Pore-Scale Grain-Binding Cement Dissolution in Sandstone Rocks and its Effect on CO2 Injectivity

CO2 geological storage in depleted oil and gas reservoirs and saline aquifers is currently one of the most economical methods to mitigate CO2 emissions, mainly because of available geological data from oil and gas exploration and production in the last decades as well as abandoned wellbore facilities. Four main factors dictate the safe and economical CO2 storage and selection of a suitable storage site: while wellbore and cap rock integrity are critical to the security of stored CO2, storage capacity and CO2 injectivity are among the main factors driving the cost of storage per ton of CO2. Low or reduced CO2 injectivity at injection wellbore can significantly increase the compression costs of the project. In worst-case scenarios, a blockage can increase injection pressures above rock fracture pressure, thus critically jeopardising further injection in the same well. The global abundance, high storage capacity and potentially low chemical reactivity of sandstone reservoirs make them great candidates for CO2 geological storage sites. The matrix of siliciclastic rocks is mainly made of quartz grains, and there is negligible chemical reactivity in contact with carbonic acid formed during the injection. However, the most common cementing agents in these rocks (e.g. calcite, dolomite and, to some extent, clays) react with carbonic acid. These reactions lead to mechanisms like cement dissolution, mineral deposition, and sand fine migration (sand mobilisation). Depending on the volumetric percentage of the binding cement, its geometrical distribution within the matrix and pore (and throat) size distribution of the rock, fine migration and binding cement dissolution can enhance or impair CO2 injectivity [1,2]. Despite numerous studies regarding the effect of CO2 injection on fine migration and injectivity, there is still no consensus regarding the impact of intergranular cement dissolution on CO2 injectivity. In this study, we have focused on the effect of binding cement dissolution on pore morphology in a typical sandstone (Berea) and demonstrating how the observations of CO2 injectivity change at the core scale alone could be misleading and explain some apparent discrepancies reported in the literature. We cut three slices of this core for X-ray micro-computed tomography (MicroCT) imaging. The first slice was cut from the original clean core. After performing a coreflood experiment with carbonated water through the rest of the core, two other pieces were cut from the inlet and outlet of the core. To represent the extreme dissolution conditions (lowest pH) during CO2 injection, a batch of North Sea brine fully saturated with CO2 was prepared and injected through the core at typical reservoir conditions (50°C and 2.6×107 Pa). Three 3D models are reconstructed from images of the prepared rock slices to represent the pore structure of the rock before injection and the inlet and outlet of the core subjected to CW injection. Avizo (a commercial software) is used to calculate the permeability and velocity fields in these models. The comparison of results from micro-CT imaging and the coreflood experiment shows how single-scale studies (either core or pore-scale) of CO2 injectivity can lead to a misleading conclusion about the effect of fine migration on permeability at larger scales.