1 Introduction
Electrochemical reduction of CO2 (CO2RR) represents an appealing alternative for recycling greenhouse gas CO2 and producing renewable clean energy sources such as CH4, C2H4, which may alleviate the environmental threats caused by the utilization of fossil fuels [1,2,3]. Cu and the Cu-based compounds (e.g., Cu2O) have been intensively studied as efficient catalysts for CO2RR. The catalyst’s surface reduction and reconstruction process lead to the coordination of d orbital of the metal atom with CO2, thus more likely to produce hydrocarbons. This property attracted the focus of investigations, including the synthesis of Cu nanostructures and the electrocatalysis of copper oxide to produce acid and alcohol [4,5,6,7]. Hazarika, et al. found that Cu2O had greater stability in producing methyl alcohol while suppressing the production of hydrogen. Gao, et al. synthesized Cu2O with different shapes and surfaces and investigated their difference in the selectivity and activity in producing ethylene. Fu, et al. then found the high-index-surface of Cu2O promoted the generation of C2+. These previous studies proposed the idea of forming different products via surface decoration of Cu2O [8,9,10].
Formic acid (HCOOH) is an essential CO2RR product, which has been commercialized due to its feasibility and large applicability spectra [11]. It is proved to be a satisfying fuel applied in fuel cells [12] and can serve as an energy-storage medium with high stability [13]. Different catalysts were utilized to generate HCOOH, including Sn, Pb, Cu2O and some metal-complex catalysts [11]. It has been reported that during the CO2RR process catalyzed by Cu2O, many oxygen atoms are precipitated in the catalyst in the reduction process [5]. Meanwhile, some of the remaining atoms can generate a portion of Cu(Cu+) with positive valence to remain on the surface of the catalyst. During the surface reduction and reconstruction of the catalyst, Cu+ and Cu in adjacent positions are favorable for coupling C-C and C-H to generate multi-carbon products with added value. The Cu-based compounds with metastable structure, crystal interface, high density and high specific surface area can expose more catalytic active sites, have higher catalytic activity, stabilize the intermediates in the reduction process, and reduce the overpotential in the CO2RR process, thus reducing energy consumption [10,14,15,16,17].
Considering these aspects, in this work, a new morphology control path for Cu2O nano-catalysts (NCs) was investigated to achieve high HCOOH Faraday efficiency (FE). We combined the facet-controlled synthesis and etching processes to construct a series of NCs with cuboctahedron or center-hollow cuboctahedron dice-like structures. The as-obtained catalysts were proved to have a high HCOOH generation performance in the CO2RR process. Characterizations of the NCs confirmed that the optimized properties were mainly attributed to the high density of oxygen defects induced during the etching process.
2 Results and discussion
2.1 Structural characterizations Cu2O NCs
Cu2O NCs were synthesized by a solution-based method, as illustrated in Fig. 1a and Fig. S1. The structures of the synthesized Cu2O NCs were etched for 1 day (1DPT), 5 days (5DPT) and 10 days (10DPT), respectively. During the synthesis process, polyvinylpyrrolidone (PVP) served as the capping agent, which could be preferably adsorbed on the {111} facets of the Cu2O NCs [17], promoting the formation of {111} facets, which are advantageous for the generation of formic acid [18]. With controlled PVP loading, the surface ratio of {100} / {111} facet was tuned to be approximately 1.73, leading to a cuboctahedron nanostructure. As shown in Fig. 1b, the 1DPT NCs exhibited the typical cuboctahedron structure without apparent etching features. After 5 days of etching, it was observed that the {100} facets were first etched from the centers, forming a hollow structure. In the 10DPT NCs, the etching on {100} facets was expanded, leaving larger round holes on the facets, while the {111} facets remained intact. When the etching duration was increased to 12 days (Fig. S2), the Cu2O NCs almost broke down, with only a part of the remained framework. Therefore, the 1DPT, 5DPT and 10DPT NCs were mainly discussed in this work, which exhibited representative properties of this series of nanoarchitectures.
Fig. 1 a Schematic illustration of the synthesis process. Field-emission scanning electron microscopy (FESEM) images of (b) 1DPT; (c) 5DPT; (d) 10DPT NCs. Inset shows the corresponding crystal structures |
Transmission electron microscopy (TEM) images in Fig. S3 confirmed the hollow structures of the etched products with a homogeneous distribution of Cu and O. Furthermore, the occupying rates of Cu and O atoms on the {111} and {100} facets were investigated by energy dispersive spectroscopy (EDS) (Figs. S4, S5), as presented in Fig. 1c, d. The atomic percentage of Cu on {100} facets increased remarkably after the 5-day etching process, indicating the formation and evolution of oxygen vacancies on the surfaces.
The crystal structures of the NCs were further investigated and the results are shown in Fig. 2. The high-resolution TEM (HRTEM) images in Fig. 2c-e clearly show lattice fringes with uniform distances of 0.31, 0.24 and 0.22 nm, corresponding to the interplanar spacing of the {110}, {111} and {200} facets in Cu2O. The X-ray diffractometry (XRD) pattern in Fig. 2f and Fig. S6 also confirm the existence of various facets, including {110}, {111}, {200}, {220} and {311}. The {111} facets dominated in all NCs because they were less engraved in the etching process.
Fig. 2 a, b TEM images of 5DPT and 10DPT NCs. c-e HRTEM images of 5DPT NCs: (c) {110}; (d) {111}; (e) {200} facets. f XRD patterns of 1DPT, 5DPT and 10DPT NCs |
2.2 CO2RR
The CO2RR performances of the 1DPT, 5DPT and 10DPT NCs were investigated, according to Fig. 3 and Fig. S7. The current densities (J) of the catalysts at different applied potentials ranging from − 0.8 to -2.2 V are presented in Fig. 3a, indicating that the current density increased with prolonged etching. The Faradic efficiencies of H2 of the catalysts (Fig. 3b) indicate that 5DPT and 10DPT NCs, which have a hollow structure distinct to the 1DPT NCs, can reduce the H2 generation at working potentials of greater than − 1.2 V. However, at the low potential range, they showed similar performance as the 1DPT NCs. All the catalysts reached a relatively high selectivity of > 60% towards formic acid (Fig. 3c). To test the stability of 1DPT, 5DPT and 10DPT NCs, we allowed the reduction reaction to last for 9 h. As shown in Fig. S8a-c, the Cu2O catalysts after CO2 electroreduction maintain the original structure morphology. There are some fine particles attached to the surface of the structure, which might come from the attachment of the electrolyte and the precipitation of copper. The binding energy is a bit weakened and the peaks shift slightly to the lower energy (Fig. S8d). As shown in Fig. S8e, all the three tested samples demonstrate good performance stability during the tests.
Fig. 3 a Current densities of 1DPT, 5DPT and 10DPT NCs, which characterize the activities of NCs; (b) H2 FE; (c) HCOOH FE of 1DPT, 5DPT and 10DPT NCs at different applied potentials (-1 to 1.5 V), which characterize the selectivities of NCs for HCOOH. Proportions of reduction products of (d) 1DPT; (e) 5DPT; (f) 10DPT NCs at different applied potentials |
Figure 3d-f further detail the particularities of the FE of all main products generated by the three catalysts. It could be observed that the 5DPT NCs showed higher HCOOH FE at lower potentials, with an optimal FE of 75.1% at -1.0 V. When a higher potential than 1.1 V was applied, the HCOOH FE of the 1DPT NCs exceeded the other two catalysts, reaching as high as 75.0% at -1.5 V. Moreover, the 5DPT and 10DPT NCs produced more by-products at higher potentials exceeding − 1.1 V, among which CO, CH4 and C2H4 dominated. Hence, our Cu2O NCs could be considered as promising yet highly efficient catalysts for HCOOH generation. A low potential is more advantageous for the more-etched 5DPT NCs, while higher potentials are better for the less-etched 1DPT NCs.
2.3 High formic acid FE
The enhanced formic acid production can be attributed to the contribution of oxygen defects on the Cu2O surfaces. As shown in Fig. 4a, the Cu 2p spectra of the 1DPT NCs show two peaks at 952 eV (Cu 2p1/2) and 932 eV (Cu 2p3/2) in all catalysts corresponding to Cu+ or Cu states, without distinct satellite Cu2+ peaks between them [19,20]. Further examinations on Cu LMM spectra (Fig. 4b) present a single peak at around 917 eV, which confirms that the Cu atoms in the catalysts are almost all in the Cu(I) state. Therefore, the enhanced catalyst performances might not be originated from the synergistic effect of Cu and Cu+ particles. The O 1s spectra in Fig. 4c show three component peaks [21]. Among the three peaks, the peak near 533 eV could be assigned to the surface oxygen adsorption. As it was already proved that no Cu2+ existed, the other two peaks in the spectra were attributed to O in the Cu2O (530 eV) and Cu2O1 − x (531 eV), indicating the existence of oxygen defects. The peaks near 530 and 531 eV of the 5DPT and 10DPT NCs shifted to lower binding energies compared to 1DPT NCs, stemming from the increased density of oxygen vacancies. The higher density of oxygen vacancies was also matched in Fig. 4d. With more oxygen vacancies in the 5DPT NCs, the peak shift of the 5DPT is larger than the 10DPT. The photoluminescence spectra of the 5DPT in Fig. 4e also show the highest intensity and the largest blue-shift among the three catalysts, as another solid evidence of the presence of numerous oxygen defects. 22.
Fig. 4 XPS spectra of (a) Cu 2p; (b) Cu LMM; and (c) O 1s. d PL spectra of 1DPT, 5DPT and 10DPT, with Lorentzian peaks divided in (e) |
The 5DPT NCs were further analyzed by high-angle annular dark-field (HAADF) - scanning transmission electron microscopy (STEM) (Fig. 5a, b) and electron-energy-loss spectroscopy (EELS) (Fig. 5c). We selected two regions at the edge and the interior of the Cu2O NC, as shown in Fig. 5b. The two sharp peaks corresponded to the fine structures of Cu L2 and L3 energy edges originated from the Cu in oxidized states, excluding zero-valence-state Cu [23,24]. The EELS characterization also accords closely with the XPS results detailed above. Cu L2 and L3 peaks acquired in the interior of Cu2O NC shift to higher energies compared with the peaks at the edge, demonstrating that Cu cations in the interior have higher valence. Therefore, it can be deduced that fewer oxygen vacancies exist in the interior of Cu2O NC. It can be further inferred that oxygen vacancies exist in all three catalysts, and the sequence of vacancy densities is: 5DPT > 1DPT > 10DPT.
Fig. 5 a HAADF-STEM micrograph of 5DPT Cu2O, with the regions marked in (b) where EELS spectra were recorded. c EELS spectra of Cu L2, L3 edges |
The vacancies on the surfaces served as favorable sites for CO2 adsorption and reduction. When applied with lower potentials, the 5DPT catalysts with high oxygen vacancies were more advantageous for *COOH production and desorption from the surface. When the potential was increased, the CO2RR process was further enhanced, and hence more by-products, such as CH4, CO, and higher-order hydrocarbon products like C2H6 and C2H4, were generated. In this high-potential range, the 1DPT NCs showed the highest HCOOH FE compared to other investigated catalysts.
3 Conclusion
A novel Cu2O nanostructure was designed by a facet-etching process, exhibiting high activity and selectivity for formic acid production. The cuboctahedron-like Cu2O NCs were etched from the centers of {100} facets, forming hollow dice-like structures. The 5DPT catalysts exhibited the highest HCOOH FE of 75.1% at a low potential of -1.0 V. When increasing the applied potential, the 1DPT catalysts showed higher potential than other catalysts, reaching an HCOOH FE of 75.0% at -1.5 V, which is much higher than previously reported result (< 50%) [10]. The enhanced CO2RR performance originated from the high-density oxygen vacancies, which served as active sites for CO2 adsorption and reaction. Our design offers a new and efficient approach to improve the activity and selectivity of Cu2O catalysts.
4 Experimental
The synthesis process of Cu2O NCs is detailed in Fig. 1a and Fig. S1. 25 First, 10 mL CuSO4 (6.8 M) was dissolved in 170 mL water, mixed with 9 g of PVP (Mw = 30,000 g mol− 1) in a round-bottomed glass flask, forming a light blue solution. 10 mL mixed solution of sodium citrate (7.4 M) and anhydrous sodium carbonate (12 M) was then added to the flask. After 10 min, 10 mL of glucose (14 M) was slowly dropped to the solution to reduce Cu2+ to Cu+, turning the solution to a darker blue. The final solution was placed in a water bath at 80 ℃ and stirred for 2 h. After cooling to room temperature, the precipitates were left in the solution and exposed to air for up to 1, 5 or 10 days (denoted as 1DPT, 5DPT, and 10DPT, respectively) at ambient temperature. The change of component concentration during the whole reaction process was measured with absorption spectroscopy and shown in Fig. S9. Then, the precipitates were collected by filtration, repeatedly washed with distilled water and absolute alcohol, and finally dried in a vacuum at 60 ℃ for 8 h.
XRD was conducted on a Rigaku SmartLab 9 kW X-ray diffractometer with CuKa radiation (λ = 1.5418 Å). FESEM images were recorded on a JEOL GeminiSEM 500 microscope operated at 15 kV. Micrographs dealing with TEM and EDS data were obtained on a JEOL JEM-2100 F microscope and JEOL ULTIM MAX microscope, respectively, operated at 200 kV. The atomic-scale morphology was characterized on an FEI Titan Cubed Themis G2 300 operated at 300 kV and equipped with a probe aberration corrector to provide a resolution of about 0.6 Å in STEM mode. The HAADF detector was 48 ~ 200 mrad. The EELS entrance aperture was 1 mm, and the energy dispersion was 0.25 eV. EELS data processing included alignment and calibration of the zero-loss position, pre-edge background subtraction, removal of multiple scattering. All EELS data were processed using the Digital Microscopy software package. 600 MHz nuclear magnetic resonance (NMR) spectroscopy (Varian INOVA) was used to identify and quantify the yield of products in the electrolytes.
Electrochemical measurements were performed to study the performance of the Cu2O NCs. Electrochemical studies of catalysts were conducted at an electrochemical workstation (CHI 660D, Shanghai CH Instruments Co., China) in an H-type cell with two compartments separated by a proton exchange membrane (Nafion 117, DuPont). The electrolyte in all experiments was 30 mL of 0.1 M NaHCO3 aqueous solution. A potassium chloride saturated Ag/AgCl, and a platinum sheet (1 × 1 cm 2) was used as the reference and counter electrodes. The working electrode was prepared as follows: ~30 mg of as-prepared catalyst was mixed with 5 mL of ultrapure water and 5 mL isopropyl alcohol under oscillator several times, followed by adding 0.1 mL 5 wt% Nafion solution (DuPont) to obtain a suspension. The electrolyte at the cathode side was bubbled with high purity CO2 (99.999%) for 2 h to exclude air and achieve CO2 saturation before testing. During the measurements, CO2 gas flow was controlled at 20 mL/min. The gas-phase products generated during CO2 electrolysis at each fixed potential were analyzed by a SHIMADZU GC-2014 equipment for gas chromatography (GC), connected online with an H-type electrochemical cell. The samples were injected into the GC using a ten-port valve system with high purity Ar (99.999%) as the carrier gas. The GC system was equipped with two columns associated separately with the two detectors. A thermal conductivity detector (TCD) was installed to detect H2 and CO, while a flame ionization detector (FID) was fabricated to detect the resulting hydrocarbons. The FEs of all gas products were calculated according to the methods reported by Varela et al. [26]. All the working potentials were estimated using the reversible hydrogen electrode (RHE) using the following equation: ERHE = EAg/AgCl + 0.197 V + 0.059 pH. The pH value of the CO2 saturated electrolyte was 6.8.
Abbreviations
CO2RR:CO2 Reduction Reaction;NCs :Nano-catalysts;FE :Faraday Efficiency;PVP :Polyvinylpyrrolidone;XRD :X-ray diffractometry;FESEM :Field-emission scanning electron microscopy;TEM :Transmission electron microscopy;HRTEM :High-resolution transmission electron microscopy;EDS :Energy dispersive spectroscopy;STEM :Scanning transmission electron microscopy;HADDF :High-angle annular dark-field;EELS :Electron-energy-loss spectroscopy;NMR :Nuclear magnetic resonance;GC :Gas chromatography;TCD :Thermal conductivity detector;FID :Flame ionization detector;RHE :Reversible hydrogen electrode
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s43979-022-00037-1.
Additional file 1.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (52172046) and the Basic Science Center Project of NSFC (51788104).
Authors’ contributions
JL and HWZ conceived the project; JL, CM, JKG and HLW performed the experiment; JL, CM, HLW, RYD and HZS analyzed the data; JL and HWZ wrote and revised the manuscript. All authors read and approved the final manuscript.
Funding
Open access funding provided by Shanghai Jiao Tong University. This work was supported by the National Natural Science Foundation of China (52172046) and the Basic Science Center Project of NSFC (51788104).
Availability of data and materials
Data and materials are available upon reasonable request.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
