Compaction localization in 4D imaged by X-ray Computed Tomography and Digital Volume Correlation

Understanding compaction localization in porous limestone in the laboratory is significantly more challenging than in sandstone because of the lack of consistent acoustic emission activity in carbonate samples. Previous studied have therefore relied on X-ray Computed Tomography imaging (CT). The first unambiguous evidence of compaction band development in limestone was provided by Huang et al. (2019), who performed synchrotron in situ CT imaging during shear-enhanced compaction in a sample of Leitha limestone. This sample was deformed in the HADES rig at the European Synchrotron Radiation Facility, in dry conditions and at a confining pressure of 20 MPa. In this study, we analysed this data set using Digital Volume Correlation (DVC). Not only could we use DVC to characterize quantitatively the spatiotemporal development of displacement and strain, we were also able to compare with direct observations to assess the stress-induced damage in multiple scales. Our new results confirm that inelastic compaction occurred in two stages in Leitha limestone: macropore collapse first and then sequential growth of compaction bands. In the pore collapse stage, DVC reveals complex and heterogeneous grain-scale strains, implying significant heterogeneity in the internal stress field. Such complexity is to be accounted for if one were to connect micromechanical and continuum models. At higher stresses, we have obtained further quantitative constraints on the spatial distribution of volumetric and shear strain during the growth of compaction bands. Our results demonstrate that compaction banding in Leitha limestone can be analysed as a bifurcation phenomenon, that would typically occur preferentially in zones of high porosity. The displacement field inferred from DVC revealed that the bands showed mostly normal displacement discontinuities, as expected for compaction bands. DVC analysis also gave more constraints on band geometric attributes. Analysis of the autocorrelation function for the strain suggested that the decay and rebound of the autocorrelation as a function of the axial separation may provide proxies for the mean width and spacing of compaction bands. The 2D autocorrelation function on the band planes also provides relevant clues on the complex sequential growths of the compaction bands.