Photochemical reactions of g-C₃N₄-based heterostructured composites in Rhodamine B degradation under visible light

Abstract

Highly efficient visible-light-driven Ag₂O/g-C₃N₄, Ag/g-C₃N₄, and BiOBr/g-C₃N₄ heterostructured photocatalysts were prepared by solution synthesis methods at room temperature. Compared with Ag/g-C₃N₄ (1 : 30 mass ratio) and Ag₂O/g-C₃N₄ (2 : 1 mass ratio) photocatalysts, the BiOBr/g-C₃N₄ composite (1 : 3) displayed enhanced photocatalytic activities for Rhodamine B (RhB) degradation under visible-light irradiation. The reaction kinetics of phenol RhB dye photodegradation of Ag₂O/g-C₃N₄ (1 : 1, 2 : 1, 4 : 1 mass ratios) and BiOBr/g-C₃N₄ were fitted with the pseudo-first-order model, ln(C₀/C) = kt, while the data of Ag/g-C₃N₄ and pure g-C₃N₄, Ag₂O/g-C₃N₄ (1 : 20 mass ratio) were fitted with the zero-order model. The former has enhanced photocatalytic activity, and BiOBr/g-C₃N₄ composites exhibit the highest degrading rate for RhB. The enhanced photocatalytic activity of two dimensional BiOBr/g-C₃N₄ photocatalysts was mainly attributed to the formation of type II p–n heterojunctions, as well as the well-matched band gap and the synergetic effects between two components, which accelerate the separation efficiency of photogenerated electrons–holes at the interface. Finally, possible photocatalytic and charge separation mechanisms of Ag₂O/g-C₃N₄ and BiOBr/g-C₃N₄ composites were proposed via active species capture experiments.

---

Introduction

Semiconductor photocatalysis has emerged as one of the most promising technologies for environmental remediation and solar energy conversion.¹ The growing concerns about visible-light-driven photocatalysts have stimulated extensive studies. Of the well-known semiconductor photocatalysts, TiO₂ has attracted more attention owing to its stability, nontoxicity, and low-cost.² However, because of relatively wide band gap and fast recombination of the photo-generated electron–hole pairs, the enhancement of the visible-light photocatalytic activity is still very limited.³ Therefore, it is still a great challenge and highly crucial to explore novel photocatalytic materials with visible light response wavelength range.

Graphitic carbon nitride (g-C₃N₄), a polymeric metal-free semiconductor has been recently focused on the visible-light photocatalytic field with the advantages of good stability and an appealing electronic structure with a medium band gap of 2.7 eV,⁴ which fulfills the basic requirements as a photocatalyst for water splitting and/or organic pollutant decomposition. g-C₃N₄ can easily be synthesized via direct polymerization of nitrogen-containing such as cyanamide, dicyandiamide, melamine, urea, and thiourea. Nevertheless, the low quantum efficiency, and the fast recombination of photogenerated electron–hole pairs, the lack of absorption above 460 nm may still restrict the photocatalytic activities of pure g-C₃N₄.⁵ In order to overcome these problems and improve the photochemical reactivity of g-C₃N₄, many strategies have been investigated to improve the photocatalytic activity of g-C₃N₄, such as transition metal doping, noble metals modification, and heterostructured nanocomposites.^6,14 Heterostructured composites have become one of the hottest research topics. Various g-C₃N₄ based hybrid photocatalysts including g-C₃N₄/Bi₂WO₆,⁶ g-C₃N₄/Co₃O₄,⁷ g-C₃N₄/Fe₂O₃,⁸ and g-C₃N₄/NiS⁹ have been developed to further extend the visible light absorption range and photogenerated carrier separation efficiency. Recently, Ag-based photocatalysts attracts growing attention which can be used to combine with g-C₃N₄ to promote the separation of photogenerated carriers and improve their photocatalytic activities. Xu et al. prepared AgX/g-C₃N₄ (X = Br and I) composites with high photocatalytic activities in methyl orange (MO) degradation.¹⁰ The preparation of a series of ternary Ag/Ag₃PO₄/g-C₃N₄ hybrid photocatalysts, which display enhanced photocatalytic activity, was reported.¹¹ These studies clearly indicated the enhanced photocatalytic activities by incorporating Ag-based materials with g-C₃N₄, and showed superior photocatalytic activities to degrade pollutants under visible-light irradiation. However, AgX and Ag₃PO₄ photocatalysts are unstable under light illumination, thus, it is hard to precise control the mass ratios between the components. Ag₂O, as a p-type semiconductor with a narrow band gap of 1.46 eV, has been found to be a self-stable and highly efficient photocatalyst under visible light.¹² It has been widely used as an efficient sensitizer to tune the light response of wide-band-gap semiconductors into the visible region and improve their photocatalytic activities, such as Ag₂O/TiO₂.¹³ The combination of g-C₃N₄ and Ag₂O that possesses well matched band structure can easily fabricate a p–n heterojunction, which will bring more effective interface transfer of photogenerated electrons and holes in comparison with pure Ag₂O and g-C₃N₄. Moreover, semiconductor photocatalysts modified by pure noble metal nanoparticles could also exhibit enhanced photocatalytic activities, such as Au/g-C₃N₄.¹⁴ Therefore, it is of vital importance to explore Ag-based photocatalysts with visible-light response. As we all known, the studies on the Ag₂O/g-C₃N₄ system are plentiful, but no study has compared Ag₂O/g-C₃N₄ with Ag/g-C₃N₄ under visible light.

Bismuth oxyhalides (BiOX, X = Cl, Br, I), new candidates of prospective photocatalysts, have exhibited enhanced photocatalytic activities in degrading organic dyes compared to common photocatalyst such as TiO₂ and ZnO as a result of their intrinsic lamellar structures.¹⁵ Although pure BiOBr materials arouse great interest for its visible light response and high stability, as well as narrow band gap, it is still necessary to construct heterostructured materials to further improve its photocatalytic activity for practical applications as a result of its rapid combination of photogenerated charge carriers of pure BiOBr. Visible-light-responsive BiOBr–ZnFe₂O₄ hetero-junction photocatalysts were prepared by a precipitation deposition method which exhibited better photocatalytic activity compared with pure BiOBr for the degradation of Rhodamine B (RhB).¹⁶ Moreover, both BiOBr/BiWO₃ (ref. 17) and BiOBr/Bi₂O₃ (ref. 18) composites showed enhanced photocatalytic activities. Due to well matched band structure, the combination of g-C₃N₄ and BiOBr can be easily fabricated into two dimensional (2D) p–n heterojunction composites, which will bring more effective separation and transfer of photogenerated charges.

Herein, highly efficient visible-light-driven heterostructured photocatalysts including Ag/g-C₃N₄, Ag₂O/g-C₃N₄, and BiOBr/g-C₃N₄ were prepared by solution reactions at room temperature, which could be quantitatively found out the optimal photocatalysts. Ag₂O/g-C₃N₄ and BiOBr/g-C₃N₄ with p–n heterostructured photocatalysts revealed high photo-degradation efficiency for RhB compared with Ag/g-C₃N₄, all of which possessed ideal photocatalytic activity and stability under visible light irradiation compared with the pure g-C₃N₄. The reaction kinetics of phenol RhB dye photodegradation of Ag₂O/g-C₃N₄ with mass ratios of 1 : 1, 2 : 1, 4 : 1 and BiOBr/g-C₃N₄ composites were fitted with the pseudo-first-order model, ln(C₀/C) = kt, while the experimental data of Ag/g-C₃N₄, Ag₂O/g-C₃N₄ with a mass ratio of 1 : 20 and pure g-C₃N₄ were fitted with the zero-order kinetics model. We then conclude that the as-prepared samples which meet the pseudo-first-order model are optimal photocatalysts. Meanwhile, we also proposed possible mechanisms for the enhanced activity of Ag₂O/g-C₃N₄, BiOBr/g-C₃N₄ composites on account of the experimental results.

Experimental

Chemicals

Melamine (C₃H₆N₆), p-benzoquinone (PBQ), triethanolamine (TEA), methanol, silver nitrate (AgNO₃), potassium bromide (KBr) and sodium hydroxide (NaOH) were obtained from Sinopharm Chemical Reagent Co., Ltd, China. Bismuth nitrate was purchased from Kermal, Tianjin. All chemicals were used as received without further purification, and all aqueous solutions were prepared with ultrapure water (∼18.25 MΩ cm) obtained from a Milli-Q synthesis system.

Instrumentation

The crystal structures of samples were identified by an X-ray diffraction (XRD) meter (Bruker D8-Advance, Germany). Photocatalytic activity was assessed by the degradation of RhB dye photodegradation under visible light irradiation (18 W λ ≥ 400 nm). UV-visible absorption spectra were obtained through the diffuse reflection method using a spectrometer (HITACHI U-4100, Japan Hitachi). The morphologies of the samples were investigated by scanning electron microscopy (SEM) (QUANTA 250 FEG).

Synthesis of samples

g-C₃N₄ powders were synthesized according to the method in literature.¹⁹ Typically, 3 g of melamine was put into a quartz boat and heated in a rate of 5 °C min⁻¹ to 550 °C in a tube furnace and then kept at this temperature for 3 h. All the experiments were performed in air conditions. The resulting yellow product was collected and ground into powder for further use. The typical preparation procedure of Ag₂O/g-C₃N₄ composite photocatalysts was as follows: 0.05 g of g-C₃N₄ was added into 25 mL of H₂O and sonicated with stirring for 2 h. Then, 0.25 g of NaOH was added into the suspension and stirred for 30 min. Further, different amounts of 0.1 M AgNO₃ were added drop by drop under stirring, and the mixture was held in the dark for 30 min with continuous stirring. All the experiments were carried out at room temperature. The obtained precipitate was collected by centrifugation and washed with distilled water for four times. Finally, the solid product was dried at 60 °C overnight. A series of Ag₂O/g-C₃N₄ composites with different mass ratios of Ag₂O and g-C₃N₄ were prepared by changing the amounts of g-C₃N₄ and marked as 4 : 1, 2 : 1, 1 : 1 and 1 : 20. As a reference, Ag/g-C₃N₄ composite photocatalyst was prepared using ethanol instead of H₂O and kept other conditions same. A series of Ag/g-C₃N₄ composites with different mass ratios of Ag and g-C₃N₄ were marked as 1 : 1, 1 : 10, 1 : 30.

For preparing BiOBr/g-C₃N₄ heterojunctions, typically, 0.05 g of g-C₃N₄ was added into 20 mL of H₂O followed by addition of 0.081 g of KBr, and the mixture was ultrasonicated for 1 h to obtain a uniform suspension (suspension A). 0.08 g of Bi(NO₃)₃·5H₂O was dissolved in 10 mL of aqueous solution containing 1 mL of acetic acid (HAc) and vigorously stirred for 30 min (solution B). The solution B was added dropwise into the suspension A and stirred magnetically for 1 h at room temperature. After aging for 6 h, the precipitates were collected by centrifugation, washed thoroughly with distilled water and ethanol, and then dried at 60 °C overnight to get the resulting samples. The different weight ratios of BiOBr/g-C₃N₄ at 1 : 2, 1 : 3 and 1 : 4 were obtained. The pure BiOBr was synthesized using same conditions in the absence of g-C₃N₄.

Photo-catalytic investigation

The photo-catalytic activity of the catalyst was tested by the decomposition of RhB solution. A RhB solution with the concentration of 10 mg L⁻¹ (pH = 5.25) was first prepared. In each test, about 0.01 g of catalyst was added to ∼25 mL of the RhB solution and sonicated for 10 min. Then, the suspension was magnetically stirred for 30 min achieving an adsorption–desorption balance. The degradation process was performed under the UV illumination. At the appropriate reaction time (15 min, 30 min, 1 h, 2 h, 3 h), the 2 mL liquid was taken out. Then, the supernate were isolated from the solution by centrifugation. The concentration of the RhB was determined by monitoring the changes in the maximal absorbance at approximately λ = 554 nm characterized by an UV-vis spectrometer.

The active species capture experiments were employed to study the photocatalysis mechanism. At the beginning, 10.0 mM CH₃OH (˙OH quencher), TEA (h⁺ quencher) and PBQ (˙O₂⁻ quencher) were added to the RhB aqueous solution.^20,21 Then the remaining experimental processes were similar to that of photocatalysis.

Results and discussion

The composition and morphology of as-prepared catalysts were characterized by XRD and SEM analysis. As shown in Fig. 1a, the XRD patterns illustrate the crystalline phase evolution of the as-prepared Ag₂O/g-C₃N₄ composites with different mass ratios. An apparent diffraction peak at 27.4° in pure g-C₃N₄ were indexed to the (002) planes of hexagonal g-C₃N₄ (JCPDS 87-1526) arising from the stacking of the conjugated aromatic system corresponding to an interlayer distance of d = 0.33 nm, while the peak at 13.1° represents in-plane structural packing motif with a period of 0.675 nm.²² In the Ag₂O/g-C₃N₄ composite, the diffraction peaks except at 2θ of 27.4° were attributed to the cubic crystal phase (JCPDF 41-1104),⁵ and a weak diffraction peak of g-C₃N₄ at 27.4° was found out owing to the relatively low percentage of g-C₃N₄ (Fig. 1a). With the growth of g-C₃N₄ content (>50%), the weak peak of Ag₂O at 26.7° may be covered by the strong peak of g-C₃N₄ at 27.4° and cannot be observed in the XRD patterns. The XRD patterns of the as-prepared Ag/g-C₃N₄ composites with the mass ratio of 1 : 30 are shown in Fig. 1a, the diffraction peaks at 2θ of 38.1°, 44.2°, 64.4° and 77.4° are attributed to the respective (111), (200), (220) and (311) planes of the crystal phase of Ag (JCPDF 04-0783).²³ Fig. 1b shows the XRD patterns of the BiOBr and BiOBr/g-C₃N₄ heterostructured composites. The typical diffraction peaks at 2θ of 10.9°, 21.9°, 25.1°, 31.7°, 32.2°, 39.4°, 46.2° and 57.1° in BiOBr/g-C₃N₄ sample are ascribed to (001), (002), (101), (102), (110), (112), (200) and (212) facets which can be indexed to the tetragonal phase of BiOBr (JCPDF no. 09-0393) except a typical diffraction peak of g-C₃N at 27.4°.

Fig. 1 XRD patterns of as-prepared samples: g-C₃N₄, Ag₂O/g-C₃N₄, Ag/g-C₃N₄ (a) and BiOBr/g-C₃N₄. (★: g-C₃N₄) composites (b).

After coupling with BiOBr, the intensity of dominate (002) peak in g-C₃N₄ and (001) peak in BiOBr decreased significantly in the composites compared with the separate sample. This suggests that BiOBr has been successfully deposited onto the surface of g-C₃N₄. The (001) peak of BiOBr in the composite was much wider than that of pure BiOBr, suggesting a stepwise thickness decrease of BiOBr in the heterojunctions, which implies that the crystallite growth of BiOBr along the c-axis would be attenuated after being deposited onto g-C₃N₄. The results above show that either p–n heterostructure composites or noble metal modified composites are successfully obtained and all exhibit good crystallization.

Fig. 2 shows the SEM images of Ag₂O/g-C₃N₄ and the BiOBr/g-C₃N₄ photocatalysts with the mass ratio of 1 : 2 (Fig. 2a) and 1 : 3 (Fig. 2b). The as-prepared g-C₃N₄ sample revealed a clearly lamellar structure. After the introduction of Ag₂O, many Ag₂O particles were evenly dispersed on the surface of g-C₃N₄. With the mass ratio of g-C₃N₄ and Ag₂O varying from 1 : 2 to 20 : 1, the average size of Ag₂O nanoparticles decreased from about 32 nm to 10 nm. The agglomeration can be emerged in pure Ag₂O. The addition of g-C₃N₄ can greatly improve the dispersibility of Ag₂O particles and efficiently decrease the particle size of Ag₂O. It may be attributed that g-C₃N₄ with a layer structure offered numerous nucleation sites to the growth of Ag₂O particles, leading to the homogeneous dispersion of Ag₂O particles on the surface of g-C₃N₄ with smaller size. All the Ag₂O particles are attached to the surface of g-C₃N₄ so that strong interface interaction could be formed between Ag₂O particles and g-C₃N₄. As reported in literature, the interplanar spacing of the Ag₂O/g-C₃N₄ composites prepared by a simple liquid phase reaction was 0.27 nm corresponded to the (111) plane of Ag₂O.²⁰ Fig. 2b clearly shows that there exist two kinds of nanosheets. The flat nanosheets can be assigned to BiOBr and the other small-size ones can be indexed to g-C₃N₄. On the basis of the crystal structure of BiOBr revealed by XRD patterns, the (BiO)⁺ layers were interlaid by double slabs of Br⁻. Such internal structure would control the growth of BiOBr at a certain axis direction to form 2D nanosheet morphology. The BiOBr nanosheets were deposited on the surface of g-C₃N₄ by situ self-assembled process, resulting in formation of strong interaction between BiOBr and g-C₃N₄. The large and closely contacted interface in Ag₂O/g-C₃N₄ and the BiOBr/g-C₃N₄ composites is favorable for charge transfer between two components.

Fig. 2 SEM images of Ag₂O/g-C₃N₄ (1 : 2) (a) and BiOBr/g-C₃N₄ (1 : 3) (b) composite.

Scheme 1 shows the formation process of Ag₂O/g-C₃N₄. Bulk graphitic carbon nitride was synthesized from melamine via the process of high temperature polymerization. After adding NaOH solution, OH⁻ ions can be attached to the surface of g-C₃N₄. Ag⁺ ions with positive charges react with hydroxyl ions to form Ag(OH) precipitates in an aqueous solvent. Ag(OH) precipitates are unstable prone to decomposition to generate Ag₂O which deposited on the surface of g-C₃N₄. However, Ag nanoparticles were dispersed on the surface of g-C₃N₄ when ethanol was used as solvent instead of aqueous solvent. Silver ions react with hydroxyl ions and ethanol to produce elemental silver under alkaline conditions. The reason for two different composites is that hydroxyl groups of ethanol exhibit weak reduction properties which could reduce silver ions into elemental silver.

Scheme 1 Formation process of Ag₂O/g-C₃N₄.

The schematic illustration of the formation process for 2D BiOBr/g-C₃N₄ heterostructure nanocomposites is presented in Scheme 2. In g-C₃N₄ aqueous solution, a large number of bromine ions are adsorbed on the surface of g-C₃N₄ nanosheets due to the electrostatic interactions. After adding bismuth source solution, (BiO)⁺ ion ionized by the Bi(NO₃)₃ solution rapidly combines with Bi⁻ to form BiOBr nanosheets and the nanosheets grow along the surface of g-C₃N₄ nanosheets, which results in the formation of BiOBr/g-C₃N₄ 2D nanojunctions. On account of lamellar structure of g-C₃N₄, Br⁻ is adsorbed on the (002) crystal plane.²⁴ By means of in situ growth process, BiOBr/g-C₃N₄ heterostructure composites have stacked layers of laminated structure through self-assembly. Basing on the above mechanisms, there is strong interface interaction existed between two components in heterostructure photocatalysts, which brings more effective interface transfer of photogenerated carriers.

Scheme 2 Formation process of BiOBr/g-C₃N₄.

The optical properties of the as-prepared g-C₃N₄, Ag₂O/g-C₃N₄, Ag/g-C₃N₄ and the BiOBr/g-C₃N₄ composites were measured via the UV-vis diffuse reflection technique as shown in Fig. 3. Fig. 3a shows the absorption edge of the pure g-C₃N₄ is at about 460 nm, which originates from its band gap of 2.7 eV and is consistent with the reported results.⁴ After the introduction of Ag₂O or Ag nanoparticles on the surface of g-C₃N₄, the optical absorption of the composites in the visible region increases. With the increasing mass ratios of Ag₂O and g-C₃N₄, the absorption intensity of these composites is strengthened. We could calculate the band edge according to the equation: ahν = A(αhν − E_g)^n/2.²⁵ Among all the samples of Ag₂O/g-C₃N₄, the band gap of the composite with the mass ratio of 2 : 1 is the narrowest with about 2.2 eV. The band gap of Ag/g-C₃N₄ composites with the increasing mass ratio of 1 : 30 is about 2.6 eV. As expected, the BiOBr sample shows its fundamental absorption edge rising at 430 nm, consistent with the reported results. Fig. 3c shows that the BiOBr/g-C₃N₄ exhibits a gradual red shift of band edge as the amount of g-C₃N₄ increased in comparison with BiOBr, which favors the absorption of visible light. The results from UV-vis absorption spectra suggest that the fabrication of the heterostructured composites can greatly improve the optical absorption property and increase the utilized efficiency of solar light, which are favorable for the enhancement of the photocatalytic activity.

Fig. 3 Diffuse reflection spectra of Ag₂O/g-C₃N₄ (2 : 1), g-C₃N₄ and Ag/g-C₃N₄ (1 : 30) composites (a) and the corresponding the band gaps (b); diffuse reflection spectra of BiOBr, BiOBr/g-C₃N₄ composites (c).

The photocatalytic activities of the as-prepared Ag₂O/g-C₃N₄, Ag/g-C₃N₄ and BiOBr/g-C₃N₄ composites with various mass ratios were evaluated by the photodegradation of RhB dye under visible light irradiation shown in Fig. 4. For comparison, the photocatalytic activities of pure g-C₃N₄, Ag₂O and BiOBr were also obtained under the same conditions. The removal percentage of RhB dye was nearly 29% by pure g-C₃N₄. According to the tendency of degradation, to quantitatively determine the reaction kinetics of RhB dye photodegradation by the as-prepared samples, the experimental data of Ag₂O/g-C₃N₄ (1 : 1, 2 : 1, 4 : 1), pure BiOBr and BiOBr/g-C₃N₄ composites were fitted with the pseudo-first-order model, ln(C₀/C) = kt, where k was the apparent first-order rate constant, C₀ is the initial concentration of RhB solution, C is the concentration of RhB solution at time t, and k is kinetic constant.²⁵ The corresponding kinetic constants (k) were calculated and given in Fig. 4b, d and f. The experimental data of Ag/g-C₃N₄, Ag₂O/g-C₃N₄ (1 : 20) and pure g-C₃N₄ were fitted with the zero-order kinetics model, which may attributed to the extremely high content of g-C₃N₄ in composite systems. The kinetic model for degradation of composites tend to be fitted with first-order model as the less content of g-C₃N₄, which is close to the reaction kinetics rules of photodegradating RhB by pure Ag₂O or BiOBr photocatalyst. In zero-order kinetics reaction, the degradation rates have nothing to do with concentration of g-C₃N₄ and Ag₂O/g-C₃N₄ (1 : 20) in RhB solution. When the ratio of Ag₂O/g-C₃N₄ (0.02204 min⁻¹) was 2 : 1, the rate constant of the as-prepared composites was approximately 12 times higher than that of g-C₃N₄ (0.00183 min⁻¹) and 3 times higher than that of pure Ag₂O (0.00732 min⁻¹) under the same conditions. However, among all as-prepared heterostructure composites, BiOBr/g-C₃N₄ (0.04 min⁻¹) with the mass ratio of 1 : 3 exhibits the highest photocatalytic activities and RhB solution could be completely degraded within 2 h. We could conclude that the as-prepared composites which meet the pseudo-first-order model are more ideal photocatalysts.

Fig. 4 Photocatalytic activity of Ag₂O/g-C₃N₄ (a), Ag/g-C₃N₄ (c), BiOBr/g-C₃N₄ (e) and the corresponding rate constant of RhB aqueous solution photodegradation of Ag₂O/g-C₃N₄ and Ag/g-C₃N₄ (b–d), BiOBr/g-C₃N₄ (f) under visible light.

The enhanced photocatalytic activities of the as-prepared Ag₂O/g-C₃N₄ (2 : 1) and BiOBr/g-C₃N₄ (1 : 3) composites which may attribute to the formation of heterostructure between two components which enhance the separation and transfer of charge carriers at the interface. Owing to the formation of type II heterojunction, BiOBr/g-C₃N₄ (1 : 3) composites exhibit the best photocatalytic activity, which promotes more effective separation and transfer of photogenerated carriers. Further adjusting the ratio of g-C₃N₄ and Ag₂O to 1 : 4, the slightly decreased photocatalytic activity may be attributed to the fact that the higher content of Ag₂O may easily result in the agglomerate of Ag₂O particles causing a low dispersibility on the surface of g-C₃N₄. In Ag₂O/g-C₃N₄ (2 : 1) composites, the intimate junctions between g-C₃N₄ and Ag₂O bring some synergistic reaction to enhance the photocatalytic activity. With respect to the Ag/g-C₃N₄ composites, the improved dispersibility and the decreased particle size of Ag nanoparticles of the as-prepared composites also play important roles in enhancing photocatalytic activity. With the mass ratio of g-C₃N₄ and Ag increasing from 30 : 1 to 10 : 1, the worse photocatalytic activity can be ascribed to the fact that the higher content of Ag may result in self-nucleation growth of Ag particles hindering rapid separation of photogenerated charge carriers. This may influence the transfer of photogenerated charge carriers as well as the separation efficiency of the photoinduced electron–hole pairs. When the ratio of BiOBr/g-C₃N₄ and Ag/g-C₃N₄ composites is 1 : 2 or 1 : 4, samples exhibited poor photocatalytic activities. High content of Ag may impede the transfer of photogenerated from g-C₃N₄ to Ag. When the content of BiOBr is more than twice as much as that of g-C₃N₄, BiOBr could be inspired to produce photogenerated electrons and holes, which limits the transfer of photogenerated electrons from g-C₃N₄ to BiOBr. Although BiOBr is favorable for separation of photogenerated hole–electron pairs, more BiOBr in the composites might reduce the photoabsorption of visible light. Therefore, in heterostructure photocatalysts, a suitable ratio between two components is significant for effectively enhancing the photocatalytic activity. Moreover, it is clearly that type II heterostructure semiconductor nanocomposites show better interface interaction.

From the viewpoint of practical applications, the stabilities of the Ag₂O/g-C₃N₄ (2 : 1) composites and BiOBr/g-C₃N₄ (1 : 3) heterojunctions were further investigated by recycling the photocatalyst for degrading RhB under visible light irradiation. As shown in Fig. 5, the as-prepared two kinds of heterostructure composites show good catalytic stability, maintaining a similar level of reactivity after four cycles. The slight decrease should derive from the inevitable loss of catalyst during the recycling process. In fact, in each cycle of tests, the RhB solution can be almost completely decolored after irradiation for 180 min. In addition, the present photocatalyst can be easily separated from the aqueous system by a simple sedimentation, due to its large size. Thus, the Ag₂O/g-C₃N₄ (2 : 1) and BiOBr/g-C₃N₄ (1 : 3) heterojunctions described in this paper hold extraordinarily high stability and recyclability for photocatalytic applications.

Fig. 5 Circling runs in the presence of the Ag₂O/g-C₃N₄ (2 : 1) composite (a) and BiOBr/g-C₃N₄ (1 : 3) (b) for photodegradation of RhB dye under visible light irradiation.

On the basic of the results above, possible photocatalytic mechanisms of the as-prepared Ag₂O/g-C₃N₄ and BiOBr/g-C₃N₄ composites under visible light irradiation were illustrated in Scheme 3. As a n-type semiconductor material, g-C₃N₄ can construct p–n heterostructure with Ag₂O (Scheme 3a). Both Ag₂O and g-C₃N₄ can be excited to generate electrons (e⁻) and holes (h⁺) under visible light irradiation. In the common p–n type II junctions, owning to the inner electric field if the photogenerated electrons tended to transfer from g-C₃N₄ to Ag₂O, the photogenerated holes had an opposite transfer.²⁶ However, the conduction band (CB) bottom of Ag₂O is more positive than that of g-C₃N₄, while the valence band (VB) top of Ag₂O is more negative.²⁷ Because of the inner electric field, the transfer of photogenerated electrons from Ag₂O to g-C₃N₄ is partly limited and the transfer of h⁺ could be accelerated which causes highly efficient separation of photogenerated charge carriers under visible light irradiation. The photogenerated electrons on the CB of g-C₃N₄ could be captured by oxygen to generate active species (e⁻ + O₂ + H⁺ = HO₂ (aq)). In the Ag₂O/g-C₃N₄ composites, ˙OH, H₂O₂ and O₂⁻ exhibit more positive potentials than the VB of Ag₂O.²⁸ The holes on the VB of Ag₂O could directly decompose the absorbed RhB dyes on the surface of the photocatalysts, which is helpful to limit the recombination of photogenerated electrons and holes.²⁸ As a result, the photogenerated electron–hole pairs can be separated efficiently, which leads to a highly enhanced photocatalytic activity than the pure g-C₃N₄ and Ag₂O. Therefore, we conclude that the enhanced photocatalytic activity of the as-prepared Ag₂O/g-C₃N₄ can be attributed not only to the improved dispersity and the appropriate mass ratio of g-C₃N₄ and Ag₂O as well as the improved optical absorption property arising from the heterostructure, but also to their synergetic effects of matched energy band structure and the inner electric field. All the reasons above can bring the rapid efficient separation of the photogenerated electron–hole pairs.

Scheme 3 Proposed photocatalytic mechanisms over the Ag₂O/g-C₃N₄ composites (a) and BiOBr/g-C₃N₄ composites (b).

In the Ag/g-C₃N₄ composites, on the one hand, the fast generation, rapid separation and transportation of these photogenerated carriers at the interface of g-C₃N₄ and Ag are the main reason for the enhanced visible-light photocatalytic activity of Ag/g-C₃N₄. As is reported, the theoretical calculated CB and VB potential of g-C₃N₄ are about −1.3 and 1.4 V (vs. NHE),²⁹ respectively, while the Fermi level of Ag is 0.4 V (vs. NHE).³⁰ After visible-light irradiating, the generated electrons on g-C₃N₄ can transfer rapidly to Ag nanoparticles, creating a Schottky barrier that narrows the band gap of g-C₃N₄. On the other hand, the visible-light harvesting ability of Ag/g-C₃N₄ is enhanced due to surface plasmon resonance (SPR) absorption of silver nanoparticles. The Ag/g-C₃N₄ with higher loading of Ag nanoparticles could exhibit stronger SPR effect, but Ag/g-C₃N₄ with excess Ag nanoparticles show worse photocatalytic activity which may attribute to the light opacity increasing, preventing g-C₃N₄ from the effective light absorption.

As is shown in Fig. 3b, the band structures of g-C₃N₄ and BiOBr are well-matched upon their intimately contacted interface and aligned with each other. The relative CB and VB edge positions of g-C₃N₄ nanosheets and BiOBr nanosheets imply that both the well-matched band energies and crystal planes can form type II heterojunctions at nanoscale level. Both g-C₃N₄ and BiOBr can be excited by visible light and the exited electrons produced by g-C₃N₄ were injected into the CB of BiOBr. The exited holes produced by BiOBr were injected into the VB of g-C₃N₄ which results in efficient separation and transport of photo-induced electrons and holes. Both electrons and holes all show extreme activity, which can decompose RhB directly. However, there are some arguments about the effect of photogenerated electrons. On the one hand, owing to RhB is active with light, the dye-sensitised photocatalysis would also exist in this system. In fact, as shown in Scheme 4, dye sensitization process cannot occur because the lowest unoccupied molecular orbital (LUMO) potential of RhB (−0.75 eV) is more positive than the CBM of g-C₃N₄.³¹ The excited electrons of RhB cannot transfer to the CB of g-C₃N₄. On the other hand, as is reported, the conduction band and the valence band of BiOBr are both more positive than the standard redox potentials of O₂/O₂⁻ and ˙OH/OH⁻. Even though a successful transfer of electrons from the LUMO of RhB to the CB of BiOBr would take place, these electrons cannot further reduce O₂ to form O₂⁻ species which would return to the HOMO of RhB, as marked by the dotted line (blue, Scheme 4). We believe that the dye-sensation which would be helpful to degradate RhB in the BiOBr/g-C₃N₄ system is impossible. Thus, it is expected that the enhanced photocatalytic activity of the BiOBr/g-C₃N₄ 2D nanojunctions can be attributed to the well-matched band structure of type II heterojunctions and the appropriate mass ratio of g-C₃N₄ and BiOBr as well as the improved optical absorption property and the special morphology of layer-by-layer self-assembly.

Scheme 4 Proposed routes for the transfer of excited electrons from RhB.

On the basis of the above discussion, compared with pure g-C₃N₄, the enhanced photocatalytic activity of Ag₂O/g-C₃N₄ and BiOBr/g-C₃N₄ toward the degradation of RhB is explained by the enhanced visible-light utilization efficiency due to the p–n heterostructures between two components. Due to superior type II heterojunction, the photocatalytic activity of BiOBr/g-C₃N₄ enhanced much more than that of Ag₂O/g-C₃N₄. One of the main reasons is that, under visible light irradiation, type II heterojunction photocatalysts can achieve more rapidly generate, separate of photogenerated charge carriers and participate in photocatalytic degradation of RhB, where electrons and holes play significant roles to degrade organic dyes. However, as a type I heterostructure, photogenerated electrons and holes of Ag₂O/g-C₃N₄ are inevitably likely to recombine on the surface of Ag₂O which affect to photocatalytic activity. Although the surface modification of noble metals and surface plasmon resonance effect can contribute to higher the photocatalytic activity in Ag/g-C₃N₄ composites, structure defects still remain in bulk g-C₃N₄ which limits the efficient separation of photogenerated charge carriers. There is no doubt that semiconductor composites photocatalysts with p–n heterostructure can be more able to improve the photocatalytic activity than noble metals modified composites.

Generally, photoinduced reactive species including trapped holes, ˙OH radicals, and O₂˙⁻ are expected to be involved in the photocatalytic process. To prove the photocatalytic mechanism, the main reactive species in the photocatalytic process were detected through the trapping experiments of radicals using CH₃OH as the hydroxyl radical scavenger, p-benzoquinone (PBQ) as the superoxide radical scavenger, and triethanolamine (TEA) as the hole radical scavenger.

Fig. 6a shows that the degradation efficiency of RhB in the Ag₂O/g-C₃N₄ composites (2 : 1) decreased obviously with the addition of a hydroxyl radical scavenger (CH₃OH; the conversion of RhB was 98.6%), which reduced slightly with the addition of a superoxide radical scavenger (PBQ; the conversion of RhB was 24%) and reduced mostly with the addition of a hole radical scavenger (TEA; the conversion of RhB was 54.3%). This results indicate that superoxide radicals were the main reactive active species and the hole radicals were secondary reactive active species. As is shown in Fig. 6b, under visible light irradiation photocatalytic activities of BiOBr/g-C₃N₄ composites (1 : 3) decreased slightly after addition of TEA (the conversion was 19.2%), which means holes radicals were the main active species.

Fig. 6 Plots of photogenerated carrier trapping during RhB degradation over the Ag₂O/g-C₃N₄ composite (2 : 1) (a) and BiOBr/g-C₃N₄ composites (1 : 3) (b) under visible light irradiation.

Conclusions

In this paper, a series of Ag₂O/g-C₃N₄, Ag/g-C₃N₄, and BiOBr/g-C₃N₄ heterostructured photocatalysts with different mass ratios of g-C₃N₄ and Ag₂O were prepared by a simple liquid phase reaction process at room temperature. The photodegradation of RhB experiments revealed that the BiOBr/g-C₃N₄ heterostructured composite with the mass ratio of 1 : 2 degraded nearly 100% of RhB within 2 h, while Ag₂O/g-C₃N₄ (2 : 1) degraded thoroughly within 3 h. The degradation rate constant of BiOBr/g-C₃N₄ (1 : 3) was 0.04 min⁻¹ under visible light, which was almost 2 and 14 times more than that of Ag₂O/g-C₃N₄ and Ag/g-C₃N₄, respectively. In composite systems, the reaction kinetics of RhB dye photo-degradation of Ag₂O/g-C₃N₄ and BiOBr/g-C₃N₄ composites were fitted with the pseudo-first-order model, ln(C₀/C) = kt, while the experimental data of Ag/g-C₃N₄ and pure g-C₃N₄ were fitted with the zero-order kinetics model. Then, we conclude that the as-prepared samples which meet the pseudo-first-order model are optimal photocatalysts. The results provided here not only offer a highly efficient and stable photocatalytic material for environmental remediation, but also showed a new insight for constructing efficient heterostructured photocatalysts.

Acknowledgements

This work was supported in part by the program for Taishan Scholars, the projects from National Natural Science Foundation of China (Grant no. 51572109, 51501071, 51302106, 51402123, and 51402124).

References

1. C. Chen, W. Ma and J. Zhao, Chem. Soc. Rev., 2010, 39, 4206.
2. S. Liu, J. Yu and M. Jaroniec, J. Am. Chem. Soc., 2010, 132, 11914.
3. S. T. Kochuveedu, D. P. Kim and D. H. Kim, J. Phys. Chem. C, 2012, 116, 2500.
4. J. Zhang, X. Chen, K. Takanabe, K. Maeda, K. Domen, J. D. Epping, X. Fu, M. Aatonietti and X. Wang, Angew. Chem., Int. Ed., 2010, 49, 441.
5. M. Xu, L. Han and S. Dong, ACS Appl. Mater. Interfaces, 2013, 5, 12533.
6. Y. Tian, B. Chang, J. Lu, J. Fu, F. Xi and X. Dong, ACS Appl. Mater. Interfaces, 2013, 5, 7079.
7. C. Han, L. Ge, C. Chen, Y. Li, X. Xiao, Y. Zhang and L. Guo, Appl. Catal., B, 2014, 147, 546.
8. L. Xu, J. Xia and H. Xu, J. Power Sources, 2014, 245, 866.
9. Z. Chen, P. Sun, B. Fan, Z. Zhang and X. Fang, J. Phys. Chem. C, 2014, 118, 7801.
10. T. Zhu, Y. Song, H. Ji, Y. Xu, Y. Song, J. Xia, S. Yin, Y. Li, H. Xu, Q. Zhang and H. Li, Chem. Eng. J., 2015, 271, 96.
11. K. Shen, M. A. Gondal, R. G. Siddique, S. Shi, S. Wang, J. Sun and Q. Xu, Chin. J. Catal., 2014, 35, 78.
12. L. M. Lyu and M. H. Huang, J. Mater. Sci., 2014, 49, 5309.
13. D. Sarkar, C. K. Ghosh, S. Mukherjee and K. K. Chattopadhyay, ACS Appl. Mater. Interfaces, 2012, 5, 331.
14. Y. Yang, Y. Guo, F. Liu, X. Yuan, Y. Guo, S. Zhang, W. Guo and M. Huo, Appl. Catal., B, 2013, 142, 828.
15. J. M. Song, C. J. Mao, H. L. Niu, Y. H. Shen and S. Y. Zhang, CrystEngComm, 2010, 12, 3875.
16. L. Kong, Z. Jiang, T. Xiao, L. Lu, M. O. Jones and P. P. Edwards, Chem. Commun., 2011, 47, 5512.
17. Y. Li, Y. Liu, J. Wang, E. Uchaker, Q. Zhang, S. Sun, Y. Huang, J. Li and G. Cao, J. Mater. Chem. A, 2013, 1, 7949.
18. S. Shenawi-Khalil, V. Uvarov, S. Fronton, I. Popov and Y. Sasson, J. Phys. Chem. C, 2012, 116, 11004.
19. S. C. Yan, Z. S. Li and Z. G. Zou, Langmuir, 2009, 25, 10397.
20. H. T. Ren, S. Y. Jia, Y. Wu, S. H. Wu, T. H. Zhang and X. Han, Ind. Eng. Chem. Res., 2014, 53, 17645.
21. K. Li, S. Guo, Q. Wang, H. Xu, Z. Wang, B. Huang, Y. Dai and J. Lu, ACS Appl. Mater. Interfaces, 2015, 7, 9023–9030.
22. P. Niu, L. Zhang, G. Liu and H. M. Cheng, Adv. Funct. Mater., 2012, 22, 4763–4770.
23. C. Su, L. Liu, M. Zhang, Y. Zhang and C. Shao, CrystEngComm, 2012, 14, 3989.
24. Y. Sun, W. Zhang, T. Xiong, Z. Zhao, F. Dong and R. Wang, J. Colloid Interface Sci., 2014, 418, 317.
25. Y. Liu, J. Wang, P. Yang and K. Matras-Postolek, RSC Adv., 2015, 5, 61657.
26. L. Li, L. Xu, W. Shi and J. Guo, Int. J. Hydrogen Energy, 2013, 38, 816.
27. S. C. Yan, S. B. Lv, Z. S. Li and Z. G. Zou, Dalton Trans., 2010, 39, 1488.
28. X. Wang, S. Li, H. Yu, J. Yu and S. Liu, Chem.–Eur. J., 2011, 17, 7777.
29. X. Wang, S. Blechert and M. Antonietti, ACS Catal., 2012, 2, 1596.
30. J. Li, S. K. Cushing, J. Bright, F. Meng, T. R. Senty, P. Zheng, A. D. Bristow and N. Wu, ACS Catal., 2012, 3, 47.
31. H. Gerischer and F. Willig, Physical and Chemical Applications of Dyestuffs, Springer, Berlin Heidelberg, 1976, vol. 31.

---