1 Introduction

Nowadays, rapid industrialization growth discharges a large number of non-biodegradable pollutants, causing great harm to the ecological environment. Especially in the manufacturing process of textile industry, 10–50% of printing and dyeing wastewater containing complex aromatic structure will be produced [1]. Therefore, how to effectively degrade printing and dyeing wastewater is one of the most urgent problems to be solved in wastewater treatment. At present, technologies have been reported to remove organic pollutants from wastewater, such as photocatalytic technology, membrane filtration, reverse osmosis, and adsorption [2, 3]. Among different technologies, semiconductor photocatalysis is considered as an efficient, environmental-friendly and low-cost treatment method to degrade contaminants, and its photocatalytic function is attributed to the light-induced generation of charge carrier pairs which can transfer to the surface of the material and exhibit unique redox properties [4, 5]. However, traditional photocatalysts, such as pure TiO2, cannot be fully applied to photocatalysis field due to its broad band gap (about 3.2 eV) only responding in the ultraviolet region and high photoelectron-hole compound rate [6, 7]. Therefore, in order to enhance the utilization rate of sunlight, it has great significance to develop new photocatalysts with superior visible light response.

The emerging graphite-like phase carbon nitride (g-C3N4) material is a polymeric layered semiconductor with defect-rich N-bridged tri-s-triazine or s-triazine as the basic structural unit. Generally, g-C3N4 has been widely used as a material for photocatalytic treatment of pollutants due to its simple preparation method, stable physical and chemical properties, and suitable optical band gap (about 2.7 eV) [8,9,10]. Nevertheless, the certain deficiencies of g-C3N4, including inadequate sunlight absorption ability, low photo response current, and high photoinduced electron–hole pairs composite rate, result in poor photocatalytic efficiency [11, 12]. To overcome the above defects, Zhou et al. have summarized a series of strategies including heteroatom doping, structure modification, shape-control synthesis, semiconductor coupling, and dye-sensitization to improve the photocatalytic activity of g-C3N4 [13]. In recent years, coupling it with the different types of semiconductor materials, such as metal oxides, metal sulfides, oxometallates, and bismuth oxyhalides [14,15,16,17,18], to fabricate II type or Z-scheme heterojunction, is an effective way to enhance the photocatalytic efficiency by broadening the light response range and prolonging the lifetime of the photoinduced charge carriers [19, 20]. Especially, Z-scheme heterojunction is helpful to maintain the superior redox ability of electrons and holes after charge transfer [21]. Moreover, many researchers are devoted to constructing ternary heterojunctions to further improve the photocatalytic performance of catalytic materials [22, 23].

Nanocomposite photocatalysts have been attracted much research attention because of their comprehensive advantages and synergistic effects which contributes to the separation of photogenerated electrons and holes. [24, 25]. As a typical layered semiconductor, Bi2S3 is endowed with a direct band gap of about 1.2–1.7 eV [26]. The narrow band gap is the direct advantage of choosing Bi2S3 as photocatalyst. Bi2S3 was successfully loaded onto g-C3N4 by ultrasonic method [27] and one-pot method [28], which can enhance the absorption of visible light and provide energy for the rapid transfer of photogenerated carriers, thus showing excellent photocatalytic performance. In addition, narrow band gap semiconductor InVO4 (Eg = 2.3 eV) is also an important visible-driven catalytic material due to its high visible light utilization rate, adjustable morphology and size, and simple synthesis method [29]. Studies have shown that the energy levels of g-C3N4 and InVO4 are well matched overlapping band structures [30]. Therefore, it is predicted that the introduction of InVO4 in Bi2S3/g-C3N4 Z-scheme system could construct dual Z-scheme heterojunction, which further expands the reduction or oxidation surfaces to improve the photocatalytic performance [31].

In order to further analyze the optical response and light catalytic properties of g-C3N4-based composites, on the basis of Bi2S3/g-C3N4, ternary dual Z-scheme InVO4/Bi2S3/g-C3N4 heterojunctions were firstly synthesized in this work. The crystal structure, micromorphology, chemical composition, and optical properties of the ternary photocatalyst were analyzed by means of various characterization methods. The photocatalytic activity was measured by degradation of anion dye Reactive Blue 19 (RB19), and the photocatalytic degradation mechanism was explored by free radical capture experiment. In comparison with pure g-C3N4, InVO4 and Bi2S3/g-C3N4, the as-prepared InVO4/Bi2S3/g-C3N4 composites displayed higher visible light absorption efficiency, stronger catalytic activity, and better stability. With the optimal content of 5 wt% InVO4 in the composite, 5wt% InVO4/Bi2S3/g-C3N4 performed the excellent visible light catalytic activity for the degradation of RB19, and the degradation rate of RB19 (15 mg/L) could reach 97% after 100 min of visible light irradiation. To prove the advantages of InVO4/Bi2S3/g-C3N4, Table 1 presents a retrospective analysis using multiple photocatalysts for earlier studies on the degradation of dye under visible light.

Table 1 Comparison of the photocatalytic degradation efficiency
Full size table

2 Experimental section

2.1 Chemicals and materials

Urea (H2NCONH2, 99%) and melamine (C3H6N6, 99%) were bought from Tianjin Damao Chemical Reagent Factory (China). Bi(NO3)3·5H2O (99%) was obtained from Tianjin Kaitong Chemical Reagent Co., Ltd. Na2S2O3·5H2O (99%) was provided from Tianjin Guangfu Technology Development Co., Ltd. InCl3·4H2O (99%) and NH4VO3 (99%) were from Tianjin Jinke Fine Chemical Research Institute. RB19 was purchased from Shanghai Maclean Biochemical Technology Co., Ltd. P-benzoquinone (BQ, 99%) was from Tianjin Qinghua Jinying Technology Co., Ltd. Tert-butanol (TBA, 99%) came from Tianjin Beichen Founder Reagent Factory. Ammonium oxalate (AO, 99%) was from Tianjin Zhiyuan Chemical Reagent Co., Ltd. Nitrotetrazolium Blue chloride (NBT) was from Hefei BSF Biology Science and Technology Co., Ltd. In this research, all of the materials were analytical grade without further purification. In addition, distilled water was used during the whole experiment process.

2.2 Catalyst preparation

2.2.1 Preparation of Bi2S3/g-C3N4

Pure g-C3N4 was prepared by calcining urea and melamine at high temperature [36]. Typically, 10.5 g urea and 4.5 g melamine were added to a covered alumina crucible and placed in a muffle furnace for 4 h at 550 °C, in which heating rate is 5 °C/min. After cooling to room temperature, the obtained light yellow g-C3N4 was ground into powder for further use.

Bi2S3 was prepared by a wet-impregnation method [37]. Firstly, 1.9403 g Bi(NO3)3·5H2O and 0.7445 g Na2S2O3·5H2O were uniformly dissolved in 50 mL methanol by ultrasonication for 10 min, and then stirred on a magnetic agitator for 60 min. Then, the suspension was transferred into a 100-mL polytetrafluoroethylene-lined stainless autoclave, heated at 180 °C for 12 h. Finally, the products were cooled to room temperature naturally, collected by centrifugation, and washed with anhydrous ethanol, distilled water for 3 times, and fully dried in an oven at 60 °C to obtain the grassy black Bi2S3 sample.

Bi2S3/g-C3N4 composite was synthesized by a simple wet-impregnation and calcination. In detail, 1 g g-C3N4 was dispersed in 80 mL methanol and treated with ultrasonic for 30 min to obtain solution A. Then, Bi2S3 with a mass doping ratio of 5% was dissolved in g-C3N4 solution, and the solution B was obtained after ultrasonic treatment for 60 min. Then, A and B were mixed through magnetic stirring for 24 h. After the reaction, the product was dried at 60 °C and finally calcined at 250 °C for 3 h. After grinding, the light yellow powdery photocatalyst of Bi2S3/g-C3N4 was collected and denoted as 5BS/CN.

2.2.2 Preparation of InVO4/Bi2S3/g-C3N4

InCl3·4H2O and NH4VO3 were used to fabricate bare InVO4. Usually, 0.0638 g InCl3·4H2O and 0.0255 g NH4VO3 were dissolved in 50 mL distilled water. Subsequently, the NaOH solution was added dropwise to the above mixture and the pH was adjusted to 6–8. The InVO4 was obtained by ultrasonic for 30 min.

A series of InVO4/Bi2S3/g-C3N4 hybrid with different doping ratios of InVO4 (denoted as xV/5BSCN, x = 3, 5, 10, 15, and 20 wt% with x representing the weight percentage of InVO4) were synthesized by the following method. 1 g 5BS/CN uniformly dispersed in 80 mL distilled water by ultrasonication for 10 min. Then, a certain amount of InVO4 was added into the above suspension. Next, the mixture was continued to be sonicated for 30 min and further magnetic stirred for 24 h at room temperature. Finally, ternary InVO4/Bi2S3/g-C3N4 photocatalyst was fully dried at 60 °C and collected by ground.

2.3 Catalyst characterization

The phase purity and crystal structure information were determined by X-ray diffraction measurement (XRD) (LABXRD-6000) using Cu-Kα source (λ = 0.154 nm) with a step size of 0.02° in the range of 10°–80° at the scan rate of 2°/min at 30 mA and 40 kV. The functional groups and chemical bonds were qualitative analyzed by Fourier transform infrared (FT-IR) spectra on thermo Scientific Nicolet iS10 spectrometer taking KBr as the reference in the range of 500–4000 cm−1. The morphologies of the photocatalysts were observed. The morphologies and microstructures of the samples were investigated by the transmission electron microscopy (TEM) (JEM-2500SE) at 200 kV and the field emission scanning electron microscopy (FE-SEM) (JSM-7100F) at 15 kV. Energy dispersive X-ray spectroscopy (EDS) attached to SEM was used to determine the chemical composition of the composite. N2 adsorption and desorption isotherms at − 196 °C were obtained using a Micromeritics ASAP 2020.The BET (Brunauer–Emmett–Teller) method and BJH (Barrett–Joyner–Halenda) method were used to calculate the specific surface area and pore size distribution, respectively. The absorption edge of the catalysts was measured by UV–Vis diffuse reflectance spectra (UV–Vis DRS) (Hitachi U-3900, Japan) using Ba2SO4 as the background in the range from 200 to 800 nm. The catalysts photoluminescence (PL) spectra were recorded on a Shimadzu RF-6000 fluorescence spectrophotometer with slit width of 3 nm in the range from 400 to 700 nm at the scan rate of 6000 nm/min. The photocurrent response and electrochemical impedance spectroscopy (EIS) were measured in Na2SO4 (0.1 M) electrolyte solution by Chi660E electrochemical work station with a three-electrode cell, and the platinum electrode, ITO glass (25 × 15 × 1.1 mm) which was coated by the samples, Ag+/AgCl electrode were used as counter electrode, working electrode, and reference electrode, respectively.

2.4 Catalytic experiments

The photocatalytic activities were evaluated by the degradation of RB19 under visible light irradiation. The visible light source was a 300 W Xenon lamp (CEL-HXF300) equipped with a 420 nm cutoff filter, 15 A current, and 14 V voltage. Typically, 200 mg of photocatalysts with different mass ratios of InVO4 was added to a 250 mL 20 mg/L of RB19 dyes solution and then placed in a 300 mL glass reactor. Before the irradiation reaction, the above suspension was firstly dispersed by ultrasound for 10 min and then stirred in the dark for 40 min to establish an adsorption–desorption equilibrium. Afterward, the suspension was placed under visible light in the photocatalytic chamber for 100 min. To eliminate the effect of temperature, the reaction vessel was endowed with a large beaker containing an appropriate amount of cold water (25 ± 1 °C). Meanwhile, 10 mL of photoreacted solution was sampled at regular irradiation time intervals (20 min) and centrifuged twice for 5 min each (10,000 rpm). Lastly, the concentration of RB19 was tested on a UV-1800PC spectrophotometer (MAPAD, Shanghai, China) (λmax = 594 nm). The photocatalytic efficiency (η) was calculated according to the following formula:

$$\eta = \frac{C}{{C_{0} }},$$
(1)

where C0 is the initial concentration of RB19 dyes solution and C is the real-time concentration of RB19 dyes solution during the reaction. Furthermore, the first-order rate constant kobs of the photocatalytic process is measured by the following formula [38]:

$$k_{{{\text{obs}}}} = - \frac{1}{t}\ln \left( {\frac{C}{{C_{0} }}} \right),$$
(2)

where kobs (min−1) is the first-order rate constant, C0 is the RB19 dyes concentration after 40 min dark adsorption (mg/L), and C is the instantaneous concentration of the RB19 solution at illuminated time t (min).

In addition to the experiment of optimal xV/BSCN doping ratio, the influence of catalyst dosage, initial concentration of RB19, initial pH, and the stability of the photocatalyst were also studied. The photocatalytic properties of xV/BSCN were explored from different perspectives.

2.5 Reactive species experiments

2.5.1 Active species trapping experiments

In order to explore the reaction mechanism and determine the generation of main active species in the process of the photocatalytic degradation of RB19, 2 mL TBA, 0.0621 g AO, and 0.0054 g BQ quenchers were used as scavengers of hydroxyl radicals (·OH), photo-holes (h+), and superoxide radicals (O2−), respectively, were added to the reaction system after the adsorption–desorption equilibrium, and the remaining steps were the same as for other degradation experiments.

2.5.2 NBT transformation experiments

NBT was often used to quantitatively analyze the·O2− in photocatalytic reaction [39]. Briefly, 50 mg as-prepared catalysts were dispersed in 250 mL distilled water by ultrasonication for 10 min. Next, 10.2205 mg NBT was added in the above dispersion to react. After a given time interval, the concentration of NBT was measured with recording variations of the absorption band maximum (259 nm) in the UV–Vis spectra to react with the amount of ·O2−.

3 Results and discussion

3.1 Optimal doping ratio selection

Figure 1a and b shows the photodegradation efficiency of 25 mg/L RB19 and the first-order rate constants of each reaction over the g-C3N4, InVO4 (V), 5BS/CN, and the ternary catalysts with different InVO4 doping ratios. It could be clearly seen from Fig. 1a that the adsorption capacity of g-C3N4 to RB19 solution was larger, reaching 41.2%, and the photocatalytic degradation rate of RB19 solution under visible light was lower and degraded to 85.3% after 100 min of light exposure. Pure InVO4 had no adsorption and degradation ability at all. The binary catalyst 5BS/CN improved the degradation capacity of pure g-C3N4 to 91.2% after 100 min of visible light irradiation. The photocatalytic performance of 5BS/CN was further increased by doping InVO4. Obviously, the catalytic activity of the xV/5BSCN composites was closely related to their doping ratios. The photocatalytic rates of 3 V/5BSCN, 5 V/5BSCN, 10 V/5BSCN, 15 V/5BSCN, and 20 V/5BSCN were 91.2%, 95.5%, 89.1%, 86.6%, and 85.3%. Clearly, as the doping ratio of InVO4 increased from 3 to 5%, the photodegradation efficiency increased gradually; however, when the loading percent of InVO4 further increased from 5 to 20%, the photocatalytic rate for RB19 tended to a significant decrease. It indicated that 5 V/5BSCN had the best photocatalytic activity. This might be due to excessive InVO4 deposition, which was not well dispersed on the surface of the material, affecting its absorption of light, reducing the number of effective heterogeneous junctions in the composite material, and influencing charge transfer between heterogeneous junction interfaces [30]. Similarly, the first-order rate constant of 5 V/5BSCN also was the largest at 40 min. Therefore, the 5 V/5BSCN were selected as the optimal photocatalyst for the following studies.

Fig. 1
figure 1

a Photocatalytic comparison of xV/5BSCN for different InVO4 doping ratios and b first-order rate constants (the dosage of catalyst = 1.0 g/L, CRB19 = 25 mg/L, pH 7.0)

Full size image

3.2 Characterization

3.2.1 Structure and morphology

The crystal structure of as-synthesized g-C3N4, Bi2S3, InVO4, 5BS/CN, 5 V/5BSCN were detected by XRD pattern and are shown in Fig. 2. Obviously, two diffraction peaks of pure g-C3N4 were observed at 13.02° and 27.52°corresponding to (100) and (002) planes of pristine g-C3N4 (JCPDS 87-1526), which were derived from the tri-s-thiazine units and the stacking of the conjugated aromatic systems, respectively [40, 41]. The crystal structures of bare Bi2S3 were orthorhombic crystal structure and their diffraction peaks at 23.44°, 29.1°, 31.94°, 45.56°, 52.46°, and 62.40° could be indexed to (130), (211), (221), (002), (351), and (152) crystal planes respectively, which matched well with the standard card (JCPDS 17-0320) [42, 43]. In the XRD pattern of 5BS/CN, the diffraction peaks of g-C3N4 and Bi2S3 appeared simultaneously, which indicated the successful loading of Bi2S3 without impurity. The XRD spectra of pure InVO4 showed that the diffraction peaks at 19.8°, 21.5°, 31.8°, 33.0°, 41.6°, 51.0°, and 60.9° could be ascribed to (110), (020), (200), (112), (202), (042), and (242) crystal planes of orthorhombic phase InVO4 (JCPDS. No 48-0898), [44]. With the addition of InVO4 to 5BSCN, it could be found that the main diffraction peaks of InVO4 and 5BS/CN existed at the same time in the XRD pattern of 5 V/5BSCN hybrid, which illustrated the successful preparation of the ternary photocatalysts. Furthermore, the weak the characteristic peak intensity of InVO4 in the composite might be due to the low content of InVO4.

Fig. 2
figure 2

XRD patterns of g-C3N4, Bi2S3, 5BS/CN, InVO4 and 5 V/5BSCN

Full size image

The functional groups and bonds of g-C3N4, 5BS/CN, and 5 V/5BSCN were tested on the FT-IR spectra, the results of which are displayed in Fig. 3. It could be seen from Fig. 3 that the characteristic peaks of the three materials were mainly concentrated in the range of 500–3500 cm−1. In the FT-IR spectrum of g-C3N4, the sharp absorption band at 807 cm−1 was from the distinctive vibration of tri-s-triazine ring units [45], the absorption peaks in the range of 1200–1700 cm−1 were caused by stretching vibrations of C–N and C=N bonds [46]. A wide absorption band at 3250 cm−1 was the stretching vibrations of O–H and N–H due to the adsorption of water molecules and residue of amino groups on the surface of g-C3N4. In the FT-IR spectra of 5BS/CN, the absorption peaks at 3254 cm−1 and 811 cm−1 which also existed in the FT-IR spectra of pristine Bi2S3, corresponding to the stretching vibrations of the residual N–H and Bi–S groups respectively [47], thus demonstrating the existence of Bi2S3 in 5BS/CN composite. The diffraction peak of 5 V/5BSCN appeared at 809 cm−1 owing to the interaction between tri-s-triazine ring units of g-C3N4 and Bi–S groups of Bi2S3, which also suggested the coexistence of g-C3N4 and Bi2S3. Moreover, the diffraction peak of 5 V/5BSCN at 3253 cm−1 was due to the stretching vibrations of N–H. However, no characteristic peaks of InVO4 could be observed in the pattern of 5 V/5BSCN, which might be attributed to low amount of InVO4 loading on 5BS/CN. In fact, the existence of InVO4 in the nanocomposite could be further verified by TEM based on the recognition of the morphology and microstructure.

Fig. 3
figure 3

FT-IR spectra of g-C3N4, 5BS/CN and 5 V/5BSCN

Full size image

Figure 4 displays TEM images of g-C3N4, Bi2S3, 5 V/5BSCN, and HR-TEM images of 5 V/5BSCN. Figure 4a shows that the internal structure of g-C3N4 was layered and overlapped. And Fig. 4b presents that Bi2S3 had a rod-like structure. TEM image of 5 V/5BSCN (Fig. 4c) showed that the bar Bi2S3 and the irregular stacked layered g-C3N4 were uniformly covered with dark-colored flake substances, indicating that the InVO4 was doped into the 5BS/CN composite material to form the ternary photocatalyst, which was consistent with the XRD analysis. In addition, Fig. 4d shows that the lattice spacing of the lattice stripes was 0.346 nm attributed to the (220) crystal plane of InVO4 [48].

Fig. 4
figure 4

TEM images of a g-C3N4, b Bi2S3, c 5 V/5BSCN and HR-TEM images of d 5 V/5BSCN

Full size image

The distribution of InVO4 in the 5 V/5BSCN composite was further demonstrated by elemental mapping. As shown in Fig. 5, C, N, O, S, Bi, In, and V elements were uniformly distributed within 5 V/5BSCN composite. Meanwhile, the weight percentages of each element in 5 V/5BSCN were detected in Fig. 5i, further implying the coexistence of InVO4, Bi2S3 and g-C3N4 without other impurities.

Fig. 5
figure 5

a SEM image of 5 V/5BSCN, and EDX layered images of b C, c N, d O, e S, f Bi, g In, h V and i EDX spectrum of 5 V/5BSCN

Full size image

The textural properties of pure g-C3N4 and 5 V/5BSCN were inferred from their N2 adsorption–desorption isotherms (Fig. 6). All photocatalysts showed type-IV adsorption isotherms with H3 type hysteresis loops, indicating the existence of slit-like pores created by the stacked nanosheets [49]. BET surface area for g-C3N4 and 5 V/5BSCN was 63.4485 m2g−1 and 59.7659 m2g−1, respectively. Compared to the pure g-C3N4, the composite materials delivered low surface area. The decrease in the surface area of g-C3N4 might be due to the fact that spherical InVO4 dispersed on the surface of CN blocks the pores of CN. These results could be affirmed by the above-mentioned SEM. It was also a good explanation that the adsorption capacity of RB19 of composite material was lower than that of pure CN. Moreover, 5 V/5BSCN showed more micropores (2 nm) that might be generated by stacked nanosheets.

Fig. 6
figure 6

BET surface area and the corresponding BJH pore size distribution curves (inset) of pure g-C3N4 and 5 V/5BSCN composite

Full size image

3.3 Optical properties

The light harvesting capacities of g-C3N4, Bi2S3, InVO4, and 5BS/CN, 5 V/5BSCN composites can be evaluated by UV–Vis diffuse reflectance (DRS) spectra. As shown in Fig. 7a, the optical absorption edges of g-C3N4, Bi2S3, and InVO4 were 475, 1130, and 620 nm respectively, while the optical absorption edges of the doped composite materials 5BS/CN and 5 V/5BSCN were 485 and 680 nm, respectively. This indicated that the light absorption edges of the composites had red shift compared with that of pure g-C3N4, and the visible light absorption region was significantly broadened, thus improving the utilization efficiency of visible light.

Fig. 7
figure 7

a UV–Visible diffuse reflectance spectrum and b band gap diagram of g-C3N4, Bi2S3, 5BS/CN, InVO4 and 5 V/5BSCN

Full size image

The band gap energy (Eg) of photocatalysts can be calculated by the following Kubelka–Munk Eq. [50]:

$$\alpha hv = A\left( {hv - E_{{\text{g}}} } \right)^{\frac{n}{2}}$$
(3)

where α is the absorption coefficient, Eg is the band gap energy, A, h, and v are a proportionality constant, plank constant, and light frequency, respectively, the value of n depends on the type of optical transition of semiconductors (n = 1 for direct transition and n = 4 for indirect transition). According to the previous research, g-C3N4, Bi2S3 and InVO4 are all direct semiconductors, so the value of n is 1 for them [41, 44, 51]. According to the plots of (αhv)1/2 versus hv (Fig. 7b), the calculated Eg values of g-C3N4, Bi2S3 and InVO4 were 2.74, 1.90, and 2.34 eV respectively. After the introduction of Bi2S3, the band gap of 5BS/CN composites was 2.25 eV, which was 0.49 eV smaller than that of pure g-C3N4. Moreover, InVO4 was further introduced, the band gap of the 5 V/5BSCN was continuously reduced to 1.96 eV. This indicated that heterogeneous junctions were constructed between g-C3N4, Bi2S3 and InVO4, which effectively reduced the interface contact barrier, strengthened the electron coupling between semiconductors, and thus enhanced the photocatalytic activity [52]. In addition, the valence band (VB) and conduction band (CB) of semiconductors can be calculated by the following formula [50]

$$E_{{{\text{CB}}}} = X - E^{{\text{e}}} - 0.5E_{{\text{g}}}$$
(4)
$$E_{{{\text{VB}}}} = E_{{{\text{CB}}}} + E_{{\text{g}}}$$
(5)

where EVB and ECB are the valence band potential and conduction band potential of semiconductors. X is the electronegativity of the semiconductor, which is the geometric average value of the electronegativity of constituent atoms, and the X values of g-C3N4, Bi2S3, and InVO4 are 4.67, 5.56, and 5.03 eV, respectively [53,54,55]. Ee is the energy of free electrons on the hydrogen scale (about 4.5 eV). Using the Eqs. (4, 5), the VB edge potentials of g-C3N4, Bi2S3, and InVO4 were 1.54 eV, 2.01 eV, and 1.70 eV, respectively, and the CB edge potentials of them were − 1.20 eV, 0.11 eV, and − 0.64 eV, respectively.

3.4 Photocatalytic activities

3.4.1 Effect of catalyst dosage

In order to explore the effect of the photocatalyst dosage on degradation of RB19, the dosage gradient of 5 V/5BSCN was set at 0.4, 0.6, 0.8, 1.0, and 1.2 g/L. Accordingly, the degradation rate of 25 mg/L RB19 and the first-order rate constant under 40 min illumination were studied, as shown in Fig. 8. Obviously, with the increase of the catalyst dosage, the adsorption capacity of the dark reaction was augmented, and when the catalyst dosage reached the maximum of 1.2 g/L, the dark adsorption capacity of RB19 achieved the maximum of 38.16% accordingly. This was because physical adsorption occurred during the dark reaction, in which the more the catalyst, the more the pollutant molecules were adsorbed. Figure 8a exhibits that the final degradation rate of RB19 also increased with the increase of dosage, which was 39.05%, 47.66%, 95.46%, 97.34%, and 97.42% respectively. Remarkably when the dosage of 5 V/5BSCN increased from 0.6 to 0.8 mg/L, the degradation rate increased significantly, and then the dosage further increased to 1.2 mg/L, the degradation rate increased slowly. It could be seen from Fig. 8b that the rate constants corresponding to the dosage were 0.351 × 10–2, 0.580 × 10–2, 2.931 × 10–2, 2.905 × 10–2, and 2.884 × 10–2 min−1, which showed a trend similar to that of degradation rate. This illustrated that an appropriate amount of catalyst could significantly improve the photocatalytic degradation efficiency, but increasing the amount of catalyst further had little effect on the degradation of pollutants and could even inhibit the process. This was because a small amount of the catalyst did not produce enough active charge carriers to afford excess amount of RB19 molecules, with the increase of its dosage, the more active sites could be contributed to enhance light utilization efficiency [56]. Nevertheless, when catalyst dosage surpassed a certain amount, the light transmittance began lower down, so less light could be irradiated to the surface of photocatalysts, which reduced the generation of the photogenerated electrons and holes, eventually resulting in the decrease of photocatalytic efficiency [56, 57]. Therefore, 0.8 g/L was chosen as the optimal dosage for 5 V/BSCN for subsequent exploratory experiments.

Fig. 8
figure 8

a Effect of dosage on photocatalysis of 5 V/5BSCN and b first-order rate constants (CRB19 = 25 mg/L, pH 7.0)

Full size image

3.4.2 Effect of initial concentration of RB19 solution

In order to investigate the effect of initial concentration of RB19 solution on degradation of RB19, the concentration gradient of RB19 was set at 10 mg/L, 15 mg/L, 20 mg/L, 25 mg/L, and 30 mg/L, and the dosage of 5 V/5BSCN was 0.8 g/L. As shown in Fig. 9, 5V/5BSCN degrading 15 mg/L and 25 mg/L RB19 performed better catalytic activity considering the dark adsorption capacity and degradation rate. And there was an inflection point when the initial concentration of the pollutant increased from 25 to 30 mg/L, the degradation rate decreased dramatically. The result indicated that RB19 solution of high concentration suppressed photocatalytic effect. This was because with the increase of the concentration of RB19, the increasing adsorbed pollutants on the surface of the catalyst consumed photoelectron-hole pairs gradually which were hindered to recombination, thus showing a high degradation rate. When the concentration of pollutant solution reached to a high level certainly, on the one hand, the massive sedimentary of RB19 molecules reduced the visible light transmittance; on the other hand, the intermediate by-products that cannot be decomposed in time competed with pollutant molecules for active sites, retarding the reaction, thus leading to the reduction of photocatalytic efficiency [57]. Figure 9b shows that when the initial concentration of RB19 was 15 mg/L, the corresponding first-order rate constant presented a maximum value. So, 15 mg/L was the most suitable initial concentration of RB19 solution among them.

Fig. 9
figure 9

a Effect of different concentrations on photocatalysis of 5 V/5BSCN and b first-order rate constants (the dosage of catalyst = 0.8 g/L, pH 7.0)

Full size image

3.4.3 Effect of initial pH

In order to study the effect of the initial pH of the reaction solution on degradation of RB19, the photocatalytic degradation efficiency of RB19 was tested by 10ABCN at initial pH 3, 5, 7, 9, and 11. The initial concentration of RB19 was 15 mg/L, and the dosage of 5 V/5BSCN was 0.8 g/L. Figure 10 presents that the dark adsorption capacities of RB19 at the initial pH of 3, 5, 7, 9, and 11 were 92.17%, 48.50%, 25.93%, 25.93%, and 31.46%, respectively. It manifested that the higher the acidity, the greater the adsorption capacity, because pH affected the electronegativity of the catalyst [50, 58]. Under acidic conditions, the surface of the catalyst was positively charged, while RB19 was a negatively charged anionic dye, then the electrostatic attraction occurred between the photocatalysts and pollutants giving rise to a plentiful adsorption of RB19 molecules on the surface of 5 V/5BSCN. On the contrary, in the alkaline environment, the surface of the catalyst was negatively charged, so there was electrostatic repulsion with the negatively charged RB19, which leaded to the weakening of the adsorption capacity of 5 V/5BSCN. As for the degradation effect, the degradation trend performed faster under alkaline conditions, especially when pH 11. This was because O2 adsorbed on the surface of the catalyst reacted with the negative charge on the surface to generate active species·O2− [59], then RB19 was rapidly degraded. However, after 40 min of photoreaction, the degradation rate of RB19 under alkaline conditions gradually slowed down, and the degradation under pH 11 reaction conditions almost stopped, mainly because the sulfur group of RB19 was substituted to form vinyl sulfone in alkaline environment and 2-hydroxyethyl sulfone [60], the stronger the basicity was, the stronger the substitution effect of the sulfur group was, thus slowing down the photocatalytic reaction process.

Fig. 10
figure 10

Effect of initial pH on photocatalysis of 5 V/5BSCN (the dosage of catalyst = 0.8 g/L, CRB19 = 15 mg/L)

Full size image

3.4.4 Stability experiment

The stability of 5 V/5BSCN was tested by photocatalytic degradation of RB19 in four cycles with the dye concentration of 15 mg/L. As shown in Fig. 11a, the degradation rate of RB19 decreased slightly after each cycle, reaching about 76% after 4 cycles, indicating that the 5 V/5BSCN composite material was relatively stable during photocatalytic degradation of RB19 dyes. Moreover, XRD (Fig. 11b) and SEM (Fig. 11c, d) displayed that the crystal structure of 5 V/5BSCN heterojunction was not destroyed in the progress of the photocatalytic reaction. The stability might be due to the tight heterojunction interface conducive to the rapid separation of photogenerated electrons and holes.

Fig. 11
figure 11

a Cyclic experiment, b XRD patterns, c, d SEM images of 5 V/5BSCN before and after photoreaction

Full size image

3.5 Photocatalytic mechanism

Fluorescence spectroscopy (PL) is an important method to study the photogenerated electron–hole separation ability. The stronger the peak of the fluorescence emission spectrum was, the higher the electron–hole recombination rate was [7, 61]. Figure 12 depicts the fluorescence spectra of pure g-C3N4, 5BS/CN, and 5 V/5BSCN. The PL spectrum of pure g-C3N4 had a strong emission peak caused by the band transition at 460 nm, which revealed that the electron–hole recombination rate was quite fast [62, 63]. Compared with pure g-C3N4, the PL peak of the 5BS/CN composite material was significantly weaker, illustrating that the construction of heterojunction could effectively inhibit the recombination of photogenerated carriers. Moreover, the PL peak of 5 V/5BSCN was further weaker, indicating that the continuous loading of InVO4 on the basis of 5BS/CN composite to construct a ternary heterojunction could further inhibit the recombination of photogenerated carriers and improve the photocatalytic degradation efficiency.

Fig. 12
figure 12

PL spectra of g-C3N4, 5BS/CN and 5 V/5BSCN

Full size image

The photocurrent response and EIS further demonstrated the separation and transfer of photoinduced charge carriers. The photocurrent response of CN, 5BS/CN, and 5 V/5BSCN was monitored for several on–off cycles at open-circuit potential. Higher photocurrent density means a stronger carrier separation capacity [64]. As shown in Fig. 13a, the 5 V/5BSCN exhibited the highest photocurrent density, which were 2.3 and 2 times greater than g-C3N4 and 5BS/CN, respectively, indicating that 5 V/5BSCN had a higher carrier separation ability. The result of EIS is shown in Fig. 13b, the smaller the radius in EIS diagram was, the smaller the charge transfer resistance was at the electrode interface, suggesting the faster interfacial electron transport [65]. The EIS measurements revealed that the radius of Nyquist curve of 5 V/5BSCN was smaller than that of CN and 5BS/CN, and therefore, 5 V/5BSCN was proved to possess higher photogenerated charge transferability than pure CN.

Fig. 13
figure 13

Photocurrent response (a) and EIS measurements (b) of CN, 5BS/CN and 5 V/5BSCN

Full size image

Figure 14 shows the results of active species trapping experiments. It could be seen that the addition of TBA had almost no effect on the degradation rate of RB19, which indicated OH slightly participated in the photocatalytic degradation process. After adding BQ acting as a scavenger of O2−, the degradation rate had a sharp decline. Also, AO slightly inhibited the reaction as h+ scavenger. Therefore, O2− played a main part during the photocatalytic reaction, and h+ was secondary. In order to further determine the main active species deciding RB19 degradation, NBT was introduced to the experiment. Figure 15 shows the spectra of the NBT transformation percentage in the presence of 5 V/5BSCN. As the photoreaction proceeded, NBT was gradually transformed, and the absorbance of NBT decreased significantly at 120 min, which implied varieties of ·O2− radicals were generated during the process.

Fig. 14
figure 14

Trapping experiment of 5 V/5BSCN photocatalysis (the dosage of catalyst = 0.8 g/L, CRB19 = 15 mg/L, pH 7.0)

Full size image
Fig. 15
figure 15

UV−Vis absorbance spectra of Nitrotetrazolium Blue chloride aqueous solution containing 0.8 g/L of 5 V/5BSCN photocatalyst

Full size image

Based on the above analysis and band structure matching, dual Z-scheme path was favorable for the photocatalytic mechanism and charge transfer pathways of 5 V/5BSCN composite, as shown in Fig. 16. Under the irradiation of visible light, the electrons on the VB of Bi2S3, g-C3N4, and InVO4 both were excited to CB, leaving holes on the VB, respectively. Then, the electrons on the CB of Bi2S3 and InVO4 recombined with the holes on the VB of g-C3N4 through the heterojunction interface, and photogenerated electrons that accumulated on the CB of g-C3N4 reduced O2 to·O2−, leading to the decomposition of RB19 molecules. Since the VB potential of Bi2S3 was positive than E0(H2O/OH) potential (1.99 eV vs. NHE),·OH radicals were generated on the VB of Bi2S3. Meanwhile, the photogenerated holes left behind on the VB of InVO4 would directly participate in the reaction of degrading RB19. In addition, because g-C3N4 accounted for the largest proportion of 5 V/5BSCN, O2− radicals generated by g-C3N4 were the main active species for oxidizing RB19 to degradation products, which was consistent with the results of the active species trapping and NBT conversion experiments. Therefore, the 5 V/5BSCN photocatalytic material realized the rapid separation of photogenerated carriers by constructing a dual Z-scheme heterojunction successfully, and improved the yield and lifetime of photogenerated carriers at the same time, which was also confirmed in the above-mentioned PL spectrum and electrochemical analysis results. The specific reaction process was shown in the following formula:

$${\text{InVO}}_{{4}} {\text{/Bi}}_{{2}} {\text{S}}_{{3}} {\text{/g - C}}_{{3}} {\text{N}}_{{4}} \mathop{\longrightarrow}\limits^{{{\text{hv}}}}{\text{InVO}}_{{4}} {\text{(e}}_{{{\text{CB}}}}^{ - } + {\text{h}}_{{{\text{VB}}}}^{ + } {\text{)/Bi}}_{{2}} {\text{S}}_{{3}} {\text{(e}}_{{{\text{CB}}}}^{ - } + {\text{h}}_{{{\text{VB}}}}^{ + } {\text{)/g - C}}_{{3}} {\text{N}}_{{4}} {\text{(e}}_{{{\text{CB}}}}^{ - } + {\text{h}}_{{{\text{VB}}}}^{ + } {)}$$
$${\text{e}}_{{{\text{CB}}}}^{ - } {\text{(g - C}}_{{3}} {\text{N}}_{{4}} {)} + {\text{O}}_{{2}} \to \cdot {\text{O}}_{{2}}^{ - }$$
$${\text{h}}_{{{\text{VB}}}}^{ + } {\text{(Bi}}_{{2}} {\text{S}}_{{3}} {)} + {\text{H}}_{{2}} {\text{O}} \to \cdot {\text{OH(small amount)}} + {\text{H}}^{ + }$$
$${\text{h}}_{{{\text{VB}}}}^{ + } {\text{(Bi}}_{{2}} {\text{S}}_{{3}} {)} + {\text{OH}}^{ - } \to \cdot {\text{OH(small amount)}}$$
$$\cdot {\text{O}}_{{2}}^{ - } + {\text{RB19}} \to {\text{Degradation products}}$$
$${\text{h}}_{{{\text{VB}}}}^{ + } {\text{(Bi}}_{{2}} {\text{S}}_{{3}} {)} + {\text{RB19}} \to {\text{Degradation products}}$$
$${\text{h}}_{{{\text{VB}}}}^{ + } {\text{(InVO}}_{{4}} {)} + {\text{RB19}} \to {\text{Degradation products}}$$
$$\cdot {\text{OH(small amount) + RB19}} \to {\text{Degradation products}}$$
Fig. 16
figure 16

Possible photocatalytic mechanism diagrams of 5 V/5BSCN

Full size image

4 Conclusions

In summary, the ternary dual Z-scheme InVO4/Bi2S3/g-C3N4 heterojunction photocatalyst was successfully synthesized by the wet impregnation method. Under visible light, the 5 V/5BSCN composite, i. e., the loading ratio of InVO4 was 5%, performed the excellent photocatalytic degradation activity for RB19. After 100 min visible light irradiation, the degradation rate of RB19 achieved 95.5%. The improvement of the degradation rate of RB19 by InVO4/Bi2S3/g-C3N4 was ascribed to generate tight heterogeneous interfaces to construct the double Z-scheme heterostructure, which effectively inhibited the recombination of photogenerated carriers on the one hand, and enhanced the oxidation on the other hand. The active species trapping experiment showed that the role of free radicals in the degradation experiment followed ·O2− > h+ > OH. According to the results of this experiment, the dual Z-scheme charges transfer mechanism could provide an effective strategy to design and fabricate g-C3N4-based nanocomposites for wastewater treatment.