Contents lists available at ScienceDirect

Materials Science in Semiconductor Processing

Materials Science in Semiconductor Processing 40 (2015) 344–350

http://d1369-80

n CorrE-m

journal homepage: www.elsevier.com/locate/mssp

Synthesis, characterization and photocatalytic propertiesof BiOBr/amidoxime fiber composites

Ting-Xian Tao n, Jun-Sen Dai, Ren-Chun Yang, Jia-Bin Xu, Wei Chu, Zhi-Chuan WuSchool of Biological and Chemical Engineering, Anhui Polytechnic University, Wuhu 241000, China

a r t i c l e i n f o

Article history:Received 27 January 2015Received in revised form29 May 2015Accepted 14 June 2015Available online 15 July 2015

Keywords:BiOBr/fiberRecycling performanceCoordination interactions

x.doi.org/10.1016/j.mssp.2015.06.04001/& 2015 Elsevier Ltd. All rights reserved.

esponding author. Tel.:+86 553 2871255; faail address: tingxiantao@126.com (T.-X. Tao)

a b s t r a c t

BiOBr/amidoxime fiber composite photocatalysts have been successfully prepared via ionschelating on fibers. The structure and composition of the products were characterized byXRD, EDS, SEM, FT-IR, XPS and TG. The results indicated that BiOBr was successfullyimmobilized on the modified commercial fibers. Photocatalytic testing showed that theBiOBr/amidoxime fiber composites exhibited an excellent recycling performance owing tothe coordination interactions between BiOBr and amidoxime fiber.

& 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Over the past decade, semiconductor materials haveattracted considerable attention owing to their intriguingpotential for diverse applications, such as environmentaland energy issues [1,2]. The photocatalytic degradation oftoxic organic pollutants is expected as an importantapproach to solve environmental problems [3–6]. It isreported that the performance of these materials areclosely related to their microstructures, particularly fortheir composition [7], size [8], and morphology [9], etc.Thus, much effort has been devoted to design the micro-structures, especially for their compositions.

Recently, studies indicate that bismuth oxyhalides (BiOX,X¼Cl, Br, and I), as a new type of semiconductor materials,have aroused great attention owing to their especial che-mical and physical structure and potential photocatalysisapplications [10–15]. Among those BiOX catalysts, BiOBr isof great research interest owing to its outstanding activeand stability [16–22]. Although BiOBr shows relatively high

x: +86 553 2871254..

stability, the application of them is limited due to itsrelatively poor regeneration performance. To solve theabove mentioned problem, much effort was explored toimmobilize the active component via the loading method.However, most of the immobilizations are a physicaladsorption process. Thus, relative poor regeneration stillexisted. Motived by above mentioned question, it should beinteresting to explore a more effective way to immobilizethe active component. Here, a facile and low cost synthesismethod was explored to immobilize BiOBr on fiber via ionschelating. The structure and composition of the productswere investigated in detail. Moreover, the photocatalyticactivity and the stability of the prepared catalysts were alsotested for the degradation of RhB.

2. Experimental

2.1. Preparation of BiOBr/amidoxime fiber

Polyacrylonitrile (PAN) fibers are obtained from com-mercial resource. All the reagents were purchased fromShanghai Chemical Company and used without furtherpurification. In a typical process, the PAN fibers (1 g) were

T.-X. Tao et al. / Materials Science in Semiconductor Processing 40 (2015) 344–350 345

first immersed in NH2OH aqueous solution (200 mL, 1 mol/L) at 65 1C for 1.5 h and named AOCF [23]. After that, theAOCF were filtrated off and then washed with distilledwater for several times. Subsequently, 1 mmol Bi(NO3)3 �5H2O were dissolved in 30 mL of ethylene glycolunder magnetic stirring for 30 min, followed by the addi-tion of 0.1 g AOCF fibers and 0.5 g CTAB (Br source). Afterstirring for 15 min, the mixture was then transferred to a50 mL Teflon-lined stainless autoclave. The autoclave wasallowed to be heated at 100 1C for 24 h under autogenouspressure, and then cooled down to room temperature. Theproduct was collected, washed with deionized water andabsolute ethanol and then dried in a vacuum at 60 1C for8 h and named BiOBr/AOCF.

2.2. Structural characterizations

The samples were characterized by X-ray powderdiffraction (XRD) with a Bruker AXS-D8X X-ray diffract-ometer with Cu Kα radiation (λ¼1.54 Å), recorded with 2θranging from 101 to 801. SEM images and EDS of thesamples were obtained on field-emission scanning elec-tron microscope (Hitachi S-4800). IR spectra wererecorded on an Aipha-Centuart FT-IR spectrometer. X-rayphotoelectron spectrum was recorded on a Thermo ESCA-LAB 250 XPS operating at the Al Kα radiation of 1486.6 eV,and all binding energy were referenced to the neutral C 1speak at 284.6 eV. Thermal gravimetric (TG) was carried outon a NETZSCH STA 449C thermoanalyzer in air atmo-sphere. The samples are heated with temperature increas-ing rate of 10 1C/min.

2.3. Catalytic testing

The photocatalytic degradation of RhB was investigatedby a BL-GHX-V photochemical reactor (Bilang Co., Shang-hai, China) equipped with a 500 W high-pressure mercurylamp. The systemwas cooled by circulating water and keptat room temperature. In a typical experiment, 0.15 g ofBiOBr/AOCF photocatalyst was dispersed into a quartzcurette containing 150 mL of RhB aqueous solution(0.02 g/L) under air at room temperature. The solutionwas continuously stirred for 60 min in the dark before thelight was turned on to ensure the establishment of anadsorption–desorption equilibrium. During irradiation,�5 mL of the suspension was continually collected fromthe cuvette at the intervals of given time. The RhBconcentration was investigated through a UV–vis spectro-photometer (YUAN XI, UV-5500) by checking the absor-bance at 553 nm. The trapping experiments of activespecies were carried out by separately adding 0.5 mlisopropanol (IPA), 0.5 mmol p-benzoquinone (p-BQ) and0.5 mmol ammonium oxalate ((NH4)2C2O4) into 50 ml RhBsolutions with 50 mg BiOBr/AOCF composites photocata-lyst, respectively. The experimental conditions were simi-lar as the former photocatalytic activity tests. In addition,after the photocatalytic tests, the reproducibility of thecatalyst was also studied only by washing and drying thecatalyst for next cycle.

3. Results and discussion

The textural analysis about the phase, composition andsuperficial area of the samples were examined by XRD,EDS and N2 adsorption–desorption. Fig. 1a shows the XRDpatterns of the BiOBr/AOCF, pure BiOBr and AOCF. As canbe seen, for the BiOBr/AOCF sample, only one relativeweak peak can be found at 2θ¼181, corresponding toamidoxime fiber. For the BiOBr sample, from the JCPDSCard 73-2061, one can easily find the sample can beindexed to a tetragonal phase of BiOBr. For the BiOBr/AOCF, not only the BiOBr diffraction peaks but also theamidoxime fiber diffraction peaks were all presented onthe XRD pattern. Moreover, the very distinct phenomenonis that the peak intensity of the amidoxime fiber in theAOCF sample is higher than that of the BiOBr/AOCFsample. Such result should be due to the covering of BiOBron the surface of the amidoxime fiber, which may promotethe chemical combining of the BiOBr and AOCF. To furtherstudy the composition, the BiOBr/AOCF sample was ana-lyzed by EDS; result is shown in Fig. 1b. As can be seen, thepeaks of Bi, Br and O can be easily found. Besides, the Cand N peaks were also presented, corresponding to ami-doxime fiber. Such results indicate that the preparedBiOBr/AOCF sample is a typical composite of BiOBr/ami-doxime fibers. The textural characteristics of the BiOBr andBiOBr/AOCF samples were studied by the nitrogen adsorp-tion–desorption and shown in Fig. 1c and d. The isothermsof the two samples presented in Fig. 1c can be classified astype IV with an H5 hysteresis loop, indicating that theobtained samples are mesoporous materials.[24,25] Fromthe pore-size distributions of the samples (Fig. 1d), one canfind that the pore-sizes of the samples all focused in 5–30 nm. According to analysis, the surface areas of theBiOBr/AOCF and BiOBr samples are 2.9 and 12.3 m2 g�1,respectively. The surface area of the sample is relativesmall, which should be ascribed to the stacking of theparticles. Obviously, the surface area of the BiOBr sample ishigh than that of the BiOBr/AOCF sample.

The morphology of the as-prepared AOCF, BiOBr/AOCF,BiOBr were observed by SEM. As can be seen in Fig. 2a, theamidoxime fibers have relatively smooth surfaces. Afterthe BiOBr nanosheets (Fig. 2d) loading, the surfaces ofamidoxime fibers were no longer smooth. As shown inFig. 2b and c, one can easily find that the amidoxime fibersare decorated with many irregular particles. Such particlesshould be ascribed to BiOBr. Clearly, BiOBr were loadedsuccessfully on the surface of the fibers, confirming theforgoing XRD analysis.

The FT-IR spectra of samples were shown in Fig. 3. For thesupport AOCF, one can find some prominent peaks at 2244,1455, 1085 and 921 cm�1, corresponding to stretching vibra-tion of nitrile groups (–CN), bending vibration of methylene(–CH2–), stretching vibration of C–N bonds, stretching vibra-tion of N–O bonds, respectively [26,27]. For the pure BiOBr, aprominent peak at 520 cm�1 can be found which is attrib-uted to Bi–O stretching mode [28]. For the BiOBr/AOCFcomposites, all the characteristic vibration bonds of AOCFand pure BiOBr were observed. Notably, the peak from Bi–Obond vibrationwas broadened and the stretching vibration ofN–O bonds was shifted to short wavenumber. Moreover, the

Fig. 2. SEM images of samples: (a) AOCF, (b, c) BiOBr/AOCF, and (d) BiOBr.

Fig. 1. (a) XRD patterns, (b) EDS, (c) N2 adsorption–desorption isotherm and (d) pore-size distributions of the samples.

T.-X. Tao et al. / Materials Science in Semiconductor Processing 40 (2015) 344–350346

stretching vibration of C–N bonds was also shifted to shortwavenumber. Such results indicate that the BiOBr particleswere combined successfully with the fiber support by che-mical effect. [28]

The XPS spectrum of BiOBr/AOCF composite was shownin Fig. 4. Distinctly, the C, N, O, Br, Bi, O and C elements canbe observed in the spectrum (Fig. 4a). As shown in Fig. 4band Table 1, the values of binding energy of N 1s in AOCF

Fig. 3. FT-IR spectra of three samples: AOCF, BiOBr/AOCF and pure BiOBr.

Fig. 4. XPS spectra of BiOBr/AOCF composites: (a) survey scan,

T.-X. Tao et al. / Materials Science in Semiconductor Processing 40 (2015) 344–350 347

were 399.3 and 400.5 eV, consistent with the two differentnitrogen atoms of the CQN and NQH for amidoximegroups, respectively [29]; however, these two values were399.0 and 400.7 eV in BiOBr/AOCF. The binding energieslocated at 164.1 and 158.8 eV were attributabled to the Bi4f5/2 and Bi 4f7/2 spin–orbital splitting photoelectrons,respectively (Fig. 4c). The splitting between these bandswas 5.3 eV, which indicates the normal state of Bi3þ inBiOBr nanoparticles.[30] The peak located at 68.3 eV canbe ascribed to Br 3d signal (Fig. 4d). Compared with thebinding energy of Bi 4f in BiOBr/AOCF and pure BiOBr, nochange can be found in the binding energy of Bi 4f. Fig. 4eshows the O 1s region of the X-ray photoelectron spectra.The sample shows three peaks around at 529.5, 531.5 and532.5 eV. The characteristic line at 529.5 eV and 531.5should be attributed to surface lattice oxygen (Olatt)

(b) N 1s, (c) Bi 4f, (d) Br 3d, (e) O 1s and (f) C 1s spectra.

T.-X. Tao et al. / Materials Science in Semiconductor Processing 40 (2015) 344–350348

species and adsorbed oxygen (Oads) species of the BiOBr,respectively, whereas the peak located at 532.5 eV wasascribed to the O species of N–O bond. Fig. 4d shows the C1s region of the X-ray photoelectron spectra. The samplepresents three peaks around at 284.6 eV, 285.9 and288.2 eV. Clearly, the characteristic line at 284.6 eV shouldbe attributed to calibration peak. The peaks located at285.9 and 288.2 eV were ascribed to the C species of C–Nbond and of the fibers support. Combined with FT-IR data,the above experimental results indicate that the Bi atom ofBiOBr may coordinate with N atom of C–N bond and Oatom of N–O bond, resulting in a novel weak interaction(Scheme 1).

To detect the content of BiOBr in BiOBr/AOCF compo-site, TG analysis was conducted. As observed in Fig. 5, theweight loss of amidoxime fiber was almost 100% when thetemperature reached to 650 1C, indicating that the fiberwas completely decomposed. For the pure BiOBr, theweight loss is about 13% from 35 to 650 1C, which shouldbe ascribed to the bound-water. For the BiOBr/AOCFcomposites, the weight loss is about 73% when thetemperature reached to 650 1C. Therefore, the value ofthe mass ratio of BiOBr in BiOBr/AOCF composite wasabout 31%. Clearly, the BiOBr/AOCF sample is consisted of

Table 1Binding energy of N, Bi, Br elements on the surfaces of BiOBr, AOCF andBiOBr/AOCF.

N 1s (eV) Bi 4f5/2 (eV) Bi 4f7/2 (eV) Br 3d (eV)

BiOBr [12] 164.2 159.0 68.5AOCF [29] 399.3

400.5BiOBr/AOCF 399.0 164.1 158.8 68.3

400.7

Scheme 1. Synthesis of BiOBr/AOCF composites.

Fig. 5. TG curves of samples: pure BiOBr, AOCF and BiOBr/AOCFcomposites.

the two constituents of BiOBr and AOCF, confirming thatBiOBr particles were successfully on the surface ofthe AOCF.

To ascertain the optical property of the samples, theoptical experiment was investigated by UV–vis DRS spec-troscopy. As shown in Fig. 6a, the pure BiOBr and AOCFsamples present very distinct absorption around 200–380 nm. It is almost no response at visible light regionfor them. However, the BiOBr/AOCF sample exhibits abroad absorption band from 200 to 650 nm. Distinctly,the optical absorption of the BiOBr/AOCF sample is stron-ger than those of the pure BiOBr and AOCF sample at thevisible light region. Such phenomenon should be assignedto the existence of coordination effect, indicating that theBiOBr/AOCF sample is more effective visible light responsematerial.

The equation (αhv)2¼K(hv�Eg) was used to calculatethe band gap of the samples, where α is the absorptioncoefficient, K is a proportionality constant, and Eg is theband-gap energy. As can be seen in Fig. 6b, the Eg valuesare estimated to be about 3.0, 2.9 and 1.7 eV for the pureBiOBr, AOCF and BiOBr/AOCF composites, respectively.Obviously, the optical band gap of the BiOBr/AOCF samplesis less than those of the pure BiOBr and AOCF samples,indicating that the band gap stronger depends on thecoordination effect. The decrease of band gap for theBiOBr/AOCF sample indicates that light harvest and chargecarrier separation can be achieve more easily, whichshould be helpful to improve photocatalytic reactionefficiencies.

The photocatalytic activity of BiOBr/AOCF composite forthe degradation of RhB was evaluated under UV lightirradiation, results are shown in Fig. 7. The absorptionpeak of RhB at λ¼553 nm was used to monitor thephotocatalytic degradation process. As can be seen inFig. 7a, with the prolonging of the irradiation time, themain absorption peak weakens. Moreover, a remarkableblue-shift of the maximum absorption occurred, suggest-ing that the conjugated structure of Rhodamine B wasdestructed and the formation of some new intermediatesmay exist in the process of degradation, deethylation ofthe fully N, N, N0, N0-tetraethylated rhodamine molecule(i.e., RhB) has the wavelength position of its major absorp-tion band moved toward the blue region, λmax, RhB,553 nm; N, N, N0-triethylated rhodamine, 541 nm; N, N0-diethylated rhodamine, 528 nm; N-ethylated rhodamine,503 nm; and rhodamine, 492 nm. RhB formed rhodamineafter being fully demethylated, and the rest was degradedthrough destruction of the conjugated structure [31,32].The completely disappearance of the absorption peaksafter illumination for 60 min indicated the complete deco-lorization of the RhB solution. As RhB is quite stable and itsself-photolysis can be neglected (Fig. 7b), the presentdecomposition of RhB mainly originated from the photo-catalysis induced by the BiOBr/AOCF composites photo-catalyst. Furthermore, the comparative experiments ofphotocatalytic degradation of RhB by BiOBr/AOCF compo-sites and pure BiOBr were performed under the sameexperimental conditions. The mass of pure BiOBr and theBiOBr in BiOBr/AOCF composites are identical. The resultswere displayed in Fig. 6b. The degradation efficiencies

Fig. 6. (a) UV–vis absorption spectrums and (b) the corresponding plots of (αhv)2 vs. hv curves for the samples: pure BiOBr, AOCF and BiOBr/AOCFcomposites.

Fig. 7. (a) UV–vis absorption spectra of the photodegradation of RhB and (b) the photodegradation efficiency.

Fig. 8. (a) Photocatalytic performance and (b) catalytic mechanism of BiOBr/AOCF composites.

T.-X. Tao et al. / Materials Science in Semiconductor Processing 40 (2015) 344–350 349

were about 100% and 87% for as-prepared and pure BiOBrat 60 min, respectively. Obviously, the photocatalytic activ-ity of BiOBr/AOCF composites was higher than that of thepure BiOBr. This may be because the coordination betweenBiOBr and AOCF is in favor of the separation of photo-generated electron–hole pairs, increasing the photocataly-tic activity of composites.

To analyze the mechanism of photodegradation of RhBby BiOBr/AOCF composites, the trapping experiments of

active species were carried out; results are shown inFig. 8a. Isopropanol (IPA, a quencher of �OH), p-benzoquinone (p-BQ, a quencher of �O2�) and ammoniumoxalate ((NH4)2C2O4, a quencher of hþ) were generallyused to detect the main active species in the photocatalyticreaction [33]. From Fig. 8a, it was found that the degrada-tion of BiOBr/AOCF composites on RhB was 97% and theaddition of isopropanol system did not change the photo-catalytic activity of BiOBr/AOCF composites in degrading

Fig. 9. Photocatalytic activity with five times of recycling testing.

T.-X. Tao et al. / Materials Science in Semiconductor Processing 40 (2015) 344–350350

RhB, which indicates that the hydroxyl radicals are not themain active species in BiOBr/AOCF composite photocata-lysts for the degradation of RhB [34]. In contrast, thephotocatalytic degradation percentage 28% and 34% ofRhB decreased obviously when ammonium oxalate andp-benzoquinone were added, respectively. The aboveexperimental results indicate that hþ and O�

2 played moreimportant roles than OH in the photodegradation of RhB[35,36]. A proposed photocatalytic mechanism for thedegradation of RhB over the BiOBr/AOCF composite photo-catalysts is presented in Fig. 8b. From the forgoing analysis,one can easily find that the AOCF not only plays a supportbut also acts as the ligands, which should be helpful toincreasing the separation efficiency of photo-generatedelectron–hole pairs.

We noted that BiOBr/AOCF composites could be easilyseparated and recovered from solution because of theirlarge length to diameter ratios and flexible property. Thischaracter confirmed their well reusable property which wasalso investigated by collected, washed with deionized waterand absolute ethanol and reusing for multiple cycles. Eachreusability experiment was carried out under the identicalcondition. As shown in Fig. 9, the photocatalytic activity ofBiOBr/AOCF composites was almost not changed after 5cycles of photodegradation of RhB. Such result confirmedthat the microstructure of BiOBr/AOCF composites is stabi-lity. The excellent stability of photocatalysts might beattributed to the strong coordination between BiOBr andAOCF molecules of BiOBr/AOCF composites. The good sta-bility and recoverable property is very helpful to enhancingthe practical applications of BiOBr/AOCF composites ineliminating organic pollutants from wastewater.

4. Conclusion

In summary, BiOBr/amidoxime fiber composite photo-catalysts have been successfully prepared via a facilesolvothermal synthesis approach. The photocatalytic activ-ity of the as-obtained photocatalysts was evaluated by thephotodegradation of RhB. The photocatalysts exhibited asignificant enhancement of photocatalytic performance inthe degradation of RhB. Notably, the BiOBr/amidoxime fibercomposite photocatalysts not only facilitate the separationof photocatalyst from dye solution because of their large

length to diameter ratios but also retain high photocatalyticstability after several cycles owing to the coordinationbetween BiOBr and amidoxime fiber molecules.

Acknowledgments

We gratefully acknowledge financial support from theNatural Foundation of Anhui Province (KJ2014ZD03) andthe National Natural Science Foundation of China(21171003 and 51202003).

References

[1] X.B. Chen, S.H. Shen, L.J. Guo, S.S. Mao, Chem. Rev. 110 (2010)6503–6570.

[2] A. Kudo, Y. Miseki, Chem. Soc. Rev. 38 (2009) 253–278.[3] F.F. Duo, C.M. Fan, Y.W. Wang, Y.Q. Cao, X.C. Zhang, Mater. Sci.

Semicond. Process. 38 (2015) 157–164.[4] V. Preethi, S. Kanmani, Mater. Sci. Semicond. Process 16 (2013)

561–575.[5] Y. Jiao, Y. Liu, F.Y. Qu, A. Umar, X. Wu, J. Colloid. Interface Sci. 451

(2015) 93–100.[6] W. Zhu, P.J. Liu, S.N. Xiao, W.C. Wang, D.Q. Zhang, H.X. Li, Appl. Catal.

B-Environ. 172 (2015) 46–51.[7] H.Y. Jiang, H.X. Dai, X. Meng, K.M. Ji, L. Zhang, J.G. Deng, Appl. Catal.

B-Environ. 105 (2011) 326–334.[8] Y. Yamada, K. Yano, S. Fukuzumi, Energy Environ. Sci. 5 (2012)

5356–5363.[9] L.L. Li, C.Y. Nan, Q. Peng, Y.D. Li, Chem. Eur. J. 18 (2012) 10491–10496.

[10] C.H. Deng, H.M. Guan, Mater. Lett. 107 (2013) 119–122.[11] X.M. Zhang, G.B. Ji, Y.S. Liu, X.G. Zhou, Y. Zhu, D.N. Shi, P. Zhang, X.

Z. Cao, B.Y. Wang, Phys. Chem. Chem. Phys. 17 (2015) 8078–8086.[12] L. Chen, R. Huang, M. Xiong, Inorg. Chem. 52 (2013) 11118–11125.[13] W.D. Zhang, Q. Zhang, F. Dong, Ind. Eng. Chem. Res. 52 (2013)

6740–6746.[14] H.F. Cheng, B.B. Huang, Y. Dai, Nanoscale 6 (2014) 2009–2026.[15] A. Dash, S. Sarkar, V.N.K.B. Adusumalli, V. Mahalingam, Langmuir 30

(2014) 1401–1409.[16] F.T. Li, Q. Wang, J.R. Ran, Y.J. Hao, X.J. Wang, D.S. Zhao, S.Z. Qiao,

Nanoscale 7 (2015) 1116–1126.[17] S. Shenawi-Khalil, V. Uvarov, S. Fronton, I. Popov, Y. Sasson, J. Phys.

Chem. C 116 (2012) 11004–11012.[18] L. Kong, Z. Jiang, T.C. Xiao, L.F. Lu, M.O. Jones, P.P. Edwards, Chem.

Commun. 47 (2011) 5512–5514.[19] H.F. Cheng, B.B. Huang, P. Wang, Z.Y. Wang, Z.Z. Lou, J.P. Wang, X.

Y. Qin, X.Y. Zhang, Y. Dai, Chem. Commun. 47 (2011) 7054–7056.[20] H.F. Cheng, B.B. Huang, Z.Y. Wang, X.Y. Qin, X.Y. Zhang, Y. Dai, Chem.

Eur. J. 17 (2011) 8039–8043.[21] L. Zhang, X.F. Cao, X.T. Chen, Z.L. Xue, J. Colloid Interface Sci. 354

(2011) 630–636.[22] Y.J. Chen, M. Wen, Q.S. Wu, CrystEngComm 13 (2011) 3035–3039.[23] Z.C. Wu, Y. Huang, Y.N. Lu, T.X. Tao, Z. Zhang, Catal. Commun. 29

(2012) 158–161.[24] A.A. Abouelsaood, J. Appl. Phys. 91 (2002) 2753–2759.[25] S. Janitabar Darzi, J. Iran. Chem. Soc. 11 (2014) 1363–1371.[26] C.Y. Su, J. Liu, C.L. Shao, J. Non-Cryst. Solids 357 (2011) 1488–1493.[27] L.W. Ji, X.W. Zhang, Mater. Lett. 62 (2008) 2161–2164.[28] Y.C. Chou, C.L. Shao, X.H. Li, C.Y. Su, H.C. Xu, M.Y. Zhang, P. Zhang,

X. Zhang, Y.C. Liu, Appl. Surf. Sci. 285P (2013) 509–516.[29] Z.C. Wu, Y. Zhang, T.X. Tao, L.F. Zhang, H. Fong, Appl. Surf. Sci. 257

(2010) 1092–1097.[30] F.D. Gao, D.W. Zeng, Q.W. Huang, S.Q Tian, C.S. Xie, Phys. Chem.

Chem. Phys. 14 (2012) 10572–10578.[31] T. Watanabe, T. Takizawa, K. Honda, J. Phys. Chem. 81 (1977)

1845–1851.[32] C. Zhang, Y.F. Zhu, Chem. Mater. 17 (2005) 3537–3545.[33] Y. Xu, S.C. Xu, S. Wang, Dalton Trans. 43 (2014) 479–485.[34] J. Li, W.H. Ma, Y.P. Huang, Appl. Catal. B 48 (2004) 17–24.[35] H. Gnayem, Y. Sasson, ACS Catal. 3 (2013) 186–191.[36] L.Q. Ye, C.Q. Gong, J.Y. Liu, J. Mater. Chem. 22 (2012) 8354–8360.