An approach for the assessment of the HLW glass source term

Reprocessing nuclear fuel produces wastes of different types and activity levels, particularly high-level aqueous ''fission product solutions''. The radionuclides are physically stabilized by melting at approximately 1000 to 1100 deg C and incorporation in the glass structure. The glass is cast into a stainless steel canister to allow handling of the nuclear glass package after cooling. Vitrified waste is currently placed in interim storage at the production site in France, i.e. the reprocessing plants operated by COGEMA at Marcoule and La Hague. The resulting sodium aluminoborosilicate glass contains oxides of the activation and fission products and actinides accounting for most of the activity initially contained in the reprocessing effluents. Each package represents between 50 and 60 kilograms of fission products and actinides in oxide form, for the R7T7 reference glass produced at La Hague. It is generally considered that these wastes will eventually represent over 99% of activity that would be suitable for disposal in a geological repository. The release of radionuclides in a repository from a nuclear glass package (or ''vitrified waste source term'') is defined by a flow of radioactive species from the pristine glass block and its alteration products. The physical and chemical processes governing the long-term behavior of the glass package are highly complex. It is generally assumed that water is both the primary nuclear glass alteration factor in an underground repository environment and the principal radioactivity dispersal vector; most of the work done in this area over the last decade has thus focused on aqueous corrosion of glass. In this context, this paper describes the current status of modeling work for the ''vitrified waste'' source term. The theoretical and experimental justifications of this approach are presented and discussed. (author). 28 refs., 2 figs