Functionalized Graphitic Carbon Nitrides for Photocatalytic H₂O₂ Production: Desired Properties Leading to Rational Catalyst Design

Abstract

Photocatalytic H₂O₂ production based on graphitic carbon nitride (g-C₃N₄) materials has been attracting increasing attention. However, it is difficult to reveal the inner relationships among the structure, properties and performance of a g-C₃N₄-based photocatalyst by simply summarizing preparation methods, properties and performances in previous works. In this review, the three most important issues for improving H₂O₂ generation based on the band diagram and physicochemical properties of pristine g-C₃N₄ are proposed. Improvement of the charge separation, promotion of the light absorption and introduction of active sites for 2e⁻ oxygen reduction reaction to suppress side reactions are the most three attractive strategies for enhancing the activities. Following discussion of these strategies, representative functionalization methods are summarized on the basis of the most desired properties for improving the photocatalytic activities for H₂O₂ production. Other influence factors for improving H₂O₂ production such as addition of electron donors and adjustment of pH value of the solution are also discussed. Future challenges for photocatalytic H₂O₂ based on g-C₃N₄ materials are also summarized to provide future directions in this field.

1. Introduction

H₂O₂ has been recognized as one of the 100 most important chemicals worldwide (Jones C.W., 1999; Pesterfield L., 2009). The reactive oxygen content in H₂O₂ (47.1 % w/w) is the highest among all chemicals. No toxic by-products are produced from H₂O₂, making H₂O₂ a highly efficient and environmental-friendly oxidant. H₂O₂ has been widely applied for disinfection and sterilization (McDonnell G., 2014; Teng Z. et al., 2019), organic synthesis (Zhan W. et al., 2018) and wastewater treatment (Zhang M.-h. et al., 2019). A one-compartment fuel cell using H₂O₂ in which H₂O₂ is used as both an oxidant and a reductant has recently been developed (Mousavi Shaegh S.A. et al., 2012). Studies on this fuel cell showed that the theoretical output potential is 1.09 V, which is comparable to that of a conventional hydrogen (H₂) fuel cell (1.23 V) (Mousavi Shaegh S.A. et al., 2012; Kim D. et al., 2015; Yamada Y. et al., 2013). Compared to H₂, H₂O₂ showed a promising prospect as a water-soluble solar fuel (H₂O₂, 3.0 MJL⁻¹, in an aqueous solution higher than the intensity of compressed H₂ gas, 2.8 MJL⁻¹, 35 MPa) (Kim D. et al., 2015). Unlike hydrogen, H₂O₂ is fully soluble in water and is easily transportable, making it an ideal energy carrier alternative to H₂. The anthraquinone (AQ) method for H₂O₂ production is the main process used in industry (more than 95 % of total production, Fig. 1) (Campos-Martin J.M. et al., 2006). The AQ process includes the following four steps: (i) in an organic solvent, hydrogenation of anthraquinone (AQ) is firstly catalyzed on an Ni/Pd catalyst; (ii) hydrogenated anthraquinone (HAQ) is oxidized with the aid of a catalyst; (iii) H₂O₂ is extracted by pure water, and HAQ is used to reproduce AQ; and iv) the as-prepared H₂O₂ is purified and concentrated for following commercial use. These hydrogenation and oxidation reactions consume a lot of energy. Additionally, the AQ process is not environmental-friendly because large amounts of wastes are generated (Campos-Martin J.M. et al., 2006).

Fig. 1

Schematic of the anthraquinone oxidation (AQ) process.

Efforts have been made to develop new methods for H₂O₂ production under more moderated reaction conditions with less waste production. (Yi Y. et al., 2016). Direct synthesis from H₂ and O₂ in the presence of a Pd catalyst or a bimetallic Au–Pd catalyst is a possible alternative route for H₂O₂ production based on the reaction of H₂ + O₂ = H₂O₂. Typical reaction conditions mainly include H₂/O₂ mixture gas catalyzed by noble metal catalysts (Xia C. et al., 2019; Freakley S.J. et al., 2016). The ratio of the H₂/O₂ mixture needs to be precisely controlled in the process to avoid the risk of explosions (Xia C. et al., 2019; Freakley S.J. et al., 2016). Recently, much attention has been given to the possibility of efficient H₂O₂ production by a photocatalytic reaction process, which would be a significant breakthrough in H₂O₂ chemistry. A photocatalytic approach would enable the use of explosive H₂ gas to be avoided and it would only need light, O₂ and water, with no pollutant generated in the overall process. Much interest has therefore been shown in photocatalytic strategies for H₂O₂ production.

In past one decade, many kinds of semiconductors including metal oxides, metal sulfides, metal organic compounds and organic semiconductors have been prepared for photocatalytic H₂O₂ production (Hou H. et al., 2020). Among these photocatalysts, graphite-like carbon nitride (g-C₃N₄), also named polymeric carbon nitride (PCN), as a metal-free polymer semiconductor (n-type), possesses a stacked two-dimensional (2D) structure of tri-s-triazine connected via tertiary amines (Wang X. et al., 2009). Because of its unique electrical, optical, structural and physicochemical properties, g-C₃N₄ has been recognized as a new class of cost-efficient multifunctional materials for electronic, catalytic and energy applications (Banerjee T. et al., 2021). Shiraishi and co-workers first found in 2014 that pristine g-C₃N₄ showed some photocatalytic activity for H₂O₂ production with the addition of ethanol as an electron donor (Shiraishi Y. et al., 2014b). Since then, more than 80 studies on photocatalytic H₂O₂ production based on g-C₃N₄ materials have been published. Several reviews summarizing some of the recent progress in photocatalytic H₂O₂ production have also been published (Hou H. et al., 2020; Zhang G. et al., 2020). Most of those reviews focused on how the various protocols for preparation of functionalized g-C₃N₄ influenced the photocatalytic activity (Hou H. et al., 2020; Zhang G. et al., 2020). However, it is difficult to reveal the inner relationships among the structure, properties and performance of a g-C₃N₄-based photocatalyst by simply summarizing preparation methods, properties and performances (Zhang G. et al., 2020; Torres-Pinto A. et al., 2019). Generally, the structure of a photocatalyst defines its properties, and the properties of the photocatalyst determine its performance. Thus, it is necessary to determine the desired properties of a g-C₃N₄-based photocatalyst for H₂O₂ production. Determination of the desired properties would lead to a comprehensive understanding of the photocatalyst from rational structure design to good performance.

In this review, we firstly show the principles of photocatalytic H₂O₂ production. Then the three most important issues for improving H₂O₂ generation based on the band diagram and physicochemical properties of pristine g-C₃N₄ are proposed. By combining the fundamental principles with preparation strategies and possible influencing factors in liquid-phase photocatalytic systems, future challenges and possible directions for H₂O₂ production using functionalized g-C₃N₄ catalysts are presented in this review with the hope that they will inspire design and preparation strategies for not only carbon nitrides but also a variety of photocatalysts with good performance.

2. Principles of photocatalytic H₂O₂ generation

Since solar energy is renewable and sustainable, photocatalysis using power systems has received much attention in the past few years and it has been applied to various fields including water splitting for H₂ and O₂ production (Wang X. et al., 2009; Wang Q. and Domen K., 2020), CO₂ reduction (Liu R. et al., 2020), nitrogen fixation (Zhang G. et al., 2020), degradation of pollutants (Koe W.S. et al., 2020) and disinfections (Teng Z. et al., 2019), organic synthesis and H₂O₂ production (Hou H. et al., 2020). Photocatalytic power systems are cost-efficient and easy to operate compared with photo-electrochemical systems and photovoltaic-electrochemical systems (Hisatomi T. and Domen K., 2019). There are three typical steps in a photocatalytic process (Fig. 2) (Nosaka Y. and Nosaka A., 2016). Sufficient light absorption, efficient charge separation, and surface reactions are crucial properties for typical efficient photocatalysts.

Fig. 2

Photoexcitation and charge decay pathway.

Photocatalytic production of H₂O₂ follows the photocatalytic reaction principles. H₂O₂ can be generated through either oxygen reduction reaction (ORR) or water oxidation reaction (WOR) (Fan W. et al., 2020) as shown by a schematic diagram of photocatalytic H₂O₂ production in Fig. 3. The light-driven 2e⁻ WOR pathway (Eqn. (3), Table 1) is difficult to be achieved because of the uphill thermodynamics (1.76 V vs. a reversible hydrogen electrode, simplified as RHE, Table 1), i.e., the decomposition of as-produced H₂O₂ will occur at this potential (1.76 V) because H₂O₂ is an excellent hole scavenger (Shi X. et al., 2017). In the case of the ORR pathway, H₂O₂ can be produced by either a 2e⁻ direct oxygen reduction (O₂ → H₂O₂) route or a sequential 1e⁻ indirect reduction (O₂ →•O₂⁻ → H₂O₂). For h⁺ in consumption, the most effective reaction route is to oxidize H₂O into O₂ and H⁺ via a 4e⁻ pathway (Eqn. (1)). Eqn. (7) to Eqn. (10) in Table 1 show production of H₂O₂ via the two-step 1e⁻ reduction pathway. First, superoxide radicals (•O₂⁻) are formed (Eqn. (7)) and further produce HO₂• radicals with the reaction of H⁺ (Eqn. (8)). Then, the HO₂• radicals could produce HO₂⁻ anions by undergoing another 1e⁻ reduction (Eqn. (9)). Finally, HO₂⁻ reacts with H⁺ to generate H₂O₂. Eqn. (2) shows the one-step 2e⁻ direct reduction for H₂O₂ production. In this process, O₂ directly reacts with two protons to form an H₂O₂ product via two-electron photoreduction. It should be noted that the back reaction of WOR via a 4e⁻ pathway is the ORR process via a 4e⁻ pathway (Eqn. (1)), which could also be a competitive reaction for H₂O₂ production via 2e⁻ ORR. The most ideal situation would be for photocatalytic production of H₂O₂ to be generated from H₂O and O₂ via 4e⁻ WOR and 2e⁻ ORR, which is an uphill reaction (ΔG₀ = 117 kJ mol⁻¹).

Fig. 3

Schematic diagram for photocatalytic H₂O₂ production. (a) Photocatalytic H₂O₂ production via water oxidation reaction. (b) Photocatalytic H₂O₂ production via oxygen reduction. CB is the abbreviation for conduction band. VB is the abbreviation for valence band.

Table 1 Standard electrode potentials for aqueous ORR half-reactions (Gao J. and Liu B., 2020). Reprinted with permission from Ref. (Gao J. and Liu B., 2020). Copyright: (2020) American Chemical Society.

Eqn. No.	Half-reactions	E0 (V vs. SHE*)	E0 (V vs. RHE**)
(1)	O2(g) + 4H+ + 4e− ↔ 2H2O	1.229	1.229
(2)	O2(g) + 2H+ + 2e− ↔ H2O2	0.695	0.695
(3)	H2O2 + 2H+ + 2e− ↔ 2H2O	1.763	1.763
(4)	O2(g) + 2H2O + 4e− ↔ 4OH−	0.401	1.229
(5)	O2 + H2O + 2e− ↔ HO2− + OH−	−0.065	0.764
(6)	HO2− + H2O + 2e− ↔ 3OH−	0.867	1.696
(7)	O2(g) + e− ↔ O2−	−0.33	−0.33
(8)	O2(g) + H+ + e− ↔ HO2 •	−0.05	−0.05
(9)	HO2• + e− ↔ HO2−	0.76	0.76
(10)	HO2• + H+ + e− ↔ H2O2	1.44	1.44
(11)	H2O2 + H+ + e− ↔ H2O + •OH	0.793	0.793
(12)	•OH + e− ↔ OH−	1.89	2.72

*  SHE: Standard hydrogen electrode.

**  RHE: Reversible hydrogen electrode.

3. Desired physicochemical properties of polymetric carbon nitride for photocatalytic H₂O₂ production

Since Wang and his co-workers first discovered the photocatalytic activities of g-C₃N₄ (Wang X. et al., 2009), g-C₃N₄-based particulate photocatalysts have attracted much interest. The band gap of g-C₃N₄ (Carbon nitrides produced by thermal treatment at 500~550 °C with nitrogen-rich precursors have various names as polymeric carbon nitride (PCN), graphitic carbon nitride (GCN), and melon) is ~2.7 eV, corresponding to an optical wavelength of ~460 nm, which makes it a promising visible-light-response photocatalyst with an appropriate band structure (Fig. 4) (Ong W.-J. et al., 2016; Hou H. et al., 2020). Additionally, the conjugated heptazine ring in the PCN matrix can provide active sites for 2e⁻ ORR (Shiraishi Y. et al., 2014a; Shiraishi Y. et al., 2014b), which is similar to the catalytic properties of graphene for 2e⁻ ORR (Kim H.W. et al., 2018; Lu Z. et al., 2018). As such, g-C₃N₄ has emerged as a star material in the field of photocatalytic H₂O₂ production.

Fig. 4

Band diagrams of g-C₃N₄ and summary of potentials for redox reactions during H₂O₂ production by using O₂ and/or H₂O. CBM is the abbreviation for conduction band minimum and VBM is the abbreviation for valence band maximum.

However, the activity of g-C₃N₄ for photocatalytic H₂O₂ production is still restricted by low efficiency because of some drawbacks of pristine g-C₃N₄. Firstly, a large concentration of defects is inevitably introduced into g-C₃N₄ during thermal polymerization (Kim D. et al., 2015; Ong W.-J. et al., 2016). This means that the inner-plane structure of g-C₃N₄ is not completely composed of covalent bonds, i.e., the interactions are hydrogen bonds between melons ([C₆N₉H₃]_n) (Fig. 5) (Kessler F.K. et al., 2017). As a result, the charge separation and charge migration are significantly hindered, resulting in a high charge recombination rate of PCN. Secondly, the large band gap (2.7 eV) and small absorption of visible light (Kubelka-Munk absorbance usually less than 2, λ = 420 nm) result in insufficient visible-light harvesting. Thirdly, although the side-on adsorption of molecular oxygen on the π-conjugated heptazine ring enables g-C₃N₄ (Shiraishi Y. et al., 2014b) to have some selectivity for 2e⁻ ORR, it is difficult for the stepwise 1e⁻ to 1e⁻ ORR reaction to be prevented since signals of superoxide radicals can still be detected during the photocatalytic reactions with electron spin resonance during the photocatalytic reactions (Li S. et al., 2016). In this case, reaction sites with higher selectivity for 2e⁻ ORR have to be constructed for improving the overall activity. To overcome these drawbacks, many protocols including tuning morphologies, defect engineering, loading cocatalysts, copolymerization of semiconductors, doping elements and hybridization have been developed. Each of these functionalization strategies overcome one or two drawbacks by changing the following properties of pristine PCN: (i) suppressing charge recombination, (ii) narrowing the bandgap or/and promoting light absorbance, and (iii) introducing active sites for selective 2e⁻ ORR (Fig. 6). Based on the improved physicochemical properties, functionalization strategies for improvement in the activity for photocatalytic H₂O₂ production can also be divided into three parts, which provide a comprehensive understanding from material preparation to physicochemical properties, thus providing the insight into the relationships among structure, properties and performance.

Fig. 5

(a) Schematic illustration of the synthesis process of g-C₃N₄ by thermal polymerization of different precursors such as melamine, cyanamide, dicyanamide, urea, and thiourea. The black, blue, white, red, and yellow balls denote C, N, H, O, and S atoms, respectively. (b) Reaction pathway for the development of g-C₃N₄ using cyanamide as the precursor. Reprinted with permission from Ref. (Ong W.-J. et al., 2016). Copyright: (2016) American Chemical Society.

Fig. 6

Solutions for overcoming the disadvantages of pristine g-C₃N₄ by manipulating physicochemical properties via 2e⁻ ORR.

3.1 Suppressing charge recombination

3.1.1 Hybridization

3.1.1.1 Hybridization with carbon-based materials

Carbonaceous nanomaterials are promising modifiers for photocatalysts because of their sp²-hybridized electronic structures (Shiraishi Y. et al., 2014a; Kim H.W. et al., 2018; Lu Z. et al., 2018; Yang S. et al., 2018). These materials usually work as electron transfer materials and photosensitizers to broaden the adsorption edge and improve the charge migration (Yang S. et al., 2018; Thakur S. et al., 2017; Zeng X. et al., 2017).

Carbon nanotubes (CNTs) with a π-conjugative structure are capable of accepting, transporting and storing electrons for g-C₃N₄ (Tasis D. et al., 2006). An amination reaction process was developed by Zhao and co-workers to introduce CNTs into nanosheets of g-C₃N₄, in which g-C₃N₄ was covalently combined with CNTs (Fig. 7a, b) (Zhao S. et al., 2018a). The CNTs covalently combined with g-C₃N₄ promoted charge migration by shifting the CB level to enhance the single-electron reduction of O₂ to ·O₂⁻, thus leading to enhanced ORR activity (Zhao S. et al., 2018a). As shown in Fig. 7c, the photocurrent of g-C₃N₄/CNTs was significantly increased compared with that pristine PCN. Shiraishi and co-workers also used reduced graphene oxide (rGO) for hybridization with a g-C₃N₄/perylene diimide (PDI) catalyst for further activity improvement (Fig. 7d, e). The improved activity may due to the ultra-thin two-dimensional nature of rGO (g-C₃N₄/PDI/rGO_0.05) with high charge carrier mobility (Kofuji Y. et al., 2016a). As shown in Fig. 7f, the photocurrent of g-C₃N₄/PDI/rGO_0.05 is significantly larger than that of g-C₃N₄/PDI. However, no current response is found by using a rGO electrode, indicating a synergic effect of g-C₃N₄/PDI and rGO. It should be noted that the formation of tight junctions between the g-C₃N₄-based photocatalyst and carbon-based material is extremely important for the formation of electron transfer pathways since the photocurrent intensity of the physical mixture of rGO and g-C₃N₄/PDI is 20 % smaller than the photocurrent density of g-C₃N₄/PDI (Kofuji Y. et al., 2016a).

Fig. 7

Hybridization with a carbon-based material for improved charge separation. (a) TEM images of g-C₃N₄-CNTs. (b) Process for preparation of g-C₃N₄-CNTs. (c) Photocurrent response of g-C₃N₄-CNTs or g-C₃N₄-loaded electrodes under light at a bias of 0.25 V vs. SCE (λ ≥ 400 nm). (d) TEM image of g-C₃N₄/PDI/rGO_0.05 and its SAED pattern (inverse contrast) with indexed graphite pattern. (e) Schematic illustration of the proposed mechanism for photocatalytic H₂O₂ production on g-C₃N₄/PDI/rGO. (f) Photocurrent responses of g-C₃N₄/PDI/rGO_0.05 and other g-C₃N₄-based catalysts measured on FTO in 0.1 M Na₂SO₄ solution under visible light (λ > 420 nm) at a bias of 0.5 V (vs Ag/AgCl). SCE is an abbreviation for saturated calomel electrode, SAED is an abbreviation for selected area electron diffraction, and FTO is an abbreviation for fluorine-doped tin oxide. Reprinted with permission from Ref. (Zhao S. et al., 2018a; Kofuji Y. et al., 2016a). Copyright: (2018) Elsevier and (2016) American Chemical Society.

3.1.1.2 Hybridization with metal and metal compounds

Fabrication of metal compounds to form a Schottky barrier between metal-based semiconductors and PCN material has been proved to be an efficient approach for enhancing the photocatalytic efficiency for several reactions (Ong W.-J. et al., 2016; Teixeira I.F. et al., 2018). The nature of the formation of Schottky barriers in PCN/metal compounds is due to the differences between the charge potentials of metal compounds and PCN under light irradiation (Zhang Z. and Yates J.T., 2012). The electrons tend to migrate to semiconductors that have more positive potentials, while holes tend to migrate to those with negative potentials (Nosaka Y. and Nosaka A., 2016). Although the overall redox potential of the photocatalyst is decreased, the charge separation of charge carriers is significantly improved, thus leading to a significantly improved charge separation (Fig. 8a). However, there are very few reports on the promotion of photocatalytic efficiency for H₂O₂ production by construction of heterojunctions based on PCN/metal compounds. There are two possible reasons. On the one hand, the potential of the conduction band minimum (0.9~1.3 V vs. SHE) of PCN (Fig. 4) is more negative than that of common metal compounds. In this case, the photoelectrons tend to accumulate on the metal oxides. Thus, the potential of the metal compounds determines the reduction potential of the heterojunction materials. As shown in Fig. 8b, the photocatalytic reduction potential of TiO₂, Cu₂O, CdS and ZnS can hardly facilitate 2e⁻ ORR because of the kinetically favored 1e⁻ ORR to form •O₂⁻, thus resulting in a stepwise ORR to generate H₂O₂. For other metal compound semiconductors, 1e⁻ ORR can be prevented because of the positively shifted CBM. However, the potential differences between the redox potentials of O₂, 2H⁺/H₂O₂ and CBM of Fe₂O₃, CuWO₄, SnO₂, WO₃, etc. are quite small, which may lead to a poor thermodynamic driving force for 2e⁻ ORR. On the other hand, formed H₂O₂ can be decomposed by metal compounds by disproportionation or photoreaction (Li X. et al., 2001). Therefore, the formation of heterojunctions by contact of PCN with one metal compound may be an unacceptable strategy unless a unique electron transfer pathway between PCN and metal compounds is formed, such as in Z-scheme catalysts, so that reduction will occur on the surface of PCN (Wang S. et al., 2020a).

Fig. 8

Schematic diagram of (a) semiconductor heterojunctions and (b) diagram of band positions based on experimental results for common metal-based photocatalysts. NHE is an abbreviation for normal hydrogen electrode.

Loading active noble metal nanoparticles (NPs) is one of the effective approaches for boosting the photocatalytic activity of a photocatalyst for H₂O₂ production (Hirakawa H. et al., 2016; Tsukamoto D. et al., 2012). Zuo and coworkers observed a similar improvement in the activity of an Au/C₃N₄ photocatalyst for H₂O₂ production (Zuo G. et al., 2019). In that study, the effects of different noble-metal cocatalysts on g-C₃N₄ for H₂O₂ production were also compared, and the results showed that g-C₃N₄ loaded with Au nanoparticles (NPs) had the highest activity for H₂O₂ production (Zuo G. et al., 2019). The enhancement of H₂O₂ production was attributed to the efficient separation of photogenerated electrons and holes between Au NPs and g-C₃N₄ (Zuo G. et al., 2019). The function of Au NPs on PCN for charge separation is similar to that of carbon materials. The Au NPs can improve the carrier mobility of photoelectrons so that the charge recombination can be suppressed.

3.1.1.3 Hybridization with metal-free semiconductors and molecules

The advantages of cost efficiency, mechanical flexibility and easy operation makes organic semiconductors promising alternatives for inorganic materials to improve the charge separation (Chung D.D.L., 2019). Triazine (Jiang X. et al., 2015) and aromatic diimides have been used as modifiers for g-C₃N₄ based photocatalysts (Shiraishi Y. et al., 2014a; Kofuji Y. et al., 2016a; Kofuji Y. et al., 2018; Kofuji Y. et al., 2017; Kofuji Y. et al., 2016b). Shiraishi and co-workers incorporated several aromatic diimides (Shiraishi Y. et al., 2014a; Kofuji Y. et al., 2018; Kofuji Y. et al., 2016b; Kofuji Y. et al., 2017) into the g-C₃N₄ matrix by thermal condensation (Fig. 9a). All of the g-C₃N₄/PDI, g-C₃N₄/biphenyl diimide (BDI) and g-C₃N₄/mellitic triimide (MTI) catalysts successfully produced H₂O₂ (millimolar level) in pure water with O₂ (Shiraishi Y. et al., 2014a; Kofuji Y. et al., 2016a; Kofuji Y. et al., 2018; Kofuji Y. et al., 2017; Kofuji Y. et al., 2016b). The photocurrent showed a significant improvement after the copolymerization of aromatic diimides. It should be noted that the introduction of more C=O groups significantly improved the photocurrent (Fig. 9b) (Kofuji Y. et al., 2017). The C=O groups in these aromatic diimides may accumulate negative charges and then increase the thickness of the space-charge region, which might improve the separation efficiency of charge carriers (Koe W.S. et al., 2020) and disinfections (Teng Z. et al., 2019). Based on this phenomenal result from the properties of C=O groups, Ohno and co-workers prepared C=O functionalized PCN by copolymerization of 2,5,8-triamino-tri-s-triazine (melem) and barbituric acid (BA) (PCNBA, Fig. 9c) (Teng Z. et al., 2020). In the photoluminescence (PL) spectra (Fig. 9d), there was a significant decrease in intensity together with a notable increase in photocurrent intensity (Teng Z. et al., 2020). The copolymerization of BA significantly suppressed recombination and accelerated the charge separation.

Fig. 9

Hybridization with electron-deficient organic semiconductors for improved charge separation. (a) Scheme of the synthesis of g-C₃N₄ hybridized by aromatic diimide using g-C₃N₄/PDI_x as an example. (b) Photocurrent responses of g-C₃N₄/MTI₄₉ and g-C₃N₄/PDI₅₁ in 0.1 M Na₂SO₄ (bias: 0.5 V vs Ag/AgCl). (c) Process for polymerization of PCNBA with barbituric acid and melem as precursors. (d) Photoluminescence spectra of PCN and PCNBA samples. Reprinted with permission from Ref. (Kofuji Y. et al., 2017; Teng Z. et al., 2020). Copyright: (2017) American Chemical Society and (2020) Elsevier.

In the case of incorporation with semiconductors, Shiraishi and co-workers combined PCN/PDI with boron nitride (BN) to separate holes since BN has a relatively negative VBM (Kofuji Y. et al., 2018). In the system of g-C₃N₄-PDI-BN_0.2-rGO, the photocurrent increased about 20 % (Kofuji Y. et al., 2018). It is notable that the oxygen evolution of g-C₃N₄-PDI-BN_0.2-rGO was increased by more than 15 % compared with that of g-C₃N₄-PDI-rGO (Kofuji Y. et al., 2016a), further confirming that the charge separation introduced by BN separated photogenerated holes, leading to an improved charge separation. Perylene imides (PIs) were modified on g-C₃N₄ nanosheets (Yang L. et al., 2017) to fabricate a Z-scheme configuration for promotion of charge separation. As a result, more electrons from the CB of the g-C₃N₄ part efficiently generated H₂O₂, while the holes of g-C₃N₄/PI oxidized OH⁻ to OH (Yang L. et al., 2017), which also subsequently reacted to produce H₂O₂.

3.1.2 Doping

3.1.2.1 Metal and non-metal doping

Optical, electronic and other physical properties of g-C₃N₄ materials could be manipulated by doping strategies that introduce additional elements and impurities into the framework (Liu X. et al., 2020). In the case of charge separation, one of the most efficient strategies is to incorporate alkaline metal ions. Alkaline metal ions were used in several works to functionalize PCN, leading to a promoted charge separation for organic synthesis (Qiu C. et al., 2019) (Fig. 10a). Hydrogen evolution reduction (HER) (Sun S. et al., 2019), carbon dioxide reduction (COR) (Wang S. et al., 2020b), and nitrogen fixation, (Li X. et al., 2018a) can all be facilitated by incorporation of alkaline metal ions into PCN. Moon and co-workers incorporated potassium (K), phosphorus (P), and oxygen (O) elements into a g-C₃N₄ matrix by thermal calcination of melamine with potassium phosphate (KPD, K₂HPO₄) (Moon G.-h. et al., 2017). The charge separation of KPD was significantly improved because of the significantly decreased PL intensity (Fig. 10b). The hector-elements in the g-C₃N₄ matrix increased the carrier lifetime to a picosecond range. Consequently, the generation rate of H₂O₂ by K-P-O-doped g-C₃N₄ was notably promoted compared to that by pure g-C₃N₄ (Moon G.-h. et al., 2017) (Fig. 10c). Halogen (Br or Cl)-doped g-C₃N₄ nanorods could also accelerate H₂O₂ production (Zhang C. et al., 2018). The charge carrier transfer in the matrix of g-C₃N₄ was promoted by the introduction of heteroatoms (Zhang C. et al., 2018). Kim et al. studied the co-doping of metal (K) and nonmetal ions (P and F) into the g-C₃N₄ framework (Kim S. et al., 2018). The co-doped g-C₃N₄ also significantly promoted charge separation. Under visible light irradiation, the co-doped g-C₃N₄ exhibited higher activity for photocatalytic production of H₂O₂ than that of bare g-C₃N₄ (Fig. 10d).

Fig. 10

Metal and/or non-metal doping for improving the charge separation. (a) XRD patterns of pristine PCN and K⁺-incorporated PCN. (b) PL emission spectra (λ = 400 nm) of bare carbon nitride (CN), KPD-CN-3, and KPD-CN-7.5. (c) Apparent quantum yield (AQY, left axis) and turnover frequency (right axis) of bare CN and those of KPD-CN-7.5 for H₂O₂ production with monochromatic light irradiation. (d) Time profiles of photocatalytic H₂O₂ generation using CN samples modified with various dopants and a control sample of a physical mixture of CN and KPF₆ (CN + KPF). The experimental conditions were 0.5 g L⁻¹ photocatalyst, 10 vol% ethanol, pH 3, continuously purged with O₂ and polychromatic light through a long pass filter (k > 420 nm). Charge distributions of (e) K-GCN and (f) KCl-GCN obtained from density functional theory (DFT) calculations. Reprinted with permission from Ref. (Qiu C. et al., 2019; Moon G.-h. et al., 2017; Kim S. et al., 2018; Zhang P. et al., 2019). Copyright: (2018) The Authors (Qiu C. et al.), published by John Wiley & Sons, (2017) American Chemical Society, (2018) Elsevier and (2019) Springer Nature Ltd.

Zhang et al. used density functional theory (DFT) calculation to reveal how the incorporated alkaline ions change the charge distribution over different atoms (Zhang P. et al., 2019). The introduction of incorporated potassium species resulted in an accumulation on the first PCN layer. With the further assistance of Cl atoms, the electron distribution between the layers was more balanced (Fig. 10e and 10f), thus leading to a significantly improved interlayer charge transfer (Zhang P. et al., 2019).

Non-metal doping of PCN also enables manipulation of the electronic configuration of PCN with significantly improved charge separation (Zhang P. et al., 2020). Wei et al. developed an oxygen-enriched g-C₃N₄ (OCN) photocatalyst that achieved enhanced H₂O₂ production (Wei Z. et al., 2018). Compared to g-C₃N₄, the OCN samples achieved significantly enhanced no-sacrificial H₂O₂ production efficiency. Electrochemical impedance spectroscopy (EIS) and photoluminescence (PL) spectra all proved that charge separation efficiency was significantly promoted after the O element had been introduced into the PCN matrix.

3.1.2.2 Doping with polyoxometalates and their derivatives

Polyoxometalates (POMs) are composed of cations and polyanion clusters with high structural diversity (Han X.-B. et al., 2015). Under light irradiation, a charge transfer from O²⁻ to Mⁿ⁺ (n = 5, 6) usually occurred in these catalysts and lead to the formation of a hole center (O⁻) and trapped electron center (M⁽ⁿ⁻¹⁾⁺) pair. POMs in the PCN matrix can serve as electron donor-acceptor pairs, which could significantly promote the charge separation or/and serve as co-catalysts (Han X.-B. et al., 2015). Zhao and coworkers synthesized a series of POMs clusters-modified g-C₃N₄ samples for efficient photocatalytic H₂O₂ evolution by the thermal condemnation method (Zhao S. et al., 2017; Zhao S. and Zhao X., 2019; Zhao S. et al., 2018b). The amount of H₂O₂ formed by g-C₃N₄/PW₁₁ reached 3.5 μmol within 60 min, while the catalytic performance of pure g-C₃N₄ was only 1.3 μmol (Zhao S. et al., 2017). That research group also used other POM clusters, including [SiW₁₁O₃₉]⁸⁻ (SiW₁₁), (NH₄)₃PW₁₂O₄₀ (NH₄-PW₁₂) and (NH₄)₈Co₂W₁₂O₄₂ (NH₄-Co₂W₁₂), for covalent combination with g-C₃N₄ (Zhao S. and Zhao X., 2019). All of the clusters significantly improved the charge separation of pristine PCN. Action spectra measurements revealed that the band excitation results in the generation of H₂O₂ in both g-C₃N₄ and g-C₃N₄-CoWO systems (Han X.-B. et al., 2015; Zhao S. et al., 2017; Zhao S. and Zhao X., 2019), indicating that the function of POM incorporation may be as the same as that of ion doping since POM clusters are all well dispersed in the PCN matrix. The inner-panel charge transfer and the charge transfer between the layers may be significantly promoted by the introduction of highly dispersed POM clusters.

3.1.3 Defect engineering

Recently, it was found that defects in semiconductors enhance photocatalytic activity if they are precisely controlled (Pei Z. et al., 2013; Xiong J. et al., 2018; Meng A. et al., 2020). Introduction of carbon vacancies (Li S. et al., 2016; Goclon J. and Winkler K., 2018) and introduction of nitrogen vacancies are usually defect engineering used in a g-C₃N₄ photocatalyst (Li X. et al., 2018b; Qu X. et al., 2018; Shi L. et al., 2018; Zhu Z. et al., 2018). These strategies can change the electronic configuration and reactant molecules to promote H₂O₂ production. For example, Li et al. synthesized g-C₃N₄ functionalized carbon vacancies (Cv-g-C₃N₄) through thermal annealing in an oxygen-deficient environment (Ar flow) (Li S. et al., 2016). This Cv-g-C₃N₄ was also functionalized with a large number of amino groups with the formation of carbon vacancies (Li S. et al., 2016). Under visible light irradiation (Kessler F.K. et al., 2017), the H₂O₂ yield of Cv-g-C₃N₄ was much higher than that of pure g-C₃N₄ (Li S. et al., 2016). The improved photocatalytic activity of Cv-g-C₃N₄ for H₂O₂ production can be ascribed to the promoted charge separation and alteration of the H₂O₂ production pathway from a stepwise 1e⁻ to 1e⁻ pathway to a 2e⁻ pathway.

Nitrogen vacancies in g-C₃N₄ can also boost the photocatalytic activity for H₂O₂ production (Zhu Z. et al., 2018). Li et al. (Li X. et al., 2018b) and Qu et al. (Qu X. et al., 2018) showed that nitrogen vacancies can be introduced in the matrix of g-C₃N₄ by dielectric barrier discharge (DBD) plasma treatment. Their results showed that H₂O₂ evolution can be improved by up to 10 times compared with that produced by pristine PCN (Li X. et al., 2018b). In another related study, Shi et al. created a holey defective C₃N₄ photocatalyst with nitrogen vacancies via a hydrazine photoreduction method. The transfer of charge carriers from the bulk to the surface was significantly improved, as confirmed by PL measurements (Shi L. et al., 2018). The introduced nitrogen vacancies narrowed the bandgap and the formation of defect states within the bandgap together with a notably suppressed electron-hole recombination. Therefore, the holey defective g-C₃N₄ photocatalyst showed much higher photocatalytic activity for visible light-driven H₂O₂ production than that of pristine g-C₃N₄.

In summary, the functionalization strategies of doping and defect engineering for improving the charge separation for PCN-based materials used for the 2e⁻ ORR process are similar to the functionalization strategies for other reactions such as HER, COR, and ORR. However, the hybridization of PCN-based materials, especially hybridization with electron-deficient semiconductors, provides a PCN-based photocatalyst with unique electronic properties for improving the charge separation. We believe that copolymerization strategies to introduce other electron-deficient groups with higher charge separation efficiency might enable the generation of π-conjugated and π-stacked donor–acceptor couples (Shiraishi Y. et al., 2019) in the PCN matrix, leading to a higher efficiency for photocatalytic H₂O₂ synthesis.

3.2 Improving light harvesting

Although pristine PCN can use visible light for a photocatalytic reaction, the large band gap (2.7 eV) results in insufficient visible light harvesting for further efficiency improvement (Ong W.-J. et al., 2016). Thus, improving the light absorbance of a g-C₃N₄-based photocatalyst and narrowing the band gap are two practical strategies for H₂O₂ production.

Copolymerization of PCN with other organic molecules/semiconductors, such as triazine, barbituric acid and aromatic diimides, not only improves the light absorbance of the photocatalyst but also narrows the bandgap. As mentioned in Section 3.1.1.3, Shiraishi and co-workers prepared a series of PCN samples hybridized with atomic diimides with a significantly narrowed band gap from 2.7 eV of pristine PCN to 2.4 eV of PCN/PDI (Shiraishi Y. et al., 2014a; Kofuji Y. et al., 2017; Kofuji Y. et al., 2016a). The light absorbance of PCN samples hybridized with aromatic diimides improved to about ~20 % at λ = 400 nm, indicating improved light absorption. Additionally, Ohno and co-workers co-polymerized melem with barbituric acid to expand the light absorption edge from 445 nm (2.76 eV) of pristine PCN to ~550 nm (2.18 eV) of PCNBA0.5 (Fig. 11a) (PCNBA is an abbreviation for polymetric carbon nitride doped by barbituric acid, 0.5 means calcination of 3 g melamine and 0.5 g barbituric acid) (Teng Z. et al., 2020). The absorbance of PCNBA0.2 at 420 nm was also increased by 200 % compared with that of pristine PCN (Fig. 11b) (Teng Z. et al., 2020).

Fig. 11

Typical strategies for improving light harvesting. (a) Photographs show the corresponding powders of PCNBA samples. (b) UV-vis spectra of PCN and PCNBA samples: absorbance spectra. (c) Diffuse reflectance UV-vis spectra (DR-UVS) of modified CN samples prepared with various KPF₆ contents. (d) UV-vis spectra and apparent quantum yield (AQY) of H₂O₂ production as a function of irradiation wavelength with 10 % EtOH (v/v) under visible light irradiation (λ ≥ 420 nm) and T = 25 °C. Reprinted with permission from Ref. (Teng Z. et al., 2020; Zhang P. 2019). Copyright: (2019) Springer Nature Ltd. and (2020) Elsevier.

Doping of alkaline metal ions can also expand the light absorption to a small extent (0.1 eV~0.3 eV) combined with significantly light improved absorbance (50 %). Alkaline metal doping and non-metal doping were also proved to be efficient strategies for improving light absorbance (Moon G.-h. et al., 2017; Kim S. et al., 2018; Zhang P. et al., 2020). Choi and co-workers found that the light absorption ability of CN samples was gradually promoted by increasing the amount of KPF₆, even though KPF₆ alone did not exhibit any visible-light absorption (Moon G.-h. et al., 2017) (Fig. 11c). A very recent work using K and S-co-doped PCN with functionalization of the –OH group (AKMT) also showed significant improvement in both light absorption and narrowing of the bandgap (Zhang P. et al., 2020). The light absorption edge expanded from ~450 nm of pristine PCN to ~520 nm of AKMT (Zhang P. et al., 2020) (Fig. 11d). With the addition of a sacrificial reagent, AKMT showed apparent quantum yields (AQYs) of 76 % and 40 % at 480 nm and 500 nm for H₂O₂ production, respectively. By comparison, pristine PCN exhibited negligible activities at the same wavelengths.

Defect engineering can also slightly improve the light absorption of photocatalysts, especially for semiconductor-based materials (Li S. et al., 2016) and Zhu et al. (Zhu Z. et al., 2018) reported that the UV-vis absorption spectra of g-C₃N₄ and carbon vacancy-functionalized g-C₃N₄ showed a red shift of absorption edges compared with that of pristine ones. Nitrogen vacancies also significantly improve light absorption (Li X. et al., 2018b; Qu X. et al., 2018; Shi L. et al., 2018). Ye and co-workers compared the optical properties of bulk PCN (BCN) and those of N-defected CN samples (DCN) (Shi L. et al., 2018). The absorption edges of the samples showed red shifts of ~40 nm compared with that of pristine g-C₃N₄ samples without thermal treatment (Shi L. et al., 2018). They also revealed that the electrons could be accommodated by forming defect states in BCN, which greatly promoted light harvesting (Shi L. et al., 2018). Compared with the copolymerization of an electron-deficient semiconductor and metal/non-metal doping, the influence of defect engineering is relatively small because defect states are usually located just below the CBM. In this case, the extent of narrowing of the bandgap by introducing defects is limited. Copolymerization of electron-deficient semiconductors seems to be the most promising functionalization strategy for both improving light absorbance and expanding the absorption edge.

3.3 Introduction of active sites for selective 2e⁻ ORR

Introduction of active sites with specific physicochemical properties can significantly improve both selectivity and activities (Bo Y. et al., 2020). To give a systematic review for this topic, three dominant strategies for introducing active sites for photocatalytic H₂O₂ production are presents as follows.

3.3.1 Increasing the surface area

PCN prepared by the thermal polymerization method is usually quite small (<15 m² g⁻¹) (Teng Z. et al., 2017). As previously mentioned, the conjugated heptazine ring in the PCN matrix can provide active sites for 2e⁻ ORR (Shiraishi Y. et al., 2014a). A plausible strategy for increasing the number of active sites for 2e⁻ ORR is to increase the surface area (Mohamed N.A. et al., 2019; Ou H. et al., 2017; Liao G. et al., 2019). However, there is still some controversy about whether the overall 2e⁻ ORR rate can be improved by increasing the specific surface area of PCN. Shiraishi and co-workers investigated the photocatalytic production of H₂O₂ with the existence of O₂ and Et-OH. A series of mesoporous g-C₃N₄ catalysts having different specific surface areas (SSA) were prepared by a hard template (silica nanoparticles with different diameters) method (Fig. 12a) (Shiraishi Y. et al., 2015). In these systems, photogenerated holes oxidize Et-OH and the conduction band electrons localized at the melem unit reduce O₂ to form H₂O₂. This g-C₃N₄ catalysts exhibit significant higher SSA of 56 and 160 m² g⁻¹ compared with that of pristine g-C₃N₄ (10 m² g⁻¹) and higher activity for H₂O₂ production (Shiraishi Y. et al., 2015). However, g-C₃N₄ with the largest SSA (228 m² g⁻¹) showed significantly decreased activity and selectivity for H₂O₂ formation (Fig. 12b) (Shiraishi Y. et al., 2015). Mesoporous g-C₃N₄ with the largest surface area inherently contains the largest number of primary amine moieties (Shiraishi Y. et al., 2015). The H₂O₂ selectivity was significantly suppressed since these defects can also serve as active sites for 4e⁻ reduction of O₂ (Fig. 12c) (Shiraishi Y. et al., 2015). These defects also further decomposed the formed H₂O₂, leading to significantly suppressed activity. Results of density function theory (DFT) calculations confirmed (i) a large distribution of lowest unoccupied molecular orbital (LUMO) electrons onto the primary amine moieties and (ii) a decrease in H₂O₂ selectivity with an increase in the amount of primary amine moieties, strongly suggesting that the primary amine moieties on the surfaces of mesopores behave as the active sites for four-electron reduction (Fig. 12d–f) (Shiraishi Y. et al., 2015). However, the specific types of those defects were not described in their reports. Specification of the types and fine structures of defects is important to further improvement by defect engineering strategies.

Fig. 12

Influence of surface defects on H₂O₂ production by using a PCN-based material. (a) Different N atoms of PCN. (b) Time-dependent change in the amounts of H₂O₂ during photoreaction on the respective GCN(x) catalysts. (c) Electron transfer numbers during the photocatalytic H₂O₂ production. Interfacial plots of the main orbitals for (d) single, (e) double, and (f) triple melem-conjugated models, calculated by the function of B3LYP at the 6–31G (d) level. Reprinted with permission from Ref. (Shiraishi Y. et al., 2015). Copyright: (2015) American Chemical Society.

3.3.2 Copolymerization with electron-deficient semiconductors

Copolymerization with electron-deficient semiconductors such as PDI, BDI and MTI has also been proved to be one of the most efficient strategies for promoting the selectivity of a photocatalyst (Shiraishi Y. et al., 2014a; Kofuji Y. et al., 2017; Kofuji Y. et al., 2016a). Calculations based on the density functional theory (DFT) were investigated to clarify the influence of the PDI unit on the electronic structure of the g-C₃N₄ framework (Shiraishi Y. et al., 2014a). Time-dependent results showed that the main transitions are S₀→S₁ (highest occupied molecular orbitals, HOMO→LUMO and HOMO→LUMO+2). The iso-surfaces showed that electron distributions of the melem-PDI model (main transitions HOMO→LUMO+2) are located mainly at the melem units with partial distribution to the PDI units (Shiraishi Y. et al., 2014a). The high electron affinity of the PDI units in g-C₃N₄/PDI may significantly improve the polarization of the whole frameworks.

Being specific, the electrons on HOMO are concentrated at the N2 and N6 atom positions of the melem unit, and those on LUMO+2 are located at the C1 and N4 atoms (Shiraishi Y. et al., 2014a). The Raman spectra of g-C₃N₄/PDI and g-C₃N₄ confirmed that the side-on adsorption of O₂ was significantly promoted on g-C₃N₄/PDI, indicating a crucial role of copolymerization of electron-deficient semiconductors.

3.3.3 Doping/Heteroatom incorporation

Incorporation of metal species or no-metal species in the PCN matrix can both improve the charge separation and introduce active sites for specific reactions (Hu S. et al., 2018; Zhang Z. et al., 2022). By the template-assisted method, Hu and co-workers fabricated hollow copper-doped g-C₃N₄ microspheres. The g-C₃N₄ with copper incorporation showed a much higher H₂O₂ production ability (~4.8 mmol·L⁻¹) than that of pristine g-C₃N₄ (0.45 mmol·L⁻¹). DFT simulations showed that the Cu(I)-N sites work as reactive sites to activate molecular O₂ (Hu S. et al., 2018). The OCN proposed by Zhu and co-workers was proven to show a high selectivity for 2e⁻ ORR for photocatalytic H₂O₂ production, which is attributed to the functionalized C-O-C groups on OCNs (Wei Z. et al., 2018). Choi and co-workers revealed by DFT calculations that the K incorporation could result in improved adsorption for O₂ molecules (Zhang P. et al., 2019) (Fig. 13a, b). They also investigated whether electrons on S-doped g-C₃N₄ can be donated to the antibonding π* orbital of adsorbed *O₂ (with –0.40 |e|) for forming efficient ORR sites, as well as AQ doping as cocatalysts (Fig. 13c). The strong electron pushing effect between carbon and oxygen facilitates subsequent protonation in ORR kinetics (Chu C. et al., 2020). Both the activity and selectivity of ORR can be changed by manipulating the properties of metallic sites (Kulkarni A. et al., 2018). As elaborated in several electrochemical catalysts, the end-on O₂ adsorption usually lead to highly selective 2e⁻ ORR. It is difficult to prevent the splitting of O-O bond breaking on surface of metal particles since end-on O₂ molecular adsorption and side-on O₂ molecular adsorption occur on metal particles (Choi C.H. et al., 2014; Chu C. et al., 2019). Benefiting from the desirable features of a single atom catalyst (SAC) (Sun H. et al., 2021), the adsorption of O₂ molecules on atomically isolated sites is often end-on type, which could reduce the possibility of O-O bond splitting (Kulkarni A. et al., 2018; Wang A. et al., 2018). For example, SACs with Pt²⁺ (Shen R. et al., 2019) and Co-N₄ (Gao J. et al., 2020; Jung E. et al., 2020) centers could electrochemically reduce O₂ to H₂O₂ via a 2e⁻ ORR pathway with ultrahigh selectivity (>96 %). However, it is difficult for Pt²⁺ and Co-N₄ sites to be coupled in the photocatalytic system due to their high charge recombination characteristics, which originate from the intermediate band formed by the half-filled d electrons (Teng Z. et al., 2021a). Very recently, Ohno and co-workers developed an Sb single atom photocatalyst (Sb-SAPC) for non-sacrificial photocatalytic H₂O₂ synthesis in a water and oxygen mixture under visible light irradiation (Teng Z. et al., 2021a). The Sb-SAPC was prepared by a wet chemical method using NaSbF₆ and melamine as precursors. As shown in the image acquired from high-angle annular dark field scanning TEM (HAADF-STEM) measurement, the bright spots with high density are uniformly dispersed in the entire carbon nitride matrix (Fig. 14a). Electron energy loss spectroscopy (EELS) (Fig. 14b) measurement revealed bright spots corresponding to Sb atoms, in which the oxidation state of Sb is regulated to +3 with a 4d¹⁰5s² electron configuration (Fig. 14c, d). The results of experimental and theoretical investigations indicated that the adsorption of O₂ on isolated Sb atomic sites is end-on type, which promotes the formation of Sb-μ-peroxide (Sb-OOH), leading to an efficient 2e⁻ ORR pathway for H₂O₂ production (Fig. 14e). The 2e⁻ ORR efficiency of Sb species improved by almost one magnitude in the presence of 2-propanol (Fig. 14f). The apparent charge recombination of Sb-SAPC was significantly suppressed by introducing Sb species into the PCN matrix (Fig. 14g, h). With highly concentrated holes, this catalyst also achieved non-sacrificial H₂O₂ production using water as an electronic donor.

Fig. 13

Typical strategies for introducing active sites for efficient 2e⁻ ORR. (a) An enlarged top view of KCl-OH-GCN, and (b) an enlarged top view of pristine GCN. The yellow and blue colors represent electron accumulation and depletion at an isosurface value of 0.002 Å e⁻³. The brown, gray, and violet colors represent carbon, nitrogen, and potassuim, respectively. (c) Spatial separation of Co single atom (as an oxidation center) and AQ (as a reduction center) cocatalysts by anchoring them in the center (i.e., pyridinic N) and on the edge (i.e., primary/secondary amine N) of 2D ultrathin C₃N₄. Reprinted with permission from Ref. (Zhang P. et al., 2019; Chu C. et al., 2020). Copyright: (2019) Springer Nature Ltd. and (2020) The Authors (Chu C. et al.), published by National Academy of Science.

Fig. 14

Single atomic sites for efficient ORR. (a) High-magnification HAADF-STEM image of Sb-SAPC15. The insert shows size distribution of the bright spots. (b) Electron energy-loss spectroscopy (EELS) spectrum of Sb-SAPC15. (c) Sb-K edge X-ray absorption near edge structure (XANES) of Sb foil, Sb₂O₅ and Sb-SAPC15. (d) Fitting of the extended X-ray absorption fine structure (EXAFS) data for Sb-SAPC15 based on the model obtained from DFT optimization. Inserted figures: optimized molecular models based on DFT for EXAFS fitting. (e) Experimental Raman spectra recorded during the photoreaction in a 2-propanol aqueous solution with saturated oxygen. Spectra a, b, c and d: PCN, Sb-SAPC1, Sb-SAPC5 and Sb-SAPC15 in 10 % (v/v) 2-propanol aqueous solutions. Spectrum e: Sb-SAPC15 in pure water. (f) Comparison of H₂O₂ production in 10 % (v/v) 2-propanol aqueous solutions catalyzed by pristine PCN, PCN_Na15 and Sb-SAPC15. (g) Photoluminescence spectra of PCN, PCN_Na15 and Sb-SAPC15 at an excitation wavelength of 380 nm (h) Electrochemical impedance spectroscopy (EIS) spectra (Nyquist plots) of pristine PCN and Sb-SAPC15 in the frequency range from 100 kHz to 0.01 Hz at 0.6 V (vs. Ag/AgCl) under visible light irradiation. Reprinted with permission from Ref. (Teng Z. et al., 2021b). Copyright: (2021) Springer Nature Ltd.

3.3.4 Cocatalyst loading

Cocatalyst loading is a traditional strategy for introducing active sites for specific reactions (Nosaka Y. and Nosaka A, 2016). For instance, pristine PCN shows quite low efficiency in photocatalytic hydrogen production, and loading of platinum species on PCN usually significantly improves the HER performance (Wang Q. and Domen K., 2020). The reason for loading a cocatalyst to boost the photocatalytic reactions is usually the decreased overpotential of the reactions on the co-catalyst surface compared with that on pristine photocatalysts (Hirakawa H. et al., 2016; Tsukamoto D. et al., 2012; Peng Y. et al., 2017). AQ species can enhance the selectivity of O₂ reduction to H₂O₂ based on the mechanism of the current industrial H₂O₂ production process (Kim H.-i. et al., 2018; Chu C. et al., 2020). Kim and co-workers prepared several AQ-modified PCN samples with significantly improved selectivity for 2e⁻ ORR (Chu C. et al., 2020). Loading an AQ cocatalyst onto ultrathin g-C₃N₄ significantly improved the selectivity of H₂O₂ production, combined with a 1.9-fold enhancement of the amount of H₂O₂ production (Chu C. et al., 2020). Several works in which Au nanoparticles were loaded onto the C₃N₄ matrix all showed significantly improved 2e⁻ ORR efficiency (Chang X. et al., 2018; Zuo G. et al., 2019). This phenomenon may be attributed to the cocatalytic effects of the Au surface on the photocatalytic 2e⁻ ORR, which has already been proved by investigating other photocatalysts and photocathodes (Hirakawa H. et al., 2016).

In summary, there are two fundamental strategies for introducing active sites for the improvement of 2e⁻ ORR of a PCN-based photocatalyst. One is to manipulate the electronic configurations of the functionalized PCN catalyst so that the selectivity and activity of the π-conjugated heptazine ring for boosting H₂O₂ production can be significantly enhanced. Similar strategies have also been used in some other carbon-based materials for promoting selective H₂O₂ production. On the contrary, the nature of melem sites is side-on adsorption of oxygen (Shiraishi Y. et al., 2015), which might result in unexpected splitting of O-O bonds, thus leading to the possible 4e⁻ ORR process and a limited selectivity for 2e⁻ ORR (Siahrostami S. et al., 2013; Verdaguer-Casadevall A. et al., 2014). The other strategy is to introduce active sites beyond the π-conjugated heptazine ring for selective 2e⁻ ORR. Up to now, only a few cocatalysts have been developed for selective ORR (Section 3.3.4). Further development of 2e⁻ ORR sites beyond the π-conjugated heptazine rings may overcome the disadvantage of side-on O₂ adsorption and shed light on further improvement for H₂O₂ production.

4. Other factors for boosting photocatalytic efficiency for H₂O₂ production

4.1 Adding electron donors or not

Sacrificial reagents are commonly used to consume the photogenerated electrons or holes so that half of the reaction can be significantly boosted (Schneider J. and Bahnemann D.W., 2013). In the case of g-C₃N₄-based photocatalysts, most researchers used electron donors so that reduction reactions could be maximized in a certain photocatalytic system since an electron donor prevents the photogenerated electrons from recombining with holes (Wang Y. et al., 2019; Wang M. et al., 2017). g-C₃N₄ has shown promising prospects for catalyzing several photochemical reduction reactions with the existence of electron donors. The reason may be as follows: the CBM of pristine PCN (–1.3 V) is quite negative, thus resulting in significantly larger redox potentials of hydrogen evolution reaction (HER), nitrogen reduction reaction (NRR), carbon dioxide reduction reaction (CRR) and 2e⁻ ORR (0.695 V vs. SHE) (Wang Y. et al., 2019). Combined with the merits of the side-on O₂ adsorption on melem units (forming 1–4 endoperoxide species) as acceptable 2e⁻ reduction sites, PCN satisfied the thermodynamic potential for 2e⁻ ORR with the existence of alcohol. However, a green and sustainable process for photocatalytic H₂O₂ synthesis requires the use of earth-abundant water as an electron donor, i.e., water oxidation reaction instead of alcohol oxidation should occur for hole consumption.

There have been several studies in which mechanisms for photoreduction reactions were systematically investigated by using g-C₃N₄ (Wang Y. et al., 2019). Since the functions of electron donors were clarified in those works, we will focus on the crucial properties and functionalization strategies for achieving H₂O₂ production with water and oxygen (Fig. 15a). Typically, water oxidation via a 4e⁻ pathway is usually recognized as an oxygen evolution reaction (OER) that is usually needed to overcome the large overpotential. VBMs of photocatalysts for OER usually have quite positive potentials (VBM_TiO2-rutile ~2.9 V vs. SHE, VBM_BiVO4 ~2.53 V) (Fan W. et al., 2020; Hirakawa H. et al., 2016). As shown in Fig. 15b, pristine PCN is unable to generate H₂O₂ in the presence of water and O₂ because the valence band maximum (VBM) lies at approximately 1.4 V, which shows an insufficient thermodynamic driving force for OER (ca. 0.8 V versus RHE, pH 7) (Shiraishi Y. et al., 2014a). To overcome this obstacle, band engineering and the use of a cocatalyst are two effective strategies.

Fig. 15

Typical strategies for preparation of a g-C₃N₄-based catalyst for photocatalytic H₂O₂ production with water and O₂. (a) Real hydrogen electrode potentials of 2e⁻ ORR and 4e⁻ WOR compared with the band diagram of PCN. (b) Electronic band structures of pristine PCN and g-C₃N₄/BDI_x. (c) Changes in the amounts of H₂O₂ formed during photoreaction with respective catalysts. For the reaction in (c), 100 mg of catalyst was used. The irradiance at 420–500 nm is 27.3 W m⁻². Blue points show the results of photoreaction with water containing ca. 100 μmol of H₂O₂. (d) Calculated HOMO and LUMO of Melem_4 model (representing PCN), Melem_4BA1 (representing PCNBA with a low motif concentration) and Melem_4BA2 (representing PCNBA with a high motif concentration). (e) Schematic diagram of PCNBA_xCo_x for accelerating water oxidation. (f) Absorption spectrum of PCNBA0.2 and action spectra for H₂O₂ formation on the respective catalysts. (g) Time course of O₂ evolution measured under Ar pressure of 0.6-kPa and 300-W xenon lamp irradiation with 0.5 g L⁻¹ of the catalyst, 1 g L⁻¹ La₂O₃, and 20 mM AgNO₃ in 100 mL of water. Reprinted with permission from Ref. (Kofuji Y., et al., 2016b; Teng Z. et al., 2020). Copyright: (2016) American Chemical Society and (2020) Elsevier.

The bandgap position of PCN can be easily manipulated by functionalization. Shiraishi and co-workers successfully manipulate the band positions of a PCN-based photocatalyst by copolymerized melamine with electron-deficient aromatic diimide (Shiraishi Y. et al., 2014a; Shiraishi Y. et al., 2014b; Kofuji Y. et al., 2016a; Kofuji Y. et al., 2017; Kofuji Y. et al., 2016b). The valence band of the as-prepared co-polymerized PCN is about 1.8~2.4 V (vs. NHE). Taking g-C₃N₄/PDI as an example, the activity results revealed a significant role of band engineering. Pristine g-C₃N₄ (VBM = 1.40 V vs. NHE) barely produces H₂O₂, and g-C₃N₄/BDI51 (VBM = 1.86 V vs. NHE) produces 31 μmol of H₂O₂ after photocatalytic reaction for 48 h (Kofuji Y. et al., 2016b) (Fig. 15c). g-C₃N₄/BDI50 (VBM = 1.83 V vs. NHE) shows a higher performance (41 μmol) (Kofuji Y. et al., 2016b). With the addition of electron acceptors, O₂ can be detected by GC, which further confirmed that water oxidation to generate oxygen consumed the generated holes during the photocatalytic reactions (Kofuji Y. et al., 2016b). Introduction of boron nitride (BN) can further improve the activity of the g-C₃N₄/PDI-rGO photocatalyst because of the enhanced charge separation efficiency (Kofuji Y. et al., 2018). Following those works, Ohno and co-workers further revealed by analysis of the density of states that the copolymerization of electron-deficient units results in the introduction of O 2p states (Teng Z. et al., 2020). The introduced O 2p states compose the valence band with the states of N 2p and C 2p, resulting in a more positive potential than that of pristine PCN for which the VB is only composed of the states of N 2p and C 2p (Fig. 15d) (Teng Z. et al., 2020). Thus, the OER activity is significantly promoted (Fig. 15e). Additionally, preparation of POM-incorporated PCN samples with ideal electronic configurations and band positions also played similar roles (shifting the VBM to a more oxidative potential) for boosting the OER, which successfully achieved the H₂O₂ production via ORR and WOR (Zhao S. et al., 2018b; Zhao S. et al., 2017; Zhao S. and Zhao X., 2019). Another strategy for improving the OER activity of PCN is loading a cocatalyst for OERs. Ohno and co-workers loaded Na₂CoP₂O₇ on the band-engineered PCNBA0.2 by a simple ball milling method (Fig. 15e) (Teng Z. et al., 2020). The time-dependent H₂O₂ production and action spectra confirmed the cocatalytic effects of Na₂CoP₂O₇ (Teng Z. et al., 2020). The action spectra showed that its bandgap excitation produces H₂O₂ (Teng Z. et al., 2020). The apparent quantum yield (AQY) of PCNBA with loading of 5 % Na₂CoP₂O₇ (PCNBA0.2Co5 %) at 420 nm (8.0 %) was two-times higher than that of PCNBA0.2 (4.0 %) (Fig. 15f) (Teng Z. et al., 2020). Additionally, no H₂O₂ was produced at the light-irradiation wavelengths between 550 nm and 650 nm, indicating that Na₂CoP₂O₇ works as a cocatalyst and does not work as a photocatalyst for production of H₂O₂ (Teng Z. et al., 2020). More recently, Chu and co-workers prepared single-atom dispersed Co as a co-catalyst for OER (Chu C. et al., 2020). X-ray absorption fine-structure spectroscopy (FT-EXAFS) showed that the oxidation state of the Co single atoms coordinated by P atoms is close to +2 (Chu C. et al., 2020). The OER activity of g-C₃N₄ was significantly increased (8.4-fold) by the introduction of Co single atomic sites (Fig. 15g). The chemical states of P-coordinated Co atoms are close to +2, indicating that the unique electronic configuration of Co(II)-P may provide active sites for OER during solar H₂O₂ synthesis (Chu C. et al., 2020). It is notable that the VBM of this single-Co dispersed PCN (~1.4 V vs. SHE) (Chu C. et al., 2020), quite close to that of PCN, is significantly more positive than that of metal oxides (>2.3 V vs. SHE). Thus, the Co(II)-P single atomic sites may significantly boost the kinetic process of water oxidation. Very recently, the Sb-SAPC proposed by Ohno and co-workers also showed good activity based on non-sacrificial photocatalytic H₂O₂ activities. Based on the results detailed characterizations and analyses, the following reaction mechanism (Fig. 16) of Sb-SAPC for photocatalytic H₂O₂ production is proposed. Firstly, efficient charge separation occurs on Sb-SAPC under visible light irradiation, resulting in the generation of photoexcited electrons and holes for ORR and WOR, respectively. Then water molecules are oxidized to evolve O₂ by photogenerated holes localized at the N atoms near the single Sb atoms. Simultaneously, O₂ dissolved in water and that generated from the WOR both participate in the ORR process to produce H₂O₂. It is notable that the efficient charge separation, ideal single atomic sites for end-on type O₂ adsorption and close spatial distribution of active sites boost both the 2e⁻ ORR and 4e⁻ WOR for efficient H₂O₂ production.

Fig. 16

Mechanism of photocatalytic H₂O₂ production. (White, gray, blue, red and magenta spheres indicate hydrogen, carbon, nitrogen, oxygen and Sb atoms, respectively.) After shining visible light, the photogenerated electrons are localized at the Sb sites (with a blue glow), and the photogenerated holes are localized at the N atoms at the melem units (with a red glow). Subsequently, the dissolved O₂ molecules are adsorbed (orange arrows) onto the Sb sites and then become reduced (blue arrows) via a 2e⁻ transfer pathway through the formation of an electron μ-peroxide as the intermediate. Simultaneously, water molecules are oxidized (pink arrows) to generate O₂ by the highly concentrated holes on the melem units. Reprinted with permission from Ref. (Teng Z. et al., 2021b). Copyright: (2021) Springer Nature Ltd.

In summary, accelerating the photocatalytic oxygen evolution reaction by using a PCN-based photocatalyst is the most crucial challenge for achieving photocatalytic H₂O₂ production with water and O₂. Band engineering and loading cocatalyst for OER may respectively improve the oxidation potential and promote the kinetics for 4e⁻ OER.

4.2 pH value

It is difficult for the light-driven 2e⁻ WOR pathway to be achieved because of the uphill thermodynamics (1.76 V vs. NHE), i.e., the as-synthesized H₂O₂ will decompose at this highly oxidative potential because H₂O₂ is an excellent hole scavenger (Fuku K. and Sayama K., 2016). Therefore, we will focus on the 2e⁻ ORR pathway. Substantial efforts have been devoted to the 2e⁻ ORR concept with the aim of achieving high-yield production of H₂O₂ in basic, neutral, and acidic media (Gao J. and Liu B., 2020; Kim H.W. et al., 2018; Lu Z. et al., 2018; Iglesias D. et al., 2018). The acid dissociation constant of H₂O₂ (H₂O₂ ↔ H⁺ + HO₂⁻, $K_a=\frac{[{\text{H}}^{+}][{\text{HO}}_2^{-}]}{[{\text{H}}_2{\text{O}}_2]}$, pK_a = 11.7) suggests that the main form of H₂O₂ would be HO₂⁻ in an alkaline water solution (pH >11.7). Alkaline H₂O₂ is widely used in the pulp bleaching industry. However, the production of H₂O₂ in basic media has the following two limitations: (1) H₂O₂ is less stable and can self-decompose in an alkaline condition (especially at pH >9), and it therefore needs to be consumed in a short time (Gao J. and Liu B., 2020; Qiang Z. et al., 2002) and (2) H₂O₂ is more widely used in natural or acidic media with stronger oxidation ability than in an alkaline solution. For instance, Fenton’s reagent, which is widely used in effluent treatment and organic synthesis, has a favorable pH range of 2.5–3.5 (Gao J. and Liu B., 2020; Kuo W.G. et al., 1992). Therefore, improving H₂O₂ production in acidic or neutral media by photocatalysts has attracted much attention for industrial applications. Over the past few years, although various g-C₃N₄ catalysts have been developed for H₂O₂ production via the 2e⁻ ORR process (Shiraishi Y. et al., 2014b), there have been very few works in which the optimized pH conditions were investigated.

Kim and co-workers found that photocatalytic activity for H₂O₂ production was higher at lower pH values (Fig. 17a) with the existence of an electron donor. This is probably due to proton conduction within the CN framework and is much more facilitated in KPF_CN. The proton conduction of KPF_CN was greatly enhanced when PF₆ ⁻ anions was introduced into the structure of the g-C₃N₄ framework (Kim S. et al., 2018). The higher charge carrier density in KPF_CN might be favorable for facilitating proton conduction. The higher proton conductivity within the KPF_CN framework promotes the production of H₂O₂, particularly when the reaction between O₂ and protons occurs within the interlayer space (Kim S. et al., 2018). Ohno and co-workers also investigated the optimal pH conditions by measuring the amount of H₂O₂ produced by the photocatalytic reaction using PCNBA0.2 (Fig. 17b) (Teng Z. et al., 2020). When ethanol was introduced into the system, the amount of produced H₂O₂ remained almost constant when the pH value was increased from 1 to 9 and then decreased drastically (Teng Z. et al., 2020). When the pH value was lower than 7, the concentration of protons had almost no effect on the pH value (Teng Z. et al., 2020). This phenomenon revealed that protonation plays a less essential role in the PCNBA system, indicating that incorporation of BA units further improved the proton conductivity compared with that of KPF_CN. They also found that the lowest [H⁺] is 10⁻⁹ mol L⁻¹ for O₂ reduction to H₂O₂ production. When no sacrificial agent was added to the system, the amount of H₂O₂ rapidly increased before the pH value reached 7 and then gradually decreased, indicating that a low concentration of OH⁻ also restricts the generation of O₂ from water (Wang Q. and Domen K., 2020). Therefore, the optimal pH condition for H₂O₂ production from O₂ and water using PCNBA is 7–9. The production of H₂O₂ significantly decreased in both of the works when pH was larger than 9, being consistent with the fact that H₂O₂ self-decomposes in alkaline conditions. Although there have been only few works in which the influence of pH on the amount of H₂O₂ generated was investigated, it was shown that the pH value did play an essential role during the photosynthesis process. Additionally, the pH value of the interfaces between materials and reaction solution might be even more important compared with the pH value of the whole solution. The rational design of Helmholtz layer by the manipulation of the surface properties are extremely important for improving the H₂O₂ production (Zeng X. et al., 2020). We believe that systematic investigations of the effects of pH on the amount of H₂O₂, surface properties and electronic configurations should be carried out to reveal the mechanisms of ORR and WOR during the photosynthesis of H₂O₂.

Fig. 17

Influence of pH values during H₂O₂ production by using a PCN-based photocatalyst. (a) H₂O₂ production in an irradiated suspension of bare CN and KPF_CN_15 at different pH values. (b) Optimization of the pH condition for H₂O₂ production when PCNBA0.2 was used as a photocatalyst. Reprinted with permission from (Kim S. et al., 2018) and (Teng Z. et al., 2020). Copyright: (2018, 2020) Elsevier.

5. Summary and perspectives

Photocatalytic H₂O₂ production by using a PCN-based material has shown a promising prospective in environmental and energy-related realms. We have summarized the principles for photocatalytic H₂O₂ production, especially for the photocatalytic H₂O₂ production using g-C₃N₄. Advantages and disadvantages for H₂O₂ production and the most favorable properties of PCN for photocatalytic H₂O₂ production were briefly summarized. Based on the intrinsic properties of pristine g-C₃N₄, the most urgent issue for overcoming the drawbacks was focused on. Notably, current PCN-based photocatalysts still suffer from relatively low H₂O₂ yield. Due to the low activity, g-C₃N₄-based photocatalysts are far from the requirements for industrial applications. Improving the charge separation, promoting the light absorption and introducing active sites for 2e⁻ ORR to suppress the side reaction are three attractive strategies for boosting the activities. Following discussion of these strategies, the most representative functionalization method was summarized on the basis of the most desired properties for improving the photocatalytic activities for H₂O₂ production. Other factors for improving H₂O₂ production such as addition of electron donors and pH value of the solution were also discussed.

There are several outlooks and suggestions for further improvement of photocatalysts. First of all, further improvement in the charge separation of the PCN is necessary. New species of elements or organic semiconductors for incorporation are highly recommended. Secondly, the absorption edge of the functionalized PCN can hardly be extended to >550 nm with a reasonable AQY for H₂O₂ production (Teng Z. et al., 2020). Further narrowing of band gap is therefore still of great importance. It is notable that the narrowing of the bandgap could reduce the thermodynamic driving force, leading to weakened capability for reduction reactions and oxidation reactions. Thus, reducing the overpotential for 2e⁻ ORR and oxidation reactions on the PCN surface is of great importance. Exploring new reaction sites with higher performance beyond the heptazine rings in the PCN matrix for 2e⁻ ORR efficiency is also of great importance for further improvement of photocatalytic activity and selectivity for H₂O₂ production. Recently, single-atom catalysts (SACs) have shown excellent selectivity for H₂O₂ production with Pt²⁺ (Shen R. et al., 2019; Gao J. et al., 2020; Jung E. et al., 2020) via a 2e⁻ ORR pathway. In the case of PCN, anchoring single-atom catalysts that have large amounts of nitrogen atoms with lone electron pairs is easy to perform. However, Pt²⁺ and Co-N₄ sites can hardly be utilized for photocatalytic systems due to poor charge separation, which origin form the half-filled d electrons of Pt²⁺ and Co-N₄ may result in a high bulk-recombination rate of PCN. The preparation of photocatalysts having atomically dispersed atoms with a d₀ or d₁₀ electronic configuration usually introduce intermediate bands with energetic levels that are slightly more negative than that of CBM, which can promote electron mobility, thus leading to a boosted performance (Inoue Y., 2009). It should be mentioned that measurement of H₂O₂ production was carried out in most studies with the addition of an organic sacrificial electron donor to consume the photogenerated holes and inhibit the recombination of photogenerated carries, which is energy-consuming and cost-inefficient. Despite of band engineering and cocatalyst loading, the construction of an all-solid-state Z-scheme photocatalytic system is another recommended strategy for achieving non-sacrificial H₂O₂ production with water and oxygen (Xu Q. et al., 2018). Additionally, it is necessary to develop approaches other than optimization of pH conditions to prevent decomposition of the generated H₂O₂. The stability of organic semiconductors during photocatalytic H₂O₂ production should also be investigated (Liu L. et al., 2021). Last not but least, attention should be given to precise determination of the amount of H₂O₂ generated because the species in the solution can influence the accuracy for determination (Wei Y. et al., 2021).

To optimize the design of functionalized PCN with maximized photocatalytic H₂O₂ production, a comprehensive understanding of the photocatalytic mechanism for H₂O₂ production is necessary. Both experimental and theoretical investigations should be simultaneously carried out. In situ measurements by Kelvin probe microscopy (Zhu J. et al., 2015) and X-ray absorption fine structure analysis (Yang H. et al., 2018) could be used to study the dynamic behavior of photogenerated carriers on the surface and the changes of oxidized states in a photocatalyst (Hou H. et al., 2020). In situ Raman spectroscopy may reveal the process from the adsorption of O₂ to dissociation of H₂O₂, thus leading to a more precise understanding of the mechanism. In the case of fundamental calculations and simulations for investigation of the mechanism, ground state properties such as density of states, adsorption energy, and the structural and electronic properties have been widely used for explanation of physicochemical properties for the design of an ideal photocatalyst for H₂O₂ production (Dong J.-C. et al., 2020). However, investigation of the excitation properties of PCN materials has not been sufficient to obtain a fundamental understanding of the excitation behavior of a photocatalyst. We suggest that excitation properties of the functionalized PCN, such as transition densities based on time-dependent DFT calculations (Laurent A.D. et al., 2013; Teng Z. et al., 2021b; Che H. et al., 2021), should be investigated in future studies in order to establish a satisfactory blueprint for designing a PCN catalyst for H₂O₂ production. Last but not least, since H₂O₂ production is a green oxidant for many chemical reactions, the development of highly value-added chemicals based on the photocatalytic systems is needed. Using the in-situ generated H₂O₂ for synthesis of chemicals could also be achieved on the basis of previous work using D₂O to produce deuterated chemicals (Zhang B. et al., 2021). We believe that better design strategies and guidelines for not only PCN-based materials but also other efficient photocatalysts can be established for H₂O₂ formation if the above-mentioned properties can be achieved.

Acknowledgements

The authors acknowledge the financial support of Mitsubishi Chemical Corporation, JSPS Grant-in-Aid for Scientific Research (B, No. 20H02847), and Grant-in-Aid for JSPS Fellows (DC2, 20J13064).

References

•  Banerjee  T.,  Podjaski  F.,  Kröger  J.,  Biswal  B.P.,  Lotsch  B.V., Polymer photocatalysts for solar-to-chemical energy conversion, Nature Reviews Materials, 6 (2021) 168–190. DOI:10.1038/s41578-020-00254-z
•  Bo  Y.,  Gao  C.,  Xiong  Y., Recent advances in engineering active sites for photocatalytic CO₂ reduction, Nanoscale, 12 (2020) 12196–12209. DOI:10.1039/D0NR02596H
•  Campos-Martin  J.M.,  Blanco-Brieva  G.,  Fierro  J.L.G., Hydrogen peroxide synthesis: an outlook beyond the anthraquinone process, Angewandte Chemie International Edition, 45 (2006) 6962–6984. DOI:10.1002/anie.200503779
•  Chang  X.,  Yang  J.,  Han  D.,  Zhang  B.,  Xiang  X.,  He  J., Enhancing light-driven production of hydrogen peroxide by anchoring Au onto C₃N₄ catalysts, Catalysts, 8 (2018) 147. DOI:10.3390/catal8040147
•  Che  H.,  Gao  X.,  Chen  J.,  Hou  J.,  Ao  Y.,  Wang  P., Iodide-induced fragmentation of polymerized hydrophilic carbon nitride for high-performance quasi-homogeneous photocatalytic H₂O₂ production, Angewandte Chemie International Edition, 60 (2021) 25546–25550. DOI:10.1002/anie.202111769
•  Choi  C.H.,  Kwon  H.C.,  Yook  S.,  Shin  H.,  Kim  H.,  Choi  M., Hydrogen peroxide synthesis via enhanced two-electron oxygen reduction pathway on carbon-coated Pt surface, The Journal of Physical Chemistry C, 118 (2014) 30063–30070. DOI:10.1021/jp5113894
•  Chu  C.,  Huang  D.,  Zhu  Q.,  Stavitski  E.,  Spies  J.A.,  Pan  Z.,  Mao  J.,  Xin  H.L.,  Schmuttenmaer  C.A.,  Hu  S.,  Kim  J.-H., Electronic tuning of metal nanoparticles for highly efficient photocatalytic hydrogen peroxide production, ACS Catalysis, 9 (2019) 626–631. DOI:10.1021/acscatal.8b03738
•  Chu  C.,  Zhu  Q.,  Pan  Z.,  Gupta  S.,  Huang  D.,  Du  Y.,  Weon  S.,  Wu  Y.,  Muhich  C.,  Stavitski  E.,  Domen  K.,  Kim  J.-H., Spatially separating redox centers on 2D carbon nitride with cobalt single atom for photocatalytic H₂O₂ production, Proceedings of the National Academy of Sciences of the United States of America, 117 (2020) 6376. DOI:10.1073/pnas.1913403117
•  Chung  D.D.L., Carbon Materials: Science and Applications, World Scientific Publishing Company, 2019, ISBN: 9789811200939.
•  Dong  J.-C.,  Su  M.,  Briega-Martos  V.,  Li  L.,  Le  J.-B.,  Radjenovic  P.,  Zhou  X.-S.,  Feliu  J.M.,  Tian  Z.-Q.,  Li  J.-F., Direct in situ Raman spectroscopic evidence of oxygen reduction reaction intermediates at high-index Pt(hkl) surfaces, Journal of the American Chemical Society, 142 (2020) 715–719. DOI:10.1021/jacs.9b12803
•  Fan  W.,  Zhang  B.,  Wang  X.,  Ma  W.,  Li  D.,  Wang  Z.,  Dupuis  M.,  Shi  J.,  Liao  S.,  Li  C., Efficient hydrogen peroxide synthesis by metal-free polyterthiophene via photoelectrocatalytic dioxygen reduction, Energy & Environmental Science, 13 (2020) 238–245. DOI:10.1039/C9EE02247C
•  Freakley  S.J.,  He  Q.,  Harrhy  J.H.,  Lu  L.,  Crole  D.A.,  Morgan  D.J.,  Ntainjua  E.N.,  Edwards  J.K.,  Carley  A.F.,  Borisevich  A.Y.,  Kiely  C.J.,  Hutchings  G.J., Palladium-tin catalysts for the direct synthesis of H₂O₂ with high selectivity, Science, 351 (2016) 965. DOI:10.1126/science.aad5705
•  Fuku  K.,  Sayama  K., Efficient oxidative hydrogen peroxide production and accumulation in photoelectrochemical water splitting using a tungsten trioxide/bismuth vanadate photoanode, Chemical Communications, 52 (2016) 5406– 5409. DOI:10.1039/C6CC01605G
•  Gao  J.,  Liu  B., Progress of electrochemical hydrogen peroxide synthesis over single atom catalysts, ACS Materials Letters, 2 (2020) 1008–1024. DOI:10.1021/acsmaterialslett.0c00189
•  Gao  J.,  Yang  H.b.,  Huang  X.,  Hung  S.-F.,  Cai  W.,  Jia  C.,  Miao  S.,  Chen  H.M.,  Yang  X.,  Huang  Y.,  Zhang  T.,  Liu  B., Enabling direct H₂O₂ production in acidic media through rational design of transition metal single atom catalyst, Chem, 6 (2020) 658–674. DOI:10.1016/j.chempr.2019.12.008
•  Goclon  J.,  Winkler  K., Computational insight into the mechanism of O₂ to H₂O₂ reduction on amino-groups-containing g-C₃N₄, Applied Surface Science, 462 (2018) 134–141. DOI:10.1016/j.apsusc.2018.08.070
•  Han  X.-B.,  Li  Y.-G.,  Zhang  Z.-M.,  Tan  H.-Q.,  Lu  Y.,  Wang  E.-B., Polyoxometalate-based nickel clusters as visible light-driven water oxidation catalysts, Journal of the American Chemical Society, 137 (2015) 5486–5493. DOI:10.1021/jacs.5b01329
•  Hirakawa  H.,  Shiota  S.,  Shiraishi  Y.,  Sakamoto  H.,  Ichikawa  S.,  Hirai  T., Au nanoparticles supported on BiVO₄: effective inorganic photocatalysts for H₂O₂ production from water and O₂ under visible light, ACS Catalysis, 6 (2016) 4976–4982. DOI:10.1021/acscatal.6b01187
•  Hisatomi  T.,  Domen  K., Reaction systems for solar hydrogen production via water splitting with particulate semiconductor photocatalysts, Nature Catalysis, 2 (2019) 387–399. DOI:10.1038/s41929-019-0242-6
•  Hou  H.,  Zeng  X.,  Zhang  X., Production of hydrogen peroxide by photocatalytic processes, Angewandte Chemie International Edition, 59 (2020) 17356–17376. DOI:10.1002/anie.201911609
•  Hu  S.,  Qu  X.,  Li  P.,  Wang  F.,  Li  Q.,  Song  L.,  Zhao  Y.,  Kang  X., Photocatalytic oxygen reduction to hydrogen peroxide over copper doped graphitic carbon nitride hollow microsphere: the effect of Cu(I)-N active sites, Chemical Engineering Journal, 334 (2018) 410–418. DOI:10.1016/j.cej.2017.10.016
•  Iglesias  D.,  Giuliani  A.,  Melchionna  M.,  Marchesan  S.,  Criado  A.,  Nasi  L.,  Bevilacqua  M.,  Tavagnacco  C.,  Vizza  F.,  Prato  M.,  Fornasiero  P., N-doped graphitized carbon nanohorns as a forefront electrocatalyst in highly selective O₂ Reduction to H₂O₂, Chem, 4 (2018) 106–123. DOI:10.1016/j.chempr.2017.10.013
•  Inoue  Y., Photocatalytic water splitting by RuO₂-loaded metal oxides and nitrides with d₀- and d₁₀ -related electronic configurations, Energy & Environmental Science, 2 (2009) 364–386. DOI:10.1039/B816677N
•  Jiang  X.,  Wang  P.,  Zhao  J., 2D covalent triazine framework: a new class of organic photocatalyst for water splitting, Journal of Materials Chemistry A, 3 (2015) 7750–7758. DOI:10.1039/C4TA03438D
•  Jones  C.W., Applications of hydrogen peroxide and derivatives, Royal Society of Chemistry, 1999, ISSN: 978-0-85404-536-5
•  Jung  E.,  Shin  H.,  Lee  B.-H.,  Efremov  V.,  Lee  S.,  Lee  H.,  Kim  J.,  Antink  W.,  Park  S.,  Lee  K.-S.,  Cho  S.-P.,  Yoo  J.S.,  Sung  Y.-E.,  Hyeon  T., Atomic-level tuning of Co–N–C catalyst for high-performance electrochemical H₂O₂ production, Nature Materials, 19 (2020) 1–7. DOI:10.1038/s41563-019-0571-5
•  Kessler  F.K.,  Zheng  Y.,  Schwarz  D.,  Merschjann  C.,  Schnick  W.,  Wang  X.,  Bojdys  M.J., Functional carbon nitride materials—design strategies for electrochemical devices, Nature Reviews Materials, 2 (2017) 17030. DOI:10.1038/natrevmats.2017.30
•  Kim  D.,  Sakimoto  K.K.,  Hong  D.,  Yang  P., Artificial photosynthesis for sustainable fuel and chemical production, Angewandte Chemie International Edition, 54 (2015) 3259–3266. DOI:10.1002/anie.201409116
•  Kim  H.-i.,  Choi  Y.,  Hu  S.,  Choi  W.,  Kim  J.-H., Photocatalytic hydrogen peroxide production by anthraquinone-augmented polymeric carbon nitride, Applied Catalysis B: Environmental, 229 (2018) 121–129. DOI:10.1016/j.apcatb.2018.01.060
•  Kim  H.W.,  Ross  M.B.,  Kornienko  N.,  Zhang  L.,  Guo  J.,  Yang  P.,  McCloskey  B.D., Efficient hydrogen peroxide generation using reduced graphene oxide-based oxygen reduction electrocatalysts, Nature Catalysis, 1 (2018) 282–290. DOI:10.1038/s41929-018-0044-2
•  Kim  S.,  Moon  G.-h.,  Kim  H.,  Mun  Y.,  Zhang  P.,  Lee  J.,  Choi  W., Selective charge transfer to dioxygen on KPF₆-modified carbon nitride for photocatalytic synthesis of H₂O₂ under visible light, Journal of Catalysis, 357 (2018) 51–58. DOI:10.1016/j.jcat.2017.10.002
•  Koe  W.S.,  Lee  J.W.,  Chong  W.C.,  Pang  Y.L.,  Sim  L.C., An overview of photocatalytic degradation: photocatalysts, mechanisms, and development of photocatalytic membrane, Environmental Science and Pollution Research, 27 (2020) 2522–2565. DOI:10.1007/s11356-019-07193-5
•  Kofuji  Y.,  Isobe  Y.,  Shiraishi  Y.,  Sakamoto  H.,  Ichikawa  S.,  Tanaka  S.,  Hirai  T., Hydrogen peroxide production on a carbon nitride–boron nitride-reduced graphene oxide hybrid photocatalyst under visible light, ChemCatChem, 10 (2018) 2070–2077. DOI:10.1002/cctc.201701683
•  Kofuji  Y.,  Isobe  Y.,  Shiraishi  Y.,  Sakamoto  H.,  Tanaka  S.,  Ichikawa  S.,  Hirai  T., Carbon nitride–aromatic diimide–graphene nanohybrids: Metal-free photocatalysts for solar-to-hydrogen peroxide energy conversion with 0.2 % efficiency, Journal of the American Chemical Society, 138 (2016a)10019–10025. DOI:10.1021/jacs.6b05806
•  Kofuji  Y.,  Ohkita  S.,  Shiraishi  Y.,  Sakamoto  H.,  Ichikawa  S.,  Tanaka  S.,  Hirai  T., Mellitic triimide-doped carbon nitride as sunlight-driven photocatalysts for hydrogen peroxide production, ACS Sustainable Chemistry & Engineering, 5 (2017) 6478–6485. DOI:10.1021/acssuschemeng.7b00575
•  Kofuji  Y.,  Ohkita  S.,  Shiraishi  Y.,  Sakamoto  H.,  Tanaka  S.,  Ichikawa  S.,  Hirai  T., Graphitic carbon nitride doped with biphenyl diimide: efficient photocatalyst for hydrogen peroxide production from water and molecular oxygen by sunlight, ACS Catalysis, 6 (2016b)7021–7029. DOI:10.1021/acscatal.6b02367
•  Kulkarni  A.,  Siahrostami  S.,  Patel  A.,  Nørskov  J.K., Understanding catalytic activity trends in the oxygen reduction reaction, Chemical Reviews, 118 (2018) 2302–2312. DOI:10.1021/acs.chemrev.7b00488
•  Kuo  W.G., Decolorizing dye wastewater with Fenton’s reagent., Water Research, 26 (1992) 881–886. DOI:10.1016/0043-1354(92)90192-7
•  Laurent  A.D.,  Jacquemin  D., TD-DFT benchmarks: a review, International Journal of Quantum Chemistry, 113 (2013) 2019–2039. DOI:10.1002/qua.24438
•  Li  S.,  Dong  G.,  Hailili  R.,  Yang  L.,  Li  Y.,  Wang  F.,  Zeng  Y.,  Wang  C., Effective photocatalytic H₂O₂ production under visible light irradiation at g-C₃N₄ modulated by carbon vacancies, Applied Catalysis B: Environmental, 190 (2016) 26–35. DOI:10.1016/j.apcatb.2016.03.004
•  Li  X.,  Chen  C.,  Zhao  J., Mechanism of photodecomposition of H₂O₂ on TiO₂ surfaces under visible light irradiation, Langmuir, 17 (2001) 4118–4122. DOI:10.1021/la010035s
•  Li  X.,  Sun  X.,  Zhang  L.,  Sun  S.,  Wang  W., Efficient photocatalytic fixation of N₂ by KOH-treated g-C₃N₄, Journal of Materials Chemistry A, 6 (2018a)3005–3011. DOI:10.1039/C7TA09762J
•  Li  X.,  Zhang  J.,  Zhou  F.,  Zhang  H.,  Bai  J.,  Wang  Y.,  Wang  H., Preparation of N-vacancy-doped g-C₃N₄ with outstanding photocatalytic H₂O₂ production ability by dielectric barrier discharge plasma treatment, Chinese Journal of Catalysis, 39 (2018b)1090–1098. DOI:10.1016/S1872-2067(18)63046-3
•  Liao  G.,  Gong  Y.,  Zhang  L.,  Gao  H.,  Yang  G.-J.,  Fang  B., Semiconductor polymeric graphitic carbon nitride photocatalysts: the “holy grail” for the photocatalytic hydrogen evolution reaction under visible light, Energy & Environmental Science, 12 (2019) 2080–2147. DOI:10.1039/C9EE00717B
•  Liu  L.,  Gao  M.-Y.,  Yang  H.,  Wang  X.,  Li  X.,  Cooper  A.I., Linear conjugated polymers for solar-driven hydrogen peroxide production: the importance of catalyst stability, Journal of the American Chemical Society, 143 (2021) 19287–19293. DOI:10.1021/jacs.1c09979
•  Liu  R.,  Chen  Z.,  Yao  Y.,  Li  Y.,  Cheema  A.W.,  Wang  D.,  Zhu  S., Recent advancements in g-C₃N₄-based photocatalysts for photocatalytic CO₂ reduction: a mini review, RSC Advances, 10 (2020) 29408–29418. DOI:10.1039/D0RA05779G
•  Liu  X.,  Ma  R.,  Zhuang  L.,  Hu  B.,  Chen  J.,  Liu  X.,  Wang  X., Recent developments of doped g-C₃N₄ photocatalysts for the degradation of organic pollutants, Critical Reviews in Environmental Science and Technology, (2020) 1–40. DOI:10.1080/10643389.2020.1734433
•  Lu  Z.,  Chen  G.,  Siahrostami  S.,  Chen  Z.,  Liu  K.,  Xie  J.,  Liao  L.,  Wu  T.,  Lin  D.,  Liu  Y.,  Jaramillo  T.F.,  Nørskov  J.K.,  Cui  Y., High-efficiency oxygen reduction to hydrogen peroxide catalysed by oxidized carbon materials, Nature Catalysis, 1 (2018) 156–162. DOI:10.1038/s41929-017-0017-x
•  McDonnell  G., The use of hydrogen peroxide for disinfection and sterilization applications, PATAI’S Chemistry of Functional Groups, (2014) 1–34. DOI:10.1002/9780470682531.pat0885
•  Meng  A.,  Teng  Z.,  Zhang  Q.,  Su  C., Intrinsic defects in polymeric carbon nitride for photocatalysis applications, Chemistry– An Asian Journal, 15 (2020) 3405–3415. DOI:10.1002/asia.202000850
•  Mohamed  N.A.,  Safaei  J.,  Ismail  A.F.,  Jailani  M.F.A.M.,  Khalid  M.N.,  Noh  M.F.M.,  Aadenan  A.,  Nasir  S.N.S.N.,  Sagu  J.S.,  Teridi  M.S.M., The influences of post-annealing temperatures on fabrication graphitic carbon nitride, (g-C₃N₄) thin film, Applied Surface Science, (2019) 92–100. DOI:10.1016/j.apsusc.2019.05.312.
•  Moon  G.-h.,  Fujitsuka  M.,  Kim  S.,  Majima  T.,  Wang  X.,  Choi  W., Eco-friendly photochemical production of H₂O₂ through O₂ reduction over carbon nitride frameworks incorporated with multiple heteroelements, ACS Catalysis, 7 (2017) 2886–2895. DOI:10.1021/acscatal.6b03334
•  Mousavi Shaegh  S.A.,  Nguyen  N.-T.,  Mousavi Ehteshami  S.M.,  Chan  S.H., A membraneless hydrogen peroxide fuel cell using Prussian Blue as cathode material, Energy & Environmental Science, 5 (2012) 8225–8228. DOI:10.1039/C2EE21806B
•  Nosaka  Y.,  Nosaka  A., Introduction to Photocatalysis: From Basic Science to Applications, 1st edition, Royal Society of Chemistry, 2016, ISBN: 1782623205.
•  Ong  W.-J.,  Tan  L.-L.,  Ng  Y.H.,  Yong  S.-T.,  Chai  S.-P., Graphitic carbon nitride (g-C₃N₄)-based photocatalysts for artificial photosynthesis and environmental remediation: are we a step closer to achieving sustainability?, Chemical Reviews, 116 (2016) 7159–7329. DOI:10.1021/acs.chemrev.6b00075
•  Ou  H.,  Yang  P.,  Lin  L.,  Anpo  M.,  Wang  X., Carbon nitride aerogels for the photoredox conversion of water, Angewandte Chemie International Edition, 56 (2017) 10905–10910. DOI:10.1002/anie.201705926
•  Pei  Z.,  Ding  L.,  Hu  J,  Weng  S.,  Zheng  Z.,  Huang  M.,  Liu  P., Defect and its dominance in ZnO films: a new insight into the role of defect over photocatalytic activity, Applied Catalysis B: Environmental, 142–143 (2013) 736–743. DOI:10.1016/j.apcatb.2013.05.055
•  Peng  Y.,  Wang  L.,  Liu  Y.,  Chen  H.,  Lei  J.,  Zhang  J., Visible-light-driven photocatalytic H₂O₂ production on g-C₃N₄ loaded with CoP as a noble metal free cocatalyst, European Journal of Inorganic Chemistry, 2017 (2017) 4797–4802. DOI:10.1002/ejic.201700930
•  Pesterfield  L., The 100 most important chemical compounds: a reference guide (by Richard L. Myers), Journal of Chemical Education, 86 (2009) 1182. DOI:10.1021/ed086p1182
•  Qiang  Z.,  Chang  J.-H.,  Huang  C.P., Electrochemical generation of hydrogen peroxide from dissolved oxygen in acidic solutions, Water Research, 36 (2002) 85–94. DOI:10.1016/S0043-1354(01)00235-4
•  Qiu  C.,  Xu  Y.,  Fan  X.,  Xu  D.,  Tandiana  R.,  Ling  X.,  Jiang  Y.,  Liu  C.,  Yu  L.,  Chen  W.,  Su  C., Highly crystalline k-intercalated polymeric carbon nitride for visible-light photocatalytic alkenes and alkynes deuterations, Advanced Science, 6 (2019) 1801403. DOI:10.1002/advs.201801403
•  Qu  X.,  Hu  S.,  Li  P.,  Li  Z.,  Wang  H.,  Ma  H.,  Li  W., The effect of embedding N vacancies into g-C₃N₄ on the photocatalytic H₂O₂ production ability via H₂ plasma treatment, Diamond and Related Materials, 86 (2018) 159–166. DOI:10.1016/j.diamond.2018.04.027
•  Schneider  J.,  Bahnemann  D.W., Undesired role of sacrificial reagents in photocatalysis, The Journal of Physical Chemistry Letters, 4 (2013) 3479–3483. DOI:10.1021/jz4018199
•  Shen  R.,  Chen  W.,  Peng  Q.,  Lu  S.,  Zheng  L.,  Cao  X.,  Wang  Y.,  Zhu  W.,  Zhang  J.,  Zhuang  Z.,  Chen  C.,  Wang  D.,  Li  Y., High-concentration single atomic Pt sites on hollow CuSx for selective O₂ reduction to H₂O₂ in acid solution, Chem, 5 (2019) 2099–2110. DOI:10.1016/j.chempr.2019.04.024
•  Shi  L.,  Yang  L.,  Zhou  W.,  Liu  Y.,  Yin  L.,  Hai  X.,  Song  H.,  Ye  J., Photoassisted construction of holey defective g-C₃N₄ photocatalysts for efficient visible-light-driven H₂O₂ production, Small, 14 (2018) 1703142. DOI:10.1002/smll.201703142
•  Shi  X.,  Siahrostami  S.,  Li  G.-L.,  Zhang  Y.,  Chakthranont  P.,  Studt  F.,  Jaramillo  T.F.,  Zheng  X.,  Nørskov  J.K., Understanding activity trends in electrochemical water oxidation to form hydrogen peroxide, Nature Communications, 8 (2017) 701. DOI:10.1038/s41467-017-00585-6
•  Shiraishi  Y.,  Kanazawa  S.,  Kofuji  Y.,  Sakamoto  H.,  Ichikawa  S.,  Tanaka  S.,  Hirai  T., Sunlight-driven hydrogen peroxide production from water and molecular oxygen by metal-free photocatalysts, Angewandte Chemie International Edition, 53 (2014a)13454–13459. DOI:10.1002/anie.201407938
•  Shiraishi  Y.,  Kanazawa  S.,  Sugano  Y.,  Tsukamoto  D.,  Sakamoto  H.,  Ichikawa  S.,  Hirai  T., Highly selective production of hydrogen peroxide on graphitic carbon nitride (g-C₃N₄) photocatalyst activated by visible light, ACS Catalysis, 4 (2014b)774–780. DOI:10.1021/cs401208c
•  Shiraishi  Y.,  Kofuji  Y.,  Sakamoto  H.,  Tanaka  S.,  Ichikawa  S.,  Hirai  T., Effects of surface defects on photocatalytic H₂O₂ production by mesoporous graphitic carbon nitride under visible light irradiation, ACS Catalysis, 5 (2015) 3058–3066. DOI:10.1021/acscatal.5b00408
•  Shiraishi  Y.,  Takii  T.,  Hagi  T.,  Mori  S.,  Kofuji  Y.,  Kitagawa  Y.,  Tanaka  S.,  Ichikawa  S.,  Hirai  T., Resorcinol– formaldehyde resins as metal-free semiconductor photocatalysts for solar-to-hydrogen peroxide energy conversion, Nature Materials, 18 (2019) 985–993. DOI:10.1038/s41563-019-0398-0
•  Siahrostami  S.,  Verdaguer-Casadevall  A.,  Karamad  M.,  Deiana  D.,  Malacrida  P.,  Wickman  B.,  Escudero-Escribano  M.,  Paoli  E.A.,  Frydendal  R.,  Hansen  T.W.,  Chorkendorff  I.,  Stephens  I.E.L.,  Rossmeisl  J., Enabling direct H₂O₂ production through rational electrocatalyst design, Nature Materials, 12 (2013) 1137–1143. DOI:10.1038/nmat3795
•  Sun  H.,  Ma  Y.,  Zhang  Q.,  Su  C., Engineering the local coordination environment of single-atom catalysts and their applications in photocatalytic water splitting: a review, Transactions of Tianjin University, 27 (2021) 313–330. DOI:10.1007/s12209-021-00295-7
•  Sun  S.,  Li  J.,  Cui  J.,  Gou  X.,  Yang  Q.,  Jiang  Y.,  Liang  S.,  Yang  Z., Simultaneously engineering K-doping and exfoliation into graphitic carbon nitride (g-C₃N₄) for enhanced photocatalytic hydrogen production, International Journal of Hydrogen Energy, 44 (2019) 778–787. DOI:10.1016/j.ijhydene.2018.11.019
•  Tasis  D.,  Tagmatarchis  N.,  Bianco  A.,  Prato  M., Chemistry of carbon nanotubes, Chemical Reviews, 106 (2006) 1105–1136. DOI:10.1021/cr050569o
•  Teixeira  I.F.,  Barbosa  E.C.M.,  Tsang  S.C.E.,  Camargo  P.H.C., Carbon nitrides and metal nanoparticles: from controlled synthesis to design principles for improved photocatalysis, Chemical Society Reviews, 47 (2018) 7783–7817. DOI:10.1039/C8CS00479J
•  Teng  Z.,  Cai  W.,  Liu  S.,  Wang  C.,  Zhang  Q.,  Chenliang  S.,  Ohno  T., Bandgap engineering of polymetric carbon nitride copolymerized by 2,5,8-triamino-tri-s-triazine (melem) and barbituric acid for efficient nonsacrificial photocatalytic H₂O₂ production, Applied Catalysis B: Environmental, 271 (2020) 118917. DOI:10.1016/j.apcatb.2020.118917
•  Teng  Z.,  Cai  W.,  Sim  W.,  Zhang  Q.,  Wang  C.,  Su  C.,  Ohno  T., Photoexcited single metal atom catalysts for heterogeneous photocatalytic H₂O₂ production: Pragmatic guidelines for predicting charge separation, Applied Catalysis B: Environmental, 282 (2021a) 119589. DOI:10.1016/j.apcatb.2020.119589
•  Teng  Z.,  Lv  H.,  Wang  C.,  Xue  H.,  Pang  H.,  Wang  G., Bandgap engineering of ultrathin graphene-like carbon nitride nanosheets with controllable oxygenous functionalization, Carbon, 113 (2017) 63–75. DOI:10.1016/j.carbon.2016.11.030
•  Teng  Z.,  Yang  N.,  Lv  H.,  Wang  S.,  Hu  M.,  Wang  C.,  Wang  D.,  Wang  G., Edge-functionalized g-C₃N₄ nanosheets as a highly efficient metal-free photocatalyst for safe drinking water, Chem, 5 (2019) 664–680. DOI:10.1016/j.chempr.2018.12.009
•  Teng  Z.,  Zhang  Q.,  Yang  H.,  Kato  K.,  Yang  W.,  Lu  Y.-R.,  Liu  S.,  Wang  C.,  Yamakata  A.,  Su  C.,  Liu  B.,  Ohno  T., Atomically dispersed antimony on carbon nitride for the artificial photosynthesis of hydrogen peroxide, Nature Catalysis, 4 (2021b)374–384. DOI:10.1038/s41929-021-00605-1
•  Thakur  S.,  Kshetri  T.,  Kim  N.H.,  Lee  J.H., Sunlight-driven sustainable production of hydrogen peroxide using a CdS–graphene hybrid photocatalyst, Journal of Catalysis, 345 (2017) 78–86. DOI:10.1016/j.jcat.2016.10.028
•  Torres-Pinto  A.,  Sampaio  M.J.,  Silva  C.G.,  Faria  J.L.,  Silva  M.T.A., Recent strategies for hydrogen peroxide production by metal-free carbon nitride photocatalysts, Catalysts, 9 (2019) 990. DOI:10.3390/catal9120990
•  Tsukamoto  D.,  Shiro  A.,  Shiraishi  Y.,  Sugano  Y.,  Ichikawa  S.,  Tanaka  S.,  Hirai  T., Photocatalytic H₂O₂ production from ethanol/O₂ system using TiO₂ loaded with Au–Ag bimetallic alloy nanoparticles, ACS Catalysis, 2 (2012) 599–603. DOI:10.1021/cs2006873
•  Verdaguer-Casadevall  A.,  Deiana  D.,  Karamad  M.,  Siahrostami  S.,  Malacrida  P.,  Hansen  T.W.,  Rossmeisl  J.,  Chorkendorff  I.,  Stephens  I.E.L., Trends in the electrochemical synthesis of H₂O₂: enhancing activity and selectivity by electrocatalytic site engineering, Nano Letters, 14 (2014) 1603–1608. DOI:10.1021/nl500037x
•  Wang  A.,  Li  J.,  Zhang  T., Heterogeneous single-atom catalysis, Nature Reviews Chemistry, 2 (2018) 65–81. DOI:10.1038/s41570-018-0010-1
•  Wang  M.,  Shen  S.,  Li  L.,  Tang  Z.,  Yang  J., Effects of sacrificial reagents on photocatalytic hydrogen evolution over different photocatalysts, Journal of Materials Science, 52 (2017) 5155–5164. DOI:10.1007/s10853-017-0752-z
•  Wang  Q.,  Domen  K., Particulate photocatalysts for light-driven water splitting: mechanisms, challenges, and design strategies, Chemical Reviews, 120 (2020) 919–985. DOI:10.1021/acs.chemrev.9b00201
•  Wang  S.,  Teng  Z.,  Xu  Y.,  Yuan  M.,  Zhong  Y.,  Liu  S.,  Wang  C.,  Wang  G.,  Ohno  T., Defect as the essential factor in engineering carbon-nitride-based visible-light-driven Z-scheme photocatalyst, Applied Catalysis B: Environmental, 260 (2020a) 118145. DOI:10.1016/j.apcatb.2019.118145
•  Wang  S.,  Zhan  J.,  Chen  K.,  Ali  A.,  Zeng  L.,  Zhao  H.,  Hu  W.,  Zhu  L.,  Xu  X., Potassium-doped g-C₃N₄ achieving efficient visible-light-driven CO₂ Reduction, ACS Sustainable Chemistry & Engineering, 8 (2020b)8214–8222. DOI:10.1021/acssuschemeng.0c01151
•  Wang  X.,  Maeda  K.,  Thomas  A.,  Takanabe  K.,  Xin  G.,  Carlsson  J.M.,  Domen  K.,  Antonietti  M., A metal-free polymeric photocatalyst for hydrogen production from water under visible light, Nature Materials, 8 (2009) 76–80. DOI:10.1038/nmat2317
•  Wang  Y.,  Vogel  A.,  Sachs  M.,  Sprick  R.S.,  Wilbraham  L.,  Moniz  S.J.A.,  Godin  R.,  Zwijnenburg  M.A.,  Durrant  J.R.,  Cooper  A.I.,  Tang  J., Current understanding and challenges of solar-driven hydrogen generation using polymeric photocatalysts, Nature Energy, 4 (2019) 746–760. DOI:10.1038/s41560-019-0456-5
•  Wei  Y.,  Zhang  J.,  Zheng  Q.,  Miao  J.,  Alvarez  P.J.,  Long  M., Quantification of photocatalytically-generated hydrogen peroxide in the presence of organic electron donors: interference and reliability considerations, Chemosphere, 279 (2021) 130556. DOI:10.1016/j.chemosphere.2021.130556
•  Wei  Z.,  Liu  M.,  Zhang  Z.,  Yao  W.,  Tan  H.,  Zhu  Y., Efficient visible-light-driven selective oxygen reduction to hydrogen peroxide by oxygen-enriched graphitic carbon nitride polymers, Energy & Environmental Science, 11 (2018) 2581–2589. DOI:10.1039/C8EE01316K
•  Xia  C.,  Xia  Y.,  Zhu  P.,  Fan  L.,  Wang  H., Direct electrosynthesis of pure aqueous H₂O₂ solutions up to 20 % by weight using a solid electrolyte, Science, 366 (2019) 226. DOI:10.1126/science.aay1844
•  Xiong  J.,  Di  J.,  Xia  J.,  Zhu  W.,  Li  H., Surface defect engineering in 2D nanomaterials for photocatalysis, Advanced Functional Materials, 28 (2018) 1801983. DOI:10.1002/adfm.201801983
•  Xu  Q.,  Zhang  L.,  Yu  J.,  Wageh  S.,  Al-Ghamdi  A.A.,  Jaroniec  M., Direct Z-scheme photocatalysts: principles, synthesis, and applications, Materials Today, 21 (2018) 1042–1063. DOI:10.1016/j.mattod.2018.04.008
•  Yamada  Y.,  Yoneda  M.,  Fukuzumi  S., A robust one-compartment fuel cell with a polynuclear cyanide complex as a cathode for utilizing H₂O₂ as a sustainable fuel at ambient conditions, Chemistry – A European Journal, 19 (2013) 11733–11741. DOI:10.1002/chem.201300783
•  Yang  H.B.,  Hung  S.-F.,  Liu  S.,  Yuan  K.,  Miao  S.,  Zhang  L.,  Huang  X.,  Wang  H.-Y.,  Cai  W.,  Chen  R.,  Gao  J.,  Yang  X.,  Chen  W.,  Huang  Y.,  Chen  H.M., et al.., Atomically dispersed Ni(i) as the active site for electrochemical CO₂ reduction, Nature Energy, 3 (2018) 140–147. DOI:10.1038/s41560-017-0078-8
•  Yang  L.,  Dong  G.,  Jacobs  D.L.,  Wang  Y.,  Zang  L.,  Wang  C., Two-channel photocatalytic production of H₂O₂ over g-C₃N₄ nanosheets modified with perylene imides, Journal of Catalysis, 352 (2017) 274–281. DOI:10.1016/j.jcat.2017.05.010
•  Yang  S.,  Verdaguer-Casadevall  A.,  Arnarson  L.,  Silvioli  L.,  Čolić  V.,  Frydendal  R.,  Rossmeisl  J.,  Chorkendorff  I.,  Stephens  I.E.L., Toward the decentralized electrochemical production of H₂O₂: a focus on the catalysis, ACS Catalysis, 8 (2018) 4064–4081. DOI:10.1021/acscatal.8b00217
•  Yi  Y.,  Wang  L.,  Li  G.,  Guo  H., A review on research progress in the direct synthesis of hydrogen peroxide from hydrogen and oxygen: noble-metal catalytic method, fuel-cell method and plasma method, Catalysis Science and Technology, 6 (2016) 1593–1610. DOI:10.1039/C5CY01567G
•  Zeng  X.,  Liu  Y.,  Kang  Y.,  Li  Q.,  Xia  Y.,  Zhu  Y.,  Hou  H.,  Uddin  M.H.,  Gengenbach  T.R.,  Xia  D.,  Sun  C.,  Mccarthy  D.T.,  Deletic  A.,  Yu  J.,  Zhang  X., Simultaneously tuning charge separation and oxygen reduction pathway on graphitic carbon nitride by polyethylenimine for boosted photocatalytic hydrogen peroxide production, ACS Catalsis, 10 (2020) 3697–3706. DOI:10.1021/acscatal.9b05247
•  Zeng  X.,  Wang  Z.,  Meng  N.,  McCarthy  D.T.,  Deletic  A.,  Pan  J.-h.,  Zhang  X., Highly dispersed TiO₂ nanocrystals and carbon dots on reduced graphene oxide: ternary nanocomposites for accelerated photocatalytic water disinfection, Applied Catalysis B: Environmental, 202 (2017) 33–41. DOI:10.1016/j.apcatb.2016.09.014
•  Zhan  W.,  Ji  L.,  Ge  Z.-m.,  Wang  X.,  Li  R.-t., A continuous-flow synthesis of primary amides from hydrolysis of nitriles using hydrogen peroxide as oxidant, Tetrahedron, 74 (2018) 1527–1532. DOI:10.1016/j.tet.2018.02.017
•  Zhang  B.,  Qiu  C.,  Wang  S.,  Gao  H.,  Yu  K.,  Zhang  Z.,  Ling  X.,  Ou  W.,  Su  C., Electrocatalytic water-splitting for the controllable and sustainable synthesis of deuterated chemicals, Science Bulletin, 66 (2021) 562–569. DOI:10.1016/j.scib.2020.09.016
•  Zhang  C.,  Bai  J.,  Ma  L.,  Lv  Y.,  Wang  F.,  Zhang  X.,  Yuan  X.,  Hu  S., Synthesis of halogen doped graphite carbon nitride nanorods with outstanding photocatalytic H₂O₂ production ability via saturated NH₄X (X = Cl, Br) solution-hydrothermal post-treatment, Diamond and Related Materials, 87 (2018) 215–222. DOI:10.1016/j.diamond.2018.06.013
•  Zhang  G.,  Yang  X.,  He  C.,  Zhang  P.,  Mi  H., Constructing a tunable defect structure in TiO₂ for photocatalytic nitrogen fixation, Journal of Materials Chemistry A, 8 (2020) 334–341. DOI:10.1039/C9TA10471B
•  Zhang  M.-h.,  Dong  H.,  Zhao  L.,  Wang  D.-x.,  Meng  D., A review on Fenton process for organic wastewater treatment based on optimization perspective, Science of the Total Environment, 670 (2019) 110–121. DOI:10.1016/j.scitotenv.2019.03.180
•  Zhang  P.,  Sun  D.,  Cho  A.,  Weon  S.,  Lee  S.,  Lee  J.,  Han  J.W.,  Kim  D.-P.,  Choi  W., Modified carbon nitride nanozyme as bifunctional glucose oxidase-peroxidase for metal-free bioinspired cascade photocatalysis, Nature Communications, 10 (2019) 940. DOI:10.1038/s41467-019-08731-y
•  Zhang  P.,  Tong  Y.,  Liu  Y.,  Vequizo  J.J.M.,  Sun  H.,  Yang  C.,  Yamakata  A.,  Fan  F.,  Lin  W.,  Wang  X., Heteroatom dopants promote two-electron O₂ reduction for photocatalytic production of H₂O₂ on polymeric carbon nitride, Angewandte Chemie, 132 (2020) 16343–16351. DOI:10.1002/ange.202006747
•  Zhang  Z.,  Xu  Y.,  Zhang  Q.,  Fang  S.,  Sun  H.,  Ou  W.,  Su  C. Semi-heterogeneous photo-Cu-dual-catalytic cross-coupling reactions using polymeric carbon nitrides. Science Bulletin, 67 (2022) 71–78. DOI:10.1016/j.scib.2021.08.001
•  Zhang  Z.,  Yates  J.T., Band bending in semiconductors: Chemical and physical consequences at surfaces and interfaces, Chemical Reviews, 112 (2012) 5520–5551. DOI:10.1021/cr3000626
•  Zhao  S.,  Guo  T.,  Li  X.,  Xu  T.,  Yang  B.,  Zhao  X., Carbon nanotubes covalent combined with graphitic carbon nitride for photocatalytic hydrogen peroxide production under visible light, Applied Catalysis B: Environmental, 224 (2018a)725–732. DOI:10.1016/j.apcatb.2017.11.005
•  Zhao  S.,  Zhao  X., Insights into the role of singlet oxygen in the photocatalytic hydrogen peroxide production over polyoxometalates-derived metal oxides incorporated into graphitic carbon nitride framework, Applied Catalysis B: Environmental, 250 (2019) 408–418. DOI:10.1016/j.apcatb.2019.02.031
•  Zhao  S.,  Zhao  X.,  Ouyang  S.,  Zhu  Y., Polyoxometalates covalently combined with graphitic carbon nitride for photocatalytic hydrogen peroxide production, Catalysis Science and Technology, 8 (2018b)1686–1695. DOI:10.1039/C8CY00043C
•  Zhao  S.,  Zhao  X.,  Zhang  H.,  Li  J.,  Zhu  Y., Covalent combination of polyoxometalate and graphitic carbon nitride for light-driven hydrogen peroxide production, Nano Energy, 35 (2017) 405–414. DOI:10.1016/j.nanoen.2017.04.017
•  Zhu  J.,  Fan  F.,  Chen  R.,  An  H.,  Feng  Z.,  Li  C., Direct imaging of highly anisotropic photogenerated charge separations on different facets of a single BiVO₄ photocatalyst, Angewandte Chemie International Edition, 54 (2015) 9111–9114. DOI:10.1002/anie.201504135
•  Zhu  Z.,  Pan  H.,  Murugananthan  M.,  Gong  J.,  Zhang  Y., Visible light-driven photocatalytically active g-C₃N₄ material for enhanced generation of H₂O₂, Applied Catalysis B: Environmental, 232 (2018) 19–25. DOI:10.1016/j.apcatb.2018.03.035
•  Zuo  G.,  Liu  S.,  Wang  L.,  Song  H.,  Zong  P.,  Hou  W.,  Li  B.,  Guo  Z.,  Meng  X.,  Du  Y.,  Wang  T.,  Roy  V.A.L., Finely dispersed Au nanoparticles on graphitic carbon nitride as highly active photocatalyst for hydrogen peroxide production, Catalysis Communications, 123 (2019) 69–72. DOI:10.1016/j.catcom.2019.02.011

Authors’ Short Biographies

Zhenyuan Teng

Dr. Zhenyuan Teng is currently a JSPS Research Fellow (PD) in the Research Initiative for Supra-Materials (RISM), Interdisciplinary Cluster for Cutting Edge Research, Shinshu University. He obtained his B.S. degree in 2016 and then acquired the Master degree (2018) from Yangzhou University. He obtained his Doctoral degree from Kyushu Institute of Technology in 2021 under the supervision of Prof. Teruhisa Ohno. His research interest focuses on the design and synthesis of single-atom organic-inorganic photocatalytic systems for energetic and environmental applications.

Wenan Cai

Wenan Cai received his B.S. in Beijing Institute of Petrochemical Technology in 2016. Currently, he is a Ph.D. student under the supervision of Prof. Ohno Teruhisa in Kyushu Institute of Technology. His research interest is the development and design of long-wavelength response photocatalysts for H₂O₂ production.

Teruhisa Ohno

Prof. Teruhisa Ohno is currently the dean of faculty of engineering in Kyushu Institute of Technology. He obtained his Doctoral degree in Kyushu University in 1988. He worked as an Associate Professor in Kyushu University from 1992 to 1994 and in Osaka University from 1994–2003. Then he moved to Kyushu Institute of Technology as a Full Professor. From 2016 to 2021, He published 57 SCI papers. Field-Weighted Citation Impact: 2.04, Citation Count: 1,108, h-index: 48, h5-index: 15. A number of papers have published on influential periodical, such as Applied Catalysis B: Environmental, ACS catalysis and Nature Catalysis. From 1983 to now, published about 202 SCI papers, these papers were totally quoted more than 11,763 times and h-index is 49 (Scopus Data). His research interests are: 1. Development of visible light sensitive photocatalysts; 2. Nanoscale surface structure control of TiO₂ photocatalysts; 3. Development of photocatalyst and photoelectrode system for CO₂ reduction; 4. Visible light responsive photocatalyst and photoelectrode for H₂O₂ production.