Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Nature Communications volume 15, Article number: 8846 (2024 ) Cite this article hydrogen peroxide cost
Water electro-oxidation to form H2O2 is an important way to produce H2O2 which is widely applied in industry. However, its mechanism is under debate and HO(ads), hydroxyl group adsorbed onto the surface of the electrode, is regarded as an important intermediate. Herein, we study the mechanism of water oxidation to H2O2 at Pt electrode using in-situ Raman spectroscopy and differential electrochemical mass spectroscopy and find peroxide bond mainly originated from the coupling of two CO32- via a C2O62- intermediate. By quantifying the 18O isotope in the product, we find that 93% of H2O2 was formed via the CO32- coupling route and 7% of H2O2 is from OH(ads)-CO3•− route. The OH(ads)-OH(ads) coupling route has a negligible contribution. The comparison of various electrodes shows that the strong adsorption of CO3(ads) at the electrode surface is essential. Combining with a commercial cathode catalyst to produce H2O2 during oxygen reduction, we assemble a flow cell in which the cathode and anode simultaneously produce H2O2. It shows a Faradaic efficiency of 150% of H2O2 at 1 A cm−2 with a cell voltage of 2.3 V.
Renewable-energy-driven water electrolysis produces hydrogen (H2) at the cathode and oxygen (O2) at the anode, with H2 gaining attention as a clean energy carrier1. However, the limited market value of O2 evolution reaction (OER) encourages researchers to explore alternative reactions at the anode2. As an alternative, water oxidation to produce hydrogen peroxide (H2O2) is a valuable reaction, because H2O2 is a potent green oxidizing agent with versatile applications in industries such as paper and textile bleaching3, water treatment4,5, chemical synthesis6,7, and electrical energy generation8. Hydrogen peroxide is currently produced almost exclusively through the anthraquinone oxidation process9. However, this method is not considered sustainable or environmentally friendly due to its complexity and multi-step nature, which involves the use of large quantities of organic solvents. In contrast, the electrocatalytic water oxidation reaction to form H2O2 is regarded as a greener and more sustainable method for producing hydrogen peroxide.
H2O2 generation at the anode is accomplished through the unconventional two-electron water oxidation reaction (2e−-WOR), which differs from traditional O2 evolution reactions. This method is challenging, considering the thermodynamic favorability of the O2 evolution reaction (Eqs. (1) and (2)).
Most studies on the two-electron water oxidation reaction focused on developing effective catalytic materials to enhance the selectivity of water oxidation towards hydrogen peroxide10,11,12. While the electrocatalyst plays a crucial role, the choice of the aqueous supporting electrolyte is equally significant as it can impact both the production and stability of H2O213,14,15. Recent research exhibited that H2O2 production rates increased in carbonate (CO32−) electrolytes, i.e. Na2CO3 solution16, K2CO3/KHCO3 mixtures17,18, and other carbonate-based electrolytes13. In carbonate solutions, using a platinum (Pt) electrode for water oxidation to produce H2O2 can achieve a Faradaic efficiency reaching about 80% at a current density >100 mA cm−219. Despite these progresses, there is significant controversy surrounding the anodic 2e−-water oxidation reaction mechanism for hydrogen peroxide production in carbonate solutions13,15,16,20,21,22,23. Many research groups believe that this reaction proceeded through a key intermediate OH(ads) adsorbed at the surface18,20,22,23,24. OH(ads) is a crucial intermediate in both ORR and OER processes, and two OH(ads) easily combine to produce H2O2 (Eqs. (3) and (4)), denoted as Route 1)
Following this concept, researchers have focused on exploring catalysts that demonstrate reduced binding strengths with intermediate oxygen species in comparison to the OER20,24. Some researchers also propose that the 2e−–water oxidation involves the coupling at the electrode surface between OH(ads) with CO32− to establish an O−O bond (Eqs. (5)–(9)), ultimately leading to the production of peroxide species (Eq. (10)), which is denoted as Route 2 here14,15,22,25.
Besides Routes 1 and 2, two CO3•− species might couple to form a C2O62− intermediate with an O−O bond (Eq. (11)), producing H2O2 eventually (Eq. (12), denoted as Route 3)19,26,27,28.
Moreover, the carbonate coupling reaction in Route 3 is believed difficult to occur due to the strong repulsion between two negatively charged ions15,29. However, in our recent study in ether-based electrolytes, we found Li2CO3 oxidized via an intermediate C2O62,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30. Currently, it is difficult to make a definitive judgment on these three reaction mechanisms due to the lack of critical quantitative evidence. Although in-situ spectroscopic results identified the surface species at electrodes21,22, it is challenging to quantitatively estimate the contribution of each pathway to the overall electrolysis reaction and H2O2 products.
To tackle this challenge, we employed in-situ Raman spectroscopy and differential electrochemical mass spectroscopy (DEMS), along with 18O isotope labeling, to quantify the contributions of various routes to H2O2 production and identify the reaction mechanisms. Our results show that the primary oxidation reaction on the Pt electrode surface proceeds via C2O62− intermediate, and Route 3 accounts for a high ratio of 93% of the overall process. The OH(ads) + CO3•− route (Route 2) contributes to ~7% of H2O2 formation, while the OH(ads)-OH(ads) coupling route (Route 1) plays a negligible role in H2O2 formation in this work. By comparing several catalysts, we found the Faradaic efficiency for H2O2 production is influenced by the competition of CO3(ads) and OH(ads) on the electrode surface.
To demonstrate its potential application, we assembled a membrane-free flow cell, in which the cathode and anode produce H2O2 simultaneously. The flow cell achieved a 150% Faradaic efficiency for H2O2 formation (FEH2O2) at a high density of 1 A cm−2. It produces H2O2 at concentrations up to 40 mM, underscoring its potential for diverse applications and its significance in the realm of electrochemistry.
In potassium carbonate (K2CO3) solution, Pt serves as an excellent catalyst for 2e−–water oxidation reaction to hydrogen peroxide and its FEH2O2 is as high as 75% and 85% in 2 and 5 M K2CO3 solution, respectively (Figs. 1a, b and S1). In-situ Raman spectra at an electrochemically roughened Pt electrode were recorded at various potentials up to 2.71 V (vs. RHE). At 2.21 V, the water oxidation started (Fig. 1a), and two Raman peaks at 883 and 1332 cm−1 appeared (Fig. 1b). These two peaks increased with the increase of potential, and typically, they are associated with the intermediates species adsorbed on the electrode surface or in near-surface solution during water oxidation. The 883 cm−1 peak typically corresponds to the O−O stretching vibration of peroxide species31, and thus, it may be assigned to an H2O2 product or an intermediate containing an O−O bond. To verify whether this peak originates from H2O2 or not, we added commercial H2O2 solution to the K2CO3 solution and found that the O−O vibration peak of H2O2 appeared at 871 cm−1 (Fig. S2), instead of 883 cm−1. Therefore, the peak at 883 cm-1 is likely to be an intermediate like C2O62− and HCO4− rather than H2O2. Furthermore, an isotope-labeled control experiment was conducted in the K2CO3–D2O electrolyte, and the peak at 883 cm−1 did not shift, indicating that this intermediate is H-free (Fig. S3). Thus, the peak at 883 cm-1 is assigned to intermediate C2O62−. Correspondingly, the peak at 1332 cm-1 is assigned to the C−O anti-symmetric stretching of the C2O62− intermediate. We have obtained the crystal of peroxodicarbonate at a low reaction temperature of −20 °C. The Raman spectrum and the digital photo of the crystal are shown in Fig. 1c. The experiment was conducted in a saturated potassium carbonate solution, but surprisingly, we observed almost no Raman peaks of potassium carbonate in the crystal. The Raman peak at 1338 cm−1 is assigned to the C−O anti-symmetric stretching of C2O62−. Additionally, the peaks at 731 and 867 cm−1 can be attributed to the stretching of O–O bonds with different coordination in the unit cell of the crystal, akin to Na2O231. The control experiments further validate the assignment of the Raman peak observed in Fig. 1b at 883 and 1332 cm−1 to intermediate C2O62−.
a Linear scan voltammetry plots for Pt electrode in 2 M K2CO3 electrolyte (pH 12.1 ± 0.2). Inset is the corresponding Faradic efficiencies of H2O2 formation. The ohmic resistance of the working electrode with a surface area of 0.150 cm−2 was 3.6 ± 0.2 Ω, and the applied potential here is not iR-corrected. b In-situ Raman spectra for a roughened Pt electrode at various potentials from 1.81 to 2.71 V (vs. RHE). The collection time for each spectrum was 60 s. c The Raman spectrum of obtained peroxodicarbonate (K2C2O6) crystal. The inset shows the digital photo of the crystal. d Time-dependent Raman spectra after turning off the applied potential. The collection time for each spectrum was 5 s.
When the applied high potential was removed, the Raman peak at 883 cm−1 shifted to 872 cm−1, which was assigned to H2O2, and this peak gradually diminished and disappeared after 20 s (Fig. 1d). The trend of this peak suggested that the C2O62− intermediate rapidly transforms into H2O2 after removing the applied high potential, and the H2O2 product diffused away from the surface-enhanced region, leading to a decrease of the 872 cm−1 peak. Meanwhile, the peak at 1332 cm−1 immediately disappeared, confirming that this peak is attributed to the intermediate species containing a C−O bond, such as C2O62−. It should be noted that in Fig. 1b, the peak of H2O2 at 871 cm−1 is weak and difficult to observe due to its low concentration and the influence of the peak at 883 cm−1. It is only observed when a large amount of H2O2 formed at a high potential of 3.01 V (Fig. 1d). The 18O-isotope-labeling experiments were carried out to further confirm the peak assignment and the source of the oxygen in the peroxide intermediate species.
To further confirm the peak assignment, 18O-labeled Raman was carried out in a 2 M K2C16O3–H218O electrolyte solution. Because 18O from H218O slowly exchanges with 16O in C16O32− in the solution, (as shown in Fig. S4)32, a fresh-prepared solution was first used. In the K2C16O3–H218O electrolyte, if H2O2 was produced via the OH(ads) intermediate (Routes 1 and 2), the peroxide intermediates would be labeled by 18O-isotope, resulting in a significant amount species containing 18O−18O (Route 1) and 16O−18O (Route 2), and the corresponding Raman peak (at 883 cm−1) would have a red shift towards a low wavenumber33,34. However, the Raman peak at 883 cm−1 did not shift (Fig. 2), indicating that OH(ads) may not involved in H2O2 formation. In the discussion below, the isotopes in the oxidation products were quantified, and the main product of oxidizing H218O was H216O2, which further confirms that the reactions are independent of the OH(ads) species.
a Comparison of in-situ Raman spectra recorded at a roughened Pt electrode in 2 M K2C16O3–H216O and 2 M K2C16O3–H218O electrolyte. The oxygen isotope exchanges after storage for some time, and the ratio (18O:16O) in K2CO3 increases with the storage time. b The Raman spectra of 2 M K2C16O3– H218O subtracts the Raman spectrum obtained in 2 M K2C16O3–H216O. Peaks at 840 and 862 cm−1 increased with storage time due to the 16O/18O exchanges in the solution.
After storing the K2C16O3–H218O solution for 8 days, the oxygen (16O and 18O) in carbonate ions and water exchanged, resulting in a K2C16/18O3–H216/18O solution (Fig. S4 and Supplementary Table 1). This method allowed us to obtain partially 18O-labeled CO32−. As shown in Fig. 2, the 883 cm−1 peak shifted to the low wavenumber, and a shoulder peak at 862 cm−1 (16O−18O) appeared. After storing for one month, 18O:16O was further exchanged and the majority of CO32− was labeled with 18O. Therefore, the 862 cm−1 peak (16O−18O) was enhanced, and a new peak appeared at 840 cm−1, corresponding to the 18O−18O signal33,34. Meanwhile, the Raman peak at 1332 cm−1 (C−16O) also had a redshift to 1316 cm−1 (C−18O) (Fig. 2b). The experimentally observed peak positions are consistent with the theoretical calculations (Supplementary Table 2). Therefore, based on the above results, the Raman peaks at 883 and 1332 cm−1 in Fig. 1b are assigned to intermediate species C2O62−.
While Raman spectroscopy confirmed the formation of C2O62− intermediate during water oxidation to hydrogen peroxide, it presents challenges in accurately quantifying the contribution of the C2O62− route to the overall reaction. Because the Raman is not a quantitative method and the intermediates that appeared in the Raman spectra might not represent the major reaction route. Herein, we employed 18O-labeled in-situ DEMS to quantify the ratio of 18O:16O in O2 evolution and H2O2 production during water oxidation. The experimental details are described in the “Methods” section. Briefly, 16O2, 18O16O, and 18O2 were quantified directly to obtain the ratio of 18O:16O. The isotope-labeled H2O2 was disproportionated to release O2, which was quantified by mass spectrometer afterward and thus obtained the ratios of H216O2:H216O18O:H218O2. To verify this method, we prepared an H2O2–K2CO3–H218O solution, in which only water was 18O-labeled, and added a piece of Pt to encourage the peroxide disproportionation. As shown in Fig. S5a, only 16O2 evolved, and no 18O16O or 18O2 was identified, indicating that the two oxygen atoms in one O2 molecule directly originate from the same hydrogen peroxide molecule. This is consistent with the previous reports, where the dominant reaction pathways for the formation of oxygen during H2O2 disproportionation on a Pd surface do not involve O–O bond cleavage35. To verify whether rapid oxygen exchange occurs between hydrogen peroxide and carbonate ions, potentially affecting the reliability of our experimental results, we conducted a control experiment in a K2C16/18O3–H218/16O solution (Fig. S5b). The results showed that the oxygen exchange reaction between hydrogen peroxide and carbonate ions is slow and can be neglected within the timescale of our in-situ DEMS experiment.
In this way, we quantified the ratio of 16O2:16O18O:18O2 and H216O2:H216O18O:H218O2 during water oxidation in the fresh K2C16O3–H218O solution. The oxygen exchange between K2C16O3 and H218O was negligible because the electrolyte was freshly prepared. Figure 3a shows that O2 evolved with a ratio of 90:9:1 (16O2:16O18O:18O2) and H2O2 formed with a ratio of 93:7:0 (H216O2: H216O18O: H218O2). Overall, almost no H218O2 was detected, and only 1% of 18O2 formed following 18OH(ads)–18OH(ads) coupling pathway, suggesting Route 1 was negligible. 18O16O and H216O18O came from the C16O3−(ads)+18OH(ads) pathway (Route 2), which contributed to <10% of the reaction. The 16O2 and H216O2 came from the CO32− coupling pathway (Route 3), and their ratio in the final products is greater than 90%, which indicates the major reaction proceeded via C2O62− intermediate as shown in the Raman results (Figs. 1 and 2). According to the ratio of isotope-labeled H2O2 after the oxidation reaction, Route 3 contributes to 93% of the peroxide formation.
a The O2 evolution in 2 M K2C16O3–H218O fresh solution at reaction current density 200 mA cm−2. The applied potential was turned off at 18 min and the later O2 evolution came from the H2O2 disproportionation, representing the ratio of H216O2:H216O18O:H218O2. b O2 evolution during water oxidation at different current densities in 2 M K2C16O3–H218O fresh solution. c O2 evolution during water oxidation at reaction current density 200 mA cm−2 after different reaction times in 2 M K2C16O3–H218O fresh solution. It should be noted that a 10 min reaction was conducted between each test. The top panel shows the I–t curves to indicate the duration of the time when the potential was applied to the cell. The mid panel was the molar flux of O2 evolution, and the bottom panel was the cumulative mole of the O2. The molar flux (mid panel) of gas evolution was denoted as ṅ, and the cumulative mole (bottom panel) of the gas was denoted as n. The flow rate of the carrier gas Ar was 2 mL min−1.
After storing the same solution for one month, the ratio (18O:16O) in the carbonate ions increased due to significant oxygen exchange in the solution. The same experiments were carried out, and the O2 evolution is shown in Fig. S6a. The ratio of O2 became 41:46:13 (16O2:16O18O:18O2), which is consistent with the in-situ Raman results (Fig. 2). To further validate the reliability of our experiments, we conducted the experiments in an H2S16O4–H218O solution. As shown in Fig. S6b, 18O2 was the major product, and almost no 16O2 was identified. It is reasonable because the production of oxygen occurs through a 4e− water oxidation in this case, and the oxygen atoms in the evolved O2 originated entirely from H218O.
Based on the results above, a reaction mechanism for water oxidation to H2O2 was proposed in Fig. 4. Initially, CO32− was adsorbed onto the Pt surface, forming CO3−(ads) (Eq. (13)). Subsequently, it underwent O−O coupling with another one, resulting in the formation of a C2O62− intermediate (Eqs. (14) and (15)). Only Eqs. (13) and (14) involved charge transfer and were driven by potential. Once C2O62− formed, it underwent spontaneous hydrolysis to yield hydrogen peroxide (Eqs. (16) and (17)).
a Main reaction pathway for catalytic two-electron water oxidation to H2O2 in a high-concentration K2CO3 solution. b Free-energy diagram of two-electron water oxidation to H2O2 on partially oxidized Pt(111) and α-Ni(OH)2(001). c The structures of OH− and CO32− adsorption on partially oxidized Pt(111), and the adatom coverage of oxygen is 0.2 ML. d α-Ni(OH)2(001) surface. Compared to OH−, CO32− exhibits strong adsorption on partially oxidized Pt(111) surface.
In this process, CO32− served as a co-catalyst and it was regenerated eventually. In the isotope labeling experiment, during its regeneration, the oxygen in CO32− exchanges with oxygen in H218O, leading to a C16/18O32− solution (Fig. 4a). This phenomenon is distinct in the in-situ DEMS with a long reaction time. In Fig. 3c, as the reaction progresses, more and more C16/18O32− accumulated in the solution, causing the ratio of 16O18O and 18O2 in the products to increase gradually. As the reaction proceeded, the initial C16O32− was converted into C216O62−, and then C216O62− reacted with H218O to form H216O2 and 18O-labeled bicarbonates HC16/18O3−. Then HC16/18O3− will couple with the OH− to regenerate C16/18O32−(Eq. (17)). It should be noted that the OH− was generated at the cathodic reaction (2H2O + 2e− → H2 + 2OH−). When the formed C16/18O32− mixed with initial C16O32− and participated in the reaction via C216/18O62− intermediate (Route 3), some 18O−16O bonds and even 18O−18O bonds formed, leading to more 18O16O and 18O2 evolution. The ratio(18O:16O) in the carbonate ions increased with the reaction time. The carbonate ions were not depleted in this process, and their total quantity remained constant. Therefore, for the overall reaction, the oxygen in H2O2 eventually all came from water, equivalent to water oxidation to H2O2. Hence, under a long reaction time, with an increased charge consumption over time, even in Route 3, a substantial amount of 16O18O evolved. However, if only the reaction current density is increased but the reaction time is not long enough, there would not be a significant accumulation of C16/18O32− in the solution. Therefore, the ratio of 16O18O and 18O2 in the products remains low (Fig. 3b).
The adsorption of CO32− at the electrode surface (Eq. (13)) is crucial because the oxidation of CO32− is the rate-determining step of this reaction (Fig. 4a, b). Considering that OH−/H2O can also adsorb on the electrode surface, the competitive adsorption between CO32− and OH− largely determines whether the reaction of 2e−-water oxidation to hydrogen peroxide occurs (Eq. (19)).
Herein, we compared the free energy of this equilibrium on partially oxidized Pt, and Ni(OH)2 surfaces. It should be noted that at a high applied potential, the Pt surface can undergo partial oxidation36,37. The free energy of CO32− adsorption on partial oxide Pt has a significant advantage (Fig. 4c, d), with a substantial energy advantage (−0.21 eV) in competition with OH−. As a result, CO32− species may have a higher coverage on Pt and play a primary role in water oxidation. On the contrary, OH− adsorption energy on the Ni(OH)2 surface is dominant, making OER the primary reaction on Ni(OH)2 surfaces. When water oxidation was conducted at Ni(OH)2 surface in 2 M K2CO3 solution, the Faradic efficiency of H2O2 was <5% (Fig. S7a). This result supports the hypothesis that the adsorption energy of CO32− at the electrode surface is crucial in this reaction. We have also conducted DFT calculations of three routes of H2O2 formation to provide a more comprehensive understanding of the reaction mechanisms (Fig. S8). The potential determining step for Route 1 is the activation of H2O as OH*, while that for Route 2 is the oxidation of CO3−* to HCO4−*. It should be noted that at high oxidation potentials, a layer of hydroxyl groups inevitably forms on the Pt electrode surface36. Although the surface-generated OHads do not directly participate in the H2O2 generation process, they may play an important role in the adsorption of CO32− ions on the electrode surface38, thereby activating Routes 2 and 3.
The adsorption energy of CO32− can guide us in designing catalysts for water oxidation to produce hydrogen peroxide. However, a strong CO32− adsorption does not guarantee a high FEH2O2 because the oxidation of the catalyst has to be considered, and we should pay attention to the status of the electrode surface at such a high potential. The theoretical potential for H2O2 production (1.78 VRHE, Eq. (2)) is 500 mV higher than that for O2 evolution (1.23 VRHE, Eq. (1)), thus a high potential at the electrode is required, and the catalyst’s resistance to oxidation could not be ignored. For instance, DFT calculations show that the adsorption of CO32− is favorable on the Au and Ru surface, but experiments show almost no H2O2 formation at the Au and Ru surface (Fig. S7a), which appears to contradict the DFT results. We found that the surfaces of Au and Ru were oxidized at high potentials and formed a layer of metal oxides (Fig. S7b), which alters the adsorption energy of CO32−. Despite CO32− adsorption being favorable at Au and Ru (Figure S9), the formation of oxides diminishes the advantage of CO32− adsorption, making them more prone to catalyze H2O2 decomposition to O2. AuOx39 and RuOx40 are good catalysts for the 4e− O2 evolution reaction; thus, OH(ads) adsorption is more favorable than CO3−(ads) at Au and Ru electrodes. In comparison, Pd was not easily oxidized, and the CO32− adsorption still had a significant advantage at high potential, making it a good H2O2 catalyst like Pt (Fig. S7a). And in-situ Raman results show that Route 3 also plays an important role in the formation of H2O2 on the Pd electrode (Fig. S10). Therefore, when designing high-efficiency catalysts for H2O2 production, we should estimate the adsorption energy of CO32− on the surface at high potential and even partially oxidized surface. To confirm this mechanism, we changed the concentration of the carbonate solution, which is a simple way to influence the adsorption competition between CO32− and OH−. When the carbonate concentration falls below 0.8 M, the Faradaic efficiency for H2O2 production decreases significantly (Fig. S1).
Some reports suggest that water oxidation to H2O2 is more efficient in bicarbonate solutions compared to carbonate solutions21,41. It is believed that in bicarbonate solutions, HCO3− loses an electron to form percarbonate (HCO4−), which then hydrolyzes to produce hydrogen peroxide14,21,22,25. However, when water oxidation was conducted in KHCO3/K2CO3 mixture solutions with a Pt electrode, the FEH2O2 decreased as the pH decreased, substantially dropping to 7% in 2 M KHCO3 solution (Fig. S11). Therefore, two factors may contribute to the significant reduction in the FEH2O2 in bicarbonate solutions: (a) HCO3− was a hydrolysis product in Route 3, and it might inhibit the forward progress of the reaction. (b) The reaction kinetics for HCO3− oxidation was slower than for CO32−42,43.
By combining water oxidation to H2O2 at the anode and oxygen reduction to H2O2 at the cathode, we constructed a membrane-free flow cell device to produce H2O2 simultaneously at both anode and cathode. Because H2O2 is aggressive and tends to attack the proton/anion exchange membrane, leading to membrane degradation, a membrane-free flow cell is preferred. As shown in Fig. 5a, Pt was used on the anodic side to oxidize water to H2O2, and oxidized carbon catalysts were used on the cathode electrode to reduce O2 to H2O2 (Fig. S12)44,45. As H2O2 is formed as both electrodes, the shuttle effect in the membrane-free cell is minimized. Under a cell voltage of 2.3 V, it delivered a high current density of 1 A cm-2anode with a total Faradaic efficiency of 150% for H2O2 production (Fig. 5b, c). This setup provided a feasible solution for on-demand hydrogen peroxide production. The concentration of H2O2 in the solution could reach 40 mM, which was already sufficient for various applications such as wastewater treatment. The H2O2 concentration is restricted by its disproportionation catalyzed by Pt. If a catalyst could be found that would not catalyze H2O2 disproportionation, then a higher concentration of hydrogen peroxide could be obtained. Additionally, if a catalyst can be designed to adsorb carbonate efficiently, a low concentration of carbonate solution is sufficient to produce H2O2. It implies that the electrochemical system for H2O2 production has the potential to contribute to the sustainability of economic development, particularly with the fast development of electrocatalysts.
a Schematic design of the H2O2 production flow cell coupling 2e− water oxidation and 2e− oxygen reduction reactions. b Cell voltage and H2O2 Faradaic efficiency and c the amount of produced H2O2 in a flow cell at a constant cell current 150 mA (1 A cm−2Pt). The data obtained from a single experiment and the applied potential here are not iR-corrected. The ohmic resistance of the flow cell was 4 ± 1 Ω.
Herein, we studied the mechanisms of 2e− water oxidation to H2O2 in carbonate solution using in-situ Raman and DEMS. By quantifying the 18O-labeled O2 and H2O2 products, we quantified the contribution of three pathways on the Pt surface. CO32−-CO32− coupling via a C2O62− intermediate is the major pathway at Pt, accounting for 93% of the H2O2 production. In contrast, the OH(ads) + CO3•− route contributed the rest 7%, while the OH(ads)–OH(ads) coupling route exhibited negligible contribution to H2O2 formation. During the water oxidation reaction, the competitive adsorption between CO3(ads) and OH(ads) at the electrode surface is the key factor influencing the CO32−–CO32− coupling and Faradaic efficiency for H2O2 production. Therefore, the strong adsorption of CO32− at the surface of catalysts is preferred. The structure change and partial oxidation at high potentials should be considered for the catalyst design, otherwise because the oxidation of the catalyst surface affects the competition of CO3(ads) and OH(ads). Furthermore, we demonstrated a membrane-free flow cell where the cathode and anode produce H2O2 simultaneously. This cell exhibits a 150% Faradaic efficiency of H2O2 production at a high current of 1 A cm−2anode under 2.3 V, demonstrating its potential for sustainable development applications.
Potassium carbonate (K2CO3, 99.5%), potassium bicarbonate (KHCO3, AR), and COOH functionalized multi-walled CNTs (>95 wt%, OD: 10–20 nm, length: 10–30 μm) were purchased from Aladdin Co., Ltd. Heavy-oxygen water (H218O, 98 atom%18O) were purchased from Energy Chemical Co., Ltd. Pt electrode (diameter: 1 mm, length: 5–38 mm) was provided by Gaossunion Co., Ltd. All chemicals were used without further purification.
The electrochemical measurements were run at 25 °C in an H-type glass cell with a Nafion 117 membrane (diameter: 3.5 cm, thickness: 183 μm). A CHI 760 workstation was employed to record the electrochemical response. In a typical three-electrode system, a platinum foil and a saturated Ag/AgCl electrode were used as the counter and reference electrode, respectively. The Pt wire with a surface area of 0.150 cm−2 was used as the working electrodes. The pH of the 2 M K2CO3 was 12.1 ± 0.2, which was measured with a pH meter (Sartorius, PB-10). The applied potential was converted to RHE by the following equation: ERHE(V) = EAg/AgCl + 0.059 × pH + 0.197. The ohmic resistance of the working electrode was measured by electrochemical impedance spectroscopy at open circuit potential, and the voltage was not iR-corrected. The volume of the anode electrolyte solution was 10 mL, and the reaction time was determined by the electrolysis current to ensure the total electrolysis coulomb was about 15–30 C. After electrolysis, the generated H2O2 was detected by using the standard potassium permanganate (0.01 N KMnO4 solution) solution, according to the following equation:
In brief, a test solution was first prepared by adding 250 μL of 0.01 N KMnO4 to 2.70 mL of 1 M H2SO4. Subsequently, 50 μL of the reacted solution was added to the freshly prepared solution. The change in KMnO4 absorption was monitored using UV–vis spectra. The FE for H2O2 production is calculated using the following equation:
The same method was used to test the 2e−-WOR to H2O2 performance for Au, Pd, Ru, and Ni(OH)2. Ni(OH)2 was in-situ grown on Ni wire through several cycles of CV scanning between 0 V and 2 V(vs. Ag/AgCl)46.
For the flow cell test, a precursor solution was prepared by adding 3.3 mg of carbon catalyst and 50 μL of Nafion solution to 1 mL of ethanol. This solution was then sprayed onto a 2 cm2 hydrophobic carbon paper electrode with a gas diffusion layer44, serving as the 2e−-ORR cathode. A 0.150 cm−2 Pt wire electrode was used as an anode. 2 M K2CO3 was used as an electrolyte. The cathode was open to the atmosphere. The flow rate of electrolytes was 5 ml min−1, as controlled by a peristaltic pump. A current of 150 mA was employed for H2O2 production. The ohmic resistance of the cell was measured by electrochemical impedance spectroscopy at open circuit potential, and the voltage was not iR-corrected. The FE o for H2O2 production is calculated using the following equations:
It should be noted when using the flow cell, the maximum FE for H2O2 production is 200%
K₂C₂O₆ was synthesized through the anodic oxidation of a 40 wt% of K2CO3 solution at −20 °C. The electrochemical reaction was conducted in an H-type glass cell without a membrane under galvanostatic conditions at a current of 1 A. A Pt wire (diameter = 1.0 mm, length = 37 mm) served as the anode, and a Pt plate (5 mm × 10 mm) was used as the cathode. The volume of the electrolyte solution in the cell was 50 mL, and the reaction time was 5 h. The resulting potassium peroxodicarbonate produced at the anode was collected, then filtered and washed with ethanol.
In-situ Raman spectra were collected on a Raman microscopy system (HORIBA Xplora Plus) with a long-focus objective lens using an excitation laser of 532 nm. The collection time of each spectrum was 10–60 s. A custom-built Raman cell was used to perform the in-situ Raman measurements (Fig. S13). A Pt wire was employed as the counter electrode, and saturated Ag/AgCl was used as the reference electrode for the in-situ cell tests. A roughened Pt rod electrode with a 1 mm diameter was used as the working electrode. To obtain the roughened surface, the Pt electrode is subjected to a square wave potential (between −0.40 and 2.2 V (vs. Ag/AgCl), 0.5–2 kHz) in a 0.5 M H2SO4 solution for a few minutes47.
A differential electrochemical mass spectrometer (DEMS, Prima BT, Thermo Scientific Ltd.) was employed to quantitatively measure the evolved gas. A homemade well-sealed cell was used for the DEMS measurements (Fig. S14). The Ar carrier gas carried the evolved gas in the cell into DEMS. The typical flow rate of Ar carrier gas is 2 mL min−1. The time resolution of DEMS is 10 s. The general gas evolution in chemical reactions was examined by the mass spectrometer (MS, Prima BT) itself. Typically, 0.5 mL of the prepared electrolyte was added to the cell. After purging with carrier gas for some time to establish a stable baseline, the potential was applied to the cell to initiate the 2e−-WOR, and the evolved gases were quantified using mass spectrometry.
We used 18O isotope-labeled water to prepare the K2CO3 electrolyte. First, 0.55 g K2C16O3 was dissolved into 2 ml of 98% H218O. After stirring for 5 min, the fresh 2.0 M K2C16O3–H218O solution was used as the electrolyte in Raman and DEMS experiments. The same solution was stored at room temperature for various durations to achieve different degrees of 18O isotope-labeled carbonate solution. Subsequently, it was used to conduct the Raman and DEMS experiments again.
A control experiment was first conducted, that the disproportionation of H216O2 in K2C16O3–H218O solution. First, the K2C16O3–H218O solution was prepared as described above, and then 5 μL of 30% H216O2 was added to the solution. Even though a solution containing H216O was added, the ratio of H216O in the H216O2–K2C16O3–H218O solution, which was ~97%, was not significantly influenced. H216O2 will be disproportionate on the Pt surface, generating numerous small gas bubbles. After several minutes of the reaction, we applied 10 s of magnetic stirring to remove the gas from the Pt surface and conducted in-situ mass spectrometry analysis of the generated gas. The same method was used to quantitatively determine the ratio of H216O2, H216/18O2, and H218O2 produced in the solution after 2e−-WOR.
The ab initio calculations were performed with Gaussian 09 W molecular orbital packages48. Geometry optimizations and vibrational frequency calculations were carried out at the B3LYP-LB level, where LB denotes the large basis set, 6-311 + G (3df, 2p). Theoretical calculations were performed using density functional theory (DFT) with the Perdew–Burke–Ernzerhof (PBE) generalized gradient approximation (GGA)49, implemented in the Vienna ab initio simulation package (VASP)50. An energy cutoff of 400 eV was applied for the plane-wave basis to describe the valence electron wave functions. Self-consistent calculations of the single-electron wavefunction at the ground state were terminated once the energy and force convergence criteria of 10−⁵ eV and 0.02 eV/Å, respectively, were met. For the Pt/Au/Pd/Ru electrodes, a 2 × 2 × 1 K-point mesh was used with a Gamma-centered Monkhorst–Pack scheme in the corresponding Brillouin zone, while a 4 × 2 × 1 K-point mesh was used for the α-Ni(OH)₂ catalyst. The (111) crystal plane for the metal electrode was chosen as the reaction surface. For simplicity, the lower two layers of atoms in Pt(111)/Au(111)/Pd(111)/Ru(111) were fixed, and the top layer fully relaxed. For α-Ni(OH)₂, the (001) facet was used as the model for calculations, based on its layered structure. A vacuum region of 15 Å was created along the z-direction to ensure negligible interaction between mirror images. The optimized structural model is provided in Supplementary Data 1. The computational hydrogen electrode (CHE) model51 was used to calculate the Gibbs free energy change (ΔG) for 2e−-WOR elementary steps. ΔG can be calculated by the formula:
Where ΔE represents the adsorption energy of reaction intermediates, obtained directly obtained from DFT calculations. ΔEZPE is the change in zero-point energy, while TΔS represents the change of entropy, derived from the vibrational frequency calculations, with T set at 298 K. The last term represents the correction of electrical potential to Gibbs free energy given by ∆GU = −neU, where U is the applied electrode potential relative to the standard hydrogen electrode (set at 0 V), e is the elementary charge, and n is the number of electrons transferred.
The data that support the conclusion of this study are available within the paper and Supplementary Information. Source data are provided with this paper.
Sadiq, M., Alshehhi, R. J., Urs, R. R. & Mayyas, A. T. Techno-economic analysis of Green-H2@Scale production. Renew. Energy 219, 119362 (2023).
Wang, T., Cao, X. & Jiao, L. Progress in hydrogen production coupled with electrochemical oxidation of small molecules. Angew. Chem. Int. Ed. 61, e202213328 (2022).
Hage, R. & Lienke, A. Applications of transition-metal catalysts to textile and wood-pulp bleaching. Angew. Chem. Int. Ed. 45, 206–222 (2005).
Chidambara Raj, C. B. & Li Quen, H. Advanced oxidation processes for wastewater treatment: optimization of UV/H2O2 process through a statistical technique. Chem. Eng. Sci. 60, 5305–5311 (2005).
Kosaka, K. et al. Evaluation of the treatment performance of a multistage ozone/hydrogen peroxide process by decomposition by-products. Water Res. 35, 3587–3594 (2001).
Article CAS PubMed Google Scholar
Lane, B. S. & Burgess, K. Metal-catalyzed epoxidations of alkenes with hydrogen peroxide. Chem. Rev. 103, 2457–2473 (2003).
Article CAS PubMed Google Scholar
Chua, S.-C., Xu, X. & Guo, Z. Emerging sustainable technology for epoxidation directed toward plant oil-based plasticizers. Process Biochem. 47, 1439–1451 (2012).
Ma, J., Choudhury, N. A. & Sahai, Y. A comprehensive review of direct borohydride fuel cells. Renew. Sustain. Energy Rev. 14, 183–199 (2010).
Campos-Martin, J. M., Blanco-Brieva, G. & Fierro, J. L. G. Hydrogen peroxide synthesis: an outlook beyond the anthraquinone process. Angew. Chem. Int. Ed. 45, 6962–6984 (2006).
Perry, S. C. et al. Electrochemical synthesis of hydrogen peroxide from water and oxygen. Nat. Rev. Chem. 3, 442–458 (2019).
Shi, X., Back, S., Gill, T. M., Siahrostami, S. & Zheng, X. Electrochemical synthesis of H2O2 by two-electron water oxidation reaction. Chem 7, 38–63 (2021).
Wang, J. et al. Prospects and promises in two-electron water oxidation for hydrogen peroxide generation. Energy Fuels 37, 17629–17651 (2023).
Mavrikis, S., Perry, S. C., Leung, P. K., Wang, L. & Ponce de León, C. Recent advances in electrochemical water oxidation to produce hydrogen peroxide: a mechanistic perspective. ACS Sustain. Chem. Eng. 9, 76–91 (2021).
Pangotra, D. et al. Anodic production of hydrogen peroxide using commercial carbon materials. Appl. Catal. B 303, 120848 (2022).
Fan, L. et al. CO2/carbonate-mediated electrochemical water oxidation to hydrogen peroxide. Nat. Commun. 13, 2668 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Xia, C. et al. Confined local oxygen gas promotes electrochemical water oxidation to hydrogen peroxide. Nat. Catal. 3, 125–134 (2020).
Mavrikis, S. et al. Effective hydrogen peroxide production from electrochemical water oxidation. ACS Energy Lett. 6, 2369–2377 (2021).
Kim, C. et al. Concurrent oxygen reduction and water oxidation at high ionic strength for scalable electrosynthesis of hydrogen peroxide. Nat. Commun. 14, 5822 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Zhang, J. & Oloman, C. W. Electro-oxidation of carbonate in aqueous solution on a platinum rotating ring disk electrode. J. Appl. Electrochem. 35, 945–953 (2005).
Shi, X. et al. Understanding activity trends in electrochemical water oxidation to form hydrogen peroxide. Nat. Commun. 8, 701 (2017).
Article ADS PubMed PubMed Central Google Scholar
Gill, T. M., Vallez, L. & Zheng, X. The role of bicarbonate-based electrolytes in H2O2 production through two-electron water oxidation. ACS Energy Lett. 6, 2854–2862 (2021).
Li, L. et al. Electrochemical generation of hydrogen peroxide from a zinc gallium oxide anode with dual active sites. Nat. Commun. 14, 1890 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Article CAS PubMed Google Scholar
Kelly, S. R. et al. ZnO as an active and selective catalyst for electrochemical water oxidation to hydrogen peroxide. ACS Catal. 9, 4593–4599 (2019).
Mavrikis, S., Göltz, M., Rosiwal, S., Wang, L. & Ponce de León, C. Carbonate-induced electrosynthesis of hydrogen peroxide via two-electron water oxidation. ChemSusChem 15, e202102137 (2022).
Article CAS PubMed Google Scholar
Chen, L.-J., Lin, C.-J., Zuo, J., Song, L.-C. & Huang, C.-M. First spectroscopic observation of peroxocarbonate/peroxodicarbonate in molten carbonate. J. Phys. Chem. B 108, 7553–7556 (2004).
Irkham et al. Electrogenerated chemiluminescence by in situ production of coreactant hydrogen peroxide in carbonate aqueous solution at a boron-doped diamond electrode. J. Am. Chem. Soc. 142, 1518–1525 (2020).
Article CAS PubMed Google Scholar
Ruiz, E. J., Ortega-Borges, R., Jurado, J. L., Chapman, T. W. & Meas, Y. Simultaneous anodic and cathodic production of sodium percarbonate in aqueous solution. Electrochem. Solid-State Lett 12, E1 (2009).
Armstrong, D. A., Waltz, W. L. & Rauk, A. Carbonate radical anion—thermochemistry. Can. J. Chem. 84, 1614–1619 (2006).
Cao, D., Tan, C. & Chen, Y. Oxidative decomposition mechanisms of lithium carbonate on carbon substrates in lithium battery chemistries. Nat. Commun. 13, 4908 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Vacque, V., Sombret, B., Huvenne, J. P., Legrand, P. & Suc, S. Characterisation of the O-O peroxide bond by vibrational spectroscopy. Spectrochim. Acta Part A 53, 55–66 (1997).
Chiba, H. & Sakai, H. Oxygen isotope exchange rate between dissolved sulfate and water at hydrothermal temperatures. Geochim. Cosmochim. Acta 49, 993–1000 (1985).
Article ADS CAS Google Scholar
Moënne-Loccoz, P. et al. The ferroxidase reaction of ferritin reveals a diferric mu−1,2 bridging peroxide intermediate in common with other O2-activating non-heme diiron proteins. Biochemistry 38, 5290–5295 (1999).
Pavlovic, Z., Ranjan, C., van Gastel, M. & Schlögl, R. The active site for the water oxidising anodic iridium oxide probed through in situ Raman spectroscopy. Chem. Commun. 53, 12414–12417 (2017).
Plauck, A., Stangland, E. E., Dumesic, J. A. & Mavrikakis, M. Active sites and mechanisms for H2O2 decomposition over Pd catalysts. Proc. Natl Acad. Sci. USA. 113, E1973–E1982 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Huang , Y.-F. , Kooyman , PJ & Koper , MTM Intermediate stages of electrochemical oxidation of single-crystalline platinum revealed by in situ Raman spectroscopy . Nat. Commun. 7, 12440 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Drnec, J., Harrington, D. A. & Magnussen, O. M. Electrooxidation of Pt(111) in acid solution. Curr. Opin. Electrochem. 4, 69–75 (2017).
You, X. et al. Relationship between oxide identity and electrocatalytic activity of platinum for ethanol electrooxidation in perchlorate acidic solution. Commun. Chem. 6, 101 (2023).
Article CAS PubMed PubMed Central Google Scholar
Burke, M. S. et al. Revised oxygen evolution reaction activity trends for first-row transition-metal (oxy)hydroxides in alkaline media. J. Phys. Chem. Lett. 6, 3737–3742 (2015).
Article CAS PubMed Google Scholar
McCrory, C. C. L. et al. Benchmarking hydrogen evolving reaction and oxygen evolving reaction electrocatalysts for solar water splitting devices. J. Am. Chem. Soc. 137, 4347–4357 (2015).
Article CAS PubMed Google Scholar
Fuku, K. & Sayama, K. Efficient oxidative hydrogen peroxide production and accumulation in photoelectrochemical water splitting using a tungsten trioxide/bismuth vanadate photoanode. Chem. Commun. 52, 5406–5409 (2016).
Zuo, Z., Cai, Z., Katsumura, Y., Chitose, N. & Muroya, Y. Reinvestigation of the acid–base equilibrium of the (bi)carbonate radical and pH dependence of its reactivity with inorganic reactants. Radiat. Phys. Chem. 55, 15–23 (1999).
Article ADS CAS Google Scholar
Haygarth, K. S. et al. Carbonate radical formation in radiolysis of sodium carbonate and bicarbonate solutions up to 250 degrees C and the mechanism of its second order decay. J. Phys. Chem. A 114, 2142–2150 (2010).
Article CAS PubMed Google Scholar
Lu, Z. et al. High-efficiency oxygen reduction to hydrogen peroxide catalysed by oxidized carbon materials. Nat. Catal. 1, 156–162 (2018).
Wu, K.-H. et al. Highly selective hydrogen peroxide electrosynthesis on carbon: In situ interface engineering with surfactants. Chem 6, 1443–1458 (2020).
Cordoba‐Torresi, S. I., Gabrielli, C., Hugot‐Le Goff, A. & Torresi, R. Electrochromic behavior of nickel oxide electrodes: I. Identification of the colored state using quartz crystal microbalance. J. Electrochem. Soc. 138, 1548–1553 (1991).
Tian, Z.-Q., Ren, B. & Wu, D.-Y. Surface-enhanced raman scattering: from noble to transition metals and from rough surfaces to ordered nanostructures. J. Phys. Chem. B 106, 9463–9483 (2002).
Frisch, M. J. et al. Gaussian 09, Revision C.01 (Gaussian, Inc., 2016).
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
Article ADS CAS PubMed Google Scholar
Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169–11186 (1996).
Article ADS CAS Google Scholar
Nørskov, J. K. et al. Origin of the overpotential for oxygen reduction at a fuel-cell cathode. J. Phys. Chem. B 108, 17886–17892 (2004).
This research was financially supported by the National Natural Science Foundation of China (52173173, 21975124, 52272217), Natural Science Foundation of Jiangsu Province (BK20220051), Jiangsu Province Carbon Peak and Neutrality Innovation Program (Industry tackling on prospect and key technology) (BE2022031-4, BE2022002-3), Scientific and Technological Innovation Project of Carbon Emission Peak and Carbon Neutrality of Jiangsu Province (BE2022028−1), The Natural Foundation of Jiangsu Higher Education Institutions of China (23KJB430021). We are grateful to the High-Performance Computing Center of Nanjing Tech University for supporting the computational resources.
State Key Laboratory of Materials-Oriented Chemical Engineering, Nanjing Tech University, 211816, Nanjing, China
Heng Zhu, Ximei Lv, Yuexu Wu, Wentao Wang & Yuhui Chen
School of Physical and Mathematical Sciences, Nanjing Tech University, 211816, Nanjing, China
Key Laboratory of Energy Thermal Conversion and Control of Ministry of Education, School of Energy and Environment, Southeast University, 210096, Nanjing, China
Collaborative Innovation Center of Advanced Microstructures, National Laboratory of Solid State Microstructures, College of Engineering and Applied Sciences, Nanjing University, 210023, Nanjing, China
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
The work was conceived and designed by Y.H.C., S.C.Y., and H.Z.; H.Z. and Y.X.W. performed the Raman measurements; H.Z. and W.T.W. carried out DEMS experiments; H.Z. and X.M.L. performed the theoretical calculations; Y.P.W. discussed the experimental ideas; Y.H.C., H.Z. and S.C.Y. drafted the manuscript and revised the manuscript; all authors discussed the results.
Correspondence to Shicheng Yan or Yuhui Chen.
The authors declare no competing interests.
Nature Communications thanks Xiaoming Sun, Luciana Vieira and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Zhu, H., Lv, X., Wu, Y. et al. Carbonate-carbonate coupling on platinum surface promotes electrochemical water oxidation to hydrogen peroxide. Nat Commun 15, 8846 (2024). https://doi.org/10.1038/s41467-024-53134-3
DOI: https://doi.org/10.1038/s41467-024-53134-3
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
Nature Communications (Nat Commun) ISSN 2041-1723 (online)
ethylene diamine tetra acetic acid Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.