Valuable techniques to measure effective diffusion coefficients in porousmedia are an indispensable prerequisite for a proper understanding of the migration of chemical-toxic and radioactive micropollutants in the subsurface and geosphere. This article discusses possible pitfalls and difficulties in the classical through-diffusion technique applied to situations where large diffusive fluxes of cations in compacted clay minerals or clay rocks occur. The results obtained from a benchmark study, in which the diffusion of 85Sr2+ tracer in compacted illite has been studied using different experimental techniques, are presented. It is shown that these techniques may yield valuable results provided that an appropriate model is used for numerical simulations. It is shown that effective diffusion coefficients may be systematically underestimated when the concentration at the downstream boundary is not taken adequately into account in modelling, even for very low concentrations. A criterion is derived for quasi steady-state situations, by which it can be decided whether the simplifying assumption of a zero-concentration at the downstream boundary in through-diffusion is justified or not. The application of the criterion requires knowledge of the effective diffusion coefficient of the clay sample. Such knowledge is often absent or only approximately available during the planning phase of a diffusion experiment.