Strain Adjustment Realizes the Photocatalytic Overall Water Splitting on Tetragonal Zircon BiVO4

Abstract Overall water splitting to generate H2 and O2 is vital in solving energy problem. It is still a great challenge to seek efficient visible light photocatalyst to realize overall water splitting. In this work, the tetragonal zircon BiVO4 is prepared by epitaxial growth on FTO substrate and its overall water splitting reaction is studied. Under the influence of epitaxial strain, the conduction band position shifts negatively and beyond H+/H2 reduction potential (0 V vs NHE), which enables it to possess the photocatalytic hydrogen evolution activity. After loading cocatalysts, the overall water splitting (λ > 400 nm) is realized (H2: ≈65.7 µmol g−1 h−1, O2: ≈32.6 µmol g−1 h−1), and the value of solar hydrogen conversion efficiency is 0.012%. The single‐particle photoluminescence (PL) spectra and PL decay kinetics tests demonstrate the cocatalysts are beneficial to the separation and transfer of carriers. The new strategy of adjusting the band structure by strain is provided.


Introduction
In recent years, solar-driven water splitting to produce clean hydrogen has been considered as a promising approach to solve energy and environmental issues. Since Honda and Fujishima used TiO 2 to produce hydrogen and oxygen in 1972, photocatalytic water splitting has been one of the research hotspots. [1][2][3][4][5][6][7][8][9] Recently, Domen's group used Al-doped SrTiO 3 to achieve a 96% external quantum efficiency of ultraviolet light from 350 to 360 nm, [10] which led the development of photocatalysis, and ignited the DOI: 10.1002/advs.202105299 hope of light-driven water splitting. Although recent research on overall water splitting makes great progress, [10][11][12] developing the efficient photocatalysts which have suitable band structure, visible light response, and high charge separation efficiency is still a great challenge.
BiVO 4 as a chemically stable and visiblelight responsive photocatalyst have drawn much attention. It has three main crystal structures containing monoclinic scheelite (s-m), tetragonal zircon (z-t), and tetragonal scheelite (s-t). Although the z-t BiVO 4 shows lower oxygen evolution activity than s-m BiVO 4 , it does not prevent scientists from further exploring it. Our group has previously reported that z-t BiVO 4 can be used as photocathode to construct a biasfree PEC cell. [13,14] In addition, no matter which crystal phase of BiVO 4 , as the conduction band minimum (CBM) is more positive than H + /H 2 reduction potential, it has basic limitation of releasing H 2 . The current research on BiVO 4 overall water splitting is mainly focused on constructing Z-scheme systems and phase transition-induced band edge engineering. [15,16] However, overall water splitting on pure z-t BiVO 4 still has not come true.
As we all know, the different lattice parameters of substrate and epitaxial layer may cause lattice mismatch, which leads to epitaxial growth with strain accumulation. The mismatch strain at interface will adjust its physical and chemical properties. [17][18][19][20][21] Zheng et al. have reported the textured substrate accumulating strain locally greatly enhances the light absorption and surface reaction of the BiVO 4 photoanode, reducing the total amount of light absorber required. [22] Minseok Choi theoretically reported the tensile strain makes the CBM energy in s-m BiVO 4 very close to H + /H 2 level, while the CBM energy in z-t BiVO 4 shifts upward or even higher than this level. [23] However, there is still lack of general understanding of how epitaxial strain experimentally affects photocatalytic property of BiVO 4 . Here, we take z-t BiVO 4 as the research object, using unstrained BiVO 4 powder and epitaxially strained BiVO 4 to study the effect of strain on the band structure and overall water splitting. during the growth process. Specifically, when epitaxially growing on a substrate with larger lattice parameters, the BiVO 4 unit cell (a = b = 0.730 nm) will suffer in-plane tensile strain. The greater the lattice mismatch is, the larger the in-plane tensile strain, leading to the change of each BiVO 4 unit cell's lattice parameters (a = b). [24] Based on the density functional theory (DFT), the influence of tensile strain on energy band structure of tetragonal BiVO 4 is calculated. As shown in Figure 1a, under tensile strain, the energy band of BiVO 4 undergoes a certain degree of negative shift, in which the valence band (VB) moves to a small extent (inset of Figure 1a), and the conduction band (CB) moves significantly. So here we can envisage that under the effect of strain, the CB position of BiVO 4 can be optimized, and stride the H + /H 2 reduction potential. Combining with its intrinsic oxygen evolution activity, the z-t BiVO 4 has promise to achieve photocatalytic overall water splitting. [15,25,26] Accordingly, the FTO glass consisting of tetragonal fluorine-doped SnO 2 (a = b = 0.476 nm) is chosen as a substrate ( Figure S1, Supporting Information) which could subject the tetragonal BiVO 4 film to the mismatch strain. [27,28] The uniform BiVO 4 film is epitaxially growing on FTO substrate, and the detailed scheme of preparing process is depicted in Figure 1b. As shown, the tetragonal zircon BiVO 4 unit cell consists of a regular VO 4 tetrahedron (V-O bond length is 1.7062 Å) and a slightly twisted BiO 8 dodecahedron (one Bi-O bond length is 2.4142 Å and another is 2.5489 Å, respectively). To make convenience for characterization and testing, the BiVO 4 film grown on FTO www.advancedsciencenews.com www.advancedscience.com substrate is scraped for collection, which is denoted as BiVO 4 -FTO sample. In contrast, the naturally nucleated BiVO 4 is prepared by coprecipitation method, hereinafter referred to BiVO 4 powder. As shown in Figure 1c, all diffraction peaks are consistent with z-t BiVO 4 (space group I41/amd, JCPDS: 14-133). Due to the same symmetry with the [101] direction of SnO 2 , the epitaxially grown BiVO 4 shows strong crystallinity along the [101] orientation. The Raman and X-ray photoelectron spectroscopy (XPS) spectra further show the z-t BiVO 4 is synthesized ( Figures  S2 and S3, Supporting Information). [29] According to the diffuse reflectance spectrum (DRS) of BiVO 4 (Figure 1d), the bandgap of BiVO 4 -FTO is 2.86 eV. As seen from the ultraviolet photoelectron spectroscopy (UPS) spectra ( Figure 1e and Figure S4a, Supporting Information), it can be determined that the E VB and E CB of BiVO 4 -FTO sample are 2.47 and −0.39 eV, respectively. The E CB is more negative than H + to H 2 reduction potential (0 V vs NHE), and E VB is more positive than H 2 O to O 2 oxidation potential (1.23 V vs NHE), indicating the BiVO 4 -FTO sample has the possibility of photocatalytic overall water splitting. The specific calculation process and band position analysis of BiVO 4 powder were depicted in the DRS and UPS spectra in Figure S4 (Supporting Information). Further, as shown in Figure 1f, there is a great difference between BiVO 4 -FTO sample and BiVO 4 powder at the edge of energy bands. The band arrangement of BiVO 4 explains why the BiVO 4 -FTO sample and BiVO 4 powder exhibit significantly different photocatalytic activities.
X-ray diffraction method was used to prove the existence of residual strain. The residual stress can be calculated from the strain which is measured by XRD diffraction. [30] According to the material parameters (Table S1, Supporting Information), the residual stress is 390.29 ± 47.44 MPa, as shown in Supporting Information 2. Although scratching will cause some stress loss, the increase in cell parameters caused by strain is irreversible, which also can be proved by XRD refinement. Through XRD data refinement (Figure 2a-c), all the diffraction peaks of BiVO 4 -FTO film and BiVO 4 -FTO sample as well as BiVO 4 powder are consistent with the Bragg position (represented by the blue-green vertical line) of tetragonal zircon BiVO 4 with space group I4 1 /amd (PDF: . It is worth noting that compared with the naturally nucleated BiVO 4 powder, the main peak of BiVO 4 -FTO sample shifts toward lower angle (Figure 2d), which corresponds to the larger lattice parameters after tensile strain. The detailed crystal structures and unit cell parameters are further established (see Table S2, Supporting Information). As known, when BiVO 4 grows on the substrate with larger lattice parameters, it will suffer in-plane tensile strain and lattice parameters will change. [24] Compared with BiVO 4 powder (a = b = 7.3007 Å), the BiVO 4 -FTO film and BiVO 4 -FTO sample collected from the substrate have the same in-plane lattice parameters (a = b = 7.3044 ≈ 7.3047 Å), which is larger than that of naturally nucleated BiVO 4 powder.

Photocatalytic Activity of BiVO 4
The light absorption of FTO was tested, which appears at about 350 nm ( Figure S8, Supporting Information). To exclude its contribution to H 2 production, the photocatalytic activities of tetragonal BiVO 4 with and without epitaxial strain were tested under visible light irradiation ( > 400 nm). As indicated in Figure S9 (Supporting Information), bare BiVO 4 powder cannot release H 2 , while bare BiVO 4 -FTO sample releases small amount of H 2 . Therefore, to ensure the reaction gas release smoothly, the dual cocatalysts Rh/Cr 2 O 3 and MnO x [10,34] were loaded by photodeposition method. The loading of cocatalysts is proved by SEM, XPS, and EDS analysis (Figures S10-S12, Supporting Information). SEM images ( Figure S10, Supporting Information) demonstrate the deposition of cocatalysts on specific crystal facets, which is consistent with SEM observations (Figure 4). Moreover, XPS and EDS results (Figures S11 and S12, Supporting Information) further confirm that the cocatalysts are successfully deposited on the surface of BiVO 4 . The LSV analysis of tetragonal BiVO 4 before and after loading cocatalysts is shown in Figure S13 (Supporting Information). After loading dual cocatalysts, BiVO 4 -Rh/Cr 2 O 3 /MnO x yields higher photocurrent than pure BiVO 4 and single cocatalyst BiVO 4 -Rh/Cr 2 O 3 , which proves that more photogenerated carriers accumulate on the surface of BiVO 4 -Rh/Cr 2 O 3 /MnO x to participate in surface reactions. The half-reaction of hydrogen evolution (HER) and oxygen evolution (OER) in the presence of sacrificial agents have been operated ( Figure S14, Supporting Information). Compared with BiVO 4 powder, BiVO 4 -FTO sample shows higher photocatalytic HER and OER activity, which is caused by improved reduction ability and carriers' separation efficiency. Further, because s-m phase BiVO 4 is usually used for photocatalytic water oxidation among the BiVO 4 polymorphs, the photocatalytic property of s-m BiVO 4 is also studied for comparison ( Figure S15, Supporting Information). Although s-m BiVO 4 has higher oxygen evolution activity in the presence of sacrificial agent, it cannot evolute H 2 due to insufficient reduction ability. In addition, s-m BiVO 4 cannot epitaxial grow on FTO substrate because of the different crystal structure and too large lattice mismatch between s-m BiVO 4 and SnO 2 . As shown in Figure 5a, under visible light irradiation ( > 400 nm), the simultaneous release of hydrogen and oxygen is achieved on BiVO 4 -FTO sample with cocatalysts, and the solar hydrogen conversion efficiency (STH) is 0.012%. Other single-step overall water splitting photocatalysts are shown in Table S3 (Supporting Information). It can be clearly seen from Figure 5b that the water splitting activity of the photocatalyst gradually decreases during the 12 h irradiation period. The origin of this phenomenon was further studied. The XRD patterns before and after the reaction didn't change significantly, indicating the BiVO 4 -FTO sample still retains the crystal structure ( Figure  S16, Supporting Information). However, the ICP-OES measurement prove that vanadium ions are dissolved as the photocatalytic reaction proceeds ( Figure S17, Supporting Information), indicating the decrease in stability is mainly due to the photoinduced dissolution of V 5+ ions, leading to the severe photocorrosion, which is consistent with the previously report. [35,36] The stability of BiVO 4 has been a hot and challenging issue for a long time, further exploration of strategies to improve the photostability of tetragonal BiVO 4 is the focus of our follow-up research. The wavelength-dependent apparent quantum efficiency (AQE) of BiVO 4 -FTO sample is depicted in Figure 5c. The AQE values match well with the diffuse reflectance spectrum, demonstrating that the overall water splitting is caused by the absorption of incident light. To determine the oxygen source of the reaction product, gas chromatography-mass spectrometer (GC-MS) test was applied to detect 16 O 2 and 18 O 2 . As shown in Figure 5d, the retention time of 16 Figure 5d) further shows that the main product with H 2 16 O as the oxygen source is 16 O 2 (m/z = 32), while the main product with H 2 18 O as the oxygen source is 18 O 2 (m/z = 36). Therefore, it is clear that the photocatalytic reaction product originates from the H 2 O splitting. To verify the effect of the crystal planes on photocatalytic hydrogen production, strong alkali was used to treat BiVO 4 -FTO film. It can be seen from Figure S18 (Supporting Information) that the etched BiVO 4 -FTO retain the crystal structure and chemical composition basically but the original smooth crystal planes are corroded ( Figure S19, Supporting Information). The etched sample are collected for photocatalytic hydrogen test and the activity is greatly reduced ( Figure S20, Supporting Information), revealing that crystal planes are beneficial to the separation of carriers. Moreover, the tetragonal BiVO 4 -FTO film with different redox crystal facets exposure ratio are provided in Figure S21 (Supporting Information), which are obtained by adjusting the pH of precursor. By comparing the water splitting activity in Figure S22 (Supporting Information), it is found that under the same loading of cocatalysts, the activity of BiVO 4 -FTO (pH = 2) is higher than those of BiVO 4 -FTO (pH = 1.6) and BiVO 4 -FTO (pH = 3), indicating that the properly exposed redox crystal facets are beneficial to the photocatalytic water splitting reaction. Surface photovoltage (SPV) measurements ( Figure S23, Supporting Information) indicate that BiVO 4 -FTO sample has higher photovoltage than BiVO 4 powder, confirming the effective separation and transportation of photogenerated carriers in BiVO 4 -FTO sample. In addition, BiVO 4 -FTO film with different thicknesses can be obtained by adjusting the hydrothermal time ( Figures S24 and S25, Supporting Information). It can be seen that as the synthesis time increases from 3 to 24 h, the hydrogen evolution rate gradually decreases ( Figure S26a, Supporting Information), which is associated with the reduction of lattice strain. In addition, after further loading the cocatalysts, the gas evolution rates are significantly increased ( Figure S26b, Supporting Information), and the best gas release rate is achieved at the synthesis time of 12 h. Furthermore, to prevent the presence of incompletely consumed organic reactants in the product from affecting the source of hydrogen, infrared analysis is carried out, and no obvious organic peaks of reactant is detected ( Figure  S27a, Supporting Information). Moreover, it can be seen from the TG analysis ( Figure S27b, Supporting Information) that from ambient temperature to 600°C, the BiVO 4 -FTO sample shows basically no change when heated in air atmosphere. The slight weight loss of BiVO 4 powder can be attributed to the evaporation of adsorbed water on the surface. The above results indicate that the tetragonal BiVO 4 has good thermodynamic stability.
Here, the single-particle photoluminescence (PL) spectra and PL decay kinetics tests were used to explore the role of cocatalysts in photocatalytic overall water splitting reaction. Compared with the ordinary photoluminescence spectra, the single particle photoluminescence has less noise and interference from other molecules or particles, so the role of cocatalyst can be studied from the perspective of single particle. PL images of single BiVO 4 -FTO sample and BiVO 4 -FTO sample with cocatalysts are depicted in Figure 6a,b. Compared with Figure 6a, the blue area in the center of Figure 6b represents the shorter average life time, which means that after loading cocatalysts, the photogenerated carriers are quickly captured by the cocatalysts and participate in the water splitting reaction. As shown in Figure 6c, the significant decreased intensity of PL peak is observed, indicating the reduced recombination of photogenerated carriers on BiVO 4 -FTO sample with cocatalysts. Further, according to the three-exponential fitting (a 1 ·exp( / 1 ) + a 2 ·exp( / 2 ) + a 3 ·exp( / 3 )) shown in Figure 6d, it is determined that the lifetime of BiVO 4 -FTO sample with cocatalysts (≈4.910 ± 0.022 ns) is shorter than that of BiVO 4 -FTO sample (≈5.490 ± 0.069 ns), demonstrating the carriers are separated and transferred more quickly after loading cocatalysts. [37] The schematic diagram of photocatalytic overall water splitting on tetragonal zircon BiVO 4 loaded with cocatalysts is shown in Figure 6e. Under the visible-light excitation, photogenerated electrons and holes are generated at the CBM and VBM of tetragonal BiVO 4 , which then migrate to the H 2evolution cocatalyst Rh/Cr 2 O 3 and O 2 -evolution cocatalyst MnO x to participate the surface redox reaction, respectively. This spatially separated structure is beneficial to inhibit the recombination of photogenerated carriers and promote photocatalytic overall water splitting reaction.

Conclusion
In summary, tetragonal BiVO 4 on FTO substrate was prepared and overall water splitting under visible light irradiation was evaluated. The BiVO 4 grown on substrate has better catalytic performance than the naturally nucleated BiVO 4 . By introducing strain, the energy band position could be effectively adjusted, so that conduction band position shifts up and exceeds the hydrogen evolution potential. The common exposure of oxidizing crystal planes and reducing crystal planes increase the separation efficiency of photogenerated carriers and promote photocatalytic reaction. Considering the s-m BiVO 4 has better light absorption and oxygen generation kinetics, the works on monoclinic BiVO 4 and other substrate are in procedure. The new initiative is put forward to adjust the band structure of photocatalyst for overall water splitting under visible light irradiation.

Preparation of BiVO 4 -FTO Film
All chemicals are of analytical grade and do not require further purification when used. The BiVO 4 -FTO film was grown by hydrothermal method. Typically, the precursor Bi (NO 3 ) 3 ·5 H 2 O (1 mmol) was dissolved in HNO 3 solution (20 mL, 2 m), and C 10 H 14 N 2 Na 2 O 8 (1 mmol) was added and stirred for 30 min to obtain solution A. At the same time, NH 4 VO 3 (1 mmol) was dissolved in NaOH solution (40 mL, 1 m), and then C 10 H 14 N 2 Na 2 O 8 (1 mmol) was added and stirred for 30 min to obtain solution B. Then A solution was slowly poured into the B solution and the solution became a clear yellow solution. After adjusting the pH to 2, continue to stir the solution for another 30 min and transfer to 100 mL Teflon-lined autoclave. The FTO glass was sonicated sequentially in acetone, ethanol, and isopropanol for 30 min. Then, put the cleaned FTO glass (3 cm × 4 cm) into 100 mL Teflon-lined autoclave at an angle, with the conductive surface facing down. After sealing the autoclave, the hydrothermal reaction was carried out at 160°C for 12 h. The autoclave was naturally cooled to room temperature, and then the FTO glass was taken out, rinsed with deionized water several times, and dried at room temperature. The obtained product was called BiVO 4 -FTO film. To make the test convenience, the scraped sample was called BiVO 4 -FTO sample.

Preparation of BiVO 4 Powder
BiVO 4 powder was synthesized by coprecipitation method. Typically, Bi (NO 3 ) 3 ·5H 2 O (2 mmol) and NH 4 VO 3 (2 mmol) were simultaneously dissolved in 60 mL of deionized water and stirred vigorously for 2 h. The reaction product was filtered, washed with deionized water and ethanol, and dried in an oven at 100°C for 5 h. The obtained product was called BiVO 4 powder.

Computational Details
All the calculations were performed by using the Vienna Ab Initio Simulation Package (VASP). [38] The ion-electron interactions were described by projector augmented wave (PAW) method. [39] The generalized gradient approximation (GGA) in the Perdew-Burke-Ernzerh (PBE) with a cutoff energy of 500 eV for exchange-correlation interactions were adopted. [40] A Monkhorst-Pack 7 × 7 × 1 k-point grid was adopted for all the calculations. The convergence criteria for the force and energy were set to be 0.01 eV Å −1 and 10 −4 eV with a vacuum space larger than 20 Å in the zdirection to avoid interactions between periodic units during the structure relaxation. The magnitude of strain is described by a/a 0 , here a 0 and a denote the lattice parameters of the unstrained and strained systems, respectively. As a result, 10% in-plane tensile strain is applied to (101) plane of BiVO 4 .

Characterization
XRD patterns were conducted on a Bruker AXS D8 diffractometer equipped with Cu K radiation to reveal the crystal structure. Morphologies were investigated by SEM (Hitachi S-4800) equipped with an EDS. The UV-vis DRS analyses were recorded by Shimadzu UV-2550 spectrophotometer using BaSO 4 as reflectance standard to explore the optical absorption. The TEM and HRTEM tests were performed with a JEOL JEM-2100F microscope to analyze the nanostructure and composition of the as-prepared BiVO 4 photocatalyst. The cross-sectional FIB-HRTEM samples were analyzed by focused ion beam (FIB) technology. UPS measurements of the as-prepared samples were performed with He I (21.2 eV) as monochromatic light source and a total instrumental energy resolution of 100 meV. XPS measurements were carried out on a Thermo ESCALAB 250XI spectrometer with a monochromatic Al-K source to explore the element composition and valence states on the surface, and the binding energies were calibrated by the C 1s peak (284.8 eV). Raman spectra were measured on LabRAM HR800 with laser excitation of 532 nm. The reaction liquid used for gas measurement was analyzed by ICP-OES/AES: Varian (720-ES). The amounts of H 2 and O 2 evolution were analyzed using gas chromatography (GC-7920) equipped with a thermal conductivity detector (Ar as carrier gas). Single particle fluorescence and fluorescence decay measurements were carried out on a confocal microscope instrument. A 375 nm continuous wave (CW) laser was used to excite the interband transition of BiVO 4 and pulse frequency was 5 MHz. To detect the source of oxygen, GCMS-QP2010 was used to detect oxygen isotope. The software used for XRD data refinement is TOPAS academic.

Photocatalytic Reactions
The photocatalytic reactions were carried out in a Pyrex reaction cell connected to a closed gas circulation and evacuation system. Typically, the area of the BiVO 4 -FTO film is 12 cm 2 , on which the mass is estimated to be ≈5 mg. Generally, 10 mg photocatalyst was immersed in 100 mL of reaction solution. The photocatalytic HER, OER, and overall water splitting reaction were carried out in 0.02 × 10 −3 m H 2 O 2 , 0.02 m AgNO 3 and pure water, respectively. The theoretical loading amounts were 2, 2, and 0.5 wt% for Rh, Cr, and MnO x . For the overall water splitting reaction, after the catalyst was ultrasonically dissolved, the cocatalyst precursor was sequentially added to the reaction solution to carry out the photodeposition reaction. First, a specific amount of RhCl 3 ·3H 2 O solution (2 mg mL −1 ) was added to the reaction system, and irradiated for 10 min. Subsequently, the K 2 CrO 4 (2 mg mL −1 ) and Mn(NO 3 ) 2 (0.1 × 10 −3 m) solution were added to the suspension and irradiated for another 5 min, respectively. The system was evacuated for 30 min to ensure complete removal of air, and then illuminated from the top surface with a 300 W Xe lamp (PLS-SXE300D, Beijing PerfectLight Technology Co., Ltd.) equipped with a 400 nm cut-off filter ( > 400 nm). A cooling water stream was used to maintain the reaction suspension at 288 K. The separated gas was analyzed by gas chromatography (GC-7290, TCD with Ar as a carrier gas). The photocatalytic stability test was performed every 4 h as a cycle. After each independent cycle, the photocatalyst was recycled by centrifugation and redispersed in new solution. The wavelength dependence of the AQE was tested under the same photocatalytic reaction conditions, through the 365, 420, 450, 500, 530, or 700 nm bandpass filters and the masked area of 1 cm 2 . The photon flux of the incident light was determined using a PL-MW2000 spectrophotometer (PerfectLight, China). The AQE was calculated from the ratio of the number of electrons reacted to the number of incident photons during the water splitting process. The calculation formula is AQE = 2 × the number of evolved hydrogen molecules the number of incident photons × 100% (1) The STH efficiency was measured under simulated sunlight (AM 1.5G, 1 cm 2 irradiation area). The calculation formula is STH = output energy as H 2 energy of incident solar light (2)

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.