In situ measurements of the hydration behavior of compacted milos (SD80) bentonite by wet-cell x-ray diffraction in an opalinus clay pore water and a diluted cap rock brine

Abstract

Compacted bentonite is currently being considered as a suitable backfill material for sealing underground repositories for radioactive waste as part of a multi-barrier concept. Alt-hough showing favorable properties for this purpose (swelling capability, low permeability, and high adsorption capacity), the best choice of material remains unclear. The goal of this study was to examine and compare the hydration behavior of a Milos (Greek) Ca-bentonite sample (SD80) in two types of simulated ground water: i) Opalinus clay pore water, and ii) a diluted saline cap rock brine using a confined volume, flow-through reaction cell adapted for in situ monitoring by X-ray diffraction. Based on wet-cell X-ray diffractometry (XRD) and calculations with the software CALCMIX of the smectite d(001) reflection, it was possible to quantify the abundance of water layers (WL) in the interlayer spaces and the amount of non-interlayer water uptake during hydration using the two types of solutions. This was done by varying WL distributions to fit the CALCMIX-simulated XRD model to the observed data. Hydrating SD80 bentonite with Opalinus clay pore water resulted in the formation of a dominant mixture of 3-and 4-WLs. The preserva-tion of ca. 10% 1-WLs and the apparent disappearance of 2-WLs in this hydrated sample are at-tributed to small quantities of interlayer K (ca. 8% of exchangeable cations). The SD80 bentonite of equivalent packing density that was hydrated in diluted cap rock brine also contained ca. 15% 1-WLs, associated with a slightly higher concentration of interlayer K. However, this sample showed notable suppression of WL thickness with 2-and 3-WLs dominating in the steady-state condition. This effect is to be expected for the higher salt content of the brine but the observed generation of CO2 gas in this experiment, derived from enhanced dissolution of calcite, may have contributed to the suppression of WL thickness. Based on a comparison with all published wet-cell bentonite hydration experiments, the ratio of packing density to the total layer charge of smec-tite is suggested as a useful proxy for predicting the relative amounts of interlayer and non-interlayer water incorporated during hydration. Such information is important for assessing the subsequent rates of chemical transport through the bentonite barrier.