Improved electrochemical conversion of CO₂ to multicarbon products by using molecular doping

The study addresses the challenge of efficiently converting CO₂ into higher-value multicarbon (C₂⁺) products via the electrochemical CO₂ reduction reaction (CO₂RR). While significant progress has been made for C₁ products (CO, CH₄), selective and efficient production of C₂⁺ species (e.g., ethylene, ethanol) on copper-based catalysts remains difficult. Prior approaches include alloying, surface doping, ligand modification, and interface engineering. The authors propose a molecular engineering strategy to tune the surface electronic state of Cu in Ag–Cu catalysts using electron-withdrawing aromatic heterocycles, hypothesizing that controlled p-doping (Cu^δ+) will promote key intermediates (*CO) and C–C coupling, enhancing activity and selectivity toward C₂⁺ products.

The paper situates its approach within several bodies of work: (1) Cu-based catalysts are uniquely capable of forming multicarbon products (Hori et al.). (2) Strategies to enhance C₂⁺ selectivity include alloying (e.g., Ag–Cu systems), induced strain, doping (e.g., B, halides), controlled oxidation states (Cu^+ or Cu^δ+), and electrolyte/interface engineering. (3) Molecular modifiers (N-aryl pyridinium salts, imidazole, thiols, cysteamine) have been shown to influence CO₂RR selectivity by stabilizing intermediates, modulating proton availability, or acting as redox mediators; hydrophobic modifications can suppress HER. (4) Literature suggests that increasing the Cu oxidation state can promote C₂⁺ formation, and that adsorbed *CO conformation (atop vs bridge) relates to C–C coupling pathways. The present work extends these concepts by using electron-withdrawing aromatic heterocycles to controllably p-dope Cu surfaces on bimetallic Ag–Cu catalysts.

Catalyst synthesis and functionalization: A two-step electrodeposition produced porous Ag–Cu on gas diffusion electrodes (GDEs). Ag was pulse-electrodeposited from 0.01 M AgNO₃, 0.6 M (NH₄)₂SO₄, 0.04 M ethylenediamine at 15 mA cm⁻² with 0.25 s on/3 s off. Cu was then deposited at 400 mA cm⁻² for 45 s from 0.2 M CuSO₄ + 1 M H₂SO₄ under continuous CO₂ bubbling. Optimal composition was 15 at.% Ag. Surface functionalization used 5 mM ethanolic solutions of thiadiazole (N₂SN), triazole (N₃N), 1,3,4-thiadiazole-2,5-dithiol (N₂SS), cysteamine (C₂N), or 1-propanethiol (C₃). 20 µL was drop-cast, allowed to react 5 min, rinsed with ethanol, and dried under Ar. Characterization: Morphology and composition were probed by SEM, HR-(S)TEM, HAADF-STEM, EDS, and EELS; an amorphous organic layer (~2.5 nm) was observed with uniform S, N, C distribution. Raman and FTIR confirmed grafting (characteristic bands assigned to C–C/C–N stretches, NH₂ scissor, C–N stretch). XPS (S 2p and N 1s) deconvolution supported thiadiazole/triazole attachment; XRD indicated distinct Ag and Cu phases (no alloying). Cu oxidation state and electronic structure were analyzed by Cu L₃M₄₅M₄₅ Auger and operando Cu K-edge XANES. Computations: DFT (VASP, PAW, PBE, 500 eV cutoff) modeled adsorption of thiadiazole on a 5-layer Cu(111) slab. The lowest-energy configuration placed N₂–N₃ atoms over Cu(111) with adsorption energy −1.08 eV (≥0.37 eV more stable than alternatives). Electrochemistry in H-cell: CO₂- or Ar-saturated 0.5 M KHCO₃; Ag/AgCl reference, Pt counter; potentials converted to RHE. LSV, product analysis by GC (Micro GC-490, TCD) and ¹H NMR (internal standard DSS) quantified gas and liquid products. ECSA determined by Pb UPD (100 mM HClO₄ + 1 mM Pb(ClO₄)₂); EIS assessed charge transfer. Operando spectroscopy: Operando Raman (633 nm, modified flow cell) tracked *CO-related modes (≈280, ≈365, 1900–2120 cm⁻¹) under CO₂RR. Operando XANES (BSRF IW1B; SOLEIL SAMBA) measured Cu K-edge during CO₂RR at fixed potentials, in fluorescence mode. MEA electrolyzer: 4 cm² MEA with Ag–Cu cathode, Ti–IrO₂ mesh anode (2 mg cm⁻²), and anion exchange membrane (Fumapem FAA-3-50). Cathode feed: humidified CO₂ (typically 10 sccm; varied 3–100 sccm). Anolyte: 0.1 M KHCO₃ at 30 mL min⁻¹. Full-cell voltage swept from −2.8 to −4.8 V; products collected via cold trap and online GC; liquids quantified by NMR. Stability tests at −4.55 V for up to 100 h.

• Molecular functionalization with electron-withdrawing aromatic heterocycles (thiadiazole N₂SN and triazole N₃N) p-dopes surface Cu on Ag–Cu, creating Cu^δ+ (0 < δ < 1), verified by Cu Auger and operando XANES (Cu oxidation states: N₂SN +0.53; N₃N +0.47; stable at +0.51 after 30 min CO₂RR at −1.2 V vs RHE).
• H-cell performance (0.5 M KHCO₃): At −1.2 V vs RHE, FE(C₂⁺) = 57.3% (N₂SN) and 51.0% (N₃N) vs 18% for pristine Ag–Cu (3.1× and 2.8× enhancements). Ethylene and ethanol were major C₂⁺ products with traces of acetate and n-propanol. Specific C₂⁺ current densities increased up to 5×; ECSA-normalized j(C₂⁺) = 5.3 mA cm⁻² for N₂SN (~5× vs pristine). N₂SN/N₃N also yielded the highest FE(C₂⁺)/FE(H₂) ratios. 20 h operation showed stable current retention (N₂SN 94%, N₃N 91%) and sustained FE(C₂⁺) (N₂SN 54%, N₃N 46.5%).
• Control functionalizations with short alkyl/amine (C₃, C₂N) did not enhance CO₂RR: larger HER, minimal C₂⁺ formation, increased charge-transfer resistance (EIS).
• Mechanistic correlations: Operando Raman showed increased intensities at ~365 cm⁻¹ (Cu–CO stretch) and 1900–2120 cm⁻¹ (C=O stretch) for N₂SN/N₃N, correlating with higher FE(C₂). A volcano-type dependence of FE(C₂) on the CO_top/CO_bridge ratio (max at 0.4–0.5) and on Cu oxidation state was observed. N₂SS (similar contact angle to pristine) still improved FE(C₂) to 43.7% (−1.2 V), indicating hydrophobicity is not the main factor.
• MEA electrolyzer performance (4 cm²): N₂SN-functionalized Ag–Cu achieved FE(C₂⁺) ≈ 80±1% with specific j(C₂⁺) = 261.4 mA cm⁻² and full-cell EE(C₂⁺) = 20.3%. FE(H₂) minimized (~14%). Ethanol and n-propanol FEs reached 16.5% and 6.1% at −4.4 V. Optimal CO₂ flow rate was 10 sccm; FE(C₂H₄) peaked at 56% at 10 sccm and decreased at lower/higher flows. CO₂-to-C₂⁺ conversion rate reached 785 µmol h⁻¹ cm⁻². Stability over 100 h at −4.55 V: average FE(C₂H₄) ~51% with ~1.6 A current; 94–99% retention of FE and current.

The findings support the hypothesis that electron-withdrawing aromatic heterocycles grafted on Ag–Cu tune the Cu surface electronic structure toward Cu^δ+ (p-doping), which stabilizes adsorbed *CO—particularly atop-bound CO—and promotes C–C coupling to form C₂⁺ products. Operando XANES confirms stable intermediate oxidation state (between Cu⁰ and Cu₂O) during operation, while operando Raman correlates increased Cu–CO interactions and optimal CO_top/CO_bridge ratios with higher FE(C₂). These molecular modifiers suppress competing HER relative to CO₂RR, re-directing selectivity toward ethylene and ethanol. The approach translates from H-cell to MEA, yielding high C₂⁺ selectivity, current densities, and energy efficiencies, along with extended stability. Control experiments (alkyl/amine modifiers and matched contact-angle N₂SS) and EIS analyses indicate the primary role of electronic effects rather than hydrophobicity or mass transfer alone.

This work demonstrates a robust molecular engineering strategy to enhance electrochemical CO₂-to-C₂⁺ conversion by grafting electron-withdrawing aromatic heterocycles on Ag–Cu catalysts. The functionalization induces Cu^δ+ on the surface, increases *CO coverage in favorable configurations, and yields FE(C₂⁺) up to 80±1% with 261.4 mA cm⁻² j(C₂⁺) and 20.3% full-cell energy efficiency in an MEA, with 100 h stability. In H-cells, FE(C₂⁺) improves up to 57.3% at −1.2 V vs RHE. The study establishes clear structure–electronic state–reactivity correlations via operando spectroscopies. Future research could explore broader families of electron-withdrawing molecular dopants, optimize modifier coverage and architecture, investigate tandem catalyst designs, and advance MEA engineering for higher-rate liquid product generation and scale-up.

• Mechanistic insights are primarily correlative (operando XANES/Raman, Auger) rather than direct observation of all intermediates; atomic-scale pathways and kinetics are inferred.
• Operando spectroscopies were conducted up to −1.2 V vs RHE (H-cell); behavior under more negative potentials or at MEA operating voltages was not probed spectroscopically.
• The study focuses on specific heterocycles (thiadiazole, triazole) and one bimetallic architecture (15 at.% Ag–Cu); generality across other supports/compositions and long-term chemical robustness beyond 100 h were not assessed.
• Product distributions still favor ethylene; control over ethanol/oxygenate selectivity, and mitigation of crossover for liquids, require further optimization.
• Mass transport and water management effects in MEA are explored via CO₂ flow rate but comprehensive modeling/diagnostics (e.g., local pH, flooding) are not provided.