CO₂/carbonate-mediated electrochemical water oxidation to hydrogen peroxide

Abstract

Electrochemical water oxidation reaction (WOR) to hydrogen peroxide (H₂O₂) via a 2e⁻ pathway provides a sustainable H₂O₂ synthetic route, but is challenged by the traditional 4e⁻ counterpart of oxygen evolution. Here we report a CO₂/carbonate mediation approach to steering the WOR pathway from 4e⁻ to 2e⁻. Using fluorine-doped tin oxide electrode in carbonate solutions, we achieved high H₂O₂ selectivity of up to 87%, and delivered unprecedented H₂O₂ partial currents of up to 1.3 A cm⁻², which represents orders of magnitude improvement compared to literature. Molecular dynamics simulations, coupled with electron paramagnetic resonance and isotope labeling experiments, suggested that carbonate mediates the WOR pathway to H₂O₂ through the formation of carbonate radical and percarbonate intermediates. The high selectivity, industrial-relevant activity, and good durability open up practical opportunities for delocalized H₂O₂ production.

Introduction

Electrochemical water (H₂O) oxidation to hydrogen peroxide (H₂O₂) via a 2e⁻ pathway (2e⁻-WOR) represents a green and sustainable route to produce H₂O₂ compared to traditional anthraquinone process, but is currently challenged by low selectivity and activity due to strong competition from the typical 4e⁻ oxygen evolution reaction pathway (OER or 4e⁻-WOR)^1,2,3,4,5,6. Traditional approaches to promoting 2e⁻-WOR to H₂O₂ have been mostly focused on exploring catalysts with relatively weak binding strength with intermediate O species compared to that in the 4e⁻ counterpart^7,8,9,10,11. These catalysts (typically made of inert metal oxides^7,8,9,10 as well as other materials^12,13,14) requires large overpotentials to activate the water oxidation step, but their 2e⁻-WOR current densities are usually limited at ~10 to 200 mA cm⁻², as the extra overpotentials to drive larger currents would start to push the water oxidation reaction all the way down to O₂ with significantly decreased H₂O₂ selectivity^{7,8,9,10,11,15,16}. As a result, the state-of-the-art 2e⁻-WOR performances are still far below the requirements in practical applications^7,17.

Reaction redox mediators as “electron shuttles” have been playing important roles in facilitating desired reaction pathways (especially for intermediate products) in electrocatalysis^18,19, electroorganic synthesis^20,21,22 and bioelectrocatalysis²³, and could become a solution to the H₂O₂ activity-selectivity dilemma. While exploring such kind of reaction mediators for selective 2e⁻-WOR is challenging, our nature may have an answer for us. H₂O₂ commonly exists in biosystems as one type of reactive oxygen species (ROS)²⁴ which are essential in signal transduction^25,26. However, researchers have found out that high concentrations of CO₂ could cause ROS burst (rapid release of ROS from cells), leading to destructions of redox-sensitive proteins and cell structures (Supplementary Note 1)^27,28.

Inspired by this phenomenon and previous studies in electrogenerated chemiluminescence^29,30 and peroxycarbonate synthesis³¹, here we hypothesize that CO₂ (or carbonate as its ionic form in water) may serve as an effective mediator to promote the H₂O₂ pathway in electrochemical water oxidation (Fig. 1a)^32,33,34. Using fluorine doped tin oxide (FTO) as a model catalyst electrode, we demonstrated high H₂O₂ selectivity up to 87%, delivered high H₂O₂ partial current densities up to 1.3 A cm⁻², and achieved a long-term stable and continuous H₂O₂ generation for 250 hours with over 80% H₂O₂ selectivity at 150 mA cm⁻² current density in carbonate solutions. The electrochemical performance of our work represents orders of magnitude improvement compared to previous works. We studied the mechanism of the carbonate mediations effects using molecular dynamics simulations, coupled with electron paramagnetic resonance and isotope labeling experiments. Simulation and experimental results suggested that carbonate mediates the WOR pathway through carbonate radical and percarbonate intermediates.

Fig. 1: CO₂/Carbonate promotion effects in 2e⁻-WOR.

a Schematic illustration of CO₂/carbonate promoted electrochemical H₂O oxidation to H₂O₂. With the mediator effects of CO₂/carbonate, the WOR reaction pathway could change from 4e⁻ towards O₂ to 2e⁻ towards H₂O₂. b H₂O₂ selectivity in Ar-saturated and CO₂-saturated 1.0 M sodium phosphate buffer solution. CO₂ serves as a promoter for 2e⁻-WOR, as the H₂O₂ selectivity increased by orders of magnitude compared to that without CO₂. c–e I–V curves, H₂O₂ FEs, and H₂O₂ partial current densities in 1.0 M NaHCO₃ and 1.0 M Na₂CO₃, respectively. The maximum H₂O₂ FE in 1.0 M NaHCO₃ was 34% with a H₂O₂ partial current density of 5.6 mA cm⁻². In comparison, the maximum H₂O₂ FE in 1.0 M Na₂CO₃ was 56%, and the maximum H₂O₂ partial current density was 68 mA cm⁻².

Results and discussion

Verification of the CO₂/carbonate mediation effects

To support the CO₂/carbonate mediator strategy, an oxidation catalyst electrode which meets the following criteria is a prerequisite: first, it should be an inert catalyst for the 4e⁻ oxygen evolution reaction (OER); second, it has a high electrical conductivity to deliver high currents; third, it should remain stable under high oxidation potentials in water. After an initial screening process, FTO (Supplementary Fig. 1) was selected as the catalytic electrode compare to other materials (Supplementary Fig. 2). While FTO has also been used in some previous studies of photoelectrochemical or electrochemical water oxidation reaction, in most cases it was used as the substrate for studying other catalytic materials^7,8. Additionally, the H₂O₂ selectivity and activity on FTO reported before are rather low⁸, and the surface reaction mechanism of 2e⁻-WOR is still unclear yet. We evaluated the electrochemical WOR performance using a standard three-electrode setup in an H-type cell. Sodium phosphate buffer (0.65 M Na₂HPO₄ and 0.35 M NaH₂PO₄, pH ~7) was chosen as the aqueous electrolyte, due to its high stability under oxidative potentials³⁵ and strong buffering capability³⁶, for investigating the possible impacts of CO₂ on WOR pathways (Fig. 1b and Supplementary Fig. 3). In Ar-saturated 1.0 M sodium phosphate buffer, only trace amount of H₂O₂ was detected within a wide range of applied water oxidation potentials, with H₂O₂ Faradaic efficiencies (FEs) less than 1% (Fig. 1b). This result suggests that FTO presents an intrinsic selectivity towards the 4e⁻-WOR pathway. Surprisingly, when the solution was saturated with CO₂, a significant jump of H₂O₂ selectivity was achieved under the same reaction conditions (Fig. 1b), supporting our hypothesis that CO₂ can play a role in steering the WOR reaction pathway towards H₂O₂.

As CO₂ exists in different types of species in aqueous solutions³⁷, including dissolved CO₂, carbonate, and bicarbonate, we therefore designed control experiments to better identify the key factors that are at play in promoting H₂O₂ generation. We first used 1.0 M NaHCO₃ as the electrolyte, where bicarbonate is the dominant species, to evaluate FTO’s 2e⁻-WOR performance. As shown in Fig. 1c–e and Supplementary Fig. 4, the peak H₂O₂ FE was ~ 34%, corresponding to a H₂O₂ partial current density of only 5.6 mA cm⁻². We further switched the electrolyte to carbonate dominated 1.0 M Na₂CO₃ solution and performed the same test, and observed a drastic change. The H₂O₂ FE jumped up to a maximal of 56% at 3.2 V versus reversible hydrogen electrode (vs. RHE) with a significantly improved H₂O₂ partial current density of over 50 mA cm⁻² (Fig. 1e), representing over one order of magnitude increase compared to that in either CO₂ saturated electrolyte or bicarbonate electrolyte. To explore if this promotion effect on 2e⁻-WOR pathway is an intrinsic catalytic property of the FTO catalyst, or specifically exists in the CO₂/bicarbonate/carbonate systems, we also tested the H₂O₂ selectivity in other commonly used electrolytes containing different anion species, including sodium sulfate, sodium nitrate, sodium hydroxide, sodium perchloride, as well as the sodium phosphate buffer we showed earlier (Fig. 2a–d and Supplementary Fig. 5). To further confirm the carbonate promotion effects, we then tested the electrochemical water oxidation performance in Na₂CO₃ + NaOH electrolyte (Supplementary Fig. 6a). In alkaline electrolytes, OH⁻ absorption plays a key role in oxygen evolution reaction³⁸. With the increase of NaOH concentration, OH⁻ absorption will decrease the coverage ratio of carbonate absorption on FTO surface, and thus leading to decreased H₂O₂ selectivity and increased OER, which further demonstrates that high pH is not the reason for high H₂O₂ selectivity when compared to bicarbonate solutions. As a result, none of them presented any preferences for 2e⁻-WOR pathway, with negligible H₂O₂ selectivity of less than 2%, which indicates the participation of CO₂-related species in the electrochemical H₂O-to-H₂O₂ conversion process on the FTO surface.

Fig. 2: Impacts of anion species and anion/cation concentrations on 2e⁻-WOR.

a–d I–V curves and H₂O₂ FEs of FTO catalyst using 1.0 M Na₂SO₄, 1.0 M NaNO₃, 1.0 M NaOH, and 1.0 M NaClO₄, respectively. The H₂O₂ FEs in these electrolytes were lower than 2%, indicating there are no promotion effects for H₂O₂ formation using these anion species. e Dependence of H₂O₂ and O₂ selectivity on Na₂CO₃ concentration at 100 mA cm⁻². The H₂O₂ FE was increased with increased Na₂CO₃ concentration, while the FE of side product O₂ was correspondingly decreased, indicating that Na₂CO₃ could be directly involved in the 2e⁻-WOR process. f H₂O₂ FE in 1.0 M Na₂CO₃ with different concentrations of NaClO₄.

Since in carbonate solution the FTO catalyst exhibited the best H₂O₂ generation performance, we thus tested the dependence of H₂O₂ selectivity over Na₂CO₃ concentration to further reveal the promotion effects of carbonate on 2e⁻-WOR. As shown in Fig. 2e, the H₂O₂ FE under 100 mA cm⁻² current presented a monotonic enhancement from 5% to 56% with increased Na₂CO₃ concentration from 0.1 M to 2.0 M, respectively, indicating the critical role of carbonate ions in promoting H₂O₂ selectivity. Possible promotion effects from sodium ions were excluded as the H₂O₂ FEs were similar under different Na⁺ concentrations (Fig. 2 f). We also examined the O₂ FEs from the 4e⁻-WOR pathway using gas chromatography quantification, which added together with H₂O₂ are close to 100%, suggesting no significant side reactions in this electrochemical system (Fig. 2e, see Methods). In addition, FTO electrodes with different thicknesses, fluorine doping levels, or surface resistivity exhibited quite similar H₂O₂ selectivity (Supplementary Fig. 6b), further confirming that the dominant factor on WOR pathway is from carbonate concentrations. Other types of common conducting metal oxides, including alumina-doped zinc oxide and indium tin oxide, were also evaluated in carbonate electrolyte but showed poor stability under the WOR conditions (Supplementary Fig. 2). Titanium mesh, which is an OER inert metal catalyst, also exhibited good 2e⁻-WOR performance in 1.0 M Na₂CO₃ with a maximum H₂O₂ FE of ~ 47% at 2.36 V (Supplementary Fig. 2e, f). However, it presented poor stability due to possible surface passivation under high oxidation potentials (Supplementary Fig. 2 g). These above experimental results strongly support our hypothesis that CO₂/carbonate may directly participate in and promote the H₂O₂ generation process as a promising 2e⁻-WOR mediator.

Electrochemical H₂O₂ generation at industrial-relevant current densities

To further amplify the carbonate mediation effect for improved H₂O₂ generation performance, we evaluated the electrochemical 2e⁻-WOR performance of FTO electrode in high concentration carbonate solutions. Figure 3a, b and Supplementary Fig. 7 show the I–V curves and corresponding H₂O₂ FEs under different potentials. In 2.0 M Na₂CO₃ solution, we achieved high H₂O₂ FEs of ~60 to 70%, while delivering large current densities of up to 800 mA cm⁻². This impressive H₂O₂ performance can be even further improved by using 5.0 M of K₂CO₃. The reason why we chose to use K₂CO₃ is due to its higher solubility in water (112.3 g in 100 g water at 25 °C) than that of Na₂CO₃ (29.4 g in 100 g water at 25 °C)³⁹. The FTO catalyst achieved a 10 mA cm⁻² onset current density at 2.75 V vs. RHE in 5.0 M K₂CO₃, which is 50 mV lower than that in 2.0 M Na₂CO₃. With the overpotentials gradually increased, the H₂O₂ FE quickly ramped up to a plateau of over 80% under a wide range of current densities of up to 1 A cm⁻² (Fig. 3b). We achieved a maximal H₂O₂ FE of 87% at 600 mA cm⁻², representing a 522 mA cm⁻² H₂O₂ partial current density (Fig. 3b). More impressively, at 3.7 V, the catalyst reached a current density of 1 A cm⁻² while still maintaining a high H₂O₂ selectivity of 78%, achieving an industrial-relevant H₂O₂ partial current density of 780 mA cm⁻². The H₂O₂ generation rate can be further boosted to a maximal partial current of 1.3 A cm⁻² (73% FE at 1.8 A cm⁻² overall current) using an extended-range power supply. This corresponds to an unprecedented H₂O₂ production rate of 24.3 mmol cm⁻² h⁻¹ and is orders of magnitude higher compared to previous reports (Fig. 3c, d). Please be noted here that this power supply was not equipped with a three-electrode system therefore the anode potential could not be accurately measured (see Methods). Such high electrochemical H₂O₂ generation rates benefit from the sufficient mass diffusions in aqueous solutions where no triple-phase boundary is needed, which shows an advantage compared to the cathodic H₂O₂ generation from 2e⁻-ORR (oxygen reduction reaction) where O₂ gas diffusions typically limit its reaction rates to hundreds of milliamps per square centimeter (Supplementary Note 2)^2,40. Finally, long-stability is usually a big challenge for 2e⁻-WOR due to the high oxidation potentials that could damage the electrode^{9,11,16,17,41,42} (Supplementary Table 1). Besides, in traditional batch reactors, as the products continuous to accumulate during the long-term stability test (Supplementary Fig. 8a), the electrolyte environment will be changed and thus the electrocatalytic performance continuously decayed. Our FTO catalyst electrode, coupling with a flow reactor (Supplementary Fig. 8b), presented excellent durability under water oxidation conditions, maintaining a stable potential and high H₂O₂ FEs of over 80% to deliver a 150 mA cm⁻² current for 250 hours (Fig. 3e and Supplementary Table 1).

Fig. 3: Electrochemical 2e⁻-WOR performance using high concentration carbonate mediator.

a I–V curves of FTO catalyst in 2.0 M Na₂CO₃ and 5.0 M K₂CO₃. K₂CO₃ was chosen due to its higher solubility than Na₂CO₃. The maximal current density in our standard three-electrode cell was cut at 1 A cm⁻². b The corresponding H₂O₂ FEs. The highest FE reached 87% in 5.0 M K₂CO₃ with a current density of 600 mA cm⁻² under 3.46 V. c H₂O₂ FEs and partial currents in 5.0 M K₂CO₃ solution with current densities greater than 1 A cm⁻². The error bars represent two independent tests. d Comparison of H₂O₂ selectivity and activity between this work and previous reports^{7,9,11,12,13,14,41}. e The stability test of FTO catalyst by maintaining a 150 mA cm⁻² WOR current in 5.0 M K₂CO₃ solution. Its H₂O₂ selectivity and potential remained stable for a 250-hour continuous operation.

The good selectivity, activity, and stability of our carbonate mediated 2e⁻-WOR makes it possible for this anode reaction to be coupled with 2e⁻-ORR cathodic reaction to double the efficiency of electrons in producing H₂O₂ from both electrodes (Supplementary Fig. 9a). On the anode side, H₂O can be oxidized to H₂O₂ via carbonate mediation by our high-performance 2e⁻-WOR catalyst. On the cathode side, we used oxidized carbon black (demonstrated in our previous study⁴⁰) as the selective 2e⁻-ORR catalyst to reduce O₂ into H₂O₂ (Supplementary Fig. 9b, c)^40,43. Based on the previous definition of H₂O₂ FE on one side of the electrode, the maximal overall H₂O₂ FE in this two-electrode system is 200%. As shown in Supplementary Fig. 9d, our system delivered a 100-mA cell current (1 cm² FTO electrode) at 2.5 V cell voltage with a high overall H₂O₂ FE of 140%, suggesting a significant improvement compared to either 2e⁻-WOR or 2e⁻-ORR system. Furthermore, to fully use generated H₂O₂ and Na₂CO₃ mediator, we designed a process (Supplementary Fig. 9a) for a continuous generation of an adduct product between Na₂CO₃ and H₂O₂ (Na₂CO₃ ∙ 1.5H₂O₂, Supplementary Fig. 9e–g)^11,44.

Mechanism studies

To gain a molecular level understanding of the reaction mechanism in our CO₂/carbonate mediated 2e⁻-WOR, we employed molecular dynamics simulations, coupled with experimental studies, to reveal the most possible reaction pathway. Reaction intermediates that could exist in electrolyte under the applied high oxidation potentials (~ 3 V vs. RHE)⁴⁵, including CO₃^•−, OH^•, and percarbonates (HCO₄⁻ or C₂O₆²⁻), were taken into consideration^34,46. We first proposed several possible reaction pathways as summarized in Supplementary Fig. 10. After an initial screening based on theoretical studies and experimental observations, we proposed that the carbonate-mediated water oxidation to H₂O₂ could proceed via the following four reaction intermediate steps with most favorable thermodynamics^7,28,47,48 CO₃^•−, HCO₄⁻, HCO₃⁻+ H₂O₂, and CO₂ + H₂O₂ + OH⁻. We used ab initio molecular dynamics (AIMD) to evaluate the thermodynamics of these intermediate steps. A 12 Å × 12 Å × 12 Å cubic supercell with 57 water molecules was used to simulate bulk water and maintain a density of 1 g cm⁻³ (Supplementary Fig.11a, b). To assess the energies of these intermediates in ionic form in aqueous solution, seven H₂O molecules were replaced with CO₃^•− or HCO₄⁻ in succession (Fig. 4a and Supplementary Fig. 11b, c), while HCO₃⁻+ H₂O₂ or CO₂ + H₂O₂ + OH⁻ replaced eight H₂O molecules (Supplementary Fig. 11d, e). Taking the simulation of CO₃^•− as an example here, we first evaluated the convergence of the average energy. Figure 4a showed running average of total energy in different averaging time windows from 0.5 ps to 2.0 ps. We found that the more stable average energy is obtained at 2.0 ps window because the fluctuations in average energy between positive and negative were less than 0.1 eV. Therefore, we took the average energy of the last 2.0 ps as the energy of CO₃^•− in Supplementary Table 2, which also applies to the other three intermediates. According to the AIMD simulation results, here we suggest a carbonate-mediated 2e⁻ water oxidation reaction mechanism as shown in Fig. 4b (all these elementary reactions are exothermic under the electrochemical potential range we operated): First, CO₃²⁻ is oxidized to CO₃^•− at high oxidation potentials on the electrode surface (CO₃²⁻ → CO₃^•−+ e⁻, ΔG = − 0.72 eV at U = 3.0 V vs. RHE). Second, the generated CO₃^•− coupled with a H₂O molecule is further oxidized to HCO₄⁻ (CO₃^•−+ H₂O → HCO₄⁻+ H⁺ + e⁻, ΔG = − 0.27 eV at U = 3.0 V vs. RHE). Subsequently, HCO₄⁻ is hydrolyzed to generate H₂O₂ and converted back to HCO₃⁻ or CO₂ (HCO₄⁻+ H₂O → HCO₃⁻+ H₂O₂ or HCO₄⁻+ H₂O → CO₂ + H₂O₂ + OH⁻, ΔG = − 0.31/0.42 eV without applied potential). As shown in Supplementary Fig. 12, the first and second elementary reactions are electrochemical steps. Under an electrode potential of 3.0 V, both reactions are exothermic. The third fundamental reaction is a non-electrochemical step, and the calculation results show that the former is more likely to occur than the latter. During this reaction step, H⁺ in H₂O combines with the –OOH group in HCO₄⁻ molecule to generate H₂O₂, and the remaining OH⁻ group from H₂O can combine with remaining CO₂ from HCO₄⁻ to form HCO₃⁻. We note here that this hydrolysis step involves O exchange between water and carbonate ions, which becomes an important evidence to be confirmed by experiments. Finally, HCO₃⁻ or CO₂ + OH⁻ generated from the anode, coupled with the remaining OH⁻ group generated from the cathode, is concerted back to CO₃²⁻ to close the mediation loop in the system (Supplementary Fig. 13). We also found out that the FTO electrode provides a conducting and stable surface to extract electrons from carbonate, and may not serve as a classic catalytic surface where proper chemical bonds with reaction intermediates were usually established to facilitate reaction steps (Supplementary Figs. 14 and 15).

Fig. 4: Molecular understanding of the reaction mechanism and experimental detection of key intermediates.

a Convergence of average energy of CO₃^•− with 50 H₂O molecules using AIMD. Evaluating average of total energy was done over a varying length of averaging time, from 0.5 ps to 2.0 ps. The time coordinate for the average energy corresponds to the point from which the averaging window begins (e.g., for an averaging window of 0.5 ps, the average that is shown at 500 fs corresponds to the average from 500 fs to 1,000 fs). Cyan, red, light pink, and gray balls represent the O in CO₃^•−_, the O in H₂O, H, and C atoms, respectively. b Reaction mechanism of carbonate-mediated 2e⁻-WOR to H₂O₂. Two possible key intermediates, CO₃^•− and HCO₄⁻, were proposed, to facilitate the 2e⁻ pathway. c EPR spectra of Na₂CO₃ electrolyte containing DMPO spin trap. Compared to the solutions of non-electrolysis or post-electrolysis, the solution obtained from the in-situ trapping method under electrolysis exhibit a clear four-line 1:2:2:1 splitting pattern characteristic of the DMPO^• − OH adduct, indicating the formation of carbonate radical intermediates under electrolysis. d ¹⁸O isotope abundance in dried carbonate (Methods) at varied isotope exchange time between Na₂CO₃ solute and deionized water without or without electrolysis. The ¹⁸O isotope abundance in dried carbonate under WOR conditions was orders of magnitude higher than that of natural exchange background, indicating the C-O bonding dissociation and formation due to the percarbonate hydrolysis step. The error bars represent two independent tests.

To support the proposed reaction mechanism, we first tested the formation of carbonate radical by electron paramagnetic resonance (EPR). Due to the short lifetime of carbonate radical, we used 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin trap and an in-situ trapping method (see Methods) to detect its generation under electrolysis (Supplementary Fig. 16a)⁴⁹. Compared to the direct mixture of DMPO and before/post-electrolysis carbonate solution, the solution obtained from the in-situ trapping method under electrolysis exhibit a clear four-line 1:2:2:1 splitting pattern characteristic of the DMPO^• − OH adduct (Fig. 4c and Supplementary Fig. 16a), indicating that carbonate radical was formed at the surface of FTO electrode under oxidation potentials. Furthermore, we performed an ¹⁸O isotope labeling experiment to support the possible formation of HCO₄⁻ intermediates. As we mentioned earlier, the hydrolysis process of HCO₄⁻, if existing in the electrolyte, will lead to oxygen exchange between carbonate and water (Supplementary Fig. 16b). Therefore, if we use ¹⁸O isotope-labeled water as the electrolyte, we would expect an increased abundance of ¹⁸O (δ¹⁸O) in Na₂CO₃ after electrolysis compared to that of natural exchange without electrolysis (Methods)¹¹. Two sets of ¹⁸O isotope experiments were therefore designed: First, to figure out the natural exchange rate as the background, we tested δ¹⁸O as a function of time in Na₂CO₃ after it was dissolved in ¹⁸O labeled water; Next, we did the same tests of δ¹⁸O at the same time points under WOR electrolysis. As clearly shown in Fig. 4d, the abundance of ¹⁸O isotope in Na₂CO₃ under electrolysis condition continues to increase over time, and is several orders of magnitude higher than that without electrolysis, suggesting the violent interaction and chemical bond reconfiguration between carbonate and water via the percarbonate intermediate pathway, instead of a direct water oxidation pathway (Supplementary Fig. 17).

In conclusion, we demonstrated a CO₂/carbonate mediated electrochemical water oxidation for high-performance H₂O₂ generation, where carbonate ions help to steer the 4e⁻ reaction pathway to 2e⁻ via intermediates such as carbonate radicals and percarbonate, and FTO electrode provides a highly conductive, stable, and 4e⁻-OER inert surface. As a result, we achieved a high H₂O₂ selectivity of up to 87%, industrial-relevant H₂O₂ generation partial current of up to 1.3 A cm⁻², as well as excellent stability, suggesting a promising route for the renewable and onsite generation of H₂O₂ using electricity. This mediation process could be further extended to other electrochemical oxidation applications, especially for those suffering from 4e⁻-OER competition such as chlorine evolution, hydrocarbon oxidation, nitrogen oxidation, etc. Future works can be focused on exploring other catalytic materials enabling this mediation reaction, improving the onset potentials, and further extending the durability.

Methods

Materials

FTO was purchased from MSE Supplies (2.2 mm 7-8 Ohm/Sq FTO TEC 7 Coated Glass Substrates, used in all experiments if there is no other noting) and Sigma (SKU: 735159 and 735256). AZO and ITO were purchased from Sigma. Titanium mesh was purchased from Kunshan GuangJiaYuan new materials Co. Ltd. The oxidized carbon was prepared according to a previous report⁴⁰. Hydrophobic carbon fiber paper (CFP, PTFE loading 60wt.%) was purchased from Fuel Cell store. Chemicals were purchased from Sigma and VWR International Company.

Electrocatalytic oxidation of H₂O¹¹

The electrochemical measurements were run at 25 °C in a customized gas-tight H-type glass cell separated by Nafion 117 membrane (Fuel Cell Store). A BioLogic VMP3 workstation was employed to record the electrochemical response. In a typical three-electrode system, a platinum foil (Beantown Chemical, 99.99%) and a saturated calomel electrode (SCE, CH Instruments) were used as the counter and reference electrode, respectively. The FTO electrodes were used as the working electrodes, and stainless-steel alligator clip was used to connect the FTO glass electrode. The alligator clip was not immerged in the electrolyte. The geometric area of the FTO electrode was 0.5 to 1 cm², which was precisely defined by an electrochemically inert, hydrophobic wax (Apiezon wax WW100) during electrochemical tests. Before electrochemical measurements, all samples were pre-stabilized at 20 mA cm⁻² to clean the FTO surface. All potentials measured against SCE (E_SCE) were converted to the RHE (E_RHE) scale in this work using E_RHE = E_SCE + 0.244 V + 0.0591 × pH, where pH values of the electrolytes were determined by an Orion 320 PerpHecT LogR Meter (Thermo Scientific). The pH values of investigated electrolytes are listed in Supplementary Table 2. The electrolyte in the anodic compartment was stirred at a rate of 1,600 r.p.m. during electrolysis. All the measured potentials were manually compensated unless stated otherwise. The overall resistance was determined by potentiostatic electrochemical impedance spectroscopy at frequencies ranging from 0.1 Hz to 200 kHz, and manually compensated as E (iR corrected versus RHE, where iR is the voltage drop from overall resistance) = E (versus RHE) – R (the overall resistance, including the electrolyte resistance and the intrinsic resistance of the electrode) × i (amps of average current). The impedance number was about 12 to 17 Ω. DC stabilized power supply (ITECH) was used for the ultra-high current density (over 1 A cm⁻²) electrolysis. The volume of the electrolyte solution was 25 mL, and the reaction time was determined by the electrolysis current to insured the total electrolysis coulomb was about 10 to 50 C. After electrolysis, the generated H₂O₂ was detected by using the standard potassium permanganate (0.1 N KMnO₄ solution, Sigma-Aldrich) titration process, according to the following equation:

$$2{{{{{{\rm{MnO}}}}}}}_{4}^{-}+5{{{{{{\rm{H}}}}}}}_{2}{{{{{{\rm{O}}}}}}}_{2}+6{{{{{{\rm{H}}}}}}}^{+}\to 2{{{{{{\rm{Mn}}}}}}}^{2+}+5{{{{{{\rm{O}}}}}}}_{2}+8{{{{{{\rm{H}}}}}}}_{2}{{{{{\rm{O}}}}}}$$

The typical quantification time is about 5~10 min, which was much longer than the lifetime of carbonate radical (the lifetime of radical is about several microseconds)⁵⁰.

In this work, sulfuric acid (2.0 N H₂SO₄, VWR International Company) was used as the H⁺ source. The FE for H₂O₂ production is calculated using the following equation:

$${{{{{\rm{FE}}}}}}=\frac{{{{{{\rm{generated}}}}}}\;{{{{{{\rm{H}}}}}}}_{2}{{{{{{\rm{O}}}}}}}_{2}\,\left({{{{{\rm{mol}}}}}}\right)\times 2\times 96485\,\left({{{{{\rm{C}}}}}}\;{{{{{{\rm{mol}}}}}}}^{-1}\right)}{{{{{{\rm{total}}}}}}\;{{{{{\rm{amount}}}}}}\;{{{{{\rm{of}}}}}}\;{{{{{\rm{charge}}}}}}\;{{{{{\rm{passed}}}}}}\;({{{{{\rm{C}}}}}})}\times 100\;({{{{{\rm{maximum}}}}}}\;100 \% )$$

For the stability test, a continuous three-electrodes flow cell was employed to continuously produce H₂O₂ (Supplementary Fig. 8b). The electrolyte in the anodic compartment was stirred at a rate of 1600 r.p.m. during electrolysis. A peristaltic pump (Longer) was used to pump in K₂CO₃ electrolyte to the WOR side, and another peristaltic pump was used to pump out generated H₂O₂ in the WOR side. The electrolyte flow rate of the WOR side was 6 mL min⁻¹. The FE for H₂O₂ production is calculated using the following equation:

$${{{{{\rm{FE}}}}}}=\, \frac{{{{{{\rm{generated}}}}}}\;{{{{{{\rm{H}}}}}}}_{2}{{{{{{\rm{O}}}}}}}_{2}\,\left({{{{{\rm{mol}}}}}}{{{{{{\rm{L}}}}}}}^{-1}\right)\times 2\times 96485\;\left({{{{{\rm{C}}}}}}\;{{{{{{\rm{mol}}}}}}}^{-1}\right)\times {{{{{\rm{flow}}}}}}\;{{{{{\rm{rate}}}}}}\;({{{{{\rm{mL}}}}}}\,{{{{{{\rm{s}}}}}}}^{-1})}{{j}_{{{{{{\rm{total}}}}}}}({{{{{\rm{mA}}}}}})}\\ \times 100\;({{{{{\rm{maximum}}}}}}\;100 \% )$$

For the 2e⁻-WOR//2e⁻-ORR electrosynthetic cell test, 0.5 mg cm⁻² oxidized carbon catalyst was air-brushed onto 2-cm² Sigracet 35 BC gas diffusion layer (Fuel Cell Store) electrodes as 2e⁻-ORR cathode. Then, a 1-cm² FTO electrode was used as anode. The two electrodes were therefore placed on opposite sides in the H cell. O₂-saturated 2.0 M Na₂CO₃ was used as electrolyte. The cathode was open to the atmosphere. The flow rate of 2.0 M Na₂CO₃ electrolyte was 3 ml min⁻¹ at both sides, as controlled by a peristaltic pump. A current of 100 mA was employed for H₂O₂ production. The FE of the electrosynthetic cell for H₂O₂ production is calculated using the following equations, respectively:

${{{{{\rm{FE}}}}}}=\frac{{{{{{\rm{generated}}}}}}\;{{{{{{\rm{H}}}}}}}_{2}{{{{{{\rm{O}}}}}}}_{2}\;\left({{{{{\rm{mol}}}}}}\;{{{{{{\rm{L}}}}}}}^{-1}\right)\times 2\times 96485\;\left({{{{{\rm{C}}}}}}\;{{{{{{\rm{mol}}}}}}}^{-1}\right)\times {{{{{\rm{flow}}}}}}\;{{{{{\rm{rate}}}}}}\;({{{{{\rm{mL}}}}}}\,{{{{{{\rm{s}}}}}}}^{-1})}{{j}_{{{{{{\rm{total}}}}}}}({{{{{\rm{mA}}}}}})}\times 100\;({{{{{\rm{maximum}}}}}}\;200 \% )$

To obtain solid H₂O₂, the electrolyte after electrolysis (two-electrode configuration) was firstly concentrated by rotary evaporator, then about 100 ml of absolute isopropanol is added into 20 ml of the electrolyte for extraction, and the mixture is mechanically stirred. The precipitate is separated by vacuum filtration and washed several times with absolute isopropanol. The isolated precipitate is then dried in a vacuum oven at room temperature for 24 h.

To quantify the gas products during electrolysis, argon gas (Airgas, 99.995%) was delivered into the anodic compartment at a rate of 50.0 standard cubic centimeters per minute (sccm, monitored by an Alicat Scientific mass flow controller) and vented into a gas chromatograph (Shimadzu gas chromatography-2014) equipped with a combination of molecular sieve 5 Å, Hayesep Q, Hayesep T and Hayesep N columns. A thermal conductivity detector was mainly used to quantify gas product concentration. The partial current density for the O₂ produced was calculated as follows:

$${j}_{i}={x}_{i}\times v\times \frac{{n}_{i}{{{{{\rm{F}}}}}}{p}_{o}}{{{{{{\rm{RT}}}}}}}\times {\left({electrode}\;{area}\right)}^{-1}$$

where x_i is the volume fraction of certain product determined by online GC referenced to calibration curves from the standard gas sample (Airgas), v is the flow rate of 50.0 sccm, n_i is the number of electrons involved, p_o = 101.3 kPa, F is the Faradaic constant, T= 298 K and R is the gas constant. The corresponding Faradaic efficiency at each potential is calculated by ${{{{{\rm{FE}}}}}}=\frac{{j}_{i}}{{j}_{{total}}}\times 100 \%$. Of note, H₂O₂ FEs and O₂ FEs are evaluated from two completely different quantification systems: one is from solution titration, and another is from gas chromatographic method. When adding together, as shown in Fig. 2e, the total FEs for H₂O₂ and O₂ are close to 100% (within testing error range), suggesting there are no side reactions other than WOR under the electrolysis conditions.

Characterizations

SEM was performed on a FEI Quanta 400 field emission scanning electron microscope. Powder X-ray diffraction data were collected using a Bruker D2 Phaser diffractometer in parallel beam geometry employing Cu Kα radiation (λ = 1.54056 Å) and a 1-dimensional LYNXEYE detector, at a scan speed of 0.02° per step and a holding time of 1 s per step. X-ray photoelectron spectroscopy was obtained with a PHI Quantera spectrometer, using a monochromatic Al Kα radiation (1486.6 eV) and a low energy flood gun as neutralizer. All XPS spectra were calibrated by shifting the detected carbon C 1 s peak to 284.6 eV.

¹⁸O isotope measurement

We used ¹⁸O isotope-labeled water to prepare the Na₂CO₃ electrolyte. First, 0.14 mol (14.8 g) Na₂CO₃ (Sigma) was dissolved into 70 ml of ultrapure Milli-Q water. Then, 0.6 ml 10% H₂O¹⁸ was added. After stirring for 5 min, the fresh 2.0 M Na₂CO₃ solution was used as the electrolyte for water oxidation. After electrolysis at 400 mA for a certain time, 1 ml of the electrolyzed 2.0 M Na₂CO₃ solution was taken out; meanwhile another 1 ml of solution was taken from the bulk 2.0 M Na₂CO₃ solution without electrolysis. Then both of the sample solutions were quickly precipitated by 5 ml 1.0 M BaCl₂. Next, the solution was centrifuged and washed with ultrapure Milli-Q water. The powder was then dried at 60 °C for ¹⁸O isotope analysis. Of note, the natural exchange time with the water was the same for the unelectrolyzed and electrolyzed Na₂CO₃. Oxygen isotope analysis was performed on a Gas Bench-Conflo-Isotope Ratio Mass Spectrometer (Delta V, Thermo Scientific) system. Then ~0.2 mg of sample powder was weighed out into an exetainer, which was flushed by helium flow for 10 min on the GasBench. Then 0.3 ml of 105% phosphoric acid was added into the exetainer for 24-h reaction. The generated CO₂ gas was then delivered to the mass spectrometry for isotope ratio analysis. The oxygen isotope composition of the carbonate was calculated based on the measured oxygen isotope composition of the CO₂ gas, based on the fractionation factor between the two at the reaction temperature. The isotope experiments to monitor the ¹⁸O isotope offers an analytical precision (1σ) of 0.05‰ for δ¹⁸O and the values are reported as standard δ notation with respect to the Vienna Standard Mean Ocean Water (VSMOW)⁵¹.

EPR measurement

Electrochemical generated carbonate radical was performed in 1.0 M Na₂CO₃ solution with a constant electrolysis current of 200 mA. Once the potential became stable, pipette contains DMPO solution (200 mM, 100 μl) was immersed into the electrolyte and make sure the pipette tip was fully contact with the FTO electrode. The radical generated on the electrode will in-situ react with the DMPO solution in the interface between pipette and FTO electrode, and the final solution (sample named solution obtained from in-situ trapping method) contains 100 μl electrolyte and 100 μl DMPO solution. For the sample named DMPO + Na₂CO₃ solution, DMPO solution (200 mM, 100 μl) was directly mixed with 1.0 M Na₂CO₃ (100 μl). For the sample named DMPO + post-electrolysis solution, DMPO solution (200 mM, 100 μl) was mixed with the electrolyte (100 μl) after electrolysis at 200 mA for 5 min. Finally, above obtained solutions were transferred to capillary for EPR test. All EPR measurements were taken at room temperature with EPR spectrometer (Bruker A300). We set typical parameters as follows: 3515 G center of field; 9.852054 GHz frequency; 1 G modulation amplitude; 40 msec conversion time; 81.92 msec time constant; 20.38 Mw microwave power; 120 G scan width; 40 s sweep time.

Molecular dynamic calculations

A cubic box 12 Å × 12 Å × 12 Å was modeled with 57 water molecules to maintain a density of water at ∼1 g cm⁻³ (Supplementary Fig. 11b). To fit the box of this size, CO₃^•−, HCO₄⁻, HCO₃⁻ + H₂O₂ and CO₂ + H₂O₂ + OH⁻ replaced seven, seven, eight, and eight water molecules, respectively. The negative charge (⁻) of the ion was neutralized by a uniform background charge. We did not consider the hydroxide ions and protons in the electrolyte during simulation. Total energy of the above four possible intermediates was calculated by AIMD at a constant temperature of 300 K (using Nose-Hoover thermostat⁵², with a time step of 0.5 fs) as implemented in the Vienna ab initio simulation program (VASP)^53,54. Perdew-Burke-Ernzerhof (PBE) functional⁵⁵ and DFT-D3⁵⁶ methods were used to describe the exchange and correlation energies and the van der Waals interactions, respectively. A plane-wave cutoff energy of 400 eV and Gamma centered k-mesh of 1 × 1 × 1 were set in MD simulations. Considering the calculation cost, we ran 7.5 ps for each structure. From these AIMD calculations, we analyzed the arithmetic average of their total energies using different averaging time windows from 0.5 ps to 2.0 ps. It is found that averaging within a time window of 2.0 ps gives a value fluctuates around the final value by ±0.1 eV, less than other time windows. Therefore, average energy was evaluated from last 2.0 ps AIMD trajectory, as shown in Supplementary Table 3. The formation of CO₃^•− and HCO₄⁻ are electrochemical oxidation steps. Supplementary Table 4 shows inorganic standard electrode potentials⁴⁵. The reaction between H₂O and CO₃^•− was calculated by two separated process including H₂O → OH^• + H⁺ + e⁻ and CO₃^•− + OH^• → HCO₄⁻ because the energy of H⁺ + e⁻ item in CO₃^•− + H₂O → HCO₄⁻ + H⁺ + e⁻ reaction step is difficult to accurately calculate by MD simulations. Therefore, we deal with this step by adding the above two steps together. To better match the results of the experiment, we convert the standard electrode potential into a reversible hydrogen electrode (U_RHE = U_SHE + 0.059 × pH, pH = 12). Take CO₃²⁻ → CO₃^•− + e⁻ reaction as an example, the standard electrode potential for this reaction is +1.57 ± 0.03 V and the potential corresponding to conversion to RHE is 2.278 V (pH=12, U_RHE = + 1.57 + 0.059 × 12 = 2.278 V), corresponding to the reaction free energy of 0 eV. When the electrode potential is increased to 3 V, the reaction is exothermic with an energy of −0.72 eV.

Gibbs free energies on the SnO₂(110) surface calculations

The exchange-correlation potential is described by the generalized gradient approximation (GGA) with spin polarized revised Perdew-Burke-Ernzerhof (RPBE)⁵⁷ functional due to that it is more adept at describing chemisorption on metals. The projector augmented wave is applied to describe the electron-ion interaction and the plane-wave energy cutoff is set to 400 eV. All structures are optimized with a convergence criterion of 1 × 10⁻⁵ eV for the energy and 0.02 eV/Å for the forces. The SnO₂(110) surface is the most common of the faces of rutile structure tin oxide (Space group: P42/mnm, No. 136)⁵⁸. Therefore, the SnO₂(110) surface is modeled using a slab with a (4 × 2) surface supercell consisting of 3 trilayers and containing 144 atoms. The vacuum spacing is set to more than 15 Å for surface isolation to prevent interaction between two neighboring surfaces. The top two trilayers are fully relaxed during the structural optimization and geometry optimizations for SnO₂(110) are performed with 1 × 1 × 1 k-mesh. The Poisson-Boltzmann (PB) implicit solvation model, Vaspsol⁵⁹, was used to describe the effect of solvation as implemented in VASP 5.4.4, with a dielectric constant ε = 80 for water. All molecules were performed in a 15 × 15 × 15 ų unit cell with a 5 × 5 × 5 k-point grid sampling.

Computational hydrogen electrode (CHE)⁶⁰ model was used to calculate the Gibbs free energy change for OER elementary reactions. At electrode potential U = 0 V (vs. reversible hydrogen electrode, RHE), the Gibbs free energy change (ΔG) can be calculated by

$$\Delta {{G}}=\Delta {{E}}+\Delta {{{E}}}_{{{ZPE}}}-{{T}}\Delta {{S}}+\Delta {{{G}}}_{{{U}}}$$

where ΔE is the energy difference between the products and reactants from DFT computations; ΔE_ZPE and ΔS are the changes in zero-point energy and entropy, respectively, which are obtained from the vibrational frequency calculations; T is the temperature at 298 K. The energy corrections of gas-phase species in this work, including zero point energies (ZPE) and entropies (TS), are listed in Supplementary Table 5. ΔG_U = −neU, where U is the electrode applied potential relative to RHE, e is the elementary charge transferred and n is the number of proton–electron pairs transferred. 2e^–-WOR path is calculated in alkaline condition. We calculated the chemical potential of hydroxides and electrons with reference to previous work⁶¹. μ(e^–)– μ(OH^–) – eU = 9.81 eV was obtained by repeated calculations under reversible hydrogen electrode potential (U₀(RHE) =1.23 V) at T = 298.15 K.

The accuracy of the above DFT methodology applied in our work is mainly based on published work. More details can be found in the Supporting Information.