CO2-3D

It is now widely accepted that the continuous global warming of the atmosphere observed in the last 150 years is due to an increase in the atmospheric concentrations of greenhouse gases, and that the storage of carbon dioxide (CO2) in deep geological formations is the only feasible medium term solution. CO2 subsurface storage relies on 4 different trapping mechanisms acting at different timescales. We focus here on solubility trapping, the principle of which can be summarized as follows. In the host formation, typically a saline aquifer at a depth larger than 800m, the CO2 is supercritical, i.e., less dense and 40 times less viscous than the resident brine. After injection it migrates to the top of the aquifer, under the impermeable cap rock. The partial dissolution of this supercritical CO2 into the brine results in a mixture that is denser than the brine underneath, which triggers a gravitational instability allowing the mixture to be transported downwards through the brine. CO2 can thus be stored at the bottom of the aquifer, where it can remain sequestrated for a very long time. Due to its complexity, this instability of the mixture-brine interface coupled to a dissolution process has been mostly investigated in two-dimensional (2D) geometries. The few existing three-dimensional (3D) studies, all numerical, suggest that predictions based on 2D models may be totally inadequate. Indeed, in 3D the convection rolls typical of 2D systems only occur in the upper part of the flow cell; below they give birth to plumes which cannot be described by a 2D geometry. This remains to be demonstrated through experiments, since all numerical simulations have been performed using continuum/Darcy scale models, thus discarding all the physics potentially occurring at the pore scale. However, 3D pore scale experiments are very challenging since they require (i) a porous medium that is transparent when saturated with the flowing fluids, and (ii) an imaging system able to capture the entire 3D structure of the convection rolls/plumes. No such experiment addressing convective dissolution has been reported in the literature up to now. We propose to investigate the gravitational instability and the 3D convection that ensues, using novel experimental setups relying on the matching of the refractive indexes of the solid and fluid phases, and providing a full 3D pore scale characterization of the concentration and velocity fields. The convective flow will first be studied in a confined cylindrical geometry in order to accurately characterize the dynamics, amplitude and stability of a solitary plume. Other setups will then allow us to investigate the dynamics of many interacting plumes. A heuristic model of the trapping efficiency will also be developed based on the experimental findings. We thus expect to characterize the geometry of the gravitational instabilities (in particular, their 2D or 3D nature), the onset time for linear and non-linear instabilities, the characteristic travel times of the brine/CO2 mixture within the plumes, and the CO2 flux, as a function of the various system parameters. The project's ultimate goal is the upscaling of the experimental results to the formation scale in order to predict the rate of CO2 storage by convective dissolution in aquifers of various properties. As the extrapolation from 2D to 3D regarding the efficiency of subsurface storage is currently quite speculative, we expect the outcome of the project to impact the scientific community significantly, and aim at providing predictions usable by geoengineers. This challenging task will be undertaken by two research teams/partners with complementary expertise in different subfields of Fluid Mechanics. Preliminary setups existing at each of the partners' laboratories have demonstrated the technical feasibility of the experimental approach. The cost consists mostly in the funding of a 1.5 year postdoc position for each of the partners.