Journal of Colloid and Interface Science 231, 289298 (2000) doi:10.1006/jcis.2000.7147, available online at http://www.idealibrary.

com on

The Effect of Alcohol Solvents on the Porosity and Phase Composition of Titania
Ki Chang Song1 and Sotiris E. Pratsinis2
Institute of Process Engineering, Swiss Federal Institute of Technology, ETH-Zentrum, CH-8092 Zurich, Switzerland Received March 22, 2000; accepted August 4, 2000

Bimodally porous titania powders were made by hydrolysis of titanium tetraisopropoxide (TTIP) dissolved in various alcohols (methanol, ethanol, isopropanol, and sec-butanol). The specic surface area (SSA) of the powders dried at 150 C ranged from 332 to 624 m2 /g as determined by nitrogen adsorption. At excess alcohol concentration, the SSA of the dried powders decreased in the order of sec-butanol, iso-propanol, ethanol, and methanol at a constant alcohol/TTIP molar ratio. The pore size distribution was bimodal with ne intraparticle pore diameters at 16 nm and larger interparticle pore diameters at 30120 nm as determined by nitrogen adsorption isotherms. The average intraparticle pore diameter decreased with increasing alcohol concentration for methanol and ethanol, while it was rather constant at 3.3 nm, irrespective of alcohol concentration for iso-propanol and sec-butanol. The evolution of particle phase composition was determined by X-ray diffraction ranging from amorphous to crystalline anatase and rutile largely proportional to the calcination temperature and to a lesser extent on the type and concentration of alcohols. C 2000 Academic Press Key Words: titania powders; alcohol; bimodal pore size distribution; phase composition.

INTRODUCTION

Titania powders are used as cosmetics, pigments, photocatalysts, adsorbents, catalytic supports, and sensors (14). In these applications, the morphology, average particle size and particle size distribution, phase composition, and porosity of the titania powders are important factors to be controlled. Titania powders are produced by liquid and gas phase reactions. In the liquid phase process, aqueous-based methods using metal salts as a precursor material, such as the forced hydrolysis method (5) and the homogeneous precipitation method (6, 7), encounter several challenges; e.g., the concentration of reacting species should be low and counterions of the starting metal salts remain in the product and deteriorate the purity of powders (8). Solgel synthesis provides a way of making nanoparticles by a chemical reaction in solution starting with metal alkoxides as a precursor. An advantage of the solgel synthesis is that it is
1 2

On leave from Konyang University, South Korea. To whom correspondence should be addressed.

possible to obtain large amounts of powder with a high level of chemical purity (9). In this synthesis route hydrolysis and polycondensation reactions take place when TTIP reacts with water (10). These two reactions form nuclei (2 to 10 nm in diameter) which develop into primary particles (50 to 100 nm in diameter). These primary particles aggregate to form secondary particles (300 to 1000 nm in diameter) during aging (11). During solvent removal processes (ltration or centrifuging) from the solution, these secondary particles agglomerate to form wet gel powders. Through drying, solvents in the wet gel powders are evaporated to give dried powders (xerogel powders) with large shrinkage. In the above particle formation process, the TTIP is dissolved in alcohol to moderate the high reactivity of the TTIP with water during hydrolysis. When different type and concentration of alcohols are added during hydrolysis, it is expected that these may affect the hydrolysis rate of TTIP and, subsequently, the aggregation rate of the primary particles. As a result, this may affect primary and secondary particle sizes, morphology, phase composition, and porosity of the powders. St ober et al. (12) synthesized monodisperse, spherical silica particles from combinations of different silicon alkoxides and alcohols. It was found that the hydrolysis rate became slower in the order of methanol, ethanol, n -propanol, and n butanol. They reported that nal particle sizes obtained under comparable conditions were smallest in methanol and biggest in n -butanol. Harris and Byers (13) studied the growth of colloidal titania by titanium ethoxide hydrolysis in ethanol, n -butanol, and tert-butanol. They found that particle growth was controlled by the surface phenomena of particles and that the morphology of precipitates formed was affected by the solvent: monodisperse spherical particles were formed in the solvents of ethanol and n -butanol, but the precipitate formed in tert-butanol had no regular shape. Park et al. (14) investigated the formation of titania powders from thermal hydrolysis of TiCl4 and the morphology of the precipitates dispersed in various alcohols. They reported that the size of secondary particles decreases as the molecular weight of alcohols increases. Recently, Vorkapic and Matsoukas (15) synthesized titania powders from various titanium alkoxides and investigated the effect of alcohol on the size of the titania particles. They claimed that the smallest size of the particles was obtained when no alcohol was added to
0021-9797/00 $35.00
Copyright C 2000 by Academic Press All rights of reproduction in any form reserved.

289

290

SONG AND PRATSINIS

the reaction mixture, while the presence of alcohol resulted in larger particles. Look and Zukoski (16) studied shear induced agglomeration during the homogeneous nucleation and growth of particles formed through the hydrolysis and condensation of tetraethylortho-titanate in aqueous ethanol solutions. They found that there is a minimum shear rate below which uniform particles are precipitated but when the precipitating suspension is sheared above this value, the resulting precipitate consists of agglomerated masses of particles that have been fused together by further reaction. Although many research groups have explored the effect of alcoholic solvent on the morphology and particle size, however, none of the above referenced studies has examined the effect of these solvents on the phase transformation and porosity of the titania powders. In this paper, we investigate the effect of the type and concentration of alcoholic solvents added during hydrolysis on the texture and phase composition of product titania powders.
EXPERIMENTAL

where X is the weight fraction of rutile in the powders, while IA and IR are the X-ray integrated intensities of the (101) reection of anatase and the (110) reection of rutile. The specic surface area of the powders was obtained by multipoint BET nitrogen adsorption (ASAP2010, Micromeritics) in the relative pressure ( P / P0 ) range of 0.050.25. All powders were degassed at 150 C prior to measurement. The desorption isotherm was used to determine the pore size distribution using the Barrett, Joyner, and Halender (BJH) method with cylindrical pore size (20).
RESULTS AND DISCUSSION

(1) Thermal Analysis Figure 1 shows the differential thermal analysis (DTA) of powders dried at 150 C and made at different ethanol concentrations. For the E5, E10, and E20 powders (Table 1) made in the presence of ethanol, a strong exothermic peak is observed at 260 C, which is associated with the oxidation of residual organic components after hydrolysis (21). On the other hand, for the N0 powder (Table 1) made in the absence of alcohol, a weak exothermic peak is observed at 260 C. This indicates that the E5, E10, and E20 powders contain more residual organic components than the N0 powder possibly from incomplete hydrolysis. For the E20 powder, another exothermic peak at 390 C is observed, corresponding to the crystallization of amorphous hydrated titania to anatase TiO2 (21). However, this peak is not found in the N0 and E5 powders and shows a weak intensity in the E10 powder. As this peak is more noticeable in intensity for amorphous powders, it is expected that the E20 powder is amorphous hydrated titania up to 350 C before the strong exothermic peak appears. In contrast, the N0, E5, and E10 powders which

Titanium tetraisopropoxide (TTIP; Ti(OC3 H7 )4 , 97%, Aldrich) was used as a precursor, while methanol, ethanol, iso-propanol, and sec-butanol were used as solvents. Prior to reaction with water, TTIP was rst dissolved in various concentrations of different alcoholic solvents. The molar ratio of alcohol/ TTIP was varied in the range of 040. When TTIP mixed with alcohols, the alcoholysis reaction (i.e., the alcohol added replaces the alcohol attached to titanium atom in TTIP) took place (17). In especial, when TTIP was dissolved in methanol, a precipitate was produced since the alcoholysis reaction produced titanium isopropoxy methoxides insoluble in methanol. However, the precipitate was not produced in the other alcohols. Hydrolysis was carried out at room temperature by adding a constant amount of distilled water slowly to the mixed solution (TTIP and alcohol) for 30 min with a vigorous stirring and maintaining the stirring for another 30 min to attain the equilibrium of the reactions (hydrolysis and particle formation). The molar ratio of H2 O to TTIP was 20. On adding water into the mixed solution, a slurry containing white hydrated titania precipitate was formed (18). Clear liquid was removed from the slurry by separation in a superspeed centrifuge (Rotina35, Hettich) at 9000 rpm for 10 min. The ltered powder was dried at 150 C for 24 h, followed by calcination at different temperatures ranging from 300 to 800 C for 1 h at a heating rate of 10 C/min. The powders were analyzed by thermogravimetric differential thermal analysis (TG-DTA; STA501, Bahr) at a heating rate of 10 C/min in owing air. An X-ray diffractometer (D5000, Siemens) was used for identication of crystalline species in the calcined powders with Ni-ltered Cu K radiation as the X-ray source for diffraction angles 2 between 20 and 60 . The weight fraction of rutile or anatase in the titania powders was calculated by (19) X = 1/[1 + 0.8 IA / IR ],

FIG. 1. Differential thermal analysis (DTA) curves of solgel made titania powders. The powders were made at ethanol/TTIP molar ratios of 0, 5, 10, and 20, corresponding to N0, E5, E10, and E20 (Table 1), respectively, and dried at 150 C.

SOLGEL MADE TITANIA IN ALCOHOL SOLVENTS

291

TABLE 1 Effect of Type and Concentration of Alcohols on the Specic Surface Area and Pore Volumes of the powders dried at 150 C
Alcohol/TTIP molar ratio 0 5 10 20 40 5 10 20 40 5 10 20 40 5 10 20 40 Specic surface area (m2 /g) 354.0 489.4 432.7 446.4 331.6 581.7 564.2 540.4 402.3 516.2 546.5 592.4 559.4 523.6 554.2 624.1 596.2 Pore volume (cm2 /g) 0.31 0.37 0.29 0.26 0.19 0.41 0.36 0.33 0.25 0.54 0.52 0.58 0.44 0.56 0.63 0.64 0.54

Sample no. N0 M5 M10 M20 M40 E5 E10 E20 E40 P5 P10 P20 P40 B5 B10 B20 B40

Alcohol type No alcohol Methanol Methanol Methanol Methanol Ethanol Ethanol Ethanol Ethanol Iso-propanol Iso-propanol Iso-propanol Iso-propanol Sec-butanol Sec-butanol Sec-butanol Sec-butanol

physically adsorbed water and alcohol. The second zone from 230 to 380 C corresponds to the oxidation of residual organic components and removal of chemically adsorbed water (21). However, for the N0 powder, only the rst zone is observed, which corresponds to the removal of physically adsorbed water and alcohol. This indicates that the E5, E10, and E20 powders have more residual organic components than the N0 powder due to the incomplete hydrolysis. This is in a good agreement with the DTA data in Fig. 1. (2) Phase Composition Figures 3a3d represent the XRD patterns of the N0, E5, E10, and E20 powders, respectively, at various calcination temperatures. The E5, E10, and E20 powders dried at 150 C are amorphous, while the N0 powder is anatase. According to Terabe et al. (22), when the amount of water and HCl addition during hydrolysis of TTIP was small, large amounts of unhydrolyzed alkyls remained in the dried powder. These alkyls prevent crystallization to anatase so that the dried powders are amorphous. However, when the water and HCl concentration were large, the amount of residual alkyls preventing crystallization to anatase was small, so the molecular structure of the dried powder resembled that of anatase, which is more stable than amorphous titania. In our study, when the TTIP is dissolved in ethanol, TTIP molecules in ethanol are more diluted, inhibiting the hydrolysis of TTIP. Thus, dried powders made in ethanol contain more residual organic components than the N0 powder, which is conrmed by TG-DTA results in Figs. 1 and 2. Hence, the dried powders made in ethanol are amorphous, while those made without ethanol are anatase. It is seen in Figs. 3b3d that the phase transformation from amorphous to anatase occurs at 300 C for the E5 and E10 powders and at 400 C for the E20 powder. On the other hand, the N0 powder shows anatase at 150 C and maintains it till 500 C in Fig. 3a. When the initial ethanol concentration increases, the phase transformation temperature of amorphous to anatase increases. This is in agreement with the TG-DTA results of Figs. 1 and 2. However, the phase transformation from anatase to rutile begins to appear at 600 C for all powders, regardless of the initial ethanol/TTIP molar ratio. This means that the initial ethanol concentration has no effect on the phase transformation temperature of anatase to rutile. Figure 4 shows the calculated rutile weight fraction in the N0, E5, E10, and E20 powders calcined at three temperatures as a function of ethanol/TTIP molar ratios. The rutile weight fraction shows almost a constant value, regardless of the initial ethanol concentration at a constant calcination temperature. This conrms that the anatase to rutile transformation rate is not inuenced by the ethanol concentration. Figures 5a5c represent the XRD patterns of the M10, P10, and B10 powders (Table 1) calcined at various temperatures after being prepared with methanol, iso-propanol, and sec-butanol, respectively. It is seen that the phase transformations from amorphous to anatase and anatase to rutile occur at 300 and 600 C,

did not show such a strong peak have already had anatase TiO2 before 390 C as shown by XRD in Figs. 3a3d. Figure 2 shows the thermogravimetric (TG) analysis of the N0, E5, E10, and E20 powders dried at 150 C. For the E5, E10, and E20 powders, two main zones in weight loss are shown. The rst zone from 50 to 230 C corresponds to the removal of

FIG. 2. The weight reduction of solgel made titania powders by thermogravimetric (TG) analysis in owing air. The powders were made at ethanol/TTIP molar ratios of 0, 5, 10, and 20, corresponding to N0, E5, E10, and E20 (Table 1), respectively, and dried at 150 C.

292

SONG AND PRATSINIS

FIG. 3. The XRD patterns of solgel made titania powders calcined at various temperatures. The powders were made at ethanol/TTIP molar ratios of (a) 0, (b) 5, (c) 10, and (d) 20.

respectively, irrespective of alcohol type. When TTIP mixed with various types of alcohols, the alcoholysis reaction, where the alcohol added replaces the alcohol attached to titanium atom in TTIP, might take place, resulting in different hydrolysis rate of TTIP. However, in this study, it is interesting to observe that the alcoholysis reaction has no signicant effect on the phase transformations from amorphous to anatase and anatase to rutile.

Figure 6 shows the specic surface area (SSA) of the N0, E5, E10, and E20 powders as a function of calcination temperature. All powders show a decrease in the SSA with increasing calcination temperature from the phase transformation and crystallite growth. It is interesting to observe that the N0 powder shows a moderate decrease without a rapid decrease in SSA during calcination, while the powders prepared with ethanol exhibit a

SOLGEL MADE TITANIA IN ALCOHOL SOLVENTS

293

FIG. 4. The calculated rutile weight fractions of the powders made at differential ethanol/TTIP molar ratios and calcined at various temperatures for 1 h.

rapid decrease in the SSA between 150 and 300 C for E5 and E10 powders and between 300 and 400 C for E20 powder. This rapid decrease in the SSA is attributed to the phase transformation from amorphous to anatase phase. (3) Texture Property Table 1 summarizes the SSA and pore volume of the powders dried at 150 C and made with different type and concentration of alcohols. The SSA ranges from 332 to 624 m2 /g and decreases along with pore volume with increasing methanol and ethanol concentration. The SSA and pore volume show a maximum at molar ratio of alcohol/TTIP = 20 and then decrease with increasing iso-propanol and sec-butanol concentration. The dried powders made with methanol show smaller SSA and pore volume than those from the other alcohols at a given alcohol/TTIP molar ratio. For low alcohol/TTIP molar ratios of 5 and 10 the SSA decreases in the order of ethanol, sec-butanol, iso-propanol, and methanol. However, the SSA decreases in the order of secbutanol, iso-propanol, ethanol, and methanol for high alcohol/ TTIP molar ratios of 20 and 40. The pore volume of the dried powders decreases in the order of sec-butanol, iso-propanol, ethanol, and methanol for all alcohol/TTIP molar ratios. Figures 7a and 7b show the adsorption isotherms of the powders dried at 150 C and made with various alcohol concentrations of ethanol and sec-butanol, respectively. According to BDDT classication (20, 23), the isotherms corresponding to the E5, E10, and E20 powders in Fig. 7a are of type IV with two hysteresis loops indicating bimodal pore size distributions. The shape of the hysteresis loops is of type H3 associated with aggregates of plate-like particles giving rise to slit-like pores (20, 23). Unlike the isotherms of the E5, E10, and E20 powders, the

isotherm of the E40 powder with excess ethanol concentration is a combination of types I and IV (20, 23) with two distinct regions: at low relative pressures, the isotherm shows high adsorption, indicating that the powder contains micropores (type I). However, at high relative pressures between 0.8 and 1.0, the curve exhibits a hysteresis loop, indicating the presence of mesopores (type IV). The shape of the hysteresis loop is also of type H3, associated with narrow slit-like pores. On the other hand, all powders made with sec-butanol show type IV isotherm with two hysteresis loops (Fig. 7b), irrespective of the alcohol concentration. The shape of the hysteresis loops is also of type H3. Figures 8a and 8b show the pore size distribution of the powders dried at 150 C and made with different concentrations of ethanol and sec-butanol, respectively. All powders show bimodal pore size distribution consisting of smaller intraparticle pores and larger interparticle pores. It is seen that the average intraparticle pore diameter decreases with increasing ethanol concentration (Fig. 8a), while that shows almost a constant value (3.3 nm), irrespective of alcohol concentration for sec-butanol (Fig. 8b). The pore size distribution of the powders made in methanol shows similar behavior with ethanol, while that made in iso-propanol exhibits similar behavior with sec-butanol. In the solgel synthesis of titania, when TTIP reacts with water, it goes through the hydrolysis and polycondensation reactions to give nuclei (2 to 10 nm in diameter) which develop primary particles (50 to 100 nm in diameter). These primary particles then organize into secondary particles (300 to 1000 nm in diameter) by aggregation (11). Further aggregation of these secondary particles gives agglomerates with several tens of micrometers in diameter (11). Look and Zukoski (16) claimed that in the absence of shear, the reactions result in the formation of uniform spherical particles. However, above a critical shear rate the nal precipitate is composed of agglomerates consisting of fused particles. Our work showed results similar to those of Look and Zukoski. In this study, we stirred the solution during particle formation vigorously. This resulted in the formation of agglomerates and not uniform particles. The bimodal pore size distributions of the titania powders in our study are closely related with the above particle formation mechanism. The intraparticle pores are associated with the voids between the primary particles within the secondary particles, while the interparticle pores are due to the voids between the secondary particles within the agglomerates (24). Thus, the observed difference in our experiment, in terms of the SSA and porosity of the dried powders, is attributed to the way in which the aggregates of primary and secondary particles are packed. When methanol and ethanol are used as solvents, higher alcohol concentration separates the primary particles formed and hinders the progress of aggregation due to the dilution effect of excess alcohol, leading to more dense packing of the primary particles within the secondary particle. This results in a decrease in the average diameter of intraparticle pores and the SSA with increasing alcohol concentration. However, for iso-propanol and sec-butanol, the dilution effect of excess alcohol seems to have

294

SONG AND PRATSINIS

FIG. 5. The XRD patterns of solgel made titania powders calcined at various temperatures. The powders were made with (a) methanol, (b) iso-propanol, and (c) sec-butanol. The molar ratio of alcohol/TTIP was 10.

little effect on the average diameter of the intraparticle pores and the SSA. Figures 9a and 9b represent the pore size distributions of the powders dried at 150 C and made with different types of alcohol for alcohol/TTIP molar ratios of 10 and 20, respectively. For both ratios, the average intraparticle pore diameters of powders made in the presence of iso-propanol and sec-butanol are larger

than those made in the presence of methanol and ethanol. It is interesting to observe that powders made in iso-propanol and sec-butanol show the same average intraparticle pore diameter (3.3 nm), while powder made in methanol has smaller average of intraparticle pore diameter than ethanol-made powder. This can be explained by the difference of capillary force during drying of wet gel powders. During drying, large shrinkage

SOLGEL MADE TITANIA IN ALCOHOL SOLVENTS

295

show bimodal distributions consisting of the intraparticle and interparticle pores at 500 C. At 600 C, the intraparticle pores of the N0 powder have not disappeared completely and show a small peak, while those of the E20 powder collapse completely.

FIG. 6. The specic surface areas of the N0, E5, E10, and E20 powders (Table 1) as a function of calcination temperatures.

in the wet gel powders takes place by solvent evaporation in the pores. This shrinkage determines the agglomeration state of the resultant dried powders and is strongly affected by the capillary force which arises from the evaporation of the solvents in the wet gel powders. The capillary force ( P ) can be expressed by (17) P = 2 cos / r, where is the surface tension, r is the pore radius, and is the contact angle. The surface tensions of the alcohol considered in this study have similar values (25). Thus, the most important factor which affects the capillary force becomes the pore radius in the wet gel powders. As the molecular weight of alcohol is increased, the dielectric constant of the alcohol decreases (25). This results in decreased stability, leading to an enhanced rate of aggregation of primary particles in the solution (14). Thus, the aggregation rate may be expected to be faster in isopropanol and sec-butanol than in methanol and ethanol, leading to a looser packing of secondary particles in the former solvents. This looser packing of secondary particles has large pores and thus the capillary force during drying becomes smaller, leading to the dried powders with looser packing. This is probably the reason why the average intraparticle pore diameters of the dried powders made with iso-propanol and sec-butanol are bigger than those made with methanol and ethanol. Similar result was reported by Vorkapic and Matsoukas (15) who made titania powders at various solvents. They found that the degree of aggregation increased in the following order: methanol, ethanol, and n -propanol. Figures 10a and 10b show the pore size distributions of the N0 and E20 powders (Table 1), respectively, as a function of calcination temperatures. For both N0 and E20 powders, the pores

FIG. 7. Nitrogen adsorption isotherms of the powders dried at 150 C and prepared with various concentrations of (a) ethanol and (b) sec-butanol.

296

SONG AND PRATSINIS

interparticle pores have completely disappeared for both powders. It is interesting to note that the intraparticle pores in the N0 powder have not disappeared completely at 600 C, compared to that in the E20 powder. That means that the intraparticle pores of

FIG. 8. Pore size distributions of the powders dried at 150 C and prepared with various concentrations of (a) ethanol and (b) sec-butanol.

It is seen in both powders that the pores exhibit monomodal distribution of the interparticle pores due to the complete collapse of the intraparticle pores at 800 C. When the calcination temperature increases from 600 to 800 C, the average interparticle pore diameter increases, suggesting pore growth. At 1000 C, the

FIG. 9. Pore size distributions of the powders dried at 150 C and prepared with various types of alcohol. The molar ratios of alcohol/TTIP were xed to (a) 10 and (b) 20.

SOLGEL MADE TITANIA IN ALCOHOL SOLVENTS

297

could be attributed to the phase transformation of anatase to rutile, as conrmed in Figs. 3a and 3d. Kumar et al. (26) observed the microstructures of titania membranes calcined at various temperatures by high-resolution scanning electron micrograph. They found that the powders before anatase to rutile phase transformation contain small anatase crystallites, but the powders after transformation have smaller anatase crystallites and bigger densied rutile regions. It is thought that smaller anatase crystallites grow into bigger rutile crystallites through the phase transformation, leading to the disappearance of the voids between anatase crystallites which results in the collapse of intraparticle pores in the bimodal pore size distributions. The effect of alcohol type on particle aggregation rate is explained by dilution and colloidal stabilization. For a given alcohol, as its concentration increases, the number concentration of primary particles is decreased and the aggregation rate is reduced, leading to more dense packing of the primary particles within the secondary particle. This results in a decrease in the average diameter of intraparticle pores with increasing alcohol concentration in Fig. 8a. However, for iso-propanol and sec-butanol, the dilution effect has little effect on the average diameter of the intraparticle pores in Fig. 8b. On the other hand, at a given alcohol/TTIP molar ratio as the molecular weight of alcohol increases, the dielectric constant of the alcohol decreases, leading to an enhanced rate of aggregation of primary particles within the secondary particle. This increases the average diameter of intraparticle pores with increasing molecular weight of alcohol as shown in Figs. 9a and 9b.

SUMMARY

FIG. 10. Pore size distributions of the (a) N0 and (b) E20 powders (Table 1) calcined at various temperatures.

the powder made with ethanol collapse more easily than those of the powder made without ethanol. Another interesting result from these gures is that the main cause of the disappearance of most intraparticle pores at 600 C

The type and concentration of alcohol added during hydrolysis of TTIP play an important role on the phase composition and texture of solgel made titania powders. Alcohol retards the hydrolysis reaction of the TTIP and results in residual organic components in the powders by incomplete hydrolysis. When ethanol was not added during hydrolysis of TTIP, the powders dried at 150 C were anatase. On the other hand, when ethanol was added, the dried powders were amorphous as residual organic components prevented the crystallization to anatase. However, the phase transformation temperature of anatase to rutile was at 600 C, independent of the type and concentration of alcohols. The porosity of the powders dried at 150 C could be controlled by the alcohol type and concentration. All dried powders showed bimodal pore size distributions consisting of ne intraparticle pores and larger interparticle pores. The average intraparticle pore diameter decreased with increasing alcohol concentration for methanol and ethanol, while that showed a constant value (3.3 nm), irrespective of alcohol concentration for iso-propanol and sec-butanol. The intraparticle pores disappeared at 600 C, whereas the interparticle pores collapsed completely at 1000 C.

298
ACKNOWLEDGMENTS

SONG AND PRATSINIS 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Yoldas, B. E., J. Mater. Sci. 21, 1087 (1986). Bowen, H. K., Mater. Sci. Eng. 65, 1574 (1986). St ober, W., Fink, A., and Bohn, E. J., J. Colloid Interface Sci. 26, 62 (1968). Harris, M. T., and Byers, C. H., J. Non-Cryst. Solids 103, 49 (1988). Park, H. K., Kim, D. K., and Kim, C. H., J. Am. Ceram. Soc. 80, 743 (1997). Vorkapic, D., and Matsoukas, T., J. Am. Ceram. Soc. 81, 2815 (1998). Look, J.-L., and Zukoski, C. F., J. Colloid Interface Sci. 153, 461 (1992). Brinker, C. J., and Scherer, G. W., Sol-Gel Science. Academic Press, San Diego, 1990. Song, K. C., and Pratsinis, S. E., J. Mater. Res., in press (2000). Spurr, R. A., and Myers, H., Anal. Chem. 29, 760 (1957). Sing, K. S. W., Everett, D. H., Haul, R. A. W., Moscou, L., Pierotti, R. A., Rouquerol, J., and Siemieniewska, T., Pure Appl. Chem. 57, 603 (1985). Montoya, I. A., Viveros, T., Dominguez, J. M. Canales, L. A., and Schifter, I., Catal. Lett. 15, 207 (1992). Terabe, K., Kato, K., Miyazaki, H., Yamaguchi, S., Imai, A., and Iguchi, Y., J. Mater. Sci. 29, 1617 (1994). Gregg, S. J., and Sing, K. S. W., Adsorption, Surfae Area, and Porosity. Academic Press, London, 1982. Song, K. C., and Pratsinis, S. E., J. Am. Ceram. Soc., in press (2000). West, R. C., Handbook of Chemistry and Physics. CRC Press, Boca Raton, FL, 1989. Kumar, K. N. P., Keizer, K., and Burggraaf, A. J., J. Mater. Chem. 3, 1141 (1993).

This research was sponsored by the SNF (Swiss National Science Foundation) and KOSEF (Korea Science and Engineering Foundation). We acknowledge the assistance and contribution of Dr. Jinsoo Kim in the preparation of the nal manuscript.

REFERENCES
1. Mezey, E. J., Vapor Deposition. Wiley, New York, 1966. 2. Barringer, E. A., and Bowen, H. K., Langmuir 1, 414 (1985). 3. Akhtar, M. K., Pratsinis, S. E., and Mastrangelo, S. V. R., J. Am. Ceram. Soc. 75, 3408 (1992). 4. Fotou, G. P., Vemury, S., and Pratsinis, S. E., Chem. Eng. Sci. 49, 4939 (1994). 5. Matijevic, E., Pure Appl. Chem. 50, 1193 (1978). 6. Sacks, M. D., Tseng, T. Y., and Lee, S. Y., Am. Ceram. Soc. Bull. 63, 301 (1984). 7. Sorddet, D., and Akinc, M., J. Colloid Interface Sci. 122, 47 (1988). 8. Kominami, H., Takada, Y., Yamagiwa, H., Kera, Y., Inoue, M., and Inui, T., J. Mater. Sci. Lett. 15, 197 (1996). 9. Koebrugge, G. W., Winnubst, L., and Burggraaf, A. J., J. Mater. Chem. 3, 1095 (1993).