One-step hydrothermal synthesis of a ternary heterojunction g-C3N4/Bi2S3/In2S3 photocatalyst and its enhanced photocatalytic performance

In recent years, photoelectrocatalysis has been one of the hotspots of research. Graphite-like carbon nitride (g-C3N4) is one of the few non-metal semiconductors known and has good potential in the field of photocatalysis due to its simple preparation method and visible light effects. In this study, a method for compounding two semiconductor materials, In2S3 and Bi2S3, on the surface of g-C3N4via a one-step hydrothermal method is reported, and it was found that this resulting material showed remarkable properties. The advantages of this method are as follows: (1) the formation of a heterojunction, which accelerates the separation efficiency of photogenerated carriers; (2) a large number of holes and defects on the surface of g-C3N4 are conducive to the nucleation, crystallisation and growth of In2S3 and Bi2S3. Compared with its counterpart catalysts, the CN/In2S3/Bi2S3 composite catalyst has significantly improved performance. Due to its high degree of crystallinity, the adsorption capacity of the catalyst itself is also significantly improved. In addition, the stability of the composite material maintains 90.9% after four cycles of use, and the structure is not damaged. In summary, CN/Bi2S3/In2S3 composite materials are believed to have broad application potential in the treatment of dye wastewater.


Introduction
With the development of science and technology, the application of photoelectrocatalytic technology 1-5 is of great benet to the water pollution caused by the of the global industrialization process. [6][7][8][9] As a result, different countries and regions worldwide have made great efforts to protect the environment on which people depend. [10][11][12] Fortunately, at the same time, there have been continuous developments in science and technology toward alleviating pollution in water. Common methods that have been developed for dealing with organic pollution in water can be divided into three categories: physical, chemical and biological methods. Adsorption is the main physical method used, 13 which mostly involves the use of activated carbon and diatomaceous earth as adsorbents. Other adsorbents have been also reported recently for wastewater treatment, e.g., sepiolite mineral nanobers. 14 However, the costs associated with this method are high and complete adsorption of the pollutants from water cannot be achieved. Oxidation reactions are the main chemical methods used in the remediation of water pollutants. 15,16 Therefore, it is urgent to develop an environmentally friendly method to solve the current water pollution issues. As photocatalytic technology has the advantages of being green, highly efficient and causing no secondary pollution, it is one of the most promising methods for degrading pollutants in water that could be applied as a future strategy. [17][18][19] At present, the reports on photocatalysts in the literature are mainly based on semiconductor materials, including titanium dioxide (TiO 2 ), 20 molybdenum sulde (MoS 2 ), 21 Bi 2 S 3 , 22 zinc oxide (ZnO), 23 graphitic carbon nitride (g-C 3 N 4 ). 24 The unique forbidden band structures of semiconductors lead to the generation of electrons and holes (e À and h + ) under the excitation of light of a certain wavelength. These e À and h + then react with water molecules and oxygen in water to form strong oxidising substances, which oxidise and decompose the pollutants present in water into environmentally friendly molecules. Among these semiconductors, in 2009 g-C 3 N 4 was rst used by Wang et al. in the photocatalytic production of hydrogen. Due to the excellent photoelectric properties of g-C 3 N 4 , it has attracted the attention of many researchers in the eld of photocatalysis. 19,24 In addition to photocatalytic application, 25 graphitic carbon nitride based materials are also promising for biomedical applications. 26 However, some studies have shown that the rapid recombination of e À and h + severely limits the photocatalytic performance of g-C 3 N 4 . 27 To solve and overcome these difficulties, many researchers have made tremendous efforts, 28,29 such as using different strategies to improve the catalytic performance of g-C 3 N 4 . For example, a higher number of active sites have been introduced in g-C 3 N 4 by increasing its specic surface area, 30 however, the coverage of dyes and other small molecules on the catalyst surface and in the pores makes the catalyst less stable. The latest research shows that a strategy of constructing heterojunction binary and multi-component catalysts not only accelerates the separation of e À and h + but also expands the range of light response of these materials has attracted great attention. [31][32][33] It has been reported in the literature that the construction of heterostructured catalysts leads to systems that have greatly improved photocatalytic performance. 34-37 C. Auttaphon et al. 38 synthesised a Z-scheme Bi 2 S 3 /ZnIn 2 S 4 heterostructure for methylene blue degradation, which showed higher photocatalytic activity than the uncombined catalysts alone. Geioushy et al. used a one-pot method to prepare BiPO 4 /Bi 2 S 3 for the reduction of harmful Cr(VI) driven by visible light. 39 ABi 2 S 3 /Bi 2 W 2 O 9 composite has also previously been prepared that shows 80% degradation of phenol aer 120 min of irradiation, a performance that is substantially higher than those of pure Bi 2 S 3 and Bi 2 W 2 O 9 under visible-light illumination. 40 An et al. synthesised a new Z-scheme In 2 S 3 /graphene heterojunction with a core-shell structure, which showed a methyl orange degradation rate of almost ve times higher than that of pure In 2 S 3 (ref. 41) and reduced Cr(VI) under visible light. For degradation of rhodamine B (RhB), these photocatalysts have been reported recently. 42,43 In summary, from the perspective of development, photocatalytic technology has great application prospects in the degradation of pollution and environmental protection. 44,45 In this study, to verify the heterostructure photocatalyst theory, the ternary heterostructured photocatalyst g-C 3 N 4 /Bi 2 S 3 / In 2 S 3 was prepared via a one-step hydrothermal method. The modication of microscale porous spherical Bi 2 S 3 and dendritic In 2 S 3 on the surface of porous g-C 3 N 4 with a large specic surface area signicantly improves its optical, physical and photoelectric conversion properties. Compared with the performances of the g-C 3 N 4 , Bi 2 S 3 and In 2 S 3 catalysts alone, and Bi 2 S 3 /In 2 S 3 binary heterojunction catalysts, the obtained g-C 3 N 4 /Bi 2 S 3 /In 2 S 3 ternary heterostructured photocatalyst shows signicantly improved degradation of rhodamine B (RhB). In addition, by carrying out powder X-ray diffraction (XRD), photoluminescence (PL), X-ray photoelectron spectroscopy (XPS) and other characterisation methods, the composition of the catalyst in terms of its elemental content, its chemical valence state, morphology, and charge carrier transfer behaviour were carefully analysed, and the possible catalytic mechanism of the catalyst was explored and is discussed in detail. This work may provide a feasible practical example of a way of improving the photocatalytic performance of g-C 3 N 4 through its combination with Bi 2 S 3 and In 2 S 3 in a heterogeneous composition as a reference for subsequent research.

Materials
Thiourea, absolute ethanol, RhB, urea, bismuth nitrate pentahydrate, indium nitrate, and other chemicals used in the experiments were of analytical purity (99%) and used without further purication. Deionised water was used throughout all of the experimental work in this study. All medicines come from Titan.

Sample preparation
2.2.1 Preparation of the g-C 3 N 4 photocatalyst. An appropriate amount of urea was added to a crucible and closed the lid, and heated in a muffle furnace at a heating rate of 2.5 C min À1 to 550 C, and held at this temperature for 4 h to obtain a yellow solid of g-C 3 N 4 , which was ground into a powder for subsequent use.
2.2.2 Pretreatment scheme of g-C 3 N 4 . An appropriate amount of g-C 3 N 4 powder was added to a 50 mL beaker, into which 20 mL of DMF (N,N-dimethylformamide) was poured and the reaction mixture was stirred at room temperature for 12 h to prepare a solution referred to as solution A.
2.2.3 Preparation of g-C 3 N 4 /Bi 2 S 3 /In 2 S 3 . 70 mL of HON 3 solution with a concentration of 1 mol L À1 was added to a 100 mL beaker and stirred for ten min before adding 0.1 mmol of Bi (NO 3 ) 5H 2 O, and stirring the resulting reaction mixture for 30 min. Then, to this mixture 0.1 mmol of In (NO 3 ) xH 2 O was added and the mixture was stirred for 30 min. Aer this time 0.2 g of cetrimonium bromide (CTAB) was added and the reaction mixture was ultrasonicated for 30 min, to produce a solution referred to as solution B. The above solution A was slowly added dropwise to solution B to mix them together uniformly, then 0.6 g of thiourea was added and the reaction mixture was stirred for 30 min, before being poured into a 100 mL polytetrauoroethylene liner and heated to 140 C under hydrothermal reaction for 12 h. The resulting liquid was washed three times with ethanol and deionised water, and then dried in an oven overnight at 60 C. g-C 3 N 4 /Bi 2 S 3 /In 2 S 3 catalysts with different mass fractions of g-C 3 N 4 were prepared, and the catalysts g-C 3 N 4 and Bi 2 S 3 /In 2 S 3 were prepared for comparison.

Characterisation of the photocatalysts
XRD patterns were recorded using a Bruker D8 Advance powder X-ray diffractometer equipped with a Cu-Ka radiation source. A Nicolet iS10 Fourier-transform infrared (FTIR) spectrometer was employed to record FTIR spectra of the samples in the wavenumber range of 500-4000 cm À1 . XPS measurements of the prepared samples were recorded using a PHI 5000 Versa Probe spectrometer, in which the binding energy positions were calibrated against the C-C bond peak that has a binding energy of 284.8 eV. The morphology characteristics of the samples were investigated using transmission electron microscopy (TEM, Tecnai G F20, Hitachi, HT7700) and scanning electron microscopy (SEM, Hitachi S-4800). UV-visible diffuse reection spectroscopy (DRS) measurements were carried out using a UVvisible spectrophotometer (Shimadzu UV-3600) across a wavelength range of 200-800 nm, employing BaSO 4 as a benchmark. Fluorescence emission and time-resolved uorescence spectra were measured using a uorescence spectrometer (Hitachi F-7000) over a wavelength range of 340-800 nm at an excitation wavelength of 320 nm, employing a 300 W xenon lamp (CEL-HXF300) as a light source.

Photocatalytic activity
To determine the photocatalytic activity of the synthesised photocatalyst, the catalytic degradation of RhB was evaluated under simulated sunlight using a 500 W xenon lamp as a light source in a photoreactor (CEL-HXF300, Beijing, China). In a typical experiment, before being exposed to light the reactants and catalyst were placed in a quartz tube and magnetically stirred for 30 min in the dark to achieve adsorption-desorption equilibrium. At a given time interval aer beginning the visible light irradiation, samples were taken from the reaction suspension (each with a volume of 5 mL), centrifuged at 8000 rpm for 10 minutes, and then the supernatant was removed. Then, the concentration of RhB in the sample solutions was analysed using a (UV-31/32/3300) UV-visible spectrometer (MAPADA) at its maximum absorption wavelength. To ensure the repeatability of the results, each run was repeated to obtain an average for each set of conditions. A blank test was also carried out by irradiating the reactant solution in the absence of the catalyst to evaluate its photo-induced selfsensitised photodegradation.
where h is the photocatalytic efficiency; C 0 is the concentration of the reactant before illumination and C t is the concentration of the reactant aer illumination for t hour.
3 Results and discussion 3.1 Morphology and structure 3.1.1 XRD analysis. As shown in Fig. 1, XRD was conducted to identify the crystallographic and phase structures of Bi 2 S 3 , CN, In 2 S 3 , Bi 2 S 3 /In 2 S 3 and CN-xBi 2 S 3 /In 2 S 3 . The XRD pattern of pure CN shows two distinct diffraction peaks at 2q ¼ 12.43 and 27.54 , which can be indexed to the (002) and (100) diffraction planes, respectively, of g-C 3 N 4 (JCPDS (87-1526)), attributed to its typical stacked aromatic rings and interlayer p-p stacking. 46 . 65-0459). 47 However, the XRD pattern of In 2 S 3 /Bi 2 S 3 only shows the diffraction peaks of Bi 2 S 3 , showing that it contains little In 2 S 3 content. In addition, one new weak characteristic peaks at 27.54 is present in the patterns of the CN/ Bi 2 S 3 /In 2 S 3 composites, which can be assigned to the (0 0 2) peak of g-C 3 N 4 . With an increase in the amount of g-C 3 N 4 present in the catalyst samples, the diffraction peak of Bi 2 S 3 / In 2 S 3 at 2q ¼ 24.92 went from being of high intensity to low intensity, reecting the inuence that it has on the phase crystallinity of Bi 2 S 3 /In 2 S 3 .
3.1.2 SEM/TEM analysis. In Fig. 2a the layered porous structure of g-C 3 N 4 can be seen in CN-xBi 2 S 3 /In 2 S 3 , which is benecial for the adsorption of dyes by the catalyst. Fig. 2b shows an image of the structure of In 2 S 3 , which is made up of a large number of hexagons with side lengths ranging from 0.3-1 mm. Fig. 2c shows an image of Bi 2 S 3 , which has a rod-shaped structure, where the rods have diameters in the range of 50-100 nm and a length in the range of 0.5-2.0 mm in size.
Compared with pure In 2 S 3 and Bi 2 S 3 , the morphology of the composite of both materials in Fig. 2d shows a spherical structure of In 2 S 3 surrounded by rod-shaped Bi 2 S 3 . This structure is more conducive to the transfer of photogenerated e À and h + than the single catalysts alone. Fig. 2e shows an SEM image of 0.05CN/Bi 2 S 3 /In 2 S 3 , which shows the combination of owerlike In 2 S 3 and a spherical Bi 2 S 3 structure. Porous layered g-C 3 N 4 has a large specic surface area and pores, so it provides more defects with which to facilitate the crystallisation, nucleation and growth of In 2 S 3 and Bi 2 S 3 . In addition, the structures of In 2 S 3 and Bi 2 S 3 are not damaged. To analyse its structure in more depth, TEM and high-resolution TEM (HRTEM) were used to characterise 0.05CN/Bi 2 S 3 /In 2 S 3 . Fig. 2f shows a TEM image of g-C 3 N 4 , from which it can be seen that the layered g-C 3 N 4 of the lm structure is consistent with the results presented in the literature. Fig. 2g shows a TEM image of Bi 2 S 3 /In 2 S 3 , which is consistent with the image shown in Fig. 2d. Fig. 2h shows a HRTEM image of 0.05CN, in which it can be clearly seen that there are two lattice fringes with different widths of 0.352 and 0.278 nm, corresponding to the (1 3 0) crystal plane of Bi 2 S 3 and the (4 0 0) crystal plane of In 2 S 3 . This result is consistent with the PXRD results.
3.1.3 FTIR analysis. To further demonstrate the surface properties and functional groups present in the samples, FTIR spectra of In 2 S 3 , Bi 2 S 3 , g-C 3 N 4 and CN-xBi 2 S 3 /In 2 S 3 were recorded and the results are presented in Fig. 3. In the FTIR spectrum of CN-xBi 2 S 3 /In 2 S 3 there is no obvious peak that is characteristic of Bi 2 S 3 . 48 The peaks at 1327, 1417, 1559, and 1631 cm À1 in the same spectrum can be assigned as the characteristic absorption bands of g-C 3 N 4 , attributed to the vibration of the C 6 N 7 skeleton and the typical stretching mode of the CN heterocycles. 49 The peak at around 809 cm À1 can be assigned to the breathing mode of the triazine units. In addition, the peaks at 772 and 1196 cm À1 are typical of In 2 S 3 and the characteristic peak of Bi 2 S 3 /In 2 S 3 is roughly consistent with In 2 S 3 . 50,51 The characteristic absorption peak of g-C 3 N 4 /In 2 S 3 /Bi 2 S 3 is consistent with that of g-C 3 N 4 , while the peak related to Bi 2 S 3 /In 2 S 3 gradually decreases in intensity, which may be a result of the g-C 3 N 4 peak being too intense. Fig. 1 The typical XRD patterns of CN, Bi 2 S 3 /In 2 S 3 and CN/Bi 2 S 3 /In 2 S 3 .

XPS analysis.
To study the surface chemical states and chemical compositions of the CN/Bi 2 S 3 /In 2 S 3 heterojunction samples, Fig. 4 shows XPS measurements of the samples, in which it can be seen that the survey spectrum of CN-In 2 S 3 /Bi 2 S 3 exhibits the typical C 1s, S 2p, S 2s, N 1s, In 3d, and Bi 4f peaks expected for CN-Bi 2 S 3 /In 2 S 3 , Bi 2 S 3 /In 2 S 3 and CN. The survey spectrum indicates that C and N exist on the surface of g C 3 N 4 , and Bi, In, and S on the surface of In 2 S 3 /Bi 2 S 3 . Besides this, as shown in Fig. 4a, Bi, In and S were found to be present on the surface of the g-C 3 N 4 nanosheets, since the peaks at around 443.61, 157.43, and 225.01 eV can be assigned to the In 3d, Bi 4f and S 2s energy levels, respectively. The highresolution C 1s, N 1s, In 3d, Bi 4f and S 2s XPS spectra of 0.05CN/Bi 2 S 3 /In 2 S 3 are presented in Fig. 4b-f. Moreover, the binding energy peak of the N-C]N groups in the 0.05CN-Bi 2 S 3 / In 2 S 3 heterostructure shis lowers by 0.02 eV to 288.14 eV, which indicates that there are weak interactions between the Bi 2 S 3 /In 2 S 3 and g-C 3 N 4 nanosheets as a result of surface modi-cation effects. 52 The N 1s Gaussian curve in Fig. 4b can be deconvolved into three peaks that have binding energies of 398.47, 400.20, and 401.90 eV, corresponding to sp 2 -hybridised nitrogen (C-N]C), the N-(C)3 groups of the skeleton, and the surface uncondensed bridging N atoms with C-N-H functional groups attached. 53,54 Compared to g-C 3 N 4 , the peak positions in the N 1s spectrum of 0.05CN/Bi 2 S 3 /In 2 S 3 are shied to higher binding energies. In the In 3d spectrum (Fig. 4d), the In 3d 3/2 Fig. 2 SEM images of (a) g-C 3 N 4 ; (b)In 2 S 3 ; (c) Bi 2 S 3 ; (d) Bi 2 S 3 /In 2 S 3 ; (e) 0.05CN/Bi 2 S 3 /In 2 S 3 ; TEM images of (f) g-C 3 N 4 ; (g) Bi 2 S 3 /In 2 S 3 and HRTEM images of (h) 0.05CN/Bi 2 S 3 /In 2 S 3 . Fig. 3 The FTIR spectra of CN, Bi 2 S 3 /In 2 S 3 and CN-xBi 2 S 3 /In 2 S 3 . and In 3d 5/2 peaks of 0.05CN/Bi 2 S 3 /In 2 S 3 are located at 452.23 and 444.58 eV, which indicates a slight shi to higher binding energies compared to pristine In 2 S 3 (451.39 and 443.87 eV). As shown in Fig. 4e, the Bi 4f XPS spectrum of Bi 2 S 3 displays clearly two peaks (162.73 and 157.43 eV), which can be attributed to the 4f 5/2 and 4f 7/2 of Bi 3+ . Moreover, compared to Bi 2 S 3 , the peak positions (163.50 and 158.18 eV) of Bi 4f in 0.05CN/Bi 2 S 3 /In 2 S 3 are shied to higher binding energies. Since the peak of Bi 4f covers the characteristic peak of S 2p, S 2s is added to illustrate the existence of S 2À . As shown in Fig. 4f, the S 2s XPS spectrum of Bi 2 S 3 /In 2 S 3 shows three peaks at 221.0, 224.0, and 224.9 eV. Compared with Bi 2 S 3 /In 2 S 3 , 0.05CN/Bi 2 S 3 /In 2 S 3 shows peaks that are slightly shied to higher binding energies (222.6, 225.01, and 227.17 eV). These results suggest the successful heterogeneous growth of In 2 S 3 and Bi 2 S 3 on the g-C 3 N 4 nanosheets.
3.1.5 UV-vis analysis. To investigate the modication effect that the g-C 3 N 4 nanosheets have on the Bi 2 S 3 /In 2 S 3 heterostructure, the optical absorption properties of the as-prepared samples were analysed using UV-vis spectroscopy, the results of which are shown in Fig. 5a. The g-C 3 N 4 shows good light absorption at a wavelength of less than 450 nm. The light absorption of In 2 S 3 and Bi 2 S 3 seem much higher than that of g-C 3 N 4 , at a wavelength of 1000 nm. Compared with Bi 2 S 3 /In 2 S 3 , the absorption edge of the g-C 3 N 4 /Bi 2 S 3 /In 2 S 3 heterostructure exhibits a distinct red shi, which may originate from the modication effect of Bi 2 S 3 /In 2 S 3 on the surface of the g-C 3 N 4 nanosheets. This signies that a strong interaction is formed between g-C 3 N 4 and Bi 2 S 3 /In 2 S 3 , which is benecial to improving the separation efficiency of charge carriers and the stability of the heterostructure. In addition, the band gap energies (E g ) of the samples were calculated using the Tauc equation: where a, hn, A and E g are the absorption coefficient, photoenergy, proportionality constant and band gap, respectively. As shown in Fig. 5b, the band gaps of g-C 3 N 4 , In 2 S 3 and Bi 2 S 3 are 2.84, 2.2, and 1.30 eV, respectively. In addition, the band gap of Bi 2 S 3 /In 2 S 3 is estimated to be 1.26 eV, and the band gaps of 0.01CN/Bi 2 S 3 /In 2 S 3 and 0.05CN/Bi 2 S 3 /In 2 S 3 are 1.20 and 1.18 eV, respectively. Compared with Bi 2 S 3 /In 2 S 3 , the band gap of 0.05CN/Bi 2 S 3 /In 2 S 3 is narrower, which indicates that the heterostructure of 0.05CN/Bi 2 S 3 /In 2 S 3 is more easily excited by visible light and is benecial for improving photocatalytic performance due to the modication of large holes and defects on the surface of the g-C 3 N 4 nanosheets.
3.1.6 PL spectra and photocurrent analysis. To further understand the reasons for the improvement in the photocatalytic activity of the composite photocatalyst, the PL spectra were measured to investigate the separation efficiency of photoinduced charge carriers. 55 The PL spectra of g-C 3 N 4 , Bi 2 S 3 /In 2 S 3 and 0.05 g-C 3 N 4 /Bi 2 S 3 /In 2 S 3 are shown in Fig. 6a. The PL emission stems from the radiative recombination between the excited e À and h + and the weaker the PL peak intensity is, the slower the recombination rate of the excited e À and h + . Comparing the PL emission spectra of the three materials, the 0.05 g-C 3 N 4 /Bi 2 S 3 /In 2 S 3 heterostructure shows low uorescence intensity relative to that shown by g-C 3 N 4 and Bi 2 S 3 /In 2 S 3 . These results indicate that the efficient interfacial charge transfer between Bi 2 S 3 /In 2 S 3 and g-C 3 N 4 hinders the recombination of e À and h + pairs. In addition, Fig. 6b is the photocurrent curve of the catalysts. We can clearly see that the photocurrent value of the composite material is the highest, and the conclusion corresponds to Fig. 6a. Fig. 7a shows the degradation RhB by the different catalysts. It can be clearly seen that the 0.05 g-C 3 N 4 composite catalyst shows the best catalytic activity, which almost completely degrades the dye aer 45 min of exposure to light. Fig. 7b shows the rst-order kinetic tting curve of the degradation, and the results obtained are consistent with the results shown in the degradation curve. Fig. 7c shows a histogram of the catalyst tting coefficient. The tting coefficient of the 0.05 g-C 3 N 4 / Bi 2 S 3 /In 2 S 3 catalyst is 0.02826, which is signicantly higher than the values for g-C 3 N 4 , In 2 S 3 and Bi 2 S 3 . Compared with recently published articles, this study has advantages in degradation efficiency. [56][57][58] 3.3 Research on the mechanism of photocatalysis Fig. 8a-c show the valence band spectra of g-C 3 N 4 , In 2 S 3 and Bi 2 S 3 respectively (1.95, 0.32, and À0.35 eV). From the results obtained and shown in Fig. 5b, the valence (VB) and conduction bands (CB), respectively, of g-C 3 N 4 (1.95 and À0.89 eV), In 2 S 3 (0.32 and À1.88 eV), and Bi 2 S 3 (À0.35 and À1.65 eV) can be calculated, which facilitate the subsequent analysis of the Fig. 5 (a) UV-vis DRS of g-C 3 N 4 , In 2 S 3 , Bi 2 S 3 , Bi 2 S 3 /In 2 S 3 , 0.01CN/ Bi 2 S 3 /In 2 S 3 , and 0.05CN/Bi 2 S 3 /In 2 S 3 and (b) plot of (ahv) n/2 versus hv. Fig. 6 PL spectra (a) and photocurrent (b) of the g-C 3 N 4 , Bi 2 S 3 /In 2 S 3 and 0.05 g-C 3 N 4 /Bi 2 S 3 /In 2 S 3 composites. mechanism of the photocatalyst degradation. Fig. 8e shows the particle capture experiment performed by the catalyst, using isopropanol (IPA), 1,4-benzoquinone (BQ), and Na 2 C 2 O 4 as the reactive species scavengers for cOH, cO 2 À , and h + , respectively. 59

Photocatalytic activity
It can be seen that BQ, IPA and EDTA-2NA all have different degrees of inuence on the catalytic performance of the catalyst. Among them, IPA has the greatest degree of inuence, with a degradation rate of 28.3%, whereas EDTA-2Na has the lowest degree of inuence and a degradation rate of 65.3%. This shows that photogenerated e À have a greater impact on the ternary heterostructured nanocomposite g-C 3 N 4 /In 2 S 3 /Bi 2 S 3 than h + . Fig. 8e shows the possible reaction pathway of the photogenerated carriers of the ternary heterostructured nanocomposite g-C 3 N 4 /In 2 S 3 /Bi 2 S 3 catalyst. When the catalyst is excited by light, e À transition from the VB to the CB. Due to the difference in electronegativity, the excited e À will be transferred from the CBs of In 2 S 3 and Bi 2 S 3 to the CB of g-C 3 N 4 . At the same time, h + will be transferred from the VB of g-C 3 N 4 to the VBs of In 2 S 3 and Bi 2 S 3 , and can also oxidise pollutants. The transfer speed of h + and e À is improved, and the recombination of photogenerated carriers is suppressed, which improves the photocatalytic performance of the material. It is speculated that the photocatalysis may proceed in the following way: CN/Bi 2 S 3 /In 2 S 3 + hn / e À + h + (1) Of these, (2), (4), and (7) are thought to be the main photocatalytic reactions that occur. Fig. 9a shows the stability test results of the 0.05 g-C 3 N 4 /Bi 2 S 3 / In 2 S 3 catalyst. Aer repeating the degradation test four times, Fig. 7 (a) Influence that the illumination time has on the photocatalytic degradation of the catalysts, (b) the pseudo-first-order dynamics, and (c) the apparent rate constant of the degradation reaction. Fig. 8 (a)-(c) show the valence band diagrams of g-C 3 N 4 , In 2 S 3 and Bi 2 S 3 ; (d) the free radical trapping experiments for the degradation of RhB over the 0.05CN/Bi 2 S 3 /In 2 S 3 composite. (e) The proposed photocatalytic mechanism for the degradation of RhB on the surface of the 0.05CN/Bi 2 S 3 /In 2 S 3 composite. the degradation performance of the catalyst still reaches 90.9. Therefore, it can be seen that the heterostructure of the composite catalyst not only improves its activity, but also its stability. To further assess whether the structure of the catalyst is damaged aer use, PXRD measurements were carried out on the catalyst before and aer its use, and the results are shown in Fig. 9b. It can be seen that there is basically no change in the PXRD peak patterns of the catalyst before and aer use. Therefore, it can therefore be concluded that the catalyst structure is not destroyed aer use.

Conclusions
In this paper, a one-step hydrothermal method was used to synthesise a 3D dendritic and porous spherical ternary heterostructure nanocomposite g-C 3 N 4 /In 2 S 3 /Bi 2 S 3 . The g-C 3 N 4 /In 2 S 3 / Bi 2 S 3 heterostructure catalyst shows signicantly improved degradation of RhB compared with the components from which it is made. The reason for the improved photocatalytic performance can be attributed to the increase in the specic surface area of the catalyst, meaning that it has more active sites; the formation of 3D dendritic and porous spherical materials enhances the adsorption performance of the catalyst and the composition of the heterostructure catalyst accelerates the separation of photogenerated carriers, thereby promoting the production of O 2 À , OH, and h + . The research herein describes the design of a new type of ternary heterostructure catalyst. The disadvantage here is that the decomposition process of the degradants was not able to be analysed in detail, which remains a subject for future work.

Conflicts of interest
The authors declare that they have no competing interests.