High purity, spherical anatase nanocrystals were prepared by a modified sol-gel method. Mixing of anhydrous TiCl(4) with ethanol at about 0 degrees C yielded a yellowish sol that was transformed into phase-pure anatase of 7.7 nm in size after baking at 87 degrees C for 3 days. This synthesis route eliminates the presence of fine seeds of the nanoscale brookite phase that frequently occurs in low-temperature formation reactions and also significantly retards the phase transformation to rutile at high temperatures. Heating the as-is 7.7 nm anatase for 2 h at temperatures up to 600 degrees C leads to an increase in grain size of the anatase nanoparticles to 32 nm. By varying the calcination time from 2 to 48 h at 300 degrees C, the particle size could be controlled between 12 and 15.3 nm. The grain growth kinetics of anatase nanoparticles was found to follow the equation, D(2) - D(0)(2) = k(0)t(m)e((-)(E)(a)/(RT)) with a time exponent m = 0.286(+/-9) and an activation energy of E(a) = 32 +/- 2 kJ x mol(-)(1). Thermogravimetric analysis in combination with infrared and X-ray photoemission spectroscopies has shown the anatase nanocrystals at different sizes to be composed of an interior anatase lattice with surfaces that are hydrogen-bonded to a wide set of energetically nonequivalent groups. With a decrease in particle size, the anatase lattice volume contracts, while the surface hydration increases. The removal of the surface hydration layers causes coarsening of the nanoparticles.

The energetics of pure-phase rutile nanorods and spherical anatase nanoparticles have been studied by high-temperature drop solution calorimetry in 3Na 2 O‚4MoO 3 solvent at 975 K and water adsorption calorimetry in a wide range of particle sizes (surface area) from 6 to 40 nm (5-270 m 2 /g). The calorimetric surface enthalpies for rutile and anatase, calculated as 2.22 ( 0.07 and 0.74 ( 0.04 J/m 2 , respectively, are in general agreement with Ranade et al. 's results (Proc. Natl. Acad. Sci. U.S.A., 2002, 99, 6476), although their numerical values are somewhat different because of impurities and unaccounted bound water in the previous work. This study supports the energy crossovers previously proposed for the TiO 2 polymorphs. The energetics of water adsorption were measured using a commercial Calvet microcalorimeter coupled with a gas dosing system. This permitted the calculation of differential and integral enthalpies of water adsorption that characterize how tightly water binds to rutile and anatase surfaces and the calculation of adsorption entropies, which reflect the surface mobility of adsorbed water. The integral enthalpy of tightly bound water (relative to liquid H 2 O standard state) is -18 kJ/mol for anatase and -40 kJ/mol for rutile. As seen previously for Al 2 O 3 , the TiO 2 polymorphs with higher surface energy bind water more tightly. The calculated entropy changes for the adsorption of water on TiO 2 are more negative than the entropy changes for the condensation of gaseous water to hexagonal ice. This finding suggests highly restricted mobility of molecules adsorbed at initial stages of adsorption (low coverage) and, possibly, dissociative adsorption on both rutile and anatase surfaces. However, nanoparticles contain both tightly bound water and loosely bound water. The latter is characterized by energetics of bulk water. The stabilizing water contribution to the overall energy of the system makes the hydrated nanophase samples more stable. The recommended transformation enthalpy for bulk anatase to bulk rutile is -1.7 ( 0.9 kJ/mol.