Atomic-level insight into super-efficient electrocatalytic oxygen evolution on iron and vanadium co-doped nickel (oxy)hydroxide

It is of great importance to understand the origin of high oxygen-evolving activity of state-of-the-art multimetal oxides/(oxy)hydroxides at atomic level. Herein we report an evident improvement of oxygen evolution reaction activity via incorporating iron and vanadium into nickel hydroxide lattices. X-ray photoelectron/absorption spectroscopies reveal the synergistic interaction between iron/vanadium dopants and nickel in the host matrix, which subtly modulates local coordination environments and electronic structures of the iron/vanadium/nickel cations. Further, in-situ X-ray absorption spectroscopic analyses manifest contraction of metal–oxygen bond lengths in the activated catalyst, with a short vanadium–oxygen bond distance. Density functional theory calculations indicate that the vanadium site of the iron/vanadium co-doped nickel (oxy)hydroxide gives near-optimal binding energies of oxygen evolution reaction intermediates and has lower overpotential compared with nickel and iron sites. These findings suggest that the doped vanadium with distorted geometric and disturbed electronic structures makes crucial contribution to high activity of the trimetallic catalyst.

the lack of peaks at high R in the EXAFS FT. d) Please show the "NiFeV" spectra at the Ni-K, Fe-K and V-K edge in a single plot to support the short V-O distance for this review. If this figure is included in revised manuscript, state clearly that phase contributions of Ni, Fe, V are not identical during FT (i.e. the reduced distances on the x-axis are not identical). e) The V-K pre-edge is lowered after OER. This could indicate a change in symmetry. The V-L spectra cannot exclude a change in symmetry without detailed analysis of the multiplet structure. f) P16. "the V−O distance is also shrunk from 1.69 to 1.66 Å" The V-O bond is not shorter within the uncertainty given in the supporting table.
3) How many electrodes were prepared for each oxide? Are the measurements reproducible? 4) Mechanism a) I like to point out that the mechanism of the LDH is still actively disused, e.g. in onlinelibrary.wiley.com/doi/10.1002/aenm.201600621/full Can the author's data support one of the previous proposals or the selection of the used reaction steps? b) The point of zero charge (PZC) of oxides is below pH 12, so at pH 14, the equilibrium should be shifted toward *O. I disagree that it should be a possible RLS since the solution thermodynamics favor its generation. The potential of charge neutrality is also much lower than the onset of the OER on oxides. Concept of PZC: http://old.iupac.org/publications/pac/1976/pdf/4804x0415.pdf and http://www.gly.uga.edu/railsback/Fundamentals/8150PointofZeroChar ge05Pt1P.pdf 5) EXAFS analysis a) Please mention for the readers that Fig. 3 and 6 show the reduced distance. The interatomic distance is determined by the fits. b) Please state the details of EXAFS FT. What was the windowing function and its parameters? What K-space range was used? What R-space range was analyzed/fit? This information is necessary for reproduction of the results c) Phases from experimental references were used. Which reference materials were used? How were they extracted. This can significantly affect the obtained fit values. Could the low V-O distance be an artefact of the experimental phase function? d) How is the uncertainty in the tables of the EXAFS analysis calculated? Is it the parameter error? e) How do the fit results compare to ref. 10 and 14? This should also be included in the revision for the readers.
6) The intensity (or better area under the curve) of soft XAS spectra does depend on the number of holes but also on other factors a) How were the spectra normalized? b) How do the soft XAS spectra look at different positions on the sample? Are the changes significant? c) Can the observed changes due to change in the total absorption when another element is added to the oxide? d) A significant change in holes on the element should also lead to a shift of the peak position, which is only observed for V. 7) DFT a) Which metal valences were found in the DFT calculations? Do they match the trends obtained from XAS? The DFT and XAS results could be better connected. b) Why are there different circles of the same color in Fig. 7c? What models do they belong to? Please provide a supporting table with detail or other means of documentation c) Are surrounding atoms and the electrolyte considered in calculations of the oxygen adsorption? Please add detail in the main text d) P.19. "The highest activity is achieved". This refers to a calculation and thus "is predicted". 8) Errors or ambiguous discussion a) P10 "For Ni3Fe (oxy)hydroxide, the unpaired electron in the π-symmetry (t2g) d-orbital of Fe3+". The Fe-L spectrum in Fig. 4e is clearly that of high spin Fe3+ (t2g3 eg2). Thus, there are three unpaired t2g electrons. I also recommend giving the t2g and eg occupancies instead of the d-electron count on P10, which would make it easier to follow. Lastly, the soft XAS spectra should be discussed in terms of t2g/eg occupancy or at least whether the ion is high or low spin, which can be done, e.g., using the branching ratio (L3/L2+L3). b) The authors use "charge transfer" frequently. It is confusing and imprecise. Which charge carrier? Electron or hole? It would be clearer to state which element would be reduced or oxidized. c) P14. The ECSA was not measured. It is obtained by dividing the capacitance of the sample by that of a perfectly flat reference of the same material, see ref.
3. The used normalization corrects for difference in roughness among the samples and is thus useful but should not be called ECSA and cannot be compared to catalysts on other substrates. d) Supplemental Figures S5+S6 do not show "atomic resolution". They are blurry and individual atomic columns are not resolved. e) Supplemental P37. NiOOH was not found by PXRD or HRTEM. It is only detected by Raman and in situ XAS. 9) Minor/typo: a) Supplemental Fig. 10. What does the label Ni stand for? Ni metal? Ni(OH)2? Please add simple Ni oxide references. b) The "Janus face property" is called amphoteric in chemistry or is something else meant? c) P14. Which of the two charge transfer resistances is reduced? d) Supplemental Table 9. Please add potential at which EIS was performed and refer to circuit d) Please index PXRD patterns. Which reflections change/shift? Are those assigned to the metal oxide slabs or to the distance between the metal oxide slabs (of the LDH)? e) I was confused by Fig. 1, where the substrate looks like a metal mesh. The carbon fiber substrate is not as ordered as shown there.

Reviewer #2 (Remarks to the Author):
This is a very comprehensive paper of very high scientific quality and considerable topical interest. The development of metal oxide electrocatalytic materials which catalyse the electro-oxidation of water to generate molecular oxygen is a grand challenge in energy science. Doping metal oxide lattices with small quantities of foreign metal atoms has been shown to be a useful partway to lower the overpotential required to drive the water oxidation at a fixed current density (say 10 -20 mA/cm2).The process of doping nickel oxyhydroxide with iron has been shown over the last few years, to significantly enhance the catalytic activity of the bimetallic catalyst.
In the present paper the authors have convincingly shown that the addition of vanadium to the bimetallic catalyst can result in even better catalytic activity. An overpotential of ca. 200 mV at 10 mA/cm2 with a Tafel slope of ca. 39 mV/dechas been measured for a Ni<sub>3</sub>Fe<sub>0.5</sub>V<sub> ;0.5</sub> oxyhydroxide material. These are impressive performance indicators. This is a very significant result.
The paper is very complete. The characterization and analysis of the material and its performance under operating conditions has been rigorously examined. Furthermore the paper contains significant DFT analysis as an aid to mechanistic interpretation.
Given the quality of the science and its inherent broad interest and topicality, I strongly recommend that this paper be accepted for publication.
Reviewer #3 (Remarks to the Author): This manuscript aims to unveil OER improvements through Fe/V co-doping in Ni(OH)2 by comprehensive spectroscopic studies including XAFS, sXAS and Raman along with advanced TEM of catalysts ex-situ and in-situ of OER. The experimental results have been combined with calculation to get a conclusion that V site is the active sits in this catalysts. the methodology is appealing and should deliver new scientific insights for this important catalyst. unfortunately the current version doesn't get there. therefore I recommend a major revision plus another round review. in the revision the author needs to carefully address those questions: 1) V in situ XAFS needs to be conducted since this is the proposed active sites. I understand that in situ work for V is a bit challenge but it is still doable.
2) from in situ XAFS, the structure of Ni/Fe in the active catalyst under OER shall be resolved; therefore this structure information shall be used in calculation rather than using models from literature.
3) the manuscript only "proves" the knowing knowledge that V is the active sites from calculation (optimal binding energy for OER intermediates at V site) and XAFS (ultra short V-O bond). a more deep discussion that how such local structural change led to the calculated results are need and this probably unveil atomic level new insights of the high OER catalysis in this known catalyst. 4) in addition to those above main concerns the claim of a uniform substitution of Fe/V into Ni(OH)2 lattice to substitute Ni sites cannot be supported by TEM-EDX mapping. EXAFS at V edge also seems against the claim.
Title: Atomic-level insight into super-efficient electrocatalytic oxygen evolution on iron and vanadium co-doped nickel (oxy)hydroxide Dear editor, Thank you for your e-mail of Feb. 17. We appreciate referees' professional comments.
The manuscript has been carefully revised according to the referees' comments. The changes in the manuscript and the responses to referees' questions are listed below.
Some new experimental results have been added to the revised manuscript. We would be grateful if you and referees find the revised manuscript acceptable for publication in Nature Communications. In addition, the NiFe 2−x V x O 4 catalysts were only characterized by IR and XRD in ref.
27, which did not provide convincible evidence for substitutionally doping of V into the lattices of host materials, let alone the information about the active sites of the multimetal catalysts under OER conditions. By contrast, we have employed many advanced techniques, such as ex-situ soft and in-situ hard XAS, in-situ EC-Raman spectroscopy, SEM, HR-TEM and atomic-resolution BF-TEM, aberration-corrected HAADF-STEM, EDX, XPS, and PXRD, as well as computational studies, to provide convincible evidence for co-doping of V and Fe into the Ni(OH) 2 lattices and to reveal how V dopant interplays with other metal ions co-existing in the trimetallic catalyst, and how the doped V cations contribute to the high OER activity of the host material. As we have stated in the Introduction, to date, there is no report on in-depth spectroscopic studies of local coordination environment and electronic structure for the V-containing bi-and trimetal (oxy)hydroxide OER catalysts in both rest and activated states. This is the key point and the new insight of the present work.
It is also worthy of note that we fabricated the Ni 3 Fe 1-x V x electrodes with a protocol completely different from that used in ref. 27, in which the NiFe 2-x V x O 4 electrodes were fabricated by multi-steps: (i) the precursor of NiFe 2−x V x O 4 was precipitated from the basic solution (pH 11) at 70 °C; (ii) the solid powder obtained was dried at 100 °C for 24 h and then treated at 600 °C for 5 h; (iii) the as-prepared NiFe 2−x V x O 4 catalyst was mixed with glycerol to form a slurry, which was painted onto a Ni substrate; finally (iv) the NiFe 2−x V x O 4 film electrode was treated at 350 °C for 1.5 h. By contrast, we directly grew Ni 3 Fe 1−x V x (oxy)hydroxides on the carbon fiber paper (CFP) by hydrothermal synthesis at 120 °C, without post high temperature treatment. Compared to the literature method, the advantages of the hydrothermal synthesis protocol are (i) more operation-convenient and less energy-consuming, (ii) forming better contact between the OER catalyst and the conductive substrate, and (iii) resulting in controllable morphology of the catalyst (an ultrathin nanosheet array on CFP) to have more active sites exposed to electrolyte. As a result, our optimized electrode of the Ni 3 Fe 0.5 V 0.5 (oxy)hydroxide displays much higher activity and better stability for OER Response 2a: Thanks for your good advice. We try to make a clear statement for this question by adding the following sentence to the manuscript at page 11: "The comparative analyses of XPS and XANES spectra suggest that co-doping of Fe together with V into Ni(OH) 2 lattices results in more electron transfer to the V in Ni 3 Fe 0.5 V 0.5 compared to that in Ni 3 V (Fig. 4c,f and Supplementary Fig. 16c)." The expression "charge transfer" has been revised to "partial electron transfer" in the entire manuscript.
An optimal binding strength is required for catalysis (i.e. Sabatier's principle) as also seen in Fig. 7c. Perhaps "optimal binding" is clearer for the readers and please refer to Response 2b: Thanks for your suggestion. To be clearer, the relevant sentence at page 4 the V atoms with higher d-band center possess less occupancy of the antibonding states with adsorbed oxygen intermediates, and thus exhibit optimal binding with regard to Ni and Fe atoms (Fig. 7c) Response 2c: In our opinion, the low t 2g and e g occupancies of V n+ (n = 3, 4, and 5) as compared to those of Fe 3+ and Ni 2+ make the redox property and the OER activity of the V sites distinguishing from those of Fe and Ni sites in the Ni 3 Fe 0.5 V 0.5 (oxy)hydroxide system.
In (ii) On the basis of these findings, it can be envisaged that by using co-doped metals other than Fe, such as Cr, Mn, and Co with different atomic radius, electronegativity, and d-band center from those of Fe, the modulation for the local coordination environment and electronic structure of V in the Ni(OH) 2 lattices could be regulated, which may further improve the OER catalytic activity of the Ni/M/V trimetallic catalysts and expand the scope of highly-active Ni(OH) 2 -based OER electrocatalysts.
The sentence, "[These findings] give some helpful hints for reasonably designing new V-containing catalysts" at page 23, has been revised according to the above paragraphs. Response 3a: Thanks for informing us this valuable reference. It reveals that in V K-edge XANES spectra, the intensity of the pre-edge peak depends predominantly on central site symmetry, while the absorption edge position is correlated to the oxidation state of central site. The intensity of the pre-edge peaks enhances in an ascending order of V 2 O 3 < V 2 O 4 < V 2 O 5 , predominantly due to the distortion of the coordination geometry from octahedral symmetry. Meanwhile, the absorption edge position of V positively shifts to higher energies with valence increase. In our case, the V K-edge XANES spectra of Ni 3 V and Ni 3 Fe 0.5 V 0.5 exhibit intense pre-edge peaks (Supplementary Fig. 15c), indicating the distorted coordination environment around V atoms in these materials. More interestingly, Ni 3 Fe 0.5 V 0.5 shows a higher pre-edge peak than that of Ni 3 V in the V K-edge XANES, implying a higher degree of octahedral geometry distortion at the V sites in Ni 3 Fe 0.5 V 0.5 compared to those in Ni 3 V. The K-edge absorption positions of Ni 3 V and Ni 3 Fe 0.5 V 0.5 are more close to those of VO 2 and V 2 O 5 than to that of V 2 O 3 (inset of Supplementary Fig. 15c), suggesting that the majority of V ions are in the formal valences of +4 and +5 in both catalysts.
Consistently, the XP spectra also show that the V atoms are predominantly in high oxidation states (+4 and +5) in Ni 3 V and Ni 3 Fe 0.5 V 0.5 , together with a minority of V 3+ (Fig. 4c) (1984)].
The related original discussion at page 8 has been revised to "Importantly, the V K-edge XANES spectra of Ni 3 V and Ni 3 Fe 0.5 V 0.5 exhibit intense pre-edge peaks ( Supplementary Fig. 15c), indicating the distorted coordination environment around V atoms in these materials 42 . More interestingly, Ni 3 Fe 0.5 V 0.5 shows a higher pre-edge peak than that of Ni 3 V in the V K-edge XANES, implying a higher degree of octahedral geometry distortion at the V sites in Ni 3 Fe 0.5 V 0.5 compared to those in Ni 3 V.
Additionally, the K-edge absorption positions of Ni 3 V and Ni 3 Fe 0.5 V 0.5 are more close to those of VO 2 and V 2 O 5 than to that of V 2 O 3 (inset of Supplementary Fig. 15c), suggesting that the majority of V ions are in the formal valences of +4 and +5 in both catalysts." In addition, the corresponding original discussion on the V 2p 3/2 peaks at page 9 was also revised as follows: "The V 2p 3/2 peak ( (ii) To further confirm the V substitution for Ni in the Ni(OH) 2 lattice, the wavelet transform (WT) analysis of the V K-edge EXAFS data was made, which reveals a similar feature of Ni 3 Fe 0.5 V 0.5 to that of Ni 3 V, i.e., the WT maximum appears at the cross point of R = 2.8 Å and k = 7.8 Å −1 ( Supplementary Fig. 7), implying the presence of V-Fe/Ni scatterings at a distance of around 2.8 Å surrounding V atoms. This affords direct evidence for the substitution of V atoms for the Ni sites in the Ni(OH) 2 lattice.
(iii) We also made the calculation of the EXAFS spectra by assuming V adsorption on the LDH layer or occupying the interstitial position. It reveals that in both cases the calculated spectra are quite different from the experimental spectra ( Supplementary Fig.   8).
(iv) Furthermore, DFT calculations suggest that V atoms initially placed on the top site of surface Ni or O atoms are relaxed to the interstitial between two LDH layers after structure optimization. The LDH structure with interstitial doping is noticeably buckled, Besides, the related discussion at page 7 has been revised to: "On the other side, from Supplementary Fig. 11, the nearest-neighbor FT peak position of V is shifted to the lower-R side and the second coordination peak to higher-R side with apparently reduced intensity as compared to that of Fe. This implies the remarkable different local environment of the substitutional V from that of Fe in Ni 3 Fe 0.5 V 0.5 ." Question 3e: "The V-K pre-edge is lowered after OER. This could indicate a change in symmetry. The V-L spectra cannot exclude a change in symmetry without detailed analysis of the multiplet structure." Response 3e: Please notice that a change in symmetry could lead to not only a change of the intensity of the pre-edge peak at V K-edge, but also a significant change of the spectral feature of the entire XANES spectrum due to the sensitivity of XANES to the geometric structure around the V atoms, just like what is shown in Phys. Rev. B 30, 5596−5610 (1984). However, after the OER at 1.75 V, except for a decrease in the pre-edge peak intensity (the decrease extent is much smaller than that caused by the change in symmetry), our V K-edge XANES shows identical spectral features (including shape and intensity of other characteristic peaks) to those measured before OER (original Fig. 6f). Therefore, we consider that the slight decrease in the intensity of the pre-edge peak at V K-edge for Ni 3 Fe 0.5 V 0.5 is most likely caused by electron transfer, rather than by a change in symmetry. Similarly, no other obvious change is visible at the V L-edge spectra after OER ( Supplementary Fig. 25), except for the decrease in the intensity of the characteristic peaks. These observations indicate no change of the coordination environment around V. That is to say, V atoms still occupy the sites of Ni after OER.
Nevertheless, as we have successfully conducted the in-situ V K-edge XAS experiment of Ni 3 Fe 0.5 V 0.5 according to the comment of Referee #3, we used the new V K-edge XANES spectra to replace the original Fig. 6f, and the related discussion on this problem has been rewritten in the revised manuscript at pages 17−18.       Fig. 29a,b) Fig. 29a, b) and will limit the reaction rate of OER process."

Question 6: "EXAFS analysis"
Question 6a: "Please mention for the readers that Fig. 3 and 6 show the reduced distance. The interatomic distance is determined by the fits." Response 6a: Thanks for reminding us of this point. In the revised manuscript, we have change the x-coordinate from "R" to "Reduced distance" in Figs 3 and 6, as well as in Response 6c: Phase-shift function is very important in extracting structural parameters from EXAFS curve-fittings. Before 1990s, this function was mostly extracted from the experimental spectra of standard materials with known structure. But it suffers from many disadvantages, e.g., in many cases it is hard to find the proper standard material, and the distance determination is not very accurate. Ever since the development of the ab initio multiple-scattering code FEFF, it could provide a reliable and convenient theoretical standard for most applications and hence overcome the disadvantages of experimental phase-shifts. Nowadays, FEFF has been used as a widely-accepted method to generate the phase-shifts for EXAFS fittings. And we also used FEFF generated phase-shifts instead of the experimental phase-shifts in the whole paper. However, the systematic errors in the data are much more difficult to estimate. For this reason, the c 2 estimating by Ifeffit usually does not include the systematic errors.
Fortunately, the high quality of X-ray beam from the advanced SR ensures this error is very small and much steady. Therefore the systematic error may hardly influence on the comparison between local structures of different samples.
The uncertainties of the variable parameters will be estimated immediately after the best-fit values of these parameters are found, by estimating the function c 2 . The uncertainty in the value of a variable is the amount by which c 2 can be increased and still have a value below some limit. Generally, a common criterion has been used in Ifeffit. That is c 2 deviates the best-fit value no more than 1. Below this limitation, the uncertainty of a variable will be calculated. In the Ifeffit, when the uncertainty in a variable is evaluated, all the other variables are allowed to vary, so that the correlations between variables can be taken into account. A brief description on the uncertainty calculation has been added to the section of "EXAFS data analysis" at page 43 in the revised Supplementary Information. Response 7b: For optimizing the XAS measurements, we collected several XAS spectra by irradiating the soft X-ray beam at different positions on each sample. No big difference was found among these XAS spectra. This is mainly due to the uniformity of our samples, which were fine powders mixed with graphite and pressed into a 13 mm diameter pellet. The sample is much larger than the beam size of the soft X-ray (3×1 mm 2 ), guaranteeing the reliability of our soft XAS spectra. This point has been mentioned at page 26 in the section of "Methods".
Question 7c: "Can the observed changes due to change in the total absorption when another element is added to the oxide?" Response 7c: When another element is added to the oxide, the total absorption will increase, because it increases the background absorption. But this does not change the characteristic peaks of the absorbing elements (here V, Fe, and Ni), because they are solely determined by the quantity of the absorbing elements. This is why XAFS could be called an "element-specific" technique.

Question 7d: "A significant change in holes on the element should also lead to a shift of the peak position, which is only observed for V."
Response 7d: Yes. The L-edge absorption arises from the transition of 2p electron to the unoccupied 3d orbitals, namely, 3d holes. Hence, L-edge XAS provides a means to detect the 3d holes. Numerous experimental and theoretical studies have shown that a significant change in 3d holes leads to changes of both intensity and position of L-edge peaks. As for our L-edge XAS spectra of Ni 3 Fe 0.5 V 0.5 , not only V edge, but also Fe and Ni edge spectra show shifts in peak positions (Fig. 4d,e,f); the shifts of the latters are less significant, but still visible. Following the referee's suggestion, we added the following information to the main text (pages 18−19) of the revised manuscript: "The model surfaces are covered by either water molecules or oxygen species that are possibly present in the reaction media. These models with different covered species give very similar results on the catalytic properties ( Fig. 7c and Supplementary Table   6)." Question 8d: "P.19. 'The highest activity is achieved'. This refers to a calculation and thus 'is predicted'." Response 8d: The sentence has been revised to "The highest activity is predicted on the V site of …" at page 22 of the revised manuscript.
Question 9: "Errors or ambiguous discussion" Question 9a: "P10 'For Ni 3 Fe (oxy)hydroxide, the unpaired electron in the π-symmetry (t 2g ) d-orbital of Fe 3+ '. The Fe-L spectrum in Fig. 4e is clearly that of high spin Fe 3+ (t 2g 3 e g 2 ). Thus, there are three unpaired t 2g electrons. I also recommend giving the t 2g and e g occupancies instead of the d-electron count on P10, which would make it easier to follow. Lastly, the soft XAS spectra should be discussed in terms of t 2g /e g occupancy or at least whether the ion is high or low spin, which can be done, e.g., using the branching ratio (L3/L2+L3)." Response 9a: Thanks for referee's valuable suggestion. Following this suggestion, we calculated the branching ratio, L 3 /(L 2 +L 3 ), at the Fe L-edge according to the literature [J.
Obviously, the calculated high-spin L 2,3 -edge XAS spectrum could well produce the experimental data, affording more evidence for the high-spin configuration of Fe 3+ substituting the Ni sites. The related calculation process of the L 2,3 -edge XAS theoretical spectra of high/low-spin models of Fe 3+ has been added to "Supplementary Obviously, the calculated high-spin L 2,3 -edge XAS spectrum could well produce the experimental data ( Supplementary Fig. 17), affording more evidence for the high-spin configuration of Fe 3+ substituting the Ni sites. Thus, the valence electronic configurations of Ni 2+ , Fe 3+ , V 4+ and V 5+ are 3d 8 (t 2g 6 e g 2 ), 3d 5 (t 2g 3 e g 2 ), 3d 1 (t 2g 1 e g 0 ) and 3d 0 (t 2g 0 e g 0 ), respectively, which are adopted in the following analysis of valence electron structures of metal ions in Ni 3 Fe, Ni 3 V, and Ni 3 Fe 0.5 V 0.5 ." Question 9b: "The authors use "charge transfer" frequently. It is confusing and imprecise. Which charge carrier? Electron or hole? It would be clearer to state which element would be reduced or oxidized." Response 9b: Thanks for the good suggestion. The expression "charge transfer" has been revised to "partial electron transfer" in the entire manuscript to avoid the ambiguity.    (2016)]. Therefore, we infer that in our experiment, the OER process is most probably catalyzed by the (101) surface of NiOOH.
The corresponding sentence at page 45 of the revised Supplementary Information has been revised to "We consider the (101)  Response 10c: Both charge transfer resistances (R ct(int) and R ct(s-l) ) are reduced (please see Supplementary Table 9). The following short paragraph has been added to the revised manuscript at page 16 to make this issue more clear in the main text: "The Nyquist plots (Fig. 5d) are fitted to a simplified Randles equivalent circuit model (Supplementary Note 3). The very small semicircles in the high frequency zone are attributed to the internal charge-transfer resistances (R ct(int) ) of electrodes, and the second semicircles represent the charge-transfer resistances (R ct(s-l) ) at the solid/liquid interface between electrode and electrolyte. Both R ct(int) and R ct(s-l) values apparently decreased as Fe and V were co-doped into the Ni(OH) 2 lattices. The total charge-transfer resistances (R ct ) measured at 300 mV overpotential are 4.2, 7.2, 10.0, and 17.2 Ω for the CFP-supported Ni 3 Fe 0.5 V 0.5 , Ni 3 V, Ni 3 Fe, and pure Ni (oxy)hydroxide catalysts, respectively (Supplementary Table 9)." Question 10d: "Supplemental Table 9. Please add potential at which EIS was performed and refer to circuit." Response 10d: The applied potential (η = 300 mV) for EIS measurement has been added to the caption of Supplementary Question 10f: "I was confused by Fig. 1, where the substrate looks like a metal mesh.
The carbon fiber substrate is not as ordered as shown there." Response 10f: The schematic diagram of the CFP substrate in Fig. 1 has been redrawn to make it look like a CFP. Besides, the figure for TOC has also been revised.
--------------------------- Given the quality of the science and its inherent broad interest and topicality, I strongly recommend that this paper be accepted for publication." Response: Thanks for the positive comments on this part of work. We will try our best to further revise the manuscript to make it suitable for publication in Nat. Commun.
---------------------------Responses to the comments of Referee #3: Interestingly, the in-situ V K-edge XANES spectra displayed that the intensity of the pre-edge peak was gradually decreased as the applied potential was increased from 1.15 to 1.75 V (inset of Fig. 6f), however, it shows identical spectral features (including shape and intensity of other characteristic peaks) to those measured before OER.
Similarly, except for the decrease in the intensity of the characteristic peaks, no other obvious change is visible at the ex-situ V L-edge spectra ( Supplementary Fig. 25 Response 3: Thanks for referee's suggestion, which helps us improve the manuscript. To the best of our knowledge, to date, there is no report on in-depth spectroscopic and theoretical studies of local coordination environment, electronic band structure and active site for the V-containing trimetal ( DFT calculations show that the V site gives near-optimal binding energies of OER intermediates and has lower overpotential compared with Ni and Fe sites in the Fe/V co-doped Ni (oxy)hydroxides. To our knowledge, this is the first in-depth spectroscopic study combined with theoretical calculation on local coordination environment and electronic structure for the V-containing bi-and trimetal (oxy)hydroxide OER catalysts in both rest and activated states, and the first time to spectroscopically demonstrate the short V−O bond distances in the activated V-containing trimetal (oxy)hydroxide OER catalysts.
Following the referee's suggestion, we added some additional discussion on the relations of high OER activity of the Fe/V co-doped Ni (oxy)hydroxide with its local structural change and electronic band structure in the revised manuscript.
At pages 17−18: "Meanwhile, the V−O1 distance is also shrunk from 1.  Fe atoms (Fig. 7c)." Question 4: "in addition to those above main concerns the claim of a uniform substitution of Fe/V into Ni(OH) 2 lattice to substitute Ni sites cannot be supported by