The application of photocatalytic technology to produce hydrogen is an important means of addressing energy and environmental issues, with significant implications for achieving carbon neutrality and a benign carbon cycle. As a semiconductor, tubular carbon nitride (TCN) is widely used in photocatalysis due to its excellent thermodynamic stability, suitable band gap (∼2.7 eV), high surface area, and low cost (Cao et al., 2015; Cheng et al., 2021). However, as with traditional semiconductors, single-phase g-C₃N₄ exhibits low mobility of photogenerated electrons and holes under light radiation, leading to high recombination rates of charge carriers and low quantum efficiency (Thomas et al., 2008; Kumar et al., 2013). Additionally, TCN requires high light excitation energy, and it can only absorb and utilize ultraviolet and a narrow range of visible light less than 460 nm, resulting in unsatisfactory photocatalytic performance (Chen et al., 2014).

To address these issues, researchers have employed methods such as atomic-level doping (Zhang et al., 2011a; Pan et al., 2011), surface chemistry via molecular-level modification (Zhang et al., 2010; Chu et al., 2013), and construction of heterojunctions (Huang et al., 2022; Lei et al., 2022; Wang et al., 2022) to enhance the photocatalytic activity of TCN. Furthermore, the 2D layered structure and polymetric nature of TCN make it a very suitable host for constructing heterojunctions with various inorganic materials (Cheng et al., 2021). Therefore, choosing an appropriate semiconductor to form a desired composite with TCN has become an important strategy for developing inexpensive and efficient photocatalysts. Compared to metal oxides, metal sulfides usually have smaller band gaps, giving them superior visible light activity (Zhu et al., 2022). Sulfides such as CdS, ZnS, CuS, and Cd_xZn_1-xS have been widely used as photocatalysts for hydrogen production due to their tunable band gaps and excellent visible light utilization (Zhang et al., 2011b; Li et al., 2013; Yuan et al., 2018). As a member of the sulfide family, Bi₂S₃ has a band gap of ∼1.3–1.7 eV (Chen et al., 2017; Liu et al., 2019) and also exhibits excellent visible light activity. When combined with other wide-bandgap semiconductors, Bi₂S₃ as a photosensitizer can effectively improve the photocatalytic performance of the composite (Chen et al., 2017). Moreover, non-toxic Bi₂S₃ has a layered structure, allowing it to easily form heterojunctions with 2D-structured TCN through surface chemical modification, which have been applied in the degradation of dyes and pollutants (Zhou et al., 2015; Chen et al., 2017; Yin et al., 2018; Hu et al., 2019; Liu et al., 2019; Gu et al., 2021; Okab and Alwared, 2023) and the reduction of CO₂ (Guo et al., 2020), as well as efficiently killing multidrug-resistant bacteria (Li et al., 2019).

Although Bi₂S₃-TCN has some applications in the above fields, there are significant differences in the photocatalytic mechanisms reported, and its application in the field of photocatalytic hydrogen production is also rather limited (Li et al., 2023). To reveal the nature of interface interactions and the relevant photocatalytic mechanisms, this study employed a green and environmentally benign route to synthesize a Z-scheme Bi₂S₃-TCN heterojunction photocatalysts. The results indicate that the formation of a heterojunction interface promotes an obvious charge transfer between the two components, enhancing the separation efficiency of charge carriers. The interface interaction also helps further stabilize the crystal structure of Bi₂S₃, thus significantly expanding the visible light absorption capability of the heterojunction while inhibiting the photo-corrosion of sulfides. In addition, the Z-scheme band alignment ensures the reduction and oxidation capabilities of photogenerated electrons and holes respectively, synergistically leading to outstanding and stable photocatalytic hydrogen production.

To determine the crystalline structure, XRD patterns of different samples are presented in Figure 1A. The results indicate that TCN exhibits two diffraction peaks at approximately 13.0° and 28.0°, corresponding to the (100) and (002) crystal planes of graphitic carbon nitride (Li et al., 2022). The former primarily originates from the ordered repetitive arrangement of the in-plane tris-S-triazine structure of TCN (Figure 1B), while the latter is associated with the characteristic stacking of conjugated aromatic groups (Li et al., 2021; Sarkar et al., 2022). XRD peaks are sharp and clear for Bi₂S₃, indicating a good crystalline property. The diffraction peaks at 22.5°, 23.9°, 25.2°, 27.7°, 28.7°, 31.9°, and 46.7° correspond to (220), (101), (130), (021), (230), (221), and (341), and the crystal structure belongs to the orthorhombic system with a space group of P_nma, which is consistent with the data in the standard card (PDF#75-1360, Figure 1A) (Lee et al., 2018; Sun et al., 2020). In addition, the diffraction peaks of both Bi₂S₃ and TCN are simultaneously observed in the Bi₂S₃-TCN sample, while the (002) diffraction peak of TCN and the (021) peak of Bi₂S₃ obviously overlap, resulting in a significant change of the intensities nearby. This demonstrates the coexistence of the two components in the composite and the successful construction of the Bi₂S₃-TCN heterojunction. The absence of XRD patterns for other impurities indicates the high purity of Bi₂S₃-TCN.

To reveal the impact of the introduction of Bi₂S₃ on the structure of TCN, Fourier-transform infrared (FT-IR) spectra of different samples are presented in Figure 1C. Corresponding to the building blocks in Figure 1B (Xiao et al., 2021), TCN exhibits a series of absorption bands (Yin et al., 2018): the absorption peak near 810 cm^-1 corresponds to the stretching vibration of the triazine ring; the fingerprint signals within the range of 1,100–1,700 cm^-1 correspond to the stretching vibrations of the C-N aromatic hetero-rings; the broad absorption band in the range of 3,000–3,600 cm^-1 is mainly attributed to the stretching modes of primary and secondary amines, as well as the O–H stretching vibration of surface-adsorbed water and intermolecular hydrogen bonding interactions. Thanks to the layered structure of Bi₂S₃ (Figure 1D), its introduction may lead to the formation of heterojunctions with the TCN through interfacial interactions. This causes a slight shift of the above IR characteristic peaks toward lower wave numbers. However, there is no significant difference in the intensity of absorption peaks in the sample compared to the pristine TCN samples. The layered structure of TCN in the heterojunctions is therefore not destroyed by the introduction of Bi₂S₃, indicating good structural integrity for Bi₂S₃-TCN.

Figure 2A shows the high-resolution electron microscopy (HR-TEM) images and elemental distribution maps of the Bi₂S₃-TCN heterojunction. The results clearly show that the Bi₂S₃-TCN has a layered structure with special morphology. After assembly of the Bi³⁺ supramolecular complex on the TCN precursor surface, an in situ sulfurization process occurs and results in the formation of a layer of nano-array structures on the heterojunction surface. This effectively enhances the specific surface area and the number of active sites of the composite while simultaneously reducing the diffusion distance of photogenerated charge carriers. This is highly favorable for improving photocatalytic activity. In addition, the test results in Figures 2B,C further clarify that the crystal planes with the interplanar distances of 0.31 nm, 0.19 nm, and 0.35 nm belong to the Bi₂S₃ (230), (002), and (310) crystalline planes, respectively (Kumar et al., 2021). The EDX mapping further demonstrates a uniform distribution of the four elements (N, C, S, and Bi) throughout the whole selected region (Figures 2D–H). These features suggest that a well-contacted interface is formed between Bi₂S₃ and TCN in the heterojunction. Therefore, it can be anticipated that this heterojunction with a unique surface morphology and 2D structure will exhibit excellent photocatalytic performance.

The photocatalytic hydrogen production experiment was applied to evaluate the photocatalytic activity of the catalyst (Li et al., 2023). The hydrogen evolution rate of Bi₂S₃-TCN is 65.2 μmol h^-1, which is 2.43 times that of TCN (26.8 μmol h^-1) and 11.64 times that of Bi₂S₃ (5.6 μmol h^-1). The mean values of the multiple experiments were Bi₂S₃-TCN (65.4 μmol h^-1), TCN (26.6 μmol h^-1), and Bi₂S₃ (5.5 μmol h^-1), respectively (Figure 3A). As shown in Supplementary Table S1, the hydrogen evolution performance of Bi₂S₃-TCN was compared with existing heterojunction photocatalytic materials, demonstrating that Bi₂S₃-TCN had better photocatalytic hydrogen evolution activity. The significant increase in hydrogen production rate demonstrates that the formation of Bi₂S₃-TCN Z-scheme heterostructure provides a new charge transfer pathway with great advantages in photocatalytic reactions. At the same time, Bi₂S₃-TCN maintained a stable H₂ yield during the 4-h test period (Figure 3B). The durability of Bi₂S₃-TCN was tested over five repeated cycles (Figure 3C). After 20 h of continuous testing, Bi₂S₃-TCN still maintained original photocatalytic activity, indicating that its heterojunction photocatalyst has excellent photostability. This indicates that the Bi₂S₃ and TCN heterojunction photocatalyst prepared can promote rapid separation and transfer of charge under light radiation while maintaining the stability of structure and morphology due to its tight heterogeneous structure.

To reveal the origin of the photocatalytic performance, UV-Vis DRS (Figure 4A) was used (Mua et al., 2021). It can be seen clearly that the absorption edge of TCN is close to 460 nm, while that of Bi₂S₃ extends significantly to 700 nm, indicating that Bi₂S₃ has much better visible light response. Compared with pristine TCN, the introduction of Bi₂S₃ leads to a red-shifted adsorption edge for the heterojunction. The interfacial interactions formed between TCN and Bi₂S₃ can modulate the positions of the valence band (VB) and conduction band (CB), obviously reducing the calculated band gap of Bi₂S₃-TCN (Figure 4B). Consequently, Bi₂S₃-TCN shows much stronger absorption of visible light, which is favorable for subsequent photocatalytic reactions. In addition to the above characteristics, the TPC plot in Figure 4C further identifies that the photocurrent of Bi₂S₃-TCN under visible light radiation is much higher than those of TCN and Bi₂S₃, and that even Bi₂S₃ possesses the smallest band gaps (Hao et al., 2020). This indicates that the intimate interface formed between the two components does facilitate the separation of the photocatalytic electrons and holes, leading to a much higher charge separation efficiency. Such a result can be further supported by the EIS data in Figure 4D (Xiao et al., 2022), in which the electric conductivity of Bi₂S₃-TCN with a minimal semi-circle is much better than Bi₂S₃ and TCN. Figures 4E,F show the Mott–Schottky plots for Bi₂S₃ and TCN, respectively. The positive slopes of the plots suggest that both materials are n-type semiconductors. The flat band potentials are found to be −0.74 and −0.43 V (vs RHE) for Bi₂S₃ and TCN, respectively, based on how the band alignment of the heterojunction is determined.

X-ray photoelectron spectroscopy (XPS) was utilized to analyze the chemical makeup and bonding structure of the final product. The shift in peak value demonstrated the transfer of electronic structure between TCN and Bi₂S₃. The XPS analysis identified the presence of Bi and S in Bi₂S₃, and C and N in TCN and Bi₂S₃-TCN, respectively. Supplementary Figures S1, S2 demonstrate a noteworthy decline in the binding energy direction of Bi₂S₃-TCN’s C 1s and N 1s XPS peaks relative to TCN. Bi₂S₃-TCN’s 2p and Bi 4f spectra also show a positive shift in binding energy (Supplementary Figure S3). The XPS spectra results indicate that there is a close contact and charge transfer between Bi₂S₃ and TCN. We utilized the electron paramagnetic resonance (EPR) technique with 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the trapping agent of hydroxyl radical (•OH) (Figure 5A) to further confirm the charge transfer mechanism between Bi₂S₃ and TCN. Under visible light irradiation, four DMPO-•OH characteristic signal peaks with 1:2:2:1 intensity ratio of TCN and Bi₂S₃-TCN were observed, while no DMPO-•OH characteristic signal peaks appeared in Bi₂S₃. This indicates that the holes are concentrated in the VB of TCN and that only the VB level of TCN has sufficient oxidation capacity to reach •OH. Conversely, as shown in Figure 5B, the CB of both TCN and Bi₂S₃ is capable of reducing O₂ to •O₂ ⁻, leading to the detection of the •O₂ ⁻ signal in all samples. Notably, the Bi₂S₃-TCN exhibits the most potent •O₂ ⁻ signal, which can be ascribed to the recombination between TCN’s electrons and Bi₂S₃’s holes, resulting in more electrons accumulating in the CB of Bi₂S₃. It is possible that the electrons on the CB of TCN transfer to the VB of Bi₂S₃, creating a Z-scheme heterojunction photocatalytic system, with higher oxidation and reduction capacities located on TCN and Bi₂S₃, respectively. This suggests that photogenerated electrons and holes in Bi₂S₃-TCN may follow a Z-scheme electron transfer pattern instead of a type II electron transfer pattern (Supplementary Figure S4).

The work function (WF) of TCN, Bi₂S₃, and Bi₂S₃-TCN has been measured by Kelvin probe (Liu et al., 2022). The charge transfer direction in the composite photocatalyst can be well verified. As shown in Figure 5C, the contact potential difference (CPD) between TCN, Bi₂S₃, and Bi₂S₃-TCN and Au probes is 218 mV, 51 mV, and 130 mV, respectively, and the WF of TCN and Bi₂S₃ is calculated to be 5.29 and 5.12 eV, respectively (Figure 5D; Luo et al., 2023). When TCN and Bi₂S₃ are in contact, electrons are transferred from Bi₂S₃ to TCN through the contact interface until their Fermi levels reach equilibrium. This transfer method of electrons induces an embedded electric field between the positively charged Bi₂S₃ and negatively charged TCN at the interface. Consequently, carrier migration between Bi₂S₃ and TCN is accelerated. Within this particular embedded electric field, electrons accumulate at the TCN interface, while the electron density decreases at the Bi₂S₃ interface. This causes the TCN bands to bend downward and those on Bi₂S₃ to bend upward. Under illumination, the internal electric field and band bending facilitate the valence band transfer of conduction band electrons of tetracyanoquinodimethane to the boundary with Bi₂S₃, where they combine with the valence band hole of Bi₂S₃ to establish a Z-scheme electron transfer mechanism. This special charge transfer process accelerates the rate of carrier separation and transfer in the two components so that Bi₂S₃-TCN has a stronger REDOX capacity. As shown in Figure 5E, when both components in the heterojunction are excited under visible light radiation, the photocatalytic electrons in the CBM of TCN will diffuse and recombine with the electrons in the VBM of Bi₂S₃. With the formation of the S-scheme heterojunction, the high energy electrons in the CB of Bi₂S₃ and the high oxidative holes in the VB of TCN are retained (Liu et al., 2021). The participation of these highly active carriers in the following photocatalytic reactions are one important reason for the excellent photocatalytic hydrogen production of Bi₂S₃-TCN.

A green and environmentally benign route was adopted to successfully synthesize a Bi₂S₃-TCN heterojunction. The sample exhibits a 2D structure with a special surface morphology, and the intimate interface formed between Bi₂S₃ and TCN not only increases the photocatalytic charge separation efficiency but also extends the visible light adsorption of the sample, leading to a Z-scheme heterojunction with excellent photocatalytic performance. Bi₂S₃-TCN exhibits a hydrogen production rate of 65.2 μmol h^-1, which is 11.64 and 2.43 times those of Bi₂S₃ and TCN.