It is now estimated that more than 90% of the whole core samples recovered in the US come from shale reservoirs. A primary reason for so much shale coring is that well log analysis requires rigorous core calibration to provide reliable data for reservoir quality, hydrocarbons-in-place, and hydraulic fracturing potential.

However, the uncertainty in interpreting shale well log data is sometimes matched or exceeded by the uncertainty observed in traditional methods of analyzing core samples. Most commercial core analysis methods in use today are 50 or more years old and were developed originally for sandstones and carbonates exceeding 1 millidarcy in permeability. High-quality organic-rich shale, on the other hand, is usually lower than 0.001 millidarcy. This extremely low permeability creates substantial challenges for existing methods and has contributed to the rapid rise of a new approach to reservoir evaluation called digital rock physics (DRP).

DRP merges three key technologies that have evolved rapidly over the last decade. One is high-resolution diagnostic imaging methods that permit detailed examination of the internal structure of rock samples over a wide range of scales. The second is advanced numerical methods for simulating complex physical phenomenon, and the third is high-speed, massively parallel computation using powerful GPUs that were originally developed for computer gaming and animation.

The DRP process

Digital rock physics analysis of shales is usually performed in three stages. Each stage provides visual and quantitative information that can be used to select a smaller but representative volume for the next stage of analysis. Stage 1 is performed on whole cores, Stage 2 uses plug-size samples, and Stage 3 is an ultra high-resolution 3-D pore-scale analysis.

This image is a cross-section of a thinly laminated, occasionally pyritic shale whole core with bed-limited, calcite-filled vertical fractures. X-ray computed tomography (CT) imaging provides the visual information for a detailed geologic description of the rock column. (Images courtesy of Ingrain Inc.)

Stage 1. In Stage 1, whole cores are analyzed using calibrated X-ray computer tomography (CT) imaging at a resolution of about 500 CT slices per linear foot of whole core. These detailed images are obtained while the fragile shale is sealed and protected inside the aluminum core barrel liner. This process has almost eliminated the need to “slab” or saw-cut the core along its length in order to photograph and describe it.

The X-ray energy data is used to determine both bulk density (RhoB) and effective atomic number (Zeff) for each CT slice. Zeff is analogous to the photoelectric wireline measurement and provides key information about lithology. Together, the bulk density and Zeff logs provide quantitative measures to assist in discriminating lithology, porosity, and rock facies. Shale formations are often comprised of stacked parasequences that are quite thin and difficult to detect from well logs. This high-resolution data from whole core provides a powerful tool to define these parasequences, assuming the entire cored section is scanned.

CoreHD data (bulk density vs. effective atomic number) is shown on two wells from the same formation. Color code shows population density where deep red is the highest concentration of data points. Well A has a wider range of lithologic variability, including substantial calcite content. The cored interval of Well B has a higher concentration of low-density, low effective atomic number data, suggesting that it has more organic rich shale than Well A.

Figure 2 shows how the RhoB and Zeff data can be plotted to separate the well into multiple facies and to determine which facies is most likely to be high-quality reservoir. In this figure, the whole core density and atomic number data have been cross-plotted for two wells. The data are color-coded by population density to show where most of the data points are concentrated. Comparing these two wells, it can be concluded that the cored interval of well B has a higher fraction of organic rich shale than the cored interval of Well A.

Stage 2. In Stage 2, plug samples are taken at multiple depths based on the high resolution RhoB and Zeff data from Stage 1. Micro-CT analysis provides information on fine-scale laminations and fracturing at a resolution of 10-20 microns. Image analysis from 2-D scanning electron microscope (SEM) data provides porosity and kerogen volume fraction at a resolution of 5-10 nanometers and is also used as a screening process to ensure representative samples for the subsequent 3-D special core analysis laboratory (SCAL) computations. An X-ray energy dispersive spectrum method provides elemental composition, which is used to compute mineralogy. This method has the advantage of giving both mineral abundance and geometric distribution. A typical sample from a calcareous shale formation is shown in Figure 3.

Figure 3a is a scanning electron microscope image of ion-beam polished shale sample. Figure 3b is an energy dispersive X-ray sprectral data can be analyzed for mineral “finger-printing.” Figure 3c shows the data from 3b summed to obtain the volume fraction of each mineral in the sample.

Stage 3.The 3-D SCAL analysis begins with nanometer-scale pore and matrix imaging. This process uses a focused ion beam SEM (FIB-SEM) system. This system acquires an SEM image of an ion-beam polished surface, then uses the ion beam to slice away a few nanometers of rock and takes another SEM image. This is repeated several hundred times for each sample. All of the individual images are aligned and combined into a single 3-D volume. Image processing and segmentation allows separation of the solid mineral, organic material, and pore space into unique 3-D objects. This resultant 3-D digital rock volume is termed a “vRock,” and is used for subsequent SCAL computation work. The outer surface of a vRock and the segmented pore volume and organic material are shown in Figure 4. Absolute permeability is computed on each vRock using a numerical method known as Lattice-Boltzmann.

The image on the left shows the outer surface of a 3-D FIB-SEM volume from an organic-rich shale. The image on the right shows a transparency view of the distribution of connected pores (blue), isolated pores (red), and organic matter (green).

Results and discussion

Pores in shale resource plays are often described as belonging to one of three classes: inter-granular, intra-granular, or organic matter. While more complex classifications have been proposed, this one seems to be quite useful and general. From work on hundreds of samples from many different shale formations, it appears that organic matter porosity (porosity associated with the diagenesis of kerogen) is especially critical in establishing unconventional reservoir permeability. On the other hand, those samples with primarily inter-granular porosity appear to have lower permeability for a given level of porosity.

Based on pore-scale images from a wide range of organic shales, it can be seen that organic material is present in a variety of forms. Three primary forms, non-porous, spongy and pendular are commonly observed. These three types are shown in a ternary diagram. Non-porous organic components (likely kerogen), shown in the lower right corner, fill all of the available non-mineral space, leaving virtually no porosity or fluid flow path. In the lower left corner is porous or “spongy” organic material commonly encountered in thermally mature gas shales (thermally altered kerogen). At the peak of the triangle is what may be called “pendular” organics. Pendular organic material appears to fill the small inter-granular and grain contact regions, leaving open pore space in the larger voids.

Three main organic matter morphologies have been observed in a wide range of formations: nonporous, spongy, and pendular. Combinations of all three types may be present in the same sample.

The pendular organic material has smooth, curved concave surfaces facing the pore and has little internal porosity. The appearance of this organic matter, potentially a migrant bitumen product, suggests that it may behave as a viscous liquid at reservoir conditions. If it is mobile under reservoir conditions, then laboratory measurements at ambient conditions may underestimate permeability. As also shown in Figure 5, there are combinations of these primary kerogen states, and numerous observations to date suggest that the non-porous type is a starting point from which increasing thermal alteration transforms the solid kerogen to the other forms. The pendular style of organic material appears to be more common in samples from oil-window shales, whereas spongy organics are more common in gas-window shales.

Additional observations

Based on performing hundreds of DRP projects from many shale basins, the following observations can be made:

  • Density (RhoB) and effective atomic number (Zeff) from high-resolution X-ray CT scanning provides detailed information on layering and facies in oil and gas shales;
  • Key facies changes can be readily observed from the CT data, and high organic content zones can be quickly identified;
  • There are three distinct organic matter morphologies: non-porous, spongy, and pendular. Combinations of all three types are often present in the same shale formations and samples; and
  • Organic matter-dominated samples have better permeability than comparable samples with mainly inter-granular porosity.

DRP is especially well-suited to analyzing low-permeability shale core samples. DRP is especially helpful in relating facies and shale pore types to porosity-permeability trends. These trends can then be integrated with facies logs from whole core CT imaging to improve net/gross, reserves, and producibility estimates.

Suggested reading

Loucks, R.G., R.M. Reed, S.C. Ruppel, and U. Hammes; Preliminary Classification of Matrix Pores in Mudrocks, presented to Gulf Coast Association of Geological Societies (GCAGS), April, 2010.

Passey, Q.R., K.M. Bohacs, W.L. Esch, R. Klimentidis, S. Sinha; From Oil-Prone Source Rock to Gas-Producing Shale Reservoir, SPE 131350, CPS/SPE Intl. Oil and Gas Conference, Beijing, China, 8-10 June, 2010.

Tolke, Jonas, C. Baldwin, Y. Mu, N. Derzhi, Q. Fian, A. Grader, J. Dvorkin; Computer simulations of fluid flow in sediment; from images to permeability, The Leading Edge, pp 68-74, January, 2010

Comments