Single-atom Pt-I3 sites on all-inorganic Cs2SnI6 perovskite for efficient photocatalytic hydrogen production

Organic-inorganic lead halide perovskites are a new class of semiconductor materials with great potential in photocatalytic hydrogen production, however, their development is greatly plagued by their low photocatalytic activity, instability of organic component and lead toxicity in particular. Herein, we report an anti-dissolution environmentally friendly Cs2SnI6 perovskite anchored with a new class of atomically dispersed Pt-I3 species (PtSA/Cs2SnI6) for achieving the highly efficient photocatalytic hydrogen production in HI aqueous solution at room temperature. Particularly, we discover that Cs2SnI6 in PtSA/Cs2SnI6 has a greatly enhanced tolerance towards HI aqueous solution, which is very important for achieving excellent photocatalytic stability in perovskite-based HI splitting system. Remarkably, the PtSA/Cs2SnI6 catalyst shows a superb photocatalytic activity for hydrogen production with a record turnover frequency of 70.6 h−1 per Pt, about 176.5 times greater than that of Pt nanoparticles supported Cs2SnI6 perovskite, along with superior cycling durability. Charge-carrier dynamics studies in combination with theory calculations reveal that the dramatically boosted photocatalytic performance on PtSA/Cs2SnI6 originates from both unique coordination structure and electronic property of Pt-I3 sites, and strong metal-support interaction effect that can not only greatly promote the charge separation and transfer, but also substantially reduce the energy barrier for hydrogen production. This work opens a new way for stimulating more research on perovskite composite materials for efficient hydrogen production.

In this manuscript, Chen and coworkers reported a high efficient photocatalyst for HI splitting based on Cs2SnI6 perovskite crystals with single atomic Pt-I3 sites for accelerating the HER process. This optimized Pt SA/Cs2SnI6 catalyst delivered high photocatalytic activity of 430 μmol h-1 g-1 and superior durability, which was comparable to MAPbI3-based photocatalysts. More interesting, the turnover frequency of 70.6 h-1 per Pt was achieved in Pt SA/Cs2SnI6, which was about 176.5 times greater than that of Pt NPs supported Cs2SnI6 perovskite. It was demonstrated that the introduction of Pt-I3 sites could promote the charge separation and transfer, and reduce the HER barrier. This work should also be inspirational for the design of perovskite-based materials with the combination of single atomic sites for other photocatalytic applications. Therefore, I would strongly recommend this work for publishing on Nat. Commun. after addressing the following concerns: (1) The solar-to-hydrogen conversion efficiency (STH) value of Pt SA/Cs2SnI6 should be provided.
(2) The authors demonstrated a poor HER activity of Pt NPs/Cs2SnI6 by both its LSV curve (Supplementary Fig. 18) and theoretical calculation (Fig. ), which was described as one of the key reasons for its low photocatalytic activity. However, I donot fully agree with this point, because it is well known that Pt-based nanomaterials are the best HER electrocatalysts. This inferior activity of Pt NPs/Cs2SnI6 may greatly relate to the possible sluggish electron transfer dynamics from CS2SnI6 to Pt NPs due to the mismatch between the conduction band of Cs2SnI6 and Fermi level of Pt NPs. The electrochemical signals of the LSV curve only reflected a generalized HER activity of the whole composite, and the actual HER activity of located Pt NPs may be suppressed due to its low content. For the theoretical calculation aspect, the provided model for Pt NPs was more like a Pt cluster, in which the underlying I-species could greatly affect the H* adsorption energy.
(3) Furthermore, please add more details of the HER measurements in the experimental part, such as the used electrolyte and scanning rates. The description of the reference electrode (NHE) in the main text should be consistent with that in supplementary Fig. 18 (RHE).
Reviewer #2 (Remarks to the Author): In this work, the authors reported the decoration of all-inorganic Cs2SnI6 perovskite with single-atom Pt-I3 sites as cocatalyst (PtSA/Cs2SnI6) towards photocatalytic H2 evolution from HI splitting reaction. The as-prepared PtSA/Cs2SnI6 catalyst showed a turnover frequency (TOF) 176.5 times higher than Cs2SnI6 decorated with Pt nanoparticle, along with excellent catalytic stability. Spectroscopy investigations together with theoretical calculations reveals that both unique coordination structure and electronic property of Pt-I3 sites and strong metal-support interaction effect are the main reasons in achieving the greatly enhanced photocatalytic performance in hydrogen production from HI aqueous solution. The manuscript is well-organized and written and the results are interesting. I would suggest the acceptance of the work after the authors have addressed the following issues.
1、The PtNP/Cs2SnI6 was prepared by mixing pre-synthesized PtNp and Cs2SnI6 powder. In this case, the contact between PtNp and Cs2SnI6 powder should be poor, which will impair the interfacial charge transfer and therefore the photocatalytic activity. To make a fair comparison between PtNP/Cs2SnI6 and PtSA/Cs2SnI6, the PtNP/Cs2SnI6 sample could be prepared with photodeposition method to allow the direct growth of PtNP on Cs2SnI6. The photocatalytic activity of the as-prepared PtNP/Cs2SnI6 should be also tested. 2. The XPS spectra suggest that the valence state of Pt element in the sample is Pt2+ or Pt4+. However, as Pt2+ or Pt4+ can be easily reduced to Pt by photogenerated electrons during reaction, XPS analysis should be performed on the PtSA/Cs2SnI6 after phototcatalytic reactions to allow the identification of the chemical state of Pt. 3. In Figure 4c, the conditions like the applied bias, the electrolyte (anything else besides 0.1 M TBAPF6 dichlormethane) for the photocurrent measurements should be provided. Moreover, as PtSA is used as the HER catalyst in this work to promote reduction reaction. I'm curious why positive photocurrents instead of negative photocurrents are used to testify the role of PtSA. 4、Some typos or errors in the manuscript should be corrected (Reference 23,24 et al.).
Reviewer #3 (Remarks to the Author): Hui Chen et al. reports on the photocatalytic properties of of a novel composite based on the Cs2SnI6 perovskite decorated with atomically dispersed Pt-I3 species. I have two main concerns about the manuscript. The first regards the possible appeal to a broad audience. In fact, the system used by the author, while possessing an improved stability in aqueous environment with respect to MAPbI3 and showing interesting properties, still implies the use of highly concentrated HI solutions, thus severely limiting the range of applicability, which is restricted (e.g. no hydrogen production in more desirable conditions). Considering that perovskites which are actually water-stable have started to be synthetized (cf. ,https://doi.org/10.1002/anie.202007584), I fear that, without any interesting and motivated interpretation of novel phenomena, the manuscript in the current form, would be more suitable for a specialized journal. Nevertheless, also in that case, there are some points that the authors need to address because the interpretation of their results appears to be shallow, in particular from the theory, and some results completely lack an explanation.
1) The computational details for the calculated results are frankly insufficient. The choice of adopted functional (which by the way does not include Van der Waals interactions that might play a role) is not motivated, convergence tests for the employed slab and the size of the vacuum layer are not reported. 2) Similarly, results are not clear. How do the authors simulate "before and after photexcitation" conditions in their calculations? Do calculation include extra electrons/holes? Then would a GGA treatment be sufficient to describe unpaired charges?
3) The authors reports charge density differences for the pristine material and PtSA but they do not include any comparison with PtNp. Furthermore, the localization of the charge in the figure is not clear. Are the author claiming that Cs2SnI6 feature a delocalized electronic state while, in presence of the additive, they observe a localized state in the band gap of the material? 4) I think that the claim that PDOS explains the difference between PtSA and PtNp is quite weak. I cannot see a dramatic difference among the two. 5) How do the authors calculate the energy diagrams in Fig. 5d? 6) The author state that "instability of the organic component" is at the root of the limited application of perovskites in photocatalysis. However, recent reports show that organic-inorganic perovskite can be water-stable. 7) The authors should comment on the trends observed for Pt loading? Why, after a maximum, the photocatalytic activity decreases? Minor points: (i) The authors should specify that the band alignment achieved via XPS is an approximation, since it does not include effects of the water-perovskite interface in the band alignment.
(ii) "The most 5d states of PtSA are below the Fermi level, indicating its strong electron-captured ability" this sentence is not clear and should be rephrased. Overall, on the basis of my comments, I cannot suggest publication of the manuscript in the present form which need to be profoundly revised in order to explain the results In this manuscript, Chen and coworkers reported a high efficient photocatalyst for HI splitting based on Cs2SnI6 perovskite crystals with single atomic Pt-I3 sites for accelerating the HER process. This optimized Pt SA/Cs2SnI6 catalyst delivered high photocatalytic activity of 430 μmol h -1 g -1 and superior durability, which was comparable to MAPbI3-based photocatalysts. More interesting, the turnover frequency of 70.6 h -1 per Pt was achieved in Pt SA/Cs2SnI6, which was about 176.5 times greater than that of Pt NPs supported Cs2SnI6 perovskite. It was demonstrated that the introduction of Pt-I3 sites could promote the charge separation and transfer, and reduce the HER barrier. This work should also be inspirational for the design of perovskite-based materials with the combination of single atomic sites for other photocatalytic applications. Therefore, I would strongly recommend this work for publishing on Nat. Commun. after addressing the following concerns: R: Thanks for your great efforts in reviewing our manuscript. We specially appreciate your valuable comments and suggestions. We have performed all the experiments suggested by you, further addressed the comments point-by-point and made the corresponding changes accordingly in the revised manuscript.

Q1:
The solar-to-hydrogen conversion efficiency (STH) value of Pt SA/ Cs2SnI6 should be provided.

R1:
Thanks for your valuable comment. In general, the STH is mainly used in the uphill photocatalytic reaction, such as overall water splitting with the production of stoichiometric hydrogen and oxygen (H2O  H2 + 0.5O2, G = +237 kJ mol -1 ). However, converting HI into hydrogen and by-product I2 in our work is not an uphill reaction according to the chemical formula (2HI  H2 + I2, G = -1.32 kJ mol -1 ), which is mainly used for the hydrogen storage instead of solar-energy storage. In this regard, the STH may be not very necessary for the photocatalytic conversion of HI into hydrogen.

Q2:
The authors demonstrated a poor HER activity of Pt NPs/Cs2SnI6 by both its LSV curve ( Supplementary Fig. 18) and theoretical calculation (Fig.), which was described as one of the key reasons for its low photocatalytic activity. (a) However, I do not fully agree with this point, because it is well known that Pt-based nanomaterials are the best HER electrocatalysts. This inferior activity of Pt NPs/Cs2SnI6 may greatly relate to the possible sluggish electron transfer dynamics from Cs2SnI6 to Pt NPs due to the mismatch between the conduction band of Cs2SnI6 and Fermi level of Pt NPs. (b) The electrochemical signals of the LSV curve only reflected a generalized HER activity of the whole composite, and the actual HER activity of located Pt NPs may be suppressed due to its low content. (c) For the theoretical calculation aspect, the provided model for Pt NPs was more like a Pt cluster, in which the underlying Ispecies could greatly affect the H* adsorption energy.

R2: Thanks for your valuable comments.
(a) Pt is considered as the best HER electrocatalysts. However, the reaction activity of Pt cocatalyst in photocatalytic reaction is also strongly determined by the interaction between photocatalyst surface and cocatalyst. The excellent activity of PtSA in our work is attributed to its special coordination structure since the PtSA is completely coordinated with the photocatalyst surface via three Pt-I bonds. This contributes to a strong interface between PtSA and Cs2SnI6, which is beneficial to the direct charge transfer from Cs2SnI6 to PtSA. However, at the PtNP-Cs2SnI6 interface, only the interfacial Pt atoms can be coordinated with Cs2SnI6 surface. Unfortunately, those surface catalytic Pt sites cannot be directly coordinated with the Cs2SnI6 surface, which undesirably increase the charge transfer distance or barrier. Thus the PtSA and PtNP with different cocatalyst-photocatalyst interactions lead to their different electronic and catalytic properties. This has been well explained in the Line 17-24 of Page 8.

(b)
The optimization on PtSA and PtNP have been done in Supplementary Fig. 11 and Supplementary Fig.  14. It is obvious that the activity of optimized PtNP-Cs2SnI6 with 3.88wt% Pt is still lower than that of PtSA-Cs2SnI6 with only 0.12wt% Pt. Hence, the difference between the activities of PtSA and PtNP on Cs2SnI6 cannot be simply attributed to their different Pt contents. Instead, the coordination structure of Pt species plays a significant role in the photocatalytic reaction. The detailed discussion has been added in the Line 3-8 of Page 7.
(c) Considering the size effect, a larger Pt nanoparticle consisting of 31 Pt atoms was used in the model, which also show a more negative hydrogen adsorption energy (-0.85 eV) than that of PtSA, as shown in Supplementary Table 3. This suggests that the PtSA on Cs2SnI6 owns a higher activity than Pt cluster or nanoparticle.
Q3: Furthermore, please add more details of the HER measurements in the experimental part, such as the used electrolyte and scanning rates. The description of the reference electrode (NHE) in the main text should be consistent with that in supplementary Fig. 18 (RHE).

R3:
Thanks for your valuable comments. The HER measurements were performed in the HI and H3PO2 mixed solution (57 wt% HI 16 mL + 50 wt% H3PO2 4 mL) at a scan rate of 50 mV s -1 . We have added these details in the revised supporting information (Please see the Line 11-14 of Page 3 of the revised supporting information). The overpotential is calculated based on the reversible hydrogen electrode (RHE), we have revised the calculation description in the revised manuscript (Please see Line 6-11 of Page 8).

To Reviewer 2:
In this work, the authors reported the decoration of all-inorganic Cs2SnI6 perovskite with single-atom Pt-I3 sites as cocatalyst (PtSA/Cs2SnI6) towards photocatalytic H2 evolution from HI splitting reaction. The as-prepared PtSA/Cs2SnI6 catalyst showed a turnover frequency (TOF) 176.5 times higher than Cs2SnI6 decorated with Pt nanoparticle, along with excellent catalytic stability. Spectroscopy investigations together with theoretical calculations reveals that both unique coordination structure and electronic property of Pt-I3 sites and strong metal-support interaction effect are the main reasons in achieving the greatly enhanced photocatalytic performance in hydrogen production from HI aqueous solution. The manuscript is well-organized and written and the results are interesting. I would suggest the acceptance of the work after the authors have addressed the following issues.

R:
Thanks for your great efforts in reviewing our manuscript. We specially appreciate your valuable comments and suggestions. We have performed all the experiments suggested by you, further addressed the comments point-by-point and made the corresponding changes accordingly in the revised manuscript.

Q1:
The PtNP/Cs2SnI6 was prepared by mixing pre-synthesized PtNP and Cs2SnI6 powder. In this case, the contact between PtNP and Cs2SnI6 powder should be poor, which will impair the interfacial charge transfer and therefore the photocatalytic activity. To make a fair comparison between PtNP/Cs2SnI6 and PtSA/Cs2SnI6, the PtNP/Cs2SnI6 sample could be prepared with photodeposition method to allow the direct growth of PtNP on Cs2SnI6. The photocatalytic activity of the as-prepared PtNP/Cs2SnI6 should be also tested.

R1:
Thanks for your valuable comment. Indeed, the interfacial contact can affect the charge transfer from Cs2SnI6 to PtNP. According to your nice suggestion, we have supplemented the photocatalytic activity of PtNPphoto/Cs2SnI6, prepared by the photo-deposition method. The rate of H2 evolution over the optimal PtNPphoto/Cs2SnI6 sample is 101 mol g -1 h -1 , which is better than that of PtNP/Cs2SnI6, but still much lower than that of PtSA/Cs2SnI6 sample. This has been added into the revised manuscript (Please see Line 8-13 of Page 7 of revised manuscript, Line 19-22 of Page 2 of revised supporting information and Supplementary Fig. 12). Supplementary Fig. 12. Photocatalytic H2 evolution rate of PtNPphoto/Cs2SnI6 depending on the loading amount of Pt.

Q2:
The XPS spectra suggest that the valence state of Pt element in the sample is Pt 2+ or Pt 4+ . However, as Pt 2+ or Pt 4+ can be easily reduced to Pt by photogenerated electrons during reaction, XPS analysis should be performed on the PtSA/Cs2SnI6 after photocatalytic reactions to allow the identification of the chemical state of Pt.

R2:
Thanks for your valuable comment. The positive-valence Pt species can accept the electron, and be reduced in the photocatalytic reaction. However, this is a dynamic process. The photogenerated electrons aggregated in Pt would be used to reduce protons into hydrogen. After photocatalytic reaction, the Pt species return to its initial chemical state. This has been demonstrated in the supplemented XPS test, in which the Pt 4f high resolution XPS spectrum of the sample after photocatalytic reaction is similar to that of the sample before photocatalytic reaction (Please see Line 23-25 of Page 7 and Supplementary Fig. 19).
Supplementary Fig. 19. The XPS spectra of Pt 4f in PtSA/Cs2SnI6 before and after photocatalytic reaction. Figure 4c, the conditions like the applied bias, the electrolyte (anything else besides 0.1 M TBAPF6 dichlormethane) for the photocurrent measurements should be provided. (b) Moreover, as PtSA is used as the HER catalyst in this work to promote reduction reaction. I'm curious why positive photocurrents instead of negative photocurrents are used to testify the role of PtSA. Intensity (a.u.)

Before reaction
After reaction a (a) The photocurrent responses were measured by utilizing 300W Xe lamp with an ultraviolet cut-off filter (λ≥420 nm) as the light source at 0 V vs. Ag/AgCl reference, and 0.1 M TBAPF6 dichlormethane solution as electrolyte. We have added these details in the revised supporting information (Please see the Line 11-14 of Page 3 of the revised supporting information).
(b) Cs2SnI6 is n-type semiconductor, which produces the positive photocurrent under present weak bias condition. The photocurrent test is mainly used to investigate the charge separation/transfer ability of photocatalyst instead of surface catalytic redox reaction. In our manuscript, the absolute value of current variation before and after illumination represents the carrier separation/transfer efficiency, which is influenced by the electronic properties of photocatalytic bulk and surface. The results show that the PtSA modification can significantly promote the charge separation/transfer efficiency between photocatalyst bulk and surface. Moreover, the surface catalytic redox reaction is mainly evaluated by the LSV test with negative current density as shown in Supplementary Fig. 20, which demonstrates the promotion effect of PtSA. This discussion has been added into the revised manuscript (please see Line 6-11 of Page 8).
Q4: Some typos or errors in the manuscript should be corrected (Reference 23, 24 et al.).

R4:
Thanks for your valuable comments. We have carefully checked and revised all typos and errors in the reference section in the revised manuscript.

To Reviewer 3:
Hui Chen et al. reports on the photocatalytic properties of a novel composite based on the Cs2SnI6 perovskite decorated with atomically dispersed Pt-I3 species. I have two main concerns about the manuscript. The first regards the possible appeal to a broad audience. In fact, the system used by the author, while possessing an improved stability in aqueous environment with respect to MAPbI3 and showing interesting properties, still implies the use of highly concentrated HI solutions, thus severely limiting the range of applicability, which is restricted (e.g. no hydrogen production in more desirable conditions). Considering that perovskites which are actually water-stable have started to be synthetized (cf. ,https://doi.org/10.1002/anie.202007584), I fear that, without any interesting and motivated interpretation of novel phenomena, the manuscript in the current form, would be more suitable for a specialized journal. Nevertheless, also in that case, there are some points that the authors need to address because the interpretation of their results appears to be shallow, in particular from the theory, and some results completely lack an explanation. R: Thanks for your great efforts in reviewing our manuscript. We specially appreciate your valuable comments and suggestions. We believe that the present work would be of great and instant importance to the following research fields: (1) Materials Science and Nanoscience. Designing the high-performance metal single atom photocatalytic catalysts is gaining dramatic attention. However, the common nitrogen or oxygen-coordinated metal single atom structure in the reported single atom catalysts still make them show the limited catalytic activity due to the high electronegativity of those ligand atoms. Herein, we firstly synthesize and demonstrate a new type of atomically dispersed iodine-coordinated Pt single atom structure (Pt-I3) with electron-rich feature, which is very promising for solar hydrogen photocatalytic production. Therefore, this work may raise research enthusiasm in the communities who are interested in single atom catalysis, materials science and nanoscience, etc.
(2) Hydrogen energy conversion. Though there are few reports on water-stable perovskite photocatalysts pointed by the reviewer, they can only use some high-cost and unstable organics (such as triethanolamine) as the hole sacrificial agent. Instead, HI acts as a carbon-free and safe hydrogen carrier, which attracts more and more attention (Energ. Environ. Sci. 2015, 8, 1484and Energ. Environ. Sci. 2009. However, those reported perovskite photocatalysts (Nature Energy, 2016, 2, 16185 andAdv. Mater. 2018, 30, 1704342) cannot stably work in HI solution. Considering those issues, we first demonstrate a HI-stable lead-free anti-dissolution environmentally friendly perovskite Cs2SnI6 photocatalyst with atomically dispersed Pt-I3 species, which achieves a stable and high activity (430 mol h -1 g -1 ) in a 24-hour HI-to-hydrogen test.
(3) Fundamental photocatalysis. By combining charge-carrier dynamics with theory calculations, this work reveals an origin for the dramatically boosted photocatalytic performance on PtSA/Cs2SnI6, which results from both unique coordination structure and electronic property of Pt-I3 sites, and strong metal-support interaction effect. This work opens a new pathway for stimulating more research on perovskite composite materials for efficient hydrogen production.
Moreover, we have performed all the experiments and theoretical calculations suggested by you, further addressed the comments point-by-point and made the corresponding changes accordingly in the revised manuscript.

Q1:
The computational details for the calculated results are frankly insufficient. The choice of adopted functional (which by the way does not include Van der Waals interactions that might play a role) is not motivated, convergence tests for the employed slab and the size of the vacuum layer are not reported.

R1:
Thanks for your valuable comment. In the initial calculations, we have considered the Van der Waals interactions, slab thickness, the vacuum thickness and supercell area in the convergence tests. The results demonstrate the accuracy of present settings. The addition of Van der Waals correction shows a limited change in the hydrogen adsorption calculation on all-inorganic Cs2SnI6 structure. Besides, the larger supercells do not produce obvious influence on the hydrogen adsorption calculation. Those parameters and results have added into the revised supporting information (Please see Line 11-13 of Page 11 and Supplementary R2: Thanks for your valuable comments. According to the previous report (Nat Mater 2016, 15, 1107-1112, an extra electron with compensating uniform background charge can be used to simulate the photogenerated electron. However, the extra electron with background charge probably leads to the undesirable effect on the electron transfer between Pt cocatalyst and photocatalyst. Hence, in order to avoid the presence of background charge in the region between adjacent slabs, a donor hydrogen atom was inserted into the bulk structure to calculate the charge density difference. The obtained charge density difference mapping also shows the localized distribution of electron in the PtSA region. Similarly, the electron is distributed on the whole PtNP. This has been added into in the revised manuscript (please see Line 17-27 of Page 8, Line 17-24 of Page 11 and Supplementary Fig. 21).
Supplementary Fig. 21. The charge density difference maps between PtSA/PtNP and Cs2SnI6: (a) PtNP/Cs2SnI6 and (b) PtSA/Cs2SnI6. The isosurface of charge density is 0.001 e Å -3 . The insets stand for the top view. The yellow region represents the additional electron distribution. An excess donor hydrogen atom was added into the models.

Q3: (a)
The authors report charge density differences for the pristine material and PtSA but they do not include any comparison with PtNP. Furthermore, the localization of the charge in the figure is not clear. (b) Are the author claiming that Cs2SnI6 feature a delocalized electronic state while, in presence of the additive, they observe a localized state in the band gap of the material?

R3:
Thanks for your valuable comments.
(a) Fig. 5a and 5b show the charge density differences between PtNP (PtSA) and Cs2SnI6, respectively. The electron is observed to be distributed on the whole PtNP and neighboring I sites, which undesirably decreases the electron density per Pt atom in PtNP. Instead, the electron in PtSA/Cs2SnI6 is only located between the PtSA and neighboring three I atoms. Thus, the electron density per Pt atom in PtNP is further lower than that of PtSA, which is considered to lead to the lower HER activity of PtNP. This explanation has been added into the revised manuscript (Please see Line 18-24 of Page 8).
(b) In the calculation of PDOS of PtNP-Cs2SnI6 and PtSA-Cs2SnI6 models, no extra electron was added. The computational details of PDOS have been added into the revised manuscript (please see the Line 15-17 of Page 11). The most localized Pt 5d states of PtSA and PtNP are observed in the band gap region of materials, which can accept the photogenerated electrons from the conduction band of materials.
Q4: I think that the claim that PDOS explains the difference between PtSA and PtNP is quite weak. I cannot see a dramatic difference among the two.

R4:
Thanks for your valuable comments. A dashed rectangle was added into Fig. 5c to label the difference between the localized Pt 5d states of PtSA and PtNP. As shown in the dashed rectangle region of Fig. 5c, the partial Pt 5d states of PtNP is above the Fermi level, which indicates its electron-deficient property. Instead, the most Pt 5d states of PtSA is below the Fermi level, indicates its electron-rich property. The higher electron density of PtSA also implies the stronger electron-captured ability, which contributes to the higher hydrogen production activity of PtSA. This discussion has been added into the revised manuscript (please see Line 1-5 of Page 9 and Fig. 5c).
Q5: How do the authors calculate the energy diagrams in Fig. 5d?

R5:
Thanks for your valuable comment. The reported standard hydrogen electrode (SHE) model (J Phys Chem B 2004,108, 17886-17892.) was adopted in the calculations of Gibbs free energy changes (ΔG) in hydrogen adsorption. The chemical potential of a proton-electron pair, µ(H + ) + µ(e − ), is equal to the half of the chemical potential of one gaseous hydrogen, 1/2µ(H2), at U = 0 V vs SHE at pH = 0. This has been added into the calculation details (please see Line 3-8 of Page 11).

Q6:
The author state that "instability of the organic component" is at the root of the limited application of perovskites in photocatalysis. However, recent reports show that organic-inorganic perovskite can be water-stable.

R6:
Thanks for your valuable comments. Though there are few reports (Nature Energy, 2016, 2, 16185 andAdv. Mater. 2018, 30, 1704342) on water-stable perovskites, they cannot stably work in HI solution. Here, we firstly report a HI-stable metal single atoms-modified lead-free perovskite photocatalyst for converting HI into hydrogen.

Q7:
The authors should comment on the trends observed for Pt loading? Why, after a maximum, the photocatalytic activity decreases?

R7:
Thanks for your valuable comments. When the content of Pt loading on Cs2SnI6 surface was higher or lower than 0.12wt%, the photocatalytic activities was decreased. This is considered from the reason that the excessive Pt species reduces the light absorption of Cs2SnI6 due to the shading effect. In contrast, the insufficient Pt species cannot provide the rich H2-releasing active sites. We have supplemented the explanation of catalytic activity trend with the increase of Pt loading in the revised manuscript. (Please see Line 3-8 of Page 7).
Minor points:

Q1:
The authors should specify that the band alignment achieved via XPS is an approximation, since it does not include effects of the water-perovskite interface in the band alignment.

R1:
Thanks for your valuable comments. We have specified that the band alignment is just an approximation since the complicated electrolyte-perovskite interface effects were not considered (Please see Line 11-12 of Page 5).
Q2: "The most 5d states of PtSA are below the Fermi level, indicating its strong electron-captured ability" this sentence is not clear and should be rephrased.

R2:
Thanks for your valuable comment. The electronic states below Fermi level are filled with electrons. According to Fig. 5c, the most 5d states of PtSA are below the Fermi level, indicating its electron-rich state.
In contrast, the partial 5d states of PtNP are above the Fermi level, indicating its electron-deficient property. This means that the PtSA species owns a stronger ability for capturing electrons from the Cs2SnI6, which contributes to the higher hydrogen production activity of PtSA/Cs2SnI6 in the above photocatalytic experiments. This explanation has been added into the revised manuscript (Please see Line 1-5 of Page 9).
Overall, on the basis of my comments, I cannot suggest publication of the manuscript in the present form which needs to be profoundly revised in order to explain the results.