English

• Carbon dioxide (CO₂) is a major contributor to global warming [1,2], which causes serious environmental issues such as ocean acidification and wildfires [3,4]. Electrocatalytic CO₂ reduction reaction (CO₂RR) driven by renewable energy is a potential approach to reduce CO₂ emissions [5–8]. The conversion of CO₂ into important fuels and chemicals by electrocatalytic reduction with great significance, which is conducive to the development of industrial decarbonization. Although the CO₂RR can produce a variety of products such as carbon monoxide (CO), formic acid, methane (CH₄), ethylene (C₂H₄), ethanol, and acetic acid, it is still extremely difficult to achieve the high efficiency and selectivity towards high value C₂ products with high energy density [9–14]. Cuprous oxide (Cu₂O) is one of the most promising catalysts for electrochemical conversion of CO₂ into value-added C₂ products [15–18]. Among these C₂ products, C₂H₄ is extremely attractive as an important raw material in the chemical industry [19]. We therefore urgently need to develop an optimal catalyst to increase the selectivity of C₂H₄ production. The rate-determining step (RDS) for the formation of C₂H₄ involves C–C coupling between adsorbed *CO or coupling of *CO hydrogenated derivatives [20–22]. However, the symmetric binding of atop-type *CO (*CO_atop) occurs on conventional low-index crystal planes of Cu₂O due to weak *CO adsorption, leading to high-energy barrier C–C coupling reactions and hindering the CO₂-to-C₂ conversion [23].

  By adjusting the crystal planes orientation of Cu₂O, the surface energy and affinity of reaction intermediates can be effectively altered, thereby influencing the adsorption and coupling of *CO species and reduction pathway [24–26]. In particular, the design of high-index crystal planes with significant amounts of atomic steps, edges, and unsaturated coordinated sites, provides more opportunities for the preparation of catalysts with high activity and selectivity [27]. High-index crystal planes, such as (311) or (522) [28], typically exhibit higher *CO binding energies because of the higher surface energy and coordination saturation [29,30]. This results in the formation of more stable adsorbed *CO species compared to *CO_atop thereby improving the C₂ products selectivity [31]. For instance, Han et al. synthesized with a controllable Cu₂O (311) crystal plane and its C₂ selectivity was about 37.5% at −1.4 V vs. RHE, which was doubled compared with 16% for the low-index crystal planes [27]. The highly chemically active high index crystal plane Cu₂O (311) is energetically favorable to reduce and simultaneous evolve Cu₂O fine grains, which benefits charge transfer and increases the local pH by confining OH⁻, resulting in high selectivity towards C₂ products.

  A conventional approach to tailoring crystal planes involves the introduction of extrinsic additives, such as surfactants and reductants, during catalyst synthesis [32,33]. However, the removal of residual surfactant requires complex procedures and may result in catalyst poisoning [34]. Although use of a reductant, e.g., NH₂OH·HCl, can address the adverse effects of residual surfactant and expose the desired crystal planes [35], the design of such synthetic protocols remains complex and requires massive trials. Recently, a self-selective strategy involving the introduction of reactants during catalyst synthesis was proposed [36]. Compared with the introduction of extrinsic additives, the proposed self-selective strategy can address the issues of additive residue and need for massive trials. For example, using CO₂ as an inducer to stabilize CO₂ favored crystal planes during self-selective catalyst preparation, thereby promoting the CO₂RR reaction [36,37]. Inspired by this strategy, we believe that the self-selective CO-favored crystal planes can be induced by introducing CO during the synthesis of Cu₂O. This self-selective strategy tends to stabilize the planes that adsorb *CO intermediates most strongly, since the adsorption of the CO molecules reduces the relative surface energies of these planes. This aims to form other types of with stronger adsorbed *CO species besides atop-type, overcoming the limitations of symmetric coupling based on *CO_atop to promote asymmetric C–C coupling.

  Here, we developed a novel Cu₂O electrocatalyst with high-index (311) crystal planes (named CO–Cu₂O) synthesized by a facile self-selective CO-induced strategy using CO as an inducer. In situ attenuated total reflection surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS) and density functional theory (DFT) calculations revealed that both atop-type and hollow-type adsorption (*CO_hollow) of *CO species occurred on CO–Cu₂O. The asymmetric C–C coupling between *CO_atop and *CO_hollow is kinetically favorable, resulting in a lower energy barrier for CO₂RR compared to the symmetric coupling of *CO_atop. Consequently, the Faradaic efficiency of the C₂ products (FE_C2) generated with CO–Cu₂O was increased by 100% than that of pristine Cu₂O.

  In typical synthesis, the pristine Cu₂O nanoparticles were synthesized under alkaline conditions using ascorbic acid as a reductant [24], while CO–Cu₂O nanoparticles were prepared following the same procedure under a CO atmosphere. Further details on the experimental section, including materials source and preparation method, can be found in Supporting information. Scanning electron microscopy (SEM) images revealed that the prepared Cu₂O and CO–Cu₂O nanoparticles exhibited cubic structural morphology with a diameter of about 600 nm (Figs. S1 and S2 in Supporting information). Based on the results of high-resolution transmission electron microscopy (HR-TEM) images, the interplanar spacings have been reduced from 0.215 nm in pristine Cu₂O to 0.130 nm in CO–Cu₂O (Figs. 1a and b). This indicates that a significant number of (311) crystal planes appear in CO–Cu₂O through the self-selective CO-induced strategy. Atomic force microscopy (AFM) images (Figs. 1c and d, Fig. S3 in Supporting information) revealed that the roughness of CO–Cu₂O was significantly higher than that of pristine Cu₂O because the (311) crystal planes possessing abundant step- and kink- sites [38]. The line profiles (Figs. 1e and f) demonstrated an increase in the height variation on the CO–Cu₂O surface compared to that on the Cu₂O surface. The average heights of the CO–Cu₂O and Cu₂O surfaces, determined via line profile extraction analysis, were approximately 6.11 nm and a mere 0.76 nm, respectively. These particulate edges represented undercoordinated sites associated with crystal microfacets, which are considered to play a crucial role in structure-activity relationship [39,40].

  Figure 1

  

  Figure 1.  Morphologies and structural characterization of the prepared catalysts. HR-TEM images of (a) Cu₂O and (b) CO–Cu₂O. AFM images of (c) Cu₂O smooth surface and (d) CO–Cu₂O, showing nanoparticle formation after CO induction. Extracted line profiles for (e) Cu₂O and (f) CO–Cu₂O.

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. S4 (Supporting information), the X-ray diffraction (XRD) spectra demonstrated that the peaks attributed to Cu₂O conformed to those of standard Cu₂O (PDF #99–0041). After self-selective induction, the diffraction peaks of CO–Cu₂O are similar to Cu₂O, confirming that no apparent difference between Cu₂O and CO–Cu₂O catalysts in the crystal structure. Besides, the grazing incidence XRD analysis indicated that the (311)/(100) crystal plane intensity ratio in CO–Cu₂O (0.68) was increased by 51% compared to that in pristine Cu₂O with only 0.45 (Fig. S5 and Table S1 in Supporting information), indicating the increase of the preferential orientation of the crystal planes after the induction. The X-ray photoelectron spectroscopy (XPS) spectra of Cu₂O and CO–Cu₂O exhibited obvious peaks at 932.1 and 952.0 eV (Figs. S6 and S7 in Supporting information), being attributed to the Cu 2p_1/2 and Cu 2p_3/2 of Cu⁺, respectively. Additionally, XPS was employed to assess the differences in the O 1s spectra of various Cu₂O catalysts (Fig. S8 in Supporting information). The O 1s spectra were decomposed into peaks at 529.6 and 531.4 eV, corresponding to surface-absorbed oxygen (O_ads) and crystal lattice oxygen species (O_latt), respectively. The peak ratio between O_latt and O_ads served as an indicator of the quantity of surface oxygen vacancies (O_v) [41]. Compared to Cu₂O, CO–Cu₂O enriched with O_v could stabilize Cu⁺ and thereby strengthen *CO adsorption to promote C–C coupling [42].

  To further elucidate the chemical state and coordination environment of Cu in Cu₂O and CO–Cu₂O at the atomic level, the X-ray near edge spectroscopy (XANES) profiles at the Cu K-edge of the prepared Cu₂O and CO–Cu₂O catalysts were compared with those of commercial Cu foil, Cu₂O, and CuO as reference materials (Fig. 2a). The energy positions of the prepared Cu₂O and CO–Cu₂O catalysts were similar to that of reference Cu₂O, confirming that the valence state of Cu was mainly +1. The peak observed at 1.4 Å (without phase correction) in the R-space (k³-weighted) extended X-ray fine structure (EXAFS) spectra could be attributed to coordination between Cu and O atoms (Fig. 2b). By contrast, the peak at 2.6 Å represented Cu-Cu interactions. Similarly, wavelet transform (WT)-EXAFS was employed to further visualize the coexistence of Cu-O and Cu-Cu coordination in Cu₂O and CO–Cu₂O (Figs. 2c and d). The fitting results (Figs. 2e and f, Fig. S9 in Supporting information) showed that the Cu-Cu coordination number in CO–Cu₂O were about 8.96, which was higher than that of Cu₂O at 8.15 (Table S2 in Supporting information). The CO–Cu₂O shows a higher mean Cu-Cu bond length of approximately 3.05 Å than Cu₂O (2.99 Å), which is due to the tensile strain effect induced by the (311) crystal plane. Notably, the ratio of Cu-O to Cu-Cu coordination strengths in CO–Cu₂O was slightly greater than that in commercial Cu₂O, leading to increased structural disorder and potentially more undercoordinated sites [43].

  Figure 2

  

  Figure 2.  X-ray fine structure analysis. (a) Normalized Cu K-edge XANES spectra. (b) k³-weighted Fourier transform EXAFS spectra without phase correction. Cu K-edge WT-EXAFS contour plots of (c) Cu₂O and (d) CO–Cu₂O. Fitting of k space and R space of Cu in (e) Cu₂O and (f) CO–Cu₂O.

  DownLoad: Full-Size Img PowerPoint

  To determine the effect of adsorbed *CO species on the high-index crystal planes of CO–Cu₂O, in situ ATR-SEIRAS was carried out (Figs. 3a and b), and peaks were classified as the corresponding intermediates according to reported studies (Fig. 3c). The spectra were collected in CO₂-saturated 0.5 mol/L KHCO₃ solution at different potentials from 0 V to −1.0 V vs. RHE. The peak intensity is expressed in terms of absorbance and the positive peak represents product formation. In particular, two peaks appeared at approximately 2050–1820 cm⁻¹ for CO–Cu₂O, which were associated with *CO_atop and *CO_hollow on the surface, respectively [44]. By comparison, the spectrum of Cu₂O mainly exhibited the *CO_atop peak in the same potential region. Two *CO_atop peaks were clearly recognized in the spectra of CO–Cu₂O at potentials above −0.4 V vs. RHE. The *CO_atop peaks could be classified into high-frequency binding *CO (CO–HFB) and low-frequency binding *CO (CO-LFB) [45,46]. Prior research reported that CO-LFB is associated with C–C coupling and C₂H₄ production, while CO–HFB is typically related to the production of gaseous CO [47,48]. Notably, as the potential shifted to more negative values, the red shift observed for the CO stretching bands in the CO–Cu₂O spectra may have been attributed to the Stark effect [49]. This finding suggested an increase in the adsorption of *CO_atop on the surface, potentially fostering further protonation or C–C coupling reactions [50]. Hence, we plotted the integrated *CO_atop peak area against the applied potentials (Figs. 3d and e).

  Figure 3

  

  Figure 3.  In situ monitoring of adsorbed *CO species. In situ ATR-SEIRAS spectra of (a) Cu₂O and (b) CO–Cu₂O during timed current scans from 0 V to −1.0 V vs. RHE. (c) The adsorption behavior of species within 2200–1600 cm⁻¹ ranges, with Cu, O, C, and H represented by orange, red, gray, and white, respectively. The integrated area of *CO_atop for (d) Cu₂O and (e) CO–Cu₂O at various wavenumbers is also shown. The experimental conditions included CO₂-saturated 0.5 mol/L KHCO₃ as the electrolyte. Peak positions were determined based on the maximum intensity of the CO stretching band. (f) Electrochemical CO stripping voltammetry curves of Cu₂O and CO–Cu₂O in saturated N₂ or CO atmospheres.

  DownLoad: Full-Size Img PowerPoint

  As the applied potential increased, the integrated area of CO–HFB on the Cu₂O surface decreased, and no CO-LFB appeared within the applied potential range. Meanwhile, CO-LFB was observed on CO–Cu₂O when the applied potential exceeds −0.4 V vs. RHE and the integration area reached a maximum at −0.7 V vs. RHE. Compared with single CO–HFB on Cu₂O, CO-LFB newly appeared on CO–Cu₂O with a more negative potential. Notably, the absence of CO-LFB signals was detected due to low C–C coupling activity at lower potentials. At higher potentials, the integrated area for CO–Cu₂O dropped owing to *CO consumption for C–C coupling. In addition, new bands emerged on the CO–Cu₂O surface at 1835–1780 cm⁻¹ when the applied potential was above −0.4 V vs. RHE, corresponding to *CO_hollow species [44]. These results suggested that the electronic structure of the catalyst was changed via CO induction, which triggered more *CO adsorption configurations. As the potential was further shifted to more negative values, the quantity of adsorbed *CO_hollow progressively augmented. Benefiting from the (311) crystal plane in CO–Cu₂O with stronger affinity for *CO, thus CO–Cu₂O catalyst has higher *CO_atop and *CO_hollow coverage at high overpotentials, further promoting asymmetric C–C coupling. Besides, the peak at 1720 cm⁻¹ can be assigned to *CHO (Fig. S10 in Supporting information), which is well accepted as the starting adsorption mode of intermediate for C–C coupling [14,51]. However, no significant *CHO adsorption peak was detected for Cu₂O at all potentials, indicating a high protonation energy barrier affecting subsequent C–C coupling. Meanwhile, an enhanced water (*H-O-H) peak at approximately 1620 cm⁻¹ was observed on Cu₂O, which may have led to a blocking effect, facilitating substitution of *H for *CO [52]. Notably, the *H coverage on Cu₂O is opposite to that on CO–Cu₂O, which is associated with higher *CO coverage, resulting in *CO occupying more active sites. In addition, we then also investigated the CO adsorption ability of Cu₂O and CO–Cu₂O via electrochemical CO stripping voltammetry [53,54]. With the activation process going on, the intensity of oxidic peak corresponding to CO adsorbed on CO–Cu₂O was significantly increased compared to that for Cu₂O (Fig. 3f), indicating stronger CO binding capacity and higher CO coverage that attributed to the increased of electrochemical adsorption of CO [49], which was consistent with the IR spectral analysis. The above results demonstrated that CO–Cu₂O possessed an asymmetric *CO coupling mode and higher *CO coverage than Cu₂O, promoting asymmetric C–C coupling on nearby sites under electrochemical reduction conditions.

  To verify the above results, we performed CO₂RR for Cu₂O and CO–Cu₂O. The electrochemical CO₂RR was performed using a flow cell with 0.5 mol/L KHCO₃ as electrolyte. The gas chromatography and ¹H nuclear magnetic resonance spectroscopy (¹H NMR) were employed to quantify the products in the gas and liquid phases, respectively. The specific relationships among different products determined from the NMR spectra are presented in Fig. S11 (Supporting information). Linear sweep voltammetry (LSV) images of the Cu₂O and CO–Cu₂O electrodes were acquired in CO₂ or N₂-saturated atmospheres electrolyte (Fig. 4a). Under CO₂ and N₂ atmospheres, Cu₂O catalysts exhibited relatively lower current density (j) values at the same applied potentials than CO–Cu₂O. The results suggested that the increment in high-index crystal planes substantially enhanced the CO₂RR activity of CO–Cu₂O. Figs. 4b and c present the selectivity of CO–Cu₂O under various imposed potentials, where the FE_C2 (62%) of CO–Cu₂O was nearly double that of Cu₂O (31%) at −1.2 V vs. RHE. For pristine Cu₂O, the main product was H₂ with nearly 30% FE, while the C₂H₄ selectivity promoted from 16.2% to 41.1% in CO–Cu₂O, whose value was 150% than pristine Cu₂O. Meanwhile, the partial current density of C₂ products generated with CO–Cu₂O reached 66 mA/cm² compared to 25 mA/cm² for pristine Cu₂O (Fig. 4d). Additionally, the electrochemical active surface areas (ECSA) of Cu₂O and CO–Cu₂O were measured (Fig. S12 in Supporting information). The ECSA of CO–Cu₂O was 65.8 cm², which was larger than that of Cu₂O (50.7 cm²). This finding indicated that CO–Cu₂O could provide more active sites than Cu₂O. The intrinsic activity of the catalysts was obtained by normalizing the current density to the ECSA [55]. The modified current densities indicated that CO–Cu₂O exhibited greater activity than Cu₂O for C₂ products formation and inhibited H₂ generation (Figs. 4e and f).

  Figure 4

  

  Figure 4.  CO₂-to-C₂ electroreduction performance. (a) Comparison of LSV for CO₂ and Ar-saturated electrolytes. Faradaic efficiency of each product generated with (b) Cu₂O and (c) CO–Cu₂O at various imposed potentials. (d) The partial current density of C₂ products at various potentials over Cu₂O and CO–Cu₂O. ECSA-normalized current density of different products of (e) Cu₂O and (f) CO–Cu₂O. (g) The FE_C2/FE_C1 ratios of Cu₂O and CO–Cu₂O under various imposed potentials. (h) Stability tests of CO–Cu₂O at −1.2 V vs. RHE in 0.5 mol/L KHCO₃.

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. 4g, CO–Cu₂O achieved a 2.35-fold increase in the FE_C2/FE_C1 ratio at −1.2 V vs. RHE compared to pristine Cu₂O. Moreover, CO–Cu₂O exhibited a lower FE_H2 than pristine Cu₂O at all potentials, and the FE_C1 of CO–Cu₂O did not significantly increase compared to that of pristine Cu₂O. This suggests that the hydrogen evolution reaction (HER) in CO–Cu₂O was effectively suppressed, and the increased in the amount of C₂ products could be attributed to the inhibition of *H conversion. This is consistent with in situ ATR-SEIRAS spectra observations of higher *CO and lower *H coverage, where elevated *CO is involved in the C–C coupling process to form C₂ products. Additionally, the long-term durability performance of CO–Cu₂O catalyst reveal that it maintained a relatively stable current density and C₂H₄ selectivity during the whole operation period (Fig. 4h). Combined with CO₂RR performance, the study findings indicated that *CO_hollow became dominant at higher cathodic potential, contributing to asymmetric *CO_atop and *CO_hollow coupling that enhanced C₂ products formation [50]. As shown in Fig. S13 (Supporting information), electrochemical impedance spectroscopy (EIS) revealed a lower charge transfer resistance (R_ct) of CO–Cu₂O at the open circuit potential compared to Cu₂O, which improved the charge transfer kinetics of the developed catalyst compared to pristine Cu₂O. However, CO₂ reduction typically occurs at much more negative applied potentials, and the arcs in the Nyquist diagrams substantially decreased from Cu₂O to CO–Cu₂O, further demonstrating the reduced the transfer resistance of CO–Cu₂O compared to pristine Cu₂O [49]. Hence, it could be concluded that the presence of high-index crystal plane also had an improvement in electron transfer performance to boost the CO₂RR performance.

  Stabilizing the Cu⁺ content during the CO₂RR process is crucial for generating the corresponding *CO intermediate and final C₂ products [56], whereas presence of Cu⁺can promote *CO to *CHO and the adsorption of OH species, causing C₂ intermediates preferably to form the sp³ hybrid C atom and undergo asymmetric hydrogenation, leading to the enhancement of C–C coupling [57]. The SEM images revealed that the surface morphology of CO–Cu₂O after CO₂RR for 1 h at −1.2 V vs. RHE had better stability than Cu₂O (Fig. S14 in Supporting information). In addition, the XRD spectra indicated the generation of reduced metal Cu (PDF #04–0836) species on Cu₂O and CO–Cu₂O (Fig. 5a). The peak intensity of Cu (111) and Cu (200) on Cu₂O was significantly higher than that on CO–Cu₂O after the reaction, indicating that more Cu⁺ was reduced on Cu₂O. Similarly, the ex-situ TEM images and selected area electron diffraction patterns revealed the presence of metallic Cu and Cu⁺ in Cu₂O and CO–Cu₂O after the reaction (Fig. S15 in Supporting information), indicating that Cu⁺ underwent reduction. Moreover, the stable presence of the (311) plane of CO–Cu₂O is still observed after the reaction, indicating that (311) plane contributes to the formation of stable atomic clusters or structures during electroreduction and that this reconfiguration contributes to the fixation of the Cu⁺ valence state. Based on the XRD spectra, the Cu LMM spectra of Cu₂O and CO–Cu₂O after 1 h CO₂RR are examined to show the changing valence state of Cu. The Cu LMM Auger spectra (Figs. 5b and c) clarified that Cu⁺ (916.6 eV) and metallic Cu (918.5 eV) species co-existed on the Cu₂O and CO–Cu₂O catalysts. The results indicated that Cu⁺ species on CO–Cu₂O remained well stabilized, with only 3% conversion to metallic Cu, whereas approximately 23% of Cu⁺ in Cu₂O was converted to metallic Cu upon electroreduction. The above results indicate that the CO–Cu₂O catalyst can stabilize the surface morphology and Cu⁺ species, which is conducive to maintaining the activity of the electroreduction reaction.

  Figure 5

  

  Figure 5.  Cu valence changes in catalysts after CO₂ electroreduction. (a) XRD spectra of Cu₂O and CO–Cu₂O after 1 h of CO₂RR at −1.2 V vs. RHE. Cu LMM spectra of (b) Cu₂O and (c) CO–Cu₂O at initial states and those after 1 h of CO₂RR at −1.2 V vs. RHE.

  DownLoad: Full-Size Img PowerPoint

  DFT calculations are conducted to verify the possible asymmetric C–C coupling mechanism on Cu₂O (311) crystal plane (Fig. 6a and Fig. S16 in Supporting information). As shown in Fig. 6a, the CO₂ molecule initially adsorbs on the Cu site, undergoing hydrogenation to form the *COOH intermediate. This is then reduced to the *CO intermediate, which combines with protons to yield the *CHO intermediate. Subsequently, the adsorbed *CHO undergoes a similar process to form 2*CHO intermediate on the surface. This 2*CHO intermediate then undergoes dimerization to form the *OHCCHO intermediate, which is finally hydrogenated to produce C₂H₄. Intriguingly, the (311) crystal plane boosts the activation of Cu site and reaches the hollow-type adsorption of *CO, which is consistent with the monitoring results from in situ spectroscopies. Consequently, the energy barrier of their asymmetric coupling of *CO_atop and *CO_hollow decreases with 47.8% compared to the conventional symmetric coupling of *CO_atop.

  Figure 6

  

  Figure 6.  Mechanistic studies by DFT calculations. (a) The free energy diagrams for CO₂ reduction to different intermediates on the Cu₂O (100) and Cu₂O (311) surfaces at U = 0 V. DCD map for the *CO intermediate adsorbed on the (b) Cu₂O (100) and (c) Cu₂O (311) surfaces. (d) ELF of Cu₂O (100) and Cu₂O (311). The WF across the (e) Cu₂O (100) slab and (f) Cu₂O (311) slab vacuum interface.

  DownLoad: Full-Size Img PowerPoint

  The two critical steps in the CO₂RR pathway to C₂H₄ are the hydrogenation of *CO to form the *CHO intermediate and the coupling of *CHO [58]. In the former *CO → *CHO step, the Gibbs free energy change on the Cu₂O (311) surface (ΔG = 0.49 eV) is slightly lower than on the Cu₂O (100) surface (ΔG = 0.58 eV). This suggests that CO–Cu₂O with (311) is more easily protonated, which is consistent with the weaker *CHO that is observed in the infrared spectra. For the *CHO dimerization step, the uphill reaction energy for the formation of 2*CHO from *CHO is an RDS for the entire reaction pathway. In addition, *CO is the pivotal intermediate in the conversion of CO₂ to C₂ products [59,60], with the adsorption energy of *CO playing an important role in their reaction rates [61]. During the activation process, the decomposition of *COOH on Cu₂O (311) has a higher free energy compared to that on Cu₂O (100), indicating that the formation of *CO is more favorable on Cu₂O (311). As shown in Figs. S17 and S18 (Supporting information), the hollow site of the (311) crystal plane exhibited the most robust CO adsorption (ΔE = −1.57 eV), and strong *CO adsorption ensured adequate *CO coverage on its surface [62,63]. In this way, *CO occupied the catalytic active sites on the (311) crystal plane, which strengthened asymmetric C–C coupling and CO₂ to C₂ production.

  To better visualize the incorporation of the *CO intermediate with the catalysts, we calculated the differential charge density (DCD) of Cu₂O (100) and Cu₂O (311) crystal planes after *CO adsorption (Figs. 6b and c). Compared with the (100) crystal plane, the charge distributed around the Cu atom on the (311) crystal plane moved more toward the C atom side, indicating enhanced binding energy between *CO and the CO–Cu₂O surface. As shown in the electron localization function (ELF) topographical analysis in Fig. 6d, the (311) crystal plane displayed higher electron delocalization around the formed *CO adsorption sites than that of (100) crystal plane. This suggests that *CO adsorbed on the (311) surface favors the formation of a more stable adsorption configuration, resulting in asymmetric C–C coupling. Work function (WF) can serve as a characteristic quantity to evaluate the catalytic activity [64]. Correspondingly, the detailed WF revealed that the WF value in the (311) crystal plane was lower than that of the (100) crystal plane (Figs. 6e and f). The lower the WF value, the more easily electrons can leave the metal, leading to more efficient electron transfer. To summarize, Cu₂O (311) surface conducive to the electron transfer to the direction of CO₂ electroreduction, thus enhancing the activity and selectivity for the CO₂RR performance.

  In summary, the asymmetric coupling of atop-type and hollow-type adsorbed *CO species occurring on the constructed high-index crystal planes was crucial for enhancing the CO₂-to-C₂ conversion. An efficient catalyst CO–Cu₂O with Cu₂O (311) crystal planes was synthesized via the facile self-selective CO-induced strategy proposed in this work. The energy barrier of their asymmetric coupling of *CO_atop and *CO_hollow decreases with 47.8% compared to the conventional symmetric coupling of *CO_atop. Benefiting from the asymmetric coupling mechanism, the FE_C2 generated with CO–Cu₂O was increased by nearly 100% at −1.2 V vs. RHE compared to that with pristine Cu₂O. This study elucidates the mechanism of asymmetric C–C coupling facilitated on high-index crystal planes and offers a method to design efficient catalysts for CO₂-to-C₂ conversion.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Wei Peng: Writing – original draft, Visualization, Formal analysis. Yao Shen: Writing – review & editing, Writing – original draft, Funding acquisition. Xiaolin Yu: Software. Chenghang Zheng: Data curation. Xiao Zhang: Resources. Jingkai Zhao: Visualization. Jiexu Ye: Resources. Shihan Zhang: Writing – review & editing, Writing – original draft, Funding acquisition. Xiang Gao: Software, Funding acquisition, Formal analysis.

  Acknowledgments

  The authors thankfully acknowledge the financial support from the National Natural Science Foundation of China (Nos. U23A20677, 22022610 and 52400137), "Pioneer" and "Leading Goose" R & D Program of Zhejiang (Nos. 2022C03146 and 2023C03017), China Postdoctoral Science Foundation (No. 2024T170805), and Zhejiang Provincial Natural Science Foundation of China (No. LDT23E06015B06). The authors acknowledge the support of the Research Computing Center in College of Chemical and Biological Engineering at Zhejiang University for assistance with the calculations.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111187.

  
  
• 1. [1]

     Z. Liu, Z. Deng, S. Davis, et al., Nat. Rev. Earth Environ. 4 (2023) 205–206. doi: 10.1038/s43017-023-00406-z

  2. [2]

     P. De Luna, C. Hahn, D. Higgins, et al., Science 364 (2019) eaav3506. doi: 10.1126/science.aav3506

  3. [3]

     K. Alter, J. Jacquemont, J. Claudet, et al., Nat. Commun. 15 (2024) 2885. doi: 10.1038/s41467-024-47064-3

  4. [4]

     C.A. Phillips, B.M. Rogers, M. Elder, et al., Sci. Adv. 8 (2022) eabl7161. doi: 10.1126/sciadv.abl7161

  5. [5]

     F. Dattila, R.R. Seemakurthi, Y.C. Zhou, et al., Chem. Rev. 122 (2022) 11085–11130. doi: 10.1021/acs.chemrev.1c00690

  6. [6]

     Y. Zhao, X.L. Zu, R.H. Chen, et al., J. Am. Chem. Soc. 144 (2022) 10446–10454. doi: 10.1021/jacs.2c02594

  7. [7]

     C. Du, J.P. Mills, A.G. Yohannes, et al., Nat. Commun. 14 (2023) 6142. doi: 10.1038/s41467-023-41871-w

  8. [8]

     Y. Sun, X. Wang, H. Zhang, et al., ACS Catal. 14 (2024) 1351–1362. doi: 10.1021/acscatal.3c04710

  9. [9]

     J. Feng, L. Wu, S. Liu, et al., J. Am. Chem. Soc. 145 (2023) 9857–9866. doi: 10.1021/jacs.3c02428

  10. [10]

      S. Kuang, M. Li, X. Chen, et al., Chin. Chem. Lett. 34 (2023) 108013. doi: 10.1016/j.cclet.2022.108013

  11. [11]

      Y. Sun, J. Xie, Z. Fu, et al., ACS Nano 17 (2023) 13974–13984. doi: 10.1021/acsnano.3c03952

  12. [12]

      H. Zhang, X. Wang, Y. Sun, et al., Appl. Catal. B 351 (2024) 123992. doi: 10.1016/j.apcatb.2024.123992

  13. [13]

      C. Sun, J. Hao, B. Wei, et al., Chin. Chem. Lett. 34 (2023) 108520. doi: 10.1016/j.cclet.2023.108520

  14. [14]

      X. Chen, A. Xu, D. Wei, et al., Chin. Chem. Lett. 36 (2025) 110175. doi: 10.1016/j.cclet.2024.110175

  15. [15]

      H. Yang, Y.W. Hu, J.J. Chen, et al., Adv. Energy Mater. 9 (2019) 1901396. doi: 10.1002/aenm.201901396

  16. [16]

      W.H. Lee, C. Lim, S.Y. Lee, et al., Nano Energy 84 (2021) 105859. doi: 10.1016/j.nanoen.2021.105859

  17. [17]

      Y. Yao, Y. Zhou, X. Liu, et al., J. Mater. Chem. A 10 (2022) 20914–20923. doi: 10.1039/d2ta05565a

  18. [18]

      Y. Zhou, Y. Yao, R. Zhao, et al., Angew. Chem. Int. Ed. 61 (2022) e202205832. doi: 10.1002/anie.202205832

  19. [19]

      F.M. Yap, J.Y. Loh, S. Yuan, et al., Adv. Funct. Mater. 35 (2025) 2407605. doi: 10.1002/adfm.202407605

  20. [20]

      W. Deng, P. Zhang, Y. Qiao, et al., Nat. Commun. 15 (2024) 892. doi: 10.1038/s41467-024-45230-1

  21. [21]

      Y. Liang, J. Zhao, Y. Yang, et al., Nat. Commun. 14 (2023) 474. doi: 10.1038/s41467-023-35993-4

  22. [22]

      T. Qin, Y. Qian, F. Zhang, et al., Chin. Chem. Lett. 30 (2019) 314–318. doi: 10.1016/j.cclet.2018.07.003

  23. [23]

      C. Wang, Z. Lv, Y. Liu, et al., Angew. Chem. Int. Ed. 63 (2024) e202411216. doi: 10.1002/anie.202411216

  24. [24]

      Y. Gao, Q. Wu, X. Liang, et al., Adv. Sci. 7 (2020) 1902820. doi: 10.1002/advs.201902820

  25. [25]

      K.D. Yang, W.R. Ko, J.H. Lee, et al., Angew. Chem. Int. Ed. 56 (2017) 796–800. doi: 10.1002/anie.201610432

  26. [26]

      H. Zhang, C. He, S. Han, et al., Chin. Chem. Lett. 33 (2022) 3641–3649. doi: 10.1016/j.cclet.2021.12.018

  27. [27]

      C. Han, V. Kundi, Z. Ma, et al., Adv. Funct. Mater. 33 (2023) 2210938. doi: 10.1002/adfm.202210938

  28. [28]

      A. Bagger, W. Ju, A.S. Varela, et al., ACS Catal. 9 (2019) 7894–7899. doi: 10.1021/acscatal.9b01899

  29. [29]

      J. Rosen, G.S. Hutchings, Q. Lu, et al., ACS Catal. 5 (2015) 4293–4299. doi: 10.1021/acscatal.5b00840

  30. [30]

      H.G. Yang, C.H. Sun, S.Z. Qiao, et al., Nature 453 (2008) 638–641. doi: 10.1038/nature06964

  31. [31]

      M. Li, S.X. Yu, H.W. Huang, et al., Angew. Chem. Int. Ed. 58 (2019) 9517–9521. doi: 10.1002/anie.201904921

  32. [32]

      J. Zhao, L.B. Sun, S. Canepa, et al., J. Mater. Chem. A 5 (2017) 11905–11916. doi: 10.1039/C7TA01871A

  33. [33]

      M.J. Siegfried, K.S. Choi, Angew. Chem. Int. Ed. 47 (2008) 368–372. doi: 10.1002/anie.200702432

  34. [34]

      M.D. Susman, Y. Feldman, A. Vaskevich, et al., ACS Nano 8 (2014) 162–174. doi: 10.1021/nn405891g

  35. [35]

      H. Luo, B. Li, J.G. Ma, et al., Angew. Chem. Int. Ed. 61 (2022) e202116736. doi: 10.1002/anie.202116736

  36. [36]

      H.X. Wang, Z. Liang, M. Tang, et al., Joule 3 (2019) 1927–1936. doi: 10.1016/j.joule.2019.05.023

  37. [37]

      H. Wu, Z. Wang, B. Tian, et al., Nanoscale 16 (2024) 3034–3042. doi: 10.1039/d3nr05023h

  38. [38]

      T. Wu, M.Z. Sun, B.L. Huang, Angew. Chem. Int. Ed. 60 (2021) 22996–23001. doi: 10.1002/anie.202110636

  39. [39]

      D.F. Gao, R.M. Arán-Ais, H.S. Jeon, et al., Nat. Catal. 2 (2019) 198–210. doi: 10.1038/s41929-019-0235-5

  40. [40]

      G.H. Simon, C.S. Kley, B.R. Cuenya, Angew. Chem. Int. Ed. 60 (2021) 2561–2568. doi: 10.1002/anie.202010449

  41. [41]

      C. Yan, W. Luo, H. Yuan, et al., Appl. Catal. B 308 (2022) 121191. doi: 10.1016/j.apcatb.2022.121191

  42. [42]

      P. Wang, S. Meng, B. Zhang, et al., J. Am. Chem. Soc. 145 (2023) 26133–26143. doi: 10.1021/jacs.3c08312

  43. [43]

      H.P. Xu, D. Rebollar, H.Y. He, et al., Nat. Energy 5 (2020) 623–632. doi: 10.1038/s41560-020-0666-x

  44. [44]

      X. Kong, J. Zhao, J. Ke, et al., Nano Lett. 22 (2022) 3801–3808. doi: 10.1021/acs.nanolett.2c00945

  45. [45]

      Y. Ou, S.D. Li, F. Wang, et al., Chin. J. Catal. 43 (2022) 2026–2033. doi: 10.1016/S1872-2067(21)63958-X

  46. [46]

      J.L. Yin, Z.Y. Gao, F.Y. Wei, et al., ACS Catal. 12 (2022) 1004–1011. doi: 10.1021/acscatal.1c04714

  47. [47]

      Y. Lum, J.W. Ager, Nat. Catal. 2 (2019) 86–93.

  48. [48]

      K.P. Kuhl, E.R. Cave, D.N. Abram, et al., Energy Environ. Sci. 5 (2012) 7050–7059. doi: 10.1039/c2ee21234j

  49. [49]

      Y. Yang, A.B. He, H. Li, et al., ACS Catal. 12 (2022) 12942–12953. doi: 10.1021/acscatal.2c03833

  50. [50]

      H.Y. An, L.F. Wu, L.D.B. Mandemaker, et al., Angew. Chem. Int. Ed. 60 (2021) 16576–16584. doi: 10.1002/anie.202104114

  51. [51]

      I.V. Chernyshova, P. Somasundaran, S. Ponnurangam, Proc. Natl. Acad. Sci. U. S. A. 115 (2018) E9261–E9270.

  52. [52]

      M. Zheng, P.T. Wang, X. Zhi, et al., J. Am. Chem. Soc. 144 (2022) 14936–14944. doi: 10.1021/jacs.2c06820

  53. [53]

      W. Sugimoto, K. Aoyama, T. Kawaguchi, et al., J. Electroanal. Chem. 576 (2005) 215–221. doi: 10.1016/j.jelechem.2004.10.018

  54. [54]

      J. Feng, L. Zhang, S. Liu, et al., Nat. Commun. 14 (2023) 4615. doi: 10.1038/s41467-023-40412-9

  55. [55]

      Z. Ji, W. Yuan, S. Zhao, et al., Chem Catal. 3 (2023) 100501.

  56. [56]

      Z.Z. Wu, F.Y. Gao, M.R. Gao, Energy Environ. Sci. 14 (2021) 1121–1139. doi: 10.1039/d0ee02747b

  57. [57]

      S.C. Lin, C.C. Chang, S.Y. Chiu, et al., Nat. Commun. 11 (2020) 3525. doi: 10.1038/s41467-020-17231-3

  58. [58]

      H.W. Deng, C.Y. Guo, P.H. Shi, et al., ChemCatChem 13 (2021) 4325–4333. doi: 10.1002/cctc.202100620

  59. [59]

      M. Asadi, K. Kim, C. Liu, et al., Science 353 (2016) 467–470. doi: 10.1126/science.aaf4767

  60. [60]

      L. Chen, C. Tang, K. Davey, et al., Chem. Sci. 12 (2021) 8079–8087. doi: 10.1039/d1sc01694f

  61. [61]

      A.J. Garza, A.T. Bell, M. Head-Gordon, ACS Catal. 8 (2018) 1490–1499. doi: 10.1021/acscatal.7b03477

  62. [62]

      Z.H. Zhao, J.R. Huang, P.Q. Liao, et al., J. Am. Chem. Soc. 145 (2023) 26783–26790. doi: 10.1021/jacs.3c08974

  63. [63]

      J. Zhang, C.X. Guo, S.S. Fang, et al., Nat. Commun. 14 (2023) 1298. doi: 10.1038/s41467-023-36926-x

  64. [64]

      M.T. Greiner, L. Chai, M.G. Helander, et al., Adv. Funct. Mater. 22 (2012) 4557–4568. doi: 10.1002/adfm.201200615

• 

  Figure 1  Morphologies and structural characterization of the prepared catalysts. HR-TEM images of (a) Cu₂O and (b) CO–Cu₂O. AFM images of (c) Cu₂O smooth surface and (d) CO–Cu₂O, showing nanoparticle formation after CO induction. Extracted line profiles for (e) Cu₂O and (f) CO–Cu₂O.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  X-ray fine structure analysis. (a) Normalized Cu K-edge XANES spectra. (b) k³-weighted Fourier transform EXAFS spectra without phase correction. Cu K-edge WT-EXAFS contour plots of (c) Cu₂O and (d) CO–Cu₂O. Fitting of k space and R space of Cu in (e) Cu₂O and (f) CO–Cu₂O.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  In situ monitoring of adsorbed *CO species. In situ ATR-SEIRAS spectra of (a) Cu₂O and (b) CO–Cu₂O during timed current scans from 0 V to −1.0 V vs. RHE. (c) The adsorption behavior of species within 2200–1600 cm⁻¹ ranges, with Cu, O, C, and H represented by orange, red, gray, and white, respectively. The integrated area of *CO_atop for (d) Cu₂O and (e) CO–Cu₂O at various wavenumbers is also shown. The experimental conditions included CO₂-saturated 0.5 mol/L KHCO₃ as the electrolyte. Peak positions were determined based on the maximum intensity of the CO stretching band. (f) Electrochemical CO stripping voltammetry curves of Cu₂O and CO–Cu₂O in saturated N₂ or CO atmospheres.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  CO₂-to-C₂ electroreduction performance. (a) Comparison of LSV for CO₂ and Ar-saturated electrolytes. Faradaic efficiency of each product generated with (b) Cu₂O and (c) CO–Cu₂O at various imposed potentials. (d) The partial current density of C₂ products at various potentials over Cu₂O and CO–Cu₂O. ECSA-normalized current density of different products of (e) Cu₂O and (f) CO–Cu₂O. (g) The FE_C2/FE_C1 ratios of Cu₂O and CO–Cu₂O under various imposed potentials. (h) Stability tests of CO–Cu₂O at −1.2 V vs. RHE in 0.5 mol/L KHCO₃.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Cu valence changes in catalysts after CO₂ electroreduction. (a) XRD spectra of Cu₂O and CO–Cu₂O after 1 h of CO₂RR at −1.2 V vs. RHE. Cu LMM spectra of (b) Cu₂O and (c) CO–Cu₂O at initial states and those after 1 h of CO₂RR at −1.2 V vs. RHE.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  Mechanistic studies by DFT calculations. (a) The free energy diagrams for CO₂ reduction to different intermediates on the Cu₂O (100) and Cu₂O (311) surfaces at U = 0 V. DCD map for the *CO intermediate adsorbed on the (b) Cu₂O (100) and (c) Cu₂O (311) surfaces. (d) ELF of Cu₂O (100) and Cu₂O (311). The WF across the (e) Cu₂O (100) slab and (f) Cu₂O (311) slab vacuum interface.

  下载: 全尺寸图片 幻灯片
•