THE EFFECT OF IRON ON MONTMORILLONITE STABILITY. (I) BACKGROUND AND THERMODYNAMIC CONSIDERATIONS

Abstract

It is envisaged that high-level nuclear waste (HLW) will be disposed of in underground repositories. Many proposed repository designs include steel waste canisters and bentonite backfill. Natural analogues and experimental data indicate that the montmorillonite component of the backfill could react with steel corrosion products to produce non-swelling Fe-rich phyllosilicates such as chamosite, berthierine, or Fe-rich smectite. In K-bearing systems, the alteration of montmorillonite to illite/glauconite could also be envisaged. If montmorillonite were altered to non-swelling minerals, the swelling capacity and self-healing properties of the bentonite backfill could be reduced, thereby diminishing backfill performance. The main aim of this paper was to investigate Fe-rich phyllosilicate mineral stability at the canister-backfill interface using thermodynamic modelling. Estimates of thermodynamic properties were made for Fe-rich clay minerals in order to construct approximate phase-relations for end-member/simplified mineral compositions in logarithmic activity space. Logarithmic activity diagrams (for the system Al2O3-FeO-Fe2 O3-MgO-Na2O-SiO2-H2O) suggest that if pore waters are supersaturated with respect to magnetite in HLW repositories, Fe(II)-rich saponite is the most likely montmorillonite alteration product (if fO2(g) values are significantly lower than magnetite-hematite equilibrium). Therefore, the alteration of montmorillonite may not be detrimental to nuclear waste repositories that include Fe, as long as the swelling behaviour of the Fe-rich smectite produced is maintained. If fO2(g) exceeds magnetite-hematite equilibrium, and solutions are saturated with respect to magnetite in HLW repositories, berthierine is likely to be more stable than smectite minerals. The alteration of montmorillonite to berthierine could be detrimental to the performance of HLW repositories. © 2006 Published by Elsevier Inc.