Tailored BiVO4 Photoanode Hydrophobic Microenvironment Enables Water Oxidative H2O2 Accumulation

Abstract Direct photoelectrochemical 2‐electron water oxidation to renewable H2O2 production on an anode increases the value of solar water splitting. BiVO4 has a theoretical thermodynamic activity trend toward highly selective water oxidation H2O2 formation, but the challenges of competing 4‐electron O2 evolution and H2O2 decomposition reaction need to overcome. The influence of surface microenvironment has never been considered as a possible activity loss factor in the BiVO4‐based system. Herein, it is theoretically and experimentally demonstrated that the situ confined O2, where coating BiVO4 with hydrophobic polymers, can regulate the thermodynamic activity aiming for water oxidation H2O2. Also, the hydrophobicity is responsible for the H2O2 production and decomposition process kinetically. Therefore, after the addition of hydrophobic polytetrafluoroethylene on BiVO4 surface, it achieves an average Faradaic efficiency (FE) of 81.6% in a wide applied bias region (0.6–2.1 V vs RHE) with the best FE of 85%, which is 4‐time higher than BiVO4 photoanode. The accumulated H2O2 concentration can reach 150 µm at 1.23 V versus RHE under AM 1.5 illumination in 2 h. This concept of modifying the catalyst surface microenvironment via stable polymers provides a new approach to tune the multiple‐electrons competitive reactions in aqueous solution.


Introduction
Photoelectrochemical (PEC) water splitting to H 2 and O 2 has been extensively investigated over five decades since the great discovery by Fujishima and Honda in 1972. [1] However, the reduction reaction efficiency of H 2 O to H 2 is limited by the oxidation reaction of H 2 O to O 2 due to its sluggish kinetic process with four steps of proton-coupled electron transfer. [2] Therefore, it's imperative to explore some alternative oxidative reactions, which not only efficiently match the reduction reaction, but also the production is valueadded. PEC water oxidative H 2 O 2 production substituting the O 2 evolution reaction (OER), by coupling with cathodic H 2 generation of solar water splitting, has been gradually considered as a promising approach because of: i) faster kinetic process of 2-electron transfer than OER of 4electron transfer; ii) more easily separation of gas-liquid products; iii) higher economic value H 2 O 2 products. [3] However, PEC water oxidative H 2 O 2 reaction also faces great challenges, which mainly need to overcome a thermodynamically favorable OER and H 2 O 2 decomposition reaction, as shown in Equations (1)-(3). [3d,4] The 2-electron water oxidation reaction (WOR) for H 2 O 2 production has been examined over many metal oxides, such as ZnO, [5] Bi 2 WO 6 , [6] TiO 2 , [7] BiVO 4 , [7,8] etc. Among these oxides, BiVO 4 is the most promising photoanode for 2-electron WOR due to its suitable bandgap for light-harvesting, deep valance band edge, and favorable thermodynamic activity trend for H 2 O 2 evolution. For this reaction, many strategies including surface passivation, [9] heterojunction construction, [10] heteroatom doping, [8b,11] crystal surface regulation, [12] etc. have been devoted to develop the highly active BiVO 4 -based photoanode. Such strategies mainly focus on tailoring the BiVO 4 surface electronic property, while the PEC water splitting reaction is still www.advancedsciencenews.com www.advancedscience.com carried out at an inherent hydrophilic solid-liquid interface, which will cause the in situ produced H 2 O 2 to be decomposed before desorption in long-term operation. [3d,4a] Therefore, the hydrophobic gas-liquid-solid three-phase interface is desired to enhance the kinetic of H 2 O 2 desorption to make the H 2 O 2 separation before the reaction reaches equilibrium. In this case, creating an appropriate hydrophobic electrode surface is a promising strategy to kinetically regulate the O 2 /H 2 O 2 production formation rate, thereby enhancing the H 2 O 2 selectivity and accumulation amount.
The hydrophobicity of electrode surface was previously tuned through polymer modification. [13] Other than that, the polymer can also accumulate the gas molecules, which can further interact with the catalysts' surface and thus effectively regulate the intermediate binding energy strength for completely different reaction paths. [14] Recently, Xia et al. [14] demonstrated that the selectivity of water oxidation could be altered from 4-electron WOR for O 2 product to 2-electron WOR for H 2 O 2 evolution by coating the electrocatalysts (C, Ni) with a hydrophobic polymer polytetrafluoroethylene (PTFE). It was suggested that the excellent H 2 O 2 selectivity attributed to the shift of the *OH intermediate binding energy through a less oxidized catalyst surface caused by the locally confined O 2 gas on PTFE surface. This catalytic concept can also be applied in photo-electrochemical field. Combined with the BiVO 4 catalyst, the confined O 2 gas molecules on polymer surface have no effect on its oxidation degree, while the possibility of their effect on the reaction intermediate species for tuning its binding energy strength need to be done to experimentally prove and support these hypotheses.
Herein, we present a hydrophobic gas-liquid-solid three-phase interface to build a tailored catalytic microenvironment to regulate the PEC water oxidation reaction pathway in thermodynamics and kinetics. We show that the hydrophobic polymer can not only in-situ confine O 2 gas close to active sites to shift the *OH intermediate thermodynamically for H 2 O 2 evolution, but also help to tune the release rate of the gas/liquid product dynamically for H 2 O 2 generation and accumulation. Using BiVO 4 as the model system, it is found that O 2 evolution from 4-electron transfer and H 2 O 2 decomposition can be significantly inhibited after the PTFE coating, achieving 4-times higher H 2 O 2 Faraday efficiency (FE) of 85% in a widely applied bias range (0.6-2.1 V vs RHE). The PEC water oxidative H 2 O 2 performance is further investigated by in situ Raman spectra, which intuitively prove that the *OH intermediates play a decisive role in promoting the PEC water oxidative H 2 O 2 activity.

Results and Discussion
The synthesis route of the aerophilic-hydrophobic PTFE/BVO photoanode is displayed in Figure S1, Supporting Information, where the BiVO 4 electrode was prepared by electrodeposition based on a previous report. [15] The PTFE, as one of the most hydrophobic materials, [16] is chosen as the overlayer framework for allowing the assemblage of the gaseous product O 2 of water on the catalyst surface in aqueous photoelectrocatalysis. The confined O 2 helps to influence the reaction intermediates to thermodynamically regulate the reaction path of PEC water oxidation. Also the hydrophobicity favors to tune the adsorption or desorption capacity of the gaseous O 2 and liquid H 2 O 2 product in water oxidation reaction (Figure 1a). In the work, we investigated the BiVO 4 coated by PTFE with different content toward PEC H 2 O oxidative H 2 O 2 evolution. In addition, we also prepared other hydrophobic material coated on BiVO 4 photoanode to further confirm the catalytic microenvironmental regulation strategy.
The successful loading of PTFE on BiVO 4 surface is confirmed by a series of systematic characterizations. As presented by XRD in Figure 1b, the peaks with "*" correspond to the PTFE diffraction peaks in PTFE/BVO photoanodes, and the other peaks are indexed to the monoclinic phase BiVO 4 . [17] The PTFE diffraction peaks are displayed in Figure S2, Supporting Information. The XPS spectra in Figure 1c,d display an additional peak pertaining to -CF x deriving from the PFTE located at 292.5 eV in the C1s and 689.7 eV in the F 1s of the PTFE-coated sample. [18] A significant shift to higher binding energies for Bi 4f, V 2p, and O 1s XPS spectra peaks after coating with PTFE indicates a decrease in electron density in metal sites of the bulk phase ( Figure S3, Supporting Information), which can be ascribe to the electronwithdrawing character of PTFE. Additionally, the FT-IR measurements in Figure 1e show that two significant peaks appear at 500-650 and 1150-1250 cm −1 ascribing to -CF 2 and -CF 3 functional group, further indicating the successful coating of PTFE. [16,19] Impressively, no peak shifts of BiVO 4 and PTFE/BVO composite samples are observed in the XRD, FT-IR, and Raman ( Figure S4, Supporting Information), which confirms the unchanged monoclinic BiVO 4 after the physical adsorption of PTFE.
The top (Figure 2a) and cross-section (inset in Figure 2a) views of microstructures of the 10PTFE/BVO electrode demonstrate the similar morphology to pure BiVO 4 ( Figure S5, Supporting Information) because of a very thin filiform PTFE film layer on the BiVO 4 surface. With the increase of PTFE content in PTFE/BVO composites, the filaments can be seen obviously ( Figure S5 (Figure 2f), illustrating that the change in electron structure of BiVO 4 caused by PTFE polymer has little effect on its band gap. In addition, the aerophilic heterogeneous interface affecting the adsorption and desorption ability of the gas-liquid products should also be considered. [20] With the addition of PTFE, the PTFE/BiVO 4 samples have good hydrophobicity, in which the water angle varies from 31.0°to 122.9°, as shown in Figure S6, Supporting Information. Visually, the O 2 gas in aqueous solution can be certified by the locally confined O 2 bubbles for 10PTFE/BVO (Movie S1, Supporting Information) and the rapid release for pristine BiVO 4 in the PEC process (Movie S2, Supporting Information). The concentration of the O 2 bubbles on the photoelectric-catalyst surface can be clearly seen in Figure S7, Supporting Information. The surface hydrophobicity would have a significant role in H 2 O 2 release kinetics, which is expected to tune the liquid-H 2 O 2 and gas-O 2 product ratios in PEC water oxidation reaction. Significantly, proper free energy of the *OH is generally desired for  thermodynamically favorable H 2 O 2 evolution in the water oxidative H 2 O 2 reaction process: too strong an OH binding would further oxidize *OH to *OOH, completing the 4-electron WOR process for O 2 ; too weak a binding would be favorable to release OH into the solution as hydroxyl radical product (·OH). [4a,14] For PTFE/BVO system, the confined O 2 molecules with the potential to donate electrons might have a significant regulatory effect on the nearby *OH and then control the *OH binding energy to a suitable position, which will be justified by the following experimental and computational part.
The PEC H 2 O 2 generation was investigated in 1 m NaHCO 3 electrolyte with a pH value of 8.3 under AM 1.5 illumination (100 mW cm −2 ), and the H 2 O 2 product was quantified by using N,N-diethyl-1,4-phenylene-diamine (DPD) method. As shown in Figure 3a, the overall current density gradually decreases with the increasing PTFE coated, since the dielectric PTFE repels water and results in a lower electrochemical surface area ( Figure  S8, Supporting Information). An oxidation peak at 0.25-0.5 V versus RHE belongs to the reoxidation of V species in BiVO 4 and the details are given in Figure S9, Supporting Information. However, the PTFE overlayer plays a positive role for the PEC H 2 O 2 selectivity after the illumination. The corresponding realtime Faraday efficiencies (FEs) for water oxidative H 2 O 2 production under applied bias ranging from 0.6 to 2.1 V versus RHE are investigated in Figure 3b. In the PEC reaction process, the real-time FEs of BiVO 4 and PTFE/BVO photoanodes maintain at a steady stage and the PTFE overlayers indeed dramatically increase the FEs of BiVO 4 . The average FEs of H 2 O 2 production for pristine BiVO 4 was 21.8%, but increased to 37.8%, 81.6%, 73.8%, and 59.2% with the increasing PTFE content, respectively. The 10PTFE/BVO photoanode exhibits a maximal 4-fold improvement in H 2 O 2 selectivity with the best FE of 85% compared to pristine BiVO 4 . In addition, with the applied bias voltage increased, the enhanced real-time FEs of PTFE overlayer suddenly drop and the performance of the PTFE hydrophobic layer to enhance the FEs of the BiVO 4 photoanode disappears, which could be ascribed to the lower dark current density where the EC reactions work. The results elucidate that the PTFE coating on the BiVO 4 surface acts a crucial role to increase the water oxidative H 2 O 2 selectivity.
The changes in the free energy of the *OH intermediate (ΔG *OH ) on PTFE/BVO catalysts can be predicted by the plausible effect of the O 2 . A suitable ΔG *OH indicates a shift from the 4-electrons route to 2-electrons pathway. [21] To unveil the underlying mechanism, the volcano plot for 2-electron WOR as a function of the ΔG *OH over BiVO 4 covered with O 2 bubbles is shown in Figure 3c, and these thermodynamic models are also given in Figures S10, S11, Supporting Information. It is found that the number of O 2 play an important role on the *OH binding, which shift the active sites towards the activity peak of the volcano plot. When the Bi sites are surrounded by 4 O 2 , a steady state with the most appropriate ΔG *OH and the highest H 2 O 2 activity can be obtained. To simulating a real reaction environment, an explicit approach in aqueous solution is also conducted (Figure 3c). We found that the *OH binding is also affected by O 2 molecules, further verifying that an appropriate amount of O 2 molecules is imperative to regulate the adsorption of *OH intermediate on the BiVO 4 catalyst and then produce a higher H 2 O 2 in the real H 2 O system. Simultaneously, the charge redistribution of BiVO 4 adsorbed by one OH intermediate caused by O 2 molecules can be analyzed through the charge density difference in Figure 3d.
In the presence of adequate O 2 , the Bi-O bond length between *OH and Bi site is 2.120 Å, slightly larger than that without O 2 (2.108 Å). The weakened Bi-O ionic bond is ascribed to the electrons transfer from the nearby O 2 molecules to *OH, then induce the ΔG *OH to adjust to a more suitable position (at the bottom of the volcano diagram). Although the thermodynamic analysis can only be taken as qualitative, it is a first step toward understanding the trend of the selectivity changes on the modified photoelectrocatalyst by locally confined O 2 molecules. The local O 2 confinement approach can be extended to other hydrophobic polymers. After silane modification (PDMS/BVO), the H 2 O 2 selectivity of BiVO 4 photoanode is also greatly promoted, confirming that the confined O 2 plays a significant role in enhancing H 2 O 2 selectivity ( Figure S12, Supporting Information). Other thermodynamic factors involving band structure and band bending, determined by valence band (VB) XPS spectra and Mott-Schottky plots, are also investigated ( Figure S13, Supporting Information). The similar VB edges and flat-band potentials of BiVO 4 and 10PTFE/BVO suggest that the two factors controlling the thermodynamic products can be excluded. Moreover, the interface charge transport and separation efficiencies of BiVO 4 and 10PTFE/BVO do not show a big difference, which is obtained by the photocurrent density measured in hole scavenger ( Figure S14, Supporting Information) The PEC water oxidation products performance is not only determined by the thermodynamic effects, but also can be associated with the kinetics-controlled processes of electron transfer. As such, the kinetics-controlled processes involve the two competing reaction, 4-electron WOR of O 2 production and stepwise 2-electron (H 2 O 2 formation)/2-electron transfer (H 2 O 2 decomposition). The surface hydrophobicity might be the reason behind the kinetics-controlled processes for selective H 2 O 2 production. The samples with hydrophobic property ( > 90°) hold higher FEs of water oxidative H 2 O 2 production compared with the wettability ( < 90°), which can be obviously seen from the plots of the FEs of water oxidative H 2 O 2 production at 1.23 V versus RHE versus different surface hydrophily-hydrophobicity in Figure 4a. It suggests that the surface hydrophobicity, especially an appropriate photoanode hydrophobic surface, is conductive to the generation of H 2 O 2 . The local hydrophobic-aerophilic environment makes the release of O 2 bubbles slowly and promotes the H 2 O 2 leaving from the PTFE/BVO surface before the competitive reactions reach equilibrium, thereby kinetically increasing H 2 O 2 production. The kinetics processes is preliminarily investigated by the ring-disk electrode technology, in which the reaction on the disk and ring electrode is H 2 O 2 production by BiVO 4 and H 2 O 2 decomposition by BiVO 4 and PTFE/BVO, respectively (Figure S15, Supporting Information). Here, the applied bias for the ring electrode is fixed at 0.7 V versus RHE, at which the only possible EC reaction is H 2 O 2 decomposition and no other side reactions occur. Figure 4b shows the J-V curves of BiVO 4 and PTFE/BVO electrodes recorded at a scan rate of 10 mV s −1 in 1 m NaHCO 3 electrolyte solution at ambient temperature at 1600 rpm. Under the same disk current generated by water oxidative H 2 O 2 production, the ring currents come from H 2 O 2 oxidation can reflect the effect of surface hydrophobicity. [22] Remarkably, a large disk current for the water oxidative H 2 O 2 evolution is observed, while a sharp drop ring current for the oxidative H 2 O 2 decom-position in PTFE/BVO system. The negligible H 2 O 2 decomposition on PTFE/BVO photoanode indicates that the hydrophobicity surface reduces the contact possibility between H 2 O 2 and electrode to a great extent, then reducing the reoxidation and increasing the release of H 2 O 2 . In addition, the photocatalytic H 2 O 2 decomposition by BiVO 4 and PTFE/BVO photoanodes is also conducted in the HCO 3 − -containing solution under AM 1.5 illumination. As shown in Figure S16, Supporting Information, the H 2 O 2 concentration decreases rapidly and is almost undetectable within 60 min for the illuminated BiVO 4 photoanode, while the retention rate of H 2 O 2 is 85% for PTFE/BVO photoanodes during irradiant 60 min, which is basically equivalent to the optical H 2 O 2 decomposition. Overall, the specific microenvironment indeed promotes the water oxidative H 2 O 2 production, in which right amounts of confined O 2 on BiVO 4 photoanode induced by PTFE overlayer thermodynamically regulate the only intermediate *OH and maintain the thermodynamic favorable water oxidative H 2 O 2 process, meanwhile the surface hydrophobicity of photoelectrode kinetically facilitates the H 2 O 2 desorption in aqueous solution.
When the surface accumulated holes are located on BiVO 4 surface, the photocorrosion of BiVO 4 is about to initiate, then these ions (V 5+ , Bi 3+ , O 2− ) may destabilize the BiVO 4 lattice and increase the solubility of BiVO 4 at the surface. Since V 5+ is more soluble than Bi 3+ in alkaline electrolyte, the dissolution of V 5+ will be dominant. [23] The PEC water oxidative H 2 O 2 stability over BiVO 4 and 10PTFE/BVO photoanodes were conducted at 1.23 V versus RHE in 1 m NaHCO 3 electrolyte, as shown in Figure 4c, from which it can be seen that the 10PTFE/BVO photoanode www.advancedsciencenews.com www.advancedscience.com maintains a relatively stable J-t curve in comparison with pristine BiVO 4 for H 2 O 2 production after 5 h illumination. It suggests that the PTFE protection surface can slow down the dissolution of V 5+ of BiVO 4 , and then weaken its photocorrosion due to the surface holes accumulation. [12] The above further verifies that the faster kinetics of 2-electron WOR in 10PTFE/BVO photoanode can reduce surface holes accumulation, thereby suppressing the dissolution of V 5+ . However, the accumulation of H 2 O 2 produced by both BiVO 4 and 10PTFE/BVO photoanode displays a nonlinear increase as the reaction time lengthens, indicating not a first-order dynamic process governed by electronhole transfer. As shown in Figure 4d, the H 2 O 2 accumulation can be divided into two stages, in which the first stage increases rapidly, followed by a gradual decreases in the second stage. Since the H 2 O 2 accumulated on the surface is easily re-oxidized by photoinduced holes, which will be competed with the produced H 2 O 2 , and finally reaching dynamic equilibrium. Therefore, the two H 2 O 2 accumulation process can be divided into 2-electron/4electron WOR competing reaction and stepwise 2-electron/2electron transfer of H 2 O 2 decomposition. In the first stage, the H 2 O 2 generation rate for 10PTFE/BVO can reach 14.4 μmol h −1 with an accumulation amount of 92 μmol L −1 , while the H 2 O 2 evolution rate for BiVO 4 is only 5.4 μmol h −1 with an accumulation amount of 38 μmol L −1 . During the second stage, despite the 10PTFE/BVO holds a slow H 2 O 2 evolution rate of 1.7 μmol h −1 accompany with H 2 O 2 decomposition, the H 2 O 2 accumulation amount continues increase, reaching 150 μmol L −1 after 2 h of illumination. The above results are fully illustrated that the PTFE layer can efficiently promote the H 2 O 2 generation and inhibit the decomposition of H 2 O 2 . Finally, the stable photocurrent density and constant FE produce a higher average solar to H 2 O 2 efficiency of 2.02% for 10PTFE/BVO compared to 0.23% for pure BiVO 4 . Meanwhile, the obtained H 2 amounts in the PEC water splitting over 10PTFE/BVO photoanode quantities as a function of the theoretical electron number calculated based on its photocurrent density ( Figure S17, Supporting Information), confirming that H 2 production rate is not affected by the H 2 O 2 decomposition. Compared to previously reported strategies, such as passivation, [3b] heterojunction, [3c] doping, [3d] crystal facets, [4a] tailoring the catalyst microenvironment is a more efficient method to tune the H 2 O 2 selectivity and accumulation, which are summarized in Table S1, Supporting Information. However, more work to prolong the H 2 O 2 accumulation stability and improve the PEC efficiency of BiVO 4 photoanode will be the topic of subsequent research for future possible application. The XRD, SEM, TEM, and XPS results of 10PTFE/BVO photoanode after the long-term H 2 O 2 evolution are shown in Figures S18-S21, Supporting Information. It can be observed that there are no detectable changes in the crystal structure, morphology, and component, confirming a good structure stability for PEC H 2 O 2 evolution.
The specific reaction pathway for the possible water oxidation reaction (4-electron WOR, 2-electron WOR) are as follows: [4a,24] The 4 − electron WOR ( In order to explore the process of producing H 2 O 2 by PEC water oxidation via BiVO 4 photoanode with PTFE hydrophobic layer in 1 m NaHCO 3 electrolyte, in situ Raman spectra were investigated. The electrochemical cell is based on a homemade square teflon disk, in which the opposite electrode and the reference electrode are slender copper foil and Ag/AgCl, respectively. The test was performed at 0-0.6 V versus RHE potentials for 10 min per potential. As shown in Figure 4e,f and Figure S22, Supporting Information, two characteristic Raman peaks at 1336 and 1605 cm −1 attributing to the stretching of HCO 3 − [8a] can be observed in both of the synthesized BiVO 4 and PTFE/BVO photoanodes. In the Raman spectra of the BiVO 4 photoanode, a strong signal at 830 cm −1 deduced as the O-O stretching bond [25] appears, indicating the 4-electron WOR dominates. However, the O-O bond in the PTFE/BVO photoanodes gradually decrease with the increasing applied bias voltage, which suggests that the 4-electron WOR path for O 2 generation is effectively suppressed after the PTFE hydrophobic layer loaded. Meanwhile, a new peak at 1020 cm −1 associating with either OH or superoxide species (·OH) [26] in PTFE/BVO photoanode rises. The EPR spectrum used to characterize ·OH free radicals [27] shows no signal in the 10PTFE/BVO sample ( Figure S23, Supporting Information). The above results indicate that loading the PTFE hydrophobic layer can effectively inhibit the transformation of *OH to *OOH intermediates and accelerate the generation of *OH to H 2 O 2 on BiVO 4 surface, then selectively turn the oxidation pathway of water oxidation from 4electron WOR to 2-electron WOR.

Conclusion
In this work, we have successfully confirmed that the surface microenvironmental modification of BiVO 4 photoanode upon polymer loading can greatly enhance the PEC H 2 O 2 selectivity and accumulation, in which the average FE of H 2 O 2 is increased by a factor of 4 and the saturated H 2 O 2 concentration can reach 150 μm at 1.23 V versus RHE under AM 1.5 illumination for continuous 2 h after the addition of PTFE. From experiments and DFT studies, we can explain this improvement by the two main factors: 1) the energetics of WOR of BiVO 4 photoanode is regulated to favor the 2-electron H 2 O 2 pathway in thermodynamics via the confined local O 2 gas on PFFE surface; 2) the H 2 O 2 desorption is greatly promoted in aqueous solution and H 2 O 2 decomposition is effectively inhibit in kinetics by the PTFE hydrophobicity. Impressively, the resultant PFFE/BVO photoanode displays better photocurrent density stability than that of BiVO 4 . This work provides a triple effect of favorable H 2 O 2 generation, suppressed OER route and H 2 O 2 decomposition via the design of BiVO 4 photoanode surface microenvironment, and the approach www.advancedsciencenews.com www.advancedscience.com has the potential to be applicable to other heterogenous photoelectrochemical system.

Experimental Section
Preparation of PTFE/BVO Photoanodes: BiVO 4 (BVO) photoanodes were prepared according to the method of Lee and Choi's method. [15] The PTFE/BVO composite photoanodes were further constructed by spincoating approach, for which BVO was spun by the same amount of PTFE solutions with different concentrations (5 wt%, 10 wt%, 15 wt%, and 20 wt%) at 1500 rpm for 150 s. Then the above precursors were annealed at 350°C for 30 min in air atmosphere to form PTFE/BiVO 4 electrodes with different proportions, denoted as xPTFE/BVO (x is the concentration of PTFE).
Preparation of PDMS/BVO Photoanodes: The BiVO 4 electrode was soaked in 9 mL tetrahydrofuran (THF) solution containing 3 g polydimethylsiloxane (PDMS) and stirred gently for 24 h. Then the above electrode precursor was illuminated with UV-light and the light intensity was controlled at 10 mW cm −2 . After illuminated for 2 h, the resulting PDMS/BVO photoanode was washed with THF and deionized water and dried at room temperature.
Characterizations: The structure and chemical properties of the assynthesized samples was confirmed with X-ray diffraction (XRD, Cu K , Rigaku SmartLab), field emission scanning electron microscope (SEM, Hitachi S-8200), transmission electron microscope (TEM) and energy dispersive X-ray spectroscopy (EDX) (JEOL JEM-ARM 200F), Raman microscope (HORIBA Xplora Plus) using a 50× objective excited by 532 nm laser light with a power of 20 mW, UV-vis diffuse reflectance spectra (DRS, Shimadzu UV-2600), X-ray photoelectron spectroscopy (XPS, Al K , Escalab 250Xi) and all the peaks calibrated with a C 1s spectrum at a binding energy of 284.6 V, electron paramagnetic resonance spectra (EPR, Bruker EMX-10/12-type spectrometer) in 1 m NaHCO 3 electrolytes under Xe lamp irradiation with a 420 nm cutoff filter, the contact angles (JC2000D1) with a NaHCO 3 droplet on the samples. The ring-disk current was carried out in N 2 -saturated 1 m NaHCO 3 solution at 1600 rpm with a scan rate of 10 m s −1 . The disk electrode was BiVO 4 with a loading mass of 0.24 mg cm −1 , and the ring electrodes were BiVO 4 , xPTFE/BVO (x = 5, 10, 15, and 20) with a loading mass of 0.1 mg cm −1 . The electrochemical cell was based on a self-made square Teflon dish. The counter electrode and the reference electrode were a slender Pt foil and an Ag/AgCl.
Photoelectrochemical Measurements: The photoelectrochemical (PEC) performance was evaluated in an H-type three-electrode quartz electrolytic cell with Ag/AgCl as reference electrode, Pt as counter electrode, Nafion film as the ion exchange membrane, and 1 m NaHCO 3 solution (pH = 8.3) as electrolyte using a CHI 660E electrochemical workstation at room temperature of 25°C. The illumination source was a 300 W Xe arc lamp with an AM 1.5G filter (100 mW cm −2 , CEL-HXF300) and all electrodes were illuminated from the back-side. All illuminated areas were 1.5 cm 2 . Linear sweep voltammetry (LSV) was monitored while sweeping the potential in the positive direction with a scan rate of 10 mV s −1 . The potential versus Ag/AgCl reference electrode was converted to the potential versus RHE according to the Nernst equation: E (vs RHE) = E (vs Ag/AgCl) + 0.0591 × pH +0.197. Flat-band potential measurements were measured using Mott-Schottky plots at potentials varying between 0 and 0.6 V with a frequency of 1 KHz.
Production Measurements: The evolution of H 2 O 2 was detected by N,N-diethyl-1,4-phenylenediamine method. The solution of DPD was prepared by dissolving 0.1 g DPD in a 10 mL of 0.05 m sulfuric acid solution. The POD solution was made by dissolving 10 mg POD in 10 mL deionized water and stored in the refrigerator for use. Potassium phosphate buffer was prepared by mixing 49.85 mL deionized water, 43.85 mL of 1 m potassium dihydrogen phosphate and 6.3 mL of 1 m potassium phosphate. After the photoelectric reaction, 1 mL of the reacted solution was put into a test tube containing 0.4 mL potassium phosphate buffer, 3 mL water, 0.05 mL DPD, and 0.05 mL POD and shaken for 2 min. The obtained solution was analyzed by UV-vis spectroscopy. The Faradaic efficiencies of H 2 O 2 were calculated as shown in Equation (4) The solar to H 2 O 2 efficiency can be calculated from the J-V curve and FE of H 2 O 2 evolution using the following Equation 5: Herein, no electrocatalysis occurs at 1.23 V versus RHE, the J sc for 10PTFE/BVO was 1.34 mA cm −2 , FE for H 2 O 2 production was 85%, P total was 100 mW cm −2 , the solar to H 2 O 2 efficiency can be calculated to 2.02%.
The H 2 evolution was measured in an on-line automatic closed circulation reactor system (Perfectlight Sci&Tech Co., Ltd., Labsolar-6A) that was connected with an online gas chromatograph (GC 9790, TCD, 5 Å molecular sieve columns and Ar carrier).
Computational method: All calculations were carried out using DFT with the Perdew-Burke-Ernzerhof (PBE) form of the generalized gradient approximation (GGA) functional. [28] The Vienna ab initio simulation package (VASP) was employed. [29] The energy cutoff for plane wave expansions was set to 450 eV, and the energy (converged to 1e −5 eV atom −1 ) and force (converged to −2e −2 eV Å −1 ) were set as the convergence criteria for geometry optimization. A BiVO 4 (111) periodic slab model which contain 48 atoms (Bi 8 V 8 O 32 ) was constructed to simulate the BiVO 4 (111) surface. The Brillouin zones were sampled with the gamma-centered Monkhorst-Pack [30] (2 × 4 × 1) k-points meshes for BiVO 4 (111) surface model. As for the slab model, a vacuum space of 15 Å was added to the slab model to avoid interaction between periodic images. In addition, the DFT-D3 method [31] was included to improve the description of the long-range weak van der Waals (vdW) interaction for all DFT calculations. The free energy correction for adsorbate is calculated by VASPKIT. [32] To investigate the effect of surface adsorbed *O 2 on catalytic activity, we calculated the *O 2 coverage-dependent *OH binding free energies (ΔG *OH ), which is a key intermediate in electrochemical H 2 O 2 . [33] Charge redistribution was defined as Δ = Slab+OH − Slab − OH , where Slab+OH , Slab , and OH denote the charge distribution of the whole adsorption system, BiVO 4 (111) surface with/without the *O 2 coverage and *OH.

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