Bi2O3 and g-C3N4 quantum dot modified anatase TiO2 heterojunction system for degradation of dyes under sunlight irradiation

A facile and feasible method was successfully utilized to incorporate Bi2O3 and g-C3N4 quantum dots on TiO2 surface to synthesize a novel composite g-C3N4/TiO2/Bi2O3. The photocatalytic activity of the composite g-C3N4/TiO2/Bi2O3 for degradation of dyes under sunlight and UV light irradiation was evaluated. It possessed the higher photocatalytic performance than that of pristine TiO2 or g-C3N4 under the same conditions. Under sunlight irradiation, the reaction rate constants of the g-C3N4/TiO2/Bi2O3 was about 4.2 times and 3.3 times higher than that of TiO2 and g-C3N4, respectively. The promising photocatalytic performance was attributed to the broader light absorption range and efficient separation of photoinduced carriers. Moreover, based on the TEM, XPS, XRD, UV-vis spectrum, radicals scavenging test and Mott–Schottky analysis systematic mechanism for photodegradation process was proposed. This work provides a promising strategy for the modification of TiO2-based semiconductors by incorporating different quantum dots and promoting the efficiency of the photocatalysts in practical application.


Introduction
At present, environmental contamination and energy shortage are the most serious problems on our planet. Hence, there is an urgent need for the development of an environmental, sustainable, economical and energy-saving technology for contamination control. Semiconductor photocatalysis provides an optimal solution for the problems with the potential that the clean solar energy can be harnessed for decomposition of pollutants. 1 To date, TiO 2 as the photocatalyst has drawn signicant attention and been widely studied because of its desirable properties including high reactivity, chemical stability, nontoxicity, and cheap price. [2][3][4][5][6][7] Nevertheless, the application of TiO 2 -based photocatalysts is hindered due to its inefficient utilization of solar energy, poor charge separation, and low adsorption of organic contaminants on the surface. Besides these intrinsic drawbacks, the complicated process of synthesis is another key factor to impede its practical application. 8 To avoid the above limitation, various modication strategies have been carried out, such as dye-sensitization, metal or non-metal doping, and transition metal doping, etc. Particularly, compositing TiO 2 with different semiconductor into heterojunction is an efficient method to enhance the photocatalytic activity. 8 Since the heterojunction structure would decrease the recombination of electron/hole by separating the charge carriers to different semiconductors surface due to the different potential barriers. These heterojunctions-based semiconductors exhibit advanced efficiency for degrading pollutants.
Presently, considerable attention has been paid to graphitic carbon nitride (g-C 3 N 4 ), which is a stable kind of polymers with a layered structure like graphene. Particularly, it possesses the strong ability to harvest solar energy due to its narrow band gap (E g ¼ 2.7 eV). Nevertheless, the quantum efficiency of pure g-C 3 N 4 is low because of the high recombination ratio of photogenerated electron/hole pair. 9 Combining the TiO 2 with g-C 3 N 4 in heterojunction is considered as an efficient method which can improve the absorption of visible light and the separation of photogenerated pairs. Yu et al. 10 synthesized g-C 3 N 4 -TiO 2 composites by calcination process using P25 and urea, achieving the high photoactivity for the degradation of HCHO. Lei et al. 11 reported the synthesis of g-C 3 N 4 /TiO 2 photocatalyst via simple calcination using cyanamide and anatase TiO 2 as precursor. The g-C 3 N 4 /TiO 2 photocatalyst exhibits a promising progress in degradation of the dye Acid Orange 7 under both visible and UV light. The g-C 3 N 4 -TiO 2 composites with prominent photoactivity for the degradation of phenol under UV light was synthesized by Colo'n et al. 12 by impregnation. Zhang et al. 13 found that well-dispersed TiO 2 nanocrystals with (001) facets prepared in situ on g-C 3 N 4 through a solvothermal method exhibit higher efficiency for photocatalytic degradation of phenol as compared to pure g-C 3 N 4 and TiO 2 . Wu et al. 14 prepared nanosheets TiO 2 /g-C 3 N 4 composite by solvothermal method, which exhibits a signicant improvement in photodegradation towards methylene blue under visible light irradiation than pristine g-C 3 N 4 and TiO 2 .
Among strategies of constructing heterojunction, synthesis of p-n junction is considered as an effective method to facilitate the photoactivity by inducing the separation of photogenerated carries due to the existence of an internal electric eld. Bismuth oxide (Bi 2 O 3 ), as a p-type semiconductor, has been extensively studied due to its suitable band gap (E g ¼ 2.8 eV) and visible light driven catalytic activity. Among the six different polymorphic phases of Bi 2 O 3 , b-Bi 2 O 3 has prominent advantages for degradation of pollutants. Since TiO 2 is a n-type semiconductor, it is feasible to combine TiO 2 with b-Bi 2 O 3 , by which catalytic efficiency can be notably enhanced. [15][16][17][18] Recently, much efforts have been made to synthesize TiO 2 /Bi 2 O 3 heterojunction. Various methods have been employed such as: pulse electrodeposition, 19 hydrothermal, sol-gel, and coprecipitation. [20][21][22][23][24][25][26] However, the complexity of these methods limits the application of photocatalysis.
Instead of above strategies, researchers have succeeded in sensitizing TiO 2 by modifying quantum dots (QDs) of low-band gap materials such as g-C 3 N 4 , CdS, CdSe, CdTe, and Bi-based materials, which can absorb light in the visible region. QDs can match the solar spectrum better due to the particle size effect. Additionally, QDs are recently reported to generate multiple excitons, which can improve the photocatalysis efficiency. [27][28][29][30] Jiao et al. 31 rstly prepared Bi 2 O 3 quantum dots decorated TiO 2 with exposed {001} facets on graphene sheets. Size-controllable were in situ synthesized on TiO 2 nanotube arrays and high activity in synergetic H 2 evolution and organics degradation. 32 In this study, we rstly use a simple ball-milling/calcination method to synthesize a g-C 3 N 4 /TiO 2 /Bi 2 O 3 heterojunction system which distinctly exhibit a promising synergetic improvement effect. The different properties of the catalyst are evaluated by the degradation of simulated dyeing wastewater under UV light and direct sunlight irradiation. While the different properties are conrmed by TEM, XPS, XRD and UV-vis diffuse reection spectra. A tentative mechanism for photocatalytic degradation of RhB by the composite g-C 3 N 4 /TiO 2 /Bi 2 O 3 is proposed.
2.1.2 Methods 1. Synthesis of pure g-C 3 N 4 . 5 g of melamine were added into crucible with a cover then heated at 520 C under air atmosphere for 2 h with a heating rate of 5 C min À1 . Aer cooling to the room temperature naturally, the resulting yellow bulks were collected and ball-milled into powders for further evaluation.
2. Synthesis of pure Bi 2 O 3 . 3 g Bi(NO 3 ) 3 $5H 2 O were added into crucible with a cover then heated at 520 C under air atmosphere for 2 h with a heating rate of 5 C min À1 . Aer cooling to the room temperature naturally, the resulting bulks were collected and ball-milled into powders for further evaluation.
3. Synthesis of g-C 3 N 4 /TiO 2 /Bi 2 O 3 . The g-C 3 N 4 /TiO 2 /Bi 2 O 3 catalysts were prepared by the processes of ball-milling and calcination. Firstly, 400 mg of anatase TiO 2 powders and 600 mg of melamine were added into the agate ball milling tank, followed by the process of ball milling for 2 h with a ball powder ratio of 10 : 1, at a speed of 400 rpm in a planetary ball mill. Then, a certain amount of Bi(NO 3 ) 3 $5H 2 O were added into the agate ball milling tank which was lled with TiO 2 and melamine. The powders mixed by ball milling for 2 h with a ball powder ratio of 10 : 1, at a speed of 400 rpm in a planetary ball mill. Subsequently, the mixed powder was placed into a crucible with a cover then heated at 520 C under air atmosphere for 2 h with a heating rate of 5 C min À1 . Aer cooling to the room temperature naturally, the resulting yellow bulks were collected and ball-milled into powders for further evaluation. The obtained powers was named as TCB-x, where the x refers to the weight percentage of Bi(NO 3 ) 3 $5H 2 O to the weight of TiO 2 and melamine.

Characterization
The crystal phases was characterized by X-ray diffraction (XRD, Bruker D8 ADVANCE A25X) with a Cu Ka radiation with a diffraction angle between  Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) were measured by Tecnai G2 F20 S-TWIN (200 kV). The samples were dispersed in ethanol and dropped on copper grids. Chemical composition and valence band (VB) were observed by X-ray photoelectron spectroscopy (XPS, Escalab 250Xi) using a monochromatic microfocused Al Ka (1486. 6 eV) source. UV-vis diffuse reectance spectroscopy (UV-vis DRS) absorption spectra of the wavelength between 200 nm and 800 nm were carried out at a Shimadzu UV-3600 spectrophotometer using BaSO 4 as a reference.
The electrochemical studies were conducted on an electrochemical workstation (CHI660C, CH Instrument Corp, Shanghai), which used catalyst-deposited FTO glass as working electrode, Pt as the counter electrode and Ag/AgCl as the reference electrode. Meanwhile, 0.5 M Na 2 SO 4 was served as the electrolyte solution. The working electrode was prepared as follows: 20 mg PEG-600 and 10 mg pure TiO 2 and Bi 2 O 3 were dispersed in 1 mL ethanol and ultrasonically scattered for 1 h. Then, the suspension above was added onto the FTO glass (1 Â 1 cm) and evaporated to dry. The Mott-Schottky measurement was performed at frequency of 1000 Hz.

Photocatalytic activity test
The aqueous Rhodamine B (RhB), methylene blue (MB) and methyl orange (MO) dyes were used as model organic contaminants to evaluate the photocatalytic activities of the synthesized catalysts. Degradation experiments were carried out under UV and direct solar light irradiation. A high-pressure 300 W mercury lamp with the radiation of 365 nm was used as the UV light source. To study the photocatalytic activity of the samples under visible light, the photoreactor were directly exposed to sunlight. And the illumination intensity was measured by solar power meter (TES-1333R). 50 mg catalyst powders were suspended in 50 mL aqueous organic dye solution and the dyes concentration was 20 mg L À1 . To reach adsorption-desorption equilibrium, the solution was stirred for 60 min in the dark. At given time intervals, 4 mL aliquots were taken and centrifuged at 10 000 rpm for 10 min to remove the particles. Then, the concentration of organic dyes in ltrates was analyzed using an Alpha-1506 UV-vis spectroscopy (Shanghai Lab-Spectrum Instruments Co. Ltd, China) at the wavelength of 554 nm, 664 nm, 465 nm for RhB, MB, MO, respectively. The experiments were repeated 3 times.

Microstructure analysis
The morphology of the prepared catalysts was characterized by TEM and high-resolution. TEM image Fig. 2b, c and d are enlarged views of Fig. 2a. As shown in Fig. 2a, TiO 2 particles have a spherical structure with an average size between 30-60 nm, which was in well accordance with the crystallite size obtained from Scherrer equation. Bi 2 O 3 and g-C 3 N 4 showed a much smaller size below 10 nm, and intimately cover on the surface of TiO 2 particles, which intensively increase the specic surface area of TiO 2 .
Clear lattice fringes for the identication of crystallographic spacing could be observed in Fig. 2c, and the lattice spacing of 0.350 nm and 0.321 nm corresponded well with (101) phase of TiO 2 and (221) phase of Bi 2 O 3 , 15-17 respectively. In Fig. 2d, the lattice spacing of 0.318 nm matched the (002) crystal phase of g-C 3 N 4 . [37][38][39] Thus, these TEM images indicated the interaction of g-C 3 N 4 / TiO 2 /TiO 2 heterojunction composite. The g-C 3 N 4 and Bi 2 O 3 located on the surface of TiO 2 as quantum dots. In Fig. 2e, the pure g-C 3 N 4 showed aggregated morphologies, which were comprised of block-based akiness. Fig. 2f displayed a spheroidal structure of Bi 2 O 3 . It was obvious that the process of ballmilling played a signicant role in synthesizing g-C 3 N 4 and Bi 2 O 3 quantum dots on TiO 2 surface. Fig. 3 showed the full survey and high-resolution spectra for the Ti 2p, O 1s, C 1s, N 1s and Bi 4f region. Fig. 3b depicts the C 1s of the prepared. The peak centered at 284.6 eV could be assigned to the adventitious carbon while the peak at 287.5 eV belongs to the N-C]N group of the g-C 3 N 4 . 40-42 Fig. 3d showed N 1s spectrum of catalyst. The N 1s peak was tted into two peaks at 399.3 eV and 400.2 eV, which were the sp 2 -hybridized nitrogen (C]N-C) and tertiary nitrogen (N-3), respectively. [40][41][42] Furthermore, two tted peaks of Ti 2p3/2 and Ti 2p1/2 at 458.9 eV and 464.4 eV suggested the presence of Ti(IV). 27,36,43 In Fig. 3a, two tted peaks at 158.9 eV and 164.3 eV belonged to Bi 4f5/2 and Bi 4f7/2, suggesting the existence of Bi 2 O 3 in the catalyst. 24,34,35 These results suggested that the composite TCB-30% consisted of g-C 3 N 4 , anatase TiO 2 and Bi 2 O 3 .

UV-vis DRS analyses
The UV-vis diffuse reectance spectra of different composites were exhibited in Fig. 4. It is obvious that the composite of TCB-30% showed a signicant red shi of band edge to 456.5 nm compared with the initial TiO 2 , Bi 2 O 3 and g-C 3 N 4 . There is also an intensively long tailing absorption in visible region, which can be explained by the synergistic effects of Bi 2 O 3 and g-C 3 N 4 quantum dots with a narrower band gap.
The band gap energies were determined using the equation below: where A, h, v, E g and a are the absorption coefficient, Planck's constant, frequency, band gap energy and a constant, respectively. For the composites in this study, n is 4 because of the indirect transition. Therefore, Tauc's plots of (ahn) 1/2 versus photon energy (hn) are obtained. As shown in Fig. 4b, the band gap energies of TiO 2 , TCB-30%, Bi 2 O 3 and g-C 3 N 4 are 3.20 eV, 2.24 eV, 2.80 eV and 2.7 eV, respectively. The TCB-30% has a lower band gap energy compared to TiO 2, which may be due to the heterojunction structure of the composite.
In order to further explain the photocatalytic activity of the ternary composite, the band structure of Bi 2 O 3 was explored according to the following empirical equation.
where E VB is the valence band edge potentials, E CB is the conduction band energy, X is the electronegativity of  Table 1.   The degradation of RhB followed the rst-order kinetics, which can be expressed by the equation

Photocatalytic activity
C 0 and C t is RhB concentration at time 0 and time t, and k a

À1
represents the rst-order reaction rate constant. As shown in Fig. 6, TCB-30% displayed a maximum value of rate constant under UV, which was about 2.2 times as high as that of TiO 2 and 2.6 times as that of g-C 3 N 4 . When under the solar irradiation (640 W m À2 ), the k a À1 of TCB-30% was 4.2 times and 3.3 times higher than that of TiO 2 and g-C 3 N 4 respectively. Therefore, the improvement of photocatalytic activity under solar irradiation is more remarkable than that in the condition of UV light. Moreover, TCB-30% was also used to degrade different kinds of dyes with the same initial concentration under sunlight irradiation, including methylene blue (MB) and methyl orange  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 1181-1190 | 1185 Paper (MO). As it shown in Fig. 7a, TCB-30% displayed a limited degradation rate for methylene blue and methyl orange compared with RhB. Fig. 8 showed the dark adsorption of three dyes over TCB-30%. It was supposed that the promising reaction rate could be contributed by the strong adsorption of the prepared heterojunction materials towards organic dyes with hydroxy groups. Thus, it is worth further research on how the surface groups of this composite affects the reaction with organic pollutants.
To gure out the contribution of photo-induced active radical in the reaction process, benzoquinone (BQ), isopropanol (IPA) and ammonium oxalate (AO) were applied as scavengers of superoxide radical ($O 2 À ), hydroxyl radical ($OH) and hole (h + ), respectively. The comparison experiments were carried out in the presence of the same molar concentration of scavengers (BQ, IPA and AO) for visible light degradation of RhB with TCB-30%. As displayed in Fig. 9, the degradation rate without any scavengers is 99.6%. By contrast, by adding BQ and IPA, the degradation rate sharply suppressed. While with the addition of AO scavengers, there is only a slight decrease of degradation rate. Therefore, it can be concluded that superoxide radical ($O 2 À ) and hydroxyl radical ($OH) play key roles in degradation towards RhB under visible light.
During the practical application of catalysts, the by-products tend to be adsorbed on the active sites of photocatalysts' surface, leading to the dramatic decrease of the photocatalytic activity. To evaluate the stability of the prepared TCB-30%, recycle experiments were conducted under solar irradiation. As shown in the Fig. 9, the TCB-30% still presents the excellent photocatalytic performance aer three cycles.

Mechanism discussion
In order to investigate the electrochemical properties of TiO 2 and Bi 2 O 3 , Mott-Schottky plots were obtained. In Fig. 10a, TiO 2 exhibit a positive slope indicating that it's n-type semiconductor. On the contrary, BiO 2 shows a negative slope, which indicts a p-type semiconductor.
Based on the results of XRD, XPS, TEM and UV-vis DRS analysis and Mott-Schottky, a tentative heterojunction system proposed and depicted in Scheme 1. TiO 2 is placed as a support and connector of Bi 2 O 3 and g-C 3 N 4 quantum dots. The promising photocatalytic performance of TCB-30% under visible light is signicantly related to the extended light adsorption spectrum and enhanced electron-hole separation.
On the one hand, the generation of g-C 3 N 4 and Bi 2 O 3 quantum dots with a narrow band gap facilitated the improvement of visible light harvesting ability, which motivates the generation of photoinduced carriers under visible light. On the   When the electrons migrated from TiO 2 , leaving a positively charged donor ion, and a positive charge region was formed on the side of the n region. In the same way, a negative charge region was formed on the side of p region. Therefore, there was no electrical neutrality on both sides of the interface of the p-n junction, and a positively and negatively charged region appeared, becoming a space charge region, which can force photoinduced electrons and holes to transfer in the opposite direction (Fig. 10b) As for the heterojunction between g-C 3 N 4 quantum dots and TiO 2 particles, since the g-C 3 N 4 has a more positive conduction band edge potential than that of TiO 2 (À0.28 eV), the photoinduced electrons under visible irradiation tend to transfer from the CB of g-C 3 N 4 to CB of TiO 2 . Based on above calculation and analysis, an energy band structure has been proposed in Scheme 1. The photogenerated electrons and holes are unable to migrate from g-C 3 N 4 to Bi 2 O 3 through TiO 2 . Besides, the Bi 2 O 3 and g-CN 4 quantum dots don't interact and they are located in different position of TiO 2 surface. Consequently, it's believed that there are just two heterojunctions (TiO 2 /g-C 3 N 4 , TiO 2 /Bi 2 O 3 ) instead of a ternary heterojunction.
According to the previous researches, there are three types of radical degradation mechanism, including superoxide radical ($O 2 À ), hydroxyl radical ($OH) and hole (h + ). 3,7,8,44,45 Steps of generation of the three radicals can be explained as follows:    OH À + h + / $OH In this work, there was a minor change in degradation performance aer eliminating the holes through the scavenging test. Thus, it was concluded that the majority dye was degraded by superoxide radical ($O 2 À ) or hydroxyl radical ($OH) induced by photo catalysis. Meanwhile, a slight number of dyes is degraded by hydroxyl radical ($OH) induced by photoinduced holes or directly oxidation by the holes. Eventually, the systematic mechanism, including the generation and transfer of carriers, generation of radicals and reaction between radicals and dyes, was proposed in Scheme 1.

Conclusion
In summary, a facile ball-milling/calcination was successively utilized to prepared a Bi 2 O 3 and g-C 3 N 4 quantum dots modied anatase TiO 2 heterojunction system. It was found that TCB-30% exhibited the highest removal rate for degradation of dyes under UV light irradiation and sunlight irradiation. Under sunlight irradiation, there was remarkable improvement of TCB-30%, the reaction rate constant was about 4.2 times and 3.3 times higher than that of TiO 2 and g-C 3 N 4 towards degradation of RhB. The notable enhancement of photocatalytic activity is mainly due to the ability of visible light harvesting and efficient separation of the carriers, which was induced by the modication of Bi 2 O 3 and g-C 3 N 4 quantum dots.

Conflicts of interest
There are no conicts to declare.