Erosion dynamics of compacted raw or homoionic MX80 bentonite in a low ionic strength synthetic water under quasi-stagnant flow conditions

Abstract

Bentonite is planned to be used as a backfill material in deep geological repositories (DGR) notably for those to be established in crystalline rock. Potential erosion processes have to be clearly understood in order to properly assess its long-term barrier function. A likely scenario over the lifetime of a DGR in certain geographical locations is the intrusion of low mineralized glacial melt water through low transmissivity bedrock fractures that could interact with bentonite and favor release of clay colloids. The objective of this work was to determine the dynamic of erosion under a glaciation scenario for Na-exchanged, Na-Ca-exchanged, Ca-exchanged and raw bentonite. The different bentonite types cover different potential materials used for backfill and different stages of cation loading via cation exchange processes upon contact with groundwater of different composition. To this end, erosion experiments were carried out with highly compacted bentonite confined by a porous filter (20 μm) under dynamic conditions simulating the presence of a hydraulically active fracture filled with fault gauge and fracture filling minerals.The <500 μm size fractions of untreated MX80 bentonite and after Na- or Ca-homoionization, were used. Two compacted clay pellets (1.6 ± 0.1 g/cm3) of identical composition were placed in separate compartments of double-side reactors. Post-mortem visual inspection and the observed washout of water soluble MX80 components indicated full equilibration of the bentonite with simulated groundwater taking place during the experiment. The erosion was investigated by circulating a low ionic strength (1.6 10−3 M) water at pH 8.4. Bentonite swelling was confined by stainless steel filter plates.The erosion (clay colloid detachment) via a low mineralized groundwater flow was quantified by determining an average eluted mass loss rate (AMLR). AMLR values integrated over 1466 days (4 years) erosion time for the raw MX80 bentonite, the Na-MX80 and the Na-Ca-MX80 are at 0.019 ± 0.003 kg/y/m2, 0.245 ± 0.007 kg/y/m2 and 0.10 ± 0.02 kg/y/m2, respectively. Only by the sensitive ICP-MS analysis of clay-bound Th and U a tiny detachment of particles from Ca-MX80 could be identified, which shows the highest resistivity against erosion (0.004 ± 0.002 kg/y/m2). The experiments pointed to the swelling pressure as main driving force for colloid generation. A possible inhibiting effect of accessory minerals on erosion can also be deduced from this study. Assuming erosion via a groundwater flow from a single fracture intersecting a deposition hole filled with raw MX80 bentonite at an angle of 90o and with a fracture aperture of up to 200 μm, the present estimations indicated that under the given scenario the mass loss limit of 1200 kg bentonite will not be reached during the first 106 years.