Metal Sulfide Photocatalysts for Hydrogen Generation: A Review of Recent Advances

Abstract

Metal-sulfide nanostructures have piqued the interest of researchers for decades due to their intriguing optoelectronic properties. Indeed, significant advances and improvements have been made in various fundamental aspects for cutting-edge applications, such as water splitting and hydrogen production. Furthermore, rising demand for low-dimensional materials due to lower material consumption and improved performance due to quantum size effects has spurred research on semiconducting metal sulfides. Consequently, size-controllable nanostructures with diverse morphologies have been fabricated and studied for potential applications. However, the photocatalytic hydrogen evolution rate is still limited mainly by fast recombination rate, poor solar energy utilization and lack of surface-active sites for H₂ reduction. This review will highlight particularly recent findings in metal-sulfide-based photocatalysts for hydrogen evolution reactions, considering the swift development and excellent research in this field. Following a brief overview of fundamental properties, we will explore state-of-the-art strategies for enhancing H₂ generation efficiencies over the pristine, heterostructured and co-catalayzed metal-sulfide photocatalysts.

1. Introduction

Increasing demand for energy consumption triggers a search for new renewable and sustainable energy sources. On the other hand, the emerging climate crisis, such as increased CO₂, NO₂, and other air pollutants, as well as global warming, steadily urges the development of new materials to cope with these problems, alongside providing high alternative energy harvesting and storage capacities. According to the United Nations (UN), energy is the dominant contributor to climate change, providing almost 60 % of total global greenhouse gas emissions [1]. Therefore, investigation of new renewable energy sources has also been included (Goal 7) in the 17 Sustainable Development Goals implemented by the UN [1]. In this regard, as a sustainable energy source, H₂ production has been the focus of increasing fundamental and applied research. In general, hydrogen can be obtained from different sources such as coal or liquid fossil fuels by employing different chemical processes, as summarized in Figure 1. Among all other methods, photocatalytic water splitting has the simplicity of using semiconducting particles in solution with incident sunlight to produce hydrogen and oxygen, as follows.

$$\begin{array}{c}{H}_{2}O\to H_2+\frac{1}{2}{O}_2\delta G^{\circ }=238kJ/mol\end{array}$$

(1)

$$\begin{array}{c}PC\stackrel{hv>Eg}{\to }P{C}^{*}+h^{+}+e^{-}\end{array}$$

(2)

$$\begin{array}{c}Wateroxidation:H_{2}O+2h^{+}\to \frac{1}{2}{O}_2+2H^{+}(+0.82VvsNHE,pH=7)\end{array}$$

(3)

$$\begin{array}{c}Waterreduction:2H^{+}+2e^{-}\to H_2(-0.41VvsNHE,pH=7)\end{array}$$

(4)

To initiate the oxidation reaction, the valence band (VB) of the photocatalyst must be more positive than the water oxidation potential (+0.82 V vs Normal Hydrogen Electrode (NHE)) (Equation (3)). Whereas, more negative conduction band (CB) in compare to hydrogen evolution potential (−0.41 V vs NHE) is required to drive the reduction reactions (Equation (4)). Ultimately, this process provides a clean and renewable source of hydrogen without producing greenhouse gases or any other harmful products. However, in practice, the simultaneous production of hydrogen and oxygen from water splitting by a single photocatalyst is a tough process since the photocatalyst must fulfill band structure and energy gap criteria. Therefore, combining co-catalysts is a commonly employed strategy for this purpose. In this regard, different catalysts, including homogeneous, heterogeneous, and hybrid, have been developed. Among all others, heterogeneous photocatalysts are known for their high chemical stability in a liquid medium, high recycling or robustness, easier recovery or separation from the reaction medium, nontoxicity, low material, and operational cost, enabling selective product formation, etc.

Semiconductor metal sulfides have recently attracted great attention due to their superior physical and chemical properties. Up to nowadays, several binary metal sulfide (MS) compounds (such as PbS, CdS, ZnS, MoS₂, SnS₂, Bi₂S₃, In₂S₃, Cu₂S, NiS/NiS₂, and CoS₂) as well as their derivatives and heterostructures have been intensively employed for versatile purposes [2,3,4,5,6,7]. In Figure 2, we have summarized the most employed metals as sulfide compounds for photocatalytic hydrogen generation. Most of these compounds have been reported with a remarkable visible light response, sufficient active sites, and appropriate reduction potential of H⁺/H₂ as an effective photocatalyst. Moreover, emerging quantum size effects enable further tunability towards fast charge transfer, enhanced excited state lifetime, etc. [8,9]. Thus, in the last decade, tremendous work has been devoted to improving solar hydrogen generation over the MS semiconductor photocatalysts [10,11,12]. Among other MS materials, ternary sulfides, such as CuInS₂, CuGaS₂ have been shown to have high selectivity and a high photocatalytic production rate, particularly when coupled to titanium dioxide (TiO₂) [13,14]. Especially, MS with reduced dimensionalities, such as 2D nanosheets, flakes, and anisotropic structures, are in the spotlight of current research. Moreover, the functionalization of low-dimensional MS structures is a growing research toward high hydrogen production [15]. For example, using metals as dopants or other compounds (i.e., see Figure 2) as co-catalyst is a prominent way of enhancing photocatalytic activity.

Metal sulfide nanostructures have been increasingly employed in hydrogen evolution reactions, especially in the last couple of years [2,16,17]. In this review article, we will discuss the recent breakthrough, mainly in the last two years, regarding nanostructured metal-sulfide photocatalysts for hydrogen evolution as an alternative energy. Following this general introduction, we will briefly discuss the fundamental properties of MS-based photocatalytic systems. This will be followed by discussions on the specific MS systems that have been the focus of very recent research regarding photocatalytic hydrogen production. In each subsection, we will present the possible and most promising routes to photocatalytic enhancement.

Metal Sulfide Photocatalysts: Design and Mechanisms

As listed in Figure 3, there are specific parameters that are important to improve photocatalytic reaction yield in addition to general environmental and industrial requirements. Certainly, the search for an excellent photocatalyst to meet all requirements has triggered tremendous research resulting in versatile new photocatalysts and photocatalytic reaction routes. Any catalytic process involves fundamental steps like adsorption, interaction, desorption, etc., which determine the overall reaction mechanism, products, and yield [18]. On the other hand, perfect parametrization of these factors enables the design of an effective photocatalyst.

Although photocatalysis based on MS semiconductors is considered an economical, environmentally benign, renewable, and clean technology, the usefulness of these photocatalysts is limited by the low solar energy utilization and the fast recombination of photoexcited electron-hole pairs. Several strategies have been developed to overcome these intrinsic limitations, that we will discuss in the following parts of this review. General strategies and the most employed combination for the related strategy for metal sulfide-based photocatalysts are listed in Figure 4. Designing new morphologies and decoration of particles are recent trends to enhance the light spectrum absorbed by the semiconducting photocatalysts. For example, hollow structures, porous morphology, nanorods (NRs), 1D nanowires (NWs), 0D quantum dots (QDs), 2D sheets, and layered structures are the focus of current research and development regarding the photocatalytic hydrogen production. In this regard, mesoporous structures have been shown to facilitate the transport of reactant and reaction products [19]. Besides that, the porous medium improves sunlight absorption due to multiple reflections of light waves in porous networks [20]. Moreover, porous structures possess a high Brunauer-Emmett-Teller (BET) surface area with more active sites, which facilitates the photocatalytic reaction. Besides enhancing solar energy utilization, photocatalytic enhancement also requires improved separation efficiency of photogenerated charge carriers, which can also be achieved by designing appropriate architectures with the help of low-dimensional nanostructures. According to Tang et al., the charge carrier recombination on the surface of semiconducting photoatalyst is ∼2 × 10⁷ fs, which is three orders of magnitude faster than chemical oxidation/reduction reactions (10¹⁰–10¹¹ fs) [21]. Therefore, single component photocatalysts are frequently ineffective for H₂ generation reaction. In order to alleviate this shortcoming combining metal sulfides with metal oxide co-catalsts has been demonstrated as a powerful approach. This type of composite photocatalyst reduces charge carrier recombination of MS while also broadening the light absorption range of wide band gap metal oxide.

Notably, size reduction improves the touching surface areas and strengthens the interparticle interactions, facilitating rapid photoexcited charge transport. Moreover, emerging quantum size effects bring about strong electronic interaction developing fast charge transfer. Also, the formed discrete states and trap states at quantum particles further improve the electron-hole separation, leading to a reduced charge recombination rate [6,22,23,24]. The robustness and recycling cycles are being improved with more rugged materials and protective surface coatings. Other than these, coupling the MS photocatalysts with plasmonic metallic structures has been intensively investigated over the last few years. When noble metal (e.g., Au, Ag, Pt) NPs are attached to the semiconductor nanostructures, the photogenerated electrons in the CB of semiconductor can transfer to the adjacent metals due to their lower Fermi energy level.

The Schottky barrier formed at the metal/semiconductor interface promote electron/hole separation in photocatalysis, while the metal act as electron sinks. The photoinduced the electrons in noble metals and holes in the VB of semiconductor participate in subsequent redox reactions. The whole process increases charge carrier lifetime and improves photocatalytic activity. The unique localized surface plasmon resonance (LSPR) of noble metal NPs can boost photocatalytic activity by increasing incoming photon absorption. Furthermore, the facile tunability of LSPR (i.e., by adjusting size and shape) allows for a broad spectrum of light absorption.

Another useful parameter to achieve higher hydrogen evolution rate (HER) is a high active crystallographic orientation or facet. In this context, zinc-blende CdS nanocubes with a well-controlled charge flow and selective transfer of photogenerated electrons and holes to the (111) and (100) facets, respectively, demonstrated better charge carrier separation and significant photocatalytic H₂ production by water splitting [25].

Usual high-performance photocatalytic materials with large band gaps, such as TiO₂, ZnO, and others, absorb light exclusively in the near-UV region ($\lambda$ = 400 nm). Most low-bandgap photocatalysts, such as CdS, on the other hand, retain photoactivity for short time due to photocorrosion, which occurs when the sulfur ions are attacked by photogenerated holes and oxidized to elemental sulfur. Also, early researches on metal chalcogenide photocatalysts revealed that hole transport plays a limiting role in these nanostructures. To tackle this shortcoming, the isolation of photoinduced holes from the reaction site has been suggested [26]. For example, in Pt/CdS-CdSe heterostructures, the holes are confined to the CdSe, whereas the delocalized electrons are transferred to the Pt tip. This was due to larger VB edge of CdSe than CdS, and consequently resulted in significant photocatalytic activity [11]. Compared to photocatalysts made up of only one type of semiconductor, heterojunction structures often possess altered band alignment at the interface, enhancing carrier separation and limiting carrier recombination for enhanced photocatalytic activity. Sulfide atoms near metal ions in typical metal sulfides potentially act as a catalytic site for the attachment and detachment of H atoms. Therefore, the increased number of sulfur sites is another parameter for enhancing photocatalytic activity. Noteworthy, compared to crystalline metal sulfides, amorphous metal sulfides often exhibit more unsaturated sulfur atoms with excessive chemical reactivity, functioning as active sites for H₂ production. In Figure 5, the band gap energy with respect to the water oxidation and reduction potential is summarized for representative semiconductor metal sulfides.

Morphological modifications and the construction of heterojunction photocatalysts are the two most employed techniques to increase the performance of photocatalytic hydrogen evolution significantly. The close distance between the two nanosheet subunits induces a strong interaction, facilitating the separation and transfer of charge carriers. In this regard, a reverse cation exchange-mediated and a ${\mathrm{Cu}}_{2-x}$S nanocube template-assisted growth strategy for fabricating hollow multinary metal sulfide was reported recently [41]. Thus, it was found that in contrast to the typical cation exchange mechanism mediated by the metal sulfide constant, the presence of tri-n-butylphosphine (TBP) can reverse cation exchange to produce a series of hollow metal sulfides. Various hollow metal sulfide cubic NSs were demonstrated using this method, including binary compounds (CdS, ZnS, Ag₂S, PbS, SnS), ternary compounds (CuInS₂, Zn_x${\mathrm{Cd}}_{1-x}$S), and quaternary compounds (single-atom platinum attached Zn_x${\mathrm{Cd}}_{1-x}$S; Zn_x${\mathrm{Cd}}_{1-x}$S-Pt₁). Also, a light-induced deposition approach to precisely load nickel sulfide nanodots onto the surface of g-C₂N₂ resulted in both enhanced photocatalytic hydrogen evolution and oxygen reduction performances over composite photocatalysts [42]. In general, as demonstrated in Equations (1)–(4), photocatalytic water splitting involves three significant steps: the generation of electron-hole pairs by photon absorption, the separation of electron-hole pairs and migration to the photocatalyst surface, and the splitting of water molecules via charge carriers on the photocatalyst surface.

One should emphasize that depending on the selected co-catalyst or heterojunction, the reaction pathway follows different routes, which are mainly dictated by the chemical potential of the components. In this regard, different routes have been reported in the literature, which are also valid for metal sulfide photocatalysts. The most pronounced photocatalytic reaction types are illustrated in Figure 6. In type-I heterojunction, a built-in electric field is established by a photogenerated charge carrier inside the large band gap semiconductor before it is transmitted into the narrower band gap semiconductors; thus, electron-hole pairs cannot be spatially separated. On the contrary, in the type-II heterojunctions between two semiconductors, rapid generation of a built-in electric field induces a prolonged lifetime of photogenerated charge carriers. Therefore, in engineering a photocatalytic system for enhanced charge carrier separation in heterojunction, facilitating a suitable type-II interfacial contact with effective carrier migration pathways is critical (i.e., see Figure 6). For instance, Liu et al. synthesized a type-II hexagonal cadmium sulfide (CdS)/molybdenum disulfide (MoS₂) heterostructure for photocatalytic hydrogen production [43]. The synergetic enhancement in the charge dynamics of the excited carriers on the heterostructure interface of the CdS on the MoS₂ film was investigated by transient absorption spectroscopy. Type-III heterostructures are called Van der Waals broken band gap heterojunctions, which are less investigated in applying two-dimensional (2D) photocatalytic materials for hydrogen production. That is because the photogenerated charge carriers are not synergistically coupled; thus, electron-hole pairs exhibit a fast decay. In the S-scheme heterojunction, strong photogenerated holes and electrons in the valence and conduction bands, respectively, participate in the redox reactions. Meanwhile, weak photogenerated carriers are recombined by the internal electric field and Coulomb gravity. As a result, more efficient electrons for photocatalytic hydrogen reduction can be used [44]. In all-solid- state Z-scheme photocatalysts, NPs such as noble and transition metals (Ag, Au, Rh, Ru, Cu, Pt, etc.) are used as electron conductors/mediators between two SCs due to their strong reduction ability. This type of heterojunction has been studied extensively over the last decade with different combinations of the SCs after the pioneering work of the anisotropic CdS-Au-TiO₂ nano junction SCs [45]. The direct Z-scheme photocatalytic process, which involves two steps of photoexcitation, mimics the natural photosynthesis process, which also requires the spatial isolation of photogenerated electrons and holes. By pooling photogenerated electrons, an electron-rich region forms in the CB of SC I, and by pooling photogenerated holes, a hole-rich region generates in the VB of SC II, which reduces bulk electron-hole recombination in this type of heterojunction. Meanwhile, the photogenerated electrons in the CB of PC II and holes in the VB of PC I with less redox power recombine.

2. Hydrogen Generation as an Energy Material

2.1. Binary Metal Sulfide Compounds

Semiconductors, with a large band gap energy, can only be activated under UV irradiation, which limits their efficiency in terms of using a narrow light spectrum. CdS, with its energy level of CB being more negative than the potential of water reduction, is one of the most suitable semiconductors for water splitting by utilizing visible-light irradiation [46,47,48,49,50,51]. However, a general drawback, photocorrosion, is also related to CdS. Thus, sulfur anions (${\mathrm{S}}^{2-}$) tend to be oxidized easier than water at the surface of the anode. Ultimately, hole scavengers, which can easily provide electrons reacting with holes such as alcohol, ${\mathrm{S}}^{2-}$, ${{\mathrm{SO}}_3}^{2-}$, etc., are required to react with h⁺ to neutralize the reaction for continuous H₂ evolution. Thus, introducing CdS as the main photocatalyst or as a component in different chemical compounds is one of the effective ways to improve its usefulness for water splitting. For example, it is known that nickel boride (Ni_xB), which is a well-known low-cost photocatalyst, suffers from the weakening of atomic H adsorption on Ni active site by the electron transfer from B to Ni atoms, which causes the inhibition of H₂ evolution. In order to enhance the photocatalytic H₂ generation over CdS, Zhang et al. have prepared a Ni_xPB-rGO/CdS hetero-photocatalyst, where the Ni_xB was doped by metalloid P to enhance its H adsorption ability [52]. The 8-10 nm Ni_xPB are loaded on reduced graphene oxide (rGO) and coupled with CdS nanoparticles. The as-synthesized photocatalysts exhibited H₂ evolution rate of 5.79 mmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, being 1.8 and 9.9 times higher than the Ni_xB-rGO/CdS and rGO/CdS photocatalysts, respectively. Similarly, CdS@Ni₂P was prepared by a simple hydrothermal method followed by in situ photodeposition [53]. The highest photocatalytic hydrogen production rate achieved with this composite was 287 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, that is ∼98.3 times that of pristine CdS. The higher activity was related to the effective interface between the Ni₂P co-catalyst and CdS nanorods, which leads to better separation and transport of photogenerated electrons.

For solar-driven water splitting, NiWO₄/CdS nanosheets-on-nanorods direct Z-Scheme heterostructure was developed by using a facile in-situ approach (see Figure 6 for Z-scheme photocatalysts) [54]. The fabricated NiWO₄/CdS heterostructure photocatalysts showed a H₂ evolution rate of 26.43 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, which is 75 times higher than that of bare CdS. It was shown that the nanostructures with multiple active sites, intimate contact interface, and improved charge separation rate are responsible for the enhanced photocatalytic performance.

Another strategy to improve the photocatalytic activity of semiconductor materials is employing metalloid compounds, which form a Schottky junction with the semiconductor at the heterogeneous interface. In this regard, tungsten nitride (WN) nanosheets, which have excellent electron conductivity, were employed as a new noble-metal-free co-catalyst to improve the H₂ evolution performance of CdS NPs via the one-pot solvothermal method [55]. The H₂ production rate of the WN/CdS composite photocatalyst under visible light irradiation reached 24.13 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, which is 9.28 times that of pure CdS NPs. Similarly, P-rich platinum diphosphide (PtP₂) nanodots were used to improve the photoactivity of CdS by a facile phytic acid-assisted pyrolysis [56]. The obtained PtP₂@C/CdS(10 wt%) photocatalyst presents an optimum H₂ evolution rate of 9.76 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with apparent quantum efficiency (AQE) of 41.67% (420 nm), which is 2.2 and 34.8 times higher than that of Pt@C/CdS and clean CdS, respectively. It was found that the electron-enriched P-rich PtP₂ can serve as the effective H-proton adsorption active site to enhance the photocatalytic activity greatly.

As discussed previously, designing porous photocatalysts is one of the major trends for HER enhancement. In this perspective, three-dimensional ordered macroporous (3DOM) structures have received increased attention because their open interconnected hierarchical macroporous architecture facilitates the transport and diffusion processes of species, leading to enhanced photocatalytic enhancement. Furthermore, the regularly ordered structure of 3DOM results in photon band gaps and the slow photon effect of photonic crystals, which can significantly improve the photoinduced carriers and boost photocatalytic activity. Additionally, changes to the pore walls typically do not modify the 3DOM structure, making it easy to build efficient photocatalysts with diverse materials. In this regard, recently, a CdS photocatalyst with a 3DOM structure using polymethylmethacrylate as a template was reported [57]. In line with the expectations, photocatalytic H₂ evolution rate of 20.1 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ was achieved, which is remarkably improved in comparison with the CdS NPs. Moreover, as shown in Figure 7f, HER was further enhanced by loading Pt metal. The schematic view of the preparation procedure, corresponding structural characterization, HER results, and proposed H₂ evolution mechanism over CdS 3DOM photonic crystals are summarized in Figure 7.

Also, a hybrid heterojunction consisting of CdS NPs and structurally defined graphene nanoribbons (GNRs) was reported with an enhanced HER [58]. Through photocurrent, electrochemical impedance, steady-state photoluminescence, and time-resolved fluorescence decay spectroscopy, it was revealed that the CdS/GNR heterojunction could efficiently reduce the photoexcited electron-hole pair recombination. Notably, the CdS/GNR hybrid displayed H₂ evolution rate up to 22.4 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ under Vis-light irradiation ($\lambda$ = 420 nm), which is much greater than that of individual CdS and GNRs. Lei et al. employed the facile solid-liquid mixing method to load CdS onto WC [59]. 3 wt% WC-loaded CdS photocatalyst (CdS/WC) exhibiting the highest HER of 9180 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 14.3% at 420 nm. Thus, incorporating WC improves the light absorption capability of CdS NPs. Moreover, a more conductive WC provides a fast electron transfer channel and water reduction site, which enhances the spatial photogenerated charge separation while facilitating the H₂ production.

Alongside metal sulfide-based heterostructures, doping the metal sulfides with metals is also a promising trend toward photocatalytic H₂ generation. For example, Ni_xCd_yS composite photocatalysts possessing porous structure, large specific surface, and highly dispersed metal active sites were synthesized using Ni-doped CdS [60]. The fabricated Ni_xCd_yS photocatalysts showed three times higher H₂ yield (i.e., 8.45 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$) than the CdS. ${\mathrm{NiS}}_{1+x}$ compounds have recently gained interest as low-cost transition-metal chalcogenides for effective photocatalytic H₂ evolution [61,62]. Co-catalyst-free NiS ($\alpha$ and $\beta$ phases) nanostructures for photocatalytic H₂ generation was reported [63]. Unfortunately, due to the limited number of exposed sulfur sites and their strong interaction with adsorbed hydrogen atoms, these materials are ineffective for H₂ generation. To overcome these limitations, Gao et al. developed a nanostructured core-shell ${\mathrm{Au@NiS}}_{1+x}$ co-catalyst and adjoined with TiO₂ by employing a two-step route [61]. The synthetic route for producing the core shell structures, as well as their characterisation, is shown in Figure 8a–g. The optimized TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.7:1.3) exhibited high HER of 9616 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 46.0% at 365 nm, which is 2.9 and 1.7 times higher than the HER obtained for the TiO₂/${\mathrm{NiS}}_{1+x}$ and TiO₂/Au photocatalysts, respectively. By XPS studies complemented by DFT calculations, it was shown that the conductive-electrons of Au can transfer to sulfur-enriched ${\mathrm{NiS}}_{1+x}$ resulting in electron-enriched active sulfur centers. This weakens the strong bonds between sulfur and adsorbed hydrogen atoms and facilitates desorption of adsorbed hydrogen for rapid hydrogen formation. Figure 8 depicts the photocatalytic performance as well as the associated charge transfer process.

Besides the aforementioned compounds, Fe₃S₄ (greigite), which is a ferromagnetic material, has also recently been extensively investigated for energy applications, including high-energy batteries, electrochemical hydrogen storage because of its low-cost, and environmental benignity and natural abundance. In addition to the sulfur ions acting as H-adsorption/desorption sites, the mixed ${\mathrm{Fe}}^{2+}$/${\mathrm{Fe}}^{3+}$ valence states in the inverse spinel structure also endow this material with relatively higher electric conductivity than the FeS and FeS₂, which makes it more suitable photocatalysts for H₂ generation. However, Fe₃S₄ crystal possesses inappropriate surface adsorption energy for H atoms, which can be altered by doping with additional metal atoms. In this regard, Fe₃S₄ nanorods were derived from a metal-organic framework by a facile method and its HER capability was further enhanced by Ni doping [64]. Thus, the organic dye (eosin-Y) sensitized, 1%Ni-doped Fe₃S₄ nanorods showed remarkably improved H₂ generation rate of 3240 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 12% under $\lambda$ = 500 nm. Due to its unique localized surface plasmon resonance in the near-IR region, the ${\mathrm{Cu}}_{2-x}$S has showeds substantial promise in the field of broad-spectrum driven photocatalyst engineering. Nonetheless, inefficient carrier separation and severe photocorrosion limit its practical uses.

Figure 8. Schematic illustration of the synthesis procedure of the core-shell ${\mathrm{Au@NiS}}_{1+x}$ combined TiO₂ photocatalyst (a) and the formation mechanism of ${\mathrm{NiS}}_{1+x}$ on Au surface (b). (c) shows the corresponding photographs for (1) TiO₂, (2) TiO₂/Au, (3) TiO₂/Au@Ni-TU (TU stands for thiourea), and (4) TiO₂/${\mathrm{Au@NiS}}_{1+x}$. (d) TEM and (e) HRTEM images of the TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.7:1.3) photocatalyst. (h) Schematic illustration for activating H-adsorbed S on Au-mediated ${\mathrm{NiS}}_{1+x}$ by forming electron-enriched active S sites. In situ irradiated XPS spectra of S 2p (f) and Ni 2p (g) for TiO₂/${\mathrm{Au@NiS}}_{1+x}$ in dark and under UV-light illumination. (h) Schematic view of charge transfer within TiO₂/${\mathrm{Au@NiS}}_{1+x}$: (i) before contact, (j) contact in the dark and (k) during UV irradiation: the (1) photoelectron trapping by the Au mediator, (2) photoelectron transfer from Au to ${\mathrm{NiS}}_{1+x}$, and (3) fast H₂ evolution on the electron-enriched S active sites of ${\mathrm{Au@NiS}}_{1+x}$ co-catalyst. (l) Photocatalytic HER over (1) TiO₂, (2) TiO₂/${\mathrm{NiS}}_{1+x}$, (3) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(2.5:0.5), (4) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(2:1), (5) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.7:1.3), (6) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.5:1.5), (7) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1:2), and (8) TiO₂/Au. (m) Recycling test for hydrogen evolution over the TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.7:1.3) photocatalyst. [61] Copyright 2022, Wiley-VCH.

Figure 8. Schematic illustration of the synthesis procedure of the core-shell ${\mathrm{Au@NiS}}_{1+x}$ combined TiO₂ photocatalyst (a) and the formation mechanism of ${\mathrm{NiS}}_{1+x}$ on Au surface (b). (c) shows the corresponding photographs for (1) TiO₂, (2) TiO₂/Au, (3) TiO₂/Au@Ni-TU (TU stands for thiourea), and (4) TiO₂/${\mathrm{Au@NiS}}_{1+x}$. (d) TEM and (e) HRTEM images of the TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.7:1.3) photocatalyst. (h) Schematic illustration for activating H-adsorbed S on Au-mediated ${\mathrm{NiS}}_{1+x}$ by forming electron-enriched active S sites. In situ irradiated XPS spectra of S 2p (f) and Ni 2p (g) for TiO₂/${\mathrm{Au@NiS}}_{1+x}$ in dark and under UV-light illumination. (h) Schematic view of charge transfer within TiO₂/${\mathrm{Au@NiS}}_{1+x}$: (i) before contact, (j) contact in the dark and (k) during UV irradiation: the (1) photoelectron trapping by the Au mediator, (2) photoelectron transfer from Au to ${\mathrm{NiS}}_{1+x}$, and (3) fast H₂ evolution on the electron-enriched S active sites of ${\mathrm{Au@NiS}}_{1+x}$ co-catalyst. (l) Photocatalytic HER over (1) TiO₂, (2) TiO₂/${\mathrm{NiS}}_{1+x}$, (3) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(2.5:0.5), (4) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(2:1), (5) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.7:1.3), (6) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.5:1.5), (7) TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1:2), and (8) TiO₂/Au. (m) Recycling test for hydrogen evolution over the TiO₂/${\mathrm{Au@NiS}}_{1+x}$(1.7:1.3) photocatalyst. [61] Copyright 2022, Wiley-VCH.

Recently, multinary compounds based on Cu_xS synthesized by a facile colloidal method were also reported as effective photocatalysts for the H_x evolution [65]. There, the authors incorporated Ni, Co, Mn, Zn metals to enhance activity of Cu_xS and reaction rates of H_x generation was found in the order of Ni:Cu_xS > Co:Cu_xS > Mn:Cu_xS > Zn:Cu_xS as 4.0, 1.2, 0.9 and 0.7 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, respectively. The higher efficiency of the Ni-doped Cu_xS nanorods can be attributed to the effective charge carrier separation and fast transport similar to the Ni-CdS, as discussed above. Molybdenum disulfide (MoS₂) is another potential low-cost photocatalyst for hydrogen production reactions. According to the previous studies, the active sites for the catalytic reaction are mainly associated with the edge sites of 2D-layered MoS₂, however, the basal plane shows poor activity toward catalytic reactions. Thus, activating the basal planes is another route for enhancing its HER. In this regard, Ni-doping and exfoliation were reported to convert the in-plane sites to active edge sites [66]. Furthermore, the activated MoS₂ was utilized for enhancing the photocatalytic activity of CdS toward H₂ production. The resulting Ni-MoS₂/CdS nanocomposite photocatalysts showed a H₂ evolution rate of 249 mmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, which is 70 times higher than pristine CdS NPs.

2.2. Ternary Metal Sulfide Compounds

Bimetals sulfides display more outstanding HER performance during continuous catalysis due to their unique bimetals synergistic effect and adjustable potential. Zinc indium sulfide (ZnIn₂S₄), one of the emblematic AB₂S₄ ternary metal chalcogenides, is a layered structure n-type semiconductor with a direct band gap. ZnIn₂S₄ can exist in three different crystal polymorphs (cubic, trigonal, and hexagonal), with the hexagonal phase being the most thermodynamically stable structure.

Tao et al. employed the wet-chemical method for the preparation of ultrathin MIn₂S₄ (M = Zn, Cd, and Ni) nanoplates with a thickness of 2 nm, using as-grown $\beta$-I₂S₃ nanoplates as a template [67]. Further optimization showed that loading of Ni NPs as co-catalysts, the 3.0 wt% Ni/ZnIn₂S₄ photocatalysts exhibited the highest H₂ evolution rate of 19.9 mmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, which is six times larger than the ZnIn₂S₄ NPs. Recently, ZnIn₂S₄ monolayers with dual defects (Ag dopants and nanoholes) for improved HER were also reported [68]. While the undoped ZnIn₂S₄ could produce HER of 32.6 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, the Ag-ZnIn₂S₄ showed 56.6 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$. Most importantly, the Ag-doped ZnIn₂S₄ could achieve 29.1 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ O₂ evolution, which was not seen for the undoped ZnIn₂S₄. The proposed mechanism and the obtained reaction yield are given in Figure 9a.

In order to overcome the intrinsic limitations of the MoS₂, Lei et al. prepared a cobalt-modified amorphous molybdenum sulfide (Co-MoS_x) co-catalyst [69]. The optimized ZnCdS/Co-MoS_x (Co/Mo = 1/4, 1 mol%) photocatalyst showed the highest HER rate of 551.48 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 21.7% at $\lambda$ = 420 nm. The achieved rate was superior to that of ZnCdS/MoS_x and pristine ZnCdS by 1.4 times and 23 times, respectively.

As stated above, the photocatalytic activity of NiS towards hydrogen generation is practically limited by the strong sulfur-hydrogen bonds. The weakening of this bond strength was achieved by introducing Cu atoms into the amorphous NiS_x (a-NiS_x) co-catalyst, which led to the formation of the electron-rich active S sites [70]. The as-synthesized co-catalyst was anchored onto TiO₂ to form a-NiCuS_x/TiO₂ composite photocatalysts. The optimized CuS_x/TiO₂ (3:1) photocatalyst exhibited an improved H₂ evolution rate of 427.9 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 34.67% using methanol as a sacrificial reagent, which is superior compared to the a-NiS_x/TiO₂ with 226.1 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. Xu et al. reported a CuGaS₂ (CGS) photocatalyst synthesized by a one-step solid-phase sintering method [71]. It was shown that when the Cu to Ga atom ratio is 1: 1.4 in the raw material, the prepared CuGaS₂ exhibited the highest visible-light ($\lambda >$ 420 nm) H₂ production rate reaching to 1.12 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. Moreover, loading with Ruthenium as a co-catalyst, the H₂ production rate of CuGaS₂@Ru was further promoted to 3.38 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. Wang et al. prepared CuInS₂ microflowers via the solvothermal method and decorated them with CdSe QDs on the surface. The prepared CdSe/CuInS₂ microflowers showed an improved photocatalytic H₂ production rate of 10,610.37 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 48.97% at 420 nm [72]. The enhanced photocatalytic H₂ evolution was related to the formation of a p-n heterostructure with improved light absorption ability, enhanced charge transfer efficiency, and a reduced recombination rate of the photogenerated electrons and holes. Also, the one-step alkaline hydrothermal strategy was used to prepare Ni-doped ${\mathrm{Mn}}_{0.5}{\mathrm{Cd}}_{0.5}$S with different Ni doping concentrations [73]. It was shown that 3% Ni-doped ${\mathrm{Mn}}_{0.5}{\mathrm{Cd}}_{0.5}$S could obtain the highest rate of 108.3 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ at $\lambda >$ 420 nm, which is nearly 14.4 times more than undoped counterpart (7.5 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$). The proposed charge transfer mechanism, together with the obtained reaction yield, is shown in Figure 9b.

Engineering sulfur vacancies supplemented by metal doping have been shown as a promising route to improve photocatalytic H₂ evolution. In this regard, ultrathin ZnIn₂S₄ nanosheets functionalized by sulfur vacancies and nickel dopants were reported [74]. It was shown that the S vacancies lower the CB minimum, enhance the photogenerated charge reduction capacity, and improve the lifetime of photoexcited states. Moreover, introducing Ni dopants shifts the Fermi level upward, leading to enhanced charge carrier separation and a reduced recombination rate. Thus, the synergistic effects lead to an 8.91 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ H₂ evolution rate under Vis-light irradiation. The proposed charge transfer mechanism, together with the obtained reaction yield, is shown in Figure 9c. In order to improve photocatalytic hydrogen yield, the biomineralization method has also been recently employed. Thus, Zn-doped CdS (${\mathrm{Cd}}_{1-x}$Zn_xS) with various ${\mathrm{Zn}}^{2+}$/${\mathrm{Cd}}^{2+}$ ratios was synthesized by engineered $Escherichiacoli$ with a surface-displayed biomineralization system [75]. More specifically, bacterial-synthesized ${\mathrm{Cd}}_{1-x}$Zn_xS with an average particle size of <50 nm was deposited on the bacterial cell surface, which led to more material loading. The optimized ${\mathrm{Cd}}_{1-x}$Zn_xS with a 0.044 molar ratio of Zn/Cd showed an improved HER of 1.146 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$.

As already given examples above, one of the main trends to improve photocatalytic H₂ production is a search for new architectures. For example, hollow Cu₂MoS₄/ZnIn₂S₄ heterostructural nanocubes with intimate-contact interface were reported as a self-template method [76]. The optimized 15 wt% Cu₂MoS₄/ZnIn₂S₄ sample exhibits the highest hydrogen production rate of 8103 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, which is four-fold improvement compared to the clean ZnIn₂S₄. The improved activity was ascribed to more active sites, increased light absorption by hollow nanocubes, and a built-in electric field, which facilitates photogenerated holes from ZnIn₂S₄ to Cu₂MoS₄ with enhanced charge separation. A sphere-like, structured ZnIn₂S₄-rGO-CuInS₂ ternary heterojunction catalyst for visible-active photocatalytic hydrogen evolution from water splitting was reported [77]. The obtained photocatalyst exhibited hydrogen generation of 2531 µmol/g after 5 h. Based on photocurrent density and photoluminescence measurements, it was discussed that rGO acts as an electron transfer medium by reducing excited charge recombination and facilitating electron transfer in the ternary composite. Chong et al. reported on atomic-level charge steering for photocatalytic water splitting in P-doped ZnIn₂S₄ (ZIS) bilayer nanosheets via Density Functional Theory [78]. It was revealed that the replacement of the S3 atom by substitutional phosphorus gave rise to the most stable P-ZIS structure. Furthermore, P-ZIS was shown to have a lower band gap energy, an upshift of the valence band maximum (VBM), an increase in electron density at the VBM, and a lower H* adsorption-desorption barrier, which are essential parameters for H₂ evolution kinetics. Thus, the p-type doping strategy was shown as a promising route for improved H₂ evolution from water splitting.

2.3. Metal Sulfide and Metal Oxide Heterostructures

In order to design an efficient interface, Lin et al. produced a 2D/2D atomic double-layer WS₂/Nb₂O₅ shell/core photocatalyst (DLWS/Nb₂O₅) [79]. It was revealed that due to the unusual interface, the average lifetime of charge carriers for DLWS/Nb₂O₅ (180.97 ps) is remarkably reduced as compared to that of Nb₂O₅ (230.50 ps). Thus, the ultrafast charge transfer from Nb₂O₅ to atomic double-layer WS₂ was achieved, which gave rise to a high photocatalytic H₂ evolution rate of 237.6 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. Also, CoS/Nb₂O₅ heterojunctions were constructed by in-situ loading CoS co-catalyst on the surface of Nb₂O₅ nanosheets [80]. The femtosecond-resolved transient absorption spectroscopy revealed that the average lifetime of charge carriers for 10 wt% CoS/Nb₂O₅ (159.6 ps) is drastically shortened by contrast with that of Nb₂O₅ (5531.9 ps), thus similar to the DLWS/Nb₂O₅ rapid charge transfer from Nb₂O₅ to CoS could be achieved. The designed photocatalyst showed H₂ evolution rate of 355 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, which is up to 17.5 times higher than the pristine Nb₂O₅. By using the hydrothermal-physical mixture method, a Mo₂S₃/Bi₂O₂CO₃ composite with a unique n-n heterojunction and 2D structure was successfully prepared [81]. The optimized co-catalyst with the mass ratio of Mo/Bi-3% showed the highest H₂ evolution rate of ∼160 µmol.

Bismuth vanadate (BiVO₄) is another promising catalyst, which has been employed for solar-driven water splitting and H₂ production [82]. Recently, a step-scheme photocatalyst was developed by spatial deposition of MnO_x and CdS NWs on a 3D decahedron BiVO₄ surface [83]. The synthesized CdS/MnO_x-BiVO₄ photocatalyst containing 5% of CdS demonstrated the highest H₂ and O₂ production rates of 1010 and 510 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, respectively, with an AQE of 11.3% without any sacrificial reagent. The existence of oxygen vacancies, effective charge separation and transport, and a significant interaction of co-catalysts (MnO_x and Pt) with photocatalysts (BiVO₄ and CdS) in the CdS/MnO_x-BiVO₄ heterojunction were shown to be the major factors for boosted photocatalytic activity. It should be noted that especially for simultaneous O₂ generation, the single MS photocatalysts are not favorable because of the severe self-photooxidation or photocorrosion, which, however, can be altered by combining with other co-catalysts. In this regard, the development of MS-based Z-scheme photocatalyst, where the photogenerated holes in MS can be successfully carried away, is an effective approach (see Figure 6 for the Z-scheme photocatalysts) [84]. Heterostructure composite photocatalyst was synthesized by combining the Nb₂O₅-MoS₂ (NM) with CdS/graphene and CuS/graphene (NM-CdS/G, NM-CuS/G) via hydrothermal method [85]. While the NM photocatalyst showed a relatively low photocatalytic activity with a H₂ evolution rate of 1.6 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, the NM-CuS/graphene and NM-CdS/graphene composites showed remarkably enhanced H₂ evolution rates of 7.1 and 8.7 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, respectively. Thus, the combination of wide and narrow band gap semiconductors resulted in a fast charge separation process and a possibly improved lifetime of the charge carriers. Also, hierarchically porous hydrangea-like In₂S₃/In₂O₃ heterostructures were reported through a simple in situ oxidization process [86]. The new heterostructures exhibited superior photocatalytic activity (683.8 µmol/g) to the In₂S₃ and In₂O₃, that was assigned to enhanced interfacial charge transfer because of heterostructures and hierarchically porous heterostructures, which increases the number of active sites and improves the permeability. See Figure 10b for the proposed photocatalytic charge transfer mechanism and HER plot.

A Vis-light active hierarchical Co₉S₈/ZnIn₂S₄ photocatalyst was designed by decorating Co₉S₈ NPs onto the ZnIn₂S₄ nanoflower through a two-step solvothermal process and applied for photocatalytic H₂ evolution [87]. Due to sulfide-sulfide integrations in a hierarchical flower-like Co₉S₈/ZnIn₂S₄ photocatalytic heterojunction, large surface areas with rich active centers, effective separation and transfer of photoinduced electrons and holes were achieved. Correspondingly, the optimized Co₉S₈/ZnIn₂S₄ photocatalyst showed remarkably improved HER of 12.67 mmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ with high stability, which is 6 times higher than that of single ZnIn₂S₄ photocatalyst under visible-light irradiation. The proposed photocatalytic charge transfer mechanism, together with the obtained yield, is shown in Figure 10b.

Ta₂O₅ is another promising photocatalyst, which has been studied in the last years [89], also for the photocatalytic H₂ generation [90]. Recently, a Cu₂S/Ta₂O₅/CdS ternary nanocomposite photocatalyst was fabricated by a solid-state approach and applied for photocatalytic hydrogen evolution [91]. It was shown that in the resulting structure, the Cu₂S, Ta₂O₅, and CdS form six-petals, nanorods, and random-spherical morphology, respectively. The optimized Cu₂S/Ta₂O₅/CdS nanocomposite demonstrated a high rate of H₂ evolution of 131 mmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ being remarkably higher than that of its constituents.

It is known that preparing photocatalysts deposited on semiconductors can effectively restrain charge recombination. In this regard, a Cu/CdS/MnO_x (CSM) heterostructured photocatalyst with a spatial separation of the photoinduced charge carriers was reported [88]. It was shown that while a reduction co-catalyst Cu NPs tend to capture electrons, the oxidation co-catalyst, MnO_x NPs, collects the photogenerated holes. Time-resolved fluorescence spectra (TRPL) and photoelectrochemical characterization showed that the lifetime of photogenerated carriers improved, and surface-reactive HER kinetics were accelerated. As a result, the optimized 1% CSM photocatalyst showed a maximum HER of 5965.03 ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ being ∼5.3 higher than the single CdS photocatalyst. The proposed photocatalytic charge transfer mechanism, together with the obtained HER, is given in Figure 10c.

Step-scheme heterojunction composite photocatalyst based on W₁₈O₄₉/${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S was produced via the physical stirring method and employed for photocatalytic hydrogen production under Vis-light illumination [44]. It was revealed that the H₂ evolution rate was highest (5373 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$) for the load ratio of W₁₈O₄₉ was 5%. Consequently, the formation of S-scheme heterojunction between W₁₈O₄₉ and ${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S contact interface was shown as the reason for the improved H₂ production. Thus, the S-scheme heterojunction promotes the consumption of ineffective electrons and holes, improves the effective charge separation and transfer, and lowers the activation energy of the catalytic reaction system, leading to enhanced photocatalytic H₂ evolution. Also, MoS₂@MoO₃ step S-scheme heterojunction photocatalyst was prepared through in-situ partial sulfidation [92]. The heterostructure interface was engineered through the in-situ vulcanization strategy, which gradually corrodes from the outside to the inside of the nanostructures. Thus, introducing S-atoms can replace oxygen atoms to build a S-rich surface and generate molybdenum sulfide. Afterward, the amount of thioacetamide (S source) was adjusted to control vulcanization and optimize the experimental conditions. The synthesized heterostructure photocatalyst showed H₂ production rate of 12,416.8 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$.

Metal Sulfide and TiO₂-Based Photocatalysts

TiO₂ is one of the most studied and applied photocatalysts for versatile purposes ranging from wastewater treatment and indoor air purification [93] to organic synthesis [94,95]. Moreover, metal sulfides/TiO₂ heterostructures with a porous architecture possess the capability to enhance the photocatalytic activity of TiO₂. First of all, the porous structure provides an abundance of carrier charge transfer routes and surface reaction sites. Moreover, the appropriate number of MS deposited on TiO₂ significantly improves the electrical conductivity and light absorption capabilities.

Lin et al. reported a sea-urchin-like ReS₂ nanosheet/TiO₂ NP heterojunctions with an unusual charge edge-collection effect leading to a significant acceleration of charge separation and transfer [96]. The designed photocatalysts yielded H₂ evolution rate of 3.71 mmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ with an AQE of 16.09%, which is ∼232 times higher than the commercial P25 TiO₂. Anatase phase TiO₂ QDs-sensitized Cu₂S (Cu₂S/TiO₂) nanocomposites with varying concentrations of TiO₂ QDs were synthesized via a facile two-stage hydrothermal-wet impregnation method [97]. The as-synthesized co-catalyst showed H₂ evolution rate of ∼45.6 mmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ under simulated solar irradiation, which is ∼5 and 2.4-fold higher than that of the pristine TiO₂ and Cu₂S, respectively.

Chen et al. synthesized the CoS_x-rGO/TiO₂ composite photocatalyst by employing two steps: first, the rGO nanosheets hydrothermally loaded onto TiO₂ NPs, and then CoS_x nanodots (0.5–2 nm) photodeposited on the rGO nanosheets [98]. The produced CoS_x-rGO/TiO₂ (10%) composite photocatalyst exhibited H₂ production efficiency of 256.97 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, being ∼12 times higher than that of rGO/TiO₂ photocatalyst and ∼2 times higher than that of CoS_x/TiO₂. It was proposed that rGO nanosheets facilitate the rapid transfer of photo-induced carriers from TiO₂ to CoS_x nanodots and that CoS_x nanodots thereafter function as active sites for H₂ production. In another study, the TiO₂@CuS double-shell nanoboxes were constructed via a multistep control strategy involving template participation [99]. The optimized TiO₂@CuS double-shell nanoboxes exhibited H₂ production rate of 2467 and 173 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ under full-spectrum and near-IR irradiation, respectively.

Bismuth sulfide (Bi₂S₃) has also recently been considered as an effective co-catalyst (i.e., with TiO₂) because of its small band gap (1.28 eV), strong visible light absorption capacity (450–700 nm), and large negative conduction band (CB) potential. The amorphous BiS_x nanodots with the size of 0.5–2 nm were loaded on TiO₂ surface to facilitate the photocatalytic H₂ generation [100]. The BiS_x/TiO₂ photocatalyst was synthesized by an Ethylenediaminetetraacetic acid (EDTA)-assisted two-step process, which included the adsorption of the Bi(III)-EDTA ions on TiO₂ and the in-situ formation of amorphous BiS_x nanodots. The optimized BiS_x/TiO₂(1.0 wt%) photocatalyst showed the maximum photocatalytic H₂ production rate of 803.2 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ with an AQE of 3.86%, which is 83.6 and 1.6 folds of the clean TiO₂ and crystalline Bi₂S₃-modified TiO₂ (c-Bi₂S₃/TiO₂), respectively. The enhanced photocatalytic H₂ production was assigned to the amorphous BiS_x nanodots providing more unsaturated active S atoms, which act as the efficient H₂ evolution active sites. Effective hydrogen evolution from water splitting under solar light irradiation was investigated using ZnS-PbS QDs loaded onto a Au/TiO₂ photocatalyst [101]. In the presence of 20 wt% methanol as a sacrificial reagent to photocatalyze for 5 h, 5011 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ HER was observed. Nguyen et al. utilized polyvinyl alcohol (PVA) as a binder compound for combining MoS₂ and TiO₂ nanotube arrays (TNAs) to improve photoelectrochemical (PEC) water splitting ability [102]. They showed that by a thermal treatment process, the formation of the $\pi$-conjunction in the PVA structure enhanced the PEC performance of MoS₂/TNAs, exhibiting linear sweeps in the anodic direction with the current density over 65 µA/cm² at zero bias voltage vs. Ag/AgCl.

In general, ternary-metal-sulfide/TiO₂ photocatalysts are expected to have lower toxicity, a tunable band structure, and favorable chemical stability, enabling a higher HER over the binary-metal-sulfide/TiO₂. In this regard, NiWS_x nanodot-modified TiO₂ (NiWS_x-ND/TiO₂) photocatalysts were prepared by evenly anchoring ultrasmall amorphous NiWS_x nanodots (∼2 nm) onto TiO₂ surface by means of a facile photodeposition strategy [103]. The optimized NiWS_x-ND/TiO₂ (3 wt%) photocatalyst manifested the maximum H₂ generation efficiency of 4580 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ with the corresponding AQE of 13%. Dai et al. reported a S-scheme heterojunction photocatalyst by coupling the ZnCo₂S₄ NPs on the surface of TiO₂ through simple solvothermal method [104]. The prepared nanocomposite photocatalyst exhibited H₂ evolution efficiency of 5580 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ with an AQE of 11.5% at 420 nm, which is 88.3 times and 54.3 times that of TiO₂ and ZnCo₂S₄ NPs, respectively. It was discussed that the S-scheme electron transfer path facilitates better electron-hole separation alongside improved charge transfer efficiency. While the ZnCo₂S₄ NPs significantly enhance the visible light absorption, leading to a high H₂ yield. The resulting efficiency and proposed photocatalytic mechanism are summarized in Figure 11.

Also, a TiO₂ nanoplate/amorphous MoS₂ QD (a-MoS₂) composite was prepared via in situ sulfuration of MoO₂/TiO₂ [105]. It was shown that in this composite, the intimate chemical bonding instead of electrostatic adsorption between TiO₂ and MoO₂ enables effective electron transfer from TiO₂ to a-MoS₂. The optimized a-MoS₂/TiO₂ nanocomposite exhibited a maximum H₂ yield of 880.3 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$. In another study, ultrathin TiO₂ nanosheets were integrated into the growth of ZnIn₂S₄ to produce TiO₂-ZnIn₂S₄ heterostructure nanoflowers in order to design a direct Z-scheme heterojunction photocatalyst with an improved PHE [106]. It was revealed that the formed direct Z-scheme heterojunction can consume residual photogenerated holes of ZnIn₂S₄ and enable water oxidation to occur solely at the TiO₂ site, improving the overall stability of the system. The optimized DZH nanoflowers showed enhanced photocatalytic water splitting activity of 214.9 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ for H₂ evolution and 81.7 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ for O₂ evolution.

2.4. Metal Sulfide/g-C₃N₄ Heterostructures

Combining suitable co-catalysts and simultaneous interface engineering is another promising strategy for photocatalytic HER enhancement. In this regard, Wang et al. reported a CN/A-MoS₂/Pt photocatalyst using amorphous MoS₂ (a-MoS₂) as the conformal multifunctional intermediate layer between g-C₃N₄ and Pt [107]. They showed that a conformal chemically bonded interface and low overpotential for the H₂ evolution reaction lead to enhanced interface charge transfer and photochemical reaction kinetics. Thus, under Vis-light illumination, 5830 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ HER was achieved, which is more than an 83-fold enhancement compared with the g-C₃N₄. The detailed parameterization and corresponding HERs, together with the proposed photocatalytic reaction mechanism, are given in Figure 12.

It should be noted that the effective surface area of the co-catalysts has a significant impact on photocatalytic performance. In this regard, it has been reported that g-C₃N₄ structures with a large surface area can be produced using a rapid high-temperature hydrothermal post treatment strategy with fast heat transfer and short processing time. Thus, the extended surface area resulted in about 5.1 higher HER compared to pristine g-C₃N₄ under Vis-light ($\lambda$ > 400 nm) irradiation [108]. A ternary composite catalyst of MoS₂/Fe₂O₃/g-C₃N₄ via hydrothermal method was reported [109]. The fabricated composite photocatalyst demonstrated H₂ generation rate of 7.82 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, which is almost 5 times higher than the single-component g-C₃N₄ photocatalyst (1.56 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$). The improvement of its photocatalytic activity can be mainly attributed to its enhanced visible light absorption and the increased specific surface area. According to the earlier DFT investigations, in the MoS₂/g-C₃N₄ composite, the CB minimum and VB maximum of the g-C₃N₄ monolayer are roughly 0.83 eV, and 0.15 eV higher than those of the MoS₂ sheet [110]. As a result of the formed type-II band alignment, photoexcited electrons can easily migrate from the g-C₃N₄ monolayer to the MoS₂ sheet, resulting in a high hydrogen evolution reaction activity. Additionally, the charge transfer between MoS₂ and g-C₃N₄ generates a polarized field at the boundary, enabling effective separation of photogenerated carriers. A transition metal sulfide (Co₃S₄) produced from MOF was used to build an ohmic contacting Co₃S₄-CN heterojunction interface for enhanced photocatalytic H₂ generation [111]. It was revealed that electron injection from the lower work function Co₃S₄ to g-CN causes a significant interfacial electric field resulting in downward band bending in a g-CN semiconductor. Furthermore, the Co₃S₄-CN heterojunction facilitates exciton dissociation while removing a possible barrier for electron extraction from g-CN, facilitating charge carrier separation synergistically. The obtained co-catalyst showed H₂ evolution of ∼217.0 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ upon visible light illumination. Chen et al. employed a light-induced deposition approach to precisely load NiS nanodots onto the surface of g-C₃N₄ to produce a bifunctional photocatalyst [42]. The obtained photocatalyst showed an enhanced photocatalytic H₂ evolution reaction and photocatalytic oxygen reduction. Particularly, the NiS@g-C₃N₄-30 displayed a H₂ evolution rate of 3297 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ under visible-light illumination. It was discussed that the close interaction of the loaded NiS nanodots with the electron transfer sites of g-C₃N₄ nanosheets provides effective separation of photogenerated charge carriers and further accelerates the chemical reaction on the surface of the photocatalyst.

Considering that loading more NiS co-catalyst to function as active sites of g-C₃N₄ for improved photocatalytic H₂ generation, a facile sulfur-mediated photodeposition approach was developed [112]. It was shown that photogenerated electrons excited by Vis light reduce the S molecules absorbed on g-C₃N₄ surface to ${\mathrm{S}}^{2-}$, and subsequently, a NiS co-catalyst is formed in situ on the g-C₃N₄ surface by a combination of ${\mathrm{Ni}}^{2+}$ and ${\mathrm{S}}^{2-}$. The resulting NiS/g-C₃N₄ photocatalyst exhibited H₂ generation rate of 244 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$.

The PtS/g-C₃N₄ co-catalyst was produced by loading 2D PtS nanorectangles with a length of ∼7 nm and width of ∼5 nm on the surface of g-C₃N₄ nanosheets utilizing in situ vapor-phase growth [113]. The constructed compound showed an unusual metal sulfide-support interaction, which promoted stabilization of the geometrical structure and energy-band structure, leading to acceleration of charge transfer, and reduction of hydrogen adsorption free energy of PtS/CN. The characteristics mentioned above gave rise to a high H₂ generation rate of 1072.6 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 45.7% at 420 nm. This is ∼13.3 and ∼1532.3 times enhancement in comparison with the Pt NPs/g-C₃N₄ nanosheets and pristine g-C₃N₄ nanosheets, respectively. In another study, ultrasmall fine Pt NPs and MoS₂ nanosheets were deposited on the surface of g-C₃N₄ as co-catalyst [114]. While MoS₂ could efficiently induce structural reconstruction of g-C₃N₄, resulting in surface charge redistribution, the boosted electrons localized in the S and N atoms were conducive to the reduction of ${\mathrm{Pt}}^{4+}$ via strong metal-support contact, hence enhancing charge separation and transfer. Consequently, the synthesized Pt/MoS₂/g-C₃N₄ composite exhibited a high photocatalytic hydrogen production rate up to 1595.3 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 30.9% at 435 nm, which is 87.6 times higher than that of the clean g-C₃N₄.

Wei et al. synthesized ternary ZnS-NiS₂/g-C₃N₄ and NiS₂-ZnS/g-C₃N₄ composites by a thermal polymerization and post-co-deposition-sulfidation methods to promote visible light absorption as well as charge separation and transfer [115]. The as-synthesized NiS₂-ZnS/CN(0.4) and ZnS-NiS₂/CN(0.4) photocatalysts showed PEH rate of 283.3 and 302.7 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, respectively. The 2D SnS/g-C₃N₄ nanosheets prepared by a facile ultrasonic and microwave heating method were also reported for enhanced H₂ evolution [116]. Furthermore, after being loaded with MoO₃ particles, the photocatalyst stability improved significantly due to the produced Z-scheme heterojunction. As a result, the enhanced light-harvesting capability of the composite, reduced charge-transfer resistance, improved electrical conductivity, and the co-catalyst effect of SnS facilitated a higher H₂ generation rate of 818.93 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ under AM1.5G irradiation. Also, g-CN/MoS₂ composite was synthesized via an in-situ gas-solid reaction where rodlike MoO₃ was sulfurized to form MoS₂ by the byproduct generated during the thermal condensation of thiourea [117]. The H₂ evolution rate improved from 0.99 to 13.31 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, which is 13.44 times higher than that of clean g-CN. Similarly, 1T/2H hybrid WS₂ was prepared as a non-noble metal co-catalyst for g-C₃N₄ using a straightforward synthesis technique based on an in situ gas-solid phase reaction [118]. Because of the hybrid crystal phase and the presence of S defects in WS₂, the photocatalytic hydrogen evolution efficiency of the co-catalyst improved noticeably, increasing from 0.32 to 34.76 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. In another study, red-phosphor-doped Co₉S₈ (P-Co₉S₈) was synthesized by a hydrothermal process [119]. It was revealed that the introduction of P-Co₉S₈ improved the electrical conductivity and surface area of the composite along with reducing the over-potential of H₂ evolution. Consequently, the optimized 25% P-Co₉S₈/g-C₃N₄ exhibited H₂ evolution rate of 4362 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ under UV and visible light, which is roughly 121.2 times improved in comparison with the pristine g-C₃N₄. According to the photoelectrochemical analysis, the designed photocatalyst was assigned as a Type-I system, where more electrons on the CB of P-Co₉S₈ to participate in the H₂ evolution processes.

In parallel to the binary metal sulfides, ternary metal sulfides were also employed to produce co-catalyst together with g-C₃N₄ for photocatalytic H₂ generation. In this regard, in order to improve the photocatalytic water splitting reactivity of g-C₃N₄, ternary ZnCo₂S₄ was coated on its surface [120]. The optimized 25 wt%-ZnCo₂S₄/g-C₃N₄ showed H₂ evolution rate of 6619 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, which is 55.2 times higher than that of g-C₃N₄ alone. Thus, the enhanced H₂ production activity was assigned to the presence of both zinc and cobalt ions in ZnCo₂S₄, which reduced the H₂ evolution overpotential and charge recombination rate. The obtained photocatalytic process was ascribed as an S-scheme charge transfer route based on trapping experiments for active species. Also, copper-nickel sulfide was anchored on the g-C₃N₄ for increasing the photocatalytic H₂ evolution [121]. It was revealed that copper-nickel sulfides can effectively suppress deep trapping states of active charge in g-C₃N₄ improving the efficiency of the shallow trapped electron transfer through a single C-S bond for enhancing Vis-light-driven photocatalytic H₂ production. The synthesized co-catalyst showed H₂ evolution rate up to 752.8 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ being 470 times higher than that of clean g-C₃N₄ [121]. Employing femtosecond time-resolved spectroscopy, it was also shown that the strong interaction between copper-nickel sulfide and g-C₃N₄ favored the hole transfer from copper-nickel sulfide to the trap state of g-C₃N₄ leading to a remarkable near-IR-driven photocatalytic H₂ production (0.32 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$, $\lambda$ > 800 nm).

2.5. Metal Sulfide/MOFs (COFs) Heterojunction Photocatalysts

One of the viable techniques for solar energy-driven water splitting is the development of metal-organic frameworks and covalent-organic framework (COF)-based heterostructure. Wang et al. fabricated a direct Z-scheme heterostructure COF-metal sulfide hybrid (T-COF@CdS) with shell-core architecture by self-polymerization of 1,3,5-benzenetricarboxaldehyde and 2,4,6-tris(4-aminophenyl)-1,3,5-triazine in situ on CdS NPs [122]. In these shell-core structures, the C-S chemical bonding generated between T-COF and CdS results in a tight and stable interface. Moreover, the T-COF shell acts as an oxidation site to avoid the photocorrosion of CdS and protects the catalytic centers of CdS against deactivation, leading to long-term stability. Thus, the fabricated Z-scheme T-COF@CdS-3 heterostructure exhibited a high AQE of 37.8% at $\lambda$ = 365 nm.

MOFs have been widely applied in photocatalytic hydrogen production because of their high surface area, semiconductor-like behavior, and great active sites. For example, ${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S NRs assembled on the surface of Ni-MOF-74 with flake morphology, where the Ni-S bonds between Ni-MOF-74 and ${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S form a unique transfer channel for photo-induced charge carriers [123]. Thus, the potential energy difference between the two compounds enables fast injection of the photogenerated electrons in the CB of ${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S into the CB of Ni-MOF-74, which remarkably inhibits the electron-hole recombination. Thus the improved carrier lifetime, fast charge transfer rate, and lower overpotential lead to a high H₂ production rate reaching 7.104 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, which is 6.96 times that of ${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S. A CdS@NiS photocatalyst using bimetallic MOFs as the precursor was also reported [124]. It was demonstrated that bimetallic Cd/Ni-MOFs can result in well-contacted heterogeneous interfaces between CdS and NiS, effectively facilitating photogenerated carrier separation and enhancing electron interface transport. The designed CdS@NiS photocatalysts showed a high photocatalytic HER of 42.7 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, and the corresponding AQE reaches 13.23% at 450 nm.

In another study, the microwave irradiation method was utilized to create the UIO-66(Ce)/ZnCdS composite using Ce-based MOF components [UIO-66(Ce)] [125]. It was revealed that ZnCdS NPs were decorated on the surface of UIO-66(Ce) and that UIO-66(Ce) considerably increased the H₂ generation ability of ZnCdS NPs by effectively enhancing the charge carrier separation. Thus, the UIO-66(Ce)/ZnCdS nanocomposite photocatalysts showed H₂ production rate of 3.958 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, which is roughly 1.95 times higher than that of ZnCdS (2.031 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$).

Also, as an efficient ZnIn₂S₄-MOFL photocatalysts for robust photocatalytic hydrogen evolution, hollow tubular ZnIn₂S₄ modified by MOF-layers with In-MOF as precursor was demonstrated [126]. The DFT studies proved the decreased density of states (DOS) of ZnIn₂S₄ after coupling with the organic ligands. The established direct Z-scheme photocatalytic system synergistically promoted the separation of electron and hole pairs leading to high photocatalytic H₂ evolution. The H₂ evolution using this photocatalyst reached 28.2 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, which is roughly 14.8 times higher than that of pristine ZnIn₂S₄. Moreover, the photocatalyst showed AQEs of 22.67% at monochromatic light (350 nm) and the stable H₂ generation capability (5.7 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$) under natural sunlight irradiation. Aiming for efficient electron-hole separation and carrier utilization in photocatalytic H₂ generation, Lian et al. employed MOF (NH₂-UiO-66) modified with inner Pt NPs and outer CdS NPs [127]. The designed ternary composite Pt@NH₂-UiO-66/CdS possessing a spatially separated, the hierarchical structure was employed for visible-light-driven H₂ evolution, which exhibited higher H₂ yields with an AQE of 40.3% at $\lambda$ = 400 nm irradiation. The proposed photocatalytic mechanism and H₂ evolution rate are given in Figure 13a.

As we already discussed through this overview, stabilizing photoexcited states by suppressing charge recombination is one of the most important parameters for being good photocatalysts. In this regard, using two-dimensional (2D) UNiMOF as a precursor and by applying the partial hydrothermal sulfidation method, a g-C₃N₄/UNiMOF/NiS₂ (GUN) ternary heterojunction (TH) photocatalyst was synthesized [128]. As expected, the optimized GUN-40 heterojunction containing 40 mg of UNiMOF (GUN-40) exhibited enhanced HER of 572 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with a longer exciton recombination lifetime of $\tau$ = 12.12 ns. This is remarkably higher than that of the pristine g-C₃N₄ (0.67 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$, $\tau$ = 7.87 ns) under Vis-light illumination. Also, the optimized GUN-40 photocatalyst showed an improved AQE up to 1.54 % at $\lambda$ = 450 nm. Thus, the GUN TH structure significantly accelerates electron transport to the NiS₂ active site, improving the excited state lifetime, and minimizes the charge recombination rate, thus enhancing photocatalytic performance. The proposed photocatalytic mechanism and H₂ evolution rate are given in Figure 13b. Similarly, a CdS@MOF-808 composite photocatalyst where CdS NPs are confined inside the nanosized pores of ${\mathrm{Zr}}^{4+}$-based MOF-808 was recently proposed [129]. It was shown that the confinement of the CdS NPs inside the MOF pores, close to the ${\mathrm{Zr}}^{4+}$ cluster, forms a shorter electron transfer route from CdS to the catalytic ${\mathrm{Zr}}^{4+}$ cluster. Thus, with an optimized load of 3.56 wt% CdS HER of 10.41 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ from the water splitting was achieved. The proposed photocatalytic mechanism and HER are shown in Figure 13c. Moreover, a similar composite containing CdS NPs on the external surface, instead of cavities of MOF-808, showed a lower photocatalytic HER of 0.15 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$.

Among other MOFs, zeolitic imidazolate frameworks (ZIFs), which are topologically isomorphic with zeolite, possess exposed metal sites, accessible carbon/nitrogen ligands, and good chemical stability, are attractive for solar-driven hydrogen generation. Recently, ZIF-67/CdS hollow sphere (HS) and ZIF-8/CdS HS heterostructures were synthesized through an in-situ self-assembly process, where the ZIF-67 and ZIF-8 are closely coated on CdS HSs to form “double-shell”-like structures with porous outer layers [130]. The optimal proportion of ZIF-67/CdS HS showed a HER of 1721 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ under Vis-light illumination, which was ∼3 times higher than that of ZIF-8/CdS HS (555 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$). The photocatalytic mechanism for the ZIF-8/CdS HS and ZIF-67/CdS HS was assigned as type-II and Z-scheme mechanisms, respectively. The synthetic route to composite formation alongside the morphological characterization, corresponding model structure and photocatalytic mechanism for the optimized ZIF-67/CdS HS are summarized in Figure 14. In this line, lollipop-shaped Co₉S₈/CdS nanocomposite via direct annealing of zeolitic imidazolate framework-67 (ZIF-67) as a template and CdS as a sulfur source was reported [131]. It was mentioned that the close proximity of these two materials facilitates electron transport and increases the number of active sites toward the desirable redox reaction. The as-prepared Co₉S₈/CdS resulted in a remarkably improved HER of 1852 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ under Vis-light illumination. In another study, the ZIF-67 derived Co₃O₄ was used as precursor to prepare ZnIn₂S₄@CoS₂ photocatalyst [132]. The optimized structure possessing 5% CoS₂ showed the highest HER performance of 2768 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ under the irradiation of simulated sunlight, which is three times higher than the ZnIn₂S₄ (879 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$).

2.6. Hierarchical Metal Sulfide Photocatalysts

Hierarchical nano- and mesostructures are potential systems for catalysis, electrocatalysis, solar cells, membranes, and photocatalysis, among other applications. Regarding photocatalytic applications the layered (e.g., 2D) and multidimensional (e.g., 1D/2D, 1D/3D, etc.) hybrid heterostructures are particularly appealing. In addition to dimensional advantages (tunable band gap, large surface and interaction area, etc.), these structures enable formation of high energy facets and easy morphological modification, such as the design of membranes [133] and photonic crystals [57], which may enable further photocatalytic enhancement. Recently promising results have been achieved in this field. For example, an ultrathin 2D/2D MoS₂/ZnIn₂S₄ co-catalyst was designed and prepared employing two steps of the in situ hydro/solvothermal method [134]. It was suggested that the reduced dimensionality of the components enables the fast charge transfer and suppression of the photogenerated charge carriers recombination. 4 wt% ut-MoS₂/ZnIn₂S₄ exhibited the highest H₂ evolution rate of 221.71 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ under Vis-light irradiation ($\lambda >$ 420 nm) with a corresponding AQE of 1.8%.

Another trend of the strategies toward photocatalytic enhancement is the improvement of sulfur vacancies (SV) in metal sulfide compounds. In this regard, CuS NPs, which possess suitable band gap energy [135,136], were employed to prepare CdS@CuS core-shell heterostructure photocatalyst by hydrothermal treatment and subsequent cation-exchange reactions [137]. The CdS-sulfur vacancies enriched by CuS via photoinduced interfacial charge transfer. The developed CdS-SV@CuS(5%) catalyst showed a significantly higher H₂ production rate of 1654.53 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. Similarly, 0D/2D NiS/CdS nanocomposite heterojunction photocatalyst with enhanced photocatalytic HER was reported [138]. There, the composite containing 25% molar ratio of NiS showed the highest HER at about 18.1 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. A Co₉S₈@CdIn₂S₄ heterostructure with hierarchical hollow construction and closely connected heterogeneous shells was also reported for photocatalytic H₂ evolution under visible light [47]. It was demonstrated that such a unique architecture may efficiently separate and transmit photoinduced charges while also providing a large surface area and exposing rich sites for photocatalytic redox processes. The designed photocatalyst showed HER of 4604 µmol ${\mathrm{h}}^{-1}$ ${\mathrm{g}}^{-1}$ under 1.5 AM simulated light irradiation. In other study, CoS_x/twinned-${\mathrm{Cd}}_{0.5}{\mathrm{Zn}}_{0.5}$S (CoS_x/T-CZS) photocatalyst was produced through an in-situ deposition method [139]. Particularly, the formation of the zinc-blende/wurtzite (ZB/WZ) homojunction in T-CZS, as well as the heterojunction between CoS_x and T-CZS promoted the charge separation in the bulk and surface of T-CZS. Consequently, T-CZS exhibited much higher activity compared to the zinc blende ${\mathrm{Cd}}_{0.5}{\mathrm{Zn}}_{0.5}$S (ZB-CZS) and wurtzite ${\mathrm{Cd}}_{0.5}{\mathrm{Zn}}_{0.5}$S (WZ-CZS) with single-phase. Introduction of amorphous CoS_x improved H₂ evolution rate of 4 wt% CoS_x/T-CZS to 76.2 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. Recently, various hollow CoS-supported CdS catalyst (H-CoS/CdS) prepared by loading CdS onto the surface of H-CoS was also reported for H₂ production in pure water and simulated seawater [140]. The H₂ production rate of the as-synthesized co-catalyst from simulated seawater reached to 572.4 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. Thus, incorporating the h-CoS co-catalyst extended the light absorption range of CdS, improved the chemical stability, and significantly enhanced the charge separation leading to high photocatalytic efficiency. In another study, cobalt sulfide was employed to produce 1D-2D CdIn₂S₄/carbon nanofibers (CNFs)/Co₄S₃ tandem Schottky heterojunction via in-situ electrospinning combined with a hydrothermal method [141]. The optimised composite showed the H₂ production activity rate of 25.87 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ with an AQE of 19.56% at $\lambda$ = 365 nm. Similarly, in order to mitigate the recombination of electrons and holes, MoS₂ nanosheets with rich active sites were modified on the surface of CdS NRs by a room-temperature sonication treatment [142]. The obtained 6% MoS₂/CdS nanocomposites co-catalyst showed improved H₂ evolution rates of 63.71 and 71.24 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ in visible light and simulated solar light irradiation, respectively. The AQE of the 6% MoS₂/CdS nanocomposite at 420 nm reached 33.62%. The enhanced photocatalytic activity was assigned to the extra photogenerated charge carriers, which enhanced the charge separation and transfer ability of the MoS₂/CdS co-catalyst. By in-situ epitaxially growing 2D ZnIn₂S₄ nanosheets onto the surface of 1D faulty CdS NRs, atomic CdS/ZnIn₂S₄ heterojunctions were produced [143]. The optimized ZIS/CdS-0.3 heterojunctions showed the highest activity (5.80 mmol/g), which is improved in comparison with the pure CdS (1.28 mmol/g) and ZIS (5.14 mmol/g). As already discussed, the efficiency of charge carrier separation and the acceleration of the surface reductive process in MS semiconductor photocatalysts are essential parameters in promoting photocatalytic H₂ evolution from water. The Ti₃C₂ (Mxene) modified all-sulfide 2D/2D S-scheme heterojunction Ti₃C₂/ZnIn₂S₄ (ZIS)/CdS composite material was produced in two steps to regulate both parameters [144]. It was demonstrated that the produced S-scheme heterojunction enhances electron-hole separation efficiency (see Figure 6 for S-scheme photocatalysts). In contrast, the intimate 2D/2D van der Waals structure provides a high interaction force and a wide contact area to promote charge transfer. Furthermore, the presence of 2D Ti₃C₂ provides the accumulation layer, minimizing electron and hole recombination. Consequently, the obtained Ti₃C₂/ZIS/CdS composite photocatalyst could reach to H₂ evolution rate of 8.93 mmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$. The proposed charge transfer mechanism and obtained H₂ evolution rate are given in Figure 15a.

Recently two-component co-catalyst, dendritic PbS@CuS core-shell photocatalysts were synthesized via a two-step method. Thus, a dendritic PbS core was formed through an ionic liquid/surfactant-assisted hydrothermal method synthesis, and a CuS shell was grown with an ion-exchange method [145]. In situ Cu L₃-edge near-edge X-ray-absorption fine-structure (NEXAFS) spectroscopy before and after exposure to solar light showed the photoelectron transfer from CuS to PbS. The CuS shell proportions in these dendritic PbS@CuS photocatalysts were adjusted to attain a maximal activity of 1736 mol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ hydrogen generation. Also, see Figure 15b for the reported HER and proposed charge transfer mechanism. Co₉S₈ is another rising member of MS photocatalysts because of its narrow band gap. Using Co₉S₈, two-component (0D)Co₉S₈/ (3D) Bi₂S₃ S-scheme heterojunction for efficient photocatalytic hydrogen evolution was reported [147]. The unique 0D/3D architecture and formed S-scheme heterojunction were shown as an efficient and convenient path for transferring photogenerated charges, which further facilitate the effective separation and diversion of the electrons. A core-shell ${\mathrm{Cu}}_{1.94}$S-MnS nanosized heterostructure photocatalyst prepared via cation exchange reaction on djurleite ${\mathrm{Cu}}_{1.94}$S nanospheres template also reported exhibiting high H₂ evolution rate of 878.1 µmol ${\mathrm{g}}^{-1}$ ${\mathrm{h}}^{-1}$ [148]. In another study, CuS nanosheets-assembled hollow cubic cages were used to construct 3D hierarchical nanocages photocatalysts. Thus, 3D hierarchical nanocages CuS@ZnIn₂S₄ photocatalyst with abundant and compact nanosheets 2D/2D hetero-interface was designed [146]. The as prepared structure exhibited an improved photocatalytic H₂ evolution rate up to 7910 µmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$ in the absence of any co-catalyst. According to further experimental and theoretical studies, the strong interaction between CuS and ZnIn₂S₄ 2D/2D hetero-interfaces can remarkably facilitate the separation of photogenerated charge carriers and accelerate the electron transfer. The photocatalytic mechanism and optimization of HER are given in Figure 15c. The CoN with high HER performance was introduced into ${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S through simple electrostatic self-assembly [48]. The inclusion of CoN boosts the absorption of Vis-light by the photocatalyst. As a result of its ideal CB location and strong conductivity, CoN receives photogenerated electrons from ${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S as an excellent electron acceptor in the form of active sites. Remarkably, CoN significantly improved the HER of photocatalysts up to 14.612 mmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$.

As stated above, one of the promising routes for photocatalytic HER enhancement is forming porous structures, which possess large surface area and enhanced excited state lifetime. However, due to the fast precipitation between the metal ions and ${\mathrm{S}}^2-$ ions and the large volume contraction during the sulfide formation, synthesizing such porous structures is not a straightforward process. To cope with this problem, Zhout et al. employed a general polymer-oriented acid-mediated self-assembly method to synthesize highly crystalline Mesoporous metal sulfides (MMSs) (e.g., ZnS, CdS, Ni₃S₄, CuS, and Zn_x${\mathrm{Cd}}_{1-x}$S) by using polyethylenimine (PEI) as pore-forming agent [149]. The synthesis procedure, together with the structural analysis, is shown in Figure 16a–c. In order to adjust the pH and control the interactions between the inorganic precursors and the PEI, acetic acid was employed. It was found that the absorption edge of ${\mathrm{Zn}}_{0.56}{\mathrm{Cd}}_{0.44}$S exhibits a slight red shift as the Cd content increases, indicating band gap narrowing. The optimized MMSs ${\mathrm{Zn}}_{0.56}{\mathrm{Cd}}_{0.44}$S photocatalyst having abundant mesopores and homojunction enabled high HER of 14.3 mmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$. The optimized HER plot and proposed charge transfer mechanism are shown in Figure 16d,e, respectively.

2.7. Plasmonic Co-Catalyst Modified Metal Sulfide Photocatalysts

Plasmons, the collective oscillation of the excited conductive charge carriers, have been in the focus of fundamental and applied sciences [150,151,152,153,154,155]. The optical response of plasmonic nanostructures can be tuned from the ultraviolet through the visible to the near-infrared wavelength, by adjusting the size, shape, composition and geometrical arrangement of the NPs. In this regard, developing new photocatalysts by employing metallic nanostructures with localized surface plasmon resonance (LSPR) is also considered state-of-the-art research. For this purpose, Au and Ag NPs are the most frequently used materials since their light absorption can effectively be shifted to Vis-light region by varying the particle size and shape [156]. In a simple picture, the enhanced electric fields around the plasmonic particle near the surface of the main MS photocatalyst increase the rate of electron-hole pair formation. As a result, the readily separated charge carriers can effectively migrate to the MS surface and drive photocatalytic process. Due to this specific pathway, combining the MS semiconductor photocatalyst with plasmonic-metal NPs can drastically reduce recombination rate. Thus, by combining enhanced light absorption capabilities through light-scattering, near-field amplification, hot-charge injection, or plasmon-induced resonance energy transfer mechanisms, such hybrid structures lead to improved photocatalytic activity and H₂ evolution rate. For example, Au-PbS-CdS nanodumbbells with double Au caps on the ends were produced using a straightforward three-step wet-chemical technique [157]. It was demonstrated that by modifying the size and spacing of the Au caps, the LSPR-induced lateral electric field distribution surrounding the hybrid structure can be beneficially produced. Thus, the dual-plasmon-induced broad spectral coupling between the metal and semiconductor could be obtained. Moreover, multi-interface structures (e.g., PbS-CdS, Au-PbS, and Au-CdS) provide fast and multiple transfer channels for the photogenerated charges. As a result, the designed Au-PbS-CdS nanodumbbells provided photocatalytic HER of 513.63 µmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$, which is superior to mixed Au, PbS, CdS NPs (97.37 µmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$). Also, the CdS/TiN core-shell structure with various thicknesses of the TiN outer layer was studied [158]. The optimized heterostructure involving 15 nm TiN showed a H₂ production rate of 86.1 mmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$, which is a 362% improvement compared to the CdS NWs. This remarkable improvement due to the TiN outer layer was assigned to a synergetic coupling of three main factors; the improvement of light absorption and electron concentration caused by LSPR, the reduction of the photo-induced carrier recombination rate, and the improved chemical stability during photocatalytic water splitting. Tu et al. have also shown that the H₂ evolution rate of Au NPs encapsulated by MoS₂ nanoflowers (Au@MoS₂ NFs) was remarkably improved due to the coupling of LSPR with the piezophototronic effect [159]. Consequently, applying simultaneous light illumination and mechanical vibration to the obtained structure enhanced the H₂ evolution rate from 2981 to 4808 µmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$ (see Figure 17). Thus, the piezoelectric generation of hot electrons through LSPR was a primary reason for the observed photocatalytic enhancement. The reported mechanism of the photocatalytic process involving LSPR and piezophototronic effects over the Au@MoS₂ NFs are described in Figure 17.

By assembling Au@CdS core-shell NPs on Ag@SiO₂ shell-isolated MPs dual-plasmonic-antenna nanocomposites to achieve the simultaneous efficient generation of energetic hot electrons and strong electromagnetic field enhancement for efficient photocatalytic H₂ evolution was reported [160]. Where, the Au@CdS antenna generates hot electrons to facilitate H₂ evolution reaction, while the Ag@SiO₂ antenna enables strong electromagnetic field enhancement to promote the generation and separation of hot carriers. The existence of both hot-electron transfer and plasmon-induced resonance energy transfer in the designed system was confirmed by transient absorption spectroscopic experiments and electromagnetic field simulations. The designed photocatalyst showed 191.2 mmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$ HER under Vis-light irradiation ($\lambda$ > 400 nm). Cheng et al. studied modification of the Schottky barriers over Ni/S vacancy-rich ${\mathrm{Mn}}_{0.3}{\mathrm{Cd}}_{0.7}$S (Ni/MCS-s) composites prepared by the self-assembly and photochemical method [161]. The optimized Ni/MCS-s heterostructures exhibited enhanced HER up to 164.1 mmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$ in simulated seawater (3.5 wt% NaCl) with the maximum QYE of 60.4% at 420 nm. They showed that coupling plasmonic Ni and S vacancies can effectively lower the Schottky barrier and enhance hot electrons across the interface for the photocatalytic process. In addition, introducing the Ni layer remarkably reduced the corrosion of the catalyst in seawater. Similarly, plasmonic quaternary CuS-AgVO₃/Ag-TNR hetero nanostructures (HNSs) were employed to improve photocatalytic H₂ generation [162]. The optimized 1 wt% CuS-AgVO₃/Ag-TNR HNS showed the highest HER of 756 µmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$. The achieved HER was related to higher charge separation, improved conductivity, LSPR, and electron-storing capacity. Au/g-C₃N₄/ZnIn₃S₄ plasma photocatalyst heterojunction composite with 3D hierarchical microarchitecture was also reported [163]. They showed that integrating Au NPs, the resulting Au/g-C₃N₄/ZnIn₂S₄ composite photocatalyst showed highest HER of 0.973 mmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$ under Vis-light irradiation and reached up to the maximum for the Au/g-C₃N₄ is 40 wt% of ZnIn₂S₄. In addition to the implementation of catalysts, the ${\mathrm{Zn}}_{1-x}$Cr_xIn₂S₄ solid-solution photocatalysts synthesized via a facial hydrothermal method was also reported with an improved HER [164]. The elevated HER was ascribed to the LSPR modes originating from amorphous domains, which facilitated interfacial charge transfer between amorphous and crystalline domains. Moreover, with the increase of Cr percentage in the stoichiometry, the highest photocatalytic activity ∼ 7 mmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$ for the ${\mathrm{Z}}_{0.75}{\mathrm{C}}_{0.25}$IS after 6 hrs reaction time was achieved. This is almost 2.2 times higher than the well-known ZnIn₂S₂ photocatalyst. As already discussed, incorporating plasmonic materials with metal sulfide photocatalysts results in enhanced hydrogen production activity. Due to the fact that established control of the plasmonic signal over the spectral range and designing effective architectures remains a promising rote for further enhancement in hydrogen evolution rate. In this regard, another parameter to optimize is the amount of co-catalyst particles. Feldmann et al., for example, found that shortening the path by increasing the number of co-catalyst particles boosts the transfer rate but decreases photocatalytic performance [165]. This is due to the fact that subsequent electron transport to the acceptor is significantly slower; as a result, the recombination rate with neighboring holes increases, compensating any gain from the enhanced electron transfer rate.

3. Summary and Outlook

In this state-of-the-art review paper, we discussed very recent, namely the last two years, developments regarding metal-sulfide-based photocatalysts for H₂ generation. This survey showed that within a short time, a tremendous amount of effort has been devoted to this topic, which has resulted in an improvement of our fundamental understanding of the MS photocatalyzed H₂ production as well as a remarkable enhancement over the H₂ evolution rate, as summarized in

Table 1. Summary of recently reported metal sulfide based photocatalysts with improved efficiency for the H₂ evolution.

Photocatalyst	Type/Junction	Rate (µmol ${\mathrm{g}}^{-1}{\mathrm{h}}^{-1}$)	QYE (%)	Incident Light ($\mathit{\lambda }$, nm)	Ref.
WN/CdS	Schottky (p-n)	24,130	19.8	>420	[55]
CdS@Ni2P	n-n	287	-	480	[53]
NiWO4/CdS nanosheets	Z-scheme	26,430	22	420	[54]
PtP2@C/CdS (10 wt%)	-	9760	41.67	420	[56]
3 wt% WC/CdS	Schottky	9180	14.3	420	[59]
red-P/Co9S8	Schottky	4362	-	UV-Vis	[119]
25 wt%-ZnCo2S44/g-C3N4	S-scheme	6619	-	Lab-solar	[120]
CuNiS/g-C3N4	p-n	752.8	-	800	[121]
T-COF/CdS	Z-scheme	500	37.8	365	[122]
${\mathrm{Mn}}_{0.2}{\mathrm{Cd}}_{0.8}$S/Ni-MOF-74	type-II	7104		>420	[123]
CdS@NiS/MOFs	p-n	42,700	13.23	450	[124]
ZnCdS/UIO-66(Ce)	type-II	3958	-	UV-Vis	[125]
ZnIn2S4-MOF layer	-	28,200	22.67	350	[126]
Pt@NH2-UiO-66/CdS	type-II	38,000	40.3	400	[127]
CN/amorphous-MoS2/Pt	Schottky	5830	8.51	400	[107]
CoSx-rGO/TiO2 (10%)	-	256.97	14.62	365	[98]
TiO2@CuS	LSPR-Schottky	2467	13.4/3.7	UV-vis/near-IR	[99]
BiSx/TiO2 (1.0 wt%)	-	803.2	3.86	365	[100]
NiWSx-ND/TiO2 (3 wt%)	Schottky	4580	13	365	[103]
ZnCo2S4/TiO2	S-scheme	5580	11.5	420	[104]
amorphous MoO2/TiO2	co-catalyst	880.3	-	UV-Vis	[105]
TiO2-ZnIn2S4	direct Z-scheme	214.9	36.7/11.6	UV/Vis	[106]
MoS2/Fe2O3/g-C3N4	Z-scheme	7820	-	Vis	[109]
Co3S4/g-CN	Ohmic	536.0	7.55	400	[111]
NiS@g-C3N4-30	co-catalyst	3297	-	Vis	[42]
NiS/g-C3N4	co-catalyst	244	-	420	[112]
PtS/g-C3N4 nanosheets	co-catalyst	1072.6	45.7	420	[113]
Pt/MoS2/g-C3N4	co-catalyst	1595.3	30.9	435	[114]
ZnS-NiS2/g-C3N4	dual co-catalyst	302.7	1.8	>420	[115]
2D SnS/g-C3N4 nanosheets	Z-scheme	818.93	0.55	Solar-simulator	[116]
NixPB-rGO/CdS	p-n	5790	9	420	[52]
MoS2 /ZnIn2S4	co-catalyst	221.71	11.8	420	[134]
CdS-SV@CuS(5%)	-	1654.53	6.51	450	[137]
W18O49/Cd0.5Zn0.5S	S-scheme	147,700	45.3	<460	[166]
MoS2@MoO3	S-scheme	12,416.8	8.43	500	[92]
CdSe/CuInS2 microflowers	p-n	10,610.37	48.97	420	[72]
3% Ni-doped ${\mathrm{Mn}}_{0.5}{\mathrm{Cd}}_{0.5}$S	-	108,300	-	420	[73]
a-NiCuSx/TiO2 (3:1)	-	427.9	34.67	UV	[167]
Co9S8@CdIn2S4	type-I	4604	-	Vis	[47]

. This is due to the fact that MS compounds, with their immense tunable stoichiometry, crystal structure, and electronic band structure, serve as superior research grounds over other semiconductors. Moreover, we saw that the interaction mechanism and strength of the MS with other co-catalyst and incorporated materials can easily be tuned by modifying the particle size and morphology. On the other hand, such size reduction and corresponding controllable electronic modification make the MS materials the perfect playground for studying light-matter interaction, which is highly important for designing and developing a new generation of effective photocatalytic systems. This survey also showed that in addition to the new emerging photocatalysts, the conventional ones also still require more studies and stay promising for vast-scale application. Especially promising is combining such well-known photocatalysts with porous materials to overcome the short excited state lifetime and enhance the light absorption capability. Evidently, hierarchical nanostructured are also promising photocatalysts for H₂ evolution, which however investigated only in a limited manner and materials. Hence, more research is required in order to find suitable co-catalysts, to optimize nanostructural morphology and finding suitable doping agent to enhance light absorption range. In this regard, developing photocatalytsts with earth-abundant metal based co-catalysts (such as NiS₂, MoS₂, CuS, etc.) and nobel metal-free co-catalysts should be more prioritized. Moreover, building hierarchical structures based on low-dimensional and metal free materials such as graphene, g-CN, graphene oxide, etc. is more promising also to reduce metal contamination and production cost. Another interesting and very promising photocatalytic systems would be membrane systems having advantage of photocatalysis and membrane separation. Eventually, a better understanding of charge generation, separation, and transport through the photocatalytic process, reactions at the surface, interaction of heterogeneous photocatalyst with the environment is necessary to design stable photocatalysts with an enhanced performance towards scalable applications. One of the most important researches in this direction should be developing facile, controllable and cost effective synthesis methods for photocatalytic systems.