Bi₄O₅Br₂/SBA-15 was prepared through the following two steps. The first step was to synthesize Bi₄O₅Br₂ with solvothermal method. Typically, under stirring at 100 rpm, cetyl trimethyl ammonium bromide (CTAB, 1.5 mmol) and Bi(NO₃)₃∙5H₂O (1.5 mmol) were added and dissolved in the glycerol solution (15 mL glycerol and 6 mL deionized water), and the stirring lasted for 30 min. Then after 0.1 g polyethylene glycol-10,000 (PEG-10,000) being added to the mixture with pH adjusted to 11.0, the obtained mixture was solvothermally treated at 140˚C for 24 h. The solid obtained after centrifugation was washed with water and ethanol for 5 times, and then dried at 50˚C for 12 h.

In the second step, Bi₄O₅Br₂/SBA-15 was prepared by sol-gel method. In a typical run, 2.0 g polyether P₁₂₃ (PEO-PPO-PEO) was mixed with 65 mL deionized water at room temperature under stirring for 3 h. 10 ml HCl (2 mol/L) was then dropped into the above mixture at 40˚C under stirring for 1 h. The synthesized Bi₄O₅Br₂ in the first step was added into the above mixture at a different mass ratio of Bi/SiO₂ of 20/100, 30/100, 40/100, and 50/100. After stirring for 1 hour, 4.5 g tetraethyl orthosilicate (TEOS) was slowly dropped into the mixture at 4˚C under stirring for 24 h. The obtained mixture was later solvothermally treated at 140˚C for 24 h. The mixture was centrifuged, washed, and dried at 60˚C for 8 h. The obtained solid samples were calcined at 550˚C in air atmosphere for 6 h, which were labeled as Bi₂₀/SBA-15, Bi₃₀/SBA-15, Bi₄₀/SBA-15, and Bi₅₀/SBA-15, respectively.

X-ray diffraction (XRD) measurement was conducted on a D/max-2550 PC X-ray diffractometer. Scanning electron microscope (SEM, Hitachi S-4800) and transmission electron microscope (TEM, JEOL JEM-2100 high-resolution transmission electron microscope (HRTEM)) were also used to characterize the obtained photocatalysts. The optical diffuse reflectance spectra (DRS) were acquired with a UV-vis-NIR scanning spectrophotometer. The photoluminescence (PL) spectra were detected by using a spectrophotometer (FS5, England). The Brunauer-Emmett-Teller (BET) surface area was determined by a micromeretics ASAP 2010 apparatus with a multipoint BET method using the adsorption data in the relative pressure (P/P₀) range of 0.05 - 0.3. The desorption isotherm was used to determine the pore size distribution according to the Barrett-Joyner-Halenda (BJH) method. X-ray photoelectron spectra (XPS) analysis was performed on a PHI-5400 instrument with Mg Kα as the X-ray source under a pressure of 1.33 × 10⁻⁷ Pa and a step of 0.05 eV. The C (1s) level (285.0 eV) was taken as the reference binding energy.

RhB solution (10 mg/L) was used to evaluate the photocatalytic activity of the obtained composites under visible light irradiation. A 300 W xenon lamp with a 420 nm cut-off filter provided visible light irradiation. A same photocatalyst dosage of 0.2 g/L was used for all the tests. Prior to visible light irradiation, the mixtures of RhB and photocatalyst were magnetically stirred in the dark for 30 min to ensure the establishment of an adsorption-desorption equilibrium of RhB on the catalyst surface. Then, the solution was irradiated by xenon lamp. At given irradiation time intervals, 4 mL of the mixtures were collected and centrifuged at 10,000 rpm for 10 min to remove the photocatalysts. The residual concentration of RhB was analyzed with a 725N UV-v is spectrophotometer. The RhB solution temperature was controlled at 19˚C ± 2˚C during the whole experiment.

To detect the active species produced during the photocatalytic process, such as superoxide radical ( • O 2 − ), hole (h⁺) and hydroxyl radical (•OH), active species trapping experiments were carried out by adding scavenger into different RhB solutions. The scavengers were p-benzoquinone (p-BQ), sodium oxalate (SO), and iso-propanol (IPA), respectively. The dosage of each scavenger was 1.0 mmol/L.

Figure 1(a) showed the small angle XRD pattern of SBA-15 with different loading ratio of Bi₄O₅Br₂. It can be seen that all the synthesized samples have a strong diffraction peak (2θ = 0.90˚ to 1.0˚) and two weak diffraction peaks (2θ = 1.3˚ to 1.8˚) in the range of small angle (2θ = 0.3˚ to 5˚). These characteristic peaks are

coincident with the crystal plane diffraction peaks of (100), (110), and (200) in the two-dimensional hexagonal structure of SBA-15, respectively. This observation reveals the presence of a typical two-dimensional hexagonal ordered mesoporous structure [17] , further indicating that the loading of Bi₄O₅Br₂ did not change the mesoporous structure of SBA-15. The diffraction peaks shifted to higher angles from Bi₂₀/SBA-15 to Bi₅₀/SBA-15, meaning that their pore diameters gradually decreased along with the increase of loading ratio, and Bi₂₀/SBA-15 has the largest pore diameter [18]. On the other hand, the diffraction peak intensity was first increased and then decreased, with the intensity of Bi₃₀/SBA-15 diffraction peak being the highest. This observation indicated that the mesoporous structure of the synthesized Bi₃₀/SBA-15 is the most regular among the synthesized samples, since the high intensity of peaks is attributable to the incorporation of Bi into the mesoporous silica matrix, creating new crystal facets on mesoporous structure [19].

Figure 1(b) showed the wide angle XRD patterns of the synthesized Bi₄O₅Br₂/SBA-15 samples. As seen, all the samples had similar characteristic peaks, and the corresponding positions of the characteristic peaks were 11.2˚, 25.0˚, 31.5˚, 39.2˚, 46.2˚, 53.5˚, 57.2˚, 67.1˚, and 76.0˚, respectively. These peaks were corresponding to the tetragonal phase of Bi₄O₅Br₂ (standard card JCPDS No. 09-0699), which were consistent with the following crystal planes of Bi₄O₅Br₂ (001), (011), (012), (110), (112), (020), (211), (220), and (130), respectively. Especially, the strongest diffraction peak intensity was detected at 31.5˚, indicating that the tetragonal Bi₄O₅Br₂ grows along with the (012) crystal plane.

Table 1 summarized the specific surface area, pore volume and average pore size of the composites. The pore diameters of Bi₂₀/SBA-15, Bi₃₀/SBA-15, Bi₄₀/SBA-15, and Bi₅₀/SBA-15 firstly increased and then decreased. This variation trend was consistent with the results from the small angle XRD. With the increase of Bi₄O₅Br₂ loading ratio, both specific surface area and pore volume of the synthesized photocatalysts showed a decreasing trend, possibly due to the relatively bigger size of Bi₄O₅Br₂ particles which are easy to block a part of the pores. This phenomenon implied that the adsorption capacity of Bi₄O₅Br₂/SBA-15 might gradually decrease when compared to SBA-15 [20]. Figure 2 illustrated the nitrogen sorption-desorption isotherm curves of the synthesized composites. All the curves exhibited the Langmuir type IV hysteresis loops, suggesting that the synthesized Bi₂₀/SBA-15, Bi₃₀/SBA-15, Bi₄₀/SBA-15, and Bi₅₀/SBA-15 have the characteristics of mesoporous materials with ordered mesoporous structures.

The SEM and TEM images of SBA-15, Bi₄O₅Br₂, and Bi₄O₅Br₂/SBA-15 showed in Figure 3. The image of pure SBA-15 clearly has well-ordered mesoporous hexagonal arrays (Figure 3(a)) and one-dimensional channels of the mesoporous structure (Figure 3(d)) with a diameter of 7.3 nm. The results demonstrated the high ordering level obtained from silica synthesis and confirmed the pore size distribution [21]. Figure 3(b) and Figure 3(e) showed the morphology of Bi₄O₅Br₂ with special nanosheets. No diffraction spots or diffraction rings were observed in Figure 3(d), revealing the non-crystallinity silica characteristics of SBA-15. Figure 3(c) and Figure 3(f) show the images of Bi₃₀/SBA-15 sample. Smaller particles were found to spread over larger particles in all the images, and the larger particles were identified as SBA-15 silica because of their characteristic lines of mesoporous material channels. The smaller particles spread on the SBA-15 reflected the same texture as the larger particles, indicating that they belong to the same type of materials. These agglomerations of nanoparticles were mainly in the range of 100 nm. This observation is consistent with the Bi₄O₅Br₂ phase from the XRD patterns. However, as it can be seen from Figure 3(c) and Figure 3(f), Bi₄O₅Br₂ as rods or spheres were distributed on the surface of the SBA-15. Figure 3(c) and Figure 3(f) also showed that Bi₄O₅Br₂ might be chemically bonded to the surface of SBA-15, not simply loaded on the surface, which may produce stable photocatalytic components thus exhibit photocatalytic activity, thereby increase the possibility of recirculation of the photocatalytic material.

In order to study the elemental composition and valence state of the photocatalytic materials, XPS analysis of Bi₃₀/SBA-15 was performed (Figure 4). The survey XPS spectrum of Bi₃₀/SBA-15 confirmed the existents of Si, C, O, Bi, and Br elements. The appearance of element C may be caused by CO₂ from the sample absorption of air during the analysis [22]. In Figure 4(b), the peak at 103.6 eV is corresponding to Si 2p spectrum, which is the characteristic peak of Si-O-Si band in SiO₂ of SBA-15 and the characteristic peak of Si⁴⁺ [23] [24] [25]. A high resolution energy spectrum of O 1s with two main peaks (Figure 4(c)),

530.6 and 533.0 eV, respectively, was obtained. The main peak 530.6 eV is assigned to the lattice oxygen of Bi-O bond in [Bi₂O₂]²⁺ in the layered Bi₄O₅Br₂/SBA-15, while the peak at 533.0 eV is chemisorbed oxygen and OH⁻ adsorbed on the surface of the material [26]. The Bi 4f spectrum displayed in Figure 4(d) contains two peaks at binding energy of 159.4 and 164.8 eV, which is attributed to Bi 4f7/2 and 4f5/2 [27] [28]. Figure 4(e) showed that the high-resolution energy spectrum belongs to Br 3d. As its two partial peaks are 68.4 and 69.5 eV [29] [30] [31] [32] , respectively, and its corresponding spectra were 3d5/2 and 3d3/2, thus its valence was −1. Results of the XPS spectra indicated that the measured elements corresponded to the elements in the synthesized samples.

Figure 5 showed the UV-vis DRS of Bi₄O₅Br₂/SBA-15 at different loading ratio of the two major materials. As seen, all the samples had a certain absorption under UV light and the absorption efficiency of Bi₂₀/SBA-15 was the highest. As shown in Figure 5(a), all the composite samples exhibited significant absorption sidebands in the visible light range around 410, 415, 430, and 435 nm, respectively. In addition, with the increase of the loading ratio, the absorption wavelength of visible light by the photocatalytic material was red-shifted, and the light response range was expanded. Among them, the bandgap of Bi₅₀/SBA-15 was determined to be the narrowest, resulting in the highest utilization rate of visible light. This could be further analyzed by using the Kubelka-Munk formula [33] :

α h v = A ( h v − E g ) n / 2

where α is the light absorption index, namely Abs, h is the Prandtl constant, v is the frequency, A is the constant, E_g is the semiconductor band gap energy. Since Bi₄O₅Br₂ is an indirect bandgap semiconductor, n = 2. E_g = hv = 1240/λ, the unit is electron volts (eV), E_g is the abscissa, (αhv)^1/n was plotted on the ordinate, and band gap energy spectra of different samples were obtained (Figure 5(b)). It can be seen from Figure 5(b) that the bandgap energy of the samples decreased from about 3.30 eV to 3.10 eV with the increase of Bi₄O₅Br₂ loading ratio, and this may be caused by the agglomeration of nanoparticles [34].

Figure 6(a) demonstrated the photodegradation curves and degradation curves for Bi₂₀/SBA-15, Bi₃₀/SBA-15, Bi₄₀/SBA-15, and Bi₅₀/SBA-15 at 25˚C, respectively. The abscissa was the reaction time for dark adsorption and photodegradation, and the ordinate was the concentration of RhB remained. From Figure 6(a), it can be seen that the Bi₄O₅Br₂ at different loading ratio possessed strong absorption capacity for RhB, which could remove RhB from 30% to 65% within 1 h of adsorption in dark, following a descending order of Bi₃₀/SBA-15 > Bi₂₀/SBA-15 > Bi₄₀/SBA-15 > Bi₅₀/SBA-15. The residual RhB in solution was removed by photodegradation under visible light. During the whole reaction process, the removal of RhB with Bi₃₀/SBA-15 was the fastest. Although the band gaps of Bi₅₀/SBA-15 and Bi₄₀/SBA-15 were narrower than that of Bi_30/SBA-15, the RhB removal rates for Bi₅₀/SBA-15 and Bi₄₀/SBA-15 were lower than that for Bi₃₀/SBA-15. It might be the structure of SBA-15 was destroyed with big size nanoparticles which blocked the porous for adsorption of RhB onto the surface of SBA-15.

Figure 6(b) showed the fluorescence spectra of synthesized materials at different loading ratio of Bi₄O₅Br₂. The test was conducted at room temperature with the excitation wavelength of 250 nm. It can be seen that the peak shapes and positions of the four samples are similar, with the largest peak around 432 nm and another two weak peaks, indicating that the Bi₄O₅Br₂/SBA-15 was excited at 432 nm by the sunlight with the occurrence of electronic transitions and generation of photogenerated electron-hole pairs. Photoluminescence spectroscopy is an important approach to characterize the photoelectron-hole recombination rate of photocatalysts. The recombination rate of photoelectron-holes is proportional to the light intensity. Therefore, the photogenerated electron-holes of photocatalysts can be compared by using the light intensities. The light intensity of Bi₅₀/SBA-15 was the largest, followed by Bi₄₀/SBA-15, Bi₃₀/SBA-15, and Bi₂₀/SBA-15. The separation effect of Bi₃₀/SBA-15 was relatively high, which might be the major reason for its better photocatalytic effect than others. The

reason was the active site was present inside and outside the SBA-15 first increasing with Bi₄O₅Br₂ loading increase, which indicated the separation rate of electron and hole was enhanced, and then decreased which explored that the active site decrease with excess loading of nanoparticles of Bi₄O₅Br₂ blocked the porous.

Bi₄O₅Br₂ is a lamellar structure composed of a [Bi₂O₂]²⁺ layer and a double Br-layer interlaced, with an internal electrostatic field established between the layers. When a semiconductor was excited under visible light irradiation, photocurrent could be formed if the photoelectrons move in a fixed direction. A stronger photocurrent indicated more effective separation of electron-holes and higher photocatalytic performance of the semiconductor as well. In the photocatalytic decolorization of RhB, the capture agents SO, IPA and p-BQ were individually added to capture the active species h⁺, •OH and • O 2 − (Figure 7(a)). As shown in Figure 7(a), the photocatalytic activity of Bi₃₀/SBA-15 slightly decreased when the capture agent IPA was added. In the case of capture agent p-BQ, the photocatalytic activity of the synthesized photocatalyst was significantly reduced after its addition in the photocatalytic process. When adding SO, the photocatalyst showed almost no effect on the decolorization of RhB, implying the catalytic activity was significantly inhibited. The result indicated that h+ played the main role in RhB degradation under visible light.

For evaluation on the stability of the photocatalytic activity of the prepared Bi₄O₅Br₂/SBA-15, sample Bi₃₀/SBA-15 was selected to remove RhB for 5 times (Figure 7(b)). Results showed that after 5 cycles of reutilization of Bi₃₀/SBA-15, its photocatalytic activity remained almost unchanged, and about 99.42% RhB was removed.