Improving CO₂ photoconversion with ionic liquid and Co single atoms

The drive toward carbon neutrality motivates efficient technologies for CO₂ conversion. Photocatalytic CO₂ reduction in water is attractive but limited by rapid charge recombination, insufficient catalytic sites, and poor selectivity. Graphitic carbon nitride (g-C₃N₄) nanosheets offer visible-light response and suitably positioned conduction bands for CO₂ reduction, yet their performance remains hampered by electron-hole recombination. Cocatalyst loading can prolong carrier lifetimes and introduce active sites. Imidazolium-based ionic liquids (ILs) possess electron-stabilizing cations and have been shown to enhance CO₂ activation and selectivity in electrocatalysis. Concurrently, single-atom cobalt sites can efficiently trap photogenerated holes and catalyze water oxidation, extending electron lifetimes. This study co-loads 1-ethyl-3-methylimidazolium tetrafluoroborate ([emim][BF₄]) and borate-anchored Co single atoms on ultrathin g-C₃N₄ to spatially separate reduction and oxidation domains, investigates their individual and synergistic effects on charge kinetics, and quantifies electron transfer efficiency during CO₂ photoreduction using in situ µs-transient absorption spectroscopy.

Imidazolium ILs feature delocalized electronic structures enabling electron extraction and stabilization, and have been widely used to catalyze and promote selective CO₂ electroreduction by lowering activation barriers and suppressing H₂ evolution. Prior work demonstrates that ILs such as [emim][BF₄] improve CO₂ conversion selectivity. Transition metal oxides, including Co oxides, serve as hole trappers and water oxidation cocatalysts; single-atom catalysts further shorten hole transfer distances and can facilitate water activation. Borate-mediated O-coordinated single-metal sites (e.g., Ni) on g-C₃N₄ have been reported, suggesting feasibility for Co single atoms. While ILs have been used as surface modifiers in some photocatalytic systems, their combination with single-atom sites to tailor both electron and hole pathways and their kinetic synergy in photocatalysis remained underexplored, particularly via in situ µs-TAS to probe multi-electron transfer processes relevant to CO₂ reduction.

Synthesis: Pristine bulk g-C₃N₄ (CN) was synthesized by H-bonding self-assembly of melamine (10 g) and cyanuric acid (4 g) in water at 80 °C for 3 h, drying at 80 °C, followed by calcination to 520 °C (1 °C min⁻¹, 4 h, N₂). Ultrathin CN was obtained by two calcinations at 500 °C for 2 h in air, then HNO₃ (5 M) treatment and vacuum drying at 80 °C. Borate-modified CN (bCN) was prepared by dispersing ultrathin CN (0.5 g) in water (30 mL), adding aqueous boric acid (0.05 M; 9–15% mass ratio to CN), stirring/ultrasonication 3 h, hydrothermal treatment at 120 °C for 2 h, washing, and drying. Co single atoms (Co-bCN) were introduced by dropwise addition of Co(NO₃)₂ aqueous solution into bCN suspension with stirring and ultrasonication, forming O-coordinated Co single atoms via ion exchange between Co²⁺ and protons of borate species. Ionic liquid [emim][BF₄] was assembled on exposed CN surfaces of Co-bCN through H-bonding interactions between [emim] hydrogen donors and sp² N atoms of CN, yielding IL/Co-bCN; IL was also loaded on bCN to form IL/bCN. Photocatalytic testing: Gas-phase CO₂ photoreduction with water vapor was conducted under UV–vis light (300 W Xe lamp). Products (CO, CH₄, H₂, O₂) were quantified by gas chromatography and mass spectrometry; ¹³CO₂ isotope labeling confirmed carbon sources. Stability tests involved repeated photocatalytic cycles. Characterization: Morphology and dispersion were probed by TEM, EDX mapping, and HAADF-STEM to visualize atomically dispersed Co. Co local structure and oxidation state were analyzed by XANES/EXAFS (Co K-edge) with fitting revealing Co(II)-O coordination. XPS assessed chemical states and IL presence (F 1s from [BF₄]⁻). Photophysical and electrochemical analyses: EPR with DMPO for •OH detection under illumination; steady-state PL; EIS; atmosphere-controlled surface photovoltage spectroscopy (AC-SPS) under N₂, air, O₂ to distinguish electron vs hole trapping roles. Electrochemical CO₂ reduction and water oxidation onset potentials were measured in CO₂/N₂-bubbled systems, respectively. Transient absorption spectroscopy: in situ µs-TAS monitored electron kinetics at 900 nm under N₂/H₂O and CO₂/H₂O atmospheres. Electron half-lifetimes (t₅₀) were extracted via power-law fits; electron transfer efficiency (ETE) during CO₂ reduction was computed from lifetimes under CO₂ vs N₂; electron transfer rates (ETR) were derived to evaluate IL loading effects. Spectroscopy of adsorption and intermediates: FTIR assessed dark adsorption of CO₂/H₂O; in situ FTIR tracked intermediates (COOH*, HCO₃⁻, CO₃²⁻, CHO⁻) under illumination, including ¹³CO₂ experiments. Theory: DFT and MD simulations modeled IL/CN interactions, Co-bCN charge distribution, CO₂ adsorption geometries and interaction energies, and differential charge densities for mechanistic insights.

• Co-loading [emim][BF₄] ionic liquid and borate-anchored Co single atoms on ultrathin g-C₃N₄ (IL/Co-bCN) yields spatially separated reduction and oxidation domains and achieves nearly 100% selectivity toward CO₂ reduction in water with no H₂ detected.
• Activity metrics under UV–vis irradiation: • bCN (optimized): CO 3.9, CH₄ 1.7, H₂ 1.7 µmol g⁻¹ h⁻¹. • IL/bCN ([emim][BF₄]): CO 12.3, CH₄ 2.8 µmol g⁻¹ h⁻¹. • Co-bCN (optimized): O₂ 867.5 µmol g⁻¹ h⁻¹; CO₂ conversion 20.6 µmol g⁻¹ h⁻¹; H₂ 9.7 µmol g⁻¹ h⁻¹ by-product. • IL/Co-bCN (optimized): CO 40.5, CH₄ 6.3 µmol g⁻¹ h⁻¹; no H₂; providing 9-fold (vs bCN) and 42-fold (vs CN) higher CO₂ conversion rates.
• Isotope labeling with ¹³CO₂ confirms formation of ¹³CO and ¹³CH₄.
• Charge kinetics and roles: • AC-SPS indicates IL attracts electrons (signal order O₂ > air > N₂), while Co single atoms trap holes (N₂ > air > O₂). • Electron half-lifetimes (t₅₀, N₂/H₂O): bCN 29.1 µs; IL/bCN 22.5 µs (reduced amplitude consistent with electron transfer to IL); Co-bCN 37.6 µs (longer lifetime due to hole trapping). bCN with MeOH shows t₅₀ 31.9 µs; Co-bCN surpasses MeOH in hole extraction strength. • Under CO₂/H₂O, IL/bCN exhibits accelerated decay (t₅₀ ≈ 16.5 µs), evidencing opened electron transfer channel to CO₂. • Electron transfer efficiency (ETE): CN ≈ 0.3%; bCN −0.7%; Co-bCN ≈ 3.2%; IL/Co-bCN ≈ 35.3% (≈93× CN), explaining large activity gains. • Electron transfer rate (ETR) increases linearly with IL loading, fitting first-order kinetics k_ET = K·c: K = 2.4 × 10⁷ M⁻¹ s⁻¹ for IL/bCN; K = 3.3 × 10⁷ M⁻¹ s⁻¹ for IL/Co-bCN; background rate k₀ = 238 s⁻¹.
• Structure and composition: • HAADF-STEM shows atomically dispersed Co; XANES/EXAFS confirm Co(II) with fivefold O coordination (Co–O peak ~1.46 Å, non-phase-corrected) forming Co(II)-O₅ single-atom sites; XPS Co 2p aligns with Co²⁺ and is unaffected by IL; F 1s from [BF₄]⁻ confirms IL presence without strong interaction with Co sites. • MD indicates stable H-bonding assembly of [emim] on CN via C2, C4, C5 hydrogens to sp² N; IL dispersion predominantly interacts with CN rather than borate species.
• Adsorption and intermediates: • FTIR adsorption shows IL/bCN significantly enhances CO₂ adsorption; Co-bCN enhances H₂O adsorption; IL/Co-bCN promotes both. • In situ FTIR detects COOH*, HCO₃⁻, CO₃²⁻, and CHO⁻ intermediates; IL increases COOH* signals; ¹³CO₂ experiments validate intermediate assignment.
• DFT/MD support: Co sites accumulate positive charge under excitation (hole trapping); electrons localize on [emim] and IL/CN interface; CO₂ adsorption is more stable at the IL/CN interface than on bare CN; differential charge density reveals electron population near [emim].

The dual-cocatalyst design decouples redox reactions: Co single atoms trap holes and catalyze water oxidation, extending electron lifetimes on g-C₃N₄, while [emim][BF₄] extracts and stabilizes electrons and catalyzes CO₂ activation and reduction, enhancing selectivity and suppressing H₂ evolution. In situ µs-TAS quantitatively links these roles to a high electron transfer efficiency (35.3%) during CO₂ reduction, much higher than single-modified or unmodified systems. Electrochemical measurements corroborate IL’s role in lowering onset potentials for CO₂ reduction and Co’s role in lowering onset for water oxidation. FTIR adsorption studies show complementary enhancements for CO₂ (by IL) and H₂O (by Co sites), and in situ FTIR captures key intermediates (COOH*, HCO₃⁻, CO₃²⁻, CHO⁻), consistent with a pathway where reduced [emim] forms CO₂–[emim] adducts, further reduced to COOH*–[emim], then proton-coupled steps generate CO (and minor CH₄), with IL possibly creating a positively charged interfacial layer repelling protons to inhibit H₂. DFT and MD substantiate electron localization on IL, hole trapping at Co sites, and favorable CO₂ adsorption at the IL/CN interface, rationalizing the observed kinetics and selectivity.

Co-loading [emim][BF₄] ionic liquid and borate-anchored Co(II)-O₅ single-atom sites on ultrathin g-C₃N₄ creates spatially separated reduction and oxidation domains that markedly improve CO₂ photoreduction in water, achieving a 42-fold enhancement over CN and nearly 100% selectivity with no H₂ evolution. In situ µs-TAS quantifies a high electron transfer efficiency of 35.3%, attributable to IL-facilitated electron capture and activation of CO₂ combined with prolonged electron lifetimes from Co-mediated hole trapping and water oxidation. The linear dependence of electron transfer rate on IL loading (first-order behavior) provides a quantitative handle to tune performance. These insights offer a strategy for designing efficient, selective artificial photosynthesis systems and a paradigm for probing charge-transfer kinetics under operating conditions.