Fei Chang a,* , Wenjing Yan a , Xiaomeng Wang a , Shijie Peng a , Sushi Li a , Xuefeng Hu b,*  a School of Environment and Architecture, University of Shanghai for Science and Technology, Shanghai 200093, PR China b School of Environmental Science and Engineering, Shaanxi University of Science and Technology, Xi’an 710021, PR China

ABSTRACT

Design and creation of photocatalytic systems with favorable microstructural and morphological merits are quite  crucial to heighten removal of environmental hormones. For such purpose, in the current study n-type Bi4O5Br2  and p-type MnO2 were specially selected to construct three-dimensional (3D) hierarchical composites Bi4O5Br2-  MnO2 (MB) via a facile one-pot procedure. Through thorough and systematical analyses, the existence of both  expected components was confirmed in composites and these heterojunction MB composites showed significantly  improved photocatalytic degradation of bisphenol A (BPA) under visible light. Particularly, the best candidate  0.5 MB owned the largest Kapp value 11.03 × 10− 3 min− 1 that was about 3.89 and 1.81 times those of bare  Bi4O5Br2 and mechanical mixture 0.5 MB, which primarily attributed to strengthened visible-light absorption,  effective integration of favorable morphologies, proper chemical composition, and efficient spatial separation of  induced charge carriers through n-p heterojunction domains and thus boosted generation of oxidative radicals in  a Z-Scheme model. In addition, other parameters possibly influenced photocatalytic degradation performance  such as catalyst dosages, initial concentrations of BPA, pH values, and involvement of inorganic anions were also  explored. Moreover, some intermediates were detected by GC-MS analysis to provide a plausible degradation  path of these robust 3D hierarchical heterojunctions with satisfactory reusability and structural stability.

1. Introduction

It is well cognized that the rapid development of chemical industries  and sharp increase of chemicals utilization have caused discovery of  various hormones in the environment [1]. These hormones are easily  accumulated in some living organisms and subsequently threaten  human health to some degree even though their contents are quite low  in watery environment. Among these hormones, BPA is one of the most  extensively utilized industrial raw-materials worldwide, mainly  employing for production of functional polymers, plasticizers, flame  retardants, antioxidants, heat stabilizers, and etc [2]. BPA molecules in  water tend to enter human bodies, causing toxic effects over reproduction and increased cancerogenic risk. Comparing to traditional treatments, semiconductor-based photocatalysis with well-known merits has  drawn both academic and industrial attentions and thus been extensively researched from both environment and energy aspects [3–8].

In recent years, BixOyXz (X = Cl, Br, I) series have caused great  concerns for their unique layered structures and favorable electronic  structures, exhibiting outstanding photocatalytic properties in comparison to corresponding BiOX counterparts [9–11]. In addition, according  to the diffusion formula t 2 = d2/k2 D (d, k, and D are the particle size,  constant, diffusion coefficient of electron–hole pairs, respectively), the  relatively small particle size and rapid transfer of charge carriers lead to  the small value t that favors the enhancement of photocatalytic performance. Thus, the ultrathin-layered structure and inherent internal static  electric fields ensure satisfactory photocatalytic nature of BixOyXz series  [12]. Among non-stoichiometric bismuth oxyhalides, n-type semiconductor Bi4O5Br2 provides a narrow band gap such as 2.33 [12], 2.3  [13], and 2.36 eV [14] with relatively negative conduction band (CB)  position comparing to BiOBr [15] and is easy to be tuned into 3D hierarchical morphologies [16,17]. With the integration of these physiochemical features, Bi4O5Br2 obviously surpasses BiOBr upon  photocatalytic degradation of BPA [15]. Nevertheless, as a singlephased semiconductor, bare Bi4O5Br2 still suffers the limited utilization of visible light and high recombination of charge carrier,  which can be conquered by multifarious microstructural modifications.  For instance, through morphological modulation hollow microspheres  Bi4O5Br2 were constructed to promote photocatalytic capability [18].  The decoration of Bi4O5Br2 surface by iodide caused excellent photocatalytic capacity of BPA degradation [19]. In addition to that, the  fabrication of composites via introducing semiconductors with suitable  band structures such as h-BN [20] and C3N4 [21] was also attempted to  enhance photocatalytic BPA removal. Contrasting to other modifications, the creation of heterojunction composites is a promising strategy  since it benefits the suppression of electron-hole pairs recombination,  light absorption, and surface properties as well, thus significantly  boosting photocatalytic capabilities [22,23]. Particularly, the integration of n-type and p-type semiconductors with appropriate band structures is advantageous because of dramatically enhanced separation of  charge carriers by self-built inner electric field at interface with simultaneous expanded absorption of light region [24]. Considering bare  Bi4O5Br2 as an n-type semiconductor with good BPA removal ability, it is  rational to choose a p-type component with favorable band structure to  construct n-p heterojunction composites in terms of ameliorating photocatalytic removal of BPA. 

As a typical transition metal oxide, MnO2 in p-type is often used to  decorate other semiconductors to prepare composites for its natural  abundance, inexpensiveness, low-toxicity, sufficient stability, and  environmentally friendliness [25,26]. In addition, MnO2 with partially  filled d-levels is able to absorb visible light via d-d transition and owns  appropriate band gaps about 1.56 [27], 1.64 [28], 1.68 V [29]. Therefore, it can be selected as a photosensitizer or component to combine  other semiconductors to achieve satisfactory catalytic behaviors [26].  Expectedly, MnO2 in p-type with aforesaid merits is desirable to couple  with Bi4O5Br2 to form n-p heterojunctions, readily steering and redistributing charge carriers driven by the equalization of Fermi levels from  both [30]. Moreover, if a Z-scheme heterojunction system is achieved,  not only more charge carries but also relatively negative CB and positive  valence band (VB) positions can be concurrently gained by recombination of electrons and holes from adjacent CB and VB in different phases,  provding sufficient charge carriers with strong redox ability and further  enhanced photocatalytic capacities [31–33]. For all we know, the facile  fabrication and visible light-driven photocatalytic removal of BPA of 3D  hierarchical n-p heterojunctions Bi4O5Br2-MnO2 have never been reported. The underlying properties, photocatalytic degradation route of  BPA molecules, and relevant mechanism speculation in a Z-Scheme  model ought to be systematically discussed in depth.

In such investigation, 3D hierarchical MB heterojunctions were  constructed through a facile one-pot procedure and subsequently characterized by various analyses. Under visible light, these composites were  subjected to photocatalytic evaluation over BPA removal. The significant enhancement of photocatalytic behavior was systematically  researched. Other parameters possibly affected degradation efficiency  such as catalyst dosage, initial concentration of BPA, pH values, and  involvement of inorganic anions were explored. In addition, some intermediates were detected by GC-MS analysis to speculate a possible  degradation path and a probable photocatalysis mechanism was eventually proposed. This research may shed a light on efficiently photocatalytic removal of environmental hormones such as BPA under mild  conditions through bismuth-based n-p heterojunction composites with  beneficial microstructural features.

2. Experimental section

2.1. Reagents and chemicalsMMolecular structures, purity grade, and sources of relevant chemicals  and reagents were collected in Table S1. These relevant reagents were  straightforwardly used without any further purifications.

2.2. Samples construction

Bare Bi4O5Br2 was fabricated via a facile chemical precipitation  process that was slightly modified on basis of a previous report [15].  Typically, Bi(NO3)3·5H2O (4.851 g, 10 mmol) was entirely dissolved in a  HNO3 aqueous solution (20 mM, 10 mL) at room temperature under  magnetic stirring for 10 min. The resultant solution was stirred in a  water bath at 70 ◦C for 2 min and then added with a CTAB solution (9  mM, 20 mL). After stirring for 2 min, the solution was introduced  dropwise by a 30 mL mixed solution containing 90 mmol NaOH and 60  mmol NaBr and the obtained mixture was continuously stirred for 10  min. Eventually, the precipitate was collected by centrifugation, washed  with deionized water and absolute ethanol for three times to remove  residual ions, and dried at 60 ◦C in air for 6 h to supply bare Bi4O5Br2  with a yield of 2.684 g.

Binary MB composites were prepared according to a similar procedure described above. In a general synthesis, MnO2 (5 mmol) was  introduced into a HNO3 solution (20 mM, 10 mL) containing Bi  (NO3)3·5H2O (10 mmol) and the resultant mixture was treated in  ultrasonication for 20 min for good dispersion. Following steps are  identical to those for bare Bi4O5Br2. Such composite was obtained as a  solid, nominated as 0.5 MB. Similarly, by changing initial dosage of  MnO2 as 1, 3, and 7 mmol, samples 0.1 MB, 0.3 MB, and 0.7 MB were  prepared accordingly. Basing on the yield of bare Bi4O5Br2, composites  0.1 MB, 0.3 MB, 0.5 MB, and 0.7 MB had theoretical mass ratios of MnO2  to Bi4O5Br2 as 3.2, 9.7, 16.1, and 22.6 wt%, respectively. Exact mass  ratios of MnO2 to Bi4O5Br2 were determined by inductively coupled  plasma mass spectrometry (Agilent 7900 ICPMS) as around 2.2, 7.5,  10.2, and 15.1 wt% for composites 0.1 MB, 0.3 MB, 0.5 MB, and 0.7 MB.  Slight distinctions between theoretical and measured values were  mainly attributed to measuring errors and possible reaction of bromide  ions and MnO2 in acidic condition to form divalent manganese [34].  Mixture 0.5 MB was synthesized by grinding Bi4O5Br2 and MnO2 in  proportion to composite 0.5 MB with an agate mortar for 30 min.

2.3. Samples characterization

Powdery X-ray diffraction (XRD) patterns of as-synthesized samples  were measured on a Bruker D8 Advance diffractometer. X-ray photoelectron spectroscopy (XPS) was recorded on a Thermo Scientific  ESCALAB 250Xi system and all binding energies were demarcated by  accidental carbon at 248.6 eV. Morphology and microstructures were  observed on a field emission scanning electron microscope (SEM, LEO-  1530) and a transmission electron microscope (TEM, FEI Tecnai G2  F20). X-ray energy dispersion spectroscope (EDS) incorporated into SEM  was used to detect distribution map of main elements. UV–Vis diffuse  reflectance spectra (UV–Vis DRS) were obtained on a Shimadzu UV-  2600 spectrophotometer with bare BaSO4 as a reference. Specific surface areas and pore size distribution were analyzed by the Brunauer Emmett-Teller (BET) method and the Barrett-Joyner-Halenda (BJH)  model on an ASAP2460 machine. Electrochemical analyses were carried  out on an electrochemical work station (CHI660E, Chenhua Instrument  Company, Shanghai). The saturated calomel electrode, a Pt wire, the  copper sheet coated with samples, and Na2SO4 solution (0.5 M) were  employed as a reference electrode, counter electrode, working electrode, and electrolyte, respectively. Zeta potentials of composite 0.5 MB  in aqueous were measured on a Malvern Zetasizer Nano ZS90 instrument. Time-resolved transient PL spectra of relevant samples were  recorded on a multifunction fluorescence spectrometer (FLS1000,  Edinburgh Instruments).

2.4. Photocatalytic performance evaluation

Photocatalytic capabilities of achieved samples were gauged by  removal of colorless BPA under visible light. A distance of 20 cm was  regulated from reaction surface to a Xenon lamp (300 W, CEL-HXF300 with an emission wavelength from 300 to 2500 nm, Aulight) that was  fitted with a 420–780 nm cutoff glass filter. The total irradiance of lamp  with filter is nearly 90.5 mW cm− 2 measured by a full-spectrum light  optical power meter (CEL-NP200-2, Aulight). Typically, 40 mg catalyst  was immersed into a BPA solution (20 mg L− 1 , 80 mL) and the gained  suspension was stirred in dark for 120 min to reach an adsorption–desorption balance. Afterwards, the suspension was exposed to  visible light and at certain time intervals a 4 mL aliquot was sampled and  centrifuged to abolish catalyst particles. Residue concentrations were  analyzed at the maximum absorption wavelength of 276 nm on a Shimadzu UV-2600 spectrophotometer. The mineralization degree was  detected on a Total Organic Carbon (TOC) Analyzer (V-CPN, Shimadzu).  For repeatability and reproducibility, each reported value was the  average of at least three parallel tests and corresponding error bars were  provided. In order to explore various factors affecting catalytic degradation of BPA, different catalyst doses (10, 20, 40, and 80 mg) of 0.5 MB,  initial concentrations (10, 20, 30, and 40 mg L− 1 ) of BPA, and pH values  (4.5, 6.5, 8.5, and 10.5) by addition of NaOH and HCl solutions were  used to conduct photocatalytic experiments. In addition, 0.05 M  different sodium salts (NaCl, Na2CO3, Na2SO4, and NaH2PO4) were  introduced into BPA solutions and photocatalytic reactions were carried  out under same condition. Moreover, photocatalytic removal ability of  samples 0.5 MB and Bi4O5Br2 for different pollutants was explored as  follows. Catalyst (40 mg) was added into TC (20 mg L− 1 , 80 mL) or RhB  (20 mg L− 1 , 80 mL) solution and during reactions residual concentrations of both were detected at the maximum absorption wavelength of  357 and 556 nm, respectively.

2.5. Intermediates derivatized and determinationQualitative analysis of intermediates was performed on a gas chromatograph (GC-2010, Shimadzu) equipped with a GCMS-QP2010 plus  mass spectrometer and a DB-5 MS column (30 m × 0.25 mm × 0.25 µm,  Agilent Technologies). The oven temperature was maintained at 60 ◦C  for 2 min and ramped to 300 ◦C at a rate of 6 ◦C min− 1 . Then, oven  temperature was at 300 ◦C held for 5 min. The carrier gas was highpurity helium that was maintained at a constant flow rate of 1 mL  min− 1 . A 1 µL sample was injected in a splitless mode at 280 ◦C of  injector temperature. Mass spectrometry was recorded in the electron  impact ionization mode at 70 eV and full scan mode over the range of  50–600 m/z was employed.

The suspension solutions were centrifuged and 30 mL sample solutions were drawn from obtained supernatants. To improve extraction  efficiency, NaCl was added to solutions to reach a concentration of 150  g L− 1 . Then, sample solutions were extracted with 5 mL dichloromethane for three times. Combined organic extracts were treated over  anhydrous Na2SO4, subjected to rotary evaporation, and further  concentrated to near-dryness under a gentle nitrogen flow. Afterwards,  samples were redissolved in 1.0 mL mixture of dichloromethane-hexane  (1:1) and silylated using 200 µL of BSTFA at 60 ◦C for 8 h after being  evaporated to dryness under nitrogen. The derivatized samples were  redissolved in 0.5 mL of hexane and then subjected to GC-MS analysis.

2.6. Entrapping experiments and ESR analyses  

5.0 mM of reagent IPA, TEA, or AgNO3 was chosen to respectively  entrap hydroxyl radicals (·OH), holes (h+), or electrons (e- ) that might  appear during photocatalytic processes. Moreover, reagent NBT was  employed in presence of composite 0.5 MB to detect superoxide radicals  (•O2− ) that reacted with NBT molecules to produce insoluble blue precipitates, which was analyzed on a Purkinje General T6 spectrophotometer at 260 nm. The existence of radicals ·OH and •O2− was further  corroborated by the entrapping reagent 5, 5-dimethyl-1-pyrroline Noxide (DMPO) in water and methanol on a JEOL JES FA200 electron spin 

resonance spectrometer (ESR).

2.7. Reusability and recyclability evaluation

The reusability and structural steadiness of composite 0.5 MB were  estimated by five successive photocatalytic cycles. After each run,  catalyst was collected by centrifugation, washed with deionized water  and ethanol for three times, and dried at 60 ◦C in air for 6 h prior to next  use. 

3. Results and discussion

3.1. Phase composition, microstructure, and morphology analyses

XRD patterns were measured to estimate phase composition and  crystal structures of bare Bi4O5Br2, MnO2, and composites MB. As  observed in Fig. 1a, Bi4O5Br2 is of high crystallinity and all diffraction  peaks can be well indexed to pure cubic phase (JCPDS card, No. 412591)  [14]. XRD pattern of bare MnO2 in Fig. S1 is in good accordance with  hexagonal ε-MnO2 phase (JCPDS card, No. 30-0820) [35]. Quite weak  and broaden peaks imply low crystallinity with relatively small size of  commercial MnO2. After combining with MnO2, MB series show similar  XRD patterns to bare Bi4O5Br2 with slightly decreased intensities,  revealing the coverage of Bi4O5Br2 surface with low content and high  dispersion of MnO2 species in poor crystallinity [20]. Moreover, significant chemical shifts of diffraction peaks in MB series are undetectable,  which excludes the occurrence of Mn-doping [36].

Morphology and microstructures of Bi4O5Br2, MnO2, and composite  0.5 MB were visualized via SEM and EDS mapping images. In Fig. 1b,  Bi4O5Br2 depicts combinatorial morphology of 3D flower-like structures  and large hierarchical clusters accumulated by numerous nanosheets.  Distinctively, SEM image of MnO2 in Fig. 1c includes a large number of  irregularly-shaped particles. Distinct morphologies of both components  are convenient for phase discrimination. As to composite 0.5 MB, 3D  flower-like structures remarked by yellow eclipses together with hierarchical clusters are also observed in Fig. 1d, indicating good preservation of original morphology without apparent destruction. In  addition, in the enlarged SEM image in Fig. 1e, lamellar structures and  particles simultaneously exist, clearly confirming coexistence of both  components in composite 0.5 MB. For clearness, some domains  remarked by red ellipse frames are attributed to MnO2 particles. EDS  element mapping from Fig. 1f shows corresponding elements Bi, Mn, O,  and Br in Fig. 1g. Apparently, these elements are evenly distributed over  the selected region of composite 0.5 MB.

For sake of further insight into microstructure of sample 0.5 MB,  TEM and HRTEM images were measured. 3D hierarchical flower-like  structures are viewed in Fig. 1h and some small particles are distributed among nanosheets remarked by red arrows, which reveals the  presence of both target components, in good accordance with SEM images. The HRTEM image of composite 0.5 MB borderline in Fig. 1i includes different lattice fringe distances of 0.242 and 0.199 nm,  respectively representing (100) and (422) crystal planes of MnO2 and  Bi4O5Br2, which reconfirms coexistence of both ingredients. As a result,  both ingredients are intimately integrated to form heterojunction  structures in a compatible morphology, by which induced charge carriers are able to be promoted to transport and separate at interfaces,  therefore boosting photocatalytic performance.

XPS spectra were achieved to study surface composition and  elemental states of Bi4O5Br2, MnO2, and composite 0.5 MB. In Fig. 2a,  survey spectrum reveals the containment of elements Bi, Br, O, and Mn  in composite 0.5 MB, suggesting the existence of both expected components. Meanwhile, fortuitous carbon at 284.8 eV is selected as a  standard to correct other binding energies. Bi 4f spectrum in Fig. 2b  consists of two signals with binding energies at 159.6 and 164.9 eV in  Bi4O5Br2, respectively assigning to Bi 4f7/2 and Bi 4f5/2 orbitals in Bi3+ species [19]. After the formation of composite 0.5 MB, both peaks are  shifted downward by 0.4 eV. As to Br spectra in Fig. S2a, binding energies of Br 3d3/2 and Br 3d5/2 in either Bi4O5Br2 or 0.5 MB are located  at around 69.8 and 68.7 eV, corresponding to Br− ions in Bi4O5Br2 lattices [19]. Similarly, O1s spectra of Bi4O5Br2, MnO2, and 0.5 MB are  depicted in Fig. S2b. O1s signals at 530.3 and 532.0 eV in MnO2 attribute to Mn-O-Mn bonds [28] and adsorbed oxygen species on surface [37]. Both O1s signals in Bi4O5Br2 at 529.6 and 531.1 eV correspond to  lattice O [38] and adsorbed oxygen species [39]. Accordingly, three  peaks can be divided from band of composite 0.5 MB at 529.6, 530.3,  and 531.1 eV, respectively indexing to Bi-O, Mn-O, and adsorbed oxygen  species. The peak at 531.1 eV in composite 0.5 MB is slightly strengthened after combining MnO2, revealing presence of more adsorbed oxygen species and probable generation of more oxidative radicals. Mn 2p  spectrum of MnO2 in Fig. 2c contains two peaks at 642.1 and 653.7 eV  with a 2p doublet deviation of 11.6 eV, assigning to Mn 2p3/2 and Mn  2p1/2 orbitals of Mn4+ species [40]. In addition, both mentioned peaks  in composite 0.5 MB are down-shifted to 641.5 and 653.1 eV. As  observed, binding energies of both Bi4f and Mn2p move to lower positions, suggesting the presence of strong chemical interaction between  both phases [41].

N2 adsorption–desorption isotherms are displayed in Fig. 2d to  research textural properties of Bi4O5Br2, MnO2, and composite 0.5 MB.  All these samples are mesoporous due to the obvious type IV curves with  hysteresis loops at relative pressures of 0.43− 0.90 in Fig. 2h [42], from  which specific surface areas (SBET) of Bi4O5Br2, MnO2, and 0.5 MB are  estimated as 8.23, 29.9, and 13.5 m2 /g, respectively. The slight increase  of SBET value of 0.5 MB in comparison to Bi4O5Br2 is possibly relevant to  the creation of new surface by incorporation of MnO2 particles and may  not be a primary factor to ameliorate photocatalytic BPA removal.

Optical feature and band structures of samples were analyzed by  UV–Vis DRS spectra. As seen in Fig. 2e, bare Bi4O5Br2 in yellow possesses absorption from ultraviolet to visible light up to 550 nm, while  MnO2 in black is able to absorb light from 200 to 800 nm and thus can be  used as a photosensitizer or a component to construct composites with  reinforced visible-light absorption and charge carries generation [27].  Obviously, MB series display enhanced visible light absorption and  apparent bathochromic-shift of absorption edges, which are unanimous  to gradient color change in inset of Fig. 2e. Band gap energies (Eg) of  samples can be estimated from the Kubelka-Munk equation αhν = A(hν-  Eg) n/2 [43]. Value n is generally as 1 and 4 respectively for direct and  indirect transition semiconductors. As reported [40,44], value n herein  is adopted as 4 and 1 for Bi4O5Br2 and MnO2. Accordingly, values Eg of  Bi4O5Br2 and MnO2 are estimated as 2.29 and 1.44 eV in Fig. 2f and  values Eg of MB series are also calculated in Fig. S3. Physiochemical  properties and photocatalytic efficiencies of as-synthesized samples are  collected in Table S2.