Recent advances in non-noble metal-based bifunctional electrocatalysts for overall seawater splitting

bifunctional electrocatalysts with high catalytic activity and durability suitable for seawater electrolysis because of the scarcity of precious metals and inadequate state-of-the-art materials for the overall reaction. The development of high-performance bifunctional electrocatalysts is crucial to the commercialization of overall seawater electrolysis and in this review, the mechanism and challenges of seawater electrolysis are introduced. Optimization strategies for different types of non-noble-metal-based electrocatalysts including structural regulation, interface regulation, doping regulation, in situ assembly, alloying, and amorphization are summarized to elucidate the re- lationship among composition, structure, and properties. Finally, the challenge and prospective for future development of non-noble-metal-based bifunctional catalysts are discussed. This paper aims at providing guidance and insights into the rational design of highly efficient catalytic materials for practical seawater splitting.


Introduction
Extensive use of fossil fuels has exacerbated environmental pollution due to the emission of a large amount of carbon dioxide (CO 2 ) which may cause the greenhouse effect [1] and more countries are committing to the transition to clean and green energy to achieve carbon neutrality [2]. Hydrogen (H 2 ) is a clean, sustainable, carbon-free energy source with high specific energy compatibility and a desirable candidate to replace fossil fuels [3,4]. At present, more than 96% of the world's H 2 still comes from the reformation of fossil fuels and this process itself emits a large amount of CO 2 . Hence, an attractive alternative is to produce hydrogen and oxygen by water splitting using thermal, photocatalytic or electrocatalytic techniques and in particular, electrocatalytic water electrolysis is a hot research topic.
Water electrolysis consists of two half reactions, the oxygen evolution reaction (OER) at the anode and hydrogen evolution reaction (HER) at the cathode. Both reactions require highly efficient electrocatalysts to offset the slow kinetics and achieve high energy efficiency. In the electrolytic cell system, the theoretical voltage required to achieve overall water splitting is 1.23 V. However, in the actual commercial seawater splitting system, a voltage of 1.8-2.0 V is required to drive water splitting to generate clean energy H 2 . In order to reduce the HER and OER overpotentials and accelerate the reactions, high-efficiency bifunctional catalysts are imperative. As shown in Fig. 1, the hydrogen adsorption free energy (ΔG *H ) directly reflects the hydrogen evolution efficiency, and its absolute value close to 0 means that the corresponding catalyst shows better activity, so Pt-based metals have better HER activity than other metals [5]. Similarly, the free energy difference of intermediate formation (ΔG *O -ΔG *OH ) reflects the oxygen evolution efficiency, and its absolute value close to 1.6 means that the corresponding catalysts act as more competitive candidates, so noble metal oxides such as IrO 2 and RuO 2 have excellent OER activity [6]. But the high price, natural scarcity, and poor stability of precious metals hinder the practical development of new energy conversion and storage technologies.
Non-noble metal-based catalysts, such as transition metal and its heterozygous structure catalysts such as carbon-based catalysts, etc., have become one of the most attractive substitutes for noble metal catalysts due to their high catalytic activity, excellent stability and low price in HER or OER. Nevertheless, there are very few bifunctional non-noble metal-based catalysts that meet the requirements for overall seawater hydrolysis. Therefore, it is urgent to develop non-noble-metal-based electrocatalysts with high activity and stability and in recent years, non-noble-metal electrocatalysts with modulated electronic structure, high conductivity, and optimized adsorption energy of the intermediates have been actively explored [7].
This review describes the mechanism of electrocatalytic water splitting and challenges encountered by commercial seawater electrolysis. Effective strategies to improve the selectivity and stability of bifunctional electrocatalysts are presented and challenges and prospective of seawater electrolysis are discussed.

Challenges of overall seawater electrolysis
The electrolysis of natural seawater is similar to pure water and is divided into two half-reactions, HER at the cathode and OER at the anode. However, seawater electrolysis is more complicated due to the existence of many side reactions.
At present, seawater electrolysis shows great promise in future hydrogen production, but the industrial application still has a long way to go because of some stoppers to be discussed in this section.

Challenges in HER
The energy barrier for HER is higher in neutral seawater compared to the acidic electrolytes rich in H + , which leads to slow HER kinetics. For HER, with the increase of electrolysis current, the local pH value near the cathode increases, resulting in the precipitation of various dissolved ions (Table 1) in seawater and covering the surface of the active site, thereby reducing the active center of the cathode. In addition, bacteria and microbes may corrode and poison the electrodes, seriously affecting their stability in seawater [8,9].

Challenges in OER
In addition to the effects of bacteria and microorganisms mentioned above, the fact that OER is a four-electron process and involves multiple intermediates results in inherently poor kinetics and thus requires higher overpotentials to drive. Another major challenge of seawater electrolysis is chlorine evolution reaction (CER), which typically occurs at the anode and competes with OER at a relatively high overpotential and chloride ions corrode the electrodes [14,15]. Dresp et al. [16] have listed the possible redox reactions during seawater electrolysis. NaCl and KCl produce CER at the anode and compete with OER on the anode against H 2 /O 2 production at overpotentials well below ClO − formation according to the following Eqs. (4) and (5): Eqs. (1) and (2) indicate that CER involves only two electrons, which may have faster dynamics than the four-electron transfer OER. However, the Pourbaix diagram (Fig. 2) shows that OER has a higher thermodynamic redox potential than the CER at all pH values [17]. In the alkaline system (blue region in Fig. 2), a standard potential of CER is 480 mV higher than that of OER, but the difference is smaller in the acidic environment. Therefore, OER is required to be kept at a low overpotential of less than 480 mV to produce O 2 and to prevent chlorine precipitation in the alkaline medium.

Structural regulation
It is well known that the current densities of catalysts increase with the active center densities and more active centers translate into higher electrocatalytic activity. Structural modulation is a promising way to enhance the density of active sites [19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37] and the optimized structures also play a key role in mitigating electrode corrosion in seawater. Seawater electrocatalysts with the sandwiched structure, core-shell structure, nanorods, and nanosheets have been proposed for overall seawater splitting ( Table 2). By designing the proper surface morphology, a large specific surface area and high surface activity can be achieved and the contact area between the electrode and electrolyte (seawater) can be improved to facilitate the transfer of electrons, ions, and reaction products.
Wood with a unique anisotropic multi-channeled structure is economical and widely available. Chen et al. [38] have removed hemicellulose and lignin from natural balsa wood and coated the NiMoP alloy (loading: 12.87 mg cm −2 ) on wood aerogel to produce the sandwiched S, P-(Ni, Mo, Fe)OOH/NiMoP/wood aerogel electrocatalyst by a one-step method (Fig. 3a). Owing to the open and aligned micro-channels (Fig. 3d), large specific surface area, good wetting ability of 3D wood aerogel, as well as high conductivity, adhesion strength, and corrosion resistance of NiMoP (Fig. 3e), the assembled electrolyzer with S, P-(Ni, Mo, Fe)OOH/NiMoP/wood aerogel as the two electrodes, shows efficient and stable catalytic characteristics such as a current density of 500 mA cm −2 at 1.861 V in 1 M KOH seawater and robust cycling.
Utilizing the anticorrosion performance of NiS x in seawater, Li et al. [39] have developed a Ni 3 S 2 -1 T-MoS 2 -Ni 3 S 2 (NMN) multilayered coat on the Ni substrate by a two-step hydrothermal process (Fig. 3f). The water-splitting cell with NMN-NF as both anode and cathode requires a voltage of 1.82 V for a current density of 100 mA/ cm −2 and stable overall water splitting can be performed in 1 M KOH seawater for 100 h at 100 mA/cm −2 . The top Ni 3 S 2 layer (a1 area) consisting of large and aggregated nanoparticles in the NMN-NF sandwiched structure (Fig. 3g) is mainly responsible for the HER process, while the bottom one (a3 area) for the OER due to the electron deficiency in the Ni oxidation state. The MoS 2 interlayer (a2 area) with microspheres boosts the HER and OER performance. The Ni 3 S 2 layer in the unique sandwiched structure serves as a shield to provide the corrosion resistance against chloride anions to enhance the long-term durability. In addition, NMN-NF has other advantages Table 1 Chemical composition of seawater [10][11][12][13].

Elements
Concentration ( such as the large surface area, large density of active sites, and small charge transfer resistance in overall seawater splitting. The core-shell structure improves the stability of seawater electrolysis. Jadhav et al. [40] have synthesized the FeOOH deposited β-Ni-Co hydroxide supported on nickel foam catalyst (GO@Fe@Ni-Co@ NF) catalyst (loading: 1.9 mg cm −2 ) with an outer graphene oxide layer by a three-step hydrothermal-annealing-electrodeposition process (Fig. 4a). The charge transfer resistance is lower and corrosion resistance is improved by the metal oxide layer underneath, β-Ni-Co LDH with s small interlayer distance, GO overlayer, and oxidized carbon layer generated in situ. To accomplish current densities of 20 mA cm −2 and 1000 mA cm −2 in 1 M KOH + 0.5 M NaCl at room temperature, the as-prepared electrolyzer with GO@Fe@Ni-Co@NF acted as both the anode and cathode, only requires very low voltages of 1.57 and 2.02 V, respectively and remarkable stability is demonstrated for 378 h at a current density of 1000 mA cm −2 .
Wang et al. [41] have synthesized a 3D core-shell electrocatalyst with the NiFe-LDH layer anchored on S-NiMoO 4 nanorods supported by the porous Ni foam (S-NiMoO 4 @NiFe-LDH) by the three-step hydrothermal-vulcanization-electrodeposition technique (Fig. 4b) and the catalyst offers abundant active sites, rapid electron transfer, and good corrosion resistance. The NiFe-LDH layer plays the primary role in OER and the S-NiMoO 4 nanorods beneath are responsible for HER in the simulated alkaline seawater electrolyte. A current density of 100 mA cm −2 is achieved by applying voltages of 1.68 and 1.73 V, and the voltages are lower than those of the IrO 2 || Pt/C pair (1.73 and 1.81 V, respectively) in simulated and natural alkaline seawater [42].
In addition to 3D structures on self-supporting materials, 2D structures have been constructed. Haq et al. [43] have proposed a strategy to assemble Au nanocluster decorated Gd-Co 2 B nanoflakes embedded in TiO 2 nanosheets on the Ti foil (Au-Gd-Co 2 B@TiO 2 ) (loading: 0.2 mg cm −2 ) for seawater electrolysis. Benefiting from the large surface area, abundant active sites, high conductivity, and excellent corrosion resistance, the electrolyzer with Au-Gd-Co 2 B@TiO 2 working as both the anode and cathode electrodes needs a low voltage of 1.74 V to attain a current density of 1000 mA cm −2 with no significant decline for over 200 h in 1 M KOH seawater.

Interface regulation
With regard to heterogeneous catalysts, it is essential to design the appropriate interface to increase the active sites [44] and promote charge transfer between the electrocatalyst and electrolyte at the interface. This can be accomplished by interfacial modulation [45][46][47][48][49][50][51][52][53][54][55][56] and a series of heterostructures have been prepared to enhance the overall seawater splitting characteristics (Table 3).
Owing to their high electrical conductivity, Nitride-and sulfidebased compounds are considered promising water splitting electrocatalysts [57,58]. However, the lack of active sites greatly hinders  the improvement of their electrocatalytic activities. It has been shown that the interface plays a significant role in accelerating dissociative adsorption of water [59] and improving the kinetics in electrochemical water electrolysis [60]. Zhao et al. [61] have fabricated a nickel nitride/sulfide (NiNS) electrode with abundant interfacial contact by simple one-step calcination of nickel foam with thiourea in a vacuum-sealed ampoule. The interface between the various planes of Ni 3 N (loading: 9.27 mg cm −2 ) and Ni 3 S 2 (loading: 10.68 mg cm −2 ) species in polycrystalline NiNS (Fig. 5a) is observed by HR-TEM ( Fig. 5b and c) and serves as electrocatalytic active sites for dissociative adsorption of water molecules and subsequent water electrolysis. The two-electrode electrolyzer with NiNS // NiNS couple shows a current density of 48.3 mA cm −2 at 1.8 V higher than that of the Ir-C // Pt-C (2.9 mA cm −2 ) in seawater. Through the study of electrochemical active sites, the electrocatalytic mechanism was explored, which suggested that the abundant interfacial regions predominantly accounted for the excellent catalytic performance. Ni 3 N has been adopted to build a strong coupling interface with Ru. Zhu et al. [62] have observed that Ru has a terrific lattice similar to Ni 3 N and the unusual epitaxial hetero-interface can be obtained by exploiting the similar lattices in the two different materials [63,64]. They have assembled cRu-Ni 3 N porous nanosheets on the conductive nickel foam to form cRu-Ni 3 N/NF (Fig. 5d). During nitridation, the tiny Ru clusters are grown epitaxially in situ on the Ni 3 N nanosheets to form strongly coupled heterointerfaces (Fig. 5e) at which the charge density is redistributed and the electron-enriched hetero-interfaces improve the intrinsic electron conduction. Consequently, adsorption of the key intermediates is modulated, which changes the rate-determining step and reduces the activation energy barrier in HER and OER ( Fig. 5f-i) [65]. The assembled electrolyzer with cRu-Ni 3 N/NF acted as both the anode and cathode, delivering excellent catalytic performance (1.62 and 1.73 V @ 50 and 100 mA cm −2 ) with few side reactions, higher selectivity, and good durability in overall seawater splitting, which is superior to the Pt/C || RuO 2 couple (1.70 and 1.84 V). This bifunctional electrode can be directly connected to commercial solar panels (Fig. 5j) and suggests the possibility of large-scale application with other external power supplies.
Heterogeneous bimetallic phosphides have become a hot research topic due to their structural and chemical advantages, for example, (Ni 0.33 Fe 0.67 ) 2 P [68], FeP/Ni 2 P [69], Ni 2 P-Cu 3 P [70], Fe-Co-P [71], and NiCoP [72], etc. Wu et al. [66] have synthesized a heterogeneous electrocatalyst with Ni 2 P-Fe 2 P micro-sheets supported on Ni foam (Ni 2 P-Fe 2 P/NF) (loading: 15.0 mg cm −2 ) by soaking and phosphating. The modification of Fe cations resulted in the formation of rough topography of some nanoparticles on the surfaces of the micro sheets ( Fig. 6a) which increase the specific surface area on the electrocatalyst. The hetero-interfaces between the phase boundary of Ni 2 P and Fe 2 P phases (Fig. 6c) produce the interfacial bonding effect, which is beneficial to exposing more active sites. Owing to the high intrinsic activity, plentiful active sites, an exceptional transfer coefficient, enhanced corrosion resistance, and hydrophilic surface, the as-prepared electrolyzer with Ni 2 P-Fe 2 P/NF as the two electrodes shows excellent catalytic activity and robust durability in overall seawater splitting requiring low voltages of 1.811 and 2.004 V to attain current densities of 100 and 500 mA cm −2 in 1 M KOH seawater, respectively.
Seenivasan et al. [67] have constructed a heterogeneous multiple transition metal sulfide NiCo 2 S 4 /NiMo 2 S 4 /NiO (NCMS/NiO) electrocatalyst by ion exchange with a layer of NiO on the catalyst surface by atomic layer deposition (ALD) to increase the active sites (Fig. 6d). The hollow cuboid NCMS/NiO electrocatalyst shows an obvious hetero-interface between NiCo 2 S 4 and NiMo 2 S 4 with close contact in the cuboid wall and NiO shell layer over the wall (Fig. 6e-g). An in situ reconstruction, occurring on the NiO and the metal sulfide (M-S), led to dual active sites of M-S and metal-oxyhydroxide (M-OOH) (Fig. 6h-j) with good corrosion resistance to chloride ions during high-temperature seawater electrolysis. The thicker NiOOH/NiO layer restricts the diffusion of the electrolyte to the NCMS core thus decreasing the number of active sites and OER activity. The NiO layer with a thickness of 0.2 nm and about the size of a NiO molecule (a) Reproduced with permission from ref. [40]. Copyright (2020), Royal Society of Chemistry. (b) Reproduced with permission from ref. [39]. Copyright (2022), Elsevier. (c) Reproduced with permission from ref. [41]. Copyright (2022), Elsevier. formed by 40 ALD cycles is the most effective in seawater electrolysis. The electrolyzer with NCMS/NiO as both the anode and cathode needs a low voltage of 1.505 V to attain a current density of 100 mA cm −2 with no apparent decline for over 30 days under industrial conditions.

Single-element doping
The catalytic activity of bimetallic phosphides in seawater electrolysis can be improved by not only constructing hetero-interfaces but also doping. Wang et al. [93] have synthesized a cobalt-doped Fe 2 P electrocatalyst (Co-Fe 2 P) (loading: 2.0 mg cm −2 ) on Ni foam by a two-step hydrothermal-phosphating process (Fig. 7a). XPS indicates that Co doping modifies the electronic properties of Fe 2 P, making the P part negatively charged (Fig. 7c-d) [94] which can attract protons to enhance the HER activity. Simultaneously, the presence of oxygencontaining functional groups (Fig. 7b) improves the hydrophilicity benefiting HER and OER [95]. The electrolyzer with Co-Fe 2 P || Co-Fe 2 P couple has excellent catalytic properties in overall seawater splitting as demonstrated by an operating voltage of 1.69 V at 100 mA cm −2 , which is lower than that of the RuO 2 // Pt-C (1.97 V).
Single-element doping can be performed in different ways. Kim et al. [96] have synchronously doped single nickel atoms (Ni SA ) and nickel phosphate clusters (Ni Pi ) on the matrix of MoS 2 nanosheets (NSs) supported by one-dimensional (1D)-TiO 2 nanorods (NRs) to produce Ni SA -Ni Pi /MoS 2 NSs/TiO 2 NRs (loading: 7.17 mg cm −2 ) by impregnation and thermal treatment (Fig. 7f-h). According to the calculated free energy diagrams (Fig. 7i-k), when Ni SA and Ni Pi coexisted in the MoS 2 NSs matrix, the electronic properties of the product could be effectively modified, which was favorable for electron transfer and thus played a role in promoting the HER and OER reactions. As a result, the as-prepared electrolyzer with Ni SA -Ni Pi /MoS 2 NSs/TiO 2 NRs acting as both the cathodic and anodic electrodes shows a current density of 10 mA cm −2 at low cell voltages of 1.52 and 1.66 V beside good stability in long-term operation in simulated seawater and natural seawater, which are lower than that of the Pt/C // RuO 2 couple (1.71 and 1.70 V).
Doping with noble metals can tune the electronic structure of catalysts [99][100][101][102][103][104][105] and Ru is more economical than other noble metals. Its binding energy to hydrogen is similar to that of Pt rendering it promising in HER [106][107][108]. Wu et al. [97] have constructed Ru-incorporated amorphous cobalt-based oxide (Ru-CoO x / NF) electrocatalyst by in-situ growth, anneal, and calcination (c) Reproduced with permission from ref. [66]. Copyright (2020), Wiley-VCH. (j) Reproduced with permission from ref. [67]. Copyright (2021), Royal Society of Chemistry. procedure (Fig. 8a). XPS analysis (Fig. 8b-c) indicated that the incorporation of Ru led to charge transfer, which enhanced the electrocatalytic activity. Therefore, the assembled electrolyzer with Ru-CoO x /NF working as both the anode and cathode electrodes requires low voltages of 2.62 V for a current density of 1 A cm −2 in seawater electrolysis.
Tran et al. [98] have proposed a catalyst (1D-Cu@Co-CoO/Rh) with continuous Co-CoO containing dispersed Rh atoms and a shell of conductive porous 1D Cu via electrodeposition and electroless deposition step (Fig. 8d). The synergistic effects arising from the uniform Rh atoms and Co-CoO hetero-structures, shown in Fig. 8e, produce rich multi-integrated active sites, optimize the electronic state and lower energy barriers of water dissociation for enhanced HER and OER activities (Fig. 8f-h). The electrolyzer made with 1D-Cu@Co-CoO/Rh as both the anode and cathode required cell voltages of 1.60, and 1.70 V at 10 mA cm −2 and robust stability in simulated seawater, and natural seawater, respectively.

Dual-element doping
Multi-element doping exploits the synergistic effects of different heteroatoms and provides the basis to generate more lattice defects, vacancies, and active sites to tailor the catalytic activity [109][110][111][112][113][114][115]. In particular, vanadium has an abundant reserve and flexible redox states. V-doped catalysts may be partially disordered structurally because of the partial dissolution of V in the electrolyte to stimulate the reactivity. In OER, V favors the formation of *O and so the activation potential of the catalyst is reduced [116]. Moreover, as the cheapest platinum group metal, Ru is similar to Pt with metal hydrogen bonding strength and abundant d-orbital electrons. It also has a superior ability to adsorb OH-and split water [117]. Consequently, the HER/OER activity of catalysts with an ultra-low amount of Ru introduction can be further promoted without adding excessive cost. Ma et al. [118] have doped RuV-CoNiP/NF catalysts (loading: 3.6 mg cm −2 ) simultaneously with Ru and V by phosphating Ru-impregnated phosphating CoV-LDH on nickel foam (Fig. 9a). The co-doping of Ru with V further facilitated the charge transfer, which was beneficial to accelerating the electrochemical kinetics. In addition, multiple valence states of V existed in the catalyst, which was also more favorable for the redox reaction (Fig. 9b-e). The assembled electrolyzer with RuV-CoNiP/NF as the two electrodes shows 20 mA cm −2 at only 1.538 V, better than that of the Pt/C/NF // RuO 2 /NF (1.678 V), which ranks about the best.
Chang et al. have studied the dual-doping effects by doping with Fe and P on nickel selenide nanoporous films (Fe, P-NiSe 2 NFs) (loading: 1.2 mg cm −2 ) [119]. The electronic structure and surface composition of the Fe, P-NiSe 2 is altered to boost the catalytic activity. According to DFT calculations, Fe doping played a key role in HER, while Ni might be responsible for OER (Fig. 9f-g). Furthermore, due to the co-incorporation of Fe and P heteroatoms in Fe, P-NiSe 2 , its adsorption energy and limiting potential were reduced while its electrical conductivity was increased, all of which boosted the activity, stability, and selectivity of high-efficiency seawater electrolysis ( Fig. 9h-j). As a result, a current density of 0.8 A cm −2 is achieved at 1.8 V for over 200 h in natural seawater feedstock and the properties are better than those of most other electrolyzers and reach the Department of Energy (DOE) 2020 target. (Table 4).

Others
There are other methods to enhance the activity of seawater splitting electrocatalysts, for instance, in situ assembly, alloying, amorphization, and so on. Using the surface etching method, in situ assembly of metal oxides/hydroxides can be formed as self-supported electrocatalysts directly on transition metal substrates. Metal corrosion often causes the formation of hierarchical metal oxides or hydroxides which provide easier access to active sites for OER. Moreover, the Mott-Schottky effect may enhance the catalysis between the metal matrix and corrosion layer [120,121] and preetching can corrode and remove unstable species from the metal matrix to improve the electrochemical stability. An in situ corrosion strategy has been proposed by Duan et al. [122] to construct NiFe hydroxides as free-standing electrodes for overall seawater splitting (Fig. 10a). Owing to the strong interaction between Cland metals, HCl was more conducive to the dissolution of Ni than H 2 SO 4 and HNO 3 , leading to the formation of NiFe hydroxides on the surface of matrixes. In-situ Raman spectroscopy confirmed that NiOOH was more likely generated in HCl-c-NiFe rather than H 2 SO 4 -c-NiFe or HNO 3 -c-NiFe (Fig. 10b-e), which played a key role in OER, contributing to enhancing activity for water splitting. In the overall water splitting electrolyzer, HCl-c-NiFe shows low working voltages of 1.62 V at 100 mA cm −2 , which is lower than that of the Pt/C // IrO 2 pair (1.72 V), together with outstanding stability for 300 h in alkaline seawater. Similarly, using a corrosion coordination method, Chen et al. [123] have prepared the 2D-3D nanostructure with metal hydroxides and Prussian blue analogus (PBA) on NiFe foam (Pt-NiFe PBA) by a facile and scalable corrosion approach. The specific morphology produces ample active sites, optimizes the reactions and accelerates mass transport. As a bifunctional electrocatalyst, the Pt-NiFe PBA ∥ Pt-NiFe PBA couple needs 1.48 V to drive 10 mA cm −2 in 1 M KOH seawater.
In addition to corrosion engineering, in situ assembly can be achieved in hydrolysis. Zhang et al. [124] have used the sol-gel method and further annealed to produce the Fe 3 O 4 /NiC x composite (NiFe-PBA-gel-cal) (loading: 1.0 mg cm −2 ) that inherits the large specific surface area of the parent structure. Operando Raman and XPS analyses indicated that in-situ generated FeO and the evolution of Ni(OH) 2 played important roles in HER activity, while in-situ generated NiOOH 2-x containing high-valence nickel cations and a large number of oxygen defects were mainly responsible for OER activity (Fig. 10f-l). When integrated electrolyzer with NiFe-PBA-gelcal as both anode and cathode, a low voltage of 1.66 V is required to provide a current density of 100 mA cm −2 in simulated seawater and there is no obvious attenuation for 50 h.
Alloying is an effective method to boost the properties of metal catalysts [127,128] by refining the grain size, improving the mechanical strength and specific surface area, and reducing the amounts of single components in order to reduce the cost. The catalytic activity and selectivity can be altered by adding other elements to form alloys to exploit the synergistic effects between components [129,130]. According to the Brewer-Engel valence bond theory [131], alloying metals with unfilled d orbitals and those with internal paired d electrons can enable hydrogen adsorption energy to be tuned on their surfaces to enhance the hydrogen evolution activity. Ros et al. [125] have fabricated an efficient and earthabundant Ni-Mo-Fe based electrocatalyst by simultaneous electrodeposition on graphitic carbon felts (Fig. 11a). As shown in Fig. 11b-g,   (e) Reproduced with permission from ref. [122]. Copyright (2020), Royal Society of Chemistry. (j) Reproduced with permission from ref. [124]. Copyright (2022), Wiley-VCH. with Ni-Mo-Fe || Ni-Mo-Fe couple, it required a low cell voltage of 1.59 V to reach a current density of 10 mA cm −2 for more than 24 h stability in 0.5 M KOH seawater, and with an energy efficiency higher than 61.5 %, which had broad application prospects and economic feasibility in seawater electrolysis. The amorphous structure can improve the catalytic properties by adjusting the arrangement on the atomic scale [132][133][134][135][136][137][138]. The shortrange atomic arrangement of the amorphous phase increases the density of active centers [139][140][141][142][143] and in seawater electrolysis, the amorphous structure delivers excellent performance. Liu et al. [126] have fabricated amorphous NiFeP/NF electrocatalysts for efficient and stable seawater hydrolysis by a simple and environmentally friendly method (Fig. 11 h). According to the in-situ Raman spectra and DFT calculations, the reasons for the excellent catalytic performance of amorphous NiFeP/NF were on the one hand attributed to the reconstruction of the surface structure of NiFeP/NF during the hydrolysis process to continuously exposed new active sites, and on the other hand, the adsorption energy of H* on metal sites was optimized due to the amorphous structure. The as-prepared electrolyzer with NiFeP/NF employed as both cathodic and anodic electrodes shows current densities of 100 and 1000 mA cm −2 at voltages of 1.57 and 1.80 V, respectively. It can be operated for over 500 h in simulated alkaline seawater. (Table 5).

Summary and outlook
In this paper, we comprehensively review the principles and challenges of seawater electrolysis and systematically summarize the existing achievements of bifunctional non-noble metal-based electrocatalysts for overall seawater splitting from the perspective of different novel strategies. The following points should be focused on in the future development of electrocatalysts for seawater splitting.
1. It is necessary to explore and develop electrocatalysts with high selectivity, activity, and stability to suppress the interference of various cations and chloride ions in natural seawater. Optimization of the electronic structure of catalyst active centers by structure regulation, interfacial regulation, doping regulation, etc., are highly efficient methods to obtain high-performance electrocatalysts for seawater electrolysis [144][145][146]. Materials with anti-corrosion properties should be selected as the protective layer. Inspired by previous potential candidate materials, artificial intelligence technology can be used as an efficient and convenient tool to screen out qualified electrocatalysts. The large-scale preparation of catalysts with uniform morphology and fine structure also needs to be studied urgently, which is of great significance for the practical application of seawater electrolysis. 2. The design of advanced membranes and progressive seawater electrolysis reactors are essential to improve the long-term stability of direct seawater electrolysis processes. The current focus is mainly on the design and synthesis of various catalysts. However, to achieve practical and efficient hydrogen production, it should be concentrated on the whole reactor not just the catalysts. Besides, seawater splitting should be integrated with interdisciplinary technologies such as advanced membranes, capacitive deionization (CDI) for seawater desalination, and rational design of the seawater electrolysis reactor to achieve the improvement of the overall reaction efficiency through the purification pretreatment. 3. A standard platform should be built for the research of direct seawater electrolysis. There are many differences in the composition of seawater around the world, which brings great difficulties to the comparison of different experimental results. The long-term stability test should also specify standardized parameters including uniform service time and current density. Existing catalysts for seawater electrolysis catalysts work well under laboratory conditions with well-defined compositions, concentrations, and pH values. However, these catalysts are still far from excellent performance under industrial and real seawater conditions. 4. The clearer mechanism of seawater electrolysis still needs further research. Although the main chlorine-related side reactions have been extensively studied, the inevitable interference and corrosion mechanisms of free metal ions and other halide ions in seawater are still poorly understood. A combination of experimental and computational research could provide a solution to this problem. The use of DFT calculations (including volcano plots, D-band center theory, and adsorption free energy) shows great promise in revealing catalyst reactions and active centers. Novel in-situ characterization methods should be developed to track the dynamic structural evolution and active site change of catalysts during practical seawater electrolysis. For example, advanced in-situ characterization techniques such as operando XAS, Raman, IR, XRD, and TEM are promising to understand the dynamic evolution of catalyst surface active sites, reaction intermediates, and corrosion-resistant layers in real working environments, providing clear principles and guidance for designing efficient catalysts.
In summary, it is highly recommended that more attention should be paid to the material design, device development, standardization of key parameters, and mechanism analysis for electrolysis under real seawater conditions, which will promote the industrialization of seawater electrolysis. This review summarizes the recent progress and perspectives of seawater electrolysis and serves as a reference and guide to foster the rational design of highly efficient catalytic materials for practical seawater splitting. We believe that the realization of large-scale seawater electrolysis for hydrogen production will be industrialized in the near future.

Data Availability
Data will be made available on request.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.