Bimetallic Tandem Strategy for Effective Modulation of CO₂ Electrocatalytic Selectivity on Relatively Inert Cu Interfaces

Abstract

The development of effective strategies to enhance the activity and selectivity of Cu‐based catalysts for the CO₂ reduction reaction (CO₂RR) remains a critical challenge, particularly on relatively inert Cu facets that are highly active in the competitive hydrogen evolution reaction (HER). In this study, a series of Cu‐M (M = Au, Ag, Pd) bimetallic tandem interfaces are fabricated on Cu nanocolumn arrays (Cu NCAs) with predominantly exposed Cu (200) facets. These interfaces excited a significant number of Cu⁺ species and modulated the adsorption behavior of critical intermediates for CO₂RR, resulting in improved CO₂RR activity and a substantial reduction in the competitive HER. Notably, the Cu‐Au, Cu‐Ag, and Cu‐Pd NCAs displayed excellent selectivity for C₂H₄, CO, and HCOOH products, with optimized faradaic efficiency (FE) of ≈43.2, ≈48.0, and ≈50.7%, respectively. According to in situ spectroscopic analysis, each Cu‐M interface exhibited distinct catalytic pathways: Cu‐Au favored *COOH and *CO adsorption followed by C‐C coupling for C₂H₄ production, Cu‐Ag promoted *CO desorption for CO generation, and Cu‐Pd facilitated *OCHO formation for HCOOH production. This study provides a strategy to design high‐performance and more practical bimetallic Cu‐based catalytic electrodes by directly modifying various commercial Cu substrates.

Bimetallic tandem interfaces of Cu‐M (M = Au, Ag, Pd) are constructed on Cu nanocolumn arrays with exposed Cu (200) facets to enhance CO₂ reduction reaction (CO₂RR) performance. In situ, spectroscopic analysis revealed specific reaction pathways for each interface.

1. Introduction

The escalating global environmental issues and energy crisis have rendered carbon dioxide (CO₂) emissions a global challenge that must be addressed urgently.^{[ 1 ]} Renewable energy‐induced CO₂ reduction reaction (CO₂RR) presents a critical pathway for alleviating the energy crisis and achieving carbon neutrality by transforming CO₂ into high‐value fuels and chemicals. This enables the generation of ethylene (C₂H₄), carbon monoxide (CO), multi‐carbon chemicals, and other valuable products under relatively mild conditions.^{[ 2 ]} Therefore, CO₂RR helps in reducing atmospheric CO₂ levels and possesses significant economic and industrial value.^{[ 3 ]} Hence, developing efficient and highly selective catalysts is crucial for achieving breakthroughs in CO₂RR technology and advancing its practical applications.^{[ 4 ]}

Copper (Cu)‐based catalysts have garnered significant attention in CO₂RR due to their high efficiency, cost‐effectiveness, and capacity to produce multi‐carbon (C₂) products, particularly high‐value chemicals like C₂H₄.^{[ 5 ]} However, commercial Cu substrates often exhibit limited selectivity and activity for CO₂RR due to the high‐temperature fabrication processes, which promote the exposure of Cu (200) and (220) facets that inherently favor the competitive hydrogen evolution reaction (HER) over CO₂RR.^{[ 6 ]} Consequently, CO₂RR selectivity is significantly compromised, necessitating energy‐intensive separation to obtain the desired products.^{[ 7 ]}

In recent years, bimetallic tandem strategies have emerged as an effective approach to optimize catalyst performance and boost product selectivity.^{[ 8 ]} Notably, the bimetallic surface can undergo electronic and structural modifications by introducing a second metal (e.g., Au, Ag, or Pd) with distinct adsorption characteristics toward key reaction intermediates (e.g., *CO, *COOH/*OCHO, or H* species) onto a Cu‐based interface, thereby steering reaction pathways more favorably toward the desired CO₂RR products.^{[ 9 ]} Nevertheless, numerous studies have primarily focused on specially prepared Cu catalysts with CO₂RR active facets, such as Cu (110), through metal/non‐metal dopants, defect engineering, or designed ligands.^{[ 10 ]} In contrast, limited efforts have been made to precisely tune the CO₂RR catalytic performance of relatively inert interfaces like Cu(200) and Cu(220).^{[ 11 ]}The underlying mechanisms of CO₂RR on interfaces based on these less active Cu facets must be explored to develop strategies that can transform the Cu (200) or (220) facets into efficient CO₂RR catalytic interfaces while simultaneously suppressing HER. This is crucial for fabricating high‐performance and more practical CO₂RR catalytic electrodes using commercial Cu substrates. Therefore, investigating the bimetallic tandem cooperative mechanism on these relatively inert Cu interfaces is crucial, despite remaining a challenge.

In this study, various typical Cu‐M (M = Au, Ag, Pd) bimetallic tandem interfaces were constructed onto free‐standing Cu nanocolumn arrays (Cu NCAs) with dominantly exposed (200) facets using an easy electroless deposition method. Cu interfaces featuring Au, Ag, or Pd as the second active metal species excited more Cu⁺ species onto Cu (200) facets, resulting in enhanced CO₂RR performance and a substantial reduction in the competitive HER. Specifically, Cu‐M NCAs were selective for producing C₂H₄, CO, and HCOOH, efficiently customized by varying the tandem secondary metal type, progressing from ‐Au to ‐Ag to ‐Pd, respectively. The in situ Raman and attenuated total reflectance Fourier transform infrared (ATR‐FTIR) spectral analysis revealed that the Cu‐Au interface efficiently excited and stably adsorbed *COOH and *CO intermediates, thereby facilitating *OCCO coupling and C₂ product formation (Faradaic efficiency (FE) of C₂H₄ was ≈43.2%). In contrast, Cu‐Ag interfaces demonstrated weaker *CO adsorption properties, leading to a preference for *CO desorption and the direct generation of CO over C‐C coupling (FE of CO was ≈48.0%). Notably, Cu‐Pd exhibited the lowest energy barrier to activate CO₂ for the formation of *OCHO primary intermediates. This, when coupled with the weaker *CO adsorption, accelerated the kinetics of HCOOH pathways (FE of HCOOH was ≈50.7%). Additionally, this study provides valuable insights into developing high‐performance Cu‐based CO₂RR catalytic electrodes by integrating diverse free‐standing Cu substrates using bimetallic tandem strategies, paving the way for enhanced catalytic selectivity and stability in practical applications.

2. Results and Discussion

2.1. Synthesis and Characterizations of Cu‐M NCAs (M = Au, Ag, Pd) Electrodes

Herein, integrated Cu‐based nanocolumn array electrodes with different bimetallic catalytic interfaces (Cu‐M (M = Au, Ag, Pd) NCAs) were synthesized through electroless deposition of the second active metal species onto the surface of Cu nanocolumn arrays (Cu NCAs). The detailed processes are depicted in the experimental section (Scheme S1, Supporting Information). Additionally, pristine Cu NCAs, processed under the same conditions without any other metal species, were used as the control sample. The X‐ray diffraction (XRD) pattern for pristine Cu and Cu‐M (M = Au, Ag, Pd) NCAs electrodes showed Cu metal diffraction peaks, with the (200) crystal plane being predominant. Notably, the Cu (200) facet is generally considered a relatively inert interface for the CO₂ reduction reaction (CO₂RR). However, it exhibits high activity toward the competitive HER. The Cu (200) facet is a commonly exposed facet of numerous commercial Cu substrates (like rolled foil, nanoarrays, and foam), which is one of the primary reasons why these commercial Cu substrates cannot be directly utilized for CO₂RR electrocatalysis.^{[ 12 ]} Furthermore, no obvious signal related to the other metal components was observed, likely due to the limited amount of the second active metal (Au, Ag, Pd) that existed only on the Cu NCAs surface (Figure S1A, Supporting Information). The changes in the short‐range alterations of the surface structure between Cu and Cu‐M NCAs were detected using Raman spectra. Moreover, Cu‐Au, Cu‐Ag, and Cu‐Pd NCAs electrodes displayed the characteristic peak signals of the second metal component oxides, AuO_x, Ag₂O, and PdO, respectively (Figure S1B, Supporting Information).

The scanning electron microscopy (SEM) images revealed that Cu NCAs consist of curved dendritic nanocolumn arrays with tip sizes in the range of ≈45–73 nm, and the spacing between nanocolumns is nearly 60 nm (Figure 1A,E). The morphology of Cu‐M NCAs (M = Au, Ag, Pd) following the deposition of the second metal component resembles that of pristine Cu NCAs (Figure 1B–D,F–H). Nevertheless, the Cu‐M NCAs electrode exhibits elevated surface roughness. The energy‐dispersive X‐ray spectrum (EDS) reveals the uniform distribution of the deposited metal elements (Au, Ag, Pd) on the Cu substrate, while only Cu and O elements are present on the pristine Cu NCAs (Figure 1I–P). Additionally, the high‐resolution transmission electron microscopy (HRTEM) images show that the main exposed facet on the pristine Cu NCAs surface is the (200) facet with a lattice spacing of 2.76 Å, which aligns with the XRD results (Figure 2A). Following the deposition of the second metal components, the lattice spacings of the nanoparticle regions formed on the surface of Cu substrates were measured to be 2.45, 2.66, and 2.17 Å for Cu‐Au, Cu‐Ag, and Cu‐Pd NCAs, associated with the (200) facet, (211) facet, and (111) facet of Au, AgO_x, and Pd, respectively (Figure 2B–D). According to the high‐resolution EDS mapping results, the Cu and the deposited Au, Ag, or Pd interfaces were uniformly mixed at the sub‐100 nm scale (Figure 2G–O). Thus, the aforementioned results confirm the successful formation of the bimetallic tandem structures for Cu‐Au, Cu‐Ag, and Cu‐Pd on the Cu NCAs electrode surface.

Figure 1.

SEM images for (A and E) Cu, (B and F) Cu‐Au, (C and G) Cu‐Ag, and (D and H) Cu‐Pd NCAs acquired from the top and cross‐section views; the SEM‐EDS elemental mappings for (I,J) Cu, (K,L) Cu‐Au, (M,N) Cu‐Ag, and (O,P) Cu‐Pd NCAs.

Figure 2.

HRTEM images for (A) Cu, B) Cu‐Au, C) Cu‐Ag, and (D) Cu‐Pd NCAs; HAADF‐STEM images of (E,F) Cu, (G–I) Cu‐Au, (J–L) Cu‐Ag, and (M–O) Cu‐Pd NCAs and corresponding element mappings.

Furthermore, the electronic structure of the Cu on the Cu‐M NCAs electrode surface was investigated by X‐ray photoelectron spectroscopy (XPS). Following peak fitting, the Cu‐Au NCAs electrode presented two characteristic peaks (932.5 and 952.6 eV) that correspond to the Cu⁰/Cu⁺ state on the Cu 2p _3/2 and Cu 2p _1/2 orbitals, respectively.^{[ 13 ]} The lower characteristic peaks (934.8 and 954.8 eV) were associated with the Cu²⁺ species,^{[ 14 ]} and the high oxidized Cu²⁺ species ratio is ≈21.4% (Figure 3A). Notably, similar Cu 2p XPS characteristic peaks were obtained for the other Cu‐M electrode samples. However, the Cu²⁺ species ratio of the Cu‐Ag NCAs significantly increased to 40.4% (Figure 3D), while this value in Cu‐Pd NCAs was the lowest among the three samples, ≈17.2% (Figure 3G). To quantitatively compare the ratio of Cu⁺ species in Cu‐M NCAs, which significantly impact Cu‐based interfaces for their CO₂RR capacity and selectivity, the XPS spectra of Cu LMM were analyzed. Additionally, the Cu⁺ species at 916.8 eV was identified as the dominant state in all Cu‐M NCAs samples, along with a specific amount of Cu²⁺ species at 917.7 eV and metallic Cu species at 918.6 eV.^{[ 15 ]} The Cu⁺ species content in Cu‐Au and Cu‐Pd NCAs reaches ≈70.0% and 74.3%, respectively (Figure 3B,H), while the Cu⁺ (≈51.1%) and Cu²⁺ (≈43.2%) species coexist on the Cu‐Ag NCAs surface, accounting for a similar proportion (Figure 3E). In contrast, the Cu⁺ species content on the pristine Cu NCAs surface is relatively lower than the Cu‐M bimetallic tandem interfaces (Figure S2, Supporting Information). Besides, the Cu‐M bimetallic tandem strategy might enhance the CO₂RR activity of the Cu interface by forming more Cu⁺ sites.^{[ 16 ]}

Figure 3.

High‐resolution XPS spectra for Cu 2p and Cu LMN of (A,B) Cu‐Au, D,E) Cu‐Ag, and (G,H) Cu‐Pd NCAs; (C) Au 4f spectra for Cu‐Au; (F) Ag 3d spectra for Cu‐Ag; (I) Pd 3d spectra for Cu‐Pd.

Additionally, the Au 4f, Ag 3d, and Pd 3d orbitals of the Cu‐M NCAs samples were analyzed. The characteristic peak at 84.4 eV in the Au 4f_7/2 orbital of Cu‐Au NCAs was attributed to the Au⁰ state (Figure 3C).^{[ 17 ]} In the Ag 3d XPS spectra of Cu‐Ag NCAs, the peak at 368.3 eV corresponds to the Ag⁰ state in the Ag 3d_5/2 orbital (Figure 3F). Compared to common metallic Au and Ag, the binding energies of Au and Ag in Cu‐Au and Cu‐Ag interfaces showed a negative shift (≈0.1–0.3 eV), indicating their strong interactions with Cu, which acts as an electron acceptor.^{[ 18 ]} Meanwhile, the Pd 3d XPS spectra for the Cu‐Pd NCAs electrode revealed Pd⁰ and Pd²⁺ states at 336.0 and 337.4 eV, respectively (Figure 3I). Notably, these values have a positive shift (≈0.1–0.2 eV) compared to the common Pd species, demonstrating strong interactions between Cu and the electron‐donating Pd.^{[ 19 ]}The aforementioned results show that the surfaces of the Cu‐M NCAs exhibit distinct intermetallic interactions between Cu and Au, Ag, or Pd, inducing significant alterations in the electronic structure of the Cu sites.

2.2. CO₂RR Catalytic Performance of Cu‐M NCAs (M = Au, Ag, Pd) Electrodes

The electrocatalytic CO₂RR activity of free‐standing Cu and Cu‐M (M = Au, Ag, Pd) NCAs electrodes in a CO₂‐saturated 0.1 M KHCO₃ solution was assessed using the H‐type electrolytic cell. In the H‐type cell, the platinum sheet, Ag/AgCl electrode, and proton exchange membrane served as the counter electrode, reference electrode, and separator, respectively. Compared to pristine Cu NCAs, the LSV curves showed that Cu‐M (M = Au, Ag, Pd) NCAs exhibited higher current densities (Figure S3A, Supporting Information). Specifically, Cu‐Pd NCAs demonstrated the highest current density at lower potentials (−0.4 ≈ −1.0 V) (vs RHE), indicating enhanced CO₂RR activity. However, its current density slightly decreased compared to Cu‐Au and Cu‐Ag NCAs electrodes at higher potentials (< −1.0 V). Additionally, the Cu‐Au and Cu‐Ag NCAs electrodes displayed similar current densities at low potentials. Nevertheless, as the cathodic potential increases, the current density of Cu‐Au NCAs significantly outperforms the other two electrodes. Subsequently, the electrochemical kinetic characteristics of the electrodes were assessed using the Tafel slope curves. Notably, all Cu‐M NCAs exhibited a smaller Tafel slope with faster reaction kinetics than pristine Cu NCAs within the potential range of the CO₂RR rate‐determining step (Figure S3B, Supporting Information).

Furthermore, double‐layer capacitance (C_dl) was measured according to cyclic voltammetry curves within the on‐Faraday region to analyze how the loading of the second active metal affected the electrochemical active surface area (ECSA) of the Cu NCAs (Figure S3C, Supporting Information). The Cu‐Au NCAs exhibited the highest double‐layer capacitance of 0.00210 mF cm⁻², followed by Cu‐Ag NCAs (0.00116 mF cm⁻²). In contrast, Cu‐Pd NCAs (0.000814 mF cm⁻²) showed only a slightly higher value than the pristine Cu NCAs (0.000692 mF cm⁻²). These results suggest that the enhanced CO₂RR catalytic activity of Cu‐M electrodes, compared to the pristine Cu NCAs, can be partly attributed to the increase in redox active sites. The electrochemical impedance spectroscopy (EIS) analysis reveals that Cu‐M NCAs have decreased charge transfer resistance compared to pristine Cu NCAs. Typically, Cu‐Au NCAs presented remarkable charge transfer resistance, indicating their more efficient charge transfer during the CO₂RR catalytic process (Figure S3D, Supporting Information).

Additionally, we evaluated Cu and Cu‐M NCAs electrodes for their selectivity at potentials ranging from −0.9 to −1.2 V due to the generation efficiency of higher CO₂RR products (the corresponding current densities were given in Table S1, Supporting Information). As expected, the pristine Cu NCAs showed high activity for the competitive HER compared to CO₂RR, with most of the current density contributing to H₂ production. The Cu NCAs had an average FE for HER of ≈71.3% at potentials in the range of − 0.9 to − 1.2 V; however, its total FE for CO₂RR remained below ≈23.7% (Figure 4A). It must be noted that this value could be achieved only after increasing the reduction potential to −1.2 V. This result demonstrated the higher reaction energy barrier of CO₂RR on the pristine Cu NCAs interface due to its main exposed facet of Cu (200), which exhibits high intrinsic HER activity and is almost inert for CO₂RR.^{[ 20 ]} The FE of Cu‐Au NCAs for HER decreased to ≈26.6 ± 3.4% after loading Au metal, while the FE for CO₂RR drastically increased to 63.0 ± 2.7% under similar conditions. The main C₂ product, C₂H₄, was generated, and its FE increased from 30% to 39.6% with increasing electroreduction potential; meanwhile, the FE of the C₁ product decreased from 32.5% to 23.4% (Figure 4B). Furthermore, a small amount of saturated hydrocarbon product (C₂H₆) was formed at −1.2 V. The total FE of the C₂ product was ≈43.2%, while the partial current density for the characteristic C₂H₄ product was ≈−8.47 mA cm⁻² (Figure S4A, Supporting Information). The high selectivity of Cu‐Au tandem interfaces to C₂ products under the large reduction potential is likely due to the enhanced reduction environment, which facilitates C‐C coupling in carbene intermediates.^{[ 21 ]} The main product for Cu‐Ag NCAs was found to be CO, with an FE of nearly 48% and a partial current density of −4.50 mA cm⁻² at −1.0 V; besides, ≈13% HCOOH was also produced (Figure 4C; Figure S4C, Supporting Information). On increasing the potential to −1.1 and −1.2 V, nearly 5% of C₂H₄ was generated, indicating that C₂ production might require higher overpotentials. Notably, the Cu‐Pd interface exhibited high selectivity for HCOOH within the broad reduction potential range, with FE achieving ≈46.5 ± 4.2%. At a potential of −1.1 V, the FE for HCOOH peaked at 50.4%, with a partial current density of −6.2 mA cm^{− 2} (Figure 4D; Figure S4D, Supporting Information). Additionally, the Cu‐Pd interface facilitated the formation of a small amount of C₂H₄, with its FE increasing from 2.5% to 7.2% as the potential varied from −0.9 to −1.2 V, demonstrating a higher efficiency of C₂H₄ production at higher potentials. Moreover, a small amount of CH₄ was generated at −1.2 V, although the mechanism of its formation remains unclear.

Figure 4.

Electrocatalytic performance for CO₂RR of Cu‐based electrodes within the 0.1 m KHCO₃ solution under CO₂ saturation: Faradaic efficiencies (FE) plots for (A) Cu, (B) Cu‐Au, (C) Cu‐Ag, and (D) Cu‐Pd NCAs to diverse CO₂RR products (Cu: typical H₂ product; Cu‐Au: typical C₂H₄ product; Cu‐Ag: typical CO product; Cu‐Pd: typical HCOOH product; (E) (A) Cu‐Au, (B) Cu‐Ag, and (C) Cu‐Pd NCAs stability at −1.0 V (vs RHE) within 24 h.

Additionally, the stability of Cu‐M (M = Au, Ag, Pd) electrodes was examined through chronoamperometry (i‐t) at the −1.0 V potential for 24 h. The current densities of Cu‐Au, Cu‐Ag, and Cu‐Pd were stable at −7.8, −8.1, and −8.4 mA cm⁻², and the current retention values during this process were ≈93.9%, ≈88.9%, and ≈91.3%, respectively (Figure 4E). Notably, after durability testing, Cu‐Au, Cu‐Ag, and Cu‐Pd NCAs maintained similar Faradaic efficiencies for their characteristic products: C₂H₄ (41.1% at 1.2 V), CO (50.1% at 1.0 V), and HCOOH (49.4% at 1.1 V), respectively. The stability of the Cu‐M tandem interface during long‐term CO₂RR electrocatalytic processes was evaluated by characterizing Cu‐M NCAs for their composition and surface structure following the 24‐h stability test (ADT). As depicted by XRD and Raman spectra, no long‐ or short‐range structural transition was detected in the prolonged CO₂RR electrocatalytic process (Figures S5 and S6, Supporting Information). Moreover, the CuO_x species levels among the Cu‐M NCAs following ADT remained largely consistent with those of the fresh samples, exhibiting a slight increase in Cu‐Au and Cu‐Pd NCAs and a slight decrease in Cu‐Ag NCAs (Figure S7, Supporting Information). These results suggest that different metal tandems efficiently modulated the selectivity and activity for the CO₂RR of Cu‐based interfaces while maintaining excellent stability.

2.3. Mechanism Analysis

To gain a comprehensive understanding of the tuning effects of diverse Cu‐M bimetallic tandem interfaces on the selectivity and activity for CO₂RR, we observed the active sites and reaction intermediates during the catalytic processes by utilizing in situ Raman spectroscopy and ATR‐FTIR spectroscopy under electrochemical test conditions. In the in situ Raman spectra for the Cu‐Au interfaces, characteristic bands corresponding to *CO²⁻ (≈1302 and 1575 cm⁻¹) and *COOH (≈1338 cm⁻¹) were observed across a wide range of reduction potentials, along with the *OCHO (≈1442 cm⁻¹) stretching vibration signal and H₂O (1602 cm⁻¹) bending vibration signal (middle panel of Figure 5A). Notably, *OCHO, a key intermediate in HCOOH formation, was detected only within a lower range (−0.2 to −0.6 V) and disappeared at higher potentials. Moreover, the limited adsorption and/or instability of the *OCHO intermediate on the Cu‐Au interface, particularly at higher potentials (−0.9 to −1.2 V), contribute to its reduced catalytic selectivity toward HCOOH production.^{[ 22 ]} In contrast, a persistent and relatively strong *COOH signal at 1338 cm⁻¹ and a pronounced *CO peak at 2062 cm⁻¹ were detected at −0.2 V. These signals remained stable as the potential increased but disappeared at −0.8 V (right panel in Figure 5A). Simultaneously, a Cu–CO band was observed at 298 cm⁻¹ over a broad potential window (−0.2 to −1.2 V) (left panel of Figure 5A). These spectral characteristics indicate that *COOH can be effectively reduced to *CO on the Cu–Au interface, leading to high coverage of Cu–CO species and thus promoting *OCCO coupling with C₂ product formation.^{[ 23 ]} Correspondingly, as the reduction potential reached −0.6 V, the *CH₂CHO signal, a key intermediate for C₂ products, was observed at 530 cm⁻¹. Notably, this signal remained stable and persistent at higher potentials, emphasizing that the bimetallic Cu‐Au tandem interfaces efficiently facilitate the production of C₂H₄.^{[ 24 ]} Here, the in situ ATR‐FTIR spectra of Cu‐Au revealed a strong R‐O* peak at 1153 cm⁻¹ at lower reduction potentials, which disappeared after −0.6 V. Beyond this potential, the bands corresponding to *COOH/*HCOO (≈1380–1398 cm⁻¹), H₂O (≈1640 cm⁻¹), and CO (≈2060 cm⁻¹) appeared, along with a pronounced inverted peak at ≈1239 cm⁻¹, which is associated with the surface reduction in the silicon reflective crystal (SiO_x) (Figure 5B).^{[ 25 ]}It is observed that the *CO*CO signal appeared at low potentials and persisted across a wide range, emphasizing the remarkable capability of the Cu‐Au interface in promoting C‐C coupling.^{[ 26 ]} As illustrated in Figure 5C, these observations underscore the high selectivity of the Cu‐Au tandem interface for C₂H₄.

Figure 5.

In situ electrochemical spectroscopy for Cu‐based catalytic electrodes within CO₂‐saturated 0.1 M KHCO₃ and possible reaction pathways for typical products obtained based on critical intermediates of CO₂RR: in situ Raman spectra for (A) Cu‐Au, (D) Cu‐Ag, and (G) Cu‐Pd NCAs obtained under diverse potentials from OCP to −1.2 V; in situ ATR‐FTIR spectra for (B) Cu‐Au, (E) Cu‐Ag, and (H) Cu‐Pd NCAs obtained under diverse potentials from OCP to −1.2 V (vs RHE); probable CO₂RR reaction pathways for typical products of C₂H₄, CO, HCOOH for (C) Cu‐Au, (F) Cu‐Ag, and (I) Cu‐Pd NCAs.

The in situ Raman spectra for Cu‐Ag interfaces exhibited close peaks to Cu‐Au before −0.6 V, although with varying peak intensities. The notable difference was that the intensity of the *CO peak (≈2060 cm⁻¹) significantly increased before −0.2 V (right panel of Figure 5D), which rapidly diminished as the potential exceeded ‐0.6 V. Besides, the Cu‐CO signal was also evident at low potentials but became nearly negligible beyond −0.6 V (left panel of Figure 5D). These observations suggest that the Cu‐Ag interface has considerably weaker *CO adsorption compared to Cu‐Au. Consequently, its desorption is facilitated, rather than further participation in C‐C coupling, favoring the direct generation of CO as the primary product.^{[ 27 ]} Nevertheless, a weaker peak signal of *OCHO emerges in the mid‐wavenumber region (middle panel of Figure 5D), reflecting its reduced selectivity for HCOOH. Additionally, a *CH₂CHO peak was detected, which attained its maximum at −1.1 V and then rapidly disappeared. This phenomenon is associated with the generation of a small amount of C₂H₄ at high potentials on Cu‐Ag. The in situ ATR‐FTIR spectra for Cu‐Ag (Figure 5E) revealed a *CO*CO signal that was limited to a low potential range (−0.1 to −0.3 V). However, the limited persistence and weak intensity of this signal indicate that the Cu‐Ag interface did not favor C‐C coupling. Instead, these findings imply that the Cu‐Ag interface promotes *CO desorption, resulting in high CO selectivity (Figure 5F).

Compared to Cu‐Au and Cu‐Ag, the Raman signal of Cu‐Pd exhibited the most pronounced peak intensity of *OCHO, corresponding to its higher selectivity for HCOOH (middle panel of Figure 5G). Moreover, the signal of *CO was weaker than the Cu‐Au and Cu‐Ag signals and is associated with its significantly lower CO selectivity. Furthermore, the signal of *CO adsorbed on Cu sites was detected within the range of −0.6 to −0.9 V. The signal disappeared at higher potentials, indicating the adsorption of a portion of the generated *CO onto the Cu surface to achieve C‐C coupling (right panel in Figure 5G). Additionally, the *CH₂CHO peak at 530 cm⁻¹ within −0.2 to −0.8 V further confirmed this trend, indicating that C‐C coupling is partly triggered onto the Cu‐Pd surface, leading to C₂ product generation (left panel of Figure 5G). This is linked to the generation of 2.5% to 7.1% C₂H₄ at the Cu‐Pd interface under strong reducing conditions. The ATR‐FTIR spectra revealed the signal of *CO*CO at 1476 cm⁻¹ within the low potential range (−0.1 to −0.3 V). Unlike the other two Cu‐M electrodes, the *HCOO/*COOH signal, observed in the broad wavenumber range of ≈1338 – 1425 cm⁻¹, is most prominent on the Cu‐Pd interface (Figure 5H). Thus, this finding demonstrates the typical behavior of Cu‐Pd NCAs, favoring the generation of HCOOH products (Figure 5I).

3. Conclusion

In conclusion, this study presents a highly efficient approach for enhancing the selectivity and activity of Cu‐based catalysts for the CO₂ reduction reaction (CO₂RR), particularly on the relatively inert Cu facets that are susceptible to the competitive hydrogen evolution reaction (HER). Furthermore, we successfully modulated the adsorption behaviors of critical intermediates for CO₂RR by constructing a series of Cu‐M (M = Au, Ag, Pd) bimetallic tandem interfaces onto Cu nanocolumn arrays (Cu NCAs), and Cu (200) facets were predominantly exposed, thereby improving both the CO₂RR activity and significantly reducing HER. Notably, the Cu‐Au, Cu‐Ag, and Cu‐Pd NCAs exhibited excellent selectivity for C₂H₄, CO, and HCOOH production, thereby achieving optimized faradaic efficiencies of ≈43.2%, ≈48.0%, and ≈50.7%, respectively. According to the in situ spectroscopic analysis, Cu‐Au favored *COOH and *CO adsorption for C‐C coupling and C₂H₄ formation, Cu‐Ag promoted *CO desorption for CO generation, and Cu‐Pd facilitated *OCHO formation for HCOOH production. These findings emphasize the potential of bimetallic tandem strategies in directing the reaction pathways for selective CO₂RR product formation. Overall, this study offers a versatile strategy for designing high‐performance bimetallic Cu‐based catalytic electrodes and provides valuable insights for developing more practical and efficient CO₂RR catalysts based on commercially available Cu substrates, and this interfacial engineering principle might be also universally applicable to other CO₂RR‐active metal interfaces (e.g., Zn, Sn, Bi).

4. Experimental Section

Preparation of Cu‐M (M = Au, Ag, Pd) NCAs

According to the previous studies,^{[ 28 ]} Cu nanocolumn arrays (Cu NCAs) were prepared using physical vapor deposition (PVD) at an oblique angle deposition (OAD) configuration technique, which mainly exposed Cu (200) facets. The shape of OAD accelerates atom deposition within early‐generated nuclei while producing shadows behind the vapor flux, resulting in the growth of nanopillar arrays (NCAs) on the film in an inclined form. Electron beam evaporation was conducted at a working pressure of nearly 10⁻⁶ mbar for deposition, with 1‐µm‐thick nanostructured Cu thin films being deposited on Cu foil at a 70° oblique vertex angle. The actual layer thickness was determined by analyzing the scanning electron microscope images.

Subsequently, Cu NCAs were deposited on Cu foil under the same conditions using the magnetron sputtering method. The precursors were immersed in 1.45 mm HAuCl₄, 1.77 mM AgNO_3, and 0.25 mM K₂PdCl₄ solutions for 10 min and subsequently calcined in Ar at 150 °C for 30 min to obtain Cu‐Au, Cu‐Ag, and Cu‐Pd NCAs electrodes. Remarkably, the pristine Cu NCAs without deposition of any other metal species were also processed under a similar thermal treatment (calcined in Ar at 150 °C for 30 min). Besides, the second metal precursor concentrations were specifically optimized to: 1) balance interfacial atomic ratios by achieving moderate secondary metal (Au/Ag/Pd) loading on Cu substrates – avoiding both insufficient active sites at low concentrations and excessive surface coverage impairing CO₂ activation; 2) enable tandem catalysis through preserving independent functionalities of Cu and M sites while promoting interfacial electronic interactions for cooperative effects.

Characterization

The electrode surface elemental distribution and morphology were analyzed using a scanning electron microscope (SEM, ZEISS Gemini SEM 300) along with the energy‐dispersive X‐ray spectroscopy (EDS) system. Moreover, the sample microstructure was examined at an accelerating voltage of 120 kV by utilizing transmission electron microscopy (TEM, JEOL JEM‐2100F), coupled with EDS. The sample composition and crystal structure were further analyzed using X‐ray diffraction (XRD, Bruker D8 Advance) at 40 kV and 40 mA using Cu‐Kα radiation. The samples were scanned at 2θ = 10°‐90° at the rate of 0.02°/s based on Bragg‐Brentano geometry. Moreover, the Thermo Scientific K‐Alpha X‐ray photoelectron spectrometer was adopted for X‐ray photoelectron spectroscopy.

Electrochemical Measurements

The electrochemical measurements were performed using the CHI760E electrochemical workstation in a sealed dual chamber, serving as the three‐electrode H‐type electrolytic cell. A proton exchange membrane (Nafion 117) was added to separate the two chambers in the H‐cell, and 18 mL of CO₂‐saturated 0.1 M KHCO₃ as the electrolyte was added to each chamber with Ag/AgCl and platinum foil electrodes being the reference and counter electrodes, respectively. The Cu or Cu‐M NCAs electrode served as the working electrode. All electrochemical potentials were referenced to the reversible hydrogen electrode (RHE), and the potential was computed as follows: E (vs RHE) = E (vs Ag/AgCl) + 0.20 V + 0.0591 × pH.

CO₂RR Product Analysis—Liquid Phase Product Analysis

The liquid‐phase products were analyzed through proton nuclear magnetic resonance (1H NMR) spectroscopy using a Bruker AVANCE HD III 400 spectrometer. Herein, the electrolyte solution (0.85 mL) was mixed in 0.1 mL deuterated water (D₂O) for field locking and 0.05 mL dimethyl sulfoxide (DMSO, ≈10.56 mM) as an endogenous reference. The 1H NMR spectra were recorded underwater suppression with the pre‐saturation sequence, and products were quantitatively analyzed via an endogenous reference method. It should be mentioned that the formic acid product generated by CO₂RR existed as formate in alkaline electrolytes, which was simply represented as HCOOH.

CO₂RR Product Analysis—Gas Phase Product Analysis

The gas phase products were measured through a gas chromatograph (GC7200). A thermal conductivity detector (TCD) was adopted for quantifying hydrogen, while a flame ionization detector (FID) was applied to quantifying carbon monoxide, methane, ethane, and ethylene. The output of gas‐phase products was determined using an external standard method by drawing the standard curve of standard gases with customized concentrations (H₂, CO, CH₄, C₂H₄, C₂H₆).

In Situ Raman Measurements

Raman spectra were obtained using an Invia Raman spectrometer with an L × 50 eyepiece and a laser wavelength of 633 nm. The in situ Raman measurements were conducted using an H‐type three‐electrode electrochemical cell (Goshilian C031). Notably, the three‐electrode system consists of a graphite electrode, an Ag/AgCl electrode, and the Cu‐M NCAs electrode, as the counter, reference, and working electrodes, respectively. The electrolyte was circulated using the peristaltic pump. The measurement range is ≈200 to 3200 cm⁻¹, with an exposure time of 10 s and 5 integration times. Furthermore, in situ Raman measurements were conducted with CO₂‐saturated 0.1 m KHCO₃ as the electrolyte at a potential range of 0 to −1.2 V (vs RHE).

In Situ ATR‐FTIR Measurements

The testing was done using a Bruker Vertex 70 Fourier transform infrared spectrometer, equipped with a liquid nitrogen‐cooled mercury cadmium tellurium (MCT) detector and attenuated total reflection accessories (Pike, Vee Max III). The experiment was conducted in a 0.1 M KHCO₃ solution saturated with CO₂, with CO₂ being continuously injected throughout the experiment. Here, platinum foil and Ag/AgCl were used as counter and reference electrodes, respectively. For the working electrode, the samples from Cu‐M NCAs were scraped off, mixed with an adhesive to form a slurry, and then dripped onto a silicon crystal coated with an Au film. The experimental conditions were as follows: a resolution of 4 cm⁻¹, a sample scanning duration of 120 s, and a measurement range from 400 to 5000 cm ⁻¹. The in situ infrared measurements were performed at intervals of 0.1 V within the potential range of 0 to −1.2 V (vs RHE).

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.