Low energy carbon capture via electrochemically induced pH swing with electrochemical rebalancing

Anthropogenic CO2 emissions are the major driver of climate change, and while low-carbon energy sources are growing, fossil fuel use will persist for the foreseeable future. Carbon removal via capture from point sources (e.g., power plants) or directly from the air or ocean is therefore increasingly important. Conventional approaches include wet amine scrubbing for point sources and strongly alkaline solutions for direct air capture (DAC), both relying on thermal regeneration with significant energy penalties (~100 kJ molCO2−1 for point sources and >150 kJ molCO2−1 for DAC) and concerns about volatility, toxicity, and corrosion. Ocean-based removal faces large water-handling challenges. Electrochemically mediated separations are attractive due to coupling with low-cost renewable electricity and mild operating conditions, but many systems are limited by low current densities due to overpotentials. The authors previously proposed a pH-swing cycle driven by proton-coupled electron transfer (PCET) using redox-active organics to raise and lower total alkalinity (TA) for CO2 absorption and release. This study reports a proof-of-concept CO2 separation using the PCET-active molecule DSPZ to achieve low electrical work at practical current densities and introduces an electrochemical rebalancing strategy to mitigate oxygen-induced side reactions and extend cell lifetime.

For point source capture, wet amine scrubbing is widely used but requires substantial thermal energy (~100 kJ molCO2−1) and presents solvent management challenges. DAC using concentrated alkaline solutions demands even higher regeneration energies (>150 kJ molCO2−1) and presents handling concerns. Alternative ocean-based capture methods reduce atmospheric CO2 by shifting ocean equilibria but require high water throughput. Electrochemical CO2 separation has emerged as an alternative benefiting from declining renewable electricity costs and ambient operation; however, many reported systems operate at low current densities (<5 mA cm−2) due to high overpotentials, implying larger capital costs. Prior work by the authors introduced a PCET-driven pH swing using redox molecules to modulate TA for CO2 capture/release, demonstrating feasibility but not yet optimized for energy at higher current density. The present work builds on these advances by employing DSPZ, quantifying energy vs. pCO2 and current density, and addressing oxygen sensitivity through electrochemical rebalancing.

Device and process: A Fe(CN)6|DSPZ electrochemical flow cell drives pH swing cycles for CO2 capture and release. The negolyte contains DSPZ (sodium (3,3'-(phenazine-2,3-diylbis(oxy))bis(propane-1-sulfonate))). PCET reduction of DSPZ (Q + 2H2O + 2e− → QH2 + 2OH−) increases pH/TA to absorb CO2 as carbonate/bicarbonate; subsequent oxidation (QH2 + 2OH− → Q + 2H2O + 2e−) lowers TA and releases CO2. The gas-phase upstream composition over the negolyte is controlled by N2/CO2 mass flow controllers; downstream flow rate and CO2 partial pressure are measured after drying. A pH probe tracks negolyte pH to infer TA and DIC (dissolved inorganic carbon). Four process steps define a cycle: (3→1) deacidification + CO2 invasion at inlet pCO2 (p1); (1→1′) switch headspace to exit pCO2 (p3 = 1 bar); (1′→3) acidification + CO2 outgassing at p3; (3→3′) switch back to p1. Electrolytes and cell hardware: Negolyte: 10 mL 0.11 M DSPZ in 1 M KCl (capacity-limiting, theoretical capacity 212 C). Posolyte: 35 mL 0.1 M K4Fe(CN)6 + 0.04 M K3Fe(CN)6 in 1 M KCl (non-capacity-limiting, theoretical capacity 473 C). Zero-gap flow cell (Fuel Cell Tech hardware), Fumasep E620(K) cation exchange membrane, 5 cm2 electrodes using stacks of Sigracet SGL 39AA porous carbon paper pre-baked at 400 °C, POCO graphite serpentine flow plates, Viton gaskets; assembly torque 80 lb-in (eight bolts). Electrolyte circulation: FEP tubing; both sides pumped at 100 mL min−1 (Masterflex pumps). Gas control by Sierra Smart Trak 50 MFCs. Downstream flowmeter Servoflo FS4001-100-V-A; CO2 sensor ExplorIR-W 100% sensor; Mettler Toledo LE422 pH electrode; drierite drying tube upstream of sensors. Antifoam B added (10 μL) to negolyte to suppress foaming. Experimental conditions: Inlet pCO2 p1 varied 0.1–0.5 bar (and 0.05 bar in a subset), exit pCO2 p3 = 1.0 bar; current densities 20–150 mA cm−2. For each new condition, posolyte was refreshed and negolyte acidified (drops of 1 M HCl) to reset oxygen-induced imbalance. Gas headspace alternated between p1 and p3 on a fixed schedule (every 3 h; 5 h for p1 = 0.05 bar). Measurements and calculations: TA changes inferred from charge passed assuming K+ is the only ion crossing the CEM. DIC changes quantified by downstream gas flow/CO2 sensor integration during capture/release: Q_CO2 = ∫ (Vt − Vbase) dt with baseline 11.6 mL min−1. DIC at states also computed from TA and pHmeas (TA–pH method) or from TA assuming gas-solution equilibrium (TA–eq; CO2(aq) = 0.035 × pCO2). Electrical work: Wcycle = Wdeacidification + Wacidification; Wprocess = Σ(Vn In A Δt), with A = 5 cm2. Molar cycle work w = Wcycle / ΔDICflow,3−1 (or equivalent ΔDICflow,1−3). Single-cycle demonstration: p1 = 0.1 bar, p3 = 1 bar at 40 mA cm−2 with deacidification to pH ~13.5, CO2 invasion continuing after current cutoff due to kinetics, and rapid outgassing on acidification. Electrochemical rebalancing: To counter oxygen-induced imbalance (DSPZH2 + 1/2 O2 → DSPZ + H2O) which depletes K4Fe(CN)6 and raises negolyte TA, apply −40 mA cm−2 with cathodic reduction [K+]3[Fe(CN)6]3− + e− → [K+]4[Fe(CN)6]4− and anodic oxygen evolution 2OH− → 1/2 O2 + H2O + 2e− until pH normalizes and capacity is restored. Materials: DSPZ synthesis adapted from prior work, using sodium hydride to deprotonate phenazine-2,3-diol. Instrumentation: Gamry Reference 3000 potentiostat; galvanostatic cycling to voltage cutoffs (deacidification 1.65 V, acidification 0.2 V) followed by potentiostatic holds to 10 mA cm−2.

• Proof-of-concept CO2 separation with DSPZ-based PCET pH-swing flow cell achieved low specific electrical work at practical current densities. • For capture from p1 = 0.1 bar (CO2/N2 mixture) and release at p3 = 1 bar CO2, the measured cycle work was 61.3 kJ molCO2−1 at 20 mA cm−2. • In a representative 40 mA cm−2 cycle (p1 = 0.1 bar, p3 = 1 bar): captured 47 mL CO2 (ΔDICflow,3′→1 = 0.20 M), released 49 mL (ΔDICflow,1′→3 = −0.20 M); net ΔDIC between states 3 and 1 was 0.17 M (1.7 mmol in 10 mL). Work components: Wdeacidification = 0.267 kJ, Wacidification = −0.119 kJ; molar cycle work = 87 kJ molCO2−1. Productivity (ΔDICflow,3−1 divided by capture+release time) = 0.085 M h−1 ≈ 8.5×10−4 molCO2 h−1 under limited gas–liquid contact conditions. • Across p1 = 0.1–0.5 bar, ΔDICflow,3−1 decreased with lower inlet pCO2 and agreed closely with equilibrium predictions (ΔDICTA−eq,3−1). • ΔDICflow,3−1 was independent of current density (20–150 mA cm−2) at fixed p1, indicating that current density altered rate but not extent when sufficient time was allowed to approach equilibrium. • Energy trends with current density: Molar cycle work increased with current density due to higher ohmic, charge-transfer, and mass-transport overpotentials; acidification work was largely independent of p1 (fixed p3), while deacidification work decreased with increasing p1 (lower average pH and voltage). • Extrapolation to DAC-like p1 = 0.4 mbar and p3 = 1 bar: with TA3 = 0.11 M and ΔTA3→1 = 0.21 M, ΔDICTA−eq,3−1 ≈ 0.049 M, yielding an estimated molar cycle work of 237.4 kJ molCO2−1 at 20 mA cm−2; if initial TA3 = 0 M (same ΔTA), ΔDICTA−eq,3−1 ≈ 0.097 M and work ≈ 121.0 kJ molCO2−1. A Tafel model suggests potential for <100 kJ molCO2−1 at lower current densities. • A subset at p1 = 0.05 bar and 40 mA cm−2 with increased negolyte flow showed average cycle work of 92.6 kJ molCO2−1. • Oxygen impact and mitigation: Under air, Coulombic efficiency dropped to ~65% and capacity was lost by ~20 cycles due to depletion of K4Fe(CN)6 and accumulation of OH−. Electrochemical rebalancing (−40 mA cm−2) restored capacity to near-original values: rebalancing charge passed 476.8 C vs theoretical 473 C; energy cost 378 J (~1.4× one deacidification half-cycle); negolyte pH returned to near-neutral; no evidence of DSPZ degradation by NMR; carbon-capture capability recovered.

The results validate electrochemically induced pH swing using PCET-active DSPZ as a low-energy pathway for CO2 capture and concentration at significantly higher current densities than many prior electrochemical approaches. Achieving 61.3 kJ molCO2−1 for capture from 0.1 bar and release to 1 bar at 20 mA cm−2 demonstrates competitiveness with thermal scrubbing while operating isothermally and at ambient pressure. The independence of ΔDIC on current density (for fixed gas equilibria) decouples throughput from thermodynamics, enabling scale-up by increasing electrode area and gas–liquid contactor performance without sacrificing efficiency. The analysis shows that initial alkalinity (TA3) strongly affects the effective ΔDIC for DAC-like conditions and hence the molar energy; maintaining low TA3 improves capture capacity at low pCO2. The electrochemical rebalancing protocol effectively counters oxygen-induced side reactions that otherwise lead to capacity fade and cell imbalance, broadening practical operating windows, especially for air-exposed systems and DAC. Together, these findings indicate that with improved membranes (lower resistance), redox mediators (lower overpotential, oxygen-insensitive), and engineered contactors (high area, fast mass transfer), electrochemical pH-swing capture can reach even lower energy costs, potentially below 100 kJ molCO2−1 for DAC-like conditions.

This work demonstrates a DSPZ-based electrochemical pH-swing flow cell for CO2 capture with low specific electrical work at practical current densities, including 61.3 kJ molCO2−1 for 0.1→1 bar separation at 20 mA cm−2. Systematic cycles across pCO2 and current densities elucidate how ΔDIC and work scale, enabling extrapolation to DAC-like 0.4 mbar conditions (121–237 kJ molCO2−1 depending on initial alkalinity) and suggesting further reductions at lower current or with improved components. An electrochemical rebalancing method successfully restores electrolyte composition and capacity after oxygen-induced side reactions, providing a general tool for maintaining long-term operation. Future work should: increase mediator concentration and develop oxygen-insensitive PCET molecules; reduce membrane and kinetic overpotentials; implement high-area gas–liquid contactors to boost productivity; and control/monitor TA3 to maximize capture at low pCO2. These advances can improve techno-economic viability of electrochemically driven CO2 capture from point sources and air.

• Direct measurement at DAC-like p1 = 0.4 mbar was not feasible due to equipment sensitivity; results rely on extrapolation using equilibrium models and measured work at higher pCO2. • The laboratory setup used simple gas bubbling with limited interfacial area (11.6–11.8 mL min−1 flow), constraining mass transfer and productivity; engineered contactors are needed for practical rates. • Oxygen sensitivity of reduced DSPZ (DSPZH2) causes side reactions, Coulombic inefficiency, and capacity fade; frequent rebalancing would add energy cost if oxygen ingress is significant. • Energy increases with current density because of ohmic and kinetic overpotentials; membrane resistance and electrocatalysis require optimization. • Maintaining low initial alkalinity (TA3) is necessary at low pCO2; oxygen exposure can raise TA3 over time. • Minor operational issues (foaming, droplet retention) introduced small capacity variations. • Assumptions of gas–solution equilibrium are used in data reduction; deviations at high currents required substituting TA–eq for TA–pH due to pH measurement artifacts.