Glass dissolution rate measurement and calculation revisited

Aqueous dissolution rate measurements of nuclear glasses are a key step in the long-term behavior study of such waste forms. These rates are routinely normalized to the glass surface area in contact with solution, and experiments are very often carried out using crushed materials. Various methods have been implemented to determine the surface area of such glass powders, leading to differing values, with the notion of the reactive surface area of crushed glass remaining vague. In this study, around forty initial dissolution rate measurements were conducted following static and flow rate (SPFT, MCFT) measurement protocols at 90 °C, pH 10. The international reference glass (ISG), in the forms of powders with different particle sizes and polished monoliths, and soda-lime glass beads were examined. Although crushed glass grains clearly cannot be assimilated with spheres, it is when using the samples geometric surface (S_geo) that the rates measured on powders are closest to those found for monoliths. Overestimation of the reactive surface when using the BET model (S_BET) may be due to small physical features at the atomic scale - contributing to BET surface area but not to AFM surface area. Such features are very small compared with the thickness of water ingress in glass (a few hundred nanometers) and should not be considered in rate calculations. With a S_BET/S_geo ratio of 2.5 ± 0.2 for ISG powders, it is shown here that rates measured on powders and normalized to S_geo should be divided by 1.3 and rates normalized to S_BET should be multiplied by 1.9 in order to be compared with rates measured on a monolith. The use of glass beads indicates that the geometric surface gives a good estimation of glass reactive surface if sample geometry can be precisely described. Although data clearly shows the repeatability of measurements, results must be given with a high uncertainty of approximately ±25%.