Seismic signatures of supercritical CO2 injection/drainage within brine saturated sandstone samples

Successful sequestration of CO2 in geological formations requires high-resolution monitoring of injected CO2 location and accurate determination of CO2 saturation in the pore space, typically using seismic methods. Although understanding of the rock physics (relationship between geophysical properties such as seismic velocities and attenuations and the physical characteristics and environmental parameters of rock including porosity and saturation) of partially saturated rock has advanced significantly in recent years, relationships between heterogeneities of rock (both inherent heterogeneity in the rock fabric and structure and distribution of multiple fluid phases in the rock) and its impact on seismic properties are complex and difficult to understand using existing models. Further, most laboratory experiments to date examining the seismic signatures of fluid substitution involving liquid or supercritical (sc-) CO2 have been conducted at ultrasonic frequencies which could result in very different results from field measurements, and also their interpretations are often made assuming sample homogeneity. We present the results of our recent laboratory measurements on the changes in sonic-frequency seismic properties of initially brine saturated sandstone samples during sc-CO2 injection and drainage. High-permeability reference sandstone sample (Berea, ~680 mD) and a medium-permeability (~15 mD) reservoir sandstone used for sequestration (Tuscaloosa formation from Cranfield, Mississippi) were tested. A modified resonant bar method (the Split Hopkinson Resonant Bar method) was used to measure near-1 kHz seismic velocities and attenuations. This method allows us to use small core samples which are typically available from boreholes. Although our measurement frequency was higher than typical field seismic measurements, it is close to the frequencies which have been used for recent cross-hole seismic monitoring of CO2 injection at several sequestration demonstration sites, including Cranfield. Preliminary results show several significant differences in the seismic signatures between the two sandstones. For example, the higher-permeability Berea sandstone exhibited a sharp peak in P-wave attenuation during sc-CO2 injection, before the breakthrough of the CO2 through the core sample occurred. In contrast, the lower-permeability Cranfield reservoir sandstone exhibited nearly monotonic increases in attenuation and no recovery in seismic amplitude during sc-CO2 injection. These differences are attributed to the differences in the scale of wave-induced diffusive pressure waves (slow P waves) and the fluid patch sizes between the two samples.