Fabrication and characterization of nano TiO2 thin films at low temperature

Article abstract of DOI:10.1016/j.materresbull.2006.02.033

Anatase TiO₂ thin films were successfully prepared on glass slide substrates via a sol-gel method from refluxed sol (RS) containing anatase TiO₂ crystals at low temperature of 100 °C. The influences of various refluxing time on crystallinity, morphology and size of the RS sol and dried TiO₂ films particles were discussed. These samples were characterized by infrared absorption spectroscopy (FT-IR), X-ray diffraction (XRD), transmission electron microscopy (TEM), field emission-scanning electron microscopy (FE-SEM) and UV-vis absorption spectroscopy (UV-vis). The photocatalytic activities of the TiO₂ thin films were assessed by the degradation of methyl orange in aqueous solution. The results indicated that titania films thus obtained were transparent and their maximal light transmittance exceeded 80% under visible light region. The TiO₂ thin films prepared from RS-6 sol showed the highest photocatalytic activity, when the calcination temperature is higher than 300 °C. The degradation of methyl orange of RS-6 thin films reached 99% after irradiated for 120 min, the results suggested that the TiO₂ thin films prepared from RS sol exhibited high photoactivities.

Full text of DOI:10.1016/j.materresbull.2006.02.033

Materials Research Bulletin 41 (2006) 1596–1603
www.elsevier.com/locate/matresbu
Fabrication and characterization of nano TiO2 thin
films at low temperature
*
Lei Ge , Mingxia Xu, Ming Sun, Haibo Fang
School of Materials Science and Engineering, Key Laboratory for Advanced Ceramics and Machining
Technology of Ministry of Education, Tianjin University, Tianjin 300072, PR China
Received 29 August 2005; received in revised form 16 January 2006; accepted 21 February 2006
Available online 20 March 2006
Abstract
Anatase TiO2 thin films were successfully prepared on glass slide substrates via a sol–gel method from refluxed sol (RS)
containing anatase TiO2 crystals at low temperature of 100 8C. The influences of various refluxing time on crystallinity, morphology
and size of the RS sol and dried TiO2 films particles were discussed. These samples were characterized by infrared absorption
spectroscopy (FT-IR), X-ray diffraction (XRD), transmission electron microscopy (TEM), field emission-scanning electron
microscopy (FE-SEM) and UV–vis absorption spectroscopy (UV–vis). The photocatalytic activities of the TiO2 thin films were
assessed by the degradation of methyl orange in aqueous solution. The results indicated that titania films thus obtained were
transparent and their maximal light transmittance exceeded 80% under visible light region. The TiO2 thin films prepared from RS-6
sol showed the highest photocatalytic activity, when the calcination temperature is higher than 300 8C. The degradation of methyl
orange of RS-6 thin films reached 99% after irradiated for 120 min, the results suggested that the TiO2 thin films prepared from RS
sol exhibited high photoactivities.
# 2006 Elsevier Ltd. All rights reserved.
Keywords: A. Oxides; A. Thin films; B. Sol–gel chemistry; C. X-ray diffraction; D. Catalytic properties
1. Introduction
Nanocrystalline TiO2, one of the most popular photocatalyst, have long been investigated in environmental
purification, decomposition of dyes in wastewater [1,2]. Many researchers have focused on the preparation and
characteristics of titanium dioxide nanoparticles as well as their applications [3]. The immobilization of TiO2 in the
form of a thin film provides an advantage over the drawbacks encountered with powder suspensions such as difficult
separation of TiO2 particles from the suspension and easy aggregation of suspended TiO2 particles. Good
photocatalytic activities of the TiO2 films necessitate efficient photo-induced electron–hole pair generation and
efficient charge separation, which requires in turn preparations of well-crystallized TiO2 films, preferably in the
anatase crystalline form [4]. To prepare titanium dioxide thin film, many papers have been published on the
preparation of titania films. Physical methods include sputtering [5], CVD [6], MOCVD [7], ion-assisted electronbeam
evaporation [8], etc. These methods require a high amount of energy that is reemitted in part as exhaust heat or in the
* Corresponding author. Tel.: +86 22 27890489; fax: +86 22 27404724.
E-mail addresses: gelei08@eyou.com (L. Ge), xumingxia@tju.edu.cn (M. Xu).
0025-5408/$ – see front matter # 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2006.02.033

L. Ge et al. / Materials Research Bulletin 41 (2006) 1596–1603
1597
exhaust gases, and an annealing post-treatment, usually at 450 8C, is necessary to remove the organics used in the
preparation of the colloidal suspension or paste. Recently, efforts have been made to produce TiO2 thin films through
chemical routes, which are less energy consuming and do not require equipment. Among these, the sol–gel method
using the peroxotitanic acid solution (PTA) is of great interest because it has neutral pH and low material cost. The
study about the formation of TiO2 film using the PTA has been recently performed by Ichinose et al. [9,10]. Gao et al.
[11] have reported the synthesis of anatase TiO2 thin films using the aqueous peroxotitanate solution obtained by
dissolving metatitanic acid (H2TiO3) in a mixture of H2O2 and ammonia in water. Lee et al. [12] prepared PTA sol
using TiCl3 as raw materials. However, the PTA sol contains no anatase crystals, the TiO2 film on substrate should be
calcined at high temperature to obtain anatase structure after the sol was coated on the substrates. It is evident the
interest of low-temperature methods for fabrication of crystalline TiO2 thin films not only from the point of view of
energy saving, as annealing post-treatment is avoided, but also as they allow the use of low thermally resistant
materials such as plastics, wood or paper as substrates. Furthermore, to make the whole chemical approach
environmental friendly, the use of aqueous solutions instead of organic solvents is desired. Recently, our group has
accomplished the deposition of anatase TiO2 films with relative ease and at low temperature using hydrothermal
processing method in which crystallization is enhanced by autoclave heating [13,14].
In this paper we show a new method in which anatase TiO2 thin films are deposited on glass slides substrates by dip-
coating method. The refluxed sol (RS) containing needle-like anatase crystals was synthesis by refluxing the PTA sol at
relative low temperature and pressure condition, and the TiO2 thin films were formed on glass slides substrates after
dried in air or heated at low temperature.
2. Experimental
For the preparation of the anatase TiO2 sol, the following materials were used: titanyl sulfate (TiOSO4ꢀnH2O,
23.2% TiO2); hydrogen peroxide (H2O2, 30%); ammonia solution (NH3ꢀH2O, 3 mol/l); methyl orange.
The anatase TiO2 sol (RS) was prepared in the following way. Titanyl sulfate (17.3 g) was slowly added to the
distilled water at room temperature, then precipitated by adding ammonia solution (NH4OH, 3 mol/l) resulting the
formation of white precipitate [Ti(OH)4]. The white precipitate was filtered and sufficiently washed with distilled
2ꢁ
water to remove the NH4+ and SO4 formed in the reaction, and then added 450 ml distilled water to disperse the
precipitate homogeneously. The precipitate was peptized in 50 ml aqueous hydrogen (30%), continual magnetic
stirring was required to avoid the immediate dense gel formation during dissolution and to keep the reactant mixed
uniformly. The obtained transparent sol was kept under reflux condition (around 100 8C) for 2 h. Finally, the pure RS
sol with anatase TiO2 crystals was obtained, which was used to prepare thin films and equivalent powders after dried in
air oven. This sample was denoted as sample RS-2. Samples RS-6 and RS-10 were prepared by the same procedure,
but the refluxing time was 6 and 10 h, respectively.
Glass slides were used as substrates. Before the deposition, substrates were ultrasonically cleaned in acetone and
absolute ethanol for 5 min, respectively. Finally, they were thoroughly rinsed with water. TiO2 thin films were deposited
on substrates by a dip-coating process at room temperature with the withdrawing speed about 5–6 cm/min. Substrates
wereimmersedintothe TiO2 solfor10 min. Uponwithdrawingfromthesol, thesubstratesweredriedunderinfraredlight
for30 min, thedryingtemperaturewasabout100 8C, the distancebetweentheIRlightandsamplesofTiO2 thinfilmswas
20 cm. TiO2 layers on substrates could be thickened by means of consecutive dip-coating processes.
The FT-IR spectra of the dried gel were measured by the KBr pellet method (BIO-RAD FTS-3000). The crystalline
structure of the RS particles dried in air and TiO2 thin films prepared from RS sols were determined by X-ray
diffractometer (Rigaka D/max 2500 v/pc, Japan) using graphite monochromatic copper radiation (Cu Ka) at 40 kV,
30 mA over the 2u range 20–808. The morphology and size of the RS sol crystals were examined by transmission
electron microscopy (JEM-1200 EX II). The adhesion of the dried TiO2 films was evaluated by ultrasonic washing and
weight measuring. The surface morphologies of dried TiO2 thin films were observed by a field emission-scanning
electron microscope (FE-SEM JEDL JSM-6700F) Spectroscopic analysis of the TiO2 films was performed using UV–
vis spectrophotometer (7230G, Shanghai, China) with wave range of 300–800 nm.
The photocatalytic activities of the prepared TiO2 thin films were evaluated from an analysis of the degradation of
methyl orange in an aqueous solution under UV illumination. The main wavelength of the UV lamp was 254 nm and
the power of the UV lamp was 25 W. Two pieces of 25 mm ꢂ 75 mm glass slides coated with TiO2 thin films were
settled in a watch glass containing 80 ml methyl orange aqueous solution (10 mg/l) and exposed to UV-irradiation. The

1598
L. Ge et al. / Materials Research Bulletin 41 (2006) 1596–1603
distance from liquid level to lamp was 20 cm. The pH value of photocatalytic solution was adjusted to 3 by nitric acid.
UV–vis absorption spectra of irradiated samples were measured every 20 min with a UV–vis spectrophotometer
(7230G, Shanghai, China). The changes in concentrations of methyl orange were estimated from the changes in
absorbance of the absorption maximum at 500 nm. The structure of methyl orange is shown below:
3. Results and discussion
Fig. 1 shows the FT-IR spectra of the RS sols refluxing at 100 8C for different hours. The absorption from 3000 to
3600 cmꢁ1 can be assigned to the stretching vibration of the hydrogen-bonded OH groups of the adsorbed water and
the titanium hydroxide obtained by hydrolysis of titanyl sulfate. The absorption around 1630 cmꢁ1 can be assigned to
the bending vibration of adsorbed H2O molecules. The peaks around 1416 cmꢁ1 are due to the stretching vibration of
the N–H bonds in NH4 that remained. The absorption at 900 cmꢁ1 in Fig. 1a can be assigned to the stretching
+
vibrations of the O–O bond in the Ti–O–O–H bond of the peroxotitanic acid. It is known that Ti4+ ions and H2O2
interact strongly and that a Ti-peroxy compound is formed from their interaction. The above results indicate that the
gel consisting of Ti-peroxy compounds was obtained by heating at refluxing at 100 8C. On the other hand, the
absorption peak due to the peroxo group obviously decreased for RS-6 and disappeared completely for RS-10. It is
concluded that the peroxo group was decomposed by reflux treatment at 100 8C for 6 and 10 h.
Fig. 2 shows the XRD patterns of TiO2 crystals from the RS sols synthesized at different time. The XRD patterns of
the three samples give five distinctive TiO2 peaks at 25.38, 37.98, 48.08, 54.68 and 62.88, corresponding to anatase
(1 0 1), (1 0 3, 0 0 4 and 1 1 2), (2 0 0), (1 0 5 and 2 1 1) and (2 0 4) crystal planes (JCPDS 21-1272), respectively.
These results indicate that all samples prepared by the refluxing method are pure anatase, the most photocatalytically
active form of TiO2. However, we prepared anatase nanocrystalline particles at 100 8C without further treatments in
the present work, in which titanyl sulfate was used as a starting material. The intensities of the XRD peaks increased
and the widths became narrower as shown in Fig. 2, compared with Fig. 2a and c. The finding indicates that the
crystallinity of the obtained anatase crystals increased as the refluxing time extended from 2 to 10 h. In general, the full
width at half maximum of XRD peak corresponds to the crystal size of porous materials. When the width was broader,
the crystallites exhibited smaller size. The Scherrer equation was used to determine the average crystal sizes of the
Fig. 1. FT-IR spectra of the RS gel prepared at 100 8C. The heating times represented are: (a) 2 h, (b) 6 h and (c) 10 h.

L. Ge et al. / Materials Research Bulletin 41 (2006) 1596–1603
1599
Fig. 2. XRD patterns of RS sol sample prepared at different refluxing time: (a) 2 h, (b) 6 h and (c) 10 h.
TiO2 nanoparticles. The average sizes of the crystallites for samples (a)–(c) were determined to be 5, 30 and 40 nm,
respectively.
Morphology and size of anatase TiO2 particles after reflux treatment were examined by transmission electron
microscopy (TEM). Fig. 3 illustrates the size and morphological development in the RS sol particles prepared at
different refluxing time. It can be seen from Fig. 3a that needle-like particles with a particle size distribution of about
10–15 nm length and 5–10 nm width, which is in agreement with the average crystal size determined by XRD, were
uniformly dispersed in the whole aqueous solution when the refluxing time was 2 h. As the refluxing time increased to
Fig. 3. TEM micrographs of anatase TiO2 sol (RS) samples prepared at different refluxing time: (a) 2 h, (b) 6 h and (c) 10 h.

1600
L. Ge et al. / Materials Research Bulletin 41 (2006) 1596–1603
Fig. 4. XRD patterns of dried TiO2 thin films prepared from RS sols: (a) glass substrates, (b) RS-2 films, (c) RS-6 films and (d) RS-10 films.
6 h, a few irregular and larger anatase TiO2 crystals of about 25–30 nm in length started to form in comparison with
Fig. 3a. As shown in Fig. 3c, some TiO2 particles of about 35–40 nm in length appeared when the refluxing time
increased to 10 h. The fact suggests that the extending refluxing time favors to the formation of larger and longer
needle-like anatase TiO2 particles.
To determine the crystallinity of the TiO2 films, the XRD patterns of the dried TiO2 thin films prepared from
different RS sols are shown in Fig. 4. For comparison, the pattern of glass substrate is given as well. It has been found
that the dried TiO2 thin films prepared from RS sols shows peaks at 2u = 25.38 and 37.98, respectively, matching with
peaks reported for anatase phase given by JCPDS card 21-1272. Hence, from the results we can conclude that the TiO2
films studied here are in the form of anatase. No peaks were found on RS-2 TiO2 films, this because the low
crystallinity of RS-2 sol and small amount of TiO2 on surface of substrates. The intensities of the XRD peaks increased
and the widths became narrower from RS-2 to RS-10 films, indicating the crystallinity of the anatase TiO2 films
increased as the refluxing time extended from 2 to 10 h.
The adhesions of the dried TiO2 films were evaluated by ultrasonic washing for 3 min and weight measuring. The
results indicated that adhesions of the dried TiO2 films were good, no powders or fragments break off from the dried
films during ultrasonic treatment, the weight of the TiO2 films did not changed after ultrasonic process, which
indicated that the adhesion of the film was good enough and can be used in the actual applications.
The SEM micrographs of the dried TiO2 thin films prepared from RS sols are shown in Fig. 5, and all films
fabricated were uniform. Almost no TiO2 particles can be found on the surface of RS-2 films (Fig. 5a), indicated that
the most of TiO2 on the surface of RS-2 films were amorphous and little anatase crystals were formed after dried under
IR light according to XRD and SEM analysis. In Fig. 5b, we can clearly see the round anatase crystals on the surface of
the RS-6 films. The as-dried RS-6 TiO2 film is composed of sphere particles of about 30–40 nm in size. However, the
particle morphologies of RS sols and TiO2 thin films were different. In our experiment, the films were dried under
infrared light, the temperature in drying process was about 100 8C, and then the length of needle-like anatase crystals
decreased in the drying treatment, the needle-like crystals agglomerated together to form secondary sphere particles,
thus the RS sols and TiO2 thin films have different particle morphologies. The size of surface particles grew with
increasing refluxing time. In Fig. 5c, as the refluxing time increased to 10 h, TiO2 particles agglomerated after infrared
irradiation and the size of RS-10 TiO2 films reached about 50–70 nm, indicated that the refluxing time have important
influence on size of TiO2 particles, and the size of nano TiO2 particles increased with extending refluxing time. The
thickness of the TiO2 films was evaluated from SEM fracture cross-section (FCS showed in Fig. 5d), it was about
200 nm in film thickness after four dip-coating cycles.
The properties of nanosized semiconductor particles have long been known to depend very sensitively on the
particle size. As the diameter of crystallite approaches the exciton Bohr diameter, a splitting of the energy bands into
discrete quantized energy level occurs. This is the well-known quantum size effect. The UV–vis absorption band edge
is a strong function of TiO2 cluster size for diameters less than 10 nm, which can be attributed to the quantum size

L. Ge et al. / Materials Research Bulletin 41 (2006) 1596–1603
1601
Fig. 5. FE-SEM micrographs of dried TiO2 thin films prepared from different RS sols: (a) RS-2, (b) RS-6, (c) RS-10 and (d) fracture cross-section
(FCS) of dried RS-6 thin films.
effect. With the decrease in particle size, the blue shift of the absorption edge is observed in the absorption spectra. The
transparent TiO2 thin films were used for the UV–vis absorption studies. Fig. 6 gives the UV–vis spectra of the TiO2
thin film samples prepared from different RS sols and heated at 300 8C. The UV–vis absorption spectra show that
absorption edge shifts to shorter wavelength with decreasing refluxing time, which shows that the sizes of TiO2
particles increase with extending refluxing time in the investigated scale. The conclusion accords with the result from
SEM analysis.
Fig. 6. UV–vis absorption spectra of TiO2 thin films prepared from different RS sols: (a) RS-2, (b) RS-6 and (c) RS-10.

1602
L. Ge et al. / Materials Research Bulletin 41 (2006) 1596–1603
Fig. 7. Plot of optical variation of a1/2 vs. hn of heated TiO2 thin films: (a) RS-2, (b) RS-6 and (c) RS-10.
The UV–vis spectra were processed in order to obtain the energy band gap Eg, from the expression corresponding to
indirect gap semiconductors:
1=2
ðahnÞ ¼ Aiðhn ꢁ EgÞ
Here a is the absorption coefficient (cmꢁ1) and hn (eV) the energy of excitation and Ai is a constant which does not
depend on photoenergy. To give correct estimation of the absorption coefficients, we heated the three film samples at
300 8C to transfer the amorphous TiO2 into anatase phase and remove the remained RS sols in the thin films. A
quantitative evaluation of the band gap energy can be performed by plotting ðahnÞ1=2 against hn and extrapolating the
absorption edge to zero as shown in Fig. 7. The thickness of the TiO2 films was 200 nm according to SEM analysis
(Fig. 5d). From these plots, the values of Eg of 3.28 eV for the TiO2 thin films prepared from RS-2 sol, 3.22 eV for TiO2
thin films from RS-6 sol and 3.19 eV for RS-10 sol are obtained. It was found that the bandgap energies of the films
decreased and red shift in the absorption edge appeared with increasing reflux time. The size of nanoparticles on
surface of TiO2 thin films increased with extending refluxing time according to the SEM analysis, the larger anatase
crystals has lower energy band gap caused by the quantum size effects. Then, the energy band gap decreased with
refluxing time.
The photocatalytic activity of the films toward the degradation of methyl orange in aqueous solutions was
investigated. Fig. 8 shows the percentage of methyl orange degraded as a function of irradiation time for TiO2 thin
Fig. 8. Photodegradation of methyl orange over TiO2 thin films as a function of UV light irradiation time.

L. Ge et al. / Materials Research Bulletin 41 (2006) 1596–1603
1603
films prepared from different RS sols. It is obvious that the TiO2 thin films prepared from RS-6 sol has the best
photocatalytic activity; the degradation of methyl orange reached 99% after irradiated after 120 min. It is worth noting
that the refluxing time is very important to photoactivity of TiO2 thin films. The correlation between photocatalytic
activity and refluxing time is well consistent with the effect of refluxing time on the crystals sizes of TiO2 thin films
according the SEM analysis (Fig. 5). It is believed the RS-6 thin films should have excellent photocatalytic activity due
to high specific surface areas and relative small crystals sizes compared with RS-2 and RS-10 thin films. The RS-2 sol
contains little anatase TiO2 crystals and most of TiO2 are amorphous according to the SEM and XRD analysis. Then,
the TiO2 films prepared from RS-2 sol have lower photoactivity than RS-6 films. The diameters anatase crystals in RS-
10 sol increased with extending refluxing time. When the crystals size becomes lager and reached 50–70 nm (RS-10
films, Fig. 5c) in diameters compared with RS-6 films, the average path length of a charge carrier to the surface is
longer, which has negative effect on the photoactivity. The TiO2 thin films prepared from RS-6 sol had better
crystallinity than RS-2 films and smaller crystal size than RS-10 films; therefore, the RS-6 films had the highest
photoactivity. The degradation of methyl orange of RS-6 thin films reached 99% after irradiated for 120 min, the
results suggested that the TiO2 thin films prepared from RS sols showed high photoactivities.
4. Conclusion
At low temperature, the crystalline TiO2 thin films were deposited on glass slides by a dip-coating method from the
RS sol containing anatase crystals, which was synthesized by peptizing the hydrolysis of titanyl sulfate in abundant
hydrogen peroxide solution and subsequent refluxing at 100 8C for different hours. Refluxing time has a significant
effect on the formation of anatase TiO2 crystallites. In this study, better crystallinity was achieved in refluxing time
range of 2–10 h. The dried TiO2 thin films contained round anatase crystals, the size of TiO2 crystals increased with
refluxing time. The TiO2 thin films, which were deposited from the RS-6 sol, exhibited a high photocatalytic activity in
degradation of methyl orange in aqueous solution. The preparation process of photocatalytic thin films was quite
simple and a low-temperature route. Therefore, this study is essential for its implementation across a wide range of
applications, involving the deposition of photocatalytic TiO2 films on low thermally resistant materials, such as
plastics.
Acknowledgment
This work was financially supported by the Doctor Foundation of China (Grant No. 20030056001).
References
[1] D.W. Thomapson, C.A. Kelli, F. Farzad, G. Mryer, Langmuir 15 (1999) 650–655.
[2] E.L. Crepaldi, G.J.A.A. Soler-Illia, D. Crosso, F. Cagnol, F. Ribot, A. Sanchez, J. Am. Chem. Soc. 125 (2003) 9770–9778.
[3] G. Subramania, K. Constant, R. Biswas, M.M. Sigalas, K.M. Ho, J. Am. Ceram. Soc. 85 (1) (2002) 383–1387.
[4] K. Shimizu, H. Imai, H. Hirashima, K. Tsukuma, Thin Solid Films 351 (1999) 220–226.
[5] D. Mardare, M. Tasca, M. Delibas, G.I. Rusu, Appl. Surf. Sci. 156 (2000) 200–208.
[6] W.G. Lee, S.I. Woo, J.C. Kim, S.H. Choi, K.H. Oh, Thin Solid Films 237 (1994) 4459–4565.
[7] E.K. Kim, M.H. Son, S.K. Min, Y.K. Han, C.H. Wang, S.S. Yom, J. Appl. Phys. 79 (1996) 4459–4566.
[8] D. Bhattacharyya, N.K. Sahoo, S. Thakur, N.C. Das, Thin Solid Films 360 (2000) 96–101.
[9] H. Ichinose, A. Terasaki, H. Katsuki, J. Ceram. Soc. Jpn. 104 (1996) 715–718.
[10] H. Ichinose, A. Kawahara, H. Katsuki, J. Ceram. Soc. Jpn. 104 (1996) 914–918.
[11] Y. Gao, Y. Masuda, Z. Peng, T. Yonezawa, K. Koumoto, J. Mater. Chem. 13 (2003) 608–612.
[12] C.K. Lee, D.K. Kim, J.H. Lee, J. Sol–Gel Sci. Technol. 31 (2004) 67–71.
[13] L. Ge, M.X. Xu, E. Lei, Y.M. Tian, H.B. Fang, K. Eng. Mater. 280–283 (2005) 809–813.
[14] L. Ge, M.X. Xu, E. Lei, H.B. Fang, Chin. J. Inorg. Chem. 21 (2005) 394–398.