Tomographic observation of injection procedures for fracture sealing

Networks of micro fissures may exist in rock salt in the excavation damage zone. Though they will not affect the integrity of the barrier function as a whole it might be advantageous to seal such likely permeable structures by means of impregnation procedures. However, the lasting effect of such procedures is difficult to judge from simple measurements of the decrease of porosity and permeability, because spatial distribution and penetration depth of the impregnation agent should be known. We therefore suggest evaluating such impregnation methods with the help of tomographic laboratory methods of macroscopic samples.
The sample size should exceed the mean fracture distance. Therefore the size of standard drill cores with a diameter in the order of 100 mm is appropriate. This rather large size limits the achievable spatial resolution and thus the detectability of fractures with the otherwise well-established µCT-method. The low contrast between the density of the fracture fill and the solid material further complicates detection and segmentation of the structures. To overcome these issues, we apply positron emission tomography (PET). This tomographic modality yields quantitative images of the concentration of a positron emitting radiotracer with molecular sensitivity and a spatial resolution of 1 mm. We obtain images of the penetration of the labelled impregnation agent, as average tracer concentration over the voxel volume of 1 µL.
As prove of principle, we applied waterglass (sodium silcate) labelled with 18F as agent that was injected into an artificially fractured sylvinite core (Z2KSTh) from Staßfurt. The sample was structurally characterized before and after impregnation with µCT, and the flow field was determined with PET process tomography of propagating saturated NaCl-solution labelled with 18F. Although we achieved a significant decrease of permeability by waterglass injection, no significant structural effect could be seen with µCT. In contrast, PET showed a superficial covering of the sample surface with solid waterglass and penetration into larger voids with a penetration depth of a few millimetres (Fig. 1). This rather small effect is a consequence of the experimental conditions – low injection rate and low pressure – that were chosen to keep the setup simple. Although it was not intended here to estimate the effectiveness of waterglass impregnation as method for improving the geological barrier, we experienced the delicate interplay between reaction kinetics of waterglass solidification and local flow rate which makes its prediction difficult. However, we could prove the applicability of the visualization method which is uncomplicated to adapt to other impregnation agents and materials, higher pressure and injection rate.