Residual trapping, a key CO2 geo-storage mechanism during the first decades of a sequestration project, immobilizes micrometre sized CO2 bubbles in the pore network of the rock. This mechanism has been proven to work in clean sandstones and carbonates; however, this mechanism has not been proven for the economically most important storage sites into which CO2 will be initially injected at industrial scale, namely oil reservoirs. The key difference is that oil reservoirs are typically oil-wet or intermediate-wet, and it is clear that associated pore-scale capillary forces are different. And this difference in capillary forces clearly reduces the capillary trapping capacity (residual trapping) as we demonstrate here. For an oil-wet rock (water contact angle θ=130°) residual CO2 saturation SCO2,r (≈8%) was approximately halved when compared to a strongly water-wet rock (θ=0°; SCO2,r≈15%). Consequently, residual trapping is less efficient in oil-wet reservoirs.

The measured contact angle of sandstone was systematically greater than that of pure quartz because of the pores present in sandstone. Moreover, the effect of pressure and temperature on the contact angle of sandstone was similar to that of pure quartz. The results showed that the contact angle increases with increase in temperature and pressure and decreases with increase in salinity.

CO2 geosequestration in oil reservoirs is an economically attractive solution as it can be combined with enhanced oil recovery (CO2‐EOR). However, the effectiveness of the associated three‐phase displacement processes has not been tested at the micrometer pore scale, which determines the overall reservoir‐scale fluid dynamics and thus CO2‐EOR project success. We thus imaged such displacement processes in situ in 3‐D with X‐ray microcomputed tomography at high resolution at reservoir conditions and found that oil extraction was enhanced substantially, while a significant residual CO2 saturation (13.5%) could be achieved in oil‐wet rock. Statistics of the residual CO2 and oil clusters are also provided; they are similar to what is found in analogue two‐phase systems although some details are different, and displacement processes are significantly more complex.

The water contact angle in a system of brine (20 wt % CaCl2) and CO2 was measured on a smooth dolomite surface [root mean square (RMS) surface roughness of 45 nm] with both hydrophilic and hydrophobic behaviors as a function of the pressure (0.1, 5, 10, 15, and 20 MPa) and temperature (308, 323, and 343 K). The experimental results show that the contact angle of brine/CO2 increases slightly with the temperature when the dolomite surface is hydrophilic but, interestingly, reduces when the surface is hydrophobic. The results also illustrate that the brine/CO2 contact angles increase with increasing pressure. We interpreted the experimental observations using the concept of alteration in van der Waals potential (substrate surface chemistry) with thermodynamic properties, including pressure and temperature.

We imaged an intermediate-wet sandstone in three dimensions at high resolution (1-3.4 mm 3 ) with X-ray microcomputed tomography (micro-CT) at various saturation states. Initially the core was at connate-water saturation and contained a large amount of oil (94%), which was produced by a waterflood [recovery factor R f ¼ 52% of original oil in place (OOIP)] or a direct gas flood (R f ¼ 66% of OOIP). Subsequent waterflooding and/or gasflooding (water-alternating-gas process) resulted in significant incremental-oil recovery (R f ¼ 71% of OOIP), whereas a substantial amount of gas could be stored (approximately 50%)-significantly more than in an analog water-wet plug. The oil-and gascluster-size distributions were measured and followed a powerlaw correlation N ! V Às , where N is the frequency with which clusters of volume V are counted, and with decays exponents s between 0.7 and 1.7. Furthermore, the cluster volume V plotted against cluster surface area A also correlated with a power-law correlation A ! V p , and p was always % 0.75. The measured sand p-values are significantly smaller than predicted by percolation theory, which predicts p % 1 and s ¼ 2.189; this raises increasing doubts regarding the applicability of simple percolation models. In addition, we measured curvatures and capillary pressures of the oil and gas bubbles in situ, and analyzed the detailed pore-scale fluid configurations. The complex variations in fluid curvatures, capillary pressures, and the fluid/fluid or fluid/fluid/ fluid pore-scale configurations (exact spatial locations also in relation to each other and the rock surface) are the origin of the wellknown complexity of three-phase flow through rock.

No abstract

Porosity and permeability of deep unmineable coal seams are key parameters in the context of (enhanced) coalbed methane recovery and CO 2 geo-storage in coal beds as they determine productivity and injection rate. Porosity and permeability are again determined by the microstructure of the coal, and the cleat network-coal matrix system. Furthermore, it is well established that swelling of the coal matrix due to water adsorption can significantly reduce permeability. However, the exact effect of swelling due to water adsorption on the coal micro-structure is only poorly understood, and how this microstructural change impacts on the permeability and porosity characteristics of the coal. We thus imaged dry coal plugs and swollen coal plugs (swollen due to brine adsorption) at high resolution (3.43 μm) 3 in 3D with an X-ray micro-computed tomograph (microCT). On the microCT images two types of cleats were identified; cleats in the coal matrix and cleats syngeneic with the mineral phase. Approximately 80% of the coal matrix cleats closed upon water adsorption, while the cleats in the mineral phase were not affected. This cleat closure by water adsorption dramatically reduced porosity and particularly permeability, consistent with dynamic permeability coreflood measurements.

No abstract