Systematically quantifying core-scale mineral heterogeneity as a controlling factor for permeability evolution

Quantifying heterogeneity in complex subsurface systems is essential to improving our understanding of long-term fate and transport of CO₂ injected into deep saline aquifers. Existing power laws, such as the Kozeny-Carman equation, have been used widely in the literature to describe porosity-permeability relationships and alterations resulting from subsurface fluid injection but fail to capture complex geochemical-geomechanical changes in pore structures of multimineral systems. This study presents a systematic approach to quantifying heterogeneity with respect to bulk mineral reactivity and reactive mineral spatial distribution in order to improve power law relationships governing porosity-permeability evolution. A set of heterogenous carbonate and sandstone cores with varying calcite distribution and abundance (0 to 100% calcite) will be examined via x-ray computed tomography (XCT), back-scattered electron microscopy (BSE), energy dispersive spectroscopy (EDS) and x- ray diffraction (XRD) to determine the spatial distribution of calcite prior to and after exposure to flow of acidic water. Bulk mineral reactivity, quantified as the fraction of calcite in each sample, will be determined by XRD. Calcite distribution will be determined by combining BSE/EDS spectroscopy with 3D XCT core data to associate known mineralogy with corresponding XCT grey scale pixel values. Cores will be segmented into pores, calcite and non-calcite fractions using the Weka 3D machine learning segmentation algorithm in Fiji, constructing a 3D mineral map of each core. Cluster statistics are used to identify reactive mineral (calcite) distribution where spatial connectivity is characterized by correlation length (λ), which quantifies the average link length connecting the largest clusters of the same mineral. Correlation lengths of each core, quantified before and after reactive fluid flow, will be used to develop a heterogeneity-based power law coefficient.