Tungsten oxide nanostructures and nanocomposites for photoelectrochemical water splitting

Abstract

Hydrogen production from photoelectrochemical (PEC) water splitting using semiconductor photocatalysts has attracted great attention to realize clean and renewable energy from solar energy. The visible light response of WO₃ with a long hole diffusion length (∼150 nm) and good electron mobility (∼12 cm² V⁻¹ s⁻¹) makes it suitable as the photoanode. However, WO₃ suffers from issues including rapid recombination of photoexcited electron–hole pairs, photo-corrosion during the photocatalytic process due to the formation of peroxo-species, sluggish kinetics of photogenerated holes, and slow charge transfer at the semiconductor/electrolyte interface. This work highlights the approaches to overcome these drawbacks of WO₃ photoanodes, including: (i) the manipulation of nanostructured WO₃ photoanodes to decrease the nanoparticle size to promote hole migration to the WO₃/electrolyte interface which benefits the charge separation; (ii) doping or introducing oxygen vacancies to improve electrical conductivity; exposing high energy crystal surfaces to promote the consumption of photogenerated holes on the high-active crystal face, thereby suppressing the recombination of photogenerated electrons and holes; (iii) decorating with co-catalysts to reduce the overpotential which inhibits the formation of peroxo-species; (iv) other methods such as coupling with narrow band semiconductors to accelerate the charge separation and controlling the crystal phase via annealing to reduce defects. These approaches are reviewed with detailed examples.

---

1. Introduction

The scarcity and growing demands of fossil fuels require our society to develop efficient, clean and renewable energy.^1–4 Photocatalytic water splitting to produce H₂ has become an effective approach to address these issues.^5,6 Since the pioneering studies by Becquerel in 1839 ⁷ and Fujishima and Honda in 1972,⁸ photocatalytic and photoelectrochemical (PEC) water splitting using semiconductors under solar irradiation have attracted remarkable attention. Afterwards, various photocatalysts including hematite,⁹ BiVO₄,^10–12 WO₃,^13–15 TiO₂,^16,17 ZnO,¹⁸ CdS,¹⁹ TaON,²⁰ IrO₂,²¹ SrTiO₃,^22,23 Ta₃N₅,²⁴ and BiFeO₃ ²⁵ have been exploited as photoanodes. Metal chalcogenides, e.g., CdS and MoS₂, have been reported as promising photocatalysts for water splitting because of their narrow band gap,²⁶ while they are more susceptible to degradation during the water oxidation reaction. Oxides are the most photochemically stable semiconductors in aqueous solution,²⁷ though TiO₂ and ZnO have a wide band gap energy that can be excited only by UV photons. Among the myriad of oxides, WO₃ with a band gap of 2.4–2.8 eV (ref. 28) has been considered to be the most promising visible light responsive photoanode for PEC water splitting. The large valence band (VB) of WO₃ located at approximately 3.0 V vs. NHE makes it appropriate for water oxidation. WO₃ possesses a moderate hole diffusion length (∼150 nm) compared with α-Fe₂O₃ (2–4 nm) and better electron mobility (∼12 cm² V⁻¹ s⁻¹) compared to TiO₂ (∼0.3 cm² V⁻¹ s⁻¹).²⁹ Moreover, the low-cost WO₃ exhibits excellent stability and the carriers move quickly under solar irradiation.^30,31 Nevertheless, the activity of WO₃ for PEC water splitting still needs to be improved. As an indirect semiconductor, WO₃ requires a relatively thick film for adequate light absorption which causes a significant recombination of electrons and holes, and thus decreases the performance of PEC water splitting.³² During water splitting, the possibly formed peroxo-species on the WO₃ surface accelerate the photo-corrosion and consequently reduce the stability of the WO₃ photoanode.^32,33 The sluggish kinetics of photogenerated holes and charge transfer at the semiconductor/electrolyte interface intensify the drawback of the WO₃ photoanode.^29,34,35 The maximum theoretical solar-to-hydrogen efficiency of WO₃ is ∼6% and the theoretical photocurrent density is 4 mA cm⁻² under Air Mass 1.5 Global (AM 1.5 G) solar illumination. The WO₃-based PEC devices cannot compete directly with coupled solar and electrolysis with an efficiency of about 10%. Nevertheless, modified WO₃-based PEC can potentially achieve efficiencies beyond 10% at a low cost.³⁶ Considerable attempts have been made to solve the issues including doping, orderly nanostructured morphologies, coupling with narrow band semiconductors, introducing oxygen vacancies, decorating with co-catalysts, etc.

Previous reviews have been presented the use of WO₃ in gas sensors, electrochromic, photochromic, and photocatalytic degradation. However, the strategies to improve the performance of WO₃ photoanodes for PEC water splitting have been rarely summarized. In this review, recent progress in the WO₃ photoanodes and the possibility of using WO₃ for direct Z-scheme overall water splitting are documented with detailed examples. The strategies to overcome the limitations of WO₃ as the photoanode are summarized to enhance the photocatalytic activity for PEC water splitting with detailed examples. (1) The PEC water splitting performance of WO₃ suffers from the recombination of electron–hole pairs. WO₃ nanostructures with small sizes and controllable morphologies can decrease the diffusion length of holes and reduce the grain boundaries to benefit the charge separation.³⁷ Doping with a hetero-element or introducing oxygen vacancies can improve electrical conductivity and facilitate charge transfer.³⁸ Moreover, exposing high energy crystal surfaces can promote the consumption of photoexcited holes on the high-active crystal face, which reduces the recombination of photogenerated electron–hole pairs.³² Coupling with a narrow band gap semiconductor can form an internal electric field at the interface which facilitates the separation of photogenerated electrons and holes and restricts the recombination.³⁹ (2) During the oxygen evolution reaction (OER), peroxo species formed and accumulated on the surface of WO₃ become kinetically competitive with O₂ production, and thus result in photo-corrosion. Decorating with co-catalysts can improve the surface kinetics and reduce the OER overpotential that inhibits the formation of peroxo species.⁴⁰ The optimized crystallinity and exposing highly active crystal surface can impede the formation of peroxo species on the surface of WO₃.⁴¹ (3) Although coupling with a narrower band gap semiconductor to form type II heterojunctions can improve the PEC water splitting performance, it is limited to realize overall water splitting without changing the original oxidation and reduction potential. To achieve overall water splitting efficiency, coupling with a narrow bandgap semiconductor to create a direct Z-scheme system can significantly promote an efficient charge separation as well as maintain the optimized redox ability.

2. Photocatalytic and PEC water splitting mechanism

The fundamental principle of one-step overall water splitting on a semiconductor is shown in Fig. 1a. Under light irradiation with an energy no less than the band gap of a semiconductor, electrons excited from the VB move to the conduction band (CB), leaving holes in the VB.⁴² Under standard conditions, the overall water splitting is an uphill reaction requiring a Gibbs free energy of ΔG = +237.2 kJ mol⁻¹ corresponding to the electrolysis cell voltage of ΔE° = 1.23 V per electron transferred according to the Nernst equation.⁴³ To split the water into H₂ and O₂, the band-gap energy of the semiconductors should be larger than 1.23 eV. In other words, the CB of the semiconductor should be more negative than 0 V vs. NHE, while the VB of the semiconductor has to be more positive than 1.23 V vs. NHE.^44,45 The band gap of the semiconductor should be narrower than 3 eV to achieve visible-light-response sufficiently.⁴⁶ The semiconductor must be stable in the photocatalytic reaction. At least, the photogenerated carrier transfer from the surface of the semiconductor to the solution must be facile enough to reduce energy loss arising from the kinetic overpotential.⁴⁷ In this regard, only a few semiconductors are suitable for one-step water splitting.

Fig. 1 Schematic illustration of (a) photocatalytic water splitting for a one-step photoexcitation system, (b) the basic principles of water splitting for a PEC cell with an n-type semiconductor photoanode and Pt cathode. CB, conduction band; VB, valence band; E_g, band gap; E_f, Fermi level.

PEC is another approach to achieve overall water splitting in which a semiconductor is employed as the photoanode and Pt always as the counter electrode, and both are immersed in an aqueous electrolyte. In this system (Fig. 1b), a bias potential is applied to facilitate the separation of the photoexcited electron–hole pairs due to the formation of a depletion layer in the photoanode. Meanwhile, the applied bias shifts the electrode potential of the metal cathode sufficiently negative of the reversible hydrogen potential and thereby enables the proton reduction.⁴⁸ It is thus likely for the narrow band gap semiconductor with a CB lower than the energy of H⁺/H₂ to produce hydrogen. Hence many semiconductors were used as photoanodes for the PEC overall water splitting including WO₃, despite more positive location of its CB (0.55–0.8 V vs. RHE (reversible hydrogen electrode), pH = 0).⁴⁹ In the PEC water-splitting cells, hydrogen is produced at the cathode and oxygen evolution takes place at the anode, thus the well physically separated chambers of hydrogen and oxygen can be achieved. This process of PEC water splitting can be summarized by eqn (1) and (2):⁵⁰

Anodic:

2H₂O + 4h⁺ → O₂ + 4H⁺; E° = 1.23 V vs. NHE (1)

Cathodic:

4H⁺ + 4e⁻ → 2H₂; E° = 0 V vs. NHE (2)

The PEC water splitting reaction normally occurs in three steps: (i) the photoanode semiconductor absorbs the photon energy that is higher than the band gap of the semiconductor to excite the electron–hole pairs in the bulk; (ii) the separation of photogenerated charges, i.e., the holes reach the surface of the photoanode to oxidize water molecules into O₂ whereas the electrons migrate to the counter electrode to reduce water molecules into H₂; (iii) the recombination of photogenerated holes and electrons, inducing a back-electron flow from the external circuit into the photoelectrode.⁵¹ The first two steps strongly depend on the structure and electronic properties of the photocatalyst. Concerning the third step, the recombination of holes and electrons restricts the photoconversion efficiency. Therefore, it is significantly necessary to develop a promising visible light response and create highly efficient photocatalysts for the overall water splitting.

In the PEC water splitting system, the photocurrent density as a function of potential is an important parameter to evaluate its performance. The theoretical photocurrent density of WO₃ is reported to be 4 mA cm⁻² under AM 1.5 G.³² In a three-electrode system using Ag/AgCl as a reference electrode, the measured electrode potentials vs. Ag/AgCl could be converted to the RHE potential according to the Nernst equation:^32,52

where E_RHE refers to the calculated potential vs. RHE, E_Ag/AgCl is the measured potential vs. Ag/AgCl, and the value of is 0.197 V at 25 °C.

Another important factor for evaluating photoactive behavior is the incident photon-to-current efficiency (IPCE), representing the photocurrent collected per incident photon flux as a function of incident-light wavelength. The IPCE could be calculated at a desired bias voltage from eqn (4):^52,53

where J is the photocurrent density (mA cm⁻²) under illumination at wavelength λ, λ is the wavelength of incident light (nm), and I corresponds to the incident light intensity (mW cm⁻²).

3. Strategies to enhance WO₃ photocatalytic activity in PEC water system

3.1. Morphological manipulation

Compared with bulk materials, nanostructured materials with a higher specific surface area and more reaction active sites can enhance the photon absorption to improve the photoelectric conversion efficiency.^37,54 The size of bulk WO₃ is larger than 150 nm, and restricts the hole migration to the surface of WO₃. Thanks to a shorter hole transport distance, the nanomaterials could improve the separation efficiency of photogenerated electron–hole pairs and yield a more rapid photoexcited hole injection to the electrolyte. However, small grains can reduce the activity in the visible light region.⁵⁵ Furthermore, the defects on the crystal surface make it possible for intensive electron–hole recombination which reduces the PEC water splitting performance.³⁴ Thus, developing an appropriate grain size, structure and morphology of the photocatalyst holds the key to improve the photocatalytic performance. The widely used methods for synthesizing WO₃ nanostructures include solvothermal,⁵⁶ hydrothermal,^39,57 anodization,⁵⁸ glancing angle deposition (GLAD),⁵⁹ pulsed laser deposition (PLD),⁶⁰ sol–gel,⁶¹ chemical vapor deposition (CVD),⁶² and electrodeposition.⁴⁰ Meanwhile, the fabrication methods were found to affect the grain boundaries and the dimension of the nanostructures, which limited the efficiency of water splitting. The morphology, fabrication method and PEC water splitting performance of the WO₃ photoanode films are illustrated in Table 1. Amongst the myriad of techniques, hydrothermal/solvothermal techniques have been commonly used to synthesize nanostructure WO₃ films. During the synthesis, introducing a very thin WO₃ dense layer has been found to be a prerequisite for the crystal growth that can hinder the interfacial recombination and decrease the photogenerated electron injection back into the electrolyte on the exposed areas of FTO.^59,63 Meanwhile, involving seed layers contributes to form grain boundaries in the nanocrystalline base film which can reduce the photocatalytic activity.^64–66

Table 1 Representative morphology of the WO₃ photoanode for PEC water splitting

Topography	Fabrication method	Film texture	Photoelectrochemistry	Ref.
Nanoplate	2-Step hydrothermal		Photocurrent: 3.7 mA cm^−2 at 1.23 V vs. RHE; IPCE: 67% at 350 nm at 1.23 V vs. RHE; electrolyte: 0.1 M Na2SO4	32
Nanoflake	Solvothermal		Photocurrent: ∼1.43 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼70% at 480 nm at 1.23 V vs. RHE; electrolyte: 0.1 M Na2SO4	56
Nanorod	Hydrothermal		Photocurrent: 2.26 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼90% at 350 nm at 1.23 V vs. RHE; electrolyte: 0.5 M H2SO4	57
Microcrystal	Hydrothermal		Photocurrent: 0.45 mA cm^−2 at 0.8 V vs. RHE; IPCE: ∼2.7% at 300 nm at 0.8 V vs. RHE; electrolyte: 1 M H2SO4	39
Nanoflake	Anodization		Photocurrent: 0.9 mA cm^−2 at 1.2 V vs. RHE; IPCE: ∼42% at 420 nm at 1.2 V vs. RHE; electrolyte: 0.5 M H2SO4	58
Nanorod	Glancing angle deposition (GAD)		Photocurrent: 2.15 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼40% at 385 nm at 1.23 V vs. RHE; electrolyte: 0.5 M KPi + 1 M Na2SO3	59
Tree-like nanoporous	Pulsed laser deposition (PLD)		Photocurrent: 1.8 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼78% at 350 nm at 1.23 V vs. RHE; electrolyte: 0.5 M KPi + 0.5 M H2SO4	60
Nanoplate	Water bath		Photocurrent: ∼1.42 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼38% at 400 nm at 1.23 V vs. RHE; electrolyte: 0.1 M Na2SO4	41
Nanoflake	Hydrothermal		Photocurrent: 2.25 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼23.25% at 440 nm at 1.23 V vs. RHE electrolyte: 0.1 M Na2SO4	67
Nanoplate	Hydrothermal		Photocurrent: 1.2 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼40% at 350 nm at 1.23 V vs. RHE electrolyte: 1 M H2SO4	68
Nanoparticle	Hydrothermal		Photocurrent: 2.7 mA cm^−2 at 1.4 V vs. RHE; electrolyte: 0.5 M H2SO4	69
Nanoneedle	Pulsed laser deposition (PLD)		Photocurrent: 2.4 mA cm^−2 at 1.23 V vs. RHE; IPCE: 50% at 410 nm at 1.23 V vs. RHE; electrolyte: 0.1 M H2SO4	70
Nanoflake	Solvothermal		Photocurrent: 1.1 mA cm^−2 at 1.23 V vs. RHE; IPCE: ∼45% at 320 nm at 1.23 V vs. RHE; electrolyte: 0.1 M Na2SO4	71
Nanoparticle	Polymer-assisted deposition		Photocurrent: 1.45 mA cm^−2 at 1.23 V vs. RHE; electrolyte: 0.1 M KPi	72

---

Note: WO₃ is only stable in acidic media of pH lower than 4. That is, the PEC properties of WO₃ may be investigated to be unstable in the electrolyte with 4 < pH ≤ 7. Researchers should investigate the PEC performance of WO₃ under strong acidic conditions in the future work.

3.1.1. 1-D WO₃ nanostructures. Due to the direct pathways for photogenerated carriers, one dimensional (1-D) semiconductor structures, e.g., nanowires,^{56,62,73–75} nanorods,^57,76–78 nanotubes,⁷⁹ and nanoneedles,⁸⁰ can efficiently facilitate the transportation of photoelectrons to FTO and thus suppress the recombination of photogenerated electron–hole pairs. These structures are capable of reducing grain boundaries and defects that result in less recombination of electron–hole pairs,⁵⁶ and consequently 1-D WO₃ nanostructure photoanodes demonstrated superior PEC properties compared to nanocrystalline particles.^81–84 For example, Chakrapani et al.⁶² fabricated WO₃ nanowire (length: 2–3 μm and diameter: 40–70 nm) arrays on the FTO substrate via chemical vapor deposition in ambient air. The blue film changed to greenish yellow corresponding to the crystal phase change from the monoclinic to orthorhombic phase. Under light irradiation AM 1.5 G, the WO₃ nanowire arrays exhibited a photocurrent of 1.2 mA cm⁻² at 1.2 V vs. SCE. These nanowires could capture the incident light more efficiently to enhance the PEC performance. Joo et al.⁵⁷ synthesized single crystalline structure WO₃ nanorods uniformly distributed on the FTO substrate using a facile hydrothermal method without the assistance of any seed layer. After being annealed at 500 °C, these nanorods exhibited an outstanding photocurrent of 2.26 mA cm⁻² at 1.23 V vs. RHE and an IPCE value of 35% under light irradiation at 400 nm. The excellent PEC performance was attributed to the direct pathway provided by the vertically aligned nanorod arrays on FTO with a desirable crystal phase.

3.1.2. 2-D WO₃ nanostructures. A large number of grain boundaries in the nanoparticle films increase the resistance and interfacial charge recombination, and thus impede the electron transfer to the back-contacted conductive substrate.⁶⁴ Different from nanoparticles and similar to 1D semiconductor structures, 2D nanostructures, e.g., nanoplates,^{64,65,85–87} nanosheets,^88,89 and nanoflakes, are favorable for highly efficient and directional transport of electrons and holes.^64,85,90 Besides, 2D semiconductors have a bigger surface/volume ratio than the 1D semiconductor.⁹¹ WO₃ nanoplates are reported to exhibit an enhanced PEC performance compared with nanorods and nanowires.^56,92 Su et al.⁵⁶ synthesized WO₃ nanowires and nanoflakes under almost the same conditions via a seed-layer-assisted solvothermal method. Both nanowires and nanoflakes were well vertically grown on the substrate (Fig. 2a–c). The WO₃ nanoflake photoanodes demonstrated a significantly higher PEC performance than WO₃ nanowires. The band gaps of the WO₃ nanowire and WO₃ nanoflake were calculated to be 2.92 and 2.51 eV, respectively. Under the same photocurrent measurement conditions, the WO₃ nanoflake film delivered about 3 times higher photocurrent than that of the nanowire film with a comparable thickness (Fig. 2e). Meanwhile, the WO₃ nanoflake films exhibited a higher IPCE value compared to the WO₃ nanowire film (Fig. 2f). The excellent photoelectric performance originated from the lower band gap of WO₃ nanoflakes and reduced light scattering in the flake array structure (Fig. 2d). Under light irradiation below 400 nm, the WO₃ nanoflakes with a thickness of 5.6 μm were found to exhibit a higher photocurrent value of 1.43 mA cm⁻² and a higher IPCE value of higher than 60%.

Fig. 2 FESEM images of (a) unannealed WO₃ nanowire, (b) unannealed WO₃ nanoflake (NF1) (c) unannealed WO₃ nanoflake (NF2), (d) photograph of WO₃ films as-prepared and after annealing, (e) current–potential plots for annealed nanowire, and two flake samples, under chopped visible light in an aqueous solution of 0.1 mol L⁻¹ sodium sulfate (Na₂SO₄), (f) IPCE of three samples with a bias of 0.5 V in a two electrode setup with Pt foil as the counter electrode. Reproduced from ref. 56. Copyright 2011 American Chemical Society.

3.1.3. Nanoporous structure. Compared with the bulk material, the nanostructured semiconductors with high porosity have shown a better photoelectric activity due to their large specific surface areas, more active sites, higher light absorption rates and excellent charge transport performance.^93–96 With a larger specific surface area, it could facilitate ion transport into nanostructured photoelectrodes. Furthermore, the nanoporous structure increases the area of the depletion layer and shortens the distance of diffusion for holes to the photoelectrode/electrolyte interface, thereby inhibiting surface recombination of electrons and holes.^60,97,98 However, there are some adverse effects when increasing the surface area. For example, the increased surface defects and grain boundaries may be formed, thereby reducing the photocatalytic activity. Therefore, it is crucial to design photoanodes with optimal active sites for photoactivities.⁵⁹ Shin et al.⁶⁰ prepared nanoporous WO₃ having a tree-like structure via a pulsed laser deposition method. The photoanodes have a monoclinic WO₃ crystal structure and are partially aligned along the [020] direction. Moreover, the nanoporous WO₃ showed 9-fold photocurrent density compared to the dense WO₃. The improved PEC performance can be mainly attributed to a higher surface/active area of the nanoporous structure, which allowed the electrolyte to permeate through the open channels to reduce the transport distance of holes, and restrained the recombination of the surface electron–hole. Song et al.⁹⁹ used a versatile foaming-assisted electrospinning method to prepare mesoporous WO₃ nanobelts which enhanced the PEC water splitting performance compared with the as-prepared WO₃ nanofiber and WO₃ nanobelt samples. Hilliard et al.¹⁰⁰ synthesized mesoporous WO₃via a facile sol–gel dip coating method. Compared with the dense WO₃ thin films, the mesoporous WO₃ photoanode exhibited a more substantial photocurrent response which was ascribed to massive specific surface area exposure to the electrolyte. This confirmed that the surface oxidation kinetics, optical absorption and conductivity of WO₃ limited the PEC activity of the mesoporous WO₃ photoanodes. Concerning neutral or basic solutions, the peroxo species formed on the surface of WO₃ reduced the stability of photoanodes.

3.1.4. Hierarchical 3D structure. Apart from the aforementioned morphologies, new structures and hierarchical 3D WO₃ photoanodes were designed to provide high solar conversion efficiencies. For example, Feng et al.²⁹ synthesized sandwich structured WO₃ nanoplate arrays via a hydrothermal method followed by annealing at 500 °C. The sandwich structure exhibited an optimal PEC performance with a photocurrent density of 1.88 mA cm⁻² at 1.3 V vs. Ag/AgCl. The highest PEC performance of the sandwich structured WO₃ compared with other samples annealed at different temperatures arose from the more active monoclinic phase, the smallest interface resistance and the largest electrochemically active surface area. Cai et al.¹⁰¹ reported another important strategy to prepare hierarchical 3D WO₃ micro–nano architectures which could be easily tuned from porous nanoparticle aggregates (WO₃ PAs) to nanowire aggregates (WO₃ NWs) by controlling the heating temperatures during the thermal oxidation (Fig. 3a–d). With increasing temperatures, the photocurrent density of three types of WO₃ nanostructures was decreased (Fig. 3e) due to the decreased exposure of (002) facets. The porous nanoparticle structure annealed at 600 °C significantly enlarged the specific surface area and promoted the electrolyte penetration into the electrode and the bottom of aggregates, thereby improving the PEC performances (Fig. 3f).

Fig. 3 (a) Schematic of the strategy of tuning WO₃ nanostructures. SEM images of (b) WO₃ PAs, (c) WO₃ PAs + NWs and (d) WO₃ NWs in different magnifications, (e) photocurrent density of three types of WO₃ nanostructures vs. potential plots, (f) schematic of the hierarchical structure with porous WO₃ nanoparticle aggregates and the corresponding charge transport/transfer and electrolyte permeation processes. The olive color and the arrows represent the electrolyte (0.1 M H₂SO₄) and the penetration of the electrolyte, respectively. Reproduced from ref. 101. Copyright 2017 American Chemical Society.

3.2. Crystal phase control

An efficient PEC water splitting performance depends on the photoanode; not only on the morphology but also the crystal phase of WO₃. The monoclinic phase of WO₃ has been demonstrated as the most stable phase at room temperature and its water oxidation activity is stronger than that of the hexagonal phase or orthorhombic phase WO₃.^{29,41,57,77,102–104} With increasing annealed temperatures, WO₃·0.33H₂O synthesized from the precursor changed from orthorhombic to anhydrous hexagonal and then to stable monoclinic (Fig. 4c). As the heating temperature increased from 400 to 500 °C, the absorption edge showed a continuous red shift and the film turned from greyish to yellow (Fig. 4a). The orthorhombic WO₃·0.33H₂O demonstrated the lowest photocurrent density, while the monoclinic WO₃ annealed at 500 °C exhibited an optimal PEC performance (Fig. 4b).^{29,41,57,64,65,77,82}

Fig. 4 (a) UV–vis absorption spectra of WO₃ films before and after annealing at different temperatures. The inset is the corresponding photographs, (b) linear sweep voltammetry of WO₃ photoanodes annealed at different temperatures under chopped incident light, (c) schematic illustration of unit cells of orthorhombic WO₃·0.33H₂O, hexagonal WO₃, and monoclinic WO₃. Reproduced from ref. 29. Copyright 2016 American Chemical Society.

Nayak et al.¹⁰⁵ synthesized WO₃·H₂O nanoplates via a facile precipitation method, which were evolved into WO₃ nanowires on subsequent solvothermal treatment (Fig. 5a–d). The WO₃·H₂O nanoplates had an orthorhombic phase, while WO₃ nanowires showed a monoclinic phase (Fig. 5e). The WO₃ nanowires exhibited outstanding PEC activity with a photocurrent density ∼21 times higher than the WO₃·H₂O nanoplates. As discussed above, (section 3.1), the nanoplates (2-D nanostructure) normally have a higher photocurrent value than the nanowire (1-D nanostructure). The significantly enhanced PEC performance of WO₃ nanowires was attributed to different phases between WO₃·H₂O nanoplates and WO₃ nanowires. The high crystallinity of WO₃ nanowires reduced the interfacial charge transfer resistance, and thereby improved the PEC performance.

Fig. 5 FESEM images of (a) stacked WO₃·H₂O square nanoplates synthesized at room temperature using 0.025 M of WCl₆ solution in ethanol for 1 h. FESEM images of WO₃ nanowires evolving from stacked WO₃·H₂O nanoplates after (b) 3 h, (c) 6 h, and (d) 12 h solvothermal treatment of WO₃·H₂O nanoplates at 200 °C in ethanol. (e) XRD patterns of these obtained samples. Reproduced from ref. 105. Copyright 2017 American Chemical Society.

Extensive research efforts have been made to explain the effects of crystal phase on the photocatalytic performance. For example, Park et al.⁸² reported that during the annealing process, with the transformation of the crystal phase, the removal of water induced the surface disorder or oxygen deficiencies of WO₃ and thus affected the PEC performance. Zeng's group considered that peroxo species were easily formed on the surface of WO₃·H₂O due to its inferior crystallinity, while WO₃ annealed at 500 °C had the highest current density because of the conversion of exposed (200) facets to highly reactive (002) facets, superior crystallinity and fewer defects to decrease the recombination of electron–hole pairs.⁴¹ This suggestion was also confirmed by Su's group.¹⁰⁴ Fan's group⁷⁷ suggested that the monoclinic WO₃ obtained at 500 °C (WO₃-500) exhibited a red shift of the absorption edge which effectively separated the photoexcited electrons and holes and migrated the carriers to the photoanode surface. The presence of both orthorhombic and monoclinic phases along with defects for the WO₃-450 sample and the collapse of nanorods for WO₃-550 restricted the efficient separation and migration of charge carriers. Yang et al. attributed the enhanced PEC properties to the higher crystallinity and the reduced PEC properties for WO₃ calcined at temperatures higher than 500 °C to the highly damaged surface of the platelets.⁶⁴ The improved crystallinity via high calcination temperature facilitating a better charge separation was also proved by Nuraje's group in the photocatalysis of SrTiO₃.²³ Thus, it can be concluded that the as-prepared hydrated tungsten oxide needs to be calcined at 500 °C to obtain monoclinic WO₃ for an optimal PEC.

3.3. Engineering of crystalline structures

Recent studies on semiconductor surface engineering have demonstrated that photoexcited electrons and holes can be driven by different crystal planes. Therefore, certain crystal faces of semiconductors will be preferentially reduced, while other crystal faces are preferentially oxidized.¹⁰⁶ Controlling and exposing some high energy crystal surfaces can improve the photocatalytic and PEC efficiency of photoelectrodes.^107,108 Compared with the (200) facet (1.43 J m⁻²) and the (020) facet (1.54 J m⁻²), the (002) facet of WO₃ possesses the highest surface energy (1.56 J m⁻²),¹⁰⁹ which is more favorable for adsorbing the reaction species on the (002) facet to decrease the surface energy. It is verified by Wang et al. through density functional theory (DFT) calculations.³² Due to the weakest W–O bond in the monoclinic WO₃, the dangling O atoms on the (002) crystal plane could provide more abundant active sites for H₂O and organics through the hydrogen bond. Therefore, photo-excited holes are more easily consumed on the (002) crystal plane involving the oxidation of water and the degradation of organics by generating active oxygen species, thereby impeding the recombination of photogenerated electrons and holes. Consequently, the exposure of highly reactive facets could effectively improve the PEC performance in water splitting and the degradation activity of organic pollution.^92,110,111

Zeng's group⁴¹ prepared WO₃ nanoplate array film photoanodes with a preferential orientation of highly reactive (002) facets by the reduction of peroxotungstates. The effects of synthesis time, temperature, and the amount of the capping agent on the morphologies and nanostructures of the as-prepared WO₃ films were discussed. The TEM images and SAED patterns indicated that the WO₃·H₂O plates grew along the crystal of (020) face which were converted into WO₃ plates with preferentially exposed (002) facets via annealing. The sample annealed at 500 °C exhibited the highest photocurrent density which was ∼1.42 mA cm⁻² at 1.23 V vs. RHE. The enhanced performance of PEC water splitting was ascribable to the superior crystallinity which could impede the formation of peroxo species on the surface of WO₃. The exposed highly reactive (002) facets of WO₃ nanoplates significantly improved the performance of PEC water splitting. WO₃ photoanodes decorated with a Co–Pi co-catalyst improved the water oxidation kinetics and reduced the photocorrosion resistance. Zhang et al.⁹² synthesized monoclinic WO₃ nanomultilayers with (002) facets exposed via a solvothermal method. The sandwiched WO₃ nanomultilayers exhibited photocurrent densities higher than the WO₃ nanorods synthesized as a reference. The specific surface area of WO₃ nanorods and nanomultilayers indicated that morphology is not the decisive factor for the enhanced photocurrent of the nanomultilayers. It is the preferentially exposed highly reactive (002) facets of WO₃ nanomultilayers that contributed to the improved PEC water splitting performance.

Although WO₃ exhibits high stability and resistance against photo-corrosion in acids under solar illumination,^52,92 WO₃ nanostructures still suffer from photocorrosion and photocatalytic deactivation because of the formation and accumulation of peroxo species on the surface of WO₃.^33,59,112 In PEC water splitting, photogenerated holes can cause a series of oxidation reactions, including the oxidation of surface hydroxyl groups of WO₃ or the oxidation of water/hydroxide. During the water oxidation reaction, oxygen and peroxo species will be formed. Thermodynamically, the photo-oxidation of water to oxygen (E° = 1.23 V) is more feasible than the formation of peroxo species (E° = 1.78 V). However, the peroxo species become kinetically competitive with oxygen evolution because of the slow kinetics of the oxygen evolution reaction. More importantly, the formation of peroxo species at the WO₃ surface can cause an obvious alternate photo-oxidation reaction, which may result in the loss of photoactivity or photodissolution of WO₃. To address the aforementioned issues, Wang et al.³² synthesized WO₃ nanoplates with highly reactive (002) oriented facets via a 2-step hydrothermal method and evaluated the PEC conversion performance (Fig. 6a–f). The 1-step-16 h WO₃ nanoplates were prepared via a hydrothermal method for 16 h, while the 2-step-16 h WO₃ nanoplates were obtained through the 1-step-8 h WO₃ nanoplates after hydrothermal treatment for another 8 h. All of the nanoplates were composed of dense and vertically aligned nanoplate arrays (Fig. 6a–c). The XRD spectra show that the 1-step-16 h WO₃ nanoplates were mainly exposed (200) facets (Fig. 6d) and the 2-step-16 h WO₃ nanoplates were enriched (002) facets (Fig. 6f). During the 1-step-16 h hydrothermal reaction, the WO₃ nanoplates mainly had the exposed (002) facets at the initial 8 h of hydrothermal reaction and were then converted into (200) facets exposed in the subsequent 8 h of hydrothermal reaction. This transformation can be explained by the theory of dissolution and recrystallization. Different from the 1-step-16 h WO₃ nanoplates, only a partial recrystallization occurred for the 2-step-16 h WO₃ nanoplates, which resulted in a larger (200)/(002) ratio. The prepared 2-step-16 h WO₃ nanoplates exhibited a remarkable photocurrent density of 3.7 mA cm⁻¹ (1.23 V vs. RHE) which corresponded to ∼93% of the theoretical photocurrent of WO₃ (Fig. 6g). The curves of chopped transient photocurrent density vs. time showed that the WO₃ nanoplates prepared by a 2-step hydrothermal method exhibited the highest photocurrent density and the best stability (Fig. 6h). DFT calculations together with experimental studies demonstrated that the exposing highly reactive (002) facet and nanoplate structures facilitated the separation of photoelectron–hole pairs and suppressed the formation of peroxo-species on the surface, and thus enhanced the PEC activity and retained a better photo-stability (Fig. 6i). As a result, the highly oriented (002) crystal plane could significantly improve the PEC water splitting performance and the stability. Accompanied by the 2-D nanostructure and highly active crystal monoclinic phase, more research should be focused on the use of 2-D monoclinic WO₃ with a preferentially exposed (002) crystal plane for PEC water splitting in the future.

Fig. 6 FESEM images and XRD patterns of (a, d) 1-step-16 h, (b, e) 1-step-8 h, (c, f) 2-step-16 h photoanodes prepared by the hydrothermal method, (g) photocurrent density vs. applied potential curves, (h) photocurrent density vs. time measured at 1.23 V vs. RHE, and (i) free energy diagram of the (200) facet and (002) facet at the potential of 1.23 V. Insets: The models that show the surface status of every reaction step: tungsten atoms are represented by gray spheres, oxygen atoms by red spheres, and protons by white spheres, while oxygen gas is not shown in the models. Reproduced from ref. 32. Copyright 2016 Elsevier.

3.4. Forming composites or heterojunctions

Since pristine WO₃ nanomaterials suffer from low solar spectrum absorption (about 12%), slow charge transfer at the electrode/electrolyte interface and rapid electron–hole recombination, it is of great interest to address these problems to improve the photoelectrocatalytic properties of WO₃ by constructing heterojunctions. By connecting two n-type materials with similar CB and VB energies, one can create a type II heterojunction which can localize electron and hole wave functions in the core or shell material, and thereby improve the charge separation.^113,114 Otherwise, coupling an n-type semiconductor together with a p-type semiconductor to create p–n junctions¹¹⁵ or Z-scheme systems can significantly promote efficient charge separation. To date, a variety of semiconductors are coupled with WO₃ to construct a heterojunction, such as WO₃/Bi₂MoO₆,¹¹⁶ α-Fe₂O₃/WO₃,^117,118 WO₃/CuWO₄,^85,119 NiWO₄/WO₃,¹²⁰ Bi₂S₃/WO₃,¹²¹ WO₃/BiVO₄,^122–125 WO₃/TiO₂,^113,126 ZnWO₄/WO₃,¹²⁷ and NiO/WO₃,¹²⁸ to enhance the photocatalyst or photoelectric conversion performance. Part of the binary heterojunction, fabrication methods and their PEC water splitting performance are summarized in Table 2. Some ternary heterojunction structures, such as WO₃/rGO/Sb₂S₃,¹²⁹ where rGO serves as the conductive layer, have been reported to enhance the charge transfer. A variety of semiconductors with conduction and valence band energies vs. NHE are shown in Fig. 7a. It is well known that the band gap positions and the Fermi level directly determine the electron transport in the heterostructure systems. In the case of WO₃, coupling with a narrow band gap semiconductor to form a type II heterojunction is an effective way to improve the photocatalytic performance. Normally, the photogenerated electrons could be easily transferred from the CB of the secondary material to the CB of WO₃ while driving photogenerated holes from the VB of WO₃ to the VB of a coupled semiconductor which possesses a narrower band gap energy (Fig. 7b). Different from the n-type semiconductors where electrons are employed as the photogenerated carriers, the number of holes is much larger than that of electrons for p-type semiconductors. Therefore, in p–n type heterojunctions or Z-scheme systems, the photo-excited electrons from the CB of WO₃ can readily recombine with the photogenerated holes from the VB of the p-type semiconductor. It is significant to optimize the reduction and oxidation ability for the PEC water splitting (Fig. 7c).

Fig. 7 (a) Bandgap energy, VB and CB positions of several semiconductors on a potential scale (V) vs. NHE/vacuum. (b) Schematic illustration of type II heterojunction and (c) p–n junction band alignments, and the correspondingly possible photoexcited electron–hole pair separation and transfer between the contacted semiconductors.

Table 2 Representative WO₃ heterojunction nanostructure for PEC water splitting

Heterostructure	Fabrication method	SEM/TEM image	Photoelectrochemistry	Ref.
CuWO4/WO3	Hydrothermal and then dipping-annealing process		Photocurrent: 1.21 mA cm^−2 (WO3: 0.64 mA cm^−2) at 1.5 V vs. Ag/AgCl; IPCE: 55.3% at ∼350 nm at 1.2 V vs. Ag/AgCl; electrolyte: 0.2 M Na2SO4	85
WO3/Sb2S3	Hydrothermal		Photocurrent: 1.79 mA cm^−2 at 0.8 V vs. RHE; IPCE: ∼12.5% at 300 nm at 0.8 V vs. RHE; electrolyte: 1 M H2SO4	39
α-Fe2O3/WO3	Hydrothermal and then followed by deposition-annealing		Photocurrent: 1 mA cm^−2 (WO3: 0.36 mA cm^−2) at 1.23 V vs. RHE; IPCE: 30.4% at ∼330 nm at 1.23 V vs. RHE; electrolyte: 0.1 M Na2SO4	117
NiWO4/WO3	Hydrothermal and followed by deposition-annealing		Photocurrent: 0.44 mA cm^−2 (WO3: ∼0.25 mA cm^−2) at 1.4 V vs. Ag/AgCl; IPCE: 40.7% at 340 nm at 1.2 V vs. Ag/AgCl; electrolyte: 0.2 M Na2SO4	120
WO3/Cu2O	Hydrothermal and then followed by electrodeposition		Photocurrent: 1.37 mA cm^−2 (WO3: 0.39 mA cm^−2) at 0.8 V vs. RHE; IPCE: 10.7% at ∼300 nm at 0.8 V vs. RHE; electrolyte: 1 M H2SO4	130
WO3/Bi2S3	Hydrothermal and followed by chemical bath deposition		Photocurrent: 5.95 mA cm^−2 (WO3: 0.17 mA cm^−2) at 0.9 V vs. RHE; IPCE: 68.8% at ∼435 nm at 0.9 V vs. RHE; electrolyte: 0.1 M Na2S + 0.1 M Na2SO3	131
WO3/BiVO4	Soaking hydrothermal-calcining		Photocurrent: 2.92 mA cm^−2 at 1.81 V vs. RHE; electrolyte: 0.5 M KH2PO4	132
WO3/FeOOH	Spray pyrolysis deposition and followed by electrodeposition		Photocurrent: 1.4 mA cm^−2 (WO3: 0.8 mA cm^−2) at 1.23 V vs. RHE; electrolyte: 0.5 M K2SO4	133
WO3/BiVO4	Hydrothermal and followed by spin coating		Photocurrent: 3.17 mA cm^−2 (WO3: 1.41 mA cm^−2) at 1.23 V vs. RHE; electrolyte: 0.1 M Na2SO4	134
WO3/BiVO4	Solvothermal followed by spin-coating		Photocurrent: 1.6 mA cm^−2 (WO3: 0.4 mA cm^−2) at 1 V vs. RHE; IPCE: 31% at ∼420 nm at 0.5 V vs. RHE; electrolyte: 0.5 M Na2SO4	135
WO3/FeOOH	Hydrothermal and followed by electrodeposition		Photocurrent: 1.3 mA cm^−2 (WO3: 0.6 mA cm^−2) at 1.23 V vs. RHE; electrolyte: 0.1 M KPi buffer	136

---

3.4.1. WO₃/chalcogenide heterojunction. Among various semiconductor materials, narrow bandgap metal sulfides have been widely used as photosensitizers for various wide bandgap semiconductor photoanodes.^137,138 As the position of the CB and VB of most metal sulfides is more negative than that of WO₃, respectively, coupling of WO₃ with metal sulfides to form heterojunctions is an optimized way to improve the PEC performance. Zhang et al. first reported the WO₃/Sb₂S₃ heterojunction photoelectrocatalyst.³⁹ They fabricated three types of WO₃ (WO₃ nanoplates, WO₃ nanorods and WO₃ microcrystals) and their Sb₂S₃ heterojunction photoelectrocatalysts by a facile hydrothermal method. According to the UV-vis absorbance spectra and Tauc formula,^85,139 the band gaps of WO₃ nanoplates, WO₃ nanorods and microcrystals were calculated to be 2.96, 2.82, and 2.75 eV, respectively. Meanwhile, the band gaps for the Sb₂S₃ heterojunction were calculated to be 2.75, 2.50 and 2.28 eV, respectively. The decreased band gaps after coupling with narrow band gap Sb₂S₃ indicated the improvement of visible light absorption which could effectively improve the photoelectrocatalytic performance. The WO₃/Sb₂S₃ heterojunction could strengthen the migration of charge carriers at the WO₃/Sb₂S₃ interface, which in turn effectively facilitated the separation of photogenerated electron–hole pairs. Compared with the pristine WO₃ nanoplates, nanorods and microcrystals, the photocurrent density of the corresponding WO₃/Sb₂S₃ heterojunction photoelectrocatalysts was improved by 258%, 308%, and 398%, respectively. Significantly, an excellent and efficient WO₃/Bi₂S₃ composite was reported by Wang's group.¹³¹ They designed a 3D WO₃ nanoplate/Bi₂S₃ nanorod heterostructure photoanode via a hydrothermal process, followed by chemical bath deposition. Based on the UV-vis absorption spectra, the bare WO₃ showed poor absorption in the visible-light region compared with the WO₃/Bi₂S₃ heterostructure (the blue plot of Fig. 8b). Moreover, the WO₃/Bi₂S₃ heterostructure with a seed layer (the red plot of Fig. 8b) further greatly enhanced the absorption in the visible-light region (Fig. 8b). Beneficial for the more negative CB edge of Bi₂S₃ compared to that of WO₃ and excellent contact quality, the photo-excited electrons in the CB of Bi₂S₃ can easily move into the CB of WO₃ and the photo-excited holes in the VB of WO₃ will in turn flow to the VB of Bi₂S₃. This process is significant to enhance the charge transport efficiency, increases the carrier density and reduces the charge recombination, which contributes to the improved PEC activity and stability (Fig. 8a). The as-prepared WO₃/Bi₂S₃ composites with a seed layer exhibited an outstanding photocurrent (5.95 mA cm⁻¹ at 0.9 V vs. RHE) which is 35 times higher than that of pristine WO₃ (0.17 mA cm⁻¹ at 0.9 V vs. RHE). The highest photoconversion efficiency was increased from 0.03% (at 0.92 V) of the pristine WO₃ to 1.70% (at 0.76 V) of the as-prepared composites (Fig. 8c and d).

Fig. 8 (a) Schematic of the WO₃/Bi₂S₃ heterojunction and the reaction process in the PEC cell, (b) UV-vis absorption spectra of bare WO₃ (black), WO₃/Bi₂S₃ without a seed layer (blue) and WO₃/Bi₂S₃ with a seed layer (red), (c) I–t curves of bare WO₃ (black), WO₃/Bi₂S₃ without a seed layer (blue), and WO₃/Bi₂S₃ with a seed layer (red) at 0.9 V vs. RHE, (d) photoconversion efficiency of pristine WO₃ (black), WO₃/Bi₂S₃ without a seed layer (blue), and WO₃/Bi₂S₃ with a seed layer (red) Reproduced from ref. 131. Copyright 2017 American Chemical Society.

3.4.2. WO₃/metal oxide heterojunction. Although metal sulfide semiconductors have smaller energy gaps, the photocorrosion limits their applications in the photocatalytic water splitting. Therefore, it is a crucial issue to develop a novel narrow bandgap semiconductor with high photocorrosion resistance. Metal oxides are known to be more stable under various reaction conditions, and consequently they appear to be suitable photocatalysts to couple with WO₃ to construct a heterojunction. For example, Zhang's group achieved a WO₃/Cu₂O heterojunction using an electrodeposition method.¹³⁰ The morphology of the as-prepared samples was rod-like (Fig. 9a and b). TEM showed that Cu₂O was evidently deposited onto the surfaces of WO₃ (Fig. 9c). A series of deposition times were investigated to disclose their impact on PEC water splitting (Fig. 9d). The heterojunction obtained with a deposition time of 240 s exhibited the highest photocurrent density which was attributed to the appropriate amount of Cu₂O nanoparticles that improved both the light absorption and the charge transfer (Fig. 9e). A smaller amount of Cu₂O on the surface of WO₃ created a large number of interface defects that intensified the recombination of photogenerated electrons and holes. In contrast, an excess amount of Cu₂O aroused lattice scattering and decreased the charge carrier concentration, and thereby reduced the PEC performance.

Fig. 9 (a) Schematic of WO₃NRs/Cu₂O array electrodes, (b) typical top view SEM images of WO₃NRs/Cu₂O, (c) TEM images of WO₃NR/Cu₂O arrays, (d) the photocurrent density–voltage curves of the samples, (e) schematic of the WO₃NR/Cu₂O array heterojunction photoanode applied in overall water splitting and simplified schematic illustration of the band-gap energy diagram, showing the enhanced light-harvesting and charge-transfer processes. Reproduced from ref. 130. Copyright 2017 Elsevier.

Zhan et al.⁸⁵ fabricated CuWO₄/WO₃ heterojunction photoanodes and demonstrated that coupling with CuWO₄ enhanced the photocurrent density of WO₃ due to the extension of visible light response, efficient separation of photo-excited carriers at the WO₃/CuWO₄ interface and better electron transport properties. The WO₃ decorated with the exfoliated Bi₂MoO₆ nanosheets showed a satisfactory photocurrent density of 2.2 mA cm⁻² at 0.8 V vs. SCE compared to WO₃ (1.1 mA cm⁻²). The improved PEC performance of the WO₃/Bi₂MoO₆ heterojunction could be attributed to the suitable band position between WO₃ and Bi₂MoO₆ that facilitated the electron–hole separation and charge transfer. The uniformly exfoliated nanosheets coated on the WO₃ surface could also provide a high electric mobility.¹¹⁶

3.4.3. WO₃/BiVO₄ heterojunction. Coating BiVO₄ on the surface of WO₃ to construct a heterojunction could enhance the migration of charge at the interface and significantly increase the light absorption.⁵⁹ BiVO₄ is a highly promising photoanode because of a narrow bandgap of 2.4 eV which leads to the visible light absorption capability, and has an advantageous portion of the CB near the thermodynamic hydrogen evolution potential.^140–142 While the short hole diffusion length and the strong recombination of photoexcited charge carriers limited its performance,¹⁰⁴ it is worth noting that when coupling with BiVO₄ on the surface of WO₃, the photogenerated electrons from the CB of BiVO₄ (+0.02 V vs. RHE) can be readily injected into the CB of WO₃ (+0.41 V vs. RHE) and the photogenerated holes left in the VB of WO₃ will be transferred to the VB of BiVO₄ due to the matched band edge positions between WO₃ and BiVO₄.^{122,142–144} This process can effectively promote the separation of photoexcited electrons and holes. In the WO₃/BiVO₄ heterojunction structure, WO₃ mainly acted as an electron conductor and BiVO₄ served as the primary light absorber.^{103,124,145,146} These combined effects could greatly inhibit the recombination of photogenerated electrons and holes, thereby further increasing the photoelectric conversion performance.¹³² Serving as a suitable oxygen evolution catalyst, coupling with BiVO₄ could restrict the formation of peroxo species and increase the stability by blocking the contact of WO₃ and corrosive electrolytes.^40,133 Since the first demonstration by P. Chatchai et al.,¹⁴⁷ many researchers have paid more attention to improve the PEC activity of WO₃/BiVO₄ heterojunction films.^{103,104,122,123,134,135,142,146,148–150} For example, Lee et al. emphasized that using highly ordered 1D WO₃ nanostructures with a high aspect ratio and active areas is an effective way to improve the photocatalytic activity of WO₃/BiVO₄ heterojunctions. They compared the WO₃/BiVO₄ heterojunction with the same set of WO₃ nanorods and showed that the WO₃/BiVO₄ anodes exhibited higher photocurrent densities than bare WO₃ nanorods.⁵⁹ The WO₃/BiVO₄ heterojunction fabricated by a simple physical deposition and calcination process may cause numerous grain boundaries and increased impedance at the interface of the WO₃/BiVO₄ films.¹⁵¹ Consequently, an in situ synthesis method was preferred to obtain high quality WO₃/BiVO₄ films. For example, Iqbal et al.¹⁵¹ synthesized the nanorod-like WO₃/BiVO₄ hetero-nanostructures with a single phase by a one-step hydrothermal reaction for the first time. In these hetero-nanostructures, BiVO₄ was embedded into the grown nanorods of WO₃. The novel nanorod-like WO₃/BiVO₄ hetero-nanostructure photoanode exhibited a significant photoresponse. The in situ transformed WO₃/BiVO₄ hetero-nanostructure photoanode¹³² exhibited the highest photocurrent density compared to WO₃, BiVO₄, Bi₂WO₆/WO₃ and BiVO₄/WO₃ composites that were prepared via a simple physical deposition. The ratio of BiVO₄ in the WO₃/BiVO₄ hetero-nanostructure influenced the PEC water splitting performance. Seo et al.¹⁴⁹ synthesized WO₃/BiVO₄ heterojunctions with various ratios of BiVO₄ and obtained the optimum ratio of complexes. When the concentration of BiVO₄ was lower, WO₃ was the major photon-absorbing semiconductor and the photogenerated holes could readily recombine at the interface of hetero-nanostructures, while the photogenerated electron–hole pairs were not fully utilized when the thickness of the BiVO₄ layer was larger than 900 nm due to the disappointing charge carrier mobility of BiVO₄. Therefore, the composites with appropriate WO₃ and BiVO₄ exhibited a better PEC water splitting performance.

It is an appealing way to enhance the performance of PEC water splitting by designing special nanostructure semiconductor composites. The core–shell structure is reported to be an efficient system to maximize each separate process optimally which is beneficial for water splitting performance.^{84,104,113,124,135} By increasing the surface and interface area of the WO₃/BiVO₄ heterojunction, the core–shell heterojunction of WO₃/BiVO₄ promoted the generation and separation of photoexcited electrons and holes. The framework of the core–shell nanostructure can capture multiple reflections of sunlight owing to its cavity which is beneficial for an efficient utilization of solar photons to improve the photocatalytic activity.¹⁴² Therefore, it is critical to develop an optimized WO₃/BiVO₄ core–shell heterojunction to enhance the PEC water splitting properties. Rao et al.¹²⁴ synthesized a core–shell structure photoanode incorporated with W-doped BiVO₄ and WO₃. The as-prepared sample exhibited excellent PEC water splitting properties with a high charge separation and surface transfer efficiency. Due to the short hole diffusion length attributed from the core–shell structure, the holes generated in the BiVO₄ shell can readily reach the semiconductor/electrolyte interface and thereby improve the photoelectric conversion efficiency. Jin's group¹⁴² prepared yolk–shell structured WO₃/BiVO₄ heterojunctions with BiVO₄ coating onto the surface and inside of WO₃ shells. As shown in Fig. 10a–f, the hollow WO₃ structure, yolk–shell WO₃ spheres and yolk–shell WO₃/BiVO₄ heterojunctions were clearly observed. The HR-TEM image (Fig. 10f) and the corresponding energy-dispersive X-ray (EDX) elemental mapping (Fig. 10g–j) indicated the intimate contact in the WO₃/BiVO₄ heterojunction structure. Compared with hollow WO₃, the enhanced photoelectrocatalytic properties of the yolk–shell WO₃ were attributed to the particular yolk–shell structure with a beneficial higher charge-carrier separation and a restricted recombination of the photogenerated electrons and holes. After introducing the FeOOH/NiOOH layer as an oxygen evolution reaction (OER) co-catalyst on the surface of the yolk–shell WO₃/BiVO₄ heterojunction, the composites showed a strong negative shift in the photocurrent onset potential and exhibited the highest photocurrent density (Fig. 10k). In this yolk–shell WO₃/BiVO₄/OER heterojunction, the yolk–shell WO₃ served as an electron conductor while BiVO₄ functioned as the main light absorber to form a type-II heterojunction structure. The photogenerated electrons migrated from the CB of BiVO₄ to the CB of WO₃. On the other hand, the photogenerated holes left in the VB of WO₃ were injected into the VB of BiVO₄. This process greatly improved the charge transfer and reduced the recombination of photoexcited electrons and holes. The OER co-catalyst layer reduced the overpotential of water oxidation and subsequently improved the PEC water splitting performance.

Fig. 10 SEM and TEM images of the hollow WO₃ (A and D) and yolk–shell WO₃ (B and E), (C and F) SEM and HR-TEM images of the yolk–shell WO₃/BiVO₄ heterojunction, (G–J) the corresponding EDX elemental mapping of the yolk–shell WO₃/BiVO₄ (scale bars are 1 mm), (K) linear sweep voltammetry (LSV) curves of the samples in 0.5 M Na₂SO₄ solution under AM 1.5G light illumination, (L) schematic diagram for the construction of the yolk–shell WO₃/BiVO₄/OER photoanode and the energy diagram for PEC water splitting. Reproduced from ref. 142. Copyright 2018 Royal Society of Chemistry.

3.5. Construction of cocatalysts

3.5.1. Hydrogen evolution cocatalysts (HEC). To overcome the drawbacks of WO₃ including the slow kinetics of OER and rapid recombination of photogenerated electron–hole pairs, modifying with cocatalysts on the surface of WO₃ is a promising way to increase the efficiency of PEC water splitting.¹⁵² Loading of cocatalysts onto WO₃ can hinder the charge recombination by providing active sites to facilitate the redox reactions. Cocatalysts increase the trapping sites for the photo-excited electrons and holes, thereby enhancing the charge separation and improving the quantum efficiency.^38,153,154 In most cases, noble metals (such as Pd, Pt, Ir, Au, Ag, and Rh) have been employed on the surface of WO₃ as the hydrogen evolution cocatalysts (HEC). When a noble metal is loaded on the surface of WO₃, a Schottky barrier can be formed. The lower Fermi level of noble metals compared to WO₃ facilitated noble metals to trap electrons more readily. Thus, the electrons from WO₃ will be injected into the noble metal to equilibrate the Fermi levels. These excess negative charges in the noble metal and excess positive charges in WO₃ create a Schottky barrier in the interfaces which can prevent the charge recombination.⁴² The high electron conductivity of noble metal helps to promote the separation of photogenerated electron–hole pairs and enhances the properties of water splitting.^155–158 However, the high cost of noble metals limited their large-scale applications. Therefore, low-cost cocatalysts, such as MoS₂, NiS, WS₂ and C, have received more attention over the past few decades.^159–161 Among the low-cost HEC cocatalysts, nanocarbon-based cocatalysts have attracted wide attention owing to their high electrical conductivity, large surface area and nontoxicity.¹⁶² Shen and coworkers¹⁵⁹ synthesized WO₃ nanoflakes decorated with carbon quantum dots (CQDs) via a solvothermal method followed by subsequent impregnation-assembling. 1.93% CQDs/WO₃ exhibited an enhanced photocatalytic activity with a photocurrent density of 1.46 mA cm⁻² at 1.0 V vs. Ag/AgCl, which was about two times higher than that of the bare WO₃ electrode. This enhancement originated from the extended photoresponse range and facilitated the charge transfer of CQD/WO₃ nanocomposites.

3.5.2 Oxygen evolution cocatalysts (OEC). In overall photocatalytic water splitting, the OER involving a four-electron transfer process is a kinetically slow uphill reaction and becomes the control step for water splitting.^42,154,163 Decorating the oxygen evolution cocatalysts (OEC) with semiconductors can enhance the efficiency of O₂ evolution because the OEC helps to reduce the OER overpotential and suppresses the formation of peroxo species. To load OEC onto WO₃ is thus an effective way to increase the efficiency for water splitting. Thus far, RuO₂,¹⁶⁴ IrO₂,¹⁶⁵ Co–Pi,^166,167 Co₃O₄,¹⁶⁸ CoO_x,¹⁶⁹ MnO_x,¹⁷⁰ and NiO_x ¹⁷⁰ have served as OEC for the photocatalytic O₂ production. Amongst them, Co–Pi has received much attention due to its outstanding OER performance.^171,172 Seabold and co-workers⁴⁰ coupled WO₃ with a layer of Co–Pi to improve the photostability of WO₃ photoanodes and enhanced the photo-oxidation reaction. After photocurrent measurement for 2 h, the appearance of pores and the decrease of crystal grains of the WO₃ electrode indicate that the OER is not the sole photo-oxidation reaction occurring at the bare WO₃ electrode. Peroxo species at the WO₃ surface were formed resulting in a loss of PEC activity of the bare WO₃ electrode. After the deposition of the Co–Pi OEC layer on the WO₃ electrodes, the photocurrent remained stable over long time since Co–Pi OEC effectively inhibited the formation of peroxo species at the surface and prevented the photochemical deactivation of WO₃. Fan et al.⁷⁷ proposed a rational design via combining WO₃ and NiFe-layered double hydroxide (LDH) nanoarrays. The WO₃ nanorod arrays were prepared via a hydrothermal process and decorated with a NiFe-LDH by electrochemical deposition (Fig. 11a–c). They found that the deposition time of the NiFe-LDH significantly affected the PEC performances and obtained the best performance measured for the WO₃@NiFe-LDH-200s sample. The as-prepared sample exhibited excellent photocurrent density and higher stability compared with bare WO₃ nanorods (Fig. 11d). As an efficient OER catalyst, the NiFe-LDH could effectively facilitate the separation of photogenerated electron–hole pairs and enhanced water splitting kinetics. After coupling with the NiFe-LDH, the photoexcited holes in WO₃ preferentially oxidized the low valence Ni^II in the redox couples of LDH to the high valence Ni^III,IV, and then the high valence Ni^III,IV further oxidized water to release oxygen. This oxidation process has been demonstrated to be kinetically faster than water oxidation directly via photogenerated holes (Fig. 11e). Due to the promotion of water splitting kinetics and hindering of the charge recombination, the WO₃@NiFe-LDH-200s sample exhibited a superior PEC water oxidation performance. However, a thick OEC layer inevitably blocked the incident light and reduced the PEC activity of WO₃.^32,40 It is highly desirable to improve the stability of the WO₃ photoanode without sacrificing the photocatalytic activity.

Fig. 11 SEM images of WO₃@NiFe-LDH NR arrays obtained with various electrodeposition durations: (a) 50 s, (b) 200 s, (c) 400 s, (d) linear sweep voltammograms at a scan rate of 10 mV s⁻¹ (solid line: under visible light, dashed line: in dark) and (e) PEC water oxidation mechanism on the WO₃@NiFe-LDH NR arrays under visible light. Reproduced from ref. 77. Copyright 2016 Elsevier.

3.6. Doping or oxygen deficiency

There exist two main kinds of defects in WO₃. One is oxygen deficiency, which accelerates the charge carrier transfer at an optimized amount.^173,174 The other is the trap states that promote the recombination of electrons and holes. Serving as shallow donors, the bulk oxygen vacancies can improve electrical conductivity and facilitate charge transfer, while surface oxygen vacancies introduce a large number of defects which can be the center of recombination of photogenerated electron–hole pairs and thereby reduce the PEC performance.³⁸ Therefore, it is one of the key objectives to reduce the surface oxygen vacancies without affecting the bulk oxygen vacancies. Li et al. synthesized a series of substoichiometric WO_3−x nanoflakes and demonstrated that the enhanced photocurrent density was related to the increased donor density after introducing oxygen vacancies. The observed anodic shift of the hydrogen-treated samples was due to excess surface oxygen vacancies which aggravated the charge carrier recombination.³³ Li et al. introduced oxygen vacancies by adding TiCl₃. They demonstrated that the oxygen vacancies were introduced in proportion to the concentration of TiCl₃. Treatment by TiCl₃ at high concentration led to excess surface oxygen vacancies that resulted in a thicker disorder layer with more recombination centers.¹⁷⁵ Similar results were obtained by the treatment of WO₃via H₂O₂ to introduce surface oxygen vacancies.¹⁷⁶ Gong et al.⁶⁷ synthesized WO₃ nanoflake photoanodes with decreased surface oxygen vacancies but suitable bulk oxygen vacancies by a facile hydrogen annealing followed by an ozone treatment (named HO-WO₃). The remarkable enhancement in the photocurrent values from 1.09 mA cm⁻² for pristine WO₃ to 2.25 mA cm⁻² for the HO-WO₃ sample at 1.23 V vs. RHE was achieved. The improved performance of the HO-WO₃ sample was due to the increased bulk oxygen vacancies that served as shallow electron donors to enhance the conductivity of WO₃ and the decreased surface oxygen vacancies that reduced the recombination of photogenerated electrons and holes. However, the introduction of oxygen vacancies using hydrogen treatment resulted in high consumption of hydrogen gas, which hindered their applications in practical fields.

Concerning the traps, modifying the passivation layer on the surface of semiconductor particles can inhibit the recombination of photogenerated electrons and holes. Doping with a hetero-element can extend the photoresponse range and increase the carrier concentration. However, due to different radii between the donor atom and the host atom, the introduction of a dopant will cause a significant increase of lattice distortion and thus results in the increase of carrier recombination centers.^177–179 Doping with a hetero-element will decrease the solubility of dopant species and reduce the structural stabilities.^179,180 Therefore, it is the most critical issue to select the optimized doping element leading to an inconspicuous structural disturbance and sparse defect sites. Transition metals and lanthanides are able to intercept the electronic structure of WO₃ to extract more visible solar spectra.¹⁸¹ Doping Ti⁴⁺ into the WO₃ semiconductor is a feasible way to improve the PEC performance because impurity states between the CB and VB of WO₃ change the band structure of WO₃. Meanwhile, similar radii of the guest and the host result in less lattice distortion.¹⁸² Yb, as one of the lanthanides, is an electron-rich f-orbital element which increases the charge carrier generation for the enhanced photocurrent densities in WO₃. Liew et al.¹⁸¹ synthesized Yb-doped WO₃ photocatalysts by the method of co-sputtering of WO₃ and Yb. They demonstrated that the samples prepared under W and Yb sputtering powers of 100 W and 5 W, respectively, exhibited the highest photocurrent density. As the ionic radius of Yb³⁺ is 28% larger than that of W⁶⁺, it is easier for Yb³⁺ to contribute electrons from its outer shell 4f¹³ to the conductive carrier path in WO₃ which is beneficial for reducing the charge transfer resistances and enhancing the efficiency of charge transfer at the photocatalyst/electrolyte interface. Nevertheless, excess Yb³⁺ caused higher densities of grain boundary which acted as carrier recombination centers to reduce the hole–electron separation efficiency. Other metals, such as Bi,¹⁸³ Al,¹⁸⁴ Ni,¹⁸⁵ Co,¹⁸⁶ and Mo,¹⁸⁷ were employed as dopants for WO₃ to enhance the PEC water splitting performance. Since the introduction of metal elements into semiconducting oxides reduced the thermal stability, doping of nonmetal elements has attracted considerable attention.¹⁸⁸ Doping N into WO₃ significantly narrowed the band gap of WO₃ to a value lower than 2.0 eV and thereby greatly enhanced the photo-absorption performance.¹⁸⁹ In S-doped WO₃, some S atoms replaced O atoms on the surface of WO₃ and some S⁶⁺ more favorably substituted W⁶⁺ to form W–O–S bonds.¹⁹⁰ Oxygen vacancies increased to reduce the resistivity and increased the extrinsic absorption. Thus, it significantly increased the absorption of visible light and enhanced the PEC water splitting performance. Despite the fact that doping is an effective strategy to introduce oxygen vacancies and enhance visible light absorption, doping with hetero-elements resulted in a highly defective lattice and seriously limited their practical applications.

4. Z-Scheme water splitting

Although coupling with other n-type semiconductors to form type II heterojunctions can significantly improve the photocatalytic performance of WO₃, to achieve overall water splitting is limited without changing the original oxidation and reduction potential. Recently, inspired by natural photosynthesis in green plants, great attention has been paid to construct Z-scheme heterojunctions for the overall water splitting under visible light irradiation due to their efficient photogenerated electron–hole separation and optimized redox ability of the photocatalytic system.^191–194 Yousaf et al.¹⁹⁵ constructed efficient Z-scheme junctions consisting of oxygen-deficient WO_3−x nanorods and Zn_0.3Cd_0.7S nanoparticles and found that H₂ was produced in a repeatable manner from aqueous solutions containing Na₂SO₃ and Na₂S as sacrificial reagents without electric energy. Zhu et al.¹⁹⁶ prepared carbon nanodots (CNDs) from glucose via a simple hydrothermal process and found that the CNDs without any modification and doping could produce relatively low hydrogen under xenon lamp irradiation. More importantly, after combining with WO₃ to form a Z-scheme through an IO₃⁻/I⁻ shuttle redox mediator, the H₂ evolution rates of the as-prepared CND/WO₃ Z-scheme reached up to 1330 μmol g⁻¹ h⁻¹, which was about 286 times higher than that of pure CNDs.

Although the more positive VB than the oxygen generation potential makes WO₃ suitable for oxidizing water to generate oxygen, the positive position of the CB limits the reduction of H⁺ to H₂. Thus, WO₃ cannot be used for overall water splitting. A strategy to overcome the inherent limits is to introduce Z-scheme junctions to produce hydrogen from water splitting. The fundamental principle of the Z-scheme water splitting is shown in Fig. 12a. In this system, semiconductor I with the CB more negative than the hydrogen generation potential is combined with semiconductor II with the VB more positive than the oxygen generation potential through an appropriate shuttle redox mediator. Compared with one-step water splitting systems, Z-scheme junctions utilize visible light more efficiently due to the reduced energy required to drive each photocatalyst. It is thus possible to apply a photocatalyst that has either a water reduction or oxidation potential to one side of the system. A reversible redox mediator, noble metal particle or graphene has been used as the shuttle electron mediator. To date, there have been many reports on the Z-scheme junctions with appropriate shuttle redox mediators for overall water splitting.^197–199 However, many problems exist for the Z-scheme photocatalysts with shuttle electron mediators in water splitting, such as the complicated synthesis process, strong photocatalytic backward reaction and the spectral shielding effect by the shuttle redox mediators.^200–202

Fig. 12 Schematic illustration of (a) Z-scheme junction band alignments with an appropriate shuttle redox, (b) direct Z-scheme without an electron mediator for overall water splitting. CB, conduction band; VB, valence band; E_g, band gap.

Over the past few decades, the direct Z-scheme has attracted great attention due to its simple synthesis and wide range of light absorption. A schematic illustration of WO₃ based direct Z-scheme junctions is shown in Fig. 12b. Both the CB and the VB of the coupled semiconductors are higher than those of WO₃. The CB position of the coupled semiconductor is higher than the position of 2H⁺/H₂. Meanwhile, a large quantity of defects could be readily accumulated at the solid–solid heterointerface, which possesses some similar properties to those of a conductor such as analogous energy levels and low electric resistance. These features prompt the interface to be apt to serve as a recombination center of e⁻–h⁺ pairs.²⁰³ Therefore, the photoexcited electrons with a lower reduction ability migrated from the CB of WO₃ to recombine with the photogenerated holes with a lower oxidation ability in the VB of the coupled semiconductor. The left photogenerated electrons and holes are able to keep strong redox states, which are promising for the overall water splitting. It is worth noting that compared to the type II heterojunction, the charge carrier migration is more feasible for the direct Z-scheme photocatalyst that is attributed to the electrostatic attraction between the electrons and holes. Therefore, photogenerated electron–hole pairs are more readily separated in the Z-scheme heterojunctions, and hence the recombination will be significantly suppressed.

The electron spin resonance (ESR) spin-trapping technique with DMPO under visible light illumination is an effective method to confirm the reaction mechanism of the electron–hole separation in the Z-scheme system. Generally, the CB potential of WO₃ located at ∼0.71 eV vs. NHE is more positive compared with the reduction potential of active superoxide (˙O²⁻) (O₂/˙O²⁻ = −0.284 eV_NHE). Thus, the photo-generated electrons on the surface of WO₃ are thermodynamically unable to react with O₂ to produce ˙O²⁻. In other words, negligible signals of DMPO–˙O²⁻ can be found for pure WO₃, whereas WO₃ can exhibit a clear DMPO–˙OH signal which is ascribed to the more positive VB potential compared with the oxidation of H₂O/or OH⁻ to the ˙OH radicals (OH⁻/˙OH = +2.4 eV). Xiao et al.¹⁹⁴ introduced g-C₃N₄ to combine with WO₃ to form a Z-scheme nanojunction. They found that the DMPO–˙O²⁻ signal was only observed over WO₃/g-C₃N₄ and g-C₃N₄ and the DMPO–˙OH signal was only observed over WO₃/g-C₃N₄ and WO₃. This result implies that both ˙O²⁻ radicals and ˙OH radicals can generate over g-C₃N₄/WO₃ heterostructures, whereas WO₃ can only produce the ˙O²⁻ radicals and g-C₃N₄ can only generate the ˙OH radicals. Therefore, in the g-C₃N₄/WO₃ Z-scheme system, electrons on the CB of g-C₃N₄ retain a high reduction capability to reduce O₂ to the ˙O²⁻ radicals as well as the holes on the VB of WO₃ still retain sufficient high oxidation capability to oxidize H₂O/or OH⁻ to the ˙OH radicals.

Xie et al.²⁰⁴ constructed a CdS/WO₃ Z-scheme nanojunction with lactate as the electron donor. The optimized CdS/WO₃ Z-scheme nanojunction showed a high rate of H₂ evolution without electric energy, which was four times higher than CdS. WO₃ exhibited no H₂ production performance. The amount of CdS on the surface of WO₃ had a significant effect on the properties of the Z-scheme nanojunction, that is, a high amount of CdS hindered the WO₃ from absorbing light and generated the recombination center of photogenerated charges, which reduced the photocatalytic activity. After introducing Pt as the cocatalyst to improve the charge transport between CdS and WO₃, H₂ evolution rates increased up to 2900 μmol g⁻¹ h⁻¹ under visible irradiation, which was about 7.9 times higher than that of the as-prepared CdS/WO₃ Z-scheme nanojunction under visible light irradiation. Dai et al.²⁰⁵ introduced diethylenetriamine (EDTA) as the structure-directing agent and designed a direct Z-scheme system of WO₃/CdS-EDTA composites to achieve an overall water splitting without using electrical energy. A distinct separation of electrons and holes was observed in the Z-scheme structure, in which the photo-excited electrons in the CB of WO₃ recombined with the photo-excited holes in the VB of CdS. This in turn avoided the enriching of holes on the VB of CdS which effectively reduced the photocorrosion of CdS as well as improved the PEC performance.

Graphitic carbon nitride (g-C₃N₄) is an organic semiconductor, and it has been intensively studied as a novel metal-free photocatalyst owing to its visible-light absorption, narrow band gap, high reduction ability and high chemical stability.^206–208 The CB edge potential of g-C₃N₄ (−1.13 eV)²⁰⁹ is more negative than the potential of 2H⁺/H₂ (−0.41 V vs. NHE at pH 7),²¹⁰ indicating that H₂ can be produced on the surface of g-C₃N₄. In contrast to WO₃, O₂ could not be generated on the surface of g-C₃N₄ because of the more negative VB edge potential (∼1.57 eV)²⁰⁹ than the potential of H₂O/O₂ (0.82 V vs. NHE).²¹⁰ The suitable band gap of the CB and VB between WO₃ and g-C₃N₄ offers appropriate driving forces to separate and transfer photogenerated electron–hole pairs and makes it possible to structure Z-scheme semiconductor nanocomposites. Therefore a composite structure is expected to improve the photocatalytic activity of WO₃ by increasing the number of photogenerated charge carriers. In recent years, constructing the Z-scheme WO₃/g-C₃N₄ composites has been reported for the complete oxidation of organic compounds or overall water splitting.^211–214 In Xiao's work,²¹⁵ Z-scheme WO₃/g-C₃N₄ composites with hollow microspheres were fabricated by in situ hydrolysis and a polymerization process consecutively. The prepared heterojunction contains numerous interfaces and highly exposed oxidation–reduction active sites. Due to the unique Z-scheme architecture, more incident photons are expected to be trapped in the hollow cavities, and thus contribute to more electrons and holes available for photocatalytic reactions. The photoluminescence (PL) decay profiles show that a prolonged lifetime of ∼2.23 ns for WO₃/g-C₃N₄ compared to ∼1.62 ns for bare WO₃ improved the reaction abilities and significantly enhanced the photocatalytic efficiency of WO₃/g-C₃N₄. Cui et al.²¹⁶ prepared direct Z-scheme WO₃/g-C₃N₄ composites via a facile one-step heating procedure, which exhibited an improved photocatalytic activity. Complexing WO₃ with g-C₃N₄ caused the red shift which promoted the separation of photogenerated electron–hole pairs. Lu and co-workers²¹⁷ established a system by combining WO₃ and g-C₃N₄. This system showed an enhanced photocurrent compared to pure WO₃, which was ascribed to the accelerated charge separation in the heterointerface because of the optimal structure of the direct Z-scheme. Zhao and coworkers²⁰⁹ combined C₃N₄ and WO₃ to form a direct Z-scheme system via a facile hydrothermal method. In this Z-scheme system, accompanied by 1 wt% Pt, as much as 30.3 and 14.8 mmol of H₂ and O₂ were generated after 20 h of visible light irradiation (λ > 420 nm), while trace amounts of H₂ and O₂ were detected over pure g-C₃N₄. After introducing rGO as the electron mediator to generate ternary composites, the as-prepared samples were able to release hydrogen and oxygen under visible light irradiation with a relatively high performance. The apparent quantum yield efficiency of the well-prepared ternary composites was ca. 0.9% while the homologous sample without rGO showed a smaller value (i.e., 0.7%) under monochromatic light at 420 nm. Due to its excellent conductivity, rGO promoted the photoexcited electron transfer from WO₃ to the photo-generated holes in C₃N₄, acting as a promising electron mediator. Peng et al.²¹⁸ designed a direct Z-scheme g-C₃N₄/WO₃ photocatalyst via partial intercalation of WO₃ within g-C₃N₄ to enhance the activities of photocatalytic H₂ production. Scanning transmission electron microscopy (STEM) clearly resolved the well-defined interfaces between g-C₃N₄ and WO₃, which prompted the charge transportation and benefited the recombination of electrons in the CB of WO₃ and holes in the VB of g-C₃N₄. The internal electric field was suggested to provide the force to drive the Z-scheme electron transportation at the interfacial junction. This increased the lifetime of excited electrons in g-C₃N₄ and excited holes in WO₃, therefore, enhancing the photocatalytic efficiency for the overall water splitting.

5. Summary and outlook

Photoelectrochemical water splitting to generate H₂ is a promising approach for our future energy sustainability and has attracted remarkable attention in the past decade. Challenges facing today are to develop photoanodes with high efficiency, stability, and low cost in practical applications. WO₃ is one of the promising materials for its visible light response, durable and efficient operation in aqueous solution. Boosting the PEC water splitting performance of WO₃ can be achieved with managing morphology, crystallinity, heterojunction, oxygen vacancy, doping, and co-catalysts. Compared to bulk structures, nanostructures are of great interest in absorbing more photons and consequently improving the photoelectric conversion efficiency. In particular, the 2-D nanostructure with more significant surface/volume ratios is favorable for highly efficient and directional transport of electrons and holes, thus promoting the electron transfer to the back-contacted conductive substrate. Concerning the crystal structure, tremendous research studies have proved that the monoclinic phase of WO₃ is the most stable phase at room temperature and more active for the OER compared with the hexagonal or orthorhombic phases of WO₃. Highly reactive (002) facets promote the consumption of photo-excited holes to split the adsorbed water molecules into hydrogen, hinder the recombination of photogenerated electron–hole pairs and suppress the formation of peroxo species to maintain the stability of WO₃.

To solve the drawbacks of pristine WO₃ with a slow charge transfer at the electrode/electrolyte interface and a rapid electron–hole recombination, coupling with a narrow band gap semiconductor to create a type II heterojunction is an effective strategy to separate photogenerated hole–electron pairs and to enhance the efficiency of PEC water splitting. Adding an optimized amount of oxygen deficiency into the catalyst can accelerate the charge carrier transfer and impede the charge carrier recombination. Doping with hetero-elements can extend the photo-response range and increase the carrier concentration. Decorating with co-catalysts helps in improving the surface kinetics and reducing the oxygen evolution overpotential that suppresses the formation of peroxo species.

The less negative CB of WO₃ compared to the potential of hydrogen generation makes it unsuitable for overall water splitting. A strategy to overcome this inherent limit is to introduce Z-scheme junctions to produce H₂ from water splitting, which not only significantly facilitates the separation of electrons and holes but also maintains the photogenerated electrons and holes in the strong redox states.

Although tremendous efforts have been made to synthesize modified nanostructured WO₃ or their composites, the PEC water splitting efficiency still fails to meet industrial applications. Researchers ignored that WO₃ is only stable in acidic media of pH lower than 4 according to the Pourbaix diagrams, and thus future research studies have to investigate the PEC performance of WO₃ under strong acidic conditions. The theoretical research of WO₃ in the PEC process should be taken into account to deeply understand the underlying physics and chemistry. Otherwise, many on-going challenges in improving the stability of WO₃ and preventing the rapid recombination of electrons and holes should be solved to improve the PEC water splitting performance in the near future.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Beijing Natural Science Foundation (2151001, 2154043), the National Natural Science Foundation (51534009, 52621003, 51402008), the Scientific Research Project of Beijing Municipal Education Commission (KM201610005026) and the New Century National Hundred, Thousand and Ten Thousand Talent Project and Beijing Municipal High-Level Innovative Team Building Program (No. IDHT 20170502). Q. G. was supported as part of the Fluid Interface Reactions, Structures and Transport (FIRST) Center, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences. A portion of this research was conducted at the Center for Nanophase Materials Sciences, which is a DOE Office of Science User Facility.

References

1. (a) D. M. Fabian, S. Hu, N. Singh, F. A. Houle, T. Hisatomi, K. Domen, F. E. Osterloh and S. Ardo, Energy Environ. Sci., 2015, 8, 2825–2850; (b) R. Li, X. Zhu, Q. Fu, G. Liang, Y. Chen, L. Luo, M. Dong, Q. Shao, C. Lin, R. Wei and Z. Guo, Chem. Commun., 2019, 55, 2493–2496; (c) C. Hou, J. Wang, W. Du, J. Wang, Y. Du, C. Liu, J. Zhang, H. Hou, F. Dang, L. Zhao and Z. Guo, J. Mater. Chem. A, 2019, 7, 13460–13472; (d) Y. Liu, M. Shi, C. Yan, Q. Zhuo, H. Wu, L. Wang, H. Liu and Z. Guo, J. Mater. Sci., 2019, 30(7), 6583–6592; (e) L. Yang, M. Shi, J. Jiang, Y. Liu, C. Yan, H. Liu and Z. Guo, Mater. Lett., 2019, 244, 27–30; (f) M. Liu, Z. Yang, H. Sun, C. Lai, X. Zhao, H. Peng and T. Liu, Nano Res., 2016, 9, 3735–3746; (g) M. Liu, B. Li, H. Zhou, C. Chen, Y. Liu and T. Liu, Chem. Commun., 2017, 53, 2810–2813; (h) J. Cai, W. Xu, Y. Liu, Z. Zhu, G. Liu, W. Ding, G. Wang, H. Wang and Y. Luo, Eng. Sci., 2019, 5, 21–29; (i) W. Zhao, X. Li, R. Yin, L. Qian, X. Huang, H. Liu, J. Zhang, J. Wang, T. Ding and Z. Guo, Nanoscale, 2019, 11, 50–59.
2. (a) Z. G. Zou, J. H. Ye, K. Sayama and H. Arakawa, Nature, 2001, 414, 625–627; (b) M. Liu, Q. Meng, Z. Yang, X. Zhao and T. Liu, Chem. Commun., 2018, 54, 5090–5093; (c) M. Liu, Y. Liu, Y. Yan, F. Wang, J. Liu and T. Liu, Chem. Commun., 2017, 53, 9097–9100; (d) Q. Hu, N. Zhou, K. Gong, H. Liu, Q. Liu, D. Sun, Q. Wang, Q. Shao, H. Liu, B. Qiu and Z. Guo, ACS Sustainable Chem. Eng., 2019, 7, 5912–5920; (e) E. Z. Shhi, T. L. Feng, J. H. Bahk, Y. Pan, W. Zheng, Z. Li, G. J. Snyder, S. T. Pantelides and Y. Wu, ES Energy Environ., 2018, 2, 43–49; (f) W. Du, X. Wang, J. Zhan, X. Sun, L. Kang, F. Jiang, X. Zhang, Q. Shao, M. Dong, H. Liu, V. Murugadoss and Z. Guo, Electrochim. Acta, 2019, 296, 907–915; (g) H. Du, C. Zhao, J. Lin, Z. Hu, Q. Shao, J. Guo, B. Wang, D. Pan, E. Wujcik and Z. Guo, Chem. Rec., 2018, 18, 1365–1372; (h) Y. Qian, Y. Yuan, H. Wang, H. Liu, J. Zhang, S. Shi, Z. Guo and N. Wang, J. Mater. Chem. A, 2018, 6, 24676–24685; (i) W. Deng, T. Kang, H. Liu, J. Zhang, N. Wang, N. Lu, Y. Ma, A. Umar and Z. Guo, Sci. Adv. Mater., 2018, 10, 937–949; (j) J. Tian, Q. Shao, X. Dong, J. Zheng, D. Pan, X. Zhang, H. Cao, L. Hao, J. Liu and Z. Guo, Electrochim. Acta, 2018, 261, 236–245.
3. (a) X. H. Cao, C. L. Tan, X. Zhang, W. Zhao and H. Zhang, Adv. Mater., 2016, 28, 6167–6196; (b) J. Ren, Q. Luo, Q. Hou, H. Chen, T. Liu, H. He, J. Wang, Q. Shao, M. Dong, S. Wu, N. Wang, J. Lin and Z. Guo, ChemElectroChem, 2019, 6, 3167–3174; (c) K. Le, Z. Wang, F. Wang, Q. Wang, Q. Shao, V. Murugadoss, S. Wu, W. Liu, J. Liu, Q. Gao and Z. Guo, Dalton Trans., 2019, 48, 5193–5202; (d) Y. Ma, C. Hou, H. Zhang, Q. Zhang, H. Liu, S. Wu and Z. Guo, Electrochim. Acta, 2019, 315, 114–123; (e) M. Idrees, S. Batool, J. Kong, Q. Zhuang, H. Liu, Q. Shao, N. Lu, Y. Feng, E. K. Wujcik, Q. Gao, T. Ding, R. Wei and Z. Guo, Electrochim. Acta, 2019, 296, 925–937; (f) B. Kirubasankar, V. Murugadoss, J. Lin, T. Ding, M. Dong, H. Liu, J. Zhang, T. Li, N. Wang, Z. Guo and S. Angaiaha, Nanoscale, 2018, 10, 20414–20425; (g) H. Wei, H. Wang, Y. Xia, D. Cui, Y. Shi, M. Dong, C. Liu, T. Ding, J. Zhang, Y. Ma, N. Wang, Z. Wang, Y. Sun, R. Wei and Z. Guo, J. Mater. Chem. C, 2018, 6, 12446–12467; (h) K. Sun, R. Fan, X. Zhang, Z. Zhang, Z. Shi, N. Wang, P. Xie, Z. Wang, G. Fan, H. Liu, C. Liu, T. Li, C. Yan and Z. Guo, J. Mater. Chem. C, 2018, 6, 2925–2943; (i) M. Idrees, S. Batool, Q. Zhuang, J. Kong, I. Seok, J. Zhang, H. Liu, V. Murugadoss, Q. Gao and Z. Guo, Ceram. Int., 2019, 45, 10572–10580.
4. (a) J. Zhang, Z. Zhang, Y. Jiao, H. Yang, Y. Li, J. Zhang and P. Gao, J. Power Sources, 2019, 419, 99–105; (b) C. Hou, Z. Tai, L. Zhao, Y. Zhai, Y. Hou, Y. Fan, F. Dang, J. Wang and H. Liu, J. Mater. Chem. A, 2018, 6, 9723–9736; (c) V. Murugadoss, P. Panneerselvam, C. Yan, Z. Guo and S. Angaiah, Electrochim. Acta, 2019, 312, 157–167; (d) C. Shen, X. Liu, H. Cao, Y. Zhou, J. Liu, J. Tang, X. Guo, H. Huang and X. Chen, Appl. Sci., 2019, 9(8), 1708; (e) X. Lou, C. Lin, Q. Luo, J. Zhao, B. Wang, J. Li, Q. Shao, X. Guo, N. Wang and Z. Guo, ChemElectroChem, 2017, 4, 3171–3180; (f) Q. Hou, J. Ren, H. Chen, P. Yang, Q. Shao, M. Zhao, X. Zhao, H. He, N. Wang, Q. Luo and Z. Guo, ChemElectroChem, 2018, 5, 726–731; (g) C. Lin, H. Hu, C. Cheng, K. Sun, X. Guo, Q. Shao, J. Li, N. Wang and Z. Guo, Electrochim. Acta, 2018, 260, 65–72; (h) L. Yang, X. Wang, X. Mai, T. Wang, C. Wang, X. Li, V. Murugadoss, Q. Shao, S. Angaiah and Z. Guo, J. Colloid Interface Sci., 2019, 534, 459–468; (i) C. Wang, B. Mo, Z. He, Q. Shao, D. Pan, E. Wujick, J. Guo, X. Xie, X. Xie and Z. Guo, J. Membr. Sci., 2018, 556, 118–125; (j) C. Wang, B. Mo, Z. He, C. X. Zhao, L. Zhang, Q. Shao, X. Guo, E. Wujcik and Z. Guo, Polymer, 2018, 138, 363–368; (k) Z. Qu, M. Shi, H. Wu, Y. Liu, J. Jiang and C. Yan, J. Power Sources, 2019, 410–411, 179–187.
5. (a) M. C. Acevedo, M. L. Stone, J. R. Schmidt, J. G. Thomas, Q. Ding, H. C. Chang, M. L. Tsai, J. H. He and S. Jin, Nat. Mater., 2015, 14, 1245–1251; (b) B. Lin, Z. Lin, S. Chen, M. Yu, W. Li, Q. Gao, M. Dong, Q. Shao, S. Wu, T. Ding and Z. Guo, Dalton Trans., 2019, 48, 8279–8287; (c) J. Brillet, J. H. Yum, M. Cornuz, T. Hisatomi, R. Solarska, J. Augustynsk, M. Graetzel and K. Sivula, Nat. Photonics, 2012, 6, 824–828.
6. (a) L. Wei, K. Lozano and Y. Mao, Eng. Sci., 2018, 3, 62–66; (b) Z. Lin, B. Lin, Z. Wang, S. Chen, C. Wang, M. Dong, Q. Gao, Q. Shao, S. Wu, T. Ding, H. Liu and Z. Guo, ChemCatChem, 2019, 11, 2217–2222; (c) C. Wang, F. Lan, Z. He, X. Xie, Y. Zhao, H. Hou, L. Guo, V. Murugadoss, H. Liu, Q. Shao and Q. Gao, ChemSusChem, 2019, 12, 1576–1590; (d) J. D. Desai, P. K. Baviskar, K. N. Hui and H. M. Pathan, ES Energy Environ., 2018, 2, 21–30; (e) M. Gratzel, Nature, 2001, 414, 338–344; (f) T. Su, Q. Shao, Z. Qin, Z. Guo and Z. Wu, ACS Catal., 2018, 8, 2253–2276.
7. E. Becquerel, Compt. Rend., 1839, 9, 561–567.
8. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
9. S. C. Riha, B. M. Klahr, E. C. Tyo, S. Seifert, S. Vajda, M. J. Pellin, T. W. Hamann and A. B. F. Martinson, ACS Nano, 2013, 3, 2396–2405.
10. J. R. Wang and F. E. Osterloh, J. Mater. Chem. A, 2014, 2, 9405–9411.
11. J. Song, J. Cha, M. G. Lee, H. W. Jeong, S. Seo, J. A. Yoo, T. L. Kim, J. Lee, H. No, D. H. Kim, S. Y. Jeong, H. An, B. H. Lee, C. W. Bark, H. Park, H. W. Jang and S. Lee, J. Mater. Chem. A, 2017, 5, 18831–18838.
12. T. W. Kim and K. S. Choi, Science, 2014, 343, 990–994.
13. G. L. Chiarello, M. Bernareggi, M. Pedroni, M. Magni, S. M. Pietralunga, A. Tagliaferri, E. Vassallo and E. Selli, J. Mater. Chem. A, 2017, 5, 12977–12989.
14. X. J. Shi, L. L. Cai, I. Y. Choi, M. Ma, K. Zhang, J. H. Zhao, J. K. Kim, J. K. Kim, X. L. Zheng and J. H. Park, J. Mater. Chem. A, 2018, 6, 19542–19546.
15. Y. Y. Bu, J. Ren, H. W. Zhang, D. J. Yang, Z. Y. Chen and J. P. Ao, J. Mater. Chem. A, 2018, 6, 8604–8611.
16. A. J. Nozik, Nature, 1975, 257, 383–386.
17. B. H. Zhou, X. F. Li, T. X. Fan, F. E. Osterloh, J. Ding, E. M. Sabio, D. Zhang and Q. X. Guo, Adv. Mater., 2010, 22, 951–956.
18. C. M. Pelicano and H. Yanagi, J. Mater. Chem. C, 2017, 5, 8059–8070.
19. H. C. Yang, J. J. Li, L. Yu, B. B. Huang, Y. D. Ma and Y. Dai, J. Mater. Chem. A, 2018, 6, 4161–4166.
20. L. Pei, Z. Xu, S. C. Yan and Z. G. Zou, J. Mater. Chem. A, 2017, 5, 12848–12855.
21. F. A. Frame, T. K. Townsend, R. L. Chamousis, E. M. Sabio, T. Dittrich, N. D. Browning and F. E. Osterloh, J. Am. Chem. Soc., 2011, 133, 7264–7267.
22. T. K. Townsend, N. D. Browning and F. E. Osterloh, ACS Nano, 2012, 6, 7420–7426.
23. M. Moniruddin, K. Afroz, Y. Shabdan, B. Bizri and N. Nuraje, Appl. Surf. Sci., 2017, 419, 886–892.
24. M. Zhong, T. Hisatomi, Y. Sasaki, S. Suzuki, K. Teshima, M. Nakabayashi, N. Shibata, H. Nishiyama, M. Katayama, T. Yamada and K. Domen, Angew. Chem., Int. Ed., 2017, 56, 4739–4743.
25. J. Song, T. L. Kim, J. Lee, S. Y. Cho, J. Cha, S. Y. Jeong, H. An, W. S. Kim, Y. S. Jung, J. Park, G. Y. Jung, D. Y. Kim, J. Y. Jo, S. D. Bu, H. W. Jang and S. H. Lee, Nano Res., 2018, 11, 642–655.
26. F. A. Frame and F. E. Osterloh, J. Phys. Chem. C, 2010, 114, 10628–10633.
27. O. Khaselev and J. A. Turner, Science, 1998, 280, 425–427.
28. B. Song, T. T. Wang, H. G. Sun, Q. Shao, J. K. Zhao, K. K. Song, L. H. Hao, L. Wang and Z. H. Guo, Dalton Trans., 2017, 46, 15769–15777.
29. X. Y. Feng, Y. B. Chen, Z. X. Qin, M. L. Wang and L. J. Guo, ACS Appl. Mater. Interfaces, 2016, 8, 18089–18096.
30. D. S. Martínez, A. M. L. Cruz and E. L. C. llar, Appl. Catal., A, 2011, 398, 179–186.
31. Y. Peng, Q. G. Chen, D. Wang, H. Y. Zhou and A. W. Xu, CrystEngComm, 2015, 17, 569–576.
32. S. C. Wang, H. J. Chen, G. P. Gao, T. Butburee, M. Q. Lyu, S. Thaweesak, J. H. Yun, A. Du, G. Liu and L. Z. Wang, Nano Energy, 2016, 24, 94–102.
33. G. M. Wang, Y. C. Ling, H. Y. Wang, X. Y. Yang, C. C. Wang, J. Z. Zhang and Y. Li, Energy Environ. Sci., 2012, 5, 6180–6187.
34. F. E. Osterloh, Chem. Soc. Rev., 2013, 42, 2294–2320.
35. X. B. Huang, G. X. Zhao and G. Wang, J. Mater. Chem. A, 2017, 5, 24631–24635.
36. P. M. Rao, I. S. Cho and X. L. Zheng, Proc. Combust. Inst., 2013, 34, 2187–2195.
37. S. C. Warren, K. Voitchovsky, H. Dotan, C. M. Leroy, M. Cornuz, F. Stellacci, C. Hebert, A. Rothschild and M. Gratzel, Nat. Mater., 2013, 12, 842–849.
38. W. J. Li, P. M. Da, Y. Y. Zhang, Y. C. Wang, X. Lin, X. G. Gong and G. F. Zheng, ACS Nano, 2014, 8, 11770–11777.
39. J. Zhang, Z. H. Liu and Z. F. Liu, ACS Appl. Mater. Interfaces, 2016, 8, 9684–9691.
40. J. A. Seabold and K. S. Choi, Chem. Mater., 2011, 23, 1105–1112.
41. Q. Y. Zeng, J. H. Li, J. Bai, X. J. Li, L. G. Xia and B. X. Zhou, Appl. Catal., B, 2017, 202, 388–396.
42. S. C. Wang, J. H. Yun, B. Luo, T. Butburee, P. Peerakiatkhajohn, S. Thaweesak, M. Xiao and L. Z. Wang, J. Mater. Sci. Technol., 2017, 33, 1–22.
43. M. D. Bhatt and J. S. Lee, J. Mater. Chem. A, 2015, 3, 10632–10659.
44. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520–7535.
45. M. Moniruddin, B. Ilyassov, X. Zhao, E. Smith, T. Serikov, N. Ibrayev, R. Asmatulu and N. Nuraje, Mater. Today Energy, 2018, 7, 246–259.
46. K. Maeda, ACS Catal., 2013, 3, 1486–1503.
47. Z. Chen, T. F. Jaramillo, T. G. Deutsch, A. Kleiman-Shwarsctein, A. J. Forman, N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, E. W. McFarland, K. Domen, E. L. Miller, J. A. Turner and H. N. Dinh, J. Mater. Res., 2010, 25, 4–16.
48. L. M. Peter and K. G. U. Wijayantha, ChemPhysChem, 2014, 15, 1983–1995.
49. M. T. Nenadovic, T. Rajh, O. I. Micic and A. J. Nozik, J. Phys. Chem., 1984, 88, 5827–5830.
50. C. A. Bignozzi, S. Caramor, V. Cristino, R. Argazzi, L. Meda and A. Tacca, Chem. Soc. Rev., 2013, 42, 2228–2246.
51. F. L. Formal, S. R. Pendlebury, M. Cornuz, S. D. Tilley, M. Gratzel and J. R. Durrant, J. Am. Chem. Soc., 2014, 136, 2564–2574.
52. J. Y. Gan, X. H. Lu and Y. X. Tong, Nanoscale, 2014, 6, 7142–7164.
53. K. Afroz, M. Moniruddin, N. Bakranov, S. Kudaibergenov and N. Nuraje, J. Mater. Chem. A, 2018, 6, 21696–21718.
54. E. M. Sabio, R. L. Chamousis, N. D. Browning and F. E. Osterloh, J. Phys. Chem. C, 2012, 116, 3161–3170.
55. M. R. Waller, T. K. Townsend, J. Zhao, E. M. Sabio, R. L. Chamousis, N. D. Browning and F. E. Osterloh, Chem. Mater., 2012, 24, 698–704.
56. J. Z. Su, X. J. Feng, J. D. Sloppy, L. J. Guo and C. A. Grimes, Nano Lett., 2011, 11, 203–208.
57. S. S. Kalanur, Y. J. Hwang, S. Y. Chae and O. S. Joo, J. Mater. Chem. A, 2013, 1, 3479–3488.
58. H. Qi, J. Wolfe, D. P. Wang, H. J. Fan, D. Fichou and Z. Chen, Nanoscale, 2014, 6, 13457–13462.
59. M. G. Lee, D. H. Kim, W. Sohn, C. W. Moon, H. Park, S. Lee and H. W. Jang, Nano Energy, 2016, 28, 250–260.
60. S. Shin, H. S. Han, J. S. Kim, I. J. Park, M. H. Lee, K. S. Hong and I. S. Cho, J. Mater. Chem. A, 2015, 3, 12920–12926.
61. C. Santato, M. Ulmann and J. Augustynski, Adv. Mater., 2001, 13, 511–514.
62. V. Chakrapani, J. Thangala and M. K. Sunkara, Int. J. Hydrogen Energy, 2009, 34, 9050–9059.
63. I. S. Cho, H. S. Han, M. Logar, J. Park and X. L. Zheng, Adv. Energy Mater., 2016, 6, 1501840–1501848.
64. J. Yang, W. Z. Li, J. Li, D. B. Sun and Q. Y. Chen, J. Mater. Chem., 2012, 22, 17744–17752.
65. F. Amano, D. Li and B. Ohtani, Chem. Commun., 2010, 46, 2769–2771.
66. D. D. Qin, C. L. Tao, S. In, Z. Y. Yang, T. E. Mallouk, N. Z. Bao and C. A. Grimes, Energy Fuels, 2011, 25, 5257–5263.
67. J. J. Zhang, X. X. Chang, C. C. Li, A. Li, S. S. Liu, T. Wang and J. L. Gong, J. Mater. Chem. A, 2018, 6, 3350–3354.
68. S. J. Hong, H. Jun, P. H. Borse and J. S. Lee, Int. J. Hydrogen Energy, 2009, 34, 3234–3242.
69. W. Z. Li, J. Li, X. Wang, J. Ma and Q. Y. Chen, Int. J. Hydrogen Energy, 2010, 35, 13137–13145.
70. C. Fàbrega, S. Murcia-López, D. Monllor-Satoca, J. D. Prades, M. D. Hernández-Alonso, G. Penelas, J. R. Morante and T. Andreu, Appl. Catal., B, 2016, 189, 133–140.
71. C. W. Wang, C. F. Tang, X. Y. Zhang, L. Qian and H. G. Yang, Prog. Nat. Sci.: Mater. Int., 2018, 28, 200–204.
72. J. H. Kim, B. J. Lee, P. Wang, M. H. Son and J. S. Lee, Appl. Catal., A, 2016, 521, 233–239.
73. M. S. Gudiksen, L. J. Lauhon, J. F. Wang, D. C. Smith and C. M. Lieber, Nature, 2002, 415, 617–620.
74. J. R. G. Navarro, A. Mayence, J. Andrade, F. Lerouge, F. Chaput, P. Oleynikov, L. Bergström, S. Parola and A. Pawlicka, Langmuir, 2014, 30, 10487–10492.
75. Q. F. Zhang and G. Z. Cao, Nano Today, 2011, 6, 91–109.
76. B. Liu and E. S. Aydil, J. Am. Chem. Soc., 2009, 131, 3985–3990.
77. X. L. Fan, B. Gao, T. Wang, X. L. Huang, H. Gong, H. R. Xue, H. Guo, L. Song, W. Xia and J. P. He, Appl. Catal., A, 2016, 528, 52–58.
78. B. X. Liu, J. S. Wang, J. S. Wu, H. Y. Li, Z. F. Li, M. L. Zhou and T. Y. Zuo, J. Mater. Chem. A, 2014, 2, 1947–1954.
79. K. Zhu, N. R. Neale, A. Miedaner and A. J. Frank, Nano Lett., 2007, 7, 69–74.
80. A. Kafizas, L. Francàs, C. S. Vazquez, M. Ling, Y. M. Li, E. Glover, L. M. Cafferty, C. Blackman, J. Darr and I. Parkin, J. Phys. Chem. C, 2017, 121, 5983–5993.
81. I. S. Cho, Z. B. Chen, A. J. Forman, D. R. Kim, P. M. Rao, T. F. Jaramillo and X. L. Zheng, Nano Lett., 2011, 11, 4978–4984.
82. M. Park, J. H. Seo, H. Song and K. M. Nam, J. Phys. Chem. C, 2016, 120, 9192–9199.
83. H. Wang, X. Quan, Y. B. Zhang and S. Chen, Nanotechnology, 2008, 19, 065704.
84. Z. Chen, M. H. Ning, G. Ma, Q. G. Meng, Y. G. Zhang, J. W. Gao, M. L. Jin, Z. H. Chen, M. Z. Yuan, X. Wang, J. M. Liu and G. F. Zhou, Nanotechnology, 2017, 28, 275401.
85. F. Q. Zhan, J. Li, W. Z. Li, Y. S. Liu, R. R. Xie, Y. H. Yang, Y. M. Li and Q. Y. Chen, Int. J. Hydrogen Energy, 2015, 40, 6512–6520.
86. M. Yagi, S. Maruyama, K. Sone, K. Nagai and T. Norimatsu, J. Solid State Chem., 2008, 181, 175–182.
87. F. Amano, M. Tian, G. S. Wu, B. Ohtani and A. Chen, ACS Appl. Mater. Interfaces, 2011, 3, 4047–4052.
88. Z. H. Jiao, J. M. Wang, L. Ke, X. W. Sun and H. V. Demir, ACS Appl. Mater. Interfaces, 2011, 3, 229–236.
89. S. Sohila, M. Geerthana, S. Prabhu, T. Pratheesya, A. Mohamed Musthafa, S. Tamilselvan and R. Ramesh, J. Mater. Sci.: Mater. Electron., 2017, 28, 15663–15667.
90. F. Amano, D. Li and B. Ohtani, J. Electrochem. Soc., 2011, 158, K42–K46.
91. Y. G. Li, Y. L. Li, B. S. Sa and R. Ahuja, Catal. Sci. Technol., 2017, 7, 545–559.
92. J. J. Zhang, P. Zhang, T. Wang and J. L. Gong, Nano Energy, 2015, 11, 189–195.
93. T. W. Kim and K. S. Choi, Science, 2014, 343, 990–994.
94. X. Zhao, W. J. Luo, J. Y. Feng, M. X. Li, Z. S. Li, T. Yu and Z. G. Zou, Adv. Energy Mater., 2014, 4, 1301785.
95. D. Chandra, K. Saito, T. Yui and M. Yagi, Angew. Chem., Int. Ed., 2013, 52, 12606–12609.
96. Z. F. Li, G. W. Zheng, J. S. Wang, H. Y. Li, J. S. Wu and Y. C. Du, J. Nanopart. Res., 2016, 18, 98.
97. C. X. Hou, Y. Hou, Y. Fan, Y. Zhai, Y. Wang, Z. Sun, R. Fan, F. Dang and J. Wang, J. Mater. Chem. A, 2018, 6, 6967–6876.
98. K. S. Choi, J. Phys. Chem. Lett., 2010, 1, 2244–2250.
99. K. Song, F. M. Gao, W. Y. Yang, E. Wang, Z. X. Wang and H. L. Hou, ChemElectroChem, 2018, 5, 322–327.
100. S. Hilliard, G. Baldinozzi, D. Friedrich, S. Kressman, H. Strub, V. Artero and C. L. Robert, Sustainable Energy Fuels, 2017, 1, 145–153.
101. M. Y. Cai, P. X. Fan, J. Y. Long, J. P. Han, Y. Lin, H. J. Zhang and M. L. Zhong, ACS Appl. Mater. Interfaces, 2017, 9, 17856–17864.
102. C. Santato, M. Odziemkowski, M. Ulmann and J. Augustynski, J. Am. Chem. Soc., 2001, 123, 10639–10649.
103. S. Y. Chae, C. S. Lee, H. Jung, O. S. Joo, B. K. Min, J. H. Kim and Y. J. Hwang, ACS Appl. Mater. Interfaces, 2017, 9, 19780–19790.
104. J. Z. Su, T. Zhang and L. Wang, J. Mater. Sci.: Mater. Electron., 2017, 28, 4481–4491.
105. A. K. Nayak, Y. Sohn and D. Pradhan, Cryst. Growth Des., 2017, 17, 4949–4957.
106. H. G. Yang, C. H. Sun, S. Z. Qiao, J. Zou, G. Liu, S. C. Smith, H. M. Cheng and G. Q. Lu, Nature, 2008, 453, 638–641.
107. X. W. Wang, G. Liu, L. Z. Wang, J. Pan, G. Q. M. Lu and H. M. Cheng, J. Mater. Chem., 2011, 21, 869–873.
108. B. H. Wu, C. Y. Guo, N. F. Zheng, Z. X. Xie and G. D. Stucky, J. Am. Chem. Soc., 2008, 130, 17563–17567.
109. D. Q. Zhang, S. L. Wang, J. Zhu, H. X. Li and Y. F. Lu, Appl. Catal., B, 2012, 123–124, 398–404.
110. P. Y. Dong, G. H. Hou, X. G. Xi, R. Shao and F. Dong, Environ. Sci.: Nano, 2017, 4, 539–557.
111. J. Y. Zheng, G. Song, J. S. Hong, T. K. Van, A. U. Pawar, D. Y. Kim, C. W. Kim, Z. Haider and Y. S. Kang, Cryst. Growth Des., 2014, 14, 6057–6066.
112. R. Liu, Y. J. Lin, L. Y. Chou, S. W. Sheehan, W. S. He, F. Zhang, H. J. M. Hou and D. W. Wang, Angew. Chem., 2011, 123, 519–522.
113. (a) W. Smith, A. Wolcott, R. C. Fitzmorris, J. Z. Zhang and Y. P. Zhao, J. Mater. Chem., 2011, 21, 10792–10800; (b) H. Sun, Z. Yang, Y. Pu, W. Dou, C. Wang, W. Wang, X. Hao, S. Chen, Q. Shao, M. Dong, S. Wu, T. Ding and Z. Guo, J. Colloid Interface Sci., 2019, 547, 40–49; (c) Y. Sheng, J. Yang, F. Wang, L. Liu, H. Liu, C. Yan and Z. Guo, Appl. Surf. Sci., 2019, 465, 154–163; (d) Z. Zhao, H. An, J. Lin, M. Feng, V. Murugadoss, T. Ding, H. Liu, Q. Shao, X. Man, N. Wang, H. Gu, S. Angaiah and Z. Guo, Chem. Rec., 2019, 19, 873–882; (e) J. Tian, Q. Shao, J. Zhao, D. Pan, M. Dong, C. Jia, T. Ding, T. Wu and Z. Guo, J. Colloid Interface Sci., 2019, 541, 18–29; (f) L. Zhang, W. Yu, C. Han, J. Guo, Q. Zhang, H. Xie, Q. Shao, Z. Sun and Z. Guo, J. Electrochem. Soc., 2017, 164, H651–H656; (g) L. Zhang, M. Qin, W. Yu, Q. Zhang, H. Xie, Z. Sun, Q. Shao, X. Guo, L. Hao, Y. Zheng and Z. Guo, J. Electrochem. Soc., 2017, 164, H1086–H1090.
114. (a) J. Johnson, N. Bakranov, M. Moniruddin, R. Iskakov, S. Kudaibergenov and N. Nuraje, New J. Chem., 2017, 41, 15528–15532; (b) B. Song, T. Wang, H. Sun, Q. Shao, J. Zhao, K. Song, L. Hao, L. Wang and Z. Guo, Dalton Trans., 2017, 46, 15769–15777.
115. Z. F. Hu, M. K. Xu, Z. R. Shen and J. C. Yu, J. Mater. Chem. A, 2015, 3, 14046–14053.
116. Y. Ma, Y. L. Jia, L. Wang, M. Yang, Y. P. Bi and Y. X. Qia, Appl. Surf. Sci., 2016, 390, 399–405.
117. Y. G. Li, J. Feng, H. J. Li, X. L. Wei, R. R. Wang and A. N. Zhou, Int. J. Hydrogen Energy, 2016, 41, 4096–4105.
118. K. H. Ng, L. J. Minggua, W. F. Mark-Lee, K. Arifin, M. H. H. Jumali and M. B. Kassim, Mater. Res. Bull., 2018, 98, 47–52.
119. D. P. Wang, P. S. Bassi, H. Qi, X. Zhao, Gurudayal, L. H. Wong, R. Xu, T. Sritharan and Z. Chen, Materials, 2016, 9, 348.
120. J. Zhu, W. Z. Li, J. Li, Y. M. Li, H. S. Hua and Y. H. Yang, Electrochim. Acta, 2013, 112, 191–198.
121. C. J. Liu, Y. H. Yang, W. Z. Li, J. Li, Y. M. Li and Q. Y. Chen, Sci. Rep., 2016, 6, 23451.
122. A. A. M. Ibrahim, I. Khan, N. Iqbal and A. Qurashi, Int. J. Hydrogen Energy, 2017, 42, 3423–3430.
123. Y. G. Zhou, L. Y. Zhang, L. H. Lin, B. R. Wygant, Y. Liu, Y. Zhu, Y. B. Zheng, C. B. Mullins, Y. Zhao, X. H. Zhang and G. H. Yu, Nano Lett., 2017, 17, 8012–8017.
124. P. M. Rao, L. L. Cai, C. Liu, I. S. Cho, C. H. Lee, J. M. Weisse, P. D. Yang and X. L. Zheng, Nano Lett., 2014, 14, 1099–1105.
125. S. Y. Chae, H. J. Jung, H. S. Jeon, B. K. Min, Y. J. Hwang and O. S. Joo, J. Mater. Chem. A, 2014, 2, 11408–11416.
126. C. W. Lai and S. Sreekantan, Int. J. Hydrogen Energy, 2013, 38, 2156–2166.
127. K. C. Leonard, K. M. Nam, H. C. Lee, S. H. Kang, H. S. Park and A. J. Bard, J. Phys. Chem. C, 2013, 117, 15901–15910.
128. P. D. Wu, Z. F. Liu, D. Chen, M. Zhou and J. D. Wei, Appl. Surf. Sci., 2018, 440, 1101–1106.
129. H. S. Lin and L. Y. Lin, Electrochim. Acta, 2017, 252, 235–244.
130. J. Zhang, H. P. Ma and Z. F. Liu, Appl. Catal., B, 2017, 201, 84–91.
131. Y. D. Wang, W. Tian, L. Chen, F. R. Cao, J. Guo and L. Li, ACS Appl. Mater. Interfaces, 2017, 46, 40235–40243.
132. C. J. Liu, Y. H. Yang, J. Li, S. Chen, W. Z. Li and X. D. Tang, Chem. Eng. J., 2017, 326, 603–611.
133. W. L. Kwong, C. C. Lee and J. Messinger, J. Phys. Chem. C, 2016, 120, 10941–10950.
134. S. S. Kalanur, I. H. Yoo, J. Park and H. Seo, J. Mater. Chem. A, 2017, 5, 1455–1461.
135. J. Z. Su, L. J. Guo, N. Z. Bao and C. A. Grimes, Nano Lett., 2011, 11, 1928–1933.
136. J. W. Huang, Y. Ding, X. Luo and Y. Y. Feng, J. Catal., 2016, 333, 200–206.
137. X. P. Qi, G. W. She, Y. Y. Liu, L. X. Mua and W. S. Shi, Chem. Commun., 2012, 48, 242–244.
138. C. J. Liu, Y. M. Li, W. Z. Li, J. Zhu, J. Li, Q. Y. Chen and Y. H. Yang, Mater. Lett., 2014, 120, 170–173.
139. Y. W. Choi, S. Kim, M. Seong, H. Yoo and J. Choi, Appl. Surf. Sci., 2015, 324, 414–418.
140. S. C. Wang, P. Chen, Y. Bai, J. H. Yun, G. Liu and L. Wang, Adv. Mater., 2018, 20, 1800486.
141. S. C. Wang, P. Chen, J. H. Yun, Y. X. Hu and L. Z. Wang, Angew. Chem., Int. Ed., 2017, 56, 8500–8504.
142. B. J. Jin, E. Jung, M. Ma, S. Kim, K. Zhang, J. I. Kim, Y. Son and J. H. Park, J. Mater. Chem. A, 2018, 6, 2585–2592.
143. M. G. Mali, H. Yoon, M. Kim, M. T. Swihart, S. S. Al-Deyab and S. S. Yoon, Appl. Phys. Lett., 2015, 106, 151603.
144. J. H. Baek, B. J. Kim, G. S. Han, S. W. Hwang, D. R. Kim, I. S. Cho and H. S. Jung, ACS Appl. Mater. Interfaces, 2017, 9, 1479–1486.
145. K. Zhang, X. J. Shi, J. K. Kima and J. H. Park, Phys. Chem. Chem. Phys., 2012, 14, 11119–11124.
146. Q. Y. Zeng, J. H. Li, L. S. Li, J. Bai, L. G. Xia and B. X. Zhou, Appl. Catal., B, 2017, 217, 21–29.
147. P. Chatchai, Y. Murakami, S. Y. Kishioka, A. Y. Nosaka and Y. Nosaka, Electrochim. Acta, 2009, 54, 1147–1152.
148. R. Saito, Y. Miseki and K. Sayama, Chem. Commun., 2012, 48, 3833–3835.
149. J. H. Seo, G. Park, K. H. Oh, S. H. Kang, H. C. Lee, S. K. Cho and K. M. Nam, J. Electroanal. Chem., 2017, 789, 17–23.
150. H. F. Zhang, W. W. Zhou, Y. P. Yang and C. W. Cheng, Small, 2017, 13, 1603840.
151. N. Iqbal, I. Khan, Z. H. A. Yamani and A. Qurashi, Sol. Energy, 2017, 144, 604–611.
152. S. J. A. Moniz, S. A. Shevlin, D. J. Martin, Z. X. Guo and J. W. Tang, Energy Environ. Sci., 2015, 8, 731–759.
153. J. H. Yang, D. G. Wang, H. X. Han and C. Li, Acc. Chem. Res., 2013, 46, 1900–1909.
154. J. R. Ran, J. Zhang, J. G. Yu, M. Jaroniec and S. Z. Qiao, Chem. Soc. Rev., 2014, 43, 7787–7812.
155. G. C. Xi, J. H. Ye, Q. Ma, N. Su, H. Bai and C. Wang, J. Am. Chem. Soc., 2012, 134, 6508–6511.
156. Z. F. Liu, J. Y. Wu and J. Zhang, Int. J. Hydrogen Energy, 2016, 41, 20529–20535.
157. D. Y. Hu, P. Diao, D. Xu and Q. Y. Wu, Nano Res., 2016, 6, 1735–1751.
158. F. Xu, Y. W. Yao, D. D. Bai, R. S. Xu, J. J. Mei, D. P. Wu, Z. Y. Gao and K. Jiang, J. Colloid Interface Sci., 2015, 458, 194–199.
159. W. Shi, X. Zhang, J. Brillet, D. Huang, M. Li, M. Wang and Y. Shen, Carbon, 2016, 105, 387–393.
160. J. Zhang, J. G. Yu, Y. M. Zhang, Q. Li and J. R. Gong, Nano Lett., 2011, 11, 4774–4779.
161. F. A. Frame and F. E. Osterloh, J. Phys. Chem. C, 2010, 114, 10628–10633.
162. L. Zhang, Q. H. Zhang, H. Y. Xie, J. Guo, H. L. Ly, Y. G. Li, Z. G. Sun, H. Z. Wang and Z. H. Guo, Appl. Catal., B, 2017, 201, 470–478.
163. H. Y. Zhang, W. J. Tian, Y. G. Li, H. Q. Sun, M. O. Tadé and S. B. Wang, J. Mater. Chem. A, 2018, 6, 6265–6272.
164. J. Sato, N. Saito, Y. Yamada, K. Maeda, T. Takata, J. N. Kondo, M. Hara, H. Kobayashi, K. Domen and Y. Inoue, J. Am. Chem. Soc., 2005, 127, 4150–4151.
165. B. J. Ma, J. H. Yang, H. X. Han, J. T. Wang, X. H. Zhang and C. Li, J. Phys. Chem. C, 2010, 114, 12818–12822.
166. R. L. Lee, P. D. Tran, S. S. Pramana, S. Y. Chiam, Y. Ren, S. Y. Meng, L. H. Wong and J. Barber, Catal. Sci. Technol., 2013, 3, 1694–1698.
167. Y. B. Wang, Y. S. Wang, R. R. Jiang and R. Xu, Ind. Eng. Chem. Res., 2012, 51, 9945–9951.
168. J. S. Zhang, M. Grzelczak, Y. D. Hou, K. Maeda, K. Domen, X. Z. Fu, M. Antonietti and X. C. Wang, Chem. Sci., 2012, 3, 443–446.
169. F. X. Zhang, A. Yamakata, K. Maeda, Y. Moriya, T. Takata, J. Kubota, K. Teshima, S. Oishi and K. Domen, J. Am. Chem. Soc., 2012, 134, 8348–8351.
170. L. C. Liu, Z. Y. Ji, W. X. Zou, X. R. Gu, Y. Deng, F. Gao, C. J. Tang and L. Dong, ACS Catal., 2013, 3, 2052–2061.
171. M. W. Kanan and D. G. Nocera, Science, 2008, 321, 1072–1075.
172. Q. Liu, Q. P. Chen, J. Bai, J. Y. Li, J. H. Li and B. X. Zhou, J. Solid State Electrochem., 2014, 18, 157–161.
173. M. Ma, K. Zhang, P. Li, M. S. Jung, M. J. Jeong and J. H. Parkm, Angew. Chem., Int. Ed., 2016, 55, 11819–11823.
174. L. Pei, Z. Xu, Z. Shi, H. Zhu, S. C. Yan and Z. G. Zou, J. Mater. Chem. A, 2017, 5, 20439–20447.
175. Y. Liu, Y. H. Yang, Q. Liu, H. Z. He, W. H. Liu, D. D. Meng, Y. M. Li, W. Z. Li and J. Li, Int. J. Hydrogen Energy, 2018, 43, 208–218.
176. Y. Liu, J. Li, H. Tang, W. Z. Li, Y. H. Yang, Y. M. Li and Q. Y. Chen, Electrochem. Commun., 2016, 68, 81–85.
177. G. M. Wang, Y. C. Ling and Y. Li, Nanoscale, 2012, 4, 6682–6691.
178. Y. Liu, J. Li, W. Z. Li, H. Z. He, Y. H. Yang, Y. M. Li and Q. Y. Chen, Electrochim. Acta, 2016, 210, 251–260.
179. B. Cole, B. Marsen, E. Miller, Y. Yan, B. To, K. Jones and M. A. Jassim, J. Phys. Chem. C, 2008, 112, 5213–5220.
180. F. Wang, C. D. Valentin and G. Pacchioni, J. Phys. Chem. C, 2012, 116, 8901–8909.
181. 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.
182. C. Feng, S. Wang and B. Geng, Nanoscale, 2011, 3, 3695.
183. S. S. Kalanur, I. H. Yoo, K. Y. Eom and H. Seo, J. Catal., 2018, 357, 127–137.
184. W. Z. Li, F. Q. Zhan, J. Li, C. J. Liu, Y. H. Yang, Y. M. Li and Q. Y. Chen, Electrochim. Acta, 2015, 160, 57–63.
185. Y. H. Xia, C. Q. Xu and W. D. Zhang, J. Solid State Electrochem., 2017, 21, 3355–3364.
186. V. Hariharan, V. Aroulmoji, K. Prabakaran, B. Gnanavel, M. Parthibavarman, R. Sathyapriya and M. Kanagaraj, J. Alloys Compd., 2016, 689, 41–47.
187. S. S. Kalanur and H. Seo, J. Colloid Interface Sci., 2018, 509, 440–447.
188. S. Sakthivel and H. Kisch, Angew. Chem., Int. Ed., 2003, 42, 4908–4911.
189. D. Paluselli, B. Marsen, E. L. Miller and R. E. Rocheleau, Electrochem. Solid-State Lett., 2005, 8, G301–G303.
190. W. Z. Li, J. Li, X. Wang and Q. Y. Chen, Appl. Surf. Sci., 2012, 263, 157–162.
191. X. D. Ma, W. X. Ma, D. L. Jiang, D. Li, S. Meng and M. Chen, J. Colloid Interface Sci., 2017, 506, 93–101.
192. H. L. Guo, H. Du, Y. F. Jiang, N. Jiang, C. C. Shen, X. Zhou, Y. N. Liu and A. W. Xu, J. Phys. Chem. C, 2017, 121, 107–114.
193. A. Meng, B. Zhu, B. Zhong, L. Y. Zhang and B. Cheng, Appl. Surf. Sci., 2017, 422, 518–527.
194. Z. Y. Liang, J. X. Wei, X. Wang, Y. Yu and F. X. Xiao, J. Mater. Chem. A, 2017, 5, 15601–15612.
195. A. B. Yousaf, M. Imran, S. J. Zaidi and P. Kasak, Sci. Rep., 2017, 7, 6574.
196. P. J. Yang, J. H. Zhao, J. Wang, B. Y. Cao, L. Li and Z. P. Zhu, J. Mater. Chem. A, 2015, 3, 8256–8259.
197. Y. J. Yuan, D. Q. Chen, S. H. Yang, L. X. Yang, J. J. Wang, D. P. Cao, W. G. Tu, Z. T. Yu and Z. G. Zou, J. Mater. Chem. A, 2017, 5, 21205–21213.
198. W. J. Yin, L. J. Bai, Y. Z. Zhu, S. X. Zhong, L. H. Zhao, Z. Q. Li and S. Bai, ACS Appl. Mater. Interfaces, 2016, 8, 23133–23142.
199. Y. G. Miseki, S. Fujiyoshi, T. Gunji and K. Sayama, J. Phys. Chem. C, 2017, 121, 9691–9697.
200. H. J. Li, W. G. Tu, Y. Zhou and Z. G. Zou, Adv. Sci., 2016, 3, 1500389.
201. P. Zhou, J. G. Yu and M. Jaroniec, Adv. Mater., 2014, 26, 4920–4935.
202. X. Y. Li, H. H. Hu, L. B. Xu, C. Cui, D. G. Qian, S. Li, W. Z. Zhu, P. Wang, P. Lin, J. Q. Pan and C. R. Li, Appl. Surf. Sci., 2018, 441, 61–68.
203. R. Ye, H. Fang, Y. Zheng, N. Li, Y. Wang and X. Tao, ACS Appl. Mater. Interfaces, 2016, 8, 13879–13889.
204. L. J. Zhang, S. Li, B. K. Liu, D. J. Wang and T. F. Xie, ACS Catal., 2014, 4, 3724–3729.
205. T. P. Hu, P. F. Li, J. F. Zhang, C. H. Liang and K. Dai, Appl. Surf. Sci., 2018, 442, 20–29.
206. J. Ding, W. Xu, H. Wan, D. S. Yuan, C. Chen, L. Wang, G. F. Guan and W. L. Dai, Appl. Catal., B, 2018, 221, 626–634.
207. X. Lu, Z. F. Liu, J. W. Li, J. Zhang and Z. G. Guo, Appl. Catal., B, 2017, 209, 657–662.
208. P. Wu, J. R. Wang, J. Zhao, L. J. Guo and F. E. Osterloh, J. Mater. Chem. A, 2014, 2, 20338–20344.
209. G. X. Zhao, X. B. Huang, F. Fina, G. Zhang and J. T. S. Irvine, Catal. Sci. Technol., 2015, 5, 3416–3422.
210. X. Li, J. G. Yu, J. X. Low, Y. P. Fang, J. Xiao and X. B. Chen, J. Mater. Chem. A, 2015, 3, 2485–2534.
211. H. J. Yan, X. J. Zhang, S. Q. Zhou, X. H. Xie, Y. L. Luo and Y. H. Yu, J. Alloys Compd., 2011, 509, L232–L235.
212. Z. Y. Jin, N. Murakami, T. Tsubota and T. Ohno, Appl. Catal., B, 2014, 150–151, 479–485.
213. H. Katsumata, Y. Tachi, T. Suzuki and S. Kaneco, RSC Adv., 2014, 4, 21405.
214. W. Y. Zhu, F. Q. Sun, R. Goei and Y. Zhou, Catal. Sci. Technol., 2017, 7, 2591.
215. T. T. Xiao, Z. Tang, Y. Yang, L. Q. Tang, Y. Zhou and Z. G. Zou, Appl. Catal., B, 2018, 220, 417–428.
216. L. F. Cui, X. Ding, Y. G. Wang, H. C. Shi, L. H. Huang, Y. H. Zuo and S. F. Kang, Appl. Surf. Sci., 2017, 391, 202–210.
217. C. H. Wang, D. D. Qin, D. L. Shan, J. Gu, Y. Yan, J. Chen, Q. H. Wang, C. H. He, Y. Li, J. J. Quan and X. Q. Lu, Phys. Chem. Chem. Phys., 2017, 19, 4507–4515.
218. W. L. Yu, J. X. Chen, T. T. Shang, L. F. Chen, L. Gu and T. Y. Peng, Appl. Catal., B, 2017, 219, 693–704.

---