One-step synthesis of single-site vanadium substitution in 1T-WS2 monolayers for enhanced hydrogen evolution catalysis

Metallic tungsten disulfide (WS2) monolayers have been demonstrated as promising electrocatalysts for hydrogen evolution reaction (HER) induced by the high intrinsic conductivity, however, the key challenges to maximize the catalytic activity are achieving the metallic WS2 with high concentration and increasing the density of the active sites. In this work, single-atom-V catalysts (V SACs) substitutions in 1T-WS2 monolayers (91% phase purity) are fabricated to significantly enhance the HER performance via a one-step chemical vapor deposition strategy. Atomic-resolution scanning transmission electron microscopy (STEM) imaging together with Raman spectroscopy confirm the atomic dispersion of V species on the 1T-WS2 monolayers instead of energetically favorable 2H-WS2 monolayers. The growth mechanism of V SACs@1T-WS2 monolayers is experimentally and theoretically demonstrated. Density functional theory (DFT) calculations demonstrate that the activated V-atom sites play vital important role in enhancing the HER activity. In this work, it opens a novel path to directly synthesize atomically dispersed single-metal catalysts on metastable materials as efficient and robust electrocatalysts.


Comment 4. Figure 3b, title for X axis: Please change "wavelenth" to "wavelength".
Response 4. We appreciate the reviewer very much for the helpful suggestion. We have changed the "wavelenth" to "wavelength" in Figure 3b, which has been highlighted on Page 7 in the revised manuscript.
The above-mentioned Figure 3 has been revised as follows: Fig. 3 Phase transition from 1T to 2H. a, Raman spectra of V 2 O 3 film (black), V SACs@1T-WS 2 /V 2 O 3 (orange) and 2H-WS 2 (blue); b, PL spectra of V 2 O 3 film (black), V SACs@1T-WS 2 /V 2 O 3 (orange) and 2H-WS 2 (blue); c, High-resolution XPS spectra of W 4f (left) and S 2p (right) core level peak regions for 2H and 1T phase, respectively. The fitting blue and pink curves represent the contributions of 1T and 2H phases, respectively; d, Schematic representation of the 1T to 2H structural phase transition occurring upon thermal annealing; e, Raman spectra of 2H-WS 2 , V SACs@1T-WS 2 , and V SACs@1T-WS 2 with different annealing temperatures in H 2 /Ar for 2h.

Comment 5. Supplementary Figure 1, caption: please delete the description "The blue and yellow balls represent …" because these balls and their meanings have been illustrated in the schematic models.
Response 5. We appreciate the reviewer very much for the helpful suggestion. We agree with the reviewer that the description of "the blue and yellow balls represent…" 30   Response 9. We appreciate the reviewer very much for these helpful suggestions.
Following your suggestions, same colors were used to represent the same samples in Supplementary Fig. 15 and the caption of "partially 1T phase transformation" has been changed to "partial 1T phase transformation".
Please note that the above-mentioned Supplementary Fig. 14  Supplementary Fig. 15 a, PL spectra of 2H-WS 2 , V SACs@1T-WS 2 , and V SACs@1T-WS 2 with different annealing temperatures in H 2 /Ar for 2h; b, Enlarged PL spectra of V SACs@1T-WS 2 samples with different annealing temperatures. For all the V SACs@1T-WS 2 samples, the intensity of PL are almost completely suppressed, indicating the typical metallic behavior 9 . Meanwhile, compared with V SACs@1T-WS 2 -200 º C sample and V SACs@1T-WS 2 -300 º C sample, the PL intensity showed much higher intensity and blue shift in the V SACs@1T-WS 2 -400 º C sample, which may be attributed to band-structure modification due to the partial 1T phase transformation to 2H phase. in the revised manuscript, which has been highlighted on Page 9-11. Therefore, we have made corresponding changes in the revised Supplementary Fig. 16.
Please note that the above-mentioned Supplementary Fig. 15 in the original supplementary information has been revised as Supplementary Fig. 16 in the revised supplementary information. The revised Supplementary Fig. 16 was displayed as follows: Supplementary Fig. 16 a, Raman spectra of pure sapphire (black), V 2 O 3 film (olive) and V 2 O 3 -Nuclei film (blue); b, XRD spectra of V 2 O 3 -Nuclei film (blue), V 2 O 3 -No Nuclei film (orange) and V 2 O 3 film (olive); c, Survey spectra of V 2 O 3 -Nuclei film on sapphire substrate; d-f, High-resolution XPS spectra showing the S 2p (d), Al 2p (e) and V 2p (f) core levels for the V 2 O 3 -Nuclei film on the sapphire substrate sample. Please note that V 2 O 3 -No Nuclei film was prepared using VCl 3 precursor only. The oxygen was deduced from the residue oxygen in the tube furnace. The Raman spectrum of V 2 O 3 film (olive) in Fig. S16a was measured on the 1T/V 2 O 3 sample where no WS 2 appeared. The XRD pattern in Fig. S16b (olive) was obtained on the 1T/V 2 O 3 sample. Response 11. We appreciate the reviewer very much for these helpful suggestions.
Following your suggestions, we have added higher resolution images for the EDX Response 13. We appreciate the reviewer very much for the helpful suggestion.
Following your suggestions, we have added the scale bars in Supplementary Fig. 19a-b.
Besides, we have labeled the 1T-and 2H-WS 2 domains produced on the V 2 O 3 film.
Please note that the above-mentioned Supplementary Fig. 18 in the original supplementary information has been revised as Supplementary Fig. 19 in the revised supplementary information. The revised Supplementary Fig. 19 was displayed as follows: Supplementary Fig. 19

V 2 O 3 -No Nuclei
1T-VS 2 /2H-WS 2 and 1T-VS 2 /1T-WS 2 " to "1T-VS 2 /1T-WS 2 and 1T-VS 2 /2H-WS 2 " in Supplementary Fig. 22. Please note that the above-mentioned Supplementary Fig. 19 in the original supplementary information has been revised as Supplementary Fig. 22 in the revised supplementary information. Figure 20,  Following your suggestion, we have changed "Synthesized strategy" to "synthesis strategies" and "Ref." to "References" in the Table S1, which has been highlighted in the revised supplementary information.  film was the constraining factor to determine the formation of 1T-WS 2 with high phase purity when the amount of VCl 3 was higher than 0.03 g. Whereas, the density of VS 2 nuclei was the constraining factor to impact the 1T-WS 2 formation when the amount of VCl 3 was less than 0.03 g.

Comment 15. Supplementary
Accordingly, the influence of amount of VCl 3 on the 1T-WS 2 growth has been added in the revised manuscript to address this question, which has been highlighted on Page 11 as follows: "Additionally, it was also demonstrated that the amount of VCl 3 could significantly affect the controllable phase growth of WS 2 (see details in Figure   The Raman spectra of as-prepared products using different amount of VCl 3 . Supplementary Fig. 25 The Raman spectra of as-prepared V 2 O 3 film using different amount of VCl 3 . The Raman spectra were measured on the areas where no WS 2 domains appeared.   Supplementary Fig. 26 High resolution XPS spectra of W 4f (a) and S 2s (b) core level peak regions for WS 2 domains using different amount of VCl 3 . The fitting blue and pink curves represent the contributions of 1T and 2H phases, respectively. Please note that obtained WS 2 domains were transferred on the sapphire substrates for the XPS measurements.
The discussion of the effect of amount of VCl 3 on the 1T-WS 2 growth has been added in the revised supplementary information as follows: "In order to demonstrate the amount of VCl 3 has a great impact on the phase purity of WS 2 , we have prepared WS 2 samples using different mass amount of VCl 3 (from 0.01 g to 0.10 g). …. Whereas, the density of VS 2 nuclei was the constraining factor to impact the 1T-WS 2 formation when the amount of VCl 3 was less than 0.03 g." in Figure R1. Secondly, the choice of the c-plane sapphire substrates was due to the similar corundum structure to the V 2 O 3 film, which could promote the epitaxial growth of V 2 O 3 film on the surface. As far as we know, both a-plane (110) and c-plane (001) sapphire substrates could promote the epitaxial growth of V 2 O 3 film on the surface due to small in-plane lattice misfits (Table R1). The XRD patterns of a-and c-plane sapphire substrates were shown in Figure R2. The WS 2 growth on the a-plane sapphire substrate was performed under the same condition with c-plane sapphire.

Comment 2. Is the substrate of c-plane sapphire play an import role in
Surprisingly, no WS 2 domains were formed on the a-plane sapphire substrate, as demonstrated in Figure R3. In particular, the surface of V 2 O 3 film on the a-plane sapphire substrate was extremely rough from the optical micrograph in Fig. R3a,

Response 3.
We appreciate the reviewer very much for these helpful suggestions.
Following your suggestions, Raman spectroscopy and STEM imaging were performed to demonstrate the stability of V SACs 1T-WS 2 electrocatalyst after HER test. As shown in Supplementary Fig. 36a, obvious metallic peaks (J 1 , J 2 , J 3 ) of V SACs 1T-WS 2 electrocatalyst after HER test were clearly observed (red plot). Moreover, the typical 1T structure of WS 2 after stability test was confirmed according to the STEM image in Supplementary Fig. 36b.
Accordingly, the description of metallic properties of V SACs 1T-WS 2 after HER test has been added in the revised manuscript to address this question, which has been highlighted on Page 13 as follows: "The metallic properties of V SACs 1T-WS 2 catalyst after stability test were also investigated by Raman spectroscopy (Figure S36a), which showed obvious metallic peaks (J 1 , J 2 , J 3 ) in the Raman spectrum (red plot). Moreover, the STEM image showed in Figure S36b confirmed the V SACs 1T-WS 2 structure after stability test. Both the Raman spectrum and STEM image indicated the robust 1T structure of V SACs 1T-WS 2 catalyst after HER test." The above-mentioned Supplementary Fig. 36 has been added in the revised supporting information as follows:

Comment 5. Although the author has cited the paper for the calculations of TOFs, please specify the details for the TOFs in this work.
Response 5. We appreciate the reviewer very much for the helpful suggestion.
Following your suggestion, the details for the TOF were specified in the revised manuscript to address this suggestion, which has been highlighted on Page 19 as follows:

"Calculation of Turnover Frequency
The TOF calculation details were specified as below, which was reported elsewhere.

TOF = ℎ
The total hydrogen turnovers were calculated from the current density in the LSV polarization curve according to the equation as below: The number of active sites in the V SACs@1T-WS 2 catalyst was obtained from the mass loading on the glass carbon electrode.

" Comment 6. The catalytic tests to demonstrate the stability of the catalysts are incomplete. Evidently, the authors have shown the stability of catalysts after continuous CV cycles. However, this is not enough. Please carry out the stability test of long-term electrolysis (at least 10 h).
Response 6. We appreciate the reviewer very much for the helpful suggestion.
Following your suggestion, the stability test was conducted at current densities higher Accordingly, the description of the electrochemical stability test has been added in the revised manuscript to address this comment, which has been highlighted on Page 13 as follows: "The stability test was conducted at current densities higher than 10 mA/cm 2

Furthermore, HER catalysis properties of the V SACs@1T-WS 2 was tested, exhibiting enhanced HER properties compared to undoped 1T-WS 2 . This is an
interesting piece of work containing both experimental and theoretical efforts.

Thus, I would recommend a rejection of the current manuscript, but if the authors address all the points, the paper could be eventually published.
Response: We are pleased that the reviewer commented our work as "interesting" piece of work containing both experimental and theoretical efforts. His/Her concerns are addressed as follows.

Comment 1. In Fig. 2f and Line 116-118 of the main manuscript, EELS was performed to identify V dopants. However, since there is a V 2 O 3 film underneath the V SACs@1T-WS 2 , signals of V can also from the V 2 O 3 . Can the authors comment on how to rule out this possibility?
Response 1. We appreciate the reviewer very much for the insightful question. EELS was a very powerful and sensitive analytical technique to determine the local coordination state of V single atoms. To identify the V dopants in the 1T-WS 2 monolayers, we have transferred the V SACs@1T-WS 2 monolayers on the TEM grid for the STEM image and EELS analysis. We have explicitly described the transferring process in both the original manuscript and revised manuscript on Page 16 as follows: "The as-grown samples were transferred onto arbitrary substrates, such as fresh sapphire, SiO 2 /Si, glass carbon, and holy-carbon nickel TEM Grid using a modified method in our lab." In particular, the EELS of V SACs@1T-WS 2 in Fig. 2e showed two major features of V peaks assigned to V 4+ , indicating the V substitutions in the 1T-WS 2 layer.
However, the valence state of V in the V 2 O 3 was +3, which was different with the V valence state in the 1T-WS 2 layer. In conclusion, the signals of V EELS from V 2 O 3 could be ruled out.

Comment 2. Was any image filtering process used for enhancing the visibility of atoms in Fig. 2c and d? This should be included in experimental
details.

Response 2.
We appreciate the reviewer very much for the helpful suggestion.
In our STEM image process, radial wiener filter was processed to enhance the visibility of atoms in Fig. 2c-2d. Following your suggestion, we have added the filtering process in the revised manuscript to address this suggestion, which has been highlighted on Page 17 as follows: "Radial wiener filter was carried out to enhance the visibility of atoms." Response 3. We appreciate the reviewer very much for these helpful suggestions.

Comment 4. For the phase purity analysis (Fig. 3c), if I understand correctly, the XPS comparison was done between pure 2H-WS 2 and V SACs@1T-WS 2 /V 2 O 3 . I think the XPS peak shift can also originate from the interaction with V 2 O 3 , and also the effect of V doping, which can not be ruled out in this case, and makes the phase purity analysis not very convincing.
Response 4. We appreciate the reviewer very much for the question. XPS was a very powerful technique to detect the oxidation state and coordination geometry of W element. To quantitatively identify the phase purity of 1T monolayers, we have transferred the V SACs@1T-WS 2 monolayers on the sapphire substrate for the XPS analyses.
To make more clearer for the XPS analysis, we have added one sentence on Page 8 to address this comment. The added sentence was described as follows: "To illustrate the high purity of the obtained 1T phase, XPS spectra were performed to quantify the 1T and 2H compositions according to the high sensitivity of tungsten signal to its oxidation state and coordination geometry. 1,47 The V@ SACs@1T-WS 2 monolayers were transferred on the fresh sapphire substrate for the XPS measurements." We have explicitly described the transferring process in both the original manuscript and revised manuscript on Page 16 as follows: "The as-grown samples were transferred onto arbitrary substrates, such as fresh sapphire, SiO 2 /Si, glass carbon, and holy-carbon nickel TEM Grid using a modified method in our lab." We have also explicitly described the details of XPS measurements in both the original manuscript and revised manuscript on Page 17 as follows: "For the XPS and Raman spectra of V SACs@1T-WS 2 annealed at different temperatures, the V SACs@1T-WS 2 monolayers were transferred on the fresh sapphire substrates and annealed in H 2 /Ar condition with different temperatures." As demonstrated by the EELS, no V 2 O 3 contamination was involved in the transferred 1T monolayers, as a result, the possibility of XPS shift caused by the V 2 O 3 film could be ruled out. Additionally, no V signals of high-resolution V 2p XPS were detected in the transferred 1T monolayers due to the detection limit of XPS ( Figure   5R). Since the V signals were not detected, its effect on the XPS peak shift was not significant. Through a series of careful analyses, our phase purity analysis of 1T-WS 2 monolayers by the XPS spectra was very convincing.

Figure R5
. High-resolution of V 2p in the transferred 1T-WS 2 monolayers on the sapphire substrate.

SACs@1T-WS 2 /V 2 O 3 . Thus, similar to the previous question, could V 2 O 3 contribute and enhance the measured catalytical performance?
Response 5. We appreciate the reviewer very much for the insightful question. In our HER measurement of V SACs@1T-WS 2 monolayers, however, it was performed on a glass carbon electrode due to the insulating sapphire substrate. In particular, the V SACs@1T-WS 2 electrocatalyst was transferred on the glass carbon electrode.
We have explicitly described the transferring process in both the original manuscript and revised manuscript on Page 16 as follows: "The as-grown samples were transferred onto arbitrary substrates, such as fresh sapphire, SiO 2 /Si, glass carbon, and holy-carbon nickel TEM Grid using a modified method in our lab." We have also explicitly described the details of HER measurement in both the original manuscript and revised manuscript on Page 11 as follows: "The as-produced V SACs@1T-WS 2 monolayers were transferred on the glass carbon (GC) electrode for the HER performance measurement using a three-electrode setup in 0.5 M H 2 SO 4 ." As demonstrated by the EELS analysis in Fig. 2e The resolution has been enhanced in the revised Supplementary Fig. 8, Supplementary Fig. 6     The peak at 406.3 cm -1 was assigned to VS 2 and different with the peak at 418.3 cm -1 , which was derived from the A 1g mode of the substrate sapphire 5 . The recorded Raman peaks of VS 2 in Figure S18 were very weak, indicating that the amount of VS 2 was in very low concentration relative to V 2 O 3 . Scale bars: a-b, 90 μm.

Comment 7. Could the authors provide the V 2p XPS? Did the authors consider the V-S bond when fitting the V 2p peak?
Response 1. We appreciate the reviewer very much for these questions. Following your suggestion, we have provided the V 2p XPS spectrum in the transferred 1T-WS 2 monolayers, which was shown in Figure 5R. Apparently, no V signals were detected due to the detection limit of XPS. And we could not gain any useful information.
However, the V signals of EELS was pronounced due to the relatively higher detection limit than XPS (Fig. 2e) and two major features of L peaks were assigned to V 4+ , affirming the V-S bond in the 1T-WS 2 layer.  Figure S18,  Accordingly, the description of 1T or 2H domains confirmed by the Raman mapping was added in the revised manuscript to address this comment, which has been highlighted on Page 11 as follows:

Comment 8. In
"The 1T-or 2H-WS 2 domains were also confirmed by the Raman mapping in Figure   S20-S21." Additionally, the Raman mapping analyses were added in the revised supplementary information as follows: "Raman mapping was directly conducted on the WS 2 @V 2 O 3 film to identify the 1T or 2H domains formed on the different V 2 O 3 films, as shown in Supplementary Fig.   20…However, the signals of A 1g mode of V 2 O 3 film was observed, demonstrating the V 2 O 3 structure." The above-mentioned Supplementary Fig. 20 Supplementary Fig. 20 a-b, Raman mapping images of a 1T-WS 2 flake obtained in the J 1 (a) and E 1 2g (b) vibrational modes, respectively. The measured WS 2 flake was grown on the V 2 O 3 -Nuclei film substrate; c-d, Raman mapping images of a 2H-WS 2 flake obtained in the J 1 (c) and E 1 2g (d) vibrational modes, respectively. The measured WS 2 flake was grown on the V 2 O 3 -No Nuclei film substrate; e-f, Raman mapping image of the fresh commercial V 2 O 3 film obtained in the E 1 2g (WS 2 ) vibrational mode (e) and A 1g (V 2 O 3 ) mode. Scale bars: a-d, 1 μm; e-f, 5 μm. Supplementary Fig. 21 Raman spectra of the selected three points (A, B, C) on their respective substrates in Supplementary Fig. 20a, 20d and 20f, respectively.

Comment 9. The sample labeling may be a bit confusing. For example, from line 351 to 352: "To obtain pure V 2 O 3 film, we used only VCl 3 as the precursors with the same growth condition. To obtain V 2 O 3 film (no W), we used only VCl 3 and sulfur as the precursors with the same growth condition."
It is a bit difficult to understand the difference between "pure V 2 O 3 film" and "V 2 O 3 film (no W)" without referring to this sentence, and it may cause misunderstanding when reading the article.

Response 9.
We appreciate the reviewer very much for these very helpful suggestions.
Following your suggestions, we have renamed the V 2 O 3 film as follows: "To obtain the V 2 O 3 -No Nuclei film, we used only VCl 3 as the as the precursors with the same growth condition. To obtain V 2 O 3 -Nuclei film, we used only VCl 3 and sulfur as the precursors with the same growth condition." Accordingly, we have made corresponding changes and highlighted these changes in the revised manuscript (Page 9-11, Page 15-16) and supplementary information ( Supplementary Figs. 16-19).

However, it is difficult to tell if the V 2 O 3 is formed during the synthesis or during the characterization, as the V compounds (VS 2 , VCl 3 ) are extremely air sensitive. If the characterization was done in air, oxidation can not be avoided, thus it may lead
to misunderstanding in the mechanism study. Response 10. We appreciate the reviewer very much for these valuable comments.
We agree with the reviewer that V compounds such as VS 2 and VCl 3 are extremely air sensitive, especially for VCl 3 . Considering the sensitivity of VCl 3 to the air and moisture, the VCl 3 was stored in the glovebox before it was used. In this work, many characterizations for V 2 O 3 film have been done to confirm the crystal structure, for example, STEM image in Fig. 2h, XRD patterns in Supplementary Fig. 2 and Raman spectra in Supplementary Fig. 10. The STEM image of 1T@V 2 O 3 /sapphire sample in The XRD patterns (Fig. S2) and Raman spectra (Fig. S10) were conducted at room temperature without heating process. From the prominent peak of V 2 O 3 (006) in Fig.   S2, the formed V 2 O 3 film was single crystalline structure, which could not be transformed by the VS 2 or VCl 3 at the room temperature. Importantly, it was well-known that Raman spectroscopy was a very sensitive technique to detect the surface structure of the nanomaterials. If the V 2 O 3 film was gradually transformed by the VS 2 or VCl 3 at room temperature, typical signals of VS 2 or VCl 3 would emerge.
However, no signals that assigned to VS 2 or VCl 3 were observed, demonstrating that the Raman signals of V 2 O 3 were not caused by the oxidation of VS 2 or VCl 3 . Through a series of careful analyses, we could confirm that the V 2 O 3 film was formed during the synthesis and the proposed growth mechanism of 1T-WS 2 was reliable based on experimental and simulated results.

Comment 11. The authors claim that it is a controllable approach to synthesize 1T WS 2 . Regarding the controllability, could the author control the 2H/1T ratio through the synthesis, not post annealing?
Response 11. We appreciate the reviewer very much for the valuable question. To demonstrate the thermal stability of 1T-WS 2 monolayers, post annealing with different temperatures were conducted, as shown in Fig. 3e and Supplementary Fig. 14. Moreover, it was found that the amount of VCl 3 played an important role in controlling the 2H/1T ratio in our exploration of controllable 1T-WS 2 growth, as shown in Supplementary Fig. 23-26. It was observed that a high 2H/1T ratio was achieved when the amount of VCl 3 was ~0.02 g and only 13% 1T-WS 2 was obtained according to the XPS analysis ( Supplementary Fig. 26). When the amount of VCl 3 was in the range of 0.05 g~0.08 g, dominated 2H-WS 2 monolayers were grown, as demonstrated in Supplementary Fig. 24 (olive and purple plots) and Supplementary   Fig. 26. A low 2H/1T ratio was obtained when 0.03 g VCl 3 was used as the co-precursors according to the XPS analysis in Fig. 3c (91% 1T phase).
The above-mentioned Supplementary Fig. 23-26 has been added in the revised supporting information as follows: Supplementary Fig. 23 Optical micrographs of as-prepared products using different amount of VCl 3 . a, 0.01g; b, 0.02 g; c, 0.03 g; d, 0.05 g; e, 0.08 g; f, 0.10 g. T 1 =860 °C, T 2 = 160 °C, Ar/H 2 = 80 sccm/20 sccm, t = 15 min; T 1 refers to the heating temperature of the furnace, T 2 refers to the heating temperature of the sulfur powder, and t refers to the growth time. All the other experimental parameters are the same. Scale bars: 5 μm. f 0.10 g Supplementary Fig. 24 The Raman spectra of as-prepared products using different amount of VCl 3 . Supplementary Fig. 25 The Raman spectra of as-prepared V 2 O 3 film using different amount of VCl 3 . The Raman spectra were measured on the areas where no WS 2 domains appeared.  Supplementary Fig. 26 High resolution XPS spectra of W 4f (a) and S 2s (b) core level peak regions for WS 2 domains using different amount of VCl 3 . The fitting blue and pink curves represent the contributions of 1T and 2H phases, respectively. Please note that obtained WS 2 domains were transferred on the sapphire substrates for the XPS measurements.
Accordingly, the growth details and related discussion were added in the revised supplementary information as follows: "In order to demonstrate the amount of VCl 3 has a great impact on the phase purity of WS 2 , we have prepared WS 2 samples using different mass amount of VCl 3 (from 0.01 g to 0.10 g)…. Whereas, the density of VS 2 nuclei was the constraining factor to impact the 1T-WS 2 formation when the amount of VCl 3 was less than 0.03 g."

Comment 12. In addition, this approach replies on the introduction of V and the formation of 1T-VS 2 and V 2 O 3 based on the proposed mechanism, could the other V precursors, such as vanadocene, give the same effect? What will happen if you change the VCl 3 amount? Do you have the tunability towards the V concentration without changing the 1T character of WS 2 ?
Response 12. We appreciate the reviewer very much for these valuable questions.
According to your questions, we have made corresponding experiments to deeply understand the 1T-WS 2 growth. Firstly, we have tried to use vanadocene as the co-precursors and found that V SACs@1T-WS 2 monolayers were also achieved using a modified condition. It was well-known that vanadocene was commonly used in the organometallic chemical vapor deposition (OMCVD) due to its less stability and non-toxicity. However, vanadocene is more moisture-and oxygen-sensitive compared to VCl 3 . To obtain the 1T-WS 2 domains, different amounts of vanadocene have been investigated. It was found the optimized amount of vanadocene was ~0.08 g. The characterizations of V SACs@1T-WS 2 monolayers were shown in Supplementary Fig.   30. From the optical micrograph, the obtained 1T-WS 2 monolayers showed similar morphology with Fig. 2b. The atomic structure of V SACs@1T-WS 2 was displayed in Supplementary Fig. 30b, confirming the 1T phase of WS 2 . The typical metallic peaks were prominently seen in the Raman spectrum ( Supplementary Fig. 30c), indicative of the metallic feature of WS 2 domains. Moreover, the negligible PL intensity of obtained WS 2 domains was also provided to demonstrate the metallic nature ( Supplementary Fig. 30d). It was reported that vanadocene could be completely Accordingly, we have added the description of using vanadocene as co-precursors for the 1T-WS 2 growth on Page 11 in the revised manuscript as follows: "Moreover, vanadocene precursors were also investigated to enrich the growth method of 1T-WS 2 monolayers ( Figure S30)." Additionally, the growth details were added in Supplementary Fig. 30 in the revised supplementary information as follows: "Apparently, 1T-WS 2 domains were also achieved if using vanadocene as the co-precursors. …As a result, it required higher amount of vanadocene to obtain the shown in Supplementary Fig. 23-26. As replied to the comment 11, it was found that the amount of VCl 3 could significantly affect the controllable phase growth of WS 2 .
The details were seen in the Supplementary Fig. 23-26 and corresponding discussion part.
Finally, the tunability towards the V concentration without changing the 1T character of WS 2 was performed by varying the heating temperatures (from 840~865 º C). Generally, two aspects could be considered to tune the V concentration during the heating process. One way is changing the amount of VCl 3 , which could control the density of VS 2 nuclei formation. However, as demonstrated in Supplementary Fig.   22-26, either 2H phase or 1T phase was obtained by varying the amount of VCl 3 . The other way is to tune the heating temperatures (from 840~865 º C), as the density of VS 2 nuclei could also be slightly affected by the heating temperature. As expected, the  Figure 2.

Comment 13. For the HER measurements, since the VWS 2 is not a continuous film, how did the authors calculate the mass loading and the area? It seems that the author assume the film is a continuous film, which abbeys the characterization shown in
Response 13. We appreciate the reviewer very much for these questions. Firstly, we assumed the WS 2 electrocatalyst was a continuous film in the original manuscript, because the density of as-grown 1T-WS 2 monolayers was different in the different growth locations. As shown in Supplementary Fig. 30 Following your comment, we strongly agree with your viewpoint that assuming of continuous WS 2 film in this work was not accurate. Apparently, the assuming obeyed the characterization in Fig. 2. In order to evaluate the actual area of WS 2 monolayers more accurately, we made statistics for the actual area of 1T-WS 2 monolayers in Figure.  As for the Fig. S32b-32c, the ratio of 2 / . was apparently less than 1.0. In particular, the relatively lower ratios of 2 / . were shown in Fig. S32d-S32f according to the lower distribution density of WS 2 monolayers. To obtain the accurate 2 / . ratios of Fig. S32d- We have deleted the description of "approximately one continuous layer of WS 2 over the electrode surface area to 6.5 µg cm 2 " in the "Electrochemical Measurements" section to avoid any confusion, which has been highlighted on Page 18 in the revised manuscript.
We have deleted the annotation of "6.5 µg cm 2 " in the Supplementary Fig. S34 to avoid any confusion in the revised supplementary information.
Additionally, we have added one sentence of "TOF values were calculated using the mass loading of 6.5 μg/cm 2 ." in the Table S7 to avoid any confusion, which has been highlighted in Table S7

REVIEWER COMMENTS
Reviewer #2 (Remarks to the Author): The authors have well addressed my concerned issues.
Reviewer #3 (Remarks to the Author): The authors addressed part of the comments. However, the main issue now is the detection of V by XPS, and the possible oxidation of the samples due to the presence of V. If the V content ranges form 4at.% to 8at% in WS2 monolayer, XPS should not have any problem in detecting V. In addition, a careful study of O and S by XPS and EELS is needed to truly confirm that substitutional V is bonded to S, instead of forming VO2/V2O3. Therefore, I would encourage the authors to address these points before I can recommend publication of this work. Detailed comments are below: 1. In Figure 2c, could the authors provide the intensity profile of V SACs@1T-WS2 with V atoms, and compare it with the simulation? Figure 2f only shows the line profile with W and S atoms. The profiles with V, W, and S can help distinguish the V@W sites to vacancy@W sites.
2. The authors stated that a modified PMMA-assisted transfer process was applied to prepare V SACs@1T-WS2 samples for XPS, STEM and EELS characterizations. However, it is my understanding that this transfer method is based on mild chemical etching followed by mechanical delamination. Thus, I am still not convinced that the transfer process can detach V SACs@1T-WS2 from V2O3 and avoid the presence of vanadium oxides on various characterizations and HER tests. In addition, although V3+ state was not observed in EELS, it is noted that EELS can only reveal local coordination states at the nanometer scale, and may not be representative to claim the absence of V2O3 in the entire sample. Even if V2O3 was removed during the transfer process, other compounds such as VO2 or VS2 could be present. How did the authors rule out the possibility of VO2/VS2 within the samples? Could the authors provide the EELS and XPS of O and S as well?
3. The authors claimed that no V signal can be detected by XPS. However, on page 6, the authors mentioned that V is 4at%. XPS`s detection limit should be lower than that, and the technique is able to detect 4at% V in WS2. Similar work can be found in "Monolayer Vanadium-doped Tungsten Disulfide: A Room-Temperature Dilute Magnetic Semiconductor". In that work, 1.5at% substitutional V can be detected in WS2 monolayer. Could the authors provide the XPS of survey and O as well?
4. If for any reason, the authors still cannot detect V by XPS on the transferred sample, could the authors deconvolute the S 2p peak to see if there is any V-S bond formation?
5. The authors claimed the stability under ambient conditions, which is confirmed by the Raman. J1, J2, J3 peaks found samples stored under ambient conditions for a year. However, one main issue for V-doped samples is the stability against oxidation. Thus, XPS is recommended to identify V, O, S, and W for the samples.
6. Since HER was performed after transferring the film to glassy carbon substrate. How did the authors ensure a good adhesion between the film and the substrate? and avoid flakes from peeling off from the glassy carbon? For stability tests, amperometric I-t stability curves is recommended.
7. For TOF calculation, the authors only considered V as the active sites, thus it may overestimate the TOF. However, W edges, vacancies, defects in WS2 are also considered as the active sites. Thus, the TOF calculation seems not that accurate. ECSA is recommended to quantify the electrochemical active area and quantify the active sites. It would be better to compare the ECSA between samples with/without V as well.
8. Based on Fig. S31, both film and flakes can be found in one sample. Does that mean the sample morphology is not uniform? As can been seen in Fig.2a, it is clear that the color is not uniform across the substrates. Could the authors comment on the homogeneity of the sample? In addition, for film region, is it still monolayer with similar V concentration?

Comment. The authors addressed part of the comments. However, the main issue now is the detection of V by XPS, and the possible oxidation of the samples due to the presence of V. If the V content ranges form 4at.% to 8at% in WS 2 monolayer, XPS should not have any problem in detecting V. In addition, a careful study of O and S by XPS and EELS is needed to truly confirm that substitutional V is bonded to S, instead of forming VO 2 /V 2 O 3 . Therefore, I would encourage the authors to address these points before I can recommend publication of this work. Detailed comments are below:
Response. We appreciate the reviewer very much for his/her comments. His/Her concerns are addressed as follows.

SACs@1T-WS 2 with V atoms, and compare it with the simulation? Figure 2f only shows the line profile with W and S atoms. The profiles with V, W, and S can help
distinguish the V@W sites to vacancy@W sites.

Response 1.
We appreciate the reviewer very much for the very good suggestion.
According to your suggestion, we have made simulated STEM images with both V SACs@1T-WS 2 and W vac @1T-WS 2 to address this comment. The simulated STEM images were obtained with the help of experts from Prof. Gu Lin's group in China. As shown in Fig. S6-S7, the V elemental identity in 1T-WS 2 from different line profile sequences were clearly verified as compared to the simulated W Vac @1T-WS 2 .
Accordingly, the V elemental identities from different line profile sequences have been added in the revised manuscript to address this comment, which has been highlighted on Page 6 as follows: "Additionally, W-S-S-V-S-S-W and W-V-W intensity profile sequences from both experimental and simulated STEM images are also achieved to verify the V atoms replacement at W sites (see details in Fig. S6-S7)." The above-mentioned Supplementary Fig. 6 and Supplementary Fig. 7 have been added in the revised supporting information as follows: Supplementary Fig. 6 a- Figure S7e).

tests. In addition, although V 3+ state was not observed in EELS, it is noted that EELS can only reveal local coordination states at the nanometer scale, and may not be representative to claim the absence of V 2 O 3 in the entire sample. Even if V 2 O 3 was removed during the transfer process, other compounds such as VO 2 or VS 2 could be present. How did the authors rule out the possibility of VO 2 /VS 2 within the samples? Could the authors provide the EELS and XPS of O and S as well?
Response 2. We appreciate the reviewer very much for these insightful questions. substrate, which was shown in Figure R1. The XPS spectra of V SACs@1T-WS 2 sample on the HOPG substrate was shown in Figure S10-S11. Except for the signals of W4f and S 2p, the V 2p signals were detected after a long-term acquisition time during the XPS scanning cycles with much weaker intensity than W 4f (Fig. S11a) and S 2p (Fig. S11b), which should be caused by the low atomic density of V atoms in the 1T sample. After deconvoluting the V 2p signals, the peaks at 516.4 eV and 523.5 eV were assigned to V 4+ from V-S bond in 1T sample. [3][4][5][6][7] Please note that if VS 2 was present in the transferred sample, a prominent V 2p peak should be detected. However, only weak V signals were detected, indicating no VS 2 contamination in the transferred sample.
Importantly, if the V-based contaminations (e.g. V 2 O 3 , VO 2 , VS 2 ) were contained in the transferred 1T sample, obviously enhanced V/W molar ratio would be measured.
However, the molar ratio of W/V was ~95:5 from the XPS analyses, which was close According to the reviewer's suggestion, the high-resolution of O 1s in the transferred 1T sample on the sapphire substrate was also shown in Fig. S18b.
However, the signal of O 1s was ascribed to the sapphire substrate. Additionally, the high-resolution of S 2p was already provided in Fig. 3c (right) in the manuscript.
Accordingly, the absence of V-based contaminations (VS 2 , V 2 O 3 , VO 2 ) has been demonstrated by XPS spectra in the revised manuscript to address this comment, which has been highlighted on Page 6 as follows: "To exclude the presence of V-based contaminations (eg. V 2 O 3 , VO 2 and VS 2 ) in the transferred 1T sample, XPS spectra of 1T sample transferred on highly oriented pyrolytic graphite (HOPG) were performed (see details in Figure S10-S11)." Accordingly, the discussion of absence of V-based contaminations (VS 2 , V 2 O 3 , VO 2 ) has been added in the revised supporting manuscript to address this comment, which has been highlighted in Supplementary Fig. 11   Supplementary Fig. 11 a- Accordingly, the description of EELS of S and O has been added in the revised manuscript to address this comment, which has been highlighted on Page 6 as follows: "The EELS of S and O spectra in Figure S5 were   reason that no V signals were detected in our first detection of high-resolution V 2p in the transferred sample was that the short time of acquisition time during the measurement. Hence, to figure out whether the V signals were detected, we have tried to measure the high-resolution of V 2p with long-term acquisition time.

Comment 3. The authors claimed that no V signal can be detected by XPS. However
Simultaneously, it was mentioned that the reviewer has raised the questions of VS Accordingly, the detection of V signals has been demonstrated by XPS spectra in the revised manuscript to address this comment, which has been highlighted on Page 7 as follows: "The V signal in Fig. S11c was ascribed to the V-S bond 41 , which was consistent with the EELS result in Fig. 2e." According to your suggestion, both the survey scan and high resolution XPS spectra are provided as Supplementary Fig. S10 and Supplementary Fig. S11 in the revised supplementary manuscript, please refer to the response to Comment 2.

Figure R2
High-resolution XPS spectra of W 4f (a), S 2p (b), V 2p (c) and O 1s (d) core levels of V SACs@1T-WS 2 transferred on HOPG substrate. The spectra were obtained at different acquisition time.

Comment 4. If for any reason, the authors still cannot detect V by XPS on the transferred sample, could the authors deconvolute the S 2p peak to see if there is any V-S bond formation?
Response 4. We appreciate the reviewer very much for the good suggestion. As demonstrated in Figure R2, the V signals were detected after longtime scanning cycles during the XPS analyses. And after deconvoluting the V 2p peak, the signal of V 2p at 516.4 eV was assigned to V 4+ , which should be from the V-S bond formation in the V is difficult to get accurate V-S bond information from the S 2p peak.

Comment 5. The authors claimed the stability under ambient conditions, which is confirmed by the Raman. J 1 , J 2 , J 3 peaks found samples stored under ambient conditions for a year. However, one main issue for V-doped samples is the stability against oxidation. Thus, XPS is recommended to identify V, O, S, and W for the samples.
Response 5. We appreciate the reviewer very much for the good suggestion. We agree the reviewer that the V-doped samples were not stable. That's why obviously enhanced intensity of E 1 2g /A 1g was seen in Fig. S19, indicative of a decreasing 1T phase. As it was known that the surface of V 2 O 3 film could be gradually oxidated upon exposure to the ambient condition. According to your suggestion, XPS spectra of 1T sample without transferring (after one year) were obtained to identify the oxidation degree of the 1T sample and V 2 O 3 film, as shown in Fig. S20. The wide-scan spectrum was shown in Fig. S20a and the main components are O, V, W, S, Al (from the sapphire). A high amount of 1T phase is still preserved even after one year (~60%) from Fig. S20b, indicating the durable stability of 1T sample. Meanwhile, the V3p was found in Fig. S20b, which could not be used to analyze the V-base oxidation states caused by a complicated interaction between V3p and 3d electrons. 8 However, three contributions were found in the V 2p XPS spectrum, as shown in Fig.   S20d: V 3+ , V 4+ , V 5+ , respectively, at 514.47 eV, 516.19 eV and 517.40 eV. 8 The main component is V 5+ , which should be caused by the oxidation of V 2 O 3 film on the surface in the air. The V 4+ signal from V-S bond in the 1T sample was also involved due to the overlapped peak position of VS 2 and VO 2 . Please note that no VO 2 or V 2 O 5 signals were detected by the Raman from Fig. S19 (red plot), which should be caused by overwhelming signals from 1T sample and V 2 O 3 film on the surface of sample.
Accordingly, the XPS spectra for the 1T/V 2 O 3 on the sapphire substrate after one year have been added in the revised manuscript to address this comment, which has been highlighted on Page 9 as follows: "Importantly, a high 1T phase (~60%) is preserved in the sample from the XPS analysis in Figure S20b, and the decrease of 1T/2H ratio should be probably caused by the oxidation of V 2 O 3 film on the surface, as demonstrated in Figure S20d." The above-mentioned Supplementary Fig. S20 has been added in the revised supporting information as follows: Supplementary Fig. 20 a, Survey spectra of V SACs@1T-WS 2 /V 2 O 3 film grown on sapphire substrate after 1 year; b-d, High-resolution XPS spectra of V 3p (b), W 4f (b), S 2p (c), O 1s (d) and V 2p (d) core level peak regions. The fitting blue and pink curves represent the contributions of 1T and 2H phases in Fig. S20b-c, respectively.
The discussion of XPS spectra for the 1T/V 2 O 3 on the sapphire substrate after one year has been added in Supplementary Fig. 20 in the revised supplementary information as follows: "The wide-scan spectrum was shown in Fig. S20a

substrate? and avoid flakes from peeling off from the glassy carbon? For stability tests, amperometric I-t stability curves is recommended.
Response 6. We appreciate the reviewer very much for the good question. It was well-known that the interaction between the transferred WS 2 film and the glassy carbon electrode was weak Van der Waals's force. To enhance the adhesion and protect the film from peeling off from the electrode, appropriate Nafion film (a conductive polymer) was casted onto the surface of WS 2 film during the HER catalytic measurements. Please note that this kind of protection was widely used in many previously published literatures. [10][11][12][13][14][15] Following your suggestion, the stability test was conducted in 0.5 M H 2 SO 4 electrolyte for 100 h. As revealed by the chronoamperometric curve of V SACs 1T-WS 2 electrocatalyst in Supplementary Fig. 47, the current density for the V SACs 1T-WS 2 electrocatalyst displayed a slight current decay of 1.0 mA cm -2 after 24 h and 3.4 mA cm -2 after 100 h, indicating a high stability of V SACs 1T-WS 2 catalyst.
Additionally, the good stability of HER performance suggested a strong adhesion between WS 2 film and glassy carbon electrode.
Accordingly, the description of the electrochemical stability test has been added in the revised manuscript to address this comment, which has been highlighted on Page 14-15 as follows: "The stability test was conducted at current densities higher than 10 mA/cm 2 in 0.5 M H 2 SO 4 electrolyte for 100 h. As revealed by the chronoamperometric curve of V SACs 1T-WS 2 electrocatalyst in Figure S47, the current density for the V SACs   Figure S46. Half of the linear slopes obtained in Figure S46b-d were equivalent to the C dl . Apparently, the C dl of V SACs@1T-WS 2 was 139.5 μF/cm 2 , which was much higher than that of 2H-WS 2 (61.7 μF/cm 2 ). The reference specific capacitance (C s ) of 40 μF/cm 2 is used in this work. [16][17][18][19] According to the equation:   Fig. 5a revealed that both V atoms and high purity 1T phase gave the contribution to the higher HER activity than 2H 1T -WS 2 or pure 2H-WS 2 .
Nevertheless, V single atoms dominantly determined the high HER performance from the DFT results, as the intrinsic activity of 1T-WS 2 was much lower than V SACs@1T-WS 2 due to the relatively larger free energy of H adsorption (∆G H ) of 1T-WS 2 (0.28 eV) than V atom sites decorated 1T-WS 2 (0.05 eV, Table S8).
Additionally, active sites from the W edge (or S edge) could be negligible due to the large free energy of H adsorption (∆GH) (Table S8). Importantly, the atomic density of V is much higher than the density of vacancy defects from the STEM image in both 1T sample (Fig. 2c) and 2H 1T (Fig. S8) Table S7 and discuss section in Fig. S43).
Accordingly, the description of ECSA has been added in the revised manuscript to address this comment, which has been highlighted on Page 13 and page 20 as follows: Does that mean the sample morphology is not uniform? As can been seen in Fig.2a  region where merged 1T film grown. To transfer the 1T samples for the investigation of HER catalytic performance, we transferred the samples at the region B (near the interface between region B and C). However, it was probable to transfer a small part of 1T merged film onto the glassy carbon electrode. To acquire a relatively accurate geometric area, we made statistics for the real geometric area of 1T samples (considering both isolated flakes and merged film, see details in Fig. S43-44).
Finally, the characterizations of 1T merged film have been provided in Fig.   S40-S42. The atomic structure of V SACs@1T-WS 2 merged film were shown in Fig.   S40. The images were randomly taken from four different areas in region C. The 1T phase of WS 2 was confirmed with V atoms distribution from the high-resolution STEM images. In particular, the average single V atoms was ~4.1 at% (2.1 wt%), which was close to the atomic density of V atoms in the growth region B (Fig. 2c,   ~2.0 wt%). The W-S-S-V-S-S-W intensity profile sequences were also achieved to identify the V atoms replacement at W sites, as shown in Fig. S40e. The typical metallic peaks were shown in the Raman spectrum ( Supplementary Fig. 41 Firstly, we have overlayed the experimental and simulated line intensity profiles together, as shown in Figure S6c. From the result, the experimental profile showed a coincident W/V peak intensity ratio with the simulated one after normalizing the peak intensity based on the intensity of the W peak. Peale note that an average experimental line intensity profile was obtained from 27 different line intensity profiles with W-S-S-V-S-S-W sequence in Figure R1a (from Figure 2c in the revised manuscript). Secondly, the reviewer mentioned that different peak intensity ratios of V and S (near to V) were observed in the previously experimental and simulated profiles. It's worth emphasizing that the peak intensity (even for the same element) was slightly different due to the limitation of detection accuracy, which has been observed in many published literatures. 1,2 For example, the W peak intensity and V peak intensity profile in the V-doped 2H-WSe 2 showed slightly different intensity in both W-Se-V-Se-W-Se sequence and W-W-V-V-W-W sequence ( Figure S3, Adv. Mater. 2020, 2003607) 2 . We have cited this paper in the revised manuscript (please see reference 36). In the case of present V VSACs@1T-WS 2 sample, the peak intensity ratio of V or S was slightly different if choosing different line sequence in our sample. Hence, to obtain a convincing experimental intensity profile, we obtain an average experimental line intensity profile from 27 different line intensity profiles with W-S-S-V-S-S-W sequences ( Figure R1a). And the final V peak intensity is slightly higher than the S peak intensity, which is generally consistent with the simulation ones. Thirdly, we appreciate your suggestions of "In to address this comment. As the much lower atomic number of V than W, V atoms in the samples were not visible as well as the simulated STEM image in Figure S6b.
Hence, it was very necessary to compare the experimental V SACs@1T-WS 2 with the simulated W vac @1T-WS 2 . So, we compared the experimental line intensity profile of V SACs@1T-WS 2 with simulated line intensity profiles of V SACs@1T-WS 2 ( Figure   S6b) and W vac @1T-WS 2 ( Figure S6c) from W-S-S-V-S-S-W ( Figure S6) and W-W-V-W-W sequences ( Figure S7), respectively. The experimental line intensity profile of V SACs@1T-WS 2 in Figure S7 was obtained from an average 28 different line intensity profiles with W-W-V-W-W sequences ( Figure R1b). Fourthly, the reviewer was wondering the coordination of W vac in the previous simulated line intensity profiles of W vac @1T-WS 2 . In that simulation, we only considered W vac defect with the removal of W atom, as shown in Figure S6d. Additionally, the W vac defect with the removal of W atom and 2S vacancy was also shown in Figure S6e.
The corresponding intensity profiles were shown in orange dots and pink dots in Figure S6h, respectively. Finally, according to your suggestion, we have also considered only one S vacancy defect with one W and S atom and compared it with 1T-WS 2 without any vacancy defect, as shown in Figure S6f and Figure S6g, respectively. The corresponding intensity profiles were shown in olive dots and blue dots in Figure S6h, respectively.