Journal of Sedimentary Research, 2009, v. 79, 848–861
Current Ripples
DOI: 10.2110/jsr.2009.092
MORPHOLOGY, GENESIS, AND DISTRIBUTION OF NANOMETER-SCALE PORES IN SILICEOUS
MUDSTONES OF THE MISSISSIPPIAN BARNETT SHALE
ROBERT G. LOUCKS,1 ROBERT M. REED,1 STEPHEN C. RUPPEL,1
AND
DANIEL M. JARVIE2
1
Bureau of Economic Geology, Jackson School of Geosciences, The University of Texas at Austin, University Station, Box X, Austin, Texas 78713-8924, U.S.A.
2
Energy Institute, Texas Christian University, Fort Worth, Texas 76109, U.S.A.
e-mail: [email protected]
ABSTRACT: Research on mudrock attributes has increased dramatically since shale-gas systems have become commercial
hydrocarbon production targets. One of the most significant research questions now being asked focuses on the nature of the
pore system in these mudrocks. Our work on siliceous mudstones from the Mississippian Barnett Shale of the Fort Worth
Basin, Texas, shows that the pores in these rocks are dominantly nanometer in scale (nanopores). We used scanning electron
microscopy to characterize Barnett pores from a number of cores and have imaged pores as small as 5 nm. Key to our success in
imaging these nanopores is the use of Ar-ion-beam milling; this methodology provides flat surfaces that lack topography related
to differential hardness and are fundamental for high-magnification imaging.
Nanopores are observed in three main modes of occurrence. Most pores are found in grains of organic matter as intraparticle
pores; many of these grains contain hundreds of pores. Intraparticle organic nanopores most commonly have irregular,
bubblelike, elliptical cross sections and range between 5 and 750 nm with the median nanopore size for all grains being
approximately 100 nm. Internal porosities of up to 20.2% have been measured for whole grains of organic matter based on
point-count data from scanning electron microscopy analysis. These nanopores in the organic matter are the predominant pore
type in the Barnett mudstones and they are related to thermal maturation.
Nanopores are also found in bedding-parallel, wispy, organic-rich laminae as intraparticle pores in organic grains and as
interparticle pores between organic matter, but this mode is not common. Although less abundant, nanopores are also locally
present in fine-grained matrix areas unassociated with organic matter and as nano- to microintercrystalline pores in pyrite
framboids.
Intraparticle organic nanopores and pyrite-framboid intercrystalline pores contribute to gas storage in Barnett mudstones.
We postulate that permeability pathways within the Barnett mudstones are along bedding-parallel layers of organic matter or a
mesh network of organic matter flakes because this material contains the most pores.
INTRODUCTION
Mudrocks (sedimentary rocks having a dominant grain size of
, 65 mm) have received renewed research focus the past few years
because of their emergence as hydrocarbon reservoirs (Montgomery et al.
2005). Numerous fundamental questions are being asked about mudrock
systems, including what are the nature and distribution of pores that
compose shale-gas reservoirs. Identification of mudrock pore networks
has become a higher research priority as the commercial value of
mudrocks has increased.
Our studies show that much of the mudrock pore system is not readily
observable by conventional sample-preparation methods. Because of their
small size, most pores are difficult to differentiate from samplepreparation artifacts, as seen in broken or conventionally mechanically
polished samples. To image mudrock pores more accurately, we have
implemented new approaches to sample preparation that allow unequivocal recognition of pores as small as 5 nm (Reed and Loucks 2007).
In this paper we report on studies of mudrock pores in siliceous
mudstones of the Barnett Shale from the Fort Worth Basin (Fig. 1). The
Barnett is a succession of organic-rich, black mudstones (Fig. 2)
Copyright E 2009, SEPM (Society for Sedimentary Geology)
deposited in a deep (below storm wave base), predominantly dysaerobic
to anoxic foreland marine basin along the south margin of the Laurussian
paleocontinent (Fig. 1) during the Mississippian Period (Gutschick and
Sandberg 1983; Loucks and Ruppel 2007; Ruppel and Loucks 2008;
Rowe et al. 2008). Geochemical studies and vitrinite reflectance (VRo)
data suggest that the Barnett strata were heated to temperatures between
100 to 180uC in some parts of the basin (Jarvie et al. 2007).
Shale-gas production in Newark East field, currently the largest gas
field in Texas and the second-largest in the United States (EIA 2006),
comes entirely from the Barnett Shale. Although Barnett gas production
depends on hydraulic fracturing stimulation to create a permeable
collection system, two major unanswered questions exist: (1) where is the
gas stored in the rock and (2) what pathways does the gas follow from the
matrix to these induced fractures that allow it to flow into the well bore?
Methodology that we developed to image nanopores and the resulting
images provide critical data for addressing these questions. The major
aims of this investigation, therefore, are to investigate the pore structure
present in this unit and consider the links between organic distribution,
gas storage, and permeability pathways (pore structure) from the hostrock matrix into the induced fracture system.
1527-1404/09/079-848/$03.00
JSR
NANOPORES IN SILICEOUS MUDSTONES
849
FIG. 1.— Paleogeographic map of the southern midcontinent region during the middle
Mississippian Period showing Barnett Shale
sample locations. Four cores are cited in this
study: Mitchell Energy Corp. T.P. Sims #2 (S),
Texas United Blakely #1 (B), Houston Oil &
Minerals Walker #D-1-1 (W), and Houston Oil
& Minerals Neal #A-1-1 (N). Also shown are 14
other wells from which samples were analyzed.
Modified from Ruppel and Loucks (2008).
FIG. 2.—Thin-section photomicrographs of Barnett mudstones. A) Sample showing a highly compacted, crudely laminated siliceous mudstone in which laminae have
varying amounts of carbonate and quartz silt. Much of the sample is composed of peloids—subspherical grains composed of micron-sized material. No burrows are
present. Porosity 5 1.3%. Blakely #1, 2,187.5 m. B) Sample composed of peloids, quartz silt, carbonate silt, and shell fragments. Porosity 5 0.7%. Blakely #1,
2,184.8 m.
850
JSR
R.G. LOUCKS ET AL.
TABLE 1.— List of core samples with corresponding depths used in
this study.
Depth
(m)
Number of
Samples
Well Name
County
57.9
117.5
197.4
359.5
390.8
539.1
731.7
733.0
1425.2
1529.1
1737.5
1877.5
1894.9
2101.9
2166.5
2167.4
2184.8
2187.6
2191.8
2196.4
2324.0
2361.9
2371.5
2450.4
2604.5
1
1
2
2
1
1
1
2
2
2
1
1
1
1
1
2
1
1
1
2
1
1
2
1
1
Houston O & M Moore #C-1-1
Houston O & M Neal #A-1-1
Houston O & M Hardy #L-A-1
Houston O & M Mullis #A-4-1
Houston O & M Walker #D-1-1
Houston O & M Petty #D-6-1
Houston O & M Godfrey #E-8-1
Houston O & G Potter #C-9-1
Proprietary well
Cities Services St. Clair #1
Proprietary well
Oxy Tarrent #A-3
Oxy Tarrent #A-3
Mitchell Energy Young #2
Mitchell Energy Young #2
Texas United Blakely #1
Texas United Blakely #1
Texas United Blakely #1
Texas United Blakely #1
Texas United Blakely #1
Mitchell Sims #2
Mitchell Sims #2
Proprietary well
Proprietary well
Proprietary well
San Saba
San Saba
Lampasas
Brown
San Saba
Brown
Brown
Brown
Erath
Erath
Archer
Jack
Jack
Wise
Wise
Wise
Wise
Wise
Wise
Wise
Wise
Wise
Hill
Montague
Hill
To further these aims, the objectives of this paper are to (1) characterize
all observable pore types in the Barnett samples, with special emphasis on
the predominant nanopore types; (2) provide images of the nanopores
and discuss the techniques to obtain images of these pores; (3) consider
origins of the nanopores based on their characteristics and occurrences;
and (4) compare petrographic data with measured petrophysical data.
The data and images we present here should provide a fundamental
starting point for continuing research on the nature, formation, and
distribution of pores in all mudrock systems.
METHODS AND SAMPLES
For this study, samples from 18 wells (Fig. 1, Table 1) were examined.
Samples range in depth from 57.9 m to 2604.5 m. Details of samples from
four of these wells are discussed in this paper (Table 2). However, the
concepts presented here are based on the entire dataset. The Texas United
Blakely #1 well and the Mitchell Energy T.P. Sims #2 well, both from
southeastern Wise County, are in the producing area of the Barnett Shale
(Fig. 1). Blakely #1 samples were taken from the upper Barnett Shale
and the upper part of the lower Barnett Shale. Sims #2 samples were
taken from the middle part of the lower Barnett Shale. All seven samples
are black, organic-rich, siliceous mudstones. Figure 2 displays two of
these mudstone samples. Lower-thermal-maturity, siliceous mudstone
samples from shallower depths (Table 2) are presented for comparison
with higher-thermal-maturity samples. The shallow samples are from the
southwestern part of the Fort Worth Basin (Fig. 1). The two samples
presented in the paper are from the Houston Oil & Minerals G.B. Walker
#D-1-1 well is in San Saba County and the Houston Oil & Minerals Neal
#A-1-1 well is in McCulloch County (Fig. 1).
For our initial studies of Barnett pore structure, we prepared 130
samples for petrographic analysis. The thin sections were impregnated
with both normal blue dye and blue fluorescent dye and ground to
30 microns thick and finished with a polished surface (0.5 mm diamond
grit). Ultraviolet-light microscopy generally can reveal clusters of
TABLE 2.— Selected core samples used in study. Blakely #1 and Sims #2
cores are from Wise County, Texas. Neal #A-1-1 is from McCulloch
County, Texas, and Walker #D-1-1 is from San Saba County, Texas. All
samples are siliceous mudstones. Values of vitrinite reflectance are based on
values obtained from material in same cores at nearby depths or from other
cores in the area and from published maps by Pollastro et al. (2007).
Sample
Depth (m)
117.3
390.7
2167.4
2184.8
2187.5
2191.8
2196.4
2324.0
2361.9
Well Name
#TOC (%)
**VRo (%)
Neal #A-1-1
Walker #D-1-1
Blakely #1
Blakely #1
Blakely #1
Blakely #1
Blakely #1
Sims #2
Sims #2
***
***
4.05
3.08
2.91
6.62
2.51
***
2.86
, 0.50*
0.52
, 1.35
, 1.35
, 1.35
, 1.35
, 1.35
, 1.6
, 1.6
*
Estimated from Pollastro et al. (2007) and comparison with Walker #D-1-1.
From Jarvie et al. (2007); USGS unpublished data.
Other samples in area are organic rich.
#
From Humble Geochemical.
**
***
micropores that are impregnated with the blue fluorescent dye; however,
no pores were visible using a petrographic microscope that was equipped
with a mercury lamp ultraviolet light. Subsequently we prepared highly
polished thin sections (microprobe quality) for scanning electron
microscopy (SEM). However, these samples also proved unsuitable for
high-resolution imaging of mudstone pores. Standard grinding and
polishing methods of preparing mudstone thin sections, using fine grit and
power, produced surface topographic irregularities because of differential
hardness of components (Fig. 3A). These irregularities greatly exceed the
size of nanopores in the Barnett Shale and other black shales (e.g., Bowker
2003; Nelson 2009; this study). Thus, even polished thin-section samples are
inadequate for pore identification using SEM-based techniques.
We also examined broken sample chips using the SEM. At first these
samples appeared to show numerous pores, but when later compared with
samples prepared by Ar-ion-milling, discussed below, it became apparent
that these pits or holes were plucked artifacts resulting from sample
breakage. The breaks in the sample were not across grains, but around
grains, providing holes that could be mistaken as pores.
To eliminate these conventional preparation limitations, we utilized
argon-ion milling to produce a much flatter surface (Fig. 3B). The
technique is similar to a common method for preparation of electrontransparent-transmission electron microscope samples (e.g., Hover et al.
1996), although the milling is confined to one surface rather than two. Arion-beam milling produces surfaces showing only minor topographic
variations unrelated to differences in hardness of the sample but, rather,
to slight variations in the path of the Ar-ion beam. Operating the ionmilling system using an accelerating voltage of 5 to 7 kV and a gun
current of about 300 mA has proven effective in producing relatively flat
surfaces on mudstone samples making them suitable for very highmagnification imaging.
For this study, 33 ion-milled cut surfaces from 25 samples from 18 wells
were prepared (Table 1). Where possible, surfaces were cut perpendicular
or at a slight angle to bedding, but in a few samples the cut had to be at a
moderate angle to bedding. We used three ion-milling machines to
prepare surfaces and to test quality of milled surfaces (machines
manufactured by JEOL USA, Inc., Leica, and the Specimen Preparation
Group of Gatan, Inc.). Several samples from the same layers were made
to check for variability within samples. Although differences in cutting
equipment produced different sample-area outlines, surface topography
of flat areas of the samples was virtually identical. One sample was given
JSR
NANOPORES IN SILICEOUS MUDSTONES
851
FIG. 3.— Secondary electron (SE) images at same scale showing the difference in surface topography between A) a mechanically polished surface and B) an Ar-ionbeam cut surface. Note that the relief of the mechanically polished surface (A) exceeds the diameter of most shale pores. Blakely #1, 2,196.4 m.
a 20-nm-thick gold coat, and other samples were given a 4-nm-thick
platinum coat. Two samples were initially examined at very low kV
without conductive coatings on the cut surface, but they have
subsequently been coated with a 4-nm thickness of platinum and
reexamined.
Samples were examined using two different scanning electron
microscopes (SEM), each of which provided different imaging advantage.
The tungsten filament model was a Philips XL30 equipped with an
Oxford Instruments ISIS EDS system. The field emission SEM was a
Zeiss Supra 40 VP. Both systems are at the University of Texas at Austin.
Initial examination of the coated samples was on the standard tungstenfilament SEM, equipped with an energy dispersive spectroscopy (EDS)
system and a backscattered electron (BSE) detector. Secondary electron
(SE) images were acquired for documenting topographic variation, and
BSE images were acquired for delineating compositional variation. This
imaging provided important information on lithologic variation and
general locations of pores throughout the whole sample. Accelerating
voltage was typically 20 kV, with spot size varying depending on the
nature of imaging being carried out.
The resolution necessary to delimit the smallest pores (, 5 nm) was
not possible on tungsten-filament SEM systems. Additional imaging
using a field-emission-gun SEM equipped with an in-lens SE detector
provided greatly increased detail of nanometer-scale features. Low
accelerating voltages (1–5 kV) were typically used on this system to
prevent beam damage and also to allow examination of uncoated
surfaces. Working distances were 3 to 6 mm.
Measurements were made from SEM photographs on all pore types
observed in the mudrock samples. Several samples were selected for size
analysis (Table 3). Detailed measurements and analyses were made
primarily for nanopores within grains of organic matter because these
pores are most common and relevant to this investigation. Pore sizes and
pore-size distribution in the grains of organic matter were determined by
using computer software (JMicrovison) to outline and measure all
individual pores in an area of interest. Pore diameters and aspect ratios
were determined for each pore, and average and median diameters were
determined for groups of pores. Measuring a three-dimensional
ellipsoidal object or space from a two-dimensional plane can create
errors in calculating the true dimensions or amounts of the object or space
(Halley 1978; Johnson 1994). These ellipsoidal features are generally
undercounted. The nanopores in the grains of organic matter in the
Barnett mudstones are multishaped, and in some organic grains the pores
have preferred alignment. Both of these factors make accurate corrections
of analyses impossible. However, the fact that the mean pore size or the
volume is underestimated is taken into account in the discussion section
on nanopore abundance later in the paper.
Image-based porosities were measured by point counts of SEM images
for selected areas (Table 4). In some cases, either multiple-image mosaics
were used for increased resolution, or parts of images were used to focus
on organic matter. Several hundred to more than 3300 points on each
image were used for pore counts. We completed analyses on 28
representative areas of interest.
Five core plugs from the Blakely #1 well were analyzed for porosity,
permeability, nuclear magnetic resonance (NMR), and capillary pressure
by Core Laboratories Advanced Technology Center. Details of porosity
and permeability are shown in Table 5. All analyses were run on the same
plug so that the analyses could be interrelated. The cores from the Blakely
852
JSR
R.G. LOUCKS ET AL.
TABLE 3.— Pore-size calculations in organic matter in Barnett mudstones.
Mean and median lengths for gains are listed. Samples were quantitatively
analyzed using JMicrovison software.
TABLE 4.—Porosities calculated within organic matter in Barnett mudstone
from SEM images.
Sample Name and Depth (m)
Sample Name
and Depth (m)
Mean Length
(nm)
Median Length
(nm)
Count
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Sims #2 2324.0
Sims #2 2324.0
Sims #2 2324.0
Sims #2 2324.0
Sims #2 2324.0
112.80
144.09
185.39
179.84
162.33
167.41
138.87
32.80
26.15
20.33
21.88
90.59
82.78
72.08
78.28
90.59
86.74
105.09
160.51
128.24
137.84
138.48
110.04
29.05
22.83
15.59
17.30
71.32
64.59
58.53
66.94
71.32
1015
269
267
60
117
100
468
30
80
227
100
197
138
52
45
197
#1 well were obtained in 1985 and were not preserved in their natural
state. Dewhurst et al. (2002) pointed out that cores which have dried can
introduce uncertainty in measurements of capillary pressure. Dehydration
of sample material may have affected NMR measurements also. The
reader is referred to Dewhurst et al. (2002) for a discussion on the effects
of drying samples.
According to Core Lab methodology, porosity and permeability were
determined from core plugs processed to remove residual reservoir fluids
using a reflux soxhlet. The samples were then dried in a vacuum oven at
220uF and cooled to room temperature in a moisture-free environment.
The samples were placed in a pulse decay permeameter to measure
porosity and permeability at both 800 psi and 2500 psi. Porosity was
calculated using Boyle’s law of gas expansion, and Klinkenberg
permeability was determined by pressure decay where the pressure was
vented at a known rate.
Nuclear magnetic resonance (NMR) analyses were conducted on three
samples. Core Lab’s process included first vacuum drying the samples at
93.3uC then pressure saturating the sample with a formation brine.
Measurements were made on a MARAN Ultra instrument operating at a
proton Lamor frequency of approximately 2 MHz. T2 (transverse
relaxation time) was measured at an interecho spacing of 0.26 ms. The
Carr-Purcell-Meiboom-Gill pulse sequence was used. The signal-to-noise
ratio for the NMR measurements was a minimum of 100:1.
Five core plugs were analyzed for analyses of capillary pressure using
high-pressure mercury injection. Core Lab used a Micromeritics
AutoPore mercury injection instrument for testing. Samples were
subjected to an injection pressure of up to 55,000 psia.
Humble Geochemical provided total organic carbon (TOC) data for
numerous mudrock samples in the Fort Worth Basin. In the northern
Fort Worth basin area, mean TOC is 4% with a range of 0.4 to 10.6%. In
the southern Fort Worth basin area, mean TOC is 5.1% with a range of
0.2 to 11.3%. Measurements of samples presented in this paper are as
follows; VRo for the Blakely #1 well core samples average 1.35% (USGS
unpublished data; Table 2), whereas in the Sims #2 well core samples,
VRo is about 1.6% (Jarvie et al. 2007; Table 2). The lower VRo sample
from the Walker #D-1-1 well measured 0.52% (USGS unpublished data;
Table 2), and the lower VRo sample from the Neal #A-1-1 well is
estimated at , 0.50% on a map from Pollastro (2007) (Table 2).
Fifty-two X-ray diffraction analyses were obtained on Barnett
mudrock samples throughout the Fort Worth Basin. Samples were
analyzed by Omni Laboratory and N. Guven, Texas Tech University,
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2167.4
Blakely #1 2196.4
Blakely #1 2196.4
Blakely #1 2196.4
Blakely #1 2196.4
Sims #2 2324.0
Sims #2 2324.0
Sims #2 2324.0
Sims #2 2324.0
Sims #2 2324.0
Sims #2 2324.0
Proprietary Well 2604.5
Porosity (%)
Count
20.15
5.60
19.43
1.94
3.30
0.25
4.40
17.00
11.30
10.10
14.64
22.50
20.15
5.60
19.43
17.00
11.30
12.40
4.41
5.00
5.20
7.77
20.24
30.00
26.45
24.00
18.20
19.33
2000
750
628
1910
1000
2000
3387
1200
1000
1406
1250
200
2000
750
628
1200
1000
500
1020
1000
500
3050
850
800
1002
400
500
750
Lubbock, Texas. Some of these data are presented in Loucks and Ruppel
(2007). Mineralogy differs between the northern and southern parts of the
Fort Worth Basin (Table 6), the southern area being generally more clay
rich and silica poor.
BARNETT SILICEOUS MUDSTONE PORE TYPES
Pore-Size Groups
Thirty-three core samples from the siliceous mudstone lithofacies of the
Barnett Shale of the northern Fort Worth Basin were analyzed using the
imaging methods described earlier (Table 1). After a detailed petrographic and SEM study of numerous samples, several types of pores were
recognized and grouped into two general categories on the basis of size:
micropores (pores having diameters $ 0.75 mm, Fig. 4) and nanopores
(pores having diameters , 0.75 mm, Figs. 4, 5, 6). Nanopores, which are
located within organic matter, and less commonly in pyrite, are by far the
most abundant pore type in our Barnett samples. Pores located in organic
matter are termed intraparticle organic nanopores in this study.
Nanopores were also observed in bedding-parallel, wispy, organic-rich
laminae in fine-grained-matrix areas in which pores exist both within
grains of organic matter and between them.
Although fracture porosity has been proposed as a storage and
transport mechanism for hydrocarbons in shales (e.g., Dewhurst et al.
1999), only one naturally occurring microfracture containing porosity has
been found in the Barnett Shale, despite extensive searching using a
variety of megascopic and microscopic investigative techniques. Cemented microfractures and fractures are present, however, particularly in
carbonate-rich mudstones (Gale et al. 2007).
Micropores
Most micropores are associated with whole microfossils, fragmentary
fossil material, or pyrite framboids. A few primary intragranular pores
JSR
NANOPORES IN SILICEOUS MUDSTONES
853
TABLE 5.— Analysis of porosity and permeability from the Blakely #1 core samples. Two samples did not produce reliable permeability analysis.
Klinkenberg permeabilities associated with capillary pressure stubs are generally more optimistic because only a thin slice of the core plug was used for
analysis and the analyses were not conducted at higher confining pressures.
Depth (m)
Confining Pressure (psi)
Porosity (%)
800
2500
800
2500
800
2500
800
2500
800
2500
7.6
6.5
0.1
2167.4
2175.7
2184.8
2187.6
2196.4
Core-plug Klinkenberg
Permeability (Microdarcys)
Capillary-pressure Klinkenberg
Permeability (Microdarcys)
4.0
1.23
1.0
0.06
0.7
6.0
Not suitable
2.5
1.3
3.2
3.2
associated with the body cavities of fossils such as foraminifera are
present. However, most fossil-related, primary intraparticle pores are
filled with carbonate, silica, and/or pyrite cement. In some shell-rich
layers of the Barnett mudstone, fossils that have been replaced by silica
have small amounts of intragranular porosity (Fig. 4A). Micropores are
also associated with diagenetic minerals, such as pyrite (Fig. 4B) or
quartz, which have incompletely filled small voids possibly left by algal
spores (e.g., Tasmanites). Secondary pores formed by dissolution along
cleavages in silt-sized feldspar have also been observed. Nano- to
microintercrystalline pores within pyrite framboids vary with the size of
the framboids. Smaller framboids (2–10 mm in size) contain pores that
typically range in size from 0.05 to 1 mm (Fig. 4B), whereas larger
framboids display pores ranging from 1 to 5 mm in diameter. Pore shapes
are generally polyhedral with straight margins. As a group, micropores
are relatively rare in Barnett mudstones except for those in pyrite
framboids.
Nanopores
Interparticle Nanopores
Interparticle nanopores (pores between grains) are rare. Those
observed were developed at the margins of larger grains (Fig. 4C) and
dispersed within very fine-grained matrix (Fig. 4D). Interparticle pores
associated with grain boundaries tend to be relatively large (long
diameters of hundreds of nanometers) and are more common in siltbearing laminae. Although these pores are uncommon overall, clusters
are observed locally in some samples.
TABLE 6.—Average composition of mudstone in the Fort Worth Basin from
X-ray diffraction analysis. Category ‘‘Other’’ in XRD data includes albite,
apatite, plagioclase, and K-feldspar. Carbonates are calcite, dolomite,
and siderite.
Mineral
Southern Area
(22 Samples)
(% Whole Rock)
Northern Area (35 Samples)
(% Whole Rock)
Quartz
Carbonate
Pyrite
Mixed layer (I/S)
Smectite
Kaolinite
Chlorite
Illite/mica
Others
25
8
5
18
0
3
8
31
2
35
17
12
20
1
1
1
13
0
5.0
0.37
Fractured
Intraparticle Organic Nanopores
Intraparticle organic nanopores (pores within grains; Fig. 5) constitute
the most widespread and numerous pore type in the Barnett Shale.
Shapes vary from nearly spherical (Fig. 5A) to irregularly polygonal
(Fig. 5B), with slightly irregular ellipsoids being the most common shape
(Fig. 5A).
CHARACTERIZATION AND DISTRIBUTION OF ORGANIC MATTER
Because most of the pores in the Barnett mudstones are nanopores and
nanopores are associated with organic matter, distribution of this organic
matter is crucial to an understanding of overall pore networks and
permeability. TOC in Barnett rocks from the producing area varies from
0.4 to 10.6%, with an average of 4.0% (Loucks and Ruppel 2007).
Samples from the southern Fort Worth Basin exhibit similar values. We
have observed from petrographic and SEM analyses that the volumetric
amount of organic matter varies widely in each sample, even at the
millimeter scale. Variations in organic matter in the Barnett mudstones
seem tied to laminations within the mudstones; some laminations are
organic matter rich and some lack organic matter almost entirely.
Size and shape of grains of organic matter vary widely. Grain diameters
observed in the samples studied vary from less than 1 micron to tens of
microns. Grain shapes vary from rounded or nonrounded equant to platy
to irregularly angular. The most complex grains have shapes suggesting
that their formation is related to early compaction and are elongate
parallel to bedding. The grains of organic matter may be pieces of
transported terrestrial plants, algal matter, or remnants of other marine
organisms. Many are structureless grains for which assigning any source
would be difficult. Geochemical analysis of the Barnett organic matter
(Montgomery et al. 2005) has shown that kerogen is dominantly type II
(algal) in the Barnett Shale. Large plant fragments, as much as 1 cm in
diameter, have been observed in core but were not intersected by any of
the milled surfaces.
DESCRIPTION OF NANOPORES
Occurrence and Distribution
Nanopores are typically found in three modes of occurrence within the
mudstone, two of which are associated with organic matter. Most
nanopores are found in discrete grains of organic matter (Fig. 5B, C).
Additional nanopores are found among bedding-plane-parallel wisps of
organic-rich matter (Fig. 5D) and in association with extremely finegrained matrix material, but not in direct contact with any grains of
organic matter (Fig. 4D). Because the vast majority of nanopores are
854
R.G. LOUCKS ET AL.
JSR
FIG. 4.— Secondary electron images of pores in nonorganic matter. A) Silicified fossil fragment (dashed line) hosting numerous intragranular micropores. Arrows
indicate area of micropore development. Blakely #1, 2,167.4 m. B) Pyrite framboid containing intercrystalline micropores and nanopores. T.P. Sims #2, , 2,324 m.
Accelerating voltage 5 4 kV; working distance 5 5 mm. C) Nanopores rimming a nonorganic grain. Blakely #1, 2,196.4 m. Accelerating voltage 5 2 kV; working
distance 5 4 mm. D) Nanopores in nonorganic matrix. Alignment of pores may indicate some grain-boundary control. Blakely #1, 2,167.4 m.
found within organic matter, particularly grains, description of this
occurrence is emphasized in the following sections.
Morphology
Intraparticle organic nanopores most commonly have somewhat
irregular, ellipsoidal shapes, but other morphologies are also present
(Fig. 5A). Elliptical sections are the most common pore outline (Figs. 5A,
6A). However, at higher magnification and increased resolution, many
pore shapes tend to appear less elliptical and show more convolute edges
(increased rugosity). In some organic grains, shapes are rounded but are
not simple ellipsoids (Fig. 6C). These complex pores may result from
coalescence of multiple ellipsoidal pores and in some cases appear
colloform to botryoidal. Polyhedral angular pores, some triangular in
cross section, are dominant in other organic grains (Fig. 6B).
Nanopores are generally nonequant to some degree. In different
organic grains, average aspect ratios of nanopores show variation from
1.8:1 to 4.1:1. Average aspect ratios (eccentricity) of sets of nanopores
within organic grains have a mean of 2.8:1. The vast majority of
intraparticle organic nanopores do not have triangular cross sections, as
do interparticle pores in clastic rocks. Nor do most nanopores resemble
the parallel-elongate secondary pores commonly formed by dissolution,
such as in feldspars.
Overall, nanopores in organic matter from the Sims #2 core samples
show better ordered, less rounded, more equant pore shapes (Fig. 6C)
than do other samples. Some nanopores from this sample show rectilinear
alignment, which seems to be related to underlying structure or
heterogeneity within grains. These patterns are similar to those observed
in some plant materials and coal macerals (i.e., Hower et al. 1999),
although pores sizes in our samples are much smaller (nanometer scale
rather than micron scale). Whereas simple internal structures are typical
(Fig. 6A), some pores can be complex. Many larger pores have internal
textures, such as columnar structures (Fig. 6B).
Connectivity between pores controls permeability of the rock and is a
key component of the movement of fluids and gases through these
siliceous mudstones. In some examples (Fig. 6C, D), long, narrow throats
were observed connecting larger nanopores. The narrow width and
shallow depth of these conduits make them difficult to resolve in images.
Traces of throats are generally straight to smoothly curving. Throat
widths are less than 20 nm, and lengths greater than 200 nm have been
JSR
NANOPORES IN SILICEOUS MUDSTONES
855
FIG. 5.— Nanopores associated with organic matter in the Barnett Shale. A) Elliptical to complexly rounded nanopores in an organic grain. Darker materials are
organics. BSE image. Blakely #1, 2,167.4 m. B) Angular nanopores in a grain of organic matter. SE image. Blakely #1, 2,167.4 m. Accelerating voltage 5 10 kV;
working distance 5 6 mm. C) Rectangular nanopores occurring in aligned convoluted structures. SE image. T.P. Sims #2, , 2,324 m. Accelerating voltage 5 2 kV;
working distance 5 3 mm. D) Nanopores associated with disseminated organic matter. Carbon-rich grains are dark gray; nanopores are black. SE image. T.P. Sims #2,
, 2,324 m. Accelerating voltage 5 2 kV; working distance 5 2 mm.
observed. These extremely small pore-throat widths are consistent with
pore-throat sizes calculated from capillary-pressure analysis, showing that
most pore throat diameters peak between 10 and 15 nm and range from
less than 5 to 100 nm in diameter (Fig. 7). Our pore-throat-size data for
Barnett Shale nanopores agrees with shale pore-throat-size data presented
by Nelson (2009). Nelson compiled studies that show pore throats range
from 8 to 17 nm in diameter in Jurassic–Cretaceous Scotian Shelf shales,
7 to 24 nm in diameter in Devonian Appalachian Basin shales, 9 to about
45 nm in diameter in Pliocene Beaufort–Mackenzie Basin shales, and 20
to about 160 nm in diameter in Pennsylvanian Anadarko Basin shales.
Pore Size
Determining average diameters of intraparticle organic nanopores is
complicated by the nonspherical shape of some pores. Many larger, more
complex nanopores appear to have composite shapes, formed by
amalgamation of smaller pores during pore growth. Most such complex
pore shapes were treated as separate pores for measurement. Also, as
mentioned in the Methods and Samples section, measuring pore sizes in a
two-dimensional plane is not totally accurate because the maximum
dimension is not always intersected.
Measurements have been made of mean and median apparent pore
diameters from various images and image mosaics. Mean and median
intraparticle organic nanopore diameters vary from grain to grain
(Table 3). Median pore diameters for nanopore groups are typically less
than mean sizes, and peaks on histograms of diameters are typically
slightly less than the median value (Fig. 8). These numerical relationships
apply for almost all histograms, even though for a few measured areas,
smaller pores may be underrepresented in the data owing to resolution
limitations.
Although some small grains of organic matter lack larger nanopore
sizes (. 30 nm), this lack in size may be a function of limited surface area
or absolute volume of the smaller grains. Large organic grains generally
show a wide variability in pore diameter, and very small nanopores
(, 30 nm) are always present.
Nanopores in grains of organic matter from the Barnett siliceous
mudstone lithofacies have an approximate median size of 100 nm and
range in size from as low as 5 nm to greater than 800 nm (Fig. 8A).
856
JSR
R.G. LOUCKS ET AL.
FIG. 6.—Secondary electron images showing variations in nanopore morphology in organic matter. A) Very small (18–46 nm diameters), nearly spherical nanopores.
Total porosity in this field of view is 5.2%. Blakely #1, 2,196.4 m. Accelerating voltage 5 4 kV; working distance 5 5 mm. B) Larger nanopore (550 nm diameter)
showing complex internal structure resembling pillars. Blakely #1, 2,167.4 m. Accelerating voltage 5 10 kV; working distance 5 6 mm. C) Tubelike pore throats
connecting elliptical pores (white arrows). Pore-throat diameter , 20 nm. Blakely #1, 2,167.4 m. D) Additional tubelike pore throats connecting elliptical pore (white
arrows). Pore-throat diameter , 20 nm. Blakely #1, 2,167.4 m.
Composite intraparticle organic nanopores rarely approach 1 micron in
diameter. Average nanopore diameters in individual grains vary from 20
to 185 nm, and median sizes range from 15 to 160 nm. Peaks on size
histograms have approximately the same range as median diameters.
Rare organic grains may contain a bimodal size distribution of pores
(Fig. 8B).
NMR measurements were conducted on three samples from the
Blakely #1 well (Fig. 9). As noted in the methods section, these samples
were resaturated with brine for analyses. T2-relaxation time curves
produced from these measurements provide an estimation of pore-size
distribution (e.g., Kenyon 1992; Basan et al. 1997), but they provide no
quantitative data on actual pore sizes. Overall shape of the point-count
histograms shown in Figure 8 reveals a relative pore-size distribution
similar to that of the NMR curves in Figure 9. This apparent correlation
suggests that the high peak on the NMR curves records the dominant
intraparticle organic nanopores, whereas the secondary lower peak at the
higher T2-relaxation times records larger intraparticle organic nanopores,
along with rarer micropores. The similarity of the NMR data to our poreimaging findings lends support to our interpretation of pore-size
distributions in the Barnett mudstones.
Pore Abundance and Size Distribution
Nanopore densities within grains of organic matter can be high
(Table 3). Grains containing hundreds of nanopores are common; in one
case a single 10.8-mm-diameter grain contained more than 1,000
nanopores of various shapes and sizes (Fig. 6C shows part of this grain).
Average intragranular nanopore diameter does not appear to be directly
related to the size of the organic grain. Both large and small nanopores
are typically found in organic grains of all sizes, although a few of the
smaller grains have only small nanopores, as discussed earlier.
In organic grains in which nanopores appear to be randomly
distributed, a slight decrease in pore density seems to occur adjacent to
grain edges. This decrease in density may be related to physical changes in
the material at grain edges so that pores are less likely to form.
In some organic grains with abundant nanopores, there are areas where
few or no pores are present (Fig. 5C). Also, a few organic grains in
samples with abundant nanopores show no nanopore development, and it
is unclear whether the lack of pore development is related to different
organic composition (i.e., inertinite) or other factors, such as late
localized compaction.
JSR
NANOPORES IN SILICEOUS MUDSTONES
857
FIG. 7.—Histograms of pore-throat diameters
calculated from four capillary-pressure sample
analyses of Barnett siliceous mudstones in the
Blakely #1 well. Note that most calculated pore
throat diameters fall in the 5–15 nanometer
range, which is consistent with pore-throat
dimensions measured from SEM images in this
paper (see Fig. 6). Porosities and permeabilities
shown on the graphs are from core-plug analyses
of the same samples.
Fewer intraparticle organic nanopores are associated with disseminated
organic matter (wisps) in the mudstone matrix (Fig. 5D). Size of this
organic matter is in the tens to hundreds of nanometers range, and it is
typically elongate parallel to bedding.
Laboratory measurements of porosity for four Blakely core-plug
samples range from 0.7 to 6.5 percent porosity (Table 5). Our calculations
of pore space from point counts of individual images and image mosaics
(tens to hundreds of square microns in area) from these same Blakely
FIG. 8.—Histograms of nanopore diameters for measurements of single grains of organic matter. A) Histogram of pore diameters for all pores within a large organic
grain. Blakely #1, 2167.4 m. B) Histogram of pore diameters for all pores within an organic grain showing bimodal distribution of nanopore sizes. T.P. Sims #2,
, 2,324 m.
858
JSR
R.G. LOUCKS ET AL.
Styles of nanopore occurrence, for example, relating to pore size, shape,
and distribution, vary among grains of organic matter. Style is consistent
within a grain, but not necessarily among grains in the same sample.
Pores in Low-Maturity Rocks
FIG. 9.— NMR T2-relaxation-time curves from three samples of Barnett
siliceous mudstones from the Blakely #1 well. Curves show relative proportions
of pore sizes measured in core plugs (, 2.5 3 5 cm). No absolute pore-size scale
is provided by this method. Combining histogram data presented in Figure 8 with
NMR data, curves suggest a large population of nanopores with fewer larger
pores. Porosities shown on the graph are from core-plug analyses of the
same samples.
Eleven siliceous mudstone samples were examined from eight wells in
the low-thermal-maturity area of the Barnett Shale in the southern Fort
Worth Basin (Fig. 1), of which two are presented here (Fig. 10).
Mudrocks in this area (southern Fort Worth Basin) have VRo values
less than 0.7% (Pollastro et al. 2007) indicating that they are at the low
end of the hydrocarbon generating window. Pore systems in these samples
contrast greatly with those of higher-maturity samples (VRo values . 0.8%) from areas in the north part of the Fort Worth Basin.
The primary difference is that grains of organic matter in these less
mature samples have few or no pores and show no internal grain
heterogeneity (Fig. 10). The nanopores that are present in these
mudstones are elongate and parallel to organic grain boundaries
(Fig. 10A).
DISCUSSION
Origin and Significance of Organic Nanopores
samples (Table 4) reveal a wide range of porosities (4.45–22.5%). In
general, porosity (pore volume) values from core-plug analyses are higher
than estimations made from our pore-imaging analyses. Methods of
calculation of porosity values from SEM analysis are presented in the
Discussion section. Our overall impression is that limited pores were
developed or preserved outside of the organic matter, except for
intercrystalline pyrite pores.
Laboratory values are matched or exceeded only in sample estimations
made from clusters of pores within grains of organic manner. Nanopore
porosity values calculated for entire grains of organic matter ranged from
5.6 to 20%. Within grains showing nonuniform pore distribution, as much
as 30% porosity was calculated in pore-rich areas. In a few image areas,
the minimum pore size was defined by image resolution, not by actual
pore sizes (which were below resolution). Thus, our porosity calculations
may be low in these areas and should be viewed as minimum values.
The strong correlation between nanopore abundance in grains of
organic matter and vitrinite reflectance suggests that pore formation is the
result of thermal maturation and conversion of the organic matter (i.e.,
kerogen). This conclusion is supported by the absence of nanopores in
grains of organic matter from our lower-thermal-maturity samples and
their abundance in more thermally mature samples. This relationship is
consistent with observations made by Hover et al. (1996), on lowthermal-maturity rocks from the Antrim and New Albany Shales using
the transmission electron microscope. They found no visible intercrystalline or interparticle matrix porosity in these shales.
Our conclusions are also supported by work of Chalmers and Bustin
(2007). In a study of Cretaceous shales, they showed a strong relationship
between micropore volume, methane sorption capacity, and organicmatter content (see their fig. 9). Their samples with highest concentrations of inertodetrinite and vitrinite had highest methane sorption
FIG. 10.— Back-scatter electron images of pores in organic grains in low maturity samples. A) Organic grain contains no nanopores except a few interparticle
micropores (white arrows) along the rims of grains. VRo of organic matter in this area is less than 0.50% (Pollastro et al. 2007). Houston Oil & Minerals Neal #A-1-1,
117.5 m. Accelerating voltage 5 4 kV; working distance 5 5 mm. B) Large organic grain showing no nanopore development. VRo of organic matter in this area is
approximately 0.52% (Pollastro et al. 2007). Houston Oil & Minerals Walker #D-1-1, 390.8 m. Accelerating voltage 5 4 kV; working distance 5 5 mm.
JSR
NANOPORES IN SILICEOUS MUDSTONES
capacities. Although Chalmers and Bustin (2007) did not directly relate
origin of micropores to maturation of organic matter, the positive
correlation between organic matter and micropore abundance is
consistent with our findings.
Collectively, these data suggest that pores are formed as kerogen is
converted to hydrocarbons, resulting in the formation of liquids and gases
that coalesce into bubbles. Jarvie et al. (2007) theorized the following
scenario for porosity development from organic-carbon decomposition.
The TOC in a hydrocarbon source rock comprises two components. The
first has been described variously as ‘‘live carbon’’ (Pepper 1992),
‘‘pyrolysable carbon’’ (Espitalie et al. 1984), or ‘‘convertible organic
carbon’’ (Jarvie 1991). The second component has been termed by Cooles
et al. (1986) ‘‘inert carbon’’ or ‘‘dead carbon.’’ The principal discriminating factor between these two components is hydrogen content, which
depends initially on organic-matter type and preservation, but also on
thermal maturity during catagenesis. Convertible organic carbon content
will yield hydrocarbons and additional dead carbon upon thermal
maturation, whereas dead carbon has only a minor capacity to generate
dry gas. During thermal maturation and conversion of convertible
organic carbon, decomposition of organic matter leads to formation of
hydrocarbons and, simultaneously, intraparticle organic nanopores.
Pore growth during hydrocarbon maturation can explain many
features seen in association with intraparticle organic nanopores.
Formation of the long, narrow pore throats may have been driven by
differential pressure buildup within evolving adjacent pores owing to
hydrocarbon generation. More complexly shaped pores may result when
closely spaced pores expand to form a connection between one another.
Origin of more ordered nanopores (Fig. 5) is less intuitive. Lack of any
ordered nanopores in grains of organic matter in low-maturity samples
suggests that the ordered nanopores in the mature samples are not the
result of inherited, organic nanopores from the grain. However, relict
nanoscale to microscale heterogeneities within organic grains may
influence nucleation sites of hydrocarbon pores.
Intraparticle organic nanopores form over a period of time, corresponding to increases in temperature with burial. The process of
conversion of kerogen to hydrocarbons and associated pores starts at
low thermal maturity (at 0.60% VRo the conversion might be 10% by
weight; Peters 1986). The temperature range for Barnett Shale hydrocarbon generation is about 100 to 160uC, given the measured rates of
decomposition (i.e., kinetics) of Barnett Shale (Jarvie et al. 2007). At 120
to 140uC (given enough time), approximately 50% conversion of
convertible organic carbon in kerogen occurs (Jarvie et al. 2007). The
result is that generation of hydrocarbons in Barnett mudstones may
create a large number of secondary pores owing to conversion of organic
matter dispersed in the rock. This conversion method and the size of
nanopore produced are consistent with studies of high-maturity kerogen
by Behar and Vandenbroucke (1987), who reported pores of 5- to 50-nm
size, depending on the type of kerogen hosting the pores.
Variability in the source of organic matter could greatly influence
amount of porosity developed. Organic matter that produces pores
consists of convertible or live carbon components of the TOC—
hydrocarbon-prone parts of the organics matter—which has varying
amounts of hydrogen, depending on organic-matter type. Because
alginites have more hydrogen than vitrinites, one would expect a larger
pore-volume yield from the decomposition of alginites.
In relating pore abundance to maturity level, the timing of pore
formation relative to compaction is important. The convolute shape of
some organic grains (Fig. 5A) suggest that they have undergone at least
some compaction, but it was probably early. Pore formation certainly
postdates most compaction because compaction would have closed pores
or at least altered their shapes. Lack of preferred orientation to pore
elongation in most samples also indicates postcompaction.
859
Upscaling Intraparticle Organic Nanopore Volumes and Size Distributions
Analysis of the volume and size distribution of intraparticle organic
nanopores requires nanometer-scale-resolution images from the SEM.
This means that only very small areas (tens to hundreds of square
microns) can be analyzed efficiently at the necessary resolution. However,
we have developed a simple method of upscaling results from these small
areas to a volumetrically significant volume.
Our procedure for upscaling limited-area porosity calculations from
point-count data was to combine porosity measurements from pointcount data with measurable TOC in the sample. Using this approach, we
first calculated average porosity from the sample image area. We then
converted TOC from weight percent to volume percent, thus allowing the
porosity value to be applied to the total organic matter in the sample. For
example, in the Blakely #1 2,167.4-m sample, median porosity within
organic matter grains is 17.0%, and TOC of the sample is 4.05 weight %.
This converts to 8.42 volume % organic matter based on the assumption
that 1.2 g/cc is the density of organic matter and 2.5 g/cc is the density of
the inorganic matter. These data convert to an estimated porosity of 1.4%
(8.42 3 0.17), which should be considered a minimum because of the
exclusion of rare pores found outside the organic matter. This calculated
value does not compare well to the substantially higher porosity value of
6.5% measured by helium porosimetery for the Blakely #1 2,167.4 m
sample (Table 5). The calculated value is much higher than any pore
volume that can be accounted for by our image-based calculations. The
difference in values may be in part the result of undercountering pores in
a 2D plane, missing pores in other parts of the plug, problems with
conventional core-plug analysis on mudrocks, and/or not including
substantial pores in pyrite framboids.
Gas Storage and Permeability Pathways within Barnett Mudstone
This is the first study to document the types, size range, and
distribution of pores in the productive mudstone lithofacies of the
Barnett Shale. It is an important step toward understanding where gas is
stored and how it might migrate from the mudrocks into induced
fractures that enable production. Up to the present, only microfractures
have been suggested as a permeable flow pathway into induced fractures
(Curtis 2002). As noted earlier, open microfractures were not found in the
productive area of the Barnett Shale.
Montgomery et al. (2005) stated that in shale-gas systems, gas is stored
as adsorbed and nonadsorbed (absorbed and free) gas. In the case of
adsorbed gas, methane molecules are attached to surfaces of organic and
mineral material, such as phyllosilicates. In the nonadsorbed state,
methane molecules reside in pore spaces as either free gas or solution gas
in a liquid. Prior to our documentation of pores within the Barnett Shale,
characteristics and distribution of pores containing adsorbed and/or free
gas were not known. We have now shown that naturally occurring pores
in Barnett mudrocks are predominantly associated with organic matter
and pyrite framboids. It should be noted, however, that even though
mean pyrite abundance is high (average 9%), only framboidal pyrite
commonly contains pores. The volume of pores associated with pyrite
was not determined, although it seems to be far less then the volume of
pores in organic matter. Work is ongoing to further resolve the relative
distribution of pores in pyrite. Understanding the magnitude of this
complex pore network is essential to quantifying gas storage in Barnett
reservoirs.
Given that most pores are associated with organic matter, permeability
pathways should be greatly influenced, if not controlled, by the threedimensional arrangement of organic-matter grains. Connected organic
matter could enable limited flow, depending on the connectivity of the
nanopores within it. Preliminary observations suggest variations in the
distribution of organic matter among samples that could account for
variations in permeability.
860
JSR
R.G. LOUCKS ET AL.
FIG. 11.— Hypothetical relationship between arrangement of grains of
nanoporous organic matter and permeable flow paths. A) Dense and laterally
continuous layers produce best permeable flow paths. B) Sparse and laterally
continuous layers provide lesser permeable connectivity. C) Sparse and laterally
discontinuous layers do not produce permeable pathways.
Figure 11 illustrates diagrammatically some hypothetical relationships
between arrangement of nanoporous organic matter and permeable flow
paths that are based on organic matter having different concentrations
and layering characteristics. If organic matter is concentrated in laterally
continuous layers, it may produce relatively good permeability pathways
within Barnett mudstones. If the organic material is sparser but laterally
continuous (mesh-like network), these layers may produce modest
permeability. Finally, if the organic matter is sparse and discontinuous
or disseminated, it will have poor to no permeability associated with it.
Core-plug permeabilities (core-plug Klinkenberg permeability ranges
from 0.06 to 1.23 microdarcys) shown in Table 5 for Barnett mudstones
are higher than permeabilities measured in other mudstones (e.g., Neuzil
1994). Montgomery et al. (2005), however, reported Barnett Shale
permeability ranges from microdarcys up to 0.01 md. Reasons for these
higher values may be that the Barnett mudstones are better quality
reservoirs or that conventional core-plug measurements of permeability
are not valid for tight mudstones. Also, the samples not being preserved
in their natural fluids may have affected the measurements (Dewhurst et
al. 1999).
CONCLUSIONS
Our studies show that pores in siliceous mudstones of the Barnett Shale
are dominantly nanometer in scale, and that most nanopores are
associated with grains of organic matter. Pores not associated with
organic matter are far less common, except for intercrystalline pores
associated with framboidal pyrite. Our data also show that the abundance
of pores in organic matter is directly related to thermal maturity. Based
on these relationships we conclude that most nanopores formed in grains
of organic matter during thermal decomposition of organic matter during
hydrocarbon generation. Other uncommon pore types consisting of
intragranular, moldic, intercrystalline, and interparticle pores do not
appear to add to the effective porosity network. We therefore conclude
that the siliceous mudstone hydrocarbon reservoirs in the Mississippian
Barnett Shale of the northern Fort Worth Basin consist of a nanopore
network that is in part controlled by abundance and distribution of
organic matter and, possibly, pyrite. Also, gas flow may be a combination
of permeable layers of organic matter and diffusion.
Our image-based pore-volume calculations suggest that conventional
laboratory core-analysis porosity determinations of pore volume may be
too high. This may indicate that new methods are needed for accurate
characterization of pore volumes. Further work is clearly needed to
determine how well our imaging results compare to analyses obtained by
traditional porosity, permeability, capillary pressure, and NMR analyses.
Our SEM-based characterization of mudrock nanopores is a critical
first step in better understanding of the distribution and causes of pore
development in shale-gas systems. More research is needed to more fully
define the interrelationships between pore formation, nanopore distribution, and fluid flow in mudrocks. To fully verify the observations and
conclusions we have made, it will be critical to examine additional
mudrocks samples from both the Barnett and other mudrock successions.
It will also be important to develop improved methods for upscaling
SEM-scale to larger rock volumes. A deeper understanding of intraparticle organic nanopore development from different organic-matter types
is also critical. Data on organic-matter volume and distribution and their
relationships to pore volume and permeability will lead to a better
petrophysical understanding of mudrocks as well as a better estimation of
gas reserves.
ACKNOWLEDGMENTS
This research was funded primarily by the State of Texas Advanced
Resource Recovery Project (STARR). Preparation of some Ar-ion-beam-cut
samples was provided by JEOL USA, Inc., Leica, and the Specimen
Preparation Group of Gatan, Inc. Kitty Milliken is acknowledged for
suggesting ion-beam milling as a sample-preparation method, for providing
one sample, and for reviewing an earlier version of this manuscript. In
addition to facilities at the Bureau of Economic Geology, imaging for this
study was also carried out at the microbeam laboratories of two other
University of Texas at Austin units: the Department of Geological Sciences
and the Institute for Cellular and Molecular Biology. Humble Geochemical
provided total-organic-carbon data for samples. Necip Guven of Texas Tech
University performed XRD-based characterization of mineralogy. JMicroVision image analysis software was used on SEM images for determining
porosity and pore diameters. Core Laboratories Advanced Technology
Center conducted the petrophysical analyses and we thank Paul Martin and
Robert Lee for insight on how the samples were analyzed. Editing was
provided by Lana Dieterich, Bureau of Economic Geology. The authors
acknowledge reviews by David Dewhurst, Joe Macquaker, Paul McCarthy,
and an anonymous reviewer. Publication authorized by the Director, Bureau
of Economic Geology, Jackson School of Geosciences, The University of
Texas at Austin.
REFERENCES
BASAN, P.B., LOWDEN, B.D., WHATTLER, P.R., AND ATTARD, J.J., 1997, Pore-size data in
petrophysics: a perspective on the measurement of pore geometry, in Lovell, M.A.,
and Harvey, P.K., eds., Developments in Petrophysics, Geological Society of London,
Special Publication 122, p. 47–67.
BEHAR, F., AND VANDENBROUCKE, M., 1987, Chemical modelling of kerogens: Organic
Geochemistry, v. 111, p. 15–24.
BOWKER, K.A., 2003, Recent development of the Barnett Shale play, Fort Worth Basin:
West Texas Geological Society, Bulletin, v. 42, no. 6, p. 1–11.
CHALMERS, G.R.L., AND BUSTIN, R.M., 2007, The organic matter distribution and
methane capacity of the Lower Cretaceous strata of northeastern British Columbia,
Canada: International Journal of Coal Geology, v. 70, p. 223–239.
COOLES, G.P., MACKENZIE, A.S., AND QUIGLEY, T.M., 1986, Calculation of petroleum
masses generated and expelled from source rocks, in Leythaeuser, D., and Rullkotter,
J., eds., Advances in Organic Geochemistry 1985, Organic Geochemistry, v. 10, p.
235–245.
CURTIS, J.B., 2002, Fractured shale-gas systems: American Association of Petroleum
Geologist, Bulletin, v. 86, p. 1921–1938.
DEWHURST, D.N., JONES, R.M., AND RAVEN, M.D., 2002, Microstructural and
petrophysical characterization of Muderong Shale: application to top seal risking:
Petroleum Geoscience, v. 8, p. 371–383.
DEWHURST, D.N., YANG, Y., AND APLIN, A.C., 1999, Permeability and fluid flow in
natural mudstones, in Aplin, A.C., Fleet, A.J., and Macquaker, J.H.S., eds., Muds
and Mudstones: Physical and Fluid Flow Properties, Geological Society of London,
Special Publication 158, p. 23–43.
EIA (ENERGY INFORMATION ADMINISTRATION), 2006, U.S. crude oil, natural gas, and
natural gas liquids reserves 2005 annual report: EIA: http://www.eia.doe.gov/pub/
oil_gas/natural_gas/data_publications/crude_oil_natural_gas_reserves/current/pdf/
appb.pdf (accessed April 2008).
JSR
NANOPORES IN SILICEOUS MUDSTONES
ESPITALIE, J., MADEC, M., AND TISSOT, B., 1984, Geochemical logging, in Voorhees, K.J.,
ed., Analytical Pyrolysis—Techniques and Applications: Boston, Butterworth, p.
276–304.
GALE, J.F.W., REED, R.M., AND HOLDER, J., 2007, Natural fractures in the Barnett Shale
and their importance for hydraulic fracture treatments: American Association of
Petroleum Geologists, Bulletin, v. 91, p. 603–622.
GUTSCHICK, R., AND SANDBERG, C., 1983, Mississippian continental margins on the
conterminous United States, in Stanley, D.J., and Moore, G.T., eds., The Shelf Break:
Critical Interface on Continental Margins, Society of Economic Paleontologists and
Mineralogists, Special Publication 33, p. 79–96.
HALLEY, R.B., 1978, Estimating pore and cement volumes in thin section: Journal of
Sedimentary Petrology, v. 48, p. 642–650.
HOVER, V.C., PEACOR, D.R., AND WALTER, L.M., 1996, Relationship between organic
matter and authigenic illite/smectite in Devonian Black Shales, Michigan and Illinois
Basins, USA, in Crossey, L.J., Loucks, R.G., and Totten, M.W., eds., Siliciclastic
Diagenesis and Fluid Flow: Concepts and Applications, SEPM, Special Publication
55, p. 73–83.
HOWER, J.C., STANTON, R.W., GAMMIDGE, L.C., AND HUTTON, A.C., 1999, Coal
petrology, in Papp, A.R., Hower, J.C., and Peters, D.C., eds., Atlas of Coal Geology.
American Association of Petroleum Geologists, Studies in Geology 45, CD-ROM.
JARVIE, D.M., 1991, Total organic carbon (TOC) analysis, in Merrill, R.K., ed., Treatise
of Petroleum Geology: Handbook of Petroleum Geology, Source and Migration
Processes and Evaluation Techniques: Tulsa, American Association of Petroleum
Geologists, p. 113–118.
JARVIE, D.M., HILL, R.J., RUBLE, T.E., AND POLLASTRO, R.M., 2007, Unconventional
shale-gas systems: the Mississippian Barnett Shale of north central Texas as one model
for thermogenic shale-gas assessment, in Hill, R.J., and Jarvie, D.M., eds., Special
Issues: Barnett Shale, American Association of Petroleum Geologists, Bulletin, v. 91,
p. 475–499.
JOHNSON, M.R., 1994, Thin section grain size analysis revisited: Sedimentology, v. 41, p.
985–999.
KENYON, W.E., 1992, Nuclear magnetic resonance as a petrophysical measurement:
Nuclear Geophysics, v. 6, p. 153–171.
861
LOUCKS, R.G., AND RUPPEL, S.C., 2007, Mississippian Barnett Shale: lithofacies and
depositional setting of a deep-water shale-gas succession in the Fort Worth Basin,
Texas: American Association of Petroleum Geologists, Bulletin, v. 91, p. 579–601.
MONTGOMERY, S.L., JARVIE, D.M., BOWKER, K.A., AND POLLASTRO, R.M., 2005,
Mississippian Barnett Shale, Fort Worth Basin, north-central Texas: gas-shale play
with multi-trillion cubic foot potential: American Association of Petroleum
Geologists, Bulletin, v. 89, p. 155–175.
NELSON, P.H., 2009, Pore throat sizes in sandstones, tight sandstones, and shales:
American Association of Petroleum Geologists, Bulletin, v. 93, p. 1–13.
NEUZIL, C.E., 1994, How permeable are clays and shales?: Water Resources Research,
v. 30, p. 145–150.
PEPPER, A.S., 1992, Estimating the petroleum expulsion behavior of source rocks: a
novel quantitative approach, in England, W.A., and Fleet, A.L., eds., Petroleum
Migration, Geological Society of London, Special Publication 59, p. 9–31.
PETERS, K.E., 1986, Guidelines for evaluating petroleum source rock using programmed
pyrolysis: American Association of Petroleum Geologists, Bulletin, v. 70, p. 318–329.
POLLASTRO, R.M., JARVIE, D.M., HILL, R.J., AND ADAMS, C.W., 2007, Geologic
framework of the Mississippian Barnett Shale, Barnett-Paleozoic total petroleum
system, Bend arch–Fort Worth Basin, Texas: American Association of Petroleum
Geologists, Bulletin, v. 91, p. 405–436.
REED, R.M., AND LOUCKS, R.G., 2007, Imaging nanoscale pores in the Mississippian
Barnett Shale of the northern Fort Worth Basin (abs): American Association of
Petroleum Geologists, Annual Convention, Abstracts Volume, v. 16, 115 p.
ROWE, H.D., LOUCKS, R.G., RUPPEL, S.C., AND RIMMER, S.M., 2008, Mississippian
Barnett Formation, Fort Worth Basin, Texas: bulk geochemical inferences and MoTOC constraints on the severity of hydrographic restriction: Chemical Geology,
v. 2571–2, p. 16–25.
RUPPEL, S.C., AND LOUCKS, R.G., 2008, Black mudrocks: lessons and questions from the
Mississippian Barnett Shale in the southern Midcontinent: The Sedimentary Record,
v. 6, no. 2, p. 4–8.
Received 28 April 2008; accepted 25 June 2009.