Osmotic efficiency in Callovo-Oxfordian argillites: Experimental vs. theoretical models

Abstract

Argillaceous media are known for their possible semi-permeable membrane behaviour which is represented by the osmotic efficiency ε. The identification of this parameter allows a characterization of the possible coupled fluxes through such media. Osmotic experiments conducted by Andra in the Callovo-Oxfordian argillite at Meuse/Haute-Marne URL have been interpreted using a numerical model. These interpretations provide values for the osmotic efficiency in this formation. These experimental data are then compared with estimates obtained using theoretical models for the osmotic efficiency. These models are the “average frictional” model by Marine and Fritz (Marine, I.W., Fritz, S.J., 1981. Osmotic model to explain anomalous hydraulic heads. Water Resources Research 17 (1), 73–82; Fritz, S.J., Marine, I.W., 1983. Experimental support for a predictive osmotic model of clay membranes. Geochimica and Cosmochimica Acta 47, 1515–1522), the “integrated frictional” model by Bresler (Bresler, E., 1973. Anion exclusion and coupling effects in nonsteady transport unsaturated soils: I. Theory. Soil Science Society of America Proceedings 37 (5), 663–669) and the “mechanical” model by Revil and Leroy (Leroy, P., Revil, A., 2004. A triple-layer of the surface electrochemical properties of clay minerals. Journal of Colloid and Interfacial Science 270, 371–380). The values for the physical and petrophysical parameters which must be introduced in such theoretical models to match the measured osmotic efficiency are compared to the available data in the literature. The “average frictional” model does not reproduce the data with realistic parameters. Despite the satisfactory results obtained with the Bresler’s model (Bresler, E., 1973. Anion exclusion and coupling effects in nonsteady transport unsaturated soils: I. Theory. Soil Science Society of America Proceedings 37 (5), 663–669), in its current form, this model is questionable since the electrical model which is implemented is not accurate known theory. The “mechanical” model gives promising results, despite a limitation for high concentrations. We show that the pore size is a crucial petrophysical parameter for the theoretical models. The pore size required in the “mechanical” model is ∼5 nm. This value is representative of the pore size where the osmotic process can exist. This value would thus suggest a throat effect.