Investigation of Percolation-Driven Fluid Transport in Rock Salt under Repository-Relevant Conditions (PeTroS)

Abstract

Abstract. According to the state of the art in mining and repository research, undisturbed rock salt is impermeable to fluids. Hence, rock salt formations are considered as host rock for nuclear waste repositories. Viscous, polycrystalline salt rock with low humidity contains no connected pore spaces. Two mechanisms are known for fluid transport: (a) damage due to large deviatoric and tensile stresses generates dilatancy, and hence permeability. (b) Fluid pressure exceeding the minor principal stress can open pathways (pressure-driven percolation, Minkley et al., 2013). To assess barrier integrity of rock salt barriers, the dilatancy and minimal stress criteria have been derived. Recently (Ghanbarzadeh et al., 2015; Lewis and Holness, 1996), high permeabilities in rock salt have been postulated under certain conditions. In particular, at high stresses and temperatures, including possible repository conditions, rock salt is claimed to develop a connected, thus permeable, pore space. In the PeTroS project (Minkley et al., 2020), we investigated fluid transport in the supposedly permeable region. Five points in pressure-temperature space were defined – pressures of 18 and 36 MPa, temperatures of 140, 160, and 180 ∘C. At each point, experiments with both nitrogen and saturated NaCl solution (brine) were performed. Samples were prepared from natural rock salt of German Zechstein formations, both bedded and domal salt. Sample material was generally relatively pure rock salt with minor impurities. Cylindrical samples (diameter 100 mm, length 200 mm) were loaded in a triaxial (Kármán) cell. Fluid pressure was applied to a central pressure chamber; any transmitted fluid was collected and extracted at the secondary side. The entire cell was heated to the specified temperature. Experiments generally comprised an isotropic phase (several stages of fluid pressure almost up to the confining stress) and a fluid breakthrough phase (lowering of axial stress by strain-controlled extension). After the test, a coloured tracer fluid was injected to visualise fluid discharge points. Fluid breakthroughs with fluid pressure above the minor principal stress were observed at all five pressure-temperature conditions. Some samples showed an approximately Darcian flow at fluid pressure below the minor principal stress, with permeabilities in the order of 10−22 m2, as is regularly observed due to the small size and initial damage from sample preparation (Popp et al., 2007). Tests consistently showed a gradual decrease of flow rate, i.e. reduction of the initial damage. A stable permeability over longer times, as would be expected due to the formation of a connected pore space network, was not observed in any of the experiments. Intriguingly, experiments with brine showed no initial permeability even though the wetting fluid should plausibly favour the formation of a stable connected pore network. Predictions of the static pore scale theory (Ghanbarzadeh et al., 2015) could thus not be confirmed. Regarding repositories for heat-generating waste, it can be concluded that from a geomechanical point of view, the dilatancy and minimal stress criteria are the relevant criteria for barrier integrity even at higher pressure and temperature.