In recent years, photocatalytic H₂ evolution from water splitting by semiconductors has been considered a promising solution for serious energy crisis and environmental issues [1-13]. Among various semiconductors, TiO₂ as a traditional photocatalyst has attracted enormous interest for photocatalytic H₂ evolution owing to its good chemical stability, low cost, and nontoxicity [14-25]. Unfortunately, bare TiO₂ exhibits low H₂-evolution activity from pure water owing to the rapid recombination of photogenerated electron–hole pairs, which severely limits its practical applications [26-32]. Recently, surface cocatalyst modification has been regarded as one of the most ideal strategies for realizing efficient photocatalytic H₂ evolution [33-39]. Noble metals (such as Pt [20, 40], Au [41, 42], Ag [43, 44], and Pd [45, 46]) as well-known cocatalysts have been extensively demonstrated to improve the photocatalytic H₂-evolution performance of TiO₂ and other photocatalysts. The enhanced activity can mainly be attributed to the fact that the noble metals can rapidly capture and transfer electrons from the semiconductor surface. In addition to rapid electron transfer, an effective interfacial catalysis reaction for H₂ production is important for achieving highly efficient cocatalyst-modified photocatalysts [36, 47, 48]. Consequently, various approaches for loading catalytic active sites on the surface of noble metals have been developed [42, 49, 50]. For example, Chio et al. [49] reported that the adsorption of SCN^– ions on the surface of Ag nanoparticles accelerated interfacial electron transfer to increase the activity of Ag in Ag/TiO₂ for photocatalytic H₂ production. Furthermore, our recent works confirmed that SCN^– and Ag₂S, which were selectively loaded on metallic Au and Ag surfaces, respectively, as efficient H₂-evolution active sites greatly promoted the interfacial H₂-evolution reaction rate [42, 50]. These reports clearly reveal that the synergistic effect of noble metals (such as Au and Ag) and interfacial catalytic active sites (such as SCN^– and Ag₂S) can greatly improve the photocatalytic H₂-evolution performance of photocatalysts (such as CdS and TiO₂).

In addition to noble metals, many studies on non-noble metals as electron-transfer mediators in photocatalytic processes have been reported. For example, low-cost and earth-abundant metals (such as Cu, Bi, Co, and Ni) have been used to modify photocatalysts [51-55]. Because almost all these metals lack interfacial catalytic active sites, methods for loading effective catalytic active sites on non-noble metals have been extensively explored [54, 56, 57]. For instance, Lv et al. [56] prepared a graphene-Cu-TiO₂ photocatalyst with Cu and graphene as cocatalysts, and the photocatalytic H₂-evolution performance of TiO₂ with the Cu-graphene synergetic cocatalyst was approximately 5 times higher than that of graphene-TiO₂. Furthermore, Jiang et al. [54] reported the successful coupling of monodisperse Bi nanoparticles with g-C₃N₄ to increase the photocatalytic performance for NO removal. In fact, a shell of amorphous bismuth oxide (BiO_x) was formed on the surface of Bi owing to the rapid oxidation of Bi when exposed to air at room temperature. As a result, the improved performance of g-C₃N₄ was attributed to the synergistic effect of Bi and BiO_x. Moreover, Wen et al. [57] demonstrated that the simultaneous loading of metallic Ni interface layers and amorphous NiS dual-layer electron cocatalysts enhanced the photocatalytic H₂-evolution performance of g-C₃N₄. The above reports strongly demonstrated a general and very effective strategy for improving the photocatalytic performance of various photocatalysts, namely, the simultaneous loading of non-noble metals (such as Cu, Bi, and Ni) and other cocatalysts (such as graphene, BiO_x, and NiS) that exhibit excellent synergistic effects. In this case, the non-noble metals acting as electron cocatalysts (or electron-transfer mediators) can rapidly transfer electrons, while the other cocatalysts function as interfacial catalytic active sites for promoting the interfacial catalytic reaction. However, very few reports on the interfacial active sites of non-noble-metal-modified photocatalysts are available, and the mechanism of the improved performance remains unclear. Therefore, it is quite important to develop new interfacial active sites and determine a reasonable mechanism for the enhanced H₂-evolution activity of metal-modified photocatalysts.

In this study, highly efficient TiO₂ photocatalysts co-modified with Ni nanoparticles as electron-transfer mediators and NiS_x as interfacial catalytic active sites (denoted as TiO₂/Ni-NiS_x) were synthesized via a two-step process including the photoinduced deposition of Ni nanoparticles on the TiO₂ surface and the subsequent formation of NiS_x by a hydrothermal reaction method. It was found that the TiO₂/Ni-NiS_x photocatalyst exhibited enhanced photocatalytic H₂-evolution activity compared with TiO₂, TiO₂/Ni, and TiO₂/NiS_x. Based on the experimental results, a synergistic effect mechanism of Ni and NiS_x was proposed to account for the improved photocatalytic performance. This work may provide new insights for the smart design and development of highly efficient photocatalytic materials for various potential applications.

All chemicals were of analytical grade. Nickel nitrate hexahydrate (Ni(NO₃)₂·6H₂O), titanium dioxide (TiO₂-P25), sodium sulfide (Na₂S), and ethanol (CH₃CH₂OH) were purchased from Shanghai Chemical Reagent Ltd. (China).

Untreated commercial TiO₂ (P25) was used as the TiO₂ photocatalyst.

The Ni-modified TiO₂ photocatalyst (TiO₂/Ni) was prepared using a photodeposition method. In a typical experiment, 50 mg of TiO₂ powder was dispersed in 80 mL of ethanol as a sacrificial reagent. To this solution, 0.5 mL of Ni(NO₃)₂/ethanol solution (3 mg mL^–1) was added, such that the amount of Ni with respect to TiO₂ was controlled at 3 wt%. After it was bubbled with nitrogen for 15 min to remove dissolved oxygen, the suspension was illuminated using four low-power LEDs (3 W, 365 nm, Shenzhen Lamplic Science Co. Ltd., China) for 120 min under stirring. Finally, the products were collected by filtration, washed three times with ethanol, and dried at 40 ℃ for 24 h to obtain the TiO₂/Ni photocatalyst.

The TiO₂/Ni-NiS_x photocatalysts were prepared via the formation of NiS_x from Ni under hydrothermal conditions. First, 50 mg of the TiO₂/Ni powder was dispersed in 15 mL of distilled water under stirring, and to this a certain amount of Na₂S solution (0.1 mol L^–1) was added. After vigorously stirring for 60 min, the resulting mixture was transferred into a Teflon-lined stainless-steel autoclave and heated at 170 ℃ for 10 h. After naturally cooling to room temperature, the sample was washed with distilled water three times and dried at 40 ℃ overnight to obtain the TiO₂/Ni-NiS photocatalysts. To investigate the effect of the NiS_x concentration on the photocatalytic performance of TiO₂/Ni photocatalysts, the amount of Ni converted to NiS_x with respect to TiO₂ was controlled at 10 wt%, 30 wt%, and 50 wt%. The resulting samples were denoted as TiO₂/Ni-NiS_x(10%), TiO₂/Ni-NiS_x(30%), and TiO₂/Ni-NiS_x(50%), respectively. For comparison, an NiS_x-modified TiO₂ (TiO₂/NiS_x) photocatalyst was prepared using a method similar to that used for the preparation of TiO₂/Ni-NiS_x, and excess Na₂S solution was added to induce the complete conversion of Ni to NiS_x.

X-ray diffraction (XRD) patterns were obtained using a Rigaku Ultima Ⅲ X-ray diffractometer (Japan) with a Cu K_α radiation source and at a scan rate (2θ) of 10° min^–1. The morphologies of various samples were observed by field emission scanning electron microscopy (FESEM, JSM-7500, JEOL, Japan), and energy-dispersive X-ray spectroscopy (EDS) results were recorded using an X-Max 50 EDS detector (Oxford Instruments, UK). The morphological features were further examined by high-resolution transmission electron microscopy (HRTEM, JEM-2100F, JEOL, Japan). Moreover, high-angle annular dark-field (HAADF) TEM images and EDS mapping were acquired using an electron microscope (Titan™ Themis 200, FEI, USA) with an energy-dispersive X-ray spectrometer fitted on the microscope. X-ray photoelectron spectroscopy (XPS) measurements were performed on an ESCALAB 250Xi system (Thermo Scientific, UK) with a Mg K_α source. All the binding energies were referenced to the C 1s peak of surface adventitious carbon at 284.8 eV. UV-vis absorption spectra were collected with a UV-vis spectrophotometer (UV-2450, Shimadzu, Japan) using BaSO₄ as a reflectance standard.

The photocatalytic H₂-evolution activity of the samples was evaluated using a method similar to that used in our previous reports [16, 50]. The experimental details were as follows. The photocatalyst (50 mg) was dispersed in 80 mL of ethanol solution (25 vol%) in a 150 mL three-necked Pyrex flask under constant stirring at room temperature. Four low-power LEDs (3 W, 365 nm, Shenzhen Lamplic Science Co. Ltd., China) placed 3 cm away from the reactor were used as the irradiation source to trigger the photocatalytic reaction. Before each experiment, the system was bubbled with nitrogen for 15 min to remove dissolved oxygen. Finally, gas (0.4 mL) was intermittently sampled through a septum, and the amount of hydrogen was analyzed using a gas chromatograph (Shimadzu GC-2014C, Japan, with nitrogen as a carrier gas) equipped with a 5 molecular sieve column and a thermal conductivity detector. The quantum efficiency (QE) was calculated according to the following equation: QE (%) = number of reacted electrons/number of incident photons x 100 = number of evolved H₂ molecules x 2/number of incident photons x 100.

Photoelectrochemical (PEC) measurements were conducted on an electrochemical workstation (CHI660E, Shanghai Chenhua, China). All the PEC measurements were performed in a 0.5 mol L^–1 Na₂SO₄ electrolyte solution using a standard three-electrode configuration (a platinum plate as the counter electrode, a standard Ag/AgCl (saturated KCl solution) electrode as the reference electrode, and the working electrode). The working electrode was prepared on fluorine-doped tin oxide (FTO) conductor glass as in our previous works [42, 50]. Linear sweep voltammetry (LSV) curves were obtained in the potential range of −0.4 to −1.4 V (vs Ag/AgCl) at a 5 mV s^–1 scan rate without light irradiation. The i–t curves were collected at the bias potential of the open circuit (approximately –0.2 V) during repeated ON/OFF illumination cycles. The electrochemical impedance spectra were collected in the frequency range of 0.001–10⁶ Hz with an ac amplitude of 10 mV at the open circuit voltage.

In this work, the steps for the preparation of TiO₂/Ni-NiS_x photocatalysts can be described as the initial photodeposition of metallic Ni nanoparticles on the TiO₂ surface and the subsequent formation of NiS_x on the Ni surface, as shown in Fig. 1. First, the TiO₂ powder (Fig. 1(A-a)) is dispersed in anhydrous ethanol to form a white suspension (Fig. 1(B-a)). After the addition of the Ni(NO₃)₂/ethanol solution, Ni(Ⅱ) ions are reduced to form metallic Ni nanoparticles and then photodeposited on the TiO₂ surface under illumination to obtain TiO₂/Ni [58] (Fig. 1(A-b)). In this case, the color of TiO₂/Ni suspension changes to gray (Fig. 1(B-b)), indicating the successful deposition of Ni nanoparticles on the TiO₂ surface. Furthermore, when the resultant TiO₂/Ni sample is redispersed in an Na₂S solution and treated in a hydrothermal autoclave at 170 ℃ under autogenous pressure, NiS_x is gradually generated on the surface of the Ni nanoparticles by the effective reaction of metallic Ni and S^2– [59] to form the TiO₂/Ni-NiS_x photocatalyst (Fig. 1(A-c)). Owing to the formation of the black-colored NiS_x phase, the final TiO₂/Ni-NiS_x sample shows an obvious gray-black color (Fig. 1(B-c)). Therefore, it is believed that TiO₂/Ni-NiS_x photocatalysts were successfully prepared using the above photodeposition and hydrothermal method.

To confirm the successful synthesis of TiO₂/Ni-NiS_x photocatalysts, the as-prepared samples were characterized by XRD, FESEM, and TEM. Fig. 2 exhibits the XRD patterns of all samples. The diffraction peaks of pure TiO₂ (Fig. 2(a)) can be indexed to anatase TiO₂ (JCPDS 21-1272) and rutile TiO₂ (JCPDS 21-1276). The TiO₂/Ni (Fig. 2(b)), TiO₂/Ni-NiS_x (Fig. 2(c–e)), and TiO₂/NiS_x (Fig. 2(f)) samples display slightly weaker peak intensities than pure TiO₂, which might be ascribed to the successful loading of Ni and NiS_x cocatalysts on the TiO₂ surface. However, the lattice and phase structures of TiO₂ are well maintained, suggesting that the loading of Ni and NiS_x cocatalysts on the TiO₂ surface has no obvious effect on the crystal structure of TiO₂. Moreover, no diffraction peak of Ni is observed for TiO₂/Ni, which may be because of the main characteristic peak of Ni at ~44.5° being shielded by the (210) peak of rutile TiO₂ at ~44.1°. Furthermore, the NiS_x-containing samples do not exhibit diffraction peaks corresponding to nickel sulfide owing to its limited amount and amorphous state.

Fig. 3 shows FESEM images of the TiO₂, TiO₂/Ni, TiO₂/Ni-NiS_x(30%), and TiO₂/NiS_x samples. The TiO₂ sample (Fig. 3(A)) exhibits an irregular morphology with a wide size range of 20–60 nm. The TiO₂/Ni sample (Fig. 3(B)) has a similar morphology to that of pure TiO₂. However, the corresponding EDS result reveals a new Ni signal in addition to Ti and O signals, indicating the successful loading of Ni. For the TiO₂/Ni-NiS_x(30%) (Fig. 3(C)) and TiO₂/NiS_x (Fig. 3(D)) samples, new S signals are observed and the amount of S increases from 0.19 at% to 0.87 at%, although these samples also show similar morphologies to those of TiO₂ and TiO₂/Ni samples. To further observe the morphology and structure of TiO₂/Ni-NiS_x, TEM images of TiO₂/Ni-NiS_x(30%) were recorded, as illustrated in Fig. 4. Fig. 4(A) and (B) reveal obvious lattice planes with spacings of ~0.352 and ~0.203 nm, which correspond to the (101) plane of anatase TiO₂ and the (111) plane of metallic Ni [15, 58], respectively. In addition, NiS_x in an amorphous state is homogeneously dispersed on the surface of metallic Ni, and Ni-NiS_x is closely connected to the TiO₂ nanoparticles. Furthermore, the elemental mapping results (Fig. 4(C–C₅)) also demonstrate the presence of Ni-NiS_x on the TiO₂ surface. It is clear that S is mainly distributed on the Ni surface and the density of Ni is obviously higher than that of S, suggesting that NiS_x has been successfully loaded on the surface of the metallic Ni nanoparticles. According to the above results, metallic Ni and NiS_x were successfully loaded on the TiO₂ surface to form TiO₂/Ni-NiS_x by the present facile strategy.

The successful loading of Ni and NiS_x on the TiO₂ surface was demonstrated by XPS, which revealed the component elements and their chemical states in the TiO₂/Ni-NiS_x photocatalyst. The XPS survey spectra in Fig. 5(A) show that the TiO₂/Ni-NiS_x photocatalyst is composed of Ti, O, C, S, and Ni. The TiO₂/Ni sample (Fig. 5(B-b)) reveals a peak at around 852.9 eV, indicating the presence of metallic Ni [58], in addition to peaks at the binding energies of 855.7 and 873.2 eV, which are attributed to Ni 2p_3/2 and Ni 2p_1/2 of Ni(Ⅱ) [15, 60], respectively. For TiO₂/Ni-NiS_x(30%) (Fig. 5(B-c)), the peak intensity of metallic Ni gradually decreases while that of Ni(Ⅱ) increases, suggesting that NiS_x is formed from Ni nanoparticles by the effective reaction of metallic Ni and S^2– under a high temperature and autogenous pressure [59]. Further investigation indicates that TiO₂/NiS_x (Fig. 5(B-d)) exhibits a lower peak intensity of metallic Ni than TiO₂/Ni-NiS_x(30%) owing to the greater extent of conversion from Ni to Ni(Ⅱ). As shown in Fig. 5(C), the peak at around 167.9 eV can be attributed to S^2– in NiS_x. With increasing amounts of NiS_x, the peak intensity of S is obviously enhanced. Therefore, the XPS results strongly demonstrate that the TiO₂/Ni-NiS_x photocatalyst was successfully prepared.

UV-vis spectra can be used to determine the optical absorption of the samples. As shown in Fig. 6, the absorption boundary of the white TiO₂ sample (Fig. 6(a)) is around 390 nm, in accordance with the band energy of TiO₂ [15, 48, 61]. Compared with pure TiO₂, the deep-gray-colored TiO₂/Ni sample (Fig. 6(b)) shows significantly enhanced visible-light absorption, which can be attributed to the absorption of metallic Ni. When Ni is partly converted into NiS_x, the color of the TiO₂/Ni-NiS_x(30%) sample (Fig. 6(c)) becomes slightly lighter than that of the TiO₂/Ni sample, leading to a decreased absorption intensity in the visible region. Once Ni is completely converted into NiS_x, the TiO₂/NiS_x sample (Fig. 6(d)) turns black and exhibits enhanced optical absorption. Based on the above results, it can be concluded that the loading of metallic Ni and NiS_x has a considerable influence on the optical absorption properties of TiO₂.

The photocatalytic performance was evaluated by detecting the H₂-evolution rate of various samples under UV-light irradiation (Fig. 7(A)). Pure TiO₂ (Fig. 7(A-a)) exhibits a low photocatalytic H₂-evolution activity with a rate of only 10.10 µmol h^–1 and a QE of 0.58%. After the surface photodeposition of metallic Ni nanoparticles, the resultant TiO₂/Ni sample (Fig. 7(A-b)) shows an improved H₂-evolution rate (27.91 µmol h^–1 with a QE of 1.59%) owing to the rapid transfer of photogenerated electrons from the TiO₂ surface to the metallic Ni nanoparticles. With the formation of NiS_x on the metallic Ni surface, the H₂-evolution performance of all the TiO₂/Ni-NiS_x samples is further improved. In particular, TiO₂/Ni-NiS_x(30%) (Fig. 7(A-d)) achieves the highest photocatalytic rate (223.74 µmol h^–1 with a QE of 12.78%), which is greater than those of TiO₂, TiO₂/Ni, and TiO₂/Ni-NiS_x (Fig. 7(A-f)) by factors of 22.0, 8.0, and 2.2, respectively. To verify the stability of H₂ evolution, a recycling test was performed with the TiO₂/Ni-NiS_x(30%) sample and the results are displayed in Fig. 7(B). After recycling four times, the TiO₂/Ni-NiS_x(30%) sample still shows steady and highly efficient H₂ evolution, indicating that this sample has good recyclability.

The above results demonstrated that the loading of Ni-NiS_x on the TiO₂ surface effectively improved the photocatalytic H₂-evolution activity of the TiO₂ photocatalyst. Hence, it is very valuable and meaningful to explore the photocatalytic H₂-evolution mechanism of TiO₂/Ni-NiS_x (Fig. 8). Pure TiO₂ usually exhibits a low photocatalytic H₂-evolution activity owing to the rapid recombination of electron-hole pairs and the slow interfacial reaction. After the deposition of highly conductive Ni on the TiO₂ surface, the TiO₂/Ni sample shows a slightly improved photocatalytic H₂-evolution activity owing to the rapid transfer of photogenerated electrons by metallic Ni. However, the TiO₂/Ni sample still shows a limited performance because of the lack of interfacial catalytic active sites for the photocatalytic H₂-evolution reaction. Interestingly, it has been reported that nickel sulfide can act as interfacial catalytic active sites for photocatalytic H₂ evolution owing to the fact that interfacial or unsaturated S atoms can efficiently capture H⁺ ions from solution [62-65]. Accordingly, loading NiS_x on the Ni surface significantly enhanced the photocatalytic activity of the resultant TiO₂/Ni-NiS_x photocatalysts. The improved performance can be attributed to the excellent synergistic effect of Ni as an electron-transfer cocatalyst and NiS_x as interfacial active sites. In this case, after the implantation of the highly conductive Ni nanoparticles between the TiO₂ photocatalyst and the NiS_x cocatalysts, the photogenerated electrons in the conduction band of TiO₂ can be rapidly transferred to metallic Ni owing to the formation of intimate Schottky junctions between metallic Ni and TiO₂. Subsequently, the collected electrons in metallic Ni with sufficiently high potentials can be driven to NiS_x for the H₂-evolution reaction because the overpotential of metallic Ni is much higher than that of NiS_x [57]. Furthermore, NiS_x functions as interfacial catalytic active sites to adsorb H⁺ ions from solution effectively and further promote the subsequent interfacial H₂-evolution reaction. Therefore, the excellent synergetic effect between metallic Ni and NiS_x on the TiO₂ surface can simultaneously achieve the effective transfer of photogenerated electrons and the interfacial H₂-evolution reaction.

To confirm that Ni and NiS_x promote the H₂-evolution reaction, the photoelectrocatalytic performance of various samples was investigated. Using LSV measurements (Fig. 9(A)), the TiO₂/Ni sample was found to show a lower overpotential than the TiO₂ sample owing to the fast transfer of electrons by metallic Ni. For the TiO₂/Ni-NiS_x(30%) sample, the overpotential dramatically decreases, suggesting that NiS_x forms better interfacial active sites than metallic Ni, thus providing excellent kinetics for efficiently reducing H⁺ to H₂. However, the TiO₂/NiS_x sample displays a slightly increased overpotential compared with the TiO₂/Ni-NiS_x(30%) sample. This difference can be attributed to the lack of an electron-transfer mediator in the TiO₂/NiS_x sample, resulting in a slow rate of electron transfer that affects the subsequent interfacial H₂-evolution reaction. In fact, efficient photocatalytic H₂-evolution materials require not only rapid electron transfer but also an effective interfacial catalysis reaction for H₂ production. Therefore, both the metallic Ni electron-transfer cocatalyst and the NiS_x interfacial active sites are important for the H₂-evolution reaction of TiO₂. Furthermore, the transient photocurrent responses (Fig. 9(B)) and electrochemical impedance spectra (Fig. 9(C)) of the TiO₂/Ni-NiS_x(30%) sample exhibit the highest photocurrent density and the smallest arc radius, respectively, among the various samples, indicating that this sample has the highest separation efficiency of photogenerated carriers. According to the above results, it can be concluded that the excellent synergistic effect of metallic Ni nanoparticles as electron-transfer cocatalysts and NiS_x as interfacial active sites in the TiO₂/Ni-NiS_x system contributes to its enhanced photocatalytic H₂-evolution activity.

TiO₂/Ni-NiS_x photocatalysts were prepared via a two-step process involving the photodeposition of Ni on the TiO₂ surface and the subsequent formation of NiS_x on the Ni surface by a hydrothermal reaction method. The resultant TiO₂/Ni-NiS_x(30%) sample achieved the highest photocatalytic activity, which was greater than those of TiO₂, TiO₂/Ni, and TiO₂/NiS_x by factors of 22.2, 8.0, and 2.2, respectively. The improved H₂-evolution performance of TiO₂/Ni-NiS_x was attributed to the excellent synergistic effect of Ni and NiS_x, where the Ni nanoparticles functioned as an effective mediator to transfer electrons from the TiO₂ surface and NiS_x served as interfacial active sites to capture H⁺ ions from solution and promote the H₂-evolution reaction. This study not only provided the promising idea of using the synergistic effect of non-noble metal cocatalysts and interfacial active sites but also highlighted the potential applications of highly efficient TiO₂ photocatalysts for energy transformation.