|Utilising New Technologies to Better Understand Relationships between Porosity & Permeability, Mineralogy and Organic Matter in Shale Gas Reservoirs|
1Fugro Robertson Ltd
Shale gas has become an increasingly important global resource. Unlike in conventional resources, shale gas reservoirs serve as the reservoir rock, source rock and seal. The natural gas produced from “gas shales” is derived from gas trapped within natural fractures or open pore spaces, and/or adsorbed onto kerogen. Many geological factors are important in any successful shale gas play, including total organic carbon, thermal maturity, brittleness and natural fractures, porosity and permeability, pore pressure, mineralogy and thickness. It is therefore important to accurately assess each of these components in any exploration or development scenario.
We have developed new applications to enhance our understanding of shale gas reservoirs, particularly the relationships between mineralogy, pore fabric and organic matter. Microporosity networks within shale gas reservoirs are often poorly understood. Although pore fabric is directly related to permeability anisotropy, it is not always clear how this fabric relates to the kerogen. Furthermore, an enhanced knowledge of the host rock lithology would also be beneficial to enhanced recovery programmes such as ‘fraccing’.
In this study we utilised a number of new technologies alongside some more conventional techniques to improve our understanding of the relationships between mineralogy, pore fabric and organic matter. This new workflow comprises a combination of automated quantitative mineralogy (QEMSCAN®) and core magnetics (MAGPORETM), with TOC (Total Organic Carbon), kerogen organofacies and optical maturity determination.
Field samples were analysed from a number of US shale gas formations:
The recorded quantitative mineralogical data were used in conjunction with conventional petrological analysis, EM analysis and magnetic susceptibility measurements (MAGPORETM) to increase understanding of the rock and organic fabric and its relationship to porosity, permeability, and permeability anisotropy.
Core samples from the Haynesville Formation in a representative well were then analysed, and the results combined with those for the field samples to illustrate downhole trends in mineralogy, elemental composition, specific gravity and organic geochemistry.
QEMSCAN® consists of a scanning electron microscope combined with four energy dispersive spectrometers and an electronic processing unit. Cleaned samples are set in a resin block or thin section, polished and carbon coated.
Each sample is scanned at 10µm resolution at 12,000 points per minute (so each sample represents approximately 900,000 point counts). Initially the back-scatter electrons (BSE) are measured. This equates surface detections to atomic weight, the brightness of the image reflecting chemical composition. A back-scatter cut-off is used to identify the rock fragments and these are then further analysed by the EDS detectors which provide spectral elemental data. The combination of elemental data and BSE value for each measurement is converted in the software into a mineral identification. The sum of all the data points then provides a digital map of the scanned lithology with each pixel representing an assigned mineral. In this particular study the mineral classification software was modified by re-defining the calcite constituent mineralogy with varying levels of silica.
The MAGPORETM technique involves injecting a ferrofluid into the pore space of the rock and measuring the anisotropy of magnetic susceptibility (AMS). The AMS response is controlled by shape alignments of the pores, and the maximum and minimum susceptibilities Kmax and Kmin correspond with the preferred orientations of long and short pore dimensions.
Figure 1: Determining pore fabric orientation by anisotropy of magnetic susceptibility (AMS) measurements.
Organic matter was isolated for microscopical analysis using standard, non-oxidative palynological techniques, including mineral dissolution using HCl and HF, and heavy liquid separation, and slides for spore colour and kerogen composition analysis were made on standard glass slides and cover slips. Polished rock samples used for QEMSCAN® were provided for photomicrographic documentation by polishing off the carbon film on the surface of the sample blocks. Bituminite was measured in Devonian samples which were devoid of vitrinite, and used to calculate the equivalent vitrinite reflectance.
FIELD SAMPLES: WOODFORD SHALE
The examples below highlight differences between dolomitic and siliceous ‘shale’ members.
Figure 2: QEMSCAN® image of dolomitic Woodford Shale.
Figure 3: QEMSCAN® image of siliceous Woodford Shale (left) and (right) split image of an acritarch from siliceous Woodford Shale. The yellow-orange colour of the acritarch in white light and strong yellow fluorescence is typical of acritarchs at early to peak oil maturity, but maturity is probably too low for significant gas generation.
Figure 4: Stereoplots of MAGPORE™ directions in the dolomite and the siliceous shale samples (unoriented). Upper hemisphere Kmax, Kint and Kmin directions are plotted as circles, squares and crosses respectively. The inferred planar feature in the dolomite is shown by the dashed line. There is a good consistency in the planar feature directions seen from the five sub-samples from the dolomite, but such planar features are more random in the siliceous shale (4 sub-samples) which is in agreement with the QEMSCAN data.
CORE SAMPLES: HAYNESVILLE FORMATION
Nine samples from a core taken in the Haynesville Formation were analysed by QEMSCAN® and organic geochemistry techniques. The results are shown on Figure 6.
Figure 6: Downhole bulk mineralogy plot, showing elemental ratios, individual mineral curves, specific gravity curve and grain size curves, with TOC and vitrinite reflectance values on the right-hand side.
The combination of techniques provides a unique insight into the relationship between porosity, permeability, organic matter and mineralogy in shale gas reservoirs.
QEMSCAN® successfully mapped and quantitatively defined the bulk mineralogy from a number of shale gas lithologies. Lithological mapping allowed detailed investigation of the micro sedimentology of each sample and – because the technique is non-destructive - allows analysis of the textural characteristics of each sample.
The definition of average grain size of key minerals such as quartz and changes in specific gravity are highlighted in this study. The cored sequence highlights the applicability of QEMSCANTM to not only quantify the mineralogy of the individual horizons but shows the changes in mineralogy downhole.
Changes in abundance of various key minerals between samples show great potential for zoning this relatively homogeneous lithological unit. By comparison to elemental ratios as used in a conventional chemostratigraphic approach, this is much more detailed and shows that it is an exceptionally powerful correlative tool.
The MAGPORETM technique shows the relationship between porosity and preferred orientation, with data from the Woodford dolomite sample showing a defined plane indicating a preferred orientation of flow. In contrast, the Woodford siliceous shale is more complex and shows that there is no preferred orientation.
QEMSCAN® can be successfully used to quantify macroporosity. It was used to verify the MAGPORETM data for the Woodford siliceous shale, which suggests that measured porosity is higher in the siliceous lithology than in the dolomite ‘shale’.
Most of the field samples are organically rich and contain mainly amorphous organic matter of sapropelic origin, and likely to have good original source potential. The presence of acritarchs in some samples and the abundant sapropel is consistent with a restricted marine source.
Woodford Shale samples are early mature for oil generation and are unlikely to have generated significant quantities of gas. The Oatka Creek Member of the Marcellus Shale and the Rhinestreet Shale are late mature for oil generation and are at the onset of gas generation. The Union Springs and Purcell Members of the Marcellus Shale, and the Genesee and Levanna Shales are post-mature; all reactive kerogen in these samples is likely to have been converted to hydrocarbons. QEMSCAN® analysis showed the presence and location of organic matter within the Woodford Dolomite and Barnett Shale, the former being directly compared to the organic geochemistry data.
All Haynesville core samples are post-mature; all reactive kerogen in these samples is likely to have been converted to hydrocarbons. The Haynesville Formation core samples are dominated by amorphous organic matter, thought to be of sapropelic origin in most of the cores. However, the finely divided and dispersed character of the organic matter in CORE-1 suggests poorer source quality, which is borne out by the relatively low TOC. Palynomorphs are too severely altered for confident identification, but a restricted marine source would be the most likely depositional environment.