Structure-Dependent Photocatalytic Performance of BiOBr x I 1 − x Nanoplate Solid Solutions

BiOXxY1−x (X, Y = Cl, Br, and I) solid solutions have been regarded as promising photocatalysts attributed to their unique layered structure, tunable band structure, and chemical and optical stability. In this study, BiOBrxI1−x nanoplate solid solutions with a high exposure of {001} crystal facets were prepared by a facile alcoholysis method at room temperature and atmospheric pressure. X-ray diffraction (XRD) peaks exhibited a slight shift to lower diffraction angle with the increase of I content in BiOBrxI1−x samples, which resulted in a gradual increase in their cell parameters. Field emission scanning electron microscopy (FESEM) and transmission electron microscope (TEM) images revealed that BiOBrxI1−x samples exhibited 2D plate-like structure with the in-plane wrinkles. The regular changes in optical absorption threshold and Eg value seen in UV-vis diffuse reflectance spectra (UV-vis DRS) indicated that the optical absorption property and band structure could be modulated by the formation of BiOBrxI1−x solid solutions. The photocatalytic degradation of active dye Rhodamine B (RhB) over BiOBrxI1−x solid solutions showed that BiOBr0.75I0.25 had the best photocatalytic activity. The RhB photodegradation processes followed a pseudo-first-order kinetic model. The synergistic effect of structural factors (including amount of exposed {001} facets, interlayer spacing of (001) plane, and energy-level position of the valence band) determined the photocatalytic performance of BiOBrxI1−x solid solutions.


Introduction
Heterogeneous photocatalysis processes have gained more attention in recent decades, attributed to their diverse applications in many fields, such as wastewater remediation, air purification, disinfection, hydrogen production, and catalysis of organic syntheses [1][2][3][4][5][6].The performance of a photocatalyst mainly depends on two factors, namely band gap and redox potential.The former requires that the energy of available photons must be larger than the band gap (E g ) of the photocatalyst; and the latter demands that the redox potential of reactants must be situated between the conduction band minimum (CBM) and the valence band maximum (VBM) of the photocatalyst [7].TiO 2 is a well-known semiconductor photocatalyst.However, the band gaps of rutile and anatase are 3.0 eV and 3.2 eV, respectively, which indicate that they only can be excited by ultraviolet light (about 4% of total solar energy) [8,9].In order to realize the excitation of TiO 2 by sunlight, many efforts have been made to introduce metal or non-metal impurities into TiO 2 structure, which can form the acceptor or donor states in TiO 2 band gap [10].Moreover, numerous novel semiconductors with narrow band gap have been developed to improve the harvesting ability across the visible region, including perovskites, organic metal halides, non-noble metal-based compounds, and transition metal oxides and sulfides [11][12][13][14][15][16][17][18][19][20].
Bismuth oxyhalide (BiOX, X = Cl, Br, and I) semiconductor materials have been regarded as promising photocatalysts because they have a unique layered structure, suitable band structure, and chemical and optical stability [21].Their crystal structures belong to the tetragonal space group P4/nmm and are defined as a matlockite (PbFCl-type) structure with the atomic positions: Bi, 00u, 1 2 1 2 u; X, 00v, 1 2 1 2 v; O, 1  2 00, 0 1 2 0 [22,23].The typical layered structure of bismuth oxyhalides is constituted by [Bi 2 O 2 ] 2+ layers intercalated with double layers of halogen ions.The atoms within [Bi 2 O 2 ] 2+ layers interact each other via covalent bonds, whereas the stack along the c-axis of [X] layers is realized by van der Waals force [24,25].This layered structure facilitates the internal electric fields (IEF) that can promote the effective separation of the photogenerated electron-hole pairs [26].The band gaps of BiOCl, BiOBr, and BiOI are 3.22 eV, 2.64 eV, and 1.77 eV, respectively.This decrease in the band gap with the electronegativity of X decreasing reflects the electronic structures of BiOX, where the CBM and VBM are dominated by the Bi 6p and X np orbitals, respectively [27].Accordingly, the light response of BiOX can be modulated from UV to visible light by changing the halogen ions.Because BiOCl, BiOBr and BiOI have the same layered structure and atomic arrangement, other halogen atoms can be easy to incorporate into their lattice, consequently realizing the substitution of each other to form BiOX x Y 1−x (X, Y = Cl, Br, and I) solid solutions [26].The formation of solid solutions might be an ideal strategy to realize the continuous adjustment of the VBM energy levels.Up until now, a series of BiOCl x Br 1−x [26][27][28][29][30][31], BiOCl x I 1−x [32][33][34][35], and BiOBr x I 1−x [36][37][38][39] solid solutions have been obtained as a highly efficient photocatalyst with a tunable band gap and visible-light response.In addition, it has been reported that BiOX nanoplates with {001} facets exposure showed higher photocatalytic activity than those with {010} facets exposure, attributed to their better adsorption property, stronger IEF, and higher separation and transfer efficiency of photogenerated electron-hole pairs [24, 30,40,41].However, the exposure of a specific crystal facet for BiOX nanomaterials strongly depended on the pH value of their growth environment.Generally, the strong acid condition (such as pH = 1) was mandatory in the hydrothermal route for {001} facet-selective growth of BiOXs [24].Hence, how to realize high {001} facets exposure in a facile and low-cost route still remains a challenging task for BiOXs, especially their solid solutions.Moreover, the formation of solid solutions can result in regular changes of crystal structure, band gap, and IEF of BiOXs and consequently affect their photocatalytic activities.The inherent correlation between crystal structure and enhanced photocatalytic activity of BiOX solid solutions still needs further insight and understanding.
In this study, BiOBr x I 1−x solid solutions were selected as the objectives of study and an alcoholysis method was developed to synthesize BiOBr x I 1−x nanoplates at room temperature and atmospheric pressure.In this route, a strong acid/alkaline environment was absolutely avoided to realize the high exposure of {001} crystal facets.Followed by the characterization analyses by XRD, FESEM, TEM, UV-vis DRS, energy dispersive spectra (EDS), and BET surface area analyzer, the obtained samples were arranged for the assessment of their photocatalytic activities under simulated sunlight irradiation using RhB as the target pollutant.Then, the atomic configuration, cell parameter, interlayer spacing, exposed facet, and energy band structure of BiOBr x I 1−x solid solutions were calculated and discussed with the aim to further understand structure-dependent enhanced photocatalytic activity of BiOBr x I 1−x nanoplate solid solutions.

Crystal Phase and Morphology of BiOBr x I 1−x Solid Solutions
The XRD results of BiOBr x I 1−x samples are presented in Figure 1, where it can be seen that all the diffraction peaks are sharp, indicating successful crystallization of BiOBr x I 1−x samples prepared by the alcoholysis route [42].The two end-member samples, i.e., BiOBr and BiOI are in good accordance with JCPDS standard cards No. 78-348 and 73-2062, respectively.As expected, all the diffraction peaks present a slight shift to lower diffraction angle with the increase of I content in BiOBr x I 1−x samples.This phenomenon can be explained by the fact that after the substitution of smaller Br ions (r Br = 0.196 nm) by larger I ions (r Br = 0.216 nm), the interlayer spacings expand.In the same way, the cell parameters of all the samples gradually increase with I content increasing, as listed in Table 1.These regular changes in XRD peaks and cell parameters as a function of I content strongly declare that the obtained BiOBr x I 1−x products belong to a group of solid solutions [27].In addition, there are no other diffraction peaks of impurity, which implies that all the atoms entered into the crystal lattice and the stoichiometric value x in BiOBr x I 1−x was accurately determined by the initial proportion of Br and I sources in the preparation procedure [36,39].The relative intensities of (001), (002), and (004) peaks gradually become stronger with increasing I content; on the contrary, that of (101) peak gradually weakens.Furthermore, in all the XRD patterns, the relative intensity of (001) peak is stronger than that of (101) peak, suggesting that the exposure percentage of {001} crystal facets is higher than that of {101} crystal facets in all the samples with highly preferred (001) orientation [43].Thus, the as-prepared samples in this work can be determined as BiOBr x I 1−x solid solutions with high {001} facets exposure.
Catalysts 2017, 7, 153 3 of 14 samples.This phenomenon can be explained by the fact that after the substitution of smaller Br ions (rBr = 0.196 nm) by larger I ions (rBr = 0.216 nm), the interlayer spacings expand.In the same way, the cell parameters of all the samples gradually increase with I content increasing, as listed in Table 1.These regular changes in XRD peaks and cell parameters as a function of I content strongly declare that the obtained BiOBrxI1−x products belong to a group of solid solutions [27].In addition, there are no other diffraction peaks of impurity, which implies that all the atoms entered into the crystal lattice and the stoichiometric value x in BiOBrxI1−x was accurately determined by the initial proportion of Br and I sources in the preparation procedure [36,39].The relative intensities of (001), (002), and (004) peaks gradually become stronger with increasing I content; on the contrary, that of (101) peak gradually weakens.Furthermore, in all the XRD patterns, the relative intensity of (001) peak is stronger than that of (101) peak, suggesting that the exposure percentage of {001} crystal facets is higher than that of {101} crystal facets in all the samples with highly preferred (001) orientation [43].Thus, the as-prepared samples in this work can be determined as BiOBrxI1−x solid solutions with high {001} facets exposure.The morphological structure of BiOBrxI1−x solid solutions was observed by FESEM and TEM, as displayed in Figures 2 and 3, respectively.In Figure 2, the FESEM images reveal that all of the BiOBrxI1−x samples exhibit a 2D plate-like structure with a diameter of 0.2-1.0μm and thickness of 40-60 nm.These nanoplates irregularly stack together without a specific architecture, such as microsphere or flower.EDS analysis indicates that typical BiOBrxI1−x (x = 0.5) has the main elements Bi, O, Br, and I. Also, as expected, the atomic ratio of Br and I is approximate to 1.In the TEM images (Figure 3), 2D nanosheet structures of BiOBrxI1−x samples can be clearly seen.These nanosheets are so thin that the overlayed shadow is visible.Moreover, the in-plane wrinkles of the nanosheets can be distinctly observed, which might be attributed to the existence of inner strain in the BiOBrxI1−x structure.The inner strain can change the band structure and facilitate the charge separation, accordingly improving the photocatalytic activity of BiOX [44].The morphological structure of BiOBr x I 1−x solid solutions was observed by FESEM and TEM, as displayed in Figures 2 and 3, respectively.In Figure 2, the FESEM images reveal that all of the BiOBr x I 1−x samples exhibit a 2D plate-like structure with a diameter of 0.2-1.0µm and thickness of 40-60 nm.These nanoplates irregularly stack together without a specific architecture, such as microsphere or flower.EDS analysis indicates that typical BiOBr x I 1−x (x = 0.5) has the main elements Bi, O, Br, and I. Also, as expected, the atomic ratio of Br and I is approximate to 1.In the TEM images (Figure 3), 2D nanosheet structures of BiOBr x I 1−x samples can be clearly seen.These nanosheets are so thin that the overlayed shadow is visible.Moreover, the in-plane wrinkles of the nanosheets can be distinctly observed, which might be attributed to the existence of inner strain in the BiOBr x I 1−x structure.The inner strain can change the band structure and facilitate the charge separation, accordingly improving the photocatalytic activity of BiOX [44].

Optical Absorption and Photocatalytic Properties
The optical properties of BiOBrxI1−x solid solutions were investigated by UV-vis DRS spectra and the results are illustrated in Figure 4a.It can be obviously seen that for the end-member samples, the absorption threshold of BiOBr is about 438.5 nm, whereas this value of BiOI is near 674.1 nm, suggesting that both of them can be excited by visible light.As expected, the absorption thresholds of the other three BiOBrxI1−x solid solutions are all located between those of BiOBr and BiOI, and regularly increase with I content increasing.The optical band gap (Eg) of a semiconductor photocatalyst can be determined by extrapolation of the linear portion of [F(R)•hυ] 0.5 curve versus photon energy hυ to [F(R)•hυ] 0.5 = 0 [45].Based on the above UV-vis DRS results, Eg values of BiOBrxI1−x solid solutions were obtained as shown in Figure 4b and Table 1.Interestingly, as I content increases, Eg values of BiOBrxI1−x samples gradually decrease.This also indicates that the as-prepared products are a group of BiOBrxI1−x solid solutions.The regular changes in optical absorption threshold and Eg value strongly state that the substitution of smaller Br ions by larger I ions can modify the optical absorption property and band structure of BiOBrxI1−x solid solutions.It has been verified that, for BiOX semiconductors, the CBM is dominated by Bi 6p orbitals and the VBM mainly originates from O 2p and X np orbitals (n = 3, 4 and 5 for X = Cl, Br and I, respectively) [24,30].Therefore, as I content increases, the contribution of I 5p orbitals dramatically increases and, correspondingly, the band gap is narrowed down.

Optical Absorption and Photocatalytic Properties
The optical properties of BiOBrxI1−x solid solutions were investigated by UV-vis DRS spectra and the results are illustrated in Figure 4a.It can be obviously seen that for the end-member samples, the absorption threshold of BiOBr is about 438.5 nm, whereas this value of BiOI is near 674.1 nm, suggesting that both of them can be excited by visible light.As expected, the absorption thresholds of the other three BiOBrxI1−x solid solutions are all located between those of BiOBr and BiOI, and regularly increase with I content increasing.The optical band gap (Eg) of a semiconductor photocatalyst can be determined by extrapolation of the linear portion of [F(R)•hυ] 0.5 curve versus photon energy hυ to [F(R)•hυ] 0.5 = 0 [45].Based on the above UV-vis DRS results, Eg values of BiOBrxI1−x solid solutions were obtained as shown in Figure 4b and Table 1.Interestingly, as I content increases, Eg values of BiOBrxI1−x samples gradually decrease.This also indicates that the as-prepared products are a group of BiOBrxI1−x solid solutions.The regular changes in optical absorption threshold and Eg value strongly state that the substitution of smaller Br ions by larger I ions can modify the optical absorption property and band structure of BiOBrxI1−x solid solutions.It has been verified that, for BiOX semiconductors, the CBM is dominated by Bi 6p orbitals and the VBM mainly originates from O 2p and X np orbitals (n = 3, 4 and 5 for X = Cl, Br and I, respectively) [24,30].Therefore, as I content increases, the contribution of I 5p orbitals dramatically increases and, correspondingly, the band gap is narrowed down.

Optical Absorption and Photocatalytic Properties
The optical properties of BiOBr x I 1−x solid solutions were investigated by UV-vis DRS spectra and the results are illustrated in Figure 4a.It can be obviously seen that for the end-member samples, the absorption threshold of BiOBr is about 438.5 nm, whereas this value of BiOI is near 674.1 nm, suggesting that both of them can be excited by visible light.As expected, the absorption thresholds of the other three BiOBr x I 1−x solid solutions are all located between those of BiOBr and BiOI, and regularly increase with I content increasing.The optical band gap (E g ) of a semiconductor photocatalyst can be determined by extrapolation of the linear portion of [F(R)•hυ] 0.5 curve versus photon energy hυ to [F(R)•hυ] 0.5 = 0 [45].Based on the above UV-vis DRS results, E g values of BiOBr x I 1−x solid solutions were obtained as shown in Figure 4b and Table 1.Interestingly, as I content increases, E g values of BiOBr x I 1−x samples gradually decrease.This also indicates that the as-prepared products are a group of BiOBr x I 1−x solid solutions.The regular changes in optical absorption threshold and E g value strongly state that the substitution of smaller Br ions by larger I ions can modify the optical absorption property and band structure of BiOBr x I 1−x solid solutions.It has been verified that, for BiOX semiconductors, the CBM is dominated by Bi 6p orbitals and the VBM mainly originates from O 2p and X np orbitals (n = 3, 4 and 5 for X = Cl, Br and I, respectively) [24,30].Therefore, as I content increases, the contribution of I 5p orbitals dramatically increases and, correspondingly, the band gap is narrowed down.The photocatalytic properties of BiOBrxI1−x solid solutions were investigated by evaluating the degradation of RhB under simulated sunlight irradiation.Figure 5a illustrates the photocatalytic degradation of RhB over BiOBrxI1−x solid solutions with 60 min adsorption in dark followed by 180 min photocatalysis under simulated sunlight irradiation.It is well known that the adsorption plays an essential role in the heterogeneous photocatalytic process [45].The results clearly reveal that the adsorbed amounts of RhB onto BiOBrxI1−x samples with different x values are almost alike at the adsorption/desorption equilibrium, consequently, the influence of adsorption capacity on photocatalytic evaluation can be neglected.It can also be found that BiOBrxI1−x solid solutions with x = 0.25, 0.5, and 0.75 exhibit higher photocatalytic activity than the end-member BiOBr and BiOI.Only 31.6% and 77.7% of RhB can be degraded over BiOI and BiOBr, respectively, within 240 min reaction time; nevertheless, almost all the RhB can be degraded over BiOBr0.75I0.25 and BiOBr0.5I0.5 within 210 and 240 min reaction time, respectively.The BET specific surface areas of BiOBr, BiOBr0.75I0.25,BiOBr0.5I0.5, BiOBr0.25I0.75, and BiOI were determined to be 7.14, 6.71, 7.30, 6.19, and 6.28 m 2 /g, respectively.Since there is no obvious difference in their surface areas, the enhanced photocatalytic activities of BiOBr0.75I0.25,BiOBr0.5I0.5, and BiOBr0.25I0.75 are not essentially related to the surface area.It is very obvious that, among them, BiOBr0.75I0.25 exhibits the best photocatalytic activity.This conclusion is similar to Zhang's report [39].Furthermore, Figure 5b indicates that the RhB photodegradation processes followed a pseudo-first-order kinetic equation: where C0 is the initial RhB concentration, C is the residual RhB concentration at time t, and the slope k is the apparent reaction rate constant.The calculated apparent reaction rate constants are listed is Table 1.The k value of BiOBr0.75I0.25 is approximately 15.1 times higher than that of BiOI and approximately 3.4 times higher than that of BiOBr, indicating an enhanced photocatalytic activity of BiOBrxI1−x solid solutions with x = 0.75.The photocatalytic properties of BiOBr x I 1−x solid solutions were investigated by evaluating the degradation of RhB under simulated sunlight irradiation.Figure 5a illustrates the photocatalytic degradation of RhB over BiOBr x I 1−x solid solutions with 60 min adsorption in dark followed by 180 min photocatalysis under simulated sunlight irradiation.It is well known that the adsorption plays an essential role in the heterogeneous photocatalytic process [45].The results clearly reveal that the adsorbed amounts of RhB onto BiOBr x I 1−x samples with different x values are almost alike at the adsorption/desorption equilibrium, consequently, the influence of adsorption capacity on photocatalytic evaluation can be neglected.It can also be found that BiOBr x I 1−x solid solutions with x = 0.25, 0.5, and 0.75 exhibit higher photocatalytic activity than the end-member BiOBr and BiOI.Only 31.6% and 77.7% of RhB can be degraded over BiOI and BiOBr, respectively, within 240 min reaction time; nevertheless, almost all the RhB can be degraded over BiOBr 0.75 I 0.25 and BiOBr 0.5 I 0.5 within 210 and 240 min reaction time, respectively.The BET specific surface areas of BiOBr, BiOBr 0.75 I 0.25 , BiOBr 0.5 I 0.5 , BiOBr 0.25 I 0.75 , and BiOI were determined to be 7.14, 6.71, 7.30, 6.19, and 6.28 m 2 /g, respectively.Since there is no obvious difference in their surface areas, the enhanced photocatalytic activities of BiOBr 0.75 I 0.25 , BiOBr 0.5 I 0.5 , and BiOBr 0.25 I 0.75 are not essentially related to the surface area.It is very obvious that, among them, BiOBr 0.75 I 0.25 exhibits the best photocatalytic activity.This conclusion is similar to Zhang's report [39].Furthermore, Figure 5b indicates that the RhB photodegradation processes followed a pseudo-first-order kinetic equation: where C 0 is the initial RhB concentration, C is the residual RhB concentration at time t, and the slope k is the apparent reaction rate constant.The calculated apparent reaction rate constants are listed is Table 1.
The k value of BiOBr 0.75 I 0.25 is approximately 15.1 times higher than that of BiOI and approximately 3.4 times higher than that of BiOBr, indicating an enhanced photocatalytic activity of BiOBr x I 1−x solid solutions with x = 0.75.The photocatalytic properties of BiOBrxI1−x solid solutions were investigated by evaluating the degradation of RhB under simulated sunlight irradiation.Figure 5a illustrates the photocatalytic degradation of RhB over BiOBrxI1−x solid solutions with 60 min adsorption in dark followed by 180 min photocatalysis under simulated sunlight irradiation.It is well known that the adsorption plays an essential role in the heterogeneous photocatalytic process [45].The results clearly reveal that the adsorbed amounts of RhB onto BiOBrxI1−x samples with different x values are almost alike at the adsorption/desorption equilibrium, consequently, the influence of adsorption capacity on photocatalytic evaluation can be neglected.It can also be found that BiOBrxI1−x solid solutions with x = 0.25, 0.5, and 0.75 exhibit higher photocatalytic activity than the end-member BiOBr and BiOI.Only 31.6% and 77.7% of RhB can be degraded over BiOI and BiOBr, respectively, within 240 min reaction time; nevertheless, almost all the RhB can be degraded over BiOBr0.75I0.25 and BiOBr0.5I0.5 within 210 and 240 min reaction time, respectively.The BET specific surface areas of BiOBr, BiOBr0.75I0.25,BiOBr0.5I0.5, BiOBr0.25I0.75, and BiOI were determined to be 7.14, 6.71, 7.30, 6.19, and 6.28 m 2 /g, respectively.Since there is no obvious difference in their surface areas, the enhanced photocatalytic activities of BiOBr0.75I0.25,BiOBr0.5I0.5, and BiOBr0.25I0.75 are not essentially related to the surface area.It is very obvious that, among them, BiOBr0.75I0.25 exhibits the best photocatalytic activity.This conclusion is similar to Zhang's report [39].Furthermore, Figure 5b indicates that the RhB photodegradation processes followed a pseudo-first-order kinetic equation: where C0 is the initial RhB concentration, C is the residual RhB concentration at time t, and the slope k is the apparent reaction rate constant.The calculated apparent reaction rate constants are listed is Table 1.The k value of BiOBr0.75I0.25 is approximately 15.1 times higher than that of BiOI and approximately 3.  The chemical oxygen demand (COD) of RhB solution after different photodegradation periods over BiOBr 0.75 I 0.25 was detected and the results are displayed in Figure 6a.It can be seen that more than 90% COD in RhB solution can be eliminated after 210 min reaction, which implies that most RhB molecules can be decomposed into H 2 O and CO 2 in this photocatalytic system.Moreover, the UV-vis absorption peaks of RhB solution after different photodegradation periods gradually weaken (Figure 6b), also indicating an almost complete decomposition of RhB molecules.Thereby, it can be inferred that almost all the RhB molecules are decomposed into H 2 O, CO 2 , and small molecule organics.Photocatalytic stability is another important parameter to evaluate a photocatalyst.The recyclability of BiOBr x I 1−x solid solution photocatalysts was evaluated, also using BiOBr 0.75 I 0.25 as the representative sample.Figure 7a reveals that, even after 5 runs, BiOBr 0.75 I 0.25 photocatalyst still has high photocatalytic activity.As shown in Figure 7b, the XRD result indicates that BiOBr 0.75 I 0.25 sample after being circularly used 5 times maintains the crystal structure of bismuth oxyhalide without any destruction.These suggest that BiOBr x I 1−x solid solutions not only have excellent photocatalytic stability but also possess good structural stability.
Catalysts 2017, 7, 153 6 of 14 The chemical oxygen demand (COD) of RhB solution after different photodegradation periods over BiOBr0.75I0.25 was detected and the results are displayed in Figure 6a.It can be seen that more than 90% COD in RhB solution can be eliminated after 210 min reaction, which implies that most RhB molecules can be decomposed into H2O and CO2 in this photocatalytic system.Moreover, the UV-vis absorption peaks of RhB solution after different photodegradation periods gradually weaken (Figure 6b), also indicating an almost complete decomposition of RhB molecules.Thereby, it can be inferred that almost all the RhB molecules are decomposed into H2O, CO2, and small molecule organics.Photocatalytic stability is another important parameter to evaluate a photocatalyst.The recyclability of BiOBrxI1−x solid solution photocatalysts was evaluated, also using BiOBr0.75I0.25 as the representative sample.Figure 7a reveals that, even after 5 runs, BiOBr0.75I0.25 photocatalyst still has high photocatalytic activity.As shown in Figure 7b, the XRD result indicates that BiOBr0.75I0.25 sample after being circularly used 5 times maintains the crystal structure of bismuth oxyhalide without any destruction.These suggest that BiOBrxI1−x solid solutions not only have excellent photocatalytic stability but also possess good structural stability.

Structure-Dependent Photocatalytic Performance
In a typical photocatalytic process, three key steps are involved in the charge kinetics.Firstly, electron-hole pairs are generated under photoexcitation; then, the photogenerated charges transfer to the catalyst surface; finally, these charges participate in the redox reactions on the catalyst surface.Thus, in order to construct highly efficient photocatalysts, the aim of structure engineering is mainly to improve the efficiency of each step in the charge kinetics.To this end, the structure engineering specifically includes surface engineering, interface engineering, and energy band engineering [46].The chemical oxygen demand (COD) of RhB solution after different photodegradation periods over BiOBr0.75I0.25 was detected and the results are displayed in Figure 6a.It can be seen that more than 90% COD in RhB solution can be eliminated after 210 min reaction, which implies that most RhB molecules can be decomposed into H2O and CO2 in this photocatalytic system.Moreover, the UV-vis absorption peaks of RhB solution after different photodegradation periods gradually weaken (Figure 6b), also indicating an almost complete decomposition of RhB molecules.Thereby, it can be inferred that almost all the RhB molecules are decomposed into H2O, CO2, and small molecule organics.Photocatalytic stability is another important parameter to evaluate a photocatalyst.The recyclability of BiOBrxI1−x solid solution photocatalysts was evaluated, also using BiOBr0.75I0.25 as the representative sample.Figure 7a reveals that, even after 5 runs, BiOBr0.75I0.25 photocatalyst still has high photocatalytic activity.As shown in Figure 7b, the XRD result indicates that BiOBr0.75I0.25 sample after being circularly used 5 times maintains the crystal structure of bismuth oxyhalide without any destruction.These suggest that BiOBrxI1−x solid solutions not only have excellent photocatalytic stability but also possess good structural stability.

Structure-Dependent Photocatalytic Performance
In a typical photocatalytic process, three key steps are involved in the charge kinetics.Firstly, electron-hole pairs are generated under photoexcitation; then, the photogenerated charges transfer to the catalyst surface; finally, these charges participate in the redox reactions on the catalyst surface.Thus, in order to construct highly efficient photocatalysts, the aim of structure engineering is mainly to improve the efficiency of each step in the charge kinetics.To this end, the structure engineering specifically includes surface engineering, interface engineering, and energy band engineering [46].

Structure-Dependent Photocatalytic Performance
In a typical photocatalytic process, three key steps are involved in the charge kinetics.Firstly, electron-hole pairs are generated under photoexcitation; then, the photogenerated charges transfer to the catalyst surface; finally, these charges participate in the redox reactions on the catalyst surface.Thus, in order to construct highly efficient photocatalysts, the aim of structure engineering is mainly to improve the efficiency of each step in the charge kinetics.To this end, the structure engineering specifically includes surface engineering, interface engineering, and energy band engineering [46].Since the facets constitute the surface and interface of a photocatalyst, they have strong influence on the structure and property of surface and interface, and should be rationally selected.The facets exposed on a photocatalyst surface may affect the photocatalytic performance through various working mechanisms [47].The photocatalytic performance of BiOBr x I 1−x solid solutions strongly depends on their crystal structures as well.
The typical layered structures of BiOX along the band c-axes are illustrated in Figure 8a,b, respectively.In this structure, all the atoms of the asymmetric unit lie on special positions, 4 mm or 4 m 2 .One bismuth atom is coordinated with four oxygen and four halogen atoms to form an asymmetric decahedral geometry, which is linked with each other by sharing the O-X edge along the aand b-axes.[Bi 2 O 2 ] 2+ layers are interleaved by double layers of halogen ions.The inhomogeneity of charge distribution between [Bi 2 O 2 ] 2+ and halogen layers in this asymmetric structure can form the IEF along the c-axis (Figure 8c) [24].IEF can efficiently separate the photogenerated electrons and holes, drive them to [Bi 2 O 2 ] and [X] layers respectively, and accommodate them within [Bi 2 O 2 ] and [X] layers to prevent their recombination [48].Moreover, the interlayer van der Waals interactions and intralayer covalent bonding can facilitate the preparation of BiOX nanoplates with highly exposed {001} facets [30].It has been verified that the IEF magnitude of BiOX nanoplates is proportional to their exposure amounts of {001} facets [41].Hence, the higher photocatalytic activity of {001} facet-dominant BiOX nanoplates profits from stronger IEF that can promote electron-hole transfer and inhibit their recombination [24].
Catalysts 2017, 7, 153 7 of 14 Since the facets constitute the surface and interface of a photocatalyst, they have strong influence on the structure and property of surface and interface, and should be rationally selected.The facets exposed on a photocatalyst surface may affect the photocatalytic performance through various working mechanisms [47].The photocatalytic performance of BiOBrxI1−x solid solutions strongly depends on their crystal structures as well.
The typical layered structures of BiOX along the b-and c-axes are illustrated in Figure 8a,b, respectively.In this structure, all the atoms of the asymmetric unit lie on special positions, 4 mm or 4 _ m 2 .One bismuth atom is coordinated with four oxygen and four halogen atoms to form an asymmetric decahedral geometry, which is linked with each other by sharing the O-X edge along the a-and b-axes.[Bi2O2] 2+ layers are interleaved by double layers of halogen ions.The inhomogeneity of charge distribution between [Bi2O2] 2+ and halogen layers in this asymmetric structure can form the IEF along the c-axis (Figure 8c) [24].IEF can efficiently separate the photogenerated electrons and holes, drive them to [Bi2O2] and [X] layers respectively, and accommodate them within [Bi2O2] and [X] layers to prevent their recombination [48].Moreover, the interlayer van der Waals interactions and intralayer covalent bonding can facilitate the preparation of BiOX nanoplates with highly exposed {001} facets [30].It has been verified that the IEF magnitude of BiOX nanoplates is proportional to their exposure amounts of {001} facets [41].Hence, the higher photocatalytic activity of {001} facet-dominant BiOX nanoplates profits from stronger IEF that can promote electron-hole transfer and inhibit their recombination [24].By carefully examining the XRD patterns in this study, we found that BiOBrxI1−x solid solutions with different x values had different diffraction intensities of (001) peak.In regard to BiOX crystals, the exposure amount of {001} facets should be in direct proportion to the relative intensity of (001) peak [43].Analogous to the method reported by Zhang's group [49], the intensity ratio of the (001) and (102) peaks were employed to approximately estimate the exposure amount of {001} facets in each sample.Figure 9a obviously indicates that the {001} facets exposure amount for BiOBrxI1−x samples is in the following order: BiOI > BiOBr0.5I0.5 > BiOBr0.25I0.75> BiOBr0.75I0.25 > BiOBr.Because the IEF magnitude is in direct proportion to the exposure amounts of {001} facets, it can be deemed that the IEF magnitude of BiOBrxI1−x samples is in the same sequence.Hereby, it is plausible that BiOI should have the best photocatalytic activity among these samples.However, the experimental data generated in this study does not follow this trend, as shown in Figure 5a.The possible explanation for this phenomenon might be ascribed to two aspects.On one hand, according to Bragg's formula, λ = 2dsinθ, the regular shift of (001) peak to smaller diffraction angle indicates an increase in the interlayer spacing of (001) crystal plane, as I content in BiOBrxI1−x solid solutions increases.Figure 9b displays the gradual increase in interlayer spacing of (001) crystal plane, d001.This result provides By carefully examining the XRD patterns in this study, we found that BiOBr x I 1−x solid solutions with different x values had different diffraction intensities of (001) peak.In regard to BiOX crystals, the exposure amount of {001} facets should be in direct proportion to the relative intensity of (001) peak [43].Analogous to the method reported by Zhang's group [49], the intensity ratio of the (001) and (102) peaks were employed to approximately estimate the exposure amount of {001} facets in each sample.Figure 9a obviously indicates that the {001} facets exposure amount for BiOBr x I 1−x samples is in the following order: BiOI > BiOBr 0.5 I 0.5 > BiOBr 0.25 I 0.75 > BiOBr 0.75 I 0.25 > BiOBr.Because the IEF magnitude is in direct proportion to the exposure amounts of {001} facets, it can be deemed that the IEF magnitude of BiOBr x I 1−x samples is in the same sequence.Hereby, it is plausible that BiOI should have the best photocatalytic activity among these samples.However, the experimental data generated in this study does not follow this trend, as shown in Figure 5a.The possible explanation for this phenomenon might be ascribed to two aspects.On one hand, according to Bragg's formula, λ = 2dsinθ, the regular shift of (001) peak to smaller diffraction angle indicates an increase in the interlayer spacing of (001) crystal plane, as I content in BiOBr x I 1−x solid solutions increases.Figure 9b displays the gradual increase in interlayer spacing of (001) crystal plane, d 001 .This result provides strong evidence to certify the formation of intercalated layered products [50].Importantly, the increase in d 001 would make the distance longer for charge transport driven by IEF, reducing the effective separation of photogenerated electron-hole pairs and thereby decreasing the photocatalytic activity.So, the largest (001) interlayer spacing of BiOI negatively affects its photocatalytic performance.
Catalysts 2017, 7, 153 8 of 14 strong evidence to certify the formation of intercalated layered products [50].Importantly, the increase in d001 would make the distance longer for charge transport driven by IEF, reducing the effective separation of photogenerated electron-hole pairs and thereby decreasing the photocatalytic activity.So, the largest (001) interlayer spacing of BiOI negatively affects its photocatalytic performance.
(a) (b) On the other hand, the different energy levels of hybrid orbitals can be formed due to the different molar ratios of Br and I in BiOBrxI1−x solid solutions, which means that their band structures can be modulated by changing the molar ratio of Br and I.In order to further know the relationship between band structure and photocatalytic activity of BiOBrxI1−x samples, the energy-level positions of VBM and CBM were calculated by Equations ( 2) and ( 3) [51][52][53]: where, ECB and EVB are the energy levels of CBM and VBM, respectively; χ is the absolute electronegativity of a semiconductor, namely the geometric mean value of the absolute electronegativity of each atom in the semiconductor; E° is the scale factor relating the reference electrode redox level to the absolute vacuum scale, determined to be −4.5 eV for normal hydrogen electrode; and Eg is the band-gap energy of the semiconductor.The calculated results are listed in Table 1.It can be obviously seen in Figure 10 that CBM potentials of all the samples are positive, implying that the photogenerated electrons in conduction band cannot react with O2 to form•O2 − radicals (O2/O2 − = −0.28V); nevertheless, VBM potentials of these samples are all more positive than +1.99 V, indicating that the photogenerated holes in the valence band are anodic enough to oxidize H2O or OH − into OH radicals (OH/OH − = +1.99V) [54].Meanwhile, as x value increases, the potential of VBM gradually increases.The deeper VBM position can result in higher oxidation power of the photogenerated holes [55], which can directly oxidize the organic molecules or react with adsorbed H2O and surface OH − to form highly reactive OH radicals.Hence, BiOI exhibits the worst photocatalytic oxidation power among BiOBrxI1−x solid solutions, attributed to its lowest VBM potential.On the other hand, the different energy levels of hybrid orbitals can be formed due to the different molar ratios of Br and I in BiOBr x I 1−x solid solutions, which means that their band structures can be modulated by changing the molar ratio of Br and I.In order to further know the relationship between band structure and photocatalytic activity of BiOBr x I 1−x samples, the energy-level positions of VBM and CBM were calculated by Equations ( 2) and ( 3) [51][52][53]: where, E CB and E VB are the energy levels of CBM and VBM, respectively; χ is the absolute electronegativity of a semiconductor, namely the geometric mean value of the absolute electronegativity of each atom in the semiconductor; E • is the scale factor relating the reference electrode redox level to the absolute vacuum scale, determined to be −4.5 eV for normal hydrogen electrode; and E g is the band-gap energy of the semiconductor.The calculated results are listed in Table 1.
It can be obviously seen in Figure 10  H2O or OH − into OH radicals (OH/OH − = +1.99V) [54].Meanwhile, as x value increases, the potential of VBM gradually increases.The deeper VBM position can result in higher oxidation power of the photogenerated holes [55], which can directly oxidize the organic molecules or react with adsorbed H2O and surface OH − to form highly reactive OH radicals.Hence, BiOI exhibits the worst photocatalytic oxidation power among BiOBrxI1−x solid solutions, attributed to its lowest VBM potential.According to above discussions, it can be concluded that the photocatalytic performance of BiOBr x I 1−x solid solutions is closely dependent on their structures, including exposure amount of {001} facets, interlayer spacing of (001) plane, and electronic band structure.When BiOBr x I 1−x semiconductor is irradiated by simulated sun light, the photogenerated electrons and holes can be produced, which are the main initiators to form the highly oxidative species.Nevertheless, they are easy to recombine together, consequently reducing the photocatalytic activity.As shown in Figure 11, driven by the IEF, photogenerated electrons and holes can be separated and transported to [Bi 2 O 2 ] and [X] layers, respectively.Since the IEF is perpendicular to (001) crystal plane, more exposure amount of {001} facets can lead to a larger IEF, which can more efficiently induce the separation and transport of photogenerated charges.However, larger interlayer spacing of (001) plane is not conducive to the separation and transport of photogenerated charges because of the longer distance along the IEF.Moreover, the deeper energy-level position of VB can make the photogenerated holes have more powerful oxidation ability.Therefore, the synergistic effect of these three structural factors determines the photocatalytic performance of BiOBr x I 1−x solid solutions.According to above discussions, it can be concluded that the photocatalytic performance of BiOBrxI1−x solid solutions is closely dependent on their structures, including exposure amount of {001} facets, interlayer spacing of (001) plane, and electronic band structure.When BiOBrxI1−x semiconductor is irradiated by simulated sun light, the photogenerated electrons and holes can be produced, which are the main initiators to form the highly oxidative species.Nevertheless, they are easy to recombine together, consequently reducing the photocatalytic activity.As shown in Figure 11, driven by the IEF, photogenerated electrons and holes can be separated and transported to [Bi2O2] and [X] layers, respectively.Since the IEF is perpendicular to (001) crystal plane, more exposure amount of {001} facets can lead to a larger IEF, which can more efficiently induce the separation and transport of photogenerated charges.However, larger interlayer spacing of (001) plane is not conducive to the separation and transport of photogenerated charges because of the longer distance along the IEF.Moreover, the deeper energy-level position of VB can make the photogenerated holes have more powerful oxidation ability.Therefore, the synergistic effect of these three structural factors determines the photocatalytic performance of BiOBrxI1−x solid solutions.

Preparation of BiOBrxI1−x Solid Solutions
All the chemical reagents used in this work were of analytical grade and were without further purification.An alcoholysis method with facile operation and low cost was employed for the preparation of BiOBrxI1−x nanoplate solid solutions at room temperature and atmospheric pressure.Bismuth nitrate pentahydrate (Bi(NO3)3•5H2O, Tianjin Ruijinte Chemical Reagent Co., Ltd., Tianjin, China), potassium bromide (KBr, Tianjin Zhiyuan Chemical Reagent Co., Ltd., Tianjin, China), and potassium iodide (KI, Tianjin Tianda Chemical Reagent Co., Ltd., Tianjin, China) were used as Bi, Br, and I sources, respectively.In a typical procedure, certain amounts of Bi(NO3)3•5H2O were put into a 200 mL beaker and some volume of absolute ethyl alcohol (Tianjin Tianli Chemical Reagent Co., Ltd., Tianjin, China) was subsequently injected.Then, the resulting suspensions underwent an alcoholysis process under ultrasonic irradiation until Bi(NO3)3•5H2O was completely dissolved in absolute ethyl

Preparation of BiOBr x I 1−x Solid Solutions
All the chemical reagents used in this work were of analytical grade and were without further purification.An alcoholysis method with facile operation and low cost was employed for the preparation of BiOBr x I 1−x nanoplate solid solutions at room temperature and atmospheric pressure.Bismuth nitrate pentahydrate (Bi(NO 3 ) 3 •5H 2 O, Tianjin Ruijinte Chemical Reagent Co., Ltd., Tianjin, China), potassium bromide (KBr, Tianjin Zhiyuan Chemical Reagent Co., Ltd., Tianjin, China), and potassium iodide (KI, Tianjin Tianda Chemical Reagent Co., Ltd., Tianjin, China) were used as Bi, Br, and I sources, respectively.In a typical procedure, certain amounts of Bi(NO 3 ) 3 •5H 2 O were put into a 200 mL beaker and some volume of absolute ethyl alcohol (Tianjin Tianli Chemical Reagent Co., Ltd., Tianjin, China) was subsequently injected.Then, the resulting suspensions underwent an alcoholysis process under ultrasonic irradiation until Bi(NO 3 ) 3 •5H 2 O was completely dissolved in absolute ethyl alcohol.At the same time, according to the stoichiometric value x (0, 0.25, 0.5, 0.75, and 1.0) in BiOBr x I 1−x solid solutions, some amount of KBr and KI was weighed and dissolved in a small quantity of deionized water.Afterwards, this solution was added to the above obtained alcoholysate and the mixture was mechanically stirred for 6 h at room temperature.Thereafter, the final suspension was centrifuged and the resulting precipitate was washed by absolute ethyl alcohol and deionized water alternately until the filtrate reached neutral.After being dried at 80 • for six hours and ground, the powder of BiOBr x I 1−x nanoplate solid solutions was obtained for later use, which was labeled as BiOBr, BiOBr 0.75 I 0.25 , BiOBr 0.5 I 0.5 , BiOBr 0.25 I 0.75 , and BiOI with x = 1.0, 0.75, 0.5, 0.25, and 0.0, respectively.

Characterization Techniques
XRD was employed for the identification of samples' crystal phase by a PANalytical X'Pert X-ray diffractometer (PANalytical Company Inc., Almelo, The Netherlands) using Cu-K α radiation (λ = 0.15418 nm) over the range of 10-80 • with a voltage of 40 kV and an electric current of 40 mA.The microscopic morphologies of BiOBr x I 1−x nanoplate solid solutions were observed by FESEM and TEM, which were implemented on FEISirion200 scanning electron microscope equipped with EDS (FEI Company Inc., Eindhoven, The Netherlands) and JEOL JEM-2100 transmission electron microscope (JEOL Ltd., Tokyo, Japan), respectively.UV-vis DRS spectra were recorded on an USB4000 UV-vis spectrometer (Ocean Optics Inc., Dunedin, FL, USA) equipped with an integral sphere using a standard template provided by South Africa Optics as the reference.The BET specific surface area of samples was obtained from nitrogen adsorption-desorption data and measured using a Sibata SA-1100 surface area analyzer (Sibata Ltd., Kyoto, Japan) at liquid nitrogen temperature.The atomic configuration of BiOX was realized by Diamond program version 3.1 (Crystal and Molecular Structure Visualization, Crystal Impact, Bonn, Germany).The cell parameters a and c of BiOBr x I 1−x solid solutions were obtained by UnitCell software (a non-linear least squares cell refinement program with regression diagnostics, designed by Tim Holland & Simon Redfern), according to a least-squares calculation from 2θ values in XRD patterns.The interlayer spacing of (001) crystal plane was calculated by the well-known Bragg's Law, λ = 2dsinθ, based on 2θ value corresponding to (001) crystal plane.

Photocatalytic Evaluation
A 300 W Xe lamp with a power of 100 mW/cm 2 was used as the simulated sunlight in the whole photocatalytic process.One hundred milligrams of BiOBr x I 1−x photocatalyst was added to 100 mL RhB solution with the initial concentration (C 0 ) of 20 mg/L.Before irradiation by simulated solar light, the suspension was stirred in the dark for 60 min to reach the adsorption and desorption equilibrium.After that, the Xe lamp was turned on and photocatalytic reaction started under continuous stirring.At regular time intervals (30 min) of the photocatalytic reaction, 5 mL of the suspension solution was taken out and centrifuged for 10 min to obtain the supernatant.The residual RhB concentration (C) in the supernatant was detected by a 722-type UV-vis spectrophotometer at RhB maximum absorption wavelength of 553 nm.The detection of chemical oxygen demand (COD) in solution was implemented by fast-digestion spectrophotometric method (Chinese Standard HJ/T 399-2007).The residual dye solution after different reaction periods was also analyzed by UV-vis absorption spectra (USB4000 UV-vis spectrometer, Ocean Optics).

Conclusions
A group of BiOBr x I 1−x solid solutions was successfully obtained at room temperature and atmospheric pressure.These samples exhibited 2D plate-like structure with high {001} facets exposure.
As I content increased, E g values of BiOBr x I 1−x samples decreased, which strongly indicated that the substitution of smaller Br ions by larger I ions could modify their optical absorption properties.BiOBr x I 1−x solid solutions with x = 0.75 exhibited the highest photocatalytic activity.The intensity ratio of the (001) and (102) peaks suggested that {001} facets exposure amount for BiOBr x I 1−x samples was in the following order: BiOI > BiOBr 0.5 I 0.5 > BiOBr 0.25 I 0.75 > BiOBr 0.75 I 0.25 > BiOBr.The regular shift of (001) peak to smaller diffraction angle indicated an increase in the interlayer spacing of (001) crystal plane with I content increasing.This might negatively affect their photocatalytic performances.As I content increased, the potentials of VBM for BiOBr x I 1−x solid solutions gradually decreased.The synergistic effect of these three structural factors determined the photocatalytic performance of BiOBr x I 1−x solid solutions.

Figure 5 .
Figure 5. (a) Photocatalytic degradation of RhB over BiOBrxI1−x solid solutions under simulated sunlight irradiation; (b) pseudo-first-order kinetic curves of RhB photodegradation processes.

Figure 5 .
Figure 5. (a) Photocatalytic degradation of RhB over BiOBrxI1−x solid solutions under simulated sunlight irradiation; (b) pseudo-first-order kinetic curves of RhB photodegradation processes.

Figure 5 .
Figure 5. (a) Photocatalytic degradation of RhB over BiOBr x I 1−x solid solutions under simulated sunlight irradiation; (b) pseudo-first-order kinetic curves of RhB photodegradation processes.

Figure 6 .
Figure 6.(a) Chemical oxygen demand (COD) removal percentage and (b) UV-vis absorption spectra of RhB solution after different photodegradation periods over BiOBr 0.75 I 0.25 .

Figure 9 .
Figure 9. (a) I 001 /I 102 and (b) d 001 of BiOBr x I 1−x solid solutions with different x values.
that CBM potentials of all the samples are positive, implying that the photogenerated electrons in conduction band cannot react with O 2 to form O 2 − radicals (O 2 /O 2 − = −0.28V); nevertheless, VBM potentials of these samples are all more positive than +1.99 V, indicating that the photogenerated holes in the valence band are anodic enough to oxidize H 2 O or OH − into OH radicals (OH/OH − = +1.99V) [54].Meanwhile, as x value increases, the potential of VBM gradually increases.The deeper VBM position can result in higher oxidation power of the photogenerated holes [55], which can directly oxidize the organic molecules or react with adsorbed H 2 O and surface OH − to form highly reactive OH radicals.Hence, BiOI exhibits the worst photocatalytic oxidation power among BiOBr x I 1−x solid solutions, attributed to its lowest VBM potential.

Figure 10 .
Figure 10.Schematic illustration of the band structures of BiOBr x I 1−x solid solutions.

Figure 10 .
Figure 10.Schematic illustration of the band structures of BiOBrxI1−x solid solutions.

Figure 11 .
Figure 11.Schematic illustration of the separation and transport of photogenerated electron-hole pairs driven by IEF.

Figure 11 .
Figure 11.Schematic illustration of the separation and transport of photogenerated electron-hole pairs driven by IEF.

Table 1 .
Cell parameters and band structure calculations of BiOBrxI1−x solid solutions, and reaction rate constants of photodegradation of Rhodamine B (RhB) over BiOBrxI1−x samples.Sample a (

Table 1 .
Cell parameters and band structure calculations of BiOBr x I 1−x solid solutions,