VIEWING SHALE RESERVOIRS IN THE MULTI-GIGAPIXEL DIMENSION

Shales are typically heterogeneous at the mm to cm scale. In order to see pores clearly, the rock has to be imaged at high resolution under the SEM (magnifications of 50,000-150,000), and the surface has to be free of physical damage. Traditional mechanical polishing invariably causes distortions (shear and smear) of the sample surface and plugging of pores by polishing debris. The application of argon ion milling overcomes this limitation and has enabled imaging of pore spaces and mineral interrelationships at the magnifications needed for shale studies. Yet, in spite of these advances, capturing heterogeneity at the pore level is hindered by (a) the small observation “footprints” of high resolution SEM images (50 to 100 µm), and (b) the fact that standard argon ion milling only produces surfaces that measure a few square millimeters at best. To overcome these limitations the Indiana University Shale Research Lab designed specialized sample holders for large format ion mill sections that fit into a GATAN 600 ion mill, and allow samples of up to half inch diameter to be milled. Whereas the milled surface is not as perfectly flat as with an edge mill design (e.g. GATAN Ilion), the surface is still very smooth and pores and mineral grain contacts can be observed in the desired detail. In order to capture heterogeneity and overcome the “footprint” limitation, we imaged areas of as large as 5x5 mm by taking large numbers of overlapping serial images. To cover a 5x5 mm square area, composites of 2500 images (4000 by 4000 pixels) were acquired at a pixel resolution of 37 nm. Smaller sized sub-mosaics have been acquired at pixel resolutions down to 10 nm per pixel. The images were assembled into large mosaics (18 gigapixels) using the software GigaPan Stitch. If such a mosaic were to be printed at publication quality (300 dpi), it would measure 38 feet square and be four stories tall. With this approach we can not only examine the various types of shale pores in context, but also see changing pore populations and densities with regard to the distribution of basic shale components, such as fecal pellets, silt streaks, microfossils, and others. A mental picture can then be developed on how the distribution of shale components and pore types is likely to affect the averaged porosity and permeability as well as overall reservoir performance.