Probing the Local Atomic Structure of In and Cu in Sphalerite by XAS Spectroscopy Enhanced by Reverse Monte Carlo Algorithm
and Kristina O. Kvashnina 2,4,5Round 1
Reviewer 1 Report
The work is sound and there are interesting aspects of the data analysis. My only concern is that the authors give no indication as to the importance of doped sphalerites. Is this an area of significant scientific interest. This need to be made clear in the manuscript.
Author Response
Dear Reviewer,
In the introduction section, it's pointed out that this Paper is followed out two papers {references 1 and 2 from the Manuscript: 1. Filimonova, a.N.; Trigub, A.L.; Tonkacheev, D.E.; Nickolsky, M.S.; Kvashnina, K.a.; Chareev, D.A.; Chaplygin, I. V.; Kovalchuk, E. V.; Lafuerza, S.; Tagirov, B.R. Substitution mechanisms in In-, Au-, and Cu-bearing sphalerites studied by X-ray absorption spectroscopy of synthetic compounds and natural minerals. Mineral. Mag. 2019, 83, 435-451. 10.1180/mgm.2019.10. 2. Trofimov, N.D.; Trigub, A.L.; Tagirov, B.R.; Filimonova, a.N.; Evstigneeva, P. V.; Chareev, D.A.; Kvashnina, K.a.; Nickolsky, M.s. The State of Trace Elements (In, Cu, Ag) in Sphalerite Studied by X-ray Absorption Spectroscopy of Synthetic Minerals. Minerals 2020, 10, 640. 10.3390jminl(070640). In the first paper Filimonova et al. provide extensive justification, literature review on the chemistry of doped sphalerites and geological impact. In the second paper Trofimov et al. performed an investigation of the synthetic doped sphalerites using conventional X-ray absorption techniques. The article of Trofimov et al. contains a detailed description of the samples synthesize and their characterization. And finally, in our present Manuscript, we tried to use state-of-the-art techniques to extract structural information from EXAFS spectra as much as possible.
Reviewer 2 Report
The manuscript of Trigub et al. describes Reverse Monte Carlo based interpretation of EXAFS (and XANES) data, in combination with DFT calculations, taken on various pure and doped forms of ZnS. The experimental and computational work are formidable, so I would be happy to recommend this work for publication -- after the authors have made some important modifications (see below).
(1) The Abstract is far too long -- should be cut to its half.
(2) The same is true for the Conclusions: this section should contain summarizing statements of the findings, and not the re-cap of the whole manuscript.
(3) It is not clear whether the RMC calculations applied all measured data, i.e., Zn, In and Cu edge EXAFS, simultaneously. For the In and Cu doped samples, there could be three independent measurements (at the Zn, Cu and In edges), so that ideally, the RMC structures could be consistent with all three datasets at the same time -- is this actually happening? (Or perhaps RMC calculations are always run on the basis of only one EXAFS dataset? This is what it looks from the presentation -- and it would be a pity...) In any case, it must be made clear how many experimental datasets are considered in the various RMC calculations.
(4) It would perhaps be polite to mention one (or some) of the original RMC publications (e.g., the one that introduced the algorithm in 1988) -- just as the works of Kuzmin et al. do, so that readers who have no background concerning the method would be able to find a reference quickly.
Author Response
Dear Reviewer,
Thank you for your careful reading of our manuscript and proposed suggestions to improve the manuscript.
- The abstract was significantly reduced.
- The conclusion part was renamed on Discussion since it contains a valuable discussion of experimental results.
- Unfortunately, we have not collected Zn K-edge XAS data. Fitting of Zn K-edge EXAFS performed just for pure ZnS to get reference radial distribution functions. Fitting of In and Cu K-edge spectra performed independently using quite small clusters of atoms (first three coordination shells) without periodic boundary conditions. We used such approximation to treat all data in the same manner. For example, samples with In differ by its concentrations (0.1 and 5 wt. %), and this means that to get smooth radial distribution functions (RDF) and to mimic the real concentrations huge supercells (nearly 100000 atoms) should be used. So, we decided to use one simplified structural model for all spectra. It should be noted that the presence of In in close vicinity of Cu or Cu near In doesn't change the results of RMC procedure (see supplementary information).
- Work by R. L. McGreevy and L. Pusztai (1988) was added to the References.
Reviewer 3 Report
This MS (# Minerals-879757) entitled “Probing the local atomic structure of In and Cu in sphalerite by XAS spectroscopy enhanced by reverse Monte-Carlo algorithm” is properly organized, well-written and informative. Therefore, it is of potential interest for the scientific community and worth to be published in the MDPI journal Minerals after minor corrections listed below:
Line 19: change “The distortion of the atomic structure around In and Cu atoms” with “The distortion of the coordination environment of In and Cu dopant atoms”
Line 23: change “using evolutionary algorithm (EA) of the RMC method.” with “using the RMC modeling coupled with the evolutionary algorithm (EA).”
Lines 25-27: change in “The RMC-EXAFS method showed that the three RDF peaks corresponding to the three coordination shells of In-bearing sphalerite are modeled by Gaussian functions.”
Line 28: change “symmetrical undisturbed structural” with “undistorted coordination”
Line 32: change “the splitting rate” with “and the splitting extent”
Line 33: please change "not symmetric" with "distorted" since the Zn replacement with Cu causes a reduction of site symmetry.
Lines 37-38: change in “better agreement between the simulated and the experimental spectra.
Lines 73-75: The authors’ statement about EXAFS spectra approximation is not quite correct. Please, see the chapter "Beyond the Gaussian approximation in EXAFS" by E.Stern for clarification. I suggest to replace with: "which allows more efficient exploration of disorder effects beyond the first coordination shell when multiple-scattering (MS) effects are not negligible."
Line 96: “European Synchrotron. Radiation Facility (ESRF) in Grenoble, France.”
Line 112: change “Ry/au” in “Ry/a.u.”
Line 192: delete “than”
Line 222, Table 1, 2nd row: change “cites” in “sites”
Line 256: change “coordination shells presented by 4 S” with “coordination shells of the probed cation with 4 S”
Line 262: change “what” with “that”
Lines 264-265: change “which represents 4 S atoms in the nearest to Cu and In coordination shells” with “which represents the mean bond length between the central atom (Cu or In) and its nearest neighbors (4 S)”
Line 265: The Authors ascribe the peak asymmetry of the fist RDF peak only to artefacts due to the limited range of the available experimental signal. However, the non-Gaussian shape of bond length distribution for the first coordination shell could be also due to pronounced disorder effects.
Line 368: change in “The In atoms are present in the”
Line 376: correct “distorts”
Line 377: change “distortion rate” in “and the distortion extent”
Line 396: change “which can be explained by Coulomb interaction” with “which can be explained by local charge balance”
Line 403: change “rate” with “extent”
Line 430: change “M.,” with “M.”
Line 431: change “under review” with “DOI: 10.3390/min10070640”
Supplementary Mat. Table S1: correct “cites” with “sites” and “Au” with “Cu”
Supplementary Mat. figure S1: change “adopted” with “taken”.
Supplementary Mat. References, Trofimov et al.: change “M.,” with “M.” and “under review” with “DOI: 10.3390/min10070640”
Comments for author File: 
Comments.pdf
Author Response
Dear Reviewer,
Thank you for your careful reading of our manuscript. We have accepted almost all the suggestions to improve the text of our manuscript.
Comment:
Correction: "The Authors ascribe the peak asymmetry of the fist RDF peak only to artefacts due to the limited range of the available experimental signal. However, the non-Gaussian shape of bond length distribution for the first coordination shell could be also due to pronounced disorder effects."
Answer: "According to DFT calculations splitting of the first coordination shell is very weak, just nearly 0.02 A, and this value couldn't be reliably determined by EXAFS. Nevertheless, significant underestimation of splitting value by DFT in comparison with EXAFS shows that real atomic structure around Cu and In impurities in ZnS is much more complicated than proposed for DFT calculations. Thus, the splitting of the first coordination shell could be explained by the short length of the spectra and by strong distortion of the nearest atomic environment."
Reviewer 4 Report
I read the manuscript "Probing the local atomic structure of In and Cu in sphalerite by XAS spectroscopy enhanced by reverse Monte-Carlo algorithm" with interest. This manuscript contains important "XAS spectroscopy enhanced by reverse Monte-Carlo algorithm". On the other hand, the introductory part is not well written, so to make it more friendly to general readers, I would like to point out the following before recommending publication.
- It is not clear why it is interesting and important to study the electronic state of dopants (In, Cu, Ag, Au, Cd) in ZnS.
- The sample list is an important information, so the authors should make a clear table rather than just showing Ref. [2].
- The sample information is lacking. Is this a thin film on a substrate?
- The reference XAS data is not shown, so it is hard to tell why the authors consider Zn2+, In3+ and Cu+.
- The real XAS data should be shown rather than starting from chi(k).
Author Response
Dear Reviewer,
- In the introduction section, it's pointed out that this Paper is followed out two papers {references 1 and 2 from the Manuscript: 1. Filimonova, a.N.; Trigub, A.L.; Tonkacheev, D.E.; Nickolsky, M.S.; Kvashnina, K.a.; Chareev, D.A.; Chaplygin, I. V.; Kovalchuk, E. V.; Lafuerza, S.; Tagirov, B.R. Substitution mechanisms in In-, Au-, and Cu-bearing sphalerites studied by X-ray absorption spectroscopy of synthetic compounds and natural minerals. Mineral. Mag. 2019, 83, 435-451. 10.1180/mgm.2019.10. 2. Trofimov, N.D.; Trigub, A.L.; Tagirov, B.R.; Filimonova, a.N.; Evstigneeva, P. V.; Chareev, D.A.; Kvashnina, K.a.; Nickolsky, M.s. The State of Trace Elements (In, Cu, Ag) in Sphalerite Studied by X-ray Absorption Spectroscopy of Synthetic Minerals. Minerals 2020, 10, 640. 10.3390jminl(070640). In the first paper Filimonova et al. provide extensive justification, literature review on the chemistry of doped sphalerites and geological impact. In the second paper Trofimov et al. performed an investigation of the synthetic doped sphalerites using conventional X-ray absorption techniques. The article of Trofimov et al. contains a detailed description of the samples synthesize and their characterization. And finally, in our present Manuscript, we tried to use state-of-the-art techniques to extract structural information from EXAFS spectra as much as possible.
- Please see answer to the question 1. In brief, this information is discussed in our previous papers (Trofimov et al., Minerals, 2020, 10, 640).
- The samples are the powders. Detail information are published in the our previous paper (Please see answer to comment of Reviewer 1. In brief, this information is discussed in our previous papers (Trofimov et al., Minerals, 2020, 10, 640).
- and 5. We prefer not to repeat this information in the present manuscript as both the papers are planned to be published in the same issue of Minerals (Trofimov et al., Minerals, 2020, 10, 640).
Reviewer 5 Report
This manuscript is a careful analysis of the EXAFS of Cu and In doped ZnS. The authors have used RMC-EXAFS analysis and DFT calculations to extract subtle features of the EXAFS around each of the three cations to contrast with and confirm a more traditional EXAFS analysis presented in Reference [2]. The primary result obtained is that the RMC-EXAFS analysis identifies additional bond distances not originally used in the fits from Reference [2]. The authors do a thorough job of justifying this result and how it is additionally able to reproduce the experimental XANES.
Overall, this paper is well written and presents a convincing case for using RMC-EXAFS for subtle feature extraction. The authors state that their DFT calculations produced substantially similar results, specifically for the distortions in the second shell about the Cu impurities, however they do not present any specific results of these calculations in any of their tabular data either in the manuscript or the SI. The paper would be much stronger with the addition of these DFT results and comparisons to the RMC-EXAFS and conventional EXAFS analysis. Finally, a bit more discussion of the underlying reasons for these second shell distortions would be a welcome addition to the paper to make it much more significant.
In addition, to the above general comments, I list below some issues in the form of the presentation which could easily be corrected:
In the SI, reference [2] is incomplete as the previous article has been published and it referenced correctly in the manuscript.
I found the caption for Figure 1 to be confusing. Labeling the plots with letters and referring to such letters would be beneficial. THesame goes for Figures 2, 3, and S2.
In Table 2, the estimated standard deviations for the EXAFS fits should be reported. The fonts and columns should be arranged in such a way as to ensure that all numerical quantities appear on a single line. Estimated standard deviations should be reported in Table 3 as well.
Figures 2 and 3 seem to be overkill. These could be included in the SI, not the main body. Figure 1 serves to present the quality of the data and fits.
In Figure S2, should the green curve in the bottom right plot (sample 4186) be Cu-In instead of Cu-Zn?
Line 421 "account" instead of "accound"
Author Response
Dear Reviewer,
Here the aswers on your comments:
"The paper would be much stronger with the addition of these DFT results and comparisons to the RMC-EXAFS and conventional EXAFS analysis."
Results of DFT calculations of Cu- and Cu-In-bearing sphalerites are given in Table S1 in the Supplementary Information. Comparision of DFT results with results of the EXAFS spectra fittings is given in the manuscript (Filimonova, a.N.; Trigub, A.L.; Tonkacheev, D.E.; Nickolsky, M.S.; Kvashnina, K.a.; Chareev, D.A.; Chaplygin, I. V.; Kovalchuk, E. V.; Lafuerza, S.; Tagirov, B.R. Substitution mechanisms in In-, Au-, and Cu-bearing sphalerites studied by X-ray absorption spectroscopy of synthetic compounds and natural minerals. Mineral. Mag. 2019, 83, 435-451. 10.1180/mgm.2019.10.). We would like to emphasise that results of DFT calculations are only auxiliary. These data helped us to interpret results of RMC-EXAFS and conventional EXAFS fittings. Therefore, extended discussion of the DFT results is out of the scope of our manuscript and would complicate reading.
Finally, a bit more discussion of the underlying reasons for these second shell distortions would be a welcome addition to the paper to make it much more significant.
The main reason of the second shell splitting is the different ionic radii for Zn2+, Cu1+ and In3+ ions. This issue is also mentioned in the section "2.4. Reverse Monte Carlo (RMC) EXAFS Spectra Fitting"
Figures 2 and 3 seem to be overkill. These could be included in the SI, not the main body. Figure 1 serves to present the quality of the data and fits.
Figures 2 and 3 present the main results of the study. We believe that these figures are necessary for readers to follow the manuscript text. Therefore, we would prefer to keep the figures in the main manuscript body.
I found the caption for Figure 1 to be confusing. Labeling the plots with letters and referring to such letters would be beneficial. THesame goes for Figures 2, 3, and S2.
Figure captions were corrected.
In Table 2, the estimated standard deviations for the EXAFS fits should be reported.
In Monte-Carlo modeling, the estimated standard deviation could be calculated like Gaussian variance divided by the square root of the number of bonds taken into calculation. In our case, the number of clusters is 432 and coordination number of the second and third coordination shells is 12. Therefore the number of connections used to calculate bond length (for example, Cu-Zn) is almost 5000 and corresponding statistical error could be estimated as 20/(1000*sqrt(5000))=0.0004. We are sure that the main source of possible error is not the collected statistic during the Monte-carlo modelling, but the chosen atomic model to fit EXAFS spectra.
We thank the Reviewer for the minor comments. The manuscript was corrected accordingly.
Round 2
Reviewer 2 Report
This is my second reading of this manuscript. As a result of the modifications made by the authors, the paper has improved.
However, there is now no 'Conclusions/summary' section -- and this makes it difficult to understand the main message(s) of the work. I ask the authors to create such a section and highlight the main findings in a concise way. (I actually recommend itemization.)
Author Response
Dear Reviewer,
"Conclusions" section has been added, and some corrections of English have been made in the Text of the Article.
Best regards,
Authors.
