Sol-gel titania particles were investigated, primarily by optical techniques, by systematically varying synthesis, sample handling, and annealing variables. The material phases investigated were amorphous titania, anatase TiO2, and rutile TiO2. Annealing-induced phase transformations from amorphous TiO2 to anatase to rutile were studied by Raman scattering, infrared reflectivity, infrared absorption, x-ray diffraction, and electron energy-loss spectroscopy (EELS). Detailed experiments were carried out on the effects of annealing on the Raman and infrared spectra of anatase nanocrystals. The frequencies of the zone-center transverse optical (TO) and longitudinal-optical (LO) phonons of anatase were determined and were used in analyzing the results obtained on composites consisting of annealed solgel particles.

The TO and LO frequencies of anatase were obtained from polarization-dependent far-infrared reflectivity measurements on single crystals. These results, which determined the dielectric functions of anatase, were used to explain infrared (IR) reflectivity spectra of titania nanoparticles pressed into pellets, as well as the grazing-incidence IR reflectivity observed for titania thin films. Because of the polycrystalline character of the titania nanoparticles, the surface roughness of the pressed pellets, and the island-structure character of the thin films, effective-medium theories (appropriate for composites) were used, along with the anatase dielectric functions, to interpret the experimental results.

The titania nanoparticles were prepared by the hydrolysis/condensation of Ti(OC2H5)4. A polymeric steric stabilizer was used in the sol-gel synthesis in order to prevent continued agglomeration during the condensation process. This yielded particles with a relatively narrow size distribution. The amount of water used in the reaction determines the final particle size. Particles as small as 80 nm and as large as 300 nm were used throughout this work. From the colloidal suspension, loose powders, pressed pellets, and thin films were formed. These samples were subjected to different annealing processes at temperatures ranging from room temperature up to 1000 C. Two different annealing atmospheres were used: air (oxygen-containing) and argon (no oxygen).

The amorphous to anatase transformation was followed by in-situ IR transmission measurements carried out during annealing. The particles as prepared are amorphous and the anatase phase could be detected, using this sensitive IR technique, at temperatures as low as 150 C. This phase transition was shown to be particle size dependent. It was also shown that introducing the stabilizer by means of the alkoxide flask instead of the water flask (during the sol-gel synthesis) decreases the anatase to rutile transformation temperature. Loose powders were found to transform more readily than dense pellets, while island-structure films were found to be the hardest to transform. Even at 1000 C, most of these films did not transform to rutile.

X-ray diffraction experiments were used to determine nanocrystal sizes in anatase samples obtained by air and argon anneals at temperatures from 300 to 800 C. A correlation was found between Raman band shape (peak position and linewidth) and crystallite size, but this correlation was different for air anneals and for argon anneals. These experiments called for an interpretation based on a stoichiometric effect rather than a finite size effect. Based on this interpretation, the as-prepared particles are slightly oxygen-deficient, with a stoichiometry corresponding to TiO1.98.

In the electron energy-loss experiments, a special data-analysis technique was used to extract the EELS spectrum of the titania nanoparticles from the observed substrate-plus-particles signal. This technique successfully resolved the titania absorption-edge peak. Which was found to be momentum independent. For low electron momentum, the results were consistent with the reported optical absorption edge.