Swelling and osmotic flow in a potential host rock

Abstract Measurements of osmotic and hydraulic permeability are reported for a series of tests conducted on Opalinus Clay samples from the Mt. Terri underground research laboratory in the Jura Mountains of NE Switzerland. Osmotic flow was observed across discs of this clayshale separating 0.245 M NaCl solution from distilled water. Pressure transients monitored during constant flow rate testing were analysed to give permeability and specific storage values. The mean permeability normal to bedding of the two Opalinus Clay specimens was 7.9 × 10 −21  m 2 . The mean specific storage based on all reliable determinations was 4.1 × 10 −4  m −1 . Values calculated from the steady-state pressure gradients established during constant flow rate testing were very close to those obtained by mathematical analysis of pressure transients. The calculation of the transients was carried out using a new model of flow and solute transport which included terms for the osmotic coupling. The form of the pressure transients and the magnitude of the strain seen during the tests lead to a revision to the definition of solid phase compressibility to incorporate a term dependent upon the osmotic coupling coefficient. Steady-state osmotic flow rates were in the range 0.1–0.6 μL h −1 when the specimens were placed between a sodium chloride solution with a theoretical osmotic pressure of 1.19 MPa and distilled water. Transient flow rates were substantially larger. Membrane efficiencies were found to be relatively low, ranging from 1% to 6% (mean around 4%). The mean osmotic permeability normal to bedding was 3.5 × 10 −22  m 2 . Specific storage and pore compressibility values were substantially larger than anticipated, suggesting that the volumetric strain of the clayshale under the conditions of laboratory testing must be largely determined by quasi-elastic deformation processes such as swelling and crack dilation. To test this hypothesis, a 3-D swelling test was performed on a cubic specimen of the same material. When immersed in distilled water, the cube swelled preferentially in a direction normal to bedding. Swelling was accompanied by visible opening of cracks. Volumetric strains of the clayshale during the swelling test were of the same order of magnitude as those during permeability testing. The presence of dilated cracks could explain the low membrane efficiency, since the dissociated ions of the salt would be able to diffuse through the crack network relatively unimpeded by electrostatic interactions with the clay mineral surfaces. The high compressibility and low membrane efficiency revealed by these experiments may be a consequence of fabric damage during sampling and specimen preparation or may actually be representative of the rock characteristics within the engineering disturbed zone (EDZ) of the tunnel niches. It is entirely possible that undisturbed rock remote from the excavations displays significantly lower compressibility and higher membrane efficiency. Even though the clayshale membrane in these experiments is inefficient, the results clearly demonstrate that Opalinus Clay is capable of supporting an osmotic flow of water. This has important implications in the specification, operation and interpretation of borehole tests aimed at the hydrogeological characterisation of this potential host rock. In addition the osmotic coupling term modifies the effective diffusion coefficient which should be accounted for when undertaking safety assessments.