After the dry measurements, the sample was evacuated and injected with low‐pressure CO2 gas for several cycles to purge the air in the pore space. After the final evacuation step, de‐aired water was slowly injected from one end of the sample. (Note that excessive clay swelling by low‐salinity water can result in disintegration of the rock and loss of permeability. In this experiment, tap water was used after confirming that it caused no significant swelling of clays in the sample.) Once the sample was water saturated, the confining stress and pore pressure were increased to 10.4 MPa and 3.5 MPa, respectively, to ensure dissolution of the remaining CO2 gas into the pore fluid over ~12 hours. During this period, the temperature of the sample and the pressure cell was also increased up to 60.5°C.
Prior to the subsequent scCO2 injection experiment, the confining stress and the pore pressure were increased again, up to 22.1 MPa and 13.6 MPa (8.5 MPa differential stress), while maintaining the sample temperature at 60.5°C. Under these conditions, CO2 is supercritical, with viscosity 0.040 cP, bulk modulus 0.040 GPa, and density 535 kg/m3 (NIST Webbook, http://webbook.nist.gov/chemistry/fluid/). In comparison, the water has viscosity 0.47 cP, bulk modulus 2.46 GPa, and density 989 kg/m3. The scCO2 was injected into the core sample at a constant rate, and the pore pressure was controlled using a back‐pressure regulator at the outlet. The injection rate varied from 0.017 to 0.043 mL/min for various test cases, and the sample cross section of 11.4 cm2 (slightly larger for a fractured sample). The pore volume replaced by the injected scCO2 was determined from the weight of the displaced water. The injection was continued until scCO2 broke through the core and no more water was produced from the system. The long‐term drying effect of the scCO2 was not examined in this experiment.
Throughout the experiment, the resonance frequencies and attenuations of the system were measured periodically to monitor changes in the seismic properties of the samples. Between the seismic measurements, during the scCO2 injection tests, X‐ray CT scans were conducted using a modified GE Lightspeed 16 slice medical X‐ray CT scanner to determine the saturation and distribution of scCO2 in the pore space. Cores were scanned at 120 kV and 160 mA, and the voxel size in the obtained images was set to 193 × 193 × 625 microns. Scans were performed under dry and water‐saturated conditions, in addition to the scans performed periodically during the CO2 injection tests, to allow computation of CO2 saturation.
Figure 5.2 Photograph of an on‐going scCO2 injection experiment. The experiment is conducted within a temperature and pressure controlled, X‐ray transparent tubular pressure cell on a CT examination table. Both seismic (acoustic) properties and fluid phase distribution changes in the sample are measured periodically while scCO2 is injected slowly into the sample. The amount of scCO2 in the sample is determined from the weight of the effluent water, and the pH of the effluent water is monitored during the experiment.
Saturation of the pore space by scCO2 in the samples can be determined from CT images by comparing the CT values (X‐ray absorption values) of water‐saturated and scCO2‐injected samples. The image resolution (the voxel size) is not sufficient for exactly resolving the pore space geometry. However, from the CT values of 100% water or scCO2‐saturated samples and a sample saturated by both water and scCO2 (
(5.1)
Similarly, fracture apertures can be determined from CT values of intact and fractured samples at the fracture location and of the fluid within the fracture (CT Intact, CT Fractured, and CT Pore fluid, respectively) (Ketcham et al., 2010). Note that the intact sample measurement can be substituted by the CT values of the rock matrix near the fracture, if the sample is reasonably homogeneous. For this case, the fracture aperture is computed by
(5.2)
where L Voxel is the dimension of the voxel in the image.
Ideally, the CT‐indicated density of the scCO2‐saturated case would also be used for saturation computations. However, because exposure of the dry rock to scCO2 has been observed to affect the rock wettability (McCool & Tripp, 2005) and thus scCO2 invasion, the scCO2‐saturated case was estimated by interpolating between the CT‐indicated density numbers of the water‐saturated and dry cases based on the relative densities of the water and scCO2. This approach is generally effective where the spatial scale of the CT noise is smaller than the spatial scale of the features of interest. For CO2 saturation in a fracture where the spatial scale of noise is larger than the fracture aperture, a segmentation approach was used in which the location of the fracture aperture was determined from the dry set of scans, and computations of differences were performed on those specific locations separately.
5.2.3. Samples and Test Cases
A series of seismic measurements with X‐ray CT imaging was conducted on two Carbon Tan sandstone core samples (Kocurek Industries, Inc., http://www.kocurekindustries.com/) with various fracture configurations (Table 5.1). The sandstone has a porosity of 16.7% and a matrix permeability of 34.5 md (measured only for the Carbon Tan #2 core. However, similar dry densities of the #1 and #2 cores (2,380 kg/m3 and 2,375 kg/m3, respectively) indicate that the #1 core has a similar porosity. The permeability was measured using the falling‐head permeability test method (e.g., Head, 1982) under unconfined conditions, using tap water at room temperature. The diameter of the cores was 38 mm, and the length varied slightly among the cores, from 98 mm to 101 mm. A single fracture was introduced in each core via extensile splitting (Brazilian loading) either parallel or perpendicular to the core axis.
Table 5.1 Descriptions of test samples with different fracture configurations
| Sample | Description | scCO2 injection rate (mL/min) | Configuration |
|---|---|---|---|
|
Carbon
|