English

• 1. Introduction

  Owing to its strong oxidizing properties across a wide pH range (E⁰ = 1.763 V at pH 0 and E⁰ = 0.878 V at pH 14), hydrogen peroxide (H₂O₂) is extensively used in disinfection, bleaching, water treatment, and chemical synthesis [1-3]. The global market demand for H₂O₂ is projected to exceed $6.2 billion by 2026 [4]. Currently, industrial H₂O₂ is mainly synthesized by a multistep anthraquinone oxidation process, which accounts for >95% of total market production [5]. Nevertheless, this method consumes a substantial amount of energy and relies on costly palladium-based hydrogenation catalysts [6]. Additionally, practical applications require vacuum distillation and purification steps to achieve high H₂O₂ concentrations, resulting in additional expenses and safety concerns during storage and transportation [7,8]. Alternatively, H₂O₂ can be synthesized directly on a small scale using catalysts such as Pd, Pt, and other metals or alloys through the reaction of H₂ and O₂ [9,10]. However, the mixture of H₂ and O₂ poses explosion risks, limiting its widespread applications [11,12]. Moreover, the produced H₂O₂ still demands additional storage and transportation procedures for practical use. Photocatalytic H₂O₂ production on polymeric carbon nitrides is also a promising method for green H₂O₂ synthesis but is still limited by low activity and efficiency [13-17]. Actually, for most industries, on-site production and decentralized application of H₂O₂ are preferred [18]. Consequently, there is a growing interest in developing “on-demand” methods for H₂O₂ production and exploring benign operating conditions [19].

  Electrochemical oxygen reduction reaction (ORR) through a 2e⁻ transfer process offers a sustainable and mild method for on-site H₂O₂ production from O₂ under ambient conditions [20-22]. However, when compared with the competitive 4e⁻ pathway (O₂ + 4H⁺ + 4e⁻ → 2H₂O, E⁰ = 1.23 V vs. RHE), the 2e⁻ pathway of O₂ (O₂ + 2H⁺ + 2e⁻ → H₂O₂, E⁰ = 0.70 V vs. RHE) demonstrates a more negative standard potential, making it thermodynamically less favorable. To this end, advanced catalysts with high selectivity for H₂O₂ electrosynthesis are highly desired [23]. To date, numerous catalysts have been investigated, including precious metals (e.g., PtHg₄, PdAu) [24,25], carbon-based materials (e.g., O-CNTs, N-FLG) [26,27], metal-organic frameworks (MOFs, e.g., Mg₃(HITP)₂, Ni MOF NSs) [28,29], transition metal oxides (e.g., ZnO, Ti₂O₃) [30,31], dichalcogenides (e.g., CoSe₂, Ni₂Mo₆S₈) [32-34], and borides (e.g., Ni₃B) [35]. Notably, these catalysts exhibit different reaction behaviors under different pH conditions. For instance, carbon-based materials excel in the selective reduction of O₂ to produce H₂O₂ in alkaline media but demonstrate limited electrocatalytic ability in acidic environments [36]. In contrast, precious metal-based materials still dominate in acidic conditions [37]. These disparities arise from variations in the reaction mechanisms under different pH conditions. Furthermore, the dissociation constant of H₂O₂ is 11.69 (pK_a1) at 25 ℃ [38], implying that H₂O₂ can exist as a molecule in acidic or neutral media, whereas in alkaline environments, it may undergo deprotonation to form HO₂⁻ [39]. Hence, the pH level of the electrolyte significantly influences H₂O₂ electrosynthesis. In terms of practical applications, different pH conditions cater to various fields. For example, alkaline media are suitable for textile and paper pulp bleaching applications. In comparison, H₂O₂ has a wider range of applications in acidic and neutral environments, such as electro-Fenton degradation of diverse pollutants in acidic conditions and bacteria killing, water disinfection, value-added chemical synthesis in neutral conditions [40-42]. Overall, electrochemical H₂O₂ production in different pH conditions is a complex, multifactorial process, and its application is a cross-disciplinary, flexible field.

  Apart from catalyst engineering and the pH effect, the production capacity of H₂O₂ is another crucial factor for its industrial applications [43]. In laboratory research, rotating ring-disk electrode (RRDE) and H-type cell setups are initially adopted to evaluate the selectivity and yield of H₂O₂, respectively [44]. However, the low solubility of O₂ in the electrolyte and the limited diffusion of O₂ to the electrode surface hinder high current densities and continuous decentralized H₂O₂ production. To address this issue, flow-type cells (assembled with a gas-diffusion electrode (GDE)), dual-membrane flow cells (with a solid electrolyte (SE)), and membrane-free gas-liquid separated cells were subsequently proposed to meet the diverse requirements of industrialization. These requirements include direct use without purification, membrane-free reactors to reduce costs, and streamlined production processes [45-48].

  Thus far, diverse catalysts and devices have been developed, and considerable progress has been achieved. However, the design principles and correlations of catalysts under different pH conditions have not been systematically studied. Moreover, the primary obstacle impeding the practical on-site application of H₂O₂ is still the insufficient capacity for H₂O₂ production. Therefore, more highly efficient and cost-effective catalysts and devices should be exploited to ameliorate these situations. Previous reviews have reported diverse two-electron ORR (2e⁻ ORR) catalysts and relevant modulation strategies. For example, Zhang et al. introduced recent progress made in different types of metal-free carbon-based catalysts from the perspective of dimensions, including three-dimensional, two-dimensional, one-dimensional, and zero-dimensional [23]. Moreover, various strategies, including heteroatom doping, structural engineering, and defect engineering, are examined for their role in enhancing catalytic efficiency. Yang et al. summarized recent methodologies (including tuning metal center sites, engineering surface functionalization and coordination environments) and achievements of single-atom catalysts (SACs), particularly M-N-C catalysts, for the selective 2e⁻ ORR to produce H₂O₂ [18]. Combined with theoretical computation and characterization, they elucidate the synthesis−structure−property correlation and highlight the basic properties that impact activity and selectivity, followed by a perspective on the effective strategies to address the stability issue during H₂O₂ production. Differentially, this review focuses on the differences in catalytic mechanisms and design principles at different pH conditions, as well as the most salient features of the production devices. The aim is to provide insights into the design and synthesis of an effective 2e⁻ ORR electrocatalyst with high selectivity, activity and yield, and to promote the development and adoption of advanced devices for scaling up the utilization of H₂O₂ and customized applications. The content of this review can be divided into three parts, as shown in Fig. 1. First, using 2e⁻ ORR under different pH conditions as a clue, we survey and analyze recent advances in catalytic mechanisms, design principles, and the potential applications of relevant catalytic systems. Second, in response to diverse on-site application demands, we elaborate on the evolution of H₂O₂ electrosynthesis devices from laboratory scale to practical implementation, explaining their advantages and disadvantages, which are beneficial for their upgrading and customized application. Third, we propose the remaining challenges and future opportunities for achieving industrial-level H₂O₂ production. We hope this review will be helpful in guiding the design and discovery of innovative and efficient catalysts toward substantial H₂O₂ electrosynthesis.

  Figure 1

  

  Figure 1.  Schematic diagram of catalytic mechanism and design principles for 2e⁻ ORR under different pH conditions and device engineering for H₂O₂ electrosynthesis.

  DownLoad: Full-Size Img PowerPoint

  2. Design principles and reaction mechanisms of catalysts under different pH conditions

  2.1 In acidic electrolyte

  2.1.1   Mechanism of acidic 2e⁻ ORR

  Owing to the decent stability and diverse applications of H₂O₂ in acidic media, acidic H₂O₂ electrosynthesis holds great promise for industrial implementation. For instance, acidic H₂O₂ electrosynthesis can proceed directly in mature proton-exchange membrane (PEM) devices and electro-Fenton installations, which are valuable for chemical transformation and environmental treatment, respectively [49,50]. Its broad applicability has also facilitated the development of 2e⁻ ORR electrocatalysts, including noble metals and SACs, in acidic environments [51-54]. In principle, 2e⁻ ORR involves the transfer of two electrons (Fig. 2a) [55], and consists of a two-step proton-coupled electron transfer process in aqueous solutions at acidic pH, which can be represented by Eq. 1:

  (1)

  (1a)

  (1b1)

  (1b2)

  Figure 2

  

  Figure 2.  (a) Pathway for 4e⁻ and 2e⁻ O₂ electroreduction. Reproduced with permission [55]. Copyright 2019, American Chemical Society. (b) Volcano plot for oxygen reduction to H₂O₂. The red solid line and blue solid line represent the theoretically calculated oxygen reduction volcano plot for the 4e⁻ and 2e⁻ ORR, with the limiting potential plotted as a function of ΔG_HO* (lower horizontal axis) and ΔG_HOO* (upper horizontal axis). In addition, η means the thermodynamic overpotential for a specific reaction. Reproduced with permission [24]. Copyright 2013, Springer Nature. (c) Possible configurations of O₂ adsorption on the catalyst surface. Reproduced with permission [7]. Copyright 2023, Springer Nature. (d) Screened metal atoms with their most stable bulk structures and a schematic illustration of all the considered SACs. Reproduced with permission [55]. Copyright 2019, American Chemical Society.

  DownLoad: Full-Size Img PowerPoint

  where * is the unoccupied surface binding site and *OOH is a critical intermediate. Eq. 1a represents the first step of proton-coupled electron transfer process to generate *OOH intermediate, and Eqs. 1b1 and 1b2 indicate that the generated *OOH undergoes a 2e⁻ pathway to produce H₂O₂ (associative 2e⁻ pathway) or a 4e⁻ pathway to produce H₂O (associative 4e⁻ pathway), respectively.

  Alternatively, the initial O−O bond in the O₂ molecule may be directly broken (dissociative pathway), leading to the formation of two *O species (without the intermediate *OOH formation). These *O species can then gain protons and electrons to ultimately form H₂O. In the context of the 2e⁻ ORR and 4e⁻ ORR, key intermediates *OOH and *OH follow a linear scaling relationship (typically ΔG_HOO* = ΔG_HO* + 3.2 eV), resulting in the creation of 2e⁻ and 4e⁻ ORR volcano-type plots (Fig. 2b) [24]. As shown in Fig. 2b, it becomes evident that the 2e⁻ ORR activity is primarily determined by the adsorption energy of *OOH (ΔG_HOO*). The equilibrium potential of the volcano peak is 0.70 V, which corresponds to a catalyst with the optimal binding strength of *OOH intermediate. At the left segment of the volcano plot, the catalyst exhibits stronger binding to *OH and *OOH species, favoring the 4e⁻ reduction pathway to produce H₂O. At the right segment of the volcano plot, the catalyst has a weaker binding to *OOH, which tends to enhance selectivity but reduce activity in the formation of H₂O₂. Guided by this volcano plot, the catalyst situated at the apex is defined as the ideal material with the highest electrocatalytic activity for 2e⁻ ORR [56,57]. Furthermore, from this free energy diagram, we can discern the propensity of different catalysts to undergo the desired 2e⁻ or competing 4e⁻ reduction pathway, allowing us to screen the most promising candidates for selective H₂O₂ production. To achieve optimal ΔG_HOO*, the O₂ adsorption mode is a crucial consideration. Typically, there are three recognized ways for oxygen adsorption, i.e., Griffiths-type, Yeager-type and Pauling-type (Fig. 2c) [7]. The 4e⁻ pathway is typically favored through a “side-on” adsorption of O₂, known as the Griffiths-type, where the O₂ molecule adsorbs side-on to a single metal atom. Another pathway, known as the Yeager-type, involves the bridge adsorption of the O₂ molecule on two adjacent metal atoms, also promoting a 4e⁻ reaction. This is primarily because both pathways involve the dissociation of the O-O bond. In contrast, the “end-on” adsorption of O₂, known as Pauling-type, prefers a 2e⁻ pathway to produce the desired H₂O₂.

  2.1.2   Electrocatalysts for acidic H₂O₂ production

  Over the past decades, platinum group metals (PGMs) have been intensively studied for their appropriate adsorption energy and mode for the 4e⁻ pathway, establishing them as state-of-the-art 4e⁻ ORR catalysts [58]. To maximize the 2e⁻ reaction route, it is essential to manipulate the local structure of PGMs and adjust the type of O₂ adsorption to the “end-on” mode. Accordingly, various strategies have been employed, including alloying with inert metals (e.g., Hg, Ag, and Au) and incorporating nonmetal elements (including O, S, Se, and P) to tailor the accessibility of active sites [59-61]. For example, by using Au as the inert metal, Zhao et al. designed bimetallic PdAu nanoframes that exhibit remarkable H₂O₂ selectivity (>90%) and rapid electro-Fenton degradation rates for Rhodamine B within minutes [25]. Moreover, with Se as the nonmetal element, Huang and co-workers fabricated amorphous PdSe₂ nanoparticles as highly selective catalysts for the electrochemical synthesis of H₂O₂ [62]. Density functional theory (DFT) calculations implied that the low-coordinated Pd sites could optimize the adsorption of oxygenated intermediates and maintain the integrity of the O—O bond, contributing to the high selectivity of 92%−98% at 0.0 − 0.3 V (vs. RHE).

  Nevertheless, the essential high cost and scarcity of precious metals constrain their large-scale applications. Hence, researchers shifted their focus to SACs, which could offer ~100% utilization of metal atoms as active sites [63-66]. Although previous studies have reported a variety of SAC synthesis methods, more studies still remain on trial-and-error methods. To this end, Wang’s group comprehensively summarized and compared the preparation methods of these trial-and-error synthesis strategies, including spatial confinement strategy, defect engineering strategy, coordination site design strategy, and atomic layer deposition strategy, putting forward some professional opinions, and aiming at guiding more rational and intelligent catalyst design and synthesis. Moreover, in light of the rapid development of SACs and the continuous expansion of application fields, they summarized the applications and research progress of SACs in different fields, such as environment catalysis, electrocatalysis, organic synthesis, photocatalysis, batteries, sensors, and enzyme catalysis. In addition, the design strategies and structure−activity relationships of SACs for specific reactions were summarized and analyzed, which are conducive to the personalized design and application of SACs [67,68].

  Moreover, there is a significant difference between SACs and bulk metals, which can be understood through the interplay between the geometrical effect and the ligand/electronic effect. Essentially, according to Eq. 1b, to achieve high catalytic activity, the adsorption of *OOH should be enhanced. By contrast, to obtain high selectivity, the adsorption of *O (the product of *OOH dissociation) should be reduced. On a metal surface, *OOH and *O intermediates normally adsorb on atop sites and hollow sites, respectively. Conversely, for a SAC, only atop sites are available for adsorption due to the discontinuous and isolated atomic site. This characteristic helps SACs specifically stabilize the *OOH and destabilize *O intermediates, breaking the linear scaling relationship between ΔG_HOO* and ΔG_HO*. This allows for superior 2e⁻ ORR selectivity and activity to be achieved synchronously. Given this circumstance, a large number of SACs, such as CoNOC [69], Co₁−NG(O) [70], penta-coordinated O−Co−N₂C₂ [71], oxygen-rich MesoC−Co nanosheets [72], β-substituted CoPorF/CNT [73], CB@Co−N−C [74], CoIn−N−C [75], Ni−N₂O₂/C [76], and Mo₁/OSG-H [77], have been exploited, and remarkable 2e⁻ ORR selectivity and activity are realized.

  Furthermore, by modifying SACs with different structural motifs, it may be possible to break the linear scaling relation and achieve more glorious electrocatalytic performances. For instance, Guo et al. constructed and screened 210 SAC models by rationally anchoring a series of single metal atoms on several representative and experimentally achievable substrates (Fig. 2d) [55]. It was elucidated that for SACs with metal atoms in groups 3 to 10, increasing the number of D-electrons of metal atoms tends to weaken the interaction between the metal atom active sites and *O, thereby improving the selectivity of H₂O₂ electrosynthesis. The host matrix and local coordination environment were found to dramatically modify the electronic structure and adsorption capability of the metal center in SACs, consequently altering the activity and selectivity of these catalysts. Notably, it was revealed that macrocyclic structures outperformed graphene- and BN-based materials in terms of overall performance. Collectively, microstructure and electrocatalytic performance are closely correlated, and by modifying the metal center, local coordination configuration, or host matrix of SACs, exceptional performances may be achievable.

  2.2 In alkaline electrolyte

  2.2.1   Mechanism of alkaline 2e⁻ ORR

  In the 1930s, Berl et al. pioneered H₂O₂ electrosynthesis in alkaline media using activated carbon cathode via 2e⁻ ORR [78]. Since then, this technology, known as Huron−Dow process, has been continuously developed and successfully implemented in the production of dilute alkaline H₂O₂ solutions used for textile and paper pulp bleaching. Essentially, the reaction equation and electron pathways for 2e⁻ ORR in an alkaline solution are illustrated in Eq. 2:

  (2)

  (2a)

  (2b)

  here, * represents the catalyst surface site. The process starts with O₂ on the catalyst surface protonating to form *OOH, which then combines with an electron to form HO₂⁻. Notably, that in alkaline media, H₂O₂ exists primarily as HO₂⁻. The proper absorption and desorption of *OOH are crucial for facilitating the desorption of HO₂⁻. Consequently, similar to acidic conditions, the selectivity of 2e⁻ ORR in alkaline environments is also contingent upon the adsorption energy of the *OOH intermediate. Furthermore, the preservation of the O-O bond on the catalyst surface site plays a decisive role in the selective H₂O₂ electrosynthesis under alkaline conditions.

  2.2.2   Electrocatalysts for alkaline H₂O₂ production

  To counter this situation, a series of catalysts have been studied, and diverse research strategies have been adopted to fine-tune the absorption and desorption balance of *OOH [79-81]. For example, carbon-based materials were initially considered due to their relatively weak interaction with *OOH, which promotes the desorption of HO₂⁻ and results in decent selectivity for H₂O₂ production [39,82]. Nevertheless, they still faced a hurdle in terms of a high O₂ adsorption barrier. Modification methods, including oxygen functionalization, defect engineering and heteroatom doping of carbon-based materials, are emphasized in Supporting information (see the modification methods of carbon-based materials and Fig. S1 for details).

  Aside from carbon-based materials, other materials such as metal oxides [83,84], dichalcogenides [34,85], and borides [35,86], have also demonstrated attractive electrocatalytic behaviors under alkaline conditions. By comparison, it was noticed that under different pH conditions, different types of catalysts are preferred for H₂O₂ production. For instance, Yang et al. observed that glassy carbon (GC) showed both high activity and selectivity in alkaline conditions, but its activity was limited in acidic environments (Fig. 3a) [44]. The Pourbaix diagram, which depicts the thermodynamically stable state of the catalyst surface as a function of pH and electrode potential in an electrochemical system, may provide an explanation for this phenomenon. Specifically, when using metal-free carbon as a catalyst, it has been demonstrated that catalyst configurations with higher oxygen content exhibit greater stability under high pH and potential conditions, whereas configurations with higher hydrogen are more stable under low pH and potential conditions (Fig. 3b) [87]. Hence, catalysts with oxygen-related groups usually tend to exhibit superior electrocatalytic performance in alkaline and neutral electrolytes [88].

  Figure 3

  

  Figure 3.  (a) ORR of GC in 0.1 mol/L HClO₄ and 0.1 mol/L KOH. Reproduced with permission [44]. Copyright 2018, American Chemical Society. (b) Pourbaix diagram showing the most energetically favorable surface species configurations as a function of reaction conditions of electrode potential and pH. The k-O, c-O and COOH represent the ketone oxygen, CO group, and COOH group, respectively. The digits before k-O, c-O and COOH represent the amount of corresponding oxygen configurations doped into the carbon substrate in relevant structural models. A larger digit represents more oxygen configurations, and less hydrogen content. Reproduced with permission [87]. Copyright 2020, American Chemical Society. (c) Schematic illustration of the double-layer structure during ORR in acidic (left) and alkaline (right) conditions. Insets (ⅰ) and (ⅱ) represent the inner sphere pathway in acidic and alkaline conditions, respectively, while inset (ⅲ) represents the outer sphere pathway in an alkaline condition. Reproduced with permission [89]. Copyright 2011, American Chemical Society.

  DownLoad: Full-Size Img PowerPoint

  In addition to considering the stability of the functional groups under reaction conditions, it is crucial to examine the electron transfer mechanisms that occur under different pH conditions, as these mechanisms can significantly impact the activity of metal-based catalysts. Ramaswamy et al. investigated the effect of traversed pH conditions (from acidic to alkaline media) on the adsorption behavior of reactants, intermediates, and products. They discovered that changes in pH led to variations in adsorption strength and the location of species adsorption within the double-layer structure (Fig. 3c) [89]. Specifically, two mechanisms, namely, the inner sphere electron transfer process and the outer sphere electron transfer process, have been proposed in relation to the ORR process. The inner-Helmholtz plane (IHP) is composed of adsorbed hydroxyl species, chemisorbed O₂, and solvent water dipoles. By contrast, the outer-Helmholtz plane (OHP) contains solvated alkali metal ions and solvated O₂ cluster O₂·(H₂O)_n in alkaline media. In the case of a typical inner-sphere electron transfer process, the electrons are first transferred to O_{2, ads} to form superoxide (O₂^•−)_ads. Subsequently, the (O₂^•−)_ads intermediates are protonated by direct proton transfer (in the acidic condition, inset (ⅰ) process) or indirect proton transfer from water molecules (in the alkaline condition, inset (ⅱ) process). However, elevated pH under alkaline conditions causes a decrease in the rate of water activation, making the proton transfer process more challenging to proceed. Therefore, an outer-sphere electron transfer process is proposed in alkaline media (inset (ⅲ) process). For an outer-sphere electron transfer process, the electron transfer process occurring in the outer sphere enables the noncovalent interaction between the solvated O₂ cluster and the OH_ads adsorbed on electrode surface. This interaction leads to the formation of HO₂⁻ anion, followed by electrostatic repulsion between the adsorbed OH_ads and HO₂⁻, resulting in the desorption of HO₂⁻ anion into the electrolyte. Under the circumstances, the nonspecific adsorption of OH_ads on the surface of nonnoble metals and metal oxides enables them to be promising candidates for 2e⁻ ORR under alkaline conditions. Importantly, in alkaline media, the potential of the working electrode decreases compared with acidic conditions, resulting in excess charge on the electrode surface. This excess charge favors the existence of HO₂⁻ intermediate, thereby contributing to superior 2e⁻ ORR activity.

  2.3 In neutral electrolyte

  Electrocatalysts for neutral H₂O₂ production: Compared with acidic and alkaline conditions, the neutral solution is noncorrosive. This is particularly advantageous for the direct production and application of pure H₂O₂ without the need for further neutralization procedures. For instance, the generated H₂O₂ can be used directly in rainwater treatment and organic synthesis [90,91]. The reaction mechanism in neutral electrolytes is similar to that in acidic electrolytes, so here we focus on the progress of 2e⁻ ORR electrocatalysts, with a particular emphasis on some unusual examples. As described in Section 2.1, isolated active sites can facilitate the selective formation of H₂O₂. Therefore, researchers have explored noble metal catalysts with discontinuous active sites first. To improve the isolation of noble metal atoms, Li et al. introduced phosphorus (P) elements and prepared PtP₂ nanocrystals (NCs) [92]. Thanks to the optimized adsorption of the key intermediate *OOH, the PtP₂ NCs, functioning as cathodic catalysts, played an important role in the neutral polymer electrolyte membrane fuel cell (PEMFC), achieving a high H₂O₂ production rate of 2.26 mmol h⁻¹ cm⁻² and long-term stability lasting up to 120 h at a high current density of 150 mA/cm² (Fig. 4a).

  Figure 4

  

  Figure 4.  (a) Time-dependent neutral H₂O₂ concentration measured at a constant potential of 0.4 V (vs. RHE) for 120 h. The accumulated H₂O₂ concentration was continuously cycled through the system (600 mL). Reproduced with permission [92]. Copyright 2020, Springer Nature. (b) Illustration of the 2e⁻ ORR to H₂O₂ over L-ZnO. (c) DOS of S- and L-ZnO absorbed *OOH. Reproduced with permission [30]. Copyright 2023, Royal Society of Chemistry. (d) Schematic structures of metastable P-hcp Ni. The different blue balls are Ni in different layers and the pink ball is P. (e) H₂O₂ yields of different catalysts in 1.0 mol/L Na₂SO₄ by using the flow cell device. Reproduced with permission [99]. Copyright 2023, Wiley–VCH. (f) Schematic illustration of 2e⁻ ORR toward H₂O₂ in acid electrolyte with or without Na⁺ under industrial-relevant current. Reproduced with permission [102]. Copyright 2022, Springer Nature.

  DownLoad: Full-Size Img PowerPoint

  In recent years, there has been a growing interest in metal compounds as alternative catalysts to the extensively studied noble metal and carbon-based catalysts for 2e⁻ ORR [28,93-96]. To enhance the 2e⁻ ORR activity of these metal compounds, tremendous efforts have been dedicated to tailoring their geometric and phase structures. From a geometrical perspective, the reduction in size from microscale to nanoscale and, eventually, to the cluster/single-atom level can increase the exposure of catalytic sites, thereby boosting H₂O₂ production activity [97,98]. Interestingly, an abnormal size effect in ZnO nanoplates was observed by Chen’s group [30]. As depicted in Figs. 4b and c, theoretical calculations and in situ Raman spectra revealed that, compared with small-thickness ZnO (S-ZnO), the d-band of large-thickness ZnO (L-ZnO) shifted toward the Fermi level. The shift stabilized the crucial reaction intermediate, ultimately resulting in ~100% selectivity and achieving ampere-level current density (1 A/cm²). In addition to the geometric effect, phase transformation has proven effective in enhancing the intrinsic activity of active sites. As shown in Figs. 4d and e, Geng et al. prepared a phosphorus-optimized metastable hexagonal-close-packed phase nickel catalyst (P-hcp Ni) [99]. This catalyst utilized P atoms as stabilizers for the metastable hcp phase while also attracting electrons from the surrounding Ni to weaken the adsorption of *OOH and contribute to a remarkable H₂O₂ productivity of 4255.9 mmol g_Ni⁻¹ h⁻¹ in a neutral electrolyte in a flow cell. Analogously, Kang and co-workers reported a metastable hexagonal SnO₂ (h-SnO₂) nanoribbon with a space group of P63/mmc (194) [100]. DFT analyses showed that the h-SnO₂ (1−210) surface, primarily the side face of the nanoribbons, had an appropriate adsorption energy for *OOH. Thus, the h-SnO₂ electrocatalysts could trigger H₂O₂ synthesis in 0.1 mol/L Na₂SO₄ with a high selectivity of 99.99% and yield of 3885.26 mmol g⁻¹ h⁻¹ at 0 V (vs. RHE).

  Notably, due to the continued consumption of protons during the ORR process, there is an alkaline shift in the local pH value of the electrode surface. Interestingly, Gyenge et al. discovered that the elevation in the local pH level could significantly improve the relevant current efficiency, increasing it from 12% to 61% in alkaline electrolyte and from 14% to 55% in acidic electrolyte [101]. Specifically, by using trialkylmethylammonium chloride (A336) as a cationic surfactant to displace protons in the electric double layer, theoretical estimations have indicated that the pH of the electrode surface can increase from 0.9 to a range of 2.0−9.4. These results indicate that by modulating the protons on the electrode surface, significantly enhanced electrocatalytic performance may be achieved. Correspondingly, Zhang et al. employed a cation-regulated interfacial engineering approach to augment 2e⁻ ORR, achieving an industrial-level rate under acidic conditions [102]. Molecular dynamic simulations have shown that in the presence of alkali metal cations, a “shielding effect” occurs, which dilutes the protons at the catalyst/electrolyte interface, thereby inhibiting the electrochemical dissociation of H₂O₂ to H₂O (Fig. 4f). Of course, if strictly neutral conditions are required, a buffer electrolyte solution is essential for evaluating 2e⁻ ORR at a neutral pH value, as proposed by Jin and Schmidt et al. [93]. Overall, understanding the local proton concentration and pH change at the electrode/electrolyte interface can provide valuable insights into molecular and mesoscopic engineering approaches for more efficient 2e⁻ ORR catalysis.

  3. Device engineering for H₂O₂ bulk generation and utilization

  In addition to the research on the development of robust electrocatalysts, there is an equally relevant pursuit of engineering proper electrochemical devices for on-site bulk generation of H₂O₂, since the “real-scenario” production of H₂O₂ is more conducive to versatile applications. In laboratory research, the RRDE setup is usually adopted for the expeditious evaluation of potential catalysts (Figs. 5a and b) [103]. Typically, with continued stirring, oxygen is transported to the GC disk electrode to occur in 2e⁻ and 4e⁻ ORR. Subsequently, the produced H₂O₂ is detected and quantified at the concentric platinum ring electrode. The applied potential for this process is set at 1.2 − 1.3 V (vs. RHE) [104]. For bulk production and storage of H₂O₂, it is conducted in an H-type cell, which features two distinct electrolyte chambers. The accumulated H₂O₂ can be quantified using the potassium permanganate (KMnO₄)-based titration method, the ceric sulfate (Ce(SO₄)₂)-based ultraviolet-visible (UV–vis) spectroscopy method, or the potassium titanium(Ⅳ) oxalate (K₂TiO(C₂O₄)₂)-based UV–vis spectroscopy method. In the KMnO₄ assay, the MnO₄⁻ anion can be reduced to Mn²⁺ by H₂O₂ (2MnO₄⁻ + 5H₂O₂ + 6H⁺ → 2Mn²⁺ + 8H₂O + 5O₂), resulting in a change in solution color from dark purple to colorless. In contrast, the (Ce(SO₄)₂)-based UV–vis spectroscopy method may provide more accurate results owing to the stable reaction between Ce(SO₄)₂ and H₂O₂ (2Ce⁴⁺ (yellow) + H₂O₂ → 2Ce³⁺ (colorless) + 2H⁺ + O₂). In this approach, H₂O₂ concentration is quantitatively measured by comparing the absorbance of Ce⁴⁺ at a wavelength of 320 nm based on their linear correlation relationship. For the potassium titanium(Ⅳ) oxalate method, colorless titanium(Ⅳ) is oxidized by H₂O₂ to an orange titanium(Ⅳ)-peroxide complex in the presence of sulfuric acid. Subsequently, the concentration of H₂O₂ can be measured at a wavelength of 400 nm [105,106].

  Figure 5

  

  Figure 5.  Schematics for (a) a RRDE, (b) an H-cell configuration with a catalyst deposited on a solid substrate in a nonflowing electrolyte and (c) a continuous flow cell with a catalyst deposited on a GDE. Reproduced with permission [103]. Copyright 2020, American Chemical Society. (d) Schematic diagram of a GDE used for H₂O₂ electrosynthesis. Reproduced with permission [86]. Copyright 2022, Wiley–VCH. (e) Dual-membrane cell with pure H₂ and O₂ streams separately introduced to the anode and cathode, respectively. Reproduced with permission [91]. Copyright 2019, Springer Nature. (f) Membrane-free cell consisting of a flow channel. Reproduced with permission [114]. Copyright 2017, Royal Society of Chemistry.

  DownLoad: Full-Size Img PowerPoint

  However, the low solubility of O₂ in the electrolyte and the limited diffusion of O₂ to the surface of submerged electrodes hinder the high current density and continuous decentralized H₂O₂ production [107]. To address these limitations, GDE (assembled in flow-type cells) was developed as an alternative electrode. As shown in Fig. 5c, the GDE consists of a porous layer containing a hydrophobic component, and the catalyst is deposited onto the GDE at the solid-liquid interface. This enables the direct delivery of oxygen to the catalyst surface without being restricted by the gas transport limitation [108,109]. During electrocatalysis, the generated H₂O₂ product is directly dissolved in the flowing electrolyte and steadily circulated within the electrode compartment. This reduces the concentration of H₂O₂ on the electrode surface, prevents the accumulation or decomposition of H₂O₂, and increases the total concentration of H₂O₂ in the electrolyte, resulting in a high concentration of H₂O₂ solution. For instance, in a recent study, Wu et al. reported a GDE coated with amorphous NiB₂ [86], and the assembled flow cell achieved a high H₂O₂ yield of 4.753 mol g⁻¹ h⁻¹ and long-term operation stability of 12 h at 125 mA/cm² (Fig. 5d). Analogously, Zhang et al. operated a natural air diffusion electrode and reached a high H₂O₂ rate of 6.72 mol g_catalyst⁻¹ h⁻¹ in neutral solution [110]. The above devices are all based on GDEs. In addition to the GDE, Zhou et al. have reported an innovative “floating” electrode, in which one side of the electrode is open to the air and the other side is submerged in the aqueous solution [111]. This special approach enables the formation of a three-phase boundary (gaseous O₂, aqueous electrolyte, and solid cathode), thus allowing synergistic utilization of O₂ from both ambient air and the electrolyte. Thus, a high H₂O₂ yield of 61.7 mg/dm³ was obtained using the “floating” electrode, which is 4.3 times higher than a conventional “submerged” cathode. Additionally, An et al. have proposed an anti-flooding air-breathing cathode (ABC, the anti-flooding herein refers to the prevention of catalyst flooding), which was prepared by a simple rolling-spraying method [112]. Then this ABC was amplified and integrated with the Ti/IrO₂-based anode to form a modular electrode system. In a 2 L prepilot scale system, the modular electrode could produce 322 ± 15.6 mg L⁻¹ h⁻¹ H₂O₂ with a current efficiency of 84.7%. Nevertheless, the H₂O₂ produced using this method is dissolved in the electrolyte and may still require additional purification steps for certain specific applications.

  To address this issue, Xia et al. reported on a two-type membrane flow cell that incorporated SE layers for the production of pure H₂O₂ solution [91]. As depicted in Fig. 5e, a porous SE layer was positioned between the cation and anion exchange membrane layers. This configuration facilitated ionic recombination, allowing H⁺ and HO₂⁻ ions to cross from the anode and cathode, respectively, resulting in the direct production of H₂O₂. Subsequently, flushing the SE layer with deionized water produced a pure H₂O₂ aqueous solution. With this device, it could achieve a high H₂O₂ selectivity of >90% at 200 mA/cm² and a decent yield of 3.4 mmol cm⁻² h⁻¹ during continuous operation lasting up to 100 h. Additionally, Wei’s group investigated the feasibility of this two-type membrane flow cell in coupling with organic oxidation [113]. In their experiment, the SE layer was mixed with an organic catalyst, titanium silicalite-1 (TS-1), and the generated H₂O₂ was captured by TS-1 to form Ti-OOH, an intermediate species with high oxidation activity. Ultimately, the organic compound underwent selective oxidization to obtain the desired value-added product.

  While employing the two-type membrane SE approach can significantly reduce product crossover and avoid energy-intensive separation processes, the introduction of additional membranes leads to increased internal resistance within the device, thereby reducing production efficiency. Moreover, the polymer electrolyte membrane is susceptible to degradation by free radicals due to the self-decomposition of H₂O₂. In response to this question, Chen et al. adopted a membrane-free gas−liquid separated device in which a flowing electrolyte is placed between the cathode and anode [114]. The cathode, featuring a hydrophobic polymer on the backside, was coated with mesoporous carbon (CMK-3) capable of reacting with O₂ flow (Fig. 5f). Subsequently, the produced H₂O₂ was captured by the flowing electrolytes within the chamber. Analogously, Wang’s group also reported a membrane-free H₂O₂ electrosynthesis flow cell [115]. Utilizing oxidized carbon nanotube and polytetrafluoroethylene polymer-decorated carbon fiber paper as the cathode and anode catalysts, respectively, this cell delivered a high cell efficiency of 153% and a large current of 50.4 mA at a cell potential of 1.7 V. This is mainly because the membrane-free cell allows for a decrease in cell voltage and overall ohmic loss, thereby opening up opportunities for practical applications [116].

  Overall, in the above-mentioned cells, the early H-type cell is difficult to meet the industrial requirements for H₂O₂ production due to the low solubility of O₂/air in an aqueous solution. In recent years, with the development of flow-type cells and the rational use of the key component GDE, the mass transfer rate of gaseous reactants has significantly improved. Meanwhile, by adjusting the reaction area of the flow cell, the shape of the gas flow channel, and the gas/electrolyte flow rate, the 2e⁻ ORR efficiency has also been significantly improved. Thus, the current density for H₂O₂ production was increased from tens of milliamps (mA) to amperes (A), thus realizing a significant breakthrough in H₂O₂ production performance. For example, Lin et al. reported a CoN₄/VG catalyst containing cobalt single atoms on vertically aligned graphene nanosheet assemblies [117]. It delivered a H₂O₂ productivity of 706 mmol g_catalyst⁻¹ h⁻¹ in an H-cell setup, whereas in a gas-diffusion flow reactor, the associated productivity was considerably increased to 4000 mmol g_catalyst⁻¹ h⁻¹, confirming the high efficiency of the flow cell. Similarly, Zhang and colleagues prepared a carbon-supported Ni^Ⅱ single-atom catalyst with Ni−N₂O₂ tetra-coordination (Ni−N₂O₂/C), which could be stabilized in a flow cell for 15 h at 70 mA/cm², far exceeding the behavior in an H-cell (8 h at 1 mA/cm²) [76]. However, these new cells also present challenges in terms of device complexity and cost. Take the dual-membrane solid-electrolyte cell as an example; the GDEs (gas diffusion and catalyst layers), membranes (anion-exchange membrane and cation-exchange membrane), and solid polymer proton conductors are tightly compressed within a flow field plate [91,113]. Although the desired concentration of H₂O₂ can be obtained by setting parameters such as external gas input, liquid circulation, and operating temperature, the complexity of the components and parameterization makes operation and maintenance considerably more difficult. From this point of view, membrane-free cells can simplify operating conditions and reduce costs, and have been adopted in ethanol oxidation [118] and water splitting [119]. In addition, the membrane-free system can regulate the fluid in a laminar flow state with a low Reynolds number to prioritize convective transport over diffusive transport, thus avoiding oxidation of the as-prepared H₂O₂ on the anode [120]. Furthermore, compressing the individual membrane-free cells together to form a stack can further increase H₂O₂ productivity and reduce operating costs.

  4. Conclusions and outlooks

  In this review, we summarized recent advances in electrocatalysts and electrocatalytic devices for the 2e⁻ ORR to produce H₂O₂, aspects of catalytic mechanisms in variable pH conditions, strategies for the rational design of electrocatalysts, and practical device engineering for scaling up H₂O₂ generation. We anticipate that the proposed mechanisms and insights may aid in the design and development of innovative catalysts and devices, thereby promoting large-scale H₂O₂ manufacturing. Overall, pH control plays an essential role in H₂O₂ production in different application environments. For example, alkaline H₂O₂ solutions can be used for textile and paper pulp bleaching, and the corresponding technology can be extended to anion-exchange membrane fuel cells to enable in-situ production of H₂O₂ while generating electrical energy. Therefore, the alkaline 2e⁻ ORR catalytic system has been extensively investigated in previous reports. In acidic media, H₂O₂ has powerful oxidizing ability and wide applicability. Correspondingly, acidic H₂O₂ electrosynthesis can in turn be conducted in Fenton installations and mature PEM devices for environmental treatment and chemical transformation, respectively, with great application prospects. In neutral conditions, H₂O₂ can be used in the green organic synthesis of value-added chemicals, including epoxidation of alkenes, alcohol or thioether selective oxidation, and phenol hydroxylation. As for the design strategy of catalysts under different pH conditions, the proton concentration or degree of water activation, the adsorption strength of O₂, and the existence of H₂O₂ (the H₂O₂ molecule or HO₂⁻ anion) all have significant effects. Specifically, under acidic and neutral conditions, separated active sites are preferred to achieve the end-on adsorption mode of O₂, thereby promoting a 2e⁻ ORR pathway. Moreover, the adsorption energy of the *OOH intermediate should be optimized to accelerate the desorption of H₂O₂. While in alkaline media, the outer-sphere transfer process should be considered first owing to the abundant adsorbed OH⁻ on the catalyst surface. To this end, non-precious metals and their oxides have become attractive alternatives for 2e⁻ ORR. Despite significant progress and the development of diverse electrocatalysts and devices (Table S1 in Supporting information), several challenges remain related to H₂O₂ selectivity, activity, yield, and stability under different pH conditions, as well as practical and on-site applications.

  First, a promising catalyst should exhibit high catalytic ability, applicability, low cost, and environmental friendliness. Currently, the reported 2e⁻ ORR catalysts are primarily focused on carbon-based materials and metal oxides in alkaline electrolytes, whereas the alkali-catalyzed decomposition of H₂O₂ limits their widespread applications. Therefore, more attention should be shifted toward the development of 2e⁻ ORR catalysts in acidic and neutral solutions to promote in-situ and stable applications of H₂O₂. Electro-Fenton degradation is an emerging application of H₂O₂ in acidic conditions, where H₂O₂ can combine with Fe²⁺ to generate hydroxyl radicals (·OH) with strong oxidizing capacity for subsequent environmental treatment applications [93,121]. Additionally, the green organic synthesis of value-added chemicals is an essential application of H₂O₂ in neutral media [122-124]. Moreover, apart from the modulation strategies mentioned in this review, pore structure, hybrid/heterogeneous configuration, and hydrophilic/hydrophobicity have significant effects on H₂O₂ electrosynthesis. Therefore, it is necessary to select the appropriate catalyst type and modulation strategy comprehensively, obeying the reaction mechanisms and characteristics under different pH conditions, to achieve flexible regulation and customized applications.

  Second, due to the pH- and catalyst type-dependent electrocatalytic properties of catalysts, along with complex operating conditions, it remains challenging to explore pH-universal 2e⁻ ORR catalysts and elucidate an explicit catalytic mechanism under tunable pH conditions. Confronted with this situation, first-principles theory and crucial characterization technologies (e.g., synchrotron radiation, operando Raman spectroscopy, operando attenuated total reflectance infrared spectroscopy, and machine learning) may allow us to delve deeper into the adsorption and desorption behaviours of intermediates on catalyst surfaces under operating conditions and to identify the real active sites. This will help us gain a thorough understanding of the pH-dependent 2e⁻ ORR performances and facilitate the development of decentralized H₂O₂ synthesis, as well as its subsequent applications in chemistry, energy, and the environment fields.

  Finally, the focus of H₂O₂ electrosynthesis should be on device engineering and the efficiency and durability of practical on-site applications. This can refer to well-established mature reaction systems and architectures, such as flow-type devices with GDEs and H₂−O₂ fuel cells. To achieve industrial-scale H₂O₂ yields, a porous, three-dimensional electrode (decorated with catalysts of appropriate thickness) is required to allow O₂ gas transport. Fluid flow systems are also needed to rinse away the generated H₂O₂ from the electrode surface for in-situ or next-step reactions. Additionally, the anodic half-reaction can be paired with other reactions (e.g., alcohol oxidation, urea oxidation, and furfural oxidation) to construct an overall electrocatalytic platform, enabling the production of high−value chemical products and maximizing energy conversion efficiency.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Xuyun Lu: Conceptualization, Data curation, Investigation, Writing – original draft. Yanan Chang: Data curation, Investigation. Shasha Wang: Validation. Xiaoxuan Li: Validation. Jianchun Bao: Funding acquisition, Project administration, Supervision. Ying Liu: Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing.

  Acknowledgment

  This work was supported by the National Natural Science Foundation of China (Nos. 22102073, 22075147).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110277.

  
  
• 1. [1]

     J. Tang, T. Zhao, D. Solanki, et al., Joule 5 (2021) 1432–1461. doi: 10.1016/j.joule.2021.04.012

  2. [2]

     J. Xie, L. Zhong, X. Yang, et al., Chin. Chem. Lett. 35 (2024) 108472. doi: 10.1016/j.cclet.2023.108472

  3. [3]

     J.M. Campos-Martin, G. Blanco-Brieva, J.L.G. Fierro, Angew. Chem. 118 (2006) 7116–7139. doi: 10.1002/ange.200503779

  4. [4]

     H.W. Kim, M.B. Ross, N. Kornienko, et al., Nat. Catal. 1 (2018) 282–290. doi: 10.1038/s41929-018-0044-2

  5. [5]

     J.M. Campos-Martin, G. Blanco-Brieva, J.L. Fierro, Angew. Chem. Int. Ed. 45 (2006) 6962–6984. doi: 10.1002/anie.200503779

  6. [6]

     S.C. Perry, D. Pangotra, L. Vieira, et al., Nat. Rev. Chem. 3 (2019) 442–458. doi: 10.1038/s41570-019-0110-6

  7. [7]

     Y. Tian, D. Deng, L. Xu, et al., Nano-Micro Lett. 15 (2023) 122. doi: 10.1007/s40820-023-01067-9

  8. [8]

     A. Byeon, W.C. Yun, J.M. Kim, J.W. Lee, Chem. Eng. J. 456 (2023) 141042. doi: 10.1016/j.cej.2022.141042

  9. [9]

     Y. Bu, Y. Wang, G.F. Han, et al., Adv. Mater. 33 (2021) 2103266. doi: 10.1002/adma.202103266

  10. [10]

      R. Ciriminna, L. Albanese, F. Meneguzzo, M. Pagliaro, ChemSusChem 9 (2016) 3374–3381. doi: 10.1002/cssc.201600895

  11. [11]

      K. Jiang, J. Zhao, H. Wang, Adv. Funct. Mater. 30 (2020) 2003321. doi: 10.1002/adfm.202003321

  12. [12]

      Y. Wang, G.I.N. Waterhouse, L. Shang, T. Zhang, Adv. Energy Mater. 11 (2020) 2003323.

  13. [13]

      J. Ma, X. Peng, Z. Zhou, Y. Shen, Y. Zhang, Chin. Chem. Lett. 34 (2023) 108784. doi: 10.1016/j.cclet.2023.108784

  14. [14]

      H. Tan, P. Zhou, M. Liu, et al., Nat. Synth. 2 (2023) 557–563. doi: 10.1038/s44160-023-00272-z

  15. [15]

      Z. Teng, Q. Zhang, H. Yang, et al., Nat. Catal. 4 (2021) 374–384. doi: 10.1038/s41929-021-00605-1

  16. [16]

      J. Ma, X. Peng, Z. Zhou, et al., Angew. Chem. Int. Ed. 61 (2022) e202210856. doi: 10.1002/anie.202210856

  17. [17]

      J. Ma, C. Peng, X. Peng, et al., J. Am. Chem. Soc. 146 (2024) 21147–21159. doi: 10.1021/jacs.4c07170

  18. [18]

      X. Yang, Y. Zeng, W. Alnoush, et al., Adv. Mater. 34 (2022) 2107954. doi: 10.1002/adma.202107954

  19. [19]

      Y. Wen, T. Zhang, J. Wang, et al., Angew. Chem. Int. Ed. 61 (2022) e202205972. doi: 10.1002/anie.202205972

  20. [20]

      Z. Zhou, Y. Kong, H. Tan, et al., Adv. Mater. 34 (2022) 2106541. doi: 10.1002/adma.202106541

  21. [21]

      Y. Zhou, G. Chen, J. Zhang, J. Mater. Chem. A 8 (2020) 20849–20869. doi: 10.1039/d0ta07900f

  22. [22]

      D. Ouyang, D. Gao, J. Hong, Z. Jiang, X. Zhao, J. Energy Chem. 79 (2023) 135–147. doi: 10.1016/j.jechem.2022.10.036

  23. [23]

      D. Zhang, E. Mitchell, X. Lu, et al., Mater. Today 63 (2023) 339–359. doi: 10.1016/j.mattod.2023.02.004

  24. [24]

      S. Siahrostami, A. Verdaguer-Casadevall, M. Karamad, et al., Nat. Mater. 12 (2013) 1137–1143. doi: 10.1038/nmat3795

  25. [25]

      X. Zhao, H. Yang, J. Xu, T. Cheng, Y. Li, ACS Mater. Lett. 3 (2021) 996–1002. doi: 10.1021/acsmaterialslett.1c00263

  26. [26]

      Z. Lu, G. Chen, S. Siahrostami, et al., Nat. Catal. 1 (2018) 156–162. doi: 10.1038/s41929-017-0017-x

  27. [27]

      L. Li, C. Tang, Y. Zheng, et al., Adv. Energy Mater. 10 (2020) 2000789. doi: 10.1002/aenm.202000789

  28. [28]

      K. Dong, J. Liang, Y. Wang, et al., ACS Catal. 12 (2022) 6092–6099. doi: 10.1021/acscatal.2c00819

  29. [29]

      M. Wang, X. Dong, Z. Meng, et al., Angew. Chem. Int. Ed. 60 (2021) 11190–11195. doi: 10.1002/anie.202100897

  30. [30]

      S. Ding, B. Xia, M. Li, et al., Energy Environ. Sci. 16 (2023) 3363–3372. doi: 10.1039/d3ee00509g

  31. [31]

      Y. Yao, H. Wang, K. Dong, et al., J. Mater. Chem. A 11 (2023) 22154–22160. doi: 10.1039/d3ta05092k

  32. [32]

      L. Zhang, J. Liang, L. Yue, et al., Nano Res. 15 (2022) 304–309. doi: 10.1007/s12274-021-3474-0

  33. [33]

      X.L. Zhang, X. Su, Y.R. Zheng, et al., Angew. Chem. Int. Ed. 60 (2021) 26922. doi: 10.1002/anie.202111075

  34. [34]

      F. Xia, B. Li, Y. Liu, et al., Adv. Funct. Mater. 31 (2021) 2104716. doi: 10.1002/adfm.202104716

  35. [35]

      F. Ma, S. Wang, X. Liang, et al., Appl. Catal. B 279 (2020) 119371. doi: 10.1016/j.apcatb.2020.119371

  36. [36]

      S. Xu, Y. Gao, T. Liang, L. Zhang, B. Wang, Chin. Chem. Lett. 33 (2022) 5152–5157. doi: 10.1016/j.cclet.2022.01.057

  37. [37]

      J.Y. Zhang, C. Xia, H.F. Wang, C. Tang, J. Energy Chem. 67 (2022) 432–450. doi: 10.1016/j.jechem.2021.10.013

  38. [38]

      J. Liu, Z. Gong, M. Yan, et al., Small 18 (2022) 2103824. doi: 10.1002/smll.202103824

  39. [39]

      Y. Pang, H. Xie, Y. Sun, M.M. Titirici, G.L. Chai, J. Mater. Chem. A 8 (2020) 24996–25016. doi: 10.1039/d0ta09122g

  40. [40]

      Q. Yang, W. Xu, S. Gong, et al., Nat. Commun. 11 (2020) 5478. doi: 10.1038/s41467-020-19309-4

  41. [41]

      Z. Chen, J. Wu, Z. Chen, et al., Angew. Chem. Int. Ed. 61 (2022) e202200086. doi: 10.1002/anie.202200086

  42. [42]

      W. Liu, J. Feng, R. Yin, et al., Chem. Eng. J. 430 (2022) 132990. doi: 10.1016/j.cej.2021.132990

  43. [43]

      X. Zhang, Y. Xia, C. Xia, H. Wang, Trends Chem. 2 (2020) 942–953. doi: 10.1016/j.trechm.2020.07.007

  44. [44]

      S. Yang, A. Verdaguer-Casadevall, L. Arnarson, et al., ACS Catal. 8 (2018) 4064–4081. doi: 10.1021/acscatal.8b00217

  45. [45]

      B. Sabri Rawah, M. Albloushi, W. Li, Chem. Eng. J. 466 (2023) 143282. doi: 10.1016/j.cej.2023.143282

  46. [46]

      A. Huang, R.S. Delima, Y. Kim, et al., J. Am. Chem. Soc. 144 (2022) 14548–14554. doi: 10.1021/jacs.2c03158

  47. [47]

      M. Li, H. Lan, X. An, et al., Appl. Catal. B 339 (2023) 123125. doi: 10.1016/j.apcatb.2023.123125

  48. [48]

      M.A. Modestino, D.Fernandez Rivas, S.M.H. Hashemi, J.G.E. Gardeniers, D. Psaltis, Energy Environ. Sci. 9 (2016) 3381–3391. doi: 10.1039/C6EE01884J

  49. [49]

      Z. Wei, H. Xu, Z. Lei, et al., Chin. Chem. Lett. 33 (2022) 920–925. doi: 10.1016/j.cclet.2021.07.006

  50. [50]

      H. Sheng, A.N. Janes, R.D. Ross, et al., Nat. Catal. 5 (2022) 716–725. doi: 10.1038/s41929-022-00826-y

  51. [51]

      J.H. Kim, D. Shin, J. Lee, et al., ACS Nano 14 (2020) 1990–2001. doi: 10.1021/acsnano.9b08494

  52. [52]

      X. Zhao, Y. Liu, J. Am. Chem. Soc. 143 (2021) 9423–9428. doi: 10.1021/jacs.1c02186

  53. [53]

      K. Liu, P. Chen, Z. Sun, et al., Nano Res. 16 (2023) 10724–10741. doi: 10.1007/s12274-023-5823-7

  54. [54]

      Y. Cheng, H. Wang, H. Song, et al., Nano Res. Energy 2 (2023) e9120082. doi: 10.26599/nre.2023.9120082

  55. [55]

      X. Guo, S. Lin, J. Gu, et al., ACS Catal. 9 (2019) 11042–11054. doi: 10.1021/acscatal.9b02778

  56. [56]

      A. Kulkarni, S. Siahrostami, A. Patel, J.K. Nørskov, Chem. Rev. 118 (2018) 2302–2312. doi: 10.1021/acs.chemrev.7b00488

  57. [57]

      V. Viswanathan, H.A. Hansen, J. Rossmeisl, J.K. Nørskov, J. Phys. Chem. Lett. 3 (2012) 2948–2951. doi: 10.1021/jz301476w

  58. [58]

      X. Wang, Z. Li, Y. Qu, et al., Chem 5 (2019) 1486–1511. doi: 10.1016/j.chempr.2019.03.002

  59. [59]

      X. Zhang, C. Wang, K. Chen, et al., Adv. Mater. 35 (2023) 2211512. doi: 10.1002/adma.202211512

  60. [60]

      S.J. Freakley, Q. He, J.H. Harrhy, et al., Science 351 (2016) 965–968. doi: 10.1126/science.aad5705

  61. [61]

      Q. Chang, P. Zhang, A.H.B. Mostaghimi, et al., Nat. Commun. 11 (2020) 2178. doi: 10.1038/s41467-020-15843-3

  62. [62]

      Z. Yu, S. Lv, Q. Yao, et al., Adv. Mater. 35 (2023) 2208101. doi: 10.1002/adma.202208101

  63. [63]

      M. Song, W. Liu, J. Zhang, et al., Adv. Funct. Mater. 33 (2023) 2212087. doi: 10.1002/adfm.202212087

  64. [64]

      Y. Sun, L. Silvioli, N.R. Sahraie, et al., J. Am. Chem. Soc. 141 (2019) 12372–12381. doi: 10.1021/jacs.9b05576

  65. [65]

      Q. Zhang, L. Zheng, F. Gu, et al., Nano Energy 116 (2023) 108798. doi: 10.1016/j.nanoen.2023.108798

  66. [66]

      Z. Zhuang, A. Huang, X. Tan, et al., Joule 7 (2023) 1003–1015. doi: 10.1016/j.joule.2023.04.005

  67. [67]

      L. Wang, J. Wu, S. Wang, et al. Nano Res. 17 (2024) 3261–3301. doi: 10.1007/s12274-023-6037-8

  68. [68]

      T. Gan, D. Wang, Nano Res. 17 (2024) 18–38. doi: 10.1007/s12274-023-5700-4

  69. [69]

      C. Tang, L. Chen, H. Li, et al., J. Am. Chem. Soc. 143 (2021) 7819–7827. doi: 10.1021/jacs.1c03135

  70. [70]

      E. Jung, H. Shin, B.H. Lee, et al., Nat. Mater. 19 (2020) 436–442. doi: 10.1038/s41563-019-0571-5

  71. [71]

      H. Shen, N. Qiu, L. Yang, et al., Small 18 (2022) 2200730. doi: 10.1002/smll.202200730

  72. [72]

      L. Jing, Q. Tian, P. Su, et al., J. Mater. Chem. A 10 (2022) 4068–4075. doi: 10.1039/d1ta10416k

  73. [73]

      C. Liu, Z. Yu, F. She, et al., Energy Environ. Sci. 16 (2023) 446–459. doi: 10.1039/d2ee02734h

  74. [74]

      Y.X. Du, Q. Yang, W.T. Lu, et al., Adv. Funct. Mater. 33 (2023) 2300895. doi: 10.1002/adfm.202300895

  75. [75]

      J. Du, G. Han, W. Zhang, et al., Nat. Commun. 14 (2023) 4766. doi: 10.1038/s41467-023-40467-8

  76. [76]

      Y. Wang, R. Shi, L. Shang, et al., Angew. Chem. Int. Ed. 59 (2020) 13057–13062. doi: 10.1002/anie.202004841

  77. [77]

      C. Tang, Y. Jiao, B. Shi, et al., Angew. Chem. Int. Ed. 59 (2020) 9171–9176. doi: 10.1002/anie.202003842

  78. [78]

      E. Berl, Trans. Electrochem. Soc. 76 (1939) 359. doi: 10.1149/1.3500291

  79. [79]

      Q. Wang, L. Ren, J. Zhang, et al., Adv. Energy Mater. 13 (2023) 2301543. doi: 10.1002/aenm.202301543

  80. [80]

      G. Wei, Y. Li, X. Liu, et al., Angew. Chem. Int. Ed. 62 (2023) e202313914. doi: 10.1002/anie.202313914

  81. [81]

      J. Huang, C. Fu, J. Chen, et al., CCS Chem. 4 (2022) 566–583. doi: 10.31635/ccschem.021.202000750

  82. [82]

      K. Dong, J. Liang, Y. Wang, et al., Angew. Chem. Int. Ed. 60 (2021) 10583–10587. doi: 10.1002/anie.202101880

  83. [83]

      Z. Bao, J. Zhao, S. Zhang, et al., Nano Res. 16 (2023) 9050–9058. doi: 10.1007/s12274-023-5682-2

  84. [84]

      H. Jiang, C. Zhang, Z. Wang, et al., Appl. Catal. A 661 (2023) 119242. doi: 10.1016/j.apcata.2023.119242

  85. [85]

      Q. Yuan, J. Zhao, D.H. Mok, et al., Nano Lett. 22 (2021) 1257–1264.

  86. [86]

      J. Wu, M. Hou, Z. Chen, et al., Adv. Mater. 34 (2022) 2202995. doi: 10.1002/adma.202202995

  87. [87]

      Y. Pang, K. Wang, H. Xie, et al., ACS Catal. 10 (2020) 7434–7442. doi: 10.1021/acscatal.0c00584

  88. [88]

      G.L. Chai, Z. Hou, T. Ikeda, K. Terakura, J. Phys. Chem. C 121 (2017) 14524–14533. doi: 10.1021/acs.jpcc.7b04959

  89. [89]

      N. Ramaswamy, S. Mukerjee, J. Phys. Chem. C 115 (2011) 18015–18026. doi: 10.1021/jp204680p

  90. [90]

      M. Ko, Y. Kim, J. Woo, et al., Nat. Catal. 5 (2022) 37–44.

  91. [91]

      C. Xia, Y. Xia, P. Zhu, L. Fan, H. Wang, Science 366 (2019) 226–231. doi: 10.1126/science.aay1844

  92. [92]

      H. Li, P. Wen, D.S. Itanze, et al., Nat. Commun. 11 (2020) 3928. doi: 10.1038/s41467-020-17584-9

  93. [93]

      H. Sheng, R.D. Ross, J.R. Schmidt, S. Jin, ACS Energy Lett. 8 (2023) 196–212. doi: 10.1021/acsenergylett.2c01945

  94. [94]

      L. Sun, X. Jin, T. Su, A.C. Fisher, X. Wang, Adv. Mater. 36 (2024) 2306336. doi: 10.1002/adma.202306336

  95. [95]

      Y. Wang, H. Yang, Z. Liu, et al., J. Energy Chem. 87 (2023) 247–255. doi: 10.54254/2754-1169/56/20231172

  96. [96]

      Z. Su, Q. Huang, Q. Guo, et al., Nano Res. Energy 2 (2023) e9120078. doi: 10.26599/nre.2023.9120078

  97. [97]

      Z. Che, X. Lu, B. Cai, et al., Nano Res. 15 (2022) 1269–1275. doi: 10.1007/s12274-021-3645-z

  98. [98]

      L.P. Yuan, T. Tang, J.S. Hu, L.J. Wan, Acc. Mater. Res. 2 (2021) 907–919. doi: 10.1021/accountsmr.1c00135

  99. [99]

      S. Geng, Y. Ji, S. Yang, et al., Adv. Funct. Mater. 33 (2023) 2300636. doi: 10.1002/adfm.202300636

  100. [100]

       Y. Zhang, M. Wang, W. Zhu, et al., Angew. Chem. Int. Ed. 62 (2023) e202218924. doi: 10.1002/anie.202218924

  101. [101]

       E.L. Gyenge, C.W. Oloman, J. Appl. Electrochem. 31 (2001) 233–243. doi: 10.1023/A:1004159102510

  102. [102]

       X. Zhang, X. Zhao, P. Zhu, et al., Nat. Commun. 13 (2022) 2880. doi: 10.1038/s41467-022-30337-0

  103. [103]

       E. Jung, H. Shin, W. Hooch Antink, Y.E. Sung, T. Hyeon, ACS Energy Lett. 5 (2020) 1881–1892. doi: 10.1021/acsenergylett.0c00812

  104. [104]

       Y. Jiang, P. Ni, C. Chen, et al., Adv. Energy Mater. 8 (2018) 1801909. doi: 10.1002/aenm.201801909

  105. [105]

       J. An, N. Li, Y. Wu, et al. Environ. Sci. Technol. 54 (2020) 10916–10925. doi: 10.1021/acs.est.0c03233

  106. [106]

       R.M. Sellers, Analyst 105 (1980) 950–954. doi: 10.1039/an9800500950

  107. [107]

       J. Xie, J. Jing, J. Gu, et al., J. Environ. Chem. Eng. 10 (2022) 107882. doi: 10.1016/j.jece.2022.107882

  108. [108]

       Z. Xing, K. Shi, Z.S. Parsons, X. Feng, ACS Catal. 13 (2023) 2780–2789. doi: 10.1021/acscatal.2c05639

  109. [109]

       Y. Wang, R. Shi, L. Shang, et al., Nano Energy 96 (2022) 107046. doi: 10.1016/j.nanoen.2022.107046

  110. [110]

       E. Zhang, L. Tao, J. An, et al., Angew. Chem. Int. Ed. 61 (2022) e202117347. doi: 10.1002/anie.202117347

  111. [111]

       W. Zhou, X. Meng, L. Rajic, et al., Electrochem. Commun. 96 (2018) 37–41. doi: 10.1016/j.elecom.2018.09.007

  112. [112]

       J. An, Y. Feng, N. Wang, et al., J. Hazard. Mater. 428 (2022) 128185. doi: 10.1016/j.jhazmat.2021.128185

  113. [113]

       H. Zhang, S. Wu, X. Huang, et al., Chem. Eng. J. 428 (2022) 131534. doi: 10.1016/j.cej.2021.131534

  114. [114]

       Z. Chen, S. Chen, S. Siahrostami, et al., React. Chem. Eng. 2 (2017) 239–245. doi: 10.1039/C6RE00195E

  115. [115]

       C. Xia, S. Back, S. Ringe, et al., Nat. Catal. 3 (2020) 125–134. doi: 10.1038/s41929-019-0402-8

  116. [116]

       B. Xia, Q. Huang, H. Wang, et al., ACS Appl. Mater. Interfaces 15 (2023) 32416–32424. doi: 10.1021/acsami.3c04772

  117. [117]

       Z. Lin, Q. Zhang, J. Pan, et al., Energy Environ. Sci. 15 (2022) 1172–1182. doi: 10.1039/d1ee02884g

  118. [118]

       E. Ruiz-López, E. Amores, A. Raquel de la Osa, F. Dorado, A. de Lucas-Consuegra, Chem. Eng. J. 379 (2020) 122289. doi: 10.1016/j.cej.2019.122289

  119. [119]

       S.M.H. Hashemi, P. Karnakov, P. Hadikhani, et al., Energy Environ. Sci. 12 (2019) 1592–1604. doi: 10.1039/C9EE00219G

  120. [120]

       P. Farinazzo Bergamo Dias Martins, I. Plazl, D. Strmcnik, B. Genorio, Curr. Opin. Electrochem. 38 (2023) 101223. doi: 10.1016/j.coelec.2023.101223

  121. [121]

       H. Sheng, A.N. Janes, R.D. Ross, et al., Energy Environ. Sci. 13 (2020) 4189–4203. doi: 10.1039/d0ee01925a

  122. [122]

       H. Yin, Y. Dou, S. Chen, et al., Adv. Mater. 32 (2020) 1904870. doi: 10.1002/adma.201904870

  123. [123]

       R.J. Lewis, K. Ueura, X. Liu, et al., ACS Catal. 13 (2023) 1934–1945. doi: 10.1021/acscatal.2c05799

  124. [124]

       S. Wu, H. Zhang, X. Huang, Z. Wei, Chem. Commun. 58 (2022) 8942–8945. doi: 10.1039/d2cc03181g

• 

  Figure 1  Schematic diagram of catalytic mechanism and design principles for 2e⁻ ORR under different pH conditions and device engineering for H₂O₂ electrosynthesis.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) Pathway for 4e⁻ and 2e⁻ O₂ electroreduction. Reproduced with permission [55]. Copyright 2019, American Chemical Society. (b) Volcano plot for oxygen reduction to H₂O₂. The red solid line and blue solid line represent the theoretically calculated oxygen reduction volcano plot for the 4e⁻ and 2e⁻ ORR, with the limiting potential plotted as a function of ΔG_HO* (lower horizontal axis) and ΔG_HOO* (upper horizontal axis). In addition, η means the thermodynamic overpotential for a specific reaction. Reproduced with permission [24]. Copyright 2013, Springer Nature. (c) Possible configurations of O₂ adsorption on the catalyst surface. Reproduced with permission [7]. Copyright 2023, Springer Nature. (d) Screened metal atoms with their most stable bulk structures and a schematic illustration of all the considered SACs. Reproduced with permission [55]. Copyright 2019, American Chemical Society.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) ORR of GC in 0.1 mol/L HClO₄ and 0.1 mol/L KOH. Reproduced with permission [44]. Copyright 2018, American Chemical Society. (b) Pourbaix diagram showing the most energetically favorable surface species configurations as a function of reaction conditions of electrode potential and pH. The k-O, c-O and COOH represent the ketone oxygen, CO group, and COOH group, respectively. The digits before k-O, c-O and COOH represent the amount of corresponding oxygen configurations doped into the carbon substrate in relevant structural models. A larger digit represents more oxygen configurations, and less hydrogen content. Reproduced with permission [87]. Copyright 2020, American Chemical Society. (c) Schematic illustration of the double-layer structure during ORR in acidic (left) and alkaline (right) conditions. Insets (ⅰ) and (ⅱ) represent the inner sphere pathway in acidic and alkaline conditions, respectively, while inset (ⅲ) represents the outer sphere pathway in an alkaline condition. Reproduced with permission [89]. Copyright 2011, American Chemical Society.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) Time-dependent neutral H₂O₂ concentration measured at a constant potential of 0.4 V (vs. RHE) for 120 h. The accumulated H₂O₂ concentration was continuously cycled through the system (600 mL). Reproduced with permission [92]. Copyright 2020, Springer Nature. (b) Illustration of the 2e⁻ ORR to H₂O₂ over L-ZnO. (c) DOS of S- and L-ZnO absorbed *OOH. Reproduced with permission [30]. Copyright 2023, Royal Society of Chemistry. (d) Schematic structures of metastable P-hcp Ni. The different blue balls are Ni in different layers and the pink ball is P. (e) H₂O₂ yields of different catalysts in 1.0 mol/L Na₂SO₄ by using the flow cell device. Reproduced with permission [99]. Copyright 2023, Wiley–VCH. (f) Schematic illustration of 2e⁻ ORR toward H₂O₂ in acid electrolyte with or without Na⁺ under industrial-relevant current. Reproduced with permission [102]. Copyright 2022, Springer Nature.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Schematics for (a) a RRDE, (b) an H-cell configuration with a catalyst deposited on a solid substrate in a nonflowing electrolyte and (c) a continuous flow cell with a catalyst deposited on a GDE. Reproduced with permission [103]. Copyright 2020, American Chemical Society. (d) Schematic diagram of a GDE used for H₂O₂ electrosynthesis. Reproduced with permission [86]. Copyright 2022, Wiley–VCH. (e) Dual-membrane cell with pure H₂ and O₂ streams separately introduced to the anode and cathode, respectively. Reproduced with permission [91]. Copyright 2019, Springer Nature. (f) Membrane-free cell consisting of a flow channel. Reproduced with permission [114]. Copyright 2017, Royal Society of Chemistry.

  下载: 全尺寸图片 幻灯片
•