Uranium associations and migration behaviour at the Needle’s Eye natural analogue site in SW Scotland
This thesis investigated uranium (U) migration behaviour at the Needle‟s Eye natural analogue site, located close to Southwick Water, South West Scotland. The results of this study are important for the prediction of U behaviour in the far-field environments of nuclear waste repositories over long time-scales. The Needle‟s Eye natural analogue site was selected because the processes involved in U mobilisation, the direction of water flow and the extent of retention of uranium in peaty soils had already been identified. To this end, previous results demonstrated that groundwater passing through the mineralisation oxidized U and transported it to the peaty area, where 80-90% of the released U has been retained. Sequential extraction of the peaty soils indicated that more than 90% of the solid phase U was bound to the organic fraction. However, in-depth characterisation of U associations within the soil porewaters and the peaty soils at this site was lacking. Therefore, the processes controlling the migration of uranium within this organic-rich system were the main focus of this study. There were five sampling trips carried out from 2007-2011, in which cave drip waters, bog waters and surface soil and soil core samples were selectively collected for analysis by a range of methods described below. The cave drip waters emerging from the mineralisation were oxidizing and slightly alkaline (7.6-7.8), U was mainly in truly dissolved (<3 kDa) forms (Ca2UO2(CO3)30, CaUO2(CO3)32- and UO2(CO3)22-). It is known that the formation of the ternary Ca-UVI-CO3 complexes inhibits the reduction of U and so it is likely that it is UVI that is present within the peaty soils and their associated porewaters. Sampling trip 1 quantified the U concentrations in cave waters and soil core porewaters. By 30 m from the cave, U concentrations in the soil porewaters had decreased by a factor of ~10. Ultrafiltration fractionated the colloidal fraction (3 kDa-0.2 μm) into large (100 kDa-0.2 μm), medium (30-100 kDa) and small (3-30 kDa) colloidal fractions. It was found that U was mainly associated with the large colloid (100 kDa-0.2 μm) but, with increasing distance from the mineralisation, the U distribution became bimodal with both large and small fractions being equally important. Iron (Fe) was exclusively associated with the large colloid fraction in the peaty soil porewaters. Gel electrophoresis and gel filtration, applied to study the interactions of U (and other elements) with humic substances (HS), showed that the associations were quite uniform with increasing depth of the cores and increasing distance from the U mineralisation. Uranium (and other elements including Fe) was associated with the largest humic molecules. Sampling trip 2 involved collection of three more soil cores and ultrafiltration again fractionated the total dissolved porewater into large, medium and small colloids. This time, the truly dissolved (<3 kDa) fraction was also analysed. Again, U was mainly associated with the large colloidal (100 kda-0.2 μm) fraction. With increasing distance and increasing depth, U was still predominantly associated with the large colloidal fraction, but the importance of the truly dissolved (<3 kDa) phase could not be neglected. At the same time, Fe was also mainly associated with the large colloidal fraction. The remainder of the experimental work on samples from trip 2 focused on determining the importance of U associations with both Fe and humic components of the solid phase. Sequential extraction of the whole soil mainly targeted different iron phases and found that U was mainly released in the sodium acetate and sodium dithionite solutions, which indicated U was associated with (i) Fe carbonates; and (ii) crystalline Fe oxides (e.g. goethite, hematite, and akaganetite). However, very little Fe was extracted in the “carbonate-bound” fraction and separate experiments showed that U was not associated with Fe carbonates but instead had been released from the surfaces of HS and humic-bound Fe surfaces. XRD spectroscopy showed that mineral compositions were in reasonable agreement with the sequential extraction results and SEM-EDX analysis indicated that U in the soil was generally not present in crystalline form, as only two particles with high U content were found after 4-hour searching. Exhaustive extraction of HS showed that >90% U was associated with organic substances, in agreement with previous work and novel experiments involving gel electrophoresis in conjunction with sequential extraction was used to study the relationships between U, Fe and the HS. It was demonstrated that ~20-25% U was weakly held by the HS or at humic-bound Fe surfaces, ~45% was incorporated into crystalline Fe oxides which were intimately associated with HS and the remainder was in the form of strong U-CO3-humic complexes. In sampling trip 3, U migration behaviour in the soil porewaters was the focus. A 30-m transect line, comprising seven0-5 cm soil samples, starting at the cave and passing through the peaty area towards the Southwick Water, was established. Soil porewaters from these surface soils were fractionated into colloidal (3 kDa-0.2 μm) fraction and truly dissolved (<3 kDa) phase. There was a major change in U speciation, from Ca2UO2(CO3)3 0,CaUO2(CO3)32- and UO2(CO3)22- in the truly dissolved fractions of waters close to the cave to a predominant association with the highly coloured colloidal fractions as soon as the boggy area was reached. With distance through the boggy area, it was clear that the colloidal U was being incorporated into the solid phase since porewater concentrations had decreased ~100-fold by 30 m from the cave. Ultrafiltration in conjunction with acetate extraction was then used to extract U from the porewater colloids isolated from a soil core (20 m from cave). In the organic-rich portion of the core (0-30 cm), ~60-70% U was colloidally associated and ~85-95% of this U was extracted from the colloidal fraction. This indicated that the interactions between U and the porewater colloids were weak. In sampling trip 4, U associations in the porewater colloids were still the main focus. Gel filtration of porewater colloids confirmed that U, Fe and humic colloids were intimately associated. It was concluded that although U in the cave drip water was mainly in truly dissolved forms, weak U----humic/Fe colloids were formed immediately when U entered the peaty area. In sampling trip 5, results for soil core porewaters showed that Fe in the whole core was mainly in the form of FeII. Thus strongly reducing conditions prevailed through the core which was situated within the peaty area. Combining the results from the five sampling trips, three zones within the peaty area were distinguished. Zone I was characterised by extremely high concentrations of dissolved HS and this was where the change in U speciation from dissolved to colloidal forms took place. Zone II contained most of the soil cores collected during this study and was characterised by strongly reducing conditions and moderate concentrations of HS. Colloidal U was removed to the solid phase as waters flow through this area. Zone III marks the transition to the saltmarsh. Focusing on Zone II, a conceptual model of U behaviour was developed: upon entering the peaty area, U is weakly held by very large humic-Fe colloids. These colloids are removed to the solid phase and over time the associations of U are transformed; some becomes incorporated into stable humic-bound crystalline oxides as a result of redox cycling of Fe, some becomes strongly complexed to HS and the remainder is weakly held by the HS and/or humic-bound Fe surfaces. The crystalline Fe oxides were transformed to Fe sulfides below 30 cm depth but the associated U was not transferred to these sulfides. Instead the weak associations became more important. In the wider context, since only UVI forms soluble complexes with acetate, UVI does not appear to be reduced even under the strongly reducing conditions encountered within waterlogged organic-rich soils. Initial interactions between UVI and porewater colloids appear to be weak but stronger interactions such as incorporation into Fe phases and complexation by HS occur once the colloids and associated U are removed to the solid phase. Waterlogged organic-rich soils appear to be a long-term sink for U but changing climatic conditions leading to the drying out of such soils may ultimately release U in association with smaller, more mobile organic-rich colloids.