Wet-chemical synthesis of two-dimensional metal nanomaterials for electrocatalysis

Abstract Two-dimensional (2D) metal nanomaterials have gained ever-growing research interest owing to their fascinating physicochemical properties and promising application, especially in the field of electrocatalysis. In this review, we briefly introduce the recent advances in wet-chemical synthesis of 2D metal nanomaterials. Subsequently, the catalytic performances of 2D metal nanomaterials in a variety of electrochemical reactions are illustrated. Finally, we summarize current challenges and highlight our perspectives on preparing high-performance 2D metal electrocatalysts.


INTRODUCTION
With the increasingly serious global energy and environmental crisis caused by the excessive consumption of non-renewable fossil fuels, it is urgent to develop clean, efficient and sustainable energy conversion technologies [1,2]. As one of the most promising methods of producing clean energy, electrocatalysis can effectively convert incombustible molecules (e.g. water and carbon dioxide (CO 2 )) into renewable fuels, including hydrogen, carbon monoxide (CO), alcohols and hydrocarbons [3,4]. However, the major challenges in electrocatalysis are the sluggish kinetics and high overpotentials in electrochemical reactions, as well as the high cost and low stability of electrocatalysts [5][6][7]. Thus, rational design and preparation of highly efficient and stable catalysts in electrochemical energy conversion, such as fuel cells, metal-air batteries and water electrolysis, have become one of the most popular research directions [8][9][10].
As known, it is difficult for the intrinsically non-layered metals to spontaneously form 2D morphology, since 2D metallic nanostructures possess the fairly large surface energy of a specific facet, which is thermodynamically unfavorable [20]. As known, capping agents used in wet-chemical synthesis can greatly reduce the surface energy of nanomaterials and thus facilitate growth along specific crystalline orientations of metal nanocrystals [15]. Hence, the wet-chemical synthetic method using various kinds of capping agents has become the most powerful strategy for synthesizing 2D metal nanomaterials. In past decades, a library of 2D noble metal nanomaterials, such as Au [21], Ag [22], Pt [23] and Pd [24], have been successfully synthesized [25]. However, practical applications of 2D noble metal nanomaterials are severely hindered by the high cost and scarcity of noble metals [26]. Hence, tremendous effort has been devoted to the preparation of 2D non-noble metal nanomaterials and 2D metal alloys, which are then used as high-performance and cost-saving electrocatalysts.
Here, we give a brief overview of the recent research advances in the wet-chemical synthesis of various 2D metal nanomaterials and their application in electrocatalysis (Fig. 1). First, we introduce typical wet-chemical synthetic methods for the preparation of 2D metal nanomaterials (e.g. noble metals, non-noble metals and bimetallic alloys), such as ligand-assisted synthesis, gasmolecule-assisted synthesis, templated synthesis, space-confined synthesis and seeded-growth synthesis. Then, we summarize the electrocatalytic applications of 2D metal nanomaterials in the hydrogen evolution reaction (HER), oxygen reduction reaction (ORR), oxygen evolution reaction (OER), ethanol oxidation reaction (EOR), methanol oxidation reaction (MOR), formic acid oxidation reaction (FAOR) and carbon dioxide reduction reaction (CO 2 RR). Finally, challenges, as well as our personal perspectives on future research directions, are presented.

SYNTHESIS OF 2D METAL NANOMATERIALS
The controlled synthesis of 2D metal nanomaterials with desired composition, size, thickness and crystal phase is essential for exploring their physicochemical properties and various applications. In this section, some representative wet-chemical methods for the synthesis of 2D metal nanomaterials are introduced. Table 1 summarizes some representative 2D metal nanomaterials prepared by various wet-chemical synthetic methods.

Ligand-assisted method
Organic ligands, e.g. organic amines [27][28][29][30][31][32], halogenic-organic ligands [23,[33][34][35][36] and polymeric ligands [37][38][39][40][41], have been widely used to synthesize 2D metal nanomaterials, and can not only lower the surface energy of their specific facets, but also stabilize their 2D structure [15]. For instance, our group first synthesized 4H-phase Au nanoribbons with a thickness of 2-6 nm using oleylamine and 1,2-dichloropropane as organic ligands in the presence of hexane. The use of 1,2-dichloropropane is critical to the formation of 2D morphology; without it, only Au nanoparticles were obtained [27]. Very recently, our group also prepared square-like, freestanding Au nanosheets with a mixture of 2H and facecentered cubic (fcc) crystal phase (Fig. 2a) through reducing HAuCl 4 ·3H 2 O in a mixture of oleylamine, hexane, squalene, 1,2-dichlorobutane and 4-tert-butylpyridine [32]. The obtained 2H/fcc Au nanosheets possessed an edge length of 416 ± 160 nm, and a thickness of ∼8 nm. Impressively, the 2H/fcc Au nanosheets possessed a unique structure, i.e. 2H/fcc edges and 2H/fcc basal planes with fccstacking faults and/or twin boundaries, and fcc edges with few or no aforementioned defects. In addition, it has been demonstrated that the organic amines are able to coordinate with complex intermediates and form a ligand-confined space to guide the formation of 2D structures. For example, Gao et al. found that n-butylamine could specifically adsorb on the surface of [Co(H 2 O) 6 ] 3+ intermediates to efficiently reduce their surface energy and avoid their aggregation, leading to the formation of 2D Co nanosheets during the subsequent condensation process [29]. The average thickness of the obtained Co nanosheets, 0.84 nm, was confirmed by atomic force microscopy (AFM) (Fig. 2b), and is close to that of a 4-atom-thick Co sheet along the [001] direction. Besides the aforementioned hydrophobic organic ligands, various kinds of hydrophilic organic ligands can also be used to produce 2D metal nanomaterials   (Fig. 2c) via reducing Cu(II) nitrate by L -ascorbic acid ( L -AA) in the presence of cetyltrimethylammonium bromide (CTAB) and hexamethylenetetramine (HMTA) [33]. The L -AA not only reduced Cu 2+ , but also protected the obtained Cu nanosheets from oxidation. The use of CTAB and HMTA provided an alkaline environment to stabilize the Cu + , so that the disproportionation reaction was suppressed and growth kinetics were controlled to favor the formation of 2D morphology. Moreover, 2D snowflake-like PdAg alloy nanodendrites (Fig. 2d) were synthesized using octadecyltrimethylammonium chloride (OTAC) as the structure-directing agent [34]. The OTAC selectively passivated the {110} plane of PdAg nanodendrites (Fig. 2d) during the fast reduction process of Pd and Ag, thus leading to the formation of such ultrathin nanodendrites. Notably, Xu et al. prepared ultrathin Pd nanosheets through the confined growth inside lamellar micelles constructed by docosylpyridinium bromide [35]. In another study, the same group also synthesized ultrathin Pt nanodendrites (Fig. 2e) using a long-chain amphiphilic surfactant (C 22 H 45 -N + (CH 3 ) 2 CH 2 COOH(Br − )) as the structure-directing template [23]. Additionally, polymeric ligands with super-long molecular chains are another category of widely used capping agents, and can reduce the reaction rate and effectively prevent the aggregation of nanostructures. For example, Kim et al. employed a combinatorial library of 24 polymers (10 cationic, 6 anionic and 8 neutral polymers) to synthesize Ag nanoplates (Fig. 2f) in an aqueous system [37]. Notably, neutral and anionic polymers tended to favor the formation of 2D Ag nanoplates, while the use of cationic polymers led to the formation of spheroidal Ag nanoparticles. Besides the negative and neutral charges, they found that the appropriate length and amphiphilic structure of polymers were the keys for the formation of Ag nanoplates. Among the various polymeric ligands, poly(vinylpyrrolidone) (PVP) has attracted great attention in the synthesis of 2D metal nanomaterials because it can not only serve as the stabilizer, but also work as the mild reducing agent to kinetically control their nucleation and growth rate. To synthesize triangular Ag nanoplates with a uniform size distribution, Wijaya et al. proposed a PVP-assisted wet-chemical reduction method [39]. PVP acted as not only a surfactant, but also a weak reducing agent during the synthesis. Impressively, after introducing polyacrylamide and acetonitrile, the yield of Ag nanoplates was increased. Additionally, 2D Cd nanosheets with a thickness of 30-50 nm (Fig. 2g) were synthesized using PVP as surfactant and diethylene glycol as solvent in low temperature [38].
Inorganic ligands, such as halide ions (e.g. Cl − , Br − and I − ), can guide the growth of nanostructures in specific directions, as they can strongly bond with the surface metal atoms and adsorb on the selective facets of metal nanocrystals, thus serving as capping agents to modulate the 2D morphology of metal nanostructures. For instance, Kim et al. synthesized 2D Ag nanoplates using halide ions, i.e. Cl − , Br − and I − , as the shaping agents [42]. Particularly, the chemisorption strength of halide ions on the (100) Ag surfaces followed the order of Cl − < Br − < I − , resulting in the different thickness of Ag nanoplates. tri-iodide ions (I 3 − ) [43]. Peng et al. synthesized hexagonal few-layer antimonene (Sb) nanosheets (Fig. 2h) in a solution phase, in which the Cl − and dodecylthiol (DDT) acted as the shape-directing agents [44]. When using SbCl 3 -DDT as precursors, rhombohedral phase Sb nanosheets with a thickness of 5-30 nm were obtained. The Cl − ions were believed to passivate the (001) basal planes of Sb nanosheets and favor the growth of 2D nanosheets. Moreover, the synthesis of intermetallic fcc Pd 3 Pb nanoplates ( Fig. 2i) with the help of Br − was reported by Wang et al. [45]. Specifically, due to the strong selective bonding on the (100) facets of fcc Pd 3 Pb, Br − served as a shape-directing agent to favor the formation of square Pd 3 Pb nanoplates with the exposure of (100) facets.

Gas-molecule-assisted method
It has been found that gas molecules can play a significant role in the anisotropic growth of 2D metal nanomaterials. Among them, CO is the most  extensively used as it can selectively adsorb on specific crystal facets of metals to form a strong bond with the surface atoms [46]. Such selective adsorption of CO significantly inhibits the growth of some specific facets of metal nanomaterials, resulting in the formation of 2D metals, including monometallic Pd [24,[47][48][49], Rh [50][51][52], Ru [53] and Ir [54,55] and bimetallic alloy nanosheets based on Pt [56,57], Pd [58][59][60][61][62][63], Ru [64,65] and Rh [52,66]. In 2011, Huang et al. reported a CO-assisted method to synthesize hexagonal-shaped Pd nanosheets with a thickness of <10 atomic layers ( Fig. 3a) by directly introducing 1 bar of CO gas to the aqueous solution [24]. The formation of 2D morphology of Pd nanosheets was ascribed to the inhibition of growth along the [111] direction due to the strong coordination of CO molecules with the (111) planes of fcc Pd. Similarly, the same group synthesized Rh nanosheets with few-atom-layer thickness by charging CO into the homogenous solution that consisted of rhodium(II) acetate, PVP and N,Ndimethylformamide (DMF) [50]. With the same strategy, Feng et al. also synthesized the intermetallic PtBi nanoplates with a regular hexagonal shape [56]. However, it is quite dangerous to directly use CO gas in reactions since CO gas is highly flammable and lethal. Therefore, many CO-releasing agents, including metal carbonyl compounds (e.g. molybdenum hexacarbonyl (Mo(CO) 6 ), tung-sten hexacarbonyl (W(CO) 6 )) [47,[58][59][60]63], formaldehyde [51][52][53]64,66,67], formic acid [54,55] and DMF [54,56], are used to replace the CO gas. These compounds can decompose and release CO gradually at a certain elevated temperature. For instance, our group has recently reported a facile one-pot wet-chemical reduction method to synthesize ultrathin Pd nanosheets with the assistance of CO decomposed from Mo(CO) 6 [47]. Importantly, the percentage of crystallinity could be tuned by changing the reaction temperature and a novel amorphous/crystalline heterophase could be obtained. With a similar strategy, our group also synthesized two types of ultrathin amorphous/crystalline heterophase PdCu nanosheets (Fig. 3b), i.e. amorphous phasedominant nanosheets (a-PdCu) and crystalline phase-dominant nanosheets (c-PdCu) [59]. Interestingly, confirmed by Fourier transform infrared spectroscopy (FTIR), a-PdCu nanosheets adsorbed more CO groups than the c-PdCu after a 14-day aging process, resulting in different surface ligands and different crystallinities between a-PdCu and c-PdCu. Moreover, our group prepared a series of unconventional face-centered tetragonal (fct) PdM (M = Zn, Cd, ZnCd) nanosheets with a thickness of <5 nm (Fig. 3c) through heating palladium(II) acetylacetonate (Pd(acac) 2 ), zinc(II) acetylacetonate (Zn(acac) 2 ) and Mo(CO) 6 in oleylamine [58].
It is worth mentioning that only PdZn nanoparticles with an irregular morphology were obtained under a similar condition without Mo(CO) 6 , indicating CO decomposed from Mo(CO) 6 played a key role in the formation of 2D nanostructures. Remarkably, Wang et al. found that the thickness of Pd nanosheets (Fig. 3d) could be precisely controlled by using either pure CO gas or CO released from cobalt carbonyl (Co 2 (CO) 8 ) [48]. Specifically, by precisely controlling the amount of CO gas and CO-releasing agent in synthetic reactions, Pd nanosheets with a thickness of 3 ± 1, 5 ± 1 and 8 ± 1 monolayers were prepared, respectively. In addition, some organics, such as formaldehyde and formic acid, could also release CO molecules during the synthetic process. For example, our group has synthesized the ultrathin flower-like Rh nanosheets with amorphous/crystalline heterophase in a diluted formaldehyde aqueous solution without any surfactant [52]. CO from the decomposition of formaldehyde adsorbed on the (111) facets of fcc Rh, leading to the formation of 2D morphology. The same synthetic strategy could also be applied to prepare a variety of amorphous/crystalline heterophase Rh-based bimetallic alloys (e.g. RhCu, RhZn and RhRu). Very recently, our group also employed formaldehyde to synthesize hierarchical Rh nanosheets with unconventional hexagonal close-packed (hcp) phase as well as novel ordered vacancies [68]. These highly voided, thermodynamically unfavorable 2D Rh nanostructures possessed a variety of vacated Barlow packings (VBPs) and orthorhombic symmetry, involving both perfect and vacated Rh nanosheets with diverse stacking sequences. In addition, Yin et al. rationally synthesized ultrathin hcp Ru triangle nanosheets from RuCl 3 in the aqueous solution containing PVP and formaldehyde [53]. CO released from formaldehyde could selectively adsorb on the exposed (0001) planes of hcp Ru with low surface energies, resulting in the formation of ultrathin Ru nanosheets. Our group synthesized multilayered RuNi nanosheets with hcp phase (Fig. 3e) using benzyl alcohol as the solvent and formaldehyde as the reduction agent [64]. Cheng et al. reported a facile solvothermal method to synthesize the partially hydroxylated fcc Ir nanosheets with a thickness of 5-6 atomic layers (Fig. 3f) in the presence of iridium chloride, PVP, N-methylpyrrolidone and formic acid [55]. Mechanism study suggested that the formic acid played a crucial role in the formation of the sheet-like morphology, since CO released from formic acid could selectively bond to the (111) facets of Ir as a surface-confining agent.
Besides, the metal carbonyl precursors can not only serve as CO-releasing agents, but also act as metal precursors of desired alloys [60,62,65].

For example, Luo et al. prepared ultrathin fcc
PdMo nanosheets with a sub-nanometer thickness of 0.88 nm (Fig. 3g) using Pd(acac) 2 and Mo(CO) 6 as metal precursors [60]. The study on the growth mechanism of PdMo nanosheets suggested that the small Pd nanosheets were first formed with the assistance of CO decomposed from Mo(CO) 6 , and then acted as seeds for the lateral growth and diffusion of Mo atoms. This approach could also be applied to synthesize PdW nanosheets after replacing Mo(CO) 6 with W(CO) 6 . Moreover, serving as both metal precursors and CO-releasing agents, triruthenium dodecacarbonyl (Ru 3 (CO) 12 ) [65] and tetrairidium dodecacarbonyl (Ir 4 (CO) 12 ) [62] were used to synthesize the triangular hcp RuRh nanosheets with a thickness of 1.9 ± 0.5 nm (Fig. 3h) and ∼1 nmthick fcc PdIr nanosheets, respectively, via a similar strategy.
Besides CO, oxygen (O 2 ) could also act as the surface-confined agent to assist the formation of 2D metal nanostructures. For instance, Luo et al. developed a facile wet-chemical approach to preparing fcc Al nanosheets via reducing aluminum chloride in a non-protonic mesitylene solution under the atmosphere of N 2 and O 2 (Fig. 3i) [69]. Through tuning the concentration of O 2 from 10 to 80 vol%, Al nanosheets with lateral sizes of 200-600 nm were obtained. The selective adsorption of O 2 on the (111) facets of fcc Al was the key factor when forming such a 2D nanostructure. Impressively, the thickness of Al nanosheets could be finely tuned from 18 nm to sub-2 nm via increasing the ratio of O 2 in the atmosphere of N 2 and O 2 .

Templated synthesis method
Recently, various kinds of 2D materials, such as graphene oxide (GO), TMDs and hydroxides, have been applied as growth templates for the synthesis of 2D metal nanomaterials [21,[70][71][72][73]. For instance, by using GO as a template, our group demonstrated the preparation of ultrathin 2D Au nanosheets with an unconventional 2H phase (Fig. 4a) [21]. Owing to the rich functional groups on GO, Au 3+ could be adsorbed on the GO surface, and then mildly reduced into Au seeds by oleylamine. The Au seeds eventually assembled into ultrathin 2D square-like nanosheets with a thickness of ∼2.4 nm. Impressively, this is the first time Au nanosheets have been directly prepared with pure 2H phase through a wet-chemical route [74,75]. It is worth mentioning that the Au square-like nanoplates with an alternating 2H and fcc crystal phase could be obtained using the aforementioned 2H Au   nanosheets as templates via secondary growth of Au [76]. Moreover, TMDs [77], one of the emerging 2D materials, can also serve as growth templates for a variety of noble metals. For example, our group has reported the epitaxial growth of Ag nanoplates on electrochemically exfoliated MoS 2 nanosheets [78], in aqueous solution (Fig. 4b)

Space-confined method
Layered materials with interlayered 2D space can be used as space-confined templates to slow down the diffusion of metal ions and constrain the growth of metals into 2D morphology. For example, Wang et al. prepared ultrathin fcc Au nanosheets between the interlayers in the Mg-Al-LDHs. The Mg-Al-LDHs could not only effectively confine the crystal growth of 2D Au, but also stabilize the obtained Au nanosheets [79]. Moreover, Jiang et al. employed the layered crystalline clay mineral montmorillonite (MMT) as a template to prepare fcc single-atom-layer (SAL) Pd nanosheets (Fig. 4c) in the angstrom-sized interlayer space of MMT [80]. Impressively, freestanding Pd SAL was successfully prepared after the removal of MMT by acid etching.
Besides the aforementioned hard templates, the interfaces of liquid/gas and liquid/liquid phases can also serve as soft templates to confine the growth of 2D metal nanomaterials. For instance, Lee et al. demonstrated a redox-active peptide templateassisted method to synthesize multilayered single-crystalline fcc Ag nanosheets. In their reaction, a tyrosine-rich α-helical peptide was used to reduce the Ag ions at the water/air interface, resulting in the formation of single-crystalline Ag nanosheets [81]. Yue et al. utilized the dodecenylsuccinic acid bilayers as the soft template for the controlled growth of {111}-oriented fcc Au nanosheets (Fig. 4d) at the liquid/liquid interface [82]. Zhu et al. developed a 1-hydroxyethylidene-1, 1-diphosphonic-acid (HEDP)-mediated method to prepare hierarchical porous Rh nanosheets at the interface between the upper-layer solution phase and the bottom-layer emulsion phase [83].

Seeded-growth method
The seeded-growth method represents a facile strategy to synthesize metal nanostructures with well-defined 2D morphology. It contains two steps, i.e. the pre-synthesis of seeds and the subsequent growth process. Generally, the lattice defects, e.g. twin boundaries, in a metal seed are more active for the heterogeneous nucleation, leading to preferential growth along the twin planes to form a sheet-like shape [22]. For instance, Khan et al. demonstrated the synthesis of large triangular Ag nanoplates from small Ag spherical seeds, and investigated the effect of poly(sodium 4-styrene-sulfonate) (PSSS) on regulating the shape of Ag [84]. The usage of PSSS generated the defective structures in Ag spherical seeds, which led to the anisotropic growth of Ag on the Ag seeds to form a 2D structure. Moreover, the kinetics control in the secondary growth of metal also plays an important role in directing the final shape of metal nanostructures when the defective metal nanocrystals are used as seeds. For example, Lin et al. employed defect-rich Au nanoparticles as seeds to direct the anisotropic growth of 2D Au nanoplates. The yield of Au nanoplates can be increased from 32.0% to 96.8% through tuning the molar ratio of PVP to HAuCl 4 during the growth process [85]. In addition, Zeng et al. elaborated that the triangular Ag seeds could grow into either thin Ag nanoplates in a lateral-growth-favorable mode or thick Ag nanoplates in a vertical-growth-favorable mode (Fig. 4e) by using two different capping agents, i.e. sodium citrate and PVP, respectively [22].

Other wet-chemical methods
Besides the aforementioned wet-chemical synthetic methods, other unique strategies have also been used to prepare 2D metal nanomaterials. For instance, our group utilized the ligand-exchange method to induce phase transformation from pre-synthesized 2H Au square sheets to fcc Au square sheets [86]. A similar crystal phase transformation from 4H Au nanoribbons to fcc Au nanoribbons was also observed [27]. Moreover, with the use of methyl orange, Ye et al. successfully synthesized the two-atomic-layer-thick 2D Au nanosheets with seaweed-like morphology (Fig. 4f), obtained via the self-assembly of small Au nanoflakes based on the oriented attachment mechanism [87]. Qiu et al. demonstrated a facile self-assembly approach for the synthesis of 2D porous fcc Pd nanosheets (Fig. 4g). Time-dependent experiments confirmed that the tiny Pd nanoparticles formed first, and gradually self-assembled into ultrathin one-dimensional nanowires with wavy orientation. Then, numerous ultrathin nanowires interweaved together and eventually formed porous Pd nanosheets [88]. Likewise, the oriented attachment assembly could also lead to the formation of Ru nanosheets [89]. Photochemical synthesis is another powerful strategy to prepare 2D metal nanostructures. For instance, the plasmon-driven synthesis of Au nanoprisms with a high yield of ∼90% has been achieved (Fig. 4h) through reducing HAuCl 4 in the presence of PVP and methanol under visible light irradiation [90]. PVP not only acted as the surfactant, but also served as a photochemical relay to direct the anisotropic growth of Au nanoprisms. Moreover, Kuang et al. utilized Ni(OH) 2 , ethylene glycol and NaOH to synthesize single-crystalline ultrathin fcc Ni nanosheets (Fig. 4i) via solvothermal method [91]. With a similar strategy, NiMo, NiFe and NiCo nanosheets were also prepared [92,93].

APPLICATION OF 2D METAL NANOMATERIALS IN ELECTROCATALYSIS
Metal nanomaterials are widely used as catalysts due to their superior activity and excellent stability in various electrochemical reactions. Recently, 2D metal nanomaterials have aroused tremendous research interest owing to their structural merits, including high surface-to-volume ratio, abundantly exposed surface active sites and unique electronic property. Here, as shown in Table 2, we summarize the recent progress in applying 2D metal nanomaterials as electrocatalysts in various electrochemical reactions, including water splitting (HER and OER), ORR, chemical fuel oxidation reactions (FAOR, MOR and EOR) and CO 2 RR.

Hydrogen evolution reaction
HER is a key half reaction for producing hydrogen via electrochemical water splitting. 2D noble metal nanomaterials, e.g. Pt [23], Pd [36], Ir [55], Rh   [67,68] and Ru [89], possess moderate hydrogen adsorption and desorption properties, thus showing superior catalytic performance in HER. For example, 2D Pt nanodendrites with a thickness of ∼2.3 nm exhibited extremely low overpotential (<10 mV) at a current density of 50 mA cm −2 in 0.5 M H 2 SO 4 , while the commercial Pt/C required 12.5 mV to achieve the same current density [23]. The enhanced performance of HER can be explained as follows. First, the dendritic morphology exposes more catalytic active sites in both corners and edges of the branches. Second, the single-crystalline nature leads to faster electron and mass transfer on 2D Pt nanodendrites. Third, the ultrathin 2D nanodendrite structure can not only increase the utilization of Pt atoms to improve HER activity, but also avoid the Ostwald ripening of the nanodendrites during electrocatalysis to ensure catalytic stability. Cheng et al. demonstrated that ultrathin Ir nanosheets outperformed Pt/C in both alkaline and acidic conditions due to their ultrathin 2D morphology and partially hydroxylated surfaces. Impressively, HER activity can be further improved by synthesizing 2D noble metal alloy nanomaterials with finely tuned electronic structures. For example, 2D IrRh [40] and RuRh [41] nanosheets are regarded as ideal electrocatalysts for HER.
Due to the scarce nature of noble metals, alloying 2D noble metal nanomaterials with non-noble metals is considered a cost-effective strategy [30,64,66]. Recently, our group prepared RuNi alloy nanostructures that were composed of multilayered nanosheets, achieving a low overpotential of 15 mV at 10 mA cm −2 (η 10 ) and a small Tafel slope of 28 mV dec −1 in 1.0 M potassium hydroxide (KOH). This surpasses commercial Pt/C and Ru/C [64]. The enhanced HER performance was ascribed to the large electrochemically active surface area (ECSA) arising from the 2D structure and the electronic effect from Ni alloying.
To reduce the use of noble metals, 2D non-noble metal alloy electrocatalysts have been prepared and used for HER [92,93]. For instance, in the alkaline condition, the ultrathin NiMo nanosheet displayed faster electron and mass transfer rate compared to the commercial Pt/C, due to its ultrathin 2D morphology and finely tuned alloy composition [92]. As is known, defects like vacancies can be more easily generated in 2D metal nanomaterials since the surface atoms are more likely to escape from the 2D lattice as the thickness decreases [11]. The hierarchical Rh nanosheets with ordered vacancies prepared recently exhibited a lowest η 10 of 37.8 mV compared with those of Rh/C (58.7 mV) and Pt/C (66.0 mV) in alkaline media. The enhanced HER performance of hierarchical Rh nanosheets was mainly attributed to the unique VBP structure and ultrathin 2D sheet-like morphology [68]. In addition, using 2D Pd nanosheets with different exposed facets as electrocatalysts, Xu et al. investigated the facet-dependent HER performance [36]. Owing to the optimal balance between adsorption and desorption of hydrogen on Pd(100) planes, Pd nanosheets exposed with the (100) planes showed a lowest η 10 of only 67 mV in acidic electrolyte, compared to those with exposed (110) planes (158 mV) and (111) planes (227 mV) (Fig. 5a). Moreover, the performance of Pd nanosheets with (100) planes only exhibited a slight degradation over 20 h. This superior HER durability could be attributed to the anisotropic ultrathin 2D structure, which greatly inhibited the dissolution and ripening process of Pd nanosheets.
2D noble metal nanomaterials can exhibit remarkable HER performance through the selective exposure of specific facets or control of their thickness. In addition, alloying with non-noble metals is a cost-effective strategy for 2D noble metal nanomaterials used for HER.

Oxygen evolution reaction
As the anodic half reaction of overall water splitting, OER plays a significant role in proton exchange membrane water electrolyzers. However, the sluggish kinetics of the multistep proton-coupled electron transfer process in OER has seriously impeded the overall efficiency of water splitting [9]. To date, Ir-based materials have been regarded as the most promising electrocatalysts towards OER due to their superior catalytic property [54,55,94]. Recently, the preparation of 2D Ir-based electrocatalysts for OERs has been extensively reported. For instance, the ultrathin mesoporous Ir nanosheets synthesized by Jiang et al. displayed superior OER activity in 0.5 M H 2 SO 4 , achieving η 10 of 240 mV, which was smaller than the non-porous Ir bulk (260 mV), commercial Ir black (268 mV) and commercial IrO 2 (360 mV) (Fig. 5b) [54]. The outstanding OER performance of the mesoporous Ir nanosheets could be ascribed to the unique electronic states of 2D structures as well as the large exposed surface area and rich active sites on the mesoporous structure. In addition, the partially hydroxylated ultrathin Ir nanosheets also exhibited improved OER activity in a wide pH range, achieving η 10 of 328 and 266 mV in 0.5 M H 2 SO 4 and 1 M KOH, respectively, which were both lower than those of the commercial IrO 2 [55]. Impressively, Ir-based bimetallic alloy (IrRh) nanosheets could also be considered as catalysts for OER in pH-universal electrolytes, and further applied to overall water splitting [40]. Additionally, other 2D bimetallic alloys, for instance, the channelrich RuCu nanosheets, reached small η 10 of 234, 276, 236 and 240 mV in 1 M KOH, 0.1 M KOH, 0.5 M H 2 SO 4 and 0.05 M H 2 SO 4 , respectively, exhibiting excellent OER catalytic activity [30]. For the overall water splitting, the RuCu nanosheets with optimized electronic structures displayed pH-universal water-splitting performance, achieving η 10 at a cell voltage of 1. State-of-the-art OER catalysts are mainly based on noble metal catalysts. 2D noble metal catalysts can exhibit high utilization of noble metal atoms and superior intrinsic activity due to their 2D structure, thus representing a promising candidate for future high-performance OER catalysts. (h) FEs versus applied potentials for CO 2 electroreduction on defected Cu nanosheets. Reproduced with permission from ref. [108]. Copyright 2020 American Chemical Society. (i) Faradaic efficiency for CO (FE CO ) of Au nanoribbons, 4H/fcc nanorods and fcc nanorods for CO 2 RR. Reproduced with permission from ref. [109]. Copyright 2020 American Chemical Society.

Oxygen reduction reaction
ORR is the cathodic half reaction operated in fuel cells. However, the commercial application of proton exchange membrane fuel cells has been severely hindered by the sluggish kinetics of current ORR catalysts [7,9]. Although the state-of-the-art Pt-based catalysts are still the most commonly used ORR catalysts owing to their high efficiency, the extensive application of Pt-based catalysts generally suffers from the high price of Pt [95]. Therefore, preparation of Pt-based catalysts with 2D morphology is an efficient strategy for increasing the atom utilization of Pt, and thus improving ORR activity.
Owing to the high specific surface area enlarged by the porous structure and the abundant active sites offered by numerous defects, the ultrathin porous Pt nanosheets synthesized by Feng et al. displayed excellent ORR performance with specific activity of 3.1 mA cm −2 and mass activity of 2.07 A mg Pt −1 at 0.90 V versus the reversible hydrogen electrode (vs. RHE) in 0.1 M HClO 4 , which were 10.7 and 9.8 times those of the commercial Pt/C, respectively [96]. Importantly, the synthesized ultrathin porous Pt nanosheets also displayed superb durability, enduring 30 000 cycles of accelerated durability testing without obvious activity decay. Moreover, the same group further prepared hexagonal PtBi nanoplates as efficient ORR catalysts, showing extraordinary tolerance to CH 3 OH, HCOOH and CO [56]. The PtBi nanoplates displayed a high mass activity of 1.04 A mg Pt −1 (vs. RHE) in 0.1 M HClO 4 , which was higher than those of the recently reported non-2D PtBi nanomaterials. Besides, the 2D Pd-based catalysts are considered as alternatives to Pt [48,60,80] [60]. Due to the large ECSA (138.7 m 2 g Pd −1 ) of the ultrathin 2D structure, the PdMo nanosheets manifested an outstanding ORR mass activity of 16.37 A mg Pt −1 at 0.9 V (vs. RHE) in 0.1 M KOH (Fig. 5c).
Impressively, the ORR performance of 2D metal electrocatalysts can be finely tuned by changing their thickness. For example, Wang et al. demonstrated the thickness-dependent ORR performance of Pd nanosheets [48]. Evidenced by the computational and experimental results, Pd nanosheets with a thickness of 5 ± 1 monolayers exhibited better ORR performance compared to Pd nanosheets with 3 ± 1 monolayers and 8 ± 1 monolayers in both acidic and alkaline media (Fig. 5d). The good performance of Pd nanosheets with 5 ± 1 monolayers originated from the optimal surface strain, allowing suitable binding with reaction intermediates on the catalyst surface.
2D noble metal electrocatalysts show intriguing ORR activity, which is related to their thickness. 2D metal nanomaterials with single-or few-atomic layers expose rich under-coordinated atoms at the surface, which have been recognized as highly active centers for ORR. Next generation 2D noble metalbased catalysts should focus on the fine control of their thickness, as well as the improvement of their stability.

Chemical fuel oxidation reaction (FAOR, MOR and EOR)
Organic acids and alcohols, such as formic acid, methanol and ethanol, are commonly used chemical fuels in direct liquid fuel cells [2,7]. 2D metal nanomaterials have been regarded as excellent electrocatalysts for the oxidation reaction of these chemical fuels, owing to their large exposed surface area and abundant active sites. Notably, 2D Pd-based nanostructures are outstanding electrocatalysts for FAOR due to their high catalytic activity and excellent anti-poisoning capacities [24,62,97,98]. Recently, Qiu et al. prepared the self-assembled porous Pd nanosheets as efficient and stable electrocatalysts for FAOR [88]. Since the ultrathin knit-like 2D morphology and porous structure provided rich active sites and a high specific surface area, the Pd nanosheets showed higher catalytic activity, superior durability and better anti-poisoning capability compared to the commercial Pd black catalyst. Our group reported that the FAOR activity could be significantly enhanced by alloying Pd with Cu [63]. The ultrathin PdCu nanosheets showed a superior mass activity of 1655.7 ± 74.6 mA mg Pd −1 in an acidic electrolyte, due to the ultrathin 2D morphology, optimized electronic structure, synergistic effect between Pd and Cu, and post-treatment with ethylenediamine (EN) (Fig. 5e).
Besides FAOR, 2D metal nanomaterials, e.g. Pd [34,58,61,99] and Rh [83], have been considered as promising Pt-alternative electrocatalysts for MOR and EOR owing to their excellent catalytic activity and stability. For instance, Zhu et al. reported hierarchical porous Rh nanosheets as efficient electrocatalysts for MOR [83]. The rich grain boundaries and ultrathin porous structure endowed the hierarchical porous Rh nanosheets with a mass activity of 333 A g −1 at 0.63 V (vs. RHE) in alkaline electrolyte, much higher than that of the commercial Pt black catalysts (57.4 A g −1 ). Ge et al. also used ultrathin Pd nanomeshes as an electrocatalyst for EOR. These exhibited high mass activity of 5.40 A mg Pd −1 and specific activity of 7.09 mA cm −2 in alkaline media due to their unique mesoporous structure and large surface area [49]. Additionally, 2D Pd-based bimetallic alloy nanostructures have also been demonstrated to be good electrocatalysts for MOR and EOR, e.g. PtPb [99] and PdCu [61], used for MOR, and PdAg [34] and PdZn [58] Fig. 5f) [58]. The enhanced EOR activity of PdZn nanosheets can be explained as follows. First, the 2D morphology of PdZn nanosheets provided a large surface area and rich uncoordinated active sites for the electrochemical EOR. Second, the introduction of Zn into Pd tuned the electronic structure of Pd, thereby improving the EOR activity. Third, lattice defects (e.g. dislocations and grain boundaries) in polycrystalline structures also altered the electronic structure, and offered extra active sites for EOR [100].
Nowadays, alloying 2D noble metal nanomaterials with other non-noble metals is a promising strategy to enrich the library of cost-effective 2D bior multi-metallic alloy electrocatalysts for FAOR, MOR and EOR. The introduction of other metals could optimize the electronic structures of 2D metal electrocatalysts, and the synergistic effect between multiple elements could further improve catalytic performance.

CO 2 reduction reaction
CO 2 RR is one of the most important strategies for solving the global energy crisis caused by excessive consumption of fossil fuels, as it converts CO 2 into high value-added products, e.g. CO, methane, methanol, formate, ethylene and ethanol [3]. 2D noble metal electrocatalysts exhibit promising application to CO 2 RR owing to their excellent selectivity, catalytic activity and stability [25]. For example, Liu et al. demonstrated that triangular Ag nanoplates could selectively reduce CO 2 to CO in 0.1 M KHCO 3 , and possessed an enhanced Faradaic efficiency (FE) of 96.8% and excellent durability of 7 days [101]. Until now, 2D non-noble metal-based electrocatalysts for CO 2 RR, such as Zn [102,103], Bi [104][105][106][107], Co [29] and Cu [108], have attracted increasing research interest because of their cost-effectiveness compared to noble metals. For instance, Liu et al. used porous Zn nanosheets as the CO 2 RR electrocatalyst to selectively produce CO. The porous Zn nanosheets exhibited a high FE of 90%, large partial current density and outstanding durability for over 24 h in 0.1 M KHCO 3 , benefiting from 2D morphology and porous architecture with increased exposure and active sites [102]. In addition, Bi-based nanosheets have considerable potential to produce formate [104][105][106][107]. Owing to the enlarged surface area and rich under-coordinated Bi sites, the ultrathin Bi nanosheets exhibited a large catalytic current density (24 mA cm −2 at −1.74 V), excellent formate selectivity (FE of >90% over a broad potential range) and superior durability (>10 h) in 0.5 M NaHCO 3 , unveiling great potential with regard to electrocatalytic conversion from CO 2 to formate [106]. Yang et al. studied the thickness-dependent CO 2 RR performance of Bi nanosheets (Fig. 5g). The monolayered Bi nanosheets with an average thickness of 0.65 nm exhibited higher FE of ∼99% and lower onset overpotential of <90 mV compared to those with a thickness of 4.2 nm and 11.3 nm, respectively. Moreover, during the CO 2 RR, the 2D atomic Co layers could also selectively and efficiently produce formate [29]. Interestingly, Cu-based nanomaterials have attracted tremendous attention due to their unique ability to convert CO 2 or CO into high valueadded multicarbon (C 2+ ) products, paving the way for the sustainable production of fuels and chemicals. Zhang et al. demonstrated the nanodefective Cu nanosheets for the electrochemical CO 2 reduction, which exhibited an ethylene FE of 83.2% with a current density of ∼60 mA cm −2 (Fig. 5h) [108]. Remarkably, 2D metal nanostructures with unconventional crystal phases may display enhanced CO 2 RR performance, since the combination of unconventional crystal phase and 2D morphology can result in the preferential exposure of under-coordinated sites. Au nanoribbons with the unconventional 4H phase have achieved >90% FE toward CO at −0.7 V (vs. RHE) in 0.1 M KHCO 3 , exhibiting better activity and selectivity for CO production, compared with the 4H/fcc nanorods and conventional fcc nanorods (Fig. 5i) [109]. Confirmed by theoretical calculations, the relatively high CO 2 RR activity of 4H nanoribbons could be attributed to the highly active and abundant 7-and 8-fold under-coordinated sites on the specific surfaces of 4H nanoribbons, i.e. the (1120) 4H and (1100) 4H ridge sites.
2D metal nanomaterials exposing well-defined specific facets favor the selective production of desired products such as CO, formate, and importantly, the C 2+ products. Notably, 2D Cu nanomaterials with unconventional crystal phase or phase boundaries are more favorable for C-C coupling due to the proper * CO adsorption energy, yielding a series of C 2+ products, such as ethylene, ethanol and acetate.

CONCLUSION AND PERSPECTIVES
With the rapid development of synthetic techniques of nanomaterials, recent decades have witnessed remarkable progress in the preparation of 2D metal nanomaterials with unique physicochemical properties. 2D metal nanomaterials with structural features like shape, thickness, size, composition and crystal phase have been successfully synthesized via various wet-chemical synthetic methods. Wetchemical synthesis can achieve massive production of 2D metal electrocatalysts with finely controlled structures and abundant surface function groups, and could enhance the performance of electrocatalysis. To date, a large quantity of research work has been devoted to preparing 2D noble metal catalysts for various electrocatalytic reactions due to their superior intrinsic catalytic activity. Recently, alloying non-noble metals with noble metals and preparing 2D non-noble metals have attracted ever-growing research interest, as these strategies can not only reduce the catalyst cost, but also boost catalytic performance. Benefiting from the large exposed surface area and abundant active sites, the ultrathin 2D metal nanomaterials exhibit superior electrocatalytic performance in various reactions, including HER, OER, ORR, FAOR, MOR, EOR and CO 2 RR. Although tremendous efforts have been devoted to the preparation of 2D metal electrocatalysts, many challenges still remain. Based on the current research progress, some challenges and potential research directions are proposed in order to inspire more fascinating research work in the near future.
(i) Investigating the formation mechanisms of 2D metal nanomaterials by in situ characterizations.
Despite the recent great progress made in the wet-chemical synthesis of numerous 2D metal nanomaterials, the formation mechanisms of these unique 2D metal nanostructures still need to be fully understood. Nowadays, most studies on the formation of 2D metal nanostructures are based on ex situ characterization techniques, such as TEM and X-ray diffraction (XRD), which are used to characterize the intermediate at a certain reaction time interval. Thanks to the rapid development of advanced in situ characterization techniques, by using the in situ TEM, Raman spectrum, XRD and synchrotron radiation spectrum, the formation mechanisms could be unraveled. For example, in situ synchrotron small-angle X-ray scattering (in situ-SAXS) can be used to reveal the formation kinetics, i.e. nucleation and subsequent growth of 2D metal nanomaterials through observing the emergence of nuclei, growth of nanocrystals and reconstruction of 2D morphology. Moreover, in situ X-ray absorption fine structure (in situ-XAFS) can provide additional information on the structural evolution of 2D metal nanomaterials, including the coordination configuration and oxidation state during their nucleation and subsequent growth processes. Importantly, by combining with theoretical calculations and simulations, a comprehensive understanding on the formation mechanisms of 2D metal nanomaterials could be realized. (ii) Enriching the library of 2D metal nanomaterials. Until now, in the periodic table of elements, less than half of the metals with 2D features have been reported. Therefore, there is still plenty of room to enrich the library of 2D metal nanomaterials. In particular, it is urgently desired to develop general synthetic methods to prepare 2D multi-metallic alloy nanostructures with tunable compositions, e.g. 2D high-entropy alloys (HEAs). Benefiting from the synergistic effect among multiple elements and multi-active sites of HEAs, the 2D HEA nanostructures may exhibit outstanding electrocatalytic performance. Therefore, through rationally tuning the elemental compositions of metal alloys, the library of 2D metal alloy nanomaterials will be greatly enriched. Impressively, high-throughput theoretical calculations, guided by rising artificial intelligence (AI) technology, e.g. machine learning algorithms, can be used to predict novel 2D functional metal and alloy nanomaterials with highly efficient electrocatalytic performance, which could be synthesized in the near future. (iii) Finely tailoring the structures of 2D metal nanomaterials. In order to develop electrocatalysts with enhanced performance, rational design and synthesis of 2D metal nanomaterials with finely tuned structural parameters, such as thickness, exposed specific facet, defect and crystal phase, are imperative. The first research direction is to finely tune the thickness of various 2D metal nanomaterials. For example, the controlled synthesis of 2D metal nanosheets with sub-nanometer thickness or even single layer is very intriguing yet challenging. The unique coordination environment in the aforementioned 2D metal nanosheets can efficiently affect their electronic structures, resulting in the enhancement of their electrocatalytic properties. Second, facet engineering has roused extensive interest in recent years. Therefore, selectively synthesizing 2D metal nanomaterials with specific exposed facets is an attractive research direction. Until now, through wetchemical synthesis, 2D metal nanomaterials only exposed limited facets, e.g. {111} planes for fcc Rh and {0001} planes for hcp Ru. Hence, the synthesis of 2D metal nanostructures with other facets, especially high-index ones, is another research topic. The exposed high-index facets possess a high density of low-coordinated atoms, edges, steps and kinks, serving as active sites for electrocatalysis, and thus showing enhanced electrocatalytic performances compared with the conventional low-index facets such as {111} and {100} planes. Third, defect engineering has also been recognized as an effective approach to designing efficient electrocatalysts. Construction of abundant defects, such as dislocations, boundaries and vacancies, in the 2D metal nanomaterials can tune the electronic structures and create more active sites, thereby enhancing the electrocatalytic performance. Last but not least, phase engineering of nanomaterials (PEN) opens a new route towards the synthesis of 2D metal nanostructures with unconventional phases, i.e. unconventional crystal phase, amorphous phase and heterophase nanomaterials for various promising applications [110]. To date, although a few 2D metal nanostructures with unconventional phases have been successfully synthesized, they are just the tip of the iceberg. A large number of unknown 2D metal nanostructures with unconventional phases still need to be explored.