Recent Development of Self‐Supported Alkaline Hydrogen Evolution Reaction Electrocatalysts for Industrial Electrolyzer

Hydrogen (H2) energy is presumed to be the most promising alternative to replacing traditional fossil fuels to achieve the global mission of carbon neutrality. Electrocatalytic water splitting driven by green electricity has been regarded as an ideal method for large‐scale green hydrogen production with a minimal CO2 footprint. However, most of the reported electrocatalysts still suffer from large overpotentials and severe activity degradation at high current density (>1000 mA cm−2). Therefore, a comprehensive review to summarize the representative alkaline hydrogen evolution reaction (HER) electrocatalysts with large current densities is essential to guide the fabrication of promising industrial electrocatalysts. In this review, starting from the fundamental of water electrolysis, the design principles to acquire alkaline electrocatalysts with large current density and high stability are elaborated. The critical factors for achieving high‐performance electrocatalysts to meet industrial H2 production are proposed. Additionally, the key processes for preparing self‐supported electrodes are clarified. Afterward, the recently advanced self‐supported transition metal‐based electrocatalysts with high current density for alkaline HER are systematically summarized. Finally, personal perspective on future opportunities and challenges is highlighted. It is hoped this review can guide the rational design of self‐supported high‐current density electrocatalysts for future commercial H2 production.

DOI: 10.1002/aesr.202200178 Hydrogen (H 2 ) energy is presumed to be the most promising alternative to replacing traditional fossil fuels to achieve the global mission of carbon neutrality. Electrocatalytic water splitting driven by green electricity has been regarded as an ideal method for large-scale green hydrogen production with a minimal CO 2 footprint. However, most of the reported electrocatalysts still suffer from large overpotentials and severe activity degradation at high current density (>1000 mA cm À2 ). Therefore, a comprehensive review to summarize the representative alkaline hydrogen evolution reaction (HER) electrocatalysts with large current densities is essential to guide the fabrication of promising industrial electrocatalysts. In this review, starting from the fundamental of water electrolysis, the design principles to acquire alkaline electrocatalysts with large current density and high stability are elaborated. The critical factors for achieving high-performance electrocatalysts to meet industrial H 2 production are proposed. Additionally, the key processes for preparing self-supported electrodes are clarified. Afterward, the recently advanced self-supported transition metal-based electrocatalysts with high current density for alkaline HER are systematically summarized. Finally, personal perspective on future opportunities and challenges is highlighted. It is hoped this review can guide the rational design of selfsupported high-current density electrocatalysts for future commercial H 2 production.
electrocatalysts for the industrial alkaline HER due to their acceptable activity, low cost, large surface area, and good stability. However, although Raney nickel operates well at a high current density of 500-1000 mA cm À2 , it needs large overpotentials of %300-500 mV at 500 mA cm À2 and requires a cell voltage of 2.4 V for overall water splitting, which considerably exceeds the thermodynamic voltage of 1.23 V. [7] Thus, innovative strategies to develop high-efficiency, low-cost, and durable electrocatalysts are crucial for reducing capital expenditure and electricity consumption for the industrial alkaline HER. Recently, plenty of nonprecious metal-based electrocatalysts have been developed for alkaline HER, which show excellent HER performance at small current densities, but mostly experience severe degradation at large current densities (>1000 mA cm À2 ). [8] Under high current densities, massive gas bubbles are rapidly formed on the interface of electrocatalysts and electrolytes, which can seriously hinder the mass transfer of liquid, slow the electrons transfer, and reduce the exposure of active sites, leading to the decreased electrocatalytic activity and durability. [9] In addition to the catalytic activity, the practical industrial applications of catalysts also suffer from several issues. First, most metal-based catalysts exist in powders, which have to be immobilized on a current collector with binders (e.g., Nafion). This process will generate a high interfacial resistance between the catalyst and current collector, bury the active sites, inhibit mass transport, and increase produce cost. Second, for the electrocatalysts with low electrical conductivity, a conductive additive such as nanocarbon material is added, which is prone to oxidation and etched at high potential, thus deteriorating the electrode performance. Third, due to the low adhesive force between the catalysts and substrate, the electrocatalyst is easily detached from the substrate during long-term operation or large current density. [10] This gas-release behavior becomes more significant as the current density increases, which can quickly accelerate the catalyst peeling problem. Moreover, the reported synthesis routes of HER electrodes are energy intensive and challenging to scale up due to the complex and sensitive operating conditions such as such as high pressure and vacuum.
To address the earlier issues, it is imperative to design a simple method to directly prepare robust HER electrocatalysts on the conducting substrate with large current density and durable stability for large-scale industrial hydrogen production. For largescale industrial hydrogen production, high current density (alkaline electrolytic cell >500 mA cm À2 ) and durability (>100 h) are crucial. [11] The self-supported transition metal (TM)-based compounds (metal sulfides, phosphides, nitrides, etc.) have been developed as promising electrocatalysts for large current density HER in alkaline media. [11,12] The self-supported electrode means that the catalytic material is directly in situ grown on the conductive substrates (C cloth, Ni foam, etc.) or is vertically grown on the same metal substrate using an oriented solid-phase synthesis (OSPS) method. The self-supported electrodes imply a simpler electrode preparation process, a lower cost, abundant catalytic sites, rapid charge transfer, and avoiding electrocatalyst shedding. [13] All these features favor boosting the catalytic activity and long-term stability of self-supported electrodes for industrial large current density electrolysis.
Some excellent reviews on electrochemical water electrolysis in energy storage and conversion systems have recently been published. [12a,14] However, a comprehensive summary of the self-supported TM-based electrocatalysts with large current density and long-term stability for industrial alkaline HER is still needed. This review briefly introduces the types of water electrolysis, primarily focusing on alkaline water electrolysis. After deeply analyzing electrocatalysts' design principles for large current density and high stability in alkaline media, critical factors for achieving high-performance electrocatalysts are proposed. Additionally, the merits and demerits, the selection of substrates, and fabrication methods for self-supported electrodes are detailed, which are helpful for designing efficient HER electrodes. As the most important part of this review, we summarize the representative works for designing self-supported TM-based electrocatalysts with large current density and high stability for industrial alkaline HER. The selected electrocatalytic materials include TM sulfides (TMSs), TM selenides (TMSes), TM hydroxides (TMHs), TM phosphides (TMPs), TM carbides (TMCs), TM nitrides (TMNs), TM alloys (TMAs), and TM oxides (TMOs). Finally, we highlight the personal perspective on future opportunities and challenges. Hence, we think this review can serve as a valuable reference for researchers in this field.

Types of Water Electrolysis
In the water electrolysis process, water molecule is the reactant, which is dissociated into hydrogen and oxygen under the influence of direct current. [15] This process can be then classified into three main types: alkaline water electrolysis (AEL), PEM (or polymer electrolyte membrane) electrolysis (PEMEL), and solid oxide electrolysis (SOEL). [16] The AEL and PEMEL technologies operate at low temperatures from 30 to 90°C and allow up to 30 bar operation pressures. [17] Alkaline water electrolysis is operated in a concentrated alkaline (KOH or NaOH) electrolyte and requires a gas-impermeable separator to prevent the product gases from mixing. The electrodes consist of non-noble metals like nickel with an electrocatalytic coating. Troostwijk and Diemann first introduced alkaline water electrolysis in 1789. It has been considered a relatively mature, simple, and low-cost hydrogen production technology. [18] However, the development of AEL is still restricted by its low working current density (less than 0.5 A cm À2 ), the low electrolyzer efficiency (63-71%), the limited load operation range (20-100%), and multiple complex equipment in the large-scale state. PEMEL is based on PEM fuel cell technology. PEMs (e.g., Nafion membrane) with lower gas permeability replace asbestos and will significantly improve the hydrogen purity compared to AEL. Besides, the high current density (1.0-3.0 A cm À2 at low voltages), compact membrane electrode assembly design, high output pressure, and fast response make PEMEL an attractive hydrogen production technology for industrial application. [19] Nonetheless, the acidic electrolyte has limited the choices of electrodes/catalysts to platinum group metals (PGMs). The expensive PGMs and Nafion membranes significantly increase the cost of electrolyzers, which is the main challenge of PEMEL. [20] SOEL is also known as hightemperature or steam electrolysis, as gaseous water is converted into hydrogen and oxygen at temperatures between 500 and 900°C. [21] The energy consumption of electrolytic cells can be reduced to 3 kW h (Nm 3 ) À1 , and the stack efficiencies can be up to %100% due to positive thermodynamic effects on power consumption at higher temperatures theoretically. However, due to technical bottlenecks in material preparation at high temperatures, SOEL products are still in the laboratory stage. Table 1 presents a summary comparison of the three water electrolysis technologies from operation to economic parameters and from system details to some nominal features. SOEL provides only small stack capacities below 10 kW, compared to 6 MW for AEL and 2 MW for PEMEL. [17] Hence, the investment costs and the lifetime determine whether AEL or PEMEL is the most favorable system design for a large-scale application.
AEL is a more attractive option in terms of cost because most non-PGM electrodes and cheaper diaphragms can be used in alkaline electrolytes. It largely decreased the cell cost and enabled scaling up the electrolyzers to megawatt levels. However, a drawback for AEL is the difficulty of delivering large current density at low voltages because of the limited activity of catalysts and the large ionic resistivity of the diaphragm. Significant advances have been made in the design of nanostructured non-PGM catalysts for AEL. These catalysts showed appealing activities at low current densities. However, their performance at high current densities has rarely been studied, although it is more important for practical applications. The rational design of HER catalysts, which can be well operated at industrial temperatures, is highly desirable for practical AEL applications. Therefore, this review will focus on alkaline water electrolysis and summarize the state-of-the-art strategies for designing high-current density HER electrocatalysts, which are the closest to actual industrial applications.

Industrial Alkaline Water Electrolysis (AEL)
An industrial alkaline water electrolyzer system is composed of a cylindrical electrolyzer, a gas-liquid separation and purification unit, and a public utility unit, as illustrated in Figure 1a. The electrolyzer, as the most fundamental component unit, is composed of two electrodes, a diaphragm, and an alkaline electrolyte solution, generally potassium or sodium hydroxide (KOH or NaOH, respectively) ( Figure 1b). Alkaline water electrolysis uses electric energy to split water into hydrogen and oxygen gases. The chemical reactions are given in Equation (1)-(3). [22] At the cathode, water molecules are reduced by electrons to hydrogen (H 2 ) and negatively charged hydroxyl ions (OH À ). The produced hydrogen emanates from the cathode surface to recombine in a gaseous form, while the hydroxyl ions transfer through the porous diaphragm to the anode under the imposed electrical potential. At the anode, hydroxide ions are oxidized to oxygen and water while releasing electrons. Overall, a water molecule reacts to hydrogen and oxygen in a ratio of 2:1.
Overall reaction∶ H 2 OðlÞ ! H 2 ðgÞ þ 1=2O 2 ðgÞ (3) Subsequently, the two-phase liquid electrolyte and the product gas mixture will enter the gas separators. The generated H 2 and O 2 can be effectively separated by the adjacent gas separators and then undergo the degassing, drying, and purification processes to reach the desired level. [23] Meanwhile, the liquid electrolyte is pumped back into the electrolysis stack. The diaphragm in the middle of the cell is usually made of asbestos or ion-conducting polymers to separate the cathode side from the anode side, avoiding mixing the produced gases and ensuring electron transfer. [24] A commercial alkaline water electrolyzer generally operates at low temperatures, around 30-90°C in high-alkali-concentration electrolytes, and must reach the current densities ranging from 200 to 1000 mA cm À2 . [25] Therefore, developing electrocatalysts that perform well at large current density is critical for large-scale use. [26]

Mechanisms of Alkaline HER
The cathodic HER is a multistep reaction with the two-electron transfer. There exist two different reaction pathways for HER in the electrolyte: Volmer-Tafel (V-T) pathway and Volmer-Heyrovsky (V-H) pathway. [27] Volmer step∶ H 2 O þ e À ! H Ã þ OH À (4) where H* indicates the hydrogen intermediate adsorbed onto the active site (* represents the active site). The rate-determining step (RDS) is commonly evaluated based on the Tafel slope for both the V-T and V-H pathways. Moreover, the HER process is also a pH-dependent reaction. The experimental results show that the HER activity of most electrocatalysts in alkaline media (pH = 13) is 2-3 orders of magnitude lower than that in acidic media (pH = 1). [28] The H þ concentration causes this large performance difference in the electrolyte in the Volmer step. In acidic media, abundant H þ ions in the electrolyte are supplied to adsorb onto the active site via Volmer reaction directly, and then the H* ads go through Tafel or Heyrovsky step to generate H 2 gas (Volmer step in acid: H þ þ e À ! H*). However, the HER process is restricted due to the lack of protons in the basic medium (alkaline or neutral). An additional water dissociation process is required to produce protons for the subsequent steps in alkaline media, which is considered the RDS for the alkaline HER.
With the rapid development related to theories of computers, density functional theory (DFT) has become the link between theory and experiment. The combination of DFT and experiment can further help to understand the process of chemical reactions and explore the mechanism of alkaline HER. For example, Zheng et al. compared the hydrogen production rates of Pt/C and Ru/C 3 N 4 /C in an alkaline medium and found that although Pt exhibited the optimal H adsorption process (ΔG H %À0.02 eV), [29] the energy barrier of water dissociation kinetics reached 0.94 eV. This result presented that the rate-limiting step of Pt metal in alkaline medium was the Volmer reaction associated with water dissociation. In comparison, the ΔG H and energy barrier of water dissociation of Ru metal were both thermodynamically feasible, which is why Ru metal has better performance than Pt metal under an alkaline medium. The free energy diagram of reaction pathways for the alkaline HER demonstrates that the Volmer and Heyrovsky steps are needed to overcome the higher energy barriers due to the sluggish water dissociation process (Figure 2a). [30] Therefore, the intrinsic activities of alkaline HER electrocatalysts are determined by the hydrogen adsorption ability and controlled by the balance of the other three major factors: H 2 O adsorption/dissociation ability, OH desorption ability, and energy barrier required for dissociation of water molecules (Figure 2b). These additional variables put forward higher requirements for the design of HER catalysts, such as the ability to dissociate water molecules, a moderate affinity for H ads, and recombination to produce H 2 .
Based on the DFT calculations, Nørskov et al. suggested that the Gibbs free energy of hydrogen adsorption (ΔG H , Figure 2c) is an important descriptor for HER catalytic reaction for both metallic and the biological catalysts. [31] According to the Sabatier principle, ΔG H % 0 explains that the optimal surface for adhesion of the reaction intermediate has the appropriate binding energy, wherein adsorption is neither strong nor weak. Both the hydrogenases and nitrogenases are recognized as effective enzymes for HER. For example, as the largest category in hydrogenases, [NiFe] hydrogenases have been widely investigated via computational studies. [32] In Siegbahn's study, one hydrogen atom can bind exothermically to each active site (Ni or Fe) for [NiFe] hydrogenases, and the calculated ΔG H value for [NiFe]hydrogenase is close to zero, which is fulfilling its excellent catalytic activity. [33] Combining ΔG H with the exchange current density ( j 0 ) can form a volcanic numerical simulation relationship (volcanic relationship). The volcano plot relationship in Figure 2d displays precious Pt group metals at the top of the volcano, for which ΔG H is %0, indicating that the hydrogen evolution rate on the surface is the fastest. The catalytic reaction is the most favorable under the equilibrium condition. [34] Thus, the PGMs are considered the most active hydrogen evolution catalysts, consistent with the experiment results.
To reduce the energy barrier of water dissociation, Markovic's group developed the Ni(OH) 2 -decorated Pt catalyst and found that the water dissociation occurred at the edge of Ni(OH) 2 cluster to produce H intermediates that adsorbed near the Pt active sites. [35] The synergistic effect on Pt and Ni(OH) 2 cluster could accelerate the reaction rate of HER and boost the recombination of hydrogen atoms. Markovic and his colleagues also evidenced that 3d-M metal hydr(oxy)oxide clusters modified the Pt (111) metal surfaces to have the activities for the HER (bi-functional) that followed the order Ni > Co > Fe > Mn. [36] Meanwhile, blocking the active site by strongly adsorbed OH groups also reduces the HER activity. By analyzing different 3d metal hydroxides, Markovic et al. proposed that the Volmer step reaction rate also depends on OH's adsorption capacity. Too strong adsorption will cause the aggregation of OH on the catalyst surface, which may restrict the number of active sites for water molecules, thus hindering the occurrence of water dissociation. Chen et al. designed an alkaline HER catalyst of 1T-MoS 2 QS/Ni(OH) 2 with highly active and stable performance. [37] The DFT results suggested that the ultralow water dissociation energy barrier, near-zero binding free energy for H, and the low affinity of OH at edge 1T-MoS2 /edge Ni(OH)2 are the key to the high HER performance. This kind of efficient bifunctional electrocatalyst constructs the effective mixed heterostructure that can not only improve the alkaline HER activity of precious metal of Pt, but also be applicable to the nonprecious metal (Ni, Cu) hybrid catalysts and nonmetallic hybrid catalysts (Ni(OH) 2 /MoS 2 , Ni(OH) 2 / NiS 2 , MoS 2 /NiCo-LDH NWs, et al.). [38] Besides, the poor water binding energies also influence the production of M─H bonds in alkaline medium. Baek et al. [11] developed Ru nanoparticles dispersed in a nitrogenated holey 2D carbon structure (Ru@C 2 N). [39] The hybrid catalyst achieved excellent HER activity, and DFT clarified the origin of the high electrocatalytic activity of Ru@C 2 N. Their calculations revealed that the strong attraction to H 2 O would increase the H 2 O capture rate of Ru nanoparticles. Thus, the much easier dissociation of H 2 O into H and OH will  In practical application, commercial alkaline water electrolyzers require high current densities (>500 mA cm À2 ) with low overpotentials (<300 mV) over a long period for large-scale H 2 production. [40] Based on the above mechanism discussion, the current density delivered by an effective HER catalyst will increase exponentially with overpotential in the Tafel region when ignoring external factors. The association of overpotential and electrode kinetics can be related to the current density at an electrode by the classical Butler-Volmer (B-V) equation. [41] j j is polarization current density, j 0 is exchange current density, η is overpotential, β is the symmetry factor (usually %0.5), which refers to a one-electron electrochemical reaction, F is Faraday constant, R is universal gas constant, and T is thermodynamic temperature. It should be noted that this BV equation is valid only for single-electron transfer processes with the elimination of all mass transport effects. In the strong polarization region, the backward reaction can be ignored if and B-V equation can be simplified as a classic Tafel relationship, that is Tafel slope corresponds to the apparent electron transfer coefficient of a complex, multielectron electrochemical reaction. Hence, the Tafel slope (b) of the overall reaction is where n 0 is the number of electrons transferred before the RDS and α is the transfer coefficient. In theory, Equation (10) specifies the significance of regulating the reaction mechanism and RDS. The effective B-V equation reveals that the exchange current density and Tafel slope (b) exhibit strong dependencies on electrode potential (or overpotential . This means it will only take 30 mV of potential to increase the current density from 10 to 100 mA cm À2 and an additional 30 mV to deliver 1000 mA cm À2 . [42] Nevertheless, the Tafel slope will increase at a large current density in actual operation and then reduce the increasing current density rate as overpotential increases. Hence, a larger overpotential is needed to acquire high current density for HER electrocatalysts, and the activity will degenerate at high current density. [11,43] It is known that the catalytic reactions are not severely mass transport limited in a well-designed cell, particularly in AEL based on earlier analysis. [44] The intrinsic electrical conductivity and reaction energy barrier are the two critical factors for high current densities, which directly determine the intrinsic catalytic activity. The intrinsic electrical conductivity is dependent on the electron transfer capability of electrocatalysts, while the reaction energy barrier is related to the surface chemisorption properties of electrocatalysts. [45] Liu's studies have shown that the electron transfer process can be regulated both on the catalyst-electrolyte interface and catalyst-substrate interface ( Figure 3a). [46] It is noticed that the large current density commonly represents a large bias applied to the electrode, and a strong polarization condition is formed. The strong polarization potential can accelerate multielectron catalytic reactions and thus fast electron consumption. Therefore, the electron transfer will act as the RDS in the catalytic process. At the catalystelectrolyte interface, catalytic activity largely depends on the energy needed for the adsorption/desorption of intermediates and the rupture/formation of chemical bonds. Moreover, massive gas bubbles generation and fast reactants consumption as a mass transfer process will occur at the gas-liquid-solid interface under large current density (Figure 3b), which will decelerate the charge transfer and cause severe degradation in current density. The reaction energy barrier is another internal limiting factor related to the electrocatalyst's surface chemisorption properties. [47] The electrocatalysts should contain multiple components that can work synergistically to address each elementary reaction step's challenges, eventually lowering the energy barriers for the catalytic reaction. For example, integrating water adsorption/dissociation sites with hydrogen adsorption active sites has been verified to be an effective method to improve the intrinsic activity of the alkali HER catalysts. [48] Consequently, the cooperative optimization of electrical conductivity and reaction energy barrier with high hydrogen coverage is crucial for achieving high current density HER electrocatalysts.

Basic Principle for Designing High-Stability HER Electrocatalysts
Stability is another critical and challenging aspect of evaluating the industrial application for HER electrocatalyst. First, promising mechanical stability for electrocatalysts can maintain the compositional and structural integrity during harsh working conditions (strong alkaline electrolyte and large current density). The mass transfer process that occurs at the gas-liquid-solid interface will generate a massive gas bubble under large current density. The rapidly formed bubbles will accumulate on the interface of the catalyst and electrolyte, thus hindering the mass transfer of the liquid, slowing down the electron transfer, reducing the exposed active sites, and decreasing electrocatalytic activity. [49] More importantly, the bubbles releasing from the catalyst surface will cause shaking forces, which results in the catalysts being detached from the substrates. In this case, the superhydrophilic surface structures have been developed by forming textured structures (rough and/or porous, arrays, et al.), [9a,50] which are expected to promote the interface wettability, accelerate the electrolyte recharge and the gas bubbles releasing under large current density.
The conventional powdery catalysts, which need to be immobilized on a current collector by electrically insulating binding agents such as Nafion, are easily detached from the substrate during long-term or large current density operations due to the low adhesive force between the catalysts and the substrate. The catalyst microstructures with high mechanical strength to respond to the electrolyte impact need to be designed and manufactured to address this problem. [13c,51] Besides, the self-supported electrodes also arouse great attention, providing a seamless contact between the current collector and catalysts, ensuring rapid charge transfer, and avoiding undesirable issues like the peeling phenomenon and catalyst aggregation with powdered catalysts. [13a,52] Finally, the structural/compositional stability also should be well considered for electrocatalysts. The active element (such as Fe, Ni, and Co) may escape from the catalyst under the harsh electrolyte environment (strong acid or alkaline electrolytes). [53] Besides, some catalysts are prone to oxidation under the applied potential in the electrochemical process. [54] 3.3. Requirements and Strategies for Designing High-Performance Electrocatalysts for Industrial HER Based on the deep analysis of fundamental principles for the design high-performance electrocatalysts, we summarized that four key aspects should be thoroughly considered for the preparation of industrial HER electrocatalysts ( Figure 4).

High Intrinsic Activity
As discussed in Section 3.1, the intrinsic catalytic activity of the electrocatalyst is determined by the intrinsic electrical conductivity and reaction energy barrier. To date, considerable efforts have been made to improve the intrinsic catalytic activity of electrocatalysts by various strategies, such as integrating active sites (heteroatom doping), composition engineering (heterostructure), and creating defects/vacancies. [3c,55] All these strategies can effectively tune electrocatalysts' physical, electronic, or surface properties, which are closely associated with intermediate adsorption/desorption, free energy, or charge transfer kinetics, eventually altering the intrinsic electrocatalytic activity. [56] The detailed strategies operated on the electrocatalysts will be discussed in Section 5.

Fast Charge Transfer and Mass Transfer
It is known that intrinsic electrical conductivity is related to the electron transfer capability of electrocatalysts. However, common non-noble metal electrocatalysts exhibit semiconductor behavior. Therefore, exploring effective methods to improve the conductivity of electrocatalysts is crucial. Constructing binder-free electrocatalysts (e.g., self-supported structure) or introducing conductive materials (e.g., carbon, metals, and some metallic compounds) has been investigated to accelerate the electron transport process. [57] Herein, the conductive substrates provide large electroactive surface areas and can enhance the electrocatalyst conductivity, thus accelerating the electron transfer rate and finally enhancing catalytic performance. In particular, the metal-based supports could directly serve as a metal source to fabricate different morphologies of electrocatalysts.

Accelerating Gas Bubble Mitigation
Under high current density operations, the rapidly generated and accumulated gas bubbles will severely restrict the mass transfer and block the active sites. Therefore, rapid bubble removal is critical in minimizing the ohmic resistance and retaining the activity  under large current density operation. For superhydrophilic electrocatalysts, the discontinuous interface of solid/liquid/gas is formed, which is beneficial for decreasing the bubble adhesion on the surface. [58] Meanwhile, the interfaces between the electrolyte and the electrode surface become richer, increasing the penetration ability and thereby improving the electrocatalytic activity.

High Mechanical/Structural/Compositional Stability
After long-term and high current density electrochemical operation, the composition and structure of the catalysts may suffer from profound change under a strong oxidation/reduction environment. Besides, the electrocatalysts are prone to peel off due to the generation of massive bubbles, which would result in severe deterioration of catalytic performance. As mentioned in Section 3.2, mechanical, thermodynamic, structural, and compositional stability are all needed for HER electrocatalysts under ideal conditions. Strengthening the adhesion force between the catalysts layer and substrate and constructing self-supported structures have been verified to be promising routes to prevent the catalysts' detachment from the substrate during gas bubble release. [59] 4

. Design and Preparation of Self-Supported Electrodes
Up to date, significant progress in advanced electrocatalysts for water electrolysis with promising performance has been reported. [60] However, most of them are not allowed to operate at high current densities (usually <200 mA cm À2 ), resulting in the urgent need to develop highly efficient electrocatalysts with large current densities for practical industrial H 2 production. [61] It is worth noting that almost all of those reported electrocatalysts working at large current density are self-supported TM-based (metal sulfides, phosphides, nitrides, etc.) electrocatalysts in alkaline electrolytes. In this section, we will briefly introduce the advantages of self-supported electrodes, compare the merits and demerits of diverse substrates, and summarize the common fabrication methods of self-supported electrodes.

Feature of Self-Supported Electrodes
Compared with conventional powdery electrocatalysts, the selfsupported electrodes benefit industrial HER electrolyzers due to the following merits ( Figure 5).
[62] 1) The self-supported architecture can be directly employed as an electrode for the catalytic reaction without any elaborate process and adding binder/ conducting agents, which can simplify the electrode preparation process and effectively decrease the fabrication cost.
2) The in situ growth of catalysts on conductive substrates (e.g., metal, FTO, C material.) will provide a seamless contact between the current collector and catalysts, ensuring rapid charge transfer and avoidance of electrocatalysts peeling or aggregation.
3) The conductive substrates (e.g., 3D scaffold, porous structure.) can offer a large surface area for the growth of the active catalysts, thus providing abundant catalytic sites. Besides, the conductive substrates also can effectively enhance the overall conductivity of the catalyst, thus accelerating the rate of electron transfer to the active site. 4) The self-supported electrode is easier to realize surface hydrophilic/hydrophobic engineering by properly regulating the morphology and microstructure, while the hydrophilic surface can accelerate the bubble detachment. 5) Monolith electrode, as a typical type of the self-supported electrode, is vertically grown on a substrate of the same metal. The metal carriers can directly function as a metal source for the growth of electrocatalysts on self-supported substrates. However, there are still some challenges for the self-supported electrodes. First, there are no criteria for evaluation of the performance of self-supported electrocatalysts compared with the powdery electrocatalysts. Due to the participation during synthesis process, the effect of substrate on electrocatalytic performance cannot be accurately explored, especially for the materials fabricated by multisteps.
In addition, the self-supported electrodes, produced via the same process with Ni foam or carbon cloth, exhibit distinctly different performance for HER. Moreover, although the most common substrates exhibit excellent conductivity, in some cases, the mechanical strength is not good enough to satisfy the long-term operation under large current density at high temperature. [63]

Selection of Substrates for Self-Supported Electrodes
The substrate plays a crucial role both as a current collector and as structural support for self-supported electrodes. The ideal substrates should satisfy the requirements of excellent electrical conductivity, mechanical flexibility, electrochemical stability, reaction activity, and large surface area. Currently, the conductive substrates are most commonly used for self-supported electrodes, which can be divided into three main categories ( Figure 6): 1) metal-based substrates (e.g., metal foil, metal mesh, metal foam); [49a,64] 2) carbon-based substrates (e.g., C cloth, C paper, graphene films); [12b,65] and 3) fluorine-doped tin oxide (FTO) or indium tin oxide (ITO) glass. [66] 3D macroporous metal foams, especially Ni foam, have been widely adopted as substrates for growth of self-supported electrodes due to their high conductivity, low cost, and large surface areas. However, their relatively poor flexibility and elasticity cannot satisfy the long-term HER operation under large current density and high temperature. Besides, the metal foams will become fragile under high electrodes' synthetic temperature (>400-500°C). [67] In this respect, carbon-based substrates may be considered as appropriate substitutes to metal-based substrates due to their high mechanical flexibility and high temperature stability. Nevertheless, the inferior conductivity and higher oxidation sensitivity compared with metal substrates eventually lead to the weaker catalytic activity of the self-supported electrodes. [68] For the FTO or ITO substrates, although the electrocatalysts show lower catalytic efficiency than the same species on other substrates, the flat surface is helpful to reveal the intrinsic catalytic activity of the deposited catalysts by eliminating the substrate disturbance. [69] Generally, more advanced substrates should be developed in the future, which is crucial for both improving the catalytic activity of the self-supported electrodes and enhancing the long-term stability in industrial application.

Fabrication Methods of Self-Supported Electrodes
Generally, ideal self-supported HER electrodes with abundant active sites, high intrinsic activity, excellent conductivity, superaerophobic surface, and high stability are critical for meeting the low-energy consumption and high performance at large current density. [70] The hydro/solvothermal growth, [71] electrodeposition, [72] chemical vapor deposition (CVD), [73] vacuum filtration, [74] and alloying/dealloying [75] are typical synthetic strategies to prepare self-supported TM-based electrodes. The selected preparation method depends on the target substrate material and catalyst component. The hydro/solvothermal method is a versatile approach for fabricating various self-supported TM-based electrodes, which generally requires a relatively moderate temperature and pressure. [76] The morphology and structure of electrocatalysts can be controlled by adjusting reaction time, temperature, solvent, or the feed mole ratio of metal ions. It is noticed that the selfsupported conductive substrates (metal foams, metal foils) also can function as the source of metal elements and provide heterogeneous nucleation sites, which assure the uniform growth of the electrocatalysts, thus resulting in the uniform coverage and strong adhesion between electrocatalysts and substrates. However, severe aggregation may occur in the absence of the substrate due to the fast nucleation and crystal growth. [77] Electrodeposition is another promising strategy to prepare self-supported TM-based electrodes. It is usually operated in a standard two or three-electrode system with conductive substrates as working electrodes. [78] The electrochemical deposition can controllably produce a large proportion of TM-based compounds since the electric field can drive almost all chemical reactions. The loading microstructure and amount of the  electrocatalysts can be tuned by deposition time and methodologies. However, it is difficult to obtain the electrocatalysts with desired size or morphology via electrodeposition approach. CVD methods also have been reported for preparing self-supported TM-based electrocatalysts on various substrates. [79] Under a high heating temperature, the as-prepared substrate supported metal precursors can react with the corresponding gas phase of nonmetal precursors such as sulfur, selenium, phosphorus (PH 3 or P), carbon (CH 4 ), and nitrogen (NH 3 or N 2 ) to form the corresponding sulfides, selenides, phosphides, carbides, and nitrides. The high purity of the compounds can be produced by the CVD technique. On the other hand, the CVD method usually involves poisonous chemical sources or produce poisonous gas, which make the requirement of the reactor be more stringent.
Alloying/dealloying is a commonly method to prepare porous self-supported monolithic electrodes without any readymade substrates, which have been extended to fabricating TM phosphides and alloys. [80] When the more reactive element is selectively dissolved or etched from the alloy, the remaining elements will form a 3D porous network. Self-supported electrodes also can be fabricated by vacuum filtration process, which is a simple and highly reproducible method without complex equipment. [81] Vacuum filtration method is beneficial for integrating various inorganic and organic materials to construct homogeneous structure, and the surface/interface microstructures can be well tuned. Typically, carbon nanotubes (CNTs) and reduced graphene oxide have been used as catalyst supports due to their high conductivity, large surface area, and low density. In Duan's study, [74] a flexible 3D monolithic electrode is fabricated by integrating reduced graphene oxide and electrocatalysts via simple vacuum filtration.

Self-Supported TM-Based Electrocatalysts for Large Current Density HER
Currently, TM elements of Fe, Co, Ni, Ti, Cu, V, Mo, and W are the most widely investigated to construct self-supported TM-based electrocatalysts due to their low cost and unique physicochemical properties. In this section, the current status of self-supported TM-based alkaline HER electrocatalysts (TM sulfides, TM selenides, TM hydroxides, TM phosphides, TM carbides, TM nitrides, TM alloys, TM oxides) operated stably at large current density was summarized.

Transition Metal Sulfides (TMSs)
Non-noble TMSs have acquired extensive interest as HER electrocatalysts, which are especially suitable for large current density operation. Molybdenum disulfide (MoS 2 ) is the most extensively developed due to its unique structural and electronic properties. [82] Previous research has verified that the electrocatalytic activity of the bulk of MoS 2 is poor, while the active sites are highly dependent on the exposure of the edge S sites with free hydrogen adsorption energy close to zero (ΔG H* ). [83] To strategically enhance the HER electrocatalytic activity of MoS 2 , morphology engineering, composition engineering, and foreign   [85] Heteroatom doping not only can effectively activate the basal plane S atoms and introduce in-plane active sites for the HER but also unavoidably lead to overactivation of the edges, resulting in overly strong hydrogen adsorption at the edge sites, which is detrimental to the HER activity. Thus, selectively activating the inert basal plane combined with stabilizing the edges without quenching the activity can maximally increase active sites of MoS 2 for the HER, but it is highly challenging due to the difficulty in balancing the activity and stability. Another efficient strategy is creating vacancies on the crystal surface of the sulfide to create more active sites at the edges. [86] In Zheng's work, [84] they reported a strategy of coconfining Se in the surface and Co in the inner layer of the MoS 2 lattice, which simultaneously activates the basal plane and stabilizes the edges for optimization of the hydrogen adsorption activity. Se-doped MoS 2 nanofoam (Se-MoS 2 -NF) possesses abundant spherical cavities, which will favor the mass transportation of reactants to access more active sites of the catalyst (Figure 7a). DFT calculation showed that the Se-saturated edge is more stable than the S-saturated edge by 0.16 eV per Se replacing an S atom. Compared with pure Se doping, Co/Se codoping can further Figure 7. a) Structural, electronic properties, and HER performance of the Co-/Se-doped MoS 2 nanofoam. Reproduced with permission. [84] Copyright 2020, Nature Publishing Group. b) Schematic illustration of the preparation process of MoS 2 /Ni 3 S 2 NWs on Ni foam. Reproduced with permission. [64a] Copyright 2020, Science China Press. c) Theoretical investigations and catalyst design for 1T-MoS 2 /Ni(OH) 2 nanohybrid. Reproduced with permission. [37] Copyright 2020, Wiley-VCH GmbH. d) High-throughput production of molybdenum disulfide (MoS 2 )-based ink-type electrocatalysts. Reproduced with permission. [49a] Copyright 2020, Nature Publishing Group. decrease the Mo-Mo and Mo-S coordination, as shown in the Mo K-edge extented X-ray absorption fine structure spectra in Figure 7a. DFT calculations demonstrated a synergy effect between inner-layer Co doping's activating effect and surface Se doping's stabilization effect in the Co-/Se-codoped MoS 2 , significantly improving HER performance. The Co/Se-MoS 2 -NF presented the highest HER activity compared with the previously reported heteroatom-doped MoS 2 catalysts under industrial-level large current density. At a high current density of 1000 mA cm À2 , the Co/SeMoS 2 -NF exhibited a much lower overpotential of 382 mV and long-term stability of more than 360 h without decay.
In addition, constructing heterostructures through the synergistic effect of the multicomponents is also considered as an effective route to improve the catalytic activity of TMSs. [64a,87] Cheng and co-workers developed a self-supported hollowlinear-heterojunction-structured molybdenum-nickel sulfide nanowire (NW) (MoS 2 /Ni 3 S 2 /NF), as shown in Figure 7b. [64a] Herein, the nickel foam was directly used as the nickel source for synthesizing MoS 2 /Ni 3 S 2 /NF, which can effectively enhance the interaction between molybdenum-nickel sulfide and the substrate. The electrochemical results showed that the prepared MoS 2 /Ni 3 S 2 /NF only requires low overpotentials of 182 and 200 mV to reach 500 and 1000 mA cm À2 , respectively. The 12 h stability test at current density of 500 and 1000 mA cm À2 presents great potential for large-scale hydrogen production. Chen et al. [37] demonstrated that 1 T-MoS 2 nanosheet edges (instead of basal planes) decorated by metal hydroxides form highly active edge 1TÀMoS 2 =edge NiðOHÞ 2 heterostructures, which significantly enhance HER performance in alkaline media. Theoretical calculations and experiments demonstrate that edge 1TÀMoS 2 =edge NiðOHÞ 2 sites can enhance the alkaline HER kinetics and are the main contributors to the improved catalytic activities of 1T-MoS 2 /Ni(OH) 2 . The edge-rich 1T-MoS 2 /Ni(OH) 2 exhibits unusual alkaline HER activity and durability (Figure 7c).
Liu's group recently demonstrated that MoS 2 has a promising HER performance at the current density of 1000 mA cm À2 by combining surface chemistry and morphology engineering (Figure 7d). [49a] The MoS 2 catalysts modified by Mo 2 C nanoparticles on their edges and surfaces can provide a high current density of 1000 mA cm À2 at 412 mV. They demonstrated the feasibility of the high-throughput production method using a cheap molybdenite concentrate from a naturally existing Earth-abundant mineral and found that the mineral catalysts also show excellent HER activity at high current densities. The production rate of the electrocatalyst is as high as 1.3 g h À1 , 1-2 orders of magnitude higher than previous results, and the catalyst price is %10 US$ m À2 , around 30 times lower than a commercial Pt/C electrocatalyst.
A metallic Mo 2 S 3 @NiMo 3 S 4 heterostructure was proposed in Huang's study [7] to satisfy the industrial demands for HER due to its superior electrical conductivity, sufficient reactive active sites, and excellent structural stability in alkaline medium at large current density. To construct the metallic Mo 2 S 3 @NiMo 3 S 4 heterostructure, the semiconductor MoS 2 was first converted to metallic Mo 2 S 3 by introducing excessive Mo with the appearance of Mo─Mo bonds. The metallic Mo 2 S 3 is utilized as the metallic support for in situ growth of NiMo 3 S 4 via a facile hydrothermal method (Figure 8a). The modulation of metallic Mo 2 S 3 and in situ epitaxial growth of bifunctional Ni-based catalyst can facilitate the charge transfer for fast Volmer H and Heyrovsky H 2 generation. The Mo 2 S 3 @NiMo 3 S 4 electrolyzer requires an ultralow voltage of 1.672 V at a large current density of 1000 mA cm À2 , with %100% retention over 100 h. This work opened up new opportunities for developing efficient and stable electrocatalysts to meet the industrial demand in the future.
In practice, stability is another key issue for hydrogen production electrodes. A monolith catalyst (MC) was developed to address this obstacle. Specifically, a metallic transition metal dichalcogenide (m-TMDC) was vertically grown on a substrate of the same metal using an OSPS method. Due to the nature of the monolith, charges can be directly transferred from the substrate to the catalyst without crossing van der Waals interfaces, providing highly efficient charge injection and excellent HER performance. In Liu's study, [88] a tantalum-tantalum sulfide (Ta-TaS 2 ) MC with a large area has been synthesized by the OSPS method and shown superior hydrogen evolution activity. The electrode achieved 2000 mA cm À2 with a small overpotential of 398 mV and continued working for >200 h under large current density without noticeable performance decay (Figure 8b). Similar results have been achieved in other MCs, such as Nb-NbS 2 and Mo-MoS 2 . In our group, the Mo-based, W-based MC was developed in alkaline media to satisfy the industrial HER demands, which will be discussed in the following parts.

Transition Metal Selenides (TMSes)
Similar to TMS, TMSes possess a layered structure, adjustable electrical properties, and exposed active edge sites being the active HER centers. [89] However, TMSes have a higher intrinsic catalytic activity than the corresponding sulfides due to their higher electrical conductivity. [90] Accordingly, there is an increased interest in developing TMSes-based HER electrocatalysts. Therefore, strategies for improving the HER activities of TMSs, such as constructing heterojunction structures, metallic doping, exposing active sites, and coupling with conductive materials, were also suitable for TMSes. [91] In Park's work, [64b] multidirectional charge transfer concept has been adopted within heterostructured catalysts to develop an efficient and robust bifunctional water electrolysis catalyst, which comprises perovskite oxides (La 0.5 Sr 0.5 CoO 3-δ , LSC) and potassium-ion-bonded MoSe 2 (K-MoSe 2 ). The LSC/K-MoSe 2 system involving twoway charge transfer from K to MoSe 2 and from LSC to MoSe 2 ( Figure 9a) led to significantly improved water electrolysis performance and operational stability. The water electrolysis performance using the LSC/K-MoSe 2 ||LSC/K-MoSe 2 couple outperformed the state-of-the-art noble metal pair of Pt/C||IrO 2 , exhibiting lower cell voltage for the overpotential at 10 and 100 mA cm À2 , improved energy efficiency, and excellent operational stability over 2500 h. In addition, the heterostructure demonstrates overwhelmingly superior durability and activity for water electrolysis at high current density. In 1.0 M KOH at 60°C, the LSC/K-MoSe 2 couple required cell voltages of 2.25 V at 500 mA cm À2 and 2.52 V at 1000 mA cm À2 , respectively. Under such conditions, it exhibited stable operation over 1200 and 800 h at 500 and 1000 mA cm À2 , respectively. In 10 M KOH at room temperature, the cell voltage needed to achieve Copyright 2021, Nature Publishing Group. b) Schematic illustration of the fabrication process, band structure and PDOS, and HER performance for selfstanding 1T 0.63 -MoSe 2 @MoP MPIC. Reproduced with permission. [92] Copyright 2022, American Chemical Society. Figure 8. a) Design and characterizations for Mo 2 S 3 @NiMo 3 S 4 heterostructure: Schematic illustration of the epitaxial construction of Mo 2 S 3 @NiMo 3 S 4 , energy band diagrams of Mo 2 S 3 and NiMo 3 S 4 , TEM image, and HER polarization curves of Mo 2 S 3 @NiMo 3 S 4 heterostructure. Reproduced with permission. [7] Copyright 2022, Wiley-VCH GmbH. b) Synthesis and characterization of the Ta-TaS 2 MC: the OSPS synthesis process of Ta-TaS 2 MC and corresponding SEM images and HER polarization curves. Reproduced with permission. [88] Copyright 2021, Nature Publishing Group. 100 mA cm À2 was 1.87 V for the LSC/K-MoSe 2 couple. In the two-electrode cell, the chronopotentiometric stability of 1600 h was achieved at 100 mA cm À2 in 10 M KOH at room temperature without any noticeable performance degradation. The optimized LSC/K-MoSe 2 catalyst exhibits significantly enhanced HER performance which can be attributed to the increased electrical conductivity of MoSe 2 . The semiconducting 2H-phase MoSe 2 was moderately converted to metallic 1T-MoSe 2 via charge transfer from potassium atoms during the potassium metal intercalation process. 1T 0.63 -MoSe 2 @MoP multiphase interface heterostructure catalyst (MPIC) was designed to promote the alkaline HER by tuning the intrinsic interfacial electronic structure. [92] The DFT calculation demonstrated that the band structure of 1T-MoSe2@MoP possesses zero bandgap (Figure 9b), providing a highly electrically conductive interface. Notably, the higher electron density of 1T phase-MoSe 2 @MoP at the Fermi level revealed faster charge transfer kinetics. Meanwhile, the increasing number of electronic interface states shown in partial density of state plots resulted in more charge carriers at the interface, thereby improving the HER performance. The atomic models with charge density difference plots of 2H phase-MoSe 2 @MoP and 1T phase-MoSe 2 @MoP revealed that the electrons are easily transferred from 1T 0.63 -MoSe 2 to the MoP interface, which significantly increases the electron density at the interface. Besides, the introduction of 1T-phase species is crucial for forming the unique and highly active interface in the 1T 0.63 -MoSe 2 @MoP MPIC. The self-standing 1T 0.63 -MoSe 2 @MoP MPIC required a small overpotential of 358 mV to reach a large current density of 1000 mA cm À2 in alkaline freshwater, along with impressive HER activity and stability at large current density in artificial alkaline seawater electrolyte. More importantly, the self-standing MPIC was vertically grown on an industrial-grade Mo mesh using a facile solid-state synthesis strategy without any sensitive conditions. This method is feasible for sizable self-standing MPIC, confined only by the growth chamber's dimensions and the Mo mesh's size. This work inspired the development of efficient HER catalysts for seawater cracking and large-scale H 2 generation.

Transition Metal Hydroxides (TMHs)
Generally, TMHs are composed of metal cations (e.g., Co, Fe, and Ni) in the trivalent or bivalent state and hydroxide anions. [93] Especially TM layered double hydroxides (LDHs) of the metal cations in the brucite-like layers and the intercalated anions have attracted considerable attention due to their Earth-abundant reserves, unique 2D layered structure, and densely distributed active sites. [94] It should be known that pure TMHs present relatively low HER activity due to their weak hydrogen adsorption energy. However, the strong OH À adsorption ability facilitates hydrolysis, which is usually the RDS for HER in alkaline water electrolysis. [53,95] To date, various strategies have been proposed to improve the HER activity of TM LDHs by regulating morphology, defect formation, and electronic structure engineering. [96] Luo et al. reported a Ni-based catalyst composed of hydroxidemediated Ni 4 Mo nanoparticles decorated with FeO x and anchored onto MoO 2 nanosheets (h-NiMoFe), which delivered an impressively good performance with a relatively low overpotential of 97 mV at the current density of 1000 mA cm À2 . [97] For practical applications, the electrochemical performance of the h-NiMoFe catalyst was operated in overall water splitting systems, where the h-NiMoFe catalyst is used as both the HER and OER catalysts. The catalyst presented a high performance, achieved a current density of 500 mA cm À2 at a record-low cell voltage of 1.56 V, and remained stable for over 40 h (Figure 10a). Impressively, the h-NiMoFe catalyst could be prepared on a meter scale, which has the prospect of industrial application.
Coupling TM LDHs with other conductive nanocomponents to construct an interface-engineered hybrid structure has been employed to enhance the activity of electrocatalysts. In Wang's study, [98] they designed bimetallic alloy/oxyhydroxide core-shell heterostructure electrocatalysts by gas-templated electrodeposition method. For NiCo/NiCo-OH and NiFe/NiFe-OH electrocatalysts (Figure 10b), the metallic core promoted the charge transport to the surface oxyhydroxide, whereas the hierarchical structure enabled fast ion diffusion and provided abundant active sites. The as-obtained NiCo/NiCo-OH and NiFe/NiFe-OH showed excellent activity toward the HER in 1.0 M KOH, especially at high current densities. More importantly, the alkaline electrolyzer assembled using the NiCo/NiCo-OH||NiFe/NiFe-OH couple only needed a small cell voltage of 1.74 V to achieve a high overall water-splitting current of 500 mA cm À2 . Further, the electrolyzer exhibited excellent durability that can stably catalyze water splitting at 1000 mA cm À2 for at least 300 h, much superior to the state-of-the-art 20% Pt/C||RuO 2 combination. Similarly, NiFe oxyhydroxide-anchored NiFe alloy NW arrays, [99] core/shell NiFe LDHs-decorated Cu NW, [100] hierarchical NiFe LDHs-anchored porous nickel particles, [101] and Cu 2 O-decorated Cu dendrites [102] also exhibited improved catalytic activities.
Moreover, Wu's group first reported that B doping-induced amorphization of LDH could effectively modify the NiCo-LDH, which achieves the goal of improving the HER catalytic activity (Figure 10c). [96b] With the doping by B atoms, the A-NiCo LDH/NF catalyst is in a disordered state at local areas, and the presence of oxygen defects makes more unsaturated metal atoms exposed on the surface, thus generating more active sites. The A-NiCo LDH/NF possessed high HER catalytic activity, while it needed low overpotentials of 286 and 381 mV to obtain the current densities of 500 and 1000 mA cm À2 , respectively. Besides, the A-NiCo LDH/NF catalyst could maintain excellent catalytic activity for at least 72 h in 1.0 M KOH. Similarly, Shen et al. used metallic Ru to modify (Fe, Ni)(OH) 2 nanosheet arrays supported on a conductive substrate of Ni foam (named Ru/(Fe, Ni)(OH) 2 /NF). [103] The superhydrophilic surface and high conductivity ensured the rapid release of gases and efficient electron transportation and mass transfer at a high current density. This electrode needed an overpotential of only 152 mV to obtain a high 1000 mA cm À2 current density for HER, along with excellent stability over 20 h under a high current density of 500 mA cm À2 in the basic electrolyte (Figure 10d).

Transition Metal Phosphides (TMPs)
TMPs usually exhibit superior catalytic activity in alkaline electrolytes compared with other TM-based electrocatalysts ( Table 2) due to the following aspects: 1) TMPs are found active in the bulk www.advancedsciencenews.com www.advenergysustres.com form, which is different from the TMSs with exposed edge active sites and [104] 2) the analogous active sites in hydrogenases with strong ability to attract protons render the TMPs to be potential electrocatalysts for the HER. Rodriguez's group performed DFT calculation to demonstrate that Ni 2 P (001) possesses the favorable H binding in hydrogenase systems and the surface P atom makes the Ni 2 P (001) behave somewhat like the hydrogenase. [105] P on the surface exhibits small negative charge and can trap protons as the bases in the hydrogenase. Consequently, the surface P can act both as proton-acceptor and hydride-acceptor centers to facilitate the HER. [106] 3) Compared with sulfur, the P atom has larger radius, and its metallic compounds are more inclined to form the isotropic structures and have more harmoniously unsaturated surface atoms. [107] To make TMPs more suitable for practical large current applications, the self-supported nanostructures in situ grown on 3D substrates have been widely exploited. [108] In addition, fabricating the electrode structure with 3D hierarchical architectures can contribute to efficiently dissipating the gaseous product and rapidly facilitating the exchange of reactants and products. Yu et al. reported a unique "superaerophobic" Ni 2 P nanoarray structure formed on a Ni foam substrate (denoted as Ni 2 P/NF), where Ni 2 P nanosheets are evenly distributed on Ni 2 P NW arrays via a facile hydrothermal-then-annealing strategy (Figure 11a). [109] The Ni 2 P/NF catalyst is highly active and stable for HER, even exceeding commercial noble metal Pt/C at large current density (>170 mA cm À2 ). The unique assembly of the Ni 2 P/NF provided a preferred "superaerophobic" structure to release the in situ-generated gas bubbles. A two-electrode lab-scale electrolyzer composed of Ni-Fe LDH anode and Ni 2 P/NF cathode also demonstrated an excellent catalytic capability for water splitting, which surpassed the performance of the Pt/C||Ir/C cell at a higher current density (>600 mA cm À2 ). Moreover, investigations found that incorporating a metal element into binary TMPs could alter the electronic structure and  [97] Copyright 2021, The Royal Society of Chemistry. b) Preparation diagram, structural characterization, and stability of the preparation of the alloy/oxyhydroxide core-shell electrocatalysts. Reproduced with permission. [98] Copyright 2020, Elsevier. c) The scheme of the preparation process, morphology characterization, and HER performance of A-NiCo LDH nanosheet arrays supported on the nickel foam. Reproduced with permission. [96b] Copyright 2020, Elsevier. d) Schematic representation of the formation of Ru/(Ni,Fe)(OH) 2 /NF and the corresponding HER performance. Reproduced with permission. [103] Copyright 2020, Wiley-VCH GmbH.  consequently optimize the adsorption-free energy of reaction intermediates, thus promoting their intrinsic catalytic activity for water splitting. [110] For example, a novel 3D hierarchical bimetallic Ni 2(1Àx) Mo 2x P NW array on Ni foams was fabricated for efficient HER under large current density in the alkaline electrolyte (Figure 11b). [111] The Ni foam worked as an efficient current collector, and the highly conductive Ni 2(1Àx) Mo 2x P NWs provided a continuous electron transfer pathway. Additionally, the highly porous NWs with a rough surface offered a large active surface area with numerous active sites and rapid release of H 2 . DFT calculations revealed that Mo substitution of Ni in Ni 2 P leads to optimal free energy of water activation and hydrogen adsorption on the catalyst surface. The Ni 2(1Àx) Mo 2x P catalyst exhibited an excellent HER activity with low overpotentials of 240 and 294 mV to acquire current densities of 500 and 1000 mA cm À2 , respectively, along with superior stability in 1.0 M KOH. This highly active and stable catalyst enabled an electrolyzer operating at 10 mA cm À2 at a voltage of 1.51 V, 100 mA cm À2 at 1.65 V, and 500 mA cm À2 at 1.82 V in 1.0 M KOH at room temperature, which are much better than the benchmark of IrO 2 /Pt. Additionally, nonmetallic element (N, O, F, B, and P) doping is an important way to improve the property of catalysts. [112] Chai et al. reported a self-supported bimetallic cobalt-iron phosphide by doping with highly electronegative fluorine atoms on iron foam (F-Co 2 P/Fe 2 P/IF). [113] The surface heteroatom (F) doping led to the strong synergistic interactions between Co 2 P and Fe 2 P, and the prepared F-Co 2 P/Fe 2 P/ IF electrode shows the ultrahigh-current test range (larger than 3000 mA cm À2 ) toward HER. The electrochemical measurements under simulated industrial conditions also showed that the F-Co 2 P/Fe 2 P/IF catalyst could be stable for a long time in 6.0 M KOH (Figure 11c). Direct phosphorization of the available metal substrate was also developed by sintering or solvothermal conditions, which could avoid the complicated precursor synthesis. [9c,12b,114] However, it is usually limited by selecting an appropriate metal substrate, particularly in synthesizing multimetal phosphides. For example, the molybdenum dioxide (MoO 2 )-molybdenum phosphide (MoP) macroporous framework catalyst grown on an industrial-grade molybdenum (Mo) plate was reported by Liu's group. [73a] Compared with the common conducting substrates such as CFP, CC, and graphite plate, the monolithic Mo plate can directly function as a molybdenum source and self-supported substrate. The direct growth of the active MoO 2 -MoP seamless electrode (SE) can avoid the bonding process and enhance the physical contact between catalysts and the conductive substrates, thereby boosting the charge transport efficiency and reducing the delamination of active species from the current collector at high catalytic current densities. Besides, the macroporous framework structure is beneficial to enhance mass transport properties, expose more active sites for the HER, and facilitate the hydrogen bubble release from the electrode surface due to its weak gas-solid interface adhesion, which are critical for the HER at high current densities. Besides, to make this catalytic system much closer to the industrial hydrogen production process, Mo mesh was also used as the substrate for in situ growth of the MoO 2 -MoP electrocatalyst ( Figure 12). The MoO 2 -MoP mesh electrode reaches the current density of 1000 mA cm À2 at 293 mV in 1.0 M KOH electrolyte, while it only requires an overpotential of 215 mV to achieve the same current density in 5.0 M KOH solution, which served as the electrolyte in industrial water-alkali electrolyzers. These results demonstrated that the HER performance of MoO 2 -MoP mesh meets the activity requirement in practical AEL technology. Meanwhile, the MoO 2 -MoP mesh is also durable, with slight decay after 100 h operations in 1.0 M KOH solution or after 50 h operations in 5.0 M KOH solution at 1000 mA cm À2 . These results indicated the huge potential of using the industrial-grade Mo plate or Mo mesh as self-supported substrates and molybdenum sources for fabricating low-cost, high-performance, and durable Mo-based electrocatalysts that are suitable for large-scale commercial hydrogen production.

Transition Metal Carbides (TMCs)
TMCs have emerged as an effective HER catalyst due to the Pt-like electronic behavior, high conductivity, and considerable stability in a wide pH range. [8a,12a] Molybdenum carbide (MoC, Mo 2 C), a representative TMC material, has attracted attention. [114] The electronic structure of the surface Mo atoms in Mo 2 C has a similar d-band center with Pt surface, which provides suitable bond strength and exhibits excellent HER activity. [115] However, their HER catalytic activities have not reached the expected performances. Element doping or material hybridization is a common way to improve the catalytic activity of MoC and Mo 2 C by tailoring their band structure. [116] Besides, the self-supported architecture of TMCs can effectively overcome the sintering and agglomeration phenomenon caused by the high carbonization temperature (above 700°C). [117] Wang et al. reported a self-standing MoC-Mo 2 C heterojunction catalytic layer on a Mo sheet in molten carbonate via a one-step electrocarbide approach using CO 2 as the feedstock (Figure 13a). [118] CO 2 was electrochemically reduced to C which simultaneously reacted with the Mo substrate to form a porous MoC-Mo 2 C heterojunction catalytic layer. This molten carbonate electrolysis used an active Mo substrate and was involved in forming the carbide layer, thereby ensuring the firm connection between the Mo substrate and the electrolytic MoC-Mo 2 C heterojunction layer. The outstanding HER performances of the MoC-Mo 2 C electrode presented low overpotentials at 500 mA cm À2 in both acidic (256 mV) and alkaline electrolytes (292 mV), longlasting lifetime of over 2400 h (100d), and Figure 11. a) Synthetic process, morphology characterization, adhesion behavior for a H 2 bubble, and HER activity for Ni 2 P nanoarrays. Reproduced with permission. [109] Copyright 2019, American Chemical Society. b) Fabrication procedure and HER activity for the self-supported 3D Ni 2(1-x) Mo 2x P electrode. Reproduced with permission. [111] Copyright 2018, Elsevier. c) Schematic illustration of the fabrication of F-Co 2 P/Fe 2 P/IF and the HER performance. Reproduced with permission. [113] Copyright 2020, Elsevier.
www.advancedsciencenews.com www.advenergysustres.com high-temperature performance (70°C). DFT calculations revealed that the superior HER performances of the carbide heterojunction electrode in both acid and alkaline conditions are attributed to the MoC(001)-Mo 2 C(101) heterojunction that has ΔG H* of À0.13 eV in acidic conditions and the energy barrier 1.15 eV for water dissociation in alkaline solution. In addition, the 3D porous structure is hydrophilic and can suppress the shielding effect of generated H 2 bubbles. The molten carbonate electrolysis is proven to be a practical surface engineering approach to tailor the hydrophilic of the surface, the composition of the catalytic layer, as well as the surface microtopography. Creating chemical bonding between the electrocatalysts and electrodes is an effective way to improve the mechanical stability of HER catalysts. Hence, fast self-heating of a conductive matrix could be used to in situ synthesize chemically bonded catalysts on the matrix and avoid the decay of their activity caused by agglomeration, which is superior to traditional methods. [119] Liu developed a low-energy-consumption method using a CNT film as a heat source and matrix, which rapidly changes its temperature in hundreds of milliseconds to in situ synthesize a robust Mo 2 C/MoC/CNT composite (Figure 13b). [115a] The as-prepared uniformly dispersed Mo 2 C/MoC heterogeneous nanoparticles can form strong chemical bonds with the CNT film. Consequently, massive Mo 2 C/MoC interfaces offered abundant active sites for HER, resulting in the Mo 2 C/MoC/CNT film with a low overpotential of 255 mV at a high current density of 1500 mA cm À2 in 1.0 M KOH. The strong chemical bonds between Mo 2 C/MoC and CNTs significantly weaken the dissolution and shedding of the Mo 2 C/MoC nanoparticles during the HER at high current densities. Hence, the overpotential of the Mo 2 C/MoC/CNT film changed by only %32 and %47 mV after working at 500 and 1000 mA cm À2 for 14 days, respectively. DFT calculations demonstrated the moderate free energy (ΔG H* ) of 0.02 eV for hydrogen adsorption at sites around Mo 2 C/MoC interfaces and strong coupling between the Mo x C and CNT matrix, which ensured the high activity and stability of the heterogeneous Mo 2 C/MoC/CNT film.
In our previous work, MoSe 2 -Mo 2 C heterostructure electrocatalysts have been constructed to improve the HER activity. [120] It was observed that the synergistic effects between MoSe 2 and Mo 2 C tune the hydrogen adsorption and desorption behavior, thereby promoting outstanding hydrogen evolution catalytic performance. DFT calculations demonstrated that the energy barrier for bond breaking of H─OH in water molecules was %0.6 eV for Mo 2 C in alkaline media. The barrier height may have a side effect on the Volmer step to some degree (Figure 14a). The heterostructure of Mo 2 C/MoSe 2 catalyst could facilitate the water dissociation process, in which MoSe 2 adsorbed OH and Mo 2 C served as the active centers for H 2 production. However, the hybrid Mo 2 C/MoSe 2 catalyst remained a big challenge since the barrier height for the cleavage of H─OH bond in H 2 O molecule on MoSe 2 edge has been calculated to be very high. Thus, it is highly desirable to endow the catalyst with the ability to split water molecules further to improve the activity of Mo 2 C catalyst in alkaline media. In our further work, a 3D macroporous framework MoS 2 -Mo 2 C heterojunction was fabricated via a facile solid-state synthesis strategy (Figure 14b). [73b] The 3D macroporous framework structure not only exposed more active sites for www.advancedsciencenews.com www.advenergysustres.com the HER, but also facilitated the release of gas evolution. Consequently, the MoS 2 -Mo 2 C/Mo electrode exhibited excellent stability and a low overpotential of 446 mV in an alkaline medium at a high current density of 1000 mA cm À2 , which enabled a superior HER activity than commercial Pt/C. Besides, a novel hybrid electrocatalyst with molybdenum carbide nanosheets and nickel hydroxide nanoparticles (Mo/Mo 2 C-Ni(OH) 2 ) was synthesized in our group to study the HER performance in alkaline media (Figure 14c). [121] The Mo 2 C nanosheets with textured surfaces not only expose large special surface area and plentiful active catalytic sites but also avoid the agglomeration of sediments for Ni(OH) 2 nanoparticles. Also, Ni(OH) 2 serves as an excellent water dissociation site, and the synergistic effect of Ni(OH) 2 and Mo 2 C can further enhance the catalytic activity.
As another promising TM-based HER catalyst, tungsten carbides (WC and W 2 C) also have provoked an increased research passion due to their similar properties with molybdenum carbides. [73c] Although W x C exhibited unusual activity in acidic media, they suffered from poor corrosion stability under high pH conditions. In our recent study, we constructed WP-W 2 C heterojunction with a nanoporous structure to complement the advantages of W 2 C and WP catalysts and synergistically enhance the HER activity and corrosion stability (Figure 14d). The results suggested that the introduction of the P precursor not only decreased the synthesis temperature of W 2 C but also enlarged the pores of the nanoporous network appropriately. Surprisingly, the self-supported nanoporous feature with a small water contact angle provided this SE with a mechanically robust Figure 13. a) Schematic of synthetic process, contact angle images, calculated ΔG H* diagram, and HER performance for MoC-Mo 2 C electrodes. Reproduced with permission. [118] Copyright 2021, Nature Publishing Group. b) Fabrication procedure, DFT calculations, and HER performance for Mo 2 C/MoC/CNT electrode. Reproduced with permission. [115a] Copyright 2022, Nature Publishing Group.
www.advancedsciencenews.com www.advenergysustres.com and more convenient gas transport channel, facilitating the electron transfer and the formation and release of H 2 bubbles. The WP-W 2 C nanoporous network presented excellent catalytic stability and activities with low overpotentials of 560 mV and 449 mV at a current density of 1000 mA cm À2 in 1.0 M KOH at 25 and 70°C, respectively. Besides, the electrode also displayed outstanding HER performance in 5.0 M KOH, which satisfied the demand of the industrial application.

Transition Metal Nitrides (TMNs)
Currently, TMNs with high intrinsic activity, small charge transfer resistance, and high stability have been studied as promising HER catalysts. [122] The N atoms introduced can change the d-band electron state density of the corresponding metal and reduce the d-band occupation shortages. [123] Compared to TMCs, the TMNs usually showed inferior HER catalytic properties. [124] Therefore, TMNs-based HER catalysts have been exploited by cation doping via redistributing the charge to activate the catalytic sites, thus improving the HER activity. [125,126] For example, Jiang's group introduced N into the MoNi alloy to prepare a Ni 0.2 Mo 0.8 N electrocatalyst supported on NF (Ni 0.2 Mo 0.8 N/NF) shown in Figure 15a, which showed excellent HER electrocatalytic performance with an overpotential of 85 mV@500 mA cm À2 in 1.0 M KOH medium. [127a] DFT calculations demonstrated that the Ni and Ni 0.2 Mo 0.8 N surfaces have favorable hydroxyl and hydrogen species adsorption energetics, respectively, which could cooperate synergistically toward alkaline hydrogen evolution. In our previous work, we reported a large-scale approach for producing low-cost and highly efficient molybdenum selenidenitride (MoSe 2 -Mo 2 N) Schottky heterojunction catalysts that work well at a high current density of up to 1000 mA cm À2 . [127] DFT calculations showed that the synergistic integration of the MoSe 2 -Mo 2 N Schottky heterojunction induces self-driven electron transfer that optimizes the electronic structure and also tunes the hydrogen adsorption and dissociation behavior (Figure 15b). The MoSe 2 -Mo 2 N/Mo electrode delivered a high current density of 1000 mA cm À2 at an overpotential of 462 mV for hydrogen evolution in alkaline media, which is superior to Figure 14. a) A schematic illustration of the preparation process and the electron differential density (EDD) contour map for Mo 2 C/MoSe 2 hybrid electrode. Reproduced with permission. [120] Copyright 2020, The Royal Society of Chemistry. b) Schematic of synthetic process, morphology characterization, and HER performance of MoS 2 -Mo 2 C heterojunction. Reproduced with permission. [73b] Copyright 2021, The Royal Society of Chemistry. c) Morphology characterization and DFT calculation of Mo 2 C-Ni(OH) 2 electrode. Reproduced with permission. [121] Copyright 2022, Elsevier. d) Digital image, contact angle behavior, and HER performance of self-supported nanoporous WP-W 2 C electrode. Reproduced with permission. [ commercial Pt/C electrodes. Besides, to make the electrochemical process much closer to the industrial hydrogen production process, the MoSe 2 -Mo 2 N/Mo mesh was used as the working electrode for HER in the homemade H 2 production system. The MoSe 2 -Mo 2 N/Mo mesh electrode afforded a current density of 1000 mA cm À2 at 386 mV in 1.0 M KOH solution, while it only needed an overpotential of only 275 mV to achieve the same current density in 5.0 M KOH solution, which is suitable for practical alkaline hydrogen production. At high current densities, the electrocatalysts in situ grown on Mo mesh presented lower overpotentials than the industrial-grade Ni mesh. Similarly, Yu et al.
proposed a multiscale modulation strategy to combine Co 3 N and N-doped WS 2 particles into a hybrid catalyst (N-WS 2 /Co 3 N), which served as water dissociation promoter and hydrogen adsorption sites, respectively (Figure 15c). [128] The metallic feature and high porosity of Co 3 N support endowed the N-WS 2 particles with good conductivity, high-density active N-WS 2 edges oriented at the surface, and efficient electron transfer between N-WS 2 and Co 3 N supports. Thus, the resultant N-WS 2 / Co 3 N exhibited remarkable HER activity in 1.0 M KOH electrolyte, requiring a small overpotential of 235 mV at 600 mA cm À2 with outstanding long-term durability at 500 mA cm À2 .

Transition Metal Alloys (TMAs)
TMAs have aroused great interest as the non-noble metal HER electrocatalysts in alkaline conditions, owing to their optimized ΔG H* and high intrinsic catalytic activity. [12a,129] Forming alloys can adjust the surface properties of the catalyst by altering the average energy of the surface d-band, and thus regulate the adsorption energy of hydrogen or oxygen intermediates, increase the active sites, and finally improve the HER performance. [130] TMAs have been extensively investigated as the most active non-noble metal HER catalysts under alkaline conditions. However, the application of TMAs for water electrolysis at high current density has rarely Reproduced with permission. [127] Copyright 2021, The Royal Society of Chemistry. c) Schematic of synthetic process and DFT calculation of N-WS 2 /Co 3 N electrode. Reproduced with permission. [128] Copyright 2022, Wiley-VCH GmbH.
www.advancedsciencenews.com www.advenergysustres.com been reported. [12c,47,131] Nickel-molybdenum (Ni-Mo) alloys or compounds with superior corrosion resistance have been widely studied in seawater splitting. Ni atoms are broadly recognized as excellent water dissociation centers, while Mo atoms have superior adsorption properties toward hydrogen. Feng et al. [131] first reported a MoNi 4 alloy electrocatalyst embedded in MoO 2 cuboids, which were vertically aligned on nickel foam (MoNi 4 /MoO 2 @Ni). The MoNi 4 /MoO 2 @Ni exhibited a high HER activity with a zero-onset overpotential and a low Tafel slope of 30 mV per decade in a 1.0 M KOH aqueous solution, and the high cathodic current density of 500 mA cm À2 required an overpotential as low as 65 mV in 5.3 wt% KOH electrolyte at ambient temperature ( Figure 16a). Experimental results revealed that the MoNi 4 electrocatalyst presented a highly active center and manifested fast Tafel step-determined HER kinetics. Furthermore, DFT calculations determined that the kinetic energy barrier of the Volmer step for the MoNi 4 electrocatalyst was as low as 0.39 eV. These results confirmed that the sluggish Volmer step was drastically accelerated for the MoNi 4 electrocatalyst. Benefiting from its scalable preparation and stability, the MoNi 4 electrocatalyst is promising for practical water-alkali electrolyzers. After that, Liu's group [132] developed a sequential reduction strategy to synthesize types of electrocatalysts made of PGMs anchored on a high-surface-area and corrosion-resistive matrix for high-performance electrolysis. The prepared Pt/Ni-Mo electrocatalyst showed a small overpotential of 113 mV at an ultrahigh current density of 2000 mA cm À2 in saline-alkaline water, which is the best performance reported so far. The electrocatalyst also showed long durability in various electrolytes under harsh conditions, including a strong alkaline electrolyte or a simulated seawater electrolyte (24 h), an ultrahigh current density of 2000 mA cm À2 (140 h), as well as temperatures up to 80°C (Figure 16b). Yin et al. [133] reported a FeIr bimetallic alloy self-supported on nickel foam prepared by hydrothermal method, Figure 16. a) Schematic of synthetic process, HER performance, and DFT calculation of MoNi 4 /MoO 2 @Ni electrode. Reproduced with permission. [131] Copyright 2017, Nature Publishing Group. b) Schematic of the synthetic process, HER performance, and Pt/Ni-Mo catalyst static contact angles. Reproduced with permission. [132] Copyright 2021, Wiley-VCH GmbH.

Transition Metal Oxides (TMOs)
TMOs, such as spinel-type oxides and perovskites, have been widely investigated as promising electrocatalysts for OER due to their attractive catalytic activity and considerable anticorrosion properties in alkaline electrolytes. However, the poor electrical conductivity and low catalytic activity of TMOs led to the inferior water dissociation and weak H adsorption capacity, which limited their further application as active electrocatalysts for HER. [134] Thus, considerable attention should be paid to improve the intrinsic electrocatalytic activity of TMO electrocatalysts. It is noticed that TMPs or TMSs are promising candidates for HER electrocatalysts, which can be integrated with TMOs to form heterostructures for the improvement of HER activity. Constructing heterostructures can regulate electron transfer and active sites due to the existence of coupling interfaces and the synergistic effect of the heterostructures. [135] For instance, plenty of heterostructures including MoO 2 -FeP, [136] MoO 3 nanodots/MoS 2 , [137] NiS 2 /N-NiMoO 4 , [138] NiFeO x /CFP, [139] Co x P@Co 3 O 4 , [140] and GDY/MoO 3 , [141] have been extensively constructed for the enhanced HER electrocatalytic performance. In Hou's work, [142] the 3D hierarchical TMO/TMS heterostructure arrays interacting with 2D MoO x /MoS 2 nanosheets attached to 1D NiO x /Ni 3 S 2 nanorods were prepared by surface reconfiguration strategy for water splitting (Figure 17a).
The NiMoO x /NiMoS heterostructure electrocatalysts achieved low overpotentials of 174 and 236 mV for HER at 500 and 1000 mA cm À2 and presented long-term stability at a large current density of 500 mA cm À2 . Especially, the assembled twoelectrode cell composed of NiMoO x /NiMoS delivered large current densities of 500 and 1000 mA cm À2 at the low cell voltages of 1.60 and 1.66 V, along with excellent stability. The excellent electrocatalytic performance could be ascribed to not only the modulation of component and geometric structure, but also the optimization of charge transfer, active sites, and synergistic effect of the heterostructure interfaces. DFT calculations revealed that the coupling interface between NiMoO x and NiMoS can optimize adsorption energies and accelerate reaction kinetics, thus enhancing the electrocatalytic performance. Yin et al. synthesized a N-doped-graphene-encapsulated Ni-MoO 2 NWs heterostructure catalyst anchored on nickel foam [(Ni-MoO 2 )@C/NF] via the two-step process for overall water splitting (Figure 17b). [143] The electrocatalyst exhibited a low overpotential 304 mV at current density of 2000 mA cm À2 for HER. Moreover, it can maintains 340 h stability under a multicurrent process condition (10-1500 mA cm À2 ) without obvious attenuation. The promising electrocatalytic performance could be ascribed to the heterostructure comprising Ni and MoO 2 , N-doped-carbon coating structure, and NW architecture anchored on nickel foam.
Moreover, it is found that the rational anion doping into TMOs is an attractive strategy for regulating the charge distribution due to the different atomic radius and electronegativity. For example, P-doped Co 3 O 4 /NF, [144] Co-doped CeO 2 nanosheets, [145] Ru-CoO x /NF, [146] and F,P-Fe 3 O 4 /IF [147] have been reported as efficient electrocatalysts for HER. Fe-based oxides have been seldom reported as electrocatalysts for HER due to their weak intrinsic activity and conductivity. In Chai's study, [148] P doping modulation was used to construct inverse spinel P-Fe 3 O 4 with dual-active sites supported on iron foam (P-Fe 3 O 4 /IF) for excellent alkaline HER (Figure 17c). The P-Fe 3 O 4 /IF only needed the overpotential of %240 mV to acquire the current density of 1000 mA cm À2 for HER in 1.0 M KOH. Moreover, the NiFe LDH/IF||P-Fe 3 O 4 /IF cell achieved an onset potential of 1.47 V for overall water splitting, with excellent stability for more than 1000 h at a current density of 1000 mA cm À2 . The obtained inverse spinel Fe-O-P derived from controllable phosphorization can provide an octahedral Fe site and O atom. The unique structure of octahedral Fe site with excellent conductivity is crucial to facilitate the water dissociation, accompanied by the highly active P site to bind the hydrogen intermediates. Meanwhile, DFT calculations demonstrated that ΔG H of the P atom in Fe-O-P as an active site is theoretically calculated to be 0.01 eV, which is beneficial for alkaline HER.

Conclusion and Perspectives
For large-scale industrial hydrogen production, high current density (alkaline electrolytic cell >500 mA cm À2 ) and durability (>100 h) are crucial. In this review, we elaborated the design principles of electrocatalysts for industrial HER and proposed the key factors for achieving high-performance electrocatalysts at large current density, including high intrinsic activity, fast charge transfer, and mass transfer, accelerated gas bubble mitigation, and high mechanical/structural/compositional stability. As the most critical part of this review, we summarized the representative works for designing self-supported TM-based electrocatalysts with large current density and long-term durability for industrial alkaline HER. Many efforts have been developed in self-supported transition-metal-based electrocatalytic materials that range from sulfides, selenides, hydroxides, phosphides, carbides, nitrides, alloys, and oxides. Table 2 shows the electrochemical performances of the corresponding representative HER performance for self-supported TM-based electrocatalysts in alkaline electrolytes. It is concluded that TMPs and NiMo-based electrocatalysts usually exhibit superior catalytic activity in alkaline electrolytes compared with other TM-based electrocatalysts ( Table 2). Moreover, considering the activity and fabrication cost, industrial-grade Mo plate or Mo mesh shows huge potential as self-supported substrates and Mo sources for fabricating lowcost, high-performance, and durable Mo-based electrocatalysts suitable for large-scale commercial hydrogen production.
Compared with the conventional powdery electrocatalysts, the self-supported electrodes provide a simple electrode preparation process, a lower cost, abundant catalytic sites, rapid charge transfer, and avoidance of electrocatalyst shedding. Despite the rapid development of self-supported TM-based HER electrocatalysts in alkaline water electrolysis, it still requires continuous efforts in the design of large-current-density and stable electrocatalysts for industrial applications. From the aspect of electrocatalytic systems, several challenges and perspectives should be pursued in the future. 1) Developing highly efficient self-supported electrodes for the industry-scale H 2 production is urgently needed in future. Monolith electrode, as the type of self-supported electrode, which is vertically grown on a substrate of the same metal, seems to exhibit huge prospects for industrial HER. For the monolith electrode, the charges can be directly transferred from the substrate to the catalyst without crossing van der Waals interfaces, leading to the highly efficient charge injection property and thus improving the HER performance. In addition, the nearly zero-resistance interfaces for the monolith electrode can accelerate the charge transfer process. Moreover, strong covalent bonds exist between the catalyst and substrate, which will offer excellent mechanical stability to prevent the peeling off of catalysts. This is crucial for the electrocatalysts operated at the large current density for efficient HER. Hence, seeking more monolith electrodes with high stability and activity is crucial.
2) The HER mechanism in the alkaline electrolyte is unclear, and it is still debatable whether hydrogen binding energy acts as the only activity descriptor. [149] Normally, the electrocatalysts depend on the ex situ characterization for the pristine and final state after catalysis, resulting in incomplete understanding of the water-splitting reactions at the multiscale level. Computational modeling (such as DFT calculation) is a powerful tool to unravel the catalysis mechanism and structural evolution of electrodes. However, more theoretical studies and experimental results are still needed to clarify the mechanism of HER in an alkaline medium. As Nørskov et al. declared [150] , in the field of catalysis, there is no more substantial evidence for the robustness and accuracy of a theoretical framework than the ability to use that framework to identify new active materials. During the experiment, some efficient operando techniques (such as X-ray absorption, Raman, and Fourier-transform infrared spectroscopy) have been developed for in situ monitoring of the catalytic reaction and tracking the surface species, which help to reveal the real catalytic active sites, preferred reaction pathways, and the reaction www.advancedsciencenews.com www.advenergysustres.com mechanism. [151] Meanwhile, DFT calculations can combine with the transition state theory to search for reaction paths in the reaction process and combine with molecular dynamics to simulate the gas transport and trajectory of H 2 O molecules, OH, and H intermediates over time. [152] . The free energy variations and energy barriers of the reaction intermediates in the reaction paths can be calculated to estimate the reaction rate of each elementary reaction and point out the RDS. Finally, the microscopic kinetic equation will be solved to clarify the alkaline HER reaction mechanism, evaluate the activity and selectivity of the catalyst, and verify the reliability of the reaction mechanism.
3) Most of the investigated self-supported TM-based electrocatalysts still suffer from unsatisfied performance despite exploring various strategies to optimize their intrinsic activity. Therefore, it is pivotal to design novel engineering strategies to boost the HER activity. Recently, an exciting way for developing highly efficient HER catalysts has emerged for TM-based hybrids at the subnanometer level. Downsizing the nanostructures to the atomic level to construct single-atom catalysts (SACs) has attracted enormous attention due to their distinct features in further improving specific activities and maximizing atom utilization efficiency. [153] It is worth noting that the catalytic center of SACs consists of the atomically dispersed metal atom, their direct neighbor atoms, or functional groups of their supports/substrates. Moreover, there is a strong charge transfer between the metal sites and the support surfaces, which significantly impacts the chemical property of the supported isolated metal atoms. [154] It is believed that the strong binding energy formed between the single atom and the coordinating atoms in the substrate is beneficial for improving the durability and HER activity. Therefore, rationally designing SACs on self-supported TM-based electrocatalysts should be a promising prospect for improving electrochemical activity. 4) The superhydrophilic or superaerophobic properties have to be considered before designing electrocatalysts. It is known that massive gas bubbles are rapidly formed and accumulated on the interface of electrocatalysts and electrolytes at large current density, decreasing the number of active sites and hindering liquid mass transport. Meanwhile, the mass transfer becomes an RDS with rapid catalytic reactions at large current density. Superhydrophilic surface structures can be assembled by forming textured structures (rough and/or porous, arrays, et al.), which are expected to promote the interface wettability and accelerate the electrolyte recharge and the gas bubbles releasing under the large current density. 5) There is still a large distance between the laboratory and industrial HER. Although many electrocatalysts have been demonstrated to be highly durable at large current density, it is still limited to the laboratoryscale test condition. The actual industrial application conditions need to be simulated to evaluate the electrocatalysts. Under rigorous industrial testing conditions (higher temperature, pressure, and electrolyte concentration), the structural stability of the catalyst, catalytically active sites, band level of the catalyst, the conductivity of the electrolyte, and the migration rate of ions will be significantly altered. Therefore, considering the huge difference between the laboratory scale and the industrial condition, implementing electrocatalysts into practical electrolyzers for the industrial application of as-developed catalysts is highly imperative.
In the past decades, enormous studies and efforts have contributed to developing self-supported electrocatalysts working at large current densities. Although there are many challenges and obstacles to overcome, we firmly believe that self-supported electrocatalysts have promising prospects to become future candidates for large-scale hydrogen production by industrial water electrocatalysis. We hope this review will help in promoting this field from basic laboratory research to practical application.