Hydrogen, which can be applied in generating abundant clean and renewable green energy, has attracted significant attention amidst the growing energy and environmental pollution crisis [1]. Photocatalytic H₂ production from water splitting has proved to be an environmentally friendly strategy for generating renewable solar energy, with the advantages of replacing fossil fuels and reducing environmental pollution [2, 3]. CdS, with a special band gap of 2.4 eV and an appropriate band level, has long been considered as a promising photocatalyst [4, 5]. However, the strong photocorrosion of CdS photocatalysts and the rapid combination of the e^––h⁺ pairs lead to their low activity and poor stability, thus hindering their wide application [6, 7]. Therefore, to improve the separation of the e^––h⁺ pairs and photocatalytic activity and stability, extensive modification studies have been carried out to boost the photocatalytic hydrogen production over CdS photocatalysts [8, 9], such as the construction of heterogeneous junctions [2, 10-13], co-catalyst loading [14-20], the design of Z-type systems [21-28], coupling with nanocarbon materials [9, 29-31], and the exploration of nanostructures [10, 14, 17, 32-40]. Typically, loading co-catalyst over nanostructured CdS photocatalysts is one of the most effective and convenient methods of improving their catalytic activity toward hydrogen production [41]. In previous researches, various morphologies of CdS photocatalysts such as nanorods [17, 34, 35, 37, 42, 43], quantum dots [44], nanosheets [45, 46], and nanowires [47] have been widely studied. However, there are few reports on the use of ultra-thin 2D CdS as a photocatalyst in hydrogen production [48], due to the difficult preparation process of 2D CdS nanosheets. In particular, it is still challenging to achieve large-scale and practical photocatalytic hydrogen generation over 2D nanostructured CdS using proper co-catalysts.

At present, the development and study of earth-abundant co-catalysts are imperative due to the low natural abundance and high cost of precious metals (Ag, Pt, Au) [3, 49]. Thus far, various low-cost co-catalysts such as MoS_x [10, 15, 33, 50-54], CoMoS_x [55], WS₂ [56, 57], NiS [58-60], CuS [61, 62], Ni₃C [63], MXene [52], Co_xP [64, 65], and MoP [65] have been immensely employed to enhance the hydrogen-production activity of CdS. Interestingly, owing to the simple preparation method, the earth-abundant Cu₇S₄ nanocrystals have aroused extensive interest in the fields of electrocatalysis and photocatalysis. Although Cu₇S₄ has also been exploited in different photocatalytic applications, it is principally concentrated on degradation organics [66-69] and photothermal applications [70]. Notably, previous researches on Cu₇S₄ nanocrystals involved using them only as templates or semiconductors, rather than as co-catalysts. Evidently, the studies on Cu₇S₄-based hydrogen-evolution co-catalysts are quite limited [71, 72].

In this work, 2D Cu₇S₄ NSs were initially used as co-catalysts in photocatalytic hydrogen production. The limitations of both Cu₇S₄ and CdS can be overcome by constructing a 2D–2D-layered heterojunction. In this study, ultra-thin 2D Cu₇S₄ and CdS nanosheets were synthesized by direct precipitation and a simple one-step hydrothermal method, respectively (as shown in Scheme 1). Consequently, we prepared 2D noble-metal-free Cu₇S₄ modified CdS nanosheet photocatalysts and intensively explored their pivotal role in enhancing the hydrogen evolution reaction. Notably, the optimally obtained CdS/Cu₇S₄ NSs demonstrate enhanced stability, high light harvesting, and enhanced activity toward photocatalysis in Na₂S–Na₂SO₃ solution. The highest hydrogen production rate was 27.8 mmol g^–1 h^–1 at ambient temperature. The excellent photocatalytic activity is mainly attributed to the ability of the loaded 2D Cu₇S₄ co-catalysts to significantly suppress e^––h⁺ recombination, enhance conductivity, accelerate electron transfer, and enhance light absorption.

All the chemicals were of analytical grade and were used directly without further purification, including thiourea (CH₄N₂S), cadmium acetate (C₄H₆CdO₄), ethylenediamine (EDA), copper(II) sulfate pentahydrate (CuSO₄·5H₂O), sodium sulfide nonahydrate (Na₂S·9H₂O), and sodium sulfite (Na₂SO₃).

The CdS NSs were synthesized by a facile one-step hydrothermal method [48]. First, 0.5330 g of C₄H₆CdO₄ and 0.4567 g of CH₄N₂S were added into the beaker containing 60 mL of EDA. EDA acts as a solvent and structure-directing compound in CdS synthesis [73]. Prior to transferring the suspension into a 100 mL Teflon-lined autoclave, the solution was stirred for 40 min, followed by heating at 100 ℃ for 8 h. The yellowish solid precipitate was collected through centrifugation at 9000 rpm for 10 min, after which it was washed with water and ethanol five times, severally. Finally, the product was dried in a vacuum oven.

CuSO₄·5H₂O (0.5 g) was added to a beaker containing 50 mL of distilled water, while 0.48 g of Na₂S·9H₂O and 0.25 g of Na₂SO₃ were added to another beaker containing 50 mL of distilled water, followed by stirring for 30 min. Thereafter, Na₂S·9H₂O and Na₂SO₃ solutions were added dropwise to the CuSO₄·5H₂O solution under continuous agitation (pH ~8.5), and a black precipitate was formed. The solution was stirred continuously for 12 h. The black precipitates were collected by centrifugation at 9000 rpm for 6 min and washed with water and ethanol three times, severally. Finally, the product was dried in a vacuum oven for 10 h.

Binary CdS/Cu₇S₄ was prepared by a simple electrostatic self-assembly method. Firstly, 8 mg of Cu₇S₄ was dispersed in 30 mL of distilled water (pH ~7.0) and dispersed by ultrasonication for 90 min at room temperature. Subsequently, 392 mg of CdS NSs was mixed with the suspension and subjected to ultrasonication for 2 h. Afterward, the solution was stirred for 10 h. Finally, CdS/2%Cu₇S₄ was obtained by centrifugation and washed two times with water and ethanol, severally. The other binary composites with varying proportions of Cu₇S₄ were prepared by this procedure.

Within the scope of 10°–80°, the structure diffraction peaks of CdS and CdS/Cu₇S₄ were recorded using an X-ray diffractometer (XRD, XD-2X/M4600 with Cu kα radiation). The UV-vis absorption spectrum was determined using a Daojin (Japan) UV-2600 PC spectrophotometer at 200–800 nm. TEM (transmission electron microscopy) and HRTEM (high-resolution TEM) images of the samples were obtained by FEI Talos F200s 200 kV STEM. XPS spectra were obtained using a K-Alpha⁺ elemental analysis system. The Daojin RF-5301PC (Japan) instrument, with an excitation wavelength of 398 nm, was used to record the photoluminescence (PL) spectra at room temperature.

At standard atmospheric pressure and room temperature, the photocatalytic experiment was performed in a three-neck flask (flat bottom) with a volume of 100 mL. To simulate solar radiation, a Xe lamp (300 W) served as the source of visible light. In a typical experiment, the as-prepared photocatalyst (5 mg) was placed in the prepared 80 mL sacrificial agent solution (0.25 mol/L Na₂S·9H₂O-Na₂SO₃). Subsequently, the suspension was uniformly dispersed by ultrasonication for 30 min and purified with N₂ for 30 min to ensure removal of dissolved oxygen. After initiation of irradiation, 400 μL of gas from the three flasks was injected into the gas chromatograph (GC-9560) at hourly intervals for analysis.

For the preparation of the working electrode, 5 mg of the photocatalyst and 20 μL of Nafion (0.25%) solution were added to 2 mL of ethanol solution, followed by ultrasonication for 2 h. A portion (500 μL) of the suspension was transferred to a fluorine-doped tin oxide (FTO, 2 cm ×3.5 cm) glass substrate in 10 drops, and each dispersion was dried under the infrared lamp. The last step involved maintaining the FTO at 150 ℃ for 1 h under Ar protection.

Transient photocurrent experiments were performed using electrochemical workstations and standard three-electrode batteries. Ag/AgCl, the prepared working electrode, Pt electrode, and 0.1 mol/L Na₂SO₄ were used as the reference electrode, working electrode, counter electrode, and electrolyte, respectively. The irradiation light source was a 300 W Xe lamp.

The electrochemical impedance spectroscopy (EIS) of the above-mentioned working electrodes in a three-electrode system was recorded via a computer-controlled electrochemical workstation. The frequency was in the range of 0.01–105 Hz; the electrolyte solution contained 0.02 mol/L Na₂SO₄ and 0.1 mol/L Na₂S, and the AC amplitude was 5 mV.

The electrocatalytic hydrogen evolution was measured at 0.5 mol/L H₂SO₄ with the standard three electrodes. The reversible hydrogen electrode (RHE), a Pt flake, and a glassy carbon electrode were the reference electrode, counter electrode, and working electrode, respectively.

To determine the crystal structures of Cu₇S₄ and CdS/Cu₇S₄, the typical XRD patterns were recorded (Fig. 1). As shown in Fig. 1a, the diffraction peaks of Cu₇S₄ were mainly concentrated in the 29.0, 32.2, 35.3, 37.8, 46.2, 54.9, 59.6, 67.3, and 68.8°, which were assigned to the (113), (220), (031), (302), (224), (026), (051), (440), and (108) crystal planes in the card of Cu₇S₄ (PDF# 33-0489), respectively, indicating the high purity of the as-prepared Cu₇S₄. In addition, the peaks can be observed in Fig. 1b at 24.8°, 26.5°, 28.2°, 36.6°, 43.7°, 47.8°, and 51.8°, which can be assigned to the (100), (002), (101), (102), (110), (103), and (112) diffraction planes of 2D CdS (JCPDS 41-1049), respectively. The results fully confirmed the synthesis of well-crystallized 2D CdS NSs. However, the peaks of Cu₇S₄ were not observed in the XRD spectra of the CdS/Cu₇S₄ composites, which is possibly due to the high dispersion and low content of Cu₇S₄. The existence of Cu₇S₄ in CdS/Cu₇S₄ needs to be further verified by TEM and other tests.

As shown in Fig. 2, to determine the morphologies and microstructures of the as-prepared Cu₇S₄, CdS, and CdS/Cu₇S₄, TEM and HRTEM were performed. Evidently, from Fig. 2a, the microstructure of CdS is almost transparent, which proves that CdS has a 2D ultrathin nano-sheet structure. As can be observed in Fig. 2b, Cu₇S₄ also exhibits the 2D thin-layer nanostructure. Subsequently, we observed and analyzed the binary composites by TEM, and the images are displayed in Figs. 2c and 2d. From Fig. 2e, it can be observed that the interplanar spacings of 0.33 nm and 0.25 nm correspond to the (002) plane in CdS (JCPDS 41-1409) and the (224) plane in Cu₇S₄ (JCPDS 33-0489), respectively. By comparing Figs. 2a, 2b, 2c, 2d, and 2e, it is evident that the Cu₇S₄ NSs were evenly dispersed on the CdS NSs. Compared with those of the 0D–1D, 1D–2D, 0D–2D, and 1D–1D interfaces, the 2D–2D interface has a larger contact area, provides more electron transfer and capture channels, inhibits e^––h⁺ recombination, and improves the efficiency of the photocatalysts [74]. Therefore, the construction of 2D CdS/Cu₇S₄ coupling and its large specific surface area are considered to be the most prominent advantages of accelerating the e^––h⁺ pairs separation and increasing hydrogen production active sites in this study. In addition, as shown in Fig. 3, elemental mapping was conducted on CdS/Cu₇S₄ to determine the distribution of elements in the sample. The EDX peaks of Cd, S, and Cu in CdS/Cu₇S₄ can be clearly seen in Fig. 3, which effectively confirms the existence of Cu₇S₄ in CdS. The existence and uniform distribution of elements S, Cu, and Cd completely confirm the 2D–2D binary coupling morphology. From the above analysis and discussion, it can be inferred that CdS/Cu₇S₄, a 2D–2D layered heterojunction, has been successfully prepared.

XPS analysis of CdS-2%Cu₇S₄ was performed to further confirm the states and chemical compositions of S, Cd, and Cu in the sample. The measured XPS spectrum (Fig. 4) of CdS-2%Cu₇S₄ shows the position of the main peaks of each element, confirming the existence of chemical elements Cd, Cu, and S. As can be observed in Fig. 4b, two peaks, 404.2 and 411.7 eV, appear in the high-resolution spectrum of Cd 3d with a peak spacing of 7.5 eV, which is attributed to the fact that Cd 3d_3/2 and Cd 3d_5/2 were the main sources of Cd²⁺ in the CdS nanosheets. Moreover, the characteristic peaks of Cu 2p at the 933.4 and 953.2 eV positions, with a peak spacing of 19.8 eV, were clearly observed in Fig. 4c, and were assigned to the Cu 2p_3/2 and Cu 2p_1/2 peaks of Cu⁺, respectively. Meanwhile, the peaks of S 2p were determined at 161.7 and 163.0 eV, with a splitting energy of 1.3 eV, which correspond to the sulfides in Cu₇S₄ and CdS, respectively. The above results strongly illustrate the coexistence of CdS and Cu₇S₄.

In general, photocatalytic reactions depend mainly on the position of the band edge and the ability of light absorption. To gain information on the absorbance, the samples were analyzed using a UV spectrophotometer (Fig. 5a). As shown in Fig. 5a, the visible light absorption edge of the prepared photocatalyst was around 500 nm, with a strong visible light absorption ability. After loading Cu₇S₄ on CdS, it can be clearly observed that the absorption capacity of the samples toward visible light was notably enhanced. Noticeably, after the loading of Cu₇S₄ on CdS, the absorption range of visible light was broadened, and the absorption capacity was enhanced notably.

The hydrogen production activity of the prepared CdS/Cu₇S₄ was verified in the Na₂S-Na₂SO₃ solution under visible light (> 420 nm) to explore the photocatalytic performance. Fig. 5b shows the relationship between the hydrogen production activities of different catalysts and time. Hydrogen generation cannot occur without a photocatalyst and irradiation, which indicates that these conditions are crucial. Intriguingly, the as-synthesized 2D CdS NSs exhibited better hydrogen production activity than those of the previously reported CdS nanorods and nanoparticles. Compared with that of the pristine CdS, the hydrogen emission of CdS-2%Cu₇S₄ is higher, which is mainly due to the synergistic effect between CdS and Cu₇S₄. The hydrogen production rate of CdS-2% Cu₇S₄ is about 27.8 mmol g^–1 h^–1, which is almost 10.69 times and 2.65 times higher than those of pure CdS (2.6 mmol g^–1 h^–1) and CdS-2%CuS (10.5 mmol g^–1 h^–1), respectively. The low hydrogen-producing activity of pure CdS was mainly due to the fact that electrons cannot be transferred rapidly; this, thus, leads to their rapid recombination with holes. The loading of Cu₇S₄ provides more active sites and accelerates the electron transfer, thus facilitating the achievement of an outstanding hydrogen-producing rate. Noticeably, 2D Cu₇S₄ has more active sites due to its ultrathin nanosheet structure, which accelerates electron transfer.

In addition, the shelf life and efficiency of the photocatalyst are important factors considered in determining the stability of the photocatalyst [75, 76]. Thus, the stability experiment of the photocatalyst against long-time irradiation (12 h) was carried out. As shown in Fig. 5c, after four cycles, the amount of generated hydrogen by CdS-2%Cu₇S₄ decreased slightly; thus, it exhibited high stability in Na₂S-Na₂SO₃ solution. Each cycle was conducted for 3 h, and N₂ was injected before the initiation of the next cycle to remove the hydrogen.

As shown in Fig. 5d, the apparent quantum efficiency (AQE) of CdS-2%Cu₇S₄ was further evaluated at incident light wavelengths of 420, 450, 500, and 550 nm. The AQE values of CdS-2%Cu₇S₄ at 420, 450, 500, and 550 nm were 14.7%, 12.3%, 9.6%, and 7.2%, respectively. The AQE values of the sample gradually decreased with the increase in the light wavelength, which implies that the light absorption performance influences the evolution of hydrogen. The above results show that the hydrogen evolution kinetics is related to the light absorption performance.

Furthermore, considering the isoelectric pH values of CdS and Cu₇S₄, as well as the reaction solution pH that might have favored the electrostatic attraction between them, we investigated the effect of the electrostatic attraction solution pH on the hydrogen production. As shown in Fig. 5e, in neutral and strong alkaline solutions, strong electrostatic adsorption ability is observed, which is beneficial for enhancing the hydrogen production efficiency.

To gain an in-depth understanding on the improvement of charge separation and transfer owing to the strong interface interaction between 2D CdS and the 2D Cu₇S₄ co-catalyst, the PL spectrum at the excitation wavelength of 398 nm was studied. Generally, the low PL peak intensity is related to efficient charge capture and separation [77-79]. As depicted in Fig. 6a, the PL spectra of CdS and CdS/Cu₇S₄ exhibit a similar emission trend, which is mainly due to the recombination of the e^––h⁺ pairs in CdS. Further observation shows that the 2D–2D CdS/Cu₇S₄ layered heterojunction exhibits better light intensity quenching effect than that of pure CdS, which means that the formation of a 2D nano-heterojunction can effectively reduce the charge-carrier recombination rate. In fact, 2D Cu₇S₄ and 2D CdS were closely combined due to their large contact areas, which can significantly promote the charge migration and separation. All these results show that the 2D Cu₇S₄ co-catalyst supported on the surface of 2D CdS can effectively promote the separation and transfer of photo-excited charge, which is the key parameter for the measurement of photocatalytic performance [80, 81].

To further study the charge transfer and separation efficiency at the interface, I-t curves (the transient photocurrent response) of all the samples fixed on the FTO substrate were obtained in a 0.1 mol/L Na₂SO₄ solution. As indicated in Fig. 6b, noticeably, the photocurrent density of CdS NSs is lower than that of CdS-2%Cu₇S₄, indicating that loading 2D Cu₇S₄ on 2D CdS significantly enhanced the interface charge transfer. The enhanced photocurrent densities of binary two-dimensional composite materials imply the improvement of charge separation and transfer efficiency, which are critical factors for advancing the photocatalytic activity [82, 83].

The electrochemical impedance spectra (EIS) is a convenient and valid method for exploring the e^––h⁺ separation and interfacial charge transfer resistance [75, 84-86]. As shown in Fig. 6c, the EIS of CdS NSs and CdS-2%Cu₇S₄ were recorded in 0.1 mol/L Na₂S and 0.02 mol/L Na₂SO₃ aqueous solutions. Compared with that of pure CdS, CdS-2%Cu₇S₄ shows a small arc, indicating the existence of a high electronic transmission rate between Cu₇S₄ NSs and CdS NSs. The appropriate proportion of the Cu₇S₄ load on CdS can effectively reduce the electron transfer obstruction, accelerate the electron transfer, and improve the efficiency of hydrogen production. To further highlight the crucial effect of the 2D Cu₇S₄ co-catalyst in promoting the photocatalytic evolution, the polarization curves of CdS NSs and CdS-2%Cu₇S₄ from –1.2 to 0.1 V vs NHE were obtained. Compared with that of CdS-2%Cu₇S₄, the pure CdS electrode in 0.5 mol/L H₂SO₄ exhibits a less ideal hydrogen evolution performance. Based on the photocatalytic mechanism, Cu₇S₄ NSs serve as the active sites of CdS NSs, which is a crucial factor in fortifying the hydrogen-evolution kinetics and decreasing the hydrogen-evolution overpotential, which can directly promote the photocatalytic hydrogen evolution [87, 88].

Based on the aforementioned analysis, the tentative mechanisms of charge transfer and visible-light hydrogen-evolution over the CdS/Cu₇S₄ NSs composites were proposed, as given in Scheme 2. Under exposure to irradiation, the electrons of CdS NSs were readily excited from their VB to CB, followed by combination with the H⁺ ions to produce hydrogen. However, the hydrogen production was hindered by the easy recombination of the e^––h⁺ pairs in 2D CdS NSs. On the contrary, due to the high electric conductivity of the CdS/Cu₇S₄ composite photocatalyst, photogenerated e^- could readily be transferred to 2D noble-metal-free Cu₇S₄ co-catalyst through the close interfacial contact between Cu₇S₄ NSs and CdS NSs, thus effectively separating the e^––h⁺ pairs. Consequently, loading Cu₇S₄ NSs on the surface of CdS can significantly improve the photocatalytic hydrogen-evolution.

2D–2D CdS/Cu₇S₄ binary-layered heterojunction photocatalysts were synthesized by the self-assembly method. The results demonstrated that the as-prepared CdS-2%Cu₇S₄ composite exhibited the highest photocatalytic hydrogen-evolution rates of 27.8 mmol g^–1 h^–1 in 0.25 mol/L Na₂S-Na₂SO₃, which were 10.69 times higher than that of pure CdS NSs. The corresponding AQEs reached 14.8% at 420 nm even with the noble-metal-free co-catalyst. The enhanced activity can be attributed to the loading of Cu₇S₄ NSs and coupling of the 2D–2D interfaces, which effectively promoted the separation of charge carriers and improved the surface hydrogen-evolution kinetics over the 2D hybrid CdS/Cu₇S₄ layered heterojunctions. Hopefully, the 2D–2D interface coupling strategy based on CdS can become a general strategy for enhancing the hydrogen-evolution activity [89].