English

• The hydrogen evolution reaction (HER) is a key step in water splitting for hydrogen production and plays a crucial role in renewable energy storage and the development of the hydrogen economy [1-4]. With the increasing global demand for clean energy, the development of efficient, stable and low-cost electrocatalysts has become a major research focus [5-7]. Transition metal dichalcogenides (TMDs) have attracted significant attention in HER research due to their unique electronic structures and excellent catalytic activities [8-10]. Among them, molybdenum disulfide (MoS₂) exhibits platinum-like catalytic characteristics owing to its near-zero Gibbs free energy change (ΔG_H*) [11]. Experimental studies have shown that MoS₂ demonstrates high activity in HER [12,13], however, its catalytic efficiency is still limited by factors such as charge injection efficiency [14,15], interfacial effects [16], and the material's microstructure [17]. Therefore, enhancing the HER catalytic performance of MoS₂ remains a key research challenge.

  Optimizing charge injection into MoS₂ is an effective strategy to improve its catalytic performance. Due to its relatively large bandgap and limited intrinsic conductivity, the catalytic efficiency of MoS₂ is restricted by electron transport. Choosing an appropriate contact electrode to reduce interfacial Schottky barriers and enhance charge carrier mobility is a feasible approach [18-21]. In recent years, tungsten ditelluride (WTe₂) has emerged as a promising contact electrode for improving the HER performance of MoS₂ [22]. WTe₂ exhibits high electrical conductivity [23], excellent interfacial contact properties, and strong spin-orbit coupling effects [24], which facilitates efficient electron injection from an external circuit into MoS₂, thereby enhancing catalytic efficiency. However, the controlled synthesis of WTe₂ remains challenging due to the low chemical reactivity of tellurium (Te) with tungsten (W) sources, which makes it difficult to grow WTe₂. Besides, the low phase stability of WTe₂ at high temperatures can lead to phase transitions or structural defects, which could degrade its performance as a contact electrode.

  To address these challenges, chemical vapor deposition (CVD) has been identified as a promising solution [25]. CVD can realize precise growth control and scalability, which enables effective control over the thickness, morphology and structural stability of WTe₂ [26-30]. Based on this, we report a salt-assisted CVD method for synthesizing high-quality WTe₂. By employing different alkali metal halide salts, we achieved morphology control of WTe₂, obtaining nanoribbons and hexagonal nanosheets. By investigating the effects of growth time and salt dosage, we map out the complete morphology evolution process. Based on the fundamental principles of salt-assisted growth, intermediate metal oxyhalides follow a vapor-solid-solid (VSS) growth mechanism, whereas molten salts derived from metal oxides follow a vapor-liquid-solid (VLS) pathway. Building on these insights, we propose a corresponding growth mechanism to clarify the distinct roles of various alkali metal salts in governing the morphological evolution of WTe₂. Additionally, we examine the stability of WTe₂ with different morphologies in air and observe that thinner samples are more prone to oxidation and degradation, suggesting that vacuum storage is a more effective preservation method. Subsequently, we evaluate the electrical properties of the synthesized WTe₂ samples and find that the nanosheets exhibit higher conductivity than the nanoribbons. Based on the excellent electrical conductivity and interfacial contact properties of WTe₂, we further construct the WTe₂−MoS₂ and graphene-MoS₂ heterostructures to investigate the effect of WTe₂ as a contact electrode on the catalytic behavior of MoS₂. Experimental results show that the WTe₂−MoS₂ heterostructure exhibits a Tafel slope of 111.57 mV/dec and an overpotential of 298 mV at −10 mA/cm², which is significantly better than that of graphene as a contact electrode. Meanwhile, we systematically investigate the effect of WTe₂ thickness on HER performance and evaluate the electrochemical durability and structural stability of the device, further validating the effectiveness of WTe₂ as a contact electrode for enhancing the HER activity of MoS₂. This study not only presents a novel strategy for enhancing the HER catalytic performance of MoS₂ but also provides an experimental basis for the controlled growth of high-quality WTe₂, laying the groundwork for the broader application of two-dimensional (2D) materials in energy catalysis.

  To prepare atomically thin WTe₂, a salt-assisted CVD method was employed, as illustrated in Fig. 1a. With sodium chloride (NaCl) salt, WTe₂ exhibits nanoribbon morphology, whereas, with potassium chloride (KCl) salt, the WTe₂ morph into hexagonal nanosheets or dendrites. The addition of alkali metal halide salt (e.g., NaCl or KCl) can significantly lower the melting point of metal oxides, thereby enhancing the precursor concentration for WTe₂ growth. This strategy has been widely used in the synthesis of other two-dimensional TMDs such as MoS₂ [31] and WS₂ [32]. Notably, the type of salt plays a critical role in controlling the morphology of WTe₂ and this mechanism will be discussed in detail later. Optical microscopy images (OM) of WTe₂ samples are presented in Figs. 1b and c, clearly showing two distinct morphologies: Nanoribbons and hexagonal nanosheets. To evaluate the thickness of the CVD-grown 2D WTe₂, atomic force microscopy (AFM) characterization was performed (Figs. 1d and e). The thickness of the WTe₂ nanoribbon is measured to be approximately 4.75 nm, while the hexagonal nanosheet has a thickness of 3.43 nm. Besides, we have also measured the thickness of other nanoribbons and hexagonal nanosheets in SiO₂/Si substrate, which are available in Fig. S1 (Supporting information). Their thicknesses are all < 10 nm. To confirm the composition and crystal structure of the obtained samples, Raman spectroscopy was conducted as shown in Fig. 1f. Seven typical Raman peaks are observed in both the nanoribbon and hexagonal nanosheet samples, located at 82.4, 88.7, 108.8, 115.9, 131.5, 161.3 and 209.0 cm^-1, which correspond to the A₁², B₁¹⁰, A₂⁴, A₂³, A₄¹, A₁⁷ and A₁⁹ modes of 1T'-WTe₂, respectively [22]. These Raman peaks can be categorized into two types: interlayer and intralayer vibration modes. Specifically, 88.7, 108.8, 115.9 and 131.5 cm^-1 are associated with intralayer vibrations, while 82.4, 161.3 and 209.0 cm^-1 correspond to interlayer vibrations. The corresponding top and side views of the atomic structure of 1T'-WTe₂ are presented in Fig. 1g. To further verify the spatial uniformity of the WTe₂ samples, Raman mappings were performed (Figs. 1h and i). Our results demonstrate that the WTe₂ samples exhibit excellent spatial homogeneity with no detectable impurities.

  Figure 1

  

  Figure 1.  Synthesis and characterizations of CVD-grown WTe₂. (a) Schematic diagram of the CVD growth process. (b, c) OM images of as-grown WTe₂ nanoribbon and hexagonal WTe₂, respectively. (d, e) AFM images of as-grown WTe₂ nanoribbon and hexagonal WTe₂, respectively. Inset: corresponding height profile analyses. (f) Raman spectra of as-grown WTe₂. (g) Top and side views of the crystal structure of 1T'-WTe₂. (h, i) Raman intensity mappings of WTe₂ nanoribbon and hexagonal WTe₂, respectively.

  DownLoad: Full-Size Img PowerPoint

  The chemical states and elemental compositions of obtained samples were characterized by X-ray photoelectron spectroscopy (XPS), with the results being shown in Figs. 2a and b. The Te 3d spectrum can be well-fitted by two peaks located at 583.7 eV (Te 3d_3/2) and 573.3 eV (Te 3d_5/2), which are associated with W-Te bonds. This confirms the successful formation of WTe₂ and demonstrates that CVD process effectively overcomes the low chemical reactivity of Te. Additionally, two oxidation peaks (binding energies at 585.4 and 575.1 eV) are clearly observed in the Te 3d spectrum, attributed to Te-O bonding. These indicate the presence of a small amount of oxidized Te in the material, which may be related to the inherent instability of WTe₂. The binding energies at ≈33.7 eV and ≈31.5 eV are attributed to W 4f_5/2 and W 4f_7/2, respectively, confirming the presence of W-Te bonds. Meanwhile, the other two peaks are attributed to W-O bonds. In addition, the X-ray diffraction (XRD) result of corresponding samples deposited on the SiO₂/Si substrate confirms the formation of WTe₂ (Fig. 2c). The main diffraction peaks in the XRD spectrum are consistent with those of orthorhombic WTe₂, with peaks at 13.00°, 38.79° and 52.35° corresponding to the (002), (006) and (008) crystal planes, respectively. Notably, the diffraction peak at 13.00° exhibits the strongest intensity, indicating that the WTe₂ surface is parallel to the SiO₂/Si substrate. Therefore, WTe₂ with good crystalline orientation can be achieved through salt-assisted atmospheric-pressure CVD. To understand the crystalline structure and quality of CVD-derived atomically thin WTe₂, the transferred samples were characterized by transmission electron microscopy (TEM). The low-magnification TEM images of the WTe₂ nanosheet are shown in Fig. 2d. The well-defined morphologies and highly transparent properties indicate the perfect crystalline qualities and ultrathin features. Fig. 2e clearly exhibits lattice fringes with spacings of 0.627 and 0.33 nm, which are attributed to the (010) and (101) crystal planes, respectively. The selected area electron diffraction (SAED) can effectively reflect the crystal quality over a large area. As shown in Fig. 2f, the diffraction spots present regular shape, like orthogonal atomic lattice. This result is highly consistent with the structural characteristics of the orthorhombic WTe₂, thus confirming that WTe₂ has good crystallinity.

  Figure 2

  

  Figure 2.  Chemical states and atomic structure of CVD-grown WTe₂. (a, b) XPS spectrum of as-grown WTe₂. (c) XRD pattern of as-grown WTe₂. (d) Low-magnification TEM image of transferred WTe₂ nanosheet. (e) Atomic-resolution TEM image of transferred WTe₂ nanosheet. (f) Corresponding SAED pattern of WTe₂.

  DownLoad: Full-Size Img PowerPoint

  To investigate the growth mode of WTe₂ in-depth, we systematically adjusted the growth conditions, focusing on the effects of alkali metal halide types and their quantities, growth time and the introduction of H₂ on the morphology of WTe₂. Experimental results indicate that the type of salt and the growth time significantly influence the morphology of WTe₂. Figs. 3a and b illustrate the changes in the size and morphology of WTe₂ with varying NaCl amounts and the effect of growth time when using KCl as a flux agent. Fig. 3a shows that without NaCl, the product on the silicon substrate appeared as granules, with no deposition of 2D material. When 50 mg of NaCl was added as flux agent, WTe₂ nanoribbons with lengths up to 100 µm were obtained. As the NaCl concentration increased, WTe₂ gradually exhibited a thin-film growth trend. When the NaCl concentration reached 100 mg, a large-area polycrystalline film formed on the substrate. However, the film exhibited cracks and grain boundaries, resulting in relatively low quality. With a fixed NaCl concentration, extending the growth time leaded to the gradual thickening of WTe₂ nanoribbons while maintaining their morphology (Fig. S2 in Supporting information). Previous studies have shown that the key to WTe₂ growth is the controlled introduction of hydrogen gas (H₂) during the reaction process. In this study, we also investigated the effect of H₂ reaction time on the growth process by controlling this single variable (Fig. S3 in Supporting information). The results show that continuous introduction of H₂ produced elongated WTe₂ nanoribbons, with lengths reaching 50 µm and widths below 3 µm. When H₂ was introduced briefly during the reaction and then turned off at the end, 2D WTe₂ structures were formed, however, etching effects were observed. These results indicate that H₂ reaction time directly affects material morphology, playing a dual role in both promoting growth and inducing etching in the CVD process of WTe₂. When KCl was used as a flux agent, the morphology of WTe₂ evolved from nanoribbons to dendritic structures and eventually to hexagonal nanosheets as the growth time increased (Fig. 3b). Fig. S4 (Supporting information) displays thin-layer samples with several branches. The six-branched WTe₂ exhibits high symmetry, with each branch forming a 60° angle with its neighbors. In the five-branched WTe₂, one angles is 120°, while the remaining angles are all 60°. Subsequently, we further confirmed the composition of dendritic WTe₂ using Raman spectroscopy. The specific Raman peaks of Fig. S5a (Supporting information) are consistent with that of 1T'-WTe₂ in Fig. 1f. Raman mappings of five and six-branched WTe₂, with peak fitting at ~162 cm^-1 and ~211 cm^-1, reveals distinct characteristics. The intensity of the A₁⁷ peak gradually increases from the center toward the branch edges, while the intensity of the A₁⁹ characteristic peak varies in different directions (Figs. S5b-e in Supporting information). The corresponding AFM and scanning electron microscope (SEM) characterization results of six-branched WTe₂ are shown in Fig. S6 (Supporting information). The sample thickness increased gradually from the edges toward the center, with a maximum thickness of 32.08 nm and an edge thickness of only 11.87 nm. Based on these observations above, we assume that salts play a crucial role in the assisted growth of WTe₂.

  Figure 3

  

  Figure 3.  Growth mechanism of CVD-grown WTe₂. (a) Influence of NaCl mass on the WTe₂ morphology. (b) Influence of growth time (with KCl) on the WTe₂ morphology (c) Schematic diagram of growth mechanism about alkali metal halide types leading to different WTe₂ morphologies.

  DownLoad: Full-Size Img PowerPoint

  Alkali metals (Li, Na, K) are essential components of alkali metal halide salts. Although NaCl and KCl differ only in their alkali metal components, they lead to distinctly different morphologies in the obtained WTe₂ samples. Herein, we propose the possible growth mechanism as shown in Fig. 3c. Firstly, alkali metal halide salts react with tungsten oxide powder to form two types of intermediates, WO_xCl_y and A_xWO_y (A = K or Na) [33,34]. These intermediates significantly reduce the melting point of WO₃, thereby facilitating its reaction with Te powder. The WO_xCl_y intermediate, formed by the reaction of halogens with metal oxides, is a metal oxyhalide compound that reacted with Te via VSS mode [31,35,36]. Since KCl and NaCl share the same halide ions, their roles in the VSS mechanism should be identical. Na_xWO_y and K_xWO_y are tungstates formed by the reaction of alkali metals with metal oxides and their reaction with Te powder follows VLS mode [31,35,36]. However, Na₂WO₄ has a significantly lower melting point than K₂WO₄ and exhibits greater solubility [37]. Consequently, in the VLS mode, Na_xWO_y has higher epitaxial mobility than K_xWO_y, making it less likely to accumulate on the substrate and more inclined to form WTe₂ nanoribbon upon reacting with Te. In contrast, K_xWO_y has a lower epitaxial mobility and a smaller molten mass than Na_xWO_y. As a result, K_xWO_y needs to accumulate to a certain concentration on the substrate before gradually diffusing outward, ultimately forming dendritic WTe₂ structures, which further evolve into hexagonal nanosheets. Notably, we find that this salt-assisted CVD method exhibits excellent universality, enabling the fabrication of nanoribbon-like, dendritic and hexagonal WTe₂ samples on different substrates such as sapphire and mica (Figs. S7 and S8 in Supporting information). This lays a solid foundation for further exploration of controlled morphology growth and the applications of WTe₂. To further validate the influence of alkali metal halide type on WTe₂ morphology, we introduced lithium chloride (LiCl) as an additional salt under the same growth conditions. As shown in Fig. S9 (Supporting information), the resulting WTe₂ structures remain as elongated nanoribbons, consistent with those synthesized using NaCl. Optical and AFM characterizations reveal narrow, flat nanoribbons with uniform thickness and Raman spectra confirms the formation of good crystalline WTe₂. The preserved WTe₂ nanoribbon morphology supports our proposed VSL growth mechanism. Compared with KCl, LiCl exhibits higher solubility and forms an intermediate (Li₂WO₄) with a lower melting point (742 ℃), which is close to that of Na₂WO₄ (698 ℃) but significantly lower than K₂WO₄ (921 ℃). This favors the generation of volatile intermediates necessary for ribbon-like growth. The consistent morphological trends observed across Li⁺, Na⁺, K⁺ salts highlight the critical role of alkali halide selection in controlling the structural evolution of WTe₂.

  To investigate the electrical properties of the synthesized WTe₂ samples, two-terminal WTe₂ devices were fabricated on a SiO₂/Si substrate. The Si substrate can modulate the electrostatic potential of the WTe₂ device, similar to a field-effect transistor (Figs. 4a and d). The output characteristics of the WTe₂ nanoribbon exhibit a linear and symmetric relationship, indicating the formation of good ohmic contacts. When the source-drain voltage (V_ds) was applied to 0.35 V, the source-drain current (I_ds) was 35 µA while when V_ds = 1.0 V, I_ds increased to 70 µA, demonstrating the good conductivity of WTe₂ nanoribbon (Fig. 4b). In the transfer curve (Fig. 4c), regardless of the gate voltage variation (from −60 V to 60 V), the current remained constant at 70 µA. This indicates that the gate voltage has no significant effect on the source-drain current, suggesting the metallic behavior of WTe₂ nanoribbons. As for the hexagonal WTe₂ nanosheet, the electrical results exhibit its metallic properties and demonstrate a higher current than the WTe₂ nanoribbon. The output curve shows that with a constant gate voltage, the I_ds exhibits a linear relationship with the V_ds (Fig. 4e). This indicates that the transistor operated in the linear region, where its output current can be controlled by the input voltage and the current variation was predictable. The output curve passed through the origin (0, 0), indicating that the transistor's leakage current can be ignored. This suggests that in the absence of an applied voltage, there is no leakage current, demonstrating good turn-off characteristics. Fig. 4f presents the transfer curve of hexagonal WTe₂ nanosheet, under V_ds = 1.0 V, V_ds = 0.5 V, and V_ds = 0.2 V, the I_ds remained unchanged regardless of the gate voltage, forming a straight line with negligible hysteresis.

  Figure 4

  

  Figure 4.  Electrical properties of WTe₂ devices and their air stability. (a) OM image of the WTe₂ nanoribbon device and (b) its corresponding output and (c) transfer characteristic curves. (d) OM image of the hexagonal WTe₂ nanosheet device and (e) its corresponding output and (f) transfer characteristic curves. (g) OM images of as-grown WTe₂ sample after exposure to air for different time periods. (h) Corresponding Raman spectra of as-grown WTe₂ sample after exposure to air for different time periods.

  DownLoad: Full-Size Img PowerPoint

  Although WTe₂ exhibits excellent electrical conductivity, thin-layer WTe₂ is highly sensitive to the environment and has poor stability, with oxidation-induced decomposition significantly reducing its conductivity. We conducted relevant studies on this issue. Figs. 4g and h present the changes in optical images and Raman spectra of WTe₂ thin films in the same region after different durations of air exposure. The as-grown WTe₂ exhibits six characteristic peaks in the Raman spectrum (85, 108, 118, 134, 162 and 214 cm^-1). After 2 h of air exposure, the intensity of the 162 and 214 cm^-1 peaks decreased significantly. After 6.5 h, these two primary peaks almost completely disappeared. This phenomenon is consistent with the two distinct Te-O oxidation peaks observed in the XPS spectra of the Te 3d, further confirming the instability of WTe₂. WTe₂ with different morphologies underwent varying degrees of decomposition after air exposure for several minutes to several days. Light pink nanosheets rapidly decomposed within 2 min, with their sharp edges becoming blurred, after which the color stabilizes. WTe₂ nanoribbons exhibited noticeable damage and material loss under an optical microscope after several days of air exposure (Fig. S10 in Supporting information). Hexagonal nanosheets, when stored in a vacuum-sealed environment within a vacuum drying chamber for 57 days, retained the same morphology as when they were first grown with clear boundaries, no significant color change, and no visible signs of oxidation (Fig. S11 in Supporting information). This indicates that vacuum storage significantly reduces the oxidation rate of the samples, which is beneficial for device fabrication and the study of the intrinsic properties of the material.

  Based on the unique metallic properties of WTe₂ and its excellent electrical transport capability, we selected WTe₂ as the electrode material to replace the traditional gold electrode, aiming to construct on-chip electrochemical devices and optimize HER catalytic performance of MoS₂. Meanwhile, graphene was introduced as a comparative electrode to compare the influence of two different electrode materials on the HER performance of MoS₂. Due to the poor adhesion of gold on the SiO₂/Si substrate, a 5 nm thick Cr wetting layer was introduced in the experiment to construct a microreactor and ensure stable electrical contact with MoS₂. During the experiment, only the basal plane of MoS₂ was exposed to the electrolyte for HER performance characterization. The configurations of heterostructures are illustrated in the schematic diagram (Fig. 5a) and the optical image (Figs. 5b and c). The on-chip electrochemical device construction process of specific WTe₂−MoS₂ heterostructure is shown in Fig. S12 (Supporting information).

  Figure 5

  

  Figure 5.  The effect of WTe₂ as electrode material on HER activity of MoS₂ (a) Cross-sectional schematic view of configurations of heterostructures. (b) OM image of graphene-MoS₂ heterostructure with MoS₂ basal plane exposed. (c) OM image of WTe₂−MoS₂ heterostructure with MoS₂ basal plane exposed. (d) Polarization curves of graphene-MoS₂ and WTe₂−MoS₂ heterostructures for the HER in 0.5 mol/L H₂SO₄. (e) Their corresponding Tafel plots of (d).

  DownLoad: Full-Size Img PowerPoint

  The experimental results further verify the superior performance of the WTe₂ electrode. Figs. 5d and e present the polarization curves and corresponding Tafel slopes of the WTe₂−MoS₂ and graphene-MoS₂ heterostructures. At the current density of −10 mA/cm², the overpotential of the WTe₂−MoS₂ heterostructure is 298 mV with a Tafel slope of 111.57 mV/dec, whereas the graphene-MoS₂ heterostructure exhibits a significantly higher overpotential of 491 mV and a Tafel slope of 157.5 mV/dec. This result indicates that WTe₂, as a contact electrode, can significantly enhance the HER catalytic activity of MoS₂, outperforming graphene electrodes.

  This performance enhancement can be attributed to the excellent electrical conductivity of WTe₂ and its effective improvement of electron injection and transport in MoS₂. Additionally, the theoretical work functions of monolayer MoS₂ and WTe₂ are 4.36 and 4.5 eV, respectively, with relatively close Fermi levels [23,38,39]. To experimentally determine the energy level alignment within the WTe₂−MoS₂ heterostructure, we performed kelvin probe force microscopy (KPFM) measurements on both materials under our specific growth conditions. As shown in Fig. S13 (Supporting information), WTe₂ and MoS₂ exhibit measured surface potentials of 223 mV and 131 mV, respectively, while the standard gold reference (Au) shows −105 mV. The work functions of WTe₂ and MoS₂ were calculated using the following equation:

  $ e V_{\text {cpd }}=\phi_{\text {tip }}-\phi_{\text {sample }} $

  (1)

  In which, V_cpd represents contact potential difference (CPD) between the measurement probe and the sample, ɸ_tip represents the work function of the probe, and ɸ_sample represents the work function of the sample. Given that the work function of Au is 5.1 eV, the work functions of WTe₂ and MoS₂ were estimated to be 4.77 and 4.86 eV, respectively. This relatively small difference results in near Fermi level alignment and a low Schottky barrier at the heterointerface, facilitating interfacial electron transport, improving electron transfer efficiency and consequently enhancing the electrocatalytic activity of MoS₂. The experimental results suggest that charge injection efficiency plays a crucial role in enhancing catalytic activity. Through this strategy, the inherently low catalytic activity of the TMDs basal plane is activated, providing a new research direction for further optimizing HER performance.

  To further investigate the effect of contact electrode thickness on the electrocatalytic performance of WTe₂−MoS₂ heterostructures, we fabricated a series of devices using WTe₂ nanoribbons with controlled thicknesses of 5.8, 12.2, 27.2, 38.6 and 56.6 nm, while keeping the exposed MoS₂ geometric configuration consistent. The optical microscope images and corresponding AFM height profiles of the devices are shown in Fig. S14 (Supporting information), confirming the uniformity in lateral dimensions and the variation in vertical thickness of WTe₂ contacts.

  HER polarization curves and Tafel plots (Fig. S15 in Supporting information) reveal a consistent enhancement in catalytic activity with increasing WTe₂ thickness. Specifically, both the overpotential at −100 mA/cm² and the Tafel slope decrease monotonically as the WTe₂ thickness increases, suggesting more favorable charge transfer kinetics. Since only MoS₂ is in direct contact with the electrolyte, the observed performance improvement is attributed to the interfacial properties modulated by WTe₂ thickness. Thicker WTe₂ layers provide enhanced structural integrity, reduced surface roughness and improved interface conformity during the dry transfer process. These factors contribute to a cleaner van der Waals interface and more efficient charge injection from the WTe₂ electrode into the MoS₂ basal plane. In contrast, ultrathin WTe₂ flakes (e.g., 5.8 nm) are more susceptible to oxidation or transfer-induced defects, which may introduce interfacial barriers or disrupt the electronic coupling.

  No performance degradation or saturation is observed within the studied thickness range, indicating that WTe₂ thickness is a key parameter for optimizing interfacial charge transport in the heterostructure. These results emphasize the importance of contact engineering in on-chip electrochemical devices, particularly when employing 2D metallic contacts such as WTe₂.

  To evaluate the electrochemical and structural stability of the WTe₂−MoS₂ heterostructure, 500 cycles cyclic voltammetry (CV) test was conducted in 0.5 mol/L H₂SO₄ within the potential range of 0.2~−0.25 V vs. RHE on the on-chip electrochemical device. As shown in Fig. S16 (Supporting information), the device maintains relatively stable HER performance throughout the cycling process, with only a slight decrease in current density, indicating good durability under repetitive operation. In addition, optical characterizations were conducted before and after HER testing. As shown in Figs. S17a and b (Supporting information), the optical images of the same device reveal that the top MoS₂ layer undergoes partial degradation after cycling, while the underlying WTe₂ remains intact and morphologically stable. Raman spectra acquired at the marked positions (blue: before HER; orange: after HER) show that the characteristic MoS₂ peaks (A_1g and E_2g¹) decrease significantly in intensity after HER, whereas WTe₂-related peaks remain well preserved (Fig. S17c in Supporting information). These results confirm the robust electrochemical performance and structural resilience of the WTe₂−MoS₂ heterostructure, highlighting WTe₂ as a stable contact layer during HER operation.

  In this study, we have successfully developed a salt-assisted CVD method for synthesizing high-quality WTe₂ with controlled morphologies. By utilizing different alkali metal halide salts (NaCl, KCl and LiCl), we achieve precise morphology control, producing both nanoribbons and hexagonal nanosheets. Systematic investigations into growth time, salt dosage, and salt types reveal the underlying growth mechanisms and demonstrate the versatility of this approach. Electrical measurements confirm that hexagonal nanosheets exhibit higher conductivity than nanoribbons. Furthermore, we constructed WTe₂−MoS₂ heterostructures and demonstrated their superior HER performance compared to graphene-based electrodes. The WTe₂−MoS₂ heterostructure achieves a Tafel slope of 111.57 mV/dec and an overpotential of 298 mV at −10 mA/cm² outperforming graphene electrodes due to the excellent conductivity of WTe₂ and improved electron injection and transport in MoS₂. Besides, we further evaluated the HER-related performance of our system by investigating the effect of WTe₂ thickness and conducting durability and structural stability tests, confirming improved interfacial charge transport and acceptable long-term stability of the WTe₂−MoS₂ heterostructure. This study provides a novel approach for enhancing HER performance through the controlled growth of WTe₂ and its application as a contact electrode. Our findings offer valuable insights into the synthesis of high-quality WTe₂ and expand the potential applications of two-dimensional materials in energy catalysis, paving the way for further advancements in hydrogen production technologies.

  Declaration of competing interest

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

  CRediT authorship contribution statement

  Zhuojun Duan: Writing – original draft, Data curation. Peiyue Jin: Writing – review & editing, Writing – original draft, Data curation. Houying Xing: Writing – review & editing. Jian Chen: Writing – review & editing. Yueting Yang: Writing – review & editing. Yawen Tan: Writing – review & editing. Song Liu: Writing – review & editing, Supervision, Project administration.

  Acknowledgment

  The work gratefully acknowledges financial support from the National Natural Science Foundation of China (No. 22175060).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111917.

  
  
• 1. [1]

     A.B. Laursen, S. Kegnæs, S. Dahl, I. Chorkendorff, Energy Environ. Sci. 5 (2012) 5577–5591. doi: 10.1039/c2ee02618j

  2. [2]

     J. Staszak-Jirkovský, C.D. Malliakas, P.P. Lopes, et al., Nat. Mater. 15 (2016) 197–203. doi: 10.1038/nmat4481

  3. [3]

     J. Zhu, L. Hu, P. Zhao, et al., Chem. Rev. 120 (2020) 851–918. doi: 10.1021/acs.chemrev.9b00248

  4. [4]

     M. Chen, N. Kitiphatpiboon, C. Feng, et al., eScience 3 (2023) 100111. doi: 10.1016/j.esci.2023.100111

  5. [5]

     J. Joo, T. Kim, J. Lee, et al., Adv. Mater. 31 (2019) 1806682. doi: 10.1002/adma.201806682

  6. [6]

     M. Wang, L. Zhang, Y. He, H. Zhu, J. Mater. Chem. A 9 (2021) 5320–5363. doi: 10.1039/d0ta12152e

  7. [7]

     H. Chen, Y. Zhou, W. Guo, B.Y. Xia, Chin. Chem. Lett. 33 (2022) 1831–1840. doi: 10.1016/j.cclet.2021.09.034

  8. [8]

     Q. Lu, Y. Yu, Q. Ma, et al., Adv. Mater. 28 (2016) 1917–1933. doi: 10.1002/adma.201503270

  9. [9]

     Q. Fu, J. Han, X. Wang, et al., Adv. Mater. 33 (2021) 1907818. doi: 10.1002/adma.201907818

  10. [10]

      A. Mondal, A. Vomiero, Adv. Funct. Mater. 32 (2022) 2208994. doi: 10.1002/adfm.202208994

  11. [11]

      B. Hinnemann, P.G. Moses, J. Bonde, et al., J. Am. Chem. Soc. 127 (2005) 5308–5309. doi: 10.1021/ja0504690

  12. [12]

      J. Xu, G. Shao, X. Tang, et al., Nat. Commun. 13 (2022) 2193. doi: 10.1038/s41467-022-29929-7

  13. [13]

      Y. Cao, ACS Nano 15 (2021) 11014–11039. doi: 10.1021/acsnano.1c01879

  14. [14]

      D. Voiry, R. Fullon, J. Yang, et al., Nat. Mater. 15 (2016) 1003–1009. doi: 10.1038/nmat4660

  15. [15]

      Y. Zhou, J.L. Silva, J.M. Woods, et al., Adv. Mater. 30 (2018) 1706076. doi: 10.1002/adma.201706076

  16. [16]

      T.F. Jaramillo, K.P. Jørgensen, J. Bonde, et al., Science 317 (2007) 100–102. doi: 10.1126/science.1141483

  17. [17]

      L. Wang, X. Liu, J. Luo, et al., Angew. Chem. Int. Ed. 56 (2017) 7610–7614. doi: 10.1002/anie.201703066

  18. [18]

      M. Li, X. Zhang, Z. Zhang, et al., Nano Res. 17 (2024) 10162–10169. doi: 10.1007/s12274-024-6942-5

  19. [19]

      S. Ghods, H. Lee, J.H. Choi, et al., ACS Nano 18 (2024) 26192–26200.

  20. [20]

      J. Meng, C. Lee, Z. Li, Sci. Bull. 69 (2024) 1342–1352. doi: 10.1016/j.scib.2024.03.003

  21. [21]

      M. Du, D. Li, S. Liu, J. Yan, Chin. Chem. Lett. 34 (2023) 108156. doi: 10.1016/j.cclet.2023.108156

  22. [22]

      Y. Zhou, J.V. Pondick, J.L. Silva, et al., Small 15 (2019) 1900078. doi: 10.1002/smll.201900078

  23. [23]

      M.J. Mleczko, R.L. Xu, K. Okabe, et al., ACS Nano 10 (2016) 7507–7514. doi: 10.1021/acsnano.6b02368

  24. [24]

      P. Li, W. Wu, Y. Wen, et al., Nat. Commun. 9 (2018) 3990. doi: 10.1038/s41467-018-06518-1

  25. [25]

      Y. Zhang, Y. Yao, M.G. Sendeku, et al., Adv. Mater. 31 (2019) 1901694. doi: 10.1002/adma.201901694

  26. [26]

      J. Zhou, F. Liu, J. Lin, et al., Adv. Mater. 29 (2017) 1603471. doi: 10.1002/adma.201603471

  27. [27]

      M. Xu, B. Tang, Y. Lu, et al., J. Am. Chem. Soc. 143 (2021) 18103–18113. doi: 10.1021/jacs.1c06786

  28. [28]

      Z. Zhao, F. Dong, Y. Wang, et al., Nanoscale 15 (2023) 11955–11962. doi: 10.1039/d3nr00849e

  29. [29]

      Y. Xiao, M. Zhou, J. Liu, et al., Chin. J. Chem. 38 (2020) 709–713. doi: 10.1002/cjoc.202000079

  30. [30]

      X. Kong, C. Fu, V. Gladkikh, F. Ding, SmartMat 4 (2023) e1152. doi: 10.1002/smm2.1152

  31. [31]

      S. Li, Y.C. Lin, W. Zhao, et al., Nat. Mater. 17 (2018) 535–542. doi: 10.1038/s41563-018-0055-z

  32. [32]

      J. Zhou, J. Lin, X. Huang, et al., Nature 556 (2018) 355–359. doi: 10.1038/s41586-018-0008-3

  33. [33]

      C. Xie, P. Yang, Y. Huan, et al., Dalton Trans. 49 (2020) 10319–10327. doi: 10.1039/d0dt01561j

  34. [34]

      W. Han, K. Liu, S. Yang, et al., Sci. China Chem. 62 (2019) 1300–1311. doi: 10.1007/s11426-019-9525-y

  35. [35]

      S. Li, iScience 24 (2021) 103229. doi: 10.1016/j.isci.2021.103229

  36. [36]

      H.R. Rasouli, N. Mehmood, O. Çakıroğlu, T.S. Kasırga, Nanoscale 11 (2019) 7317–7323. doi: 10.1039/c9nr00614a

  37. [37]

      D. Li, S. Yoshida, B. Siritanaratkul, et al., ACS Catal. 11 (2021) 14237–14248. doi: 10.1021/acscatal.1c04206

  38. [38]

      J.M. Woods, J. Shen, P. Kumaravadivel, et al., ACS Appl. Mater. Interfaces 9 (2017) 23175–23180. doi: 10.1021/acsami.7b04934

  39. [39]

      S.Y. Lee, U.J. Kim, J. Chung, et al., ACS Nano 10 (2016) 6100–6107. doi: 10.1021/acsnano.6b01742

• 

  Figure 1  Synthesis and characterizations of CVD-grown WTe₂. (a) Schematic diagram of the CVD growth process. (b, c) OM images of as-grown WTe₂ nanoribbon and hexagonal WTe₂, respectively. (d, e) AFM images of as-grown WTe₂ nanoribbon and hexagonal WTe₂, respectively. Inset: corresponding height profile analyses. (f) Raman spectra of as-grown WTe₂. (g) Top and side views of the crystal structure of 1T'-WTe₂. (h, i) Raman intensity mappings of WTe₂ nanoribbon and hexagonal WTe₂, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Chemical states and atomic structure of CVD-grown WTe₂. (a, b) XPS spectrum of as-grown WTe₂. (c) XRD pattern of as-grown WTe₂. (d) Low-magnification TEM image of transferred WTe₂ nanosheet. (e) Atomic-resolution TEM image of transferred WTe₂ nanosheet. (f) Corresponding SAED pattern of WTe₂.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Growth mechanism of CVD-grown WTe₂. (a) Influence of NaCl mass on the WTe₂ morphology. (b) Influence of growth time (with KCl) on the WTe₂ morphology (c) Schematic diagram of growth mechanism about alkali metal halide types leading to different WTe₂ morphologies.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Electrical properties of WTe₂ devices and their air stability. (a) OM image of the WTe₂ nanoribbon device and (b) its corresponding output and (c) transfer characteristic curves. (d) OM image of the hexagonal WTe₂ nanosheet device and (e) its corresponding output and (f) transfer characteristic curves. (g) OM images of as-grown WTe₂ sample after exposure to air for different time periods. (h) Corresponding Raman spectra of as-grown WTe₂ sample after exposure to air for different time periods.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  The effect of WTe₂ as electrode material on HER activity of MoS₂ (a) Cross-sectional schematic view of configurations of heterostructures. (b) OM image of graphene-MoS₂ heterostructure with MoS₂ basal plane exposed. (c) OM image of WTe₂−MoS₂ heterostructure with MoS₂ basal plane exposed. (d) Polarization curves of graphene-MoS₂ and WTe₂−MoS₂ heterostructures for the HER in 0.5 mol/L H₂SO₄. (e) Their corresponding Tafel plots of (d).

  下载: 全尺寸图片 幻灯片
•