Recent advances in and comprehensive consideration of the oxidation half reaction in photocatalytic CO₂ conversion

Abstract

Photocatalytic conversion of CO₂ into value-added chemicals is a promising method to tackle the global warming issue and ease the need for sustainable energy. To date, much effort has been devoted to improving the electron behavior of the reduction process for effective CO₂ reduction. However, it is often overlooked that the simultaneous control of electrons and holes is more meaningful and beneficial for an overall photocatalytic reaction, since the two half reactions driven respectively by electrons and holes are equally important. Considering that much less attention has been paid to the oxidation half reaction, this review specifically highlights the approaches and strategies for promoting the oxidation half reaction that accompanies CO₂ conversion, which can in turn improve the CO₂ photoreduction performance. In detail, a comprehensive discussion on regulating the hole behavior (e.g. lifetime and driving force) is presented in three parts, viz. materials design (element modification, oxidation co-catalysts, and Z-scheme structures), electron donors (H₂O and sacrificial electron donors), and reaction system management (high value-added oxidation reaction). We anticipate that this review will offer a systematic summary of the past achievements and general considerations for the oxidation half reaction occurring during CO₂ reduction, which will provide possible routes and/or implications for scientists to further improve the CO₂ conversion efficiency.

---

Yanjie Wang

Yanjie Wang received her PhD in physical chemistry from the National Center for Nanoscience and Technology, China, in 2018. After that she joined the National Center for Nanoscience and Technology. Her current research focuses on the rational design of nanomaterials and their application in sustainable energy.

Tao He

Tao He received his PhD from the Institute of Chemistry, Chinese Academy of Sciences in 2002. After postdoctoral training at the Weizmann Institute of Science and Rice University, he joined the National Center for Nanoscience and Technology, China in 2009. He is now a professor at the National Center for Nanoscience and Technology, China. He was selected by the Hundred Talents Program (Chinese Academy of Sciences) in 2009. His current research interest mainly focuses on R&D of novel photoelectric functional nanomaterials and related devices that can efficiently utilize light energy at a relatively low cost, specifically for the photocatalytic reduction of CO2.

---

1 Introduction

Considering the ever-growing threat from CO₂ emission into the atmosphere, it is critical to balance the carbon cycle in nature. Carbon conversion is a feasible strategy to reduce CO₂ emission and simultaneously meet the increasing demand for sustainable energy. To this end, photocatalytic CO₂ reduction to value-added chemicals has attracted ever-increasing attention for decades as a green technique. Because the CO₂ molecule is extremely stable (ΔG^θ_f = −394.4 kJ mol⁻¹), a significant amount of energy input is required for activation and hence a highly efficient photocatalyst is needed.¹ In the past decades, enormous efforts have been made to realize photocatalytic CO₂ conversion by developing semiconductor photocatalysts, such as metal oxides, metal sulfides and metal nitrides, which possess a conduction band potential negative enough to drive the CO₂ reduction reaction (CO₂RR) (Fig. 1). To achieve high conversion efficiency, extensive efforts have been devoted to regulating the photogenerated electron behavior to boost the CO₂RR both thermodynamically and kinetically. Approaches like doping, loading cocatalysts and fabricating heterojunctions have been widely employed.^2,3 However, the photocatalytic CO₂RR performance is still below expectations and far from commercial applications, implying that other factors should be considered.

Fig. 1 Energy band diagram for semiconductors used in the CO₂ reduction. Reprinted with permission from ref. 4. Copyright 2019, American Chemical Society.

It is to be noted that an overall photocatalytic reaction involves both reduction and oxidation half-reactions driven respectively by photogenerated electrons and holes. So the photocatalysis is highly dependent on the behavior of electrons and holes (such as generation, separation/recombination and transport) and surface redox reaction, with electron–hole recombination as the dominant loss and surface reaction as the rate-determining step. The overall catalytic efficiency can be improved by promoting the charge separation and surface redox reaction via controlling the behavior of either electrons or holes. In this regard, modulation of hole behavior is as important as that of electrons for the photocatalysis. Taking CO₂ photoconversion in H₂O as an example,⁵ in the oxidation part (2H₂O + 4h⁺ → O₂ + 4H⁺), it requires four holes to remove four protons from two H₂O molecules for the production of one O₂ molecule, with a minimum required potential of 0.82 V vs. SHE. Hence, besides the CO₂RR, the oxidation part is also a bottleneck for improved solar-to-chemical conversion efficiency.

Moreover, the mobility of photogenerated holes is usually lower than that of electrons in a semiconductor.⁶ Increasing the hole mobility can accelerate the oxidation half reaction (OHR) and leave more electrons and protons for the reduction half reaction. Zou et al. proved this by combining experimental and theoretical calculation results, i.e., higher hole mobility on the {100} facets of ZnGa₂O₄ is favorable for efficient H₂O oxidation, leading to more protons for CO₂ reduction to CH₄.⁷ In addition, some of the oxidation processes are relatively slow (around 0.3 s for H₂O),⁸ and the holes accumulated on the photocatalyst surface may accelerate the charge recombination, which will slow down the overall photocatalytic process. Kamat pointed out the importance of trapping or scavenging either charge carrier to accelerate the dynamics of the opposite carrier.⁹ Hence, promoting hole consumption will also help suppress the electron–hole recombination and thus accelerate the CO₂RR.

For the time being, most of the attention is focused on regulating the electron behavior in CO₂RR research, while much less attention is focused on the modulation of the hole behavior. Some approaches have been used to promote the OHR in the CO₂RR like introducing sacrificial electron donors, loading oxidation cocatalysts, and establishing Z-scheme structures.^10–12 All of these approaches can either enhance the hole-involving reactions or promote the behavior of both electrons and holes synergistically. Despite the availability of many review articles on the general aspects of the CO₂RR like photocatalyst design and mechanism study,^5,13–17 there is a lack of comprehensive summary of the general considerations and progress of these studies on the OHR.

Herein, we present a systematic review of the state of the art of the OHR involved in the CO₂RR study. An overview of the topics covered in this article is provided in Fig. 2. Firstly, materials design, which aims to improve the trapping ability or maintain the high driving force of photogenerated holes, is presented in detail. Then, electron donors are discussed, including the widely used H₂O and various other sacrificial electron donors. The CO₂RR occurring in aqueous solution is highlighted, and studies on the stoichiometric yield of reduction and oxidation products are systematically analyzed. Furthermore, tandem systems aiming to fully utilize the overall reaction, for which both oxidation and reduction half reactions are meaningful, are covered. This review aims to present a detailed discussion on the OHR in the CO₂RR study, which can help researchers understand CO₂ reduction at the full scale and further inspire them to achieve high conversion efficiency.

Fig. 2 Overview of the topics discussed in this review.

2 Materials design for boosting oxidation reactions

Materials design is a common way to boost the photocatalytic activity by tuning the charge behavior, such as redox potential and charge generation/separation efficiency. The performance of a single pure photocatalyst is limited by the possible backward reactions on neighboring active sites and its intrinsic properties that cannot meet all the requirements. To further improve the CO₂RR activity, modification of the photocatalytic systems is desired, which can increase the driving force of the charge carrier and spatially separate the two half reactions. The common approaches include functionalizing the photocatalyst with other elements, loading cocatalysts and interfacial engineering (forming heterostructures). Since advances in materials design for improved CO₂RR have already been thoroughly summarized in the literature, here we only focus on the strategies and approaches of managing the associated OHR.

2.1 Element modification

Introducing foreign elements (i.e., doping) into the lattice of a photocatalyst is an effective means to narrow down the bandgap by creating acceptor levels above the valence band (VB) or donor levels below the conduction band (CB). Various semiconductors, especially those with wide bandgap values, have been employed as host materials, such as TiO₂, SrTiO₃, WO₃, C₃N₄ and Fe₂O₃.^18–22 However, the influence of doping on the charge behavior, especially on the hole behavior, is still a subject of debate. Even though the visible-light response can be enhanced and it was claimed that most photogenerated holes can avoid being trapped at the N-induced gap,²³ a number of reports have pointed out that undesired effects may emerge as a result of doping, resulting in decreased photoactivity. Many doping methods rely on raising the energy position of the VB. For example, it is popular to partially replace O in metal oxides by N or C to raise the VB position for narrowing the bandgap. Taking N doping in TiO₂ as an example, since the N 2p orbital has higher energy than the O 2p orbital, the additional intraband state from N 2p lies above the VB of TiO₂ (Fig. 3). Dynamics of the charge carriers derived from time-resolved absorption spectra show that the states induced by N doping may result in electron–hole recombination and reduce the oxidation power of photogenerated holes,¹⁸ leading to decreased water oxidation performance.^5,18 Therefore, when introducing a foreign element into a photocatalyst, it is desirable to consider its influence on the VB position of the host material and thereby the hole behavior simultaneously.

Fig. 3 Semiconductor with dopant states above the VB. Reprinted with permission from ref. 5. Copyright 2013, Wiley-VCH.

Halogen elements (especially Cl⁻) incorporated on semiconductors are able to help trap the photogenerated holes, facilitate the charge separation and thereby improve the photocatalytic CO₂RR; however, the explanations in various studies on how Cl⁻ works are still controversial. Li et al. claimed that Cl⁻ adsorbed on TiO₂ can lower the activation energy for water oxidation by reacting with the OHR intermediate (Fig. 4a),²⁴ while Jing et al. proposed that the Cl⁻ incorporated on rutile TiO₂ nanorods can trap holes due to its lone-pair electrons and the appropriate electronegativity of the Cl atom intermediate which is favorable for water oxidation (Fig. 4b).²⁵ Moreover, they reported that rutile TiO₂ nanorods co-modified with Cl⁻ and phosphate groups, both of which are demonstrated to help trap holes, can lead to improved CO₂ conversion to CH₄ and CO.²⁶ They further indicated that the strong electronegativity of halogens (F⁻, Cl⁻, and Br⁻) can induce surface polarization on g-C₃N₄ and hence the positively charged holes can be effectively captured (Fig. 4c).²⁷ Therefore, the halogen modified g-C₃N₄ shows simultaneously enhanced reduction and oxidation activity in the photocatalytic CO₂RR and 2,4-dichlorophenol degradation, respectively. Recently, Cl was doped into Cu₂O.²⁸ Interestingly, the authors proposed that Cl shifts the VB of Cu₂O to a more positive potential for more efficient water oxidation (Fig. 4d). Cl doped Cu₂O exhibits much higher photocatalytic CO₂RR performance than pure Cu₂O, and the O₂ evolution reaction (OER) occurs following an almost stoichiometric ratio to the reduction product (Fig. 4e). In addition, an ¹⁸O isotope labelling experiment was employed to verify that the evolved O₂ comes from H₂O rather than the photocatalyst.²⁸

Fig. 4 Schematic illustrations of (a) a possible mechanism for how Cl⁻ lowers the water oxidation activation energy on a Cl⁻ adsorbed TiO₂ surface. Reprinted with permission from ref. 24. Copyright 2014, Royal Society of Chemistry. (b) Photogenerated hole transfer and reaction pathway with water on Cl-modified rutile TiO₂. Reprinted with permission from ref. 25. Copyright 2016, Springer Nature. (c) Photogenerated charge process over halogen-modified g-C₃N₄. Reprinted with permission from ref. 27. Copyright 2017, Elsevier. (d) Band structure alignments of pure Cu₂O and Cu₂O–Cl-4. Reprinted with permission from ref. 28. Copyright 2019, Elsevier. (e) Photocatalytic CO₂ reduction performance over pure Cu₂O and Cl-doped Cu₂O samples under 6 h visible-light irradiation. Reprinted with permission from ref. 28. Copyright 2019, Elsevier.

2.2 Oxidation cocatalysts

Cocatalysts usually help lower the activation energy of catalytic reactions, extract/trap superfluous photogenerated carriers to inhibit recombination, and provide active sites for the redox reactions.¹¹ In this case, both reduction cocatalysts for electron capture and oxidation cocatalysts for hole capture have been developed (Fig. 5). Many efforts have been devoted to developing reduction cocatalysts for the photocatalytic CO₂RR including various metals (Pt, Ag, Au, etc.), semiconductors (Cu₂O, MoS₂, etc.) and carbonaceous materials (Fig. 5), while much less attention has been paid to oxidation cocatalysts. In essence, the photogenerated holes usually suffer from low oxidation ability, especially for catalysts with relatively high VB positions. Since oxidation cocatalysts can help lower the activation energy and promote hole trapping, it is desirable to develop oxidation cocatalysts for efficient oxidation reactions accompanying the CO₂RR.^29–32 So far, some oxidation cocatalysts, such as CoO_x, MnO_x, RuO_x, IrO_x and CoPi, have been applied to CO₂RR systems, and they exhibit high performance in promoting the oxidation reactions.

Fig. 5 Photocatalytic CO₂ reduction over TiO₂ loaded with reduction and oxidation cocatalysts (potentials in the plot at pH = 0 vs. NHE). Reprinted with permission from ref. 11. Copyright 2019, Wiley-VCH.

2.2.1 Single oxidation cocatalysts. The improved catalytic performance brought about by oxidation cocatalysts mainly relies on their promotion of hole behavior (transfer and lifetime), surface redox reaction kinetics and photocatalyst stability. A spectroscopic study on the kinetics of photogenerated hole transfer influenced by IrO₂ for a TiO₂/IrO₂ system indicates that the hole transfer rate from TiO₂ to IrO₂ (6 × 10⁵ s⁻¹) is much faster than the direct transfer from the VB of TiO₂ and the catalytic turnover rate.³³ Such a rapid charge transfer across the semiconductor/cocatalyst interface makes it possible to separate the electron and hole spatially. Furthermore, the lifetime of holes is also increased by loading oxidation cocatalysts. Co–Pi has become a popular oxidation cocatalyst that can be easily formed from earth-abundant ions since its first application in water oxidation in 2008.³⁴ Transient absorption spectrum results indicate that the addition of Co–Pi can increase the lifetime of photogenerated holes in α-Fe₂O₃ by 3 orders of magnitude.³⁵ Moreover, oxidation cocatalysts can enhance the stability of the semiconductors via preventing their oxidative corrosion by photogenerated holes. For instance, the employment of RuO_x can increase the stability of Cu₂O by trapping holes.³⁶ It is worth noting that the RuO_x loading on Cu₂O can also improve the photocatalytic CO₂RR selectivity compared to Cu₂O, mainly due to blocking of the H₂ evolution sites on Cu₂O by RuO_x.

There are some materials that possess the required reduction power for the CO₂RR, but the VB position is not positive enough or the overpotential is not large enough to drive the OHR efficiently. In this case, the adoption of oxidation cocatalysts can be helpful. For example, CoO_x particles incorporated on g-C₃N₄ act as an effective oxidative promoter to accelerate the charge transfer kinetics and thus enhance the CO₂ to CO conversion.³⁷ Similarly, the introduction of polyoxometalates (Co4) onto g-C₃N₄ not only facilitates the charge transfer but also increases the surface catalytic oxidative ability of the system, giving rise to high CO yield.³⁸

2.2.2 Redox dual cocatalysts. Simultaneously modulating the photogenerated electrons and holes is an effective way to greatly improve the photocatalytic CO₂RR performance, which is favorable for the charge separation and catalyzing the respective reactions. Thus, a combination of both reduction and oxidation cocatalysts (i.e., dual cocatalysts) has been developed for the photocatalytic CO₂RR, and it displays great potential in separating photogenerated electrons and holes and extracting them to the respective sites for the reactions.¹¹ Jing et al. simultaneously introduced Ni and Co species onto g-C₃N₄ to modulate the respective behavior of electrons and holes, with 3.1-times higher photocatalytic CO₂ reduction activity than that of g-C₃N₄.²⁹ Shoji et al. reported a SrTiO₃ semiconductor loaded with Cu_xO as the reduction cocatalyst and CoPi as the oxidation cocatalyst, which shows high CO₂ to CO conversion ability as well as stoichiometric O₂ evolution (Fig. 6a).³⁹ This system mimics the natural leaf that realizes the reduction and oxidation reactions in water without any external bias. In another study, BiOI loaded with Au and MnO_x respectively as the reduction and oxidation cocatalysts (i.e., Au/BiOI/MnO_x) shows enhanced performance for CO₂ reduction in H₂O compared to BiOI only loaded with Au or MnO_x (Fig. 6b).⁴⁰ In addition, it is to be noted that the hole trapping ability of an oxidation cocatalyst is often unequal to the electron trapping ability of a reduction cocatalyst. The photocatalytic CO₂RR performance of Au/BiOI is superior to that of MnO_x/BiOI,⁴⁰ partly because the enhancement caused by the reduction cocatalyst Au is higher than that caused by the oxidation cocatalyst MnO_x. However, it is reported that the dual cocatalysts of Pt and RuO_x on Zn₂GeO₄ exhibit a reverse promotion trend for the photocatalytic CO₂RR, with the oxidation cocatalyst showing higher enhancement in the CO₂ reduction properties.⁴¹ Specifically, Zn₂GeO₄ co-loaded with Pt and RuO₂ shows a significantly enhanced CH₄ generation rate, and the CH₄ yield from RuO₂/Zn₂GeO₄ almost triples compared to that of Pt/Zn₂GeO₄ (Fig. 6c).⁴¹ Similar results are also reported for the Pt/Zn_1.7GeN_1.8O/RuO₂ system.⁴² Hence, different types of cocatalysts can result in a distinct carrier migration and influence on the ultimate performance. Moreover, the amount of oxidation cocatalyst also has an impact on the CO₂RR activity and selectivity, as confirmed with a RuO₂ loaded TiO₂–SiO₂ framework.⁴³ By optimizing loading amounts of RuO₂ and Au, a high CH₄ yield (5.9 μmol g⁻¹ h⁻¹) and selectivity (86.4%) are achieved, due to the optimized charge separation and surface reaction pathway influenced by the number of active sites and electrons generated.

Fig. 6 (a) Product yield of photocatalytic CO₂ reduction over Cu_xO/SrTiO₃/CoPi. Reprinted with permission from ref. 39. Copyright 2017, American Chemical Society. (b) Photocatalytic CO₂RR performance of BiOI, Au/BiOI/MnO_x, Au/BiOI and MnO_x/BiOI under Xe lamp irradiation; the inset shows the results under visible light irradiation. Reprinted with permission from ref. 40. Copyright 2016, Royal Society of Chemistry. (c) Photocatalytic CO₂RR performance of (a) bulk Zn₂GeO₂, (b) Zn₂GeO₂ nanoribbons, (c) Pt/Zn₂GeO₂ nanoribbons, (d) RuO₂/Zn₂GeO₂ nanoribbons and (e) Pt/Zn₂GeO₂ nanoribbons/RuO₂. Reprinted with permission from ref. 41. Copyright 2010, American Chemical Society.

Although the parallel utilization of oxidation and reduction cocatalysts is favorable for the CO₂RR, their random distribution may lead to contact between them, giving rise to increased chances of charge recombination and reverse reactions and thus lowered overall efficiency. Accordingly, researchers started to consider designing photocatalysts with a well-defined structure (facet, morphology, etc.) and carefully tuning the loading sites of the reduction and oxidation cocatalysts on the materials.

Hierarchical structures are useful in separating the spatial location of dual cocatalysts. The core–shell structure is the most frequently employed structure among them. For example, TiO₂/In₂O₃ mesoporous hollow spheres with Pt and MnO_x loaded on the inner and outer surfaces show high activity for water oxidation because of the opposite flow direction of photogenerated electrons and holes driven by the respective cocatalyst.⁴⁴ Similarly, TiO₂ hollow spheres decorated with spatially separated MnO_x and CuPt (Fig. 7a) exhibit a high CO yield of 84.2 μmol g⁻¹ h⁻¹.⁴⁵ Furthermore, it is also possible to separate the locations of the dual cocatalysts in a 3D porous structure. TiO₂–SiO₂ with spatially separated CoO_x and Pt on the inner and outer surfaces of the skeleton (Fig. 7b) shows a good CH₄ production rate and selectivity for the photocatalytic CO₂RR in water vapor, which is higher than that of TiO₂–SiO₂ decorated with randomly scattered CoO_x and Pt (Fig. 7c).⁴⁶

Fig. 7 (a) Schematic of MnO_x@TiO₂@CuPt alloy mesoporous hollow spheres and the charge behavior. Reprinted with permission from ref. 45. Copyright 2018, Royal Society of Chemistry. (b) TEM image of 2%Pt/TiO₂–SiO₂/CoO_x; the inset is the HRTEM image of the red square in the TEM image. Reprinted with permission from ref. 46. Copyright 2016, Royal Society of Chemistry. (c) Comparison of the charge transfer route in the TiO₂–SiO₂ with separated Pt–CoO_x locations (Pt/HCTSO) and random Pt–CoO_x locations (Pt–CoO_x/HTSO). Reprinted with permission from ref. 46. Copyright 2016, Royal Society of Chemistry.

Photocatalysts with different exposed crystal facets are interesting because it is possible to separate reduction and oxidation locations on a single crystal. A facet heterojunction can be established to drive the photogenerated electrons and holes to migrate toward the respective facet owing to the different work functions of different facets. The reduction and oxidation cocatalysts can be synthesized on the corresponding facets to trap the respective charge carriers so as to further increase the charge separation efficiency. TiO₂ is a typical semiconductor that can be prepared with a co-exposed hole-rich {001} facet and electron-rich {101} facet to form the facet heterojunction. The oxidation and reduction cocatalysts are selectively deposited on the corresponding facet to trap the electrons and holes, respectively. Ag/TiO₂{101}/TiO₂{001}/MnO_x^47,48 and Pt/TiO₂{101}/TiO₂{001}/MnO_x⁴⁹ have been demonstrated to be efficient systems for the photocatalytic CO₂RR. Taking the Pt/TiO₂{101}/TiO₂{001}/MnO_x system as an example, Pt and MnO_x are selectively deposited on the respective facets of TiO₂{101} and {001} (Fig. 8a–d). TiO₂ with the dual cocatalysts shows enhanced photocatalytic CO₂RR performance (Fig. 8e) due to the efficient carrier migration induced by the facet junction and joint electron–hole extraction by the cocatalysts (Fig. 8f).

Fig. 8 (a–d) SEM images and (e) photocatalytic CO₂RR performance of TiO₂ (T), TiO₂/MnO_x (TM), TiO₂/Pt (TP) and TiO₂/MnO_x/Pt (TMP), (f) schematic diagram of the photocatalytic CO₂RR mechanism of TiO₂/MnO_x/Pt. Reprinted with permission from ref. 49. Copyright 2019, American Chemical Society.

2.3 Z-Scheme structures

Combining different semiconductors to form a heterostructure offers a feasible way to realize the photocatalysis with a tunable light response, efficient charge separation, prolonged charge carrier lifetime, and long-term stability of the photocatalysts. Typically, two types of heterostructures are most frequently used for photocatalysis, type II and Z-scheme systems. The type II heterojunction is a classical structure with enhanced catalytic performance compared to the single semiconductor, due to an extended light response and efficient spatial charge separation via the favorable staggered band alignment. However, in the type II structure the CO₂RR and OHR are respectively driven by charges with weakened reduction and oxidation power (Fig. 9a). The Z-scheme heterostructure stands out as a more efficient “artificial photosynthesis” system than the type II heterojunction, which exhibits stronger redox capability by simultaneously retaining the electrons and holes in the respective higher levels of the VB and CB (Fig. 9b). Generally, in a Z-scheme system the holes with less positive oxidation potentials recombine with the electrons with less negative reduction potentials, and thus the high redox power can be maintained. Therefore, Z-scheme systems usually deliver higher photocatalytic performance than their single material counterparts, due to long carrier lifetime, no energy loss, and high stability owing to the timely consumption of photogenerated charge carriers. Since the Z-scheme systems for the photocatalytic CO₂RR have been systematically summarized recently,¹² here we just discuss some of them to demonstrate the importance of the Z-scheme in boosting the OHR and thus enhancing the CO₂RR performance.

Fig. 9 Photocatalytic CO₂ reduction on (a) a type II heterojunction, and (b) Z-scheme system. Reprinted with permission from ref. 12. Copyright 2020, Wiley-VCH.

In a Z-scheme photocatalytic system, one semiconductor has a high CB position and another has a low VB position, making it possible to couple the CO₂RR active semiconductor with that usually active in the OHR. Thus, many Z-scheme systems can be used to photoreduce CO₂ in aqueous solution. The common Z-scheme systems include indirect and direct ones. In an indirect Z-scheme system, there is a redox mediator such as Fe³⁺/Fe²⁺ or IO³⁻/I⁻ or an electron bridge like a noble metal or RGO, while a direct Z-scheme system consists of two semiconductors in direct contact.

g-C₃N₄/Bi₄O₅I₂ with I³⁻/I⁻ as the redox mediator is an efficient system for the photocatalytic CO₂RR in H₂O.⁵⁰ CO is found to be the major reduction product with an optimized yield 7.9 and 2.3 times higher than that using pure g-C₃N₄ and Bi₄O₅I₂, respectively (Fig. 10a). Interestingly, O₂ is produced as the oxidation product over g-C₃N₄/Bi₄O₅I₂ and pure C₃N₄, while no O₂ is observed over Bi₄O₅I₂. This phenomenon indicates that the high oxidation capability is preserved in C₃N₄/Bi₄O₅I₂, which has also been confirmed via hydroxyl radical (˙OH) quantification experiments (Fig. 10b). Another example is all-solid-state Z-scheme systems with Au in between two semiconductors as the electron bridge. WO₃ is usually used as the oxidation photocatalyst, with high oxidation power but no CO₂ reduction capability. The Z-scheme charge transfer mode is realized between WO₃ and semiconductors with a high CB position for enhanced photocatalytic CO₂RR, such as CdS,⁵¹ Cu₂O⁵² and In₂S₃.⁵³ For instance, WO₃ is coupled with In₂S₃, which exhibits a high enough reduction power for the CO₂RR, to construct the WO₃/Au/In₂S₃ Z-scheme system.⁵³ The results of Kelvin probe force microscopy (KPFM) indicate that the surface potential of WO₃/Au/In₂S₃ is much higher than that of pure WO₃ and WO₃/In₂S₃, demonstrating the highest number of photogenerated holes accumulated in WO₃/Au/In₂S₃ (Fig. 11a). This work clearly indicates that the Z-scheme charge transfer is effective in promoting the high oxidation ability of WO₃ and thereby extracting more electrons from In₂S₃ for the CO₂RR (Fig. 11b).

Fig. 10 (a) Photocatalytic CO₂RR rate for g-C₃N₄/x-Bi₄O₅I₂ (x = 0, 10, 20, 50, 70 and 100), and (b) reactive oxygen species quantification experiments: TA represents ˙OH. Reprinted with permission from ref. 50. Copyright 2016, Elsevier.

Fig. 11 (a) Surface potential of samples under test (1–4 and 5 under λ = 410 and 550 nm illumination), and (b) photocatalytic CO₂RR results over different samples. Reprinted with permission from ref. 53. Copyright 2016, American Chemical Society.

Similar results are obtained over the CoZnAl-LDH/RGO/g-C₃N₄ Z-scheme, which has superior activity for both reduction and oxidation in selective conversion of CO₂ to CO and methylene blue degradation, respectively.⁵⁴ Recently, the common oxidation semiconductor α-Fe₂O₃ was combined with CsPbBr₃, using amine functionalized RGO as an efficient mediator to regulate the interfacial interactions (Fig. 12a–d). Such α-Fe₂O₃/amine-RGO/CsPbBr₃ Z-scheme system makes it possible to take advantage of the high reduction power of CsPbBr₃ for the CO₂RR and high oxidation power of α-Fe₂O₃ for the OER, resulting in significantly higher activity for the CO₂RR than that using CsPbBr₃ and α-Fe₂O₃/CsPbBr₃ with almost stoichiometric O₂ production (Fig. 12e).⁵⁵ Control experiments have confirmed that the sources of reduction and oxidation products are CO₂ and H₂O, respectively.

Fig. 12 a) SEM, (b) TEM and (c) HRTEM images of α-Fe₂O₃/amine-RGO/CsPbBr₃, (d) energy band diagram and the reaction mechanism in type II and Z-scheme heterojunctions (pH = 0), and (e) photocatalytic CO₂RR performance after 15 h. Reprinted with permission from ref. 55. Copyright 2020, Elsevier.

The direct Z-scheme system consisting of two semiconductors is more easily achieved considering the preparation simplicity. It is to be noted that it is always necessary to verify the charge carrier migration path to differentiate it from the type II heterojunction. Although CdS/Au/TiO₂ is an efficient Z-scheme system,⁵⁶ it is also possible to directly construct an interface between CdS and TiO₂ that follows the Z-scheme charge transfer mechanism. For example, Low et al. reported a direct Z-scheme TiO₂/CdS film.⁵⁷ Both in situ irradiated X-ray photoelectron spectroscopy (ISI-XPS) and photoluminescence (PL) spectroscopy are used to prove the Z-scheme charge migration in TiO₂/CdS that exhibits much higher performance for the photocatalytic CO₂RR in H₂O, compared to that of pure CdS, TiO₂ and commercial P25 (Fig. 13a). The binding energy of Ti 2p and Cd 3d respectively shows a blue and a red shift under illumination compared to that without illumination, corresponding respectively to decreased and increased electron density under illumination (Fig. 13b and c). Moreover, probing the oxidation capability is another means to study the Z-scheme transfer path. With the help of a probe molecule for ˙OH detection, the PL spectrum results indicate significantly enhanced production of oxidative ˙OH in TiO₂/CdS, but a very low PL signal for CdS due to its less positive VB potential for ˙OH radical generation (Fig. 13d). Thus, the holes accumulated in this composite come from Z-scheme transfer rather than the type II mode (Fig. 13e and f). Another structure, InVO₄/β-AgVO₃, is also proved by ISI-XPS to be an effective Z-scheme system for the photocatalytic CO₂RR with O₂ as the oxidation product from H₂O. The carbon source of the reduction products and oxygen source of the obtained O₂ was confirmed by isotopically labelled ¹³CO₂ and H₂¹⁸O, respectively.⁵⁸

Fig. 13 (a) The photocatalytic CO₂RR over CdS, TiO₂, TiO₂/CdS and P25, high resolution XPS spectra of (b) Ti 2p and (c) Cd 3d under irradiation, (d) PL intensity of 7-hydroxycoumarin over CdS, TiO₂ and TiO₂/CdS under illumination, and comparison of the charge carrier migration paths following (e) type II and (f) Z-scheme mechanisms for TiO₂/CdS. Reprinted with permission from ref. 57. Copyright 2018, Wiley-VCH.

3D-SiC@2D-MoS₂ is another efficient direct Z-scheme system for simultaneous photocatalytic CO₂ reduction to CH₄ and H₂O oxidation to O₂.⁵⁹ The high performance is not only due to the Z-scheme charge separation, but also the high hole mobility of MoS₂ and electron mobility of SiC that respectively accelerate the water oxidation and CO₂ reduction, with ¹⁸O and ¹³C tracer experiments confirming the product origin.

Apart from the favorable charge separation and driving force, the Z-scheme system is also effective in improving the photocatalyst stability. Among the semiconductors active for the CO₂RR, the narrow bandgap ones that possess high reduction power usually undergo self-oxidation by the photogenerated holes and rapid charge recombination, such as metal sulfides and Cu₂O. The Z-scheme mechanism provides a recombination drive-thru for the photogenerated charge carriers with inferior redox potentials, resulting in fast consumption of the holes and accordingly protecting the photocatalyst. The stability of Cu₂O is enhanced by constructing a Z-scheme system with TiO₂, and the obtained Cu₂O/TiO₂ showed a 4 times higher production rate of CO than pure Cu₂O.⁶⁰ CuGaS₂/RGO/TiO₂ is another efficient and stable Z-scheme system for the photocatalytic CO₂RR, with a stoichiometric ratio of O₂ and reduction products.⁶¹

Besides the systems mentioned in Sections 2.1, 2.2 and 2.3, other strategies have also been considered to tune the hole behavior. For instance, the formation of a Schottky junction between Au and a p-GaN photocathode can favor the collection of plasmon-induced hot holes with strong oxidation power, resulting in enhanced activity and selectivity of CO production from photoelectrochemical CO₂ reduction in aqueous electrolyte.⁶²

Despite the remarkable progress in tuning hole behavior through photocatalyst design like doping, loading cocatalysts and establishing Z-scheme structures, there are still obstacles that remain to be overcome. For example, compared to the large number of reductive semiconductors and cocatalysts for the CO₂RR, only a few oxidative candidates are available to construct effective Z-scheme or semiconductor/cocatalyst systems. Moreover, the activity of oxidation cocatalysts is still not equivalent to that of the reduction cocatalysts, which needs further study to balance the reduction and oxidation performance.² In addition, the contact between different materials in a composite photocatalyst is always important for efficient electron–hole transfer. Thus, special attention should also be paid to the rational design of the interface between semiconductors and/or cocatalysts and deliberate consideration of the lattice mismatch, as well as an in-depth understanding of the structure–property relationship.

3 Electron donors

To prevent the accumulation of holes on the photocatalyst surface and to facilitate the charge separation and CO₂RR as well, electron donors are essential for donating electrons and protons during the reactions. Generally, electron donors exhibit different donating properties, and range from strong donors like triethanolamine (TEOA) to weak donors like H₂O. The CO₂RR is an uphill reaction (ΔG^θ > 0) when H₂O serves as the electron donor, making the overall reaction “nonsacrificial”, while the adoption of a strong electron donor can make the overall reaction “sacrificial” (ΔG^θ < 0).⁶³

3.1 H₂O

H₂O is the most frequently adopted electron donor for CO₂ conversion to value-added chemicals since it provides a “green chemistry” route, given that no external sacrificial reagents are consumed.⁶⁴ Ideally, the reaction between H₂O and holes can generate O₂ or H₂O₂, as well as protons that can also participate in the CO₂RR.⁶⁵

In terms of standard electrode potentials of the reduction and oxidation half reactions, one can deduce the minimum driving force for the CO₂RR in H₂O (eqn (1)–(5)), where ΔG^θ is the free energy and ΔE^θ is the standard redox potential.^66,67 Taking the conversion of CO₂ into CO as an example, it requires a bandgap value of at least 1.33 eV for a photocatalyst to drive this reaction (Eqn (1)).

CO₂ → CO + O₂, ΔG^θ = 257 kJ mol⁻¹, ΔE^θ = 1.33 V (1)

CO₂ + H₂O → HCOOH + 1/2O₂, ΔG^θ = 286 kJ mol⁻¹, ΔE^θ = 1.48 V (2)

CO₂ + H₂O → HCHO + O₂, ΔG^θ = 522 kJ mol⁻¹, ΔE^θ = 1.35 V (3)

CO₂ + 2H₂O → CH₃OH + 3/2O₂, ΔG^θ = 703 kJ mol⁻¹, ΔE^θ = 1.21 V (4)

CO₂ + 2H₂O → CH₄ + 2O₂, ΔG^θ = 818 kJ mol⁻¹, ΔE^θ = 1.06 V (5)

It seems that the bandgap requirement is not very harsh. However, considering that the power of both electrons and holes should respectively meet the CO₂RR and OER potential, and an overpotential is usually required, suitable semiconductors for the CO₂RR in H₂O usually have a wide bandgap. Pioneering work for the photocatalytic CO₂RR in H₂O was reported in 1979, in which several photocatalysts (including TiO₂, ZnO, CdS, GaP and SiC) were used to reduce CO₂ to organic compounds in H₂O.⁶⁸ Substantial efforts have been devoted thereafter to photocatalysts for the CO₂RR with H₂O as the electron donor, including the aforementioned TiO₂,^69,70 and other materials like K₂Ti₆O₁₃,⁷¹ BiOCl,⁷² Mg–In layered double hydroxides⁷³ and g-C₃N₄.⁷⁴

The efficiency for photocatalytic CO₂ reduction with H₂O has been unsatisfactory hitherto, making it difficult to reproduce reliable results. Moreover, even though it is well accepted that the CO₂RR is accompanied by the OER from the water oxidation, the activity is often evaluated only by the CO₂RR product yield. In many cases, the evolution of O₂ is not observed or reported, leading to vagueness in the consumption of photogenerated holes or the stability of the semiconductors, since holes can be consumed by either the electron donors, photocatalysts, or residual carbon species. In this respect, the study of the OER is necessary and evidence is required to demonstrate that the produced O₂ is generated from H₂O oxidation. The isotope tracer experiment using H₂¹⁸O as the reagent is one method to verify that H₂O is the oxygen source.^75,76 Another method is to calculate the ratio of reacted electrons to holes, which should be unity to indicate that H₂O is the sole electron donor (i.e. e⁻/h⁺ = 1).⁷⁷

Stoichiometric chemistry is always a challenge for the CO₂RR in H₂O, even though overall H₂O splitting has been widely proved to be feasible in which O₂ and H₂ are evolved simultaneously. It is difficult to ideally couple the CO₂RR with H₂O oxidation since it requires high reduction and oxidation power for the photogenerated charge carriers to drive the two half reactions, and there is a competitive H₂O reduction as well.

So far most of the research has only focused on characterizing the reduction products. The O₂ amount is disregarded in most cases, or no stoichiometric ratio of the reduction products/O₂ is reached. The first report on the CO₂ photoconversion accompanied by the OER dates back to 1992 over the CeO₂/TiO₂ composite.⁷⁸ So far, only a limited number of materials have clearly been demonstrated to be capable of producing O₂ as the oxidation product. Table 1 shows those that can realize the photocatalytic CO₂RR in H₂O with simultaneous production of O₂ and CO₂RR products.

Table 1 Photocatalysts for photocatalytic reduction of CO₂ in H₂O

No.	Photocatalyst	Weight/g	Bandgap/eV	Reduction products^a	Oxidation products (O2)^a	e^−/h^+^b	Conditions (solution & illumination)	Ref.
				Unit is μmol g h^−1 unless otherwise noted	Unit is μmol g h^−1 unless otherwise noted
a The amount of reduction and oxidation products is converted to the same unit by processing the original data, which are just used for reference. It is meaningless to directly compare them, as the experimental conditions can be quite different among them, such as the catalyst amount, electrolyte, illumination source and reaction time, all of which have a strong impact on evaluating the photocatalysts. b e^−/h^+ = (number of electrons consumed for reduction products)/(number of holes consumed for O2 formation). c ‘—’ means that the corresponding data are not mentioned in the related reference.
1	1.0 wt% Cu/ZrO2	1.0	5.0	CO: 2.5 ± 0.1, H2: 19.5 ± 0.3	10.8 ± 0.2	1.02	NaHCO3 aqueous solution, 400 W high-pressure Hg lamp	(1993)^101
2	2.0 wt% Ag/BaLa4Ti4O15	0.3	3.9	CO: 73.33, HCOOH: 2.33, H2: 33.33	53.33	1.02	H2O, 400 W high-pressure mercury lamp	(2011)^79
3	1.0 wt% Ag/La2Ti2O7	1.0	4.05	CO: 5.2, H2: 4.9	5.3	0.95	H2O, 400 W high-pressure mercury lamp	(2015)^102
4	0.5 wt% Ag/CaTiO3	0.2	3.6	CO: 1.75, H2: 2.15	1.8	1.08	NaHCO3 aqueous solution (1.1 mol L^−1), 300 W xenon lamp	(2015)^103
5	3.5 wt%Ag/CaTiO3	0.3	3.6	CO: 180, H2: 10.33	83.33	1.14	NaHCO3 aqueous solution (1 mol L^−1), 100 W high pressure mercury lamp	(2017)^104
6	1.0 wt% Ag/4 mol% Al–SrTiO3	0.5	—	CO: 14.4, H2: 0.3	7.16	1.03	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure Hg lamp (λ > 300 nm)	(2020)^19
7	0.5 wt% Ag/SrNb2O6 nanorods	0.5	3.86	CO: 102.4, H2: 2.2	49.6	1.05	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure Hg lamp	(2017)^105
8	KTaO3	0.1	3.7	CO: 61.98 ppm g^−1 h^−1, H2: 1323.00 ppm g^−1 h^−1	501.87 ppm g^−1 h^−1	1.38	H2O, 300 W xenon arc lamp	(2014)^106
9	0.5 wt% Ag/KCaSrTa5O15	0.5	4.1	CO: 16.2, H2: 112	74	0.83	H2O, 400 W high-pressure mercury lamp	(2014)^80
10	0.5 wt% Ag/KCaSrTa5O15	0.5	4.12	CO: 194, H2: 30	92	1.22	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2015)^81
11	(a) 2.0 wt% Ag/NaTaO3:Ca (5.0%)	0.25–0.5	4.1	(a) CO: 394.67, H2: 40	(a) 224	(a) 0.97	H2O, 400 W high-pressure Hg lamp	(2017)^82
	(b) 2 wt% Ag/NaTaO3:Sr (5%)			(b) CO: 469.33, H2: 74.67	(b) 272	(b) 1.00
	(c) 3.0 wt% Ag/NaTaO3:Ba (5.0%)			(c) CO: 333.33, H2: 64 (0.375 g for calculation)	(c) 202.67	(c) 0.98
12	3.0 wt% Ag/ZnTa2O6	0.5	4.5	CO: 38.6, H2: 50.2	37.2	1.19	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure Hg lamp	(2016)^107
13	Ag/SrO/Ta2O5	1.0	—	CO: 6.8, H2: 3.8	5.1	1.04	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2015)^108
14	1.0 wt% Ag–Sr2KTa5O15	1.0	4.0	CO: 65.5, H2: 8	35	1.05	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2016)^109
15	1.0 wt% Ag/Sr1.6K0.37Na1.43Ta5O15	1.0	3.95	CO: 94.6, H2: 16	53.8	1.03	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2017)^110
16	1.0 wt% Ag/K2YTa5O15	1.0	3.86	CO: 91.9, H2: 16.3	43.2	1.25	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2018)^111
17	0.1 wt% Ag/Ga2O3	0.2	4.1	CO: 10, H2: 47	23	1.24	NaHCO3 aqueous solution (1 mol L^−1), 300 W xenon lamp	(2015)^112
18	Ag/Zn-Doped Ga2O3	1.0	—	CO: 117.0, H2: 16.9	70.1	0.96	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2014)^113
19	1.0 wt% Ag/Ga2O3 with a (3.0 mol%) ZnGa2O4 layer	1.0	Ga2O3: 4.6, ZnGa2O4: 4.2	CO: 108, H2: 8.9	60.6	0.96	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2016)^114
20	Ag/3.0 mol% Yb-3.0 mol% Zn/Ga2O3	0.5	4.68	CO: 300, H2: 75.2	206	0.91	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W Hg lamp	(2017)^115
21	Ag/3.0 mol% Pr/Ga2O3	0.5	—	CO: 498, H2: 129.4	300.8	1.04	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2017)^116
22	0.25 wt% Ag/95-MgAl LDH/Ga2O3	1.0	—	CO: 211.7, H2: 130	169	1.01	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2017)^117
23	1.0 mol% Ag-1.0 mol% Cr/Ga2O3	0.5	—	CO: 960, H2: 185.8	562	1.02	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2018)^118
24	0.25 mol% Ag@Cr/Ga2O3	0.5	—	CO: 1050.6, H2: 180	650	0.95	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2018)^119
25	Ag–Cr/Ga2O3:Rh (0.7 mol%)	0.5	Ga2O3: 4.7, Ga2O3:Rh: 3.6 & 2.9	CO: 7.8, H2: 20.6	13	1.09	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp (λ > 300 nm)	(2018)^120
26	ZnGa2O4 nanocubes	0.1	4.4	CH4: 1.6	12.0	0.27	H2O vapor, UV-enhanced (200 to 350 nm) 300 W xenon arc lamp	(2013)^7
27	1.0 wt% Ag/ZnGa2O4	1.0	4.67	CO: 155, H2: 8.5	74.3	1.10	NaHCO3 aqueous solution (0.1 mol L^−1), 400 W high-pressure mercury lamp	(2015)^121
28	Mg–In LDH	0.1	—^c	CO: 3.21	17	0.09	H2O, 200 W Hg–Xe lamp	(2012)^73
29	Bi2O2(OH) (NO3)–Br	0.02	3.15	CO: 6.5	3.0	1.08	H2O vapor, 300 W xenon lamp	(2019)^122
30	InVO4 sheets	0.1	2.3	CO: 18.3, CH4: 0.3	8.5	1.14	H2O vapor, xenon lamp	(2019)^85
31	InVO4/β-AgVO3	0.05	β-AgVO3: 2.15, InVO4: 2.36	CO: 12.61, H2: 0.13	5400	0.0012	H2O vapor, 300 W xenon lamp (λ > 420 nm)	(2019)^58
32	WO3	—	2.36	CO: 2.8	1.7	0.85	H2O, 40 W silicon nitride lamp (0.8 to 17 μm)	(2018)^86
33	WO3/Au/In2S3	10 cm^2		CH4: 0.42	0.86	0.98	H2O vapor, 300 W xenon arc lamp (λ > 420 nm)	(2016)^53
34	Hollow multi-shelled structures of Co3O4 dodecahedron	0.005	2.33	CO: 46.3	22	1.05	H2O vapor, 200 W Xe lamp with a standard AM 1.5 filter and a 780 nm reflector	(2019)^123
35	CoN	—	1.44	CO: 0.29	0.15	0.97	H2O, xenon lamp (λ > 800 nm)	(2020)^87
36	Mg2(DOBDC)/TiO2, DOBDC = 2,5-dioxido-1,4-benzenedicarboxylate	0.01	—	CO: 4.1, CH4: 2.4	6.3	1.07	H2O vapor, four 4 W UV lamps with a wavelength centered at 365 nm	(2016)^88
37	TTCOF–Zn	0.01	1.49	CO: 2.06	0.98	1.05	H2O, 300 W xenon lamp (λ = 420–800 nm)	(2019)^89
38	BiOI/g-C3N4	0.1	BiOI: 1.75, g-C3N4: 2.69	CO: 4.86, CH4: 0.18	0.78	3.58	H2O vapor, 300 W xenon arc lamp (λ > 400 nm)	(2016)^65
39	Bi4O5I2/g-C3N4	0.1	Bi4O5I2: 1.45, g-C3N4: 2.31	CO: 45.6, CH4: 6, H2: 2.2	9	3.99	H2O vapor, 300 W xenon lamp (λ > 400 nm)	(2016)^50
40	Bi2WO6/RGO/g-C3N4	0.05	g-C3N4: 2.65, Bi2WO6: 2.82	CO: 16, CH4: 2.5, H2: 2.25	10.5	1.35	H2O vapor, 300 W xenon arc lamp (λ > 420 nm)	(2018)^124
41	Nb–TiO2/g-C3N4	0.1	Nb–TiO2: 2.91, g-C3N4: 2.68	CH4: 562, CO: 420, HCOOH: 698	1702	0.99	H2O vapor, two 30W white bulbs	(2019)^125
42	CuGaS2/RGO/TiO2	0.1	CuGaS2: 2.3	CO: 1.5, H2: 288	112	1.29	H2O, 300 W xenon lamp (λ > 330 nm)	(2017)^61
43	Cl-Doped Cu2O nanorods	—	Cu2O–Cl-4: 2	CO: 1.74 μmol cm^−2, CH4: 0.39 μmol cm^−2	1.41 μmol cm^−2	1.17	H2O vapor, 350 W xenon lamp (λ > 420 nm)	(2019)^28
44	CuxO/SrTiO3/CoPi	—	SrTiO3: 3.2	CO: 0.225 μmol cm^−2, H2: 0.06 μmol cm^−2	0.155 μmol cm^−2	0.92	Aqueous solution containing KHCO3 (0.5 mol L^−1), KH2PO4 (0.025 mol L^−1), and K2HPO4 (0.025 mol L^−1), Hg- xenon lamp (UV light irradiation)	(2017)^39
45	Cu2O/WO3	0.085	WO3: 2.75, Cu2O: 2.05	CO: 137.65, H2: 67.06	8.24	12.4	H2O vapor, 300 W xenon arc lamp (λ > 400 nm)	(2019)^52
46	α-Fe2O3/amine-RGO/CsPbBr3	—	CsPbBr3: 2.36, α-Fe2O3: 2.2	CH4: 9.45, CO: 2.36, H2: 0.29	18.67	1.08	H2O vapor, 150 W xenon lamp (λ > 420 nm)	(2020)^55
47	Fe2V4O13/RGO/CdS	0.025	Fe2V4O13: 1.83, CdS: 2.4	CH4: 156	92	3.39	H2O vapor, 300 W xenon lamp	(2015)^126
48	SiC@MoS2	0.01	SiC: 2.48, MoS2: 1.77	CH4: 323 μL g^−1 h^−1	621 μL g^−1 h^−1	1.04	H2O vapor, 300 W xenon lamp (λ ≥ 420 nm)	(2018)^59

---

It is to be noted that Kudo et al. have reported a series of Ag loaded ALa₄Ti₄O₁₅ (A = Ca, Sr and Ba) photocatalysts with a wide bandgap, showing simultaneous CO₂ reduction and H₂O oxidation. Specifically, Ag/BaLa₄Ti₄O₁₅ exhibits selective CO₂ reduction over H₂O splitting with stoichiometric O₂ production (Fig. 14).⁷⁹ Ag/KCaSrTa₅O₁₅ (ref. 80 and 81) and Ag/(Ca, Sr, Ba) doped NaTaO₃ (ref. 82) also show good performance for CO₂ reduction with the ratio of reacted electrons/holes as unity. Tanaka et al. have developed a bunch of metal oxides modified by Ag or doped with other metals, including Ga₂O₃, ZnGa₂O₄, Ta₂O₅, ZnTa₂O₆, La₂Ti₂O₇, SrNb₂O₆, Sr₂KTa₅O₁₅, K₂YTa₅O₁₅, layered double hydroxides (LDHs) and SrTiO₃, which can realize stoichiometric CO₂ reduction and H₂O oxidation (see Table 1).

Fig. 14 Photocatalytic CO₂RR performance over Ag/BaLa₄Ti₄O₁₅: amounts of H₂ (○), O₂ (●), and CO (△) produced. Reprinted with permission from ref. 79. Copyright 2011, American Chemical Society.

Even though some materials have been reported to show CO₂ reduction activity in H₂O, not all of them exhibit higher selectivity for CO₂ reduction than H₂ evolution, as listed in Table 1. Moreover, most of the materials in Table 1 possess a wide bandgap so as to meet the requirement for reduction and oxidation potentials, while they are only responsive to UV light that accounts for ca. 4% of the total solar energy. Recently, some efforts have been devoted to developing photocatalysts with an extended light-response spectrum. Zou et al. reported simultaneous photocatalytic conversion of CO₂ and H₂O under visible-light irradiation over ultrathin WO₃ nanosheets⁸³ and over Bi₂WO₆ nanoplates,⁸⁴ while no O₂ evolution information was reported. They also prepared atomically thin InVO₄ nanosheets (E_g = 2.3 eV) with highly selective CO₂ reduction performance, exhibiting almost stoichiometric CO and CH₄ with O₂.⁸⁵ Furthermore, considering that infrared (IR) light occupies around 50% of the solar radiation, IR-activated CO₂ reduction in H₂O is appealing. Xie et al. reported IR-driven CO₂ reduction to CO and O₂ over oxygen-deficient WO₃ atomic layers with an intermediate band (Fig. 15a–f).⁸⁶ They also designed metallic CoN porous atomic layers with a narrow bandgap (E_g = 1.44 eV) for CO₂ reduction in H₂O under IR light illumination, displaying a 2 : 1 ratio of CO to O₂, but the evolution rate is very low without using sacrificial reagents (CO: 0.29 μmol g⁻¹ h⁻¹, O₂: 0.15 μmol g⁻¹ h⁻¹).⁸⁷ Hence, further development of active photocatalysts with a broadband spectrum response is imperative. Except for the aforementioned photocatalysts, other materials and structures have also shown successful CO₂RR and OER, such as lead halide perovskites,⁷⁵ metal–organic frameworks (entry 36, Table 1)⁸⁸ and covalent organic frameworks (entry 37, Table 1).⁸⁹ In order to tune the properties of the photocatalyst system in a more flexible way, heterostructures consisting of different materials are employed (e.g. entries 33, 38, 42, 45, 47 and 48 in Table 1), which can either extend the light response region, favor the charge separation, or meet the requirement for the CO₂ reduction and H₂O oxidation reaction simultaneously. Some of these systems have been analyzed in detail in Section 2.

Fig. 15 (a) TEM and (b) HRTEM images, (c) height profiles and the corresponding (d) AFM image, (e) schematic electronic band structure, and (f) photocatalytic CO₂RR activity of oxygen vacancy-rich WO₃ atomic layers in H₂O. Reprinted with permission from ref. 86. Copyright 2018, Elsevier.

Although the stoichiometric generation of reduction and oxidation products is an important indicator for authentic photocatalytic CO₂RR in H₂O, it is to be noted that sometimes the OER cannot be successfully observed, albeit CO₂ is indeed converted to the chemicals.⁸ First, it is possible that back reactions occur between the CO₂RR products and O₂, but in this case both reduction products and O₂ will be simultaneously consumed. Secondly, free O₂ can be adsorbed on the photocatalyst surface in the presence of H₂O, which makes it difficult to determine the actual amount of O₂ produced.^90–92 It is also evidenced by theoretical calculations that the adsorption affinity of O₂ on C₃N₄ is larger than that of CO.⁹³ Especially, oxygen vacancies on the catalyst surface can favor the adsorption of O₂ and O_ad species⁹⁴ and formation of O²⁻ or H₂O₂via reactions with electrons,^95,96 and thus prevent O₂ evolution from the solution. Alternatively, O₂ molecules can further form a peroxide complex, except that some of them are adsorbed as physisorbed, chemisorbed and O²⁻ species.^97,98 It is worth mentioning that the surface peroxo species may also hold back the holes for H₂O oxidation and increase the electron–hole recombination possibility. Moreover, the holes can be consumed by other chemical species besides H₂O. Organic surfactant is frequently used during the preparation of the photocatalysts, resulting in the inevitable carbon residuals on the surface. It is possible that the organic residuals (e.g. PVP) on the photocatalysts can consume the photogenerated holes,⁹⁹ resulting in no O₂ generated at the early stage. Evolved O₂ or photogenerated holes with lower quasi Fermi energy than the oxidation potential of the photocatalysts can also react with the photocatalysts, leading to changes in the catalysts upon the reaction.^78,100 Moreover, the detection accuracy of O₂ also needs to be considered since there is probable O₂ leakage from the atmosphere, which can be reasonably deduced from the accompanying N₂ amount also from the atmosphere.

Therefore, achieving stoichiometric CO₂ reduction and H₂O oxidation is still a challenge for many reasons, such as the photocatalytic activity and other chemical species that may consume the photogenerated holes, which need to be distinguished when analyzing the system activity.

3.2 Sacrificial electron donors

As mentioned above, considering all the requirements for a photocatalyst, as well as the competitive H₂O reduction, it is a great challenge to close the photosynthetic cycle of CO₂ reduction and H₂O oxidation, especially the conversion of CO₂ into four-electron (and above) reduction products. Moreover, when H₂O serves as the electron donor, the resultant O₂ or hydroxyl radicals (˙OH) may partially participate in oxidizing the reduction products like formic acid and/or the CO₂RR intermediates. The VB position for many semiconductors with a narrow bandgap is usually unsuitable for the H₂O oxidation, making it difficult to oxidize H₂O, but easy to undergo self-photocorrosion.^127,128 Hence, the H₂O oxidation half reaction can be achieved by surrogate reactions with sacrificial electron donors that usually possess good electron donating properties, which can pick up the photogenerated holes quickly and thus improve the efficiency and photocatalyst stability, such as inorganic anions (e.g. SO₃²⁻, Cl⁻) and organic compounds (e.g. triethylamine (TEA), 2-propanol).¹²⁹ In addition, the adoption of sacrificial electron donors also allows the utilization of semiconductors with low stability like non-oxides (i.e., CdS, CdSe, GaAs, GaP, etc.) by preventing them from being oxidized.

3.2.1 Influence factors on the activity of sacrificial electron donors. It is crucial to understand the influence mechanism of the sacrificial electron donor before analyzing its application. The redox potential is usually an important factor that reflects the electron donating ability. Theoretically, sacrificial electron donors should have a more negative or less positive redox potential than that of H₂O to preferentially capture the photoinduced holes. Fig. 16a displays the oxidation potential of several common sacrificial electron donors. Berr et al. proposed an important concept that the redox potential at a high position on the scale (weak electron binding ability) can result in fast hole consumption and thus has a great impact on the photocatalytic activity.¹³⁰ Among the four electron donors of SO₃²⁻, TEA, EDTA⁴⁻ and methanol that are used in the Pt/CdS photocatalytic system, the inorganic SO₃²⁻ is the most efficient for H₂ generation and helps to maintain the photocatalyst stability as it has the highest oxidation potential position.¹³⁰ Another study on the photocatalytic reduction of bicarbonate over Cu₂O using SO₃²⁻, TEA and EDTA as the electron donors also follows the same trend (Fig. 16b).¹³¹

Fig. 16 (a) Oxidation potential for SO₃²⁻, TEA, EDTA, 2-propanol, ethylene glycol and glycerol, and (b) formate production from bicarbonate photocatalytic reduction over Cu₂O in the above sacrificial electron donors. Reprinted with permission from ref. 131. Copyright 2018, Cambridge University Press.

Besides the thermodynamic issue of the oxidation potential, kinetic factors also play a critical role in the proton dissociation process for the electron donors, which can be determined by the number of available hydroxyl groups. For example, glycerol is considered the most efficient electron donor among several candidates with one secondary and two primary alcohol groups, which are the potential sites for oxidation. The results of ultrafast transient absorption spectroscopy indicate that the hole scavenging efficiency of glycerol (rate constant k_s = 1.0 × 10¹⁰ M⁻¹ s⁻¹) is higher than that of methanol (k_s = 1.4 × 10⁹ M⁻¹ s⁻¹), formaldehyde (k_s = 3.1 × 10⁹ M⁻¹ s⁻¹) and H₂O.¹³² A DFT study suggests that the scavenging power of five different organic compounds follows the order of glycerol > tert-butanol > 2-propanol > methanol > formic acid.¹³³ Moreover, for the photocatalytic experiments over Cu₂O, TiO₂ and ZnS, glycerol is also more efficient in formate production compared to ethylene glycol and 2-propanol that possess similar oxidation potentials.^131,134

However, it is not always true that the activity of electron donors follows the above rules derived from the influence of redox potential and number of oxidation sites. For instance, unlike the trend mentioned above, it is reported that methanol shows higher electron donor ability than glycerol in a TiO₂/NiS_x system.¹³⁵ Moreover, Yanagida et al. proposed a joint influence from oxidation potential and CO₂ adsorption of different aliphatic amine electron donors on CO production over CdS.¹³⁶ Specifically, for TEA, Et₂NMe and Et₂NH with similar numbers of alkyl groups, the increased CO production yield follows the decreasing order of oxidation potential value: TEA > Et₂NMe > Et₂NH. For TEA (0.66 V vs. SCE), (N–Pr)₃N (0.64 V vs. SCE) and (n-Bu)₃N (0.62 V vs. SCE) with similar oxidation potentials, the CO yield presents a reverse trend to the ease of oxidation, with (n-Bu)₃N showing the lowest performance due to the hindering effect on CO₂ adsorption from the increased number of alkyl groups.

Therefore, choosing an appropriate sacrificial electron donor is complicated since its activity can be influenced by various factors. The photocatalytic CO₂RR is a very complex process and the reaction conditions can be different among systems (such as photocatalyst, solvent, temperature, etc.), which can influence the charge transfer process and/or trapping ability of the electron donors. For instance, since the VB positions of distinct photocatalysts are different, the choice of electron donors is diverse considering the oxidation potential. Moreover, the solvent also has an impact on screening the electron donors as the interactions between electron donors and solvent may influence the activity. Currently, electron donors that are universally effective for all systems are still unavailable. Some common sacrificial electron donors and the corresponding applications are presented in this section.

3.2.2 Organic sacrificial electron donors. Since organic compounds such as tertiary amines and alcohols are strong electron donors, they are usually favorable for improving conversion efficiency. Tertiary amines, especially TEOA and TEA, are demonstrated to be very efficient and popular electron donors that undergo single-electron oxidation by trapping holes to form radical cation TEOA˙⁺ or TEA˙⁺. Theoretical calculations indicate that during photocatalysis one hydrogen relocates from an α carbon of radical cation TEA˙⁺ to the oxygen in radical anion CO˙⁻ to form an Et₂N⁺=CHCH₃/HCO²⁻ ion pair, and then form Et₂NCH=CH₂ and HCO₂H.¹³⁷ A slightly different oxidation process for TEA over CdS is presented in entry 1 of Table 2, indicating that TEA serves as an electron and proton donor, and is finally decomposed into Et₂NH and CH₃CHO. With a strong ability for trapping holes, tertiary amines are widely used in semiconductor systems like Pt/C–In₂O₃,¹³⁸ CdS,^136,139 metal complex/semiconductor^140,141 and COF.¹⁴² Specifically, a suitable electron donor like TEOA is required for metal complex photocatalysts since they cannot directly extract electrons from H₂O.¹⁴⁰ Moreover, the electron donor can also influence the CO₂RR selectivity as evidenced by the work on tertiary amines. CO₂ is reduced over the CoTe photocatalyst to a small amount of CH₄ in H₂O, while more CO is generated when TEOA acts as the electron donor, which is probably due to the low CO adsorption on CoTe in the presence of TEOA.¹⁴³ In addition, the CO₂RR selectivity may also be influenced by the pH of solution caused by different electron donors.¹⁴⁴ Maeda et al. showed that the main product over RuP/C₃N₄ hybrids is formic acid in a mixed DMA and TEOA basic solution,¹⁴⁵ while it is CO in DMA and methanol acidic solution.⁶³ Moreover, the selectivity of the CO₂RR product can also be influenced by the nature of tertiary amines. Lehn et al. tested the photochemical reduction of CO₂ and H₂O over the Ru(2,2′-bipyridine)₃²⁺/Co²⁺ system using different tertiary amines as electron donors and the product ratio CO/H₂ increases from 0.79 to 400, following the order of trimethylamine, TEA, tripropylamine, and TEOA.¹⁴⁶ They suggested that these tertiary amines may influence the reaction by coordination to cobalt. Still, no detailed mechanism is available on the influence of electron donors on the photocatalytic process.

Table 2 Possible reactions of several electron donors

No.	Electron donor	Reaction	Ref.
1	TEA		136
2	2-Propanol	(CH3)2CHOH + 2h^+ → (CH3)2CO + 2H^+	148, 149 and 156
3	Cyclohexanol	C6H12O + 2h^+ → C6H10O + 2H^+	150
4	SO3^2−	2h^+ + Na2SO3 + H2O → Na2SO4 + 2H^+	155
5	Cl^−	2Cl^− + 2h^+ → Cl2, Cl2 + H2O → HClO + HCl	157
6	NH3	2/3NH3 → H2 + 1/3N2	158

---

2-Propanol is claimed to be a better hole scavenger than methanol,¹⁴⁷ with acetone as the photooxidation product (entry 2, Table 2). It is to be noted that the CO₂ reduction product yield initially increases with increasing concentration of 2-propanol, and then decreases at higher concentration. This optimized scavenger concentration reflects the less hindered CO₂ adsorption due to the large scavenger amount.^148,149

Cyclohexanol can function as both solvent and electron donor, due to its better CO₂ solubility and less positive oxidation potential (+0.85 V vs. NHE) than that of water oxidation.¹⁵⁰ Song et al. studied the photocatalytic CO₂RR activity in cyclohexanol over a series of materials, such as CdS/TiO₂,¹⁵⁰ ZnFe₂O₄/TiO₂ (ref. 151) and ZnFe₂O₄/BiOCl.¹⁵² Formic acid is the reduction product, which further reacts with cyclohexanol to produce cyclohexyl formate, and cyclohexanol is simultaneously oxidized to cyclohexanone by the photogenerated holes (entry 3, Table 2).

3.2.3 Inorganic sacrificial electron donors. The adoption of organic compounds may cause environmental concern and vagueness in terms of the carbon source of CO₂RR products, as most organic sacrificial reagents are valuable chemicals with high energy density and usually undergo irreversible oxidation during the photocatalysis. Accordingly, inorganic candidates are more desirable. Various inorganic electron donors have been used to scavenge the photogenerated holes hitherto, such as H₂PO₂⁻ and sulfur-containing species (e.g. SO₃²⁻, S₂O₃²⁻ and S²⁻).

The low-cost Na₂SO₃ is often employed in photocatalytic systems that are at the risk of photooxidation like Cu₂O,^36,127 CdS/Cu₂S¹⁵³ and ZnS,¹⁵⁴ with Na₂SO₄ produced (entry 4, Table 2). An enhanced yield of the photocatalytic CO₂RR product is achieved over the TiO₂–CoPc system with Na₂SO₃ as the electron donor compared to that without Na₂SO₃.¹⁵⁵ Generally, a higher concentration of Na₂SO₃ leads to a more effective photoreduction reaction, while a concentration plateau may exist sometimes.¹⁵⁵ Moreover, the concentration of Na₂SO₃ also plays an important role in the reduction product selectivity. With P25 as the photocatalyst, gas products (H₂ and CO) decrease with the increase of Na₂SO₃ concentration (1.66 to 6.68 g L⁻¹), while formic acid yield follows the opposite trend with increments of up to 9 times from low to high Na₂SO₃ concentration.¹⁵⁹ This may be because the Na₂SO₃ hole scavenger can prohibit the oxidation of the obtained organic products and thus favor the further formation of formic acid. In the case of Na₂SO₃ with low concentration, as soon as Na₂SO₃ runs out, the generated organic products will act as scavengers with gas products evolved.^160,161 So the photoreduction products can be tuned by adjusting the electron donor concentration.

The S²⁻ ion is another popular electron donor. Xie et al. analyzed the carrier dynamics with and without the S²⁻ hole scavenger by femtosecond time-resolved transient absorption (fs-TA) spectroscopy, indicating that the charge recombination time in Na₂S solution increases by 1.6-fold than that in H₂O.⁸⁷ Hence the lifetime of electrons is prolonged, as the photogenerated holes are consumed via the reaction with S²⁻ ions (2S²⁻ + 2h⁺ → S₂²⁻), leading to an increment of CO evolution rate by 50 times in Na₂S solution compared to that in H₂O.

Halide anions can also trap the photogenerated holes. For instance, Cl⁻ has been used to improve the conversion efficiency in CO₂ reduction. Tanaka et al. proved that the addition of NaCl to water can increase the selective photocatalytic conversion of CO₂ to CO by two times compared to that in pure water over Ni–Al layered double hydroxides (LDHs).¹⁵⁷ For the first time, HClO is proposed to be the oxidation product in H₂O based on the experimental results (entry 5, Table 2), whose amount is in a stoichiometric ratio to CO. The photocatalytic CO₂RR efficiency using NaCl as the electron donor can be further improved by controlling the morphology of the LDH-based photocatalysts at the nanoscale¹⁶² or tuning their chemical composition.¹⁶³ Interestingly, Ni–V LDH shows significantly high photocatalytic CO₂ reduction properties due to the use of sacrificial Cl⁻ and the regeneration of Cl₂ by V³⁺.¹⁶³

NH₃ is a typical environmental contaminant and can be removed by oxidation from water.¹⁶⁴ The photooxidation of NH₃ or NH⁴⁺ is much less demanding than H₂O oxidation since ΔG^θ for NH₃ → N₂ (18 kJ mol⁻¹) is much less than that for H₂O → O₂ (237 kJ mol⁻¹),¹⁵⁸ which makes it more favorable as the hole scavenger for the photocatalytic CO₂RR. Tanaka et al. indicated that NH₃ can work as an effective electron donor for the CO₂RR.¹⁶⁵ CO and H₂ are the reduction products with a rate of 0.5 mmol h⁻¹ for CO, and stoichiometric N₂ is found to be the oxidation product. They further reported enhanced CO₂RR performance compared to H₂O using ammonia as the reducing reagent over Ag/Ba₂KTa₅O₁₅ under illumination of λ > 280 nm, with a CO production yield of 117 μmol h⁻¹ and selectivity up to 98.6% in 0.7 mol L⁻¹ NH₃·H₂O.¹⁶⁶ Recently NH₄HCO₃ was also used as the electron donor, and it shows significantly higher selective CO production than other additives like NaOH, Na₂CO₃, KHCO₃ and a Na₂HPO₄/NaH₂PO₄ mixture.¹⁶⁷ During the photooxidation of NH₃ or NH₄⁺, it is claimed that N₂H₄ is the first intermediate that is more active than NH₃.^166,168 So N₂H₄·H₂O can also work as an electron donor, as manifested in the Au–Cu/SrTiO₃/TiO₂ system.¹⁶⁹ Besides being the electron donor, in this work N₂H₄·H₂O also serves as the H source and reducing agent for protecting the metal from oxidation. Thus, this system shows a much higher CH₄ evolution rate than that in H₂O.

Besides the aforementioned organic and inorganic electron donors, other chemicals can also facilitate hole trapping and thereby enhance the photocatalytic CO₂RR performance, such as ascorbic acid,¹⁷⁰ methanol,¹⁷¹ EDTA¹⁷² and NaH₂PO₂.¹⁷³ Moreover, apart from those employed in the liquid phase, H₂ and CH₄ are the representative gas-phase reductants.

In summary, sacrificial chemicals acting as the electron donors can consume the superfluous photogenerated holes in the semiconductors, resulting in more efficient electron utilization. However, the sacrificial electron donor is not recyclable, and needs to be supplemented during the reaction to maintain the high performance. Therefore, in most cases the product amount tends to increase nonlinearly with time, as the sacrificial electron donors are consumed during the reaction. Accordingly, designing a recyclable electron donor that can be regenerated from its oxidized state is desired. A cyclic reaction was proposed using carbon-coated In₂O₃ as the photocatalyst, H₂O as the reductant, and TEOA as the electron donor.¹³⁸ TEOA is oxidized by the photogenerated holes during the reaction and regenerated by the hydroxyl ions. However, relevant evidence is needed to support this, since water oxidation is more formidable than TEOA oxidation. Moreover, the radical cations for the tertiary amines decompose easily into a mixture, which is difficult to convert back to the initial state. Based on the computational results, Richardson et al. revealed two remote C–H activation modes for amine radical cations that might make it possible to regenerate the electron donor.¹⁷⁴ A general strategy is thus put forward theoretically in photochemical CO₂ reduction by introducing a reducing mediating reagent that can recycle amines via hydrogenation (eqn (6)–(9)).¹⁷⁵ However, further experiments are needed to confirm this.

Overall reaction:

4hν + 2CO₂ + 2H₂O → 2HCOOH + O₂ (9)

4 Multifunctional CO₂RR systems with effective use of the OHR

As discussed in Section 3.1, the photocatalytic CO₂RR using H₂O as an electron donor is an ideal choice to mimic photosynthesis. However, H₂O oxidation needs a high overpotential. In addition, the evolved O₂ is an undesired by-product and is difficult to separate from the CO₂ reduction products. As for the sacrificial electron donors discussed in Section 3.2 (such as TEA, TEOA and 2-propanol), they are of high value, while the related oxidation products are usually useless and may cause environmental concern as well. Therefore, it is highly desirable to effectively utilize the oxidative holes to drive meaningful reactions, via integrating with the CO₂RR consuming the electrons. Since waste treatment and organic transformation are common oxidation processes, efforts have been devoted to coupling the photocatalytic CO₂RR with these processes.

Some systems have been reported to show activity in both the photocatalytic CO₂RR and useful OHR, especially the Z-scheme and cocatalyst-modified systems. For instance, Al–O-linked porous-g-C₃N₄/TiO₂-nanotube Z-scheme composites,¹⁷⁶ SnO₂ and Ag co-modified g-C₃N₄,¹⁷⁷ TiO₂-coupled N-doped porous perovskite LaFeO₃ nanocomposites,¹⁷⁸ chloride and phosphate anion modified P25 TiO₂ (ref. 179) and TiO₂ nanotubes co-modified with MnO_x and RGO¹⁸⁰ have all shown high performance in both the photocatalytic CO₂RR and degradation of 2,4-dichlorophenol. Other g-C₃N₄ containing heterogeneous systems^181–189 or semiconductor systems like zinc silicate,¹⁹⁰ SnO₂-coupled α-Fe₂O₃,¹⁹¹ SnO₂/Ag/MoS₂,¹⁹² CeO₂/Ti₃C₂-MXene,¹⁹³ LaDySn₂O₇/SnSe,¹⁹⁴ N-doped graphene/Zn_1.231Ge_0.689N_1.218O_0.782,¹⁹⁵ and Pt,N co-doped BiOCl¹⁹⁶ have also exhibited activity for both the CO₂RR and degradation of various organic pollutants (such as rhodamine B, methyl orange, methylene blue (MB), phenol, tetracycline, acetaldehyde, 4-chlorophenol, and benzyl alcohol). However, it should be pointed out that in the aforementioned work the CO₂RR and OHR are tested in separate experiments. For real applications, it is highly desirable to carry out the two half reactions simultaneously in a single multifunctional system, which is helpful for decreasing energy demand and promoting the overall reaction for a sustainable chemistry. In this regard, since the concept of (photo)electrochemical setup is useful for the photocatalytic system design, the related advances will be presented in this section for a deep and broad reference.

4.1 CO₂ reduction combined with waste treatment

For CO₂ reduction in tandem with waste treatment by the (photo)electrochemical method, the waste oxidation can accelerate the overall conversion efficiency via lowering the required input energy as compared to the OER. A recent analysis of Gibbs free energy indicates that for the electrocatalytic CO₂RR in H₂O, ∼90% of the overall energy is used to drive the H₂O oxidation.¹⁹⁷ Using an alternative OHR with superior thermodynamics can lower the energy requirement for a specific system. For instance, the replacement of H₂O oxidation by glycerol oxidation can reduce the electricity input by 53% during the CO₂ electroreduction.¹⁹⁷ Another study by Guo et al. employed an integrated cell with Cu nanowires as the cathode for CO₂ reduction and Sb/SnO₂–TiO₂ nanotubes as the anode for photoelectrocatalytic phenol oxidation (Fig. 17a).¹⁹⁸ Coupling these two reactions leads to a decreased energy consumption by 51.33% compared to the sum of the two half reduction and oxidation reactions (Fig. 17b). Moreover, Bevilacqua et al. developed an electrocatalytic device for the CO₂RR to yield useful chemicals (CH₄, C₂H₄, and HCOO⁻) at the cathode and ethanol oxidation to acetate at the anode, for which the operation potential decreases as compared to the OER at the anode.¹⁹⁹

Fig. 17 (a) Schematic diagram of the integrated cell with Cu nanowires as the cathode for CO₂ reduction and Sb/SnO₂–TiO₂ nanotubes as the anode for photoelectrocatalytic phenol oxidation, and (b) energy consumption for different reaction systems. Reprinted with permission from ref. 198. Copyright 2019, Wiley-VCH.

Apart from the organic degradation, other waste removal reactions can also be used as the OHR. Poisonous SO₂ removal and CO₂ reduction are integrated in a photoelectrochemical cell, with improved CO₂RR efficiency and faradaic efficiency of the corresponding products. The high performance benefits from the introduction of SO₂ because of its lower activation energy and faster kinetics than those of H₂O oxidation.²⁰⁰ Moreover, simultaneous efficient As(III) oxidation to As(V) over Ti/Ir_1−xTa_xO_y/TiO₂ and CO₂ reduction to formate on a Bi electrode has been realized, driven by a photovoltaic device.²⁰¹ In addition, Li et al. proposed a solar-driven electrochemical system to simultaneously convert CO₂ and H₂S, the usual contaminants in natural gas, respectively, to CO and S, with the configuration shown in Fig. 18.²⁰² The oxidation of H₂S is achieved with the aid of EDTA–Fe²⁺ on graphene, and the CO₂RR occurs over the graphene-encapsulated ZnO catalyst.

Fig. 18 Configuration used for simultaneous conversion of CO₂ and H₂S. Reprinted with permission from ref. 202. Copyright 2018, Wiley-VCH.

4.2 One-pot conversion of pollutants to value-added chemicals

Considering that the oxidation of organic pollutants usually generates CO₂ as the final product, utilization of the obtained CO₂ can make it possible to convert organic pollutants to useful chemicals. Such a carbon cycle is an important resource utilization method, and may realize the real applications of reduction and oxidation reactions in a synergistic way.

4.2.1 (Photo)electrochemical approach. Cui et al. reported a single chamber setup consisting of a Pd/TiO₂ anode and Pd–Cu cathode, in which the in situ generated CO₂ from phenol oxidation is reduced at the Pd–Cu cathode.²⁰³ Since the constitution of the syngas obtained here is tunable, it is possible to directly convert the pollutant to valuable products. In addition, direct conversion of pollutants to value-added chemicals can also be realized in an H-type cell, in which the CO₂ generated at the anode is transported to the cathode by a carrier gas. Zou et al. prepared a Co₃O₄ anode for 4-nitrophenol (4-NP) oxidation, coupled with a CuO cathode for the generated CO₂ to be reduced to liquid products (methanol and ethanol) (Fig. 19).²⁰⁴ The respective maximum yield of methanol and ethanol is 98.29 and 40.95 μmol L⁻¹, which means a conversion of 41.8% of 4-NP into liquid fuels. The conversion of 4-NP to formate is also achieved with a 24.1% faradaic efficiency at −1.3 V, by using 3D Co₃O₄@C as the anode and SnO₂/carbon cloth as the cathode.²⁰⁵

Fig. 19 Setup used for one-pot conversion of 4-nitrophenol to methanol and ethanol and the possible reaction processes. Reprinted with permission from ref. 204. Copyright 2019, Elsevier.

4.2.2 Photochemical approach. One-pot conversion of organic pollutants to valuable chemicals over a powder suspension is a cost-effective choice that is easier to operate, but is more challenging and has low conversion efficiency. Zou et al. developed a graphene quantum dot/V-doped TiO₂ composite for direct conversion of MB to methanol, ethanol and CH₄, with optimized production rates of 13.24, 5.65 and 0.445 μmol g⁻¹ h⁻¹, respectively.²⁰⁶ It is proposed that this process is realized via two steps, MB oxidation to CO₂ followed by CO₂ reduction to the final products. Similarly, the rGO/SrTi_0.95Fe_0.05O_3−δ catalyst is also able to couple the pollutant degradation and CO₂ reduction in one pot, converting rhodamine into methanol and ethanol with the respective product rates of 10.76 and 6.14 μmol g⁻¹ h⁻¹.²⁰⁷ Xie et al. realized direct photoconversion of the major components of waste plastics (polyethylene (PE), polypropylene (PP), and polyvinyl chloride (PVC)) to acetic acid over Nb₂O₅ in a short time in natural simulated environments (Fig. 20a and b).²⁰⁸ The PE to acetic acid photoconversion includes two sequential reactions, C–C cleavage to CO₂ through an oxidation process and C–C coupling to acetic acid by photoinduced electrons (Fig. 20c and d). Therefore, the one-pot conversion of pollutants to useful products can provide an ideal scenario to couple the degradation and reduction and thus treat organic pollutants via simultaneous resource recycling.

Fig. 20 The yield of (a) CO₂ and (b) acetic acid during photoconversion of PE, PP and PVC over Nb₂O₅, (c) band edge positions for Nb₂O₅ and related redox potentials, and (d) mechanism for C–C cleavage and coupling during PE photoconversion to acetic acid. Reprinted with permission from ref. 208. Copyright 2020, Wiley-VCH.

4.3 CO₂RR integrated with chemical oxidative synthesis

Forming value-added products from both the reduction and oxidation half reactions is an appealing concept, which makes full utilization of the electrons and holes. There are some reports on tandem CO₂ reduction and selective organic oxidation to produce value-added chemicals. For example, the photocatalytic CO₂RR coupled with selective oxidation of alcohols is realized over a Ag/TiO₂ composite, with a maximum methanol yield of 358.7 μmol g⁻¹ for the CO₂RR and 91.7% aromatic alcohol oxidation to aldehyde.²⁰⁹ Moreover, simultaneous photocatalytic CO₂ reduction and benzyl alcohol oxidation to benzyl acetate (or benzaldehyde) is achieved over a Cu₂O/Cu (or Cu₂O-RGO/BiVO₄) composite.^210,211 Li et al. reported an OHR involving conversion of four alcohols to high value-added hydrocarbon products with >78% yield and CO₂ reduction to CO with a high faradaic efficiency.²¹² However, some alcohols may not be the ideal resource for the oxidation synthesis considering their value. Alternatively, using a waste product or common synthetic material for valuable chemical oxidative synthesis is usually preferred.

Llorente et al. designed a system consisting of CO₂ reduction to CO at the cathode coupled with syringaldehyde and o-phenylenediamine oxidative condensation to 2-(3,5-dimethoxy-4-hydroxyphenyl) benzimidazole at the anode.²¹³ Furthermore, the direct oxidation of organic materials to useful chemicals is more frequently used. A BiOBr photoanode coupled with a CuO photocathode can yield useful chemicals by merging the oxidation of the tetracycline pollutant with photoelectrocatalytic reduction of CO₂.²¹⁴ Wu et al. combined the photocatalytic CO₂RR with oxidative synthesis of pinacol, an important pharmaceutical intermediate, from 1-phenylethanol over CdSe/CdS quantum dots, with the utilization ratio of photogenerated electrons and holes close to 1.²¹⁵ The oxidative conversion of pinacol from 1-phenylethanol can reach 98% yield, which also provides electrons and protons for highly selective conversion of CO₂ to CO. Moreover, the reduction product CO and oxidation product pinacol can spontaneously separate in different phases. Another pharmaceutical intermediate, dihydroisoquinoline (DHIQ) can be obtained by dehydrogenation of the corresponding tetrahydroisoquinoline (THIQ) precursors. The lower oxidation potential of THIQ than water makes it a possible candidate to boost the OHR. Zhao et al. integrated the photocatalytic CO₂RR with selective THIQ oxidation to DHIQ over an InP–In₂O₃ junction with a CO yield of 170 μmol g⁻¹ h⁻¹.²¹⁶

5 Conclusion & perspectives

In summary, this review highlights the progress on managing the OHR in photocatalytic CO₂ conversion. The related studies cover materials design, electron donors and valuable oxidation integrated with CO₂ reduction.

Investigations on materials design mainly aim at regulating the hole behavior, such as accelerating the transfer process and/or preserving the high driving force. Element modification like with Cl⁻ can help trap photogenerated holes, facilitate the charge separation, and thereby improve the photocatalytic CO₂RR. Similarly, loading oxidation cocatalysts can promote extraction and trapping of the photogenerated holes, and serve as active sites for the OHR. Furthermore, dual cocatalysts stand out as a better choice due to parallel extraction of the electrons and holes to the respective reaction sites. Different from dual cocatalysts, Z-scheme systems simultaneously modulate the behavior of electrons and holes by retaining their high driving force and enhancing the carrier lifetime.

Electron donors play an important role in accelerating the reaction process. The photocatalytic CO₂RR in H₂O is an ideal case but is very challenging for the time being. So far, only a few studies clearly reported the overall CO₂ reduction with O₂ as the oxidation product, but with unsatisfactory efficiency. Higher performance can be achieved by replacing H₂O with sacrificial electron donors, which can react with the photogenerated holes very quickly. However, there is still a lack of detailed and systematic study on the oxidation mechanism of the electron donor and its impact on the CO₂RR, which makes it difficult to rationally design highly efficient systems by matching the appropriate electron donor. Moreover, sacrificial electron donors are not the ultimate solution, as the experiments will slow down as soon as they run out and many sacrificial electron donors are more valuable compared with the useless oxidation products. Accordingly, it is hard to say whether the benefit from the CO₂RR can make up for the cost of the sacrificial chemicals, but one major strength of sacrificial agents is their contribution to the studies that only focus on the reduction half reactions.

Integrating the CO₂RR with a meaningful OHR offers a possible approach to address the above concern. Deliberate integration of the two half reactions can result in the maximum use of the electrons and holes, and thus high efficiency of the full reaction. This has been demonstrated by combining the CO₂RR with waste treatment or oxidative organic transformation. However, so far only a few studies have been devoted to this. More efforts are needed to construct effective systems to further increase the activity and product selectivity for both half reactions. Specifically, a better understanding of both half reactions is highly desired so that a perfect match between the reduction and oxidation reactions can be established and, accordingly, a high overall efficiency can be achieved.

To conclude, study of the OHR that accompanies the CO₂RR is quite necessary for long-term research. Although significant achievements have been made hitherto, there is still a long way to go to realize the expected practical applications. Advances in the photocatalytic CO₂RR are often hindered by the sophisticated multi-electron–hole involving reaction processes, low quantum efficiency, and poor photocatalyst stability. In future studies, a more rational design of the photocatalyst and better understanding of the mechanism are needed. It is highly desirable to explore new photocatalysts for the overall photocatalytic CO₂ reduction, especially for synchronously managing the behavior of electrons and holes. Systematic investigations should be focused on the mechanism of CO₂ reduction, especially that coupled with a sacrificial electron donor or meaningful oxidation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Belt and Road Initiative by the Chinese Academy of Sciences (121D11KYSB20170050), National Natural Science Foundation of China (21972029), and Strategic Priority Research Program of the Chinese Academy of Sciences (XDB36000000).

References

1. T. Sakakura, J. C. Choi and H. Yasuda, Chem. Rev., 2007, 107, 2365–2387.
2. N. N. Vu, S. Kaliaguine and T. O. Do, Adv. Funct. Mater., 2019, 29, 1901825.
3. J. L. White, M. F. Baruch, J. E. Pander III, Y. Hu, I. C. Fortmeyer, J. E. Park, T. Zhang, K. Liao, J. Gu, Y. Yan, T. W. Shaw, E. Abelev and A. B. Bocarsly, Chem. Rev., 2015, 115, 12888–12935.
4. X. Li, J. Yu, M. Jaroniec and X. Chen, Chem. Rev., 2019, 119, 3962–4179.
5. S. N. Habisreutinger, L. Schmidt-Mende and J. K. Stolarczyk, Angew. Chem., Int. Ed., 2013, 52, 7372–7408.
6. J. Fu, K. Jiang, X. Qiu, J. Yu and M. Liu, Mater. Today, 2020, 32, 222–243.
7. S. Yan, J. Wang, H. Gao, N. Wang, H. Yu, Z. Li, Y. Zhou and Z. Zou, Adv. Funct. Mater., 2013, 23, 758–763.
8. J. Tang, J. R. Durrant and D. R. Klug, J. Am. Chem. Soc., 2008, 130, 13885–13891.
9. P. V. Kamat, Chem. Rev., 1993, 93, 267–300.
10. J. Z. Y. Tan and M. M. Maroto-Valer, J. Mater. Chem. A, 2019, 7, 9368–9385.
11. A. Meng, L. Zhang, B. Cheng and J. Yu, Adv. Mater., 2019, 31, e1807660.
12. W. Zhang, A. R. Mohamed and W. J. Ong, Angew. Chem., Int. Ed., 2020 DOI:10.1002/anie.201914925.
13. L. J. Guo, Y. J. Wang and T. He, Chem. Rec., 2016, 16, 1918–1933.
14. K. Li, X. An, K. H. Park, M. Khraisheh and J. Tang, Catal. Today, 2014, 224, 3–12.
15. E. Karamian and S. Sharifnia, J. CO₂ Util., 2016, 16, 194–203.
16. W. Tu, Y. Zhou and Z. Zou, Adv. Mater., 2014, 26, 4607–4626.
17. A. D. Handoko, K. Li and J. Tang, Curr. Opin. Chem. Eng., 2013, 2, 200–206.
18. J. Tang, A. J. Cowan, J. R. Durrant and D. R. Klug, J. Phys. Chem. C, 2011, 115, 3143–3150.
19. S. Wang, K. Teramura, T. Hisatomi, K. Domen, H. Asakura, S. Hosokawa and T. Tanaka, ACS Appl. Energy Mater., 2020, 3, 1468–1475.
20. S. L. Liew, Z. Zhang, T. W. G. Goh, G. S. Subramanian, H. L. D. Seng, T. S. A. Hor, H. K. Luo and D. Z. Chi, Int. J. Hydrogen Energy, 2014, 39, 4291–4298.
21. K. Wang, J. Fu and Y. Zheng, Appl. Catal., B, 2019, 254, 270–282.
22. R. Zhang, Y. Y. Fang, T. Chen, F. L. Qu, Z. Liu, G. Du, A. M. Asiri, T. Gao and X. P. Sun, ACS Sustainable Chem. Eng., 2017, 5, 7502–7506.
23. R. Nakamura, T. Tanaka and Y. Nakato, J. Phys. Chem. B, 2004, 108, 10617–10620.
24. L. Huang, R. Li, R. Chong, G. Liu, J. Han and C. Li, Catal. Sci. Technol., 2014, 4, 2913–2918.
25. X. Zhang, H. Cui, M. Humayun, Y. Qu, N. Fan, X. Sun and L. Jing, Sci. Rep., 2016, 6, 21430.
26. J. Wu, H. Lu, X. Zhang, F. Raziq, Y. Qu and L. Jing, Chem. Commun., 2016, 52, 5027–5029.
27. J. Li, X. Zhang, F. Raziq, J. Wang, C. Liu, Y. Liu, J. Sun, R. Yan, B. Qu, C. Qin and L. Jing, Appl. Catal., B, 2017, 218, 60–67.
28. L. Yu, X. Ba, M. Qiu, Y. Li, L. Shuai, W. Zhang, Z. Ren and Y. Yu, Nano Energy, 2019, 60, 576–582.
29. X. Zhang, X. Zhang, W. Ali, X. Chen, K. Hu, Z. Li, Y. Qu, L. Bai, Y. Gao and L. Jing, ChemCatChem, 2019, 11, 6282–6287.
30. T. Yoshinaga, M. Saruyama, A. Xiong, Y. Ham, Y. Kuang, R. Niishiro, S. Akiyama, M. Sakamoto, T. Hisatomi, K. Domen and T. Teranishi, Nanoscale, 2018, 10, 10420–10427.
31. D. Liu, L. Jing, P. Luan, J. Tang and H. Fu, ACS Appl. Mater. Interfaces, 2013, 5, 4046–4052.
32. Y. Zheng, Y. Jiao, Y. Zhu, Q. Cai, A. Vasileff, L. H. Li, Y. Han, Y. Chen and S. Z. Qiao, J. Am. Chem. Soc., 2017, 139, 3336–3339.
33. B. H. Meekins and P. V. Kamat, J. Phys. Chem. Lett., 2011, 2, 2304–2310.
34. M. W. Kanan and D. G. Nocera, Science, 2008, 321, 1072–1075.
35. M. Barroso, A. J. Cowan, S. R. Pendlebury, M. Gratzel, D. R. Klug and J. R. Durrant, J. Am. Chem. Soc., 2011, 133, 14868–14871.
36. A. D. Handoko and J. Tang, Int. J. Hydrogen Energy, 2013, 38, 13017–13022.
37. J. Lin, Z. Pan and X. Wang, ACS Sustainable Chem. Eng., 2013, 2, 353–358.
38. J. Zhou, W. Chen, C. Sun, L. Han, C. Qin, M. Chen, X. Wang, E. Wang and Z. Su, ACS Appl. Mater. Interfaces, 2017, 9, 11689–11695.
39. S. Shoji, A. Yamaguchi, E. Sakai and M. Miyauchi, ACS Appl. Mater. Interfaces, 2017, 9, 20613–20619.
40. Y. Bai, L. Ye, L. Wang, X. Shi, P. Wang and W. Bai, Environ. Sci.: Nano, 2016, 3, 902–909.
41. Q. Liu, Y. Zhou, J. Kou, X. Chen, Z. Tian, J. Gao, S. Yan and Z. Zou, J. Am. Chem. Soc., 2010, 132, 14385–14387.
42. Q. Liu, Y. Zhou, Z. Tian, X. Chen, J. Gao and Z. Zou, J. Mater. Chem., 2012, 22, 2033–2038.
43. C. Dong, S. Hu, M. Xing and J. Zhang, Nanotechnology, 2018, 29, 154005.
44. A. Li, X. Chang, Z. Huang, C. Li, Y. Wei, L. Zhang, T. Wang and J. Gong, Angew. Chem., Int. Ed., 2016, 55, 13734–13738.
45. A. Li, T. Wang, X. Chang, Z. J. Zhao, C. Li, Z. Huang, P. Yang, G. Zhou and J. Gong, Chem. Sci., 2018, 9, 5334–5340.
46. C. Dong, M. Xing and J. Zhang, Mater. Horiz., 2016, 3, 608–612.
47. Z. Jiang, D. Ding, L. Wang, Y. Zhang and L. Zan, Catal. Sci. Technol., 2017, 7, 3065–3072.
48. M. S. Akple, J. Low, S. Liu, B. Cheng, J. Yu and W. Ho, J. CO₂ Util., 2016, 16, 442–449.
49. A. Meng, L. Zhang, B. Cheng and J. Yu, ACS Appl. Mater. Interfaces, 2019, 11, 5581–5589.
50. Y. Bai, L. Ye, L. Wang, X. Shi, P. Wang, W. Bai and P. K. Wong, Appl. Catal., B, 2016, 194, 98–104.
51. J. Jin, J. Yu, D. Guo, C. Cui and W. Ho, Small, 2015, 11, 5262–5271.
52. W. Shi, X. Guo, C. Cui, K. Jiang, Z. Li, L. Qu and J.-C. Wang, Appl. Catal., B, 2019, 243, 236–242.
53. H. Li, Y. Gao, Y. Zhou, F. Fan, Q. Han, Q. Xu, X. Wang, M. Xiao, C. Li and Z. Zou, Nano Lett., 2016, 16, 5547–5552.
54. Y. Yang, J. Wu, T. Xiao, Z. Tang, J. Shen, H. Li, Y. Zhou and Z. Zou, Appl. Catal., B, 2019, 255, 117771.
55. Y. Jiang, J. Liao, H. Chen, H. Zhang, J. Li, X. Wang and D. Kuang, Chem, 2020, 6, 766–780.
56. H. Tada, T. Mitsui, T. Kiyonaga, T. Akita and K. Tanaka, Nat. Mater., 2006, 5, 782–786.
57. J. Low, B. Dai, T. Tong, C. Jiang and J. Yu, Adv. Mater., 2019, 31, e1802981.
58. J. Yang, J. Hao, S. Xu, Q. Wang, J. Dai, A. Zhang and X. Pang, ACS Appl. Mater. Interfaces, 2019, 11, 32025–32037.
59. Y. Wang, Z. Zhang, L. Zhang, Z. Luo, J. Shen, H. Lin, J. Long, J. C. S. Wu, X. Fu, X. Wang and C. Li, J. Am. Chem. Soc., 2018, 140, 14595–14598.
60. M. E. Aguirre, R. Zhou, A. J. Eugene, M. I. Guzman and M. A. Grela, Appl. Catal., B, 2017, 217, 485–493.
61. T. Takayama, K. Sato, T. Fujimura, Y. Kojima, A. Iwase and A. Kudo, Faraday Discuss., 2017, 198, 397–407.
62. J. S. DuChene, G. Tagliabue, A. J. Welch, W. H. Cheng and H. A. Atwater, Nano Lett., 2018, 18, 2545–2550.
63. R. Kuriki, O. Ishitani and K. Maeda, ACS Appl. Mater. Interfaces, 2016, 8, 6011–6018.
64. N. M. Dimitrijevic, B. K. Vijayan, O. G. Poluektov, T. Rajh, K. A. Gray, H. He and P. Zapol, J. Am. Chem. Soc., 2011, 133, 3964–3971.
65. J. C. Wang, H. C. Yao, Z. Y. Fan, L. Zhang, J. S. Wang, S. Q. Zang and Z. J. Li, ACS Appl. Mater. Interfaces, 2016, 8, 3765–3775.
66. X. Li, J. Wen, J. Low, Y. Fang and J. Yu, Sci. China: Chem., 2014, 57, 70–100.
67. Z. Chen, J. J. Concepcion, M. K. Brennaman, P. Kang, M. R. Norris, P. G. Hoertz and T. J. Meyer, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 15606–15611.
68. T. Inoue, A. Fujishima, S. Konishi and K. Honda, Nature, 1979, 277, 637–638.
69. H. Wang, Y. Wang, L. Guo, X. Zhang, C. Ribeiro and T. He, Chin. J. Catal., 2020, 41, 131–139.
70. T. Di, J. Zhang, B. Cheng, J. Yu and J. Xu, Sci. China: Chem., 2018, 61, 344–350.
71. L. F. Garay-Rodríguez, L. M. Torres-Martínez and E. Moctezuma, J. Energy Chem., 2019, 37, 18–28.
72. J. Jin, Y. Wang and T. He, RSC Adv., 2015, 5, 100244–100250.
73. K. Teramura, S. Iguchi, Y. Mizuno, T. Shishido and T. Tanaka, Angew. Chem., Int. Ed., 2012, 51, 8008–8011.
74. Y. Huang, Y. Wang, Y. Bi, J. Jin, M. F. Ehsan, M. Fu and T. He, RSC Adv., 2015, 5, 33254–33261.
75. Y. F. Mu, W. Zhang, X. X. Guo, G. X. Dong, M. Zhang and T. B. Lu, ChemSusChem, 2019, 12, 4769–4774.
76. Y. A. Wu, I. McNulty, C. Liu, K. C. Lau, Q. Liu, A. P. Paulikas, C.-J. Sun, Z. Cai, J. R. Guest, Y. Ren, V. Stamenkovic, L. A. Curtiss, Y. Liu and T. Rajh, Nat. Energy, 2019, 4, 957–968.
77. K. Teramura and T. Tanaka, Phys. Chem. Chem. Phys., 2018, 20, 8423–8431.
78. K. Ogura, M. Kawano, J. Yano and Y. Sakata, J. Photochem. Photobiol., A, 1992, 66, 91–97.
79. K. Iizuka, T. Wato, Y. Miseki, K. Saito and A. Kudo, J. Am. Chem. Soc., 2011, 133, 20863–20868.
80. T. Takayama, K. Tanabe, K. Saito, A. Iwase and A. Kudo, Phys. Chem. Chem. Phys., 2014, 16, 24417–24422.
81. T. Takayama, A. Iwase and A. Kudo, Bull. Chem. Soc. Jpn., 2015, 88, 538–543.
82. H. Nakanishi, K. Iizuka, T. Takayama, A. Iwase and A. Kudo, ChemSusChem, 2017, 10, 112–118.
83. X. Chen, Y. Zhou, Q. Liu, Z. Li, J. Liu and Z. Zou, ACS Appl. Mater. Interfaces, 2012, 4, 3372–3377.
84. Y. Zhou, Z. Tian, Z. Zhao, Q. Liu, J. Kou, X. Chen, J. Gao, S. Yan and Z. Zou, ACS Appl. Mater. Interfaces, 2011, 3, 3594–3601.
85. Q. Han, X. Bai, Z. Man, H. He, L. Li, J. Hu, A. Alsaedi, T. Hayat, Z. Yu, W. Zhang, J. Wang, Y. Zhou and Z. Zou, J. Am. Chem. Soc., 2019, 141, 4209–4213.
86. L. Liang, X. Li, Y. Sun, Y. Tan, X. Jiao, H. Ju, Z. Qi, J. Zhu and Y. Xie, Joule, 2018, 2, 1004–1016.
87. L. Liang, X. Li, J. Zhang, P. Ling, Y. Sun, C. Wang, Q. Zhang, Y. Pan, Q. Xu, J. Zhu, Y. Luo and Y. Xie, Nano Energy, 2020, 69, 104421.
88. M. Wang, D. Wang and Z. Li, Appl. Catal., B, 2016, 183, 47–52.
89. M. Lu, J. Liu, Q. Li, M. Zhang, M. Liu, J. L. Wang, D. Q. Yuan and Y. Q. Lan, Angew. Chem., Int. Ed., 2019, 58, 12392–12397.
90. I. H. Tseng, W.-C. Chang and J. C. S. Wu, Appl. Catal., B, 2002, 37, 37–48.
91. H. Yamashita, H. Nishiguchi, N. Kamada, M. Anpo, Y. Teraoka, H. Hatano, S. Ehara, K. Kikui, L. Palmisano, A. Sclafani, M. Schiavello and M. A. Fox, Res. Chem. Intermed., 1994, 20, 815–823.
92. G. Munuera, V. Rives-Arnau and A. Saucedo, J. Chem. Soc., Faraday Trans. 1, 1979, 75, 736–747.
93. B. Zhu, S. Wageh, A. A. Al-Ghamdi, S. Yang, Z. Tian and J. Yu, Catal. Today, 2019, 335, 117–127.
94. E. Lira, S. Wendt, P. Huo, J. O. Hansen, R. Streber, S. Porsgaard, Y. Wei, R. Bechstein, E. Laegsgaard and F. Besenbacher, J. Am. Chem. Soc., 2011, 133, 6529–6532.
95. P. Salvador and C. Gutiérrez, Surf. Sci., 1983, 124, 398–406.
96. M. V. Rao, K. Rajeshwar, V. R. P. Verneker and J. DuBow, J. Phys. Chem., 1980, 84, 1987–1991.
97. J. Kiwi and M. Graetzel, J. Phys. Chem., 1984, 88, 1302–1307.
98. R. Nakamura and Y. Nakato, J. Am. Chem. Soc., 2004, 126, 1290–1298.
99. K. L. Bae, J. Kim, C. K. Lim, K. M. Nam and H. Song, Nat. Commun., 2017, 8, 1156.
100. T. Abe, E. Suzuki, K. Nagoshi, K. Miyashita and M. Kaneko, J. Phys. Chem. B, 1999, 103, 1119–1123.
101. K. Sayama and H. Arakawa, J. Phys. Chem., 1993, 97, 531–533.
102. Z. Wang, K. Teramura, S. Hosokawa and T. Tanaka, Appl. Catal., B, 2015, 163, 241–247.
103. H. Yoshida, L. Zhang, M. Sato, T. Morikawa, T. Kajino, T. Sekito, S. Matsumoto and H. Hirata, Catal. Today, 2015, 251, 132–139.
104. A. Anzai, N. Fukuo, A. Yamamoto and H. Yoshida, Catal. Commun., 2017, 100, 134–138.
105. R. Pang, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, Appl. Catal., B, 2017, 218, 770–778.
106. K. Li, A. D. Handoko, M. Khraisheh and J. Tang, Nanoscale, 2014, 6, 9767–9773.
107. S. Iguchi, K. Teramura, S. Hosokawa and T. Tanaka, Catal. Sci. Technol., 2016, 6, 4978–4985.
108. K. Teramura, H. Tatsumi, Z. Wang, S. Hosokawa and T. Tanaka, Bull. Chem. Soc. Jpn., 2015, 88, 431–437.
109. Z. Huang, K. Teramura, S. Hosokawa and T. Tanaka, Appl. Catal., B, 2016, 199, 272–281.
110. Z. Huang, S. Yoshizawa, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, ACS Omega, 2017, 2, 8187–8197.
111. Z. Huang, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, Catal. Today, 2018, 300, 173–182.
112. M. Yamamoto, T. Yoshida, N. Yamamoto, T. Nomoto, Y. Yamamoto, S. Yagi and H. Yoshida, J. Mater. Chem. A, 2015, 3, 16810–16816.
113. K. Teramura, Z. Wang, S. Hosokawa, Y. Sakata and T. Tanaka, Chemistry, 2014, 20, 9906–9909.
114. Z. Wang, K. Teramura, Z. Huang, S. Hosokawa, Y. Sakata and T. Tanaka, Catal. Sci. Technol., 2016, 6, 1025–1032.
115. H. Tatsumi, K. Teramura, Z. Huang, Z. Wang, H. Asakura, S. Hosokawa and T. Tanaka, Langmuir, 2017, 33, 13929–13935.
116. Z. Huang, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, J. Mater. Chem. A, 2017, 5, 19351–19357.
117. S. Iguchi, Y. Hasegawa, K. Teramura, S. Kidera, S. Kikkawa, S. Hosokawa, H. Asakura and T. Tanaka, Sustainable Energy Fuels, 2017, 1, 1740–1747.
118. R. Pang, K. Teramura, H. Tatsumi, H. Asakura, S. Hosokawa and T. Tanaka, Chem. Commun., 2018, 54, 1053–1056.
119. R. Pang, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, ACS Sustainable Chem. Eng., 2018, 7, 2083–2090.
120. S. Kikkawa, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, J. Phys. Chem. C, 2018, 122, 21132–21139.
121. Z. Wang, K. Teramura, S. Hosokawa and T. Tanaka, J. Mater. Chem. A, 2015, 3, 11313–11319.
122. L. Hao, L. Kang, H. Huang, L. Ye, K. Han, S. Yang, H. Yu, M. Batmunkh, Y. Zhang and T. Ma, Adv. Mater., 2019, 31, e1900546.
123. L. Wang, J. Wan, Y. Zhao, N. Yang and D. Wang, J. Am. Chem. Soc., 2019, 141, 2238–2241.
124. W.-K. Jo, S. Kumar, S. Eslava and S. Tonda, Appl. Catal., B, 2018, 239, 586–598.
125. N. T. Thanh Truc, L. Giang Bach, N. Thi Hanh, T. D. Pham, N. Thi Phuong Le Chi, D. T. Tran, M. V. Nguyen and V. N. Nguyen, J. Colloid Interface Sci., 2019, 540, 1–8.
126. P. Li, Y. Zhou, H. Li, Q. Xu, X. Meng, X. Wang, M. Xiao and Z. Zou, Chem. Commun., 2015, 51, 800–803.
127. X. An, K. Li and J. Tang, ChemSusChem, 2014, 7, 1086–1093.
128. J. A. Mendoza, H. K. Kim, H. K. Park and K. Y. Park, Korean J. Chem. Eng., 2012, 29, 1483–1486.
129. H. Yoneyama, Catal. Today, 1997, 39, 169–175.
130. M. J. Berr, P. Wagner, S. Fischbach, A. Vaneski, J. Schneider, A. S. Susha, A. L. Rogach, F. Jäckel and J. Feldmann, Appl. Phys. Lett., 2012, 100, 223903.
131. H. Pan and M. D. Heagy, MRS Commun., 2018, 8, 1173–1177.
132. N. M. Dimitrijevic, I. A. Shkrob, D. J. Gosztola and T. Rajh, J. Phys. Chem. C, 2011, 116, 878–885.
133. C. Di Valentin and D. Fittipaldi, J. Phys. Chem. Lett., 2013, 4, 1901–1906.
134. D. P. Leonard, H. Pan and M. D. Heagy, ACS Appl. Mater. Interfaces, 2015, 7, 24543–24549.
135. Y. Wei, G. Cheng, J. Xiong, J. Zhu, Y. Gan, M. Zhang, Z. Li and S. Dou, J. Energy Chem., 2019, 32, 45–56.
136. S. Yanagida, M. Kanemoto, K.-i. Ishihara, Y. Wada, T. Sakata and H. Mori, Bull. Chem. Soc. Jpn., 1997, 70, 2063–2070.
137. B. K. Carpenter, J. Phys. Chem. A, 2007, 111, 3719–3726.
138. Y. X. Pan, Y. You, S. Xin, Y. Li, G. Fu, Z. Cui, Y. L. Men, F. F. Cao, S. H. Yu and J. B. Goodenough, J. Am. Chem. Soc., 2017, 139, 4123–4129.
139. H. Fujiwara, H. Hosokawa, K. Murakoshi, Y. Wada, S. Yanagida, T. Okada and H. Kobayashi, J. Phys. Chem. B, 1997, 101, 8270–8278.
140. S. Sato, T. Morikawa, S. Saeki, T. Kajino and T. Motohiro, Angew. Chem., Int. Ed., 2010, 49, 5101–5105.
141. L. Lin, C. Hou, X. Zhang, Y. Wang, Y. Chen and T. He, Appl. Catal., B, 2018, 221, 312–319.
142. W. Zhong, R. Sa, L. Li, Y. He, L. Li, J. Bi, Z. Zhuang, Y. Yu and Z. Zou, J. Am. Chem. Soc., 2019, 141, 7615–7621.
143. L. Zhou, X. H. Zhang, L. Lin, P. Li, K. J. Shao, C. Z. Li and T. He, Acta Phys.-Chim. Sin., 2017, 33, 1884–1890.
144. H. Ishida, K. Tanaka and T. Tanaka, Organometallics, 1987, 6, 181–186.
145. R. Kuriki, K. Sekizawa, O. Ishitani and K. Maeda, Angew. Chem., Int. Ed., 2015, 54, 2406–2409.
146. J. M. Lehn and R. Ziessel, Proc. Natl. Acad. Sci. U. S. A., 1982, 79, 701–704.
147. W. Choi and M. R. Hoffmann, Environ. Sci. Technol., 1997, 31, 89–95.
148. N. Sasirekha, S. Basha and K. Shanthi, Appl. Catal., B, 2006, 62, 169–180.
149. S. Kaneco, Y. Shimizu, K. Ohta and T. Mizuno, J. Photochem. Photobiol., A, 1998, 115, 223–226.
150. G. Song, F. Xin, J. Chen and X. Yin, Appl. Catal., A, 2014, 473, 90–95.
151. G. Song, F. Xin and X. Yin, J. Colloid Interface Sci., 2015, 442, 60–66.
152. G. Song, X. Wu, F. Xin and X. Yin, J. Nanosci. Nanotechnol., 2017, 17, 2438–2446.
153. A. Manzi, T. Simon, C. Sonnleitner, M. Doblinger, R. Wyrwich, O. Stern, J. K. Stolarczyk and J. Feldmann, J. Am. Chem. Soc., 2015, 137, 14007–14010.
154. H. Pang, X. Meng, P. Li, K. Chang, W. Zhou, X. Wang, X. Zhang, W. Jevasuwan, N. Fukata, D. Wang and J. Ye, ACS Energy Lett., 2019, 4, 1387–1393.
155. S. Liu, Z. Zhao and Z. Wang, Photochem. Photobiol. Sci., 2007, 6, 695–700.
156. G. R. Dey, J. Nat. Gas Chem., 2007, 16, 217–226.
157. S. Iguchi, K. Teramura, S. Hosokawa and T. Tanaka, Phys. Chem. Chem. Phys., 2015, 17, 17995–18003.
158. H. Kominami, H. Nishimune, Y. Ohta, Y. Arakawa and T. Inaba, Appl. Catal., B, 2012, 111–112, 297–302.
159. E. Bahadori, A. Tripodi, A. Villa, C. Pirola, L. Prati, G. Ramis and I. Rossetti, Catalysts, 2018, 8, 430.
160. M. Compagnoni, G. Ramis, F. S. Freyria, M. Armandi, B. Bonelli and I. Rossetti, Rend. Fis. Acc. Lincei, 2017, 28, 151–158.
161. F. Galli, M. Compagnoni, D. Vitali, C. Pirola, C. L. Bianchi, A. Villa, L. Prati and I. Rossetti, Appl. Catal., B, 2017, 200, 386–391.
162. Y. Tokudome, M. Fukui, S. Iguchi, Y. Hasegawa, K. Teramura, T. Tanaka, M. Takemoto, R. Katsura and M. Takahashi, J. Mater. Chem. A, 2018, 6, 9684–9690.
163. S. Iguchi, Y. Hasegawa, K. Teramura, S. Hosokawa and T. Tanaka, J. CO₂ Util., 2016, 15, 6–14.
164. H. Wang, Y. Su, H. Zhao, H. Yu, S. Chen, Y. Zhang and X. Quan, Environ. Sci. Technol., 2014, 48, 11984–11990.
165. Z. Huang, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, Chem. Sci., 2017, 8, 5797–5801.
166. Z. Huang, S. Yoshizawa, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, ACS Sustainable Chem. Eng., 2018, 6, 8247–8255.
167. R. Pang, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, ACS Appl. Energy Mater., 2019, 2, 5397–5405.
168. H. Yuzawa, T. Mori, H. Itoh and H. Yoshida, J. Phys. Chem. C, 2012, 116, 4126–4136.
169. Q. Kang, T. Wang, P. Li, L. Liu, K. Chang, M. Li and J. Ye, Angew. Chem., Int. Ed., 2015, 54, 841–845.
170. J. C. Cardoso, S. Stulp, M. K. R. de Souza, F. F. Hudari, J. R. Gubiani, R. C. G. Frem and M. V. B. Zanoni, J. Electroanal. Chem., 2020, 856, 113384.
171. K. Sekizawa, K. Maeda, K. Domen, K. Koike and O. Ishitani, J. Am. Chem. Soc., 2013, 135, 4596–4599.
172. G. Kim and W. Choi, Appl. Catal., B, 2010, 100, 77–83.
173. M. Kanemoto, T. Shiragami, C. Pac and S. Yanagida, J. Phys. Chem., 1992, 96, 3521–3526.
174. R. D. Richardson and B. K. Carpenter, J. Am. Chem. Soc., 2008, 130, 3169–3180.
175. R. D. Richardson, E. J. Holland and B. K. Carpenter, Nat. Chem., 2011, 3, 301–303.
176. J. Wu, Y. Feng, B. E. Logan, C. Dai, X. Han, D. Li and J. Liu, ACS Sustainable Chem. Eng., 2019, 7, 15289–15296.
177. W. Ali, X. Zhang, X. Zhang, S. Ali, L. Zhao, S. Shaheen and L. Jing, Mater. Res. Bull., 2020, 122, 110676.
178. M. Humayun, Y. Qu, F. Raziq, R. Yan, Z. Li, X. Zhang and L. Jing, Environ. Sci. Technol., 2016, 50, 13600–13610.
179. Y. Liu, J. Wu, H. Lu, J. Wang, Y. Qu and L. Jing, Mater. Res. Bull., 2017, 92, 23–28.
180. J. Wu, Y. Tian, D. Li, X. Han, J. Liu and Y. Feng, Chemosphere, 2019, 226, 704–714.
181. Y. He, L. Zhang, M. Fan, X. Wang, M. L. Walbridge, Q. Nong, Y. Wu and L. Zhao, Sol. Energy Mater. Sol. Cells, 2015, 137, 175–184.
182. X. Li, C. Liu, D. Wu, J. Li, P. Huo and H. Wang, Chin. J. Catal., 2019, 40, 928–939.
183. A. Zada, N. Ali, F. Subhan, N. Anwar, M. I. Ali Shah, M. Ateeq, Z. Hussain, K. Zaman and M. Khan, Prog. Nat. Sci.: Mater. Int., 2019, 29, 138–144.
184. Y. Bai, T. Chen, P. Wang, L. Wang, L. Ye, X. Shi and W. Bai, Sol. Energy Mater. Sol. Cells, 2016, 157, 406–414.
185. H. Li, S. Gan, H. Wang, D. Han and L. Niu, Adv. Mater., 2015, 27, 6906–6913.
186. Z. Feng, L. Zeng, Y. Chen, Y. Ma, C. Zhao, R. Jin, Y. Lu, Y. Wu and Y. He, J. Mater. Res., 2017, 32, 3660–3668.
187. Y. Xu, Y. You, H. Huang, Y. Guo and Y. Zhang, J. Hazard. Mater., 2020, 381, 121159.
188. F. Raziq, Y. Qu, M. Humayun, A. Zada, H. Yu and L. Jing, Appl. Catal., B, 2017, 201, 486–494.
189. X. Hu, J. Hu, Q. Peng, X. Ma, S. Dong and H. Wang, Mater. Res. Bull., 2020, 122, 110682.
190. L. Wang, D. W. Bahnemann, L. Bian, G. Dong, J. Zhao and C. Wang, Angew. Chem., Int. Ed., 2019, 58, 8103–8108.
191. Z. Li, P. Luan, X. Zhang, Y. Qu, F. Raziq, J. Wang and L. Jing, Nano Res., 2017, 10, 2321–2331.
192. B. Khan, F. Raziq, M. Bilal Faheem, M. Umar Farooq, S. Hussain, F. Ali, A. Ullah, A. Mavlonov, Y. Zhao, Z. Liu, H. Tian, H. Shen, X. Zu, S. Li, H. Xiao, X. Xiang and L. Qiao, J. Hazard. Mater., 2020, 381, 120972.
193. J. Shen, J. Shen, W. Zhang, X. Yu, H. Tang, M. Zhang, Zulfiqar and Q. Liu, Ceram. Int., 2019, 45, 24146–24153.
194. N. Karamat, M. N. Ashiq, K. S. Joya, S. Ijaz, M. Sharif, M. F. Ehsan, M. Sher and N. Ul-Haq, ChemistrySelect, 2019, 4, 11511–11517.
195. J. Liang and L. Li, J. Mater. Chem. A, 2017, 5, 10998–11008.
196. H. Maimaitizi, A. Abulizi, K. Kadeer, D. Talifu and Y. Tursun, Appl. Surf. Sci., 2020, 502, 144083.
197. S. Verma, S. Lu and P. J. A. Kenis, Nat. Energy, 2019, 4, 466–474.
198. C. Guo, P. He, R. Cui, Q. Shen, N. Yang and G. Zhao, Adv. Energy Mater., 2019, 9, 1900364.
199. M. Bevilacqua, J. Filippi, A. Lavacchi, A. Marchionni, H. A. Miller, W. Oberhauser, E. Vesselli and F. Vizza, Energy Technol., 2014, 2, 522–525.
200. K. Li, J. Han, Y. Yang, T. Wang, Y. Feng, S. Ajmal, Y. Liu, Y. Deng, M. A. Tahir and L. Zhang, Chem. Eng. J., 2019, 355, 11–21.
201. W. Choi, M. Kim, B.-j. Kim, Y. Park, D. S. Han, M. R. Hoffmann and H. Park, Appl. Catal., B, 2020, 265, 118607.
202. W. Ma, H. Wang, W. Yu, X. Wang, Z. Xu, X. Zong and C. Li, Angew. Chem., Int. Ed., 2018, 57, 3473–3477.
203. R. Cui, Q. Shen, C. Guo, B. Tang, N. Yang and G. Zhao, Appl. Catal., B, 2020, 261, 118253.
204. J. P. Zou, Y. Chen, S. S. Liu, Q. J. Xing, W. H. Dong, X. B. Luo, W. L. Dai, X. Xiao, J. M. Luo and J. Crittenden, Water Res., 2019, 150, 330–339.
205. M. Zhu, L. Zhang, S. Liu, D. Wang, Y. Qin, Y. Chen, W. Dai, Y. Wang, Q. Xing and J. Zou, Chin. Chem. Lett., 2020, 31, 1961–1965.
206. J.-P. Zou, D.-D. Wu, J. Luo, Q.-J. Xing, X.-B. Luo, W.-H. Dong, S.-L. Luo, H.-M. Du and S. L. Suib, ACS Catal., 2016, 6, 6861–6867.
207. W.-H. Dong, D.-D. Wu, J.-M. Luo, Q.-J. Xing, H. Liu, J.-P. Zou, X.-B. Luo, X.-B. Min, H.-L. Liu, S.-L. Luo and C.-T. Au, J. Catal., 2017, 349, 218–225.
208. X. Jiao, K. Zheng, Q. Chen, X. Li, Y. Li, W. Shao, J. Xu, J. Zhu, Y. Pan, Y. Sun and Y. Xie, Angew. Chem., Int. Ed., 2020, 56, 15497–15501.
209. L. Wang, X. Zhang, L. Yang, C. Wang and H. Wang, Catal. Sci. Technol., 2015, 5, 4800–4805.
210. Y. Chen, M. Wang, Y. Ma, Y. Li, J. Cai and Z. Li, Catal. Sci. Technol., 2018, 8, 2218–2223.
211. X. Li, D. Wei, L. Ye and Z. Li, Inorg. Chem. Commun., 2019, 104, 171–177.
212. T. Li, Y. Cao, J. He and C. P. Berlinguette, ACS Cent. Sci., 2017, 3, 778–783.
213. M. J. Llorente, B. H. Nguyen, C. P. Kubiak and K. D. Moeller, J. Am. Chem. Soc., 2016, 138, 15110–15113.
214. S. Liu, Q. Xing, Y. Chen, M. Zhu, X. Jiang, S. Wu, W. Dai and J. Zou, ACS Sustainable Chem. Eng., 2018, 7, 1250–1259.
215. Q. Guo, F. Liang, X. Li, Y. Gao, M. Huang, Y. Wang, S. Xia, X. Gao, Q. Gan, Z. Lin, C. Tung and L. Wu, Chem, 2019, 5, 2605–2616.
216. B. Zhao, Y. Huang, D. Liu, Y. Yu and B. Zhang, Sci. China: Chem., 2019, 63, 28–34.

---