Impact of Postprocessing Approaches and Interface Cocatalysts Regulation on Photocatalytic Hydrogen Evolution of Protonic Titanate Derived TiO2 Nanostructures

TiO2–based photocatalysis system for splitting water into hydrogen offers a sustainable and green technology to produce clean hydrogen energy. However, pristine TiO2 still exists inherent shortcomings restricting its practical applications. Herein, the impact of postprocessing approaches of protonic titanate on engineering of oxygen vacancy and photocatalytic hydrogen evolution of TiO2−x is studied. Subsequently, interfacial cocatalysts are successfully involved in the optimized TiO2−x for enhanced photocatalytic hydrogen evolution. TiO2−x with the highest photocatalytic hydrogen evolution performance of 3112.09 μmol g−1 h−1, denoted as TiO2–C, is selected to adjust the interface with Cu and MoS2 respectively. Cu–TiO2–C and MoS2–TiO2–C composites are synthesized to enhance the separation ability of photogenerated electron‐hole pairs and significantly improve the photocatalytic hydrogen evolution performance. The photocatalytic hydrogen evolution rates of 5 wt% Cu–TiO2–C and 40 wt% MoS2–TiO2–C are 9225.75 and 5765.48 μmol g−1 h−1, respectively. It is proved that different postprocessing methods can tune the content of oxygen vacancy in TiO2−x and regulate the photocatalytic hydrogen evolution performance of TiO2−x materials. The interface regulation of the cocatalyst also contributes to the separation of photogenerated electron‐hole pairs and serves as active sites to enhance hydrogen evolution performance.

DOI: 10.1002/aesr.202300002 TiO 2 -based photocatalysis system for splitting water into hydrogen offers a sustainable and green technology to produce clean hydrogen energy. However, pristine TiO 2 still exists inherent shortcomings restricting its practical applications. Herein, the impact of postprocessing approaches of protonic titanate on engineering of oxygen vacancy and photocatalytic hydrogen evolution of TiO 2Àx is studied. Subsequently, interfacial cocatalysts are successfully involved in the optimized TiO 2Àx for enhanced photocatalytic hydrogen evolution. TiO 2Àx with the highest photocatalytic hydrogen evolution performance of 3112.09 μmol g À1 h À1 , denoted as TiO 2 -C, is selected to adjust the interface with Cu and MoS 2 respectively. Cu-TiO 2 -C and MoS 2 -TiO 2 -C composites are synthesized to enhance the separation ability of photogenerated electron-hole pairs and significantly improve the photocatalytic hydrogen evolution performance. The photocatalytic hydrogen evolution rates of 5 wt% Cu-TiO 2 -C and 40 wt% MoS 2 -TiO 2 -C are 9225.75 and 5765.48 μmol g À1 h À1 , respectively. It is proved that different postprocessing methods can tune the content of oxygen vacancy in TiO 2Àx and regulate the photocatalytic hydrogen evolution performance of TiO 2Àx materials. The interface regulation of the cocatalyst also contributes to the separation of photogenerated electron-hole pairs and serves as active sites to enhance hydrogen evolution performance.
recombination sites of photogenerated electrons and holes, which will adversely affect the photocatalysis reaction.
Supporting cocatalysts on semiconductor materials is one of the promising interface regulations to enhance their photocatalytic activity. The commonly-used cocatalysts are noble metals, [24][25][26][27] which are favorable for photocatalytic hydrogen production due to their high work function, exchange current density, and low hydrogen adsorption free energy. However, the high cost of noble metals limits their application in photocatalysis. Copper, as the most common nonnoble metal cocatalyst, has been widely studied. Copper has a high work function (4.65 eV), [28][29][30] and it can promote the directional aggregation of photogenerated electrons to the surface of the semiconductor. The free energy of hydrogen adsorption (ΔG H ) of Cu is close to 0, which is conducive to the adsorption and desorption of hydrogen, and the occurrence of hydrogen evolution by photocatalytic water splitting. [31][32][33] Molybdenum disulfide (MoS 2 ) has the advantages of low cost, abundant reserves, and easy structure control. [34] More importantly, the ΔG H of MoS 2 edge sites is only 0.08 eV, even better than noble metal Pt. [35,36] In this regard, MoS 2 is considered a promising alternative for noble metal Pt in photocatalytic water splitting for hydrogen evolution. In photocatalysis, it can be used as a cocatalyst, which can provide active sites at the edge, [37] promote the interfacial transport of photoinduced electrons, reduce the overpotential, and improve the photocatalytic hydrogen production activity. [38,39] Herein, four different TiO 2Àx materials are synthesized with the same protonic titanate as the precursor through different postprocessing methods. In the posttreatment process, oxygen vacancy content is controlled by tuning the ratio of water to diethylene glycol (DEG) in the solvothermal process. Then nano-TiO 2Àx material with the best activity is obtained by photocatalytic hydrogen evolution test. On this basis, to improve the performance of photocatalytic hydrogen evolution, Cu and MoS 2 nanoparticles are used to regulate the interface of TiO 2Àx to accelerate the transfer rate of photogenerated electrons and improve the rate of photocatalytic hydrogen evolution.

Materials and Chemicals
All the chemical reagents used in this experiment are of analytical grade, purchased directly, and used without any purification treatment. Tetra butyl titanate (TBT), CuCl 2 ·2H 2 O, and thioacetamide (TAA) were purchased from Aladdin Chemical Reagents Co., Ltd. NaOH (flake and granular) was purchased from Zhengzhou Paini Chemical Reagent Co. Ltd. Na 2 MoO 4 ·2H 2 O, NaBH 4 and diethylene glycol (DEG) were purchased from Sinopharm Chemical Reagent Co. Ltd. Hydrochloric acid was purchased from Kaifeng Dongda Chemical Co. Ltd.

Titanium Glycolate Precursor (TGP)
Five milliliters TBT was dissolved in 180 mL EG and kept refluxing at 120°C for 1 h. [40,41] After cooling down to room temperature, the resulting suspension was centrifuged and washed with deionized water and ethanol respectively. Finally, the obtained white product was dried at 60°C in an oven for 12 h and denoted as TGP.
2.2.2. Synthesis of H 2 Ti 5 O 11 ·3H 2 O Nanoflakes 0.5 g TGP was ultrasonically dispersed in 40 mL deionized water, and 10 mL 5 mol L À1 NaOH solution was added into the solution and stirred for 1 h. Then the mixture was transferred to a 100 mL Teflon-lined autoclave and maintained at 180°C for 2 h. After cooling down to room temperature, white precipitate was washed several times with deionized water and ethanol by centrifuge and dried at 60°C in an oven for 12 h. One gram of white powder was stirred in 60 mL 1 mol L À1 of HCl solution for 12 h to ensure that the Na þ ions were removed by proton exchange. Then the protonic titanate was centrifuged and washed with plenty of distilled water until the pH of supernatant reached 7 and then dried at 60°C. [42]

Synthesis of Nano-TiO 2Àx
All TiO 2 materials are processed with H 2 Ti 5 O 11 ⋅3H 2 O as a precursor. 0.2 g H 2 Ti 5 O 11 ·3H 2 O was dispersed in 40 mL deionized water and treated in 50 mL Teflon-lined autoclave at 180°C for 12 h. The products were washed with distilled water several times and then dried at 60°C, denoted as TiO 2 -H40. TiO 2 -DEG10 was prepared with the same method except for the solution of the reaction system consists of 30 mL deionized water and 10 mL DEG. TiO 2 -DEG20 was prepared with the same method except for the volume of deionized water and DEG were 20 and 20 mL, respectively. TiO 2 -C was obtained through a calcination method. Typically, 0.3 g H 2 Ti 5 O 11 ⋅3H 2 O was calcined in a Muffle furnace at 500°C for 2 h with a heating rate of 5°C min À1 .

Preparation of Cu-TiO 2 -C Materials
Fifty milligrams TiO 2 -C was ultrasonically dispersed in 40 mL deionized water, then 88 μL of 0.1 mol L À1 CuCl 2 ·3H 2 O solution was added under stirring. Next, 10 mL deionized water containing 29 mg NaBH 4 was added to the homogeneous solution by drop. The product 3 wt% Cu-TiO 2 -C was obtained after washing several times with ethanol and dried in vacuum drying oven at 60°C for 12 h. Cu-TiO 2 -C (5 wt%) and Cu-TiO 2 -C (10 wt%) were obtained by adjusting the amount of copper source and reducing agent NaBH 4 .

Preparation of MoS 2 -TiO 2 -C Materials
Fifty milligrams TiO 2 -C was ultrasonically dispersed in 20 mL ethylene glycol. Next, 6 mg Na 2 MoO 4 ·2H 2 O and 12 mg TAA were dissolved in the above solution. The mixture was transferred in a 50 mL Teflon autoclave at 180°C and lasted for 12 h. The resultant black precipitate was washed thoroughly with deionized water and anhydrous ethanol several times. MoS 2 -TiO 2 -C (4 wt%) was obtained after drying in vacuum drying oven process at 60°C for 12 h. MoS 2 -TiO 2 -C (20 wt%), MoS 2 @TiO 2 (40 wt%), and MoS 2 -TiO 2 -C (80 wt%) were obtained with the same method by changing the amount of Na 2 MoO 4 ⋅2H 2 O and TAA, and the mass ratio of Na 2 MoO 4 ⋅2H 2 O and TAA was kept at 1:2. For comparison, pure MoS 2 was synthesized under identical conditions but in the absence of TiO 2 -C.

Material Characterizations
The crystallographic structure was characterized by X-ray diffraction (XRD) (Shimadzu XRD-6000) with Cu Kα radiation at a diffraction angle of 10°-80°with a scanning rate of 30°min À1 . The field emission scanning electron microscopy (FESEM) images and X-ray energy dispersive spectroscopy (EDS) were obtained by field emission scanning electron microscope (SEM, Thermo Scientific Apreo S Hi Vac) with the acceleration voltage of 30 kV. The UV-vis diffuse reflectance spectra (UV-DRS) were measured with Shimadzu UV-2550 spectrophotometer. The surface composition and valence state of the material are determined by X-ray energy spectrum (XPS), which was characterized on AN X-ray energy spectrometer (Thermo Scientific K-Alphaþ) with Al Kα (1486.6 eV) as the radiation source and the binding energy was corrected by the C 1s photoelectron peak (284.8 eV). Photocurrent (PC) and electrochemical impedance spectroscopy (EIS) were measured on an electrochemical workstation (CHI760E) with a standard three-electrode system, in which the Pt wire, the saturated calomel electrode, and the photocatalyst were used as the counter electrode, the reference electrode, and the working electrode, respectively. The prepared products were coated on FTO conductive glass and tested on the working electrode. Oxygen vacancies were analyzed by electron spin resonance spectroscopy (ESR, German Bruker EMXplus).

Photocatalytic Hydrogen Evolution
The photocatalytic hydrogen evolution was performed in a 100 mL quartz reactor with a 3 cm thick top quartz window to allow light irradiation (CEL-SPH2N, CEAULIGHT Beijing) under vacuum atmosphere. In the test, 25 mg of photocatalyst was dispersed by a magnetic stirrer in an up-irradiated photocatalytic reactor containing an aqueous solution of methanol (50 mL, volume ratio of water to methanol is 4:1). The reaction cell was connected to a gas circulation system, and the hydrogen evolved was analyzed by an online gas chromatograph (high-purity N 2 as carrier gas, thermal conductivity detector). The reaction temperature was kept at about 10°C by a circulating water jacket. A 300 W Xe lamp was used as the light source. The gas produced was automatically sampled and analyzed by online gas chromatography. Before the photocatalytic reaction, the reactor was flushed with nitrogen for 30 min to ensure complete removal oxygen.

Results and Discussion
All samples are prepared from H 2 Ti 5 O 11 ·3H 2 O through a series of post-processing methods. As shown in Figure 1, TiO 2 -H40, TiO 2 -DEG10, TiO 2 -DEG20 are obtained by one-step hydrothermal method by changing the volume ratio of water to DEG.      TiO 2 -DEG20 samples gradually turn yellow as the amount of DEG increases, which may be due to the increase of oxygen vacancies. TiO 2 -C, TiO 2 -H40, TiO 2 -DEG10, and TiO 2 -DEG20 are well crystallized and all of the diffraction peaks could be indexed to TiO 2 (JCPDS 4-477) as determined XRD in Figure 3b. X-ray photoelectron spectroscopy (XPS) measurement is further performed to obtain insight into the elemental composition and oxidation states of TiO 2 samples. The survey spectrum shows distinctive peaks of Ti, O, and C elements in Figure 3c, indicating the successful preparation of TiO 2 . The O 1s spectrum in Figure 3d is deconvoluted into three peaks centered at about 529.8, 530.33, and 531.3 eV, which attributes to lattice oxygen (Ti-O), bridged hydroxyl group, and physically adsorbed water, respectively. [43] These surface-bridged hydroxyl groups and physically adsorbed water are usually closely related to the surface oxygen vacancy. The Ti 2p spectrum in Figure 3e, two convoluted peaks at 458.6 and 464.4 eV correspond to Ti 4þ 2p 3/2 and Ti 4þ 2p 1/2 , respectively, whereas the convoluted peaks at 459 and 462 eV correspond to Ti 3þ 2p 3/2 and Ti 3þ 2p 1/2 , respectively.
The appearance of Ti 3þ is to balance the potential difference caused by the presence of oxygen vacancies.
As shown in Figure 4a,b, the color of Cu-TiO 2 -C samples gradually turn blue with the increase of copper loading. The diffraction peaks at 43.3°and 50.5°in Cu-TiO 2 -C, correspond to the (111) and (200) planes of Cu (JCPDS 4-836), respectively, indicating the existence of Cu. The peaks of Cu are not found in the XRD patterns of 3% Cu-TiO 2 -C and 5% Cu-TiO 2 -C, because the content is lower than the detection limit of XRD.
We performed XPS spectra to investigate the chemical compositions and chemical states of 5% Cu-TiO 2 -C. The survey spectrum is shown in Figure 4c, Cu, Ti, O, C, four peaks are detected, indicating the successful preparation of Cu-TiO 2 -C compounds. In Figure 4d, the peaks of Cu spectrum at 932.64 and 952.61 eV are ascribed to Cu 2p 3/2 and Cu 2p 1/2 of Cu, [44,45] while the ones at 933.61 are assigned to Cu 2þ . The peak intensity of Cu 2þ is far weaker than Cu 0 , which implies the slight oxidation of Cu nanoparticles. Ti 2p spectrum in Figure 4e Figure 4f implies the appearance of oxygen vacancies at 532.10 eV.
As shown in Figure 5a,b, the color of MoS 2 -TiO 2 -C samples gradually turns black with the increase of MoS 2 ratio. Restricted by the lower amount and crystallinity, the diffraction peaks of the MoS 2 could be hardly observed in MoS 2 -TiO 2 -C series samples. [46] As shown in Figure 5b, the diffraction peaks of the pure MoS 2 match with MoS 2 (JCPDS 37-1492) with a low crystallinity. All XRD characterization indicates that the sample has been successfully prepared.
The valence state of various species presented in the composite can be known from the XPS analysis. The results of the survey scan of MoS 2 -TiO 2 -C composite in Figure 5c show the presence of C, Ti, O, Mo, and S elements. In Figure 5d,e, the presence of Ti 3þ and oxygen vacancies can be clearly seen in the XPS Ti 2p and O 1s spectra. In Figure 5f, the peaks at 162.3 and 163.5 eV correspond to S 2p 3/2 and S 2p 1/2 . [47] At about 169.1 eV, there is an obvious new low-intensity peak, which is caused by the oxidation of SO 4 2À and S 2À in the air. [48,49] In Figure 5g, two characteristic peaks of MoS 2 at 228.33 and 232.28 eV are attributed to Mo 4þ 3d 5/2 and Mo 4þ 3d 3/2 , respectively. [50,51] The peak at 225.63 eV corresponds to the S2s orbital of www.advancedsciencenews.com www.advenergysustres.com MoS 2 . [47] Compared with MoS 2 , the peak strength of Mo and S in the composite is weaker and the noise is larger, indicating the low load of MoS 2 . UV-vis diffuse reflectance spectra are employed to track the change of light absorbance characteristics of the as-prepared products. As shown in Figure 6a, the absorption edges of TiO 2 -C and TiO 2 -H40 are around 400 nm, which is caused by the photoinduced electrons of O 2p to Ti 2d orbital. The absorption edges of TiO 2 -DEG10 and TiO 2 -DEG20 are around 480 nm, and the enhanced absorption may be due to the increase of oxygen vacancy. As shown in Figure 6b, there is only a steep absorption edge at the UV region for the pristine TiO 2 -C. However, after incorporating Cu species into TiO 2 -C, the hybrids show enhanced absorption of visible light. In Figure 6c, with MoS 2 incorporated into TiO 2 -C, the absorption capability of visible light upon the material is also significantly enhanced by the MoS 2 .
The morphology and microstructure of TiO 2Àx -based samples are investigated by SEM images. Figure 7a,b are the SEM images of TiO 2 -C with low and high magnification, respectively. It can be seen that TiO 2 -C is composed of nanoflakes with 100 nm length and 20 nm width, which assemble into a hollow tubular structure of 3 μm length and 1 μm width. Figure 7c-h are SEM images of TiO 2 -H40, TiO 2 -DEG10, and TiO 2 -DEG20, respectively. The TiO 2 samples obtained by the hydrothermal method are composed of nanoparticles, and the particle size gradually decreases with the increase in the volume of DEG.
To further study the successful preparation of the Cu-TiO 2 -C composite, EDS spectra of the 5% Cu-TiO 2 -C sample is also performed. As shown in Figure 8a-e, Cu, O, and Ti elements are clearly observed. Figure 8b indicates the distribution of Cu on the TiO 2 -C. The morphology of the Cu-TiO 2 -C composite is characterized by SEM images. Figure 8f,g displays the microscopic morphology of the 5 wt% Cu-TiO 2 -C sample in different areas. From the low magnification SEM image in Figure 8f,g, it can be seen that Cu nanoparticles are distributed on the surface of TiO 2 -C and a tight connection is formed between these two materials. Figure 8h,i shows the enlarged images in the dotted box in Figure 8f,g. To have a better and clearer view of the   microstructure of the composite, we enlarge the circle of Figure 8f,g again and get Figure 8j,k. It can be observed that Cu is a coarse surface particle in Figure 8j with a diameter of about 100 nm, while the TiO 2 -C still shows a hollow tube structure consisting of nanosheets in Figure 8k. The MoS 2 -TiO 2 -C samples are obtained by adding S and Mo sources in the hydrothermal process using TiO 2 -C as the support. The structure of as-prepared MoS 2 -TiO 2 -C samples is confirmed by SEM images. As shown in Figure 9a, the MoS 2 is a particle with a diameter of about 200 nm. The morphology of the 40% MoS 2 -TiO 2 -C sample is shown in Figure 9b, and the enlarged images of the circles are shown in Figure 9b To evaluate the photocatalytic hydrogen evolution performance of different TiO 2Àx samples, the line and column chart of photocatalytic hydrogen evolution rate are shown in Figure 10a,b. TiO 2 -C has the best photocatalytic activity of hydrogen evolution, and the hydrogen production rate is up to 3,112.09 μmol g À1 h À1 . For the TiO 2 -DEG10, the hydrogen evolution rates decrease greatly after 60 min, and it might be due to the equilibration of H 2 yield. The photocatalytic hydrogen evolution of the composites test is also performed. As shown in Figure 10c,d, the photocatalytic hydrogen evolution on TiO 2 -C is greatly enhanced by the introduction of copper as a cocatalyst. Among those composites, the one with 5% copper loading has the highest photocatalytic hydrogen evolution rate of 9,225.75 μmol g À1 h À1 , which is twice higher than that of TiO 2 -C. As shown in Figure 10e,f, the involvement of MoS 2 cocatalyst also can enhance the photocatalytic hydrogen generation of the TiO 2 -C sample. It can be seen that the 40 wt% MoS 2 -TiO 2 -C displays the best photocatalytic hydrogen evolution rate of 5765.48 μmol g À1 h À1 ( Table 1).
In addition, we have performed the cycle test in the photocatalytic hydrogen evolution upon the 5 wt% Cu-TiO 2 -C and 40 wt% MoS 2 -TiO 2 -C in the updated manuscript. As shown in Figure S1, Supporting Information, after four cycle test, the 5 wt% Cu-TiO 2 -C sample still has photocatalytic hydrogen evolution rate of 4040 μmol g À1 h À1 , which is higher than that of pristine TiO 2 -C. However, when two cycle test is performed upon the 40 wt% MoS 2 -TiO 2 -C, as shown in Figure S2, Supporting Information, it only shows photocatalytic hydrogen evolution rate of 2100 μmol g À1 h À1 , and further study on this poor stability is underway. As shown in Figure S3 and S4, Supporting Information, the cycled 5 wt% Cu-TiO 2 -C and 40 wt% MoS 2 -TiO 2 -C still keep the main crystal phase of anatase TiO 2 , indicating its good composition stability. Figure S5 and S6, Supporting Information, show the SEM and EDS mapping images of the cycled 5 wt% Cu-TiO 2 -C and 40 wt% MoS 2 -TiO 2 -C samples, and it can be observed that the photocatalysts   Figure 11a, according to the results of UV-DRS, the band gap (E g ) is calculated according to Kubellka-Munk formula. [52][53][54] The bandgap of TiO 2 -C, TiO 2 -H40, TiO 2 -DEG10, and TiO 2 -DEG20 are 3.04, 2.94, 2.52, and 2.05 eV, respectively. The addition of DEG can significantly narrow the bandgap of TiO 2 . As shown in Figure 11b, the VB edge of TiO 2 -C, TiO 2 -H40, TiO 2 -DEG10, and TiO 2 -DEG20 are 2.76, 2.59, 2.46, and 2.53 eV, respectively. According to the correlation  between conduction band and band gap E CB = E VB À E g , the CB-VB potential diagram is shown in Figure 11d. Compared with the other three samples, the TiO 2 -C sample possesses the most positive VB potential and more negative CB potential, which is in favor of the photocatalytic hydrogen evolution.
To better explain the difference in oxygen vacancy concentration, ESR is used to characterize the samples. As shown in Figure 11c, the oxygen vacancy concentration of the TiO 2 samples from high to low is TiO 2 -DEG20, TiO 2 -DEG10, TiO 2 -C, and TiO 2 -H40. With the increase of the volume of DEG in the reaction system, the concentration of oxygen vacancies significantly increases, in which DEG as a reducing agent is beneficial for the generation of oxygen vacancies.
Photocurrent response is conducted to verify the photoinduced electron-hole pairs' separation efficiency. As shown in Figure 11e, the TiO 2 -C sample shows the strongest photocurrent response in each on/off light cycle, indicating of the best separation rate among all the TiO 2 samples. The smallest arc radius of TiO 2 -C in Figure 11f also proves its lowest electron transport resistance. Figure 12a shows the periodic on/off photocurrent response of the TiO 2 -based composites. In comparison with pure TiO 2 -C, dramatic improvement of photocurrent density can be observed in the 5% Cu-TiO 2 -C composite, which suggests that surface oxygen vacancy and Cu cocatalyst contribute to inhibiting the recombination of electron-hole pairs. [41,55,56] At the same time, as shown in Figure 12b, it is obvious that the 5% Cu-TiO 2 -C sample displays a smaller semicircle in the Nyquist plots than that of pure TiO 2 -C, indicating its lower interfacial chargetransfer resistance. As shown in Figure 12c,d, the photocurrent intensity of the 40% MoS 2 -TiO 2 -C composite is higher than that of pristine TiO 2 -C, while the radius of the arc is smaller than TiO 2 -C. This result demonstrates that the formation of oxygen vacancy and the modification of MoS 2 are favorable to interface charge transfer and migration. On the basis of the above results, the enhanced photocatalytic hydrogen evolution upon the TiO 2 -C, Cu-TiO 2 -C, and MoS 2 -TiO 2 -C photocatalysts is proposed. As described in Figure 13a, upon the excitation by the light irradiation, TiO 2 -C would absorb photons and generate electron-hole pairs. Appropriate oxygen vacancy is beneficial to capture electrons for reduction reaction and hydrogen evolution, while excessive oxygen vacancy would become the recombination sites of photogenerated electrons and holes. Different postprocessing methods of protonic titanate could tailor the photocatalytic hydrogen production activity by engineering oxygen vacancy concentration. With involvement of the cocatalysts, the obtained Cu-TiO 2 -C and MoS 2 -TiO 2 -C composite could have binary channels for the transfer of the photoinduced electron-hole pairs. On one hand, the cocatalyst on the TiO 2 -C acts as active site to capture the electrons for H 2 generation. On the other hand, the oxygen vacancy on the surface of TiO 2 -C could facilitate electron transfer and thus speeds up the electron-hole separation rate. Simultaneously, the photoinduced holes will migrate to their surface and are consumed by methanol sacrificial reagents. As a consequence, the oxygen vacancies on the surface and the modification of Cu or MoS 2 cocatalyst on the interface can act as electron traps to attract electrons from TiO 2 -C, which can effectively improve the charge transfer and separation, thus improving the photocatalytic hydrogen evolution rates of Cu-TiO 2 -C and MoS 2 -TiO 2 -C composites.

Conclusion
In summary, protonic titanate-derived oxygen vacancy TiO 2 photocatalysts have been prepared by calcination and hydrothermal treatment. The concentration of oxygen vacancy can be regulated by adjusting the proportion of DEG in the hydrothermal process. Interestingly, the postprocess approaches not only enable to tailor the oxygen vacancies but also reduce the band gap of TiO 2 , making it responsive to visible light. Through the regulation of surface oxygen vacancy, it is found that the TiO 2 -C prepared by calcination has appropriate oxygen vacancy concentration for enhanced photocatalytic hydrogen evolution. Accordingly, by introducing Cu or MoS 2 cocatalysts into the TiO 2 -C, the tight contact between the cocatalyst and the semiconductor is beneficial for the transfer of photogenerated electrons and holes. In this regard, in the obtained Cu-TiO 2 -C and MoS 2 -TiO 2 -C composite, the oxygen vacancy and the cocatalyst on the surface of TiO 2 -C could facilitate electron transfer and acts as active sites for efficient H 2 generation. Indeed, the 5 wt% Cu-TiO 2 -C and 40 wt% MoS 2 -TiO 2 -C have the highest hydrogen rates of 9225.75 and 5765.48 μmol g À1 h À1 , respectively. It is hoped this work could offer guidance for surface oxygen vacancy and interface cocatalyst regulation for solar-to-hydrogen evolution. www.advancedsciencenews.com www.advenergysustres.com