Stirring-Assisted In Situ Construction of Highly Dispersed MoS₂/g-C₃N₄ Heterojunctions with Enhanced Edge Exposure for Efficient Photocatalytic Hydrogen Evolution

Abstract

Constructing heterojunction photocatalysts with efficient interfacial charge transfer is critical for solar-driven hydrogen evolution. In this study, a highly dispersed MoS₂/g-C₃N₄ composite was successfully synthesized via a stirring-assisted hydrothermal in situ growth strategy. The introduction of stirring during synthesis significantly enhanced the uniform dispersion of MoS₂ nanosheets and exposed abundant edge sites, leading to well-integrated heterojunctions with enhanced interfacial contact. Comprehensive structural and photoelectronic characterizations (XRD, SEM, TEM, EDS mapping, UV–Vis, TRPL, EIS, EPR) confirmed that the composite exhibited improved visible-light absorption, accelerated charge separation, and suppressed recombination. Under simulated solar irradiation with triethanolamine (TEOA) as a sacrificial agent, the optimized 24% MoS₂/g-C₃N₄-S catalyst achieved a high hydrogen evolution rate of 14.33 mmol·g⁻¹·h⁻¹ at a catalyst loading of 3.2 mg, significantly outperforming the unstirred and pristine components, and demonstrating excellent cycling stability. Mechanistic studies revealed that the performance enhancement is attributed to the synergistic effects of Type-II heterojunction formation and edge-site-rich MoS₂ co-catalysis. This work provides a scalable approach for non-noble metal interface engineering and offers insight into the design of efficient and durable photocatalysts for solar hydrogen production.

1. Introduction

The escalating global energy crisis and environmental degradation have intensified the pursuit of clean and sustainable energy solutions. Hydrogen energy, characterized by its high energy density, zero carbon emissions, and renewability, stands out as a promising alternative. Among various hydrogen production methods, photocatalytic water splitting using semiconductor materials offers a green and efficient pathway, directly harnessing solar energy to convert water into hydrogen [1]. However, the practical application of this technology is hindered by challenges such as limited visible light absorption, low quantum efficiency, and poor stability of existing photocatalysts [2].

Graphitic carbon nitride (g-C₃N₄), a metal-free semiconductor, has garnered significant attention due to its suitable bandgap (~2.7 eV), excellent chemical stability, abundant raw materials, and facile synthesis [3]. These attributes make it a compelling candidate for applications in photocatalytic water splitting, CO₂ reduction, and pollutant degradation. The conduction band position of g-C₃N₄ (~−1.1 eV) endows it with strong reducing power, sufficient for driving the hydrogen evolution reaction (HER). Nevertheless, its photocatalytic performance is limited by rapid recombination of photogenerated electron-hole pairs, low specific surface area, and inadequate visible light absorption. To overcome these limitations, strategies such as heterojunction construction, heteroatom doping, and cocatalyst loading have been explored to enhance its photocatalytic efficiency.

Transition metal dichalcogenides (TMDs), particularly molybdenum disulfide (MoS₂), have emerged as effective cocatalysts in photocatalytic systems due to their unique electronic properties and layered structures. MoS₂ exhibits a sandwich-like S–Mo–S layered configuration, where the basal planes are catalytically inert, but the edge sites demonstrate remarkable HER activity. Density functional theory (DFT) calculations reveal that the Gibbs free energy for hydrogen adsorption at MoS₂ edge sites is close to zero, indicating high catalytic activity comparable to that of platinum [4,5,6]. Additionally, the conduction band position of MoS₂ (~−0.1 eV) facilitates the acceptance of electrons from photoexcited semiconductors like g-C₃N₄, promoting efficient charge separation and transfer.

Integrating MoS₂ with g-C₃N₄ to form heterojunctions not only provides additional active sites but also enhances the separation of photogenerated charge carriers. Studies have demonstrated that MoS₂/g-C₃N₄ composites exhibit superior photocatalytic performance compared to their individual components [7,8,9,10,11], attributed to the effective spatial separation of electrons and holes. In these systems, electrons generated in g-C₃N₄ under light irradiation can transfer to MoS₂, where they participate in the HER, while holes remain in g-C₃N₄ to drive oxidation reactions. Moreover, the incorporation of MoS₂ can improve the light absorption properties of g-C₃N₄ and create internal electric fields at the interface, further facilitating charge separation.

Despite the potential of MoS₂/g-C₃N₄ heterojunctions, their photocatalytic efficiency is highly dependent on the quality of the interface and the dispersion of MoS₂. Conventional synthesis methods, such as physical mixing in solutions(DI water or N,N-dimethylformamide (DMF)), deposition and calcination [7,8,11,12,13,14,15,16,17,18,19,20], often result in poor interfacial contact and aggregation of MoS₂, leading to suboptimal performance. These issues underscore the need for advanced synthesis strategies that ensure uniform dispersion of MoS₂ and strong interfacial bonding with g-C₃N₄.

In situ synthesis approaches, particularly hydrothermal methods [10,21,22,23,24,25], photo-deposition [26,27,28], and annealing [9,29,30,31,32,33,34,35], have shown promise in achieving well-dispersed MoS₂ on g-C₃N₄ surfaces. By controlling reaction parameters such as temperature, time, and precursor concentration, it is possible to tailor the morphology and crystalline structure of MoS₂, enhancing its integration with g-C₃N₄ [9,21]. However, traditional hydrothermal synthesis without agitation can lead to issues like precursor sedimentation and uneven nucleation, resulting in non-uniform MoS₂ distribution and reduced exposure of active edge sites.

To address these challenges, this study proposes a stirring-assisted in situ hydrothermal synthesis strategy for fabricating MoS₂/g-C₃N₄ composites. Unlike traditional static hydrothermal methods, the introduction of stirring during the synthesis process enhances mass and heat transfer, ensuring a more homogeneous distribution of precursors and facilitating uniform nucleation and growth of MoS₂ on g-C₃N₄. This novel approach leads to the formation of highly dispersed MoS₂ nanosheets with abundant edge sites, significantly improving the interface interactions between MoS₂ and g-C₃N₄. As a result, the composite exhibits enhanced charge separation efficiency, which is crucial for improving the photocatalytic hydrogen production performance. This research systematically investigates the impact of the stirring-assisted synthesis method on the structural uniformity and optical properties of MoS₂/g-C₃N₄ composites, providing valuable insights into the design of efficient, noble-metal-free photocatalysts. The findings highlight the unique advantage of this approach in achieving superior photocatalytic performance through the synergistic effects of MoS₂ and g-C₃N₄, with the potential to drive sustainable hydrogen generation.

2. Results and Discussion

2.1. Analysis of Synthesis, Phase, Morphology, and Element Distribution

2.1.1. Structure and Phase Characterization

The phase structures of the synthesized samples were examined by X-ray diffraction, and the results are shown in Figure 1. Pure g-C₃N₄ exhibits two characteristic diffraction peaks located at 13.1° and 27.4°, corresponding to the (100) and (002) planes, respectively, which reflect the in-plane structural packing of tri-s-triazine units and interlayer stacking of conjugated aromatic systems (JCPDS No. 87-1526). The pure MoS₂ sample shows broad reflections around 14°, 33°, and 58°, consistent with the (002), (100), and (110) planes of hexagonal MoS₂, albeit with low intensity, suggesting its poor crystallinity under the current synthesis conditions. Distinct structural differences are observed between the stirred and unstirred composite samples. In the 24% MoS₂/g-C₃N₄-U sample (unstirred), the characteristic peak at 27.4° becomes very weak, and the (100) peak at 13.1° almost disappears, indicating partial structural degradation or disordering of g-C₃N₄ during hydrothermal treatment without stirring. Meanwhile, the MoS₂-related features are clearly retained, and the pattern closely resembles that of the pristine MoS₂, suggesting that MoS₂ nucleates and grows independently without strong interaction with the g-C₃N₄ matrix. In contrast, the 24% MoS₂/g-C₃N₄-S sample (stirred) shows a well-resolved (002) peak at 27.4°, with only slightly reduced intensity compared to pure g-C₃N₄, and a weak but visible (100) peak at 13.1°. This suggests that the g-C₃N₄ framework remains largely intact during the stirring-assisted hydrothermal process. The diffraction features associated with MoS₂ in this sample are significantly weaker and broader than those in both the unstirred sample and pure MoS₂, implying a high dispersion of MoS₂ and possible formation of ultrathin or poorly crystalline nanosheets intimately anchored on the g-C₃N₄ surface. These results suggest that magnetic stirring during hydrothermal synthesis plays a critical role in preserving the crystallinity of g-C₃N₄ and ensuring a more homogeneous integration of MoS₂ onto the host matrix, which is expected to enhance interfacial contact and improve photocatalytic performance.

2.1.2. Morphology Analysis

The surface morphologies of pure g-C₃N₄, MoS₂, and their composites were investigated by scanning electron microscopy (SEM), as shown in Figure 2. The pristine g-C₃N₄ sample (Figure 2a) displays a typical layered and crumpled sheet-like morphology with thin, stacked nanosheets loosely aggregated into a porous structure. This morphology is consistent with the thermally exfoliated character of graphitic carbon nitride. In the case of the 24% MoS₂/g-C₃N₄-S composite synthesized under stirring conditions (Figure 2b), the sheet-like structure of g-C₃N₄ is largely preserved, with small MoS₂ nanoparticles appearing to be homogeneously anchored on the surface of the g-C₃N₄ layers. The good dispersion of MoS₂ is attributed to the uniform mixing and nucleation environment facilitated by continuous stirring during the hydrothermal process, which promotes better interfacial contact between the two components. By contrast, the morphology of the 24% MoS₂/g-C₃N₄-U sample prepared without stirring (Figure 2c) reveals obvious structural disruption. The g-C₃N₄ nanosheets appear more fragmented and partially collapsed, and MoS₂ aggregates are irregularly deposited, often forming visible clusters with less uniformity. This suggests that the absence of stirring leads to localized supersaturation and non-uniform growth of MoS₂, resulting in poorer dispersion and reduced interaction with the g-C₃N₄ substrate. The isolated MoS₂ sample (Figure 2d) consists of irregularly stacked, petal-like nanosheets forming dense spherical aggregates, a morphology commonly observed in MoS₂ synthesized by hydrothermal methods. Compared to the composites, the morphology here is more compact and lacks the layered architecture provided by the g-C₃N₄ scaffold. Overall, these observations indicate that magnetic stirring during synthesis plays a crucial role in preserving the lamellar structure of g-C₃N₄ and achieving uniform MoS₂ distribution, which is expected to contribute positively to charge separation and photocatalytic activity.

The microstructures of pure g-C₃N₄, MoS₂, and MoS₂/g-C₃N₄ composites were investigated by transmission electron microscopy (TEM) and high-resolution TEM (HRTEM), as shown in Figure 3. The pristine g-C₃N₄ sample (Figure 3a,e) exhibits a typical stacked nanosheet morphology with curled and transparent edges, reflecting its characteristic layered structure and low electron density.

For the 24% MoS₂/g-C₃N₄-S composite synthesized under stirring conditions (Figure 3b,f), the lamellar morphology of g-C₃N₄ is largely preserved, as evidenced by the sharp and planar edges visible in the TEM image. However, no well-defined lattice fringes corresponding to MoS₂ are observed in the HRTEM image (Figure 3f), suggesting that MoS₂ is highly dispersed and possibly amorphous or extremely thin, which renders its crystallinity difficult to resolve. Such high dispersion is favorable for increasing the active surface area and enhancing interfacial electron transfer.

In contrast, the 24% MoS₂/g-C₃N₄-U sample (Figure 3c,g) prepared without stirring displays a distinctly different morphology. The nanosheets appear more clustered and curled into flower-like stacks, closely resembling the morphology of pure MoS₂ (Figure 3d,h). Furthermore, the HRTEM image (Figure 3g) reveals clear lattice fringes, indicating locally aggregated and well-crystallized MoS₂ domains. These observations suggest that without stirring, MoS₂ preferentially nucleates and grows independently, leading to phase segregation and insufficient integration with the g-C₃N₄ matrix.

Elemental distributions in the composites were further examined via energy-dispersive X-ray spectroscopy (EDS) mapping (Figure 3i,j). In the stirred sample (Figure 3i), Mo, S, C, and N elements are evenly dispersed throughout the field of view, confirming the uniform incorporation of MoS₂ into the g-C₃N₄ framework. By contrast, the unstirred composite (Figure 3j) exhibits pronounced aggregation of Mo and S elements, consistent with localized MoS₂ clustering. These results reaffirm the essential role of stirring in promoting homogeneous precursor distribution and enabling uniform MoS₂ nucleation during hydrothermal synthesis.

2.1.3. Elemental Composition

Table 1. The atomic and mass ration of composite elements in 24% MoS₂/g-C₃N₄-S and 24% MoS₂/g-C₃N₄-U samples obtained from EDS Mapping and ICP-MS.

Samples	Ratio Mode	C	N	O	S	Mo
24% MoS2/g-C3N4-S	atom%-EDS	38.59	56.27	1.7	2.67	0.77
	wt%-EDS	32.23	54.81	1.89	5.95	5.11
24% MoS2/g-C3N4-U	atom%-EDS	16.28	14.72	11.32	46.28	11.4
	wt%-EDS	6.19	6.52	5.73	46.95	34.6
24% MoS2/g-C3N4-S	wt%-ICP					6.25

summarizes the atomic and weight percentages of key elements (C, N, O, S, Mo) in 24% MoS₂/g-C₃N₄ composites prepared under stirring (S) and unstirring (U) conditions, as obtained by EDS mapping and ICP-MS analysis. These quantitative findings provide important insights into the structural and photocatalytic differences revealed in XRD, SEM, and TEM results.

In the stirred sample (24% MoS₂/g-C₃N₄-S), EDS analysis shows that nitrogen and carbon dominate the atomic composition (56.27% and 38.59%, respectively), while Mo and S account for only 0.77% and 2.67%, respectively. The corresponding weight ratios of Mo (5.11%) and S (5.95%) confirm that MoS₂ is present in a small but uniformly distributed amount across the g-C₃N₄ surface. The Mo content measured by ICP-MS (6.25%) is in excellent agreement with the EDS result. These values are consistent with the XRD pattern showing strong g-C₃N₄ peaks and nearly invisible MoS₂ signals, as well as the TEM images where lattice fringes of MoS₂ are difficult to resolve—supporting the conclusion that MoS₂ is highly dispersed, possibly amorphous, and forms a strong interface with the g-C₃N₄ substrate.

In contrast, the unstirred sample (24% MoS₂/g-C₃N₄-U) shows a drastically different profile. The atomic percentages of Mo (11.4%) and S (46.28%) are significantly higher, while those of C (16.28%) and N (14.72%) are much lower. In terms of weight ratio, Mo and S reach 34.6% and 46.95%, respectively, indicating a strong enrichment of MoS₂ within the detected region. This aligns with the intense MoS₂ diffraction peaks in XRD, clearly visible MoS₂ lattice fringes in HRTEM, and dense, flower-like morphology in SEM and TEM images—suggesting severe aggregation and poor integration with g-C₃N₄.

These structural and compositional differences directly correlate with photocatalytic performance. The sample synthesized with stirring exhibits superior hydrogen evolution activity, owing to the highly dispersed MoS₂ nanostructures that form extensive heterojunction interfaces with g-C₃N₄, enhancing charge separation and surface reactivity. On the other hand, in the absence of stirring, MoS₂ tends to aggregate and dominates the surface, reducing the number of effective interfacial contact sites and thus compromising the charge-transfer efficiency and catalytic activity.

2.1.4. Specific Surface Area Analysis and Correlation with Structural Properties

Table 2. BET surface areas of prepared samples.

Samples	g-C3N4	24% MoS2/g-C3N4-S	24% MoS2/g-C3N4-U	MoS2
BET surface area (m2/g)	62.6027	48.9016	27.6771	17.6937

summarizes the Brunauer–Emmett–Teller (BET) surface areas of the prepared samples. The pure g-C₃N₄ sample exhibits the highest specific surface area of 62.60 m²/g, reflecting its intrinsic lamellar structure with loosely stacked nanosheets and high porosity. After incorporation of 24 wt% MoS₂, the BET surface area decreases slightly to 48.90 m²/g for the stirred composite (24% MoS₂/g-C₃N₄-S), while a more drastic reduction to 27.68 m²/g is observed for the unstirred composite (24% MoS₂/g-C₃N₄-U). The pure MoS₂ sample exhibits the lowest BET surface area of 17.69 m²/g, which is attributed to its characteristic flower-like microsphere morphology. This structure consists of tightly stacked multilayered MoS₂ nanosheets forming dense spherical aggregates, as observed in the SEM and TEM images. Such a morphology limits the accessible surface area and reduces the number of exposed active sites for photocatalytic reactions.

These surface area differences closely align with the previously discussed morphological features. The stirred sample maintains a relatively high surface area due to the uniform deposition of highly dispersed MoS₂ nanostructures onto the g-C₃N₄ matrix, as evidenced by EDS mapping and the absence of significant MoS₂ diffraction peaks in XRD. The preservation of the lamellar g-C₃N₄ morphology in SEM and TEM further supports this structural integrity. In contrast, the unstirred sample exhibits severe MoS₂ agglomeration, leading to the collapse or blockage of g-C₃N₄’s porous network, thereby reducing the accessible surface area. This aggregation is also reflected in the high Mo and S elemental content detected in localized regions by EDS.

From a photocatalytic perspective, a larger specific surface area generally provides more active sites and improves light harvesting and reactant adsorption. Thus, the relatively higher BET surface area of the stirred composite is an important contributing factor to its superior hydrogen evolution performance. The combination of structural uniformity, better interfacial contact, and higher accessible surface reinforces the advantages of stirring-assisted hydrothermal synthesis in optimizing both material structure and catalytic activity.

2.2. Characterization of Photoelectrochemical Test Results

2.2.1. UV–Vis, PL, and TRPL Analysis

As shown in Figure 4a, pure g-C₃N₄ exhibits a clear absorption edge at ~460 nm, indicating limited visible-light utilization. MoS₂ shows broad absorption across the UV–Vis range. The 24% MoS₂/g-C₃N₄-S composite demonstrates enhanced absorption intensity, especially in the visible region, compared to pure g-C₃N₄. The enhancement is less significant in the unstirred composite, suggesting weaker light-harvesting efficiency.

Figure 4b shows the Tauc plots (n = 1/2) of the prepared samples for estimating their indirect optical bandgaps (Eg). The pure g-C₃N₄ sample exhibits an Eg of approximately 2.54 eV, consistent with reported values for this metal-free photocatalyst. For the stirred composite (24% MoS₂/g-C₃N₄-S), the bandgap is reduced to 1.77 eV, accompanied by a significant enhancement in optical absorption, indicating stronger interfacial interaction between g-C₃N₄ and MoS₂. In contrast, the unstirred composite (24% MoS₂/g-C₃N₄-U) shows no well-defined MoS₂-related indirect transition in the n = 1/2 plot, suggesting that MoS₂ in this sample does not exhibit a clear indirect band edge due to aggregation effects. Since MoS₂ in the composites is ultrathin and highly dispersed, its optical transition characteristics are better captured in the n = 2 Tauc plots (Figure S1), corresponding to direct-like transitions. In this case, the extracted Eg values are 1.47 eV for pure MoS₂, 1.75 eV for the unstirred composite, and 1.87 eV for the stirred composite. The blue-shift in the stirred sample compared with pure MoS₂ reflects enhanced dispersion and thickness reduction, leading to a more pronounced direct-like transition onset.

PL spectra (Figure 4c) provide insight into carrier recombination behavior. Pure g-C₃N₄ displays the strongest emission, indicating rapid charge recombination. The PL intensity of the stirred composite is significantly quenched, suggesting enhanced separation of photogenerated carriers. The unstirred composite shows less PL suppression, confirming that insufficient mixing limits interface formation and charge extraction.

TRPL results (Figure 4d,

Table 3. PL lifetimes of g-C₃N₄, 24% MoS₂/g-C₃N₄-S, 24% MoS₂/g-C₃N₄-U and MoS₂.

Samples	τ1 (ns)	α1	τ2 (ns)	α2	τave. (ns)
g-C3N4	3.015	64.14	21.93	35.86	4.89
24% MoS2/g-C3N4-S	2.134	51.57	42.7	48.43	10.89
24% MoS2/g-C3N4-U	1.817	66.87	11.35	33.13	2.48
MoS2	5.485	51.02	11.83	48.98	4.29

) reveal more details of carrier dynamics. The average carrier lifetime (τ_ave) of g-C₃N₄ is 4.89 ns, which increases to 10.89 ns in the stirred sample. However, average lifetime alone does not fully represent separation efficiency. Notably, in the stirred composite, the slow-decay component τ₂ reaches 42.7 ns and accounts for 48.43% (α₂), indicating a large portion of electrons are effectively extracted before recombination. In contrast, the unstirred sample has a much shorter τ₂ (11.35 ns) and lower α₂ (33.13%), implying faster recombination and inefficient charge utilization—likely caused by MoS₂ aggregation and interfacial discontinuity, as supported by EDS and TEM data.

These findings are consistent with XRD, SEM, and TEM analyses. The stirred sample exhibits homogeneous MoS₂ distribution, better crystallinity of g-C₃N4, and intimate MoS₂/g-C₃N₄ contact—creating effective heterojunctions that facilitate bandgap tuning, broaden absorption, and prolong carrier lifetimes. In contrast, the unstirred composite exhibits flower-like MoS₂ aggregation and limited heterointerface formation, which lead to stronger recombination and inferior optical and catalytic performance.

2.2.2. EPR, EIS, Photocurrent, Mott–Schottky

Figure 5a shows the electron paramagnetic resonance (EPR) spectra of g-C₃N₄ and 24% MoS₂/g-C₃N₄-S samples, measured under dark conditions. The observed signal at g ≈ 2.0027 typically corresponds to unpaired electrons associated with structural defects, such as nitrogen vacancies, or electrons trapped in localized π-conjugated defect states within the g-C₃N₄ matrix. Compared to pure g-C₃N₄, the stirred composite exhibits a significantly enhanced EPR signal intensity, indicating a higher density of defect-related paramagnetic centers. This suggests that the introduction of MoS₂ and the formation of a heterojunction facilitate charge redistribution and generate more surface traps or localized electron states. These trapped or defect-bound electrons are beneficial for suppressing rapid electron–hole recombination, thereby contributing to the prolonged lifetime and enhanced reduction ability of photogenerated electrons. Electrochemical impedance spectroscopy (EIS) results are presented in Figure 5b. The Nyquist plots show that the stirred composite exhibits the smallest arc radius, indicating the lowest interfacial charge transfer resistance. In contrast, both pristine g-C₃N₄ and the unstirred composite exhibit larger arcs, implying poorer charge transfer kinetics across the interface.

Figure 5c displays photocurrent response curves under chopped light illumination. The stirred composite shows the highest and most stable current density over repeated light on–off cycles, reflecting superior photo response, charge carrier separation, and surface reaction kinetics. The unstirred composite shows moderate enhancement but still lower photocurrent than the stirred sample.

Band edge analysis based on Mott–Schottky plots (Figure 5d) shows that the conduction band (CB) of g-C₃N₄ remains stable at −1.45 eV, which is consistent with previous studies [8,21,24,32], while the CB positions for the stirred and unstirred composites are −0.44 eV and −0.69 eV, respectively. The upward CB shift in the stirred sample is consistent with more efficient electron transport and improved heterojunction formation.

Collectively, these results confirm that the 24% MoS₂/g-C₃N₄-S composite synthesized under stirring possesses superior charge transfer efficiency, enhanced photogenerated electron density, and improved redox potential—all contributing to its outstanding photocatalytic hydrogen evolution performance.

2.3. Photocatalytic H₂ Evolution Performance

Figure 6a shows the time-dependent H₂ production curves over 5 h. Both pristine g-C₃N₄ and MoS₂ exhibit poor hydrogen evolution activity, confirming their limited standalone photocatalytic performance. In contrast, MoS₂/g-C₃N₄ composites with different loadings show enhanced H₂ production, with the 24% MoS₂/g-C₃N₄-S sample achieving the highest H₂ yield. The unstirred 24% sample performs significantly worse, indicating that stirring facilitates homogeneous MoS₂ dispersion and effective heterojunction formation, which promotes carrier mobility and exposure of active sites.

Figure 6b summarizes the corresponding hydrogen evolution rates. The rate increases from 8% to 24% MoS₂ loading, peaking at over 10 mmol·g⁻¹·h⁻¹, then slightly decreases at 32%, likely due to excessive MoS₂ coverage causing light-shielding or blocking catalytic sites. Notably, the unstirred 24% sample exhibits a much lower rate than its stirred counterpart, reaffirming the critical role of stirring in structural optimization.

Figure 6c explores the effect of catalyst dosage using the optimal 24% MoS₂/g-C₃N₄-S sample. The H₂ evolution rate reaches a maximum of 14.33 mmol·g⁻¹·h⁻¹ at a loading of 3.2 mg. Higher amounts result in reduced performance, possibly due to light scattering, mass transfer limitations, or reduced photon penetration. Compared to the HER rate values provided in Table S1, this performance is notably higher than many other reported composites, including those synthesized through conventional methods like hydrothermal calcination or ultrasonic dispersion. For example, the HER rate for similar composites such as MoS₂/g-C₃N₄ (no stirring) and MoS₂/g-C₃N₄ with calcination falls in the range of 0.2 to 12 mmol·g⁻¹·h⁻¹, while your composite demonstrates a significant enhancement in photocatalytic efficiency, clearly surpassing these values. This highlights the effectiveness of the stirring-assisted in situ synthesis method in optimizing MoS₂ dispersion and enhancing charge separation, leading to superior hydrogen evolution performance.

While compared with previous studies, the performance lies within a moderate range compared to other advanced photocatalysts. Several factors could be limiting the enhancement of HER activity in our system: (1) MoS₂ dispersion and edge exposure, (2) charge recombination, (3) electronic conductivity and the MoS₂-g-C₃N₄ interface, and (4) catalyst stability. Future work will focus on optimizing MoS₂ loading, enhancing charge separation efficiency, improving electron transport, and enhancing long-term stability.

2.4. Mechanism Explanation

Figure 7 depicts the mechanism of photocatalytic hydrogen evolution over MoS₂/g-C₃N₄ composites synthesized under stirred and unstirred hydrothermal conditions, focusing on band alignment, charge dynamics, and interface interactions. Mott–Schottky measurements show that pristine g-C₃N₄ has a flat-band potential of −1.45 V (vs. Ag/AgCl), indicating a highly negative conduction band edge suitable for efficient proton reduction. MoS₂, on the other hand, exhibits a conduction band (CB) edge at −0.44 V for the stirred sample and −0.69 V for the unstirred composite, with bandgaps of 1.47 eV (MoS₂), 1.75 eV (unstirred composite), and 1.87 eV (stirred composite), respectively. These findings suggest that MoS₂ in the stirred sample adopts a more direct-like transition due to its ultrathin, highly dispersed structure. Band position comparisons reveal that both samples exhibit characteristics consistent with a Type-II heterojunction structure [36,37,38], where g-C₃N₄ valence band (VB) and CB are positioned entirely above those of MoS₂, facilitating efficient charge separation. The stirred composite (24% MoS₂/g-C₃N₄-S) demonstrates a positively shifted flat-band potential (−0.44 V) compared to the unstirred sample, which shows a more negative −0.69 V, suggesting improved interfacial band alignment in the stirred sample.

In addition, photocatalytic tests show a significant performance gap: the stirred sample exhibits a much higher H₂ evolution rate (up to 14.33 mmol·g⁻¹·h⁻¹), far surpassing the unstirred counterpart. This discrepancy is primarily attributed to the enhanced MoS₂ dispersion and the presence of active co-catalytic edge sites in the stirred sample, rather than differences in band structure alone.

XRD, SEM, TEM, and EDS mapping confirm that the stirred sample retains the crystallinity of g-C₃N₄, with highly dispersed MoS₂ nanosheets and abundant edge exposure on the g-C₃N₄ surface. BET measurements further show an increased surface area, providing more catalytic sites for the hydrogen evolution reaction.

Electrochemical analyses reinforce these findings: EIS reveals reduced charge transfer resistance, and transient photocurrent responses indicate enhanced carrier separation. The Mott–Schottky plot suggests beneficial interfacial band bending in the stirred sample, facilitating electron transfer. EPR and TRPL data further support the improved charge carrier density and lifetime in the stirred composite.

While both systems are classified as Type-II heterojunctions, the stirred composite introduces a synergistic enhancement mechanism: highly dispersed MoS₂ nanosheets act as co-catalytic edge sites, enabling fast photogenerated electron transfer and providing efficient active sites for H₂ evolution. These edge sites overcome the recombination limitations typically associated with Type-I heterojunctions and further enhance the catalytic activity.

Therefore, the significantly improved photocatalytic performance of the MoS₂/g-C₃N₄-S composite can be attributed to the synergistic effects of multiple structural and electronic factors. First, the formation of a structurally favorable Type-II heterojunction facilitates efficient charge migration, with photogenerated electrons transferring smoothly from the g-C₃N₄ conduction band into MoS₂. Second, the highly dispersed MoS₂ nanosheets expose abundant edge sites that act as active co-catalytic centers for the H₂ evolution reaction, enhancing the overall catalytic performance. Third, the optimized morphology and improved interfacial contact between g-C₃N₄ and MoS₂ significantly promote charge carrier separation and mobility. These combined effects suppress electron–hole recombination and optimize electron transfer, leading to a remarkable improvement in photocatalytic hydrogen evolution efficiency.

3. Materials and Methods

3.1. Materials

Thiourea (CH₄N₂S, analytical reagent grade), ammonium molybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O, analytical reagent grade), triethanolamine (C₆H₁₅NO₃, analytical reagent grade), and absolute ethanol (C₂H₆O, analytical reagent grade) were all purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China. N,N-Dimethylformamide (C₃H₇NO, analytical reagent grade) was obtained from Aladdin Industrial Corporation, City, China. All chemicals were used as received without further purification.

3.2. Synthesis of the Photocatalyst

All chemicals used in this study were of analytical reagent grade and used without further purification.

Figure 8 illustrates the synthesis procedure used in this study. We employed a stirred hydrothermal method to improve uniformity during the growth of MoS₂/g-C₃N₄ composites. We hypothesize that stirring inhibits MoS₂ crystallization in solution, allowing nucleation mainly on the g-C₃N₄ substrate, thus forming highly dispersed and strongly interacting MoS₂/g-C₃N₄ composites. In contrast, traditional hydrothermal synthesis permits MoS₂ nucleation both on g-C₃N₄ and in solution, leading to flower-like MoS₂ structures attached to g-C₃N₄. The detailed information are as follow.

3.2.1. Synthesis of g-C₃N₄

As shown in Figure 8, the bulk g-C₃N₄ was synthesized via direct thermal condensation of melamine. Specifically, 0.6 g of melamine powder was placed in a covered alumina crucible and heated in a tubular furnace under continuous air flow. The sample was heated to 550 °C at a rate of 5 °C min⁻¹ and held at this temperature for 4 h. After natural cooling to room temperature, the resulting light-yellow solid was collected and ground into fine powder for subsequent use.

3.2.2. Synthesis of MoS₂@g-C₃N₄ Composite Catalysts

In a typical synthesis as show in Figure 8, 0.6 g of as-prepared g-C₃N₄ powder was dispersed in 90 mL of DMF under magnetic stirring and then ultrasonicated for 30 min. Subsequently, 120 mg of ammonium molybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O) was added to the suspension and ultrasonicated for another 30 min. Then, 240 mg of thiourea was introduced and the suspension was again ultrasonicated for 30 min to ensure thorough mixing.

After the final sonication, the suspension was gently shaken and immediately transferred to a Teflon-lined stainless steel autoclave equipped with a magnetic stirrer. The mixture was first stirred at room temperature at 600 rpm for 4 h, followed by hydrothermal treatment at 200 °C for 24 h under continuous stirring (600 rpm). After natural cooling to room temperature (~2 h), the product was collected by vacuum filtration through moistened filter paper. The solid was sequentially washed with DMF, absolute ethanol, and deionized water to remove unreacted precursors and residues. Finally, the sample was dried in a vacuum oven at 60 °C for 12 h and ground into fine powder for storage and further characterization.

3.2.3. Sample Variation and Control Design

To investigate the effect of MoS₂ loading and stirring conditions on photocatalytic performance, a series of MoS₂@g-C₃N₄ composites were prepared by varying the precursor amounts to achieve different MoS₂ loadings: 8%, 16%, 24%, and 32% by weight. All samples were prepared under the same conditions except for the amount of precursors. Samples synthesized with magnetic stirring during the hydrothermal process were labeled as “-S” (e.g., 24% MoS₂/g-C₃N₄-S), while those synthesized without stirring were denoted as “-U” (e.g., 24% MoS₂/g-C₃N₄-U), serving as the unstirred control. Pure g-C₃N₄ and pure MoS₂ were also synthesized and included as reference samples.

3.3. Characterization

The phase structures of the as-prepared samples were identified by X-ray diffraction (XRD) using a Bruker D8 Advance diffractometer (Bruker AXS GmbH, Karlsruhe, Germany) equipped with Cu Kα radiation (λ = 1.5406 Å), operated at 40 kV and 40 mA. Surface morphology and microstructural features were examined through field-emission scanning electron microscopy (FESEM, Hitachi S-4800, Hitachi High-Technologies Corporation, Tokyo, Japan) and high-resolution transmission electron microscopy (HRTEM, JEOL JEM-2100F, JEOL Ltd., Tokyo, Japan). Elemental mapping and local composition were further characterized by energy-dispersive X-ray spectroscopy (EDS) integrated into the TEM system.

The specific surface area and porosity were evaluated via nitrogen adsorption–desorption isotherms collected at 77 K using a Micromeritics ASAP 2020 system (Micromeritics Instrument Corporation, Norcross, GA, USA). Molybdenum content in the composite materials was quantitatively determined by inductively coupled plasma optical emission spectrometry (ICP-OES, Agilent 5110, Agilent Technologies Inc., Santa Clara, CA, USA).

3.4. Photoelectrochemical Test

To investigate optical and photoelectrochemical behaviors, UV–Vis diffuse reflectance spectra (DRS) were recorded on a Shimadzu UV-2600 instrument (Shimadzu Corporation, Kyoto, Japan), and the bandgap energy was calculated based on the corresponding Tauc plots. Steady-state photoluminescence (PL) spectra were measured with a Hitachi F-7000 spectrofluorometer (excitation wavelength: 325 nm, Hitachi, Tokyo, Japan) to study carrier recombination. Carrier lifetimes were estimated through time-resolved PL (TRPL) spectroscopy using an Edinburgh FLS980 fluorescence spectrometer (Edinburgh Instruments, Edinburgh, UK).

Electron paramagnetic resonance (EPR) was employed using a Bruker EMXplus (Bruker Corporation, Billerica, United States) to detect paramagnetic surface species and defect signals. Electrochemical impedance spectroscopy (EIS), transient photocurrent measurements, and Mott–Schottky (M–S) analyses were conducted on a CHI660E electrochemical workstation (CH Instruments, Shanghai, China) in a standard three-electrode configuration, where a Pt wire served as the counter electrode, Ag/AgCl as the reference electrode, and the sample-coated glassy carbon electrode as the working electrode. All electrochemical tests were carried out in 0.2 M Na₂SO₄ solution under dark or illuminated conditions, with illumination provided by a 300 W xenon lamp equipped with an AM 1.5G filter.

3.5. Photocatalytic Activity Evaluation

Photocatalytic hydrogen evolution experiments were carried out using a Labsolar-6A automated photocatalytic analysis workstation (Perfectlight, Beijing, China). In a typical run, the catalyst was dispersed in 50 mL of aqueous solution containing 20 vol% triethanolamine (TEOA), which served as a sacrificial electron donor. Prior to illumination, the suspension was sonicated for 10 min and purged with high-purity nitrogen gas for 30 min to remove residual oxygen.

A 300 W xenon arc lamp (PLS-SXE300D, Perfectlight, Beijing, China) was used as the irradiation source without any optical filter, thus providing full-spectrum light. The reaction suspension was continuously stirred using a magnetic stirrer to maintain homogeneity and prevent catalyst sedimentation. To examine the influence of photocatalyst dosage on hydrogen production, a series of tests were conducted using catalyst masses ranging from 2.0 to 5.6 mg under otherwise identical conditions.

The evolved hydrogen was collected and quantified using the gas analysis interface of the Labsolar-6A system, which is equipped with a built-in gas chromatograph (Fuli GC-9790Plus, Fuli, Hangzhou, China). The GC system utilized a 5A molecular sieve column and a thermal conductivity detector (TCD), with nitrogen as the carrier gas. The amount of hydrogen generated was calculated by comparing the GC peak areas with standard calibration curves.

The rate of hydrogen evolution (R_H2) was calculated using the following equation:

$${{\displaystyle R}}_{{{\displaystyle H}}_2}={\displaystyle \frac{{{\displaystyle n}}_{{{\displaystyle H}}_2}}{m\times t}}$$

(1)

where n_H2 is the amount of hydrogen (in mmol), m is the mass of the photocatalyst (in grams), and t is the irradiation time (in hours).

For catalyst stability evaluation, the optimized sample was subjected to four consecutive 5-h photocatalytic cycles under identical experimental conditions. After each cycle, the photocatalyst was recovered by centrifugation, washed thoroughly with deionized water and ethanol, dried at 60 °C, and reused in the subsequent run. The durability of the catalyst was assessed by comparing the hydrogen production rates across the repeated cycles.

4. Conclusions

This study proposes a stirring-assisted hydrothermal in situ synthesis strategy to construct a MoS₂/g-C₃N₄ heterojunction photocatalyst with highly dispersed MoS₂ nanosheets. Unlike conventional static hydrothermal methods, the introduction of agitation during synthesis significantly improves precursor mixing, facilitates the uniform nucleation and growth of MoS₂, and ensures the exposure of abundant catalytic edge sites on the g-C₃N₄ surface. Structural characterizations confirm that the stirred samples exhibit superior interfacial contact, reduced aggregation, and well-defined lattice continuity, while the unstirred samples suffer from MoS₂ clustering and poor interface formation. Multidimensional characterizations, including XRD, SEM/TEM, EDS mapping, BET, and ICP analyses, together with optoelectronic measurements (UV–Vis, TRPL, EIS, EPR, Mott–Schottky), collectively demonstrate that the high dispersion of MoS₂ facilitates charge separation, improves electron mobility, and provides abundant active sites at the MoS₂ edges for hydrogen evolution. As a result, the optimized 24% MoS₂/g-C₃N₄-S catalyst achieves a maximum hydrogen production rate of 14.33 mmol·g⁻¹·h⁻¹ at a catalyst dosage of 3.2 mg, accompanied by excellent cycling stability. The superior performance is attributed to three critical factors: (i) the stirring-assisted in situ method enables controlled interface construction and heterojunction formation; (ii) the highly exposed edge-rich MoS₂ structure functions as an effective co-catalyst and electron extraction channel; and (iii) the engineered morphology enhances visible-light absorption and charge migration pathways. Overall, this work offers a novel and scalable approach to co-catalyst dispersion and interface engineering, providing mechanistic insights into heterojunction design for efficient solar-to-hydrogen energy conversion using non-noble metal materials.