The appeal of gas as a clean-burning fuel factors heavily into future demand and current interest in gas shale reservoirs. By some estimates, the US has a secure domestic supply of natural gas for at least 200 years. The offset in conventional gas well decline has come from unconventional gas resources, specifically coalbed methane, tight gas sands, and gas shales.

The future growth in supply is projected from recent success in gas shales. The Barnett shale of the Forth Worth Basin is the poster child for success. With experience, meaningful, fundamental questions are being asked:
• Where is the gas?
• How is gas transported through shale?
• How is gas estimated in place?
• How much gas is adsorbed, and how much gas is free gas?
• How effective are stimulations?
• What stimulation geometries are optimal?

Answers to these questions can be found in the microstructure of shales. The first and most important thing to learn is all shales are not the same and are not economic gas shales. Most successful gas shale plays have extensive acreage; they range in thickness, total organic content (TOC), porosity, permeability, and fracture density. Critical of many plays are timing and sealing of fracture systems. Shales can be fractured during generation of methane and/or through post-depositional tectonic deformation. If fracture systems are not filled through mineralization, the hydrocarbon escapes and charges overlying reservoirs; however, if the fractures seal, gas becomes trapped in the shale.

Barnett, Haynesville, Woodford, Horn River, photomicrographs

Ion-milled SEM photomicrographs depict four gas shales: Barnett, Haynesville, Woodford, and Horn River. Microstructure is variable from shale to shale, and each controls the gas in place and production. (Images courtesy of Mark Curtis)

Within a shale play there are opportunities for breaching and compartmentalization. Focusing on microstructural characteristics of gas shales, porosities range from 1% to 12%, TOC content ranges from 1% to 20%, and permeabilities vary over several orders of magnitude. Porosity measurements, while trivial in concept, are difficult in practice. This leads to multiple values often ranging by factors of 2 to 5 from various commercial laboratories. Some laboratories refuse to supply a description of measurement procedures. Errors in porosity could immediately imply erroneous gas-in-place estimations and recovery efficiencies.

The TOC values obtained on cores calibrate log-based estimates. Porosity in TOC and its contribution to free and absorbed gas estimates are not captured. Permeabilities are low, requiring untenable times for steady-state measurements. Most measurement techniques are, therefore, pulse decay or pressure decay/buildup. These techniques can be applied to crushed core samples or core plugs. Crushed sample measurements mitigate induced and natural fracture contributions. However, measurement on crushed samples is demonstrated to depend almost linearly on particle size. Few commercial vendors sieve crushed samples prior to measurement, and none provide the range of sizes used in the test. Additionally, values are an average
of vertical and horizontal permeabilities.

Permeability measurements on core plugs have demonstrated dependency on fractures, which, with sufficient pressure, can be minimized. However, few permeability measurements are made at conditions approaching reservoir conditions. Furthermore, because shales contain organics, pulse measurement techniques are susceptible to gas adsorption during the pressure transient, which leads to overestimation of permeability.

All shales are not shale gas plays
Economic reality is not all shales are or will be shale gas plays due to factors such as infrastructure, market, price, depth, producibility, and containment.
Most shales are fractured; however, successful gas shales have fractures that have been cemented. Timing cementation is critical. Evidence suggests some fractures are cogenerated during hydrocarbon generation. Shales have been historically recognized as source rocks that upon heating generate hydrocarbons that charge overlying reservoirs; making shales reservoirs requires containment.

Microstructure holds the secrets to development of gas shales. New imaging capabilities visualize and measure dimension of pores controlling performance of a gas shale. Images can be taken with a dual-beam scanning electron microscope. Integrated ion and electron beams and the resolution at low electron beam-driving voltages make the system suitable for shale microstructure investigations. Integrated ion and electron beams provide images of freshly ion-milled surfaces with absolute geometric registration, and low electron beam-driving voltages produce consistent image quality while minimizing charge buildup.

The disadvantage of high magnification imaging systems is the limited field of view. In a slice and view mode, a series (250 to 1,000) of such images can be collected as the ion-beam slices its way 10 nanometers at a time through the sample. The slices can be aligned, and features (gray scales) can be connected from image to image, rendering a 3-D volume that can be quantitatively analyzed for pore, kerogen, grain dimensions, etc.

What can be observed from these pictures? First and foremost, the Haynesville shale image shows a radically different pore structure from others; pore structure appears more “crack-like,” and porosity appears to be located between grains. The significance of the features is production decline can be rapid because as the pore pressure drops, features collapse, reducing permeability. If storage is limited to the crack-like pores, there is limited or nonexistent longer term contribution from adsorbed gas on the organics.

In contrast, the Barnett and Woodford shales each have a different mix of organic (dark materials) and grains (lighter gray). The very bright features are pyrite. Predominant porosity is visible in the organics in the shales. This suggests additional free gas is associated with the organics, and organic densities are less than pure end members. Both insights factor into gas-in-place estimates. Shale from the Horn River is a quartz-rich system that lacks substantial organics.

Dimensions of the pores and their connectivity change the physics of flow models needed in reservoir simulators. While images are small compared to field size, they are providing new insights into performance, produciblity, and modeling of strategic resources. Statistical sampling can increase with time, and subtleties can be documented and linked to performance. Shales require tools and technologies individual to their scale; nanometer imaging is one such technology. Recent success in relating nanoscale microstructural observations to “petrotypes” and production define a promising formulaic approach to shale reservoir description.

Acknowledgements
We would like to thank Devon Energy for its generous support and establishment of the comprehensive shale research capabilities at the University of Oklahoma.