Montmorillonite stability. With special respect to KBS-3 conditions

Abstract

The basic advantageous properties, e.g. low hydraulic conductivity and high swelling pressure, of the bentonite buffer in a KBS- repository stem from a strong interaction between water and the montmorillonite mineral in the bentonite. Minerals similar in structure but with substantially lower mineral-water interaction exist in nature. Transformations from montmorillonite to such minerals are observed e.g. in burial diagenesis and in contact metamorphism. A thermodynamic consideration confirms that medium and low charged montmorillonite is not in chemical equilibrium with quartz. From a safety assessment perspective it is therefore of vital importance to quantify the montmorillonite transformation under KBS- conditions. Silica release from the montmorillonite tetrahedral layers is the initial process for several possible transformations. Replacement of silica by aluminum increases the layer charge but maintains the basic atomic structure. A sufficiently high layer charge results in an irreversible collapse of the clay-water structure, i.e. a non-swelling mineral is formed. Compared to other cations, potassium as counter ion leads to a collapse at lower layer charge and the produced phase is generally termed illite. Montmorillonite-to-illite transformation is the most frequently found alteration process in nature. Three different kinetic illitization models are reviewed and the model proposed by Huang et al. is considered the most suitable for quantification in a KBS- repository, since the kinetic rate expression and its associated parameters are systematically determined by laboratory work. The model takes into account temperature, montmorillonite fraction and potassium concentration, but do not include relevant parameters such as pH, temperature gradients and water content. Calculations by use of the Huang illitization model applied for repository conditions yield insignificant montmorillonite transformation also under very pessimistic assumptions. Other non-swelling minerals, e.g. chlorite and rectorite, may be formed in processes similar to illitization. However, there are no indications, neither from laboratory work nor from natural analogs, that these processes should be faster than illitization. A relatively fast layer charge increase may be caused by reduction of montmorillonite structural iron. The maximum effect of the reduction on mineral-water interaction depends on the initial layer charge and the total amount of structural iron. Migration of small cations into the montmorillonite octahedral layer reduces the layer charge at temperatures significantly higher than the maximum temperature in a KBS- repository. The extent of the migration at repository temperature remains to be determined. Processes causing breakdown of the montmorillonite atomic structure takes place under extreme conditions not expected in a KBS- repository. Substantial dissolution of montmorillonite is expected above pH 11 since new aqueous silica species start to dominate in this pH range. Recent laboratory studies also indicate that close contact with metallic iron may result in a destabilization of the montmorillonite structure. Based on the present knowledge, none of the identified transformation processes are expected to lead to significant reduction of the buffer performance. However, uncertainties remain concerning silica diffusivity, kinetics of octahedral layer charge reduction, and interaction between metallic iron and montmorillonite.