Effects of salinity and high ph on crushed rock and bentonite - experimental work and modelling

In the research work presented here the results obtained in Finland in the three-year ECOCLAY II project are presented. Two types of experiments were carried out; flow-through experiments, which simulated one suggested backfill concept for the disposal tunnel of spent nuclear fuel consisting of compacted bentonite and crushed rock, and batch experiments including sorption studies. Sorption was studied with two radionuclides, 45Ca and 22Na. Both types of experiments were conducted in anaerobic CO2-free atmosphere. In both experiments the solid materials used were the same; MX-80 bentonite and crushed crystalline rock powder from Olkiluoto disposal site in Finland. The solid materials were attacked by simple simulated groundwater compositions; synthetic aqueous solutions covering fresh alkaline (pH=12.5), saline alkaline (pH=12.5), and saline hyperalkaline (pH=13.5) conditions in order to simulate the effects of cement in the repository, but also fresh (pH=8.8) and saline (pH=8.3) conditions were included for comparison purposes. Chemical changes were analysed in the various solution phases and solid phases were characterised prior to and after the solution attacks (both experiments included). The experimental results were bases in the modelling task containing both forward and inverse modelling. The flow-through experiment results showed clear increasing trends of the propagation of the attacking solutions in bentonite in the case of saline and saline alkaline experiments. Trends in the pH value and/or the element concentrations (Na, Ca, K, Mg, Si, SO4, Cl) of the out-flow solutions progressed with increasing experimental time reflecting the alteration processes occurring in bentonite (e.g., cation exchange, dissolution). Only in the saline alkaline experiment the out-flow solution pH showed an increasing trend with increasing experimental time (9.5 at completion), while in the other two experiments constant pH levels were maintained (9.5 for fresh and 8 for saline experiment). In the saline alkaline experiment the pH value in bentonite porewater was highest (around 11) at the in-flow interface and showed a steep decreasing trend towards the out-flow interface (down around 8). Corresponding trends were seen in the exchangeable cation (Na, Ca, K and Mg) capacities. Based on the results obtained it could be concluded that the alkaline attack had affected the bentonite column to some extent throughout the entire column. Sorption of 22Na was low on crushed rock (Kd 0.02-1.2x10-3 m3/kg) and bentonite (Kd 0.65-2.1x10-3 m3/kg). An exception was the fresh alkaline experiment with bentonite (Kd 1.5-2.7x10-2 m3/kg). High sorption of 45Ca on crushed rock was measured only in the saline hyper alkaline experiment (Kd 0.81-9.1x10-3 m3/kg), whereas very high sorption on bentonite was observed (Kd 0.015-100 m3/kg) in all three alkaline experiments. Effect of the initial solution Na and Ca on sorption was also studied. Only minor effects were observed. Main mineral alterations of the solids were detected in the batch experiments. By alkaline attack Nax-montmorillonite was partially altered to Nax-bedellite, and bentonite interactions with alkaline solutions produced CSH-phases (less in flowthrough experiments) structurally analogous to 14Aa-tobermorite. In the flow-through experiments the saline attack introduced smectite alteration towards Cax-montmorillonite and the fresh solution attack caused some purification within the smectite layers. The experimentally observed uptake of Na and Ca was adequately explained by the simple thermodynamic model developed for montmorillonite, provided that Ca uptake occurred predominantly by precipitation of Ca-bearing mineral phases. In the inverse modelling some problems were encountered, e.g., multiple independent sources and sinks for the same element could not be handled and the inverse cation exchange calculations failed. (orig.)