An investigation on a WO₃/MoO_3−x heterojunction photocatalyst for excellent photocatalytic performance and enhanced molecular oxygen activation ability

Abstract

The activation capacity of molecular oxygen is an important indicator to evaluate the photocatalytic efficiency of a catalyst. In this paper, MoO_3−x nanosheets with oxygen vacancies were deposited on WO₃ nanoflowers to form a WOM heterojunction, which improved the photocatalytic performance and molecular oxygen activation ability of the catalyst. This novel WOM heterojunction exhibited excellent photoactivity for degrading rhodamine B (RhB), photocatalytic water splitting ability and molecular oxygen activation ability under sunlight irradiation. Among all the samples, WOM_83% could degrade 98% of 30 ppm RhB in 40 min. Meanwhile, WOM_83% exhibited the highest hydrogen generation rate (4214.2 μmol h⁻¹ g⁻¹) and the strongest TMB oxidation capacity, and can generate 2.23 μmol of ·O₂⁻ in 50 min. The enhanced photocatalytic performance via heterojunction construction may be attributed to these following reasons: (i) higher light absorption achieved by the MoO_3−x and WO₃ composite; (ii) the matched energy band gap of MoO_3−x and WO₃ led to higher photogenerated carrier mobility; (iii) MoO_3−x with oxygen vacancies as electron traps suppressed the photogenerated carrier recombination; (iv) enhanced molecular oxygen activation ability of the WOM heterojunction, such as the production of ·O₂⁻. The measurements for TMB photo-oxidation and ·O₂⁻ showed excellent molecular oxygen activation ability of WOM heterojunctions. Finally, a possible photocatalytic mechanism was proposed. This study provided a viable option for the application of materials with oxygen vacancies to remove pollutants and improve molecular oxygen activation ability.

---

1. Introduction

With the rapid development of industry in recent decades, more and more factory wastewater is being discharged into rivers and oceans, posing a great threat to the survival of humans and other organisms.^1–3 Dye wastewater has a more serious impact. Dye wastewater has many problems such as dark colour, staining that is difficult to remove, carcinogenicity and teratogenicity, low biodegradability, high stability and high emissions, so researchers have made a number of efforts to eliminate dyes from wastewater.⁴ The usual methods of water treatment are natural sedimentation, electrostatic adsorption and membrane separation,^5,6 but these methods have high costs, low treatment efficiency, incomplete treatment and are prone to secondary pollution.^7,8 Meanwhile, over-consumption of fossil fuels has triggered an energy crisis,^9,10 while hydrogen energy is a green, environmentally friendly and safe source of clean energy.^11–13 Due to the high cost of hydrogen production, photocatalysis was considered as the best way to produce hydrogen in large quantities,¹⁴ and photocatalysis is also an emerging, cutting-edge, green, inexpensive and efficient technology.^15–17 However, the photocatalytic performance of semiconductors depends on the size of the band gap. High band gap semiconductors, such as TiO₂ and ZnO,^18–20 can only be excited by ultraviolet light, which accounts for 5% of the sunlight and has a low utilisation of light energy.²¹ Narrow bandgap semiconductors such as PbS, PbTe, etc. have a narrow bandgap, which makes it easy for photogenerated carriers to recombine and reduces photocatalytic efficiency.

Semiconductors with moderate band gaps have become increasingly popular in recent decades for photocatalytic applications such as BiOIO₃,²² g-C₃N₄,^23,24 CdS,²⁵ Ag₃PO₄,²⁶etc. Among them, tungsten trioxide (WO₃) is a semiconductor with a moderate band gap, high chemical stability and good solar response,^27–30 environmentally friendly and non-toxic, which was widely used in photocatalytic hydrogen production, carbon dioxide reduction and pollutant degradation.³¹ However, WO₃ has many disadvantages such as a small specific surface area and fast photogenerated charge recombination, which inhibits its photocatalytic applications.³² Usually, WO₃ is combined with other semiconductors to build heterojunctions for improving the photocatalytic performance, such as WO₃/TiO₂,^33,34 Ag@WO₃,³⁵ WO₃/CNT³⁶ and others.

Molybdenum trioxide (MoO₃), similar to WO₃, is a moderate band gap semiconductor with important applications in batteries, thin films, and sensors. Due to the low solar light utilization and high photogenerated carrier recombination efficiency of MoO₃, its application in photocatalysis is limited. Recent studies have shown that defect engineering can improve the photocatalytic performance of MoO₃.³⁷ MoO₃ with oxygen vacancies exhibited localized surface plasmon resonance (LSPR) effects, and the LSPR effect can broaden the light absorption range of semiconductors, thus improving the light absorption capacity.^38–41 In addition, oxygen vacancies can act as an electron trap to impede the recombination of photogenerated carriers.⁴² Therefore, MoO_3−x also has a lower band gap and higher photogenerated charge separation efficiency. As a non-precious metal and a semiconductor with the LSPR effect, MoO_3−x has a wider range of photocatalytic applications. MoO_3−x is often used to build heterojunctions with other semiconductors to further enhance the photocatalytic performance, such as MoS₂/MoO_3−x,⁴³ TiO₂/MoO_3−x,⁴⁴ g-C₃N₄/MoO_3−x,^45,46 SnO₂/MoO_3−x,⁴⁷ WO_3−x/MoO_3−x,⁴⁸etc.

Molecular oxygen is a green molecular oxidant. Activated molecular oxygen is called reactive oxygen species (ROS), such as superoxide radicals (·O₂⁻), hydrogen peroxide (·OH), etc. Photocatalysis is a gentle method of generating reactive oxygen species, in which photogenerated electrons and holes interact with molecular oxygen in a series of interactions to produce ROS under light illumination.⁴⁹

There is little research on the photocatalytic performances of MoO₃ with oxygen-vacancies and its heterojunction. In this study, a novel WOM heterojunction photocatalyst was constructed using MoO₃ with oxygen-vacancies and WO₃ by the solvothermal method. To evaluate its photocatalytic performance, the molecular oxygen activation capacity, the photodegradation of RhB and NBT, the photo-oxidation of TMB and the photocatalytic water splitting ability under sunlight irradiation were also investigated.

2. Experiment

2.1 Reagents

The chemicals used in the experiment were all of analytical grade and used without further purification.

N,N′-Dimethylformamide (DMF), ethanol (C₂H₆O), 30% hydrogen peroxide (H₂O₂), ethylenediaminetetraacetic acid (EDTA), tert-butanol (t-BuOH), benzoquinone (BQ), and carbon tetrachloride (CCl₄) were purchased from Shanghai Sinopharm Chemical Reagent Co. Tungsten chloride (WCl₆), nitrotetrazolium blue chloride (NBT), rhodamine B (RhB), and molybdenum powder (Mo) were obtained from Macklin Reagent. L-Histidine (L-His) was bought from Ge Ao Chemical Reagent Co. 3,3′,5,5′-Tetramethylbenzidine (TMB) was bought from Aladdin Reagent.

2.2 Synthesis of the WO₃ nanoflowers

WO₃ nanoflowers were synthesized by the solvothermal method.⁵⁰ 1.0 g of WCl₆ was dissolved in 30 ml DMF, then after the mixture was completely dissolved, it was magnetically stirred for 20 h. Then the solution was transferred to a Teflon high-pressure reactor and heated at 120 °C for 4 h. After cooling to room temperature, a white precipitate was collected and washed three times with deionized and ethanol, respectively. In the end, it was dried in a vacuum drying oven at 60 °C for 12 h. The white solid was annealed in a muffle furnace at 450 °C for 5 h.

2.3 Synthesis of the MoO_3−x nanosheets

MoO_3−x nanosheets were synthesized by a previously reported hydrothermal method.⁵¹ 0.192 g of molybdenum powder was dispersed in 24 ml of ethanol and 88 mmol of 30% hydrogen peroxide was added, after stirring for 30 minutes, the molybdenum powder dissolved and formed a yellow solution, then the solution was transferred to a Teflon high-pressure reactor and heated at 160 °C for 12 h. After cooling to room temperature, the blue-black product was washed three times with ethanol, then dried under air at 60 °C for 12 h.

2.4 Synthesis of WO₃/MoO_3−x heterojunctions

WO₃/MoO_3−x heterojunctions were obtained by a one-step simple solvothermal method. 0.1 g of WO₃ was dispersed in 24 ml of ethanol, then an amount of molybdenum powder and 30% hydrogen peroxide were added, followed by stirring for 30 minutes to obtain a yellow solution. The solution was transferred to a Teflon high-pressure reactor and heated at 160 °C for 12 h. After cooling to ambient temperature, the obtained blue-black precipitate was washed three times with ethanol and dried under air at 60 °C for 12 h. Thereinto, different mass ratios of MoO_3−x were denoted as WO₃/MoO_(3−x)y (y = 30%, 50% and 83%), abbreviated as WOM_y, respectively. The synthesis process of WO₃/MoO_3−x is illustrated in Scheme 1.

Scheme 1 Schematic diagram of the synthesis of the WOM_83% heterojunction.

2.5 Photoactivity and cycling experiments

The photodegradation efficiency of all the catalysts was evaluated by degradation of RhB under sunlight irradiation. Before illumination, 10 mg of photocatalyst was added to 50 ml of 30 ppm RhB solution and then stirred for 30 minutes in the dark to gain adsorption–desorption equilibrium. Then, the RhB mixture was photodegraded under a 300 W Xe lamp. After 5 min intervals, 3 ml of the solution was aspirated, the solid particles were removed by a 0.22 μm microporous filter and the actual concentration of RhB was measured with a UV-visible spectrophotometer at the maximum absorption peak at 550 nm.

In order to determine the active species produced in photocatalysis, nitrotetrazolium blue chloride (NBT), ethylenediaminetetraacetic acid (EDTA), L-histidine, and tert-butanol were used as scavengers to capture the superoxide radical (·O₂⁻), hole (h⁺), singlet oxygen (¹O₂), and hydroxyl radical (·OH), respectively. During the photocatalytic reaction, 10 mg EDTA, 10 mg L-histidine,1 ml tert-butanol and 1 ml 0.4 mM NBT were added to the RhB solution, respectively. Then the operations were the same as the photocatalytic degradation TC experiment.

The stability of the WOM_83% heterojunction photocatalyst was measured by six consecutive photodegradation experiments of RhB.

2.6 Photocatalytic hydrogen production test

Before illumination, 30 mg of photocatalyst, 80 ml deionized water, 10 ml 0.25 M sodium sulfide solution, and 10 ml 0.35 M sodium sulfite solution were added to a sealed quartz reactor. After stirring completely, N₂ was slowly passed through the reactor for 30 minutes until interfering gases (like O₂, CO₂) were removed from the reactor. The mixture was irradiated under sunlight with a 300 W Xe lamp, the gas in the reactor was extracted every hour and the amount of hydrogen was measured by a gas chromatograph. Usually, hydrogen was continuously collected over 4 hours; meanwhile, four cycles of hydrogen generation experiments were also conducted.

2.7 Photocatalytic ROS detection

Molecular oxygen activation capacity was measured by 3,3′,5,5′-tetramethylbenzidine (TMB) photo-oxidation.⁵² Usually, 4 mg of photocatalyst was ultrasonically dispersed in 1 ml of pure water, then 0.2 ml of the mixture was added to 50 ml of 0.5 mM TMB HAc/NaAc buffer, the 3,3′,5,5′-tetramethylbenzidine (TMB). The mixture was irradiated under a 300 W Xe lamp. 3 ml solution was taken every 3 minutes and the solid particles were removed by a 0.22 μm microporous filter. The TMB photo-oxidation product concentrations were recorded by a UV-visible spectrophotometer at 370 nm. In order to explore the reactive oxygen species during the photocatalytic reaction, 10 mg EDTA, 10 mg L-histidine, 1 ml 0.4 mM NBT, 1 ml tert-butanol and 1 ml CCl₄ were added to the TMB solution, respectively. Then the operations were the same as the photo-oxidation TMB experiment.

20 ppm NBT was used to measure the ·O₂⁻ concentration, because the reaction molar ratio of NBT with ·O₂⁻ is 1 : 4. Commonly, 10 mg of WO₃, MoO_3−x and WOM_83% was added to 50 ml of 20 ppm NBT solution and then stirred for 30 min in the dark to gain adsorption–desorption equilibrium. Then, the mixture was photodegraded under a 300 W Xe lamp. Every 10 min, 3 ml of the solution was aspirated, the solid particles were removed by a 0.22 μm microporous filter and the actual concentration of NBT was measured by a UV-visible spectrophotometer at the maximum absorption peak at 259 nm.

2.8 Electrochemical tests

Electrochemical measurements were carried out on an electrochemical workstation (CHI 760E) connected to a standard three-electrode system, namely the working electrode prepared from the test sample, a platinum plate as the counter electrode and a saturated Ag/AgCl electrode as the reference electrode. The working electrode was prepared as follows: 10 mg of sample was weighed and poured into a 5 mL EP tube, then 2 mL of Nafion injection (Nafion : isopropanol = 1 : 25) was added and sonicated for 15 min. 200 μL of each drop was applied to the conductive glass surface using a micro syringe, dispersed evenly and dried under a sun lamp. The test system for both was 0.5 M Na₂SO₄ solution, which was required to submerge the three electrodes. Electrochemical impedance spectroscopy (EIS) data were obtained from the electrochemical workstation with a circuit potential in the frequency range of 0.1 Hz–100 kHz and a test voltage of 1.5 eV. Samples were subjected to transient photocurrent response (PC) under sun light conditions (300 W xenon lamp) for 30 s with the lamp on and 30 s with the lamp off for 5 consecutive cycles.

2.9 Characterization

The crystal structures of the photocatalysts were characterized by powder X-ray diffraction (XRD) on a Bruker D8 Advance diffractometer using Cu Kα₁ radiation at an operating voltage of 40 kV and current of 40 mA. The scanning speed was 2° per minute and the angle range was 5–90°. The morphology of the photocatalysts was investigated through scanning electron microscopy (SEM, TESCAN MIRA LMS at an accelerating voltage of 15 kV) and transmission electron microscopy (TEM, JEM-2100 at an accelerating voltage of 200 kV). Elemental distributions of the material were studied with EDS and element mapping images. The element valance states of the photocatalysts were measured through XPS (Thermo Scientific K-Alpha) and using Mg Kα radiation as the non-monochromatized source (hν = 1253.6 eV). Meanwhile, the electron binding energies of the elements were corrected using C 1s as the reference (284.8 eV). The UV-vis diffuse reflectance spectra (DRS) of the photocatalysts were measured using a Shimadzu UV-3600 spectrometer with BaSO₄ as the reference and the wavelength range of the test was 200–800 nm. The Mott–Schottky curves (M–S) of the samples were recorded on a CHI 760E electrochemical system (Chenhua, Shanghai, China) at the frequency of 1000 and 1500 Hz in 0.5 M Na₂SO₄ solution with the Ag/AgCl potential varying from −1.0 to 1.0 eV. ESR was carried out on a Bruker EMXplus-6/1 self-selecting resonator with DMPO and TEMP as trapping reagents and electron paramagnetic resonance (EPR) spectra were recorded on a Bruker EMXplus-6/1 at room temperature and the magnetic field modulation frequency was 100 kHz. The photodegradation products of RhB were recorded by high-performance liquid chromatography-mass spectrometry (HPLC-MS, Agilent-1290/6550-QTOF LC/MS). The Brunauer–Emmett–Teller (BET) specific surface area was measured by the nitrogen adsorption–desorption method using a Micromeritics ASAP 2460. Photoluminescence (PL) spectra were measured with an Edinburgh FLS1000 and the excitation wavelength was 455 nm.

3. 3.Results and discussion

3.1 Characterization

The XRD patterns of the photocatalysts are shown in Fig. 1. For WO₃, the peaks at 22.7°, 24.0°, 28.2°, 33.2° and 34.1° were attributed to the (001), (110), (101), (111) and (200) planes, respectively. The diffraction peak results were consistent with that of WO₃ (PDF#05-0388), indicating the successful synthesis of WO₃. For MoO_3−x, the main diffraction peaks at 12.6°, 23.7°, 25.4°, 27.1°, 33.9°, and 38.5° were attributed to the (020), (110), (040), (021), (111) and (060) crystal planes of MoO_3−x, respectively, which was consistent with the results reported in the literature.⁵¹ For WOM_83%, the diffraction peaks appeared at 12.6°, 23.7°, 25.4°, 27.1°, 33.9°, and 38.5°, which corresponded to MoO_3−x. Meanwhile, the diffraction peaks of WO₃ were discovered at 22.7°, 24.0°, 28.2° and 33.2°, revealing that both WO₃ and MoO_3−x appeared in the heterojunction.

Fig. 1 XRD patterns of WO₃, MoO_3−x, and WOM_83%.

The microscopy morphologies of the monomers and the heterojunctions were studied by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As shown in Fig. 2(a) and (d), spherical WO₃ nanoflowers self-assembled from some nanosheets with a length of tens to hundreds of nm. As shown in Fig. 2(b) and (e), MoO_3−x showed an irregular sheet-like stacked structure, and the average length was about 2–3 μm, which was consistent with the reported literature.⁵¹ As shown in Fig. 2(c) and S1,† the WOM_83% composite exhibited a distinct nanoflower shape, and MoO_3−x nanosheets with smaller size grew tightly on the surface of the larger WO₃ nanoflowers. Meanwhile, MoO_3−x nanosheets also self-assembled into a nano-flower shape, increasing the specific surface area of the catalyst. In addition, the lattice spacing was obtained from the HRTEM image of Fig. 2(f), and the lattice pitch of 0.39 and 0.38 nm was attributed to the (001) crystal planes of WO₃ and the (110) crystal planes of MoO_3−x, respectively, which indicated that WO₃ and MoO_3−x existed and interacted with each other in the heterojunctions. In order to further research the elements in WOM_83%, EDS mapping was performed and shown in Fig. S2.† The results showed the presence of W, Mo and O elements, and these elements were evenly distributed throughout the composites, indicating the successful synthesis of the WOM_83% heterojunction.

Fig. 2 (a) SEM image of WO₃, (b) SEM image of MoO_3−x, (c) SEM image of WOM_83%, (d) TEM image of WO₃, (e) TEM image of MoO_3−x and (f) HRTEM image of WOM_83%.

In order to investigate the specific surface area and pore size distribution of WO₃, MoO_3−x and WOM_83%, BET analysis was conducted. The results are shown in Fig. 3(a). WO₃, MoO_3−x and WOM_83% exhibited type IV isotherms and H3 hysteresis loops, indicating the mesoporous properties of these photocatalysts. Meanwhile, the specific surface areas of WO₃, MoO_3−x and WOM_83% were 21.31 m² g⁻¹, 14.11 m² g⁻¹ and 19.64 m² g⁻¹, respectively. Besides, the pore size distribution shown in Fig. 3(b), (c) and (d) indicated that the pore size of prepared WO₃, WOM_83% and MoO_3−x was about 10.92, 23.47 and 25.98 nm, respectively. The average pore volume increased sequentially and could provide more reactive active sites for adsorption and photocatalysis of RhB.

Fig. 3 (a) N₂ adsorption–desorption isotherm of WO₃, MoO_3−x and WOM_83%; (b), (c) and (d) pore size distribution curves for WO₃, MoO_3−x and WOM_83%.

X-ray photoelectron spectroscopy (XPS spectroscopy) was used to study the major elements and chemical bonding energy changes on the surface of the monomers and the composites. The full XPS elemental spectra and high resolution XPS spectra of all the catalysts are shown in Fig. S3† and 4, respectively. The full XPS spectrum exhibited the presence of O W, Mo elements in WOM_83%, indicating the successful synthesis of the composites. Fig. 4(a) shows the fine spectra of W 4f in WO₃ and WOM_83%. For WO₃, the peaks of W 4f_5/2 (37.8 eV) and W 4f_7/2 (35.67 eV) were attributed to W⁶⁺. In WOM_83%, the peaks of W 4f_5/2 (41.21 eV) and W 4f_7/2 (39.14 eV) were attributed to W⁶⁺. Meanwhile, the peaks of WOM_83% shifted to higher binding energy by 3.41 and 3.47 eV, respectively, indicating a strong interaction of MoO_3−x and WO₃ in WOM_83%.

Fig. 4 XPS spectra of WO₃, MoO_3−x and WOM_83%: (a) W 4f, (b) Mo 3d, (c) and (d) O 1s.

As shown in Fig. 4(b), in MoO_3−x, the peaks of Mo 3d_3/2 (235.86 eV) and Mo 3d_5/2 (232.73 eV) were attributed to Mo⁶⁺. However, the peaks of Mo 3d_3/2 (234.75 eV) and W 3d_5/2 (231.57 eV) were attributed to Mo⁵⁺, indicating the presence of Mo⁵⁺ in MoO_3−x.⁵¹ For WOM_83%, the peaks of Mo⁶⁺ and Mo⁵⁺ shifted to lower binding energy by 0.08, 0.09, 0.12 and 0.12 eV, respectively.

All of these further supported the existence of the interaction between WO₃ and MoO_3−x in the WOM_83% heterojunction. O 1s spectra are shown in Fig. 4(c) and (d). The peaks of O (530.48 eV) in MoO_3−x, the peaks of O (530.51 eV) in WOM_83% and the peaks of O (530.38 eV) in WO₃ were attributed to metal oxides, respectively. However, both WOM_83% and MoO_3−x showed a faint oxygen peak (531.8 eV), which was attributed to oxygen vacancies. The presence of oxygen vacancies was demonstrated by XPS. To further prove the existence of oxygen vacancies, the oxygen vacancies of MoO_3−x and WOM_83% were measured through EPR, and the results are shown in Fig. 5(a). Both MoO_3−x and WOM_83% exhibited a signal peak at the g value of 2.003, which confirmed the presence of oxygen vacancies in MoO_3−x and the WOM_83% heterojunction.⁵¹ Due to the electron trapping effect of oxygen vacancies, the electron cloud was biased from WO₃ to MoO_3−x, strengthening the interaction between WO₃ and MoO_3−x in the heterojunctions.

Fig. 5 (a) EPR spectra of MoO_3−x and WOM_83%; (b) PL spectra of WO₃, MoO_3−x, and WOM_83%.

Meanwhile, in order to explore the photogenerated carrier separation ability of the photocatalysts, PL was measured, as shown in Fig. 5(b). WO₃ exhibited the highest emission peaks due to having the strongest recombination of photogenerated carriers. Compared with WO₃, MoO_3−x and WOM_83% exhibited low emission peak intensity. Due to the WOM_83% was constructed, which showing the lowest emission peaks intensity. Therefore, the presence of oxygen vacancies promoted the photogenerated carrier separation, and construction of heterojunctions can further enhance the photogenerated charge separation efficiency.

The light absorption capacity and absorption range of the photocatalysts were researched by UV-vis spectroscopy. The UV-vis DRS spectra of WO₃, MoO_3−x and WOM_83% are shown in Fig. 6(a). The light absorption ability of MoO_3−x was stronger than that of WO₃, and WOM_83% exhibited strong light absorption ability at 200–800 nm, indicating that MoO_3−x was an excellent photoresponsive catalyst due to the LSPR effect. In addition, compared to MoO_3−x, the absorption ability of WOM_83% was enhanced at 200–800 nm, which indicated that MoO_3−x interacted with WO₃. The band gap of the catalyst was calculated by the equation αhv = (hv − E_g)^1/2, in which α, v and E_g are the absorption coefficient, photon frequency and band gap of the photocatalyst, respectively. As shown in Fig. 6(b), the band gaps of WO₃, MoO_3−x, and WOM_83% were 2.97, 2.81 and 2.78 eV, respectively. The conduction band values of WO₃ and MoO_3−x were measured from MS curves. When the tangent line of the M–S curve of a semiconductor has a positive slope, the semiconductor is an n-type semiconductor. Fig. 6(c) and (d) showed that both WO₃ and MoO_3−x were n-type semiconductors, and the E_fb of WO₃ and MoO_3−x was −0.4 V and −0.62 V versus Ag/AgCl, respectively. As the electrode potential of Ag/AgCl was 0.2224 V, the E_fb of WO₃ and MoO_3−x was −0.18 V and −0.4 V vs. NHE, respectively. Meanwhile, because WO₃ and MoO_3−x were n-type semiconductors, the position of E_fb was very close to that of the CB. According to E_VB = E_CB + E_g, the VB values of WO₃ and MoO_3−x were 2.79 and 2.41 eV, respectively. As shown in Fig. S4,† the VB values for WO₃ and MoO_3−x were 2.79 and 2.41 eV, respectively, which was consistent with the calculated values.

Fig. 6 (a) UV-vis DRS spectra of the photocatalysts, (b) band gap diagrams of the photocatalysts, (c) and (d) M–S curves of WO₃ and MoO_3−x.

In order to study the generation and separation efficiency of photogenerated carriers, photocurrent (PC) response and electrical impedance spectroscopy (EIS) were performed on the photocatalysts. As shown in Fig. 7(a), 5 irradiation cycles were conducted with each cycle being 60 seconds under sunlight irradiation. Usually, the higher the value of photocurrent response, the higher the generation and separation efficiency of the photogenerated electrons and holes. WO₃ showed weak photocurrent response, and WOM_83% exhibited a stronger photocurrent response than all the photocatalysts, indicating that WOM_83% had the strongest photogenerated charge generation and separation efficiency. With increasing the content of MoO_3−x in the heterojunction, the photocurrent increased gradually and was very stable, which further verified that the presence of oxygen vacancies as electron traps inhibited the recombination of photogenerated carriers. The electrical impedance spectroscopy (EIS) results are shown in Fig. 7(b). Component-based simulations show that the diameters of the Nyquist semicircle for the three heterojunctions WO₃/MoO_(3−x)y (y = 30%, 50% and 83%) in the high-frequency region were much smaller than that of WO₃, suggesting the higher efficiency of interfacial photocharge-transfer of the heterojunctions formed between MoO_3−x nanosheets and WO₃. Meanwhile, WOM_83% showed the smallest radius among the heterojunctions and monomers, which indicated that it had the lowest interfacial carrier transfer impedance, thus WOM_83% has a high photocatalytic capacity.

Fig. 7 (a) Photocurrent response graph of WO₃, MoO_3−x and WOM_y (y = 30%, 50%, 83%) and (b) EIS diagram of WO₃, MoO_3−x and WOM_y (y = 30%, 50%, 83%).

3.2 Photocatalytic performance

The photocatalytic ability of the as-prepared photocatalysts was evaluated by the degradation of RhB in an aqueous solution under sunlight irradiation at room temperature. In the RhB degradation experiment, 10 mg of the photocatalyst was added to 50 ml of 30 ppm RhB. In addition, a blank experiment was also performed. Before photocatalysis operations, all the photocatalysts were stirred for 30 minutes in RhB solution in the dark to establish adsorption–desorption equilibrium. After dark treatment, 3 ml of RhB mixture solution was taken and marked as C₀. In photocatalytic degradation, 3 ml of mixture solution was taken every 5 minutes and marked as C₅, C₁₀, etc. The results are displayed in Fig. 8(a).

Fig. 8 (a) Photocatalytic degradation curves of RhB over different photocatalysts; (b) pseudo-first-order kinetic curves of photocatalytic degradation performance of all the photocatalysts.

As shown in Fig. 8(a), the concentration of RhB has no evident change in the blank experiment, indicating that the degradation of RhB was negligible under visible light without photocatalysts. WO₃ exhibited negligible photocatalytic activity, while the degradation rate can reach 42% over MoO_3−x within 40 min. After forming the WOM heterojunction, WOM_83% displayed the highest photocatalytic performance, the degradation rate increased to 98% in 40 min and the degradation velocity was faster than the other photocatalysts.

In addition, a kinetic study was used to further illustrate the photocatalytic activity of RhB degradation, and it followed the pseudo-first-order dynamics model below:

−ln(C_t/C₀) = k·t (1)

in which C_t and C₀ are the actual and initial concentrations of RhB, respectively, and k is the kinetic rate constant. The value of k was calculated from the slope and intercept of the linear graph. As shown in Fig. 8(b), apparently, WOM_83% (0.09173 min⁻¹) revealed the highest k value among the photocatalysts, which was about 4.83 and 20.8 times that of the pure MoO_3−x (0.01896 min⁻¹) and WO₃ (0.00441 min⁻¹), respectively. In summary, loading MoO_3−x with oxygen vacancies on WO₃ helps to expand the sunlight absorption and separate the photogenerated carriers, thus enhancing the photocatalytic activity.

3.3 Stability

The stability of the WOM_83% heterojunction was assessed by recycling tests for the photocatalytic degradation of RhB. The XRD of WOM_83% after the recycling tests is displayed in Fig. 9(a). The peak pattern of the XRD spectra before and after the recycling tests was almost unchanged. As shown in Fig. 9(b), after 6 cycles of experiments, the photodegradation efficiency of WOM_83% reduced slightly, which further proved the stability of WOM_83% in photodegradation.

Fig. 9 (a) XRD results of WOM_83% before and after six cycles; (b) degradation performance of WOM_83% in six consecutive photocatalytic degradation experiments for RhB.

3.4 Photocatalytic hydrogen production and cyclic stability

To further evaluate the photocatalytic performance of the catalysts, the photocatalytic hydrogen production was researched in deionized water under sunlight irradiation. The results are shown in Fig. 10(a); the amount of hydrogen produced by the monomer WO₃, MoO_3−x and WOM_83% heterojunction in 4 hours was 2222.3 μmol g⁻¹, 984.6 μmol g⁻¹ and 16855.07 μmol g⁻¹, respectively. WOM_83% had the strongest hydrogen production in 4 hours, which was 7.6 and 17.1 times higher than those of pure WO₃ and MoO_3−x, respectively. Meanwhile, the hydrogen generation rate of the photocatalysts is exhibited in Fig. 10(b). The hydrogen generation rate of WO₃, MoO_3−x and WOM_83% in 4 hours was 555.6 μmol h⁻¹ g⁻¹, 246.1 μmol h⁻¹ g⁻¹ and 4214.2 μmol h⁻¹ g⁻¹, respectively. WOM_83% also had the highest hydrogen generation rate in all the photocatalysts, indicating that after building the WOM heterojunction, the presence of oxygen vacancies facilitates the separation of photogenerated carriers, thus, WOM_83% had the strongest hydrogen production and highest hydrogen production rate within 4 hours, and constructing heterojunctions can enhance the photocatalytic splitting ability of water.

Fig. 10 (a) Hydrogen production curves and (b) hydrogen generation rates of all the photocatalysts.

A cyclic photocatalytic hydrogen evolution experiment was carried out to study the stability of the WOM_83% heterojunction. After each cycle, the reaction mixture was bubbled with N₂ and then used for subsequent runs. As shown in Fig. S5,† the successive H₂ production demonstrated that the WOM_83% heterojunction had high stability and efficiency.

3.5 Photocatalytic mechanism and test of molecular oxygen activation

In order to understand the full photodegradation process and research the molecular oxygen activation capacity of the catalysts, a free radical trapping experiment was conducted. Usually, the scavengers EDTA, t-BuOH, NBT, and L-His were used to remove h⁺, ·OH, ·O₂⁻, and ¹O₂, respectively. As shown in Fig. 11, the addition of EDTA, NBT, and L-His greatly inhibited the photocatalytic performance of the catalyst. Meanwhile, the effect of different doses of trapping agents on the photocatalytic performance was studied (Fig. 12). The results suggested that h⁺, ·O₂⁻ and ¹O₂ were the main active species in the RhB photodegradation system. To demonstrate the existence of ·O₂⁻ in the photocatalytic process, the ESR spin capture technique was performed. As shown in Fig. 11b–d, no ·O₂⁻ or ·OH signal peaks were produced under dark conditions and the signal peaks gradually increased with the extension of illumination time. DMPO–·O₂⁻ showed strong characteristic signal peaks under sunlight irradiation. It should be noted that ·OH characteristic signal peaks were also detected, which was mainly because the water molecule was oxidated by h⁺ to generate ·OH. In addition, ¹O₂ was not present in the ¹O₂ ESR profile, indicating that ¹O₂ was produced by photosensitization of RhB. Combining the free radical trapping tests and the ESR results, h⁺ and ·O₂⁻ were the main active species in this photodegradation system.

Fig. 11 (a) Radical trapping test for WOM_83%; ESR spectra of (b) DMPO–·O₂⁻, (c) DMPO–·OH and (d) TEMP–¹O₂ over WOM_83%.

Fig. 12 (a–d) Radical trapping test for WOM_83%.

Moreover, the molecular oxygen activation ability of the catalysts was also studied. The TMB oxidation activity of the photocatalysts was determined under sunlight illumination (Fig. 13a). The results showed that WO₃ and MoO_3−x exhibited common oxidizing ability, but WOM_83% had enhanced TMB photooxidation ability; meanwhile, WOM_83% has the highest activity to photooxidize TMB, which indicated that constructing heterojunctions can enhance the photocatalytic ability. In order to explore the oxygen species in TMB photooxidation, different scavengers were added in the photooxidation process of TMB under sunlight. The scavengers BQ, L-His, EDTA, t-BuOH and CCl₄ were used to capture ·O₂⁻, ¹O₂, h⁺, ·OH and e⁻, respectively, and the result is displayed in Fig. 13(b). When different trapping agents were added, the photooxidation capacity of TMB over WOM_83% was inhibited. Evidently, BQ and EDTA greatly inhibited the photooxidation of TMB, which indicated that ·O₂⁻ and h⁺ were the main active species in TMB oxidation.

Fig. 13 (a) Absorbance evolution of TMB oxidation with different photocatalysts at 371 nm, (b) absorbance of TMB oxidation with WOM_83% at 371 nm with different scavengers, (c) time-dependent degradation curves of NBT over the catalysts, and (d) time-dependent concentration plots of ·O₂⁻.

In order to study the production of ·O₂⁻ during photocatalysis, NBT photodegradation over WOM_83% was performed. As the reaction molar ratio of NBT to ·O₂⁻ is 1 : 4, the ·O₂⁻ concentration could be quantitatively calculated by NBT photodegradation. The UV-vis maximum absorption peak of NBT at 259 nm gradually decreased with the increase of photodegradation time. As shown in Fig. 13(c) and (d), NBT was almost not degraded in the presence of WO₃, which was similar to the result of the blank condition, indicating that WO₃ cannot produce ·O₂⁻. For WOM₈₃ and MoO_3−x, about 45.7% and 15.6% of NBT were degraded, and the corresponding generation amounts of ·O₂⁻ were 2.23 and 0.76 μmol, respectively. This result indicated that WOM_83% had stronger activity to produce ·O₂⁻ than pure WO₃ and MoO_3−x, and constructing heterojunctions could enhance the production capacity of ·O₂⁻.

In order to investigate the specific process of RhB photodegradation, the HPLC-MS system was used to determine the degradation products at different time intervals. Several major intermediates were studied and the variation in relative peak intensities is shown in Fig. S6.† As the degradation time increased, the RhB peak at m/z = 479 disappeared and a number of fragment peaks (m/z = 359, 331, 301, 270, 218, 142, 122, 102 and 60) appeared. The results showed that the relative intensities of the fragment peaks (m/z = 359, 331, 301 and 102) decreased and the relative intensities of the fragment peaks (m/z = 270, 218 and 60) increased, indicating that the degradation of the RhB intermediate had occurred. After analysis, a potential degradation mechanism for RhB was proposed. As shown in Fig. 14, the RhB molecule was directly demethylated to give intermediate A (m/z = 359), and A was demethylated to yield intermediate B (m/z = 331). Removal of a nitrogen atom produced intermediate C (m/z = 301), which was further demethylated to give intermediate D (m/z = 270). Subsequently, due to the progressive rupture of the hexameric ring, intermediate E (m/z = 218) was produced, and intermediates F (m/z = 142) and G (m/z = 122) were obtained by hydrogenated carbon bond breakage and dehydroxylation. The final rupture of the hexameric ring gave small fragmented molecules m/z = 102 and 60. After 35 minutes, there was almost no molecular ion peak for RhB, indicating that RhB (m/z = 479) could be completely degraded within 35 min.

Fig. 14 Possible photodegradation pathway of RhB over WOM_83%.

Based on the above experimental results, a possible mechanism for the degradation of RhB over WOM heterojunctions is proposed in Scheme 2. Based on the results of DRS, M–S curves and VB-XPS, the CB of WO₃ and MoO_3−x was −0.18 eV and −0.40 eV, respectively; meanwhile, the VB of WO₃ and MoO_3−x was 2.79 eV and 2.41 eV, respectively.

Scheme 2 Diagram of the photocatalytic mechanism of WOM_83% under sunlight.

Based on the band structure of the heterojunction, we proposed a possible type-II heterojunction between MoO_3−x and WO₃. Under sunlight irradiation, both MoO_3−x and WO₃ produced photogenerated electron–hole pairs. Due to the presence of oxygen vacancies in MoO_3−x, photogenerated electrons can be more easily excited to the CB by the defective energy level.⁵³ Because of the electron trapping effect of the oxygen vacancies, the recombination of photogenerated carriers was restricted. Meanwhile, the presence of oxygen vacancies may lead to the formation of intermediate energy bands below the CB. Therefore, there were three possible ways for transfer of photogenerated electrons: (1) photogenerated electrons excited from the VB to the CB; (2) photogenerated electrons excited from the VB to the oxygen-deficient states; and (3) photogenerated electrons excited from the oxygen-deficient states to the CB. Thus, the presence of oxygen-deficient states further inhibited the photogenerated carrier recombination and reduced the direct excitation energy of photogenerated electrons. If the traditional type II electron–hole transfer mechanism occurred, electrons in the CB of MoO_3−x were injected into the CB of WO₃ and holes in the VB of WO₃ were injected into the VB of MoO_3−x. However, as the conduction bands of WO₃ were higher than that of the superoxide radical (O₂/˙O₂⁻ = −0.33 eV vs. NHE), the WOM_83% heterojunction cannot produce ·O₂⁻, which did not correspond to the results of the experiments. Therefore, the Z scheme was exploited to explain the photocatalytic mechanism. The photogenerated electrons in the CB of WO₃ migrate to the VB of MoO_3−x to combine with photogenerated holes. In addition, the CB value of MoO_3−x (−0.4 eV) was more negative than the potential of ·O₂⁻ (O₂/·O₂⁻ = −0.33 eV vs. NHE), which allowed the photogenerated electrons to interact with oxygen in RhB solution to generate ·O₂⁻. Meanwhile, the VB of WO₃ (2.79 eV) was more positive than the hydroxyl radical generation potential (·OH/H₂O = 2.38 eV, vs. NHE), and the water in solution was oxidized to ·OH by h⁺, while h⁺ also participated in the oxidation of RhB. Moreover, due to the photosensitivity of RhB, ¹O₂ was also produced. Eventually, these reactive oxygen species (·O₂⁻, and ·OH), along with h⁺ and ¹O₂, were involved in the degradation of RhB, which was consistent with the results of molecular oxygen activation experiments, free radical capture tests and ESR.

The process of RhB photodegradation by WOM_83% may be described as follows:

WO₃/MoO_3−x + hv → [WO₃(e⁻ + h⁺)/MoO_3−x(e⁻ + h⁺)] (2)

MoO_3−x(e⁻)VB → CB (3)

MoO_3−x(e⁻)VB → OV (4)

MoO_3−x(e⁻)OV → CB (5)

WO₃(e⁻) → MoO_3−x(h⁺) (6)

O₂ + e⁻ → ·O₂⁻ (7)

H₂O + h⁺ → ·OH (8)

RhB + h⁺ → Degradation products(CO₂ + H₂O + others) (9)

RhB + ·O₂⁻ + ·OH → Degradation products(CO₂ + H₂O + others) (10)

4. Conclusion

In this work, a Z type WO₃/MoO_3−x heterojunction was constructed by in situ growth of MoO_3−x nanosheets on the surface of WO₃ nanoparticles via a simple solvothermal method. The WO₃/MoO_3−x heterojunction with abundant oxygen vacancies can significantly improve the photocatalytic performance and molecular oxygen activation, and the H₂ production and the RhB degradation rate were 17.1 and 4.83 times higher than those of the pure MoO_3−x nanosheets, respectively. The outstanding photo-oxidative reduction ability of WO₃/MoO_3−x should be attributed to the formation of the heterojunction structure with strong charge transport ability. The photoelectrochemical measurements and PL spectroscopy further confirmed the substantial increase in the electron–hole pair separation efficiency for the WO₃/MoO_3−x heterojunction. In addition, the presence of oxygen vacancies enhanced the molecular oxygen activation capacity, further facilitating the photocatalytic reaction. This work provides new opportunities for the design of other novel oxygen vacancy catalysts without precious metals.

Author contributions

Yuxuan Shao and Dan You: conceptualization; Yuqi Wan: investigation; Qingrong Cheng: funding acquisition; Zhiquan Pan: administration.

Conflicts of interest

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

Acknowledgements

This work was supported by the Graduate Innovative Fund of Wuhan Institute of Technology [CX2022453].

References

1. Z. Li, G. Huang, K. Liu, X. Tang, Q. Peng, J. Huang, M. Ao and G. Zhang, J. Cleaner Prod., 2020, 272, 122892.
2. G. Palanisamy, K. Bhuvaneswari, T. Pazhanivel and G. Bharathi, Optik, 2020, 204, 164171.
3. A. Singh, A. K. Singh, J. Liu and A. Kumar, Catal. Sci. Technol., 2021, 11, 3946–3989.
4. H. Zhang, H. Yang, Z. Pan and Q. Cheng, React. Chem. Eng., 2022, 7, 1626–1639.
5. A. Ali, U. Garg, K. U. Khan and Y. Azim, J. Polym. Environ., 2022, 30, 4435–4451.
6. E. Matei, C. I. Covaliu-Mierla, A. A. Turcanu, M. Rapa, A. M. Predescu and C. Predescu, Membranes, 2022, 12, 1–35.
7. J. Gan, X. Li, K. Rizwan, M. Adeel, M. Bilal, T. Rasheed and H. M. N. Iqbal, Chemosphere, 2022, 286, 131710.
8. H. Zhao, Z. Xing, S. Su, S. Song, T. Xu, Z. Li and W. Zhou, Appl. Mater. Today, 2020, 21, 100821.
9. J. Ke, J. Liu, H. Sun, H. Zhang, X. Duan, P. Liang, X. Li, M. O. Tade, S. Liu and S. Wang, Appl. Catal., B, 2017, 200, 47–55.
10. X. Yang, A. Banerjee, Z. Xu, Z. Wang and R. Ahuja, J. Mater. Chem. A, 2019, 7, 27441–27449.
11. Z. Kong, J. Zhang, H. Wang, B. Cui, J. Zeng, X. Chen, P. Tian, G. Huang, J. Xi and Z. Ji, Int. J. Energy Res., 2019, 44, 1205–1217.
12. L. Su, L. Luo, H. Song, Z. Wu, W. Tu, Z.-j. Wang and J. Ye, Chem. Eng. J., 2020, 388, 124346.
13. V. Navakoteswara Rao, P. Ravi, M. Sathish, M. Vijayakumar, M. Sakar, M. Karthik, S. Balakumar, K. R. Reddy, N. P. Shetti, T. M. Aminabhavi and M. V. Shankar, J. Hazard. Mater., 2021, 415, 125588.
14. S. Mao, Y. Zou, G. Sun, L. Zeng, Z. Wang, D. Ma, Y. Guo, Y. Cheng, C. Wang and J. W. Shi, J. Colloid Interface Sci., 2021, 581, 1–10.
15. N. Chang, Y.-R. Chen, F. Xie, Y.-P. Liu and H.-T. Wang, Colloids Surf., A, 2021, 616, 126351.
16. L. Sun, Y. Yuan, F. Wang, Y. Zhao, W. Zhan and X. Han, Nano Energy, 2020, 74, 104909.
17. Q. Xi, J. Liu, Z. Wu, H. Bi, Z. Li, K. Zhu, J. Zhuang, J. Chen, S. Lu, Y. Huang and G. Qian, Appl. Surf. Sci., 2019, 480, 427–437.
18. Y. Yang, M. Liu, S. Han, H. Xi, C. Xu, R. Yuan, J. Long and Z. Li, Appl. Surf. Sci., 2021, 537, 147991.
19. H.-B. Zheng, D. Wu, Y.-L. Wang, X.-P. Liu, P.-Z. Gao, W. Liu, J. Wen and E. V. Rebrov, J. Alloys Compd., 2020, 838, 155219.
20. L. Wang, G. Huang, L. Zhang, R. Lian, J. Huang, H. She, C. Liu and Q. Wang, J. Energy Chem., 2022, 64, 85–92.
21. P. Gholami, A. Khataee and A. Bhatnagar, J. Cleaner Prod., 2020, 275, 124157.
22. Z. Zhu, C. Zhu, C. Hu and B. Liu, J. Colloid Interface Sci., 2022, 607, 595–606.
23. Z. Lu, D. Zeng, H. Zheng, Q. Liu, X. Gao, X. He, L. Wei and W.-J. Ong, J. Mater. Sci., 2020, 55, 13114–13126.
24. Z. You, X. Yue, D. Zhang, J. Fan and Q. Xiang, J. Colloid Interface Sci., 2022, 607, 662–675.
25. D. You, Z. Pan and Q. Cheng, J. Alloys Compd., 2023, 930, 167069.
26. Y. Zhu, Y. Zhuang, L. Wang, H. Tang, X. Meng and X. She, Chin. J. Catal., 2022, 43, 2558–2568.
27. Y. Ishida, S. Motono, W. Doshin, T. Tokunaga, H. Tsukamoto and T. Yonezawa, ACS Omega, 2017, 2, 5104–5110.
28. I. M. Szilágyi, B. Fórizs, O. Rosseler, Á. Szegedi, P. Németh, P. Király, G. Tárkányi, B. Vajna, K. Varga-Josepovits, K. László, A. L. Tóth, P. Baranyai and M. Leskelä, J. Catal., 2012, 294, 119–127.
29. F. Pei, S. Feng, Y. Wu, X. Lv, H. Wang, S. M. Chen, Q. Hao, Y. Cao, W. Lei and Z. Tong, Biosens. Bioelectron., 2021, 189, 113373.
30. G. Zheng, J. Wang, H. Li, Y. Li and P. Hu, Appl. Catal., B, 2020, 265, 118561.
31. D. Pan, S. Xiao, X. Chen, R. Li, Y. Cao, D. Zhang, S. Pu, Z. Li, G. Li and H. Li, Environ. Sci. Technol., 2019, 53, 3697–3706.
32. Q. Liu, F. Wang, H. Lin, Y. Xie, N. Tong, J. Lin, X. Zhang, Z. Zhang and X. Wang, Catal. Sci. Technol., 2018, 8, 4399–4406.
33. Q. Wang, W. Zhang, X. Hu, L. Xu, G. Chen and X. Li, J. Water Process. Eng., 2021, 40, 101943.
34. S. Higashimoto, Y. Ushiroda and M. Azuma, Top. Catal., 2008, 47, 148–154.
35. J. Zhang, X. Fu, H. Hao and W. Gan, J. Alloys Compd., 2018, 757, 134–141.
36. A. A. Isari, M. Mehregan, S. Mehregan, F. Hayati, R. Rezaei Kalantary and B. Kakavandi, J. Hazard. Mater., 2020, 390, 122050.
37. Y. Zhang, X. Yu, H. Liu, X. Lian, B. Shang, Y. Zhan, T. Fan, Z. Chen and X. Yi, Environ. Sci.: Nano, 2021, 8, 2049–2058.
38. P. Wang, M. Li, T. Song, P. Yang and G. Gao, Nanotechnology, 2020, 31, 425605.
39. Y. Ren, C. Li, Q. Xu, J. Yan, Y. Li, P. Yuan, H. Xia, C. Niu, X. Yang and Y. Jia, Appl. Catal., B, 2019, 245, 648–655.
40. D. Luo, L. Peng, Y. Wang, X. Lu, C. Yang, X. Xu, Y. Huang and Y. Ni, J. Mater. Chem. A, 2021, 9, 908–914.
41. Z. Sun, Y. Gao, C. Ban, J. Meng, J. Wang, K. Wang and L. Gan, ACS Appl. Nano Mater., 2023, 6, 8635–8642.
42. J. Wu, Y. Tao, C. Zhang, Q. Zhu, D. Zhang and G. Li, J. Hazard. Mater., 2023, 443, 130363.
43. J. Li, X. Xu, B. Huang, Z. Lou and B. Li, ACS Appl. Mater. Interfaces, 2021, 13, 10047–10053.
44. J. Shang, X. Xu, K. Liu, Y. Bao, Y. Yang and M. He, Ceram. Int., 2019, 45, 16625–16630.
45. F. Zheng, F. Dong, Z. Lv, H. Li, L. Zhou, Y. Chen, T. Huo and X. Luo, Colloid Interface Sci. Commun., 2021, 43, 100434.
46. Y. Guo, B. Chang, T. Wen, S. Zhang, M. Zeng, N. Hu, Y. Su, Z. Yang and B. Yang, J. Colloid Interface Sci., 2020, 567, 213–223.
47. H. Zhang, G. Wang, G. Dai and X. Xu, Res. Chem. Intermed., 2019, 45, 2369–2381.
48. Y. Liu, X. Dong, Q. Yuan, J. Liang, Y. Zhou, X. Qu and B. Dong, Colloids Surf., A, 2021, 621, 126582.
49. Y. Yang, C. Zhang, D. Huang, G. Zeng, J. Huang, C. Lai, C. Zhou, W. Wang, H. Guo, W. Xue, R. Deng, M. Cheng and W. Xiong, Appl. Catal., B, 2019, 245, 87–99.
50. J.-Y. Shen, L. Zhang, J. Ren, J.-C. Wang, H.-C. Yao and Z.-J. Li, Sens. Actuators, B, 2017, 239, 597–607.
51. Q. Liu, Y. Wu, J. Zhang, K. Chen, C. Huang, H. Chen and X. Qiu, Appl. Surf. Sci., 2019, 490, 395–402.
52. M. Zhang, C. Lai, B. Li, F. Xu, D. Huang, S. Liu, L. Qin, Y. Fu, X. Liu, H. Yi, Y. Zhang, J. He and L. Chen, Chem. Eng. J., 2020, 396, 125343.
53. L. Liang, Q. Chang, T. Cai, N. Li, C. Xue, J. Yang and S. Hu, Chem. Eng. J., 2022, 428, 131139.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3re00396e

---