Introduction

Hydrogen peroxide (H2O2) is an important chemical, which has been widely used for various bleaching, sanitary, and ecological remediation purposes1,2. Unfortunately, the current H2O2 production is still dominated by the energy-intensive and centralized anthraquinone oxidation method, which is associated with significant transportation/storage risks and costs3,4. Considering this, the electrocatalytic two-electron oxygen reduction reaction (2e− ORR) has been regarded as a promising alternative due to its high safety, energy efficiency, eco-friendliness, and decentralized features5,6. Although great progress has been made for H2O2 production via the 2e− ORR under alkaline or acidic conditions recently, the significant difficulty in H2O2 separation from such electrolytes thus narrows down its application to very limited scenarios7. Consequently, developing a technology that can directly synthesize H2O2 in neutral conditions or even in natural water bodies without the requirement of post-synthesis separation is particularly attractive.

Brine, which mainly contains NaCl, is the most earth-abundant neutral medium and biocompatible with the human body. It has been identified as a primary research objective in the sustainable energy, biology, and environment-related fields8. The ever-increasing ecocrisis in relation to seawater pollution and eutrophication further signifies the necessity of direct H2O2 production in brine9. In addition, the coupled anodic chlorine evolution reaction (CER), when brine is used in the anode side of the electrolyzer, also generates value-added products, i.e., chlorine gas (Cl2)10, hypochlorous acid (HClO), and hypochlorite salts (ClO−)11, in which the ClO−/HClO is another chemical for water purification/disinfection.

Unlike conventional acid or alkaline electrolytes, electrosynthesis of H2O2 via 2e− ORR in neutral brine is particularly sluggish and unstable. On the one hand, neutral brine cannot provide enough protons on the absorbed O2 to form *OOH12, due to the sluggish water dissociation (WD) and the poor proton supply rate, collaboratively resulting in a significantly slower ORR kinetics and lower H2O2 yield rate than in acidic/alkaline electrolytes13. On the other hand, the chloride ions in salt water may be preferably adsorbed on the active sites and suppress the desired oxygen adsorption, thus compromising the ORR activity14. These characteristics thus impose significant challenges in developing high-efficiency and high-activity electrocatalysts for 2e− ORR in neutral brine6,7,15,16,17.

Interestingly, nickel hydroxide (Ni(OH)2) possesses unique features that may tackle these two issues in one batch. Firstly, it already shows tolerance against the chloride-containing species, which has been proven in its excellent stability in brine electrocatalysis18. Secondly, this 3d transition metal hydroxide is also capable of catalytic water dissociation, proton generation, and their transportation, which may significantly reduce the energy barrier of the proton-involving processes, e.g., forming *OOH intermediate during 2e− ORR. Moreover, Ni(OH)2 has also been proven inactive for hydrogen evolution reaction (HER), potentially offering additional benefits for achieving high-rate H2O2 generation over a wide potential range (below 0 V vs. RHE) in brine12,19.

On the basis of these considerations, we herein present a multifunctional Ni(OH)2 nanoplates anchored on carbon nanotubes (CNTs) for 2e− ORR in neutral brine. The unsaturated coordination Ni atoms exposed on O-vacancies or at the edge over the Ni(OH)2 nanoplates act as the major active sites for 2e− ORR electrocatalysis20,21, while the CNTs improve the electronic conductivity (Fig. 1). These merits, together with its suppression of hydrogen evolution, collaboratively result in a high H2O2 selectivity of over 85% in a wide potential range of 0.3 to −1.0 V vs. RHE, a high H2O2 yield rate of 30.86 ± 0.03 mol g−1 h−1 (3.086 mmol cm−2 h−1) at 200 mA cm−2 and 141 mol g−1 h−1 (14.1 mmol cm−2 h−1) at 1 A cm−2, and operational stability of over 156 h at 200 mA cm−2 with a high FE over 80%. Even in natural seawater, this catalyst still shows an excellent electrocatalytic performance with a H2O2 yield of 14.2 mol g−1 h−1 (1.42 mmol cm−2 h−1) at 100 mA cm−2 with FE over 70%. In a prototype electrolyzer with our catalyst, the electro-synthesized H2O2 demonstrates rapid water treatment and disinfection capability. This study proves the feasibility of the electrosynthesis H2O2 in seawater coupled with the CER in the anodic tank with extra valuable industrial chemicals production, i.e., ClO− disinfectants, which endows it huge potential in practical application, i.e., to alleviate seawater pollution.

Fig. 1: Illustration of the advantages of Ni(OH)2/CNT to selective H2O2 electrosynthesis, and the possible tanks with 2e− ORR and CER.
Fig. 1: Illustration of the advantages of Ni(OH)2/CNT to selective H2O2 electrosynthesis, and the possible tanks with 2e− ORR and CER.
Full size image

Schematic diagram on the proton transfer and water dissociation of Ni(OH)2 to rapid *OOH formation, and the undesirable 4e− ORR and hydrogen evolution process in neutral environment over Ni(OH)2/CNT (left); Schematic diagram on the tanks with 2e− ORR coupled with the CER reaction in the anodic tank (right).

Results

Theoretical prediction guided material synthesis

Firstly, density functional theory (DFT) calculations were performed to confirm the intrinsic capability of Ni(OH)2 for ORR in brine. The adsorption free energy of the key *OOH intermediates (ΔG*OOH) at the exposed Ni sites was calculated to be 4.14 eV, which is very close to the ideal value of ΔG*OOH (4.23 eV) for 2e− ORR. It corresponds to a small overpotential of only 0.09 eV, indicating its nearly barrier-free 2e− ORR activity (Fig. 2a, Supplementary Fig. 1 and Supplementary Data 1). As for the comparative Ni sites in NiO, they show a lower ΔG*OOH value of 3.06 eV, which is far from the ideal value, corresponding to a huge overpotential of 1.17 eV for 2e− ORR and indicating its preferred 4e− pathway for ORR. (and the activation barrier of *OOH formation on the active sites of NiO and Ni(OH)2 will be discussed later). To illustrate the chlorine resistance capability of Ni(OH)₂ in seawater, we conducted 5 ns molecular dynamics simulations using the large-scale atomic/molecular massively parallel simulator (LAMMPS). As shown in Fig. 2b and Supplementary Data 1, during the simulation process after reaching equilibrium under different electric fields, the distribution of chloride ions on the surface of the material was rather sparse, with a concentration not exceeding 0.1 g cm−3. Compared with Ni and NiO, there were fewer chloride ions on the surface of the material due to the repulsion of OH groups (Supplementary Fig. 2 and Supplementary Data 1), which demonstrated the better chlorine resistance performance of Ni(OH)₂. In addition, the optimized models for H* adsorption on Ni sites also show that hydrogen adsorption is weak on the surface of Ni(OH)2 with an expected large overpotential of ~0.51 eV for HER (Fig. 2c). Meanwhile, the large kinetic barrier of *H absorbed at the Ni site also confirms its inactiveness for catalyzing HER (Supplementary Fig. 3), enabling the possible 2e− ORR at negative potentials below 0 V vs. RHE to achieve a high H2O2 yield rate.

Fig. 2: Theoretical prediction guided Ni(OH)2/CNT synthesis and the structural characterizations.
Fig. 2: Theoretical prediction guided Ni(OH)2/CNT synthesis and the structural characterizations.
Full size image

a Gibbs free energy diagram for 2e− ORR at 0.7 V vs. RHE. b The density distributions of chloride ions in the z-axis direction for Ni(OH)2 (left) and equilibrium configuration of electrolyte system (3.6 wt% NaCl) above the electrode surface of Ni(OH)2, viewed from XZ cross-section. c Calculated free energy diagram for HER on Ni(OH)2. d The TEM image of Ni(OH)2/CNT. e STEM image of Ni(OH)2/CNT and corresponding elemental mapping images of Ni, C and O. f XRD pattern of Ni(OH)2/CNT, Ni(OH)2 and NiO/CNT. g Ni K-edge XANES of Ni(OH)2/CNT, standard Ni(OH)2 and Ni foil. h Ni K-edge EXAFS-FT signals of Ni(OH)2/CNT and standard Ni(OH)2. i EXAFS fitting results using the model with a path of Ni-O and Ni-Ni in R space (FT magnitude and imaginary part). The k3-weighted and not phase-corrected. The blue, pink and white represent nickel, oxygen and hydrogen, respectively.

Based on these predictions, Ni(OH)2 nanoplates, exposed sufficient surface and edge Ni sites, were grown on CNTs via a simple hydrothermal process. Transmission electron microscopy (TEM) images show that the Ni(OH)2 nanoplates, with a typical two-dimensional structure of 20–50 nm, are attached to CNTs (Fig. 2d). Meanwhile, the pure Ni(OH)2 and NiO/CNT are also prepared for comparison, which also possess a well-distributed and 2D nanoplate morphology (Supplementary Fig. 4). Under high resolutions, the lattice fringes show a planar spacing of 0.236 nm, which can be ascribed to the (101) facet of Ni(OH)2 (Supplementary Fig. 5), confirming the successful formation of β-Ni(OH)2. Meanwhile, the obvious lattice fringe of 0.209 nm and 0.240 nm can be ascribed to the (200) facet of NiO/CNT and the (101) facet of Ni(OH)222,23, respectively (Supplementary Figs. 6 and 7). The energy-dispersive X-ray spectroscopy (EDS) elemental mappings present the distribution of Ni, O, and C elements, respectively conformal with the Ni(OH)2 nanoplates and CNTs (Fig. 2e). The X-ray diffraction (XRD) patterns confirm the formation of β-Ni(OH)2 phase in Ni(OH)2/CNT, similar to the sole Ni(OH)2 nanoplates without CNTs (Fig. 2f)22 and in agreement with the TEM observations. In addition, the obvious electron paramagnetic resonance (EPR) signals for the Ni(OH)2/CNT, NiO/CNT, and Ni(OH)2 indicate the existence of O vacancies trapped with extra electrons (Supplementary Fig. 8)24, which would improve the number of exposed Ni sites at the O vacancies for ORR electrocatalysis.

To probe the chemical structure of the materials, X-ray photoelectron spectroscopy (XPS) was carried out. As shown in Supplementary Figs. 9 and 10, the survey scan confirms the coexistence of Ni and O for Ni(OH)2/CNT and NiO/CNT25. For the Ni 2p spectra in Ni(OH)2/CNT, it can be deconvolved into Ni2+ 2p3/2 species at 855.58 eV and a Ni2+ 2p1/2 species at 873.22 eV26, slightly shifted to a higher binding energy compared with Ni(OH)2 (855.28 vs. 872.92 eV), indicating there is a charge transfer from Ni(OH)2 to the oxygen-containing functional groups in Ni(OH)2/CNT27. However, there is no obvious effect on the performance of 2e− ORR in Ni(OH)2/CNT, because the major active sites are the unsaturated coordination Ni atoms exposed on O-vacancies or at the edge over the Ni(OH)2 nanoplates20,21 rather than the Ni sites close to CNTs.

X-ray absorption spectroscopy (XAS), including X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) measurements, was then carried out to clarify the coordination environment of the Ni sites in Ni(OH)2/CNT and NiO/CNT (Supplementary Figs. 11 and 12). The absorption edge of Ni K-edge XANES of Ni(OH)2/CNT and NiO/CNT is slightly red-shifted in comparison with the standard Ni(OH)2 and NiO sample (Fig. 2g), which may be caused by oxygen vacancies28. Meanwhile, in the Fourier transform EXAFS spectra of Ni(OH)2/CNT and the comparative Ni(OH)2, they both exhibit a main peak at 1.66 Å corresponding to the Ni-O species and a secondary peak at 2.69 Å corresponding to the Ni-Ni species29 (Fig. 2h), which indicates their similar Ni-O coordination configurations. Of note, the intensities of the Ni-O and Ni-Ni peaks of Ni(OH)2/CNT are considerably lower than those of the standard Ni(OH)2 sample, again suggesting the existence of structural defects in it30. The EXAFS fitting results confirm that the number of O atoms in the first coordination shell of the Ni sites is 5.7 (i.e., Ni-O5.7) with a distance of 2.07 Å (Fig. 2i and Supplementary Table 1), which also confirms the existence of O vacancies in Ni(OH)2/CNT, and the exposed unsaturated Ni atoms on these sites would be the potential active sites for 2e− ORR electrocatalysis.

Electrocatalytic performance for 2e− ORR under brine

The electrocatalytic 2e− ORR performance of the material was first evaluated in 3.6 wt% NaCl aqueous solution (i.e., simulated seawater) by linear sweep voltammetry (LSV) on a rotating ring-disk electrode, in which the collection coefficient is 0.37 (Fig. 3a and Supplementary Fig. 13). Ni(OH)2/CNT delivers the largest ring current, indicating its favorable two-electron pathway for ORR and the excellent chloride-tolerating capacity. Correspondingly, it possesses a high H2O2 selectivity of over 90% (Fig. 3b). Besides, the pure Ni(OH)2 also maintains the high H2O2 selectivity but with much lower current densities on both the disk and ring compared with Ni(OH)2/CNT. This confirms that the major role of introducing CNTs is to facilitate the reaction kinetics by significantly enhancing the electronic conductivity of the catalyst, while the intrinsic capability of Ni(OH)2 for highly selective 2e− ORR catalysis can be fully exploited in this case for Ni(OH)2/CNT (Supplementary Fig. 14).

Fig. 3: Electrochemical 2e− ORR catalytic performance of Ni(OH)2/CNT.
Fig. 3: Electrochemical 2e− ORR catalytic performance of Ni(OH)2/CNT.
Full size image

a LSV curves of different samples, tested at 1600 rpm and a rate of 10.0 mV s−1, showing a the current (densities) on the disk (jdisk) and ring (Iring) electrodes and b H2O2 selectivity and electron number of the materials in 3.6 wt. % NaCl. c Comparison of Iring and selectivity of Ni(OH)2/CNT in 0.9 wt. % NaCl, 3.6 wt. % NaCl and 0.1 M Na2SO4. d LSV and e H2O2 selectivity and electron number of Ni(OH)2/CNT at wide potential range. f Polarization curves of Ni(OH)2/CNT and NiO/CNT in Ar-saturated neutral electrolyte.

In addition, either in diluted NaCl solution (0.9 wt% NaCl solutions, i.e., physiological saline), or in the 0.1 M Na2SO4 solution, the Ni(OH)2/CNT material both shows high ORR current densities and H2O2 selectivity of over 80% (Fig. 3c). Similarly, Ni(OH)2/CNT also shows excellent 2e− ORR selectivity in other electrolytes, such as 0.1 M KOH (FE = 94%), simulated alkaline seawater (3.6 wt% NaCl + 0.1 M KOH, FE = 96%), natural seawater from Bohai Sea (FE = 80%), and alkaline seawater (natural seawater + 0.1 M KOH, FE = 85%) (Supplementary Figs. 15 and 16), which further confirms the huge potential of the H2O2 electrosynthesis using this Ni(OH)2/CNT material and its applicational readiness in various practical conditions.

In addition to the aforementioned merits, Ni(OH)2/CNT also shows stable current plateaus on both the ring and disk electrodes till as negative as −1.0 V vs. RHE, with stably high H2O2 selectivity up to 98% in the O2-saturated neutral electrolyte (3.6 wt% NaCl, 0.9 wt% NaCl, and 0.1 M Na2SO4) (Fig. 3d, e). Such stable 2e− ORR performance at low potentials, which is rarely achieved previously, clearly confirms the superiority of the material for selectively catalyzing 2e− ORR and suppressing HER. This is further proven by the fact that Ni(OH)2/CNT exhibits a high overpotential of over 1.0 V for the HER in all neutral media, according to the LSV curves in the Ar atmosphere (Fig. 3f), which is also consistent with the DFT results (Fig. 2c)31. In addition, Ni(OH)2/CNT also showed low catalytic activity towards H2O2 reduction reaction (H2O2RR, Supplementary Fig. 17) as well as the smallest Tafel slope (82.94 mV dec−1) among the three samples (Supplementary Fig. 18). These unique features of the material further guarantee its feasibility to be used at high overpotentials for achieving high-rate H2O2 generation while maintaining a high efficiency simultaneously.

H2O2 production performance in flow cells and applicational perspectives

A homemade prototype electrolyzer was subsequently constructed to evaluate the ORR-to-H2O2 performance of Ni(OH)2/CNT in neutral electrolytes for long-term operation. In this electrolyzer, Ni(OH)2/CNT was coated on a carbon paper with a gas-diffusion layer (GDL, SGL-39BB) and was used as the working electrode. Ag/AgCl and Ru/Ir-coated Ti mesh were used as the reference electrode and counter electrode. A Nafion film (117) was used as the separator (Supplementary Fig. 19). The accumulated H2O2 concentration in the electrolyte was quantified by the potassium oxalate colorimetric method according to the standard curve in Supplementary Fig. 20. In 1 M NaCl electrolyte, Ni(OH)2/CNT delivers a high FE of over 80% with increasing ORR current densities until 200 mA cm−2. In other electrolytes with lower NaCl concentrations, including 3.6 wt% NaCl (simulated seawater) and 0.9 wt% NaCl (physiological saline), the ORR current density could still exceed 100 and 50 mA cm−2 while maintaining a high Faraday efficiency of over 80% until −0.96 V vs. RHE, respectively (Fig. 4a and Supplementary Fig. 19a). Correspondingly, high H2O2 yields of 30, 18 and 10.86 mol g−1 h−1 (3, 1.8, 1.086 mmol cm−2 h−1) were achieved in 1 M NaCl, 3.6 wt% NaCl and 0.9 wt% NaCl electrolytes (Supplementary Fig. 21), respectively. In addition, Ni(OH)2/CNT also exhibits a high yield of 18.71 mol g−1 h−1 (1.871 mmol cm−2 h−1) at 100 mA cm−2 with FE of 80% at −0.96 V vs. RHE in 0.5 M Na2SO4 electrolyte (Supplementary Fig. 22). Besides, Ni(OH)2/CNT also has high yield in alkaline electrolyte (Supplementary Fig. 19a and Supplementary Figs. 23). These solidly validate the practical feasibility of electrocatalytic H2O2 synthesis under various conditions.

Fig. 4: H2O2 production performance of Ni(OH)2/CNT materials in the flow cell.
Fig. 4: H2O2 production performance of Ni(OH)2/CNT materials in the flow cell.
Full size image

a LSV polarization curve of Ni(OH)2/CNT in different concentration NaCl solutions and FE of Ni(OH)2/CNT in 1 M NaCl. b FE and linear fitted lines for the accumulation of H2O2 yields in 1 M NaCl for 156 h electrolysis. c The stability test of Ni(OH)2/CNT at 1 A cm−2 in 1 M NaCl. d The stability test of Ni(OH)2/CNT tested in natural seawater from Bohai Sea. e A comparison of stability reported with the most advanced electrocatalysts newly reported under various electrolysis configurations (cyan symbols: neutral electrolyte; pink symbols: alkaline electrolyte; green symbols: acid electrolyte. CoPc-CNT(O)27, Co-N5 SAC65, ZnCo-ZIF-C366, h-SnO229, CoNCB67, Pd/MCS-868, B-C69, JUC-60070, OCNS90071, Co1-NGO34, CoNOC37, CoTPP@RGO72, Data sources are provided in Table S2). f Techno-economic analysis of H2O2 electrosynthesis with three different conditions. g Bactericidal rate of S. aureus in 0.9 wt. % NaCl. h UV-vis spectra during the decomposition process of Rhodamine B.

The electrochemical stability of Ni(OH)2/CNT was examined at a constant current density of 200 mA cm−2. The FE of Ni(OH)2/CNT maintained over 80% during the whole 156 h testing period, with an average apparent H2O2 yield rate of 30.86 ± 0.03 mol g−1 h−1 (3.086 mmol cm−2 h−1), corresponding to a specific H2O2 yield up to 4830.43 mol gcat−1 during the whole test (Fig. 4b and Supplementary Figs. 19a and 24). Besides, as shown in Supplementary Figs. 25-27, the SEM, XRD, and XPS results of Ni(OH)2/CNT show that the morphology and structure of Ni(OH)2/CNT remain intact after the 100 h i-t test at 200 mA cm−2 in 1 M NaCl electrolyte, which indicates that Ni(OH)2/CNT structure is fairly stable. More importantly, at a high current density of 1 A cm−2, Ni(OH)2/CNT delivers a high H2O2 yield of 141 mol g−1 h−1 (14.1 mmol cm−2 h−1, FE = 75.58%) and a cell voltage (Ecell) of 6.4 V (non-iR-corrected)27 was required (Fig. 4c and Supplementary Figs. 19b and 28). This high current density dominates among state-of-the-art electrocatalysts for 2e− ORR (Supplementary Table 2). Moreover, in the natural seawater from Bohai Sea, Ni(OH)2/CNT also shows excellent electrochemical properties considering the high H2O2 yield of 14.2 mol g−1 h−1 (1.42 mmol cm−2 h−1) at 100 mA cm−2 with FE over 70% (Fig. 4d and Supplementary Fig. 29). In this test, 1 M H3PO4 is used as the anolyte to prevent the formation of precipitate in the seawater during the reaction (Supplementary Fig. 19c). Besides, Ni(OH)2/CNT has higher H2O2 yield (30 mol g−1 h−1) and FE (>80%) in 1 M NaCl compared with NiO/CNT (12.3 mol g−1 h−1, 50%), which highlights the excellent performance of Ni(OH)2/CNT (Supplementary Figs. 19a and 30). In addition, the stability of the material, tested at 200 mA cm−2 and 1 A cm−2, surpasses most of the recently published catalysts for electro-producing H2O2 (Fig. 4e and Supplementary Table 2), which clearly demonstrates the superiority of this material.

Techno-economic analysis was further conducted in flow cell to estimate the plant-gate levelized cost for H2O2 production using Ni(OH)2/CNT at current densities of 1 A cm−2 (US $901 tonne−1) and 200 mA cm-2 (US$1085.65 tonne−1) in 1 M NaCl and 100 mA cm−2 (US $882.77 tonne−1) in natural seawater, both of which are lower than the market price of H2O227,32 (US$ 1200 tonne−1, Fig. 4f, Supplementary Fig. 31, Supplementary Notes 1 and 2 contain details of parameters and calculation processes used in economic analyzes), which can be mainly attributed to its zero separation cost, cheap raw materials, and readiness for scaling up (Supplementary Fig. 32).

The unique feature of this technology endows it with significant feasibility for practical applications, including disinfection and wastewater treatment, etc. To verify this, S. aureus was selected, and it was shown that this technology can effectively eliminate S. aureus in 0.9 wt% NaCl solution after only 5 min of H2O2 electrosynthesis, with a 99.99% sterilizing ratio (Fig. 4g, Supplementary Fig. 33 and Supplementary Table 3). In addition, 1 M NaCl solution after 20 min of H2O2 electrosynthesis at 200 mA cm−2 can be directly utilized as a Fenton reagent under sunlight irradiation, which can effectively eliminate Rhodamine B in 15 min and Methylene Blue in 90 min (Fig. 4h, Supplementary Figs. 34 and 35), demonstrating the effectiveness of generated H2O2 for water treatment and disinfection33.

Of note, when NaCl solution was used for both electrodes, the anodic half-reaction was the dominant chlorine evolution reaction (CER), rather than the common oxygen evolution reaction, which was confirmed by the lower onset potential and large current density of Ru/Ir-coated Ti mesh anode in the NaCl solutions than that of in Na2SO4 solutions from the anodic LSV curves (Supplementary Fig. 36). In addition, the CER process would also generate value-added products, i.e., chlorine gas (Cl2), hypochlorous acid (HClO), and hypochlorite salts (ClO−), in which the ClO−/HClO is the major product of CER in brine and is another chemical for water purification/disinfection. To prove this, anolyte running at 200 mA cm−2 could oxidize the I− ions into I2, thus changing the transparent NaI solution into a dark yellow color (Supplementary Fig. 37), confirming the generation of ClO−. The 2e− ORR towards H2O2 production, coupled with the anodic CER reactions, would more effectively produce disinfectants and provide a possible method to alleviate serious marine pollution.

An original mechanism for 2e− ORR: the water dissociation and proton transfer in Ni(OH)2/CNT

Apart from tailoring the electronic structure of the catalysts to alter *OOH intermediates adsorption to achieve high ORR activity and H2O2 selectivity34, as previously calculated in a thermodynamics aspect (Fig. 2a), the kinetics of the *OOH intermediate formation is another key factor in 2e− ORR towards high-rate H2O2 electrosynthesis, which has often been neglected. The formation rate of *OOH intermediates depends on the proton supply rate. To further explore the 2e− ORR process in Ni(OH)2/CNT, DFT calculations were performed to obtain the kinetic barrier of proton transfer and water dissociation.

As shown in Fig. 5a, the kinetic barrier of proton migration via hydroxyl groups on Ni(OH)2 to the adjacent adsorbed oxygen molecules was calculated to be 0.35 eV at 0.7 V vs. RHE (the equilibrium potential of H2O2). When the applied potential was reduced to 0.5 V vs. RHE, the proton transfer barrier further decreased to 0.12 eV, indicating a highly facile proton transfer on our material (Supplementary Fig. 38 and Supplementary Data 1). However, the kinetic barrier of proton migration from H2O in NiO at 0.7 V vs. RHE was calculated to be 1.21 eV (Supplementary Fig. 39), much higher than that of Ni(OH)2/CNT (0.35 eV), indicating that efficient proton transfer from nearby hydroxy groups to the absorbed O2 can thus reduce the formation energy of intermediate *OOH as well. Besides, the energy barriers of water dissociation on Ni(OH)2 surface were calculated to be 0.44 eV (Supplementary Fig. 40 and Supplementary Data 1). In 2e− ORR process, OOH is formed when O2 adsorbed at unsaturated Ni sites attracts nearby hydrogen from hydroxy groups, Ni-O sites were exposed when the hydroxy hydrogen was consumed. Then, the OH group and H atom in the H2O molecule were attracted by the unsaturated Ni sites and Ni-O sites on Ni(OH)2/CNT, respectively. This simultaneous absorption of O and H atoms in the water molecule destabilizes the H-OH bond23, which would greatly facilitate the dissociation of water30. Meanwhile, Ni-OH reforms due to the Ni-O site traps hydrogen and Ni-OH− forms due to the Ni sites traps OH from water molecule, respectively. The Ni-OH species can provide sufficient proton supply for catalytic oxygen reduction, which accelerates the protonation of absorbed O2 to form *OOH species during ORR, in a streamline manner (Fig. 5b and Supplementary Data 1). Subsequently, the OH− at Ni-OH− strips off into the electrolyte because of the repulsive effect of negative potential applied to cathode30,35, which recombines with a proton from the anode side to form a water molecule. In this regard, continuous proton generation and their migration to the absorbed O2 molecules can be achieved, which is favorable for the *OOH formation and subsequent H2O2 formation17.

Fig. 5: Theoretical insight for excellent ORR performances.
Fig. 5: Theoretical insight for excellent ORR performances.
Full size image

DFT calculations of the Ni(OH)2 interface. a I, Proton transfer (PT) at U = 0.5 V vs. RHE and U = 0.7 V vs. RHE. II, III, Water dissociation process on the outer surface of Ni(OH)2 and the optimized geometries of the initial, transition and final state. IV, *OOH protonation for 2e– ORR pathway (*OOH + H→H2O2, solid line) and *OOH dissociation for 4e– ORR pathway (*OOH + H→*O + H2O, dotted line) of Ni(OH) at U = 0.5 V and U = 0.7 V vs. RHE. The pink and yellow represent oxygen and hydrogen in water, respectively. b The optimized structure of ORR process. The blue, pink, white and yellow represent nickel, oxygen in Ni(OH)2 and hydrogen in Ni(OH)2 and H in water, respectively. c In situ ATR-SEIRAS spectra for Ni(OH)2/CNT and NiO/CNT at different potentials in O2-saturated 3.6 wt. % NaCl solution. d Intensity ratio of OOH and O2 from ATR-SEIRAS. e FT-IR spectra of Ni(OH)2/CNT before and after oxygen reduction reaction, respectively.

In addition, avoiding the cleavage of O-O bond in the adsorbed *OOH intermediate is another important requirement for H2O2 generation via 2e− ORR36. Therefore, the kinetic barriers of *OOH protonation for 2e– ORR pathway (*OOH + H→H2O2) and *OOH dissociation for 4e– ORR pathway (*OOH + H→*O + H2O)16,37 over Ni(OH)2 surface were calculated to better understand the origin of the excellent selectivity of H2O2 electrosynthesis. For the 2e− ORR pathway, Ni(OH)2 shows low kinetic barriers of 0.46 and 0.58 eV at 0.5 and 0.7 V vs. RHE (described by the Ni-O2 bond elongation until breaking), respectively. For the 4e− ORR pathway to produce H2O (described by the O-O bond elongation until breaking in *OOH intermediate), however, obviously higher kinetic barriers of 0.65 and 0.73 eV at 0.5 and 0.7 V vs. RHE were obtained, respectively (Supplementary Fig. 41 and Supplementary Data 1). It thus indicates the binding of the O-O bond is stronger than that of the Ni-O bond, allowing the key *OOH intermediate to be well reserved for forming H2O2. It also clearly confirms the advantages of Ni(OH)2 in stabilizing O2* adsorbents and preventing the over-protonation of OOH* intermediates, thereby leading to a highly selective and rapid H2O2 production38,39.

To verify these simulation results, in-situ attenuated total reflectance surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS) and ex-situ Fourier transform infrared spectroscopy (FTIR) were performed for Ni(OH)2/CNT and NiO/CNT40. As shown in Fig. 5c, both materials exhibit peaks at 1120 and 1450 cm−1 from 0.5 V to 0 V vs. RHE during the ORR process, and these two prominent peaks could be assigned to the surface-adsorbed OOH (*OOH) and O2 species41,42,43, respectively. It is clear that the intensities of the peak at 1450 cm−1 (i.e., adsorbed O2) for both samples are comparable, while the intensities at 1120 cm−1 (i.e., adsorbed OOH) for Ni(OH)2/CNT are much higher than that for NiO/CNT at all testing potentials (Fig. 5d), indicating the higher *OOH formation rate on Ni(OH)2/CNT44,45. This result agrees well with the above simulation conclusions that the efficient proton transfer and water dissociation accelerate the formation of *OOH intermediates. Besides, ex-situ FTIR further elucidates the role of Ni(OH)2 as proton donors, which was conducted after ORR in the O2-saturated 3.6 wt% NaCl D2O solution. As shown in Fig. 5e, a new peak occurs at 2600–2700 cm−1, which can be attributed to the OD groups46, confirming that the H atoms in the hydroxy groups of Ni(OH)2 have been partially replaced by D atoms from the dissociation of D2O. It thus can be deduced that the protons obtained from water dissociation supplement the protons in the hydroxy group in Ni(OH)2, which was consumed for the protonation of absorbed O2 to OOH species during ORR, in a pipeline manner, as shown in path a–c in Fig. 1. This accelerates the formation of *OOH intermediates and is conducive to achieving a rapid H2O2 electrosynthesis.

Discussion

In this work, we present a Ni(OH)2-based catalyst with Ni(OH)2 nanoplates grown on carbon nanotubes for efficient and rapid H2O2 electrosynthesis in salt water. In this material, Ni(OH)2 shows a unique capability of swift proton transfer and rapid water dissociation, even under neutral conditions, ensuring a sufficient proton supply to the Ni active sites for rapid *OOH formation and H2O2 generation. Meanwhile, the strong inhibition against the undesirable 4e− ORR/HER and excellent chloride resistance of the material collaboratively guarantee a high 2e− ORR selectivity in Cl−-containing brine. Consequently, the as-prepared Ni(OH)2/CNT exhibits a high H2O2 selectivity consistently exceeding 85% in salt water over a wide potential range of −1 to 0.3 V vs. RHE. In 1 M NaCl electrolyte, it exhibits high H2O2 yield rate (30.86 mol g−1 h−1 @ 200 mA cm−2; 141 mol g−1 h−1 @ 1 A cm−2), long stability of 156 h at 200 mA cm−2 with FE over 80%, resulting in a specific H2O2 yield up to 4830.43 mol g.−1. Besides, in the natural seawater from Bohai Sea, this catalyst exhibits a high H2O2 yield (14.2 mol g−1 h−1 @ 100 mA cm−2) with FE over 70%, enabling it to directly synthesize H2O2 without the requirement of post-synthesis separation. This study also provides an alternative pathway for 2e− ORR catalyst design oriented by enhancing the water dissociation kinetics.

Methods

Materials

All chemicals were utilized without further purification. Ni(NO3)2·6H2O (AR, 98.0%), Co (NO3)2·6H2O (AR, 98.0%), FeCl3·6H2O (AR, 99.0%), KOH (AR, 95.0%), HNO3 and Na2SO4 (AR, 85%) were purchased from Aladdin Reagents. NaCl (GR, 99.8%) were procured from Macklin (AR Analytical reagent, GR Guaranteed reagent). Nafion (5.0 wt%) was procured from Sigma-Aldrich. Nafion-117 membrane and Pt-C (20.0 wt%) were acquired from Johnson Mattey.

Synthesis of O-CNT

The reaction flask equipped with reflux condenser, magnetic stirrer and thermometer was mounted in the preheated oil bath. Next, predetermined quantities of raw MWCNTs (2.5 g) and nitric acid (200 ml) were added into a two-necked, round-bottomed glass flask at 80 °C for 12 h. The sample was filtered on a membrane filter, washed to neutral pH, dried at 60 °C for 24 h to get about 1 g O-CNTs47.

Synthesis of Ni(OH)2/CNT

2.5 mmol Ni(NO3)2·6H2O (726.963 mg) and 30 mg O-CNTs were dissolved in 30 ml deionized water. After slow addition of 10 mL 1 M KOH aqueous solution, the mixed suspension was transferred to a 50 ml polytetrafluoroethylene inner liner, and reacted at 180 °C for 12 h. The solid obtained was washed with a large amount of deionized water, and dried at 60 °C overnight to get about 229 mg Ni(OH)2/CNT.

Synthesis of NiO/CNT

2.5 mmol Ni(NO3)2·6H2O (726.963 mg) and 30 mg O-CNTs were dissolved in 30 ml deionized water. After slow addition of 10 mL 1 M KOH aqueous solution, the mixed suspension was transferred to a 50 ml polytetrafluoroethylene inner liner, and reacted at 180 °C for 12 h. The solid obtained was washed with a large amount of deionized water, and dried at 60 °C overnight to get 1; To synthesize the NiO/CNT, the obtained 1 was placed in the middle of the quartz crucible and calcined at 500 °C with a heat rate of 5 °C/min under flowing Ar and maintained at this temperature for 2 h in a muffle furnace and NiO/CNT was obtained.

Synthesis of Ni(OH)2

2.5 mmol Ni(NO3)2·6H2O (726.963 mg) was dissolved 30 ml deionized water. After quick addition of 10 mL 1 M KOH aqueous solution, the mixed suspension was transferred to a 50 ml polytetrafluoroethylene inner liner, and reacted at 180 °C for 12 h. The green solid obtained was washed with a large amount of deionized water, and dried at 60 °C overnight to get Ni(OH)2.

Characterization

Hitachi S-4800 scanning electron microscopy (SEM), Hitachi Jem-2100F transmission electron microscopy (TEM) were used for the morphological and structural characterizations. The composition of the samples and elemental mapping was analyzed by energy-dispersive X-ray spectroscope (EDX) attached to the TEM instrument. Bruker D8 Advanced X-ray diffraction (XRD) with Cu Ka radiation (k = 1.5418 Å) was used for analyzing the crystal structure of as-prepared materials. X-ray photoelectron spectroscopy (XPS, Thermo Scientific ESCALAB 250Xi) with Al Ka X-rays was utilized to investigate the chemical state of the elements of the as-prepared products. XAFS spectra were measured at the beamline BL14W1 station of the Shanghai Synchrotron Radiation Facility. The Ni K-edge XANES data were recorded in transmission mode with Ni foil, standard Ni(OH)2 as references. The EXAFS raw data were background-subtracted, normalized, and Fourier transformed by standard procedures using the Athena and Artemis implemented in the IFEFFIT software packages. The existence of O-vacancy was analyzed by EPR spectra (Bruker EMXplus).

Electrochemical tests

The electrochemical measurements were conducted at room temperature (~25 °C). All electrochemical tests were conducted on the CHI 760E workstation with an RRDE-3A Rotating Ring Disk Electrode (RRDE) Apparatus (ALS Co, Ltd, Japan), using a typical three-electrode system at room temperature. An RRDE loaded with catalysts was employed as the working electrode. An Ag/AgCl electrode and a carbon rod were used as the reference electrode and counter electrode, respectively. 1 mg of catalyst was dispersed in a mixture of 1 mL of 700 µL deionized water, 200 µL isopropanol and 100 µL PTFE (0.2 wt%) solution to form the ink (1 mg/mL). After that, 20 µL of catalyst (catalyst loading: 0.808 mg cm−2) was drop-casted on the disk electrode (area: 0.2475 cm2), and the H2O2 produced during the reaction was measured by the ring electrode (area: 0.1866 cm2). Cyclic voltammetry (CV) was performed in 0 to 1.0 V vs. RHE for ~10 cycles before linear scanning voltammetry (LSV) testing to activate the catalysts with a scan rate of 10 mV s −1 and rotation rate of 1600 revolutions per minute (rpm). The scan voltage range of LSV in 0.9 wt% NaCl and 0.1 M Na2SO4 is from 0.6 to −0.35 V vs. RHE, in 3.6 wt% NaCl is from 0.6 to −1.1 V vs. RHE. the scan voltage range in 0.1 M KOH, 0.1 M KOH + 3.6 wt% NaCl, seawater + 0.1 M KOH is from 0.9 to 0 V vs. RHE. Ring electrode potential was set at 1.2 V vs. RHE to detect the amount of generated H2O2. At this voltage, only H2O2 formed at the disk electrode is oxidized without affecting the ORR current at the ring electrode. The H2O2 selectivity (H2O2%) and transfer electron number (n) of the catalysts were calculated based on the current of ring (IR, mA) and disk (ID, mA):

$${H}_{2}{O}_{2}\%=\frac{2{I}_{R}/N}{{I}_{D}+({I}_{R}/N)}\,\times 100\%$$
(1)
$$n=4 \times \frac{{I}_{D}}{{I}_{D}+({I}_{R}/N)}$$
(2)

Where N is the collection efficiency, which was calibrated by the reversible [Fe(CN)6]4−/[Fe(CN)6]3− redox couple, N = 0.37

CER was measured by using single electrolyzer, in which Ru/Ir-coated Ti mesh (commercial) was used as the working electrode, Ag/AgCl and platinum plate as the reference electrode and counter electrode. The H2O2 reduction performance was analyzed in the Ar-saturated 3.6 wt% NaCl solution with 10 mM H2O2.

H2O2 yield measurements

H2O2 generated by continuous ORR electrolysis in the electrolyte can be accumulated and quantified. Electrochemical synthesis of H2O2 and quantification of H2O2 production rate were performed in a flow cell. 20 µL catalysts ink of 5 mg/mL (5.0 mg catalyst powder in the mixture of 700 uL of deionized water, 200 µL isopropanol and 100 μL of 0.2 wt% PTFE solution) was dripped on the gas-diffusion layer (GDL, SGL-39BB), which was used as the cathode electrode with a catalyst loading of 0.1 mg cm−2, Ru/Ir-coated Ti mesh (commercial, 2 × 2 cm2), and Ag/AgCl electrode were used as the counter electrode and reference electrode, respectively. Nafion117 membrane (thickness: 183 µm) was used to separate the cathode from the anode chamber. The membrane was treated by protonating it in a 5 wt% aqueous H2O2 solution at 80 °C for 1 h then boiling it in a 5 wt% H2SO4 at 80 °C for 1 h. Next, washing it several times with deionized water, then it was soaked in deionized water. The H2O2 concentration in the final electrolyte was determined by the potassium titanium oxalate colorimetric method. Standard solutions were prepared with 100 µL H2O2 (variable concentrations), 0.5 mL 0.05 M potassium titanium oxalate, 0.5 mL 3 M H2SO4, and 3.9 mL DI water. Absorbance at 400 nm (UV-vis) was fitted with H2O2 concentration to generate the calibration curve48. The LSV test at different potentials from 0.6 to −1.4 V vs. RHE and the i-t test at 200 mA cm−2 current densities. The FE of H2O2 generation in the flow cell was calculated as follows:

$${FE}\left(\%\right)=\frac{2{cVF}}{Q}\times 100\%$$
(3)

Where c and V are the H2O2 concentration calculated and the volume of the electrolyte of the cathode, F is the Faraday constant and Q is the total charge passed.

In situ ATR-SEIRAS measurements

In situ ATR-SEIRAS was performed on a Nicolet iS50 spectrometer equipped with a VeeMax III (PIKE technologies) accessory. The electrochemical test was conducted in a custom-made three-electrode electrochemical single cell. A Pt wire (commercial) and a saturated Ag/AgCl were used as the counter and reference electrodes, respectively. 0.1 mg electrocatalyst was sprayed on a Si wafer coated with Au layer. This Si wafer was used to load the catalysts and served as the working electrode. The in-siut ATR-SEIRAS spectrum were recorded by varying the potential stepwise from 0.5 V to 0 V vs. RHE in O2-saturated 3.6 wt% NaCl and run time was 60 s.

Density functional theory (DFT) calculations

Spin-polarized DFT calculations were performed with the Vienna Ab Initio Simulation Package (VASP)49. The exchange-correlation interaction was described using Perdew-Burke-Ernzerhof functional50. The Projector Augmented Wave potentials were used to treat electronic cores51. The plane-wave cutoff energy of geometry optimizations was set to 450 eV. All the slab were separated with a 15 Å vacuum in z-direction. The k-points of 3 × 2 × 1 Gamma-centered grids were used during the optimization. The convergence of atomic relaxations was applied with the max force of 0.02 eV Å–1.

The computational hydrogen electrode (CHE) model52 was applied to calculate free energies of absorbed species:

$$G=E+{ZPE}-{TS}$$
(4)

Where \(E\) is the total energy computed by DFT optimization. \({ZPE}\) and \({TS}\) represent zero-point energy correction and entropy correction.

The climbing-image nudged elastic band (CI-NEB)53 was used to calculate kinetic barriers of chemical reaction process, convergence thresholds of energy and force were 1 × 10–4 eV and 0.05 eV Å–1, respectively. The implicit solvent effect was applied by using VASPsol with the linearized Poisson-Boltzman model54. The Debye screening length was adjusted to 3.0 Å, and the relative permittivity was set to 78.4. In this work, we employed the python code developed by Duan’s group for GC-DFT calculations. Different voltage was applied through the change of electrons using the counter-charge distribution55,56,57, The potential of the electrode was calculated by correlating the work function (Wf) of the system with the experimental Wf of the standard hydrogen electrode (SHE). The potential−dependent electrochemical energy can be expressed as58,59

$$E={E}_{{vasp}}+{e\phi }_{{{{\rm{elyte}}}}}{\Delta }_{q}-\left({\mu }_{e}+{e\phi }_{{{{\rm{elyte}}}}}\right){\Delta }_{q}={E}_{{vasp}}-{E}_{F}{\Delta }_{q}$$
(5)

Where \({E}_{{vasp}}\) is the DFT energy determined by VASPsol, \({\phi }_{{{{\rm{elyte}}}}}\) signifies the inherent electrostatic potential within the bulk electrolyte. \({E}_{F}\) represents Fermi energy, also coincides with the electron chemical potential \({\mu }_{e}\), \({\Delta }_{q}\) represents the extra numbers of electrons. The relation between \({E}_{F}\) and the corresponding electrode potential referenced to the standard hydrogen electrode scale U (V vs SHE) is formulated as:

$$U\left({{{\rm{V}}}}\;{vs}.\;{{{\rm{SHE}}}}\right)=-4.6V-\left({E}_{F}+{E}_{{Femishift}}\right)/e$$
(6)

Where \({E}_{{Femishift}}\) is the Femi energy shift given by VASPsol.

Molecular dynamics (MD) simulations

Molecular dynamics (MD) simulations were carried out by utilizing the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS)60 for the purpose of exploring the ion distributions on the catalyst surfaces of Ni(OH)2, Ni, and NiO. During the entirety of the simulations, the system was run at a temperature of 300 K within the NVT ensemble for a period of 4 ns to achieve equilibrium. A variable electric field in the Z direction was applied to each atom, with the selected voltage values being 0, 0.1, and 1 V Å⁻¹ respectively. Periodic boundary conditions were applied within the horizontal plane. To simulate the electrolyte employed in the experiment (3.6 wt% NaCl), the solvent model consisted of 1500 H₂O molecules, 28 Na⁺ ions, and 28 Cl⁻ ions. The interatomic interactions between the catalysts and H₂O as well as various ions were described by means of Coulomb and Lennard-Jones potentials. The corresponding parameters for Na⁺ and Cl⁻ were developed by In Suk Joung and Thomas E Cheatham61. The interaction among H₂O molecules was depicted by the SPC/E model62. The IFF interaction parameters of Ni(OH)₂, Ni, and NiO were referred to following previous reports63. Furthermore, the detailed simulation models of both Ni(OH)₂, Ni, and NiO catalysts are shown in Supplementary Figs. 2 and 3. The cutoff distance was set to 10 Å for both of the aforesaid potentials, and the time step was 1 fs. All the configurations were visualized via the use of OVITO software64.

Statistics & reproducibility

No data were excluded from the analyses.