1. Introduction
Photocatalytic water splitting, a promising solution for addressing the limitations of fossil fuels, utilizes abundant and renewable energy sources [
1]. This process harnesses solar energy to produce hydrogen in a clean and environmentally friendly manner, offering economic and ecological advantages. However, the efficient utilization of the solar spectrum is crucial, and develo** cost-effective, stable, and efficient photocatalysts that can harness visible and UV light is a crucial challenge [
2]. Factors such as crystallinity, particle size, band gap, and redox stability influence the performance of photocatalysts in water splitting [
3,
4]. Achieving efficient, scalable, and sustainable hydrogen production requires photocatalysts that meet specific criteria, including suitable band gaps, precise band alignment, stability, low production cost, recyclability, abundance, corrosion resistance, and suitability for large-scale production [
5,
6].
Graphite-like carbon nitride (g-C
3N
4) exhibits exceptional electric, optical, structural, and physiochemical properties, rendering it a versatile nano-platform for diverse electronic, catalytic, and energy applications [
7,
8,
9]. Since its discovery in 2009, g-C
3N
4-based photocatalysts, particularly for generating H
2 and O
2, have garnered significant attention, leading to a surge in publications and citations [
10,
11]. These nanostructures and graphene-based photocatalysts are increasingly considered prime contenders for numerous energy and environmental endeavors, including photocatalytic water splitting, pollutant degradation, and carbon dioxide mitigation [
12,
13,
14,
15]. The insufficient photocatalytic activity of bulk g-C
3N
4, attributed to limited solar-light absorption, low electrical conductivity, and fast charge carrier recombination, poses challenges for practical applications [
11,
16,
17,
18]. Extensive research has been conducted to enhance its photoactivity through various modification approaches, such as do**, nanostructure design, and integration with different materials [
19,
20,
21,
22,
23]. Promising outcomes have been achieved by constructing nanostructures through heterojunction formation or controlling morphology and incorporating g-C
3N
4 with semiconductors, carbon-based materials, or metal nanoparticles [
24,
25,
26,
27,
28]. Moreover, achieving superior performance in g-C
3N
4 modification relies on precise control of interfacial contacts at heterojunctions and the customization of nanostructures with surface functionality [
19,
29].
Ternary semiconductor chalcogenides, such as ZnIn
2S
4, are gaining attention for their potential as visible light active photocatalysts due to their excellent chemical stability and optical band gap [
30,
31]. ZnIn
2S
4 stands out with its layered structure, where Zn and In atoms are arranged in distinct environments, leading to improved photocatalytic performance. In
3+ metal ions with a d
10 configuration further enhance the photocatalytic activity, making ZnIn
2S
4 a promising candidate for water splitting [
30,
32,
33]. The photocatalytic activity of ZnIn
2S
4 nanosheets needs improvement due to the short lifetime and high recombination rate of electron-hole pairs and the agglomeration of nanosheets [
34,
35]. To address these issues, coupling ZnIn
2S
4 nanosheets with hollow nanostructures can alleviate agglomeration and promote charge separation [
36,
37]. One promising approach is the construction of a composite hollow heterojunction photocatalyst based on ZnIn
2S
4 nanosheets and g-C
3N
4, which has a favorable energy band structure and easy loading capability [
37,
38]. Previous studies have reported the successful synthesis of ZnIn
2S
4/g-C
3N
4 nanocomposites with enhanced photocatalytic performance for various essential applications [
39,
40,
41,
42]. Despite their remarkable efficiency in photocatalytic hydrogen production, no research has investigated the extended durability of g-C
3N
4@ZnIn
2S
4 heterostructures under low-wattage light sources. By exploring this aspect, we can optimize photocatalyst utilization and achieve significant energy conservation from the light source.
In this study, we successfully synthesized g-C3N4 nanostructures and g-C3N4@ZnIn2S4 heterostructures and thoroughly investigated their photocatalytic performance for water splitting. The results demonstrate that the flower-like C3N4@ZnIn2S4 heterostructures exhibit significantly enhanced photocatalytic activity compared to g-C3N4 nanostructures. Furthermore, we elucidated the mechanism underlying the superior photocatalytic performance of C3N4@ZnIn2S4 heterostructures.
2. Results and Discussion
Figure 1 shows the synthesis processes of g-C
3N
4@ZnIn
2S
4 heterostructures prepared by combining thermal annealing and hydrothermal methods. First, g-C
3N
4 nanostructures were synthesized by thermal annealing at a high temperature of 550 °C for 3 h. Second, the different weights of g-C
3N
4 nanostructures were dispersed in the ZnIn
2S
4 reaction precursors (1 mM ZnCl
2, 2.5 mM InCl
3, and 5 mM TAA) and heated at 160 °C for 12 h by a facile hydrothermal method. In order to investigate the morphological features of the synthesized g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures, field-emission scanning electron microscopy (FESEM) was utilized and depicted in
Figure 2.
Figure 2a illustrates the stacked arrangement of g-C
3N
4 nanostructures composed of irregularly layered nanosheets. The FESEM image in
Figure 2b depicts the flower-like microsphere structure of the g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures. These microspheres exhibit a hierarchical architecture composed of numerous ultrathin nanosheets. A sheet-like structure enhances the availability of active surface sites, making it highly favorable for facilitating photocatalytic reactions [
43].
Figure 3 displays the XRD patterns, providing insights into the composition and structure of the synthesized g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures. Notably, the diffraction peak observed in the XRD pattern of g-C
3N
4 nanostructures (
Figure 3a) is positioned at 27.3°, corresponding to the crystal plane (002) of g-C
3N
4 (JCPDS No. 87–1526). In the g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures (
Figure 3b), the XRD analysis reveals distinct characteristic peaks corresponding to various crystal planes of the hexagonal phase ZnIn
2S
4 (JCPDS No. 72–0773). These peaks are observed at 2θ angles of 21.6°, 27.7°, 30.5°, 39.8°, 47.2°, 52.4°, and 55.6°, corresponding to the (006), (102), (104), (108), (110), (116), and (202) crystal planes, respectively. However, in the g-C
3N
4@ZnIn
2S
4 heterostructures, the diffraction peak of g-C
3N
4 at 2θ = 27.3° is either of weak intensity or obscured by other diffraction peaks. These characteristic peaks and the absence of impurity peaks indicate the successful synthesis of g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures.
The N
2 adsorption–desorption isotherms were utilized to investigate the specific surface areas of g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures.
Figure 4a,b show that both g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures exhibit an H3 model hysteresis loop (P/P
0 = 1) in their N
2 adsorption–desorption isotherms, indicating the presence of mesoporous features in the synthesized materials, respectively. In addition, the Brunauer–Emmett–Teller (BET) of g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures are 73.13 m
2g
−1 and 91.65 m
2g
−1, respectively. g-C
3N
4 nanostructures decorated with ZnIn
2S
4 nanostructures to form g-C
3N
4@ZnIn
2S
4 heterostructures can increase their specific surface area. The Fourier Transform Infrared spectroscopy (FTIR) spectra of g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures are depicted in
Figure 4c. In these FTIR spectra, a distinct peak at 811 cm
−1 originates from the characteristic breathing vibration of the triazine ring in g-C
3N
4 [
37,
44]. Additionally, absorption bands within the 1200–1640 cm
−1 range signify the stretching vibration modes of the C, N-heterocyclic groups present in g-C
3N
4 [
45,
46]. Weaker peaks appearing in the 3000–3600 cm
−1 range can be attributed to the N–H characteristic vibration of -NHx and the O–H characteristic vibration of residual hydroxyl groups or absorbed water molecules [
46,
47]. These observations validate the presence of g-C
3N
4 nanostructures within the g-C
3N
4@ZnIn
2S
4 heterostructures, as evidenced by the FTIR spectrum.
Figure 5a displays the FETEM image of g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures, exhibiting a consistent morphology with the nanosheet structure observed in the FESEM image formed by lamellar stacks. Furthermore,
Figure 5b illustrates the selected area electron diffraction (SAED) patterns, revealing distinct polycrystalline diffraction rings that substantiate the polycrystalline nature of the g-C
3N
4@ZnIn
2S
4 heterostructures. These diffraction rings correspond to crystal planes (006), (102), (104), (108), (110), (116), and (202), indicating the typical hexagonal structure of ZnIn
2S
4 (JCPDS card No. 72-0773), which aligns with the XRD analysis results and confirms the polycrystalline properties of the ZnIn
2S
4 nanosheets. In
Figure 5c, the high-resolution transmission electron microscopy (HRTEM) image reveals the presence of lattice fringes exhibiting spacings of 0.226 nm and 0.193 nm, corresponding to the (108) and (110) planes of hexagonal ZnIn
2S
4, respectively. The amorphous regions observed in the image are attributed to g-C
3N
4. Furthermore,
Figure 5d displays the map** images obtained from energy dispersive spectroscopy (EDS) analysis, which demonstrate a uniform and interconnected distribution of nitrogen (N), zinc (Zn), indium (In), and sulfur (S) elements. These comprehensive findings provide conclusive evidence for successfully synthesizing g-C
3N
4@ZnIn
2S
4 heterostructures, showcasing the well-defined interface between the two materials.
X-ray photoelectron spectroscopy (XPS) was performed to analyze the chemical states of the synthesized samples, including g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures. The survey spectra (
Figure 6a) revealed the presence of primary elements, such as C, N, Zn, In, S, and O. The appearance of the O 1s signal suggests the adsorption of oxygen atoms [
48]. The XPS spectrum (
Figure 6b) illustrates the C 1s signal for g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures. Three distinct peaks are observed, corresponding to C-C bonds, sp
3-hybridized C, and sp
2-hybridized C, respectively [
42]. The XPS spectra of N 1s for g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures are shown in
Figure 6c. Three distinct peaks are observed in the XPS spectra, corresponding to the N-H group, tertiary N in the aromatic ring, and sp
2-hybridized N in the triazine ring, respectively [
49,
50]. The high-resolution XPS spectrum (
Figure 6d) of Zn 2p displays two distinct peaks at 1021.4 eV and 1044.4 eV, corresponding to Zn 2p
3/2 and Zn 2p
1/2, respectively [
51]. These peaks indicate the presence of Zn
2+ in the ZnIn
2S
4 structure. The XPS spectrum of In 3d (
Figure 6e) reveals the presence of two distinct peaks at 444.6 and 452.1 eV, corresponding to the In 3d
3/2 and In 3d
5/2 states, respectively, indicating the presence of In
3+ in ZnIn
2S
4 [
51]. The high-resolution XPS spectrum (
Figure 6f) of S 2p reveals the presence of two prominent peaks at 161.3 and 162.6 eV, which can be attributed to the S 2p
3/2 and S 2p
1/2 states in ZnIn
2S
4, respectively [
52].
Figure 7a illustrates the assessment of the recombination ability of electrons and holes based on the photoluminescence intensity of the g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures. A higher intensity in the PL emission spectra indicates a diminished carrier separation ability. Remarkably, the PL spectrum of g-C
3N
4@ZnIn
2S
4 heterostructures exhibited lower emission intensity than that of g-C
3N
4 nanostructures, reducing the recombination frequency of electrons and holes within the heterostructures. A lower peak in the photoluminescence spectra suggests the potential for superior catalytic performance. Therefore, g-C
3N
4@ZnIn
2S
4 heterostructures may outperform g-C
3N
4 nanostructures regarding photocatalytic efficiency.
UV-visible diffuse reflectance spectroscopy (UV-vis DRS) was employed to investigate the optical properties of g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures. As depicted in
Figure 7b, the g-C
3N
4@ZnIn
2S
4 heterostructures exhibit significantly enhanced light absorption capability compared to g-C
3N
4 nanostructures across the wavelength range of 300 nm to 800 nm. This broader light harvesting ability of g-C
3N
4@ZnIn
2S
4 heterostructures is advantageous for efficient solar utilization, facilitating photocatalytic hydrogen production.
The energy band gaps (
Eg) were determined using the Kubelka–Munk method, which is expressed as follows [
53]:
In this equation, A, α, ν, Eg, and h represent constants, the absorption coefficient, the frequency of light, the band gap energy, and Planck’s constant, respectively. The parameter “n” denotes the characteristic of the semiconductor material, where it equals 1 for indirect bandgap semiconductors and 1/2 for direct bandgap semiconductors. Previous literature indicates that g-C3N4 possesses a direct band gap, thus setting the value of “n” as 1/2. The energy band gap value of g-C3N4 nanostructures was calculated at about 2.90 eV. The g-C3N4@ZnIn2S4 heterostructures exhibited a narrower band gap (2.48 eV), indicating the influence of incorporating ZnIn2S4 and g-C3N4.
Figure 8a depicts the results of electrochemical impedance spectroscopy (EIS) for g-C
3N
4 nanostructures, g-C
3N
4@ZnIn
2S
4 heterostructures with different weights of g-C
3N
4 nanostructures, and ZnIn
2S
4 nanostructures. A notable difference in the EIS Nyquist curves’ arc radii is that g-C
3N
4 nanostructures exhibit larger arc radii than g-C
3N
4@ZnIn
2S
4 heterostructures. The g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures reveal the lowest arc radii. In addition, the charge transfer resistance values of g-C
3N
4 nanostructures, g-C
3N
4@ZnIn
2S
4 heterostructures with different weights of g-C
3N
4 nanostructures, and ZnIn
2S
4 nanostructures, as shown in
Table S1. The g-C
3N
4@ZnIn
2S
4 heterostructure with 0.01 g g-C
3N
4 nanostructure exhibits the lowest charge transfer resistance values. This result indicates that g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures possess the lowest charge transfer resistance to enhance efficiency in separating charge carriers, facilitating the fastest electron transfer process. The reduced arc radius in g-C
3N
4@ZnIn
2S
4 heterostructures signifies improved charge transfer kinetics and highlights the potential for efficient photocatalytic performance [
54,
55].
In order to investigate the charge-transfer properties, transient photocurrent responses of g-C
3N
4 nanostructures, g-C
3N
4@ZnIn
2S
4 heterostructures with different weights of g-C
3N
4 nanostructures, and ZnIn
2S
4 nanostructures were conducted, as depicted in
Figure 8b. Notably, the g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures exhibited significantly higher photocurrent density than g-C
3N
4 nanostructures and g-C
3N
4@ZnIn
2S
4 heterostructures with other weights, indicating a substantial improvement in the efficiency of separating photogenerated electrons and holes. The obtained result aligns with the earlier findings from PL and EIS measurements, further supporting the conclusion. Incorporating ZnIn
2S
4 onto g-C
3N
4 can promote the efficient transfer of electron-hole pairs and enhance the potential for photocatalytic hydrogen production.
The photocatalytic performance of the synthesized g-C
3N
4 nanostructures, g-C
3N
4@ZnIn
2S
4 heterostructures with different weights of g-C
3N
4 nanostructures, and ZnIn
2S
4 nanostructures was evaluated by measuring the hydrogen evolution rate (HER) in a 50 mL deionized (DI) water solution with 50% triethanolamine (TEOA) serving as a scavenger under visible light irradiation, as shown in
Figure 9a. The pH value of the solution was not adjusted. The average HER values of as-prepared photocatalysts were 10.4 (g-C
3N
4 nanostructures), 2056.2 (0.005 g g-C
3N
4 nanostructures), 2377.6 (0.01 g g-C
3N
4 nanostructures), 1355.1 (0.025 g g-C
3N
4 nanostructures), 448.7 (0.05 g g-C
3N
4 nanostructures), and 921.2 μmolh
−1g
−1L
−1 ZnIn
2S
4 nanostructures, respectively. The average HER of g-C
3N
4@ZnIn
2S
4 heterostructures gradually increased as the weight of g-C
3N
4 nanostructures increased. However, a notable decline in the average HER was observed when the weight of g-C
3N
4 nanostructures exceeded 0.01 g. This outcome could be attributed to the higher weights of g-C
3N
4 nanostructures, which might cause an excessive generation of g-C
3N
4 nanostructures and subsequently decrease the efficiency of electron-hole pair transfer, thereby inhibiting the overall photocatalytic hydrogen production efficiency. This result is consistent with the above EIS and photocurrent response measurements. In addition, g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures revealed almost 228.6 and 2.58 times higher than g-C
3N
4 nanostructures and ZnIn
2S
4 nanostructures, respectively.
In order to investigate the impact of sacrificial agents on the hydrogen production performance of the hybrid system, various sacrificial agents, including methanol, ethanol, ethylene glycol (EG), and TEOA, were employed in the g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures.
Figure 9b demonstrates the average HER of g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures in the presence of methanol, ethanol, ethylene glycol (EG), and TEOA solutions, which were recorded as 16.3, 96.5, 10.9, and 2377.6 μmolh
−1g
−1L
−1, respectively. These results indicate that TEOA is a more suitable choice as a sacrificial agent for the g-C
3N
4@ZnIn
2S
4 heterojunction photocatalyst in hydrogen evolution. The observed outcome can be attributed to the effective binding of TEOA on the catalyst surface, which helps prevent photocorrosion and degradation of the g-C
3N
4 base photocatalysts [
56,
57,
58]. TEOA efficiently scavenges photogenerated holes, enhances the dispersion of photocatalysts, and acts as a binding ligand to improve the interaction between g-C
3N
4 and water molecules [
58].
Using concentrated sacrificial reagents is advantageous in promoting efficient diffusion of reacting species towards the surface of photocatalysts [
59]. However, it is essential to consider that achieving the highest hydrogen evolution rate is impossible with diluted or highly concentrated sacrificial reagents due to their respective limitations [
48]. Balancing the concentration of sacrificial reagents is crucial to optimize the performance of the photocatalytic system.
Figure 9c illustrates the influence of TEOA concentrations on the photocatalytic efficiency of g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures. The average HER values for the g-C
3N
4@ZnIn
2S
4 heterostructures were recorded as follows: 0 (without TEOA), 1152.3 (20% TEOA), 2377.6 (50% TEOA), and 671.5 μmolh
−1g
−1L
−1 (100% TEOA). It was observed that g-C
3N
4@ZnIn
2S
4 heterostructures with appropriate TEOA concentrations displayed the highest HER under visible-light irradiation.
The efficient recycling of catalysts is a crucial consideration in photocatalysis [
60]. In order to reduce waste and ensure sustainable processes, it is desirable to design photocatalysts that maintain consistent photoactivity throughout each cycle [
61]. If photocatalysts gradually lose activity over time, they contribute to waste generation during their life cycle. Furthermore, it is essential to develop easily separable and recyclable photocatalysts to prevent the loss of valuable materials in the waste stream [
62]. Herein, eight cycles of photocatalytic hydrogen generation were conducted to assess the stability of g-C
3N
4@ZnIn
2S
4 heterostructures with 0.01 g g-C
3N
4 nanostructures. The average HER of the g-C
3N
4@ZnIn
2S
4 heterostructures remained consistently high throughout the eight cycles (
Figure 10a). Furthermore, the XRD pattern of the sample after eight cycles (
Figure 10b) showed no new peaks. The FESEM (
Figure 10c) image and FESEM-EDS map** image (
Figure 10d) of the sample after eight cycles still reveal a similar morphology and uniform element distribution. These results indicate the exceptional stability of the g-C
3N
4@ZnIn
2S
4 heterostructures. Comparative analysis of the photocatalytic hydrogen production activity of g-C
3N
4@ZnIn
2S
4 heterostructures with other reported photocatalysts strongly supports their favorable application prospects, as shown in
Table S2 [
25,
63,
64,
65]. The promotion of stable performance can be attributed to the enhanced photoinduced charge separation achieved through efficient electron transfer from g-C
3N
4 to ZnIn
2S
4. These findings confirm the excellent stability and reusability of g-C
3N
4@ZnIn
2S
4 heterostructures and highlight their potential for broader and diverse applications in various fields.
Figure 11 presents the photocatalytic hydrogen production mechanism of g-C
3N
4@ZnIn
2S
4 heterostructures, as deduced from the results above. In this study, we employ ion exchange resin to coat the materials (g-C
3N
4 and ZnIn
2S
4) onto indium tin oxide (ITO) glass, followed by measuring the flat band potential using cyclic voltammetry [
66,
67]. The conduction band (CB) positions of g-C
3N
4 and ZnIn
2S
4 are −1.40 eV and −0.58 eV, respectively, while their valence band (VB) positions are 1.50 eV and 1.78 eV [
68,
69]. Under visible light irradiation (λ
max = 420 nm), g-C
3N
4 and ZnIn
2S
4 materials undergo excitation, generating photogenerated electrons in the VB transitioning to the CB, consequently creating holes in the VB. The construction of the heterojunction facilitates the transfer of electrons from the conduction band (CB) of g-C
3N
4 to the lower surface of ZnIn
2S
4, while the photogenerated holes are transferred from the VB of ZnIn
2S
4 to the VB of g-C
3N
4. This formation of a type II heterojunction at the interface between g-C
3N
4 and ZnIn
2S
4 greatly enhances the separation efficiency of the photogenerated electron-hole pairs in g-C
3N
4. Additionally, due to the more negative conduction potential of ZnIn
2S
4 compared to the reduction potential of H
+/H
2, the photogenerated electrons accumulated in the CB of ZnIn
2S
4 undergo a reduction reaction, effectively reducing H
+ in an aqueous solution to produce H
2. Simultaneously, the presence of TEOA serves to consume the photogenerated holes. This synergistic effect significantly reduces the possibility of carrier recombination, resulting in a notable increase in the photocatalytic activity of g-C
3N
4@ ZnIn
2S
4 and promoting efficient hydrogen generation. This intricate photocatalytic process facilitates efficient charge carrier separation and promotes hydrogen production. Additionally, the g-C
3N
4@ZnIn
2S
4 heterostructures substantially enhance light-harvesting capacity, further boosting the efficiency of photocatalytic hydrogen production.