Caesium is one of the main radionuclides present in radioactive waste. 135Cs has a very long half-life (~2.3 106 yr) and Cs is, therefore, one of the main safety relevant elements (Kissane, 2009). For these reasons, Cs is one of the most widely studied radionuclides in environmental sciences. Numerous works have been devoted to developing sorption models enabling the prediction of Cs sorption in clay-rich media. These models consider the presence of different exchange sites at clay surfaces having various relative affinities for Cs and other cations present in solution (Bolt et al., 1963; Jackson, 1968; Sposito, 1980; Brouwer et al., 1983; Bradbury and Baeyens, 2000; Steefel et al., 2003; Wang et al., 2011). Illite has been identified as the main mineral responsible for the uptake of Cs due to the high affinity of its surfaces for Cs ions. Caesium sorption on illite surfaces can be successfully modelled by considering up to three different exchange sites denoted as frayed edge sites (FES), planar sites on the basal surfaces and type II sites whose morphological or structural origin is not well defined (Bradbury and Baeyens, 2000). Of these sites, frayed edge sites are responsible for the very high Cs affinity for illite surfaces at low Cs concentrations. Concepts of a radioactive waste repository in deep clay formations take advantage of this Cs retention by the geological barrier. Nevertheless, although many studies have focused on the prediction of Cs sorption and mobility in this context (Poinssot et al., 1999; Bradbury and Baeyens, 2000; Van Loon et al., 2009), few have been dedicated to the study of Cs retention by the clay-rich media under perturbations due to the presence of the repository structure, the engineered barriers and the waste containers. According to the numerous data produced on alkaline perturbation (Gaucher and Blanc, 2006; Fernandez et al., 2009; Dauzeres et al., 2010; Gaboreau et al., 2011), the use of cement materials in the repository design is assumed to alter the confining properties of the clay buffer plug. Over a long time scale, the degradation of the clays could induce a diminution of their physical and chemical properties by the dissolution of the clay minerals at high pH and by the release of competitive cations for Cs retention such as K+, mainly originating from the cement material alkali-rich pore-water. The quantification and the modeling prediction of these effects are important issues in order to attribute a retention performance to the barrier as a function of pertinent markers of the disturbance (changes in mineral phase composition, concentration levels of competing ions). The potential effect of a near-field alkaline perturbation is therefore explored in the present study. Batch experiments were conducted to measure the Cs sorption in alkaline perturbed conditions. These experiments aimed at mimicking the effect of an alkaline plume intrusion on COx with accompanying changes in pH and other solution parameters such as K concentration. The second system mimic an interface (clayrock/cement materials) during early reaction time and the last system was an assemblage of minerals representative of the concrete/clay interfaces during long reaction time. A predictive sorption model taken from the literature was applied to identify discrepancies between measured and predicted retention properties. The experimental data, for Cs sorption on Callovian-Oxfordian argillaceous rock under short term (one day) alkaline perturbation, shows that K+ competition for Cs+ sorption on illite FES sites is the main factor affecting Cs retention capacities. This competitive effect related to the intrusion of hydrated cement pore water in the argillaceous rock is well predicted by the model of Bradbury and Baeyens (2000) without considering any corrections/adjustment in the original model. It is thus possible to make reliable prediction of Cs retention level as a function of K+ intrusion in the rock. For instance, from coupled geochemistry/transport modeling K+ diffusion in the rock after 10 000 years should lead to a K+ concentration up to 20 mmol/L ten centimeter after the interface (corresponding Kd ~ 90 L/kg) whereas this concentration increase should be limited to 3 mmol/L at 0.5 m of the interface (corresponding Kd ~ 160 L/kg). For one month experiments, an additional contribution to the overall Cs sorption is measured and can probably be attributed to sorption or co-precipitation on newly formed minerals, e.g. C-S-H phases (Figure 1). Experiments on a mineralogical assemblage representative of a Callovian-Oxfordian rock/concrete interface showed that this mineralogical assemblage has a similar efficiency for Cs uptake as the intact clay rock. The chosen mineral assemblage is representative of a highly perturbed material and corresponds to an interface composition modeled after thousand years of cement clay interaction (Marty et al. 2009). This result shows that clay barrier material transformation under long term alkaline perturbation can lead to similar barrier performance as for undisturbed barriers with regard to Cs retention.