In the context of photocatalytically self-cleaning windows and windshields, clear, abrasion resistant, thin (60 +/- 10 nm) photocatalytic films of TiO2 were formed by a sol-gel process on (a) soda lime glass, (b) the proton-exchanged surface of soda lime glass, and (c) fused silica. The hypothesis that diffusion of sodium oxide from the soda lime glass into the titanium dioxide layer during the calcination step causes the lower photoefficiency in films on glass was tested and proven. At high concentration sodium prevented formation of the photoactive anatase phase and, at low concentration, introduced surface and bulk recombination centers. Sodium transport was efficiently blocked by a thin layer at the interface of proton-exchanged (''hydrogen'') glass and nascent TiO2, formed at 400 degrees C of a poly(titanyl acetylacetonate) TiO2 precursor. The sodium transport blocking layer did not form and the highly photocatalytic film was not obtained when the TiO2-precursor film was applied to glass that was not proton exchanged. Furthermore, only a much less effective sodium transport blocking layer was formed on glass that was proton-exchanged, but was calcined at 400 degrees C prior to application of the TiO2 precursor layer, showing that the sodium depleted glass surface, by itself, was a less effective barrier against sodium transport than the interfacial product of hydrogen glass and the TiO2 precursor.
