1. IntroductionSemiconductor photocatalysis has attracted tremendous interest as a green and efficient technique to carry out the degradation of organic pollutants [1] [2]. During the typical photocatalytic process, the low efficiency of carriers (i.e., electron and hole pairs) separation is one of the key factors limiting the activity of photocatalytic materials [3]. On account of the limitation of the carriers’ separation, even photocatalyst with a broad spectrum of light absorption cannot exhibit a high photocatalytic activity [4]. The enhancement of the built-in electric field in semiconductors has been found to have a great effect on the improvement of carriers separation [5] [6] and the transport efficiency. Accordingly, various strategies, such as hetero/homojunction [7] [8], elemental doping [9] [10], crystal facet engineering [11] [12], and so on, have been employed to enhance the built-in electric field.
In recent years, various lamellar bismuth based compounds, e.g. BiOX (X = Br, I, Cl) [7], BiOIO3 [13], Bi2O2(OH)(NO3) [14], Bi-M-O (M = V, Mo, W) [15] [16], and so on, have been synthesized and gradually developed to be a series of materials that widely applied in the field of photocatalysis [17]. Compared to the conventional semiconductor photocatalysts, these series of layered bismuth-based materials not only support the forming of an internal electric field to allow the charges diffusion between layers, but also provide abundant distance to polarize orbitals and atoms, enabling the electron-hole pairs separate efficiently [11]. Jiang et al. demonstrated that on single-crystal BiOCl nanosheets, the intensity of the internal electric field was dependent on the exposing ratio of the {001} facet, in which BiOCl nanosheets with more exposed {001} facets exhibited higher activity for photocatalytic pollutant degradation under UV light [12]. Sun and co-workers constructed a heterostructure of BiOI(001)/BiOCl(010). They demonstrated that beyond achievement of lattice and band match, the shorter photogenerated electron diffusion distance in the self-induced internal electric fields of BiOCl slabs leads to a higher charge injection of BiOI(001)/BiOCl(010) [7]. These observations indicate that the crystal facet combination playsa key factorfor enhancing the photocatalytic. Zhanget al. reported that incorporating carbon into the Bi3O4Cl lattice could increase internal electric field (IEF), and strong IEF could separate electrons and holes effectively, and confine them within [Bi3O4] and [Cl] slices to restrict their recombination during their migration from the bulk to the surface [9]. Ren and co-workers synthesized the fluorine isomorphously substituted BiOBrxI1−xsolid solution with various OVs concentrations, and assumed that the excellent photocatalytic performance was due to the synergistic effect of IEF and OVs by fluorine substitution, which could change the charge distribution and promote the separation of photogenerated carriers [10]. Many efforts have been done in tuning IEF or polarizing the related atoms and orbitals, but it still remains unclear to address the key issue of IEF tuning for the photocatalysts with layered structure.
In this work, a series of BiOBr1−xIx/BiOBr hierarchical microsphere photocatalyst were designed and controllably synthesized by a conventional solvothermal process. The visible-light photocatalytic activity was evaluated by the degradation of phenol, which was selected as a colorless probe organic contaminant. An enhanced efficiency of photocatalytic phenol degradation under visible light has been achieved over the optimal BiOBr1−xIx/BiOBr (x = 0.25) sample. It elucidated that the local distortion increases due to elemental doping and heterojunction construction resulted in the spontaneous polarization enhancement and therefore the variation in the built-in electrical field. It is believed that the design of such layered materials could bring a new common strategy to create advance photocatalytic materials for environmental remediation.
2. Experimental Section2.1. MaterialsAll chemical reagents used in this work were analytical reagent grade and without further purification. Bismuth nitrate pentahydrate (Bi(NO3)3·5H2O), potassium bromide (KBr), potassium iodide (KI), ethylene glycol (EG, C2H6O2), phenol, bisphenol A (BPA), 2,4-Dichlorophenol (2,4-DCP), sodium sulphate (Na2SO4), acetylacetone (C5H8O2) and Triton X-100 were purchased from Sinopharm Chemical Reagent Co. Ltd. Deionized water was used throughout the experiments.
2.2. Catalysts PreparationThe photocatalysts were synthesized through a facile solvothermal method. Firstly,the amount of KBr and KI were calculated with different molar Br:I ratios of 2:0, 1.5:0.5, 1:1, 0.5:1.5 and 0:2. In a typical procedure, 2.425 g Bi(NO3)3·5H2O and a certain amount of KBr were dissolved in 20 mL and 10 mL of EG solution under ultrasonication at room temperature, respectively. Then, the KBr solution was added to the Bi(NO3)3 solution under continuous stirring for 30 min. 10 mL of an ethylene glycol (EG) solution containing a certain amount of KI was added dropwise and stirred for another 30 min. Then the resulting mixture was transferred to a 100 mL Teflon-coated autoclave and kept at 100˚C for 8 h. After being cooled to room temperature, the precipitates formed were collected through centrifugation and washed with water and ethanol for three times, respectively. The products were then dried at 60˚C for 6 h. The obtained powder was denoted as BiOBr, BiOBr1−xIx/BiOBr (x = 0.25), BiOBr1−xIx/BiOBr (x = 0.5), BiOBr1−xIx/BiOBr (x = 0.75) and BiOI, respectively.
2.3. CharacterizationTo identify the material composition and the crystal phase of the synthesized samples, XRD patterns were collected by using a Bruker D8 X-ray diffractometer with Cu Kα radiation (λ = 0.15418 nm). The scanning electron microscope (SEM) characterizations were performed on JSM-6700 F field emission scanning electron microscope. The energy dispersive spectrometer (EDS) images were used to analysis elemental composition of the samples. Transmission electron microscopy (TEM) and High-resolution transmission electron microscopy (HRTEM) were performed on a JEOL JEM-2100 microscope at an acceleration voltage of 200 kV. XPS studies were carried out at a ESCALAB250 electronic spectrometer (Thermo, US) using an Al Kα excitation source (hν = 1486.6 eV). The position of the C 1s peak was taken as a standard (with a banding energy of 284.8 eV). Radicals detection experiments were examined by JES-FA200 electron spin-paramagnetic resonance (ESR) instrument (JEOL). The Brunauere-Emmette-Teller (BET) surface areas of the samples were measured by a N2 sorption analyzer (BEISHIDE INSTRUMENT CORP, 3H-2000PM2) at a liquid nitrogen temperature. The UV-Vis spectra were recorded on a Hitachi U-41000 spectrophotometer.
2.4. Photocatalytic ActivityThe visible-light-driven photocatalytic performances of the samples were estimated by measuring the degradation of phenol (10 mg/L). A 300W Xe lamp (Perfect Light, PLS-SXE300) equipped with a 420 nm cut-off filter provided visible-light irradiation. 50 mg of catalysts was dispersed into 50 ml aqueous solution containing the pollutants. The suspension was stirred in the darkness for 60 min to accomplish absorption-desorption equilibrium. After illumination, 4 ml solution was collected at some time intervals and centrifuged to remove the powder. The absorption spectrum and absorption peak of centrifuged solution were analyzed using a UV-vis spectrometer (Shimadzu 3600PLUS).
To investigate the universal applicability of BiOBr1−xIx/BiOBr (x = 0.25) sample, several industrial contaminants (BPA (10 mg/L) and 2,4-DCP (10 mg/L)) were also degraded in the similar condition of phenol, and their concentrations were also analyzed by UV-vis spectrometer. The mineralization of phenol, BPA and 2,4-DCP in aqueous solution was analyzed using total organic carbon (TOC) analyzer (Shimadzu TOC-L-CPN).
2.5. Photoelectrochemical MeasurementThe transient photo-current measurements and electrochemical impedance spectroscopy (EIS) were recorded in a three-electrode quartz cell by using an Electrochemical Workstation (B.V Vertex. C. EIS, Ivium Technologies, Netherlands). The 0.5 M Na2SO4 aqueous solution was used as the electrolyte. The fabricated photoanodes, a Pt foil, and Ag/AgCl electrode were used as working, counter and reference electrode, respectively. The working electrodes were prepared by doctor blade method. A mixture of 20 mg photocatalyst powder, 40 μLdeionized water and a certain ratio of acetylacetone and triton X-100 was coated on a 1 cm × 2 cm FTO glass, followed by annealing at 200˚C for 120 min in atmosphere. A short-arc xenon lamp (Perkin-Elmer, PE300BF) with a cutoff filter (λ ≥ 420 nm) was used as the visible-light source.
3. Results and Discussion3.1. Morphology and Structure of the Hierarchical MicrosphereX-ray powder diffraction (XRD) was carried out to investigate the crystal phase of the as-prepared samples [18]. Figure 1 shows the crystalline structures of the synthesized samples. It can be seen from pure BiOBr that characteristic diffraction peaks detected at 10.90˚, 25.21˚, 31.72˚, 32.27˚, 46.28˚ and 57.20˚ were clearly observed. There characteristic diffraction peaks can be attributed to the (001), (101), (102), (110), (200) and (212) crystal planes of tetragonal phase of BiOBr (PDF#09-0393), respectively [19]. While for pure BiOI, strong diffraction peaks, which can be identified at 29.74˚, 31.73˚, 45.49˚ and 55.30˚, and ascribed to the (012), (110), (020) and (122) planes of the tetragonal BiOI (PDF#73-2062), respectively [20]. Characteristic peaks of both BiOBr and BiOI were clearly identified in the BiOBr1−xIx/BiOBr composites. Remarkably, as shown in the enlarged XRD patterns (Figure 1(b)), when the iodine was introduced in the microspheres structures, the diffraction peaks significantly shift toward lower angles, resulting from the larger radius of I− with respect to Br− [21]. These results indicate that part of bromide atoms had been replaced by the iodine atoms in the lattice and the series of BiOBr1−xIx/BiOBr composites had been successfully synthesized.
Figure 2 shows the morphologies and element analysis of pure BiOBr, pure BiOI and BiOBr1−xIx/BiOBr (x = 0.25) composite. It can be clearly observed from SEM image (Figures 2(a)-(c)) that all samples showd a flower-like morphology, which were composed of nanosheets with a thickness of about 10 - 20 nm. The diameters of microspheres are about 3 μm, and these nanosheets contained many gaps which were in favor for the interfacial charge transfer [22]. It is interesting that the nanosheets of iodine doped composite BiOBr1−xIx/BiOBr (x = 0.25) are rather thinner and more unconsolidated than that of pure BiOBr. To further illustrate the distribution of BiOBr1−xIx/BiOBr (x = 0.25), energy dispersive spectroscopy (EDS) elemental mapping was performed and the results were
shown in Figures 2(g)-(k). The elements of Bi, O, Br and I all can be obviously seen and their distributions are highly homogeneous on the surface of BiOBr1−xIx/BiOBr (x = 0.25) microspheres.
From the corresponding HRTEM image (Figure 3), good crystalline structure and clear lattice fringes can be observed. The interplanar d-spacing of about 0.279 nm and 0.299 nm in Figure 3(a) and Figure 3(c) could be ascribed to the (110) lattice plane of BiOBr and BiOI nanosheets, respectively [23]. As shown in Figure 3(b), the fringes with the lattice spacing of 0.285 nm matched well with the (110) lattice plane of BiOBr1−xIx, and the lattice spacing of 0.279 nm in the outer layer was assigned to the (110) facet of BiOBr [24]. The SAED patterns in the inset picture of Figure 3(b) illustrated that a polycrystalline structure has been fabricated. The above results suggested the heterojunction structures with intimate interface contact between BiOBr1−xIx and BiOBr had been formed. This heterojunction structure is beneficial for the electron transfer process during photocatalytic processes [7].
To measure the chemical composition of as-prepared photocatalysts, X-ray photoelectron spectroscopy (XPS) measurement was performed to determine the exact surface state [25]. As shown in Figure 4(a), the survey scan spectra listed the peaks of Bi 4f, O 1s, Br 3d and/or I 3d elements which were calibrated by the C 1s at a binding energy of 284.8 eV [23]. This indicates the high purity of these samples except the slight adventitious hydrocarbons [20]. As shown in
Figure 4(b), the binding energy to Bi 4f7/2 peak at 159.2 eV and Bi 4f5/2 peak at 164.5 eV were detected, which demonstrated that Bi existed in the form of Bi3+ [26] [27]. Two typical peaks were observed in Br 3d spectra at 68.2 and 69.3 eV (Figure 4(c)), which can be assigned to Br 3d5/2 and Br 3d3/2 in the composites, respectively [28]. That can be confirmed that the presence of Br−. Two strong peaks were observed at 619.2 and 630.7 eV (Figure 4(d)), which could be assigned to I 3d5/2 and I 3d3/2, indicating the existence of I− in the BiOBr1−xIx/BiOBr (x = 0.25) samples [26]. Moreover, a minor shift of the Br 3d3/2 andBr 3d5/2 binding energies was observed for the BiOBr1−xIx/BiOBr (x = 0.25) composite, which increased to 69.6 eV and 68.5 eV, respectively. Similarly, the binding energies of I 3d3/2 and I 3d5/2 for the BiOBr1−xIx/BiOBr (x = 0.25) composite were also slightly shifted to 630.9 eV and 619.3 eV, respectively. The slight shift of Br and I peaks in BiOBr1−xIx/BiOBr (x = 0.25) is probably due to the interaction between different halogen atoms [29].
Raman spectroscopy is an appropriate technique for probing the local structure of materials because the bonding states in the coordination polyhedra of a material can be deduced directly from the Raman vibrational spectrum. Raman spectra of the synthesized BiOBr, BiOI and BiOBr1−xIx/BiOBr with various Bi/I ratio were recorded to identify the local structure information (Figure 5). Pure BiOBr shows two distinctive bands (110 and 158 cm−1). The A1g internal Bi-Br stretching is probably overlapped by the strong band at 110 cm−1, and the band at 158 cm−1 can be ascribed to Eg internal Bi-Br stretching mode [30] [31]. The band at 85 cm−1 (assigned to the A1g internal Bi-I stretching mode) and the band at 148 cm−1 (assigned to the Eg internal Bi-I stretching mode) of BiOI can be observed [32] [33].
As for the BiOBr1−xIx/BiOBr (x = 0.25) composite, the Raman peak assigned to A1g internal Bi-Br stretching shifts to higher wavenumbers, which can be attributed to the fact thatiodine doping might produce intrinsic stresses on the crystal structure in which therefore alter the periodicity of the lattice [30] [34] and result in the distortion of the local structure. The distortion in the local structure leads to overlap between the Bi 6s and the p orbital of halogen. The greater the degree of distortion of the local structure, the more intrinsic stresses on the crystal structurewill be produced, which enhances the migration of photogenerated holes. This result can also be deduced from the results obtained by XRD and XPS.
Figure 6 is the N2 adsorption-desorption isotherms of the photocatalysts, and all samples showed a type IV curve [35]. And the BiOI had the highest surface area of 52.5 m2/g among all the samples. It may result from the thinner and more unconsolidated structure of BiOI. Generally, there was a positive correlation between the BET surface area and the photocatalytic activity. Simply judging from the surface areas of the as-synthesized samples, the BiOI should have better photocatalytic activity. However, the BiOBr1−xIx/BiOBr (x = 0.25) composite has a higher photocatalytic activity. Therefore, it is reasonable to extrapolate
that the SBET value of the samples is a possible factor that may influence the capability of a photocatalyst, but not the only reason in this system.
3.2. Photocatalytic ActivityPhenolphotodegradation was used as a model reaction to evaluate the photoactivities of the as-prepared samples. As depicted in Figure 7(a), without irradiation, it can be observed that the phenol and catalyst suspensions reach absorption-desorption equilibrium at 60 min, and all samples show a small adsorption ability of phenol. Under visible light irradiation, the blank experiment reveals that phenol is barely decomposed without photocatalysts. For the pure BiOBr and BiOI, the concentration of phenol was only reduced by about 40.8% and 63.9% after 150 min of visible-light irradiation. In contrast, the photodegradation activity of as-prepared BiOBr1−xIx/BiOBr (x = 0.25) composite is significantly enhanced, and 92.5% of phenol can be degraded after 150 min irradiation. As shown in insert of Figure 7(b), the pseudo first-order linear relationship is revealed by the plots of ln(C/C0) vs irradiation time (t). The apparent reaction rate constant (k) of BiOBr1−xIx/BiOBr (x = 0.25) is calculated as 0.01943 min−1. Moreover, during the photocatalysis characterization of the as-synthesized samples, the concentration of total organic carbon (TOC) was evaluated. The TOC value of phenol was reduced by about 56.4% after 150 min of visible-light irradiation for BiOBr1−xIx/BiOBr (x = 0.25) photocatalysts, which showed the evidence of the decomposition of phenol into carbon dioxide [36].
Furthermore, the recyclability of the BiOBr1−xIx/BiOBr (x = 0.25) composite in the degradation of phenol is considered. After the complete degradation of phenol, the catalyst was collected by centrifugation, and reused in the next cycle. As shown in Figure 8(a), after four cycles of repeated use, a phenol degradation efficiency of 87.6% is still obtained. And no distinct changes in the morphology
and structure were found for the BiOBr1−xIx/BiOBr (x = 0.25) photocatalyst after 4 runs of cycling experience, as shown in XRD patterns (Figure 8(b)), which indicates the structure of the photocatalyst is extremely stable and has not been significantly changed during the reaction process. Figure 8(c) exhibits the SEM of BiOBr1−xIx/BiOBr (x = 0.25) composite before and after 4runs cycling reaction. It can be seen that the composite still keeps a good hierarchical microsphereshape.
Moreover, other persistence organic contaminants, such as BPA and 2,4-DCP, can also be degraded efficiently, verifying the common applicability of BiOBr1−xIx/BiOBr (x = 0.25) as effective visible-light photocatalyst using in the decomposition of organic compound (Figure 9). The total organic carbon (TOC) results show that more than 66% and 44% TOC in the BPA and 2,4-DCP solution are removed within 180 min of irradiation, respectively. All these results confirm the high mineralization activity of BiOBr1−xIx/BiOBr (x = 0.25) under visible-light irradiation.
3.3. Dependence of Photocatalytic Activity on the StructureUV-Vis DRS was employed to investigate the photophysical properties of the as-synthesized samples (Figure 10(a)). It can be found that all samples show a strong absorption in the visible light region. Compared with BiOBr, with the introduction of iodine into the crystalline BiOBr, the absorption edge of BiOBr1−xIx/BiOBr samples exhibited red-shift, but band edge shift is still smaller than that of BiOI. This observation suggests that the formation of BiOBr1−xIx/BiOBr composites can indeed modulate the capacity of visible light absorption to some extent [28]. The band gap energies (Eg) of all samples were obtained by the following formula [37]:
α h υ = A ( h υ − E g ) n 2 (1)
where α, h, ν and A represent the absorption coefficient, Planck constant and light frequency and proportionality, respectively. And the fundamental absorption of both BiOBr and BiOI are indirect transition between bands, so n is equal to 4.
As shown in Figure 10(b), the band gaps of BiOBr, BiOBr1−xIx/BiOBr (x = 0.25) and BiOI are 2.66, 2.37 and 1.93 eV, respectively. After the introduction of
iodine ions into the BiOBrlattice, localized levels in the forbidden energy gap were formed, leading to a decrease in the band gap energy [38]. Further, the valence band X-ray photoelectron spectra (VB-XPS) suggested that the positions of valence band of BiOBr, BiOBr1−xIx/BiOBr (x = 0.25) and BiOI were 1.98, 1.69 and 1.37 eV, respectively (Figure 10(c)). Based on the equation of ECB = Eg − EVB [37], the conduction band (CB) edges of BiOBr, BiOBr1−xIx/BiOBr (x = 0.25) and BiOI were calculated to be −0.68, −0.68 and −0.56 eV, respectively. According to the above analysis, the band position schematic of BiOBr, BiOBr1−xIx/BiOBr (x = 0.25) and BiOI was obtained, with the results shown in Figure 10(d). The BiOBr1−xIx/BiOBr (x = 0.25) has a narrower Eg than that of BiOBr, demonstrating that BiOBr1−xIx/BiOBr (x = 0.25) is conducive to improve the visible-light absorptive capacity [39]. And the valence band edge of BiOBr1−xIx/BiOBr (x = 0.25) is higher than that of BiOI simultaneously, unveiling that BiOBr1−xIx/BiOBr (x = 0.25) possesses elevated oxidizing ability [40].
To explore the primary active species generated in the reaction system, trapping experiments of active species were conducted as shown in Figure 11(a). Sodium oxalate (SO), N2 and isopropanol (IPA) were used as the quenchers for h+, • O 2 − and •OH, respectively. The addition of SO or bubbling N2 caused the degradation efficiency of phenol decreased to 32.3% and 37.5%. After IPA was added, the photocatalytic efficiency was almost invariable. ESR measurements were employed to evaluate the main oxidative species during the photodegradation
process [41]. As shown in Figure 11(b) and Figure 11(c), ESR signals of DMPO- • O 2 − adduct and DMPO-•OH adduct are not detected indark. However, strong DMPO- • O 2 − signals were observed after the same irradiated by visible-light, confirming that the photogenerated electrons (e−) on the CB of BiOBr1−xIx/BiOBr (x = 0.25) could be trapped to produce • O 2 − [42]. And the characteristic peak of DMPO-•OH cannot be obviously detected under the identical conditions, which is consistent with the scavenger experiments. These indicated that h+ and • O 2 − were significant active species in the photocatalytic process of BiOBr1−xIx/BiOBr (x = 0.25), while •OH was not the major reactive species.
To investigate the transportation behaviors of the photoexcited charges in these samples, the photoelectrochemical analyses were carried out. Figure 12(a) records the transient photocurrent responses of the catalysts under visible light irradiation. Among all of the samples, BiOBr1−xIx/BiOBr (x = 0.25) shows the most intensity signal of photocurrent density. This result revealed that less recombination and longer lifetime of photoinduced charges than those of others, which unveils that charge separation efficiency can be enhanced by the internal electric fields and band match [43]. Further, Figure 12(b) shows the the electrochemical impedance spectra (EIS) of the samples. Normally, the smaller radius in an EIS Nyquist is, the better charge transfer ability would be, which would impede the recombination of photogenerated charge carriers [44]. The BiOBr1−xIx/BiOBr (x = 0.25) composite has the smallest radius with respect to
other samples, indicating much lower electron transfer resistance than that of the other samples, and therefore, defines higher charge transfer efficiency [43] [45]. As shown in Figure 12(c), all samples are regarded as n-type semiconductors because of the positive slope of the Mott-Schottky curve [46]. If the slope in a Mott-Schottky plot is lower, the carrier concentration is higher [47] [48]. It is indicated that the BiOBr1−xIx/BiOBr (x = 0.25) composite has the highest carrier concentration. Accordingly, the enhancement of separation efficiency of photoinduced electron-hole pairs (e−-h+) ought to be another origin of excellent photoactivity of BiOBr1−xIx/BiOBr (x = 0.25).
The enhanced photocatalytic activity of BiOBr1−xIx/BiOBr (x = 0.25) originated from the fast separation of photo generated charges under visible light. On the basis of the above results and previous reports [15] [46], a possible mechanism for the enhanced photocatalytic activity over the BiOBr1−xIx/BiOBr (x = 0.25) is shown schematically in Figure 13. On the one hand, since the iodine element has the optimum size and valence, with the introduction of iodine into the crystalline BiOBr, the local tension around the iodine in the crystal lattice changes, resulting in a change in the delocalization energy. And so on, the IEF increases [49]. The IEF drives e− and h+ move from the separation sites to [Bi2O2] and [Br] slices [9]. Therefore, the charge recombination was drastically inhibited, which was of great benefit for enhancing the photocatalytic activity [36] (Figure 13(a)). On the other hand, as shown in Figure 13(b), under visible light, bothBiOBr1−xIx and BiOBr can be simultaneously excited, the photoexcited electrons transferred
from CB of the BiOBr1−xIx to that of BiOBr, and since the VB of BiOBr (1.98 eV) was more positive than that of BiOBr1−xIx (1.69 eV), the photoexcited holes on BiOBr could transfer to the VB of BiOBr1−xIx driven by CB offset of 0.29 eV. The potential difference was the main driving force for efficient charge separation and transfer [50]. As the photogenerated electrons and holes were spatially separated into two different components, the charge recombination was drastically inhibited, which was of great benefit for enhancing the photocatalytic activity [51]. Because the redox potential of O2/ • O 2 − was −0.33 V vs. NHE at pH about 7 [52], the electrons assembled on the conduction band of BiOBr had enough driving force (−0.68 V) to react with O2 to form • O 2 − reactive species for degrading pollutants. At the same time, the photoinduced holes gathered at the valence band of BiOBr1−xIx directly oxidize pollutants to the final degradation products. Therefore, BiOBr1−xIx/BiOBr (x = 0.25) exhibits the improved photocatalytic activity by the synergistic effect of IEF and heterojunction, which could be utilized as the manipulating strategy for high effective catalyst.