Photocatalytic degradation of X-3B by titania-coated magnetic activated carbon under UV and visible irradiation

Journal of Alloys and Compounds 471 (2009) 33–38

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jallcom

Photocatalytic degradation of X-3B by titania-coated magnetic activated carbon under UV and visible irradiation Yanhui Ao a,b,c , Jingjing Xu a,b,c , Degang Fu a,b,c,∗ , Chunwei Yuan a,b a

School of Chemistry and Chemical Engineering, Southeast University, Nanjing 210096, China State Key Laboratory of Bioelectronics, Southeast University, Nanjing 210096, China c Key Laboratory of Environmental and Bio-Safety in Suzhou, Research Intitute of Southeast University, Dushu Lake Hogher Education Town, Suzhou 215123, China b

a r t i c l e

i n f o

Article history: Received 15 January 2008 Received in revised form 21 March 2008 Accepted 1 April 2008 Available online 15 May 2008 Keywords: Magnetic activated carbon Magnetically separable Photocatalysis Titania X-3B

a b s t r a c t A novel magnetically separable composite photocatalyst, titania-coated magnetic activated carbon (TMAC), was prepared by depositing of anatase titania onto the surface of magnetic activated carbon (MAC). The MAC was prepared by a simple route: directly adsorbing of magnetic Fe3 O4 nanoparticles onto activated carbon. The prepared samples were characterized by XRD, scanning electron microscopy (SEM) and vibrating sample magnetometer (VSM). The photocatalytic activity of the samples was determined by degradation of reactive brilliant red X-3B under either UV or visible irradiation, and compared to Degussa P25. The composite photocatalyst can be separated easily from solution by a magnet, its photocatalytic activity in degradation of X-3B also has dramatic enhancement compared to that of P25 titania. The composite photocatalyst can also be reused with a little reduction of its photocatalytic activity. The degradation rates of X-3B were both higher than 80% after six cycles under UV and visible irradiation. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Titania-mediated photocatalytic oxidation and reduction offers potentially a facile and cheap method for removing inorganic and organic pollutants from waste waters. It has been an area of intense interest for the past 20 years, particularly for removing organic compounds as they can be completely mineralized under photocatalytic oxidation [1–4]. Typically, a photocatalytic reaction is conducted in a suspension of submicrometer-sized titania, and therefore requires an additional separation step to remove the catalyst from the treated water. Removing such fine particles from large volumes of water involves further expense and manpower. This is a major drawback for the practical application of the photocatalytic processes [5]. Pozzo et al. stated that the cost of the separation may invalidate altogether the claimed energy savings for a solar-induced decontamination process due to the small particle size of the used photocatalyst [6]. To solve the separation problem of photocatalyst, researches have been carried out by immobilizing titania onto various substrates, such as glass beads [7], sand, silica gel, activated carbon

∗ Corresponding author at: School of Chemistry and Chemical Engineering, Southeast University, Nanjing 210096, China. Tel.: +86 25 83794310; fax: +86 25 83793091. E-mail address: [email protected] (D. Fu). 0925-8388/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jallcom.2008.04.001

fibers [8], and quartz optical fibers [6]. Although this approach provides a solution to the solid–liquid separation problem, slurry-type reactors offer significant advantages over immobilized-catalysttype reactors because of the catalyst surface availability and superior mass-transfer properties [9]. Some investigators prepared nanoparticles with magnetic core and photoactive shell using magnetic granules and titania [10–13]. They proved that the nanoparticles had magnetic property and could be separated easily by magnetic materials. But the photocatalytic activity of the nanoparticles declined. Matos et al. [14] investigated the photocatalytic activity of a suspended mixture of titania and activated carbon. They found that there were a synergistic effect and a common interface, which contributed to the higher photocatalytic activity, between titania and activated carbon. Some others investigated the photocatalytic degradation of aqueous organic pollutants by titania-coated activated carbon [15–18]. They also found enhanced photocatalytic activity of these composite photocatalysts. The separation problem still exists because the catalyst is size in micron, although it can be separated more easily than the slurry titania system. Our work is concerned on fabricating composite photocatalyst with enhanced photoactivity and recyclability. So, we associate the adsorption activity of activated carbon with the magnetic separability of Fe3 O4 and photocatalytic activity of titania. The separation problem of the photocatalysts will be solved easily and the photocatalytic activity of them enhance a lot at the same time.

34

Y. Ao et al. / Journal of Alloys and Compounds 471 (2009) 33–38 FEI). The magnetic measurements were carried out with a vibrating sample magnetometer (VSM, PARR, Model 4500). UV–vis absorption spectra of irradiated samples were measured every 20 min with a UV–vis spectrophotometer (8500, China). The Fe content of the final water was analysised by ICP-AES (PerkinElmer, ELAN9000) to investigate whether iron is entering solution after the irradiation under both UV and visible light. 2.6. Photocatalytic studies

Scheme 1. Molecular structure mula = C19 H10 O7 N6 C12 S2 Na2 ).

of

dye

X-3B

(chemical

for-

The production and consumption of azo dyes account for over 50% of the total dyes around the world. Reactive brilliant red X3B is a kind of typical azo dye. Therefore, X-3B was chosen as a representative model compound. The photocatalytic degradation of X-3B was examined in an aqueous dispersion of the composite photocatalyst under either UV or visible irradiation. The degradation behavior of X-3B was also investigated for macroscopic degradation kinetics. Furthermore, the efficiency of the photocatalyst has also been evaluated by recycling the photocatalyst to check the economic viability of this method. 2. Experimental 2.1. Materials Activated carbon (AC) was purchased from ShangHai Activated Carbon Ltd. Titanium dioxide (TiO2 ) powder sample used in the experiment was commercial Degussa P25 with 80% anatase, 20% rutile and BET area of ca. 50 m2 g−1 , which was produced by the Degussa AG Company in Germany. Ti(OBu)4 was chosen as a Ti precursor, which is less reactive than titanium chloride and titanium isopropoxide. Ti(OBu)4 used in the experiment was chemical pure grade. Nitric acid (HNO3 ) and isopropanol (PrOH) were analytical reagent grade. Distilled water was additionally cleaned prior to its use with a quartz sub-boil high purity water purification system. The substrate of X-3B dye was obtained from Shanghai Dyestuff Chemical Plant and used without further purification. Scheme 1 displays the structure of reactive brilliant red dye X-3B. 2.2. Preparation of titania sol For preparation of TiO2 sol, the detailed procedure was as following: Ti(OBu)4 diluted with PrOH was added dropwise into water under vigorous stirring, whose acidity was adjusted with HNO3 . The molar ratios of PrOH and water to Ti(OBu)4 were 1.42 and 151, respectively. Then, the solution was kept under reflux condition (around 75 ◦ C) for 24 h. Finally, pure TiO2 sol was obtained, which was used to prepare thin films and equivalent powders, after PrOH and n-butyl alcohol were removed from the solution in a rotatory evaporator under vacuum. 2.3. Preparation of magnetic activated carbon (MAC) Magnetic magnetite particles were prepared by chemical co-precipitation method as the following procedure: definite amount of FeSO4 and FeCl3 were dissolved in distilled water which was bubbled with N2 , the molar ratio of Fe3+ and Fe2+ was 5:3. Then, definite NH4 OH was added dropwise into the solution at vigorous stirring till the pH 9. Finally, black magnetite sol was obtained. MAC were prepared in the following way: 3 g of activated carbon powder, which washed for three times by distilled water, was added into 200 mL of diluted Fe3 O4 sol whose concentration is 3 g L−1 . Then the admixture was stirred for 1 h. The solid phase of Fe3 O4 /AC was separated by a magnet and then dried at 40 ◦ C.

The potoreactor was a glass beaker (250 mL capacity) covered by silver paper, UV and visible irradiation was provided by a 200-W ultraviolet lamp and a halogen 500-W lamp (Instrumental corporation of Beijing Normal University) with a light filter cutting the light below 400 nm, respectively. In an ordinary photocatalytic test performed at room temperature, 0.2 g photocatalyst was added under stirring into 200 mL of X-3B aqueous solution whose concentration was 50 mg L−1 and maintained in the dark for 0.5 h to reach complete adsorption and desorption equilibrium before the irradiation start. Afterward samples of the suspension (5 mL) were removed and filtered through one layer of Millipore 0.22 ␮m films at regular intervals of 20 min before spectrum analysis.

3. Results and discussion 3.1. Characterization results XRD patterns of prepared samples are presented in Fig. 1. Fig. 1(a) shows XRD pattern of MAC, presenting the characteristic peaks signed by M (30.4◦ , 35.7◦ , 43.4◦ , 53.8◦ , 57.4◦ and 63.0◦ ) of cubic spinel structure. The iron oxide can be present as magnetite and maghemite because part of Fe3 O4 particles may be oxidized to ␥-Fe2 O3 . It can also be seen from Fig. 1(b) that the magnetic particles maintain cubic spinel structure in MAC and TMAC samples. This implies that the magnetic properties of the magnetic particles are basically invariable. The TiO2 coating layer has an anatase structure in TMAC sample, determined from XRD pattern in Fig. 1(b) where the peaks signed by A (25.4◦ , 38.0◦ , 48.0◦ and 54.7◦ ) are the characteristic peaks of anatase structural TiO2 . It is believed that the photocatalytic effects of TiO2 are mainly due to anatase structural TiO2 . In the previous reported magnetic iron oxide/titanium dioxide photocatalysts, some structural characteristics formed during calcination at high temperature have been considered to lead a decrease in the photoactivity [19–22]. For example, it had been found that the presence of the iron oxide could cause anatase to rutile phase transformation of the coating [22,23], or the possible formation of a mixed iron/titanium oxide (pseudobrookite, Fe2 TiO5 ) [24,25]. Furthermore, the calcination at high temperature would also lead to the aggregation and enlargement of TiO2 nanoparticles, which involve decrease of surface area and losses of species such as hydroxyl and adsorbed water that dominate the surface chemistry and adsorption activity of titania [26].

2.4. Preparation of titania-coated magnetic activated carbon (TMAC) To obtain titania-coated magnetic activated carbon (TMAC), the coating procedure of titania onto the magnetic activated carbon was as following: 3 g of MAC was dispersed in titania sols in an ultrasonic bath for 1 h. Then, it was dried into powders in a rotatory evaporator under vacuum at 75 ◦ C. At last, the mass ratio of Fe3 O4 /AC/TiO2 was 1:5:12 in TMAC. 2.5. Instruments and analysis methods The crystalline structure of the prepared samples were determined by X-ray diffractometer (XD-3A, Shimadazu Corporation, Japan) using graphite monochromatic copper radiation (Cu K␣) at 40 kV, 30 mA over the 2 range 20–80◦ . The morphologies were characterized with a scanning electron microscopy (SEM, Sirion,

Fig. 1. XRD patterns of prepared samples: (a) MAC and (b) TMAC.

Y. Ao et al. / Journal of Alloys and Compounds 471 (2009) 33–38

35

Fig. 2. SEM micrograph of (a) MAC and (b) TMAC.

Our preparation and coating process at low temperature avoided or reduced the structural changes of TiO2 sol when it deposited onto MAC as demonstrated from XRD patterns. The results also show that the AC did not influence the crystal structure of TiO2 and iron oxide deposited on it during our preparation. In obtained composites, there is no new compound has been formed. Compared to XRD, Raman spectroscopy is a much more sensitive technique for the detection of nanosized crystalline domain [27]. XRD is a useful tool to determine the period crystal structure with a long range, but its detection limit is relative high, about 5 mol%. Raman spectroscopy is sensitive to some vibration modes in crystal structure [28]. Fig. 2 shows Raman spectra of TMAC. As shown in the figure, the Raman bands around 146, 398, 516 and 639 cm−1 attributed to the single-crystal anatase titania [29]. Raman band presented around 360 cm−1 attributed to the Fe–O streching modes of the maghemite phase [30]. Other bands can be attributed to magnetite phase. It illustrates that part of magnetite was oxidized to maghemite phase. After the deposition of Fe3 O4 , part of AC surface has been coated by Fe3 O4 nanoparticles with little aggregation (Fig. 3(a)). Then, the titania nanoparticles were deposited. Some titania particles deposited on AC and some deposited on Fe3 O4 nanoparticles which coated on AC. The SEM study showed a homogeneous distribution of titania nanoparticles with uniform size (Fig. 3(b)). Although the puckered morphology of AC has been smoothed to a certain extent, the titania layer is still porous as seen from SEM images (Fig. 3(b)). Table 1 shows the results from surface area measurements of the different samples. As it can be seen, the higher AC contents of the catalysts the larger surface area. The results illustrates that TMAC still shows high pore volume and surface area, which is in agreement with SEM results. The magnetic properties of MAC and TMAC were measured by VSM, as shown in Fig. 4. The magnetic parameters such as saturation magnetization Ms , coercivity Hc and remanent magnetization Mr were given in Table 2. The decrease of saturation magnetization Ms of TMAC compared to MAC is consistent with their Fe3 O4 content in unit weight sample. The low values of Hc and Mr which are close to

Fig. 3. Magnetization vs. applied magnetic field for the (a) MAC and (b) TMAC.

zero indicate that the prepared samples exhibited superparamagnetic behaviors at room temperature [31]. The superparamagnetic behaviors of the prepared TMAC make the photocatalyst can be separated more easily by a magnet or an applied magnetic field. In the meantime, the very low remanent magnetization largely reduced the aggregation of catalyst after it was separated by applied magnetic field from original reaction solution, so the photocatalyst can be easily redispersed in a solution for reuse. 3.2. Photocatalytic activity In the previous articles, we can know the occurrence of photodissolution in titanium/iron oxides has been frequently encountered [5,13]. To investigate whether this happened in our

Table 1 Surface area measurement of the different samples Sample

BET-specific surface (m2 g−1 )

Pore volume (cm3 g−1 )

AC MAC TMAC

1093.07 870.50 373.56

0.54 0.46 0.21

Fig. 4. Kinetics of X-3B disappearance in the presence of TMAC under (a) visible and (b) UV irradiation, in the presence of MAC under (c) visible and (d) UV irradiation.

36

Y. Ao et al. / Journal of Alloys and Compounds 471 (2009) 33–38

Table 2 Magnetic properties of the different samples Samples

Coercivity, Hc (Oe)

Remanent magnetization, Mr (emu g−1 )

Saturation magnetization Ms (emu g−1 )

Fe3 O4 MAC TMAC

50.245 11.757 12.552

4.26 0.2 0.17

80.79 9.5 6.45

Table 3 Conduction band potentials, valence band potentials and band gap of titania, magnetite, maghemite at pH 3.5

Conduction band potential (V) Valence band potential (V) Band gap (V)

TiO2

Fe3 O4

␥-Fe2 O3

−0.04 3.16 3.2

0.17 0.27 0.10

0.29 2.59 2.30

system, an ICP analysis of the final water has been carried out. The results show that the concentrations of iron ions were all lower than 0.60 mg L−1 for UV and visible light-irradiated samples. It illustrates that the X-3B was photocatalyticly degraded but not oxidized by photo-Fenton reaction because the amount of Fe is negligible. The photoactivity of the prepared photocatalyst was assessed by applying it to degrade model contaminated water of X-3B aqueous solution whose initial concentration was 50 mg L−1 . For comparison, the photocatalytic degradation of X-3B by P25 under UV or visible irradiation was also investigated. The curves of the photocatalytic degradation of X-3B by MAC and TMAC under UV and visible light are ploted in Fig. 5. It can be seen that MAC has little photocatalytic activity. The results by Degussa P25 were shown in Fig. 6. The usage amount of the catalyst was both 1 g L−1 for TMAC and P25, respectively. It can be seen that MAC shows little photocatalytic activity under both UV and visible irradiation. The apparent first-order kinetic equation ln(C0 /C) = kapp t was used to fit experimental data, where kapp is apparent rate constant, C is the solution-phase concentration of X-3B, and C0 is the initial concentration at t = 0 [14]. The linear transforms in ln(C0 /C) as a function of irradiation time given in Fig. 7 (for Fig. 5) and Fig. 8 (for Fig. 6) confirmed that the kinetic curves in Figs. 5 and 6 are of apparent first order. Therefore, the apparent rate constant was chosen as the basic kinetic parameter for the different photocatalysts since it enables one to determine a photocatalytic activity independent of the previous adsorption period in the dark and the concentration of X-3B remaining in the solution. It can be seen from the figure that the photocatalytic activity of TMAC was higher than P25 under both UV and visible irradiation. The apparent rate con-

Fig. 5. Kinetics of X-3B disappearance in the presence of P25 under (a) visible and (b) UV irradiation, blank reaction under (c) UV and (d) visible irradiation.

Fig. 6. Linear transform ln(C0 /C) = f(t) of the kinetic curves of X-3B disappearance for TMAC from Fig. 4.

stant were 0.02884 and 0.01961 min−1 for TMAC and P25 under UV irradiation, 0.02251 and 0.0088 min−1 for TMAC and P25 under visible irradiation, respectively. This can be attributed to the difference in the photocatalytic mechanisms of X-3B under UV and visible irradiation. Under visible irradiation, the X-3B molecule can be excited to appropriated excited states, then the excited X-3B undergoes reactions by two concurrent routes: (1) a route involving singlet oxygen, which was generated by energy transfer from the excited X-3B molecule to ground state oxygen, then the X-3B molecules are oxidized by the singlet oxygen; it can be seen from the data of degraded X-3B in the blank reaction (in the absence of any solids). (2) electron transfer from an excited state of X-3B to the conduction band of titania particles, and the injected electrons of the semiconductor can react with the adsorbed oxygen to form oxidizing species which can bring about the photooxidation of X-3B. In the presence of TMAC and P25, it may be both undergo the photosensitized degradation of X-3B molecule. In addition, we

Fig. 7. Linear transform ln(C0 /C) = f(t) of the kinetic curves of X-3B disappearance for P25 from Fig. 5.

Y. Ao et al. / Journal of Alloys and Compounds 471 (2009) 33–38

37

Table 4 Photocatalytic data for different initial concentration of X-3B Sample

Initial X-3B concentration (mg L−1 )

UV

Visible

Equilibrium concentration

Concentration after 80 min irradiation

kapp (min−1 )

Equilibrium concentration

Concentration after 80 min irradiation

kapp (min−1 )

P25

75 100

29.31 45.73

9.89 19.21

0.015 0.011

29.13 45.55

17.88 32.51

0.0064 0.0044

TMAC

75 100

40.88 59.22

7.64 16.98

0.021 0.016

41.22 59.03

12.69 24.38

0.015 0.011

Fig. 8. Recycle of TMAC under UV irradiation.

can observe that the TMAC may be show the activity of iron doped titania which had been proved to has visible-light-induced photocatalytic activity [32,33]. Therefore, the photocatalytic activity of P25 was much lower than TMAC under visible irradiation. Under UV irradiation, although the photo-sensitization process of the X-3B still exists to some extent, it is titania but not the X3B molecule that mainly absorb the photon with energy larger than the band gap of titania to generate electrons and holes in the conduction and valence bands, respectively. Then the photogenerated holes can react with adsorbed hydroxy or H2 O to form hydroxy radicals which can oxidize the X-3B molecule. While the photogenerated electron can react with the adsorbed oxygen yield superoxide radical anion which can also oxidize the X-3B molecule. From the results we can observed that P25 adsorbed much more X-3B than TMAC. So, more active site of P25 was occupied by X-3B

Fig. 9. Recycle of TMAC under UV irradiation.

molecule and less UV light can be absorbed by P25. This results is in agreement with the investigation of Xu et al. [34]. In addition, when two or more semiconductors are in contact, electronic interactions occur at the point of contact of the different phases (heterojunction), leading to the transfer of charge carriers across this junction if this is thermodynamically feasible. It is important to note that the magnetite were partly oxidized to maghemite when the titania particles were deposited onto the magnetic particles. Therefore, the magnetic particles of iron oxide will be present as a mixture of magnetite and maghemite phase. When comparing the relative positions of the conduction band and valence bands in Table 3, it is evident that it is thermodynamically feasible for the generated electrons and holes in titania to be transferred to the lower lying conduction band and upper lying valence band of the iron oxides of both phase, respectively. So, the recombination of the light induced holes and electrons can be reduced. All in all, the higher photoactivity of TMAC compared to P25 under UV irradiation can be ascribed to the above reasons. The photocatalytic degradation of other initial concentrations (75 and 100 mg L−1 ) was also investigated, results are shown in Table 4. It can be seen from the table that apparent rate constant of TMAC are all higher than that of P25 either under UV or visible irradiation. Furthermore, we can observe that absolute degraded amount of X-3B increases with the increasing of initial concentration for both TMAC and P25. But the apparent rate constant decreases with the increasing of initial concentration of X-3B. The results are in agreement with Xu et al. [34]. All together, the photocatalytic activity of TMAC is higher than P25. 3.3. Recycle of the photocatalyst The regeneration of TiO2 photocatalyst was one of the key steps to develop heterogeneous photocatalysis technology for practical applications. An examination of the photocatalytic activity of the

Fig. 10. Recycle of TMAC under visible irradiation.

38

Y. Ao et al. / Journal of Alloys and Compounds 471 (2009) 33–38

recycled TMAC by a magnetic field was carried out. No particles were detected in the degraded solution after the TMAC photocatalyst had been separated by a magnetic field. The separated TMAC was utilized to photodegrade the X-3B repeatedly under UV or visible irradiation as presented in Figs. 9 and 10, respectively. The photocatalytic activity of the recycled photocatalyst decreased a little after six cycles. It shows that the separation of photocatalysts by the magnetic field was effective and the reproducibility of photocatalytic activity of the recycled photocatalyst TMAC was very good. 4. Conclusion We have illustrated a procedure for successfully fabricating MAC through the adsorption of iron oxide magnetic nanoparticles onto the surface of a commercial activated carbon. This method suggests a simple route for the synthesis of magnetically separable activated carbons. Then the anatase titania particles prepared at low temperature were deposited onto the MAC. The TMAC thus prepared shows super-paramagnetic properties and can be separated easily by an external magnetic field. It can also be redispersed into aqueous solution by ultrasonic vibrating after removing the external magnetic field. The photocatalytic activity of TMAC and P25 titania was determined using reactive brilliant red X-3B as a model compound under either UV or visible irradiation. The experiment results show that the photocatalytic activity of TMAC was greater than Degussa P25 under both UV and visible irradiation. The recycle of TMAC was also investigated. The degradation rates of X-3B were 83 and 81% under UV and visible irradiation after six cycles, respectively. Acknowledgements This work was supported by the National Natural Science Foundation of China (No. 60121101) and Joint Project of Guangdong Province and Education Department (No. 2007A090302018). We are also very grateful to Dr. Liu Ji-wei in Southeast University for his help in VSM experiments.

References [1] A.L. Linsebigler, G.Q. Lu, J.T. Yates, Chem. Rev. 95 (1995) 735–758. [2] M.R. Hoffmann, S.T. Martin, W.Y. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995) 69–96. [3] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C 1 (2000) 1–21. [4] S. Malato, J. Blanco, A. Vidal, C. Richter, Appl. Catal. B 37 (2002) 1–15. [5] D. Beydoun, R. Amal, J. Phys. Chem. B 104 (2000) 4387–4396. [6] R.L. Pozzo, M. Baltanas, A. Cassano, Catal. Today 39 (1997) 219–231. [7] N.B. Jackson, C.W. Wang, Z. Luo, J. Schwitzgebel, J.G. Ekerdt, J.R. Brock, A. Heller, J. Electrochem. Sci. 138 (1991) 3660–3664. [8] P.F. Fu, Y. Luan, X.G. Dai, J. Mol. Catal. A 221 (2004) 81–88. [9] R. Matthews, Water. Res. 24 (1990) 653–660. [10] D. Beydoun, R. Amal, Mater. Sci. Eng. B 94 (2002) 71–81. [11] Y.S. Chung, S.B. Park, D.W. Kang, Mater. Chem. Phys. 86 (2004) 375–381. [12] F. Chen, Y.D. Xie, J.C. Zhao, G.X. Lu, Chemosphere 44 (2001) 1159–1168. [13] D. Beydoun, R. Amal, G. Low, S. McEvoy, J. Mol. Catal. A 180 (2002) 193–200. [14] J. Matos, J. Laine, J.M. Herrmann, Appl. Catal. B 18 (1998) 281–291. [15] B. Tryba, A.W. Morawski, M. Inagaki, Appl. Catal. B 41 (2003) 427–433. [16] M. Toyoda, Y. Nanbu, T. Kito, M. Hirano, M. Inagaki, Desalination 159 (2003) 273–282. [17] E. Carpio, P. Zuniga, S. Ponce, J. Solis, J. Rodriguez, W. Estrada, J. Mol. Catal. A 228 (2005) 293–298. [18] J. Arana, J.M. Dona-Rodriguez, E.T. Rendon, C.G.I. Cabo, O. Gonzalez-Diaz, J.A. Herrera-Melian, et al., Appl. Catal. B 44 (2003) 161–172. [19] J.A. Navio, M. Macias, M. Gonzalez-Catalan, A. Justo, J. Mater. Sci. 27 (1992) 3036–3042. [20] R.I. Bickley, T. Gonzalez-Carreno, L. Palmisano, R.J.D. Tilley, J. Williams, Mater. Chem. Phys. 51 (1997) 47–53. [21] A. Milis, J. Peral, X. Domenech, J. Mol. Catal. 87 (1994) 67–74. [22] M.I. Litter, J.A. Navio, J. Photochem. Photobiol. A 84 (1994) 183–193. [23] R.P. Madhusudhan, B. Viswanathan, P.P. Viswanath, J. Mater. Sci. 30 (1995) 4980–4985. [24] J. Yuan, S. Tsujikawa, Zairyo-to-Kankyo 44 (1995) 534–542. [25] J. Huang, T. Shinohara, S. Tsujikawa, Zairyo-to-Kankyo 47 (1997) 651–661. [26] R. Howe, M. Gratzel, J. Phys. Chem. 91 (1987) 3906–3909. [27] M. Ivanda, S. Music, M. Gotic, A. Turkovic, A.M. Tonejc, O. Gamulin, J. Mol. Struct. 480 (1999) 641–644. [28] G. Liu, Z.G. Chen, C.L. Dong, Y.N. Zhao, F. Li, G.Q. Lu, H.M. Cheng, J. Phys. Chem. B 110 (2006) 20823–20828. [29] T. Ohsaka, J. Phys. Soc. Jpn. 48 (1980) 1661–1668. [30] M.H. Sousa, J.C. Rubim, P.G. Sobrinho, F.A. Tourinho, J. Magn. Magn. Mater. 225 (2001) 67–72. [31] Q.A. Pankhurst, J. Connolly, S.K. Jones, J. Dobson, J. Phys. D: Appl. Phys. 36 (2003) R167–R181. [32] X.H. Wang, J.G. Li, H. Kamiyama, Y. Moriyoshi, T. Ishigaki, J. Phys. Chem. B 110 (2006) 6804–6809. [33] X.W. Zhang, M.H. Zhou, L.C. Lei, Catal. Commun. 7 (2006) 427–431. [34] Y.M. Xu, C.H. Langford, Langmuir 17 (2001) 897–902.