Taking the synthesis of Cu₂O (100) nanoparticles (NPs) as an example^[20]. Add 90 mL of ultrapure water to a round-bottom flask, then add 10 mL of 0.1 mol/L CuCl₂ solution dropwise. After stirring thoroughly (at 300 rpm for 5 min), place the mixture in an oil bath and heat it to 55 °C. Denote this as Solution A. Add 10 mL of 2 mol/L NaOH solution dropwise to Solution A, and maintain the mixture at 55 °C for 0.5 h. Denote this as Solution B. Add 10 mL of 0.6 mol/L ascorbic acid solution dropwise to Solution B; stir at 300 rpm for 30 min, then age the mixture at 55 °C for 3 h. Finally, filter the product under vacuum, wash it three times with ultrapure water and anhydrous ethanol, respectively, and then dry it overnight at 55 °C.

Using synthesized Cu₂O (100) nanoparticles, long-chain hydrophobic alkane groups were introduced onto their surfaces, yielding HY-Cu₂O (100) NPs. The specific synthesis procedure is as follows: The catalyst was immersed in a solution of 10 mM octadecanethiol in argon-saturated CH₃CH₂OH for 15 min in a 60 °C water bath to perform thiol treatment via self-assembly. Continuous argon flushing was maintained throughout the immersion process to ensure the desired results. The electrode was then repeatedly rinsed with anhydrous ethanol to remove excess octadecanethiol and air-dried.

To investigate the effect of the water microenvironment at the electrode interface on the performance of CO₂ electroreduction, we selected widely studied cubic Cu₂O nanoparticles [Figure 1] as a model catalyst^[21-23], designated as Cu₂O (100), which exhibits hydrophilic surface properties. The synthesis procedure for this catalyst is shown in Figure 1. To enhance the hydrophobicity of the catalyst, we surface-modified Cu₂O (100) with octadecanethiol mercaptan, and the resulting sample was named HY-Cu₂O (100). The modification process was carried out under heated conditions: the pristine Cu₂O nanoparticles were immersed in a CH₃CH₂OH solution containing octadecanethiol mercaptan (structural formula shown in Figure 1) and stirred continuously. During heating, sulfur atoms formed stable coordination bonds with copper atoms; furthermore, due to the low steric hindrance of the linear alkane chain, the self-assembly process was more efficient. This confirms the existence of chemical interactions between the S atom at the terminal end of the octadecanethiol molecule and the Cu atoms on the catalyst surface.

As shown in Figure 2, the cubic Cu₂O nanoparticles fully expose their (100) crystal faces. The Cu₂O (100) surface is relatively smooth. In contrast, the surface of HY-Cu₂O (100) shows a degree of roughness and a visible coating, indicating that the material has been successfully modified with long-chain alkanes. At the same magnification, the image clarity of Cu₂O (100) [Figure 2A-C] is higher than that of HY-Cu₂O (100) [Figure 2D-F], indicating that thiol modification lowers the electrical conductivity of the catalyst material.

Selective Energy-Dispersive X-ray Spectroscopy (EDS) analysis of Cu₂O (100) and HY-Cu₂O (100) revealed [Supplementary Figure 1] that sulfur was detected in HY-Cu₂O (100) but not in Cu₂O (100); Additionally, the carbon content on the surface of HY-Cu₂O (100) was higher than that of Cu₂O (100), further confirming the successful modification with hydrophobic long-chain alkanes. Since both samples were sputtered with Au before testing, Au signals were present in the EDS spectra of both samples. Under the same conditions, the Cu content in HY-Cu₂O (100) was lower than that in Cu₂O (100), which is attributed to the surface of HY-Cu₂O (100) being covered by saturated long-chain alkanes.

Transmission Electron Microscopy (TEM) results [Figure 2G] show that the prepared HY-Cu₂O (100) nanostructures are intact, with a particle size of approximately 300 nm, and retain a cubic morphology without significant changes. Figure 2H and I shows that the hydrophobic long-chain alkanes uniformly coat the Cu₂O (100) surface, forming an amorphous structure, which also accounts for the increased surface roughness.

Selective-area electron diffraction [Figure 2J] of HY-Cu₂O (100) was performed from the direction shown in Figure 2I. The diffraction spots formed a regular two-dimensional lattice pattern, with no polycrystalline rings or diffuse spots observed, indicating that these Cu₂O nanoparticles have a single-crystal structure. The square arrangement of the spots further confirms their cubic crystal system. The High-Resolution Transmission Electron Microscopy (HRTEM) image [Figure 2I] reveals an interplanar spacing of 0.242 nm, corresponding to the (111) plane of Cu₂O. HAADF-STEM and EDS analysis [Figure 2K-P] show that S and C elements are uniformly distributed on the Cu₂O outer layer.

Infrared spectroscopy further confirmed the successful modification of the saturated long-chain alkane layer [Figure 3A-C]. The spectrum of HY-Cu₂O (100) exhibits characteristic peaks at 2,917 cm^-1 and 2,849 cm^-1, whereas Cu₂O (100) shows no such peaks at these wavelengths; the literature indicates that these peaks correspond to the antisymmetric and symmetric stretching vibrations of the -CH₂ group, and a peak corresponding to C-C stretching vibrations (1,468 cm^-1) is also observed. This confirms the presence of long-chain alkanes on the HY-Cu₂O (100) surface (as shown in Figure 3A)^[24].

The X-ray Diffraction (XRD) pattern [Figure 3D] shows that both samples exhibit the strongest diffraction peak at 36.5°, corresponding to the Cu₂O (111) crystal plane, consistent with the HRTEM analysis [Figure 2I]. The subtle differences in the XRD patterns indicate that the crystal structure of HY-Cu₂O (100) remains consistent with that of Cu₂O (100), suggesting that the long-chain alkane modification primarily affects the surface properties of the material rather than its crystalline phase composition. Hydrophobicity test results [Figure 3E and F] show that Cu₂O (100) exhibits hydrophilic properties (contact angle of 46.1°), whereas HY-Cu₂O (100) exhibits enhanced hydrophobicity (contact angle of 133.5°). The increased hydrophobicity of the catalyst surface plays a crucial role in promoting the effective diffusion of reactant gases during the electrochemical reduction of CO₂.

X-ray Photoelectron Spectroscopy (XPS) characterization was used to verify the surface elemental chemistry of the prepared catalyst [Figure 4]. As shown in Supplementary Figure 2, Cu, O, and C were detected in the full XPS spectra of both HY-Cu₂O (100) and Cu₂O (100), whereas S was detected only in HY-Cu₂O (100), suggesting that Cu-S bonds may have formed on its surface. The full-scan range was 0-1,250 eV.

The S 2p fine-line spectrum [Figure 4E and F] shows that HY-Cu₂O (100) exhibits distinct characteristic peaks at 162.3 eV and 163.4 eV, attributed to S 2p_3/2 and S 2p_1/2, respectively, corresponding to the formation of Cu-S bonds. In contrast, no S signals were detected in Cu₂O (100)^[25,26]. Combining Infrared Spectroscopy (IR) spectroscopy and XPS data, it can be clearly confirmed that the Cu₂O (100) surface has been successfully functionalized with octadecanethiol chains.

In the O 1s spectrum [Figure 4C and D], the characteristic peaks at 530.6 eV and 530.3 eV are attributed to lattice oxygen, while the peaks at 532.2 eV and 531.6 eV correspond to adsorbed oxygen^[27,28]. The ratio of lattice oxygen to adsorbed oxygen in HY-Cu₂O (100) is significantly higher than that in Cu₂O (100). This is not due to a decrease in the absolute amount of adsorbed oxygen, but rather stems from the formation of a Cu-S coordination protective layer by thiol molecules with surface Cu atoms, which effectively passivates surface defect sites prone to oxygen loss, thereby suppressing the loss of lattice oxygen and increasing the relative signal intensity of lattice oxygen. This phenomenon reflects microstructural changes in the Cu₂O (100) surface following long-chain alkane modification, characterized by a reduced concentration of oxygen vacancies and improved lattice integrity. Furthermore, the characteristic peak at 284.8 eV in the C 1s spectrum [Figure 4I] further confirms the successful modification of the HY-Cu₂O (100) surface with long-chain alkanes.

To reveal how hydrophobic long alkane chain modification regulates the electrochemistry of Cu-based catalysts, cyclic voltammetry measurements in an H-type cell and CO₂ electroreduction evaluations in a flow cell were carried out for bare Cu₂O (100) and hydrophobically modified HY-Cu₂O (100) [Figure 5].

By comparing the Cyclic Voltammetry (CV) curves of the two catalysts in Ar and CO₂ atmospheres [Figure 5A and B], we found that in the potential range from 0 to -0.5 V [vs. Reversible Hydrogen Electrode (RHE)], the response curve under CO₂ atmosphere was closer to that under Ar atmosphere. The multiple redox peaks observed in the figure are attributed to the interconversion of copper ions in different oxidation states. At the same time, the presence of CO₂ inhibits the reversible redox process observed in argon. As shown in Figure 5C and D, in the low-potential range, the absolute value of the cathodic current density of HY-Cu₂O (100) in the CO₂ atmosphere exceeds -300 mA cm^-2, which is higher than that of Cu₂O (100) (approximately -200 mA cm^-2). In contrast, the current densities of both materials in the Ar atmosphere are similar (approximately -180 mA cm^-2). This difference indicates that HY-Cu₂O (100) exhibits a higher reaction current in a CO₂ atmosphere, reflecting its superior CO₂ electro-reduction activity.

Figure 5E further confirms that hydrophobic modification increases the C₂H₄ yield. In constant-current testing, the C₂H₄ yield [Faradaic Efficiency (FE) (C₂H₄)] of HY-Cu₂O (100) increased gradually with current density, peaking at 45.2% at -250 mA cm^-2, outperforming the yield of Cu₂O (100) (26.0%). Compared with relevant studies on hydrophobicity, this study demonstrates superior performance in current control, catalyst selection, and selectivity for the multi-carbon product ethylene [Supplementary Table 1]. The difference between the two was particularly pronounced in the -150 to -250 mA cm^-2. Significant C₂H₄ production was detected at all four current densities, indicating that this potential range satisfies the kinetic requirements for the C-C coupling reaction.

At the same time, the hydrophobic modification effectively suppressed HER [Figure 5F]. At all tested current densities, the FE(H₂) of HY-Cu₂O (100) was lower than that of Cu₂O (100), with the greatest difference observed at -150 mA cm^-2. This phenomenon indicates that thiol modification effectively suppressed the competitive hydrogen-evolution pathway while enhancing selectivity for C₂H₄. Furthermore, the distribution of other CO₂ reduction products on the modified catalyst also changed [Figure 5G and H]: the CO production efficiency of HY-Cu₂O (100) increased markedly. Still, it then gradually decreased with increasing current density, contrary to the trend observed for C₂H₄. This phenomenon suggests that more surface-adsorbed *CO intermediates participate in the C-C coupling pathway, thereby reducing the proportion that desorbs to form CO.

In terms of partial current densities [Figure 5I and J], the C₂H₄ partial current of HY-Cu₂O (100) is higher than that of Cu₂O (100), while the H₂ partial current is effectively suppressed. At the same catalyst loading, the C₂H₄ yield of HY-Cu₂O (100) was increased [Figure 5K], while the H₂ yield correspondingly decreased [Figure 5L], further confirming the synergistic effect of hydrophobic modification in enhancing the selectivity and activity of CO₂ electroreduction. Meanwhile, stability tests [Supplementary Figures 3 and 4] demonstrated that hydrophobic modification also improves the stability of CO₂ electroreduction.

Electrochemical impedance spectroscopy (EIS) was further employed to investigate how the surface modification with thiol molecules alters the charge transfer resistance (R_ct) of the Cu₂O (100) electrode [Figure 6A-F]. The amplitude of the arc in the high-frequency region of the measured data reflects the magnitude of the interfacial transfer resistance. It is generally accepted that a smaller arc indicates a lower R_ct and, consequently, a lower reaction resistance at the electrochemical interface^[24,29-32]. The increased charge transfer resistance of HY-Cu₂O compared to pure Cu₂O at high potentials (less negative potentials) is primarily due to the hydrophobic nature of the octadecanethiol mercaptan modification. This long-chain alkane layer acts as an insulating barrier on the electrode surface, which hinders the transport of electrolyte ions to the active sites and impedes electron transfer. We performed circuit fitting on the impedance data obtained at the open-circuit potential (OCP) and at 0.55 V vs. RHE, with the results shown in Figure 6A, B, D and E. The fitted circuit resembles the classic electrochemical interfacial transfer circuit, including solution resistance (R_s), constant phase element (CPE), R_ct, and warburg element (W).

At the OCP point [Figure 6A and D], HY-Cu₂O (100) exhibited a slightly larger arc radius than Cu₂O (100) in the high-frequency region, which implies that the R_ct value of HY-Cu₂O (100) was higher than that of Cu₂O (100); however, when the potential was reduced to 0.55 V (relative to RHE) [Figure 6B and E], the R_ct value of HY-Cu₂O (100) was comparable to that of Cu₂O (100); When the potential was further reduced to 0.35 V and 0.15 V (vs. RHE) [Figure 6C and F], HY-Cu₂O (100) exhibited a slightly smaller arc radius than Cu₂O (100) in the high-frequency region; it can be concluded that at these two potentials, the interfacial resistance of HY-Cu₂O (100) is lower than that of Cu₂O (100). Comparing the gradual decrease in potential reveals that the arc radius of HY-Cu₂O (100) gradually decreases, indicating that during the E-CO₂RR reaction, the reaction kinetics of HY-Cu₂O (100) gradually accelerate as the potential decreases, suggesting that the catalytic activity of HY-Cu₂O (100) is more readily enhanced.

To evaluate the electrochemically active surface area (ECASA) of these catalysts, we measured the double-layer capacitance in the non-Faradaic region of the catalysts [Figure 6G-I]^[29]. The results show that the cyclic voltammogram of Cu₂O (100) exhibits a rugby-ball shape [Figure 6G], indicating certain capacitive properties; in contrast, the curve for HY-Cu₂O (100) is relatively flat [Figure 6H]. In Figure 6I, the double-layer capacitance (C_dl) value of HY-Cu₂O (100) is 28.0 μF cm^-2, which is lower than the 71.5 μF cm^-2 value for Cu₂O (100). The long alkyl chain modification of the Cu-based catalyst increased the hydrophobicity of HY-Cu₂O (100), reducing the extent of contact between the catalyst surface and the liquid-phase electrolyte; this also explains the lower C_dl value. Since ECASA shows a positive correlation with C_dl, the ECASA of HY-Cu₂O (100) is also lower than that of Cu₂O (100)^[30]. This is a direct consequence of the hydrophobic octadecanethiol mercaptan layer. This long-chain alkane modification renders the surface hydrophobic, reducing its wettability and thus decreasing the effective contact area with the aqueous electrolyte.

In situ infrared spectroscopy was used to investigate the mechanism by which hydrophobic long-chain alkanes act as intermediates in the electrocatalytic reduction of carbon dioxide. The adsorption behavior and evolution of intermediates on the surfaces of HY-Cu₂O (100) and Cu₂O (100) during the electrocatalytic reduction of carbon dioxide were monitored^[31,32].

According to existing research, the O-H stretching vibrations of water molecules in the 3,600-3,000 cm^-1 range can be broadly classified into three categories: “ice-like water” (4HB-H₂O) at 3,250 cm^-1 with four hydrogen bonds; “asymmetric hydrogen-bonded water” (asymmetric HB-H₂O) at approximately 3,450 cm^-1 with one to three hydrogen bonds; and “cationic water” (K∙H₂O) at 3,610 cm^-1 that does not form hydrogen bonds, which in this system specifically refers to water molecules coordinated with K⁺. It is also called “isolated water”^[33]. This spectral region is highly sensitive to hydrogen-bonding structures; a negative shift in potential typically enhances hydrogen-bonding interactions, leading to a red shift of the spectral peak.

In situ IR spectra recorded in a CO₂ atmosphere [Figure 7A and B] show that as the potential becomes more negative, the K∙H₂O content at the HY-Cu₂O (100) interface increases, whereas no obvious changes are observed at the Cu₂O (100) interface. At the same time, the increase in 4HB-H₂O and asymmetric HB-H₂O at the Cu₂O (100) interface was much higher than that at the HY-Cu₂O (100) interface. This indicates that hydrophobic modification effectively regulates the water structure at the electrode-electrolyte interface, promoting the enrichment of K∙H₂O. Furthermore, we observed a red shift in the interface water spectral peaks as the potential became more negative, which is attributed to the Vibrational Stark Effect (VSE), a shift in molecular vibrational energy levels under the influence of an external electric field. The more negative the potential, the higher the electron density on the electrode surface; this increases the transfer of lone pair electrons from oxygen in water molecules to the electrode surface, leading to a reduction in the O-H bond order and a decrease in the vibrational frequency, thereby causing the red shift^[34]. Test results under an Ar atmosphere [Figure 7C and D] further demonstrate that even in the absence of CO₂, the K∙H₂O content at the HY-Cu₂O (100) interface remains higher than that at the Cu₂O (100) interface, and its trend is consistent with that observed under a CO₂ atmosphere. This indicates that the regulatory effect of the hydrophobic chains on the interfacial water structure is independent of the atmosphere, further validating their modification effect.

It is worth noting that during the in situ ATR-SEIRAS measurements [Figure 7A and C], we also detected characteristic peaks of the -CH₂ antisymmetric/symmetric stretching vibrations and the -CH₃ antisymmetric stretching vibration at 2,925 cm^-1, 2,850 cm^-1, and 2,960 cm^-1, respectively, further confirming the presence of long-chain alkanes on the HY-Cu₂O (100) surface (The shift in the position corresponding to the ATR-FTIR spectrum in Figure 3 is due to differences in testing conditions, specifically the absence of electrolyte and the application of an external bias). These peaks were more pronounced in an Ar atmosphere than in a CO₂ atmosphere, possibly because CO₂-reduction intermediates occupy active surface sites, disrupting the ordered arrangement and orientation of the alkyl chains and partially masking their infrared signals.

Figure 8A-F shows the in situ ATR-SEIRAS spectra of HY-Cu₂O (100) and Cu₂O (100) under different applied potentials. The peak near 1,608 cm^-1 in Figure 8B is assigned to the O-H bending vibration mode (δH₂O) of water molecules at the electrode interface, and this characteristic feature is observable in all four sets of spectra^[34,35]. Under the inert Ar atmosphere [Figure 8A], the bending vibration peak of interfacial water on hydrophobic HY-Cu₂O appears at 1,661 cm^-1, which is significantly higher than that at 1,613 cm^-1 on hydrophilic Cu₂O, indicating the formation of a more ordered hydrogen-bond network on the hydrophobic surface^[35]. The characteristic signals of the key intermediates *CO, *COO^-, *CHO, and O*C*CHO were detected in Figure 8B and E, whereas no corresponding signals appeared in Figure 8A and D, confirming the presence of these intermediates during the CO₂ reduction process. The positive peak at approximately 1,382 cm^-1 is attributed to the *COO^- intermediate [Figure 8B], which is formed by the adsorption and activation of CO₂ on the catalyst surface and is presumed to be further converted into key intermediates such as *COOH^[21,36].

The positive peak observed in the 2,100-2,000 cm^-1 range [Figure 8C and F] was identified as the top-site adsorption species of CO (*CO_atop), which is formed by the further reduction of *COOH. The position of this spectral band shifts toward the red or blue as the potential becomes more negative, a phenomenon attributable to the VSE^[37]. *CO_atop can be hydrogenated to form *CHO (1,673 cm^-1), which is considered a key intermediate in the CH₄ formation pathway. However, based on the E-CO₂RR test results discussed earlier, the FE (CH₄) values for both Cu₂O (100) and HY-Cu₂O (100) are extremely low. Therefore, *CHO does not primarily convert to methane; instead, it participates in other reaction pathways. Given the high energy barrier for *CO dimerization on the surface, it is generally accepted in the literature that *CO couples with *CHO to form O*C*CHO (1,446 cm^-1 and 1,326 cm^-1) [Figure 8E]^[38,39], which serves as an important precursor for the formation of C₂H₄ and other C₂₊ products.

Combining this with electrocatalytic performance data, the trends in K∙H₂O content and C₂H₄ production efficiency as a function of potential are similar, suggesting a possible correlation between the two. Based on the preceding discussion, we propose the following hypothesis: if the content of 4HB-H₂O and asymmetric HB-H₂O at the interface is high, *CO will be less likely to interact with isolated water molecules, which will hinder its protonation; conversely, if the concentration of K∙H₂O at the interface is high, the hydrogen bonding between water molecules is weakened, which is more conducive to the hydrogenation of *CO, thereby promoting subsequent C-C coupling. That is, K∙H₂O plays a key role in facilitating the conversion of CO₂ to C₂H₄.

The catalysts used in the electrochemical tests were all modified with Nafion (via ultrasonic mixing) to leverage its binding and mass-transfer properties. Because Nafion contains organic functional groups such as sulfonic acid (-SO₃H), C-F, and C-C bonds, it exhibits distinct absorption peaks in the infrared spectrum^[40]. By comparing the test results under the two atmospheres, we can effectively distinguish the characteristic peaks of the adsorbed intermediates (marked in green in Figure 8) from other background signals (marked in blue).

The formation of C₂₊ products is closely related to the local *CO_atop coverage. Under the same test conditions, the *CO_atop signal intensity of HY-Cu₂O (100) was higher than that of Cu₂O (100) [Figure 8C and F], indicating that it can more effectively stabilize the *COatop intermediate and promote the C-C coupling reaction^[41-43]. As shown in Figure 8G, the peak area of the *CO_atop signal for both catalysts first increases and then decreases with potential, but the peak for HY-Cu₂O (100) appears at a more negative potential. Taking HY-Cu₂O (100) as an example, its *CO_atop peak area continued to increase up to -0.81 V (relative to RHE); as shown in Figure 8G, at the same potential, the integrated *CO_atop peak area on the HY-Cu₂O (100) surface was greater than that on the Cu₂O (100) surface. This indicates that the surface CO coverage increases in tandem and reaches a maximum at this potential, demonstrating that hydrophobic modification enhances the adsorption capacity of the *CO intermediate. This high local coverage of *CO is a prerequisite for the subsequent C-C coupling step. At more negative potentials, the peak area gradually decreases as *CO_atop is consumed by hydrogenation and C-C coupling reactions.

In the range of -0.4 V to -1.26 V (vs. RHE), the peak area of *CO_atop on HY-Cu₂O (100) consistently exceeds that on Cu₂O (100) [Figure 8G and H], and the intensity ratio remains greater than 1 throughout this range [Figure 8H], particularly in the range of -1.0 V to -1.26 V. Combined with electrochemical performance data, this potential range corresponds precisely to the critical region where C-C coupling occurs. Higher *CO_atop coverage, together with optimized K∙H₂O supply, jointly promotes *CO_atop conversion, thereby enhancing its hydrogenation and coupling reactions with *CHO.

In the range of -0.45 V to -1.05 V (vs. RHE), the vibrational wavenumbers of CO adsorbed atop the HY-Cu₂O(100) surface (*CO_atop) are generally higher than those on Cu₂O(100). As shown in Figure 8C and F, no characteristic signals of bridge-site adsorbed CO were observed when species other than the top-site adsorbed *CO_atop were excluded, ruling out peak position differences caused by changes in adsorption configuration. Two distinct Stark-shifted regions were observed in Figure 8I, with a critical point at -0.7 V (vs. RHE). Between -0.45 and -0.7 V vs. RHE, the *CO stretching vibration undergoes a blue shift. Combined with Figure 8G and H, the higher peak area of *CO_atop compared to that of the pristine Cu₂O indicates that the hydrophobic interface of HY-Cu₂O(100) enhances the local CO coverage. Adjacent adsorbed CO dipoles repel each other, increasing the effective force constant of the C≡O bond, indicating that the dipole–dipole coupling effect between adjacent adsorbed CO molecules dominates this blue shift behavior, corresponding to a tendency for CO to desorb; When the potential becomes more negative than -0.7 V, νCO shifts toward red, while the peak area of CO_atop reaches a maximum near -0.81 V and then gradually decreases. The red shift of ν(*CO) results from the vibrational Stark effect under strong interfacial electric fields interacting with the intrinsic Cu-CO orbitals; the C-O bond is weakened, making *CO more prone to hydrogenation and coupling with *CHO. Higher local *CO coverage kinetically promotes C-C coupling between *CO and *CHO to form the O*C*CHO intermediate; this result is consistent with electrochemical performance and DFT theoretical calculations.

To elucidate the underlying mechanism by which hydrophobic modification enhances the ethylene production efficiency on Cu₂O (100) electrodes, clarify the regulatory role of the K∙H₂O species in a hydrophobic environment, and reveal its reaction pathway, this study employed density functional theory (DFT) to conduct theoretical calculations^[44], comparing and analyzing the 4HB-H₂O, asymmetric HB-H₂O, and K∙H₂O - on the Cu₂O (100) surface. These calculations not only systematically reveal the formation mechanisms of ethylene and C₂₊ products on the Cu₂O (100) surface but also clarify the key role of K∙H₂O in the electroreduction of CO₂ to ethylene.

Based on the research approach outlined above, this paper constructs an electrode model based on the Cu₂O (100) crystal plane and introduces three types of hydration structures - K∙H₂O, asymmetric hydrogen-bonded water, and 4HB-H₂O - above the electrode interface, corresponding to the interfacial water environment before and after hydrophobic modification, respectively. The free energy calculation models for the three structures are shown in Figure 9A-C. The formation process of the key intermediate O*C*CHO in the K∙H₂O environment is illustrated in Figure 9D, while the corresponding reaction processes for the other two comparison models are shown in Supplementary Figures 5 and 6 in the Supplementary Materials.

As shown in Figure 9E, the calculation results indicate that on a hydrophobic Cu₂O (100) surface enriched with K∙H₂O, the reaction energy barrier (ΔG = 0.38 eV) for the hydrogenation of *CO to *CHO is lower than that of the two unmodified control systems (ΔG = 0.51 eV/0.70 eV). This indicates that K∙H₂O enriched at the interface effectively promotes the hydrogenation of *CO, lowering the reaction energy barrier for Rate-Determining Step (RDS) and thereby facilitating the subsequent C-C coupling reaction. Based on this, it can be inferred that the presence of K∙H₂O promotes the conversion of *CO adsorbed on the surface (*CO_atop)^[24,45,46]. At the same time, the formation of O*C*CHO is an exothermic process, indicating that C-C coupling does not determine the rate of multi-carbon product formation. It can be confirmed that the hydrogenation of *CO to form *CHO is the rate-determining step for the formation of ethylene and C₂₊ products. Furthermore, on hydrophobic interfaces enriched with K∙H₂O, the reaction energy barrier for the coupling of *CO and *CHO to form O*C*CHO is lower than in the two unmodified systems, indicating that K∙H₂O not only accelerates the *CO-*CHO coupling process but also promotes the formation of the key intermediate O*C*CHO. These theoretical calculation results are consistent with experimental observations.

In summary, hydrophobic interfaces can enrich “free water” species (K∙H₂O) that coordinate with K⁺, thereby inhibiting the formation of “ice-like water” structures characterized by strong hydrogen bonding. The K∙H₂O species enriched at the interface can effectively accelerate the protonation and hydrogenation of *CO, lower the RDS energy barrier, thereby promoting subsequent C-C coupling reactions and driving the dimerization of *CO with *CHO, ultimately facilitating the formation of key intermediates for ethylene production.