Cocatalyst designing: a binary noble-metal-free cocatalyst system consisting of ZnIn2S4 and In(OH)3 for efficient visible-light photocatalytic water splitting

A binary noble-metal-free cocatalyst consisting of ZnIn2S4 and In(OH)3 was developed via a facile one-step hydrothermal method. The ZnIn2S4/In(OH)3 modified ZnWO4 nanocomposite exhibited enhanced photocatalytic H2 evolution activity compared to all the related pure samples and binary composite photocatalysts under visible light irradiation. The enhanced photocatalytic hydrogen production activities can be attributed to the synergistic effects of the favorable light trapping ability and efficient spatial charge separation. The photocatalytic hydrogen evolution activity over other semiconductors, such as Zn2SnO4 and TiO2, can also be significantly increased by loading ZnIn2S4/In(OH)3 as a cocatalyst. The results clearly demonstrated that ZnIn2S4/In(OH)3 is discovered as a new class of earth-abundant cocatalyst for water-splitting under visible light irradiation. It is expected that our work could provide a new strategy to improve the visible light response of semiconductors and facilitate their application in water splitting.


Introduction
Photocatalytic water-splitting is a promising approach for the clean, economical, and environmentally benign production of hydrogen by utilizing solar energy. [1][2][3][4][5][6][7][8] During the past few decades, many photocatalytic materials, such as sulde-, oxideand oxynitride based materials, have been extensively explored for hydrogen generation through splitting water. [9][10][11] However, none of them can meet the requirements for practical largescale application to photocatalytic H 2 evolution under visible light. Hence, the highly efficient visible light photocatalysts for hydrogen production is still highly demanded for practical applications.
It is well-known that a cocatalyst plays a signicant role in improving both the activity and stability of semiconductor photocatalysts. Noble metals such as Pt, Rh, Au and oxides such as RuO 2 are generally used as cocatalysts for photocatalytic hydrogen evolution. However, the above noble-metal based cocatalysts are too scarce and expensive. It is necessary to explore alternative cocatalysts based on inexpensive transition metals to facilitate the up-scaling of renewable H 2 production.
Recently, ZnIn 2 S 4 with a suitable band gap (2.06-2.58 eV), 12,13 have been demonstrated to be excellent cocatalyst for CdIn 2 S 4 and K 2 La 2 Ti 3 O 10 in the photocatalytic hydrogen evolution. [14][15][16] It has been also proved to be an efficient visible-light driven photocatalyst or sensitizer to extend the absorption spectra of some wide-gap photocatalysts from the UV region to the visible region. 17,18 Furthermore, the unique 2D layer structure nanostructure of ZnIn 2 S 4 enables its application as an excellent supporting matrix in the construction of heterostructured nanocomposite photocatalysts with high photocatalytic activity. [19][20][21][22] In addition, it was reported that indium ions can hydrolyze into In(OH) 3 when the pH value of the reaction system was above 2.5. 23 So it is convenient to obtain ZnIn 2 S 4 /In(OH) 3 nanocomposite by a simple hydrothermal strategy by adjusting the pH value of the solution. In(OH) 3 is an important wide-gap semiconductor material with a direct bandgap of 5.15 eV. [23][24][25] The wide band gap provides photogenerated charge with stronger redox capability, favoring photocatalytic performance. Therefore it is expected that ZnIn 2 S 4 /In(OH) 3 can be an ideal candidate as a nonnoble-metal cocatalyst system for water splitting under visible light irradiation.
Herein, the ZnIn 2 S 4 /In(OH) 3 modied ZnWO 4 nanocomposite was fabricated by a facile three-step hydrothermal strategy. The results revealed that as-prepared ternary ZnIn 2 S 4 / In(OH) 3 /ZnWO 4 nanocomposite exhibited enhanced photocatalytic H 2 -evolution activity than all the related pure samples and binary composite photocatalysts. Based on the experimental results, a possible synergistic mechanism for the enhanced photocatalytic H 2 evolution activity was proposed. The photocatalytic hydrogen evolution activity over other semiconductor such as Zn 2 SnO 4 26 and TiO 2 , 27 can also be signicantly increased by loading ZnIn 2 S 4 /In(OH) 3 as a cocatalyst. The results clearly demonstrated that ZnIn 2 S 4 /In(OH) 3 is discovered as a new class of earth-abundant cocatalyst for watersplitting under visible light irradiation. It is expected that our work could provide deep insight into design and application of new system with high activity for solar hydrogen generation in the future.

Experimental
Preparation of photocatalysts Synthesis of ZnIn 2 S 4 , In(OH) 3 and ZnIn 2 S 4 /In(OH) 3 . ZnIn 2 S 4 was synthesized by a hydrothermal method. In a typical procedure, 0.300 g Zn(NO 3 ) 2 $6H 2 O, 0.782 g In(NO 3 ) 3 $4.5H 2 O and 0.616 g thiourea were dissolved in deionized water. The pH value of the solution was adjusted to 1.00 by adding 0.1 mol L À1 HCl. The mixed solution was then transferred into a 100 mL Teon-lined autoclave and heated at 180 C for 12 h. Aer being cooled to room temperature, the yellow precipitate was collected by centrifugation and washed with ethanol and the distilled water for three times, respectively. And then it was dried at 80 C to obtain the product.
The preparation of In(OH) 3 was similar to that of ZnIn 2 S 4 . In a typical procedure, 0.782 g In(NO 3 ) 3 $4.5H 2 O was added into the solution. The pH value of the solution was adjusted to 6 by adding 0.2 mol L À1 NaOH. The mixed solution was then transferred into a 100 mL Teon-lined autoclave and heated at 180 C for 12 h. The remaining steps are the same as ZnIn 2 S 4 .
The synthesis of ZnIn 2 S 4 /In(OH) 3 is similar to that of ZnIn 2 S 4 . All of the experimental conditions are same except the initial pH value. The pH value of the solution was adjusted to 3.5, 4.5 and 5.5 with 0.2 mol L À1 NaOH, respectively. The products synthesized at different pH values were labeled as ZnIn 2 S 4 /In(OH) 3 -3.5, ZnIn 2 S 4 /In(OH) 3 -4.5, ZnIn 2 S 4 /In(OH) 3 -5.5, respectively.
Preparation of ZnWO 4 . 5 mmol Na 2 WO 4 $2H 2 O and 5 mmol Zn(NO 3 ) 2 $6H 2 O were added to 70 mL deionized water with magnetic stirring to form a homogeneous solution. The reaction mixture was then sealed in a 100 mL Teon-lined autoclave and kept at 180 C for 12 h. Aer cooling, the product was ltered, washed, and dried at 80 C.
Fabrication of the ZnIn 2 S 4 /ZnWO 4 nanocomposite. 0.2 g ZnIn 2 S 4 was ultrasonicated in 70 mL deionized water for 20 min. Aer that, 0.15 g ZnWO 4 was added into the solution to prepare the ZnIn 2 S 4 /ZnWO 4 composite. The mixture was continued to be ultrasonicated for 20 min to obtain a homogeneous suspension and then it was stirred vigorously. Next, this suspension was transferred into a 100 mL Teon-lined autoclave and maintained at 180 C for 12 h. Aer cooling, the product was ltered, washed, and dried at 80 C.
Preparation of In(OH) 3 /ZnWO 4 nanocomposite. 0.2 g In(OH) 3 was ultrasonicated in 70 mL deionized water for 20 min. Aer that, 0.15 g ZnWO 4 was added into the solution to prepare the In(OH) 3 /ZnWO 4 composite. The mixture was continued to be ultrasonicated for 20 min to obtain a homogeneous suspension and then it was stirred vigorously. Next, this suspension was transferred into a 100 mL Teon-lined autoclave and maintained at 180 C for 12 h. Aer cooling, the product was ltered, washed, and dried at 80 C.
Synthesis of ZnIn 2 S 4 /In(OH) 3 /ZnWO 4 (ZIZ) nanocomposite. 0.2 g ZnIn 2 S 4 /In(OH) 3 synthesized at different pH values was ultrasonicated in 70 mL of deionized water for 20 min. Aer that, 0.15 g ZnWO 4 was added into the solution to prepare a series of ZIZ composites. The mixture was continued to be ultrasonicated for 20 min to obtain a homogeneous suspension and then it was stirred vigorously. Next, this suspension was transferred into a 100 mL Teon-sealed autoclave and maintained at 180 C for 12 h. Aer being cooled to room temperature, the precipitate was collected by centrifugation and washed with ethanol and the distilled water for three times, respectively. And then it was dried at 80 C to obtain the product. The obtained samples were ZnIn 2 S 4 /In(OH) 3 /ZnWO 4 -3.5, ZnIn 2 S 4 / In(OH) 3 /ZnWO 4 -4.5 and ZnIn 2 S 4 /In(OH) 3 /ZnWO 4 -5.5, which were labeled as ZIZ-3.5, ZIZ-4.5, ZIZ-5.5, respectively.

Characterization
The phase structure and crystallinity of the products was measured by X-ray diffraction (XRD) patterns, which were recorded on a Bruker D8 Advance X-ray diffractometer with Cu Ka radiation at 40 kV and 40 mA. The morphology of the samples was detected by eld emission scanning electron microscopy (FESEM, Hitachi S-4800) and transmission electronic micrograph (TEM, Hitachi, H-7650). High resolution transmission electron microscopy (HR-TEM) images were collected on a JEM-2100 electron microscope, operated at an acceleration voltage of 200 kV. XPS studies were carried out by means of a VG-ESCALAB250 electronic spectrometer using an Al Ka excitation source (hn ¼ 1486.6 eV). The position of the C 1s peak was taken as a standard (with a banding energy of 284.6 eV). The Brunauer-Emmett-Teller (BET) surface areas (S BET ) of the samples were measured by a N 2 sorption analyzer (Quantachrome, NOVA 4000e) at a liquid nitrogen temperature. UV-vis diffuse reectance spectra of the samples were evaluated using a UV-vis spectrophotometer (U-3010, Hitachi, Japan). Photoluminescence (PL) spectra of the samples were performed using an Hitachi F-4600 uorescence spectrophotometer.

Photocatalytic H 2 -production activity
The photocatalytic hydrogen-production experiment was conducted in a closed gas circulation and evacuation system tted with a top window Pyrex cell. A 300 W xenon lamp (PLS-SXE300C, Beijing Perfectlight Co. Ltd., China) which was positioned 10 cm away from the solution surface, coupled with a UV cut-off lter (l > 420 nm) was used to provide the visible light. The focus intensity measured by a spectroradiometer (PM100D, Thorlabs, America) was 0.54 W cm À2 . In each run, 0.05 g photocatalyst was dispersed in a 100 mL aqueous solution containing 0.25 M Na 2 SO 3 and 0.35 M Na 2 S, which served as a sacricial agent. Before the irradiation with visible light, the suspension was degassed with N 2 for 1 h to drive away the O 2 in the system. The reaction cell was kept at room temperature with cooling water. The produced H 2 was detected using an online gas chromatography (GC7900, N 2 carrier, 5 A molecular sieve column, TCD detector). The reaction was continued for 3 h. The apparent quantum efficiency (QE) was measured by the similar condition, but the light source was changed into high power LED (PLS-LED100B, Beijing Perfectlight Co. Ltd., China). The QE was calculated based on the following equation:

Photoelectrochemical measurements
The working electrodes were prepared as follow: 0.05 g of the sample, 0.05 g of polyethyleneglycol (molecular weight 20 000) and 0.5 mL of ethanol were mixed and ground into the slurry. Then the slurry was evenly coated onto a 2 cm Â 1.2 cm F-doped SnO 2 -coated (FTO) glass electrode by the doctor blade method. The area and the thickness of the coating layer were limited to 1 cm 2 and 0.01 mm, respectively. Finally, the prepared electrodes were dried in an oven at 100 C for 60 min. The measurement of the photocurrent and electrochemical impedance spectroscopy (EIS) were performed on an electrochemical analyzer (CHI660E, Shanghai Chenhua Limited, China) in a standard three-electrode system. In this system, the prepared glass electrodes were used as the working electrodes. Pt wire and Ag/AgCl (saturated 0.3 M KCl) were used as counter electrode and reference electrode, respectively. The three electrodes were connected by the electrolyte of 0.5 M Na 2 SO 4 . The source of visible light is a 300 W Xe lamp with a UV cut-off lter (l > 420 nm). The EIS were tested in the same standard three-electrode system, the phase frequency and amplitudefrequency characteristics were recorded over frequency range from 0.01 Hz to 10 5 Hz, and then the spectrum was transferred into Nyquist diagram.

Results and discussion
Structure and properties characterization  3 and ZnWO 4 coupled together successfully and without other phases. The morphology of the samples was investigated by SEM. Fig. 2a shows the SEM image of the as-prepared ZnIn 2 S 4 at a higher magnication. It is shown that the ZnIn 2 S 4 microsphere has a unique marigold-like spherical superstructure composed of numerous nanosheets. As shown in Fig. 2b, the asprepared pure In(OH) 3 was made up of rectangular blocks with an edge of about 0.7-2 mm. It is also clearly seen that some small nanoparticles are attached to the edge of rectangular blocks, indicating that In(OH) 3 rectangular blocks may be formed by the adsorption of small In(OH) 3 nanoparticles via an Oswald ripening mechanism. 28 As shown in Fig. 2c, the as-prepared ZnWO 4 was composed of nanoparticles with a wide size distribution. Fig. 2d and e show the SEM image of the ZIZ-4.5 nanocomposite. It is clearly observed that In(OH) 3 cubic blocks and ZnWO 4 nanoparticles were attached to the surface of ZnIn 2 S 4 microspheres. Fig. 2f shows the HRTEM images of as prepared ZIZ-4.5 sample. As shown in Fig. 2f, the fringes of 0.33 nm, 0.29 nm and 0.48 nm correspond to the (102), (220) and (100) plane of ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 , respectively. [29][30][31] The EDS spectrum (Fig. 3) of the composite clearly conrms the presence of Zn, In, S, O and W elements. From the above results, it can be seen that ZIZ composites with close contact have been successfully fabricated.
To characterize the chemical composition of as-prepared ZIZ photocatalyst, X-ray photoelectron spectroscopy (XPS) measurement was used to determine the exact surface state. The XPS survey spectrum conrms that the composite is mainly composed of Zn, In, S, W and O (Fig. 4a). The high-resolution XPS spectra of Zn 2p region of the catalyst was shown in Fig. 4b. The XPS signals of Zn 2p were observed at binding energies at around 1021.9 eV (Zn 2p 3/2 ) and 1044.9 eV (Zn 2p 1/2 ). The high resolution XPS spectra of In 3d region of the catalyst was shown in Fig. 4c. The XPS signals of In 3d were observed at binding energies at around 445.0 eV (In 3d 5/2 ) and 452.6 eV (In 3d 3/2 ), which are in good agreement with the previous works. 32 As illustrated in Fig. 4d, the appearance of S 2p 3/2 and S 2p 1/2 peaks at 161.9 and 163.0 eV can be assigned to the S 2À in ZnIn 2 S 4 . 33 The XPS spectrum in the W 4f region exhibits binding energy at 35.8 eV for W 4f 7/2 and 37.9 eV for W 4f 5/2 (Fig. 4e), the gap between them is about 2.1 eV, which is close to the value of reference, 34 indicating that the W element exhibits in the chemical formation of W 6+ . The O 1s region can be tted by three peaks with binding energies of 530.5 eV, 531.5 eV, and 533.5 eV. The peaks at 530.5 eV and 531.5 eV correspond to the lattice oxygen in ZnWO 4 and In(OH) 3 , 30,35 respectively. While the peaks at 533.5 eV can be attributed to the oxygen in the adsorbed H 2 O on the surface of sample. The above results show that the prepared sample is ZIZ.
The nitrogen adsorption-desorption isotherms of ZIZ-4.5, ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 are shown in Fig. 5. The obtained results from the sorption isotherms suggest that all samples exhibit a type IV isotherm with a H3 hysteresis loop according to the IUPAC standard, 30,36,37 indicating the existence of a mesoporous structure or slit-like pore. The porous structure parameters of pure ZnIn 2 S 4 , In(OH) 3 , ZnWO 4 and ZIZ-4.5 composites are listed in Table 1. The BET surface area (S BET ) of the pure ZnIn 2 S 4 , In(OH) 3 , ZnWO 4 was 33.730, 13.736, and 11.876 m 2 g À1 , respectively. The BET surface area of the ZIZ-4.5 nanocomposite was 27.274 m 2 g À1 , which is lower than that of ZnIn 2 S 4 . The reduction of S BET in the ternary sample may be ascribed to the blockage of partial macropores due to the    The optical absorption properties of as-prepared samples were evaluated by UV-vis diffuse reection spectra (DRS) measurement (Fig. S1a †). The pristine ZnWO 4 showed a narrowed light absorption edge of no more than 380 nm, due to its intrinsic wide band gap. 38 The pure In(OH) 3 has weak absorption ability, with an absorption edge of ca. 250 nm. 24 The pure ZnIn 2 S 4 has a strong absorption in the wavelength ranging from 200 to 550 nm. 39 The introduction of ZnIn 2 S 4 increases the light absorption of composite materials in the range of the visible light region. ZIZ-4.5 composite has a strong absorption both in the UV region and visible light region (Fig. S1a †).
The energy level and band gap of the semiconductors play a crucial role in determining their physical properties. The band gap energy of a semiconductor can be calculated by the following formula: where a, h, v, E g and A are the absorption coefficient, Planck constant, light frequency, band gap energy, and a constant, respectively. Among them, n is determined from the type of optical transition of a semiconductor (n ¼ 1 for direct transition and n ¼ 4 for indirect transition). The n value of ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 is 1 for the direct transition. 30,40,41 Therefore, the band gap energy (E g value) of the samples can be estimated from a plot of (ahn) 2 -(hn). The band gap energy of ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 is 2.50 eV, 4.99 eV and 3.84 eV, respectively (Fig. S1b †). All the values for the band gap of ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 are very close to previously reported results. [42][43][44] The valence band edge potential and the conduction band edge potential of a semiconductor material can be determined by using the following equation: 45,46 where E VB represents valence band edge potential, X is the electronegativity of the semiconductor estimated by the geometric mean of the electronegativity of the constituent atoms, E e is the energy of free electrons on the hydrogen scale (4.5 eV), E g is the band gap energy of the semiconductor, E CB is the conduction band edge potential. The X value of pure ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 is about 4.82, 6.155 and 5.55 eV, 24,44,47 respectively. The E VB of bare ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 can be assigned to be +1.57, +4.15 and +2.97 eV, respectively. The corresponding E CB of ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 can be estimated to be À0.93, À0.84 and À0.87 eV, respectively.

Photocatalytic activities
Photocatalytic H 2 production experiments were carried out over these as-prepared photocatalysts in the presence of Na 2 SO 3 and Na 2 S as the sacricial reagents under visible light irradiation (l > 420 nm). 48 As shown in Fig. 6, pure ZnWO 4 (116.7 mmol g À1 h À1 ), pure In(OH) 3 (13.4 mmol g À1 h À1 ) and pure ZnIn 2 S 4 (187.6 mmol g À1 h À1 ) show their low photocatalytic activity, respectively. The ternary ZnIn 2 S 4 /In(OH) 3 /ZnWO 4 (ZIZ) nanocomposite exhibited enhanced photocatalytic H 2 evolution activity as compared to all the related pure samples and binary composite photocatalysts. Especially, ZIZ-4.5 shows the highest photocatalytic activity with a H 2 evolution rate of 1030.1 mmol g À1 h À1 , which was 8.83 times higher than that of ZnWO 4 alone. Moreover, the apparent quantum efficiency (QE) for H 2 evolution was examined at 420 nm to further evaluate the photocatalytic activity of ZIZ-4.5. The corresponding QE reached 8.1%, which indicated that the ZIZ-4.5 is a highly-efficient   photocatalyst for hydrogen evolution under visible light irradiation. A further increase of pH value will lead to a signicant reduction of the photocatalytic H 2 -evolution rate. With increasing pH value, more ZnIn 2 S 4 transformed into In(OH) 3 through the hydrolysis reaction of In 3+ ions, which is evidenced from the XRD measurements ( Fig. 1e-g). Excessive amount of In(OH) 3 particles may cover the active sites on the surface of composite, shield the incident light and act as charge recombination centers. 49 In addition, the decrease of ZnIn 2 S 4 can result in the reduction of the utilization of visible light, leading to the decrease of hydrogen production activity.
To conrm the stability of our composite catalysts, we performed four consecutive hydrogen evolution runs over ZIZ-4.5 under the same conditions (Fig. S2 †). 50 Each cycle was performed under visible light irradiation for 3 h. Aer each run, the reaction system was re-evacuated. The H 2 evolution rate aer the fourth cycle can keep ca. 91% of the initial rate. This decrease of the H 2 evolution rate from 3090 to 2805 mmol g À1 aer four cycles could be attributed to the continuous consumption of Na 2 S and Na 2 SO 3 . 51 The above results indicate that the prepared ZIZ-4.5 composite has good stability during photocatalytic reactions.

Possible mechanism of photocatalytic H 2 evolution
A schematic illustration of the tentative mechanism proposed for the high H 2 production activity of the ternary ZIZ composite is shown in Fig. 7. Under visible light illumination, ZnIn 2 S 4 with narrow bandgap energy (2.50 eV in this work) can be easily excited and generate electrons and holes. Normally, these photogenerated electrons and holes recombine rapidly, resulting in a poor H 2 evolution activity of pure ZnIn 2 S 4 . However, in the ternary ZIZ composite system, since the CB of ZnIn 2 S 4 (À0.93 eV) is more negative than that of In(OH) 3 (À0.84 eV) and ZnWO 4 (À0.87 eV), the photogenerated electrons in the CB of ZnIn 2 S 4 are transferred to the CB of In(OH) 3 and ZnWO 4 (Fig. 7), where the electrons can effectively reduce H + and produce H 2 molecules. 52 Meanwhile, holes accumulated at the VB of ZnIn 2 S 4 could be consumed by the sacricial agent (S 2À , SO 3 2À ). 53 Owning to the close contact of ZIZ composite, photogenerated electron can be transferred quickly, so as to extend the charge carriers lifetime and reduce the recombination of electron-hole pairs. Consequently, the high photocatalytic H 2production activity is achieved over the ternary ZIZ composite. The charge recombination and transfer can be investigated by photoluminescence (PL) spectra. It can also help understanding the fate of photogenerated electrons and holes in semiconductor because PL emission results from the recombination of free carriers. 54 The PL spectra of ZIZ-4.5, ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 samples were recorded at room temperature with an excitation wavelength of 215 nm (Fig. S3 †). It can be observed that the bare ZnIn 2 S 4 shows a strong peak at about 550 nm, which can be attributed to the direct recombination of the photo-generated electron-hole pairs in ZnIn 2 S 4 nanocrystals. When introducing both ZnWO 4 and an optimum concentration of In(OH) 3 content, ZIZ nanocomposite exhibited much lower PL emission peak at about 470 nm. The results implied that ZIZ has remarkably enhanced separation efficiency of photo-induced electron-hole pairs, which is qualitatively in good agreement with the trend in photo-activity enhancement.
To further deeply understand the photocatalytic enhancement mechanism for H 2 evolution, the transfer and separation of photo-generated charge carriers were also investigated through the photoelectrochemical (PEC) analysis. The transient Fig. 7 Possible mechanism of the photocatalytic H 2 -production activity of the ternary ZIZ nanocomposite under visible-light irradiation. Fig. 8 The photocatalytic activity for hydrogen production of (a) ZnIn 2 S 4 , In(OH) 3  photocurrent responses of ZnIn 2 S 4 , In(OH) 3 , ZnWO 4 and ZIZ-4.5 with typical on-off cycles of intermittent irradiation were measured (Fig. S4a †). Notably, the photocurrent of ZIZ-4.5 sample is much higher than the other samples, conrming the more efficient interfacial mobility and separation of photogenerated electron-hole pairs for the ZIZ-4.5 composite sample. In addition, the electrochemical impedance spectroscopy (EIS) has also been investigated to provide sufficient evidence for the interfacial charge transfer resistance and separation efficiency. The electrochemical impedance spectroscopy (EIS) of these four samples were shown (Fig. S4b †). It is observed that the arc radius in the Nyquist plots of ZIZ-4.5 is much smaller than those of ZnIn 2 S 4 , In(OH) 3 and ZnWO 4 , indicating that a more effective separation of photogenerated electron-hole pairs and a faster interfacial charge transfer had occurred on the surface of ZIZ composite. The PL and PEC analysis results are in good accordance with the photocatalytic-activity measurements, which further conrm that introducing ZnIn 2 S 4 /In(OH) 3 system is an effective way to improve the photocatalytic efficiency of ZnWO 4 .
Although the results here obtained certainly open up the possibility to design a noble-metal-free photocatalyst for water splitting, the deposition of the ZnIn 2 S 4 /In(OH) 3 cocatalyst on other semiconductor is certainly required. As shown in Fig. 8, the photocatalytic H 2 -production activity of Zn 2 SnO 4 and TiO 2 were both tremendously improved by loading the specied amount of ZnIn 2 S 4 /In(OH) 3 as cocatalyst. It is expected that the combination of ZnIn 2 S 4 and In(OH) 3 as a new class of earthabundant cocatalyst can become a general strategy to improve the H 2 -evolution activity over various kinds of conventional semiconductors.

Conclusions
In conclusion, the combination of ZnIn 2 S 4 with In(OH) 3 formed a new class of earth-abundant cocatalyst for water-splitting under visible light irradiation, which greatly advanced the photocatalytic activity of ZnWO 4 promoted by synergetic effects. The roles of ZnIn 2 S 4 and In(OH) 3 in boosting the H 2 evolution activity of ZnWO 4 were revealed. This work could not only provide an excellent candidate for visible-light H 2 evolution, but also opens new possibilities to provide insight into the design of new system with high activity for solar hydrogen generation.

Conflicts of interest
There are no conicts to declare.