Enhanced Photo-Assisted Fenton Degradation of Antibiotics over Iron-Doped Bi-Rich Bismuth Oxybromide Photocatalyst

Herein, combining photocatalysis and Fenton oxidation, a photo-assisted Fenton system was conducted using Fe-doped Bi4O5Br2 as a highly efficient photocatalyst to realize the complete degradation of Tetracycline antibiotics under visible light. It has been observed that the optimized photocatalyst 5%Fe-doped Bi4O5Br2 exhibits a degradation efficiency of 100% for Tetracycline with H2O2 after 3 h visible-light irradiation, while a degradation percentage of 59.8% over the same photocatalyst and 46.6% over pure Bi4O5Br2 were obtained without the addition of H2O2 (non-Fenton process). It is unambiguous that a boost photo-assisted Fenton system for the degradation of Tetracycline has been established. Based on structural analysis, it demonstrated that the Fe atoms in place of the Bi sites may result in the distortion of the local structure, which induced the occurrence of the spontaneous polarization and thus enhanced the built-in electric field. The charge separation efficiency is enhanced, and the recombination of electrons and holes is inhabited so that more charges are generated to reach the surface of the photocatalyst and therefore improve the photocatalytic degradation efficiency. Moreover, more Fe (II) sites formed on the 5%Fe-Bi4O5Br2 photocatalyst and facilitated the activation of H2O2 to form oxidative species, which greatly enhanced the degradation efficiency of Tetracycline.


Introduction
Antibiotics are quite effective in resisting microorganisms and treating biological diseases [1,2]. Tetracycline (TC), the quintessential antibiotic, is commonly employed, possessing the broadest antibacterial spectrum as well as a tough chemical structure [3]. However, the massive misuse of TC must be accompanied by incomplete metabolism and accidental spills due to its intrinsic toxicity and constancy. Tiny TC can seriously poison the aquatic environment and conversely threaten life on Earth [4,5]. Therefore, removing tetracycline and other antibiotics has become one of the important research topics in the environmental field.
The photocatalytic technique as an advanced oxidation processes (AOPs) stands out among various technologies to purify antibiotic sewage, which is apparently a "green" process driven by sustainable solar energy [6]. It is presently significant to harvest visible light (at least 50% of the solar spectrum) for efficient photocatalytic elimination of TC. As a novel photocatalyst, bismuth oxybromide (BiOBr) exhibits potential applications on energy generation, environmental remediation, bacterial disinfection and so on, particularly in wastewater treatment [7,8]. Pristine BiOBr possesses a bandgap of~2.8 eV determined by a sandwich-like molecular structure with [Bi 2 O 2 ] 2+ and a double layer of Br; however, the weak visible-light harvest and slow carrier separation efficiency hindered its wide applications [9]. To alleviate the above symptom, a "Bi-rich" strategy is employed to optimize the unique molecular structure, and a series of Bi-rich bismuth oxybromide (Bi 4 [10][11][12][13][14]. The suitable bandgap (~2.5 eV) and and iron nitrate nonahydrate were dissolved in ethylene glycol (25 mL) at various mass ratios (Bi:Fe = 99:1, 97:3, 95:5 and 93:7) and labeled as solution A, and 2 mmol potassium bromide was dissolved into 10 mL ethylene glycol, labeled as solution B. Then solution B was slowly added into solution A by stirring vigorously for 30 min until they were homogeneously mixed. The pH of the mixed solution was regulated to 10.5 using a 2.0 mol/L sodium hydroxide solution. Finally, the mixture was charged into a 100 mL Teflon-lined autoclave and kept at 160 • C for 12 h. The product was washed with deionized water and anhydrous ethanol 3 times, respectively, and then dried at 70 • C. The resulting products were labeled as Bi 4 O 5 Br 2 , 1%Fe-Bi 4 O 5 Br 2 , 3%Fe-Bi 4 O 5 Br 2 , 5%Fe-Bi 4 O 5 Br 2 and 7%Fe-Bi 4 O 5 Br 2 , respectively.
All the chemical reagents used in the experiments, including bismuth nitrate pentahydrate, iron nitrate nonahydrate, potassium bromide, ethylene glycol, sodium hydroxide and ethanol absolute are analytically pure and purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) without any further purification.

Preparation of the Photocatalysts
The one-step solvothermal method was used for the preparation of all the photocatalysts in this study. The steps are illustrated as Scheme 1: first, bismuth nitrate pentahydrate and iron nitrate nonahydrate were dissolved in ethylene glycol (25 mL) at various mass ratios (Bi:Fe = 99:1, 97:3, 95:5 and 93:7) and labeled as solution A, and 2 mmol potassium bromide was dissolved into 10 mL ethylene glycol, labeled as solution B. Then solution B was slowly added into solution A by stirring vigorously for 30 min until they were homogeneously mixed. The pH of the mixed solution was regulated to 10.5 using a 2.0 mol/L sodium hydroxide solution. Finally, the mixture was charged into a 100 mL Teflonlined autoclave and kept at 160 °C for 12 h. The product was washed with deionized water and anhydrous ethanol 3 times, respectively, and then dried at 70 °C. The resulting products were labeled as Bi4O5Br2, 1%Fe-Bi4O5Br2, 3%Fe-Bi4O5Br2, 5%Fe-Bi4O5Br2 and 7%Fe-Bi4O5Br2, respectively. Scheme 1. Schematic illustration of the Fe-Bi4O5Br2 preparation processes.

Characterizations
Raman vibrational spectra were achieved on a DXR2 spectrometer (Thermo, Waltham MA, USA), which was excited by a 532 nm laser. The crystal structures of the samples were performed on an Ultima IV (XRD, Rigaku, Tokyo, Japan) with Cu Kα irradiation (λ = 0.15418 nm), and the range of diffraction angles was controlled at 5-80°. The morphology and element mapping images were taken on a Regulus 8100 (SEM, JOEL,Tokyo, Japan). The local images were taken on a 2100F (TEM, HRTEM, JOEL, Tokyo, Japan). The XPS spectra were conducted on Escalab Xi + (Thermo) with Al-Kα (hν = 1486.6 eV), which were calibrated based on C1s of adventitious carbon at 284.8 eV. The diffuse reflectance UV-Vis spectra were obtained on a UV-2700 (Shimadzu, Tokyo, Japan) using BaSO4 as background. The nitrogen adsorption and desorption isotherm and pore size distribution were measured and analyzed by Autosorb-iQ-MP-C (Quantachrome, Norcross, GA, USA). The PL spectra were conducted on a FLS 980 (Ediburgh, livingston, Scotland, UK) using a Xenon lamp with an excitation wavelength of 375 nm at room temperature (293 K). The in-situ ESR signals were collected on an EMX plus (Bruker, Berlin, Germany) under visible-light irradiation (λ ≥ 420 nm). The photoelectrochemical measurements were carried out on PARSTAT 4000A (Princeton, NJ, USA). The ICP-MS test was measured on an Agilent 730ES (Palo Alto, CA, USA). Scheme 1. Schematic illustration of the Fe-Bi 4 O 5 Br 2 preparation processes.

Characterizations
Raman vibrational spectra were achieved on a DXR2 spectrometer (Thermo, Waltham MA, USA), which was excited by a 532 nm laser. The crystal structures of the samples were performed on an Ultima IV (XRD, Rigaku, Tokyo, Japan) with Cu Kα irradiation (λ = 0.15418 nm), and the range of diffraction angles was controlled at 5-80 • . The morphology and element mapping images were taken on a Regulus 8100 (SEM, JOEL, Tokyo, Japan). The local images were taken on a 2100F (TEM, HRTEM, JOEL, Tokyo, Japan). The XPS spectra were conducted on Escalab Xi + (Thermo) with Al-Kα (hν = 1486.6 eV), which were calibrated based on C1s of adventitious carbon at 284.8 eV. The diffuse reflectance UV-Vis spectra were obtained on a UV-2700 (Shimadzu, Tokyo, Japan) using BaSO 4 as background. The nitrogen adsorption and desorption isotherm and pore size distribution were measured and analyzed by Autosorb-iQ-MP-C (Quantachrome, Norcross, GA, USA). The PL spectra were conducted on a FLS 980 (Ediburgh, livingston, Scotland, UK) using a Xenon lamp with an excitation wavelength of 375 nm at room temperature (293 K). The in-situ ESR signals were collected on an EMX plus (Bruker, Berlin, Germany) under visible-light irradiation (λ ≥ 420 nm). The photoelectrochemical measurements were carried out on PARSTAT 4000A (Princeton, NJ, USA). The ICP-MS test was measured on an Agilent 730ES (Palo Alto, CA, USA).

Degradation Experiments of Tetracycline Hydrochloride
The photocatalytic activity of the catalyst was evaluated by the degradation of Tetracycline hydrochloride (TC) under visible-light irradiation. Specifically, 30 mg of photocatalyst was dispersed into a 20 mg/L 50 mL TC solution and the suspension was stirred at a certain speed. Firstly, a 4 mL reaction mixture was taken every 30 min under dark conditions to ensure that the degradation reached the adsorption-desorption equilibrium. The light source used in this experiment was a 300 W Xe lamp with a 420 nm cutoff filter. Following this step, 4 mL samples were taken every 30 min after turning on the light. The obtained reaction mixture was centrifuged, and then the concentration of the tetracycline was analyzed by measuring the absorbance at 356 nm on a UV-Vis spectrophotometer. The photo-Fenton process was operated similarly to the above photocatalytic degradation of TC, except that a concentration of H 2 O 2 was 0.28 mol/L in the reaction system. The experiments under darkness are the same as the above steps without light.

Results and Discussion
3.1. The Structure of the Synthesized Samples X-ray diffraction (XRD) patterns were carried out to determine the crystal structure of the samples. Figure 1a clearly shows five characteristic peaks located at 24 the range of 80-140 nm. The formation of these mesopores is beneficial for the adsorption of pollutants in wastewater, and then effectively decomposed the pollutants [41]. As displayed in Table 1, it was found that the pore volume of the sample increased due to iron doping. Comparatively, the pore volume of 5%Fe-Bi4O5Br2 is 0.431 m 2 /g compared to that of pure Bi4O5Br2 (0.328 m 2 /g).  The FESEM, TEM and HRTEM are conducted to show the morphologies and the microstructure of the samples. It is obvious that both Bi4O5Br2 (Figures 2a and 3a) and 5%Fe-Bi4O5Br2 (Figures 2b and 3b) show nanosheet-interlaced morphologies; the thickness of 5%Fe-Bi4O5Br2 is close to 8.3 nm, which is much thinner than that of Bi4O5Br2 (about 12.5 nm). This observation indicates that the doping of Fe atoms reduced the thickness of the nanosheet, which thus reduced the transfer distance of photogenerated carriers to the surface and therefore facilitates more carriers to participate in the reaction. The lattice fringes of the samples can be clearly observed in the HRTEM of Figure 3c,d, and the lattice spacing of both Bi4O5Br2 and 5%Fe-Bi4O5Br2 are 0.28 nm, which is consistent with the index of the (11-3) lattice plane. This result indicates that Fe-doped Bi4O5Br2 with high crystallinity is successfully synthesized. The elemental mappings (Figure 2c) show that the O, Fe, Bi and Br elements are uniformly dispersed in 5%Fe-Bi4O5Br2, which indicates that the Fe element is highly distributed in the sample. The molecular structure variation and the lattice distortion induced by the Fe atom doping can also be confirmed by the Raman spectra. As illustrated in Figure 1b, the characteristic peaks at 96.6 cm −1 and 156.3 cm −1 could be assigned to A 1g and E 1g of the internal Bi−Br stretching mode of Bi 4 O 5 Br 2 , respectively [38]. The peak intensity represents the content of the vibratory group, and the half-peak width is related to the number of layers in the same space [39]. It is clear that variations in the half peak width and shift in the vibrational frequency of the peak can be observed with the introduction of Fe atoms into the framework of Bi 4 O 5 Br 2 . Moreover, the intensity of peaks assigned to A 1g and E 1g modes gradually reduced with the increasing amount of Fe dopant, and even these peaks almost disappeared in the 7%Fe-Bi 4 O 5 Br 2 sample. This result indicates that the number of Bi-Br groups decreased due to the substitution of the Fe atom doping [40]. In contrast, the widening in the half-peak of Fe-doped Bi 4 O 5 Br 2 indicates that the Bi-Br layers in the same space reduced in number and thus the lattice distortion is formed, which resulted from the doping of Fe atoms with a smaller radius. The specific surface area (S BET ) and pore distribution of the samples were investigated by low temperature N 2 physical adsorption. As shown in Figure 1c, it indicated that all samples showed a type IV N 2 adsorption-desorption isotherm with an H 3 hysteresis loop. This result shows that there might be a large number of active vacancies on the surface of the sample, which further confirmed that the doping of surface Fe atoms resulted in the increase in vacancies. The S BET of pure Bi 4 O 5 Br 2 is 24.6 m 2 /g, while all the Fe-doped Bi 4 O 5 Br 2 increased in value. Among them, the S BET of 7%Fe-Bi 4 O 5 Br 2 is the largest one attaining to 75.0 m 2 /g. Similarly, the 5%Fe-Bi 4 O 5 Br 2 increases 2.63-fold in the S BET than that of pure Bi 4 O 5 Br 2 , which is 64.7 m 2 /g. The increases in the S BET due to the doping of Fe atoms could offer a large number of reactive sites to participate in photodegradation reactions.
The pore size distribution of samples is mainly in the range of 2-50 nm (mesoporous) as shown in Figure 1d. Notably, the sample of 5%Fe-Bi 4 O 5 Br 2 shows many macropores in the range of 80-140 nm. The formation of these mesopores is beneficial for the adsorption of pollutants in wastewater, and then effectively decomposed the pollutants [41]. As displayed in Table 1, it was found that the pore volume of the sample increased due to iron doping. Comparatively, the pore volume of 5%Fe-Bi 4 O 5 Br 2 is 0.431 m 2 /g compared to that of pure Bi 4 O 5 Br 2 (0.328 m 2 /g).  (Figures 2b and 3b) show nanosheet-interlaced morphologies; the thickness of 5%Fe-Bi 4 O 5 Br 2 is close to 8.3 nm, which is much thinner than that of Bi 4 O 5 Br 2 (about 12.5 nm). This observation indicates that the doping of Fe atoms reduced the thickness of the nanosheet, which thus reduced the transfer distance of photogenerated carriers to the surface and therefore facilitates more carriers to participate in the reaction. The lattice fringes of the samples can be clearly observed in the HRTEM of Figure 3c  The elemental composition and states on the surface of samples are determined by X-ray photoelectron spectroscopy (XPS). As shown in Figure 4a, the obvious signals of Bi, Br and O are collected in the survey spectrum of Bi 4 O 5 Br 2 , and on the basis, an extra signal of Fe is captured on 5%Fe-Bi 4 O 5 Br 2 . In detail, the Bi 4f 5/2 and Bi 4f 7/2 peaks at 164.5 eV and 159.2 eV, representing the presence of Bi 3+ , are detected in the Bi 4f refined spectra for Bi 4 O 5 Br 2 (Figure 4b). In comparison, these peaks shift~0.15 eV to low binding energy for 5%Fe-Bi 4 O 5 Br 2 , confirming that the density of the electron cloud around the Bi atom increased. This result can be explained by the fact that the electronegativity of the Fe element (1.83) is lower than that of the Bi element (1.9) [42,43]. Therefore, the electron from the doping of Fe is easily injected into the Bi sites. As shown in Figure 4c, the Br 3d signal in Bi 4 O 5 Br 2 can be deconvoluted into Br 3d 5/2 and Br 3d 3/2 at 68.6 eV and 69.6 eV, belonging to the Br − , respectively. Similarly, a low binding energy shift of about 0.15 eV occurred in the Br 3d spectrum of 5%Fe-Bi 4 O 5 Br 2 . Therefore, the electron density of the Br atom plate is stronger, which is affected by the doping of Fe possessing more valence electrons compared to the Bi atom [44]. Furthermore, in the high-resolution O 1s spectra (Figure 4d), two peaks at 529.8 eV and 531.2 eV are observed in Bi 4 O 5 Br 2 and 5%Fe-Bi 4 O 5 Br 2 , which are ascribed to the intrinsic lattice oxygen and absorbed oxygen [9]. In addition, Bi 4 O 5 Br 2 possessed an additional peak at 533.1 eV assigned to the absorbed water [45]. Deeply, the signal of Fe 2p can be fitted into four peaks for 5%Fe-Bi 4 O 5 Br 2 corresponding to Fe 2p 1/2 and Fe 2p 3/2 , each of which contains two different valence states of Fe ( Figure 4e). Among them, the two peaks at 725.8 eV and 712.5 eV are assigned to the Fe 3+ . Meanwhile, the other two peaks at 722.6 eV and 710.0 eV are assigned to the Fe 2+ [45,46]. In terms of contributing to the Fe 2p, the content of Fe 2+ accounted for 0.587%, which is larger than Fe 3+ (0.413%), indicating the surface content of Fe 2+ is higher. Therefore, the abundant redox Fe 2+ /Fe 3+ on the surface are facilitated to excite H 2 O 2 to boost the degradation performance of the TC.  The elemental composition and states on the surface of samples are determined by X-ray photoelectron spectroscopy (XPS). As shown in Figure 4a, the obvious signals of Bi, Br and O are collected in the survey spectrum of Bi4O5Br2, and on the basis, an extra signal of Fe is captured on 5%Fe-Bi4O5Br2. In detail, the Bi 4f5/2 and Bi 4f7/2 peaks at 164.5 eV and 159.2 eV, representing the presence of Bi 3+ , are detected in the Bi 4f refined spectra for Bi4O5Br2 (Figure 4b). In comparison, these peaks shift ~0.15 eV to low binding energy for 5%Fe-Bi4O5Br2, confirming that the density of the electron cloud around the Bi atom increased. This result can be explained by the fact that the electronegativity of the Fe element (1.83) is lower than that of the Bi element (1.9) [42,43]. Therefore, the electron from the doping of Fe is easily injected into the Bi sites. As shown in Figure 4c, the Br 3d signal in Bi4O5Br2 can be deconvoluted into Br 3d5/2 and Br 3d3/2 at 68.6 eV and 69.6 eV, belonging to the Br − , respectively. Similarly, a low binding energy shift of about 0.15 eV occurred in The photophysical properties of the samples are measured by UV-vis diffuse reflectance spectroscopy. As shown in Figure 5a, the absorbance edge of pure Bi 4 O 5 Br 2 is around 500 nm, indicating that this sample is a visible-light responsive narrow bandgap photocatalyst. It is obvious that the red-shift in the absorbance edge with the introduction of the Fe atom can be observed, demonstrating that the absorbance spectrum in the range of visible light is broadened with the increasing amount of dopant. In addition, the band gap of Fe-Bi 4 O 5 Br 2 becomes narrow compared to that of pure Bi 4 O 5 Br 2 , reflected in the  and Fe 2p3/2, each of which contains two different valence states of Fe (Figure 4f). Among them, the two peaks at 725.8 eV and 712.5 eV are assigned to the Fe 3+ . Meanwhile, the other two peaks at 722.6 eV and 710.0 eV are assigned to the Fe 2+ [45,46]. In terms of contributing to the Fe 2p, the content of Fe 2+ accounted for 0.587%, which is larger than Fe 3+ (0.413%), indicating the surface content of Fe 2+ is higher. Therefore, the abundant redox Fe 2+ /Fe 3+ on the surface are facilitated to excite H2O2 to boost the degradation performance of the TC. The photophysical properties of the samples are measured by UV-vis diffuse reflectance spectroscopy. As shown in Figure 5a, the absorbance edge of pure Bi4O5Br2 is around 500 nm, indicating that this sample is a visible-light responsive narrow bandgap photocatalyst. It is obvious that the red-shift in the absorbance edge with the introduction of the Fe atom can be observed, demonstrating that the absorbance spectrum in the range of visible light is broadened with the increasing amount of dopant. In addition, the band gap of Fe-Bi4O5Br2 becomes narrow compared to that of pure Bi4O5Br2, reflected in the Tauc plot (Figure 5b). The bandgaps of Bi4O5Br2, 1%Fe-Bi4O5Br2, 3%Fe-Bi4O5Br2, 5%Fe-Bi4O5Br2 and 7%Fe-Bi4O5Br2 were calculated to be about 2.   The enhanced photocurrent confirmed that the effective separation of the photogenerated electron−hole pairs is driven by a boosted internal electric field generated by the doping of Fe. Furthermore, the transfer resistance of carriers within the samples could be evaluated from the electrochemical impedance spectroscopy (EIS). Generally, the shorter semi-cycle arc radius in the Nyquist plot means a lower resistance to the charge transfer. Therefore, as shown in Figure 6b, the semicircular radius in the Nyquist plots of the Fe-Bi 4 O 5 Br 2 series samples decreases with the introduction of Fe, which confirmed that the transfer resistances of carriers in the Fe-Bi 4 O 5 Br 2 series samples are less than that in Bi 4 O 5 Br 2 . The semi−cycle arc radius in the Nyquist plot of 5%Fe-Bi 4 O 5 Br 2 is the shortest one, indicative of the quickest carrier transferring rate of this sample being generated. In summary, the doping of Fe atoms into the framework of Bi 4 O 5 Br 2 resulted in the occurrence of spontaneous polarization and therefore enhancement in the internal electric field, which effectively induced the separation and rapid migration of the photogenerated electron-hole pairs [50].  Figure 6a shows the transient photocurrent density response of Bi4O5Br2 and Fedoped Bi4O5Br2 samples. It can be observed that the series of Fe-doped samples produce intensive current responses compared to Bi4O5Br2 under visible-light irradiation. The photocurrent intensity of the 5%Fe-Bi4O5Br2 photocatalyst exhibited much stronger than that of pure Bi4O5Br2. The enhanced photocurrent confirmed that the effective separation of the photogenerated electron−hole pairs is driven by a boosted internal electric field generated by the doping of Fe. Furthermore, the transfer resistance of carriers within the samples could be evaluated from the electrochemical impedance spectroscopy (EIS). Generally, the shorter semi-cycle arc radius in the Nyquist plot means a lower resistance to the charge transfer. Therefore, as shown in Figure 6b, the semicircular radius in the Nyquist plots of the Fe-Bi4O5Br2 series samples decreases with the introduction of Fe, which confirmed that the transfer resistances of carriers in the Fe-Bi4O5Br2 series samples are less than that in Bi4O5Br2. The semi−cycle arc radius in the Nyquist plot of 5%Fe-Bi4O5Br2 is the shortest one, indicative of the quickest carrier transferring rate of this sample being generated. In summary, the doping of Fe atoms into the framework of Bi4O5Br2 resulted in the occurrence of spontaneous polarization and therefore enhancement in the internal electric field, which effectively induced the separation and rapid migration of the photogenerated electron-hole pairs [50]. The recombination kinetics of photogenerated carriers can be comprehensively simulated by the measurement of fluorescence spectra. In the steady-state photoluminescence (PL) spectra (Figure 6c), 5%Fe-Bi4O5Br2 responds with a weaker fluorescence intensity compared to Bi4O5Br2 upon excitation at a 375 nm exciting light source. The result indi-  The recombination kinetics of photogenerated carriers can be comprehensively simulated by the measurement of fluorescence spectra. In the steady-state photoluminescence (PL) spectra (Figure 6c), 5%Fe-Bi 4 O 5 Br 2 responds with a weaker fluorescence intensity compared to Bi 4 O 5 Br 2 upon excitation at a 375 nm exciting light source. The result indicated that the recombination of photogenerated carriers was inhibited in 5%Fe-Bi 4 O 5 Br 2 , which is consistent with the result observed in the photocurrent. In addition, the average fluorescence lifetime (τ) of photogenerated carriers is evaluated in the time−resolved photoluminescence spectra (TRPL). As shown in Figure 6d, the τ value of 5%Fe-Bi 4 O 5 Br 2 is 3.112 ns, which is effectively extended than that of Bi 4 O 5 Br 2 (2.909 ns), suggesting that facilitated carriers with longer residence time are generated in the sample. These results further demonstrated that the photogenerated carriers can be effectively separated and transferred driven by the built−in electron field enhanced by 5%Fe-Bi 4 O 5 Br 2 .

Degradation Performance of TC
The photocatalytic activities of the Bi 4 O 5 Br 2 and Fe-Bi 4 O 5 Br 2 samples were investigated by the degradation of TC under visible-light irradiation. As shown in Figure 7a, it is clear that almost no loss of TC in the absence of the photocatalyst in the control experiment can be observed, which confirmed that the degradation of TC is entirely derived from the action of the photocatalysts. Moreover, a dark treatment for 1 h is enough for the TC to reach an adsorption-desorption equilibrium on the surface of all the samples. The 5%Fe-Bi 4 O 5 Br 2 sample shows the strongest adsorption for TC, almost three times as much as Bi 4 O 5 Br 2 , which is attributed to a larger specific surface area exposing more adsorption sites. Significantly, 5%Fe-Bi 4 O 5 Br 2 possesses the best photocatalytic degradation activity for TC under visible-light irradiation for 3 h, reaching 60%. However, a further increase in the amount of the dopant into Bi 4 O 5 Br 2 induces an adverse effect on the photocatalytic performance. It can be observed that the photocatalytic activity of 7%Fe-Bi 4 O 5 Br 2 for the removal of TC decreases to only 28.9%, even worse than that of pure Bi 4 O 5 Br 2 (46.6%). Therefore, the appropriate ratio of Fe doping can induce the spontaneous polarization and enhance the built-in electric field, thus achieving effective separation of the photogenerated carrier. Moreover, the photocatalytic degradation of TC is evaluated in the first-order reaction by Equation (3). Further, as shown in Figure 7b, the reaction rate constant (k) of 5%Fe-Bi 4 O 5 Br 2 is fitted to 0.006 min −1 , which increases nearly 1.5-fold than that of Bi 4 O 5 Br 2 (0.004 min −1 ).
Furthermore, the photo−Fenton system is employed to improve the degradation efficiency of TC, and the dependence of the activity on the structure of the photocatalyst in the presence of H 2 O 2 is studied. As shown in Figure 7c, without the photocatalyst while only the H 2 O 2 system is added, 29.4% of TC degradation activity is achieved under visible-light irradiation. It is interesting that when the Fe-Bi 4 O 5 Br 2 series of samples are supplied, the degradation performance of TC in the presence of H 2 O 2 under visible-light irradiation is promoted significantly and is much higher than that of Bi 4 O 5 Br 2 with the H 2 O 2 system (57.9%). The highest photocatalytic activity over the optimized sample, 5%Fe-Bi 4 O 5 Br 2 , can reach the entire removal of TC in 3 h under visible-light irradiation. Similarly, the degradation reaction rate constant of TC in the photo−Fenton system is also estimated according to the first-order reaction. As shown in Figure 7d, the highest k is 0.016 min −1 for the 5%Fe-Bi 4 O 5 Br 2 -Fenton system, which is almost 3.2 times greater compared to the Bi 4 O 5 Br 2 -H 2 O 2 system (0.005 min −1 ). This result indicates that the optimized recombination kinetics of the photogenerated carrier over 5%Fe-Bi 4 O 5 Br 2 facilitates to activate H 2 O 2 sufficiently to generate strong oxidation • OH and expedites the degradation reaction kinetic.
As shown in Figure 7e, the Fenton process for the degradation of TC is evaluated under darkness. Specifically, almost 17% of TC is removed only with H 2 O 2 , and the Bi 4 O 5 Br 2 -H 2 O 2 shows the degradation ratio of TC is about 27%, which is comparable to that of only with H 2 O 2 regardless of the adsorption capacity of the photocatalyst. Apparently, the degradation performance is enhanced in the 5%Fe-Bi 4   In order to explore the active species over Fe-Bi4O5Br2 in the photocatalytic process, an in-situ ESR technique is employed. The radical capture reagent TEMPO (2,2,6,6-tetramethyl-1-piperidinyloxy) is utilized in the photocatalytic system to remove the h + . As shown in Figure 8a,b, the triple peak in the ESR signal shows the characteristic species of TEMPO with h + . The decreases in the peak intensity represent that the generated h + is captured by TEMPO. It is obvious that the decrease in peak intensity of the 5%Fe-Bi4O5Br2 is more prominent than that of the Bi4O5Br2 after 10 min visible-light irradiation, indicating that the 5%Fe-Bi4O5Br2 may be able to generate holes more efficiently during the photocatalytic degradation. The generated • OH is captured by the 5,5-dimethyl-1-pyrroline (DMPO) in the 0.28 M H2O2 aqueous solution. As shown in Figure 8c,d, four characteristic From the above results, the 5%Fe-Bi 4 O 5 Br 2 photocatalyst showed excellent activity with light and H 2 O 2 . Finally, in order to study the photocatalytic stability of the catalyst, a cyclic experiment was carried out, as shown in Figure 7f. After five cycles for the degradation of TC, the 5%Fe-Bi 4 O 5 Br 2 still retains its excellent performance (85%), with only a 15% loss of degradation ratio compared to the fresh reaction. In addition, it is found that without obvious variations of the crystal structure of the samples after five cycles can be observed. Moreover, the concentration of Fe ions in the solution was negligible (about 0.022 mg/L) in the used 5%Fe-Bi 4 O 5 Br 2 after the cycle experiments in Figure 7h.
In order to explore the active species over Fe-Bi 4 O 5 Br 2 in the photocatalytic process, an in-situ ESR technique is employed. The radical capture reagent TEMPO (2,2,6,6-tetramethyl-1-piperidinyloxy) is utilized in the photocatalytic system to remove the h + . As shown in Figure 8a,b, the triple peak in the ESR signal shows the characteristic species of TEMPO with h + . The decreases in the peak intensity represent that the generated h + is captured by TEMPO. It is obvious that the decrease in peak intensity of the 5%Fe-Bi 4 O 5 Br 2 is more prominent than that of the Bi 4 O 5 Br 2 after 10 min visible-light irradiation, indicating that the 5%Fe-Bi 4 O 5 Br 2 may be able to generate holes more efficiently during the photocatalytic degradation. The generated • OH is captured by the 5,5-dimethyl-1-pyrroline (DMPO) in the 0.28 M H 2 O 2 aqueous solution. As shown in Figure 8c, peaks with an intensity of 1:2:2:1 are collected, demonstrating that the Bi4O5Br2 and the 5%Fe-Bi4O5Br2 can activate H2O2 efficiently under visible-light irradiation [51]. Additionally, the more intensive signal of DMPO confirmed that more photogenerated carriers in the 5%Fe-Bi4O5Br2 could be used to activate H2O2 to produce a greater number of • OH. As shown in Figure 8e, the characteristic signal of DMPO-• O2 − was not observed on the 5%Fe-Bi4O5Br2 regardless of the irradiation or not. This result indicated that the • O2 − species could not be generated through the single−electron reduction in oxygen under visible-light irradiation; for that the CB potential of Fe-Bi4O5Br2 is −0.21 eV vs. NHE (E [O2/ • O2 − ] = −0.33 eV vs. NHE). Interestingly, the ESR signal of • O2 − can be observed with 10 min irradiation in the presence of H2O2 in Figure 8f. It is therefore demonstrated that the • O2 − species could be generated by the activation of H2O2 through excited 5%Fe-Bi4O5Br2. So, both the • OH and • O2 − could be generated by visible-light irradiation of the Fe-Bi4O5Br2 photocatalyst to realize the efficient degradation of tetracycline. In summary, combined with the above discussion of free radicals, the mechanism and diagram can be presumed and shown in Equations (4)-(10) and Figure 9. Under the irradiation of visible light, the carriers are separated and shifted efficiently to the photocatalyst surface (Equation (4)). Due to the enhancement in the spontaneous polarization In summary, combined with the above discussion of free radicals, the mechanism and diagram can be presumed and shown in Equations (4)-(10) and Figure 9. Under the irradiation of visible light, the carriers are separated and shifted efficiently to the photocatalyst surface (Equation (4)). Due to the enhancement in the spontaneous polarization effect, a great number of photogenerated charges reach the surface of 5%Fe-Bi 4 O 5 Br 2 . More holes are transferred to the surface of the 5%Fe-Bi 4 O 5 Br 2 photocatalyst from the in−situ ESR spectrum. Photogenerated holes may oxidize TC (Equation (5)), while the Fe 3+ species is reduced to Fe 2+ by the photogenerated electrons (Equation (6)). Moreover, more Fe(II) sites were synthetized on the 5%Fe-Bi 4 O 5 Br 2 from the XPS spectrum . Due to the addition of hydrogen peroxide, the Fenton effect in the reaction, that is, Fe 2+ reacts with H 2 O 2 to form Fe 3+ and produce • OH (Equation (7)), and generated • OH can degrade TC (Equation (9)). Finally, Fe 3+ combined with hydrogen peroxide will produce • O 2 − (Equation (8)) [52], which can further degrade TC pollutants (Equation (10)).

Conclusions
In conclusion, the photo-Fenton degradation system was constructed via Fe-doped Bi4O5Br2 nanosheet. The obtained 5%Fe-Bi4O5Br2-photo-Fenton system realized excellent degradation performance (almost 100%) for the removal of Tetracycline under visiblelight irradiation. The apparent reaction rate constant of 5%Fe-Bi4O5Br2 reaches 0.016 min −1 , which is almost 3.2 times faster than that of Bi4O5Br2 (0.005 min −1 ). The complete removal of Tetracycline and the enhanced reaction rate are primarily attributed to the multiple effects of Fe doping in the 5%Fe-Bi4O5Br2-photo-Fenton system. On one hand, the doping of Fe induces the spontaneous polarization, which enhances the built-in electric field to

Conclusions
In conclusion, the photo-Fenton degradation system was constructed via Fe-doped Bi 4 O 5 Br 2 nanosheet. The obtained 5%Fe-Bi 4 O 5 Br 2 -photo-Fenton system realized excellent degradation performance (almost 100%) for the removal of Tetracycline under visible-light irradiation. The apparent reaction rate constant of 5%Fe-Bi 4 O 5 Br 2 reaches 0.016 min −1 , which is almost 3.2 times faster than that of Bi 4 O 5 Br 2 (0.005 min −1 ). The complete removal of Tetracycline and the enhanced reaction rate are primarily attributed to the multiple effects of Fe doping in the 5%Fe-Bi 4 O 5 Br 2 -photo-Fenton system. On one hand, the doping of Fe induces the spontaneous polarization, which enhances the built-in electric field to modulate the photogenerated carrier separation kinetics of 5%Fe-Bi 4 O 5 Br 2 . On the other hand, the Fe(II) in the molecular framework becomes the activation center of H 2 O 2 and the abundant photogenerated carriers accelerate the transition between Fe(II) and Fe(III), thus efficiently activating H 2 O 2 to generate • O 2 − and more • OH. The holes also play a decisive role to oxidize Tetracycline under visible-light irradiation. These active species are confirmed by in-situ ESR. Therefore, the 5%Fe-Bi 4 O 5 Br 2 -photo-Fenton system is a potential candidate for "green" photocatalytic removal of antibiotics.