Based on the only reliable temperature profile available in the deepest ∼10 km layer above Venus' surface (obtained by the VeGa-2 landing probe), the mixing conditions of the main constituents of Venus's atmosphere, CO 2 and N 2 , have been questioned. In this work, we report the results of a series of experiments that were done in the GEER facility at Glenn Research Center to investigate the homogeneity of CO 2 /N 2 gas mixtures at 100 bars and temperatures ranging from ∼296 K to ∼735 K. When the gas mixtures are initially well-mixed, separation of the two gases based on their molecular mass does not occur over the time scales observed; although, small systematic variations in composition remain to be fully interpreted. However, when N 2 is injected on top of CO 2 (layered fill), the very large density ratio makes it more difficult to mix the two chemical species. Timescales of mixing are of the order of 10 2 hours over the height of the test vessel (roughly 60 cm), and even longer when the gas mixture is at rest and only molecular diffusion is occurring. At room temperature, close to the critical point of the mixture, large pressure variations are obtained for the layered fill, as N 2 slowly mixes into CO 2. This can be explained by large density variations induced by the mixing. For conditions relevant to the near-surface atmosphere of Venus, separation of CO 2 and N 2 based on their molecular mass and due to physical properties of the gas mixture is not demonstrated, but cannot be firmly excluded either. This suggests that if the compositional vertical gradient deduced from the VeGa-2 temperature profile is to be trusted, it would most probably be due to some extrinsic processes (not related to gas properties, e.g. CO 2 volcanic inputs) and large mixing time constants.