SPE 124253 Pore Networks and Fluid Flow in Gas Shales
F.P. Wang and R.M. Reed, Bureau of Economic Geology John A. and Katherine G. Jackson School of Geosciences, The University of Texas
Copyright 2009, Society of Petroleum Engineers This paper was prepared for presentation at the 2009 SPE Annual Technical Conference and Exhibition held in New Orleans, Louisiana, USA, 4–7 October 2009. This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.
Abstract Favorable gas content, depth, and thickness, along with high brittleness of the Barnett Shale in the Fort Worth Basin, North Texas, have made the basin one of the best shale-gas plays in North America. Using recent pore images and geochemical data for the Barnett Shale, we investigated potential effects of organic matter on petrophysical properties, pore networks and fluid flow in gas-shale systems. Four types of porous media are present in productive gas-shale systems: nonorganic matrix, organic matter, natural fractures, and hydraulic fractures. Organic-matter pores, ranging from 5 to 1,000 nm, are especially important because they can adsorb gases as well as store free gases. Gas content and adsorption data from Barnett Shale also suggest that a significant amount of free gas is stored in organic matter. Porosity in organic matter can be five times higher than that in the nonorganic matrix. Organic matter is oil wet, and associated pores work as nanofilters for hydrocarbon flow, suggesting that fluid flow in organic matter is predominantly single phase. Owing to high porosity, predominantly single-phase flow, and the gas slippage effect, gas permeability in organic matter, significantly higher than that in the nonorganic matrix, tends to enhance gas permeability in gas shale. In addition, the pore network in organic matter, can be larger than that in the fractures, could be the hidden pathway to high gas production in gas shale when connected to natural and hydraulic fractures.
Introduction Success in gas production in the Barnett Shale, Fort Worth Basin (FWB), can be attributed to its high shale quality, evolution of production technology, and market conditions. The high shale-gas quality stems from the basin’s favorable gas content, depth, thickness, pore networks and reservoir quality, and high brittleness. The pore network, reservoir quality, and mechanisms of fluid flow in gas shales, which are significantly different from those of conventional reservoirs and nonorganic fine-grain sediments. Not been well understood, they are complex and challenging. Studies of Devonian shales in the Appalachian Basin (Komar and others, 1981; Soeder, 1988; Davies and others, 1991; Luffel and others, 1992) have led to many important breakthroughs concerning shale gas systems, as well as technologies used today. Davies and others (1991) revealed that pores in Devonian shales are intergranular and that pore diameters increase with detrital particle size. Soeder (1988) showed that permeability of Devonian shales is a function of effective stress and pore pressure. Luffel and Guidry (1992) and Luffel and others (1993) developed the technique of measuring true matrix permeability of Devonian Shale using crushed samples to eliminate natural and drilling-induced microfracture. The role of organic matter in gas-shale, which is poorly understood, can be important in terms of petrophysical properties, as well as migration and production. Leverson (1954) and Perrodon (1983) envisioned that organic substances are oil wet and that hydrocarbons could travel along them in a continuous phase. Using advanced sample preparation and image technology, Reed and others (2007) imaged pores in organic matter from Barnett Shale samples with sizes ranging from 5 to 1,000 nm, which adsorb and store free gas. Objectives of this study are to look into potential effects of organic matter (OM) on ? Pore types, ? Pore networks and permeability, and ? Fluid flow mechanisms. Pore Types At depths typical of oil and gas reservoirs, porosity values of gas-shale systems range from 2 to 15% (Curtis, 2002). Productive gas-shale systems are composed of four types of porous media: nonorganic, organic, natural fractures, and hydraulically induced fractures. The shale matrix is comprised of predominantly clay minerals, quartz, pyrite, and organic matter. Two types of matrix pores are nano-scale pores (Davies and others, 1991, Reed and others, 2007; Bowker, 2007; Bustin and others, 2008) and micro-scale pores (Davies and others, 1991; Bustin and others, 2008). Davies and others (1991) found that the diameter of intergranular pores of the Devonian Shale in the Appalachian Basin increases with detrital particle size. Nano-scale pores have been reported in organic matter (Reed and others, 2007) and clay-rich mudrock (Bustin and others, 2008) and micro-scale pores are seen mostly in silica-rich mudrock (Bustin and others, 2008). Reed and others (2007) and Loucks and others (in print) found that most pores in the Barnett Shale, Fort Worth Basin, are in organic matter or are associated with pyrite. Organic Matter. Organic-matter fragments may act as a separate porous medium in shale. Pore spaces in organic matter are thought to have formed when oil and gas were generated. With pore size ranging from 5 to 1,000 nm (Fig. 1), these pore spaces can adsorb methane (molecule size of 0.38 nm) and store free methane at the same time. Reed and others (2007) estimated that porosity in organic matter ranges from 0 to 25%. Free gas stored in organic matter cannot be separated from that in nonorganic matrix but is indirectly indicated by data from the T. P Sims #2 well, Newark field, FWB. In Fig. 2, gas content and adsorption data of the Barnett Shale from the T. P Sims #2 well were plotted with respect to total organic content (Jarvie, 2004). The graph shows that adsorbed gas, free gas stored in the matrix, increases linearly with total organic content (TOC). Gas content at zero TOC can be considered free gas stored in the nonorganic matrix, and for specific TOC, gas stored in the organic matrix is free gas subtracted from that stored in the nonorganic matrix. Gas stored in the organic matrix increases with TOC as well. Mechanisms of free gas that flow through organic matter are speculative. Leverson (1954) considered organic substances to be the media in which hydrocarbon traveled in source rocks because organic substances and their adjacent areas were oil wet (Boyer and others, 2006). Leverson (1954) suggested that hydrocarbon can flow either along surfaces of, or through, organic substances. This suggests that these substances are porous media. Perrodon (1983) argued that hydrocarbon in organic-rich shale can migrate as a continuous phase instead of being carried by water because not only does water not flow well itself, but it also significantly reduces hydrocarbon permeability in most deep, organic-rich shales.
Figure 1. Scanning electron microimage of Ar-ion-beam milled surface showing pores in organic matter from the T. P. Sims #2 Well, Barnett Shale, FWB.
Gas content (scf/ton)
Total gas Adsorbed gas
Free gas in organic matter Free gas in nonorganic matrix
Free gas in matrix
0 0 2 4
Total organic carbon (wt%)
Figure 2. Adsorbed and total gas contents with respect to total organic content of Barnett Shale, from the T. P. Sims #2 well, Newark East field, FWB. Data from Jarvie (2004).
In contemporary terminology, embedded, oil-wet organic fragments work as nanofilters of hydrocarbon flow. Therefore, gas or oil flow through organic matter is predominately single-phase without residual water, and gas permeability in organic matter is expected to be high. Additionally, the high porosity of organic matter, as well as the Klinkenberg (1941) slippage effect, enhance the gas permeability of organic matter. The wettability of minerals in organic shale, although speculative, is altered toward oil wetness because of the adsorption of polar hydrocarbon (Perrodon, 1983; England and Fleet, 1991).
Size, geometry, stacking pattern, and connectivity of organic flakes are all poorly understood. Organic matter fragaments in shales can be sparsely scattered to layered, with varying connectivity (Loucks and others, in print). However, total pore volume in tiny and poorly connected organic matter is often significant. Pore-Volume Estimation. Pore volume of organic matter is a function of porosity and TOC. TOC, normally measured in weight percent, needs to be converted into volume percent. Using SEM images, Reed and others (2007) and Loucks and others (in review) reported 0 to 25% porosity in organic matter in the Barnett Shale, North Texas. We use data from Barnett shale to estimate pore volumes in organic matter and compare them to pore volumes in fractures in gas shale systems. Using an assumed porosity of 10% in organic matter, and total porosity and TOC values for Barnett Shale in FWB, Marcellus Shale in Appalacian Basin (Chesapeake, 2008) and Haynesville Shale in Louisiana (Stoneburner, 2009) shown in Table 1, estimated porosities in organic matter are 1%. 1.2% and 0.7% of the bulk volume respectively. That is, 20%, 18.5% and 6% of all pores in the Barnett Shale, Marcellus Shale and Haynesville Shale are in organic matter. Haynesville Shale is dominated by pores nonorganic matrix more than Barnett and Marcellus Shales. These values increases with the assumed porosity in oranic matter.
Table 1. Estimation of pore volume in organic matter.
Item Total porosity of shale Assumed porosity in natural fractures Assumed porosity in hydraulic fractures Total organic content (TOC) Assumed porosity in organic matter Porosity of organics in shale Porosity in nonorganic rock matrix
Barnett Shale, FWB 5% <0.5% <0.2% ~10 vol% (5 wt%) 10% ~1.0% >3.3%
Marcellus Shale 6.5% <0.5% <0.2% ~12% (6 wt%) 10% ~1.2% >4.6%
Hayneville Shale 12% <0.5% <0.2% ~7.0% (3.5 wt%) 10% ~0.7% >10.6%
Porosity and pore-volume values in natural and hydraulic fractures are difficult to measure or estimate. Nelson (1991) considered that the porosity in nondissolution-type, naturally fractured reservoirs is generally thought to be <0.5%. This value is used in Table 1 because the porosity in natural fractures for organic-rich shales are speculative. Water fracs may stimulate an additional <0.2% porosity in induced fractures. Therefore, pore volume in organic matter is a function of TOC and porosity in organic matter. It can be equal to or larger than those in natural and hydraulic fractures. In addition, the large weak boundaries between organic matter and the nonorganic matrix can be open for gas to flow during hydraulic fracturing. This hidden pore network in organic matter may play an important role in the higher-than-expected gas production from shales. Permeability Permeability of mudrocks, which is extremely low, from subnanodarcys to microdarcys, is a function of shale type, sample type, porosity, confining pressure, and pore pressure (Soeder, 1988; Davies and others, 1991; Luffel and others, 1992; Guidry and others, 1995; Bustin and others, 2008). Permeability of deep organic-lean mudrocks can be <0.1 nd to tens of nanodarcys (Davies and others, 1991), and permeability values in organic-rich gas shales range from subnanodarcys (Guidry and others, 1995; Cluff and others, 2007) to tens of microdarcys (Soeder, 1988; Davies and others, 1991; Bustin and others, 2008). Two types of samples used are core plug or crushed sample. Shown in Fig. 3, permeability values measured in core plugs (Soeder, 1988; Bustin and others, 2008) are significantly higher than those reported by Guidry and others in 1995 (green squares) using crushed samples and Cluff and others in 2007 (red circles) which might have been measured in crushed samples. Cluff and others (2007) reported that another core sample with 16 μD of plug permeability has only 6 nD of crushed rock matrix permeability. The technique of measuring shale permeability using crushed samples was designed by Luffel and others (1992 and 1993) to measure matrix permeability only by eliminating natural and drilling induced microfractures. Althogh drilling induced fractures are common (Boyer and others, 2006), they can be minimized by selecting core plugs at locations without drilling induced fractures. Because pore networks in organic matter are most likely connected through microfractures, the connectivity of organic pore network can be significantly reduced in crushed samples. Although pore networks in organic matter and natural microfractures are important propertie of shale and critical to shale-gas production, they are too small to be properly quantified in the laboratory or reservoir simulation. The
compromisary but easy way is to include them in core permeability measurements as part of a lumped permeability value. By including important organic and microfracture pore networks, this lumped permeability can characterize gas shales better than the true matrix permeability,
BC, Canada Plugs
Marcellus Shale WV #6
le ha tS et B rn Ba FW
Crushed Samples Crushed Samples?
1.0E-12 0 2 4 6 8 10
Figure 3. Porosity and permeability relationships of shale gas plays in north America measured using core plugs and crushed samples. Data from the Marcellus Shale, West Virginia (Soeder, 1988) and from Canada (Bustin and others, 2008) using core plugs, and the Barnett Shale, FWB as red circles (Cluff and others, 2007) and black shale as green squares (Guidry and others, 1995) using crushed samples.
Soeder (1988) showed that permeability in the Marcellus Shale (Fig. 4) is pressure dependent and decreases with an increase in confining of pore pressure. The effect of confining pressure is caused by a reduction of porosity, and the effect of pore pressure is caused by the Klinkenberg (1941) slippage effect. Bustin and others (2008) also reported the effect of effective stress (confining pressure) in Barnett, Muskwa, Ohio, and Woodford shales (Fig. 4). Note that degree of permeability reduction with confining pressure is significantly higher in shale than in consolidated sandstone or carbonate. It is critical to measure shale-gas permeability at a confining pressure close to lithostatic pressure at reservoir depth.
Gas Permeability (mD)
Barnett Ohio Woodford-1 Huron
Muskwa Marcellus Woodford-2
1E-05 0 1600 3200 4800 6400
Effective Stress (psi)
Figure 4. Effect of confining pressure on gas permeability in gas shales. Data of Marcellus Shale from Soeder (1988) and data of other shales from Bustin and others (2008).
The Klinkenberg (1941) slippage effect maintains that gas flow becomes nonviscous when pore sizes are only 1 to 1,000 times larger than the mean free path of gas molecules (Javadpour and others, 2007). Fig. 5 plots Soeder’s data (1988) using a core plug at 7,448.5 ft from the West Virgina #6 well at a confining pressure of 3,000 psi. Gas permeability increases from 19.6 μD at 1,000 psi to 54 μD at 80 psi because the free path of gas and the
slippage effect increases inversely with pore pressure. However, the effect is significant only at pore pressures <500 psi, and gas permeability changes <15% at pore pressures >500 psi.
0 0 1,000 2,000 3,000
Pore Pressure (psi)
Figure 5. Effect of pore pressure on gas permeability in the Marcellus Shale, with a confining pressure of 3,000 psi. Data measured by Soeder (1988) using a core plug at 7,448.5 ft from the West Virginia #6 well.
Production Fairways As shown in Fig. 6, numerous organic flakes will be intersected by a natural fracture, and high-permeability organic flakes can behave like high-gas-content microfractures. Therefore, high-permeability networks in gas shale can most likely form when nanopores in the organic matter connected by natural micro- and macrofractures are accessed by hydraulic fractures (Wang, 2008).
Figure 6. Schematic diagram showing high-permeability elements in gas shale: organic matter, natural fractures, and hydraulic fractures.
The production fairway in stimulated-reservoir-control, early high gas production and gas in poorly connected organic matter helps mature low gas production to last. The gas rate depends on (1) the ratio of porosity in organic matter to total porosity, permeability, geometry, distribution, and connectivity of organic flakes and (2) how organic matter is connected to natural and hydraulic fractures. Fluid-Flow Mechanisms Different from conventional gas reservoirs, fluid flow in gas shales is controlled by flow mechanisms at all scales, from molecular to macroscopic. Fluid-flow mechanisms include (1) free gas flow, (2) desorption, (3) diffusion, and (4) imbibition suction. Free gas flow can be a non-Darcy type in both organic and nonorganic matrices as a result of the slippage effect, but a Darcy type in natural and hydraulic fractures. Adsorbed gas on organic matter could have an adverse effect on permeability because the layer of adsorbed gas increases drag on gas molecules. If organic matter is not a porous medium, only adsorbed gases adjacent to matrix pores or microfractures can be released, and adsorbed gases far from pores or fractures can be moved
only by diffusion along the organic surface during production. When organic matter is a porous medium, part of the adsorbed gas can be released directly into organic pores, and, as a result the role of diffusion can be markedly reduced. Water production is one of the critical factors affecting the success of shale gas wells. Water is normally from adjacent aquifers rather than shales themselves and produced through fractures in shales. High-quality gas shales typically have high initial gas saturation and less than 30% of initial water saturation (Boyer and others, 2006; Stoneburner, 2009), which is much lower than that measured in the laboratory. The low initial water saturation is referred to as subirreducible initial water saturation (Bennion and Thomas, 2005), which has two important effects on fluid flow in gas shales: preventing water from production and creating a powerful capillary suction of water. The subirreducible initial water saturation was probably created by the excessive drying at high paleo-temperatures and pressure, and during subsequent cooling, there was insufficient water supply to increase irreducible water saturation which made the reservoir subirreducible. An organic-rich shale at subirreducible water saturaton such as Bakken Shale in North Dakota (Perrodon, 1983), Marcellus and Ohio Shales in the Appalachian Basin (Soeder, 1988; Curtis, 2002), the Lewis Shale in the San Juan Basin, and the Barnett Shale in North Texas (Curtis, 2002) produce nil to little water. Imbibition suction is a phenomenon occurring in frac-water flow in tight-gas sandstones (Bennion and Thomas, 2005). In some productive shale-gas plays, such as the Haynesville in Louisiana (Stoneburner, 2009) and the Barnett in North Texas, <50% of frac water flows back during production. This situation could be caused by the combined effects of gravity segregation in fractures and imbibition suction. Gas tends to flow through the upper part of fractures, leaving a significant amount of water in the lower part of fractures during production. During drilling and stimulation, some water can penetrate into shale formation and. part of the penetrated water become residual water in the nonorganic matrix as a result of its being filled to normal irreducible water saturation. The cooling by drilling and frac water around stimulated areas can increase the irreducible water saturation and further escalate the effect of imbibition suction. Conclusions Four types of porous media are present in productive gas-shale systems: the nonorganic matrix, the organic matrix, natural fractures, and hydraulic fractures. Organic-matter pores, ranging from 5 to 1,000 nm, are especially important because they can adsorb gases, as well as store free gases. Gas content and adsorption data from Barnett Shale also suggest that a significant amount of free gas is stored in organic matter. Organic matter is oil wet, and associated pores work as nanofilters of hydrocarbon flow and water blocking, suggesting that fluid flow in organic matter is predominantly single phase. Because most pores in organics are only 10 to 1,000 times larger than the mean free path of gas molecules at reservoir conditions, the nonviscous Klinkenberg slippage effect, can increase gas permeability by only 15% at producing pressures >500 psi. Owing to high porosity, predominantly single-phase flow, and the gas slippage effect, gas permeability in organic matter is significantly higher than that in the nonorganic matrix. In addition, high permeability, high gas content, and the pore network in organic matter, estimated to be equal to or larger than that in the fracture, could be the hidden pathways to high gas production in gas shale when connected with natural and hydraulic fractures. Permeability values measured using core plugs can be orders of magnitude higher than those measured using crushed samples, in which effects of organic and microfracture pore networks on permeability may have been significantly reduced. Different from tight gas sandstones gas flow in organic and nonorganic matrices can be nonDarcy. Many organicrich shales are at subirreducible water saturatons. These shales produce no water and can create powerful capillary suctions of water during drilling and hydraulic fracturing. Acknowledgments Publication is authorized by the Director of Bureau of Economic Geology, John A. and Katherine G. Jackson School of Geosciences, The University of Texas at Austin. Special thanks are to Dr. John Ward, PetroEdge Energy Inc., Houston, Texas for his review and invaluable suggestions, and extended to Chuck Vavra, North Star Geological Services, LLC, Lucas, Texas for the discussion on capillary suction. The manuscript was edited by Lana Dieterich.
Bennion, B. D. and Thomas, F. B., 2005, Formation Damage Issues Impacting the Productivity of Low Permeability, Low Initial Water Saturation Gas Producing Formations: Trans. AIME, v. 127, p. 240-247. Bowker, K. A., 2007, Barnett Shale Gas Production, Fort Worth Basin: Issues and Discussion: AAPG Bulletin, v. 91, no. 4, p. 523-533. Boyer, C., Keischnick, J., Lewis, R. E., and Waters, G., 2006, Producing Gas from Its Source: Oilfield Review, v. 18, no. 3, p. 36-49. Bustin, R. M., Bustin, A. M. M., Cui, X., Ross, D. J. K. and Pathi, V. S. M., 2008, Impact of Shale Properties on Pore Structure and Storage Characteristics: Society of Petroleum Engineers, SPE Paper No. 119892, 28 p. Chesapeake Energy, Inc., 2008, Presentation: in Investor an Analyst Meeting. Cluff, R. M., Shanley, K, W., and Miller, M. A., 2007, Three Things We Thought We Knew about Shale Gas but were Afraid to Ask…: AAPG Poster, 1 p. Curtis, J. B., 2002, Fractured Shale-Gas Systems: AAPG Bulletin, v. 86, no. 11, p. 1921–1938. Davies, D., Bryant, W. R., Vessel, R. K., and Burkett, P. J., 1991, Porosity, Permeability and Microfabric of Devonian Shale, in Bennett, R. H., Bryant, W. R., and Hulbert, M. H., eds., Microstructure of Fine-Grain Sediments: from Mudstone to Shale: New York, SpringerVerlag, p. 109-119. England, W. A., and Fleet, A. J., 1991, Petroleum Migration: London, Geological Society, 280 p. Guidry, K., Luffel, D., and Curtis, J., 1995, Development of Laboratory and Petrophysical Techniques for Evaluating Shale Reservoirs, final report: Gas Research Institute Report GRI-95/0496, 286 p. Jarvie, D., 2004, Evaluation of Hydrocarbon Generation and Storage in Barnett Shale, Fort Worth Basin, Texas: The University of Texas at Austin, Bureau of Economic Geology/PTTC, 116 p. Javadpour, F., Fisher, D., and Unsworth, M., 2007, Nanoscale Gas flow in Shale Gas Sediments, J. Canadian Pet. Tech., v. 46, no. 10, p. 55-61. Klinkenberg, L. J., 1941, The Permeability of Porous Media to Liquids and Gases, in Drilling and Productions Practices: American Petroleum Institute, p. 200-213. Komar, C.A., A. E. Hunt, and A. B. Yost, A Technical Review of DOE Activities in the Eastern Gas Shales: Department of Energy Report UGR_455, 1981, Morgantwon, West Virginia, 18 p. Leverson, A. L., 1954, Geology of Petroleum, 2001 ed.: Tulsa, AAPG Foundation, 724 p. Loucks, R. G., Reed, R. M., Jarvie, D. M., and Ruppel, S. C., in press, Morphology, Genesis and Distribution of Nano-Scale Pores in Siliceous Mudstones of the Mississippian Barnett Shale: Journal of Sedimentary Research. Luffel, D. L, and Girdry, 1992, New Core Analysis Methods for Measuring Reservoir Rock Properties of Devonian Shale: J. of Pet. Tech. (November), p. 1184-1190. Luffel, D. L, Hopkins, C. W., and Schettler, P. D., 1993, Matrix Permeability Measurement of Gas producing Shale: Society of Petroleum Engineers, SPE Paper No. 26633, 10 p. Nelson, R., 1991, Geologic Analysis of Naturally Fracture Reservoirs, 2nd ed.: Houston, Gulf Publishing, 332 p. Perrodon, A., 1983, Dynamics of Oil and Gas Accumulation: Elf Aquitaine, 368 p. Reed, R. M., Loucks, R. G., Jarvie, D. M., and Ruppel, S. C., 2007, Nanopores in the Mississippian Barnett Shale: Distribution, Morphology, and Possible genesis (abs.): Geological Society of America Abstracts with Programs, v. 39, no. 6, p. 358. Soeder, D.J., 1988, Porosity and Permeability of Eastern Devonian Gas Shale: Society of Petroleum Engineers, SPE Formation Evaluation, March, p. 116-124. Stoneburner, R., 2009, The Haynesville Shale: a look back at the first year, in 8th Gas Shales Summit, Dallas, 23 p. Wang, F. P., 2008, Production fairway: speed rails in gas Shales, in 7th Annual Gas Shales Summit, May 6–7, Dallas, 22 p.