BiO 2-x Nanosheets with Surface Electron Localizations for Efficient Electrocatalytic CO 2 Reduction to Formate

Open AccessCCS ChemistryCOMMUNICATIONS1 Jul 2022BiO2-x Nanosheets with Surface Electron Localizations for Efficient Electrocatalytic CO2 Reduction to Formate Zhonghao Tan, Jianling Zhang, Yisen Yang, Yufei Sha, Ran Duan, Jiajun Zhong, Buxing Han, Jingyang Hu and Yingzhe Zhao Zhonghao Tan Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid, Interface and Chemical Thermodynamics, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100190 , Jianling Zhang *Corresponding author: E-mail Address: [email protected] Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid, Interface and Chemical Thermodynamics, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100190 , Yisen Yang Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid, Interface and Chemical Thermodynamics, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100190 , Yufei Sha Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid, Interface and Chemical Thermodynamics, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100190 , Ran Duan Key Laboratory of Photochemistry, Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 , Jiajun Zhong Beijing Synchrotron Radiation Facility (BSRF), Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100190 , Buxing Han Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid, Interface and Chemical Thermodynamics, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100190 , Jingyang Hu Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid, Interface and Chemical Thermodynamics, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100190 and Yingzhe Zhao Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid, Interface and Chemical Thermodynamics, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 School of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100190 https://doi.org/10.31635/ccschem.022.202202068 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail To enhance the activity and selectivity of electrocatalytic CO2 reduction to formate is of great importance from both environmental and economical viewpoints. Herein, the BiO2-x nanosheets with surface electron localizations were constructed and utilized for the efficient CO2-to-formate conversion. The formate Faraday efficiency reaches 99.1% with current density of 12 mA cm−2 at −1.1 V versus the reversible hydrogen electrode (RHE) in an H-type cell while those in the flow cell are 91.3% and 319 mA cm−2 at −1.0 V versus RHE, respectively. Theoretical calculations indicate that the electron localization presenting in the BiO2-x nanosheet favors OCHO* intermediate stabilization and suppresses H* intermediate adsorption, thus improving the CO2-to-formate efficiency. The BiO2-x electrocatalyst is nondopant, easily prepared, low-cost, highly active and selective for CO2RR to formate, which has demonstrated potential for application in the Zn-CO2 battery. The maximum power density can reach 2.33 mW cm−2, and the charge/discharge cycling stability is >100 h (300 cycles) at 4.5 mA cm−2. Download figure Download PowerPoint Introduction The high CO2 concentration in the atmosphere caused by excessive consumption of fossil fuels has led to a series of environmental problems and disrupted the natural carbon cycle.1–4 CO2 reduction reaction (CO2RR) to value-added chemicals using electricity is a promising way to reduce CO2 concentration in the atmosphere and achieve artificial carbon cycles.5–15 In particular, the electrochemical CO2RR to formate is very important because formate is a high-value liquid product, widely used in chemical production and fuel cells.16–18 Compared with the CO2RR to other carbon-containing products (e.g., methane, methanol, ethylene, ethanol, etc.), the CO2RR to formate has been regarded as the most economically viable route due to the need for only two electrons and its low equilibrium potential.19–25 Despite these advantages, the CO2RR to formate is restricted by the high energy barrier for CO2 conversion to OCHO* and the competition with the hydrogen evolution reaction (HER) in aqueous solution.26,27 Up to now, various metals such as Sn,28 In,29 Bi,30 Pb,31 and Co32 have been utilized for electrocatalytic CO2RR to formate. Among these electrocatalysts, the Bi-based materials have attracted much attention because Bi has advantages of being nontoxic, inexpensive, and abundantly available in the earth.33,34 Diverse kinds of Bi-based catalysts have been synthesized for the electrocatalytic CO2 conversion to formate, including Bi,35 Bi-Sn,36 Pd3Bi,37 Bi2Te3,38 Bi2WO6,39 and their composites with a secondary phase, like Bi/CeOx,40 [email protected] carbon nanotubes,41 S-Bi2O3/carbon nanotubes,42 Bi2O3@carbon,43 Bi2O3/carbon nanofiber,44 and so on. The formate Faraday efficiencies (FEs) can reach values >90.0%, but the current densities mostly remain low. It is still a challenge to develop Bi-based catalysts with both high formate selectivity and current density that can meet the industrial requirement of being >300 mA cm−2. For the first time in this work, we constructed BiO2-x nanosheets with surface electron localizations that exhibit high selectivity and activity for electrochemical CO2RR to formate. The formate FE can reach 99.1% at −1.1 V versus a reversible hydrogen electrode (RHE) in a H-type cell. The current density is up to 319 mA cm−2 with formate FE of 91.3% at −1.0 V versus RHE in flow cell. Density functional theory (DFT) calculations reveal that the electron localization of BiO2-x lowers the energy barrier for the production of OCHO* intermediates, making it easier to generate than H* intermediates. Compared with the widely adopted heteroatom doping method to induce electron localization of catalyst for boosting electrochemical CO2RR,45–47 the BiO2-x electrocatalyst is nondopant, easily prepared, and low-cost. To explore its potential application, a Zn-CO2 battery using the BiO2-x catalyst for cathode was assembled, which exhibits high power density of 2.33 mW cm−2 with a long-term stability >100 h (300 cycles) at 4.5 mA cm−2. Experimental Section Chemicals Ascorbic acid and sodium hydroxide (purity, 96.0%) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Potassium bicarbonate (purity, 99.5%) and potassium hydroxide (purity, 85.0%) were provided by Aladdin Reagent Co., Ltd. (Shanghai, China). Zn plate, sodium bismuthate dihydrate (purity, 80.0%), Nafion D-521 dispersion, Nafion N-117 membrane, and Toray carbon paper (TGP-H-060) were supplied by Alfa Aesar Reagent Co., Ltd. (Shanghai, China). The gas diffusion layer (YLS-30T), Fumasep FAA-3-PK-130 membrane, and Fumasep FBM-PK membrane were obtained from Suzhou Sinero Technology Co. Ltd. (Suzhou, China). D2O (purity, 99.9%) and dimethyl sulfoxide (DMSO; purity, 99.0%) were bought from Innochem Reagent Co., Ltd. (Beijing, China). High purity CO2 gas (99.999%), high purity Ar gas (99.999%), and deionized water were provided by Beijing Analytical Instrument Company (Beijing, China). All reagents were used directly without further treatment. Synthesis of p-BiO2-x p-BiO2-x was synthesized by a hydrothermal method. First, 3.0 g sodium bismuthate dihydrate and 2.4 g sodium hydroxide were dissolved in 60 mL deionized water and then stirred vigorously for 0.5 h. The above solution was transferred into a 100 mL Teflon-lined autoclave and heated to 180 °C for 5 h. After naturally cooling down to room temperature, the solid product was separated by centrifugation and washed three times by deionized water and dried in vacuum at 80 °C for 6 h. Synthesis of m-BiO2-x m-BiO2-x was synthesized by a solid-phase grinding method for p-BiO2-x. A mixture of p-BiO2-x and ascorbic acid with a mass ratio of 1:1 was vigorously ground in an agate mortar for 0.5 h. Then the solid was washed six times by deionized water to remove ascorbic acid and dried in vacuum at 80 °C for 6 h. Characterizations X-ray diffraction (XRD) patter was determined by a Rigaku D/max-2500 X-ray diffractometer (Rigaku Corporation, Tokyo, Japan) with Cu Kα radiation. X-ray photoelectron spectroscopy (XPS) was determined by a Thermo Fisher Scientific ESCALAB 250 Xi (Thermo Fisher Scientific, Waltham, MA, USA) using 200 W Al Kα radiation. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images were obtained from a HITACHI S-4800 scanning electron microscope (Hitachi, Tokyo, Japan) and a JEOL-1011 field-emission transmission electron microscope. High-resolution TEM (HRTEM) image was obtained from a JEOL-2100F field-emission transmission electron microscope (JEOL, Tokyo, Japan). High-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) and STEM electron energy loss spectroscopy (EELS) were characterized by Cryo-TEM (Thermo Scientific Themis 300; Thermo Fisher Scientific, Waltham, MA, USA). X-ray absorption fine structure (XAFS) data were collected at the 1W2B station at the Beijing Synchrotron Radiation Facility (BSRF, Beijing, China). Raman spectra were determined by a Horiba LabRAM HR Evolution Raman microscope (HORIBA Scientific, Paris, France; 512 nm). Electrocatalytic CO2 reduction All electrochemical tests were performed using a CHI-660E electrochemical workstation equipped with a high current amplifier CHI 680c in a H-type cell and flow cell, respectively. All electrode potentials were transformed into standard RHE potentials by using the following formula: E ( RHE ) = E ( Ag / AgCl ) + 0.197 + 0.0591 × pH The iR compensation was conducted in flow cell experiment at each potential. In a H-type cell, 0.1 M KHCO3 aqueous solution and Nafion-117 membrane were used as electrolyte and proton exchange membrane, respectively. Before each electrochemical experiment, 0.1 M KHCO3 aqueous electrolyte was saturated with CO2 for at least 0.5 h. The hydrophobic carbon paper (TGP-H-060, 1 × 1 cm−2) coated with catalyst was used as the working electrode. Ag/AgCl electrode and Pt net (1 × 1 cm−2) were used as reference electrode and counter electrode, respectively. For the preparation of the working electrode, 1 mg catalyst powder and 10 μL Nafion were distributed into 200 μL ethanol to form a homogeneous catalyst ink, which was then spread onto carbon paper. The flow cell was obtained from Gaoss Union (Tianjin) Photoelectric Technology Company (Tianjin, China). It was constructed with an anode chamber and a cathode chamber. An anion exchange membrane (Fumasep FAA-3-PK-130) was used to separate the anode and cathode chambers. In the flow cell, 1 M KOH aqueous solution and the Fumasep FAA-3-PK-130 membrane were used as electrolyte and anion exchange membrane, respectively. The gas diffusion layer (YLS-30T, 0.5 × 2 cm−2) coated with catalyst was used as a gas diffusion electrode. Ag/AgCl electrode and Pt net (0.5 × 2 cm−2) were used as reference electrode and counter electrode, respectively. To prepare the gas diffusion electrode, 5 mg catalyst powder and 50 μL Nafion were dispersed into 750 μL ethanol to form a homogeneous catalyst ink. Then 160 μL catalyst ink was dropped on the gas diffusion layer with the eventually loading mass of ∼1 mg cm−2. The flow rate of cathode peristaltic pump and anode peristaltic pump were adjusted to be 30 mL min−1, and the flow rate of CO2 was controlled at 20 sccm by gas flow meter. Product analysis The gas product of electrochemical experiment was collected by a gas bag (each gas bag was collected for 2000 s). The gas product was detected by gas chromatography (Agilent 8890; Agilent Technologies Inc., CA, USA) with a thermal conductivity detector (TCD) detector using high purity argon as carrier gas. The liquid product was determined by 1H NMR (Bruker AVANCE III 400 HD; Bruker, Germany). After each electrochemical test, 200 μL reaction electrolyte was mixed with 200 μL D2O and 100 μL 6 mM DMSO solution, and then detected by 1H NMR. The Faraday efficiency was calculated by the following formulation: FE = Moles of product Q / n F × 100 % where Q: charge (C); F: 96485 C/mol; n: number of electrons required to generate the product. Zn-CO2 battery The Zn-CO2 battery test was performed in the flow cell. The gas diffusion layer (YLS-30T, 0.5 × 2 cm−2) coated with m-BiO2-x and the polished Zn plate were used as battery electrodes. The 1 M KHCO3 aqueous solution, and 2 M KOH/0.02 M Zn(CH3COO)2 aqueous solution served as electrolytes for cathode and anode, respectively, which were separated by a bipolar membrane (Fumasep FBM-PK). CO2 flow rate was controlled at 20 sccm during the test. The galvanostatic discharge curves were measured by galvanostatic discharge at 1.5, 3.0, and 4.5 mA cm−2. For charge/discharge cycles, the current density was set to 4.5 mA cm−2. In situ Raman spectroscopy The in situ Raman experiment was conducted by using a Horiba LabRAM HR Evolution Raman microscope (HORIBA Scientific, Paris, France). The laser wavelength was controlled at 785 nm. 0.1 M KHCO3 aqueous solution was used as electrolyte. Carbon paper coated with catalyst, a Ag/AgCl electrode, and a carbon rod were used as working electrode, reference electrode, and counter electrode, respectively. The in situ Raman electrolytic cell was purchased from Gaoss Union (Tianjin) Photoelectric Technology Company (Tianjin, China). Computational method All the computations were performed by using the Vienna ab initio simulation package.48,49 The ion-electron interactions were described by the projector augmented wave method,50 and the general gradient approximation in the Perdew–Burke–Ernzerhof form was used.51,52 A cutoff energy of 450 eV for the plane-wave basis set was adopted. During structural relaxation, the convergence criterion was set to be 0.03 eV/Å and 10−5 eV for the residual force and energy, respectively. A 3 × 3 × 1 BiO2-x (111) slab was used as the model, and the Brillouin zone was sampled by a Monkhorst–Pack 3 × 3 × 1 k-point grid. To avoid the interaction between two periodic units, a vacuum space of 15 Å was employed. The free energy change (ΔG) of each elementary reaction was calculated as Δ G = Δ E + Δ E ZPE − T Δ S where ΔE, EZPE, T, and S are reaction energy difference, zero-point energy, temperature, and entropy, respectively. Results and Discussion The BiO2-x electrocatalyst was prepared by a two-step method. First, the normal BiO2-x was synthesized by a hydrothermal route for sodium bismuthate dihydrate and sodium hydroxide at 180 °C for 5 h (see its characterizations in Supporting Information Figures S1–S3). Then a surface modification step was applied to the above BiO2-x by grinding it with ascorbic acid for 0.5 h. The pristine BiO2-x synthesized in the first step and the modified BiO2-x were named as p-BiO2-x and m-BiO2-x, respectively. The XRD pattern of m-BiO2-x coincides completely with that of BiO2-x (JCPDS PDF#47-1057) and p-BiO2-x (Figure 1a). This indicates that the crystalline structure of m-BiO2-x remains unchanged after modification from p-BiO2-x. Notably, the crystallinity of m-BiO2-x decreases compared with that of p-BiO2-x, which might be attributed to the generation of oxygen defects as discussed in following.53 SEM and TEM images reveal that m-BiO2-x keeps the nanosheet morphology of p-BiO2-x (Figure 1b,c). The HRTEM image shows lattice spacings of 0.317 and 0.119 nm (Figure 1d), corresponding to the (111) and (220) crystal planes of BiO2-x, respectively. Bi M4,5-edge EELS characterization displays that the M4,5-edge gradually shifts to higher energy areas by 12 eV from spot 1 to spot 4 (Figure 1e,f). This suggests the gradual increase of oxygen content from surface to bulk, which is indicative of a large number of surface oxygen defects in m-BiO2-x.54 In contrast, the M4,5-edge of p-BiO2-x has a shift of 1 eV from spot 1 to spot 4 ( Supporting Information Figure S4). The Brunauer–Emmett–Teller surface areas of m-BiO2-x and p-BiO2-x are 7.73 and 6.37 m2 g−1, respectively, as determined by the N2 adsorption–desorption method ( Supporting Information Figure S5). Figure 1 | Morphology characterizations of m-BiO2-x. (a) XRD patterns of p-BiO2-x and m-BiO2-x. (b) SEM image, (c) TEM image, (d) HRTEM image, (e) STEM-EELS sampling region, and (f) EELS spectra of m-BiO2-x. Scale bars: 200 nm in (b), 400 nm in (c), 5 nm in (d) and 100 nm in (e). Download figure Download PowerPoint The Raman spectrum of m-BiO2-x exhibits a lowered intensity and a red shift compared with that of p-BiO2-x (Figure 2a). This indicates a smaller Bi-O bond force constant in m-BiO2-x according to Hooke's law, which is consistent with the large number of oxygen defects in m-BiO2-x.55 By electron paramagnetic resonance (EPR), the signal of m-BiO2-x is much stronger than that of p-BiO2-x (Figure 2b), indicating a higher defect degree. The g-factor (g) values for m-BiO2-x and p-BiO2-x were determined to be 2.007 and 2.004, which are typical oxygen defect signals.56 XPS was utilized to further characterize p-BiO2-x and m-BiO2-x ( Supporting Information Figure S6). For high-resolution O1s XPS spectra (Figure 2c), the three peaks represent lattice oxygen (O1, 529.2–529.5 eV), defect oxygen (O2, 531.1–531.7 eV), and absorbed oxygen (O3, 533.4–533.6 eV), respectively. The content of defect oxygen relative to the total oxygen in m-BiO2-x was calculated to be 62.2%, which is much higher than that of p-BiO2-x (25.5%, Supporting Information Table S1). For high-resolution Bi 4f XPS spectra, the Bi 4f7/2 and Bi 4f5/2 peaks present at 158.15 and 163.40 eV for m-BiO2-x and 158.30 and 163.55 eV for p-BiO2-x, respectively (Figure 2d). The results suggest a higher electron density around Bi atoms of m-BiO2-x than p-BiO2-x. Figure 2 | Structural characterizations of m-BiO2-x. (a) Raman spectra, (b) EPR spectra, (c) O1s XPS spectra, (d) Bi 4f XPS spectra, (e) Bi L3-edge XANES spectra, and (f) Bi L3-edge EXAFS spectra of p-BiO2-x and m-BiO2-x. Download figure Download PowerPoint The fine structure of m-BiO2-x was investigated by Bi L3 edge X-ray absorption near edge structure (XANES). Commercial Bi, Bi2O3, and p-BiO2-x were used as contrast samples. The XANES spectra of m-BiO2-x and p-BiO2-x are roughly similar but different from those of commercial Bi and Bi2O3 powder (Figure 2e). Compared with p-BiO2-x, the white line peak of m-BiO2-x is weaker and shifts to a lower energy region, indicating the weaker oxidation state in m-BiO2-x. From extended X-ray absorption fine structure (EXAFS), the Bi-O bond lengths of p-BiO2-x and m-BiO2-x are 1.56 Å while that for commercial Bi2O3 is 1.65 Å (Figure 2f). According to EXAFS fitting curves ( Supporting Information Figure S7 and Table S2), the Bi coordination numbers of p-BiO2-x and m-BiO2-x were determined to be 4.4 and 3.1, respectively. The smaller coordination number of m-BiO2-x confirms the presence of a large number of oxygen defects. Further, the wavelet-transform EXAFS (WT-EXAFS) gives support that m-BiO2-x has an unsaturated coordination environment ( Supporting Information Figure S8). To investigate the catalytic activity of m-BiO2-x for CO2RR, linear sweep voltammetry (LSV) measurements were carried out. A H-type cell was used with 0.1 M KHCO3 aqueous solution as electrolyte. Obviously, the current densities in CO2-saturated electrolyte are higher than those in Ar-saturated electrolyte (Figure 3a). This indicates that m-BiO2-x is more active for CO2RR than HER. Compared with p-BiO2-x, m-BiO2-x exhibits larger current density at the same reduction potential, implying higher CO2RR activity of m-BiO2-x. Figure 3 | Electrochemical CO2 reduction measurements in H-type cell. (a) LSV curves of p-BiO2-x and m-BiO2-x in Ar- or CO2-saturated 0.1 M KHCO3 aqueous electrolyte. (b) Formate FE values at different applied potentials in CO2-saturated 0.1 M KHCO3 aqueous electrolyte. (c) Formate partial current densities at different applied potentials. (d) Charging current density differences (Δj) under different scan rates. (e) Nyquist plots for p-BiO2-x and m-BiO2-x. (f) Stability test of m-BiO2-x at −1.1 V versus RHE. Download figure Download PowerPoint The performance of m-BiO2-x for electrocatalytic CO2RR was tested. The gas and liquid products were determined by gas chromatography and 1H nuclear magnetic resonance, respectively. Formate was detected as the major product along with the production of small amounts of CO and H2. The formate FE values maintain >93.0% in an electrochemical window range from −1.0 to −1.4 V versus RHE (Figure 3b). At −1.1 V versus RHE, the maximum formate FE can reaches 99.1%, indicating the efficient suppression of HER and CO production ( Supporting Information Figure S9). Compared with p-BiO2-x, the selectivity to formate is improved by m-BiO2-x as well as the formate partial current density (jformate, Figure 3c). To gain a deeper insight into the superior catalytic activity of m-BiO2-x, cyclic voltammetry experiments under various sweep rates were conducted ( Supporting Information Figure S10). The double-layer capacitances (CdI) of p-BiO2-x and m-BiO2-x were determined to be 0.94 and 2.83 mF cm−2, respectively (Figure 3d). This indicates that m-BiO2-x has a larger electrochemical surface area and more exposed active sites. To compare the intrinsic HER activities of p-BiO2-x and m-BiO2-x, the LSV curves under Ar situation were normalized ( Supporting Information Figure S11). It is clear that m-BiO2-x has lower HER performance. The electrochemical impedance spectra of p-BiO2-x and m-BiO2-x were measured for researching electron transfer characteristics. From the Nyquist plots, m-BiO2-x has much smaller electron transfer resistance than p-BiO2-x (Figure 3e). Both the greater exposure of active sites and smaller electron transfer resistance in m-BiO2-x are favourable for CO2RR. The stability of m-BiO2-x was investigated at −1.1 V versus RHE for 15 h (Figure 3f). The current density can maintain at ∼12 mA cm−2 during the whole test while formate FE slightly drops to 94.9% over 10 h and remains stable up to 15 h. Furthermore, the flow cell was used as an alternative to the above H-type cell for electrocatalytic CO2RR to improve the current density to meet the requirement of commercial application ( Supporting Information Figure S12).57–59 1 M KOH aqueous solution was used as electrolyte in accordance with the literature.60,61 From the LSV curves (Figure 4a), we can see that the current densities are remarkably higher than those in the H-type cell (Figure 2a). For example, a larger current density of 400 mA cm−2 was achieved at −1.2 V catalyzed by m-BiO2-x, which is higher than that by p-BiO2-x (304 mA cm−2). The electrocatalytic CO2RR was carried out in a flow cell at potentials from −0.6 to −1.0 V versus RHE. As catalyzed by m-BiO2-x, the maximum formate FE reaches 95.1% at −0.8 V versus RHE and the current density is 220 mA cm−2 (Figure 4b). The current density can be improved to 319 mA cm−2 at −1.0 V versus RHE, with formate FE of 91.3%. In contrast, the formate FE and current density over p-BiO2-x at −1.0 V versus RHE are 78.9% and 230 mA cm−2, respectively. In addition, the jformate values catalyzed by m-BiO2-x are significantly higher than those catalyzed by p-BiO2-x (Figure 4c). The stability of CO2RR over m-BiO2-x in the flow cell was also investigated at −0.8 V versus RHE. Through 10 h testing, the formate selectivity and current density maintain high values of >94.0% and >200 mA cm−2, respectively (Figure 4d). Compared with the reported electrocatalysts for CO2RR to formate, m-BiO2-x exhibits both superior selectivity and activity ( Supporting Information Table S3). Figure 4 | Electrochemical CO2 reduction measurements in flow cell. (a) LSV curves of p-BiO2-x and m-BiO2-x in 1 M KOH aqueous electrolyte in flow cell. (b) Formate FE values at different applied potentials in 1 M KOH aqueous electrolyte. (c) Formate partial current densities at different applied potentials. (d) Stability test of m-BiO2-x at −0.8 V versus RHE. Download figure Download PowerPoint The reaction pathway of CO2 electroreduction to formate on m-BiO2-x was investigated by an in situ Raman experiment. It was carried out in an in situ Raman electrolytic cell, using 0.1 M KHCO3 aqueous solution as electrolyte. As shown in Figure 5a, there are two peaks located at 1351 and 1544 cm−1, which correspond to the OCHO* intermediate and the CO2*− intermediate, respectively.62,63 With the gradual increase of potential from −0.3 to −1.3 V versus RHE, the intensity of the peak at 1544 cm−1 increases. It indicates that the CO2 molecules adsorbed by the active site (CO2*) get electrons to form CO2*− intermediate. Simultaneously, the peak at 1351 cm−1 becomes stronger with increased potential, implying the rapid protonation of CO2*− intermediate to form OCHO*. The in situ Raman spectra of p-BiO2-x were also determined, which show a similar trend with that of m-BiO2-x ( Supporting Information Figure S13). This indicates the same reaction pathway over the two electrocatalysts. Figure 5 | Mechanistic studies of the electrochemical CO2-to-formate conversion on p-BiO2-x and m-BiO2-x. (a) In situ Raman spectra of m-BiO2-x at different applied potentials. (b) Free energy diagrams for CO2RR and HER on p-BiO2-x and m-BiO2-x. (c) PDOS of Bi atom p orbital of p-BiO2-x and m-BiO2-x. (d) Charge density difference for the adsorption of OCHO* intermediate on p-BiO2-x and m-BiO2-x. Isosurface value is set to be 0.005 e/Å3, and the charge accumulation and depletion are shown in yellow and cyan, respectively. (e and f) ELF plots of p-BiO2-x and m-BiO2-x. Download figure Download PowerPoint DFT calculations were applied to investigate the catalytic mechanism of m-BiO2-x for CO2RR to formate. The free energy diagrams of CO2 reduction to formate and the H2 formation on m-BiO2-x and p-BiO2-x were calculated. The relevant models are shown in Supporting Information Figures S14 and S15. For CO2 conversion into formate, the first elementary reaction, that is, reduction of CO2 into OCHO*, is the potential determining step. As shown in Figure 5b, the reaction free energy over m-BiO2-x was calculated to be 1.12 eV, which is much lower than that on p-BiO2-x (1.48 eV). The reduced free energy over m-BiO2-x suggests that the key intermediate OCHO* formation can be facilitated, thus improving the selectivity for formate generation. For HER, the calculated free energy for hydrogen adsorption (ΔGH) on m-BiO2-x is higher than that on p-BiO2-x (Figure 5b). This implies that the hydrogen production can be suppressed on m-BiO2-x, which is in agreement with the experimental results. Partial density of states (PDOS) analyses were performed to study the electron structure of m-BiO2-x. The p-band center value of m-BiO2-x was calculated to be −0.07 eV while that for p-BiO2-x is −0.82 eV (Figure 5c). Evidently, the p-band center of m-BiO2-x is much closer to the Fermi energy level than that of p-BiO2-x. This implies the localization of Bi p-orbitals electrons in m-BiO2-x,64 which is favorable for the formation of more bonding orbitals between the active center and the OCHO* intermediate and thus stabilizes the OCHO* intermediate for producing formate.65 The charge density differences for the adsorption of OCHO* intermediate on the two catalysts were calculated (Figure 5d). Compared with p-BiO2-x, more electron transfer between m-BiO2-x and OCHO* can be observed. This demonstrates stronger electron interaction between m-BiO2-x and OCHO*. From the electron localization function (ELF) plot, it is obvious that the electrons around the