Elaboration and study of the properties of nano- crystalline TiO 2 thin films prepared by sol–gel dip
coating
Y. Bouachiba1, 2,*, A. Bouabellou1,F. Hanini1, A. Taabouche1,2, H. Nezzari2.
1 Thin Films and Interfaces Laboratory, University of Constantine 1, 25000 Constantine, Algeria
2 Welding and NDT Research Center (CSC), BP 64, Cheraga, Algeria
*ybouachiba@gmail.com
Abstract— TiO2 thin films for gas sensing applications are prepared by non-aqueous sol–gel method on glass substrates. The structural evolution of TiO2 films with thickness is investigated by X-ray diffraction and Raman Spectroscopy. Prepared films are in anatase phase. The grain size calculated from XRD patterns increases with thickness from 14 to 22 nm. Uv–vis transmission spectra show that the absorption edge shifts to longer wavelength as the thickness increases, which is correlated with the change in the optical band gap value. The determination of the refractive index and thickness of TiO2 thin films by m-lines spectroscopy is presented. Refractive index is found to vary slightly with thickness. Calculation of the film density confirms the behavior of refractive index. The films exhibit one guided TE0 and TM0 polarized modes.
Index Terms— Sol–gel, TiO2, Anatase, Refractive index, Optical band gap, m-lines.
I. INTRODUCTION
Titanium oxide (TiO2) has been widely investigated for various applications such as photocatalysis [1, 2], dye- sensitive solar cells [3, 4], electrochromic displays [5–7], gas sensing [8], superhydrophilicity [9] and optoelectronic materials [10].
Determination of the optical constants of thin films, e.g., refractive index and band gap energy is a topic of fundamental and technological importance. Film thickness is also an important factor that can influence the properties of TiO2 thin films. The studies indicate that the optical and structural properties of TiO2 films depend on the process conditions and the materials used in the sol-gel process. In some studies, effect of film thickness on the performance of thin TiO2 films has been investigated [11, 12].
For accurate measurement of optical constants and thickness, many attempts and theories were established.
Spectroscopic ellipsometry is a non-destructive optical technique. It could be used to extract thickness and optical
constants through proper optical modeling [13]. Refractive index and thickness of a transparent homogeneous film on a non-absorbing flat substrate can be determined accurately from its reflectance and transmittance spectra [14]. In general, optical thin films constants may be determined either by multiple wavelength methods or by single wavelength methods [15]. One of the most important single wavelength techniques is the m-lines spectroscopy [16], which has been found favorably for the non-destructive characterization of thin solid films, especially semiconductors. This technique is of high sensitivity, high accuracy, of being able to easily in- situ measure both optical constants and thickness simultaneously. Several rigorous conditions have to be satisfied in this technique. Indeed, the films have to have high refractive index related to the substrate, exhibit low level of defects, be absorption-free in the used wavelength range and possess sufficient thickness.
TiO2 films have been prepared by various techniques such as sol-gel [17, 18], reactive sputtering [19], pulsed laser deposition [20] and spray pyrolysis [21]. Sol-gel technique has emerged as one of the main techniques for growing anatase thin films, since optical and other properties of the films can be controlled with relative ease by changing the solution composition and deposition process details. Moreover, equipment costs are moderate. Titanium alkoxides have been frequently used as a precursor material for sol-gel deposition of such TiO2 films. Hydrolysis of titanium alkoxides is usually achieved by adding water. However, when precursors with strong reactivity toward water (like titanium alkoxides) are used, it is difficult to form stable sols.
This can be overcome by using non-aqueous sol-gel chemistry with no added water. In this case, the water needed for hydrolysis comes from esterification reaction between alcohol and catalysis [22]. Consequently, non-aqueous sol-gel processes have been developed to achieve TiO2 sols with lower reactivity and long-term stability.
In this paper, we report the study of the structural and optical properties of TiO2 thin films deposited on glass substrates by sol–gel dip coating technique as a function of the film thickness. Structural evolution is investigated by X-ray Diffraction (XRD) and confirmed by Raman Spectroscopy measurements. Optical band gap Eg for the respective films extrapolated by Tauc plot is discussed. The values of refractive index are compared with those of literatures for TiO2 thin films and the reasons accounting for these values are inferred.
II. MATERIALS AND METHODS
A. Experimental
The TiO2 thin films were prepared by non-aqueous sol–gel process, which is based on the hydrolysis of alkoxydes in alcoholic solutions in the presence of an acid catalyst. The procedure of preparation includes the dissolution of titanium isopropoxide in isopropanol alcohol. The solution was left under closed stirring during 10 minutes. Then, acetic acid was poured, stirred during 15 minutes. Finally, methanol were added and stirred during 2 hours. This obtained solution is transparent, of yellowish color. The glass substrates carefully cleaned were dipped into the solution and were pulled up at a constant rate of 10 cm/min. After each dipping, the obtained films were dried for 15 min. Finally, TiO2 thin films were annealed at 500°C for 2 hours.
XRD patterns were recorded on Siemens D8 diffractometer using Copper Kα radiation. The Raman spectra were recorded at room temperature with Renishaw invia Raman microscope equipped with a motorized xy stage and autofocus. The observation of surface morphology for TiO2
films in a region of 10.14×10.14 μm2 areawas carried out using an atomic force microscopy (Pacific Nanotechnology) operating in contact mode. The optical transmittance was measured by using Shimadzu 3101 PC UV–visible spectrophotometer. The waveguiding properties were investigated by a Metricon 2010/M prism coupler. A right- angle rutile prism P2 (nTE = 2.8639, nTM = 2.5822) was used for coupling light of a He–Ne laser with a wavelength λ = 632.8 nm into the waveguide.
B. Results and discussion B.1. Structural properties
The effect of the film thickness on structural properties of Sol–gel TiO2 films has not been got main attention until now.
Since our study tries to clarify this relation we deposited TiO2
films with a thickness ranging approximately from 120 to 220 nm. The film thickness was varied by the change of the dip- coating times. The XRD patterns of TiO2 films dip-coated three to five times are displayed in Figure 1. The XRD measurements indicate that the films exhibit (1 0 1) diffraction peak as the main line of the polycrystalline anatase. This peak becomes more intense and sharper as the thickness increases, suggesting that the average particles size is enhanced.
Generally, when the film thickness increases the X-ray radiation comes in a larger volume of the film through, and thus the measured signal has a higher intensity [23].
Moreover, the peak intensity is also in relation to the crystallinity. The diffraction pattern of the films dip-coated five times shows, besides the peak assigned to the (1 0 1) plane, an other weak peak assigned to (2 0 0) plane.
The grain size in the films can be estimated by Scherrer formula [24]:
θ β
λ cos
90 . 0
) (nm
=
D
(1)Where D is the crystallite size, λ the wavelength of X-ray (Cu Kα= 1.5406 Å), β the peak full-width at half-maximum and θ the diffraction angle. The results are tabulated in table 1. The grain size values increase quasi linearly as the thickness increases.
TABLE I GRAIN SIZE OF THE TIO2 THIN FILMS FOR VARIOUS LAYERS.
Sample grain size (nm) (hkl)
3 layers 14 (101)
4 layers 17 (101)
5 layers 21 (101)
To confirm the results of XRD measurements, Raman spectroscopy was used to quantify the vibration modes of TiO2 thin films. Figure 2 shows the Raman spectra in the range of 100–800 cm-1. It is well known that generally the TiO2 anatase phase is characterized by six raman modes: A1g
+2B1g +3Eg. The Raman spectrum for anatase TiO2 was identified by Ohsaka [25] at 144 (Eg), 197 (Eg), 399 (B1g), 513 (A1g), 519 (B1g) and 639 (Eg) cm−1.
The recorded spectra in this study show five identified symmetric vibration modes localized at 144 (Eg), 198 (Eg), 399 (B1g), 516 (B1g), and 639 (Eg) cm−1. No peaks relating to rutile phase can be found in the Raman spectra, indicating the pure anatase structure of the films.
Figure 1: X-ray diffraction pattern of TiO2 thin films: 3 layers (a), 4 layers (b), and 5 layers (c).
Figure 2: The Raman spectra of TiO2 thin films: 3 layers (a), 4 layers (b), and 5 layers (c).
B2. Surface morphology
Figure 3 shows the AFM images of dip-coated TiO2 films.
From the Analysis of AFM images, the root mean squared (rms) roughness of three, four and five layers TiO2 films is about 1.246 nm, 1.69 nm and 2.11 nm, respectively. The relatively uniform growth of the deposited film across a 10 μm2 area and the low surface roughness observed in the samples allow to expect satisfying optical properties.
Figure 3: AFM images of TiO2 films: 3 layers (a), 4 layers (b), and 5 layers (c).
B3. Optical properties
B3.1. UV transmittance analysis
UV–Vis transmission spectra of studied TiO2 films are shown in Fig. 4. Results indicate that all prepared films possess a good transmission of light in the visible region. As a general feature, the spectra of the films show waveforms (ripples) as the thickness increases, which are characteristic of the interference of light [26].
Due to the fundamental absorption in the vicinity of band gap, the transmission decreases sharply as the wavelength reaches the ultraviolet radiation. It can be found that the absorption edge shifts to longer wavelength as the thickness increases, which is correlated with the change in the optical band gap value.
It is well known that the dispersion of the absorbance of a semiconductor is described by the following relationship [27]:
p
Eg
h A
hυ= ( υ− )
α (2) (a)
(b)
(c)
where α is the absorbance, hυ is the photon energy, Eg is the band gap of the material and p = 1/2, 2, 3/2 or 3 for allowed direct, allowed indirect, forbidden direct or forbidden indirect transitions, respectively. Since allowed indirect transition dominates the absorption in the optical region for TiO2, the value of p is taken as 2.
The absorbance is mainly influenced by two factors:
scattering losses and fundamental absorption. At shorter wavelengths close to the optical band gap, the influence of fundamental absorption on α is more prominent than to scattering losses and α may be obtained by [28]:
⎟⎠
⎜ ⎞
⎝
= ⎛ T d
ln 1
α 1 (3)
where d is the thickness of the film and T the transmittance.
The plot of (αhυ)1/2 vs hυ is expected to exhibit a linear Tauc region just above the absorption edge, allowing us to determine the optical band gap of TiO2 films. An example of the plot of (αhν)1/2 versus photon energy of the films, including an extrapolation from the linear curve, is illustrated in Fig.5. As we can see, The Tauc plot of the films graphically illustrates how the absorption edge shifts to lower photon energies as the thickness is increased. The determined values of the optical band gap are 3.41, 3.32 and 3.28 eV for three, four and five TiO2 layers respectively. However, one can observe a weak variation in the optical band gap (~ 0.1 eV). It appears that Eg approaches the bulk value as TiO2 film becomes thicker. The decrease in the optical band gap of the TiO2 films with thickness might be the result of the increasing in grain size. These results are in good agreement with the previously reports for optical absorption–thickness variation [29].
Figure 4. UV–Vis transmittance spectra of the TiO2 thin film, for various layers: 3 layers (a), 4 layers (b), and 5 layers (c).
Figure 5. Optical absorption spectra plotted as (αhν)1/2 versus hν of 3 layers (a), 4 layers (b), and
5 layers (c).
B4. Waveguiding measurements
Prism coupler technique is a useful method to determine optogeometric parameters of waveguiding thin films, such as thickness and refractive index [30]. For discrete values of the incidence angle, the evanescent wave, existing in the air gap between the prism and the film, may excite the guided modes in the film. The coupling is detected by the apparition of dark lines in the reflected beam and the effective index may be calculated for each mode. In our study, the dip-coated waveguide films can just support one propagating mode of each polarization (one transverse electric mode: TE0 and one transverse magnetic mode: TM0). For example figure 5 exhibits the optical guided modes of the five layers film. Under the assumption of that the film is not be birefringent, the refractive index and the film thickness are calculated using the TE/TM combination option (using TE0 and TM0 modes). The obtained values are reported in table 2.
TABLE 2 MEASURED VALUES OF THE OPTO-GEOMETRIC PARAMETERS OF TIO2 FOR VARIOUS LAYERS.
Sample
Film thickness (± 0.1 nm)
Refractive index (± 10-4)
3 layers 122.3 2.0820
4 layers 161.6 2.0876
5 layers 222.9 2.0907
As shown in Table 2, no significant change is observed in the values of n with respect to film thickness. In the present study, the values of refractive index lie in the range ~ 2.08–
2.09 for the TiO2 films. The deduced values of refractive index are in good agreement with that obtained by sol–gel methods [31–33]. Refractive indices of TiO2 films prepared at 500 °C by sol–gel process were obtained in the range 2.08–
2.17 by Hu et al. [31]. In the work of Sreemany et al. [32], the values of n lie in the range ~ 2.03–2.13. Refractive index of
dip-coated TiO2 films using m-lines measurements was found to be slightly greater than 2.06 by Mechiakh et al. [33]. In our study, no appreciable increment in the values of n with thickness may be explained by the fact that the film densification almost attains its saturation point. Therefore, after three coatings, refractive index of multiple-coated TiO2
film remains almost unchanged with further increment in the film thickness.
This explanation is confirmed by calculation of the film density using the well known Clausius–Mossotti relation [34]:
⎟⎟⎠
⎜⎜ ⎞
⎝
⎛
−
⎟ +
⎟
⎠
⎞
⎜⎜
⎝
⎛ +
= −
⎟⎟⎠
⎜⎜ ⎞
⎝
=⎛
1 2 2
1
2 2 2
2
b b f
f b f
n n n
P n ρ
ρ (4)
Where nf and nb are the refractive index of the film and the bulk material of TiO2. The ρf and ρb are the density of the film and bulk material, respectively. In the present study, nb is taken as 2.57 [35]. The determined values of the density are 82.1, 82.4 and 82.6 % for three, four and five TiO2 layers respectively. This behavior also supports our above result that one of the reasons for the weak decrease in band gap with thickness (~ 0.1 eV) can be the weak variation in the densification of the films.
In the sol–gel method, the values of n are always smaller than that of bulk material, which is typical to thin films due to lower atomic densities of thin film materials comparing to bulk densities [36]. As a consequence, calculations using only one TE0 mode and one TM0 mode are acceptable.
Figure 6: Optical guided modes of the TiO2 five layers film.
III. CONCLUSIONS
The influence of film thickness on the transparency, refractive index, packing density, band gap energy and structural of sol–
gel derived TiO2 films has been studied. X-ray diffraction analysis show that the obtained thin films crystallize into tetragonal titanium oxide anatase. The Raman spectra confirm the presence of anatase phase. The grain size and the root mean squared (RMS) roughness values increase as the thickness increases. The analysis of the UV–Vis transmission spectra show that TiO2 thin films are transparent in the visible range and opaque in the UV region and the absorption edge shifts to longer wavelength as the thickness increases. This is due to the decrease in the optical band gap. The investigation of waveguiding measurements reveals the excitation of one guided TE0 and one TM0 polarized modes. The refractive index and the thickness have been determined using TE0 and TM0
modes. The values of refractive index are slightly greater than
~ 2.08. This may be explained by the fact that the film densification almost attains its saturation point. The values of refractive index are close to that reported in literatures for TiO2
thin films.
REFERENCES
[1] J. Zitaa, J. Krysa, U. Cernigoj, U. Lavrencic-Stangar, J.
Jirkovsky, J. Rathousky, Catalysis Today, 161 (2011) 29–34.
[2] C.H. Huang, C.C. Tsao, C.Y. Hsu, Ceramics International, 37 (2011) 2781–2788.
[3] Y. Liu, X. Sun, Q. Tai, H. Hu, B. Chen, N. Huang, B. Sebo, X-Z.
Zhao, Journal of Alloys and Compounds, 509 (2011) 9264–
9270.
[4] X. Tang, J. Qian, Z. Wang, H. Wang, Q. Feng, G. Liu, Journal of Colloid and Interface Science, 330 (2009) 386–391.
[5] A. Verma and S.A. Agnihotry, Electrochimica Acta, 52 (2007) 2701–2709.
[6] A. Verma, M. Kar, S.A. Agnihotry, Solar Energy Materials &
Solar Cells, 91 (2007) 1305–1312.
[7] A. Verma, A. Basu, A.K. Bakhshi, S.A. Agnihotry, Solid State Ionics, 176 (2005) 2285– 2295.
[8] E. Gyorgy, G. Socol, E. Axente, I.N. Mihailescu, C. Ducu, S.
Ciuca, Applied Surface Science, 247 (2005) 429–433.
[9] Y. Tsuge, J. Kim, Y. Sone, O. Kuwaki, S. Shiratori, Thin Solid Films, 516 (2008) 2463–2468.
[10] N.E. Stankova, I.G. Dimitrov, T.R. Stoyanchov, P.A. Atanasov, D. Kovacheva, Applied Surface Science, 255 (2008) 5275–5279.
[11] H. Tada, M. Tanaka, Langmuir, 13 (1997) 360.
[12] M. Takahashi, K. Tsukigi, T. Uchino, T. Yoko, Thin Solid Films, 388 (2001) 231.
[13] H.-Q. Jiang, Q. Wei, Q.-Xi. Cao, X. Yao, Ceramics International, 34 (2008) 1039–1042.
[14] D.J. Kim, S.H. Hahn, S.H. Oh, E.J. Kim, Materials Letters, 57 (2002) 355–360.
[15] J.A. Dobrowolski, F.C. Ho, A. Waldrof, Appl. Opt. 22 (1983) 319.
[16] R. Chassagnon. O. Marty, P. Moretti, C. Urlacher, J. Mugnier, Nuclear Instruments and Methods in Physics Research, B I22 (1997) 550–552.
[17] D.V. Wellia, Q.C. Xu, M.A. Sk, K.H. Lim, T.M. Lim, T.T.Y. Tan, Applied Catalysis A: General, 401 (2011) 98–105.
[18] M. Bockmeyer, B. Herbig, P. Löbmann, Thin Solid Films 517 (2009) 1596–1600.
[19] C.H. Huang, C.C. Tsao, C.Y. Hsu, Ceramics International 37, (2011) 2781–2788.
[20] B. Farkas, J. Budai, I. Kabalci, P. Heszler, Zs. Geretovszky, Applied Surface Science 254 (2008) 3484–3488.
[21] I. Oja, A. Mere, M. Krunks, R. Nisumaa, C.-H. Solterbeck, M. Es- Souni, Thin Solid Films, 515 (2006) 674 – 677.
[22] H. Choi, E. Stathatos, D. D. Dionysiou, Thin Solid Films 510 (2006) 107 – 114.
[23] I.A. Al-Homoudi, J.S. Thakur, R. Naik, G.W. Auner, G. Newaz, Applied Surface Science 253 (2007) 8607–8614.
[24] B.D. Culity, S.R. Stock, Elements of X-Ray Diffraction, Prentice Hall, Inc., New Jersey, 2001.
[25] T. Ohsaka, S. Yamahoka, O. Shimomura, Solid State Commun. 30 (1979) 345.
[26] M. Addamo, V. Augugliaro, A. Di Paola, E. García-López, V. Loddo, G. Marcì, L. Palmisano, Thin Solid Films 516 (2008) 3802–3807.
[27] D.G. Syarif, A. Miyashita, T. Yamaki, T. Sumita, Y. Choi, H.
Itoh, Applied Surface Science 193 (2002) 287–292.
[28] D. Mardare, M. Tasca, M. Delibas, G.I. Rusu, Applied Surface Science 156 (2000) 200–206.
[29] A.M. More, J.L. Gunjkar, C.D. Lokhande, R.S. Mane, S.-H. Han, Micron 38 (2007) 500–504.
[30] M. Ulrich and R. Torge, Appl. Opt. 12 (1973) 2901–2908.
[31] L. Hu, T. Yoko, H. Kozuka, S. Sakka, Thin Solid Films 219 (1992) 18.
[32] M. Sreemany and S. Sen, Materials Research Bulletin 42 (2007) 177–189.
[33] R. Mechiakh, F. Meriche, R. Kremer , R. Bensaha , B. Boudine , A. Boudrioua, Optical Materials 30 (2007) 645–651.
[34] M. Subramanian, S. Vijayalakshmi, S. Venkataraj, R. Jayavel, Thin Solid Films 516 (2008) 3776–3782.
[35] A. Palik, Handbook of Optical Constants, Academic, New York, 1985.
[36] E. Haimi, H. Lipsonen, J. Larismaa, M. Kapulainen, J. Krzak-Ros, S.-P. Hannula, Thin Solid Films 519 (2011) 5882–5886.
