DOI 10.1007/s00339-011-6749-6
Phase transformation of nanostructured titanium dioxide thin
films grown by sol–gel method
Sava¸s Sönmezo˘glu· Güven Çankaya · Necmi Serin
Received: 21 June 2011 / Accepted: 16 December 2011 / Published online: 7 January 2012 © Springer-Verlag 2012
Abstract Nanostructured TiO2 thin films were deposited on quartz glass at room temperature by sol–gel dip coat-ing method. The effects of annealcoat-ing temperature between 200◦C to 1100◦C were investigated on the structural, mor-phological, and optical properties of these films. The X-ray diffraction results showed that nanostructured TiO2thin film annealed at between 200◦C to 600◦C was amorphous trans-formed into the anatase phase at 700◦C, and further into ru-tile phase at 1000◦C. The crystallite size of TiO2thin films was increased with increasing annealing temperature. From atomic force microscopy images it was confirmed that the microstructure of annealed thin films changed from column to nubbly. Besides, surface roughness of the thin films in-creases from 1.82 to 5.20 nm, and at the same time, average grain size as well grows up from about 39 to 313 nm with increase of the annealing temperature. The transmittance of the thin films annealed at 1000 and 1100◦C was reduced sig-nificantly in the wavelength range of about 300–700 nm due to the change of crystallite phase. Refractive index and opti-cal high dielectric constant of the n-TiO2thin films were in-creased with increasing annealing temperature, and the film thickness and the optical band gap of nanostructured TiO2 thin films were decreased.
S. Sönmezo˘glu (
)Faculty of Kamil Ozdag Science, Department of Physics, Karamano˘glu Mehmetbey University, 70100, Karaman, Turkey e-mail:svssonmezoglu@kmu.edu.tr
Fax: +90-338-2262116 G. Çankaya
Faculty of Engineering and Natural Sciences, Materials Engineering, Yıldırım Beyazıt University, Ankara, Turkey N. Serin
Faculty of Engineering, Department of Physics Engineering, Ankara University, 06100, Ankara, Turkey
Keywords Nanostructured titanium dioxide· Sol–gel method· Anatase-to-rutile phase transformation
1 Introduction
Nanostructured titanium dioxide (n-TiO2) materials have attracted great interest due to their potential applications, such as in photocatalysis, energy storage and transfer, pho-tovoltaic solar cell production, sensor design, pigment pro-duction, optical coatings, ceramic manufacturing, wastewa-ter purification, and self-cleaning coating [1–12]. However, the phase and morphology of n-TiO2have been found to be critical parameters in determining their stability for special applications [13–17]. Furthermore, the properties of the n-TiO2thin films strongly depend on the deposition method and annealing temperature. Thermal annealing is widely used to improve crystallinity, which affects electrical and structural properties. During the annealing process, disloca-tions and other structural defects may smear in the material and adsorption/decomposition and, especially at the surface regions, may change the structural, stoichiometric, and elec-trical properties of the material. Therefore, understanding the effects of annealing processes on n-TiO2 surfaces and films is of interest for various technologies that employ this material [18].
TiO2thin films grow in an amorphous and in three main crystalline structures, namely, rutile (tetragonal, P42/mnm), anatase (tetragonal, I4/amd), and brookite (orthorhombic, Pcab) [6], among which rutile is the most stable phase, whereas anatase and brookite are metastable and can irre-versibly transform to rutile when they are thermally treated. The local order in each phase is described by representative octahedra made of O atoms in its vertices and Ti atom near the center. Different phase structures lead to differences in
physical and chemical properties, and then to different appli-cations. The films with dense structure are good in solar cell applications, whereas porous films are better for gas sensors. Rutile phase has good thermal and chemical stability, high dielectric constant and high refractive index which makes it suitable for protective coating lenses [19], gate oxides of metal-oxide-semiconductor field effect transistor [20]. The anatase phase, which has better response with ultraviolet photons, is used for photocatalysis [21, 22], gas detectors [23], and self-cleaning windows [24], whereas amorphous TiO2films are utilized in biomedical fields due to its blood compatibility [25]. Intensive efforts, therefore, have been made to prepare amorphous thin films, and phase pure ru-tile and anatase thin films using various technique.
A number of processing techniques including sol–gel [26–28], screen printing [29], electron-beam evaporation [30], metal organic chemical vapor deposition [31], elec-trospinning [32], mechanical alloying [33], sputtering [12, 15], pulsed laser deposition [9], and aerosol- and ultrasonic-spray pyrolysis [34,35] have been employed for preparing nanostructured TiO2thin films using different raw materials. Hence, the resultant thin films may have similar (nominally the same) composition but will be of different morphologies and functionalities. However, sol–gel method has some ad-ditional advantages over other techniques. The advantages include: (i) low cost of operation, (ii) simple experimental setup, (iii) no requirement of vacuum system, (iv) conve-nience of use (not requiring specialized equipment or tech-niques), (v) potential for mass production (allowing both continuous and large-surface-area production), (vi) ease of doping in solution, (vii) reproducibility of films, and (viii) rapid film growth rates.
The aim of our work is to obtain n-TiO2thin films pro-duced by sol–gel dip coating method and to investigate the effects of annealing temperature on the structural modifi-cation and surface morphology evolution. We also discuss the influence of annealing on the optical characteristics of
n-TiO2 thin films based on optical parameters such as re-fractive index, optical band gap, and relative high dielectric constant.
2 Experimental procedure
In order to prepare a TiO2solution, firstly 2.4 mL titanium tetraispropoxide [Ti(OC3H7)4, ex. Ti≥ 98%, Merck] was added in 25 mL ethanol [C2H6O, 99.9%, Merck], and the solution was kept in a magnetic stirrer for 1 h. Then, 5 mL glacial acetic acide [C2H4O2,99.9%, Merck] and 25 mL ethanol were added in the solution, and after each additive component is added, it was mixed in the magnetic stirrer for 1 h. As a final step, 1.5 mL trietilamine [(C2H5)3N, 99%, Merck] was added in the solution, and the final solution
was subjected to the magnetic stirrer for 2 h. The solution was aged at room temperature for 1 day before deposition. Quartz glass (15× 15 mm) was used as the substrate, and it was cleaned with solvent and rinsed with distilled wa-ter and then blown with nitrogen gas prior to deposition. The dipping process was performed using a home-made mo-torized unit, and each sample was dipped into the solution five times. After each dipping process, samples were sub-jected to repeated annealing processes at the temperature of 200◦C, 300◦C, 400◦C, 500◦C, 600◦C, 700◦C, 800◦C, 900◦C, 1000◦C, and 1100◦C for 5-min period. Finally, the samples were post-annealed at the temperature of 500◦C for 1 hour in air using an electric oven (Vecstar VCTF-4).
The crystalline properties of the TiO2thin films were an-alyzed by an X-ray diffractometer (Model-D8 Advanced, Bruker) using CuKα radiations (λ= 1, 5405 Å) over the range of 2θ= 10–60◦at room temperature. For morpholog-ical investigations, AFM images were recorded using SPM Solver-PRO (NT-MDT) atomic force microscopy controller in a tapping mode. The transmission spectra of TiO2 thin films were measured, at room temperature, by Shimadzu UV-VIS-NIR–3600 spectrophotometer in the range of 300– 1500 nm.
3 Results and discussion
3.1 Phase confirmation by X-ray diffraction analysis Figure1shows the XRD patterns of the n-TiO2 thin films as functions of the annealing temperature. The XRD results indicated that the n-TiO2 thin films annealed from 200◦C to 600◦C are amorphous. The anatase peaks appeared at 700◦C, resulting from a phase transition from amorphous phase to the anatase phase. When the n-TiO2films were de-posited on quartz glass from 700◦C to 900◦C, they showed the anatase phase crystal plane with a (101) preferred orien-tation and the randomly oriented polycrystalline structure. As the temperature increased from 700◦C to 900◦C, the in-tensities of the anatase peaks were increased, implying an improvement in crystallinity. On increasing the temperature to 1000◦C, the intensities of the anatase peaks nearly dis-appeared, while the intensities of the rutile peaks greatly increased. Few weak peaks representing rutile (110) and (211) planes are also observed in the thin films annealed at 1100◦C. The appearance of those new peaks suggests a slight improvement of crystallinity. However, one peak rep-resenting SiO2(101) plane at 1100◦C appears, which corre-sponds to substrate as seen from the corresponding JCPDS files. It has been reported that the onset temperature of ther-mally activated transformation from anatase to rutile took place at temperatures of 600–800◦C. In this work, the phase transformation occurred at 1000◦C. This discrepancy may
Fig. 1 X-ray diffraction patterns of TiO2thin films for
annealed at 200◦C (a), 300◦C (b), 400◦C (c), 500◦C (d), 600◦C (e), 700◦C (f), 800◦C (g), 900◦C (h), 1000◦C (i), and 1100◦C (k)
be due to the difference in crystallite structure and size of TiO2 films [36, 37]. Also, crystalline phase transition temperature depends on a catalyst used in the sol prepara-tion [38].
The crystallite size of n-TiO2thin films can be deduced from XRD line broadening using the Scherrer relation [23]:
D= 0.9λ
βcos θ (1)
where D is the crystallite size (nm), λ is the wavelength of CuKα radiation (nm), θ is the Bragg angle (◦), and β is the full width at half-maximum (FWHM) of diffraction peak. The results are listed in Table1. As the annealing temper-ature increases, the TiO2crystallites continue to grow. The crystallite size of the anatase phase increased from 21.42 to 28.91 nm as the annealing temperature increased from 700◦C to 900◦C. The crystallite size of the rutile phase
at 1000◦C was 30.13 nm and increased to 37.95 nm at 1100◦C. These observations can be explained in terms of higher adatom mobility with increasing annealing tempera-ture, which results in the larger crystallite size and enhances the crystallinity of the thin films.
3.2 Morphological characterization
The evolution of surface morphology, roughness and gran-ularity has been characterized by AFM. Figure 2 shows the two- and three-dimensional AFM images of the thin films annealed at different temperatures. All TiO2thin films exhibit a smooth surface with uniform grains. Although having the same granularity when deposited, the original nanograins transform into spherical particles with different dimensions, accounting for the crystallization of n-TiO2into a polycrystalline material. As shown in Fig.2, the surface
Table 1 The values of thickness, energy band gap, high dielectric constant, refractive index, crystallite size, roughness and average grain size of TiO2thin films at different annealing temperature
200◦C 300◦C 400◦C 500◦C 600◦C 700◦C 800◦C 900◦C 1000◦C 1100◦C
Thickness, d (nm) 444 295 272 251 218 179 168 148 – –
Energy band gap (eV) 3.69 3.63 3.60 3.53 3.52 3.46 3.42 3.40 2.93 2.79
High dielectric constant, εs 5.13 5.71 5.82 7.17 7.43 6.49 6.70 5.64 – –
Refractive index, n 2.38 2.53 2.71 2.83 2.94 3.09 3.16 2.46 – –
Crystallite size (nm) – – – – – 21.42 25.61 28.91 30.13 37.95
Roughness (nm) 1.82 2.06 2.24 3.58 3.78 3.88 3.97 4.18 4.64 5.20
Average grain size (nm) 39 43 49 55 59 70 118 153 195 313
morphology reveals the n-TiO2 grains, which combine to make denser films significantly with the increased anneal-ing temperatures. As shown in Table1, average grain size of thin films increases from about 39 to 313 nm with increas-ing annealincreas-ing temperature. This may be due to the bigger clusters formed by the coalescence of two or more grains, and increasing of roughness. The root mean square (rms) is the most widely used parameter to characterize surface roughness. The rms roughness of the thin films increases from about 1.82 to 5.20 nm with increase of the annealing temperature (see Table 1). As seen from Table1, the sur-face roughness of the thin amorphous films is smaller than that of the crystallized films, a common observation for the amorphous oxides [39,40].
3.3 Optical characterization
Figure 3 indicates the transmittance spectra of TiO2 thin films in the wavelength range from 300 nm to 1500 nm as functions of the annealing temperature. The transmission spectrum can be roughly divided into two regions: a trans-parent one with the interference pattern in the visible and IR region and a zone of strong absorption in the near-UV region, where the transmittance T decreases drastically owing to the effect of the absorption coefficient. TiO2thin films exhibit high transparency about 97% at 700 nm. As shown in Fig. 2, the transmittance maxima increase with the increasing annealing temperature up to 900◦C and after-wards decrease slightly. This is because TiO2films annealed at 1000◦C and 1100◦C are purely rutile, which has a more intense absorbance and a smaller bandgap than anatase. Ad-ditionally, the reduction of transmittance observed in the transmittance spectrum when the annealing temperature in-creased is consistent with the increase of surface scattering related to the surface roughness [41].
As we all know, a typical transmission spectrum at nor-mal incidence has two spectral regions: the region of weak and medium absorption and the strong absorption region. In the weak and medium absorption region, the refractive index
(n)of the film can be calculated by the following expression [42]: n=N+N2− n1/2s 1/21/2 (2) where N=1 2 1+ n2s+2ns(TM− Tm) TMTm (3) where ns is the refractive index of quartz substrate, and TM
and Tm are points of the maxima and the minima of
trans-mission spectrum, respectively. The basic equation for inter-ference fringes is
2nd= mλ (4)
where the order numbers m is the integer for maxima and half integer for minima. If n(λ1)and n(λ2)are the refractive indices at two adjacent maxima (or minima) at λ1and λ2, the film thickness (d) can be expressed by
d= λ1λ2
2[n(λ1)λ2− n(λ2)λ1]
(5) The thicknesses of n-TiO2thin films were also determined from transmittance measurements and given in Table1. As seen from Table1, the film thickness decreases with increas-ing annealincreas-ing temperature.
Following the procedure described above, the physical parameters n and d were determined for all the annealed films. The curves and values of refractive index for as grown
n-TiO2 thin film are shown in Fig.4and Table1, respec-tively. As seen in Fig.4, the refractive index values increase with increasing the annealing temperature. The increment of the refractive index up to 900◦C can be explained by the increase of packing density as confirmed by the AFM sults. If the XRD results are combined, the increase of the re-fractive index at higher temperatures above 600◦C could be related to the anatase-to-rutile phase transition at the same wavelength.
Near the absorption edge or in the strong absorption zone of the transmittance spectra, the absorption coefficient α is
Fig. 2 2-D and 3-D AFM images of TiO2thin films for annealed at 200◦C (a), 600◦C (b), 700◦C (c), 900◦C (d), 1000◦C (e), and 1100◦C (f),
Fig. 3 Transmission spectra for TiO2thin films at different annealing
temperatures
Fig. 4 Refractive index dispersion spectrum of TiO2thin films at
dif-ferent annealing temperatures
related to the optical energy gap Egfollowing the power-law
behavior of Tauc’s relation [43]
(αhv)= A(hv − Eg)r (6)
where α is the absorption coefficient, A is an energy-independent constant between 107and 108m−1, Eg is the
optical band gap, and r is a constant, which determines type of optical transition; r= 1/2, 2, 3/2 or 3 for allowed direct, allowed indirect, forbidden direct, and forbidden indirect
Fig. 5 (αhv)2versus (hv) plots of TiO
2thin films at different
anneal-ing temperatures
electronic transitions, respectively. Figure5shows the plot of (αhv)2versus energy (hv) according to (6). The optical energy gaps Eg of the thin films determined by
extrapolat-ing the linear portion of this plot at (αhv)2= 0 for r = 1/2, which indicates that the direct transition dominates in the n-TiO2thin films. The direct band gaps (Eg) of thin films were
calculated from Fig.5and are given in Table1. As clearly shown in Fig.5 and Table1, it is evident that an increase in annealing temperature leads to a decrease in optical band gaps. The decrease in the optical band gap is attributed to the lowering of the interatomic spacing, which may be associ-ated with a decrease in the amplitude of atomic oscillations around their equilibrium positions [44].
The properties of the investigated TiO2thin films could be treated as a single oscillator at wavelength λ0at high fre-quency. The high-frequency dielectric constant (ε∞)can be calculated by applying the following simple classical disper-sion relation [45]: (n20− 1) (n2− 1)= 1 − λ0 λ 2 (7) where n0 is the refractive index at infinite wavelength λ0 (average interband oscillator wavelength), n the refractive index, and λ the wavelength of the incident photon. Plot-ting (n2− 1)−1against λ−2, which showed linear part, was below the absorption edge as shown in Fig.6. The intersec-tion with (n2− 1)−1axis is (n20− 1)−1and hence, n20at λ0 equal ε∞(high-frequency dielectric constant). The values of
ε∞for TiO2thin films are given in Table1. As clearly seen in Fig.6and Table1, the optical high-frequency dielectric
Fig. 6 Plots of (n2− 1)−1against λ−2for TiO2thin films at different
annealing temperatures
constant increases with increasing the annealing tempera-ture. The increase may be attributed to higher packing den-sity within the film, lattice vibrations, and bounded carriers in an empty lattice.
4 Conclusions
In this study, we show that the as grown amorphous films crystallize in anatase structure on annealing above 600◦C and partially convert to rutile phase in the temperature range of 1000–1100◦C. Annealing at higher temperature, however, meets two competing effects, growth of anatase grains and their conversion to rutile phase, thus suppressing the con-version rate beyond 700◦C and inhibiting complete conver-sion to rutile phase even at 1000◦C. Also, the thin films an-nealed at different temperatures consisted of spherical parti-cles having submicron diameters and had important change upon annealing. The surface roughness (rms) of the thin films increases from 1.82 to 5.20 nm, and at the same time, average grain size as well grows up from about 39 to 313 nm with increase of the annealing temperature.
Two types of optical transmission data were observed, where anatase films showed relatively high and constant UV-VIS transmission, but rutile films showed lower UV-VIS transmission, which increased with increasing wavelength. The preceding data are attributed to the effect of the pho-tonmean free paths, where the grain size of the anatase was lower than visible light wavelengths, but the grain sizes of rutile traversed the scale of visible lightwavelengths. In the
latter case, owing to the effect of light scattering at the grain boundaries, the transmission and grain size changed propor-tionately. Besides, the refractive index, optical band gap, and high dielectric constant were influenced by both annealing process and the anatase-to-rutile phase transition in the TiO2 films.
The anatase-to-rutile phase transition and the modifica-tion of microstructure during the annealing process were key factors causing the variations of optical and structural prop-erties such as transmittance, refractive index, high dielec-tric constant, optical band gap and crystallite size, surface roughness, and average grain size. It was demonstrated that the annealing was a controllable process in order to make the suitable structural and optical properties of n-TiO2films for various applications.
References
1. A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C 1, 1 (2000)
2. B. O’Regan, M. Grätzel, Nature 353, 737 (1991) 3. E.W. Mcfarland, J. Tang, Nature 421, 616 (2003)
4. H. Tang, K. Prasad, R. Sanjines, F. Levy, Sens. Actuators B 26, 71 (1995)
5. J.H. Braun, J. Coat. Technol. 69, 59 (1997) 6. U. Diebold, Surf. Sci. Rep. 48, 53 (2003)
7. R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. Kojima, A. Kitamura, M. Shimohigoshi, T. Watanabe, Nature 388, 431 (1997) 8. A. Fujishima, K. Honda, Nature 238, 37 (1972)
9. G.-B. Wang, M.-G. Fu, B. Lu, G.-P. Du, L. Li, X.-M. Qin, W.-Z. Shi, Appl. Phys. A 100, 1169 (2010)
10. I.N. Kholmanov, E. Barborini, E. Vinati, P. Piseri, A. Podesta, C. Ducati, C. Lenardi, P. Milani, Nanotechnology 14, 1168 (2003) 11. B. Karunagaran, P. Uthirakumar, S.J. Chung, S. Velumani, E.K.
Suh, Mater. Charact. 58, 680 (2007)
12. J. Alami, K. Sarakinos, F. Uslu, C. Klever, J. Dukwen, M. Wuttig, J. Phys. D, Appl. Phys. 42, 115204 (2009)
13. A.S. Barnard, L.A. Curtiss, Nano Lett. 5, 1261 (2005)
14. L.X. Chen, T. Rajh, W. Jäger, J. Nedeljkovic, M.C. Thurnauer, J. Synchrotron Radiat. 6, 445 (1999)
15. F. Armani, M. Gougis, S.A. Impey, A.C. James, K. Lawson, L. Lihrmann, M. Stock, S. Dunn, Mater. Lett. 64, 140 (2010) 16. H.A. Bullen, S.J. Garrett, Nano Lett. 2, 739 (2002)
17. H.D. Jang, S.K. Kim, S.J. Kim, J. Nanopart. Res. 3, 141 (2001) 18. Y.Q. Hou, D.M. Zhuang, G. Zhang, M. Zhao, M.S. Wu, Appl.
Surf. Sci. 218, 98 (2003)
19. H. Takikawa, T. Matsui, T. Sakakibara, A. Bendavid, P.J. Martin, Thin Solid Films 348, 145 (1999)
20. H.S. Kim, D.C. Gilmer, S.A. Campbell, D.L. Polla, Appl. Phys. Lett. 69, 3860 (1996)
21. M.R. Hoffmann, S.R. Martin, D.W. Choi, D.W. Bahnemann, Chem. Rev. 95, 69 (1995)
22. N.R. Mathews, E.R. Morales, M.A. Cortes-Jacome, J.A.T. Anto-nio, Sol. Energy 83, 1499 (2009)
23. M. Ferroni, V. Guidi, G. Martinelli, P. Nelli, G. Sberveglieri, Sens. Actuators B 44, 499 (1997)
24. T. Hanawa, M. Ota, Appl. Surf. Sci. 55, 269 (1992)
25. J.X. Liu, D.Z. Yang, F. Shi, Y.J. Cai, Thin Solid Films 429, 225 (2003)
26. J.K. Lee, H.S. Jung, Y.Q. Wang, N.D. Theodore, T.L. Alford, M. Nastasi, Appl. Phys. A 103, 179 (2011)
27. R. Poyato, M.L. Calzada, L. Pardo, Appl. Phys. A 80, 369 (2005) 28. M. Zaharescu, M. Crisan, I. Musevic, J. Sol-Gel Sci. Technol. 13,
769 (1998)
29. I. Seigo, P. Chen, P. Comte, M.K. Nazeeruddin, P. Liska, P. Péchy, M. Grätzel, Prog. Photovolt. Res. Appl. 15, 603 (2007)
30. V. Mikhelashvili, G. Eisenstein, J. Appl. Phys. 89, 3256 (2001) 31. X.H. Xu, M. Wang, Y. Hou, S.R. Zhao, H. Wang, D. Wang, S.X.
Shang, Cryst. Res. Technol. 37, 431 (2002)
32. C.W. Jia, E.Q. Xie, J.G. Zhao, H.G. Duan, J. Appl. Phys. 101, 093509 (2007)
33. D.H. Kim, H.S. Hong, S.J. Kim, J.S. Song, K.S. Lee, J. Alloys Compd. 375, 259 (2004)
34. M. Okuya, K. Nakade, S. Kaneko, Sol. Energy Mater. Sol. Cells 70, 425 (2002)
35. A. Nakaruk, D. Ragazzon, C.C. Sorrell, Thin Solid Films 518, 3735 (2010)
36. N. Negishi, K. Takeuchi, Mater. Lett. 38, 150 (1999)
37. D.J. Won, C.H. Wang, H.K. Jang, D.J. Choi, Appl. Phys. A 73, 595 (2001)
38. T. Hashimoto, T. Yoko, S. Saka, Bull. Chem. Soc. Jpn. 67, 653 (1994)
39. D. Mardare, G.I. Rusu, Mater. Lett. 56, 210 (2002)
40. H. Tang, K. Prasad, R. Sanjines, P.E. Schmid, F. Levy, J. Appl. Phys. 75(4), 2042 (1994)
41. M.H. Suhail, G.M. Rao, S. Mohan, J. Appl. Phys. 71, 1421 (1992) 42. R. Swanepoel, J. Phys. E, Sci. Instrum. 16, 1214 (1983) 43. J. Tauc, Mater. Res. Bull. 5, 721 (1970)
44. S. Sönmezo˘glu, G. Çankaya, N. Serin, Int. J. Nat. Eng. Sci. 5(13), 57 (2011)
