Ce travail s’inscrit dans le cadre du projet Cigéo (Centre industriel de stockage géologique de déchets radioactifs) et a porté sur la valorisation des argilites du Callovo-Oxfordien sous forme de coulis d’injection activé alcalinement. Cette utilisation requiert un liant présentant une importante durée pratique d’utilisation (~24 heures) et une valeur de pH modérée (10-11). Ainsi l’activation thermique de l’argilite a dû être maitrisée afin d’accroitre la réactivité des espèces argileuses tout en limitant la décarbonatation, pour éviter la présence de calcium réactif. Une solution alcaline silicatée potassique, de réactivité modérée (Si/K = 0,70), a été utilisée pour l’activation alcaline des argilites calcinées. Afin de développer le coulis, il a été nécessaire de favoriser la réactivité de l’argilite par addition de métakaolin et d’incorporer des additifs à base de bore. Un coulis dont les caractéristiques correspondent à celles visées a ainsi été obtenu. Enfin, une étude du liant dans des conditions proches de celles de l’utilisation finale (injection entre la couche géologique d’argilites et le chemisage de déchets radioactifs) a été menée. Pour cela, le liant a été mis en contact avec un bloc d’argilites du Callovo-Oxfordien, à une température de 90 °C (température maximale en condition d’usage). Il a ainsi été mis en évidence qu’à 90 °C le réseau géopolymère se réorganise (intégration du bore dans le réseau et formation éventuelle de zéolite). La mise en contact du liant avec l’argilite géologique a aussi permis d’identifier une diffusion du potassium dans les argilites du Callovo-Oxfordien. / In the context of Cigéo project (geological industrial disposal for radioactive wastes), this work focuses on the valorization of Callovo-Oxfordian argillite by the formulation of an alkali-activated grout. This application required a high setting time (close to 24 hours) and a moderate pH value (10-11). Thus, the argillite thermal treatment has been adapted to alter clays minerals without decomposing carbonate species (to avoid the presence of reactive calcium). A potassium-based silicate solution, with moderated reactivity (Si/K = 0.70), has been used for the alkali-activation step. Concerning the grout formulation, it has been necessary to enhance the argillite reactivity by the insertion of metakaolin and to add boron-based compounds. The resulting grout was in accordance with the specification. Finally, a study of the binder in an environment close to the using conditions (injection between geological layer and a radioactive waste lining) has been investigated. In this aim, the binder has been casted in contact with the geological argillite at 90 °C. It was evidenced a network reorganization at 90 °C (boron integration in the network and possible formation of zeolite). Potassium diffusion from the binder to the argillite has also been observed.

The focus of this work is to investigate the dissolution of MW25, a non-radioactive simulant of UK high-level nuclear waste borosilicate glass, and to predict its performance in the near field of a geological repository. A single-pass flow-through (SPFT) experimental system was used to measure the forward dissolution rates of MW25. Experiments were conducted in two parts. Experiment Part 1 considers the dissolution of the waste glass in deionised water at 40 and 90 oC and circum-neutral pH. Experiment Part 2 considers the dissolution of the waste glass in simulant groundwaters, with similar compositions to groundwaters of Callovo-Oxfordian clay (lower-strength sedimentary rock (LSSR)) and Borrowdale Volcanic Group rocks (higher-strength rock (HSR)), at 40 oC and pH 7. The forward dissolution rate measured in deionised water was found to be approximately one order of magnitude higher at 90 oC than at 40 oC. A similar release was observed for Si, Mg and Al at 40 oC and 90 oC, whereas the B, Cs, Na, Li and Mo showed an order of magnitude increase when the temperature was increased from 40 to 90 oC for low q/S values. The activation energy (Ea) of the reactions shows that the dissolution process is a surface phenomenon. At 90 oC the net effect of the processes governing MW25 dissolution led to the preferential release of boron and alkali metals relative to the release of Si during the transient dissolution stage, accompanied by an increase in the concentration of silicic acid. This suggests that the solution activity of silicic acid at a higher temperature has a weak influence on the release of the mobile elements. The forward dissolution rate measured in LSSR simulant groundwater was found to be slightly higher than that measured in HSR simulant groundwater. The dissolution behaviour of MW25 in both groundwaters is consistent with its behaviour in deionised water at 40 oC, with the dissolution rates of elements increasing as flow rates were increased. However, forward dissolution rates measured in the simulant groundwaters were lower than the forward dissolution rates measured in deionised water under these experimental conditions. This is attributable to the interaction of the components of the simulant groundwaters with the glass, as revealed by post-reaction surface analyses, and a consequential lower alkalinity of the leachates collected in the experiments with simulant groundwater than in deionised water. Reactive chemical transport simulations of waste glass dissolution and radionuclide release in a hypothetical near field were conducted over a time span of a million years with GoldSim. The results showed that enclosing the waste glass in a steel canister covered by a copper canister and emplacing the waste package in a granite host rock is optimal for the long-term isolation of the radionuclides. The waste glass was found to play a significant role in the overall performance of the near field. This study features a new method for estimating the surface area of reacted glass powder more accurately than the geometric surface area estimate, which is the preferred standard method among researchers.