Enhancing understanding of mineral accessible surface area to improve simulation of mineral reaction rates in porous media
Type of DegreePhD Dissertation
Civil and Environmental Engineering
MetadataShow full item record
Geological formations have great potential for large scale carbon sequestration, where the CO2 is mineralized through geochemical reactions. These reactions are rather complex and can potentially alter the formation properties including porosity and permeability. Reactive transport modeling is a powerful tool to enhance the understanding of the impact of CO2 injection on formation properties by simulating the CO2-brine-mineral reactions over both laboratory- and geological- time scales. Mineral reactive surface area (RSA) is one of the rate-controlling parameters in geochemical reactions, where 1 – 5 orders of magnitude variations in mineral reactive surface area have been reported in the literature. These surface areas are measured/estimated through different methods, such as Brunauer–Emmett–Teller (BET) adsorption method, geometry approximations and imaging techniques. Previous work has found that simulations carried out using mineral accessible surface areas calculated from imaging better reproduces the changes observed in core-flood experiments. However, how accessible mineral surface area is impacted by image resolution, how simulations are impacted by surface area variations and if mineral accessible surface area is predictable remain to be answered. This work aims to improve our understanding of mineral accessible surface area and explore the possibility of predicting mineral accessible surface areas from other mineral properties such as mineral abundance. To understand the impact of image resolution on quantified mineral properties, a rock sample extracted from the Paluxy formation (Kemper, Mississippi) was imaged using scanning electron microscope under varying resolutions (0.34 µm to 5.71 µm). Porosities and mineral abundances agree relatively well with changing resolutions, while less than one order of magnitude variations were observed in mineral accessible surface areas. Accessible surface areas from high (0.34 µm) and low (5.71 µm) resolution images were then used in continuum-scale reactive transport simulations to study the impact of image resolution on simulated mineral reactions. Only minor differences were observed between the simulations at both short (1 week) and long time-scales (20 years), indicating the impact of variations in surface area resulting from image resolution on simulated reactions is small. Simulations were also carried out using geometric surface area and BET specific surface area from the literature to understand how surface area variation affects simulated reactions. Small variations were observed at short times, where carbonate minerals are slightly impacted by surface area variations. At longer times, large differences occurred in simulated reactions for non-carbonate mineral phases. Depending on mineral composition and simulation duration, it is not always necessary to quantify mineral accessible surface area from imaging. While imaging approaches are a promising means of estimating reactive surface area values, this approach is time and resource intensive. Here, further investigation of the relationships among the quantified mineral properties including mineral abundance, accessibility, accessible surface area, clay content and connected surface area is explored as a means of predicting mineral accessible surface area without extensive image processing. Ten sandstone samples were imaged and processed using similar approaches to quantify mineral properties, and the results were compared to explore the relationships among these properties. No distinct pattern was found to directly link mineral accessible surface area to other properties determined from imaging. However, the accessibilities of quartz and feldspars were found to be predictable based on their abundances as well as clay content. Lastly, efforts were made to understand the evolution of accessible mineral surface area as mineral dissolution reactions progress. Core-flood experiments were performed on Bandera Grey and Kentucky sandstone samples using 0.01 M hydrochloric acid (HCl) as the reacting fluid. Mineral accessible surface areas before and after the experiments were calculated using the imaging approach and compared to evaluate the changes of surface area for different mineral phases.