In situ catalyst activation and regeneration enable energy-efficient high-current CO₂ reduction to ethanol-rich C₂₊ mixtures

Abstract

Electrochemical conversion of dissolved CO₂ in bicarbonate electrolytes, i.e., bicarbonate electrolysis, offers distinct advantages over gas diffusion electrode systems by enabling direct utilization of the CO₂ capture electrolyte while bypassing the energy-intensive CO₂ release step. However, bicarbonate electrolysis faces challenges such as CO₂ mass-transfer limitation, local pH-driven CO₂ depletion, and high cathodic potentials. The higher potential often causes catalyst surface reorganization, leading to a gradual loss of active sites and variations in selectivity during CO₂ reduction. Here, we report a directed, in situ activation and regeneration method that allows precatalysts to equilibrate under dynamic (pulsed) electrolysis conditions. We demonstrate in situ activation of a scalable Cu₂O/Cu mesh that, under short-width (t = 4 s) pulsed electrolysis, provides stable mixed oxidation states of Cu, favoring the formation of an ethanol-rich crude mixture. The pulsed electrolysis waveform, consisting of six distinct segments, is tuned to form Cu⁺ oxides, which are then reduced to generate local alkaline conditions favoring C–C coupling. This synergistic effect results in FEs of 73% for C₂₊ products and 39% for ethanol at an applied current density of −150 mA cm⁻² and a cathodic potential of −1.45 V (vs. RHE). The overall half-cell energy efficiency is ∼30% for C₂₊ products. The in situ Raman experiments confirm the role of pCO₂R in dynamically regenerating Cu⁺-containing surface species during pulsed operation, thereby steering selectivity towards C₂₊ products. A comprehensive multiscale, multiphysics model is developed to investigate the dynamic behavior of copper surface species (Cu, Cu⁺, and Cu²⁺) and local microenvironmental conditions during the pCO₂R. The results reveal that the coexistence of different copper oxidation states, especially the Cu⁺ intermediate, is critical in steering selectivity towards multicarbon (C₂₊) products. The dynamic modulation of surface redox states via tailored pulsing strategies favors C–C coupling pathways by inducing localized alkaline conditions and stabilizing reactive intermediates. This work establishes a predictive modeling platform that links pulse waveform design with mechanistic insights into catalyst state evolution and product selectivity. Overall, this study provides valuable insights into the synergistic effect of in situ activation of pre-catalysts and pulsed electrolysis for higher selectivity towards C₂₊ products.

---

Green foundation

1. The electrochemical CO₂ reduction reaction (eCO₂R) is a promising route to carbon neutrality, converting CO₂ into valuable C₂₊ products such as ethanol and ethylene using Cu-based catalysts. A mixture of eCO₂R products, referred to as electrochemical crude (e-crude), could be a sustainable alternative to traditional crude oil, supporting global initiatives such as the U.S. Department of Energy's Clean Fuels and Products Shot, which prioritizes the development of low-carbon fuels to achieve a net-zero carbon future.

2. The synergistic effect of in situ pre-catalyst activation under pulsed electrolysis results in faradaic efficiencies (FEs) of 72% for C₂₊ products and 39% for ethanol at an applied current density of 150 mA cm⁻² and a cathodic potential of −1.45 V (vs. RHE). Overall, the half-cell energy efficiency is ∼30% for C₂₊ products.

3. Further improvements in FEs and energy consumption can be achieved by identifying an ideal electrolyte with desirable properties, such as CO₂ solubility, higher ionic conductivity, pH control, catalyst complexation, stabilization of desirable intermediates, reduced overpotential, higher current density, higher duty cycle, and long-term stability.

1. Introduction

The electrochemical CO₂ reduction reaction (eCO₂R) is a promising route to carbon neutrality by converting CO₂ into valuable C₂₊ products such as ethanol and ethylene¹ using Cu-based catalysts.² However, a mixture of eCO₂R products, referred to as electrochemical crude (e-crude), is often produced, which could be a sustainable alternative to traditional crude oil and support global initiatives such as the U.S. Department of Energy's Clean Fuels and Products Shot, which prioritizes the development of low-carbon fuels to achieve a net-zero carbon future.³ Ethanol is a particularly desirable component of e-crude due to its market demand, high energy density, and various industrial applications.⁴ However, ethylene is typically the dominant C₂ product⁵ in Cu-based eCO₂Rs due to thermodynamic and kinetic factors, which lead to ethanol selectivity challenges.⁶

The higher selectivity toward both major C₂ products, ethylene and ethanol, on Cu-based catalysts is due to their ability to stabilize the CO intermediate and effectively facilitate C–C coupling.^7–9 The key factors that promote C–C coupling¹⁰ are crystal facets (especially the (100) facet), higher CO coverage, and more alkaline conditions. However, a significant challenge remains in understanding the factors governing the split of reaction pathways leading to these two products.¹¹ A detailed understanding of the reaction pathway is required to control the selective hydrogenation of key intermediate(s) at the onset of pathway divergence.¹² Ethanol and ethylene share a common intermediate (*CHCOH), where ethylene is favored by C–O bond cleavage, while ethanol formation requires C–O bond retention and further hydrogenation.¹³ Ethylene formation is thermodynamically favored on Cu-based catalysts due to lower energy barriers for C–O bond cleavage in the *CHCOH intermediate.¹² Density functional theory (DFT) calculations indicate that the energy required for ethylene formation is approximately 0.2 eV lower than that for ethanol.¹⁴

Several strategies, including catalyst design,^15,16 electrolyte composition,^17–19 oxidation-state tuning,²⁰ and tandem catalysis,²¹ have been explored to shift the selectivity toward ethanol. One practical approach is to reduce the thermodynamic barrier by increasing the local pH,²² which can be achieved at higher current density or with low-buffered electrolytes.^7–9 Nanocavity confinement in porous CuO catalysts was recently shown to increase alkalinity.²³ By tuning the nanocavity size to approximately 12.5 nm, researchers have significantly enhanced ethanol selectivity by increasing the surface coverage of hydroxyl species (*OH). These hydroxyl species play a crucial role in stabilizing the *CHCOH intermediate, shifting its hydrogenation pathway toward *CHCHOH, leading to ethanol formation rather than dehydration to ethylene. The binding energy of *H decreases, and that of *CO increases with increasing pH, reducing the hydrogen evolution reaction (HER) and improving selectivity towards the CO₂R.²³ Another way to control *H coverage is to tune interfacial wettability. By modifying the Cu catalyst surface with alkanethiols of varying chain lengths, researchers have achieved a controlled balance between *CO and *H coverages, directly influencing the ethanol-to-ethylene ratio. Hydrophobic modifications optimize mass transport by regulating CO₂ and H₂O availability at the reaction interface, ensuring a sufficient supply of kinetically controlled *CO and *H.²⁴ This strategy has successfully increased the ethanol-to-ethylene ratio from 0.9 to 1.92, with ethanol faradaic efficiency (FE) reaching 53.7%. Overall, C₂₊ selectivity has been enhanced to 86.1%, although at lower current densities.²⁵

Another critical factor in improving ethanol selectivity is controlling the oxidation states in catalysts. Studies^5,20,24–26 have demonstrated that a balanced coexistence of Cu⁺ and Cu oxidation states stabilizes oxygenated intermediates, making ethanol formation more favorable. Zheng et al.⁷⁴ demonstrated that weakening the Cu–O interaction (with *CHCOH) relative to the O–C interaction could favor the protonation pathway toward ethanol. Catalysts such as Cu₂O-derived materials and Cu-based oxides,^27,28 including CuAl₂O₄,²⁹ have been shown to promote ethanol production by enhancing the catalyst's oxygen affinity. This stabilization effect counteracts the tendency of Cu surfaces to favor ethylene formation, as C–O bond cleavage is thermodynamically preferred.

Recent studies have further highlighted the importance of Cu oxidation-state regulation, interfacial stabilization, and dynamic electrolysis protocols in steering the CO₂RR toward C₂₊ products. For example, operando studies have shown that stabilized Cu²⁺ active sites can promote efficient CO₂-to-C₂H₄ conversion by maintaining neighboring Cu sites that favor C–C coupling intermediates.³⁰ In parallel, oxide-derived Cu catalysts operated under pulsed electrolysis have demonstrated that transient potential “spikes” and dynamic surface reconstruction can enhance C–C coupling.^31,32 Organic-functionalized Cu₂O nanoparticles have also shown that the Cu₂O surface environment and shell thickness can strongly influence C₂⁺ product distribution.^33,34 These advances demonstrate that maintaining favorable Cu oxidation states, stabilizing catalyst interfaces, and dynamically controlling the reaction microenvironment are critical for improving C₂₊ selectivity and durability.

Despite significant progress, challenges remain in retaining an ensemble of active sites and microenvironments for extended periods to maintain catalyst stability. The Maxwell stress and electron wind force play crucial roles in inducing surface heterogeneities that influence catalyst morphology and the distribution of active sites.³⁵ In addition to these effects, the depletion of CO₂ and the increase of local pH cause the microenvironment to vary. Therefore, even in static potential or current experiments, the catalyst structure and microenvironment change continuously, leading to (total) loss of catalyst activity and selectivity over time. There are two promising strategies for addressing these issues. The first approach is to develop robust catalyst mixtures, such as multimetallic systems or alloys, that are resilient to structural drifts induced by electrical forces.³⁶ The second and most commonly used technique is pulsing (or oscillating) the potential or current, where the idea is to reverse catalyst surface reorganization and microenvironment changes by continuously switching the potential or current.³⁷ The pulse duartion and amplitude are critical factors that can be optimized to counteract the negative structural and compositional changes. Since the timescales are different for electrochemical processes such as charging/discharging (∼3 ms), catalyst oxidation/reduction (1–100 ms), and diffusion/migration in the electrolyte (>0.1 s), the pulse duration can be varied to separate redox and diffusional processes. Most pulsed electrolysis methods reported to date have focused on pulses with durations > 10 s, up to minutes, to allow regeneration of depleted CO₂ in the electric double layer.³⁸ We argue that narrower-width pulses are better for preventing irreversible damage to the catalyst composition and providing favorable alkaline conditions to improve ethanol selectivity. For instance, in this article, we pulsed the current from a mildly oxidative region to CO₂-reducing conditions, thereby maintaining a balance between Cu^+/Cu²⁺ and Cu states, essential for ethanol formation. The reduction of Cu²⁺/Cu⁺ to Cu when pulsing from oxidative to reductive conditions also generates OH⁻ and thereby increases local pH in the reaction layer (non-equilibrium layer) without reducing CO₂ concentrations, thereby improving C₂₊ selectivity. In such pulsed electrolysis, the catalyst structure, composition, microenvironment, and CO₂R currents vary dynamically within a narrower bound. Usually, in such cases, a pre-catalyst would evolve over a relatively longer duration to equilibrate with dynamic electrolysis conditions, resulting in substantially different catalyst structure/composition and non-reproducible product distributions.

In this article, a mixed Cu oxide pre-catalyst is developed for the electrochemical conversion of dissolved CO₂ in bicarbonate electrolytes to liquid fuels. We previously reported the advantages of bicarbonate electrolytes and demonstrated their integration with the CO₂ capture process.⁵ To regenerate the catalyst and stabilize the oxide layer on the catalyst, pulsed electrochemical CO₂ reduction (pCO₂R) is employed by carefully controlling parameters such as (i) reduction current (I_R), (ii) oxidation current (I_O), (iii) reduction time (t_R), and (iv) oxidation time (t_O).³⁸ By employing a Cu₂O-coated Cu pre-catalyst, this article, for the very first time, demonstrates the importance of in situ catalyst activation and regeneration to achieve an optimized distribution of Cu, Cu⁺, and Cu²⁺ species.^39,40 This approach allows for high FEs of 73% and 62% for C₂₊ and overall liquid products, respectively (Fig. 1 and Table S2), at a current density of −150 mA cm⁻² and a potential of −1.45 V vs. RHE. The FEs for C₂H₄ and C₂H₅OH are 17% and 39%, respectively. The insights were obtained using in situ Raman spectroscopy and cyclic voltammetry (CV), enabled by a modified electrochemical cell, along with ex situ SEM, EDS, XPS, XRD, and ICP-MS analyses. These results provide strong experimental evidence supporting dynamic surface restructuring of copper as a key driver of enhanced C₂₊ selectivity under pulsed conditions. In parallel, a multiscale, microkinetic model captures the dynamic evolution of copper surface species (Cu, Cu⁺, and Cu²⁺) and local microenvironmental variables, supporting the roles of transient pH regulation and CO₂ concentration gradients. Overall, these findings provide critical insights into the design and operation of Cu-based catalysts for scalable, sustainable CO₂ reduction processes, advancing the feasibility of ethanol-rich e-crude as a next-generation fuel.

Fig. 1 Comparison of the faradaic efficiency (FE) for C₂₊ and ethanol (EtOH), current density (I_R), and voltage (V) of the current work with those of similar systems reported in the literature. The red stars and green triangles represent the FEs for EtOH and C₂₊, respectively. The black squares and blue circles represent I_R and voltage, respectively. The pink area shows the FEs obtained in the present work. The marked references are ref. 41–52.

2. Experimental methods

2.1 Electrode and electrolyte preparation and electrochemical cell operation

The Cu₂O/Cu mesh is prepared using the sol–gel method. The five-step procedure (Fig. S1) for fabricating a Cu₂O film on a Cu mesh is described in section S1.1 of the SI. The details of the experimental set-up are given in section S1.3 of the SI. A 25 : 75 v/v ratio of 0.1 M KHCO₃ and 1 M KCl is used as the catholyte for all experiments. The anolyte consists of 5 M KOH. All the experiments are performed at room temperature. Ni foam was used as the counter electrode, and a leakless Ag/AgCl electrode was used as the reference. The two compartments were separated by a bipolar membrane (Fig. S2). The bipolar membrane offers significant advantages for enhancing C₂₊ selectivity, efficiency, and stability by controlling pH, reducing CO₂ losses,^53,54 and enabling high-performance electrolysis. Before initiating electrochemical measurements, CO₂ was sparged into the catholyte reservoir at 30 SCCM for at least 45 minutes to saturate the catholyte. CO₂ sparging in the electrolyte was maintained throughout the experiment to maintain a high CO₂ concentration.¹¹

2.2 In situ activation of the pre-catalyst

The in situ activation of the Cu₂O/Cu catalyst is achieved by performing chronopotentiometry experiments with oscillating currents at different reduction currents (I_R = −50, −100 and −150 mA cm⁻²). The other parameters, such as I_O = 10 mA cm⁻² and t_R = t_O = 2 s, are kept constant in all these experiments. Each experiment is performed for 45 minutes with a total oxidation time (t_O) and reduction time (t_R) of 22.5 min each. The electrolyte (catholyte and anolyte) is collected after each experiment to quantify liquid products. The electrochemical cell is then cleaned with deionized (DI) water for 15 min to remove any residual liquid products or impurities on the catalyst. After cleaning, the reservoir and tubings are dried with air before introducing 13 mL of CO₂-saturated fresh electrolyte into the reservoir.

2.3 Catalyst characterization

The post-ex situ characterization of different oxidation states of Cu is challenging because Cu undergoes dynamic redox transitions (Cu ↔ Cu⁺ ↔ Cu²⁺) that depend on the potential and electrolyte. After pCO₂R, the catalyst may revert to a state different from that during the reaction. Furthermore, exposing samples to air can lead to a change in the oxidation state from Cu ↔ Cu⁺ ↔ Cu²⁺, introducing artifacts. In addition, the sample preparation steps may wash away critical surface features or intermediates that indicate the actual oxidation state. Therefore, great care was taken during the sample preparation for pre- and post-pCO₂R catalyst characterization. To provide mechanistic insight into pCO₂R, in situ Raman spectroscopy, enabled by a modified electrochemical cell (custom-built at the University of Illinois Chicago), was performed, along with ex situ Scanning Electron Microscopy (SEM), Energy Dispersive X-ray Spectroscopy (EDS), X-ray Diffraction (XRD), X-ray Photoelectron Spectroscopy (XPS), and ICP-MS analyses. These results provide strong experimental evidence supporting dynamic surface restructuring of copper as a key driver of enhanced C₂₊ selectivity under pulsed conditions.

3. Mathematical modeling and simulation

A comprehensive multiscale, multiphysics model is developed to investigate the dynamic behavior of copper surface species (Cu, Cu⁺, and Cu²⁺) and local microenvironmental conditions during the pulsed electrochemical reduction of carbon dioxide (pCO₂R) on copper-based catalysts. This framework integrates experimental measurements of dynamic potential waveforms with microkinetic modeling to predict the temporal evolution of adsorbed copper species (Cu, Cu⁺, and Cu²⁺) and local pH variations at the cathode–electrolyte interface using the continuum model. The model employs experimentally obtained faradaic efficiencies (FEs) for various CO₂ reduction (CO₂R) products to estimate equilibrium constants and rate coefficients for key elementary steps occurring on the catalyst surface under non-steady-state conditions. These kinetic parameters, along with the applied time-varying potential and the resulting cathodic pH, are incorporated into a dynamic microkinetic framework to simulate transient surface-coverage profiles of copper species. The continuum and microkinetic models are explained in greater detail in SI sections S4 and S5, respectively.

4. Results and discussion

While pulsed electrolysis and Cu⁺ rich catalysts have been proposed in section 1, achieving high C₂⁺ faradaic efficiency at high current density and low overpotential remains a challenge. Furthermore, multiple competing hypotheses have been proposed to explain the observed performance enhancements. The primary objective of this work is therefore to deconvolute the individual and synergistic effects of these strategies through a systematic experimental sequence: (a) static CO₂ reduction (sCO₂R) and (b) pulsed CO₂ reduction (pCO₂R) on Cu. The pCO₂R demonstrated enhanced selectivity toward C₂₊ products and revealed the critical role of the Cu⁺ oxidation state in steering product distribution. Based on these findings, we synthesized a Cu₂O-coated Cu (Cu₂O/Cu) catalyst using the sol–gel method. Pulsed operation is markedly more effective for the Cu₂O/Cu catalyst than for metallic Cu. Building on these results, we introduce a novel in situ catalyst activation strategy. This approach, combined with pulsed electrolysis, is used to assess its role in steering selectivity toward C₂₊ products.

4.1 Static vs. pulsed current electrolysis with Cu mesh and Cu₂O/Cu mesh catalysts

4.1.1 Static current electrolysis on the Cu mesh catalyst. We first investigated the performance of a pure Cu mesh by performing static reduction current (I_R) experiments over the range of −50 < I_R < −200 mA cm⁻². The sCO₂R experiments were performed in a 3-electrode flow cell, with a slightly acidic pH electrolyte (0.75 M KCl and 0.025 M KHCO₃) for the catholyte and 5 M KOH for the anolyte. The selection of the catholyte is based on our previous work.⁵ However, the anolyte is selected by performing systematic studies over a wide range of KOH concentrations (0–5 M). Fig. S5 shows a clear dependence of product selectivity and cell energetics on electrolyte concentration. Specifically, the FE for C₂H₄ and ethanol increases with increasing KOH concentration, reaching a maximum at 5 M KOH, while the FE for the HER decreases correspondingly. This trend is attributed to enhanced ionic conductivity at higher concentrations. Additionally, the potential drop across the working electrode remains nearly unchanged over this concentration range, whereas that across the counter electrode decreases significantly with increasing KOH concentration, indicating reduced ohmic losses and improved overall cell efficiency. Based on this combined improvement in selectivity and reduced energy loss, 5 M KOH was selected as the optimal anolyte for the reported experiments.

The objective here is to benchmark the catalyst (Cu mesh) activity for CO₂RR against previously reported literature data. Fig. S6 shows the presence of four gaseous (CO, CH₄, C₂H₄, and H₂) and three liquid (HCOOH, C₂H₅OH, and C₃H₇OH) products. The experimental results are in close agreement with those reported in the literature.^25,26,55,56 The experimental results reveal that for static reduction current, the HER dominates at all currents. The FE towards C₂ products (C₂H₄ and C₂H₅OH) increases with an increase in I_R and reaches maxima at I_R = −150 mA cm⁻². As I_R increases further, the FE towards C₂₊ products decreases from 37% to 25%, whereas that for the HER increases from 36% to 47%. This behavior can be attributed to inherent issues related to bicarbonate electrolysis, as shown in Fig. 2. (1) Local pH control: the eCO₂R of dissolved CO₂ leads to an increase in local pH near the catalyst/electrolyte interface. The elevated pH shifts the equilibrium towards HCO₃⁻ or CO₃²⁻, hence reducing the availability of dissolved CO₂. (2) High overpotential: the CO₂ desorption from HCO₃⁻ requires additional energy (0.21 V vs. SHE), leading to a high thermodynamic potential, which leads to larger overpotentials during electrolysis. Hence, bicarbonate electrolysis needs a higher potential compared to the gas diffusion electrode (GDE) system; (3) Mass transfer limitation: the combined effect of eCO₂R of dissolved CO₂ and HCO₃⁻ shifts the equilibrium towards higher pH and hence reduces the CO₂ diffusion flux towards the catalyst interface; (4) Reduced copper: in the presence of continuous eCO₂R, the copper oxides (Cu⁺ and Cu²⁺) converted into reduced copper (Cu). Cu⁺ oxides have been shown to promote C₂₊ product formation and suppress side reactions, such as the HER. Hence, the absence of these copper oxides reduces selectivity towards C₂₊ products; (5) Higher HCO₃⁻ concentration: elevated pH promotes the reverse reactions, CO₂ reacts with OH⁻ to form bicarbonate (HCO₃⁻), which reduces the CO₂ availability and also increases the overpotential for eCO₂R; (6) Higher CO₃²⁻ concentration: HCO₃⁻ further reacts with OH⁻ to produce carbonate (CO₃²⁻) and water, reducing the dissolved CO₂ available for conversion. At high current densities, the local pH can reach 12, exacerbating these challenges; (7) Promotion of the hydrogen evolution reaction (HER): the elevated concentrations of HCO₃⁻ facilitate the competing HER via the Volmer–Heyrovsky mechanism.⁵⁷ In this pathway, HCO₃⁻ donates a proton, which reacts with an adsorbed hydrogen atom to form molecular hydrogen (H₂). This competition reduces the efficiency of eCO₂R; (8) Catalyst degradation: the catalyst undergoes dynamic structural and morphological changes under continuous eCO₂R. It may reduce the catalyst lifetime and lower the FE for the target C₂₊ products.

Fig. 2 Issues related to aqueous bicarbonate electrolysis of CO₂ to multi-carbon products.

4.1.2 Pulsed electrolysis on the Cu catalyst. To improve the selectivity and stability of catalysts towards C₂ products, a methodology for catalyst regeneration in controlled microenvironments is adopted by applying pulsating currents. A square wave pulse (Fig. 3a) of alternate reduction (I_R) and oxidation current (I_O) is applied for equal duration (t_R = t_O = 2 s). Fig. 3b and Fig. S7 demonstrate that the HER is reduced by 70% in the presence of an oxidation pulse. In contrast, the FE towards C₂ products (C₂H₄ and C₂H₅OH) increases from 34% to 38%, whereas the FE for C₁ products (CO, CH₄, and HCOOH) increases from 21% to 40%. However, HCOOH is replaced by CH₄. This enhancement in C₂ selectivity during pulsed electrolysis can be attributed to multiple hypotheses reported in the published literature, such as (i) dynamic restructuring of the catalyst surface due to a change in the oxidation state of copper;⁵⁸ (ii) dynamic local pH due to the abundance of OH⁻ ions during the reduction phase and their consumption during the oxidation phase;^40,59 (iii) enhanced CO₂ concentration gradients overcome mass transfer limitations at higher currents;⁶⁰ (iv) oxidative destruction of the accumulated products on the catalyst surface prevents catalyst degradation;⁶¹ (v) unstable double layer;⁶² and (vi) stabilization of intermediates for enhanced C₂₊ products.⁶³

Fig. 3 (a) Schematic of static vs. pulsed CO₂R; (b) comparison of the faradaic efficiencies (FEs) for multi-carbon products for Cu (static vs. pulsed) and Cu₂O/Cu (pulsed) catalysts; in situ cyclic voltammetry (CV) analysis for pre and post CO₂R of (c) the Cu catalyst with (d) an enlarged version to show the change in oxidation peaks and (e) the Cu₂O/Cu catalyst with (f) an enlarged version to show the change in oxidation peaks.

To deconvolute which hypothesis plays an essential role in steering the selectivity towards C₂₊ products, in situ cyclic voltammetry (CV) is performed before (pre) and after (post) pulsed electrolysis is complete. The scanning potential range is −1.5 ≤ V ≤ 0 at a scan rate of 10 mV s⁻¹. The comparison between in situ CV curves (Fig. 3c) obtained for pre and post-pulsating experiments confirms the validity of hypothesis 1, which states that catalysts undergo dynamic structural modifications due to alternate reduction and oxidation cycles. The post-CV curve exhibits a distinct Cu → Cu²⁺ peak at −0.2 V (vs. Ag/AgCl), indicating the oxidation of the catalyst during pCO₂R. The expanded CV curve in Fig. 3d also exhibits a small peak associated with Cu → Cu⁺ transition at −0.5 V (vs. Ag/AgCl). Together, these peaks indicate the formation of mixed copper oxidation states (Cu/Cu⁺/Cu²⁺) during pulsed electrolysis. The selectivity of C₁ and C₂ products is highly dependent upon the nature of these oxides. For instance, Cu⁺/Cu combination leads to higher selectivity towards C₂₊ products. However, the presence of Cu²⁺/Cu species leads to higher FEs towards C₁ products.^64,65 Therefore, the selection of I_O (for chronopotentiometry experiments) or V_O (for chronoamperometry experiments) is very critical for steering the selectivity towards C₁ or C₂ products. The bottom half of CV curve shows a small peak at −0.85 V (vs. Ag/AgCl), indicating Cu⁺ → Cu. After that, the current increases continuously with increasing reduction potential, highlighting CO₂R. These oxidation (Cu → Cu⁺ → Cu²⁺) and reduction (Cu²⁺ → Cu⁺ → Cu) peaks are not very evident in pre-CV curves compared to post-CV curves, which indicates the absence of oxide species. Since oxide is known to reduce electrical conductivity,⁶⁶ a higher reduction current in the pre-CV curve further confirms the presence of only copper metal in its pure form (Cu).

To evaluate the applicability of Hypothesis 1 (dynamic restructuring of catalyst surface), Hypothesis 2 (dynamic local pH), and Hypothesis 3 (enhanced CO₂ concentration gradients), a comprehensive multiscale, multiphysics model is developed as explained in Section 3. The model helps to understand the dynamic behavior of copper surface species, local pH, and CO₂ concentration.

The continuum model predicts the dynamic variations in local pH and concentration of CO₂, as well as buffer species (HCO₃⁻ and CO₃²⁻), as shown in Fig. 4a–c, for static CO₂ reduction (sCO₂R), and Fig. 4d–f, for pulsed CO₂ reduction (pCO₂R). Under sCO₂R, continuous charge transfer to the catalyst results in sustained generation of OH⁻ anions, rapid increase in surface pH, and the formation of a highly alkaline microkinetic environment. Simultaneously, the local CO₂ concentration decreases due to its constant consumption. The combination of high local pH and low CO₂ concentration shifts the buffer equilibrium toward CO₃²⁻, reducing the availability of CO₂ at the catalyst–electrolyte interface and consequently lowering the faradaic efficiency (FE) toward C₂₊ products. This observation aligns with previous reports indicating that excessive alkalinity near the catalyst surface leads to CO₂ depletion and promotes the formation of carbonate species, thereby suppressing C₂₊ product selectivity.^59,67 In contrast, the alternating oxidation and reduction phases in pCO₂R prevent the accumulation of OH⁻ anions by consuming them during the oxidation period. During each cycle, this dynamic control of pH results in a highly acidic state (pH ∼ 3.5) and a highly basic state (pH ∼ 9.3), demonstrating that periodic current modulation can regulate local pH and sustain CO₂ availability, thereby enhancing C₂₊ product formation. An interesting pH buffering effect is observed between 4 s < t < 6 s. During the initial phase of oxidation (4 s < t < 4.2 s), the decrease in pH is gradual due to the buffering capacity of HCO₃⁻ in the electrolyte, followed by a sharp drop to pH = 3.5 as the oxidation phase progresses. Such a delayed pH shift has been reported in systems utilizing bicarbonate-based electrolytes, where the buffer effect mitigates rapid pH fluctuations and influences product selectivity.⁶⁸ Beyond preventing excessive local pH increase, pCO₂R also provides a recovery period for CO₂, allowing it to replenish to its equilibrium concentration of 34 mM, as observed in previous modeling studies.⁵⁹ This synergistic effect of dynamic pH control and CO₂ concentration maintenance enhances C₂₊ product selectivity while preventing a shift in equilibrium toward CO₃²⁻, thereby ensuring a consistent CO₂ supply throughout the reduction process.

Fig. 4 Comparison of static (a–c) vs. pulsed (d–f) CO₂R using the coupled multiphase-microkinetic model for (a and d) dynamic variation in pH; (b and e) dynamic variation in the concentration of CO₂; (c and f) dynamic variation in the concentration of KHCO₃ and CO₃²⁻; (g) dynamic variation in fractional coverage of Cu, Cu⁺, and Cu²⁺ during pCO₂R; and (h) dynamic variation in equilibrium potential and applied potential during pCO₂R.

The fractional coverage of copper species (Cu, Cu⁺, and Cu²⁺) predicted by the microkinetic model in Fig. 4g confirms that the catalyst undergoes dynamic structural modifications during alternating reduction and oxidation cycles. Initially, the Cu⁺ species is dominant because the Cu₂O/Cu mesh serves as a catalyst. During the reduction cycle, the Cu⁺ species is depleted due to the conversion of Cu⁺ → Cu. However, the Cu⁺ regenerates to some extent during the next oxidation cycle. This behavior can be explained by the voltage curve shown in Fig. 4h. During the oxidation period, the applied voltage is in the range of E₀₁ < V < E₀₂. Therefore, the probability of Cu → Cu⁺ is higher than that of Cu⁺ → Cu²⁺. The presence of mixed copper species (Cu and Cu⁺) leads to higher FE of C₂₊ products.

4.1.3 Pulsed electrolysis on the Cu₂O/Cu mesh. The previous section shows that Cu⁺ oxides play a pivotal role in enhancing selectivity toward C₂⁺ products (e.g., ethylene and ethanol) during CO₂R. This preference stems from their unique electronic structure, surface properties, and interaction with reaction intermediates. Therefore, it is desirable to increase the Cu⁺/Cu ratio. This ratio depends on the oxidation current (I_O) and duration (t_O). The small I_O during a short t_O allows for fast oxidation of the Cu metal to its oxide. Higher I_O and/or longer t_O led to complete dissolution of the catalyst. Therefore, pulsed electrolysis imposes a limit on the maximum Cu⁺/Cu. Therefore, it was desirable desirable to find an alternative route to enhance Cu⁺/Cu. A sol–gel method was used to incorporate Cu₂O coating on the Cu mesh. The synthesis method is explained in greater detail in section S1.2 of the SI. Fig. 3b shows that the novel catalyst (Cu₂O/Cu) has shown 41% enhancement in the FE towards C₂₊ products (C₂H₄ and C₂H₅OH) compared to the Cu mesh. On the other hand, the FE towards C₁ products (CO, CH₄, and HCOOH) reduced by 55%. In order to optimize the process parameters, the effect of oxidation current (I_O) is studied by keeping other parameters fixed. Fig. S8 clearly demonstrates that the HER increases significantly beyond I_O > 10 mA. Also, the FE towards C₂ products decreases for I_O > 10 mA. Hence, I_O > 10 mA is used for all the remaining experiments.

To get insights into the role of Cu₂O/Cu in steering selectivity towards C₂₊ products, in situ cyclic voltammetry (CV) was performed before and after the pulsed electrolysis is complete. The scanning potential range is −1.5 ≤ V ≤ 0 at a scan rate of 10 mV s⁻¹. Fig. 3e shows that the pre-pCO₂R CV curve shows a very large peak near −0.8 V, which is a signature for Cu⁺ → Cu transition. It clearly indicates the presence of Cu₂O species obtained after the sol–gel method. The transition leads to very active Cu species, resulting in very high current and, hence, steering selectivity towards C₂₊ products C₂H₄ and C₂H₅OH.

4.2 In situ activation of the pre-catalyst (Cu₂O/Cu mesh)

The motivation for introducing this activation procedure stems from the limitations of conventional ex situ catalyst pretreatment methods such as ultrasonication and electropolishing. Although these techniques can modify the catalyst surface prior to electrolysis, they are performed outside the electrochemical reactor and therefore do not reproduce the actual reaction environment experienced during CO₂ reduction. In contrast, the proposed activation strategy conditions the catalyst directly inside the flow cell under realistic electrochemical, chemical, and hydrodynamic conditions.

4.2.1 Pulsed electrolysis on the Cu₂O/Cu mesh with in situ activation. In the present work, we have developed a novel in situ catalyst activation technique for pre-treating the catalyst within a flow cell. In conventional pCO₂R experiments, the freshly prepared Cu₂O/Cu catalyst is directly subjected to target pulsed electrolysis conditions (I_R = −150 mA cm⁻², I_O = 10 mA cm⁻², and t_R = t_O = 2 s). In contrast, A-pCO₂R (Activated-pCO₂R) incorporates an additional in situ catalyst activation step (Fig. 5a) before the catalyst is evaluated at the target operating current density. Specifically, the catalyst is first exposed to pCO₂R at a low reduction current density and subsequently conditioned through a series of progressively increasing reduction current densities with an optimum step size of ΔI_R until the desired operating current density is reached. After completion of this activation sequence, the catalyst is tested under the desired pulsed electrolysis conditions used for conventional pCO₂R experiments.

Fig. 5 (a) Schematic of pulsed electrolysis for in situ pre-catalyst activation of the Cu₂O/Cu catalyst; (b and c) effect of varying reduction current density (I_R) on the FEs for multi-carbon products for reduction current step sizes (ΔI_R) of 25 and 50 mA cm⁻²; (d) effect of varying reduction current step size (ΔI_R) on (d) the FEs for multicarbon products and (e) the dynamic variation in applied potential.

The gradual increase in reduction current density during A-pCO₂R provides several advantages. First, it enables controlled surface reconstruction and facilitates the formation of catalytically active Cu⁺/Cu interfaces without inducing abrupt reduction of surface oxide species. The periodic anodic pulses continuously regenerate Cu⁺ species, while the progressively increasing cathodic current allows the catalyst to evolve toward a stable Cu⁺-rich active state. Second, gradual current ramping minimizes sudden changes in local pH, CO₂ concentration, and intermediate coverage, thereby preventing severe local CO₂ depletion and excessive HER during activation. Third, the activation sequence promotes controlled accumulation of adsorbed CO intermediates and the progressive development of active sites favorable for C–C coupling and ethanol formation. Finally, the repeated oxidation–reduction cycles occurring during activation generate a catalyst surface that is already adapted to the local reaction microenvironment before the catalyst is exposed to the final operating conditions.

Therefore, although the catalyst composition (Cu₂O/Cu) and the final pulsed electrolysis conditions are identical for both pCO₂R and A-pCO₂R, the catalyst surface state at the beginning of the performance test is fundamentally different. A-pCO₂R utilizes a catalyst that has been pre-conditioned and activated in situ under pulsed electrolysis conditions, whereas pCO₂R employs the freshly prepared catalyst without this activation sequence. As a result, during pCO₂R, the freshly prepared catalyst is immediately exposed to the target current density, resulting in rapid surface reduction and less controlled catalyst evolution.

The first set of experiments is performed by gradually increasing the I_R at a step size (ΔI_R) of 25 mA cm⁻². The other operating parameters, such as I_O, t_O and t_R, are kept constant throughout the experiment. The experiment at each reduction current is performed for 45 min. After that, the electrolyte is collected for quantification of liquid products using the NMR technique. Fig. 5b shows that the FEs towards C₁, C₂, and C₃ products enhance with I_R reach maxima at I_R = −150 mA cm⁻². The FE for C₂₊ products was 60%, comprising 29% C₂H₅OH and 25% C₂H₄. On the other hand, the FEs for C₁ and H₂ were 12% and 20%, respectively. For I_R > −150 mA cm⁻², the FE for C₂₊ products reduces significantly. The enhancement in the FE towards C₂₊ products at I_R > −150 mA cm⁻²can be attributed to the in situ directed pre-catalyst evolution of the Cu₂O/Cu catalyst inside the flow cell.

The next step is to study the effect of increasing the current step size (ΔI_R) and its relationship with the selectivity towards C₂₊ products. Fig. 5c shows that the increase in the step size (ΔI_R) to 50 mA cm⁻² leads to similar behavior to that at I_R = 150 mA cm⁻². The FE towards C₂₊ products reaches maxima at I_R = −150 mA cm⁻². It is enhanced by 7%, whereas that for the HER is reduced by 52%. The maximum FEs for C₂H₅OH and C₂H₄ were 39% and 17%, respectively. With a further increase in ΔI_R, the FE toward C₂₊ products decreases, as shown in Fig. 5d and Fig. S9. Analysis of the dynamic potential profiles in Fig. 5e reveals that increasing ΔI_R shortens the duration of regime 2, while extending regime 3. Additionally, the oxidation potential (V_O) increases with increasing ΔI_R to maintain a constant oxidation current density (I_O = 10 mA cm⁻²). CV analysis indicates that higher V_O promotes the transition from Cu⁺ → Cu²⁺, which is detrimental to C₂₊ selectivity. Therefore, the optimal performance at ΔI_R = 50 mA cm⁻² results from a balance between favorable Cu⁺ formation and controlled oxidation potential.

4.2.2 Dynamics of applied potential, electric double layer (EDL), and local pH. It is well known that the charging and discharging of the electrical double layer (EDL) plays a significant role in pCO₂R, particularly by influencing the local environment at the electrode–electrolyte interface.^40,60,63 In this section, the dynamics of applied potential and its relationship with the EDL and local pH are investigated. Fig. 6a shows the transient variation in applied potential for −50 ≤ I_R ≤ −300 mA cm⁻² at a data sampling rate of 0.1 s. In all these experiments I_O = 10 mA cm⁻² and t_O = t_R = 2 s. Each pulse, which consists of one cycle of oxidation and reduction, can be categorized into six different regimes, as shown in Fig. 6b.

Fig. 6 (a) Dynamic variation of applied potential for varying reduction currents during A-pCO₂R; (b) classification of different regimes during A-pCO₂R for the Cu₂O/Cu catalyst.

Regime 1 is labeled as an oxidative non-faradaic regime as it indicates the onset of the oxidation phase of the pulse, and there is no net charge transfer or chemical reactions occurring during this phase. Instead, these processes contribute to phenomena such as charging of the electrical double layer (EDL), adsorption/desorption of species, and ion migration within the electrolyte. There is a sharp reduction in applied potential within a very short time (<0.2 s).

Regime 2 delays the rate of reduction in applied potential and highlights the change in the oxidation state of Cu. Accordingly, this regime is known as the oxidative faradaic regime. This regime begins at −0.8 V (vs. Ag/AgCl), which coincides with the CV results and indicates the Cu → Cu⁺ transition. After that, the Cu⁺ → Cu²⁺ transition occurs at −0.2 V. The duration of this regime (0.8 s) is longer than that of regime 1. During this regime, electrons are removed from the cathode surface or from species near the cathode/electrolyte interface, leading to a net positive charge near the cathode. The positively charged electrode surface attracts negatively charged ions (anions) from the electrolyte to the inner Helmholtz layer, while cations are repelled. As a result, the EDL becomes more compact because the stronger electric field pulls anions closer to the electrode surface.

Regime 3 is characterized by an almost steady potential. This regime is called an oxidative equilibrium phase, in which the oxide species are in complete equilibrium. The oxide species participate in chemical reactions with hydroxide ions (OH⁻) and protons (H⁺). Hence, the local pH near the electrode/electrolyte interface is lowered. Furthermore, copper oxide species also play an important role in stabilizing intermediates crucial for CO dimerization and in steering selectivity towards C₂₊ products. This regime persists for a longer period (>1.2 s), and its duration decreases with increasing I_R.

Regime 4 is known as a reductive non-faradaic regime because there is no chemical reaction or charge transfer during this regime. The applied potential rises rapidly (<0.2 s). The entire energy input from the applied potential is consumed in charging the EDL.

Regime 5 is very similar to regime 2; it lasts for a very short time (<0.2 s) and is known as a reductive faradaic regime. The transfer of electrons near the electrode/electrolyte interface leads to the reduction of copper oxides to copper metal (Cu²⁺ → Cu⁺ → Cu). The accumulation of negative charges near the electrode attracts positively charged ions (cations) from the electrolyte into the Helmholtz layer, while anions are repelled. Furthermore, it leads to the expansion of the EDL, as cations are less tightly held than anions due to their larger hydration shells.

Regime 6 is characterized by an almost steady potential and is named the CO₂R regime because the electrochemical reduction of CO₂ occurs during this phase. Also, the HER occurs during this phase. The CO₂R and HER generate OH⁻ ions, thereby increasing the local pH near the electrode/electrolyte interface. This regime persists for a longer period (>1.6 s), and is independent of I_R.

4.2.3 Effect of varying pulse duration. This section focuses on the effect of pulse duration (0.5 s ≤ t_R = t_O ≤ 3 s) on the dynamic potential response and product selectivity. The results in Fig. S14a reveal that increasing the pulse duration primarily extends the durations of regime 3 (oxidative equilibrium regime) and regime 6 (CO₂R regime), while the durations of the non-faradaic (regimes 1 and 4) and faradaic transition regimes (regimes 2 and 5) remain nearly unchanged. The transition from regime 2 to regime 3 marks the completion of the rapid oxidation process and the establishment of a quasi-equilibrium oxide layer on the Cu surface. While regime 2 is associated with the electrochemical formation of Cu⁺ and Cu²⁺ species, the subsequent regime 3 provides sufficient time for oxide redistribution, restructuring, and stabilization. Therefore, the oxidative pulse duration (t_O) directly influences the residence time of the catalyst in this equilibrium state and consequently affects the thickness and stoichiometry of the regenerated oxide layer. At short pulse durations (t_O < 1 s), regime 3 is relatively brief, limiting the extent of oxide stabilization and yielding a thinner, less stable CuO_x layer. Under these conditions, C₁ products (CH₄ and HCOOH) and the HER are favored, indicating insufficient surface restructuring and limited availability of Cu⁺ species required for efficient C–C coupling.

As t_O increases, regime 3 allows the oxide layer to mature and stabilize, promoting the persistence of Cu⁺ species, which have been widely reported to enhance *CO adsorption and facilitate CO dimerization. Consequently, Fig. S14b shows that the selectivity shifts from C₁ products and the HER toward C₂₊ products. In addition to extending regime 3, increasing the pulse duration also prolongs regime 6, the CO₂R regime where the majority of CO₂ electroreduction occurs. The extended residence time in regime 6 provides a longer period for the accumulation of key reaction intermediates (*CO) on the catalyst surface, thereby increasing the probability of C–C coupling reactions. Furthermore, the continuous generation of OH⁻ in this regime elevates the local pH near the catalyst surface, thereby suppressing the competing hydrogen evolution reaction (HER) and favoring pathways to multicarbon products. Therefore, the combined effect of a stabilized CuO_x-derived surface generated during regime 3 and an extended CO₂R environment during regime 6 creates favorable conditions for enhanced C₂₊ formation.

At shorter pulse durations, the limited duration of both regime 3 and regime 6 prevents sufficient oxide stabilization and restricts the buildup of surface-bound CO intermediates, resulting in higher selectivity toward C₁ products and the HER. In contrast, increasing the pulse duration enhances both surface conditioning and reaction time for CO dimerization, thereby increasing C₂₊ selectivity.

However, for pulse durations exceeding 2 s, a decrease in C₂₊ selectivity is observed. Also, a gradual increase in the HER is detected, indicating the onset of CO₂ mass-transport limitations near the catalyst surface. Although longer pulse durations extend both regime 3 and regime 6, the prolonged CO₂ reduction period can deplete the local CO₂ concentration faster than it can be replenished from the bulk electrolyte. Consequently, the surface coverage of carbon-containing intermediates becomes limited, while proton- and water-reduction pathways become increasingly competitive. As a result, the HER is promoted at the expense of CO₂ reduction. These observations suggest that pulse durations of approximately 2 s provide an optimal balance between oxide-layer stabilization, accumulation of C–C coupling intermediates, and local CO₂ availability. Beyond this duration, the benefits of longer oxidative equilibrium and CO₂ reduction regimes are offset by local CO₂ depletion, leading to a plateau in C₂₊ selectivity and a concomitant increase in the HER.

4.2.4 Performance of the Cu₂O/Cu mesh w and w/o pre-catalyst activation. Section 4.2.1 highlights the importance of pre-catalyst activation and reduction current step size (ΔI_R) in directing the selectivity towards C₂⁺ products. The results indicate that the maximum FE towards C₂₊ products is obtained at ΔI_R = 50 and I_R = −150 mA cm⁻². Since each experiment (at I_R = −50 mA cm⁻² and −100 mA cm⁻²) was performed for 45 minutes, a total of 90 minutes of pre-catalyst activation time is needed to activate the catalyst before running it for optimum I_R = −150 mA cm⁻². To estimate the reduction in the FE for C₂₊ products by avoiding pre-catalyst activation, pCO₂R is performed for I_R = −150 mA cm⁻². The comparison between the activated pulsed electrochemical CO₂ reduction (A-pCO₂R) and direct pCO₂R in Fig. 7a and Fig. S11 clearly reveals a 35% enhancement in the FE towards C₂H₅OH due to the activation. However, the FE for the HER decreases by 59%. However, the FE for C₂H₄ remains almost unchanged. Insights into this behavior can be obtained by investigating the dynamics of applied potential. Fig. 7b shows that in the absence of pre-catalyst activation, the slope is steep in regime 2. As a result, the duration of regime 3 is extended compared to activated A-pCO₂R. Even the oxidation potential (V_O) is higher for achieving the same I_O = 10 mA cm⁻². As we know from the CV analysis, a higher V_O leads to Cu⁺ → Cu²⁺ transition, which reduces the selectivity towards C₂₊ products. In addition, a higher V_R results in higher energy consumption than A-pCO₂R. The total duration for A-pCO₂R is 135 min (45 × 3), whereas it is only 45 min for pCO₂R without any pre-catalyst activation. To deconvolute the role of total duration vs. reduction current step size (ΔI_R), pCO₂R is studied by performing three experiments for 45 min each, keeping all other parameters fixed (I_R = −150 mA cm⁻²; I_O = 10 mA cm⁻²; and t_R = t_O = 2 s). Fig. S13 clearly demonstrates that the FE toward C₂H₅OH remains stable at 28–29%. Similarly, that for the HER remains consistent at 20–21%. Therefore, it is evident that the step size (ΔI_R) rather than the total duration of pCO₂R plays an important role in steering selectivity towards C₂₊ products. Overall, the comparison between static, pulsed, and pre-catalyst activation CO₂R in Fig. S15 shows the clear advantage of A-pCO₂ for the Cu₂O/Cu catalyst compared to the only static or pulsed CO₂R.

Fig. 7 Comparison of A-pCO₂R and pCO₂R for the Cu₂O/Cu catalyst with regard to (a) FE for multi-carbon products and (b) dynamic variation of applied potential.

4.3 Energy efficiency of bicarbonate electrolysis

Bicarbonate electrolysis is inherently an integrated CO₂ capture and conversion process in which CO₂ is captured as bicarbonate and then converted. Since the reduction of bicarbonate is substantially uphill energetically, it must be converted to CO₂ (e.g., via a pH swing) for further reduction to CO₂RR products. The energy efficiency of electrolysis is a product of FE and voltage efficiency (VE). VE is a ratio of equilibrium potential to cell potential. However, the VE for bicarbonate electrolysis must account for the thermodynamic energy required to convert bicarbonate to CO₂. This requirement is roughly 59 mV per unit pH change to convert bicarbonate to CO₂. In addition to this pH swing, the Nernstian potential losses due to the CO₂ saturation ratio relative to its depleted state should be considered. Since these potential requirements are specific to operating current, we have considered a total minimum potential requirement of 0.515 V to release CO₂. This value is an estimate based on previously reported data of CO₂ capture potential (V).⁶⁹ We now define the pseudo-energy efficiency for a half-cell, which includes the thermodynamic penalty for CO₂ release during bicarbonate electrolysis.

Fig. 8 shows the pseudo-energy efficiency of different bicarbonate electrolysis systems reported to date. Notably, in situ catalyst activation and regeneration enable the half-cell pseudo-energy efficiency for C₂₊ products to reach a maximum of 30% at a current density of −150 mA cm⁻².

Fig. 8 Comparison of the pseudo energy efficiency at different reduction current densities. The references are ref. 41–52.

The half-cell pseudo-energy efficiency did not explicitly account for the BPM-associated voltage loss. Therefore, additional measurements have been performed using a modified four-electrode cell configuration (Fig. S16) that enables direct quantification of the potential drop across the BPM during operation. Specifically, an additional reference electrode was incorporated into the flow cell, allowing the membrane potential drop to be determined independently from the cathode potential. The BPM voltage drop was quantified using two complementary approaches: (i) by comparing the measured cathode potentials obtained from conventional three-electrode and four-electrode configurations, and (ii) by directly measuring the potential difference between two reference electrodes in otherwise identical cells operated with and without the BPM. The measurements reveal that the voltage drop across the BPM increases with current density and reaches approximately 1.3 V vs. RHE at a reduction current density of −150 mA cm⁻². This additional voltage loss is attributed to the combined effects of ionic transport resistance within the membrane and the overpotential associated with water dissociation at the bipolar junction. The observed values are consistent with those reported for BPM-based CO₂ electrolyzers operating at similar current densities. After incorporating the BPM potential drop into the energy-efficiency calculation, the half-cell energy efficiency decreases from approximately 30% to 16% under pulsed operation at 150 mA cm⁻². This reduction reflects the energetic cost associated with maintaining water dissociation and ion transport across the BPM.

While the BPM imposes an additional energy penalty, it simultaneously enables the establishment of distinct pH environments at the cathode and anode, suppresses CO₂ crossover, and maintains a favorable local reaction environment for C₂₊ product formation. Therefore, the observed efficiency loss represents a trade-off between energy consumption and the enhanced selectivity and carbon utilization afforded by BPM operation.

4.4 Insights into pCO₂R on the Cu₂O/Cu catalyst

To elucidate the mechanistic differences between static CO₂ reduction (sCO₂R) and pulsed CO₂ reduction (pCO₂R) on Cu₂O/Cu catalysts, we employed in situ Raman spectroscopy using a custom-designed electrochemical flow cell (Fig. S4), complemented by ex situ SEM, EDS, XPS, XRD, and ICP-MS analyses. Together, these results provide compelling evidence that dynamic surface restructuring of copper under pCO₂R operation plays a critical role in enhancing C₂₊ selectivity.

The in situ Raman measurements were first conducted under dry conditions (no electrolyte) at open circuit potential (OCP). As shown in Fig. 9a, four characteristic peaks at ∼149, 219, 419, and 630 cm⁻¹ are observed, consistent with the reported^70–73 first-order and second-order vibrational modes of Cu₂O. Upon introduction of electrolyte flow (30 mL min⁻¹), a decrease in the intensity of the 149 cm⁻¹ peak and a relative increase in the 219 cm⁻¹ peak are observed, which can be attributed to refractive index changes and/or subtle surface restructuring. Under static reduction (I_R = −50 mA cm⁻²), the Cu₂O-related peaks diminish, indicating progressive reduction of Cu⁺ species to metallic Cu (Fig. 9b and c). Increasing the I_R to −100 mA cm⁻² further suppresses these features (Fig. 9b and d), although weak Cu⁺ persists. At I_R = −150 mA cm⁻², the Cu⁺ peaks are no longer detectable, confirming near-complete reduction to Cu and depletion of Cu⁺-rich surface species. In contrast, under pulsed electrolysis conditions (I_R = −150 mA cm⁻²; I_O = 10 mA cm⁻²; and t_R = t_O = 6 s), the Raman spectra reveal periodic disappearance and reappearance of Cu⁺ bands during reduction (cathodic) and oxidation (anodic) phases, respectively. This behavior indicates rapid and reversible redox cycling between Cu and Cu⁺ on a timescale of a few seconds. Although the temporal resolution of Raman acquisition (∼6 s per spectrum, which consists of 5 s of exposure, 1 s of acquisition, and 1 s of processing) limits direct observation at shorter pulse durations (t_R = t_O = 2 s), the results clearly demonstrate that pCO₂R sustains a dynamically regenerated Cu⁺/Cu surface population. Since the Raman acquisition time (6 s) is longer than the individual reduction or oxidation pulse width (2 s), the collected spectra represent a time-averaged response over multiple pulse segments rather than a fully time-resolved snapshot of the instantaneous catalyst surface state.

Fig. 9 Time-resolved Raman spectra recorded through the optical window while focusing on the Cu₂O/Cu catalyst surface of the electrochemical cell during chronopotentiometry (CP) experiments in CO₂-saturated 0.75 M KCl and 0.025 M KHCO₃ solution. The experiments are performed under different conditions: (a) dry cell (no electrolyte at the open circuit potential), wet cell (cell is filled with the electrolyte without any forced flow), and flow cell (the electrolyte flows at 30 mL min⁻¹); (b) transient Raman spectrum (single scan snapshots) for static CO₂R at varying reduction current densities (I_R) of −50, −100 and −150 mA cm⁻²; pulsed CO₂R with I_R = −150 mA cm⁻²; I_O = 10 mA cm⁻², t_R = t_O = 6 s; transient peak intensities at (c) I_R = −50 mA cm⁻² and (d) I_R = −100 mA cm⁻² for wavenumbers 149, 219, and 630 cm⁻¹.

Ex situ characterization further corroborates these findings. SEM images (Fig. S17) reveal increased surface roughness after pCO₂R compared to that after sCO₂R, suggesting enhanced morphological evolution under pulsed conditions. EDS elemental mapping (Table S5) shows a relative decrease in oxygen content after sCO₂R and a higher oxygen content after pCO₂R, consistent with partial reoxidation during oxidation (anodic) pulses.

XPS analysis (Fig. S18B), calibrated using reference Cu and Cu₂O samples, indicates a strong Cu₂O signature prior to electrolysis, in agreement with cyclic voltammetry results. After pCO₂R, the Cu₂O contribution diminishes, and the metallic Cu signal becomes dominant, indicating an overall Cu⁺ → Cu transition. Similarly, XRD patterns (Fig. S18A) confirm the structural evolution from Cu₂O to Cu during electrolysis.

To assess the role of dissolution–redeposition processes, ICP-MS was used to quantify copper ions in the electrolyte. Under reduction-only conditions, the Cu concentration remains minimal (∼0.5 ppm), indicating negligible dissolution. In contrast, pulsed electrolysis results in measurable Cu dissolution (∼5 ppm), attributable to transient oxidation during the anodic pulse. This effect is enhanced in the presence of chloride ions (0.75 M KCl, 0.025 M KHCO₃), which stabilize soluble Cu–Cl complexes. Importantly, the absence of significant Cu dissolution during the cathodic phase suggests that redeposition-driven restructuring is not the dominant mechanism governing catalytic behavior.

Overall, these results demonstrate that pCO₂R promotes a dynamic Cu⁺/Cu redox environment, coupled with surface restructuring and controlled oxidation events, which together contribute to sustained C₂₊ selectivity and improved catalytic performance relative to steady-state operation.

4.5 Long-term stability test of static, pulsed, and activated pulsed CO₂ electrolysis

To evaluate the effectiveness of the proposed in situ catalyst activation strategy in enhancing catalyst durability and maintaining C₂₊ selectivity, long-term stability tests were performed for (i) static CO₂ reduction (sCO₂R), (ii) pulsed CO₂ reduction (pCO₂R), and (iii) activated pulsed CO₂ reduction (A-pCO₂R) for 48 h, as shown in Fig. 10. All experiments were conducted under practical operating conditions with continuous electrolyte recirculation to mimic realistic electrolyzer operation.

Fig. 10 Long-term stability tests for (a) static reduction (sCO₂R); (b) pulsed CO₂R (pCO₂R); and (c) activated pulsed CO₂R (A-pCO₂R).

A comparison of Fig. 10a–c highlights the limitations of both sCO₂R and conventional pCO₂R with respect to C₂₊ selectivity and long-term operational stability. Under static electrolysis, the catalyst progressively evolves toward a metallic Cu surface, resulting in dominant HER activity and relatively low C₂₊ selectivity throughout the experiment. Although pCO₂R initially improves C₂₊ formation through periodic regeneration of Cu⁺ species, the product distribution gradually shifts during extended operation. Specifically, the FE toward C₂₊ products continuously decreases, while the FE toward CH₄ and H₂ increases, indicating progressive deterioration of the catalyst surface and local reaction environment.

In contrast, A-pCO₂R exhibits both superior C₂₊ selectivity and enhanced long-term stability. During the initial stage of electrolysis, the FE for ethanol remained relatively stable at 34–38%, with high overall C₂₊ selectivity. The improved performance of A-pCO₂R is attributed to the in situ activation protocol, which gradually conditions the catalyst under pulsed electrolysis and promotes the formation of a stable Cu⁺/Cu interfacial structure adapted to the local reaction microenvironment. This activated catalyst state facilitates sustained CO adsorption, enhanced C–C coupling, and HER suppression compared with both sCO₂R and pCO₂R.

Despite the improved stability, a gradual decline in C₂₊ selectivity was observed during extended operation. After approximately 23 h of electrolysis, the FE toward C₂₊ products decreased to ∼35%, accompanied by a corresponding increase in HCOOH and H₂ formation. Several factors may contribute to this performance decay, including catalyst restructuring, gradual loss of active Cu⁺ species, accumulation of liquid products in the catholyte, and salt deposition within the catalyst layer and the porous electrode structure. The accumulation of liquid products can further reduce CO₂ solubility and hinder mass transport, thereby promoting competing reaction pathways. The ex situ SEM/EDS analysis in Fig. S19 shows noticeable changes in surface composition after a 48-hour test. Specifically, the Cu content decreased from 90.3 wt% to 83.2 wt%, while the O content increased from 9.97 wt% to 15.1 wt%. The increased oxygen content suggests that part of the Cu₂O surface may have been further oxidized to CuO during prolonged electrolysis. The XPS results (Fig. S19) confirm changes in the Cu oxidation state after extended operation. This surface oxidation could modify the active Cu/Cu₂O interface and reduce the availability of active sites favorable for C–C coupling and C₂₊ product formation, thereby contributing to the observed decline in selectivity.

Taken together, the SEM, EDS, and XPS results indicate that the performance degradation during the 48-hour stability test is likely due to the combined effects of salt accumulation, partial blockage of active sites, surface oxidation, and morphological/compositional changes in the Cu₂O/Cu catalyst.

To isolate the contribution of salt accumulation from other degradation mechanisms, the flow cell underwent an in situ cleaning procedure after 23 h of operation. Deionized water was circulated through the electrolyzer at a high flow rate (150 mL min⁻¹), followed by replacement with fresh electrolyte. Remarkably, the FE toward C₂₊ products was substantially restored after the cleaning step, demonstrating that salt accumulation is a major contributor to the observed loss of selectivity. The recovery of performance after salt removal indicates that catalyst deactivation is not entirely irreversible and that the catalyst retains a significant fraction of its intrinsic activity following extended operation.

These results demonstrate that A-pCO₂R significantly outperforms both sCO₂R and pCO₂R in terms of maintaining high C₂₊ selectivity during prolonged electrolysis. The combination of gradual in situ catalyst activation and periodic Cu⁺ regeneration effectively delays catalyst degradation and preserves the favorable reaction microenvironment required for multicarbon product formation. However, the results also reveal that salt accumulation remains a critical challenge for long-term continuous operation, particularly when substantial quantities of liquid products are generated.

From a practical perspective, these findings suggest that periodic electrolyte management or intermittent cleaning protocols may be necessary for stable long-term operation of industrial CO₂ electrolyzers. In addition, future optimization of the pulse waveform, such as using higher oxidation currents or longer oxidation pulses, may further mitigate salt accumulation and enhance catalyst regeneration. However, such approaches must be carefully balanced against the risk of catalyst dissolution or excessive oxidation of the Cu surface. Therefore, developing pulsed protocols that simultaneously maximize catalyst regeneration while minimizing catalyst degradation represents an important direction for future research.

5. Conclusion

An in situ catalyst activation strategy is developed here that pre-treats copper-based catalysts under operating microenvironments, overcoming critical limitations of conventional ex situ methods. By initiating partial CO₂ reduction at low current densities and progressively increasing the reduction current, combined with pulsed electrolysis, we stabilize Cu₂O species and suppress catalyst degradation. This approach enables faradaic efficiencies of 73% for C₂₊ products, 62% for liquid products, and a pseudo-energy efficiency of 30% for C₂₊ products at 150 mA cm⁻² and −1.45 V vs. RHE, with ethanol and ethylene selectivities of 39% and 20%, respectively. Comprehensive in situ Raman spectroscopy, cyclic voltammetry (CV), and ex situ structural characterization techniques (SEM, EDS, XPS, and XRD), along with continuum and microkinetic modeling, reveal that microenvironment engineering during catalyst activation and regeneration plays a decisive role in steering product selectivity. Dynamic potential modulation during pulsed electrolysis further reveals critical regimes associated with electric double-layer dynamics and local pH control. Our findings establish a generalizable framework for in situ catalyst activation, dynamic regeneration of Cu⁺/Cu species, and microenvironment modulation in electrochemical CO₂ conversion. This work provides a roadmap for designing durable, highly selective CO₂ reduction systems, advancing the commercial viability of electrochemical CO₂ valorization. Future efforts will focus on investigating the influence of halide ions (Cl⁻, Br⁻, and I⁻) and in situ generated liquid products on the catalyst surface state, and on optimizing duty cycles (t_R/(t_R + t_O)) to maximize C₂₊ yield and operational stability.

Author contributions

Nitin Minocha: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), software (equal), validation (lead), visualization (lead), writing – original draft (lead), and writing – review & editing (lead); Yancun Qi: conceptualization (support), data curation (equal), validation (equal), writing – original draft (support), formal analysis (support), and writing – review & editing (support); Ahmed A. Farghaly: resources (support), data curation (support), formal analysis (support), methodology (support), and validation (support); Haitham M. El-Bery: conceptualization (support), data curation (support), validation (support), writing – original draft (support), and methodology (equal); Rohan Sartape: conceptualization (support) and writing – original draft (support); Ayush Karwa: data curation (support) and validation (support); Prem K. R. Podupu: software (support); Alexey Izgorodin: conceptualization (support), resources (lead), formal analysis (lead), investigation (support), methodology (lead), writing – review & editing (support), and project administration (support); Husain Naji: conceptualization (support), resources (lead), formal analysis (lead), investigation (support), methodology (lead), writing – review & editing (support), and project administration (support); Aqil Jamal: conceptualization (support), resources (lead), formal analysis (lead), investigation (support), methodology (lead), writing – review & editing (support), and project administration (support); Meenesh R. Singh: project administration (lead), software (lead), supervision (lead), conceptualization (support), data curation (support), and writing – review & editing (equal).

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6gc02368a.

Acknowledgements

The work was performed in the Materials & Systems Engineering Laboratory at the University of Illinois, Chicago. M. R. S. acknowledges funding support from Saudi Aramco and the U.S. National Science Foundation – ECO-CBET program (award no. 2420733). Funding for the development of the in situ Raman capability was provided by the U.S. Department of Energy (DOE), Energy Efficiency and Renewable Energy, Hydrogen and Fuel Cell Technologies Office (HFTO) under the auspices of the Electrocatalysis Consortium (ElectroCat 2.0). Argonne is managed for the U.S. Department of Energy by the University of Chicago Argonne, LLC, under Contract DE-AC-02-06CH11357.

The work performed at the Center for Nanoscale Materials, a U.S. Department of Energy Office of Science User Facility, was supported by the U.S. DOE, Office of Basic Energy Sciences, under Contract No. DE-AC-02-06CH11357.

N. M. acknowledges the use of ChatGPT (OpenAI, version GPT-5.5) for assistance with language editing and rephrasing to improve the clarity and readability of this manuscript. Following the use of this tool, the authors reviewed and edited the content as needed and take full responsibility for the publication's final content.

References

1. R. Chauhan, R. Sartape, N. Minocha, I. Goyal and M. R. Singh, Energy Fuels, 2023, 37, 12589–12622.
2. S. Fankhauser, S. M. Smith, M. Allen, K. Axelsson, T. Hale, C. Hepburn, J. M. Kendall, R. Khosla, J. Lezaun and E. Mitchell-Larson, Nat. Clim. Change, 2022, 12, 15–21.
3. S. Zhang, Q. Fan, R. Xia and T. J. Meyer, Acc. Chem. Res., 2020, 53, 255–264.
4. M.-Y. Lee, K. T. Park, W. Lee, H. Lim, Y. Kwon and S. Kang, Crit. Rev. Environ. Sci. Technol., 2020, 50, 769–815.
5. A. Prajapati, N. C. Kani, J. A. Gauthier, R. Sartape, J. Xie, I. Bessa, M. T. Galante, S. L. Leung, M. H. S. Andrade, R. T. Somich, M. V. Rebouças, G. T. Hutras, N. Diniz and M. R. Singh, Cell Rep. Phys. Sci., 2022, 3, 101053.
6. Z. Wang, Y. Li, Z. Ma, D. Wang and X. Ren, iScience, 2024, 27, 110437.
7. J. Qiao, Y. Liu, F. Hong and J. Zhang, Chem. Soc. Rev., 2014, 43, 631–675.
8. D. Gao, R. M. Arán-Ais, H. S. Jeon and B. Roldan Cuenya, Nat. Catal., 2019, 2, 198–210.
9. C.-T. Dinh, T. Burdyny, M. G. Kibria, A. Seifitokaldani, C. M. Gabardo, F. P. García de Arquer, A. Kiani, J. P. Edwards, P. De Luna and O. S. Bushuyev, Science, 2018, 360, 783–787.
10. J. Y. Kim, W. Park, C. Choi, G. Kim, K. M. Cho, J. Lim, S. J. Kim, A. Al-Saggaf, I. Gereige and H. Lee, ACS Catal., 2021, 11, 5658–5665.
11. Z. Zhang, E. W. Lees, F. Habibzadeh, D. A. Salvatore, S. Ren, G. L. Simpson, D. G. Wheeler, A. Liu and C. P. Berlinguette, Energy Environ. Sci., 2022, 15, 705–713.
12. Y. Y. Birdja, E. Pérez-Gallent, M. C. Figueiredo, A. J. Göttle, F. Calle-Vallejo and M. T. Koper, Nat. Energy, 2019, 4, 732–745.
13. K. Zheng, Y. Li, B. Liu, F. Jiang, Y. Xu and X. Liu, Angew. Chem., Int. Ed., 2022, 61, e202210991.
14. Y. Wang, M. Zheng, X. Wang and X. Zhou, Catalysts, 2022, 12, 1617.
15. A. Prajapati, R. Sartape, N. C. Kani, J. A. Gauthier and M. R. Singh, ACS Catal., 2022, 12, 14321–14329.
16. L. Bian, Z.-Y. Zhang, H. Tian, N.-N. Tian, Z. Ma and Z.-L. Wang, Chin. J. Catal., 2023, 54, 199–211.
17. M. R. Singh, Y. Kwon, Y. Lum, J. W. Ager 3rd and A. T. Bell, J. Am. Chem. Soc., 2016, 138, 13006–13012.
18. A. Prajapati, R. Sartape, M. T. Galante, J. Xie, S. L. Leung, I. Bessa, M. H. S. Andrade, R. T. Somich, M. V. Rebouças, G. T. Hutras, N. Diniz and M. R. Singh, Energy Environ. Sci., 2022, 15, 5105–5117.
19. H. Tian, J.-T. Yang, X. Wang, H. Jiao, Z.-F. Gao, K.-Y. Zhu, Q. He and Z.-L. Wang, Appl. Catal., B, 2025, 375.
20. Y. Wu, H. Zhu, W. Wang, J. Shi, Y. Sun, F. Bai, F. Yu, Y. Wu and Y. Chen, J. Energy Chem., 2025, 107, 416–426.
21. L. Bian, Y. Bai, J. Y. Chen, H. K. Guo, S. Liu, H. Tian, N. Tian and Z. L. Wang, ACS Nano, 2025, 19, 9304–9316.
22. J. C. Bui, C. Kim, A. J. King, O. Romiluyi, A. Kusoglu, A. Z. Weber and A. T. Bell, Acc. Chem. Res., 2022, 55, 484–494.
23. A. S. Varela, Curr. Opin. Green Sustainable Chem., 2020, 26, 100371.
24. L. Guo and S. Guo, Int. J. Hydrogen Energy, 2021, 46, 33120–33131.
25. L. Cheng, R. Wang, W. Si, Y. Deng, J. Li and Y. Peng, ACS Catal., 2024, 14, 10324–10333.
26. R. Cai, M. Sun, F. Yang, M. Ju, Y. Chen, M. D. Gu, B. Huang and S. Yang, Chem, 2024, 10, 211–233.
27. X. D. Shi, R. T. Guo, H. F. Cui, C. Liu and W. G. Pan, Nanoscale, 2025, 17, 4338–4364.
28. P. Grosse, A. Yoon, C. Rettenmaier, A. Herzog, S. W. Chee and B. Roldan Cuenya, Nat. Commun., 2021, 12, 6736.
29. T. Zhang, B. Yuan, W. Wang, J. He and X. Xiang, Angew. Chem., 2023, 135, e202302096.
30. Z. Zhang, Q. Xu, J. Han, K. Ren, Y. Hu, R. Zhao, H. Sun, J.-S. Qin and H. Rao, Nat. Commun., 2026, 17, 2654.
31. C. Thammaniphit, J. Santatiwongchai, S. Impeng, P. Khemthong, K. Meesombad, K. Faungnawakij, T. Hanrath, R. Methaapanon and P. Chakthranont, ACS Catal., 2025, 15, 17003–17014.
32. Q. Chen, Y. Tan, X. Wang, Q. Wang, H. Li, K. Liu, J. Fu, L. Chai, M. Liu and Z. Lin, Nat. Commun., 2025, 16, 8468.
33. J. Hu, S. Osella, J. Albero and H. García, Adv. Funct. Mater., 2024, 34, 2404566.
34. H. Liu, Y. Yu, Y. Bai, Y. Yang, Y. Wang, W. Li, L. Li, J. Hao and W. Shi, Nat. Commun., 2025, 16, 6186.
35. H. Liu, T. Yan, S. Tan, L. Sun, Z. Zhang, S. Hu, S.-H. Li, X. Kang, Y. Lei and L. Jiang, J. Am. Chem. Soc., 2024, 146, 5333–5342.
36. L. Liu, H. Akhoundzadeh, M. Li and H. Huang, Small Methods, 2023, 7, 2300482.
37. Y. Jännsch, J. J. Leung, M. Hämmerle, E. Magori, K. Wiesner-Fleischer, E. Simon, M. Fleischer and R. Moos, Electrochem. Commun., 2020, 121, 106861.
38. R. Kas, R. Kortlever, A. Milbrat, M. T. Koper, G. Mul and J. Baltrusaitis, Phys. Chem. Chem. Phys., 2014, 16, 12194–12201.
39. J. Yano and S. Yamasaki, J. Appl. Electrochem., 2008, 38, 1721–1726.
40. K. W. Kimura, R. Casebolt, J. Cimada DaSilva, E. Kauffman, J. Kim, T. A. Dunbar, C. J. Pollock, J. Suntivich and T. Hanrath, ACS Catal., 2020, 10, 8632–8639.
41. G. Lee, A. S. Rasouli, B.-H. Lee, J. Zhang, D. H. Won, Y. C. Xiao, J. P. Edwards, M. G. Lee, E. D. Jung, F. Arabyarmohammadi, H. Liu, I. Grigioni, J. Abed, T. Alkayyali, S. Liu, K. Xie, R. K. Miao, S. Park, R. Dorakhan, Y. Zhao, C. P. O'Brien, Z. Chen, D. Sinton and E. Sargent, Joule, 2023, 7, 1277–1288.
42. H. Lu, G. Wang, Y. Zhou, A. S. Wotango, J. Wu, Q. Meng and P. Li, Int. J. Mol. Sci., 2022, 23, 9373.
43. H. Wang, X. Bi, Y. Yan, Y. Zhao, Z. Yang, H. Ning and M. Wu, Adv. Funct. Mater., 2023, 33, 2214946.
44. B. Yang, L. Chen, S. Xue, H. Sun, K. Feng, Y. Chen, X. Zhang, L. Xiao, Y. Qin, J. Zhong, Z. Deng, Y. Jiao and Y. Peng, Nat. Commun., 2022, 13, 5122.
45. Z. Yang, D. Ji, Z. Li, Z. He, Y. Hu, J. Yin, Y. Hou, P. Xi and C. H. Yan, Small, 2023, 19, e2303099.
46. H. Xu, D. Rebollar, H. He, L. Chong, Y. Liu, C. Liu, C.-J. Sun, T. Li, J. V. Muntean, R. E. Winans, D.-J. Liu and T. Xu, Nat. Energy, 2020, 5, 623–632.
47. P. Iyengar, M. J. Kolb, J. Pankhurst, F. Calle-Vallejo and R. Buonsanti, ACS Catal., 2021, 11, 13330–13336.
48. S. Kuang, M. Li, R. Xia, L. Xing, Y. Su, Q. Fan, J. Liu, E. J. M. Hensen, X. Ma and S. Zhang, ACS Appl. Nano Mater., 2020, 3, 8328–8334.
49. M. Wang, H. Chen, M. Wang, J. Wang, Y. Tuo, W. Li, S. Zhou, L. Kong, G. Liu, L. Jiang and G. Wang, Angew. Chem., Int. Ed., 2023, 62, e202306456.
50. W. Sun, P. Wang, Y. Jiang, Z. Jiang, R. Long, Z. Chen, P. Song, T. Sheng, Z. Wu and Y. Xiong, Adv. Mater., 2022, 34, e2207691.
51. X. Su, Z. Jiang, J. Zhou, H. Liu, D. Zhou, H. Shang, X. Ni, Z. Peng, F. Yang, W. Chen, Z. Qi, D. Wang and Y. Wang, Nat. Commun., 2022, 13, 1322.
52. S. C. Zignani and A. S. Aricò, Mater. Renewable Sustainable Energy, 2024, 13, 233–241.
53. M. Hesselmann, J. K. Lee, S. Chae, A. Tricker, R. G. Keller, M. Wessling, J. Su, D. Kushner, A. Z. Weber and X. Peng, ACS Appl. Mater. Interfaces, 2024, 16, 24649–24659.
54. M. Lin, L. Han, M. R. Singh and C. Xiang, ACS Appl. Energy Mater., 2019, 2, 5843–5850.
55. J. Y. Kim, G. Kim, H. Won, I. Gereige, W. B. Jung and H. T. Jung, Adv. Mater., 2022, 34, 2106028.
56. J. Kok, J. de Ruiter, W. van der Stam and T. Burdyny, J. Am. Chem. Soc., 2024, 146, 19509–19520.
57. R. Kronberg, H. Lappalainen and K. Laasonen, Phys. Chem. Chem. Phys., 2020, 22, 10536–10549.
58. T. Yan, X. Chen, L. Kumari, J. Lin, M. Li, Q. Fan, H. Chi, T. J. Meyer, S. Zhang and X. Ma, Chem. Rev., 2023, 123, 10530–10583.
59. J. C. Bui, C. Kim, A. Z. Weber and A. T. Bell, ACS Energy Lett., 2021, 1181–1188,  DOI:10.1021/acsenergylett.1c00364.
60. K. Ye, T.-W. Jiang, H. D. Jung, P. Shen, S. M. Jang, Z. Weng, S. Back, W.-B. Cai and K. Jiang, Nat. Commun., 2024, 15, 9781.
61. G. Marcandalli, M. C. Monteiro, A. Goyal and M. T. Koper, Acc. Chem. Res., 2022, 55, 1900–1911.
62. Z. Tang, E. Nishiwaki, K. E. Fritz, T. Hanrath and J. Suntivich, ACS Appl. Mater. Interfaces, 2021, 13, 14050–14055.
63. C. Kim, L.-C. Weng and A. T. Bell, ACS Catal., 2020, 10, 12403–12413.
64. R. M. Arán-Ais, F. Scholten, S. Kunze, R. Rizo and B. Roldan Cuenya, Nat. Energy, 2020, 5, 317–325.
65. J. Timoshenko, A. Bergmann, C. Rettenmaier, A. Herzog, R. M. Arán-Ais, H. S. Jeon, F. T. Haase, U. Hejral, P. Grosse, S. Kühl, E. M. Davis, J. Tian, O. Magnussen and B. Roldan Cuenya, Nat. Catal., 2022, 5, 259–267.
66. P. Ctibor, L. Straka, F. Lukáč, J. Sedláček and I. Králová, Bol. Soc. Esp. Ceram. Vidrio, 2024, 63, 413–424.
67. F. P. García de Arquer, C.-T. Dinh, A. Ozden, J. Wicks, C. McCallum, A. R. Kirmani, D.-H. Nam, C. Gabardo, A. Seifitokaldani, X. Wang, Y. C. Li, F. Li, J. Edwards, L. J. Richter, S. J. Thorpe, D. Sinton and E. H. Sargent, Science, 2020, 367, 661–666.
68. M. R. Singh, E. L. Clark and A. T. Bell, Phys. Chem. Chem. Phys., 2015, 17, 18924–18936.
69. A. Prajapati and M. R. Singh, ACS Sustainable Chem. Eng., 2019, 7, 5993–6003.
70. H. Li, P. Wei, D. Gao and G. Wang, Curr. Opin. Green Sustainable Chem., 2022, 34, 100589.
71. S. Mu, H. Lu, Q. Wu, L. Li, R. Zhao, C. Long and C. Cui, Nat. Commun., 2022, 13, 3694.
72. X. Chen, D. A. Henckel, U. O. Nwabara, Y. Li, A. I. Frenkel, T. T. Fister, P. J. A. Kenis and A. A. Gewirth, ACS Catal., 2019, 10, 672–682.
73. H. An, L. Wu, L. D. B. Mandemaker, S. Yang, J. de Ruiter, J. H. J. Wijten, J. C. L. Janssens, T. Hartman, W. van der Stam and B. M. Weckhuysen, Angew. Chem., Int. Ed., 2021, 60, 16576–16584.
74. Z. Liu, L. Song, X. Lv, M. Liu, Q. Wen, L. Qian, H. Wang, M. Wang, Q. Han and G. Zheng, J. Am. Chem. Soc., 2024, 146, 14260–14266.

---