Technical Report NTB 03-06

Project Opalinus Clay: Integrated Approach for the Developmentof Geochemical Databases Used for Safety Assessment

Chemical retention plays a central role in the Swiss repository concept for spent fuel/highl-level waste (SF/HLW) and intermediate level waste (ILW). Chemical retention is taken into account in the safety assessment calculations by applying the concept of solubility limits and Kd values for the safety-relevant nuclides. The necessary data were compiled in five geochemical databases, the derivation of which is described in detail in the corresponding reports (Berner 2002a; 2003; Bradbury & Baeyens 2003a and b; Wieland & Van Loon 2002).

The elaboration of the geochemical databases (GDBs) was done by a team of scientists from the Paul Scherrer Institute and the Safety Assessment Group at Nagra in a two years' effort based on many years of extensive scientific investigations. An integrated approach was applied, which was based on the principles of chemical thermodynamics, sound experimental sorption and diffusion data and expert judgement. A consistent procedure and a number of quality assurance measures contributed to obtaining high quality retention data. A strong emphasis was on the derivation of transparent and traceable "best estimate" data and associated uncertainties.

The applied methodology can be separated into three parts. The first part consisted of elaborating the geochemical foundations, which included a detailed update of the Nagra/PSI thermodynamic database, derivation of the geochemical in-situ boundary conditions (e.g. pH and Eh) in the different compartments and generation of experimental sorption data in clay and cement systems. The second part involved the derivation of scientifically sound retention data, i.e. the solubility limits and sorption values for safety-relevant radionuclides under relevant repository conditions. The last part in the overall procedure involved the critical evaluation and, where necessary, adaptation of the values for their use in safety assessment calculations. This also included comparison with recent databases from other countries and evaluation of natural analogue data.

The solubility limit databases for the canister-bentonite and cementitious environments show reasonable agreement with other databases from recent safety assessments in spite of the differences in the underlying thermodynamic data and assumed geochemical conditions.

Our approach for deriving Kd values and apparent diffusion coefficients in the bentonite and the clay host rock differs from those applied in most other assessments. Whereas the latter are principally based on diffusion measurements in compacted clays, we systematically derived Kd values from well-controlled batch experiments and adapted these to the compacted in-situ conditions. Nevertheless, the proposed Kd and Da values agree fairly well with those used in the other assessments, except for tetravalent species. This agreement is further supported by comparison of sorption values derived in batch systems with those obtained from Japanese diffusion measurements on Kunigel bentonite. This result strongly suggests that compaction and swelling do not have a strong effect on the retention properties of the clay.

In general, higher sorption and lower Da values for tetravalent metals are used in our assessment for the bentonite near field and clay host rock relative to other assessment studies. This discrepancy is particularly large in the case of the redox-sensitive Tc(IV), U(IV) and Np(IV) species. The reasons for this are not clear, but might arise from (i) the more conservative treatment of uncertainties in other assessments, (ii) poorly constrained redox conditions in diffusion experiments, or (iii) extrapolation of batch sorption data to compacted in-situ conditions. The last point, however, is rather unlikely from the indications pointed out above. In case of the cementitious near field the difference in Kd values between our and other assessments is less pronounced than for the clay cases.

A major identified uncertainty is related to the porewater composition in the clay compartments. This was accounted for in the derivation of GDBs by including a large range of pH/pCO2 and Eh in the uncertainty treatment of solubility limits and Kd values. Furthermore, important conceptual model uncertainties identified in the present work include the effect of carbonate on solubility and sorption of tetravalent metals, the redox chemistry of Pu, the nature and crystallinity of solubility-controlling phases, kinetics of redox-sensitive radionuclides and physico-chemical properties of water in compacted clays. Since the derived databases are based on a reference temperature of 25 °C, further uncertainty arises from the slightly higher temperature (≈50 °C) predicted for the conditions relevant to safety assessment. From a preliminary assessment we expect, however, the temperature uncertainty to be of less relevance relative to those mentioned above.

The conceptual model uncertainties in the retention data were accounted for in the geochemical databases by the conservative approach for deriving the uncertainty range and for the estimation of the pessimistic values. In addition, the high uncertainty related to the transport behaviour of redox-sensitive RN was implicitly considered by analysing the effects of an oxidising near field as a "what-if?" case.