REDEVA

REDEVA is an INTERREG funded project granted to Stefan Bruns, supervised by H.O. Sørensen and co-supervised by Dag K. Dysthe (University of Oslo, Norway), running from 02/2017 to 08/2017.

Title:

“Resolving density variations in chalk percolated by CO2 saturated brine using X-ray microtomography at multiple energies”

Description:

Time-resolved X-ray microtomography combined with a flow cell setup allows studying dynamic pore scale processes like the formation of heterogeneous dissolution patterns in chalk upon injection of CO2 acidified brine. This is relevant to judging the suitability and long term safety of potential geological carbon storage sites in the North Sea region.

The limitation of absorption based X-ray imaging is that it only provides information on local electron density of the imaged materials and chalk is a very fine grained material. Consequently, the observed image intensities are always a mixture of varying contributions of solid calcium carbonate and seawater with varying CO2 concentration, i.e., standard absorption imaging can only provide a qualitative representation of the evolving microstructure (Fig. 1A). In the REDEVA project we explore the potential of adding an additional dimension of information to the reconstructed water-gas-solid system by using microtomography at multiple X-ray energies. This provides an energy dispersive footprint for every voxel with a unique solution to its density and the elemental composition of the materials mixed within (Fig. 1B).

Figure 1: Horizontal slice through a typical reconstructed chalk sample at micrometer resolution (A) and tentative solution to the distribution of CO2, water and material in a false colored vertical slice (B). Green denotes gas phase, blue water phase and white carbonate phase.

The key challenge of REDEVA is to provide a quantitative reconstruction from incomplete and imperfect datasets. A flow cell setup imaged with micrometer resolution only allows for field of view tomography, i.e., the datasets are incomplete. Furthermore, synchrotrons need to reinject electrons permanently to maintain a constant energy flux. Variations in energy density cause temperature fluctuations in the imaging setup. At micrometer resolution the consequences are visible as non-linear time-dependent artifacts that are only insufficiently addressed by traditional sinogram normalization and correction techniques. REDEVA addresses these artifacts with a self-optimizing dynamic flatfield correction (Fig. 2).

Figure 2: A sinogram acquired from a microtomography experiment with a flow cell setup corrupted by time-dependent flatfield variations normalized by ‘traditional’ linear flatfield interpolation (A) and self-optimizing dynamic flatfield correction (B).