Characterizing Disproportionate Permeability Reduction Using Synchrotron X-Ray Computed Microtomography

Summary X-ray computed microtomography was used to investigate why gels reduce permeability to water more than that to oil in strongly water-wet Berea sandstone and in an oil-wet porous polyethylene core. Although the two porous media had very different porosities (22% vs. 40%), the distributions of pore sizes and aspect ratios were similar. A Cr(III)-acetate-HPAM gel caused comparable oil and water permeability reductions in both porous media. In both cores, the gel reduced permeability to water by a factor 80 to 90 times more than that to oil. However, the distributions of water and oil saturations (vs. pore size) were substantially different before, during, and after gel placement. This paper examines the mechanism for the disproportionate permeability reduction in the two porous media. Introduction Many polymers and gels can reduce the permeability to water more than that to oil or gas.1–15 This property is critical to the success of water-shutoff treatments in production wells if hydrocarbon- productive zones cannot be protected during polymer or gelant placement.2,3 However, the magnitude of the effect has been unpredictable from one application to the next. Presumably, the effect would be more predictable and controllable if we understood why the phenomenon occurs. Although many mechanisms have been considered (see Table 1), the underlying cause of the disproportionate permeability reduction remains elusive. Previously, we used NMR imaging to observe disproportionate permeability reduction on a microscopic scale.16 Results from these experiments revealed that the imaging technique had many limitations that prevented us from obtaining reliable pore-level images. Most importantly, the spatial resolution was on the order of hundreds of micrometers, which was too low to clearly distinguish fluid pathways on the pore level. In this paper, we describe imaging experiments using high-resolution computed X-ray microtomography (XMT) to compare the oil and water pathways and fluid distributions before and after gel treatment. The current generation of synchrotron-based XMT scanners provides the ability to obtain 3D pore-level images of rock samples with a spatial resolution on the order of micrometers. 17–23 For this study, we used the ExxonMobil beamline X2-B at the Natl. Synchrotron Light Source.18 X2-B is a dedicated XMT imaging facility capable of producing continuous registered stacks of 2,048×2,048×1,024 14-bit 3D images of X-ray linear attenuation coefficients at energies tunable from 8 to 40 keV. The highly collimated synchrotron X-rays permit the reconstruction of a 3D image from 2D projections taken at uniformly spaced angles between 0 and 180°. X2-B converts the pattern X-rays transmitted by the specimen (projections) to a visible light image with a thin single crystal of CsI(Na). This image was magnified by an optical microscope objective onto a 1,024×1,024 charge coupled device (CCD). Using Fourier methods, the set of angular projections at each row of pixels in the CCD was used to reconstruct the crosssectional slice at that row. These slices were stacked to form the 3D image. In this work, a 5×microscope objective was used to provide a pixel size of 4.1 µm and a 4.1-mm field of view. Because part of the core was outside the imaged area, a profile extension method was used to supress edge artifacts. Several authors used XMT to characterize the microscopic structure of porous media.17,19,23 For a 15-darcy sandstone, Coles et al.19 found a mean tortuosity of 2.7, with a range from 1.5 to 4.5. Along a 2.2-mm-long section of this core, porosity varied only a few percent around the average value (26.4%). After oilflooding, this core was waterflooded to a water saturation of 25.1%. Interestingly, large variations in water saturation were observed along the 2.2-mm-long section, ranging from 12 to 39%. A 3D view showed the nonwetting phase (water, in this case) to exist as large ganglia (blobs of nonwetting phase that extend over multiple pores, often exhibiting a branched structure).19 Chatzis et al.24,25 suggested that rock heterogeneity can be responsible for saturation variations within a porous medium. Nonwetting phase saturations that are lower than expected can occur when clusters of small pores are dispersed in a matrix dominated by large pores. In contrast, nonwetting phase saturations that are higher than expected can occur when clusters of large pores are dispersed in a matrix dominated by small pores.24 However, significant saturation variations can occur even in homogeneous porous media, depending on the pore-body/pore-throat aspect ratio. For homogeneous 2D micromodels, Chatzis et al.24 reported piston-like displacements with very little trapping of the nonwetting phase when the aspect ratio was 2 or less. However, for aspect ratios around 3, large nonwetting phase clusters formed as the wetting phase formed fingers while displacing the nonwetting phase. At higher aspect ratios, the nonwetting phase tended to be trapped in individual pores rather than in large clusters of pores. The pore coordination number had a minor effect on nonwetting phase residual saturations.24 Using XMT data, Lindquist et al.23 extensively characterized pore- and throat-size distributions for Fontainebleau sandstones. As core porosity increased from 7.5 to 22%, they found that the average pore coordination number increased from 3.4 to 3.8; the average channel length decreased from 200 to 130 µm; the average throat area increased from 1,600 to 2,200 mum2; and the average pore volume remained fairly constant at approximately 0.0004 mm3. The aspect ratio (effective average pore radius/effective average throat radius) was greater than 2 in 65% of pores and greater than 3 in 40%. The aspect ratios tended to increase slightly as porosity decreased.