Selective electroreduction of carbon dioxide to methanol on copper selenide nanocatalysts

Production of methanol from electrochemical reduction of carbon dioxide is very attractive. However, achieving high Faradaic efficiency with high current density using facile prepared catalysts remains to be a challenge. Herein we report that copper selenide nanocatalysts have outstanding performance for electrochemical reduction of carbon dioxide to methanol, and the current density can be as high as 41.5 mA cm−2 with a Faradaic efficiency of 77.6% at a low overpotential of 285 mV. The copper and selenium in the catalysts cooperate very well for the formation of methanol. The current density is higher than those reported up to date with very high Faradaic efficiency for producing methanol. As far as we know, this is the first work for electrochemical reduction of carbon dioxide using copper selenide as the catalyst.

In this manuscript, the authors reported the synthesis and characterization of copper selenide nanocatalysts for electrochemical reduction of CO2 into methanol. They claim that the cooperative effect of Cu and Se accelerates the reaction by electrochemical study, including linear sweep voltammetry and electrochemical impedance spectroscopy. The presented copper selenide catalyst exhibited a high faradaic efficiency of over 76.2% at a low overpotential of 285 mV. The results are interesting in terms of synthesis and applications of copper selenide towards CO2 reduction. Unfortunately, this manuscript lacks some key structural characterizations, and the theoretical investigation is not touched to clarify the mechanism. There are still some critical problems that should be clarified before this manuscript can be considered for publication. Some questions and suggestions are listed below.
1. The authors synthesized the nonstoichiometric Cu2-xSe-y nanocatalysts and obtained a series of samples with different atomic ratio by controlling the value of VDETA/VH2O. However, the structure of Cu2-xSe-y is not clear. What is the influence of the ratio of VDETA/VH2O on the Cu2-xSe-y? They should provide sufficient critical structural characterizations (e.g., XPS, HRTEM, XAFS etc.) to confirm and discuss the relationship between the electrochemical performance and structure.
1) The need to operate in BmimPF6/ACN/H2O electrolyte compromises the utility of the catalyst. The overpotential at the cathode may be low, but it is not clear how the electrolysis could be performed with a low cell potential. The cell potential is determined by the electrode overpotentials and the voltage required for ion transport (iR). The anode is operating in acid, which will require >+1.2 V vs Ag/AgCl. Combined with the -2 V at the cathode, and an unspecified but likely large iR, the overall cell voltage will be ~3.5 V, which makes the process very inefficient. In addition, separating methanol from the electrolyte would be very challenging without a large energy input.
Optimizing catalysis under conditions that are not amenable to practical electrosynthesis is perhaps fundamentally interesting but, in my opinion, not suitable for this journal.
2) The comparisons in Fig.3a and S14 do not take into account surface area. It is very difficult to assess differences in activity across materials when the electrode morphologies/surface areas are significantly different.
3) The LSVs in Fig. 3b are used to probe "surface adsorption" of sulfate. It is not clear how to interpret the data. What does the current correspond to? These are not adsorption features and represent either oxidation of the electrode itself or water. In any case, the adsorption of sulfate under strong anodic potential has nothing to do with the adsorption of CO2 reduction intermediates.
4) The authors propose hydrogen bonding between CO2 and the Bmim+ electrolyte. CO2 is a terrible hydrogen bond acceptor and bmim is a rather weak donor. What is the evidence for this interaction?

Responses to the comments and revisions made
Referee 1

Remarks to the Author
Yang et al produce a very interesting report on the use of copper selenide nanoparticles as electrode materials for the electrochemical conversion of carbon dioxide to methanol in aqueous ionic liquid solution. While the use of nanostructured transition metal chalcogenides in electrochemical CO 2 reduction has precedent (see refs 26 and 27, for example) the findings presented here are impactful as the use of copper selenides appears to bias product formation towards the 6e -, 6H + reduction product methanol, of interest as a liquid energy carrier and fuel precursor. In comparison, previously reported metal selenide electrocatalysts showed high selectivity for carbon monoxide (ref 26). The catalysts are well characterized, and the products are supported by a variety of chemical analysis techniques (NMR, GC-MS).
In general, this is novel, exciting work and strong evidence is provided for its conclusions. I believe it will very important to researchers in the field moving forward, as engineering of binary Cu/Se particles may now be fine tuned to further improve catalyst performance to the point where it may be practical for fuel-forming purposes, and therefore interesting to researchers in related disciplines as well.
To help influence thinking in the field, the manuscript would be greatly strengthened, in my opinion, with some careful consideration of the potential mechanism and role of Se. Better referencing of work on metallic copper, known to mediate the reduction of CO 2 to gaseous hydrocarbons, in particular ethylene and methane (Hori, Chem. Lett. 1985, 11, 1695, and surface modified Cu (see for example, ACS Cent. Sci. 2017, 3, 853−859), would help bolster the impact of this work. Pending this revision, I would recommend acceptance for publication.

Response:
We thank the referee for the comment. On the basis of the comment, we have made the following modifications.
1) To explain the role of Se in the catalysts, the experimental results for electroreduction CO 2 to methanol over various catalysts, including Cu, CuO, Cu 2 O, CuS, Cu 2 S, CuSe, Cu 2 Se and Cu 1.63 Se(1/3), have been provided in Fig. 3a. It was found that Cu x Se catalysts had better performance for CO 2 reduction to methanol. We can deduce that Se in the catalysts is crucial for efficient CO 2 reduction to methanol.
In the revised manuscript, we have discussed this by "Thus, on the basis of above results, we can deduce that Se in the catalysts is crucial for efficient CO 2 reduction to methanol.". (Please see Page 10 of the revised manuscript).
2) To explain the role of Se in the catalyst, the EXAFS experiments were also conducted to characterize the catalysts. We have discussed this by adding "We also carried out extended X-ray absorption fine structure spectroscopy (EXAFS) experiments to study Cu K-edge, which can disclose the local atomic arrangements of the catalysts. The Cu K-edge k 2 χ(k) oscillation curve for Cu 3) The suggested references (Chem. Lett. 1985, 11, 1695and ACS Cent. Sci. 2017 have also been discussed in the revised manuscript by "Metal and metal-based catalysts have been used for CO 2 electroreduction to CO, hydrocarbons and alcohols 15,16 .". Please see page 2 in the revised manuscript.
Other suggested revisions/corrections: 1. Line 23-"store" Response: We thank the referee for the comment. The "store" has been changed to "transform" in the revised manuscript.

2.
Line 40-what is meant by "reasonable current density"? Can you give a range and a justification? This is an important point to make, given the justification of the importance of the findings presented here being the formation of MeOH with both high J and high Faradaic efficiency. On this note, a justification for the range given is also important, as to support the assertion that this set up may be useful for fuel forming purposes, it might be relevant to discuss what makes for a practically useful current density. I take similar issue with the idea of CuSe materials being described as "low-cost" (line 46).

Response:
We thank the referee for the comment. We agree we the referee that the meaning of "reasonable current density" is not clear. We agree with the referee that "low-cost" is confusing without comparing with other materials. Thus, according to the comment, we have also changed "They are also low-cost materials that have structure stability and composition-dependent optical/electrical properties 25,33 ." into "They also have structure stability and composition-dependent optical/electrical properties 33 . In addition, they are also low-cost materials comparing with many other materials 25 , especially noble metals." (Please see Page 3 in the revised manuscript) 3. Line 51-I find the nomenclature"Cu 2-x Se-y… where y represents the V DETA /V H2O " confusing, especially the use of a dash to represent both a minus sign (2-x) and a hyphen (-y) in the same breath. Is it possible to use an alternative nomenclature or to leave the solvent ratio out?

Response:
We thank the referee for the comment. As suggested by the reviewer, in the revised manuscript, we have changed the nomenclature "Cu 2-x Se-y" to "Cu 2-x Se(y)".

Line 90-"formulae"
Response: We thank the referee very much for the comment. According the suggestion of referee, "formulas" has been modified to "formulae". (Please see Page 5 in the revised manuscript) 5. Line 104-"AcN"-acetonitrile? "catholyte"-electrolyte? In general, I'm not sure this is the appropriate nomenclature. Please check. See also line 236.

Response:
We thank the referee for the comment. In this manuscript, "AcN" stands for acetonitrile. To avoid confusion, we have changed "AcN" into "CH 3 CN" and "catholyte" into "electrolyte".
6. Line 106-with respect to CO 2 binding (later sulfate is tested as well, Figure 3b), this feels less important perhaps than the binding affinity of further products, as multi-electron, multi-proton steps are required to be orchestrated by the CuSe electrodes to manage to form methanol. Can the authors comment on CO binding affinity? See, for example, discussion in J. Am. Chem. Soc., 2014, 136 (40), pp 14107-14113. See also Figure 4, which suggests a mechanism that relies on strong binding of intermediates (such as CO) to the electrode surface to explain the selectivity for MeOH.

Response:
We thank the referee for the comment. We agree with the referee that the binding affinity of CO is important for producing further products. Our control experiments indicate that CO clearly promoted the formation of methanol and thus it is a possible intermediate in the reaction. This is also consistent with result of our density functional theory (DFT). Previous reports also indicate that product yield and selectivity of the CO 2 reduction reaction depend on binding energy of CO and catalysts. Catalyst that binds CO strongly produces few CO 2 reduction reaction products, whilst that with low binding energy produces mostly CO. Therefore, the moderate CO binding affinity on Cu selenide is crucial to enhancing the selectivity of known. Therefore, we used the extrapolation method, which is also commonly used 43-45 to obtain the thermodynamic potential of this reaction. In this method, the thermodynamic potential is obtained by extrapolation of partial current density vs potential curve to zero partial current density. As suggested by the referee, we have given a more detailed description on the overpotential. We have emphasized this by "The equilibrium (thermodynamic) potential for CH 3 OH was -1.815 V vs Ag/Ag + , which was obtained by extrapolation of partial current density vs potential curve to zero partial current density ( Supplementary Fig. 11) [43][44][45] ." Please see page 7 in the revised manuscript.
At the same time, in the Supplementary Information, the detailed method has been discussed by "Overpotential (η) is the difference between the equilibrium potential and the actual potential for the transformation of the substrate CO 2 into the product methanol: η=E-E 0 CO2→methanol . Here, the E 0 CO2→methanol referred to the equilibrium potential for CO 2 transformation to CH 3 OH, which can be obtained by extrapolation method [43][44][45] . Taking the Cu 1.63 Se(1/3) electrode as example, stepped potential electrolysis experiments between -1.8 V and -2.1 V were carried out and the electrolysis products were collected and characterized. The current densities for CH 3 OH at each potential are shown in Supplementary Fig. 11, and the potential at j CH3OH =0 by extrapolation method is the equilibrium potential. Therefore, the overpotential can be obtained. The method to calculate the overpotential over other electrodes was similar.". (Please see page 2 in Supplementary Information).

Line 118-"replace"
Response: We thank the referee for the comment. The wrong spelling has been corrected.
9. Line 124-can the authors comment on surface-bound stabilizing ligands on their nanoparticles? Does the amine persist on the surface of the NPs? Do they ripen over time?
Response: We thank the referee for the comment. According to the XPS results before and after electrolysis ( Supplementary Fig. 15), it can be known that no amine persists in the catalysts. After reading the comment, we also carried out the TG analysis to further verify this. TG curve is shown in Supplementary Fig. 3. 10. Line 132-I'm not sure I agree with the assertion that the Tafel slope reveals a fast 1epre-equilibrium step to form the CO 2 radical anion. The Tafel slope, for example, of the Sn/graphene catalysts mentioned in reference 40 (cited here) is just over half as large. Is the comparison here between various CuSe nanoparticle materials prepared here? Please clarify language if so.

Response:
We thank the referee for the comment. As suggested by the referee, we have modified the explanation with clarified language by "The Tafel slope of Cu 1.63 Se(1/3) was smaller than other Cu 2-x Se(y) catalysts, which leads to faster increment of CO 2 reduction rate with increasing overpotential 26,47 ." Please see page 7 in the revised manuscript.

Line 142-"stabilizing"
Response: We thank the referee for the comment. The wrong spelling has been corrected.
12. Line 164-sulfate seems like an odd choice. Firstly, it is dianionic, with a very different geometry to reduced CO 2 radical anion. Can the authors elaborate more on their choice of proxy adsorbent?

Response:
We thank the referee for the comment. We agree with the referee that the property of sulfate is different from that of reduced CO 2 radical anion and the adsorption of sulfate is not equal to the adsorption of reduced CO 2 radical anion quantitatively. However, sulfate is often used as the analogs of the CO 2 ion to qualitatively compare the variation of binding energy for CO 2 intermediates on various materials 21, 49, 51 because it is very difficult to find better analogs considering the various factors, such as chemical stability at the electrochemical reaction condition. Therefore, we used sulfate in this work, and the results support our conclusion. Besides sulfate, we also tried to use both formate and acetate, but they are not stable at the experimental condition. However, this is only one of experimental evidences to explain why Cu 1.63 Se(1/3) showed better performance. If the referee insists on removing this, we would like to do this because the conclusion of the paper is the same without these results. In revised manuscript, we have discussed this as following.
"It is known that smaller overpotential of hydroxyls and sulfate adsorption indicates larger binding energy of intermediates 51 . Therefore, hydroxyls and sulfates (or bisulfates) can be used as the analogs of the CO 2 ion to study the binding energy of intermediates. In this work, we studied the binding affinity for CO 2 − intermediates on various electrodes using sulfate, and the method is similar to that in the previous reports 21, 49, 51 . Fig. 3b shows LSVs of the sulfate adsorption peaks on different catalysts in 0.1 M sulfuric acid. It can be known that the sulfate binds more strongly on Cu 1.63 Se(1/3)." Please see page 9 in the revised manuscript.

Line 171-"proposed"
Response: We thank the referee for the comment. The wrong spelling has been corrected.
14. Line 177-Reference 43 does not support this assertion " … ionic liquid containing electrolyte and CO 2 formed a complex-[Bmim-CO 2 ] + " . Both modeling and spectroscopic studies (cited in reference 43) describe that with respect to solutions of CO 2 imidazolium ionic liquids, it is the anion that dominates the interactions with the CO 2 , with the cation playing a secondary role. In fact, this is at odds slightly with reference 2, which employs a Ag cathode and sees CO as the dominant product. We also agree with the referee that anion in the ionic liquid also plays an important role for interaction between CO 2 and ionic liquids 52 , which in turn influences the electrochemical reaction. Ionic liquids can not only enhance local CO 2 concentration by complexation, but also influence the catalytic activity of CO 2 reduction. The statement in the original manuscript has been modified as "The electrolyte containing ionic liquids and CO 2 formed complex CO 2 -[Bmim]PF 6 , which can enhance the CO 2 concentration in electrolyte and transport CO 2 to the catalyst surface to improve further reduction of CO 2 2 ." Please see page 11 in the revised manuscript.

Line 177-"electron"
Response: We thank the referee very much for the comment. The wrong spelling has been corrected.
16. Line 187-This seems a missed opportunity to discuss in more depth the work in references 26 and 27, which indeed demonstrate that "other transition metal selenides"can be efficient electrocatalysts for CO 2 reduction, albeit with selectivity for CO rather than methanol.

Response:
We thank the referee for the comment. Some discussions have been added in the revised manuscript by "WSe 2 nanoflakes was reported as an efficient catalyst for CO 2 electroreduction to CO 29 . Density functional theory (DFT) calculation indicated that molybdenum sulfides and selenides were also possible catalysts for CO 2 electroreduction 30 , which showed that the intermediates COOH and CHO were more easily adsorbed on the S and Se atoms at the edges than the intermediate CO.
Therefore, transition-metal selenides may be a class of promising catalysts for CO 2 electroreduction." Please see page 3 in the revised manuscript.
Response: We thank the referee very much for the comment. The meaning of "AcN-d 6 "was CD 3 CN, and"AcN-d 6 "has been changed to "CD 3 CN" in the revised manuscript.

Line 290-incorrect journal abbreviation in reference
Response: We thank the referee very much for the comment. The journal abbreviation has been corrected.

Remarks to the Author
In this manuscript, the authors reported the synthesis and characterization of copper selenide nanocatalysts for electrochemical reduction of CO 2 into methanol.  Please see page 10 in the revised manuscript.
We have also given the EXAFS experimental details in the revised manuscript by "The homogeneously mixed samples (20 mg) and graphite (100 mg to more reduced products that require more than a two-electron reduction 20, 53 It can also be seen that the step of *CO reduction to *CHO was an endothermic and rate-limiting process. Compared with Cu 2 Se and CuSe, the free energy of *CHO over Cu 3. In Figure 4, the authors propose the reaction mechanism for electrocatalytic CO 2 reduction to methanol on Cu 2-x Se-y electrode. However, the experimental data and results are not enough to support the author's statement. Additional experimental measurements are needed to further determine the reaction pathway, such as IR spectroscopy study during the reaction to observe the CO 2 • − intermediates directly.

Response:
We thank the referee for the comment. We agree with the referee that in-situ IR spectroscopy is a very useful technique to study the intermediate during a reaction. However, we cannot carry out the experiment because the composite electrode which include catalyst and carbon black, has strong adsorption of IR light.
We also found researches in the literatures are concentrated on using pure metal  Table 7). From the production rates of methanol, it can be seen that CO and formaldehyde clearly promoted the formation of methanol, and thus they are reasonable intermediates in the formation of methanol. We have discussed this by "To understand the reaction pathway for the formation of methanol, some control experiments were conducted in the presence of the possible reaction intermediates, such as formic acid, CO and formaldehyde (Supplementary Table 7). From the production rates of methanol, it can be seen that CO and formaldehyde clearly promoted the formation of methanol and thus they are possible intermediates in the formation of methanol.". Please see page 11 in the revised manuscript.
3) To get insight into the CO 2 reduction reaction at the microscopic level, we carried out density functional theory (DFT) to study the mechanism. Theoretical calculation results were demonstrated in Fig. 4b and The Cu 1.63 Se(1/3) catalyst also has a moderate binding energy for *CO among the three catalysts, which is beneficial for CO 2 transformation to more reduced products that require more than a two-electron reduction 20, 53 It can also be seen that the step of *CO reduction to *CHO was an endothermic and rate-limiting process. Compared with Cu 2 Se and CuSe, the free energy of *CHO over Cu 1.63 Se(1/3) catalyst is more negative, which may be mainly originated from the moderately strong binding energy for *CO intermediate. In addition, the C-Cu bond ( Supplementary Fig. 33) between Cu 1.63 Se(1/3)-CHO is 1.926 Å, which is shorter than those of Cu 2 Se-CHO (2.188 Å) and CuSe-CHO (2.002 Å), indicating that *CHO is easier to adsorb on the surface of the catalyst to accept electrons and protons to form *OCH 2 and *OCH 3 , and then is reduced to methanol. These results illustrate that the structure distortion of Cu 1.63 Se(1/3) was beneficial for CO 2 electroreduction to methanol." Please see page 11 in the revised manuscript.
4. The author should check the typos and grammer more carefully. For example: (1) Line 118 on page 7, the text "relace" should be "replace".

Response:
We thank the referee for the comment. The wrong typos and grammar have been modified in the revised manuscript.

Remarks to the Author
The manuscript by Han and co-workers describes Cu selenide electrodes for CO 2 reduction in electrolytes composed of ionic liquid, acetonitrile, and water. The work appears technically sound, but the significance of the results is not high enough to warrant publication in Nat. Commun. In addition, some of the conclusions are not fully supported by the data. In particular: 1. The need to operate in BmimPF 6 /AcN/H 2 O electrolyte compromises the utility of the catalyst. The overpotential at the cathode may be low, but it is not clear how the electrolysis could be performed with a low cell potential. The cell potential is determined by the electrode overpotentials and the voltage required for ion transport (iR). The anode is operating in acid, which will require >+1.2 V vs Ag/AgCl.
Combined with the-2 V at the cathode, and an unspecified but likely large iR, the overall cell voltage will be ~3.5 V, which makes the process very inefficient. In addition, separating methanol from the electrolyte would be very challenging without a large energy input. Optimizing catalysis under conditions that are not amenable to practical electrosynthesis is perhaps fundamentally interesting but, in my opinion, not suitable for this journal.

Response:
We thank the referee for the comment. The two questions are discussed separately below.
1) In this study, we carried the experiments using commonly used H-type electrolysis cell. We agree with the referee that the cell potential is determined by the electrode overpotentials and the voltage required for ion transport (iR) in an electrochemical reaction system, and we cannot change the theoretical voltage for the reaction. So researches in the literature are concentrated on increasing current density and Faradaic efficiency for CO 2 reduction on the cathode, and reducing the overpotentials. Here, we report the first work for electrochemical reduction of CO 2 using copper selenide as the catalyst. It was discovered that the Cu 2) We also agree with the referee that separation of a polar product is a very important factor for practical application. In this work, we mainly report our discovery that Cu 1.63 Se(1/3) nanocatalysts is an outstanding electrocatalyst as discussed above, and the separation of methanol is not studied. Here we would like to discussion the separation briefly in response to the comment. We used to ~98% for practical application. Here, we would like to emphasize that, while this paper focuses on reporting the interesting results for the catalytic reaction, the separation of the product is crucial for practical application, as mentioned by the referee, which should be studied and optimized systemically in case the electrochemical method to reduce CO 2 is used in industry in future. Fig. 3a and S14 do not take into account surface area. It is very difficult to assess differences in activity across materials when the electrode morphologies/surface areas are significantly different.

Response:
We thank the referee for the comment. As suggested by the referee, to 3. The LSVs in Fig. 3b are used to probe "surface adsorption" of sulfate. It is not clear how to interpret the data. What does the current correspond to? These are not adsorption features and represent either oxidation of the electrode itself or water. In any case, the adsorption of sulfate under strong anodic potential has nothing to do with the adsorption of CO 2 reduction intermediates.

Response:
We thank the referee for the comment. We agree with the referee that the property of sulfate is different from that of reduced CO 2 radical anion and the adsorption of sulfate is not equal to the adsorption of reduced CO 2 radical anion quantitatively. However, sulfate is often used as the analogs of the CO 2 ion to qualitatively compare the variation of binding energy for CO 2 intermediates on various materials 21, 49, 51 because it is very difficult to find better analogs considering the various factors, such as chemical stability at the electrochemical reaction condition. Therefore, we used sulfate in this work, and the results support our conclusion. Besides sulfate, we also tried to use both formate and acetate, but they are not stable at the experimental condition. However, this is only one of experimental evidences to explain why Cu 1.63 Se(1/3) showed excellent performance. If the referee insists on removing this, we would like to do this because the conclusion of the paper is the same without these results. In revised manuscript, we have discussed this as following.
"It is known that smaller overpotential of hydroxyls and sulfate adsorption indicates larger binding energy on intermediates 51 . Therefore, hydroxyls and sulfates (or bisulfates) can be used as the analogs of the CO 2 ion to study the binding energy of intermediates. In this work, we studied the binding affinity for CO 2 − intermediates on various electrodes using sulfate, and the method is similar to that in the previous reports 21, 49, 51 . Fig. 3b Commun. 20, 2047Commun. 20, -2048Commun. 20, , (2000.). In the revised manuscript, we removed "hydrogen bonding between CO 2 and the Bmim + ", and discussed the interaction of CO 2 and IL [Bmim]PF 6 by "The electrolyte containing ionic liquids and CO 2 formed complex CO 2 -[Bmim]PF 6 , which can enhance the CO 2 concentration in electrolyte and transport CO 2 to the catalyst surface to improve further reduction of CO 2 2 ." Please see page 11 in the revised manuscript.

Reviewers' comments:
Reviewer #1 (Remarks to the Author): Yang et al have provided a very interesting report on the use of CuSe nanoparticles as electrode materials for the selective electrochemical conversion of CO2 to methanol in ternary ionic liquid mixtures. Cu is a very good electrode material for CO2 reduction, though without much selectivity. This report therefore is very interesting in terms of the apparent ability of non-stoichiometric CuSe nanoparticles to orchestrate a complex series of 6e-, 6H+ steps. The materials are wellcharacterized, and product identify supported by a number of techniques, including IR (added in revised version).
Having considered the revised manuscript in light of the authors' additions and reviewers comments, I still believe this is novel, exciting work, and warrants communication to Nat. Comm.'s broad readership as this fundamental study suggests opportunities now for engineering to take over to potentially lead to practical systems.
In terms of the revisions, the authors' have largely addressed my concerns from their initial submission. I would recommend acceptance pending addressing the following minor revisions. Also, perhaps as a result of the revision, the manuscript feels repetitive in parts and might benefit from careful editing.
Line 212-213 and Fig4a: the adsorption of CO2 to the surface followed by reduction is shown as a concerted process in Fig4a, but described with some certainty as a step-wise process in lines 212-213. Please clarify.
Line 225: "rate-limiting". Figure 4b does not appear to show barriers. Were these (transition states) calculated? Is something known about the relative barriers from DFT? Please discuss if so.

Reviewer #3 Comments and Evaluation of Response
Demonstrating the significance of an electrochemical process whose merits rest on reporting high current density and Faradaic efficiency (less so overpotential, really) requires careful attention to the practical implementation of a system. The work does appear technically sound, and the answer provided in the response to the reviewer comments addresses issues with respect to practical implementation (specifically, the importance of maximizing FE and current density, and the impossibility of really addressing overall cell potential).
To properly address the fact that a reader may leave with a first impression that this is not a practical system for fuel formation, these arguments given in response to the reviewer comment should be evident to some extent in the text. That is, can you explain why, given the opportunity to address cell potential, you focus on maximizing FE and current density by using CuSe (novel approach) in mixed organic/aqueous media (not a novel approach) despite the organic solvents likely large contribution to iR drop? Can you explicitly describe for readers who might suspect low process efficiency a rationale for why given the likelihood of high iR it's still worth it to use your system? Same comment with respect to separation of products. Comment #2 appears to be well-addressed (surface area). With respect to comment #3, while a justification for the use of sulfate is given, the authors have not responded to the question of what the current responds to (why adsorption and not oxidation?) Without addressing this properly, the example should likely be removed. Comment #4 is addressed satisfactorily.
Reviewer #2 (Remarks to the Author): Through carefully reading all the comments of three reviewers, the replies to all reviewers, and the revised manuscript and supplementary information, I noted that the authors have made major revisions and conducted additional experimental measurements and DFT calculations to meet the suggestions and criticisms raised by the three reviewers. After adding more structural characterizations, such as HRETM, XPS, XAFS, TG measurements, the structure of the assynthesized Cu1.63Se(1/3) electrocatalyst is more clear. To investigate the reaction mechanism for electrocatalytic CO2 reduction, FTIR measurements and other control experiments in the presence of formic acid, CO, and formaldehyde were conducted. Necessary DFT calculations were also carried out, yielding the free energy diagram for various intermediates. The authors have addressed most of the questions raised by Reviewer #3. The first concern of Reviewer #3 is about the significance of the result. In my opinion, this work reports the electrochemical reduction of CO2 to methanol using copper selenide as the catalyst is of broad interest to researchers working on fundamental issues in materials, energy, and catalysis, although the industry application of this reaction is still a little far. In order to respond to Reviewer #3's concern on the electrode surface areas, the authors have also evaluated the intrinsic activity of various electrodes through comparing the current densities of various electrodes vs. electrochemical active surface areas (ECSA). But there are still some unanswered or not-well-addressed questions. The first is on the efficiency of the cell with an overall voltage of ~3.5 eV (Question 1 by Reviewer #3). The authors mentioned many times that the Cu1.63Se(1/3) is the most efficient catalyst, at least a brief comment on the cell efficiency should be made. Reviewer #3's another concern is on why to use the adsorption of sulfate to analogize the adsorption of CO2 radical anion (Question 3 of Reviewer #3). I agree with Reviewer #3 that surface adsorption of sulfate does not directly correlate with the adsorption features of the electrode itself or water. Hence, it is better to remove this result from this work, and try their best to give a reasonable experimental explanation. In addition, a careful inspection of the EXAFS results reveals that the fitting results are quite questionable. The FT curves in Fig. S25(c) indicate the same Cu-Se peak positions for the three sample, in contrast to the large difference of the extracted Cu-Se bond lengths in Table 6. Besides, the extracted coordination numbers for the three samples are all around 4.0, although their FT peak intensity differs significantly. This coordination number is inconsistent with the structure models used for DFT calculations. As the free energies of the reaction intermediates rely on the adsorption configuration, this inconsistence may affect the correctness of the mechanistic interpretation.
To summarize, if these above mentioned issues could be satisfactorily clarified, I consider this work is worthy of publication in Nature Communications.
changed "nuclear magnetic resonance" into "nuclear magnetic resonance spectroscopy". (Please see Page 7 in the revised manuscript).
6. Line 170: This sentence should include a reference. Reference 43?

Response:
We thank the referee for the comment. To better illustrate this sentence, we have added the reference as following.
"which resulted partially from the difference of the interaction between CO 2 and the anions of the ILs 55 ." (Please see Page 9 in the revised manuscript). Fig 4a: the adsorption of CO 2 to the surface followed by reduction is shown as a concerted process in Fig 4a, but described with some certainty as a step-wise process in lines 212-213. Please clarify.

Response:
We thank the referee for the comment. According to the advice of referee, the adsorption of CO 2 to the surface followed by reduction is a concerted process. To better explain the phenomenon, we have modified the statement as following.
"In the initial stage of the reduction, the electrolyte containing ionic liquids and CO 2 formed complex CO 2 -[Bmim]PF 6 , which can enhance the concentration of CO 2 in electrolyte and transport of CO 2 to the catalyst surface to improve further transform of CO 2 into CO 2 •− 2, 55 ." (Please see Page 11 in the revised manuscript). In addition, based on the Brønsted-Evans-Polanyi (BEP) relationship (Ref, Trans. Faraday Soc. 1938, 34, 11;J. Phys. Chem. C, 2008, 112, 1308, the reaction barrier has a linear relationship to the reaction energy, and this method does not affect the results of DFT calculation considerably. In the revised manuscript, we have modified the related sentences as following. "Based on the Brønsted-Evans-Polanyi (BEP) relationship 60, 61 , the reaction barrier has a linear relationship to the reaction energy, and it can also be seen that the step of *CO reduction to *CHO was an endothermic and rate-limiting step since the highest energy potential (0.56 eV) is needed in this step." (Please see Page 12 in the revised manuscript). The authors have addressed most of the questions raised by Reviewer #3. The first concern of Reviewer #3 is about the significance of the result. In my opinion, this work reports the electrochemical reduction of CO 2 to methanol using copper selenide as the catalyst is of broad interest to researchers working on fundamental issues in materials, energy, and catalysis, although the industry application of this reaction is still a little far. In order to respond to Reviewer #3's concern on the electrode surface areas, the authors have also evaluated the intrinsic activity of various electrodes through comparing the current densities of various electrodes vs. electrochemical active surface areas (ECSA).

Response:
We thank the referee for the comment. On the basis of the comment, we have made point-to-point modifications as following.
1. But there are still some unanswered or not-well-addressed questions. The first is on the efficiency of the cell with an overall voltage of ~3.5 eV (Question 1 by Reviewer #3). The authors mentioned many times that the Cu 1.63 Se(1/3) is the most efficient catalyst, at least a brief comment on the cell efficiency should be made.

Response:
We thank the referee for the comment. According to the advice of referee, we have added some statements about cell voltage and cell efficiency, and a brief comment has been made in the revised manuscript.
First, we calculated the cell voltage and cell efficiency using the corresponding reported method (Ref, Nat. Catal., 2018，1, 11;Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 5526;Angew. Chem. Int. Ed., 2018, 57, 6883.). Second, we have given a systematic summary about cell voltage of CO 2 reduction to different products reported in the literature (Supplementary Table 3). Third, a brief comment on the above results has been made. The overall cell voltage of our system at optimized reaction condition calculated using the reported method (Ref, Nat. Catal., 2018， 1, 11;Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 5526.) is 2.67 V, which is in the range of cell voltage values Supplementary Table 3). The highest energy efficiency was 61.7% was obtained at cell voltage of 2.67 V. In the revised manuscript, as suggested by the referee, we have discussed this as following.
"The cell voltage is an important factor for practical application, which depends mainly on the performances of the electrocatalysts. In this study, we calculated the cell voltage using the reported method 46, 47 , and the cell voltage of our system was 2.67 V, which is in the range of reported values Supplementary Table 3).
In Supplementary Fig. 12, the FE for methanol production increased with the cell voltage to reach the maximum value of 77.6% at 2.67 V. We also calculated the energy efficiency (EE) for methanol production at different cell voltages using the reported method 48 , and the results are given in Supplementary Fig. 12 Please see Page 7 in the revised manuscript.
The above results are given in Supplementary Fig. 12, and the methods to calculate cell voltage and cell efficiency are also shown Page 2 in the revised Supplementary Information, together with the corresponding references.
2. Reviewer #3's another concern is on why to use the adsorption of sulfate to analogize the adsorption of CO 2 radical anion (Question 3 of Reviewer #3 were obvious lower ( Fig. 3 and Supplementary Fig. 24 To summarize, if these above mentioned issues could be satisfactorily clarified, I consider this work is worthy of publication in Nature Communications.

Response:
We thank very much the referee for the comment.

Referee 3
Remarks to the Author 1. Demonstrating the significance of an electrochemical process whose merits rest on reporting high current density and Faradaic efficiency (less so overpotential, really) requires careful attention to the practical implementation of a system.
The work does appear technically sound, and the answer provided in the response to the reviewer comments addresses issues with respect to practical implementation (specifically, the importance of maximizing FE and current density, and the impossibility of really addressing overall cell potential).
To properly address the fact that a reader may leave with a first impression that this is not a practical system for fuel formation, these arguments given in response to the reviewer comment should be evident to some extent in the text. That is, can you explain why, given the opportunity to address cell potential, you focus on maximizing FE and current density by using CuSe (novel approach) in mixed organic/aqueous media (not a novel approach) despite the organic solvents likely large contribution to iR drop? Can you explicitly describe for readers who might suspect low process efficiency a rationale for why given the likelihood of high iR it's still worth it to use your system? Same comment with respect to separation of products.

Response:
We thank the reviewer for the very instructive comment. We have addressed all the concerns, as can be known from the answers to the comments as discussed in the following.

1) About practical application
According to the comment of the referee, in the Conclusion (Summary) section, we have added "Despite the catalytic system is far from industrial production, it is still very interesting that Cu 1.63 Se(1/3) nanocatalysts can yield highest current density up to date at very high Faradaic efficiency.". Please see Page 12 in the revised manuscript.
2) About the cell potential.
As mentioned by the referee, cell voltage is an important factor for practical application. In the revised manuscript, we calculated cell voltage of our system using the reported method (Ref, Nat. Catal., 2018, 1, 11;Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 5526.). The cell voltage of our system at optimized reaction condition is 2.67 V. We also give a table (Supplementary Table 3) to list the cell voltage values.
We found in the literature for CO 2 reduction to different products (most papers on CO 2 reduction do not report cell voltage, and they pay attention to current density, Faradaic efficiency, and overpotential), and the values of cell voltage are in the range of 2.2-3.7 V (Supplementary Table 3). In the revised manuscript, as suggested by the referee, we have discussed this as following.
"The cell voltage is an important factor for practical application, which depends mainly on the performances of the electrocatalysts. In this study, we calculated the cell voltage using the reported method 46, 47 , and the cell voltage of our system was 2.67 V, which is in the range of reported values (2.2-3.7 V, Supplementary Table 3)." Please see Page 7 in the revised manuscript.
The above results are given in Supplementary Fig. 12, and the method to calculate cell voltage is also shown Page 2 in the revised Supplementary Information, together with the corresponding references.
3) Can you explain why, given the opportunity to address cell potential, you focus on maximizing FE and current density by using CuSe (novel approach) in mixed organic/aqueous media (not a novel approach) despite the organic solvents likely large contribution to iR drop? Can you explicitly describe for readers who might suspect low process efficiency a rationale for why given the likelihood of high iR it's still worth it to use your system?

Answer:
We agree with the referee that the organic solvents generate relatively large iR drop. Many researchers have reported the electroreduction CO 2 in aqueous electrolyte as catholyte. Although water is cheaper, easily available, and greener, the evolution of hydrogen in aqueous electrolyte is much easier than reduction of CO 2 , and thus the Faradaic efficiency is usually low, especially at larger current density.  57,2427). Therefore, despite organic system may contribute to iR drop, we believe that this system is suitable because multi-electron/proton coupling steps can occur in this system to obtain valuable products with high current density and high selectivity.
In our work, combination of copper selenide and ionic liquid-CH 3  In this revised manuscript, we have discussed this as following.
"We also calculated the energy efficiency (EE) for methanol production at different cell voltages and the results are given in Supplementary Fig. 12 , and they provide more opportunity to produce various valuable products 1,2,8,53,54 ." Please see Page 9 in the revised manuscript.

4) About separation of products.
We also agree with the referee that separation of a polar product is a very important factor for practical application. As discussed above, we mainly report our discovery that Cu 1.63 Se(1/3) nanocatalyst is an outstanding electrocatalyst, and the separation of methanol is not studied in this work. Here we would like to discuss this briefly in response to the comment. towers, a rectification column, condenser, and a fluid transfer tube. The separation efficiency of methanol can reach to ~98% for practical application. Here, we would like to emphasize that, while this paper focuses on reporting the interesting results for the catalytic reaction, the separation of the product is crucial for practical application as mentioned by the referee, which should be studied and optimized systemically in case the electrochemical method to reduce CO 2 is used in industry in the future. In the revised manuscript, we have discussed this as following "The separation of the reaction mixture is crucial for practical application. Although this is out of the scope of this work, we would like to discuss this very briefly. 2. Comment #2 appears to be well-addressed (surface area).

Response:
We thank the referee for the comment.
3. With respect to comment #3, while a justification for the use of sulfate is given, the authors have not responded to the question of what the current responds to (why adsorption and not oxidation?) Without addressing this properly, the example should likely be removed.

Response:
We thank the referee for the comment. According to the advice of referee, we have removed the statements about the sulfate adsorption in the revised manuscript.

Response:
We thank the referee for the comment.