Recently, the continued use of fossil sources of C has led to increased concerns about the depletion of fossil fuel reserves, the increase in greenhouse gas emissions, and other environmental problems. As a C-free and environmentally friendly energy carrier, sustainable H₂ production by photocatalytic water splitting has been widely investigated since the pioneering work on photoelectrochemical water splitting by using TiO₂ photoanode by Fujishima et al. in 1972 [1]. However, owing to the sluggish multielectron reduction kinetics and a large positive change in the thermodynamic Gibbs free energy (△G⁰ = +237.2 kJ mol^-1) for the uphill water-splitting reaction [6], most of the photocatalysts for H₂ evolution generally show low quantum efficiencies and stability [2-4]. Therefore, to significantly accelerate the reaction kinetics and decrease the thermodynamic barrier, electrocatalysts (used as cocatalysts) are usually loaded onto the surfaces of different semiconductors to achieve highly active and durable photocatalytic H₂ generation by water splitting [5, 6]. In general, noble metal nanoparticles, such as those of Pt, Au, and Ag, could be used as effective cocatalysts to boost the photocatalytic H₂ evolution over semiconductors. However, the high cost and low abundance of noble metals significantly limit their widespread applications. Among the noble-metal-free cocatalysts, Ni-based cocatalysts have attracted extensive attention owing to their low costs, high activity, and stability [7-11]. In 1980, Domen et al. [12] demonstrated for the first time that SrTiO₃ modified with NiO cocatalyst could realize steady overall splitting of water vapor into H₂ and O₂. Subsequently, they also demonstrated that the existence of Ni metal at the interface of NiO and SrTiO₃ played an important role in promoting electron transfer from SrTiO₃ to NiO [13]. Since then, Ni-based H₂-evolution electrocatalysts and cocatalysts have attracted increasing interest from different research communities [14-21]. As shown in Fig. 1, almost 50 and 120 papers on Ni-based H₂-evolution electrocatalysts and cocatalysts, respectively, have been published since 2015. Indeed, a significantly increasing trend in the number of publications concerning Ni-based H₂-evolution electrocatalysts and cocatalysts is expected over the next few years. Consequently, it is urgently required, though challenging, to develop more efficient and durable Ni-based cocatalysts for photocatalytic H₂ generation, which will continuously inspire new ideas and discussions and stimulate new and exciting research advances in the field of photocatalytic H₂ generation.

Significant progress has been made in photocatalytic or electrocatalytic H₂ evolution by using Ni-based cocatalysts and electrocatalysts over the past decade [22, 23]. In 2011, Dubois and his coworkers first reported a synthetic Ni complex electrocatalyst, [Ni(P₂^PhN^Ph)₂](BF₄)₂, (P₂^PhN^Ph = 1, 3, 6-triphenyl-1-aza-3, 6-diphosphacycloheptane), that exhibited superhigh H₂-production activity in protonated dimethylformamide solution [24]. Surprisingly, the electrocatalyst could achieve turnover frequencies of 33, 000 and 106, 000 s^-1 in dry acetonitrile and in the presence of 1.2 mol/L of water, respectively, at a potential of -1.13 V (Fig. 2(A)). Subsequently, in 2012, Krauss and his coworkers reported a robust and highly active solar H₂-generation system with > 600, 000 turnovers at pH = 4.5 (Fig. 2(B)), in which CdSe nanocrystals capped with dihydrolipoic acid (DHLA), ascorbic acid, and soluble Ni²⁺-DHLA were employed as the light absorber, electron donor, and H₂-evolution cocatalyst, respectively [25]. Since the publication of the two aforementioned famous reports, Ni-based H₂-evolution cocatalysts and electrocatalysts have been extensively investigated by using different modification strategies [16, 17, 22, 23]. However, reviews on Ni-based H₂-evolution cocatalysts for photocatalytic H₂-evolution over different semiconductors are still limited [22, 23]. Most of the reviews published only summarized the progresses in the development of the various Ni-based cocatalysts used in semiconductor photocatalytic H₂-generation systems. However, the modification strategies for further enhancing the activity and stability of Ni-based cocatalysts and the identification of the exact active sites and surface reaction mechanisms have been nearly completely neglected. Therefore, it is timely to summarize the past research efforts in order to promote further developments in this area. We believe that it is also of significant importance to exploit new Ni-based H₂-evolution cocatalysts and develop new modification strategies, and deeply investigate the exact active sites and surface reaction mechanisms to boost the selectivity, stability, and activity towards photocatalytic H₂ evolution. Hence, in this review, we will focus on the progress of different kinds of Ni-based H₂-evolution cocatalysts. The design principle, modification strategies, and functional mechanisms of these Ni-based H₂-evolution cocatalysts will be thoroughly discussed and summarized.

In general, semiconductor-based photocatalytic reactions are strongly dependent on three reaction processes: light harvesting, charge transport and separation, and surface catalytic reaction processes [26-32]. For photocatalytic H₂ evolution over powdered semiconductors with Ni-based cocatalysts, there are four processes during the photocatalysis (Fig. 3). They include (1) light harvesting; (2) separation of photogenerated charge carriers; (3) bulk and interface (between the semiconductor and cocatalysts) transport of charge carriers; and (4) the trapping of charge carriers by the cocatalysts and surface electrocatalytic reactions occurring on them. Clearly, the charge recombinations observed in the bulk, interface, and surface of semiconductors are unfavorable factors, which could significantly decrease the quantum yield of H₂ evolved [33-35]. Thus, in comparison with increasing the light harvesting by narrowing the bandgap of the semiconductor, both improving the transfer/separation and accelerating the surface reaction kinetics of the photogenerated charge carriers seem to be much more crucial in enhancing the overall photocatalytic efficiency [36, 37]. Therefore, on one hand, research efforts are continuously taken to design various kinds of nanostructured heterojunction photocatalysts to enhance the photocatalytic activity for H₂ evolution by promoting the bulk and interfacial separation of the photogenerated charge carriers [38-40]. Until now, a large number of semiconductor heterojunctions [37, 38, 41-43] such as type Ⅰ heterojunctions, type Ⅱ heterojunctions, p-n heterojunctions, Schottky-based junctions, surface heterojunctions [44], phase junctions [45-50], and direct Z-scheme [51-53] and S-scheme [54] heterojunctions have been widely designed to produce favorable built-in electrical fields for boosting the charge separation, thus achieving enhanced photocatalytic H₂ evolution. On the other hand, more attention has also been paid to accelerating the sluggish surface reaction kinetics by loading proper cocatalysts over semiconductors. Up to now, noble-metal-based [55-57] (i.e., Pt, Pd, Au, and Ag), earth-abundant-metal-based [6, 58-61] (i.e., Fe, Co, Ni, Cu, Mo, and W), metal-free C-based [26, 62-67], and their hybrid cocatalysts have been widely exploited and applied in photocatalytic H₂ evolution. Notably, among these cocatalysts, the Ni-based cocatalysts have attracted growing interest in photocatalytic H₂ evolution owing to their superior features, which include low cost, easy fabrication and modification, and high efficiency [23]. In particular, these surface H₂-evolution cocatalysts play multifunctional roles in enhancing the charge separation and improving the H₂-evolution activity and stability of photocatalysts under illumination [68-71]. Accordingly, wholistic consideration and optimization of all these above processes during photocatalysis are crucial for developing highly efficient H₂-evolution photocatalysts. Notably, in this review, the integration design of other modification strategies and the loading of suitable Ni-based cocatalysts will be thoroughly proposed and discussed.

Ni-based H₂-production cocatalysts over semiconductor light harvesters are generally selected from the excellent electrocatalysts. In theory, any excellent H₂-evolution electrocatalyst is a potential cocatalyst candidate that can boost the photocatalytic H₂ evolution over semiconductors. Hence, the development of promising H₂-evolution electrocatalysts on different electrodes should be highly encouraged, which can facilitate the exploitation of H₂-evolution cocatalysts. Herein, the fundamentals of Ni-based H₂-evolution electrocatalysts will be simply discussed to provide a direction for rational selection and design of the Ni-based H₂-evolution cocatalysts. Since the early 1960s, the mechanism of alkaline HER on metallic Ni has been widely proposed to be Volmer-Tafel or Volmer-Heyrovsky, depending on the electrolyte pH (Eq. (1)-(5)) [16, 17]. Notably, both alkaline and acidic media reveal similar Tafel mechanisms (Eq. (3)). However, the Tafel process in acidic media occurs 2 orders of magnitude faster than that in alkaline electrolytes. Particularly, in the alkaline HER process, an HO-H dissociation reaction (Volmer step, Eq. (2)) is generally involved prior to the Tafel step (Eq. (3)). Clearly, the chemisorption of H_ad is crucial for the Tafel step. Thus, the selection of proper HER electrocatalysts could be guided by optimizing the surface chemisorption energy of H, H_ad, on different electrode materials. As shown in Fig. 4, Trasatti's volcano plot (cathodic HER current densities at a given overpotential vs. the metal-H binding energy), is a straightforward method of comparing and predicting the HER activity of a new metal or alloy [72]. In fact, as observed in Fig. 4, metallic Ni is the most active non-precious electrocatalyst in the alkaline HER process [16, 17, 73]. However, severe deactivation of metallic Ni, caused by the formation of hydride species, and rapid corrosion in acidic media have limited its largescale application [74]. Consequently, continuous efforts have been devoted to exploit more active, stable, and overall pH-applicable Ni-based H₂-evolution electrocatalysts in recent years [9, 10, 14]. The reduction of a proton (Eq. (1), (3), and (4)) or a water molecule (Eq. (2), (3), and (5)) is the key step in the mechanism of H₂ evolution in an acidic or alkaline electrolyte, respectively. For example, the famous NiO/Ni-CNT HER electrocatalyst exhibited 10 mA cm^-2 current density at ~81 mV overpotential in alkaline media [75]. Ni, NiO, and oxidized CNT have been identified as the efficient active sites of the H adsorption step, the preferential absorbing region of the OH^- produced by the Ni sites (i.e., the positive charge and higher number of empty d orbitals of Ni²⁺, stronger electrostatic affinity of NiO for OH^- ions), and a substantial substrate for growing the NiO/Ni heterostructure, respectively. Until now, a combination of experimental and computational studies has revealed numerous Ni-based H₂-evolution electrocatalysts [76]. Some of the popular representative Ni-based H₂-evolution electrocatalysts are summarized in Table 1 according to their different compositions, including alloys/metals, carbides, nitrides, oxides, phosphides [77-79], sulfides, selenides, and complexes.

(1)

(2)

(3)

(4)

(5)

Moreover, to further enhance the activity and stability of these H₂-evolution electrocatalysts, five designing strategies have been developed, namely composition engineering, [149-157] nanostructure engineering [158-167], interfacial engineering [168-174], surface engineering, [175], and hybrid engineering [176-181] (Fig. 5). Composition engineering can be divided into three types, including doping (metal, nonmetal, and dual doping), introduction of defects/vacancies, and design of crystal phases and alloys. Clearly, Ni-based H₂-evolution electrocatalysts composed of binary metal alloys or compounds, such as PtNi [80, 182], Ni-WN [82], Ni-Mo [89], Ni-Mo [17, 89], Ni-Co [183, 184], NiMoN [97], MoNiNC [105], CoNi(OH)_x [110], W-doped Ni-P [116], NiP_1.93Se_0.07 [123], Ni-Co-P [124], NiFeP [18], Fe-Ni-S [136], Ni-W-S [20, 138], NiCo₂S₄ [140], Ni_0.33Co_0.67S₂ [143], Ni_0.33Co_0.67Se₂ [129], Fe(Ⅲ)-doped NiS₂ nanosheets [144], NiCo₂S₄ double-shelled hollow nanospheres [141], Ni_1-xCo_x Se₂ [146], and Ni_0.89Co_0.11Se₂ [146], have been widely exploited. Secondly, nanostructure engineering can be performed by controlling the facets, nanostructures, and pores or by tuning the dispersion, shape, and nanosize. Typically, various nanostructures, including Ni-Fe LDH hollow microspheres [83], NiCo_2.7(OH_)x nanocages [109], Ni₂P nanosheets [114], NiP_x nanospheres [122], and Ni₃S₂ nanosheets (210) [135], have been utilized in electrocatalytic H₂ generation. Thirdly, interfacial engineering can be accomplished by constructing high-quality heterojunction interfaces. Commonly, several heterojunction interfaces such as MoNi₄/MoO₂@Ni [22], Ni(OH)₂/MoS₂[113], Ni₃N/Ni [185], Pt/NiO@Ni [186], MoS₂/Ni₃S₂ [187], hierarchical Ni₃S₂-NiOOH/Ni foam[112], Ni-P/MoS_x [128], Co₉S₈-Ni_x S_y [130], and Ni/Ni₃S₂ [133] have been successfully constructed by coupling Ni-based and other types of electrocatalysts. As a typical example, Dai's group synthesized nanoscale NiO/Ni-CNT heterostructures via a low-pressure heat calcination route (Fig. 6) [75]. The NiO/Ni core shell-like structure formed over the mildly oxidized CNT could be clearly confirmed. Impressively, the NiO/Ni-CNT heterostructures could achieve an overpotential of less than 100 mV at the large current density of 100 mA cm^-2. The interface of NiO and Ni was found to serve as active sites for the HER. The interfacial NiO sites with positive charge and greater number of unfilled d orbitals acted as adsorption sites for the OH^- generated by H₂O splitting preferentially through the strong electrostatic interactions, whereas a nearby Ni site would facilitate H adsorption and thus promote the Volmer step. However, the stability of NiO/Ni-CNT is not satisfactory. Subsequently, the chemically stable Cr₂O₃ shell has been demonstrated to improve the HER stability of the Ni/NiO heterostructure [108, 188], which could protect the Ni core from oxidation and prevent the NiO component from being reduced to metallic Ni, thus retaining the active Ni/NiO interface sites for durable HER in an alkaline electrolyzer. Similarly, Kuang et al. [131] prepared a bimetal (Ni and Mo) sulfide-based hybrid nanowire (NiS₂/MoS₂ HNW) heterostructure composed of NiS₂ nanoparticles and MoS₂ nanoplates via the in situ chemical vapor deposition process (Fig. 7(A) and (B)). The resulting NiS₂/MoS₂ HNW on a simple glassy C electrode shows a much smaller overpotential of 204 mV, along with a smaller Tafel slope of 65 mV dec^-1 in alkaline media, owing to the formation of lattice interfaces of NiS₂-MoS₂ heterojunctions. Fourthly, surface engineering can be carried out by tailoring the surface oxidation state/roughness or creating functional groups or dangling bonds [189]. Finally, hybrid engineering of the Ni-based electrocatalysts can be performed through a hybrid of different cocatalysts or through a hybrid of electrocatalysts and nanocarbons [176]. For example, various nanocarbon materials have been used as conductive supports to develop a number of advanced hybrid H₂-evolution electrocatalysts, such as FeNi/NC [81], Ni-RGO [84], CoNi@NC [90], Ce-Ni/RGO [91], NiMoN_x/C [96], Ni-MoC_x/C [103], Ni/NiO/N-rGO [107], Pt/Ni-MoC_x/C black [190], Ni₂P/C cloth [115], Co_0.5Ni_0.5P/CC, [121], Ni_2-xCo_x P/rGO [129], Ni_x S_y nanowalls/N-doped graphene foam [132], S-Ni₅P₄ NPA/CP [134], WS₂/graphene/Ni [139], Ni-W-S/CF [20], Ni-Mo-S/C, [142], and porous Ni₃S₂@NPC [145]. Notably, it is also expected that density functional theory (DFT) calculations can be used to aid in the design of new cocatalyst materials that render photocatalytic H₂ generation more active and durable. Undoubtedly, these design and development strategies of Ni-based electrocatalysts could be imitated and extended to the rational construction and fabrication of Ni-based cocatalysts for highly efficient and durable photocatalytic H₂ evolution.

As is well known, according to the active site, hydrogenases, which can convert the released protons into H₂, can be classified into [NiFe], [FeFe], and [Fe] hydrogenases. The structures of [NiFe] and [FeFe] hydrogenases are shown in Fig. 8 [191]. Clearly, as observed from this figure, the [NiFe] center in [NiFe] hydrogenase contains the Ni ion, implying that Ni is active towards the production of molecular H₂. Luckily, Ni-based H₂-production electrocatalysts have been widely investigated, and can be also employed as H₂-production cocatalysts on photocatalyst particles for realizing highly improved photocatalytic activities for H₂ evolution. The roles of Ni-based cocatalysts in the photocatalytic H₂ evolution are summarized in Fig. 9. Clearly, apart from increasing the number of active sites for H₂ evolution, the Ni-based cocatalysts can also play various important roles in improving the photocatalytic efficiency, such as promoting charge separation and transport, lowering the electrochemical H₂-evolution overpotential, enhancing the catalytic H₂-evolution kinetics by decreasing the activation energy, and reducing the fabrication costs. Thus, it is obvious that suitable Ni-based cocatalysts play decisive roles in achieving highly active low-cost photocatalytic H₂ evolution. Hence, the design, exploitation, and application of Ni-based earth-abundant H₂-production electrocatalysts in photocatalytic H₂ evolution should be further studied.

However, the design of highly efficient Ni-based H₂-production cocatalysts for photocatalysis is more challenging than the design of Ni-based H₂-production electrocatalysts, owing to the complex separation and transfer of charge carriers between the semiconductor light harvesters and the cocatalysts during the photocatalytic processes. In other words, electrons must first be generated and separated in the semiconductor light harvesters, and then transferred into the Ni-based H₂-production cocatalysts to drive the final surface electrocatalytic reactions. At this point, for designing highly efficient photocatalytic H₂-evolution systems, surface electrocatalytic H₂ production over Ni-based electrocatalysts, charge separation/transfer involving the light-harvesting semiconductors, and the interfacial charge separation between them should be simultaneously considered and optimized. However, previous investigations were more focused on the design and optimization of Ni-based cocatalysts. For example, many Ni-based H₂-evolution cocatalysts are simply loaded onto the surfaces of semiconductors to boost the photocatalytic H₂ generation. Therefore, it is not surprising that the H₂-evolution activity is still very low in many photocatalytic systems involving the Ni-based cocatalysts. This review will highlight the various potential modification strategies based on system integration to simultaneously optimize the charge separation/transfer in photocatalytic processes and the active sites in the Ni-based cocatalysts, which will be thoroughly discussed in the following sections. It is expected that these considerations will provide helpful insights and practical guidance on the design and development of highly efficient photocatalytic H₂-evolution systems involving Ni-based cocatalysts.

Precise identification of the H₂-evolution active sites is crucial for rationally designing and synthesizing highly active and durable Ni-based electrocatalysts or cocatalysts for high-efficiency electrocatalytic or photocatalytic H₂ production. Previously, for electrocatalytic H₂ production by using molecular complexes, it was revealed that the reduction and initial protonation of the metallic center could generate a common metal hydride intermediate, which could further lead to the formation of H₂ by a second protonation step.[192-195] However, studies are limited on the nature and structural features of the H₂-evolution active sites in Ni-based electrocatalysts, as well as on the structural and bonding scenarios possible at the active sites of the cocatalysts under light irradiation. Recently, various in situ observation technologies, such as in situ XAS studies, in situ EPR measurements, and operando Raman spectroscopy, have been used for identifying the exact active sites for H₂ evolution over Ni-based cocatalysts. These in situ technologies can provide more direct experimental evidences for confirming the exact charge separation and transfer mechanisms. On one hand, several in situ techniques have been recently employed to precisely identify the actual active sites of Ni-based electrocatalysts and the active species involved in the electrocatalytic H₂-evolution process. For example, to identify the actual active sites, Hu et al. successfully synthesized 2D metallic Ni nanosheets by transforming Ni-thiolate coordination polymer by using 1, 4-benzenedithiol as the ligand (Fig. 10(A)) [111]. The resulting catalyst could reveal a current density of 10 mA cm^-2 at an extremely low overpotential of 80 mV for the electrocatalytic HER (Fig. 10(B)). In situ XAS studies of the X-ray absorption near-edge structure (XANES) (Fig. 10(C)) and X-ray absorption fine structure (EXAFS) (Fig. 10(D)) spectra revealed that ultrathin Ni NSs with large areas acted as the actual catalytically active sites for H₂ generation. Notably, the trace S adsorbed on the Ni surface plays a crucial role in promoting water dissociation (Fig. 10(E)). This study highlights the deep investigations and the importance of the surface coordination chemistry of electrocatalysts in increasing the HER activity. Similarly, Ma et al. synthesized 2D Ni disulfide (NiS₂) by using Ni(OH)₂ on a Ni foam as the substrate for electrocatalytic HER (Fig. 11(A)) [196]. The resulting catalyst exhibited superior activity, with an overpotential of 67 mV at 10 mA cm^-2 (Fig. 11(B)). XANES/EXAFS spectra were used to reveal the different coordination environments of the samples before and after activation (Fig. 11(C) and (D)). The in situ XAS studies confirmed the presence of Ni-S and Ni-Ni bonds in the active sample, further suggesting that the metallic Ni⁰ with a trace amount of sulfide should be the possible active sites. All these results highlight the significance of in situ techniques in identifying the actual HER active sites. On the other hand, the in situ techniques could also be applied to the identification of the active sites of Ni-based cocatalysts for photocatalytic H₂ evolution. For example, Indra et al. demonstrated that Ni-based cocatalysts could serve as the active catalytic species on the surface of porous g-C₃N₄. They photodeposited Ni species as the cocatalysts on the surface of g-C₃N₄, based on XPS and EPR tests (Fig. 12(A) and (B)), which significantly prolonged the illumination time. A continuous increase in the Ni⁰ signal intensity was detected, though not all the Ni²⁺ species were reduced to Ni⁰ at the same time. Ni nanoparticles are formed only during the photocatalytic process, and reoxidize quickly when exposed to air during isolation. Therefore, the Ni²⁺ species could act as a cocatalyst to reduce protons to H₂ (Fig. 12(C) and 7(D)) [197]. Additionally, it should be pointed out that a Raman peak was recorded at 2530 cm^-1 at the potentials relevant to H₂ evolution, which corresponded to the S-H stretching vibration of the amorphous MoS_x -H moieties, as revealed by the operando Raman spectroscopy [198]. The absence of Mo-H (or Mo-D) stretching vibrations rules out the possibility of Mo centers acting as catalytic H₂-production sites. These results directly confirmed for the first time that the S atoms in amorphous MoS_x are the actual active H₂-evolution sites. In theory, the exact active sites of Ni-based cocatalysts could also be revealed by operando Raman spectroscopy, which is highly desirable. Additionally, based on first-principles method, it was demonstrated that increasing the coverage of O atoms on the surface of Ni atoms could lead to the increase of ΔG_H* [199], suggesting that partially oxidized Ni(111) surfaces with a considerably low coverage of O are theoretically the best catalytic centers for the electrocatalytic HER owing to the optimal electronic properties. In future, the combination of various in situ observation technologies and first-principles method should be a promising approach to identifying the exact H₂-evolution active sites over Ni-based cocatalysts.

Investigation of the reaction pathways over the active sites of Ni-based electrocatalysts or cocatalysts guides the design and development of more efficient electrocatalytic or photocatalytic systems for H₂ generation. In this regard, DFT should be a promising tool for revealing the exact reaction pathways leading to H₂ evolution. For instance, DFT calculations clearly revealed that the threefold hollow (TFH) Ni site and the neighboring P site are responsible for the adsorption of H [200]. Importantly, the H atom is mostly adsorbed on the TFH Ni sites on the (0 0 1) or (0 0 2) surface, due to unstable H adsorption on the P sites of the (0 0 2) and (0 0 1) surfaces [200]. Particularly, a high electron density in the d-band that extends to near the Fermi level was confirmed for the Ni₂P (001) surface, resulting in a strong interaction with H. In comparison with the simultaneous H adsorption reaction pathway over the Ni₂P (001) surface (Fig. 13(A)), a lower reaction barrier or activation energy was observed for consecutive H adsorptions on TFH Ni sites and then on P(Ⅱ) sites on the Ni₂P (001) surface (Fig. 13(B)), which indicated that the consecutive adsorption reaction pathway over the Ni₂P surface might actually occur before the Volmer step. In future, it is expected that a combination of DFT calculations and in situ techniques should be applied to identify the exact active sites and reaction pathways of Ni-based cocatalysts under light irradiation, which could be beneficial for precise exploitation of efficient next-generation semiconductor-cocatalyst composite H₂-generation systems.

To date, various Ni-based electrocatalysts, including Ni [197, 201-204], Ni(OH)_x [9, 205-218], NiS_x [219-232], [Ni(TEOA)₂]Cl₂ [233], NiO_x [234], NiB_x [235, 236], Ni₃C [237, 238], NiP_x [197, 239-245], and Ni(dmgH)₂ [246, 247] have been widely utilized as cocatalysts to increase the photocatalytic H₂-evolution activity over various organic and inorganic semiconductors [23]. Their activities are summarized in Table 2. From the table, it should be obvious that a simple combination of a cocatalyst and a semiconductor is involved in many studies. Herein, several original binary systems should be highlighted. For example, in 2010, Xu and coworkers demonstrated for the first time that 1.2 mol% of NiS modified CdS NPs showed the highest H₂ production rate of 2.18 mmol h^-1, with a QE of 51.3% at 420 nm (Fig. 14) [248]. Subsequently, in 2013, the same group first reported that 1.1 wt% NiS cocatalyst on g-C₃N₄ could achieve a H₂-production rate of 48.2 mmol h^-1, corresponding to an apparent quantum efficiency of 1.9% at 440 nm (Fig. 15) [219]. In 2011, Yu et al. found that CdS nanorods loaded with 23 mol% Ni(OH)₂ cocatalyst revealed the highest visible-light H₂-production rate of 5085 μmol h^-1 g^-1, with a QE of 28% at 420 nm (Fig. 16(A) and (B)) [212]. In 2013, Yu et al. first demonstrated that 0.5 mol% Ni(OH)₂-modified g-C₃N₄ composite photocatalyst exhibited the highest H₂-production rate of 7.6 μmol h^-1 (with an AQE of 1.1% at 420 nm, which is very close to that of 1.0 wt% Pt/g-C₃N₄ (8.2 μmol h^-1) (Fig. 16(C) and (D)) [205]. In 2014, Cao et al. for the first time reported that a combination of monodispersed Ni₂P nanoparticles, as the cocatalyst, and CdS nanorods, as the photosensitizer, was found to result in highly active and stable photocatalytic H₂ evolution from an acidic aqueous solution, with a TON of 26, 300 for the first 20 h of irradiation (Fig. 17) [249]. In 2018, Li's group for the first time demonstrated that a hybrid of 1% Ni₃C cocatalyst and CdS nanosheets exhibited the highest photocatalytic H₂ evolution rates of 357 and 450 μmol h^-1 in 0.25 mol/L Na₂S-Na₂SO₃ and lactic acid, with apparent quantum yields (AQYs) of 7.58% and 8.72% at 420 nm, respectively (Fig. 18(A) and (B)) [238]. More recently, the same group reported that 15 wt% Ni₃C cocatalyst nanoparticles decorated on g-C₃N₄ could display the maximum H₂-production rate of 15.18 mol h^-1, corresponding to an AQY of 0.40% at 420 nm [237]. These investigations perfectly highlighted the innovation of direct integration of Ni-based cocatalysts, inspired by the novel Ni-based H₂-evolution electrocatalysts. However, more attention should be paid to further improving these simple hybrids of cocatalysts and semiconductors for achieving more efficient and durable photocatalytic H₂ evolution. Therefore, in this section, the various strategies for improving H₂ production over Ni-based cocatalysts will be thoroughly discussed, which should be fully integrated and well designed to maximize the H₂-evolution over different cocatalysts.

As is well known, the unique structural features of hierarchical micro-/nano- photocatalysts can simultaneously enhance the light harvesting and improve the various types of kinetic factors in photocatalysis (such as the charge carrier kinetics, adsorption and diffusion kinetics of reactants, and surface reaction kinetics) [28, 359-362]. Therefore, the fabrication of hierarchical photocatalysts has become a popular strategy to improve the photocatalytic H₂ production over Ni-based cocatalysts. For example, Zhang et al. prepared hierarchical core/shell hydrogenated TiO₂ (H-TiO₂) nanothorns/C nanofiber (CNF) composites through the solvothermal method (Fig. 19(A) and (B)). The optimal NiS quantum dot (QD)-decorated H-TiO₂/CNF composite could achieve an enhancement of about 12.3 times in the solar-driven H₂ generation rate under AM 1.5 irradiation, in comparison with that of TiO₂/CNF nanostructures (Fig. 19(C)). It is believed that the unique architecture leads to many advantageous features such as greatly improved charge transfer and separation, increased accessible surface area and surface donor density of the composites, and increased light harvesting, all of which result in enhanced photoactivity (Fig. 19(D)). In another study, Li et al. [208] synthesized layer-like Ni(OH)₂-decorated ZnIn₂S₄ sub-microsphere photocatalysts for increasing the photocatalytic H₂ production via a facile in situ deposition method (Fig. 20(A)). Loading of 5 mol% Ni(OH)₂ cocatalysts over the ZnIn₂S₄ sample could achieve an 18-fold activity enhancement for photocatalytic H₂ production (Fig. 20(B)). Apparently, after introducing Ni(OH)₂ into ZnIn₂S₄, the interfacial charge transfer process, H₂-evolution kinetics, and light harvesting were simultaneously improved (Fig. 20(C) and (D)), which were responsible for the activity enhancement. Similarly, Chen et al. synthesized CdS photocatalysts modified with hierarchical flowerlike Ni/Ni(OH)₂ cocatalysts. The optimal H₂-evolution rate of Ni/Ni(OH)₂/CdS (3735 μmol g^-1 h^-1) photocatalyst is 1.7 times higher than that of CdS/Pt photocatalyst. The UV-vis diffuse reflection spectrum reveals an improved absorbance in the visible-light region, indicating that the hierarchical flowerlike Ni/Ni(OH)₂ cocatalysts could improve the light-absorption ability of the CdS photocatalysts. Furthermore, the CB level of pure CdS (-0.7 V) is of a higher magnitude than the potential of Ni²⁺/Ni (-0.23 V vs. SHE), indicating that Ni can serve as a cocatalyst to accelerate the separation and transfer of photogenerated electrons from the CB of CdS to the Ni/Ni(OH)₂ cocatalyst [262]. Additionally, hierarchical Ni-doped ZnIn₂S₄ microspheres [340], Ni (Ni²⁺)-salt-modified CdE (E = Te, Se, and S) QD hollow nanospheres [348], Ni-supported CdIn₂S₄ spongy-like spheres [251], and Ni_x P cocatalyst embedded 3D N-TiO₂/g-C₃N₄ heterojunctions [322] have been also demonstrated to exhibit significantly enhanced photocatalytic H₂-evolution activity owing to the improved light harvesting and charge separation. In future, it is expected that hierarchical photocatalysts with Ni-based cocatalysts could be widely designed and utilized in photocatalytic H₂ generation.

Besides the hierarchical nanostructures, utilization of surface sensitization has been demonstrated to be an efficient strategy to increase the light harvesting for enhancing the photocatalytic H₂-generation activity over wide bandgap semiconductors. Clearly, three typical sensitization strategies, namely, dye and QD sensitization, and introduction of surface plasmon resonance (SPR) effect, have been developed for the fabrication of visible-light-driven photocatalysts for various applications [364]. Among them, dye sensitization has been demonstrated to be the simplest way to enhance the visible-light absorption of wide bandgap semiconductors. For example, Kong et al. demonstrated that the hybrid of graphene with NiS_x cocatalyst obtained by in situ chemical deposition method could result in a two-fold improvement in the H₂ generation rate (Fig. 21(A)), with the highest quantum efficiency of 32.5% at 430 nm, compared with that of pristine NiS_x under visible light irradiation [365]. Thus, confined growth of NiS_x on a 2D conductive graphene substrate could greatly improve the transfer of photoelectrons from the excited Eosin Y (EY) to the NiS_x cocatalyst, leading to greatly enhanced photocatalytic H₂ evolution (Fig. 21(B)). In this work, the earth-abundant NiS_x/G nanohybrid represents a highly efficient cocatalyst for low-cost photocatalytic H₂ generation in dye sensitization systems. In another example, Peng et al. used the Ni complex ([BzPyN(CH₃)₂]₂[Ni(mnt)₂]1(mnt^2-= maleonitriledithiolate)) as the photosensitizer to enhance the photocatalytic performance of CdS nanorods. The results revealed that the optimized photocatalytic H₂-evolution performance of the CdS-Ni complex could reach 350 μmol h^-1, with an average value of the AQY of ~18% after 10 h of irradiation. According to the UV-vis diffuse reflectance spectra obtained, the loading of the Ni complex could significantly improve the light absorption of the CdS nanorods, thus leading to the photogeneration of more electron-hole pairs [350]. Furthermore, CdS QDs have been employed to sensitize various semiconductors with wide bandgaps, such as TiO₂ and g-C₃N₄, for enhanced photocatalytic H₂ evolution [366-370]. Wang et al. [315] fabricated CdS QD-sensitized TiO₂ hybrid photocatalysts with dual Ni(OH)₂ and CNT cocatalysts that were anchored on the TiO₂ surface for photocatalytic H₂ generation. The synergetic effect of the introduced Ni(OH)₂ and CNT cocatalysts resulted in an optimal photocatalytic H₂-evolution rate of 12 mmol g^-1 h^-1 (Fig. 22(A)). The UV-vis diffuse reflectance spectra clearly confirmed the role of the CdS QDs in visible-light sensitization (Fig. 22(B)), whereas the time-resolved PL emission decay curve revealed improved interfacial charge separation and transfer (Fig. 22(C)). As a result, increased separation efficiency and enhanced visible-light absorbance led to a fundamental improvement in the visible-light H₂-evolution activity (Fig. 22D). This study might provide some interesting insights into designing and constructing visible-light-responsive photocatalyst composites for realizing enhanced solar-to-hydrogen conversion efficiencies. Additionally, it is known that the localized surface plasmon resonance (LSPR) in several metals (e.g., Au, Ag, Cu, and Bi) [371-383] and nonmetal compounds (e.g., W₁₈O₄₉ and Mo_x W_1-xO_3-y) [384-388] could be used to sensitize wide bandgap semiconductors and achieve visible-light or NIR photocatalysis. Therefore, combining plasmonic materials and Ni-based cocatalysts should be an interesting strategy to boost visible-light H₂ production. For instance, Luna et al. successfully synthesized Au and/or Ni nanoparticle-loaded TiO₂ (commercial P25) photocatalysts for enhanced H₂ production [275]. The highest H₂-production rate from aqueous methanol solution was obtained over NiAu/TiO₂ catalyst (Ni:Au atomic ratio of 5:1 and total metal content 0.5-1 at%) (Fig. 23(A) and (B)). A plasmonic band in the wavelength region of 540-560 nm was clearly observed for NiAu/TiO₂, suggesting that LSPR could improve the absorption of visible light (Fig. 23(C)). Thus, the synergetic effect between the Au NPs and Ni(O) clusters on TiO₂ could be attributed to the enhanced photocatalytic H₂ evolution. Similarly, the enhanced H₂-evolution photoactivities of the Au-Cu₂O and Au-NiO nanoparticles deposited on TiO₂ were confirmed, and attributed to a synergetic effect between the enhanced electron transfer from TiO₂ to the Au-M_xO_y system and the LSPR of the Au nanoparticles [303]. In future, all of these sensitization strategies are expected to be coupled with the Ni-based cocatalysts and deeply studied with respect to the mechanism.

Developing wide spectrum photocatalysts is another strategy to directly increase the visible- and/or NIR-light harvesting for enhanced photocatalytic H₂ evolution. Simply speaking, several popular strategies, including impurity doping, [258, 286, 306, 326, 336, 340, 344] introduction of structural defects [9, 344, 389, 390], and formation of solid solutions [231, 280, 391], have been used to develop wide spectrum semiconductors. Typically, various metals and non-metals have been doped into different semiconductors such as TiO₂ [392-397] and g-C₃N₄ [398-405] to fundamentally narrow their bandgaps and thus enhance the visible-light absorbance and photoactivity. In fact, in some cases, the doped Ni species in sulfide and oxide semiconductors themselves could act as cocatalysts in increasing the photocatalytic H₂ evolution.[336, 340, 406-408] Further loading of the Ni-based cocatalyst over these wide spectrum semiconductors could lead to improved photocatalytic H₂ evolution [258]. More interestingly, Ni²⁺-to-metal charge-transfer excitation and the substitution of Ni²⁺ in semiconductor photocatalysts can narrow the bandgap energy of a semiconductor, resulting in visible-light photocatalytic activity [409, 410]. For example, Baliarsingh and his coworkers prepared Ni-Zn/Cr-CO₃^2- LDH photocatalysts for visible light photocatalytic H₂ evolution (Fig. 24(A)) [411]. The Ni-Zn/Cr-CO₃^2- LDHs with the Ni: Zn ratio of 75: 25 (LDH4) show the highest photocatalytic activity of 1915 μmol g^-1 h^-1 (Fig. 24(B)). The diffuse reflectance spectra indicated that the metal-to-metal charge-transfer excitation from Zn^Ⅱ/Ni^Ⅱ-O-Cr^Ⅲ to Zn^Ⅰ/N^iⅠ-O-Cr^Ⅳ could improve the absorption in the visible-light region. Furthermore, the partially filled 3d orbitals of Ni²⁺ played an important role in water splitting. The bands of Ni-3d-t_2g (-0.23 V) are energetically lower than the H⁺/H₂ reduction level. The photogenerated electrons can be easily excited from Ni-3d-t_2g to Ni-3deg owing to the d-d transition in the MO₆ octahedron (Fig. 24(C)). In addition, the oxo-bridged bimetallic linkages Ni(_Ⅱ)-O-Cr(_Ⅲ) could serve as catalytic redox centers. Therefore, the substitution of Ni²⁺ could narrow thebandgap energy and improve the photoactivity [411] (Fig. 24(D)). Apart from the modification of wide bandgap semiconductors, developing new wide spectrum photocatalysts, such as upconversion nanocrystals [412-414] and C QDs [415-418], has also attracted increasing attention. Considering the unique tunable bandgaps of graphene QDs (GQDs) with high water-solubility, biocompatibility, and photostability, Zhu et al. for the first time designed and fabricated the hybrid of Ni₂P cocatalyst nanoparticles and OH-functionalized GQD (OH-GQD) photosensitizer for photocatalytic H₂ production under λ > 420 nm light [7]. Under optimal conditions, the highest H₂ evolution rate of 1567 µmol h^-1 g^-1 was achieved, which is ~94 times higher than that of bare OH-GQDs and comparable with the catalytic activity of 1 wt% Pt/OH-GQDs sample (1683 µmol h^-1 g^-1) (Fig. 25(A) and (B)). The absorption spectrum of the OH-GQDs reveals a strong emission peak at ~540 nm, which corresponds to a bandgap of about 2.3 eV (Fig. 25(C)). After adding Ni₂P to the OH-GQD solution, the decrease in the emission intensity clearly indicates that electron-hole separation is promoted. It was demonstrated that Ni₂P can efficiently promote the charge-carrier separation and the photocatalytic H₂-evolution activity of OH-GQDs (Fig. 25(D)). This study suggested that metal-free OH-GQDs could be used to develop highly effective photocatalytic H₂ evolution systems by loading with suitable low-cost cocatalysts. Additionally, black P nanosheets have been widely employed as an exciting wide spectrum semiconductor for NIR photocatalytic H₂ production [419-428]. More importantly, earth-abundant cocatalysts such as CoP [429] and WS₂ [430] have been coupled with black P nanosheets for enhanced photocatalytic H₂ evolution. Thus, it is expected that an increasing number of wide spectrum semiconductors can be integrated with Ni-based cocatalysts for high-efficiency photocatalytic H₂ evolution.

Fabricating unique nanostructures has long been considered to be an advanced approach to shortening the transport distance and improving the separation of the photogenerated charge carriers, which could fundamentally improve the overall photocatalytic efficiency [27, 431-434]. Therefore, designing unique semiconductor nanostructures is also beneficial in enhancing the photocatalytic H₂ evolution over loaded Ni-based cocatalysts. To date, various kinds of semiconductor nanostructures, such as 0D QDs, 1D nanorods/nanowires, 2D nanosheets/nanoplates, and their assembled 3D structures, have been widely coupled with Ni-based cocatalysts to achieve enhanced photocatalytic H₂ generation. Typically, Ni-based cocatalysts have been widely utilized to boost the photocatalytic H₂ generation over different sulfide QDs. For example, Li and his coworkers synthesized Ni_h-CdSe/CdS core/shell hybrid, which was obtained from CdSe QDs stabilized by MPA and NiCl₂·6H₂O [227]. The maximum H₂-evolution rate of Ni_h-CdSe/CdS core/shell hybrid could reach 153 μmol mg^-1 h^-1 (Fig. 26(A)). The TON of the CdSe QDs and Ni could reach 15, 340 and 18, 000, respectively. The results revealed that the introduced Ni could enhance the absorptions of MPA-CdSe QDs and produce a redshift in them. Thus, the excellent photocatalytic performance could be attributed to the strong electronic interaction between the MPA-CdSe QDs and Ni²⁺ ions (Fig. 26(B) and (C)). The nanosecond transient absorption results in effective photogenerated charge transfer from the excited MPA-CdSe QDs to the Ni²⁺ active sites to produce H₂ (Fig. 26D). Notably, in this example, the doped Ni²⁺ ions played bifunctional roles in enhancing the photocatalytic H₂ evolution, acting as both the doping element and the cocatalyst. In another example, Li et al. [348] synthesized CdTe hollow nanospheres with Ni_h-QDs for photocatalytic H₂ evolution. The nanospheres with the diameter of 10 to 20 nm and a shell thickness of 3 to 6 nm showed excellent photocatalytic performances (Fig. 27(A) and (B)). It was found that the formation of the hollow structure resulted from H₂ gas bubbles. More importantly, compared with other ions, Ni²⁺ shows the best performance towards H₂ production. For the other ions, no hollow nanospheres could be found in the TEM images. These results suggest that the photocatalytic H₂ bubbles are important for assembling QDs into hollow nanospheres. Furthermore, the CB of the MPA-CdTe QDs is more negative than the redox potential of the Ni²⁺/Ni⁰ couple. Therefore, electrons could effectively transfer from the MPA-CdTe QDs to the Ni²⁺ to generate H₂ (Fig. 27(C)). Thus, the photocatalyst exhibited an excellent performance (Fig. 27(D)). Notably, the H₂ bubbles produced in situ under visible light irradiation may act as the centers of Nih-QD assembly (Fig. 27(E)). In other words, the Ni²⁺ species as cocatalysts over various semiconductor QDs have been demonstrated to be excellent photocatalytic H₂-generation systems, and should find more extensive application for practical photocatalytic H₂ evolution.

Apart from QDs, hybrids of 1D semiconductor nanorods/nanowires and Ni-based cocatalysts have been extensively constructed and applied in photocatalytic H₂ production. In general, the unique 1D semiconductor nanostructures possess several favorable features, including high accessible surface area, preferential electrical transport in the axial direction, and small transport distance in the radial direction, for improving the separation and transfer of photogenerated charge carriers from the semiconductor to the cocatalyst, thus achieving improved overall photoactivity [10, 27, 435-439]. For instance, Sun et al. synthesized crystalline Ni₂P cocatalysts to modify CdS nanorods for photocatalytic water splitting leading to H₂ evolution [329]. Loading of crystalline Ni₂P cocatalysts of size 10 nm at 0.5 wt% could result in an activity of about 1200 μmol mg^-1 h^-1 (Fig. 28(A) and (B)) [329]. The TON of Ni₂P could reach 3.27×10⁶ in 90 h, with a TOF of 36, 400. Time-resolved photoluminescence suggested that the photogenerated charges could be efficiently transferred from the CdS nanorods to the surface of the crystalline Ni₂P cocatalyst (Fig. 28(C) and (D)). The results confirmed that the Ni₂P cocatalyst could markedly improve the electron transfer and utilization in 1D CdS nanorods, thus resulting in significantly enhanced photocatalytic activity. More interestingly, enhancement in the hole transfer and utilization has also been demonstrated to be crucial in obtaining a much superior H₂-evolution photoactivity. Simon et al. impressively used Ni cocatalysts to accelerate the hole transfer on the oxidation side of 1D CdS nanorods. They replaced the slow hole-transfer process with two faster processes that involved a redox mediator and Ni cocatalysts (Fig. 29(A)) [255]. In this system, the value of pH plays an important role in achieving the highest formation rate of 63 mmol g^-1 h^-1, with the AQY exceeding 53% at 447 nm (Fig. 29(B)). Apparently, the photoexcited hole transfer from the semiconductor photocatalyst surface to the scavenger strongly depends on the redox couple ·OH/-OH operating as a shuttle, in which the ·OH is generated as an intermediate during the oxidation process. The decreasing driving force for H⁺ reduction by the CB electrons seems to not affect the higher H₂-production rate, confirming that the rate-limiting step is the hole transfer (Fig. 29(C)). An enhanced PL intensity corresponding to the production of ·OH and the accumulation of the fluorescent TA derivative (Fig. 29(D)) was clearly detected. Thus, a two-step mechanism of hole transfer employing a redox shuttle leads to a significantly enhanced photocatalytic H₂ evolution with Ni cocatalyst. This study opens new pathways to improve the activity and long-term photostability by boosting the hole transfer on the oxidation side, which could provide references for the design and fabrication of more efficient Ni-based photocatalytic H₂-evolution systems.

Additionally, Ni-based cocatalysts have been widely integrated with 2D semiconductor nanosheets or 3D micro/nanostructures for increased photocatalytic H₂ evolution [100, 208, 216, 250, 440-442]. It should be pointed out that Ni-based cocatalysts should be selectively deposited on electron-enrichment facets or locations of various nanostructured semiconductors, which could efficiently enhance the utilization rate of electrons to fundamentally increase the photocatalytic H₂ production. More importantly, the new fabrication methods should be proposed so as to achieve much stronger interfacial interactions between the semiconductors and Ni-based cocatalysts, which can significantly improve both the stability and the activity of Ni-based cocatalysts.

In general, Schottky junctions can be produced by combining metallic materials with higher Fermi levels or lower work functions and n-type semiconductors having lower Fermi levels or higher work functions, which results in upward band bending toward the semiconductor surface [31, 443]. Although too high a Schottky barrier is unfavorable for charge separation, a Schottky barrier of suitable height could serve as an efficient electron trap that prevents unexpected migration of electrons from the cocatalyst back to the semiconductor and thereby boosts the electrocatalytic activity of the cocatalyst with a high partial electron density [31]. On the contrary, metallic materials with lower Fermi levels or higher work functions can be used to combine with a p-type semiconductor to construct Schottky heterojunctions for facilitating hole separation from the p-type semiconductor to the contacted metal [443]. Here, a metal or nanocarbon with a higher work function could be employed to combine with n-type semiconductors to fabricate Schottky barriers for enhanced charge separation and H₂-evolution photoactivity. Until now, various metallic nanocarbon materials with higher work functions, such as graphene, C nanotubes, C QDs, C black, and C fibers, have been employed as cocatalysts to form Schottky heterojunctions and further improve the surface electrocatalytic H₂-evolution activity of different Ni-based cocatalysts [29, 215, 263, 297, 444-446]. For example, Li and coworkers [222] synthesized g-C₃N₄/C black Schottky junctions loaded with NiS cocatalyst. The resulting ternary g-C₃N₄/0.5% CB/1.5% NiS system could yield an average H₂-evolution rate of 992 μmol g^-1 h^-1, which was almost 2.51 times higher than that of g-C₃N₄/1.5% NiS (395 μmol g^-1 h^-1, Fig. 30). In fact, after inserting the C black interface layer between g-C₃N₄ photocatalyst and NiS cocatalyst, the Schottky junction introduced between the photocatalyst and C black could significantly promote the separation and transfer of the photogenerated electrons from the conduction band of g-C₃N₄ to the NiS cocatalyst, as the active sites, thus achieving enhanced photoactivity. Inspired by this work, the same group successively fabricated ternary Schottky-based heterojunctions (Fig. 31), including mesoporous g-C₃N₄/C nanotubes/NiS [224], g-C₃N₄/acetylene black (AB)/Ni(OH)₂ [206], CdS nanosheets/metallic C black/NiS₂ [11], and g-C₃N₄/C black/Co_1.4Ni_0.6P [317], for highly improved photocatalytic H₂ evolution. Among these systems, the metallic C-based Schottky heterojunctions play crucial roles in improving the charge separation from the semiconductor to the Ni-based cocatalyst and subsequently enhancing the electrocatalytic H₂-evolution activity and stability of Ni-based cocatalysts, thus resulting in fundamentally improved photocatalytic H₂ generation. To deeply reveal the multifunctional roles of the different nanocarbon interface layers, Li et al. introduced different nanocarbons into the interfaces between g-C₃N₄ and NiS for boosting the photocatalytic H₂-generation activity, such as graphite (G), C nanotubes (CNTs), reduced graphene oxide (rGO), C black, and AB. It was found that the introduction of the metallic amorphous AB and C black with higher electrical conductivity and weaker electrocatalytic H₂-evolution activity could enhance the H₂-generation activity by 3.17 and 2.57 times, respectively (Fig. 32(A)), in comparison with that of g-C₃N₄-1.0% NiS, which is in turn much higher than that of g-C₃N₄-0.5% rGO-1.0% NiS. Furthermore, the Nyquist plots (Fig. 32(B)), polarization curves (Fig. 32(C)), and Mott-Schottky plots (Fig. 32(D)) further confirmed the trifunctional roles of nanocarbons as conductive electron bridges, namely, promoting charge separation, boosting the H₂-evolution kinetics, and improving TEOA oxidation ability (Fig. 32(E)), thus greatly increasing the photoactivity. These results fully revealed that inserting a C layer between the semiconductors and Ni-based cocatalysts should be an appealing strategy for constructing highly efficient earth-abundant Ni-based photocatalytic H₂-evolution systems. However, to date, the combination of conductive graphdiyne C additive and Ni-based H₂-evolution cocatalysts has not been reported, but which is highly expected in the field of photocatalytic H₂ generation [447]. More interestingly, doped or hybrid nanocarbon materials with much superior conductivity and electrocatalytic activity are encouraged to be in situ or intimately integrated with Ni-based cocatalysts to develop highly durable and effective Schottky-based H₂-evolution photocatalysts [448-451].

Besides nanocarbons, metallic metals could also be employed to form Schottky heterojunctions to boost the surface electrocatalytic H₂ evolution over Ni-based cocatalysts. Fundamentally speaking, metallic elements as conductive electron bridges can also play multifunctional roles in increasing the overall photoactivity. For example, Li's group inserted multifunctional metallic Ni interface layers in g-C₃N₄ nanosheets/amorphous NiS heterojunctions to improve the photoactivity based on the mechanism of Schottky heterojunctions [69]. The results demonstrated that optimized g-C₃N₄/0.5 wt% Ni/1.0 wt% NiS could achieve the highest H₂ evolution of 515 μmol g^-1 h^-1, which was about 2.8 and 4.6 times higher than those of binary g-C₃N₄-1.0% NiS and g-C₃N₄-0.5% Ni, respectively (Fig. 33(A) and (B)). Similarly, the multifunctional roles of the metallic Ni interface layers, namely, improving the charge separation, accelerating the H₂-evolution kinetics, and enhancing TEOA oxidation ability, were further confirmed from the Nyquist plots (Fig. 33(C)), polarization curves (Fig. 33(D)), and Mott-Schottky plots (Fig. 33(E) and (F0), respectively. Consequently, the excellent synergistic effects of the metallic Ni interface layers and amorphous NiS cocatalyst result in significantly enhanced H₂-evolution photoactivity (Fig. 33(G)). Additionally, Wang and his coworkers confirmed the increased photocatalytic H₂ evolution activity and stability of Ni/NiO core-shell cocatalysts decorated on g-C₃N₄ sample, which was attributed to the formation of a Schottky heterojunction between the g-C₃N₄ and metallic Ni [234]. In this regard, other metallic elements such as the cost-acceptable Fe, Co, Bi, Mo, and W are also expected to form Schottky heterojunctions that can boost the activities of Ni-based H₂-evolution cocatalysts. More interestingly, in theory, other novel metallic materials such as 1T-MoS₂ [452-454], 1T-WS₂, MoO₂ [455, 456], MXene [457, 458], and Fe-Ni sulfide ultrathin nanosheets [136] can not only act as excellent photocatalytic H₂-evolution cocatalysts but also serve as electron acceptors to form high-quality Schottky heterojunctions that can markedly improve the photocatalytic H₂-evolution over various Ni-based cocatalysts with lower electrical conductivities.

Construction of type Ⅱ/p-n hetero(homo)junctions has been proven to be one of the appealing strategies for boosting charge separation and transfer through the formation of built-in electrical fields [31, 38, 41, 42, 46, 459, 460]. The deposition of Ni-based cocatalysts over these type Ⅱ/p-n hetero(homo)junctions could further improve the transfer of the separated charge carriers from semiconductors to cocatalysts for enhanced photocatalytic H₂ generation activity. For instance, Li and coworkers decorated NiS cocatalysts on CdS nanorods/g-C₃N₄ type Ⅱ heterojunction for photocatalytic H₂ evolution.[223] The resulting ternary g-C₃N₄-CdS-NiS composite could achieve the highest H₂-production activity of 2563 μmol g^-1 h^-1, which is about 1528 times higher than that of pure g-C₃N₄ (Fig. 34(A) and (B)). The unexpectedly high activity of the ternary g-C₃N₄-CdS-NiS hybrid can be attributed to the increased separation of the photogenerated electron-hole pairs as a result of the fabrication of the 1D/2D type Ⅱ heterojunction and the accelerated H₂-evolution kinetics upon loading the NiS cocatalyst (Fig. 34(C)). In another study, Liu et al. [461] designed a Cd_0.5Zn_0.5S (CZS) nanotwinned nanorod homojunction with an unanchored NiS_x cocatalyst for activation for H₂ generation.. The resulting system containing 0.03 wt% Ni could yield the highest internal H₂-production quantum efficiency of close to 100% at 425 nm by using Na₂S/Na₂SO₃ as the hole-scavenging electron donor (Fig. 35(A) and (B)). It is proposed that collision-contact mechanism between the CZS and NiS_x clusters in the suspension and type Ⅱ charge separation in the homojunction are responsible for the enhanced transfer of the photogenerated electrons to the clusters for photocatalytic H₂ generation (Fig. 35(C) and (D)). This work demonstrated the promise of photocatalyst-cocatalyst collision-contact interactions and CZS twinned homojunction for enhanced photoinduced H₂ production activity. Yan and his coworkers [207] employed Ni(OH)₂ as a cocatalyst in modified CdS/g-C₃N₄ core/shell nanorod heterojunction photocatalysts (Fig. 36(A)-(D)). The Ni(OH)₂ cocatalyst showed excellent stability during photocatalysis. The photocurrent of CdS/g-C₃N₄/Ni(OH)₂ photocatalyst was much higher than that of CdS/g-C₃N₄, suggesting that Ni(OH)₂ could serve as an efficient cocatalyst to improve the separation of the photogenerated electron-hole pairs (Fig. 36(E)). Thus, the electrons in the CB of CdS could be efficiently transferred to the surface of Ni(OH)₂ cocatalyst for participation in the photocatalytic reaction (Fig. 36(F)). Therefore, the ternary photocatalyst showed excellent photocatalytic performance (Fig. 36(G)).

Besides type Ⅱ hetero(homo)junctions, p-n heterojunctions have been widely constructed to promote charge separation through interfacial electrical fields [40, 462-465]. As shown in Fig. 37, the contact between p-type and n-type semiconductors with suitable Femi levels could result in p-n junctions under illumination, which could further induce the transfer of the photogenerated electrons and holes to the n-type and p-type semiconductors, respectively, according to the level of band bending and in-built electrical fields [52]. Thus, the formation of p-n junctions could lead to highly enhanced photocatalytic efficiency. For example, Zhang et al. designed the p-n heterojunction of NiS-modified CdS nanorods (Fig. 38(A)) [229]. The resulting NiS-CdS p-n heterojunction exhibited enhanced photocatalytic H₂-evolution activity (Fig. 38(B)). It is believed that the photogenerated electrons in p-type NiS could be injected into the n-type CdS nanorods under light irradiation owing to the formation of in-built electrical fields, induced by the p-n heterojunctions (Fig. 38(C) and (D)), thus resulting in highly increased photocatalytic H₂-evolution activity. However, here, the NiS cocatalyst plays the photosensitive role in boosting the photoactivity, instead of acting as active H₂-evolution sites. Zhang and his coworkers incorporated Ni nanoparticles in modified Co₃O₄/CdS p-n heterojunction photocatalysts [40]. The photocatalyst showed an excellent H₂-production rate of 5.12 mmol^-1 g^-1, which was 14.2 times higher than that of pure CdS (Fig. 39(A)). The steady-state PL spectra and the time-resolved PL spectra of Co₃O₄/CdS/Ni were much superior than those of Co₃O₄/CdS (Fig. 39(B) and (C)). It is believed that Ni nanoparticles could serve as active sites to enhance the separation of the photogenerated electron-hole pairs, thus resulting in a highly improved photocatalytic H₂-evolution activity (Fig. 39(D)). In future, it is expected that more and more type Ⅱ/p-n hetero(homo)junctions could be developed and combined with Ni-based cocatalysts for increasing the photocatalytic H₂-generation activity. Notably, Ni-based cocatalysts as active sites should be selectively and precisely loaded onto semiconductors containing the collected photogenerated electrons.

Direct Z-scheme junctions have also attracted increasing interest in heterogeneous semiconductor photocatalysis [39, 41, 51-53, 460, 466-468]. The concept of a third-generation Z-scheme photocatalytic system, namely the direct Z-scheme photocatalyst, was originally and specifically defined by Yu's group in 2013 [53, 469]. The formation mechanism of direct Z-scheme heterojunctions with staggered band structures (W₁ < W₂) is shown in Fig. 40. Clearly, suitable band structures and band bending result in favorable direct Z-scheme heterojunctions (Fig. 40(A)-(C)), rather than type Ⅱ heterojunctions (Fig. 40(D)) and impossible charge separation (Fig. 40(E)) [52]. In comparison with type Ⅱ/p-n heterojunctions, the direct Z-scheme heterojunctions not only possess electrons and holes with higher reduction and oxidation capacities for photocatalytic reactions, but also significantly reduce the construction cost, with the light-shielding effect removed [52, 53]. Until now, various direct Z-scheme heterojunctions, including g-C₃N₄/TiO₂ [469-473], g-C₃N₄/WO₃ [9, 474, 475], g-C₃N₄/Ag₂WO₄ [476], Zn-doped g-C₃N₄/BiVO₄ [477], Bi₂O₃/g-C₃N₄ [474, 478-481], g-C₃N₄/ZnO [482-485], AgCl@g-C₃N₄ [486], TiO₂/CdS [487-489], WO₃/Ag₃PO₄ [490], TiO₂/WO₃ [491], WO₃/CdS [492, 493], g-C₃N₄/MnO₂, [494], g-C₃N₄/MoS₂, [495], BiVO₄/g-C₃N₄, [496], g-C₃N₄/SnS₂ [497], and g-C₃N₄/BiOI [498, 499], have been extensively designed, fabricated, and applied in the different fields of photocatalysis. Interestingly, combining these direct Z-scheme systems with Ni-based cocatalysts should be a promising strategy to achieve significantly increased photocatalytic H₂ generation. For example, He et al. first fabricated novel WO₃/g-C₃N₄ hybrids through high-temperature calcination and subsequently loaded earth-abundant Ni(OH)_x cocatalyst by the method of in situ photodeposition.[9] The resulting ternary WO₃ (20 wt%)/g-C₃N₄/Ni(OH)_x (4.8 wt%) hybrid exhibited the highest photocatalytic H₂ generation rate of 576 μmol g^-1 h^-1) (Fig. 41(A)). The polarization curves clearly confirmed the cocatalytic function of Ni(OH)_x (Fig. 41B). More interestingly, the WO₃ (20 wt%)/g-C₃N₄/Ni(OH)_x (4.8 wt%) hybrid revealed much stronger ESR signals of ·O₂^- and ·OH radicals than WO₃ (20 wt%)/g-C₃N₄ and g-C₃N₄/Ni(OH)_x (4.8 wt%) (Fig. 41(C)), clearly confirming the direct Z-scheme charge separation mechanism during photocatalytic H₂ generation. Thus, the remarkably improved H₂ evolution performance was attributed to the combined effects of increased separation of the photoexcited electron-hole pairs and enhanced TEOA oxidation kinetics through the construction of the Z-scheme heterojunction (WO₃/g-C₃N₄), and improved H₂ evolution as a result of the loading of the cocatalyst (Ni(OH)_x) (Fig. 41D). More recently, a new concept of a step-scheme (S-scheme) heterojunction between WO₃ and g-C₃N₄ has been proposed and applied in photocatalytic H₂ generation in the presence of Pt cocatalyst [54]. However, so far, there have been limited reports on the integration of Ni-based cocatalysts and direct Z-scheme/S-scheme heterojunctions for photocatalytic H₂ production. Therefore, the development of Ni-based-cocatalyst-modified direct Z-scheme/S-scheme heterojunctions is highly expected in future studies.

Strengthening semiconductor/cocatalyst interfacial interaction is another appealing research direction to boost the electron transfer from the semiconductor to the cocatalyst and subsequently accelerate the surface electrocatalytic H₂ evolution over the cocatalysts. Generally, stronger interfacial interactions between semiconductors and cocatalysts could lead to a favorable interfacial effect on the electron-transfer process during photocatalysis, and are dependent on several important parameters that include the size, morphology, crystallinity, and composition of both the cocatalysts and the semiconductors, and their interfacial bonding and contact area [383, 500-502]. A simple method is to fabricate tight heterojunction interfaces through in situ high-temperature reactions. For instance, He et al. fabricated an intimate g-C₃N₄ nanosheet/NiS cocatalyst heterojunction through one-step in situ annealing of urea, thiourea, and Ni acetate (Fig. 42A) [228]. Here, thiourea could serve as both the S source for NiS and the precursor for g-C₃N₄. The resulting optimal g-C₃N₄ nanosheet/NiS cocatalyst heterojunction with intimate interfaces could achieve the maximum H₂-production rate of 29.68 μmol h^-1, which is comparable with that of a 0.5 wt% Pt-loaded sample (Fig. 42(B)). The remarkably enhanced H₂ evolution could be attributed to the acceleration of the sluggish H₂-evolution kinetics (Fig. 42C), promoted charge separation (Fig. 42(D)), and increased oxidizing ability of TEOA [228]. Therefore, this study highlights the role of high-quality semiconductor/cocatalyst interfaces in improving the H₂-generation photoactivities of binary hybrid photocatalytic systems.

Besides increasing the tightness of interfaces, construction of chemically bonded semiconductor/cocatalyst interfaces is also an effective method to strengthen interfacial interactions. Compared to the conventional interfacial interactions involving electrostatic or Van der Waals forces, the stronger chemical bonding interactions between the cocatalyst and the host semiconductor could provide firmer interfaces, thus essentially favoring synergistic enhancements in the charge separation and the photocatalytic activity and stability. For example, Wu et al. fabricated a chemically bonded Ni cocatalyst over S-doped g-C₃N₄ nanosheets (Ni-S:g-C₃N₄/Ni-SCN) based on S bonds for enhanced H₂ production under sunlight irradiation [264]. The resulting Ni-SCN could achieve excellent photostability and photocatalytic H₂ production rate of 3628 μmol g^-1 h^-1, with the apparent quantum efficiency being 17.2% at 420 nm (Fig. 43(A) and (B)). It is believed that the chemical bonding of Ni through the S could largely facilitate the charge separation during the photocatalytic process and promote surface electrocatalytic reactions over the Ni cocatalyst, thus resulting in the highly enhanced efficiency (Fig. 43(C) and (D)). This work highlights that the fabrication of a chemical bonding interface between a cocatalyst and the host photocatalyst is a relatively appealing strategy to strengthen the interfacial interactions for boosting the photoactivity and photostability, and is expected to be widely used to fabricate advanced photocatalysts based on Ni-based H₂-generation cocatalysts.

Additionally, increasing the effective interface contact area and in situ ion-exchange fabrication of interfaces [220] are also effective strategies for strengthening the interfacial interactions. Previously, it has been proved that 2D-2D interfaces with larger interface contact areas, in comparison with the 0D-1D, 0D-2D, and 1D-2D interfaces, could provide more charge transfer channels for boosting the photocatalytic activity [494, 503-505]. In future, it is expected that 2D-2D coupling and in situ ion-exchange interfaces could be extensively used in photocatalytic H₂ evolution by employing Ni-based H₂-evolution cocatalysts.

Increasing the number of active sites is another strategy to boost the surface electrocatalytic activity over semiconductors, driven by the photogenerated charge carriers. Generally, the catalytically active sites in cocatalysts are located at their surface coordination unsaturated sites, basal edge and structural defect sites, highly reactive facets, and surface dangling bonds [506-513]. Therefore, in comparison with a simple increase in the surface area, an improvement in the density of the intrinsic active sites of cocatalysts is much more crucial and appealing for enhancing their overall catalytic activity [33, 514-518]. In this regard, suitable modification strategies, such as advanced heteroatom doping and creating defect sites and alloys, are available for effectively increasing the intrinsic number of surface H₂-evolution active sites over Ni-based electrocatalysts. For example, DFT shows that a Ni-Mo alloy exhibits a lower H₂ adsorption free energy than Ni or Mo itself. Zhen and his coworkers fabricated NiMo@MIL-101 cocatalyst for photocatalytic H₂ evolution (Fig. 44(A)) [276]. As shown in Fig. 44(B), the NiMo@MIL-101 cocatalyst displays an excellent HER performance in dye-sensitization systems. The overpotential of ITO/EY-NiMo@MIL-101 electrode was -0.34 V, which was lower than that of ITO/EY-MIL-101. The lower overpotential suggests faster photogenerated electron transfer from the sur face of MoNi₄ to the H₂O (Fig. 44(C)). Consequently, the photocatalytic performance over ITO/EY-NiMo@MIL-101 was much higher than that over ITO/EY-MIL-101. The results were consistent with those of the photocurrent test (Fig. 44(D)) [276]. Additionally, improving the dispersion [206, 241, 256, 268, 269, 299] and exposing the surface atoms of the highly reactive facets [268] and porous structures [259, 265, 300, 301, 304, 309] of Ni-based cocatalysts have been used to increase the number of surface H₂-evolution active sites, and are expected to be extensively applied in the design for more efficient photocatalysis. Exposing the surface atoms of the highly reactive facets can increase the number of active sites. Zhen et al. used the highly exposed (111) facets of the Ni nanoparticles inserted in the frameworks of MOF-5 as a cocatalyst (Fig. 45(A)) [268]. The photocatalytic H₂ evolution rate of Ni@MOF-5 with the highly exposed (111) facets of Ni nanoparticles could reach 30.22 mmol h^-1 g^-1 (Fig. 45(B)). EIS test showed that the highly exposed (111) facets of the Ni nanoparticles significantly outperformed the highly exposed (200) facets in terms of electron transfer from MOF-5 (Fig. 45(C)). Furthermore, Ni@MOF-5 showed a lower overpotential of -0.37 V, which was attributed to the high dispersion (41.8%) and small sizes of the Ni particles. Thus, the photoexcited electrons would transfer to the H₂ production active sites (Ni), resulting in an efficient photocatalytic H₂ evolution performance (Fig. 45(D)). Yang and his coworkers designed a highly dispersed NiP_x cocatalyst coupled with g-C₃N₄ nanosheets via the precipitation and solid/gas-phase phosphorization method (Fig. 46(A)) [241]. The results showed that the NiP_x nanoparticles with a high dispersion were well loaded onto the surface of the g-C₃N₄ nanosheets. The maximum H₂ evolution rate of the g-C₃N₄ was about 4068.84 μmol g^-1h^-1 (Fig. 46(B)). The enhanced photocurrent further confirmed that the highly dispersed NiP_x cocatalyst could serve as effective active sites to decrease the recombination of the photogenerated electron-hole pairs, leading to a great improvement in the H₂ evolution (Fig. 46(C)). Furthermore, the highly dispersed NiP_x -cocatalyst-modified g-C₃N₄ nanosheets showed an excellent stable performance (Fig. 46(D)), indicating that the highly dispersed NiP_x could effectively serve as a cocatalyst to improve the photocatalytic performance. Furthermore, Chen et al. demonstrated that highly dispersed Ni cocatalysts could significantly enhance the photocatalytic H₂-evolution performances of TiO₂ photocatalysts. At lower loadings, the Ni nanoparticles were almost 1 nm in size, which suggests a very strong metal-support interaction between Ni⁰ and TiO₂ (Fig. 47(A) and (B)). On the other hand, at a Ni loading of 4 wt%, the size of the Ni nanoparticles was about 1-2 nm. The Ni nanoparticles could be easily observed at the Ni loading of 4 wt% (Fig. 47(C)), whereas, at a lower loading, it was difficult to detect the nanoparticles in the TEM images. The result of the photocatalytic H₂-evolution performance showed that a lower Ni loading resulted in a better photocatalytic performance (Fig. 47(D)), indicating that a highly dispersed cocatalyst is favorable for improving the photocatalytic H₂-evolution performance of TiO_2. In theory, any strategy that increases the number of active sites could be effectively used to improve the intrinsic electrocatalytic activities of Ni-based cocatalysts, and are highly expected to be applied in photocatalytic H₂ generation over Ni-based cocatalysts.

Besides increasing the number of active sites, enhancing the intrinsic activity of each active site is also an effective strategy to improve the surface electrocatalytic activity. Generally, all of the abovementioned (in section 2) five modification strategies for Ni-based electrocatalysts, namely, composition engineering, nanostructure engineering, interfacial engineering, surface engineering, and hybrid engineering, can be employed to boost the surface electrocatalytic H₂ evolution over Ni-based cocatalysts. Herein, some typical examples will be highlighted to provide ideas for the rational design and fabrication of Ni-based cocatalysts with enhanced intrinsic activity of each active site. For example, Fan and his coworkers developed Ni_x Co_1-xS₂-g-C₃N₄ for photocatalytic H₂ evolution (Fig. 48(A) and (B)) [330]. The optimal amount of photocatalytic H₂ produced over Ni_x Co_1-xS₂-g-C₃N₄ could reach 400.81 μmol under continuous visible light irradiation for 4 h, which was 2.5 times higher than that of NiS-g-C₃N₄ photocatalyst (Fig. 48(C)), suggesting that with different amounts of addition of Co, the Ni-Co boundary active surface activity could be enhanced to varying degrees. The synergistic effect between Ni and Co could effectively decrease the recombination of the photogenerated electron-hole pairs, thus achieving efficient photocatalytic water splitting leading to H₂ evolution. These results could also be confirmed from the fluorescence results (Fig. 48(D)). Additionally, Ma et al. used Ni-doped MoS₂ to enhance the photocatalytic performance of CdS nanorods (Fig. 49(A)-(D)) [519]. The resulting highest photocatalytic H₂-evolution rate of CdS/NiMoS was 24 mmol h^-1 g^-1, which was 2.4 times higher than that of CdS/MoS₂ (Fig. 49(E)). According to DFT calculation, the S-Mo Mulliken population in the NiMoS system was 0.35, which was slightly smaller than that of the MoS system. They demonstrated that Ni doping can increase the number of uncoordinated atoms at the edge sites of MoS₂ nanosheets, thus resulting in enhanced electron transfer across the CdS/MoS₂ interface, as well as increased H₂ reduction (Fig. 49(F)-(K)). Furthermore, Tian and his coworkers prepared Cu and Ni comodified TiO₂ photocatalysts via hydrothermal method (Fig. 50(A)-(C)) [271]. The photocatalytic H₂-evolution rate of Cu-Ni/TiO₂ was 13.5 mmol h^-1 g^-1, which was 2 and 2.5 times higher than those of Ni/TiO₂ and Cu/TiO₂, respectively (Fig. 50(D)). The photocatalytic performances of the photocatalysts were strongly affected by the molar ratio of Ni to Cu, which suggested that Ni could improve the intrinsic activities of the cocatalysts. The results of the Nyquist plots and linear scan voltammograms further confirm this point of view (Fig. 50(E) and (F)). Enhancing the intrinsic activity was also applied to transition metal phosphide cocatalysts for improved photocatalytic H₂ evolution. Man et al. found that Co-doped Ni₂P cocatalyst showed excellent photocatalytic performance with a dye photosensitizer (Fig. 51(A) and (B)) [520]. According to the calculation, the free energy value of CoNiP was close to 0 eV, and was more negative than that of Ni₂P. This result represents the strong H-adsorption ability of CoNiP, which suggests optimal HER performance (Fig. 51(C)). Therefore, the overpotential of CoNiP is much lower than that of Ni₂P, suggesting an improvement in the photocatalytic HER performance (Fig. 51(D)). Lu et al. improved the photocatalytic performance of g-C₃N₄ nanosheets through a synergistic effect between Ni₃B and Ni(OH)₂ cocatalysts. It was found that the photocatalytic H₂-evolution rate of the ternary photocatalyst could reach 352.4 μmol g^-1h^-1, which was 4.6 and 2 times higher than those of g-C₃N₄/Ni₃B and g-C₃N₄/Ni(OH)₂ binary photocatalysts, respectively (Fig. 52(A)). The much smaller EIS diameter of the ternary photocatalyst further confirms the better interfacial electron transfer rate (Fig. 52(A)). More importantly, the ternary photocatalyst showed an excellent HER overpotential, compared with that of the binary photocatalyst, which suggested that the synergistic effect between the Ni₃B and Ni(OH)₂ cocatalysts could enhance the intrinsic activity of the binary photocatalyst (Fig. 52(C)). Compared with the binary photocatalyst, the synergistic effect between the Ni₃B and Ni(OH)₂ cocatalysts could significantly improve the light absorption, decrease the recombination of the photogenerated electron-hole pairs, and accelerate the transfer of the photoexcited charges (Fig. 52(D)), thus achieving enhanced photoactivity. Xu et al. enhanced the intrinsic activity by hybridizing C, N-codoped Ni₂P and Fe₂P cocatalysts (Fig. 53(A)) [327]. The H₂-evolution rate of EY-g-C₃N₄-NiFeP photocatalyst could reach 13.81 mmol h^-1 g^-1, which was 10 and 5 times higher than those of EY-g-C₃N₄ and EY-g-C₃N₄-FeP, respectively (Fig. 53(B)). Compared with pure Fe₂P, the hybrid materials of Ni₂P and Fe₂P cocatalysts could significantly enhance the charge carrier lifetimes of g-C₃N₄, which suggested the generation of more photoexcited electrons during the photocatalytic reaction leading to H₂ production (Fig. 53(C)). The photocurrent result further confirmed the improved transfer of the charge carriers for the Ni₂P and Fe₂P hybrid cocatalysts, compared with that of pure Fe₂P (Fig. 53(D)). In future, besides composition engineering, it is expected that other modification strategies such as nanostructure engineering, interfacial engineering, surface engineering, and nanostructure engineering could be used to enhance the intrinsic activities of cocatalysts.

Commonly, the random deposition of cocatalysts on supports could result in the formation of large particles with irregular morphologies, due to unexpected agglomeration, thus leading to poor photoactivity owing to the low utilization rate of the active sites [522]. Recently, based on surface coordination and confinement growth, the rational design and fabrication of a highly atomically dispersed cocatalyst with a special shape or a precisely controlled size has attracted increased attention in photocatalysis [218, 241, 523]. For example, to obtain a far better photocatalytic performance, the interlayered spatial steric inhibition effect of layered materials has been employed to design highly dispersed cocatalysts on a layered photocatalyst by suppressing the growth of the cocatalyst particles [218]. Layered HNb₃O₈ was fabricated by a proton-exchange reaction (Fig. 54(A)) and further exfoliated to synthesize a-few-layered HNb₃O₈ nanosheets through intercalation of tetrabutylammonium cations into the layered HNb₃O₈ (Fig. 54(B)). The confined Ni(OH)₂ cocatalyst with ultrasmall size and very strong interactions could be realized in the interlayer region of layered HNb₃O₈ by a self-assembly approach (Fig. 54(C)). The resulting Ni(OH)₂/HNb₃O₈ sample showed the highest photocatalytic H₂-evolution rate of 951.0 μL after 4 h of irradiation, which was about 15.7 times higher than that of the sample prepared by a traditional cocatalyst modification method (Fig. 54(D)). In this work, the efficient strategy for designing a cocatalyst with a large number of active surface sites might guide the design and development of a highly dispersed cocatalyst-modified semiconductor for highly efficient photocatalytic H₂-evolution systems. Similarly, EY-sensitized Ni(111)@MOF-5 frameworks could achieve highly active and durable photocatalytic H₂ evolution under visible light irradiation owing to their high dispersion (41.8%) and small size of the Ni particles (~9 nm), and the high specific surface area of MOF-5 (2973 m²/g) [268]. More recently, Zeng et al. loaded monodisperse sub-15 nm Ni₂P nanoparticles on porous g-C₃N₄ nanosheets to fabricate 0D-2D heterointerfaces for achieving the highest photocatalytic H₂ evolution rate of 474.7 μmol g^-1 h^-1, with an AQY of 3.2% at 435 nm; this was possible due to the good electrical conductivity and electrocatalytic H₂-evolution activity of Ni₂P nanoparticles [523]. More interestingly, Ye and coworkers anchored well-dispersed hexacoordinated Ni(Ⅱ) species on CdS through a facile wet-chemistry approach by using the promising precursor metal-ethylenediaminetetraacetate complex (Fig. 55(A)-(C)) [253]. The resulting hexacoordinated Ni(Ⅱ) on CdS showed a high activity of 4.3 mmol h^-1 (with an AQE of 67.5% at 420 nm), which was much higher than that of Pt-CdS (Fig. 55(D) and (E)). Theoretical modeling indicated that the free energy of atomic H adsorption on the remaining carboxylate and Ni²⁺ is 0.065 eV, which is quite close to that of Pt (Fig. 55(F)). Dehydrogenation of the "carboxylic acid-Ni-H" intermediate was revealed to be the turnover-limiting step in photocatalytic H₂ evolution, based on thermodynamic modeling (Fig. 55(F)). [253] This study highlights that the design of Ni-based cocatalysts at the atomic/molecular scale could offer new opportunities for developing highly efficient photocatalytic H₂-production systems. In future, it is expected that promising single-atom cocatalysts [57, 524-526], especially those involving Ni-based materials, could be used in photocatalytic H₂ evolution.

Besides dispersion, nanoconfinement can not only induce strong interface interactions and greatly improve the electronic and band structures of cocatalysts, but also be beneficial in enhancing the stability of the interior component and increasing the number of exposed active surface sites for photocatalytic reactions. Therefore, the development of nanoconfined Ni-based cocatalyst architectures has been demonstrated to be an appealing and promising strategy to boost the high-efficiency photocatalytic H₂-production activity. Owing to the simple fabrication and outstanding performance, nanoconfined Ni@C core-shell nanoparticles have attracted wide interest for specific photocatalytic H₂-generation applications. For example, Zhang et al. synthesized Ni@C core-shell nanoparticle-cocatalyst-modified CdS nanorods to boost the photocatalytic H₂-evolution activity and stability by a simple one-pot strategy and self-assembly process (Fig. 56(A) and (B)) [281] The resulting nanoconfined Ni@C-modified CdS nanorods could achieve the highest visible-light photocatalytic H₂-production rate of 76.1 μmol g^-1 h^-1, which corresponded to an excellent quantum efficiency of 31.2% at 420 nm (Fig. 56(C)). The results revealed that the nanoconfinement effect in Ni@C core-shell cocatalyst pronouncedly resulted in a controlled Ni core size, intimate interfacial contact and enhanced charge transfer, an optimized electronic structure, and suppressed chemical corrosion (Fig. 56(D)-(F)). This work highlights the potential application of a highly active and low-cost Ni@C core-shell structure with significant nanoconfinement effect in boosting photocatalytic H₂ evolution. Similarly, C- [277-280] and NiO- [234, 282-284]coated core-shell cocatalysts have also been extensively employed to enhance the photocatalytic H₂ evolution over other photocatalyst systems. It is expected that greater attention should be paid to enhanced photocatalysis based on Ni-based cocatalysts with excellent nanoconfinement effects in future studies.

This paper provides a comprehensive review on the significant advances in Ni-based cocatalysts for boosting the heterogeneous photocatalytic H₂-generation activity. Until now, various Ni-based cocatalysts, including alloys/metals, oxides, hydroxides, sulfides, borides, carbides, nitrides, and complexes, have been developed and applied in photocatalytic H₂ evolution. Meanwhile, important design strategies based on increasing the light harvesting, enhancing the charge separation, strengthening the interfacial interaction, and improving the electrocatalytic activity are also thoroughly discussed. Obviously, Ni-based H₂-evolution cocatalysts show great potential owing to their unique cost and performance advantages. Therefore, it is not surprising that Ni-based H₂-evolution cocatalysts have received significant attention in the field of photocatalytic H₂ generation. Until now, although considerable progress has been made in the past decade, it is still urgently required to develop highly efficient Ni-based H₂-evolution cocatalysts and reveal the exact active sites and the underlying enhancement mechanism in these photocatalysts at the atomic and molecular levels.

On one hand, novel Ni-based H₂-evolution cocatalysts and semiconductor heterojunctions should be constantly developed and investigated. Recently, new Ni-based H₂-evolution electrocatalysts have provided great opportunities for designing new cocatalysts for photocatalytic H₂ evolution. In theory, all the modification strategies for developing highly efficient Ni-based electrocatalysts could be extended to the exploitation of Ni-based H₂-evolution cocatalysts. However, the difference between electrocatalysts and cocatalysts should be carefully noted in designing Ni-based cocatalysts. In other words, the separation of the photogenerated charge carriers is the biggest challenge, owing to the absence of external electrical fields. Thus, the construction of photocatalysts with effective in-built electrical fields is of great importance to enhance the photocatalytic activity. Clearly, the developments of Schottky junctions, p-n heterojunctions, direct Z-scheme heterojunctions, and their combinations by experimental and theoretical techniques should be significantly encouraged in the near future. Meanwhile, the separated electrons must quickly reach the active surface sites of Ni-based cocatalysts to drive the photocatalytic H₂ evolution. In this regard, the semiconductor/cocatalyst interfaces and the electrical conductivities of the cocatalysts must be carefully controlled and optimized to achieve maximum photocatalysis. In future studies, increasingly metallic, multifunctional and ultrathin 2D Ni-based nanosheets, core-shell, and confined nanostructures, single-atom systems, and stronger interfacial interactions should be realized through simple and scaleable fabrication methods, which are highly expected to be applied in high-efficiency and durable photocatalytic H₂ generation.

On the other hand, detailed investigations on the exact active sites and reaction pathways should be performed. Interestingly, various in situ observation technologies, such as in situ XAS studies, in situ EPR measurements, and operando Raman spectroscopy, are required, which are highly expected to be applied in identifying the exact active sites for H₂ evolution over Ni-based cocatalysts. Furthermore, the charge carrier transfer/separation dynamics in various heterojunctions and cocatalysts should be further identified by using the in situ technologies to provide more direct experimental evidences for confirming the exact charge separation and transfer mechanisms. In particular, DFT calculation is an effective and advanced supplementary tool that can provide highly useful indirect information, such as the exact reaction barrier, the adsorption energy of H, and the adsorption/dissociation properties of water molecules, which offer insights into the underlying charge carrier dynamics and reaction pathways. It is worth pointing out that the mechanisms in Ni-based H₂-evolution electrocatalysts still need to be investigated in future studies, as they can provide useful information for understanding the electrocatalytic reaction mechanism over the active surface sites of Ni-based cocatalysts.

Nevertheless, Ni-based H₂-evolution cocatalysts are interesting and will continue to be one of the hottest and most important research topics in photocatalytic H₂ generation. In particular, more attention should be paid to the application of Ni-based cocatalysts in overall water splitting leading to H₂ evolution. An increasing number of exciting Ni-based cocatalysts will be continually designed and applied in photocatalytic H₂ generation. It is expected that precise design and development of the active sites of Ni-based H₂-evolution cocatalysts can be readily achieved based on detailed structure/performance relationships in the near future, which will significantly advance the appeal of the promising field of photocatalytic H₂ production.

Acknowledgment

Dr. X. Li would like to thank the National Natural Science Foundation of China (51672089), Specical Funding on Applied Science and Technology in Guangdong (2017B020238005) and the State Key Laboratory of Advanced Technology for Material Synthesis and Processing (Wuhan University of Technology) (2015-KF-7) for their support. Dr. Q. Xiang would like to thank the National Natural Science Foundation of China (51672099). X. Chen appreciates the financial support from the U.S. National Science Foundation (DMR-1609061), the College of Arts and Sciences, University of Missouri-Kansan City and University of Missouri Research Board.