English

• 1. Introduction

  Since Thenard first discovered H₂O₂ by reacting BaO₂ with HNO₃ in 1818, it has been extensively utilized as a highly efficient and environmentally friendly oxidant in various applications such as wastewater treatment, medicine disinfection, and chemical industries [1-3]. The global demand for H₂O₂ is rapidly increasing. Reports indicate that the global H₂O₂ market is estimated to grow at a compound annual growth rate of 5.7%, reaching 4.05 billion USD by 2027 from 3.24 billion USD in 2023 [4]. A straightforward route for synthesizing H₂O₂ involves the reaction of H₂ and O₂ in the presence of noble metal catalysts, such as Au and Pd (H₂ + O₂ → H₂O₂, the standard Gibbs free energy ΔG⁰ = −120 kJ/mol) [5-7]. Due to safety concerns and the need for optimal selectivity, large quantities of N₂ must be added during the actual production, significantly reducing the yield of H₂O₂ [8]. Currently, H₂O₂ is primarily produced on an industrial scale through anthraquinone oxidation, which accounts for approximately 95% of the world's total output [1,9]. However, this process requires huge energy input and generates chemical waste. Therefore, the development of an inexpensive and environment-friendly route for H₂O₂ synthesis is desired.

  Photocatalytic H₂O₂ synthesis offers a potentially more sustainable and cleaner alternative to the anthraquinone oxidation method [10-16]. Due to its utilization of solar energy and green reactants (H₂O and O₂), the photocatalytic process generates no pollution. Therefore, there has been a significant increase in studies focusing on photocatalytic H₂O₂ production [17-20]. Among the various investigated photocatalysts, 2D photocatalysts have received tremendous research interest due to their planar structure and ultrathin thickness, which endow several unique physicochemical features [21,22]. For instance, 2D photocatalysts with greatly exposed surfaces offer abundant active sites; the nanoscale thickness dramatically shortens the migration distance of photoinduced charge carriers from the bulk phase to the photocatalysts on the surface; the 2D flexible planar structure facilitates the design and engineering of photocatalysts at molecular-level such as heterojunction construction, defects engineering, surface modification, thereby tuning their electronic structures, active sites number and charge carriers behaviour for photocatalysis [23].

  Figure 1

  

  Figure 1.  Schematic illustration of (a) photocatalytic processes for H2O2 production and (b) corresponding energy diagrams. (c) Reaction pathways and corresponding energy diagrams for ORR and WOR pathways.

  DownLoad: Full-Size Img PowerPoint

  Herein, we review the recent progress in photocatalytic synthesis of H₂O₂ from H₂O and O₂ using 2D photocatalysts, focusing on molecular-level regulation strategies. Beginning with the fundamentals of photocatalytic H₂O₂ synthesis, this review provides a detailed summary of the pathways for producing H₂O₂ and briefly introduces common methods for H₂O₂ detection. Subsequently, in alignment with modification strategies, the latest advancements in 2D photocatalysts that significantly facilitate charge separation/transfer to enhance H₂O₂ generation are exhaustively discussed. Key issues and prospective outlooks for achieving efficient and sustainable generation of H₂O₂ in scale-up industrial production are also proposed. This review aims to provide comprehensive insight and valuable guidance for the design of high-efficiency, stable and promising photocatalysts for H₂O₂ production.

  2. Fundamentals of photocatalytic H₂O₂ synthesis

  2.1 Reaction mechanisms for photocatalytic H₂O₂ synthesis

  The photocatalytic synthesis of H₂O₂ is based on the fundamental principles of photocatalysis ^[14]. In general, the light-driven photocatalytic reaction using a photocatalyst primarily involves three steps: photoexcitation, charge carriers separation/migration and surface redox reactions, as illustrated in Fig. 1a. Firstly, photons are absorbed by the photocatalyst. When the energy of photons exceeds the band gap (E_g) of the photocatalyst, electrons (e^–) and holes (h⁺) with certain redox capacity are generated on the conduction band (CB) and valence band (VB), respectively. Subsequently, the two types of charge carriers separate and transfer to the surface of the photocatalyst. During this process, most of these photoinduced electrons and holes will recombine with each other without involving in any chemical reactions, and the remaining charge carriers trigger subsequent redox reactions on the photocatalyst surface. In this context, the light absorption capability and charge carrier separation/transfer of the photocatalyst are critically important for the generation of H₂O₂.

  H₂O₂ can be generated through the oxygen reduction reaction (ORR) or water oxidation reaction (WOR) (Figs. 1a and b) ^[18,24]. For example, the ORR (Fig. 1c) pathways can be direct one-step 2e^– ORR pathway, indirect two-step e^– ORR pathway and 4e^– ORR ^[25,26]. Thermodynamically, the direct 2e^– ORR (+0.68 V) is more favourable than 4e^– ORR (+1.23 V). In the indirect two-step e^– ORR pathway, photoinduced electrons are first captured by electrophilic O₂ to form O₂^•– radicals, which subsequently generate H₂O₂ through proton-coupled electron transfer (PCET) reactions. Since each step requires only one electron, it is kinetically desirable. However, compared to the direct one-step 2e^– ORR pathway, the single-electron reduction of O₂ to O₂^•– radical needs a more negative potential (−0.33 V), necessitating the photocatalyst to have a more negative CB position. Alternatively, H₂O₂ can be produced via oxidation reactions by photogenerated holes, such as the direct one-step 2e^– WOR pathway and indirect two-step e^– WOR pathway ^[27,28]. During the direct 2e^– WOR pathway, H₂O can be oxidized into H₂O₂ by two photoexcited holes. In the indirect two-step e^– WOR pathway, H₂O is oxidized into ^•OH radical via one photoinduced hole. Subsequently, two ^•OH radicals combine with each other to form one H₂O₂ molecule. Due to the indirect two-step e^– WOR pathway only requiring one hole, it is kinetically favourable compared to the direct one-step 2e^– WOR pathway. However, from a thermodynamic perspective, the direct one-step 2e^– WOR pathway (+1.76 V) is more favourable than the indirect two-step e^– WOR pathway (+2.73 V). In addition, owing to the more negative reaction potential of the 4e^– WOR pathway (+1.23 V) to form O₂ (Fig. 1c), it will compete with the direct 2e^– WOR pathway, thereby impacting the selectivity of H₂O₂ production. Hence, finding ways to inhibit the competition of the 4e^– water-to-O₂ conversion, while also enhancing the efficiency of both the direct one-step 2e^– WOR pathway and the indirect two-step e^– WOR pathway, is undeniably crucial for addressing these challenges.

  Given the presence of multiple ORR and WOR pathways, an indepth understanding of the reaction mechanisms and precise controlling of these pathways are paramount for enhancing the photocatalytic H₂O₂ production rate and solar-to-chemical conversion (SCC) efficiency. Experimental studies have confirmed the existence of five potential routes for photocatalytic H₂O₂ synthesis: direct one-step 2e^– ORR pathway, indirect two-step e^– ORR pathway, direct one-step 2e^– WOR pathway, indirect two-step e^– WOR pathway, simultaneous ORR and WOR two-channel pathways ^[25,27–29].

  Especially, the simultaneous ORR and WOR two-channel pathways can achieve the photocatalytic generation of H₂O₂ from both O₂ reduction and H₂O oxidation. The photoinduced electrons directly reduce O₂ to H₂O₂ via 2e^– ORR. Simultaneously, the holes directly oxidize H₂O to produce H₂O₂ via 2e^– WOR, providing protons for the 2e^– ORR. Since there are no additional sacrificial reagents besides H₂O and O₂, the theoretical atom utilization efficiency can reach 100% ^[27]. This process is also defined as a photosynthesis process (2H₂O + O₂ → 2H₂O₂, △G₀ = 204 kJ/mol), in which the solar energy is converted and stored in H₂O₂ ^[25]. However, since the 4e^– WOR pathway is thermodynamically more favorable than the direct one-step 2e^– WOR, achieving selectivity for H₂O₂ production becomes a significant issue. Compared with the direct one-step 2e^– ORR and direct one-step 2e^– WOR dual pathways, the atomic utilization efficiency of this direct one-step 2e^– ORR and 4e^– WOR pathways is only 68% ^[30]. From another perspective, if this is realized, oxidation of H₂O by the photoinduced h⁺ provides O₂ and H⁺ for oxygen reduction, which in turn would facilitate the direct one-step 2e^– reduction of O₂ for the efficient generation of H₂O₂.

  2.2 Protocols for photocatalytic H₂O₂ synthesis measurements

  The production rate of H₂O₂ via photocatalytic H₂O₂ synthesis is typically low, usually within the micromolar range. Therefore, accurate detection and quantification of H₂O₂ is crucial for the development of photocatalytic H₂O₂ synthesis. As early as 1927, Baur et al. used an iodide method (2KI+ H₂O₂ →2KOH + I₂, I₂ caused the starch solution to adopt a blue colour) to analyze H₂O₂ produced over ZnO via a photocatalytic reaction ^[31]. Up to now, several approaches have been applied to determine the H₂O₂ concentration in photocatalysis, including colourimetric test strip ^[32], direct optical absorption ^[33], colouration ^[34–37], chemiluminescence ^[33,38,39], enzymatic assay ^[11,40], fluorescence probe ^[41,42], and chemical titration methods ^[3,43–48]. The advantages and disadvantages of different methods for H₂O₂ detection are summarized in Table 1.

  Table 1

  

  Table 1.  Pros and cons of various H₂O₂ detection methods.

  DownLoad: CSV

  The colourimetric test strip method is a simple and convenient way to measure the concentration of H₂O₂ ^[32,49]. The principle of redox colour change of strip is similar to that of the chemical titration method, but they utilize solid phase assays ^[32]. When using a colourimetric test strip, it is dipped into the solution containing H₂O₂ for approximately one second and then exposed to air for a period of time. During this time, the colour of the strip changes from blue to yellow/brown. The concentration of H₂O₂ can be determined according to the colour of the strip. There are many commercially available H₂O₂ colourimetric test strips on the market, such as Quantofix has three types of strips with maximum detection limits of 25, 100 and 1000 ppm, respectively. It is worth noting that the concentration of H₂O₂ is determined by observing the colour of the strip with the naked eye, and the systematic error is large. Additionally, prolonged exposure of the strip to air also affects the accuracy of measurement results. Therefore, the colourimetric test strip method is better as a semi-quantitative method.

  For the direct optical absorption method ^[33], the molar absorption coefficient of H₂O₂ is 0.01 M^–1 cm^–1 at the wavelength of 360 nm. Even at a shorter wavelength of 260 nm, the molar absorption coefficient is only 13 M^–1 cm^–1. Although this method is simple to operate, due to the small absorption coefficient, it is difficult to directly detect the small amount of H₂O₂ in the reaction system using a UV–vis spectrophotometer.

  In photocatalysis, the colouration method is the most widely used for quantifying H₂O₂. In this method, iodide (I^–), Ti⁴⁺,N,N–diethyl-p-phenylenediamine (DPD), metavanadate (VO₃^–) or Ti^IV-tetrapyridylporphyrin (TiTPyP) are often used as color-development reagents to react with H₂O₂, forming a compound that strongly absorbs in the ultraviolet or visible range. These compounds are then measured using a UV–vis spectrophotometer to indirectly determine H₂O₂ concentration. For example, 1 mL of sample is added to the mixture of KI (1 mL, 0.4 mol/L) and potassium hydrogen phthalate (C₈H₅KO₄, 1 mL, 0.1 mol/L) solutions, which is then kept for 30 min. Under acidic conditions, I^– can be oxidized by H₂O₂ to form triiodide anions (I₃^–), and the resulting amount of I₃^– (3I^– + 2H⁺ + H₂O₂ → I₃^– + 2H₂O) can be measured using UV–vis spectrophotometer at 350 nm, which further determines H₂O₂ concentration ^[30,50–52]. When using Ti⁴⁺ as indicator, H₂O₂ reacts with Ti⁴⁺ to form the yellow peroxo-titanic acid (Ti⁴⁺ + 2H₂O + H₂O₂ → H₂TiO₄ + 4H⁺). Then the solution containing H₂O₂ can be monitored indirectly by detecting the yellow H₂TiO₄ at the wavelength of 410 nm using a UV–vis spectrophotometer ^[36]. Furthermore, Choi and co-workers detected the concentration of H₂O₂ using N,N–diethyl-1,4-phenylene-diamine (DPD) and peroxidase (POD), which is based on the POD-catalyzed oxidation of DPD by H₂O₂ (). The absorption wavelength maxima of the DPD^•+ is at 551 nm ^[37,53]. Nogueira et al. measured H₂O₂ in acidic conditions using VO₃^–, which reacts with H₂O₂ (VO₃^– + 4H⁺ + H₂O₂ → VO₂³⁺ + 3H₂O). The resulting red-orange-coloured peroxovanadium cation (VO₂³⁺) can be detected by a UV–vis spectrophotometer with a maximum absorbance at 450 nm ^[54]. When TiTPyP solution (3× 10^–5 mol/L in 1 mol/L H₂SO₄) is used, TiTPyP is oxidized by H₂O₂. The resulting product can be detected by a UV–vis spectrophotometer at an absorption wavelength of 450 nm, thereby indirectly determining the amount of H₂O₂ ^[55].

  In the chemiluminescence method, a luminol chemiluminescence probe is employed to monitor the formation of H₂O₂. During the experiment, oxidized luminol (5-amino-2,3-diaza-1,4-naphthoquinone) reacts with H₂O₂ to generate an excited state of ion (3-amino phthalate^*), whose chemiluminescence intensity can be measured by a photomultiplier tube. For example, Hirakawa et al. used the chemiluminescence method to determine the amount of H₂O₂ when investigating the reaction mechanism of selective generation of O₂^•– and H₂O₂ by N- and S-doped TiO₂ photocatalysts in aqueous solution ^[39]. It should be noted that the limit of detection for H₂O₂ by the chemiluminescence method is about 1 nmol/L ^[33,39].

  Enzymatic assay for H₂O₂ detection is highly sensitive, making them suitable for measuring trace amounts of H₂O₂ in various applications ^[56,57]. Horseradish peroxidase is a commonly used enzyme reagent. For example, resorufin, produced from the reaction of H₂O₂ with Ampliflu Red in the presence of horseradish peroxidase, can be detected using high-performance liquid chromatography with a photo-diode array detector (λ_abs. = 560 nm) ^[11,40]. Calibration of the resorufin peak area with a standard H₂O₂ solution allows for quantitative analysis of H₂O₂ concentration. At present, this is one of the most accurate ways to determine H₂O₂ ^[11]. It should be noted that when horseradish peroxidase catalyzes the reaction of H₂O₂, there are certain restrictions on the pH and temperature of the reaction solution. During analysis, avoid storing the enzyme for extended periods or exposing it to unfavourable conditions to prevent enzyme deactivation, which affects the accuracy of H₂O₂ detection.

  The fluorescence probe method for the quantification of H₂O₂ is frequently reported in biological systems ^[58]. The principle of this method is to convert H₂O₂ into fluorescence products, and the concentration of fluorescence products is determined by fluorescence spectroscopy. For instance, in the catalysis of horseradish peroxidase (HRP), H₂O₂ reacts with p-hydroxyphenylacetic acid (POHPAA) to form a fluorescent dimer, which can be measured with a fluorescence spectrophotometer at an emission wavelength of 408.5 nm and an excitation wavelength at 316.5 nm ^[42].

  Chemical titration is the most commonly used method for quantifying H₂O₂. In this approach, titration standard solutions such as yellow-coloured Ce(SO₄)₂ and purple-coloured KMnO₄ are typically employed ^[3,43–47]. For yellow-coloured Ce(SO₄)₂, H₂O₂ reacts with Ce⁴⁺ to form the colourless Ce³⁺ (2Ce⁴⁺ + H₂O₂ → 2Ce³⁺ + O₂ + 2H⁺), which can be analyzed by the UV–vis spectrophotometer. Specifically ^[44,46–48], the yellowcolored Ce(SO₄)₂ transparent solution (1 mmol/L) is prepared by dissolving Ce(SO₄)₂ (33.2 mg) in H₂SO₄ solution (100 mL, 0.5 mol/L). A series of known concentrations of H₂O₂ are then added to this solution to produce a standard curve, which is measured by UV–vis spectrophotometer at 320 nm. According to the linear relationship between signal intensity and Ce⁴⁺ concentration, the H₂O₂ concentration of the sample can be obtained. When using purple-coloured KMnO₄ as the titrant, under acidic conditions, the purple colour of MnO₄^– will be reduced to colourless Mn²⁺ by H₂O₂ (2MnO₄^– + 6H⁺ + 5H₂O₂ → 2Mn²⁺ + 5O₂ + 8H₂O) ^[45]. Before H₂O₂ is completely consumed, the solution to be tested will quickly turn colourless upon the addition of MnO₄^–. After a complete reaction of H₂O₂, further addition of MnO₄^– leads to the titration endpoint, indicated by a persistent (about 30 s) light pink colour. The concentration of H₂O₂ in the solution can be determined based on the amount of MnO₄^– consumed. It should be noted that since the KMnO₄ assay relies on the detection of the purple colour produced by a slight excess of KMnO₄ to determine the titration endpoint, choosing Ce(SO₄)₂ as the titration standard solutions will yield more accurate results than KMnO₄.

  3. 2D photocatalysts for H₂O₂ production

  To elucidate the relationship between structure and photocatalytic activity for H₂O₂ production using 2D photocatalysts, the primarily studied materials have been classified into two types in this review based on their composition and chemical properties: Inorganic materials and organic polymers. To date, inorganic materials such as metal oxides, metal chalcogenides and bismuth-based materials as well as organic polymers including g-C₃N₄, metal organic frameworks (MOFs) and covalent organic frameworks (COFs) have been explored as photocatalysts to produce H₂O₂.

  3.1 Metal oxides

  Constructing metal oxides, such as TiO₂ [59] and ZnO, with 2D structures can effectively increase their specific surface area as well as maximize exposure of active surfaces/sites, thereby improving their photocatalytic activity [21]. The pioneering study dates back to 1927 when Baur et al. first reported the production of H₂O₂ using ZnO photocatalyst [31]. Approximately 50 years later, Harbour et al. observed that photocatalytic generation of H₂O₂ via indirect 2e^– ORR pathway when using particulate TiO₂ as the photocatalysts, yet the presence of H₂O₂ was transient due to its rapid decomposition [60]. Since then, various strategies have been employed to improve the photocatalytic production performance of H₂O₂ in terms of efficiency and selectivity. For instance, surface fluorination of TiO₂ enhanced the production of H₂O₂ with steady-state concentration to millimolar levels (1−1.3 mmol/L) by inhibiting the degradation [61]. Additionally, the deposition of Au nanoparticles [62] or silver-gold clusters [63] on TiO₂ has been effective in enhancing the rate of H₂O₂ production. This improvement is attributed to the separation of photogenerated carriers and the transfer of electrons from TiO₂ to the loaded metal nanoparticles/clusters. The representative metal oxide photocatalysts [64-86], along with various strategies that enhance H₂O₂ production at the molecular level are summarized in Table S1 (Supporting information).

  The creation of a heterojunction structure is one effective strategy in the pursuit of efficient photocatalytic processes. Such a structure is constructed by combining two kinds of photocatalysts with different band positions and Fermi levels [87]. Typically, heterojunction structures involve conventional type-Ⅰ, type-Ⅱ and type-Ⅲ heterojunctions, p-n heterojunctions, Schottky heterojunctions and Z-scheme heterojunctions [88,89]. The non-uniform distribution of charge at the interface of the heterojunction construction results in the generation of an internal electric field, driving the directional migration of charge carriers to achieve effective charge separation, utilization, and extension of lifetime [90]. For instance, Zhang's group designed a 2D g-C₃N₄/TiO₂ nanocomposites (HT-CN/TiO₂) as a vertical Z-scheme heterojunction [64]. Compared with the physically mixed sample, HT-CN/TiO₂ generated H₂O₂ more rapidly. The authors suggested that the enhanced performance was attributed to the Z-scheme transfer mechanism and the integrated effects of this specific face-to-face interfacial morphology. Driven by the internal electric field, photoinduced e^– from CB of TiO₂ combine with photoinduced h⁺ of g-C₃N₄ at the inter-plane contact interface, leading to an accumulation of e^– in the CB of g-C₃N₄ and h⁺ in the VB of TiO₂. As a result, the accumulated holes in the VB of TiO₂ oxidized H₂O to generate H₂O₂ via an indirect two-step e^– WOR pathway, while electrons in the CB of g-C₃N₄ reduced O₂ to generate H₂O₂ by indirect two-step e^– ORR pathway (Fig. 2a). They also reported a WO₃-TiO₂ vertical heterojunction via in situ assembling of WO₃ nanosheets on TiO₂ nanosheets and found the formation of an internal electric field between WO₃ and TiO₂ [65]. This internal electric field formed at the interface resulted in a Z-scheme photocatalytic mechanism for H₂O₂ formation without the addition of sacrificial reagents. The Z-scheme charge transfer route enabled the spatial separation of the remaining electrons in the CB of TiO₂ and holes in the VB of WO₃ with high redox potentials. Therefore, the electrons in CB of TiO₂ triggered a powerful production of H₂O₂ by indirect two-step e^– ORR pathway and holes in the VB of WO₃ oxidized H₂O to yield H₂O₂ via indirect two-step e^– WOR pathway.

  Figure 2

  

  Figure 2.  (a) Photocatalytic H₂O₂ production over HT-CN/TiO₂ featuring the Z-scheme charge transfer mechanism. Reproduced with permission [64]. Copyright 2019, Elsevier. (b) The photocatalytic production of H₂O₂ pathway over In₂S₃@In₂O₃ and In₂S₃@O_v/In₂O₃ composites. Reproduced with permission [86]. Copyright 2021, American Chemical Society. (c) Schematic diagram of the reaction mechanism over Au/WO₃. Reproduced with permission [69]. Copyright 2021, Elsevier. (d) Mechanism of Gd³⁺ doped WO₃ photocatalysts for photocatalytic H₂O₂ production. Reproduced with permission [76]. Copyright 2023, Elsevier.

  DownLoad: Full-Size Img PowerPoint

  In addition to constructing heterojunctions, introducing defects has been proven to be an effective method to modulate the electronic band structure of a photocatalyst, which has a significant impact on its photocatalytic performance [91]. For example, as reported by Zhu et al. [67], C-supported oxygen vacancy (O_v) rich Co₃O₄ nanoplate (C—O_vCo-400) performed selective and efficient H₂O₂ production (3.78 mmol g^–1 h^–1) under visible light irradiation (λ ≥ 420 nm) in the absence of sacrificial agent. It has been confirmed that the presence of O_v reduced the band gap and enhanced the donor density. Additionally, O_v can act as the active sites for WOR, while the C support serves as reaction centres for the ORR, thereby markedly facilitating the separation and migration of charge carriers. Consequently, H₂O adsorbed on the surface of O_vCo was oxidized to form H₂O₂ via direct one-step 2e^– WOR, while O₂ was reduced into H₂O₂ via direct one-step 2e^– ORR. As another example, Fu et al. found that the photocatalytic production rate of H₂O₂ from TiO₂ with bulk defects and surface defects was 30 and 45 times higher than that of pristine TiO₂ and P25, respectively [68]. A synergistic effect caused by the bulk and surface defects in TiO₂ was observed: (1) The bulk defects served as h⁺ acceptors to drive directional h⁺ transfer and efficiently boosted the separation and transfer of charge carriers; (2) The surface defects strengthened the adsorption of O₂, thereby dramatically improving the generation of H₂O₂ via indirect two-step e^– ORR. Chen et al. also reported that O_v sites in the In₂S₃@In₂O₃ can change the H₂O₂ production pathway from the indirect two-step e^– ORR to the direct one-step 2e^– ORR, resulting in a significantly improved photocatalytic performance (Fig. 2b) [86].

  Surface modification for the photocatalyst is another efficient approach to improve the photocatalytic activity for H₂O₂ production. Apart from regulating the electronic structure of photocatalysts, the heteroatom on the photocatalyst surface can promote interfacial electron transfer [92]. For instance, Xu's group modified WO₃ with Au nanoparticles (NPs) and achieved superior performance for photocatalytic H₂O₂ production, exhibiting rates 63.6 and 1511.1 times higher than pristine WO₃ and Au/SiO₂, respectively (Fig. 2c) [69]. When WO₃ was excited by visible light (λ ≥ 420 nm), the photogenerated e^– transferred from the CB of WO₃ to Au NPs, followed by direct one-step 2e^– ORR on Au NPs. However, when WO₃ was irradiated by light with a wavelength longer than 535 nm, the hot e^– transferred from Au NPs to the CB of WO₃, thereby generating H₂O₂ via direct one-step 2e^– ORR pathway on WO₃. Similarly, Zhou's team reported photocatalytic generation of H₂O₂ via direct one-step 2e^– WOR pathway over 1.0 wt% Pt/WO₃ and suggested that Pt loading significantly enhanced the separation and transfer of charge carriers [70]. Additionally, the ratios of Pt⁰ and PtO_x affected the photocatalytic activities. Specifically, Pt⁰ species acted as active sites to reduce O₂ to H₂O, while PtO_x species served as oxidation centres to oxidize H₂O to H₂O₂.

  Doping is also adopted to modulate the electron transfer process. For instance, Wang et al. doped Gd³⁺ into WO₃ for promoted photocatalytic degradation of tetracycline and simultaneous in situ production of H₂O₂ [76]. They found that doping Gd³⁺ into WO₃ not only expanded the light absorption range of the photocatalyst to the near-infrared range but also greatly facilitated the charge separation/transfer from photocatalyst to O₂, thereby stimulating the generation of H₂O₂ via direct one-step 2e^– ORR pathway (Fig. 2d).

  3.2 Metal chalcogenides

  As a significant class of 2D materials, layered metal chalcogenides have garnered considerable attention owing to their diverse structures and rich electronic characteristics [93]. Importantly, metal chalcogenides exhibit broad and tunable light absorption bands, which can be modified by changing the morphology and replacing the metal elements. Therefore, chalcogenides exhibit fascinating photocatalytic properties in various photocatalytic reactions such as oxygen or hydrogen evolution from water. In recent years, researchers have discovered the potential of metal chalcogenides in sunlight-driven H₂O₂ production and have made significant efforts to explore efficient methods for generating H₂O₂ (Table S2 in Supporting information) [94-108].

  Among metal chalcogenides, molybdenum disulfide (MoS₂) stands out as one of the most popular photocatalytic materials [109]. It is composed of S-Mo-S modules held together by weak van der Waals interactions, enabling the production of MoS₂ ranging from bulk to ultrathin crystals, and even down to atomistic thin layers [110]. It has been reported that the sunlight absorption ability can be effectively improved by decreasing the MoS₂ layers [111]. Besides, its electrical conductivity can be enhanced by adding other conducting materials, such as graphene and noble metals [94,112]. For instance, Song et al. modified MoS₂ nanosheets with atomic Au⁰, which not only increased the separation and transfer of photoexcited carriers but also endowed a more negative flat band potential availably to drive the generation of H₂O₂ from O₂ and H₂O through a multichannel pathway [94]. Consequently, the photocatalytic H₂O₂ production rate over MoS₂ loaded with 0.50 wt% Au was 2.5 times than that of bare MoS₂. Interestingly, doping Mn²⁺ in MoS₂, along with the Au modification, yielded a H₂O₂ generation twice higher than that of Au@MoS₂ (Fig. 3a).

  Figure 3

  

  Figure 3.  (a) Mn²⁺ doping and Au⁰ modification on MoS₂ nanosheets, and in-situ-Fenton process using photogenerated H₂O₂. Reproduced with permission [94]. Copyright 2019, Elsevier. (b) Mechanism of S_v−MoS₂/BN heterojunction toward sacrificial agent-free photocatalytic H₂O₂ production for in situ water disinfection. Reproduced with permission [96]. Copyright 2024, Elsevier. (c) Schematic diagram of H₂O₂ production over CBO@MoS₂ Z-scheme heterojunction Reproduced with permission [97]. Copyright 2023, Royal Society of Chemistry. (d). Mechanism of CDs modified SnS₂/In₂S₃ type Ⅱ heterostructure for photocatalytic H₂O₂ production by indirect two-step e^– ORR and 4e^– WOR. Reproduced with permission [99]. Copyright 2021, Royal Society of Chemistry.

  DownLoad: Full-Size Img PowerPoint

  The S-vacancies in metal chalcogenides have been reported to affect the work function and regulate the interfacial barrier of the heterojunction, thereby effectively boosting the separation/transfer of charge carriers [113]. As an example, Jiang et al. prepared a defective MoS_2-v/carbon nitride nanotube (TCN) heterojunction and observed that the combination of MoS_2-v and TCN enhanced O₂ adsorption and facilitated electron exchange between O₂ and MoS_2-v/TCN, which was beneficial to the ORR pathway [95]. Additionally, the creation of heterojunction between MoS_2-v and TCN not only facilitated the rapid separation of charge carriers but also reduced the energy barrier for the production of *OOH and *HO₂^– toward H₂O₂, thereby effectively improving the photocatalytic H₂O₂ generation. Yang et al. adopted defect and interfacial engineering coupled dual strategies to design and synthesize S_v−MoS₂/BN heterojunction for sacrificial agent-free photocatalytic H₂O₂ production, aiming at in situ water disinfection [96]. The experimental and density functional theory (DFT) calculations results demonstrated that the presentence of S-vacancies effectively regulated the interfacial barrier by decreasing the work function of MoS₂, which changed the photocatalytic H₂O₂ production from indirect two-step e^– ORR pathway to direct one-step 2e^– ORR pathway (Fig. 3b). Therefore, S_v−MoS₂/BN exhibited excellent photocatalytic activity for H₂O₂ production.

  In another example, Meduri's team fabricated CBO@MoS₂ Z-scheme heterojunction through a facile hydrothermal method [97]. As a result, the photocatalyst exhibited a significantly high H₂O₂ production rate of 1457 mmol L^–1 h^–1 under one sun illumination via simultaneous indirect two-step e^– ORR and 4e^– WOR two-channel pathways (Fig. 3c). The excellent performance was attributed to the efficient separation and transfer of charge carriers, facilitated by the staggered band alignment and the construction of a Z-scheme heterojunction, resulting in a stronger redox capacity.

  Apart from MoS₂, other metal chalcogenides have also been developed in the field of photocatalytic H₂O₂ synthesis, such as CdS, In₂S₃, SnS₂, MoSe₂, ZrS₃ (Table S2). For example, Fang et al. prepared a Bi/BiOBr-CdS heterojunction through the solvothermal method [98]. It has been reported that the synergistic effect of the heterojunction and the introduction of Bi⁰ decreased the band gap, facilitating the separation and transfer of charge carriers while maintaining a high redox potential for simultaneous ORR and WOR. Consequently, such a Bi/BiOBr-CdS heterojunction showed a high H₂O₂ production rate of 346.4 µmol L^–1 h^–1 (λ > 420 nm) via direct one-step 2e^– ORR and indirect two-step e^– WOR dual channel pathways. Recently, Li et al. reported SnS₂/In₂S₃ type Ⅱ heterostructure modified by carbon dots (CDs) for high-efficiency H₂O₂ production [99]. CDs not only stored the e^– induced by SnS₂ and In₂S₃, thereby regulating the interface photo-charge kinetics of SnS₂/In₂S₃/CDs composite but also acted as the catalytic centres for ORR. As a result, the photocatalytic activity of H₂O₂ production over SnS₂/In₂S₃/CDs composite increased to 1111.89 µmol h^–1 g^–1 (λ ≥ 420 nm) without sacrificial agents by simultaneous indirect two-step e^– ORR and 4e^– WOR two-channel pathways (Fig. 3d).

  3.3 Bismuth-based materials

  Bismuth-based materials have received tremendous research interest due to their unique layered structure and tunable electronic properties. These characteristics make them promising candidates for constructing ultrathin 2D structures, which have been demonstrated to have promising photocatalytic abilities [114-117]. Up to now, many bismuth-based materials, including bismuth vanadate (BiVO₄), bismuth tungstate (Bi₂WO₆), bismuth oxyhalides and bismuth oxycarbonate (Bi₂O₂CO₃), have been reported for the photocatalytic generation of H₂O₂ [11,118-136], and the details are summarized in Table S3 (Supporting information).

  As a visible light-responsive photocatalyst, BiVO₄ (band gap is about 2.4 eV) has been widely used in photocatalytic water oxidation for O₂ evolution [137]. However, the absence of two-electron oxygen reduction active sites in pristine BiVO₄ renders it unfavourable for the generation of H₂O₂ from O₂. To address this challenge, various strategies have been adopted, including surface modification, doping, defect engineering and heterojunction construction. For example, Hirakawa et al. deposited metallic Au nanoparticles on BiVO₄ (Au/BiVO₄) as the cocatalyst, resulting in enhanced H₂O₂ production from O₂ and H₂O via direct one-step 2e^– ORR and 4e^– WOR dual-channel pathways [118]. The formation of a Schottky barrier between BiVO₄ and Au facilitated the separation and transfer of photoinduced electrons to Au, thereby hindering their recombination with holes. However, Wang et al. proposed that Au nanoparticle-modified BiVO₄ could only enhance the photocatalytic H₂O₂ production performance to a certain extent because the built-in potential generated by the upward bending of BiVO₄ might hamper the transfer of electrons [119]. Additionally, due to the different work functions of Au and BiVO₄, when Au interacts with BiVO₄, the negative charge density of Au increases, which weakens the adsorption of O₂ and HOO* on the Au surface, ultimately leading to a decrease in photocatalytic H₂O₂ production via direct one-step 2e^– ORR. Therefore, they introduced Cu in Au/BiVO₄ to form Cu@Au/BiVO₄ wherein the metallic Cu core not only promoted the photoinduced e^– transfer of BiVO₄ but also decreased the negative charge density of the metallic Au shell. As a result, Cu@Au/BiVO₄ possessed better photocatalytic activity for H₂O₂ production compared to Au/BiVO₄ (Fig. 4a). In another study, Yu's team used ultrathin g-C₃N₄ (CN) to modify Au/BiVO₄ [121]. The optimized CN-Au/BiVO₄ showed a H₂O₂ yield of 1.35 mmol/L in 2 h (420 nm LED, 50 mW/cm²), which was 2.65 times that of Au/BiVO₄. The authors proposed that the modification of BiVO₄ with g-C₃N₄ promoted the hole transfer and restrained H₂O₂ decomposition, while metallic Au promoted the electron transfer and boosted H₂O₂ production through a direct one-step 2e^– ORR pathway. Wang and co-workers prepared a serial of different Y doping contents BiVO₄ (Y: BVO) photocatalysts using a simple hydrothermal method [124]. The doping of Y promoted the adsorption of O₂ on the BiVO₄ surface and induced the formation of monoclinic/tetragonal BiVO₄ heterojunction, which greatly boosted the separation efficiency of charge carriers (Fig. 4b). Consequently, the optimized Y: BVO exhibited much higher activity for H₂O₂ production (114 µmol g^–1 h^–1) compared to pristine BiVO₄ (26 µmol g^–1 h^–1) under simulated sunlight irradiation via direct one-step 2e^– ORR pathway. Zhang's team prepared defective BiVO₄ for the effective generation of H₂O₂ from O₂ [125]. Due to the presence of oxygen vacancies, the adsorption of O₂ and interfacial electron transfer rate were 19- and 23-fold higher than that of pristine BiVO₄, respectively. Therefore, the coupling of charge carriers to chemisorbed species was effectively boosted, promoting the formation of H₂O₂ by direct one-step 2e^– ORR pathway under visible light (Fig. 4c). Wei et al. constructed NiCo₂O₄/BiVO₄ Z-scheme heterojunction and reported that the H₂O₂ yield rate over NiCo₂O₄/BiVO₄ under visible light was 16 times than that using pristine BiVO₄ [126]. The authors claimed that the desired electronic conductivity and electron transport characteristics of Ni^Ⅱ and Co^Ⅱ effectively delayed the recombination of photoinduced e^– and h⁺ and promoted the preferable indirect two-step e^– ORR pathway for improved H₂O₂ generation. Meanwhile, O₂^•– radical and ^•OH radical generated by such a heterojunction system improved the efficiency for degradation of methylene blue (Fig. 4d).

  Figure 4

  

  Figure 4.  (a) Schematic diagram of efficient photocatalytic H₂O₂ production by Cu@Au/BiVO₄. Reproduced with permission [119]. Copyright 2021, American Chemical Society. (b) Schematic illustration for the H₂O₂ production mechanism of Y-doped BiVO₄ photocatalyst. Reproduced with permission [124]. Copyright 2023, John Wiley and Sons. (c) H₂O₂ production on defective BiVO₄ via the direct one-step 2e^– ORR pathway. Reproduced with permission [125]. Copyright 2023, Elsevier. (d) Energy diagram for NiCo₂O₄/BiVO₄ heterojunction depicting the production of H₂O₂ and degradation of organic pollutants, simultaneously. Reproduced with permission [126]. Copyright 2022, American Chemical Society.

  DownLoad: Full-Size Img PowerPoint

  Bi₂WO₆ (band gap is about 2.8 eV) is another important visible-light-driven bismuth-based photocatalyst [138]. However, pristine Bi₂WO₆ exhibits unsatisfactory performance in H₂O₂ production due to poor separation of photoinduced charge carriers and in situ H₂O₂ consumption [139,140]. In order to enhance the photocatalytic activity of Bi₂WO₆ for the generation of H₂O₂, Zeng et al. in situ grew a red phosphorus/black phosphorus (RP/BP) heterostructure on Bi₂WO₆ nanosheets via a simple ultrasonic agitation method and ultimately constructed RP/BP-Bi₂WO₆ dual heterostructure photocatalyst [127]. The optimal 20% RP/BP-Bi₂WO₆ photocatalyst presented the highest H₂O₂ production rate, being 15-fold greater than that of pristine Bi₂WO₆ nanosheets under visible light. The enhanced H₂O₂ production was mainly ascribed to the heterojunctions, which facilitated highly efficient separation and transfer of charge carriers. Moreover, the authors confirmed that the RP/BP-Bi₂WO₆ photocatalyst produced H₂O₂ via both indirect two-step e^– ORR and indirect two-step e^– WOR.

  Bismuth oxyhalides are another important photocatalyst with a formula of Bi_aO_bX_c, where X = Cl, Br and I. Their VB top is mainly occupied by O 2p states as well as a considerable contribution of X δ_p states (δ = 3, 4 and 5, corresponding to X = Cl, Br, and I, respectively), while their CB bottom is dominated by Bi 6p states [141,142]. Due to the unique electronic structure, the light absorption range of Bi_aO_bX_c can extend from the ultraviolet to visible light range [143,144]. Especially, the internal electric field induced by Bi_aO_bX_c can greatly boost the separation and transfer of charge carriers [145]. Therefore, bismuth oxyhalides exhibit excellent photocatalytic performance in pollutant degradation, CO₂ reduction and water oxidation. Recently, Xu et al. investigated the facet-dependent photocatalytic H₂O₂ synthesis on BiOCl and found that the H₂O₂ production rate of BiOCl with exposed (001) facet is about two times than that of (010) facet, which was attributed to the higher charge separation efficiency of BiOCl (001) facet [128]. Electron paramagnetic resonance (EPR) measurements and trapping experiments revealed that the BiOCl (001) produced H₂O₂ mainly via indirect two-step e^– ORR, while photocatalytic generation of H₂O₂ on BiOCl (010) involved indirect two-step e^– ORR and indirect two-step e^– WOR pathways simultaneously. In another study, Gong et al. constructed a BiOCl/BiOBr heterojunction by in situ generating BiOCl nanosheets on the BiOBr surface, followed by NH₄Cl etching [129]. This heterojunction boosted the separation of charge carriers, thereby significantly enhancing the productivity of H₂O₂. Besides, Zhang et al. synthesized BiOBr/COF heterojunction photocatalysts [133], which facilitated a spatial separation and transfer of charge carriers and enhanced redox power. Additionally, by introducing COF, O₂ was adsorbed on the TpPaCl₂ surface in a lying-down geometry, which led to the reduction and protonation of O₂ to form H₂O₂ via direct one-step 2e^– ORR pathway in the presence of ethanol as a sacrificial agent. Consequently, BiOBr/COF displayed the H₂O₂ production yield of 3749 µmol g^–1 h^–1 (under 300 W Xe irradiation), which was 27 times higher than BiOBr and 1.85 times higher than COF.

  Bi₂O₂CO₃ is also highly promising for photocatalysis owing to its unique layered crystal structure, which induces an internal electric field and enhances the separation and transfer of photogenerated carriers [117,146]. As reported by Dong et al., Bi₂O₂CO₃ possesses a wide band gap of about 3.1−3.5 eV [147]. Additionally, the appropriate band potential of Bi₂O₂CO₃ makes it suitable for the photocatalytic synthesis of H₂O₂ [136]. Recently, Wang et al. prepared a Bi₂O₂CO₃/Bi-MOF Z-scheme heterojunction, significantly improving the photocatalytic performance of H₂O₂ formation via the direct one-step 2e^– ORR pathway [136]. With the close-contact heterostructure and internal electric field under simulated sunlight irradiation and methanol as a sacrificial agent, the H₂O₂ yield rate (280.5 µmol g^–1 h^–1) was 11.23- and 6.06-fold higher than that of pristine Bi₂O₂CO₃ (24.97 µmol g^–1 h^–1) and Bi-MOF (46.31 µmol g^–1 h^–1), respectively.

  3.4 g-C₃N₄

  g-C₃N₄, as one of the important organic polymer materials [148-150], has been thoroughly studied for different photocatalytic applications owing to its awesome optical, electronic, and chemical properties. The photocatalytic generation of H₂O₂ using g-C₃N₄ was first reported by Hirai's group as early as 2014 [151]. Efficiently H₂O₂ was generated via direct one-step 2e^– ORR using alcohol as an electron donor. It is also uncovered that the generation of 1,4-endoperoxide species on photoactivated g-C₃N₄ restrained indirect two-step e^– ORR and favored direct one-step 2e^– ORR pathway to the generation of H₂O₂. Nevertheless, pristine g-C₃N₄ suffered from low photoinduced charge carrier separation efficiency and slow kinetics [152,153]. Therefore, the photocatalytic performance of g-C₃N₄ was not satisfactory, and the H₂O₂ produced in a 12-h reaction (λ > 420 nm) was only 30 µmol. Therefore, strategies such as structural modification, doping, defect engineering, and heterojunction construction were applied to improve the photocatalytic ability of g-C₃N₄ to produce H₂O₂ (Table S4 in Supporting information) [15,40,151,154-203].

  Subsequently, Hirai's group modified g-C₃N₄ with pyromellitic diimide (PDI) [154], biphenyl diimide (BDI) [155] and mellitic triimide (MTI) [156], achieving efficiently H₂O₂ production from H₂O and O₂ without using any sacrificial agents. When modified with PDI, the formation of H₂O₂ (λ > 420 nm, 48 h) over g-C₃N₄/PDI₅₁ (50.6 µmol) was more than 250 times that of pristine g-C₃N₄ (< 0.2 µmol). Such attractive performance originated from the incorporation of PDI, which shifted the redox potentials of g-C₃N₄, enabling g-C₃N₄/PDI to oxidize water while maintaining high selectivity for 2e^– reduction of O₂. Consequently, H₂O₂ was produced from H₂O and O₂ over g-C₃N₄/PDI via direct one-step 2e^– ORR pathway coupled with 4e^– WOR pathway (Fig. 5a) [154]. When g-C₃N₄ was modified with BDI instead of PDI, the charge separation was enhanced, and the apparent quantum yield at 420 nm was increased from 2.6% to 4.6% [155]. After incorporating MTI into g-C₃N₄, the SCC efficiency reached 0.18% and the apparent quantum yield (AQY) was increased to 6.2% at 420 nm. The enhancement was attributed to the creation of a condensed melem layer, which improved intra- and interlayer transfer of the photoinduced charge carriers [156]. Furthermore, the authors prepared g-C₃N₄/PDI/rGO nanohybrids, which further enhanced the formation of H₂O₂ with an SCC efficiency of 0.20% [157]. It is revealed that the photoinduced e^– from g-C₃N₄/PDI was transferred to rGO, promoting the separation of photoinduced charge carriers while preserving high selectivity for the 2e^– reduction of oxygen. As a result, the h⁺ in g-C₃N₄/PDI oxidized H₂O through 4e^– WOR pathway, while the e^– on rGO promoted the reduction of O₂ via a direct one-step 2e^– ORR pathway. In another study, Zhang's team modified g-C₃N₄ with polyethylenimine (PEI), which tuned the separation of charge carriers and the pathway of oxygen reduction [159]. The optimal PEI/g-C₃N₄ photocatalyst showed efficient H₂O₂ production with a production rate of 208.1 µmol g^–1 h^–1 (under simulated solar light), being 25 times greater than that of pristine g-C₃N₄. Specifically, the incorporating of PEI promoted the generation of H₂O₂ via the indirect two-step e^– ORR pathway, which was different from the direct one-step 2e^– ORR pathway when using pristine g-C₃N₄.

  Figure 5

  

  Figure 5.  (a) Mechanism of PDI-modified g-C₃N₄ for photocatalytic H₂O₂ production from H₂O and O₂. Reproduced with permission [154]. Copyright 2014, John Wiley and Sons. (b) The different adsorption modes of O₂ adsorption on metal surfaces and the mechanism of photocatalytic H₂O₂ production. Reproduced with permission [172]. Copyright 2021, Springer Nature. (c) Mechanism of an aryl amino-substituted g-C₃N₄ photocatalyst with atomically dispersed Mn for photocatalytic H₂O₂ production from seawater. Reproduced with permission [163]. Copyright 2023, American Chemical Society. (d) Mechanism of heteroatom doped g-C₃N₄ (AKMT) for photocatalytic H₂O₂ production. Reproduced with permission [174]. Copyright 2020, John Wiley and Sons. (e) The presence of C-vacancies changed the photocatalytic H₂O₂ production from the indirect two-step e^– ORR pathway to the direct one-step 2e^– ORR pathway. Reproduced with permission [181]. Copyright 2016, Elsevier. (f) Mechanism of PI_x-NCN heterojunction for photocatalytic H₂O₂ production. Reproduced with permission [197]. Copyright 2017, Elsevier. (g) Mechanism of heptazine/triazine layer stacked g-C₃N₄ heterojunction with F/K dual sites modified for photocatalytic H₂O₂ production. Reproduced with permission [200]. Copyright 2022, American Chemical Society.

  DownLoad: Full-Size Img PowerPoint

  Single-atom catalysts, possessing unique electronic characteristics and theoretically enabling 100% metal atom utilization, show great potential as cocatalysts for enhancing the efficiency of photocatalysis when combined with semiconductors [204,205]. Incorporating single atoms into g-C₃N₄ has been reported to boost the activity and selectivity of the 2e^– ORR pathway. In general, there are three types of O₂ adsorption on metal surfaces: Pauling-type (end-on), Griffiths-type (side-on) and Yeager-type (side-on) [206]. Among these, side-on adsorption facilitates the cleavage of O−O bond, resulting in high selectivity for 4e^– ORR. Conversely, end-on adsorption minimizes O−O bond cleavage, inhibiting 4e^– ORR and promoting 2e^– ORR [172]. For single atoms, O₂ molecules typically undergo end-on adsorption, reducing the possibility of O−O bond cleavage. For instance, Teng et al. deposited various single atoms on g-C₃N₄ including Fe, Co, Ni, In, Sn and Sb, leading to markedly promoted charge separation and facilitated end-on adsorption of O₂ (Fig. 5b) and thus enhanced selectivity of 2e^– ORR pathway [172,206]. In another example, Zhang et al. prepared highly loaded Ni single atoms on g-C₃N₄ and observed the structural evolution of active sites (Ni-N₃ → O₁Ni-N₂ → HOO—Ni-N₂) upon O₂ adsorption. As the energy barrier of key *OOH intermediate was remarkably reduced, the formation of H₂O₂ was promoted [15]. Very recently, Ren et al. prepared an Mn atomically dispersed on aryl amino-substituted g-C₃N₄ photocatalyst (Mn/AB-C₃N₄), exhibiting excellent photocatalytic activity and stability for the generation of H₂O₂ from seawater without the addition of sacrificial agents [163]. By introducing aryl amino units, the band gap was shortened, and the light absorption was boosted. Additionally, the presence of coordinatively unsaturated Mn-N_x sites not only stimulated the adsorption and activation of O₂ but also promoted the separation of photoinduced charge carriers and suppressed their recombination. Hence, the Mn/AB-C₃N₄ exhibited outstanding photocatalytic performance of H₂O₂ production. Most importantly, different from the usual direct or indirect ORR pathway using g-C₃N₄-based photocatalysts, H₂O₂ produced over Mn/AB-C₃N₄ via a less common WOR route in which direct one-step 2e^– WOR and indirect two-step e^– WOR pathways occurred simultaneously (Fig. 5c).

  Apart from surface modification, the production of H₂O₂ over g-C₃N₄ can also be enhanced by ion doping. Choi's team demonstrated that in situ doping of K, P and O into g-C₃N₄ (KPD-CN-x, x represents the amount of substance added dibasic potassium phosphate) remarkably improved the photocatalytic performance of H₂O₂ formation via direct one-step 2e^– ORR without loading noble-metal cocatalysts [169]. The AQY at 420 nm for the formation of H₂O₂ obtained by using the optimized KPD-CN-7.5 photocatalyst was 25-fold higher than that of pristine g-C₃N₄. The authors also prepared heteroatom (S and K)-doped polymeric C₃N₄, which exhibited an AQY at 450 nm approaching 100% for the H₂O₂ formation [174]. The synergistic association of K and S dopants in polymeric C₃N₄ was suggested to boost the separation of interlayer charge carriers and the polarization of trapped electrons, which is beneficial for the capture and reduction of O₂ in ORR kinetics, as indicated by the spectroscopic analysis in spatiotemporal resolution and DFT calculations (Fig. 5d). In another study, Liu et al. doped polymeric C₃N₄ with dual-heteroatom (O and K), and verified that the as-prepared O/K-CN facilitated the electron transfer from the β spin-orbital to π* orbital of O₂ compared to other C₃N₄, thereby optimizing the activation of O₂ to enhance the formation of *OOH intermediate and lowering the energy barrier for the generation of H₂O₂ [176].

  Creating surface defects in g-C₃N₄ is another important strategy to improve their photocatalytic activity. Two types of surface defects, C- and N-vacancies, in g-C₃N₄, have shown advantages in H₂O₂ production [181,184]. For instance, Wang et al. introduced C-vacancies on the surface of g-C₃N₄ via calcination treatment and investigated the effects of C-vacancies on the O₂ reduction over g-C₃N₄ [181]. It is revealed that the C-vacancies significantly enhanced the activity of g-C₃N₄ in reducing O₂ to generate H₂O₂ without the addition of any organic sacrificial agents. The authors observed that the adsorption strength of O₂ on C_v-g-C₃N₄ was stronger than that on pristine g-C₃N₄ by temperature-programmed desorption spectra. Additionally, the nitrogen atoms of the amino group around C-vacancies acted as a Lewis base. Since O₂ functions as a Lewis acid, these nitrogen atoms can interact with O₂ through Lewis acid-base interactions, thereby facilitating electron transfer. As a result, the route for reducing O₂ to product H₂O₂ changed from indirect two-step e^– ORR to direct one-step 2e^– ORR (Fig. 5e). In another study, Xie et al. prepared g-C₃N₄ with two types of cooperative N-vacancies (NH_x and N₂C) through simple one-pot potassium hydroxide assisted calcination [186]. The optimized photocatalyst exhibited a H₂O₂ yield of 152.6 µmol/h under simulated sunlight irradiation, which was 15-fold higher than that of pristine g-C₃N₄. Both the experimental and DFT results uncovered that NH_x and N₂C vacancies boosted the separation of photoinduced charge carriers as well as the activation of oxygen via an indirect two-step e^– ORR pathway. Furthermore, Zheng and his co-authors designed a g-C₃N₄ with dual defect sites (–C≡N groups and N-vacancies) for the overall photosynthesis of H₂O₂ [179]. Based on the experimental results and DFT calculations, the authors discovered that the synergistic effect between –C≡N groups and N-vacancies enhanced the absorption of visible light and facilitated carrier separation. Specifically, the N-vacancies enhanced the adsorption and activation of O₂, while the –C≡N groups facilitated the adsorption of H⁺, thereby synergistically promoting H₂O₂ production through the direct one-step 2e^– ORR. However, it has recently been reported that constructing g-C₃N₄ with a high crystalline structure and eliminating defects can also significantly improve the photocatalytic activity of g-C₃N₄ [187]. For example, a highly crystalline g-C₃N₄ with extremely low defect concentrations was synthesized by a molten salt method. The highly crystalline g-C₃N₄ obtained by using thiourea as the precursor showed the highest H₂O₂ production rate (2.48 mmol g^–1 h^–1) and AQY of 22% at 400 nm. In another work, Shiraishi et al. explored the effect of surface defects of g-C₃N₄ on the photocatalytic activity of H₂O₂ synthesis [182]. It was found that the surface of mesopore g-C₃N₄ with a larger surface area has abundant primary amine groups and these defects served as active sites not only for the 4e^– ORR to reduce the selectivity of H₂O₂, but also for decomposing the formed H₂O₂. Therefore, g-C₃N₄ with a relatively large surface area but extremely low defect concentrations is considered to promote efficient H₂O₂ production. Controlling defects on the g-C₃N₄ surface is therefore crucial for enhancing photocatalytic H₂O₂ production.

  Constructing heterojunction is an effective approach to regulate the behaviour of charge carriers in g-C₃N₄, enhancing their spatial separation and migration. For example, Wang and his co-authors built a Z-scheme heterojunction PI_x-NCN by assembling perylene imide (PI) on g-C₃N₄ nanosheets (NCN) [197]. The authors claimed that the photoexcited electrons transferred from the CB of perylene imide to the VB of NCN, promoting the charge separation and enabling more electrons for the reduction of O₂ to generate H₂O₂. Most importantly, the perylene imide changed the photocatalytic H₂O₂ production from an indirect two-step e^– ORR single approach to indirect two-step e^– ORR and indirect two-step e^– WOR hybrid approaches, thus significantly enhancing the H₂O₂ generation under visible light (Fig. 5f). In another study, Li et al. prepared a heptazine/triazine layer stacked g-C₃N₄ heterojunction with F/K dual sites (FKHTCN) [200]. It has been found that the formation of a heptazine/triazine heterostructure leads to a reduction in the interlayer distance of aromatic units, while the incorporation of F/K dual sites adjusts the layer-component ratio. As a result, the dissociation of interlayer exciton was promoted and the lifetime of charge carriers was extended. The introduced F/K dual sites were also found to enhance the adsorption of O₂ and intermediate *OOH, optimizing the ORR pathway and consequently improving the photocatalytic production of H₂O₂ by 2e^– ORR (Fig. 5g).

  3.5 MOFs

  MOFs are a novel class of structurally defined porous organic polymers consisting of metal clusters (Ti, Zr, Ni, Cu, etc.) and organic ligands [207,208]. Since the first reported application of MOF-5 for photocatalytic reaction in 2007 [209], MOFs have been widely employed in photocatalytic elimination of organic contaminants [210], overall water splitting [211], CO₂ reduction [212], and ammonia production [213] due to their high structural and functional tunability. In addition, some MOFs possess the properties of visible light response resulting from their amine-functionalized linkers [214-219], which greatly boost the utilization of solar energy. Therefore, MOFs are regarded as promising photocatalysts for H₂O₂ production.

  The generation of H₂O₂ using MOF photocatalysts was first reported by Yamashita's group in 2018 [220]. The authors deposited NiO nanoparticles on MIL-125-NH₂, a commonly used MOF composed of Ti₈O₈(OH)₄ clusters and 2-aminoterephthalic acid ligands, to synthesize Ni/MIL-125-NH₂. In the presence of electron donors (such as H₂O, and benzyl alcohol), Ni/MIL-125-NH₂ generated H₂O₂ via an indirect two-step e^– ORR pathway, and the deposition of NiO accelerated the formation rate of H₂O₂. However, the photocatalytic activity of pristine MOFs is limited due to their short light adsorption edge (at 345 nm) [221].

  Therefore, in pursuit of higher efficiency for H₂O₂ production, it is particularly important to develop visible-light-driven MOF-based photocatalysts that can both perform efficient H₂O₂ formation and suppress H₂O₂ decomposition. To this end, several strategies have been explored including surface modification, defect engineering, ion doping, and interfacial engineering to improve the photocatalytic activity of visible-light-driven MOFs for H₂O₂ production (Table S5 in Supporting information) [220,222-240].

  To enhance the visible light absorption and photocatalytic activity of MIL-125-NH₂ for H₂O₂ generation, Yan et al. synthesized Ce-doped MIL-125-NH₂ by impregnation-coordination method [235]. The results demonstrated that the introduction of Ce optimized the electronic structure and band gap of MIL-125-NH₂, and the formed Ce-O-Ti charge-transfer transitions caused a band-to-band red shift of absorption edge. Additionally, the Ce⁴⁺/Ce³⁺ and Ti⁴⁺/Ti³⁺ redox pairs acted as reaction sites and facilitated the transfer of charge carriers, thereby boosting the generation rate of H₂O₂ via indirect two-step e^– ORR (Fig. 6a).

  Figure 6

  

  Figure 6.  (a) Proposed reaction mechanism for the generation of H₂O₂ over Ce-doped MIL-125-NH₂. Reproduced with permission [235]. Copyright 2023, John Wiley and Sons. (b) The proposed mechanism of H₂O₂ production over alkylated MIL-125-R7 in a two-phase system. Reproduced with permission [224]. Copyright 2019, John Wiley and Sons. (c) Reaction mechanism of H₂O₂ production over missing-linker defective UiO-66-NH₂-X photocatalyst. Reproduced with permission [236]. Copyright 2021, American Chemical Society. (d) The relationship between the yield of H₂O₂ production and heterojunction coverage, as well as the proposed mechanism over NH₂−MIL-125@ZnS. Reproduced with permission [238]. Copyright 2022, John Wiley and Sons.

  DownLoad: Full-Size Img PowerPoint

  Recently, Yamashita and co-authors modified MIL-125-NH₂ with various alkyl chains and achieved further-enhanced photocatalytic activity for H₂O₂ production [224,225]. It has been suggested that O₂ was first reduced to O₂^•– radicals by Ti^Ⅲ species in the benzyl alcohol phase, and subsequently the formed O₂^•– radicals transferred to aqueous phase to generate H₂O₂. Besides, the production of H₂O₂ was accelerated in the presence of H⁺ and Na⁺ [224]. Therefore, by utilizing hydrophobic MOFs, spontaneous separation of products was achieved, thereby favouring the production of H₂O₂ (Fig. 6b).

  In another example, Kondo et al. investigated the effect of missing-linker defects on the photocatalytic H₂O₂ production performance over UiO-66-NH₂-X [236]. By adjusting the amount of acetic acid before solvothermal treatment, the concentration of missing linker defects introduced into the UiO-66-NH₂ was altered. As a result, the formation of H₂O₂ via an indirect two-step e^– ORR pathway was dramatically improved, while the decomposition of H₂O₂ was restrained (Fig. 6c).

  Very recently, Liu et al. anchored ZnS on the corners of the NH₂−MIL-125 body and constructed a NH₂−MIL-125@ZnS heterojunction with tunable heterojunction coverage [238]. The obtained NH₂−MIL-125@ZnS heterojunction with strong interface interactions is critical to photocatalytic H₂O₂ production. The heterojunction coverage serves as a vital parameter that affects the formation of e^–-h⁺ pairs and charge separation. Additionally, it determines light absorption and the accessibility of photoinduced electrons. Consequently, the NH₂−MIL-125@ZnS heterojunction, with an average heterojunction coverage of 45.1%, exhibited optimal photocatalytic activity, achieving a H₂O₂ production rate of 120 mM g^–1 h^–1 under visible light via indirect two-step e^– ORR pathway (Fig. 6d).

  3.6 COFs

  Since their discovery in 2005, COFs have proven to be promising metal-free materials for their diverse structures and precise functionalities [241-243]. Of particular interest, 2D COFs structurally resemble graphite, with building blocks covalently bonded within the walls to form 2D planes and close π-π stacking between layers [244]. Such stacked 2D layered COFs form conductive columns, providing ideal channels for the rapid diffusion and transfer of photoinduced charge carriers [245].

  In 2014, Stegbauer et al. first reported a hydrazone-based COF for the photocatalytic H₂ evolution, with quantum efficiencies similar to g-C₃N₄ [246]. Subsequently, various COFs have been synthesized and used for a wide variety of photocatalytic reactions [26,247-250]. COFs emerged as extremely promising candidates for photocatalytic H₂O₂ production [251-253], exhibiting comparable photocatalytic activity to inorganic materials. In 2019, Xu's group utilized COFs for photocatalytic H₂O₂ production for the first time [251]. They designed three different 2D covalent triazine frameworks (CTFs), a branch of the COFs, by embedding acetylene or diacetylene moieties into CTFs (CTF-(1,1′-biphenyl)-4,4′-dicarbonitrile (CTF-BPDCN), CTF-4,4′-(ethyne-1,2-diyl)dibenzonitrile (CTF-EDDBN) and CTF-4,4′-(buta-1,3-diyne-1,4-diyl)dibenzonitrile (CTF-BDDBN)) to extend π-conjugation, thereby finely tuning the electronic structure of CTFs. The rotating disk electrode and EPR measurements results demonstrated all of these three CTFs reduced O₂ into H₂O₂ by direct one-step 2e^– ORR. The H₂¹⁸O isotopic labelling experiments identified that CTF-BPDCN oxidized H₂O to O₂ via 4e^– WOR while CTF-EDDBN and CTF-BDDBN directly oxidized H₂O to H₂O₂ through direct one-step 2e^– WOR pathway. The authors claimed that acetylene and diacetylene units on CFTs significantly reduced the Gibbs free energy (ΔG) for the formation of OH* intermediates, enabling CTF-EDDBN and CTF-BDDBN to produce H₂O₂ from H₂O via direct one-step 2e^– WOR pathway. Therefore, different reaction pathways toward H₂O₂ photosynthesis using CTF-BPDCN, CTF-EDDBN, and CTF-BDDBN were proposed. Among them, CTF-BDDBN exhibited the highest photocatalytic H₂O₂ generation without any sacrificial agents. These acetylene and diacetylene functionalized CTFs realized 100% atom utilization efficiency, yet the SCC efficiency is still limited to 0.1%. Hence, much effort has been devoted to further advancing COFs and optimizing the photocatalytic generation of H₂O₂ (Table S6 in Supporting information) [30,46-48,50-52,251,252,254-294].

  An effective strategy to functionalize COFs at a molecular level is to optimize their structures and properties by modifying the starting linker with functional groups, such as thiourea [261], piperazine [256], sulfone [262,295], thiophene [259] and diazine [260]. For example, Han and co-workers found that thiourea functionalized CTF (Bpt-CTF) functioned as an active photocatalyst for H₂O₂ production even without sacrificial agents [261]. By introducing the thiourea functional group into COFs, the charge separation/transport and proton transfer were greatly improved, resulting in a high H₂O₂ production rate (3268.1 µmol h^–1 g^–1) with an impressive AQY of 8.6% (at 400 nm). H₂O₂ was produced via coupled direct one-step 2e^– ORR with the 4e^– WOR pathway. In another example, Zhi et al. synthesized piperazine-linked porous CoPc-based COFs, which showed strong visible light absorption (at 400–1000 nm) and enhanced the separation/transfer efficiency of photoinduced charge carriers, thereby boosting the H₂O₂ production via indirect two-step e^– ORR with a rate of 2096 µmol h^–1 g^–1 [256]. Luo et al. also achieved an optimized visible-light-driven photocatalytic H₂O₂ production yield of 3904.2 µmol h^–1 g^–1 without any sacrificial agents by grafting sulfone functional group onto COFs (FS-COFs) [262]. Based on experimental results and theoretical calculations, the authors found that modification of the sulfone functional group not only accelerated the separation of photoinduced charge carriers but also enhanced the protonation of COFs and boosted the adsorption of O₂ molecular in a "side-on" Yeager-type. As a result, FS-COFs altered the photocatalytic ORR process from indirect two-step e^– ORR to direct one-step 2e^– ORR, achieving efficient H₂O₂ production (Fig. 7a). Moreover, Yue et al. designed and prepared two thiophene-containing COFs (TDCOF and TT-COF), which realized overall photosynthetic H₂O₂ production by indirect two-step e^– ORR and direct one-step 2e^– WOR dual-channel pathways in the absence of sacrificial agents [259]. By adjusting the N-heterocycle units (pyridine and triazine) in COFs, the generation of H₂O₂ was regulated. Using only air and water as the reactants without the addition of sacrificial agents, the photocatalytic H₂O₂ generation rates over TDCOF in natural seawater and deionized water were 3364 µmol g^–1 h^–1 and 4060 µmol g^–1 h^–1, respectively. The H₂O₂ yield in natural seawater is slightly lower than that in deionized water due to the presence of plenty of ions, which may interfere with the production or cause the decomposition of H₂O₂. In addition, theoretical calculations confirmed that thiophene was the primary photoreduction unit for ORR, while the benzene ring connected to thiophene through an imine bond was the main photooxidation unit for WOR. Very recently, Xi and co-workers synthesized a serious of COFs functionalized with different diazines, including pyridazine, pyrazine, and pyrimidine (denoted as TpDz, TpPz and TpMd) and in-depth studied the effects of relative nitrogen locations of diazines functionalized COFs on the photocatalytic production of H₂O₂ from O₂ and pure water [260]. The intercalation of pyridazine in TpDz preferred to stabilize the endoperoxide intermediate compared to pyrimidine and pyrazine, leading to a more efficient direct one-step 2e^– ORR pathway. In addition, TpDz possessed higher charge carrier separation efficiency and lower charge transfer resistance. Consequently, TpDz produced H₂O₂ via the direct one-step 2e^– ORR and 4e^– WOR pathways with a photocatalytic H₂O₂ yield of 7327 µmol h^–1 g^–1 and 0.62% SCC efficiency under visible light irradiation (λ > 420 nm).

  Figure 7

  

  Figure 7.  (a) Reaction mechanism of photocatalytic H₂O₂ production over sulfone functional group modified COF (FS-COFs). Reproduced with permission [262]. Copyright 2023, John Wiley and Sons. (b) Proposed reaction mechanism for the generation of H₂O₂ over fluorine-doped COF (TF₅₀–COF). Reproduced with permission [283]. Copyright 2022, John Wiley and Sons. (c) Schematic illustration of photocatalytic reaction pathways toward H₂O₂ production over Z-scheme WO₃/COF heterojunction. Reproduced with permission [288]. Copyright 2023, Royal Society of Chemistry. (d) The schematic diagram of photocatalytic H₂O₂ production over CHF-DPA and CHF-DPDA with separated redox centres. Reproduced with permission [46]. Copyright 2022, John Wiley and Sons.

  DownLoad: Full-Size Img PowerPoint

  As another alternative strategy, ion doping is a traditional method to modulate the electronic band filling and the transport of charge carriers in organic polymers, thereby improving the photocatalytic activity for H₂O₂ production [296]. For instance, Wang et al. developed and designed a highly crystalline fluorine-doped COF (TF-COF) [283]. By partially fluorinating the edge of aromatic parts of 2D COF (TF₅₀–COF), plenty of Lewis acid sites were generated, adjusting the electron distribution of adjacent C atoms. TF₅₀–COF with AA stacking mode was found to exhibit the strongest π-π interactions and the highest crystallinity compared to non-fluorinated COF and TF-COF, thereby effectively promoting the separation and transfer of charge carriers. The UV–vis absorption spectra showed that the visible-light-responsive range of TF₅₀–COF was also widened. Furthermore, the adsorption of O₂ was significantly enhanced due to the formation of *OOH intermediates. Consequently, the TF₅₀–COF exhibited remarkable photocatalytic H₂O₂ performance (1739 µmol h^–1 g^–1) with an AQY of 5.1% at 400 nm via indirect two-step e^– ORR H₂O₂ production approach (Fig. 7b). Very recently, Guo's team rationally engineered F-containing COFs (TAPT-TFPA COFs) to strongly confine Pd metal-isolated clusters (ICs) for enhanced H₂O₂ photosynthesis [284]. The results showed that the strong electronegative F in TAPT-TFPA COFs@Pd ICs not only strengthened the metal-support interaction for improving the photochemical stability but also optimized the D-band centre of Pd ICs. Therefore, TAPT-TFPA COFs@Pd ICs exhibited an outstanding photocatalytic H₂O₂ production activity (2143 µmol h^–1 g^–1), spectacular SCC efficiency (0.82%), and high stability over 100 h in an aqueous solution containing 10% ethanol under simulated solar irradiation (AM 1.5). H₂O₂ was formed via an indirect two-step e^– ORR pathway.

  Additionally, the photocatalytic H₂O₂ production using COFs can be significantly boosted by the construction of heterojunctions. Yao and co-workers constructed ZnIn₂S₄/TpPa-1 heterostructure for photocatalytic H₂O₂ evolution in air [285]. Under visible light irradiation, this hybrid photocatalyst efficiently absorbed visible light and captured O₂ in air, achieving effective H₂O₂ production via an indirect two-step e^– ORR pathway. In another study, Yang et al. prepared WO₃/Tp-TAPB heterojunction and observed a high H₂O₂ production rate in pure water, which was 72.3- and 2.8-fold than that of WO₃ and Tp-TAPB, respectively [288]. The heterojunction structure possessed a large specific surface area, providing increased active sites. Additionally, it facilitated the separation and transfer of charge carriers and maintained a high redox potential. Ultimately, the WO₃/Tp-TAPB achieved H₂O₂ formation with high efficiency via indirect two-step e^– ORR and indirect two-step e^– WOR two-channel pathways (Fig. 7c).

  Construction donor-acceptor (D-A) structure is one of the common and valid approaches to affect the efficiency of photo-capture, and photoinduced excitons splitting in COFs [297,298]. This is because the existence of a polarization electric field facilitates the separation of photoinduced charge carriers in a D-A structure through intramolecular charge transfer. Additionally, this unique spatial separation feature greatly reduces the exciton binding energy, thereby inhibiting the recombination of the charge carriers [299]. Thus, in recent years, various donor and acceptor units, especially alternating donor and acceptor building blocks with planar structure, have been designed to build D-A configuration in COFs, achieving considerable photocatalytic activity in H₂O₂ production. For example, Xu and coworkers first reported heptazine-based COFs (CHF-DPA and CHF-DPDA) in 2021, which possessed spatially separated redox centres and efficiently produced H₂O₂ from H₂O and O₂ under visible light irradiation without additional sacrificial agents [46]. They used femtosecond time-resolved transient absorption and X-ray absorption near-edge structure (XANES) spectroscopy to study the photoexcited carrier dynamics and charge transfer process in these heptazine-based COFs and found that photoinduced e^– was concentrated in s-heptazine units, whereas photoinduced h⁺ was accumulated in acetylene or diacetylene parts. Moreover, theoretical and experimental results confirmed that the C atoms in s-heptazine units acted as the reduction centres favouring the formation of *H and *OO* intermediates, leading to direct one-step 2e^– ORR pathway, while the C atoms in acetylene or diacetylene groups most likely served as oxidation active sites, beneficial for the generation of OH* intermediates, resulting in direct one-step 2e^– WOR pathway (Fig. 7d). As reported, the high-N-content s-heptazine and triazine moieties often serve as the O₂ reduction centres for H₂O₂ generation due to their ability to stabilize suitable intermediates [261,290]. Besides, benzene groups were proven to be the reaction centre for the oxidation of H₂O to generate O₂ [300]. Inspired by this, Chen et al. designed and prepared a 2D s-heptazine-based COFs (HEP-TAPT-COF) with spatially separated redox centres at the molecular level for the production of H₂O₂ from O₂ and pure water [48]. HEP-TAPB-COF, integrating dual O₂ reduction centres of triazine and s-heptazine units, showed excellent photocatalytic performance in producing H₂O₂ with a 0.65% SCC efficiency and an AQY of 15.35% at 420 nm. The experimental and DFT investigations demonstrated that the spatially separated redox centres facilitated the separation of charge carriers in this conjugated structure. As a result, HEP-TAPB-COF reduced O₂ toward H₂O₂ with high selectivity via a direct one-step 2e^– ORR pathway and directly oxidised H₂O into O₂ by 4e^– WOR pathway. Additionally, the H₂¹⁸O isotopic labelling experiments confirmed that the O₂ generated from H₂O oxidation in turn as a reactant to promote H₂O₂ generation. In another study, Chang et al. prepared an oxidation–reduction molecular junction TTF-BT-COF by covalent coupling of tetrathiafulvalene (TTF) and benzothiazole (BT) [52]. Interestingly, TTF groups served as oxidation sites and BT parts acted as reduction sites, achieving the full reaction of photosynthesis for H₂O₂ generation through direct one-step 2e^– ORR and indirect two-step e^– WOR two-channel pathways. Tong's group also designed and synthesized a novel D-A configuration COFs by optimizing the intramolecular polarity of COFs by introducing an appropriate amount of phenyl moieties as the electron donors, which significantly increased the direct photosynthesis of H₂O₂ from pure water and air [50]. Theoretical and experimental results revealed that appropriate N 2p and C 2p states with optimal intramolecular polarity in COF-N32 effectively lowered the energy barrier for H₂O oxidation and O₂ reduction, respectively. Without the use of any sacrificial agent, COF-N32 exhibited a high photocatalytic H₂O₂ production rate (605 µmol g^–1 h^–1), SCC efficiency of 0.31% and AQY of 6.2% at 459 nm. H₂O₂ was formed by coupling the indirect two-step e^– ORR and indirect two-step e^– WOR dual pathways. Very recently, Wang and coworkers prepared a benzotrithiophene-based D-A configuration COFs (TaptBtt) for H₂O₂ photosynthesis from H₂O and O₂ under visible light irradiation without additional sacrificial agents [30]. Due to the spatially separated redox centres, the reduction sites were predominantly centred on C atoms in the imine bond of TaptBtt, which regulated the direction of charge transfer and the energy difference of intramolecular D-A. This further reduced the Gibbs free energy of OH* and OOH* intermediates, facilitating the generation of H₂O₂ via the synchronous indirect two-step e^– ORR and direct one-step 2e^– WOR two-channel pathways.

  4. Conclusions and perspectives

  As the global demand for H₂O₂ rises sharply, it is particularly important to develop stable and efficient photocatalysts for H₂O₂ production. In this review, we thoroughly outline the recent progress in designing 2D photocatalysts, including inorganic materials (e.g., metal oxides, metal chalcogenides and bismuth-based materials) and organic polymers (e.g., g-C₃N₄, MOFs and COFs), for the high-efficiency photocatalytic H₂O₂ production. Moreover, common methods for detecting and quantifying H₂O₂ are presented. We have also highlighted various rational and effective strategies to promote photocatalytic H₂O₂ production and discussed the related mechanisms. Specifically, strategies including surface modification, ion doping, defect engineering, and heterojunction construction are employed to expand the light absorption range, retain the high redox capacity, boost the separation/transfer of charge carriers and enhance the adsorption of O₂ and redox sites entrenchment.

  Although significant progress has been achieved in photocatalytic H₂O₂ production, practical and large-scale industrial production remains elusive, and several key challenges still need to be addressed.

  Firstly, there is a need to gain a comprehensive understanding of the mechanisms underlying efficient and sustainable generation of H₂O₂ [27,301]. The interplay between oxygen reduction and water oxidation half-reactions plays a crucial role in governing the kinetics and efficiency of photocatalytic H₂O₂ production [293]. It is worth noting that previous studies have primarily focused on ORR, while the significance of WOR has often been overlooked, despite its equal importance [25,302-305]. Up to now, the high-efficiency photocatalytic ORR for producing H₂O₂ has been achieved, but it typically requires the addition of sacrificial agents like methanol, ethanol, benzyl alcohol, or isopropanol. However, the use of these sacrificial agents may lead to the generation of by-products and complicate the subsequent separation and purification of H₂O₂ [306]. Alternatively, exploring the coupling of ORR with other oxidation reactions to simultaneously generate value-added products could be considered [293]. Therefore, more attention should be given to WOR and non-sacrificial H₂O₂ photosynthesis in the future, even though water oxidation presents more challenges [17]. In this scenario, it is imperative to investigate both ORR and WOR to optimize the reaction process, particularly the meticulous integration of them, to achieve simultaneous ORR and WOR pathways for efficient photocatalytic H₂O₂ generation [307,308]. Furthermore, employing advanced time-resolved or space-resolved characterization methods (e.g., time-resolved photoluminescence, transient absorption spectroscopy, transient electron paramagnetic resonance, in situ X-ray photoelectron spectroscopy, and Kelvin probe force microscopy) is crucial to observe the charge transfer behaviour [309]. Additionally, conducting in situ monitoring of reaction intermediates (e.g., in situ diffuse reflectance infrared Fourier transform spectroscopy) during the photocatalytic H₂O₂ production process is beneficial [30,50]. If necessary, theoretical calculations should be combined with experimental observations to better guide the fabrication of photocatalysts for efficient H₂O₂ generation.

  Secondly, there is a need to explore and develop 2D photocatalysts with high solar-to-H₂O₂ conversion efficiency. Despite the enormous variety of 2D photocatalysts reported to date, including metal oxides, metal chalcogenides, bismuth-based materials, g-C₃N₄, MOFs and COFs, many face challenges such as insufficient light absorption ability, rapid charge recombination, and kinetically sluggish WOR half-reaction (Table 2). As a result, their SCC efficiency generally remains lower than 0.5% [308]. One approach to address this challenge is to precisely engineer 2D photocatalysts with suitable band positions and bandgaps to ensure broad light absorption ranges, high efficiency in charge carrier separation/transfer, and strong redox capabilities [255,260]. In this context, the detection of the electronic band structures of 2D photocatalysts becomes imperative. Angle-resolved photoemission spectroscopy, an emerging energy- and momentum-resolved technique, directly visualizes electronic band structures and provides information on band gaps, band edge positions, Fermi surface, and spin polarization properties [310]. Additionally, it is crucial to rationally design and modify 2D photocatalysts to simultaneously accelerate both the oxygen reduction and water oxidation reactions [46]. This design strategy is grounded in a deep understanding of the mechanism underlying H₂O₂ production. Incorporating multiple strategies, including modification, doping, defect, and interfacial engineering, without limitation to a particular approach, is essential to simultaneously modulate the three key steps of light absorption, charge separation/transfer, and surface catalytic reactions in preparing photocatalysts for high-efficiency H₂O₂ photosynthesis [178,200].

  Table 2

  

  Table 2.  The main advantages and challenges of different 2D photocatalysts.

  DownLoad: CSV

  Thirdly, accurately evaluating and reporting the activity of photocatalysts for H₂O₂ production is crucial. Despite recent progress in developing efficient 2D photocatalysts for H₂O₂ synthesis, the absence of a standardized evaluation method impedes meaningful comparisons between different catalyst systems and even across laboratories, posing a significant obstacle to the advancement of 2D photocatalysts [311,312]. Currently, in the study of photocatalytic H₂O₂ synthesis, the main evaluation parameter is the reaction rate [253]. However, the H₂O₂ production rate is influenced by factors such as light source intensity, incident light wavelength range, reactor type, and reaction temperature. Therefore, subsequent studies need to provide sufficient and accurate experimental details, including the AQY and SCC efficiency.

  Fourthly, the stability of photocatalysts is paramount. Over extended periods of light exposure, photocatalysts often encounter issues such as photocorrosion, aggregation, or other stability concerns, especially given that the product, H₂O₂, is a potent oxidant [313]. Inorganic photocatalysts, such as metal oxides, may experience corrosion or metal leaching upon prolonged contact with accumulated H₂O₂ [17]. Additionally, radicals like O₂^•– generated through the indirect two-step e^– ORR pathway and ^•OH produced by the indirect two-step e^– WOR pathway may lead to the decomposition of polymer photocatalysts [314]. We believe that further exploration of photocatalysts with long-term stability will be beneficial for the practical applications of photocatalytic H₂O₂ production.

  Fifthly, the photocatalytic H₂O₂ production process is often accompanied by in situ H₂O₂ decomposition [315]. The generated H₂O₂ absorbs UV light and directly decomposes under the illumination of sunlight (2H₂O₂ → 2H₂O + O₂) [33]. Furthermore, H₂O₂ can undergo redox to generate ^•OH radical (H₂O₂ + e^– → ^•OH + OH^–) and O₂^•– radical (H₂O₂ + h⁺ → O₂^•– + 2H⁺) [306]. Hence, future research should focus on developing photocatalytic systems driven by visible light and avoiding the decomposition of in situ formed H₂O₂ [231,233]. Alternatively, combining with Fenton technology to convert the formed H₂O₂ into reactive oxygen species [316], such as ^•OH and O₂^•– radicals, could enable rapid sterilization [317], photodegradation of volatile organic compounds [318], and cancer therapeutic [319].

  The comprehensive summary and insightful discussion provided in this timely review on the photocatalytic production of H₂O₂ by 2D photocatalysts would contribute to advancing the future of efficient and sustainable H₂O₂ generation.

  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

  Liyong Ding: Writing – review & editing, Writing – original draft. Zhenhua Pan: Conceptualization, Writing – review & editing. Qian Wang: Writing – review & editing, Writing – original draft, Supervision.

  Acknowledgments

  This work was supported by the National Natural Science Foundation of China (No. 22106087), the Joint Funds of the Zhejiang Provincial Natural Science Foundation of China (No. LZY22B070001), the Foundation of China Scholarship Council (No. 202208330186), the JST Fusion Oriented Research for disruptive Science and Technology Program (No. JPMJFR213D), and JSPS Leading Initiative for Excellent Young Researchers program. In addition, Liyong Ding thanked Yu Chen for her understanding and support.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110125.

  
  
• 1. [1]

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

  2. [2]

     L. Pi, J. Cai, L. Xiong, et al., Chem. Eng. J. 389 (2020) 123420. doi: 10.1016/j.cej.2019.123420

  3. [3]

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

  4. [4]

     Research and Markets, Hydrogen Peroxide Global Market Report 2023, https://www.researchandmarkets.com/reports/5741554/hydrogen-peroxide-global-market-report, February 2023.

  5. [5]

     P. Landon, P. Collier, A. Papworth, et al., Chem. Commun. (2002) 2058–2059.

  6. [6]

     J. Lunsford, J. Catal. 216 (2003) 455–460. doi: 10.1016/S0021-9517(02)00070-2

  7. [7]

     Y. Sun, L. Han, P. Strasser, Chem. Soc. Rev. 49 (2020) 6605–6631. doi: 10.1039/d0cs00458h

  8. [8]

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

  9. [9]

     S. Qu, H. Wu, Y. Ng, Adv. Energy Mater. 13 (2023) 2301047. doi: 10.1002/aenm.202301047

  10. [10]

      N. Kaynan, B. Berke, O. Hazut, R. Yerushalmi, J. Mater. Chem. A 2 (2014) 13822–13826. doi: 10.1039/C4TA03004D

  11. [11]

      T. Liu, Z. Pan, J. Vequizo, et al., Nat. Commun. 13 (2022) 1034. doi: 10.1038/s41467-022-28686-x

  12. [12]

      Y. Zhang, C. Pan, G. Bian, et al., Nat. Energy 8 (2023) 361–371. doi: 10.1038/s41560-023-01218-7

  13. [13]

      Y. Ye, J. Pan, F. Xie, et al., Proc. Natl. Acad. Sci. U. S. A. 118 (2021) e2103964118. doi: 10.1073/pnas.2103964118

  14. [14]

      H. Hou, X. Zeng, X. Zhang, Angew. Chem. Int. Ed. 59 (2020) 17356–17376. doi: 10.1002/anie.201911609

  15. [15]

      X. Zhang, H. Su, P. Cui, et al., Nat. Commun. 14 (2023) 7115. doi: 10.1038/s41467-023-42887-y

  16. [16]

      H. Huang, Q. Zhang, R. Shi, et al., Appl. Catal. B: Environ. 317 (2022) 121731. doi: 10.1016/j.apcatb.2022.121731

  17. [17]

      S. Wu, X. Quan, ACS ES&T Eng. 2 (2022) 1068–1079. doi: 10.1021/acsestengg.1c00456

  18. [18]

      L. Wang, J. Zhang, Y. Zhang, et al., Small 18 (2022) 2104561. doi: 10.1002/smll.202104561

  19. [19]

      L. Zhang, J. Zhang, H. Yu, J. Yu, Adv. Mater. 34 (2022) e2107668. doi: 10.1002/adma.202107668

  20. [20]

      X. Xu, Y. Sui, W. Chen, et al., Appl. Catal. B: Environ. 341 (2024) 123271. doi: 10.1016/j.apcatb.2023.123271

  21. [21]

      C. Feng, Z. Wu, K. Huang, et al., Adv. Mater. 34 (2022) 2200180. doi: 10.1002/adma.202200180

  22. [22]

      L. Wang, H. Xu, Prog. Poly. Sci. 145 (2023) 101734. doi: 10.1016/j.progpolymsci.2023.101734

  23. [23]

      Y. Zhou, Z. Wang, L. Huang, et al., Adv. Energy Mater. 11 (2021) 2003159. doi: 10.1002/aenm.202003159

  24. [24]

      Y. Guo, X. Tong, N. Yang, Nano-Micro Lett. 15 (2023) 77. doi: 10.1007/s40820-023-01052-2

  25. [25]

      X. Zeng, Y. Liu, X. Hu, X. Zhang, Green Chem. 23 (2021) 1466–1494. doi: 10.1039/d0gc04236f

  26. [26]

      Z. Yong, T. Ma, Angew. Chem. Int. Ed. 62 (2023) e202308980. doi: 10.1002/anie.202308980

  27. [27]

      H. Cheng, J. Cheng, L. Wang, H. Xu, Chem. Mater. 34 (2022) 4259–4273. doi: 10.1021/acs.chemmater.2c00936

  28. [28]

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

  29. [29]

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

  30. [30]

      C. Qin, X. Wu, L. Tang, et al., Nat. Commun. 14 (2023) 5238. doi: 10.1038/s41467-023-40991-7

  31. [31]

      E. Baur, C. Neuweiler, Helv. Chim. Acta 10 (1927) 901–907. doi: 10.1002/hlca.192701001113

  32. [32]

      T. Gill, X. Zheng, Chem. Mater. 32 (2020) 6285–6294. doi: 10.1021/acs.chemmater.0c02010

  33. [33]

      Y. Nosaka, A. Nosaka, Chem. Rev. 117 (2017) 11302–11336. doi: 10.1021/acs.chemrev.7b00161

  34. [34]

      G. Eisenberg, Ind. Eng. Chem. Anal. Ed. 15 (1943) 327–328. doi: 10.1021/i560117a011

  35. [35]

      V. Diesen, M. Jonsson, J. Phys. Chem. C 118 (2014) 10083–10087. doi: 10.1021/jp500315u

  36. [36]

      K. Sahel, L. Elsellami, I. Mirali, et al., Appl. Catal. B: Environ. 188 (2016) 106–112. doi: 10.1016/j.apcatb.2015.12.044

  37. [37]

      Y. Zhao, P. Zhang, Z. Yang, et al., Nat. Commun. 12 (2021) 3701. doi: 10.1038/s41467-021-24048-1

  38. [38]

      C. Marquette, L. Blum, Anal. Bioanal. Chem. 385 (2006) 546–554. doi: 10.1007/s00216-006-0439-9

  39. [39]

      T. Hirakawa, Y. Nosaka, J. Phys. Chem. C 112 (2008) 15818–15823. doi: 10.1021/jp8055015

  40. [40]

      C. Chu, Q. Zhu, Z. Pan, et al., Proc. Natl. Acad. Sci. U. S. A. 117 (2020) 6376–6382. doi: 10.1073/pnas.1913403117

  41. [41]

      A. Keston, R. Brandt, Anal. Biochem. 11 (1965) 1–5. doi: 10.1016/0003-2697(65)90034-5

  42. [42]

      C. Bernardini, G. Cappelletti, M. Dozzi, E. Selli, J. Photochem. Photobiol. A 211 (2010) 185–192. doi: 10.1016/j.jphotochem.2010.03.006

  43. [43]

      C. Huckaba, F. Keyes, J. Am. Chem. Soc. 70 (1948) 1640–1644. doi: 10.1021/ja01184a098

  44. [44]

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

  45. [45]

      X. Shi, S. Siahrostami, G. Li, et al., Nat. Commun. 8 (2017) 701. doi: 10.1038/s41467-017-00585-6

  46. [46]

      H. Cheng, H. Lv, J. Cheng, et al., Adv. Mater. 34 (2022) 2107480. doi: 10.1002/adma.202107480

  47. [47]

      G. Li, P. Fu, Q. Yue, et al., Chem. Catal. 2 (2022) 1734–1747. doi: 10.1016/j.checat.2022.05.002

  48. [48]

      D. Chen, W. Chen, Y. Wu, et al., Angew. Chem. Int. Ed. 62 (2023) e202217479. doi: 10.1002/anie.202217479

  49. [49]

      X. Shi, Y. Zhang, S. Siahrostami, X. Zheng, Adv. Energy Mater. 8 (2018) 1801158. doi: 10.1002/aenm.201801158

  50. [50]

      F. Liu, P. Zhou, Y. Hou, et al., Nat. Commun. 14 (2023) 4344. doi: 10.1038/s41467-023-40007-4

  51. [51]

      Y. Mou, X. Wu, C. Qin, et al., Angew. Chem. Int. Ed. 62 (2023) e202309480. doi: 10.1002/anie.202309480

  52. [52]

      J. Chang, Q. Li, J. Shi, et al., Angew. Chem. Int. Ed. 62 (2023) e202218868. doi: 10.1002/anie.202218868

  53. [53]

      G. Moon, W. Kim, A. Bokare, et al., Energy Environ. Sci. 7 (2014) 4023–4028. doi: 10.1039/C4EE02757D

  54. [54]

      R. Nogueira, M. Oliveira, W. Paterlini, Talanta 66 (2005) 86–91. doi: 10.1016/j.talanta.2004.10.001

  55. [55]

      F. Kuttassery, Y. Ohsaki, A. Thomas, et al., Angew. Chem. Int. Ed. 62 (2023) e202308956. doi: 10.1002/anie.202308956

  56. [56]

      V. Sanz, S. Marcos, J. Castillo, J. Galbán, J. Am. Chem. Soc. 127 (2005) 1038–1048. doi: 10.1021/ja046830k

  57. [57]

      M. Zhou, Z. Diwu, N. Panchuk-Voloshina, R. Haugland, Anal. Biochem. 253 (1997) 162–168. doi: 10.1006/abio.1997.2391

  58. [58]

      Z. Wu, M. Liu, Z. Liu, Y. Tian, J. Am. Chem. Soc. 142 (2020) 7532–7541. doi: 10.1021/jacs.0c00771

  59. [59]

      S. Du, J. Lian, F. Zhang, Trans. Tianjin Univ. 28 (2021) 33–52.

  60. [60]

      J. Harbour, J. Tromp, M. Hair, Can. J. Chem. 63 (1985) 204–208. doi: 10.1139/v85-032

  61. [61]

      V. Maurino, C. Minero, G. Mariella, E. Pelizzetti, Chem. Commun. (2005) 2627–2629. doi: 10.1039/b418789j

  62. [62]

      M. Teranishi, S. Naya, H. Tada, J. Am. Chem. Soc. 132 (2010) 7850–7851. doi: 10.1021/ja102651g

  63. [63]

      E. Goliaei, N. Seriani, J. Phys. Chem. C 123 (2019) 2855–2863. doi: 10.1021/acs.jpcc.8b09300

  64. [64]

      Y. Liu, X. Zeng, X. Hu, et al., Catal. Today 335 (2019) 243–251. doi: 10.1016/j.cattod.2018.11.053

  65. [65]

      Y. Liu, X. Zeng, C. Easton, et al., Nanoscale 12 (2020) 8775–8784. doi: 10.1039/d0nr01611j

  66. [66]

      B. Liu, C. Bie, Y. Zhang, et al., Langmuir 37 (2021) 14114–14124. doi: 10.1021/acs.langmuir.1c02360

  67. [67]

      C. Zhu, M. Zhu, Y. Sun, et al., ACS Appl. Energy Mater. 2 (2019) 8737–8746. doi: 10.1021/acsaem.9b01712

  68. [68]

      C. Fu, L. Liu, Z. Li, et al., J. Phys. Chem. Lett. 14 (2023) 7690–7696. doi: 10.1021/acs.jpclett.3c01865

  69. [69]

      Y. Wang, Y. Wang, J. Zhao, et al., Appl. Catal. B: Environ. 284 (2021) 119691. doi: 10.1016/j.apcatb.2020.119691

  70. [70]

      W. Xie, Z. Huang, R. Wang, et al., J. Mater. Sci. 55 (2020) 11829–11840. doi: 10.1007/s10853-020-04837-7

  71. [71]

      D. Tsukamoto, A. Shiro, Y. Shiraishi, et al., ACS Catal. 2 (2012) 599–603. doi: 10.1021/cs2006873

  72. [72]

      M. Teranishi, R. Hoshino, S. Naya, H. Tada, Angew. Chem. Int. Ed. 55 (2016) 12773–12777. doi: 10.1002/anie.201606734

  73. [73]

      S. Cao, T. Chan, Y. Lu, et al., Nano Energy 67 (2020) 104287. doi: 10.1016/j.nanoen.2019.104287

  74. [74]

      X. Zeng, Z. Wang, N. Meng, et al., Appl. Catal. B: Environ. 202 (2017) 33–41. doi: 10.1016/j.apcatb.2016.09.014

  75. [75]

      Y. Yang, Q. Wang, X. Zhang, et al., J. Mater. Chem. A 11 (2023) 1991–2001. doi: 10.1039/d2ta08145h

  76. [76]

      J. Wang, J. Wang, S. Zuo, et al., Chin. Chem. Lett. 34 (2023) 108157. doi: 10.1016/j.cclet.2023.108157

  77. [77]

      B. He, Z. Wang, P. Xiao, et al., Adv. Mater. 34 (2022) 2203225. doi: 10.1002/adma.202203225

  78. [78]

      X. Dang, S. Wu, H. Zhao, ACS Sustain. Chem. Eng. 10 (2022) 4161–4172. doi: 10.1021/acssuschemeng.1c07985

  79. [79]

      G. Han, F. Xu, B. Cheng, et al., Acta Phys. Chim. Sin. 38 (2022) 2112037. doi: 10.3866/pku.whxb202112037

  80. [80]

      X. Meng, P. Zong, L. Wang, et al., Catal. Commun. 134 (2020) 105860. doi: 10.1016/j.catcom.2019.105860

  81. [81]

      Z. Jiang, B. Cheng, Y. Zhang, et al., J. Mater. Sci. Technol. 124 (2022) 193–201. doi: 10.1016/j.jmst.2022.01.029

  82. [82]

      B. Mishra, L. Biswal, S. Das, et al., Langmuir 39 (2023) 957–971. doi: 10.1021/acs.langmuir.2c02315

  83. [83]

      X. Chen, W. Zhang, L. Zhang, et al., J. Mater. Chem. A 8 (2020) 18816–18825. doi: 10.1039/d0ta05753c

  84. [84]

      L. Feng, B. Li, Y. Xiao, et al., Catal. Commun. 155 (2021) 106315. doi: 10.1016/j.catcom.2021.106315

  85. [85]

      X. Yan, G. Yu, C. Xing, et al., Catal. Sci. Technol. 13 (2023) 3094–3105. doi: 10.1039/d3cy00235g

  86. [86]

      X. Chen, W. Zhang, L. Zhang, et al., ACS Appl. Mater. Interfaces 13 (2021) 25868–25878. doi: 10.1021/acsami.1c02953

  87. [87]

      S. Mansingh, K. Das, N. Priyadarshini, et al., Energy Fuels 37 (2023) 9873–9894. doi: 10.1021/acs.energyfuels.3c00717

  88. [88]

      J. Low, J. Yu, M. Jaroniec, et al., Adv. Mater. 29 (2017) 1601694. doi: 10.1002/adma.201601694

  89. [89]

      C. Pan, Z. Mao, X. Yuan, et al., Adv. Sci. 9 (2022) 2105747. doi: 10.1002/advs.202105747

  90. [90]

      X. Yue, J. Fan, Q. Xiang, Adv. Funct. Mater. 32 (2021) 2110258.

  91. [91]

      X. Mao, X. Lang, Z. Wang, et al., J. Phys. Chem. Lett. 4 (2013) 3839–3844. doi: 10.1021/jz402053p

  92. [92]

      W. Wang, Q. Song, Q. Luo, et al., Nat. Commun. 14 (2023) 2493. doi: 10.1002/ps.7431

  93. [93]

      F. Haque, T. Daeneke, K. Kalantar-zadeh, J. Ou, Nano-Micro Lett. 10 (2018) 23. doi: 10.1007/s40820-017-0176-y

  94. [94]

      H. Song, L. Wei, C. Chen, et al., J. Catal. 376 (2019) 198–208. doi: 10.1016/j.jcat.2019.06.015

  95. [95]

      G. Jiang, X. You, B. An, et al., Appl. Surf. Sci. 618 (2023) 156656. doi: 10.1016/j.apsusc.2023.156656

  96. [96]

      Y. Yang, H. Guo, D. Huang, et al., Chem. Eng. J. 479 (2024) 147863. doi: 10.1016/j.cej.2023.147863

  97. [97]

      A. Tikoo, N. Lohia, S. Kondeti, P. Meduri, J. Mater. Chem. A 11 (2023) 14887–14899. doi: 10.1039/d3ta02875e

  98. [98]

      W. Fang, L. Wang, X. Meng, C. Li, J. Alloy. Compd. 947 (2023) 169606. doi: 10.1016/j.jallcom.2023.169606

  99. [99]

      Y. Li, Y. Zhao, H. Nie, et al., J. Mater. Chem. A 9 (2021) 515–522. doi: 10.1039/d0ta10231h

  100. [100]

       R. Li, D. Zhang, Y. Shi, et al., J. Catal. 416 (2022) 322–331. doi: 10.1016/j.jcat.2022.11.016

  101. [101]

       Y. Yang, H. Yu, M. Wu, et al., Appl. Catal. B: Environ. 325 (2023) 122307. doi: 10.1016/j.apcatb.2022.122307

  102. [102]

       Y. Li, R. Jia, H. Lin, et al., Adv. Funct. Mater. 31 (2020) 2008420.

  103. [103]

       S. Thakur, T. Kshetri, N. Kim, J. Lee, J. Catal. 345 (2017) 78–86. doi: 10.5958/2455-7129.2017.00011.5

  104. [104]

       Z. Tian, C. Han, Y. Zhao, et al., Nat. Commun. 12 (2021) 2039. doi: 10.1038/s41467-021-22394-8

  105. [105]

       B. Jiang, D. Chen, N. Li, et al., Ind. Eng. Chem. Res. 62 (2023) 12974–12984. doi: 10.1021/acs.iecr.3c01800

  106. [106]

       A. Tikoo, A. Koushik, P. Meduri, ACS Appl. Eng. Mater. 1 (2023) 1397–1407. doi: 10.1021/acsaenm.3c00077

  107. [107]

       H. Chen, Y. Xing, S. Liu, et al., Chem. Eng. J. 462 (2023) 142038. doi: 10.1016/j.cej.2023.142038

  108. [108]

       Y. Yang, B. Cheng, J. Yu, et al., Nano Res. 16 (2021) 4506–4514. doi: 10.3390/app11104506

  109. [109]

       Z. Wang, B. Mi, Environ. Sci. Technol. 51 (2017) 8229–8244. doi: 10.1021/acs.est.7b01466

  110. [110]

       K. Mak, C. Lee, J. Hone, et al., Phys. Rev. Lett. 105 (2010) 136805. doi: 10.1103/PhysRevLett.105.136805

  111. [111]

       M. Bernardi, M. Palummo, J. Grossman, Nano Lett. 13 (2013) 3664–3670. doi: 10.1021/nl401544y

  112. [112]

       U. Maitra, U. Gupta, M. De, et al., Angew. Chem. Int. Ed. 52 (2013) 13057–13061. doi: 10.1002/anie.201306918

  113. [113]

       F. Bussolotti, J. Yang, H. Kawai, et al., ACS Nano 15 (2021) 2686–2697. doi: 10.1021/acsnano.0c07982

  114. [114]

       L. Ding, R. Wei, H. Chen, et al., Appl. Catal. B: Environ. 172 (2015) 91–99.

  115. [115]

       L. Ding, M. Li, Y. Zhao, et al., Appl. Catal. B: Environ. 266 (2020) 118634. doi: 10.1016/j.apcatb.2020.118634

  116. [116]

       X. Wu, H. Tan, C. Zhang, et al., Prog. Mater. Sci. 133 (2023) 101047. doi: 10.1016/j.pmatsci.2022.101047

  117. [117]

       P. Chen, H. Liu, W. Cui, et al., EcoMat 2 (2020) e12047. doi: 10.1002/eom2.12047

  118. [118]

       H. Hirakawa, S. Shiota, Y. Shiraishi, et al., ACS Catal. 6 (2016) 4976–4982. doi: 10.1021/acscatal.6b01187

  119. [119]

       K. Wang, M. Wang, J. Yu, et al., ACS Appl. Nano Mater. 4 (2021) 13158–13166. doi: 10.1021/acsanm.1c02688

  120. [120]

       H. Shi, Y. Li, K. Wang, et al., Chem. Eng. J. 443 (2022) 136429. doi: 10.1016/j.cej.2022.136429

  121. [121]

       H. Shi, Y. Li, X. Wang, et al., Appl. Catal. B: Environ. 297 (2021) 120414. doi: 10.1016/j.apcatb.2021.120414

  122. [122]

       K. Wang, Y. Li, H. Shi, et al., Adv. Sustain. Syst. 6 (2022) 2200144. doi: 10.1002/adsu.202200144

  123. [123]

       K. Fuku, R. Takioka, K. Iwamura, et al., Appl. Catal. B: Environ. 272 (2020) 119003. doi: 10.1016/j.apcatb.2020.119003

  124. [124]

       D. Dai, X. Bao, Q. Zhang, et al., Chem. Eur. J. 29 (2023) e202203765. doi: 10.1002/chem.202203765

  125. [125]

       M. Sun, X. Wang, H. Pan, et al., J. Colloid Interf. Sci. 629 (2023) 215–224. doi: 10.1016/j.jcis.2022.08.142

  126. [126]

       L. Wei, S. Liu, H.P. Wang, ACS Appl. Nano Mater. 5 (2022) 15378–15388. doi: 10.1021/acsanm.2c03420

  127. [127]

       Y. Zeng, H. Fu, L. Liu, et al., ChemistrySelect 8 (2023) e202204286. doi: 10.1002/slct.202204286

  128. [128]

       Y. Xu, H. Fu, L. Zhao, et al., New J. Chem. 45 (2021) 3335–3342. doi: 10.1039/d0nj05506a

  129. [129]

       Y. Gong, Y. Yuan, B. Wang, et al., Mater. Lett. 351 (2023) 135122. doi: 10.1016/j.matlet.2023.135122

  130. [130]

       W. Wang, X. Li, F. Deng, et al., Chin. Chem. Lett. 33 (2022) 5200–5207. doi: 10.1016/j.cclet.2022.01.058

  131. [131]

       X. Li, J. Xiong, X. Gao, et al., J. Hazard. Mater. 387 (2020) 121690. doi: 10.1016/j.jhazmat.2019.121690

  132. [132]

       S. Shen, X. Li, Y. Zhou, et al., J. Mater. Sci. Technol. 155 (2023) 148–159. doi: 10.1016/j.jmst.2023.03.006

  133. [133]

       H. Zhang, J. Liu, Y. Zhang, et al., J. Mater. Sci. Technol. 166 (2023) 241–249. doi: 10.1016/j.jmst.2023.05.030

  134. [134]

       X. Zhao, Y. You, S. Huang, et al., Appl. Catal. B: Environ. 278 (2020) 119251. doi: 10.1016/j.apcatb.2020.119251

  135. [135]

       R. An, Y. Zhao, H. Bai, et al., J. Solid State Chem. 306 (2022) 122722. doi: 10.1016/j.jssc.2021.122722

  136. [136]

       J. Wang, Y. Gong, M. Gao, et al., ACS Appl. Nano Mater. 7 (2023) 1067–1077.

  137. [137]

       A. Kudo, K. Omori, H. Kato, J. Am. Chem. Soc. 121 (1999) 11459–11467. doi: 10.1021/ja992541y

  138. [138]

       Z. Xing, J. Hu, M. Ma, et al., J. Am. Chem. Soc. 141 (2019) 19715–19727. doi: 10.1021/jacs.9b08651

  139. [139]

       J. Sheng, X. Li, Y. Xu, ACS Catal. 4 (2014) 732–737. doi: 10.1021/cs400927w

  140. [140]

       Y. Ding, S. Maitra, S. Halder, et al., Matter 5 (2022) 2119–2167. doi: 10.1016/j.matt.2022.05.011

  141. [141]

       J. Li, H. Li, G. Zhan, L. Zhang, Acc. Chem. Res. 50 (2017) 112–121. doi: 10.1021/acs.accounts.6b00523

  142. [142]

       H. Li, J. Li, Z. Ai, et al., Angew. Chem. Int. Ed. 57 (2018) 122–138. doi: 10.1002/anie.201705628

  143. [143]

       L. Ding, H. Chen, Q. Wang, et al., Chem. Commun. 52 (2016) 994–997. doi: 10.1039/C5CC08146G

  144. [144]

       X. Xiong, L. Ding, Q. Wang, et al., Appl. Catal. B: Environ. 188 (2016) 283–291. doi: 10.1016/j.apcatb.2016.02.018

  145. [145]

       J. Li, L. Cai, J. Shang, et al., Adv. Mater. 28 (2016) 4059–4064. doi: 10.1002/adma.201600301

  146. [146]

       Z. Zhao, Y. Zhou, F. Wang, et al., ACS Appl. Mater. Interfaces 7 (2015) 730–737. doi: 10.1021/am507089x

  147. [147]

       Z. Ni, Y. Sun, Y. Zhang, F. Dong, Appl. Surf. Sci. 365 (2016) 314–335. doi: 10.1016/j.apsusc.2015.12.231

  148. [148]

       H. Ou, P. Yang, L. Lin, et al., Angew. Chem. Int. Ed. 56 (2017) 10905–10910. doi: 10.1002/anie.201705926

  149. [149]

       Z. Wei, M. Liu, Z. Zhang, et al., Energy Environ. Sci. 11 (2018) 2581–2589. doi: 10.1039/c8ee01316k

  150. [150]

       B. Feng, Y. Liu, K. Wan, et al., Angew. Chem. Int. Ed. 63 (2024) e202401884. doi: 10.1002/anie.202401884

  151. [151]

       Y. Shiraishi, S. Kanazawa, Y. Sugano, et al., ACS Catal. 4 (2014) 774–780. doi: 10.1021/cs401208c

  152. [152]

       J. Fu, J. Yu, C. Jiang, B. Cheng, Adv. Energy Mater. 8 (2017) 1701503.

  153. [153]

       R. Du, K. Xiao, B. Li, et al., Chem. Eng. J. 441 (2022) 135999. doi: 10.1016/j.cej.2022.135999

  154. [154]

       Y. Shiraishi, S. Kanazawa, Y. Kofuji, et al., Angew. Chem. Int. Ed. 53 (2014) 13454–13459. doi: 10.1002/anie.201407938

  155. [155]

       Y. Kofuji, S. Ohkita, Y. Shiraishi, et al., ACS Catal. 6 (2016) 7021–7029. doi: 10.1021/acscatal.6b02367

  156. [156]

       Y. Kofuji, S. Ohkita, Y. Shiraishi, et al., ACS Sustain. Chem. Eng. 5 (2017) 6478–6485. doi: 10.1021/acssuschemeng.7b00575

  157. [157]

       Y. Kofuji, Y. Isobe, Y. Shiraishi, et al., J. Am. Chem. Soc. 138 (2016) 10019–10025. doi: 10.1021/jacs.6b05806

  158. [158]

       Y. Kofuji, Y. Isobe, Y. Shiraishi, et al., ChemCatChem 10 (2018) 2070–2077. doi: 10.1002/cctc.201701683

  159. [159]

       X. Zeng, Y. Liu, Y. Kang, et al., ACS Catal. 10 (2020) 3697–3706. doi: 10.1021/acscatal.9b05247

  160. [160]

       S. Yao, T. Tang, Y. Shen, et al., Sci. China Mater. 66 (2022) 672–678.

  161. [161]

       H. Che, X. Gao, J. Chen, et al., Angew. Chem. Int. Ed. 60 (2021) 25546–25550. doi: 10.1002/anie.202111769

  162. [162]

       Y. Peng, L. Wang, Y. Liu, et al., Eur. J. Inorg. Chem. 2017 (2017) 4797–4802. doi: 10.1002/ejic.201700930

  163. [163]

       P. Ren, T. Zhang, N. Jain, et al., J. Am. Chem. Soc. 145 (2023) 16584–16596. doi: 10.1021/jacs.3c03785

  164. [164]

       H. Liang, A. Wang, R. Cheng, et al., Small 19 (2023) 2303813. doi: 10.1002/smll.202303813

  165. [165]

       Y. Ye, C. Wen, J. Pan, et al., Appl. Catal. B: Environ. 285 (2021) 119726. doi: 10.1016/j.apcatb.2020.119726

  166. [166]

       S. Kim, G. Moon, H. Kim, et al., J. Catal. 357 (2018) 51–58. doi: 10.1016/j.jcat.2017.10.002

  167. [167]

       P. Sha, L. Huang, J. Zhao, et al., ACS Catal. 13 (2023) 10474–10486. doi: 10.1021/acscatal.3c02118

  168. [168]

       J. Cai, J. Huang, S. Wang, et al., Adv. Mater. 31 (2019) 1806314. doi: 10.1002/adma.201806314

  169. [169]

       G. Moon, M. Fujitsuka, S. Kim, et al., ACS Catal. 7 (2017) 2886–2895. doi: 10.1021/acscatal.6b03334

  170. [170]

       Y. Zhao, Y. Liu, J. Cao, et al., Appl. Catal. B: Environ. 278 (2020) 119289. doi: 10.1016/j.apcatb.2020.119289

  171. [171]

       Y. Yang, G. Zeng, D. Huang, et al., Appl. Catal. B: Environ. 272 (2020) 118970. doi: 10.1016/j.apcatb.2020.118970

  172. [172]

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

  173. [173]

       L. Zhou, J. Feng, B. Qiu, et al., Appl. Catal. B: Environ. 267 (2020) 118396. doi: 10.1016/j.apcatb.2019.118396

  174. [174]

       P. Zhang, Y. Tong, Y. Liu, et al., Angew. Chem. Int. Ed. 59 (2020) 16209–16217. doi: 10.1002/anie.202006747

  175. [175]

       L. Chen, C. Chen, Z. Yang, et al., Adv. Funct. Mater. 31 (2021) 2105731. doi: 10.1002/adfm.202105731

  176. [176]

       W. Liu, P. Wang, J. Chen, et al., Adv. Funct. Mater. 32 (2022) 2205119. doi: 10.1002/adfm.202205119

  177. [177]

       D. Li, C. Wen, J. Huang, et al., Appl. Catal. B: Environ. 307 (2022) 121099. doi: 10.1016/j.apcatb.2022.121099

  178. [178]

       S. Wu, H. Yu, S. Chen, X. Quan, ACS Catal. 10 (2020) 14380–14389. doi: 10.1021/acscatal.0c03359

  179. [179]

       X. Zhang, P. Ma, C. Wang, et al., Energy Environ. Sci. 15 (2022) 830–842. doi: 10.1039/d1ee02369a

  180. [180]

       Z. Zhu, H. Pan, M. Murugananthan, et al., Appl. Catal. B: Environ. 232 (2018) 19–25. doi: 10.1504/ijipt.2018.10012654

  181. [181]

       S. Li, G. Dong, R. Hailili, et al., Appl. Catal. B: Environ. 190 (2016) 26–35. doi: 10.1016/j.apcatb.2016.03.004

  182. [182]

       Y. Shiraishi, Y. Kofuji, H. Sakamoto, et al., ACS Catal. 5 (2015) 3058–3066. doi: 10.1021/acscatal.5b00408

  183. [183]

       L. Shi, L. Yang, W. Zhou, et al., Small 14 (2018) 1703142. doi: 10.1002/smll.201703142

  184. [184]

       X. Li, J. Zhang, F. Zhou, et al., Chin. J. Catal. 39 (2018) 1090–1098. doi: 10.1016/S1872-2067(18)63046-3

  185. [185]

       Q. He, J. Ding, H. Tsai, et al., J. Colloid. Interf. Sci. 651 (2023) 18–26. doi: 10.1016/j.jcis.2023.07.168

  186. [186]

       Y. Xie, Y. Li, Z. Huang, et al., Appl. Catal. B: Environ. 265 (2020) 118581. doi: 10.1016/j.apcatb.2019.118581

  187. [187]

       J. Chen, W. Gao, Y. Lu, et al., ACS Appl. Nano Mater. 6 (2023) 3927–3935. doi: 10.1021/acsanm.3c00092

  188. [188]

       Y. Wang, D. Meng, X. Zhao, Appl. Catal. B: Environ. 273 (2020) 119064. doi: 10.1016/j.apcatb.2020.119064

  189. [189]

       H. Fattahimoghaddam, T. Mahvelati-Shamsabadi, B. Lee, ACS Sustain. Chem. Eng. 9 (2021) 4520–4530. doi: 10.1021/acssuschemeng.0c08884

  190. [190]

       Y. Zheng, Y. Luo, Q. Ruan, et al., Appl. Catal. B: Environ. 311 (2022) 121372. doi: 10.1016/j.apcatb.2022.121372

  191. [191]

       J. Yang, Z. Ji, S. Zhang, ACS Appl. Energy Mater. 6 (2023) 3401–3412. doi: 10.1021/acsaem.2c04124

  192. [192]

       J. Lei, B. Chen, W. Lv, et al., ACS Sustain. Chem. Eng. 7 (2019) 16467–16473. doi: 10.1021/acssuschemeng.9b03678

  193. [193]

       H. Xie, Y. Zheng, X. Guo, et al., ACS Sustain. Chem. Eng. 9 (2021) 6788–6798. doi: 10.1021/acssuschemeng.1c01012

  194. [194]

       Q. Li, Y. Jiao, Y. Tang, et al., J. Am. Chem. Soc. 145 (2023) 20837–20848. doi: 10.1021/jacs.3c05234

  195. [195]

       Y. Zhu, X. Liu, H. Liu, et al., SusMat 2 (2022) 617–629. doi: 10.1002/sus2.83

  196. [196]

       R. Li, Q. Deng, A. Chen, et al., ChemSusChem 16 (2023) e202300763. doi: 10.1002/cssc.202300763

  197. [197]

       L. Yang, G. Dong, D. Jacobs, et al., J. Catal. 352 (2017) 274–281. doi: 10.1016/j.jcat.2017.05.010

  198. [198]

       X. Zhang, J. Yu, W. Macyk, et al., Adv. Sustain. Syst. 7 (2022) 2200113.

  199. [199]

       X. Li, J. Liu, J. Huang, et al., Acta Phys. Chim. Sin. 37 (2020) 2010030. doi: 10.3866/pku.whxb202010030

  200. [200]

       B. Li, Z. Guo, Y. Feng, et al., ACS Appl. Mater. Interfaces 14 (2022) 43328–43338. doi: 10.1021/acsami.2c12038

  201. [201]

       X. Wang, Z. Han, L. Yu, et al., ACS Sustain. Chem. Eng. 6 (2018) 14542–14553. doi: 10.1021/acssuschemeng.8b03171

  202. [202]

       W. Zhang, S. Zhao, Y. Xing, et al., Chem. Eng. J. 442 (2022) 136151. doi: 10.1016/j.cej.2022.136151

  203. [203]

       K. Das, S. Mansingh, R. Mohanty, et al., J. Phys. Chem. C 127 (2022) 22–40.

  204. [204]

       P. Zhou, M. Luo, S. Guo, Nat. Rev. Chem. 6 (2022) 823–838. doi: 10.1038/s41570-022-00434-1

  205. [205]

       P. Zhou, H. Chen, Y. Chao, et al., Nat. Commun. 12 (2021) 4412. doi: 10.1038/s41467-021-24702-8

  206. [206]

       Z. Teng, W. Cai, W. Sim, et al., Appl. Catal. B: Environ. 282 (2021) 119589. doi: 10.1016/j.apcatb.2020.119589

  207. [207]

       G. Chakraborty, I. Park, R. Medishetty, J. Vittal, Chem. Rev. 121 (2021) 3751–3891. doi: 10.1021/acs.chemrev.0c01049

  208. [208]

       J. Choi, B. Check, X. Fang, et al., J. Am. Chem. Soc. 146 (2024) 11319–11327.

  209. [209]

       M. Alvaro, E. Carbonell, B. Ferrer, et al., Chem. Eur. J. 13 (2007) 5106–5112. doi: 10.1002/chem.200601003

  210. [210]

       S. Rojas, P. Horcajada, Chem. Rev. 120 (2020) 8378–8415. doi: 10.1021/acs.chemrev.9b00797

  211. [211]

       S. Navalon, A. Dhakshinamoorthy, M. Alvaro, et al., Chem. Rev. 123 (2023) 445–490. doi: 10.1021/acs.chemrev.2c00460

  212. [212]

       K. Sun, Y. Qian, H. Jiang, Angew. Chem. Int. Ed. 62 (2023) e202217565. doi: 10.1002/anie.202217565

  213. [213]

       W. Huang, M. Chai, R. Lin, et al., Ind. Eng. Chem. Res. 62 (2023) 14130–14143. doi: 10.1021/acs.iecr.3c02177

  214. [214]

       Y. Fu, D. Sun, Y. Chen, et al., Angew. Chem. Int. Ed. 51 (2012) 3364–3367. doi: 10.1002/anie.201108357

  215. [215]

       M. Wen, K. Mori, Y. Kuwahara, H. Yamashita, ACS Energy Lett. 2 (2016) 1–7.

  216. [216]

       Y. Horiuchi, T. Toyao, M. Saito, et al., J. Phys. Chem. C 116 (2012) 20848–20853. doi: 10.1021/jp3046005

  217. [217]

       C. Hendon, D. Tiana, M. Fontecave, et al., J. Am. Chem. Soc. 135 (2013) 10942–10945. doi: 10.1021/ja405350u

  218. [218]

       T. Wang, M. Chen, Y. Lai, et al., ACS Sustain. Chem. Eng. 11 (2023) 6465–6473. doi: 10.1021/acssuschemeng.3c00742

  219. [219]

       Q. Lan, S. Jin, B. Yang, et al., Trans. Tianjin Univ. 28 (2022) 214–225. doi: 10.1007/s12209-022-00324-z

  220. [220]

       Y. Isaka, Y. Kondo, Y. Kawase, et al., Chem. Commun. 54 (2018) 9270–9273. doi: 10.1039/c8cc02679c

  221. [221]

       K. Yue, X. Zhang, S. Jiang, et al., J. Mol. Liq. 335 (2021) 116108. doi: 10.1016/j.molliq.2021.116108

  222. [222]

       X. Chen, Y. Kondo, Y. Kuwahara, et al., Phys. Chem. Chem. Phys. 22 (2020) 14404–14414. doi: 10.1039/d0cp01759k

  223. [223]

       Y. Kondo, Y. Kuwahara, K. Mori, H. Yamashita, Chem 8 (2022) 2924–2938. doi: 10.1016/j.chempr.2022.10.007

  224. [224]

       Y. Isaka, Y. Kawase, Y. Kuwahara, et al., Angew. Chem. Int. Ed. 58 (2019) 5402–5406. doi: 10.1002/anie.201901961

  225. [225]

       Y. Kawase, Y. Isaka, Y. Kuwahara, et al., Chem. Commun. 55 (2019) 6743–6746. doi: 10.1039/c9cc02380a

  226. [226]

       L. Yuan, Y. Zou, L. Zhao, et al., Appl. Catal. B: Environ. 318 (2022) 121859. doi: 10.1016/j.apcatb.2022.121859

  227. [227]

       Y. Li, Y. Liu, Z. Wang, et al., Chin. J. Catal. 45 (2023) 132–140. doi: 10.3390/cryst13010132

  228. [228]

       X. Chen, Y. Kuwahara, K. Mori, et al., J. Mater. Chem. A 9 (2021) 2815–2821. doi: 10.1039/d0ta10944d

  229. [229]

       X. Chen, Y. Kondo, S. Li, et al., J. Mater. Chem. A 9 (2021) 26371–26380. doi: 10.1039/d1ta08036a

  230. [230]

       Y. Hao, L. Chen, J. Li, et al., Nat. Commun. 12 (2021) 2682. doi: 10.1038/s41467-021-22991-7

  231. [231]

       Y. Li, Y. Guo, D. Luan, et al., Angew. Chem. Int. Ed. 62 (2023) e202310847. doi: 10.1002/anie.202310847

  232. [232]

       X. Chen, Y. Kuwahara, K. Mori, et al., J. Mater. Chem. A 8 (2020) 1904–1910. doi: 10.1039/C9TA11120D

  233. [233]

       X. Chen, Y. Kuwahara, K. Mori, et al., ACS Appl. Energy Mater. 4 (2021) 4823–4830. doi: 10.1021/acsaem.1c00371

  234. [234]

       Y. Li, F. Ma, L. Zheng, et al., Mater. Horiz. 8 (2021) 2842–2850. doi: 10.1039/d1mh00869b

  235. [235]

       B. Yan, S. Hu, C. Bu, et al., ChemNanoMat 9 (2023) e202300079. doi: 10.1002/cnma.202300079

  236. [236]

       Y. Kondo, Y. Kuwahara, K. Mori, H. Yamashita, J. Phys. Chem. C 125 (2021) 27909–27918. doi: 10.1021/acs.jpcc.1c07735

  237. [237]

       Y. Kondo, K. Honda, Y. Kuwahara, et al., ACS Catal. 12 (2022) 14825–14835. doi: 10.1021/acscatal.2c04940

  238. [238]

       C. Liu, T. Bao, L. Yuan, et al., Adv. Funct. Mater. 32 (2022) 2111404. doi: 10.1002/adfm.202111404

  239. [239]

       R. Bariki, S. Pradhan, S. Panda, et al., Langmuir 39 (2023) 7707–7722. doi: 10.1021/acs.langmuir.3c00519

  240. [240]

       Y. Wu, X. Li, Q. Yang, et al., Chem. Eng. J. 390 (2020) 124519. doi: 10.1016/j.cej.2020.124519

  241. [241]

       A. Côté, A. Benin, N. Ockwig, et al., Science 310 (2005) 1166–1170. doi: 10.1126/science.1120411

  242. [242]

       S. Wang, Z. Xie, D. Zhu, et al., Nat. Commun. 14 (2023) 6891. doi: 10.1038/s41467-023-42720-6

  243. [243]

       P. Zhou, Y. Li, T. Zeng, et al., Angew. Chem. Int. Ed. 63 (2024) e202402911. doi: 10.1002/anie.202402911

  244. [244]

       S. Qi, R. Guo, Z. Bi, et al., Small 19 (2023) 2303632. doi: 10.1002/smll.202303632

  245. [245]

       Z. Li, J. Wang, S. Ma, et al., Appl. Catal. B: Environ. 310 (2022) 121335. doi: 10.1016/j.apcatb.2022.121335

  246. [246]

       L. Stegbauer, K. Schwinghammer, B. Lotsch, Chem. Sci. 5 (2014) 2789–2793. doi: 10.1039/C4SC00016A

  247. [247]

       H. Chen, H. Jena, X. Feng, et al., Angew. Chem. Int. Ed. 61 (2022) e202204938. doi: 10.1002/anie.202204938

  248. [248]

       R. Freund, O. Zaremba, G. Arnauts, et al., Angew. Chem. Int. Ed. 60 (2021) 23975–24001. doi: 10.1002/anie.202106259

  249. [249]

       T. Banerjee, F. Podjaski, J. Kröger, et al., Nat. Rev. Mater. 6 (2020) 168–190. doi: 10.1038/s41578-020-00254-z

  250. [250]

       K. Geng, T. He, R. Liu, et al., Chem. Rev. 120 (2020) 8814–8933. doi: 10.1021/acs.chemrev.9b00550

  251. [251]

       L. Chen, L. Wang, Y. Wan, et al., Adv. Mater. 32 (2020) 1904433. doi: 10.1002/adma.201904433

  252. [252]

       C. Krishnaraj, H. Jena, L. Bourda, et al., J. Am. Chem. Soc. 142 (2020) 20107–20116. doi: 10.1021/jacs.0c09684

  253. [253]

       D. Tan, R. Zhuang, R. Chen, et al., Adv. Funct. Mater. 34 (2024) 2311655. doi: 10.1002/adfm.202311655

  254. [254]

       Y. Cong, X. Li, S. Zhang, et al., ACS Appl. Mater. Interfaces 15 (2023) 43799–43809. doi: 10.1021/acsami.3c09039

  255. [255]

       M. Kou, Y. Wang, Y. Xu, et al., Angew. Chem. Int. Ed. 61 (2022) e202200413. doi: 10.1002/anie.202200413

  256. [256]

       Q. Zhi, W. Liu, R. Jiang, et al., J. Am. Chem. Soc. 144 (2022) 21328–21336. doi: 10.1021/jacs.2c09482

  257. [257]

       W. Zhao, P. Yan, B. Li, et al., J. Am. Chem. Soc. 144 (2022) 9902–9909. doi: 10.1021/jacs.2c02666

  258. [258]

       G. Pan, X. Hou, Z. Liu, et al., ACS Catal. 12 (2022) 14911–14917. doi: 10.1021/acscatal.2c03878

  259. [259]

       J. Yue, L. Song, Y. Fan, et al., Angew. Chem. Int. Ed. 62 (2023) e202309624. doi: 10.1002/anie.202309624

  260. [260]

       Q. Liao, Q. Sun, H. Xu, et al., Angew. Chem. Int. Ed. 62 (2023) e202310556. doi: 10.1002/anie.202310556

  261. [261]

       C. Wu, Z. Teng, C. Yang, et al., Adv. Mater. 34 (2022) 2110266. doi: 10.1002/adma.202110266

  262. [262]

       Y. Luo, B. Zhang, C. Liu, et al., Angew. Chem. Int. Ed. 62 (2023) e202305355. doi: 10.1002/anie.202305355

  263. [263]

       X. Di, X. Lv, H. Wang, et al., Chem. Eng. J. 455 (2023) 140124. doi: 10.1016/j.cej.2022.140124

  264. [264]

       P. Das, G. Chakraborty, J. Roeser, et al., J. Am. Chem. Soc. 145 (2023) 2975–2984. doi: 10.1021/jacs.2c11454

  265. [265]

       P. Das, J. Roeser, A. Thomas, Angew. Chem. Int. Ed. 62 (2023) e202304349. doi: 10.1002/anie.202304349

  266. [266]

       P. Gao, C. Wu, S. Wang, et al., J. Colloid. Interf. Sci. 650 (2023) 40–46. doi: 10.1016/j.jcis.2023.06.186

  267. [267]

       H. Hu, Y. Tao, D. Wang, et al., J. Colloid. Interf. Sci. 629 (2023) 750–762. doi: 10.1016/j.jcis.2022.09.111

  268. [268]

       C. Shao, Q. He, M. Zhang, et al., Chin. J. Catal. 46 (2023) 28–35. doi: 10.1016/S1872-2067(22)64205-0

  269. [269]

       L. Li, L. Xu, Z. Hu, J. Yu, Adv. Funct. Mater. 31 (2021) 2106120. doi: 10.1002/adfm.202106120

  270. [270]

       Q. Shang, Y. Liu, J. Ai, et al., J. Mater. Chem. A 11 (2023) 21109–21122. doi: 10.1039/d3ta03966h

  271. [271]

       S. Zhou, H. Hu, H. Hu, et al., Sci. China Mater. 66 (2023) 1837–1846. doi: 10.1007/s40843-022-2337-7

  272. [272]

       Y. Yang, J. Kang, Y. Li, et al., New J. Chem. 46 (2022) 21605–21614. doi: 10.1039/d2nj03744k

  273. [273]

       M. Deng, J. Sun, A. Laemont, et al., Green Chem. 25 (2023) 3069–3076. doi: 10.1039/d2gc04459e

  274. [274]

       Y. Liu, W. Han, W. Chi, et al., Appl. Catal. B: Environ. 331 (2023) 122691. doi: 10.1016/j.apcatb.2023.122691

  275. [275]

       J. Chu, Z. Liu, T. Yang, A. Kong, Appl. Surf. Sci. 611 (2023) 155717. doi: 10.1016/j.apsusc.2022.155717

  276. [276]

       T. Yang, Y. Wang, Y. Chen, et al., CrystEngComm 25 (2023) 4511–4520. doi: 10.1039/d3ce00607g

  277. [277]

       J. Sun, H. Jena, C. Krishnaraj, et al., Angew. Chem. Int. Ed. 62 (2023) e202216719. doi: 10.1002/anie.202216719

  278. [278]

       L. Wang, J. Sun, M. Deng, et al., Catal. Sci. Technol. 13 (2023) 6463–6471. doi: 10.1039/D3CY01175E

  279. [279]

       T. Yang, Y. Chen, Y. Wang, et al., ACS Appl. Mater. Interfaces 15 (2023) 8066–8075. doi: 10.1021/acsami.2c20506

  280. [280]

       F. Hao, C. Yang, X. Lv, et al., Angew. Chem. Int. Ed. 62 (2023) e202315456. doi: 10.1002/anie.202315456

  281. [281]

       Z. Zhou, M. Sun, Y. Zhu, et al., App. Catal. B: Environ. 334 (2023) 122862. doi: 10.1016/j.apcatb.2023.122862

  282. [282]

       X. Yu, B. Viengkeo, Q. He, et al., Adv. Sustain. Syst. 5 (2021) 2100184. doi: 10.1002/adsu.202100184

  283. [283]

       H. Wang, C. Yang, F. Chen, et al., Angew. Chem. Int. Ed. 61 (2022) e202202328. doi: 10.1002/anie.202202328

  284. [284]

       Y. Liu, L. Li, H. Tan, et al., J. Am. Chem. Soc. 145 (2023) 19877–19884. doi: 10.1021/jacs.3c05914

  285. [285]

       G. Xia, J. Qiu, L. Zhang, et al., Colloid Surf. A: Physicochem. Eng. Asp. 664 (2023) 131124. doi: 10.1016/j.colsurfa.2023.131124

  286. [286]

       Y. Yang, J. Liu, M. Gu, et al., Appl. Catal. B: Environ. 333 (2023) 122780. doi: 10.1016/j.apcatb.2023.122780

  287. [287]

       Y. Zhang, J. Qiu, B. Zhu, et al., Chem. Eng. J. 444 (2022) 136584. doi: 10.1016/j.cej.2022.136584

  288. [288]

       Y. Yang, Y. Li, X. Ma, et al., Catal. Sci. Technol. 13 (2023) 5599–5609. doi: 10.1039/d3cy00878a

  289. [289]

       H. Chen, S. Gao, G. Huang, et al., Appl. Catal. B: Environ. 343 (2024) 123545. doi: 10.1016/j.apcatb.2023.123545

  290. [290]

       L. Zhai, Z. Xie, C. Cui, et al., Chem. Mater. 34 (2022) 5232–5240. doi: 10.1021/acs.chemmater.2c00910

  291. [291]

       S. Chai, X. Chen, X. Zhang, et al., Environ. Sci. 9 (2022) 2464–2469 Nano. doi: 10.1039/d2en00135g

  292. [292]

       X. Zhang, J. Zhang, J. Miao, et al., Chem. Eng. J. 466 (2023) 143085. doi: 10.1016/j.cej.2023.143085

  293. [293]

       J. Chang, J. Shi, Q. Li, et al., Angew. Chem. Int. Ed. 62 (2023) e202303606. doi: 10.1002/anie.202303606

  294. [294]

       R. Liu, Y. Chen, H. Yu, et al., Nat. Catal. 7 (2024) 195–206. doi: 10.1038/s41929-023-01102-3

  295. [295]

       C. Shu, X. Yang, L. Liu, et al., Angew. Chem. Int. Ed. 63 (2024) e202403926. doi: 10.1002/anie.202403926

  296. [296]

       H. Li, H. Li, S. Xun, J. Brédas, Chem. Mater. 32 (2020) 9228–9237. doi: 10.1021/acs.chemmater.0c02913

  297. [297]

       Z. Lan, M. Wu, Z. Fang, et al., Angew. Chem. Int. Ed. 60 (2021) 16355–16359. doi: 10.1002/anie.202103992

  298. [298]

       Z. Luo, X. Chen, Y. Hu, et al., Angew. Chem. Int. Ed. 62 (2023) e202304875. doi: 10.1002/anie.202304875

  299. [299]

       Y. Qian, Y. Han, X. Zhang, et al., Nat. Commun. 14 (2023) 3083. doi: 10.1038/s41467-023-38884-w

  300. [300]

       Y. Wan, L. Wang, H. Xu, et al., J. Am. Chem. Soc. 142 (2020) 4508–4516. doi: 10.1021/jacs.0c00564

  301. [301]

       X. Wang, X. Yang, C. Zhao, et al., Angew. Chem. Int. Ed. 62 (2023) e202302829. doi: 10.1002/anie.202302829

  302. [302]

       L. Liu, F. Chen, J. Wu, et al., Proc. Natl. Acad. Sci. U. S. A. 120 (2023) e2215305120. doi: 10.1073/pnas.2215305120

  303. [303]

       A. Gopakumar, P. Ren, J. Chen, et al., J. Am. Chem. Soc. 144 (2022) 2603–2613. doi: 10.1021/jacs.1c10786

  304. [304]

       Y. Zheng, Z. Yu, H. Ou, et al., Adv. Funct. Mater. 28 (2018) 1705407. doi: 10.1002/adfm.201705407

  305. [305]

       B. Sheng, Y. Xie, Q. Zhao, et al., Energy Environ. Sci. 16 (2023) 4612–4619. doi: 10.1039/d3ee02200e

  306. [306]

       T. Freese, J. Meijer, B. Feringa, S. Beil, Nat. Catal. 6 (2023) 553–558. doi: 10.1038/s41929-023-00980-x

  307. [307]

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

  308. [308]

       C. Zhao, X. Wang, Y. Yin, et al., Angew. Chem. Int. Ed. 62 (2023) e202218318. doi: 10.1002/anie.202218318

  309. [309]

       L. Wang, L. Liu, Y. Li, et al., Adv. Energy Mater. 14 (2023) 2303346.

  310. [310]

       R. Yang, Y. Fan, J. Hu, et al., Chem. Soc. Rev. 52 (2023) 7687–7706. doi: 10.1039/d2cs00205a

  311. [311]

       Z. Wang, C. Li, K. Domen, Chem. Soc. Rev. 48 (2019) 2109–2125. doi: 10.1039/C8CS00542G

  312. [312]

       S. Cao, L. Piao, Angew. Chem. Int. Ed. 59 (2020) 18312–18320. doi: 10.1002/anie.202009633

  313. [313]

       L. Liu, M. Gao, H. Yang, et al., J. Am. Chem. Soc. 143 (2021) 19287–19293. doi: 10.1021/jacs.1c09979

  314. [314]

       Q. Wang, X. Kong, Y. Wang, et al., ChemSusChem 15 (2022) e202201514. doi: 10.1002/cssc.202201514

  315. [315]

       H. Zhao, Q. Jin, M. Khan, et al., Chem. Catal. 2 (2022) 1720–1733. doi: 10.1016/j.checat.2022.04.015

  316. [316]

       Y. Li, S. Ouyang, H. Xu, et al., J. Am. Chem. Soc. 138 (2016) 13289–13297. doi: 10.1021/jacs.6b07272

  317. [317]

       Q. Lian, Z. Liang, X. Guan, et al., Appl. Catal. B: Environ. 331 (2023) 122724. doi: 10.1016/j.apcatb.2023.122724

  318. [318]

       S. Zhou, R. He, J. Pei, et al., Environ. Sci. Technol. 56 (2022) 10474–10482. doi: 10.1021/acs.est.2c02067

  319. [319]

       Z. Chu, J. Yang, W. Zheng, et al., Coord. Chem. Rev. 481 (2023) 215049. doi: 10.1016/j.ccr.2023.215049

• 

  Figure 1  Schematic illustration of (a) photocatalytic processes for H2O2 production and (b) corresponding energy diagrams. (c) Reaction pathways and corresponding energy diagrams for ORR and WOR pathways.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) Photocatalytic H₂O₂ production over HT-CN/TiO₂ featuring the Z-scheme charge transfer mechanism. Reproduced with permission [64]. Copyright 2019, Elsevier. (b) The photocatalytic production of H₂O₂ pathway over In₂S₃@In₂O₃ and In₂S₃@O_v/In₂O₃ composites. Reproduced with permission [86]. Copyright 2021, American Chemical Society. (c) Schematic diagram of the reaction mechanism over Au/WO₃. Reproduced with permission [69]. Copyright 2021, Elsevier. (d) Mechanism of Gd³⁺ doped WO₃ photocatalysts for photocatalytic H₂O₂ production. Reproduced with permission [76]. Copyright 2023, Elsevier.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Mn²⁺ doping and Au⁰ modification on MoS₂ nanosheets, and in-situ-Fenton process using photogenerated H₂O₂. Reproduced with permission [94]. Copyright 2019, Elsevier. (b) Mechanism of S_v−MoS₂/BN heterojunction toward sacrificial agent-free photocatalytic H₂O₂ production for in situ water disinfection. Reproduced with permission [96]. Copyright 2024, Elsevier. (c) Schematic diagram of H₂O₂ production over CBO@MoS₂ Z-scheme heterojunction Reproduced with permission [97]. Copyright 2023, Royal Society of Chemistry. (d). Mechanism of CDs modified SnS₂/In₂S₃ type Ⅱ heterostructure for photocatalytic H₂O₂ production by indirect two-step e^– ORR and 4e^– WOR. Reproduced with permission [99]. Copyright 2021, Royal Society of Chemistry.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) Schematic diagram of efficient photocatalytic H₂O₂ production by Cu@Au/BiVO₄. Reproduced with permission [119]. Copyright 2021, American Chemical Society. (b) Schematic illustration for the H₂O₂ production mechanism of Y-doped BiVO₄ photocatalyst. Reproduced with permission [124]. Copyright 2023, John Wiley and Sons. (c) H₂O₂ production on defective BiVO₄ via the direct one-step 2e^– ORR pathway. Reproduced with permission [125]. Copyright 2023, Elsevier. (d) Energy diagram for NiCo₂O₄/BiVO₄ heterojunction depicting the production of H₂O₂ and degradation of organic pollutants, simultaneously. Reproduced with permission [126]. Copyright 2022, American Chemical Society.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) Mechanism of PDI-modified g-C₃N₄ for photocatalytic H₂O₂ production from H₂O and O₂. Reproduced with permission [154]. Copyright 2014, John Wiley and Sons. (b) The different adsorption modes of O₂ adsorption on metal surfaces and the mechanism of photocatalytic H₂O₂ production. Reproduced with permission [172]. Copyright 2021, Springer Nature. (c) Mechanism of an aryl amino-substituted g-C₃N₄ photocatalyst with atomically dispersed Mn for photocatalytic H₂O₂ production from seawater. Reproduced with permission [163]. Copyright 2023, American Chemical Society. (d) Mechanism of heteroatom doped g-C₃N₄ (AKMT) for photocatalytic H₂O₂ production. Reproduced with permission [174]. Copyright 2020, John Wiley and Sons. (e) The presence of C-vacancies changed the photocatalytic H₂O₂ production from the indirect two-step e^– ORR pathway to the direct one-step 2e^– ORR pathway. Reproduced with permission [181]. Copyright 2016, Elsevier. (f) Mechanism of PI_x-NCN heterojunction for photocatalytic H₂O₂ production. Reproduced with permission [197]. Copyright 2017, Elsevier. (g) Mechanism of heptazine/triazine layer stacked g-C₃N₄ heterojunction with F/K dual sites modified for photocatalytic H₂O₂ production. Reproduced with permission [200]. Copyright 2022, American Chemical Society.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  (a) Proposed reaction mechanism for the generation of H₂O₂ over Ce-doped MIL-125-NH₂. Reproduced with permission [235]. Copyright 2023, John Wiley and Sons. (b) The proposed mechanism of H₂O₂ production over alkylated MIL-125-R7 in a two-phase system. Reproduced with permission [224]. Copyright 2019, John Wiley and Sons. (c) Reaction mechanism of H₂O₂ production over missing-linker defective UiO-66-NH₂-X photocatalyst. Reproduced with permission [236]. Copyright 2021, American Chemical Society. (d) The relationship between the yield of H₂O₂ production and heterojunction coverage, as well as the proposed mechanism over NH₂−MIL-125@ZnS. Reproduced with permission [238]. Copyright 2022, John Wiley and Sons.

  下载: 全尺寸图片 幻灯片
  

  Figure 7  (a) Reaction mechanism of photocatalytic H₂O₂ production over sulfone functional group modified COF (FS-COFs). Reproduced with permission [262]. Copyright 2023, John Wiley and Sons. (b) Proposed reaction mechanism for the generation of H₂O₂ over fluorine-doped COF (TF₅₀–COF). Reproduced with permission [283]. Copyright 2022, John Wiley and Sons. (c) Schematic illustration of photocatalytic reaction pathways toward H₂O₂ production over Z-scheme WO₃/COF heterojunction. Reproduced with permission [288]. Copyright 2023, Royal Society of Chemistry. (d) The schematic diagram of photocatalytic H₂O₂ production over CHF-DPA and CHF-DPDA with separated redox centres. Reproduced with permission [46]. Copyright 2022, John Wiley and Sons.

  下载: 全尺寸图片 幻灯片

  Table 1.  Pros and cons of various H₂O₂ detection methods.

  下载: 导出CSV

  Table 2.  The main advantages and challenges of different 2D photocatalysts.

  下载: 导出CSV
•