Engineering Active Sites of 2D Materials for Active Hydrogen Evolution Reaction

Hydrogen evolution reaction (HER) is a promising clean and sustainable energy source with zero carbon emissions. Numerous studies have been conducted with versatile low dimensional materials, and the development of highly active electrochemical catalysts for HER is one of the most important applications of the materials in these studies. Despite such extensive research, the physical origin of the active catalytic performance of low dimensional materials remains unclear, and is distinguished from that of classical transition metal‐based catalysts. Here, recent studies on the intrinsic catalytic activity of 2D semimetals are reviewed, particularly among transition metal dichalcogenides (TMDs), highlighting promising strategies for the design of materials to further enhance their catalytic performance. One attractive approach for active HER involves fabricating single‐atom catalysts in the framework of TMDs. The electrochemical reaction at a catalytic atom for hydrogen evolution has typically been described by the Sabatier principle. Recent studies have focused on optimizing the Gibbs free energy for hydrogen adsorption via down‐sizing, alloying, hybridizing, hetero‐structuring, and phase boundary engineering, mostly with TMDs. The unique advantages of TMDs and their derivatives for HER are summarized, suggesting promising research directions for the design of low dimensional electrochemical catalysts for efficient HER and their energy applications.


Introduction
Global environmental issues such as climate change, global warming, and air pollution, which are related to the vast use of fossil fuels, are negatively impacting human life. [1,2] Clean, renewable, and sustainable energy sources such as solar, wind, geothermal, and biomass are increasingly in demand to address these global environmental problems, and to respond to rapidly growing energy consumption. Among the various alternatives to fossil fuels, hydrogen is considered an ideal energy source because of its high energy density and carbon-free final product. [3][4][5][6] Hydrogen can be categorized into three different types, grey, blue, and green hydrogen, depending on the hydrogen production technology. When water electrolysis is performed using sustainable energy sources, it can produce green hydrogen without emitting carbon by-products. In this way, with a highly efficient electrocatalyst, water electrolysis with highly efficient electrocatalyst can be a promising zero-carbon technology for economic hydrogen production. [7] Hydrogen evolution reaction (HER) in 2D materials has attracted significant attention because of their extraordinary electrocatalytic activity, which has been attributed to their highly activated surface area and exotic physical and chemical phenomena. [8][9][10][11][12] Following the discovery of graphene and its derivatives (e.g., graphene oxides), remarkable scientific achievements have been reported, and advances in nanoscience have recently introduced quantum phenomena to the field of electrocatalysis. Previously, the HER performance of catalysts has mainly been understood using the Sabatier principle, which intuitively states that the bonding strength between hydrogen atoms and active atomic sites determines HER activity. [13] The macroscopic explanation describes various molecular motions and carrier transport under electric fields using an equivalent circuit with resistance, capacitance, and impedance.
More recently, studies of electrochemical catalytic activity in 2D materials have focused on individual atomic sites. They have revealed that atoms at the edges or atomic defects in semiconducting transition metal dichalcogenides (TMDs), such as MoS 2 , MoSe 2 , and WS 2 , exhibit efficient catalytic activity. [14][15][16] Theoretical calculations based on the Sabatier principle predict that atoms at the edge or atomic defects in semiconducting TMDs would  [74] Pt-Ru dimer on nitrogen doped carbon nanotube 50 28.9 0.5 m H 2 SO 4 [36] Single atom Pt on defective graphene matrix C x N y 41 40 1 m KOH [38] Rh nanoparticles on boron (B) nanosheet 66 [103] have almost zero Gibbs free energy for hydrogen atom adsorption, which is a prerequisite for ideal and reversible hydrogen adsorption and desorption reactions. [13] However, it has been difficult to develop semiconducting TMDs for industrial applications because the number of active sites in 2D materials is limited. Active HER has also been observed in the basal plane of metallic 2D materials. In particular, 2D metallic TMDs (e.g., MoTe 2 ) have shown extraordinary HER performance, with a high and stable current density and conversion rate at atomic sites on their basal plane, which are not easily explained by the Sabatier principle and Volcano plot. [17][18][19] Several experimental and theoretical studies have reported unprecedented HER resulting from quantum phases in the metallic 2D materials, which can affect the physics and chemistry at the interface between the catalyst and electrolyte during HER, opening a new era of HER with quantum phenomena.
Here, we review recent HER studies based on 2D electrocatalysts, and strategies to design efficient HER catalysts with 2D semimetals. In chapter II, the catalytic activities of transition metal-based catalysts are introduced and explained using the traditional Sabatier principle. The strategies to obtain 2D elec-trocatalysts with excellent HER performances are discussed in chapters III and VII. In chapter III, we discuss quantum size effects in sub-nanometer-scaled catalysts, covering the recent growing interest in single-atom catalysts. We review basal plane activities in metallic 2D materials, especially TMDs, and their intriguing surface states resulting from low dimensionality, from the viewpoint of surface catalytic reactions, in chapter IV. Several strategies to further enhance intrinsic catalytic activity are presented in chapters V, VI, VII, and VIII: These include down-sizing, alloying, hybridizing, hetero-structuring, and phase boundary engineering with TMDs for emerging 2D electrocatalysts for HER (Table 1). Finally, we suggest material designs based on TMDs for green hydrogen production via efficient HER process.

Principle of the Hydrogen Evolution Reaction
HER is a cathodic reduction process of water electrolysis, which produces hydrogen gas molecules (H 2 ) by receiving electrons with the overall reactions of 2H + + 2e − → H 2 in acid solution and 2H 2 O + 2e − → H 2 + 2OH − in alkaline solution. Unlike In the Volmer step, the protons (H + ) are bound to the catalytic atom and after receiving electrons are converted to an adsorbed hydrogen atom (H*, purple spheres), that is ready to produce hydrogen gas molecules via the Heyrovsky and Tafel step. b) Volcano plot with Gibbs free energies of a hydrogen atom (ΔG H* ) at the surface of noble metal catalysts. Reproduced with permission. [22] Copyright 2005, Electrochemical Society. c) Schematics of bonding ( ) and antibonding ( *) states via orbital hybridization between a metal atom and hydrogen atom. This shows the strong relationship between ΔG H* and the antibonding state E * as well as the charge transfer. Reproduced with permission. [31] Copyright 2019, Wiley. d) The Gibbs free energy of hydrogen adsorption (ΔG H* ) varies with d-band center of d-metal diborides, MB 2 . Inset shows that the center of the d-bands was downshifted when the d-metal atoms were chemically bonded with boron atoms. Reproduced with permission. [32] Copyright 2018, Wiley.
the multiple-step processes of other catalytic reactions, HER has been understood to involve three simple steps, as described in Figure 1a. In acid solution, a hydrogen ion (H + ) adsorbs on the surface of the catalyst, H + + e − → H* (Volmer step). The adsorbed hydrogen atom (H*) reacts with an approaching hydrogen ion (H + ) and turns into a hydrogen molecule (H 2 ), H* + H + + e − → H 2 (Heyrovsky step), or two adsorbed hydrogen atoms react, H*+ H* → H 2 (Tafel step). In alkaline solution, a water molecule (H 2 O) is broken into a hydroxide ion (OH − ) and a hydrogen ion (H + ), and by receiving an electron, a hydrogen ion adsorbs onto the surface of the catalyst, H 2 O + e → OH − + H* (Volmer step). To produce hydrogen gas, the adsorbed hydrogen atom (H*) reacts with another hydrogen ion dissociated from another water molecule, H* + H 2 O + e → OH − + H 2 (Heyrovsky step), or it reacts with another adsorbed hydrogen atom, H*+ H* → H 2 (Tafel step). [20,21] The Volmer step has been considered the rate determining step in most HER using noble metal catalysts and their derivatives. Hydrogen ions must be adsorbed at catalytic atoms to receive electrons to be converted to hydrogen gas, and then des-orbed from the catalytic atoms. Accordingly, the adsorption and desorption processes of hydrogen atoms at the surface of the catalyst must undergo a reversible reaction with zero Gibbs free energy. Figure 1b shows a graph of Gibbs free energies (ΔG H* ) plotted as a function of exchange current density (j 0 ), which is known as a Volcano plot. [22] The volcano plot has two distinct regions; in the left part (ΔG H* < 0), the adsorption of hydrogen ions on the catalyst is unstable due to the weak bond strength, and thus, electron transfer from a catalytic atom to hydrogen ions is difficult. Increasing ΔG H* toward the left further decreases the reaction rate. On the other hand, in the right part (ΔG H > 0), the adsorption of hydrogen ions on the catalyst is strong and maintained for a long time. In this case, preparing vacant sites on the catalytic atoms for upcoming hydrogen ions is not efficient, leading to a decrease in reaction rate. Many theoretical calculations have reported various values of the Gibbs free energy of hydrogen adsorption at the surface of metal catalysts. To date, the state-ofthe-art catalysts for HER with almost zero ∆G H* are mainly composed of noble metal atoms, which are located at the apex of the volcano plot. www.advancedsciencenews.com

www.advphysicsres.com
In many theoretical studies, the Gibbs free energy (ΔG H* ) is calculated from Equation (1) below.
where T is temperature, and ΔE H* , ΔE ZPE , and ΔS H* represents the energy, zero-point energy and entropy differences of the adsorbed hydrogen on a catalytic atom and the hydrogen molecule in gas phase, respectively. In most cases, the catalytic materials undergo negligible entropy changes when hydrogen atoms are adsorbed on the catalytic surface. Therefore, assuming the entropy change (ΔS H* ) is similar to −½ ΔS H2 , the term of TΔS H* in Equation (1) is around −0.20 eV at room temperature. Thus, the ΔG H* can be estimated from the ΔE H* . The bonding strength between a catalytic atom and hydrogen atom is strongly related to catalytic activity. [23][24][25] On the other hand, catalytic materials with a large entropy change (ΔS H* ) during the HER process do not follow the conventional HER trend. [26] The bonding strength at the catalytic surface is closely related to the electronic structure close to the Fermi level. It has been reported that the bond strength between a hydrogen atom and a catalytic atom varies with the work function (Φ) of the metal catalyst. [27,28] In 3d and 4d transition metal catalysts, the exchange current density increases with the number of unpaired d electrons or the energy level of the d-band center. [29,30] Figure 1c shows that the antibonding state (E * ) determines the catalytic activity for HER by optimizing the charge transfer process. [31] Li et al. reported that the HER activity of metal diborides increased as the center position of the d-band becomes farther away from the Fermi energy. This broadened d-band helps enhance HER performance by adjusting the bonding strength, setting the ΔG H* close to zero (see Figure 1d). [32] The downshift of the d-band center makes the bonding between a catalytic metal atom and hydrogen atom moderately strong for efficient HER.
Metals and metal compounds with delocalized d-orbitals close to the Fermi level have been considered ideal catalysts that adsorb and desorb hydrogen atoms reversibly with near zero ΔG H* during the HER process. To date, this ability to adsorb and desorb hydrogen ions has been explained mainly by the characteristics of the d-orbital electrons of the metal atoms in the catalytic materials. In addition, in compound catalysts such as metal borides, carbides, and chalcogenides, the characteristics of the d-orbital electrons have been modified by hybridized s-p orbital electrons from non-metal anions. [33] Many metal boride, carbides, and chalcogenides are known to be low-dimensional materials with fewatom-thick layers, which provide an extremely large active area for HER. We will discuss the size effects of the catalytic materials for active HER from a morphological view in chapter III.

Sub-Nanometer-Scaled Catalysts
In chapter II, we discussed metal catalysts in terms of bonding strength for hydrogen adsorption on the catalytic surface. Although the ability to capture hydrogen atoms (ions) is important for electrochemical reactions at the catalytic surface, the morphology of the catalyst, which can maximize the catalytic surface is critical to achieve high catalytic performance. Reducing the dimension of the catalysts down to single atom scale has advan-tages, including extremely small mass loading, a high density of surface atoms per unit area, and greater opportunities to react hydrogen atoms with unsaturated surface atoms.
Reducing surface energy and dispersing catalytic atoms on a conductive support to make nano-structured catalytic materials has been challenging. As shown in Figure 2a, even though the surface energy increases with the size of the catalysts (black lines), the interaction between catalytic and support atoms can be stabilized by the significant decrease in surface energy that occurs as we reduce the dimensions of the catalyst (red lines). [34] Since the interaction between catalytic atoms and the support is critical to stabilize the nanostructured catalysts, the physical and chemical properties of the support also need to be seriously considered. To overcome the difficulties to dispersing catalytic atoms on the support, catalytic atoms can be selectively bonded to certain atomic sites such as defects, vacancies, and substituted atoms. [35,14] Zhang et al. reported that a Pt-Ru dimer catalyst was prepared by adsorbing a Ru atom to a Pt atom, which was strongly bonded with a substituted N atom in carbon nanotubes. [36] The interaction between catalytic atoms and support atoms also affects catalytic performance, resulting in a trade-off relationship between the bonding strength and catalytic activity, as shown in Figure 2b. [37] The phase stability of these small-dimension catalysts and single atom catalysts is achieved by the strong bonding between the catalytic atoms and support atoms; however, their catalytic activities tend to be weaker with stronger bonding. In contrast, weak bonding between the catalytic atoms and support atoms results in the agglomeration of atoms into nanoparticles on the support.
Among various forms of Pt (e.g., single atoms, nanoparticles, and clusters) Pt single-atoms exhibit the largest active site density and highest mass activity, with almost zero ∆G H* . Figure 2c shows the atomic arrangements of Pt (i.e., bulk, clusters, and single atoms) on a nitrogen-doped defective carbon (C x N y ) matrix with corresponding ΔG H* estimated by density functional theory (DFT) calculations. Pt single atoms anchored on C x N y exhibit a moderate ΔG H* close to zero, which facilitates HER. [38] Recently, sub-nanometer-scaled metal clusters down to the atomic scale have been prepared with high coverage and uniform distribution. The atomic scale catalysts have shown extraordinary catalytic activity for many electrochemical reactions including HER. [39,40] For practical HER, the catalysts were prepared as nanoparticles and deposited on conductive 1D or 2D supports (see Figure 3a). [41][42][43] Such nanoparticulate catalysts have a high proportion of surface atoms, and conductive supports such as carbon nanotube or graphene oxide act as a pathway (i.e., electrode) for charge transport in HER. The enhanced catalytic activity of nanoparticles can be explained by the low coordination number of surface atoms, [44,45] the quantum size effect, [46,47] and metalsupport interaction. [10,48] Noble and transition metal nanoparticles have been successfully distributed on 1D or 2D matrices to achieve single atom catalysts with excellent HER performance. [24] Highly dispersed multi-component-nanoparticles on a porous and micro-structured template have shown excellent HER performance. [49,50] These dispersed catalytic atoms or clusters with high coverage and phase stability are highly active with almost zero ∆G H* by regulating the interaction between the Figure 2. a) Changes in surface energy and catalytic activity with decreasing metal particle size. The interaction between metal atoms and supports significantly decreases the surface energy. Reproduced with permission. [34] Copyright 2013, American Chemical Society. b) The competitive relationship between the catalytic activity and the phase stability of single atom catalysts. Reproduced with permission. [37] Copyright 2019, Elsevier. c) The adsorption arrangement and Gibbs free-energy barriers by DFT calculation for HER on Pt single atom, Pt cluster and Pt-C x N y (x + y = 4). The dark blue, gray, blue, red, and white spheres are Pt, C, N, O, and H atoms, respectively. Reproduced with permission. [38] Copyright 2021, CCS Chemistry. Synergistic effect of platinum single atoms and nanoclusters boosting electrocatalytic hydrogen evolution is available online at https://doi.org/10.31635/ccschem.020. 202000497. Copyright 2021, Chinese Chemical Society.
catalytic metal atoms and support atoms, modulating the bonding strength (E H* ) of hydrogen atoms at the catalytic atoms for ideal HER. Cobalt (Co) atoms are known to be inactive for HER. However, single Co atoms incorporated in a 2D MoS 2 support have served as highly active catalytic sites. The interaction between the Co atoms and the MoS 2 support is significant, resulting in a phase transition of the MoS 2 from a semiconducting 2H phase to a metallic 1T phase (see Figure 3b. [39] According to Zhu et al, catalytic activity increases with a larger density of Pt single atoms and decreases when the Pt atoms become aggregated, forming clusters (see Figure 3c). [51] The adsorption energy of hydrogen atoms on the surface strongly depends on the size of the Pt nanoparticles. [52] However, nanoparticles are more easily aggregated as their size increases (often referred to as Ostwald ripening), and therefore, methods to uniformly distribute catalytic nanoparticles are highly necessary for stable HER. [53,54] Jiang et al. reported that a single-atom-layer Pd-Co alloy film with absolute 2D geometry showed six times higher mass activity for electrochemical reaction than a Pt catalyst (see Figure 3d). [55] The strong 2D interaction of the Pd and Co in the single-atom-layer resulted from the splitting of d or-bitals into a lower-energy band in the x-y plane and a higherenergy band in the z-direction. This produced a 2D geometry with structural stability and catalytic activity, respectively. Freestanding, self-stabilized single-atom-layer catalysts can be an ideal catalyst with a large electrochemical surface area, high phase stability, and abundant dangling bonds in the out-of-plane direction for electrochemical reaction.
Even though nano-structured catalysts have shown excellent HER performances, it is difficult to obtain 2D catalysts, which are composed of uniformly distributed catalytic atoms (or clusters) on conductive 1D or 2D supports with high coverage and phase stability. Moreover, these sub-nano or single atom catalysts sometime reconstruct their phases when hydrogen atoms are adsorbed or desorbed. [52] In addition, the conductive 1D or 2D supports have considerable Joule heating when the exchange current increases due to numerous defect sites, which are generated by incorporating catalytic atoms into the supports. Therefore, allin-one 2D materials with catalytically active atoms on a conductive 2D geometry could be a promising candidate for hydrogen production. In chapter IV, we discuss metallic 2D stoichiometric compounds, which exhibit selective catalytically active atoms in  [41] Copyright 2019, Springer Nature. Middle figure) Reproduced with permission. [42] Copyright 2018, Springer Nature. Right figure) Reproduced with permission. [43] Copyright 2019, Wiley. b) Single atom catalysts of Co atom selectively decorated on distorted 1T-MoS 2 . Reproduced with permission. [39] Copyright 2019, Springer Nature. c) HAADF-STEM image of Pt nanoparticles distributed on VS 2 nanosheet and its particle size statistics. Reproduced with permission. [51] Copyright 2020, American Chemical Society. d) Splitting d orbitals upon 2D strong interaction in a Pd-Co single atom layered structure with unique 2D-oriented coordination. Reproduced with permission. [55] Copyright 2020, Elsevier.
their basal planes with outstanding transport behaviors, leading to all-in-one catalysts for robust HER.

Semimetallic Transition Metal Dichalcogenide Catalysts
TMDs have several thermodynamically stable crystal structures with a variety of electronic phases ranging from insulator to metal. Figure 4a shows representative crystal structures of TMDs, a hexagonal (2H) and a distorted octahedral (1T') structure with a chemical composition MX 2 where M and X represent a transition metal and chalcogen atoms, respectively. Their electronic properties are determined by transition metal and chalcogen atoms. For examples, most group 6 2H-TMDs (MoS 2 , MoSe 2 , MoTe 2 , WS 2 , and WSe 2 ) are semiconductors while group 6 1T' TMDs (WTe 2 and MoTe 2 ) are semimetals. [56,57] Recently, semimetallic The " " site in Td (or 1T') TMDs exhibit the best value of ΔG H* among other Te atoms and the " " site can be an active site for hydrogen conversion, resulting in basal plane HER activity in metallic TMDs Reproduced with permission. [59] Copyright 2020, IOP Science. c) The basal planes activity of three metallic TMDs and their HER performances. Reproduced with permission. [60] Copyright 2020, MDPI. d) Comparison of catalytic activities taken from three different metallic TMDs. Reproduced with permission. [60] Copyright 2020, MDPI.
TMDs have received much attention in condensed matter physics because of their intriguing quantum phases, including a quantum spin Hall insulator, topological superconductor, and Weyl semimetal.
Some of these semimetallic TMDs have shown intrinsic HER activities on their basal planes, raising interest in HER with quan-tum phenomena in 2D materials. [58] Seok et al. reported that certain Te atoms in the basal plane of TMDs had lower ΔG H* than others and the Te atoms behaved as active sites for hydrogen evolution (Figure 4b). This indicates that the adsorbed hydrogen atoms at these active Te atoms are more likely to be converted to hydrogen gas molecules. [59] Unlike nanoparticulate www.advancedsciencenews.com www.advphysicsres.com noble metal catalysts deposited on a conducting support or template, semimetallic TMDs naturally possess conductive basal planes with selective atomic sites for hydrogen adsorption. Therefore, semimetallic TMDs exhibit not only a uniform distribution of catalytic active sites, but also excellent carrier transport. From this viewpoint, semimetallic 2D TMDs in the form of active sites on a conductive support are promising candidates for HER catalysts, due to HER on their basal planes.
To study the HER mechanism in basal plane-active semimetallic TMDs, the surfaces of single crystalline TMDs with various compositions were used. In Figure 4c, Kwon et al. compared the HER activities of three TMDs semimetals (VTe 2 , WTe 2 , and MoTe 2 ) by considering several parameters that affect HER performance; the number of active sites on the catalyst, the electrolyte interface, charge transfer from active sites to hydrogen, and the hydrogen conversion capability of the active sites (turnover frequency, TOF), which are summarized in Figure 4d. [60] The HER activities of the three semimetallic TMDs were investigated with their single crystalline surfaces in the same flat geometry, assuming that they possessed selective HER active Te atoms on their basal planes. The best HER performance was observed on the basal plane of VTe 2 , even though VTe 2 had less HER active Te atoms than other semimetallic TMDs. They claimed that the HER activity of the semimetallic TMDs on their basal planes was mostly determined by charge transfer efficiency, rather than the hydrogen conversion rate at the catalytic atoms. Another HER study of semimetallic TMDs was conducted with a series of Pt dichalcogenides (PtS 2 , PtSe 2 , and PtTe 2 ), and the best HER performance was observed in PtTe 2 . [61] They claimed that the electrocatalytic properties of the Pt dichalcogenides were affected not only by the size of the chalcogen atoms but also by the charge transfer efficiency. The size of the chalcogen atoms can change the bonding strength between the hydrogen and chalcogen atoms, while the charge transfer efficiency can be highly enhanced by changing the chalcogen atoms, such as from semiconducting (VS 2 and VSe 2 ) to metal (VTe 2 ). [62] Even though semimetallic TMDs are quantum materials that operate with intriguing physics and chemistry, their HER activities are still less attractive from an industrial point of view because of their low HER performances with rather large ΔG H* . [60] Enhancing the catalytic activity of the atoms in semimetallic TMDs by engineering their surface state via morphology and chemistry has become a central issue. In many semiconducting TMDs, creating defects and edge atoms has been the main approach to enhancing their catalytic activities. [14,15] For semimetallic TMDs, we expect to observe changes in catalytic activity, charge transfer ability and morphology resulting from defect generation in the catalytic materials. Theoretical studies have shown that the ΔG H* in active chalcogen atoms in the basal plane of semimetallic TMDs can be significantly reduced by introducing chalcogen vacancies (Figure 5a). [59,60] From the transport point of view, a fast electron supply to the adsorbed hydrogen should increase the production rate of hydrogen; therefore, enhancing the carrier transport at the interface of the catalyst-electrolyte and the carrier transport of the catalytic materials or supports could be a promising strategy to obtain practical HER performances. The carrier transport of semimetallic catalysts is less affected by the strategies that have been used to enhance the charge transport of semiconducting catalysts, such as introducing defects (including dopants, substitutional atoms, interstitial atoms, or vacant atomic sites).
Since most semimetallic TMDs belong to the category of semimetals with quasi-gaps near the Fermi energy (E F ), the density of state (DOS) at the E F can be modulated to some extent by introducing defects. Many experimental and theoretical works have observed enhanced HER performances in semimetallic TMDs with certain defects. Kwon et al. reported that point defects in the basal plane of semimetallic TMDs affect their work function, resulting in a reduced overpotential in the polarization curves of HER (Figure 5b,c). [63] On the other hand, creating enormous defects in the catalyst may also change its morphology. Kwon et al. reported that the electrochemical surface area for HER increased as the single crystalline surface of a semimetallic TMD become chapped during the electrochemical reduction process in acidic condition (Figure 5d). [63] Since the electrochemical surface area was significantly increased, the HER performance of electrochemical-reduced WTe 2 was enhanced by the abundant active sites (Figure 5e,f). [63] Semimetallic TMDs exhibit complex effects with defect formation, such as an increase in charge carriers, improved ΔG H* , and increased electrochemical active area for HER.
Having an HER active surface with dangling bonds and vacancies is considered an important prerequisite for high HER activity. However, inevitable disadvantages result from the instability of dangling bonds or vacancies, and these have hindered the longterm stability of HER performance. Semimetallic TMDs have intriguing physics and chemistry originating from their low dimensionality, and the surface state of the semimetallic TMD catalysts plays an important role in determining the level of carrier transport and surface reaction at the interface between the catalyst and electrolyte.

Transition Metal Dichalcogenide Alloy Catalysts
Introducing heteroatoms into TMDs at high concentration, to produce what are known as TMD alloys, has resulted in good HER performance with better phase stability than defective TMDs, which were discussed in the chapter IV. Sometimes, these TMD alloys have an interesting structural property called polymorphism, where the crystal structures vary with the contents of heteroatoms. [64][65][66] These polymorphic 2D TMD alloys provide a unique platform for exploring the structural aspects which influence the catalytic activity. Sometimes the substitution of heteroatoms into the TMD matrix can result in defective TMDs with vacancies or interstitial defects. [67][68][69] Therefore, alloying TMDs can also be an effective way to modulate the catalytic activity of the TMDs.
DFT calculations predicted that ΔG H* significantly decreases at vacant atomic sites in TMD alloys where the DOS at the Fermi level is high. [70,71] Angle resolved photoemission spectroscopy measurements of the Mo 1-x W x Te 2 alloy showed an increased DOS at the Fermi level due to the overlap between valence and conduction bands with increasing W content, resulting in enhanced charge transfer from the catalyst to the electrolyte (see Figure 6a). Excellent HER was also found in the W 1-x V x Se 2 alloy with x = 0.1 where d-band center was positioned at the apex of the Volcano plot with appropriate d-orbital filling (Figure 6b). [66] MoS 2(1-x) Se 2x nanosheets with abundant S and Se  [60] Copyright 2020, MDPI. b) UV photoemission spectra (UPS) of pristine and annealed WTe 2 single crystal. [63] c) Linear sweep voltammetry (LSV) polarization curves and Tafel plots acquired for pristine and annealed WTe 2 single crystals. [63] d) SEM images and Te/W ratio of the surface of pristine crystal and after CV cycles. [63] e) Comparison of Tafel slopes and overpotential (at 10 mA cm −2 ) according to the number of CV cycles. [63] f) Double layer capacitance (C dl ) of the pristine and the electrochemically reduced WTe 2 . Panels (b)-(f)) Reproduced with permission. [63] Copyright 2020, Elsevier. Figure 6. a) Electronic structure evolution for Mo 1-x W x Te 2 alloys with various W concentrations. Reproduced with permission. [70] Copyright 2018, American Chemical Society. b) Atomic-resolution STEM images and overpotentials ( ) of W 1-x V x Se 2 as a function of V concentration. Reproduced with permission. [66] Copyright 2022, American Chemical Society. c) DFT calculation of hydrogen adsorption free-energies of S and Se vacancies in MoS 2(1−x) Se 2x alloy nanosheets and catalytic performance of MoS 2(1−x) Se 2x alloy. Reproduced with permission. [71] Copyright 2013, Royal Society of Chemistry. d) EIS measurements from the Nyquist plots and Tafel analyses of Mo 1-x W x Se 2 alloys. Reproduced with permission. [72] Copyright 2009, Royal Society of Chemistry. www.advancedsciencenews.com www.advphysicsres.com vacancies, which was explained by the fact that ΔG H* changes with the content of S and Se vacancies. The optimal contents of vacancies in the MoS 2(1-x) Se 2x alloy produced the best HER performance (Figure 6c)). [71] Meiron et al. reported that the Mo 1-x W x Se 2 alloy became a good HER electrocatalyst in alkaline solution with an overpotential of 262 mV and a Tafel slope of 101 mV dec −1 , and in an acidic environment with an overpotential of 209 mV and Tafel slope of 73 mV dec −1 , while pristine MoSe 2 and WSe 2 were inactive in alkaline solution. [72] The study of electrode kinetics of Mo 1-x W x Se 2 alloys using electrochemical impedance spectroscopy showed that the lowest Tafel slope of the Mo 1-x W x Se 2 alloys occurred with the optimal content of W atoms needed for efficient charge transfer (Figure 6d).

Hybrid Catalysts
Semimetallic TMDs and conductive 2D supports with high carrier mobility are promising substrates for the fabrication of nanoparticulate catalysts, leading to a new type of hybrid catalysts with the synergetic effects of semimetallic TMDs. The unique catalysts have been fabricated on semimetallic TMDs using hydrogen bubble-assisted electrochemical deposition, and the catalysts showed efficient HER with a large double layer capacitance and lower Tafel slope. Ling et al. reported dynamic stacking of reduced graphene oxide (rGO) on WTe 2 by electrochemical deposition through spontaneously generated hydrogen bubbles (see Figure 7a). [17] The rGO catalyst with high porosity exhibited a large catalytic surface area with optimized, chemically treated edges, which exposed numerous active sites for HER. Similar to catalytically inactive supports like carbon nanotubes or graphene oxides, some semimetallic TMDs exhibit weak basal plane catalytic activity for HER. In that case, the hydrogen bubbleassisted growth of nanoporous catalytic materials could be a solution. [17,73,74] Nanoporous AgTe, synthesized by hydrogen bubble template with WTe 2 , was found to be active for HER (Figure 7b). [74] The DOS and ΔG H* locally vary at the atomic sites of Ag and Te in a AgTe slab, and the atomic sites with high DOS exhibit almost zero ΔG H* . In particular, hybridization of the p orbital of Te atoms with the d orbital of Ag atoms influences the strength of hydrogen adsorption, leading to excellent HER performance with a low overpotential and Tafel slope value, despite the poor catalytic activity of Ag. Gao et al. also reported that the S atoms in Ag-imbedded MoS 2 exhibited modulated ΔG H* by changing the distance between the two Ag atoms in the MoS 2 (Figure 7c). The local DOS near the Fermi energy was enhanced at HER-active S atoms with almost zero ΔG H* with an optimal distance between Ag atoms and Ag-S hybridization in Ag-imbedded MoS 2 electrocatalyst. [75] Seok et al. reported that Pt-Te alloy on semimetallic MoTe 2 was a good HER catalyst with a Tafel slope of 22 mV dec −1 , which was the best value among those reported for TMD-based catalysts. [76] They claimed that Pt atoms on atomically distorted semimetallic TMDs can have different HER activity with varied ΔG H* , as shown in Figure 7d. It is noted that chemically bonded Pt─Te with a coverage between 1 and 1.5 ML showed better HER performance than the Pt catalyst with the smallest ΔG H* . The hybrid catalyst, combined with semimetallic TMDs, provides rich catalytic active sites on large basal planes for efficient hydrogen production.
Other semimetallic 2D materials, composed of noble metal and chalcogen atoms, have shown excellent HER performance, similar to that of semimetallic TMDs. Bae et al. reported excellent HER performance of Pt 3 Te 4 nanorods grown on HERactive MoTe 2 . [73,77] Pt 3 Te 4 is a noble metal chalcogenide with a layered structure, where the two monolayers of PtTe and PtTe 2 are sequentially stacked (Figure 8a). The theoretically predicted ΔG H* on the Pt 3 Te 4 catalyst indicated that the edge atoms of the PtTe layer in Pt 3 Te 4 had the smallest ΔG H* compared with other atomic sites (Figure 8b). [77] In Figure 8c, the Pt 3 Te 4 catalyst showed an overpotential of 39.6 mV and a Tafel slope of 32.7 mV dec −1 . [73,[77][78][79][80] Abundant edge sites, resulting from the geometrical shape of the Pt 3 Te 4 nanorod, exhibited high HER performance, with the smallest overpotential and the lowest Tafel slope among former Pt chalcogenides. [77] Interestingly, Fujii et al. reported a topological semimetallic surface state in Pt 3 Te 4 with strong spin polarization, [81] and an extremely high current density exceeding 6-7 A cm −2 was observed in a Pt 3 Te 4 catalyst, [73] highlighting the possibility of a stable HER catalyst for mass hydrogen production (Figure 8d).
It has recently it has been reported that topological surfaces states from nontrivial band topology can be used as catalysts for various electrochemical reactions such as CO oxidation, hydrogen/oxygen evolution, and so on. [82][83][84] Topological catalysts have shown intriguing catalytic behaviors with extraordinary charge transport, originating from their massless fermions. [85][86][87] Indeed, topological surfaces states are known to be robust surface states against electrochemical reactions because topological surface states are generated by the inversion of bulk bands at the surface. Many topological insulators (Bi 2 Se 3 ), Dirac semimetals (PdSn 4 ), and Weyl semimetals (NbP, TaP, NbAs, and TaAs) with robust topological surface states, high current density, and large room temperature carrier mobility have shown excellent HER performance, opening a new era of topological catalysts for HER. [85,88,89]

Micro-and Atomic-Scale Catalysts
Recently, extraordinary HER activity has been observed at the interface, surface and grain (or phase) boundaries of heterogenous TMD catalysts, and their derivatives. [90,91] This performance has been explained by enhanced charge transfer via phase engineering or charge accumulation at the localized areas. To reveal the electrochemical properties of the interface, surface, grain (or phase) boundaries, micrometer-scale areas including interfaces, and boundaries should be selectively exposed to the electrolyte for HER, essentially creating HER microreactors. HER microreactors have been precisely designed on micrometer-scale catalytic areas for local electrochemical measurements.
For example, a HER microreactor was fabricated by e-beam lithography as described in Figure 9a. [92] 2D materials have been prepared by synthesis, or chemical/mechanical exfoliation on a dielectric substrate for gate-tunable on-chip electrocatalytic devices. The HER windows are selectively opened with insulating polymethyl methacrylate layer. Figure 9b shows an experimental HER microreactor set-up with 35 nm thick MoS 2 with gating voltage (V bg ). Reference and counter electrodes were placed into the droplet of electrolyte on the HER window, and the working electrode was connected to the probe tip. [91] Figure 7. a) HR-TEM images of the rGO-WTe 2 , TEM image of cross-sectional Pt 3 Te 4 on MoTe 2 template, SEM images of nanoporous AgTe grown on WTe 2 , respectively. Reproduced with permission. [17] Copyright 2019, American Chemical Society. Reproduced with permission. [73] Copyright 2021, Elsevier. Reproduced with permission. [74] Copyright 2021, American Chemical Society. b) Top and side views of an AgTe slab. The projected density of state (PDOS) of six atomic sites for AgTe slab (red solid line) and bulk AgTe (black dashed line), and its corresponding Gibbs free energy changes (ΔG H ) of hydrogen adsorption at each site. Reproduced with permission. [74] Copyright 2021, American Chemical Society. c) Calculated free energy (ΔG H* ) diagram for HER on MoS 2 and Ag-MoS 2 electrocatalysts and its relationship with the distance between the two doped Ag atoms. Reproduced with permission. [75] Copyright 2021, Elsevier. d) Illustration showing the preparation of the Pt/MoTe 2 hybrid catalyst via an electrochemical method. Polarization curves of Pt/MoTe 2 hybrid catalysts with various treatment times and comparison with a bulk Pt rod, Pt film, and pristine MoTe 2 . Schematics of the Pt/MoTe 2 hybrid structure at different Pt coverages, along with respective ΔG H values. Reproduced with permission. [76] Copyright 2019, American Institute of Physics.  [73] Copyright 2021, Elsevier. b) Comparison of hydrogen adsorption free-energy changes for the surface of a monolayer, the surface with a Te vacancy and the edge of PtTe 2 and PtTe, respectively, along with a Pt (111) surface. Reproduced with permission. [77] Copyright 2021, American Chemical Society. c) Polarization curves and corresponding Tafel slopes of Pt 3 Te 4 nanocrystals, pristine MoTe 2 , Pt film, and Pt/C (20 wt% Pt) for HER in acidic condition. Reproduced with permission. [77] Copyright 2021, American Chemical Society. d) Experimental results and theoretical calculations for the PtTe 2 and Pt 2 Te 2 termination. Each graph indicates the band dispersion along the Γ-K direction, with spin spectra and spin polarization. Reproduced with permission. [81] Copyright 2021, American Chemical Society.
The interfacial electric transport of TMDs also plays an important role in electrocatalytic performance. HER microreactors can provide insights into local electrochemical reactions at a specific area, and the contribution of the carrier density and charge transfer to the HER performance can be precisely modulated using an electric field. As shown in Figure 9c, the HER activity of monolayer MoS 2 could be modulating by selfgating, enhancing charge transfer at the interface of MoS 2 and electrolyte. [91] The surface carrier concentration in MoS 2 can be modulated by the applied electrochemical potential, which is similar to the gate voltage-driven modulation of field-effect transistors. [93] HER microreactors with selected sample areas enable rigorous study of the role of edges and basal plane of 2D materials. Figure 9d shows that the edge sites in TMDs behaved as catalytically active sites for HER while its basal plane was inert for HER. [93] Ling et al. fabricated a HER microreactor to study the effect of oxidation on catalytic activity in semimetallic WTe 2 using O 2 plasma-treatment. The plasma-treated WTe 2 edges in the HER microreactor exhibited an improved overpotential (538-325 mV at 10 mA cm −2 ) and Tafel slope (145-96 mV dec −1 ). [94] Compared with previous bulk studies, the HER microreactor can reveal the effects of microstructures such as grains and heterophase boundary more precisely. The HER microreactor is useful for Figure 9. a) Fabrication process of HER microreactor. Reproduced with permission. [92] Copyright 2020, Royal Society Chemistry. b) Experimental setup to unveil HER active sites at the basal plane of MoS 2 using gated HER microreactor. Reproduced with permission. [91] Copyright 2019, Springer Nature. c) Gating voltage (V BG )-dependent current density (i W ) measured by HER microreactors and the surface conductance of MoS 2 electrocatalyst enhanced by self-gating. Reproduced with permission. [91] Copyright 2019, Springer Nature. Reproduced with permission. [93] Copyright 2019, American Chemical Society. d) Comparison of basal-plane-activity and edge-activity for HER when the surface of WTe 2 was treated by O 2 plasma. Reproduced with permission. [94] Copyright 2021, American Institute of Physics.
uncovering the origin of catalytic reactions in micro-structured 2D materials.
Zhu et al. reported that Ar-plasma treatment induced a phase transition from a 2H to 1T phase in MoS 2 with many grains and grain boundaries. [95] It has been reported that Ar-plasma treated MoS 2 contained a nano-structured heterophase boundary, via HER microreactor. [96,97] The MoS 2 with 2H and 1T' phase was opened for HER and the exposed MoS 2 areas exhibited abundant heterophase boundaries of 2H and 1T', leading to better HER performance than samples with no grain boundary. The Ar plasma induced heterophase samples showed clear enhancement in HER performance compared with the pure phase samples. The catalytic activity of the heterophase boundary in 2D catalysts will be discussed in the following chapter VIII.

Heterophase Boundary Catalysts
Heterophase boundary engineering, described in Figure 10a, can be a promising way to obtain high HER performance without creating defects in the system. The systematic study of heterophase boundaries in 2D catalysts requires large scale sample synthe-sis with well-optimized heterophase boundaries, or small scale HER probing technique for atomic catalysts. Most HER studies have been conducted at macroscopic scale with massive catalysts without defining the detailed HER mechanism which occurs at the phase boundary. Here, we introduce a rigorous study of the catalytic activity of heterophase in 2D catalysts using phasecontrolled 2D materials and a HER microreactor together with theoretical calculations.
The atomic sties at the heterophase boundary exhibit superior ΔG H* compared with pristine TMDs with single phases of 2H, 1T, or 1T'. Zhao et al. reported theoretical calculations of the catalytic activity of heterophase boundaries. [98] Figure 10b shows the atomic structures of possible heterostructures with the 2H-1T' phase in MoS 2 . The calculated catalytic activities of several active sites on the atomic scale are plotted in a volcano plot in Figure 10c. In the volcano plot, some of the active sites at the heterophase boundaries are almost at the top of the plot, which is even better than bulk Pt.
The enhanced HER performance at the heterophase boundary (ZZ) can be explained by the enhanced DOS near the Fermi level. Compared with a pure 2H phase, both armchair (AC) and Figure 10. a) Schematics of heterophase boundary with HER active atoms (yellow spheres). b) Atomic structures for various 2H-1T' heterophase boundaries. The red and black circles indicate the active atomic sites at the heterophase boundary for HER. [98] c) Volcano plot of the heterophase boundary (ZZ) and AC structured heterophase boundary atoms, 1T' and Pt. [98] d) Density of states for the 2H phase, AC 2H-1T' heterophase, and ZZ heterophase interface of MoS 2 . Panels (b)-(d)) Reproduced with permission. [98] Copyright 2019, American Chemical Society.
heterophase boundaries (ZZ) (Figure 10d) showed much higher DOS near the Fermi level, and that could be related to the efficient charge transfer at the interface between the atomic sites at the heterophase boundaries and the electrolyte. [98] Previous research has mentioned that the 2H-1T (1T') boundary shows semimetallic behaviors, with charge accumulation, because of the differences in the work functions of the 2H and 1T phases. This was confirmed by calculations, scanning tunneling microscopy, and Kelvin probe force microscopy (KPFM). [99][100][101] Therefore, efficient charge transfer at the interface between the catalyst and electrolyte could boost the catalytic activity more than on the basal plane of pristine TMDs.
Theoretical calculations have indicated that the heterophase boundaries in MoS 2 had almost zero ΔG H* , compared with the grain boundaries in polycrystalline 2H-, and 1T MoS 2 . Huang et al. reported enhanced HER performance by the in situ generation of heterophase boundaries using Li intercalation in exfoliated MoS 2 nanosheets via HER microreactor. [97] Compared with MoS 2 nanosheets with single phases of 2H and 1T, the 2H/1T mixed MoS 2 nanosheets exhibited improved HER performance, as shown in the polarization curves and Tafel plots in Figure 11a. [95] The phase stability of the mixed phase MoS 2 with 2H and 1T seemed to be robust in acidic electrolyte compared to the MoS 2 with a metastable 1T phase.
It has recently been reported that the heterophase boundaries of polymorphic TMDs exhibit enhanced HER activity compared to pristine TMDs. Lee et al. showed that MoTe 2 with abundant heterophase boundaries of 1T'-2H exhibits a high exchange current (right side of Figure 11b) and a large potential difference at the heterophase boundary. [102] The rate of hydrogen production at the heterophase boundary was estimated using the TOF, 317 s −1 , which is 10 3 times larger than that of the pristine 1T' MoTe 2 . Nguyen et al. showed that the heterophase boundary of 1T' Re x Mo 1-x S 2 -2H MoS 2 has superior catalytic performance compared with pure phase MoS 2 , ReS 2 , and Re x Mo 1-x S 2 as shown in Figure 11c. [103] They further showed sharp band bending at the heterophase boundary using the KPFM (see the middle panel in Figure 11c). The modulation of band structure at the heterophase boundary can enhance the catalytic activity of the boundary atoms. As shown in the right panel of Figure 11c, DFT calculations showed that locally accumulated charges can be located at the S atoms in the heterophase boundary and high electrical conductivity can be achieved because of the largely enhanced DOS near Fermi energy. Recently, crystal symmetry breaking at the heterophase boundary of 2H-1T' was shown to enhance catalytic activity. [104,105] Piezoelectricity originating from crystal symmetry breaking helps to form surface dipoles. [106][107][108]

Conclusion and Perspectives
We have discussed recent efforts to enhance the HER catalytic activity of semimetallic 2D TMDs, focusing on the design of semimetallic and semiconducting low dimensional materials. We introduced earlier strategies to achieve high HER performance using traditional noble metal catalysts and their derivatives based on the Sabatier principle. By reducing their dimensions to sub-nanometer-scale, it was discovered that single atoms can generate effective HER with low mass loads and abundant active sites. In addition, the intrinsic HER that occurs on the basal plane of semimetallic TMDs with their intriguing surface states was discussed. From physics and chemistry perspectives, the material design (e.g., alloying, hybridizing, phase boundary engineering, and heterostructuring) of emerging 2D electrocata- Figure 11. a) Optical microscopy images of MoS 2 catalysts with many grains. LSV curves and Tafel slopes of the pristine MoS 2 (black and red lines) and heterophase MoS 2 (blue and green lines). Reproduced with permission. [95] Copyright 2019, Springer Nature. b) Schematic of a three-electrode HER microreactor system. Optical microscope images of two microreactors with different heterophase boundaries. Exchange current density of 1T'-2H heterophase as a function of the boundary length in a unit area. Reproduced with permission. [102] Copyright 2021, Wiley. c) LSV curves of heterophase boundary with 1T' Re x Mo 1-x S 2 and 2H MoS 2 (blue line) with references of Pt/Gr (yellow line), graphene (black line), MoS 2 (green line), ReS 2 (magenta line), and Re x Mo 1-x S 2 (red line). KPFM mapping image and the line profile at the heterophase boundary of 1T'-2H. Projected DOS of S atom at the Mo-rich heterophase boundary. Reproduced with permission. [103] Copyright 2022, Wiley.
lysts is a promising strategy for HER and green energy sources of hydrogen.