Journal of Crystal Growth 41(1977) 41—44 © North-Holland Publishing Company
INTERMITTENT CHEMICAL VAPOUR DEPOSITION OF ANATASE FILMS Shinsuke HAYASHI and Toshio HIRA! The Research Institute for Iron, Steel and Other Metals, Tohoku University, Sendai, 980, Japan Received 16 May 1977;revised manuscript received 15 June 1977
An intermittent CVD (chemical vapour deposition) method has been applied to the production of hO
2 deposits on a substrate heated at 400 and 900°C,using a TiC14—H20 system; TiCl4 vapour was admitted intermittently into the CVD chamber, while H20 vapour was conducted into it continuously, with Ar gas. The effects of the intermittent CVD conditions, i.e. the period of TiC14 vapour introduction (t1) and its intermission (t2) on the crystal structure of the Ti02 deposits have been investigated. The anatase content in the Ti02 deposits varied with t1 and t2 and is maximum (about 95 wt%) at t1 = t2 = S mm.
vapour for an intermission of TiC14 vapour introduction, t2. This cycle is repeated at different intervals of t1 and t2. The details of the apparatus are already described in a previous paper . The deposition conditions are shown in table 1. In most of the experiments, the vapour pressure of TiCl4 (P-~1c14)is 10.0 Torr.
In the ordinary chemical vapour deposition (continuous CVD), flow rates of reactant gases, deposition temperatures and other CVD conditions are maintamed in a steady state. However, pulsed CVD or intermittent CVD, in which the above CVD conditions are varied periodically [1—5], has received increasing attention in recent years. One feature of this method is to form a metastable phase [1,5]. Up to the present, the CVD of Ti02 has been carned out by the continuous CVD method alone [6—12]. In the present work anatase, a metastable phase of TiO2 , has been prepared on the substrate heated at deposition temperatures (Tdep) of 400 and 900°C,by admitting TiCl4 vapour intermittently (intermittent CVD method) using a TiC14— H20 system. CVD This conditions paper describes effect of the the mtermittent on thethe structure
2.2. Characterization of the Ti02 deposits The anatase content in the deposits was calculated from the X-ray powder diffraction intensities of the anatase (101) and rutile (110) reflections by means of Table 1 Intermittent CVD conditions Deposition temperature, Tdep with TiC1 Flow rate of 4Arvapour gas saturated
Flow rateHof Ar gas saturated with 20 vapour Vapour pressure of TiC14,
2. Experimental procedure
~TiC14 Vapour pressure of H20, 2as pressure ~l~i 0 Total Periodg of T1C1 4 vapour introduction, t5 Intermission of TiCl4 vapour introduction, t2 Substrate
2.1. Intermittent CVD method H20 vapour is carried continuously with Ar gas into the CVD chamber. TiCl4 vapour is admitted with Ar gas for a period of TiCl4 vapour introduction, t1~ followed by the admission of Ar gas without TiCl4 41
mm~ 400 and3 900 C 300 cm 3 min~ 200 cm 3.2 and 10.0 Torr 18 Torr 760 Torr 1, 2, 3,5, 10 and 20 min 1, 2, 3, 5, 10 and 20 mm Silicon polycrystal
S. Hayashi, T. Hirai / Intermittent CVD of anatase films
Spurr’s method . The experimental accuracy of the anatase content is ±3%of wt% values. Infrared absorption was measured by a KBr method using an infrared spectrometer (Hitachi 285) in order to examine the presence of OH and Cl in the deposits. The deposition rates, densities and microstructures of the deposits were also studied by the same method described previously . The precision of the density measurement is estimated as ±0.5%.
3. Results 3.1. Anatase content in the Ti02 deposits
Fig. 1 shows the anatase content in the TiO2 deposits obtained at t~= t2 = 1, 3, 5, 10 and 20 mm, and Tdep = 400 and 900°C. At Tdep = 400°C,the anatase content increases from 45 wt% at t1 = = 1 mm to 96 wt% at t1 = t2 = 5 mm and decreases above t1 = t2 = 5 mm, as indicated by circles. As with the case of Tdep = 400°C,the anatase content in the deposits prepared at Tdep = 900°C is maximum (95 wt%) at t1 = t2 = 5 mm. At t1 = 5 mm, the anatase content is 92 to 93 wt% at t2 10 mm and is 76 to 83 wt% at t2 = 2 mm. The present results imply that anatase is obtained predominantly at t1~5mm and t2~5 mm. Neither brookite nor amorphous Ti02 is formed in the deposits.
0 t (= 12)
Fig. 2. (a) Deposition surface of the intermittent CYDanatase prepared at t1 = t2 = 5 mm and Tdep = 400°C.(b) Fracture surface of the intermittent CVD-anatase prepared at = t2 5 mm and Tdep 400°C.
Figs. 2a and 2b show the scanning electron micrographs of the intermittent CVD-anatase prepared at Tdep = 400°C. Fine anatase grains of about 0.2 pm size seem to form a cone structure, in which primary and secondary cones appear. For the intermittent CVD-anatase obtained at Tdep = 900°C,grains are 2 to 7pm in size and appear to be loosely packed as shown in figs. 3a and 3b, which differ markedly
Fig. 1. Relation between t 1 ( t2) and the anatase content in the intermittent CVD-Ti02 deposits.
the microstructures of the continuous CVD.
S. Hayashi, T. Hirai Intermittent CVD of anatase films - -
The intermittent CVD-anatase is white at Tdep 900°Cand is blue at Tdep = 400°C.It is said that the blue anatase contains oxygen vacancies  and easily transforms to rutile [19.]. When heat-treated at 900°Cin air for 6 h, 64% of the intermittent CVDanatase obtained at Tdep = 400°Cwas transformed to rutile. The optimum intermittent CVD condition for the preparation of anatase is t1 ~ 5 mm and t2 ~ 5 mm as described above. As the continuous CVD condition arises at t1 > 5 mm, rutile formation occurs as reported previously . At t2 <5 mm, the peaks of TiCl4 vapour introduction will overlap each other; a perfect intermittent CVD condition is not attained resulting in the decrease in the anatase content. In order to examine whether the formation of anatase in the intermittent CVD process is due to the low concentration of TiC14 vapour in the initial and latter stages of TiC14 vapour introduction, the continuous CVD was carried out at a low FTICi4 = 3.2 Torr. The anatase contents ranged from 4 to 5 wt% and rutile was formed predominantly even at a low concentration of TiCl4 vapour. This fact seems to support the conception that the formation of anatase is caused not by the low concentration of TmC14 vapour but by the intermittent admission of TiCl4 vapour. In the CVD process clusters normally form at first under a supersaturation condition and grow to partides . It was reported that a freshly-formed cluster has the anatase structure in the CVD of Ti02 . In the present experiments, anatase was formed in the initial stage of TiCl4 vapour introduction and then rutile was formed when TiCl4 vapour was steadily carried into the CVD chamber. This result suggests that the anatase formed in the initial stage of the vapour introduction seems not to transform to rutile during isthe Theand intermittent CVD-anatase notdeposition stabilized run. by OH Cl. Thus, the structure of the anatase cluster formed in the mitial stage of the vapour introduction is considered to =
5 p.m I
Fig. 3. (a) Deposition surface of the intermittent CVDanatase prepared at t1 = t2 = 5 mm and Tdep = 900 C. (b) Fracture surface of the intermittent CVD-anatase prepared at t1 = t2 = 5 mm and Tdep = 900°C.
3.3. Density, deposition rate and infrared spectra The densities of the intermittent CVD-anatase prepared at Tdep = 400Here, and 900°Care 3.77 and 3.86 3,respectively. the theoretical densities g cm~ of anatase and rutile are 3.94 and 4.27 g cm3, respectively [161. The deposition rates of the intermittent CVD-TiO 2 deposits are independent of t1 and t2, and are almost equal to those of the continuous CVD-rutile [141 at Tdep = 400 and 900°C.Infrared spectra of the intermittent CVD-anatase show no absorption peaks of OH and Cl .
be different from that formed in the steady state. As the crystal structures of Ti02 (rutile and anatase) can be controlled by the intermittent CVD method, an attempt has been made to prepare a layertype composite, i.e. a rutile—anatase—rutile com-
S. Hayashi, T. Hiral / Intermittent CVD of anatase films
~iUT LE ANATASE P TILE
Fig. 4. Fracture surface of the rutile-anatase-rutile composite. The rutile layer was prepared by the continuous CVD method and the anatase layer was prepared at ti = t 2 = 5 min by the intermittent CVD method (Tdep = 400°C).
posite. Fig. 4 shows the fracture surface of the rutile—anatase—rutile composite, in which the rutile layer was obtained by the continuous CVD method and the anatase layer was prepared at t1 = t2 = 5 mm by the intermittent CVD method at Tdep = 400°C.
Acknowledgements The authors wish to thank Dr. M. Kikuchi for the IR measurements. This research was supported in part by the scientific research fund from the Ministry of Education, Contract No. 175423.
[1)      
J.M. BY. (1975) Blocher, Derjaguin 102. Jr., and MF. D.B. Browning, Fedoseev, Wilson, Amer. V.M. 233 Secrest, Nuci. WA. E. Bryant, Sci. A.C. Eng. and J.Secrest, 20 Less~ommonMetals H. (1964) Nickel, RB. 153. Carbon Landrigan 8 W.J. (1970) 45Sci. and (1976) J.H. 400. 37. Oxley, T. (1977) Hirai 205. and S. Hayashi, Bull. Inst. Metals 16 M. Gyarmati Yokozawa, H. Iwasa and I. Japan Teramoto, Japan. J. Appl. Phys. 7 (1968) 96. M. Balog and M. Schieber, J. Crystal Growth 17 (1972) 298.  RN. Ghoshtagore, J. Electrochem. Soc. 117 (1970) 529.
 RN. Ghoshtagore and A.J. Noreika, J. Electrochem. Soc. 117 (1970) 1310.  M.P. Lepie and R.A. Pucel, Raytheon Co. Tech. Mem. T-697 (May 1966).  Y. Suyama, K. Ito and A. Kato, J. Inorg. Nucl. Chem. 37 (1975) 1883.  Y. Suyama, K. Ohmura and A. Kato, J. Chem. Soc. Japan No. 4 (1976) 584.  J.C. Jamieson and B. Olinger, Am. Mineralogist 54 (1969) 1477.  S. Hayashi and T. Hirai, J. Crystal Growth 36 (1976) 157.
R.A. and H. Myers, Anal. Chem. 29 (1957) 760.J. E.T. Spurr Fitzgibbons, K.J. Sladek and W.H. Hartwig, Electrochem. Soc. 119 (1972) 735.  GD. Parfitt, J. Ramsbotham and C.H. Rochester, J. Chem. Soc., Faraday Trans. I, 68 (1972) 17.  K.L. Hardee and A.J. Bard, J. Electrochem. Soc. 122 (1975) 739.  R.D. Shannon and J.A. Pask, J. Am. Ceram. Soc. 48 (1965) 391.  R.F. Strickland-Constable, Kinetics and Mechanism of Crystallization (Academic Press, New York, 1968).