Accepted Manuscript Anatase phase evolution and its stabilization in ion beam sputtered TiO2 thin films Nalin Prashant Poddar, S.K. Mukherjee, Mukul Gupta PII: DOI: Reference: S0040-6090(18)30637-0 doi:10.1016/j.tsf.2018.09.038 TSF 36900 To appear in: Thin Solid Films Received date: Revised date: Accepted date: 31 December 2017 26 July 2018 17 September 2018 Please cite this article as: Nalin Prashant Poddar, S.K. Mukherjee, Mukul Gupta , Anatase phase evolution and its stabilization in ion beam sputtered TiO2 thin films. Tsf (2018), doi:10.1016/j.tsf.2018.09.038 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Anatase phase evolution and its stabilization in ion beam sputtered TiO2 thin films Nalin Prashant Poddar1, S. K. Mukherjee1*, Mukul Gupta2 1 Department of Physics, Birla Institute of Technology, Mesra, Ranchi, Jharkhand PT 835215, India RI nalinppoddar@gmail.com; sanat_aphy@yahoo.co.in +91-9431363464* UGC-DAE Consortium for Scientific Research, University Campus, Khandwa Road, SC 2 NU Indore 452 001, India MA dr.mukul.gupta@gmail.com EP T Abstract ED Corresponding author - S. K. Mukherjee Thin films of titanium oxide (TiO2) were prepared by ion beam sputtering at AC C room temperature under various oxygen partial pressure and annealed at 350 ºC and higher. Complete target oxidation is observed at O2/Ar pressures much lower than the conventional sputtering. The films are analyzed using X-ray, Raman spectroscopy, optical transmittance, and soft X-ray absorption spectroscopy. As-deposited thin films are all amorphous. Their Ti coordination number is less than that of crystalline phases, anatase and rutile. Upon post-heating above 350 ºC in vacuum, films prepared with a ratio O2/Ar between 0.25 and 0.66 remain amorphous. Films prepared below this range, develop Ti6O11 and Ti3O5 grains, those prepared above this range become anatase. The band gap of the films varied between 3.12 and 3.36 eV, their refractive 1 ACCEPTED MANUSCRIPT index between 2.07 and 2.81. Dielectric modeling of the transmittance spectra shows a broad Gaussian distribution of resonance frequency oscillators suggesting the asdeposited films to be disordered. Key Words: - Thin films, TiO2, ion beam sputtering, XRR, XRD, Raman RI PT spectroscopy 1. Introduction SC Titanium dioxide (TiO2) is one of the most widely used metal oxides with NU wide band gap and has been extensively studied in recent years. High transmittance of thin films make them ideal for optical coatings [1]. Such coatings are widely used in MA mirror technology [2-3]. In addition, TiO2 has a large dielectric constant, which is useful in various electronic applications. It can be used as gas-sensing agents, ED photocatalyst and as an active component in dye-sensitized solar cells [4]. In bulk form, titanium dioxide is known to exist in three crystalline structures: two tetragonal EP T phases, anatase (a = 3.785 Å, c = 9.514 Å) and rutile (a = 4.593 Å, c = 2.959 Å); and an orthorhombic phase, brookite (a = 5.456 Å, b = 9.182 Å, c = 5.143 Å) [5]. Only AC C anatase and rutile structures are commonly observed in thin film form. Anatase is a low temperature phase (< 873 K) of TiO2 [6]. TiO2 thin films can be prepared by a variety of methods such as ion beam sputtering (IBS) [2-3, 7-10], rf/dc magnetron sputtering [11-13], sol–gel method [14], pulsed laser deposition [15], chemical vapor deposition [16], e-beam evaporation [17-18], atomic layer deposition [19], spray pyrolysis [20] and Supersonic Cluster Beam Deposition (SCBD) [21]. 2 ACCEPTED MANUSCRIPT Ion beam sputtering has advantages over others deposition techniques as there a separate ion source is used to produce focused ion beam, which is responsible for the sputtering of the target. Therefore, it provides a better control over the energy and current density of ion beam. The discharge is confined within the cross-section of the ion beam, which reduces the gas consumption during deposition [22]. The density of PT high energy ion beam sputtered films can be close to that of the bulk material. The ion RI beam technique is commonly used to produce amorphous TiO2 optical thin film coatings [7, 12]. In addition, full target utilization makes IBS a cost effective method SC as compared to that of e.g. magnetron sputtering. NU Process parameters have a considerable impact on film properties, especially while sputtering metal targets and in the presence of reactive gases. Effect of these MA parameters on the properties of TiO2 films has been investigated, particularly using conventional magnetron [5-6, 13, 23-26] and ion beam sputtering [2-3, 7-10]. Most ED of these investigations are on amorphous TiO2 films deposited on unheated substrates for optical coatings. Some groups have also explored the effect of substrate heating EP T [9, 27], post-deposition annealing [28-29] and reactive gas (O2) on structural and optical properties of ion beam sputtered TiO2 films [8, 27, 30]. However, the growth AC C of the anatase phase in ion beam sputtered TiO2 films is still not well documented. The aim of this work is to explore anatase phase evolution in TiO2 thin films deposited with ion beam sputtering. The content of various phase of TiO2 films strongly depends on the oxygen content during film formation [4, 27]. Thus, this work focuses on optimization of the oxygen content for the evolution of the anatase phase in ion beam sputtered TiO2 films. The effect of reactive gas on the growth of TiO2 thin films in terms of structural and optical properties is investigated. 3 ACCEPTED MANUSCRIPT 2. Experimental details Titanium dioxide films were deposited using an in-house built reactive ion beam sputtering (IBS) system [22] with 10×10 cm square shaped, 1 mm thick titanium target of 99.99 % purity. The ion gun used was a radio frequency (rf) broad beam Kaufman-type hot cathode gun. Oxygen (99.99 %) and argon (99.99 %) gases PT were used in the ion gun for sputtering. The vacuum chamber was first evacuated RI down to 2×10-5 Pa with a turbo-molecular pump. The chamber was then flushed with argon (Ar) gas to remove any contamination of other gases. Subsequently, oxygen SC (O2) gas was also introduced so as to have overall total pressure of argon and oxygen NU of 2×10-3 Pa during deposition. The ratio of O2 and Ar gases were varied for the deposition of different samples while keeping the working pressure and total input gas MA constant (5 sccm). The titanium (Ti) target was bombarded with an ion beam containing a ED mixture of argon and oxygen ions, neutralized by nearly a similar number of electrons. The beam current and voltage during the thin film deposition were 40 mA ; EP T and 1000 V, respectively. The ion beam was accelerated with the accelerator running with a current of 2 mA and a voltage of 300 V. The system was connected with a rf AC C neutralizer having an emission current of 100 mA and a power of 30 W. The rf source was operated in a power range of 80-100 W during deposition. During the film formation process, the oxygen gas reacts with titanium atoms and forms oxides of titanium which are deposited onto various substrates. In the same run, depositions onto polished silicon (111) (2×2 cm), a quartz plate (2×2×0.2 cm) and corning glass substrates (2.5×7×0.2 cm) were done. The stationary substrate holder was kept at room temperature (300 K) during deposition. Before each deposition, a pre-sputter cleaning of the Ti target was done for 5 min in the presence 4 ACCEPTED MANUSCRIPT of 5 sccm Ar gas. The TiO2 thin films were grown without any external substrate heating. Post-deposition annealing was performed in a vacuum (~2 Pa) furnace for one hour at 350-650 ºC. X-ray reflectivity (XRR) measurements were carried out with a Bruker D8 Discover diffractometer by using Cu Kα (0.154 nm) radiation. The scanning was done PT from 0.01º to 0.30º. X-ray diffraction (XRD) analysis of deposited films was carried RI out with a Bruker D8 Advance X-ray diffractometer by using Cu Kα (0.154 nm) radiation. The scanning was done from 20º to 80º in 2θ steps of 0.04º and a scanning SC rate of 5 s/step. NU The transmittance spectra of as-deposited TiO2 films were determined within the range 200-1000 nm using a UV-Vis spectrometer (Perkin-Elmer, model Lamda- MA 20). Theoretical transmittance spectra were generated using dielectric modelling and fitted to the experimentally obtained curves with the help of the SCOUT [40] optical ED simulation program. Raman active modes in the annealed film were investigated by a Reinshaw EP T Invia Raman spectrometer using a 514.6 nm Ar laser for a scan time of 20 seconds in the range of 100-800 cm-1. AC C Soft X-ray absorption spectroscopy (SXAS) measurements were done using the INDUS-2 beamline (BL-01) in the synchrotron facility at Raja Ramanna Centre for Advanced Technology, Indore, India. The Ti L-edge absorption spectra were recorded to identify the Ti-coordination in the films and to estimate its population. 3. Results and discussion 3.1. Deposition rate 5 ACCEPTED MANUSCRIPT The deposition rate of samples deposited at different gas ratios were deduced from the sample thickness that was obtained from X-ray reflectivity (XRR) measurements. The XRR data of films deposited at different O2/Ar gas ratio is shown in Fig. 1. From the periodicity of the signal the thickness of the films can be deduced. The experimental data were fitted using Parratt32 software (version 1.6, Informer PT Technologies, Inc.), taking standard density of TiO2 (4.23 g/cm3) (Scattering Length RI Density is Re = 3.410-06 Å-2, Im = 1.7410-06 Å-2) (see Fig. 1). The oxygen flow rate SC determine the deposition rate of thin films. In our experiment we assure that films of similar thickness (about 200 nm) are deposited at different gas ratios. NU Fig. 2 represents the deposition rate as a function of the O2 concentration of the amorphous during deposition. The deposition rate is observed to decrease with MA increase in the O2/Ar ratio. This deposition rate variation is similar to that proposed by J. Heller et al. [31]. They observed that above a critical oxygen partial pressure the ED target oxidation decreases the sputtering rate to a constant value. This leads to a decrease in deposition rate with an increase in O2 concentration. In the present case EP T also, the deposition rate is decreasing with the O2 content. From Fig. 2 it can be concluded that even a small amount of oxygen is enough to lead to complete Ti target AC C oxidation (target poisoning). The deposition rate is decided by two factors – the sputtering rate of the target and the rate at which the oxide layer forms on the target due to reaction with the oxygen present in the deposition chamber. Contrary to conventional reactive sputtering, the sputtering area in the ion beam sputtering is localized and hence, the sputter rates are more sensitive to target oxidation. With increasing oxygen flow the target poisoning rate dominates leading to a decrease in the deposition rate. In region I the deposition rate varies while in region II the rate stabilizes (fig. 2). It has been 6 ACCEPTED MANUSCRIPT observed that in ion beam sputtering the Ti target oxides completely beyond an O2/Ar ratio of 0.4. The observed deposition rate variation is similar to H. Demiryont et al. [30] (see Fig. 2). They deposited TiO2 using ion beam sputtering and varied the oxygen percentage in the amorphous from 5-50 %. Although the trend is similar, their PT deposition rate decreases faster with the increase in O2 and attains a constant value at RI lower oxygen flow. Also, the deposition rate is slightly lower as compared to the present case. H. Ohsaki et al. [32] also observed a similar trend of the deposition rate SC with respect to the O2 concentration. Similar trends were verified by L. F. Donaghey NU et al. [23] when depositing TiOx films using rf reactive sputtering. Our results at low O2/Ar values do not match with Jim Cherng Hsu et al. [10]. MA They observed an increase in sputter rate in a single ion beam sputtering of Ti at very low oxygen partial pressures (<5.310-4 Pa). Such an increase is explained to be due EP T ED to the formation of sub-oxides loosely bound to the target surface [10]. 3.2. X-ray diffraction studies Fig. 3 represents the XRD pattern of all thin films deposited at different O2/Ar AC C ratio and annealed at 350 ºC. The as-deposited films are XRD amorphous (not shown in the figure) except for Ti thin films (O2/Ar = 0). This could be due to the low energy and low mobility of particles impinging on the substrate at room temperature. XRD peaks of Ti at 34.9º (100) and 37.8º (002) were observed in as-deposited film [33-34]. Mixed phase of Ti6O11 (130) and Ti3O5 (204) were obtained for a O2/Ar ratio of 0.11. Further oxygen addition leads to other phases. Anatase with (101) orientation [27] evolves when the O2/Ar ratio increase beyond 0.66. In Fig. 3, for O2/Ar = 0.66, 1.00 and 1.50 the anatase phase of TiO2 is observed. TiO2 films 7 ACCEPTED MANUSCRIPT preferentially oriented along (101). All observed XRD peaks were well matched with JCPDS:84-1285. In the literature, it was observed that as-deposited TiO2 films are xray amorphous and clear peaks of crystalline phases can only be seen after annealing at 300-350 ºC [5]. We applied different annealing temperature TiO2 films (deposited under O2/Ar PT = 1.50) to investigate crystalline growth with annealing. Fig. 4 represents XRD of RI annealed TiO2 (O2/Ar = 1.50) films at different temperature (350-650 ºC). The average crystallite size of annealed TiO2 films is estimated using Debye K  cos  NU Dav  SC Scherer’s equation given by, (1) MA where Dav is the average crystallite size, K is constant (0.9 for spherical grains), λ is the x-ray wavelength (1.541 nm), β is the full width half maximum (FWHM) of ED strongest intensity peak. No significant variation in crystallite size with annealing temperature is observed. All crystalline films exhibit a grain size of about 60 nm. This EP T suggests that the film crystallization occurs by transforming the amorphous phase into separate crystallite and not by coalescence of smaller crystallites into bigger ones. AC C Such a phenomenon is possible only when the initially existing crystallites are widely separated from one another. 3.3. Raman Spectroscopy Fig. 5 depicts the Raman spectra of TiO2 thin films prepared under different O2/Ar ratio. The as-deposited samples show no Raman peaks (plot not shown) and thus, lack lattice ordering. Raman active modes are observed only after annealing at 350 ºC where the films are crystalline according to Fig. 4. [5]. According to the factor group analysis, TiO2 has 15 optical modes, 1A1g  1A2u  2B1g  1B2u  3Eg  2Eu. 8 ACCEPTED MANUSCRIPT The modes, A1g, B1g and Eg are Raman active and A2u and Eu are infrared active. The B2u mode is inactive in Raman and infrared spectra [35]. Films deposited at O2/Ar ratios ranging between 0.42-1.50 show well defined Raman peaks at around 142.4 cm1 (Eg), 394.5 cm-1 (B1g), 515.8 cm-1 (A1g) and 637.6 cm-1 (Eg) [36-40]. A shift in peak position is observed compared to the reported Raman peaks of bulk crystalline TiO2. PT Table-I showed the variation in the peak position of the present work and of reported RI Raman peaks. No rutile peaks are observed. The peak at 142.4 cm-1 (Eg) has the maximum intensity. This peak corresponds to the symmetric stretching vibrations of SC O-Ti-O along the a-axis. The B1g mode is caused by the symmetric bending vibration NU of O-Ti-O, and the A1g mode is caused by the anti-symmetric bending vibration of OTi-O [41]. All peaks are symmetric. No significant variation in peak intensity or peak MA broadening is observed with the variation of the gas ratio; the FWHM always at 11.20 ± 0.53. This confirms that the stretching vibrations are unhindered along the a-axis ED and are the same within the above specified gas ratio range. EP T Films deposited at lower O2/Ar ratio viz. 0.25 and 0.11 are showing broad yet detectable humps. The peak positions are not very clear and could only be guessed upon fitting with Gaussian curves. These Raman plots along with their fitting are AC C shown in Fig. 6. Humps observed within 443 cm-1 and 615 cm-1 are at the position of the Eg and A1g mode respectively [35,42]. The humps are too broad to be assigned to crystallite of rutile. They rather indicate a rutilelike neighbourhood in the amorphous phase. A few extra peaks broad humps at 183, 203, 260 and 294 cm-1 are also realized upon fitting. They may correspond to mixed titanium oxide phases [35, 42]. These modes do not match with those of Ti2O3 confirming that the unidentified phases are not Ti2O3 [43]. Some other phases of titanium oxide are identified in the XRD data of the film deposited at a O2/Ar gas ratio of 0.11. Due to the lack of standard data, the 9 ACCEPTED MANUSCRIPT observed unidentified Raman shifts cannot be related to these phases. Such phases may appear due to the scarcity of oxygen. From the above data (Table-I), it can be inferred that the O2/Ar ratio affects TiO2 nucleation and seed layer growth of separate phases. The present annealing temperature (350 °C) cannot lead to phase transformation to rutile [3] [5]. Only PT existing nuclei in the film can grow either by coalescence or at the sacrifice of the RI amorphous phase. Thus, the above results support that for deposition at low oxygen pressure nucleation and formation of nuclei in the film is non-uniform resulting in SC mixed phase phases. At high O2/Ar gas ratio (0.42) the nuclei phases are commonly NU anatase. MA 3.4. Optical studies Transmittance measurements using UV-Vis spectroscopy of the TiO2 thin ED films are shown in Fig. 7. Samples prepared at O2/Ar = 0 and 0.11 were opaque (metallic) and dark respectively, probably due to strong absorption by the free charge EP T carriers in metallic Ti and enhanced scattering at vacant oxygen sites. For all other films, transmittance increases from 60 to 90 % with an increase in the O2/Ar ratio. AC C The transmittance values of TiO2 films with anatase phase, developed with an O2/Ar ratio ranging between 0.66-1.50, are similar to the sputtered films developed by YaQi Hou et al. [24]. The relatively high transmittance of the films at a higher O2/Ar ratio indicates low scattering suggesting low surface roughness and good homogeneity of the deposited films. The strong absorption in the UV regime ( < 400 nm) can be assigned to electronic inter-band transitions. The experimental transmittance data were fitted theoretically using dielectric modeling [40]. The transmittance of one sample (O2/Ar = 1.50) together with the 10 ACCEPTED MANUSCRIPT fitting curves and the imaginary part of the dielectric function is shown in Fig. 8. The dielectric function of our samples,𝜀() = 𝜀1 + 𝑖𝜀2 , where 𝜀1 and 𝜀2 are the real and the imaginary parts of the dielectric function, is derived using a dielectric background, an O’Leary-Johnson-Lim (OJL) model [44] and a Brendel oscillator [45]. A dielectric background signifies a constant and real contribution to the dielectric function. OJL PT model simulates the optical transitions from the valence band to the conduction band RI near the band edge. It was originally proposed to model the strong electronic interband transitions of amorphous Si [44]. It contains four free parameters: band gap SC energy, the overall strength of the transition, the exponent gamma describing the NU decay of tail states into the band gap and a decay parameter which reduces the imaginary part of the dielectric function to zero at high frequencies. In addition to the MA OJL function, a damped harmonic oscillator is often used to represent the inter-band transitions to conduction levels well above the band edge [44]. ED However, in our case, we didn’t get a good fit using one damped harmonic oscillator. Instead, a reliable fit is observed (see Fig. 8) using a Gaussian distribution EP T of damped harmonic oscillators often termed as Brendel oscillator. A Brendel oscillator uses another parameter, distribution width, in addition to the common fitting AC C parameters of a damped harmonic oscillator: resonance frequency, oscillator strength, and damping. Brendel et al. [45] convoluted a Gaussian function with the damped harmonic oscillator model (Drude model) accounting for a statistical distribution of a resonance frequency of a vibrational mode in amorphous solids. Although such a Gaussian distribution of harmonic oscillators was originally proposed to model the infra-red spectra of amorphous solids, like, silicon, silicon nitride and aluminium oxide films, this function is quite effective as an inter-band dielectric function model in simulating the UV-Vis spectra of porous Si [46]. 11 ACCEPTED MANUSCRIPT The thicknesses of as-deposited films obtained from the theoretical model match well with those estimated from XRR. The proposed model (OJL function and a Brendel oscillator) suggests that the films may not have a periodic array of TiO6 octahedra. Their local environment is disordered probably due to the evolution of a random phase in an amorphous film. This may lead to variations in bond angles and PT charge distributions and hence, no definite resonance frequency can be realized. RI The width of the Brendel oscillator can be used to represent the amount of disorderness in the films qualitatively. In our case, we have not observed any definite SC trend in the variation of the width of the Brendel oscillator with the O2/Ar ratio. The NU width of the Brendel oscillator in our films averages around 1801 ± 175 cm-1. The band gaps (Eg) of the films were determined by the theoretical fit and MA plotted as a function of the O2/Ar ratio (see Fig. 9). It is observed that the band gap increases nonlinearly with the increase in the O2/Ar gas ratio. Irrespective of the ED occurrence of phases, the band gap of bulk TiO2 ranges between 3.1-3.3 eV [1]. The band gap of nanostructured TiO2 varies from 3.59-3.68 eV with different processing EP T conditions [47]. Optical band gaps in the range 3.41-3.44 eV with increase in sputtering pressure have been reported earlier by T.M. Wang et al. [48]. Preetam AC C Singh et al. [25] also reported variation in band gap from 3.20-3.28 eV with sputtering pressure and dc power. We obtained band gap ranging between 3.12-3.36 eV. They are consistent with other reported band gap values [25, 48] but are less than those reported by P. B. Nair et al. [49] (3.58-3.75 eV) for TiO2 films prepared by rf magnetron sputtering. The refractive index (n) is determined from the theoretical fit as a square root of the dielectric function. The variation of refractive index of the films with O2/Ar ratio is shown in the Fig. 9. For films deposited with O2/Ar  0.25, the n – values 12 ACCEPTED MANUSCRIPT range between 2.07-2.81. However, a very low n – value (1.74) is estimated from the theoretical fit of the film deposited at O2/Ar = 0.11. This film is dark and has almost zero transmittance (Fig. 7). The XRD plot of this film shows peaks at 2 values other than regular TiO2 positions (see Fig. 3). This confirms that this film is structurally different from the rest of the films (see Fig. 9). Thus, the proposed theoretical model PT is not appropriate for this film and the estimated refractive index value is not reliable. RI The n – values of our films (2.07-2.81) are comparable to the theoretically fitted SC values reported by D. Mergel et al. [50]. That study was based on films of variable mass densities modeled as films with dispersed crystalline phases in an amorphous NU TiO2 matrix (Bruggeman’s effective medium approximation). A similar assumption has been made in our case but instead of considering two phases (crystalline and MA amorphous) of different bulk densities and refractive indices, we have considered distortions in the modeled harmonic oscillators, probably due to local disorder in bond ED length, bond angle and bond strength. Such an assumption is more suitable (in the present case) than a clear separation between crystalline and amorphous phase. EP T The refractive indices of different phases of TiO2 (amorphous, rutile and anatase) in films prepared using different physical deposition technique are also listed AC C in Table-II for comparison. The n-values of our TiO2 films (except deposited at O2/Ar = 0.11) are comparable to the experimentally determined values of most of the research groups. M. Laube et al. [52], D. Mergel et al. [53], E. Khawaja et al. [54] and S.Y. Kim et al. [55] deposited TiO2 using evaporation and reported refractive indices between 1.9-2.51 at 550 nm. M G Krishna et al. [17] and M. Jerman et al. [18] used e-beam evaporation and obtained the refractive indices between 2.3-2.5 at 550 nm. TiO2 films prepared using conventional sputtering show refractive indices ranging from 2.05-2.73 at 550 nm [5, 24, 26, 32, 52, 56-58]. S. M. Rossnagel et al. [8] 13 ACCEPTED MANUSCRIPT used ion beam sputtering and reported refractive index 2.48 of TiO2 at 630 nm. H. Demiryont et al. [30] also reported refractive index 1.98-2.52 at 630 nm for TiO2 films deposited using ion beam sputtering. Some researchers also reported very low refractive index (1.3-1.8) of TiO2 films [59-60]. The refractive index (1.74) of our film deposited at O2/Ar = 0.11 lie within this range. J. Q. Xi et al. [60] reported n- PT values between 2.7-1.3 for their films made up of TiO2 nanorods. These films were RI obtained by oblique – angle deposition using electron beam evaporation. G. K. Mor et al. [61] also reported an average refractive index of 1.66 for their TiO2 nanotube array NU SC films. Such low n – values were attributed to the nano-porous nature of the films. 3.5. SXAS Studies MA Fig. 10 represents the Soft X-ray absorption spectroscopy plots of the asdeposited TiO2 thin films at the Ti-edge. The Ti 2p X-ray absorption spectra consist ED of L3 and L2 - absorption edges corresponding to 2p3/2 and 2p1/2 excitations of 3𝑑𝑜 to 2𝑝5 3𝑑1 transition [62]. EP T The L3 edge is at lower energy and has two peaks, a (464.9 ± 0.14) and b (466.7 ± 0.11). The L2 edge is at higher energy and also has two peaks, c (470.5 ± AC C 0.14) and d (472 ± 0.10). The splitting of each edge is due to the t2g and eg symmetry of the d – orbital. This phenomena is called crystal field splitting [63]. In anatase, rutile and Ti2O3 phases, the peak b of L3 edge is a convolution of two peaks. In rutile, the low intensity peak, often termed as shoulder, is on the lower energy side while in anatase, it is on the higher energy side. In Ti2O3, both peaks are of similar intensity [62, 64-65]. The coordination number of Ti in rutile and anatase is 6 (Ti6) and in Ti2O3 is 5 (Ti5). The nature of the peak b of the SXAS plot is, thus, often used to indicate the coordination environment of Ti in a sample. If there is no splitting in 14 ACCEPTED MANUSCRIPT peak b then it indicates a coordination environment, Ti4 [62, 66]. In order to estimate the Ti coordination environment in as-deposited TiO2 films, their SXAS plots were fitted using Gaussian curves. Fig. 11 shows the fitting of the SXAS data of a TiO 2 film deposited at O2/Ar ratio of 1.50. The peaks of L3 edge (a, b) and L2 edge (c, d) fits well with single Gaussian profiles. However, we have observed a pre-edge peak PT (e) in almost all of our samples. Such pre-edge peaks are due to multiplet core hole-d RI electron interactions [67]. Since, the peak b in the SXAS spectra of the as-deposited film can be fitted well with one Gaussian profile, we conclude that our as-deposited SC TiO2 films have a coordination environment less than that for rutile, anatase or Ti2O3. NU These films may contain sub-oxides of Ti. In the present case, these compositions are not known but may be estimated quantitatively using X-ray photoelectron MA spectroscopy. It is worth mentioning here that a similar splitting is expected for peak d of the ED L2 edge due to crystal field splitting. However, such a splitting of this high-energy EP T peak is not evident because of vibrational and dispersion broadening of the peak. 4. Conclusion AC C In this work, the occurrence of phases in reactive ion beam sputtered TiO2 thin films has been investigated. Films were developed with variable O2/Ar ratio and annealed at 350 C. The ion beam sputtering is different from conventional sputtering since the deposition rate starts varying at much low oxygen content in the chamber. The variation could clearly separate films deposited with the partially oxidized target and with fully oxidized targets. The oxygen content has a major impact on the structure and optical properties of the reactive ion beam sputtered films. 15 ACCEPTED MANUSCRIPT From the present study, we can conclude that TiO2 films deposited using ion beam sputtering are amorphous irrespective of the variations in O2/Ar ratio during deposition. Titanium L-edge soft X-ray absorption spectroscopy of as-deposited films confirmed that the Ti coordination environment in the films is less than that of rutile and anatase phases. Crystalline phases, particularly anatase, can evolve after PT annealing the samples at higher temperatures ( 350 ºC) but only in films deposited RI with an O2/Ar ratio higher than  0.66. At low O2/Ar values ( 0.11), the annealed SC films seem to be a mixture of Ti6O11 and Ti3O5. With moderate values (0.25-0.42) films are amorphous. Variation in annealing temperature (450-650 ºC) has no NU influence on average crystallite size (60 nm) of the films. This rules out the possibility of coalescence of crystallites in the films and supports the model of dispersed and MA separated crystallites in an amorphous matrix. Raman spectra confirmed the presence of anatase phase in the annealed films. Raman measurements also indicate rutile like ED neighborhoods in the amorphous phase for films deposited at the low O2/Ar value (0.11-0.25). However, XRD has no strong evidence of rutile phase in such films. As- EP T deposited films showed high transmittance (~90%) in the visible range, still increasing with the increase in the O2/Ar ratio. AC C Transmittance spectra of the films are theoretically fitted using dielectric modeling. The dielectric function includes an O’Leary-Johnson-Lim (OJL) function for band gap transitions and a Brendel oscillator for other inter-band transitions. The derived band gap values of the films vary between 3.12 and 3.36 eV and the refractive indices lie between 2.07 and 2.81. Acknowledgements 16 ACCEPTED MANUSCRIPT The authors wish to express their sincere thanks to UGC DAE Consortium for Scientific Research, Indore for help in thin films deposition using ion beam sputtering and for the XRR and XRD measurements of the deposited films. The authors would like to acknowledge the Raja Ramanna Centre for Advanced Technology (RRCAT) for SXAS measurements. The authors also want to acknowledge the cooperation of PT the Central Instrumental Facility (CIF) Birla Institute of Technology, Ranchi for SC RI characterization work. [1] NU References X. Chen and S. S. Mao, Titanium Dioxide Nanomaterials: Synthesis, MA Properties, Modifications, and Applications, Chem. Rev. 107 (7) (2007) 28912959. [2] M. Magnozzi, S. Terreni, L. Anghinolfi, S. Uttiya, M. Carnasciali, G. Gemme, ED M. Neri, M. Principe, I. Pinto, L.C. Kuo, S. Chao and M. Canepa, Optical properties of amorphous SiO2-TiO2 multi-nanolayered coatings for 1064-nm [3] EP T mirror technology, Opt. Mater., 75 (2018) 94-101. H. W. Pan, S. J. Wang, L. C. Kuo, S. Chao, M. Principe, I. M. Pinto and R. DeSalvo, Thickness-dependent crystallization on thermal anneal for AC C titania/silica nm-layer composites deposited by ion beam sputter method, Opt. Express, 22 (24) (2014) 29847-29854. [4] U. Diebold, The surface science of titanium dioxide, Surf. Sci. Rep. 48 (2003) 53-229. [5] M. H. Suhail, G. M. Rao and S. Mohan, dc reactive magnetron sputtering of titanium-structural and optical characterization of TiO2 films, J. Appl. Phys. 71 (3) (1992) 1421-1427. [6] M. D. Wiggins, M. C. Nelson and C. R. Aita, Phase development in sputter deposited titanium dioxide, J. Vac. Sci. Technol., A 14, (3) (1996) 772-776. [7] C. Bundesmann, T. Lautenschläger, D. Spemann, A. Finzel, E. Thelander, M. Mensing and F. Frost, Systematic investigation of the properties of TiO2 films 17 ACCEPTED MANUSCRIPT grown by reactive ion beam sputter deposition, Appl. Surf. Sci., 421, (2017) 331-340. [8] S. M. Rossnagel and J. R. Sites, X-ray photoelectron spectroscopy of ion beam sputter deposited SiO2, TiO2, and Ta2O5, J. Vac. Sci. Technol., A 2 (2) (1984) 376-379. [9] P. A. M. Hotsenpiller, G. A. Wilson, A. Roshko, J. B. Rothman and G. S. Rohrer, Heteroepitaxial growth of TiO2 films by ion-beam sputter deposition, [10] PT J. Cryst. Growth 66 (1) (1996) 779-785. J. C. Hsu and C. C. Lee, Single-and dual-ion-beam sputter deposition of C. H. Heo, S. B. Lee and J. H. Boo, Deposition of TiO2 thin films using RF SC [11] RI titanium oxide films, Appl. Opt. 37 (7) (1998) 1171-1176. magnetron sputtering method and study of their surface characteristics, Thin [12] NU Solid Films 474 (1) (2005) 183-188. J. M. Bennett, E. Pelletier, G. Albrand, J. P. Borgogno, B. Lazarides, C. K. Carniglia, R. A. Schmell, T. H. Allen, T. Tuttle-Hart, K. H. Guenther and A. MA Saxer, Comparison of the properties of titanium dioxide films prepared by various techniques, Appl. Opt. 28 (16) (1989) 3303-3317. B. R. Weinberger and R. B. Garber, Titanium dioxide photocatalysts produced ED [13] by reactive magnetron sputtering, Appl. Phys. Lett. 66 (18) (1995) 2409-2411. [14] J. Yu, X. Zhao, J. Du and W. Chen, Preparation, microstructure and EP T photocatalytic activity of the porous TiO2 anatase coating by sol-gel processing, J. Sol-Gel Sci. Technol. 17 (2) (2000) 163-171. [15] J. H. Kim, S. Lee and H. S. Im, The effect of target density and its AC C morphology on TiO2 thin films grown on Si (100) by PLD, Appl. Surf. Sci. 151 (1) (1999) 6-16. [16] K. L. Siefering and G. L. Griffin, Growth kinetics of CVD TiO2: influence of carrier gas, J. Electrochem. Soc. 137 (4) (1990) 1206-1208. [17] M. G. Krishna, S. Kanakaraju and S. Mohan, Structure and composition related properties of titania thin films, Vacuum 46 (1) (1995) 33-36. [18] M. Jerman and D. Mergel, Structural investigation of thin TiO2 films prepared by evaporation and post-heating, Thin Solid Films 515 (17) (2007) 6904-6908. [19] J. Aarik, A. Aidla, A. A. Kiisler, T. Uustare, V. Sammelselg, Effect of crystal structure on optical properties of TiO2 films grown by atomic layer deposition, Thin Solid Films 305 (1-2) (1997) 270-273. 18 ACCEPTED MANUSCRIPT [20] C. Natarajan, N. Fukunaga and G. Nogami, Titanium dioxide thin film deposited by spray pyrolysis of aqueous solution, Thin Solid Films 322 (1) (1998) 6-8. [21] C. Toccafondi, S. Uttiya, O. Cavalleri, G. Gemme, E. Barborini, F. Bisio and M. Canepa, Optical properties of nanogranular and highly porous TiO2 thin films, J. Phys. D, 47 (48) (2014) 485301. [22] M. Gupta, A. Gupta, D. M. Phase, S. M. Chaudhari and B. A. Dasannacharya, PT Development of an ion-beam sputtering system for depositing thin films and multilayers of alloys and compounds, Appl. Surf. Sci. 205 (1) (2003) 309-322. L. F. Donaghey and K. G. Geraghty, Effect of target oxidation on reactive RI [23] SC sputtering rates of titanium in argon-oxygen plasmas, Thin Solid Films 38 (3) (1976) 271-280. Y. Q. Hou, D. M. Zhuang, G. Zhang, M. Zhao and M. S. Wu, Influence of NU [24] annealing temperature on the properties of titanium oxide thin film, Appl. Surf. Sci. 218 (1) (2003) 98-106. P. Singh and D. Kaur, Room temperature growth of nanocrystalline anatase MA [25] TiO2 thin films by dc magnetron sputtering, Physica B Condens. Matter 405 [26] ED (5) (2010) 1258-1266. P. Löbl, M. Huppertz, D. Mergel, Nucleation and growth in TiO2 films 79. [27] EP T prepared by sputtering and evaporation, Thin Solid Films 251 (1) (1994) 72M. C. Marchi, S. A. Bilmes, C. T. M. Ribeiro, E. A. Ochoa, M. Kleinke and F. Alvarez, A comprehensive study of the influence of the stoichiometry on the AC C physical properties of TiOx films prepared by ion beam deposition, J. Appl. Phys. 108 (6) (2010) 064912. [28] W.H. Wang and S. Chao, Annealing effect on ion-beam-sputtered titanium dioxide film, Opt. Lett., 23 (18) (1998) 1417-1419. [29] S. Chao, W. H. Wang, M.-Y. Hsu and L. C. Wang, Characteristics of ionbeam-sputtered high-refractive-index TiO2-SiO2 mixed films, J. Opt. Soc. Am. A, 16 (6) (1999) 1477-1483. [30] H. Demiryont and J. R. Sites, Effects of oxygen in ion-beam sputter deposition of titanium oxides, J. Vac. Sci. Technol., A 2 (4) (1984) 1457-1460. [31] J. Heller, Reactive sputtering of metals in oxidizing atmospheres, Thin Solid Films 17 (2) (1973) 163-176. 19 ACCEPTED MANUSCRIPT [32] H. Ohsaki, Y. Tachibana, A. Hayashi, A. Mitsui and Y. Hayashi, High rate sputter deposition of TiO2 from TiO2-x target, Thin Solid Films 331 (1) (1999) 57-60. [33] M. J. Jung, K. H. Nam, L. R. Shaginyan and J. G. Han, Deposition of Ti thin film using the magnetron sputtering method, Thin Solid Films, 435 (1-2) (2003) 145-149. [34] K. Cai, M. Müller, J. Bossert, A. Rechtenbach and K. D. Jandt, Surface PT structure and composition of flat titanium thin films as a function of film thickness and evaporation rate, Appl. Surf. Sci., 250 (1-4) (2005) 252-267. U. Balachandran and N. G. Eror, Raman spectra of titanium dioxide, J. Solid RI [35] [36] SC State Chem. 42 (3) (1982) 276-282. V. V. Yakovlev, G. Scarel, C. R. Aita and S. Mochizuki, Short-range order in Lett. 76 (9) (2000) 1107-1109. [37] NU ultrathin film titanium dioxide studied by Raman spectroscopy, Appl. Phys. T. Sekiya, S. Ohta, S. Kamei, M. Hanakawa and S. Kurita, Raman MA spectroscopy and phase transition of anatase TiO2 under high pressure, J. Phys. Chem. Solids 62 (4) (2001) 717-721. T. Ohsaka, F. Izumi and Y. Fujiki, Raman spectrum of anatase, TiO2, J. ED [38] Raman Spectrosc. 7 (6) (1978) 321-324. [39] W. X. Xu, S. Zhu, X. C. Fu and Q. Chen, The structure of TiO x thin film EP T studied by Raman spectroscopy and XRD, Appl. Surf. Sci. 143 (3) (1999) 253-262. [40] S. Mukherjee and D. Mergel, Thickness dependence of the growth of AC C magnetron-sputtered TiO2 films studied by Raman and optical transmittance spectroscopy, J. Appl. Phys. 114 (1) (2013) 013501. [41] D. Bersani, P. P. Lottici and X. Z. Ding, Phonon confinement effects in the Raman scattering by TiO2 nanocrystals, Appl. Phys. Lett. 72 (1998) 73. [42] W. Ma, Z. Lu and M. Zhang, Investigation of structural transformations in nanophase titanium dioxide by Raman spectroscopy, Appl. Phys. A Mater. Sci. Process. 66 (6) (1998) 621-627. [43] T. Ohtsuka, J. Guo and N. Sato, Raman spectra of the anodic oxide film on titanium in acidic sulfate and neutral phosphate solutions, J. Electrochem. Soc. 133 (12) (1986) 2473-2476. 20 ACCEPTED MANUSCRIPT [44] S. K. O’Leary, S. R. Johnson and P. Lim, The relationship between the distribution of electronic states and the optical absorption spectrum of an amorphous semiconductor: An empirical analysis, J. Appl. Phys. 82 (7) (1997) 3334-3340. [45] R. Brendel and D. Bormann, An infrared dielectric function model for amorphous solids, J. Appl. Phys. 71 (1) (1992) 130001-06. [46] W. Theiβ, Optical properties of porous silicon, Surf. Sci. Rep., 29 (3-4) (1997) [47] PT 91-192. M. H. Habibi, N. Talebian and J. H. Choi, The effect of annealing on RI photocatalytic properties of nanostructured titanium dioxide thin films, Dyes [48] SC Pigm. 73 (1) (2007) 103-110. T. M. Wang, S. K. Zheng, W. C. Hao and C. Wang, Studies on photocatalytic NU activity and transmittance spectra of TiO2 thin films prepared by rf magnetron sputtering method, Surf. Coat. Technol. 155 (2) (2002) 141-145. [49] P. B. Nair, V. B. Justinvictor, G. P. Daniel, K. Joy, V. Ramakrishnan and P. V. MA Thomas, Effect of RF power and sputtering pressure on the structural and optical properties of TiO2 thin films prepared by RF magnetron sputtering, [50] ED Appl. Surf. Sci. 257 (24) (2011) 10869-10875. D. Mergel, Modeling thin TiO2 films of various densities as an effective optical medium, Thin Solid Films 397 (1) (2001) 216-222. A. Bendavid, P. Martin and H. Takikawa, Deposition and modification of EP T [51] titanium dioxide thin films by filtered arc deposition, Thin Solid Films 360 (1) (2000) 241-249. M. Laube, F. Rauch, C. Ottermann, O. Anderson and K. Bange, Density of AC C [52] thin TiO2 films, Nucl. Instrr. Meth. Phys. Res. B 113 (1-4) (1996) 288-292. [53] D. Mergel, D. Buschendorf, S. Eggert, R. Grammes and B. Samset, Density and refractive index of TiO2 films prepared by reactive evaporation, Thin Solid Films 371 (1) (2000) 218-224. [54] E. Khawaja, F. Bouamrane, F. Al-Adel, A. B. Hallak, M. A. Daous and M. A. Salim, Study of the Lorentz-Lorenz law and the energy loss of 4He ions in titanium oxide films, Thin Solid Films 240 (1-2) (1994) 121-130. [55] S. Kim, Simultaneous determination of refractive index, extinction coefficient, and void distribution of titanium dioxide thin film by optical methods, Appl. Opt. 35 (34) (1996) 6703-6707. s 21 ACCEPTED MANUSCRIPT [56] S. B. Amor, G. Baud, J. Besse and M. Jacquet, Structural and optical properties of sputtered Titania films, Mater. Sci. Eng., B 47 (2) (1997) 110118. [57] R. Dannenberg and P. Greene, Reactive sputter deposition of titanium dioxide, Thin Solid Films 360 (1) (2000) 122-127. [58] J. Szczyrbowski, G. Bräuer, M. Ruske, J. Bartella, J. Schroeder and A. Zmelty, Some properties of TiO2 layers prepared by medium frequency [59] PT reactive sputtering, Surf. Coat. Technol. 112 (1) (1999) 261-266. J. Q. Xi, M. F. Schubert, J. K. Kim, E. F. Schubert, M. Chen, S.-Y. Lin, W. RI Liu and J. A. Smart, Optical thin-film materials with low refractive index for [60] SC broadband elimination of Fresnel reflection, Nat. Photonics, 1 (3) (2007) 176. G. K. Mor, O. K. Varghese, M. Paulose and C. A. Grimes, Transparent highly NU ordered TiO2 nanotube arrays via anodization of titanium thin films, Adv. Funct. Mater., 15 (8) (2005) 1291-1296. [61] S. Lee and J. Hong, Comparison of various parameterization models for MA optical functions of amorphous materials: application for sputtered titanium dioxide thin films, Jpn. J. Appl. Phys. 39 (1R) (2000) 241. G. S. Henderson, X. Liu and M. E. Fleet, A Ti L-edge X-ray absorption study ED [62] of Ti-silicate glasses, Phys. Chem. Miner. 29 (1) (2002) 32-42. [63] Y. Hwu, Y. Yao, N. Cheng, C. Tung and H. M. Lin, X-ray absorption of [64] EP T nanocrystal TiO2, Nanostruct. Mater., 9 (1-8) (1997) 355-358. R. Ruus, A. Kikas, A. Saar, A. Ausmees, E. Nommiste, J. Aarik, A. Aidla, T. Uustare and I. Martinson, Ti 2p and O 1s X-ray absorption of TiO2 AC C polymorphs, Solid State Commun. 104 (4) (1997) 199-203. [65] A. Thomas, W. Flavell, A. Mallick, A. Kumarasinghe, D. Tsoutsou, N. Khan, C. Chatwin, S. Rayner, G. Smith, R. Stockbauer, S. Warren, T. K. Johal, S. Patel, D. Holland, A. Taleb, and F. Wiame, Comparison of the electronic structure of anatase and rutile TiO2 single-crystal surfaces using resonant photoemission and X-ray absorption spectroscopy, Phys. Rev. B, 75 (3) (2007) 035105. [66] F. Farges, G. E. Brown Jr. and J. J. Rehr, Ti K-edge XANES studies of Ti coordination and disorder in oxide compounds: Comparison between theory and experiment, Phys. Rev. B, 56 (4) (1997) 1809. 22 ACCEPTED MANUSCRIPT [67] J. P. Crocombette and F. Jollet, Ti 2p X-ray absorption in titanium dioxides (TiO2): the influence of the cation site environment, J. Phys.: Condens. Matter SC RI PT 3 (49) (1994) 10811. Fig. 1. XRR plot of all as-deposited TiO2 thin films with fitting using MA Parratt32 software Fig. 2. NU List of the figure caption The deposition rate of TiO2 thin films with the different gas ratio. Region I and II represent partially and fully oxidized deposition rate. XRD plot of all annealed (350 ºC) TiO2 thin films deposited at the ED Fig. 3. different O2/Ar gas ratio. XRD plot of annealed TiO2 (O2/Ar = 1.50) thin films at the different EP T Fig. 4. temperature viz. 350-650 ºC. Fig. 5. Raman spectra of annealed TiO2 thin films deposited at the different AC C O2/Ar gas ratio. Standard anatase (A) and rutile (R) shifts are marked as dotted lines for comparison. Fig. 6. Raman spectra (together with fitting curves) of the annealed (350 ºC) TiO2 thin films with O2/Ar = 0.25 and 0.11. Fig. 7. Transmittance spectra of as-deposited TiO2 thin films deposited at the different O2/Ar gas ratio. Fig. 8. Transmittance spectrum of the TiO2 thin film (O2/Ar = 1.50) together with fitting curves generated using two different dielectric models, one with Brendel oscillator (BO) and the other one with Harmonic oscillator (HO). 23 ACCEPTED MANUSCRIPT Fig. 9. Band gap and refractive index of TiO2 thin films as a function of O2/Ar gas ratio. The data points within the dotted circle show band gap and refractive index of TiO2 film deposited at O2/Ar = 0.11. Fig. 10. SXAS plot of TiO2 thin films at Ti-edge deposited at the different O2/Ar gas ratio. Fig. 11. SXAS plot (together with fitting curves) of TiO2 thin film with O2/Ar PT = 1.50. List of the table caption RI Table – I : Raman peak position of the present work and references SC Table – II : Refractive index (n) of TiO2 deposited using the different methods at the different wavelength Present Sekiya et Ohsaka et Xu et al. W. Ma et Balachandran Work al. [37] al. [38] [39] al. [42] et al. [35] 144.5 144 145 -- 147 MA Band 142.4 Eg 196.6 197.2 197 195 -- 198 B1g 394.5 396.6 399 396 -- 398 A1g 515.8 515.2 513 512 -- 515 Eg 637.6 638.6 639 639 -- 640 183 -- -- -- -- -- 202 -- -- -- -- -- 260 -- -- -- 238, 250 235 294 -- -- -- -- -- 566 -- -- -- -- -- Eg 443 -- -- -- 450 448 A1g 615 -- -- -- 606-613 612 Unidentified Rutile EP T Anatase ED Eg AC C Phase NU Table – I 24 NU SC RI PT ACCEPTED MANUSCRIPT Deposition method Filtered arc deposition 2.32 - 2.72 550 A. Bendavid et al. [51] Evaporation 2.23 - 2.35 550 M. Laube et al. [52] 2.0 - 2.3 550 D. Mergel et al. [53] 1.9, 2.3 550 E. Khawaja et al. [54] 2.28 - 2.51 550 S.Y. Kim et al. [55] E-beam evaporation 2.3 - 2.5 550 M. G. Krishna et al. [17] E-beam evaporation 2.30 - 2.46 550 M. Jerman et al. [18] E-beam evaporation 1.3 - 2.7 633 J. Q. Xi et al. [59] Atomic layer deposition 2.2 - 2.3 633 J. Aarik et al. [19] Sputtering 1.66 380-800 G. K. Mor et al. [60] Sputtering 2.24 - 2.46 550 M. H. Suhail et al. [5] Sputtering 2.25 - 2.7 550 H. Ohsaki et al. [32] Sputtering 2.05 - 2.59 550 Ya-Qi Hou et al. [24] Sputtering 2.1 - 2.5 550 S. Ben Amor et al. [56] DC Sputtering 2.1 - 2.35 550 P. Löbl et al. [26] DC Sputtering 2.4 - 2.6 550 R. Dannenberg et al. [57] Twin magnetron sputtering Twin magnetron sputtering 2.45 - 2.73 550 J. Szcyrbowski et al. [58] 2.45 550 S. Lee et al. [61] Evaporation AC C Evaporation EP T Evaporation MA Wavelength (nm) 546 Reference Theoretical Fit Data Refractive Index (n) 2.52 - 2.95 ED Table – II D. Mergel et al. [50] 25 ACCEPTED MANUSCRIPT 2.42 - 2.56 550 M. Laube et al. [52] Ion beam sputtering 2.48 630 S. M. Rossnagel et al. [8] Ion beam sputtering 1.98 - 2.52 633 H. Demiryont et al. [30] Ion bean sputtering 1.74, 2.07 - 2.81 550 Present Case RI PT Sputtered and ion plated  SC Highlights: - TiO2 thin films deposited using reactive ion beam sputtering at room Complete target oxidation at O2/Ar pressures much lower than for MA  NU temperature. conventional sputtering. Ti coordination in the as-deposited films is less than anatase and rutile.  Annealed films prepared with the O2/Ar ratio between 0.25-0.66 ED   EP T remain amorphous. Below this range, develop Ti6O11and Ti3O5 grains and above this AC C anatase occurs. 26