Modulating adsorbed hydrogen drives electrochemical CO2-to-C2 products

Electrocatalytic CO2 reduction is a typical reaction involving two reactants (CO2 and H2O). However, the role of H2O dissociation, which provides active *H species to multiple protonation steps, is usually overlooked. Herein, we construct a dual-active sites catalyst comprising atomic Cu sites and Cu nanoparticles supported on N-doped carbon matrix. Efficient electrosynthesis of multi-carbon products is achieved with Faradaic efficiency approaching 75.4% with a partial current density of 289.2 mA cm−2 at −0.6 V. Experimental and theoretical studies reveal that Cu nanoparticles facilitate the C-C coupling step through *CHO dimerization, while the atomic Cu sites boost H2O dissociation to form *H. The generated *H migrate to Cu nanoparticles and modulate the *H coverage on Cu NPs, and thus promote *CO-to-*CHO. The dual-active sites effect of Cu single-sites and Cu nanoparticles gives rise to the catalytic performance.


INTRODUCTION
The introduction should be revised providing more information and state-of-the-art knowledge about multiactive sites for multiple reactants in CO2RR (L33). A lot of studies have been performed to study the role of particle size, distance or shape. A more through summary on the literature should be presented with a focus on the relevant parameters for this work (L49-51). The cited literature should show similar reaction systems, e.g. in L56, the authors claim H2O activation on Cu single atoms, but the literature presents their data in an acidic system, while the authors study alkaline systems. A discussion why this result like this is transferable, is thus required. A specification on the ratio of CuNP compared to Cu single atoms should be given already in L60.

TEM ANALYSIS
The Cu nanoparticle size is a relevant parameter for the catalytic performance. It would thus be relevant to mention the obtained sizes in the main text and present particle size distribution for all catalysts. The authors may also provide a NP size distribution for the R-Cu1/CuNP as they have nicely done for the other two samples. Additionally a more exact measure for the ratio between Cu NP and single atom sizes from the TEM data should be given and discussed against the EXAFS fitted data and the observed percentages between the coordination numbers. In this regard, the HAADF-STEM images showing the single atom Cu sites are missing in Figure S1 and S2 and are asked to be provided. systems. The manuscript has started well by presenting three different samples and analyzing the electrocatalytic data of each quite well. However, some more work has to be done on the spectroscopic site to understand the catalytic systems and to extract a relevant conclusion from it.

EXPERIMENTAL
In general, the experimental section lacks several details that are important to follow the experiments. E.g., different types of flow cells exist, a more explicit description should be provided. If a gas-fed CO2 type flow cell was used, please also add the side from which the gas products were determined and how often gas products were injected into the GC (if an online-GC was used). Please provide a scheme or a reference of the used cell. Please state type and company of membrane and used chemicals, type of the used gas diffusion electrode was used (type, content of PTFE, with or without microporous layer), type of the Ni foam. Please state the total volume of electrolyte (re)circulated, and how big the electrolyte volume in the cell is. The authors may also provide more details of the electrolyte (purity of the water, purity and company of the KOH, possible presaturation with CO2). Experimental details on the ohmic resistance should be added. The authors may also provide the experimental details the double layer capacitance measurements. Ni foam as counter electrode not the state of the art to be used as counter electrode for CO2RR, the authors may comment on their choice. The authors may also comment on their choice of and Hg/HgO reference electrode instead of other stable and non-toxic reference electrodes, such as RHE or Ag/AgCl. The authors may provide schemes or references to the used operando and in situ XAS, Raman and ATR-SEIRAS cells and a description the catalysts are exposed to the spectroscopic techniques. The authors may state which definition are the terms operando and in-situ chosen for each technique. The authors may also combine the Methods in the main manuscript and the methods in the Supporting information into one section.
Reviewer #3 (Remarks to the Author): Authors reported the dual-active sites catalyst for electrocatalytic CO2 reduction reaction. The atomic Cu sites and Cu NPs exhibited an efficient CO2-to-C2 reaction. M-Cu1/CuNP delivered a high C2 products FE of 75.4% with a partial current density of 289.2 mA cm-2 and remarkable long-term stability. Based on the experimental and theoretical studies, the boosted H2O dissociation and formed*CHO play the crucial role in the CO2-to-C2 reaction. However, the description of X-ray and FTIR spectroscopy is not convincing. The reviewer has several questions regarding the mechanistic insights into the proposed reaction model. Please address these issues in detail.

1.
Line 92: XPS results show that the valance state of Cu in the catalysts increased in the order of P-Cu1/CuNP< M-Cu1/CuNP < R-Cu1/CuNP. Can authors provide quantitative analysis? 2. Line 100: " Figure 1F shows that the characteristic peaks of Cu Kβ1,3 for Cu1/CuNP samples lied between 0 and +2 in the sequence of P-Cu1/CuNP < M-Cu1/CuNP < R-Cu1/CuNP". In fact, it is difficult to distinguish the position of these characteristic peaks in figure 1F. Authors should present this figure clearly and provide detailed description. 3. In Figure 3C, the increase of Cu valence state was studied using the absorption energy of white line position of XANES spectra. However, the average Cu valence state of pristine materials was examined using first derivative of XANES spectra ( Figure 1H). Please provide the detailed XANES spectra of pristine materials in the supporting information. 4. In Figure 3C, the feature of Cu K-edge XANES spectrum of M-Cu1/CuNP at OCP is very similar with that of Cu foil (Cu0). Also, the shift in the absorption energy of white line position of M-Cu1/CuNP and Cu foil is close. These results suggest that Cu valence state of M-Cu1/CuNP could be metallic Cu (or close to Cu0) at OCP, which is not consistent with the XPS/XAS results in Figure 1H and Figure S3. Authors should add XANES spectrum of pristine M-Cu1/CuNP and Cu foil in figure 3C. The detailed discussion should be provided. More information could be obtained using linear combination fitting. 5. , authors claimed that the increase of Cu valence state could be attributed to the adsorption of CO2 and H2O molecular on Cu NPs and Cu sites. It is surprising that Cu valence state increases during CO2RR. This result is not consistent with previous studies. Also, Cu K-edge XAS provides bulk information not surface information. These results and statements are not convincing. If the adsorption of CO2 and H2O does cause the oxidation process, authors should provide a serious of operando XANES spectra obtained from positive potentials to negative potentials. The shifts in the absorption energy of white line position can be therefore discussed. 6. Figure S12-S14 show the TEM/XPS/XRD results of M-Cu1/CuNP after CO2RR. Is it after 5 hours or 40 hours? Please provide the information. 7. The experimental description of in situ XAS/Raman/SEIRAS is not clear. Were all in situ XAS/Raman/SEIRAS experiments performed in the flow cell with KOH electrolyte? Since cell configuration of traditional SEIRAS measurement is H-type cell, how did authors perform SEIRAS experiment with flow cell? Please provide the cell configuration. The changes in the OH-associated peak should be further addressed.
If authors performed the SEIRAS experiment with traditional cell configuration (H-type cell), the pH value of electrolyte changes during CO2RR. Also, the pH value of electrolyte used in SEIRAS measurement is different from that in XAS/Raman experiments. How did authors correlate the FTIR results with XAS/Raman results? 8. Cu will be oxidized in the alkaline solution. Why did not XAS ( Figure 3C)/Raman ( Figure 3D) results show the formation of CuO/Cu(+2)? 9. Since the S/N ratio of CHO-associated peak is poor in figure 3F, the role of CHO in the CO2 reduction mechanism is questionable. To prove the proposed reaction model, authors should provide SEIRAS of P-Cu1/CuNP and R-Cu1/CuNP materials. The potential dependence of CHO-associated peak (at 1548 cm-1) in different materials should be further discussed.
Reviewer #4 (Remarks to the Author): In this work, the authors report the design of catalysts for electrochemical CO2 reduction to C2+ products based on a dual-active site strategy. They combine a water dissociation catalyst (single atom Cu) and a C-C coupling one (Cu nanoparticles) to produce Cu nanoparticles supported on Cu-N-C single atom matrix. They found that this combination enhances CO2 reduction to C2+, achieving a C2 Faradaic efficiency of 75% and a partial current density of 289 mA with a total FE at -0.6 V vs RHE. While the fabrication of the dual catalysts could be interesting, I do not find the designed catalysts represent a significant advance in the field of electrochemical CO2 conversion. My conclusion comes from both fundamental and applied perspectives as I explain below: 1. The electrochemical CO2 conversion performance of the designed catalysts in this work is comparable to most of normal Cu nanoparticles catalysts. In a basic electrolyte, C2 selectivity of 70-80% in the current density range of 200-500 mA/cm2 has been frequently reported using Cu nanoparticle catalysts. A good catalyst for C2 production should exhibit an FE of at least over 80%. 2. The high performance in this work was achieved with very basic electrolyte (5M KOH). Many recent analyses have pointed out the cost related to carbonate formation with alkaline electrolyte and suggest that this condition is not suitable for practical application. In this work, the author designed dual-active site catalysts that work well only in strong alkaline electrolyte ( Figure S15). Therefore, it is unclear how the designed catalysts can be used in practical systems. 3. The mechanism for enhanced C2+ production based on dual-active sites is rather speculative as there are not enough experimental and theoretical evidence to support that claim. For example, the authors claim that Cu-N-C matrix can enhance water dissociation to provide protons for C2 formation, but they did not provide a clear experimental and theoretical evidence for the enhanced water dissociation. Some obvious questions are: How does the water dissociate on Cu sites? What happens to the OH-? How does the H+ transport to the surface of Cu nanoparticles (in the form of H+ or H)? 4. The conversion of Cu single atom to Cu nanoparticle/clusters during electrochemical CO2 reduction are well-documented (Nat. Commun. 9, 415 (2018); Angewandte Chemie, 58, 15098 (2019)). The presence of Cu single site with operando XAS could be originated from non-electrically conductive sites (Nat. Commun. 13, 4190 (2022). Therefore, it is unclear how the Cu single sites can help CO2 reduction if they are not stable. 5. Additional controlled samples should be tested to confirm the proposed mechanism. For example, mixing Cu-N-C with a commercially available Cu nanoparticles or other metals such as Ag or Au to see if this mechanism is universal (i.e., for CO2 conversion to other products).

Reviewer 1:
In this work, Feng et al reported a dual-active sites catalyst comprising atomic Cu sites and Cu nanoparticles for CO2RR. Among them, the Cu nanoparticles facilitated the C-C coupling step by *CHO dimerization, and the atomic Cu sites boosted H2O dissociation to form *H and further migrated to Cu nanoparticles. As a result, the catalyst shows the CO2-to-C2+ performance with a FE of 75.4%. Some concerns remain regarding the suitability of the authors' conclusions, for which the reviewer would like to see additional data to address and then can justify whether this work is suitable for publication.

Response:
We thank the referee very much for the comment, which help and guide us to think about and polish our work greatly. We have tried our best to answer the questions from the reviewers. We hope that we have addressed all of the questions satisfactorily. Figure 3B. The CO stripping peak is determined at approximately -0.7 V vs. SHE (Nature Catalysis 2020, 3, 797-803). The peak around 0.8-0.9 V vs RHE should not be classified as CO stripping.

Response:
We thank the referee again for the comment. The potential of CO stripping peak in the as-mentioned paper (Nat. Catal. 2020, 3, 797-803) was presented versus Standard Hydrogen Electrode (SHE) and 0.1 M NaOH was used as electrolyte, while the CO stripping experiments in our work were conducted in 0.1 M Na2SO4 electrolyte and the potential of CO stripping peak was presented versus Reversible Hydrogen Electrode (RHE). Unlike the SHE, the potential measured by RHE does not change with the pH, so it is a more widely used (J. Phys. Chem. B 2004, 108, 28, 9829-9833;Proc. Natl. Acad. Sci. U.S.A. 1919, 5, 160-163). In many literature (Angew. Chem. Int. Ed. 2017, 56, 8828-8833;Nat. Commun. 2021, 12, 6261;Adv. Mater. 2023, 35, 2208799), the CO stripping peak on metalbased catalyst located at range of 0.6-0.9 V vs. RHE. Moreover, we have added the LSV curve without CO adsorption on the catalyst as Figure S22 in the revised supporting information, and no stripping peak was observed at around 0.8-0.9 V vs. RHE. Therefore, the peak around 0.8-0.9 V vs. RHE in Figure 3B could be assigned to CO stripping.
In the revised manuscript, we have added the above literature as references and discussed them as "On the other hand, the electrochemical CO stripping voltammetry tests of R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP were also performed to study the CO adsorption ability. Figure 3B shows a peak in the potential range of 0.8-0.9 V for R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. No peak was observed in the LSV curve without CO adsorption ( Figure   S22). Further according to pervious literature, 32-34 the peak at around 0.8-0.9 V can be attributed to the CO stripping peak. Interestingly, the CO stripping peak occurred around 0.89 V for M-Cu1/CuNP and P-Cu1/CuNP, while it was around 0.82 V for R-Cu1/CuNP. The positive shift suggested that M-Cu1/CuNP and P-Cu1/CuNP had stronger CO binding ability than R-Cu1/CuNP, which have higher Cu NP content. The results above indicated that the Cu NPs was beneficial for the adsorption and activation of CO2 and CO." Please see them in page 13 of the revised manuscript.
2. It is uncertain that why the author chose different reference samples in the tests. M-Cu1/CuNP was compared with P-Cu1/CuNP in KIE experiment, but compared with R-Cu1/CuNP in CO stripping experiment. Why not compare the three samples together?

Response:
We thank the referee again for the comment. According to the comment, we have conducted the KIE experiments, gas electro-response experiments and CO stripping experiments over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP, and discussed them in the revised manuscript by "The kinetic isotope effect (KIE) of H2O/D2O (H/D) experiments were performed to get insights into the role of H2O dissociation in CO2-to-C2 products ( Figure 3A). When H2O was replaced by D2O in the electrolyte, the formation rate of the product (for example C2H4) decreased over P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP, and the level of its decrease was closely related to the content ratio of Cu1/CuNP. If the KIE value (defined as the ratio of C2H4 formation rates in H2O and D2O) closes to 1, H2O dissociation is not the rate-determining step over the catalyst. The KIE value for R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP were 1.12, 1.21 and 2.87, respectively, which indicated that H2O dissociation was accelerated gradually with the increase of Cu1/CuNP content ratio. These results confirmed that the atomic Cu sites were responsible for accelerating H2O dissociation and providing proton to adjacent Cu NPs, thus affecting CO2-to-C2 products.
Meanwhile, the N-doped carbon matrix catalysts has been reported to favor the migration of proton. 30 In addition, the H2 formation rate in 5 M KOH H2O solution and D2O solution was presented in Figure S17. An obvious decrease was observed over all the catalysts, while the decreasing degree followed the sequence of P-Cu1/CuNP > M-Cu1/CuNP > R-Cu1/CuNP, suggesting that the increase of Cu single atom content can reduce the influence of isotope effect on H2 formation rate. Furthermore, we studied the influence of electrolyte pH on CO2RR performance over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. Three different concentrations of KOH aqueous electrolytes, i.e., 0.1 M, 3 M and 5 M, were employed to adjust the pH environment at the electrode/electrolyte interface. The results in Figure S18 show that the FE and formation rate of C2 products increased with the increasing pH value of electrolyte, suggesting that strong basic local environment favored the C2 products formation. Although CO2RR activity of M-Cu1/CuNP in low concentration KOH electrolytes was lower than that in 5 M KOH electrolyte, M-Cu1/CuNP exhibited higher FE and formation rate of C2 products than P-Cu1/CuNP and R-Cu1/CuNP in low concentration KOH electrolytes, which suggested that the cooperative effect of Cu NPs and atomic Cu sites still facilitated C2 products formation. The ratio of C2 products formation rate over M-Cu1/CuNP to P-Cu1/CuNP, i.e., RateM/RateP, in 5 M KOH (3.4) is higher than those in 3 M KOH (3.0) and 0.1 M KOH (2.3). It demonstrated that the role of atomic Cu sites in accelerating the H2O dissociation process was significantly more pronounced at higher pH values, leading to the enhanced C2 products formation, even if the dissociation of H2O in higher pH alkaline electrolyte was a sluggish step.
We then investigated the capacity adsorption and activation of CO2 and CO molecules on R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP via gas electro-response experiments in a self-designed gas adsorption electroresponse device ( Figure S19). 31 The results in Figure S20 and Figure S21 show that the CO2 and CO adsorption responses changed in the sequence of R-Cu1/CuNP < M-Cu1/CuNP < P-Cu1/CuNP, suggesting that the adsorption and activation of CO2 and CO were promoted as the ratio of Cu1 to CuNP decreased. On the other hand, the electrochemical CO stripping voltammetry tests of R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP were also performed to study the CO adsorption ability. Figure 3B shows a peak in the potential range of 0.8-0.9 V for R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. No peak was observed in the LSV curve without CO adsorption ( Figure S22). Furthermore, according to pervious literature, 32-34 the peak at around 0.8-0.9 V can be attributed to the CO stripping peak. Interestingly, the CO stripping peak occurred around 0.89 V for M-Cu1/CuNP and P-Cu1/CuNP, while it was around 0.82 V for R-Cu1/CuNP.
The positive shift suggested that M-Cu1/CuNP and P-Cu1/CuNP had stronger CO binding ability than R-Cu1/CuNP, which have higher Cu NP content. The results above indicated that the Cu NPs was beneficial for the adsorption and activation of CO2 and CO." Please see them in page 12-13 of the revised manuscript.
3. The assignment of C≡O Raman bands at 2000-2100 cm -1 is problematic. In fact, the CO bridge appears 1830-1900 cm -1 . (ACS Central Science 2016, 2, 522-528, The Journal of Physical Chemistry C 2017, 121, 12337-12344, ACS Catal. 2018, 8, 7507-7516, ACS Catal. 2019, 9, 6305-6319, Nat. Commun. 2022, 13, 2656 Response: We thank the referee again for the comment. We agree with the referee. The Raman band of CO bridge configuration locates at 1830-1900 cm -1 , but the peak at 2000-2100 cm -1 can be assigned to the CO atop configuration (ACS Catal. 2018, 8, 7507-7516;ACS Catal. 2019, 9, 6305-6319;Nat. Commun. 2022, 13, 2656. In the revised manuscript, we have added the above literature as references and we modified the description of in situ SERS as "In situ surface-enhanced Raman spectroscopy (SERS) was employed to reveal the interaction between Cu species and the reaction intermediates during CO2RR ( Figure S25). [35][36][37][38] As shown in the Raman spectra under different potentials ( Figures 3D-3F), the peaks assigned to the restricted rotation of adsorbed CO (298 cm -1 , ν(Cu-CO)), Cu-CO stretching (365 cm -1 , ν(Cu-CO)) and C-O stretching of atop *CO (2000-2100 cm -1 , ν(CO)) were observed from -0.3 to -0.8 V over M-Cu1/CuNP, while ν(Cu-CO) and ν(CO) peaks over P-Cu1/CuNP appeared from -0.4 V and -0.5 V, respectively. This observation could be related to the lower onset potential for C2 products for M-Cu1/CuNP. Only CO rotation peaks can be observed on R-Cu1/CuNP from -0.4 to -0.8 V, meaning poor CO2 reduction activity. In situ SERS was also conducted over Cu-N-C ( Figure S26), whereas no peaks associated with *CO intermediate were observed, suggesting that single atomic Cu was not conducive to the conversion of CO2 to adsorbed CO. The results of in situ SERS suggested that CO2-to-C2 products proceeded via CO intermediate and moderate ratio of Cu1 to CuNP facilitated CO2 conversion of adsorbed CO." Please see them in Page14 of the revised manuscript.
4. It is not clear about the details of the in situ Raman/ATR-SEIRAS process. What electrolyte did the authors use for in-situ tests? The CO2RR performance of the catalysts were evaluated in KOH aqueous solution. If the authors also used alkaline electrolytes for the in situ experiments, why did the CO3 2signal appear?
Response: We thank the referee again for the comment, and we answer the two questions as follows.
(1) 5 M KOH aqueous solution was used as electrolyte during the in situ Raman spectroscopy experiments.
Considering strong basic electrolyte would damage the germanium ATR crystal, 3 M KOH aqueous solution was used as electrolyte during the in situ ATR-SEIRAS experiments. The details of in situ surface-enhanced Raman and ATR-SEIRAS experiment have been added in the Methods section of the revised manuscript: "In situ Raman measurements. A flow cell with a quartz window by GaossUnion (Tianjin) Photoelectric Technology Company was used to carry out in-situ Raman measurements using a Horiba LabRAM HR Evolution Raman microscope ( Figure   S25). A 785 nm laser was used and signals were recorded using a 20 s integration and by averaging two scans. The gas diffusion electrode sprayed with the catalyst was used as working electrode and a graphite rod and a Hg/HgO electrode were used as counter and reference electrodes, respectively. The counter electrode was separated from the working electrode by anion exchange membrane. The 5 M KOH aqueous solution was used as electrolyte and were recirculated by pump with flow rates of 20 mL min -1 . Meanwhile, CO2 gas was continuously supplied to the gas chamber with a flow rate of 40 mL min -1 .
In situ ATR-SEIRAS measurements: The experiments were conducted in a modified electrochemical cell that integrated into a Nicolet 6700 FTIR spectrometer equipped with MCT detector cooled by liquid nitrogen ( Figure   S27). The catalysts ink was dropped on a germanium ATR crystal deposited with Au film. A platinum wire and a Hg/HgO electrode were used as counter and reference electrodes. Each spectrum was collected with 32 times with a resolution of 4 cm -1 . The CO2-saturated 3 M KOH aqueous solution was used as electrolyte to avoid that the germanium ATR crystal was damaged by strong basic electrolyte. The background spectrum was collected at the potential of +0.1 V vs RHE." Please see them in Page 21-22 of the revised manuscript.
The optical photograph and schemes of cells used in in situ surface-enhanced Raman and in situ ATR-SEIRAS have been also provided in the revised supporting information as Figure S25 and Figure S27, respectively.
CO2,aq + H2O ↔HCO3 -+ H + HCO3 -↔ CO3 2-+ H + HCO3 -+ OH -↔ CO3 2-+ H2O According to above reactions, it can be concluded that CO3 2exists as the main species in alkaline condition. The pH value of CO2-saturated 5 M KOH and 3 M KOH aqueous solution are 11.02 and 10.21, respectively, so CO3 2signal appeared in in situ Raman spectra. It was also reported in many literature (J. Am. Chem. Soc. 2022, 144, 14936-14944;Angew. Chem. Int. Ed. 2022, 61, e202110657). 5. The authors compared the catalytic performance with other literatures ( Figure 2D), but the performances of this work are not outstanding, especially in alkaline electrolytes. Moreover, this kind of comparison does not help much with the advance of the research community, as when researchers get used to them, they just target higher metrics but ignoring science and insights behind the experiments.

Response:
We thank the referee very much for the comment. We appreciate the reviewer's comment regarding the significance of revealing the science and insights behind the experiments, rather than solely pursuing high metrics.
We have conducted additional experiments to further reveal the mechanism and revised the manuscript carefully.
Additionally, it is significant to obtained high CO2RR activity at low applied potential, which means low energy consumption. The as-prepared M-Cu1/CuNP exhibited large C2+ products partial current density at applied potential, which is lower than most of the reported electrocatalysts for CO2RR to C2+ products in literature. In response to the reviewer's suggestion to focus less on the catalytic performance, we have removed the figure comparing performance and instead included Table S7 in the revised supporting information and discussed them in the revised manuscript by "Compared with reported electrocatalysts for CO2RR to C2+ products in the literature, M-Cu1/CuNP exhibited higher CO2RR to C2+ products activity at low applied potential, especially in term of C2+ partial current density (Table S7)." Please see them in Page 10 of the revised manuscript.
6. Regarding catalytic properties, the reviewer is not sure whether the authors have adopted rigorous protocol to ensure the accuracy of CO2RR performance data. The outlet flowrates could be very different from the inlet flowrates due to the reaction between CO2 and KOH, leading to problematic FE data of the gas products.

Response:
We thank the referee again for the comment. Adopting rigorous protocol is important for obtaining accurate CO2RR performance data. Our CO2RR activity testing procedure is consistent with the methods described in the literature. We took into account that the outlet flow rate is not consistent with inlet flow rate due to the reaction between CO2 and KOH. Therefore, we calculated the product FEs using the outlet flow rate to ensure accuracy.
In the revised manuscript, we have added further experiments details: "Electrocatalytic CO2 reduction. To prepared gas diffusion electrode, 1 mg catalyst was suspended in 500 μL isopropanol with 10 μL Nafion D-521 dispersion (5 wt%) to form a homogeneous ink. Then the catalyst ink was spread onto the gas diffusion electrode of 0.5 × 2 cm 2 in area by a micropipette and then dried under room temperature. All the electrochemical experiments were conducted on the electrochemical workstation (CHI 660E) equipped with a high current amplifier (CHI 680C) and the CO2RR performance was investigated in flow cell ( Figure S7). The prepared gas diffusion electrode and Ni foam were used as the cathode and anode, respectively. An anion exchange membrane was used to separate the cathode and anode compartment. Aqueous KOH solution (5 M) was used as the electrolyte solution and the electrolyte volume was 30 mL. The catholyte solution and anolyte solution were recirculated by two pumps with flow rates of 10 mL min -1 and 30 mL min -1 , respectively. Meanwhile, CO2 gas was continuously supplied to the gas chamber by using a mass flow controller with a flow rate of 40 mL min -1 . All the potentials were measured against a Hg/HgO reference electrode and converted to versus RHE with iR (80%) compensations, i represents the current obtained at corresponding potential, R is ohmic resistance of the cell measured by electrochemical workstation: The chronopotentiometry method is used for evaluating long-term CO2RR activity stability of electrocatalyst in a flow cell. In order to address the issues of flooding and carbonation accumulation of GDE, the CO2RR was interrupted every 5 hours and the GDE were removed, washed with deionized water and followed by dryness under N2 atmosphere. Meanwhile, the electrolyte solution was refreshed for each interval.
Product Analysis. The gaseous products were collected using a gas bag and quantified by gas chromatography (GC 7890B). A thermal conductivity detector (TCD) and a flame ionization detector (FID) were used to quantify H2, CO, and other alkane contents, respectively. The Faradaic efficiency (FE) of gaseous products was calculated by the equation: Where z represents the number of electrons transferred for product formation, n is the volume concentration from GC, V is the total volume calculated by outlet flow rate, F is Faraday constant (96485 C/mol) and the Q is the amount of cumulative charge recorded by the electrochemical workstation.
The liquid product was analyzed by 1 H NMR (Bruker Advance III 400 HD spectrometer) in deuteroxide. To accurately integrate the products in NMR analysis, the sodium 2, 2-dimethyl-2-silapentane-5-sulfonate (DSS) was the reference for ethanol and acetic acid, and phenol was the reference for formate. The FE of liquid products was calculated by the equation: Where z represents the number of electrons transferred for product formation, C is the liquid concentration obtained from NMR, V is electrolyte volume, F is Faraday constant (96485 C/mol) and the Q is the amount of cumulative charge recorded by the electrochemical workstation." Please see them in pages 19-20 of the revised manuscript.

Reviewer 2:
The authors present a multi-active site catalyst by mixing Cu Nanoparticles and atomic Cu sites, both supported on an N-doped carbon matrix. Additionally, they have wisely chosen to additionally prepare catalysts comprising only Cu NPs or Cu atomic sites. The catalysts were thoroughly prepared and fully characterized. The CO2 reaction was performed in alkaline media, in a flow cell, focusing on the detection of C2 products. A wide variety of in situ/operando techniques have been used to understand the evolution of the catalysts under reaction conditions. The authors claim a mixture of CuNP and Cu atomic sites to be more beneficial than both individual systems comprising a synergistic effect. By characterizing the catalysts in the as prepared state and during reaction, the authors claim to find that C-C coupling occurs mostly on CuNP, while H2O dissociation appears mostly at atomic Cu sites, *H would migrate to the CuNP and thus promote the protonation of *CO. To achieve these results, the authors have comprehensively used different techniques to understand the catalysts under reaction conditions. The authors have started strongly demonstrating the interesting nature of the catalysts and characterizing them well in the pristine state. The data sets and their analysis of the measured data under reaction conditions or after reaction is, however, incomplete. The data for at least one catalyst for most of the presented in situ/operando techniques is not presented.
Thus, the interpretation is weakly supported on the basis of these investigations. It is suggested to integrate the missing data and to revise the presented interpretation in this manuscript majorly.

Response:
We thank the referee very much for the comment, and we have carefully revised the manuscript based on the comments. The datapoint values, error bars and total current density of CO2RR activity test have been added.
The characterization of the catalyst after reaction have been supplemented and discussed in detail, such as TEM, XRD, XPS, HAADF-TEM. More details of experiment and cell structure have been included in the method section of the manuscript. The KIE experiments, gas electro-response experiments and CO stripping experiments, in situ spectroscopic characterizations over all the three catalysts have been presented and discussed in details. Based on the results presented above, we have revised the manuscript carefully, which can be known from the detailed discussion below.
Finally, we would like to take this chance to thank the referee for the invaluable advice and suggestions, which help and guide us to improve and polish our work greatly. We have tried our best to answer the questions from the reviewers and modify the manuscript. We hope that we have addressed all of the questions satisfactorily.
1. INTRODUCTION ： The introduction should be revised providing more information and state-of-the-art knowledge about multi-active sites for multiple reactants in CO2RR (L33). A lot of studies have been performed to study the role of particle size, distance or shape. A more through summary on the literature should be presented with a focus on the relevant parameters for this work (L49-51). The cited literature should show similar reaction systems, e.g. in L56, the authors claim H2O activation on Cu single atoms, but the literature presents their data in an acidic system, while the authors study alkaline systems. A discussion why this result like this is transferable, is thus required.
A specification on the ratio of CuNP compared to Cu single atoms should be given already in L60.

Response:
We thank the referee again for the comment. According to the reviewer's suggestion, we have revised the Introduction section: "Significant efforts have been devoted to the development of dual-active sites catalysts for boosting CO2RR performance. Many reported catalysts are composed the active sites that are responsible for CO2 activation (e.g. Au, Ag and Zn) and further hydrogenation or C-C coupling (Cu), respectively. It has been demonstrated that the intermediate migration (e.g. *CO) plays a crucial role on the selectivity of the reaction, which greatly depends on the optimal site distribution and distance between the dual active sites. Meanwhile, an increase in the *CO coverage over Cu sites has been shown to promote the formation of C2+ products. 9-11 In addition, considering H2O molecule serves as the proton source for hydrogenation step, the active site for H2O dissociation has been designed to be part of the dual-active sites catalysts. 12, 13 The introduction of H2O dissociation sites, such as metal single atoms, sulfur, and oxygen vacancies, has been shown to accelerate H2O dissociation into *H species, which are subsequently fed to CO2 conversion sites through *H spillover. 14-16 However, these findings have mainly focused on the influence of H2O dissociation in CO2-to-C1 product process (formate or methane), while it is of greater significance for the C2 products formation process. This is due to the commonly used alkaline electrolyte in CO2-to-C2 product, which slows down the H2O dissociation step, resulting in a sluggish reaction kinetic procedure. 15, 17 Therefore, constructing dual-active sites electrocatalyst to accelerate CO2 reduction and H2O dissociation respectively is a feasible strategy for achieving the desired electrochemical CO2-to-C2 products performance.
Cu is the most promising electrocatalyst for converting CO2 into C2 products, due to its moderate adsorption capacity for the key intermediates. 18-20 Diverse strategies have been investigated to improve the C2 selectivity and current density, such as the manipulation of crystal facets, morphology, particle size, and oxidation state. [21][22][23] However, most of these studies mainly focus on regulating the Cu structure on CO2 activation and C-C coupling, and the reports about the effect of surface *H coverage through accelerating H2O dissociation are very limited. It is important to note that *H coverage must be controlled rigorously, since excessive *H favors H2 production via the competing hydrogen evolution reaction (HER). Recently, single-atom catalyst characterized by the uniform dispersion of isolated metal atoms on a substrate has received significant attention in HER, owing to utmost metal utilization, tunable electronic structures and structural stability. Among them, the HER activity of single-atom Cu catalyst in alkaline electrolyte can be regulated through modifying coordinated environment and support type. [24][25][26] Therefore, designing a dual-active sites catalyst with co-loaded Cu nanoparticles (NPs) and single-atom Cu sites could realize CO2 conversion and H2O activation simultaneously, favoring the high performance for electroproduction of C2 products." Please see them in Page 4 of the revised manuscript. In addition, the content ratio of atomic Cu sites to Cu NPs of M-Cu1/CuNP has been added in the last paragraph of the Introduction section.
2. TEM ANALYSIS：The Cu nanoparticle size is a relevant parameter for the catalytic performance. It would thus be relevant to mention the obtained sizes in the main text and present particle size distribution for all catalysts. The authors may also provide a NP size distribution for the R-Cu1/CuNP as they have nicely done for the other two samples. Additionally, a more exact measure for the ratio between Cu NP and single atom sizes from the TEM data should be given and discussed against the EXAFS fitted data and the observed percentages between the coordination numbers. In this regard, the HAADF-STEM images showing the single atom Cu sites are missing in Figure S1 and S2 and are asked to be provided. According to the comment, the nanoparticle size distribution of the R-Cu1/CuNP was presented as Figure S2 in the revised supporting information, which showed that the average nanoparticle size is around 5 nm. The HAADF-STEM images of P-Cu1/CuNP and R-Cu1/CuNP have been added in Figure S1 and S2 of the revised supporting information. We have discussed them in the revised manuscript by "In addition, the TEM and aberration-corrected HAADF-STEM images of P-Cu1/CuNP and R-Cu1/CuNP are shown in Figure S1 and Figure  3. XPS ANALYSIS：Please comment further on the observed differences and its relevance in the binding energies of the Cu2p3/2 peaks (L94). Additionally, fits that indicate the small amounts of Cu 2+ species in the M and P catalysts should be presented. The authors also may provide an analysis of the Auger peaks for distinguishing the metallic Cu and the Cu 1+ oxidation states of the catalysts, if possible. The content of Cu should be set into context with the content of C and N in the catalysts (L95), and any impurities should be ruled out (e.g. sulfur from guanidine thiocyanate. The analysis of N1s for the other two catalysts should be presented ( Figure S4).

Response:
We thank the referee again for the comment. To avoid the influence of long-term exposure in air, we have resynthesized the three catalysts and conducted XPS characterization. The XPS survey spectra of the catalysts have been added in the revised supporting information as Figure S3, which showed that only Cu, N, C and O elements existed in P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP. No other element could be detected. The existence of O element can be attributed to the adsorbed oxygen on surface. The contents of Cu in P-Cu1/CuNP, M-Cu1/CuNP, and R-Cu1/CuNP were 7.4, 4.3, and 1.9 at%, respectively, and the corresponding N and C content were also displayed in Table S1 of the revised supporting information. The Cu 2p peaks were fit and displayed in Figure 1F in the revised manuscript. Only Cu 2p3/2 peak attributed to Cu 0/+ existed in the spectra, and the peak moved to lower binding energy in the order of R-Cu1/CuNP (933.0 eV), M-Cu1/CuNP (932.7 eV) and P-Cu1/CuNP (932.5 eV), indicating that the valance state of Cu in the catalysts increased gradually. We have also provided the Cu LMM Auger spectra in the revised supporting information as Figure S4, which confirmed that Cu + and Cu 0 co-existed in P-Cu1/CuNP, M-  Table   S1. The Cu 2p spectra ( Figure 1F) suggest that only Cu 2p3/2 peak attributed to Cu 0/+ existed in the spectra, which moved to lower binding energy in the order of R-Cu1/CuNP (933.0 eV), M-Cu1/CuNP (932.7 eV) and P-Cu1/CuNP (932.5 eV), indicating that the valance state of Cu in the catalysts increased gradually. The Cu LMM Auger spectra confirmed that Cu + and Cu 0 coexisted in P-Cu1/CuNP M-Cu1/CuNP and R-Cu1/CuNP ( Figure S4). Considering the valance state of Cu in Cu NP is 0, while that of atomically dispersed Cu site is Cu δ+ , thereby the content of Cu1 site in the catalyst changed in order of P-Cu1/CuNP < M-Cu1/CuNP < R-Cu1/CuNP. Meanwhile, the high resolution XPS N 1s spectra of P-Cu1/CuNP, M-Cu1/CuNP, and R-Cu1/CuNP are displayed in Figure S5. All the spectra showed obvious Cu-N peak at around 399.2 eV, which suggested that the Cu1 sites were coordinated by N atoms and other peaks could be attributed to pyridinic N (~398.5 eV), pyrrolic N (~399.9 eV), and graphitic (~401.0 eV),  Figure   1G: The authors may also comment why their data quality is so low compared to their mentioned reference literature and what they could significantly learn from the XES experiments, which they did not learn from XAS or XPS.

Response:
We thank the referee again for the comment. The XES experiments was adopted to identify the coordination atoms of Cu single atoms and Cu valence state. However, the data quality was much lower than that reported in literature due to the limitation of the detector accuracy in XES experiments, which affected the validity of the conclusions. Additionally, according to the results of XPS and XAS, it can be identified that Cu valence state varied in the sequence of P-Cu1/CuNP < M-Cu1/CuNP < R-Cu1/CuNP, and Cu single atoms were coordinated by N atoms. Therefore, according to the comment, we have deleted the XES experiments data, because we have already obtained the same conclusion by the more accurate technique (XPS and XAS).
5. XAS ANALYSIS：The XAS analysis has been thoroughly performed. However, the descriptive text starting from L106 is not clear. E.g., in L108, it is unclear, which catalyst is meant. A much as the Cu valence state increases in a sequence it is more vital in the reviewer's opinion, that the P-and the R-catalyst show very similar intensity transients than the two references, indicating that both could be used to represent the reference catalysts. Please also comment on the artifact(?) peaks around 1Å (L112, Figure 1I) and on the calculation of the percentage between Cu NP and single Cu sites (L117). Please also comment on the maximum observed coordination number of the Cu NPs and if an estimation of the particle size can be extracted from it.

Response:
We thank the referee again for the comment. The XAS characterization of R-Cu1/CuNP has been reperformed and Cu valence state was analyzed through XANES spectra and the results were presented as Figure   1G in the revised manuscript. The Cu K-edge XANES of the three catalysts was compared with reference samples to determine the average Cu valence state, which was found to lie between 0 and +2. The average Cu valence state varied in the order of P-Cu1/CuNP < M-Cu1/CuNP < R-Cu1/CuNP, consistent with XPS results. The FT EXAFS of the three catalysts was analyzed and Figure 1H in the revised manuscript shows that all three catalysts exhibited peaks at around 1.5 Å and 2.2 Å, attributed to Cu-N and Cu-Cu coordination, respectively. Therefore, atomic Cu and metallic Cu NP co-existed in the three catalysts. Notably, the intensity ratio of the Cu-N peak to the Cu-Cu peak increased with the sequence progressed from P-Cu1/CuNP, M-Cu1/CuNP to R-Cu1/CuNP, indicating a gradual increase in the content ratio of Cu1 to Cu NP. The peak located below 1 Å, i.e., low-frequency structure, can be ascribed to the resonant scattering from the electronic states within the central atom. The structure is not part of the structural information. These non-structural contributions to the total absorption are considered to be part of the background (Phys. Rev. B 1993, 47,14126-14131). The percentage of Cu NP and single Cu site was calculated according to the method reported in the literature (J. Phys. Chem. C 2012, 116, 24999-25003). The structural information provided by XAFS is the average value and the coordination number obtained through EXAFS fitting is the information of atomic structure. It has not been reported that estimating particle size from EXAFS fitting coordination number.
In the revised manuscript, we have discussed them by "To further determine the chemical state and local coordination environment, X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) measurements were conducted. The Cu K-edge XANES edge of R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP with the reference samples Cu foil and Cu phthalocyanine (CuPc) were presented in Figure 1G, the adsorption edge of all catalysts located between that of Cu foil and CuPc, indicating the average Cu valence state of Cu in the catalysts laid between 0 and +2. Meanwhile, adsorption edge moved to higher energy in the sequence of P-Cu1/CuNP < M-Cu1/CuNP < R-Cu1/CuNP, confirming that the average Cu valence state increased with the increase of Cu1/CuNP ratio. It is in accordance with the XPS results. The Fourier-transformed (FT) EXAFS spectra in the Rspace of the three catalysts was processed and displayed in Figure 1H, all three catalysts exhibited peaks at around 1.5 Å and 2.2 Å, attributed to Cu-N and Cu-Cu coordination, respectively. Therefore, atomic Cu and metallic Cu NP co-existed in R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. Notably, the intensity ratio of the Cu-N peak to the Cu-Cu peak increased in the order of P-Cu1/CuNP < M-Cu1/CuNP < R-Cu1/CuNP, indicating a gradual increase in the content ratio of atomic Cu to Cu NP. In order to obtain the local structure of Cu species and the content ratio of atomic Cu to Cu NP, the quantitative analysis by the least-squares EXAFS fittings was performed ( Figure S6). The percentages (P) of Cu-N4 (Cu1) and Cu NPs in total Cu species of catalysts were displayed in Table S2. Thus, the content ratio of Cu1 to CuNP of P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP were 0.39, 0.25 and 0.05, respectively, which is consistent with our catalyst design expectation." Please see them in Page 6-7 of the revised manuscript.
6. ELECTROCATALYTIC CO2RR PERFORMANCE：The authors present very interesting and carefully studied CO2 reduction reaction results. The P-and R-catalysts should however be set into the context of literature, as Cu NPs and Cu single atoms are already well studied.

Response:
We thank the referee again for the comment. We have discussed them in the revised manuscript by "The control potential electrolysis was then performed to analyze the reduction products and the catalyst loading was 1 mg cm -2 ( Figure S9). The gas-phase and liquid-phase products were analyzed by gas chromatography and 1 H nuclear magnetic resonance spectroscopy, respectively. H2, CO, CH4, formate, C2H4, ethanol and acetate were formed. Figure 2A shows that the FE of C2 products (FEC2) of M-Cu1/CuNP exhibited a volcano-shaped dependence on the applied potential ( Figure S10, Tables S3-S5), and a maximum FEC2 could reach 75.4% at -0.6 V, which is much higher than that over R-Cu1/CuNP and P-Cu1/CuNP. The maximum FEC2 of P-Cu1/CuNP was 47.3% at -0.7 V, which closed to the performance of Cu nanoparticles reported in the literature. 22, 27 Moreover, the M-Cu1/CuNP had a lower onset potential for C2 products formation. The FEC2 could reach 20.5% at -0.4 V over M-Cu1/CuNP, while C2 products cannot be detected over R-Cu1/CuNP and P-Cu1/CuNP under the same potential. The ratio of FEC2 to FEC1 could keep > 4.5 from -0.6 to -0.9 V over M-Cu1/CuNP, while those of R-Cu1/CuNP and P-Cu1/CuNP were below 2 in the whole applied potential range ( Figure 2B). Furthermore, M-Cu1/CuNP was treated by sulfuric acid solution to completely remove Cu NPs (denoted as Cu-N-C) and the results of TEM, XRD and XAS indicated that only Cu single atoms existed in the Cu-N-C catalyst (Figures S11, S12). H2 was the dominant product over Cu-N-C in the whole applied potentials range, suggesting that the atomic Cu sites mainly facilitated H2O dissociation ( Figure 2C and Table S6).
The formation of CO and CH4 was also observed, in agreement with previous reports that Cu single atoms could generate H2, CO, and CH4. 25,28,29 These results suggest that there is synergistic effect between Cu sites and Cu NPs, and the proper ratio of Cu1 to CuNP would obviously enhance the selectivity for C2 products." Please see them in Page 9-10 of the revised manuscript.
7. How were the presented LSVs in Figure S6 normalized? The reviewer assumes that a normalization to the ECSA results, which are presented in Figure S10, would change the results and thus the interpretation. A thoughtful analysis of the electrochemical surface area is necessary here to evaluate the observed current densities. Additionally, Figure   S6, and in general for most Figures, specifically the electrochemistry figures, lack a complete description of the used electrolyte, scan rate and normalization, which should be provided.

Response:
We thank the referee again for the comment. The LSV curves presented in Figure S6 were not normalized by ECSA, while the potentials were treated by 80% ohmic compensation. Although the Hg/HgO electrode was used as reference electrode during experiments, the applied potentials were converted to versus reversible hydrogen electrode (RHE) with 80% ohmic compensation in figures. The formula for the conversion as: where EHg/HgO represents the potential versus Hg/HgO electrode, pH is the pH value of electrolyte, i represents the current obtained at corresponding potential, R is ohmic resistance of the cell measured by electrochemical workstation. The experiment details have been added in the Method section of revised manuscript and we have marked the test conditions in the caption of electrochemistry figures in revised manuscript and supporting information.
The LSV curves are commonly used to judge whether CO2RR occurred and measure the onset potential. It is seldom reported to normalize LSV curve by using ECSA. As shown in Figure S8 of the revised supporting information, R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP showed significant increase in cathodic current density when feed gas was change from N2 to CO2, indicating that CO2RR occurred over all the three catalysts. Meanwhile, M-Cu1/CuNP exhibited the most positive onset potential and highest current density in the presence of CO2. The current density reached 1093.0 mA cm -2 over M-Cu1/CuNP at -0.8 V, which was roughly 2.5 and 1.7 times higher than that over R-Cu1/CuNP and P-Cu1/CuNP, respectively. The results suggested that M-Cu1/CuNP has higher CO2 activity than R-Cu1/CuNP and P-Cu1/CuNP, which could be attributed to the moderate content ratio of Cu1 to CuNP.
In the revised manuscript, we have discussed them by "The CO2RR performance of the as-prepared catalysts were first evaluated by linear sweep voltammetry (LSV) in 5 M KOH aqueous solution in a flow cell ( Figure S7). As shown in Figure S8, all the catalysts exhibited significant increase in cathodic current density when feed gas was change from N2 to CO2, indicating that CO2RR occurred over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. Meanwhile, M-Cu1/CuNP exhibited the most positive onset potential and highest cathodic current density in the presence of CO2 gas, and taking -0.8 V as an example, the current density reached up to 1093.0 mA cm -2 over M-Cu1/CuNP, which was roughly 2.5 and 1.7 times higher than that of R-Cu1/CuNP and P-Cu1/CuNP, respectively. Therefore, the results of LSV experiments suggested that M-Cu1/CuNP has higher CO2 activity than R-Cu1/CuNP and P-Cu1/CuNP, which could be attributed to the moderate content ratio of Cu1 to CuNP." Please see them in page 9 of the revised manuscript.
8. It is unclear, which figure is described here in L136. The data in Figure 2A are nicely evaluated. Please also add the obtained total currents to the graphs. For transparency, please also add a table with the obtained values and error bars (or add them to the plots). Describe in the experimental data how many datapoints are comprised in one error bar.

Response:
We thank the referee again for the comment. The figure is described in L136 is Figure 2A. We apologize for the confusion caused by unclear expression and have revised the manuscript as " Figure 2A shows that the FE of C2 products (FEC2) of M-Cu1/CuNP exhibited a volcano-shaped dependence on the applied potential ( Figure S10, Tables S3-S5), and a maximum FEC2 could reach 75.4% at -0.6 V, which is much higher than that over R-Cu1/CuNP and P-Cu1/CuNP. The maximum FEC2 of P-Cu1/CuNP was 47.3% at -0.7 V, which closed to the performance of Cu nanoparticles reported in the literature. 22, 27 " According to the comment, we have added total current density in Figure S13 in the revised supporting information. Three datapoints are comprised in one error bar. We have described it in the method section of the revised manuscript and added tables with datapoint values and error bars (Tables S3-S5) in the revised supporting information.
9. The additional synthesis procedure for the Cu-N-C catalyst should be described in the experimental data and in the figure descriptions Figures S7-S9). It looks unorganized if this catalyst is only introduced here, but not in the general catalyst characterization in Figure 1. It is proposed that this Cu-N-C catalyst is used as a 4th main catalyst, and the whole characterization is also performed on this one. Please comment on the huge H2 production on the Cu-N-C and the much better performance of the R-Cu1/CuNP (L145). Please also relate the results of Cu-N-C to literature.
Discuss that Cu-N-C catalysts should in principle hamper any C-C coupling, but still small amounts of ethylene are formed around -0.6 V.

Response:
We thank the referee again for the comment. The synthesis procedure of Cu-N-C has been supplemented in the Methods section of the revised manuscript and figure description of Figures S11, S12 in the revised supporting information. The Cu-N-C catalyst was not used as a 4th main catalyst, and it was only a sample obtained by acid treatment of M-Cu1/CuNP. The activity of Cu-N-C catalyst was measured to confirm the role of Cu single atom, so the Cu nanoparticles or clusters in Cu-N-C catalyst must be completely removed. We have re-optimized the conditions of acid treatment and the related part of the manuscript is changed into "The M-Cu1/CuNP was added into 50 mL 1 M sulfuric acid aqueous solution and heated at 80 °C for 48 h, then washed with deionized water several times and dried at 80°C overnight." in the Methods section of revised manuscript and the figure descriptions of Figures S11, S12 in the revised supporting information.
TEM, XRD, XPS and XAS characterization have been conducted for Cu-N-C catalyst. No obvious nanoparticles were observed in TEM image and a large amount of isolated bright dots recognized as Cu single atoms were observed in aberration-corrected HAADF-STEM image. No diffraction peaks related to crystalline Cu species can be identified in XRD pattern. XPS spectra showed that Cu element existed in Cu-N-C catalyst and Cu + was main species. More importantly, the EXAFS profiles in the R-space of Cu-N-C suggested that only the peak attributed to Cu-N coordination located at around 1.4 Å could be observed, while the Cu-Cu coordination peak at around 2.2 Å disappeared, which confirmed that the Cu species only existed as single atom form. The results of least-squares EXAFS fitting confirmed that the Cu-N coordination numbers in Cu-N-C was 4.0 (Table S2 of the revised supporting information), implying that the atomic Cu species mainly existed as Cu-N4 structure. We have discussed the characterization results of Cu-N-C in the figure description of Figures S11 and S12.
The activity of the Cu-N-C obtained through acid treatment was remeasured and the data was presented in the revised manuscript as Figure 2C and the revised supporting information as Table S6. The results indicated that H2 was the dominant product in the whole applied potentials range. CO and CH4 was also detected, which was in agreement with previous reports in the literature (Small Struct. 2021, 2, 2000058;Nat. Commun. 2021, 12, 586;Appl. Surf. Sci. 2023, 610, 155506). It indicated that Cu single atoms could generate H2, CO, and CH4. We have discussed them in the revised manuscript as "Furthermore, M-Cu1/CuNP was treated by sulfuric acid solution to completely remove Cu NPs (denoted as Cu-N-C) and the results of TEM, XRD, XPS and XAS indicated that only Cu single atoms existed in the Cu-N-C catalyst (Figures S11, S12). H2 was the dominant product over Cu-N-C in the whole applied potentials range, suggesting that the atomic Cu sites mainly facilitated H2O dissociation ( Figure   2C and Table S6). The formation of CO and CH4 was also observed, in agreement with previous reports that Cu single atoms could generate H2, CO, and CH4. 25, 28, 29 ". Please see them in pages 9-10 of the revised manuscript. As for R-Cu1/CuNP, we have reperformed characterization over R-Cu1/CuNP, and found Cu nanoparticles on R-Cu1/CuNP through TEM ( Figure S2 in the revised supporting information). The results of XAS were showed in the Figures 1G and 1H of the revised manuscript, which indicated that Cu single atoms and Cu nanoparticles co-existed in R-Cu1/CuNP and the content ratio of Cu1 to Cu NP was 0.39 (Table S2 in the revised supporting information).
10. The discrepancy of the obtained results from DL Capacitance are huge (L145). If the M-and P-catalysts would be normalized to the R-catalyst, roughness factors of 1, 2 and 5 would emerge (R, M, P). It is thus not enough to just normalize the CO2RR data to these values due to many other parameters that could bias the results (loading, particle distances, agglomeration, ...). Please justify your results by remeasuring one or two potentials of your catalysts with similar ECSA (e.g. by lowering the loading)

Response:
We thank the referee again for the comment. We agree with reviewer's opinion that just normalizing CO2RR data by ECSA is limit. Therefore, the normalized C2 products partial current density by ECSA was abandoned in the revised manuscript. Meanwhile, we have investigated the influence of M-Cu1/CuNP loading on CO2RR activity. The CO2RR activity of the working electrode with 0.25, 0.5, 1, 1.5 and 2 mg cm -2 M-Cu1/CuNP were measured in 5 M KOH electrolyte at -0.6 V in the flow cell. The catholyte solution and anolyte solution were recirculated by two pumps with flow rates of 10 mL min -1 and 30 mL min -1 , respectively. CO2 gas was continuously supplied to the gas chamber by using a mass flow controller with a flow rate of 40 mL min -1 . The results were added as Figure S9 in the revised supporting information. Both C2 products selectivity and total current density increased as the increasing loading of M-Cu1/CuNP until 1 mg cm -2 , and no obvious difference was observed among 1, 1.5 and 2 mg cm -2 . Therefore, we chose 1 mg cm -2 as the catalyst loading to conduct activity experiments.
In the revised manuscript, we have discussed them by "The control potential electrolysis was then performed to analyze the reduction products and the catalyst loading was 1 mg cm -2 ( Figure S9)." Please see them in Page 9 of the revised manuscript.
11. Please establish your choice to measure chronopotentiometry for the long-term stability test instead of chronoamperometry (L155). Discuss the long-term stability results on the results of the other two (or three catalysts).
How would you expect that the R-and P-catalysts behave during 40 h? Please also show the results of the other products, especially hydrogen. Unfortunately, the presented TEM, XPS and XRD data (Figures S12-S14) after the long-term stability test are not analyzed. For the TEM, an evaluation of the NP size is lacking, as well as an HAADF-STEM image showing the presence and amount of the single atom sites. The authors should provide the same type of data, that they have nicely provided in Figure 1A-D. XPS: An evaluation of the oxidation state (Cu2p with fits and, if possible, Auger data) and chemical composition (Cu/C/N ratio) should be given. Differences, e.g. of the N1s composition change before and after reaction should be discussed. Here, the CO2RR conditions should be described in the descriptions of these Figures. Similar for all shown data after reaction. Please specify how the regular refreshing of the catalyst could influence the stability and product distribution (L164). This important experimental detail should be noted in the experimental data, and maybe even in the main text as it may be deceptive for the reader of this manuscript. Please discuss and demonstrate that flooding might not alter the catalyst, but that carbonation does and that it can be reversed.

Response:
We thank the referee again for the comment. In order to diminish the impact of solution resistance during long-term electroreduction, chronopotentiometry is a common method used for evaluating long-term CO2RR activity stability of electrocatalyst in a flow cell (J. Am. Chem. Soc. 2022, 144, 259-269;Angew. Chem. Int. Ed. 2022,61, e202113498;Angew. Chem. Int. Ed. 2022,61, e202110657). The selectivity of C1, C2 products and H2 during the long-term stability test were added and the long-term stability results of the other two catalysts have been presented as Figure S14 in the revised supporting information. R-Cu1/CuNP and P-Cu1/CuNP showed lower C2 products selectivity than that of R-Cu1/CuNP and no obvious changes in product distribution or potential were observed over the course of the long-term stability tests.
In the revised manuscript, we have discussed them by "The long-term stability test was conducted through chronopotentiometry for 40 h and the electrode was washed, then dried and the electrolyte was refreshed at intervals 5 h to address the issues of flooding and carbonation. Figure 2E and Figure S14 showed that no obvious change was observed over the potential and products selectivity, and the high FE of C2 products could be held over 70% on M-Cu1/CuNP during the electrolysis." Please see them in page 10 of the revised manuscript.
According to the comment, we have added more characterization over M-Cu1/CuNP after long-term CO2RR The CO2RR stability is affected by catalyst, electrolyte, gas diffusion electrode (GDE) and so on. In flow cell, CO2RR occurred at the three-phase interface formed by GDE with CO2 gas channel and hydrophobic layer. CO2 can react with hydroxide in electrolyte to form carbonate precipitation during CO2RR process, and the accumulation of carbonate at three-phase interface would block CO2 gas channel and the gas diffusion electrode suffers from losing hydrophobicity and being flooded over time. Both CO2 gas channel blockage and flooding would destroy GDE stability, and thus affect the catalytic performance. Meanwhile, the water dissociation, accumulation of liquid products and carbonate formation would change the properties of electrolyte, and also affect the catalyst performance.
Therefore, in order to verify the stability of the catalyst, the influence of GDE and electrolyte should be minimized.
To issue the problem of gas channel blockage and flooding, the CO2RR was interrupted every 5 hours and the GDE were removed, washed with deionized water and followed by dryness under N2 atmosphere. Meanwhile, the fresh electrolyte solution was pumped into flow cell for each interval to keep the electrolyte property stable. The results of long-term CO2RR experiment showed that C2 products selectivity and potential had no obvious change after minimizing effect of GDE and electrolyte, demonstrating the catalyst had high stability. In the revised manuscript, we have supplemented the experiment details in Methods section of the revised manuscript: "The chronopotentiometry method was used for evaluating long-term stability of electrocatalyst in 5 M KOH electrolyte.
In order to address the issues of flooding and carbonation accumulation of GDE, the CO2RR was interrupted every 5 hours and the GDE were removed, washed with deionized water and followed by dryness under N2 atmosphere.
Meanwhile, the electrolyte solution was refreshed for each interval." Please see them in page 20 of the revised manuscript.
12. KINETIC ISOTOPE EFFECT：The authors provide a decent study on the origin of the hydrogen used for C2 products. These results could be justified by KIE results of R-Cu1/CuNP and Cu-NC (L172). The authors should provide these investigations. It should also be discussed how the hydrogen formation is altered with D2O experiments. Again, please specify the experimental conditions in Figure 3. (potential/scan rate, electrolyte, CO2 saturation). The authors also present a small study on the influence of the electrolyte concentration on the formation of ethylene. Please describe why the formation rate instead of the FE or partial current densities are chosen here for the analysis and how it justifies the interpretation. Results for R-Cu1/CuNP should also be given. The potential and electrochemical details should be provided in Figure S15.

Response:
We thank the referee again for the comment, and we answer the two questions as follows.
(1) The KIE results of R-Cu1/CuNP have been added in Figure 3A of the revised manuscript. The KIE value of R-Cu1/CuNP was 1.12, which was less than that of P-Cu1/CuNP (2.87) and M-Cu1/CuNP (1.21), further indicating that H2O dissociation was accelerated gradually as ratio of Cu1 to CuNP increased. Additionally, the change of H2 formation rate have been displayed as Figure S17 in the revised supporting information. When the H2O was replaced by D2O in the electrolyte, an obvious decrease was observed over all the catalysts, while the decreasing degree followed the sequence of P-Cu1/CuNP > M-Cu1/CuNP > R-Cu1/CuNP, suggesting that the increase of Cu single atom content can reduce the influence of isotope effect on H2 formation rate.
In the revised manuscript, we have discussed them by "The kinetic isotope effect (KIE) of H2O/D2O (H/D) experiments were performed to get insights into the role of H2O dissociation in CO2-to-C2 products ( Figure 3A).
When the H2O was replaced by D2O in the electrolyte, the formation rate of the product (for example C2H4) decreased over P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP, and the level of its decrease was closely related to the content ratio of Cu1/CuNP. If the KIE value (defined as the ratio of C2H4 formation rates in H2O and D2O) closes to 1, H2O dissociation is not the rate-determining step over the catalyst. The KIE value for R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP were 1.12, 1.21 and 2.87, respectively, which indicated that H2O dissociation was accelerated gradually as the increasing of Cu1/CuNP content ratio. These results confirmed that the atomic Cu sites were responsible for accelerating H2O dissociation and provided proton to adjacent Cu NPs, thus affecting CO2-to-C2 products. Meanwhile, the N-doped carbon matrix catalysts has been reported to favor the migration of proton. 30 In addition, the H2 formation rate in 5 M KOH H2O solution and D2O solution was presented in Figure S17. An obvious decrease was observed over all the catalysts, while the decreasing degree followed the sequence of P-Cu1/CuNP > M-Cu1/CuNP > R-Cu1/CuNP suggesting that the increase of Cu single atom content can reduce the influence of isotope effect on H2 formation rate." Please see them in page 12 of the revised manuscript. Besides, we have added the experiments conditions in the caption of Figure 3.
(2) The results of the electrolyte concentration influence over R-Cu1/CuNP have been added in Figure S18 of the revised supporting information. We have discussed the results in the revised manuscript as "Furthermore, we studied the influence of electrolyte pH on CO2RR performance over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. Three different concentrations of KOH aqueous electrolytes, i.e., 0.1M, 3M and 5M, were employed to adjust the pH environment at the electrode/electrolyte interface. The results in Figure S18 show that the FE and formation rate of Both FE and formation rate were chosen here for the analysis, and the C2 products FE and formation rate over P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP increased with the increasing pH value of electrolyte, suggesting that strong basic local environment favored the C2 products formation. FE reflects the products selectivity of catalyst, while formation rate reflects products formation activity. The products formation rate is calculated from the products partial current density considering the time parameter, so they reflect similar property of electrocatalyst. However, the products formation rate is more intuitive than the partial current density, so that we used "formation rate". It is also a commonly used parameter in the literature (Nat. Commun. 2022, 13, 1965Nat. Commun. 2019, 10, 892). In the revised supporting information, the potential and electrochemical details were provided in Figure S18.
13. ELECTRORESPONSE & CO Stripping：Discuss why the used Cu foam as support does not take part in the experiment. Discuss the relevance of catalyst loading and ECSA for this experiment. Please provide the results for the missing catalysts. The showed results indicate intrinsic differences between the catalysts, but could also very likely be just a roughening/loading effect. Please add experimental details also in the experimental section and add a scheme of the device ( Figure S17). Similarly, please provide experimental details of the CO stripping voltammetry in the experimental section. Explain, how the CO was adsorbed on the surface and please add the LSV of the double layer without the CO stripping peak. Again, please also provide the results for the missing catalyst. The interpretation of the electroresponse and CO stripping experiments are vague, and need clarification, providing that the "reference catalyst" P-Cu1/CuNP was left out. Please set the results in the context of literature.

Response:
We thank the referee again for the comment, and we answer the two questions as follows.
(1) In the process of the gas electro-response experiments, the used Cu foam took part in the gas electro-response experiments. We kept the same Cu foam size and catalyst loading for each experiment, thereby the differences in current density shown in the experiment was attributed to the intrinsic property of the catalysts. The results of P-Cu1/CuNP and the device scheme have been added in the revised supporting information (Figures S19-S21). Since the experiment is a gas adsorption process on solid surface, the concept of ECSA in not involved. In the revised manuscript, we have discussed them by "We then investigated the capacity adsorption and activation of CO2 and CO molecules on R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP via gas electro-response experiments in a self-designed gas adsorption electroresponse device ( Figure S19). 31 The results in Figure S20 and Figure S21 show that the CO2 and CO adsorption responses changed in the sequence of R-Cu1/CuNP < M-Cu1/CuNP < P-Cu1/CuNP, suggesting that the adsorption and activation of CO2 and CO were promoted as the ratio of Cu1 to CuNP decreased." Please see them in page 13 of the revised manuscript.
In addition, the experimental details of gas electro-response experiments have been added in the Methods section of the revised manuscript: "A self-designed gas adsorption electroresponse device ( Figure S19) was used to perform gas electroresponse experiments. To prepare the electrode, 5 mg catalyst was suspended in 1 mL isopropanol to form a homogeneous ink. Then the catalyst ink was spread onto the Cu foam with area of 1 × 2 cm 2 by a micropipette and then dried under room temperature. The as-prepared electrode was put into a sealed container and connected with electrochemical workstation though two electrode system. Before Ar, CO2 or CO gas was injected (2) During the CO stripping experiments, the CO adsorption procedure was accomplished by polarizing the working electrode at +0.2 V and bubbling the electrolyte with CO for 10 min. The experimental details of the CO stripping test have been added in the Methods section of the revised manuscript: "CO stripping experiments were conducted in a single-chamber electrolytic cell with three electrode system through linear sweep voltammetry method. To prepare the working electrode, 1 mg catalyst was suspended in 500 μL isopropanol with 10 μL Nafion D-521 dispersion (5 wt%) to form a homogeneous ink. Then the catalyst ink was spread onto the carbon paper of 1 × 1 cm 2 in area and then dried under room temperature. Ag/AgCl electrode and graphite rod were used as reference electrode and counter electrode, respectively. The electrolyte was 0.1 M Na2SO4 aqueous solution. The CO adsorption procedure was accomplished by polarizing the working electrode at +0.2 V and bubbling the electrolyte with CO for 10 min and subsequently with N2 for another 10 min. Then the linear sweep voltammetry was conducted from 0.5 to 1.2 V with scan rate of 10 mV s -1 ." Please see them in page 21 of the revised manuscript.
According to the comment, the CO stripping result of P-Cu1/CuNP has been added in Figure 3B of the revised manuscript and the LSV curve without CO stripping has displayed as Figure S22 in the revised supporting information. In the revised manuscript, we have discussed them by "On the other hand, the electrochemical CO stripping voltammetry tests of R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP were also performed to study the CO adsorption ability. Figure 3B shows a peak in the potential range of 0.8-0.9 V for R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. No peak was observed in the LSV curve without CO adsorption ( Figure S22). Further according to pervious literature, 32-34 the peak at around 0.8-0.9 V can be attributed to the CO stripping peak. Interestingly, the CO stripping peak occurred around 0.89 V for M-Cu1/CuNP and P-Cu1/CuNP, while it was around 0.82 V for R-Cu1/CuNP.
The positive shift suggested that M-Cu1/CuNP and P-Cu1/CuNP had stronger CO binding ability than R-Cu1/CuNP, which have higher Cu NP content. The results above indicated that the Cu NPs was beneficial for the adsorption and activation of CO2 and CO." Please see them in page 13 of the revised manuscript.
14. OPERANDO XAS：The show results for operando XAS are incomplete and the interpretation is based on weak presented data. The authors should quantify the increase of the Cu valence state for example with Linear-Combination fitting (L194). It is not clear, how long the catalyst was set at -0.6 V before the shown spectrum was measured. It could very likely be that the experiment did not perform as wished. Different (experimental) parameters can play a role, why the catalyst did not reduce completely or oxidizes again. Parameters that could show a working system could be the following: i) current similar to the experimental conditions in the lab ii) similar product detection iii) different potentials to see the evolution of the catalyst. Please discuss. Thus the assumption towards adsorbed CO2 and H2O on the basis of the shown data is unfortunately highly overinterpreted. Please show the data of several potentials and justify with the other catalysts e.g. that Cu NPs can reduce completely, and/or that Cu single atoms behave different. L196: The authors should also provide the least-square fittings (L196) and clarify how you attribute/correlate the increased Cu-N/O coordination number to adsorbed CO2 and H2O (L198). It is quite hard from EXAFS data to see the local coordination of adsorbed species on the data. In some cases, however, it is evaluable from XANES data or from XES data.

Response:
We thank the referee again for the comment. An optical photo of in situ XAS experiment has been included as Figure S23 in the revised supporting information. In order to overcome the condition limitation of in situ XAS test and ensure data quality, the flow cell used in the in situ XAS experiments was designed differently from that used in CO2RR activity tests. The differences between the two flow cells included the distance between the working electrode and counter electrode, as well as the site of the electrolyte inlet and outlet, resulting in variations in the measured current. We endeavored to simulate laboratory test conditions as closely as possible. We have reconducted in situ XAS experiments over M-Cu1/CuNP, and the spectra of XANES and FT EXAFS were presented as Figure S24 in the revised supporting information and Figure 3C in the revised manuscript. The Cu K-edge adsorption spectra did not show any significant difference under different potentials, indicating the stable nature of the average Cu valence state during CO2RR. Moreover, the peaks attributed to Cu-N and Cu-Cu coordination still existed in FT EXAFS, with no notable changes observed in peak intensity, indicating the stability of the content ratio of Cu1 to CuNP during CO2RR. These results suggested that the structure and content of Cu1 and CuNP remained stable during CO2RR. The in situ XAS experiments of P-Cu1/CuNP and R-Cu1/CuNP were omitted, considering that the difference among the three catalysts is Cu1 and CuNP content and the limited XAS testing time.
In the revised manuscript, we have discussed the in situ XAS experiments by "In situ X-ray absorption spectroscopy (XAS) experiments were performed on M-Cu1/CuNP to investigate the changes of Cu valence state and structure during CO2RR ( Figure S23). In the XANES spectra ( Figure S24), the Cu K-edge adsorption spectra did not show obvious difference under different potentials, indicating that the average Cu valence state kept stable during the reaction. Moreover, the peaks attributed to Cu-N and Cu-Cu coordination still existed in FT EXAFS spectra ( Figure   3C), with no notable changes observed in peak intensity, suggesting the stability of the content ratio of Cu1 to CuNP.
The results of in situ XAS experiments showed that the structure and content of Cu1 and CuNP remained stable during CO2RR." Please see them in pages 13-14 of the revised manuscript.

INSITU-SERS and ATR-SEIRAS：The authors should specify the CO stretching and CO rotation in L201 and
provide the results of the other two catalysts (L203). It would be intriguing to see how the CO adsorption changes for the other catalysts. What did the authors learn from these experiments? It is well known that CO is an intermediate for CO2 reduction. Would the COads be lacking for the R-Cu1/CuNP or the Cu-N-C catalyst? Are differences in the CO onset potential seen that would be relatable to the earlier onset for C2+ products for the M-catalyst? The interpretation on the experiments of ATR-SEIRAS are weak without the measurements of the other catalysts. The authors should provide complete data sets of the other catalysts.

Response:
We thank the referee again for the comment, and we answer the two questions as follows.
(1) The photograph and scheme of cell used for in situ SERS spectroscopy were showed as Figure S25 in the revised supporting information. The in-situ SERS spectroscopy of R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP were presented in Figures 3D-3F in the revised manuscript. The peaks at round 298, 365 and 2100 cm -1 were assigned to adsorbed CO restricted rotation (ν(Cu-CO)), Cu-CO stretching (ν(Cu-CO)) and C-O stretching of atop *CO (ν(CO)), respectively. The peaks can be observed from -0.3 to -0.8 V over M-Cu1/CuNP, while ν(Cu-CO) and ν(CO) peaks appeared at -0.4 V and -0.5 V, respectively, on P-Cu1/CuNP. This observation could be relatable to the lower onset potential for C2 products for M-Cu1/CuNP. Only CO rotation peaks can be observed on R-Cu1/CuNP from -0.4 to -0.8 V, meaning that high ratio of Cu1 to CuNP did not facilitate CO2 reduction. In situ SERS of Cu-N-C has been added as Figure S26 in the revised supporting information, and no peaks associated with *CO intermediate were observed on Cu-N-C, suggesting that single atomic Cu was not conducive to the conversion of CO2 to adsorbed CO. In the revised manuscript, we have discussed them by "In situ surface-enhanced Raman spectroscopy (SERS) was employed to reveal the interactions between Cu species and the reaction intermediates during CO2RR ( Figure   S25). [35][36][37][38] As shown in the Raman spectra under different potentials ( Figures 3D-3F), the peaks assigned to the restricted rotation of adsorbed CO (298 cm -1 , ν(Cu-CO)), Cu-CO stretching (365 cm -1 , ν(Cu-CO)) and C-O stretching of atop *CO (2000-2100 cm -1 , ν(CO)) were observed from -0.3 to -0.8 V over M-Cu1/CuNP, while ν(Cu-CO) and ν(CO) peaks over P-Cu1/CuNP appeared from -0.4 V and -0.5 V, respectively. This observation could be relatable to the lower onset potential for C2 products for M-Cu1/CuNP. Only CO rotation peaks can be observed on R-Cu1/CuNP from -0.4 to -0.8 V, meaning poor CO2 reduction activity. In situ SERS was also conducted over Cu-N-C ( Figure S26), whereas no peaks associated with *CO intermediate were observed, suggesting that single atomic Cu was not conducive to the conversion of CO2 to adsorbed CO. The results of in situ SERS indicated that CO2-to-C2 products proceeded a CO intermediate process, and moderate content ratio of Cu1 to CuNP facilitated CO2 conversion to adsorbed CO." Please see them in page 14 of the revised manuscript.
(2) The photograph and scheme of cell used for in situ ATR-SEIRAS spectroscopy are showed as Figure S27 in the revised supporting information. In situ ATR-SEIRAS spectroscopy of R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP are presented in Figures 3G-3I in the revised manuscript. The peak attributed to *CO existed on all the three catalysts.
However, as the potentials moved to negative value, the *CO peak over M-Cu1/CuNP appeared earlier than that of R-Cu1/CuNP and P-Cu1/CuNP, and the peak intensity was stronger, which indicated that M-Cu1/CuNP facilitated CO2 conversion to adsorbed CO. Additionally, the peak assigned *CHO only appeared over M-Cu1/CuNP, and its change trend was similar to that of *CO intermediate, indicating that the *CHO was originated from the hydrogenation of *CO with the assistance of *H. The results of in situ ATR-SEIRAS suggested that proper content ratio of Cu1 to CuNP is necessary for *CHO formation. In the revised manuscript, we have discussed them by "In situ attenuated total reflection-surface-enhanced IR absorption spectroscopy (ATR-SEIRAS) spectra were further collected to trace the evolution of reaction intermediates during CO2RR from -0.2 to -0.9 V over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP ( Figures 3G-3I, Figure S27). The peak at around 2100 cm -1 can be attributed to electrogenerated CO adsorbed (*CO) on catalyst surface, 39, 40 which did not appear until the applied potential reached -0.5 V on R-Cu1/CuNP and P-Cu1/CuNP. However, the ν(*CO) peak over M-Cu1/CuNP obviously existed from -0.3 to -0.9 V, and the peak intensity stronger than that of R-Cu1/CuNP and P-Cu1/CuNP, suggesting that M-Cu1/CuNP preferred to generate *CO intermediate. Meanwhile, a peak at 1748 cm -1 observed over M-Cu1/CuNP from -0.3 to -0.9 V can be ascribed to the *CHO intermediate, 12, 38, 39 and its change trend was similar to that of *CO intermediate, indicating that the *CHO was originated from the hydrogenation of *CO with the assistance of *H. Nevertheless, the *CHO peak was not observed on R-Cu1/CuNP and P-Cu1/CuNP, meaning that proper content ratio of Cu1 to CuNP is necessary for *CHO formation." Please see them in pages 14-15 of the revised manuscript.
16. DISCUSSION & SUMMARY：The discussion is weak, providing that the assumptions were mostly made on the basis of just one sample. Mixing Cu NPs and Cu atomic sites require good knowledge and verification of both individual systems. The manuscript has started well by presenting three different samples and analyzing the electrocatalytic data of each quite well. However, some more work has to be done on the spectroscopic site to understand the catalytic systems and to extract a relevant conclusion from it.

Response:
We thank the referee again for the comment. As the suggestion of the referee, we have supplemented more data on spectroscopic characterization to further understand the catalytic system and discussed them in the revised manuscript.
For in situ surface-enhanced Raman spectroscopy, the spectra of P-Cu1/CuNP and R-Cu1/CuNP have been added and presented as Figures 3D-3F in the revised manuscript. The peaks associated with *CO intermediate appeared at more positive potential in the spectra of M-Cu1/CuNP than in those of P-Cu1/CuNP spectra, indicating that an increased content ratio of Cu1 to CuNP facilitated CO2 conversion to adsorbed CO. However, the spectra of R-Cu1/CuNP only showed a CO rotation peak within a narrow potential range and no peaks associated with *CO intermediate were observed in the spectra of Cu-N-C ( Figure S26 in the revised supporting information), suggesting that high content of single atomic Cu is not conducive to the conversion of CO2 to adsorbed CO. Therefore, the results of in situ surface-enhanced Raman spectroscopy suggested that CO2-to-C2 products proceeded a CO intermediate and moderate ratio of Cu1 to CuNP facilitated CO2 conversion to adsorbed CO.
For in situ attenuated total reflection-surface-enhanced IR absorption spectroscopy, the spectra of P-Cu1/CuNP and R-Cu1/CuNP have been also added and presented as Figures 3G-3I in the revised manuscript. As the potentials shifted to negative, the peak attributed to *CO in M-Cu1/CuNP spectra appeared earlier than in those of R-Cu1/CuNP and P-Cu1/CuNP. Moreover, M-Cu1/CuNP showed the highest *CO peak intensity among the three catalysts at the same potentials, which implied that moderate content ratio of Cu1 to CuNP facilitated CO2 conversion to adsorbed CO, as confirmed by in situ surface-enhanced Raman spectroscopy results. Additionally, a peak ascribed to *CHO intermediate was observed in the M-Cu1/CuNP spectra, while not in the P-Cu1/CuNP and R-Cu1/CuNP spectra, and its change trend was similar to that of *CO intermediate. These findings suggest that the *CHO intermediate is produced via the hydrogenation of *CO with the assistance of *H, and a proper content ratio of Cu1 to CuNP was necessary for *CHO formation.
For in situ X-ray adsorption spectroscopy, we have reconducted the characterization for M-Cu1/CuNP and displayed the results as Figure 3C in the revised manuscript and Figure S24 in the revised supporting information.
No obvious differences were observed in both XANES and EXAFS, which implied the structural stability of single atomic Cu and Cu nanoparticles.
In the revised manuscript, we have discussed the results of in situ spectroscopic characterization in detail and summarized them in the Discussion section by "In summary, a series of dual-active sites catalysts with co-loaded Cu sites promoted H2O dissociation to provide *H, which in turn migrated to Cu NPs and facilitated *CO protonation to form *CHO by modulating the *H coverage on Cu NPs, leading to high activity for C2 products production. This work puts forward rational concept for promoting conversion CO2 to C2 products through modulate the adsorbed hydrogen coverage on Cu-based catalysts. We believe that it will inspire the design of more dual-/multi-active sites catalysts for multi-step reactions." Please see them in page 18 of the revised manuscript.
17. EXPERIMENTAL：In general, the experimental section lacks several details that are important to follow the experiments. E.g., different types of flow cells exist, a more explicit description should be provided. If a gas-fed CO2 type flow cell was used, please also add the side from which the gas products were determined and how often gas products were injected into the GC (if an online-GC was used). Please provide a scheme or a reference of the

Response:
We thank the referee again for the comment. The scheme of flow cell has been added in the revised supporting information, and gas products were collected using a gas bag from CO2 outlet, then analyzed by using an off-line GC. We have combined the methods in the main manuscript and supporting information into one section and more experiment details have been supplemented in the revised manuscript as "Materials. Electrocatalytic CO2 reduction. To prepared gas diffusion electrode, 1 mg catalyst was suspended in 500 μL isopropanol with 10 μL Nafion D-521 dispersion (5 wt%) to form a homogeneous ink. Then the catalyst ink was spread onto the gas diffusion electrode of 0.5 × 2 cm 2 in area by a micropipette and then dried under room temperature. All the electrochemical experiments were conducted on the electrochemical workstation (CHI 660E) equipped with a high current amplifier (CHI 680C) and the CO2RR performance was investigated in flow cell ( Figure S7). The as-prepared gas diffusion electrode and Ni foam were used as the cathode and anode, respectively.
An anion exchange membrane was used to separate the cathode and anode. Aqueous KOH solution (5 M) was used as the electrolyte solution and the electrolyte volume was 30 mL, and the electrolyte volume in the cell is 0.5 mL.
The catholyte solution and anolyte solution were recirculated by two pumps with flow rates of 10 mL min -1 and 30 mL min -1 , respectively. Meanwhile, CO2 gas was continuously supplied to the gas chamber by using a mass flow controller with a flow rate of 40 mL min -1 . All the potentials were measured against a Hg/HgO reference electrode and converted to versus RHE with iR (80%) compensations, i represents the current obtained at corresponding potential, R is ohmic resistance of the cell measured by electrochemical workstation: The chronopotentiometry method is used for evaluating long-term CO2RR activity stability of electrocatalyst in 5 M KOH electrolyte. In order to address the issues of flooding and carbonation accumulation of GDE, the CO2RR was interrupted every 5 hours and the GDE were removed, washed with deionized water and followed by dryness under N2 atmosphere. Meanwhile, the electrolyte solution was refreshed for each interval.
Product Analysis. The gaseous products were collected using a gas bag and quantified by gas chromatography (GC 7890B). A thermal conductivity detector (TCD) and a flame ionization detector (FID) were used to quantify H2, CO, and other alkane contents, respectively. The Faradaic efficiency (FE) of gaseous products was calculated by the equation: Where z represents the number of electrons transferred for product formation, n is the volume concentration from GC, V is the total volume calculated by outlet flow rate, F is Faraday constant (96485 C/mol) and the Q is the amount of cumulative charge recorded by the electrochemical workstation.
The liquid product was analyzed by 1 H NMR (Bruker Advance III 400 HD spectrometer) in deuteroxide. To accurately integrate the products in NMR analysis, the sodium 2, 2-dimethyl-2-silapentane-5-sulfonate (DSS) was the reference for ethanol and acetic acid, and phenol was the reference for formate. The FE of liquid products was calculated by the equation: Where z represents the number of electrons transferred for product formation, C is the liquid concentration obtained from NMR, V is electrolyte volume, F is Faraday constant (96485 C/mol) and the Q is the amount of cumulative charge recorded by the electrochemical workstation." Please see them in pages 19-20 of the revised manuscript.
The oxygen evolution reaction (OER) occurred in anode chamber of flow cell during CO2RR occurred in cathode chamber. Therefore, the low-cost, earth-abundant, and efficient OER catalyst is necessary for CO2RR investigation. Although IrO2 and Pt anode catalysts are high activity for OER, the scarcity and high cost limit their widespread use. Ni foam has the advantage of low cost, easy accessibility and high OER activity in an alkaline environment, which has been widely used as the anode for OER in flow cell with alkaline electrolyte (Science 2020, 367, 661-666;Nat. Energy 2020, 5, 478-486;Nat. Commun. 2021, 12, 5745;J. Am. Chem. Soc. 2022, 144, 3039-3049  However, operando characterization indicates the operation under genuine reaction conditions, where the catalyst structure (active sites and intermediates), as well as activity and selectivity, is measured simultaneously in real time (Catal. Today, 2005, 100, 71-77;ACS Catal., 2021, 11, 1136-1178. Based on the above definition, we have used in situ to describe XAS, Raman and ATR-SEIRAS characterization in the revised manuscript.

Reviewer 3:
Authors reported the dual-active sites catalyst for electrocatalytic CO2 reduction reaction. The atomic Cu sites and Cu NPs exhibited an efficient CO2-to-C2 reaction. M-Cu1/CuNP delivered a high C2 products FE of 75.4% with a partial current density of 289.2 mA cm -2 and remarkable long-term stability. Based on the experimental and theoretical studies, the boosted H2O dissociation and formed *CHO play the crucial role in the CO2-to-C2 reaction.
However, the description of X-ray and FTIR spectroscopy is not convincing. The reviewer has several questions regarding the mechanistic insights into the proposed reaction model. Please address these issues in detail.

Response:
We thank the referee very much for the comment, which help and guide us to improve and polish our work greatly. We have tried our best to answer the questions from the reviewers. coordination still existed in FT EXAFS spectra ( Figure 3C), with no notable change was observed in peak intensity, suggesting the stability of the content ratio of Cu1 to CuNP. The results of in situ XAS experiments showed that the structure and content of Cu1 and CuNP remained stable during CO2RR." Please see them in page 13-14 of the revised manuscript.
5. Line 193-195 ( Figure 3C), authors claimed that the increase of Cu valence state could be attributed to the adsorption of CO2 and H2O molecular on Cu NPs and Cu sites. It is surprising that Cu valence state increases during CO2RR. This result is not consistent with previous studies. Also, Cu K-edge XAS provides bulk information not surface information. These results and statements are not convincing. If the adsorption of CO2 and H2O does cause the oxidation process, authors should provide a serious of operando XANES spectra obtained from positive potentials to negative potentials. The shifts in the absorption energy of white line position can be therefore discussed.

Response:
We thank the referee again for the comment. We have reconducted in situ XAS experiments carefully and made the data more precise. The results of in situ XAS experiments showed that no obvious difference was observed in both XANES and EXAFS, which implied the structure stability of single atomic Cu and Cu nanoparticle.
There results were in agreement with the previous literature.
In the revised manuscript, we have shown the modified Figure and discussed them by "In situ X-ray absorption spectroscopy (XAS) experiments were performed on M-Cu1/CuNP to investigate the changes of Cu valence state and structure during CO2RR ( Figure S23). In the XANES spectra ( Figure S24), the Cu K-edge adsorption spectra did not show obvious difference under different potentials, indicating that the average Cu valence state kept stable during the reaction. Moreover, the peaks attributed to Cu-N and Cu-Cu coordination still existed in FT EXAFS spectra ( Figure 3C), with no notable changes observed in peak intensity, suggesting the stability of the content ratio of Cu1 to CuNP. The results of in situ XAS experiments showed that the structure and content of Cu1 and CuNP remained stable during CO2RR." Please see them in page 13-14 of the revised manuscript.
Please provide the information.

Response:
We thank the referee again for the comment. The morphology, valence states and crystal structure of M-Cu1/CuNP were characterized after 40 h CO2RR experiment. More characterizations have been adopted to investigate the stability of M-Cu1/CuNP. The TEM image, HRTEM image and aberration-corrected HAADF-STEM image of M-Cu1/CuNP after 40 h electrolysis were added as Figure S15 in the revised supporting information. The Cu LMM Auger spectrum, XPS spectra of N 1s orbits and XRD pattern of M-Cu1/CuNP after 40 h electrolysis were added as Figure S16 in the revised supporting information. In the revised manuscript, we have discussed them as "The morphology, valence states and crystal structure of M-Cu1/CuNP after 40 h CO2RR experiment were characterized ( Figures S15, S16). TEM image suggested the size Cu NPs is similar to that before used and the aberration-corrected HAADF-STEM image confirmed that Cu1 site and Cu NPs still coexisted in M-Cu1/CuNP. The Cu LMM Auger spectra show that Cu + and Cu 0 coexisted in M-Cu1/CuNP. Meanwhile, no obvious difference was observed in N 1s XPS spectra between before and after CO2RR experiment, and only the peaks corresponding to Cu were observed in XRD spectrum after CO2RR experiment. All these results indicated the excellent stability of the catalyst." Please see them in Page 10 of the revised manuscript. Data were recorded in fluorescence excitation mode using a Lytle detector. During the in situ XAS experiments, the spectrum was recorded following the application of a corresponding potential to the electrode for a duration of 600 s. Beamline of Beijing Synchrotron Radiation Facility (BSRF) and the radiation was monochromatized by a Si (111) double-crystal monochromator. All collected spectra were analyzed using Athena and Artemis program within the IFEFFIT software packages.
In situ Raman measurements. A flow cell with a quartz window by GaossUnion (Tianjin) Photoelectric Technology Company was used to carry out in situ Raman measurements using a Horiba LabRAM HR Evolution Raman microscope ( Figure S25). A 785 nm laser was used and signals were recorded using a 20 s integration and by averaging two scans. The gas diffusion electrode sprayed with the catalyst was used as working electrode and a graphite rod and a Hg/HgO electrode were used as counter and reference electrodes, respectively. The counter electrode was separated from the working electrode by anion exchange membrane. The 5 M KOH aqueous solution was used as electrolyte and were recirculated by pump with flow rates of 20 mL min -1 . Meanwhile, CO2 gas was continuously supplied to the gas chamber with a flow rate of 40 mL min -1 .
In situ ATR-SEIRAS measurements. The experiments were conducted in a modified electrochemical cell that integrated into a Nicolet 6700 FTIR spectrometer equipped with MCT detector cooled by liquid nitrogen ( Figure   S27). The catalysts ink was dropped on a germanium ATR crystal deposited with Au film. A platinum wire and a Hg/HgO electrode were used as counter and reference electrodes. Each spectrum was collected with 32 times with a resolution of 4 cm -1 . The CO2-saturated 3 M KOH aqueous solution was used as electrolyte to avoid that the germanium ATR crystal was damaged by strong basic electrolyte. The background spectrum was collected at the potential of +0.1 V vs RHE." Please see them in page 22-23 of the revised manuscript.
The in situ XAS and Raman experiments were performed in flow cell with 5 M KOH aqueous solution as electrolyte. The in situ ATR-SEIRAS experiments were performed in single-chamber electrolytic cell. As the strong basic electrolyte would damage the germanium ATR crystal, 3 M KOH aqueous was used as electrolyte during the in situ ATR-SEIRAS experiments. It is difficult to distinguish that the OH-associated peak is caused by hydroxide ion in electrolyte or formation from water dissolution, because there were abundant hydroxide ions in the alkaline electrolyte, Therefore, the changes in the OH-associated peak did not investigate in the manuscript.
(2) A mass of bubbles would be generated on the catalyst surface under high current, which affected the signal of in situ ATR-SEIRAS. Thus, in order to ensure the signal quality of in situ ATR-SEIRAS experiments, the highest current during the experiments kept a low value (< 4 mA). Meanwhile, the time of collection each spectrum was quick. The pH value of electrolyte changed from 10.21 to 10.15 after in situ ATR-SEIRAS test. Therefore, the effect of electrolyte pH change on CO2RR was almost negligible. Although the pH value of electrolyte in in situ ATR-SEIRAS experiments was different from that in XAS and Raman experiments, the CO2RR conditions were all in alkaline electrolyte, and the same intermediates were involved in CO2RR process during in situ ATR-SEIRAS and Raman experiments. Thereby, it has a good correlation between the in situ ATR-SEIRAS and Raman/XAS results, which has gradually developed into useful techniques for the mechanism analysis for CO2RR (Nat. Catal. 2020, 3, 75-82;Nat. Catal. 2018, 1, 922-934;J. Am. Chem. Soc. 2022, 144, 2079-2084J. Am. Chem. Soc. 2020, 142, 9567-9581;).
8. Cu will be oxidized in the alkaline solution. Why did not XAS ( Figure 3C)/Raman ( Figure 3D) results show the formation of CuO/Cu(+2)?
Response: We thank the referee again for the comment. Previous works have reported that Cu 2+ species can be easily reduced to Cu 0 or Cu + species when reduce potentials were applied on catalysts (J. Am. Chem. Soc. 2019, 141, 6986−6994;J. Am. Chem. Soc. 2022, 144, 2079−2084Nat. Commun. 2022, 13, 4857). Therefore, the average Cu oxidation state was between 0 and +1 from in situ XAS analysis, and thus CuO/Cu 2+ could not be formed under the electrolysis conditions. 9. Since the S/N ratio of CHO-associated peak is poor in figure 3F, the role of CHO in the CO2 reduction mechanism is questionable. To prove the proposed reaction model, authors should provide SEIRAS of P-Cu1/CuNP and R-Cu1/CuNP materials. The potential dependence of CHO-associated peak (at 1548 cm -1 ) in different materials should be further discussed.

Response:
We thank the referee again for the comment. According to the comment, we reconducted the in situ ATR-SEIRAS experiments over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. The peak of *CO existed on all the three catalysts. However, as the potentials moved to negative, the *CO peak over M-Cu1/CuNP appeared earlier than that of R-Cu1/CuNP and P-Cu1/CuNP, and had a stronger intensity. It indicated that M-Cu1/CuNP facilitated CO2 conversion to adsorbed *CO. On the other hand, the peak assigned *CHO only appeared over M-Cu1/CuNP, and its change trend was similar to that of *CO intermediate, indicating that the *CHO was originated from the hydrogenation of *CO with the assistance of *H. The results suggested that proper ratio of Cu1/CuNP is necessary for *CHO formation.
In the revised manuscript, we have discussed them by "In situ attenuated total reflection-surface-enhanced IR absorption spectroscopy (ATR-SEIRAS) spectra were further collected to trace the evolution of reaction intermediates during CO2RR from -0.2 to -0.9 V over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP ( Figures 3G-3I, Figure S27). The peak at around 2100 cm -1 can be attributed to electrogenerated CO adsorbed (*CO) on catalyst surface, 39, 40 which did not appear until the applied potential reached -0.5 V on R-Cu1/CuNP and P-Cu1/CuNP.
However, the ν(*CO) peak over M-Cu1/CuNP obviously existed from -0.3 to -0.9 V and the peak intensity was stronger than that of R-Cu1/CuNP and P-Cu1/CuNP, suggesting that M-Cu1/CuNP preferred to generate *CO intermediate. Meanwhile, a peak at 1748 cm -1 observed over M-Cu1/CuNP from -0.3 to -0.9 V can be ascribed to the *CHO intermediate, 12,38,39 and its change trend was similar to that of *CO intermediate, indicating that the *CHO was originated from the hydrogenation of *CO with the assistance of *H. Nevertheless, the *CHO peak was not observed on R-Cu1/CuNP and P-Cu1/CuNP, meaning that proper content ratio of Cu1 to CuNP is necessary for *CHO formation." Please see them in Page 14-15 of the revised manuscript.

Reviewer 4:
In this work, the authors report the design of catalysts for electrochemical CO2 reduction to C2+ products based on a dual-active site strategy. They combine a water dissociation catalyst (single atom Cu) and a C-C coupling one (Cu nanoparticles) to produce Cu nanoparticles supported on Cu-N-C single atom matrix. They found that this combination enhances CO2 reduction to C2+, achieving a C2 Faradaic efficiency of 75% and a partial current density of 289 mA with a total FE at -0.6 V vs RHE. While the fabrication of the dual catalysts could be interesting, I do not find the designed catalysts represent a significant advance in the field of electrochemical CO2 conversion. My conclusion comes from both fundamental and applied perspectives as I explain below: Response: We thank the referee very much for the comment, which help and guide us to think about and polish our work greatly. We have tried our best to answer the questions from the reviewers. We hope that we have addressed all of the questions satisfactorily. Many thanks for your time and efforts. We believe our manuscript meets the high standard of the journal, and also hope the reviewer could agree with us.
1. The electrochemical CO2 conversion performance of the designed catalysts in this work is comparable to most of normal Cu nanoparticles catalysts. In a basic electrolyte, C2 selectivity of 70-80% in the current density range of 200-500 mA/cm 2 has been frequently reported using Cu nanoparticle catalysts. A good catalyst for C2 production should exhibit an FE of at least over 80%.

Response:
We thank the referee again for the comment. Although 70%-80% C2 products selectivity with 200-500 mA cm -2 C2 products partial current density has been reported in some papers, the performance was always obtained at high applied potential.  Int. Ed. 2021, 61, e202113498). Obtaining good performance at low applied potential is significant for CO2RR. In this work, the prepared M-Cu1/CuNP exhibited a C2 products Faradaic efficiency (FEC2) of 75.4% with partial current density of C2 products of 289.2 mA cm -2 at low applied potential (-0.6 V), which lower than the reported electrocatalysts for CO2RR to C2+ products in literature because the excellent synergistic effect between Cu1 and CuNP. In the revised manuscript, we have discussed them by "Compared with reported electrocatalysts for CO2RR to C2+ products in the literature, M-Cu1/CuNP exhibited higher CO2RR to C2+ products activity at low applied potential, especially in term of C2+ partial current density (Table S7)." Please see them in page 10 of the revised manuscript.
More importantly, this work has shown a dual-active sites catalyst model and proved that inducing an active site for generating moderate *H can efficiently improve the performance for CO2RR through modifying *H coverage.
In order to analyze the get insight into catalytic mechanism in detailed, a series of additional characterizations and experiments, including HAADF-STEM, XPS, in situ ATR-SEIRAS, Raman, XAS, etc., have been supplemented in the revised manuscript. We believe that more catalysts with higher activity for CO2RR would be fabricated in the future according this catalyst model, through further optimizing the activity of H2O dissociation site and CO2 reduction site.
2. The high performance in this work was achieved with very basic electrolyte (5M KOH). Many recent analyses have pointed out the cost related to carbonate formation with alkaline electrolyte and suggest that this condition is not suitable for practical application. In this work, the author designed dual-active site catalysts that work well only in strong alkaline electrolyte ( Figure S15). Therefore, it is unclear how the designed catalysts can be used in practical systems.

Response:
We thank the referee again for the comment. Alkaline aqueous solution is often used as electrolyte for producing C2+ products, which can inhibit hydrogen evolution reaction and facilitate C-C coupling reaction (Science 2018, 360, 783-787;Science 2020, 367, 661-666;Nat. Catal. 2020, 3, 98-106). Carbonate formation is a common problem in alkaline electrolyte and neutral electrolyte. Although acid electrolyte can efficiently avoid carbonate formation, the existing equipment corrosion, metal-based catalyst instability problems also make it not suitable for practical application. Actually, the practical application of CO2RR is a complex systematic problem, involving catalyst, electrolyte, ion exchange membrane, cell structure, mass transfer, et al. Recently, some works have been reported to try to solve these problems. It can be anticipated that the problem of carbonate formation in basic electrolyte would be solved through other methods and the practical application of CO2RR would eventually be realized.
Although the CO2RR activity in low concentration KOH electrolyte was lower than that in 5 M KOH electrolyte, the C2 products selectivity of M-Cu1/CuNP was still higher than that of R-Cu1/CuNP and P-Cu1/CuNP, suggesting that the cooperative effect of Cu NPs and atomic Cu sites still facilitated C2 products formation. The dual-active site catalyst model that we proposed was still suitable for low concentration KOH electrolyte. In the revised manuscript, we have discussed them by "Furthermore, we studied the influence of electrolyte pH on CO2RR performance over R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP. Three different concentrations of KOH aqueous electrolytes, i.e., 0.1M, 3M and 5M, were employed to adjust the pH environment at the electrode/electrolyte interface. The results in Figure   S18 show that the FE and formation rate of C2 products increased with the increasing pH value of electrolyte, suggesting that strong basic local environment favored the C2 products formation. Although CO2RR activity of M-Cu1/CuNP in low concentration KOH electrolytes was lower than that in 5 M KOH electrolyte, M-Cu1/CuNP exhibited higher FE and formation rate of C2 products than P-Cu1/CuNP and R-Cu1/CuNP in low concentration KOH electrolytes, which suggested that the cooperative effect of Cu NPs and atomic Cu sites still facilitated C2 products formation.
The ratio of C2 products formation rate over M-Cu1/CuNP to P-Cu1/CuNP, i.e., RateM/RateP, in 5 M KOH (3.4) is higher than those in 3M KOH (3.0) and 0.1 M KOH (2.3). It demonstrated that the role of atomic Cu sites in accelerating the H2O dissociation process was significantly more pronounced at higher pH values, leading to the enhanced C2 products formation, even if the dissociation of H2O in higher pH electrolyte was a sluggish step." Please see them in page 12-13 of the revised manuscript.
3. The mechanism for enhanced C2+ production based on dual-active sites is rather speculative as there are not enough experimental and theoretical evidence to support that claim. For example, the authors claim that Cu-N-C matrix can enhance water dissociation to provide protons for C2 formation, but they did not provide that Cu element existed in Cu-N-C catalyst and Cu + was main specie. More importantly, the EXAFS profiles in the R-space of Cu-N-C suggested that only the peak attributed to Cu-N coordination located at around 1.4 Å, while the Cu-Cu coordination peak at around 2.2 Å disappeared, which confirmed that the Cu species only existed as single atom form. The results of least-squares EXAFS fitting confirmed that the Cu-N coordination numbers in Cu-N-C was 4.0 (Table S2), implying that the atomic Cu species mainly existed as Cu-N4 structure." Please see them in the figure description of Figure S11 and Figure S12 in the revised supporting information. The activity test of the Cu-N-C catalyst showed that H2 was main products in the whole applied potentials range, and H2O dissociation provided proton source in basic electrolyte, thereby the Cu-N-C catalyst enhanced H2O dissociation even CO2 presented in the reaction system.
Additionally, the kinetic isotope effect (KIE) test is an efficient method to justify whether H2O dissociation is involved in the rate-determining step during CO2 protonation process or not (J. Am. Chem. Soc. 2022, 144, 12807-12815;Nat. Commun. 2019, 12, 892;Adv. Mater. 2022, 34, 2200057). We have performed KIE test over all three catalysts and displayed the results in Figure 3A in the revised manuscript. If the KIE value (defined as the ratio of C2H4 formation rates in H2O and D2O) closes to 1, H2O dissociation is not the rate-determining step over the catalyst.
The KIE value of P-Cu1/CuNP was calculated to be 2.87, demonstrating that proton transfer was limited. When the Cu single atoms content was increased, the KIE value over M-Cu1/CuNP and R-Cu1/CuNP sharply dropped to 1.21 and 1.12 respectively, indicating that H2O dissociation was accelerated, which is not involved in the rate-determining step. Therefore, the activity test and KIE results revealed that Cu single atoms were responsible for the dissociation of water to accelerate the proton transfer process. In the revised manuscript, we have discussed them by "The kinetic isotope effect (KIE) of H2O/D2O (H/D) experiments were performed to get insights into the role of H2O dissociation in CO2-to-C2 products ( Figure 3A). When the H2O was replaced by D2O in the electrolyte, the formation rate of the product (for example C2H4) decreased over P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP, and the level of its decrease was closely related to the content ratio of Cu1/CuNP. If the KIE value (defined as the ratio of C2H4 formation rates in H2O and D2O) closes to 1, H2O dissociation is not the rate-determining step over the catalyst. The KIE value for R-Cu1/CuNP, M-Cu1/CuNP and P-Cu1/CuNP were 1.12, 1.21 and 2.87, respectively, which indicated that H2O dissociation was accelerated gradually as the increasing of Cu1/CuNP content ratio. These results confirmed that the atomic Cu sites were responsible for accelerating H2O dissociation and provided proton to adjacent Cu NPs, thus affecting CO2-to-C2 products. Meanwhile, the N-doped carbon matrix catalysts has been reported to favor the migration of proton. 30 " Please see them in page 12 of the revised manuscript.
Although the OH-associated peak can be observed on in situ Raman spectra, considering that abundant hydroxide ion existed in alkaline electrolyte, it is difficult to distinguish that the OH-associated peak is caused by hydroxide ion in electrolyte or formation from water dissolution. Therefore, the changes in the OHdid not investigate in the manuscript. H + is not stable in the basic environment, which is easily neutralized by OHto generate H2O again, so proton would be adsorbed on the surface of catalyst in the form of *H. It has been reported that hydrogen spillover can be occurred on N-doped carbon matrix (Adv. Mater. 2022, 34, 2200057), and doping N atoms into carbon matrix would induce C atoms possess positive charge, the positively charged C atoms could theoretically enhance the proton transfer through the repulsive interaction between the generated protons and the positively charged C atoms. Therefore, the proton is transported to adjacent Cu nanoparticles mainly via hydrogen spillover as  Figure S24 in the revised supporting information. The Cu K-edge XANES of M-Cu1/CuNP under different potentials still located between 0 and +2 and no obvious difference was observed, suggesting that the Cu average valence state kept stable during CO2RR. Moreover, the peaks attributed to Cu-N and Cu-Cu coordination still existed in FT EXAFS ( Figure 3C in the revised manuscript), with no notable changes observed in peak intensity, indicating the stability of the content ratio of Cu1 to CuNP during CO2RR. These results suggested that the structure and content of Cu1 and CuNP remained stable during CO2RR.
In the revised manuscript, we have discussed them by "In situ X-ray absorption spectroscopy (XAS) experiments were performed on M-Cu1/CuNP to investigate the changes of Cu valence state and structure during CO2RR ( Figure S23). In the XANES spectra ( Figure S24), the Cu K-edge adsorption spectra did not show obvious difference under different potentials, indicating that the average Cu valence state kept stable during the reaction.
Moreover, the peaks attributed to Cu-N and Cu-Cu coordination still existed in FT EXAFS spectra ( Figure 3C), with no notable change was observed in peak intensity, suggesting the stability of the content ratio of Cu1 to CuNP. The results of in situ XAS experiments showed that the structure and content of Cu1 and CuNP remained stable during CO2RR." Please see them in page 13-14 of the revised manuscript.
5. Additional controlled samples should be tested to confirm the proposed mechanism. For example, mixing Cu-N-C with a commercially available Cu nanoparticles or other metals such as Ag or Au to see if this mechanism is universal (i.e., for CO2 conversion to other products).

Response:
We thank the referee again for the comment. It has been demonstrated that an optimal distribution and distance between the dual active sites is crucial for efficient active species (*CO or *H) transformation, which significantly impacted the synergistic effect of dual active sites. According to the comment, more controlled samples have been used in the CO2RR to confirm the proposed mechanism.
The size of Cu nanoparticles in our fabricated dual-active site catalyst was around 4 nm. To exclude the size effect, we have prepared Cu and Ag nanoparticles with around 4 nm diameter according to the literature (Nanoscale 2021, 13, 4835-4844), and measured CO2 activity of pure Cu nanoparticles and the catalyst fabricated via physically mixing the Cu/Ag nanoparticles with the Cu-N-C. After the Cu nanoparticles were mixed with the Cu-N-C, C2+ products selectivity decreased from 32.4% to 19.6%, while that of hydrogen increased from 42.5% to 63.7%.
Meanwhile, mixing Ag nanoparticles with Cu-N-C resulted in decreased CO selectivity from 88.2% to 64.1% and an obvious enhancement in hydrogen selectivity. The dual-active site catalysts fabricated through physical mixing Cu-N-C with Cu/Ag nanoparticles were highly heterogeneous, with a considerable distance between the dual active sites that impeded the efficient transformation of *H species, which are dimerized to produce hydrogen. However, the distribution and distance between the dual active sites of the catalysts synthesized using one-step chemical methods can be effectively modified by altering the ratio of precursors used in the synthesis process, thus obtaining highly effective catalytic performance. Therefore, the mechanism proposed in the manuscript is only applicable to catalysts synthesized using chemical methods with optimal distribution and distance between the dual active sites.

REVIEWER COMMENTS</B>
Reviewer #1 (Remarks to the Author): The reviewer appreciate the efforts from the authors to address the concerns and questions. The quality of manuscript has greatly improved. Thus, the reviewer recommend its publication in Nature Communications as it is.
Reviewer #2 (Remarks to the Author): The authors revised their manuscript thoroughly by complementing their presented experimental data with the missing ones. To the reviewer's opinion, data quality, comprehensiveness and completeness has improved significantly and the datasets are now complete to allow any interpretation. Furthermore, the authors have addressed questions and concerns of the reviewers with a lot of detail.
Looking at the complete data sets, the reviewer is concerned that the improved catalytic activity of the M-sample is due to higher electrocatalytic surface area and the majority of the discussion and interpretation could thus be invalid. For example, the signal-to-noise ratio for the Raman data seems to be much lower for the M-sample than the P-sample. As Cu is active for surface enhancement, this might indicate a significant increase in surface area. Also, the CO-stripping voltammetry for the M-sample provides double of the current density compared to the other samples (and compared to the LSV after CO stripping in the supplementary information).
To study the sample loading-dependent catalytic selectivities is a creative way to circumvent the analysis of ECSA, it is however not convincing. First, the optimal loading for the M-sample was studied, but what happened with the other catalysts, where they adapted to the same loading? Unfortunately, it is best practice in the electrocatalysis community to analyse the ECSA method in order to avoid wrong interpretation of the catalytic activity. The reviewer expects a reasonable correction on this topic.
In the following, there are some minor comments: - Figure  Reviewer #3 (Remarks to the Author): Although most of the minor questions have been addressed properly, the reviewer has quite substantial concerns in the mechanism study in the present study. The reviewer can not recommend the present study for further publication. These serious concerns should be further addressed.
It is agreed that the CO2 reduction mechanism and electrochemical performance are different in the Htype cell and flow cell. In the flow cell, the CO2 reduction reaction is carried out at GDE, which includes the CO2 (gas)/electrocatalyst (solid)/electrolyte (liquid) reaction. It is a three-phase reaction. In the Hcell, CO2 gas is dissolved in the electrolyte and forms bicarbonate/carbonate anions, which affect the pH of electrolyte during cycling. Also, mass diffusion is another problem. Thus, the product distribution and electrochemical performance are different in the H-type cell and flow cell, suggesting that the reaction mechanism is different in the H-type cell and flow cell. That is why the authors prepared GDE and perform in situ XAS/Raman spectroscopy in the flow cell in the present study ( Figure S23 and figure S25). However, the authors performed in situ FTIR (SEIRAS) in the H-type cell and claimed that the results (mechanistic insights) obtained in the H-type cell and flow cell should be similar (for example: the correlation of SEIRAS/Raman/XAS, the negligible changes in the pH value). In fact, the pH of electrolyte in the SEIRAS (pH= ~10) is not the same with that in the Raman/XAS studies (concentrated KOH). The presence of anions is different in these cells. Moreover, a three-phase reaction is carried out in the flow cell with good mass diffusion. It is not about the different cells. It is about the cell configuration that affects the reaction mechanism.
If the authors insist that the results (mechanistic insights) obtained in the H-type cell and flow cell should be similar. The authors should perform the in situ Raman/XAS measurements/electrochemical performance in the "H-type cell" and provide a detailed comparison. To have a huge impact in the research community, authors should provide the mechanism insights precisely.
The reviewer understands that in situ SEIRAS measurement is available with the H-type cell. Since both Raman and SEIRAS measurements are used to study the formation of surface-adsorbed intermediates such as CO during CO2 reduction reaction, do in situ SEIRAS results really provide crucial/additional mechanistic insight? Compared to the in situ Raman results, in situ SEIRAS also shows the formation of CHO species during CO2 reduction reaction. Can this species be obtained in the Raman or ex-situ FTIR measurements (samples are prepared in the flow cell)? If so, the authors can consider to remove the SEIRAS results in the present study.
Reviewer #4 (Remarks to the Author): The author have addressed most of my concerns.
I still believe that strong basic (KOH) electrolyte is not a suitalbe electrolyte to evaluate the intrinsic activiity and selectiivty of the catalysts for CO2 conversion to multiple carbon products. As the authors shown in talble S7, most of the catalysts tested using alkaline electrolyte show C2+ selectivity in the range of 75-90%, which is comparable and higher than the perormance reported in this work. Even commercial Cu nanoparticles show high C2+ selectivity with alkaline electrolyte. Based on the performance of the catalysts in strong alkaline electrolyte, I am not convinced that this represents a significnat advance in the field.

Reviewer 1:
The reviewer appreciates the efforts from the authors to address the concerns and questions. The quality of manuscript has greatly improved. Thus, the reviewer recommends its publication in Nature Communications as it is.

Response:
We thank the reviewer very much for the positive comment.

Reviewer 2:
The authors revised their manuscript thoroughly by complementing their presented experimental data with the missing ones. To the reviewer's opinion, data quality, comprehensiveness and completeness has improved significantly and the datasets are now complete to allow any interpretation. Furthermore, the authors have addressed questions and concerns of the reviewers with a lot of detail.

Response:
We thank the referee very much for the comment, and we have carefully revised the manuscript based on the comments.
1. Looking at the complete data sets, the reviewer is concerned that the improved catalytic activity of the M-sample is due to higher electrocatalytic surface area and the majority of the discussion and interpretation could thus be invalid. For example, the signal-to-noise ratio for the Raman data seems to be much lower for the M-sample than the P-sample. As Cu is active for surface enhancement, this might indicate a significant increase in surface area.
Also, the CO-stripping voltammetry for the M-sample provides double of the current density compared to the other samples (and compared to the LSV after CO stripping in the supplementary information). To study the sample loading-dependent catalytic selectivity is a creative way to circumvent the analysis of ECSA, it is however not convincing. First, the optimal loading for the M-sample was studied, but what happened with the other catalysts, where they adapted to the same loading? Unfortunately, it is best practice in the electrocatalysis community to analyses the ECSA method in order to avoid wrong interpretation of the catalytic activity. The reviewer expects a reasonable correction on this topic.

Response:
We thank the referee again for the comment. The electrochemical active surface area (ECSA) is proportional to double layer capacitance, which measures the capacitive current associated with double-layer charging from the scan-rate dependence of cyclic voltammogram. As suggested by the referee, the ECSA values of the P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP have been added as Figure S14 in the revised supporting information, and the results suggested that M-Cu1/CuNP exhibited the largest ECSA, consistent with the phenomenon of COstripping voltammetry experiments. To exclude the influence of ECSA on catalytic activity, we further normalized the jC2 based on the ECSA. The results suggested that M-Cu1/CuNP still exhibited the highest normalized jC2 among the catalysts, which were displayed in Figure S15 of revised supporting information. Thus, the intrinsic catalytic activity of M-Cu1/CuNP is higher than that of P-Cu1/CuNP and R-Cu1/CuNP.
The catalytic activity and ECSA were tested under the optimal loading of P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP, respectively. To further eliminate the potential impact of differences in ECSA on catalytic activity, we increased the loading of P-Cu1/CuNP and R-Cu1/CuNP in accordance with the ECSA results, ensuring the catalytic activity were measured under similar ECSA conditions. The results were displayed in Figure S16 of the revised supporting information and indicated that M-Cu1/CuNP still exhibited largest C2 products selectivity and partial current density, further confirming that the intrinsic catalytic performance of M-Cu1/CuNP is better than that of both P-Cu1/CuNP and R-Cu1/CuNP. In the revised manuscript, we have discussed them by "We further normalized the jC2 on the basis of the electrochemical active surface area (ECSA), which was measured by the double-layer capacitance (Cdl) method ( Figure S14). M-Cu1/CuNP still exhibited the highest normalized jC2 among the catalysts ( Figure S15), confirming that the intrinsic catalytic activity of M-Cu1/CuNP is higher than that of P-Cu1/CuNP and R-Cu1/CuNP. Moreover, we also increased the loading of P-Cu1/CuNP and R-Cu1/CuNP in accordance with the ECSA results, ensuring the catalytic activity were measured under similar ECSA condition. The results suggested that M-Cu1/CuNP still exhibited largest C2 products selectivity and partial current density ( Figure S16), further confirming that the intrinsic catalytic activity of M-Cu1/CuNP exceeds that of both P-Cu1/CuNP and R-Cu1/CuNP." Please see them in page 10 of the revised manuscript.
We have added the experimental details of double layer capacitance measurements in the Methods section of the revised manuscript as "The electrochemical active surface area is proportional to double layer capacitance, which measures the capacitive current associated with double-layer charging from the scan-rate dependence of cyclic voltammogram. The double layer capacitance was determined in a single-compartment electrolytic cell with 0.5 M KHCO3 aqueous solution as electrolyte. Ag/AgCl electrode and graphite rod were used as reference electrode and counter electrode, respectively. The scan rates of cyclic voltammogram were 20, 40, 60, 80, 100 mV s -1 ." Please see them in pages 21-22 of the revised manuscript.
2. In the following, there are some minor comments: -  Figure S33 of the revised supporting information. Additionally, the content ratios in L121/122 were reversed and we have modified them as "Thus, the content ratio of Cu1 to CuNP of P-Cu1/CuNP, M-Cu1/CuNP and R-Cu1/CuNP were 0.05, 0.25 and 0.39, respectively, which is consistent with our catalyst design expectation." Please see them in Page 6 of the revised manuscript.

Reviewer 3:
Although most of the minor questions have been addressed properly, the reviewer has quite substantial concerns in the mechanism study in the present study. The reviewer cannot recommend the present study for further publication.
These serious concerns should be further addressed.
It is agreed that the CO2 reduction mechanism and electrochemical performance are different in the H-type cell and flow cell. In the flow cell, the CO2 reduction reaction is carried out at GDE, which includes the CO2 (gas)/electrocatalyst (solid)/electrolyte (liquid) reaction. It is a three-phase reaction. In the H-cell, CO2 gas is dissolved in the electrolyte and forms bicarbonate/carbonate anions, which affect the pH of electrolyte during cycling.
Also, mass diffusion is another problem. Thus, the product distribution and electrochemical performance are different in the H-type cell and flow cell, suggesting that the reaction mechanism is different in the H-type cell and flow cell. That is why the authors prepared GDE and perform in situ XAS/Raman spectroscopy in the flow cell in the present study ( Figure S23 and figure S25). However, the authors performed in situ FTIR (SEIRAS) in the Htype cell and claimed that the results (mechanistic insights) obtained in the H-type cell and flow cell should be similar (for example: the correlation of SEIRAS/Raman/XAS, the negligible changes in the pH value). In fact, the pH of electrolyte in the SEIRAS (pH= ~10) is not the same with that in the Raman/XAS studies (concentrated KOH).
The presence of anions is different in these cells. Moreover, a three-phase reaction is carried out in the flow cell with good mass diffusion. It is not about the different cells. It is about the cell configuration that affects the reaction mechanism.
If the authors insist that the results (mechanistic insights) obtained in the H-type cell and flow cell should be similar. The authors should perform the in situ Raman/XAS measurements/electrochemical performance in the "Htype cell" and provide a detailed comparison. To have a huge impact in the research community, authors should provide the mechanism insights precisely.
The reviewer understands that in situ SEIRAS measurement is available with the H-type cell. Since both Raman and SEIRAS measurements are used to study the formation of surface-adsorbed intermediates such as CO during CO2 reduction reaction, do in situ SEIRAS results really provide crucial/additional mechanistic insight? Compared to the in situ Raman results, in situ SEIRAS also shows the formation of CHO species during CO2 reduction reaction.
Can this species be obtained in the Raman or ex-situ FTIR measurements (samples are prepared in the flow cell)? If so, the authors can consider to remove the SEIRAS results in the present study.

Response:
We thank the referee very much for the comment and agree with reviewer's opinion. Due to the inherent limitations of the test method, in situ SEIRAS can only be conducted in the H-type cell in electrocatalytic field.
Therefore, in situ SEIRAS cannot directly verify the reaction mechanism in flow cell. In order to elucidate the mechanism with higher precision, we have performed online differential electrochemical mass spectrometry (DEMS) to monitor the intermediates during the reaction process. To maintain experimental conditions closely resembling those of the actual reaction, the online DEMS experiments were conducted in a flow cell equipped with a gas diffusion electrode and using 5 M KOH as the electrolyte. The scheme of the cell used for online DEMS is displayed in Figure S30 of the revised supporting information. As depicted in Figure S31 of the revised supporting information, the m/z signal of 29, corresponding to CHO, can be detected during five continuous cycles at -0.6 V vs. RHE for M-Cu1/CuNP, while it was absent in both P-Cu1/CuNP and R-Cu1/CuNP. The results of online DEMS demonstrated that the proper content ratio of Cu1 to CuNP is necessary for CHO formation. We have discussed them in the revised manuscript as "The online differential electrochemical mass spectrometry (DEMS) was conducted to further detected intermediate ( Figure S30). During five continuous cycles at -0.6 V vs. RHE, the m/z signal of 29 that correspond to CHO could be detected over M-Cu1/CuNP, while it was not detected over P-Cu1/CuNP and R-Cu1/CuNP ( Figure S31). The results of online DEMS demonstrated that the crucial role of the proper Cu1 to CuNP content ratio in CHO formation." Please see them in page 15 of the revised manuscript.

Reviewer 4:
The author has addressed most of my concerns. I still believe that strong basic (KOH) electrolyte is not a suitable electrolyte to evaluate the intrinsic activity and selectivity of the catalysts for CO2 conversion to multiple carbon products. As the authors shown in table S7, most of the catalysts tested using alkaline electrolyte show C2+ selectivity in the range of 75-90%, which is comparable and higher than the performance reported in this work. Even commercial Cu nanoparticles show high C2+ selectivity with alkaline electrolyte. Based on the performance of the catalysts in strong alkaline electrolyte, I am not convinced that this represents a significant advance in the field.

Response:
We thank the referee very much for the comment. We would like to emphasize the significance and novelty of our work briefly. It is known that design of novel catalysts and explore the reaction mechanism for CO2RR is a very interesting topic, and many important issues need to be studied. In this study, we present a novel dual-active site catalyst model comprising of atomic Cu sites and Cu NPs, and have proved that the substantial enhancement of CO2RR performance through the synergistic effect between atomic Cu sites and Cu NPs. Meanwhile, the synergistic effect of atomic Cu sites and Cu NPs is also suitable for low concentration KOH electrolyte. We have combined experimental and theoretical studies to explore the reaction mechanism. It was revealed that Cu NPs facilitated the C-C coupling step through *CHO dimerization reaction to produce C2 products, while the atomic Cu sites boosted H2O dissociation to form *H. The generated *H migrated to Cu NPs and modulated the *H coverage on Cu NPs, and thus promoted *CO protonation pathway toward *CHO. Therefore, the synergistic effect of atomic Cu sites and Cu NPs resulted in the excellent performance of the catalyst. We believe that the catalyst model and mechanism will inspire the design of more dual-/multi-active sites catalysts for multi-step reactions.
As above, we believe our manuscript meets the high standard of the journal, and also hope the reviewer could agree with us.