The transformation of CO2 into value-added products using renewable power offers a promising strategy to solve the energy and environment crisis, and therefore has drawn increasing interest in academic and industry.1–6 Various carbon-based chemicals can be obtained through the electrochemical CO2 reduction reaction (eCO2RR), among which, the deeply reduced ones, such as multi-carbon (C2+) molecules, are preferred. Copper (Cu) based catalysts are known as the only family which can produce considerable amounts of C2+ products in eCO2RR.7–9 However, it is difficult to achieve high C2+ selectivity because the formation of such products involves an eight or more electron transfer process, while the less-electron transfer involved products, such as CO and CHOOH, are easier to form.10,11
The C−C coupling of eCO2RR intermediates (CO*, CHO* and et.al) results in producing C2+ molecules or else to C1 products.10–16 Oxide-derived Cu materials have been reported to have good selectivity for C2+ products. Surface residual O on Cu has been suggested to stabilize CO*, promoting C−C coupling.17–20 However, in some oxide-derived Cu involved systems, O is found to disappear in eCO2RR.21–22 In addition to O, residual Cu (ǀ) has also been revealed to assist the C−C coupling possess.23–27 Up to now, Cu2O modified Cu catalysts are the main oxide-derived Cu material type which have been extensively studied in eCO2RR. Besides oxide-derived Cu, Cu materials with abundant high-index facets or defects, also have showed good selectivity for C2+ products.28–35
Although valuable progresses have been made, numerous challenges are still existing. Firstly, some reported catalysts have showed excellent Faraday efficiency (FE) for multi-carbon (C2+) products (> 80%), but selectively converting CO2 at high current densities (> 1 A cm−2) remains as a challenge. Secondly, some oxide-derived Cu materials have showed good C2+ selectivity, but the role of O or Cu (ǀ) is still on debate. Thirdly, most of the reported efficient catalysts need complicated preparation process with limited synthetic quantity, and it is still difficult to scale up the production. Additionally, sub-nano structured Cu or Cu/Cu2O, especially those with high-index facets are easy to degenerate in air, which makes the materials hard to store and sale as commodities. In this aspect, in-situ forming active Cu species in the course of eCO2RR from stable precursors could be a good choice.
Herein, we developed a wet chemistry method for the preparation of copper oxide nanosheets (CuO-st). Field emission scanning electron microscopy (FE-SEM), transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) revealed that CuO-st possessed a rough and defective surface. CuO-st showed excellent C2+ selectivity (mainly C2H4, ethanol and propanol) at high current density, with FE (C2+) of 86 and 80 % at 1.0 and 1.8 A cm−2, respectively. Outstandingly, excellent catalytic stability was also achieved (@ ~580 mA cm−2, 50 h). The C−C coupling rate was found positively correlated with the concentration of KCl once the reaction initiated in flow cell. Mechanistic study indicated the existence of Cu (ǀ) in the CuO-st derived copper species during eCO2RR, and the amount of Cu (ǀ) varied with the reaction. Moreover, CuO-st showed good storage stability, without distinctive changes in powder X-ray diffraction (PXRD) and eCO2RR performance after more than 10 months exposed to air.
Synthesis and characterization of the catalysts. Copper oxide (CuO-st) was synthesized in gram scale with hydrazine hydrate as an initiator (see the Methods for details). By adjusting the reaction condition, cuprous oxide (Cu2O-pt) and copper (Cu-pt) were prepared for comparison. The composition of the materials were firstly confirmed by powder X-ray diffraction (PXRD) (Supplementary Fig. 1-3). The morphologies were investigated using field emission scanning electron microscopy (FE-SEM). Cu2O-pt was revealed to have octahedron structures (Supplementary Fig. 6), while Cu-pt was amorphous particles (Supplementary Fig. 7). CuO-st possessed a sheet-like structure under low magnification (Fig. 1a), but the high power SEM images revealed that the nanosheets had rough surface, forming by the accumulation of nano-particulates with an average diameter of 20 ± 3 nm (Fig. 1b). CuO-st was further studied by transmission electron microscopy (TEM). The low-resolution image showed a sheet-like structure (Fig. 1c and d). The HR-TEM image exhibited that the spacing of the lattice planes were ~0.27 and 0.23 nm, which corresponded well to (110) and (111) planes of copper oxide, respectively (Fig. 1e). Interestingly, lattice dislocation could be observed in the HR-TEM images (Fig. 1e and Supplementary Fig. 5), indicating the existence of rich defects. Such a nature was further confirmed by X-ray photoelectron spectroscopy (XPS) analysis. As shown in Fig. 1f, the O 1s region of CuO-st can be fitted into three peaks with binding energy (BE) of 592.6, 531.3 and 533.0 eV, respectively. The peak at 592.6 eV was attributed to lattice oxygen, the minor peak at 533.0 eV was assigned to the surface hydroxide or chemisorbed oxygen, and the peak between them (531.3 eV) was attributed to defective oxygen.36–38 The radio of defective oxygen/ lattice oxygen is close to 1/2.1 (Supplementary Table 1).
Electrochemical performance evaluation. The eCO2RR performance of the catalysts were firstly studied with an H-cell. We analyzed the gas and liquid products at the applied potentials between −0.85 and −1.40 V (vs RHE) in CO2-saturated 0.1 M KCl (pH 3.8). It could be observed that the C-C coupling efficiency improved steadily once the voltage decreased from -0.9 to -1.2 V (vs RHE) for all the materials. For CuO-st, the FEs of C2H4 and C2+ products reached up to 58 and 82 % at −1.21 V (vs RHE), respectively, while the formation of CH4 and H2 remained at very low levels at all the tested potential range. Cu-pt and Cu2O-pt showed maximum FE (C2+)s of 73 and 68 % at -1.24 and - 1.28 V (vs RHE), respectively, lower than the one of CuO-st. Additionally, the partial current of C2+ products (j(C2+) ) of CuO-st was the highest among the three materials achieved (Supplementary Fig. 8-10). The results indicate that CuO-st has better performance than Cu-pt and Cu2O-pt in selective conversion of CO2 to C2+ products.
To explore the possible use of CuO-st in industry, we further evaluated the eCO2RR selectivity in a flow cell. The FEs distribution of cathode products vs currents in 1 M KCl, 2M KCl and saturated KCl (sa. KCl) are shown in Fig. 3a-c. It was revealed that the FE (C2+) showed an increase in all the solutions when the constant current density was raised from 600 mA cm−2 to 1.0 A cm−2. The FE (C2+) seemed to have a positive correlation with the KCl concentration. In sa. KCl, the FE (C2+) at 1.0 and 1.8 A cm−2 were achieved 86 and 80 %, respectively. Such an excellent C2+ selectivity at high reaction rate almost represents the state-of-the-art performance of Cu-based catalysts. (Fig. 3f and Supplementary Table 2). Potentiostatic electrolysis was chose for the stability test. It was observed that during the operation of the reaction at -0.98 V (vs RHE, with iR correction) in 1 M KCl for ~50 h, the current density kept ~580 mA cm−2 and the FE (C2H4) showed no significant change (Supplementary Fig. 17). Moreover, storability of CuO-st was also evaluated. It was revealed that no distinctive change happened in the PXRD and eCO2RR performance of CuO-st after in air for ~10 months (Supplementary Fig. 18 and 19).