Pt-O bond as an active site superior to Pt0 in hydrogen evolution reaction

The oxidized platinum (Pt) can exhibit better electrocatalytic activity than metallic Pt0 in the hydrogen evolution reaction (HER), which has aroused great interest in exploring the role of oxygen in Pt-based catalysts. Herein, we select two structurally well-defined polyoxometalates Na5[H3Pt(IV)W6O24] (PtW6O24) and Na3K5[Pt(II)2(W5O18)2] (Pt2(W5O18)2) as the platinum oxide model to investigate the HER performance. Electrocatalytic experiments show the mass activities of PtW6O24/C and Pt2(W5O18)2/C are 20.175 A mg−1 and 10.976 A mg−1 at 77 mV, respectively, which are better than that of commercial 20% Pt/C (0.398 A mg−1). The in situ synchrotron radiation experiments and DFT calculations suggest that the elongated Pt-O bond acts as the active site during the HER process, which can accelerate the coupling of proton and electron and the rapid release of H2. This work complements the knowledge boundary of Pt-based electrocatalytic HER, and suggests another way to update the state-of-the-art electrocatalyst.

The manuscript submitted by Li et al report on the remarkable HER efficiency of two platinumcontaining polyoxometalates (POMs), abbreviated PtW6O24 and Pt2(W5O18)2. In a general point of view, the presented work appears well-done and highly convincing. The authors demonstrate that containing-platinum POMs as soluble oxide analogues are able to mimic the platinum oxide behavior toward hydrogen evolution giving highly efficient HER catalysts. The manuscript proposes a multiscale characterization of the catalyst using electrochemistry, XPS, X-ray absorption spectroscopy (XANES and EXAFS), TEM and STEM, vibrational spectroscopies (Raman and infrared, impedance spectroscopy…. Furthermore, experimental data were supported by theoretical calculation at the DFT level, thus giving a set of reliable arguments consistent with the hypotheses. Undoubtely, these two Pt-containing POMs are stable, highly processable, very efficient and resistant against usual Pt-contaminant. This submitted manuscript could display the requested scientific quality to be published in NatureCOMM. The first critical point of the MS is certainly the repeated comparison of the HER performances of Pt-POM/catalysts with those of the commercial 20% Pt/C. The sentence "better than that of commercial 20% Pt/C" appears at least twelve times in the text. The reader understand that the commercial 20% Pt/C is a very bad HER catalyst, with not optimized Pt-dispersion and probably containing platinum-contaminant agents. Furthermore, the procedures of preparation of both type of catalyst are necessarily different. For instance, the nature of the used carbon is different, such as Ketjen carbon for Pt-POM/catalysts leading to important changes of the surface area and conductivity. Actually, this is easy to get a poorly HER active platinum-based catalyst. This point must be corrected and the highest HER performances of the two Pt-POMs/C should be commented objectively and scientifically. Furthermore, the paper is divided in two redundancy part dealing with the HER properties of the PtW6O24/C, and those of the Pt2(W5O18)2/C, respectively. The first part appears very interesting, dealing with the electrochemical behavior of the Pt(IV) center embedded within the {W6O24} framework. Calculations are consistent with the expected quasi-square Pt(II) center, while the HER process resulted from the presence of Pt(I)-H hydrid group. The second part (Pt2(W5O18)2/C) brings nothing new and should be deleted. The authors did not give any arguments about the selection of this Pt2(W5O18)2 POM species. Less minor point : In the "Methods" section, the synthetic procedure of the two Pt-POMs/C, given in the text is not consistent with the scheme 1 (supp Info). In the text Nafion is used while in the supp Info, Melamine-Formaldehyde is added. This should be cleared.
In conclusion, this paper should disserve publication in NatureCOMM, therefore some revisions are still needed.

Reviewer 1
The authors convincingly show that (in acetonitrile) Pt(IV) in PtW 6 O 24 is reduced to Pt(II) after which the third electron reduces W(VI). As such, catalysis in aqueous acid is reasonably due to Pt(II) (and perhaps Pt(I) as shown later), rather than to Pt(0). This is also consistent with reductions of numerous other transition-metal (TM) substituted heteropolytungstates, in which reduction of bound TM cations to their metallic state does not occur due to more favorable reduction of W(VI). Stability to repeated cycling and resistance to poisoning by added TM cations (referred to by the authors as "anti-toxicity") further support the presence of a redox-active molecular catalyst, as opposed to the reductive degradation of the complexes and formation of Pt(0) clusters. Findings from DFT calculations argue that the kinetically competent species responsible for H 2 evolution actually involves reduction of the W(VI)-based LUMO by two electrons. This is not only reasonable, but quite interesting in its implications for further catalyst design. Notably, quite similar results are obtained for the Pt(II) complex (Pt 2 (W 5 O 18 ) 2 ). In summary, the key finding is that the partial reduction of Pt cations bound via oxide ligands to reduced W ions within molecular heteropolytungstate clusters are more effective than Pt(0), and comparable to some of the most effective reported in the literature (wrt overpotential at 10 mA cm^-1 ) in the hydrogen evolution reaction. Moreover, the W-based clusters appear to serve as stable, redox catalysts, consistent with reversible reduction / reoxidation under turnover conditions. This an interesting, clearly presented, high-quality and significant work, potentially suitable for publication in Nature Comm.

Our response:
We do thank your positive comments.

Comment 1:
Before publication, the authors should provide definitive evidence that the molecular complexes are indeed stable under turnover conditions. One way to do this would be to run bulk-electrolytic H 2 productions reactions using solutions of the complexes in sulfuric acid (as done to obtain the CV data in water). After a reasonable number of turnovers, the complexes in the solution can be characterized, for example by electrospray-ionization mass spectroscopy (perhaps after precipitation from water using quaternary ammonium cation salts, and redissolution in MeCN). It might be necessary to first oxidize the reduced complexes by addition of Br 2 , to avoid electron transfer to O 2 and the formation of hydroxide (which could cause hydrolytic degradation). This simple experiment would an intriguing dimension to the work, and increase it's generality and significance. Other approaches would also be acceptable; the main point is to demonstrate that the complexes behave as reversible redox-active catalysts.

Our response:
According to the suggestion, the electrospray-ionization mass spectra (EIS) of PtW 6 O 22 compound (before and after HER in 0.5 M H 2 SO 4 aqueous solution as well as its re-oxidized species by Br 2 ) were measured by precipitating the polyoxoanion with tetrabutylammonium (TBA) bromide and dissolved in CH 3 CN (see Supplementary Figure 7  Furthermore, the capillary electrophoresis have also been used to monitor this process (see Supplementary Figure 8). The retention time of PtW 6 O 24 before HER and its re-oxidized species by Br 2 after HER almost unchanged, further confirming that PtW 6 O 24 compound is a reversible redox-active catalyst. The retention time of PtW 6 O 24 before HER and its re-oxidized species by Br 2 after HER almost unchanged, further confirming that PtW 6 O 24 compound is a reversible redox-active catalyst.
We made the modification in the supporting information as shown below.  (pH=3) before HER (a), and its re-oxidized species by Br 2 after HER (b). The retention time of PtW 6 O 24 before HER and its re-oxidized species by Br 2 after HER almost unchanged, further confirming that PtW 6 O 24 compound is a reversible redox-active catalyst. (see Supplementary Information pages 4-5)

Reviewer 2
Revision after more data is provided.
This reviewer would like to get access to the DFT computed geometries and frequencies of all species included in the manuscript, particularly those corresponding to the Tafel-like H-H forming transition state. The barrier reported is very low, and has to be verified. Normally, this data is included in the Supplementary Information section, but it is missing here. Alternatively, the authors may upload input and output files to any digital repository and provide links to access the data.

Minor questions:
The version of the program used needs to be properly mentioned, as well as properly cited.

Our response:
Now a data set of computational results in the manuscript is available in the ioChem-BD repository and can be accessed via https://doi.org/10.19061/iochem-bd-6-27. The optimized geometries are also provided as the Supplementary Information. The reviewer could check the Tafel-like H-H forming transition state from the above link which named as TS-4e-4H. The transition states (TS) have been confirmed by the existence of only one imaginary frequency along the reaction coordinate and intrinsic reaction coordinates (IRC) calculations which indeed connect the right reactants and products (Supplementary Figure 56). Finally, the right version of the program has been included and cited in the section of method "DFT computational details".
We made the corresponding modification in the supporting information as shown below. We also made the corresponding modification in the manuscript as shown below.
DFT computational details. All calculations were performed through the facilities provided by the Gaussian09 package (revision D.01) 41 . Geometry optimizations for all intermediates and transition states were carried out at the M06 level without symmetry restrictions 42 . The LANL2DZ basis set was employed for the Pt and W, whereas the 6-31G(d, p) basis set was used for the O and H 43,44 . To confirm the stability of all structures, frequency calculations were performed at the same level as optimization. The transition states (TS) were confirmed by the existence of only one imaginary frequency along the reaction coordinate and intrinsic reaction coordinates (IRC) calculations which indeed connect the right reactants and products ( Supplementary Fig. S56)

Reviewer 3
The manuscript submitted by Li et al report on the remarkable HER efficiency of two platinum-containing polyoxometalates (POMs), abbreviated PtW 6 O 24 and Pt 2 (W 5 O 18 ) 2 . In a general point of view, the presented work appears well-done and highly convincing. The authors demonstrate that containing-platinum POMs as soluble oxide analogues are able to mimic the platinum oxide behavior toward hydrogen evolution giving highly efficient HER catalysts. The manuscript proposes a multi-scale characterization of the catalyst using electrochemistry, XPS, X-ray absorption spectroscopy (XANES and EXAFS), TEM and STEM, vibrational spectroscopies (Raman and infrared, impedance spectroscopy…. Furthermore, experimental data were supported by theoretical calculation at the DFT level, thus giving a set of reliable arguments consistent with the hypotheses. Undoubtedly, these two Pt-containing POMs are stable, highly processable, very efficient and resistant against usual Pt-contaminant. This submitted manuscript could display the requested scientific quality to be published in NatureCOMM. The first critical point of the MS is certainly the repeated comparison of the HER performances of Pt-POM/catalysts with those of the commercial 20% Pt/C. The sentence "better than that of commercial 20% Pt/C" appears at least twelve times in the text. The reader understand that the commercial 20% Pt/C is a very bad HER catalyst, with not optimized Pt-dispersion and probably containing platinum-contaminant agents. Furthermore, the procedures of preparation of both type of catalyst are necessarily different. For instance, the nature of the used carbon is different, such as Ketjen carbon for Pt-POM/catalysts leading to important changes of the surface area and conductivity. Actually, this is easy to get a poorly HER active platinum-based catalyst. This point must be corrected and the highest HER performances of the two Pt-POMs/C should be commented objectively and scientifically. Furthermore, the paper is divided in two redundancy part dealing with the HER properties of the PtW 6 O 24 /C, and those of the Pt 2 (W 5 O 18 ) 2 /C, respectively. The first part appears very interesting, dealing with the electrochemical behavior of the Pt(IV) center embedded within the {W 6 O 24 } framework. Calculations are consistent with the expected quasi-square Pt(II) center, while the HER process resulted from the presence of Pt(I)-H hydrid group. The second part (Pt 2 (W 5 O 18 ) 2 /C) brings nothing new and should be deleted. The authors did not give any arguments about the selection of this Pt 2 (W 5 O 18 ) 2 POM species. Less minor point: In the "Methods" section, the synthetic procedure of the two Pt-POMs/C, given in the text is not consistent with the scheme 1 (supp Info). In the text Nafion is used while in the supp Info, Melamine-Formaldehyde is added. This should be cleared.
In conclusion, this paper should disserve publication in NatureCOMM, therefore some revisions are still needed.

Our response:
We do thank your positive comments.

Comment 1:
The first critical point of the MS is certainly the repeated comparison of the HER performances of Pt-POM/catalysts with those of the commercial 20% Pt/C. The sentence "better than that of commercial 20% Pt/C" appears at least twelve times in the text. The reader understand that the commercial 20% Pt/C is a very bad HER catalyst, with not optimized Pt-dispersion and probably containing platinum-contaminant agents. Furthermore, the procedures of preparation of both type of catalyst are necessarily different. For instance, the nature of the used carbon is different, such as Ketjen carbon for Pt-POM/catalysts leading to important changes of the surface area and conductivity. Actually, this is easy to get a poorly HER active platinum-based catalyst. This point must be corrected and the highest HER performances of the two Pt-POMs/C should be commented objectively and scientifically.

Our response:
We agree with the reviewer's comments. Although commercial 20% Pt/C is widely used as the reference electrocatalyst for studying HER reactions, its catalytic activity is already not the best as the reviewer commented. No matter in size, morphology, or dispersion of metal Pt-based electrocatalysts on various loading materials, there still has enough space to improve its catalytic activity, which has been proved by a number of literature reports in recent five years 9-14，39，40 . It also becomes an important driving force for scientists to continue developing new Pt-based electrocatalysts. Therefore, "PtW 6 and Pt 2 W 10 are more active than commercial Pt/C", which was over emphasized in our manuscript, should just be an experimental fact but not an unusual concern. We agree with the reviewer's comment and correct the relevant statement in the revised manuscript. The main contribution of this work is to reveal the critical role of O atom in the oxidized platinum-based catalysts, further complement the knowledge boundary of Pt-based electrocatalytic HER, and may provide a new way to design more efficient and lower content Pt-based catalysts for HER.

We made the modification in the manuscript as shown below.
The mass activity and specific activity were normalized by the mass loading and the ECSA of Pt. As depicted in Fig. 2c，at an overpotential of 77 mV, PtW 6 O 24 /C displays a mass activity of 20.175 A mg -1 , while the mass activity of 20% Pt/C is 0.398 A mg -1 . Furthermore, 1% PtW 6 O 24 /C displays a specific activity of 35.266 mA cm -2 at 50 mV, and the value of 20% Pt/C is 0.132 mA cm -2 under the same condition. (see manuscript page 9) Herein, it should be also clarified that although commercial Pt/C is widely used as a standard reference for HER research, its performance is already not the best. No matter in size, morphology, and dispersion of metal Pt, there exists enough space to improve its catalytic activity 9-14,39,40 . Thus, surpassing commercial Pt/C does not mean that metal Pt-based catalysts are out of date, which exactly suggests an important driving force for deeply developing such state-of-the-art catalysts.  We made the corresponding modification in supporting information as shown below.
We further explore the HER performance of electrocatalyst Pt 2 (W 5 O 18 ) 2 /C. As shown in Supplementary Figure. 42a (6) Å. The distance between two Pt atoms is 3.1315(8) Å, which is obviously longer than the metal-metal bond distance. The 1% Pt 2 (W 5 O 18 ) 2 /C also exhibits an excellent electrocatalytic hydrogen evolution performance with an overpotential of 26 mV at 10 mA cm -2 (Supplementary Figure 42b and Figure 43), which is similar to that of 1% PtW 6 O 24 /C. The Tafel slope is 29.8 mV dec -1 with the Volmer-Tafel mechanism (Supplementary Figure 42c and Figure 44). The exchange current density is 1.42 mA cm -2 and the mass activity is 10.976 mg -1 at 77 mV, respectively. The TOFs at 100 mV is 16.63 s -1 . All these results are similar to 1% PtW As noted in review of the original manuscript, the key finding is that the partial reduction of Pt cations bound via oxide ligands to reduced W ions within molecular heteropolytungstate clusters are more effective than Pt(0), and comparable to some of the most effective reported in the literature in the hydrogen evolution reaction. Importantly, the W-based clusters appear to serve as stable, redox catalysts, consistent with reversible reduction / reoxidation under turnover conditions. This latter point is sufficiently important that additional experimental support for this conclusion was requested for inclusion in the revised manuscript. The authors have now included this additional data, and it indeed makes a strong case for catalyst stability under turnover conditions. As such, publication of the revised manuscript is now recommended (with compliments to the authors on the quality of their newly added experiments and data). Ira A. Weinstock Reviewer #2 (Remarks to the Author): This reviewer acknowledges author's revision.
In particular, I'd like to congratulate them for the way that they have made available the DFT data through such ioChem-BD repository. This is really a nice form of publishing computational chemistry results.
Publish as it is.