Repositories for radioactive wastes are sited in the environment with very low permeability. One of the most important processes
leading to the release of radionuclides to the environment is therefore diffusion of radionuclides in both natural and engineered
barriers. Data for its description are crucial for the results of safety assessment of these repositories. They are obtained
usually by comparison of the results of laboratory diffusion experiments with analytical and/or numerical solution of the
diffusion equation with specified initial and boundary conditions. Results of the through-diffusion experiments are obviously
evaluated by the “time-lag” method that needs for most of sorbing species unfortunately very long time of the experiment duration.
In this paper a modified approach is proposed for the evaluation of diffusion data for safety assessment, which decreases
the influence of propagation uncertainties using incorrect data and reduces time for acquiring data for safety assessment.
This approach consist in the following steps: (i) experimental measurement of material diffusion parameters under various
conditions using non-sorbing tritiated water or chlorine for which it is easy to reach conditions under which the “time-lag”
method of evaluation of the result of the through-diffusion experiment is applicable—this step provides well established diffusion
characteristics of materials for neutral species and anions, then (ii) to evaluate sorption isotherms for sorbing radionuclides
from batch experiments under conditions corresponding to composition of material pore water, (iii) to assess the values of
effective and apparent diffusion coefficients for sorbing radionuclides from well-defined diffusion coefficients of species
in free water and (iv) to verify the obtained results using relatively short-term diffusion experiments with sorbing radionuclides,
which will be evaluated using the time dependent decrease of the concentration in the input reservoir of the diffusion cell.
A numerical model of the diffusion cell can model the decrease of concentration of species. The code, which we used for this
type of evaluation of diffusion experiment, is based on the same approach to modelling that we use for modelling of the diffusion
transport in performance assessment studies. The advantage of this approach consists in the compatibility of the approaches
used for both evaluation and verification of the laboratory diffusion experiments and performance assessments and in considerable
less time needed for experiments.