Facile preparation of novel quaternary g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites: magnetically separable visible-light-driven photocatalysts with significantly enhanced activity

Abstract

In this paper, we successfully fabricated quaternary g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites with various weight percents of Bi₂S₃, as novel magnetically separable visible-light-driven photocatalysts, using a facile refluxing method at 96 °C. The prepared samples were characterized by XRD, EDX, SEM, TEM, UV-vis DRS, FT-IR, TGA, BET, PL, and VSM techniques. Photocatalytic activity of the samples was investigated by degradation of rhodamine B, methylene blue, and fuchsine under visible-light illumination. It was found that photocatalytic activity of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite, as the optimal photocatalyst, in the degradation of RhB is nearly 56, 44, 6.5, and 16-folds higher than those of the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) samples, respectively. Furthermore, the effects of refluxing time, calcination temperature, scavengers of the reactive species, and number of recycling runs on the photocatalytic activity were investigated. Based on the results, the improved photocatalytic activity was attributed to the enhanced visible-light absorption ability and matching energy levels of the counterparts that facilitate generation and separation of electron–hole pairs, respectively.

---

Introduction

In recent decades, a major cause of water pollution has been organic compounds from different industrial sources. These persistent organic pollutants, introduced into natural water resources or wastewaters, are highly toxic and hazardous to living organisms.¹ Hence, removal of these pollutants from water prior to discharge into the environment is highly desirable. Among different treatment strategies such as physical, chemical, and biological, heterogeneous photocatalytic processes are promising technology for decontamination of water containing stable organic compounds.^2,3 In these processes, after irradiating a photocatalyst with proper photons (their energy equal to or greater than band gap of the photocatalyst), electrons are moved from the valence band (VB) to the conduction band (CB) of photocatalyst to create some electron–hole pairs. After that, the photogenerated charge carriers start up different oxidation and reduction reactions over the surface of the photocatalyst to produce different reactive species.⁴ Many semiconductors including TiO₂, ZnO, and SnO₂ are commonly used in photocatalytic processes.^2,5,6 However, these photocatalysts can be activated only by the light with a wavelength less than about 390 nm.³ It is well known that this radiation accounts for about 4% of the solar energy, resulting in restricted large-scale utilization of this technology in different disciplines. Therefore, in recent years, extensive efforts have been devoted to looking for novel visible-light-driven photocatalysts with considerable activity.^6–16

Graphitic carbon nitride (g-C₃N₄), a new kind of conjugated polymeric semiconductor, has been discovered as a fascinating metal-free organic photocatalyst with moderate activity under visible-light irradiation.¹⁷ This semiconductor has found broad applications, such as splitting of water to produce hydrogen gas, degradation of different pollutants, reduction of carbon dioxide to different fuels, catalysis of different reactions, solar cells, and sensing owing to its high stability, appealing electronic structure, and medium band gap.^18–22 However, pure g-C₃N₄ suffers from some drawbacks, due to recombination of the charge carriers with high rate, low visible-light harvesting ability, and difficulties in recycling from the treated solutions.²³ Therefore, it is very important to design and prepare more efficient g-C₃N₄-based photocatalysts by expanding the light absorption ability further into the visible range. Combining g-C₃N₄ with narrow band gap semiconductors, having matched energy levels, is the most effective strategy to overcome the first and second shortcomings.^24–29 Furthermore, deposition of magnetic materials over g-C₃N₄ sheets is fascinating method to tackle the third drawback by application of an external magnetic field.^30–32

On the other hand, considerable attention has been recently paid to the bismuth-based semiconductors.^33–35 Bismuth sulfide (Bi₂S₃) is a promising photocatalyst and it has attracted more attention due to its narrow band gap of 1.4 eV.^35,36 Hence, Bi₂S₃ absorbs a wide range of visible-light irradiation. In this regard, different Bi₂S₃-containing photocatalysts have been reported.^37–39 Very recently, we prepared g-C₃N₄/Fe₃O₄/AgI (20%) nanocomposite as visible-light-driven photocatalysts with high photocatalytic activity.⁴⁰ However, because of its weak ability for absorption of visible light, this nanocomposite has moderate activity under visible-light illumination. Hence, it seems that by combining g-C₃N₄/Fe₃O₄/AgI (20%) nanocomposite with Bi₂S₃ particles, we could prepare novel magnetically separable quaternary nanocomposites with significantly enhanced activity. For the best of our knowledge, there is not any report about magnetically separable photocatalysts containing g-C₃N₄ and Bi₂S₃. Furthermore, preparation and investigation photocatalytic activity of the quaternary g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites have not reported.

In the present study, we successfully prepared g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites with different weight percents of Bi₂S₃, as novel visible-light-driven photocatalysts, using a facile large-scale method and they were characterized using different techniques. Photocatalytic activity of the prepared samples was investigated by degradation of rhodamine B (RhB), methylene blue (MB), and fuchsine under visible-light irradiation. The results indicated that the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite showed the highest photocatalytic activity. We also illustrated that different parameters such as refluxing time, calcination temperature, and scavengers of the reactive species played particularly important roles on the photocatalytic performance of the nanocomposites. Finally, the recycling runs for the degradation reaction of RhB were performed to evaluate stability of the optimal nanocomposite.

Experimental

Materials

Ferric chloride (FeCl₃·6H₂O, 99.5%), ferrous chloride (FeCl₂·4H₂O, 98.0%), ammonia solution (30%), melamine (C₃H₆N₆, 99.2%), silver nitrate (99.9%), sodium hydroxide (98%), and benzoquinone were purchased from Loba Chemie and used as received. Bismuth nitrate (Bi(NO₃)₃·4H₂O, 99%), potassium iodide (99.89%), thioacetamide (C₂H₅NS, 99%), 2-propanol, ammonium oxalate, RhB, MB, fuchsine, and absolute ethanol with high quality were obtained from Merck and used as received. Deionized water was used throughout this work.

Instruments

The X-ray diffraction (XRD) patterns were recorded by a Philips Xpert X-ray diffractometer with Cu Kα radiation (λ = 0.15406 nm). Surface morphology and distribution of particles were studied by LEO1430VP scanning electron microscopy (SEM) of model LEO1430VP, using an accelerating voltage of 15 kV. The purity and elemental analysis of the products were obtained by energy dispersive analysis of X-rays (EDX) on the same SEM instrument. The transmission electron microscopy (TEM) investigations were performed by a Zeiss-EM10C instrument with an acceleration voltage of 80 kV. The UV-vis diffuse reflectance spectroscopy (DRS) data were recorded by a Scinco 4100 apparatus. The Fourier transform-infrared (FT-IR) spectra were obtained using a Perkin-Elmer Spectrum RXI apparatus. The thermogravimetric analysis (TGA) of the samples was performed on Linseis STAPT1000 by heating under air atmosphere from room temperature to 700 °C at 10 °C min⁻¹. The specific surface area and pore properties of the samples were calculated using the Brunauer–Emmett–Teller (BET) and Barret–Joyner–Halenda (BJH) models with nitrogen adsorption–desorption isotherms collected by Belsorp apparatus at −196 °C. Prior to the experiments, the samples were degassed at 120 °C for 15 h. The photoluminescence (PL) spectra of the samples were studied using a Perkin-Elmer (LS55) fluorescence spectrophotometer with an excitation wavelength of 300 nm. The conditions such as amount of the sample, volume of solvent, ultrasonic irradiation time and its power (applied for dispersion of the sample) were fixed in order to compare the PL intensities. UV-vis spectra of the degradation reaction were studied using a Cecile 9000 spectrophotometer. Magnetic properties of the samples were obtained using vibrating sample magnetometer (VSM, Meghnatics Kavir Kashan Co., Iran). The ultrasound radiation was performed using a Bandeline ultrasound processor HD3100 (12 mm diameter Ti horn, 75 W, 20 kHz).

Preparation of the samples

Pyrolysis of melamine, as the starting material, was employed to prepare the g-C₃N₄ sample in a semi closed system.¹⁸ The g-C₃N₄/Fe₃O₄ (2 : 1), where 2 : 1 is weight ratio of g-C₃N₄ to Fe₃O₄, and g-C₃N₄/Fe₃O₄/AgI (20%), where 20% is weight percent of AgI, were prepared using the reported methods by our research group.⁴⁰ The g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites were simply prepared by refluxing method at 96 °C. Typically for preparation of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite, where 30% is weight percent of Bi₂S₃, 0.3 g of the g-C₃N₄/Fe₃O₄/AgI (20%) nanocomposite was dispersed in 150 mL of water via ultrasonication for 10 min and then 0.241 g of bismuth nitrate were added to the suspension. After stirring for 120 min, an aqueous solution of thioacetamide (0.056 g in 20 mL of water) was dropwise added to the suspension with subsequent stirring for 30 min and then refluxed at 96 °C for 60 min. After that, the product was collected by centrifugation, washed with water and absolute ethanol, and dried in an oven at 60 °C for 24 h. According to this method, different weight percents of Bi₂S₃ were loaded and they were labeled as g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (10%), g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (20%), g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (40%) nanocomposites. For comparison purposes, the g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) sample was synthesized by the same method. Typically, 0.241 g of bismuth nitrate was added to 0.3 g of the ultrasonicated g-C₃N₄/Fe₃O₄ (2 : 1) nanocomposite. Next, the suspension was stirred for 120 min and thioacetamide solution (0.056 g in 20 mL of water) was dropwise added under stirring for 30 min. After refluxing for 60 min, the resultant suspension washed with water and ethanol for two times and dried in an oven at 60 °C for 24 h (Scheme 1).

Scheme 1 The procedure for preparation of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites.

Photocatalysis experiments

To investigate photocatalytic activity of the prepared samples, several photocatalytic experiments were comparatively carried out by degradation 250 mL of RhB (2.50 × 10⁻⁵ M), MB (2.50 × 10⁻⁵ M), and fuchsine (9.2 × 10⁻⁶ M) solutions under visible-light irradiation. Due to high absorption coefficient of fuchsine, concentration of fuchsine was very lower than those of RhB and MB dyes. For this purpose, 0.1 g of the powdered photocatalyst was dispersed under ultrasonication for 6 min. For the irradiation system, an LED lamp with 50 W was used as visible-light source at the distance of 20 cm from the solution in a dark box. The emission spectrum of the source has high intensity in visible range and its intensity rapidly decreases in wavelengths near to UV and IR ranges.⁴¹ During the degradation reactions, different samples were withdrawn regularly from the reactor and the catalyst was removed by a magnet to analyze the transparent solutions spectrophotometrically. The absorbances of RhB, MB, and fuchsine solutions were determined as a function of the irradiation time at wavelengths of 553, 664, and 540 nm, respectively.

Results and discussion

Fig. 1a shows XRD patterns of the g-C₃N_4, g-C₃N₄/Fe₃O₄, and quaternary g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites with different amounts of Bi₂S₃ along with the patterns for the g-C₃N₄/Fe₃O₄/AgI (20%) and g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) samples. Two distinct diffraction peaks at 2θ of 13.1 and 27.3° can be found for pure g-C₃N₄, which correspond to the (100) and (002) planes of hexagonal g-C₃N₄ (JPCDS no. 87-1526), respectively.¹⁸ All diffraction peaks of Bi₂S₃ can be well indexed to orthorhombic phase (JCPDS no. 84-0279).⁴² Due to presenting XRD patterns of many samples in Fig. 1a, some of the peaks are not clearly seen. For this reason, the XRD pattern of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite was shown in Fig. 1b. As can be seen, diffraction peaks of g-C₃N₄, Fe₃O₄, AgI, and Bi₂S₃ counterparts are clearly seen, suggesting the successful deposition of Fe₃O₄, AgI, and Bi₂S₃ particles on the surface of g-C₃N₄.

Fig. 1 (a) XRD patterns for the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites with different weight percents of Bi₂S₃. (b) XRD pattern for the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite.

The EDX spectra of g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples were provided to further illustrate purity of the prepared samples. Fig. 2a shows the EDX spectra of the samples, which reveals that the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite possesses C, N, Fe, O, Ag, I, Bi, and S elements. Moreover, distributions of g-C₃N₄, Fe₃O₄, AgI, and Bi₂S₃ counterparts in the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite were analyzed through EDX mapping of C, N, Fe, O, Ag, I, Bi, and S elements, as seen in Fig. 2b–i. It can be observed that all of the elements are detectable and distributed evenly, confirming formation of heterojunction between counterparts of the nanocomposite. In addition, weight percents of the elements in the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite are 15.2, 21.8, 12.2, 7.20, 6.48, 6.83, 24.4, and 5.92% for C, N, Fe, O, Ag, I, Bi, and S elements, respectively. The theoretical weight percents of C, N, Fe, O, Ag, I, Bi, and S elements are 15.3, 21.8, 12.1, 7.24, 6.51, 6.82, 24.3, and 5.92%, respectively. Hence, as can be observed, the experimental and theoretical values of the weight percents are comparable. Based on the experimental values, weight percents of g-C₃N₄, Fe₃O₄, AgI, and Bi₂S₃ counterparts in the nanocomposite were calculated to be 36.7, 18.7, 14.2, and 30.4%, respectively.

Fig. 2 (a) EDX spectra for the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples. (b–i) EDX mapping of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite.

Fig. 3 shows the SEM and TEM images of the quaternary g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite. In this figure, aggregates of deposited particles are observed over the g-C₃N₄ sheet.

Fig. 3 (a) SEM and (b) TEM images of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite.

Fig. 4a displays UV-vis DRS spectra of the as-prepared samples. The pristine g-C₃N₄ has an absorption edge at about 470 nm. Due to low band gap of about 1.4 eV for Bi₂S₃, absorption edges of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites gradually shift toward longer wavelengths and the absorption intensities obviously increase in the visible region with the increase of Bi₂S₃ amount. The band gap energies (E_g) of the prepared samples were estimated using equation αhν = B(hν − E_g)^n/2, in which α, ν, and B are absorption coefficient, the light frequency, and proportionality constant, respectively.⁴³ The value of n depends on the characteristics of the transition in the semiconductor. It is well known that g-C₃N₄, AgI, and Bi₂S₃ are direct band gap semiconductors.^18,35,44 The E_g values for the prepared samples were estimated by extrapolation of the linear part of the curves obtained by plotting (αhν)² versus hν (Fig. 4b). It can be observed that the E_g values of samples are between 1.56 and 2.73 eV and band gap of the quaternary nanocomposites decreases with increasing weight percent of Bi₂S₃.

Fig. 4 (a) UV-vis DRS for the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites with different weight percents of Bi₂S₃. (b) Plots for calculation of band gap of the prepared samples. (c) Magnetization curves for the Fe₃O₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples.

Magnetization versus magnetic field curves (M–H) for the Fe₃O₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples were measured at room temperature using a VSM instrument with an applied field of 8500 Oe and the results are shown in Fig. 4c. The saturation magnetizations of these samples were measured to be 55.5, 20.6, 16.9, and 7.13 emu g⁻¹ for the Fe₃O₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples, respectively. In particular, it is evident from Fig. 4b that saturation magnetizations of these samples increased with an increase in Fe₃O₄ contents in the samples. The magnetic separability of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite was also examined by placing a magnet near the glass bottle. The grey particles were tightly attracted toward the magnet (inset of Fig. 4b), revealing that the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite possesses excellent magnetic separability. These results clearly state that the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites may be an excellent candidate as novel type of magnetically separable photocatalysts for potential environmental applications.

To further confirm composition of the as-prepared samples, FT-IR spectra were provided and the results are shown in Fig. 1S.† The strong IR absorption bands in the pure g-C₃N₄ sample revealed a typical molecular structure of g-C₃N₄.¹⁸ Several bands in the 1240–1650 cm⁻¹ region are corresponding to the typical stretching modes of CN heterocycles. Moreover, a peak at 806 cm⁻¹ is related to the out-of-plane breathing vibration of triazine units.¹⁸ For the Fe₃O₄ containing samples, the intense signal at 630 cm⁻¹ is assigned to vibration of the Fe–O bond.⁴⁰ Also, similar to the literature, there is not any peak for Ag–I bond.⁴⁰ For the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite, a new peak was appeared at 462 cm⁻¹, due to the presence of Bi₂S₃.⁴⁵

The amounts of g-C₃N₄ in the nanocomposites were estimated by TGA analysis. Fig. 2S† shows TGA curves of the g-C₃N₄, g-C₃N₄/Fe₃O₄, and quaternary g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites along with those of the AgI and Bi₂S₃ samples from room temperature to 700 °C in the air condition. For the pure g-C₃N₄ sample, a 97% of weight loss is observed. The deviation from 100% may be related to the baseline changes during the heating process. This weight loss is due to burning of g-C₃N₄ starts from nearly 530 °C to produce carbon dioxide and different oxides of nitrogen.^18,19 This implies that the g-C₃N₄ can be almost completely burned in air up to 700 °C. As can be seen, similar to many reports, pure g-C₃N₄ is stable up to 500 °C.¹⁸ It is evident that the AgI and Bi₂S₃ samples have considerable stability in the heating process. In the case of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) sample, the total weight loss is about 35% in the temperature range of 25–700 °C, indicating that this sample contains about 35% of g-C₃N₄. The actual loading amounts of g-C₃N₄ in the g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (10%), g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (40%) samples were estimated to be about 67, 45, 40, and 30 wt%, respectively. In addition, it is evident that all of the prepared samples are thermally stable up to 430 °C. Similar to many reports, stability of g-C₃N₄ decreases with loading different materials over that.^31,32

The textural properties of the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples were characterized using nitrogen adsorption–desorption experiments and the results are displayed in Fig. 5. It is evident that the samples have typical IV isotherms with H1 hysteresis, representing that the samples are mesoporous. The specific surface area and pore properties of the samples were calculated by BET and BJH models, respectively and the results are shown in Table 1. As can be seen, the specific surface area of the g-C₃N₄/Fe₃O₄ sample has increased compared with the g-C₃N₄, due to formation of the hierarchical structure. However, after loading of AgI and Bi₂S₃ particles on the g-C₃N₄/Fe₃O₄ nanocomposite, the specific surface area was decreased. This decrease may be ascribed to covering of the g-C₃N₄/Fe₃O₄ surface by AgI and Bi₂S₃ particles, resulting in blocking some parts of the active sites. As can be seen, there are not major differences between textural properties of the g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposites. Hence, textural properties cannot describe the highly enhanced photocatalytic activity of the quaternary nanocomposite relative to the ternary nanocomposites.

Fig. 5 Nitrogen adsorption–desorption data for the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples.

Table 1 The textural properties of g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples

Photocatalyst	Surface area (m^2 g^−1)	Mean pore diameter (nm)	Total pore volume (cm^3 g^−1)
g-C3N4	14.6	14.2	0.0500
g-C3N4/Fe3O4	29.4	21.6	0.159
g-C3N4/Fe3O4/AgI (20%)	16.0	14.5	0.0581
g-C3N4/Fe3O4/Bi2S3 (30%)	22.3	17.9	0.0998
g-C3N4/Fe3O4/AgI/Bi2S3 (30%)	14.9	19.6	0.0728

---

Photocatalytic activity of the prepared samples was investigated by degradation of RhB under visible-light irradiation and the results are depicted in Fig. 6a. As it can be observed, the pure g-C₃N₄ presents poor photocatalytic activity. For comparison, photocatalytic activities of the g-C₃N₄/Fe₃O₄/AgI (20%) and g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) nanocomposites were also tested under the same conditions, which can degrade 64 and 41% of RhB within 60 min, respectively. It is evident that all of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites exhibit higher photocatalytic activity than those of the g-C₃N₄/Fe₃O₄/AgI (20%) and g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) samples, indicating that a synergistic effect exists between AgI and Bi₂S₃ particles in enhancing photocatalytic activity. As can be seen, with the increase of Bi₂S₃ content, photocatalytic activity of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites exhibits a rise followed by a decline and the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) sample demonstrates the highest photocatalytic activity, which can degrade 99% of RhB after 60 min. By loading more amount of Bi₂S₃, they could destruct the heterojunction between the counterparts due to aggregation of its particles, leading to the decrease of the photocatalytic activity.

Fig. 6 (a) Photodegradation of RhB over the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites with different weight percents of Bi₂S₃. The UV-vis spectra for degradation of RhB under visible-light irradiation over the (b) g-C₃N₄, (c) g-C₃N₄/Fe₃O₄, (d) g-C₃N₄/Fe₃O₄/AgI (20%), (e) g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and (f) g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples.

Fig. 6b–f show spectral changes of RhB during the degradation reaction over the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples under visible-light irradiation. It is clear that intensity of the main absorption peak of RhB, located at 553 nm, decreases rapidly with increasing irradiation time and stable intermediates are not formed during the degradation reaction.^46,47 Moreover, the quaternary nanocomposite completely degrades RhB after 60 min and compared with the other samples, this photocatalyst exhibits superior activity.

As compared in Fig. 7a, the rate constant for degradation of RhB over the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite is 731.8 × 10⁻⁴ min⁻¹, which is significantly higher than those of the g-C₃N₄ (16.0 × 10⁻⁴ min⁻¹), g-C₃N₄/Fe₃O₄ (20.2 × 10⁻⁴ min⁻¹), g-C₃N₄/Fe₃O₄/AgI (20%) (139 × 10⁻⁴ min⁻¹), and g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) (54.9 × 10⁻⁴ min⁻¹) samples. Hence, activity of the quaternary nanocomposite is about 56, 44, 6.5, and 16-folds higher than those of the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) samples, respectively. It is well known that PL spectra can be applied to display recombination rate of photogenerated charge carriers. Herein, we conducted PL measurements to study the effect of Bi₂S₃ on recombination rate of the electron–hole pairs in the g-C₃N₄/Fe₃O₄/AgI (20%) nanocomposite. It can be seen that the g-C₃N₄ sample exhibits the strongest PL spectrum with an emission peak at about 435 nm (Fig. 7b). This strong peak is attributed to the band–band PL phenomenon with the energy of light approximately equal to the band gap energy of g-C₃N₄.¹⁸ Compared with the pristine g-C₃N₄, the PL intensities of the g-C₃N₄/Fe₃O₄ and g-C₃N₄/Fe₃O₄/AgI (20%) samples are weaker, reflecting their slow recombination rates of the charge carriers. The weakest intensity of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite suggests that the presence of Bi₂S₃ can sharply reduce recombination rate of the photogenerated charge carriers in the quaternary nanocomposite.

Fig. 7 (a) The degradation rate constants of RhB over different samples. (b) PL spectra for the g-C₃N₄, g-C₃N₄/Fe₃O₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples.

Preparation time of a photocatalyst is a very important parameter, affecting crystallinity, size, aggregation, and morphology of particles. Thus a series of experiments were conducted by preparation of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite at different refluxing times. Fig. 3S† demonstrates the preparation time dependence of the degradation rate constant of RhB over the quaternary nanocomposite under visible-light irradiation. As illustrated, the nanocomposite prepared by refluxing for 2 h has the highest photocatalytic activity and with more increasing of the preparation time, the degradation rate constant starts to decrease. In addition, Fig. 3S† shows SEM image of the nanocomposite prepared by refluxing for 4 h. As can be seen, the deposited particles over g-C₃N₄ considerably aggregated relative to the sample prepared by refluxing for 2 h (see Fig. 3a). Hence, by more aggregation of the deposited particles over g-C₃N₄, heterojunction formed between the counterparts were destructed, leading to decrease of the photogenerated electron–hole pair separations. Therefore, the nanocomposite prepared by refluxing for 2 h was selected for further investigations.

It is notable that calcination temperature of photocatalysts could remarkably affect on crystallization and size of particles. In this step, the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite prepared by refluxing for 120 min was calcined at 200, 300, 400, and 500 °C for 2 h. As can be seen in Fig. 3S,† the degradation rate constant of RhB over the nanocomposite decreases with an increase in the calcination temperature. Moreover, Fig. 3S† shows SEM image of the nanocomposite calcined at 500 °C. Compared with the uncalcined sample (Fig. 3a), it is evident that the deposited particles have considerably aggregated and formed bigger sizes.⁴⁸ The decrease of the rate constant is attributed to increase the aggregation of AgI and Bi₂S₃ particles deposited on the g-C₃N₄ sheet, resulting in destruction of heterojunction between counterparts of the nanocomposite. Hence, photogenerated electrons could not easily transfer from g-C₃N₄ to AgI and Bi₂S₃ particles, leading to decrease of the photocatalytic activity.

Depending on the reaction mechanism, a photocatalytic degradation reaction involves formation of various active species that can be different from those of the other reactions.^2,3 To understand the reaction mechanism, it is important to identify the active species participating in the degradation reaction. Using trapping experiments, the roles of holes (h⁺), hydroxyl radicals (˙OH), and superoxide ion radicals (˙O₂⁻) were identified using ammonium oxalate, 2-propanol, and benzoquinone, respectively. It can be seen in Fig. 4S† that the degradation rate constant significantly decreased upon addition of benzoquinone. On the contrary, the photocatalytic activity was slightly decreased when ammonium oxalate and 2-propanol were added to the reaction system. The above results indicate that superoxide ion radicals are the main reactive species in the degradation reaction of RhB over the quaternary nanocomposite.

The band structures and matching of the band energies are responsible for the efficient generation and separation of the charge carriers in photocatalysts. Hence, VB potentials of the counterparts were estimated by the following empirical equation: E_VB = X − E^e + 0.5E_g, where E_VB is the VB potential, X is the electronegativity of the semiconductor (the geometric mean of the electronegativity of the constituent atoms), E^e is the energy of free electrons on the hydrogen scale (∼4.5 eV), E_g is the band gap energy of the photocatalyst.⁴⁹ Moreover, the E_CB potentials were calculated by E_CB = E_VB − E_g formula and the results were tabulated (Table 2). Based on the experimental and calculation results, the plausible mechanism for the enhanced photocatalytic activity over the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites is presented in Fig. 8. It is well known that g-C₃N_4, AgI, and Bi₂S₃ are n-type semiconductors and their Fermi levels are close to their CB levels. Thus, after contacting these semiconductors to each other, n–n heterojunctions will be formed in the contacting regions.^50–54 Band gaps of g-C₃N₄, AgI, and Bi₂S₃ are 2.7, 2.8, and 1.4 eV, respectively. As a result, they are simply activated under visible-light irradiation to produce electron–hole pairs. The photogenerated electrons transfer from the CB of AgI and Bi₂S₃ to that of g-C₃N₄ under the driving force of the produced electrostatic fields.^11,13 As can be seen in Table 1, the CB potential of g-C₃N₄ is more negative than the potential of O₂/˙O₂⁻ (E⁰(O₂/˙O₂⁻) = −0.33 eV/NHE),⁵⁵ hence, photogenerated electron on g-C₃N₄ react with adsorbed O₂ to form ˙O₂⁻ species, as a major reactive species participating in the degradation reaction. Furthermore, the VB energies for AgI and Bi₂S₃ are less positive than that of g-C₃N₄. As a consequence, the photogenerated holes transfer from the VB of g-C₃N₄ to those of the AgI and Bi₂S₃. Hence, the photogenerated electrons and holes are gathered on the CB of g-C₃N₄ and VB of AgI and Bi₂S₃, respectively. Therefore, due to formation of tandem n–n heterojunctions between these semiconductors, the charge carriers are effectively separated from each other, resulting in enhanced photocatalytic activity in the quaternary nanocomposites.^50–54

Table 2 The VB, CB, and E_g values for counterparts of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites

Semiconductor	VB energy (eV)	CB energy (eV)	Eg (eV)
g-C3N4	+1.58	−1.12	2.70
AgI	+2.38	−0.42	2.80
Bi2S3	+1.50	+0.10	1.40

---

Fig. 8 The mechanism for the enhanced photocatalytic activity of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites.

Additionally, another two dyes (MB and fuchsine) were chosen as target pollutants to further evaluate photocatalytic activity of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite relative to its counterparts. As can be seen in Fig. 9a and b, similar to RhB degradation, the quaternary nanocomposite exhibited the highest photocatalytic activity in degradations of MB and fuchsine. The degradation rate constants of MB over the g-C₃N_4, g-C₃N₄/Fe₃O₄/AgI, and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples are 19.2 × 10⁻⁴, 44.5 × 10⁻⁴, and 83.5 × 10⁻⁴ min⁻¹, respectively. Also, the rate constants for degradation of fuchsine over the g-C₃N₄, g-C₃N₄/Fe₃O₄/AgI, and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples are 20.8 × 10⁻⁴, 27.6 × 10⁻⁴, and 172.4 × 10⁻⁴ min⁻¹, respectively. Therefore, it is evident that activities of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite in degradation of MB and fuchsine are about 4.3 and 8.2-folds greater than those of the g-C₃N₄.

Fig. 9 The degradation rate constants of (a) MB and (b) fuchsine over the g-C₃N₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) samples.

In addition to photocatalytic activity, the stability of a photocatalyst is also very important parameter in widespread practical applications. The results shown in Fig. 10 revealed that the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite did not exhibit significant loss of its activity after four runs in degradation of RhB. Only a slight decrease was observed, which can be attributed to adsorption of degradation intermediates over the surface of the photocatalyst during the degradation process. Hence, this work showed that the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites could be an excellent candidate as novel type of magnetically separable photocatalysts for potential environmental applications.

Fig. 10 Reusability of the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite for four successive runs.

In order to investigate the effect of solution pH on the degradation reaction of RhB over the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ (30%) nanocomposite, a series of experiments were carried out by changing initial pH of the solution between 2 and 10 by adding HCl and NaOH and the results are shown in Fig. 5S.† As can be seen, the degradation rate constant enhances with increasing pH of solution. Very recently, it was found that isoelectric point of g-C₃N₄ is nearly in solutions with pH = 5.⁵⁵ Consequently, in solutions with pH < 5, surface of g-C₃N₄ is positively charged due to protonation of its amine groups.⁵⁶ On the other hand, RhB is a cationic pollutant. Hence, in solutions with low pH, electrostatic repulsion between the pollutant and the catalyst restricts adsorption of RhB. As a result, the degradation rate constant decreases in acidic solutions. However, in alkaline solutions, surface of g-C₃N₄ is negatively charged through reaction with hydroxide ions.⁵⁵ Therefore, the positively charged pollutant is easily adsorbed over the catalyst in solutions with high pH, resulting in enhanced activity in these solutions.

Conclusions

Quaternary g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites, as novel magnetically separable visible-light-driven photocatalysts, were synthesized via a simple method for the first time and their physicochemical properties were systematically characterized using different techniques. The optimal content of Bi₂S₃ in the g-C₃N₄/Fe₃O₄/AgI/Bi₂S₃ nanocomposites was found to be 30 wt% and the optimal refluxing time was 120 min. This nanocomposite showed 56, 6.5, and 16-folds enhanced activity in degradation of RhB compared to the g-C₃N₄, g-C₃N₄/Fe₃O₄/AgI (20%), and g-C₃N₄/Fe₃O₄/Bi₂S₃ (30%) samples, respectively. Moreover, it was found that superoxide ions played a major role, whereas hydroxyl radicals and holes had minor roles. The recycling of the photocatalyst was also investigated and the activity had little decrease after four cycles. Due to the presence of narrow band gap semiconductor and formation of tandem n–n heterojunctions, the results demonstrated that the introduction of Bi₂S₃ could modify the visible-light absorption ability and separation of the charge carriers in the ternary g-C₃N₄/Fe₃O₄/AgI photocatalyst.

Acknowledgements

The authors wish to acknowledge University of Mohaghegh Ardabili-Iran, for financial support of this work.

References

1. T. Reemtsma and M. Jekel, Organic Pollutants in the Water Cycle: Properties, Occurrence, Analysis and Environmental Relevance of Polar Compounds, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2006.
2. V. Etacheri, C. D. Valentin, J. Schneider, D. Bahnemann and S. C. Pillai, J. Photochem. Photobiol., C, 2015, 25, 1.
3. S. Dong, J. Feng, M. Fan, Y. Pi, L. Hu, X. Han, M. Liu, J. Sun and J. Sun, RSC Adv., 2015, 5, 14610.
4. M. Pelaez, N. T. Nolan, S. C. Pillai, M. K. Seery, P. Falaras, A. G. Kontos, P. S. M. Dunlop, J. W. J. Hamilton, J. A. Byrne, K. O'Shea, M. H. Entezari and D. D. Dionysiou, Appl. Catal., B, 2012, 125, 331.
5. R. Fagan, D. E. McCormack, D. D. Dionysiou and S. C. Pillai, Mater. Sci. Semicond. Process., 2016, 42, 2.
6. K. M. Lee, C. W. Lai, K. S. Ngai and J. C. Juan, Water Res., 2016, 88, 428.
7. M. Pirhashemi and A. Habibi-Yangjeh, Appl. Surf. Sci., 2013, 283, 1080.
8. S. Khanchandani, P. K. Srivastava, S. Kumar, S. Ghosh and A. K. Ganguli, Inorg. Chem., 2014, 53, 8902.
9. S. Pal, S. Maiti, U. N. Maiti and K. K. Chattopadhyay, CrystEngComm, 2015, 17, 1464.
10. S. Shaker-Agjekandy and A. Habibi-Yangjeh, Mater. Sci. Semicond. Process., 2015, 34, 74.
11. F. Kiantazh and A. Habibi-Yangjeh, Mater. Sci. Semicond. Process., 2015, 39, 671.
12. C. Dong, X. Xiao, G. Chen, H. Guan and Y. Wang, Mater. Chem. Phys., 2015, 155, 1.
13. M. Pirhashemi and A. Habibi-Yangjeh, J. Mater. Sci.: Mater. Electron., 2016, 27, 4098.
14. H.-P. Lin, W. W. Lee, S.-T. Huang, L.-W. Chen, T.-W. Yeh, J.-Y. Fu and C.-C. Chen, J. Mol. Catal. A: Chem., 2016, 417, 168.
15. X. Lv, J. Wang, Z. Yan, D. Jiang and J. Liu, J. Mol. Catal. A: Chem., 2016, 418, 146.
16. W. Zhao, Y. Liu, Z. Wei, S. Yang, H. He and C. Sun, Appl. Catal., B, 2016, 185, 242.
17. S. Cao, J. Low, J. Yu and M. Jaroniec, Adv. Mater., 2015, 27, 2150.
18. Y. Zheng, J. Liu, J. Liang, M. Jaroniec and S. Z. Qiao, Energy Environ. Sci., 2012, 5, 6717.
19. G. Dong, Y. Zhang, Q. Pan and J. Qiu, J. Photochem. Photobiol., C, 2014, 20, 33.
20. J. Zhu, P. Xiao, H. Li and S. A. C. Carabineiro, ACS Appl. Mater. Interfaces, 2014, 6, 16449.
21. Y. Gong, M. Li, H. Li and Y. Wang, Green Chem., 2015, 17, 715.
22. S. Ye, R. Wang, M.-Z. Wu and Y.-P. Yuan, Appl. Surf. Sci., 2015, 358, 15.
23. Z. Zhao, Y. Sun and F. Dong, Nanoscale, 2015, 7, 15.
24. C. Xing, Z. Wu, D. Jiang and M. Chen, J. Colloid Interface Sci., 2014, 433, 9.
25. H. Xu, Y. Song, Y. Song, J. Zhu, T. Zhu, C. Liu, D. Zhao, Q. Zhang and H. Li, RSC Adv., 2014, 4, 34539.
26. X. Rong, F. Qiu, J. Yan, H. Zhao, X. Zhu and D. Yang, RSC Adv., 2015, 5, 24944.
27. H. Xu, H. Zhao, Y. Song, W. Yan, Y. Xu, H. Li, L. Huang, S. Yin, Y. Li, Q. Zhang and H. Li, Mater. Sci. Semicond. Process., 2015, 39, 726.
28. N. Wang, Y. Zhou, C. Chen, L. Cheng and H. Ding, Catal. Commun., 2016, 73, 74.
29. D. Ma, J. Wu, M. Gao, Y. Xin, T. Ma and Y. Sun, Chem. Eng. J., 2016, 290, 136.
30. K. Vignesh, A. Suganthi, B.-K. Min and M. Kang, J. Mol. Catal. A: Chem., 2014, 395, 373.
31. M. Mousavi and A. Habibi-Yangjeh, Mater. Chem. Phys., 2015, 163, 421.
32. A. Habibi-Yangjeh and A. Akhundi, J. Mol. Catal. A: Chem., 2016, 415, 122.
33. M. Gao, D. Zhang, X. Pu, H. Li, D. Lv, B. Zhang and X. Shao, Sep. Purif. Technol., 2015, 154, 211.
34. C. Li, G. Chen, J. Sun, J. Rao, Z. Han, Y. Hu, W. Xing and C. Zhang, Appl. Catal., B, 2016, 188, 39.
35. R. P. Panmand, Y. A. Sethi, R. S. Deokar, D. J. Late, H. M. Gholap, J.-Q. Baeg and B. B. Kale, RSC Adv., 2016, 6, 23508.
36. J. Zhang, L. Zhang, N. Yu, K. Xu, S. Li, H. Wang and J. Liu, RSC Adv., 2015, 5, 75081.
37. W. Xu, J. Fang, Y. Chen, S. Lu, G. Zhou, X. Zhu and Z. Fang, Mater. Chem. Phys., 2015, 154, 30.
38. T. J. Entradas, J. F. Cabrita, B. Barrocas, M. R. Nunes, A. J. Silvestre and O. C. Monteiro, Mater. Res. Bull., 2015, 72, 20.
39. N. Qin, Y. Liu, W. Wu, L. Shen, X. Chen, Z. Li and L. Wu, Langmuir, 2015, 31, 1203.
40. A. Akhundi and A. Habibi-Yangjeh, Mater. Chem. Phys., 2016, 174, 59.
41. M. Shekofteh-Gohari and A. Habibi-Yangjeh, Ceram. Int., 2015, 41, 1467.
42. Z. Xuan, Y. Shiyue, L. Yumei, W. Zuoshan and L. Weifeng, Mater. Lett., 2015, 145, 23.
43. X. Li and J. Ye, J. Phys. Chem. C, 2007, 111, 13109.
44. R. H. Victora, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 56, 4417.
45. T. K. Patil and M. I. Talele, Adv. Appl. Sci. Res., 2012, 3, 1702.
46. X. Li and J. Ye, J. Phys. Chem. C, 2007, 111, 13109.
47. A. Martinez-dela Cruz and U. M. Garcia Perez, Mater. Res. Bull., 2010, 45, 135.
48. M. Shekofteh-Gohari and A. Habibi-Yangjeh, RSC Adv., 2016, 6, 2402.
49. S. R. Morrison, Electrochemistry at semiconductor and oxidized metal electrode, Plenum, New York, 1980.
50. Y. Yan, H. Guan, S. Liu and R. Jiang, Ceram. Int., 2014, 40, 9095.
51. M. Xu, T. Ye, F. Dai, J. Yang, J. Shen, Q. He, W. Chen, N. Liang, J. Zai and X. Qian, ChemSusChem, 2015, 8, 1218.
52. Y.-K. Hsu, Y.-C. Chen and Y.-G. Lin, ACS Appl. Mater. Interfaces, 2015, 7, 14157.
53. M. Pirhashemi and A. Habibi-Yangjeh, Ceram. Int., 2015, 41, 14383.
54. M. Shekofteh-Gohari and A. Habibi-Yangjeh, J. Ind. Eng. Chem., 2016, 44, 174.
55. B. Zhu, P. Xia, W. Ho and J. Yu, Appl. Surf. Sci., 2015, 344, 188.
56. Y.-P. Sun, W. Ha, J. Chen, H.-Y. Qi and Y.-P. Shi, TrAC, Trends Anal. Chem., 2016, 84, 12.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra12414c

---