Intimate atomic Cu-Ag interfaces for high CO2RR selectivity towards CH4 at low over potential

Developing highly efficient electrochemical catalysts for carbon dioxide reduction reaction (CO2RR) provides a solution to battle global warming issues resulting from ever-increasing carbon footprint due to human activities. Copper (Cu) is known for its efficiency in CO2RR towards value-added hydrocarbons; hence its unique structural properties along with various Cu alloys have been extensively explored in the past decade. Here, we demonstrate a two-step approach to achieve intimate atomic Cu-Ag interfaces on the surface of Cu nanowires, which show greatly improved CO2RR selectivity towards methane (CH4). The specially designed Cu-Ag interfaces showed an impressive maximum Faradaic efficiency (FE) of 72% towards CH4 production at −1.17 V (vs. reversible hydrogen electrode (RHE)).


Introduction
Humanity is at the brink of fossil fuel exhaustion and faces challenges of global climate change. Carbon dioxide (CO2) emission is a primary driver of global warming and the reducing pH levels of the ocean. Meanwhile fossil fuels are not renewable and will eventually deplete. Creating a closed-loop process to recycle CO2 to value-added fuels is a promising option to mitigate global warming and grant inexhaustible energy sources [1][2][3][4]. Developing efficient electrochemical catalysts for CO2 reduction reaction (CO2RR) is a prerequisite for establishing a carbon recycle loop and renewable energy technologies.
In the past decade, electrochemical CO2RR has remarked prominent achievements in both scientific apprehension and technological developments. Among many electrocatalysts, copper (Cu) is the only known electrochemical catalyst to convert CO2 to alternative energy fuels and hydrocarbons (especially methane (CH4)) with sufficient current density and selectivity [5]. However, a mixture of primary products, competition with hydrogen evolution reaction (HER), and required high overpotential for CO2RR on monometallic Cu still poses challenges. Therefore, designing Cu-based catalysts with high selectivity at low overpotentials is of great interest [6].
The prior art of research has improved Cu catalyst' s modulation of structure defects [7][8][9][10], shapes [11][12][13], size [14], and chemical states [15][16][17][18][19][20] of Cu. For example, grain boundaries (GBs) exhibited ~ 2.5 times higher CO2RR activity with less competitive reaction (HER) [21]. Cheng et al. reported that Cu's surface steps, having a combination of one strong and one weak CO binding site, enhance C2 productions with reduced formation energy of the rate determining step (*OCCOH) to 0.52 eV [22]. Alloying Cu with a second metal is another attractive way to design catalysts [23]. However, Cu usually lost its extraordinary CO2RR capability of producing hydrocarbon and oxygenates, having C2 and C3 carbons by forming an alloy with other elements [24][25][26][27]. Thus, interface Cu with neighboring unmixable second elements has been proposed to design Cu-based catalysts to retain the Cu's unique CO2RR capability [27,28]. Because of this complexity, the research in CuM (M denotes another metal element) alloy catalysts for electrochemical CO2RR has not been sufficiently explored or compared with pure Cu. Silver (Ag) is a promising candidate to achieve such an unmixable Cu-M interface design because Ag and Cu are known for their thermodynamical immiscibility over all compositions at room temperature [29][30][31][32][33][34]. For example, Huang et al. reported that the interface between Cu catalysts and Ag catalysts was the crucial active site to enhance CO2RR over pure Cu catalysts [34]. However, the interface between Cu and Ag has been limited because the boundary was miserly obtained between Cu and Ag catalysts. Maximizing the Cu and Ag interfaces at the atomic level is highly desired but challenging.
Herein, we report a two-step approach to build the interface between Cu and Ag at the atomic level ( Fig. 1). Cu nanowires (CuNWs) were first synthesized and followed by galvanic replacement from Cu to Ag to achieve in situ formation of CuAg ensembles, which builds CuNWs with rich Cu-Ag interfaces. The attractive Cu-Ag interfaces showed a dramatic change in CO2RR selectivity from C2H4 to CH4, which remarked a 63.29% ± 4.85% FECH 4 (FE means Faradaic efficiency) at −1.12 ± 0.01 V (vs. reversible hydrogen electrode (RHE), referenced to all potentials) and an impressive maximum FECH 4 of 72% at low potential of −1.17 V

Results
We synthesized bimetallic CuAg nanowires (CuAgNWs) through a synthesis of CuNWs followed by galvanic replacement of Cu to Ag (Figs. 1(a) and 1(b)). To be specific, 22 mg of CuCl2·2H2O, 50 mg of glucose, and 180 mg of hexadecylamine (HDA) were mixed in 10 mL of deionized (DI) water (18.2 MΩ/cm) under sonication for 15 min, then heated at 100 °C for 8 h in an oil bath. After the reaction solution cooled down to room temperature, 1.21 mg of AgCH3CO2 and 0.84 mg of imidazole were added to the CuNW solution, which was kept at 50 °C for 25 or 60 min without stirring for galvanic replacements from Cu to Ag. The relatively high standard reduction potential of Ag ((Ag + + e − → Ag(s), E° = 0.80 V) (standard hydrogen electrode)) compared to Cu ((Cu 2+ + 2e − → Cu(s), E° = 0.34 V (standard hydrogen electrode)) [35] drives replacements from Cu to Ag on the surface of CuNWs. The synthesized CuAgNWs were washed five times with hexane/ethanol mixture and were collected by centrifuge. The CuAgNWs were characterized by transmission electron microscopy (TEM), Cs-corrected highangle annular dark-field imaging scanning transmission electron microscope (HAADF STEM), energy-dispersive X-ray spectroscopy (EDX), and powder X-ray diffraction (PXRD). The average size of CuNWs (25 ± 7.7 nm) was obtained by averaging more than 100 NWs in width ( Fig. S1 in the Electronic Supplementary Material (ESM)). Imidazole, a small molecular capping agent, was pivotal in slowing down the fast Ag galvanic replacement reaction on the surface of CuNWs and keeping one-dimensional structure of CuAgNWs without disintegrating into CuAg nanoparticles (NPs) (Fig. S2 in the ESM). The pure CuNWs showed flat and clean surfaces, whereas the uneven surface was observed after the Ag galvanic replacement on the surface of CuNWs (Fig. S3 in the ESM). The PXRD peaks of CuAgNWs indicate no Bragg angle shift of Cu{111} and Ag{111} after the galvanic replacement of Ag for 25 min (Fig. 2), which indicates an unmixable pure Cu and Ag phase system without forming the CuAg alloy. High intensity of Ag{111} plane in CuAgNWs with the galvanic replacement for 60 min illustrates more replacements from Cu to Ag element for longer galvanic reaction times.
To quantify the bulk and surface composition of CuAgNWs, we carried out inductively coupled plasma atomic emission spectroscopy (ICP-AES) and X-ray photoelectron spectroscopy (XPS). Table 1 shows a summary of ICP and XPS composition analyses. The bulk composition from ICP analysis showed 82% of Cu and 18% of Ag (Cu8.2Ag1.8NWs) in CuAgNWs with the galvanic replacement for 60 min; and 90% of Cu and 10% of Ag (Cu9Ag1NWs) with the galvanic replacement for 25 min. In the case of XPS, the calculated electron inelastic mean free path of Cu at 992.3 eV and Ag at 365 eV are 1.67 and 0.727 nm, respectively [36], which reveals a surface limited component analysis of XPS for Cu and Ag metal. XPS analysis of Cu8.2Ag1.8NWs (60 min) illustrates two times higher Ag concentration than the ICP analysis, indicating richer Ag concentration on the surface of Cu8.2Ag1.8NWs. While, Cu9Ag1 which is synthesized from 25 min galvanic replacements shows comparable Ag concentration at bulk and surface of Cu9Ag1NWs, indicating that the galvanic replacement only occurs at the CuNW surface and Ag mainly stays near the surface.
Indeed Cs-corrected HAADF STEM images of CuNWs and EDX maps of the Cu9Ag1NWs confirmed that galvanic replacement of Cu to Ag on the surface (Fig. S4 in the ESM). The EDX maps showed Ag mainly located on the surface of CuNWs, and formed separate Ag phases from Cu phases in Cu9Ag1NWs (Figs. S4(c) and S4(d) in the ESM). Cu K and Ag L EDX maps illustrated that a thin layer of Ag covers the surface of CuNWs (Figs. S4(e) and S4(h) in the ESM). However, with longer (60 min) galvanic replacement, the Cu8.2Ag1.8NWs showed more uneven surfaces and much thicker surface coverage of Ag compared to Cu9Ag1NWs (25 min galvanic reaction) as shown in Fig. S5 in the ESM.  We further found that electrochemical treatment of the as-prepared Cu9Ag1NWs can drive Cu to the surface of Cu9Ag1NWs and generate atomic Cu-Ag interfaces with more exposed Cu surface. The Cu9Ag1NW catalysts inks were prepared by mixing 4 mg of the CuAgNWs in 1 mL ethanol and 10 μL of 5% Nafion 117. We dropped 10 μL catalyst inks on the 1 cm diameter glassy carbon electrode. Subsequently, by using a high reduction bias (V = −1.05 V), we activated the CuAgNWs in CO2-saturated 0.1 M KHCO3 solution for 2 h. After such treatment, Fig. 3 showed the generated atomic Cu-Ag interfaces on the surface of Cu9Ag1NWs. EDX maps showed that Cu rose above the top of Ag-covered surface of as-prepared Cu9Ag1 NWs (Figs. 3(a) and 3(d), and Fig. S6 in the ESM). The interfaces between Cu ensembles and Ag layers were quite prominent in EDX maps of Cu component in Figs. 3

(b) and 3(e), and Ag component in Figs. 3(c) and 3(f).
The migration of Cu atoms outward to the surface of the NWs can be understood based on the differential binding strengths of Cu and Ag to CO2RR intermediates or products. Back et al. [37] reported that Cu (211) Figure S7 in the ESM illustrates that the Cu component went to the top surface at all purging conditions. Interestingly, in the Cu8.2Ag1.8NWs, the Cu and Ag elements separated into Cu particles and Ag straw structures (Fig. S8 in the ESM). This indicates a thick enough Ag surface layer might limit the movement of Cu towards the surface to rise above the Ag surface.
We conducted CO2RR with CO2-saturated 0.1 M KHCO3 (pH 6.8) in a gas-tight H-cell at room temperature under atmospheric pressure. We analyzed effluent gas/liquid products at different applied potentials between −1.02 and −1.25 V (vs. RHE). To compare the CO2RR performance of pure Ag, we also synthesized and conducted CO2RR test of AgNPs (Fig. S9 in the ESM).
The reaction pathways of CH4 and C2H4 share a common *COH intermediate, which deviates to CH4 with *HCOH and C2H4 with *OC-COH (*active sites of catalysts) [39]. Thus, catalyst's availability to adsorb H (Had) compared to adsorb CO (COad) induces higher CH4 selectivity over C2H4 selectivity. Chen et al. reported that hydrogen binding energy increases by ~ 1 eV/V while the CO binding energy varies little with applied potentials [39]. Assuming dominant hydrogen coverage at more negative than −0.8 V (vs. RHE) [39], at −1.12 V (vs. RHE) on the surface of CuAgNWs, the Cu portion would be covered by hydrogen (Cu-H*) while the surface of Ag was still dominant by CO (Ag-CO*). CO dominance on Ag surface was consistent with our observation of FECO over 90% at ~ −1.13 V (vs. RHE) on AgNPs (Fig. 4(d)). On CuAgNWs, the diffusion of CO from Ag section to the Cu section is likely more efficient due to the short diffusion length. Thus, efficient feeding of CO to hydrogen-covered Cu surface could make it favorable to generate *COH, and the dominant hydrogen coverage on the Cu surface might drive *COH to *HCOH and finally CH4. To further investigate the importance of the different interfaces between Cu and Ag components for high CH4 generation, we compared FEs of Cu9Ag1NWs and Cu8.2Ag1.8NWs. Both Cu9Ag1NWs and Cu8.2Ag1.8NWs showed higher selectivity of CO ( > 60% FECO) at ~ -1.05 V (vs. RHE) than pure CuNWs (Figs. 4(a)-4(c)), confirming *CO contribution from the Ag component. However, the Cu8.2Ag1.8NWs still showed high CO selectivity (53.21% ± 20.00%) with low CH4 production (10.55% ± 9.15%) at −1.13% ± 0.02% V (vs. RHE) (Tables S1 and S2 in the ESM), while Cu9Ag1NWs already demonstrated over 60% FECH 4 . HRTEM and EDX showed larger separations between Cu and Ag components in Cu8.2Ag1.8NWs (Fig. S8 in the ESM) than Cu9Ag1NWs (Fig. S3 in the ESM).
This indicates that an intimate atomic level Cu-Ag interface between the Cu and Ag components is necessary to promote the synergistic effect of CO-Ag* and H-Cu* for high CH4 selectivity. To the best of our knowledge, the Cu9Ag1NWs provide the highest FECH 4 at the lowest applied potential 72% of FECH 4 at −1.17 V (vs. RHE) in H-cell with standard glassy carbon electrode at room temperature and atmospheric pressure (0.1 M KHCO3) compared to all other materials reported in the literature to date (Table 2).

Conclusion
We successfully generated atomic Cu-Ag ensembles via ex-situ galvanic replacement from Cu to Ag, followed by an in-situ electrochemical activation approach. The atomic Cu-Ag ensemble interfaces showed a change of CO2RR selectivity from C2H4 to CH4, which remarked the highest FECH 4 at the lowest applied potential 72% of FECH 4 at −1.17 V (vs. RHE) in H-cell. These findings suggest an effective way to generate unmixable atomic ensemble Cu-Ag interfaces to enhance CH4 selectivity with lower over potential under operando conditions. This approach can be expanded to other unmixable metal atoms to engineer the atomic ensemble interfaces for desired catalytic properties. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.