Hydrologic Responses to CO2 Injection in Basalts Based on Flow-through Experiments

Abstract:

Experimental studies of basalt-CO₂ interactions have increased our ability to predict geochemical responses within a mafic reservoir during geologic CO₂ sequestration. However, the lack of flow-through experiments prevents the use of coupled hydrologic-geochemical models to predict evolution of permeability and porosity, critical parameters for assessing storage feasibility. We present here results of three flow-through experiments on an intact basalt core during which we employed X-ray Computed Tomography (CT) to quantify porosity evolution and fluid flow. Using a single core of glassy basaltic tuff from the Snake River Plain (Menan Buttes complex), we performed tracer tests using a solution of NaI (~100,000 ppm) before and after injection of CO₂-saturated water at reservoir conditions (90 bar, 50°C) to image porosity and flow path distribution. During the tracer tests, CT scans were taken at 2.5-minute intervals, and outlet fluid was discretely sampled at the same intervals and subsequently measured via ICP-MS, enabling interpretation of the tracer breakthrough curve through both imaging and geochemical analyses. Comparison of the porosity distribution from before and after injection of CO₂ shows an overall decrease in core-averaged porosity from 34% to 31.1%. Permeability decreased exponentially from ~4.9x10^-12 m² to 1.18 x10^-12 m². The decrease in porosity and permeability suggests geochemical transformations in the mineral assemblage of the core, which we observe through petrographic analysis of an unaltered sample of the same lithology in contrast with the altered core. There is a significant increase in grain coatings, as well as reduction in the grain size, suggesting dissolution re-precipitation mechanisms. Finally, to develop a framework for the coupled geochemical and hydrologic responses observed experimentally, we have calibrated a reactive transport model at the core scale using the TOUGHREACT simulator [1].

[1] Xu et al. (2011) Comput. Geosci.