Since the discovery of photo-induced water splitting on TiO₂ electrodes 1, solar-driven photocatalytic hydrogen production from water using a semiconductor catalyst has attracted a tremendous amount of interest 23. To efficiently utilize solar energy, numerous attempts have been made in recent years to realize different visible light-active photocatalysts 45678. Among them, sulfides, especially CdS-based photocatalysts with narrow band gaps, proved to be good candidates for photocatalytic hydrogen evolution from water under visible light irradiation 9101112. However, CdS itself is not stable for water splitting, and Cd²⁺ is hazardous to environment and human health. A number of nontoxic multicomponent sulfides without Cd²⁺ ions have been developed to show comparable photocatalytic efficiency for hydrogen evolution 1314151617. In our previous work 1819202122, hydrothermally synthesized ZnIn₂S₄ was found to have photocatalytic and photoelectrochemical properties that made it a good candidate for hydrogen production from water under visible light irradiation. On the other hand, a solid co-catalyst, typically noble metal (e.g., Pt, Ru, Rh) 23 or transition-metal oxide (e.g., NiO 24, Rh_2-yCr_yO₃ 25, RuO₂ 26, IrO₂27), loaded on the surface of the base photocatalyst can be beneficial to photocatalytic H₂ and/or O₂ evolution for water splitting 25. Nevertheless, there have been only a limited number of studies in which metal sulfides acted as co-catalysts to enhance the activity of a semiconducting photocatalyst 282930. For instance, Li and co-workers observed that dual co-catalysts consisting of noble metals (Pt, Pd) and noble-metal sulfides (PdS, Ru₂S₃, Rh₂S₃) played a crucial role in achieving very high efficiency for hydrogen evolution over CdS photocatalyst 2930. In this study, a series of transition-metal sulfides (MS: Ag₂S, SnS, CoS, CuS, NiS, and MnS) were deposited on hydrothermally synthesized ZnIn₂S₄ by a simple precipitation process. The photocatalytic activities for hydrogen evolution over these MS/ZnIn₂S₄ products were investigated. We demonstrated that transition-metal sulfides, such as CuS, combined with Pt could act as dual co-catalysts for improving photocatalytic activity for hydrogen evolution from a Na₂SO₃/Na₂S aqueous solution under simulated sunlight.

The ZnIn₂S₄ products prepared by the CTAB-assisted hydrothermal method possessed a hexagonal structure and morphology of microspheres comprising of numerous petals, and showed an absorption edge at about 510 nm (Additional file 1, Figure S1-3). Compared to pure ZnIn₂S₄, the obtained MS/ZnIn₂S₄ (MS = metal sulfide: Ag₂S, SnS, CoS, CuS, NiS, and MnS) displayed different absorption profiles (Additional file 1, Figure S4), with enhanced absorption in the visible light region from 550 to 800 nm. Such additional broad band (λ > 550 nm) can be assigned to the absorption of transition-metal sulfides.

We investigated the photocatalytic activity for hydrogen evolution over MS/ZnIn₂S₄ (MS = metal sulfide: Ag₂S, SnS, CoS, CuS, NiS, and MnS). Photocatalytic activities for hydrogen evolution over MS/ZnIn₂S₄ were evaluated by loading 1 wt.% Pt as co-catalyst. Figure 1 shows the average rates of H₂ evolution over Pt-loaded MS/ZnIn₂S₄ photocatalysts under simulated solar irradiation in the initial 20-h period. The Pt-ZnIn₂S₄ showed a photocatalytic activity for H₂ production at the rate of 126.7 μmol·h^-1, which is comparable to reported values in previous literatures 181920. The hydrogen production rates of Pt-MS/ZnIn₂S₄ photocatalysts varied with different kinds of loaded transition-metal sulfides. The Pt-MS/ZnIn₂S₄ (MS = Ag₂S, SnS, and CuS) photocatalysts displayed enhanced activities for hydrogen evolution under solar irradiation. In particular, the H₂ evolution rate greatly increased to 200 μmol·h^-1 after loading 1.0 wt.% of CuS on ZnIn₂S₄. In this CuS/ZnIn₂S₄ sample, the formation of CuS (copper monosulfide) could be evidenced by XPS analysis results shown in Figure S5 (Additional file 1). The survey scan spectrum (Figure S5A of Additional file 1) indicated the presence of Cu, Zn, In, and S in the sample 2131. The binding energies shown in Figure S5E (Additional file 1) for Cu 2p_3/2 and Cu 2p_1/2 were 952.5 and 932.5 eV, respectively, which are close to the reported value of Cu²⁺31. The actual molar ratio of Cu:Zn:In:S was 0.011:0.2:0.39:1.01 as confirmed by XRF analysis result, with weight content of CuS calculated to be 1.15 wt.%, which is quite close to the proposed stoichiometric ratio. The photocatalytic activities for hydrogen evolution over Pt-MS/ZnIn₂S₄ (MS = Ag₂S, SnS, and CuS) in the initial 20-h period were measured to increase in the order of SnS <Ag₂S <CuS. Generally, these transition-metal sulfides (SnS, Ag₂S, and CuS) alone are not photocatalytically active for H₂ evolution, as no H₂ was detected when they were used as the catalysts. Thus, the improvement of photocatalytic performances of Pt-MS/ZnIn₂S₄ (MS = Ag₂S, SnS, and CuS) can be related to the enhanced separation of photo-generated electrons and holes induced by the hybridization of MS with ZnIn₂S₄. In this photocatalysis system, transition-metal sulfides (MS = Ag₂S, SnS, and CuS) combined with noble-metal Pt acted as dual co-catalysts for photocatalytic hydrogen evolution. However, when transition-metal sulfides (MS = CoS, NiS, and MnS) were loaded on ZnIn₂S₄, the rates of H₂ evolution over Pt-MS/ZnIn₂S₄ (MS = CoS, NiS, and MnS) were sharply decreased. Instead of the role as effective co-catalysts, these transition-metal sulfides (i.e., CoS, NiS, and MnS) may work as the recombination center of photo-generated electron-hole pairs, which lowered the photocatalytic activity for hydrogen evolution over Pt-MS/ZnIn₂S₄ (MS = CoS, NiS, and MnS). Further investigation is needed and also under way to provide enough supporting information to evidence the negative effects of CoS, NiS, and MnS, although main attention has focused on the more effective co-catalysts such as Ag₂S, SnS, and CuS in the following discussion.

Figure 2 shows the reaction time depended H₂ evolution over Pt-loaded MS/ZnIn₂S₄ (MS = Ag₂S, SnS, and CuS) under solar irradiation. Pt-SnS/ZnIn₂S₄ and Pt-CuS/ZnIn₂S₄ exhibited stable activity in the period of 34-h experiment. However, the rate of H₂ production over Pt-Ag₂S/ZnIn₂S₄ had a significant drop after irradiation for approximately 20 h. This deactivation may result from gradual reduction of Ag₂S particles loaded on the surface of ZnIn₂S₄ to metallic Ag by photo-generated electrons during the reaction. Similar deactivation of photocatalyst was previously observed for CdS modified with Ag₂S 32. However, this result is quite different from our previous report on Pt-Ag₂S/CdS, in which the high dispersion of Ag₂S in the nanostructure of CdS contributed to stable photocatalytic activity for hydrogen evolution 33. Taking into account the reduction potential (vs. normal hydrogen electrode (NHE)) of Ag⁺/Ag (0.80 V), Cu²⁺/Cu (0.34 V), and Sn²⁺/Sn (-0.14 V), reduction of Ag₂S by photo-generated electrons is easier than photoreduction of CuS and SnS. Therefore, Pt-MS/ZnIn₂S₄ (MS = SnS and CuS) turned to be more stable than Pt-Ag₂S/ZnIn₂S₄ during the photocatalytic reaction for hydrogen evolution.

Table 1 shows the dependence of photocatalytic activity for H₂ evolution over Pt-loaded MS/ZnIn₂S₄ (MS = SnS and CuS) on the loading amount of transition-metal sulfides. With the increase of SnS-loading from 0 to 2.0 wt.%, the rate of H₂ evolution over Pt-SnS/ZnIn₂S₄ does not show significant changes. In contrast, the photocatalytic performance of Pt-CuS/ZnIn₂S₄ depends strongly on the amount of CuS-loading, and the optimum loading amount of CuS is approximately1.0 wt.%. The progressive regression of photocatalytic activity observed with the amount of CuS increasing from 1.0 to 2.0 wt.% may be due to the fact that excess CuS particles loaded on the surface of ZnIn₂S₄ could act as the optical filter or charge recombination center instead of co-catalyst for charge separation 1932.

To visualize hybridization of CuS with ZnIn₂S₄, ZnIn₂S₄, and CuS/ZnIn₂S₄ photocatalysts were investigated by TEM. A representative TEM image of ZnIn₂S₄ is shown in Figure 3A, which shows the formation of microspheres, 1-2 μm in diameter and comprised of a circle of micro-petals. The ED pattern (inset of Figure 3A) substantiates that the ZnIn₂S₄ microsphere is of a hexagonal phase. The TEM image in Figure 3B shows that some nanoparticles are loaded on the surface of ZnIn₂S₄ microspheres. Such nanoparticles were confirmed by the ED pattern (inset in Figure 3B) to be CuS with typical orthorhombic structure. Thus, nanosized CuS particles dispersed on the ZnIn₂S₄ surface would act as the charge-transfer co-catalyst, together with photodeposited Pt particles. The Pt-CuS dual co-catalysts improved the charge separation and therefore increased the photocatalytic activity.

Figure 4 illustrates the process of photo-generated charge transfer for photocatalytic hydrogen evolution over Pt-CuS/ZnIn₂S₄ in an aqueous solution containing Na₂SO₃/Na₂S under simulated sunlight. Band gap excitation produces electron-hole pairs in ZnIn₂S₄ particles. The excited electrons are subsequently channeled to Pt sites, which reduce protons to generate hydrogen. On the other hand, the valence band potential of ZnIn₂S₄, deduced from the conduction band potential (0.29 V vs. NHE) 22 and the band gap energy (2.43 eV), is about 2.72 V vs. NHE, which is more positive than the OH^-/O₂ redox potential 4. The valence band potential of CuS is less positive than the OH^-/O₂ redox potential 34. Such a difference of valence band potentials makes it possible for the excited holes to transfer from ZnIn₂S₄ to CuS to react with Na₂S/Na₂SO₃ electron donor in the solution. Therefore, Pt and CuS are supposed to act as the reduction and oxidation co-catalyst, respectively, which leads to more efficient charge separation, thus improves photocatalytic activity of Pt-CuS/ZnIn₂S₄. Similar benefits of dual co-catalysts on photocatalytic activity have been observed for CdS loaded with noble metals as reduction catalysts and noble-metal sulfides as oxidation catalysts 2930. It is noteworthy that replacing noble-metal sulfides (such as PdS) by transition-metal sulfides (such as CuS) as the co-catalysts would help lower the cost of photocatalysts for solar-hydrogen production. Moreover, seeking effective co-catalyst candidates could be expanded to other transition-metal sulfides such as FeS and SnS₂, etc. Detailed research on this subject is still an ongoing progress in our group.

In summary, a series of Pt-loaded MS/ZnIn₂S₄ (MS = transition-metal sulfides: Ag₂S, SnS, CoS, CuS, NiS, and MnS) photocatalysts were developed. It is found that Ag₂S, SnS, and CuS could enhance the photocatalytic activity of hydrogen evolution over ZnIn₂S₄ to varying degrees, while SnS, CoS, and NiS would reduce the activity. Among them, the Pt-CuS/ZnIn₂S₄ photocatalyst exhibited the most efficient and stable activity for hydrogen evolution. This can be attributed to the fact that the dual co-catalysts of Pt and CuS entrapped photo-induced electrons and holes for reduction and oxidation reaction, respectively, improving charge separation and hence the photocatalytic activity. Application of transition-metal sulfides as co-catalysts opens up an opportunity toward realizing high-efficiency, low-cost photocatalysts for solar-hydrogen conversion.

The authors declare that they have no competing interests.

SS carried out experiments except SEM and TEM characterization, and drafted the manuscript. XC participated in the design of the study. FR performed the TEM characterization. CXK performed the SEM characterization and improved English writing. SSM provide financial support and participated in the design and coordination of this study. LG conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.