Prior to any CO 2 geological storage operation, the caprock's ability to prevent CO 2 leakage must be carefully assessed. This ability is primarily related to the caprock's pore structure and to the interfacial properties of the caprock and the fluids in place, namely the brine that imbibes the caprock and the CO 2 stored in the underlying reservoir. This paper reports an experimental effort to characterize some of these parameters, using as a working example the carbonate-rich caprock and reservoir conditions of the Rousse depleted gas field in the South-West of France, where an estimated 120000 tons of CO 2 will be injected during two years. The parameters examined are: (i) the caprock wetting behavior in the presence of CO 2 , (ii) the caprock intrinsic (single-phase) permeability, (iii) the CO 2 breakthrough (or displacement) overpressure, i.e., the minimum pressure difference between CO 2 and brine required for CO 2 to penetrate and flow through the brine-saturated caprock, and (iv) the caprock effective permeability to CO 2 after breakthrough. The latter two parameters are indicative of, respectively, the caprock's capillary sealing efficiency and CO 2 leakage rate once breakthrough has occurred.The main observations and results are as follows. (i) The water-wet character of the Rousse caprock in the presence of CO 2 in storage conditions is confirmed through a series of contact angle measurements on substrates. (ii) Single-phase (brine) permeability coefficients measured in steady-state conditions are extremely small and strongly sensitive to the effective stress, i.e., to the difference between the confining pressure and the pore fluid pressure. They do not exceed 20-25 nanodarcy (10 -21 m 2 ) for effective stresses below 4 MPa, and 1 nanodarcy for effective stresses in the range of 10 MPa and above. (iii) Gas breakthrough in two different brine-saturated caprock samples, using either CO 2 or N 2 as the displacing gas phase, occurs in one sample for a CO 2 overpressure in excess of 7.6 MPa, and in the other sample for a N 2 overpressure in the interval of 4.5-6 MPa. In the latter sample, the effective permeability to N 2 after breakthrough is extremely low, below 1 nanodarcy; some "aging" effects are observed upon subsequent resaturations with brine and CO 2 breakthrough experiments, with a CO 2 breakthrough overpressure lower than expected and an increasing effective permeability to gas, yet still lower than 2 nanodarcy after two consecutive brine resaturations and CO 2 breakthroughs.