1. INTRODUCTION

Hydrogen energy has emerged as a highly promising alternative to fossil fuels, and an efficient way to address the environmental crisis and rising energy demand, particularly due to its advantages in storage and transportation. Among the various hydrogen-production technologies, photoelectrochemical (PEC) water splitting, which utilizes sunlight to generate hydrogen, has attracted significant research attention as a sustainable and environmentally friendly method^{[1, 2]}. Conventional PEC materials, such as TiO₂, Fe₂O₃, ZnO, and WO₃, have been extensively studied due to their excellent stability^[3-^6]. However, their wide bandgaps limit visible-light absorption. Although strategies such as doping and heterostructure formation have been explored to overcome this limitation, improvements in PEC efficiency remain limited^[7-^10]. To fully leverage solar energy, semiconductor materials with an appropriate bandgap for visible-light absorption are essential. Two-dimensional (2D) transition metal dichalcogenides (TMDs) have attracted considerable attention as next-generation functional materials for applications including solar cells and PEC photoelectrodes because of their favorable optoelectronic properties, high catalytic activity, and suitable bandgap (E_g)^[10-^14]. In particular, MoS₂, a prototypical 2D TMD, has been reported as a viable PEC electrode material due to its high carrier mobility and favorable E_g for visible-light-driven hydrogen generation^{[15, 16]}. The PEC performance of MoS₂ can be further enhanced by heterostructure engineering, to facilitate photogenerated electron–hole separation and transfer. In this regard, MoS₂/graphene heterostructures exhibit an efficient 2D/2D synergistic effect.

Graphene is particularly attractive as a conductive support layer due to its exceptional electron mobility (~2 × 10⁴ cm²V^-1s^-1) and its compatibility for forming efficient heterojunctions with 2D TMDs^{[16, 17]}. However, most MoS₂/graphene heterostructures are synthesized via wet chemical or mechanical exfoliation methods, which often result in aggregation. This aggregation not only inhibits the ideal formation of nanostructures on graphene but also diminishes electron-transfer properties and catalytic active-site availability due to strong van der Waals and π–π interactions^{[18, 19]}. In contrast, as a promising alternative, vacuum-based fabrication techniques for MoS₂/graphene heterostructures enable precise control over MoS₂ morphology, facilitating the design of efficient PEC electrodes. Additionally, control of MoS₂ morphology and thickness is critical, as both parameters directly influence its surface area and charge transport characteristics. While charge transport in the basal planes of MoS₂ is relatively poor, the edge sites exhibit superior catalytic activity and charge carrier dynamics^[15]. Therefore, fabricating nanostructured MoS₂ with maximized edge-site exposure is desirable for efficient PEC performance^{[12, 20]}.

In this study, we demonstrate the morphology-controlled growth of MoS₂ on graphene substrates via metal–organic chemical vapor deposition (MOCVD), using deposition time as the primary tuning parameter. The synthesis method preserved the underlying graphene in the resulting MoS₂/graphene heterostructures, ranging from thin films to vertically aligned nanosheets, highlighting the capability of this approach for constructing tunable hybrid nanostructures. Comprehensive analyses of MoS₂ morphology and thickness were performed, and the optimized MoS₂/graphene heterostructure for PEC water-splitting applications was identified.

2. EXPERIMENTAL

3. RESULTS AND DISCUSSION

Figure 1(a) and 1(b) show cross-sectional SEM images of MoS₂ grown for 10 min on ITO and graphene substrates, respectively, illustrating the morphological characteristics of the nanosheet structures. The insets provide top-view SEM images for additional perspective. As shown in Figure 1(a), MoS₂ grown at 290 °C for 10 min (hereafter referred to as pristine MoS₂) exhibits vertically aligned, densely packed nanosheets with sharp edges. The average size of the nanosheets is approximately 200–240 nm. Similarly, MoS₂ grown on graphene (Figure 1(b)) forms vertically oriented nanosheets that uniformly cover the substrate, demonstrating morphological features comparable to those of pristine MoS₂.

Figure 1(c) and 1(d) present the Raman spectra of the graphene, pristine MoS₂, and the MoS₂/graphene heterostructure. The pristine MoS₂ shows two prominent peaks at 381.2 and 404.7 cm^–1, corresponding to the in-plane E¹ _2g and out-of-plane A_1g vibrational modes of 2H-MoS₂, respectively^{[16, 22]}. The MoS₂/graphene heterostructure displays similar E¹ _2g and A_1g peaks, while no noticeable Raman features are observed from the graphene substrate in this spectral region (300–500 cm^–1). However, within the Raman range of 1200–2800 cm^–1, the presence of characteristic G and 2D bands (Figure 1(d)) confirms that the graphene layer remained following the MOCVD growth of MoS₂ ^[23]. This result indicates that the proposed synthesis method is suitable for fabricating high-quality MoS₂/graphene heterostructures while preserving the underlying graphene.

X-ray photoelectron spectroscopy (XPS) was employed to investigate the chemical bonding states of the sample. Figure 1(e) and (f) present the deconvoluted XPS core-level spectra of Mo 3d and S 2p obtained from the MoS₂/graphene sample. All spectra were fitted using a Gaussian–Lorentzian function with a Shirley background, enabling detailed analysis of the characteristic peaks. As shown in Figure 1(e), the Mo 3d core-level spectrum of the MoS₂/graphene sample exhibits two prominent peaks at 229.5 and 232.6 eV, which are assigned to the Mo 3d_5/2 and 3d_3/2 states of the 2H phase, respectively^[12]. In the same spectrum, the S 2s peak appears at 226.7 eV, while a weak peak at 235.9 eV corresponds to Mo⁶⁺ species, indicating slight surface oxidation. This oxidation is likely caused by unavoidable exposure to ambient air during sample transfer for analysis^[16]. Furthermore, the S 2p core-level spectrum shown in Figure 1(f) displays two characteristic peaks at 162.4 and 163.6 eV, which are attributed to the S 2p_3/2 and S 2p_1/2 states of MoS₂, respectively. The binding energies of the Mo 3d, S 2s, and S 2p peaks are consistent with previously reported values for 2H-MoS₂ ^[12], further confirming the successful formation of MoS₂ on graphene, in agreement with the Raman analysis.

Fig. 1. SEM images of (a) pristine MoS₂ and (b) MoS₂/graphene heterostructure. Raman spectra of various samples (graphene, pristine MoS₂, and MoS₂/graphene) in the (c) 300–500 cm^–1 and (d) 1200–2800 cm^–1 region. XPS spectra of (e) Mo 3d, and (f) S 2p core levels for MoS₂/graphene sample

Figure 2(a)–(d) presents the morphological characteristics of the MoS₂/graphene heterostructures with different MoS₂ growth times (2, 5, 10, and 30 min, hereafter referred to as MG-2m, MG-5m, MG-10m, and MG-30m, respectively). The insets show corresponding top-view SEM images. At a growth time of 2 min (MG-2m), a uniform MoS₂ thin film with a thickness of approximately 80 nm was formed, as shown in Figure 2(a). In addition, nanoparticles with diameters ranging from ~90 to 110 nm could be partially observed (inset of Figure 2(a)). When the growth time increased to 5 min (MG-5m), the MoS₂ morphology changed to vertically aligned nanosheets with sizes of approximately 150–180 nm (Figure 2(b)). When growth times were further increased to 10 min and 30 min (MG-10m and MG-30m), the nanosheet structures continued to grow, reaching average sizes of approximately 230–270 nm and 880–960 nm, as shown in Figure 2(c) and 2(d), respectively. Figure 2(e) displays a TEM image of the MG-10m sample, clearly showing uniformly distributed vertical nanosheets. A high-resolution TEM (HRTEM) image taken from the yellow-boxed region in Figure 2(e) is shown in Figure 2(f), revealing a lattice spacing of approximately 0.63 nm, which corresponds to the (002) plane of the 2H-phase MoS₂ ^[23].

Fig. 2. (a) SEM images of MoS₂ synthesized on graphene substrate at various growth times: (a) MG-2m at 2 min, (b) MG-5m at 5 min, (c) MG-10m at 10 min, and (d) MG-30m at 30 min. (e) Low-magnification TEM image of MG-10m and (f) HRTEM image.

To investigate the optical properties of the samples, a series of systematic analyses, including UV–Vis absorbance, PL, and TRPL, were carried out. Figure 3(a) presents the UV–Vis absorption spectra of the graphene, pristine MoS₂, and the MoS₂/graphene heterostructure (MG-10m). Both pristine MoS₂ and MG-10m exhibit two characteristic absorbance peaks at approximately 604 nm and 661 nm, corresponding to direct bandgap transitions at the K point in the Brillouin zone of MoS₂ ^{[24, 25]}. The slight difference in absorbance between the two samples is attributed to the weak absorbance of the underlying graphene layer. To gain further insights into the photoexcited carrier dynamics, PL and TRPL measurements were performed on pristine MoS₂ and the MoS₂/graphene heterostructures (MG-2m, MG-5m, MG-10m, and MG-30m). As shown in Figure 3(b), the pristine MoS₂ sample exhibits a strong PL emission peak centered at approximately 675 nm, characteristic of direct bandgap transitions, in agreement with previous reports^[23]. For the MoS₂/graphene heterostructures with vertically aligned nanosheet structures (MG-5m, MG-10m, and MG-30m), PL peaks were observed at approximately 682 nm. In contrast, the MG-2m sample, which has a thin-film structure, exhibited a broader PL emission in the 700–750 nm range, likely originating from an indirect bandgap transition^{[26, 27]}. Notably, all of the MoS₂/graphene heterostructures exhibited significantly lower PL intensities compared to pristine MoS₂, indicating recombination of photoexcited electron–hole pairs is being suppressed^{[23, 26]}. This decreased PL intensity is attributed to the MoS₂/graphene heterojunction effect, which facilitates efficient charge separation and carrier transport across the MoS₂/graphene interface, potentially enhancing PEC performance^{[26, 28]}. The TRPL decay profiles of pristine MoS₂ and the MoS₂/graphene heterostructures are presented in Figure 3(c), with the extracted carrier lifetimes summarized in the inset table. Pristine MoS₂ exhibited an average carrier lifetime of 11.7 ns, whereas the MoS₂/graphene samples demonstrated shorter lifetimes (MG-2m: 10.0 ns, MG-5m: 9.2 ns, MG-10m: 8.9 ns, and MG-30m: 10.1 ns). These shortened lifetimes can be attributed to the MoS₂/graphene heterostructure, which facilitates the efficient separation and transport of photogenerated carriers to the semiconductor–liquid interface. In particular, the vertically aligned nanosheet architecture further enhances carrier extraction by providing effective charge transport pathways through their highly conductive edges^{[12, 26]}. Meanwhile, MG-30m showed a slight increase in carrier lifetime compared to MG-10m, which is attributed to increased recombination resulting from the limited diffusion length of photogenerated carriers in the larger nanosheet structures^[29]. This trend is consistent with the PL results.

Fig. 3. (a) UV–Vis absorbance spectra of graphene, pristine MoS₂, and the MoS₂/graphene heterostructure (MG-10m). (b) PL spectra and (c) TRPL results of MoS₂, MG-2m, MG-5m, MG-10m, and MG-30m. The inset table in (c) lists the photogenerated carrier lifetimes based on corresponding TRPL measurements.

The PEC activities of pristine MoS₂ and MoS₂/graphene (MG-10m) heterostructures were evaluated by recording linear sweep voltammograms under dark conditions and simulated AM 1.5G illumination. As shown in Figure 4(a), MG-10m exhibited an enhanced photocurrent density (2.44 mA cm^–2 at 0.8 V) compared to pristine MoS₂ (1.65 mA cm^–2 at 0.8 V) across the entire potential range, while their dark currents (dotted lines) remained comparable. The onset potential, defined as the potential at the intersection of the dark current and the tangent to the region of photocurrent slope, was approximately 0.49 V for pristine MoS₂. In contrast, MG-10m showed a cathodic shift in onset potential to approximately 0.44 V, indicating improved catalytic activity. The enhanced photocurrent and reduced onset potential observed in MG-10m are attributed to more effective separation and the transfer of photogenerated carriers at the MoS₂/graphene heterojunction^{[23, 28]}, which was further corroborated by EIS. Figure 4(b) presents the EIS Nyquist plots of pristine MoS₂ and MG-10m under illumination. These Nyquist plots were fitted using a simplified Randles circuit, as shown in the inset of Figure 4(b), which consists of solution resistance (R_s), charge transfer resistance (R_ct), and a constant phase element (C). According to prior studies, a smaller semicircle diameter in Nyquist plots corresponds to a lower charge transfer resistance (R_ct), indicating more efficient carrier transport^[30-^32]. Consistent with this, MG-10m displayed a smaller semicircle than pristine MoS₂, confirming the enhanced separation and transport of photogenerated electron–hole pairs across the heterojunction.

Furthermore, the PEC performance was significantly affected by the morphology and nanosheet size of MoS₂ grown via the morphology-controlled MOCVD method. Figure 4(c) displays the photocurrent (solid lines) and dark current (dotted lines) densities of MoS₂/graphene heterostructures synthesized at various growth times (MG-2m, MG-5m, MG-10m, and MG-30m). The MG-5m sample showed a notable increase in photocurrent compared to MG-2m, and MG-10m exhibited the highest PEC performance. However, when the growth time was extended to 30 min (MG-30m), the PEC performance did not improve, and instead deteriorated compared to MG-10m. This tendency may be due to the following factors. (i) Small-sized particles and nanosheets showed relatively unsatisfactory PEC performances because their specific surface areas were not significantly improved. (ii) Large and excessively packed vertical nanosheets can limit the diffusion length of the carriers and reduce active site exposure owing to their high overlap with each other, which leads to the degradation in PEC reactivity^{[12, 33]}. Previous studies have shown that a significant portion of photogenerated charge carriers generated in the upper regions of MoS₂ nanosheets beyond the diffusion length contributes minimally to the PEC reaction^[29], which accounts for the low photocurrent observed in MG-30m with excessively large nanosheets. These findings indicate why MG-10m, which features vertically aligned nanosheets of optimal size and density, achieved the best PEC performance. Figure 4(d) presents the EIS Nyquist plots for all of the MoS₂/graphene heterostructures. MG-2m exhibited the largest semicircle diameter due to its limited surface area and poor PEC activity. Conversely, MG-10m showed the smallest semicircle, consistent with its superior PEC photocurrent performance, confirming efficient carrier separation and transport facilitated by its optimized nanostructure.

Fig. 4. (a) Photocurrent and dark density-potential curves of pristine MoS₂ and MG-10m PEC cells. (b) EIS Nyquist plots of MoS₂ and MG-10m under illumination. (c) Current density-potential curves and (d) EIS Nyquist plots of PEC cells with various working electrodes (MG-2m, MG-5m, MG-10m, and MG-30m).

4. CONCLUSIONS

In this study, vertically aligned MoS₂ nanosheets were successfully synthesized and practically optimized on graphene using a controllable MOCVD system. The resulting MoS₂/graphene heterostructures, featuring vertically oriented nanosheets, demonstrated enhanced PEC water splitting performance. This improvement can be attributed to the following main factors: (i) the formation of an ideal nanostructure, where edge-rich vertical MoS₂ nanosheets grown on graphene expose a large number of active sites; and (ii) the efficient separation and suppressed recombination of photogenerated charge carriers enabled by the heterojunction effect between MoS₂ and graphene. Furthermore, the morphology and size of the MoS₂ nanosheets were found to play a critical role in determining the PEC performance and carrier transport behavior of the heterostructures, and optimal PEC performance was achieved in structures with appropriately sized vertical nanosheets. These results highlight a promising strategy that combines morphology-controlled synthesis with heterojunction engineering of TMD nanosheets and graphene, offering potential for efficient PEC and other photocatalytic applications.