Selective methane electrosynthesis enabled by a hydrophobic carbon coated copper core–shell architecture

Abstract

The electrosynthesis of valuable chemicals via carbon dioxide reduction reaction (CO₂RR) has provided a promising way to address global energy and sustainability problems. However, the selectivity and activity of deep-reduction products (DRPs) still remain as big challenges. Here, a copper–carbon-based catalyst with a hydrophobic core–shell architecture has been constructed and was found to exhibit excellent DRPs of methane generation with a faradaic efficiency of 81 ± 3% in a neutral medium and a maximum partial current density of −434 mA cm⁻² in a flow cell configuration, which is among the best of CO₂-to-CH₄ electrocatalysts. Density functional theory calculations suggest that the hydrophobic structure decreasing the water coverage on the catalyst surface can promote the protonation of the *CO intermediate and block CO production, further favoring the generation of methane. These results provide a new insight into the electrosynthesis of DRPs via constructing a hydrophobic core–shell architecture for tuning the surface water coverage.

---

Broader context

Fossil fuel has greatly driven the economic growth and improved the living standards of humans, but this has occurred at the expense of ever-increasing anthropogenic carbon dioxide (CO₂) emissions, leading to irreversible climatic changes. For alleviating anthropogenic CO₂ emissions, electrochemical conversion of CO₂ into value-added fuels and chemicals utilizing renewable electricity is a practical and promising strategy. However, achieving high product selectivity and partial current density for desirable deep-reduction products (DPRs) like methane (CH₄) is a big challenge. Due to the high susceptibility of the CO₂ reduction reaction (CO₂RR) to the local environment and mass transportation near the gas–solid–liquid interface, tuning interfacial properties by engineering the surface wettability of the catalysts holds great promise to regulate the CO₂RR performance. In this work, we open up a new avenue for narrowing the selectivity for DRPs by controlling the hydrophobicity for tuning water availability on the core–shell architecture. The rationally designed and constructed hydrophobic core–shell Cu-based catalyst possesses remarkable CH₄ formation activity and selectivity compared to the reported Cu-based catalysts.

Introduction

The renewable-electricity-powered carbon dioxide reduction reaction (CO₂RR) provides an effective strategy to utilize CO₂ and mitigate CO₂ emissions.^1–3 Among the developed electrocatalytic material candidates, copper-based materials are well known as the group that can selectively convert CO₂ to hydrocarbons and alcohols in an aqueous solution.⁴ Unfortunately, their poor selectivity and low partial current density, especially towards deep-reduction products (DRPs, e.g., eight-electron transfer for methane (CH₄), and twelve-electron transfer for ethylene (C₂H₄) and ethanol (C₂H₅OH)), significantly limit the CO₂ single-pass conversion, which is among the largest challenges for electrochemical CO₂ utilization.⁵ Hitherto tremendous research efforts have been devoted to catalyst design via tuning the morphology,^6,7 manipulating the oxidation states,^8,9 or creating defects,^10,11 with optimized adsorbate energies thereby promoting the selectivity of copper-based electrocatalysts.^12–14

Besides the modulation of the catalyst electronic structures, constraining the available CO₂ reactant can also regulate product selectivity.^15,16 Recent reports have exploited the idea that electrode design of triple-phase boundary construction can concentrate CO₂ and accommodate the local alkali environment to enhance the CO₂RR activity.^17,18 For instance, increasing the hydrophobicity of the catalyst electrode for constructing abundant gas–liquid–solid interfaces has been proposed as the governing factor to boost DRP production.¹⁹ In parallel, co-feeding CO₂ with other gases (e.g., CO, N₂, and O₂) to decrease the local CO₂ concentration can also narrow the product distribution for DRPs.^20,21 Moreover, it has been recently reported that the molar fraction of water (H₂O) at the cathodic GDE/membrane interface does not change as a function of humidification of the CO₂ feed.²² On the basis of the above results, we attempt to initially tune the availability of local H₂O, which has a substantial effect not only on the competing hydrogen evolution reaction (HER) but also on the CO₂RR, via the catalyst structure design. The elaborately designed hydrophobic structure is more sophisticated to form a gas–solid–liquid interface on the surface of each catalyst particle, and is conducive to the effective regulation of electron–proton transfer, significantly affecting the overall CO₂RR performance.

Herein, we explore a new strategy to increase the DRP activity of CH₄, the deepest C₁ and the simplest hydrocarbon product, via controlling the local H₂O availability on the surface of the copper-based materials over the hydrophobic core–shell architecture. A density functional theory (DFT) study illustrates the favourable pathway for CH₄ generation on the Cu surface with low local H₂O concentration compared with the competing *CO desorption and C–C coupling process. In experiments, we rationally crafted the hydrophobic carbon shell coated on the copper/cuprous oxide core (H-CuO_x@C, although oxidized Cu would be reduced to metallic Cu during CO₂RR, the obtained samples were all named according to their pristine structure) to push away the aqueous electrolyte for constraining the H₂O coverage on the surface (Fig. 1a). The resultant H-CuO_x@C can sustain the hydrophobicity under CO₂RR conditions (aqueous solution and secure negative-potential biased) and deliver a partial current density of CH₄ (j_CH4) of −39 ± 3 mA cm⁻² with a peak CH₄ faradaic efficiency of 81 ± 3% at the potential of −1.60 V vs. reversible hydrogen electrode (vs. RHE, all potentials were referenced to RHE unless mentioned otherwise; the potentials were not iR-corrected if not mentioned) in the neutral medium. Through using a flow cell system in the alkaline medium, the crafted catalyst can reach a maximum FE_CH4 of 73.3 ± 5.5% at a j of −500 mA cm⁻². Notably, j_CH4 of H-CuO_x@C is as high as −434 mA cm⁻² at an applied j of −700 mA cm⁻², reaching a maximum CO₂-to-CH₄ conversion rate of 0.56 μmol cm⁻² s⁻¹. These performances exceed those of the wettable CuO_x (W-CuO_x) controls and most-reported methane-selective catalysts. We highlight that controlling the hydrophobicity for tuning water availability on the core–shell architecture might open up a new avenue for narrowing the selectivity for DRPs.

Fig. 1 (a) Schematic illustration of the hydrophobic core–shell architecture, which can constrain H₂O availability via hydrophobic carbon coating. (b) Three different pathways from *CO desorption to CO_(g), *CO protonation to *CHO or *COH, and *CO coupling to *OCCO, with the active Cu site denoted by an asterisk (*). Orange, red, grey, and white balls represent copper, oxygen, carbon, and hydrogen atoms, respectively. (c) Reaction energies for subtraction of three competing reactions (*CO desorption, protonation of *CO to *CHO or *COH, and C–C coupling to *OCCO) under different H₂O quantities considering the potential of −0.74 V vs. RHE.

Experimental

Materials and chemicals

Copper nitrate trihydrate (Cu(NO₃)₂·3H₂O) was purchased from Aladdin, and chitosan was obtained from Adamas Reagent Co., Ltd. Polyvinylpyrrolidone (PVP, M. W. 8000), carbon paper (TGP-H-060) and Nafion 115 proton exchange membrane were bought from Alfa Aesar. Ethylene glycol (AR, ≥99.0%) was obtained from Shanghai Titan Scientific Co., Ltd. Potassium bicarbonate (KHCO₃) and isopropanol (AR, ≥99.0%) were obtained from Shanghai Chemical Reagent Co., Ltd. Nafion (5 wt% in a mixture of lower aliphatic alcohols and water) was obtained from Sigma-Aldrich. Carbon dioxide (CO₂, 99.9999%) was bought from Shanghai Jiajie Special Gas Co., Ltd. All reagents were commercially available and of analytical grade. The water used was purified using a Millipore system (typically 18.2 MΩ cm resistivity).

Materials characterization

The crystal structure was determined using X-ray diffraction (Bruker D8 Advanced Diffractometer with Cu Kα radiation). The morphology and structure were characterized using a scanning electron microscope (Hitachi S4800) and a transmission electron microscope (TEM, JEM 2100, 200 kV). Scanning transmission electron microscopy (STEM) characterization was performed using a ThermoFisher Talos F200X instrument. High-angle annular dark field (HAADF)-STEM images were recorded using a convergence semi angle of 11 mrad, and inner- and outer collection angles of 59 and 200 mrad, respectively. Energy dispersive X-ray spectroscopy (EDS) was carried out using 4 in-column Super-X detectors. Fourier transform infrared spectroscopy (FT-IR) was performed using a Nicolet 6700 spectrometer with a spectral range of 4000–400 cm⁻¹. Volumetric CO₂ adsorption measurement (Micromeritics ASAP 2020) was analyzed at 298 K and 1 atm. Brunauer–Emmett–Teller (BET, Micromeritics ASAP 2460) surface area measurement was performed at 77 K in the N₂-adsorption mode. The contact angles were measured using a contact angle system (Shanghai Zhongchen, JC2000D3) at ambient temperature, with the probe liquid being 2 μL of water. The chemical state was analyzed by X-ray photoelectron spectroscopy (XPS, Thermo Escalab 250), and the binding energy of the C 1s peak at 284.8 eV was taken as the internal standard. Raman analysis was carried out by using a Leica DMLM microscope (Renishaw) with a 514 nm laser. XAFS spectra at the Cu K-edge were performed on the 1W1B beamline station of the Beijing Synchrotron Radiation Facility (BSRF), China. Cu foil, Cu₂O, and CuO were used as references.

The in situ electrochemical attenuated total reflectance infrared spectroscopy (ATR-IR) was performed in a homemade cell attached to a PerkinElmer spectrum 100 spectrometer. Catalyst powders were deposited onto the carbon fiber paper electrode as the working electrode, tightly contacting the surface of the silica prism. A Pt counter electrode, an Ag/AgCl reference electrode, and a gas inlet to purge CO₂ were also contained in the cell. Prior to electrochemical measurements, the electrolyte (0.1 M KHCO₃) was injected into the cells and purged with high-purity CO₂ (99.9999%). A CHI1242C potentiostat was employed to record the electrochemical response. The spectrum was collected every minute at −1.40 V vs. RHE with a dwell time of 30 min.

H-CuO_x@C synthesis

The H-CuO_x@C sample was constructed by a hydrothermal method. Firstly, 0.435 g of Cu(NO₃)₂·3H₂O and 0.676 g of PVP were dissolved in 15 mL and 30 mL of ethylene glycol, respectively. Then, the Cu(NO₃)₂ solution was slowly dropped into the PVP solution under vigorous stirring to form homogeneous solutions, followed by adding 0.322 g chitosan. The autoclave was sealed and heated at 180 °C for 2 h. After the hydrothermal reaction, the obtained samples were thoroughly washed with ethanol and water three times each.

W-CuO_x synthesis

The preparation method of W-CuO_x was the same as the above procedure except without the addition of chitosan.

Preparation of electrodes

The sample ink was prepared by dispersing 5 mg of catalyst (H-CuO_x@C or W-CuO_x) and 50 μL of Nafion solution into 0.125 mL of H₂O and 0.375 mL of isopropanol followed by sonication for more than 30 min. 25 μL of ink was dropped onto a carbon paper (CP, 1.0 × 1.0 cm²) and dried in the air 2 times. Then H-CuO_x@C/CP and W-CuO_x/CP were used for XRD, XPS, Raman, and XAFS measurements.

A glassy carbon electrode (GCE) was used as the working electrode for electrochemical measurements in an H-type gas-tight reactor (H-Cell). The sample ink was prepared as mentioned above. 2.5 μL of the ink was dropped on the GCE (with a diameter of 3 mm) and dried in the air 2 times with the total catalyst loading amount of about 0.7 mg cm⁻².

The gas diffusion electrodes (GDE) with the H-CuO_x@C catalyst deposited on the Cu coated polytetrafluoroethylene (PTFE) gas diffusion layers were constructed and used as the working electrodes to evaluate CO₂RR performance in a flow cell. To prepare Cu/PTFE, approximately 200 nm nominal thick Cu films were constructed by the vacuum evaporation method on the PTFE substrate (with an average pore diameter of 0.22 μm) using Cu targets (99.999%) at an evaporation rate of around 0.5 Å s⁻¹ in an OMV FS300-S6 evaporating tool at a base pressure of <6 × 10⁻⁴ Torr. The sample ink was prepared by dispersing 5 mg of the catalyst and 20 μL of Nafion solution into 0.25 mL of H₂O and 0.75 mL of ethanol followed by sonication for more than 1 h. Then the ink was sprayed onto a Cu/PTFE film electrode using an air-brush, with the total catalyst loading amount of about 0.5 mg cm⁻².

Electrochemical measurements

All electrochemical studies were performed using an electrochemical station (CHI 660E). The H-Cell consists of two compartments which are separated by a Nafion 115 membrane. In the three-electrode system, a customized GCE with a surface area of 0.07 cm² and a catalyst mass loading of 0.7 mg cm⁻² was used as the working electrode. An Ag/AgCl electrode (3.5 M KCl) and platinum mesh were used as the reference electrode and the counter electrode, respectively. CO₂-saturated 0.1 M KHCO₃ (pH ≈ 6.75) was used as the electrolyte; CO₂ was continuously purged through the electrolyte (both catholyte and anolyte were 30 mL in each compartment) at a rate of 20 sccm and was routed into the gas chromatograph. All potentials measured were calibrated to the RHE reference scale using E_RHE = E_Ag/AgCl + 0.059 × pH + 0.205 (the potentials were not iR-corrected if not mentioned). The linear sweep voltammetry (LSV) at a scan rate of 1 mV s⁻¹ was performed in Ar-saturated and CO₂-saturated 0.1 M KHCO₃ aqueous solution. The current density was calculated by normalizing the current to the corresponding geometric surface area.

The large current density performance of H-CuO_x@C was determined in a flow cell configuration. A Sustainion® membrane (Dioxide Materials) was activated in 1.0 M aqueous KOH solution for 24 h, washed with H₂O, and used as the anion-exchange membrane. The anode consisted of IrTaO_x supported on a Ti mesh was prepared by a dip coating and thermal decomposition method. 36 mg of IrCl₃·was dissolved in 4.75 mL of H₂O and 1.5 mL of concentrated HCl. 18.9 mg of TaCl₅ was dissolved in 1.5 mL of methanol and added into the IrCl₃ solution. A Ti mesh was dip coated with the mixed solution by sonication for 5 min, then dried at 120 °C for 0.5 h and calcinated at 500 °C for 1 h in air. The procedure was repeated ten times to increase the loading amount. The GDE was attached on the cathode using a copper tape at the edge of the electrode to electrically connect with the cathode, whose exposure area was 0.5 cm². PTFE spacers were used for assembling the flow cell. 1.0 M KOH flowed (10 mL per min) in the chambers between the membrane and the working electrode, and the membrane and the anode as a liquid electrolyte. CO₂ gas flowed behind the PTFE layer of GDE at a rate of 20 sccm. Chronopotentiometry experiments were performed at currents ranging from −300 to −800 mA cm⁻², showing continuous gas product distributions.

Product analysis

The gas products of CO₂ electroreduction were analyzed using a gas chromatography instrument (GC online test, RAMIN, GC2060), equipped with a flame ionization detector (FID for CO, CH₄, and C₂H₄) and a thermal conductivity detector (TCD for H₂). Ar was used as the carrier. Faradaic efficiencies were tested on-line and collected several times to obtain the average value. Liquid products were quantified by ¹H nuclear magnetic resonance (NMR) (Varian 700 MHz spectrometer, 16.4T). In a typical analysis, after 5 hours’ CO₂RR at an applied potential of −1.40 V vs. RHE, a mixture of 500 mg of the electrolyte and 100 mg of 4.8 ppm DMSO (used as internal standard) in D₂O solution was used as the measured sample. The ¹H spectra were obtained by using a pre-saturation method to suppress the water peak.

Assuming that eight electrons are needed to produce one methane molecule, the FE and the partial current densities of CH₄ formation were calculated as follows:

where v = volume concentration of CH₄ obtained from the gas chromatography (GC) data, p₀ = 1.013 bar, T (K) is the temperature of the test environment, G = 20 sccm is the CO₂ flow rate, i_total (mA) = steady-state cell current, F = 96 485 C mol⁻¹, and R = 8.314 J mol⁻¹ K⁻¹.then:

j_CH4 = FE_CH4 × i_total × (electrode area)⁻¹

DFT calculation method

All the spin-polarized DFT computations were performed by using the Vienna ab initio simulation package (VASP)²³ and the ion cores represented by the projector augmented wave (PAW) potentials. The generalized gradient approximation (GGA) in the form of Perdew–Burke–Ernzerhof (PBE) functional was used in this work.²⁴ A cutoff energy of 450 eV for the plane-wave basis set was adopted. The convergence threshold for the structure relaxation was set to be 0.05 eV Å⁻¹ in force and 10⁻⁵ eV in energy. A vacuum space exceeding 15 Å was employed. The computations on Cu(111) were performed in a p(2 × 2) periodic slab with four layers with a 2 × 2 × 1 Monkhorst–Pack k-point mesh for all the calculation models. The van der Waals (vdW) interactions were taken into consideration by the method of Grimme (DFT-D3).²⁵

Gibbs free energy calculation method

The reaction free energies of the four competing steps (*CO desorption, protonation of *CO to *CHO or *COH, and C–C coupling to *OCCO) were calculated using the following equations:

ΔG(CO) = E(*) + E(CO_(g)) − E(*CO) + (ΔZPE − TΔS)_CO

ΔG(*CHO) = E(*CHO) − E(H⁺ + e⁻) − E(*CO) + (ΔZPE − TΔS)_*CHO + eU

ΔG(*COH) = E(*COH) − E(H⁺ + e⁻) − E(*CO) + (ΔZPE − TΔS)_*CHO + eU

ΔG(*OCCO) = E(*OCCO) − 2E(*CO) + (ΔZPE − TΔS)_*OCCO

where E(CO_(g)), E(*), E(*CO), E(*CHO) and E(*OCCO) are the total energies of the gaseous molecule CO, clean surface, and the adsorbed *CO, *CHO and *OCCO on the surface, respectively. In our calculation, zero point energy (ΔZPE) and entropic contributions (TΔS) were also considered, which resulted from the vibrational frequencies; for the gaseous molecules (CO and H₂), the entropy contributions of gaseous molecules (TΔS) were obtained with the experimental values.²⁶ Notably, the computational hydrogen electrode (CHE) model was used to express the chemical potential of a proton–electron pair at 0 V (vs. RHE).²⁷

Results and discussion

We first sought to investigate the availability of the local H₂O concentration on CO₂RR selectivity over Cu catalysts with a relatively high CO₂ concentration (corresponding to high *CO coverage¹⁵) using DFT. With *CO as an important intermediate in CO₂RR, as shown in Fig. 1b, three competing conversion pathways occur: (i) *CO desorption to form the gaseous CO(g), (ii) *CO protonation into *CH₄, *CO → *CHO (or *COH) →⋯→ *CH₄, and (iii) the C–C bond coupling in two *CO to form multi-carbon products.²⁸

We studied the influence of local H₂O quantities for CO₂RR on common Cu(111) surface with high *CO coverage by the above competing pathways (Fig. S1, ESI†). The multi-point averaging molecular dynamics (MPA-MD) calculations²⁹ were performed to reduce the energy difference originating from the different structures of water (see details in the ESI† and Fig. S2). After calculating the reaction free energy of *CO desorbing to CO_(g) (ΔG(CO)), *CO to *CHO (ΔG(*CHO)), *CO to *COH (ΔG(*COH)), and C–C coupling to *OCCO (ΔG(*OCCO)) at different local H₂O quantities (Fig. 1c and see the detailed data in Table S1, ESI†), we found that the formation of *CHO is more exothermic than that of *COH (ΔG(*CHO) − ΔG(*CO) < ΔG(*COH) − ΔG(*CO)), indicating that *CHO is the key intermediate to form CH₄, which is similar to the reported results.^28,30 Our calculation also shows that ΔG(*OCCO) − ΔG(CO) (or ΔG(*OCCO) − ΔG(*CHO)) is always positive, showing that ΔG(*OCCO) is larger than both ΔG(CO) and ΔG(*CHO). This result indicates that the C–C coupling into C₂₊ products is not favorable on the pristine Cu(111) surface. Furthermore, comparing ΔG(*CHO) and ΔG(CO), we found that ΔG(*CHO) − ΔG(CO) decreases with the increase in H₂O content on the surface and gradually becomes more negative at the beginning. At a low H₂O content with about 2 or 3 local molecules, ΔG(*CHO) − ΔG(CO) is the lowest, i.e., *CO protonation to form CH₄ is more feasible, implying a higher CH₄ selectivity vs. CO under these conditions. The whole energy change confirms what we proposed before (Fig. S3, ESI†). With further increase in the surface H₂O content, ΔG(*CHO) − ΔG(CO) increases rapidly and becomes positive after the H₂O content reaches a certain level (e.g., 15 molecules), indicating that CO desorption becomes favored instead of *CHO formation owing to the steric squeezing from the local H₂O molecules (see the dynamic process in Fig. S4, ESI†), which seriously affects the product distribution of CH₄ and CO. In this case, the selectivity of CO might be higher compared with that of CH₄.

Taken together, we anticipate that controlling the local H₂O content at a low coverage on Cu catalysts with high CO₂ concentration could promote the selectivity to CH₄ in CO₂RR. Motivated by the DFT results, we rationally synthesized hydrophobic core–shell nanohybrids to reduce the local H₂O coverage, containing a copper/cuprous oxide sphere core and a nanocarbon shell, via a modified one-step hydrothermal method with chitosan as the reductive agent and carbon source, which has abundant amino and hydroxyl functional groups for strong chelation of Cu ions. Carbon nanoshells stemmed from the hydrothermal carbonization of chitosan, which is consistent with past reports.³¹ In comparison, the cuprous oxide sphere without carbon coating was constructed as the control sample without the addition of chitosan. We investigated the crystal structure of the synthesized samples using X-ray diffraction (XRD) (Fig. 2a), which revealed that H-CuO_x@C contains metallic Cu (JCPDS #04-0836), a small amount of Cu₂O (JCPDS #65-3288), and a broad diffraction peak between 20° and 30° which is attributed to amorphous carbon.³² For the bare pristine W-CuO_x control, all peaks could be indexed to Cu₂O (JCPDS #65-3288, Fig. S5, ESI†), which is attributed to the lack of the reductive agent, chitosan.³³ H-CuO_x@C and W-CuO_x were both reduced to metallic Cu with obvious Cu(111) diffraction peaks during the CO₂RR process (Fig. S6, ESI†). Fig. S7a (ESI†) displays the scanning electron microscopy (SEM) image of H-CuO_x@C in an enlarged view with the particle size ranging from 200 nm to 1.5 μm, and the corresponding energy-dispersive spectroscopy (EDS) mapping images suggest a sphere-like nanostructure with Cu, O, and C elements distributed universally over the sample (Fig. S7b–e, ESI†). In contrast, W-CuO_x nanospheres show a relatively rougher surface, and only Cu and O elements exist in the sample (Fig. S8, ESI†).

Fig. 2 (a) XRD pattern of H-CuO_x@C, indicating the coexistence of Cu, Cu₂O, and amorphous carbon components. (b) HAADF-STEM image of H-CuO_x@C and (c–e) the corresponding EDS elemental mapping images of (c) Cu, (d) O, and (e) C elements, respectively, displaying the carbon coating core–shell architecture. (f) Enlarged HAADF-STEM image of H-CuO_x@C, showing the ∼50 nm thick carbon shell. (g) The CA measurements of the H-CuO_x@C and W-CuO_x electrodes in their original state, after being staged for 15 min, and after CO₂RR at −1.40 V for 2 h, respectively. (h) CO₂ adsorption isotherms of H-CuO_x@C and W-CuO_x, respectively, indicating the higher CO₂ adsorption capability of H-CuO_x@C. (i) N₂ adsorption–desorption isotherm of H-CuO_x@C and W-CuO_x samples, respectively.

Transmission electron microscopy (TEM) images of H-CuO_x@C (Fig. S9 and S10, ESI†) confirm that carbon layers exist stably before and after 2 hours’ (h) electrolysis at −1.40 V. Please note that when the applied potential increased to −1.60 V, carbon layers were destroyed during the CO₂RR process (Fig. S11, ESI†), indicating the tolerance of the carbon shell under the electrolyte with appropriate bias, which would also affect the hydrophobicity. No carbon shell could be found in the bare W-CuO_x control (Fig. S12 and S13, ESI†). Both H-CuO_x@C and W-CuO_x samples were reduced to metallic Cu during the test with the obvious diffraction ring of Cu(111), which is consistent with the XRD patterns and demonstrates the validity of the DFT calculations. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was also conducted to gather detailed structure information about H-CuO_x@C (Fig. 2b–f). The HAADF-STEM image of the typical H-CuO_x@C particle (Fig. 2b) along with its corresponding EDS elemental maps demonstrate the core–shell architecture with a carbon shell. The enlarged HAADF-STEM image further confirms that the thickness of the carbon shell is about 50 nm (Fig. 2f). To characterize the cross-section structure and the crystal phase of H-CuO_x@C, the sample was embedded in epoxy resin and then cut into ultrathin slices by ultramicrotomy. HRTEM imaging was acquired and the corresponding FFT was used for analysis (Fig. S14, ESI†), indicating the existence of an amorphous carbon shell and Cu₂O component in the core. Furthermore, the FT-IR spectrum of H-CuO_x@C identified the existence of oxy-hydrogenated amorphous carbon and Cu₂O (Fig. S15, ESI†).^34,35

Contact angle (CA) measurements were implemented to investigate the wettability of the samples. The pristine bare W-CuO_x surface exhibits an external water CA of 103.6°, whereas the CA would be decreased under a solution environment, with CA of 85.4° staged for 15 min and 49.4° after 2 h of electrolysis at −1.40 V, which demonstrates its hydrophily under the real CO₂RR conditions (Fig. 2g). In contrast, by introducing a carbon shell, the H-CuO_x@C surface turns hydrophobic with a drastically increased CA of 148.2°, and could maintain the hydrophobicity both under the static state with surface water and CO₂RR conditions biased with a secure negative potential applied on the electrode in the electrolyte (Fig. 2g). The detailed time dependencies of surface wettability of H-CuO_x@C and W-CuO_x were further studied, shown in Fig. S16 (ESI†), indicating the durable hydrophobicity of H-CuO_x@C. Moreover, the CA of H-CuO_x@C will decrease to 103.8° after which the carbon shell is destroyed, which demonstrates the importance of the carbon shell for hydrophobicity (Fig. S17, ESI†). We obtained in situ diffusion reflectance infrared Fourier transform spectra (DRIFTS) of H₂O adsorption to detect the variation of H₂O on the catalyst surface (measurement details are described in the ESI†). It has been reported that the more hydrophobic the material is, the more enhanced ordering and stronger H₂O–H₂O interactions via hydrogen-bonding can occur during sorption so that the maximum of band 3000–3500 cm⁻¹ will shift to a lower wavenumber.³⁶ These conclusions are consistent with our results shown in Fig. S18 (ESI†), which further proves that H-CuO_x@C has stronger hydrophobicity. The hydrophilicity of the catalyst structure also allows the competition with adsorption of H₂O, which affects their CO₂ capture abilities significantly.^37,38 In general, the hydrophobic surface is able to enhance the local CO₂ concentration by keeping water away from the electrode surface and leaving water-free space to store gaseous CO₂. In this work, the CO₂ adsorption isotherm and N₂ adsorption–desorption isotherm further confirm that the introduction of a carbon shell could enhance the CO₂ adsorption capacity by 2 times and the CO₂/N₂ adsorption selectivity by 10 times, which might profit from the accelerated gas diffusion brought about by the hydrophobic carbon shell, thereby contributing to the enhanced generation of CH₄³⁹ (Fig. 2h and i).

In order to identify the chemical composition and surface valence states of H-CuO_x@C and W-CuO_x samples, X-ray photoelectron spectroscopy (XPS) was conducted (Fig. 3a and Fig. S19, S20 (ESI†)). The peak of Cu 2p could be fitted into three constituent peaks (Fig. 3a), which are ascribed to Cu⁺/Cu⁰ (932.6 eV), Cu²⁺ (933.8 eV), and Cu(OH)₂/CuCO₃ (935.0 eV), and the surface Cu²⁺ species indicates that the Cu⁺/Cu⁰ components were partially oxidized, ascribed to slight surface oxidation in the air.⁴⁰ In comparison, the W-CuO_x sample exhibits the main component of oxidized Cu species, which results from the lack of a reductive agent (Fig. S20, ESI†), and these results are coincident with those of the XRD analysis. After CO₂RR at a potential of −1.40 V for 2 h, the surface Cu²⁺ components of H-CuO_x@C were reduced to Cu⁺/Cu⁰ (Fig. 3a). We also employed Raman spectroscopy to probe the existence of the carbon shell. In Fig. 3b and Fig. S21 (ESI†), Raman spectra show the Cu₂O components in both H-CuO_x@C and W-CuO_x, probed by the characteristic vibrational fingerprints at 146, 217, 526, and 632 cm⁻¹ (ref. 41). Additionally, two bands located at 1327 and 1593 cm⁻¹ were only found in H-CuO_x@C, which are indexed to the D band and the G band of carbon, respectively, demonstrating the specific carbon shell.⁴² After being applied at a cathodic potential of −1.40 V for 2 h, only carbon characteristic peaks could be observed in the Raman spectra, without obvious oxide Cu species (Fig. 3b). This also proves that Cu₂O was reduced to metallic copper during the CO₂RR, but the carbon shell was undamaged, featuring the ability to continuously tune the hydrophobicity.

Fig. 3 (a) XPS spectra of the Cu 2p region and (b) Raman spectra of H-CuO_x@C collected before and after CO₂RR at −1.40 V for 2 h. (c) Normalized Cu K-edge operando XANES spectra and (d) Morlet WT of the k³-weighted operando EXAFS data for H-CuO_x@C with standard Cu foil, Cu₂O, or CuO powders as controls. The XPS and Raman spectra along with operando XAFS analysis demonstrate the reductive process of oxide Cu species to the metallic Cu⁰ states. (e) Time-dependent in situ ATR-IR differential spectra of H-CuO_x@C carried out at −1.40 V under the CO₂RR in CO₂-saturated 0.1 M KHCO₃, indicating that the hydrophobicity of H-CuO_x@C has a minor change over half an hour operation time.

We further performed X-ray absorption fine structure spectroscopy (XAFS) measurements, which can provide further insights into the electronic structures and chemical bonding of bulk metal atoms in their local environment. The ex situ X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectra indicate that H-CuO_x@C is composed of Cu and Cu₂O, while W-CuO_x is consistent with those of Cu₂O reference (Fig. S22, ESI†). To probe the oxidation state evolution of these electrocatalysts during the CO₂RR, we carried out operando XANES measurements under the same electrochemical conditions. At −1.40 V, the operando XANES spectra clearly show that both samples of H-CuO_x@C and W-CuO_x exhibited metallic Cu⁰ during the CO₂RR and kept their metallic Cu⁰ states after the test (Fig. 3c and Fig. S23, ESI†). The additional analysis of the Morlet wavelet transform (WT) was also performed to provide solid supports for the generation of metallic Cu⁰ during the CO₂RR process. As shown in Fig. 3d, a WT maximum at 3–9 Å⁻¹, which represents the existence of a Cu–O and Cu–Cu bond, is visible in H-CuO_x@C, whereas it exhibits a distinct feature at 6–8 Å⁻¹ during the CO₂RR at a working potential of −1.40 V, corresponding to the Cu–Cu bond, which suggests that the metallic Cu⁰ species occupy the main composition in the bulk during the test,⁴³ while for the W-CuO_x catalyst, the pristine main component of Cu₂O was also reduced to metallic Cu⁰ under the same reaction conditions (Fig. S24, ESI†). These XAFS results together with Raman analysis unambiguously prove the negligible changes in the valence states between H-CuO_x@C and W-CuO_x under the CO₂RR conditions which could affect selectivity. We also performed in situ electrochemical attenuated total reflectance infrared spectroscopy (ATR-IR) to prove that the hydrophobicity of the carbon shell would persist over a significant operation time. A digital photograph of a homemade cell attached to a PerkinElmer spectrum 100 spectrometer for ATR-IR measurement is shown in Fig. S25 (ESI†). We can observe in Fig. 3e that the intensity of hydrogen-bonded water peak between 3100 and 3500 cm⁻¹ remains almost the same with additional electrolysis time, which indicates that the amount of water bound on the H-CuO_x@C catalyst surface would not increase gradually during the CO₂RR process. This result also suggests that the electrowetting effect on the wettability of H-CuO_x@C during the CO₂RR experiment is minimal.

The activity of H-CuO_x@C for electrocatalytic CO₂RR was evaluated in a standard three-electrode configuration in a gastight two-compartment electrochemical cell (Fig. S26, ESI†). The electrodes were made by drop casting an ink solution containing a certain concentration of catalysts, isopropanol, H₂O and Nafion onto the glassy carbon electrodes (GCE, for more details see the ESI†). The linear sweep voltammetry (LSV) curves of H-CuO_x@C and W-CuO_x were collected in a CO₂- and Ar-saturated 0.1 M KHCO₃ solution, respectively. As shown in Fig. 4a, the LSV curve of H-CuO_x@C in a CO₂-saturated solution shows a positive shift of the onset potential and significant increase in current density compared with that measured in an Ar-saturated solution, suggesting the occurrence of CO₂RR.⁴⁴ Thereafter, the H-CuO_x@C catalyst was tested over a range of potentials for catalytic activity, and showed a marked selectivity for methane (FE_CH4) at potentials from −1.30 to −1.70 V, with a peak FE_CH4 reaching 81 ± 3% at −1.60 V (Fig. 4b). To further determine the selectivity of CO₂RR, a nuclear magnetic resonance (NMR) test was performed, and no detectable liquid product was found in ¹H NMR, which also suggests that the total FE of all the products is nearly 100% (Fig. S27, ESI†). However, at −1.60 V, the FE_H2, FE_CO, FE_CH4 and FE_C2H4 of W-CuO_x are 44%, 40%, 12% and 3%, respectively (Fig. 4b). Notably, the partial current density towards CH₄ (j_CH4) of H-CuO_x@C is obviously higher than that of the W-CuO_x control from −1.30 to −1.70 V: its j_CH4 is as high as −39 ± 3 mA cm⁻² and 15-folder larger than that of W-CuO_x at −1.60 V (Fig. S28, ESI†). As shown in Fig. 4c, H-CuO_x@C produced nearly exclusively CH₄ among the C₁ products; in contrast, the C₁ selectivity of W-CuO_x shifted from hydrocarbon to CO over the course of cathodic potential increase. Moreover, the ratios of C₂H₄ and C₁ products for H-CuO_x@C and W-CuO_x are fairly close (Fig. 4c), and these results also coincide with those of the DFT study. As illustrated in Fig. S29 (ESI†), H-CuO_x@C features a C_dl of 0.69 mF cm⁻², which is close to that of 0.52 mF cm⁻² for W-CuO_x, suggesting comparative active surface areas, which ambiguously demonstrates that the H-CuO_x@C catalyst exhibits a higher intrinsic CH₄ activity than the W-CuO_x control. To identify the role of the hydrophobic core–shell structure without destroying this architecture or changing the particle size, we employed a hydrophilic organic modifier of cetyltrimethylammonium bromide (CTAB) to change the local hydrophilicity over the catalysts (details are described in the ESI†).⁴⁵ As shown in Fig. S30a and b (ESI†), the CH₄ selectivity will be decreased when the local hydrophobicity becomes weaker (without changing the morphology), with a maximum FE_CH4 of 38.9 ± 1.3% at −1.40 V. When the catalyst is converted to being hydrophilic, it shows a major selectivity for H₂ at potentials from −1.20 to −1.70 V, with no detectable CH₄ (Fig. S30c, ESI†). The above results further demonstrate the important role of local hydrophobicity of the core–shell structure in promoting CH₄ selectivity.

Fig. 4 Electrochemical performance evaluation of H-CuO_x@C and W-CuO_x catalysts for CO₂RR. (a) The LSV curves of H-CuO_x@C and W-CuO_x catalysts with respect to RHE at a scan rate of 1 mV s⁻¹ in the CO₂- or Ar-saturated 0.1 M KHCO₃ electrolyte, indicating the occurrence of the CO₂RR. (b) Faradaic efficiencies of gas products on H-CuO_x@C and W-CuO_x, exhibiting a higher CH₄ selectivity for H-CuO_x@C. (c) CH₄/CO and C₂H₄/C₁ ratios of H-CuO_x@C and W-CuO_x catalysts at different applied potentials, suggesting the superior CH₄ selectivity for H-CuO_x@C. (d) Faradaic efficiencies for gas products and j_CH4 over H-CuO_x@C at various applied current densities in a flow cell reactor with 1.0 M KOH as the electrolyte. Error bars were all calculated based on the standard deviation of three measurements under each test condition. (e) Stability test of electrochemical CO₂ methanation during 30 000 s of electrolysis under a j of −700 mA cm⁻². Red dots represent FE_CH4, and the blue line represents the applied potential curve. Faradaic efficiencies of detailed gas products are shown in Table S2 (ESI†). (f) Maximum j_CH4 of H-CuO_x@C and recently reported Cu-based CO₂RR catalysts (details are shown in Tables S3 and S4, ESI†).

To further improve the activity of CO₂-to-CH₄, we evaluated CO₂RR performance in a flow cell configuration. The sprayed gas diffusion electrode (GDE) and flow cell device are shown in Fig. S31 (ESI†). H-CuO_x@C is evaluated at different j from −300 to −800 mA cm⁻² and CH₄ is detected as a major CO₂RR product in a wide j range (Fig. 4d). The representative gas chromatography result of H-CuO_x@C is shown in Fig. S32 (ESI†), showing the main CO₂RR product of CH₄. Notably, H-CuO_x@C reaches a maximum FE_CH4 of 73.3 ± 5.5% at an applied j of −500 mA cm⁻², which results in a j_CH4 of −366.5 mA cm⁻². In addition, the j_CH4 of H-CuO_x is as high as −434 mA cm⁻² at an applied j of −700 mA cm⁻², reaching a maximum CO₂-to-CH₄ conversion rate of 0.56 μmol cm⁻² s⁻¹ (Fig. 4d). The long-term stability test of H-CuO_x@C electrode at the applied j of −700 mA cm⁻² was also implemented in the flow cell configuration, which exhibits FE_CH4 of more than 50% over a time range of up to 23 700 s, suggesting the j_CH4 can be maintained over −350 mA cm⁻² for more than 6 h (Fig. 4e and Table S2, ESI†). The attenuation of CO₂RR performance might be influenced by the GDL fouling and the side reaction between the electrolyte and CO₂ in flow cells. We also compared the maximum j_CH4 for H-CuO_x@C with that for other Cu/Cu-based catalysts (Fig. 4f and Tables S3, S4, ESI†), and H-CuO_x@C with a j_CH4 of −434 mA cm⁻² was found to possess a remarkable CH₄ formation activity compared to the reported Cu-based catalysts.

Conclusions

In summary, a hydrophobic core–shell architecture was designed and constructed to tune the H₂O availability over the Cu-based catalyst surface, which could provide a high faradaic efficiency of 81 ± 3% towards CH₄ in a neutral medium and a maximum j_CH4 of −434 mA cm⁻² in a flow cell configuration, which is among the best of the Cu-based CO₂-to-CH₄ electrocatalysts. The superior DRP selectivity originates from the improved CO₂ capacity and decreased H₂O coverage induced by the hydrophobic carbon shell, which is demonstrated using experimental data and computational techniques. We highlight that numerous Cu-based catalysts can be expected to benefit from this study, with appropriate H₂O availability via catalyst architecture design to block CO generation, further improving the DRP selectivity along with other surface modulation strategies.

Author contributions

H. G. Y. and P. F. L. directed the research. X. Y. Z. performed the experiments and data analyses. W. J. L. performed the DFT calculations. Z. J. and Y. W. L. performed the XAFS test and analyses. S. D. and X. F. W. provided assistance with the HAADF-STEM characterization. M. Z. and J. C. assisted with the in situ DRIFTS and ATR-IR tests and data analyses. H. F. W. and H. Y. Y. guided the DFT calculations. All the authors participated in writing and editing the manuscript, and contributed their efforts to the discussion.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Funds for Distinguished Young Scholars (51725201), the International (Regional) Cooperation and Exchange Projects of the National Natural Science Foundation of China (51920105003), the Innovation Program of Shanghai Municipal Education Commission (E00014), the National Natural Science Foundation of China (51902105, 21902048), the Fundamental Research Funds for the Central Universities (JKD01211519), the Shanghai Engineering Research Center of Hierarchical Nanomaterials (18DZ2252400) and the Shanghai Sailing Program (19YF1411600). The authors acknowledge the support by Shanghai Rising-star Program (20QA1402400). Additional support was provided by the Feringa Nobel Prize Scientist Joint Research Center. The authors also thank the Frontiers Science Center for Materiobiology and Dynamic Chemistry. The authors also thank the crew of the BL14W1 beamline at the Shanghai Synchrotron Radiation Facility (SSRF) and the 1W1B beamline of Beijing Synchrotron Radiation Facility (BSRF) for their constructive assistance with the XAFS measurements and data analyses.

References

1. C. Hepburn, E. Adlen, J. Beddington, E. A. Carter, S. Fuss, N. M. Dowell, J. C. Minx, P. Smith and C. K. Williams, Nature, 2019, 575, 87–97.
2. R.-B. Song, W. Zhu, J. Fu, Y. Chen, L. Liu, J.-R. Zhang, Y. Lin and J.-J. Zhu, Adv. Mater., 2019, 32, 1903796.
3. R. G. Grim, Z. Huang, M. T. Guarnieri, J. R. Ferrell, L. Tao and J. A. Schaidle, Energy Environ. Sci., 2020, 13, 472–494.
4. S. Nitopi, E. Bertheussen, S. B. Scott, X. Liu, A. K. Engstfeld, S. Horch, B. Seger, I. E. L. Stephens, K. Chan, C. Hahn, J. K. Nørskov, T. F. Jaramillo and I. Chorkendorff, Chem. Rev., 2019, 119, 7610–7672.
5. J. A. Rabinowitz and M. W. Kanan, Nat. Commun., 2020, 11, 5231.
6. T. Möller, F. Scholten, T. N. Thanh, I. Sinev, J. Timoshenko, X. Wang, Z. Jovanov, M. Gliech, B. R. Cuenya, A. S. Varela and P. Strasser, Angew. Chem., Int. Ed., 2020, 59, 17974–17983.
7. F. Pan and Y. Yang, Energy Environ. Sci., 2020, 13, 2275–2309.
8. M. G. Kibria, C.-T. Dinh, A. Seifitokaldani, P. D. Luna, T. Burdyny, R. Quintero-Bermudez, M. B. Ross, O. S. Bushuyev, F. P. G. de Arquer, P. Yang, D. Sinton and E. H. Sargent, Adv. Mater., 2018, 30, 1804867.
9. D. Gao, I. Sinev, F. Scholten, R. M. Arán-Ais, N. J. Divins, K. Kvashnina, J. Timoshenko and B. R. Cuenya, Angew. Chem., Int. Ed., 2019, 58, 17047–17053.
10. T. Kim and G. T. R. Palmore, Nat. Commun., 2020, 11, 3622.
11. B. Zhang, J. Zhang, M. Hua, Q. Wan, Z. Su, X. Tan, L. Liu, F. Zhang, G. Chen, D. Tan, X. Cheng, B. Han, L. Zheng and G. Mo, J. Am. Chem. Soc., 2020, 142, 13606–13613.
12. X. Chang, T. Wang, Z.-J. Zhao, P. Yang, J. Greeley, R. Mu, G. Zhang, Z. Gong, Z. Luo, J. Chen, Y. Cui, G. A. Ozin and J. Gong, Angew. Chem., Int. Ed., 2018, 57, 15415–15419.
13. M. Luo, Z. Wang, Y. C. Li, J. Li, F. Li, Y. Lum, D.-H. Nam, B. Chen, J. Wicks, A. Xu, T. Zhuang, W. R. Leow, X. Wang, C.-T. Dinh, Y. Wang, Y. Wang, D. Sinton and E. H. Sargent, Nat. Commun., 2019, 10, 5814.
14. A. Guan, Z. Chen, Y. Quan, C. Peng, Z. Wang, T.-K. Sham, C. Yang, Y. Ji, L. Qian, X. Xu and G. Zheng, ACS Energy Lett., 2020, 5, 1044–1053.
15. X. Wang, A. Xu, F. Li, S.-F. Hung, D.-H. Nam, C. M. Gabardo, Z. Wang, Y. Xu, A. Ozden, A. S. Rasouli, A. H. Ip, D. Sinton and E. H. Sargent, J. Am. Chem. Soc., 2020, 142, 3525–3531.
16. S. He, F. Ni, Y. Ji, L. Wang, Y. Wen, H. Bai, G. Liu, Y. Zhang, Y. Li, B. Zhang and H. Peng, Angew. Chem., Int. Ed., 2018, 57, 16114–16119.
17. F. Li, A. Thevenon, A. Rosas-Hernández, Z. Wang, Y. Li, C. M. Gabardo, A. Ozden, C. T. Dinh, J. Li, Y. Wang, J. P. Edwards, Y. Xu, C. McCallum, L. Tao, Z.-Q. Liang, M. Luo, X. Wang, H. Li, C. P. O’Brien, C.-S. Tan, D.-H. Nam, R. Quintero-Bermudez, T.-T. Zhuang, Y. C. Li, Z. Han, R. D. Britt, D. Sinton, T. Agapie, J. C. Peters and E. H. Sargent, Nature, 2019, 577, 509–513.
18. C. T. Dinh, T. Burdyny, M. G. Kibria, A. Seifitokaldani, C. M. Gabardo, F. P. García de Arquer, A. Kiani, J. P. Edwards, P. De Luna, O. S. Bushuyev, C. Zou, R. Quintero-Bermudez, Y. Pang, D. Sinton and E. H. Sargent, Science, 2018, 360, 783–787.
19. W. H. Lee, Y.-J. Ko, Y. Choi, S. Y. Lee, C. H. Choi, Y. J. Hwang, B. K. Min, P. Strasser and H.-S. Oh, Nano Energy, 2020, 76, 105030.
20. Y. Xu, J. P. Edwards, J. Zhong, C. O'Brien, C. M. Gabardo, C. McCallum, J. Li, T. C. Dinh, E. H. Sargent and D. Sinton, Energy Environ. Sci., 2020, 13, 554–561.
21. X. Wang, J. F. de Araújo, W. Ju, A. Bagger, H. Schmies, S. Kühl, J. Rossmeisl and P. Strasser, Nat. Nanotechnol., 2019, 14, 1063–1070.
22. D. G. Wheeler, B. A. W. Mowbray, A. Reyes, F. Habibzadeh, J. He and C. P. Berlinguette, Energy Environ. Sci., 2020, 13, 5126–5134.
23. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758–1775.
24. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
25. S. Grimme, J. Comput. Chem., 2006, 27, 1787–1799.
26. M. W. J. Chase, NIST-JANAF Themochemical Tables, Fourth Edition, J. Phys. Chem. Ref. Data, 1998, 9, 1–1951.
27. A. A. Peterson, F. Abild-Pedersen, F. Studt, J. Rossmeisl and J. K. Nørskov, Energy Environ. Sci., 2010, 3, 1311–1315.
28. T. Cheng, H. Xiao and W. A. Goddard III, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 1795–1800.
29. D. Wang, T. Sheng, J. Chen, H.-F. Wang and P. Hu, Nat. Catal., 2018, 1, 291–299.
30. X. Liu, J. Xiao, H. Peng, X. Hong, K. Chan and J. K. Nørskov, Nat. Commun., 2017, 8, 15438.
31. Y. Yang, X. Liang, F. Li, S. Li, X. Li, S.-P. Ng, C.-M. L. Wu and R. Li, ChemSusChem, 2018, 11, 376–388.
32. B. Priyanto, M. Saleh, S. Tunmee, C. Euaruksakul, Y. Cahyono, T. Triwikantoro and D. Darminto, Mater. Sci. Forum, 2019, 966, 95–99.
33. M. S. Almughamisi, Z. A. Khan, W. Alshitari and K. Z. Elwakeel, J. Polym. Environ., 2020, 28, 47–60.
34. V. Ţucureanu, A. Matei and A. M. Avram, Crit. Rev. Anal. Chem., 2016, 46, 502–520.
35. A. O. Moghanlou, A. Bezaatpour, M. H. Sadr, M. Yosefi and F. Salimi, Mater. Sci. Semicond. Process., 2021, 130, 105838.
36. J.-L. Xu and A. A. Gowen, Spectrochim. Acta, Part A, 2021, 250, 119371.
37. J. G. Nguyen and S. M. Cohen, J. Am. Chem. Soc., 2010, 132, 4560–4561.
38. N. Ding, H. Li, X. Feng, Q. Wang, S. Wang, L. Ma, J. Zhou and B. Wang, J. Am. Chem. Soc., 2016, 138, 10100–10103.
39. L. Han, S. Song, M. Liu, S. Yao, Z. Liang, H. Cheng, Z. Ren, W. Liu, R. Lin, G. Qi, X. Liu, Q. Wu, J. Luo and H. L. Xin, J. Am. Chem. Soc., 2020, 142, 12563–12567.
40. I. Platzman, R. Brener, H. Haick and R. Tannenbaum, J. Phys. Chem. C, 2008, 112, 1101–1108.
41. D. Ren, Y. L. Deng, A. D. Handoko, C. S. Chen, S. Malkhandi and B. S. Yeo, ACS Catal., 2015, 5, 2814–2821.
42. J. Wu, D. Wang, S. Wan, H. Liu, C. Wang and X. Wang, Small, 2020, 16, 1900550.
43. W. Zhang, C. Huang, Q. Xiao, L. Yu, L. Shuai, P. An, J. Zhang, M. Qiu, Z. Ren and Y. Yu, J. Am. Chem. Soc., 2020, 142, 11417–11427.
44. W. Ren, X. Tan, W. Yang, C. Jia, S. Xu, K. Wang, S. C. Smith and C. Zhao, Angew. Chem., Int. Ed., 2019, 58, 6972–6976.
45. A. K. Buckley, M. Lee, T. Cheng, R. V. Kazantsev, D. M. Larson, W. A. Goddard III, F. D. Toste and F. M. Toma, J. Am. Chem. Soc., 2019, 141, 7355–7364.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental and computational details, TEM and SEM images, XRD, XPS, Raman, XAFS and performance comparison. See DOI: 10.1039/d1ee01493e

‡ These authors contributed equally.

---