Comparative Study on Electronic Structure and Optical Properties of α-Fe2O3, Ag/α-Fe2O3 and S/α-Fe2O3
Round 1
Reviewer 1 Report
Dear authors,
This paper reports calculated results of hematite and doped hematite related materials. However, this report seems to be just a report on calculated results not to have enough scientific discussion for the results. The comments are below.
- ‘Doping’ means an introduction of another atom to a material. The amount of doping elements should be very small. ‘small’ means that unit cell type does not change by doping.
- Title and abstract contain grammatical error and misspelling.
- What kind of resource would be shortage should be described in the top line of introduction.
- The difference between your calculated results and experimental results of lattice constants of alpha-Fe2O3 should be discussed. The calculated results were +3.90% in a-axis and +3.27% in c-axis overestimated compared to known experimental results. This value is very large. The referee consider that subsequent description cannot be reliable for any discussion.
Sincerely Yours,
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
The authors present a computational study of the electronic and optical properties of Ag- and S-doped hematite.
The title is incorrect and refers to Au.
Methodological problems - convergence with no. of k-points is not discussed. Convergence with respect to supercell size is not discussed.
While S may replace O in hematite, it is not obvious that Ag with 1 valence electron can replace Fe with3 valence electrons. The authors did not consider charged defects.
The authors present optical absorption spectra in Frg.7 There is no explanation how these were obtained.
Conclusion 2 makes no sense in English and needs to be revised.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
In the manuscript "Comparative study on electronic structure and optical properties of α-Fe2O3, Au/α-Fe2O3 and S/α-Fe2O3", the authors report on structural, electronic, and optical properties of silver- and sulfur-doped alpha-hematite obtained using density functional theory calculations. Although there has been numerous studies on doped hematites up to now, the results presented in this manuscript could be of interest from photo-electrochemical point of view. However, the authors should address the following points before the manuscript is suitable for publication.
- It has been known for some time now that the effect of exact exchange could be very important in correct description of electronic and optical properties of hematite. Here, the authors use a non-hybrid GGA-PW91 XC functional. I believe that the authors should comment more on the general reliability of their results.
- The change in the high symmetry points in the Brillouin zone is a mere consequence of the modifications to the unit cell brought about by the doping. I think it should not be considered as a new finding.
- It would be helpful if the authors could comment on the defect (dopant) formation energies.
- It would also be good if the authors could explain how the half-metallicity caused by doping could affect the performance of the considered systems from a device perspective.
-
Just above the Conclusions section, the authors state "The doping of Ag and S doping can prevent the recombination of electrons and holes, which enhances the photocatalytic activity of α-Fe2O3." The hope is , of course, to reach such an effect with doping. However, this statement is not fully supported by the performed calculations/ presented discussions in the manuscript. Could the authors comment more on this?
Minor issues:
- I believe that the manuscript could benefit from an overall more scientific language. Therefore, I highly recommend a thorough English language editing to avoid the problems/typos which the reader could detect here and there in the current manuscript.
- In the title "Au" should be changed to "Ag".
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
The manuscript “Comparative study on electronic structure and optical properties of α-Fe2O3, Au/α-Fe2O3 and S/α-Fe2O3” by Cuihua Zhao, Baishi Li, Xi Zhou, Jianhua Chen and Hongqun Tang is devoted to ab-initio study of α-Fe2O3 dopped with Au or S. Although the study of this photocatalyst is a hot topic, it seems to me that the work does not correctly describe the magnetic state of this substance. First, when describing the calculation method, it is not discussed in any way how magnetism was taken into account (collinear / noncollinear, with or without spin-orbit interaction). The obtained band structures show that a ferromagnetic state is being calculated in which the band structure for spin-up and spin-down states of electrons are very different. At the same time, for example, in [Tasaki et al., J.Phys.Soc.Japan 18, 1148 (1963); https://materialsproject.org/materials/mp-19770/; Piccinin2019 doi:10.1039/C8CP07132B] it is stated that up to 260K hematite is in the antiferromagnetic state, and at higher temperatures it is in a weak ferromagnetic state.
In my opinion, this leads to an incorrect description of the band structure, which also leads to an incorrect description of the optical properties. In the literature, the calculations usually made for antiferromagnetic state and, less often, for the state of a weak ferromagnet.
It seems to me that the authors need to clarify this point. Only then can the manuscript be considered for publication. In any case, at least, a major revision of the article is needed or it should be rejected.
In addition to this basic objection, there are also other remarks:
- Atomic positions are not given.
- Number of atoms in unit cell is not given. So, the concentration of dopants is impossible to determine. In any case, the change in the crystal structure observed in the calculations (and the change in the Brillouin zone, which the authors emphasize) is an artifact of the supercell method used, since in a real doped material the position of impurity atoms in different cells will be random. In addition, at a low concentration of impurities, the crystal structure is most often preserved. Usually, at low concentrations, the cell parameters of the starting material are used in the calculations and, if necessary, the atomic positions are relaxed. The case of high concentration (solid solution) is more complicated: it is necessary to understand whether the type of crystal structure is preserved for the entire series of solid solutions or not.
- I suggest rotate vertically labels on PDOS plot.
- A large number of strange looking e-mail addresses is noteworthy.
- The method of absorption coefficient calculation is not described in Methods.
- I suggest also improve or correct following phrases:
“the peak of optical absorption peaks” – in Abstract
“In current applications and researches, researchers usually increase the utilization of solar energy...” - 1st paragraph of Introduction
“Monk-horst Pack“ - 1st paragraph of 2.1
“ The energy level across the Fermi level is split, and forms two DOS peaks, which corresponds to Fe 3d and O 2p.” -p.7 (cannot get the meaning)
“As we known, only the light above band-gap hits α-Fe2O3, the electrons in the valence band are excited to the conduction band leaving behind holes.” - p.10 (cannot get the meaning)
“Ag atom loses obtains more electrons “ - p.10 Conclusions
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors have improved their manuscript considerably in line with the reviewers comments and it is now above the threshold for acceptance.
Author Response
Thank you very much for your comment.
Reviewer 4 Report
In revised version of manuscript authors partially addressed some of questions and suggestions. I suggest to revise the manuscript, taking into account also following considerations.
Two important points should be thought of, in my opinion. Both of them are related to the attitude to the simulations and their results. At first, it is expected that simulations should model the system under study as close, as possible. Secondly, if there are differences between modeled and real physical system, they should be clearly stated. Therefore,
1) As it seems from literature that non-ferromagnetic state is most stable for considered material near room temperature [Tasaki et al., J.Phys.Soc.Japan 18, 1148 (1963); Piccinin2019 doi:10.1039/C8CP07132B], the results of calculations, obtained in manuscript, can be correct by themselves, but may be not related to real physical system. I understand, that the study of magnetic properties was not the purpose of the work, but, anyway, the magnetic state of material will affect the band structure and optical properties. At least, I suggest to mention in the section on the calculation method that the ferromagnetic state of the material was studied and if spin-orbital interaction was considered. The manuscript would benefit if some reasoning on the magnetic state selected was added, if possible.
2) Concerning structural relaxation in supercell approach, I would refer, e.g. to bookchapter “Supercell Methods for Defect Calculations” by R. Nieminen [https://www.researchgate.net/publication/226789999_Supercell_Methods_for_Defect_Calculations], especially, the section 4 “Issues with the Supercell Method” and its last paragraph on the p. 36:
“Finally, it is important to note that defect and impurity calculations should, as a rule, be carried out using the theoretical lattice constant, optimized for the bulk unit cell. This is crucial in order to avoid spurious elastic interactions with defects or impurities in the neighboring supercells. The purpose is to investigate properties of isolated defects or impurities in the dilute limit. If the volume of the defect-containing supercell is relaxed (in addition to relaxing the positions of the atoms near the defect), the calculation would in fact correspond to finding the lattice parameter of the system containing an ordered array of defects at a high concentration.”
In this respect, the substitution, of course, will change atomic geometry, but the modelling system, again, will not mimic real physical material with small impurity concentration. I am not sure that it should be stressed in the manuscript.
3) It seems that the number of Fe atoms in pure Fe2O3 should be 18 and not 17 (p. 3).
Also, in the same paragraph, why in pure Fe2O3 a is not equal to b?
4) Also, I suggest to consider the following correction “light with the photon energy larger than band-gap can excite electron into the conduction band” (if this is what the authors meant) in the phrase (p.12) “As we known, only the light above band-gap hits α-Fe2O3, an electron is excited into the conduction band, leaving a hole in the valence band”
5) p. 5 “One is 2.022 Å (O-Fe, O-Fe”
Author Response
In revised version of manuscript authors partially addressed some of questions and suggestions. I suggest to revise the manuscript, taking into account also following considerations.
Two important points should be thought of, in my opinion. Both of them are related to the attitude to the simulations and their results. At first, it is expected that simulations should model the system under study as close, as possible. Secondly, if there are differences between modeled and real physical system, they should be clearly stated. Therefore,
1) As it seems from literature that non-ferromagnetic state is most stable for considered material near room temperature [Tasaki et al., J.Phys.Soc.Japan 18, 1148 (1963); Piccinin2019 doi:10.1039/C8CP07132B], the results of calculations, obtained in manuscript, can be correct by themselves, but may be not related to real physical system. I understand, that the study of magnetic properties was not the purpose of the work, but, anyway, the magnetic state of material will affect the band structure and optical properties. At least, I suggest to mention in the section on the calculation method that the ferromagnetic state of the material was studied and if spin-orbital interaction was considered. The manuscript would benefit if some reasoning on the magnetic state selected was added, if possible.
Our answer: Thank you very much for your comments and suggestion. Iron oxide is a transition metal oxide which has different stoichiometric and crystalline structures, including wüstite (FeO), hematite (α-Fe2O3), maghemite (ν-Fe2O3), and magnetite (Fe3O4). Out of all the phases, hematite (α-Fe2O3) is the most stable state of iron oxide at ambient conditions. Hematite is a weak ferromagnet at room temperature and is antiferromagnetic at temperatures below 260 K (Maneesha Mishra, Doo-Man Chun, Applied Catalysis A: General 498, 126 (2015)). The calculated magnetic moment of every iron atom for pure α-Fe2O3 is 4.16 µB, which is the same with result of Pozum (Pozun et. al. Journal of Chemical Physics, 134, 224706 (2011)), but lower than the experimental value (4.6 µB) (Krén et. al. Physics Letters, 19, 103(1965)). After Ag and S doping, the magnetic moments of different iron atoms are different, however, the values become small, and magnetic moments of iron atoms for S/α-Fe2O3 is smaller than those for Ag/α-Fe2O3, suggesting that the magnetic property of S/α-Fe2O3 is weaker than that of Ag/α-Fe2O3.
Page 3 line 2 after “Table 3” added “In addition, the spin polarization calculations were conducted to consider the magnetic moments of the individual Fe atoms”. Page 4 line 2 after “S/α-Fe2O3 added “The calculated magnetic moment of every iron atom for pure α-Fe2O3 is 4.16 µB, which is the same with result of Pozum [23], but lower than the experimental value (4.6 µB)[25]. After Ag and S doping, the magnetic moments of different iron atoms are different, however, the values become small, and magnetic moments of iron atoms for S/α-Fe2O3 is smaller than those for Ag/α-Fe2O3, suggesting that the magnetic property of S/α-Fe2O3 is weaker than that of Ag/α-Fe2O3”.
2) Concerning structural relaxation in supercell approach, I would refer, e.g. to bookchapter “Supercell Methods for Defect Calculations” by R. Nieminen [https://www.researchgate.net/publication/226789999_Supercell_Methods_for_Defect_Calculations], especially, the section 4 “Issues with the Supercell Method” and its last paragraph on the p. 36:
“Finally, it is important to note that defect and impurity calculations should, as a rule, be carried out using the theoretical lattice constant, optimized for the bulk unit cell. This is crucial in order to avoid spurious elastic interactions with defects or impurities in the neighboring supercells. The purpose is to investigate properties of isolated defects or impurities in the dilute limit. If the volume of the defect-containing supercell is relaxed (in addition to relaxing the positions of the atoms near the defect), the calculation would in fact correspond to finding the lattice parameter of the system containing an ordered array of defects at a high concentration.”
In this respect, the substitution, of course, will change atomic geometry, but the modelling system, again, will not mimic real physical material with small impurity concentration. I am not sure that it should be stressed in the manuscript.
Our answer: Thank you very much for your comments. You are right. Maybe the modelling system cannot mimic real physical material with small impurity. Anyway, in the calculation, this change happened. I hope our results can provide a reference for other study on the doping system in the future. However, we deleted a small part of the description of the structural change.
Page 3 lines 5-7 from the bottom “The lattice type after Ag doping changed into 3D Triclinic from 3D Hexagonal, which may be due to the larger covalent radius of Ag atom than Fe atom” was deleted, line 2 from the bottom “The lattice type of α-Fe2O3 also changed into 3D Triclinic after S doping” was deleted.
3) It seems that the number of Fe atoms in pure Fe2O3 should be 18 and not 17 (p. 3).
Also, in the same paragraph, why in pure Fe2O3 a is not equal to b?
Our answer: Thank you very much for your comments. I am sorry for our negligence. Page 3 line 14 from the bottom “17 oxygen atoms” was revised to “18 oxygen atoms”. Page 3 line 12 from the bottom b=5.058 Å was revised to “b=5.088 Å”
4) Also, I suggest to consider the following correction “light with the photon energy larger than band-gap can excite electron into the conduction band” (if this is what the authors meant) in the phrase (p.12) “As we known, only the light above band-gap hits α-Fe2O3, an electron is excited into the conduction band, leaving a hole in the valence band”
Our answer: Thank you very much for your suggestion. Page 12 line 7 “As we known, only the light above band-gap hits α-Fe2O3, an electron is excited into the conduction band, leaving a hole in the valence band” was revised to “As we known, the light with the photon energy larger than band-gap can excite electron into the conduction band”
5) p. 5 “One is 2.022 Å (O-Fe, O-Fe”
Our answer: Thank you very much for your comments. Page 5 “2.022 Å (O-Fe, O-Fe, O5-Fe, O4-Fe)” was revised to “2.022 Å (O-Fe, O-Fe1, O5-Fe, O4-Fe)”
Round 3
Reviewer 4 Report
Authors address reviewer's comments, so I think that the manuscript can be published in the present form.
