Sorption of Cs(I), Ni(II), Eu(III), Th(IV) and U(VI) on rock samples of Opalinus Clay and confining units from deep bore holes at the potential siting regions Jura Ost, Nördlich Lägern and Zürich Nordost: measurements and predictive sorption modelling

pdf NTB 23-01 Sorption of Cs(I), Ni(II), Eu(III), Th(IV) and U(VI) on rock samples of Opalinus Clay and confining units from deep bore holes at the potential siting regions Jura Ost, Nördlich Lägern and Zürich Nordost: measurements and predictive sorption modelling (7.49 MB)

In Stage 2 of the Sectoral Plan for Deep Geological Repositories (SGT), three regions, namely Jura Ost (JO), Nördlich Lägern (NL) and Zürich Nordost (ZNO) have been identified as suitable to host the deep geological repository for radioactive waste. To assess the potential of radionuclide (RN) retention and perform safety analyses during SGT Stage 3, sorption data bases covering the various combinations of porewater and mineralogical composition for the argillaceous rock formations in these three regions are required.

Because radionuclide sorption in complex geologic rock formations is inherently complex to be easily predicted by considering all processes, a methodology that accounts for the major retention pathways is needed. The simplified component additive modelling approach, known as the "bottom-up" approach, assumes that retention in clay-mineral-rich environments is controlled by sorption onto 2:1 clay minerals (e.g., montmorillonite, illite or illite/smectite mixed layers). This approach allows the derivation of solid-liquid distribution coefficients (Rd values) for argillaceous rocks over a wide range of porewater and mineralogical compositions, using thermodynamic sorption models specifically developed for individual clay minerals.

In this report, the predictive capability of the ClaySor model, which is an upgraded version of the generalised Cs sorption (GCS) model and the 2-site protolysis non-electrostatic surface complexation and cation exchange (2SPNE SC/CE) model developed for illite is assessed on experimental sorption data obtained on rock samples from the three siting regions. Sorption isotherms for Cs(I), Ni(II), Eu(III), Th(IV) and U(VI) were measured on drill core samples from the boreholes Bözberg-1-1 in JO, Bülach-1-1 in NL and Trüllikon-1-1 in ZNO in their corresponding porewater compositions and compared to the Rd values calculated with the bottom-up approach. A total of 100 sorption isotherms were measured. For each of these isotherms, a blind prediction was calculated by scaling the ClaySor model, in which the GCS and 2SPNE SC/CE models together with the updated PSI/Nagra thermodynamic data base (TDB) 2020 were implemented to the 2:1 clay mineral content, considering the aqueous speciation of the RNs in the porewater. A comparison between measured and predicted Rd values for each case was then made.

For most clay-mineral-rich samples, the prediction and experimental data for Cs, Eu and Th agree well. The sorption of Ni is, however, mostly overestimated in the Ni equilibrium concentration below 10^-5.5 M, probably due to the presence of competing cations such as Fe(II) or Mn(II). This was not considered in the modelling of the experimental data since the analytical determination of the competing cations, particularly Fe(II), was inconsistent. In contrast, the sorption of U is systematically underestimated, which was not the case in previous studies. The main reason is that the formation constant for Ca₂UO₂(CO₃)_3(aq) (major aqueous U complex in the porewaters) leading to U sorption reduction in the modelling, was an order of magnitude lower in the previous studies compared to the one selected in the most recent PSI/Nagra TDB 2020. Another explanation could be the reduction of small amounts of U(VI) to U(IV) by Fe(II)-containing trace minerals, which would also lead to increased sorption. In clay-mineral-poor rocks, other minerals, especially calcite, might (partly) control the retention of given radionuclides.

All in all, the bottom-up approach provides a good and conservative methodology to predict the sorption on argillaceous rocks of several radionuclides, with oxidation states from I to IV and VI. In the cases where the blind prediction underestimates the observed sorption data, the bottom-up approach gives mostly conservative R_d values for the safety assessment.