Cu–Pd Bimetallic Gas Diffusion Electrodes for Electrochemical Reduction of CO2 to C2+ Products

Carbon dioxide electroreduction driven by renewable electricity into high‐value chemicals enables energy storage while contributing to climate change mitigation. Herein, a CuPd bimetallic catalyst is developed by electrodeposition for efficient CO2 electroreduction, achieving a C2+ Faradaic efficiency as high as 75.6% with a current density of −200 mA cm−2 at −1.15 V versus reversible hydrogen electrode. Upon incorporation of Pd, the average d‐band center of the CuPd bimetallic catalyst downward shifts relative to the Fermi level, making the adsorbed CO intermediates active for further C–C coupling. Theoretical calculations confirm that CO generated on the Pd domain can spill over to the CuPd interface for C–C coupling with a lower energy barrier, further promoting the formation of C2+ compounds. This study provides insights into the rational design of Cu‐based bimetallic catalysts for highly efficient CO2 electroreduction to multicarbon products.


Introduction
Excessive CO 2 emissions from fossil energy consumption have caused serious environmental problems and energy crises. [1][2][3][4] Electrochemical CO 2 reduction reaction (eCO 2 RR) to high value-added chemicals is a potential strategy to help close the anthropogenic carbon cycle while realizing intermittent electric energy storage. [5][6][7] At present, relatively proven strategies have been developed to electroreduce CO 2 to CO or formic acid (C 1 products) with high activity and selectivity. [8][9][10][11][12][13][14] However, the generation of the higher energy density hydrocarbons and alcohols (C 2þ products) that can alleviate the heavy dependence of energy and chemical industries on fossil fuels is still suboptimal in activity and selectivity. [15][16][17] Cu was recognized as the most effective catalyst to promote the electroreduction of CO 2 to C 2þ products due to its moderate *CO binding. However, the slow C-C coupling kinetics on the Cu surface limited the efficient generation of C 2þ products. [18][19][20] Therefore, it is of crucial significance to focus on improving the activity and selectivity of Cu-based catalysts for CO 2 electroreduction to C 2þ products.
The physicochemical properties of Cu-based catalysts directly affect the catalytic performance, so various modification strategies have been proposed to improve the selectivity of C 2þ products. [21,22] Among them, the bimetal strategy was reported to be able to break the scaling relationship and stabilize the reaction intermediates, thus improving the catalytic activity. [23] Recently, many efforts have been devoted to developing various Cu-based bimetallic catalysts for efficient eCO 2 RR. Buonsanti et al. revealed the critical role of Ag in modulating the electronic structure of the Cu interface, which ultimately promotes CO 2 electroreduction to multicarbon products. [24] Jaramillo et al. proposed a tandem catalytic mechanism for Au/Cu bimetallic catalysts, where the Au sites promoted CO formation and enriched the CO density around Cu sites, thereby facilitating the highly active reduction of CO 2 to alcohols. [25] Despite these efforts recently, the eCO 2 RR performance of Cu-based bimetallic catalysts is still insufficient for industrial applications, and the synergistic effect between bimetals is undefined. Therefore, it is necessary to design refined bimetallic catalysts to further enhance the eCO 2 RR activity and clarify the relationship between interfacial effects and catalytic performance.
Herein, a series of CuPd bimetallic catalysts were developed via electrodeposition for highly efficient eCO 2 RR to C 2þ products. The CuPd bimetallic catalyst with a Pd content of around 2% exhibited a high selectivity for C 2þ products formation that a high Faradaic efficiency of 75.6% and a high current density of À200 mA cm À2 was achieved at a low potential of À1.15 V versus DOI: 10.1002/sstr.202200328 Carbon dioxide electroreduction driven by renewable electricity into high-value chemicals enables energy storage while contributing to climate change mitigation. Herein, a CuPd bimetallic catalyst is developed by electrodeposition for efficient CO 2 electroreduction, achieving a C 2þ Faradaic efficiency as high as 75.6% with a current density of À200 mA cm À2 at À1.15 V versus reversible hydrogen electrode. Upon incorporation of Pd, the average d-band center of the CuPd bimetallic catalyst downward shifts relative to the Fermi level, making the adsorbed CO intermediates active for further C-C coupling. Theoretical calculations confirm that CO generated on the Pd domain can spill over to the CuPd interface for C-C coupling with a lower energy barrier, further promoting the formation of C 2þ compounds. This study provides insights into the rational design of Cu-based bimetallic catalysts for highly efficient CO 2 electroreduction to multicarbon products. reversible hydrogen electrode (RHE). The valence band spectra showed a downward shift of the average d-band center of the catalyst relative to the Fermi level upon the incorporation of Pd, which facilitates activating of the adsorbed CO intermediates for the further C-C coupling reaction. As revealed by the theoretical investigation, the generated CO on the Pd domain will spill over to the CuPd interface for C-C coupling with a lower energy barrier, thus promoting the generation of C 2þ products.

Synthesis and Characterizations of Catalysts
Monometallic Cu and three CuPd bimetallic catalysts were synthesized by galvanostatic electrodeposition in the electrolytes with controlled Pd 2þ concentrations (Figure 1a and S1, Supporting Information), which were denoted as Cu 100 Pd 0 , Cu 99 Pd 1 , Cu 98 Pd 2 , and Cu 97 Pd 3 , respectively ( Figure S2, Supporting Information), according to the element contents determined by inductively coupled plasma (ICP) measurements (Table S1, Supporting Information). Typical scanning electron microscope (SEM) images showed that the morphology of a series of CuPd catalysts transformed gradually from irregularly aggregated particles to well-aligned wires with Pd content increasing ( Figure S3, Supporting Information). The diffraction peaks at 43.29°, 50.43°, and 74.13°observed in X-ray diffraction (XRD) patterns of all samples (Figure 1b) are assigned to the metallic Cu (111), (200), and (220) planes, respectively. The Pd-associated peaks were very faint and invisible in XRD patterns due to the low Pd contents. No signal was observed for copper oxide and CuPd alloys. The energy-dispersive X-ray spectroscopy (EDS) elements mapping showed that Cu and Pd elements were evenly distributed in bimetal catalysts, and the Pd concentration gradually increased from Cu 100 Pd 0 to Cu 97 Pd 3 (Figure 1d and S3, Supporting Information), which is consistent with the ICP results. Based on the cross-sectional SEM observations, the electrodeposited catalysts were evenly distributed across the GDE, and the thickness of the catalyst layer on the top of the GDL gradually increased from 32.31 to 39.23 μm with the increasing Pd contents ( Figure S4, Supporting Information). High-resolution transmission electron microscopy (HRTEM) images showed the lattice spacings of 0.208 and 0.225 nm assigned to Cu(111) and Pd(111) planes, respectively, and the CuPd interface existed between the two grains ( Figure 1c and S5, Supporting Information). All these results indicate that the obtained CuPd catalysts are composed of phase-separated bimetals with CuPd(111) interfaces rather than alloys.
X-ray photoelectron spectroscopy (XPS) was performed to investigate the chemical states and composition of the catalysts. The Cu 2p peaks of CuPd bimetallic catalysts exhibited a positive shift of 0.28 eV compared with Cu (Figure 2a), while the Pd 3d peaks shifted to lower binding energy with the increase of Pd content (Figure 2b). The slight change in the binding energy of Cu 2p and Pd 3d has resulted from the electron interaction between Cu and Pd. [26] In addition, the sharp Cu 2p 3/2 peak at 932.6 eV is assigned to Cu 0 or Cu þ species, and the Cu LMM Auger spectrum further confirmed the existence of Cu þ on the surfaces apart from Cu 0 ( Figure S6, Supporting Information), which may be attributed to the oxidation upon exposure to the air. However, numerous studies have shown that this naturally formed oxide can be rapidly reduced and does not significantly affect the product distribution in eCO 2 RR. [27,28] Both Cu 2p and Cu LMM Auger peaks confirm that Cu 0 is the main species. The Pd 3d spectrum in catalysts was fitted with two components. The main 3d 5/2 peak located at a binding energy of 335.4 eV is the characteristic of metallic Pd 0 , while the very weak one at 336.9 eV is assigned to Pd II in the PdO state ( Figure S7, Supporting Information). However, PdO was not seen in XRD and HRTEM characterization, which means that the surface oxide layer proved by XPS is too thin to be detected. The metal surface exposed to air is easily oxidized, but due to the dynamic limit of surface oxidation, the bulk is still metallic. X-ray absorption spectroscopy (XAS) was conducted to further ascertain the chemical nature and structure of Cu species in CuPd bimetallic catalysts. As shown in the Cu K-edge X-ray absorption near-edge fine structure (XANES) spectroscopy (Figure 2c), the Cu K-edge positions of CuPd bimetallic catalysts are all close to that of Cu foil, confirming that the states of Cu in CuPd bimetallic catalysts are mainly composed of Cu 0 , which is consistent with the XRD results. The coordination environment of Cu was certified by the extended X-ray absorption fine structure (EXAFS). The Fourier-transform (FT) k 3 -weighted EXAFS spectra of CuPd bimetallic catalysts showed the same main peak as Cu foil at 2.2 Å, corresponding to a metallic Cu-Cu coordination shell (Figure 2d). With increasing the Pd content, the intensity of the metallic Cu-Cu peak gradually decreased, indicating the strong electronic interaction between Cu and Pd. This can be further confirmed by the fitting results ( Figure S8 and Table S2, Supporting Information) that the average coordination number of metallic Cu-Cu shell gradually decreased with the Pd incorporated increases. Additionally, no Cu-Pd coordination environment was found whether in FT-EXAFS curves or the fitting results further confirming that the obtained CuPd samples are composed of phase-separated Cu and Pd bimetals.

CO 2 Electrochemical Reduction Performance
The catalytic performance of a series of CuPd bimetallic catalysts for eCO 2 RR was evaluated in a flow cell with 1 M KCl electrolyte through the galvanostatic electrolysis ( Figure S9, Supporting Information). All the catalysts exhibited similar catalytic behavior, that is, the main reduction products were formate and CO (C 1 products) at low applied potentials, and the C 2þ products including C 2 H 4 and CH 3 CH 2 OH gradually dominated at high applied potentials. In the meantime, the product distribution, especially C 2þ , also varied with the incorporated Pd contents (Figure S10, Supporting Information). With the introduction of Pd into the catalysts, the C 2þ products gradually increased until the maximum C 2þ Faradaic efficiency was obtained over the Cu 98 Pd 2 catalyst. In regard to Cu 98 Pd 2 (Figure 3a), the Faradaic efficiency of C 2þ products dominated by C 2 H 4 and C 2 H 5 OH gradually increased as the potential shifted from À0.79 to À1.15 V in 1 M KCl solution, while the C 1 Faradaic efficiency gradually decreased and the H 2 Faradaic efficiency decreased from 35.3% to 16.1%. The Faradaic efficiency of the C 2þ products reached a maximum of 75.6% at À1.15 V, specifically, the Faradaic efficiencies of C 2 H 4 and C 2 H 5 OH were 32.2% and 38.5%, respectively, and the remaining products were a small amount of n-propanol and acetate. When the potential further negatively shifted to À1.32 V, the Faradaic efficiency of C 2þ products decreased to 56.3% at the expense of H 2 Faradaic efficiency increasing to 21.5%, indicating that the hydrogen evolution reaction was gradually dominant at more negative potentials and thus unfavorable for eCO 2 RR.  The Faradaic efficiencies of H 2 , C 1 , and C 2þ products are compared at a current density of À200 mA cm À2 in Figure 3b. The Faradaic efficiency of H 2 was not significantly different between the four samples, while C 1 and C 2þ showed different distributions. In contrast to Cu 100 Pd 0 without Pd incorporation, the selectivity of C 2þ products on CuPd bimetallic catalysts increased with the Pd content increasing and reached the maximum value at Cu 98 Pd 2 . At this time, the Faradaic efficiency of C 2þ products reached 75.6%, while that of C 1 products was inhibited to 7.0%. When the Pd content further increased to 3%, the C 2þ Faradaic efficiency dropped to 47.6% at the expense of increased H 2 and C 1 Faradaic efficiencies. Therefore, the electronic structure of Cu should be delicately modulated with moderate Pd content to make it conducive to eCO 2 RR.
The selectivity of eCO 2 RR to C 2þ products was further evaluated by the C 2þ /C 1 Faradaic efficiency ratio. As shown in Figure 3c, the value of C 2þ /C 1 gradually increased to a maximum and then decreased with the potential negatively increasing. The maximum C 2þ /C 1 value of Cu 98 Pd 2 (%10.76 at À1.15 V) is 4.8, 2.0, and 4.2 times higher than that of Cu 100 Pd 0 (%2.22 at À1.29 V), Cu 99 Pd 1 (%5.32 at À1.20 V), and Cu 97 Pd 3 (%2.54 at À1.15 V), respectively, demonstrating the highest selectivity of C 2þ formation on Cu 98 Pd 2 . The selectivity of CO 2 conversion to C 2þ products on Cu 98 Pd 2 was up to 91.5% at À1.15 V, which is significantly higher than that of Cu 100 Pd 0 (68.9%), Cu 99 Pd 1 (84.2%), and Cu 97 Pd 3 (71.7%) ( Figure S11, Supporting Information). The eCO 2 RR activities of CuPd bimetallic catalysts were evaluated by the C 2þ partial current density. Figure 3d shows the C 2þ partial current density as a function of the applied potential. The onset potential on Cu 98 Pd 2 was more positive than Cu 100 Pd 0 , Cu 99 Pd 1 , and Cu 97 Pd 3 , indicating that electrons on Cu 98 Pd 2 more easily transferred to CO 2 to form the initially adsorbed intermediates. Moreover, electrochemical impedance spectroscopy (EIS) measurements were performed at opencircuit potential (OCP) to study the charge transfer behavior. Cu 98 Pd 2 shows smaller charge transfer resistance (R ct ) than other catalysts, indicating the more favorable charge transfer kinetics for eCO 2 RR ( Figure S12, Supporting Information). [29] The C 2þ partial current density of Cu 98 Pd 2 reached À225 mA cm À2 at À1.32 V, which is higher than that of Cu and other CuPd bimetallic catalysts. The C 2þ partial current density of Cu 99 Pd 1 and Cu 97 Pd 3 increased to the maximum and then decreased with the increase of applied potential, while the C 2þ partial current density of Cu 98 Pd 2 showed a consistently increasing trend within the applied potential range, indicating a high potential of Cu 98 Pd 2 for eCO 2 RR. To exclude the effect of the number of active sites, electrochemically active surface areas (ECSA) of the Cu 100 Pd 0 , Cu 99 Pd 1 , Cu 98 Pd 2 , and Cu 97 Pd 3 were  evaluated by the electrochemical double-layer capacitance (C dl ) ( Figure S13, Supporting Information). Although the C dl increases a little with the Pd content increases (Table S3, Supporting Information), the Cu 98 Pd 2 still shows the highest ECSA-normalized C 2þ partial current density ( Figure S14, Supporting Information). Furthermore, the eCO 2 RR activity of CuPd bimetallic catalysts in 1 M KHCO 3 and 1 M KOH was also investigated to shed light on the effect of anions in the electrolyte ( Figure S15 and S16, Supporting Information). Cu 98 Pd 2 exhibited superior C 2þ Faradaic efficiency to other catalysts in various electrolytes. On Cu 98 Pd 2 catalyst, the C 2þ Faradaic efficiency was only 9.8% at À200 mA cm À2 in KHCO 3 solution, while it could reach 71.2% at À200 mA cm À2 in KOH solution. Both of them are lower than that in the KCl solution (75.6%). In contrast to the KHCO 3 solution, KCl and KOH solutions exhibited comparably excellent eCO 2 RR activity and selectivity within the applied potential range. The C 2þ Faradaic efficiency of Cu 98 Pd 2 in KHCO 3 , KOH, and KCl solutions gradually increased following KHCO 3 < KOH < KCl (Figure 4a), which is the same with other catalysts ( Figure S17, Supporting Information). All these results indicate that the high ionic conductivity and local buffering capacity could enhance the proton transfer reaction rate and thus improve the activity of eCO 2 RR. [30] In near-neutral 1.0 M KHCO 3 electrolyte with the same CO 2 diffusion path, CuPd electrode showed poor activity and selectivity for eCO 2 RR. The relevant calculations show that Cl À will provide negative charge to the Cu surface in the chemisorption state, resulting in a dipole moment and the change of surface local electronic environment. On the other hand, KCl and KHCO 3 exhibit different buffering capacities result in different electrode surface pH values. For KHCO 3 electrolyte with strong buffer capacity, the pH before and after electrolysis will basically remain unchanged. The low buffering capacity of the KCl electrolyte allows the electrode surface pH to increase to the weak alkalinity. The increase of surface pH value in KCl electrolyte is conducive to the production of C 2þ because initial potential of C 2þ formation shifts positively with the increase of pH value, which is consistent with previous reports. [31][32][33][34] Long-term stability tests of Cu 98 Pd 2 were performed in the KCl and KOH solutions (Figure 4b). In KCl electrolyte, the C 2þ Faradaic efficiency decreased from 75.61% to 73.29% after 10 h electrolysis. Even after 20 h electrolysis, the C 2þ Faradaic efficiency decreases to 67.24%, indicating that most of the initial electrocatalytic activity is preserved. However, the Faradaic efficiency of C 2þ in KOH catholyte decreased rapidly from 71.5% to 58.5% after 10 h electrolysis, while the Faradaic efficiency of H 2 increased from 17.7% to 27.2%. The reaction stability of Cu 98 Pd 2 in the KCl solution was better than that in the KOH solution. Inevitably, CO 2 gas can pass through the gas diffusion layer and react with the KOH solution to generate carbonate. The continuous accumulation of carbonate will block the CO 2 transfer channels, which may be the reason for the sharp decline of CO 2 reduction activity in KOH solution. [35,36]

Theoretical Calculations and Mechanism Analysis
The electronic effect between Cu and Pd may be the crucial factor affecting the eCO 2 RR activity. [26,37] To confirm this possibility, we calculated the d-band center relative to the Fermi level based on the surface valence band photoelectric emission spectra of Cu and CuPd bimetallic catalysts. It can be seen from Figure 5a that the d-band center of Cu 99 Pd 1 was more negative than that of Cu and further negatively shifted with the Pd content increased, implying the electron redistribution upon incorporation of Pd in CuPd bimetallic catalysts. As discussed by Norskov et al., the lower d-band position relative to the Fermi level was related to the weaker interaction between the intermediate and the catalyst surface due to the occupation of antibonding states. [38] The highest d-band center position of the Cu sample indicated the strongest binding with CO, while the moderate d-band center of Cu 98 Pd 2 could activate adjacent CO intermediates and promote C-C coupling. Overall, the addition of Pd optimized the electronic structure of Cu and improved the activity of eCO 2 RR. It has been proved that the electron-depleted Cu increases the CO binding on the catalyst surface, which, in turn, promotes the coupling of CO to C 2þ . This phenomenon can only occur when there is a shared interface on the catalyst, which has been proved in other studies. [24,[39][40][41] Apart from a surface with moderate binding strength to CO, a localized CO-rich environment has been confirmed to be another important factor that contributes to eCO 2 RR to  C 2þ products. [42,43] The eCO 2 RR activity of the CuPd bimetallic catalyst was higher than that of Cu because of the synergistic effect of Cu and Pd atoms. CO 2 was easier to be converted into CO on the Pd domain, while C-C coupling is difficult to occur. The generated CO on the Pd domain will spill over to the CuPd interface and/or Cu domain driven by thermochemical force to enrich the local CO concentration for C-C coupling (Figure 5b). The CuPd interface and Cu surface had moderate adsorption behavior of key intermediates, [26,44] which facilitates the production of C 2þ products and leads to the superior eCO 2 RR activity of CuPd bimetallic catalysts to most Cu-based bimetallic catalysts (Table S4, Supporting Information). The intrinsic activity reflected by the ECSA-normalized C 2þ partial current density has also been improved on CuPd catalysts, which indicates that the enhancement of C 2þ formation on CuPd may be mainly attributed to the incorporation of Pd promoting the formation of CO. Pd is well known to generate more CO than Cu at a given potential. Therefore, one possibility is that the Pd domain supplies a high flux of CO to the adjacent Cu domain, which sequentially transforms CO into C 2þ . Theoretical calculations were carried out to better understand the mechanism of eCO 2 RR on the CuPd bimetallic catalysts. It was reported that the coupling of two *CO is the ratedetermining step (RDS) in the formation of C 2 products. [45,46] We then calculated the energy barrier of the *CO coupling at the possibly active Cu surface and the CuPd interface. Based on the characteristic results, Cu(111) and CuPd(111) interfaces were selected as the models for calculations. The free energy of C-C coupling to form *OCCO on the CuPd interface (0.60 eV) was significantly lower than that on the Cu surface (1.66 eV). In addition, the activation energy barriers of C-C coupling at the Cu surface and the CuPd interfaces were 1.72 and 1.14 eV, respectively. These results indicate that the CuPd interface was more likely to kinetically promote the C-C coupling (Figure 5c). Then, we explored the possible reaction paths of C 2 products (C 2 H 4 and C 2 H 5 OH) formation on the Cu surface and Figure 5. a) Surface valence band photoemission spectra of various CuPd catalysts. b) Schematic illustration of the mechanism of Cu-Pd catalyst for promoting eCO 2 RR to C 2þ products. c) The energy profiles of *CO coupling on the Cu domain and the phase-separated Cu-Pd interface. d) The calculated free energy diagram for *CO coupling to *CH 2 CHO at the phase-separated Cu-Pd interface.
www.advancedsciencenews.com www.small-structures.com CuPd interface ( Figure S18, Supporting Information). [47,48] As shown in Figure 5d, the coupling of *CO at the CuPd interface requires free energy of 0.60 eV and the hydrogenation of *OCCO to generate *COCHO needs 0.89 eV, which is a little endothermic. Subsequently, *COCHO removes a water molecule to form *CCO with a reaction energy of À0.91 eV, which are thermodynamically very favorable. The subsequent multiple proton-electron transfer reactions proceed to form the intermediate *CH 2 CHO. After that, the protonation of the C atom bonded to the O atom will result in the rapid cleavage of the C─O bond and the final formation of C 2 H 4 in accompany by H 2 O. Alternatively, the protonation of the α-carbon in *CH 2 CHO will give the final generation of C 2 H 5 OH. Overall, both of the reaction pathways to generate C 2 H 4 and C 2 H 5 OH are quite thermodynamically favorable, which contributes to the similar formation trends of these two products during eCO 2 RR. The energetic pathways for the formation of C 2 H 4 and CH 3 CH 2 OH paths on the Cu surface have also been examined, and the calculated results indicate that the *CO coupling to C 2 products can occur on both the Cu surface and the CuPd interface ( Figure S19, Supporting Information). However, the high energy barrier of *CO coupled to *OCCO on the Cu surface indicates that the reaction is thermodynamically unfavorable. Therefore, the CuPd interface promotes the conversion of CO 2 to C 2 products more effectively than the Cu interface.

Conclusion
In summary, we successfully developed a series of CuPd bimetallic catalysts for highly efficient electroreduction of CO 2 to C 2þ products. Especially, the Cu 98 Pd 2 catalyst exhibited an excellent C 2þ products Faradaic efficiency of 75.6% and a current density of À200 mA cm À2 at À1.15 V in 1 M KCl solution. The synergistic effect of Cu and Pd in the CuPd bimetallic catalyst significantly improves the charge transfer kinetics and the local CO concentration, both of which are favorable for eCO 2 RR to C 2þ products. Furthermore, the abundant CuPd interface possesses unique electronic properties and adsorption behavior of key intermediates, which reduced the energy barrier for C-C coupling and facilitated the formation of C 2þ products. This work contributes to understanding the interfacial effects of CuPd bimetallic catalysts and provides important insights and strategies for the conversion of CO 2 to C 2þ products via Cu-based bimetallic catalysts.

Experimental Section
Preparation of Catalysts: Electrochemical deposition was carried out in a self-made quartz electrolytic cell. A commercial carbon paper (YLS-30T, Toray Industries Inc., 1 cm Â 1 cm) with coating a hydrophobic microporous layer as the support substrate was used for electrodepositing catalysts. The hydrophobic microporous layer was acted as the gas diffusion layer (GDL) during CO 2 electroreduction. The carbon fiber side of carbon paper was covered by adhesive tape to only expose the GDL for electrodeposition. The electrolyte used for the preparation of Cu 100 Pd 0 sample was 0.1 M CuSO 4 and 0.1 M H 2 SO 4 solution. The concentration of Pd 2þ added in the electrolyte (0.1 M CuSO 4 and 0.1 M H 2 SO 4 ) for the preparation of Cu 99 Pd 1 , Cu 98 Pd 2 , and Cu 97 Pd 3 samples was 0.5, 1, and 2 mM, respectively. Electrodeposition was performed at a constant current density of À2.5 mA cm À2 for 500 s. According to the quantity of electric charge applied during electrodeposition, the mass load of Cu 100 Pd 0 , Cu 99 Pd 1 , Cu 98 Pd 2 , and Cu 97 Pd 3 is about 0.415, 0.416, 0.418, and 0.419 mg cm À2 , respectively (assuming a 100% Faradaic efficiency for deposition). The Cu-Pd bimetallic gas diffusion electrodes with different Pd contents were then obtained after washing with deionized water and drying overnight.
Materials Characterization: SEM images were collected with a ZEISS microscope (SIGMA 500, Germany). XRD measurements were performed on a Rigatku Ultima 4 X-ray diffractometer with Cu Kα radiation, operating at 40 kV and 40 mA and in the step mode (0.0167°). XPS was conducted using a Quantum 2000 Scanning ESCA Microprobe instrument with a monochromatic Al Kα source (1486.6 eV). The peaks of Cu XPS and Cu auger were recorded in the binding energy ranges of 925-965 and 555-585 eV, respectively. The binding energies in all XPS spectra were calibrated according to the C 1s peak (284.8 eV). The d-band center was determined based on the valence band spectra recorded in the binding energy range of 2 to À9.0 eV by ∫ NðεÞεdε=∫ NðεÞdε, where N(ε) is the density of states. TEM was conducted on a JEOL microscope (2100F) operated at 200 kV. The ICP optical emission spectrometry tests were performed on a Thermo Fisher iCAP PRO using RF Power under 1150 W. EXAFS spectroscopy data were collected by a BL14W1 at 3.5 GeV in transmission mode at the Shanghai Synchrotron Radiation Facility (SSRF). The EXAFS spectra were measured at the Cu K-edge (7709 eV) under ambient conditions. The energy was calibrated according to the absorption edge of pure Cu foil, and data extraction was performed in Athena. The raw data were corrected and normalized by IFEFFIT software.
Electrochemical Measurement for CO 2 Reduction: A typical GDE-based flow cell was used for eCO 2 RR experiments ( Figure S9, Supporting Information). A three-electrode system consisting of a working electrode, an Ag/AgCl reference electrode, and a counter electrode (Ni foam) was adopted. The GDE-based electrolyzer adopted a flow-through configuration for gas flow. The gaseous CO 2 diffuses into the GDE pores and arrives at the vicinity of the catalyst to form a three-phase interface with the catalyst and electrolyte. The gaseous products were collected on the same GDE side as the gaseous feed, and the liquid products entered the electrolyte. The cathodic and anodic compartments were separated by a Nafion 117 membrane. A 1.0 M KOH aqueous solution was used as anolyte; 1.0 M KCl, 1.0 M KOH, and 1.0 M KHCO 3 solutions were used as catholytes, respectively. Both anolyte and catholyte were cycled at a fixed flow rate of 20 mL min À1 by using two identical peristaltic pumps (Jihpump BT-50EA 153YX), and the flow rate of CO 2 was controlled at 20 mL min À1 by a mass flow controller. CO 2 reduction was performed by chronopotentiometry for 30 min at each fixed current density. The ECSA of the electrode was evaluated by the double-layer capacitance (C dl ). Cyclic voltammetry was carried out at in a potential range where Faradaic processes are absent and at four different scan rates (40,60,80, and 100 mV s À1 ). The geometric current density was then plotted as a function of scan rate. The C dl , which was proportional to the ECSA, was obtained by linearly fitting the absolute value of the slope of Δj (the difference of cathodic and anodic current densities of the cyclic voltammetry curves) against the scan rates.
To determine the solution resistance, potentiostatic EIS tests were performed at OCP after the electrocatalytic performance tests at each applied potential, and the frequency limits typically set in the range of 100 kHz to 20 Hz, with a voltage amplitude of 10 mV. All the current densities presented in this work were calculated on the basis of the electrode geometric area unless otherwise stated. All potentials were corrected for iR loss using the uncompensated resistance derived from the EIS and converted to the potential with the RHE. All applied potentials in the main text and Supporting Information referred to the RHE unless otherwise stated.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.