AAPPS Bulletin
 
 
> home > Feature Articles
 
Strategies for Efficient Generation of Hydrogen in Monolayer Transition Metal Dichalcogenides
Frederick Osei-Tutu Agyapong-F
File 1 : Vol30_No2_Feature Articles-3.pdf (0 byte)

DOI: 10.22661/AAPPSBL.2020.30.2.10

Strategies for Efficient Generation of Hydrogen in Monolayer Transition Metal Dichalcogenides

FREDERICK OSEI-TUTU AGYAPONG-FORDJOUR†§, ADOFO LAUD ANIM†§, SOO MIN KIM, KI KANG KIM†#*
DEPARTMENT OF ENERGY SCIENCE, SUNGKYUNKWAN UNIVERSITY (SKKU), SUWON 16419,
REPUBLIC OF KOREA.
#CENTER FOR INTEGRATED NANOSTRUCTURE PHYSICS, INSTITUTE FOR BASIC SCIENCE (IBS),
SUWON 16419, REPUBLIC OF KOREA.
DEPARTMENT OF CHEMISTRY, SOOKMYUNG WOMEN'S UNIVERSITY, SEOUL 140742, REPUBLIC OF KOREA.

*Corresponding Author: kikangkim@skku.edu

ABSTRACT

Hydrogen has been highly regarded as the ultimate clean energy carrier with zero carbon content and the highest mass-energy density of any fuel. Pt is an ideal electrocatalyst for producing hydrogen via water electrolysis, but its high cost and rarity in the earth's crust restricts commercialization. As an alternative, two-dimensional (2D) materials have been widely investigated to replace Pt due to the optimized Gibbs free energy of hydrogen ion adsorption at a specific catalytic site. This review provides extensive insights into understanding the role of monolayer 2D materials for catalyzing the hydrogen evolution reaction at the atomic level. Furthermore, several approaches by means of computational simulation and strategic experimental methods such as phase transition (2H-semiconductor to 1T-metal), induction of active sites through substitutional doping and chalcogen vacancies, electrical coupling in heterostructures and purposeful control of morphology are addressed. Lastly, we discuss the major bottlenecks and future perspectives in advancing toward suitable alternatives to Pt.

INTRODUCTION

Hydrogen is a sustainable and eco-friendly alternative energy source free of carbon dioxide emission. It is estimated that the energy in 2.2 pounds (1 kilogram) of hydrogen gas is about the same as the energy in 1 gallon (6.2 pounds, 2.8 kilograms) of gasoline; hence, there is the need to efficiently produce hydrogen and store it as a viable fuel source. Currently, the majority of hydrogen (~ 95%) is produced from fossil fuels by steam reforming, partial oxidation of methane or coal gasification with only a small quantity by other routes such as biomass gasification or electrolysis of water. Among the various methods of hydrogen production, water electrolysis has attracted tremendous attention because of related advantages such as abundant water (71% in the earth's surface), its high efficiency, and it is non-polluting, unlike coal gasification and other forms of hydrogen production. In a typical water electrolysis system, H2 and O2 are produced at the cathode, and the anode through the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER), respectively, and an external current is applied to overcome the 237 kJ/mol energy barrier of the reaction [1]. To reduce the activation barrier, a catalyst is highly desirable. Pt is regarded as a perfect electrocatalyst, exhibiting low overpotentials, high exchange current densities, and remarkable stability in the water reduction reaction. Its unique activity has been explained by its high electrical conductivity and its almost perfect thermoneutrality at the equilibrium potential of hydrogen ion adsorption [2, 3]. However, the scarcity of Pt in the earth's crust makes scaling Pt-based processes challenging for industrial applications. To date, various catalysts such as cobalt phosphide (CoP) and metal nitride catalyst (NiMoN) have been extensively explored to realize suitable substitutes for Pt [4, 5]. Two-dimensional (2D) transition metal dichalcogenides (TMdCs) are of interest in H2 evolution because of their unique structural and electronic properties [6]. The atomically thin and flat nature of monolayer TMdCs makes them perfect platforms to study and understand the mechanism of HER catalysis [7]. Since the performance of 2D bulk electrocatalysts strongly depends on the number of layers, where the hopping of electrons to active sites occurs through the van der Waals gap, monolayer TMdCs hold an advantage over their bulk counterparts due to faster electron injection from conductive substrates.

Group V metallic monolayer TMdCs such as vanadium disulfide (VS2), tantalum disulfide (TaS2) and niobium disulfide (NbS2) have been touted as ideal electrocatalysts for HER due to their favorable Gibbs free energy (ΔGH) for adsorption of hydrogen ions, possessing numerous active sites at the basal plane and edge, and an intrinsically metallic nature to reduce the loss by electron transfer from electrode [8-10]. However, they may oxidize in aqueous media; hence, their instability is a drawback for long-term catalytic activities [11]. Therefore, group VI semiconducting TMdCs such as molybdenum disulfide (MoS2) and tungsten disulfide (WS2) that possess high stability have attracted so much attention. But, it has been revealed that only the edge site is active for hydrogen evolution [12]. To maximize the catalytic activity, the basal plane needs to be active. Hence, extensive studies of processes such as oxidation, doping, strain, phase transition from 2H to 1T structure and construction of heterostructures have been conducted [13-17]. Though these techniques improve the restricting factors for the potential of monolayer TMdCs in HER, achieving the industrially required current densities of 500 - 1000 mAcm-2 at low voltages of 1.5 - 1.7 V is a daunting task [5].

In this review, the progress of the hydrogen evolution reaction for 2D monolayer TMdCs is highlighted as a guide to related research. Several remarkable achievements of catalysts will be discussed. Finally, we present the significant challenges in tuning the surface properties of these catalysts to enhance activity. We hope that this comprehensive review will inspire more prospective studies in pursuit of alternatives to Pt.

 



Fig. 1: Schematic diagram of sustainable hydrogen economy and usage.

HYDROGEN ECONOMY AND USAGE

The hydrogen economy involves replacement of fossil fuel with hydrogen energy in human activities [1]. Hydrogen can be produced, stored and used as energy source when necessary. Seasonal energy sources such as solar power, wind and tidal energy can be used as energy sources for the generation of hydrogen by water electrolysis (Fig. 1). The generated hydrogen can be compressed for storage in tanks and used in various fields including the generation of electricity via a fuel cell for electric vehicles and domestic purposes and as a fuel for space exploration. In the chemical industry, hydrogen reacts with nitrogen gas to produce ammonia for use as an agricultural fertilizer (the Haber process). Hydrogen can also be used to produce cyclohexane and methanol, which are intermediates in the production of plastics and pharmaceuticals. Hydrogen is used in other industrial applications; such as a protecting gas, and a reductant and etchant during material fabrication process. The versatility of hydrogen and its multiple uses is why hydrogen can contribute to a green and more sustainable future.

WHY 2D MATERIALS FOR HER?

Electrochemical reactions are based on the transfer of one or more electrons, typically at the electrode surface. As a result, generated active sites and modifications located close to the surface of the electrocatalyst can play a significant role in the reaction process. The interplay between catalytic activity and atomic active sites can be clearly observed and quantified in large areas of monolayer TMdCs using analytical tools such as tunneling electron microscopy (TEM), scanning electron microscopy (SEM) and scanning tunneling microscopy (STM). 2D TMdCs have been intensely studied for HER catalysis, thus making them our primary focus. A typical TMdCs with the general formulae MX2 consists of a three-atomic thickness structure with one transition metal atom layer sandwiched by two chalcogen atom layers [7]. The atomic layers are bonded to each other by the relatively weak van der Waals attractions, which allows for easy separation of monolayer TMdCs from their bulk materials [3]. Their atomically thin nature endows them with many distinctive properties with respect to their bulk counterparts such as abundant uncoordinated surface atoms, tunable electron mobility and mechanical flexibility [6]. Various studies have shown that monolayer TMdCs possess tunable ΔGH, which allows for extensive studies to elucidate ways to fine-tune the active sites to approach the optimal value (ΔGH ≈ 0 eV) [15]. These features make the 2D TMdCs ideal candidates (or parts for hybrid architectures) with improved electrocatalytic performance to substitute for their parent materials. To understand why there is a surge in hydrogen evolution research among TMdCs, some superior physical properties for catalysis are highlighted in Fig. 2. They have been summarized in two groups: self-structure changing and hybrid structure modification.

 

Fig. 2: Superior physical properties of 2D dimensional materials as leading candidates for catalysis. a. Size control with growth conditions b. Layer control c. Vacancy control. d. Chalcogen loading as alloys. e. Substitutional metal loading. f. Construction of vertical and lateral heterostructure.

Self-Structure changing

Several conventional preparation techniques such as physical vapor deposition, chemical vapor transport (CVT) and chemical vapor deposition (CVD) have demonstrated the production of high-quality large-area monolayer TMdCs synthesized by carefully controlling deposition parameters. Among these methods, CVD has aroused significant attention due to its superior advantages such as controlled structure, large area, high throughput and low cost. CVD based methods provide high-quality flakes with controllable size and number of layers and thickness (Fig. 2a) [18]. In addition, layer-controlled CVD growth of large-area MoS2 films is achieved by pre-treatment of the SiO2 substrate with O2 plasma before the introduction of Mo and S precursors (Fig. 2b) [19]. Several vacancies can be induced in the basal planes of TMdCs through various plasma treatment strategies such as O2 plasma treatment, helium plasma and annealing in H2 [13, 14]. Recently, ultrarich S vacancies by stripping S from the MoS2 framework with thermal annealing (~ 600C) under H2 atmosphere has been achieved. Specifically, defects in the form of S vacancies are expected to induce a reduction of the free energy of hydrogen adsorption at the basal plane (Fig. 2c) [20].

Hybrid Structure Modification

As discussed earlier, 2D TMdCs consist of one-layer metal atoms sandwiched between two-layer chalcogen atoms. In contrast to typical 2D TMdCs, Janus 2D monolayer TMdCs is formed with a different chalcogen in each layer (Fig. 2d) [21]. Zhang et al. synthesized the Janus monolayer MoSSe through the sulfurization of monolayer MoSe2. The breakage of the out-of-plane structural symmetry induces a significant difference in the physical and chemical properties compared to the pristine case. By DFT calculations, while pristine TMdCs require a considerably large amount of tensile strain or S-vacancies to activate their basal planes, Janus asymmetry WSSe monolayers require minimal or no applied strain to achieve the optimal 0 eV for hydrogen adsorption [22]. This effect was attributed to the intrinsic lattice strain of WSSe as a result of its asymmetrical nature. Recent CVT work has demonstrated that Niobium (Nb), which has one less valence electron than Mo, can be introduced in the host TMdCs as a covalently bonded dopant (Fig. 2e) [23]. Theoretical study further predicts that the inactive S-atoms in the basal plane similarly become active for hydrogen adsorption upon metallic substitutional doping by Zn, like the Mo-edge of pristine MoS2 [24]. Furthermore, strong van der Waals (vdW) forces allow 2D-dimensional nanomaterial units to assemble into heterostructures, exhibiting unprecedented physical and chemical properties by electronic coupling or orbital hybridization between layers [17]. Gong et al. demonstrated that the growth of in-plane a WS2/MoS2 monolayer (lateral heterostructure) dominates at 650 C, but at elevated temperature above 850 C, a sequential growth of MoS2 and WS2 leads to a vertically stacked bilayer (vertical heterostructure) (Fig. 2f) [25]. Well-defined, sharp and clean interfaces in these heterostructures allow for improved electron transfer kinetics at the catalyst's surface. These self-structure changes and hybrid structure modifications will play beneficial roles in enhancing the overall catalytic activity of the monolayer TMdCs for hydrogen evolution.

CATALYTIC ACTIVITY PARAMETERS FOR THE EVALUATION OF HER

Hydrogen evolution is a two-electron reaction process which takes place at the surface of various catalysts in aqueous media. In an acidic medium, the cathodic reaction of water electrocatalysis is the reduction of hydronium ions (H3O+) to hydrogen gas. From a thermodynamic point of view, this multi-step electrode reaction should occur at the potential of the reference hydrogen electrode (RHE).

 



Fig. 3: The mechanism of hydrogen evolution on the surface of an electrode in an acidic solution. (Ref. 3)

Absorbed hydrogen is produced from the reaction between hydronium and an electron in the Volmer step 1.1 (Fig. 3). The subsequent procedure is termed a hydrogen evolution reaction, which can proceed through either a Heyrovsky step (1.2), where the absorbed hydrogen atoms combine with another proton from the solution to form H2, or the Tafel step (1.3), the combination of two hydrogen atoms adsorbed on the surface of the electrode [3].

(Volmer) H3O+ + e- + * → H* + H2O

(1.1)

(Heyrovsky) H* + H3O+ + e-→ H2 + H2O

(1.2)

(Tafel) H* + H* → H2

(1.3)

Experimental HER studies are performed in a three- electrode setup that includes working, counter and reference electrodes. The counter electrode should especially have no significant influence on the electrochemical properties of the working electrode typically during long-duration tests in an acidic medium. Graphite rod and carbon cloth are used as counter electrode. A common reference electrode used is the silver chloride electrode or saturated calomel electrode [26].

Onset potential and overpotential

The applied potential where there is apparent cathodic current is termed the onset potential. This could be calculated at 1 mA cm-2 current density. Overpotential can be defined as the additional potential above the thermodynamic requirement needed to drive a reaction at a certain rate. Overpotential value for 10 mV cm-2 current density (ɳ10) is the overpotential expected for 12.3 % efficiency of a solar water splitting device and this method is used to compare the performance of electrocatalysts quantitatively (Fig. 4a) [28]. A good catalyst will deliver this current at a very low overpotential.

 

Fig. 4: HER characterization (a) Polarization curves for carbon fiber cloth, NiSx/C, MoS2/C, and different Ni-Mo-S/C compositions in a neutral electrolyte. (b) Corresponding Tafel plots obtained using slow-scan rate polarization curves. (c) Electrochemical impedance spectra of different electrodes at -0.3 V versus RHE. (d) LSV curves of Ni-Mo-S/C (1:1) before and after continuous cycling. (e) Comparison of the detected amount of evolved H2 as compared to the theoretical amount in Faraday efficiency measurement (ref. 26). (f) Difference in current density (ja - jc) at 0.24 V vs RHE plotted against scan rate fitted to a linear regression showing the extraction of double-layer capacitance (Cdl) for the estimation of relative electrochemically active surface area. Inset: voltammograms of a VS2 electrode at various scan rates (ref. 29).

Tafel slope and exchange current density

The reaction kinetics for HER can be revealed by the Tafel slope and exchange current density. Tafel curves can be obtained from the polarization curves by simply using the Tafel equation:

ɳ = a + blog(j)

Where ɳ, b and j are the overpotential, the Tafel slope and the current density, respectively [28]. The Tafel slope b gives insight into the reaction mechanism. If the Tafel slope is ~ 0.029 V/dec, then chemical combination of adsorbed hydrogen (equation 1.3) is the rate determining step (Volmer-Tafel mechanism). If it is ~ 0.038 V/dec, the electrochemical desorption (equation 1.2) is the rate determining step (Volmer- Heyrovsky mechanism). A Tafel slope of 0.116 V/dec suggests discharge reaction of the Volmer step (equation 1.1) is the rate limiting step. The exchange current density (j0) suggests the catalytic activity of a material can be estimated by extrapolating the Tafel curve to 0 V at ɳ = 0. A large exchange current density indicates high reaction kinetics.

Electrochemical impedance spectroscopy

The charge transfer resistance (Rct) is related to the electrode electrolyte interface charge transfer process. This information can be obtained by the fitted Nyquist plot from the electrochemical impedance spectroscopy measurement at the diameter of the semicircle in the high frequency region (Fig. 4c) [26]. A lower Rct indicates a faster reaction rate. The intercept of the semicircle in the low frequency region gives an estimation of the series resistance.

Stability and durability

Durability and long-term stability are very crucial in assessing the catalytic performance of a material. Usually, cyclic voltammetry (CV) tests and chronoamperometry (static potential-time curve) or chronopotentiometry (static current-time curve) are used. Concerning durability testing, polarization curves before and after continuous CV cycles (1000-10000 cycles) are carried out. If the final polarization curve almost overlays the initial curve or there is a less than 10 % increase in overpotential then the material exhibits high durability (Fig. 4d). For long term stability testing, either chronoamperometry or chronopotentiometry is measured for certain hours (normally for more than 10 hours) with a given applied potential or current density, respectively. Chronoamperometry is more common among researchers and in most cases the potential for 10 mA cm-2 current density is applied. Discharging a stable current at the applied potential for long hours shows the stability of the catalyst. Long-term stability, which involves discharging a stable current for many hours (normally for more than 10 hours), is ideal for stability comparison.

Faradaic Efficiency

Faradaic efficiency (FE) describes the transfer efficiency of electric charge in an electrochemical reaction. FE plays a role in detecting the complete decomposition of water. This is calculated by comparing experimentally produced H2 gas (gas chromatography) with theoretically calculated H2 (Fig. 4e).

Electrochemical active surface area

To evaluate the electrochemical active surface area, electrochemical double layer capacitance (Cdl) is measured. The most effective way is to measure Cdl at the solid-liquid interface within a potential window where there are no apparent faradaic processes at a series of sweep speeds (20-200 mV/s). By plotting the current difference between anodic and cathodic current densities (∆j = janodic - jcathodic) against each scan rate at an overpotential (intermediate of the potential range), a linear fitting can be conducted to estimate the Cdl, where the slope of the linear fitting is twice the Cdl (Fig. 4f) [26]. A larger Cdl shows more exposed surface reactive sites and thus much higher current density.

Turnover frequency (TOF)

It is very difficult to judge catalytic performance precisely because independent examinations are always with different electrocatalyst mass loadings. TOF is defined as the number of reactant molecules transformed per active site in a second. A higher TOF value shows better catalytic performance. Comparison with other materials are usually done at the same overpotential [27].

IMPROVEMENT OF HER PERFORMANCE WITH ENGINEERING ACTIVE SITES

Both theoretical and experimental studies have indicated that the edges of monolayer semiconducting TMdCs are the active sites for HER, exhibiting a Gibbs free energy of absorbed H ion close to zero [12]. As a result, many recent efforts have been focused on increasing the density of exposed defective edges in TMdCs using oxygen plasma treatment, Ar plasma bombarding and hydrogen plasma treatments [16]. Gonglan Ye et al. demonstrated that oxygen plasma etches the surface of monolayer MoS2 and hence creates a large number of defects that act as catalytic active sites for HER (Fig. 5a). Results show catalytic activity is strongly dependent on the duration of plasma exposure [14]. The sample exposed for 20s showed the smallest onset overpotential. This corresponds to the increasing electrochemical active sites for MoS2 with increasing exposure time. However, it is worth noting that overexposure to plasma treatment could lead to deterioration of the intrinsic active sites of the catalyst.

 

Fig. 5: Improvement of HER performance with active sites (a) Polarization curves of MoS2 before and after oxygen plasma treatment (ref. 14). (b) Polarization curves of different MoS2 catalysts (ref. 15). Inset: schematic of the upper panel and lower panel views of strained basal MoS2 plane S-vacancies. (c) Polarization curves of monolayer MoS2(1-x)Se2x. Inset: diagrammatic view of the catalytic activity of monolayer MoS2(1-x)Se2x for HER. (ref. 30). (d) Polarization curves of monolayer Co doped MoS2 (ref. 31). (e) Polarization curves of the heterophase monolayer MoS2 pristine type-I MoS2 (No domain boundaries), pristine type-III MoS2 (2H-2H), heterophase type-I MoS2 (2H-1T), and heterophase type-III MoS2 (2H-2H and 2H-1T), respectively (ref. 16). (f) Polarization curves of self-optimized H-NbS2 after different cycles. Inset: schematic proposed mechanism for the morphology change (ref. 9).

Applying strain in addition to generation of S vacancies in the basal plane of monolayer 2H-TMdCs is a competent way of increasing the active sites for HER. Density functional theory (DFT) calculations indicate that S vacancies introduce gap states between conduction and valence bands that allow favorable hydrogen adsorption. Straining the S-vacancy sites further shifts the gap states even closer to the Fermi level, reaching the optimal ΔGH*. Experimentally, optimized monolayers of strained MoS2 with S vacancies (SV-MoS2) exhibit high HER activity (Fig. 5b) [15]. A tandem application of strain and S vacancies reduces the overpotential from 250 mV (vacancy only) to 170 mV for 10 mAcm-2 current density with a TOF per S-vacancy site of 0.31 H2 molecules per second at 0 V (vs RHE).

Substitution of chalcogenides (S to Se) and metal element doping (Mo to Co) are also valid strategies for triggering the catalytic activity of the basal plane [24]. Yang et al. reported that monolayer MoS2 (1-x) Se2x with different Se contents (x = 0.39, 0.51 and 0.61) exhibits higher catalytic activity compared with monolayer MoS2 and MoSe2 (Fig. 5c). This improved result is attributed to the optimal ΔGH* and narrowing of the MoS2 band gap by induced impurity states [30]. Particularly, the corresponding polarization curve shows that the onset potential of MoS2 (1-x) Se2x (~273 mV) is much smaller than that of pristine MoS2 (~335 mV) or MoSe2 (~303 mV). Based on DFT calculations, HER performance of piezoelectric Janus TMdCs will be further improved by introducing chalcogen vacancies due to the intrinsic strain and an internal electric field caused by an asymmetric structure [22].

Like chalcogen substitution, impurity metal atoms can also be introduced to transition metal sites in TMdCs to achieve higher efficiency for H2 adsorption. Hai et al. demonstrated that when cobalt (Co) is substituted in monolayer MoS2, the in-plane MoS2 experiences a compressive strain [31]. This causes a change of ΔGH*, resulting in significant improvement of HER activity with a reduction in overpotential from 283 mV to 156 mV at a current density of 10mAcm-2 (Fig. 5d).

Recently, the phase boundary between the semiconducting domain (2H) and the metallic domain (1T) and the domain boundary between 2H and 1T across the basal plane of MoS2 have been found to influence the catalytic activity (Fig. 5e). Through CVD coupled with plasma induction, 2H and 1T phase boundaries were induced in poly-crystalline 2H-MoS2 [16]. The overpotential at 10 mAcm-2 current density decreased due to the presence of domain boundaries from ~ 375 mV (pristine type-I (no domain boundary)) to ~ 325 mV (pristine type-III). It is reduced to ~ 260 mV with phase boundaries (heterophase type I). The high density of phase and domain boundaries further improve the catalytic activity, eventually reaching ~ 200 mV (heterophase type-III).

The key factors that influence HER performance are active sites and their densities in the basal surface. Metallic TMdCs such as VS2 and NbS2 tend to show more active sites due to populated states at or near the lowest unoccupied state (琯LUS) i.e., Fermi level during H* adsorption. For the metallic system, H adsorption does not change the overall density of states (DOS); instead, it shifts the Fermi level slightly to reflect complete charge transfer from the H adsorbate [8-10]. The charge density distribution shows that the transferred electrons are delocalized throughout the M layer. Interestingly metallic TMdCs exhibit a steep increase in HER activity; thus, the term self-optimized catalytic performance. H-NbS2 flakes exhibit continual dramatic improvement in catalytic performance from a CV test of 1000 cycles to a steady state after 12000 cycles (Fig. 5f) [9]. This improved catalytic activity upon cycling originated from the shortening of the interlayer electron-transfer pathways and increased accessibility of electrons to active sites due to the thinning of flakes during CV tests. Notably, there is a morphology change in metallic TMdCs from layered flakes to monolayer flakes resulting in beneficial consequences for catalytic activities. It can be seen that remarkable progress has been achieved in the search for the best catalyst for HER, however there are still more issues and challenges such as stability and prolonged efficiency to overcome. Hence, further research is needed in this field to find a more efficient and stable electrocatalyst.

IMPROVEMENT OF HER PERFORMANCE WITH VERTICAL HETEROSTRUCTURE

From DFT calculations and simulations, the basal plane and edges of 2D TMdCs can be tuned merely by varying the underlying substrate (Fig. 6a) [17, 32]. The ΔGH of TaS2/HfSe2 almost approaches the optimal value of 0 eV, being comparable or better than Pt/C, while that of the basal plane of pristine MoS2 can be tuned from 2.03 eV to 0.972 eV only by using NbS2 as the support substrate (MoS2/NbS2). This effect may be related to the linear correlation between hydrogen adsorption energies and adhesion energy difference (ΔEadh) before and after the hydrogen adsorption. The stronger the hydrogen atom adsorption is, the tighter the adhesion energy difference is. This implies that the supporting TMdCs substantially contributes to hydrogen adsorption in the overlying catalyst. Theoretical calculations further predict that formation of heterostructures by coupling of MoS2 with one S vacancy to other 2D structures such as hexagonal boron nitride (h-BN), graphene, MXene and MXene-OH, an optimal ΔGH ≈ 0 eV can be achieved [32]. Particularly, Ti2C-OH (hydroxide terminated-titanium carbide) is a desirable supporting layer for MoS2. The interfacial interaction (deq) between MoS2 and those supporting layers play an important role for determining the ΔGH. The optimal ΔGH is associated with the presence of the density of states located just below 琯LUS of Mo d orbitals when H is absorbed on the overlying MoS2 layer.

 

Fig. 6: Vertical heterostructure assembly (a) Schematic description of assembling 2D van der Waals heterostructures and the evolution of hydrogen on the basal plane of the overlying 2D TMDs (ref.17, 32). (b, c) Polarization curves and corresponding Tafel plots of stacked MoS2/graphene (ref. 33). (d, e) Polarization curves of graphene/h-BN heterostructure. Inset: Schematic illustration of defect free graphene decorated with small hBN sheets (ref. 34).

The hybridization of MoS2 with graphene was proved to be effective for improving the electrocatalytic HER efficiency. The extra high conductivity of graphene and its electrical coupling to MoS2 facilitate the rapid electron transfer from the electrode to the active sites of MoS2. Experimentally, Boandoh et al. constructed a heterostructure comprising a clean interface of monolayer MoS2 stacked with graphene (MoS2/Gr_Ul) (Fig. 6b, c) [33]. The ultraclean interface between the MoS2 and graphene facilitates faster transfer of electrons to active sites thereby promoting superior HER catalytic performance, whereas limited catalytic activity is observed when there are contaminants at the interface of the heterostructure (MoS2/Gr_Cl). As a result, the Tafel slope reaches ~ 73 mV/dec with a clean interface.

The stacking order in van der Waal heterostructures also influences HER performance. Bawari et al. prepared shear exfoliated graphene (SEG) on h-BN and vice versa [34]. From the polarization curves (Fig. 6d), a reduced onset potential was observed for SEG on the H-BN (~ 330 mV) compared to that for h-BN on SEG (~ 400 mV). Inspired by this result, SEG was replaced with reduced graphene oxide (rGO), which has various types of defects in its honeycomb lattice. Eventually, the onset potential was further shifted to a lower potential of ~ 220 mV (Fig. 6e), indicating that defective graphene on hBN further enhances the catalytic activity. DFT calculations reveal that the improved activity is attributed to the electrostatic induction effect as a result of the graphene/h-BN heterostructure. Due to the stacking of the graphene on hBN, there is an induction of partial positive charge on the carbon atoms at the interface of the heterostructure. This phenomenon causes the surface of graphene to be partially negative, facilitating the adsorption of hydrogen ions. Though these results for heterostructures were promising, it must be noted that research in this field is still in its infant stage. Therefore, more work is required since this opens a pathway to explore other modifications for enhancing HER activity by breaking away from the traditional defect induction methods.

VOLCANO PLOT

The Sabatier principle is a qualitative way of predicting the activity of catalysts. The principle states that to achieve high catalytic activity, the interaction between reactants and catalysts should neither be too strong nor too weak. If the interaction is too weak, there will be no reaction on the surface because it is difficult for the catalyst surface to bind to the reactants. If the interaction is too strong, then the reactant or product is difficult to desorb from the catalyst surface, which also lowers the catalytic activity [35].

 

Fig. 7: The volcano plot of the experimentally measured exchange current density (j0) versus the DFT calculated hydrogen binding energy of various phases of 2D TMDs compared with metallic Pt. Short dash lines (black) indicate the hypothetical j0 region of the TMDs.

Integration of this prediction into hydrogen evolution shows that adsorption of hydrogen is favored for some catalysts, preferably Pt, but not too much for others, thus giving rise to the so-called Volcano plot (Fig. 7). In particular, the basal planes of most 2D TMdCs, such as 2H-MoS2 show weak H bonds with large positive ΔGH* ≈ 2 eV, indicating limited activity from the Volmer reaction, and thus exhibiting poor activity for HER. However basal planes with chalcogen vacancies such as Vse-MoSe2 and VS-MoS2 have ΔGH* significantly reduced to ≈ 0.02 eV and ≈ 0 eV respectively, indicating that the chalcogen vacancies provide an effective hydrogen adsorption site for the Volmer reaction [20, 27]. With the phase transition of 2H- MoS2 semiconductor to 1T semimetal MoS2, the free energy of adsorption is close to 0.13 eV indicating that the 1T transformed surface is amenable for catalysis [36]. Additionally, the metallic 3R-NbS2 basal plane locates close to the peak of the volcano plot (ΔGH* close to thermoneutral, comparable with Pt), in good agreement with the Sabatier principle. The other members of group V metallic TMdCs like 2H-TaS2 [37] also have highly active basal planes that are inferior only to 3R-NbS2 among the studied candidates. Cases of heteroatom substitutional doping when Li was substituted into ReS2 (Li-ReS2) and vacancy (V) defects such as VRe-ReS2, also tunes the ΔGH* closer to 0 eV as indicated on the volcano plot [38].

SUMMARY AND PERSPECTIVE

In this review, we have highlighted some critical modifications of 2D monolayer TMdCs for enhancing the catalytic activity for highly efficient HER. The superiority of 2D monolayer TMdCs with maximally exposed active sites and ultra-short diffusion paths ensure high catalytic performance, which is even comparable to that of Pt catalysts. While significant progress has been made in the past two decades on improving the intrinsic catalytic activity of monolayer TMdCs, many challenges in the development of catalysts still remain: such as controlling the density of induced active sites, utilization of the entire basal plane of TMdCs, stabilization of the metallic TMdCs during HER and precise integration of computational results into experimental tests. We propose the following approaches to effectively engineer the surface chemistry of monolayer TMdCs to advance their catalytic activities:

a. Correlating the revealed active sites observed to the HER mechanism: Though it has widely been used to explain the hydrogen evolution phenomenon, further understanding is needed to associate the role these active sites play in hydrogen adsorption and evolution. For instance, in the metallic doping of monolayer MoS2, there could be ambiguity in the exact mechanism for catalysis. Since there is a potential for a metallic TMdCs formation, both the Volmer-Tafel and the Volmer-Heyrovsky mechanisms could play interchangeable roles in the evolution mechanism. Hence, there is a need for a more mechanistic approach to clarifying the HER process further.

b. Transferring computational knowledge of TMdCs heterostructures to experimental tests: While they have been predicted to possess close to zero value for Gibbs free hydrogen adsorption energy, no experimental work to the best of our knowledge has been done on the HER performance of TMdCs heterostructures. A combination of experimental tests and computational results is required to improve research efficiency in this novel approach. Further studies are needed to understand how the TMdC heterostructures could be utilized to tune the ΔGH for hydrogen adsorption of the overlying catalyst.

c. Finding ideal dopant and heterostructure with artificial intelligence (AI): In mimicking the HER performance of Pt, we propose the use of AI to design and study the functionality of Pt as an electrocatalyst. Clearly, to describe the activity of Pt needs a more sophisticated level of knowledge by considering all potential side reactions in the hydrogen evolution process. By using artificial intelligence, all data for the suitable electrocatalysis processes can sorted and analyzed to find the ideal dopant for TMdCs and the precise heterostructure for hydrogen evolution.

d. Study of the effect of electrode on monolayer TMdCs for HER: Finally, we propose an intensive study of the effect of substrates on the optimization of ΔGH of 2D monolayer TMdCs. In electrocatalysis, the selection of a supporting substrate is of vital importance since it promotes the injection and collection of charge carriers to the electrocatalyst. To date, very little research has been done on understanding the crucial role of substrates in tuning the ΔGH of 2D monolayer electrocatalysts. We believe that a critical look at the correlation between the substrates and HER activity will provide a new direction for designing electrodes that will ultimately maximize the catalytic performance of monolayer TMdCs.

AUTHOR INFORMATION

Corresponding Author: kikangkim@skku.edu
Authors Contributions: §Frederick Osei-Tutu Agyapong-Fordjour and Adofo Laud Anim contributed equally.
Notes: The authors declare no competing financial interest.

Acknowledgments: K.K.K. acknowledges support from the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning (2018R1A2B2002302).

References

[1] J. A. Turner Science, 305, 972-974 (2004).
[2] E. Santos et al., ChemPhysChem, 12, 2274-2279 (2011).
[3] C. G. Morales-Guio, L.-A. Stern, X. Hu, Chem. Soc. Rev., 43, 6555-6569 (2014).
[4] L. A. King et al., Nature Nanotechnology, 14, 1071-1074 (2019).
[5] L. Yu et al., Nature Comm. 10, 5106 (2019).
[6] D. Voiry et al., Nature Reviews Chemistry 2, 0105 (2018).
[7] D. Voiry, J. Yang, M. Chhowalla, Adv. Mater. 28, 6197-6206 (2016).
[8] J. Yuan et al., Adv. Mater. 27, 5605-5609 (2015).
[9] Y. Liu et al., Nature Energy, 2, 17127 (2017).
[10] J. Yang et al., Nature Mater. 18, 1309-1314 (2019).
[11] X. Chia et al., J.Mater. Chem. A., 4, 14241 (2016).
[12] T.F. Jaramillo, et al., Science 317, 100-102 (2007).
[13] H. Li et al., Adv. Mater., 28, 8945 (2016).
[14] G. Ye et al., Nano Lett. 16, 2, 1097-1103 (2016).
[15] H. Li et al., Nat. Mater., 15, 48-53 (2016).
[16] J. Zhu et al., Nature Comm. 10, 1348 (2019).
[17] S.H. Noh et al., J. Mater. Chem. A, 6, 20005-20014 (2018).
[18] S.J. Yun et al., ACS Nano, 9, 5, 5510-5519 (2015).
[19] J. Jeon et al., Nanoscale, 7, 1688 (2015).
[20] J. Lee et al., Nano Energy, 63, 103846 (2019).
[21] A.Y. Lu et al., Nature Nanotechnology 12,744 - 749 (2017).
[22] D. Er et al., Nano lett. 18, 3943-3949 (2018).
[23] J. Suh Nano Lett. 14, 6976 - 6982 (2014).
[24] W. Wu et al., Angew. Chem. Int. Ed., 58, 2029-2033 (2019).
[25] Y. Gong et al., Nature Mater. 13, 1135-1142 (2014).
[26] J. Miao et al., Sci. Adv., 1, 7, e1500259 (2015).
[27] L. Li et al., ACS Nano 13, 6824 - 6834 (2019).
[28] M. G. Walter et al., Chem Rev, 110, 6446-6463 (2010).
[29] H. Liang et al., Chem. Mater. 28, 5587-5591 (2016).
[30] L. Yang et al., Nanoscale, 7, 10490-10497 (2015).
[31] X. Hai et al., Nano Energy, 39,409-417 (2017).
[32] F. Ling et al., npj Computational Materials 20 (2019).
[33] S. Boandoh et al., ACS Appl. Mater. Interfaces, 11, 1 (2019).
[34] S. Bawari et al., Phys. Chem. Chem. Phys., 20, 15007-15014 (2018).
[35] J. K. N첩rskov et al., J. Electrochem. Soc., 152, J23. (2005).
[36] J. Seok et al., 2D Mater. 4 025061 (2017).
[37] J. Zhang et al., Materials Today, 25, 28-34 (2019).
[38] Y. Zhou et al., ACS Nano, 12, 4486 - 4493 (2018).

 

Agyapong-Fordjour Frederick Osei-Tutu is currently a PhD student at the Energy Science Department, Sungkyunkwan University, Suwon, South-Korea and his research focuses on 2D transition metal dichacogenides (TMdCs) for electrochemical application in water splitting.

Adofo Laud Anim is currently a Ph.D candidate at the Energy Science Department, Sungkyunkwan University, Suwon, South-Korea. His current research focuses on 2D material synthesis for energy harvesting applications.

Soo Min Kim is an Assistant Professor in the Department of Chemistry, Sookmyung Women's University, Korea. She received her M.S. and Ph.D. in SKKU Advanced Institute of Nano Technology in the year 2011 under the supervision of Prof. Young Hee Lee. After one-year postdoctoral work under the supervision of Prof. Jing Kong at Massachusetts Institute of Technology, USA and 7-years as a senior researcher in Korea Institute of Science and Technology, she joined the current position in 2020. Her research interests are the synthesis of single-crystal 2D materials and their applications.

Ki Kang Kim is an Associate Professor in the Department of Energy Science, Sungkyunkwan University (SKKU). He received his Ph.D in Physics from Sungkyunkwan University, Korea (SKKU), under the supervision of Prof. Young Hee Lee. He held a post-doctoral position at SKKU and Massachusetts Institute of Technology, USA for 4 years and as Assistant/Associate Professor at Dongguk University, Korea for 7 years. He is one of the pioneers in the growth of 2D materials. His group in SKKU and collaborators are currently working on developing novel synthesis techniques for diverse 2D materials and their applications such as HER.

 
AAPPS Bulletin        ISSN: 2309-4710
Copyright 짤 2018 Association of Asia Pacific Physical Societies. All Rights Reserved.
Hogil Kim Memorial Building #501 POSTECH, 67 Cheongam-ro, Nam-gu, Pohang-si, Gyeongsangbuk-do, 37673, Korea
Tel: +82-54-279-8663Fax: +82-54-279-8679e-mail: aapps@apctp.org