Electronic metal–support interaction modulates single-atom platinum catalysis for hydrogen evolution reaction

Tuning metal–support interaction has been considered as an effective approach to modulate the electronic structure and catalytic activity of supported metal catalysts. At the atomic level, the understanding of the structure–activity relationship still remains obscure in heterogeneous catalysis, such as the conversion of water (alkaline) or hydronium ions (acid) to hydrogen (hydrogen evolution reaction, HER). Here, we reveal that the fine control over the oxidation states of single-atom Pt catalysts through electronic metal–support interaction significantly modulates the catalytic activities in either acidic or alkaline HER. Combined with detailed spectroscopic and electrochemical characterizations, the structure–activity relationship is established by correlating the acidic/alkaline HER activity with the average oxidation state of single-atom Pt and the Pt–H/Pt–OH interaction. This study sheds light on the atomic-level mechanistic understanding of acidic and alkaline HER, and further provides guidelines for the rational design of high-performance single-atom catalysts.

1. In Fig. 4a are shown positions of d-bands from UPS valence-band spectra. The authors should include DOS projected on Pt-d orbitals and the d-states occupation numbers, calculated using DFT. Table 3 of the SI are listed Pt binding energies. Even more useful information would be the energy barriers for Pt diffusion between the nearest adsorption sites. The calculated activation energies will demonstrate the structural stability of the studied single-atom catalysts. This is an important issue, studied by Li et al. (ACS Nano 11, 3392 (2017)), where they claimed that Pt atoms at the ideal 2H-MoS2 were very mobile in the room temperature and only those trapped at the S-vacancies were structurally stable. In the present study, one of the supports was 1T phase of MoS2, so the picture might be different from 2H-MoS2, but the authors should clarify this by carrying out the corresponding DFT calculations.

Responses to Reviewers
Reviewer #1: In this manuscript, the authors reported four kinds of single-atom Pt supported on different transition metal dichalcogenides (MoS2, WS2, MoSe2, and WSe2) as efficient electrocatalysts for HER. Fundamental morphology, structure, catalytic performance and stability of the material in both alkaline and acidic media were studied. Detailed spectroscopic and electrochemical characterizations showed that the fine tailoring of the oxidation states of single-atom Pt catalysts through EMSI significantly could modulate the catalytic activities in either acidic or alkaline HER. The authors also conducted different experimental measurements to study the HER intermediates (H* and OH*), and revealed the structure-activity relationship by UPS and DFT calculations.
After reading this manuscript, I suggest this manuscript should be resubmitted after addressing the following issues.
Response: Thank you for your summary. We appreciate your efforts in reviewing our manuscript. We have revised the manuscript accordingly. Our point-by-point responses are detailed below.
1. In the Introduction, the author should give explanation about why TMD was chosen as the support of single atoms. Besides, it is well known that transition metal dichalcogenides as semiconductors possess novel band structure. Their conductivity can be effectively tuned by heteroatom doping. In this study, the incorporation of Pt atom into MoS2, MoSe2, WS2 and WSe2 substrates may induce significant change in the conductivity, which plays a key role in the electrocatalytic process. This study should not attribute the catalytic performance enhancement to the modulation of platinum before confirming the Pt incorporation influence on the dichalcogenides substrates.
Response: We thank this reviewer for the valuable comments and concern. To highlight our design concept and make the current work more precise, we have made the following revisions and discussion: (1) explanation of why TMD was chosen as the support of single atoms in the Introduction (paragraph 4, Fig. 1a): "Transition metal dichalcogenides (TMDs) have been widely used as the supports for immobilizing SAMCs in heterogeneous catalysis [Chem. Rev. 2019, 119, 1806Chem. Rev. 2020, 120, 11810;Chem 2020, 6, 885;Energy Environ. Sci., 2015, 8, 1594Nat. Chem. 2017, 9, 810;Nat. Commun., 2019, 10, 5231]. Compared with SAMCs supported on carbon-based materials [Angew. Chem. Int. Ed., 2020, DOI: 10.1002J. Am. Chem. Soc. 2020, 142, 8431;J. Am. Chem. Soc. 2019, 141, 20118], the electronic structure of single-atom metals supported on TMDs is usually adjusted by both the anchoring atom and the neighboring transition metal atoms with relatively high atomic number, which affords a more flexible and complex coordination environment to regulate the catalytic activity [Nano Res., 2020, 13, 1842Adv. Mater. 2020, 32, 2003300]. Owing to the various well-defined band structures of TMDs (e.g. MoS2, WS2, MoSe2, and WSe2, Fig. 1a), the core anchoring chalcogen (S, Se) and the neighboring transition metal (Mo, W) can synergistically regulate the electronic structure of SAMC through EMSI. The tuneable d-orbital state of singleatom Pt changes the adsorption energy of reactants on metal atoms and thus influences the catalytic activity of HER (Fig. 1a)." R-3 Left: schematic structure of single-atom Pt on TMDs material. The grey, purple, and green spheres represent the chalcogen (sulphur/selenium), transition metal (molybdenum/tungsten), and platinum, respectively. The electronic structure of singleatom Pt was modulated by two-dimensional TMDs through charge delocalization, enabling the single-atom Pt to take slightly positive charge (Pt + ). The structural unit of single-atom Pt was circled by the orange dashed line and further enlarged above. Top right: schematic diagram of the band edges of TMDs. The conduction band minimum R-4 (CBM)/valence band maximum (VBM) band edges of TMDs (theoretical values) refer to ref. 42 [Chem. Soc. Rev. 2018, 47, 6845]. The schematic band structure-showing the electron affinity and ionization potential of various TMDs-provides a guideline for rationalizing the EMSI modulation of single-atom Pt. Bottom right: schematic illustrating that the d-state shift of single-atom Pt induced by EMSI regulates the catalytic performance of HER.
(2) We agree with the reviewer that heteroatom doping effectively increases the conductivity of substrate and further enhances the catalytic activity [e.g. J. Am. Chem. Soc. 2017, 139, 15479;Adv. Mater. 2013, 25, 5807;Chem. Sci. 2011, 2, 1262Energy Environ. Sci. 2015, 8, 1594. However, the structure in our work is protrusion-shaped single-atom Pt straddled atop Mo/W, which is different from the single-atom Pt incorporation into TMD lattice [for example, Energy Environ. Sci. 2015, 8, 1594 (detailed discussion and comparison shown below). The catalytically active Pt is adsorbed atop TMDs and can expose more unsaturated active sites to reactive species, which reduces the influence of TMDs support for HER enhancement.
We further conducted the thiocyanate ion (SCN − ) poison experiment (Supplementary Fig. 26), which suggested that the HER activity dominantly derives from the singleatom Pt and thus the catalytic performance enhancement is mainly attributed to the EMSI modulation of Pt.  Environ. Sci. 2015, 8, 1594. Additionally, the Tafel behavior of Pt-SAs/TMDs in acidic HER (~30 mV dec −1 ) resembles that of the commercial Pt, indicating that the catalytic reaction on single-atom Pt contributed mostly to the HER. On the contrary, it has been reported that Pt-doped MoS2 showed a Tafel slope of 96 mV dec −1 , close to that of pure MoS2 (~100 mV dec −1 ) [Energy Environ. Sci. 2015, 8, 1594."

Supplementary Note 3. Comparison of single-atom Pt adsorbed atop MoS2 and Ptdoped MoS2 cases
In order to elucidate the advantage of single-atom Pt supported atop TMDs (Pt-SAs/TMDs in this work) for the atomic-level electronic modulation for HER enhancement, we further compared the adsorbed single-atom Pt with the previously reported Pt-doped MoS2 material [Energy Environ. Sci., 2015, 8, 1594]. In our system, we can see that single-atom Pt adsorbed atop MoS2 (Pt-SAs/MoS2) was dramatically poisoned by SCN − owing to the blocking of active Pt sites, resulting in a large decrease of HER current approaching zero ( Supplementary Fig. 26). Similarly, Bao et al. also found that the activity of Pt adsorbed on MoS2 drops quickly after adding methanol because the exposed Pt atoms atop MoS2 support were easily poisoned by methanol ( Fig. R1). On the contrary, Pt-doped MoS2 showed negligible current decrease and excellent poisoning resistance ability with the introduction of methanol (Fig. R1    Commun. 2019, 10, 4500]. Specifically, the differential XANES (ΔXANES) spectra were achieved by subtracting the spectra of Pt-SAs/TMDs from that of Pt foil reference ( Supplementary Fig. 12). Owing to the linear relationship between the white-line area R-12 and the oxidation states/formal d-band hole counts, these two parameters of single-atom Pt can be fitted by correlating the ΔXANES area of the references (Pt foil and PtO2) and the single-atom Pt catalysts. For example, the formal d-band hole count was calculated based on the slope of 1.166 unit area per d-band hole obtained from Pt 0 foil (5d 9 6s 1 ) and Pt IV O2 (5d 6 6s 0 ) standards [Nat. Commun. 2020, 11, 1029." (4) To further confirm that our data are within a reasonable range for HER activity, we also listed the previously reported HER activity in many literatures for double-check:   Fig. 19). The thiocyanate ions (SCN − ) poison experiment further confirmed that the HER activity dominantly R-18 derives from the single-atom Pt and the catalytic performance enhancement is mainly attributed to the EMSI modulation of Pt (see Question 1 or Supplementary Fig. 26).

Supplementary
Detailed discussion has also been provided (see Question 1 or Supplementary Note 3) to compare the single-atom Pt adsorbed atop TMDs with the Pt-doped TMDs case. The active sites of our system are Pt atoms whereas the active sites of Pt-doped case are S atoms [Energy Environ. Sci., 2015, 8, 1594. All these results indicate that single-atom Pt plays a central role in the HER, implying that TMDs have less effect on the construction of the atomic-level structure-activity relationship.
We further correlated the loading amount of single-atom Pt (for example, Pt-SAs/MoS2) with the HER activity. A nearly linear correlation (Fig. R4) corroborated that singleatom Pt contributed mostly to the HER, which rationalizes the EMSI modulation of our single-atom system for the mechanistic investigation of structure-activity relationship.   Fig. 28). CV measurements under non-catalytic conditions were usually used to measure the Pt-H adsorption/desorption and Pt-O formation/reduction. Unfortunately, these characteristic peaks of Pt cannot be detected on single-atom Pt ( Supplementary Fig. 8 Fig. 23).

Supplementary
4. More discussion is needed to emphasize the advantages of the EMSI effect in the structure and activity relationship section.

Response:
We thank the reviewer for this constructive and valuable advice. We have added more discussion to emphasize the advantages of the EMSI effect in the structure and activity relationship section (Discussion section in the manuscript): "In this work, we demonstrate that EMSI modulation of single-atom Pt significantly regulates the HER activity over a wide pH range, and systematically unravel the relationship between oxidation state and HER activity of single-atom Pt. The EMSIacting as a bridge between electronic study and catalyst design-provides a detailed explanation of the enhanced properties of supported catalysts at the electronic scale.
With the length scale of catalysts shrinking to atomic level, the EMSI effect becomes stronger and can predominate the reaction rate [Adv. Mater. 2020, 32, 2003300;Nano Res., 2020, 13, 1842.

R-27
Reviewer #3: Although the studied systems and results are very interesting and the methodology combined approaches able to illuminate a microscopic picture of investigated catalysts, unfortunately, I can not recommend the publication of the paper due to significant similarity to the Ref. 31 -"Site-specific electrodeposition enables self-terminating growth of atomically dispersed metal catalysts" -recently published work of several co-authors of the present manuscript (Nature Communication 11, 4558 (2020)).
In Ref. 31, the authors mainly studied Pt atoms at MoS2 but also included results for Pt at WS2 and MoSe2, as well as a few other metals at MoS2. In this paper, they focused on Pt-based catalysts. In addition to three substrates considered in Ref. 31, WSe2 was included as the 4th support.
In the present paper the main results, highlighted in the abstract, are increased activity of Pt1/WS2 and Pt1/MoSe2, compared to Pt/C and the correlation of the activity and the Pt interaction with the support. However, the activity of several of the inspected catalysts was already compared in Fig. 24, Supported Information (SI) of Ref. 31. In Table 4, SI of Ref. 31 were also presented DFT results for the binding energies of Pt adatom at MoS2, WS2 and MoSe2 (these results are partially presented again in Table   3, the SI of the present manuscript).
Due to apparent similarity in studied systems, methodology, and results, the present manuscript could be considered as a follow up to Ref. 31, and thus lacks sufficient novelty compulsory for the publication in Nature Communications.

Response:
We thank the reviewer for the comments. We found that we didn't clearly convey the information that precisely distinguish our current work with previous work described in ref. 31. We provided a detailed explanation here and revised our manuscript accordingly.

R-28
The previous work (Ref. 31) developed a site-specific electrodeposition method for synthesis of single-atom materials and focused on the growth mechanism of singleatom metals. Different from the main idea of previous work (single-atom growth), the aim of the current work is to unveil the mechanism of wide-pH-range HER and mechanistic structure-activity relationship for catalysts design. We used many different advanced spectroscopies (UHV-XPS, UPS, XAS and NAP-XPS) and electrochemical techniques to characterize the EMSI-induced changes in the electronic structure of single-atom and its corresponding alkaline and acidic HER activities.
Our current work used experimental techniques to explore and quantify the EMSI effects on the wide-pH-range HER at the atomic level, which was realized by correlating the acidic/alkaline HER activities with the average oxidation state of

single-atom Pt and the Pt-H/Pt-OH interaction.
These results help to understand the HER mechanism and satisfies the high demand for deep understanding of structureactivity relationship. We show that one descriptor (H*) governs the acidic HER while two descriptors (OH* and H*) co-determined the rate of alkaline HER, and the structure-activity relationship can be used as a universal guideline for the design of single-atom metal catalysts.
In view of the reviewer's concern, to better highlight the concept of the current work, we revised the paper from the following aspects: (1) We have revised the abstract and emphasized the EMSI modulation to avoid the confusion of the main results: "Here, we reveal that the fine control over the oxidation states of single-atom Pt catalysts through electronic metal-support interaction significantly modulates the catalytic activities in either acidic or alkaline HER".
(2) We have also cited the ref. 31 to the investigation of acidic HER, which makes our description more precise. "Similarly, we also showed the various acidic HER activities R-29 on different supported single-atom Pt samples, which followed the order distinctive from that in alkaline HER: Pt-SAs/WS2 > Pt-SAs/MoS2 > Pt-SAs/MoSe2 > Pt-SAs/WSe2 ( Fig. 3e and 3f), in consistence with the previously reported results [Nat.
Commun. 2020, 11, 4558]." "From the structure-activity relationship, the fine control over the oxidation state of single-atom Pt catalysts enables to achieve the optimal catalytic activity in either acidic or alkaline HER. In acidic environment, the HER performance could be well optimized through properly decreasing the oxidation state of single-atom Pt (Fig. 5a), accelerating the hydrogen desorption process. Similar to our finding, Liang's and Yao's groups have recently reported that the electron-enriched or near-zero-valence atomically dispersed Pt species are more active than the high-valence single-atom Pt for catalyzing acidic HER [Nat. Commun. 2020, 11, 1029Nat. Commun. 2019, 10, 4977]. In alkaline environment, as the oxidation state of single-atom Pt increases, the HER activity also increases initially (left side of the volcano plot, Fig. 4e). Under the electrochemical condition, the charge of the metal plays a decisive role with regard to water dissociation [Nat. Energ. 2017, 2, 17031;J. Electroanal. Chem. 1996, 411, 95]. Choi's group reported the counterintuitive promoting effect of CO molecule on alkaline HER of single-atom Pt [J. Am. Chem. Soc. 2018, 140, 16198], which could be also explained by our proposed model (Fig. 5b). After the coordination of CO (strong electron acceptor), metal-to-ligand charge transfer from Pt was increased, endowing the single-  Am. Chem. Soc. 2018, 140, 16198;Nat. Mater. 2016, 15, 197].
We further suggest that single-atom Pt with optimal valence state is simultaneously favourable for water dissociation, adsorption/desorption of OH* and H* (Fig. 5b, and R-31 Supplementary Fig. 34). It should be noted that with the increased oxidation state of single-atom Pt, the adsorption of H* and OH* on the catalyst were both strengthened.
Although single-atom Pt with high oxidation state energetically favors the adsorption of electron-rich H2O and OH*, too strong hydrogen adsorption will also lead to the slow release of active sites and hence the sluggish HER kinetics. From a fundamental point of view, the present work shows the atomic-level enhancement of HER thermodynamically and kinetically by optimizing the catalyst-H interaction and accelerating the water dissociation. The single-atom Pt catalyst with optimal oxidation state (~+2), showed neither too strong Pt-H interaction to release hydrogen, nor too weak Pt-OH interaction to dissociate water, which dramatically contributes to the overall alkaline HER." (5) More description and discussion about the debate on HER mechanism have been added (Discussion section in Manuscript): "To date, three widely adopted theories have emerged to explain the alkaline HER mechanism: water dissociation theory (hydroxyl binding energy, OHBE), hydrogen binding energy (HBE) theory, or interface water and/or anion transfer theory [Mater.
Today 2020,36,125]. It still remains unclear which descriptor governs the alkaline HER. At the atomic level, our results show that the two descriptors (OH* and H*) codetermined the rate of alkaline HER (Fig. 5b), which sheds light on the long-standing puzzle about HER mechanism." The present manuscript, which I consider as a nice work, could be improved by including a better description of the Pt interaction with TMDs. The effect of the support should be presented in more detail. These details can be easily obtained from DFT < Pt-SAs/MoS2 < Pt-SAs/MoSe2 < Pt-SAs/WSe2, whose overall trend is consistent with the theoretical results. The only discrepancy lies in the higher position of Pt-SAs/MoS2.
We considered that the discrepancy is mainly caused by the approximations adopted in our calculations for achieving the compromise between accuracy and affordable computational cost. First, the TMD systems used in calculation are simplified to the perfect surface and different from the real situation in the experimental study. Second, the dipole of the whole system is not considered owing to that only one Pt atom was involved in the calculation. Nevertheless, the DFT-calculated results still rationalize the experimental trends (Pt-SAs/WS2 or Pt-SAs/MoS2 < Pt-SAs/MoSe2 < Pt-SAs/WSe2), which is of great reference value to our work.

R-33
Supplementary Figure 28. Density of states projected on Pt-d orbitals of Pt-SAs/TMDs. The white dashed lines represent the d-band center calculated by DFT. Table 3 Fig. 16)".

R-34
We have calculated the activation energy of the Pt atom diffusion on the 1T-MoS2 along two directions. The energy barrier of the Pt atom diffusion between the nearest adsorption sites is 1.0-1.1 eV ( Supplementary Fig. 16 Reviewer #2 (Remarks to the Author): The revised version can be accepted.
Reviewer #3 (Remarks to the Author): I am very pleased with the authors' answer to my comments and the corresponding changes brought to the manuscript. In the revised paper and the response to my comments, they emphasized differences between this work and former Ref. 31 (now Ref. 43), particularly detailed description of the mechanism behind favorable catalytic properties of the single-atom metal catalysts over a wide-pH-range. Since they applied the same methods in two papers, it is difficult to avoid the similarity in the Figures with results, and it is me who did not see this very clear when reading the original version of the paper.
Therefore, I recommend the publication of the paper in the present form.
A minor comment to Figure 4c and 4e: The y-axis in the graph in Figure 4c,d labeled as "H/OH adsorption ability" should include a quantity that can be measured or calculated -for example, d-band center. The "adsorption ability" is a loose term.

REVIEWERS' COMMENTS:
Reviewer #1 (Remarks to the Author): I am satisfied with the revision, and it can be accepted at the present state.
Response: We thank the reviewer for his/her positive comments.
Reviewer #2 (Remarks to the Author): The revised version can be accepted.
Response: We thank the reviewer for his/her positive comments.
Reviewer #3 (Remarks to the Author): I am very pleased with the authors' answer to my comments and the corresponding changes brought to the manuscript. In the revised paper and the response to my comments, they emphasized differences between this work and former Ref. 31 (now Ref. 43), particularly detailed description of the mechanism behind favorable catalytic properties of the single-atom metal catalysts over a wide-pH-range. Since they applied the same methods in two papers, it is difficult to avoid the similarity in the Figures with results, and it is me who did not see this very clear when reading the original version of the paper. Therefore, I recommend the publication of the paper in the present form.
A minor comment to Figure 4c and 4e: The y-axis in the graph in Figure 4c,d labeled as "H/OH adsorption ability" should include a quantity that can be measured or calculated -for example, d-band center. The "adsorption ability" is a loose term.
Response: We thank the reviewer for this minor comment. The H adsorption ability is quantified by the position of the d-band center (Fig. 4a), while the Pt-OH interaction is quantified by the CO oxidation potentials obtained from CO stripping voltammetry (Fig. 4d).
We have highlighted the relevant description in the Figure legend. The scanning potential value for Pt-SAs/TMDs samples could only reach as high as 0.7 V, since higher potential resulted in the oxidation of ce-TMDs ( Supplementary Fig. 36). (e) Relationship of average oxidation state, Pt-OH interaction and alkaline HER activity of Pt-SAs/TMDs. The circle and triangle are the average oxidation states of single-atom Pt obtained from XPS and XANES, respectively. The Pt-OH interaction is quantified by the CO oxidation potentials obtained from CO stripping voltammetry.