synthesis and characteristics of fe3+-doped sno2 nanoparticles via

Compared with the sol–gel-calcination route, sol–gel-hydrothermal route led to better dispersed nanoparticles with a narrower size distribution and a larger Brunauer–Emmett–Teller BET su

Synthesis and characteristics of Fe 3+ -doped SnO 2 nanoparticles via

sol–gel-calcination or sol–gel-hydrothermal route

L.M Fanga, X.T Zua,b,∗, Z.J Lia, S Zhuc, C.M Liua, W.L Zhoue, L.M Wangc,d

aDepartment of Applied Physics, University of Electronic Science and Technology of China, Chengdu 610054, PR China

bInternational Center for Material Physics, Chinese Academy of Sciences, Shengyang 110015, PR China

cDepartment of Nuclear Engineering and Radiological Sciences, University of Michigan, Ann Arbor, MI 48109-2104, USA

dDepartment of Materials Science and Engineering, University of Michigan, Ann Arbor, MI 48109-2104, USA

eAdvanced Materials Research Institute, University of New Orleans, New Orleans, LA 70148, USA

Received 15 October 2006; received in revised form 3 December 2006; accepted 5 December 2006

Available online 29 December 2006

Abstract

Fe3+-doped SnO2nanoparticles were prepared by sol–gel-calcination and sol–gel-hydrothermal routes, respectively, and their microstructure as well as physical and chemical properties have been characterized and compared Based on XRD, TEM, and Fourier transform infrared (FT-IR) anal-yses, the SnO2crystallites with the tetragonal rutile structure formed directly during a hydrothermal process Compared with the sol–gel-calcination route, sol–gel-hydrothermal route led to better dispersed nanoparticles with a narrower size distribution and a larger Brunauer–Emmett–Teller (BET) surface area Also, the Fe3+-doped SnO2nanoparticles prepared by sol–gel-hydrothermal route had a better thermal stability against agglomeration and crystalline grain size growth than those prepared by the sol–gel-calcination route XRD, EDS, and diffuse reflectance spectra (DRS) analyses proved that the Fe3+and SnO2formed a solid solution in the nanoparticles with both processing routes A significant red shift in the UV absorbing band edge was observed with the increasing Fe3+content

© 2006 Elsevier B.V All rights reserved

Keywords: SnO2; Sol–gel; Hydrothermal; Calcination; Doping; Nanoparticles

1 Introduction

Nano-sized SnO2particles, as an n-type semiconductor with

a wide band gap (Eg= 3.6 eV, at 25◦C), have been attracting

much attention due to their transparency and sensitivity to

reduc-ing gases They have a wide range of applications in gas sensors

[1–3], optoelectronic devices[4–5], dye-base solar cells[6],

sec-ondary lithium batteries electrode materials [7], and catalysts

[8] One area of primary importance is the field of solid state

gas sensors for environmental monitoring, such as for CO, NO,

and C2H5OH, where SnO2has been established as the

predom-inant sensing material It is generally accepted that increasing

the surface/bulk ratio by decreasing the grain size of rutile SnO2

nanoparticles is crucial for achieving high-sensitivity in gas

sen-∗Corresponding author at: Department of Applied Physics, University of

Elec-tronic Science and Technology of China, Chengdu 610054, PR China.

Tel.: +86 28 83201939; fax: +86 28 83201939.

E-mail address:xiaotaozu@yahoo.com (X.T Zu).

sors One of the other most common ways to modify the charac-teristics of the material is introducing dopants into the structure Many results have shown that several additives (cations: Fe, Cu,

Co, Cr, Al, Mn, Mg; anions: P, S) can lead to an increase of the surface area of SnO2-based powders[3,9–14] The added active elements stabilize the SnO2surface, and promote a decrease in grain size It has been found that Fe3+-doped SnO2 nanopar-ticles with lower crystallinity and high surface area have the higher catalytic activity and sensor signal than both pure SnO2 and Fe2O3[15,16] Thus, nanocrystalline Fe3+/SnO2have been investigated as a material for oxygen, ethanol, hydrogen sulfide, carbon monoxide, and water vapor gas sensors[3,9,17] The properties of metal ions (including transition metal ions and rare-earth elements)-doped SnO2 can be altered to differ-ent states due to many factors during the preparation process [18,19] Thus, the final products are related to both composi-tion and the processing method used for synthesize the material

To take advantage of the properties of metal ions-doped SnO2 nanoparticles, a number of attempts have been made by devel-oping effective synthetic techniques in the preparation of metal 0925-8388/$ – see front matter © 2006 Elsevier B.V All rights reserved.

doi: 10.1016/j.jallcom.2006.12.014

ions-doped SnO2 nanoparticles, such as mechanical alloying

[20], sol–gel[3,11,14,18], co-precipitation[13,16], and

chemi-cal vapor deposition[21,22] The sol–gel method is the most

effective method due to its capability in controlling the

tex-tural and surface properties of metal oxides This method is

mainly based on the hydrolysis and polycondensation of a metal

salt, which ultimately yields hydroxide or oxide under

cer-tain conditions[19] Generally, sol–gel-derived precipitates are

amorphous in nature So calcination in air is needed for the

trans-formation from amorphous to rutile phase to complete However,

the calcination process generally lead to serious particle

agglom-eration, grain growth, and thus, a small total surface area, which

decreases the sensitivity to a number of reducing gases The

hydrothermal processing represents an alternative to the

calci-nation for the crystallization of SnO2under mild temperatures

In the hydrothermal treatment, grain size, particle

morphol-ogy, crystallinity, and surface chemistry can be controlled via

processing variables such as sol composition, pH, reaction

tem-perature and pressure, aging time, and nature of solvent and

additive It provides a facilitated route to prepare uniform,

well-dispersed, and well-crystallized SnO2nanoparticles However,

the hydrothermal method has not been used for the preparation of

Fe3+-doped SnO2nanoparticles It is well known that the method

of preparing Fe3+-doped SnO2nanoparticles can determine the

physico-chemical, optoelectronic properties and sensitivity for

gasses, and the effects of Fe3+ doping on SnO2 can be quite

different according to the synthetic procedures, so a

compara-tive study on the properties of Fe3+-doped SnO2nanoparticles

prepared by different routes is necessary

In the present work, Fe3+-doped SnO2 nanoparticles were

prepared by sol–gel calcination and sol–gel-hydrothermal

routes, respectively, and the physico-chemical and

optoelec-tronic properties were compared The results demonstrate

that Fe3+-doped SnO2 nanoparticles prepared by

sol–gel-hydrothermal route are better dispersed and have a narrower size

distribution with larger specific surface area than that prepared

by the normal sol–gel route

2 Experimental

2.1 Synthesis of samples

SnCl4·5H2O and FeCl3·6H2 O were used as tin and iron sources, respectively.

First, citric acid was added to 96 ml of H2O until the pH 1.5 at 50 ◦C with

magnetic stirring; then, 17.529 g of SnCl4·5H2O were added and dissolved.

Secondly, FeCl3·6H2O (the Fe molar content in samples was 0, 1.0, 5.0, 10.0,

15.0, and 20.0 mol%) and 10 ml polyglycol were added to the above solution It

was stirred for 10 min and a sol formed Thirdly, 30 ml NH3 ·H2O (15 mol/l) was

added dropwise to the above sol under magnetic stirring within 30 min Then,

the hydrolysis product was stirred for 30 min to form a gel Finally, the gel was

treated by the following two routes, respectively.

2.1.1 The calcination process

The gel was filtered, and dried at 110 ◦C for 12 h After that, the powders were

further calcined at 350 ◦C in air for 1 h Then, the Fe3+ -doped SnO2 nanoparticles

were obtained The products were denoted as SGC-series.

2.1.2 The hydrothermal process

The gel was transferred into a Teflon-lined autoclave for hydrothermal

reac-tion at 150 ◦C (the pressure in the hydrothermal process should about 0.4 MPa)

for 12 h After that, the hydrothermal product was filtered and dried at 110 ◦C for 10 h Then, the Fe 3+ -doped SnO2 nanoparticles were obtained The products were marked SGH-series.

According to the Fe molar content in samples (0, 1.0, 5.0, 10.0, 15.0, and 20.0 mol%), the products were designed as SGC0, SGC1, SGC5, SGC10, SGC15, and SGC20 for SGC series and SGH0, SGH1, SGH5, SGH10, SGH15, and SGH20 for SGH series, respectively To compare the effect of temperature

on the nanoparticle properties, both the SGC and SGH series of samples were calcined at 350, 600, and 800 ◦C in air for 1 h.

2.2 Characterization of samples

The crystalline phase of Fe 3+ -doped SnO2 particles were determined by X-ray diffraction (XRD, Cu K ␣, 40 kV, 60 mA, Rigaku D/max-2400) The mor-phologies of samples were characterized using a JEOL 2010 F field emission gun electron microscope (TEM) with an accelerating voltage of 200 keV The specific surface area of the prepared samples was calculated from the adsorp-tion isotherm of nitrogen at 77 K on the basis of the Brunauer–Emmett–Teller method (BET; Tristar3000, Micromeritics) The chemical structure information

of the particles was collected by FT-IR spectra (Nicolet 560 Spectrom-eter) Diffuse reflectance spectra (DRS) was recorded by a Shimadzu UV-2101 apparatus, equipped with an integrating sphere, using BaSO4 as reference.

3 Results and discussion

3.1 XRD analyses

The XRD patterns of Fe3+-doped SnO2 nanoparticles are shown in Fig 1a and b, respectively It clearly shows that the rutile phase formed after calcined at 350◦C in the SGC series of samples (seeFig 1a) However, rutile structured SnO2 crystallites have also been obtained by the SGH route with-out calcination (seeFig 1b) The Fe-doping not only makes the degradation of the crystallinity but also makes effects on the size

of the nanoparticles, similar results had been reported[23–25] The crystallite grain size of SnO2 can be estimated according

to Scherrer formula L = K λ/(β cos θ), where λ is the wavelength

of the X-ray radiation (Cu K␣ = 0.15406 nm), K the constant taken as 0.89, β the line width at half maximum height and

θ is the diffracting angle While the crystalline grain sizes of

SnO2of the SGH series decreased from 5.98 to 3.85 nm when the Fe3+ content increased from 0.0 to 20% (seeFig 2), the crystalline grain size of the SGC-series decreased from 17.55

to 7.20 nm (seeFig 2) It is clear that the addition of Fe pre-vented the growth of crystalline grains of SnO2, and the grain size for the SnO2by the sol–gel-calcination route with the same

Fe3+content was larger than that by the sol–gel-hydrothermal route

In addition, no characteristic peaks for the ferric oxide crystal phase were observed in all samples derived from both preparing routes At the same time, the addition of Fe3+significantly influ-enced the crystallinity of SnO2.Fig 1suggested a decrease of crystallinity in the Fe3+-doped SnO2nanoparticles in compar-ison with the undoped sample by the decrease in the intensity

of SnO2peaks This means that a portion of Fe3+formed stable solid solutions with SnO2, it could occupy regular lattice site in SnO2and that may cause the introduction of point defects and change in stoichiometry due to charge imbalance The effect leads to a distortion of the crystal structure of the host

com-Fig 1 XRD patterns of Fe 3+ -doped SnO2 nanoparticles: (a) SGC series and (b)

SGH series.

pound Because of the interference of Fe3+with SnO2lattice, the

crystallinity of the Fe3+-doped SnO2nanoparticles were worse

than that of pure SnO2 But when the Fe3+ content was high

(for example, 15 and 20%), a portion of Fe3+ should exist as

amorphous ferric oxide aggregates The amorphous ferric oxide

Fig 2 The crystalline grain sizes of SnO2 calculated from Scherrer formula.

aggregates also retarded the growth of SnO2nanoparticles This means that a portion of Fe3+formed stable solid solutions with SnO2 In addition, no diffraction peaks of ferric oxide could be detected even after calcination at 800◦C.

The above results suggested that the grain sizes of the Fe3+ -doped SnO2nanoparticles prepared by both methods depended

on the Fe3+content, and Fe3+could prevent the growth of crys-tal grains of SnO2 Moreover, with sol–gel-hydrothermal route, rutile phase was directly synthesized and the addition of Fe3+ was more effective in preventing the growth of crystalline grain size, suggesting that sol–gel-hydrothermal route was a more effective process for preparing Fe3+-doped SnO2nanoparticles compared with the sol–gel-calcination route

It is well known that calcination can improve the crys-tallinity of the samples, but increase the crystallite grain size

of SnO2 at the same time In order to investigate the possible crystallization process of SnO2 by increasing the calcinations temperature, XRD analysis was carried out on the samples cal-cined at 350, 600, and 800◦C, respectively Typical diffraction patterns obtained from samples of pure SnO2 and 10% Fe3+ -doped SnO2 are shown in Figs 3 and 4 The grain sizes of SnO2 calculated by Scherrer formula are showed inTable 1 Compared to pure SnO2, the crystalline grain size of 10% Fe3+

-Fig 3 XRD patterns of samples from sol–gel-calcination route calcined at different temperature: (a) SGC0 and (b) SGC10.

Fig 4 XRD patterns of samples from sol–gel-hydrothermal route calcined at

different temperature: (a) SGH0 and (b) SGH10.

Table 1

The crystalline grain sizes of SnO2 calcined at different temperature

Samples Crystalline grain sizes (nm)

110 ◦C 350◦C 600◦C 800◦C

doped SnO2was smaller On the other hand, for the SGH series,

the addition of Fe3+showed more significant inhibiting effect

on the increase of crystallite size After calcination at 800◦C,

the grain size of SGH10 was 17.71 nm, but it was 45.62 nm

for the SGC10 It is clear that the Fe3+-doped SnO2

nanopar-ticles prepared by sol–gel-hydrothermal process has a better thermal stability against grain growth than that prepared by the calcination process This result suggested that sol–gel-hydrothermal route led to a better atomically mixed structure between Fe3+and SnO2, which retarded agglomeration of grains due to the strong Fe–O–Sn linkage In addition, no diffraction peaks of ferric oxide could be detected even after calcination at

800◦C.

3.2 TEM and BET surface areas

The bright-field TEM and high resolution TEM (HRTEM) images of pure SnO2 nanoparticles and Fe3+-doped SnO2 nanoparticles prepared by SGC or SGH routes are shown in Figs 5 and 6, respectively It could be seen that all SnO2 nanopar-ticles were not larger than 6 nm in diameter for SGH1, SGH5, SGH15, and SGC15 Selected-area diffraction exhibited the polycrystalline rings that were all from the rutile SnO2 struc-ture as indexed inFig 5a These rings are in agreement with the peaks in the power XRD spectra as shown inFig 1 More-over, there are no extra peaks from Fe or ferric oxides Thus, the

Fe3+ions are believed to be dissolved in the SnO2structure The HRTEM image showed the crystallinity of the nanoparticles are well-developed The Fe3+-doped SnO2 nanoparticles prepared

by sol–gel-hydrothermal route were uniform in size with narrow size distribution, and were well dispersed The average grain size decreased when the Fe3+content increased It decreased from 6.0

to 3.0 nm when the Fe3+content increased from 1 to 15% This result is similar to that obtained from the XRD analysis But for the SGC15 nanoparticles derived from sol–gel-calcination route, it could be seen that the average grain size was not uni-form, range from 2 to 6 nm And unfavorable agglomeration occurred (seeFig 6)

The BET surface areas of the samples doped with differ-ent Fe3+ content are listed in Table 2 It could be seen that the surface areas were strongly dependent on the preparation route and the Fe3+ content It increased with the increasing

of Fe3+ content, and it was clear that the BET surface areas

of samples obtained from sol–gel-hydrothermal method were remarkably higher than that from sol–gel-calcination route That

is because the samples prepared by sol–gel-hydrothermal route were well-dispersed and less size with narrow distribution, but the samples prepared by sol–gel-calcination route were agglom-erated and larger in grain size This result indicated that the addition of Fe3+into SnO2matrix suppressed the reduction of surface area and the suppression was more effective with higher

Fe3+content Compared with the sol–gel-calcination route, the sol–gel-hydrothermal route was a more effective process to pro-duce the SnO2with large surface areas

Table 2

The BET surface area of samples with different Fe 3+ content

Fig 5 Bright field TEM and HRTEM micrographs of samples: (a) SGH1, (b) SGH5, and (c and d) SGH15.

SnO2nanoparticles owned much greater surface-to-bulk ratio

than that of bulk materials A large fraction of the atoms

pre-sented at the surface of SnO2nanoparticles An important effect

is associated with the depth of the surface space charge region

that is affected by gas adsorption in relation to the particle size

When the particle size gets into the nano-range, the depletion

layer takes over the bulk, and it becomes difficult to distinguish

surface from bulk conduction So the narrow size distribution,

excellent dispersibility and large surface areas are all favorable

for application in gas sensors and optoelectronic devices The

characteristics of Fe3+-doped SnO2 nanoparticles prepared by

sol–gel-hydrothermal route make the processing route

particu-larly appealing

3.3 Fourier transform infrared (FT-IR) spectroscopy analyses

The FT-IR transmission spectra of samples are given inFig 7 The absorption peaks at 3400 and 1640 cm−1were attributed to vibration of hydroxyl due to the fact that SnO2 retained cer-tain adsorbed water Additionally, bands at 2930, 2850, and

1390 cm−1 were assigned to C–H vibrations The C–H could

be attributed to the organic trace residuals, which remained in nanoparticles even after calcinations for the sol–gel-calcination route The band at approximately 2365 cm−1resulted from the adsorption and interaction of atmospheric carbon dioxide with water according to literature [26] The bands observed in the

Fig 6 HRTEM micrographs of SGC15.

Fig 7 FT-IR transmission spectra of samples: (a) SGC series and (b) SGH

series.

range of 1059 cm−1were assigned to the vibration of different

types of surface hydroxyl groups Lenaerts[27]had observed

that the bands at 1068 and 970 cm−1 most strongly appeared

in the spectra of SnO2taken in O2at room temperature versus

450◦C spectra They ascribed these absorptions to Sn O and

Sn O modes of surface cation–oxygen bonds The Sn–O–Sn

vibration appeared in the range of 400–700 cm−1as the result

of condensation reaction The peaks at 660 and 560 cm−1 are

attributed to the Sn–O–Sn antisymmetric vibrations Compared

with the SGC series samples, the absorption peaks of 3400 cm−1

were higher for the SGH series samples, suggesting that there

were more O H bonds on the Fe3+-doped SnO2 nanoparticle

prepared by sol–gel-hydrothermal route than that prepared by

sol–gel-calcination route

3.4 UV–vis diffuse reflectance spectroscopy (DRS)

analyses

The diffuse reflectance spectra of samples with different Fe3+

content are depicted inFig 8 The spectra of Fe3+-doped SnO2

nanoparticles displayed a red shift in the band gap transition

with the increasing dopant content It was reported that similar

shift was resulted from the incorporation of Fe3+into the SnO

Fig 8 The diffuse reflectance spectra of samples: (a) SGC series and (b) SGH series.

nanoparticles[28–30] Red shifts of this type could be attributed

to the charge-transfer transitions between the Fe3+d-electrons and the SnO2 conduction or valence band On the other hand, there was no separated phase of ferric oxide because its broad band centered at 500 nm was not present[31] This is another evidence showing the Fe3+were doped in the SnO2nanoparticles rather than existed as ferric oxide

After comparing the DRS spectra, it can be seen that the SGH-series had larger band gap than SGC-series The larger band gap was resulted from the well-known quantum size effect

of semiconductors[32] The results of XRD and HRTEM have proved that with same Fe3+content, the crystalline grain size of SnO2from sol–gel-hydrothermal route was more uniform and smaller than that prepared with the sol–gel-calcination route According to the size quantization, an increase in the band gap happened with the decrease in particle dimensions[33]

4 Conclusion

Fe3+-doped SnO2 nanoparticles were prepared by sol–gel-calcination and sol–gel-hydrothermal routes, respectively The SnO2crystallites with the tetragonal rutile structure could form directly during the hydrothermal process without calcination

Fe3+formed stable solid solutions in SnO2nanoparticles

Com-pared with sol–gel-calcination route, sol–gel-hydrothermal

route led to better dispersed spherical Fe3+-doped SnO2

nanoparticles with narrower size distribution and larger specific

surface area The composite nanoparticles prepared by the

sol–gel-hydrothermal route have better thermal stability against

agglomeration and particle growth than those prepared by

sol–gel-calcination route The BET surface area of the SnO2

sample with 15% Fe3+prepared by sol–gel-hydrothermal route

was increased to 259.8 m2/g from 206.1 m2/g of SGH0 Narrow

size distribution nanoparticles, excellent dispersibility and large

surface areas of Fe3+-doped SnO2 nanoparticle prepared by

sol–gel-hydrothermal route make it particularly appealing in

applications for gas sensors and optoelectronic devices

Acknowledgements

This study was supported financially by the NSAF Joint

Foundation of China (10376006) and by the Program for New

Century Excellent Talents in University (NCET-04-0899) and by

the Ph.D Founding Support Program of Education Ministry of

China (20050614013) and by Program for Innovative Research

Team in UESTC

References

[1] A Teeramongkonrasmee, M Sriyudthsak, Sens Actuators B 66 (2000)

256.

[2] G Ansari, D Boroojerdian, S.R Sainker, R.N Karekar, R.C Aiyer, S.K.

Kulkarni, Thin Solid Films 295 (1997) 271.

[3] D Kotsikau, M Ivanovskaya, D Orlik, M Falasconi, Sens Actuators B

101 (2004) 199.

[4] F Gu, S.F Wang, M.K Lu, Y.X Qi, G.J Zhou, D Xu, D.R Yuan, Opt.

Mater 25 (2004) 59.

[5] Tomokatsu Hayakawa, Masayuki Nogami, Sci Technol Adv Mater 6

(2005) 66.

[6] S Ferrere, A Zaban, B.A Gsegg, J Phys Chem B 101 (1997) 4490.

[7] V Subramanian, K.I Gnanasekar, B Rambabu, Solid State Ionics 175

(2004) 181.

[8] S.R Stampfl, Y Chen, J.A Dumesis, Ch Niu, C.G Hill, J Catal 105

(1987) 445.

[9] O.K Tana, W Caoa, W Zhua, J.W Chaib, J.S Pan, Sens Actuators B 93

(2003) 396.

[10] Shunji Abea, U.Sung Choib, Kengo Shimanoeb, Noboru Yamazoeb, Sens Actuators B 107 (2005) 516.

[11] F Gu, S.F Wang, M.K Lu, Y.X Qi, G.J Zhou, D Xu, D.R Yuan, Inorg Chem Commun 6 (2003) 882.

[12] H.Y Jin, Y.H Xu, G.S Pang, W.J Dong, Q Wan, Y Sun, S.H Feng, Mater Chem Phys 85 (2004) 58.

[13] Y.D Wang, X.H Wu, Q Sun, Y.F Li, Z.L Zhou, Solid State Electron 45 (2001) 3470.

[14] Lincoln A Kurihara, Sergio T Fujiwara, Ren´ı V.S Alfaya, Yoshitaka Gushikem, Antonio A.S Alfaya, Sandra C de Castro, J Colloid Interface Sci 274 (2004) 579.

[15] Ricardo H.R Castro, Pilar Hidalgo, R Muccillo, Douglas Gouvea, Appl Surf Sci 214 (2003) 172.

[16] M.N Rumyantseva, V.V Kovalenko, A.M Gaskov, T Pagnier, D Machon,

J Arbiol, J.R Morante, Sens Actuators B 109 (2005) 64.

[17] O Tan, W Zhu, Q Yan, Sens Actuators B: Chem 65 (2000) 361 [18] Mira Risti, Mile Ivanda, Stanko Popovi, Svetozar Musi, J Non-Cryst Solids 303 (2002) 270–280.

[19] A Chandra Bose, P Thangadurai, S Ramasamy, Mater Chem Phys 95 (2006) 72.

[20] A.S Albuquerque, J.D Ardisson, W.A.A Macedo, T.S Plivelic, I.L Tor-rianib, J Larrea, E.B Saitovitch, J Magn Magn Mater 272–276 (2004) 2211.

[21] M.Y Huh, S.H Kim, J.P Ahn, J.K Park, B.K Kim, Nanostruct Mater 11 (1999) 211.

[22] J.X Wang, H.Y Chen, Y Gao, D.F Liu, L Song, Z.X Zhang, X.W Zhao, X.Y Dou, S.D Luo, W.Y Zhou, G Wang, S.S Xie, J Cryst Growth 284 (2005) 73.

[23] Hua Yanga, Weichun Jina, Li Wang, Mater Lett 57 (2003) 3686– 3689.

[24] X.Y Fu, H.W Zhang, S.Y Niu, Q Xin, J Solid State Chem 178 (2005) 603–607.

[25] Masanori Hirano, Keisuke Ota, Hiroyuki Iwata, Chem Mater 16 (2004) 3725–3732.

[26] S Emiroglu, N Barsan, U Weimar, V Hoffmann, Thin Solid Films 391 (2001) 176.

[27] S Lenaerts, J Roggen, G Maes, Spectrochim Acta 51A (5) (1995) 883.

[28] W Choi, A Termin, M.R Hoffmann, J Phys Chem 98 (1994) 13669 [29] C Wang, D.W Bahnemann, J.K Dohrmann, Chem Commun (2000) 1539.

[30] Z.J Li, B Hou, Y Xu, D Wu, Y.H Sun, J Mater Sci 40 (2005) 3939 [31] J.A Navio, G Colon, M.I Litter, G.N Bianco, J Mol Catal 106 (1996) 267.

[32] M Anpo, T Shima, S Kodama, Y Kobokawa, J Phys Chem 91 (1987) 4305.

[33] K.M Reddy, C.V.G Reddy, S.V Manorama, J Solid State Chem 158 (2001) 180–186.