This dissertation was presented to the Graduate Council of the University of Florida in partial fulfillment of the requirements for the degree of Doctor of Philosophy. An extensive critical review of the glass corrosion and related literature is presented. Many irregularities are noted and some of the previously reported information is reinterpreted. Specific questions are raised which are discussed in detail. Differences are shown to exist between the corrosion of glass powders and the corrosion of bulk glass surfaces. For glass grains,dissolution kinetics of t 2 and t are found to be only limiting conditions at short and long times respectively. Bulk glass surfaces, however, have these two conditions for long periods of time and have J' 1 a sharp transition from t 2 to t kinetics when the pH exceeds approximately 9.5. The data from the corrosion of glass grains indicate that concentration cell effects exist which can produce anomalous results. Variations of the surface area of glass to volume of solution (A/V) on glass corrosion kinetics are investigated. For bulk glass surfaces graphs of log Q vs. log A/V (at a given time of corrosion) and graphs of log t (time to reach a given concentration of glass constituent in solution) vs. log A/V yield results which are in agreement with derived equations. The corrosion of glass grains, however, yields results inconsistent with the derived equations. This is attributed to the change in the effective surface area of the glass grains as corrosion proceeds. Many glasses in the L^O-, Na-0-, and K^O-SiOp systems are corroded using bulk glass surfaces and static test conditions. Reaction rate constants are determined for the leaching of alkali ions and for the extraction of silica during the i2 regime. Concentrations of alkali ions and silica in solution are used to calculate the corrosion parameters alpha and epsilon. Corroded glass surfaces are examined by infrared reflection spectroscopy (IRRS) and techniques are developed for compositional analysis of corroded binary alkali silicate glass surfaces and for the determination of the leached layer compositional profiles. An equation is developed for calculating the thickness of the leached layer (X) from the surface composition and epsilon and calculated values of X are verified to be within experimental error. Another equation is developed for calculating the thickness of glass totally dissolved from the surface (X-,). Trie kinetics of binary alkali silicate glass corrosion are followed with solution data, corrosion parameters, IRRS, surface composition, leached layer profiles, X, X-j and X/X-j . A general sequence of events is described for the corrosion of these glasses. The temperature dependent corrosion kinetics for two glasses are presented. The same kinetic sequence of events occurs at 30°C, 50°C, and 100°C and only the time sequence to obtain equivalent extents of corrosion is changed. The time-temperature dependence does, however, depend on the composition of the glass. Graphs of log (time to reach a critical extent of corrosion) vs. log A/V and temperature studies are discussed in terms of the possible long term prediction of corrosion and the study of nuclear waste containing glasses. Evidence is presented which indicates that the surface of most corroded glasses undergoes a transformation induced by strain from the ion exchange reaction in which many silanol groups combine during a dehydration reaction to form new silicon bridging oxygen bonds. Interdiffusion coefficients are calculated for the ion exchange reaction. The results indicate that the interdiffusion process is controlled by the diffusion of alkali ions out of the glass and by the openness of the leached layer. A model is proposed for the & dependent release of silica into acidic and neutral solutions. Diffusion coefficients are calculated which support a model in which Si-0 bonds are attacked by OH ions from solution and free silanol groups are produced which diffuse through the transformed porous vitreous silica like leached layer.
