The aqueous corrosion of potash-lime-silica glass in the range 10–250°C

Abstract

It is difficult to measure accurately the natural long-term corrosion rates of highly durable nuclear waste glasses. Care should be taken when using data from high temperature experiments to predict corrosion rates under ambient repository conditions as there are many factors (such as the precipitation of secondary compounds, the chemistry of the water in contact with the glass, or circulation of this water through the repository) that can influence the temperature dependence of aqueous reactions. In this study, some standard leach tests using pure water, in continuous flow and direct sampling autoclaves, were performed on a synthetic medieval glass. Archaeological glass samples of a similar composition and which had suffered several centuries of corrosion in damp soil had been the subject of previous studies. The corrosion of the archaeological samples in damp soil was well understood and the aim was to determine how well standard leach tests would predict the observed corrosion. The experimentally measured long-term corrosion rates were not simply dependent on temperature, therefore they could not be used to predict either the corrosion rates or products found on the archaeological samples. In continuous flow experiments this was largely because the release of some cations (e.g. Mg 2+) was proportional to the flow rates used, and so the long-term corrosion was dependent on both flow rate and temperature. In the autoclave experiments, concentrations of dissolved ions climbed rapidly in the small water volume, leading to the precipitation of complex mineral phase assemblages. The experiments did not reach silica saturation, therefore the long-term rates measured were a function of both the temperature and the extent of reaction. The initial release rates of silica and of the network modifying cations showed an Arrhenius temperature dependence with an activation energy of about 78.5 kJmol −1. These results would have predicted the soil corrosion of the archaeological samples to within about 30%, i.e. the initial corrosion rate in pure water was similar to the long-term rate in damp soil. As an explanation of the experimental observations it is suggested that clay minerals in the soil consumed silica as the glass dissolved, maintaining under-saturated conditions and allowing corrosion at a rate similar to the initial rate in pure water.