Metal–Organic Framework-Derived CuS Nanocages for Selective CO 2 Electroreduction to Formate

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Dec 2021Metal–Organic Framework-Derived CuS Nanocages for Selective CO2 Electroreduction to Formate Xing Zhang, Rongjian Sa, Feng Zhou, Yuan Rui, Ruixia Liu, Zhenhai Wen and Ruihu Wang Xing Zhang State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fujian 350002 University of Chinese Academy of Sciences, Beijing 100049 Google Scholar More articles by this author , Rongjian Sa State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fujian 350002 Institute of Oceanography, Ocean College, Minjiang University, Fujian 350108 Google Scholar More articles by this author , Feng Zhou State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fujian 350002 University of Chinese Academy of Sciences, Beijing 100049 Google Scholar More articles by this author , Yuan Rui State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fujian 350002 Google Scholar More articles by this author , Ruixia Liu Innovation Academy for Green Manufacture, Chinese Academy of Sciences, Beijing 100190 Google Scholar More articles by this author , Zhenhai Wen State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fujian 350002 Google Scholar More articles by this author and Ruihu Wang *Corresponding author: E-mail Address: [email protected] State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fujian 350002 University of Chinese Academy of Sciences, Beijing 100049 Innovation Academy for Green Manufacture, Chinese Academy of Sciences, Beijing 100190 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.020.202000589 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Electroreduction of CO2 to target products with high activity and selectivity has techno-economic importance for renewable energy storage and CO2 utilization. Herein, we report a hierarchical CuS hollow polyhedron (CuS-HP) for electrocatalytic CO2 reduction (E-CO2R) in neutral pH aqueous media. Under E-CO2R conditions, CuS-HP undergoes structural reconstruction into sulfur-doped metallic Cu catalyst, which promotes formate production with Faradaic efficiency >90% in a wide potential range. The durability of the in situ evolved catalyst has been demonstrated by stable operation for 36 h at a formate partial current density of ∼16 mA cm−2 at −0.6 V versus reversible hydrogen electrode. Theoretical thermodynamic analysis reveals that the sulfur-doped Cu(111) facet is responsible for high formate selectivity by promoting formate production and suppressing hydrogen evolution. Download figure Download PowerPoint Introduction Electrocatalytic CO2 reduction (E-CO2R) powered by renewable electricity is a promising route to mitigate net CO2 emissions and simultaneously synthesize value-added products.1 CO2 reduction is a highly complex process as it involves multiple reaction paths and products, such as CO, formate, ethylene, ethanol, and propanol.2–6 Therefore, active and product-selective catalysts are urgently required for E-CO2R to enhance energy efficiency and reduce the cost of product separation. Formate is an important commodity chemical, which can be directly used as fuel or feedstock in fine chemical synthesis.7,8 Catalysts based on main group metals, including indium (In), tin (Sn), lead (Pb), and bismuth (Bi), have been experimentally identified as highly selective for the conversion of CO2 into formate in neutral pH aqueous media.9–19 However, many of the reported catalysts promote formate production with Faradaic efficiency >80% only at very negative potentials [<−0.8 V vs reversible hydrogen electrode (RHE)]. Recently, transition-metal cobalt (Co)-based catalysts with several atomic-layer thickness have been shown to be active for carbon dioxide-to-formate conversion at low overpotentials.20–22 Unfortunately, this performance merit could only be achieved at very low cathodic current density (<10 mA cm−2). Metallic copper (Cu) is an appealing candidate for the construction of E-CO2R catalysts because its moderate CO binding strength enables the reduction of CO2 into multicarbon products at potentials below −1.0 V versus RHE.23–25 At low overpotentials, Cu-based catalysts mainly catalyze the reduction of CO2 into CO and formate with relatively low Faradaic efficiency.8,26,27 Controlling the morphology, oxidation state, and exposed facets of Cu can tune the binding strength of the intermediates and thus the selectivity of target products.28–32 For example, sulfur (S)-doped Cu catalysts derived from CuSx precursors have exhibited remarkably enhanced formate selectivity under E-CO2R conditions.33–37 However, Faradaic efficiencies toward formate remain <80% and a mechanistic understanding of the actual active sites is still elusive. Structural reconstruction of catalyst materials under real working conditions is a common phenomenon.3,25,38 The initial morphology, composition, and coordination environment of Cu-based catalysts significantly affect structural evolution of the active catalysts under E-CO2R conditions.39–43 Therefore, it is essential to engineer the composition, morphology, and structure of Cu-based catalyst precursors or precatalysts and track their structural evolution under working conditions for understanding reaction mechanisms and improving E-CO2R performance. Herein, we present a hierarchical CuS hollow polyhedron (CuS-HP), which was synthesized via a metal–organic framework self-sacrificial template method. Using operando and postcatalysis spectroscopic characterizations, we found that the CuS phase in CuS-HP completely disappeared and a S-doped metallic Cu catalyst [Cu(S)] was generated under E-CO2R conditions. The in situ evolved Cu(S) affords formate Faradaic efficiency of 94% at −0.5 V versus RHE and partial current density of 57 mA cm−2 at −0.8 V versus RHE. Theoretical calculations identified that the S-doped Cu(111) surface is necessary for high formate selectivity due to its lower thermodynamic energy barrier for formate production and weaker hydrogen binding strength compared with those on Cu(111). Experimental Methods Synthesis of CuS-HP HKUST-1 was initially prepared per the modified literature method.44 Using ultrasonication for 3 min, 60 mg of HKUST-1 was dispersed in 30 mL of ethanol. After 2 mL of 0.5 M thioacetamide ethanol solution was added, the mixture was sealed in a gas-tight glass vessel (48 mL) and heated with stirring at 90 °C for 1 h. The black precipitate was collected by centrifugation and washed with ethanol three times. Synthesis of CuS nanosheets To 6 mL of 0.5 M thioacetamide ethanol solution was added 25 mL of 0.2 M Cu(NO3)2 ethanol solution . The mixture was then sealed in a gas-tight glass vessel (48 mL) and heated with stirring at 90 °C for 3 h. The black precipitate was collected by centrifugation and washed with ethanol three times. Synthesis of Cu nanoparticles 1-Hexadecylamine (200 mg) and Cu(OAc)2 (30 mg) were dissolved in 30 mL of ethanol. After 80 mg of l-ascorbic acid was added, the mixture was sealed in a gas-tight glass vessel (48 mL) and heated with stirring at 90 °C for 0.5 h. The reddish precipitate was collected by centrifugation and washed with ethanol three times. Other details related to material characterization, electrochemical measurements, and computational methods are available in the Supporting Information. Results and Discussion CuS-HP was readily synthesized from the HKUST-1 template via the solvothermal sulfidation method. As illustrated in Figure 1a, the octahedron-shaped HKUST-1 was initially prepared by coordination assembly of 1,3,5-benzenetricarboxylic acid and Cu2+ in methanol ( Supporting Information Figure S1).44 The formation of CuS-HP follows the Kirkendall-type mechanism. Specifically, thioacetamide releases S2−, which etches the surface of HKUST-1 to form a thin CuS seed layer. The dissolution of metastable HKUST-1 results in fast outward diffusion of Cu2+, and subsequent reaction with S2− enables the resultant CuS to continuously deposit on the outer surface, which generates hierarchical HP consisting of CuS nanosheets. The powder X-ray diffraction (XRD) pattern of CuS-HP fit well with the hexagonal CuS phase (Figure 1b). Transmission electron microscopy (TEM) and high-angle annular dark-field scanning TEM (HAADF-STEM) images show that CuS-HP has a mean size of ∼1.3 μm and is composed of crystalline CuS nanosheets with a thickness of ∼3 nm (Figures 1c–1f and Supporting Information Figure S2). The HAADF-STEM image and coupled energy-dispersive X-ray spectroscopy (EDS) elemental maps show the uniform distribution of Cu and S over CuS-HP (Figures 1g–1j). The Raman spectrum of CuS-HP exhibits a strong resonance peak centered at 470 cm−1, which is assigned to the S–S bond stretching vibration in CuSx ( Supporting Information Figure S3a).45,46 X-ray photoelectron spectroscopy (XPS) characterizations confirm that the binding energies of Cu 2p and S 2p correspond to those of CuSx ( Supporting Information Figures S3b–S3d).37 The Cu:S ratio quantified by XPS is 1.06, which conforms to the stoichiometry of CuS. Figure 1 | Structural characterization of CuS-HP. (a) Schematic illustration of the synthetic process. (b) XRD pattern. (c–f) TEM images. (g–j) HAADF-STEM image and corresponding EDS elemental mapping images. Download figure Download PowerPoint The E-CO2R performance of CuS-HP was evaluated in an H-type electrolyzer filled with CO2-saturated 0.5 M K2SO4 aqueous solution as the electrolyte ( Supporting Information Figure S4). The catalyst electrode was prepared by pasting CuS-HP on carbon fiber paper (CFP) with a mass loading of 1 mg cm−2. After the catalyst electrode was activated by 10 cyclic voltammetry (CV) scans from 0.2 to −0.8 V versus RHE at a scan rate of 20 mV s−1, repeatable CV curves were observed. The polarization curve at 2 mV s−1 shows a rapid increase of cathodic current as the potential is more negative than −0.3 V versus RHE, indicative of high electrode activity (Figure 2a). The formate Faradaic efficiency exceeds 90% in the surveyed potential region from −0.5 to −0.8 V versus RHE, while the CO production rate is negligible (Figure 2b). The corresponding formate cathodic energy efficiencies in the surveyed potential region are >50% ( Supporting Information Figure S5). For comparison, commercially available CuS nanoparticles (NPs) with mean size of 50 nm were tested under the same conditions ( Supporting Information Figure S6). The Brunauer–Emmett–Teller (BET) specific surface area of CuS-HP (20.2 m2 g−1) is two times higher that of CuS NPs (9.8 m2 g−1; Supporting Information Figure S7). However, the electrochemical active surface area (ECSA) of CuS-HP is nearly three times larger than that of CuS NPs ( Supporting Information Figures S8a–S8c). The difference in BET surface area and ECSA can be attributed to different surface wettability to gas and liquid or structural changes under electrochemical conditions. It was found that both the electrode activity and formate selectivity of CuS NPs are significantly inferior to those of CuS-HP (Figures 2a and 2b). By normalizing to their ECSA, the specific activity of CuS-HP is slightly lower than that of CuS NPs ( Supporting Information Figure S8d). Thus, this suggests that there is a significant difference in catalytically active structures derived from CuS-HP and CuS NPs under E-CO2R conditions. The robust catalytic durability of CuS-HP was further corroborated by continuous operation at −0.6 V versus RHE over 36 h (Figure 2c). When evaluated in a gas-diffusion flow cell, which reduces mass transport limitation, CuS-HP electrode affords a formate Faradaic efficiency of 95% at a current density of −300 mA cm−2 with an applied cathodic potential of −0.91 V ( Supporting Information Figure S9). The excellent catalytic performance ranks CuS-HP in the first class of reported nonprecious metal-based formate-selective E-CO2R catalysts (Figure 2d and Supporting Information Table S1). Figure 2 | E-CO2R performance of CuS-HP. (a) Polarization curves and (b) product selectivity of CuS-HP and CuS NPs. (c) Stability test and formate Faradaic efficiencies of CuS-HP at −0.6 V vs RHE. (d) Comparison of potential-dependent formate partial current densities with those of NTD-Bi,10 S-In,12 SELF-CAT-Pb,13 Sn(S)/Au,15 SnO2 QWs,17 and AC-CuSx.35 Download figure Download PowerPoint Structural reconstruction of Cu-based catalysts under E-CO2R conditions is common and the derived active structure generally correlates with the initial morphology and structure of precatalyst.39–41 The operando Raman spectroscopic characterization was performed to track structural evolution of CuS-HP under E-CO2R conditions. During the measurements, the potential applied on the electrode was reduced by steps from open circuit potential (OCP) to −0.2 V versus RHE. Each potential was held for 15 min before recording the Raman spectra. The results show that the characteristic S–S bond stretching vibration peak totally disappeared once the applied potential was more negative than −0.15 V versus RHE, indicative of the destruction of the CuS phase (Figure 3a). The CV curves of fresh CuS-HP/CFP electrode show a cathodic reduction peak started at −0.15 V versus RHE ( Supporting Information Figure S10). This cathodic reduction process is ascribed to the transformation of CuS into Cu.34 Figure 3 | Characterizations of the structural evolution of CuS-HP under E-CO2R conditions. (a) Operando Raman spectra at different applied potentials. (b) XRD patterns, (c) S 2p, and (d) Cu 2p XPS spectra before and after constant-potential electrolysis for 1 h. Download figure Download PowerPoint Ex situ XRD and XPS were performed to investigate structural and compositional changes of CuS-HP after constant-potential E-CO2R for 1 h at −0.6 V versus RHE. The XRD pattern of CuS-HP (−0.6 V) reveals the appearance of metallic Cu phase and complete disappearance of CuS phase (Figure 3b). High-resolution S 2p XPS spectra of CuS-HP (−0.6 V) show the survival of minimal Sδ− (0 ≤ δ ≤ 2) in the derived catalysts (Figure 3c). It is noted that the signals at binding energy >164 eV in S 2p spectra are ascribed to Nafion binder or adsorbed sulfate ions from electrolyte ( Supporting Information Figure S11), which will not be involved in subsequent discussion.34 Based on the XPS analyses, the content of Sδ− is 14.1 atom % in CuS-HP (−0.6 V). When the electrolysis potential was further decreased to −0.8 V versus RHE, the sulfur content remained 12.3 atom %, suggesting high stability of sulfur dopants in Cu(S) in the surveyed potential region. The Cu 2p XPS spectra and Cu LMM auger spectra show that the mean oxidation state of Cu in both CuS-HP (−0.6 V) and CuS-HP (−0.8 V) is higher than those in CuS and metallic Cu, which could be ascribed to extensive oxidation of superficial Cu atoms of in situ evolved Cu(S) upon exposure to air (Figure 3d and Supporting Information Figure S12). Thus, these operando and postcatalysis characterization results demonstrate that under E-CO2R conditions CuS-HP is reconstructed into Cu(S), which is the active species for highly selective conversion of CO2 to formate. Sulfur-free HKUST-1 and Cu NPs were also tested under the same conditions. They mainly promote H2 generation and the formate Faradaic efficiencies are <40% over the surveyed potential region ( Supporting Information Figures S13 and S14). These observations demonstrate that S-doping in Cu is important for high formate selectivity. CuS nanosheets with disorderly stacked morphology were also synthesized by direct treatment of Cu(NO3)2 with thioacetamide. The as-synthesized CuS nanosheets have similar microstructure to the assembly unit of CuS-HP ( Supporting Information Figure S15). However, the E-CO2R activity of CuS nanosheets is inferior to that of CuS-HP ( Supporting Information Figure S16), suggesting the advantages of the regularly assembled hollow morphology of CuS-HP derived from HKUST-1. It has been widely recognized that metal–organic frameworks are one kind of promising precursor for constructing functional nanomaterials with well-defined composition, morphology, and microstructure.47,48 To better understand the effects of S-doping on the formate selectivity of Cu(S), density functional theory (DFT) calculations were performed to study reaction thermodynamics of E-CO2R and competitive hydrogen evolution reaction (HER) on Cu, Cu0.96S0.04, and Cu0.87S0.13. The Cu0.87S0.13 model slab was chosen due to its similar composition to that of CuS-HP-derived Cu(S). The Cu0.96S0.04 slab was also chosen as a model for the sample with a low S-doping amount. First, we performed the calculations on the Cu(111) surface based on two reasons: (1) Cu(111) is the most stable facet in polycrystalline Cu due to its lowest surface energy among various Cu facets.49 (2) The Cu(111) facet is selective for C1 product.50 Formate production on catalyst surfaces can proceed through either COOH* or OCHO* intermediates, whereas CO and H2 evolution occur through COOH* and H* intermediates, respectively ( Supporting Information Figures S17 and S18).15,51 On the (111) surface of all model slabs, free-energy diagrams show that COOH* is the relevant intermediate for formate production and the potential-limiting step is the second proton-coupled electron-transfer process, which results in the formation of HCOOH from COOH* (Figures 4a–4c). The theoretical overpotential for formate production is 0.22 V, which is much smaller than that of Cu (0.35 V) and Cu0.96S0.04 (0.3 V). CO evolution is suppressed due to the strong binding strength of CO* on all surfaces. With the increase of S-doping amount, H* binding strength on the (111) surface becomes weaker, and Cu0.87S0.13 shows much larger theoretical HER overpotential (0.25 V) than that of pure Cu (0.17 V) (Figure 4d). We also performed DFT calculations on the (100) surface of these model slabs ( Supporting Information Figures S19 and S20). The results show that the formation of both formate and CO are unfavorable ( Supporting Information Figures S21a–S21c). Like the case of the (111) surface, the incorporation of S atom in Cu significantly weakens H* binding strength on the (100) surface ( Supporting Information Figure S21d). Therefore, the Cu(S)(111) surface is responsible for high formate selectivity of Cu(S). Figure 4 | Theoretical thermodynamic calculations for E-CO2R on Cu(111) and Cu(S)(111) surfaces. Free-energy diagrams for HCOOH and CO on (a) Cu, (b) Cu0.96S0.04, and (c) Cu0.87S0.13. (d) Comparison of thermodynamic energy barrier for HER on Cu, Cu0.96S0.04, and Cu0.87S0.13. Download figure Download PowerPoint Conclusions We have synthesized CuS-HP consisting of crystalline CuS nanosheets and explored its electrocatalytic performance for CO2 reduction. We have demonstrated that potential-induced reconstruction of CuS-HP under E-CO2R conditions generates Cu(S), which exhibits the highest formate selectivity among the reported Cu-base catalysts. Theoretical studies have revealed that S-doping in metallic Cu does not switch the reaction paths, but lowers the energy barriers for formate production and simultaneously suppresses HER. This work highlights that engineering the initial morphology and structure of precatalysts is one effective pathway to tailor in situ evolution of catalytically active structure and thus product selectivity of E-CO2R. Conflicts of Interest The authors declare no competing financial interests. Supporting Information Supporting Information is available, including additional experimental methods, Figures S1–S21, and Table S1. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (nos. 22002168, 21975259, and 21673241), the Innovation Academy for Green Manufacture of the Chinese Academy of Sciences (no. IAGM2020C17), and the Strategic Priority Research Program of the Chinese Academy of Sciences (no. XDB20000000). The authors thank Prof. Y. Liang from Southern University of Science and Technology for the kind use of TEM and Raman instruments.