same order of magnitude, two positrons, being produced in a fluctuation together with two electrons, go off to infinity and those two electrons fill the stationary negative-energy state, as was explained.
If the ground state level was initially occupied by two electrons of opposite spins, then, at adiabatic change of the potential (in the sense clarified above), they remain on this level = 0 . At the adiabatic change of the potential, electrons have no energy to escape anywhere from this level. The production of pairs does not occur, since the level is occupied by electrons. During a time ∼ 1/Γ, their charge 2e < 0 is redistributed over the range of energies 0 − Γ( 0 )/2 < ∼ < ∼ 0 + Γ( 0 )/2. This charge is localized at distances (∼ 1/(3m) that are typical for the ground state in the Coulomb field [7]). In this sense, one formally requires a many-particle description of the stationary electron with Re 0 < −m at Γ = 0. However, neglecting a tiny Γ correction, for the finding of (Z), one may continue to employ the one-particle description. If the experimenter scatters an external positron with e + − 0 > m on such a potential, the positron annihilates with one of the two electrons have occupied the ground-state level. After the passage of a time ∼ 1/Γ, there occurs spontaneous production of the one new pair, the electron fills empty state (after that, again, two electrons occupy the ground-state level) and the positron goes to infinity.
Semiclassical Approximation
Semiclassical approximation is one of the most important approximate methods of quantum mechanics [37]. Classical and semiclassical ideas are widely used in quantum field theory in problems dealing with the spontaneous vacuum symmetry breaking for bosons, cf. [21,36,38], in condensed matter physics, cf. [39][40][41], and in physics of nuclear matter [42,43].
As a consequence of the instability of the boson vacuum in a strong external field, there appears a reconstruction of the ground state and there arises a condensate of the classical boson field [44,45]. Many-particle repulsion of particles in the condensate provides the stability of the ground state. After that, excitations prove to be stable, cf. [42,43]. They are also successfully described using semiclassical methods, e.g., such as the loop expansion [36,46].
For fermions, there exist two possibilities. In the first situation, fermions heaving attractive interaction, being rather close to each other, may form Cooper pairs, cf. [40]. In the second situation, which we focus on here, electron-positron pairs, being produced in a strong static electric field, are well separated from each other by the potential barrier. Consequently, the electric potential attracts particles of one sign of the charge and repels antiparticles. Because of the Pauli principle, each unstable single-particle state is occupied by only one fermion. Therefore, it is natural to prolong a single-particle description in a overcritical region (until there appeared still not too many dangerous states). Classical approximation does not work for fermions, but semiclassical methods prove to be working. As is known, the semiclassical approach yields correct results for the values of the energy levels with big quantum numbers and in the case of spatially smooth potentials, when dλ/dx 1, whereλ = 1/p(x) is the reduced electron De Broglie length, p(x) is the momentum, and x is the coordinate. For the Coulomb field for the ground-state level, a rough estimate yields dλ/dr ∼ 1/(Ze 2 ) for r → 0. However, even for dλ/dx ∼ 1, semiclassical approximation continues to work not bad in calculation of the energy levels, with an error not larger than 10% due to the presence of a numerically small parameter ∼ 1/π 2 , cf. [37].
Instability of the vacuum near a nucleus heaving a supercritical charge. It proves to be that the semiclassical approximation is applicable with an appropriate accuracy for the description of the electron energy levels in the supercritical field of a nucleus with the supercritical charge Z > (170 − 173). Semiclassical approximation allows for finding rather simple expressions for the critical value of the charge, cf. Refs. [8,47,48], for energies of deep levels as a function of Z and for the probabilities of the penetration of the barrier between continua, cf. [3][4][5][6][7].
The spontaneous positron production in low-energy heavy-ion collisions. A comparison of the theory and experiment should check the application of QED in the region of strong fields outside the applicability of the perturbation theory. The description of the spontaneous production of positrons in heavy-ion collisions needs a solution of the two-center problem for the Dirac equation. Because variables are not separated in this case, the problem does not allow for the analytical treatment and numerical calculations are cumbersome. However, the use of the semiclassical approximation results in simple analytical expressions for the energies of the electron levels, cf. [6,7], valid with error less than few %. Thereby, this is one more example of the efficiency of the semiclassical approach.
Electron condensation in a field of a supercharged nucleus. In supercritical fields, many energy levels cross the boundary of the lower continuum and the problem of the finding of the vacuum charge density becomes of purely many-particle origin. It can be considered within the relativistic Thomas-Fermi method, cf. [2]. All of the initially empty states, which crossed the boundary = −m, are filled after a while. In this sense, one may speak about "electron condensate".
Vacuum polarization and electron condensation at super-short distances from Coulomb center. In spite of the successes in explanation of all purely electrodynamical phenomena, QED is a principally unsatisfactory theory, since relations between the bare mass and charge and observable ones contain divergent integrals [49,50]. As the result, as one thinks, there is no not contradictive manner to pass from super-short to long distances. In spite of this, as is well known, it is possible to remove divergencies from all observable quantities with the help of the renormalization procedure.
The problem of the so-called "zero charge" or Moscow zero, cf. [51,52], is one of central problems related to renormalization of the charge. When considering the square of the charge of electron e 2 (r) as a function of the radius r and assuming finite value of the bare charge e 2 (r 0 ) = e 2 0 > 0 for the source-size r 0 → 0, one derives e 2 (r → ∞) → 0 instead of an expected value e 2 (r → ∞) → e 2 = 1/137. The same problem appears, when one considers the screening of the central source with the charge density n ext = Z 0 δ(r − r 0 ) for r 0 → 0, cf. [3]. The problem of a distribution of the charge near an external source of the charge with the radius R 1/m, as well as the problem of the distribution of the charge of the electron at distances r 1/m are the key principal problems of QED. The semiclassical approach proves to be very promising in the calculation of the vacuum dielectric permittivity in strong inhomogeneous electric fields [53]. The density of the polarized charge is supplemented by the density from the electron condensation [3,42]. The problem proves to be specific and it depends on whether the radius of the external source of the charge is larger than a distance r pole , where the dielectric permittivity decreases to zero, or smaller r pole , cf. [54,55]. References [54,55] argued for the condensation of electron states in the upper continuum at distances larger than r pole for r 0 > r pole and for the condensation of electron states originated in the lower continuum at distances smaller than r pole (for r 0 < r pole ), at which the dielectric permittivity proves to be negative and e 2 0 < 0. The semiclassical consideration of this problem allows for presenting arguments in favor of a logical consistency of QED. Similar effects in semimetals and in stack of graphene layers. The existence of the Weyl semimetals, i.e., materials with the points in Brillouin zone, where the completely filled valence and completely empty conduction bands meet with a linear dispersion law, = v F p, where the Fermi velocity is v F ∼ 10 −2 , has been predicted in [56]. Systems with the relativistic dispersion law are likely to be realized in some doped silver chalcogenides, pyrochlore iridates, and in topological insulator multilayer structures. Weyl semimetals are three-dimensional analogs of graphene [57], where the energy of excitations is also approximately presented by the linear function of the momentum, but the electron subsystem is a two-dimensional one, whereas the photon subsystem remains three-dimensional. Even though the mass of excitations m = 0 for ideal graphene and Weyl semimetals without interactions, a non-zero mass, m = 0, can be induced in many ways [58], resulting in a dispersion relation characterized by a gap, i.e. 2 = p 2 v 2 F + m 2 v 4 F . In difference with a small value of the fine structure constant in QED, e 2 = 1/137, the effective coupling in Weyl semimetals and in graphene is α ef = e 2 /v F ε 0 , where ε 0 is the dielectric permittivity of the substance. The coupling constant α ef can be as 1 as > ∼ 1, depending on the substance, and both weak and strong coupling regimes are experimentally accessible. Thus, Weyl semimetals and an infinite stack of graphene layers make it possible to experimentally study various effects have been considered in 3+1 quantum electrodynamics (QED) for weak and effectively strong couplings, cf. [59,60].
Not concerning spontaneous production of positrons of our interest here, the electronpositron production in heavy-ion collisions was studied in many papers, cf. [61][62][63][64].
Electric fields with the strength E m 2 may exist in astrophysical environments, e.g., they may occur at phase transitions in neutron and hybrid stars [43,74] and in neutron star mergers [75], and they also exist at surfaces of hypothetical nuclearites and abnormal superheavy nuclei [43,53,[76][77][78].
Various radiative corrections to the deeply bound electron levels should certainly be taken into account, e.g., cf. [79][80][81] and the references therein. These higher-order corrections will not be considered in the given paper.
Below, attention is focused on a semiclassical description. I describe the instabilities of the boson and fermion vacua in static potentials, in particular in the Coulomb field. Afterwards, focus is concentrated on the description of the spontaneous positron produc-tion in low-energy heavy-ion collisions. Next, a many-particle semiclassical description of the electron condensation is considered. Finally, modification of the Coulomb field at super-short distances due to the vacuum polarization and electron condensation is studied.
The paper is organized, as follows. Section 2 starts with a brief discussion of instability for the charged bosons in static electric fields, in particular in the Coulomb field of a point-like nucleus with the charge Z > Z cr = 1/(2e 2 ). The behavior of deeply bound electrons obeying the Dirac equation in the strong static electric fields is considered in Section 3. First, I consider the case of a one-dimensional field and then of a spherically symmetric field. The Dirac equation is transformed to equivalent Schrödinger form in an effective potential and the interpretation of the solutions is discussed. Subsequently, in Section 3.5, I demonstrate exact solution of the problem of bound states in the strong Coulomb field of a point-like center. The focus is made on the problem of the falling of the electron to the center for a nucleus with the charge Z ≥ 1/e 2 . Section 3.6 describes how the problem is resolved while taking into account that nuclei have a finite size. In Section 4, I introduce a semiclassical approach to the Dirac equation, being transformed to the second-order differential equation. Electron levels crossed the boundary of the lower continuum are considered. The mean radius of the K-electron shell and the critical charge of the nucleus are found for = −m, as well as the number of levels that crossed the boundary of the lower continuum and their energies. The critical charge of the nucleus for the muon is also found. A comparison of semiclassical expressions with much more cumbersome exact expressions permits understanding the merits of the semiclassical approach. In Section 5, a semiclassical approximation is developed for the system of linear Dirac equations. Semiclassical wave functions in classically allowed and forbidden regions are introduced, and the Bohr-Sommerfeld quantization rule is formulated. Next, the probability of the positron production is calculated. Subsequently, semiclassical approximation is applied to non-central potentials. In Section 6, focus is concentrated on problems of the spontaneous positron production in low-energy collisions of heavy ions. The energies of deep levels as a function of the distance between colliding nuclei and the angular distribution of the positron production are found while employing semiclassical approach. Subsequently, I consider a screening of the charge at collisions of not fully striped nuclei. Semiclassical approximation (imaginary time method) is adequate for describing dynamics of the tunneling of electrons from the lower continuum to the upper one. In such a way, a correction on non-adiabaticity to the probability of the production of positrons is found. The electron condensation in the field of a supercharged nucleus is considered in Section 7. Section 8 presents the effects that are associated with the polarization of the electron-positron vacuum in weak and strong fields. Subsequently, in Section 9, I focus on the description of the charge distribution at super-short distances from the charge source. The effects of polarization of the vacuum and the electron condensation in the upper and lower continua will be considered. Section 10 contains a conclusion.
. Reduction of Klein-Gordon-Fock Equation to Schrödinger Equation
Consider a spinless negatively charged boson placed in a stationary attractive potential well V. The Klein-Gordon-Fock equation renders With the help of notations we may rewrite Equation (1) in the form of the Schrödinger equation, As we see from Equation (2), for relativistic particles there appears to be an attractive term in the effective potential −V 2 /(2mc 2 ), even for a purely repulsive potential V. In the limit case E m and |V| m, we have m + E and U ef V, and we recover the Schrödinger equation for a nonrelativistic particle. For | | < m the "nonrelativistic" energy is E < 0, which corresponds to bound states in the interval of energies −m < < m. For a sufficiently deep potential well, the energy of the ground state level may cross the boundary = −m. In a deeper potential, other levels cross this boundary. For < −m, here ReE > 0, the levels become quasistationary, see Figure 1.
A comment is in order (D. N. Voskresensky 1974, see comment in [82]). For a spinless particle under consideration, the ground-state single-particle level only crosses the boundary = −m for far-distant potentials, when −V(r → ∞) > C cr /r 2 , for a constant C cr > 0. For potentials obeying condition −V(r → ∞) < C cr /r 2 , there appears to be a bound state for the antiparticle. In both cases for a broad potential well of a typical radius R 1/m the vacuum instability occurs at |V| |V| cr 2m( ). In the case of a broad potential well, solutions of many-particle problems in both cases are almost the same, cf. [36]. For −V > −V cr there appears production of pairs. Positively charged antiparticles go to infinity and negatively charged particles form a condensate, see [36,42].
Let us illustrate how the deformation of boundaries of upper and lower continua occurs in a static electric field forming a broad potential well for a negatively charged particle, cf. [2]. To be specific, consider a spherically symmetric field. Boundaries of continua, ± , are determined by They are shown in Figure 2. In upper and lower continua p 2 (r) > 0, these are classically allowed regions. In the gap between continua p 2 (r) < 0. This is a classically forbidden region. For V < V cr = −2m − O(1/(m 2 R 2 )), there arises a region of the overlapping of the continua that means that the negatively charged particle may penetrate from the lower continuum (from the exterior of the potential well) to the upper one (to the interior of the well).
With an exponential accuracy, the probability of a passage of the one-dimensional barrier is determined by where x 1 and x 2 are the turning points at which p(x) = 0. This expression is applicable for W 1. As example, consider a uniform static electric field eE = −∇V = const, |eE| m 2 . Then we have p ( + eEx) 2 − m 2 . From Equation (5), we immediately obtain This expression coincides with the first term of the infinite series solution [83]. A question arises as to whether it is possible to observe a process of the production of pairs already in a weak attractive electric field with the strength |E| m 2 at −δV > 2m? The critical difference −δV −2m can be easily reached in the field of the capacitor, where ∇A 0 = const, at the increase of the distance d between plates. Employing |∇A 0 | = | E| ∼ 10 4 V/cm, the value, which is easily produced in electrical engineering, we estimate |δV| > 2m π already for d > ∼ 10 3 cm. Here, m π 140 MeV is the mass of the lightest charged boson, the pion. However the probability of the production of the pairs W ∼ e −2ImS , ImS = x 2 x 1 |p|dx, is negligibly small at these conditions. Indeed, for V = −eEx, we get For pions E 0 10 21 V/cm. For electrons E 0 1.3 · 10 16 V/cm.
Relativistic Spinless Charged Particle in Coulomb Field of Point-Like Center
In the case of the Coulomb field of a point-like nucleus, V = −Ze 2 /r, with the help of the replacement φ( r) = R(r)Y lm , we obtain equation for the radial wave function R(r) in the form where E = 2 −m 2 2m is the effective nonrelativistic Schrödinger energy of the particle, is the effective potential, now, depending on l. Equation (7) and the ordinary Schrödinger equation for the radial function in the effective potential coincide after undertaking replacements l(l + 1) − (Ze 2 ) 2 = λ(λ + 1) , in the former one. Thus, instead of the expression for the energy of the Schrödinger particle in the Coulomb field, we derive E n r ,l = − (Z e 2 ) 2 m 2(n r + λ + 1) 2 .
For Z > Z cr = 1/(2e 2 ) the particle, being in the ground state (n = 1), falls down to the center. Let Ze 2 = 1/2 + δ for 0 < δ 1. Subsequenty, choosing positive-sign square root of solution (11) we have for Ze 2 = 1/2 + δ, m(1+iδ) √ 2 and the wave function is not normalized, reflecting the fact of the falling of the negatively charged particle to the Coulomb center with Z > 0 and the falling of the positively charged particle to the Coulomb center at Z < 0. We dropped the negative-root solution of Equation (11) as not physical one, since it arises at −m already for small Z > 0. However, note that the negative-root solution of Equation (11), − m(1+iδ) √ 2 , for the negatively charged particle near the Coulomb center for Z > Z cr = 1/(2e 2 ) yields φ ∝ e −mδt/ √ 2 , i.e., decreasing at t → ∞. This implies a possibility of a multi-particle interpretation of the < 0 solution for the negatively charged particle in the field Z > 0. We return to this question in Section 9.2.
The value Z cr = 68.5. It means that the Mendeleev table would be closed on element with Z cr = 68, if the nuclei were point-like. As we have mentioned, the lightest spinless meson is the pion. The radius of the real nucleus with atomic number A is found from the condition 4πρ 0 R 3 /3 = A, where ρ 0 0.16 fm −3 0.48m 3 π . For a symmetric nucleus A 2Z we estimate R > a π 1B = 1/(m π Ze 2 ) (radius of the ground-state orbit for the pion) already for Z > 40. Subsequently, the lowest pion orbit enters inside the nucleus and approximation of a point-like nucleus becomes invalid.
Note that, for Z = Z cr , part + a.part = m √ 2 > 0, and thereby pairs are not produced at such conditions. This peculiarity appears only for the case of the point-like Coulomb field. For a field, being cut at R = 0 (R 1/m π , such that V = −Ze 2 /r for r > R and V = −Ze 2 /R, the model I, or for V = − Ze 2 R ( 3 2 − r 2 R 2 ), the model II at r < R, the ground state particle level continues to decrease with increasing Z and decreasing R and for Z = Z cr (R) > Z cr , it reaches = −m. At Z = Z cr (R), the sum part + a.part is zero, corresponding to the spontaneous production of the pairs for Z ≥ Z cr (R), at R < R cr .
Sommerfeld formula for electron. Electron has spin 1/2. In the absence of the magnetic field spin and orbital spaces are orthogonal. Thus one may expect that expression (11) continues to hold also for electron after replacement l + 1/2 → | J| + 1/2 = |κ|, where κ = −1, 0, 1... is integer number, since axial vectors of angular momentum and spin are summed up, L → J = L + s. Subsequently, we have 2 n r ,κ = where n r = n − |κ| = 0, 1, ... is a radial quantum number. Now, falling to the center appears when the ground state level reaches the value = 0. It occurs for Z = Z cr = 1/e 2 = 137.
For a field cutted at R = 0, e.g., for the case V = −Ze 2 /r for r > R and V = −Ze 2 /R for r < R, the ground state level continues to decrease with increasing Z and for Z = Z cr (R) > Z cr it reaches = −m. After that, the sum part + a.part reaches zero, corresponding to the spontaneous production of the electron-positron pairs. Two electrons occupy the ground-state level and two positrons with − > m move to infinity. Note that the same expression (12) is derived from the exact solution of the Dirac equation in the Coulomb field, as we will see in Section 3.5.
Dirac Equation for Particle in Static Electric
We are now at the position to focus on the problem of our main interest in this paper, i.e., to describe the behavior of electrons in a strong static electric field.
Interaction with 4-vector field A µ = (A 0 , A) is constructed with the help of minimal coupling p µ = i∂ µ , γ µ are ordinary Dirac matrices.
Dirac System in Case of One-Dimensional Electric Field
In the case of a static one-dimensional electric field ( A = 0) using replacement we rewrite Equation (13) as We may rewrite Equation (15) as For further convenience, here we retained dependence onh.
Reduction of Dirac System to Schrödinger Equation
With the help of the replacement Equation (18) is reduced to the equation of the second-order in r-derivative, similar to the Schrödinger equation, where is the term appeared due to the spin. If U s were zero, after the replacement κ → l we would recover the Klein-Gordon-Fock equation for a spinless particle. At r → 0, for V = −Ze 2 /r, we have U s → − 1+4κ 8mr 2 . For 1 s level κ = −1, U s → 3 8mr 2 . In the latter case for r → 0. The falling to the center in such a Schrödinger potential occurs when U ef (r) < −1/(8mr 2 ), cf. [84], which corresponds to Ze 2 > 1.
Interpretation of Bound States in a Weak Field
The Dirac equation describes the electron and positron simultaneously. Therefore at appearance of the bound state in a potential well there arises a question regarding whether it relates to the electron or to the positron. As example, consider the case of a weak external static central-symmetric electric field produced by a static source of a positive charge distributed in a range r. Subsequently, V = −ζv(r) < 0 for the electron, where ζ > 0 is a parameter proportional to the depth of the potential well. As is known, for sufficiently small ζ, the Dirac equation, as the Klein-Gordon-Fock equation, can be transformed to the Schrödinger equation for a nonrelativistic particle. The bound state for the electron appears first at a certain value of ζ. At decreasing ζ, this state is diluted in the continues spectrum with ≥ m.
The system of Dirac Equation (18) is symmetric in respect to replacements → − , V → −V, κ → −κ, G → F. Equation describing energy levels does not depend on G and F. Thereby, it is symmetric, respectively, replacements → − , V → −V, κ → −κ. In the case of the source of a positive charge, the electron undergoes attraction. In the field of the opposite-sign charge (V → −V), the electron undergoes repulsion. Because, in the attractive field, there appears the electron energy level going from the upper continuum, in the repulsive field there appears the electron energy level originating from the lower continuum. However, because the Dirac equation simultaneously describes electron and positron, if the electron moves in a repulsive field, then the positron moves in an attractive one. Thereby, the electron level moving in a repulsive field from the lower continuum can be interpreted as the positron level ( → − , κ → −κ) going from the upper continuum (now in the field of attraction to the positron). It is natural to think that in a weak repulsive field for the electron for a small ζ < 0 a deeply bound level with −m should not exist. Because such a state nevertheless exists in the full set of solutions of the Dirac equation, after the replacement → − , κ → −κ, it should be interpreted as the positron state. This interpretation is confirmed experimentally. In the field of a proton, there are electron bound states lying near the boundary of the upper continuum but there are no positron states with −m. Vise versa, in the field of an antiproton, there exist positron levels with m, but there are no electron levels with −m. This picture is also established by the minimization of the energy in the mentioned cases. Namely, in the field of a positive charge, the presence of the bound electron is more energetically favorable when compared to the presence of the positron.
Statements done above seem obvious except the case, which I shall consider below in Section 9.2, when polarization of the vacuum may result in a negative dielectric permittivity and attraction is replaced by repulsion.
Exact Solution for Electron in Coulomb Field of Point-Like Center
Consider the discrete spectrum < m of the Dirac equation in the potential V = −Ze 2 /r. We search G and F in Equation (18) as This form of the solution, cf. [49], follows from asymptotic behavior of G, F ∼ r ±g at r → 0 and G, F ∼ e −r/2 at r → ∞. Solutions G, F ∼ C 1,2 r −g are dropped (i.e., we put C 1,2 = 0) due to the divergence of their contribution to the probability ( |ψ| 2 dr → ∞).
Setting (24) in Equation (18), we obtain a system of equations These equations are reduced tõ As is seen, Equation (27) are symmetric under simultaneous replacement → − and Ze 2 → −Ze 2 .
The finite solution forr → 0 gets the form where F(α, β, z) is the degenerate hypergeometric function. Settingr = 0 in one of Equation (26), we find relation Both of the hypergeometrical functions in (28) are reduced to polynomials, otherwise they would grow as er forr → ∞, which results in the divergence of the probability. From this requirement follows that α in F(α, β, z) equals a non-positive integer number, i.e., For n r = 0, only one of two functions is reduced to a polynomial. Subsequenty, g = Ze 2 √ m 2 − 2 and Ze 2 m √ m 2 − 2 = |κ|. If κ < 0, then B = 0 in Equation (29) and Q 2 = 0, and the required condition is fulfilled. If κ > 0, then B = −A and Q 2 is a divergent function at n r = 0. Thereby, permitted states are n r = 0, 1, ... for κ < 0 and n r = 1, 2, ... for κ > 0. From (30), it also follows the solution for the negatively charged particle with < 0 for Z < 0. In a single particle problem under consideration, one should drop such a solution, since it describes a strongly bound particle already in a weak field. However, such a solution can be appropriately treated within a many-particle picture with taking the vacuum polarization and the electron condensation that originated in the lower continuum into account, as we argue below in Section 9.2.
The ground state 1 s-level of the electron in the field of the positively charged Coulomb center (Z > 0) corresponds to κ = −1, n r = 0. Its energy is At Ze 2 ≥ 1, there occurs falling of the electron to the center. Indeed, for r → 0 following (24), (27) we get For Ze 2 = 1 + δ > 1, the value g = i √ 2δ becomes imaginary and solutions oscillate as that corresponds to not normalized probability ∞ 0 |ψ| 2 dr. At Ze 2 = 1 + δ, 0 < δ 1, solution of Equation (32) yields = +im √ 2δ and the electron wave function grows as Ψ ∝ e +m √ 2δt , indicating the falling of the electron to the center. The solution of opposite sign (see Equation (31)) arises from the lower continuum at V → 0. In the single-particle problem a negative-energy solution should be dropped. Note that at Ze 2 = 1 + δ, it yields = −im √ 2δ and Ψ → 0 at t → ∞ that may suggest an interpretation. However, an appropriate interpretation proves to be possible only beyond the single-particle problem, as will be shown in Section 9.2.
Solutions (31) and (32) hold formally for the positron in the Coulomb potential of the nucleus with the charge Z < 0. Within the single-particle problem under consideration, appropriate interpretation again exists for the solution, where energy originates from the upper continuum decreasing with increasing −Z, rather than the negative-energy solution, similarly to that happened for the electron at Z > 0.
For Z > 0, only two electrons (due to Pauli principle), if they have occupied the ground state, undergo falling to the Coulomb center for Ze 2 = 1. For levels with the quantum number n r > 0, we have n r ,κ > 0 for Z = 1/e 2 . Now, assume that the groundstate level was empty and we adiabatically increase Z. There is no appropriate solution of the single-particle problem for the point-like nucleus with Z > 1/e 2 in this case.
Avoiding problem of falling to the center. A reasonable interpretation may appear, only if one assumes that the nucleus has a size R = 0, and then we may safely decrease R. First assume that R r Λ = 1/m. In the limit Λ = ln(r Λ /R) 1 for the ground-state level of the electron, one gets [15,16] For ζ < 1, Λg 0 1, the value th(Λg 0 ) 1 − 2e −2Λg 0 rapidly tends to unity and Equation (35) coincides with (32). For R = 0, the point ζ = 1 is already not a singular point for the function 0 (ζ). Equation (35) is analytically continued in the region ζ > 1. For ζ close to unity, we have whereg 0 = ζ 2 − 1. At any R = 0 the curve 0 (ζ > 1) continues to decrease with increasing ζ and reaches the boundary of the lower continuum. It occurs at ζ cr A comment is in order. The single-particle solution for R → 0 should be modified. Indeed, for R as small as R ∼ r L r Λ e −3π/(2e 2 ) , the multi-particle effects of the polarization of the vacuum should be included, and the problem goes beyond the single-particle one, see the below consideration in Section 8.
Avoiding Problem of Falling to Center in Realistic Treatment. Spherical Nucleus of Finite Size
For the Coulomb field with the charge Z < 1/e 2 , the electron in the ground state is typically situated at distances ∼ a 1B = 1/(Z obs e 2 m) > 1/m and distribution of the charge Z(r) at distances r ∼ R nucl a 1B almost does not affect the electron motion. In the realistic problem, the nucleus has a finite size, R nucl r N A 1/3 a 1B , where A is the atomic number, r N 1.2 fm, and, thereby, the potential is smoothen at r < R nucl . The falling to the centrum does not occur, as it has been mentioned. Even for Z 1/e 2 , the electron density remains to be distributed at finite distances.
Taking into account of the distribution of the charge inside the nucleus, we have Two models have been employed in the literature: model I, when f (x < 1) = 1, that corresponds to the surface distribution of the charge, and model II, when f (x < 1) = (3 − x 2 )/2, which describes distribution of protons with the constant volume density.
The energy shift of the electron level can be found with the help of the perturbation theory that is applied to the Dirac system (18). Following [16], i.e., the curve (ζ) decreases monotonically with increasing ζ and crosses the boundary of the lower continuum with a finite value β. After that, (ζ) acquires an exponentially small imaginary part.
Because the exact solution of the Coulomb problem for r > R looks rather cumbersome and for r < R is impossible for a realistic cut of the potential, it is natural to use approximate methods. Most economical is a semiclassical approach. Here, we should notice that the replacement (19) becomes singular for < −m in the point V(r 1 ) = m + < 0. Because to this, the effective potential and semiclassical expressions loose their sense due to the divergency of the integral r 2m(E − U ef (r, )) 1/2 dr. However, this is only a formal problem, since the initial Dirac system (18) has no singularity at r → r 1 .
To avoid the problem one should bypass the singular point in the complex plane, as one usually does bypassing turning points, or one may apply the semiclassical consideration straight to the linear Dirac equations. Note that, in the one-dimensional case corresponding to κ = 0, see Equation (16), the mentioned singularity occurs in the turning points, and one may use standard semiclassical methods. The probability of the spontaneous production of positrons is determined by the width of the corresponding electron level, Im , for Re < −m. Thus the width is found from the solution of the Dirac equation. The value Γ, which determines probability of the positron production, W ∼ e Γt , can be expressed directly through components of the Dirac bispinor (G and F). It yields the flux of particles going to infinity (at normalization on one particle):
Accuracy of Calculation of Energy Levels in Semiclassical Approximation
Substituting ψ = Ae iS/h , where A and S are real quantities, in equation we find two equations For a convenience, the dependence onh is recovered here. The Hamilton-Jacobi equation for the action (∇S) 2 = p 2 is obtained provided where l is the typical size of the potential V. For the Coulomb potential at typical distances r ∼ 1/(2m) characterizing ground-state electron with −m we have p ∼g/r. From estimate (43), we see that the semiclassical approximation for the wave function for such distances is accurate up to terms 1/g 2 ,g = ζ 2 − κ 2 for ζ > |κ|.
Using the Bohr-Sommerfeld quantization rule, we havē where the phase γ ∼ 1, n r = 0, 1, ..., r 0 , and r − are the turning points separating the classically allowed region. Thus even in calculation of the energy of the levels with small quantum numbers one may consider on the error not larger that 10% . Finally, let us notice that the transition from the Dirac equation in the external field to the corresponding more simple Hamilton-Jacobi equation has been used in many investigations, cf. [85][86][87]. The case of the deep electron levels, with the energy < ∼ −m, was studied in [3][4][5][6][7].
Semiclassical Approximation to Coulomb Field of Point-Like Nucleus
In the field V = −ζ/r, for ζ < |κ|, the semiclassical method results in exact expression for the energy spectrum. Let us show this. For that, we do replacements Subsequently, the system of two Dirac Equation (18) reduces to equations with Adding the Langer correction to the effective potential results in replacements p i → p * i , we find Subsequently, applying the Bohr-Sommerfeld quantization rule, we have From here, we recover the exact result (31). To get (31) from an exact solution of the Dirac equations, we have performed a cumbersome analysis of hypergeometric functions, whereas the semiclassical approach needs taking only one simple integral.
After (48) is also valid for spinless bosons. Performing integration leads us to the exact expression (11).
Finite Nucleus. Semiclassical Wave Functions and Quantization Rule
Certainly , it is also possible to apply semiclassical approach to Equation (20) with effective potential in the form (21), (22). In the range, where the parameter of applicability of semiclassical approximation is |dλ/dr| ∼ 1, the usage of Dirac equations presented in different forms leads to slightly different results. For instance, applying (20) to the Coulomb field does not yield the exact result for the energy of the levels, although the accuracy of the approximation proves to be appropriate. For the electron energy < −m the variable replacement (19) leads to the singularity in the point r 1 , where V(r 1 ) = m + < 0 . Near this point, semiclassical expressions become invalid due to divergence of the contribution to the action [2m(E − U ef )] 1/2 dr. However, as it was mentioned, this circumstance is not reflected on the calculation of the energy levels, since r 1 is situated under the barrier, where wave functions prove to be exponentially small.
The electron energy levels can be found with the help of the Bohr-Sommerfeld quantization rule [3] applied to the Dirac equation presented in the form (20) with effective potential in the form (21), (22). We have Value p * is obtained from expression (20) after taking the Langer correction into account, i.e., after doing the replacement κ(1 + κ)/r 2 → (κ + 1/2) 2 /r 2 in the expression for the effective potential. The value of the phase γ depends on whether the turning point is inside the nucleus or outside it. In the latter case, the potential is V = −ζ/r and γ = 3/4 for κ = −1 and γ = 1/2 for κ = −1.
The contribution to the normalization of the semiclassical wave function from the classically forbidden region is usually dropped. In order to understand accuracy of this approximation consider the probability of the presence of the electron in sub-barrier region r − < r < r + : To be specific, let us put = −m and consider ζ |κ|. The wave function in the classically allowed region is [37]: Constant c 0 is found from the normalization condition [2], Subsequently, we expand the effective potential (21) near the turning point. For V = −ζ/r, we obtain The solution of Equation (20) in potential (53) is expressed through the Airy function The probability of finding the particle in the sub-barrier region is where Thus, the probability of a penetration of the electron in classically forbidden region is numerically small for ζ ∼ 1, and it falls down with increasing ζ. This justifies that we neglected the contribution of the region r > r − at the normalization of the wave functions (taking r 0 < r < r − ). Note that the quantization rule remains applicable with a larger accuracy, 1/ζ 2 , since, at its derivation, it was not used how wave functions are normalized. Strictly speaking, in the case of quasistationary levels, the quantization rule is slightly modified, due to Im = 0, cf. [88]. However, changes of the energy levels are exponentially small, due to the exponential smallness of the penetrability of the barrier.
With the semiclassical χ function, we obtain an expression for the averages r λ . For = −m and ζ |κ|, one has [3], Γ(x) is the Euler Γ-function. For ζ ∼ 1, the accuracy of this expression is not as good, but it increases appreciably with increasing ζ.
The quantity r characterizes the mean radius of the bound state at = −m, values r λ at λ = 1/2, 3/2, 2 are met in the problem of the modification of the value Z cr due to a screening of the charge by other electrons of the ion (if they are), see below in Section 6.4. A comparison of the semiclassical expressions with the exact solutions numerically found shows an appropriate accuracy of the semiclassical results, even for ζ ∼ |κ| ∼ 1. For ζ κ 1, the result (56) coincides with the corresponding asymptotic of the exact solution.
Critical Charge of the Nucleus
Let us calculate the critical charge of the nucleus (when the electron level with quantum numbers n, κ reaches = −m). Using the Bohr-Sommerfeld quantization rule in the form (50), one obtains, cf. [48], where y is positive root of the equation n r = 0, 1, ... radial quantum number, γ 1 = 3/4 for ns levels and γ 1 = 1/2 for κ = −1. In Ref. [48], quantity Ξ was found from matching of the exact solution inside the nucleus and semiclassical one outside the nucleus. As was shown in [8], usage of the semiclassical solutions both inside and outside the nucleus does not spoil the accuracy of the result. Therefore we further follow consideration of [8].
For the model I, the semiclassical solution inside the nucleus coincides with the exact one and we find Here, note that a first estimate of R cr in this model was performed in [47], where it was takenγ = ζ, that differs from that follows from (58), (59).
For the model II, an analytical expression can be found expanding p(r < R) in the parameter ζ,γ where f (x) follows Equation (37), here for the model II. Although the parameter of applicability of semiclassical expressions to the Coulomb field isg 1, the difference of the above obtained expression with the result of the exact calculation is less than few percents, even at ζ = ζ cr 1.24.
For ζ ∼ 1, expanding (58) in 1/y and dropping numerically small term e −2y , from Equation (57), we finally find from where we find Z cr (R nucl ).
Number of Levels Which Crossed Boundary of Lower Continuum
Now, let us find the number of levels n κ with fixed quantum number κ and the total number of levels N, which have crossed the boundary = −m. For this aim [5], we need to use the Bohr-Sommerfeld quantization rule at = −m. Forg 1, we have dλ/dr 1. For ζ 1, this means that ζ − |κ| ζ −1 , i.e., semiclassical approximation can only be violated for states with the momenta at which ζ − |κ| < ∼ ζ −1 . The accuracy of the semiclassical expressions for the wave function is ∼ 1/ζ 2 , cf. [2]. Taking these approximations into account, employing the Bohr-Sommerfeld quantization rule, we obtain For the potential that is given by Equation (37), for R nucl m 1, we obtain where ρ = |κ|/ζ , η = R nucl /r − = 2R nucl m/(ζ(1 − ρ 2 )), r − is the turning point in the effective potential, h(ρ) takes into account integral over the interior region of the nucleus 0 < r < R nucl , For Ze 3 1 (at this condition distribution of electrons, which fill the vacuum shell, only slightly modifies the bare potential, as we shall see below), Equation (63) correctly determines the distribution of electrons with < −m of the supercritical atom over the momenta j = |κ| − 1/2. The maximum value of j corresponds to r − = R nucl , η = 1, The total number of levels with < −m, can be found by replacing the summation by the integration. We should take into account that, in the Dirac equation, |κ| ≥ 1. Thereby, we still should subtract spurious term κ = 0. Thus, . For the model II, the result of this calculation is shown in Figure 3. Again, we observe an excellent accuracy of the semiclassical result, even for ζ ∼ 1. Expand the effective potential in m + , cf. [3]: where U ef (r, ) can be taken following Equation (21). Here, u 0 (r) = U ef (r, = −m). For n ≥ 1, where δ n1 is the Kronecker symbol. The energy of the levels is found from the Bohr-Sommerfeld quantization condition As before, γ = 3/4 for levels with κ = −1 and γ = 1/2 for κ = −1. With the help of (71) we find A comparison of numerical calculation done following these expressions with that for the exact Dirac equation again shows a good agreement. Note that the value β determines the threshold behavior of the probability of the production of positrons.
Energy Spectrum for | | −m
This spectrum has been found in [5]. For ζ ζ cr , many levels have energies | | −m. In this case, as follows from Equation (21) and (22), the terms ∝ κ in the centrifugal potential and in the spin term cancel each other. Approximately, we have For k = √ 2 − m 2 < ζ/R nucl , the turning point r − lies outside the nucleus, r − > R nucl . Employing the Bohr-Sommerfeld quantization condition, we get For deeper levels, k > ζR −1 nucl , classically permitted region r 0 < r < r − is completely inside the nucleus. Thereby, the spectrum is entirely determined by the expression for f (x): where Ξ n is the root of equation For example, for the model II at 1 n n * , we have From these expressions, it is easy to find expression for the level density dn/d . For model II, we find dn/d = Cy −1 , for 0 < y < 1 (78) and for y = kR nucl /ζ, C = const. From here, we see the accumulation of levels toward the boundary = −m (k → 0). For levels with arbitrary angular momenta the "Coulomb" part of the spectrum gets the form where ρ = |κ|/ζ, 0 < ρ < 1. Pre-exponential factor where h(ρ) that is given by Equation (65) depends on the f (x), e = 2.718... is the Euler number. The function c(ρ) monotonically decreases with increase of ρ from 1 for model I and from 1.4 for model II at ρ = 0 up to zero for ρ = 1 in both models. Equation (80) is obtained at the condition that the turning point r − lies inside the nucleus. The condition of applicability of Equation (80) isg/π n < n κ . Because n κ (g/π) ln(ζ/R nucl ), then, due to large values of the logarithm, this equation describes most of the levels crossed the boundary < −m.
The exponential dependence of n on n and the accumulation of levels near = −m, as follows from Equations (74) and (80), are related to the fact that U ef −g 2 /r 2 for r → 0. If R was zero, the electrons would collapse to the center. The spectrum of the Schrödinger equation in such a potential behaves as [89], where E 0 is the energy of the lowest level. In our case, E 2 /2m, and thereby we recover Equation (80) for c(ρ) = 1.
Exponential Estimate of Probability of Spontaneous Production of Positrons
Because, following Dirac the process of the production of e − e + pairs can be treated as the penetration of electrons of the lower continuum into the upper continuum through the classically forbidden region (p 2 < 0), the probability of this process is, as in case of spinless particles, determined by Equation (5). Equivalently, one can find the coefficient of transmission of the barrier in the effective potential or find semiclassical asymptotic of the functions G and F for r → ∞. This single-particle picture is distorted with a deepening of the level and with the increase of the number of levels crossed the boundary = −m. We may use Equations (20)- (22) while taking the Langer correction into account, which improves the application of semiclassical expressions.
In the threshold region of positron energies setting −m in the expression for the spin term U s , we obtain cf. with Equation (73) we have used for a description of the very deep levels. In case of the Coulomb field V = −ζ/r, replacing (83) in (5), we obtain that coincides with the asymptotic of the exact solution of the Coulomb problem.
Critical Charge of Nucleus for Muon
For the electron, one has R nucl 1/m, since 1/m 386 fm and R nucl r 0 A 1/3 A 1/3 /m π , m π 280m. For muon R nucl 1/m µ , m µ 207m e . In order to find the critical charge for the muon, ζ µ cr , when µ − level reaches = −m µ , we continue to apply the semiclassical approximation. For the model I, the turning point lies outside the nucleus. Let us expand U ef (r, ) near the turning point. Using Equation (54), after the replacement r − → r 0 , and matching solutions G /G at r = R nucl , we find [3]: for the ns level. From here follows that coincides with expression, which follows from the direct solution of the Dirac equation at = −m µ . In the model II we obtain ζ µ cr 16.7 that corresponds to Z µ cr 2300, and in the model I, respectively Z µ cr 3700.
Semiclassical Wave Functions
Let us apply semiclassical expansion to Equation (18), cf. [7]. The parameter of expansionλ/l is ∝h, where l is the typical length for the change of the potential. We present and arrive at the chain of equations for y n and φ (n) : One usually restricts expansion by consideration of first two terms. Because semiclassical series is an asymptotic one, retaining of too many terms may worsen the convergence of the series to the exact solution.
In order the system of homogeneous Equation (89) to have nontrivial solution, y −1 (r) should be an eigenvalue and φ (0) ≡ φ i , i = 1, 2, the eigenfunction of one of two-component eigenvectors of the matrixD(r). From the condition detD = 0, we get Replacing y −1 back to Equation (89), we obtain where A and A 1 are normalization constants. Because the matrixD is not symmetrical, besides the right-hand eigenvectors φ i , we should introduce the left-hand eigenvectorsφ i : Note that the left eigenvectors do not coincide with transposed right eigenvectors (φ i = φ T i ) and the left-hand and right-hand vectors are mutually orthogonal, To determine y 0 , let us put φ (0) = φ i in Equation (89) and multiply both sides of equation from the left byφ i . As follows from the first Equation (92), the term with φ (1) vanishes, and we obtain Further calculations entail no difficulty, cf. [7,90]. The resulting wave functions of the quasistationary state with energy < −m in the region of classically permitted motion r 0 < r < r − to have the form: Here, C 1 is normalization constant. As it was discussed, semiclassical wave functions can be normalized neglecting penetration of the particle into the classically forbidden regions r < r 0 and r > r − , i.e., r − r 0 (G 2 + F 2 )dr = 1. Thus, we find where T is the period of the particle motion in the classically allowed region.
In the sub-barrier region r − < r < r + , where p 2 < 0, p = iq and q, y −1 and y 0 are real, wave functions attenuate exponentially with increasing r. The resulting expressions have different forms in dependence on the sign of κ. For κ < 0, i.e., for κ = −1, we have with Q = q − κ/r. For κ > 0, we have with Q = q + κ/r, C 2± are normalization constants.
In the region r > r + , the quasistationary state describes outgoing positron and represents a diverging wave. For κ < 0: with P = p − iκ/r. The flux of particles moving to infinity is then given by Γ = lim Im(F * G) at r → ∞.
For κ > 0: with P = p + iκ/r. C 2± are normalization constants. The obtained formulas are valid for all r, except regions δr ∝ 1/ζ 2/3 near the turning points. The usual procedure is employed to match semiclassical solutions. The solution is either expressed in terms of an Airy function or one may use the Zwaan's method. Consequently, we have Note that the effective potential, which we have used in (20), can be presented while employing function w that appeared in (95): The terms in Equation (102), which contain the function w, are due to the electron spin. For |V| m, they are small compared to the first three terms. Subsequently, the expression for the effective potential takes the same form as for a scalar particle. At the turning points r − and r + , the effective potential is not singular.
Nonrelativistic Limit
To be specific consider case κ < 0 and the classically allowed region. Introducing a nonrelativistic energy˜ = − m and the variableq = (q 2 + κ/r 2 ) 1/2 q + κ 2qr 2 , let us transform the factor in exponent (97) as where Q = q − κ/r. The latter term in the integral cancels with the pre-exponential factor Q −1/2 . Now let us take into account that κ(1 + κ) = l(1 + l). Subsequently, we havẽ where C = const that reproduces the Schrödinger wave function in this region. Note that q(r) enters not κ = ∓(j + 1/2), but orbital moment l. We formally considered case κ < 0 just to be specific. Case κ > 0 is considered similarly. Additionally, note that, for κ = −1, one should add toq(r) the Langer correction.
Equation (108) determines the real part of the energy nκ . It differs from the ordinary Bohr-Sommerfeld rule used in nonrelativistic quantum mechanics by expression for relativistic momentum p(r) and by the term ∝ w appeared due to the spin-orbital interaction. Taking into account of the term ∝ w is legitimate within semiclassical scheme. Let us show it on an example of the Coulomb field V = −ζ/r. Subsequently, w(r) = − m+ 2(ζ+(m+ )r) and p(r) is determined by Equation (95). For r 0 < r < r − , the momentum p(r) ∼g/r and the ratio | κw p 2 r | ∼ |κg −2 rw| ∼ |κ|/ζ 2 for deep levels. Because semiclassical approximation for wave functions is valid up to 1/ζ 2 , the second term in the integral (108) should be retained in the case of deep levels | | m for |κ| 1, but it can be dropped for |κ| ∼ 1. For = −m, we have w = 0.
Note that the results of calculations performed with the help of the quantization rules (50) and (108) differ only in correction terms. For instance, from (108), we derive exactly the same electron energy spectrum as that given by Equations (80) and (81), with the help of the quantization rule in the form (50).
Probability of Spontaneous Production of Positrons
Let us calculate the probability of spontaneous production of positrons, Γ = −2Im . Replacing (99), (100) in (40), we find The last integral is understood in the sense of the principal value, being denoted as Pr, due to singularity at the point where V(r) = m + . In the nonrelativistic limit, the value Γ 0 = 1/T has the meaning of the number of impacts per unit time of the particle (localized inside the region r 0 < r < r − ) against the potential barrier at r = r − , and the exponential is the probability of the penetration of the barrier in each impact. The allowance for the relativistic effects and the spin change the expression for the period of the oscillations and add to (109) a factor depending on the sign of κ.
While taking into account that in the region of the barrier V is the purely Coulomb field, for w = 0 all of the integrals are calculated exactly: For the positron momentum k = For k ζ 1/2 m the width Γ is exponentially small for any κ. For |κ| (ζ/π) 1/2 expression simplifies as For |κ| < ∼ κ 0 = (ζ/π) 1/2 , the exponential factor in Γ becomes of the order of unity, and the semiclassical approximation becomes invalid. Note that κ 0 /κ max = 1/ √ πζ. Therefore, a number of levels diffused in the continuum, for which Γ is not exponentially small, is tiny for ζ 1.
Semiclassical Method for Noncentral Potentials Obeying System of Linear Dirac Equations
We described the spectrum of the quasistationary levels in the lower continuum for a spherical nucleus with the charge Z > Z cr . The results can be generalized to the case, when the potential does not obey spherical symmetry [7]. Let us present the Dirac equation as whereˆ α = γ 0 γ,β = γ 0 are Dirac matrices, and we recovered dependence onh. Let us present bispinor ψ as ψ = φe iσ and expand real quantities φ and σ in the parameter that is proportional toh: Replacing these series to Equation (112), we obtain the chain of equations The condition of existence of a nontrivial solution φ (0) , results in the Hamilton-Jacobi equation In difference with the spherically-symmetric case, the matrix is Hermitian; therefore, its left-hand,φ i , and right-hand, φ i , eigenvectors are Hermitian conjugates,φ i = φ † i , and With the help of this equation, from (114), we find a system of equations for σ 0 , Bispinors φ i are found by diagonalizing the matrixD −ˆ α∇S, so that the right-hand side of Equation (119) contains known quantities. Determining from this equation σ 0 , we obtain the quasiclassical solution of the Dirac equation In practice, the calculation of the functions σ −1 and σ 0 for noncentral potentials is a complicated mathematical problem requiring the solution of first-order differential equations in partial derivatives. In contrast to the case when V is spherically symmetric, in general case the result is not expressed in quadratures. If a parameter of a "non-sphericity" is small, then one may develop a perturbation theory.
Approach to the Problem
The minimal distance between colliding nuclei with charges Z 1 and Z 2 is as follows [18,91], where E c.m. is the kinetic energy of colliding nuclei in c.m. reference frame, b is the impact parameter. In order the energy of the electron, 1s , in the quasi-molecule would become < −m the colliding heavy nuclei should reach distances | r 1 − r 2 | = R < R cr , where R cr 33 fm for central U+U collisions, see below. Thus, R cr is approximately twice larger than 2R nucl , where R nucl 1.2A 1/3 fm is the radius of the single nucleus 7 fm. On the other hand, R cr r K 0.3ζ cr , where r is estimated using Equation (105). For U+U collisions, R cr /r K ∼ 0.2. Nuclei move with the velocity v A ∼ (0.025 − 0.07), cf. [18], whereas the electron of the K-shell has a typical velocity v e 1. Thereby, one may use adiabatic approximation, i.e., we may use (R(t)). Because R cr /r K ∼ 0. 2 1, the anisotropy of the potential is not as large, and we may present where Z = Z 1 + Z 2 , r 1,2 = | r ± R/2|, P 2 is the second Legendre polinomical, R(t) is the distance between centers of nuclei. In the second equation and, further we for simplicity, consider the case Z 1 = Z 2 . Otherwise, odd-power terms appear in the expansion. In inclusive experiments, this anisotropy disappears due to the averaging. However for eventby-event collisions such terms may lead to the forward-backward anisotropy reflecting in some observable effects. In the first approximation in (R/2r K ) 2 , the problem is reduced to that we have considered above for the spherical nucleus with the charge Z = Z 1 + Z 2 . The effective nucleus radius now is 2R nucl . The process of the spontaneous production of positrons can also be described in adiabatic approximation, since, as we have argued, we may use that (R(t)) and, since 1/Γ( (R(t))) τ col > ∼ 2R cr /v A . The most serious experimental problem is to separate spontaneous production of positrons in the tunneling process from the frequency dependent processes also resulting in a production of positrons. For example, the parameter 2R nucl /R cr ∼ (1/2 − 1/3) is not as small. Therefore, a serious competing time-dependent process is associated with an induced production of positrons occurring due to excitation of the nuclear levels, cf. [20,92] and the references therein. However, the difference between characteristics of the induced and spontaneous production of positrons is significant. The induced positron production exists in both subcritical and supercritical regimes. When the electron level crosses the boundary = −m, there appears a narrow energy-line in the positron spectrum owing to the switching on of the spontaneous positron production occurring in the tunneling process. Thus, there is a principal difference between the subcritical and the supercritical regimes that may help in the experimental identification of the spontaneous positron production.
Another effect is associated with the presence of a magnetic component of the field. First, an indication on presence of strong magnetic fields in heavy ion collisions was performed in [93]. For peripheral collisions of heavy ions at collision energies < ∼ GeV·A it yields h ∼ H π (Ze 6 ) 1/3 for R A 1/3 /m π , v A ∼ 1, H π = m 2 π /e. More generally, replacing For collisions with low energies E ∼ (5 − 10)MeV·A of our interest here, it follows that h ∼ 10 15 G, for R ∼ R cr (30 − 50) fm, and v A ∼ 10 −1 , cf. also [94].
In the presence of a "weak" homogeneous magnetic field, the reduction of Z cr in the case of the supercritical atom has been found by using the perturbation theory [95], H 0 = m 2 /e 4.4 · 10 13 G, µ 1/3 for ζ = ζ cr .
For strong fields, numerical evaluations [95], see also [19], yielded Z cr = 165 for h = H 0 , and Z cr = 96 at h = 10 2 H 0 . For h = 10 18 G, one gets Z cr = 41. This effect appears because of the exact compensation of the diamagnetic and paramagnetic contributions to the ground state for the electron. Although these estimates are performed for the case of purely uniform static magnetic field, they show that a magnetic effect also should be carefully studied for the case of realistic time-space configuration of the field.
Below, I only focus on the description of the spontaneous production of positrons and, simplifying this consideration, I also ignore the mentioned magnetic effects.
Electron Energy as a Function of Distance between Nuclei
Usage of the Bohr-Sommerfeld quantization rule allows for considering the problem analytically [8], cf. [7]. From (19)- (21), taking into account of the Langer correction resulting in the replacement p → p * , we have Here, r is the turning point for the given and r −m is the turning point for = −m. I used that in integration over the regions r < R/2, r < R cr /2 dependence on can be dropped, since at | | ∼ m of our interest, we have |V| | |. Thereby, the specifics of the behavior V(r) in the region r < R cr /2 almost does not affect the result. To be specific, we may use V = const for r < R cr /2. Integrals undergo logarithmic diverge at the lower limit. After their regularization, the dependence on R and R cr is separated in the explicit form: Integrals in (126) are calculated numerically. A comparison with the exact solution of two-center Dirac problem shows that the error of the semiclassical result does not exceed 0.1%. We can proceed further using that r|m + |/ζ < r |m + |/ζ 1 at least for | | ∼ m of our interest. Thereby, we expandã in Equation (124) in the series of r. As the result, we find F(r, ) = (g 2 + br + cr 2 ) 1/2 , From (126) and (127), we obtain For | + m| m, we find For U+U collisions for the ground-state level, we find ζ 1.343 and β 0.79. The slope-parameter β determines the probability of the production of positrons for | + m| m. The semiclassical approximation reproduces the Z dependence of β correctly, the difference with exact calculation done within solution of the two-center problem for the Dirac equation [96] is approximately (3-4)%.
Setting c = 0 in Equation (128), we obtain a very simple and accurate result [6][7][8]: The difference of this simple expression with exact solution of the two-center Dirac Equation [96] is less than (1-2)% already for ζ → 1 when the parameter of applicability of the semiclassical approximation is 1. Such an accuracy is sufficient; therefore, here I do not present a more accurate semiclassical expression [7] obtained without using expansion in c 2 , which has still higher accuracy. It may be curious to notice that, when in 1976 I showed the result (130) to Vladimir Stepanovich Popov, he did not believe in it, saying that one of his collaborators during a year is trying to solve the Dirac equation for the two-center problem numerically on ITEP big computer and, yet, only obtained the result for ζ = 1. He took the slide rule (that time there were no PCs) and confirmed that for ζ = 1 the whole curve (130) fully coincides with the result of the exact numerical calculation. Because the criterion of applicability of the semiclassical approximation for the ground state isg 0 1, it became clear that, for ζ > 1, the accuracy of approximate solution (130) should at least not be worse than in case ζ = 1.
Subsequently, the result (130) was reflected in our publications [6,7]. Result (130) is shown in Figure 4. For ζ = 1.343, κ = −1, we get − (R/R cr ) = 0.705(R cr /R) + 0.295. The expression for the critical distance between nuclei, R cr , can be found from Equation (62) for a spherical nucleus after replacement of the nucleus radius R nucl by R/2, where, now, R is the distance between nuclei and Z → Z 1 + Z 2 . Consequently, we find For the case of U+U collisions, in the model I that we obtain R cr 33 fm, whereas exact solution of the Dirac equation [96] yields R cr 34.3 fm.
Tunneling in the Two-Center Problem. Angular Distribution of Positrons
The potential of the system of two nuclei (121) contains, at r R, a quadrupole correction. In the sub-barrier region, the correction is < ∼ (R cr /(2r − )) 2 < ∼ 10 −2 . Therefore, the problem is reduced to the calculation of the penetrability of a three-dimensional barrier that only differs little from a spherically symmetrical one. Thus, we may use expansion We substitute these expressions to the Hamilton-Jacobi equation and obtain The first equation is easily integrated, resulting in Taking the first term into account leads to exponential term in Equation (110). Second term in (134) is due to anisotropy of the potential. Equation for S 1 in the under-barrier region r − < r < r + gets the form and it is solved by the method of separation of the variables. Supposing and taking into account the boundary condition ImS 1 (r − , θ) = 0, for r = r + we obtain For the angular asymmetry of the positron production, the constant a 1 is immaterial. A remarkable fact is that the expression for a acquires a hyperbolic cosine that enhances the angular anisotropy of the emitted particles when compared with the anisotropy of the potential. The cause of this effect is that the sub-barrier trajectory of a tunneling particle with nonzero angular momentum is not a straight line due to κ = 0. This leads to a substantial difference in the description of the three-dimensional and the one-dimensional tunneling of particles.
For the Coulomb field integrals (137) can be calculated exactly. However, the result looks cumbersome. An estimate shows that W(θ) exp(−2ImS) = Cexp(αP 2 (cos θ)), where C is a constant, α ∼ m 2 R 2 η −1 shη m 2 R 2 , η = 2πκ/g. For U+U collisions α ∼ 1/3, and we can expect a noticeable angular anisotropy. The positrons are predominantly emitted along the axis joining the nuclei at the instant of their closest approach. This question is worthy of experimental study.
Concluding, note that we needed the applicability of semiclassical approximation for both the radial motion and the angular motion. Strictly speaking, the latter takes place only for |κ| 1. However, as it always occurs, even for |κ| ∼ 1, one may expect good accuracy of semiclassical expressions.
Screening of K-Electron by Electron Cloud of Not Fully Stripped Quasi-Molecule
If the colliding nuclei are not fully stripped, the quasi-molecule is surrounded by an electron cloud. Screening weakens the attraction of the K-electron to the nuclei in the quasi-molecule. Consequently, the critical distance R cr , at which the K-electron level crosses the boundary = −m, is decreased. This effect can be calculated using nonrelativistic many-particle semiclassical approximation (Thomas-Fermi method), cf. [7,8]. Let us use that R cr r K a TF = (9π 2 /128) 1/3 (Ze 6 ) −1/3 /m 30ζ −1/3 /m , where a TF is the mean radius of the Thomas-Fermi atom. The shift of the ground-state electron energy level can be found with the help of the perturbation theory. We have where V 0 ( r) is the potential of the two striped nuclei (121) and V( r) is the potential of the two not fully striped ions. The typical size for the change of δV is a TF . Therefore, with the accuracy ∼ (R cr /a TF ) 2 ∼ 10 −5 , the perturbation can be considered to be spherically symmetric. Thus, r i = x 0 a TF is the radius of the ion, φ(r) is the solution of the Thomas-Fermi equation [84], with boundary conditions φ(0) = 1, φ(x 0 ) = 0, x = r/a TF , and Z 1 = −Zx 0 φ x (x 0 ) is the observed charge of the two partially screened nuclei. Expansion φ(x → 0) yields [84]: For the case of neutral atoms φ x (0) = −1.588. From (140) and (142) for the shift of the ground-state level, we obtain Values φ x (0) and φ x (x 0 ) are tabulated. We estimate |∆R cr /R cr | ∼ |∆ 0 / 0 | 10% for the ionization parameter q = (Z 1 + Z 2 − N)/(Z 1 + Z 2 ) 0.5, and 12% for q = 0, where N is the total number of electrons in the quasi-molecule.
Calculation of Positron Production
Employing the Imaginary-Time Method 6.5.1. General Description of the Method First, consider the problem of the one-dimensional motion of a relativistic particle in the potential V(x, t). The Lagrangian is as follows The constant is added to recover Lorentz invariance of the action since t is not a scalar. At the initial time-moment particle was in the point x 1 (t 1 ) and, at the final moment, in x 2 (t 2 ). In the semiclassical approximation, the wave function is The action is found from the Hamilton-Jacobi equation.
In the imaginary-time method, the sub-barrier motion is formally considered at imaginary values of the time variable. Performing the variable replacement τ = it, we arrive at the Euclidian action The trajectory x(τ) in the under-barrier motion, where S E is real, is determined by the condition δS = 0. From here, one finds the equation of motion, which has a meaning of the Newton equation With exponential accuracy, the probability to find the particle in the turning point of the exit from the barrier, if it initially were in the point of the entrance of the barrier, is given by This expression can be generalized to take the pre-exponential coefficient into account. However, we will restrict ourself by consideration of the exponential term. It is essential that the sub-barrier trajectory satisfies the classical equation of motion, but now in the Euclidian time. To find it and to calculate S and W, we may formally use the known equations of the classical physics.
Tunneling in Slowly Time-Dependent Potential
The case of space-dependent and slowly time-dependent fields was considered in [7], cf. [68]. For simplicity, consider a scalar particle in a one-dimensional field. Let the probability of the tunneling in the static limit be known, where x 1 and x 2 are the entrance and exit turning points, i.e., p(x 1 ) = p(x 2 ) = 0. Variation of the action due to a weak dependence of the potential on time V(x, t) yields We used equation of motion and integration by parts. The last integral can be calculated while using imaginary-time method. Thus, we obtain Dependence x(τ) is determined from (148) as where we used relationp 2 = m 2 − ( − V) 2 and that may only adiabatically change with time, i.e., it may depend on τ only via the dependence of one of the parameters.
Correction on Non-Adiabaticity to the Spontaneous Positron Production in Low-Energy Heavy-Ion Collisions
As a specific example, consider the probability of the spontaneous positron production in low-energy heavy-ion collisions. Deriving Equations (110) and (137), we assumed that, during a time of the tunneling ((r + − r − ) √ m 2 + k 2 /k), the potential V and did not have a time to change. Here, please do not mix typical time, for which the particle passes the barrier, cf. [97], and time 1/Γ, with an inversed probability to observe the positron. As we see from this simple estimate, adiabatic approximation does not hold at least for k → 0, i.e., in the vicinity of the boundary of the continua, | | m.
Let us find a correction to the penetrability of the Coulomb barrier due to finite speed of the colliding nuclei [7]. Following (121), the R(t) dependent correction to the static Coulomb potential is as follows Further consider the case when positrons are emitted along the axis that joins the nuclei, P 2 (0) = P 2 (π) = 1. Subsequently, the probability of their production is maximal. Expanding R(t) near the closest approach point, we obtain From (154) and (155), we have The imaginary time τ = it is found from Equation (153). Thus, we obtain where we introduced variable φ = 2arcsin[(r + − r)/(r − r − )] 1/2 , 0 ≤ φ ≤ π, r = r + cos 2 (φ/2) + r − sin 2 (φ/2), values τ = 0 and φ = 0 correspond to the instant of emergence from under the barrier. The total imaginary tunneling time is τ t = πζm 2 /k 3 , i.e., τ t → ∞ for the electron energy → −m, whereas, for deep electron levels, τ t strongly diminishes. The replacement of (157) in (152) yields where p = − , v p = (1 − m 2 / 2 p ) 1/2 is the speed of the positron, The ratio where for the collisions U+U (ζ = 1.343) is shown in Figure 5 as a function of the positron energy p . It is seen that δ < 0.1 for p > 1.65m. The adiabatic approximation in the problem of spontaneous production of positrons becomes invalid near p = m, where the positron production cross section is, in any case, tiny. Numerical calculations [18,48] have shown that R cr rapidly increases with increasing charge Z = Z 1 + Z 2 of colliding nuclei. The cross section of the spontaneous production of positrons increases in this case ∝ R 7/2 cr , while the correction for the non-adiabaticity of the tunneling decreases as 1/R cr at a fixed p . Therefore, it would be more convenient to perform experiments with heavier nuclei, for which R cr is larger.
Screening of a Source of Positive Charge in Presence of External Electrons
In a many-particle problem, most of the electrons in spherically symmetric potential well, V < 0, have angular momenta l 1. Thereby, to find distribution of the charge, we may deal with a more simple Klein-Gordon-Fock Equation (1) while assuming j l. The value of the maximum momentum, at which the electron placed in the positively charged ion where all levels with energies less than bound are already occupied is bound, satisfies the condition with bound ≥ −m. If there is a sufficient amount of external electrons, the resulting system is charge-neutral. In this case, we should put bound = m. Subsequently, p max = √ −2mV + V 2 , and taking into account that each cell of the phase space can only be occupied by two electrons of opposite spin, we have Thus, the relativistic Thomas-Fermi equation renders n nucl is the charged density of the nucleus. It is curious to note that such an equation for neutral atom has been introduced long ago [98], but a relativistic term was then treated as a small correction in nonrelativistic limit |V| m.
Filling of the Vacuum Shell by Electrons
Note that, even in the absence of external electrons, which may fill the empty states, in case when the potential well V < −2mc 2 electrons and positrons can be created already from the vacuum in the absence of any external electrons. Positrons go off to infinity, whereas electrons screen the initial positive charge of the source. In this case, we should put bound = −m. Subsequently, the relativistic Thomas-Fermi equation renders, cf. [1,2,99], where θ(x) is the step-function, with the boundary conditions on the boarder of the ion and with V(r) = −Z i e 2 /r for r > r i . Reference [99] presented numerical solutions. The thorough analytical and numerical study of the problem of the filling of the vacuum shell by many electrons was performed in an independent study [1,2]. This phenomenon was called "electron condensation", demonstrating that all of the vacuum levels are filled by electrons of the lower continuum, cf. [42].
A Detailed Derivation of Relativistic Thomas-Fermi Equation
The electron density can be found by direct summation of the moduli squared of the wave functions [2]: where ψ nκm are semiclassical wave functions presented in Equations (95)- (100). Actually, we need wave functions in the classically allowed region given by (95). Differentiating quantization rule (108) over n, we obtain ∂ ∂n where we dropped the term ∂ ∂n κw pr , which only leads to a small correction |w|/V 2 ∼ 1/ζ 2 , cf. [5].
Taking the exchange and correlation corrections in the relativistic Thomas-Fermi equation into account is conveniently done by means of a variational method analogously to that is performed for the nonrelativistic Thomas-Fermi equation [100]. We arrive at ν e 2 /π. For Ze 3 1 this correction can be safely dropped. For Ze 3 1 it can be taken into account in Equation (164) by introducing the renormalized coupling constant e 2 → e 2 (1 + 3e 2 /π), cf. [2].
Additionally, a correction appears due to that the dielectric permittivity of the vacuum, ε(eE), differs from unity, e E = −∇V. Thus, one should replace ∆V → ∇(ε(E)∇V) in Equation (164). However, this correction, as the correlation correction, is tiny, since ε(eE) = 1 − (e 2 /(3π)) ln(eE/m 2 ), and at distances r > ∼ 1/(a Z ) of our interest ε(eE) 1 + O(e 2 /(3π)), cf. [49] and Equation (252) Consider the screening of the positively charged nucleus of the initial proton number Z and the radius R (typically R nucl A 1/3 /m π , A ∼ 2Z). Assume that, inside the nucleus, the proton charge density is n 0 p = const. Introducing ψ = −V/m − 1 in the region V < −m (ψ ≥ 1), where the electrons of the vacuum shell give some contribution to the screening of the charge Z, from Equation (164) we obtain θ(x) is the step-function. For r > R nucl , with the help of the replacement x = r/r i , we obtain . (177) Here, Z obs is the charge seen at infinity. Because µ 1, we may use expansion Subsequently, we have equations At the edge of the nucleus x = 2mR/ζ 1. At x 1, we derive Inside the nucleus at the condition Ze 3 1, the potential is close to the bare one. Setting ψ = ζy(Ξ)/(R nucl m), Ξ = r/R, we obtain Using that inside the nucleus |V| ∼ ζ/R nucl ∼ Z 2/3 m m and 4π 3 R 3 nucl n 0 p = Z, we get Because ν 1, we expand and get Matching of V and V at the edge of the nucleus yields and 7.5. Strong Screening, Ze 3 1 Continue to consider a nucleus with Z ∼ A/2 and R nucl Z 1/3 /m π . Because R grows with Z, one may expect that, for a sufficiently large Z, most of the electrons enter the nucleus and the interior becomes charge-neutral, as infinite matter. For the bare nucleus, the energy that is associated with the electric field, increases with Z more sharply when compared to the binding energy ∼ A ∼ Z, thereby the volume-charged systems do not exist. The charge, if it exists, is repelled to the surface. To approximately solve Equation (164), we now introduce variables x = (r − R nucl )/l and V = −V 0 χ(x). Constant V 0 is found from the condition of the charge neutrality at x → −∞, i.e., V 3 0 /(3π 2 ) = n 0 p for V 0 m. Thus, in new variables, Equation (164) renders with boundary conditions χ(−∞) = 1, χ(∞) = 0. The latter condition just means that typical decrease of the potential occurs already at x ∼ l near the nucleus boundary, whereas the transition to the Coulomb law occurs at x l. The solution at such large distances can only be found numerically.
Because, in dimensionless equation with dimensionless boundary conditions typical |x| ∼ 1, for R nucl l, which we assume, we can neglect the second term in l.h.s. of Equation (190). In this case, geometry becomes one-dimensional and Equation (190) reduces to where we determined the length l, as Taking the boundary conditions into account, the first integral of Equation (191) is as follows and the final solution is Note that Equation (191) allows for very simple approximate solution for x < 0. To get it, we write χ = 1 + ψ, ψ 1 and, from (191), find Using the boundary conditions at x = 0, we find C 0.24. This solution with an error less than 1.5% coincides with the exact solution.
The maximal strength of the electric field is reached at the edge of the nucleus, 3 π 1/6 (n 0 p ) 2/3 8.2 · 10 19 V/cm , that 6000 times exceeds the electron QED unit E QED = m 2 c 3 /(eh) 1.3 · 10 16 V/cm. Note that, to obtain this conclusion, we essentially used the relation R nucl ∼ Z 1/3 /m π .
The energy of the system can be recovered by the integration of Equation (164). For |V| m, we have Expression (189) is obtained, after one puts to zero the term V 4 12π 2 related to the electron condensation and employs the partial integration and Poisson equation.
In our case, ∇V = 0 inside the system for R nucl l and V 0 = (3π 2 n 0 p ) 1/3 . With these values, Equation (197) yields Accordingly, the energy is reduced to the kinetic energy of the degenerate relativistic electron gas filling all energy levels of the vacuum shell with < −m. One should add to it the energy that is associated with the strong interaction of nucleons resulting in the binding of the ordinary atomic nuclei. In such a way, we get transition to the description of infinite matter. We see that, not taking into account a pion condensate or some other complex processes, we have E > 0 and such a matter, without inclusion of the gravity, is unstable, cf. [2,43].
If, instead of the usage that A ∼ 2Z, we assumed the validity of the β equilibrium conditions, n ↔ p + e +ν, we would get A Z, and taking into account the gravity and the filling of all electron levels up to = m, we would recover the description of the ordinary neutron-star matter, cf. [43].
Falling to the Center in Relativistic Thomas-Fermi Equation
For V = −Ze 2 /r, the number of electrons filling the vacuum shell is for r → 0. Now, consider a formal solution of Equation (164) at r < r i with boundary conditions (165) corresponding to that for r > r i , we deal with the Coulomb law with the charge equal to the observable charge Z obs . As we shall see, such a problem has a unique solution independently on the charge Z 0 put in the center, i.e., at r → 0. It proves to be that the exact solution of Equation (164) has the pole singularity already at a finite value r = r pole (µ). In a weak screening limit from Equation (176), for r → r pole (µ), in the dimensionless variable x = r/r i , x pole = r pole /r i , we get [3], .. The substitution of (200) in Equation (176) allows for finding coefficients a n , but does not allow for recovering dependence x pole (µ). To obtain a full solution of the problem, we need to solve Equation (176) with the boundary conditions (165) in the whole interval 1 > x > x pole (µ). The numerical solution yields x pole (µ) = r pole (µ)/r i = D(µ)e −1/(8µ) , µ → 0 .
Pre-exponential factor D(µ) is shown in Figure 6. For Z obs 1/(2e 2 ) with increasing Z obs the pole moves towards the value 1/m. We conclude that, in the many-particle problem, including the electron condensation but not including the polarization of the vacuum, the falling to the center manifests itself in the presence of the pole at a distance r pole (µ). Accordingly, in the problem of the distribution of the charge at r 1/m, there appeared a typical size r pole (µ), which characterizes the electron condensation, where all of the states are occupied according to the Pauli principle. Thus, we have found a relation between Z obs and Z(r 0 (µ)), for the size of the source r 0 > r pole (µ). To match this exterior solution with the interior solution for r < r 0 , we may use either model I or model II. It is important that r 0 should be larger than r pole (µ).
At this instance, we should remind about the existence of the Landau pole for r = r L e −3π/(2e 2 ) /m, which appears within the multi-particle problem of the polarization of the electron-positron vacuum near the Coulomb center, cf. [49]. Comparison of the exponential factors shows that, for Z obs < 1/(2e 2 ), we have r L > r pole (µ) and, for Z obs > 1/(2e 2 ), we have r L < r pole (µ). Thus, in the case Z obs < 1/(2e 2 ), with decreasing r, first the polarization of the vacuum becomes effective and only at r in a narrow vicinity of r L , where Z(r) > 1/e 2 , the electron condensation becomes to be efficient. For Z obs > 1/e 2 , the electron condensation first becomes effective and only at r in a narrow vicinity of r pole (µ) > r L the polarization of the vacuum begins to contribute, see a detailed discussion below in Section 9.
Note that the value Z obs e 2 plays a role of an effective coupling in description of semimetals and effects under discussion might be relevant in this case, cf. [60].
It is curious to note that the inclusion of gravitational field of the source into consideration modifies the QED problem of the distribution of the charge while taking the electron condensation into account, cf. [101]. Solution (200) is modified at r approaching r pole . After a growth, solution continues up to r → 0 as V → −Z 0 e 2 /r with Z 0 ∼ Z 2/3 obs /(eGm 2 ) 1/3 , where G is the gravitational constant. Additionally, the pole solution (200) disappears in case of the electron condensation in presence of a strong uniform magnetic field, cf. [102].
At the end, note [3] that Equation (164) can be solved within the main logarithmic approximation [49,103], being broadly used in different problems of the quantum field theory, see a discussion below in Section 8.1. Introducing variables ψ = φ(x)/x, t = − ln x, x = r/r i , in ultra-relativistic limit |V| m, we obtain Assume φ = ∑ ∞ n=1 µ n φ n with φ n = C n t n + O(t n−1 ) for t → ∞. Subsequently, we get solution C n = 2 n+1 (2n)!/(n!) 2 that finally yields ψ(x → 0) = C n x −1 (− ln x) n + ... A summation of these terms yields solution which has a spurious square-root singularity at x → x 0 = e −1/(8µ) , whereas the exact solution has the pole. Thus, this example demonstrates the possible deficiencies of the main logarithmic approximation in cases when we deal with divergent series.
Polarization of Vacuum in Uniform Stationary Electric and Magnetic Fields
In the absence of external electromagnetic fields, electrons of the lower continuum have infinite energy where 0,− pσ = − m 2 + p 2 are negative-sign solutions of the dispersion relation of the free Dirac equation. In pure QED, i.e., at ignorance of gravitational effects, infinite constant (204) has no sense, being subtracted within renormalization procedure. In the presence of the electric and magnetic fields energy levels of the lower continuum, − pσ are changed. The difference has the physical meaning.
The ground-state corresponds to the "−" sign solution. To calculate the sum (205), one uses that the number of states in the interval dp z in the uniform magnetic field is given by cf. [39]. Taking into account the double degeneracy of levels with n, σ = 1 and n + 1, σ = −1 excluding ground state n = 0, σ = −1, with − pσ solution, one obtains The divergence of integrals is removed by the subtraction of E 0 . To do this renormalization, it is convenient to calculate a convergent derivative of the energy After double integration and subtraction of the value E 0 , we obtain The contr-terms C 1 and C 2 do not depend on m 2 , but may depend on H.
In the case of uniform stationary fields E and H, the Lagrangian density L = −E can only be a function of Lorentz invariants E 2 − H 2 and E H. Note here that, in the presence of the sources of the current, the Lagrangian density additionally depends on j µ A µ .
In the case under consideration employing arguments of dimensionality and parity in H, one can write The first term is the ordinary Lagrangian density in the magnetic field, whereas the second term is the contribution of the polarization of the vacuum in the magnetic field. In Equation (211), there are no terms odd in m 2 , so C 2 = 0. Using that cthx = x −1 + x/3 for x → 0, we may see that the absence of H 2 term L (H) corresponds to the choice In the case of uniform static magnetic and electric fields, function f (H) in (211) should be replaced by At H = 0, thereby f (0, E) = f (−E 2 , 0). At E = 0, f (H, 0) = f (H 2 , 0). From here, we see that f (0, E) = f (H = iE, 0), i.e., the expression (211) for the case H = 0, E = 0, remains valid after replacement H → iE. Note that f (−E 2 , 0) has a small imaginary part associated with a possibility of the tunneling of a part of electrons, which initially occupied levels of the lower continuum, to the upper continuum. Created in a sufficient number, the electron-positron pairs change the spatial dependence of the electric field. In a realistic treatment of the problem one should consider electron and positron condensates occurred near the plates of the capacitor, which produced initially uniform electric field.
In case of strong uniform electric and magnetic fields |eE|/m 2 1 and |eH|/m 2 1, with a logarithmic accuracy from Equations (210), (212), one finds expressions for the dielectric and magnetic permittivities [49,104]: The corresponding contributions to the energy of the lower continuum are Note that expressions (214) are derived with the logarithmic accuracy, i.e., at the assumption that ln(|eE|/m 2 )| 1 and ln(|eH|/m 2 )| 1. Thereby, they are also formally applicable for negative values of ε and µ provided for the calculation of the vacuum energy in stationary uniform electric and magnetic fields one may employ the singleparticle Dirac equation. At this assumption they are invalid only in a narrow region of fields, where | e 2 3π ln(|eE|/m 2 )| ∼ O(e 2 ) and | e 2 3π ln(|eH|/m 2 )| ∼ O(e 2 ). The result (214) also follows from the Dyson equation for the photon propagator that was calculated at one-loop, but with the electron Green functions that are dressed by the background field. In such an approximation, the radiative photon corrections to the electron Green function and vertices in the photon polarization operator are dropped. Figure 7 shows the effective action with one-particle irreducible (1PI) diagrams presented up to twoloops. The same result (214) is also recovered within the so-called main logarithmic resummation, when e 2l ln l (eE), e 2l ln l (eH) terms in the Dyson equation for the photon Green function are summed up, whereas terms e 2l ln l−1 (eE), e 2l ln l−1 (eH) are disregarded, cf. [49,103,[106][107][108][109]. The radiative photon corrections to the electron Green function continue to be disregarded. The difference between two approximations is only manifested in the region where e 4 ln 2 (eE) > ∼ e 2 , e 4 ln 2 (eH) > ∼ e 2 . At the two-loop order, the term that is included in the effective action is given by the sandwich diagram (the one-particleirreducible (1PI) contribution). The resulting dielectric and magnetic permittivities up to correction terms are Recently, Refs. [110][111][112] studied the role of the one-particle reducible (1PR) loop diagrams. In this scheme, Figure 8 shows the effective action up to four loops. These 1PR diagrams yield zero contribution in the case of constant fields [107], since, in the case of purely constant classical fields, the four-current term is absent. However, the argument for the vanishing of the current no longer holds as soon as the external field supports a slightest inhomogeneity somewhere in the space-time [110]. In the latter case, all possible 1PR loop diagrams, being included, can be constructed from the 1PI one-loop constant-field diagram. The result of such a resummation of the diagrams in the strong-field limit yields [112], Note that, although, formally, these expressions are derived in the approximation ln(|eE|/m 2 ), ln(|eH|/m 2 ) 1, as noticed in [112], they cannot be valid at least in the region where |1 − (e 2 /(3π)) ln(|eE|/m 2 )|, |1 − (e 2 /(3π)) ln(|eH|/m 2 )| < ∼ e 2 , due to the presence of the pole in expressions (217). For example, the dielectric permittivity ε 1PR (E) → −∞ for (e 2 /3π) ln(|eE|/m 2 ) → 1 − δ and ε 1PR (E) → +∞ for (e 2 /3π) ln(|eE|/m 2 ) → 1 + δ for δ → 0. Conversely, (214) and (216) do not produce any non-physical singularities, yielding zero, rather than the pole at (e 2 /3π) ln(|eE|/m 2 ) → 1. In the region where ln(|eE|/m 2 ), ln(|eH|/m 2 ) 3π/e 2 expressions (214) In the one-loop order, results (214)-(217) coincide. Beyond the one-loop approximation, various partial resummation schemes produce different results.
To proceed further, we will use expression With ν = 1, we deal with the result [49,104,106], for e 2 3π ln(|eE|/m 2 ) < 1 being recovered within the main logarithmic approximation for the 1PI diagrams and, for e 2 3π ln(|eE|/m 2 ) 1, being also recovered within the main logarithmic approximation applied for the 1PR diagrams. With ν, being a very smooth function of the tortoise variable ln(|eE|/m 2 ) varying from 1 at |eE| ∼ m 2 to 1/2 for ln(|eE|/m 2 ) 3π/e 2 , we recover the asymptotic behavior that was derived in [112] with the included 1PR loop diagrams.
At the end, we stress that both main-logarithmic resummation schemes considered above may be not valid for (e 2 /3π) ln(|eE|/m 2 ), (e 2 /3π) ln(|eH|/m 2 ) → ∞, since the dropped sub-series of the diagrams may yield divergent contributions. We have demonstrated examples of such a kind in Section 7.6, cf. [3]. A summation of the 1PR diagrams leads to the appearance of the pole in expressions ε 1PR (E) and µ 1PR (H) for (e 2 /3π) ln(|eE|/m 2 ) = 1, (e 2 /3π) ln(|eH|/m 2 ) = 1. Moreover, recall that the expansion in the number of loops is a semiclassical series. The latter series is an asymptotic one, and retaining too many terms may worsen the convergence of the series to the exact solution. Bearing this in mind, the result that is given by ε HE (E), µ HE (H) looks more physically motivated. Nevertheless, further on, we use Equation (219) varying parameter ν in the interval (1/2, 1) to recover both asymptotics in Equation (218).
Noninteracting Photon, Electron, and Spin-Zero Boson Propagators
The Green function of the free photon is given by T is the ordinary time ordering, operators are in interaction picture, cf. [49]. The most general form is as follows, g µν is the metric tensor. One usually uses the Feynmann gauge condition D 0 (l) = 0.
In the Feynmann gauge, The free propagator of spin 1/2 electron is where Ψ = Ψ † γ 0 and Ψ 0 i (x) satisfy the Dirac equation (γ µp µ − m)Ψ 0 i (x) = 0. Thus, the Fourier transform is We may turn the contour in p 0 plane against clock arrow not touching poles and, then, we perform replacements ip 0 = p 4 , ix 0 = x 4 , px = −px = −(p 4 x 4 + p x),p = ( p, p 4 ), x = ( x, x 4 ), dp 0 /i → dp 4 . Let us present u = 1/α. Forxm 1, we may put m = 0 and find Forxm 1, we may use the pass method and present and we find For Dirac electrons Thus, forxm 1, we obtain G 0 (x) = γ µ x µ 2π 2 x 4 , the electron Green function is odd function of its coordinate argument. The power law increase of G 0 for r → 0 reflects the fact that there is no scale of the length, which could describe the free particle at r 1/m. For r 1/m, processes of polarization of the vacuum in the absence of external fields are suppressed as follows from Equation (229).
Dyson Equation for Photon Propagator
Taking the vacuum polarization diagrams in the first order perturbation theory in e 2 into account, the Dyson equation gets the form In the momentum representation, we obtain The last factor (−1) comes from the closed fermion loop. The next terms in the full Dyson equation are constructed analogously. The sum of all irreducible diagrams (which cannot be separated by a single photon line) is called the photon polarization operator, −iΠ µν . Thereby, in the lowest order −iΠ λρ 0 = Tr[(−ieγ λ )iG 0 (p + k)(−ieγ ρ )iG 0 (p)]. In brief, notations Dyson equation renders In the lowest order in e 2 one has Π = Π 0 .
Case of a Weak Static Electric Field. Renormalization of Charge
To remove divergencies in observables, one employs renormalization procedure. Below, we demonstrate this procedure on an example of renormalization of the charge. One assumes that, initially, the action enters the bare coupling e 2 0 rather than physical one, e 2 = 1/137. As we shall see, the polarization characteristics are divergent for r → 0. At the same time, the r → 0 limit is legitimate, because QED is the theory with the local interaction. To proceed, one introduces the cut-value r 0 , with performing the limit r 0 → 0 in final expressions. According to diagrammatic rules in the first non-vanishing order At r > 1/m, in the case of weak external fields, the effects of polarization of vacuum should be suppressed, since the electron Green function and, thereby, the photon polarization operator decrease exponentially in Euclidean variables, cf. Equation (229). Therefore, consider the opposite limit casex 1/m when the effects of the polarization of the vacuum can be significant. We recognize that at short distances there is no scale of length, except the Compton wave length. Thus, G 0 and Π 0 µν should be power-law functions ofx. We have In mixed ω, R representation: Using Equation (221) with D (l) = 0, we have Multiplying Equation (233) by e 2 0 n ext ( r) and integrating, we arrive at the Poisson equation for the static field V( r) = e 0 A n.ren 0 = eA ren 0 , being expressed in terms of nonrenormalized quantities, where in case of weak fields we took the polarization operator in the lowest order, i.e., K 0 00 (ω = 0, R) = Π 0 00 (ω = 0, R)/e 2 0 . K 0 00 (ω = 0, R) does not depend on e 2 0 . As will be shown below, K 0 00 (ω = 0, R) diverges for r 0 → 0. Now, our aim is to rewrite the Poisson Equation (239) in the form ∆V = −4πe 2 n ext ( r) .
To perform this procedure of renormalization of the charge, we continue to consider the polarization of the vacuum in a weak field, i.e assuming n ext to be small. Subsequently, we may use expansion We may drop convergent terms in the expansion (240) irrelevant for the renormalization procedure. The term K 0 00 (ω = 0, R)d 3 RV( r) should be put zero, since constant potential cannot produce polarization charges due to gauge invariance. The term K 0 00 (ω = 0, R) Rd 3 R∇V( r) = 0 due to isotropy of the vacuum in the weak field. Hence, we obtain ∆V = −e 2 ren 4πn ext , e 2 ren = e 2 = Finally, we derived a formal relation between the bare coupling constant e 2 0 and the physical one e 2 = 1/137. After this procedure is performed, we may say that all physical values already depend only on e 2 . Thus, in the lowest approximation over e 2 0 using Equation (237) and relation between Π 0 00 and K 0 00 , we obtain The formal solution of the first equation for any e 2 0 > 0 yields e 2 → 0, rather than e 2 . This is known as "the problem of the zero charge", (or "Moscow zero"), cf. [49]. Strictly speaking, such a consideration suffers from inconsistency, since the inverse relation given by the second equation has so called Landau pole for From the second Equation (242), for r 0 → 0, follows the solution corresponding to e 2 0 < 0 and imaginary e 0 . A similar procedure could be performed in four-invariant form for the 4-potential e 0 A µ , instead of e 0 A 0 .
Case of a Strong Static Electric field
In the presence of a strong static electric field the electron polarization operator, even being considered with the only one-loop diagram, should be calculated with full electron Green functions, G, instead of free ones [42,53]. In this approximation, expression (234) is replaced by At this level, the Ward-Takahashi identity is only satisfied approximately. It can be fulfilled exactly after taking the higher order diagrams into account. Multiplying Equation (233) by e 2 0 n ext ( r), we derive the Poisson equation for the static field V( r) = e 0 A n.ren 0 = eA ren 0 , expressed in terms of non-renormalized quantities, where K 00 (ω = 0, r, R, e 2 0 ) = Π 00 (ω = 0, r, R, e 2 0 )/e 2 0 . Being expressed in non-renormalized terms, both of these quantities depend on e 2 0 . For G → G 0 , they transform to K 0 00 (ω = 0, r, R) = Π 0 00 (ω = 0, r, R, e 2 0 )/e 2 0 . We again use expansion (240). The term K 00 (ω = 0, r, R)d 3 RV( r) should be put to zero, since the constant potential cannot produce polarization charges due to the gauge invariance. The term K 00 (ω = 0, r, R) Rd 3 R = 0 due to the symmetry respectively replacement r ↔ r . Accordingly, we obtain where we retained the residual convergent term δn 1 . Let the field E( r) be locally directed in the z direction. Subsequently, we rewrite where ρ 2 = x 2 + y 2 . The renormalization of the charge is performed by addition and subtraction to n 1 the term where we used isotropy of the quantity K 0 00 (ω = 0, R). Thus, we obtain We may also use another intuitive argument in favor of a formal validity of this expression at ε(E) < 0. For this, let us consider theory with N 1 number of charged species with masses ∼ m and let the coupling is e 2 /N, cf. [113]. Afterwards, instead of Equation (252), we immediately arrive at expression being valid in the region, where ε(E) > 0, as well as for ε(E) < 0. Note that, obviously, expressions (214) that are derived by Heisenberg and Euler for the cases of purely uniform fields [104] also continue to hold for slightly inhomogeneous fields provided where R H = 1/ |eH| is the typical radius of the curvature of the charged particle trajectory in the magnetic field (Larmor radius) and R E = 1/ |eE| is the typical radius of the curvature of the charged particle trajectory in the electric field. Thus, for the electric field of the form criterion of applicability of approximation of a uniform field coincides with inequality Q(r) 1 provided rQ 1. Accordingly, the expression for the dielectric permittivity of the vacuum (214) derived for the case of the uniform field coincides with (252) with the logarithmic accuracy and with the same accuracy we may write interpolation expression Here, we additionally inserted a smooth function ν varying within the interval (1/2, 1). With ν = 1/2, we recover the asymptotic behavior that is found by a resummation of the sub-set of 1PR diagrams [112], as we have discussed above. Once more, notice that we will use Equations (214), (252), and (257) for both ε(E) > 0 and ε(E) < 0. There exist corrections to Equation (257) in the region, where |ε(E)| ∼ e 2 ; however, as we have discussed, there are no physical reasons to expect the presence of any singularities in this region. Therefore, it seems reasonable to use the same expression (257) at all distances.
Polarization of Vacuum and Electron Condensation
In the presence of charge sources, the Lagrangian density is already not only a function of E 2 , as was the case in the purely uniform field, but it contains the term n ext V. The charge sources always exist in a realistic problem. Indeed, the uniform electric field can only be constructed in a limited region of space, namely inside the capacitor with the length of plates l d, where d is the distance between the plates. Outside the capacitor, the field decreases to zero. The electron-positron pairs produced in the tunneling process inside the capacitor go to the plates. The electrons are localized near the positively charged plate and positrons, near the negatively charged one.
Recall that the energy of the electron in a smooth field V in the classical approximation is given by cf. Figure 2, demonstrating the boundaries of the upper and lower continua in the field V < 0. The upper sign solution corresponds to states that originate in the upper continuum, which can be occupied in an attractive field for electrons, V < −2m in the case of a broad potential well, after the tunneling of electrons from the lower continuum. In the standard interpretation, see the discussion in Section 3.4, the lower sign solution corresponds to positrons after replacement → − . Let us also study another interpretation when the lower sign solution corresponds to electron states that originate in the lower continuum, being occupied by the electrons. As we show below, this interpretation might be relevant in a specific case, when ε < 0 in some region and, thereby, the resulting potential V > 0. The introduction of the electric field in the Dirac equation for electron corresponds to the replacement → − V. Let us expand the potential V( r ) near a point r: Assuming V( r) to be very smooth function of coordinates, we may only retain these two terms in the expansion. It is easy to ascertain the consequences of the replacement −e E r → V( r) − e E r. The term − ∑ ψ * e E Rψd 3 R was already taken into account in the problem solved by Heisenberg and Euler in the case of purely uniform electric field. The expressions for the Lagrangian and the energy of the lower continuum in uniform fields are more easily calculated for the case of purely magnetic field as we have mentioned. We found Equation (208), where typical momenta p z contributing to the sum are p z ∼ |eH|. In case of purely electric field the typical momenta contributing to the sum are p z ∼ |eE|. Performing summation in Equation (208) Refs. [49,104] derived expression (211) and with the help of invariants recovered Equation (213). After doing replacement H → iE, |eH| → |eE| one arrived at expressions (214). Now, see Equation (258), in the expression for the energy, there appears an additional potential term since ∑ njm |ψ njm | 2 = |V 3 | 3π 2 > 0. The upper sign is for V < −2m and the lower sign is for V > 0 and we, for simplicity, assume |V| m. There is still a kinetic term in the energy, see Equation (258), which we should add while considering the condensation of electrons, corresponding to the region of momenta | p| ∼ |V| m rather than to | p| ∼ |eE|, the latter term we have included. At least in limit cases V 2 |eE| and V 2 |eE|, the mentioned contributions are not overlapped. As a result, the kinetic term is The upper sign corresponds to the electron condensation on levels of the upper continuum that is occupied during the tunneling of electrons from the lower continuum in the field V < 0. We have studied this case in Section 7. The lower sign solution corresponds to the electron condensation on levels of the lower continuum, may be possible for V > 0, compare with the first term in Equation (208), which was summed up in the case of the magnetic field.
Finally, in the case of a weakly inhomogeneous electric field we obtain From the semiclassical derivation, one may see the difference between the condensation of electrons on levels of upper continuum crossed the boundary = −m, cf. Equation (197), and condensation on levels in the lower continuum in a repulsive field.
In the former case, vacant states with < −m are occupied only in the process of the tunneling of electrons from the lower continuum. In the upper continuum, the kinetic energy of electrons is positive E kin = + ∑ ψ * | p|ψd 3 x, | p| > 0, whereas the kinetic energy of electrons occupying levels of the lower continuum is negative, E kin = − ∑ ψ * | p|ψd 3 x, | p| > 0, cf. the first term in Equation (208), has been used in the case of the uniform magnetic field.
Variation of the energy yields the Poisson equation, cf. Equation (164), which described the electron condensation in the attractive potential of a supercharged nucleus at ε 1. Although we are interested in the case |V| m, we recovered the dependence on m in Equation (263). Now, for ε > 0 and V < −2m, we deal with the electron condensation on levels of the upper continuum crossed the boundary = −m with increasing |V|, as it follows from the standard interpretation of the levels, appearing from the upper continuum during an adiabatic increase of |V|. Below, we will argue for a possibility of the condensation of electrons that originated in the lower continuum in the problem of the screening of the positively charged source at ultrashort distances from it (at r < r L ), ε ren (r L ) = 0, ε ren (r < r L ) < 0 and the potential is repulsive due to that.
For r > ∼ 1/m, we can set ε(r, Q(r)) 1 and, thereby, we may put C = Z obs e 2 . The potential V is easily recovered in the case of a smooth variation of the charge This condition is fulfilled for |Q 1 | |Q 1 |/r that yields |ε(r)| e 2 /(3π). The solution of Equation (265) has two branches, one corresponds to ε(r, Q(r)) > 0, other relates to ε(r, Q(r)) < 0. We assume Z = Z obs for r > ∼ 1/m and find Q(r) for decreasing r. Subsequetly, we obtain Q(r) = Z obs e 2 /ε(r, Q(r)) on the positive branch of ε(r, Q(r)). Expression (267) has a kink at r =r m , ε(r m ) ∼ e 2 /(3π) and Q(r m ) ∼ 3πZ obs 1. Therefore, Equation (267) only has a meaning for r 0 >r m . Only then can one find a relation between Z obs and Z 0 . However, note that, actually, Equation (267) already becomes invalid at a slightly larger r thanr m , when (r, Q(r)) reaches values ∼ e 2 . At these distances, Equation (265) for ε becomes invalid and approximation (266), which we have used, also fails.
A comment is in order. Consider what would be, if we used Equations (251) and (265) for r <r m . Subsequently, we would get Q 1 (r) = −Z 0 e 2 /ε(Q 1 ) > 0, ε(Q 1 ) < 0. This solution becomes invalid in the vicinity ofr m , where −ε ∼ e 2 , now for r <r m , and it cannot be smoothly matched with the solution we have derived for r >r m .
Electron Condensation on Levels of Upper Continuum Is Included
In the region, where Q(r) > 1, besides the vacuum polarization, cf. Equation (251), we should include the electron condensation on levels of the upper continuum crossed the boundary < −m, cf. Equation (263). Thus, we have −V m. The solution of this equation can be easily obtained in the approximation (266). We have [42], To be specific, consider the case Q obs 1. Constant C is determined from the condition Q(r > ∼ 1/m) Q obs = Z obs e 2 , since ε(r > ∼ 1/m) 1. Thus, we obtain This solution shows an apparent pole at r = r ap pole . Near this point, in the region where ε(r, Q(r)) − √ 2Q obs < ∼ e 2 /(3π), the condition (266) is no longer fulfilled and solution given by Equation (270) loses its meaning. Now, let r pole < r 0 < r ap pole . To determine Q(r) in immediate vicinity of the point r pole (at r 0 approaching r pole ) we, as before, assume that ε(r, Q(r)) is a smooth function of coordinates but now Q(r) Q 1 (r). Above we have found the pole solution of the relativistic Thomas-Fermi equation for ε(r, Q(r)) = 1, cf. Equation (200) and [3]. Now with ε(r, Q(r)) const < 1 assuming Q(r) Q 1 (r) we similarly get [54], The valuer m is now irrelevant, because solution (267) is modified due to inclusion of the electron condensation. Solution (271) with η = 1 is valid for ε(r, Q(r)) e 2 /(3π). At very short distances from r pole , at which 0 < ε(r, Q(r)) < ∼ e 2 /(3π), the condition that ε(r, Q(r)) varies smoothly with r is violated. In this region, we may present ε(r, Q(r)) a(r − r pole ), for a = const and then solution (271) continues to be valid, but now for η = 1/8.
Finally, we stress that solution (271) corresponds to the charge distribution near the bare charge Z 0 for r 0 > r pole . It looses the meaning for r 0 < r pole . At fixed Z obs for r > ∼ 1/m, the charge Z 0 (r 0 ) that is related to this Z obs is increased with decreasing r 0 . Even for Z obs 1/e 2 , at tiny distances, r ∼ r pole , the charge Q(r) becomes very large, Q(r) 1, and at these distances the electron condensation on levels of the upper continuum crossed the boundary < −m comes into play. Our solution does not exist for r 0 < r pole , r pole = r L , where ε(r L ) = 0. The value of r pole essentially depends on the value of Z obs . For Q obs > ∼ 1 the value r pole increases considerably, see Figure 6 and Equation (201), being derived for ε 1. Because QED is the theory with a local interaction, the charge sources can be of arbitrary sizes, including r 0 → 0. To attack the zero-charge problem, let us reconsider the interpretation of the electron condensation in the field of the charged source of a very small size.
Because the Dirac equation in the spherically symmetric field does not change under simultaneous replacements → − and e → −e, i.e., V → −V and κ → −κ, in the Coulomb field of a negative charge Z 0 < 0, there are electron levels (and in the field of a positive charge Z 0 > 0, there are positron levels), which originate in the lower continuum. With increasing |Z 0 |, the energy of such level, e , goes up and at a value |Z 0 | > 137 − 170 (depending on r 0 ), the level intersects the boundary of the upper continuum e = m. According to the traditional interpretation, which we have used while considering r 0 > r pole , the electron states with e > −m, which appeared from the lower continuum already in a weak field of repulsion to the electron, should be regarded as unphysical, and they should be reinterpreted as positron states with energies e + = − e . As a consequence of such reinterpretation, for a nucleus with −Z 0 > 1/e 2 , upon decreasing r 0 , the lowest positron level reaches the energy e + = −m. Subsequently, two positrons, after tunnelling from the lower continuum, occupy this empty level and two electrons move to infinity. Similarly, positron states with e + > −m appeared from the lower continuum already in a weak field of attraction to electron (for Z 0 > 0) are regarded as unphysical, being interpreted as electron states with energies e = − e + . As we have demonstrated, such an interpretation allows for solving the problem of the charge distribution only for r > r 0 > r pole , even while taking such multiparticle effects into account, such as the polarization of the vacuum and (for Z > 0) the electron condensation on levels of the upper continuum crossed the boundary e = −m.
However, beyond the framework of a single-particle problem, there appears to be a possibility of another interpretation [54,55]. Following this possibility, we may interpret the electron levels that originated in the lower continuum in the weak repulsive field (for Z 0 < 0), as levels have been occupied by electrons of the lower continuum, while taking into account that dielectric permittivity ε(r) can be negative at small distances. Subsequently, no preliminary tunneling occurs from one continuum to another. Near the positively charged center of radius r 0 < r pole , the desired repulsive potential for the electrons appears, since the dielectric permittivity of the vacuum expressed in terms of the physical charge e 2 > 0 becomes negative at r < r pole . In terms of a not renormalized charge ε n.ren (r → r 0 → 0) → 1 but e 2 0 < 0 leading to the same result, V(r) > 0, cf. (244). Passage of the pole with decreasing r becomes possible because of the phenomenon of electron condensation on levels originated in the lower continuum even in a weak field.
Above, dealing with the electron condensation on levels of the upper continuum, due to presence of the pole, we could not get a continues solution for all r. Now, dealing with ε < 0 at r → r 0 → 0, we are able to find an appropriate solution connecting Q(r > r 0 → 0) and Q obs = Q(r → ∞).
For ε < 0 and Z 0 > 0, the resulting potential V proves to be repulsive. Thus, for a positively charged center, due to change of the sign of ε there are electron levels coming from the lower continuum. Since the quantity |Z 0 /ε(r)| increases with increasing r, in a certain range of r, where −Q(r) > 1, in the bare potential there are many such levels.
Using explicit expression (257) with approximately constant value ν, and integrating further, we find Q 2 (r) = C 2 Choosing an appropriate sign of the solution corresponding to the repulsive potential for the electron due to ε < 0 for r < r pole , we arrive at For r → r 0 → 0, for any finite value of Q 0 > 0 we obtain Q(r) −Q 0 /|ε| → 0. Thus, a test particle does not interact with the nucleus at ultrashort distances. Recall the asymptotic freedom property in the QCD for r → 0. For r ∼ 1/m, we have ε 1 and Q(r) = Z obs e 2 . Thus, we obtain a relation between the bare and observed charges Z obs = −Z 0 /(1 + 2(Z 0 e 2 ) 2 /ν) 1/2 .
It is important that, at distances r r pole , the potential looks like an ordinary Coulomb potential. Individual charges situated at these distances, each with Z obs 1/e 2 , can be summed up to the total charge Z > Z cr ∼ 1/e 2 . At these distances ε > 0 and it is close to unity for r r pole , and there may appear the electron condensation on the levels in the upper continuum crossed the boundary = −m. These levels become occupied by electrons, after the tunneling from the lower continuum, as we have demonstrated in Section 7. Thus reconstruction of the interaction at r < r pole does not affect any phenomena that can be observed experimentally occurring at much larger distances.
Note that solution (276) is similar to the solution obtained within QCD in the model [102], which took a possibility of the quark condensation near the external color-charge source into account. The essential difference is in the dependencies of ε(r) in QCD and in QED. In QCD within a logarithmic approximation ε QCD (r) b 0 ln(r 2 Λ /r 2 ) where b 0 and r Λ are some positive constants, i.e., ε QCD (r → 0) → ∞ and ε QCD (r → ∞) → −∞, whereas, within QED, we employed that ε QED (r → 0) → −∞ and ε QED (r → 1/m) → 1. In QCD, there appears to be condensation of quarks on levels that originate in the upper continuum and in the case under consideration in QED for r 0 → 0, we included the electron condensation on levels that originate in the lower continuum.
Additionally, recall that the Hamiltonian, where one replaced p µ → p µ − e 0 A n.ren µ should be Hermitian operator, as well as the same Hamiltonian that is expressed in terms of the renormalized charge, where one uses the replacement p µ → p µ − eA ren µ . Within the ordinary second quantization scheme, one expands µ in series of plane waves, where the creation and annihilation operators appear, considering A µ as the real quantity. Because e 0 is imaginary, A n.ren µ should be considered as purely imaginary quantity. Now, we should perform expansion for e 0 A n.ren µ , being real quantity. The energy is reduced to the energy of stable oscillators only after performing renormalization, i.e., being expressed in terms of eA ren µ .
Distribution of Charge of Electron
Up to now, we considered the charge distribution near the external charge source, which was assumed to be infinitely massive. For description of the electron mass distribution, m(r), one needs to study Dyson equation for the electron Green function, cf. [103]. At distances of our interest |V| m(r) and the dependence of m(r) does not influence the charge distribution in the logarithmic approximation that we have used. Equation for the mass is given by [114], where d t is the so called d-function of the photon and Ξ = ln(1/(r 2 m 2 )) is the tortoise coordinate introduced above. A clarification is in order. As is known, the presence of a zero in the expression for the dielectric permittivityε(Ξ) defined via the photon d-function, e 2 0 d t = e 2 (Ξ) ,ε(Ξ) = e 2 /e 2 (Ξ) , according to the Källen-Lehmann expansion, would correspond either to the violation of the causality or to the instability of the vacuum [115]. However, note that, in our case, the quantity˜ (Ξ) does not have zero, ε(Ξ) = e 2 e 2 (Ξ) = (ε 2 (Ξ) + 2e 4 /ν) 1/2 , as follows from (275) for Q 0 = Z 0 e 2 , Z 0 = 1. Thus, the quantity˜ (Ξ) does not coincide with ε(Ξ). The latter quantity may vanish and it can even be negative, whereas the "true" valueε(Ξ) > 0. Integrating (279), we obtain [54], m(Ξ) = m ε(Ξ) + (ε 2 (Ξ) + 2e 4 /ν) 1/2 1 + (1 + 2e 4 /ν) 1/2 where m is the observed electron mass. Thus m(Ξ → ∞) → 0 and m(Ξ → 1) → m, i.e., in this case, the entire electron mass is of purely electromagnetic origin. Concluding, we presented some arguments for the logical consistency of QED.
Conclusions
Most actively, the problem of a spontaneous production of positrons from the QED vacuum in strong fields has been attacked in theoretical works in Moscow (in the group of V. S. Popov in 1970s) and in Frankfurt (in the group of W. Greiner in 70s and 80s of the previous century). The experiments performed at GSI Darmstadt in 1980s had turned out puzzling line structures in the energy spectra. These results were not confirmed by the subsequent experiments performed in the 1990s. Questions regarding the experimental confirmation of existence of the spontaneous positron production in low-energy heavy-ion collisions remained open. Now, interest in this problem is renewed [34], in connection with the possibility to perform new experiments at the upcoming accelerator facilities in Germany, Russia, and China [29][30][31]. The study of many-particle effects in description of the QED vacuum in strong fields is of of principal interest. The problem of the zero-charge remains one of the most important fundamental problems of QED already about 70 years. In the given paper, these problems were studied within a common relativistic semiclassical approach that was developed in the reviewed papers.
In the given paper, first, the problems of the falling to the Coulomb center for the charged spinless boson and for the fermion were considered within the single-particle picture. Subsequently, focus was concentrated on a case of the spontaneous positron production in the field of a finite supercritical nucleus with the charge Z > Z cr (170 − 173). The behavior of deep electron levels that crossed the boundary of the lower continuum and the probability of the spontaneous positron production were studied. Subsequently, similar effects were considered in application to the low-energy collisions of heavy ions, when, for a short time, the electron level of the quasi-molecule crosses the boundary of the lower continuum = −m. Next, the phenomenon of the electron condensation on levels of the upper continuum crossed the boundary of the lower continuum in the field of a supercharged nucleus with Z Z cr was studied. Subsequently, focus was concentrated on many-particle problems of the polarization of the QED vacuum and the electron condensation at ultra-short distances from the source of the charge. Arguments were presented for the important difference of the cases, when the size of the source is larger than the pole size r pole = r L , at which the dielectric permittivity of the vacuum reaches zero, and smaller r pole . Subsequently, distributions of the charge and mass of the electron were considered and arguments were given in favor of the logical consistency of QED. Additionally, I believe that at least some of the results reviewed in this paper can find applications in the description of semi-metals and stack of layers of graphene. | 33,616.8 | 2021-02-14T00:00:00.000 | [
"Physics"
] |
Stress in Metastatic Breast Cancer: To the Bone and Beyond
Simple Summary Breast cancer is the most common cancer affecting women of all ages worldwide. In spite of the encouraging advances made in early diagnosis, 10% of breast cancer patients are still affected with metastatic breast cancer at the time of diagnosis. The available therapeutic options are predominantly palliative, and thus this unfavourable prognosis is associated with a low survival rate. Intriguingly, stress has been shown to promote the growth of breast tumours and the incidence of metastasis. Herein, we describe the contribution of the sympathetic hyperactivation induced by stress to the progression of breast cancer and its dissemination to distant organs, specifically to the bone, but also to the lung, liver and brain. The putative sympathetic adrenergic signalling mechanisms responsible for this modulation are also summarised. The knowledge gathered highlights the therapeutic potential of targeting sympathetic signalling to tackle cancer progression and metastasis. Abstract Breast cancer (BRCA) remains as one the most prevalent cancers diagnosed in industrialised countries. Although the overall survival rate is high, the dissemination of BRCA cells to distant organs correlates with a significantly poor prognosis. This is due to the fact that there are no efficient therapeutic strategies designed to overcome the progression of the metastasis. Over the past decade, critical associations between stress and the prevalence of BRCA metastases were uncovered. Chronic stress and the concomitant sympathetic hyperactivation have been shown to accelerate the progression of the disease and the metastases incidence, specifically to the bone. In this review, we provide a summary of the sympathetic profile on BRCA. Additionally, the current knowledge regarding the sympathetic hyperactivity, and the underlying adrenergic signalling pathways, involved on the development of BRCA metastasis to distant organs (i.e., bone, lung, liver and brain) will be revealed. Since bone is a preferential target site for BRCA metastases, greater emphasis will be given to the contribution of α2- and β-adrenergic signalling in BRCA bone tropism and the occurrence of osteolytic lesions.
Introduction
Breast cancer (BRCA) is the most frequently diagnosed cancer in women worldwide and is expected to represent around 25% of all new cancer cases diagnosed in females. In 2020, more than 2 million people were diagnosed with BRCA, with Europe accounting for nearly 24% of all new cancer cases [1,2]. Despite the large number of new cases, the mortality rate has been slowly decreasing in developed countries with the implementation of earlier diagnosis and the improvement of adjuvant therapies [3][4][5][6]. Nevertheless, metastatic BRCA still affects 6-10% of women at the time of diagnosis and presents a 5-year survival rate of 27% [6][7][8].
The development of metastasis and its prediction can be dictated by specific risk factors, such as grade, nodal involvement and size of the tumour. However, these factors do not predict the specific sites or patterns of metastasis, characteristic of BRCA tumours. Interestingly, it has been hypothesised that the primary tumour can provide insight about the organ that BRCA-disseminating cells eventually home to, partaking in the possibility to influence the therapeutic and survey strategies for each patient since the time of primary diagnosis. Although BRCA subtypes present a known organotropism [9], this process remains still largely unexplained.
In addition, patients diagnosed with BRCA, undergoing surgery or therapy, are at a higher risk of feeling emotional stress [10][11][12][13][14]. These symptoms may lead up to a psychiatric disorder, such as anxiety or depression, and can develop several years after the diagnosis of the disease. The link between cancer and emotional disorders has been suggested since the ancient times, however the nature of this association has only started to be revealed during the last two decades [14][15][16][17][18][19]. Although a few studies demonstrate an association between stress-induced sympathetic hyperactivation and the incidence of cancer and dissemination [14], there seems to be a stronger and more consistent relationship between psychological factors and the progression of already-existing tumours [20][21][22][23][24][25]. Nevertheless, greater emphasis has been granted to the use of drugs targeting stress-induced signalling pathways in cancer initiation and progression [26], and more importantly in metastatic BRCA [27].
Fortunately, the nature of this association between stress and cancer progression is being gradually uncovered in order to better understand how chronic stress and the sympathetic tone influences the metastatic cascade and how it can affect the destination of BRCA cells. In this review, BRCA heterogeneity will be discussed, focusing on its effects on metastatic site predisposition. Likewise, the consequences of a sympathetic hyperactivation, owing to chronically stressed conditions, on BRCA metastasis will also be reviewed.
BRCA Molecular Subtypes
The heterogeneity and complexity of BRCA has long been noted by accessing histologic samples and patients' outcomes [28,29]. The development of molecular profiling techniques has further ensured this heterogeneity, and it is now possible to classify BRCA within, at least, three main subtypes: luminal, HER2-enriched (HER2 + ) and triple-negative breast cancer (TNBC). As detailed in Figure 1, each molecular subtype has different biological characteristics, including risk factors, prognosis, response to therapies and a preferential metastatic site [30][31][32][33][34]. Luminal tumours express receptors for oestrogen (ER) and progesterone (PR) hormones, and are divided into two subtypes: luminal A and luminal B [35]. The other subtypes represent hormonal-negative tumours and usually portray a worse prognosis than the luminal subtypes. While the HER2-enriched subtype illustrates the tumours that have a high expression of the HER2 gene and other genes related to its pathway [36], the TNBC subtype, often referred to as basal-like tumours, mimics the expression profile of the myoepithelial cells and usually lacks the expression of both hormonal receptors (HR) and HER2. Moreover, not all TNBC are basal-like tumours [37,38] and a new subclass of TNBC lacking cell-cell adhesion and tight junction's markers (e.g., claudins) has emerged, denominated the claudin-low subtype [39][40][41]. Overall, the novel information provided by molecular profiling techniques has allowed to better understand each breast tumour subtype, generating new treatment approaches to be used as individualised therapies, which will be translated in the improvement of BRCA patient outcomes.
Patterns of Metastatic BRCA
In the end, metastasis is a disorganised multifactorial process where the ability for a primary tumour to metastasize to a specific organ depends on a variety of factors, including the cancer cell type, the primary organ and the microenvironment of the secondary site [42]. The intrinsic characteristics of cancer cells and the cellular and cytokine profile of the tissue of origin dictate how these cells will migrate, survive and proliferate. The tissue microenvironment to which metastatic cells eventually home also plays a significant role in this process. Most importantly, the interaction between the primary organ and the secondary site commands the success of metastasis [43,44].
Organ-specific metastasis was firstly described by Paget in 1889 in the "seed and soil" hypothesis, where he, after evaluating BRCA patient autopsies, stated that "in cancer of the breast, the bones suffer in a special way" [45]. The author proposed that these patterns were due to the seed dependency (cancer cell) for the soil (environment factors in the new organ), suggesting, therefore, that the distribution of metastatic sites is not a random act.
In general, BRCA cells commonly metastasize to bone, lungs, liver and brain [46]. Interestingly, the majority of studies showed that the luminal subtypes, tumours with positive ER and PR expression, metastasize preferentially to the bone, taking a longer time to relapse. On the other hand, tumours with negative HR expression, such as TNBC subtypes, present a smaller tropism to bone and are usually present in the brain and lungs, and rapidly recur. The HER2 + subtype mostly metastasizes to visceral organs ( Figure 1) [9,47,48]. The median overall survival of metastatic BRCA patients ranges from approximately 1 year for metastatic TNBC, to 5 years for HR and HER2 + BRCA subtypes [49]. Thus, each BRCA subtype not only displays specific primary tumour characteristics, such as aggressiveness and response to treatments, but can also exhibit different metastatic behaviour. This knowledge is very important since it can help in the development of follow-up and surveillance strategies for newly diagnosed patients, allowing for different options of adjuvant therapies. However, the characteristics and the mechanisms that determine the location of BRCA metastatic spreading are still largely unknown, and thus more research is required in this field.
Stress and the Activation of the SNS
Behavioural stress has been pointed out as an accelerator of cancer progression. Indeed, stress is a complex process where both psychosocial and environmental factors trigger a cascade of information-processing pathways in the central and peripheral nervous system [50]. Both the hypothalamic-pituitary-adrenal (HPA) axis and the SNS systems have been implicated in cancer, supported by an increasing body of studies linking the "fight-or-flight" stress response of SNS mediators with cancer progression [21,22,[51][52][53]. For this reason, the investigation on the role played by the SNS on cancer biology has been largely encouraged.
By controlling involuntary body functions, SNS virtually regulates all human organs. The activation of SNS leads to the release of catecholamines, such as norepinephrine (NE) and epinephrine (E), which regulate these functions through two possible pathways. On one hand, there is the localised release of NE from the sympathetic nervous terminals that directly innervate the target organs, along with the co-release of sympathetic nonadrenergic neurotransmitters (e.g., neuropeptide-Y and ATP), whereas the other pathway mainly involves the systemic release of NE and E (in a proportion of 20:80, respectively) by the adrenal glands to the circulation [54]. It is noteworthy that growing evidence suggests that the local release of NE from SNS nerve terminals is the dominant driving force in the sympathetic control of cancer progression [55][56][57].
During the SNS acute "fight-or-flight" responses, E and NE levels can increase by 10 times. This causes rapid physiological changes in respiratory, cardiovascular, muscular, immune and neural systems, increasing the blood flow to muscles and lungs, preparing the body for alert situations. The catecholamines levels return to baseline in a very short amount of time (20-60 min), and therefore the activation of the acute stress responses is considered adaptive [53]. On the other hand, in chronic stress conditions, the physiological systems are exposed to glucocorticoids and catecholamines for long periods of time. This increased exposure leads to a deterioration of health conditions, such as increased risk of infections and cardiac diseases, decreased wound healing and eventually death [59]. Importantly, in the case of cancer, catecholamines have the potential to induce a panoply of physiologic effects, both deleterious and beneficial, termed the cancer catecholamine conundrum. These variable effects were described to be dependent on several factors, such as: (i) catecholamine concentration in the blood, (ii) exposure time, (iii) physical activity, (iv) activation of nine different ADR and (v) the duration and recurrence of the stress, among others [60]. Overall, a correlation between stress and tumorigenesis/cancer progression endures in different types of tumours, namely BRCA [61].
Effects of Stress in Primary BRCA
Since it has been reported that the stress response is partly mediated by the activation of ADRs, several studies have confirmed that BRCA cancer cells and BRCA tissues generally express αor β-ARs [62][63][64]. The ADRs expression and its associated correlations with BRCA are summarised in Table 1. Abbreviations: ER-oestrogen receptor; HER2-human epidermal growth factor 2; PR-progesterone receptor.
The most described α-ADR in BRCA is the α2-ADR subtype, which was shown to cause increased proliferation of BRCA cell lines [69,70], tumour growth [70][71][72], metastasis [73][74][75] and chemoresistance of cancer cells [76]. Interestingly, not only can E and NE act directly on tumour-express α2-ADR, but SNS activation can also influence tumour growth and metastasis through α2-ADR present on stromal cells of the tumour microenvironment, even when specific BRCA cell lines did not express functional α-ADRs [74].
The prognostic significance of αand β-ADRs expression in BRCA tumours has also been explored, suggesting a possible role for targeted therapy using ADR antagonists. While α-ADR has been described to be expressed in poor-prognosis tumours, β2-ADRexpressing tumours are mostly associated with good prognosis [65][66][67]. Furthermore, a retrospective study observed that HER2 + tumours expressing β2-ADR presented a significantly lower disease-free survival rate and lymph node metastasis incidence [62]. Interestingly, besides the putative effect of ADR expression on BRCA prognosis, the ADR expression profile seems to be dependent on the BRCA subtype. For instance, luminallike tumours strongly express β2-ADR, while α-ADRs, such as α1Band α2C-ADR, are overexpressed in basal-like breast tumours [65].
It is important to note that ADR expression in BRCA is still not consistent. Besides different expression profiles between tissues and cell lines, the same cell line in different studies may not express the same ADRs, and this is probably due to the use of distinct quantification methodologies. For instance, MCF-7, a luminal A BRCA cell line, has been shown to express β2-ADR in some studies at the mRNA level, while in another study, β2-ADR was suggested to be negatively expressed [69,81]. Additionally, MCF-7 has been shown to exhibit the highest β2-ADR protein expression level among the tested BRCA cell lines [86]. Thus, further investigation is required to ensure correct characterisation of the adrenergic profile of commonly used cell lines and match it with clinical human biopsies. Even though the presence and the effects of ADRs in BRCA remains to be clarified, these studies clearly suggest that SNS activation, through the signalling of ADRs, influences BRCA progression and metastasis. Furthermore, the link between stress and BRCA progression was strengthened when pharmaco-epidemiological studies showed that the use of β-blockers, at the time of diagnosis, was associated with improved survival, decreased tumour invasion, metastasis, recurrence and mortality [95][96][97][98]. These studies present some limitations, such as the limited size of the analysed patient cohort, and the benefits of β-blocker usage on improved survival was not replicated in other epidemiologic studies [99,100]. Additionally, the role of β-blockers (e.g., propranolol), as a neoadjuvant therapy, has been assessed in a small number of clinical trials, yet these studies did not deliver major conclusive results (NCT01847001; NCT02596867). Interestingly, a phase II randomised controlled trial highlighted the benefits of using propranolol in BRCA patients, since only one week of treatment was associated with decreased expression of pro-metastatic biomarkers [101]. Nevertheless, further pre-clinical and clinical studies clarifying the importance of ADRs on different subtypes of BRCA could potentiate the development of novel therapeutic strategies.
Targeting the Bone Microenvironment
BRCA bone metastatic foci are often characterised by the formation of osteolytic lesions, where the interaction between tumour cells and the bone niche leads to the establishment of a vicious cycle of bone destruction and, subsequently, complications such as fractures, hypocalcaemia and severe bone pain [102][103][104]. When BRCA-disseminated cells arrive at the bone microenvironment, they disrupt the intricate cascade of events that regulate bone remodelling to ultimately favour bone resorption. In fact, at the site, BRCA cells can directly activate osteoclasts (the bone-resorbing cells) or act indirectly through osteoblasts (the bone-forming cells) by stimulating osteoblast-derived receptor activator of NF-κB ligand (RANKL), which is a master regulator of osteoclastogenesis [105]. Additionally, BRCA cells can also inhibit the differentiation and adhesion of osteoblasts, increase their apoptosis and delay collagen synthesis, thus impairing osteoblasts' capacity to fully replace the resorbed bone. The over-activation of osteoclast bone resorption will lead to the release of ionised calcium and growth factors entrapped in the bone matrix, which will further stimulate the growth and survival of cancer cells. This self-perpetuating cycle leads to bone loss and tumour growth and is designated as the "osteolytic vicious cycle" of BRCA bone metastasis [106].
Some studies have explored the effects of SNS activation on bone remodelling diseases, where increased incidence of fractures and decreased bone mass were observed [121][122][123][124][125][126]. Osteoclasts express ADRs, and SNS activation has been shown to promote bone loss, by directly affecting bone resorption in mice [127]. Furthermore, β-blockers [128] and α2-ADRs agonists [112] inhibited the mRNA expression of the osteoclast-related genes such as TRAP and cathepsin K, and decreased the number of TRAP-positive multi-nucleated osteoclasts. In humans, NE was described to inhibit osteoclast-precursor cell proliferation on osteoclast-precursor cells and increased osteoclast maturation and TRAP activity [129]. These studies suggest that ADRs are involved in the regulation of osteoclastogenesis by directly affecting osteoclast activity. However, the majority of the studies mainly focus on the role of osteoblasts, suggesting that local NE release and binding to osteoblastic β2-ADR leads to inhibition of bone formation and stimulation of bone resorption, mainly due to augmented RANKL expression [130][131][132].
Several preclinical and epidemiologic studies have shown that β-blockers, drugs that inhibit β-adrenergic signalling and are commonly used to treat hypertension, have been linked with reduced BRCA metastasis and improvement of the patient survival [95][96][97][98][133][134][135][136]. Furthermore, a vast number of in vitro and in vivo studies have explored the presence and role of αand β-ADRs in BRCA, highlighting an association between chronic SNS stimulation and BRCA progression [137]. Adrenergic signalling seems to either directly affect BRCA cells [67,[73][74][75]90,138] or indirectly influence cells in the pre-metastatic niche to facilitate disease progression [89,93,[139][140][141][142][143][144]. Studies performed in other common sites for metastasis also suggest a major indirect effect of SNS in stromal cells, facilitating the migration and colonisation of cancer cells in these organs during chronically stressed states [89,140,142,144].
During the last decade, multiple studies that focused on BRCA showed that SNS activation can modulate BRCA bone metastasis. Besides the direct stimulation of BRCA ADRs, some studies have shown that SNS can also act on bone marrow stromal cells to indirectly promote the colonisation of bone by metastatic BRCA cells. SNS activation, via β-AR signalling in osteoblasts, induced RANKL production, which promoted BRCA metastasis to bone via its pro-migratory effect on RANK-expressing BRCA cells. Moreover, blocking SNS activation with a β-blocker inhibited the stimulatory effect of sympathetic activation on bone metastasis [139]. The activation of β2-AR expressed in osteoblasts also seems to play a crucial role in mediating the SNS signals in bone, through the production of vascular endothelial growth factor (VEGF) [143]. Moreover, β2-ADR stimulation in osteoblasts triggers the release of soluble factors, such as IL-1β, that favour BRCA cell engraftment within the skeleton, by the upregulation of E-and P-selectin expression by endothelial cells [141]. Overall, the role that these cytokines play in bone metastasis suggests that when chronic stress activates the SNS, the bone marrow microenvironment is transformed into a more favourable tissue for the establishment of metastasis [145].
Other Metastatic Sites: Beyond the Bone
Even though the skeleton is the preferential site for metastasis in a vast majority of BRCA tumours, other organs such as lungs, brain, lymphatic nodes and liver can also be targeted and colonised by metastatic BRCA cells. Although there is a scarce number of studies exploring the role of SNS in these organs, the influence of chronic stress is beginning to be noticed, with special interest in the β2-adrenergic signalling pathway.
For instance, β-adrenergic signalling activation was shown to promote lung metastatic colonisation by BRCA cells. Treatment with the β-adrenergic antagonist propranolol supressed the stress-induced metastasis, while stimulation of β-adrenergic signalling with isoproterenol (ISO) increased the number of monocytes and infiltration of macrophages into the pre-metastatic lung. Thus, under chronic stress, β-ADR stimulation promotes metastatic homing and seeding of circulating BRCA cells through remodelling of the pre-metastatic lung microenvironment [140]. Previously, Shakhar et al. demonstrated that the systemic β-adrenergic stimulation in rats suppressed the activity of NK cells and caused an increase in tumour cell retention in the lungs, that was later translated in increased lung metastases. Importantly, the use of a β-adrenergic antagonist reversed these effects [144]. Additionally, clinical trials have shown that propranolol combined with etodolac, a COX-2 inhibitor, decreased the risk of lung metastasis. This combination also enhanced tumour clearance from the lungs and improved long-term recurrence-free survival rates, showing its potential in limiting post-operative immunosuppression and metastatic progression [146]. However, other studies have demonstrated that treatment with a β2-ADR agonist decreased the number and size of BRCA lung metastases [94]. Despite the controversial results, these studies highlight the role of β-ADRs in lung metastasis development. Finally, although there are only a few studies addressing the role of α-ADR signalling on lung metastasis, it was previously demonstrated that treatment with dexmedetomidine, an α2-ADR agonist, leads to increased tumour-cell retention and growth of BRCA lung metastases in vivo [74,75].
Besides the action of SNS on stromal and immune cellular components of the premetastatic niche, another study demonstrated that an augmented tumour invasiveness and metastasis in lung and lymph nodes was caused by neuroendocrine signalling directly affecting BRCA cells [89]. Additionally, using the RNA interference methodology, Chang et al. showed that β2-ADR knockdown in BRCA cells reduced tumour cell invasion in vitro and significantly reduced the impact of stress on lung and lymphatic metastasis in vivo [90].
Interestingly, chronic stress was found to be a pathophysiological regulator of lymphatic remodelling in cancer, facilitating BRCA dissemination to nearby lymph nodes. Through tumour-derived VEGFC and macrophage-derived COX-2, chronic stress restructured lymphatic networks within and around tumours, providing pathways for tumour cell escape. Inhibition of SNS signalling using propranolol blocked chronic stress effects and reduced metastasis to lymph nodes. Conversely, SNS activation with ISO was sufficient to increase lymph node metastasis. Furthermore, in agreement with the pre-clinical evidence discussed above, β-blocker use reduced the risk of lymph node and distant metastasis in a group of BRCA patients [142].
Among the different subtypes of BRCA, the prevalence of brain metastasis in TNBC is the highest [147]. The SNS can also modulate brain metastasis progression since a retrospective study showed that perioperative β-blockade was associated with decreased cancer recurrence in stage II BRCA patients [138]. Moreover, TNBC metastatic cells presented increased β2-ADR mRNA and protein expression levels relative to TNBC from the primary tumour. When a β2-ADR agonist was used to mimic stress conditions in vitro, TNBC brain-metastatic cells exhibited increased cell proliferation and migration, which was abrogated by propranolol. In addition, propranolol also decreased the establishment of brain metastases in vivo [138].
Liver is also another common metastatic site in BRCA, and even though the effects of chronic stress in BRCA liver metastasis have not been assessed, SNS activation was previously shown to influence tumour dissemination to the liver in other cancer types, i.e., colon cancer [148]. In a study with socially isolated mice, an enhancement of liver metastasis from a colon carcinoma cell line was observed. Chronic stress was used as a model of endogenous SNS activation characterised by elevated levels of catecholamines, NE and E. These stressed mice developed metastatic foci at earlier time points, presented a decreased survival time and displayed worse chemotherapeutic responses than control mice. Interestingly, β-blocker treatment reversed these effects [149].
Chronic stress was also associated with increased infiltration of tumour-associated macrophages into the primary tumour and increased the expression of metastatic genes [148]. In addition, propranolol inhibited proliferation and induced apoptosis in liver cancer cells, confirming that the SNS activation may have a role in liver cancer, particularly through the β-adrenergic signalling pathway [150]. Since similar mechanisms might be in play in BRCA, further studies exploring the effects of SNS activation in BRCA liver metastasis are warranted.
Concluding Remarks and Future Perspectives
Several reports support the hypothesis that chronic SNS activation plays a critical role in the establishment of metastatic BRCA, specifically in the bone microenvironment. The information gathered so far on this association is described in Figure 2. Nevertheless, the results obtained so far are scarce, in some cases contradictory, and the mechanisms remain poorly understood. Thus, more studies need to be performed, particularly, a better characterisation of the adrenergic profile of the different BRCA subtypes, and its influence in metastasis of chronically stressed patients, would be of great importance. Additionally, the role of other adrenergic signalling pathways (besides β-adrenergic) should also be clarified. Figure 2. A brief summary of the established correlations between stress, BRCA and metastases. In a stress situation, catecholamines (NE and E) are released by the sympathetic fibres and sympathoadrenal system. These catecholamines stimulate the activation of adrenergic receptors, locally in BRCA primary tissue and also by affecting the BRCA metastatic organs. Abbreviations: ADR-adrenergic receptor; BRCA-breast cancer; E-epinephrine; HER2-human epidermal growth factor 2; NEnorepinephrine; RANKL-receptor activator of NF-κB ligand; TNBC-triple-negative breast cancer; VEGF-vascular endothelial growth factor.
It is now clear that unveiling the mechanisms behind the correlations between stress, BRCA and metastasis may allow for the development of specific therapies for metastatic BRCA. The use of αand β-blockers in BRCA animal models suggests a role for pharmacologic inhibition of ADRs signalling in helping to control the metastatic progression [73,118,139]. Moreover, since cancer diagnosis, surgery and associated treatment is a highly stressful experience, potentially worsening the progression of the disease, the general use of ADR-blockers systemically (specifically β-blockers) as an adjuvant therapy has the ability to improve the effectiveness of cancer treatment [26,151]. In addition, the use of alternative non-pharmacological approaches has also been suggested to reduce the sympathetic activity via the modulation of local neuronal activity and improve BRCA patients' survival outcomes [151]. Overall, further investigation will be needed to determine the best time for stress-management interventions [14] and to further analyse the clinical benefits of such approaches. | 5,812.8 | 2022-04-01T00:00:00.000 | [
"Biology",
"Medicine"
] |
Sentiment Analysis and Text Analysis of the Public Discourse on Twitter about COVID-19 and MPox
Mining and analysis of the big data of Twitter conversations have been of significant interest to the scientific community in the fields of healthcare, epidemiology, big data, data science, computer science, and their related areas, as can be seen from several works in the last few years that focused on sentiment analysis and other forms of text analysis of tweets related to Ebola, E-Coli, Dengue, Human Papillomavirus, Middle East Respiratory Syndrome, Measles, Zika virus, H1N1, influenza like illness, swine flu, flu, Cholera, Listeriosis, cancer, Liver Disease, Inflammatory Bowel Disease, kidney disease, lupus, Parkinsons, Diphtheria, and West Nile virus. The recent outbreaks of COVID-19 and MPox have served as catalysts for Twitter usage related to seeking and sharing information, views, opinions, and sentiments involving both of these viruses. None of the prior works in this field analyzed tweets focusing on both COVID-19 and MPox simultaneously. To address this research gap, a total of 61,862 tweets that focused on MPox and COVID-19 simultaneously, posted between 7 May 2022 and 3 March 2023, were studied. The findings and contributions of this study are manifold. First, the results of sentiment analysis using the VADER approach show that nearly half the tweets had a negative sentiment. It was followed by tweets that had a positive sentiment and tweets that had a neutral sentiment, respectively. Second, this paper presents the top 50 hashtags used in these tweets. Third, it presents the top 100 most frequently used words in these tweets after performing tokenization, removal of stopwords, and word frequency analysis. Finally, a comprehensive comparative study that compares the contributions of this paper with 49 prior works in this field is presented to further uphold the relevance and novelty of this work.
The recent outbreaks of COVID-19 and MPox have served as "catalysts", leading to the usage of Twitter for the sharing and exchanging information on diverse topics related to these viruses, leading to the generation of tremendous amounts of big data.No prior work in this field has focused on studying and analyzing tweets that focused on both of these viruses simultaneously to understand and interpret the underlying paradigms of conversations.Therefore, this serves as the main motivation for this work.
In December 2019, there was an outbreak of an unknown respiratory disease in a seafood market in Wuhan, China.This outbreak affected about 66% of the people in the market.A prompt investigation from the healthcare and medical sectors revealed that a novel coronavirus was responsible for this disease, and this virus was named severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2, 2019-nCoV), as it was observed to have a high homology of about 80% with SARS-CoV [61].The disease that humans suffer from after getting infected by this virus is known as COVID-19 [62].Despite the best efforts of the Chinese Government to contain the spread of this virus, it soon spread to other parts of the world while undergoing multiple mutations, and several variants, such as Alpha (B.1.1.[63] led to an increase in COVID-19 cases.At present, there have been a total of 681,518,412 cases and 6,811,869 deaths on account of COVID-19 on a global scale [64].Respiratory systems are the primary target of the SARS-CoV-2 virus, although infections in other organs of the body have been reported in some cases.The symptoms of COVID-19 usually include fever, dry cough, dyspnea, headache, dizziness, exhaustion, vomiting, and diarrhea [65].However, studies have shown that symptoms can vary from person to person based on user diversity, such as age group, preexisting conditions, disabilities, etc. [66,67]. MPox (monkeypox) is a re-emerging zoonotic disease.It is caused by the MPox (monkeypox) virus, which belongs to the Poxviridae family, the Chordopoxvirinae subfamily, and the Orthopoxvirus genus [68].This virus was originally identified in monkeys in 1958 [69], and the first case of this virus in humans was recorded in 1970.The MPox virus is closely related to the variola virus and causes a smallpox-like disease in humans.The common symptoms of MPox include fever, headache, and myalgia.A distinguishing feature of MPox is the presence of swelling at the maxillary, cervical, or inguinal lymph nodes [70,71].The MPox virus was endemic in the Democratic Republic of the Congo (DRC) and a few African countries for a long time, and a few cases outside these geographic regions were recorded only twice-first in 2003 [72] and then in 2017-2018 [73,74].However, since May 2022, the world has been experiencing an outbreak of the MPox virus.At present, there have been a total of 86,231 cases of MPox, with 84,858 of these cases being recorded in regions that have not historically reported MPox [75].
In the context of recent works related to Twitter data mining and analysis, a number of works have focused on the sentiment analysis of tweets.Sentiment analysis [76] is the computational analysis of people's attitudes, views, and sentiments regarding an entity, which may represent an individual, concept, topic, event, or scenario.Sentiment analysis can be considered a classification process.The three primary classification levels in sentiment analysis are the document level, sentence level, and aspect level.The goal of document-level sentiment analysis is to categorize an opinion document as expressing a positive or negative sentiment.The entire document is viewed as a single fundamental informational unit in this process.Sentence-level sentiment analysis seeks to categorize the sentiment that each sentence expresses.In order to categorize the sentiment in relation to particular features of entities, aspect-level sentiment analysis is used.While there have been prior works in this field, however, those works focused on performing sentiment analysis of tweets about COVID-19 or MPox and did not perform sentiment analysis of tweets that focused on both of these viruses simultaneously.The outbreak of MPox during the ongoing outbreak of COVID-19 has resulted in several tweets involving the views, opinions, concerns, and perspectives of the public regarding both of these viruses.A few examples of such tweets (obtained by using the Advanced Search feature of Twitter) are shown in Table 1.As can be seen from these tweets, these two ongoing virus outbreaks prompted the sharing and exchange of views, information, concerns, and perspectives on a wide range of topics (related and unrelated to these viruses) that reflect sentiments of varying polarities related to those topics.No prior work in this field thus far has focused on studying and analyzing tweets that involved conversations about both COVID-19 and MPox.This work aims to address this research gap in this field.The work in this paper involved performing sentiment analysis and text analysis on 61,862 tweets that focused on MPox and COVID-19 simultaneously, posted between 7 May 2022 and 3 March 2023.The findings and contributions of this paper are summarized as follows:
•
The results of sentiment analysis using the VADER (Valence Aware Dictionary for sEntiment Reasoning) approach shows that nearly half the tweets (46.88%) had a negative sentiment.It was followed by tweets that had a positive sentiment (31.97%) and tweets that had a neutral sentiment (21.14%), respectively.
•
Using concepts of text analysis, the top 50 hashtags associated with these tweets were obtained.These hashtags are presented in this paper.
•
The top 100 most frequently used words that featured in these tweets were obtained after performing tokenization, removal of stopwords, and word frequency analysis of these tweets.The findings show that some of the commonly used words involved Twitter users directly referring to either or both of these viruses.In addition to this, the presence of words such as "Polio", "Biden", "Ukraine", "HIV", "climate", and "Ebola" in the list of the top 100 most frequent words indicates that topics of conversations on Twitter in the context of COVID-19 and MPox also included a high level of interest related to other viruses, President Biden, and Ukraine.Another lockdown is incoming.They are trying to make monkey pox look like a pandemic.Their media tools are ready, their vaccines were ready before the pox was introduced.These were the same people that played the COVID19 play.They just changed the name of the movie.Failure!Tweet #8 Monkey Pox new Covid.Election is coming.Coincidence?No Tweet #9 First it was maga.Then there came covid.Now, it's Monkey Pox.When will these horrors end?!? Tweet #10 No longer scared of disease be it Covid or Monkey pox; I'm scared of loosing more years of my life ... 1 These Tweets are presented here in "as is" form after obtaining the same from the Advanced Search feature of Twitter.These Tweets do not represent or reflect the views or opinions or beliefs or political stance of the author of this paper.
In addition to the above, a comprehensive comparative study that compares the contributions of this paper with 49 prior works in this field to uphold its relevance and novelty is also presented in this paper.This paper is organized as follows.In Section 2, an overview of recent works related to sentiment analysis and text analysis of tweets about COVID-19 and MPox is presented.Section 3 discusses the detailed methodology and the specific steps that were followed in this work.In Section 4, the results are presented.Section 5 concludes the paper and outlines the scope for future work in this field.It is followed by references.
Literature Review
This section is divided into two parts.Section 2.1 presents a review of recent works in this field that focused on sentiment analysis and text analysis of tweets about COVID-19.Section 2.2 presents a review of the recent works in this field that focused on sentiment analysis and text analysis of tweets about MPox.
Recent Works that Focused on Sentiment Analysis and Text Analysis of Tweets about COVID-19
The study by Vijay et al. [77] examined the impact of COVID-19-related tweets from November 2019 to May 2020 in India.Three categories were created for all tweets (positive, negative, and neutral).To assess how people would react to the COVID-19 lockdown in June 2020, the authors also created many datasets, which were organized by state and month and pooled across all states.The findings showed that most individuals started off tweeting negatively, but as time went on, more and more people began to post positively and neutrally.The work by Mansoor et al. [78] examined the global sentiment analysis of tweets about COVID-19 and the evolution of global sentiment over time.The authors also studied tweets focusing on Work From Home (WFH) and Online Learning to gauge the effect of COVID-19 on daily areas of life.They used different machinelearning models, such as Long Short-Term Memory (LSTM) and Artificial Neural Networks (ANNs), to perform sentiment analysis and reported the accuracy of these models.Pokharel [79] used Google Collab to perform text mining and sentiment analysis of tweets focusing on COVID-19.The study involved collecting tweets posted between 21 May 2020 and 31 May 2020 by Twitter users who shared Nepal as their location.According to the study's findings, while the majority of people had a positive attitude toward COVID-19, there were also situations where fear, grief, and disdain were expressed in the tweets.
In the study by Chakraborty et al. [80], two categories of tweets related to COVID-19 were studied.In one instance, the top 23,000 retweeted tweets over the period of 1 January 2020 to 23 March 2020 were studied.According to the findings presented by the authors, the majority of the tweets expressed neutral or negative emotions.The paper also reports the findings from the analysis of a dataset encompassing 226,668 tweets from the period between December 2019 and May 2020.The findings show that there was a disproportionately high number of neutral and positive tweets posted by internet users.The study also showed that despite the majority of COVID-19-related tweets being positive, internet users were preoccupied with retweeting the negative ones.The objective of the work by Shofiya et al. [81] was to comprehend and examine perceptions of social distancing in the context of COVID-19, as expressed on Twitter.The study focused on analyzing tweets emerging from Canada and containing social-distancing-related keywords.The authors used the SentiStrength tool to determine the sentiment polarity of tweets.Basiri et al. [82] proposed a methodology based on the fusion of four deep-learning models and one classical supervised machine-learning model for sentiment analysis of COVID-19 tweets.They applied their methodology to tweets originating from eight countries.Cheeti et al. [83] used a Naïve Bayes classifier to perform sentiment analysis of tweets focusing on COVID-19, with a specific focus on tweets related to education and learning.In their study, Ridhwan et al. [84] performed sentiment analysis of tweets about COVID-19 posted between 1 February 2020 and 31 August 2020, with a specific focus on tweets that originated from Singapore.The findings showed that the majority of the tweets had positive emotions.Tripathi [85] and Situala et al. [86] used multiple machine-learning approaches to perform sentiment analysis of COVID-19-focused tweets that were posted by people who stated their location as Nepal on Twitter.The purpose of the work by Gupta et al. [87] was to examine the perceptions of Indians, as expressed on Twitter, towards the Indian Government's countrywide lockdown, which was implemented to slow the spread of COVID-19.In this context, the authors used the LinearSVC classifier to perform sentiment analysis, and their classifier achieved a performance accuracy of 84.4%.The work by Alanezi et al. [88] focused on performing sentiment analysis of tweets originating from multiple countries.The results of the study showed that most tweets originating from the USA, Australia, Nigeria, Canada, and England had a neutral sentiment.A similar study that focused on performing sentiment analysis of tweets originating from multiple countries was performed by Dubey [89].In addition to the above, several studies focused on performing sentiment analyses of tweets about COVID-19 originating from different countries, such as the United Kingdom [90][91][92][93][94][95], the United States [92,93,[96][97][98][99][100], Canada [101][102][103][104][105], India [106][107][108][109][110], Australia [111][112][113], and Brazil [114][115][116].
Recent Works that Focused on Sentiment Analysis and Text Analysis of Tweets about MPox
Iparraguirre-Villanueva et al. [117] aimed to examine people's emotions, including positive, negative, and neutral sentiments, towards the MPox outbreak by analyzing tweets containing the hashtag #Monkeypox.The findings of the study indicated that 45.42% of individuals did not express any discernible positive or negative opinions, whereas 19.45% conveyed negative and apprehensive sentiments related to the outbreak.The objective of the study by Mohbey et al. [118] was to infer the range of reactions of the general public in response to the MPox outbreak.The methodology was based on using CNN and LSTM to study relevant tweets to infer these specific characteristics.Farahat et al. [119] conducted a study involving sentiment analysis and topic modeling of tweets associated with MPox.The tweets that were analyzed in this study were posted on Twitter between 22 May 2022 and 5 August 2022.The authors utilized the concept of keyword search to mine tweets containing the keywords "monkeypox", "Monkeypox cases", and "Monkeypox virus".The findings of the sentiment analysis indicated that 48% of the tweets were neutral, 37% were positive, and 15% were negative.The authors used LDA to extract 12 topics that were present in these tweets.Sv et al. [120] focused on understanding the attitude of the general public towards MPox, as expressed on Twitter.They performed sentiment analysis of tweets containing the keyword "monkeypox" that were posted between 1 June 2022 and 25 June 2022.The results of sentiment analysis showed that the percentage of positive tweets was higher as compared to the percentage of negative tweets.The results of topic detection revealed multiple subject matters associated with both positive and negative tweets.The work by Bengesi et al. [121] was performed primarily in two steps.The first step of their work involved mining relevant tweets related to MPox.Thereafter, in the next step, they developed and used multiple categorization models to perform sentiment analysis of these tweets.Dsouza et al.'s work [122] focused on performing sentiment analysis of specific tweets related to MPox to detect any stigmatization of the LGBTQ+ community on Twitter.They retrieved tweets posted between 1 May 2022 and 7 September 2022 containing the hashtags "#monkeypox", "#MPVS", "stigma", and "#LGBTQ+".The study involved the analysis of a total of 70,832 tweets.
Zuhanda et al. [123] performed sentiment analysis on 5000 tweets about MPox posted on 5 August 2022.The study showed that the terms "health", "emergency", "public", "covid", and "declares" were often used by Twitter users in the context of tweeting about MPox.The NRC lexicon comparison categorization revealed that fear was the most often expressed emotion, with a representation rate of 19.73%.This was followed by sorrow at 14.77%, trust at 13.90%, anger at 9.99%, shock at 9.14%, disgust at 8.12%, and happiness at 7.90%.In the work by Cooper et al. [124], tweets about MPox posted between 1 May 2022 and 23 July 2022 were studied.The results showed that LGBTQ+ advocates or allies posted a total of 48,330 tweets, and the average sentiment score for all the tweets was −0.413 on a scale of −4 to +4.Ng et al. [125] collected tweets that contained "monkeypox" or "monkey pox" posted on Twitter between 6 May 2022 and 23 July 2022.They used concepts of topic modeling and sentiment analysis to infer the characteristics of the communication expressed in these tweets.The authors identified five topics, which they divided into three main themes.These included worries about safety, the stigmatization of minority populations, and a general loss of confidence in governmental institutions.The public sentiments highlighted increasing and existing partisanship, personal health concerns related to the changing situation, and worries about how the media portrayed minority and LGBTQ communities.
As can be seen from these works that focused on sentiment analysis and text analysis of tweets related to MPox and COVID-19, none of them focused on mining and analyzing tweets that focused on COVID-19 and MPox at the same time to infer the underlying patterns of sentiments.The work presented in this paper aims to address this research gap.The methodology that was followed is discussed in Section 3, and the results are presented in Section 4.
Methodology
This section outlines the methodology that was followed for the development and implementation of the proposed framework for performing sentiment analysis and text analysis of tweets that focused on COVID-19 and MPox simultaneously.
First of all, a relevant Twitter dataset had to be selected.The dataset that was selected for this study is MonkeyPox2022Tweets [126].This dataset presents more than 600,000 Tweet IDs of tweets about the 2022 outbreak of MPox.These tweets were posted between 7 May 2022 and 3 March 2023.The dataset comprises tweets in 34 languages, with English being the most common language in which the Tweets are available.The tweets in the dataset include 5470 distinct hashtags related to MPox, out of which #monkeypox is the most frequent hashtag.As this dataset comprises only Tweet IDs, the Hydrator app [127] was used to hydrate this dataset.The process of hydration refers to the process of obtaining the tweets and related information corresponding to each of the Tweet IDs.The Hydrator app works by complying with the policies of accessing the Twitter API, as well as the specific rate limits in terms of accessing the Twitter API.The following steps were followed for hydrating the Tweet IDs present in this dataset: 1.The desktop version of Hydrator was downloaded and installed on a computer with a Microsoft Windows 10 Pro operating system (Version 10.0.19043Build 19043) comprising Intel(R) Core (TM) i7-7600U CPU @ 2.80 GHz, 2904 Mhz, 2 Core(s) and 4 Logical Processor(s) 2. The Hydrator app was then connected to the Twitter API by clicking on the "Link Twitter Account" button on the app's interface.3.This next step involved uploading a dataset file to the Hydrator app for hydration.
As the Hydrator app allows only one file to be uploaded at a time, all the dataset files (containing only Tweet IDs) were merged to create one .txtfile, which was uploaded to the app.4.Then, specific information about the uploaded dataset file (such as Title, Creator, Publisher, and URL) was entered in the Hydrator app, and then the "Add Dataset" button was clicked to complete the process of dataset upload.5. Thereafter, in the "Datasets" tab of the Hydrator app, the "Start" button was clicked to initiate the process of hydration.
Figure 1 is a screenshot from the Hydrator app obtained after the completion of this hydration task.The output of the Hydrator app provided 509,248 tweets about MPox.Upon obtaining these tweets, it was crucial to perform text filtering to obtain tweets that contained keywords related to COVID-19.The specific keywords that were selected for text filtering were "COVID", "COVID19", "coronavirus", "coronavirus pandemic", "COVID-19", "corona", "corona outbreak", "omicron variant", "SARS-CoV-2", "corona virus", and "Omicron".These keywords were selected based on the findings of [128].The text filtering task produced a set of 61,862 Tweets, i.e., each of these Tweets focused on MPox and COVID-19 at the same time.This set of 61,862 tweets was selected for performing sentiment analysis and text analysis.It is worth mentioning here that Twitter introduced multiple changes to the Twitter API in April 2023, as a result of which the Hydrator app is not functional at present.However, the work that involved the usage of the Hydrator app was completed by the first week of March 2023.So, the recent changes to the Twitter API did not have any effect on this study.There are various approaches for performing sentiment analysis, such as manual classification, Linguistic Inquiry and Word Count (LIWC), Affective Norms for English Words (ANEW), the General Inquirer (GI), SentiWordNet, and machine-learningoriented techniques relying on Naïve Bayes, Maximum Entropy, and Support Vector Machine (SVM) algorithms.However, the specific approach that was used in this study was VADER (Valence Aware Dictionary for sEntiment Reasoning).VADER was used because it has been reported to outperform manual classification, and it addresses the limitations in similar approaches for sentiment analysis, as outlined below [129]: a. VADER distinguishes itself from LIWC, as it is more sensitive to sentiment expressions in social media contexts.b.The General Inquirer suffers from a lack of coverage of sentiment-relevant lexical features common to social text.c.The ANEW lexicon is also insensitive to common sentiment-relevant lexical features in social text.d.The SentiWordNet lexicon is very noisy; a large majority of synsets have no positive or negative polarity.e.The Naïve Bayes classifier involves the naïve assumption that feature probabilities are independent of one another.f.
The Maximum Entropy approach makes no conditional independence assumption between features and thereby accounts for information entropy (feature weightings).g.In general, machine-learning classifiers require (often extensive) training data, which are, as with validated sentiment lexicons, sometimes troublesome to acquire.h.In general, machine-learning classifiers also depend on the training set to represent as many features as possible.
VADER uses sparse-rule-based modeling to build a computational sentiment analysis engine that performs well on the social-media-style text while easily generalizing to multiple domains, needs no training data but is built from a generalizable, valencebased, human-curated gold-standard sentiment lexicon, is quick enough to utilize online with streaming data, does not suffer significantly from a speed-performance tradeoff, has a time complexity of O(N), and is freely available without any subscription or purchase costs.In addition to detecting the polarity (positive, negative, and neutral), VADER is also able to detect the intensity of the sentiment expressed in the texts.To develop the system architecture for sentiment analysis and text analysis, RapidMiner was used.RapidMiner, formerly known as Yet Another Learning Environment (YALE) [130], is a data science platform that enables the development, implementation, and utilization of several algorithms and models related to machine learning, data science, artificial intelligence, and big data.RapidMiner is utilized for both academic research and the creation of business-related applications and solutions.RapidMiner is available as an integrated development environment that consists of (1) RapidMiner Studio, (2) RapidMiner Auto Model, (3) RapidMiner Turbo Prep, (4) RapidMiner Go, (5) RapidMiner Server, and (6) RapidMiner Radoop.For all the work related to the methodologies proposed in this paper, RapidMiner Studio was used.For the remainder of this paper, wherever the phrase "RapidMiner" is used, it refers to "RapidMiner Studio" and not any of the other development environments associated with this software tool.RapidMiner is created as an open-core model with a powerful Graphical User Interface (GUI) that enables developers to create numerous applications and workflows and develop and implement algorithms.In the RapidMiner development environment, specific operations or functions are referred to as "operators" and a collection of "operators" (connected linearly or hierarchically or as a combination of both) to achieve a desired task or goal is referred to as a "process".For the creation of a particular "process", RapidMiner offers a variety of built-in "operators" that may be utilized directly.A particular class of "operators" can also be utilized to change the distinguishing qualities of other "operators".Moreover, the development environment also allows developers to develop their own "operators", which can then be shared and made accessible to all other RapidMiner users via the RapidMiner Marketplace.
The VADER approach for performing sentiment analysis is available as an "operator" in RapidMiner, which can be directly used in a "process".This "operator" calculates and then outputs the sum of all sentiment word scores in a given text(s) by following the VADER approach.If the advanced output option of this "operator" is selected, then it also outputs a nominal attribute with all words taking part in the scoring, the sum of positive components, the sum of negative components, and the number of used and unused tokens.The "process" that was developed in RapidMiner involving the use of this "operator" and other "operators" connected to it is shown in Figure 2. The description of all the "operators" used in this "process" is presented next.The "Dataset" "operator" was used to import the original dataset of 509,248 tweets about MPox (obtained from the output of the Hydrator app).The "Filter Tweets" "operator" was used to perform text filtering on the text of the tweets.Specifically, tweets that contained one or more of these keywords -"COVID", "COVID19", "coronavirus", "coronavirus pandemic", "COVID-19", "corona", "corona outbreak", "omicron variant", "SARS-CoV-2", "corona virus", and "Omicron" were filtered.Thereafter, the "Select Attributes" "operator" was used to select only that specific attribute from the dataset that would be used for sentiment analysis.The specific attribute in this context was the text of the tweets.The output of this "operator" was provided as an input to the "Extract Sentiment" "operator", which performed sentiment analysis according to the VADER approach.The output of this "operator" comprised a score associated with each tweet, classifying it as a positive, neutral, or negative tweet.To compute the number of positive, neutral, or negative tweets, additional data filters were used.However, this required creating multiple copies of the output.To achieve this, the "Multiply" "operator" was used.Specifically, three copies of the output from the VADER "operator" were created by using this operator.These copies of the output were passed through data filters that had been set up to filter out the positive, neutral, and negative tweets based on specific rules as per the working of the VADER approach.These rules were-a tweet with a score greater than 0 was filtered as a positive tweet, a tweet with a score equal to 0 was filtered as a neutral tweet, and a tweet with a score less than 0 was filtered as a negative tweet.Thereafter, an analysis of the number of tweets from these respective data filters was performed to infer the percentages of positive, neutral, and negative tweets.These results are discussed in Section 4.
In addition to performing sentiment analysis, this study also involved the detection of some of the commonly used hashtags and words in the 61,862 tweets that were considered for this study.The RapidMiner "process" that was developed to implement the same is shown in Figure 3.The description of all the "operators" used in this "process" is presented next.The "Dataset" "operator" was used to import the original dataset of 509,248 tweets about MPox (obtained from the output of the Hydrator app).The "Filter Tweets" "operator" was used to perform text filtering on the text of the tweets.Specifically, tweets that contained one or more of these keywords -"COVID", "COVID19", "coronavirus", "coronavirus pandemic", "COVID-19", "corona", "corona outbreak", "omicron variant", "SARS-CoV-2", "corona virus", and "Omicron" were filtered.Thereafter, the "Select Attributes" "operator" was used to select only that specific attribute from the dataset that would be used for sentiment analysis.The specific attribute in this context was the text of the tweets.The output of this "operator" was provided as an input to the "Nominal to Text" operator.Thereafter, the "sub-process" "Process Documents" was used.This "sub-process" comprised specific operators to perform tokenization and elimination of stopwords.The output of this "operator" was provided as an input to the "WordList to Data" operator to display the results for detection and analysis of the commonly used hashtags and words in these tweets.The results of this "process" are also discussed in Section 4. It is worth mentioning here that the VADER "operator" performs tokenization and elimination of stopwords automatically, so the "sub-process" "Process Documents" was not used in the RapidMiner "process" (shown in Figure 2) to perform the sentiment analysis.
Results and Discussion
This section is divided into three parts.Section 4.1 presents the results of sentiment analysis of 61,862 tweets that focused on MPox and COVID-19 at the same time.In Section 4.2, the results of the text analysis of the tweets are presented.Specifically, this section reports some of the commonly used hashtags and words that were present in these tweets.Section 4.3 presents a comprehensive comparative study with all the prior works in this field (reviewed in Section 2) to further uphold the scientific contributions of this paper.
Results of Sentiment Analysis
The sentiment analysis of this set of 61,862 tweets that focused on MPox and COVID-19 at the same time was performed using the VADER approach.The output of the VADER "operator" presented multiple new attributes, and each attribute provided specific information related to the sentiment associated with the tweets that were analyzed.Figure 4 shows the output that was produced by the RapidMiner "process" shown in Figure 2. To avoid presenting an image with 61,862 rows, Figure 4 shows a random selection of 19 rows from the output table.In this Figure, the columns marked in yellow were introduced by the VADER "operator" and were not originally present in the dataset.For each tweet, the VADER approach performed tokenization at first.This is represented in the attributes "Total Tokens" and "Uncovered Tokens" in Figure 4. Thereafter, it captured those tokens from the tweets which expressed either a positive or negative sentiment and then assigned a sentiment score to these respective tokens.This score was assigned on a scale of −4 to +4, where −4 meant highly negative, and +4 meant highly positive.These sets of tokens and their respective sentiment scores comprised the value of the "scoring string" (as shown in Figure 4) for each tweet.Thereafter, for each tweet, the VADER approach grouped all those tokens that had a positive sentiment and computed the sum of the sentiment scores for those tokens.This comprised the "Positivity" value of that tweet.In a similar manner, the VADER approach grouped all those tokens that had a negative sentiment and computed the sum of the sentiment scores for those tokens.This comprised the "Negativity" value of that tweet.The difference between the "Positivity" value and the "Negativity" value was thereafter computed by the VADER approach to display the overall score of the tweet.If this score was negative, the tweet was considered to have an overall negative sentiment.If this score was positive, the tweet was considered to have an overall positive sentiment.Finally, if this score was zero, the tweet was considered to have a neutral sentiment.Based on this analysis, the number of tweets with a positive sentiment was observed to be 29,000, the number of tweets with a negative sentiment was observed to be 19,780, and the number of tweets with a neutral sentiment was observed to be 13,082.This is illustrated in Figure 5.According to these findings, it can be concluded that almost half the tweets (46.88%) that focused on COVID-19 and MPox simultaneously had a negative sentiment.It was followed by tweets that had a positive sentiment (31.97%) and tweets that had a neutral sentiment (21.14%), respectively.
Results of Text Analysis
The results of the text analysis of the tweets are presented in this section.The steps included tokenization, removal of stopwords, and word frequency analysis.Table 2 shows the list of the top 50 hashtags and their frequencies.Here, frequency refers to the number of times each of these hashtags was present in the total number of tweets.The results of tokenization are presented next.In view of the large number of tokens obtained from the set of tweets, this analysis was performed by including the top 100 tokens in terms of their respective frequencies.Table 3 shows these tokens, and a visual representation of the same in the form of a word cloud is shown in Figure 6.As can be seen from this table, several words directly related to these respective viruses are in the list of the top 100 used words.This was expected, as this study focuses on tweets about COVID-19 and MPox.At the same time, the fact that this analysis shows several words that are not directly related to any of these viruses, such as "Polio", "Biden", "Ukraine", "HIV", "climate", and "Ebola", in the list of top 100 most frequent words that featured in these tweets underlines the fact that topics of conversations on Twitter in the context of COVID-19 and MPox also included a high level of interest related to other viruses, President Biden, and Ukraine.
Comparative Study with Prior Works
This section presents a comparative study with prior works in this field (reviewed in Section 2).This comparative study is represented in Table 4.As can be seen from Table 4, the work presented in this paper is the first work in this area that focuses on sentiment analysis of tweets that focused on COVID-19 and MPox at the same time.✓ ✓
Conclusions
The big data of Twitter conversations holds the potential for the inference of the views, opinions, perspectives, mindsets, sentiments, and feedback of the general public towards pandemics, epidemics, viruses, and diseases.This has attracted the attention of researchers in the fields of computer science, big data, data science, epidemiology, healthcare, medicine, and their interrelated areas in the last few years.Various forms of analysis of this big data, such as sentiment analysis, hashtag analysis, and frequentkeyword analysis, can be seen in prior works in this field that focused on studying tweets involving some of the virus outbreaks of the past, such as Ebola, E-Coli, Dengue, Human Papillomavirus, Middle East Respiratory Syndrome, Measles, Zika virus, H1N1, influenza-like illness, swine flu, flu, Cholera, COVID, Listeriosis, cancer, Liver Disease, Inflammatory Bowel Disease, kidney disease, lupus, Parkinson's, Diphtheria, and West Nile virus.The recent outbreaks of COVID-19 and MPox have escalated the use of Twitter for conversations related to these respective viruses.While there have been a few works published in the last few months that focused on performing sentiment analysis of tweets related to either COVID-19 or MPox, none of the prior works in this field thus far focused on the analysis of tweets focusing on both COVID-19 and MPox at the same.To address this research gap, this study presents the findings from a comprehensive sentiment analysis of 61,862 tweets that focused on MPox and COVID-19 at the same time.The VADER approach was used to perform the sentiment analysis.The results show that almost half the tweets (46.88%) involving COVID-19 and MPox had a negative sentiment.It was followed by tweets that had a positive sentiment (31.97%) and tweets that had a neutral sentiment (21.14%), respectively.This study also presents the findings from hashtag analysis and keyword analysis of these tweets.The top 50 hashtags that featured in all these tweets were detected and are presented in this paper.The top 100 most frequently used words that featured in all these tweets were also detected using concepts of tokenization.The findings of frequent word analysis show that some of the commonly used words directly referred to either or both of these viruses.In addition to this, the presence of words such as "Polio", "Biden", "Ukraine", "HIV", "climate", and "Ebola" in the list of the top 100 most frequent words indicate that topics of conversations on Twitter in the context of COVID-19 and MPox also included a high level of interest related to other viruses, President Biden, and Ukraine.A limitation of this study is that the data preprocessing and analysis did not involve the detection and elimination of tweets posted by bot accounts on Twitter.Future work would involve addressing this limitation and collecting more tweets over the next months to repeat this study, with an aim to infer and analyze any potential evolution or trends of public sentiment related to these viruses over the course of time.
Figure 1 .
Figure 1.Screenshot from the Hydrator app after completion of the Hydration process.
Figure 2 .
Figure 2. The RapidMiner process developed for performing sentiment analysis.
Figure 4 .
Figure 4.A random selection of 19 rows from the output table generated by RapidMiner.
Figure 5 .
Figure 5. Representation of the percentage of positive, negative, and neutral tweets present in this dataset.
Figure 6 .
Figure 6.Representation of some of the most frequently used words in these tweets in the form of a word cloud.
Table 1 .
A random collection of 10 Tweets that focused on COVID-19 and MPox simultaneously.
Table 2 .
The list of top 50 hashtags and their frequencies in the given tweets.
Table 3 .
The list of the top 100 words from these tweets and their respective frequencies.
Table 4 .
Comparative study with prior works in this field. | 8,621.2 | 2023-06-09T00:00:00.000 | [
"Computer Science"
] |
New acylphloroglucinols from a crude acetone extract of Eucalyptus camaldulensis Dehnh. leaf
Abstract Acylphloroglucinols are well-known Eucalyptus secondary metabolites which exhibit a variety of structures and bioactivities. The investigation of a crude acetone extract of Eucalyptus camaldulensis leaves led to the isolation of two new acylphloroglucinols, eucalypcamals O and P (1 and 2) together with seven phloroglucinols (3–9), and a benzene derivative (10). Their chemical structures were elucidated by 1D and 2D nuclear magnetic resonance (NMR) spectroscopy and mass spectroscopy. The absolute configurations of compounds 1 and 2 were established by comparison of experimental and calculated electronic circular dichroism (ECD) data. In the putative biosynthetic pathway, eucalypcamals O and P should be derived from hetero-Diels-Alder reaction between grandinol and trans-isoeugenol. Graphical Abstract
Introduction
Acylphloroglucinols are secondary metabolites that have been shown to exert a variety of bioactivities.In various studies, they have exhibited cytotoxicity (Wang et al. 2012), anti-inflammatory (Gu et al. 2020), and acetylcholinesterase inhibitory activities (Qin et al. 2018a).The Eucalyptus plants, belonging to the family Myrtaceae, are the important natural sourced of acylphloroglucinol metabolites with different type of acyl groups as isovaleryl (Kokumai et al. 1991), isobutyryl (Qin et al. 2018a), and 2-methylbutanoyl (Cheng and Snyder 1991).Eucalyptus species have also yielded acyphloroglucinols coupled with terpenes (Shang et al. 2019), b-triketones (Wang et al. 2014) or other phloroglucinol (Qin et al. 2018a) moieties providing more complex natural products with multiple bioactivities.
Eucalyptus camaldulensis (Myrtaceae) is known to contain bioactive compounds that display antibacterial (Cimanga et al. 2002;Salem et al. 2015;Nasr et al. 2019), antifungal (Su et al. 2006, Elansary et al. 2017), analgesic, anti-inflammatory effects (Silva et al. 2003), antioxidative and antiradical activities (Siramon and Ohtani 2007).In our previous study of a non-polar extract of the leaves of E. camaldulensis, we reported a series of new acylphloroglucinol-meroterpenoids that displayed cytotoxicity and antimicrobial activity (Daus et al. 2022).In the course of our continued investigation, we isolated two new acylphlorglucinols, eucalypcamals O and P (1 and 2) together with eight known compounds from a polar extract of E. camaldulensis leaves.Herein, we described the structure characterisation and putative biosynthetic pathway of the new compounds.In addition, the purified known compounds were evaluated for cytotoxicity and antimicrobial activity.
Compound (1) was obtained as a pale yellowish gum.The HRESIMS of 1 revealed an m/z at 415.1751 ) implied that methine carbon C-8 0 was substituted by a methyl group (Supplementary material, Figure S2).The complete structure of 1 was finalised by an HMBC experiment, HMBC correlations of 1 0 -OH/H-3 0 to C-1 0 (d C 146.4) and 2 0 -OCH 3 /H-6 0 to C-2 0 (d C 147.0) placed the hydroxy and methoxy groups at C-1 0 and C-2 0 , respectively.The 1,2,4-trisubtituted aromatic ring was connected to C-7 0 which was confirmed by the HMBC correlations of H-3 0 to C-7 0 (d C 86.5) and H-7 0 to C-3 0 (d C 109.6), C-4 0 (d C 130.1), and C-5 0 (d C 121.2).The relative stereochemistry of 1 was determined by the magnitude of the coupling constant (J ¼ 10.0 Hz) between H-7 0 and H-8 0 which indicated a trans relationship between these two protons (Satoh et al. 1992;Singh et al. 1997).The trans relationship was supported by NOESY correlations of H-7 0 to H 3 -9 0 and H-7a which confirmed the b-orientation of H-7a, H 3 -9 and H-7 0 , and a-orientation of H-7b, H-8 0 and phenyl ring.The electronic circular dichroism (ECD) spectrum displayed a positive Cotton effect at 241 (De þ 8.80) nm and negative Cotton effects at 212 (De À 21.35), and 270 (De À 8.23) nm.The absolute configuration of 1 as 7 0 S,8 0 S was determined by the comparison of calculated and experimental ECD data (Supplementary material, Figure S3).Thus, 1 was a new phloroglucinol and was given the name eucalypcamal O.
Eucalypcamal P (2) was obtained as a pale yellowish gum.Compound 2 possessed the same molecular formula C 23 H 26 O 7 as that of 1 which was established by HRESIMS The similar UV, IR and NMR data (Supplementary material, Table S1) of 1 and 2 demonstrated that they were structural isomers; the only difference being the alternation of isovaleryl and formyl , respectively (Supplementary material, Figure S2).Compound 2 showed the same NOESY correlations as 1, confirming the b-orientation of H-7a, H 3 -9 0 and H-7 0 and the a-orientation of the phenyl ring.The comparison of calculated and experimental data and calculated ECD spectra confirmed the absolute configuration of 2 to be 7 0 S,8 0 S (Supplementary material, Figure S3).Compound 2 was thus a new phloroglucinol and given the name eucalypcamal P. Acylphloroglucinols in Eucalyptus species are produced via different biosynthetic pathways.As a result, these compounds exhibit a certain structural diversity.In this study, the plausible biosynthetic pathway of eucalypcamals O and P (1-2) were proposed.The key o-quinone methines (i and ii) are generated from grandinol (9) (Singh et al. 1997) and couple with trans-isoeugenol, the essential oil of Eucalyptus plants (Rencoret et al. 2011), via the hetero-Diels-Alder reaction (Tran and Cramer 2014).The reaction produces the new O-C-7 0 and C-7-C-8 0 bonds that complete the structure of compounds 1 and 2 (Supplementary material, Figure S4).
Compounds 3-10 were evaluated for cytotoxicity against human cancer cell lines including the colorectal cancer cells LoVo and SW48, the cervical cancer cell HeLa, and the bone cancer cell SW1353.However, all the compounds exhibited weak cytotoxicity against these cancer cell lines with IC 50 values >100 lg/mL.Compounds 3-10 were further tested for antimicrobial activity against both Gram positive (Staphylococcus aureus) and Gram negative (Escherichia coli) bacterial strains and a fungus (Candida albicans).All were inactive against the tested microbes.
Plant material
The leaves of E. camadulensis Dehnh.were collected from Nong Song Hong village, Huai Haeng subdistrict, Kaeng Khoi district, Saraburi Province, Thailand in April 2017 and identified by Dr. Phattaravee Prommanut.A voucher specimen (BK No. 070170) was deposited in the Bangkok Herbarium, Bangkok, Thailand.
ECD calculations
All configurations of compounds 1 and 2 in this work were optimised using the density functional theory (DFT) approach at the B3LYP/6-31G(d, p) level of theory.The excited states were calculated by using TD-DFT at the B3LYP functional with the 6-311þþG(d, p) basis.The geometry optimisation and TD-DFT calculations were both performed using a polarisable continuum salvation model (PCM) with methanol.The rotary strengths of 80 excited states were calculated.All calculations were performed using the Gaussian09 program package with a fitting parameter of r ¼ 0.3 eV and were processed with the SpecDis 1.64 program (Frisch et al. 2010).
Cytotoxicity assay
Cell proliferation or cell viability of the cell lines was measured by MTT colorimetric assay.Human cancer cell lines including colorectal cancer cells (LoVo and SW48), cervical cancer cell (HeLa), and bone cancer (SW1353) were obtained from the American Type Culture Collection.This study was carried out in accordance with a previously reported procedure (Daus et al. 2022).
Antimicrobial assay
This study was carried out in accordance with a previously reported procedure (Daus et al. 2022).
Conclusions
This investigation of a polar extract (acetone) of the leaves of E. camaldulensis yielded two new acylphloroglucinols together with seven known phloroglucinols and a known benzene derivative.The structure of the new compounds was likely derived from the hetero-Diels-Alder reaction involving the o-quinone methine of grandinol and transisoeugenol.However, the known compounds (3-10) exhibited neither cytotoxicity nor antimicrobial activity.
Figure 1 .
Figure 1.Structure of compounds 1 and 2 from the leaves of E. camaldulensis. | 1,729 | 2022-09-02T00:00:00.000 | [
"Chemistry"
] |
Art Design of the Real-Time Image Interactive Interface of the Advertising Screen Based on Augmented Reality and Visual Communication
Augmented reality refers to the use of new display technology, multimedia technology, and human-computer interaction technology to superimpose computer-generated virtual objects, virtual scenes, or various system prompts into the real scene, so as to achieve integration with the surrounding environment of the user. The user can visually feel that the superimposed virtual information is a part of the real environment around him. Under the technology of augmented reality and visual communication, this article achieves the art design of the real-time image interactive interface of the advertising screen, digitally enhances the real situation, enriches the visual sense of the advertising audience, and turns the advertisement into an interactive form. This article analyzes the specific content of the composition, graphics, color, proportion, brightness, and design principles in the advertisement. We conduct a questionnaire survey and combine them according to the above 6 indicators. Query a large number of documents for analysis, and conduct a theoretical analysis of the real-time interactive image interface design of the advertising screen based on augmented reality technology. According to the experimental results obtained in this study, the data shows the P value of the scores of the six age groups on the text in the advertisement is less than 0.05. There is a significant difference; at the same time, the P value of each index in the advertisement is also less than 0.05. The significant difference indicates that the legibility of the text is an important factor in the text interaction in the interactive interface.
Introduction
Augmented reality technology is a new technology that uses virtual information to enhance real scenes. Augmented reality technology has attracted great attention from academia and industry with its good sense of reality, interest, and practicality and has also been applied and developed in many different fields. It is the most popular and most popular in the global science and technology community today one of the areas of concern. In recent years, research results such as tracking technology, interactive technology combined with computer graphics, and multimedia technology have been comprehensively used in augmented reality systems. While enabling the augmented reality system to achieve a realistic visual effect of seamless integration of the real environment and 3D virtual objects, allowing users to interact with virtual world objects in a natural way has gradually become the focus of research. Secondly, interactive advertising design promotes the faster development of online interactive advertising, humanized multilevel experience, and better improves the emotional coordination of the audience. As a product of the information age, online interactive advertising is still immature in theoretical research, but its practical application has begun extensively.
Augmented reality technology is dedicated to digitally adding computer-synthesized information such as visual images and auditory sounds to the real environment. Augmented reality technology can enhance various sensory perceptions, but because visual information accounts for the largest proportion of information obtained by humans, it is generally reflected in the enhancement of visual perception. Nah and Lee proposed that the functionality of images is of a certain aesthetic significance, so the way images are made will also lead to great changes in artistic creation [1]. Chen et al. proposed a template-based gesture recognition framework, which can detect and segment actions in time during video-based human-computer interaction. This method can process online video sequences in real time [2]. Laskari uses a TOF camera to connect to a laptop, record real-time respiratory processing results, and transmit them directly to the doctor's terminal through wireless technology to perform respiratory diagnostic functions without actually intruding the device [3]. Combining the history of interaction design development, El Ammari and Hammad deeply analyze the importance of interaction planning, study interaction design methods, and discuss how to design interactive experience prototypes [4].
Although the domestic start is relatively late, it is also rapidly researching augmented reality technology and it is very effective. Augmented reality technology has attracted the attention of major universities and research institutions at home and abroad with its huge application prospects and economic benefits, and it has spent a lot of manpower and material resources. Research on this technology has been conducted, and many laboratories have been established to study augmented reality-related technologies. Lim et al. used a TOF camera to perform 3D gesture posture imaging, and it was successfully applied in medical imaging. Collect data images through the TOF camera on the endoscope, and show a visual interface to the medical staff [5]. Deng et al. proposed that the general understanding of augmented reality in modern times has changed, emphasizing the two indications of three-dimensional and interactive, and effectively distinguished between plane postproduction technology and film and television packaging technology and augmented reality technology [6]. Gogolin and Gogolin proposed that augmented reality can promote the exchange of experience in augmented reality technology and, at the same time, provide guiding opinions for determining future research fields and future development trends [7]. Murakami et al. proposed the relationship between illustration and market economy and the important role of illustration in market development [8]. In practice, these studies have played a great role in promoting the research in the field of image enhancement technology and visual communication and have provided great value for supplementing the literature in this field, but the methods of these studies are not innovative enough. The experimental data is not very complete.
The definition and development of advertising and the characteristics and advantages of augmented reality advertising are thoroughly analyzed, and a feasibility study based on augmented reality advertising is proposed to verify the rationality and correctness of the proposed views. In terms of theoretical research and analysis, based on the current limitations of traditional advertising, research the characteristics and advantages of augmented reality technology advertising, and combine the two to explore the feasibility of augmented reality-based advertising and demonstrate that augmented reality technology can be used in advertising expression. Put forward some ideas for the application and development of interactive interface design based on augmented reality and visual communication in the field of advertising screens, and look forward to the research direction and significance of future topics.
Art Design of Real-Time Image Interactive
Interface of Advertising Screen Based on Augmented Reality and Visual Communication
Key Augmented Reality Technology
(1) Three-dimensional registration technology One of the most important requirements of augmented reality technology is to identify and locate objects in the real world. Accurate registration allows virtual information to accurately overlay real objects [9,10]. Three-dimensional registration technology is used to register and superimpose real-life scene information in a virtual three-dimensional scene modeled by a computer to enhance the computer's cognition of the environment. The virtual scene is the mainstay and the real scene is the supplement.
(2) Camera calibration technology The pictures captured by the camera lens are only twodimensional images composed of many pixels. Establish a coordinate system for the two-dimensional image generated by these many pixels. The computer calibration technology actually obtains the corresponding conversion matrix through the projection conversion of the above-mentioned coordinate system to obtain the internal and external parameters of the camera. Finally, the camera position and orientation are calculated to provide accurate information for the final registration [11,12]. Next, we will introduce the conversion process and mathematical derivation between these coordinate systems used during registration. As far as advertising is concerned, the current advertising is mostly in a two-dimensional era, which is a flat state. It is well known that the presentation of a flat surface will be affected by many factors, such as light and inclined surfaces, and the methods and techniques to use these factors are limited. Three-dimensional technology is a breakthrough in twodimensional technology, and the operation is simple, and the technical level is higher. A deeper combination analysis can bring better visual effects.
Rotation transformation matrix between two-dimensional coordinate systems: Rotation matrix between coordinate systems in threedimensional space: : The conversion relationship between point X w ðx w , y w , z w Þ T in the world coordinate system and coordinate X c ðx c , y c , z c Þ T in the camera coordinate system is as follows: : ð6Þ The external parameter of the M exp value camera is the 4 × 4 matrix, and 4 × 4 represents the rotation matrix from the world coordinate system to the camera coordinate system. Incorporating Formula (3) and Formula (6) into Formula (5), the conversion from the two-dimensional pixel plane coordinate system to the three-dimensional world coordinate system is obtained: where a x = f p x , a y = f p y , a x , a y , u 0 , v 0 is the internal parameters of the camera, written in a simplified form: M int refers to the internal parameters of the camera, and M exp refers to the external parameters of the camera, which are related to the position of the camera. Suppose the threedimensional coordinates of the template plane is ðx w , y w , z w Þ T , and the point on the screen image is ðu, vÞ T . Regardless of the special case z w = 0, we choose the template plane, then Formula (7) can be written as 3 Journal of Sensors where s is an arbitrary value, r 1 and r 2 are the rotation directions of the x and y directions, respectively, and H is the homography matrix of 3 × 3 that is sought. In the twoview geometry, it can be understood that two cameras take two images AB in the same space. One of the images A has a transformation in the other B, and there is a one-to-one correspondence. A matrix can be used between them homography. H can be expressed as Decompose the equation to get Since the rotation vectors r 1 and r 2 are mutually orthogonal vectors in the construction process, we have So we can get the formula: Formula (14) is the external parameters of the camera obtained by the homography matrix. : ð15Þ B contains the internal parameter information of the matrix. At the same time, according to the general form of Formula (15), rearrange the elements of B to form a new vector b. The following formula is obtained: : ð16Þ If there are multiple equations at the same time, combine these equations into a vector representation form, namely, Vb = 0.
(3) Augmented reality tracking registration technology It mainly monitors the position and movement of people or objects and then accurately places virtual information in the real scene. Augmented reality surveillance technology uses a variety of methods to locate objects and process the detected messages. The processed information provides input technology with conversion information from the image coordinate system to the actual coordinate system [13,14], as shown in Figure 1. Figure 1 is the construction of the operation process of the three-dimensional registration interface, which mainly includes three parts: hybrid registration, visual interface, and hardware. The visual interface has a single perspective, a dual perspective, and a triple perspective. The computer visionbased three-dimensional registration method is accurate. The characteristics of high performance and low hardware cost have attracted the attention of more and more researchers.
Augmented Reality Interface Requirements
(1) Human-computer interface development The man-machine system consists of three parts: man, machine, and environment. They are connected together to form a whole. There is a horizontal area of humancomputer interaction, called the human-machine interface, and the exchange takes the human-machine interface as the carrier. People receive information through the screen and make corresponding judgments and decisions in the brain [15,16]. Then, they control the controller and input a command to make the machine receive, and then, the machine executes the corresponding program of the command and displays the program. The server processes and analyzes the information provided by the man-machine operation interface and then transfers it to the command page. This process of data conversion is called an interface. This interface has an important connection in the information transmission between humans and machines in the circulatory system.
(2) Interactive interface based on augmented reality (i) Augmented reality technology Augmented reality technology can be regarded as one of the virtual reality technologies, but the creation of a new virtual world by virtual reality is different from virtual reality, while augmented reality emphasizes the combination of virtual reality and the real environment and virtual objects 4 Journal of Sensors created. Augmented reality uses computers to create a realistic virtual environment of vision, hearing, power, touch, and motion perception, enabling users to see a virtual and real combined space of real scenes seamlessly integrated into virtual objects. Through the projector, you can immerse yourself in the environment to realize the direct physical interaction between the user and the environment. It can simulate the real scenery of the scene and is an advanced human-computer interaction interface with basic interaction and capture functions [17,18]. Users can not only experience the realism experienced in the objective physical world through the virtual reality system but also penetrate the space, time, and other objective constraints of the virtual environment, so that users can experience the experience they cannot experience in the real world.
(ii) The prospect of augmented reality technology and the development trend of man-machine interface Augmented reality technology has broad prospects, especially in today's internet era; the prospects for augmented reality services on mobile platforms are very broad. The increasing emphasis of various industries on the integration of augmented reality technology in mobile applications will promote the rapid development of augmented reality technology [19,20]. Among many advertisers, the application of augmented reality technology to the positioning of advertising business information services has been highly recommended. This application can maximize the close connection between advertising products and customers and make full use of the various information of customers to closely connect with the market to maximize the benefits.
Illustrations in Print Ads in the Age of Electronic Images
(1) Graphic advertising illustration 5
Journal of Sensors
As an art form, illustration is now widely used in various fields of design. Particularly in books, advertisements, and packaging, the role of illustrations is very important. At the same time, images have gradually become an important part of our design. As an art form of information transmission, visualization can convey information in real time through language and accurately describe the function of the product. The narrative feature of illustration design in print advertisements is to spread important advertising information and ideas to the public.
(2) The role of illustrations in print ads The uniqueness of images has become the key to the dissemination of print advertisements and creation of works. When dealing with various graphic advertisement images, consumers' understanding and reading of product information will first focus on the graphics and images in the promotional items and then read the relevant product text content; the dissemination of printed advertising information with custom images can make consumers more receptive to products. Therefore, the unique realization and intuitiveness of graphic advertisements strongly reflect the importance and characteristics of advertising products and enhance the persuasiveness and influence of graphic advertisements.
(3) Design guidelines for illustrations in print advertisements in the electronic image age (i) Establish a clear advertising purpose The purpose of advertising is to spread, expand the awareness of the product in the eyes of the public, attract consumers to pay attention to the advertising product and have a strong interest in it, and stimulate people's sense of demand for the product. The purpose of print advertising is to induce consumers to consume. The design of illustrations in print advertising needs to be based on this fundamental purpose. The creative thinking of print advertising needs to focus on corporate publicity and product marketing. The use of illustrations in print advertisements in the era of electronic images makes the spread of advertising information faster. The essence of illustration in print advertising is practical art. For the purpose of illustration in print advertising, it is a direct and effective visual language. Artisticity is the creative method of illustration in print advertising. Only the purpose of print advertising can be accurately expressed so that the audience can feel the art of advertising illustrations while receiving advertising information and bring spiritual enjoyment to people.
(ii) Reflect a distinctive design theme The design theme is an important part of all art works, embodied as a kind of idea and matter. Illustrations in print ads need to have a clear theme in the design process and deal with it through simplified means to achieve a prominent and vivid effect and attract public attention and be easy to grasp by the public. Advertising illustration is a unique way to Journal of Sensors convey the connotation of goods, as the so-called first think about the problem objectively, and then express the content reasonably. In the era of electronic imagery, the competition among commercial brands in the market is very fierce. In order to make one's own products stand out among the many competing products, it is necessary to have a very unique creative expression.
(iii) Stimulate the emotional resonance of consumers Advertisements can give special symbolic meaning and connotation to the product brand, so that the target consumers will associate after seeing the advertisement, arouse emotional resonance, and finally achieve a deep and strong communication effect. In the age of electronic images, illustration design has become more and more focused on practical emotional resonance in advertising. It mainly uses warm, cheerful, and positive emotional images to achieve emotional resonance for target consumers and let people indulge in what the advertising gives. Memories and easier to accept the services and goods in the advertisement. Illustrations in print advertisements should pay attention to expressing human touch, appropriately and delicately portraying the details of life, so that illustrations can be more accessible to people.
(iv) Meet the psychological demands of consumers In the age of electronic images, information gradually becomes visualized images, and the design of illustrations in print advertisements needs to meet the psychological demands of the audience. When designing advertising illustrations, it is necessary to highlight the main body's image and sharp color contrast to achieve the purpose of attracting the attention of the audience. Through a series of consecutively changing comic strips, illustrations with continuity can be formed, creating a continuous sense of time and space, and fully expressing the activity process in the advertisement. At the same time, it also has a strong sense of movement and rhythm, achieving a special attention. An excellent graphic advertising illustration design must have a reading effect that can convey the connotation of the design and make the target audience understand and accept it.
Experimental Design of Real-Time Image
Interactive Interface of Advertising Screen 7 Journal of Sensors user experience and user needs, according to the theory of user experience elements, the required conceived augmented reality advertising design is based on five levels: strategy, scope, structure, framework, and presentation. Advertisement content design and planning plan production: according to the content of user experience elements, preliminary design, determine the general process of task completion and divide the steps, design according to the functional goals to be achieved in each step, determine the general style, here build the project based on the foundation, and finally carry out the test, as shown in Figure 2. 3.2. Test Subject. This article uses a real-time image interactive system for advertising screens based on augmented reality and visual communication technology. Analyze from the composition in the advertisement, the graphics in the advertisement, the color in the advertisement, the proportion in the advertisement, the brightness in the advertisement, and the design principles in the advertisement. Based on the above 6 indicators, 100 questionnaires were issued for each age group. All questions are divided into 5 levels. The questionnaire response rate was 100%.
Experimental Method.
In the preexperiment stage, we need to enter a large amount of sample data and classify and analyze the real-time image interaction system of the advertising screen based on augmented reality and visual communication technology. We also need to query a large amount of literature to design-related questionnaires.
In the postexperiment stage, we conducted statistics on the collected questionnaires and followed the six indicators of the composition in the advertisement, the graphics in the advertisement, the color in the advertisement, the scale in the advertisement, the brightness in the advertisement, and the design principles in the advertisement. And relevant knowledge to analyze these data to get a conclusion.
Data
Collection. Pixel depth refers to the number of bits used to store each pixel, and it is also used to measure the resolution of an image. Pixel depth determines the number of colors that each pixel of a color image may have or determines the number of gray levels that each pixel of a grayscale image may have. In the random forest training sample data, each sample has features. In this algorithm, simple pixel depth features are used for comparison. The feature calculation formula for pixel x is as follows: where d 1 ðxÞ represents the depth value of the x pixel in the image, the parameter θ = ðu, vÞ describes the offset of 3.5. Statistical Data Processing Method. SPSS23.0 software was used for data processing, and the count data was expressed in percentage (%), k is the number of data in this experiment, σ 2 is the variance of all survey results, and P < 0:05 indicates that the difference is statistically significant. The formula for calculating reliability is shown in 4. Real-Time Image Interactive Interface of Advertising Screen Based on Augmented Reality and Visual Communication
Evaluation Index System Based on Index Reliability
Testing. The results of the survey on the stability and incumbency of the questionnaire are shown in Table 1. It can be seen from Table 1 the composition of the advertisement, the graphics in the advertisement, the color in the advertisement, the scale in the advertisement, the brightness in the advertisement, and the design principles in the advertisement. The impact is acceptable (α > 0:7). This shows that in the era of electronic images, the forms of illustrations in print advertisements have become more and more diversified, and the space for development and creation has become wider.
Visual Communication Language in Advertising
(1) Analyze the composition in the advertisement Image symbols are the use of images to interpret and show the behaviors, objects and concepts to be expressed, so that they can be easily recognized, learned, and remembered. Image symbols help reduce the design efficiency load, save display area, and make identification and control easy to understand in all cultures. The results are shown in Table 2, and we make a bar graph based on this result, as shown in Figure 3. Through the paired sample t-test, it can be found that after the experiment, the six age groups have significant differences in the P value of the composition score in the advertisement less than 0.05; at the same time, the P value of each index in the composition in the advertisement is also less than 0.05, which also shows the difference is significant. This image representation method is usually the most efficient. The specific situation is shown in Figure 3.
(2) Analyze the graphics in the ad An interface with a mixed arrangement of pictures and text makes it easier to recall pictures rather than words. It is also found that when users browse the interface, they will spend more time on pictures than words. The results are shown in Table 3. We make a line chart based on this result, as shown in Figure 4.
Through experiments, it can be found that the P value of the graphic score in the advertisement is less than 0.05, and there are significant differences between the 6 age groups; at the same time, the P value of each indicator in the graphic in 6 Figure 6: Data chart of ratings of six age groups on the scale in advertising. the advertisement is also less than 0.05, which is also significant. The difference shows that it is very important to appropriately use the advantages of pictures to design the interactive interface. It can make the key priority information of the interface quickly become the focus of the user's attention. The specific situation is shown in Figure 4.
(3) Analyze the colors in advertising As a visual language, color not only corresponds to the real object but also connects with the human spirit and emotion. The results are shown in Table 4, and we made a bar graph based on the results, as shown in Figure 5.
Through experiments, it can be seen that the P values of color scores in the advertisements of the 6 age groups are all less than 0.05, and the difference is significant; at the same time, the P values of the color indicators in the advertisement are also less than 0.05, which is also significant.
(4) Analyze the metrics in advertising
The importance of proportion is one of the other visual communication elements based on the real-time image of the advertising screen. The results are shown in Table 5, and we made a bar graph based on the results, as shown in Figure 6.
Through experiments, it can be seen that there is a significant difference in the scale scores. P values of advertisements of 6 age groups are all less than 0.05; at the same time, the P value of each indicator of the advertising mesoscale is also less than 0.05.
(5) Analyze the text in the ad 10 Journal of Sensors In the text design of the interactive interface, you should choose to use simple sentences that are familiar to the user with clear expression and avoid using too many professional terms. Try to use verbs that can accurately and concisely describe the operation, so that it is clear and easy to understand by the user. The results are shown in Table 6. We make a histogram based on this result, as shown in Figure 7.
Through experiments, it can be seen that the P value of the text in the advertisement is less than 0.05, and there are significant differences between the 6 age groups; at the same time, the P value of each index in the advertisement is also less than 0.05. There is also a significant difference, which shows that the legibility of the text is an important factor in the text interaction in the interactive interface. If the legibility of the text is reduced, the user's reading comprehension, degree, and speed will also be reduced. In terms of page design, it is necessary to reduce the difficulty of text browsing and operational design page layout so that the text has a sense of beauty, and the user's text interaction efficiency in the interactive interface is improved. The specific situation is shown in Figure 7.
Conclusions
With the rapid development of human society's economy and technology, people are also seeking new humancomputer interaction methods that are more realistic and more immersive than traditional screen-based interaction methods. Augmented reality, as a new type of humancomputer interaction technology aimed at "seamlessly" superimposing virtual information in real scenes, has become one of the most concerned scientific and technological development directions in the scientific and technologi-cal circles. The main research goal of this paper is the artistic design of interactive real-time image interface of advertising screen based on augmented reality and visual communication technology. Through the classification and analysis of the literature, the real-time interactive image interface design of the advertising screen based on augmented reality technology is analyzed theoretically. Starting from the definition and format of advertising, it studies the functions and benefits of augmented reality technology advertising and demonstrates that augmented reality technology can be used for advertising screen expression. Design advertising screens based on augmented reality technology, through interactive multisensor design, to convey information or safety warnings to users. Finally, by evaluating the advertising effect, determine whether the design is successful, optimize the drawbacks, and improve the details of the user experience. The emergence and development of augmented reality technology have brought new opportunities for advertising and can create excellent advertising. Advertising based on augmented reality will become a new reference direction for future advertising development.
Interface design research mainly focuses on the visual aspects and content. The design process and principles virtualize the design of physical products, and there are actually no major changes. With the development of information technology and the maturity of human-computer interaction technology, the author believes that this is the watershed between traditional interface design and interactive interface design. From the research progress at home and abroad, it can be seen that even for the recently published academic results, the research content is still traditional software interface design, and there is no theoretical combining for interactive interface design. Through understanding the 11 Journal of Sensors process design and principles of the interactive interface, this article discusses how to augment the realistic interactive interface and studies its feasibility and value.
Data Availability
No data were used to support this study.
Conflicts of Interest
The author declares that there are no conflicts of interest regarding the publication of this article. | 6,825 | 2021-12-13T00:00:00.000 | [
"Art",
"Computer Science"
] |
CD18 (ITGB2) expression in chronic lymphocytic leukaemia is regulated by DNA methylation-dependent and -independent mechanisms
The integrin lymphocyte function-associated antigen 1 (LFA-1; CD11A/CD18; ITGAL/ITGB2), is a key regulator of lymphocyte trafficking, activation and prolonged lymphocyte residence in lymph nodes (LNs) (Reichardt et al, 2013); yet, little is known about the regulation of ITGAL (CD11A) and ITGB2 (CD18) transcription.
Chronic lymphocytic leukaemia (CLL) cells strongly depend on the lymphoid microenvironment, where they transiently localize to receive supportive signals by accessory cells (Burger & Gribben, 2014). Previously, we found that the majority of CLL cells expressed low surface LFA-1 levels as the result of reduced ITGB2 transcription, which was accompanied by impaired LN homing (Hartmann et al, 2009).
Dissecting LFA-1 expression in CLL subgroups of different cytogenetics, we observed comparable and correlating CD11A and CD18 surface expression, allowing cytometrical measurements of each subunit as a surrogate for the other (Fig 1A). We found a prominent increase of CD11A and CD18 expression in Trisomy 12 (tri12) harbouring CLL (Fig 1B,C), in accordance to a recent observation (Riches et al, 2014). The data confirmed our previous observation of associated expression of LFA-1 with CD38 and CD49D (ITGA4) (Hartmann et al, 2009), while we found no differences in CLL subgroups defined by the IGHV mutational status or ZAP70 expression ( Figure S1A,B). Curiously, we observed lower CD11A levels in subgroups harbouring the unfavourable 17p deletion or the favourable 13q deletion but no association of 11q deletion with CD11A expression ( Figure S1C).
Tri12 defines a CLL subgroup with high cell proliferation and an increased frequency of Richter transformation, which manifests with lymphadenopathy and LN infiltration (Tsimberidou & Keating, 2005). Recently, we reported that tri12 CLL cells are also characterized by high CD49D expression due to ITGA4 (CD49D) promoter hypomethylation (Zucchetto et al, 2013). Hypothesizing a similar mechanism responsible for regulation of LFA-1 expression, we studied promoter methylation of the rate-limiting CD18 subunit in CLL.
The ITGB2 promoter contains 23 potentially methylated CpG motifs (CpG1-CpG23) (Agura et al, 1992). We analysed these sites by bisulfite conversion of genomic DNA from anti-CD19 purified CLL cells followed by nested polymerase chain reaction, cloning and sequencing. Notably, whereas the region closer to the transcription start site (CpG16-CpG23) was unmethylated in CLL, the region from À603 bp to À241 bp upstream of the transcription start site (CpG4-CpG15) was variably methylated ( Figure S2). The grade of methylation in these sites inversely correlated with CD18 surface expression ( Fig 1D). A high grade of ITGB2 promoter methylation of CpG4-CpG15 was found in CLL samples with low CD18 expression, whereas CD18 high expressing CLL cells, overrepresented in the tri12 CLL group even in presence of a low percentage of tri12+ cells, were mostly unmethylated at the same sites (Figs 1D and S2). Our data indicate regulation of ITGB2 transcription by DNA hypomethylation in quiescent CLL cells of this specific subgroup.
Next, we addressed whether CLL cell activation could influence CD18 expression. Therefore, CLL cells were co-cultured in vitro with autologous T cells on a layer of murine fibroblasts and stimulated with IL2/CpG. After a 5-d co-culture, levels of the activation marker CD86 were measured and an up-regulation of CD18 in CD86-positive CLL cells was detected ( Figure S3). Particularly, in tri12 CLL, CD86-positive CLL cells in both unstimulated and IL2/CpG-stimulated samples expressed significantly higher CD18 than CD86-negative sub-clones, indicating a higher propensity of tri12 CLL to become activated and up-regulate CD18 expression, even in the absence of strong stimulation ( Figure S3). Furthermore, actual cell division occurred on day 5, with significantly higher proliferation rates of tri12 CLL cells (Fig 2A).
Within individual samples, CLL cell subpopulations that had either progressed through the cell cycle or remained quiescent showed higher CD18 expression in proliferating CLL cells ( Fig 2B). Remarkably, tri12 CLL were capable of undergoing more cell cycles in vitro than non-tri12 ( Fig 2C), with continuously increased CD18 levels in each proliferation round (Fig 2D), suggesting a higher intrinsic proliferative capacity of CD18-high-expressing sub-clones or an up-regulation of CD18 expression by signals from the co-culture.
In addition to IL2/CpG, other co-culture systems (Asslaber et al, 2013) effectively induced CD18 and CD11A expression in CLL, including activated autologous or allogeneic T cells (in presence of a fibroblast layer) but not CD40LG-overexpressing fibroblasts alone ( Figure S4A,B). Thus, secreted cytokines rather than direct CD40/CD40LG interactions seem responsible for this phenomenon.
Next, proliferative and quiescent CLL fractions were sorted according to their Cell Trace Violet TM (CTV) intensity using fluorescence-activated cell sorting. The levels of ITGB2 promoter methylation, however, were similar in non-proliferating and proliferating CLL cells, despite their significant up-regulation of CD18 expression (Fig 2E). Thus, while basal CD18 levels in resting CLL cells appear to be regulated by promoter methylation, CD18 up-regulation during proliferation is independent of demethylation. In line with these results, similar methylation patterns of CLL cells isolated from peripheral blood and LNs were recently observed (Cahill et al, 2013). Thus, rather than being a general regulatory mechanism of ITGB2 transcription, promoter hypomethylation appears to be specific for tri12 CLL.
As the ITGB2 promoter contains a potential NFKB1 binding site ( Figure S2) (Bottinger et al, 1994), located in the generally unmethylated region close to the transcription start site, we analysed the correlation between active NFKB1 levels and CD18 expression in resting CLL cells. RELA levels were therefore determined in nuclear fractions. Indeed, CD18 expression directly correlated with expression of the DNAbinding NFKB1 subunit RELA (Fig 2F), suggesting that NFKB1 inducing signals can influence CD18 levels in CLL. As non-tri12 CLL cells mainly display low NFKB1 levels, this pathway might be particularly important in the tri12 CLL subgroup.
In summary, we demonstrated that: (i) CD11A/CD18 is constitutively overexpressed in tri12 CLL and correlates with ITGB2 promoter hypomethylation; (ii) an activating microenvironment up-regulates CD11A/CD18 in both non-tri12 and tri12 CLL cells via a DNA methylation-independent process; (iii) tri12 CLL hereby display a higher propensity to proliferate and up-regulate CD18 in response to activation; (iv) levels of active NFKB1 and CD18 expression correlate in CLL.
Our study highlights that LFA-1 expression in CLL is subject to multiple transcriptional regulatory mechanisms, particularly in the peculiar CLL subgroup harbouring tri12. . DNA methylation and CD18 expression was analysed in the so defined CLL cells. (F) CD18 levels of CLL cells were correlated with nuclear NFKB1 subunit RELA, measured by an enzyme-linked immunosorbent assay. Statistics were performed using GRAPHPAD PRISM 5.0. After normality testing, Mann-Whitney t-test (for unpaired data; A) or Wilcoxon signed rank test (for paired data; B, E) was used to compare groups. Linear regression (D) or Spearman correlation (F) was used for correlation analysis. P-values are depicted as: P ≥ 0Á05 ns (not significant), P < 0Á05 *, P < 0Á01 **, P < 0Á001 ***. Medians are marked as lines (A). Corresponding numbers (n) of samples are given for each group in the graphs. Closed circles (•) indicate non-tri12 CLL samples, open circles (○) indicate tri12 CLL samples. non-prolif, non-proliferating; prolif, proliferating; MFIR, median fluorescence intensity ratio (MFI-specific Antibody/MFI corresponding isotype control).
Supporting Information
Additional Supporting Information may be found in the online version of this article: Fig S1. LFA-1 expression, prognostic markers and cytogenetic aberrations in CLL. Fig S2. Schematic presentation of the ITGB2 promoter. Fig S3. Influence of activation inducing stimuli on CLL cells and their LFA-1 expression. Fig S4. Increase of LFA-1 expression on CLL cells upon different proliferation inducing stimuli. | 1,727.6 | 2014-10-17T00:00:00.000 | [
"Biology",
"Medicine"
] |
Effective Usage of Biochar and Microorganisms for the Removal of Heavy Metal Ions and Pesticides
The bioremediation of heavy metal ions and pesticides is both cost-effective and environmentally friendly. Microbial remediation is considered superior to conventional abiotic remediation processes, due to its cost-effectiveness, decrement of biological and chemical sludge, selectivity toward specific metal ions, and high removal efficiency in dilute effluents. Immobilization technology using biochar as a carrier is one important approach for advancing microbial remediation. This article provides an overview of biochar-based materials, including their design and production strategies, physicochemical properties, and applications as adsorbents and support for microorganisms. Microorganisms that can cope with the various heavy metal ions and/or pesticides that enter the environment are also outlined in this review. Pesticide and heavy metal bioremediation can be influenced by microbial activity, pollutant bioavailability, and environmental factors, such as pH and temperature. Furthermore, by elucidating the interaction mechanisms, this paper summarizes the microbe-mediated remediation of heavy metals and pesticides. In this review, we also compile and discuss those works focusing on the study of various bioremediation strategies utilizing biochar and microorganisms and how the immobilized bacteria on biochar contribute to the improvement of bioremediation strategies. There is also a summary of the sources and harmful effects of pesticides and heavy metals. Finally, based on the research described above, this study outlines the future scope of this field.
Introduction
The rapid expansion of industrialization has resulted in the depletion of natural resources and the production of vast volumes of hazardous waste that pollute water and soil, threatening the environment and human health [1]. The deterioration of soil and water quality due to releasing toxic pollutants has become a serious threat around the world. The release of these harmful wastes into the environment occurs in different forms; for example, atmospheric pollutants include noxious gases such as sulfur oxides and nitrogen oxides, while soil and water can be contaminated by organic pollutants (pesticides, hydrocarbons, phenols, etc.) and heavy metals (cadmium, arsenic, lead, chromium, mercury, etc.). Human health can be adversely affected by these environmental pollutants [2] through inhalation or ingestion ( Figure 1). Additionally, some pollutants, such as heavy metal ions, can bioaccumulate in the food chain, and these persistent organic pollutants present significant risks to humans and other living creatures. The accumulation of pesticides and their derivatives is becoming more prevalent, due to the rising population and rapid industrialization. As much as 80% to 90% of pesticides applied to crops in agricultural fields affect non-target life forms; they can relocate or volatilize from the treated area to pollute the air and soil and negatively affect non-target plants. The leaching of these accumulated pesticides leads to the contamination of groundwater and soil [3].
During the last few decades, the separation of pollutants from water systems and soil via several methods has been developed and successfully applied. Recently, technologies such as membrane filtration, ion exchange, and chemical precipitation have been utilized in real-life applications to remove pollutants such as metal ions from polluted areas. Chemical precipitation is a frequently used method for treating heavy metals because it is simple, inexpensive, and effective. However, chemical precipitation results in secondary pollution and eventually leads to additional difficulties in cleaning up the trace contaminants from large areas. Ion-exchange resin offers fast kinetics and is highly efficient for pollutant removal. However, the need for an acidic environment restricts their application in various contexts. Membrane filtration-based technologies can remove toxic substances with high efficiency, but the manufacture of membrane material is usually very complex and at a high cost. Conventional pollutant remediation methods are not ecofriendly and produce toxic chemical sludge. Therefore, there is a serious need to develop efficient and sustainable technologies for remediating toxic environmental pollutants.
Biochar is a carbonaceous material produced through the thermal treatment of different types of biomass, such as crop residues and biosolids [4,5]. Biochar production can be achieved via various processes, including slow or fast pyrolysis, flash carbonization, gasification, hydrothermal carbonization, torrefaction, etc. [6]. The key goal when designing the synthesis of the biochar is that the final material should possess high porosity, a large specific surface area, and elevated surface chemistry heterogeneity, as with oxygencontaining functional groups and minerals [7]. The physicochemical properties of the final obtained biochar can also be tuned by altering the microstructure. Such characteristics encourage biochar's rising application in (waste) water treatment, soil improvement, and its use in general air, water, and soil remediation [8]. Mechanisms such as physisorption, complexation, precipitation, ion exchange, and electrostatic interaction are involved in the removal of pollutants from aqueous solutions using biochar. Biochar with a high surface area and pore volume exhibits a higher metal-ion philicity because it can be physically The accumulation of pesticides and their derivatives is becoming more prevalent, due to the rising population and rapid industrialization. As much as 80% to 90% of pesticides applied to crops in agricultural fields affect non-target life forms; they can relocate or volatilize from the treated area to pollute the air and soil and negatively affect non-target plants. The leaching of these accumulated pesticides leads to the contamination of groundwater and soil [3].
During the last few decades, the separation of pollutants from water systems and soil via several methods has been developed and successfully applied. Recently, technologies such as membrane filtration, ion exchange, and chemical precipitation have been utilized in real-life applications to remove pollutants such as metal ions from polluted areas. Chemical precipitation is a frequently used method for treating heavy metals because it is simple, inexpensive, and effective. However, chemical precipitation results in secondary pollution and eventually leads to additional difficulties in cleaning up the trace contaminants from large areas. Ion-exchange resin offers fast kinetics and is highly efficient for pollutant removal. However, the need for an acidic environment restricts their application in various contexts. Membrane filtration-based technologies can remove toxic substances with high efficiency, but the manufacture of membrane material is usually very complex and at a high cost. Conventional pollutant remediation methods are not eco-friendly and produce toxic chemical sludge. Therefore, there is a serious need to develop efficient and sustainable technologies for remediating toxic environmental pollutants.
Biochar is a carbonaceous material produced through the thermal treatment of different types of biomass, such as crop residues and biosolids [4,5]. Biochar production can be achieved via various processes, including slow or fast pyrolysis, flash carbonization, gasification, hydrothermal carbonization, torrefaction, etc. [6]. The key goal when designing the synthesis of the biochar is that the final material should possess high porosity, a large specific surface area, and elevated surface chemistry heterogeneity, as with oxygencontaining functional groups and minerals [7]. The physicochemical properties of the final obtained biochar can also be tuned by altering the microstructure. Such characteristics encourage biochar's rising application in (waste) water treatment, soil improvement, and its use in general air, water, and soil remediation [8]. Mechanisms such as physisorption, complexation, precipitation, ion exchange, and electrostatic interaction are involved in the removal of pollutants from aqueous solutions using biochar. Biochar with a high surface area and pore volume exhibits a higher metal-ion philicity because it can be physically entrapped within the pores on its surface [9]. The negatively charged surfaces of biochar can adsorb positively charged metal ions via electrostatic attraction. Compared to other adsorbents/microbial supports, biochar is a low-cost option and a promising candidate for pesticide and heavy metal treatment.
Bioremediation by microorganisms is considered a green technology that is acceptable to the general public. Microorganisms can bioadsorb, bioaccumulate, or biotransform the pollutants permanently at a low operating cost and without the generation of harmful secondary products [10]. Bioremediation can be effective even for contaminants in low concentrations that cannot otherwise be removed by chemical (e.g., incineration) or physical methods. According to some studies, microbial remediation has also been combined with other physical and chemical treatment methods. Hence, bioremediation reduces the health hazards to workers [11]. The microbial degradation of harmful and recalcitrant pesticides is efficient, cost-effective, and eco-friendly, with minimum hazards. The microbial consortia used for remediation have the additional advantage of promoting plant growth in the contaminated site.
Microorganisms can rapidly mutate and evolve in order to withstand environmental stress. The diversity and metabolic activity of the microorganisms are influenced by the presence of heavy metal ions and/or metalloids, which compels the microorganisms to develop resistance systems for overcoming this toxic metal ion stress. Furthermore, microorganisms convert toxic metal ions into inactive forms and can thus be utilized for bioremediation. Pesticide bioremediation involves biodegradation and biotransformation. In biodegradation, biological reactions modify the compound's chemical structure, decreasing its toxicity.
Microorganism immobilization on biochar is an efficient technology for treating wastewater and soil pollutants [12]. However, less information is available about the degradation of antibiotics, pesticides, heavy metals, PAHs, and other macromolecular organic pollutants that are immobilized on biochar by the microorganisms. Such pollutants are usually remediated via chemical methods or photocatalysis using biochar, which results in the production of free radicals that pose an ecotoxicological risk. As a result, biodegradation is becoming more important, and it is necessary to further investigate unexploited microorganisms for immobilization technology, based on biochar for pollutant biodegradation. However, biochar may cause toxicity to microorganisms according to their particle size [13]. To reduce this toxicity, the appropriate size of biochar must be chosen as a carrier for microorganism immobilization. Therefore, this work will evaluate the properties, influencing factors, strategies of immobilization, and removal efficiency of microbial cell-immobilized biochar (MCB) for the remediation of heavy metals and pesticides. The mechanisms involved in the bioremediation process will be explored. In the published research into the increasing environmental pollution caused by heavy metals and pesticides, much importance is given to remediation techniques. For biological remediation, there are many articles that discuss the role of microorganisms as an effective agent for the remediation of heavy metals and pesticides, as well as the role of biochar as an excellent adsorbent for the above pollutants. However, this review article focuses on the emerging role of biochar as an immobilization support for microbial cells.
Biochar Production, Properties, and Characterization
In general, biochar is a carbon-rich material derived from biomass (such as wood, manure, or leaves) upon thermal treatment at high temperatures in a closed container with minimal or in the absence of air [14]. Various processes, such as pyrolysis, gasification, and hydrothermal carbonization, are applied in biochar generation [15]. Biochar uses include (but are not limited to) carbon capture and storage, capacitive deionization, the Fenton process, microbial fuel cell electrodes, and electrochemical storage [16][17][18].
Biochar has been well established as a low-cost adsorbent that has adsorption capacities similar to carbon-based adsorbents, such as activated carbon, porous graphitic carbon nitride, graphene oxide, etc. The most crucial benefits are: (a) low cost of production, Molecules 2023, 28, 719 4 of 28 (b) porous structure, (c) simple fabrication on a large scale, (d) eco-friendly nature promoting the cycle of the (bio)economy, (e) multiple surface-functional groups, especially oxygen-containing groups (thus enabling both hydrophobic and polar interactions), (f) ease of modification, etc. [19,20]. In addition, the preference for biochar as a catalyst support for photocatalysis and Fenton/photo-Fenton processes has become prevalent, due to its low cost and high surface area characteristics. The aromatic and other hetero-atom-containing functional groups that are present in biochar also provide moieties that are capable of electron transfer and facilitate the faster and more efficient degradation/reduction of pollutants because of electron delocalization and photo-induced e − /h + pairs separation.
Biochar Production
Biochar production usually involves biomass collected from various plant/animal sources or wastewater sludge and thermal treatment using oxygen-deficient conditions, particularly pyrolysis. For instance, plant sources include olive pomace and rapeseed straw cereal waste, whereas animal sources include crustacean shells and animal manure [21][22][23][24][25]. Additionally, municipal wastewater sludge has also been used as biomass for biochar production [26]. The basic composition of biochar predominantly comprises amorphous phases and graphene sheets, as well as various aliphatic cyclic and aromatic groups as a matrix. The temperature of the treatment and the biomass source influence the final physicochemical features. For example, fibrous biomass sources such as wheat/rice straw generate tubular structures [27]. In contrast, the usage of sludge biochar prevents the formation of such structures in the biochar matrix [28].
Pyrolysis in oxygen-free conditions comprises the decomposition of lignocellulosic material, volatile matter release, and the reduction of carbonaceous material for plant biomass [29]. The types of pyrolysis include slow, fast, microwave-assisted, hydro-and copyrolysis. Slow pyrolysis operates for hours at lower temperature conditions (300-700 • C), resulting in a higher output percentage of biochar content compared to fast pyrolysis with lower residence time (<2-5 s), higher temperature conditions, and a lower output percentage of biochar. Increasing the temperature can lead to higher carbon content, alkalinity, and elevated specific surface area. In contrast, higher residence time can increase the specific surface area, due to prolonged temperature application.
Variations in high-temperature processes have also been tested in the context of biochar production. Microwave-assisted pyrolysis for biochar generation has also been demonstrated, with variations in absorbable power observed for biochar property analysis, with the demonstrated advantages of larger surface area and improved porosity characteristics [22]. Hydro-pyrolysis is usually conducted within a temperature range of 250-550 • C, with hydrogen gas application, ensuring the hydrocracking of the biomass [19]. Co-pyrolysis involves multiple biomass sources for biochar pyrolysis. The resultant physicochemical properties mainly depend on the biomass sources' blending ratios and pyrolysis temperature, improving the biochar sample's pore structure [30]. Gasification is another method of generating biochar in the presence of steam/oxygen at 750-900 • C, with the products being syngas and a low biochar yield. Torrefaction is conducted under oxygendeficient conditions similar to those for biochar, apart from a temperature of 200-300 • C and a residence time of less than 30 min. Another method explored extensively for biochar production is hydrothermal carbonization, with an operating temperature range from 160 to 800 • C (preferably at lower temperatures) in the presence of water. The low-temperature environment results in higher O/C and H/C content, along with the creation of functional groups on the biochar surface; the process yields a low aromaticity level and low-porosity biochar (hydrochar). The conversion of the non-carbonized (amorphous) part of the biomass into a carbonized form can be enhanced by increasing the pyrolysis temperature, which also increases the aromaticity, π electron availability, etc. [30]. Both the negative effect of pore-size thermal shrinkage due to the collapse of micropore walls and the positive effect of pore-size increment due to the removal of volatile matter can be observed with increasing temperature conditions. Increasing the pyrolysis temperatures also decreases the biochar's stability in terms of chemical oxidation resistance [31].
Biochar: Physicochemical Properties and Characterization
Several characterization analyses can be conducted to elucidate biochar's physical and chemical properties. The proximate analysis involves the quantification of ash, fixed carbon, volatile matter, and moisture. High ash and fixed carbon contents are good indicators of high adsorbent capacity. Ultimate analysis, i.e., the quantification of C, H, N, and O composition in biochar samples, especially the H/C ratio and O/C and (O + N)/C ratios, is an indicator of the aromaticity and polarity of biochar [26].
The textural features, with an emphasis on the sizes and the volume of the pores and the specific surface area (S BET ), are usually estimated via N 2 sorption tests at 77 K, using the Barrett-Joyner-Halenda (BJH) or density functional theory (DFT) methods for the pore analysis and the Brunauer-Emmett-Teller (BET) theory for the S BET . The definition of pore size category (micro-, meso-, and macropores) decides the interaction ability of biochar with the required moiety. For instance, biochar systems with microporous structures would show the lower adsorption capacity of higher molecular-weight pesticides, although a higher one is needed for metal cations [15].
Surface pH analysis, zeta potential, and electrical conductivity can define the range in which biochar-pesticide and biochar-metal ion interactions are maximized. The graphitization and alkalinity of the produced char increase at higher pyrolysis temperatures [32]. Surface functional group analyses, such as cation exchange capacity, Boehm titration, and humic substance analysis, are also used to evaluate the biochar's adsorption capacity and microbial support. Fourier transform infrared spectroscopic (FTIR) analysis also provides insight into the biochar matrix's multiple bond formation, with additional information on post-adsorption studies. The solid-state C-nuclear magnetic resonance technique can be used to study the relative abundance of the functional groups and the aliphatic and aromatic hydrocarbon contents [33].
The morphological and structural properties can be explored by scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), and atomic force microscopy (AFM). The surface chemistry can be analyzed by IR, Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), potentiometric titration, and Boehm titration. It is always crucial to determine the surface pH and the point of zero charge, since they play a key role in the adsorption performance and activity when biochar is used as an adsorbent in aqueous phases.
The most important techniques for the physicochemical characterization of biochar are presented in Figure 2.
Biochar as Adsorbents
Heavy metals and pesticides can be directly adsorbed onto the biochar's surface. Modifying the outer surface of biochar via activation by tuning the chemical heterogeneity
Biochar as Adsorbents
Heavy metals and pesticides can be directly adsorbed onto the biochar's surface. Modifying the outer surface of biochar via activation by tuning the chemical heterogeneity and/or by anchoring decorating different active species can lead to elevated and selective adsorption efficiency, exceptional stability, easy separation efficiency, and better recyclability [19]. Modification, including physical/thermal activation such as steam (for -OH functional group increment) and CO 2 activation, ball milling and sonication/ultrasonication, acid treatment (for deashing and demineralization) and base treatment, functional group activation, such as amine-functionalization, impregnation with metal oxides, doping, electrochemical treatment, plasma treatment, etc., can enhance the properties of biochar as an adsorbent [34].
Regarding the analysis of metal ions or pesticide removal using biochar-based materials, Langmuir and Freundlich's isotherm models are the most established ones. In general, a Langmuir versus Freundlich isotherm comparison explains monolayer adsorption vs. mono/multilayer adsorption, even though this approach is not absolutely correct in the case of studying adsorption in aqueous phases. Other isotherm models/approaches, such as the Jovanovich, Elovich, and Dubinin-Radushkevich (D-R) models, are also used in order to present an additional understanding of the role of adsorption conditions [35,36]. In addition, pseudo-first-order and pseudo-second-order models are the most widely applied models for kinetic studies for biochar-heavy metal/pesticide systems [37].
Removal of Metal Ions
The application of adsorbents for the removal of heavy metal ions involves physical and/or chemical adsorption via electrostatic interactions, ion exchange, complexation, reactions that have taken place on the material's surface, and/or precipitation [38]. When interacting with biochar, some metal ions undergo reduction and oxidation reactions, precipitation, and co-precipitation [39].
Multiple experimental condition parameters can affect the adsorption and removal capacity. The elevated adsorption of metal ions can be due to an increase in the specific surface area of the biochar as a result of optimizing the synthetic protocol, for instance, modifying the pyrolysis temperature [32]. A high pH directly affects the adsorbent's surface due to protonation, thus competing with metal ion adsorption [40]. Conversely, in alkaline pH conditions, hydroxy-complex formations can compete with other ions and impede adsorption [21]. Preferably, the point of zero-charge pH should be in the acidic region to efficiently adsorb metal ions and form complexes with a negative surface charge [41]. Cation-exchanging capacity also plays a crucial role in metal ion adsorption. For instance, Ma et al. [42] discovered that cation exchange significantly contributed to removing Cu 2+ from lobster-shell-derived (HCl-treated) biochar, with 53-74% removal contributed by the cation exchange.
Biochar surface modifications are primarily conducted to improve the adsorption efficiency, and some of them are summarized in Table 1. A zirconium and iron composite with sludge biochar was generated to increase As 5+ adsorption via complexation. The Zr-Fe biochar composite had a maximum adsorption capacity of 62.5 mg/g, compared to the pristine biochar capacity of 15.2 mg/g. The probable mechanism was suggested as the inner-sphere complexation of As 5+ on the Zr-O-Fe surface [36]. Khan et al. [43] studied MoS 2 -modified magnetic biochar with a maximum adsorption capacity of 139 mg/g, and hypothesized the presence of complexation, cation exchange, and Cd-π interactions. The deashing of biochar with acid solutions and potassium acetate improved lead adsorption, due to the pore size increment (unblocking SiO 2 particles out of biochar) and complexation of Pb 2+ and C=C (π-electrons) [40].
Adsorption/Removal of Pesticides
Studies indicate that increased pesticide concentration and adsorption time has an asymptotic effect on adsorption capacity, whereas the adsorption capacity is enhanced by the increases in biochar concentration. The common mechanisms for pesticide adsorption onto biochar are the hydrophobic effect, π-π electron donor-acceptor interaction, pore filling, electrostatic interactions, ionic bonding, and H-bonding [26,52].
Several studies regarding the adsorption/removal of pesticides by biochar have been evaluated in Table 2, wherein the parameters of biochar pyrolysis temperature and surface modifications have been compiled, along with the adsorption capacity values. The pyrolysis temperature has a similar effect on biochar-pesticide adsorption as on biochar-heavy metal adsorption. The adsorption of carbendazim on dewatered sludge biochar was at a maximum at 700 • C, owing to the increased surface area and the increment in the partition coefficient [26]. Pore size governs the definitive adsorption capacity for pesticide-biochar interaction. Dichlorvos and pymetrozine had molecular sizes that were comparable to pore diameter; thus, adsorption was facile in both cases [53]. A decrease in the original biochar's H/C and O/C atomic ratios is expected to enhance the π-π electron donor-acceptor interactions, contributing to the sorption of certain pesticides, such as oxytetracycline and carbaryl [24]. Binh and Nguyen [52] concluded that a pH of 2 is a more favorable condition for the adsorption of 2,4 dichlorophenoxy acetic acid on corn-cob biochar based on the electrostatic interactions. In addition to the inherent functional groups and mechanisms involved in metolachlor adsorption onto biochar, Liu et al. [54] incorporated fulvic acid and citric acid into walnut-shell biochar that augmented the functional groups with oxygen, as shown in Figure 3. The removal capacity was also observed to decrease after 3 cycles in the metolachlor-simulated sewage biochar system. Several studies regarding the adsorption/removal of pesticides by biochar have been evaluated in Table 2, wherein the parameters of biochar pyrolysis temperature and surface modifications have been compiled, along with the adsorption capacity values. The pyrolysis temperature has a similar effect on biochar-pesticide adsorption as on biocharheavy metal adsorption. The adsorption of carbendazim on dewatered sludge biochar was at a maximum at 700 °C, owing to the increased surface area and the increment in the partition coefficient [26]. Pore size governs the definitive adsorption capacity for pesticide-biochar interaction. Dichlorvos and pymetrozine had molecular sizes that were comparable to pore diameter; thus, adsorption was facile in both cases [53]. A decrease in the original biochar's H/C and O/C atomic ratios is expected to enhance the π-π electron donor-acceptor interactions, contributing to the sorption of certain pesticides, such as oxytetracycline and carbaryl [24]. Binh and Nguyen [52] concluded that a pH of 2 is a more favorable condition for the adsorption of 2,4 dichlorophenoxy acetic acid on corncob biochar based on the electrostatic interactions. In addition to the inherent functional groups and mechanisms involved in metolachlor adsorption onto biochar, Liu et al. [54] incorporated fulvic acid and citric acid into walnut-shell biochar that augmented the functional groups with oxygen, as shown in Figure 3. The removal capacity was also observed to decrease after 3 cycles in the metolachlor-simulated sewage biochar system. The fulvic acid-and citric acid-modified biochar adsorption mechanism for metolachlor in water Reprinted/adapted with permission from Ref. [54]. Copyright 2021, Elsevier. . The fulvic acid-and citric acid-modified biochar adsorption mechanism for metolachlor in water Reprinted/adapted with permission from Ref. [54]. Copyright 2021, Elsevier.
Biochar as a Bioremediation Catalyst Support
The physicochemical properties of biochar that enable it to be an effective catalyst support include its large surface area, multi-scale porous structure, and surface functional group. Chen et al. [57] studied the volatilization of Hg 2+ using the Pseudomonas strain, DC-B1, with biochar. The combined application of biochar and microbial strain resulted in the greatest Hg removal. Qiao et al. [58] demonstrated the stimulation of the microbial reduction of As 5+ and Fe 3+ , using oil palm fiber-derived biochar in synergy with soil microbes extracted from paddy for both studies. Biochar amended microcosm possessed a higher As 5+ concentration than the control, indicating that biochar had an affinity to As 5+ and Fe 3+ . Both moieties were reduced in the biochar-amended microcosms since microbes drove the reduction reactions, and biochar behaved similarly to an electron shuttle. Qiao et al. [59] summarized the As 5+ reduction with biochar and lactate. This reduction resulted in the identification of three ways: (i) Fe 3+ reduction by microbial cells facilitated As 5+ release; (ii) expression of As 5+ -respiring gene transcripts in dissimilatory As 5+ -reducing bacteria; (iii) the functioning electron transfer between the metal and As 5+reducing bacteria.
Biochar is often employed as a good carrier in improving the photocatalytic activity of metal oxides. As a stable and inexpensive carbonaceous material, biochar effectively reduces the recombination rate of photogenerated electron-hole pairs, due to its excellent conductive property. An et al. [60] developed biochar-supported α-Fe 2 O 3 /MgO composites for photocatalytic degradation of organophosphorus pesticides and obtained a degradation efficiency of 90% in 80 min. Huang et al. [61] utilized pristine and manganese ferritemodified biochar for Cu removal, confirming the role of biochar being principally an oxide carrier instead of an adsorbent. In addition, a preference for biochar as a carrier for photocatalysis and Fenton/photo-Fenton processes has been prevalent, due to its low cost and high surface area characteristics. The utilization of lignin-biochar as a catalyst support for LaFeO 3 in the catalytic photo-Fenton process had a positive effect on the degradation efficiency of pollutants, owing to enhanced adsorption capacity, a reduction in the charge transport resistance between LaFeO 3 and lignin-biochar, and the presence of oxygen-containing functional groups [62].
Several studies have been conducted for enzyme-immobilized biochar, particularly with laccase utilized as an enzyme to degrade pollutants [33]. The basic biochar-enzyme immobilization techniques are adsorption and covalent bonding. Comparatively, fewer instances of enzyme-biochar systems for the degradation of pesticides have been studied. Wang et al. [63] used laccase-immobilized biochar to degrade 2,4-dichlorophenol and obtained 64.6% degradation. The immobilized laccase improved the cation exchange capacity, organic matter content, stability, and catalytic degradation effect. A general outline for adsorption and the removal mechanisms for metals and pesticides via biochar systems are shown in Figure 4.
Role of Microorganisms in the Removal of Metal Ions and Pesticides
The surface area of microorganisms exhibits higher biological activity relative to their volume, resulting in greater interaction with their immediate environment. Thus, they can adapt and survive in polluted areas with the subsequent removal or detoxification of the pollutant [64]. The microorganisms use different strategies for their survival, including surface adsorption, micro-precipitation, extracellular or intracellular sequestration, reduction, enzymatic degradation, etc. Bioremediation is possible only when microbial activity and growth are allowed by environmental conditions. In certain situations, environmental factors can be altered to allow microbial population growth to eliminate
Role of Microorganisms in the Removal of Metal Ions and Pesticides
The surface area of microorganisms exhibits higher biological activity relative to their volume, resulting in greater interaction with their immediate environment. Thus, they can adapt and survive in polluted areas with the subsequent removal or detoxification of the pollutant [64]. The microorganisms use different strategies for their survival, including surface adsorption, micro-precipitation, extracellular or intracellular sequestration, reduction, enzymatic degradation, etc. Bioremediation is possible only when microbial activity and growth are allowed by environmental conditions. In certain situations, environmental factors can be altered to allow microbial population growth to eliminate contaminants [11]. As shown in Figure 5, various factors influence microbial degradation: The pH can affect bioremediation by changing metal bioavailability; for instance, a decrease in soil pH value generally causes an increase in metal bioavailability [65]. This is because, at lower pH, the exchangeable capacity between metal cations and H + on the surface of soil particles is more prominent than at higher pH. Additionally, an optimum pH is essential for microbial growth, and some microbial degradation processes can be inhibited at an extreme pH. Temperature is another crucial factor influencing the bioremediation of metals and pesticides [65]. The solubility of these contaminants is increased at higher temperatures, which leads to their increased bioavailability. The physical nature and chemical composition of several organic pollutants and their adsorption-desorption mechanism are governed by temperature. Temperature also influences microbial growth, activity, and degradation potential. Furthermore, the soil moisture content is another parameter that affects the bioremediation process. A low soil moisture content limits the growth and metabolism of microorganisms, while high values can reduce soil aeration.
(ii) Type of microorganism and degradation capacity
The microorganism that is selected for biodegradation should be able to survive in a high-contamination environment and should be evaluated first for its degradation capacity before employing it for in situ remediation. The survival of these strains can be ensured by providing favorable growth conditions. It is also important to note that microbial strains selected for pollutant removal may need to meet certain ecological requirements. One such requirement is that the strains should be non-pathogenic. For instance, Staphylococcus aureus, as a typical pathogen, was resistant to many antibiotics and showed high bioremediation efficiency for heavy metals such as Cr and U through bioprecipitation [66]. However, certain metabolites that formed during the degradation of contaminants can be toxic. Therefore, deeper investigations of ecological security and the metabolic functions of microbial cells are indispensable before their possible application in environmental pollution control.
(iii) Bioavailability of the contaminants The bioavailability of the contaminants can be defined as the fraction of a contaminant in a specific environment that is either adsorbed or degraded by the microbial cells within a given time. The control of bioavailability is dependent on the diffusion, uptake, and desorption of the contaminants. The slow mass transfer of contaminants into degrading microbes reduces their bioavailability. The significance of bioavailability depends very much on the properties of the pollutant, microorganism, and characteristics of the contaminated site [11].
(iv) Aerobic or anaerobic operating conditions Depending on the type of organism and contaminant, bioremediation can be either aerobic or anaerobic. Most bioremediation systems work under aerobic conditions, but to effectively degrade the recalcitrant molecules, it is better to run the microbial degradation tests under anaerobic conditions. Apart from the abovementioned factors, the properties of the contaminated site (soil type, soil porosity, soil nutrients) and the properties of the contaminants (structure, hydrophobicity, recalcitrance, toxicity, solubility, and leaching ability) are also important in bioremediation.
Depending on the type of organism and contaminant, bioremediation can be either aerobic or anaerobic. Most bioremediation systems work under aerobic conditions, but to effectively degrade the recalcitrant molecules, it is better to run the microbial degradation tests under anaerobic conditions. Apart from the abovementioned factors, the properties of the contaminated site (soil type, soil porosity, soil nutrients) and the properties of the contaminants (structure, hydrophobicity, recalcitrance, toxicity, solubility, and leaching ability) are also important in bioremediation.
Removal of Heavy Metals Using Microorganisms
The removal of heavy metal ions by microorganisms is considered economical and sustainable. Any environmental stress can be withstood by microorganisms through rapid mutation and evolution, leading to toxic heavy metal resistance. They can sequester heavy metal ions, either intracellularly or extracellularly. Additionally, microorganisms can transform and reduce the metal ions to inactive forms. Table 3 summarizes the microorganisms used for various metal ion remediation conditions in recent years.
Removal of Heavy Metals Using Microorganisms
The removal of heavy metal ions by microorganisms is considered economical and sustainable. Any environmental stress can be withstood by microorganisms through rapid mutation and evolution, leading to toxic heavy metal resistance. They can sequester heavy metal ions, either intracellularly or extracellularly. Additionally, microorganisms can transform and reduce the metal ions to inactive forms. Table 3 summarizes the microorganisms used for various metal ion remediation conditions in recent years.
The factors influencing heavy metal remediation by microbes generally include pH, temperature, biomass concentration, the presence of other pollutants, etc. The inherent pH of the system defines the charges of the surface functional groups present on the microbial surfaces; pH in an unsuitable range may affect microbial growth. This shows pH to be an essential parameter in the degradation and removal of heavy metals by live biomass [81]. The pH also has an effect on the solubility of metal ions in the microbe-heavy metal system. A decrease in soil pH leads to an increase in the bioavailability of metals, thereby resulting in higher biosorption efficiency, as studied by Zhang et al. [73]. Another essential parameter in microbial growth and proliferation is the system's ambient temperature. With an increase in temperature, the solubility of metal ions increases; thus, the bioavailability of metals also increases [81]. High biomass or sorbent concentration will increase the overall biosorption efficiency, but any interference between binding sites reduces the specific metal ion uptake. The removal or adsorption of a particular heavy metal by microorganisms can also be positively or negatively affected by the co-existence of other metal ions.
The Mechanism of Heavy Metal Removal by Microorganisms
Microorganisms can adopt several mechanisms in order to survive in heavy-metal toxicity conditions. These mechanisms are depicted in Figure 6 and include biotransformation, extracellular polymeric substances secretion, metallothionein synthesis, etc. Heavy metal degradation by microorganisms can be described in two ways: biosorption and bioaccumulation.
Biosorption is the reversible physicochemical interaction of living (or dead) biomass or biomass-secreted products that act as biosorbents with sorbate molecules (e.g., metal ions). It was previously categorized as metabolism-dependent and metabolism-independent biosorption. Recently, the former has been widely accepted as bioaccumulation (also called active biosorption), and only the metabolism-independent processes are considered to be biosorption [82]. A metabolism-independent mechanism occurs passively on the dead or the living biomass cell surface. However, the biosorption of metal ions carried out by dead biomass is superior to that carried out by living cells. Cheng et al. [78] studied the biosorption of Cd 2+ in the living and dead cells of the microalgae Chlorella vulgaris. The dead algal biomass removed 96.8% of the total cadmium, while the live algal biomass achieved 95.2% of cadmium adsorption. The steps involved in toxic heavy metal biosorption include binding the metal ions to various extracellular functional groups present on the microbial cell wall, via surface precipitation, chemical bonding (complexation/chelation), adsorption, or ion exchange. Physical adsorption depends on intermolecular or inter-ionic attraction forces. Complexation or chelation occurs due to the dative covalent bonds between metal ions, surface functional groups, and the ligands of biomass. When metal ion concentrations are higher than the solubility limit, surface precipitation or micro-precipitation has been observed. The exchange involves electrostatic interaction between the metal cations and the negatively charged functional groups on the cell surface; the interchange of the cations resulted in the metal ion being bound to the surface [82,83]. Surface-binding is found to be the principal phenomenon governing the biosorption of metal ions [84]. Physical modifications have been suggested to provide a cumulative effect on the biosorption capacity of the microorganisms by removing surface impurities or through the production of metal-binding sites. Li et al. [74] investigated the biosorption ability of a lactic acid bacterium, Weissella viridescens ZY-6, for Cd 2+ removal from the aqueous solution, and achieved a 69.45-79.91% removal of Cd 2+ from three kinds of juices: tomato, apple, and pear juices.
The extracellular sequestration of metal ions often occurs due to various biological structures produced by microbial cells, including extra-cellular polymeric substances, siderophores, glutathione, and biosurfactants. Under heavy metal stress, microorganisms often secrete extra-cellular polymeric substances or exopolysaccharides (EPS) as a protective response. EPS are constituted of proteins, lipids, complex carbohydrates, nucleic acids, uronic acid, humic acid, etc., which prevent the entrance of heavy metals into the cell [68,85]. Generally, EPS contain negatively charged functional groups and can interact electrostatically with heavy metals, resulting in the immobilization of the metal ions within the EPS. Some examples include the accumulation of Pb 2+ and Zn 2+ in the soluble EPS secreted by Oceanobacillus profundus KBZ 3-2 [68], and Pb 2+ adsorption onto EPS of Enterobacter sp. FM-1 [69] and Cd 2+ adsorption onto the EPS secreted by a living cyanobacteria, Synechocystis sp. PCC6803 [86]. Siderophores are secreted by microbes and act as metal chelators, with an extreme affinity for ferric iron. They can reduce the metal's bioavailability and toxicity by binding metal ions with variable affinities that have a similar chemistry to that of iron [87].
Biosurfactants are amphiphilic compounds that are produced extracellularly by microorganisms for the solubilization, desorption, complexation, and mobilization of pollutants in solutions. The induction of biosurfactants in microbe-heavy metal systems facilitates the extracellular sequestration and formation of biosurfactant-metal complex [88]. Rhamnolipids produced by Pseudomonas aeruginosa showed 53% As, 90% Cd, and 80% Zn extraction capacity from contaminated soil [89]. Ayangbenro and Babalola [90] observed that a lipopeptide biosurfactant generated by Bacillus cereus NWUABO1 could remove 69% of Pb, 54% of Cd, and 43% of Cr from the soil. Several microorganisms, including Pseudomonas sp., Bacillus subtilis, Candida tropicalis, Candida sp., Burkholderia sp., and Citrobacter freundii can produce biosurfactants, demonstrating heavy metal removal capacity [88]. Biosorption has been determined to be simple, fast, reversible, and inexpensive compared to bioaccumulation and can concentrate heavy metals, even from a very dilute aqueous solution. The advantageous properties of biosorption include the presence of multi-functional groups and the uniform distribution of binding sites on the cell surface, low operational cost, the absence of metal toxicity limitations, minimal preparatory steps, high efficiency and selectivity for metal ions, no production of secondary waste, and the possibility of the toxic heavy metal recovery and reusability of the biosorbent [84,91]. Several microbial strains have been identified to show multi-metal resistance and remediation abilities. Nokman et al. [92] isolated a Pseudomonas putida strain from effluent water generated from a tannery that exhibited resistance to Ag 2+ and Co 2+ and enhanced resistance to lead and chromium. Conversely, bioaccumulation is the metabolism-dependent active transportation of metal ions across the membrane into the living cell, as represented in Figure 6. The microorganisms selected for bioaccumulation should have specific properties, such as adaptation to the polluted environment, resistance to high loads of metal ions, and a mechanism of intracellular binding [93]. The mechanism consists of two steps; the first step is identical to biosorption and involves the attachment of heavy metals to charged functional groups on the cell surface. The second step is metabolism-dependent, relatively slow, and involves the penetration/transport of a metal-ligand complex into the cell membrane. The subsequent interaction of the complexes with intracellular metal-binding proteins (such as metallothionein and phytochelatins) occurs within the cell, leading to bioaccumulation [85]. Metallothioneins (MTs) help to regulate the intracellular metabolism of metals and protect against oxidative stress and toxic heavy metals [86,87]. Engineered recombinant E. coli expressed the Corynebacterium glutamicum metallothionein gene and achieved improved intracellular biosorption of Pb 2+ and Zn 2+ . Hu et al. [86] constructed a bio composite of immobilizing metallothionein, expressing Pseudomonas putida for the sorption of Cu 2+ . Similarly, phytochelatins are metal-binding proteins that are analogous to the metallothioneins produced from microalgae, which can also chelate and detoxify heavy metal ions intracellularly.
Molecules 2023, 28, x FOR PEER REVIEW 13 of 29 will increase the overall biosorption efficiency, but any interference between binding sites reduces the specific metal ion uptake. The removal or adsorption of a particular heavy metal by microorganisms can also be positively or negatively affected by the co-existence of other metal ions.
The Mechanism of Heavy Metal Removal by Microorganisms
Microorganisms can adopt several mechanisms in order to survive in heavy-metal toxicity conditions. These mechanisms are depicted in Figure 6 and include biotransformation, extracellular polymeric substances secretion, metallothionein synthesis, etc. Heavy metal degradation by microorganisms can be described in two ways: biosorption and bioaccumulation. Biosorption is the reversible physicochemical interaction of living (or dead) biomass or biomass-secreted products that act as biosorbents with sorbate molecules (e.g., metal ions). It was previously categorized as metabolism-dependent and metabolismindependent biosorption. Recently, the former has been widely accepted as bioaccumulation (also called active biosorption), and only the metabolism-independent processes are considered to be biosorption [82]. A metabolism-independent mechanism occurs passively on the dead or the living biomass cell surface. However, the biosorption of metal ions carried out by dead biomass is superior to that carried out by living cells. Cheng et al. [78] studied the biosorption of Cd 2+ in the living and dead cells of the microalgae Chlorella vulgaris. The dead algal biomass removed 96.8% of the total cadmium, while the live algal biomass achieved 95.2% of cadmium adsorption. The steps involved in toxic heavy metal biosorption include binding the metal ions to various extracellular functional groups present on the microbial cell wall, via surface precipitation, chemical bonding (complexation/chelation), adsorption, or ion exchange. Physical adsorption depends on intermolecular or inter-ionic attraction forces. Complexation or chelation occurs due to the dative covalent bonds between metal ions, surface functional groups, and the ligands of biomass. When metal ion concentrations are higher than the solubility limit, surface precipitation or micro-precipitation has been observed. The exchange involves electrostatic interaction between the metal cations and the negatively charged functional groups on the cell surface; the interchange of the cations resulted in the metal ion being bound to the surface [82,83]. Surface-binding is found to be the principal
Removal of Pesticides Using Microorganisms
The major types of pesticides and persistent organic pollutants include insecticides, herbicides, and fungicides. As with heavy metals, the microbial remediation of these persistent pesticides is economical and sustainable, compared to physical or chemical removal processes. It involves the degradation of complex pesticide molecules into simpler inorganic chemicals. Table 4 includes the commonly used microorganisms for the removal of pesticides. Indigenous soil microbial consortia have been more effective for the microbial degradation of pesticides than the non-indigenous strains, as non-indigenous strains are exposed to pesticide-contaminated regions exhibiting unfamiliar conditions. Several studies have reported on the degrading ability of indigenous microbes. Some of them show organophosphate degradation by indigenous Kosakinia oryzae [94], herbicide glyphosate degradation by Providencia rettgeri [95], and herbicide atrazine remediation by indigenous microbial consortia [96]. Individual or mixed microbial cultures can degrade the various sources of pesticides. Single microbial cells abide by their metabolic pathways for pesticide degradation, whereas mixed microbial cultures can achieve the same result through coupled metabolic pathways [97]. Thus, pesticides can rapidly be degraded by applying the combined microbial consortia isolated from indigenous sites. However, certain recalcitrant pesticides have resilience against biodegradation by the indigenous microbial community. In such situations, bio-augmentation and biostimulation are considered promising approaches for the remediation of contaminated sites. Bio-augmentation involves the introduction of specific exogenous microbes to improve the degradative capacity of the contaminated sites. The two main strategies of bio-augmentation are autochthonous bio-augmentation, where the microbes are isolated from the same site and then re-injected, and allochthonous bio-augmentation, where the microbes are cultured from another site [107]. In one study, bio-augmentation with Paenarthrobacter sp. W11 significantly accelerated the degradation rate of atrazine in soil and dampened its toxic effect on wheat growth [108]. The success of bio-augmentation strategies depends on several factors, including the selection of appropriate microorganisms, the target pollutant's bioavailability, and the inoculum's survival capability in the toxic environment [11,109].
Bio-stimulation can be performed by providing the necessary nutrients or electron acceptors, such as oxygen or nitrate, to promote the proliferation of indigenous microbes. Aldas-Vargas et al. [110] investigated the biodegradation of herbicides, namely, mecopropp and 2,4-dichlorophenoxyacetic acid (2,4-D), in groundwater. They concluded that biostimulation with oxygen and dissolved organic carbon had the potential for field application. Raimondo et al. [111] bio-augmented lindane-contaminated soil with actinobacteria (mixed culture) and bio-stimulated it with sugarcane filter cake, further noticing enhanced lindane removal, along with microbial cell counts and enzyme activities.
The removal of pesticides depends, firstly, on the optimal conditions of the biomass, its survival and activity, and, secondly, on the pesticide's chemical structure, along with several biotic and abiotic factors, such as suitable microbial strains, nutrient availability, salinity, pH, temperature, etc. [112]. In the case of the in situ remediation of soil contaminated by the extensive use or overuse of pesticides for agricultural purposes, the growth of pesticide-degrading soil microbes depends on the soil characteristics [11].
Mechanisms Involved in Pesticide Removal by Microorganisms
There are several mechanisms by which microorganisms transform pesticides into their non-toxic forms in a contaminated site. Some include the surface adsorption, enzymatic degradation, or co-metabolism of the pesticide molecules, as depicted in Figure 7. Adsorption of the pesticide molecules is categorized as a passive process and involves the direct interaction of molecules with the microbial cell surface. As a result, the efficiency of pesticide adsorption by microorganisms is primarily determined by the available surface-active groups. The ultimate result of adsorption is the reduced mobility of the toxic pesticides. The extent of removal and the degradation efficiency are influenced by various components, such as the charge, polarity, solubility, volatility, and solubility of the pesticide molecules. Extra-cellular polymeric substances (EPS) and biosurfactants produced by the microorganisms also aid in the removal of pesticides. EPS can be produced by the microbial cell as a byproduct of pesticide degradation. This approach can have two benefits: (i) the reduction of excess toxic pesticides, and (ii) the production of EPS, which can have further environmental applications. Gupta et al. [113] observed 98% carbofuran degradation within 96 h by Cupriavidus sp. with simultaneous EPS production. Satapute and Jogaiah [114] reported that surfactin, a biosurfactant produced by a bacterial strain, could degrade 91% of difenoconazole.
Molecules 2023, 28, x FOR PEER REVIEW 17 of 29 02 cell degradation was completed from 99.0% to 95.0% within 4 h. However, the extracted enzymes can be affected by solution properties, such as pH, temperature, etc. Depending on the environmental factors, enzymes may lose their degradation potential due to varied ambient conditions [117]. Oxidation, hydrolysis, alkylation, and dealkylation reactions have been predominantly observed in the microbial degradation process [118]. Some studies that have reported enzymatic degradation are on cypermethrin by esterase and laccase [119], carbendazim by carbendazim hydrolase [120], malathion by phosphotriesterase [121], and isoproturon, procymidone, chlorpyrifos, dichlorophos, and monocrotophos by laccase [122][123][124]. The enzymatic biodegradation mechanism of pesticides is often complex, and this diverse biodegradation pathway needs further investigation to understand enzyme involvement properly.
Challenges of Using Microorganisms as a Catalyst
The microbial degradation of metal ions and pesticides tends to be an appealing approach for bioremediation, even though certain challenges hinder their commercial application. These include: (i) the loss of microorganisms or reduced microbial survival because of the toxicity to microorganisms at a higher metal ion or pesticide concentration, (ii) reduced microbial proliferation, (iii) uneven microbial growth with high concentrations of the pollutant, (iv) the washing out of the microbial cells during the application, (v) the longer time required for the completion of the process, (vi) the presence of other co-existing metal ions and organics that can positively or negatively affect the remediation process.
Microbial immobilization on a support material can overcome the above drawbacks by fixing the free microbial cells to a specific carrier, either chemically or physically, and keeping them active for longer. An ideal carrier provides operational stability and cell protection from the toxic external environment, leading to efficient biodegradation. A support material retains the microbes and contributes to the sorption of the pollutants [125]. Hence, immobilizing the microorganism accelerates the pollutant's biodegradation capacity, enhances the robustness of the immobilized strains, and improves their tolerance to high pollutant concentrations. Microbial enzymes can catalyze the breakdown of pesticides. The enzymatic degradation processes may include an alteration in the structural components, the removal of undesirable pesticide properties, oxidation, and reduction [115]. Dash and Osborne [116] investigated monocrotophos degradation by Bacillus aryabhattai (VITNNDJ5) instead of the bacterial enzyme. The enzymatic degradation of pesticides can either be performed by intracellular enzymes that are present in the microbial cell or by extracting the enzymes capable of degradation from the cells. Sirajuddin et al. [100] isolated the E. coli IES-02 strain from a site contaminated with the organophosphate malathion, and the strain showed efficient degradation, utilizing it as the sole carbon source. They also purified carboxylesterase enzyme from the IES-02 strain and achieved 81% malathion degradation under optimized conditions within 20 min, whereas the IES-02 cell degradation was completed from 99.0% to 95.0% within 4 h. However, the extracted enzymes can be affected by solution properties, such as pH, temperature, etc. Depending on the environmental factors, enzymes may lose their degradation potential due to varied ambient conditions [117]. Oxidation, hydrolysis, alkylation, and dealkylation reactions have been predominantly observed in the microbial degradation process [118]. Some studies that have reported enzymatic degradation are on cypermethrin by esterase and laccase [119], carbendazim by carbendazim hydrolase [120], malathion by phosphotriesterase [121], and isoproturon, procymidone, chlorpyrifos, dichlorophos, and monocrotophos by laccase [122][123][124]. The enzymatic biodegradation mechanism of pesticides is often complex, and this diverse biodegradation pathway needs further investigation to understand enzyme involvement properly.
Challenges of Using Microorganisms as a Catalyst
The microbial degradation of metal ions and pesticides tends to be an appealing approach for bioremediation, even though certain challenges hinder their commercial application. These include: (i) the loss of microorganisms or reduced microbial survival because of the toxicity to microorganisms at a higher metal ion or pesticide concentration, (ii) reduced microbial proliferation, (iii) uneven microbial growth with high concentrations of the pollutant, (iv) the washing out of the microbial cells during the application, (v) the longer time required for the completion of the process, (vi) the presence of other co-existing metal ions and organics that can positively or negatively affect the remediation process.
Microbial immobilization on a support material can overcome the above drawbacks by fixing the free microbial cells to a specific carrier, either chemically or physically, and keeping them active for longer. An ideal carrier provides operational stability and cell protection from the toxic external environment, leading to efficient biodegradation. A support material retains the microbes and contributes to the sorption of the pollutants [125]. Hence, immobilizing the microorganism accelerates the pollutant's biodegradation capacity, enhances the robustness of the immobilized strains, and improves their tolerance to high pollutant concentrations.
Microbial Cell-Immobilized Biochar for the Removal of Metal Ions and Pesticides
Bioremediation with free microbial cells is generally inefficient, due to the lesser amount of microbes utilized for degradation, microbial loss, and the inhibition of growth and functioning from indigenous microorganisms [126]. Immobilizing the microorganisms creates a safe environment for microbial cells to perform specific functions, such as highly efficient physiochemical sorption and microbial metabolism. Pollutant adsorption/binding on the carrier material allows the degrading cells to outcompete indigenous microbes, overcoming the limitations of using free cells for bioremediation [127]. Biochar has been a prominent carrier for microbial cell immobilization, due to its minimal toxicity and abundant generation. Immobilized microbes have commonly been observed for better remediation efficiency than pristine biochar or free cell [128].
Immobilization Methods
Biochar-immobilized microorganisms are produced through the adsorption of microbes on biochar, entrapment with the help of crosslinking materials, or a combination of both methods. Adsorption is a simple and inexpensive method for immobilizing microorganisms [129,130]. Adsorbed cells colonize the biochar after being transferred from a bulk solution to its surface. The adsorption technique involves physical interactions, such as van der Waals forces, ionic interactions, and hydrogen bonding between the surface functional groups of microorganisms and functional groups on the surface of carriers, particularly the oxygen-containing groups, such as carboxylic, phenolic, and sulphonate groups. Microorganisms have a low affinity for carriers; there will thus be a high rate of desorption of cells from carriers [125]. As a result, appropriate carriers with high cell-binding characteristics are required for improved remediation. With a relatively weak interaction between the carrier and microbial cells, immobilization does not affect the intrinsic structure of the original microbes if the adsorption method is utilized. As a result, this method is better suited for immobilizing viable cells and biodegrading pollutants in the laboratory. Entrapment is a standard method of physical immobilization that is irreversible and provides better stability of microbes than adsorption [126]. Due to the improved stability of the thus-prepared immobilized cells, the entrapment method is preferred and is exercised in industrial applications for pollution abatement.
Factors that Influence Bioremediation Using Immobilized Microorganisms
The effective pollutant removal capacity of MCB is affected by pollutant concentration and its bioavailability, the incubation time of the cell, and various parameters, such as temperature, pH, etc. The biochar-immobilized microorganism technology requires a thorough understanding of the best conditions for maximum contamination removal.
Initial pollutant concentration influences the removal of pollutants, wherein setting the initial pollutant concentration until the saturation point increases the adsorption capacity of the biosorbed pollutants per unit weight of MCB [131]. The bioavailability of pollutants is defined as the total amount of a contaminant that is either available or that may be made available for uptake by microorganisms from its surroundings within a given period. The significance of bioavailability depends on the pollutant's physicochemical properties, microorganisms, and contaminated site characteristics [11]. Incubation time is another critical parameter affecting bioremediation because it has been observed to affect the growth pattern of microorganisms directly. Proteus mirabilis YC801, immobilized on biochar, achieved a 42.5% Cr bioreduction and adsorption capacity after 6 h of incubation [132]. The temporal requirement is high for microbial degradation, and the reaction time for complete degradation is higher than that for other removal processes. The time scale of the microbial degradation process can be reduced by selecting suitable microorganisms with quicker growth phases for pollutant degradation or removal. However, choosing biochar with a high adsorption potential for pollutants is critical for reducing the bacterial adaptation time.
The pH value also influences microbial metabolic processes, particularly growth, cell membrane transport, the zeta potential of sorbate, and changes in the sorbent surface characteristics [133]. Huang et al. [132] observed an increase in Cr 6+ reduction with a pH increment from 6.0 to 7.0, showing a maximum removal of 83.7% at pH 7.0. However, alkalifying the Cr 6+ -MCB system from pH = 8.0 to pH = 10.0 inhibited the removal capacity of MCB for Cr 6+ . Similarly, the highest Cr bioreduction was found at 30°C, similar to the optimal culture temperature for the strain. Bioreduction significantly decreased with a further increase in temperature above 30 • C, which might be attributed to the loss of cell viability and the inhibition of the essential enzymes and proteins responsible for microbial growth and biodegradation at elevated temperatures [12,132].
Similarly, temperature and pH significantly influenced tebuconazole degradation by Alcaligenes faecalis WZ2, and degradation efficiency was strongly correlated with bacterial growth [125]. Tebuconazole degradation efficiency reached 88.5% under ideal conditions (a temperature of 30-35 • C and a pH of 6-8). Because of bacterial growth inhibition and a decrease in the catalytic activities of microbial enzymes involved in tebuconazole degradation, the efficiency was significantly reduced below the ideal temperature and pH.
Heavy Metal Ions and Pesticide Removal Using MCB
The advantages of immobilizing cell systems onto carriers in the bioremediation of metal ions and pesticides are far superior to those of using biochar or free cells alone [134]. Pollutant transfer into the microbial community from the contaminated sites can be enhanced by immobilizing the microbial strains onto biochar. Biochar can enhance the biological community composition of the soil through physisorption; in return, these microorganisms, adsorbed on the biochar surface, have a metabolizing capability for the pollutants present in the soil [135]. The porous structure of biochar enhances the growth and reproduction of the microorganisms and can also act as a source of nutrients for the microorganisms [136]. The immobilization also ensures that the microorganisms are assimilated for degradation to form biofilms around the porous structure complex of the biochar microbes [136]. Biochar can alleviate the contaminant concentration and reduce the inhibitory effect of these contaminants on the growth of microorganisms via the adsorption and subsequent decrease in contaminant concentration in soil/aqueous medium [137].
Using biochar and bioremediation in tandem with functional microbial strains is a viable and emerging strategy for the long-term remediation of contaminated water and soil. Numerous microbial strains with strong metal tolerance or adsorption capability have been isolated and used for bioremediation, either as free-living cells or by immobilizing a microbial cell with a specific carrier substance. Metal-tolerant microorganisms immobilized on biochar have been used as a bio-augmentation method to improve heavy metal phytoremediation, indirectly reducing heavy metal contamination in soil. Incorporating bacteria immobilized on biochar into the soil may indirectly improve Cd removal by promoting plant growth and the phytoremediation effect [138]. Cd-resistant bacteria immobilized on biochar improved the phytoextraction efficiency by Chlorophytum laxum R. Br. via cadmium phytoaccumulation in the shoots and roots, and Cd translocation from the roots to the shoots. Insoluble phosphate solubilization can be achieved via microbial phosphate solubilizers (PSB). Teng et al. [134] observed that combining PSB and biochar improved Pb 2+ immobilization by forming a stable crystal texture on its surface. Zhang et al. [139] used the PSB bacteria Pseudomonas chlororaphis for lead removal. However, the organism could not proliferate in indigenous bacteria, whereas the addition of PSB-immobilized biochar (PIB) improved bacterial growth and reduced Pb concentrations to less than 1 mg/kg. As a result, soil inoculation with PIB can be used as a substitute for Pb immobilization, avoiding the secondary pollution caused by phosphorus toxicity.
The microorganism immobilization with biochar carrier was also influential in remediating soil polluted with a combination of heavy metals. Tu et al. [140] introduced Pseudomonas sp. NT-2, loaded onto maize straw biochar, into Cd-Cu mixed soil. The application of Pseudomonas sp. NT-2-loaded biochar effectively reduced the bioavailability of Cd and Cu and increased the soil enzymatic activities in the soil system. Qi et al. [135] used three strains of mixed bacteria, Bacillus subtilis, Bacillus cereus, and Citrobacter sp.loaded biochar for U and Cd removal. They discovered that MCB promoted growth in celery and reduced the U and Cd phytoaccumulation, compared to free cell and biochar treatments. Research on Cr 6+ removal by immobilized microorganisms with biochar has attracted increased interest recently. The metal ion-resistant bacterium, Proteus mirabilis YC80, was immobilized using biochar derived from the bloom-forming cyanobacterium, D. flos-aquae [132]. The ability of biochar-immobilized Proteus mirabilis PC801 to remove Cr 6+ was superior, compared to a free cell. The removal efficiency of Cr 6+ by PC801-immobilized biochar was 100%, with 87.7% total Cr immobilized on the carrier and only 12.3% Cr 3+ remaining in the solution. Table 5 includes microbial cell immobilized biochar reported for heavy metal and pesticide abatement. Physical adsorption, ion exchange, surface complexation, precipitation, and biotransformation are some of the mechanisms involved in MCB-mediated heavy metals removal (Figure 7). Biochar containing oxygen functional groups, mineral components such as carbonates and phosphates, and microbial surface functional groups contribute to the removal of metal cations. Shen et al. [143] investigated the mechanism of cadmium removal using biochar-immobilized microalgae. They discovered that electrostatic attraction, surface complexation, and ion exchange were responsible for cadmium removal (maximum adsorption 217.41 mg/g) from wastewater. Similarly, Tu et al. [140] noted that surface complexation with different functional groups on cells, cation exchange, and surface complexation on biochar contributed to the enhanced stability of Cd 2+ and Cu 2+ in the contaminated soil. Microorganisms secrete enzymes that mediate redox reactions and surface complexation. These are the mechanisms involved in removing As 3+ , As 5+ , Cr 6+ , U 6+ , and Mn 2+ . Youngwilai et al. [147] examined the mechanism of Mn 2+ removal by the Streptomyces violarus strain, immobilized on biochar. They found that the two processes, namely, biological oxidation by the immobilized strain and adsorption by biochar, work together.
The presence of multiple contaminants at a particular contaminated site is a widespread phenomenon that could severely affect the microorganisms' remediation potential [148]. This limitation can be addressed by the associative effect of the benefits of biochar and the microorganisms via the immobilization of functional bacteria (such as organic contaminantsdegraders) on biochar, as this could potentially remediate various types of contaminants. The application of biochar-microbial complex also increased the soil microbial and enzymatic activity, along with conducting the simultaneous bioremediation of multiple contaminants in several studies [126,148]. Several studies report that the degradation efficiency by biochar-immobilized bacterial consortia in co-contaminated sites is significantly enhanced compared to the free bacteria, due to their bioaugmentation abilities. For instance, Li et al. [149] immobilized PAH-degrading bacteria (Citrobacter sp.) into biochar, increased the degradation rate of PAH and reduced the toxicity of Ni by bio-transforming the available Ni into a stable form.
Pesticide degradation can be enhanced by introducing exogenous free cells to polluted soil. However, this method has several drawbacks, including the growth and survival of microbial cells, inadequate nutrients, lesser adaptability to surroundings, and competition with native microorganisms [139,150]. Immobilizing the exogenous pollutant-degrading bacteria on a support material can be an alternative strategy. This can be an ideal environment for their survival in different soil conditions [151]. Microorganisms immobilized in biochar have the potential to directly or indirectly reduce environmental pollution, while also allowing for the long-term maintenance of catalytic activity. Due to its superior porosity, ample surface area, and functional groups, biochar is an ideal medium and a rich nutrient composition for immobilizing and reproducing microbial cells [125].
Biochar can improve the soil's pollutant adsorption capacity while providing the nutrients for microbial growth and function [152]. Adsorption and covalent-binding methods were used to immobilize Pseudomonas putida onto coconut fiber-derived biochar. The efficacy of MCB in paraquat removal from contaminated water was studied by Ha et al. [129]. After 48 h of incubation, MCB could convert paraquat to 4,4-bipyridyl and malic acid. According to Wahla et al. [148], the immobilization of the MB3R consortium was achieved on biochar-remediated soil contaminated with metribuzin. The immobilization of a microbial consortium on biochar increased the rate of cypermethrin degradation and removal efficiency while lowering the cypermethrin's bioavailability to indigenous organisms [153]. Sun et al. [125] isolated and identified Alcaligenes faecalis WZ-2 as a tebuconazole-degrading strain and supported it on wheat straw biochar as a carrier. The biochar-immobilized WZ-2 reduced the half-life of tebuconazole in soil from 40.8 to 13.3 days and affected the microbial population and enzyme activities in polluted soil.
Conclusions and Future Prospective
Recent research on removing heavy metal ions and/or pesticides using biochar and microorganisms has revealed their enormous potential. Biomass-derived materials, such as biochar, have gradually been established as a viable platform for advancing the design and development of carbon-based materials and their suitability for various uses, such as, for instance, environmental remediation applications. A plethora of biochar production and activation approaches can be used, depending on the final application. In this review, the role of microorganisms and biochar in bioremediation is thoroughly discussed. Along with its usage as an adsorbent for heavy metal ions and pesticides, biochar can also be utilized as an immobilization support for microorganisms. Carbonaceous materials have been frequently used as carrier materials for bacterial immobilization, to enhance the bioremediation efficiency of organic pollutants. Compared with expensive carbon materials, biochar is more competitive as a carrier material, as it is cheaper but has acceptably high porosity, which could provide shelter and nutrients for microbial cells, facilitating the colonization of microbial cells and the formation of microbial hot spots on the surface and in the pores of biochar. According to previous research, adsorption and entrapment are the most common methods for preparing the MCB. Toxic metal ions and pesticides have been successfully removed using immobilized cells. The key factors influencing the removal efficiencies are the pollutant's concentration, incubation time, temperature, and pH.
The physical and chemical properties of biochar make it a suitable carrier/platform for microbial cell immobilization; however, this research area is still in its initial stages. The limitations related to the loss of activity of MCB and mass transfer potential have not been studied widely. Even though the immobilization of metal ions and pesticide-degrading microorganisms are cost-effective, stable, and environmentally friendly approaches, research can be conducted to enhance the treatment efficiency and improve the stability of microbial cells. The regeneration of the immobilized cells and recovery of the adsorbed pollutant can be improved. Most of the research focusing on immobilized microbes on biochar is mainly laboratory-based and involves the remediation of soil or an aqueous environment. The practical application of this in situ method is restricted, as the actual contamination sites are usually complicated. Research can be conducted to elucidate the heavy metal and pesticide degradation ability of a particular MCB from the soil and aqueous environment. The practical use of MCB can be further improved by increasing the efficacy and viability of the immobilized microbial cells and exploring approaches that would make the usage of MCB easier in contaminated sites. Moreover, the microbe-immobilized biochar can be employed in co-contaminated sites with heavy metals and pesticides for remediation. Genetically modified microorganisms are of increasing interest for the treatment of targeted pollutants. Therefore, further studies can be performed to genetically modify the microorganism for the targeted remediation of metal ions and pesticides, as well as to study the immobilization characteristics of these microbes on biochar. | 15,236.2 | 2023-01-01T00:00:00.000 | [
"Environmental Science",
"Chemistry"
] |
Magnetic Force-Assisted Nonlinear Three-Dimensional Wideband Energy Harvester Using Magnetostrictive/Piezoelectric Composite Transducers
This paper presents a nonlinear magnetoelectric energy harvester which has the potential to harvest vibrational energy over a wide bandwidth in arbitrary motion directions. Three springs with equal intersection angles are adopted to absorb the multi-directional vibration energy. Magnetic interaction between the magnets and ME transducers allows the nonlinear motion with enhanced harvesting frequency range. Very good agreement is observed between the numerical and experimental open-circuit voltage output frequency response curves. The experimental results show that the harvester can harvest vibrational energy in an arbitrary direction, exhibiting a further bandwidth of 5.2 Hz. This study provides a new solution to effectively use the magnetoelectric energy harvester for multi-directional and bandwidth vibrational energy scavenging in the surrounding environment.
Introduction
Energy harvesting provides a promising solution to implement self-sustained lowerpower electronic devices and which is regarded as a promising alternative to traditional batteries [1]. Vibration to electricity energy conversion strategies have received extensive attention over the past decade. The basic mechanisms of vibration energy harvesting include piezoelectric, refs. [2,3] electromagnetic, refs. [4,5] magnetostrictive, refs. [6,7] and triboelectric transduction [8][9][10]. Nevertheless, many reported devices are not easy to implement in practice due to a number of critical issues, such as diverse mechanical vibrations or varying excitation frequency scenarios, resulting in drastic reduction of the power output. To remedy these key issues of conventional energy harvesters, energyharvesting systems with three-dimensional (3D) motion directions have been carried out in many researches [11][12][13][14]. Aktakka et al. [11] reported a piezoelectric energy harvester based on transverse-mode piezoelectric crab-leg suspensions and partitioned top electrodes to scavenge ambient vibration energy from all three axes. Liu et al. [12] developed an electromagnetic energy harvester with multiple vibration modes characterized using 3D excitation at different frequencies. However, due to the complicated structure and increased system volumes, power densities and efficiency of energy conversion in such designs are reduced. A tri-directional piezoelectric energy harvester that exhibits the capability of harvesting vibration energy from three orthogonal directions was demonstrated by Su et al. [13] Nevertheless, such energy harvester can only work in a narrow frequency bandwidth due to the ambient vibration with a time-variant frequency.
In this study, we propose a nonlinear magnetoelectric energy harvester which can achieve multi-directional sensitivity and remarkable broad bandwidth. In the proposed Micromachines 2022, 13, 1633 2 of 8 energy harvester, three spiral springs at equal central angle with each other make the cylindrical magnet vibrate in arbitrary directions. Magnetic interaction between the magnets is adopted for broadband energy harvesting because of the nonlinearity on the system. Numerical simulations and experimental results are carried out that nonlinear harvester can sustain large-amplitude oscillations over a wide frequency range, and it can generate power efficiently in an arbitrary direction.
Harvester Design and Analysis Model
The novel 3D wideband energy harvester using magnetostrictive/piezoelectric composite (ME) transducers has been designed, which is composed of three identical springs, three ME transducers, a harvester frame, and a magnetic circuit provided by a cylindrical magnet and a ring magnet, as shown in Figure 1a. The cylindrical magnet is mobile and suspended by the three identical springs, resulting in oscillating in three-dimensional space. The magnetic circuit comprises a cylindrical magnet and a ring magnet and this arrangement produces a concentrated magnetic flux gradient through the ME transducers, as demonstrated in Figure 1b. The ME transducer is a sandwich of one piezoelectric ((Pb(Zr 1−x ,Tix)O 3 )) layer bonded between two magnetostrictive layers. The magnetostrictive layers are magnetized along the longitudinal direction (L mode), and the piezoelectric layer is polarized in its thickness direction (T mode). Figure 1c shows the operation model of the ME transducer. ME transducers are fixed symmetrically on the acrylic supporting frame. Owing to an external acceleration from ambient vibration, there will exist a relative motion between the ME transducers and the magnetic circuit. Then ME transducers experience the various magnetic field variations, leading to the stress in the magnetostrictive layers. As a result, the stress is transmitted to the piezoelectric layers, and then generating electrical power due to piezoelectric induction.
Micromachines 2022, 13, 1633 2 of 8 [13] Nevertheless, such energy harvester can only work in a narrow frequency bandwidth due to the ambient vibration with a time-variant frequency.
In this study, we propose a nonlinear magnetoelectric energy harvester which can achieve multi-directional sensitivity and remarkable broad bandwidth. In the proposed energy harvester, three spiral springs at equal central angle with each other make the cylindrical magnet vibrate in arbitrary directions. Magnetic interaction between the magnets is adopted for broadband energy harvesting because of the nonlinearity on the system. Numerical simulations and experimental results are carried out that nonlinear harvester can sustain large-amplitude oscillations over a wide frequency range, and it can generate power efficiently in an arbitrary direction.
Harvester Design and Analysis Model
The novel 3D wideband energy harvester using magnetostrictive/piezoelectric composite (ME) transducers has been designed, which is composed of three identical springs, three ME transducers, a harvester frame, and a magnetic circuit provided by a cylindrical magnet and a ring magnet, as shown in Figure 1a. The cylindrical magnet is mobile and suspended by the three identical springs, resulting in oscillating in three-dimensional space. The magnetic circuit comprises a cylindrical magnet and a ring magnet and this arrangement produces a concentrated magnetic flux gradient through the ME transducers, as demonstrated in Figure 1b. The ME transducer is a sandwich of one piezoelectric ((Pb(Zr1−x,Tix)O3)) layer bonded between two magnetostrictive layers. The magnetostrictive layers are magnetized along the longitudinal direction (L mode), and the piezoelectric layer is polarized in its thickness direction (T mode). Figure 1c shows the operation model of the ME transducer. ME transducers are fixed symmetrically on the acrylic supporting frame. Owing to an external acceleration from ambient vibration, there will exist a relative motion between the ME transducers and the magnetic circuit. Then ME transducers experience the various magnetic field variations, leading to the stress in the magnetostrictive layers. As a result, the stress is transmitted to the piezoelectric layers, and then generating electrical power due to piezoelectric induction. In order to design a vibration energy scavenger that meets 3D energy harvesting from diverse mechanical vibrations, theoretically, simulations were conducted by COMSOL Multiphysics in this work. For material data, steel (Young's modulus of 2.05 GPa and Poisson's ratio of 0.28) was used for three springs. Note that the magnetic interaction between the ME transducers and magnets is not calculated in the process of simulations in order to simplify the model. The instant motion states of the cylindrical magnetic are shown in Figure 2a,b corresponding to resonant frequencies of 8.4 Hz and 9.2 Hz, respectively, which indicate the mechanical motion behaviors of the harvester under different mode shapes. There are two mode shapes: out-of-plane mode and in-plane mode for the vibration energy harvesting system due to the rationally designed structure. As a result, the proposed harvester is capable of extracting vibration energy from arbitrary directions. As mentioned above, magnetic interaction between the ME transducers and the magnets exists in the vibration energy harvesting system. Thus, the motion behaviors of the cylindrical magnet in an arbitrary excitation direction will be affected by the magnetic force. The proposed 3D energy harvester could be modeled as the mechanical spring mass-damper system to analyze the frequency response in out-of-plane mode (note that the harvester excited in in-plane mode shares the same model), the governing dynamic equation of motion for this system can be given by where m is the mass of the cylindrical magnet, c and k are the mechanical damping coefficient and spring constant of the spring, respectively. x is the relative displacement of the cylindrical magnet to the acrylic supporting frame, .
x and ..
x are the relative velocity and acceleration of the cylindrical magnet with respect to the frame. the supporting frame along the z-axis direction. F(x) denotes the vertical component of the magnetic forces between the ME transducers, which is expressed as a summation of the magnetic forces acting on the cylindrical magnet by the ring magnet, F mag (x), and the mounted ME transducers, F ME (x), i.e., To investigate the influence of F(x) on the motion behaviors of the cylindrical magnet, Ansoft Maxwell 3D software was used as in further study of the relationship between magnetic force and displacement, as demonstrated in Figure 2c. It is clear that the magnetic forces are nonlinear functions of displacement and which are different from each other. It is reasonable to simplify the function of magnetic force into a third-order polynomial and can be written as F(u) = k l x + k n x 3 (see Reference [15] for details), where the coefficients k l and k n are the fitting parameters, which can be calculated by using a least square procedure. Then, the nonlinear factor related to the motion of the cylindrical magnet evaluated as where Y 2 = mg/k [16]. The working bandwidth of harvester is determined by the nonlinear factor, with a larger value resulting in a wider bandwidth. To further investigate the nonlinear behaviors of the magnetic on the bandwidth, Equation (1) can be given in nondimensional form as ..
where ξ is the damping ratio, ε is the third order coefficient of the magnetic force expansion. ω 0 is the natural frequency, A is motion amplitude of the acrylic supporting frame, ω and θ are the angular frequency and the phase angle, respectively. Using the Harmonic Balance method (x = X cos ωt is assumed), Equation (4) can be solved. The frequency-amplitude relationship for the harvester can be predicted by where γ = ω/ω 0 . Subsequently, the typical frequency-response curves (FRC) for a hardening system are plotted using Equations (5) in out-of-plane and in-plane modes, respectively, as depicted in Figure 2d. An intensely nonlinear behavior appears in each FRC with the magnetic force, which contributes the harvester to operating in a wider bandwidth. Thus, the amplitude of frequency-response jumps to a higher level due to the nonlinear behavior of the cylindrical magnetic, leading to a wider resonance range and a higher resonance drop point.
Owing to the motion of the cylindrical magnetic, the fixed ME transducer undergoes a changing magnetic field, then electrical power will be generated. The ME output voltage can be written as where ∆H is the magnetic field variation induced by the ME transducer, α ME is the ME voltage coefficient under a DC magnetic bias field along longitudinal direction. In order to investigate the influence of the magnetic field distribution on the ME voltage output, Ansoft Maxwell 3D software was used to elevate the distributions of magnetic field and variations of the magnetic circuit, as illustrated in Figure 3. Figure 3a,b show that, in the range of −10 mm to 10 mm, the average magnetic fields provided by magnetic circuit are~230.7 Oe and~410.2 Oe for the two modes, and the one in in-plane mode is close to the measured optimal bias magnetic field of the ME transducer,~405 Oe [17]. By contrast, the magnetic field variation in in-plane mode is larger compared with the magnetic field variation in out-of-plane mode. In this case, the output voltage would be larger in in-plane mode, because the ME transducer not only works in a good condition of the DC bias magnetic field but also experiences a large enough magnetic field variation.
Fabrication
The fabrication process of the designed energy harvester using magnetostrictive/piezoelectric composite transducers was composed of three parts, ME transducers, energy harvester frame and a magnetic circuit. First, three ME transdusers were fabricated by piezoelectric and magnetostrictive layers, one piezoelectric layer is bonded between two magnetostrictive layers using insulated epoxy adhesive, and both the dimensions of the piezoelectric and magnetostrictive layers are 12 mm × 6 mm × 1 mm, being 12 mm in the longitudinal direction. An acrylic sheet was cut into the desired ring shape by a laser cutter (PLS6.75) with the inner and outer diameters of 54 mm and 64 mm, respectively. Three grooves with an included angle of 120° between each other were cut on the outside of the ring acrylic, forming the harvester frame. Then a cylindrical magnet with diameter of 20 mm and height of 8 mm was fixed by three spiral springs with equal intersection angles. And a ring magnet with the inner and outer diameters of 64 mm and 74 mm, was mounted on the outside acrylic frame. Finally, Three ME transducers were mounted symmetrically on the prepared grooves of the acrylic frame. And the wires was led out of the ME transducers to the experimental test system.
Experimental Results and Discussion
To measure the output characteristics of the prototype, experimental platform was established, Figure 1d shows the schematic illustrations of experimental test system clearly. We measured the open-circuit voltages for all three ME transducers in out-ofplane mode, as shown in Figure 4a. The voltage outputs of those follow a similar trend and are fairly close to each other. The reason is that three ME transduces experience a uniform magnetic field variation. Want's more, hardening system responses of the harvester can be observed due to the nonlinear behavior of the magnetic force applied to the movable magnet. As a result, the working bandwidth was broadened, which is well match with the model predictions shown in Figure 2d. If the half peak voltage point is adopted as the criteria of the working bandwidth [18], an approximate working bandwidth of 4.6 Hz can be achieved.
For the capability of vibration energy harvesting in in-plane mode, the experimental voltage responses of the harvester at different excitation angles are measured, as displayed in Figure 4b−d, respectively. It can be found that the proposed harvester could scavenge and convert mechanical vibrations into electricity at different excitation angles, and the experimental results could be extended to the full range of 360° because of the structural symmetry. Besides, one ME transducer with larger output voltage at different excitation angles can be observed, owing to the difference of magnetic field variations. Moreover, the nonlinear behavior of the energy harvesting system leads to broadening the working bandwidth and the maximum working bandwidth can get up to 5.2 Hz, which is well consistent with the simulated results in Figure 2d and theoretical analysis.
Fabrication
The fabrication process of the designed energy harvester using magnetostrictive/piezoelectric composite transducers was composed of three parts, ME transducers, energy harvester frame and a magnetic circuit. First, three ME transdusers were fabricated by piezoelectric and magnetostrictive layers, one piezoelectric layer is bonded between two magnetostrictive layers using insulated epoxy adhesive, and both the dimensions of the piezoelectric and magnetostrictive layers are 12 mm × 6 mm × 1 mm, being 12 mm in the longitudinal direction. An acrylic sheet was cut into the desired ring shape by a laser cutter (PLS6.75) with the inner and outer diameters of 54 mm and 64 mm, respectively. Three grooves with an included angle of 120 • between each other were cut on the outside of the ring acrylic, forming the harvester frame. Then a cylindrical magnet with diameter of 20 mm and height of 8 mm was fixed by three spiral springs with equal intersection angles. And a ring magnet with the inner and outer diameters of 64 mm and 74 mm, was mounted on the outside acrylic frame. Finally, Three ME transducers were mounted symmetrically on the prepared grooves of the acrylic frame. And the wires was led out of the ME transducers to the experimental test system.
Experimental Results and Discussion
To measure the output characteristics of the prototype, experimental platform was established, Figure 1d shows the schematic illustrations of experimental test system clearly.
We measured the open-circuit voltages for all three ME transducers in out-of-plane mode, as shown in Figure 4a. The voltage outputs of those follow a similar trend and are fairly close to each other. The reason is that three ME transduces experience a uniform magnetic field variation. Want's more, hardening system responses of the harvester can be observed due to the nonlinear behavior of the magnetic force applied to the movable magnet. As a result, the working bandwidth was broadened, which is well match with the model predictions shown in Figure 2d. If the half peak voltage point is adopted as the criteria of the working bandwidth [18], an approximate working bandwidth of 4.6 Hz can be achieved.
For the capability of vibration energy harvesting in in-plane mode, the experimental voltage responses of the harvester at different excitation angles are measured, as displayed in Figure 4b−d, respectively. It can be found that the proposed harvester could scavenge and convert mechanical vibrations into electricity at different excitation angles, and the experimental results could be extended to the full range of 360 • because of the structural symmetry. Besides, one ME transducer with larger output voltage at different excitation angles can be observed, owing to the difference of magnetic field variations. Moreover, the nonlinear behavior of the energy harvesting system leads to broadening the working bandwidth and the maximum working bandwidth can get up to 5.2 Hz, which is well consistent with the simulated results in Figure 2d and theoretical analysis.
The electrical output power of the harvester was further characterized at the driving frequency of 12 Hz in out-of-plane mode and 14 Hz in in-plane mode, external load was connected to the harvester in series, and the output voltage and power across the resistor were measured, as shown in Figure 5. As the increase of the resistor, the output power calculated by P = V 2 ⁄R will increase dramatically and then decrease slowly. The peak power arrives at the maximum value of 20.7 µW at 1.8 MΩ in out-of-plane mode, while the corresponding maximum output power density of 8.24 µW/cm 3 can be obtained (Figure 5a). Due to the structural symmetry of the harvester working at in in-plane mode, the experiments of the output power at the excitation angle of 60 • was employed, as demonstrated in Figure 5b. The measured maximum output power reaches 160.8 µW with the matched resistance of 4 MΩ. And the corresponding output power in in-plane mode at its matched resistance gets up to 64 µW/cm 3 . Therefore, the above experimental results reveal the output performance of the proposed device for vibrational emery harvesting. The electrical output power of the harvester was further characterized at the driving frequency of 12 Hz in out-of-plane mode and 14 Hz in in-plane mode, external load was connected to the harvester in series, and the output voltage and power across the resistor were measured, as shown in Figure 5. As the increase of the resistor, the output power calculated by P=V 2 ⁄R will increase dramatically and then decrease slowly. The peak power arrives at the maximum value of 20.7 μW at 1.8 MΩ in out-of-plane mode, while the corresponding maximum output power density of 8.24 μW/cm 3 can be obtained (Figure 5a). Due to the structural symmetry of the harvester working at in in-plane mode, the experiments of the output power at the excitation angle of 60° was employed, as demonstrated in Figure 5b. The measured maximum output power reaches 160.8 μW with the matched resistance of 4 MΩ. And the corresponding output power in in-plane mode at its matched resistance gets up to 64 μW/cm 3 . Therefore, the above experimental results reveal the output performance of the proposed device for vibrational emery harvesting. The electrical output power of the harvester was further characterized at the driving frequency of 12 Hz in out-of-plane mode and 14 Hz in in-plane mode, external load was connected to the harvester in series, and the output voltage and power across the resistor were measured, as shown in Figure 5. As the increase of the resistor, the output power calculated by P=V 2 ⁄R will increase dramatically and then decrease slowly. The peak power arrives at the maximum value of 20.7 μW at 1.8 MΩ in out-of-plane mode, while the corresponding maximum output power density of 8.24 μW/cm 3 can be obtained (Figure 5a). Due to the structural symmetry of the harvester working at in in-plane mode, the experiments of the output power at the excitation angle of 60° was employed, as demonstrated in Figure 5b. The measured maximum output power reaches 160.8 μW with the matched resistance of 4 MΩ. And the corresponding output power in in-plane mode at its matched resistance gets up to 64 μW/cm 3 . Therefore, the above experimental results reveal the output performance of the proposed device for vibrational emery harvesting.
Conclusions
In summary, we proposed a novel design of three dimensional vibration energy harvester with a broad operating frequency range. The theoretical analysis and experimental Micromachines 2022, 13, 1633 7 of 8 characterization are conducted for the proposed harvester. By mean of unique structural design, the harvester can scavenge ambient vibration energy in arbitrary directions. Assisted with the nonlinear behaviors of the movable magnet, the operating bandwidth of the harvester was shown to be substantially enhanced. With further optimization of the dimensional parameters, we believe the proposed design could be promising for practical applications to vibration energy scavenging from the environments for realizing the self-powered wireless sensor networks.
Data Availability Statement:
The data presented in this study are available upon request from the corresponding author.
Conflicts of Interest:
The authors declare no conflict of interest. | 4,978.4 | 2022-09-29T00:00:00.000 | [
"Engineering"
] |
Fourier Transform for Locally Integrable Functions with Rotational and Dilation Symmetry
: The Fourier transform for slowly increasing functions is defined by the Parseval equation for tempered distributions. This definition was supplemented by a novel method of performing practical calculations by computing the Fourier transform for a suitably tempered function and then by integration by parts. The application of this method is illustrated both for the toy case, in which the function is integrable, so its Fourier transform can also be computed using the standard formula, and for the case of Coulomb-like potentials, which are only locally integrable functions. All of them have spherical symmetry, and two of them additionally have dilation symmetry. The proposed novel method does not violate these symmetries at any stage of the calculation.
Introduction
The Fourier transform (FT), defined for an integrable function, i.e., ψ ∈ L 1 (R 3 ) as [1] is a very useful concept that simplifies solutions of many problems in physical sciences, engineering, etc. Unfortunately, for several important physical quantities, the above definition cannot be applied, e.g., for the Coulomb potential V c (x) = 1 r , with r = |x| = √ x · x, which plays an important role in the description of electrostatic [2] and gravitational interactions, but evidently V c / ∈ L 1 (R 3 ). Not surprisingly, the standard definition of FT, when integrated over spherical angles in the position space R 3 , leads to an ill-defined integral where R + = {x ∈ R : x ≥ 0} and k = |k|. To eliminate this deficiency, a damping factor e −αr , with α > 0, is commonly added; for example, see [3][4][5], which converts the long-range Coulomb potential into the short-range Yukawa potential V α (x) Then, it is argued that the FT of the Coulomb potential V c can be defined by the limit α 0 of F {V α }, which is based on the analogy The advantage of such a procedure is that the spherically symmetric Coulomb potential is replaced by a spherically symmetric Yukawa potential, which means that rotational sym-metry remains unbroken. However, this does not apply to the dilation symmetry associated with the transformation x → x = λx, where the Coulomb potential is dilational covariant whereas the Yukawa potential explicitly breaks this symmetry. It seems obvious that any damping factor will break the dilation symmetry, so you have to invent a completely different computational procedure to preserve the dilation symmetry. This trouble of calculating the FT for the Coulomb potential can be avoided by saying that this FT is determined uniquely by the Poisson equation [6], but we did not pursue such reasoning further.
In this work, we present a novel method for calculating FT, which, when applied directly to the Coulomb potential, will reproduce the above result but without recourse to the Yukawa potential, so that both rotational and dilational symmetries remain intact at each stage of the procedure. Here, the dilational symmetry represents a physical symmetry that should be easily controlled at each stage of the computation. For more complicated cases, recovering such symmetry in the final result can lead to quite serious problems such as noncommutativity of limits, etc.). It will use the distributions with the test functions of the Schwartz class S (R 3 ), which due to the Parseval equation for the tempered distributions will lead to FT in the sense of distributions. For checking the consistency of this method, one may take a toy case W(x) = 1 x 2 ∈ L 1 (R 3 ), which gives the same FT, as the standard formula (1). Next, we consider the Coulomb-like potentials, which are slow-growing functions and thus have no standard FT, according to (2). First, we considered the softened Coulomb potential with a decaying exponential term V p (x) = 1 r e −c/r , with c > 0, [7], which vanishes for r → 0. Then, we studied the soft-core Coulomb potential U a (x) = 1 √ r 2 +a 2 , with an arbitrary parameter a ∈ R + [8], which is a special case of a generic soft-core Coulomb potential U a,q (x) = 1 (r q +a q ) 1/q , for q ≥ 1 [9][10][11], which is used for modeling potentials in atomic and molecular physics. At last, we considered the pure Coulomb potential V c , which has both rotational and dilation symmetry, unlike the previous two cases, which have only rotational symmetry. Our novel method is shown to be equivalent to the standard one, with the exponential damping factor, when the distributional derivatives are the standard derivatives-this is presented in the Appendix A. Finally, we mention how our method can be compared to the one used in [12], where dilation symmetry is used as a key step in the computation.
Calculation of Distributional FT-Methods
We begin by recalling some important properties of the Fourier transform and tempered decompositions that can be found in the vast literature, where the fundamental monograph by Gelfand and Shilov (see [12] pp. 71-74 and pp. 190-200) should be mentioned, but we can also select some items relevant to physics: [13][14][15]. The standard Fourier transform defined by (1) for integrable functions in L 1 (R 3 ) is a continuous and bounded function on R 3 , and it vanishes at infinity. For φ, ψ ∈ L 1 (R 3 ) one finds ψF {φ} ∈ L 1 (R 3 ) and φF {ψ} ∈ L 1 (R 3 ), so the Parseval theorem follows from Fubbini's theorem, where we used a compact notation for x, y ∈ R 3 . Evidently, this Parseval theorem is valid also for the Schwartz test function φ ∈ S (R 3 ), because S is a dense subspace of L p with 1 ≤ p < ∞. Moreover in this case both sides of (6) are the regular distributions S (R 3 ). This is a consequence of the fact that the Fourier transform is a bijection map F : S (R 3 ) → S (R 3 ) from the space S (R 3 ) unto itself, and hence F {φ} ∈ S (R 3 ). The situation changes for φ ∈ S (R 3 ) and L 1 loc F {ψ} cannot be given by the standard definition (1), so we cannot calculate the left-hand side in (6). In contrast, the right-hand side in (6) is a tempered distribution S (R 3 ) provided ψ is a slow-growing function, i.e., it has polynomial growth as |x| → ∞, which gives ψF {φ} ∈ S (R 3 ). This allows us to reinterpret (6) as a definition of distributional FT in the sense of distributions S (R 3 ).
Definition 1.
For a slow-growing locally integrable function ψ its distributional Fourier transform F D {ψ} is given by the relation for tempered distributions S (R 3 ) This equation will be called the Parseval equation for the distributional FT or the Parseval equation for short. Evidently if ψ ∈ L 1 (R 3 ), then its distributional Fourier transform F D {ψ} coincides with the standard one F {ψ} given by (1).
Before we move on to practical computations with the help of (8), we can slightly change the notation for the Schwartz test function by introducing where the bar denotes the complex conjugate. Thus, we may re-express (7) and (8) as with explicit notation for vectors x, k ∈ R 3 . Furthermore, in the following discussion, without loss of generality, we can drop the subscript in φ F = φ. The calculation of F D {ψ} is based on the analysis of the tempered distribution ψ[φ], and the standard procedure starts with a suitable modification of ψ[φ], where a damping factor is inserted under the integral sign. Such a calculation can be illustrated using the Coulomb potential ψ = V c / ∈ L 1 (R 3 ) as an example, where relations (9) and (10) look as However, for the tempered distribution we can introduce a damping factor e −αr , α > 0 and calculate using the complex conjugation of (1) for φ(x), where we can switch the order of integrations by Fubini's theorem, because V α (x) = V c (x)e −αr ∈ L 1 (R 3 ) and take F {V α } from (3). Then, for the last convergent integral, we can insert the limit α 0 under the sign of the integral, thus reaching the left-hand side of the Parseval equation in the form (10) and yielding the distributional FT for the Coulomb potential Accordingly, the ad hoc calculations presented in the introduction are closely related to the standard calculation of the distributional FT. Therefore, the previous drawbacks mentioned in the introduction also apply to this method.
We propose a different computational procedure, (a comparison of these two computational methods is presented in Appendix A), that relies on specific properties of Schwartz where n ∈ N and M ∈ R + were chosen arbitrarily [14] (also see pp. 16-17 in [12]). From the definition (1), one checks the formula for FT: k is the Laplace operator in the space of k ∈ R 3 . All this leads to the tempered distribution where we can take n ∈ N 0 and M ∈ R + arbitrarily, provided that Then, one can integrate by parts with respect to k, which gives the final equivalent form, which can be compared with the Parseval equation for the distributional FT (10) where and it has a polynomial growth for |x| → ∞ and ψ/ (r 2 + M 2 ) n ∈ L 1 (R 3 ) with appropriate choice of n and M, then its distributional FT is given in the sense of distributions S (R 3 ) by In case of rotational symmetry one has ψ(x) = ψ(r), so the integral over the unit sphere embedded in R 3 can be easily performed done leading to the result, which depends only on the radial variable k = |k| This in turn allows us to move on to integrals with respect to k ∈ R 3 , which can be expressed as the radial integral where the radial FT F rad {φ} is defined as the integral over the unit sphere embedded in R 3 where dω k is the hypersurface element on the unit sphere, and Ω 3 = 4π is the hypersurface area of the unit sphere in R 3 (see [12] p. 71). (Evidently F rad {φ}(k) is the mean value of F rad {φ}(k) on the sphere of radius k (see [12] p. 71), and one may express (18) by means of the mean value of φ(x) on the sphere of radius k which agrees with the FT for the spherically symmetric functions.) In practical calculations, one must first make the definite integral with respect to r, which gives a well-defined function of k. Then, integrating by parts with respect to k one finally obtains thus indicating that the distributional FT F D {ψ}(k) is also rotationally symmetric.
Although the tempered distribution ψ[φ] is defined in (13) as an expression independent of M, its final form explicitly includes the parameter M. It is therefore desirable to see how this presence effectively cancels itself, and this can be done by computing the derivative of the last line in (13) with respect to the parameter M where in (16) one may push the derivative under the integral sign in F ψ/(r 2 + M 2 ) n that is given by a convergent integral. These two tempered distributions S (R 3 ) are equal to each other, which can be easily checked by integrating by parts in the second case, where ∇ 2 k can be inserted under the integral sign F ψ/(r 2 + M 2 ) n+1 , which is given by the convergent integral. This brings us to the expected result ∂ ∂M 2 ψ[φ] = 0, but we discovered the crucial role played by integration by parts for checking the M-independence of ψ [φ].
If one imposes the dilation transformation on the function ψ(r) → ψ λ (r) = ψ(λr), then the radial Parseval equation (17) for the distribution ψ λ [φ] becomes where we changed the integral variables: r = ξλ −1 , k = pλ and the parameter M = M 1 λ. Thus, in the sense of distributions, we have the distributional FT for the dilation transformed function F D {ψ λ }(k) = λ −3 F D {ψ}(kλ −1 ). However, if the function ψ has dilation symmetry, then its distributional FT also has dilation symmetry This shows that our novel method of computing FT for functions in R 3 does not conflict with rotational and dilational symmetry.
Example 3. One may check the above tools for a toy function W(x) = 1 r 2 , which is both rotationally symmetric and dilation covariant W λ (x) = W(λx) = λ −2 W(x) and W ∈ L 1 (R 3 ), so its FT can be also computed from the standard definition (1) Then, we used the Parseval equation (17) with n = 1, for the functional W[φ], which gives the distribution, where first we performed the integral with respect to r, and then we integrated by parts with respect to k , so it agrees with the standard FT, as expected. Moreover, the dilation symmetry relation for FT is in agreement with the Formula (23).
Results for Coulomb-like Potentials
Next, we calculated FT for Coulomb-type potentials in turn: the smooth potential V P (x) = 1 r e −b 2 /r , the soft-core potential U a (x) = 1 √ r 2 +a 2 for a ∈ R + , and the plain Coulomb potential V c (x) = 1 r . All of them are locally integrable in R 3 ; thus, one can define the respective tempered distributions, for test function φ ∈ S (R 3 ), , we may take the simplest form of the Parseval equation (17) with n = 1 and M = 0, because the exponential factor smooths the vicinity of r = 0; then, we computed the integral with respect to r using formula (3.957.1) in [16] where K n (z) = K −n (z) is the modified Bessel function of the first kind of the n-th order. The integration by parts with respect to k introduces no boundary terms and one needs to take the second order derivative for the modified Bessel function Therefore, by analogy with (20), we obtain differential equation with the distributional FT This FT agrees with the corresponding result in [7], provided we use (iz) = (z) and take into account the slightly different definition of FT therein. We also note that in [7] the damping factor e −κr with κ > 0 was added before computing the FT, and the limit of κ 0 was taken for the result obtained. Then, we calculated the distributional FT for the soft-core Coulomb potential U a (x) = 1 √ r 2 +a 2 , where we can take (17) with n = 1 and M = 0, which gives the distribution that allows to compute the integral with respect to r by using formula (3.754 The integration by parts with respect to k introduces no boundary terms; therefore, we obtained with the distributional FT For this soft-core Coulomb potential we may make alternative calculations that begin with separation of the leading term (for r a) as U a (x) = V c (x) + U a sub (x), where the subtracted soft-core potential U a sub (x) follows from the relation This leads to the subtracted functional where we integrated by parts twice with respect to k, inserting the differentiation under the sign of the integral. Then, the integrals with respect to r and µ can be calculated successively, using formula (3.754.3) in [16] a 0 R + µ r sin(kr) giving the final form of the subtracted functional From the formula U a (x) = V c (x) + U a sub (x), one directly obtains the formula for the functionals V c [φ] = U a [φ] − U a sub [φ] and next the formula for FTs In this way, we obtained the distributional FT for the Coulomb potential, which agrees with the expression obtained earlier by different methods. This clearly shows that a FT that is symmetric under dilation can appear as a difference of two FTs that are not symmetric under dilation. At last, for checking consistency, we calculated directly the distributional FT for the Coulomb potential, so we took (17) with n = 1 and M > 0, and, as before, first we needed to calculate the integral with respect to r. Using formula (8.217.1) in [16], we obtained where Ei(z) is the exponential integral function and then we may integrate by parts with respect to k (Above we have used the differential properties of F ± , that are discussed in Appendix B). Thus, the distributional FT is F D {V c }(k) = 4π k 2 , which agrees with the previous result (35). The distributional FT for the pure Coulomb potential V c can be used to prove the Poisson equation for V c in the sense of distribution S (R 3 ). To demonstrate this, we calculates a new distribution V c [∇ 2 φ] for test function φ ∈ S (R 3 ), which allows for performing distributional derivatives, applying the Parseval equation for tempered distributions and the definition of Dirac delta singular function, successively Thus, we may integrate by parts, which leads to distributional derivatives,
Discussion and Further Research
The proposed novel method of computing FT for slowly increasing functions can be applied to functions that are locally integrable and can be tempered with a polynomial of a finite order. The most troublesome part of this procedure is computation of a convergent improper integral, which may additionally depend on some arbitrary auxiliary parameterin this study, we used the formulas for integrals from the table of integrals [16].
Due to the rotational symmetry of the analyzed functions, we obtained radial integrals and radial test functions. The dilation symmetry manifested itself in a less evident way but was still not broken at any stage of the calculation.
For the softened Coulomb potential V p (x) and the soft-core Coulomb potential U a (x), due to a smooth dependence on r → 0, the Parseval equation may be taken with M = 0, which simplifies our computation and leads to differentiable auxiliary functions. Then, the distributional FT for each potential becomes the second-order derivative of the respective auxiliary function. Both the auxiliary functions and the FTs depend on the modified Bessel functions. However, the plain Coulomb potential V c (x) is singular for r → 0, so we needed to take M = 0, and further steps of computation are more difficult. We obtained the auxiliary function that depends on the exponential integral function of Ei(z). However, it appeared only in an intermediate step of the computation, while in the final result this special function disappeared.
Furthermore, we showed the usefulness of calculating the FT for a subtracted function, U a sub (x), that is already an integrable function, at the cost of introducing an additional proper integral with respect to an auxiliary parameter, in our case µ.
In general, it can be said that the increased rigor of the method comes at a price, as the calculations become somewhat more complicated. The applicability of our new method, according to the lemma in Section 2, is limited to locally integrable functions on R 3 that grow slowly as |x| → ∞. However, these functions can also contain additional variables that are not integrated when computing FT. In physics, such an additional variable might be a time coordinate. For example, in quantum field theory, our method can be used for calculation of the Wightman functions [17], for a free scalar field. The resulting expression has the Lorentz symmetry and contains distributional derivatives-this will be published soon in a separate article. This new result deviates from previous approaches, where one either uses mathematically inaccurate tricks, as in [18], or introduces damping factors that break an important physical symmetry (here, the Lorentz symmetry) [19] or uses Lorentz symmetry to analyze divergent integrals [20]. The main feature of the new expressions for Wightman functions is the appearance of distributional derivatives, which has not been used before. We plan to apply our novel approach to FT calculations in the context of light front quantization [21], where existing previous results for Wightman functions [22,23] should be checked and/or improved by analysis from a different point of view. Our computational method can be applied to other mathematical problems arising in quantum mechanics that require the computation of FT for functions that are not in L 1 (R n ); for example, regarding the time-evolution operator in the free Schrödinger equation, see [24].
We also need to comment on the FT calculated for the spherically symmetric generalized function r λ for λ = −n, −n − 2, . . . in R n in [12]. A key part of this calculation is the dilation covariance (homogeneity), so it cannot be used for the case of Coulomb potentials considered in this study. Therefore, we plan to apply our novel procedure to compute the FT for the above generalized function in a separate study, which should provide a basis for comparing the two methods.
which can be used for doing partial integration in the linear functionals where primes denote derivatives with respect to k. | 5,087.6 | 2022-01-26T00:00:00.000 | [
"Physics"
] |
Hidden Nambu mechanics II: Quantum/semiclassical dynamics
Nambu mechanics is a generalized Hamiltonian dynamics characterized by an extended phase space and multiple Hamiltonians. In a previous paper [Prog. Theor. Exp. Phys. 2013, 073A01 (2013)] we revealed that the Nambu mechanical structure is hidden in Hamiltonian dynamics, that is, the classical time evolution of variables including redundant degrees of freedom can be formulated as Nambu mechanics. In the present paper we show that the Nambu mechanical structure is also hidden in some quantum or semiclassical dynamics, that is, in some cases the quantum or semiclassical time evolution of expectation values of quantum mechanical operators, including composite operators, can be formulated as Nambu mechanics. We present a procedure to find hidden Nambu structures in quantum/semiclassical systems of one degree of freedom, and give two examples: the exact quantum dynamics of a harmonic oscillator, and semiclassical wave packet dynamics. Our formalism can be extended to many-degrees-of-freedom systems; however, there is a serious difficulty in this case due to interactions between degrees of freedom. To illustrate our formalism we present two sets of numerical results on semiclassical dynamics: from a one-dimensional metastable potential model and a simplified Henon--Heiles model of two interacting oscillators.
Introduction
In 1973, Nambu proposed a generalization of the classical Hamiltonian dynamics [1] that is nowadays referred to as the Nambu mechanics. In his formulation, the phase space spanned by the canonical doublet (q, p) is extended to that spanned by N (≥ 3) variables (x 1 , x 2 , ..., x N ), the Nambu N -plet, and the Hamilton equations of motion are generalized to the Nambu equations. In order for the Liouville theorem to hold in the N -dimensional extended phase space, the Nambu equations are defined by N − 1 Nambu Hamiltonians and the Nambu bracket, an N -ary generalization of the Poisson bracket. The structure of Nambu mechanics has impressed many authors, who have reported studies on its fundamental properties and possible applications, including quantization of the Nambu bracket [2][3][4][5][6][7][8][9][10][11][12]. However, the applications to date have been limited to particular systems, because Nambu systems generally require multiple conserved quantities as Hamiltonians and the Nambu bracket exhibits serious difficulties in systems with many degrees of freedom or quantization [1,2,11].
In 2013 we proposed a new approach to Nambu mechanics [13]. We revealed that the Nambu mechanical structure is hidden in a Hamiltonian system which has redundant degrees of freedom. For example, in a Hamiltonian system with a Hamiltonian H(q, p), if we take three variables as (x 1 , x 2 , x 3 ) = (q, p, q 2 ), their classical time evolution can be given by N = 3 Nambu equations with two Hamiltonians F (x 1 , x 2 , x 3 ) and G(x 1 , x 2 , x 3 ). Here x 3 = q 2 is a redundant degree of freedom in the original Hamiltonian system, and the Nambu Hamiltonians are given by the original Hamiltonian F (x 1 , x 2 , x 3 ) = H(q, p) and the constraint G(x 1 , x 2 , x 3 ) = x 3 − x 2 1 = 0, which is induced due to the consistency between the three variables. We derived the consistency condition to determine the induced constraints.
In the present paper we show that the Nambu mechanical structure is also hidden in some quantum or semiclassical systems. The key idea is as follows. In our previous work, the Nambu multiplet is given as a function of classical variables (q, p), and therefore the induced constraints are always trivial, i.e. set to zero [13]. However, if we take the Nambu multiplet as a set of expectation values of quantum mechanical operators including composite operators (q 2 ,p 2 , ...), the constraints become nontrivial because of quantum fluctuation. Furthermore, if these constraints are constants of motion, the time evolution of the Nambu multiplet could be given by the Nambu equations. For example, consider a classical system with a Hamiltonian H(q, p) and a corresponding quantum system with the HamiltonianĤ = H(q,p). If we take three variables as (x 1 , x 2 , x 3 ) = (q, p, q 2 ), the trivial constraint G = x 3 − x 2 1 = 0 is induced. Then, if we replace these variables with (x 1 , x 2 , x 3 ) = ( q , p , q 2 ), the same function G = x 3 − x 2 1 has a nonzero value in general because of quantum fluctuation. Furthermore, in the case of frozen Gaussian wave packet dynamics [14], which is the dynamics of a Gaussian wave packet with a fixed width σ, the function G = x 3 − x 2 1 = σ 2 is constant in time and therefore the quantum or semiclassical time evolution of the Nambu triplet can be given by the N =3 Nambu equations with Nambu Hamiltonians F and G. Here, F is equal to or approximately equal to the expectation value of the Hamiltonian operator F = Ĥ or F ≃ Ĥ . We present a general procedure to find the Nambu mechanical structure in quantum or semiclassical systems of one degree of freedom with some specific examples. It should be noted that our formulation is not a quantization of the Nambu bracket. We just propose a prescription to describe ordinary quantum or semiclassical dynamics in a classical Nambu mechanical manner.
The Nambu mechanical structure is hidden not only in one-degree-of-freedom systems. It is straightforward to extend our formalism to many-degrees-of-freedom systems by extending the definition of the Nambu bracket. However, the resulting hidden Nambu mechanics becomes pathological because in many-degrees-of-freedom systems the Nambu bracket does not satisfy the fundamental identity, which is an important property of the Nambu bracket and corresponds to the Jacobi identity in Hamiltonian dynamics [2,11]. Without the Jacobi identity the canonical transformation of the canonical doublet cannot be properly defined, and the dynamics becomes anomalous [15]. Similarly, without the fundamental identity we cannot define the canonical transformation of Nambu multiplets including their consistent time evolution. The hidden Nambu mechanics in many-degrees-of-freedom systems is an example of dynamics without the canonical structure.
The outline of this article is as follows. In Sect. 2 we briefly review our previous work on the hidden Nambu mechanics in classical Hamiltonian systems [13]. As preparation for the next section, we give a detailed description of some examples. In Sect. 3 we present a procedure to find the Nambu mechanical structure hidden in some quantum or semiclassical 2/18 dynamics. Two examples are given: the exact quantum dynamics of a harmonic oscillator and the semiclassical nonlinear dynamics of a frozen Gaussian wave packet. In Sect. 4 we give an extension of our formalism to many degrees of freedom. In Sect. 5 we present two numerical results to illustrate our formalism: the semiclassical tunneling dynamics in a onedimensional metastable system and the semiclassical energy exchange dynamics between two coupled oscillators in a simplified Henon-Heiles model. In the last section we give our conclusions and discuss the direction of future work.
Hidden Nambu mechanics
We begin with a brief review of Hamiltonian dynamics, Nambu mechanics [1], and hidden Nambu mechanics [13]. We describe two examples in detail to prepare for the next section. In this and the next section we consider only one-degree-of-freedom systems.
Hamiltonian dynamics
Hamiltonian dynamics is the classical dynamics of the canonical doublet (q(t), p(t)), which is given by a Hamiltonian H = H(q, p) and the Poisson bracket defined by the two-dimensional Jacobian, where A = A(q, p) and B = B(q, p) are any functions of the canonical doublet. The Poisson bracket should satisfy the Jacobi identity, where A 1 = A 1 (q, p), A 2 = A 2 (q, p), and B = B(q, p) are any functions. In terms of the Poisson bracket, the Hamilton equation of motion for any function f = f (p, q) can be written as The time evolution according to this equation preserves the phase space volume because of the divergenceless property, This is the Liouville theorem in Hamiltonian dynamics.
Nambu mechanics
Nambu mechanics is a generalized Hamiltonian dynamics of N (≥ 3) variables (x 1 , x 2 , ..., x N ) [1]. In Nambu mechanics the canonical doublet is generalized to the Nambu N -plet, and the Poisson bracket, Eq. (1), is generalized to the Nambu bracket defined by means of the 3/18 N -dimensional Jacobian, where A a = A a (x 1 , x 2 , ..., x N ) (a = 1, ..., N ) are any functions of the Nambu multiplet and ε i1i2···iN is the N -dimensional Levi-Civita symbol, the antisymmetric tensor with ε 12···N = 1. The Nambu bracket should satisfy the following fundamental identity [2], an N -ary generalization of the Jacobi identity in Eq. (2): .., N − 1) are any functions. In terms of the Nambu bracket, the Nambu equation for any function f = f (x 1 , x 2 , ..., x N ) can be written as where .., N − 1) are Nambu Hamiltonians. The time evolution according to this equation preserves the N -dimensional phase space volume because of the divergenceless property, Therefore the Liouville theorem also holds in Nambu mechanics.
Hidden Nambu mechanics
Consider a Hamiltonian system of a canonical doublet (q, p) with a Hamiltonian H = H(q, p).
The key idea of hidden Nambu mechanics is to describe this system by means of N (≥ 3) variables x i = x i (q, p) (i = 1, ..., N ). We assume that at least N − 1 of {x i , x j } PB do not vanish, so that the time evolution of any functionf (x 1 , ..., x N ) = f (q, p) can be written via Hamilton equation of motion in Eq. (3), where F (x 1 , ..., x N ) = H(q, p). We introduce the functions G c = G c (x 1 , ..., x N ) (c = 1, ..., N − 2) which satisfy the consistency conditions Then, Eq. (9) can be rewritten as the Nambu equation in the form of Eq. (7), 4/18 where we have used the following formula concerning Jacobians: The functions G c are constants in motion and can be set to zero by redefining G c . This is a natural choice because the functions G c work as constraints for the Nambu multiplet (x 1 , x 2 , ..., x N ). We refer to G c = 0 as induced constraints, because they are induced by enlarging the phase space from (q, p) to (x 1 , x 2 , ..., x N ).
Examples
Here we present detailed descriptions of two simple examples to show how induced constraints are obtained for given multiplets. We adopt the same choice of N -plets in the next section. Finally we comment on the functional forms of the Nambu Hamiltonians.
(a) N = 3: classical harmonic oscillator Consider three composite variables of the canonical doublet, which satisfy the following relations: Then, the conditions in Eq. (10) become and G is solved as G = 2x 2 3 − 2x 1 x 2 + C with a constant C. Redefining G to eliminate the constant, we obtain the induced constraint As an example of the dynamics of the Nambu triplet in Eq. (13), consider a one-dimensional harmonic oscillator whose Hamiltonian is given by The Hamilton equations of motion for the triplet are as follows: Let us derive these equations from the N = 3 Nambu equations with two Nambu Hamiltonians (F, G). One of the Hamiltonians, F , is equal to the original Hamiltonian H(q, p), Eq. 5/18 (17), and the other Hamiltonian, G, is given by the induced constraint, Eq. (16). The N = 3 Nambu equations are and each equation is given by These equations are equivalent to the Hamilton equations of motion in Eq. (18).
(b) N = 4: classical nonlinear systems Consider four variables, two of them being composites, which satisfy the following relations: Then, the conditions in Eq. (10) become and G 1 and G 2 are given by where C 1 and C 2 are constants. By redefining G 1 and G 2 , we obtain the induced constraints As an example of the dynamics of the Nambu quartet in Eq. (22), consider a onedimensional nonlinear system whose Hamiltonian is given by where V (q) is an anharmonic potential. The Hamilton equations of motion for the quartet are written as follows: Let us derive these equations from the N = 4 Nambu equations with three Nambu Hamiltonians (F, G 1 , G 2 ). One of the Hamiltonians, F , is equal to the original Hamiltonian whereṼ (x 1 , x 3 ) = V (q). The other two Hamiltonians, G 1 and G 2 , are given by the induced constraints in Eqs. (25) and (26). The N = 4 Nambu equations are and each equation is given by These equations are equivalent to the Hamilton equations of motion in Eq. (28).
Some comments
Here we make some comments on the functional forms of Nambu Hamiltonians. In some cases, the functional form of (G 1 , ..., G N −2 ) cannot be determined uniquely. For example, for (x 1 , x 2 , x 3 , x 4 ) = (q, p, q 2 , q 3 ), one of the Poisson brackets in the consistency condition in Eq. (10) is given by 3 1 , respectively. Although we can choose either expression in classical mechanics, we must choose the latter expression, {x 2 , x 4 } PB = −3x 3 , in quantum or semiclassical mechanics. As shown in the next section, we must express the Poisson brackets in the consistency condition of Eq. (10) using variables of the highest order possible.
Also, in some cases we cannot uniquely determine the functional form of the Hamiltonian F in classical mechanics. In the next section we present a prescription to determine the functional form of F in quantum or semiclassical systems.
3. Hidden Nambu mechanics in quantum/semiclassical systems 3.1. Quantum/semiclassical dynamics Consider a quantum system of a doublet (q,p) with a Hamiltonian operator The dynamics of a quantum operator = A(q,p) is given by the Heisenberg equation, In this work we focus on the dynamics of the expectation value of the quantum operator,  (t) = ψ|Â(t)|ψ , where |ψ is a quantum state. The time evolution of  (t) is given by 7/18 taking the expectation value of both sides of the Heisenberg equation, This equation gives the exact quantum dynamics, and we can consider several approximated dynamics. The lowest-order approximation is simply the classical Hamiltonian dynamics, A systematic approximation scheme to derive higher-order semiclassical dynamics, the quantized Hamiltonian dynamics [16,17], has been developed.
In quantum or semiclassical systems there might exist conserved quantities other than Ĥ = H(q,p) , the expectation value of the original Hamiltonian. Moreover, in some cases such conserved quantities might be identified as the constraints G c in the hidden Nambu mechanics. This means that the Nambu structure could be hidden in quantum or semiclassical systems with nontrivial constraints G c = 0, which are trivial (G c = 0) in classical systems.
How to find the hidden Nambu structure
The procedure to find the Nambu structure hidden in quantum or semiclassical systems is as follows.
Step (1) Step (2): Consider a quantum system of a doublet (q,p) with a HamiltonianĤ = H(q,p), Eq. (32), which corresponds to the classical Hamiltonian H(q, p). Replace the Nambu N -plet with the corresponding expectation values of quantum operators. For example, Step (3): Determine the functional form of F (x 1 , x 2 , ..., x N ) by representing Ĥ as a function of the Nambu N -plet. If Ĥ includes an expectation value Ô which is not a member of the Nambu N -plet, we reduce Ô to a function of the Nambu N -plet by means of the zero-cumulant approximation that ignores the cumulant, For example, for (x 1 , x 2 , x 3 ) = ( q , p , q 2 ), if Ĥ includes q 4 , it is approximated as q 4 ≃ 3 q 2 2 − 2 q 4 = 3x 2 3 − 2x 4 1 by means of q 4 c ≃ 0 followed by q 3 c ≃ 0.
Step (4): The other Nambu Hamiltonians G c (c = 1, ..., N − 2) are given by the same functional forms as the trivial constraints. They are in general nontrivial, G c = 0, because of quantum fluctuation. 8/18 Step (5): If the quantities (F, G 1 , ..., G N −2 ) are all conserved in quantum or some semiclassical dynamics, the dynamics of the Nambu N -plet can be cast into the Nambu form in Eq. (11).
The zero-cumulant approximation is similar to the approximation adopted in the quantized Hamiltonian dynamics [16]. However, it is not the only approximation for the Hamiltonian F in the hidden Nambu mechanics. It is also possible to consider an approximation that ignores the quantum fluctuation, for example (q − q ) n ≃ 0. As for the example shown in Step (3), this approximation leads to q 4 ≃ 6 q 2 q 2 − 5 q 4 = 6x 3 x 2 1 − 5x 4 1 . The resulting Nambu equations for quantum/semiclassical systems are the same as the ones for classical systems, and quantum properties are introduced through nonzero constraints G c = 0 (c = 1, ..., N − 2). This might imply that the replacement where C c are nonzero constants, could be regarded as a kind of "quantization" scheme for the Nambu mechanics. However, as opposed to various attempts to quantize the Nambu mechanics proposed so far [3][4][5][6]11], this replacement only gives a scheme to quantize the hidden Nambu mechanics. Furthermore, this replacement is incomplete even as a quantization scheme for the hidden Nambu mechanics, because the constants C c in general depend on the models and the initial conditions. Therefore the procedure presented here is not for quantizing the Nambu mechanics, but just for finding the hidden Nambu structures in quantum/semiclassical systems. The resulting hidden Nambu mechanics is a volume-preserving dynamics of the expectation values of the quantum operators. If the Nambu multiplet includes the variables ( q , p ), the Nambu mechanics can be regarded as a kind of quantized Hamiltonian dynamics [16]. In some cases, such Nambu mechanics can be reduced to the effective Hamiltonian dynamics by explicitly solving the constraints G c = 0. In the next subsection we will see an example of such reduction.
Examples
Here we present two examples; one is an example of exact quantum dynamics and the other is semiclassical. In the latter example, the resulting Nambu mechanics can be reduced to the Hamiltonian dynamics with the effective Hamiltonian. They correspond to the examples shown in Sect. 2.4, and therefore we will show the procedure after Step (2).
(a) N = 3: quantum harmonic oscillator Consider three expectation values which correspond to the Nambu triplet in Eq. (13), The quantum Hamiltonian of a one-dimensional harmonic oscillator is given bŷ Then, one of the Nambu Hamiltonians, F , can be obtained without any approximation, The other Nambu Hamiltonian, G, is given by the same functional form as Eq. (16), which is nonzero in general due to quantum fluctuation. We can see that both F and G are conserved in the exact quantum dynamics, The Nambu equations of Eq. (21) are equivalent to these exact equations; that is, the Nambu structure is hidden in the exact quantum dynamics of a harmonic oscillator.
(b) N = 4: semiclassical nonlinear systems Consider four expectation values which correspond to the Nambu quartet in Eq. (22), The quantum Hamiltonian of a one-dimensional nonlinear system is given byĤ = H(q,p), Eq. (32), with an anharmonic potentialV = V (q). The Nambu Hamiltonian F can be obtained as an approximation of Ĥ , where the functional form of the reduced potentialṼ (x 1 , x 3 ) is uniquely determined by means of the zero-cumulant approximation, Eq. (36). The other Nambu Hamiltonians are given by the same functional forms as Eqs. (25) and (26), G 1 = x 3 − x 2 1 and G 2 = x 4 − x 2 2 , which are nonzero in general due to quantum fluctuation. We can see that all of F , G 1 , and G 2 are conserved in the following approximated dynamics: wheref is determined by the zero-cumulant approximation in Eq. (36) if necessary. This is a semiclassical dynamics which corresponds to the lowest order of the quantized Hamiltonian dynamics [17]. We can also see that the N = 4 Nambu equations of Eq. (31) are equivalent to these semiclassical equations; that is, the Nambu structure is hidden in the semiclassical nonlinear dynamics. This semiclassical dynamics can be regarded as the frozen Gaussian wave packet dynamics [14], where the quantum wave function is approximated by a Gaussian wave packet with a constant width σ, Here q c is the center of the wave packet and p c is that in momentum space. The time evolution of the variables (q c , p c ) can be determined by means of the time-dependent variational 10/18 principle [18], by taking the frozen Gaussian wave function of Eq. (46) as a trial function. The resulting variational equations are semiclassical equations which have the same forms as the Hamilton equations of motion in Eq. (3), Here H c (q c , p c ) = ψ FG |Ĥ|ψ FG is the effective Hamiltonian modified by the quantum correction.
Evaluating the expectation values in Eq. (43) by means of the state |ψ FG (t) , we obtain . Then, G 1 and G 2 are given by Using these nontrivial constraints, we can show that the Nambu Hamiltonian F in Eq. (44) is equivalent to H c (q c , p c ). This is because the zero-cumulant approximation for the Nambu quartet in Eq. (43) is exact in the case of the frozen Gaussian wave packet. Using Eqs. (49) and (50), we can also show that the Nambu equations of Eq. (31) are reduced to the variational equations of Eq. (48). That is, the Nambu structure is hidden in the semiclassical dynamics of the frozen Gaussian wave packet. In Sect. 5.1 we present a numerical demonstration of the semiclassical tunneling dynamics in a metastable system.
Many-degrees-of-freedom extension
It is straightforward to extend our formalism to many-degrees-of-freedom systems. However, the resulting classical or quantum/semiclassical hidden Nambu mechanics becomes pathological, because the Nambu bracket itself has a serious problem in interacting systems [1,2,11].
Difficulties in the Nambu bracket
where A and B are any functions of the 2n variables. Since the dynamics is divergenceless, the Liouville theorem holds. The Poisson bracket in Eq. (51) satisfies the Jacobi identity of Eq. (2) and therefore we can define canonical transformations of the 2n variables in a consistent manner.
11/18
On the other hand, the Nambu mechanics has a problem in the many-degrees-of-freedom extension. Consider a system of n Nambu N -plets (x where A a (a = 1, ..., N ) are any functions of the N × n variables. Because of the divergenceless property, the Liouville theorem holds. The Nambu bracket of Eq. (53), however, fails to satisfy the fundamental identity in Eq. (6) if the N -plets interact with each other [2,11]. Therefore we cannot define consistent canonical transformations of the N × n variables in general [1].
Hidden Nambu mechanics in many-degrees-of-freedom systems
Although the Nambu bracket has a serious problem in many-degrees-of-freedom systems, it is still possible to extend our hidden Nambu formalism to such systems. The resulting hidden Nambu mechanics is the Nambu mechanics without the fundamental identity.
We start from a Hamiltonian system of n canonical doublets with a Hamiltonian H = H(q (1) , p (1) , ..., q (n) , p (n) ). Then we introduce N variables x i (q (α) , p (α) ) (i = 1, ..., N ) for each α. We assume that at least N − 1 of ∂(x N ) = f (q (1) , p (1) , ..., q (n) , p (n) ) can be written via the Hamilton equation of motion, where F (x N ) (c = 1, ..., N − 2) that satisfy the consistency conditions for each α, then Eq. (55) can be rewritten in the same form as Eq. (7), where the Hamiltonians G c are defined as the sum of each G The Nambu mechanics is hidden in classical many-degrees-of-freedom systems. The Liouville theorem holds in the hidden mechanics, though the fundamental identity does not hold. Such hidden Nambu mechanics is an example of dynamics without the canonical structure [15]. Taking the same procedure presented in Sect. 3.2, we can find the Nambu structure in quantum/semiclassical many-degrees-of-freedom systems: Step (1): Start from the classical hidden Nambu mechanics shown above.
Step (2): Replace the Nambu N -plets with the corresponding expectation values of quantum operators. For example, Step (3): Determine the functional form of F . Use the zero-cumulant approximation if necessary.
In Sect. 5.2 we present a numerical demonstration of the semiclassical dynamics of a twobody system.
Numerical results for semiclassical dynamics
Here we give two numerical results for N = 4 hidden Nambu mechanics equivalent to the semiclassical frozen Gaussian wave packet dynamics in one-and two-degrees-of-freedom systems. We choose the same systems as used in the applications of the quantized Hamiltonian dynamics [16]. We compare the results with the corresponding quantum and classical results. In both systems, the time developments in the Nambu and classical mechanics are numerically evaluated by using the fourth-order Runge-Kutta integrator, 1 while the propagation of the quantum wave function is numerically evaluated by a split-operator method, which is a hybrid of Cayley's form and the Suzuki-Trotter decomposition [20]. As the initial wave function for the quantum dynamics, we take the Gaussian wave packet ψ FG (q, 0), Eq. (46). The initial conditions for the Nambu mechanics are given by the expectation values of the quantum operators with respect to that initial state, and those for the classical mechanics are given by the center of the initial wave packet (q c (0), p c (0)). We choose the width of the initial wave packet as σ = /(2mω), for which the frozen Gaussian wave packet dynamics becomes exact for a harmonic oscillator.
Metastable cubic potential
The first model is a quantum system which exhibits tunneling. Consider a one-dimensional metastable system whose quantum Hamiltonian is given by 2 The corresponding classical Hamiltonian is H = (1/2m)p 2 + V (q), where V (q) is the classical potential, V (q) = (mω 2 /2)q 2 + (g/3)q 3 . We choose N = 4 Nambu variables, as in Eq. (43), and the Nambu Hamiltonian F is then determined by the zero-cumulant approximation in Eq. (36), For the frozen Gaussian wave packet dynamics, the Nambu Hamiltonians F and (G 1 , G 2 ), Eqs. (49) and (50), are conserved in the time evolution according to the semiclassical equations of Eq. (45), which are equivalent to the N = 4 Nambu equations of Eq. (31). The initial conditions for the Nambu mechanics are given as follows: The classical potential V (q) is plotted in Fig. 1a, where we set the parameters ω = 1 and g = 0.3 with the units = m = 1. These parameters are the same as in Ref. [16]. The initial wave function given by |ψ(q, 0)| 2 = |ψ FG (q, 0)| 2 is also shown in Fig. 1a, where we choose the initial conditions as (q c (0), p c (0)) = (0, 1.8). The initial wave packet is located at the local minimum of the classical potential V (q) and moves to the right. The calculated trajectories of the quantum, Nambu, and classical mechanics are shown in Fig. 1b. The quantum mechanical expectation value q(t) moves to the right, bounces off the wall, and moves to the left through the potential barrier, the top of which is located at q = −3.3. This is an instance of quantum mechanical tunneling because the classical variable q(t) fails to go through the potential barrier and oscillates around the local minimum of V (q). On the other hand, the Nambu variable x 1 (t) can reproduce the quantum mechanical tunneling, although it deviates from the quantum result as time increases. This semiclassical behavior of the Nambu variable can be understood as follows. The N = 4 Nambu mechanics discussed here is equivalent to the variational dynamics of (q c , p c ), whose time evolution is given by Eq. (48). The effective Hamiltonian is H c (q c , p c ) = (1/2m)p 2 c + V c (q c ), where V c (q c ) 2 This metastable system has also been used in the applications of the symplectic semiclassical wave packet dynamics [21]. 14/18 is the effective potential shown in Fig. 1a, The last two terms are proportional to the Planck constant and generated by the quantum correction. As shown in Fig. 1a, these terms lower the height of the potential barrier, and there exists a region of initial values (q c (0), p c (0)) where the Nambu mechanics can tunnel but the classical mechanics cannot. The initial conditions adopted here, (q c (0), p c (0)) = (0, 1.8), are in such a region.
Simplified Henon-Heiles model
The second model is a quantum system which exhibits nonlinear energy exchange dynamics between coupled oscillators. Consider a one-dimensional quantum system of two oscillators whose Hamiltonian is given bŷ 6) is zero, whereas the right-hand side is −λ. That is, the interaction between two oscillators violates the fundamental identity and the canonical structure is broken in the hidden Nambu mechanics. However, by explicitly solving the constraints in Eqs. (68) and (69), this Nambu mechanics can be reduced to the effective Hamiltonian dynamics where the time evolution of the canonical doublets can be properly defined. Therefore the dynamics of two oscillators considered here is anomalous as the Nambu mechanics, but not anomalous as the Hamiltonian dynamics.
Conclusions and future work
We have shown that the Nambu mechanical structure is hidden not only in classical Hamiltonian dynamics but also in some quantum or semiclassical dynamics. We focused on the dynamics defined in an extended phase space spanned by N (≥ 3) quantum mechanical expectation values. 3 The dynamics of variables such as ( q , p , q 2 , p 2 ) cannot be described by the Hamilton equations of motion; however, if the system has a sufficient number of conserved quantities, (F, G 1 , G 2 ), their dynamics could be described by the N = 4 Nambu equations. We gave some quantum/semiclassical examples of hidden Nambu mechanics, including a many-degrees-of-freedom system. It would be interesting to investigate other examples.
In many-degrees-of-freedom systems, however, the hidden Nambu mechanics become anomalous, because interactions between multiple degrees of freedom violate the fundamental identity of Eq. (6) [2,11]. Since the fundamental identity would play a similar role to the Jacobi identity in the Hamiltonian dynamics, its violation implies that it would be difficult to formulate the Nambu statistical mechanics or quantize the Nambu mechanics. On the other hand, in Hamiltonian dynamics there also exists anomalous dynamics known 3 Our formalism could also be applied to statistical-mechanical expectation values.
17/18
as nonholonomic dynamics [23], where the Jacobi identity is violated and the Hamiltonian structure is broken [15]. Recently, a procedure has been proposed to recover the Hamiltonian structure and formulate a statistical theory of the nonholonomic dynamics [24]. This work might provide guidance for formulating a statistical theory of Nambu mechanics, and our formalism presented in this article might provide example systems suitable for Nambu statistical mechanics to be tested. | 7,544.6 | 2019-08-09T00:00:00.000 | [
"Physics"
] |
Synthesis, In Vitro α-Glucosidase Inhibitory Activity and Molecular Docking Study of New Benzotriazole-Based Bis-Schiff Base Derivatives
This study was carried out to synthesize benzotriazole-based bis-Schiff base scaffolds (1–20) and assess them in vitro for α-glucosidase inhibitory potentials. All the synthetics analogs based on benzotriazole-based bis-Schiff base scaffolds were found to display an outstanding inhibition profile on screening against the α-glucosidase enzyme. The synthetic scaffolds showed a varied range of inhibition profiles having IC50 values ranging from 1.10 ± 0.05 µM to 28.30 ± 0.60 µM when compared to acarbose as a standard drug (IC50 = 10.30 ± 0.20 µM). Among the series, fifteen scaffolds 1–3, 5, 6, 9–16, 18–20 were identified to be more potent than standard acarbose, while the five remaining scaffolds 4, 7, 8, 16, and 17, also showed potency against the α-glucosidase enzyme but were found to be less potent than standard acarbose. The structure of all the newly synthesized scaffolds was confirmed using different spectroscopic techniques such as HREI-MS and 1H- and 13C- NMR spectroscopy. To find a structure-activity relationship, molecular docking studies were carried out to understand the binding mode of the active inhibitors with the active sites of the enzyme and the results supported the experimental data.
Introduction
Diabetes mellitus (DM) is a well-known, progressive endocrine disorder associated with increased morbidity and mortality, as well as high healthcare costs. There were approximately 171 million cases of DM in 2000, and this number is expected to more than double over the next 25 years, reaching 366 million by 2030 [1,2]. This rising trend has prompted the development of new therapeutic agents as a serious global medical concern. Due to deficiencies in insulin production or action, DM is characterized by hyperglycemia and changes in the metabolisms of carbohydrates, proteins, and lipids [3]. Postprandial hyperglycemia is a prominent defect that occurs early in diabetes and may lead to various secondary complications, including elevated risk for cardiovascular diseases [4], atherosclerosis, cataracts, retinopathy, neuropathy, nephropathy, and impaired wound healing [5]. α-glucosidase (EC 3.2.1. 20), an enzyme found in the epithelial mucosa of the small intestine that cleaves the glycosidic bonds in complex carbohydrates to release absorbable monosaccharide, aids in the rapid absorption of carbohydrates, which causes elevated blood glucose levels when consumed [6].
One therapeutic approach for treating diabetes involves controlling postprandial hyperglycemia by inhibiting the α-glucosidase in the digestive tract, delaying and prolonging the overall carbohydrate digestion time. Slowing carbohydrate digestion should reduce the rate of glucose absorption and consequently prevent spikes in postprandial blood glucose and insulin levels [7,8]. Using α-glucosidase inhibitors has become a promising therapeutic strategy for reducing the risks of diabetes and other carbohydrate-mediated diseases, including hyperlipoproteinemia and obesity [9][10][11][12][13][14][15].
Benzotriazole scaffolds were reported as key precursors in designing and developing biologically more-active drugs and were known to have a broad range of biological activities such as antibacterial [16], antifungal [17,18], antihistaminic, anti-adrenergic, and analgesic [19], anti-cancer [20] and anti-convulsant activities [21].
Viewing the biological importance of benzotriazole [22] and bis-Schiff base analogs [23], herein this study we plan to synthesize a library of new benzotriazole-bearing bis-Schiff base derivatives as potent α-glucosidase inhibitors. Therefore, in the current study, we synthesized unreported benzotriazole-bearing bis-Schiff base analogs. We believe such compounds are more lipophilic and will increase activity and easily pass through the cell wall ( Figure 1).
In Vitro α-Glucosidase Inhibition Profile
To treat various diseases, the discovery and development of more potent drugs through the inhibition of enzymes has received much attention from medicinal chemists in the past few years. Benzotriazole-based bis-Schiff base scaffolds (1-20) were synthesized and then investigated in vitro for α-glucosidase inhibitory potentials. All the synthetic analogs based on benzotriazole-based bis-Schiff base scaffolds were found to display outstanding inhibition profiles on screening against the α-glucosidase enzyme. The synthetic scaffolds showed a varied range of inhibition profiles having IC 50 values ranging from 1.10 ± 0.05 µM to 28.30 ± 0.60 µM when compared to acarbose as a standard drug (IC 50 = 10.30 ± 0.20 µM) (Table 1). Among the series, fifteen scaffolds 1-3, 5, 6, 9-16, 18-20 were identified to be more potent than standard acarbose, while the five remaining scaffolds 4, 7, 8, 16, and 17 also showed potency against the α-glucosidase enzyme but were found to be less potent than standard acarbose.
Structure-Activity Relationship (SAR) for α-Glucosidase Enzyme Structure-activity relationship (SAR) studies suggested that the variation in number/s, position, and nature of substituents around both aryl parts 'B' and 'C' of the benzotriazole-based bis-Schiff base scaffolds (1-20) greatly affect the inhibitory potentials of the α-glucosidase enzyme. Compounds (1, 9, 10, and 13), bearing tri-substitutions such as di-Cl at the 3,5-position, -OH group at the 2-position of aryl part 'C' and a variety of other groups such as -OCH 3 , di-Cl and -NO 2 at various position of aryl part 'B', were shown to display considerable inhibitory potentials against the α-glucosidase enzyme. Scaffold (9), which has a -NO 2 moiety at the 3-position of aryl part 'B' and tri-substitutions -OH and di-Cl moieties present at the 2,3,5-position of aryl part 'C', was identified as the most active inhibitor of the αglucosidase enzyme with nine-fold greater potency than standard acarbose. However, scaffold (10) showed somewhat less potency than scaffold (9) even though it also bears a substituent of the same nature; a slight difference in inhibitory potentials may be attributed to an alteration in the position of the -NO 2 moiety on aryl part 'B'. The enhanced inhibitory potentials of both these analogs (9) and (10) against α-glucosidase may perhaps be due to greater numbers of e-withdrawing groups (di-Cl and -NO 2 ) attached to both aryl parts 'B' and 'C', which make both aryl rings more susceptible for interactions with the active residues of the α-glucosidase enzyme. Moreover, attached substituents such as -OH and -NO 2 groups (capable of interactions through H-bonding with the active site of targeted enzyme) also enhanced the enzymatic activity of scaffolds (9) and (10). Compound (1) displayed a four-fold lower potency than compound (9). This decline in the inhibition profile of analog (1) was due to the replacement of the -NO 2 moiety with a -OCH 3 group, showing that the nature of attached substituents greatly affects the potency of the synthe-sized scaffolds against the α-glucosidase enzyme. It was also observed that replacement of the -OCH 3 group of compound (1) with a hydrogen atom followed by subsequent addition of di-Cl moieties at the 2,4-position of aryl part 'B' enhanced the α-glucosidase activity by three-fold, as in scaffold (13).
It seemed from the SAR studies ( Figure 2) that the attachment of groups of a bulky nature to either side of both aryl parts 'B' and 'C' resulted in decreased inhibitory potentials against the α-glucosidase enzyme. When the -NO 2 moiety of scaffold (10) present at the 4-position of aryl part 'B' was replaced by a bulkier -Br moiety, as in compound (5), the enzymatic activity against the α-glucosidase enzyme was lowered four-fold, showing that bulky natured substituents cause steric hindrance because of which the interaction caused by the synthetic analogs with the active residues of the α-glucosidase enzyme was diminished and, hence, the activity reduced. Among the scaffolds (3, 5, 6, and 17), analog (5) bearing the -OH moiety at the 2-position and di-Cl moieties at the 3,5-position of aryl part 'C' along with a -Br group at the 4-position of aryl part 'B' was identified as the most potent inhibitor of α-glucosidase. Scaffold (3), which holds a -Br moiety at the 4-position of aryl part 'B' and a -NO 2 group also at the 3-position of aryl part 'C', seemed to be a better competitor of α-glucosidase. However, its counterpart (6), which, although it holds substituents of the same nature, has a -NO 2 group at the 4-position of aryl part 'B', demonstrated half the potency of compound (3) against the α-glucosidase enzyme. Moreover, the activity of scaffold (5) was decreased ten-fold by de-attachment of the -OH group followed by shifting of the 5-Cl to the 4-position of aryl part 'C', as in scaffold (17). This lowered activity of compound (17) in comparison to scaffold (5) may be attributed to the lack of a group (-OH) that can cause a strong interaction with the active amino acid residues through conventional hydrogen bonding. The scaffolds bearing the -Br moiety at aryl part 'B' and a -NO 2 group either at the 4-position of aryl part 'C', as in compound (6), or at the 3-position of aryl part 'C' (3), displayed better potency against the α-glucosidase enzyme than compound (17) bearing di-Cl moieties at the 3,4-position of aryl part 'C' along with a -Br moiety at the 4-position of aryl part 'B'. This superior potency of scaffolds (6) and (3) in comparison to scaffold (17) was due to better interaction of the -NO 2 moiety through conventional hydrogen bonding with the active amino acid residues as well as the -NO 2 group providing a strong e-withdrawing effect making the aryl part 'C' have a pi-cationic interaction with the active residues of α-glucosidase enzyme and, hence, enhanced the enzymatic activity. In addition, the attachment of substituents of stronger e-withdrawing effect (di-Cl) and a substituent (-OH) capable of interaction through stronger forces with the active residues of α-glucosidase, as in compound (5), enhanced the inhibitory potentials more greatly than compounds (3), (6) and (17) that lack substituents of this nature.
It was noteworthy that encouraging inhibitory potentials against the α-glucosidase enzymes were shown by scaffolds (2, 11, 12, and 14) that hold a -NO 2 moiety either at the 4-position or the 3-position of aryl part 'C' along with a variety of substitutions (-NO 2 , -OCH 3 and di-Cl) present at different position of aryl part 'B' (Figure 3). Among these scaffolds, scaffold (11) proved to be a potent inhibitor of the α-glucosidase enzyme. This scaffold holds a -NO 2 moiety on both aryl parts 'B' and 'C', but they have different positions on both sides of aryl parts 'B' and 'C'. Aryl part 'B' has the -NO 2 group at the 4-position while aryl part 'C' has the -NO 2 moiety at the 3-position. The inhibition profile of scaffold (11) fell by half by shifting the -NO 2 moiety of aryl part 'C' from the 3-position to the 4-position, as in compound (12), indicating that inhibitory potentials were affected by changing the position of the substituent around aryl part 'C'. Moreover, it was also observed that the nature of the substituent also greatly affects the inhibitory activity of the synthetic scaffold; therefore, compound (2), which holds a -OCH 3 moiety at the 3-position of aryl part 'B' along with a -NO 2 moiety at the 3-position of aryl part 'C', displayed a two-fold lower potency than compound (12). Both compounds (2) and (12) hold the same -NO 2 moiety at the 3-position of aryl part 'C' but have different substituents (-OCH 3 and -NO 2 ) around aryl part 'B'. Furthermore, it was noted that the addition of di-Cl moieties at aryl part 'B' instead of the -OCH 3 moiety resulted in enhanced activity, as in the case of compound (14), showing that the introduction of a substituent with a strong e-withdrawing nature in greater numbers makes the ring susceptible for interaction with the active residues of amino acid through pi-cationic interaction and, hence, enhanced the enzymatic activity ( Figure 4). Analog (20), which holds the -NO 2 moiety at the 4-position of aryl ring 'B' and the -OH moiety also at the 4-position of aryl part 'C', emerged as a better competitor of α-glucosidase. This better potency shown by the -NO 2 and -OH moieties-bearing analog was due to strong interaction with the active amino acid residues. An oxygen atom of both -OH and -NO 2 has a strong tendency to bind well via strong conventional H-bonding with the inner cavity of α-glucosidase and hence, enhanced α-glucosidase activity. However, a decrease in the α-glucosidase activity of analog (20) was seemed by de-attachment of -OH moiety of aryl part 'C' and consequent attachment of di-Cl moieties at the 3,4-position of aryl part 'C' as in case of scaffold (19). This discrepancy in the inhibition profile of these scaffolds was because di-Cl moieties are incapable of interactions through a conventional hydrogen bond with the α-glucosidase active residues. Moreover, the activity was further decreased by attachment of a substituent of a bulky nature that is unable to interact better with the active part of the amino acid. Therefore, scaffold (16), which is different from analog (20) only in the nature of the substituent present at the 4-position of aryl part 'B', was found to be a low competitor of the α-glucosidase enzyme. This inferior potential of compound (16) was due to the attached bulky -Br moiety at the 4-position of aryl part 'B', even though this analog holds a -OH moiety at aryl part 'C'. The bulky group causes steric hindrance and, hence, lowers the enzymatic inhibitory potentials. Furthermore, the inhibitory potential of scaffold (18) is better than that of analog (16) even though both these scaffolds hold the same substituents (-Br and -OH) around aryl parts 'B' and 'C' but have different positions of the -Br moiety at the aryl part 'B', indicating that the enzymatic activity was also dependent on the position of substituent around both aryl parts 'B' and 'C'. Scaffold (20), bearing a -NO 2 moiety at the 4-position of aryl part 'B' and a -OH moiety also at the 4-position of aryl part 'C', displayed better potency than scaffold (18), which holds a -OH moiety at the 4-position of aryl part 'C' but a -Br moiety at the 3-position of aryl part 'B' (Figure 5).
Overall, it was concluded that the inhibition profile of the benzotriazole-based bis-Schiff base was dependent mainly on the position and nature of substituents on both aryl parts 'B' and 'C'. It was also observed that the alteration of the number/s of substituents around both aryl parts greatly affects the inhibitory potentials of the synthetic analogs.
Molecular Docking Study
The synthetic scaffolds and their measured inhibitions against α-glucosidase enzyme are listed in Table 1. It appeared from the IC 50 values of benzotriazole-based bis-Schiff base scaffolds that α-glucosidase inhibition is related strongly to the position, nature, and number of attached functional moieties of both aryl parts 'B' and 'C' at the benzotriazolebased bis-Schiff base basic skeleton (Scheme 1, Table 1). However, molecular docking was carried out to observe the position, nature, and number of attached substituents and enzymatic inhibition and further to develop the binding interactions of newly synthesized scaffolds with the active residues of the targeted α-glucosidase enzyme. It was observed from a detailed protein-ligand interaction (PLI) study of both potent scaffolds 38 and 10 against α-glucosidase that they established several key interactions with the active residues of the targeted enzyme, which may help in the improvement of the inhibition profile of these potent scaffolds against the targeted α-glucosidase enzyme. It was noteworthy that scaffold 9 adopted numerous interactions with the active amino acid residues of the α-glucosidase enzyme including Lys506 (pi-alkylation), Phe476 (pi-alkylation and pi-pi stacking), Asp232 (conventional hydrogen bond), Ala234 (conventional hydrogen bond and pi-alkylation), Asp568 (pi-anion), Arg552 (conventional hydrogen bond), ILE358 (pi-alkylation), Trp329 (pi-pi T shaped) and Trp432 (pi-pi stacking and pi-sigma) ( Figure 6A).
Similarly, the PLI profile of the second most active scaffold 10 against the α-glucosidase enzyme also revealed numerous important interactions with the active sites of α-glucosidase including residues Ala234 (conventional hydrogen bond and pi-alkylation), Asp232 (conventional hydrogen bond), ILE233 (pi-alkylation), Phe236 (pi-alkylation and pi-pi-stacking), Phe476 (pi-pi stacking), Arg552 (conventional hydrogen bond), Asp568 (pi-anion), Trp329 (pi-pi stacking), Ph2601 (pi-pi T shaped), and Trp432 (pi-sigma) ( Figure 7B). The high potentials of these active scaffolds might be due to attached -NO 2 and di-Cl moieties. The -NO 2 and di-Cl moieties on both sides of aryl parts 'B' and 'C' withdraw most of the electronic density from the aryl rings making them electron-deficient species, which further regain stability by adopting several key interactions with the active residues of the targeted enzyme and hence enhanced the enzymatic potential. Moreover, the attached -OH moiety also enhanced the enzymatic inhibition through participation in H-bonding with one of the active amino acid residues.
General Information
NMR spectra of benzotriazole derivatives that had been synthesized were acquired using a Bruker Ultra Shield FT NMR 600 MHz spectrometer operating at 600 MHz. HREI-MS was used to determine the mass, and fragmentation patterns were determined by analysis carried out using a Finnigan-MAT-311-A instrument. Thin layer chromatography was used to monitor the progress of reactions (Merck, Kieselgel 60F-254, 0.20 mm) and visualized using a UV lamp at 254 nm (UVGL58; Upland, CA, USA).
Formation of S-Substituted Benzotriazole Substrate (III)
In the first step, the pre-synthesized 1,2,3-benzotriazole (I) was treated with various phenacyl bromide (II) in ethanol and triethylamine, and the resulting mixture was refluxed with gentle stirring until the formation of the 2-(1H-benzo[d][1-3]triazol-1-yl)-1phenylethan-1-one substrate (III) was complete (the conversion was monitored by TLC, reflux, 3 h). After completion of the reaction, the residue was stored overnight on cooling for crystallization of the desired substrate (III).
Synthesis of Benzotriazole-Based Bis-Schiff Base Scaffolds (1-20)
In the next step, the carbonyl residue of substrate (III) underwent condensation with hydrated hydrazine in acetic acid, and the resulting mixture was stirred for 3 hrs to achieve the formation of (E)-1-(2-hydrazono-2-phenylethyl)-1H-benzo[d][1-3]triazole (IV), which was treated with various benzaldehydes in acetic acid to afford the synthesis of 1,2,3benzotriazole based bis-Schiff base derivatives (1-20). The synthesized scaffolds were characterized by 1 H NMR, 13 C NMR, and HREI-MS spectroscopy. 1.0 mM (0.2 mg/mL) concentration were prepared in de-ionized water with the enzyme. PIPES buffer was used to adjust the pH of the solution. This solution was incubated for 30 min at 25 • C and the absorbance was recorded on an ELISA reader in 96-well plates. Kinetic parameters like Vmax, AICs, Km, and R2 were calculated using sigma plot enzyme kinetic software [24][25][26][27].
Molecular Docking Protocol
The molecular docking study was performed using discovery studio visualizer (DSV) and autodoc tools 1.5.7. The synthesized compounds were docked against α-glucosidase and their structures were obtained from the protein data bank (PDB) by searching for codes such as 3w37. Initially, the protein was prepared using DSV to maintain the structure by removing water and the structure of the protein, as well as that of the selected analog, was saved in PDB format. The structure of α-glucosidase was made open in autodoc tools by adding polar hydrogen to the protein as well as Kollman and Gasteiger charges. Ligand preparation was done by using a torsion tree to detect the root and it was saved in PDBQT format. A configuration file was generated along with the X, Y, and Z axes, and the protein structures were saved in PDBQT. In order to generate different poses of molecules with varied energies, a command prompt was employed. The top-ranked molecules were docked with protein PDBQT format in DSV [28].
Conclusions
Benzotriazole-based bis-Schiff base scaffolds (1-20) were synthesized and then investigated in vitro for α-glucosidase inhibitory potentials. All the synthetic analogs based on benzotriazole-based bis-Schiff base scaffolds were found to display outstanding inhibition profiles on screening against the α-glucosidase enzyme. The synthesized scaffolds showed a range of inhibition profiles, having IC 50 values ranging from 1.10 ± 0.05 µM to 28.30 ± 0.60 µM when compared to acarbose as a standard drug (IC 50 = 10.30 ± 0.20 µM). Among the series, fifteen scaffolds 1-3, 5, 6, 9-16, 18-20 were identified to be more potent than standard acarbose, while the five remaining scaffolds 4, 7, 8, 16, and 17 also showed potency against the α-glucosidase enzyme but were found to be less potent than standard acarbose. The structures of all the newly synthesized scaffolds were confirmed using different spectroscopic techniques such as HREI-MS and 1 H-and 13 C-NMR spectroscopy. To find a structure-activity relationship, molecular docking studies were carried out to understand the binding mode of the active inhibitors with the active sites of the enzyme and the results supported the experimental data. | 4,690 | 2022-12-22T00:00:00.000 | [
"Chemistry"
] |
Comparisons of Synchronous and Asynchronous Discussions in an Online Roleplaying Simulation to Teach Middle School Written Argumentation Skills
In this study, different degrees of synchronous and asynchronous online social interactions are investigated in the context of an online educational roleplaying simulation game that is played across multiple classrooms simultaneously to teach argumentation skills and social studies. Results from 45 K–12 middle school social studies teachers and 867 students over 3 study conditions were compared based on the degree of real-time discussion that was embedded in each condition’s version of game (i.e., two scheduled live conferences, one scheduled live conference, and asynchronous-only interactions or zero live conferences). All conditions exhibited significant small to moderate-level pre-post effect sizes, including the condition featuring asynchronous-only discussions. Additionally, the “mid-range” 1 live conference condition exhibited the greatest pre-post effect size in comparison to the other two conditions. This study demonstrates evidence for the benefits of implementing asynchronous-only discussions in digital interventions in comparison to live discussions when synchronous interaction may not be feasible. For designers, implementing both asynchronous and synchronous interactions based on available resources and feasibility can be used to maximize social presence among participants in educational roleplaying games and other virtual learning environments.
Riel, J., Lawless, K. A., & Oren, J. B., (2020).Comparisons of synchronous and asynchronous discussions in an online roleplaying simulation to teach middle school written argumentation skills.Online Learning, 26(4), 146-167. DOI: 10.24059/olj.v26i4.3468For over 20 years, a central policy initiative for K-12 education has been the effort to promote student skills and interest within the STEM disciplines (Committee on STEM Education, 2018;NRC, 2014).Researchers and policymakers have repeatedly issued warnings of a great shortage of workers to meet STEM career openings and that working within the modern knowledge economy requires development in strong scientific and technological literacy skills that should begin as early as the elementary and middle grades (English, 2017;NRC, 2011NRC, , 2022;;van den Hurk, Meelissen, & van Langen, 2018).To meet this need, governments, researchers, and policymakers worldwide have continually advocated for more STEM education offerings to engage students with socio-scientific content (Newcombe et al., 2009;Scogin et al., 2017).Specifically, these groups have called for teaching students not just the content of STEM disciplines, but also to develop essential cognitive skills for using content, such as critical thinking, problem-solving, and argumentation (Van Laar et al., 2017).Such skills are frequently cited as necessary for success in the STEM and knowledge-economy workforce where digital information is now ubiquitous, of varying quality, and from multiple perspectives (Noroozi, Dehghanzadeh, & Talee, 2020).
Among this call for critical STEM skills training within schools is the mastery of argumentation and the skills for evaluating and generating arguments to succeed in navigating the deluge of information that is encountered in everyday life (NRC, 2014).To this end, argumentation is often cited as an essential life skill for success during this age of information ubiquity (Bathgate et al., 2015;Kuhn, Hemberger, & Khait, 2016a;Özdem Yilmaz, Cakiroglu, Ertepinar, & Erduran, 2017).Additionally, it has been argued that the teaching of argumentation skills provides opportunities for robust learning experiences in any discipline and for any career, as argumentation establishes relevant active learning contexts for teaching subject content instead of teaching through rote memorization of facts and conceptual definitions, particularly in social studies (Cavagnetto, 2010;Iordanou, Kuhn, Matos, Shi, & Hemberger, 2019).
Research on the differences between asynchronous and synchronous social interactions is particularly important for providing insights toward the design of learning environments.This is especially the case in which the learning objectives are skills that are best developed in social situations like argumentation training, as it takes at least two people to hold an argument.Although asynchronous activities have always existed in K-12 through homework assignments, or, more recently, through out-of-class communications with teachers via media applications, the effects of asynchronous-only interactions in educational interventions that are deployed in K-12 schools are only recently becoming more regularly studied (Loncar, Barrett, & Liu, 2014;Lowenthal & Dunlap, 2020).
To contribute toward this literature, this study examined GlobalEd, an online educational roleplaying simulation game designed for middle school social studies classrooms.Originally designed to have both synchronous and asynchronous components for play among students across multiple classrooms, a recent edition of the game featured and investigated the effects of exclusively asynchronous-only discussions without any synchronous component.For this study, we evaluated whether an asynchronous-only condition was beneficial to students in comparison to versions of the game with synchronous discussions.Specifically, we experimentally investigated how two different live-discussion conditions compared to an asynchronous-only condition in terms of observed effects on students' argumentation skills.As argumentation is best learned in a social space that allows for regular dialogue between participants, the efficacy of an asynchronous-only design could dramatically increase the flexibility and design potential for social learning interventions.
Background Argumentation as a Cross-disciplinary, Socially Learned Skill and Mechanism for Learning Disciplinary Content
Of the many skills that are necessary for scientists to be successful, mastery of argumentation and scientific reasoning are often cited as priorities for STEM instruction (Kuhn, Hemberger, & Khait, 2016b;McNeill, Lizotte, Krajcik, & Marx, 2006;Sandoval, Enyedy, Redman, & Xiao, 2019).Argumentation, as it is frequently used in the STEM disciplines, is more than just having disagreements with people (Andriessen, Baker, & Suthers, 2003).As the research on scientific argumentation and STEM career skills has grown over the last three decades, argumentation skills and the ability to critically analyze arguments have increasingly been cited as required critical skills within large-scale educational reforms and standards for socio-scientific literacy and competency within STEM disciplines, such as the Next Generation Science Standards (NGSS Lead States, 2013), the Common Core State Standards (CCSS, 2010), and the National Curriculum Standards for Social Studies (NCSS, 2010).
Indeed, the practical aspect of scientific communication of findings and persuasion through argumentation achieves a core function of the scientific process.However, additional benefits can also emerge when students are engaged with argumentation.Participants not only persuade others of their explanations, but they also engage in a collaborative and social process of understanding the content being argued (Coffin, Hewings, and North, 2012).Importantly, engaging with argumentation encourages students to confront, analyze, and refine their own understandings as well, such as that which has been demonstrated in the growing body of research that adopts the approach of Arguing to Learn (Andriessen, Baker, & Suthers, 2003;Bathgate et al., 2015).Within this approach, although students are simultaneously developing their argumentation skills, they have also been observed to develop critical thinking skills, writing skills, and the ability to learn content knowledge across domains as a direct result of engaging with argumentation processes (Kuhn, Hemberger, & Khait, 2016a;Suephatthima & Faikhamta, 2018).Additionally, because information is more readily available for retrieval at a moment's notice in today's digital landscape, it has even been suggested that the ability to interpret and analyze facts and concepts is perhaps more important than simply knowing these facts (Van Laar et al., 2017), a role for which argumentation training is well poised to support.
When learning skills like argumentation that are inherently grounded in social interaction and require the consideration of multiple perspectives, repeated practice within authentic social contexts is often seen as a necessary condition for learning such skills (Crowell & Kuhn, 2014;Iordanou et al., 2019).Otherwise, as argumentation is fundamentally a process that occurs between two or more people, any attempts at learning these skills without discussion or collaboration deprives learners of experiencing the authentic, situated contexts in which the skills are used (Noroozi et al., 2012).For instance, simply learning facts about argumentation or its structure does not sufficiently prepare students for engaging with actual argumentative tasks, as it lacks the opportunity to experience the transactive back-and-forth dialogue that underlies the process (Mercier, Boudry, Paglieri, & Trouche, 2016).Therefore, argumentation instruction is necessarily situated in social interaction: the practice of making and analyzing arguments always occurs between at least two people (Mercier et al., 2016;Scardamalia & Beriter, 2006).As a result, a consensus among argumentation scholars is that these skills are necessarily taught in socially rich environments in which participants regularly engage in dialogue with each other and conduct argument analysis, construction, and feedback in a back-and-forth, transactive way (Henderson et al., 2018).
To this end, social processes such as argumentation require learning environments that enable social interaction to fully learn how to perform the skill.Especially in the post-pandemic educational environment, it has become increasingly important for researchers and instructional designers to create learning environments that can leverage the unique opportunities provided by digital technologies to enable authentic discussions and other social interactions, albeit at a distance.When people cannot be physically present together, synchronous and asynchronous online social discussions can be employed to provide spaces for socially intensive learning activities (Mercier et al., 2016;Noroozi et al., 2012).
Furthermore, online interactive approaches might afford unique conditions, opportunities, and motivations for learners that are not otherwise present in face-to-face learning contexts.In recent reviews, highly social online learning environments for teaching social skills such as argumentation have shown promising results; however, there has been virtually no research performed on the modality differences between face-to-face and various online, computermediated social interactive modalities for teaching argumentation (Asterhan & Schwarz, 2016;Lowenthal & Dunlap, 2020).The unique technological affordances for online socialization, including synchronous and asynchronous online discussions, should thus be further researched to maximize the potential for online learning in both K-12 and higher education (Henderson et al., 2018;Nussbaum, 2021).
Considering Simultaneity of Social Interaction and Social Presence for Online Learning Designs
The timing by which someone interacts in an online space may matter just as much as whether it is socially interactive in the first place.Knowing not just whether someone is expected to interact in a learning space, but also when someone is expected to interact are both primary components of the degree of "social presence" within an online Community of Learning (Garrison, 2016).The construct of social presence within a Community of Learning framework argues for the required presence of rich social interactions among learners in online learning environments.Opportunities for social interaction can activate the interpersonal and transactive processes that are essential for learning and meaning-making processes, such as discussing and determining the meaning of phenomena and concepts, debating concepts, and encountering other points of view to refine one's own understanding (Kozan & Richardson, 2014).Toward this focus on social presence, it has been regularly observed that the expectation of the degree and timing of which participants will interact will often influence variations in the type of behaviors that are exhibited in learning environments (Chen, Park, & Hand, 2016;Coffin, Hewings, & North, 2012;Koehler et al., 2020).
Varied expectations by the learner of the timing and simultaneity of responsiveness from peers in the social setting may determine the types of responses, depth of thinking, and included content associated with a given learner's participation (Cui, Lockee, & Meng, 2012;Foo & Quek, 2019;Larrain, Freire, Lopez, & Grau, 2019;Peterson, Beymer, & Putnam, 2018).Additionally, technology-based supports and scaffolding may be more readily implemented in asynchronous online activities than those requiring more real-time adaptations and assistance (Jeong & Joung, 2007;Jeong & Fraiser, 2008;Lin, Hong, & Lawrenz, 2012).Furthermore, although the inclusion of real-time interactions might create a more immersive and engaging environment that requires the participant to be cognitively attentive, such real-time expectations could demand more of the learner's attention, as well as be taxing on teachers who face various classroom and scheduling constraints when implementing live, synchronous interventions (Cui, Lockee, & Meng, 2012;Nieuwoudt, 2020).
It has become increasingly necessary given the post-pandemic educational landscape to investigate the effects and mechanisms connected to different levels of social presence within online learning environments that rely on social interactions.Although live social interactions in an online intervention have regularly been assumed to yield better results, such interactions may not always be feasible for a teacher to implement.This is especially true in situations where students may be having discussions or otherwise collaborating with people outside of a physical classroom.Various classroom constraints are typically present and teachers often need flexible options, or at least options for students to engage with environments outside of their scheduled classroom time or in a virtual manner.
Online Educational Simulation Games (ESGs) and Roleplaying: Enabling Flexible Implementation of both Synchronous and Asynchronous Discussions
The use of educational simulation games (ESGs) and interactive roleplaying is one approach that is well-suited to provide rich contexts for social interactions and exposure to social studies concepts in an authentic way (Devlin-Scherer & Sardone, 2010;Liu, Cheng, & Huang, 2011).The use of simulations as educational interventions is certainly not new, but advances in digital technologies over the last two decades have enabled the virtualization of both physical and social processes in ways never possible before.ESGs and roleplaying games that specifically model social processes (Gredler, 2013) can allow players to interact with social forces and assume the role of actors within the system through authentic roleplaying.In such games, players are assigned roles with specific goals within a simulated social event or system that models realworld social phenomena (Sauve et al., 2007).When a social simulation is additionally integrated with game mechanics, players, as agents in the game, gain clear goals on how to win the game, a set of rules for interactions and allowed player "moves" in the game, and feedback mechanisms (e.g., points, penalties) to guide their play and improve motivation (Brom, Stárková, Bromová, & Děchtěrenko, 2019;Vlachopoulos & Makri, 2017).Thus, authentic roleplaying in this manner allows for deep and authentic investigation of the forces and concepts under study within the game and to foster opportunities for social interaction to grapple with skills that are socially learned, like argumentation (Squazzoni et al., 2014).
Although modern ESGs and roleplaying games that model social processes can be played both in-person and online, online games are particularly timely for social studies education in today's post-pandemic world due to their ability to provide uninterrupted continuation of gameplay both inside and outside of the classroom.As seen from the widespread school closures as a result of the global COVID-19 pandemic, effective online interventions that facilitate ongoing interactions among students and teachers can be valuable in the situation of school closures or student absences from school.As they are educational interventions that can enable motivating synchronous and asynchronous modes of social interactivity, ESGs are well-poised to permit continuous dialogue and collaboration among students in their own class based on the teacher's pedagogical needs.
The Present Study: Observing Effects of Variations in Simultaneity in the GlobalEd Game
Studies have been performed recently between the varying degrees of simultaneity in online social interactions in K-12 learning environments, generally showing that both synchronous and asynchronous interactions, such as online written discussions, among participants have shown benefits based on the intended learning goals for which they were implemented (Gašević et al., 2015;Lowenthal, Dunlap, & Snelson, 2017;Yamagata-Lynch, 2014).Fewer studies, however, have been performed comparing the varying types, levels, and benefits of asynchronous-only and live, real-time discussions specifically in the context of online roleplaying and social simulations and how they can foster student achievement.
This study reports an experiment on multiple designs of GlobalEd, an online educational roleplaying simulation game for middle school social studies classrooms.GlobalEd simulates a social process of a complex international crisis in which students play the roles of different countries that come together to research and develop proposals to solve a given real-world problem scenario (Lawless et al., 2018;Riel & Lawless, 2022).Through gameplay, social interactions like discussion are a fundamental principle to the design of GlobalEd as a pedagogical approach for developing students' argumentation skills (Mercier, Boudry, Paglieri, & Trouche, 2016;Scardamalia & Beriter, 2006).
Specifically, because previous iterations of the GlobalEd game over its ten-year history had always included a synchronous discussion opportunity to online players, we were particularly interested if the game could be played in an asynchronous-only way and still generate an observable effect on the argumentation skills learning outcome.We wanted to investigate if increasing levels of simultaneity or synchronous play had a positively trending effect in comparison to asynchronous play.This would help test an assumption of whether including the most or highest-level live discussion is the best option in online and socially intensive learning interventions, such as social simulations or roleplaying games.
The following two research questions guided this study to respond to the need for additional research on comparing the differences in the effects on learning outcomes between synchronous and asynchronous discussions in online simulations and games that prioritize social interaction for learning: RQ1: Does an asynchronous-only version of the GlobalEd intervention demonstrate either comparable or higher effects in written argumentation skills (i.e., the primary learning objective of GlobalEd) than two other versions of GlobalEd that emphasize synchronous discussions among players?RQ2: Do increased levels of synchronous discussions in GlobalEd demonstrate progressively higher effects in written argumentation skills (i.e., the primary learning objective of GlobalEd).
Context for the Study-Description of the Intervention The GlobalEd Online Roleplaying Simulation
The intervention in this study is an online roleplaying simulation called GlobalEd.GlobalEd is designed for play across multiple social studies classrooms simultaneously to simulate complex international social interactions and systems in an authentic way (Lawless et al., 2018;Riel & Lawless, 2022).This allows for players to discover and apply real-world knowledge related to socio-scientific issues that do not often have a "correct answer" solution.Such ill-defined challenges mirror the authentic problems that scientists, technologists, diplomatic professionals, and policymakers face with solving authentic global issues.
In the game, students play the roles of scientific advisors to an assigned country.Each country that is roleplayed by students in the game is invited to an international summit (represented by synchronous or asynchronous discussions) to solve an assigned problem scenario.Up to 20 countries (i.e., different classrooms) play in a single GlobalEd game.
Interactive Discussions within GlobalEd
Play of GlobalEd progresses over three phases during a multi-week period: an initial research phase, an interactive discussion phase, and a summary debriefing phase.The primary goal of play is for each team to develop a single final proposal that has been co-sponsored by at least two other country teams (i.e., other classrooms).When the final proposals are submitted, they are voted upon by all teams, with the winner of the game being the one who has received the most votes.The essential feature of GlobalEd is the dialogue that is generated by students during both asynchronous messaging and live synchronous conferences across teams.In the first type of dialogue, players solve the assigned problem scenario via live, real-time conferences between classroom teams in collaboration on solutions to the problem scenario.The live conferences take place within a synchronous, instant-messaging-like online communications system where all players meet at a scheduled time.Before each live conference, students are provided with an agenda of the topics that will be discussed, which allows the students to prepare their ideas, solutions, and evidence to submit to the other teams for consideration.All student dialogue is moderated by a trained coordinator for both appropriate content, for prompting students to maintain their assigned roles in the game, and for coaching students in the use of argumentation skills.An example screenshot from a live conference is provided in Figure 1.
Figure 1 Screenshot from conference
In the second form of dialogue, students also interact with each other via asynchronous messaging (i.e., email-like messages) throughout the entire duration of the game.In asynchronous messages, players negotiate their positions and perform collaborative research over the full duration of the interactive phase.The asynchronous messaging is performed in an email-like interface with which students can log on at any time, including outside-of-classroom time or at home.An example asynchronous message and reply between two country teams from the actual game environment is illustrated in Figure 2.
Figure 2 Screenshot of asynchronous messaging between teams in actual GlobalEd play
Through the asynchronous messages, players continue the conversation and to debate issues with teams as they work toward developing well-argued proposals that will gain cosponsorships and alliances with other teams.Both the asynchronous and synchronous messaging discussions in the simulated international summit are facilitated in an online communications platform that moderates all communications between players, hosts scheduled events, and promotes interaction among players.Within both types of discussions, players regularly are encouraged to challenge each other to strengthen their arguments, to provide more evidence about their claims, or to provide additional context for the solutions that they are proposing.
GlobalEd has been in continual development and iteration over the last 10 years and has repeatedly demonstrated high levels of efficacy in development of student argumentation skills, content knowledge, and interest and self-efficacy in social studies and science topics and careers (Lawless et al., 2018(Lawless et al., , 2019;;Yukhymenko, 2011).However, live synchronous discussions have been the highlight for each iteration of the game for the past ten years.For this study, we attempted a game version that only used asynchronous communications for player discussion, with no live synchronous discussions.Additionally, we also wanted to identify if more live discussion opportunities had a stronger effect than the asynchronous-only alternative.
Participants
In the present study, 45 middle school social studies teachers in the United States participated, along with the students (n = 867) in each of their classrooms.Teachers each played a version of the GlobalEd game with their students based on the condition to which they were assigned.The simulation's program, content, and structure among conditions were identical except for the number of scheduled real-time, live conferences in which students would participate.Table 1 provides a breakdown of participants (students and teachers) by condition.
Teachers from different schools in both suburban and urban classrooms were randomly divided into one of three study conditions, which represent the level of live, real-time synchronous discussions (i.e., live conferences) that their assigned simulation would have: two scheduled live conferences (n = 17 teachers, 341 students), one scheduled live conference (n = 13 teachers, 263 students), and no scheduled live conferences or asynchronous-only discussions (n = 15 teachers, 260 students).Table 1 provides a breakdown on participant totals by condition.
Data and Instruments
Students were presented with identical pre-and post-intervention essay assignments to demonstrate their skill with written argumentation and to exercise their knowledge of the social studies concepts they encountered.In this assignment, students were presented with a prompt related to the simulation that they were tasked with writing about.The text used in the essay assignment for both the pre-and post-instruments is featured in Figure 5.
Figure 5 Pre-and post-essay assignment
The assigned problem scenario for all students in each of the three conditions was a global water scarcity dilemma to solve collaboratively with other teams, so it was expected that students would improve in the post assessment in both content knowledge of social studies as well as their written argumentation skills in response to the assessment prompt.We intentionally used instruments that captured students' writing as they made and defended a claim, as the instrument specifically prompted students to demonstrate their skill in complex thinking and argumentation.Thus, direct evidence of students' written argumentation skills and content knowledge were captured with a high degree of resolution for identifying the connections between the content knowledge and use of argumentation (Albanese, 2000;Savin-Baden, 2004).
The research team developed a rubric before implementation to analyze the pre-and postessay writing instruments.This rubric measured the level of argumentation skills on multiple parameters, including the presence and quality of students' use of claim, evidence, reasoning, and addressing the opposition, as well as to capture evidence of the use of social studies concepts that students encountered during the game.The rubric scored essays on seven items related to argumentation skills, with the post-coding values for each item being combined into a single summative scale value for each the pre-and post-essay.
Each essay was scored by three graduate-level students who were trained on the rubric and had 100% interrater agreement on a test set of essays after conference.After completing the test set, each coder graded each essay, pre and post.Because the instruments were identical, the pre and post versions of the essays were blinded to the coders as to reveal whether it was a pre or post during scoring.Although each of the three coders coded each essay, for data imputation purposes each essay was randomly assigned by computer to two of the coders.Each item was analyzed for alignment by computer between the coders.Any disagreements within 1 point between the two coders on the spreadsheet were resolved by adding the third coder's score and
ESSAY WRITING ASSIGNMENT Prompt
The world is in danger of running out of fresh water.Do you think this is true?Do you agree or disagree with this statement?Why?
Assignment
Write a persuasive essay stating your point of view on the prompt above.Give evidence to support your answer and provide your reasoning why this evidence supports your claim.Use your knowledge about water, science, world geography and cultures to help you write your response.You will have a total of 30 minutes to complete your essay.
taking the mean, averaging to the nearest half-point.No additional coding disagreements emerged after a third coder was introduced.Coding reliability between raters was > 0.80.Table 2 presents the scoring parameters in the rubric that were used for coding the identical pre-and post-essays.
Data Analysis
We conducted a hierarchical linear modeling (HLM) analysis (mixed) with the pre-and post-essay writing scores to compare the three conditions of the study and account for pre-test skills exhibited by students, as well as any classroom-or teacher-level effects that might be observed.HLM is a type of mixed-level multiple regression analysis that accounts for multiple "nested" levels of data and potential effects on the dependent variable that could occur at the different levels.HLM uses maximum-likelihood estimation to estimate the coefficients for each fixed effect that is entered into the model as the model predicts the output dependent variable.
HLM is increasingly used in educational research due to its robustness to detect classroom-or teacher-level effects among student achievement and other outcome variables (Raudenbush & Bryk, 2002).HLM is well-suited for education research as its models account for the moderating effects of teachers or even schools that are within different hierarchical levels (i.e., students within classrooms within schools).Furthermore, like ordinary multiple regression, HLM can account for other independent mediating or moderating factors within the analysis as fixed effects or random effects.
We employed the HLM 7 software suite (Raudenbush, Bryk, Cheong, Congdon, & du Toit, 2011) to conduct the analysis.Due to the naturally stratified nature of educational research data originating from multiple authentic classroom sites, student participants (at level 1-L1) were nested in the HLM model within teacher classrooms (at level 2-L2).In this multilevel analysis, a nested structure allows for the researchers to account for any possible teacher effects via inclusion of the pretest of students' writing performance at L2 centered around the grand mean to account for students' skill level at the outset of the intervention and their growth over time (Raudenbush & Bryk, 2002).A third nested level (L3) that represents the schools in which classrooms are nested was not necessary in this analysis, as there were no school-level effects to observe with multiple classrooms within single schools participating in the study.Different schools participated in the analysis.
The three experimental conditions were each coded as binary variables (0/1) that represented whether a student participated a given condition.In the model, the conditions of "2 live conferences" and "1 live conference" were entered as fixed effects in the conditional model.The binary coding scheme for each condition's variable assigned a value of 1 if a student was a part of the condition, or 0 if not.Thus, if a student was in the 1 live conference condition, the variable would be value = 1, otherwise it would be 0. The condition of "0 live conferencesasynchronous only" represented the baseline comparison for the model and was therefore not entered as a fixed effects term.The 0-conference condition is instead represented in the model's intercepts (i.e., when the "1 live conference" and "2 live conference" conditions are both value = 0).These comparison conditions were entered at L2 to represent each classroom's experimental condition to which they were randomly assigned.
Additionally, students' pre-scores on the essay instrument were entered as an L1 fixed effect that was centered around the group mean at L1 to account for students' prior knowledge and skills with the instrument and to identify the degree of pre and post student gains.Groupmean centering at this level is appropriate due to the potential classroom-level effects that might be observed within each classroom group.Furthermore, teacher-or classroom-level effects were also accounted for in the model, which was represented by students' pre-test scores centered around the grand mean at L2 to consider pre-scores between groups.
The results from the HLM analyses were then used to determine the effect size of each condition.The HLM equation for this study is provided in Equation 1.
Equation 1 Expanded 2-Level Equation for Hierarchical Linear Model Analysis
Post-achievement (Y) = g00 + g01*1Conf + g02*2Conf + g03*TC_achievement + g10 *SC_achievement + u0 + u1 + r In the model, Y represents the dependent variable for student achievement, as measured by student written argumentation scores on the post-essay instrument.The fixed effects terms for the experimental conditions are 2Conf (2 live conferences) and 1Conf (1 live conference), which were binary terms that indicated participation in the particular condition or not.The 0 live conference condition is represented in the model as the baseline measure through the intercept g00 when both 2Conf and 1Conf are value = 0. TC_achievement represents the level-2 teacher-centered grand-mean value for the pre-essay instrument to account for teacher-level classroom effects, SC_achievement represents the student-centered group-mean value for the pre-essay instrument, and u0 , u1 , and r collectively are random effects terms in the model.
Results
Table 3 displays the descriptive statistics on essay writing scores (as a summative scale score of the seven items on the essay rubric) for all conditions.Table 4 displays the results of the HLM analysis.The fixed effects of 1-conference and 2conference are in comparison to the 0-conference condition, which is represented as the baseline in the model.Comparatively, the 1-conference condition yielded higher positive results in comparison to the 0-conference condition, as indicated by a positive coefficient estimate.Because of its negative coefficient, the 2-conference condition fixed effect demonstrated that the 0-conference asynchronous condition outperformed the 2-live conference condition.It is important to take care with interpreting the 1-to-0 conference comparison (i.e., the 1conference term), as it was observed at p = .062and thus the observed differences may be due to chance.Although the comparison between 0 conference (asynchronous) and 1 conference closely approached significance at the p < .05threshold commonly accepted in social science research, there could also be no difference between the two, or instead interpreted as roughly equal groups.
Additionally, through the inclusion of the pre-writing assessment at both L1 (student) and L2 (teacher), the model also accounts for students' skills prior to starting the intervention.A significant L2 teacher-level pre-writing assessment term indicates that there were classroom-level effects observed and that students performed differently between collective classrooms.The HLM model accounts for these potential effects in calculating the overall estimates of the coefficients and their relationships to the dependent variable of written argumentation achievement.
Table 5 further interprets differences between the comparison conditions by providing pre-post effect sizes for each condition (reported as Cohen's d) to compare which condition had the highest pre-post effects across the study.For each condition, pre-post effect size was calculated as the difference between the means between the pre-and the post-tests divided by the pooled standard deviation of the condition.The comparison of pre-post effect sizes, otherwise known as a standardized difference of means, is appropriate in situations where identical instrumentation is used in educational pre-post assessment and effect sizes are thus interpretable in a standardized, comparable way (Morris, 2008).Each of the three conditions were confirmed to have been effective as intended, as each condition demonstrated significant positive mean differences favoring the post-test within confirmatory paired-samples t-tests (p < .001for all).This indicated that within each condition, the students performed better in the post-than the preassessment, Subsequently, this can be interpreted as having demonstrated learning and growth (or, alternatively, that the intervention achieved its learning objective goals).Note.Pre-post differences in means for each condition were confirmed by paired-samples t-tests, all of which were observed to be p < .001.Effect sizes reported as Cohen's d.
In Table 5, the 1-Conference condition was observed to yield superior pre-post student achievement effects in comparison to the other two conditions.The 2-Conferences and No-Conference also demonstrated effects in the HLM model and were confirmed by paired-samples t-tests, but to a lesser degree than the 1-Conference condition.These results indicate evidence for the efficacy of the intervention regardless of condition.In a conventional interpretation effect size, each condition can be seen as having a small to moderate effect (0.3-0.5) on student achievement.Indeed, the 1-conference condition yielded the highest effect, but the 2-conference and asynchronous-only 0-conference conditions both also yielded effects that trend toward moderate levels.
Because the difference between 0 and 1 live conference was not observed to be significant at the p < .05threshold generally accepted by the education field, these two effects are relatively the same.Although the difference was not significant in the HLM model, this study does suggest that some degree of combined live discussion and asynchronous-only discussion might provide a boost to student learning outcomes in comparison to asynchronous-only discussion, especially when the learning outcomes are highly social in nature (such as from learning argumentation skills).
Also of note is the significant negative difference between the 0-conference condition and the 2-conference condition in the HLM model, providing evidence that higher levels of live discussions may not always be the best option in virtual learning environments in comparison to providing asynchronous-only discussions.This observation is corroborated by observing a lower effect size between the 2-conference and 0-conference conditions, with 0-conference demonstrating a higher effect size.
Conclusion
Each condition in the study yielded a moderate effect size, providing evidence for flexibility in how designers develop socially intensive online spaces and for teachers in the degree to which they choose to engage with online social activity for their students synchronously.For this study, it was useful to identify evidence for designers that when course time is limited, an asynchronous-only condition can still be feasible and yielded a moderate effect in the achievement of learning outcomes.In many cases in the post-pandemic landscape, virtual asynchronous social interactions may be a teacher's best or only option.In this study, the asynchronous-only condition of the GlobalEd intervention was demonstrated to be effective.
More study and theorization on this concept are certainly needed to understand how and why the higher degree of live discussion was observed to have a lesser effect than the mid-range live-discussion condition and the asynchronous-only condition.In terms of social presence, live interactions are thought of to be a "richer" learning experience but may not always be necessary to indicate the presence of other individuals and groups (Chen, Park, & Hand, 2016;Garrison, 2016;Koehler et al., 2020).In today's digital ecosystem, a high degree of live discussions may serve to be distracting for some individuals or demand a high level of cognitive load, which may actually counter the benefits of the learning activity.Live interactions, particularly over time, might be mentally taxing to some learners but invigorating to others (Cui, Lockee, & Meng, 2012;Nieuwoudt, 2020).
Additionally, in virtual discussion, social presence also is dictated by the level of expectation of a person's behavior in the learning experience, as well as how the learning environment facilitates both asynchronous and synchronous discussion (Chen, Park, & Hand, 2016;Coffin, Hewings, & North, 2012).As such, the expectations of learners' social presence when interacting in a virtual space may be different than the expectations of the instructional designers and game developers who design activities and interactions for play (Cui, Lockee, & Meng, 2012;Larrain et al., 2019).
If real-time interaction and synchronous social presence are deemed the most desirable in online and hybrid learning environments, further study should be pursued in virtual learning contexts to investigate if and why students might perform better with only some but not the highest number of real-time interactions possible.
However, with the evidence from this study, it is heartening for instructional designers and teachers alike that any level of social interaction chosen still elicited the desired learning outcomes.Additional studies on the level of simultaneity of effective virtual interventions should be conducted to investigate whether asynchronous-only, mixed, or high-synchronous discussions all work effectively at achieving learning objectives, as to give educators increased choice in the implementation of virtual learning products with varying levels of required social presence.This is particularly important in the post-pandemic landscape where teachers may need to move rapidly from a synchronous learning context to an asynchronous-only context.Research on the efficacy of innovations tested with varying levels of simultaneity will help decision makers with selecting robust curricular materials.
This study is limited in scope related to asynchronous and synchronous learning conditions as it investigated just one single roleplaying game, one context in which discussions occurred by students, and one set of learning objectives.Additionally, the intervention is a simulation roleplaying game and not another type of online learning activity, preventing too broad of claims about simultaneity of discussion.Despite these classic limitations that are common in educational research, what has been demonstrated is that there was value to the asynchronous-only version of play as it yielded a beneficial effect.Additionally, the most live discussions were not found to be the condition to have the highest impact.Primary research like this study that richly describes the intervention design and evaluates the effectiveness of single intervention designs are necessary for teachers, policymakers, and instructional designers to make sound decisions on development and implementation of interventions.
In our reflection as instructional designers and researchers of the GlobalEd project after over ten years of implementation of the GlobalEd game in hundreds of classrooms, one of teachers' biggest hurdles was the scheduling of live discussions during constrained curricular time.Within the classroom, teachers have only limited time to get students to interact together, especially if working in small groups.Additionally, GlobalEd players are afforded the opportunity to interact across classrooms through extended play.Thus, the GlobalEd roleplaying game enables two layers of discussions, both of which are enabled through asynchronous interactions that can be performed outside of class through homework, small group work, or even remote learning at home.The results of a substantial effect size for the asynchronous-only condition confirmed for us the value in providing teachers flexibility in the play and implementation of GlobalEd.When designed in a principled way, asynchronous discussions can still promote social presence among participants, including those in the K-12 age range.However, this study also highlights the importance of evaluating whether designs work as intended and if learning objectives are met, otherwise designers risk the intervention yielding no effect and possibly a disappointing social experience for participants.
In the post-pandemic educational landscape where shifts to virtual learning can happen in an instant, online learning activities such as games and simulations that model social processes can continue to foster inquiry and development of key social studies skills without any interruption.Online games and simulations can be played in face-to-face classrooms, when possible, but also can allow for the virtual game platform to facilitate and organize high-impact play discussion regardless of whether the game is played in the classroom or online, or whether it is played synchronously or asynchronously.
Table 1
Number of Participants by Condition
Table 2
Essay Grading Scoring Parameters for the Identical Pre-and Post-Writing Assignments
Table 3
Descriptive Statistics
Table 4
HLM Analysis Results: Model Statistics
Table 5
Pre-Post Effect Size Results for Synchronous and Asynchronous Interaction Conditions | 9,136.4 | 2022-12-01T00:00:00.000 | [
"Education",
"Computer Science"
] |
Atmosphere-Induced Transient Structural Transformations of Pd–Cu and Pt–Cu Alloy Nanocrystals
We have investigated the transformations of colloidal Pd–Cu and Pt–Cu bimetallic alloy nanocrystals (NCs) supported on γ-Al2O3 when exposed to a sequence of oxidizing and then reducing atmospheres, in both cases at high temperature (350 °C). A combination of in situ diffuse reflectance infrared Fourier transform spectroscopy and X-ray absorption spectroscopy was employed to probe the NC surface chemistry and structural/compositional variations in response to the different test conditions. Depending on the type of noble metal in the bimetallic NCs (whether Pd or Pt), different outcomes were observed. The oxidizing treatment on Pd–Cu NCs led to the formation of a PdCuO mixed oxide and PdO along with a minor fraction of CuOx species on the support. The same treatment on Pt–Cu NCs caused a complete dealloying between Pt and Cu, forming separate Pt NCs with a minor fraction of PtO NCs and CuOx species, the latter finely dispersed on the support. The reducing treatment that followed the oxidizing treatment largely restored the Pd–Cu alloy NCs, although with a residual fraction of CuOx species remaining. Similarly, Pt–Cu NCs were partially restored but with a large fraction of CuOx species still located on the support. Our results indicate that the noble metal present in the bimetallic Cu-based alloy NCs has a strong influence on the dealloying/migrations/realloying processes occurring under typical heterogeneous catalytic reactions, elucidating the structural/compositional variations of these NCs depending on the atmospheres to which they are exposed.
■ INTRODUCTION
Heterogeneous catalytic processes involving bimetallic alloy nanocrystals (NCs), including oxidation and reduction reactions, environmental catalysis, electrocatalysis, biomass conversion, and energy storage, have been the subject of several studies. 1−3 Consequently, significant progress has been made in the synthesis of well-defined alloy NCs by tuning their size, shape, and composition. 4,5 However, bimetallic alloy NCs can undergo extensive structural transformation, surface segregation and change in their chemical state, structure, and reactivity, upon exposure to reactants and high temperatures. 6−13 In analogy with monometallic NCs, bimetallic NCs can also suffer from sintering and can be redispersed in certain conditions and depending on gas atmospheres. 14 Ostwald ripening and particle disintegration induced by reactants are processes that imply different interactions between adsorbate-metal, metal−metal, and metal−support. 15 Major research efforts have focused on the investigation of complex dynamic processes occurring in bimetallic-based NCs under reaction conditions. 16,17 Efforts have been made to alloy noble metals (e.g., Au, Pd, Pt) with nonprecious transition metals, such as Cu, also with the aim to alter the electronic properties of the noble metal and to attain cost-affordable, stable NCs with enhanced catalytic performance. 18 For example, Shan et al. 8 studied how the surface structure affected the dispersion of metal sites on the surface of shapecontrolled Pt−Cu NCs. Their study applied different treatment conditions to vary the NC surface composition and structure, which correspondingly changed their catalytic performance in the CO oxidation reaction. Kalyva et al. 19 observed a mechanism of Cu leaching out from the Pt−Cu NCs under oxidation−reduction cycles at elevated temperatures. Additionally, it was found that the local composition of alloy NCs can be affected by the presence of the support due to the preferential interaction of one of its elements with the support at the interface. Huang et al. 20 showed that Pt−Cu alloyed NCs supported on TiO 2 undergo extensive transformations when exposed to either an oxidizing or a reducing environment, resulting in a surface reconstruction that is different from that of the bulk, highlighting the important role of the support in the dispersion and morphology of the alloyed NCs. Xi et al. 21 demonstrated that in the case of Pd−Cu NCs supported on WO 2.72 nanorods, the strong interaction between the NCs and the support stabilized Cu in the NCs in an acidic environment.
Despite the weak interaction of nonreducible oxides such as Al 2 O 3 , SiO 2 , ZrO 2 , and MgO with metals, 22,23 only a few studies have investigated the mechanism of electronic and/or geometric modifications of noble metal-Cu alloyed NCs in the presence of these supports, generally used to enhance the NCs stability, by controlling the process conditions in selected atmospheres. A recent study, conducted by Le et al., 24 demonstrated the role of different supports, such as TiO 2 , SiO 2 , and γ-Al 2 O 3 , in controlling the crystal structure of Pd− Cu NCs, which was reflected in different catalytic activities and selectivities in the hydrogenation of succinic acid.
Previous studies 25−27 have elucidated the impact of the support on the structural evolution of Au 1−x −Cu x colloidal NCs with well-controlled size and composition supported on γ-Al 2 O 3 and SiO 2 at high temperature (350°C) in selected atmospheres. Specifically, it was proven how the type of support impacts the phase segregation between Au and Cu. Through different characterization techniques, it was shown that, under oxidizing conditions, Cu was dealloyed from Au and the formed CuO x species had different fates, depending on the support: while they were finely dispersed on alumina and partially migrated away from the Au NCs, the CuO x species on silica formed small clusters located in the proximity of the Au NCs, with limited Cu migration on the support. Changing the gas atmosphere to a reducing one restored the Au−Cu alloyed NCs when supported on alumina. This means that changing the gas atmosphere reversed the migration process of Cu. A partial realloying also occurred on the silica-supported Au−Cu NCs, resulting in the formation of Cu-depleted alloy NCs and isolated metallic Cu. For such bimetallic NCs, and under the investigated conditions, Cu was the only element that underwent the dealloying/migration/realloying process as a function of the reacting atmosphere.
Considering the mobility of Cu on different supports and the role of Au as a reversibility anchor in the alloying/ dealloying/migration processes for Au−Cu NCs, we studied here the effect of other two noble metals, i.e., Pd and Pt, in the compositional rearrangements of two families of supported noble metal-Cu alloyed NCs in response to oxidizing and reducing atmospheres at 350°C, i.e., the same conditions used in the previous studies to effectively remove the ligands from the surface of deposited NCs. 25, 26 Moreover, a detailed characterization was carried out on the initial state of the NCs before exposing them to the reaction mixture for the subsequent catalytic application (CO oxidation reaction, not reported here) along with their transformations upon exposure to the above-mentioned treatments. In this regard, the effects of temperatures beyond 350°C, i.e., the maximum temperature of the treatment, were not explored. Two types of Cubased bimetallic alloyed NCs, namely Pd−Cu and Pt−Cu, were synthesized through colloidal synthesis methods, which allowed fine adjustment of size and atomic ratio between the Cu and the noble metal (atomic noble metal: Cu = 50:50). The as-synthesized NCs were deposited on γ-Al 2 O 3 and calcined in static air to remove the organic ligands present on the NC surface. Their transformations upon oxidative or reductive conditions were studied, applying the same treatment protocols and using the same supports of the previous papers. 25, 26 Among other techniques, in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) and X-ray absorption spectroscopy (XAS) were specifically used to characterize the evolution of the chemical states and compositions (surface and bulk) after and during the treatments. A varying extent of the alloy/dealloy process, depending on the type of noble metal involved, was observed. During the oxidizing treatment on the Pd−Cu NCs, Cu was mostly retained in the NCs, with the formation of a PdCuO mixed oxide (Scheme 1, left panels), while, in the Pt−Cu NCs case, Cu was found finely dispersed on the alumina support as CuO x species, and Pt stayed in the NCs (Scheme 1, right panels). In a reducing atmosphere, an almost complete realloying of the Pd−Cu NCs was found, compared to the partial restoration of the Pt−Cu NCs. These observations highlight the capability of the Pd−Cu system to regenerate its initial structure thanks to the formation of the CuO-Pd interface that promotes the realloying of the Cu.
Bimetallic Alloy Nanocrystals. NCs were synthesized by employing wet chemistry methods, which allowed the control of their size, shape, and composition. Pd−Cu NCs were prepared according to modified one-pot procedures reported by Shan et al. 7 and Yin et al. 28 while Pt−Cu NCs were synthesized using the procedure reported by Yu et al. 29 with minor modifications. In a typical synthesis of alloy NCs with an atomic Pd/Cu ratio equal to 50/50, 0.25 mmol palladium(II) acetylacetonate and 0.25 mmol copper(II) acetylacetonate were dissolved in 20 mL of benzyl ether. One mmol of 1,2hexadecanediol dissolved in 5 mL of benzyl ether was added as the reducing agent. The mixture was heated slowly to 105°C under N 2 atmosphere, followed by the addition of 0.714 mL of OlAc and 0.741 mL of OlAm as capping agents to the as-formed dark solution. After the injection, N 2 purging was stopped, and the temperature was increased to 220°C keeping the reaction mixture at reflux for 30 min. The final product was cooled down to room temperature and Scheme 1. Sketch of the Pd−Cu and Pt−Cu Alloy NC Transformations as a Function of the Reaction Environments transferred to the glovebox. The NCs were precipitated out by adding 25 mL of anhydrous isopropanol and centrifuging at 1000 rcf for 30 min. In the second washing step, 200 μL of anhydrous chloroform was used to wash the wall of the vials, then 100 μL of OlAm and 25 mL of isopropanol were added. After a second centrifugation at 1000 rcf for 10 min, the NCs were dispersed in anhydrous chloroform.
Pt−Cu NCs with an atomic Pt/Cu ratio of 50/50 were synthesized at room temperature by mixing 0.5 mmol of copper(II) acetylacetonate, 0.5 mmol of platinum(II) acetylacetonate and 5 mL of OlAm. The solution was heated to 280°C at a rate of 5°C min −1 , and then it was cooled down to room temperature. The solution was diluted with 5 mL of ODE at 80°C to avoid the agglomeration and the coalescence of NCs that may occur during the cooling. The black product was precipitated by adding 40 mL of ethanol and separated by centrifugation at 8421 rcf for 5 min. In the second washing step, 200 μL of toluene was used to wash the wall of the vials, then 40 mL of ethanol was added, and the NCs were precipitated by centrifugation at 8421 rcf for 5 min. Finally, the NCs were dispersed in toluene.
Monometallic Nanocrystals. Monometallic Cu, Pd, and Pt NCs were prepared and supported on γ-Al 2 O 3 as a basis of comparison with the bimetallic ones. Cu NCs were synthesized using a method reported in the literature: 30 4 mmol of copper(I) acetate was mixed with 6.6 mmol of oleic acid and 15 mL of trioctylamine and degassed at 180°C under an inert atmosphere of N 2 for 1 h. Then, the solution was quickly heated to 270°C and kept at this temperature for 15 min. The NCs were precipitated by adding 25 mL of ethanol and separated by centrifugation at 6654 rcf for 5 min, and then dispersed in hexane, generating a final green oxidized Cu 2 O nanocrystal solution. Pd NCs were prepared according to the procedure reported by Jin et al. 31 Typically, a solution containing 0.33 mmol of palladium(II) acetylacetonate, 8 mL of ODE, and 10 mL of OlAm was heated to 100°C under N 2 flux. 0.33 mmol of MB dissolved in 2 mL of OlAm was added into the above solution. The resulting solution was heated to 130°C and was further kept at this temperature for 20 min. Then, the solution was cooled to room temperature. The NCs were washed by adding 30 mL of anhydrous ethanol and precipitated by centrifugation at 1000 rcf for 5 min. The final product was collected and dispersed in hexane. Pt NCs were synthesized using the procedure reported in ref 32. In a typical synthesis, 0.13 mmol of platinum(II) acetylacetonate was mixed with 10 mL of OlAm. The mixture was heated to 100°C under argon atmosphere. After 20 min, 3 mmol of TBAB dissolved in 5 mL of OlAm was injected into the above solution. The temperature was raised to 120°C and kept for 1 h before cooling down to room temperature. For the washing procedure, repeated twice, 30 mL of ethanol were added, and the NCs were precipitated by centrifugation at 6654 rcf for 5 min. The final product was collected and dispersed in hexane.
Alumina-Supported Nanocrystals. Typically, a dispersion of γ-Al 2 O 3 powder, in hexane for Cu, Pd, and Pt NCs, in chloroform for Pd−Cu alloy NCs and toluene for Pt−Cu alloy NCs, was sonicated for 5 min. A solution containing an appropriate volume of NCs was added to the support dispersion and left under fast stirring for 2 h. The powder was recovered by centrifugation at 1000 rcf for 5 min. The sample was finally dried in a vacuum oven at 40°C for 1 h. The resulting powder was calcined at 450°C for 3 h in a muffle furnace under static air to completely remove any capping ligand from the surface of the NCs. The calcination conditions were chosen based on the results of the thermogravimetric analysis (TGA) performed on colloidal NC solutions ( Figure S1 of the Supporting Information, SI). The differences in the weight loss observed among the samples could be attributed to the variation in residual solvents and ligands uptake on the different samples.
Redox Treatments. The experiments were performed in a flow reactor consisting of a vertical quartz tube (6 mm internal diameter) where the calcined alumina-supported NCs were placed between two beds of quartz wool. All gases were introduced into the reactor via calibrated mass flow controllers, and a tubular furnace was used to heat the reactor. One thermocouple was placed inside the sample to monitor the temperature. The supported NCs were exposed to an oxidizing (6% v/v O 2 diluted with He) and a reducing (5% v/v H 2 diluted with He) atmospheres up to 350°C for 1 h. The initial oxidation after loading was done to ensure a clean surface of the samples free of any adsorbed water or formed carbonates due to storage in an open environment. The heating rate used in both treatments was 5°C min −1 , and the total flow rate was 60 mL min −1 .
Nanocrystal Characterization Techniques. Transmission Electron Microscopy. The bright-field transmission electron microscopy (BF-TEM) images of NCs were recorded using a JEOL JEM-1011 instrument with a thermionic W source, operated at 100 kV. The samples were prepared by drop-casting the diluted NC solution or the alumina-supported NC powders suspended in hexane, toluene, or chloroform onto a carbon-coated 200 mesh Cu grids. Highresolution TEM images (HRTEM) were collected by a JEOL JEM-2200FS microscope (Schottky emitter), operated at 200 kV, equipped with a CEOS spherical aberration corrector for the objective lens, and an in-column Omega energy filter. The chemical composition of the NCs was determined by energy-dispersive X-ray (EDX) spectroscopy performed in the high-angle annular scanning TEM (HAADF-STEM) mode with a Bruker XFlash 5060 detector. The STEM-EDX maps were acquired using Cu Kα and Pd/Pt Lα lines, taking care to select signals only from the element of interest, without contributions from neighboring X-ray peaks. For HRTEM and STEM-EDX analyses, the samples were prepared by drop-casting NCs solutions onto ultrathin carbon-coated Ni grids. HAADF-STEM images of Cu NCs supported on γ-Al 2 O 3 were acquired using a FEI Tecnai G2 F20 microscope (Schottky emitter), operated at 200 kV acceleration voltage. The selected area electron diffraction (SAED) patterns were acquired using the same microscope with a constant camera length, bringing the sample to eucentric height and eucentric focus. The camera length was calibrated using a nanocrystalline Au sputtered film on a standard carbon-coated Cu grid. The NC average size and distribution were obtained using ImageJ software. 33 Alumina supported NCs were processed by manual counting using Gatan Digital Micrograph software from analysis of the HAADF-STEM images to determine the NC size distribution.
Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) Analysis. The measurements were carried on an iCAP 6000 Thermo Scientific spectrometer for quantification of the elemental composition of NCs and the metal loading. The samples (specific volume of NCs colloidal solution or weight of the alumina supported NC powder) were digested in aqua regia HCl:HNO 3 3:1 v/v (Sigma-Aldrich for trace analysis) overnight. Ultrapure Milli-Q water (18.2 MΩ cm) was added to the sample, and any remaining solids were filtered using a PTFE filter before the measurements. All chemical analyses were affected by a systematic error of about 5%.
X-ray Powder Diffraction (XRD). XRD patterns of the NCs were collected on a PANalytical Empyrean X-ray diffractometer equipped with a 1.8 kW Cu Kα ceramic X-ray tube, PIXcel 3D 2 × 2 area detector and operating at 45 kV and 40 mA. The samples were prepared by dropping a concentrated NC solution or by directly depositing a powder onto a zero-diffraction silicon substrate. The diffraction patterns were performed at ambient conditions in a parallel-beam geometry and symmetric reflection mode over an angular range 30°−90°, with a step size of 0.05°. High Score 4.1 software from PANalytical was used for phase identification.
In Situ Characterization Techniques. In Situ DRIFT. The measurements were performed using a Bruker Optics Vertex 70 FTIR spectrometer, equipped with a Praying Mantis DRIFT cell. Liquid nitrogen cooled Mercury Cadmium Telluride (MCT) detector was used for data acquisition and OPUS software for data processing. The outlet gaseous species were analyzed with a mass spectrometer (Pfeiffer Omnistar). A four-port selector valve was used to switch between two different gas streams, one used for the treatments and the other containing the CO probe. In a typical experiment, the DRIFT cell was loaded with 30 mg of the sample packed on top of about 80 mg of γ-Al 2 O 3 (fine powder with particle size <63 μm). The supported NCs were treated under oxidizing and reducing atmospheres prior to the test, under the same gas compositions and temperatures described in the Redox treatments Section, except for the heating rate, which was set to 10°C min −1 , and the flow rate (80 mL min −1 ). The effect of changing the set parameters was verified and did not lead to different results (not reported here). The measurement sequence was the following: the sample was cooled to room temperature under a He flow, and a background spectrum was recorded at 25°C directly after the treatment. After collection of the background, the He gas stream was switched to a stream containing 0.2% v/v CO balanced with He. Nine absorption spectra were collected every 10 s from the gas switch. After 6 min from the beginning of the adsorption process, additional 5 spectra were collected every 60 s, approaching the surface saturation. Then, the sample was purged with He and desorption spectra were acquired with the same frequency as in the adsorption phase.
In Situ XAS. Data were recorded in transmission mode at the Cu Kedge (8979 eV), Pt L III -edge (11564 eV) and Pd K-edge (24350 eV) on the ROCK (Rocking Optics for Chemical Kinetics) beamline of the synchrotron SOLEIL (France). For the measurements during the reduction/oxidation treatments, the calcined bimetallic supported NCs were exposed to 60 mL min −1 of a mixture of 5% v/v H 2 /He and then to 6% v/v O 2 /He. Spectral acquisition was also done on the samples in He at room temperature before each treatment for data comparison. Thanks to the edge jumping capability of the quickscanning extended X-ray absorption fine structure (EXAFS) monochromator, 34 both the edges of elements composing the NCs were characterized simultaneously. Indeed, Si(111) and Si(311) crystals were alternatively used as a monochromator for the Cu and Pd/Pt edges. Data were recorded for 1 min at one edge before changing the monochromator to the other one (with about 30 s of dead time for the exchange) and then keeping the subsequent minute at the other edge of interest. The cell allowed to perform in situ treatments of the supported NCs under controlled conditions such as temperature, pressure, and chemical environments during the experiment according to the assembled system described by La Fontaine et al. 35 The calibration of the energy scale was ensured by the simultaneous measurement of the absorption spectrum of the correspondent metallic foil of the elements composing the NCs set between the second and third ionization chambers. The reference PtO, CuO, and Cu 2 O XAS spectra were measured in transmission mode at the SuperXAS beamline of the Swiss Light Source (SLS) at Paul Scherrer Institute (PSI Synchrotron). After normalization, the collected spectra were used as qualitative comparison of XANES part features with those of the pure XANES spectra of the formed compounds during the two treatments.
XAS Data Processing. The collected XAS spectra were initially extracted, subsequently calibrated and normalized by means of a graphical user interface (GUI) within Python developed at the ROCK beamline, as described in the literature. 36 The normalized XAS data sets were then processed using the methodology of multivariate curve resolution-alternating least-squares (MCR-ALS) to unravel the number and the concentration profiles of the species evolving throughout the treatments. The number of variability sources, related to the number of chemical species, was analyzed using singular value decomposition (SVD), and evolving factor analysis (EFA). 37,38 Assuming that the experimental data follow a linear model, this statistical analysis of minimization was carried out using the MCR-ALS GUI 2.0 developed by Tauler et al. 39 on the Matlab platform, which decomposes the series of time-resolved spectra (D), recorded during reaction into pure species whose relative concentrations vary with time, as follows: where C is the matrix containing pure concentration profiles and S T is the transposed matrix of S containing pure XAS spectra of the k species of the mixtures. E is the matrix of residuals, which contains the variability not explained by the model, ideally close to the experimental error. Non-negativity and unimodality constraints were applied on the matrices C and S, responsible for the observed data variance, to help the convergence of the multivariate curve resolution. EXAFS extraction, Fourier transform (FT), and EXAFS simulation were performed using the software ATHENA and ARTEMIS within the Demeter package to obtain structural parameters: degeneracy of selected path (N) for each shell, interatomic distance (R), and Debye−Waller factor (σ 2 ). Specifically, after data conversion into k space, the k 2 -weighted EXAFS functions were Fourier transformed and fitted in R space simulating the experimental signal. The fitting was performed for the first and second shell scattering at both edges to determine the identity, number and positions of the nearest neighbors and thus to generate the cluster around the selected absorbers. Initially, the ATOMS and FEFF packages implemented inside the program were employed to generate ab initio the scattering paths for the defined clusters starting from model compounds of Metal load obtained from ICP-OES and the size after the two treatments for the mono-and bimetallic alumina supported NCs (NM = noble metal). b Standard deviation of four repeated tests. Figure S5. The Pt−Cu NCs were homogeneously distributed on the support, regardless of the treatments, with no NC morphology change; their average size remained unchanged (Table 1) and about equal to that of the assynthesized NCs. Pd−Cu NCs were also well dispersed on the grains of the alumina support after both oxidizing and reducing treatments, confirming that no significant sintering phenomena had occurred with time on stream. However, it is not possible to completely exclude the formation of a small fraction of large particles possibly related to the sintering that occurred during calcination. However, a significant increase of the average size of the Pd−Cu NCs to about 53% was observed after oxidation, mainly due to the formation of a flocky shell on the NC surface, poorer in Pd than the NC core, as evidenced by the STEM-EDX analysis (Table 1 and Figures 2, S5a, and S6). Relative to the size of the supported monometallic NCs, no significant changes were observed in the case of Pd NCs after the two treatments, as well as for Pt NCs after reduction. After oxidation, the Pt NC size increased by 47%. A notable change was especially found for oxidized and reduced Cu NCs with triple and double size increase, respectively ( Figure S7a,b and Table 1). This can be explained as a result of NC agglomeration induced by reactants and/or sintering phenomena under redox treatments by coalescence of smaller NCs or by Ostwald ripening, 44 which were altered in the presence of a second metal. Nevertheless, the presence of larger NCs in the as-synthesized Cu NCs could enhance the above-mentioned process, due to the detachment of metal atoms from small particles with high chemical potential as transient monomers, diffusion onto the support as oxygen−metal complexes, and subsequent attachment to larger particles with lower chemical potential. The bimetallic NCs were analyzed by STEM-EDX to study the spatial distribution of the elements after the two treatments. After oxidation, STEM-EDX maps showed that Cu was predominantly associated with Pd within the NCs, and only a small amount of Cu had diffused onto the alumina support. In particular, larger NCs exhibited a Cu-rich shell surrounding a Pd-rich core (Figure 2a). The comparison between EDX spectra (not shown) extracted from two identical areas in the map in Figure 2a, one from the shell and the other one from the core region of one of NCs on the surface of the alumina support fragment, returned a Cu/Pd atomic ratio of 0.8 in the NC shell region, to be compared to 0.3 in the NC core. However, while Pt was still localized in NCs entities, a higher concentration of Cu was detected from all over the alumina support (Figure 3a), as found for Au−Cu NCs on Al 2 O 3 . 25 After reduction, Pd and Cu were localized in the same region of the Pd−Cu NCs, as well as Pt and Cu in the Pt−Cu NCs, with only a weak Cu signal detected on the support (Figures 2b and 3b), as in the case of the Au−Cu NCs. 25 Due to the low metal loading, the small NCs, and the presence of γ-Al 2 O 3 , the XRD profiles obtained after the treatments were mostly dominated by the support contribution, thus limiting their interpretation. For this reason, SAED patterns of supported Pd−Cu and Pt−Cu NCs were collected after the treatments and compared with the pattern of the assynthesized NCs (Figures 2c and 3c). Since the peaks of γ-Al 2 O 3 partially overlapped with the peaks of the NCs, the SAED pattern of pure γ-Al 2 O 3 is also reported. After oxidation of the supported Pd−Cu NCs (Figure 2c), the collected SAED Figure S8). On the basis of these results, it was possible to estimate whether the measured increase of NC size could be ascribed to the formation of Pd 0.7 Cu 0.3 O NCs. From the experimental data, the average size of the starting Pd 0.5 Cu 0.5 NCs was 4.9 ± 1.8 nm, while that calculated after oxidation with eq S3 was 7.0 nm, a value in agreement with that observed in the HAADF-STEM images (7.8 ± 2.0 nm). Thus, an increase in NC average size of 43% was observed compared to the initial NC size. After reduction, the supported Pd−Cu NCs pattern (Figure 2c) showed characteristic peaks related to the fcc phase of Pd−Cu. Indeed, the comparison with the pattern of the as-synthesized NCs suggested that the Pd−Cu alloy composition with 36 at. % of Cu (estimated by Vegard's law) 40 was restored. The SAED pattern of Pt−Cu/Al 2 O 3 after oxidation (Figure 3c) featured diffraction peaks that were characteristic of a cubic structure, with a 3.9 Å lattice parameter corresponding to Pt; no Cu or CuO x peaks were detected. By coupling this information with the STEM-EDX maps, we could conclude that Cu was highly dispersed onto the support either as an amorphous phase or as a finely dispersed phase. The reduced Pt−Cu/Al 2 O 3 SAED pattern (Figure 3c) resembled the one after oxidation, with no Chemistry of Materials pubs.acs.org/cm Article significant shift of the peak positions toward higher 2θ angles due to the reincorporation of Cu into the NCs. Indeed, the obtained lattice spacing of 3.9 Å was comparable to that obtained after oxidation, suggesting no difference between oxidized and reduced Pt−Cu NC samples. This result can be explained by considering that it is not possible to rule out the exposure of the reduced sample to air when unloaded from the reactor. For this reason, as soon as the system was exposed to air, it was immediately oxidized. Additional experimental proofs for this conclusion will be provided in the next sections describing the EXAFS results. Therefore, the calculated value of Cu incorporation in the alloy after reduction (12 at. % from Vegard's law) represented a rough estimate of the state of the Pt−Cu system after this treatment.
NC Surface Chemistry Evolution. DRIFT tests on supported bimetallic NCs after oxidizing and reducing treatments were performed to probe the NC surface composition and obtain information about the electronic features of the metals in NC surface. It was found that the different sequences of reduction and oxidation on calcined supported bimetallic NCs do not influence the spectra in terms of number and position of the bands observed (not reported), meaning that the surface transformations are fully reversible. In addition, the effect of oxidative treatment with or without the initial calcination, which was typically done to remove the capping agents, was investigated and no difference was found (not shown here). This proved that the in situ oxidation treatment was also efficient in removing the capping agents and induced the same transformation when O 2 is present in the atmosphere. Figures 4 and 5 report the DRIFT spectra in the carbonyl region at different exposure times during room temperature absorption and desorption of CO on/from the two bimetallic NCs. Under the same conditions, CO adsorption spectra for supported Cu, Pd, and Pt NCs were also acquired ( Figures S9, S11, and S12). The CO adsorption on the oxidized Pd−Cu NCs (Figure 4a,b) resulted in the appearance of two overlapping bands at 2153 and 2137 cm −1 (note the difference in the intensity compared to the reduced ones). As it is reasonable to assume that the Cu is mostly oxidized to Cu 2+ at the conditions of the oxidative treatment, and since carbonyl on Cu 2+ are not stable and not detectable at room temperature, 46 we assigned these two main bands to the carbonyl on Pd 2+ species. The difference in the band positions could be attributed to the Pd adsorption sites in different coordinating environments, i.e., either localized as PdO phase or in close proximity with the Cu in the Pd 0.7 Cu 0.3 O phase. However, we could not rule out the presence of Cu + species with respect to the data recorded on the monometallic Cu NCs (Figure S9a,b). To have a better band separation, the 10 min spectrum of oxidized PdCu was magnified and deconvoluted, which resulted in 5 bands ( Figure S10a). The major components at 2153 and 2131 cm −1 could be assigned to carbonyl on Pd 2+ , in PdO and Pd 0.7 Cu 0.3 O phases, respectively, as stated above. The minor band 2102 cm −1 could then be attributed to carbonyls on Cu + in close contact with Pd. 47,48 The above description is consistent with a prior investigation of supported Pt−Cu NCs that highlighted how the atomic closeness of Cu atoms with Pt ones affected the mutual environment and, thus, the positions of the bands. 49 The presence of Cu + −CO species could be inferred from the stability of this band during the desorption at room temperature (see the deconvoluted 10 min spectrum in Figure S10b), although, since the desorption at high vacuum and over a longer time could not be performed, bands on Pd 2+ were still present.
The Cu + −CO band position in the CO desorption spectrum was shifted to 2110 cm −1 due to the strong dynamic coupling between the adsorbed molecules with increasing CO surface coverage. 46 The additional and very low-intensity bands at 2069 and 1990 cm −1 could be related to linearly bonded and bridge-bonded Pd 0 −CO, respectively, suggesting partial oxidation of the supported Pd−Cu NCs, as found for the monometallic ones ( Figure S11a), possibly due to a temperature deviation of the sample inside the cell vs the set point. In addition, the high-frequency band at 2153 cm −1 which we assigned to CO linearly adsorbed on Pd 2+ atoms, could be corroborated by performing the same experiment on the monometallic Pd NCs (Figure S11a,b). 46 After reduction (Figure 4c,d), three main bands centered at about 2122, 2070, and 1991 cm −1 with a shoulder at 1942 cm −1 appeared. The three latter vibrational bands can be straightforwardly attributed to CO bonded to surface Pd atoms in the linear, bridged and 3-fold hollow bonded forms, respectively. 50 As the surface coverage increased by exposure time, the CO stretching frequency of Pd-bounded CO shifted to a higher frequency (about 10 cm −1 ) due to the dipole−dipole coupling between adsorbed CO molecules on Pd surface. Although the frequency of the band at 2122 cm −1 was in the range of CO adsorbed on Cu + (2110−2140 cm −1 ), the attribution of this band to CO adsorbed on Cu 0 can be justified by assuming an electronic modification by neighboring Pd atoms that increases the CO− Cu 0 strength and, thus, shifts the band position. Indeed, the frequencies of the bands assignable to linearly bonded Pd 0 − CO and Cu 0 −CO carbonyls in the spectra of reduced Pd ( Figure S11c,d) and Cu ( Figure S9c,d) NCs were about 16 cm −1 higher and 4 cm −1 lower, respectively, than those recorded for the bimetallic NCs. This confirmed the presence of an electron transfer from Cu to Pd such that the electronic properties of Pd atoms were strongly modified by Cu addition and vice versa. After 10 min of evacuation (Figure 4d), the incomplete disappearance of the band at 2122 cm −1 implied that the Pd−Cu NCs might not be fully reduced due to the presence of residual Cu + species on which CO was strongly adsorbed in addition to reduced Cu 0 . This was in agreement with SAED characterization, which suggested that the copper was not completely reincorporated into the alloy. However, as in the case of the oxidized sample, the incomplete disappearance of the bands could be related to the fact that the desorption could not be performed either under higher vacuum or for a longer time. The schematic visualization of the band assignment of carbonyl species on Pd−Cu system is reported in Figure 4e.
CO adsorbed on the oxidized Pt−Cu NCs (Figure 5a,b) resulted in a main absorption band at 2090 cm −1 and a second weak band at 2122 cm −1 . The former was attributed to the linear CO species on Pt atoms 46 and the latter to those on Cu + atoms, 51 in agreement with what obtained for monometallic Pt (Figure S12a,b) and Cu ( Figure S9a,b) NCs. One should note that most of the Cu atoms are reasonably oxidized to the 2+ state, and the presence of Cu + could be due to the deviation of the temperature in the DRIFT cell and the temperature set point. The spectra after reduction (Figure 5c 52 In this way, the band around 2066 cm −1 was attributed to this latter species. The high-energy band at 2121 cm −1 resulted from CO bound to Cu 0 , even if the position of this band would be more consistent with Cu + rather than Cu 0 . 53 The blue-shift of about 28 cm −1 of the frequency of CO on Pt in Pt−Cu NCs compared with that in monometallic Pt NCs ( Figure S12c) was attributed to the electronic interaction between the two metals and a decreased dipole−dipole coupling due to the dilution of Pt by Cu. 54 Likewise, the frequency of CO on Cu 0 in the bimetallic NCs was about 4 cm −1 higher than the one recorded for reduced Cu NCs ( Figure S9c). Indeed, this effect enhanced the electron density on Pt atoms ensuring that, during the evacuation, the adsorption of CO on Pt was much stronger than that on Cu. 55 Furthermore, the evidence that the Pt and Cu were atomically mixed in the NCs was provided by two effects visible in the spectra: the first one was associated with the appearance of the band at 2013 cm −1 related to CO bridged between Pt and Cu atoms; 56 the second was visible in the spectra during the evacuation (Figure 5d) in which the intensity of the Pt 0 −CO band increased at the expense of that of Cu 0 −CO due to an energy intensity redistribution effect. 49 The infrared band assignments for the Pt−Cu system are summarized in Figure 5e.
NC Geometric and Electronic Structural Modifications. XAS was applied to obtain information about the local geometric and/or electronic structure of NCs in response to the variation of the gas environments and temperature. In this regard, the transformations of Pd−Cu and Pt−Cu NCs were in situ monitored at the Pd K-edge, Pt L III -edge, and Cu K-edge X-ray absorption near edge structure (XANES) spectra while the NCs were subjected to the reducing and oxidizing treatments. The overview of all XANES spectra recorded during the experiments is presented in Figures S14 and S15. Initial XAS spectra were recorded at the Pd K-, Pt-L III -, and Cu K-edges of the as-synthesized Pd−Cu and Pt−Cu NCs (not shown). The results of the EXAFS analysis for Pd−Cu NCs (Tables S1 and S2) indicated the presence of alloy nanoclusters by the detection of the Pd−Cu 1 and Cu−Pd 1 coordination signals at 2.63 and 2.65 Å, respectively, for both Pd-and Cu K-edges, in agreement with what observed by STEM-EDX, HRTEM, and XRD analysis. For Pt−Cu NCs, the fitting of EXAFS (Tables S3 and S4) indicated the formation of Pt−Cu alloy nanoclusters at both edges along with the formation of the CuO phase at the Cu K-edge. A higher total coordination number (CN) for Pt (Pt−Pt 1 and Pt−Cu 1 9.08) compared to that one for Cu (Cu−Cu 1 and Cu−Pt 1 3.28) was found in the Pt−Cu alloy, indicating that Cu was preferentially segregated. This result can be justified considering a possible alloy inhomogeneity or a beginning of metal segregation due to the difference in atomic radius between Pt (1.39 Å) and Cu (1.28 Å).
XAS spectra of alumina-supported Pd−Cu and Pt−Cu NCs were collected in He as the starting point of the measurement ( Figure S13). The Pd and Cu K-edge XANES spectra of Pd− Cu NCs were similar to those of PdO and CuO standards ( Figure S13a,b). The EXAFS analysis (Tables S1 and S2) confirmed the presence of PdO and CuO along with the formation of PdCuO mixed oxides species with Cu−O and (Tables S3 and S4) confirmed the formation of PtO, with a major contribution of metallic Pt and CuO. These results were consistent with the oxidation occurring during the calcination step, necessary for the removal of the organic ligands from the NCs surface, during which NC dealloying occurred. The concentration profiles of the two obtained principal components named D r (descending component, i.e., the concentration of which was decreasing) and A r (ascending component, i.e., the concentration of which was increasing) formed during the reduction treatment at the Pd K-edge of the Pd−Cu NCs are reported in Figure 6a,b. The composition profile evidenced a significant reduction of the component D r and a corresponding increase of the component A r at T > 50°C . By comparing the D r and A r component spectra ( Figure 6b) with different standards, it is possible to claim that the D r spectrum displayed XANES edge features resembling those of the PdO reference, 57 while the A r spectrum showed features similar to the metallic Pd reference. From the fitting procedure of the Fourier transform (FT) of the EXAFS spectra of the two components D r and A r , the D r is described by two major contributions corresponding to the Pd−O and Pd−Pd coordination shells at distances of 2.03 and 3.06 Å, related to PdO (Table S1). The additional formation of PdCuO mixed oxide-like was only observed during the EXAFS data refinement. The best EXAFS fitting for the A r component was instead obtained with a Pd−Pd contribution at 2.70 Å and Pd−Cu one at 2.64 Å, characteristic of Pd−Cu. The large contribution of Pd−Pd distances found in the first coordination shell suggested the formation of a Pd-rich Pd−Cu disordered alloy. This observation is in line with the results obtained from the SAED data, in which the composition of the reduced NCs does not fully match with the starting NC alloy, while a Pd-rich alloy was instead found.
During the oxidative treatment at the Pd K-edge, the evolution of the concentration profiles of the two bimetallic NCs as a function of temperature, along with their pure spectra, are presented in Figure 6c,d. Specifically, the concentration of the component D o started to decrease gradually at about 150°C under exposure to oxygen and becoming increasingly evident during the heating ramp in the oxidation atmosphere with a simultaneous increase of the component A o . An initial visual inspection of the pure XANES spectra of the components with respect to the references ones ( Figure 6d) showed that the shape of the features of the XANES part of D o was consistent with what has been seen in bulk Pd−Cu 57 while the A o component features were similar to those of the PdO reference. The qualitative description of these two principal components was then corroborated by the analysis of the FT of Pd K-edge EXAFS oscillation of the representative components during oxidation. The best agreement between the observed and calculated EXAFS for D o was achieved by using a structural model derived from the Pd−Cu alloy. In particular, a Pd-rich Pd−Cu disordered alloy was obtained from the fit. The component A o was instead fitted using the structural models of PdO and PdCuO mixed oxide (Table S1). In situ XAS at the Cu K-edge confirmed the Cu reduction and partial realloying under reducing conditions and then the dealloying during the oxidative treatment (see Figure S16 in the SI for additional details). To summarize, the fitting carried out at the Pd-and Cu K-edges during reduction indicates the formation of PdO and CuO with a minor contribution of PdCuO mixed oxide at the initial stage of the reduction. As the temperature was increased, Cu 2 O and Pd−Cu alloy were formed as intermediate species. The presence of Cu + as an intermediate is commonly evidenced by XAS as was found for the systems in which Cu is supported on Al 2 O 3 , SiO 2 , and ZrO 2 compared to the bulk powders in which the direct reduction of CuO to metallic Cu is reported. 58,59 Finally, Pdrich Pd−Cu disordered alloy and metallic Cu with a minor contribution of Cu 2 O were formed in the last step of the reduction, indicating that the initial alloy structure was not fully recovered. During the oxidation, the Pd-rich Pd−Cu alloy was already segregated at room temperature, with the formation of Cu 2 O as intermediate species to yield PdO, CuO, and, to a minor extent, a PdCuO mixed oxide. A heat treatment at 350°C was required to fully oxidize the Cu oxide species in the +2 oxidation state, restoring the system to the initial situation recorded after the calcination. This outcome is partially in agreement with the picture obtained by DRIFT measurements. Indeed, due to the DRIFT setup limitations and the intrinsic deviation from the actual temperature set point reached inside the cell at high temperatures, complete oxidation of Cu was not observed under DRIFT. In addition, one should note that the DRIFT experiment is more surfacesensitive while the XAS elucidates the bulk properties, which could be overlooked by the DRIFT. XAS data were also examined at the Pt L III -and Cu K-edges for the aluminasupported Pt−Cu NCs during the oxidative and reductive treatments. On the basis of the concentration plot of the principal components during the reduction at the Pt L III -edge (Figure 7a), the component D r rapidly decreased already at room temperature under the exposure to H 2 with the consequent increase of the A r component. By comparison with the references XANES spectra (Figure 7b), the D r spectrum visually resembled the Pt one with a characteristic shape and edge shifted to higher energy with respect to the Pt reference one, 60 to which the A r spectrum approached. From the FT of the EXAFS part (Table S3), the best fit of the D r EXAFS data confirmed the assignment to metallic Pt (CN 4.24 ± 0.6 at 2.8 Å) with a minor contribution of PtO (first shell CN 1.79 ± 0.0.2 at 2.0 Å) at the initial stage of the reduction, as already observed for the supported NCs after calcination. Referring to the A r component, the high coordination number of Pt around Pt atoms in the first shell of Pt−Cu structure (CN 8.00 ± 0.6) suggested the formation of Pt-rich Pt−Cu alloy in the second stage of the reduction. The formation of these species during the exposure of the sample to a reductive environment was confirmed by the results obtained at the Cu K-edge (see Figure S17a,b in SI for more details). Two principal components were chosen to explain the variance of the data set during the oxidative treatment at the Pt L III -edge. From the concentration profile (Figure 7c), the descending D o component was found to be stable and present at room temperature under exposure to oxygen.
At the beginning of the heating ramp, the D o decrease was associated with the appearance of the ascending A o component. The two components showed similar XANES features related to metallic Pt by comparison with the standard Pt ( Figure 7d). In particular, the A o edge was shifted slightly to lower energies, indicating stronger metallic features. 61 Despite the similarity found by the visual inspection of the XANES spectra, the fit of the FT EXAFS part of the two components suggested an assignment to different species. Specifically, the presence of Pt-rich disordered Pt−Cu alloy at the initial stage of the oxidation as component D o was evident from the presence of the Pt−Pt 1 first shell at 2.72 Å with a coordination number of 6.96 ± 0.8 and Pt−Cu 1 one at 2.68 Å with a CN of 2.83 ± 0.7. The fitting of the A o FT EXAFS spectrum indicated that the Pt atom was surrounded by nine Pt atoms at a distance of 2.73 Å with a minor contribution of the PtO phase (CN Pt− O 0.36 in the first shell). The dealloying process was confirmed by the Cu K-edge data as well (see in Figure S17c,d the SI for additional details). To sum up, two principal components were found to be relevant for the supported Pt−Cu NCs during the exposure to H 2 . Metallic Pt and a minor fraction of PtO mixture phases were transformed into a Pt-rich Pt−Cu alloy already at room temperature, which remained unchanged until the end of the process. The formation of this latter species was in agreement with the results obtained from the SAED analysis, in which the initial NC alloy was found not to be fully restored. Complete segregation occurred during the oxidation treatment starting from a Pt-rich Pt−Cu alloy to metallic Pt and CuO phases at the end of the process, as confirmed by the previous techniques. Considering the previous results, it is possible to claim that, as for the case of the Au−Cu NCs, in which the CuO x was found as isolated species away from the Au NCs on the support after the oxidizing treatment, 25,26 the same situation was observed for the Pt−Cu NCs. This is different from the Pd−Cu NCs case, in which the oxidized Cu was retained by the palladium with the formation of PdCuO mixed oxide, along with a small fraction of CuO x . In general, during the reduction, the CuO x species, found around the NCs or dispersed on the support, were partially realloyed within the Pt NCs to a minor extent compared to the Pd−Cu NCs. Indeed, 36 at. % of Cu was found in the Pd−Cu NCs on Al 2 O 3 versus 12 at. % for Pt−Cu on the same support. Note that the latter value reported for Pt−Cu could deviate from the actual composition as mentioned before due to the limitation of Vegard's law in determining the composition of NCs having strain. In the latter case, the EXAFS results obtained at the end of the reductive treatment showed higher coordination numbers of 2.66 ± 0.5 in the Pd−Cu 1 first shell and 1.04 ± 0.63 in the Pd−Cu 2 s shell for Pd−Cu compared to those obtained for the Pt−Cu alloy (CN Pt−Cu 1 = 1.52 ± 0.6, CN Pt−Cu 2 = 0.76 ± 0.3). This suggested that a greater number of Cu atoms are present around the Pd atoms after the reductive treatment.
Thus, the different extent of NC dealloying/migration/ realloying process depending on the type of noble metal for these two systems could be explained considering the following considerations: (i) the slightly higher oxophilicity 62 of Cu compared to Pd and Pt makes it a sacrificial element toward oxidation. Indeed, in the case of Pd−Cu NCs system, the formation of CuO was favored by the more negative Gibbs free energy of formation of CuO with respect to PdO (ΔG CuO = −127 ± 4 kJ/mol and ΔG PdO = −65 ± 8 kJ/mol, respectively). 63,64 Then, the oxidation of Pd NCs was driven by the formation of the CuO−Pd interface due to the lower energy barrier for the penetration of oxygen in the Pd lattice at the CuO−Pd interface. 65 For Pt−Cu NCs, it has been shown previously that the adsorption of oxygen onto non-supported Chemistry of Materials pubs.acs.org/cm Article Pt−Cu NCs could induce segregation of the 3d transition metal onto the NC surface due to the strong interaction with oxygen (heat of formation of oxide for Cu −150 kJ/mol and for Pt −50 kJ/mol). 19,63,66 In particular, the exposure to oxygen causes outward diffusion of Cu and encapsulation of the particles by the formed Cu oxide layer due to the lower surface energy of CuO (<1 J/m 2 ) compared to that of Pt (1.9 J/m 2 ). 67 Furthermore, due to the larger atomic radius of Pt compared to Cu and the presence of more strain in the Pt−Cu NCs, the dealloying process is more favored in this system compared to the Pd−Cu one, in which the formation of PdCuO mixed oxide was able to retain the oxidized Cu. (ii) After reduction, the heat of formation of the Pd−Cu solid solution bulk alloy is −14 kJ/mol, 68 while that of the Pt−Cu solid solution is −11 kJ/mol. 68 This indicates that the formation of these alloys is favored in both cases, with a slightly increased stability for the Pd−Cu alloy. However, the incomplete restoration of the initial Pt−Cu NC alloy compared to the Pd−Cu one occurred due to the partial segregation of Pt driven by minimization of strain energy (due to the difference in surface energies and atomic radius between Pt and Cu). 19,66,69 A similar behavior was found in other Ptbased bimetallic alloys such as Pt−Cu, Pt−Ni, Pt−Fe, and Pt− Co. 66 Additionally, the presence of a larger fraction of CuO x species onto the support in the Pt−Cu system might result in a considerable fraction of Cu unavailable for alloying with Pt. In the context of the studies carried out so far in our research group addressing bimetallic alloys of noble and nonnoble metals, 25−27 we speculate that similar synergies are expected to operate for nanoalloys containing other noble metals, depending on the actual compositions and phase structures. Specifically, the non-noble part acts like a sacrificial component that segregates from the alloy and prevents sintering of the noble metal. The newly formed species on the surface of the non-noble metal could have a potential role in the catalytic properties. However, the extent of this outcome needs to be further explored in future works to rationalize the effect on the catalytic properties of such materials for the desired reaction.
■ CONCLUSIONS
In this work, we have studied the dynamics of structural transformations of Pd−Cu and Pt−Cu NCs supported on γ-Al 2 O 3 while heating them under exposure to a sequence of oxidizing and reducing gas atmospheres. Specifically, the oxidizing treatment led to a different scenario depending on the type of noble metal used. PdCuO mixed oxide was found in the case of supported Pd−Cu alloy NCs, along with PdO and a small fraction of CuO x , while for Pt−Cu alloy NCs, a larger amount of CuO x species migrated away from the Pt/PtO NCs on the support. The reducing treatment largely restored the Pd−Cu alloy NCs, and partially also the Pt−Cu ones, highlighting the different behavior depending on the type of noble metal and spatial distribution of CuO x .
We, therefore, concluded that the noble metal present in the bimetallic Cu-based alloy NCs has a strong influence on the dealloying/migrations/realloying processes occurring under typical heterogeneous catalytic reactions. Hence, the present work provides useful insights into the preparation of materials for catalysis. Indeed, the addition of nonprecious metal to Pd/ Pt reduces the mobility of the active Pd/Pt metal sites onto the alumina support, stabilizing them against sintering into large clusters under reaction conditions. The implications of these findings on elucidating the transformations of supported noble metal−Cu alloyed NCs upon different activation methods are significant and motivate our ongoing investigations on the finetuning catalytic activity of these NCs in the CO oxidation reaction. | 12,712 | 2021-11-10T00:00:00.000 | [
"Materials Science",
"Chemistry"
] |
A REVIEW OF RAILWAY TRANSPORTATION IN CENTRAL ASIA FOR CORRIDORS AND THE REVIVAL OF THE GREAT SILK ROAD
This review examines the state of railway transportation in Central Asia within the context of the Silk Road initiative, tracing its origins from the United States' New Silk Road strategy. Despite recent setbacks, the Silk Route has become a crucial geopolitical tool for military, political, and economic interests, with corridors like Kazakhstan's Terminal and the China-Kyrgyzstan-Uzbekistan-Afghanistan route identified for potential revitalization. Uzbekistan, particularly in the China-Kyrgyzstan-Uzbekistan corridor, emerges as pivotal for advancing the Silk Road initiative. A comprehensive analysis of data collection methodologies employed in studying these corridors is also included to ensure the reliability and validity of the findings. Employing a multifaceted methodology, including quantitative analysis of infrastructure metrics, case studies of specific projects, a literature review synthesizing existing knowledge, and comparative analysis drawing parallels and contrasts, this study explores opportunities and challenges in enhancing these corridors, aiming to contribute to the broader strategy's rejuvenation in Central Asia. By analyzing current conditions and proposing improvements, the research seeks to optimize railway transport, aligning with the Silk Road's historical significance in the contemporary geopolitical landscape.
INTRODUCTION
The historic Silk Road, which has connected Asia and Europe for centuries, has comprised various routes spanning a vast geography (Winter 2022).This unique path facilitated trade, prosperity, and the exchange of knowledge, promoting cultural integration and information sharing.The trade-driven interaction along the Silk Road exposed small villages to new and different beliefs, knowledge, and ideas.Initially a consequence of multifaceted expectations, supplies, and demands from diverse sources, the Silk Road eventually declined in significance due to political changes and advancements in maritime transportation (Frankopan 2017).
The U.S. New Silk Road Strategy aimed to boost trade-economic cooperation and liberalize trade in Central and South Asia.A key objective was to provide an economic uplift to Central Asian republics, which, despite abundant resources, remain among the least integrated regions globally (Zimmerman 2015).The strategy had the potential to spur economic growth in the region; however, implementation efforts significantly slowed down due to geopolitical circumstances and the pandemic.
Railway transportation is crucial for Central Asia's stability, as landlocked economies heavily rely on it for trade due to the absence of maritime access.The rail networks, being safer, costeffective, and less prone to delays, have immense potential to facilitate regional and international trade, contributing to the revival of the Silk Road (Karimova 2022).The existing rail network connects all economically essential areas and capitals, making railway transportation more critical than other modes (Kulipanova 2012).However, there is a shortage of research identifying relevant corridors for the revival of the Silk Road in Central Asian countries (Barisitz 2017).This article reviews corridors in Uzbekistan, Kazakhstan, Kyrgyzstan, Tajikistan, and Turkmenistan, pinpointing those with potential for further development, and emphasizing the corridors requiring attention for the Silk Road's revival and growth.
LITERATURE REVIEW
Central Asia's strategic importance has drawn global attention, leading to collaborative efforts to establish and enhance transport networks in the region.This literature review delves into the nuanced evolution of these networks, examining early collaborations, China's infrastructure development, European initiatives, challenges faced, and the modern global context, especially focusing on China's "One Belt, One Road" (OBOR) initiative.
Early Collaborations (1950s-1960s) According to a study by (Otsuka et al. 2017), the mid-20th century witnessed the initiation of efforts by the People's Republic of China (PRC) to link its railway system with the Soviet Union through Central Asia.PRC's emphasis on railway connectivity with Xinxiang in the western province marked a significant turning point, leading to the establishment of the Organization for the Cooperation of Railways in 2016 (Górski, 2016).This early collaboration laid the groundwork for regional cooperation and laid the foundation for future developments.
China's Infrastructure Development (1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992) A detailed analysis of China's infrastructure endeavors emerged from the study by Otsuka et al. (2017), detailing the construction of the railway line connecting the PRC and the Soviet Union.This study provided insights into the challenges faced and the eventual connection achieved in 1990, fostering regional integration.(Shu 1997) shed light on the broader vision of the "Europe Asia Land Bridge," underscoring the significance of the railway in facilitating transcontinental trade.
European Union's Initiatives (1990s)
The European Union's active role in the 1990s is highlighted through a comprehensive examination of the Transit Corridor Europe-Caucasus-Asia (TRACECA) in the study by Teimuraz Gorshkov (2001).This initiative, funded by the EU, aimed to improve road systems and establish overland transport corridors by contributing to the ambitious goal of connecting European and Central Asian train networks.The study emphasized the geopolitical implications of this initiative and its impact on regional dynamics.
Challenges and Obstacles (Late 1990s-2000s)
The slowing down of TRACECA due to the influence of the Eurasian Customs Union (Eurasec) is scrutinized in Lúcio Vinhas de Souza (2011).This study unraveled the geopolitical challenges posed by Eurasec, providing a comprehensive understanding of the obstacles faced by the ambitious European initiative.The persistence of advocates within China during this period was explored in (Shu 1997), demonstrating the ongoing efforts to promote the Eurasian land bridge despite regional challenges.
China's Strategic Vision and the "One Belt, One Road" Initiative (2013 onwards) A detailed exploration of President Xi Jinping's announcement of the Silk Route Economic Belt (SREB) and the 21st Century Maritime Silk Road (MSR) in 2013, as well as the subsequent "One Belt, One Road" (OBOR) initiative, was presented in a study by Foo, Lean, and Salim (2020).That study underscored China's long-term global power shift and the role of OBOR in fostering global trade and connectivity, with particular attention to the establishment of the Silk Road Fund.
Global Perspectives and Challenges
Wang and Wang (2022) introduced a global perspective by examining Japan's efforts to counterbalance China's influence in Southeast and Central Asia.This comparative analysis adds depth to the understanding of the geopolitical dynamics in the region, emphasizing the multidimensional aspects of infrastructure development.
The evolution of Central Asian transport networks reflects a complex interplay of historical collaborations, regional initiatives, and contemporary global strategies.From early partnerships to modern geopolitical dynamics, this in-depth analysis provides a nuanced understanding of the factors shaping the connectivity landscape in Central Asia.
Railway transportation in Central Asia
This research provides a comprehensive analysis of the role of railway transportation in shaping trade corridors within the Central Asian region.This section outlines the methodological approach employed in conducting the study.
Data Collection: The primary focus of this research is to utilize quantitative data to gain insights into the dynamics of railway transportation in Central Asia.
Literature Review: A thorough review of existing literature forms an integral part of the methodology, enabling the contextualization of the historical significance of railways in Central Asia.Scholarly articles, reports, and academic publications have been explored to understand the evolution of railway transportation, its impact on economic processes, and the historical context of the Great Silk Road.
Analysis Framework: Quantitative analysis is employed to evaluate and interpret the collected data, emphasizing the distribution of freight transportation across different rail routes and countries in Central Asia.The presentation of findings in a tabular format (Table 1) facilitates a clear representation of the freight volumes in domestic, export, import, and transit categories.
KAZAKHSTAN
The current density of railways is around 5.9 kilometers per 1000 sq.km, which shows a prominent lag in comparison with developed nations.For instance, it is 1.5 to 3 times lower than other countries such as India and Vietnam, and dozens of times to developed European countries.In recent times, the development and application of new technologies for the organization of container traffic have made it easier to enhance the average speed of container trains in the country (Carbajo and Sakatsume, 2004).KYRGYZSTAN Kyrgyzstan Railways is the exclusive operator for both freight and passenger transport in Kyrgyzstan.The main northern railway line spans from Dzhambul's Lugovaya station through Bishkek to Balykchy, transporting over 7 million tons of freight, including mineral fertilizers, oil products, and metals.International railways are primarily used for transporting goods like industrial raw materials, minerals, building materials, lubricants, and fuels (ESCAP 2022).Domestic rail freight is minimal; however, it is limited to small amounts of sugar and coal.
Road transportation dominates the movement of goods within the country.
TAJIKISTAN
In the country, Gorno Badakhshan lacks rail tracks, and the main railway spans 617.5 kilometers.Of this, 62.1 kilometers are double track, and 555.7 kilometers are single track.Transit rail traffic constitutes about two-thirds of overall rail freight, mostly through the northern section, particularly in the cities of Khuzhand and Kanibadam (Max Ee and Eshonov Boymurod 2009).Tajikistan transports critical products like cement, wheat, cotton, and building materials by rail.The east and north lines are unconnected, requiring passage through enclaves.Building a railway to connect these lines would be a costly project (World Bank 2011).
TURKMENISTAN
Turkmen Railways is accountable for operating the railway system and belongs to the Ministry of Railways.Currently, efforts are being made by the country to expand its railway network and cover around 5,256 kilometers more by 2025.In 2020, the country had around 6,561 kilometers of rail lines close to the southern and northern borders (Gao 2016).It is important to note that the Tejen-Sarahs-Mashdad railway, established in 1966 by Iran and Turkmenistan, links European, Russian, and Central Asian rail systems with the Persian Gulf, South Asia, and Turkey.
Impact of the pandemic on transportation in Uzbekistan
The global pandemic significantly impacted transportation sectors worldwide.In Uzbekistan, rail passenger transport decreased by 81% in Q1 2020, and air carriers saw an additional 83.4% decline.Quarantine measures did not affect traffic but sharply reduced road transport passenger traffic (Sitora Primova 2020).Kuryk Port currently collaborates exclusively with the Azerbaijani port of Alat (Hoh 2019).While there are intentions to expand ferry services with Iran and Turkmenistan, progress remains pending.Improving the corridor is crucial for enhancing collaboration with other ports and contributing to the Silk Road's revival.Another vital corridor is the Kashgar -Irkeshtam -Osh -Andijan -Tashkent highway, initially discussed in the late 20th century and postponed for economic and political reasons (Sahakyan 2022).Since its launch, this highway has facilitated the transportation of goods.The Silk Road International venture between China and Uzbekistan has similarly aided the transportation of goods along the China-Kyrgyzstan-Uzbekistan-Afghanistan transport corridor.
Uzbekistan plays a crucial role in shaping the region's image, mainly through the China-Kyrgyzstan-Uzbekistan railway corridor.This corridor not only connects the region and its cities with the outside world but also contributes significantly to reviving the Silk Road.The project has the potential to alter the geoeconomics of the region, reinforcing Uzbekistan's position as a regional transportation hub.Additionally, it fosters closer ties between Uzbekistan, Kyrgyzstan, and China, facilitating enhanced interaction.
Central Asian economic corridor -challenges and opportunities
As part of the Belt and Road Initiative (BRI) and the Central Asian Economic Corridor (CAEC), the BRI seeks to enhance trade and transportation between Europe and China via the Middle East and Central Asia.Among the six BRI economic corridors, the China-Arab Economic Corridor (CAEC) is considered the most complex due to its involvement with most countries.The CAEC has faced new challenges amid the COVID-19 pandemic, including the "Cold War mentality" and anti-globalization movements.This study suggests that China should collaborate with countries along the route to improve the economic structure, establish the Digital Silk Road, promote economic corridor growth, and build a community with a shared vision for the future.
Countries in the CAEC have a unique opportunity for economic growth by leveraging the corridor to facilitate increased trade and investment within the region and with neighboring countries.The CAEC is expected to play a significant role in enhancing regional infrastructure, a crucial factor for economic development (Jiang et al. 2016).Improving the governance system is another crucial step, requiring standardized rules and norms for trade and business, fostering collaboration among countries, and creating a legal framework that encourages investment and protects intellectual property rights.
The CAEC provides member countries with a distinctive chance to promote economic growth and progress.However, addressing existing challenges is imperative for effective corridor utilization.Collaborative efforts are needed to establish a better regulatory environment, formulate a shared vision for the CAEC, develop a detailed growth plan, and enhance the overall business climate in the region.Joint efforts among participating countries are essential to overcome challenges and capitalize on opportunities.
CONCLUSION AND RECOMMENDATION
Central Asia can revitalize the historic Silk Road by establishing crucial transportation networks fostering enhanced trade with other countries through economic integration.However, improving the existence of transit lines is essential for these routes to meet their objectives.Cooperation, coordination, and resource pooling among key players in the region are vital.Critical thoroughfares, such as the Kazakhstan Terminal at Lianyungang Port, the Kashgar-Irkeshtam-Osh-Andijon-Tashkent Corridor, the China-Kyrgyzstan-Uzbekistan railroad, and the Kuryk Port, could significantly contribute to regional economic growth.Overcoming challenges like red tape, customs processes, and links to nearby cities requires international collaboration.Collaborations between the Kuryk and Alat terminals in Azerbaijan are underway, while progress with ports in Iran and Turkmenistan is still pending.The China-Kyrgyzstan-Uzbekistan railway has the potential to be a key artery connecting China to Europe and the Middle East, but challenges like global competitiveness and corruption need addressing for successful completion.
For the efficient operation of transportation networks, regional leaders must exhibit political will and agree to connect economies and address historical tensions among Central Asian governments.
Close collaboration with organizations like the Eurasian Economic Union (EAEU), Shanghai Cooperation Organization (SCO), and Conference on Interaction and Confidence-Building Measures in Asia (CICA) is expected from China, Russia, and Turkey, given their strong ties to the region.China's Belt and Road Initiative can provide both energy and funds, but tackling challenges like corruption and local ownership remains crucial.Collaboration among Central Asian states is essential not only for transportation but also for energy, agriculture, tourism, internet access, and crosscultural interaction.Overcoming historical animosities rooted in complex factors is necessary for a comprehensive alliance, unlocking economic potential and fully reestablishing Silk Road ties.A deeper exploration of these historical tensions can reveal the challenges to achieving a more integrated and collaborative Central Asia.
Given its strategic location, growing economic influence, and stable political atmosphere, Uzbekistan can play a crucial role in connecting various sectors, addressing challenges, and navigating historical tensions for a more cooperative regional environment.Moving forward, Central Asian governments should overcome past animosities and recognize each other as true partners in a mutually beneficial joint endeavor.This approach positions Central Asia as a vital node for international commerce, facilitating transactions and fostering ties among significant countries such China, India, and those in Europe.Revitalizing and integrating both physical and soft networks can pave the way for a more interconnected and prosperous Central Asia globally.However, it is crucial to acknowledge certain limitations that may pose challenges to the successful implementation of the recommended strategies.One significant limitation lies in the potential geopolitical tensions and historical animosities among Central Asian governments.Despite the call for political will and regional cooperation, deeply rooted historical tensions may hinder the seamless connectivity of economies and the establishment of effective transportation networks.Addressing these historical animosities requires a nuanced and delicate approach, and the process may encounter resistance from various stakeholders.
Another limitation worth considering is the inherent complexities associated with international collaboration, particularly in addressing challenges such as red tape, customs processes, and corruption.The proposed collaborations with countries such as Azerbaijan, Iran, and Turkmenistan are positive steps, but the varying political and bureaucratic landscapes in these nations may introduce delays and obstacles to the smooth execution of joint initiatives.It is imperative for the stakeholders involved to navigate these limitations effectively and adopt flexible strategies that account for the diverse challenges inherent in the geopolitical, economic, and cultural landscape of Central Asia.By recognizing and proactively addressing these limitations, the region can maximize its potential and truly re-establish the Silk Road as a vibrant and interconnected economic corridor.
Cargo transportation, however, was affected with a 35.5% decrease in Q1 2020 due to restrictions.Public transport in Tashkent closed in March 2020, and all automobile trips were restricted in April.The substantial losses resulted in a 1.5-2.5% GDP decline, as reported by the Central Bank of Uzbekistan (Tursunbaevich, Bulturbayevich, and Rahmat 2021).The transportation sector in Uzbekistan faced severe consequences, leading to increased unemployment, with a 35-57% decrease in employment in 2021.Transport corridors and opportunities for Silk Road Picture 2. Transport corridors and opportunities for Silk Road Note: Connecting CAREC: A Corridor Network Source: CAREC Secretariat (View full-sized map) The following are the economic corridors in Central Asia, and the most important ones -the Kazakh Terminal, Kuryk's Port, and Uzbekistanwill be discussed in this section: • Kazakhstan's Terminal in Lianyungang • Kuryk Port • Kashgar -Irkeshtam -Osh -Andijan -Tashkent Highway • China-Kyrgyzstan-Uzbekistan Railway • China-Kyrgyzstan-Tajikistan-Afghanistan-Iran RailwayRailway • Lapis -Lazuli Transport Corridor • Mazar-e-Sharif -Herat Railway • Uzbekistan -Turkmenistan -Iran -Oman -India Corridor The Kazakh Terminal in Lianyungang Port is a vital transportation corridor in Central Asia (Peter Golden 2011).It could serve as a crucial link in Kazakhstan's logistics chain and facilitate the movement of goods to and from Central Asia.However, extensive collaboration is necessary among not only Kazakhstan but also other Central Asian nations.Implementing preferences for Central Asian countries at the port and streamlining customs procedures at specific locations like Dostyk and Khorgos can reduce bureaucratic barriers and save time.The terminal faces stiff competition from ports and corridors in inaugurated in 2016 as a railway ferry terminal, plays a crucial role in enhancing the Silk Road and trans-Caspian multimodal transport (Indeo 2018).With a capacity of 4 million tons, the port facilitates efficient logistics for cargo transshipment, including chemicals, metal products, consumer goods, and oil products.This trans-Caspian route serves as a key link for exports and imports between the Ural-Siberian region of Russia and China.The completion of the Beineu-Akzhigit highway reconstruction by the end of 2019 significantly increased the route's capacity, fostering road transit through Kuryk Port and creating a new connection between Uzbekistan and the Trans-Caspian corridor (Shaikova, Dronzina, and Zholdasbekova 2023).
Table 1 :
Freight Transportation through Railways in Central Asia (in Million Tonnes), 2021 Limitations:Recognizing the study's limitations is essential.Using data from 2021 may not reflect recent changes.Additionally, this quantitative analysis mostly needs to consider Uzbekistan.Ziyoda specializes in railway transport, logistics, and the development of technologies for the transportation of goods by rail.She has developed various tools and technologies for organizing cargo transportation.She is the head of the state-applied grant in Uzbekistan.In 2023, she worked as an expert on the development of transportation corridors at the Asian Development Bank, Uzbekistan.She also developed two instructions for extending the service life of rail service cars in Uzbekistan.Since 2017, psychology, and the study of issues of the psychology of human resources in management.She is the author of 5 textbooks and, four manuals, two monographs and has published more than 75 scientific articles.Her research is aimed at studying current issues of psychological characteristics of personnel training and processes in the organization of professional activities.Associate Professor Gavkhar Fuzailova is a lecturer at the Psychology of Religion and Pedagogy Department of the International Islamic Academy of Uzbekistan and specializes in the methodology of teaching humanities, the study of modernization of contemporary education, and the improvement of pedagogical activity.She is the author of 4 textbooks and eight manuals | 4,417 | 2024-08-03T00:00:00.000 | [
"Geography",
"Political Science",
"Engineering"
] |
Nonlinear neural patterns are revealed in high frequency functional near infrared spectroscopy analysis
Functional Near Infrared Spectroscopy (fNIRS) is a useful tool for measuring hemoglobin concentration. Linear theory of the hemodynamic response function supports low frequency analysis ( < 0.2 Hz). However, we hypothesized that nonlinearities, arising from the complex neurovascular interactions sustaining vasomotor tone, may be revealed in higher frequency components of fNIRS signals. To test this hypothesis, we simulated nonlinear hemodynamic models to explore how blood flow autoregulation changes may alter evoked neuro-vascular signals in high frequencies. Next, we analyzed experimental fNIRS data to compare neural representations between fast (0.2 – 0.6 Hz) and slow ( < 0.2 Hz) waves, demonstrating that only nonlinear representations quantified by sample entropy are distinct between these frequency bands. Finally, we performed group-level distance correlation analysis to show that the cortical distribution of activity is independent only in the nonlinear analysis of fast and slow waves. Our study highlights the importance of analyzing nonlinear higher frequency effects seen in fNIRS for a comprehensive analysis of cortical neurovascular activity. Furthermore, it motivates further exploration of the nonlinear dynamics driving regional blood flow and hemoglobin concentrations
Introduction
Functional near-infrared spectroscopy (fNIRS) is a non-invasive method used to measure brain tissue hemodynamics as a proxy for neural activity (Strangman et al., 2002b).Slow hemodynamic response signals in fNIRS oscillations ( < 0.2 Hz) are well-suited for linear (i.e., spectral) analysis, particularly when the time intervals between stimuli are long (Glover, 1999).However, such a conservative cut-off frequency is aimed at avoiding overlap with frequency components in vascular dynamics that are independent of brain function, such as cardiac pulsatility or respiratory waves (Pinti et al., 2019).High frequency signals are often filtered out because they are assumed to correspond to systemic physiological or instrumentation noise (Huppert et al., 2009).However, disentangling fNIRS artifacts from systemic brain-heart and, more in general, brain-body interaction effects may be difficult (Vikner et al., 2021;Candia-Rivera et al., 2022a,b).
Recent studies suggest that high frequency ( > 0.2 Hz) signals in fNIRS may provide additional information about neurovascular activity (Yücel et al., 2021;Santosa et al., 2018;Ghouse et al., 2020).Indeed, the nonlinear nature of the autonomic nervous system (Goldberger et al., 2002;Marmarelis, 2004;Sunagawa et al., 1998;Barbieri et al., 2017) may impact the hemodynamic response in higher frequencies.Recent research by (Ghouse et al., 2020) assessed entropy estimates of fNIRS signals that contained an upper bound of 0.6 Hz and demonstrated complementary areas of activity when compared to neural correlates observed using linear analysis of fNIRS.This finding supports recent suggestions to use cutoff filters as high as 0.5 Hz (Yücel et al., 2021).Some studies even suggest minimal fNIRS preprocessing, avoiding filtering and instead opting for robust statistics (Santosa et al., 2018).These developments underscore the importance of investigating the potential contribution of higher frequency signals in fNIRS analysis to improve on our understanding of neural and neurovascular activity.
More specifically, to investigate the potential contribution of nonlinearities in higher frequency of fNIRS signals, this study assesses differences between neural representations of fNIRS in the traditional slow wave ( < 0.2 Hz) and the proposed fast wave (0.2-0.6 Hz) frequencies using nonlinear information-theoretic methods.We hypothesize that to fully characterize the cognitive phenomena reflected in fNIRS signals, we need to assess its full spectrum with additional nonlinear analysis.This study uses sample entropy (SampEn) (Richman and Moorman, 2000) to assess fNIRS signals irregularity (and so predictability) based on embeddings of the fNIRS signals (Sauer et al., 1991).
First, we simulated plausible modulations of autoregulatory feedback on blood flow control with stochastic dynamics in the hemodynamic model (Friston et al., 2000) to motivate fast wave nonlinearity analysis.Then, using a mental arithmetic paradigm and a control motor imagery paradigm (Berntson et al., 1996), we investigated the additional information that high frequency components may provide in experimental data when a subject is under cognitive stress.We performed a representational similarity analysis (RSA) (Kriegeskorte et al., 2008;Carlin et al., 2011) on data obtained from a public database (Shin et al., 2017), comparing fast and slow wave fNIRS modalities.Our hypothesis was that fast wave activity would add significant complementary regions, particularly in nonlinear analysis, and that the correlations between fast and slow wave fNIRS activity would be lower during mental arithmetic compared to motor imagery due to regional nonlinear interactions.We assessed multivariate correlation at each detector and hypothesized that mental arithmetic topoplots of activity would be less correlated between fast and slow wave fNIRS.
Simulations
For initial validation of plausible effects of autoregulation dynamics on hemoglobin, we first simulate a mechanistic balloon model of hemodynamic activity (Cui et al., 2010;Friston et al., 2000).First, a flow inducing signal (s), i.e. the response of the vascular system to a neural metabolic demand, is linearly described for the sustenance of incoming blood flow (f in ).Particularly: u is the control signal which is the neural signal generating the flow.ϵ is the efficacy by which the neural signal can sustain a flow-inducing signal, i.e. a response that may dilate vessels to increase blood flow into.This is related to the vascular resistance (r), whose dynamics are ṙ = − r 2 s (Friston et al., 2000).τ s is the time constant for this flow-inducing signal and τ f is the time constant describing the effects of autoregulatory feedback from blood flow.Considering τ f comprises information on autoregulatory effects of blood flow (the time constant for returning back to baseline), this is the parameter we later model to assess its nonlinear effects on hemoglobin concentrations.
From the sustaining blood flow generating signal, a so-called "balloon model" describes the dynamics of the volume of a blood vessel (the balloon), and the permeation of hemoglobin inside and outside the vessel (Buxton et al., 1998).Explicitly, the rate of decay of blood volume is related to blood flowing in and blood flowing out of a vessel: (2) α describes the capacity at which a balloon can expel water, having been distended by the surge of inflow.τ o is the time constant which governs the rate of change of the volume, which is similarly intertwined with the rate of change of deoxyhemoglobin in the venous compartment (q): (3) The function E describes the oxygen extraction coefficient, or efficacy of the tissue in extracting oxygen from the incoming blood, while E o is the resting oxygen extraction coefficient.In (Cui et al., 2010), an extension was proposed to relate the blood volume and total hemoglobin concentration as: Then, oxyhemoglobin is merely o = p − q.
We reiterate the observation in eq. ( 1) that the parameter τ f is particularly related to the autoregulatory feedback.Exactly how its value relates to the true autoregulatory changes is little understood, though literature states that responses to autoregulatory changes occur over a period of 1-2 min (or less than 0.02 Hz) (Lemkuil et al., 2013).To maintain this expected periodicity, while allowing random deviations due to uncertain dynamics, we designed a 2nd order stochastic dynamical model whose power spectral density (PSD) on average has a peak at the frequency of hypothesized autoregulatory changes.
The steady state value of τ f is at 0.8 (normalized units), while the steady state of its change is 0. A is a parameter that modulates how fast it returns to steady state, and the diffusion term is random variations outside the potential normal oscillatory behavior of τ f .When there is no autoregulatory activity, it quickly returns to steady state, with A= 10.When there is autoregulatory activity, it is more free to change with A = .1.black Eq. ( 5) was devised to model the realizations of τ f through simple, non-trivial stochastic differential equations.Such a generative model embeds interpretable 2nd order linear dynamics showing emergent nonlinar oscillatory properties in the observed variable.blackWe integrate the stochastic differential equations with the Euler-Maruyama method (Platen and Bruti-Liberati, 2010).
We numerically simulated 100 τ f time courses to illustrate expected spectral properties of the stochastic dynamics.This was achieved by obtaining an estimate of the power spectral density (PSD) using the Welch method (Welch, 1967).The units of the PSD for τ f is in arbitrary units, considering its the time constant of a normalized blood flow signal seen in eq. ( 1).We simulated a block design experiment using these generated τ f time courses, where each block is 40 s long with a minute long rest, and 5 stimulus blocks would either induce τ f modulations (i.e. where A = 0.1) or 5 stimulus blocks would not induce the modulations.The final simulated hemoglobin concentration time courses are corrupted with a signal-to-noise (SNR) ratio of 0 dB.
Experiment data
A publicly available dataset was used to obtain fNIRS signals with the desired experimental protocol for this study, as reported in (Shin et al., 2017).In summary, the experiment recruited twenty-nine healthy subjects (aged 28.5 ± 3.7), fifteen of which were female and fourteen male.Of these twenty-nine healthy subjects, all were right-handed expect for one.As according to Shin et al. (2017), all participants were free from neurological, psychiatric, or any brain-related disorders.They were fully informed about the experimental process, and written consent was obtained from each volunteer.After the experiment, they received financial compensation.The study adhered to the guidelines of the declaration of Helsinki and received approval from the Ethics Committee of the Institute of Psychology and Ergonomics, Technical University of Berlin (approval number: SH_01_20150330).Three trials were performed with ten repetitions of mental arithmetic and baseline events, and three trials were performed with ten repetitions of right-hand and A. Ghouse et al. left-hand motor imagery for each subject.We note that the motor imagery tasks and mental arithmetic tasks were done independently, not concurrently.Thirty-six fNIRS series were acquired for each subject with 10 Hz sampling rate.
The experiment design had sixty seconds of resting state to start data acquisition from a subject, after which an instruction was shown on the screen indicating which task was to be performed-either an arithmetic problem, a " − " for a baseline, or a left or right arrow for motor imagery.The subject performed the indicated task for ten seconds, with a subsequent fifteen second resting phase before the next instruction.After twenty repetitions of these instructions and tasks (ten repetitions for each task in an experiment run), a sixty second rest was performed.A total of three trials were performed, for a total of thirty repetitions per event.
fNIRS signals
Thirty six channels of optical densities (OD) were resolved from source detector pairs comprising 760 nm and 850 nm wavelengths at distances of 3 cm covering the frontal, lateral parietal and posterior cortical regions as seen in Fig. 1a.The modified Beer Lambert law was used to convert the ODs to deoxyhemoglobin (Hb) and oxyhemoglobin (HbO) (Strangman et al., 2002a).
Fig. 2 illustrates the preprocessing pipeline.After applying the modified Beer-Lambert law to resolve the 36 channels seen in Fig. 2b from the source-detector pairings in Fig. 2a, band-pass frequency filters were applied to extract traditional hemodynamic bands ( < 0.2 Hz) (Strangman et al., 2002a;Pinti et al., 2019) or the proposed increased hemodynamic band (0.2-0.6 Hz).A wavelet filtering approach using a Daubechies 5 wavelet, nine level decomposition was used to further reduce instrumentation noise such as movement in the oxy-and deoxyhemoglobin signals (Molavi and Dumont, 2012).Detrending and prewhitening with an AR(1) model was then performed to remove temporally structured noise in the signal (Huppert, 2016).The signals were separated into epochs of 30 s (comprising the 15 s task phase, and the 15 s rest phase), with each channel at each activity block being referenced to the mean of the previous 5 s.We exclude the 2 second instruction phase before the task-phase from the analysis to reduce potential confabulation of task-evoked responses from processing the instructions.Total hemoglobin was computed as the addition of both Hb and HbO.The total hemoglobin is important considering the integration of the concentrations may present unique temporal dynamics revealing distinct nonlinear temporal effects.For all three signals, entropies and mean estimates were extracted then passed into the 1st and 2nd level analysis.Before second level analysis, the results were spatially smoothed to increase the sensitivity of random effects from detector locations (Tak et al., 2016).A Gaussian smoothing kernel with a full-width-half-max of 1.5 cm (half the source-detector separation distance) was used as seen in Fig. 1b.
Entropy analysis
To analyze nonlinearity and regularity of the fNIRS signals, we exploited sample entropy (SampEn) (Richman and Moorman, 2000).SampEn is a method to calculate the entropy of a dynamical system in its phase space.In other words, it assesses how much information it takes to characterize the dynamics of a system.It shares similarities with methods like approximate entropy (ApEn) (Pincus, 1991), except the SampEn algorithm alleviates biases that are introduced in ApEn by not considering self-matches in its calculation of the correlation integral used in the definition of entropy in dynamical systems (Delgado-Bonal and Marshak, 2019b;Sinai, 1959).There are a whole slew of other entropy estimation methods that could have been used instead, however we have previously performed a study comparing a whole battery of entropy estimation methods for assessing fNIRS signals, with the conclusion that SampEn provides similar results as the other entropy estimation methods (Ghouse et al., 2020).
A delay-time τ and embedding dimension m are needed to reconstruct manifolds using delay-coordinates.τ was selected as the first zero of the autocorrelation, while m was found using the false nearest neighbors approach (Abarbanel et al., 1993).
For calculating SampEn, radius R = 0.2 × σ x was used as the threshold to determine whether states were neighbors, where σ x is the standard deviation of the fNIRS time series (Delgado-Bonal and Marshak, 2019a).The particular equation describing how to calculate SampEn is: X in this equation is a state space reconstructed using a time series.The superscript denotes the embedding dimension while the subscript denotes the state index, for which there exists N states.
Representational similarity analysis
Representational similarity analysis is a method used to compare multivariate data such as different brain data types (Kriegeskorte et al., 2008).This paper calculates similarities between tasks using the distance correlation in order to construct the representational similarity matrices for fast or slow wave (Székely et al., 2007;Geerligs et al., 2016).
In brief, given random vector X and random vector Y with dimensions R p and R q respectively, distance correlation evaluates independence by integrating the distance between the random vectors in conjunction with a weighting function w(t, s) = (c p c q ⃒ ⃒ ⃒t| Γ( 1+d2 ) corresponds to half the surface area of a unit sphere in the given dimensionality d.This leads to the following statistic: As the dimensions of the random vectors, p and q, approach infinity, the statistic converges to a Student's t-distribution and can be approximated as such for hypothesis testing (Székely and Rizzo, 2013).Given a ij = |x i − x j | and b ij = |y i − y j |, where i and j represent the ith and jth observations of x or y, the sample covariance can be estimated as: In our application, random vectors X and Y have equivalent dimensionality R 3 as we are looking at multivariate correlations between oxy, deoxy and total hemoglobin.Furthermore, due to there being 4 experimental conditions (baseline, mental arithmetic, left-and right-hand imagery), the resulting representational dissimilarity matrix (RDM) is a 4 × 4 matrix.Representational dissimilarity outputs were obtained for each detector location for each measure (entropies or mean value).Extracting the upper triangle, we have two 29 matrices for slow and fast wave fNIRS respectively to perform statistical analysis upon.
Fast vs. slow wave spatial analysis
For each subject, the data obtained from each detector had a shape of N repetitions × N concentrations.By calculating the median over the N repetitions for either fast or slow wave fNIRS, we obtain a vector of size N Concentrations x N detectors for each subject.This analysis enables us to determine whether the spatial fNIRS arrays for fast and slow waves are dependent, or whether the null hypothesis that they are independent can be rejected.We expect that the slow and fast wave analysis will become more independent when there is a task inducing changes in autoregulatory activity.We also perform a Wilcoxon paired analysis at the detector level to compare which medians are significantly different in the group of subjects for each concentration.
Statistical analysis
In order to perform statistical analysis of the data, we exploited the Scipy statistics package in Python (Virtanen et al., 2020).
Fig. 2.
Overview of the analysis pipeline used for each fNIRS signal in the dataset, which includes computation of hemoglobin concentrations, band pass filtering, motion artifact correction, removal of linear trends and serial correlations, total hemoglobin calculation, state space reconstructions using delay-coordinate embeddings, estimation of sample entropy or mean value for each trial, first level analysis using median value over trials, and spatial smoothing to improve sensitivity in group analysis results (Tak et al., 2016).
A. Ghouse et al.To compare RDMs between fast and slow wave fNIRS, a group analysis was conducted using a Friedman test, with a Bonferroni correction to account for multiple comparisons.A family-wise p-value of less than 0.05 was considered significant.Nonparametric Friedman statistical tests were used to avoid assumptions of normality in the dataset when performing intergroup analysis (Friedman, 1937).Then, the Bonferroni correction was applied to be as conservative as possible when correcting for family-wise error rates such that any inferences are robust to Type I errors (Tukey, 1953;Dunn, 1961).
For median analysis, rainclouds were generated from standardized Z transformations to compare the shapes of distributions between fast and slow wave analysis.A Kolmogorov-Smirnov test was applied to compare concentrations analyzed from fast and slow wave analysis, correcting for multiple comparisons.A paired sample Wilcoxon signed rank test was performed on the difference of median value for each concentration between mental arithmetic vs baseline or right-hand vs left-hand motor imagery.Detectors were retained for further analysis if any of the 3 concentrations returned significant with a false alarm rate of α = 0.05∕6 = 0.008.For remaining detectors, distance correlation was performed between fast and slow wave fNIRS.
Results
The first set of results are from simulations of neurovascular signals generated by a hemodynamic model, evoked from events that either induce or do not induce autoregulatory changes.The second set of results are from group statistics of representational similarity analysis between fast wave and slow wave cortical distribution of activity, and median analysis of cortical activity.Standard fNIRS analysis refers to time averaged value of fNIRS activity during a task, while nonlinear fNIRS analysis refers to SampEn during the time duration of the task.
Simulations
Fig. 3 illustrates the results of realizations of the hypothesized autoregulatory feedback changes as according to eq. ( 5).The median peak frequency was at 0.02 Hz, corresponding to periodicity of 50 s, with periodicity ranges from 12 ss to 100 s.Fig. 4 illustrates the effects that these autoregulatory changes may have on the hemodynamics, where HbO is the oxyhemoglobin and HbR is the deoxyhemoglobin.Over the 100 realizations of the block design experiment, the mean entropy of the hemodynamic response in the high frequency (0.2-0.6 Hz) regime without autoregulation modulations was at 0.212 (±0.015) bits compared to 0.163 (±0.011) bits when the modulations occur; through a paired Wilcoxon signed rank test, the SampEns were found to be significantly different (p ⋘ 0.001).On the other hand, the power of this high frequency band without autoregulation modulation was found to be 0.0368 (±0.00441) as compared to 0.0372 (±0.00385) during autoregulation modulation; the power bands were not significantly different according to the Wilcoxon signed rank test (p = 0.46).
Representational similarity analysis
Results from the Friedman analysis comparing the upper triangle of the RDMs constructed using either fast wave or slow wave fNIRS signals according to the methods described in section 2.5 can be seen in Fig. 5, for mean and SampEn respectively.For the mean value, no detector has significantly different representations between slow and fast wave fNIRS, whereas each quadrant of the sensor space on the cortex contained significantly different detectors for SampEn.Generally, SampEn contained higher Friedman test scores than the mean analysis.
Fast vs slow wave spatial analysis
Using the methods described in 2.6, we obtained group level median analysis results for fast and slow wave fNIRS.To reiterate, this entailed calculating the median value of the mean or SampEn signal over the repetition of the task performed by the participant, and then performing a contrast analysis between the effects of mental arithmetic (MA) and baselines (BL), or left-hand motor imagery (LH0 and right-hand motor imagery (RH).A distillation of the spatial results can be seen in the raincloud Fig. 6 to visualize distributions of contrasts between the task comparisons across subjects; Kolmogorov-Smirnov tests were performed to ascertain whether the distributions were significantly different.Supplementary fig.S1 shows the spatial cortical map of contrasts between mental arithmetic (MA) and baseline (BL), or right-(RH) or lefthand (LH) motor imagery tasks.
From the raincloud plots in Fig. 6, distributions for mean results consistently appear to be overlapped when comparing slow and fast wave different in medians, whereas SampEn demonstrates a flatter distribution for slow wave analysis as compared to fast wave analysis.Motor imagery also presents a flatter distribution in SampEn when using slow wave analysis compared to fast wave analysis, however this greater variance is similarly observed in mean value analysis.Performing a Kolmogorov-Smirnov two sample test, we found that the distribution of fast and slow wave analysis was significantly different between SampEn Fig. 3. Spectral analysis of realizations of τ f simulated from the stochastic differential equation in eq. ( 5).a) Represents the average power spectral density (PSD) of τ f as estimated by the Welch method over 100 realizations.The units of the PSD are arbitrary as τ f is the time constant of f in seen in eq. ( 1).b) on the other hand is a histogram of the peak frequency of the PSD of τ f over the 100 realizations.in oxy and deoxyhemoglobin, while in no mean analysis were they significantly different.
For mental arithmetic and baseline Wilcoxon paired analysis, mean estimates show significant detectors only in slow wave fNIRS, with a cluster in the right lateral cortical areas and detectors bilaterally in the frontal cortex.SampEn mainly determined significant detectors in the frontal right cortex and left parietal cortex in the slow wave analysis, while fast wave analysis complemented information in the right parietal cortical areas.
Left-and right-hand motor imagery Wilcoxon paired analysis demonstrated activity bilaterally in the parietal cortical areas with both SampEn and mean estimates.Both mean and SampEn demonstrated slow activity predominantly in the right parietal cortex, while fast activity was found in the left lateral parietal cortex.Mean also contained significant detectors in the medial left parietal cortex in slow wave analysis and in the occipital regions for fast wave analysis.
As a sanity check, we also compared the neural representations between slow frequency bands (0 Hz, 0.2 Hz) and cardiac frequency bands (0.8 Hz, 3 Hz).The distance correlation of topolots of fast wave vs slow wave was not significant, indicating that the fast wave fNIRS analysis in freqs (0.2 Hz, 0.6 Hz) was not a result of systemic physiological confounders (see also supplementary fig.S2).
Discussion and conclusion
In this study, we compared neural activity representations between fast wave (0.2-0.6 Hz) fNIRS and the standard slow wave ( < 0.2 Hz) fNIRS signal using both linear measures of mean value and nonlinear measures derived from sample entropy.The fNIRS signals were obtained from a publicly available dataset of 29 subjects performing mental arithmetic and left-/right-hand motor imagery tasks (Shin et al., 2017).Our hypothesis was that mental arithmetic tasks, which induce cognitive stress, would provide more variation in areas of cortical activation than motor imagery when comparing slow and fast wave fNIRS using nonlinear analysis.This hypothesis was motivated by literature suggesting the nonlinear modulation of fNIRS dynamics sustained by various factors, including those induced by autonomic nervous system activity (Friston et al., 2000;Gianaros et al., 2012;Sheng and Zhu, 2018;Candia-Rivera et al., 2023).Hemodynamic models are also known to be nonlinear in nature (Friston et al., 2000;Buxton et al., 1998;Friston, 2001).
To investigate the potential effects of vasomotor property dynamics on autoregulation modulations, we first conducted a simulation study using equations ( 1) and ( 5).Our theoretical model for modulations of vasomotor tone closely matched the expected spectral profile of realworld vasomotor tone variations at frequencies below 0.2 Hz, as shown in Fig. 3. Additionally, we found that autoregulation variations in flow signals that propagate hemoglobin concentration dynamics (as depicted in Fig. 4) reduce entropy in the observed signal even with an SNR of 0 dB.This suggests that with proper statistical power, these autoregulatory dynamics could potentially be observed in real-world data using nonlinear analysis methods.It is important to note that our empirical data analysis should not be interpreted as evidence that high frequency signal entropy is a measure of autoregulation.Rather, we hypothesize that the nonlinear activity observed in high frequency fNIRS provides complementary information that is crucial for a comprehensive characterization of cognitive states, such as those associated with stress processing during mental arithmetic tasks.
To explore potential distinctions in neural signal representations between fast and slow waves in fNIRS data, we conducted a Friedman test to compare results on the tasks of baseline, mental arithmetic, and motor imagery, as shown in Fig. 5.The mean value analysis did not reveal any significant differences in task combinations, as expected since the mean corresponds to the DC value of the signal and high pass filtering only affects attenuation of the DC value.In contrast, nonlinar SampEn analysis demonstrated significant differences between fast and slow wave fNIRS detectors in all four quadrants of the sensor space covering cortex, with a higher Friedman score across the cortex than the mean analysis.While it is unclear how filtering affects state space regularity, high pass filtering may result in lower mean entropy and higher variance than low-pass filtering (Borges et al., 2020).This effect may correspond to the significantly different SampEn representational Upon examining motor imagery results we observed changes in laterality in SampEn analysis depending on the frequency band of focus, with fast wave analysis being more sensitive to left hemisphere activity and slow wave analysis being more sensitive to right hemisphere activity (see Supplementary fig.S1).A similar phenomenon was seen in the mean analysis, although an additional medial left hemispheric detector was observed for slow waves.We expect this activity to correspond to the somatosensory cortex, located in the central cortical regions, where the hands of the cortical homunculus are located between the lateral and medial portions of the cortex (Grodd et al., 2001).By considering both frequency bands using SampEn analysis, we may have discovered its bilateral effects.This bilateral difference between left-and right-hand motor imagery was only found using slow wave analysis with just the mean analysis.It is important to mention that the handedness of a participant plays a critical role in the effect seen in motor imagery.Right-handed individuals appear to have a stronger lateralization effect specific to the side of the hand imagined; left-handed individuals on the other hand have bilateral responses (Crotti et al., 2022).Considering the set of participants used in this study were primarily right-handed, the results presented in this study may be specific to this group.Nonetheless, considering we do paired analysis of the same participants performing left-and right-hand motor imagery, the resulting differences of cortical signal effects observed when comparing fast and slow wave analyses appear to indeed be a neurovascular result.This suggests that to fully describe the complementary areas of activity that entropy provides, we must consider the full frequency profile of potential hemodynamic activity.
We also conducted a distance correlation analysis of the difference of median maps and found that all mean maps were significantly correlated, as expected.The cortical representations reflected in median difference maps between tasks should be similar if we scale the signal by merely a DC value, as previously discussed.However, our hypothesis was that mental arithmetic should not be significantly correlated in entropy analysis while motor imagery should be significant.Instead, we found that neither was significant, although mental arithmetic vs. baseline had a lower correlation than the motor imagery distinction.This may imply that the nonlinear irregular effects induced by vasomotor mechanical dynamics are always occurring, although the degree to which they affect the resulting neural signal varies depending on the task.It's known that a significant portion of cerebral vascular resistance comes from vasomotor control in arterioles (Mandeville et al., 1999).The nonlinear transform of the constant dynamics of vasomotor control to hemoglobin concentrations may always be present, yet their scale is modulated by a stress task such as mental arithmetic, making their high frequency contributions to hemoglobin concentrations more irregular in the case of mental arithmetic.We finally remark that the proposed fast wave analysis within (0.2 Hz, 0.6 Hz) is not affected by non-cognitive phenomena associated with oscillations in the (0.8 Hz, 3 Hz) band (Hoshi, 2016).
We recognize that this study may have its limitations.First, we did not perform simultaneous fNIRS and autonomic signals analysis.Respiration and heart rate, typically occurring within the fast-wave band, could be confounding factors in our data analysis since a multitude of regulatory processes may change during cognitive tasks.However, the heterogeneity in the response across the cortex suggests that such oscillations may not systematically affect fNIRS series all over the scalp.Literature suggests possible differential respiratory effects on the scalp, indicating the need for future studies to take actual respiratory and heart rate signals into account (Candia-Rivera et al., 2022c;Dubois et al., 2016).Additionally, signals corresponding to blood pressure in Mayer waves contaminate fNIRS signals within the low frequency bands (Pinti et al., 2015;Kirilina et al., 2013), and nonlinearity evoked in blood pressure variations could similarly reflect in high frequency fNIRS.Hence, future studies should also monitor end-tidal CO2 as partial pressure of CO2 has been shown to affect vascular dynamics (Mas et al., 2000).Despite these limitations may introduce some uncertainty regarding the origin of nonlinear fNIRS dynamics in the fast wave band, our study's outcomes remain robust and reliable.This is further substantiated by the corroborative evidence provided through our comprehensive simulation-based analysis.
In conclusion, our study shows that nonlinear analysis can detect distinct neural representations in fast (0.2-0.6 Hz) waves compared to conventional slow (<0.2 Hz) wave frequencies in fNIRS data.Traditional methods like mean estimates could not reveal such distinct representations.While the origin of neural hemodynamic activity from a behavioral stimulus may be associated with oscillations at frequencies <0.2 Hz, nonlinear neurovascular interactions may generate fNIRS oscillations at higher frequencies.Thus, when assessing effects on fNIRS, comprehensive characterizations should also consider the nonlinear properties of high-frequency band oscillations.
Fig. 1 .
Fig. 1.Illustration of spatial analysis of fNIRS signals.(a) shows the position of the optodes used to derive the 36 fNIRS channels seen in (b).(b) shows an example of the full-width-half-max of the smoothing Gaussian kernel prior to second level group analysis (Tak et al., 2016) in the color green.
Fig. 4 .
Fig. 4. (a) demonstrates a single realization of a band passed oxyhemoglobin power spectrum density when either autoregulatory activity is occurring with a task, or when there is no autoregulatory activity with a task.(b) displays a representative simulation of the signals in the model and the events that influence them (whether a task induces autoregulatory behavior or not).
Fig. 5 .
Fig. 5. Results of a group level Friedman test comparison between the upper triangle of the RDM for fast (0.2-0.6 Hz) and slow wave ( < 0.2 Hz) fNIRS on the group level, where significance indicates at least one element in the upper triangle had a significantly different median between fast and slow wave fNIRS.Channels that are significant are marked by a thick outline on the marker.
Fig. 6 .
Fig. 6.Rainclouds illustrating the standardized Z transformed group level distributions of the absolute value of the difference between medians for mental arithmetic (a) and motor imagery (b) for either fast wave (0.2-0.6 Hz) or slow wave (0-0.2Hz) fNIRS signals.The "*" represents significant difference between the fast and slow wave distribution for the given measure. | 7,429.8 | 2023-09-01T00:00:00.000 | [
"Physics"
] |
Recent Advances in C–C and C–N Bond Forming Reactions Catalysed by Polystyrene-Supported Copper Complexes
This present mini-review covers recently published results on Cu(I) and Cu(II) complexes immobilized on polystyrene carriers, which are used as heterogeneous, eco-friendly reusable catalysts applied for carbon–carbon and carbon–nitrogen forming reactions. Recent advances and trends in this area are demonstrated in the examples of oxidative homocoupling of terminal alkynes, the synthesis of propargylamines, nitroaldolization reactions, azide alkyne cycloaddition, N-arylation of nitrogen containing compounds, aza-Michael additions, asymmetric Friedel–Crafts reactions, asymmetric Mukaiyama aldol reactions, and asymmetric 1,3-dipolar cycloaddition of azomethine ylides. The type of polystyrene matrix used for the immobilization of complexes is discussed in this paper, and particularly, the efficiency of the catalysts from the point of view of the overall reaction yield, and possible enantioselectivity and potential reusing, is reviewed.
Introduction
Preservation of the current standard of living for the human population is not further possible without the development of ecologically sustainable chemical processes and technologies [1,2]. One of the methods for contributing to responsible handling with raw materials and energy resources is the utilization of recyclable catalysts [1][2][3][4][5]. The strategy of how to obtain recyclable catalysts belongs the immobilization of known homogeneous catalysts to the solid carriers [3][4][5]. Organic polymers, especially commercially available cross-linked pearl-like copolymers of styrene e.g., Merrifield resin™ [6] and JandaJel resin™ [7] belong among the most relevant carriers. The remarkable advantage of polymer-supported catalysts lies in their simple separation and reusability, their low price, and also in the possibility of their usage in selective continuous-flow procedures [5,7]. The immobilization of the catalyst can be performed by the reaction of ligands or catalysts with suitably chemically activated polystyrene (post-modification strategy) [7]. Another possibility is the grafting of the ligand or catalyst containing a double bond from the reaction, with monomers generating the scaffold of the polymeric carrier (copolymerization strategy) [7]. The convenience of one or other strategies cannot be generally stated-it depends on many factors for each concrete case. Nowadays, there are many examples in chemical databases of recyclable catalysts based on complex compounds immobilized on different polymers. This review is focused on the area of highly efficient recyclable catalysts based on Cu(I) and Cu(II) complexes immobilized on polystyrene carriers. The aim of this review was to summarize, discuss and evaluate recently published results of catalysts, which were designed particularly for reactions, where new carbon-carbon respective carbon-nitrogen bonds are formed.
Oxidative Homocoupling of Terminal Alkynes
Carbon-carbon bond forming reactions are essential tools for building the carbon skeleton of organic compounds, and therefore they represent an elemental content of organic synthesis. Copper-catalysed oxidative homocoupling of terminal alkynes known also as Glaser coupling [8] introduce a suitable method for the preparation of 1,3-diynes, which are needful basic building blocks in the synthesis of many natural and pharmaceutical products, as well as some advanced functional materials [9]. A successful application of the heterogeneous catalyst intended for this reaction can be documented on the example of recyclable catalyst 1, which is based on the complex consisting of N,N,N′,N′tetraethyldiethylenetriamine and CuSO4, anchored on polystyrene cross-linked with 1% divinylbenzene (DVB) (Scheme 1) [10]. The efficiency of catalyst 1 was confirmed from the syntheses of 20 substituted symmetrical 1,3-diynes, where high yields of isolated products were achieved (58-99%). During the reuse of catalyst 1, release of small amount of Cu(II) ions was observed, however this did not affect significantly the yield of 1,4-diphenylbutane-1,3-diyne, even after nine cycles [10]. Another recently published successful heterogeneous catalyst [11] for Glaser coupling was based on the Cu(II) complex of copolymer N- (1,10-phenanthroline-5-yl)acrylamide and DVB. This catalyst was also used for Huisgen 1,3-dipolar cycloaddition [11]. For oxidative C-C homocoupling of acetylenes, a heterogeneous catalyst based on copper nanoparticles (3-9 nm), stabilized onto nitro functionalized polystyrene resin, was prepared. This catalyst was tested in the synthesis of seven 1,4-disubstituted-1,3-diynes (92-95%) [12].
The main advantage of this system lies in easy preparation of the catalyst from inexpensive and commercially available starting materials. Moreover, a significant benefit is the possibility of reusing of the catalyst, and the eco-friendly performance of the reaction in aqueous media [13].
Carbene Transfer Reaction
Reactions where carbenes were transferred onto unsaturated substrates (alkenes, alkynes, arenes, imines, aldehydes etc.), have provided a number of useful products [14][15][16][17][18][19]. The mentioned carbenes can be generated in situ from diazo compounds by the application of transition metal-based catalysis [14]. The most widespread example of such reaction implemented at large scale is copper-catalysed cyclopropanation of substituted alkenes with alkyl diazoacetates. Therefore, many homogeneous as well as heterogeneous catalytic systems have been developed for this reaction [14][15][16][17][18][19][20]. In the case of Cu(I) complexes, bisoxazoline ligands [15][16][17][18] and azabisoxazoline ligands [19] were used, anchored on pearl-like polystyrenes. For the immobilization of ligands either a copolymerization strategy (catalysts 3 and 4) or post-modification strategy (catalysts 5-7) was utilized. The catalysts were generated in situ by the coordination of copper(II) triflate and consequent reduction with phenylhydrazine (Scheme 3) [15][16][17][18][19][20]. The main advantage of this system lies in easy preparation of the catalyst from inexpensive and commercially available starting materials. Moreover, a significant benefit is the possibility of reusing of the catalyst, and the eco-friendly performance of the reaction in aqueous media [13].
Nitroaldolization Reaction
The nitroaldolization (Henry) reaction catalysed by chiral optically pure catalysts is important among reactions; it involves the formation of a new carbon-carbon bond [32][33][34][35][36][37][38][39]. This reaction occurs in the synthesis of substituted 2-nitroalcohols used e.g., for the preparation of biologically active compounds and medicinal drugs [34]. A recently published comprehensive paper [35] covers works concerning catalysts based on supported transition metal complexes aimed for an asymmetric variant of the Henry reaction. Three recent works were not covered in the above mentioned review [36,38,39]. The first work [39] included a study of a library of substituted polystyrylsulfonyl-imidazolineaminophenol copper complexes as solid supported catalysts for the Henry reaction. The selected complex of copper(II) acetate, based on (S,S)-diphenylethylenediamine-derived imidazoline, Scheme 6. Addition of terminal alkynes to imines catalysed by catalysts 2 and 11.
From the presented results, the prepared heterogeneous catalysts 2, 9-11 represent highly efficient catalytic systems. Particularly, catalyst 11 showed itself to be an eco-friendly variant that allowed the synthesis of propargylamines in aqueous media. From the chemical structure of catalyst 9, it contains a defined stereogenic centre. However, the eventual optical purity of chiral propargylamines prepared with the use of this catalyst was not discussed in the cited paper [23]. Elsewhere, e.g., in [27][28][29][30], several enantioselective homogeneous catalysts based on Cu(II) complexes designed for the synthesis of optically pure propargylamines were described. For the heterogeneous variant of enantioselective synthesis of propargylamines, only a version of a Cu(I) pybox complex anchored on magnetic nanoparticles (Fe 3 O 4 @SiO 2 ) have been described so far [31].
Nitroaldolization Reaction
The nitroaldolization (Henry) reaction catalysed by chiral optically pure catalysts is important among reactions; it involves the formation of a new carbon-carbon bond [32][33][34][35][36][37][38][39]. This reaction occurs in the synthesis of substituted 2-nitroalcohols used e.g., for the preparation of biologically active compounds and medicinal drugs [34]. A recently published comprehensive paper [35] covers works concerning catalysts based on supported transition metal complexes aimed for an asymmetric variant of the Henry reaction. Three recent works were not covered in the above mentioned review [36,38,39]. The first work [39] included a study of a library of substituted polystyrylsulfonyl-imidazoline-aminophenol copper complexes as solid supported catalysts for the Henry reaction. The selected complex of copper(II) acetate, based on (S,S)-diphenylethylenediamine-derived imidazoline, (S)-phenylethylamine, and dibromophenol, was highly efficient and provided the products of the reaction (nitromethane with substituted aldehydes) in high yields (76-99%) and high enantioselectivity (75-95% ee). The same ligand as a complex with copper(II) triflate was examined in the Friedel-Crafts alkylation of indole by substituted nitroalkenes, to give the adducts in high yields (86-99%) and high enantioselectivity-up to 83% ee [39]. The second variant of immobilization represented recyclable the heterogeneous catalyst 12 (Scheme 7), containing imidazolidine-4-one covalently bound on swellable pearl-like polystyrene having an -SH group (200-800 µm) [36]. Covalent anchoring of the ligand was performed by radical thiol-alkene click reaction, initiated thermally (azobisisobutyronitrile (AIBN), toluene, reflux 24 h), as well as photochemically (2,2-dimethoxy-2-phenylacetophenone, λ max = 365 nm, 40 W, DCM, 25 • C, 24 h). For the subsequent reaction of the copolymer with copper(II) acetate, heterogeneous catalyst 12 was prepared, and was characterized by elemental microanalysis and Raman spectroscopy [36]. Henry reactions proceeded in the polymeric matrix of swellable catalyst 12 with a rate comparable with the use of catalyst 13 in homogeneous media [37]. The corresponding substituted (R)-2-nitroethanols were produced in high yields (65-99%) and with high enantioselectivity (up to 92% ee) comparable with the homogeneous catalyst 13 (up to 96% ee). The reuse of catalyst 12 was studied for the reaction of pivalaldehyde with nitromethane. After five-fold recovery and reuse of heterogeneous catalyst 12, a decrease in yield of about 24% occurred; on the other hand, the decrease in enantioselectivity was negligible (3% ee) (Scheme 7) [36]. (S)-phenylethylamine, and dibromophenol, was highly efficient and provided the products of the reaction (nitromethane with substituted aldehydes) in high yields (76-99%) and high enantioselectivity (75-95% ee). The same ligand as a complex with copper(II) triflate was examined in the Friedel-Crafts alkylation of indole by substituted nitroalkenes, to give the adducts in high yields (86-99%) and high enantioselectivity-up to 83% ee [39]. The second variant of immobilization represented recyclable the heterogeneous catalyst 12 (Scheme 7), containing imidazolidine-4-one covalently bound on swellable pearl-like polystyrene having an -SH group (200-800 µm) [36]. Covalent anchoring of the ligand was performed by radical thiol-alkene click reaction, initiated thermally (azobisisobutyronitrile (AIBN), toluene, reflux 24 h), as well as photochemically (2,2-dimethoxy-2-phenylacetophenone, λmax = 365 nm, 40 W, DCM, 25 °C, 24 h). For the subsequent reaction of the copolymer with copper(II) acetate, heterogeneous catalyst 12 was prepared, and was characterized by elemental microanalysis and Raman spectroscopy [36]. Henry reactions proceeded in the polymeric matrix of swellable catalyst 12 with a rate comparable with the use of catalyst 13 in homogeneous media [37]. The corresponding substituted (R)-2-nitroethanols were produced in high yields (65-99%) and with high enantioselectivity (up to 92% ee) comparable with the homogeneous catalyst 13 (up to 96% ee). The reuse of catalyst 12 was studied for the reaction of pivalaldehyde with nitromethane. After five-fold recovery and reuse of heterogeneous catalyst 12, a decrease in yield of about 24% occurred; on the other hand, the decrease in enantioselectivity was negligible (3% ee) (Scheme 7) [36]. The third variant represented heterogeneous catalysts 14a,b based on Cu(II) complexes of substituted imidazolidine-4-thiones immobilized by means of sulfur atoms on different types of polystyrene carriers (poly[styrene-co-4-vinylbenzyl chloride-co-tetra(ethylene glycol)-bis(4-vinylbenzyl)ether), Merrifield resin™ and JandaJel resin™ (Scheme 8) [38]. It was found, that catalysts 14a,b catalyse the Henry reaction with high yields and high enantioselectivity, regardless of the polymeric carrier used (up to 98% ee). The obtained yields and enantioselectivities were comparable with the individual cases of reactions catalysed by analogical homogeneous catalysts 15a,b (Scheme 8). Beside monitoring the yields and enantioselectivity, the possibility of recovery and reusing of catalysts 14a,b was also studied. For instance, in the reaction of 2-methoxybenzaldehyde with nitromethane, the catalyst was reused more than 10 times without any decrease in enantioselectivity (81% ee) (Scheme 8) [38]. The third variant represented heterogeneous catalysts 14a,b based on Cu(II) complexes of substituted imidazolidine-4-thiones immobilized by means of sulfur atoms on different types of polystyrene carriers (polystyrene-co-4-vinylbenzyl chloride-co-tetra(ethylene glycol)-bis(4-vinylbenzyl) ether), Merrifield resin™ and JandaJel resin™ (Scheme 8) [38]. It was found, that catalysts 14a,b catalyse the Henry reaction with high yields and high enantioselectivity, regardless of the polymeric carrier used (up to 98% ee). The obtained yields and enantioselectivities were comparable with the individual cases of reactions catalysed by analogical homogeneous catalysts 15a,b (Scheme 8). Beside monitoring the yields and enantioselectivity, the possibility of recovery and reusing of catalysts 14a,b was also studied. For instance, in the reaction of 2-methoxybenzaldehyde with nitromethane, the catalyst was reused more than 10 times without any decrease in enantioselectivity (81% ee) (Scheme 8) [38].
Asymmetric Friedel-Crafts Reaction
Beletskaya et al. [40] studied asymmetric Friedel-Crafts alkylation of indole and its derivatives in the presence of complex of Cu(OTf)2 with chiral isopropyl bis(oxazoline) ligand immobilized on Merrifield resin™ according to the "click" procedure 16. The target products were obtained with a yield of up to 99% and up to 97% ee. Catalyst 16 ensured high yield and enantioselectivity even after five-fold reuse (Scheme 9) [40]. The heterogeneous catalyst 16 was also tested in the alkylation of N-methylindole by methyl (E)-2-oxo-4-phenylbut-3-enoate with yields of 22-94% and lower enantioselectivity (49% ee) [40].
Asymmetric Friedel-Crafts Reaction
Beletskaya et al. [40] studied asymmetric Friedel-Crafts alkylation of indole and its derivatives in the presence of complex of Cu(OTf) 2 with chiral isopropyl bis(oxazoline) ligand immobilized on Merrifield resin™ according to the "click" procedure 16. The target products were obtained with a yield of up to 99% and up to 97% ee. Catalyst 16 ensured high yield and enantioselectivity even after five-fold reuse (Scheme 9) [40].
Asymmetric Friedel-Crafts Reaction
Beletskaya et al. [40] studied asymmetric Friedel-Crafts alkylation of indole and its derivatives in the presence of complex of Cu(OTf)2 with chiral isopropyl bis(oxazoline) ligand immobilized on Merrifield resin™ according to the "click" procedure 16. The target products were obtained with a yield of up to 99% and up to 97% ee. Catalyst 16 ensured high yield and enantioselectivity even after five-fold reuse (Scheme 9) [40]. The heterogeneous catalyst 16 was also tested in the alkylation of N-methylindole by methyl (E)-2-oxo-4-phenylbut-3-enoate with yields of 22-94% and lower enantioselectivity (49% ee) [40].
Asymmetric Mukaiyama Aldol Reaction
The enantioselective Mukaiyama aldol reaction of methyl pyruvate with silylated ketene thioacetal was tested with a heterogeneous polystyrene copper complex catalyst based on chiral bis(oxazoline) [41] (Scheme 10). The highly-crosslinked polymer 3 was prepared by a copolymerization strategy in the presence of toluene as a porogen agent [41]. The heterogeneous catalyst 3/Cu(OTf)2 was less active than the analogous homogeneous catalysts, and aldol products were formed after a longer time; however, high overall yield (90%) and with an enantiomeric excess comparable to those obtained with the soluble ligands (90% ee), was obtained. Moreover, the catalyst based on polymer complex 3 was easily recovered by simple filtration and was reused several times without any loss of activity or stereoselectivity (Scheme 10) [41]. In the second work, the azabis(oxazoline) complexes 17 bonded to Merrifield resin™ were confirmed for the aldol reaction of different α-ketoesters with silylated ketone enolates [42]. The catalysis results showed that these heterogeneous complexes 17 accomplished enantioselectivities only slightly lower than those obtained in solution. Catalyst recovery depends on the ligand, so whereas polymers (R 3 : Ph) and (R 3 : i-Pr) were recovered with the same unsatisfactory results (51%, 38-39% ee), the recovered catalyst coming from (R 3 : t-Bu) led to even slightly lower yields (25%) but with the higher enantioselectivity (85% ee) (Scheme 10) [42].
Copper(I)-Catalysed Azide Alkyne Cycloaddition
Beside the reactions leading to carbon-carbon bond formation, the reactions providing new carbon-nitrogen bond also play a relevant role in organic synthesis. Particularly, this includes the Huisgen 1,3-dipolar cycloaddition of alkynes with azides providing 1,4-and 1,5-isomers of 1,2,3triazole [43]. Copper(I)-catalysed azide alkyne cycloaddition (CuAAC) represents a current variant, which converse to the original thermic version, proceeds quickly, regiospecifically (1,4-), and with high yields [44,45]. These properties rank this reaction among the most significant "near-perfect" bond-forming reactions known as "click" reactions [46]. The importance of this reaction consists not only in the efficient preparation of substituted 1,2,3-triazoles, but especially in the possibility for connecting different molecules and macromolecules, or to functionalize surfaces in a controlled way. The first polystyrene-supported copper(I) catalyst for CuAAC was based on the Amberlyst A21-CuI complex and was published in 2006 [47]. This and many further papers are discussed in a recently published review concerning immobilized copper complexes anchored on polystyrene and other heterogeneous carriers [48]. Also, two other papers deal with CuAAC [49,50].
The authors [49] prepared the catalyst 18 based on copper(I) iodide complex with cryptand-22 immobilized on commercial chloromethylated polystyrene (1% DVB) (Scheme 11). This heterogeneous heat-and air-stable catalyst 18 was used for a two-component CuAAC reaction of substituted organic azides with substituted alkynes, and this provided excellent yields of 1,4-disubstituted 1,2,3-triazoles (78-99%). The second three-component variant of this reaction involving organic halides, sodium The highly-crosslinked polymer 3 was prepared by a copolymerization strategy in the presence of toluene as a porogen agent [41]. The heterogeneous catalyst 3/Cu(OTf) 2 was less active than the analogous homogeneous catalysts, and aldol products were formed after a longer time; however, high overall yield (90%) and with an enantiomeric excess comparable to those obtained with the soluble ligands (90% ee), was obtained. Moreover, the catalyst based on polymer complex 3 was easily recovered by simple filtration and was reused several times without any loss of activity or stereoselectivity (Scheme 10) [41]. In the second work, the azabis(oxazoline) complexes 17 bonded to Merrifield resin™ were confirmed for the aldol reaction of different α-ketoesters with silylated ketone enolates [42]. The catalysis results showed that these heterogeneous complexes 17 accomplished enantioselectivities only slightly lower than those obtained in solution. Catalyst recovery depends on the ligand, so whereas polymers (R 3 : Ph) and (R 3 : i-Pr) were recovered with the same unsatisfactory results (51%, 38-39% ee), the recovered catalyst coming from (R 3 : t-Bu) led to even slightly lower yields (25%) but with the higher enantioselectivity (85% ee) (Scheme 10) [42].
Copper(I)-Catalysed Azide Alkyne Cycloaddition
Beside the reactions leading to carbon-carbon bond formation, the reactions providing new carbon-nitrogen bond also play a relevant role in organic synthesis. Particularly, this includes the Huisgen 1,3-dipolar cycloaddition of alkynes with azides providing 1,4-and 1,5-isomers of 1,2,3-triazole [43]. Copper(I)-catalysed azide alkyne cycloaddition (CuAAC) represents a current variant, which converse to the original thermic version, proceeds quickly, regiospecifically (1,4-), and with high yields [44,45]. These properties rank this reaction among the most significant "near-perfect" bond-forming reactions known as "click" reactions [46]. The importance of this reaction consists not only in the efficient preparation of substituted 1,2,3-triazoles, but especially in the possibility for connecting different molecules and macromolecules, or to functionalize surfaces in a controlled way. The first polystyrene-supported copper(I) catalyst for CuAAC was based on the Amberlyst A21-CuI complex and was published in 2006 [47]. This and many further papers are discussed in a recently published review concerning immobilized copper complexes anchored on polystyrene and other heterogeneous carriers [48]. Also, two other papers deal with CuAAC [49,50].
Halligudi et al. studied biomimetic oxidative coupling of 2-aminophenol to phenoxazine-2-one catalysed by bis(2′-[2-hydroxyethyl]benzimidazolato)copper(II) complex anchored onto chloromethylated polystyrene [53]. The kinetics and the effects of the catalyst concentration, substrate concentration, air pressure and temperature on the reaction course were studied. 2-Aminophenol in the presence of air (70 °C, 800 psig, DMF, 8 h) gave 62% conversion and the immobilized catalyst was more stable and more active than its homogeneous analogue. The catalyst recycling study demonstrated that the anchored catalyst did not leach out the metal complex during the reaction, and thus its catalytic activity was reproducible [53].
N-Arylation of Amines, Amides and Nitrogen Containing Heterocycles
Aryl-nitrogen bonds formed via cross-coupling reactions represent a powerful tool for the production of a number of compounds with practical uses [54]. The original variant of the Ullmann reaction is still utilized for this purpose [55]. However, this method of performing of the reaction is constrained by several disadvantages, e.g., it needs high temperature (150-200 °C) and the use of stoichiometric amounts of copper [54,55]. A remarkable simplification was achieved by the use of substituted phenylboronic acids for N-arylation of imidazole catalysed by copper(II) salts [56]. The next facilitation was the utilization of heterogeneous catalyst 22 (Scheme 14) [57]. Catalyst 22 was prepared from commercial chloromethylated polystyrene (5.5% DVB), on which β-alanine was attached through Interestingly, 20 showed much better catalytic performance than 21 (in all cases from 91 to >99% ee). Furthermore, the polystyrene-supported Fesulphos ligand 20 were recovered and reused in successive catalytic 1,3-dipolar cycloadditions without further addition of Cu(I), maintaining excellent enantioselectivity in up to three runs (Scheme 13) [52].
Halligudi et al. studied biomimetic oxidative coupling of 2-aminophenol to phenoxazine-2-one catalysed by bis(2 -[2-hydroxyethyl]benzimidazolato)copper(II) complex anchored onto chloromethylated polystyrene [53]. The kinetics and the effects of the catalyst concentration, substrate concentration, air pressure and temperature on the reaction course were studied. 2-Aminophenol in the presence of air (70 • C, 800 psig, DMF, 8 h) gave 62% conversion and the immobilized catalyst was more stable and more active than its homogeneous analogue. The catalyst recycling study demonstrated that the anchored catalyst did not leach out the metal complex during the reaction, and thus its catalytic activity was reproducible [53].
N-Arylation of Amines, Amides and Nitrogen Containing Heterocycles
Aryl-nitrogen bonds formed via cross-coupling reactions represent a powerful tool for the production of a number of compounds with practical uses [54]. The original variant of the Ullmann reaction is still utilized for this purpose [55]. However, this method of performing of the reaction is constrained by several disadvantages, e.g., it needs high temperature (150-200 • C) and the use of stoichiometric amounts of copper [54,55]. A remarkable simplification was achieved by the use of substituted phenylboronic acids for N-arylation of imidazole catalysed by copper(II) salts [56]. The next facilitation was the utilization of heterogeneous catalyst 22 (Scheme 14) [57]. Catalyst 22 was prepared from commercial chloromethylated polystyrene (5.5% DVB), on which β-alanine was attached through the oxygen atom, and copper(II) acetate was coordinated. Less efficient variants of this catalyst were prepared by the coordination with CuCl 2 and CuI. Catalyst 22 was successfully tested for the reaction of substituted phenylboronic acids with nitrogen containing heterocycles (imidazole, benzimidazole; 65-98%) and with amides (phthalimide, benzamide, benzenesulfonamide; 37-98%) under very mild conditions (methanol, 40 • C, 10-20 h). The reaction with amines was performed in DMSO in the presence of K 2 CO 3 at 140 • C (aniline, benzylamine; 58-91%) (Scheme 14) [57]. The recovery and reuse of catalyst 22 was studied for the reaction of phenylboronic acid with imidazole (Cycle 1: 98%; Cycle 6: 90%) and aniline (Cycle 1: 86%; Cycle 4: 82%). In both cases, only a slight decrease in yields was observed (Scheme 14) [57]. Under comparable conditions as in [57] The recovery and reuse of catalyst 22 was studied for the reaction of phenylboronic acid with imidazole (Cycle 1: 98%; Cycle 6: 90%) and aniline (Cycle 1: 86%; Cycle 4: 82%). In both cases, only a slight decrease in yields was observed (Scheme 14) [57]. Under comparable conditions as in [57] The recovery and reuse of catalyst 22 was studied for the reaction of phenylboronic acid with imidazole (Cycle 1: 98%; Cycle 6: 90%) and aniline (Cycle 1: 86%; Cycle 4: 82%). In both cases, only a slight decrease in yields was observed (Scheme 14) [57]. Under comparable conditions as in [57] Catalyst 23 was prepared from commercial chloromethylated polystyrene (2% DVB), on which 5-phenyl-1H-tetrazole was attached, and CuCl 2 was coordinated (Scheme 15). The reactions were performed in DMSO at 120 • C for 12-16 h in the presence of K 2 CO 3 (N-heterocycles; 67-95%), or in the presence of KOH (amines; 67-85%). In different solvents (THF, MeCN, DME, N-methylpyrrolidone (NMP), DMF) and under different conditions (Cs 2 CO 3 , NaHCO 3 , K 3 PO 4 , TEA; 80 • C, 130 • C) N-arylation of 1H-imidazole with iodobenzene occurred with lower yields than the reaction in DMSO. The recovery and reusing of catalyst 23 was tested with success for the reaction of imidazole with iodobenzene (Cycle 1: 94%; Cycle 6: 89%). The N-arylation of amides and lactams with substituted iodobenzenes, bromobenzene, 2-bromopyridine or 2-iodopyridine catalysed by 24 was also successful (Scheme 16) [61]. Catalyst 24 was prepared from commercial chloromethylated polystyrene, where the chloromethyl group was transformed to an aldehydic group. Subsequent reaction with β-alanine led to the production of an imine, which was then coordinated with Cu(II) acetate. N-arylations were performed without a solvent and in the presence of K2CO3 at 120 °C, for 15-20 h, with yields in the range of 62-92%. The recovery and reusing of catalyst 24 was successfully tested for the reaction of benzamide with iodobenzene (Cycle 1: 92%; Cycle 8: 89%) (Scheme 16) [61]. Further examples of reactions proceeding in aqueous media were N-arylation of substituted bromobenzenes with substituted anilines and (cyclo)alkylamines, catalysed by catalyst 25 in the presence of tetrabutylammonium bromide (TBAB) (Scheme 17) [62]. Catalyst 25 was prepared by the sequence of reactions from atactic polystyrene. Firstly, polystyrene was nitrated, reduced to amine, diazotated, and reduced to hydrazine, which was then acylated by 1H-pyrrole-2-carboxylic acid. The catalyst 25 itself was generated directly in the reaction mixture from the functionalized polystyrene and Cu(I) iodide. Catalyst 24 was prepared from commercial chloromethylated polystyrene, where the chloromethyl group was transformed to an aldehydic group. Subsequent reaction with β-alanine led to the production of an imine, which was then coordinated with Cu(II) acetate. N-arylations were performed without a solvent and in the presence of K 2 CO 3 at 120 • C, for 15-20 h, with yields in the range of 62-92%. The recovery and reusing of catalyst 24 was successfully tested for the reaction of benzamide with iodobenzene (Cycle 1: 92%; Cycle 8: 89%) (Scheme 16) [61]. Further examples of reactions proceeding in aqueous media were N-arylation of substituted bromobenzenes with substituted anilines and (cyclo)alkylamines, catalysed by catalyst 25 in the presence of tetrabutylammonium bromide (TBAB) (Scheme 17) [62]. Catalyst 25 was prepared by the sequence of reactions from atactic polystyrene. Firstly, polystyrene was nitrated, reduced to amine, diazotated, and reduced to hydrazine, which was then acylated by 1H-pyrrole-2-carboxylic acid. The catalyst 25 itself was generated directly in the reaction mixture from the functionalized polystyrene and Cu(I) iodide. Catalyst 24 was prepared from commercial chloromethylated polystyrene, where the chloromethyl group was transformed to an aldehydic group. Subsequent reaction with β-alanine led to the production of an imine, which was then coordinated with Cu(II) acetate. N-arylations were performed without a solvent and in the presence of K2CO3 at 120 °C, for 15-20 h, with yields in the range of 62-92%. The recovery and reusing of catalyst 24 was successfully tested for the reaction of benzamide with iodobenzene (Cycle 1: 92%; Cycle 8: 89%) (Scheme 16) [61]. Further examples of reactions proceeding in aqueous media were N-arylation of substituted bromobenzenes with substituted anilines and (cyclo)alkylamines, catalysed by catalyst 25 in the presence of tetrabutylammonium bromide (TBAB) (Scheme 17) [62]. Catalyst 25 was prepared by the sequence of reactions from atactic polystyrene. Firstly, polystyrene was nitrated, reduced to amine, diazotated, and reduced to hydrazine, which was then acylated by 1H-pyrrole-2-carboxylic acid. The catalyst 25 itself was generated directly in the reaction mixture from the functionalized polystyrene and Cu(I) iodide. The reaction of substituted bromobenzenes with benzylamine and substituted anilines (70 • C, 16 h) proceeded with very good yields of up to 82%. The lowest yield was attained in the reaction of 4-bromonitrobenzene with benzylamine (44%), while the reaction with sterically demanding 2,4,6-trimethylaniline did not proceed. Analogical reactions with (cyclo)alkylamines (90 • C, 16 h) provided yields of up to 93%, while the lowest yield was achieved in the reaction of 4-bromonitrobenzene with butylamine (51%). Substituted bromopyridines provided products with yields of 33-85%. The reactions of substituted iodobenzenes with imidazole (120 • C, 16 h) led to products with yields ranging from 65% to 87%. The lowest yield was obtained for iodobenzene itself (65%), while the reaction with 2-methoxyiodobenzene did not proceed. Catalyst 25 was also applied in the intramolecular N-arylation of N-(2-iodophenyl)-1H-imidazole-2-carboxamide giving imidazo[1,2-a]quinoxaline. The recovery and reusing of catalyst 25 was successfully tested for the reaction of 4-methoxyiodobenzene with imidazole (Cycle 1: 79%; Cycle 5: 82%) (Scheme 17) [62].
Aza-Michael Addition
Nucleophilic addition of N-nucleophiles on activated multiple bonds (aza-Michael addition) is a significant reaction in the synthesis of β-amino carbonyl compounds as crucial intermediates in the synthesis of β-amino acids, β-lactam antibiotics, and chiral auxiliaries [63]. Simple, green and efficient preparation of N-alkylated imidazoles was tested, with a reaction of substituted imidazoles with different α,β-unsaturated compounds catalysed by polystyrene-supported CuI-imidazole complex catalyst 26 (Scheme 18) [64]. Catalyst 26 was prepared via surfactant free emulsion polymerization with 4-VBC (4-vinylbenzylchloride) and DVB (2%), whereby uniform globular particles (400-500 nm) with high Cl content (21.5%) were obtained initially. Afterwards, these particles were highly functionalized by imidazole and coordinated by CuI. The efficiency of catalyst 26 was firstly confirmed for the addition of imidazole on acrylonitrile (8.5 mol% Cu; 60 °C; 4-10 h); while the influence of the solvent was studied: methanol (58%); MeCN (63%); DMSO (52%); DMF (71%); acetone (47%); THF (58%). The reactions were then performed in DMF and the temperature optimized to 75 °C. The addition of substituted imidazoles on α,β-unsaturated substrates proceeded with very good yields up to 74-84%. The lowest yield was achieved for the addition of 4-nitro-1H-imidazole on acrylonitrile. Catalyst 26 was separated by filtration and was further recycled for the reaction of imidazole with acrylonitrile (Cycle 1: 92%; Cycle 2: 91%; Cycle 5: 89%). The given results show that catalyst 26 represents an excellent reusable catalyst; which could be potentially applied at an industrial scale [64].
Conclusions
In this mini-review, we focused on Cu(I) and Cu(II) complexes anchored on pearl-like polystyrenes. The prepared heterogeneous catalysts were beneficially used for oxidative homocoupling of terminal alkynes, synthesis of propargylamines, nitroaldolization reaction, azide alkyne cycloaddition, Narylation of nitrogen containing compounds, aza-Michael addition asymmetric Friedel-Crafts reactions, asymmetric Mukaiyama aldol reactions, and asymmetric 1,3-dipolar cycloaddition of azomethine ylides. In several given examples, any difference between the efficiency of individual catalysts was not found depending on the method of their preparation (copolymerization strategy or post-modification strategy) [15][16][17][18][19]. In a prevalent number of examples, any significant decrease in chemical yields was not observed in the comparison of heterogeneous catalysts with corresponding The efficiency of catalyst 26 was firstly confirmed for the addition of imidazole on acrylonitrile (8.5 mol % Cu; 60 • C; 4-10 h); while the influence of the solvent was studied: methanol (58%); MeCN (63%); DMSO (52%); DMF (71%); acetone (47%); THF (58%). The reactions were then performed in DMF and the temperature optimized to 75 • C. The addition of substituted imidazoles on α,β-unsaturated substrates proceeded with very good yields up to 74-84%. The lowest yield was achieved for the addition of 4-nitro-1H-imidazole on acrylonitrile. Catalyst 26 was separated by filtration and was further recycled for the reaction of imidazole with acrylonitrile (Cycle 1: 92%; Cycle 2: 91%; Cycle 5: 89%). The given results show that catalyst 26 represents an excellent reusable catalyst; which could be potentially applied at an industrial scale [64].
Conclusions
In this mini-review, we focused on Cu(I) and Cu(II) complexes anchored on pearl-like polystyrenes. The prepared heterogeneous catalysts were beneficially used for oxidative homocoupling of terminal alkynes, synthesis of propargylamines, nitroaldolization reaction, azide alkyne cycloaddition, N-arylation of nitrogen containing compounds, aza-Michael addition asymmetric Friedel-Crafts reactions, asymmetric Mukaiyama aldol reactions, and asymmetric 1,3-dipolar cycloaddition of azomethine ylides. In several given examples, any difference between the efficiency of individual catalysts was not found depending on the method of their preparation (copolymerization strategy or post-modification strategy) [15][16][17][18][19]. In a prevalent number of examples, any significant decrease in chemical yields was not observed in the comparison of heterogeneous catalysts with corresponding homogeneous catalysts. The advantages of discussed heterogeneous catalysts consisted in their simple preparation from inexpensive and easily available starting materials, their simple isolation, and the possibility of reuse. Easy performance of the reaction in aqueous media brings great benefits in the case of cross-coupling reactions [13], Huisgen cycloaddition [49], N-arylation [62], and solvent-free N-arylation [61]. Most of the mentioned reactions possess high potential for industrial usage, hence the catalytic systems presented in this review significantly contribute to further development of ecologically sustainable chemical processes and technologies. | 7,420 | 2017-05-24T00:00:00.000 | [
"Chemistry",
"Materials Science"
] |