Preliminary results from mineralogical and chemical investigation of the prototype repository at Aespoe hard rock laboratory, Sweden

Document available in extended abstract form only. The Prototype Repository is a large-scale experiment installed at Aespoe Hard Rock Laboratory, Sweden. The experiment consists of six full-scale deposition holes at 450 meters depth. Each deposition hole is installed with full-scale bentonite buffer and a full-scale copper canister are installed. The canisters are equipped with heaters to simulate the heat generation from radioactive waste. The experiment is divided into two sections; the inner section consists of four deposition holes and the outer section of two deposition holes. The deposition tunnel is backfilled with a mixture of crushed rock and bentonite. The experiment also includes two concrete plugs at the end of each section. Furthermore, sensors are installed in the rock, backfill and buffers to monitor the pressure build-up and wetting of buffer and backfill. The installation of buffers, canisters with heaters, packing of backfill, and also the making of concrete plugs were done during the period 2001-2003. Heating of the canisters in the two sections was started successively after the filling of the tunnel had reached the deposition holes. The heaters in canister 1 were activated in October 2001 and the heaters in canister 6 were activated in May 2003. Natural wetting from the rock of buffer and backfill continued from that date until the dismantling stared at the end of 2010. The heaters in the two canisters in outer section were switched off in January 2011 and July 2011, respectively. During 2011 the bentonite was excavated and sampled in vacuum sealed aluminium laminate bags prior to the analysis. The hydro-mechanical, chemical, mineralogical and microbiological properties of the bentonite buffer are investigated. The formation water and gas samples are analysed and also the rock and the canister. Here we present preliminary results from the mineralogical and chemical investigations of selected blocks from the bentonite buffer material excavated from the outer section, together with some visual observations of the buffer. Early observations of the buffer showed a dark discoloration of the buffer in the innermost 1 cm towards the canister. It was noted that this discoloration disappeared upon drying in air. This suggested the possibility of a redox process, such as the oxidation of iron in the bentonite. The iron-redox chemistry was studied by X-ray absorption spectroscopy, Moessbauer spectroscopy and wet chemical analysis. All three methods showed a significant increase in Fe(II) in the warmer part of the clay towards the canister. With X-ray absorption spectroscopy also copper was studied (Cu K edge) in the samples close to the canister. The signal-to-noise ratio with Cu was much lower than with Fe due to the very low levels present. However, the curve did not match with the Cu(II)-references investigated. Possibly a Cu(I) phase was present or a Cu-sulphide. In a very small part of the bentonite - canister interface, cracks were found with a white crystalline precipitate. X-ray diffraction (XRD) identified the precipitate as gypsum, a mineral found naturally in the bentonite. Due to high water solubility, calcium sulphate can be transported from the peripheral parts during the saturation of the bentonite and precipitated towards the hot canister. Small amounts of anhydrite were found in one of the other blocks investigated, possibly indicating a locally drier or hotter environment. XRD indicated an increased basal spacing of the montmorillonite, indicating replacement of some Na by Ca, which can be expected due to the composition of the Aespoe ground water. No change was observed in the hk reflection at 4.5 A or in the 060 reflection at 1.5 A, which indicated that no significant mineral transformation had occurred. The clay fraction was intercalated with ethylene glycol and the amount of illite interlayers in the clay mineral was evaluated using the Moore and Reynolds method, and no significant changes were detected. Infrared spectroscopy and chemical analysis indicated an increased amount of organic material in samples close to the canister, probably due to the lubricant oil used during the pressing of the buffer blocks. Optical microscopy proved to be a useful complement to XRD in mineral identification. No significant changes in cation exchange capacity was observed, except a small decrease in the innermost sample. Transport of calcium sulphate and cation exchange has also been observed in previous experiments. However this is, to our knowledge, the first time a significant increase in Fe(II) is noticed in this kind of test. This has to be carefully evaluated to identify the reductant in the system or the source of additional Fe(II). A large scale experiment of this size is very complex and several sources must be evaluated. Final reporting of the results will be done in 2013