MATERIALS CHEMISTRY COMMUNICATION Physical stabilization of anatase (TiO2) by freeze-drying Hiroyuki Izutsu,a Padmakumar K. Nairb and Fujio Mizukamib aT aki Chemical Co., L td. 2,Midorimachi, Befu-cho, Kakogawa, 675-01 Hyogo, Japan bNational Institute of Materials and Chemical Research, 1-1 Higashi, T sukuba, 305 Ibaraki, Japan the primary particles of the anatase phase sinter together to The anatase phase of titania has been stabilized without reach the critical nucleus size.10 From circumstantial evidence, chemically modifying the system, by changing the level of it is expected that the critical nucleus size of rutile is at least packing within the aggregates by freeze-drying the water- three times larger than the crystallites present in the anatase washed precipitate obtained by the hydrolysis of titanium phase.11 This means that if sintering of anatase particles is isopropoxide.After calcination at 700 °C for 8 h, freeze-dried retarded by a suitable technique (the probability of reaching samples showed more than 97% anatase phase with a surface the critical nucleus size is lowered), we can retain a high area and porosity of 29 m2 g-1 and 37% respectively. The porosity and surface area as well as retard the transformation oven-dried samples showed a surface area and porosity of 5 m2 to rutile.In this study we have retarded the sintering by g-1 and 10% respectively. eectively decreasing the packing of anatase particles by freeze-drying. Titania powder was prepared by the hydrolysis of titanium isopropoxide (Wako Chemicals, Japan). The alkoxide (0.18 mol) was dissolved in 350 g (5.8 mol) of isopropyl alcohol and The anatase modification of titania is an important material this solution was added dropwise to 350 g of distilled water for catalysis.1 Anatase is a metastable phase and it will under vigorous stirring at room temperature.After separation transform irreversibly to the stable rutile phase.2–5 For many of the precipitated gel by decantation and centrifugation, the applications this transformation is not favoured, for two main precipitate was washed five times with a total of 5 l of deionized reasons: (1) anatase has distinctly better catalytic properties water.This washed precipitate was then divided into two than rutile,1 and (2) this transformation always results in a portions: one portion was dried in an oven at 110 °C for 24 h dense (non-porous) rutile phase, which is not useful as a (designated as WWOD); the other portion was freeze-dried catalyst or as a ceramic membrane material.5,6 It is generally (WWFD). observed that the anatase-to-rutile transformation temperature From the thermal analysis data it was found that both depends on many factors like preparation conditions, nature samples were amorphous in the as-prepared state and on of the precursor, minor impurities, morphology of the primary heating transforms first to anatase and then to rutile.Table 1 particles etc.6 Almost all the attempts to stabilize the anatase gives the DTA transformation temperatures and the textural phase have been based on changing the chemistry of the properties (surface area and porosity) of the oven-dried and system.1,6,7 Even in those studies in which a chemical modifi- freeze-dried samples.As expected, both transformation tem- cation was not the primary aim, the stabilization technique peratures, amorphous-to-anatase and anatase-to-rutile, are ultimately resulted in modifying the chemistry of the system.8,9 higher for the freeze-dried samples. During drying the WWOD However, the degree of modification is thought to be much samples experience a large compressive stress, owing to the smaller than for other chemical methods of stabilization.8,9 In surface tension of the pore fluid, compared to the WWFD this paper we report, for the first time, a purely physical means samples.Therefore the level of packing of primary particles of stabilization of the anatase phase. The major advantage of within the aggregate and also the packing of aggregates this is that unlike chemical modification, this technique will themselves will be higher compared to the WWFD. It should not change the chemistry of the system, which means that the be noted that during freezing, the first step in the freeze-drying catalytic properties are not changed.process, the WWFD samples will also experience compressive The anatase-to-rutile transformation is a nucleation growth stresses similar to drying stresses.12,13 However, this stress will type of transformation and the temperature and rate at which not be acting throughout the sample and, moreover, the it happens, from a physical point of view, depends on how fast possibility for the primary particles to rearrange during freezing is much lower compared to the rearrangement during the Table 1 Eect of drying method on the phase transformation tempera- oven-drying of WWOD.A lower level of packing will lead to ture and BET surface area large porosity (Table 1) and a broad pore size distribution with a higher average pore size.This can be seen from the phase transformation temp./°Ca surface area/m2 g-1 pore size distribution data given in Fig. 1. The average pore size of WWOD samples is lower than that of WWFD. method am.�ana. ana.�rut. 400 °C 700°C Moreover, WWOD samples show a broader pore size distribution with some pores in the microporous (pore radius oven-dried 399 727 116 (51)b 5 (10) <1 nm) range.freeze-dried 416 895 75 (57) 29 (37) A lower level of packing in WWFD samples will result in a aPhase transformation temperaures from amorphous (am.) to anatase lower number of particle–particle contacts and this in turn (ana.) and anatase to rutile (rut.) obtained from exothermic peaks of results in slower particle growth.Therefore WWFD samples DTA curves. bValues in parentheses show porosities (%) evaluated by will transform more slowly than WWOD samples. Fig. 2 shows using total N2 adsorption amounts and densities of titania (3.8g ml-1 XRD patterns of the samples heated at 400 and 700 °C. After for oven-dried and 400 °C heated sample and freeze-dried samples, 4.2g ml-1 for oven-dried and 700°C heated sample).heating at 400 °C for 8 h both the samples show the typical J. Mater. Chem., 1997, 7(6), 855–856 855Fig. 1 Pore size distributions of WWOD and WWFD before heating (a) and after heating to 400 °C (b) and 700 °C (c). Upper traces: freeze-dried; lower traces: oven-dried samples the WWFD sample shows higher porosity indicating a lower level of packing. At 700 °C WWFD has a much higher surface area (29 m2 g-1) and porosity (37%) with more than 97% anatase phase.From the above results it can be seen that freeze-drying is a very eective way to stabilize the catalytically important anatase phase with high surface area and porosity at relatively high temperatures. References 1 K. Foger and J. R. Anderson, Appl. Catal., 1986, 23, 139. 2 R. D.Shannon and J. A. Pask, J. Am. Ceram. Soc., 1965, 48, 391. 3 C. N. R. Rao, Can. J. Chem., 1961, 39, 498. 4 F. Dachille, P. Y. Simens and R. Roy, Am.Mineral., 1968, 53, 1929. 5 K-N. P. Kumar, K. Keizer, A. J. Burggraaf, T. Okubo, S. Morooka and H. Nagamoto, Nature (L ondon), 1992, 358, 48. 6 K-N. P. Kumar, K. Keizer and A. J. Burggraaf, J. Mater. Chem., 1993, 3, 1141. 7 Y-S. Lin, C-H.Chang and R. Gopalan, Ind. Eng. Chem. Res., 1994, 33, 860. 8 K-N. P. Kumar, Jalajakumari Kumar, K. Keizer, T. Okubo, Fig. 2 X-Ray diraction patterns of the titania samples: (a) WWOD M. Sadakata and J. Engell, J. Mater. Sci. L ett., 1995, 14, 1784. heated at 400°C; (b) WWFD heated at 400 °C; (c) WWOD heated at 9 K-N. P. Kumar, Appl. Catal., 1994, 119, 163. 700 °C; (d) WWFD heated at 700°C. A=anatase, R=rutile 10 K-N. P. Kumar, Jalajakumari Kumar and K. Keizer, J. Am. Ceram. Soc., 1994, 77, 1396. 11 K-N. P. Kumar, Scr.Metall.Mater., 1995, 32, 873. anatase pattern [Fig. 2 (a) and (b)], but at 700 °C WWFD 12 G. W. Scherer, J. Non.-Cryst. Solids, 1993, 155, 1. [Fig. 2(d)] practically remains as anatase and WWOD 13 J. Kumar, PhD Thesis, University of Twente, The Netherlands, [Fig. 2(c)] transforms to rutile. 1995. The higher surfarea of WWOD at 400°C is probably Communication 7/01755C; Received 13thMarch, 1997 due to the presence of micropores in the WWOD. However, 856 J. Mater. Chem., 1997, 7(6), 855–856