Crystalline rocks are considered as adequate host formations for deep geological repositories of radioactive waste. A sound estimation of retention parameters in these rocks, for a defendable choice of distribution coefficients, Kd, values for performance assessment calculations, strongly requires a better insight on radionuclide retention processes. Furthermore, for such complex and heterogeneous systems, the main sources of uncertainty for distribution coefficients must be evidenced, evaluated and quantified in a precise way.
Many data on Cs sorption onto granite of various origin, were obtained over the past years in the frame of different national and international projects. More recently, new tests were carried out in a systematic way, trying to keep the experimental and chemical conditions as similar as possible, to directly compare sorption results obtained with different solids. The main objective of this document is summarizing all the data obtained and evaluating the main causes of their spreading. The effects of parameters as contact time, Cs concentration, pH and water salinity were discussed.
In all the studied rocks, Cs sorption showed a non-linear behaviour, negligible dependence on pH, and a more significant dependence on the ionic strength. These characteristics suggest ionic exchange as the most probable mechanism for Cs retention.
A simplified top-down approach was adopted to simulate the sorption data, considering that the main sorption mechanism is cation-exchange and that at least two sorption sites must exist on the solids surface. The sorption behaviour as a function of Cs concentration could be reasonably fit in all the cases, this being an important result as the one of the main sources of variability in Cs distribution coefficients is precisely its concentration. Other possible sources of uncertainty on the determination of Kd values in crystalline rocks have been outlined and discussed.