Zhang Cong; Wang Shuaiyang,2; Rong Junfeng; Mi Wanliang

(1. SINOPEC Research Institute of Petroleum Processing, Research Center of Renewable Energy, Beijing, 102299, China;2. China Uniνersity of Petroleum (Beijing), School of Chemical Engineering and Enνironment, Beijing, 102249, China)

Abstract: Hydrogen production by water electrolysis is an important route for generating green hydrogen. However, the development of efficient and inexpensive electrocatalysts is crucial for future industrial applications. Amorphous catalysts possess a large specific surface area with abundant structural defects. In addition, their structures can be tuned to provide more efficient active sites, leading to superior water electrolysis activity compared with their crystalline counterparts. In this review, we summarize recent progress on amorphous electrocatalysts for water splitting, with a focus on the reaction mechanisms under both acidic and alkaline conditions, catalyst synthesis, and the application of these catalysts to the hydrogen and oxygen evolution reactions. Moreover, we highlight the current challenges and promising opportunities relating to amorphous catalysts for electrochemical water splitting.

Key words: electrocatalysis; water splitting; amorphous; synthesis; mechanism

1 Introduction

The development of clean energy is crucial for solving the growing energy crisis and mitigating environmental pollution. Hydrogen is considered the most ideal energy carrier because of its high calorific value, lack of pollution, and renewable nature[1]. Hydrogen production by water electrolysis based on renewable energies has the advantages of high efficiency, high purity, and environmental friendliness and represents an important approach to green hydrogen generation, carbon emission reduction, and the realization of a sustainable hydrogen energy economy[2]. The key to its further development toward industrial application lies in the design of highefficiency and low-cost electrocatalysts.

A diverse range of crystalline catalysts have been studied in an effort to overcome the slow kinetics of the hydrogen evolution reaction (HER) and oxygen evolution reaction(OER). The main approaches have included the following:(1) tuning the catalyst morphology to increase the conductivity and specific surface area[3]; (2) doping with metals to optimize the adsorption strength of the reaction intermediates OOH* and H*[4]; (3) using appropriate supports to facilitate electron transfer and reduce the required amount of noble metals[5]; (4) introducing defects to improve the inherent activity by optimizing the electronic structure[6]; and (5) forming heterostructures to regulate the surface strain of the catalyst, thus promoting electron transfer and durability[7]. These strategies have improved our basic understanding of the catalytic process,which can be exploited to guide the rational design of efficient water-splitting electrocatalysts. However, the development of inexpensive catalysts featuring high performance and good stability remains a considerable challenge.

In recent years, it has been found that amorphous catalysts outperform their crystalline counterparts in electrochemical water splitting[8]. As shown in Figure 1, amorphous catalysts, which simultaneously possess short-range order and long-range disorder, can affect catalytic processes in distinct ways. First, crystalline materials only provide sites for surface catalysis,whereas amorphous catalysts can mediate both internal and surface electrocatalysis owing to their structural and chemical disorder[9]. Second, the abundant defects and dangling bonds present in amorphous catalysts often result in enhanced intrinsic activity and electrochemical stability[10]. Third, the structural flexibility of amorphous catalysts facilitates their selfhealing according to catalytic conditions and promotes the transformation of inert sites into active species[11].In addition, under the reaction conditions crystalline structures can change into amorphous structures that serve as the actual active species for catalyzing water splitting[12].

Figure 1 Advantages of amorphous catalysts over crystalline catalysts

Figure 2 (a) Volcano plot of exchange current density (j0) versus hydrogen adsorption energy (EM–Hads) for the acidic HER[14].(b) Schematic representation of the alkaline HER on a M-Ni(OH)2 surface and AC+ represents hydrated cations[15]

In this review, we aim to provide an overview of recent progress in amorphous electrocatalysts for water splitting. In Section 2, we introduce the mechanisms of the anodic and cathodic reactions of electrochemical water splitting and the key adsorbed species governing the catalytic performance. In Section 3, we describe the main approaches for synthesizing amorphous catalysts.In Section 4, we discuss the applications of amorphous materials for the HER and OER under acidic and alkaline conditions to help beginners in this field get exposed to the basics. Finally, in Section 5, we provide a summary of the current status of the field and highlight some unresolved issues and future challenges.

2 Mechanisms of Electrochemical Water Splitting

2.1 Hydrogen evolution reaction

The HER is generally considered to be favored in an acidic electrolyte containing a large number of protons[13].According to the Volmer-Tafel mechanism, protons and electrons combine on the surface to form adsorbed hydrogen atoms. Subsequently, two adsorbed hydrogen atoms react to afford a hydrogen molecule, as shown in equations (1) and (2):

Alternatively, according to the Volmer-Heyrovsky mechanism, an adsorbed hydrogen atom combines with another proton and electron to generate the hydrogen molecule in the second step, as shown in equations (3)and (4):

As shown in Figure 2a, acidic HER activity is related to the adsorption energy of a hydrogen atom on the catalyst surface[14]. It can be seen that a moderate hydrogen adsorption energy is required for a catalyst to exhibit excellent acidic HER activity.

Under alkaline conditions, the HER proceeds via a different Volmer reaction and Heyrovsky reaction[16].According to the Volmer-Tafel mechanism, molecular water rather than a proton is coupled with an electron,resulting in an adsorbed hydrogen atom and a hydroxide anion. Two adsorbed hydrogen atoms then combine to form molecular hydrogen, as shown in equations (5) and (6):

According to the Volmer-Heyrovsky mechanism, the adsorbed hydrogen atom combines with another water molecule and electron to generate the hydrogen molecule in the second step, as shown in equations (7) and (8):

It is known that alkaline HER activity is not only related to the adsorption energy of hydrogen on the surface but also depends on the decomposition ability of water molecules. As depicted in Figure 2b, transition-metal hydroxides (Ni(OH)2) have strong adsorption ability for OH, which promotes the decomposition of water molecules, while platinum has strong adsorption ability for H, which is conducive to the generation of hydrogen.Therefore, a series of Pt-M(OH)2(M = Fe, Co, Ni)catalysts were designed to enhance the alkaline HER activity[17].

Figure 3 (a) Adsorption energy of OOH＊ (ΔEOOH) plotted against adsorption energy of OH＊ (ΔEOH) for various oxides[20]. (b)Volcano plot of theoretical OER overpotential (ηthe) using the second charge-transfer reaction as a descriptor[21]. GS1 and GS2 represents the Gibbs free energy of first electron transfer and second electron transfer, respectively

2.2 Oxygen evolution reaction

The sluggish kinetics of the OER at the anode greatly hamper the efficiency of electrochemical water splitting.The most widely accepted OER mechanisms are the traditional adsorbate evolution mechanism (AEM) and the lattice oxygen mechanism (LOM).

(1) Adsorbate evolution mechanism

The conventional AEM involves four concerted proton/electron transfer steps. Under alkaline conditions, the dissociated hydroxide ion is adsorbed on the surface and deprotonated to form OH*, O*, and OOH*. Finally,OOH* combines with OH?to liberate molecular oxygen,as shown in equations (9)-(12)[18]:

Under acidic conditions, two molecules of water are adsorbed on the catalytic site, followed by three subsequent deprotonation steps to form an OOH*intermediate, which is finally oxidized to O2, as shown in equations (13)-(16)[19]:

Because OOH* formation or OH* deprotonation is generally considered the rate-limiting step in the OER, the binding energy difference between O* and OH* is used as a descriptor for predicting activity. As shown in Figure 3a, the adsorption energies of OH*, O*, and OOH* are linearly correlated. In particular, the gap between ΔGOOHand ΔGOHis a constant value of 3.2 eV, resulting in an estimated minimum theoretical overpotential of 0.37 eV[20]. Therefore, breaking the scaling relationship is the key to going beyond the activity limitation. For example, this limitation can be removed by incorporating nickel or cobalt into the RuO2surface, which introduces a proton donor/acceptor to the otherwise inert bridging sites (Figure 3b)[21].

Figure 4 Comparison of the AEM and LOM mechanisms

(2) Lattice oxygen mechanism

The LOM can bypass the scaling relation limitation by enabling lattice oxygen redox reactions. Rong et al.proposed a basic reaction pathway where lattice oxygens participate by forming surface oxygen vacancies[22]. First,an OH* intermediate combines with a metal ion and then a surface lattice oxygen atom reacts with OH* to form an oxygen vacancy and an OO* intermediate. Subsequently,an oxygen molecule is released and the lattice oxygen atom is replenished by adsorbing another water molecule to complete the OER cycle.

In contrast to the four concerted proton/electron transfer steps in the AEM, the LOM consists of five elementary steps[23]. Specifically, the adsorbed O* species formed in the first two steps combines with a lattice oxygen atom(OL) in the catalyst to release an oxygen molecule. This simultaneously generates an oxygen vacancy (Vo), which reacts with a water molecule to regenerate the active site for the next catalytic cycle, as shown in equations(17)-(21):

(3) Difference between the AEM and LOM

The main difference between the LOM and AEM is that the oxygen atoms in the former originate not only from water molecules but also from lattice oxygens in the oxide. In addition, as shown in Figure 4, (1) the active sites are coordinating unsaturated metal ions for the AEM but coordinating unsaturated oxygen ions for the LOM[22];(2) the reaction energy barrier is lower for the LOM than for the AEM based on DFT calculations, resulting in higher catalytic activity[24]; (3) most LOM mechanisms suggest four non-concerted proton/electron transfer steps, which account for the observed pH-dependent OER activity[25]; and (4) the LOM usually suffers from metal dissolution and Voformation, leading to structural instability, whereas the AEM does not involve such surface changes or structural collapse[26]. It should also be noted that these two mechanisms may coexist and compete with each other in some cases. Thus, balancing the AEM and LOM is still a great challenge when attempting to develop efficient and stable OER catalysts.

Figure 5 (a) Schematic illustration of the preparation process for A-NiCo LDH nanosheet arrays supported on nickel foam[36]. (b) Schematic illustration of the fabrication of α-Ni(OH)2 and α-NixFey(OH)2 for enhanced OER catalysis[37]

3 Synthesis of Amorphous Catalysts

There are two main strategies for preparing amorphous nanomaterials: (1) the direct synthesis of structures with short-range order and long-range disorder and (2) the treatment of crystalline structures to break the long-range order while maintaining the short-range order.

3.1 Direct synthesis of amorphous catalysts

Several methods have been developed for directly preparing amorphous structures. Among them,electrodeposition and co-precipitation have emerged as efficient techniques owing to their simplicity, low cost,and scalability. An amorphous OER catalyst based on cobalt and phosphate was synthesized in situ by the electrochemical oxidation of an indium tin oxide electrode in phosphate buffer solution containing cobalt ions[27]. An aerosol-spray-assisted approach has also been investigated for continuous industrial-scale synthesis, which allows for accurate control over the catalyst composition and resulting performance by adjusting the proportion of the metal precursors. For example, this method was applied to prepare various amorphous Ni-Fe-Oxcompounds and the best activity was obtained for an Fe:Ni ratio of 6:10[28].The inverse micelle/sol-gel technique can also be used to synthesize amorphous nanocatalysts with mesoporous structures. Miao et al. reported the synthesis of amorphous Fe2O3catalysts that were rich in mesopores by annealing precursors at 150 °C under air[29]. Furthermore,hydrothermal and oxidation/reduction methods are regarded as efficient strategies for preparing amorphous nanocatalysts featuring abundant pores and favorable morphologies. For example, an amorphous nickel cobalt phosphate bifunctional electrocatalyst was obtained by the hydrothermal method[30]and amorphous cobalt iron oxyhydroxide nanosheet arrays were rapidly grown in situ on iron foam by oxidizing FeN and Co(NO3)2at room temperature[31].

3.2 Indirect synthesis of amorphous catalysts

In addition to direct synthesis methods, the amorphization of nanocrystals by physical or chemical treatment is also an effective strategy. In terms of physical treatment,the long-range order can be broken by certain physical processes, such as pressure, mechanical force, thermal treatment, and irradiation. For example, Poryvaev et al.investigated the transformation of the zeolitic imidazolate framework ZIF-8 to amorphous materials under various pressures[32]. Qiao and co-workers found that crystalline AgS4can be converted to amorphous Ag3PS4by ball milling with P2S5to afford a threefold increase in the conductivity of Ag+compared with the crystalline structure[33]. Moreover, crystalline materials can be converted to amorphous phases by low-temperature radiation. For example, Widmer et al. reported that synchrotron X-ray radiation induced the amorphization of ZIF-4, ZIF-62, and ZIF-zni[34]. In situ microwave-induced amorphization is a new technology with the potential to solve the stability problem of amorphous solid dispersions during preparation and storage[35].Compared with physical treatment, chemical methods are often more facile. These methods exploit chemical reactions to induce a controllable phase transition or changes in the coordination environment by introducing defects and disorder. The first such method is the formation of amorphous structures by electrochemical reactions, during which the thin surface layer is converted in situ into an amorphous phase at the electrode potential[38]. For instance, atoms with a low binding energy (F, P, and S) can be easily dissolved during electrocatalysis, resulting in disintegration of the original crystalline structure and its transformation into a porous amorphous structure[39-40]. It is important to note that the amorphous layer formed in situ can protect the internal structure from erosion by the electrolyte, thus the leaching effect only promotes amorphization on the surface. Second, heteroatom doping can also effectively induce partial or complete amorphization. As depicted in Figure 5a, the amorphization of a layered double hydroxide (LDH) was induced by doping boron atoms[36].Compared with its corresponding crystalline structure,this amorphous catalyst with rich oxygen vacancies and a partially disordered structure had the advantage of exposing more coordination unsaturated atoms, resulting in a large number of catalytically active sites. Sun and co-workers prepared an ultrathin α-Ni(OH)2amorphous phase by doping with iron, and its amorphous degree could be controlled in the range of 10%-78%, as shown in Figure 5b[37].

4 Applications of Amorphous Catalysts in Electrochemical Water Splitting

4.1 Amorphous catalysts for acidic HER

The synthesis of amorphous catalysts is an effective approach to increase the density of active centers for electrochemical water splitting. Among various amorphous HER catalysts, amorphous metal sulfides and phosphides have been widely studied. In particular,amorphous MoSxexhibits outstanding HER activity due to its disordered structure and large number of defective active sites[41-42]. For example, MoSxlayers were successfully obtained by room-temperature deposition and the overpotential corresponding to a current density of 10 mA/cm2was only 180 mV after electrochemical activation[43]. However, determination of the actual active site of MoSxremains a challenge. It was proposed that the active site of MoSxis a Mo(III) species connected by two S22?species[44]. However, other researchers have asserted that the active site is a MoVdefective site rather than MoV(S2)[45]. In addition, the activity of amorphous molybdenum sulfide can be further improved by elemental doping or loading. For example, owing to the high specific surface area and conductivity of nanoporous gold, the current density of amorphous molybdenum sulfide thin films was increased by more than six times after their deposition on this material[46]. The number of active sites and conductivity of amorphous MoSxwere also increased through nitrogen doping, and the resulting overpotential at a current density of 10 mA/cm2was only 143 mV[47].

Compared with metal sulfides, metal phosphides contain more coordination unsaturated atoms on their surfaces.As shown in Figure 6a, an amorphous CoPi catalyst was immobilized on functionalized carbon nanotubes by electrodeposition[48]. The initial potential of the CoPi/PF-CNT catalyst was as low as 29 mV, while the overpotential corresponding to a current density of 10 mA/cm2was only 105 mV. As depicted in Figure 6b, a CoMoP@PS/NCNT catalyst featuring a unique core-shell structure was formed by coating a thin layer of amorphous phosphorus sulfide around CoMoP crystals, which were then fixed on nitrogen-doped carbon nanotubes[49]. The overpotential of the composite catalyst was only 80 mV at a current density of 10 mA/cm2. Similarly, a thin layer of amorphous FexP was coated on the surface of Fe2N[50].The electronic state and coordination environment of the iron atoms were adjusted by decoration of phosphorus atoms, which further improved the activity and stability of the acidic HER.

Figure 6 (a) Synthetic procedure and polarization curves for amorphous cobalt phosphate anchored on tri (4-fluorophenyl)phosphane functionalizing carbon nanotube (CoPi/PF-CNT) nanocomposites[48]. (b) Structural characterization and HER activity of a core-shell hybrid catalysts composed of transition metal phosphosulfide decorated at N-doped carbon nanotubes(CoMoP@PS/NCNT)[49]

4.2 Amorphous catalysts for alkaline HER

According to the reaction mechanism of the HER under alkaline conditions, hydrogen production requires additional hydrolysis steps. Amorphous metal sulfides and phosphides are commonly used in the alkaline HER[51]. The activity of highly conductive CoP nanowires was reported to be similar to that of Pt/C, with an overpotential at 100 mA/cm2of only 114 mV[52]. The synergistic effects of sulfur doping and phosphorus vacancies not only weaken the hydrogen adsorption but also strengthen the water adsorption, thus improving the alkaline HER activity. Moreover, the construction of crystalline/amorphous interfaces is an effective strategy for improving the alkaline HER activity[53]. For example,Chen et al. synthesized crystalline/amorphous Co2P/CoMoPxnanoparticles supported by nickel foam with an overpotential of 121 mV at a current density of 10 mA/cm2[54]. This excellent HER activity was ascribed to the large interface between crystalline Co2P and amorphous CoMoPx, which effectively exposed active sites and promoted electron transfer. An effective interface between crystalline cobalt molybdate and amorphous cobalt phosphide (CoMoO4@a-CoPx) was also constructed by incomplete amorphous phosphating on the surface of crystalline cobalt molybdate[55]. The amorphous CoPxshell provided sufficient active sites, while the CoMoO4crystalline core promoted charge transfer. This unique core-shell structure accelerated the alkaline HER activity,resulting in an overpotential of 74.7 mV at 10 mA/cm2.

Transition-metal hydroxides display strong OH adsorption ability, which can promote the cleavage of H-OH bonds under alkaline conditions. The combination of transitionmetal hydroxides with sulfides or phosphides, which have strong hydrogen adsorption ability, is an efficient approach for improving alkaline HER performance. For example, Liu et al. reported novel hollow nanospheres composed of amorphous Ni(OH)2and a crystalline CuS mesoporous material, which exhibited a low overpotential of 95 mV at a current density of 10 mA/cm2[56]. On the one hand, amorphous Ni(OH)2was beneficial for providing abundant active centers. On the other hand,CuS accelerated the charge conversion and enhanced the structural effect of the porous nanospheres. Similarly,owing to the interface effects between phosphides and metal hydroxides, Ni-P/Ni(OH)2afforded excellent alkaline HER activity with an overpotential of 54.7 mV at a current density of 10 mA/cm2[57].

4.3 Amorphous catalysts for acidic OER

Numerous crystalline structures have been demonstrated to undergo surface reconstruction to form amorphous structures, which have been identified as the actual active sites, especially in the OER process. Iridium/ruthenium-based amorphous materials have been found to be superior catalysts for the acidic OER with minimum overpotentials. For example, Blakemore et al. reported that the activity of amorphous IrOxsynthesized by electrodeposition was better than that of crystalline IrOxfilms[58]. Li and co-workers synthesized a series of singlecomponent (Ir, Ru, and Rh), binary (IrNi, IrFe, IrCo,and RhRu), and ternary (IrRhRu) amorphous nanosheets to improve the acidic OER activity[59]. In situ X-ray absorption fine structure spectra indicated that the valence of the iridium in the catalysts remained below +4 during the OER process. The nanosheet structure was also retained after stability tests.

Considering the two different OER mechanisms, it is easier for oxygen atoms to insert into or detach from an amorphous structure compared with catalysts possessing high crystallinity. It can be concluded that the OER tends to proceed through the AEM process on crystalline catalysts, while the LOM with its lower reaction energy barrier is more suitable for amorphous catalysts. The results of Stoerzinger et al. and Fabbri et al. support this conclusion[60-61]. It was demonstrated that the unique“oxygen atom exchange” phenomenon of the LOM would not occur on a perfect crystalline RuO2film, whereas the lattice oxygen atoms of the catalyst would participate in the catalytic reaction for amorphous RuO2containing unsaturated coordination edges and abundant defects.Pfeifer et al. first studied the properties of iridium and oxygen species in the OER process by an in situ technique based on the use of a proton-exchange membrane[12].They reported that the redox reactions of electrophilic OI-species played a key role in the catalysis of the OER.The amorphous iridium hydroxide formed during the OER process provided a large amount of active oxygen species for forming O-O bonds. Wang et al. proposed that the local strain effect in amorphous RuTe2could improve the coupling ability of p/π orbitals of Te and the electron transfer ability of Ru[18]. The high density of defect sites in amorphous structures is beneficial for adsorbing oxygen atoms to form RuOxHyspecies with superior activity.

4.4 Amorphous catalysts for alkaline OER

In recent years, amorphous transition-metal oxides have attracted considerable attention as alkaline OER catalysts. Compared with crystalline materials, the disordered atomic structure of amorphous catalysts is beneficial for strengthening the electronic interactions and interatomic bonding[62]. For example, amorphous CoOxdisplays a larger electrochemical effective area and lower electrochemical impedance compared with the corresponding crystalline structure[10]. Moreover,the adaptive relationship between elements and active centers can be adjusted by introducing transition metals with various valence states, which is conductive to improving the efficiency of water electrolysis. The overpotential corresponding to a current density of 10 mA/cm2was reduced from 378 to 309 mV by the in situ doping of Fe3+into an amorphous CoOxcatalyst[63].In addition to altering the coordination environment,the doped metal ions in amorphous catalysts have the advantages of overcoming the higher diffusion energy barrier in nanocrystals and providing more active sites for the alkaline OER.

Amorphous transition-metal hydroxides are also used for the alkaline OER on account of their good ability to cleave the H-OH bonds of water. As shown in Figure 7, amorphous NiFe(oxy)hydroxide nanosheets with various Ni:Fe ratios were directly grown on pre-peeled 3D graphite foil[64]. Because of the synergistic effect of the high conductivity of graphite, the gas diffusion ability of the 3D graphite foil, and the high activity of the NiFe(oxy)hydroxide, the amorphous catalyst exhibited low overpotentials of 210 and 251 mV at current densities of 10 and 500 mA/cm2, respectively. An amorphous nanoelectrocatalyst that is composed of amorphous Ni-Fe bimetallic hydroxide film-coated, nickel foam (NF)-supported, Ni3S2nanosheet arrays (denoted as Ni-Fe-OH@Ni3S2/NF) was successfully prepared by a simple and ultrafast (5 s) ion-exchange method[62]. The catalyst displayed high OER catalytic activity at a high current density of 1000 mA/cm2, and the activity and structure showed no discernible change after a 50 h stability test.In addition, an amorphous Co-Fe-OH nanosheet array was directly grown on the surface of a macroporous iron foam and this composite exhibited high OER activity with a Faraday efficiency of 93%[31]. The corresponding overpotentials were 208 and 298 mV at current densities of 10 and 500 mA/cm2, respectively.

Figure 7 (a) Fabrication process, (b) scanning electron microscopy image, (c) transmission electron microscopy image, and(d) selected-area electron diffraction pattern of the NiFe(oxy)hydroxide nanosheet integrated partially exfoliated graphite foil(NiFe/EG) electrode. (e) iR-corrected linear sweep voltammetry curves, (f) Tafel plot, and (g) overpotentials at 10 and 500 mA/cm2 of the NiFe/EG-1.2 V (upper potential cutoff = 1.2 V) electrodes with various Ni/Fe ratios[64]

5 Summary and Outlook

The development of efficient and stable catalysts is crucial for promoting the commercialization of electrochemical water splitting. Compared with crystals, amorphous catalysts have demonstrated improved performance for the electrolysis of water. On the one hand, amorphous catalysts contain abundant defect sites, which is beneficial for optimizing the electronic structure and improving the intrinsic activity. On the other hand, they can provide larger specific surface areas and expose more active sites. In addition, the structural flexibility of amorphous catalysts enables them to self-heal during the reaction.

Despite the considerable progress made in amorphous electrodes for water splitting, some issues have yet to be resolved. First, the exact identities of the active sites of amorphous materials still remain elusive owing to the uncertainty regarding the precise local geometry.Thus, the exploration of operando measurements to elucidate the valence state, coordination environment,and connection mode will be essential in the future.Second, despite the widespread use of theoretical simulations to evaluate the electrocatalytic reactions of crystalline structures, it is still difficult to establish such models to accurately represent the long-range disordered structure of amorphous materials, which hinders the use of these simulations to understand amorphous catalysis. Therefore, advanced techniques are urgently needed to model the short-range ordered and long-range disordered structures of amorphous materials. Third,the relationship between the amorphous phase and electrochemical stability must be revealed to develop optimized catalysts. Thus, it is necessary to conduct comprehensive analysis of the amorphous geometry,which may allow the stability to be effectively improved via local passivation or doping/alloying after predicting the sites of electrolyte attack. Finally, even though the amorphous state is beneficial for the intrinsic activity, a concomitant reduction in conductivity is unavoidable.Therefore, new techniques for adjusting the relationship between the degree of disorder and conductivity should be developed. The incorporation of metal-metal bonds or coupling to conductive substrates may facilitate efficient electron transport across the catalyst surface. Despite these challenges, there is growing interest in the use of amorphous catalysts for electrochemical water splitting.More effort should be devoted to the design of efficient amorphous materials, the study of their formation processes, and the catalytic reaction mechanisms of water electrolysis.

Acknowledgments:We are thankful for financial support from the National Natural Science Foundation of China (U20B6002)and the National Key Research and Development Program of China (2021YFB4000205).