English

• Zero-gap membrane electrode assembly electrolyzer significantly reduces the inter-electrode distance and markedly improve energy efficiency, which is conducive to promote industrial electro-synthesis [1-5]. The electrocatalytic reduction of CO₂ (CO₂RR) to valuable fuels and feedstocks using renewable electricity presents a carbon-neutral pathway for sustainable fuel generation [6]. Among various hydrocarbons and oxygenates, ethanol (C₂H₅OH) stands out as an important liquid product due to its widespread applications as an organic solvent, disinfectant, and fuel additive [7,8]. Thus, the efficient electrosynthesis of C₂H₅OH from CO₂ holds considerable economic and environmental significance.

  Recently, numerous copper-based catalytic sites, including homo-Cu sites and multi-Cu sites structures, have been designed to enhance the CO₂RR performance toward C₂H₅OH [9-11]. However, in most cases, the Faradaic efficiency for C₂H₅OH (FE_C2H5OH) remains lower than that for ethylene (C₂H₄), another major C₂ product. This discrepancy primarily stems from the higher energy barrier associated with C₂H₅OH formation [12,13]. Specifically, during CO₂RR, both C₂H₅OH and C₂H₄ share a common key intermediate, ^*CO–^*COH (where * denotes an adsorption site) [14]. It is crucial to note that for C₂H₅OH to be produced, one of the two C–O bonds in the ^*CO–^*COH intermediate must be preserved [15].

  However, during the CO₂RR process, the key intermediate ^*CO–^*COH becomes coordinated to two Cu sites via each of its C atoms. This bidentate coordination facilitates the hydrogenation of O atoms, promoting facile cleavage of the C–O bond and leading preferentially to C₂H₄ formation (Scheme 1a) [16]. Moreover, the adsorbed ^*CO−^*COH species may undergo structural rearrangement, transforming into an enol-like configuration where one of the C atoms originally bound to Cu is replaced by an O atom. Nevertheless, even in this enol-form intermediate, C–O bond rupture remains favorable, resulting again in C₂H₄ as the dominant product [17]. Consequently, the FE_C2H5OH is typically two to three times lower than that for FE_C2H4, highlighting the significant challenge in achieving high selectivity of C₂H₅OH by Cu-based catalysts with homo-sites and multi-sites. Therefore, developing more efficient catalyst architectures and gaining deeper mechanistic insights into the pathways are essential for advancing CO₂RR toward practical C₂H₅OH synthesis.

  Scheme 1

  

  Scheme 1.  Poroposed reaction paths for yielding C₂H₄ vs. C₂H₅OH on the surface of (a) symmetric Cu¹-Cu¹ sites and (b) asymmetric Cu¹-Cu² sites in this work.

  DownLoad: Full-Size Img PowerPoint

  A promising strategy to suppress C₂H₄ formation and promote ethanol selectivity involves steering the CO₂RR pathway toward distinct mechanistic routes for C₂H₅OH production [18,19]. This alternative mechanism proceeds through unique C₁ and C₂ intermediates and involves an unconventional C–C coupling step. By enabling one of the C₁ intermediates to coordinate to the metal center via its O atom, rather than the C atom, premature hydrogenation of O can be circumvented, thereby preserving the C–O bond and favoring C₂H₅OH formation [20]. Theoretical studies have identified several such O-bound intermediates, including ^*OC, ^*OCHO, ^*OCH₂OH, and ^*OCH₂ [21,22]. Furthermore, an asymmetric C–C coupling between ^*OCH₂ and another C₁ species, ^*CO, can generate the critical intermediate CO–OCH₂ (Scheme 1b), which exhibits a higher propensity toward C₂H₅OH than C₂H₄ [23-25]. Consequently, it is imperative that adjacent copper atoms in the catalytic site possess asymmetric electronic or geometric properties to facilitate this distinct reaction pathway.

  Cu-based catalysts with strong affinity for O atoms are conducive to the formation of ^*OCH₂ during CO₂RR process. In particular, cuprous oxide (Cu₂O) has been reported to preferentially stabilize ^*OCHO with respect to ^*CO species [26,27], which accounts for its widespread application in the electrosynthesis of HCOOH. Recent studies show that introducing Cu⁺ vacancies into Cu₂O via surface modification can selectively enhance the adsorption and fixation of ^*CO [28]. These findings indicate that adjacent active sites within copper catalysts, featuring distinct affinities for carbon and oxygen, respectively, can simultaneously promote the generation of ^*CO and ^*OCH₂ species. This asymmetric environment facilitates the critical C–C coupling between ^*CO and ^*OCH₂, forming the ^*CO–^*OCH₂ intermediate and ultimately promoting the selective production of C₂H₅OH from CO₂, enriching the local reaction concentration at the catalytic surface.

  To verify the above hypothesis, the alkanethiols molecule-modified Cu₂O nanowire array (OTT-Cu₂O) with unprecedented asymmetric Cu sites (namely, a pair of Cu–O and Cu–S sites) was designed to overcome aforementioned limitation, which exhibits an impressive selectivity of CO₂RR to C₂H₅OH reaches up to 45% at 300 mA/cm², representing a more than two-fold improvement compared to the Cu₂O electrode. Mechanistic studies indicate that, in contrast to the Cu–O site which exhibits higher oxygen affinity, the Cu–S site facilitates unique carbon bonding, playing a critical role in stabilizing key intermediates such as ^*OCH₂ and ^*CO. Consequently, the asymmetric S–Cu–O sites are shown to be thermodynamically more favorable for the asymmetric C–C coupling between ^*CO and ^*OCH₂, leading to the formation of the critical CO–OCH₂ intermediate, compared to symmetric O–Cu–O sites that mainly produce HCOOH. This configuration significantly promotes the selective formation of C₂H₅OH. This work not only establishes a promising strategy for achieving high selectivity in CO₂-to-C₂H₅OH conversion through the rational design of multi-active-site catalysts, but also provides deeper mechanistic insights into the pathway of CO₂ reduction toward C₂H₅OH.

  Copper mesh (CM, 200 mesh) was purchased from anping hangying silkscreen Co., Ltd. Potassium bicarbonate (KHCO₃, > 99.5%), potassium hydroxide (KOH, > 90%), hydrochloric acid (HCl, 36%−38%) and anhydrous ethanol (C₂H₅OH, > 99.7%) were purchased from Chengdu Kelong Chemical Co., Ltd. Phosphoric acid (H₃PO₄, > 85%) was obtained from Chongqing Chuandong Chemical Co., Ltd. Dimethylsulfoxide (DMSO, 99.8%), deuterium oxide (D₂O, 99.9%), and 1-octadecanethiol (C₁₈H₃₈S, 97%) was acquired from Aladdin Co., Ltd. All reagents were used as received without further purification. Deionized water (18.2 MΩ/cm) was generated from an ultra-pure purified water system.

  Pretreatment of Cu mesh: The purchased Cu mesh cleaned in a variety of ways. Firstly, Cu mesh was washed with HCl to remove surface impurities. Whereafter, surface electropolish was conducted in a two-electrode system by using 85% H₃PO₄ solution as electrolyte, Cu mesh as working electrode (2 × 2 cm²), and a platinum plate as counter electrode, respectively. A constant voltage of 4 V was applied to the Cu surface for 75 s. After polishing, the electrode was washed with deionized water several times.

  Preparation of Cu₂O nanowires (Cu₂O electrode): The dense Cu₂O nanoneedles layers can synthesize by electrochemical oxidation coupled with heating dehydration strategy. Firstly, the cleaned Cu mesh as anode electrode was immersed in 3 mol/L KOH electrolyte, a platinum plate as counter electrode, and Ag/AgCl as reference electrode, respectively. A constant current of 11 mA/cm² was applied for 350 s to obtain Cu(OH)₂ nanoneedles. Afterward, Cu(OH)₂ nanoneedles were dehydrated at 180 ℃ for 1 h, and the Cu₂O electrode was prepared.

  Preparation of OTT modified Cu₂O nanowires (Cu₂O-OTT electrode): The hydrophilic Cu₂O was immersed in an N₂-saturated ethanol solution containing 10 mmol/L of 1-octadecanethiol at 60 ℃ in continuous N₂ flow for natural drying, and the Cu₂O@OTT electrode was prepared. Subsequently, the Cu₂O@OTT precursor was electrochemically converted to OTT-Cu₂O by applying a potential of −1.4 V vs. Ag/AgCl for 30 min in a CO₂-saturated 0.5 mol/L KHCO₃ electrolyte.

  X-ray diffraction (XRD) measurements were performed on D8 ADVANCE X-ray diffractometer from Bruker AG at 40 kV and 40 mA using Cu Kα radiation. X-ray photoemission spectroscopy (XPS) data were collected on a ESCALAB250Xi spectrometer equipped with an Al-Kα radiation source (1486.6 eV) using C 1s (284.8 eV) as the reference. Fourier transform infrared spectroscopy (FTIR) measurements were recorded on Nicolet IS 10. Scanning electron microscopy (SEM) images were acquired on GeminiSEM 300 with an accelerating voltage of 3 kV. Transmission electron microscope (TEM) images and EDX mapping images were taken on JEM1200EX with an operated voltage at 200 kV. X-ray absorption fine structure (XAFS) measurement was collected on an Easy XAFS300+ system, and the acquired data were processed via the standard procedure using the ATHENA module of Demeter software packages. The contact angle was measured using a contact angle measuring instrument XC-CAM. The gas-phase product was detected by gas chromatography (GC) from Agilent 8890. The liquid phase product was tested by nuclear magnetic resonance (NMR) from Bruker AVANCE Ⅲ. 600 M. In-situ ATR-SEIRAS spectra were collected by an FT-IR spectrometer (IRTracer100+CHI 660E) equipped with an MCT-A detector at different applied potentials. In-situ Raman spectra obtained by the LabRAM HR Evolution at different applied potentials.

  Electrochemical characterization of the prepared electrodes was performed on an electrochemical workstation (VMP3, Biologic) in a two-electrode MEA system (Wuhan Zhisheng New Energy Co., Ltd.), which is comprised of the Cu₂O@OTT electrode cathode, an nickel foam anode, and a piece of Nafion 115 membrane. During electrolysis, the CO₂-saturated 1 mol/L KHCO₃ was directly fed to cathodic compartment at a rate of 100 mL/min, and 1 mol/L KOH was forced to continuously circulate through the anode compartment with a rate of 40 mL/min. Chronopotentiometry measurements were performed in the current density of 50 mA/cm² to 600 mA/cm², and full-cell voltages reported were non-iR corrected full-cell potentials.

  In general, the thiol end group are among the most stable, suggesting that, even under electrochemical reducing conditions, the ligand could remain chemisorbed to the catalyst surface. To address the simultaneous the requirements, we engineered a hierarchical Cu₂O nanowires array with a thiol-modified surface (Fig. 1). Initially, copper hydroxide (Cu(OH)₂) nanowires on Cu foam (Cu(OH)₂ NWs) were prepared via a straightforward oxidizing process. After annealing under argon condition, these were transformed into Cu₂O NWs, subsequently converted into OTT-Cu₂O NWs array by simple wet-chemical adsorption combined with electrochemical activation strategies. As shown in Fig. S1 (Supporting information), for Cu₂O NWs, SEM images feature its one-dimensional wire with a smooth surface, and the fine wire average diameter circa of 100 nm. Afterward, the resulting Cu₂O NWs were immersed in an C₂H₅OH solution containing 1-octadecanethiol (OTT) for the synthesis of Cu₂O@OTT, and then subjected to an electrochemical activation to remove excess weak molecular layer, leading to a strongly bonded OTT molecular surface on the one-dimensional wire (Cu₂O-OTT), which feature asymmetric Cu–S–Cu–O sites for significantly enhancing CO₂-to-C₂H₅OH conversion.

  Figure 1

  

  Figure 1.  Schematic illustration of the synthesis of asymmetric Cu–S–Cu–O sites.

  DownLoad: Full-Size Img PowerPoint

  The powder X-ray diffraction (XRD) pattern of as-prepared Cu₂O electrode exhibited a characteristic peak corresponding to the (111) plane of Cu₂O (Fig. 2a), which is more favorable to bond ^*OCHO species. After OTT modification, no additional diffraction peaks emerged in the XRD pattern, indicating that the crystalline structure of Cu₂O remained unaltered. To confirm the existence of the OTT layer, contact angle measurements were conducted and results shown in Fig. 2b, the electrode contact angle to increase dramatically from 6.1° to 130.1° after OTT modification. Moreover, the strong Fourier transform infrared spectroscopy (FTIR) absorption bands at 785,1473 and 2067 cm⁻¹ (Fig. 2c) further suggested that OTT covered the surface of Cu₂O electrode. The successful functionalization with the OTT was further corroborated through X-ray photoelectron spectroscopy (XPS) spectroscopy. As shown in Fig. S3 (Supporting information), C, S, O, and Cu elements in Cu₂O@OTT electrode. Prior to treatment, Cu₂O exhibited characteristic peaks consistent with Cu₂O at 932.5 and 952.4 eV, while corresponding to Cu²⁺ environments at 942.7 and 962.9 eV. After thiolation of the OTT layer, the Cu²⁺ environment on the surface was removed, leaving Cu-S bonds and forming a Cu⁺ surface (Fig. S4 in Supporting information). Additionally, a new S 2p peak appeared at 163.0 eV, indicating the presence of Cu-S bonds (Fig. 2d).

  Figure 2

  

  Figure 2.  Composition and structure characterizations: (a) XRD pattern, (b) the contact angle, (c) FTIR spectra, (d) S 2p XPS spectra of Cu₂O and Cu₂O@OTT. (e) SEM image, (f) TEM image, (g) HAADF-STEM image of Cu₂O@OTT. (h) In-situ Raman spectra of Cu₂O@OTT at electrolysis from −1.6 V to −2.0 V vs. Ag/AgCl in CO₂-saturated 0.5 mol/L KHCO₃. (i) SEM image, (j) TEM image, (k) HAADF-STEM image of Cu₂O-OTT. Equivalent images from i-to-k after electrolysis at –1.4 V vs. Ag/AgCl in CO₂-saturated 0.5 mol/L KHCO₃ for 30 min.

  DownLoad: Full-Size Img PowerPoint

  SEM of the Cu₂O NWs after treatment confirmed that the nanostructure remained intact (Fig. 2e and Fig. S2 in Supporting information) and was coated with a monolayer between 4 nm and 5 nm in thickness (Fig. 2f). Through HAADF-STEM element-specific signal analysis, S elemental mapping analysis visually demonstrates uniform coverage of OTT on the Cu₂O surface (Fig. 2g). After maintaining a reduction potential in CO₂-saturated 0.5 mol/L KHCO₃ electrolyte for over 30 min, the activation of the Cu₂O@OTT electrode was converted into a Cu₂O-OTT electrode. Meanwhile, the presence of Cu₂O species are confirmed by in-situ Raman spectra during electrolysis (Fig. 2h and Fig. S6 in Supporting information). SEM characterization confirmed that the activated Cu₂O-OTT electrode retained its integrity (Fig. 2i and Fig. S5 in Supporting information), the Cu₂O (111) crystal plane still presented, and the Cu₂O surface still retains an approximately 2 nm-thick OTT layer (Fig. 2j), consisting with a surface of 1-octadecanethiol molecules bound upright (the chain length is 2.3 nm between the surface-bound S and terminal C). HAADF-STEM elemental analysis revealed that although the OTT coverage decreased after electrochemical activation, it maintained uniform distribution (Fig. 2k), which indicates selective peeling of partial OTT layers during potential application. Preliminary electrochemical characterization demonstrates that the ECSA decreases after OTT coating, the active area of Cu₂O-OTT electrodes (1.98 mF/cm²) is smaller than that of Cu₂O electrodes (6.26 mF/cm²) (Fig. S7c in Supporting information), which stems from reduced electrolyte contact area due to hydrophobic surface properties.

  X-ray absorption spectroscopy (XAS) was employed to characterize the chemical states and local fine structures of the Cu in Cu₂O before and after OTT modification. The X-ray absorption near-edge structure (XANES) spectra in Fig. 3a reveal that the absorption edge positions of Cu₂O-OTT shifting to the higher energy side with respect to Cu₂O, which indicates that Cu₂O-OTT had a different coordination environment from that of Cu₂O. As shown in Fig. 3b, the valence of Cu₂O lies between the characteristic values of Cu₂O Re and Cu Re, whereas the Cu₂O-OTT is sized of a valence between Cu₂O Re and standard CuO Re, implying that the oxidation state of Cu in Cu₂O increase after the modification of OTT coating layers. The Fourier transform extended k-edge X-ray absorption fine structure (EXAFS) spectra in Fig. 3c reveals that Cu₂O-OTT consisted of Cu–O, Cu–S, and Cu–Cu bonds with simulated coordination numbers of 1.13, 1.85, and 2.43 Å, respectively. While the fitting results of Cu₂O show that coordination numbers of Cu-O and Cu-Cu bonds are 1.75 and 2.61 Å, suggesting the formation of many unsaturated sites in Cu₂O after OTT modification. Wavelet transform extended XAFS (WT-EXAFS) spectra was conducted to further highlight the local structural changes in Cu₂O (Fig. 3d). The reduced backscattering amplitude at Cu–O in Cu₂O-OTT compared with Cu₂O, indicating the lower coordination state of Cu₂O-OTT. As a result of OTT modification, unprecedented asymmetric Cu sites (namely, a pair of Cu–O and Cu–S sites) was successfully created for the further selective reduction of CO₂.

  Figure 3

  

  Figure 3.  XAS characterizations: (a) Normalized Cu K-edge XANES spectra of Cu₂O and Cu₂O@OTT with the references. (b) the mean chemical valence analysis of Cu species. (c) FT-EXAFS spectra of Cu₂O and Cu₂O@OTT with the references. (d) wavelet transform for the Cu K-edge EXAFS spectra of Cu₂O Re, Cu₂O, Cu₂O@OTT and Cu₂S Re.

  DownLoad: Full-Size Img PowerPoint

  To evaluate the CO₂RR activity of the OTT-modified Cu₂O electrode, the electrocatalytic CO₂RR performance was measured in a zero-distance electrolyzer with a CO₂-saturated 0.5 mol/L KHCO₃ electrolyte at applied current density from 50 mA/cm² to 500 mA/cm² (Figs. S8 and S9 in Supporting information). The products were characterized by nuclear magnetic resonance (NMR) and gas chromatography (GC), including CO, HCOO-, CH₄, and C₂₊ products (C₂H₄, C₂H₅OH, CH₃COOH and C₃H₇OH) (Fig. S10 in Supporting information). Fig. 4a shows the LSV of the hydrophobic Cu₂O@OTT electrode and equivalent wettable Cu₂O electrode in a CO₂-saturated KHCO₃ electrolyte (0.5 mol/L, pH 6.8). To achieve a current density of −300 mA/cm², the Cu₂O electrode required a voltage of 2.78 V, whereas the Cu₂O-OTT electrode required a significantly higher voltage of 3.15 V. This reduction in current density at a given voltage is partially attributable to the lower ECSA of the hydrophobic electrode.

  Figure 4

  

  Figure 4.  CO₂RR performance in a MEA electrolyzer: (a) I-V curves of Cu₂O and Cu₂O-OTT; FE of products of the (b) Cu₂O and (c) Cu₂O-OTT. (d) FE of C₂₊ and C₂H₅OH. (e) Ratios of C₂₊ to-HCOO- for Cu₂O and Cu₂O-OTT. (f) Performance comparison of Cu-based catalysts for C₂H₅OH production by CO₂RR pathway. (g) Stability test of Cu₂O-OTT in 0.5 mol/L KHCO₃ electrolyte.

  DownLoad: Full-Size Img PowerPoint

  The distribution of gases and liquid products generated are presented in Fig. 4 and c. For Cu₂O electrode, HCOO⁻ is the primary product along with little C₂H₅OH. As the current density becomes large, C₂H₄ and CH₄ products begin to appear, while the maximum FE_C2H5OH reached only 18% at 100 mA/cm² (Fig. 4b). However, C₂H₅OH is the primary product for Cu₂O-OTT electrode, reaching a peak Faradaic efficiency of 45% at 300 mA/cm² (Fig. 4c). It is worth noting that the selectivity of H₂ over Cu₂O electrode does not change significantly after OTT modification. To visually represent the influence of OTT on the reduction products, the selectivity data is re-plotted in Fig. 4d. The FE_C2+ of the CuO-OTT electrode is 60%, contrasting with 30% for the Cu₂O electrode, with C₂H₅OH comprising the majority of C₂₊ products, showcasing an enhancement of more than 2-fold. Besides, the ratio of C₂₊/HCOO⁻ is 4.8 for CuO-OTT at 600 mA/cm², which is much higher than that of Cu₂O (Fig. 4e). Furthermore, energy efficiency of C₂H₅OH for Cu₂O-OTT is also significantly higher with respect to Cu₂O (Fig. S11 in Supporting information). Notably, to the best of our knowledge, the outstanding CO₂RR performance of Cu₂O-OTT for C₂H₅OH production is record high in comparison with the previously reported (Fig. 4f and Table S1 in Supporting information).

  As shown in Fig. 4g, the Cu₂O-OTT electrode exhibited a relatively stable bias under a constant current density of 200 mA/cm², with a negligible performance degradation within 10 h, evidencing the excellent stability of Cu₂O-OTT electrode under CO₂RR operating conditions. In addition, SEM image illustrates that the nanowire structure is maintained, XRD analysis confirmed retention of the Cu₂O (111) crystal plane in the Cu₂O-OTT electrode, while XPS detected persistent S species. Contact angle measurements further demonstrated maintained hydrophobicity, collectively illustrating excellent stability and activity (Fig. S12 in Supporting information).

  Typical Cu-based electrocatalysts cannot achieve high FE_C2H5OH due to the competitive thermodynamics of the products of C₂H₄. To further investigate the possible reaction pathways and intermediates of the CO₂RR to C₂H₅OH on the Cu₂O-OTT surface, in situ spectroscopy are performed in CO₂-saturated 0.5 mol/L KHCO₃ electrolyte (Fig. 5a, Figs. S13 and S14 in Supporting information). In our observations, we noted five distinct infrared signals at 1120, 1226, 1473, 1583, and 1646 cm⁻¹ for Cu₂O-OTT during continuous operation (Fig. 5b). The first peak signal at 1120 cm⁻¹ corresponds to the key intermediate ^*OC₂H₅ involved in the conversion of CO₂ to C₂H₅OH [29]. The simultaneous observation of C–C (1226 cm⁻¹) and C–O (1253 cm⁻¹) vibrational peak signals further verified the formation of the C₂H₅OH [30]. It is worth noting that the characteristic peak signal at 1583 cm⁻¹ can be attributed to the C₂H₅OH key coupling ^*CO−^*OCH₂ intermediate over the asymmetric S-Cu-O sites, which is generated by C–C coupling through ^*OCH and ^*CO [31]. By contrast, the main characteristic peak signals of the Cu₂O appear at the ^*OCHO intermediate displayed at 1450 cm⁻¹ and the ^*H₂O around 1650 cm⁻¹ (Fig. 5c).

  Figure 5

  

  Figure 5.  Mechanistic investigation: (a) The schematic illustration of in-situ FTIR and Raman electrolytic cell. (b, c) In-situ FTIR spectra at different potentials vs. Ag/AgCl. (d, e) In-situ Raman spectra at different potentials vs. Ag/AgCl. (f) The ratio of ^*CO_bridge to ^*CO_atop obtained from Raman spectra. (g) Schematic illustration for the catalytic mechanism for CO₂RR to C₂H₅OH on Cu₂O-OTT.

  DownLoad: Full-Size Img PowerPoint

  In Raman spectra (Fig. 5d), two regions appearing at about 283 and 374 cm⁻¹ for Cu₂O-OTT are associated with the surface-absorbed ^*CO, belonging to the Cu–CO frustrated rotation and Cu–CO stretch, respectively [10,13,15]. At the same time, two extra bands appearing at 1435 and 1580 cm⁻¹ for the Cu₂O-OTT were also noticed at all testing potentials. The former is tentatively assigned to the symmetric stretching vibration of the carboxylate group (namely, v–O–(CH)–O), corresponding to the surface-adsorbed ^*OCHO intermediates [32]. While the band at 1580 cm⁻¹ is assigned to the C═O stretching mode of surface-adsorbed *CH_xCO species [33], a pivotal intermediate responsible for the yielding of C₂H₅OH, which maybe originate from the asymmetric C–C coupling between the ^*CO and ^*OCHO species on Cu₂O-OTT with Cu–O and Cu–S pairs. In comparison, only a peak at 1435 cm⁻¹ corresponding to ^*OCHO emerged in the Raman spectra of Cu₂O (Fig. 5e).

  Additionally, there appear two peaks at 1923 and 2044 cm⁻¹ for Cu₂O and Cu₂O-OTT, associating with the bridge–adsorbed ^*CO (^*CO_bridge) and atop–adsorbed ^*CO (^*CO_atop), respectively. In general, ^*CO_bridge protonation to ^*OCH₂ is proved to be more energetically favorable than that of ^*CO_atop [34-36], and the ratio ^*CO_bridge/^*CO_atop for Cu₂O@OTT is significantly greater than that for Cu₂O over the potential range (Fig. 5f), which could be in favor of C₂H₅OH generated from CO₂ (Fig. S16 in Supporting information).

  Based on the above results, we proposed a comprehensive reaction mechanism, as outlined in Fig. 5g. The Cu₂O primarily follows the pathway of CO₂-to-HCOOH to generate HCOO⁻ due to its inherent crystal characteristics (Fig. S15 in Supporting information). OTT layer not only creates a favorable microenvironment for CO₂/CO enrichment, but also allow the formation of S–Cu–O asymmetric site in Cu₂O-OTT electrode, which induces more ^*CO absorbed on the catalyst surface. Afterward, the coupling of ^*CO and ^*OCHO intermediates via asymmetric active sites of Cu–S and Cu–O enables significant enhancement of C₂H₅OH selectivity through synergistic dual-sites pathway. Additionally, the OTT layer enriches CO, allowing C–C coupling between different adsorbed CO states to produce C₂H₅OH. These combined mechanisms achieve the switch of products from HCOO⁻ to C₂H₅OH, and surface modification engineering strategy provides a novel perspective for designing efficient CO₂RR electrocatalysts.

  In summary, this work demonstrates the rational design of an alkanethiol-modified Cu₂O nanowire array electrocatalyst featuring asymmetric Cu–O/Cu–S dual sites for significantly enhancing CO₂-to-C₂H₅OH conversion. The catalyst exhibits a notable C₂H₅OH Faradaic efficiency of 45% at 300 mA/cm², representing a more than two-fold improvement over bare Cu₂O. Mechanistic studies reveal that the O–Cu–S site facilitates critical carbon-bonding interactions that stabilize key ^*CO and ^*OCH₂ intermediates, complementing the oxygen-affinitive nature of the Cu–O site. This asymmetric configuration enables thermodynamically favored C–C coupling between ^*CO and ^*OCH₂ via the asymmetric S–Cu–O motif, leading to the formation of the CO–OCH₂ intermediate and selectively promoting C₂H₅OH production, in contrast to symmetric O–Cu–O sites that favor HCOOH formation. Our findings provide both a promising catalyst design strategy for multi-site synergistic CO₂RR and fundamental insights into the mechanism of asymmetric C–C coupling toward C₂H₅OH.

  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

  Min Zhang: Writing – review & editing, Writing – original draft, Visualization, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Weimin Wang: Visualization, Data curation. Jun Li: Supervision, Resources. Xun Zhu: Supervision, Resources, Funding acquisition. Qian Fu: Writing – review & editing, Supervision, Resources, Funding acquisition.

  Acknowledgments

  The authors are grateful for the financial supports of the National Natural Science Foundation of China (NSFC, Nos. 52394202, 52476056, and 52301232) and the Natural Science Foundation of Chongqing Province (No. 2024NSCQ-MSX1109).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111926.

  
  
• 1. [1]

     M. Zhang, A. Cao, Y. Xiang, et al., Nano-Micro Lett. 16 (2024) 50. doi: 10.1007/s40820-023-01264-6

  2. [2]

     M. Zhang, W. Wei, S. Zhou, et al., Energy Environ. Sci. 14 (2021) 4998–5008. doi: 10.1039/D1EE01495A

  3. [3]

     M. Zhang, X. Wang, J. Ding, et al., Nanoscale 16 (2024) 7076–7084. doi: 10.1039/D4NR00361F

  4. [4]

     M. Zhang, S. Zhou, W. Wei, et al., Chem. Catal. 2 (2022) 3528–3545.

  5. [5]

     M. Zhang, Z. Tan, M. Xing, et al., J. Energy Chem. 112 (2026) 90–96. doi: 10.1016/j.jechem.2025.08.050

  6. [6]

     X. Wang, Z. Wang, F.P. García de Arquer, et al., Nat. Energy 5 (2020) 478–486. doi: 10.1038/s41560-020-0607-8

  7. [7]

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

  8. [8]

     J. Bi, P. Li, J. Liu, et al., Nat. Commun. 14 (2023) 2823. doi: 10.1038/s41467-023-38524-3

  9. [9]

     Z. Chen, Z. Ma, G. Fan, et al., ACS Appl. Mater. Interfaces 16 (2024) 35143–35154. doi: 10.1021/acsami.4c05973

  10. [10]

      G. Feng, S. Wang, S. Li, et al., Angew. Chem. Inter. Ed. 62 (2023) e202218664. doi: 10.1002/anie.202218664

  11. [11]

      A. Herzog, M.L. Luna, H.S. Jeon, et al., Nat. Commun. 15 (2024) 3986. doi: 10.1038/s41467-024-48052-3

  12. [12]

      Y. Lin, T. Wang, L. Zhang, et al., Nat. Commun. 14 (2023) 3575. doi: 10.1038/s41467-023-39351-2

  13. [13]

      P. Luan, X. Dong, L. Liu, et al., ACS Catal. 14 (2024) 8776–8785. doi: 10.1021/acscatal.4c01579

  14. [14]

      C. Peng, J. Ma, G. Luo, et al., Angew. Chem. Int. Ed. 63 (2024) e202316907. doi: 10.1002/anie.202316907

  15. [15]

      D. Wakerley, S. Lamaison, F. Ozanam, et al., Nat. Mater. 18 (2019) 1222–1227. doi: 10.1038/s41563-019-0445-x

  16. [16]

      J. Wang, H. Yang, Q. Liu, et al., ACS Energy Lett. 6 (2021) 437–444. doi: 10.1021/acsenergylett.0c02364

  17. [17]

      P. Wang, H. Yang, C. Tang, et al., Nat. Commun. 13 (2022) 3754. doi: 10.1038/s41467-022-31427-9

  18. [18]

      X. Wang, Z. Jiang, P. Wang, et al., Angew. Chem. Int. Ed. 62 (2023) e202313646. doi: 10.1002/anie.202313646

  19. [19]

      A. Xu, S.F. Hung, A. Cao, et al., Nat. Catal. 5 (2022) 1081–1088. doi: 10.1038/s41929-022-00880-6

  20. [20]

      F. Xu, B. Feng, Z. Shen, et al., J. Am. Chem. Soc. 146 (2024) 9365–9374. doi: 10.1021/jacs.4c01610

  21. [21]

      H. Zhang, Y. Sun, J. Wang, et al., ACS Appl. Energy Mater. 6 (2023) 11448–11457. doi: 10.1021/acsaem.3c01463

  22. [22]

      T. Zhang, S. Xu, D.L. Chen, et al., Angew. Chem. Int. Ed. 63 (2024) e202407748. doi: 10.1002/anie.202407748

  23. [23]

      Y. Zhang, Y. Chen, X. Wang, et al., Nat. Commun. 15 (2024) 5172. doi: 10.1038/s41467-024-49247-4

  24. [24]

      J. Zhou, Q. Liang, P. Huang, et al., Nanoscale 16 (2024) 13011–13018. doi: 10.1039/D4NR01173B

  25. [25]

      J. Li, R. Cai, H. Mu, et al., ACS Catal. 14 (2024) 3266–3277. doi: 10.1021/acscatal.3c05857

  26. [26]

      X. Ma, Y. Zhang, T. Fan, et al., Adv. Funct. Mater. 33 (2023) 2213145. doi: 10.1002/adfm.202213145

  27. [27]

      D. Tan, B. Wulan, J. Ma, et al., Nano Lett. 22 (2022) 6298–6305. doi: 10.1021/acs.nanolett.2c01942

  28. [28]

      S. Li, H. Duan, J. Yu, et al., ACS Catal. 12 (2022) 9074–9082. doi: 10.1021/acscatal.2c01750

  29. [29]

      D. Wakerley, S. Lamaison, J. Wicks, et al., Nat. Energy 7 (2022) 130–143. doi: 10.1038/s41560-021-00973-9

  30. [30]

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

  31. [31]

      Y.C. Li, Z. Wang, T. Yuan, et al., J. Am. Chem. Soc. 141 (2019) 8584–8591. doi: 10.1021/jacs.9b02945

  32. [32]

      H. Li, P. Wei, D. Gao, et al., Curr. Opin. Green Sustain. Chem. 34 (2022) 100589. doi: 10.1016/j.cogsc.2022.100589

  33. [33]

      F. Li, Z. Men, S. Li, et al., Spectrochim. Acta A 189 (2018) 621–624. doi: 10.1016/j.saa.2017.08.077

  34. [34]

      S. Jiang, K. Klingan, C. Pasquini, et al., J. Chem. Phys. 150 (2019) 041718. doi: 10.1063/1.5054109

  35. [35]

      H.H. Heenen, H. Shin, G. Kastlunger, et al., Energy Environ. Sci. 15 (2022) 3978–3990. doi: 10.1039/D2EE01485H

  36. [36]

      D. Karapinar, C.E. Creissen, J.G. Rivera de la Cruz, et al., ACS Energy Lett. 6 (2021) 694–706. doi: 10.1021/acsenergylett.0c02610

• 

  Scheme 1  Poroposed reaction paths for yielding C₂H₄ vs. C₂H₅OH on the surface of (a) symmetric Cu¹-Cu¹ sites and (b) asymmetric Cu¹-Cu² sites in this work.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  Schematic illustration of the synthesis of asymmetric Cu–S–Cu–O sites.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Composition and structure characterizations: (a) XRD pattern, (b) the contact angle, (c) FTIR spectra, (d) S 2p XPS spectra of Cu₂O and Cu₂O@OTT. (e) SEM image, (f) TEM image, (g) HAADF-STEM image of Cu₂O@OTT. (h) In-situ Raman spectra of Cu₂O@OTT at electrolysis from −1.6 V to −2.0 V vs. Ag/AgCl in CO₂-saturated 0.5 mol/L KHCO₃. (i) SEM image, (j) TEM image, (k) HAADF-STEM image of Cu₂O-OTT. Equivalent images from i-to-k after electrolysis at –1.4 V vs. Ag/AgCl in CO₂-saturated 0.5 mol/L KHCO₃ for 30 min.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  XAS characterizations: (a) Normalized Cu K-edge XANES spectra of Cu₂O and Cu₂O@OTT with the references. (b) the mean chemical valence analysis of Cu species. (c) FT-EXAFS spectra of Cu₂O and Cu₂O@OTT with the references. (d) wavelet transform for the Cu K-edge EXAFS spectra of Cu₂O Re, Cu₂O, Cu₂O@OTT and Cu₂S Re.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  CO₂RR performance in a MEA electrolyzer: (a) I-V curves of Cu₂O and Cu₂O-OTT; FE of products of the (b) Cu₂O and (c) Cu₂O-OTT. (d) FE of C₂₊ and C₂H₅OH. (e) Ratios of C₂₊ to-HCOO- for Cu₂O and Cu₂O-OTT. (f) Performance comparison of Cu-based catalysts for C₂H₅OH production by CO₂RR pathway. (g) Stability test of Cu₂O-OTT in 0.5 mol/L KHCO₃ electrolyte.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Mechanistic investigation: (a) The schematic illustration of in-situ FTIR and Raman electrolytic cell. (b, c) In-situ FTIR spectra at different potentials vs. Ag/AgCl. (d, e) In-situ Raman spectra at different potentials vs. Ag/AgCl. (f) The ratio of ^*CO_bridge to ^*CO_atop obtained from Raman spectra. (g) Schematic illustration for the catalytic mechanism for CO₂RR to C₂H₅OH on Cu₂O-OTT.

  下载: 全尺寸图片 幻灯片
•