Metallic W/WO2 solid-acid catalyst boosts hydrogen evolution reaction in alkaline electrolyte

The lack of available protons severely lowers the activity of alkaline hydrogen evolution reaction process than that in acids, which can be efficiently accelerated by tuning the coverage and chemical environment of protons on catalyst surface. However, the cycling of active sites by proton transfer is largely dependent on the utilization of noble metal catalysts because of the appealing electronic interaction between noble metal atoms and protons. Herein, an all-non-noble W/WO2 metallic heterostructure serving as an efficient solid-acid catalyst exhibits remarkable hydrogen evolution reaction performance with an ultra-low overpotential of −35 mV at −10 mA/cm2 and a small Tafel slope (−34 mV/dec), as well as long-term durability of hydrogen production (>50 h) at current densities of −10 and −50 mA/cm2 in alkaline electrolyte. Multiple in situ and ex situ spectroscopy characterizations combining with first-principle density functional theory calculations discover that a dynamic proton-concentrated surface can be constructed on W/WO2 solid-acid catalyst under ultra-low overpotentials, which enables W/WO2 catalyzing alkaline hydrogen production to follow a kinetically fast Volmer-Tafel pathway with two neighboring protons recombining into a hydrogen molecule. Our strategy of solid-acid catalyst and utilization of multiple spectroscopy characterizations may provide an interesting route for designing advanced all-non-noble catalytic system towards boosting hydrogen evolution reaction performance in alkaline electrolyte.


REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): Chen et al. presented an interesting study where W/WO2 catalyst was produced and demonstrated as highly effective for H2 evolution in alkaline media.The experimental side of the work is very detailed and presents a concerted use of different advanced characterization techniques intended to unveil the mechanism of high HER activity of W/WO2 catalysts.Moreover, the experimental findings were supplemented by DFT calculations.The presented picture is plausible but lacks an explanation of how dissolution was prevented.Also, there are numerous other points that should be resolved, as indicated below, before considering the present work for publication in Nature Communications.
-Language can be improved to make the manuscript easier to read -Authors mentioned "appealing electronic interaction between zero-valence W atoms and protons" citing the work which actually refers to WC as H2 evolution catalyst.It is accepted that strong metallic W-Hads interactions render poor HER activity of tungsten.Could Authors be more specific about their claims?-Looking at the Pourbaix plot of tungsten, at high pH, metallic W actively dissolves.The Authors also mentioned that their catalyst is less prone to dissolution, so a detailed justification of the twist of W dissolution behavior is needed.
-What is the role of carbon and the carbon sources in the W/WO2 synthesis?Please make sure to clearly present your strategy from the very beginning.
-Why PtRu/C was selected as a benchmark?-Please re-do the Tafel analysis at potentials that are far enough from 0 V vs. RHE (zero overvoltage).As a note, HER is the cathodic reaction, thus, current densities, overvoltages, and Tafel slopes should all be negative.-EIS is always in operando, it is not possible to do it ex-situ.
-Could Authors show EIS at lower frequencies (minimum was 0.1 Hz) and discriminate between capacitance or some additional resistance components?Also, the discussion is very descriptive, without proper quantification of experimental data.Which circuits were used to model EIS and extract discussed parameters?-Stability is not actually very good, but I was also wondering why electrolyte was replaced for the measurements at -50 mA cm-2?Also, are there any tungsten species in the solution, suggesting dissolution?-HER polarization curves go to relatively high currents, but stability was tested at lower currents.Why not providing stability tests at higher currents (at least 200 mA cm-2, which is the lower limit for industrial alkaline electrolysis)?-Going back to the poor HER activity of W, what is the barrier for the formation of H2 from 2Hads adsorbed on W, and how this relates to the Authors' findings?-Much more detail are needed for the description of DFT calculations.As presented now, the calculations cannot be reproduced.
Reviewer #2 (Remarks to the Author): Electrocatalytic hydrogen evolution is an important technology to convert renewable electricity into hydrogen fuel.Developing highly efficient, cheap HER catalysts are key to make this technology widely useful.In this work, the authors have synthesized a W/WO2 catalyst which shows decent performance in alkaline electrolytes, demonstrate an excellent water dissociation performance.The work is 1.There is a discrepancy between the terminology used within the manuscript and that used in the accompanying figures when referring to the prepared tungsten nanoparticles and WO2 nanorods.For instance, the term 'phase-pure WO2' is utilized in the manuscript, whereas 'bare WO2' appears in the figures (for example, Figure 2b or 2c).For consistency, it is recommended to revise the terminology across both the manuscript and the figures.
2. The W/WO2 catalyst was calcined using a 5% H2/Ar and at 750 oC, and carbon residue was visible in the TEM images (Figure 1 and S4).Could there be a possibility of tungsten carbide (WC) synthesis?WC is typically studied as another type of HER catalyst.To conclude that improved catalytic performance is originated from the W/WO2, evidence is needed to exclude the presence of WC.C 1s XPS spectra would help confirm that the influence of any potential WC catalyst on the HER can be disregarded.
3. The authors attributed the enhanced catalytic performance of the W/WO2 heterostructure to the abundance of oxygen vacancies, as revealed by ESR and NEXAFS analyses at the O K edge.However, the XPS O 1s spectra, which also contain the information about the oxygen vacancies, were not presented in this manuscript.XPS results could provide additional support by confirming the quantity of oxygen vacancies formed within the prepared catalyst.4. In Figure S11, a new NEXAFS peak emerges for the used W/WO2 at 540 eV.Commonly, an O K edge NEXAFS spectra peak around 540 eV is indicative of interaction between the O 2s orbital and the metal sp state.Please provide further explanation about the significance and implications of this new peak.
5. The illustration of Figure 4e and f was omitted in the manuscript and caption.Please describe the meaning of that figure.
6. What about assessing the catalytic activity of the W, WO2, and W/WO2 catalysts using the Rotating Disk Electrode (RDE) technique?Evaluating the HER activity of the prepared catalysts through RDE measurements could provide a more quantitative comparison of their catalytic activity relative to previously reported catalysts.
7. The authors computed the transition energy for the water dissociation pathway on W, WO2, and at the W/WO2 interface, with the corresponding simulated models displayed in Figures 5 and S18.Could the authors specify the bond length between the oxygen and dissociated hydrogen atom after dissociation?It appears that this bond length on the W/WO2 interface may be shorter than in the other two cases, which might contribute to the smaller energy barrier.Additionally, more comprehensive information, such as top and side images of the calculated structure, would be beneficial to better illustrate the W/WO2 interface.
8. The author insisted that heterojunction between the W nanoparticle and WO2 structure is the reason of the improved catalytic activity.How about the HER activity of the catalyst that synthesized by physical mixing of the W nanoparticle and WO2 nanorod?

Answer to the comments of reviewers:
Replies on comments of reviewer 1: Chen et al. presented an interesting study where W/WO2 catalyst was produced and demonstrated as highly effective for H2 evolution in alkaline media.The experimental side of the work is very detailed and presents a concerted use of different advanced characterization techniques intended to unveil the mechanism of high HER activity of W/WO2 catalysts.Moreover, the experimental findings were supplemented by DFT calculations.The presented picture is plausible but lacks an explanation of how dissolution was prevented.Also, there are numerous other points that should be resolved, as indicated below, before considering the present work for publication in Nature Communications.
Author reply: Thanks for the referee's valuable comments.We have supplemented additional experiments and demonstrations to address the referee's concerns, in particular, detailed characterizations have been performed to better understand the twist of tungsten dissolution for W/WO2 materials in high-pH solutions.

Comments 1:
Language can be improved to make the manuscript easier to read.
Author reply: Thanks for the referee's valuable comment.The language of this manuscript has been polished thoroughly.

Comments 2:
Authors mentioned"appealing electronic interaction between zero-valence W atoms and protons" citing the work which actually refers to WC as H2 evolution catalyst.It is accepted that strong metallic W-Hads interactions render poor HER activity of tungsten.Could authors be more specific about their claims?
Author reply: Thanks for the referee's careful reading and professional comments.Generally, our cited work demonstrates that the introduction of metallic W into WC materials can efficiently regulate the hydrogen adsorption energies on catalyst surface when the produced hydrogen intermediates locate at the bridging and/or hollow sites between zero-valence W atoms in alkaline HER process (Adv.Sci. 9, 2106029 (2022)).In our work, we mainly would like to illustrate that the high-valence W 4+ oxidation species (WO2) benefits the construction of Brønsted acid sites for achieving a proton-concentrated catalyst surface, while the interfacial zero-valence W (W 0 ) sites prefer to accelerate the deprotonation kinetics of neighboring Brønsted acid sites.The significant advantage of interfacial W 0 sites for hydrogen desorption is predicted by the subsequent DFT calculations (Fig. 5c).In addition, for the viewpoint that strong metallic W-Hads interactions render poor HER activity of tungsten, we would like to point out that the metallic W-Hads interactions are largely depended on the chemical and electronic environments of metallic W atoms (Fig. 5c), bare W metal exhibits a relatively strong W-Hads interaction with the ΔGH up to -0.51 eV, while the value is substantially decreased (-0.41 eV) when H intermediates are located at the interfacial W 0 atoms (Fig. 5c), and can be further optimized with the increase of proton coverage (Brønsted acid sites) at the W/WO2 interface (Fig. 5e).Accordingly, the more specific text has been added in the revised manuscript."(iii) considering the relatively sluggish hydrogen desorption kinetics of HxWOy intermediates, the introduction of zero-valence W (W 0 ) sites can further accelerate the deprotonation kinetics of Brønsted acids for the cycling of active sites due to the optimized electronic interactions between W 0 atoms and protons at the W/WO2 interface."Looking at the Pourbaix plot of tungsten, at high pH, metallic W actively dissolves.The authors also mentioned that their catalyst is less prone to dissolution, so a detailed justification of the twist of W dissolution behavior is needed.
Author reply: Thanks for the referee's constructive suggestion.Prior to discuss the dissolution behaviors of the as-prepared three types of tungsten-based materials, we would like to point out that all tungsten-based materials are supported and/or encapsulated by graphite carbon layers (synthetic strategy and morphology characterizations), thus surface carbon layers cannot be neglected for the protection of inner tungsten species (Adv.Mater.34, 2202743 (2022); Nano   Energy 69, 104455 (2020)).In order to better illustrate the advantage of the as-prepared W/WO2 in anti-dissolution property under high-pH solutions, we have detected the concentrations of the dissolved W species for W, WO2, and W/WO2 powders in 1 M KOH solutions.In detail, W (1 mg), WO2 (1 mg), and W/WO2 (1 mg) were added to 5 mL KOH (1 mol/L) solutions, respectively.
After 20 mins, the concentration of dissolved W species was determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES, Agilent ICP-OES730), and the values of dissolved W species are 7.1 × 10 -4 , 1.1 × 10 -4 , and 6.3 × 10 -5 mol/L for W, WO2, and W/WO2, respectively.Compared to WO2 and W/WO2 materials, W powder exhibits a much higher concentration of dissolved W species, probably because W nanoparticles with sizes less than 5 nm have poor oxidation resistance in alkaline solutions.After 2 days, the detected concentration of dissolved W is 3.4 × 10 -4 and 9.6 × 10 -5 mol/L for WO2 and W/WO2 samples, respectively, while the value is as high as 4.7 × 10 -3 for W nanoparticles, respectively, suggesting the long-term alkaline leaching has severely caused the dissolution of W powder.Based on above experimental evidences, two significant points can be concluded: (i) WO2 and W/WO2 heterostructures have relatively strong anti-dissolution property in high-pH solutions, which may be attributed to their intrinsically good oxidation resistance with co-existence of metal and oxide features (J.Mater. Chem.C, 6, 3200-3205 (2018); Adv.Energy Mater. 12, 2103301 (2022)) and the protection of surface carbon layers, meanwhile, the strong chemical and electronic interactions of W and WO2 components within W/WO2 may further improve the structural robustness in alkaline solutions (ACS Catal. 10, 13227-13235 (2020); Mater.Today Nano 6, 100038 (2019)); (ii) the ultra-small W NPs have poor alkaline leaching resistance, but we cannot observe the complete dissolution of W powders due to the protection of carbon layers.In addition, we would like to point out that partial tungsten species of initial ICP-OES detection may be originated from the naturally dissolution of inevitable high-valence tungsten oxides on low-valence WO2 and W/WO2 samples (Adv.Energy  Author reply: Thanks for the referee's valuable comment.The carbon sources are organic polyethylene oxide-co-polypropylene oxide-co-polyethylene oxide (P123) and dopamine (DA).In the pyrolysis process, the pyrolyzed carbon can cause the reduction reaction (C+W18O49→W+WO2+CO2↑) of high-valence W18O49 parent materials for the synthesis of desired W and/or WO2 products, meanwhile, the produced graphite carbon layers not only improve the electron transfer in following alkaline HER process, but also alleviate the alkaline-leaching rate of inner W/WO2 materials.Accordingly, the following text has been added in the revised manuscript.
"where the whole reduction reactions can be simplified as following equations: C+W18O49→W + WO2 + CO2↑, H2 + W18O49→ W + WO2 + H2O↑.In the pyrolysis process, carbon can cause the reduction process of high-valence W18O49 parent materials for the synthesis of desired W and WO2 products.In the following alkaline HER process, the produced graphite carbon layers not only improve the electrocatalytic charge transfer, but also alleviate the alkaline-leaching rate of inner W/WO2 materials."2021)).In our work, from the viewpoint of constructing solid-acid catalyst surface using tungsten-based materials, we mainly declare that the energy barriers of water dissociation and hydrogen desorption steps can be substantially decreased on W/WO2 heterostructure catalyst in alkaline HER process.Accordingly, commercial PtRu/C is selected as the benchmark HER catalyst in our work.

Comments 6:
Please re-do the Tafel analysis at potentials that are far enough from 0 V vs. RHE (zero overvoltage).As a note, HER is the cathodic reaction, thus, current densities, overvoltages, and Tafel slopes should all be negative.
Author reply: Thanks for the referee's valuable comment.As suggested by the referee, we have re-done the Tafel analysis with regions near 0 V (Fig. 3d).As can be seen, the current responses are extremely weak for all catalysts when overpotentials are positive (>0 V), and such weak signals may be originated from the capacitance of naturally adsorbed electrolyte on Ni-foam electrode (0.5 cm × 0.5 cm, Ni foam).However, slow current responses are observed on all large-area free-standing catalysts (slow begin of HER process on large-area electrode) when overpotentials just entering to negative regions (<0 V).Afterwards, the alkaline HER process starts to enter the Tafel region with the fastest reaction rate.Similar Tafel regions are also observed in other free-standing systems (Nat.Commun.12, 6776 (2021); Nat.Commun.9, 924 (2018); Adv.Mater. 31, 1904989 (2019)).In addition, we have also tried to eliminate the influence of slow begin of HER process for the selection of Tafel regions by using rotating disk electrode (RDE) technique with glassy carbon working electrode (diameter: 3 mm) (measurement details seen in the method section), where current responses are sensitive and smooth when overpotentials entering to negative potential regions.W/WO2 catalyst still exhibits a remarkable alkaline HER activity with a low overpotential (-60 mV) at -10 mA/cm 2 and a small Tafel slope (-54 mV/dec) (Supplementary Fig. 9), which are more excellent than those of W and WO2 counterparts, and still excelling most previously reported metal oxides (Fig. 3c).Accordingly, the following text has been added in the revised manuscript."Moreover, the HER performance of W, WO2, W/WO2 powders were also examined using the rotating disk electrode (RDE) technique in 1 M KOH electrolyte.As can be seen, W/WO2 catalyst still exhibits a remarkable alkaline HER activity with a low overpotential (-60 mV) at -10 mA/cm 2 and a small Tafel slope (-54 mV/dec) (Supplementary Fig. 9), which are more excellent than those of W and WO2 counterparts, and still excelling most previously reported metal oxides (Fig. 3c)."EIS is always in operando, it is not possible to do it ex-situ.
Author reply: Thanks for the referee's valuable comment.All inappropriate descriptions have been corrected in the revised manuscript.

Comments 8:
Could authors show EIS at lower frequencies (minimum was 0.1 Hz) and discriminate between capacitance or some additional resistance components?Also, the discussion is very descriptive, without proper quantification of experimental data.Which circuits were used to model EIS and extract discussed parameters?
Author reply: Thanks for the referee's valuable comments.Generally, all presented EIS curves were collected in the frequency range of 10 -1 ~10 5 Hz (Fig. 3f), which can be confirmed by the corresponding Bode phase plots in Fig. 3g 2020) ), and the second parallel ones of R2 and Cφ are attributed to the hydrogen adsorption resistance and pseudo-capacitance at high frequencies, respectively.As expected, all electrochemically treated W/WO2 samples exhibit the similar R1 value (~4.0 Ω), and the small values of R2 for all W/WO2 catalysts suggest the fast charge transfer kinetics between catalyst surface and H2O molecules.In particular, R2 decreases to 4.4 Ω sharply with a negligible water diffusion resistance at applied overpotential of -20 mV, indicating the adsorption and activation of water molecules on the W/WO2 catalyst surface can be achieved under low overpotentials.Further, we also notice that R3 and Cφ are largely overpotential-dependent, where W/WO2 catalysts exhibit significantly decreased R3 with increased Cφ when increasing the applied overpotentials, in particular, the value of R3 can be as low as approximately 1.8 Ω, while Cφ is up to 0.017 F at an overpotential of -30 mV, suggesting the hydrogen adsorption resistance (R3) is very small, and the capacitance of proton coverage is very large on W/WO2 catalyst surface under low overpotentials.
Accordingly, the relevant discussions have been added in the revised manuscript."All Nyquist plots of W/WO2 samples were simulated by a double-parallel equivalent circuit model in accordance with previous reports (Energy Environ. Sci., 12, 2298-2304(2019);J. Power Sources, 158, 464-476 (2006)), where R1 represents the uncompensated solution resistance, the first parallel components (constant phase element (CPE) and R2) indicate the charge transfer resistance caused by the adsorption and activation of water molecules at low frequencies (Energy Environ. Sci., 14, 6428-6440 (2021); ACS Appl.Energy Mater. 3, 66-98 ( 2020) ), and the second parallel ones of R2 and Cφ are attributed to the hydrogen adsorption resistance and pseudo-capacitance at high frequencies, respectively (Fig. 3f and Supplementary Table 1).As expected, all electrochemically treated W/WO2 samples exhibit the similar R1 value (~4.0 Ω), and the small values of R2 for all W/WO2 catalysts suggest the fast charge transfer kinetics between catalyst surface and H2O molecules.In particular, R2 decreases to 4.4 Ω sharply with a negligible water diffusion resistance at applied overpotential of -20 mV, indicating the adsorption and activation of water molecules on the W/WO2 catalyst surface can be achieved under low overpotentials.Further, we also notice that R3 and Cφ are largely overpotential-dependent, where W/WO2 catalysts exhibit significantly decreased R3 with increased Cφ when increasing the applied overpotentials, in particular, the value of R3 can be as low as approximately 1.8 Ω, while Cφ is up to 0.017 F at an overpotential of -30 mV, suggesting the hydrogen adsorption resistance (R3) is very small, and the capacitance of proton coverage is very large on W/WO2 catalyst surface under low overpotentials."Stability is not actually very good, but I was also wondering why electrolyte was replaced for the measurements at -50 mA cm -2 ?Also, are there any tungsten species in the solution, suggesting dissolution?
Author reply: Thanks for the referee's valuable comments.The larger current density means the higher consumption of water molecules, and the replacement of electrolyte can guarantee the supply of HER reactant, which meanwhile prevents the W/WO2 catalyst from the possible alkaline leaching by the continuously concentrated OH -intermediates near cathodic electrode after long-term hydrogen production.The operations of refreshing electrolyte have been widely used in electrolysis (Angew.Chem.Int. Ed. 62, e202300390 (2023);Energy Environ. Sci., 13, 119-126 (2020);Adv. Funct. Mater. 31, 2010367 (2021)), and the activity can be recovered when refreshing the electrolyte in our work, suggesting the relatively stable HER process using W/WO2 solid-acid catalyst.On the other hand, the larger current density also means the higher hydrogen production, and the release of H2 gas bubbles also can cause the fluctuation of the collected stability curves (Nat.Commun.12, 3540 (2021); Nat.Commun.9, 924 (2018); J. Mater.Chem.A, 10, 6242-6250 ( 2022)).However, the overpotentials at -10 and -50 mA/cm 2 of the Chronopotentiometry measurements are still maintained near the values of -35 mV and -115 mV in our work, which also suggests the good stability performance of W/WO2 solid-acid catalyst in alkaline HER process.
In addition, similar to previous report (Nat.Commun.9, 2609 (2018)), inductively coupled plasma-optical emission spectroscopy (ICP-OES) was performed to detect the possibly dissolved tungsten species of W/WO2 catalyst in the used KOH electrolyte (50 mL, 1 mol/L KOH).As can be seen, the concentration of W is determined to be approximately 7.5 × 10 -6 mol/L when the catalyst electrode is immersed in KOH electrolyte under open circuit potential condition, partial of which can be attributed to the dissolution of inevitable high-valence tungsten oxides (e.g., WO3) on W/WO2 catalyst surface.After long-term hydrogen production, the detected dissolved W species in the used and refreshed electrolytes (indicated by the red dashed circle, Fig. 3i) are approximately 9.8 × 10 -6 and 6.3 × 10 -7 mol/L (refreshed electrolyte), respectively, suggesting the extremely slow dissolution of W/WO2 catalyst in alkaline HER process.Meanwhile, unlike the naturally alkaline leaching under circuit potential condition, we would like to point out that the negative potentials at cathode can provide rich electrons to avoid the oxidation and dissolution of low-valence tungsten species during alkaline HER process, which causes rather low concentration (6.3 × 10 -7 mol/L) of dissolved tungsten species detected in the refreshed electrolyte, suggesting structural robustness of W/WO2 catalyst after long-term hydrogen production in alkaline electrolyte.
Therefore, combining the detailed characterizations in Comments 3 of Reviewer 1, the significantly improved anti-dissolution property of our well-designed W/WO2 composites in high-pH electrolyte can be understood by the following two reasons: (i) W/WO2 heterostructures have intrinsically good oxidation resistance with co-existence of metal and oxide features (J. Mater.Chem. C, 6, 3200-3205 (2018); Adv.Energy Mater. 12, 2103301 ( 2022)) and the protection of surface carbon layers (Adv. Mater. 34, 2202743 (2022); Nano Energy 69, 104455 (2020)), meanwhile, the strong chemical and electronic interactions of W and WO2 components within W/WO2 may further improve the structural robustness in alkaline solutions (ACS Catal. 10, 13227-13235 (2020); Mater.Today Nano 6, 100038 (2019)); (ii) unlike the naturally alkaline leaching under circuit potential condition, the negative potentials at cathode can provide rich electrons to avoid the oxidation and dissolution of low-valence tungsten species during alkaline HER process.
Accordingly, the relevant discussions have been added in the revised manuscript.
"The relatively good stability of W/WO2 solid-acid catalyst in alkaline electrolyte is also confirmed by the inductively coupled plasma-optical emission spectroscopy (ICP-OES).After long-term hydrogen production, the initial electrolyte, used electrolyte, and refreshed electrolyte (indicated by red dashed circle, Fig. 3i) shows no significant increase in the concentrations of dissolved W species with values of 7.5 × 10 -6 , 9.8 × 10 -6 , and 6.3 × 10 -7 mol/L, respectively.In particular, the rather low concentration (6.3 × 10 -7 mol/L) of dissolved W species in the refreshed electrolyte directly suggests the structural robustness of W/WO2 catalyst after long-term hydrogen production in alkaline electrolyte.""Therefore, based on the comprehensive evaluations of alkaline HER activity and stability on W/WO2 materials, the significantly improved structural robustness of our well-designed W/WO2 composites in high-pH solutions can be attributed to the following two reasons: (i) W/WO2 heterostructures have intrinsically good oxidation resistance with co-existence of metal and oxide features (J.Mater.Chem. C, 6, 3200-3205 (2018); Adv.Energy Mater. 12, 2103301 (2022)) and the protection of surface carbon layers (Adv. Mater. 34, 2202743 (2022); Nano Energy 69, 104455 (2020)), meanwhile, the strong chemical and electronic interactions of W and WO2 components within W/WO2 may further improve the structural robustness in alkaline solutions (ACS Catal. 10, 13227-13235 (2020); Mater.Today Nano 6, 100038 (2019)); (ii) unlike the naturally alkaline leaching under circuit potential condition, the negative potentials at cathode can provide rich electrons to avoid the oxidation and dissolution of low-valence tungsten species during alkaline HER process." Comments 10: HER polarization curves go to relatively high currents, but stability was tested at lower currents.
Why not providing stability tests at higher currents (at least 200 mA cm -2 , which is the lower limit for industrial alkaline electrolysis)?
Author reply: Thanks for the referee's valuable comment.According to the constructive suggestion of the referee, the electrocatalytic stability of W/WO2 catalyst was evaluated in alkaline electrolyte at a current density of -200 mA/cm 2 .Compared to the relatively stable profile collected at -50 mA/cm 2 , although W/WO2 catalyst exhibits much stronger fluctuation in activity at current density of -200 mA/cm 2 , the activity can be recovered after refreshing the electrolyte, suggesting the good catalytic stability at industrial current density.
Accordingly, the following text has been added in the revised manuscript."In addition, we also evaluate the stability of W/WO2 catalyzing alkaline HER process at a current density of -200 mA/cm 2 (the lower limit for industrial water electrolysis).No significant activity loss can be observed on W/WO2 catalysts after long-term hydrogen production, suggesting the good catalytic stability at industrial current density."2020)).Meanwhile, the DFT calculations also give the consistent information, which suggests that pure W is generally inefficient in the cleavage of H-OH bonds with energy barrier up to 0.84 eV (Fig. 5b), whereas the energy barrier of hydrogen desorption can be substantially decreased to -0.51 eV (-0.51 eV).Conversely, metallic WO2 exhibits an extremely low energy barrier (0.06 eV) in water dissociation, but is poor at resulting in the desorption (-1.24 eV) of adsorbed H* intermediates.Therefore, an optimal catalyst can be designed by combining the catalytic proficiencies of W and WO2 materials, where the introduction of W component can accelerates the desorption kinetics of protons on neighboring WO2 catalyst surface.Author reply: Thanks for the referee's valuable comment.More detailed information of DFT calculations has been provided for the reproduction of simulation results in the revised manuscript.
Accordingly, the following text has been added in the method section for DFT calculations."The projector augmented wave potentials with the Perdew-Burke-Ernzerhof form of the exchange-correlation functional were employed in all the simulations with the energy cut-off of 450 eV, and the long-range van der Waals (vdW) interaction was described with the DFT-D3 method.The k-point meshes were generated using the VASPKIT tool with the grid separation of 0.04 Å −1 for the geometry optimizations and self-consistent field energy calculations, in which the total energy convergence and interaction force were set to be 10 -6 eV and 10 -2 eV/Å, respectively.
Based on the experimental TEM analysis, the structures of W and WO2 substrates were built by cleaving the (110) plane of bulk W and (01-1) bulk WO2, respectively.A vacuum region of 15 Å was set along the z direction to avoid the interaction between periodic images.Since the lattice parameters of WO2 (01-1) plane were 7.38 Å and 5.58 Å, a 2 x 2 supercell was built as the WO2 substrate for constructing the W/WO2 interface.The (110) plane of W with a lattice parameter of 2.70 Å was enlarged to build a 3 x 4 supercell, which was then cleaved (100) surface to build a W slab whose ( 110) and ( 110) surfaces were exposed.The W/WO2 interface was constructed by placing the W slab on the WO2 substrate.The lattice mismatch between W slab (10.8 Å) and WO2 substrate (11.2Å) was only about 3%.The structures of constructed W/WO2 interface with front, side, and top images can be observed in Supplementary Fig. 21."

Replies on comments of reviewer 2:
Electrocatalytic hydrogen evolution is an important technology to convert renewable electricity into hydrogen fuel.Developing highly efficient, cheap HER catalysts are key to make this technology widely useful.In this work, the authors have synthesized a W/WO2 catalyst which shows decent performance in alkaline electrolytes, demonstrate an excellent water dissociation performance.The work is interesting, cleanly done, and well organized.I would like to recommend its publication in Nature Communications after the following minor revisions.
Author reply: We appreciate the reviewer's comments and recommendation very much.We have supplemented additional experiments and demonstrations to address the reviewer's concerns.
Comments 1: In the step of thermal decomposition precursor, is it possible for C to infiltrate into the material lattice and form low-crystallinity WC or W2C? Please demonstrate that there is no infiltration of C during the synthesis of W/WO2, and prove the existence of graphite carbon in the material through the D/G peak in Raman spectra.

Figure .
Figure.The investigation of anti-dissolution property of W/WO2 in comparison with W and WO2 counterparts.Comments 4:What is the role of carbon and the carbon sources in the W/WO2 synthesis?Please make sure to clearly present your strategy from the very beginning.

Fig 3 .
Fig 3. (a) The raw data including the capacitance of naturally adsorbed electrolyte, slow begin of HER process, and Tafel regions for all catalysts.(b) Tafel plots of all catalysts extracted from (a).
(c) Polarization curves and (d) the corresponding Tafel plots of W, WO2, W/WO2 catalysts examined by RDE technique.Comments 7:

Fig. 3
Fig. 3 (f) Nyquist plots of W/WO2 catalysts with the increase of applied overpotentials, the inset shows the equivalent circuit for the simulation.Note that inhomogeneities in the surface of metal oxide electrodes usually result in non-ideal capacitance in the double-layer at the solid/electrolyte interface.Thus, CPEs (CPE-T and CPE-P) are routinely used in place of pure capacitors to model this interfacial layer.Comments 9:

Fig. 5
Fig. 5 The calculated free energy diagrams of (b) water adsorption and dissociation, and (c) hydrogen desorption steps on W, WO2, and W/WO2 catalyst surface.Comments 12: Much more detail are needed for the description of DFT calculations.As presented now, the calculations cannot be reproduced.
. All Nyquist plots of W/WO2 samples were simulated by a double-parallel equivalent circuit model in accordance with previous reports (Energy Environ.