Facile synthesis of two-dimensional copper terephthalate for efficient electrocatalytic CO2 reduction to ethylene

Abstract Electrochemical CO2 reduction (ECR) powered by renewable electricity is reckoned to provide an effective strategy to alleviate environmental issues and energy crisis, enabling a potential carbon neutral economy. To boost the ECR to fuels and value-added chemicals, the design of highly active and selective catalysts is important. In this study, we demonstrate facile ultrasonication-facilitated synthesis of two-dimensional copper terephthalate for efficient ECR. High faradaic efficiencies (FEs) of up to 72.9% for hydrocarbons are achieved at a mild overpotential in an H-type cell. In particular, the FE for ethylene (C2H4) formation approaches 50.0% at an applied potential of −1.1 V (vs. the reversible hydrogen electrode), outperforming commercial Cu, Cu2O, CuO, Cu(OH)2 and many recently reported Cu-based materials. The C2H4 partial geometric current density is as high as ∼6.0 mA·cm−2. This work offers a simple avenue to developing advanced electrocatalysts for converting CO2 into high-value hydrocarbons.


Introduction
CO 2 concentration has been progressively increasing since the preindustrial era ($280 ppm) due to anthropogenic activities, reaching over 419 ppm levels in 2021 [1][2][3][4]. This intensifies environmental issues. Electrochemical CO 2 reduction (ECR) driven by electricity generated from intermittent renewable resources provides a potential strategy to produce fuels and value-added chemicals, enabling a sustainable future [5]. Since the pioneering work dated back to the 1950s, massive efforts have been made to design and screen materials for efficient ECR [6][7][8][9][10].
In comparison to methane (CH 4 ) [11], carbon monoxide (CO) [12][13][14][15][16][17], or formic acid (HCOOH or formate [HCOO -] in alkaline electrolyte) [18,19] that are typically the major C 1 products of CO 2 reduction, converting CO 2 into C 2þ (encompassing two or more carbon atoms) hydrocarbons and oxygenates is desirable from both ecological and economic viewpoints [20][21][22][23]. Among various materials, Cu is the only active metal to electrochemically catalyse CO 2 into multi-electron transferred products because of its negative adsorption energy for CO Ã and positive adsorption energy for H Ã [24]. Cu has a suitable affinity for Ã CO in accordance with the Sabatier principle, thus facilitating further stepwise transformation to yield up to 12 C 2þ products [4,[25][26][27]. However, Cu is intrinsically limited by the scaling relations between the binding energies (BEs) of different reaction intermediates on the metallic surface. This results in wide product distributions and undesired H 2 evolution, therefore inhibiting large-scale implementation. From this scenario, endeavours have been initiated to improve the activity and selectivity by crystal facet engineering [28], heteroatom doping [29,30], alloying [31,32], construction of hybrids [33][34][35][36] or through control of morphology [37], oxidation state [38] of Cu catalysts. Despite rapid progress that has been made in ECR after decades of exploration, the viability of CO 2 electrolysis from a commercial perspective still needs to overcome key hurdles such as large kinetic overpotential (up to 1.0 V for C 2þ products), insufficient current density and low selectivity for C 2þ products, among others. To this end, the design and development of new and advanced electrocatalysts are important.
Herein, we reported the synthesis of a two-dimensional (2D) Cu-based porous structure through a facile ultrasonication-assisted coordination strategy at room temperature. The as-obtained materials facilitated the ECR to hydrocarbons at mild overpotentials. Especially, the faradaic efficiency (FE) for the formation of ethylene (C 2 H 4 ) reached up to 50.0% at an applied potential of À1.1 V vs. the reversible hydrogen electrode (RHE) in an H-type cell with 0.1 M aqueous KHCO 3 electrolyte solution, superior to commercial Cu, Cu 2 O, CuO, Cu(OH) 2 and many other Cu-based electrocatalysts demonstrated in prior literature.

Synthesis of Cu/TPA-DMF and Cu/BTC-DMF
To obtain Cu/TPA-DMF, 32 mL of DMF, 2 mL of ethanol and 2 mL of deionised water were first mixed in a 100 mL flask at 25 C under magnetic stirring. Then 0.75 mmol of TPA was dissolved into the above mixture under bath ultrasonication to yield a homogeneous dispersion. Subsequently, 0.75 mmol of CuCl 2 2H 2 O was added and the system was subjected to bath ultrasonication (40 kHz) for 8 h. Finally, the products were obtained after repeating washing with water and ethanol followed by vacuum drying at 60 C overnight. Cu/BTC-DMF was prepared via a similar procedure except the use of BTC rather than TPA.

Characterisation
X-ray powder diffraction (XRD) was performed with a D/MAX-RC diffractometer operated at 30 kV and 100 mA with Cu Ka radiation. X-ray photoelectron spectroscopy (XPS) experiments were carried out using Thermo Scientific ESCALAB 250Xi instrument. The instrument was equipped with an electron flood and scanning ion gun. All spectra were calibrated to the C 1s BE at 284.8 eV. Scanning electron microscopy (SEM) was conducted using a field emission microscope (JEOL JSM-7800F) operated at 5 kV. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was performed using a JEM-ARM200F microscope with 200 kV accelerating voltage. STEM samples were prepared by depositing a droplet of suspension onto an Au grid coated with a Lacey Carbon film.

Electrochemical tests
Typically, 10 mg of a catalyst was dispersed in 2 mL of a solution containing IPA, deionised water and Nafion solution (5.0 wt%) with a corresponding volume ratio of 100:100:1 under bath ultrasonication for 30 min to form a homogeneous ink. 200 mL of the dispersion was then deposited onto a carbon paper electrode with a size of 1 cm Â 1 cm, which was dried in the air. For linear sweep voltammograms (LSVs) in Ar-or CO 2 -saturated 0.1 M KHCO 3 electrolyte, 6 mg of a catalyst was dispersed in a mixture of ethanol (600 mL), deionised water (600 mL) and Nafion solution (5.0 wt%, 600 mL). Subsequently, the mixture was ultrasonicated for 30 min to form a homogeneous ink. 7.95 mL of the dispersion was then dropped onto a glassy carbon electrode (geometric area: 0.19625 cm 2 ; catalyst loading content: 0.135 mg cm À2 ) and dried at room temperature.
Controlled potential electrolysis of CO 2 was carried out in an H-type cell separated by a Nafion 117 membrane. Before ECR experiments, the Nafion membrane was pre-treated by heating in H 2 O 2 aqueous solution (5.0%) and H 2 SO 4 (0.5 M) at 80 C for 1 h, respectively. Then the Nafion membrane was immersed in deionised water under ambient conditions for 30 min and then washed with deionised water. Toray Carbon fibre paper with a size of 1 cm Â1 cm was used as working electrode. Pt wire and Ag/AgCl electrodes were used as counter electrode and reference electrode, respectively. Applied potentials were controlled by an electrochemical working station (CHI 760E, Shanghai CH Instruments Co., China). All potentials in this study were measured against the Ag/AgCl reference electrode (in KCl solution, 3.5 M) and converted to the RHE reference scale by ECR was initiated in CO 2 -saturated KHCO 3 solution (0.1 M) at room temperature and atmospheric pressure. CO 2 was purged into the KHCO 3 solution for 30 min to expel residual air in the reservoir, then controlled potential electrolysis was conducted at each voltage for 60 min. LSVs in Ar-or CO 2 atmosphere were carried out in a three-electrode system using Ag/AgCl as reference electrode, Pt wire as counter electrode, and glassy carbon as working electrode on a CHI 760E potentiostat. Rotating disk electrode (RDE) experiments were run on an AFMSRCE RDE control system (Pine Inc., USA). The electrolyte is 0.1 M KHCO 3 solution purged with Ar or CO 2 for 30 min.

Quantitative analysis of gaseous and liquid products
The ECR gas-phase products were probed by an Agilent 7890B gas chromatography (GC) with two thermal conductivity detectors (TCD) and one flame ionisation detector (FID). The liquid products were quantified by 1 H NMR (nuclear magnetic resonance, Bruker Avance III 400 HD spectrometer) using a solvent presaturation technique to suppress the water peak. NMR samples were prepared by mixing 0.5 mL of the product-containing electrolyte and 0.1 mL DMSO-d 6 as the internal standard. FE was determined from the amount of charge passed to produce each product divided by the total amount of charge passed at a specific time or during the overall run.

Results and discussion
In this work, we developed a very simple method for the preparation of 2D crystalline Cu/ TPA-DMF via ultrasonication treatment of a mixture of CuCl 2 2H 2 O and TPA in DMF.

Structural characterisation
The phase and structural features of the as-made samples were characterised using XRD analysis. The XRD patterns given in Figure 1 TPA-DMF with good crystallinity. No reflection peaks originating from the CuCl 2 precursor, CuO and Cu 2 O were found.
XPS was employed to investigate the surface composition and chemical state of the Cu moieties. Depicted in Figure 1(b-d) are the wide-survey, Cu 2p and O 1s XP spectra of Cu/TPA-DMF, respectively. The sample was predominantly composed of copper, oxygen, carbon and nitrogen. No other heteroelement including element Cl was detectable, suggesting that there are no unreacted precursors or byproducts in the product. The Cu 2p core-level spectrum of Cu/TPA-DMF indicated the presence of Cu 2þ in the Cu/TPA-DMF, manifesting prominent Cu 2p 1/2 and 2p 3/2 peaks with respective BEs centred at around 955.1 and 935.4 eV, along with their strong satellite peaks [39]. No apparent peaks ascribed to Cu þ can be observed. The O 1s XP spectrum was deconvoluted into three components, corresponding to C-O (533.4 eV), C ¼ O (532.5 eV) and Cu-O (531.9 eV) (Figure 1(d)). The Cu-O bond can be attributed to the coordination between copper metal ions and carboxylate-based ligands.
To decipher the morphology and microstructure of Cu/TPA-DMF, SEM (Figure 2(a)) and aberration-corrected HAADF-STEM along with energy-dispersive X-ray analysis (Figure 2(b-j)) were conducted. Large quantities of flakes that stacked on top of each other with lateral sizes of 300.0 nm À 3.0 mm were observed. Further high-resolution STEM imaging revealed that some nanosheets were decorated with CuO nanocrystals of about 7.0 nm plausibly resulting from oxidation of copper precursor in DMF. Furthermore, the energy-dispersive X-ray spectroscopy (EDS) maps (Figure 2(g-j)) together with the EDS spectrum (Figure 2(d)) confirmed that the nanosheets were composed of C, Cu, O and N elements.

Electrochemical measurements
ECR is sensitive to operating environments including the nature and properties of electrocatalyst, electrolyte type and concentration and electrochemical cell type [40]. To probe the intrinsic catalytic properties of the Cu/TPA-DMF, we performed the ECR in CO 2 -saturated 0.1 M KHCO 3 aqueous solution (pH 6.8) using an H-type cell with continuous CO 2 bubbling at room temperature and atmospheric pressure [41]. The potential-dependent geometric current densities of Cu/TPA-DMF in the voltage range of 0.0 to À1.3 V were measured by LSV, as illustrated in Figure 3(a). Invariably larger cathodic currents were observed in a CO 2 environment than in an Ar environment throughout the potential range. No liquid compounds were detected over the applied potential region. Products of CO, H 2 , CH 4 and C 2 H 4 were identified in the potential range of À1.0 to À1.3 V in a CO 2 -saturated 0.1 M KHCO 3 electrolyte. The ECR dominated over hydrogen evolution reaction (HER) at potentials between À1.0 and À1.15 V (Figure 3(b,c)). The FEs for hydrocarbons including C 2 H 4 and CH 4 reached up to 70.0% at À1.1 V with about 8.4 mA cm À2 partial geometric current density. Further elevation of overpotential resulted in decrease of C 2 H 4 FE likely owing to more severe competition with hydrogen evolution reaction at more negative voltages. It is worth noting that the FE for C 2 H 4 formation is as high as 50.0%, substantially outperforming that of TPA (FE C2H4 % 0.0%), carbon paper (FE C2H4 % 0.0%), commercial Cu (FE C2H4 % 10.6%), Cu 2 O (FE C2H4 % 7.9%), CuO (FE C2H4 % 23.8%) and Cu(OH) 2 (FE C2H4 % 13.3%) (Figure 3(d)), and many recently reported Cu-based electrocatalysts at similar overpotentials (Figure 3(e) and Supplementary Table S1). The Tafel slope for C 2 H 4 production over Cu/TPA-DMF was measured to be about 160.9 mVÁdec À1 (Supplementary Figure S1) lower than defective CuO demonstrated in prior literature, 23 indicating that the as-synthesised catalyst has good reaction kinetics for ECR. The formation of the Ã CO intermediate for tandem catalysis on the surface of the catalyst appeared to control the reaction rate. Chronoamperometric measurements showed that the FE for C 2 H 4 remained above 40.0% even after 7.0 h of successive polarisation at À1.1 V (Figure 3(f)).
The ECR performance was readily tuned by controlling the added amount of the linker and metal precursor and ultrasonic temperature (Figure 3(g)). The optimal mole ratio of TPA and CuCl 2 for ECR was found to be 1:1. The catalyst synthesised at 25 C exhibited superior ECR activity than those obtained at both 5 C and 45 C. Likewise, manipulating the type of solvent during materials synthesis allowed one to regulate the ECR property. The Cu/TPA-DMF produced in a mixture of DMF, ethanol, and water displayed considerably better activity for ECR than either in pure ethanol or water (Figure 3(g)). The type of ligand would affect the bonding mode and structure of materials, thus affecting its catalytic performance. The linker species in a metal-organic frameworks (MOF) structure can also modify the adsorption and binding of water and CO 2 reduction intermediates, providing the opportunity for fine-tuning of the ECR selectivity. Indeed, the TPA ligand imparted remarkably better activity in terms of ECR FE and C 2 H 4 selectivity than the BTC, probably due to the concomitant reduction in the numbers of unsaturated active Cu sites in the Cu/BTC-DMF (Figure 3(g)). We also  Supplementary Table S1). (f) Current density and CH 4 FE, C 2 H 4 FE, H 2 FE time response of Cu/TPA-DMF at -1.1 V. (g) C 2 H 4 FEs for CuO/Cu-MOF attained at different conditions by modulating the linker-metal precursor mole ratio, ultrasonic bath temperature, solvent, linker type and synthesis method.
prepared Cu/TPA-DMF by using agnetic fixing and hydrothermal treatment at 120 C at equivalent periods akin to the protocol of ultrasonication, which exhibited lower ECR activity in both cases (Figure 3(g)).

Conclusions
In summary, we showed for the first time a facile and effective ultrasonication-facilitated strategy to synthesise 2D Cu/TPA-DMF. The as-obtained materials significantly promoted the ECR to hydrocarbons at low overpotentials in an H-type cell with 0.1 M KHCO 3 solution, affording FEs exceeding 72.0%. The catalytic property was readily tuned by manipulating reaction conditions. A high FE for C 2 H 4 of up to 50.0% and a C 2 H 4 partial geometric current density of about 6.0 mA cm À2 were attained at À1.1 V (vs. RHE), surpassing many reported Cu-based electrocatalysts. This work provides a new possibility for designing 2D copper terephthalate to enhance the conversion of CO 2 into C 2 H 4 .

Disclosure statement
No potential conflict of interest was reported by the authors.