Construction of 3D copper-chitosan-gas diffusion layer electrode for highly efficient CO2 electrolysis to C2+ alcohols

High-rate electrolysis of CO2 to C2+ alcohols is of particular interest, but the performance remains far from the desired values to be economically feasible. Coupling gas diffusion electrode (GDE) and 3D nanostructured catalysts may improve the efficiency in a flow cell of CO2 electrolysis. Herein, we propose a route to prepare 3D Cu-chitosan (CS)-GDL electrode. The CS acts as a “transition layer” between Cu catalyst and the GDL. The highly interconnected network induces growth of 3D Cu film, and the as-prepared integrated structure facilitates rapid electrons transport and mitigates mass diffusion limitations in the electrolysis. At optimum conditions, the C2+ Faradaic efficiency (FE) can reach 88.2% with a current density (geometrically normalized) as high as 900 mA cm−2 at the potential of −0.87 V vs. reversible hydrogen electrode (RHE), of which the C2+ alcohols selectivity is 51.4% with a partial current density of 462.6 mA cm−2, which is very efficient for C2+ alcohols production. Experimental and theoretical study indicates that CS induces growth of 3D hexagonal prismatic Cu microrods with abundant Cu (111)/Cu (200) crystal faces, which are favorable for the alcohol pathway. Our work represents a novel example to design efficient GDEs for electrocatalytic CO2 reduction (CO2RR).

normalised current density represents the CO2RR activity per unit of electroactive area (also referred to as "intrinsic activity"). A larger electrochemical surface area corresponds to a lower current density, hence lower CO2RR activity per unit area. For example, assuming that the values in Figure S19 are geometric current densities (please specify), the relative values of the LSV curves change when ESCA-normalised currents are used. In particular in Figure S19, the geometric current density of Cu-CS-GDL and Cu/CS-GDL are about -0.6 and -0.5 A/cm2, respectively, at -0.6V. However, the ESCA of Cu-CS-GDL should be about 5 times larger than that of Cu/CS-GDL (based on the ratio of Cdl values 12.0:2.5, from Figure S26e), it follows that the ESCA-normalised current density of Cu-CS-GDL is significantly lower than that of Cu/CS-GDL, meaning that the intrinsic activity of Cu-CS-GDL is less than that of Cu/cs-GDL. This is a critical point that needs to be clarified in the manuscript, for all electrocatalysts using their corresponding ECSA-normalised current densities.
In all cases, please specify which current densities are geometric or ECSA normalised.
In all cases, please use either "Cu-CS-GDL" or "3D Cu-CS-GDL" but not both otherwise there is uncertainty regarding being either the same GDE or two different ones.

Supporting information
In addition to the experimental details given in the manuscript, please provide a detailed description (possibly with photos) of the flow cell reactor used for CO2 electrolysis. In particular, provide the dimensions of the CO2RR GDE, and which ion transport membrane was used. This study reports a 3D Cu-chitosan (CS)-GDL electrode for CO2 reduction. The authors claimed that the CS acts as a transition layer between catalyst and the GDL which can facilitate rapid electron transport and mitigate mass diffusion limitations. This design resulted in a C2+ Faradaic efficiency (FE) of 88.2% at 900 mA/cm2 with a C2+ alcohols selectivity of 51.4% and a notable partial current density of 462.6 mA/cm2. I see value in the work. However, the following must be addressed for the work to be suitable for Nature Communications.
Major comments: 1-Adding a polymer layer on top of GDL can significantly decrease the electrical conductivity and impeding the electron transfer between the GDL and catalyst. More experiment/characterization and explanation are needed to support the authors' claim: "the CS acts as a transition layer between catalyst and the GDL which can facilitate rapid electrons transport and mitigates mass diffusion limitations".
2-It is not clear the extent to which this layer differs -in make-up and function -from the many polymer, ionomer, binder etc. layers that are commonly applied in this field.
3-A high EtOH selectivity at high current densities was achieved in a flow cell. Achieving record performance in membrane electrode assembly (MEA) would support the claim that this is a broadly applicable approach and one that advances the field toward practical application.
4-This study was performed in an alkaline flow cell which leads to a severe carbonate formation, and the many associated issues documented well in this journal by the perspective by Kanan. How can the authors address the Kanan's concerns?
5-Most of the commercially available GDLs are hydrophobic. The reported contact angle in this study (95.1) is less than some previous reports. More explanation is needed to support why Cuchitosan (CS)-GDL mitigates mass diffusion limitations and improves the CO2 reduction performance at higher current densities.
6-Cu NPs-GDL electrode shows a good performance at higher current densities (600-1000 mA/cm2). Is this related to the synthesis approach? A comparison between commercial CuNP and the Cu NPs-GDL is needed. 7-A deeper mechanistic investigation is needed to support the high EtOH FE. For example, Raman analysis can be compared to these works to further investigate the reaction pathways: J. Am. Chem. Soc. 2019, 141, 21, 8584, Nat Catal 3, 75-82 (2020), J. Mater. Chem. A, 2022 8-XAS analysis shows some interesting results. I do recommend revisiting the results and combining them with the mechanistic study in the main text. 9-XRD needs to be revised, I suggest removing graphitic peaks and focus on the Cu peaks. If the interface of Cu (200) and Cu (111) is very important in EtOH pathway and production, this needs to be experimentally shown.
Minor 1-The authors claimed that "Unfortunately, construction of 3D nanostructured catalysts on porous gas diffusion layer (GDL) is very difficult". Spraying a mixture of a catalyst and ionomer (which is very common) leads to a 3D structure on GDL. Therefore, making a 3D structure is not challenging.
2-In some case, the total FE is higher than 100%, indicating an error.

Response:
We thank the reviewer very much for positive and valuable comment. We have addressed all the concerns, including those about ECSA, as can be known from the answers to the following detailed comments.
Comment 1: Line 122: The statement "which was not only stabilized the hydrophobic surface of the GDL" is not clear. Do the authors mean "which was not only stabilized by the hydrophobic surface of the GDL" or "which was not only stabilizing the hydrophobic surface of the GDL"? Please clarify.
Response 1: We thank the reviewer for the comment. In the revised manuscript, we have clarified it to "which not only stabilized the hydrophobic surface of the GDL".
Please see Page 6 in the revised manuscript. Figure 2g and all other XRD patterns presented in the study: What is the main peak at about 24 2Theta due to? And the other two smaller peaks around it?

Comment 2:
Response 2: We thank the reviewer for the comment. As there is a layer of graphite in the hydrophobic carbon paper, the main peak mentioned by the reviewer was due to the (002) lattice plane of graphite, which was located at 26.1° in the XRD patterns.
The other two smaller peaks around it were also attributed to graphite. To avoid confusion, in the revised manuscript, we only give the pattern in the useful angle range from 30° to 60°, which shows Cu peaks clearly. Please see Figure 2g, S9, S13, S17 and S33 in the revised manuscript.
Response 3: We thank the reviewer for the comment. The Cu/CS composite was made of "Cu NPs + CS", as can been know from the characterizations in the manuscript ( Figure S11-13). We have also emphasized this by "The Cu/CS composite was made of Cu NPs and CS.". Please see Page 7 in the revised manuscript.
Comment 4: Line 150: The authors state " Figure S12 illustrated that Cu NPs were uniformly distributed in the CS network structure", but where can Cu NPs be seen in Figure S12? There may be some NPs visible in image "e" but it is not apparent.

Response 4:
We thank the reviewer for the comment. After reading the comment, we have added the TEM with high-magnification and the particle size distribution of Cu NPs in catalyst. As shown in Figure S12f and g, The TEM images indicate that the average Cu NPs size was around 5 nm, and Cu NPs were uniformly distributed in the CS network structure. Please see Page 11 in the Supplementary information of the revised manuscript.
In the revised manuscript, we have also emphasized this by "Supplementary Figure S12f and g further illustrated that Cu NPs with average particle size around 5 nm were uniformly distributed in the CS network structure.". Please see Page 7 in the revised manuscript.
Comment 5: Line 172: "1 M KOH aqueous solution" in the air or saturated with CO2?
Please specify.
Response 5: We thank the reviewer for the comment. As the electrolysis was carried out in a separated flow cell with three chambers, 1 M KOH aqueous solution was in the air and CO2 was then diffused to the catholyte through a gas diffusion layer. section of the revised manuscript. Accordingly, the diagram of the flow cell we used was also added, which was shown in Figure S19.
Comment 6: Line 211: Where are the values of ECSA? Figure S26 reports the values of electric double-layer capacitance for different GDEs but not of electrochemical surface area.

Response 6:
We thank the reviewer for the comment. As suggested by the referee, in the revised manuscript, we have calculated the ECSA, and the specific calculation method is supplemented in the "Electrochemical active surface area (ECSA) measurements" section. Please see Page 20 in the revised manuscript.
In the revised manuscript, we have also discussed this by "As shown in Comment 7: Line 214: The statement "It was obvious that 3D Cu-CS-GDL electrode had the largest ECSA, indicating that the 3D Cu structure was responsible for promoting high CO2RR activity via generating more active sites" may need to be revised based on the following considerations. The ECSA-normalised current density represents the CO2RR activity per unit of electroactive area (also referred to as "intrinsic activity"). A larger electrochemical surface area corresponds to a lower current density, hence lower CO2RR activity per unit area. For example, assuming that the values in Figure S19 are geometric current densities (please specify), the relative values of the LSV curves change when ESCA-normalised currents are used.
In particular in Figure S19, the geometric current density of Cu-CS-GDL and Cu/CS-GDL are about -0.6 and -0.5 A/cm2, respectively, at -0.6 V. However, the ESCA of Cu-CS-GDL should be about 5 times larger than that of Cu/CS-GDL (based on the ratio of Cdl values 12.0:2.5, from Figure S26e), it follows that the ESCA-normalised current density of Cu-CS-GDL is significantly lower than that of Cu/CS-GDL, meaning that the intrinsic activity of Cu-CS-GDL is less than that of Cu/cs-GDL. This is a critical point that needs to be clarified in the manuscript, for all electrocatalysts using their corresponding ECSA-normalised current densities.
In all cases, please specify which current densities are geometric or ECSA normalised.

Response 7:
We thank the reviewer for the very instructive comment. In the revised manuscript, we have measured the ECSA and the current densities were also ECSA normalized. The results were provided in Supplementary Figure S31f. After normalizing the current density to ECSA, 3D Cu-CS-GDL still exhibited the largest partial current densities of C2+ alcohols at the potential of -0.87 V vs RHE, which indicates that the 3D structure could improve the intrinsic activity of the catalyst.
In the revised manuscript, we have revised the discussion by "After normalizing the current density to ECSA, 3D Cu-CS-GDL still exhibited the largest partial current densities of C2+ alcohols at the potential of -0.87 V vs RHE, which indicates that the 3D structure could improve the intrinsic activity for producing C2+ alcohols in CO2RR ( Figure S31f).". Please see Page 11 in the revised manuscript.
In other cased, we have used current densities with geometric normalized, in order to compare it with other catalysts in the literature. In the revised manuscript, we have emphasized this by "the C2+ Faradaic efficiency (FE) could reach 88.2% with a current density (geometrically normalized) as high as 900 mA · cm -2 " and "Comparison of FE of C2+ products and current density (j, geometric normalized) over 3D Cu-CS-GDL architecture with some typical Cu-based catalysts in CO2RR".
Please see Page 1, Page 3 and Table S1 in the revised manuscript.
Comment 8: In all cases, please use either "Cu-CS-GDL" or "3D Cu-CS-GDL" but not both otherwise there is uncertainty regarding being either the same GDE or two different ones.

Response 8:
We thank the reviewer for the comment. According to the comment, we have replaced "Cu-CS-GDL" with "3D Cu-CS-GDL" in the revised manuscript.

Supporting information
Comment 9: In addition to the experimental details given in the manuscript, please provide a detailed description (possibly with photos) of the flow cell reactor used for CO2 electrolysis. In particular, provide the dimensions of the CO2RR GDE, and which ion transport membrane was used.
Response 9: We thank the reviewer for the comment. According to the comment, the diagram of the flow cell we used was added, which was shown in Figure S19.
In the revised manuscript, we have also added the description by "The electrocatalysis experiments were carried out in a separated flow cell with three chambers. As shown in Supplementary Figure  Comment 10: Figure S8 & S12: What are the rods seen in "e"?
Response 10: We thank the reviewer for the comment. The rods seen in FigureS8e and Figure S12e are graphite (or carbon fiber) in the hydrophobic carbon paper, which is represented by gray rods in the schematic illustration of Figure 1.

Comments to the Author
This study reports a 3D Cu-chitosan (CS)-GDL electrode for CO2 reduction. The authors claimed that the CS acts as a transition layer between catalyst and the GDL which can facilitate rapid electron transport and mitigate mass diffusion limitations.
This design resulted in a C2+ Faradaic efficiency (FE) of 88.2% at 900 mA/cm2 with a C2+ alcohols selectivity of 51.4% and a notable partial current density of 462.6 mA/cm2. I see value in the work. However, the following must be addressed for the work to be suitable for Nature Communications.

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

Major comments:
Comment 1: Adding a polymer layer on top of GDL can significantly decrease the electrical conductivity and impeding the electron transfer between the GDL and catalyst. More experiment/characterization and explanation are needed to support the authors' claim: "the CS acts as a transition layer between catalyst and the GDL which can facilitate rapid electrons transport and mitigates mass diffusion limitations".
Response 1: We thank the reviewer for the comment. After reading the comment, we have made the following modifications.
(i) To elucidate this point, we use electrochemical impedance spectroscopy (EIS) to study the interfacial properties of the electrodes (Figure S29). On one hand, Figure   S29a shows the Nyquist plot of various electrodes. It indicates that the charge transfer resistance (Rct) of 3D Cu-CS-GDL was much smaller than that of others. A reasonable interpretation of the result is that coupling 3D structure and CS can enhance electron mobility and accelerate the charge transfer rate on 3D Cu-CS-GDL interface, which is conducive to enhance the activity of CO2RR. On the other hand, the Bode plots  (2002)). It has been found that CS-derived adsorbents are attractive in the CO2 capture process because of the presence of amino groups in their structure. In addition, some researchers have made innovative applications in environmental, medical and other fields by using the capture CO2 ability of CS. Therefore, when CS synergized with other components to form integrated structure, the electrons transport and mitigates mass diffusion could be improved.
In the revised manuscript, we have added the discussion by "On the other hand, the electrochemical impedance spectroscopy (EIS) was also carried out to study the Biomacromolecules 5, 2340-2346 (2004)). Consequently, we think it is interesting to develop a new paradigm for CS application and propose a new route to prepare 3D Cu-CS-GDL electrode for CO2 reduction, and we believe that novel use of CS in the GDEs to tune the architecture is applicable to design of other efficient electrodes for CO2RR.
In the revised manuscript, we have added the discussion by "For catalyst layer, the most straight forward way is coating of powder-type electrocatalysts onto a gas diffusion layer (GDL) using commonly used polymers/binders, such as polyaniline (PANI), polypyrrole (PPy) and Nafion D-521. 12-14 However, the additive binders would inevitably decrease the CO2RR performance and considerably increase the overpotential, which are due to the obstruction of gas transport, insufficient exposure of active sites, and detachment of catalyst from electrode surface by binder degradation in the reaction. 15-17 " and "Chitosan (CS), an abundant amino polysaccharide, obtained from the carapaces of shrimp and crabs, containing a carbon skeleton with amino functional groups. [25][26][27] It has the advantages of low cost, non-toxic, renewable, degradable and abundant reserves, which has some unique advantages comparing with commonly used polymers/binders. The hydroxyl group and amino group in CS structure make it has strong affinity, especially has good chelation ability for transition metals, coordinating with metal ions to form complexes, this property also provides a basis for dispersing metal active sites. [25][26][27][28] In addition, CS has been proved to have the ability of structure guidance and good adsorption of CO2. 29, 30 These features of CS made it as interesting materials in designing electrocatalysts for CO2RR.". Please see Page 2 and 3 in the revised manuscript.

Comment 3: A high EtOH selectivity at high current densities was achieved in a flow
cell. Achieving record performance in membrane electrode assembly (MEA) would support the claim that this is a broadly applicable approach and one that advances the field toward practical application.

Response 3:
We thank the reviewer for the comment. On the basis of the comment, we have conducted CO2 electrolysis using MEA. The schematic diagram of MEA and the results were shown in Figure S34 and S35, respectively. As expected, it also achieved a high overall current of 1.2 A· cm -2 with C2+ alcohols FE of 36.7% at -3.6 V cell voltage. The production rates of EtOH and PrOH could reach 1.54 mmol· h -1 · cm -2 and 0.50 mmol· h -1 · cm -2 , respectively. In the revised manuscript, we have emphasized this by "We also carried out CO2 electrolysis in membrane electrode assembly (MEA) (Supplementary Figure S34 and S35) 36,43,44 . A high overall current of 1.2 A·cm -2 with C2+ alcohols FE of 36.7% was achieved at -3.6 V cell voltage, and the production rates of EtOH and PrOH were 1.54 mmol·h -1 ·cm -2 and 0.50 mmol·h -1 ·cm -2 , respectively.". Please see Page 11-12 in the revised manuscript.
Accordingly, the specific experiment operations have also been supplemented in "CO2 Electrolysis in membrane electrode assembly (MEA)" section of the revised manuscript. Please see Page 20 in the revised manuscript.

Comment 4:
This study was performed in an alkaline flow cell which leads to a severe carbonate formation, and the many associated issues documented well in this journal by the perspective by Kanan. How can the authors address the Kanan's concerns?
Response 4: We thank the reviewer for the comment. Currently, KOH and KHCO3 solution are both commonly used electrolytes for CO2RR. We agree with the referee that performing in an alkaline flow cell leads to a severe carbonate formation, because we also observed this phenomenon in our experiment. Currently, intermittent cleaning is used to remove salt precipitation if long-term electrolysis is carried out.
Alternatively, we think that CO2 electrolysis in MEA is a promising approach for practical application.
Comment 5: Most of the commercially available GDLs are hydrophobic. The reported contact angle in this study (95.1) is less than some previous reports. More explanation is needed to support why Cu-chitosan (CS)-GDL mitigates mass diffusion limitations and improves the CO2 reduction performance at higher current densities.

Response 5:
We thank the reviewer for the comment. Here we would like to discussion this briefly in response to the comment. Even though the commercially available GDLs are hydrophobic, we found that it is not the decisive factor to improve the catalytic activity. It has been found that controlling the electrode surface with appropriate contact angle was more conducive to form abundant gas-liquid-solid three-phase interface, which is favorable to improve the CO2RR performance (Nat  (2018)). The result shows clearly that ECSA values of 3D Cu-CS-GDL is larger than Cu NPs-GDL, indicating that more exposing active sites were formed in the abundant gas-liquid-solid three-phase interface. In addition, we have also added EIS experiment to explain the enhancing electrons transport and mass diffusion of 3D Cu-CS-GDL, and the results support our conclusions.
In the revised manuscript, we have emphasized this by "This also can be known from the fact that when CS synergized with other components to form integrated 3D structure, it could form abundant gas-liquid-solid three-phase interface with more exposing active sites. This phenomenon leads to the lower contact angle of 3D Cu-CS-GDL (95.1°) than that of Cu NPs-GDL (138.4°) and Cu/CS-GDL (135.8°).
However, a much lower contact angle of De-Cu-GDL (70°) leads to loss of gas diffusion ability (Supplementary Figure S30). The above result suggests that controlling the electrode surface with appropriate contact angle was more conducive to form abundant gas-liquid-solid three-phase interface with more exposing active sites, which is favorable to improve the CO2RR performance. 38 Comment 6: Cu NPs-GDL electrode shows a good performance at higher current densities (600-1000 mA/cm 2 ). Is this related to the synthesis approach? A comparison between commercial Cu NP and the Cu NPs-GDL is needed.

Response 6:
We thank the reviewer for the comment. We agree with the referee that the high performance is related to the synthesis approach. As suggested by the referee, we have also used commercial Cu NPs for comparison ( Figure S26). The electrochemical CO2RR performance was investigated ( Figure S27). As a result, the FE of C2+ was only 10.3% and H2 was the major product at 900 mA·cm -2 , indicating that the good performance of Cu NPs-GDL at high current density is related to the synthesis approach. In the revised manuscript, we have added the discussion by "In addition, we have also used commercial Cu NPs for comparison ( Figure S26). The real size of commercial Cu NPs was approximately 60 to 400 nm. As a result, the FE of C2+ was only 10.3% and H2 was the major product at 900 mA· cm -2 ( Figure S27).".
Please see Page 10 in the revised manuscript. Cu-CS-GDL and Cu NPs-GDL, which can be deconvolved into top-bound CO and bridge-bound, suggesting that the pathway of generating C2+ products was in progress.

51
". Please see Page 12 and 13 in the revised manuscript. The related reference has also been cited (Ref. 46, 48 and 51).
Comment 8: XAS analysis shows some interesting results. I do recommend revisiting the results and combining them with the mechanistic study in the main text.
Response 8: We thank the reviewer for the comment. As suggest by the referee, in the revised manuscript, we have added the discussion by "In Supplementary Figure S36, the XAFS data of 3D Cu-CS-GDL are provided at OCV, -0.4 V, -0.8 V vs RHE during CO2RR and after reaction. The K-edge XANES spectra and the derivative K-edge XANES spectra indicated that 3D Cu-CS-GDL presented zero valence Cu in the whole process of CO2RR. All curves in k space also followed the trend of the curve of Cu foil, which also proved that Cu (0) was maintained in CO2RR. In R space, the ever-present Cu-Cu bond confirmed to the above conclusion, but its corresponding radial distance was shifted with applied potential, which is caused by surface adsorption and lattice vibration in the reaction environment. Therefore, the internal bonding of 3D Cu-CS-GDL was constant.". Please see Page 12 in the revised manuscript.
Comment 9: XRD needs to be revised, I suggest removing graphitic peaks and focus on the Cu peaks. If the interface of Cu (200) and Cu (111) is very important in EtOH pathway and production, this needs to be experimentally shown.

Response 9:
We thank the reviewer for the comment. As suggested by the referee, we have removed the graphitic peaks and focus on the Cu peaks with the range from 30° to 60°. Please see Figure 2g, S9, S13, S17 and S33 in the revised manuscript.
Response 10: We thank the reviewer for the comment. In the revised manuscript, we have changed the statement to "which was very efficient for C2+ alcohols production" and "the as-synthesized 3D Cu-CS-GDL electrode was among the outstanding catalysts for C2+ products, especially for the high-rate production of C2+ alcohols (Figure 3d, Supplementary Table S1).". Please see Page 1 and 10 in the revised manuscript. The related reference has also been cited (Ref. 36, 37).
Minor comments: