Colloidal Synthesis of MoSe₂/WSe₂ Heterostructure Nanoflowers via Two-Step Growth

Abstract

The ability to control the active edge sites of transition metal dichalcogenides (TMDs) is crucial for modulating their chemical activity for various electrochemical applications, including hydrogen evolution reactions. In this study, we demonstrate a colloidal synthetic method to prepare core-shell-like heterostructures composed of MoSe₂ and WSe₂ via a two-step sequential growth. By overgrowing WSe₂ on the surface of preexisting MoSe₂ nanosheet edges, MoSe₂-core/WSe₂-shell heterostructures were successfully obtained. Systematic comparisons of the secondary growth time and sequential order of growth suggest that the low synthetic temperature conditions allow the stable overgrowth of shells rich in WSe₂ on top of the core of MoSe₂ with low Gibbs formation energy. The electrochemical analysis confirms that the catalytic activity correlates to the core-shell composition variation. Our results propose a new strategy to control the edge site activity of TMD materials prepared by colloidal synthesis, which is applicable to diverse electrochemical applications.

1. Introduction

Transition metal dichalcogenides (TMDs) have attracted tremendous attention owing to their interesting properties originating from their two-dimensional (2D) structures, which generally consist of a central transition metal layer sandwiched between chalcogen atomic layers. Along with the geometry, a diverse selection of constituents (metals and chalcogens), crystal phases (e.g., trigonal prismatic and octahedral), and a number of layers impart characteristic physical properties, depending on the TMD structures. As an example, group VI TMDs, including MoS₂ or WSe₂, primarily show semiconducting properties with a bandgap of −1 eV, while group V TMDs, such as the semimetallic VSe₂ or NbS₂ exhibit superconductivity [1,2]. The phase transition between metallic 1T and semiconducting 2H-phases occurs via chemical intercalation [3,4] or laser excitation [5], without changing the TMD layer constituents. Moreover, the optical bandgap of semiconducting TMD layers exhibits a direct-to-indirect transition depending on the number of layers.

More importantly, the dangling bond-free van der Waals (vdW) interactions, together with the analogous crystal structures, allow the preparation of TMD heterostructures with diverse compositions and geometries. The formation of alloyed TMDs with entire solid solution ranges has been demonstrated via tunable bandgap engineering [6,7]. Furthermore, vertically stacked and laterally extended heterostructures with variable optical/electronic properties have been reported [8,9]. Accordingly, numerous preparation techniques, including mechanical exfoliation or physical/chemical vapor deposition, have been employed to prepare TMD-based heterostructures [10,11,12,13].

While these methods lead to complex heterostructures with a high level of control from atomically sharp junctions to alloys with arbitrary compositions, their applications are limited by complicated preparation procedures and low scalability. Alternatively, solution-based synthetic approaches overcome these limitations with scalable growth protocols at relatively mild temperatures [14,15]. Furthermore, a homogenous reaction environment enables the formation of TMD nanosheets with radially outward oriented edges, which is advantageous for catalysis. Regarding the heterostructure formation, TMD catalytic performance is enhanced by preparing type II heterojunctions, because the electron-hole separation facilitates the charge transfer [16,17,18]. Nevertheless, growing core-shell like TMD heterostructures in the solution phase remains challenging because the thermodynamic control of stepwise nucleation is hindered in the homogeneous precursor mixture.

Here, we present a synthetic approach for the preparation of core-shell-like TMD nanoflower heterostructures based on colloidal crystal growth in the solution phase. We selected MoSe₂ and WSe₂ as model systems for the heterostructures because of their comparable lattice constants (2.7 Å for MoSe₂ (100) and 2.8 Å for WSe₂ (100)). The sequential colloidal growth using primarily grown MoSe₂ nanosheets as the host material for the subsequent nucleation and growth of WSe₂ resulted in the formation of phase-segregated, core-shell-like nanosheets. The systematic comparison of the heterostructure compositional distribution as a function of secondary growth time and growth order provides important information on the thermodynamic and kinetic factors dominating the growth. Specifically, low-temperature conditions and the Gibbs formation energy difference between MoSe₂ and WSe₂ enable the stable growth of MoSe₂-core/WSe₂-shell heterostructure nanosheets. Moreover, electrochemical measurements based on the hydrogen evolution reaction evaluate the catalytic activity correlation to the core-shell composition. These results suggest that tuning the active site density of colloidally synthesized TMD materials is a crucial strategy for various electrochemical applications.

2. Materials and Methods

2.1. Materials

Molybdenum hexacarbonyl (Mo(CO)₆, 98%), tungsten hexacarbonyl (W(CO)₆, 99.99 %), tetradecylphosphonic acid (TDPA, 97%), oleic acid (OA, technical, 90%), and diphenyl diselenide (Ph₂Se₂, 98%) were purchased from Sigma-Aldrich (Darmstadt, Germany). Solvents, including toluene (99.5%) and n-butanol (99%), were purchased from Duksan (Ansan, South Korea). All chemicals were used as received, without further purification.

2.2. Synthesis of MoSe₂, WSe₂, and MoSe₂/WSe₂ Heterostructures

MoSe₂ and WSe₂ nanosheets were grown using a colloidal hot-injection method. 108 mg (0.4 mmol) of Mo(CO)₆ or 141 mg (0.4 mmol) of W(CO)₆ were added into a 100 mL three-neck flask which was pre-loaded with 78.5 mg (0.28 mmol) of TDPA dissolved into 6 g of OA. 250 mg Ph₂Se₂, the selenium precursor, was dissolved in 4 g of OA in a separate vial. The Mo or W precursor mixtures were heated to 130 °C for 10 min and kept for 20 min under Ar atmosphere. The temperature was increased to 350 °C (for MoSe₂) and 330 °C (for WSe₂) at a rate of 4 °C/min, followed by the Se precursor solution injection. The nanosheet growth was maintained for 2 h, and then the flask was removed from the heating mantle to gradually cool down to room temperature. The crude solution was diluted with a solvent mixture composed of toluene and n-butanol in a 2:1 volume ratio. The suspension was then centrifuged for 30 min at 4000 rpm. The black powdery product was washed three times with a solvent mixture of toluene and n-butanol. Core–shell heterostructures were prepared by sequential growth of WSe₂ on MoSe₂ nanosheets or vice versa. The as-synthesized MoSe₂ and WSe₂ nanosheets were loaded into a 100 mL three-neck flask with 0.4 mmol of W(CO)₆ and Mo(CO)₆, respectively, and 78.5 mg of TDPA dissolved in 6 g of OA. The temperature profiles and the injection amount of the Se precursor were identical to those of the individual nanosheet growth, except that the secondary growth time varied from 30 min to 6 h.

2.3. Characterization

The as-grown nanosheets dispersed in toluene were drop-casted onto Si and SiO₂/Si substrates, and their morphologies were examined using a scanning electron microscope (SEM, SU8010, Hitachi, Tokyo, Japan). The same substrates were used for the investigation of structures using a custom-built Raman spectrometer equipped with a charge-coupled device (DV420A-OE, Toshiba, Tokyo, Japan). The wavelength and power of the laser were 532 nm and 5 mW, respectively. The sample suspension was drop-cast onto a TEM grid (#01824, Ted Pella, Redding, CA, USA) for transmission electron microscopy (TEM, JEM2100F, JEOL, Tokyo, Japan). Microscopic images and selected area diffraction patterns were obtained under an operating voltage of 200 kV. The crystallinity and chemical composition of the samples were measured using powder X-ray diffraction (PXRD, X’Pert-PRO MRD, PANalytical, Malvern, United Kingdom) and X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo Scientific, Seoul Korea). Prior to the measurements, the samples were dried at 300 °C for 2 h under N₂ atmosphere.

2.4. Electrochemical Measurements

The hydrogen evolution reaction (HER) was performed with a three-electrode-configuration HER potentiostat (PGSTAT 302N, Autolab, Utrecht, The Netherlands). Graphite rod and KCl-saturated Ag/AgCl were used as the counter and reference electrodes, respectively. The catalyst ink was prepared by mixing 4 mg of the catalyst (sample powders after drying), 30 μL of Nafion solution, and 1 mL of an ethanol-deionized water mixture (1:4 ratio). The mixed ink solution was ultrasonicated for 1 h. The catalyst ink (5 μL) was deposited onto a polished glassy carbon electrode and dried overnight. Linear sweep voltammetry measurements were conducted in 0.5 M H₂SO₄ solution with a sweep rate of 5 mV/s from 0.42 V to −0.78 V vs. reversible hydrogen electrode (RHE).

3. Results and Discussion

The MoSe₂/WSe₂ heterostructures were prepared via a sequential two-step colloidal synthesis, as illustrated in Scheme 1. First, MoSe₂ nanosheets were primarily grown upon hot injection of diphenyl diselenide dissolved in oleylamine (OA) into a solution of Mo(CO)₆ dissolved in OA with TDPA at 350 °C and reacted for 2 h. The as-grown MoSe₂ nanosheets served as the substrate for the secondary growth of WSe₂; therefore, MoSe₂, W(CO)₆, and TDPA were mixed in a three-neck flask, followed by a hot injection of diphenyl diselenide at 350 °C. After 2–6 h growth, the final MoSe₂/WSe₂ heterostructures were obtained.

Figure 1a–c illustrates the SEM images of the representative MoSe₂, WSe₂, and MoSe₂/WSe₂ heterostructures. The agglomerates of nanosheets with flower-like morphologies are confirmed from all the samples, analogously to previous studies [15,19]. The average diameter of the individual domains were in the range 200–300 nm. The powder XRD analysis shown in Figure 1d–f confirmed the correspondence of the agglomerated nanosheet diffraction pattern to the reference data (JCPDS #: 03-065-3999 (MoSe₂), 01-089-5257(WSe₂)). Five peaks located at 2θ = 13.6° (or 13.7°), 31.4°, 37.8° (or 37.9°), 47.4°, and 55.9° (or 56.0°) were assigned to the (002), (100), (103), (105), and (110) planes of MoSe₂ (WSe₂), respectively, indicating that the as-grown materials displayed the characteristic crystal structures of the 2D TMDs prepared via colloidal synthesis. Notably, since MoSe₂ and WSe₂ are characterized by comparable crystal structures (2H phase) and lattice constants [15,20], the as-obtained MoSe₂/WSe₂ hetero-nanosheets exhibited analogous morphologies and XRD peak patterns to the corresponding single TMD counterparts (Figure 1f). Despite the indistinguishable shapes and structures, XPS measurements of the MoSe₂/WSe₂ hetero-nanosheets (Figure 1g–i) confirmed the concurrent presence of both Mo and W. The peaks located at 229 and 232 eV (Figure 1g) corresponding to 2H MoSe₂ (Mo⁴⁺ 3d_5/2 and Mo⁴⁺ 3d_3/2, respectively) suggested that Mo in the core region of the heterostructure largely retained its unoxidized state [21]. Conversely, the core-level W 4f spectrum of the same sample (Figure 1h) indicated the presence of both oxidized (WO₃, W⁶⁺ 4f_5/2 at 38 eV and W⁶⁺ 4f_7/2 at 36 eV) and unoxidized (2H-WSe₂, W⁴⁺ 4f_5/2 at 35 eV and W⁴⁺ 4f_7/2 at 33 eV) states [22]. The core-level Se²⁻ spectrum (Se²⁻ 3d_5/2 at 54.5 eV and Se²⁻ 3d_3/2 at 55.3 eV) shows no direct evidence of Se oxidation, as shown in Figure 1i [23]. The observed distinct oxidized states between Mo and W suggest a selective oxidation of the outer regions, rich in WSe₂, whereas the inner domains, primarily consisting of MoSe₂, were preserved. Therefore, the results presented in Figure 1 provide initial evidence of the uneven distribution of Mo and W in the MoSe₂/WSe₂ heterostructures.

TEM measurements were conducted to investigate the as-grown-nanosheet crystal structure. Figure 2a–c display low-magnification TEM images of individual MoSe₂, WSe₂, and MoSe₂/WSe₂ nanosheets, respectively, confirming their nanoflower-like morphologies with noticeable thin edges. High-resolution TEM (HRTEM) images obtained near the edges of each sample (highlighted by dashed red boxes) are shown in Figure 2d–f. Lattice fringes spaced by 2.7 Å (Figure 2d) and 2.8 Å (Figure 2e), corresponding to the (100) planes of MoSe₂ and WSe₂, showed six-fold symmetry [15,20]. The selected area electron diffraction (SAED) patterns in the insets exhibited clear diffraction spot separation with the zone axis along the [001] direction, confirming that the edges of both MoSe₂ and WSe₂ are primarily composed of highly crystalline monolayers. While similar six-fold symmetric lattice fringes were observed for the MoSe₂/WSe₂ heterostructures, the fast Fourier transform (FFT) pattern (inset of Figure 2f) was stretched along the edges, reflecting the presence of curved monolayers facing outwards.

Figure 3a,b show the energy-dispersive X-ray spectroscopy (EDS) measurements of a representative MoSe₂/WSe₂ heterostructure. The elemental mappings of Mo, W, and Se (Figure 3b) confirmed that the spatial distributions of Mo and W are distinct in a 350 nm-sized hetero-nanosheet. While the signals of Mo are prominent only within the core region (200 nm), those of W appear uniformly distributed across all the domains (350 nm). The spatially distinct distributions of Mo and W were further validated by superimposing the mapping images (lower right panel of Figure 3b). Hence, the MoSe₂/WSe₂ hetero-nanosheets prepared via the two-step method exhibited core-shell-like structures. As the SEM and TEM images of MoSe₂ and WSe₂ in Figure 1 and Figure 2 indicate that the as-grown individual nanosheets exhibit three-dimensionally connected planar edges, we speculate that the sequential growth will result in the formation of WSe₂ secondary layers both on top of the planes and along the edges of MoSe₂, as illustrated in Figure 3c.

To corroborate the core-shell-like growth of WSe₂ on MoSe₂, we systemically investigated the distribution of W in the heterostructures as a function of the secondary growth time. Three samples were prepared by varying the secondary growth times, as shown in the TEM images in Figure 4a. Here, MW22, MW24, and MW26 denote the heterostructures of WSe₂ grown for 2, 4, and 6 h, respectively, over the primary MoSe₂ nanoflowers obtained after 2 h of growth. The profiles of W:Mo and Se:Mo atomic ratios corresponding to the 11 linear yellow dots crossing each nanoflower are shown in Figure 4b. For comparison, the ratios were normalized to the point with the highest Mo concentration. Notably, W atomic ratio increased in the outer domain region with the secondary growth time, consistently with the core-shell-like growth of WSe₂. Figure 4c shows the W:Mo ratio distribution across the nanoflower domains. While the ratio near the outer edges drastically increased with the growth time, near the center it remained lower than 1. The change in the overall W:Mo ratio of each sample (Figure 4d) further corroborated the increase in W concentration proportionally to the secondary growth time. Remarkably, WSe₂ layer heterogeneous nucleation was favored on the surface of the pre-grown MoSe₂ domains for the core-shell structures observed in this study. The activation energy barrier for WSe₂ nucleation on the preexisting MoSe₂ surface was lower than that for homogeneous nucleation because of the reduced surface free energy [24]. Since the individual WSe₂ nanosheets were absent in the samples grown via this two-step growth, it is speculated that the core-shell-like growth was favored over the independent nucleation, especially considering the maximum growth temperature of colloidal synthesis of 400 °C.

While the formation of the phase-segregated structures was preferred to the mixed alloys under our selected growth conditions, the difference in the formation enthalpy between MoSe₂ and WSe₂ may affect the spatial distribution of the components (i.e., core vs. shell), according to the chronological order of the growth protocol. Figure 5 displays the results of the heterostructure growth in reverse order for WSe₂ growth for 2 h, followed by MoSe₂ growth for 30 min (marked as WM2½). While both TEM and HRTEM images indicate that the nanosheet size and the monolayer edge configurations are similar to those of the MW samples (Figure 5a,b), the elemental analysis across the domain points out striking differences. The concentration of W was substantially lower than that of Mo, and only a small amount was detected near the edge (Figure 5c), indicating that W in the primary nanosheet structure can dynamically interchange with Mo during the secondary growth.

Two possible explanations of the observed Mo-rich nanosheets from the reverse order growth can be considered. First, the MoSe₂ nanosheet particles grew homogeneously without the pre-grown WSe₂ surface assistance. According to this assumption, both MoSe₂ and WSe₂ independently exist in the final product. However, no WSe₂ nanosheets were observed in this sample, suggesting that the hypothesis of homogeneous nucleation can be excluded. The second suggested scenario considers the substitution via interdiffusion of W in the primarily-grown-WSe₂ nanosheet by Mo during the second step. The lower Gibbs free energy for the MoSe₂ formation compared to WSe₂ under the selected growth temperature (<700 °C) [25] suggests that MoSe₂ tends to reduce its surface energy relatively to WSe₂. Thus, Mo diffused towards the nanosheet core upon shell growth on the preexisting WSe₂ nanosheet surfaces. Noticeably, the formation of the Mo_1−xW_xSe₂ alloy was unlikely because of the low-temperature conditions used in colloidal synthesis, with a consequent reduction of the entropic contribution to the formation energy of mixed alloys. A similar behavior was observed in the two-step sequential growth of the WS₂/MoS₂ heterostructure employing the chemical vapor deposition (CVD) technique, as reported by Bogaert et al. [26]. In that case, the sequential growth of WS₂-MoS₂ showed the formation of Mo_0.5W_0.5S₂ at high temperatures (>680 °C), whereas the MoS₂ core-WS₂ shell heterostructure was found at a low temperature (650 °C), which was reversed from the growth order. Hence, our observations indicate that the compositional distribution of the core-shell heterostructures is dominated by the thermodynamic preferences of their constituents.

Figure 6a shows the Raman spectra of the core-shell heterostructures prepared with various MoSe₂ secondary growth times. The Raman peaks of the pure MoSe₂ nanosheet (black) are located at 240 and 283 cm⁻¹, corresponding to the out-of-plane A_1g and in-plane E¹_2g modes, respectively, whereas the WSe₂ nanosheet (red) exhibits a degenerated E¹_2g-like peak centered at 249 cm⁻¹ [26,27]. While the hetero-nanosheets (MW21, MW22, MW24, and MW26) displayed Raman spectra largely similar to the core-only MoSe₂, the emergence of the shoulder peaks at 249 cm⁻¹ and their increase in intensity were evident as the secondary growth time increased from 1 h to 6 h. Notably, the MoSe₂ peaks of the heterostructures did not exhibit a shift in wavenumbers, suggesting that the WSe₂ shell growth did not affect the lattice strain of the MoSe₂ core. Furthermore, the typical features of the A_1g blue-shift and E¹_2g red-shift corresponding to the formation of Mo_xW_1−xSe₂ alloy were absent [28], supporting the core-shell-like structure for all the samples. To confirm the compositional dependence of the heterostructures on the secondary growth time, the Raman intensity ratio of the characteristic E¹_2g peak of WSe₂ (249 cm⁻¹) to MoSe₂ (240 cm⁻¹) was plotted as a function of WSe₂ growth time, as illustrated in Figure 6b. The tendency of the asymptotic increase in the intensity ratio with the secondary growth time confirmed the tunable overgrowth of the WSe₂ shell, which is immiscible with the MoSe₂ core.

The obtained MoSe₂/WSe₂ core-shell heterostructures with variable shell thicknesses were further applied to the HER electrochemical reaction. We compared the HER performance of the as-grown heterostructures using a three-electrode configuration, as shown in Figure 7. Figure 7a shows the linear sweep voltammetry (LSV) curves of MW22 and MW24 transferred onto glassy carbon (GC) electrodes. The data from pure MoSe₂ and WSe₂ are shown together with the Pt electrode, as a reference. The overpotential values at a current density of 10 mA/cm² were 80, 340, 410, 580, and 650 mV for Pt, MoSe₂, MW22, MW24, and WSe₂, respectively.

While MoSe₂ required a lower overpotential than WSe₂, the heterostructures exhibited intermediate performance according to the WSe₂ shell growth time. The corresponding Tafel slopes (Figure 7b) of 83 (Pt), 105 (MoSe₂), 150 (MW22), 238 (MW24), and 178 (WSe₂) mV/decade additionally reflected the relative HER performance. Finally, the Nyquist plot obtained from the electrochemical impedance spectroscopy (EIS) shown in Figure 7c further confirmed that the resistance of MW22 (580 Ω) and MW24 (713 Ω) is between that of MoSe₂ (86 Ω) and WSe₂ (1652 Ω).

Table 1. HER performance comparison of MoSe₂-based nanostructure catalysts showing the overpotentials and Tafel slopes.

Materials	Loading	η at j = 10 mA/cm2 (mV)	Tafel Slope (mV/decade)	Ref.
MoSe2 nanoflowers	0.28 mg cm−2 on GC	340	105	This work
MW22 nanoflowers		410	150
MW24 nanoflowers		580	238
WSe2 nanoflowers		650	178
Bulk MoSe2	0.28 mg cm−2 on GC	>600	346	[14]
Porous MoSe2 nanosheets	0.47 mg cm−2 on GC	150	80	[29]
Vertically aligned MoSe2 layers	0.0135 mg cm−2 on GC	−450	105–120	[30]
Hierarchical MoSe2−x nanosheets	0.28 mg cm−2 on GC	288	98	[14]
MoSe2 nanoflakes (2:1 ratio to Ketjen black)	0.28 mg cm−2 on GC	181 ± 1	45 ± 1	[31]
MoSe2 nanonetworks (8:1 ratio to Ketjen black)	0.28 mg cm−2 on GC	302	109	[19]
MoSe2 nanoflowers (8:1 ratio to Ketjen black)	0.28 mg cm−2 on GC	233	89	[19]

summarizes the electrochemical performances of the colloidal TMDs prepared in this study compared with previously reported data. While the absolute performance values are inferior to the published values, presumably owing to the presence of residual precursors, our results provide important implications regarding the dependence of electrochemical activity on structural characteristics of MoSe₂/WSe₂ core-shell heterostructures. Given that the edges of MoSe₂ exhibit lower hydrogen adsorption energy relative to WSe₂, the increase in the shell growth time would result in the reduction of MoSe₂ edges and a subsequent decrease in the electrochemical activity. Thus, the intermediate HER performances of heterostructures, dependent on the secondary growth time, suggest that the core-shell growth protocol proposed in this study can be applied to tune the electrochemical activities of colloidal TMD materials.

4. Conclusions

We presented core-shell heterostructures of MoSe₂/WSe₂ TMDs obtained via a two-step colloidal synthesis. Comprehensive structural and chemical analyses confirmed that the as-grown heterostructures were composed of a spatially distinct 2H-MoSe₂ core and WSe₂ shell layers. Due to the lower regime of synthetic temperatures relative to the vapor-phase-based approach, the colloidal growth of TMDs in solution enables the formation of phase-segregated, core-shell-like structures rather than uniformly distributed alloys. The composition of the shell layer was tunable by changing the secondary growth time, and the spatial distribution of constituents in the heterostructure was thermodynamically determined as such MoSe₂ resides in the core region. Notably, the electrochemical properties measured from the as-grown samples correlated to the variety of core-shell compositions. Our results offer an opportunity to control the properties of colloidal TMDs, depending on the edge structure. The ability to modulate electrochemistry by engineering active edge sites is critical for developing highly efficient HER electrocatalysts.