On the Crystal Structural Control of Sputtered TiO2 Thin Films
© The Author(s). 2016
Received: 4 May 2016
Accepted: 24 June 2016
Published: 7 July 2016
In this study, we focused on the origin on the selective deposition of rutile and anatase TiO2 thin films during the sputtering process. The observation on microstructural evolution of the TiO2 films by transmission electron microscopy revealed the coexistence of rutile and anatase TiO2 phases in the initial stage under the preferential growth conditions for the anatase TiO2; the observations further revealed that the anatase phase gradually dominated the crystal structure with increasing film thickness. These results suggest that the bombardment during the sputtering deposition did not obviously affect the TiO2 crystal structure, and this was also confirmed by off-axis magnetron sputtering experiments. We also investigated the mechanism of the effect of Sn impurity doping on the crystal structure using first-principles calculations. It is found that the formation energy of Sn-doped rutile TiO2 is lower than that of Sn-doped anatase TiO2; this suggests that the Sn-doped TiO2 favours the rutile phase. These results offer a guideline for the utilization of selective deposition of rutile and anatase TiO2 thin films in various industrial applications.
Rutile and anatase TiO2 films are widely used in various industrial applications [1–3]. For example, rutile TiO2 films are used as an optical coating material because of their high refractive index whereas anatase TiO2 films are utilised as photocatalysts or transparent electrodes . Rutile TiO2 is the most common phase in nature, and anatase TiO2 transforms to rutile at temperatures above 400–600 °C .
Conventional wet processes such as the sol–gel method can be used to produce pure-phase TiO2 films; however, fabricating dense TiO2 films is difficult by this method . Sputtering deposition can be used to produce uniform TiO2 thin films with a large area, high packing density and strong adhesion [1, 6]. However, TiO2 films deposited by magnetron sputtering are often a mixture of anatase and rutile phases. As a practical measure, controlling the phase content of TiO2 films is necessary for films used in precise optical applications. Therefore, the deposition of pure-phase TiO2by magnetron sputtering has attracted much attention. Currently, two approaches can be used to fabricate pure-phase sputtered TiO2 films. One is to control the sputtering conditions such as the total gas pressure, substrate temperature and type of sputtering gas to selectively fabricate uniform coatings of rutile or anatase TiO2 films ; the other is to use impurity doping to induce a phase transformation between the anatase and rutile phases [1, 7–14].
In the sputtering process, rutile and anatase TiO2 films are easily fabricated under low and high total gas pressure, respectively [1, 15, 16]. A general explanation for this observation is that bombardment by high-energy particles such as negative oxygen ions can lead to a dense rutile TiO2 phase . However, the bombardment effects have not been confirmed experimentally, and the mechanism by which the rutile or anatase TiO2 phase grows during the sputtering process remains unknown. In the case of impurity doping, elements such as Mn , Fe , Cu , Ag , Ni  and Co  have been reported to enhance the phase transition from the anatase to the rutile phase, whereas other elements such as W , V , Si , Nb , Ta  and Cr  have been reported to suppress the anatase-to-rutile phase transition. On the basis of first-principles calculations, the room-temperature phase conversion of anatase to rutile TiO2 using Co or Ni doping is attributed to the increased interaction between Co and Ni atoms, which results in the formation of a linear chain in the rutile phase . In a previous study, we demonstrated that Sn doping can induce the anatase-to-rutile transformation in a sputtered TiO2 film . However, the related mechanism for the transformation induced by Sn doping has not yet been elucidated.
In this study, we first use transmission electron microscopy (TEM) to observe the microstructural evolution of TiO2 films during sputtering. Second, we investigate the bombardment effects of high-energy particles on the crystal structure of TiO2 films using the off-axis sputtering method and discuss the effect of sputtered Ti particles on the crystal structure. Finally, we reveal the origin of the Sn-doping-induced anatase-to-rutile phase transformation on the basis of first-principles calculations.
To investigate the microstructural evolution of the TiO2 films, TiO2 films with thicknesses of 50, 100, 200 and 500 nm were deposited by rf magnetron sputtering using a 3-in. diameter Ti metal target (99.99 %, Furu-uchi Kagaku), where the sputtering gas was Ar and the total gas pressure was set to 3.0 Pa. Moreover, we changed the sputtering gas from Ar to Kr or Ne to study the effect of sputtered Ti particles on the crystal structure at different total gas pressures (0.5, 1.0, 2.0 and 3.0 Pa). The oxygen flow ratio (O2/(Ar + O2)) was maintained at 60 %. The distance between the target and the substrate was 55 mm.
The film thickness was measured using a surface profiler (Dektak3, Sloan Tech). X-ray diffraction (XRD, XRD-6000, Shimadzu) analysis was performed using Cu Kα1 radiation generated at 40 kV and 20 mA. Microstructural studies were performed using TEM (JEM-4010, JEOL).
Sn K-edge X-ray absorption fine structure (XAFS) spectra of the Sn-doped TiO2 thin films was measured at beamline BL07 of the SAGA Light Source  using convergent electron yield (CEY) mode. We also collected the spectra of 0.05-mm-thick Sn foil and SnO and SnO2 powders diluted with high-purity hexagonal BN powder as reference samples using transmission mode. All the measurements were conducted in air at room temperature.
In our previous study, we used X-ray absorption near edge structure (XANES) spectra to demonstrate that Sn doping induces the anatase-to-rutile phase transformation in sputtered TiO2 films. In the present study, we used first-principles calculations to investigate the geometrical structure of Sn doping and the transformation mechanism from anatase to rutile.
All the first-principles calculations were based on a plane-wave pseudopotential method using the CASTEP code [19, 20]. Vanderbilt ultrasoft pseudopotentials were employed, and the generalised gradient approximation (GGA-PBE)  was used as an exchange-correlation functional. After careful convergence tests with respect to the number of k-points and the plane-wave cutoff, a Monkhorst–Pack k-point grid with a special resolution of 0.5 nm−1 and a plane-wave cutoff energy of 380 eV was used for all calculations.
Results and Discussion
TEM Observation of Thin-Film Growth
On the basis of the aforementioned experimental observations, we concluded that the anatase and rutile phases coexist during the initial growth stage and that the anatase phase gradually dominates the crystal structure to form large crystallites with increasing film thickness. These films were deposited at a total gas pressure of 3.0 Pa, which is the preferential growth condition for anatase TiO2 films in the sputtering process, as shown in Fig. 7. In our previous report, we attributed the formation of anatase TiO2 films at a total gas pressure of 3.0 Pa to the suppressed bombardment effect from the high-energy particles due to gas scattering. However, such an explanation appears to be inconsistent with the present TEM observations that the rutile phase is still observed during the initial growth stage even under the anatase-preferential growth conditions because, in principle, the bombardment from the high-energy particles should homogeneously affect the film in the direction of the film thickness. Thus, the bombardment effect may not be a main factor for the formation of rutile phase during the sputtering process (we subsequently confirmed the weak influence of the bombardment effect by off-axis sputtering, as discussed in the following section). Such a growth behaviour can likely be attributed to the strong thermodynamic driving force towards the I41/amd phase, which results in smaller critical radii of the crystal nuclei and accelerates nucleation over the growth process .
Because the rutile phase is more dense (4.250 g · cm−3) than the anatase phase (3.894 g · cm−3), the rutile phase is considered to be stable at both high temperatures and high pressures [4, 25]. The bombardment effect of high-energy particles during the sputtering process has been reported to lead to a dense film . Thus, the bombardment effect is considered a possible reason for the production of rutile phase. In this study, off-axis sputtering was used to confirm this possible mechanism. In general, negative oxygen ions are considered to be the origin of the bombardment effect because of their high energy (approximately several hundred electron volts) . In the sputtering process, negative oxygen ions are accelerated by the cathode sheath and move towards the substrate in a straight path; thus, the bombardment effect from the high-energy particles should be observed in the facing position, as shown in Fig. 1. Consequently, more rutile phase is expected to be produced in the facing position.
Deposition rate (nm · min−1) of TiO2 films at different oxygen flow ratios
O2 flow ratio
Off-axis position 1
Off-axis position 2
Energy of Sputtered Ti particles
Mechanism of Sn-Doping-Induced Transformation from Anatase to Rutile
In addition to control of the sputtering conditions, impurity doping is an effective method of controlling the crystal structure of the sputtered TiO2 films. Li et al. attributed the room-temperature phase conversion of anatase to rutile TiO2 using Co or Ni to increased interaction between Co and Ni atoms forming a linear chain in the rutile phase . In our previous reports, we observed that impurity Sn doping also induced the phase transformation from the anatase to the rutile phase in sputtered TiO2 films . The XANES spectra suggest that Sn4+ ions are doped into Ti sites in TiO2 films. However, the geometrical doping structure of the Sn dopant and its effect on the growth of the crystal structure remain unknown.
First-shell EXAFS fitting results for the Sn–O direct bond in a Sn-doped TiO2 film and in a SnO2 reference sample. In these fittings, the coordination number was fixed at 6. Here, r is the phase-corrected atomic distance and σ 2 is the Debye–Waller disorder factor
σ 2 (Å2)
where E t is the total energy of the unit cell of TiO2 and SnO2 and the super supercell with/without the dopant. The total energies of TiO2 (I21/amd)  and SnO2 (P42/mnm)  were obtained after their crystal structures were optimised by the same computational methods previously described. The obtained formation energies are 0.66 and 0.25 eV for anatase and rutile TiO2, respectively. The formation energy for rutile TiO2 is lower than that for anatase TiO2 by 0.41 eV. The rutile TiO2 is more suitable for the Sn-doped TiO2 system. That is, Sn-doped TiO2 favours the rutile phase. This conclusion is consistent with that from our previous experimental report.
In this study, TiO2 films were prepared on unheated glass substrates using dc off-axis and rf magnetron sputtering methods, and the mechanism of the selective deposition of rutile and anatase TiO2 films during the sputtering process was investigated. TEM observations of the microstructural evolution of the TiO2 films showed the coexistence of rutile and anatase TiO2 phases in the initial stage, even in under anatase-preferential growth conditions; the anatase phase gradually dominated the crystal structural with increasing film thickness. These results suggest that the bombardment had no obvious effect on the TiO2 crystal structure during the sputtering process, which was also confirmed by off-axis magnetron sputtering experiments. Moreover, we studied the relationship between the kinetic energy of sputtered Ti particles and the crystal structure of TiO2 films and observed that the anatase TiO2 thin film was easily formed when the kinetic energy of sputtered Ti particles was less than 0.1 eV.
The mechanism of the effect of Sn impurity doping on the crystal structure was investigated by first-principles calculations. We observed that the formation of Sn-doped rutile TiO2 was lower than that of Sn-doped anatase TiO2, suggesting that Sn-doped TiO2 favours the rutile phase. These results offer a guideline for the selective deposition of rutile and anatase TiO2 thin films for industrial applications.
XAFS spectra measurements using synchrotron radiation were performed at beamlines BL07 and BL11 of SAGA-LS (Proposal No. 1504023F).
JJ and YS designed the experiments, and HY carried out the sputtering deposition experiments. TO performed the XAFS spectra measurements and the first-principle calculations. JJ worte the manuscript. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Kotake H, Jia J, Nakamura S, Okajia T, Shigesato T (2015) J Vac Sci Technol A 33:041505Google Scholar
- Bennett JM, Pelletier E, Albrand G, Borgogno JP, Lazarides B, Carniglia CK, Schmell RA, Allen TH, Tuttle-Hart T, Guenther KH, Saxer A (1989) Applied Optics 28:3303Google Scholar
- Furubayashi Y, Hitosugi T, Yamamoto Y, Inaba K, Kinoda G, Hirose Y, Shimada T, Hasegawa T (2005) Appl Phys Lett 86:252101Google Scholar
- Smith SJ, Stevens R, Liu S, Li G, Navrotsky A, Boerio-Goates J, Woodfield BF (2009) Am Mineralogist 94:236Google Scholar
- Naceur JB, Gaidi M, Bousbih F, Mechiakh R, Chtourou R (2012) Curr Appl Phys 12:422Google Scholar
- Jia J, Torigoshi Y, Shigesato Y (2013) Appl Phys Lett 103:013501Google Scholar
- Iida Y, Ozaki S (1961) J Am Ceram Soc 44:120Google Scholar
- Chao HE, Yun YU, Xingfanga HU, Larbot A (2003) J Eur Ceram Soc 23:1457Google Scholar
- Gole JL, Prokes SM, Glembocki OJ (2008) J Phys Chem C 112:1782Google Scholar
- Jr WA, Arenal F, de Mendonca VR, Ribeiro C, Longo E (2014) CrystEngComm 16:5021Google Scholar
- Okada K, Yamamoto N, Kameshima Y, Yasumori A, MacKenzie KJD (2001) J. Am Ceram Soc 84:1591Google Scholar
- Rimoldi L, Ambrosi C, Liberto GD, Presti LL, Ceotto M, Oliva C, Meroni D, Cappelli S, Cappelletti G, Soliveri G, Ardizzone S (2015) J Phys Chem C 119:24104Google Scholar
- Peng Y, Huang G, Huang W (2012) Adv Powder Technol 23:8Google Scholar
- Li S, Jena P (2009) Phys Rev B 79:201204Google Scholar
- Song PK, Shigesato Y, Kamei M, Yasui I (1999) Jpn J Appl Phys 38:2921Google Scholar
- Shigesato Y, Kamei M, Yasui I (1999) Jpn J Appl Phys 38:2921Google Scholar
- Jia J, Yoshimura A, Kagoya Y, Oka N, Shigesato Y (2014) Thin solid films 559:69Google Scholar
- Kawamoto M, Sumitani K, Okajima T (2010) AIP Conf Proc 1234:355Google Scholar
- Rez P, Bruley J, Brohan P, Payne M, Garvie LAJ (1995) Ultramicroscopy. 59:159Google Scholar
- Clark SJ, Segall MD, Pickard CJ, Hasnip PJ, Probert MIJ, Refson K, Payne MC, Krystallogr Z (2005). 220:567Google Scholar
- Perdew JP, Burker K, Erzerhof M (1996) Phys Rev Lett 77:3865Google Scholar
- Horn M, Schwerdtfeger CF, Meagher EP (1972) Zeitschrift fur Kristallographie 136: 273Google Scholar
- Baur WH, Khan AA (1971) Acta Crystallographica B 27:2133Google Scholar
- Jia J, Nakamura S, Shigesato Y, (2016) Phys Status Solidi A doi:10.1002/pssa.201532887
- Burdet JK, Hughbanks T, Miller GJ, Richardson JW, Smith JV (1987) J Am Chem Soc 109:3639Google Scholar
- Okimura K (2001) Surf Coatings Technol 135:286Google Scholar
- Behrisch R, Eckstein W (2007) Sputtering by particle bombardment: experiments and computer calculations from threshold to MeV energies. Springer Berlin Heidelberg, New York, p 4. ISBN: 978-3-540-44500-5Google Scholar
- Meyer K, Schuller IK, Faico CM (1981) J Appl Phys 52:5803Google Scholar
- Pannetier J, Denes G (1980) Acta Crystallographica B 36:2763Google Scholar
- Ravel R, Newville M. J (2005) Synchrotron Rad 12:537Google Scholar
- Grzeta B, Tkalcec E, Goebbert C, Iakeda M, Takahashi M, Nomura K, Jaksic M (2002) J Phys Chem Solids 63:765.Google Scholar