Retardation mechanism studies for performance assessments

Performance assessments of geological repositories for nuclear waste require information about how possibly escaping radionuclides will be transported through the bedrock. Retardation mechanism studies provide an important part of the database required. So far Finnish performance assessments have considered sorption and matrix diffusion as retardation mechanisms. In this report we discuss how the data supplied by experimental retardation mechanism studies meet those required by performance assessments in the Finnish case. Experimental structural studies on small-scale pore network are discussed in detail. The C-14-PMMA method and helium-gas methods provide porosity, diffusivity and permeability data. The effort within the Palmottu natural analogue project that aimed to locate large-scale flow channels is described briefly. Specific retardation mechanism studies are discussed. The laboratory studies cover column experiments and anion exclusion studies. Column experiments provide data on sorption and matrix diffusion. The natural analogue studies cover work done in the Palmottu, the Hämeenlinna boulder, and the Hästholmen projects. Studies at Palmottu provide in situ data about sorption and matrix diffusion. This review indicates that data required by current migration models in performance assessments can be met by existing experimental methods. But the situation changes if more detailed modelling is wanted. In that case, a strategic question must be answered: what is a reasonable level of detail and conceptual rigour in performance assessment modelling, considering the unavoidable overall uncertainties due to long time scales and heterogeneous bedrock? The results obtained so far in natural analogue studies indicate that there is room for improvements in sorption modelling. The current Kd based approach could be supported by parallel mechanistic sorption modelling. It is acknowledged, however, that mechanistic sorption modelling is possible for neither all nuclides, nor all minerals, nor all geochemical conditions. Notwithstanding this, it would nevertheless better utilise the current state-of-the-art understanding of sorption.