Neutron and X-ray µ-tomography-based 3D imaging of alteration phases in faulted granodiorite at Nojima (Japan)

The study of the evolution of petrophysical properties and alteration of host rock in an active fault system is essential for understanding the mechanisms of deformation localization. The distribution of alterations is closely linked to fluid flow paths, while the formation of new deformation structures depends on the mechanical contrasts induced by these alterations. Our study focuses on granodiorite samples from a borehole drilled in 1996 at Hirabayashi by the Geological Survey of Japan, one year after the Nanbu-Kobe earthquake. Crossing the active Nojima fault, this borehole intersects the fault core at 625 m. The analyses include imaging of samples using neutron and X-ray microtomography (ILL – NeXT) and thin sections, as well as mineralogical quantification by X-ray diffraction (XRD). X-ray diffractograms on oriented slides and Rietveld analyses of XRD data acquired on disoriented powders reveal the presence of secondary mineral phases (e.g., montmorillonite, kaolinite, laumontite, siderite, ankerite), representative of different fluid-rock interaction conditions during the exhumation of the massif. Their proportions, which increase as they approach the fault, reach more than 30% of the volume of a sample at 625 m. Whereas X-ray µ-tomography imaging allows us to observe density contrasts within the samples (e.g., mineral phases and fracture network). On the other hand, neutron imaging allows us to observe the distribution of hydrated mineral phases due to the high neutron absorption coefficient of hydrogen (e.g. for 25 meV neutrons, hydrogen attenuation is 3.44 vs 0.17 and 0.11 for oxygen and silicon, respectively). Neutron and X-ray image registration in the same reference frame allows us to perform joint image segmentation, using gaussian-mixture-model to quantify uncertainties, based on the neutron and X-ray coefficients of absorption of the pre-identified mineral phases. The volumes segmented in this way enable us to (i) quantify in a non-destructive way the volume of secondary mineral phases present in the samples along the fault damage profile and (ii) obtain their spatial distribution and assess the anisotropies of distribution in relation to the deformation structures. This work will subsequently enable us to understand the impact of both the distribution of secondary mineral phases and the network of microfractures on the evolution of the petrophysical and mechanical properties of a seismogenic fault.