COGEL-02129; No of Pages 11 International Journal of Coal Geology xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

International Journal of Coal Geology journal homepage: www.elsevier.com/locate/ijcoalgeo

Application of BIB–SEM technology to characterize macropore morphology in coal S. Giffin a, R. Littke a,⁎, J. Klaver b, J.L. Urai b a Institute of Geology and Geochemistry of Petroleum and Coal, Energy and Mineral Resources Group (EMR), RWTH Aachen University, Lochnerstrasse 4-20, Haus B, 52056 Aachen, Germany b Structural Geology, Tectonics and Geomechanics, Energy and Mineral Resources Group (EMR), RWTH Aachen University, Lochnerstrasse 4-20, 52056, Aachen, Germany

a r t i c l e

i n f o

Article history: Received 21 December 2012 Received in revised form 18 February 2013 Accepted 19 February 2013 Available online xxxx Keywords: Coal porosity Broad-ion-beam milling SEM imaging Pore morphology Pore size distribution Vitrinite maceral Inertinite maceral

a b s t r a c t We use broad ion beam (BIB) milling to prepare low-relief polished surfaces of coal samples for high-resolution SEM imaging, in a study of the morphology and distribution of macro- and mesopores. Results show that the BIB-sections of a few square millimeters are not large enough to be statistically representative so that porosity was investigated as a function of maceral type. For a vitrinite maceral type, we found comparably little visible macroporosity within the resolution limits of the SE detector. Less than 2% of all the meso- and macropores studied were found in vitrinites. Pore morphology in an inertinite maceral is dependent on the original maceral. Fusinite yields large, elongated pores (often filled with mineral phases), while macrinite shows comparatively smaller, rounder pores. The distribution of pore sizes follows a similar power law at different magnifications. Our results show that micropores and macropores in coal belong to different populations, with different size distributions and morphologies. BIB–SEM imaging is a useful tool to study meso- and macropore morphology, especially in the size range between 10 nm and 10 μm, but more maceral types should be characterized for a better characterization of maceral porosity at different stages of coalification. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Coal is a naturally fractured reservoir, and the nature of these fractures plays an important role in the development and production of coalbed methane. Natural gas flow in a coalbed is commonly modeled in two phases (Karacan and Okandan, 2000; Yao and Liu, 2009). First, adsorbed gas in the coal matrix must diffuse to the fracture, or cleat, network. Once the gas molecules reach the cleat network, the second phase of flow, namely viscous flow, becomes the dominant parameter. Assessing the structural features of the coal matrix and cleat network facilitates in characterizing coal as a porous medium for natural gas recovery. Coal is an extremely heterogeneous material that can be characterized by its matrix structure and the pores within this structure. Porosity evaluation would be improved by a better understanding of its microstructural controls. For example, Gamson et al. (1993) described two types of matrix porosity: 1, pores found within the macerals and 2, pores located in between macerals, clays, or mineral matter. Other studies have shown that the relative abundance of micro-, meso-, and macropores is related to organic composition and coal type (Clarkson and Bustin, 1999; Gan et al., 1972; Harris and Yust, 1976; Unsworth et al., 1989). Micropores are less than 2 nm in diameter, while mesopores cover a range of 2–50 nm in diameter; macropores are greater than

⁎ Corresponding author. Tel.: +49 241 809 5748; fax: +49 241 809 2152. E-mail address: [email protected] (R. Littke).

50 nm in width (Rouquerol et al., 1994). Micropores make up the bulk of coal porosity, but mesopores and macropores can contribute significantly to the total pore volume (Clarkson and Bustin, 1996; Gan et al., 1972). Microporosity is contained in vitrinite-rich coals, determined by adsorption techniques (Clarkson and Bustin, 1996), but little work has been done to visually prove that vitrinite macerals contain very few meso- or macropores. Porosity in coal is often described by the pore volume distribution, determined by either mercury porosimetry, gas adsorption, liquid adsorption, or varied optical methods, used to ultimately determine the total porosity for a certain class of pore size. Integrating the results of several methods is still not without limitations. For example, gas adsorption and mercury porosimetry only provide information regarding connected pores with minimum pore and throat radii (Prinz and Littke, 2005). Scattering and diffraction, such as small angle neutron scattering (SANS), compute pore size distributions based on a spherical pore geometry, and porosity is the sum of both open and closed pores (Prinz et al., 2004; Radlinski et al., 2004). The pore size distribution results obtained by nuclear magnetic resonance (NMR) are difficult to interpret (Yao et al., 2010a) or require a sample specific fitting parameter (Yao et al., 2010b). Since nitrogen adsorption is often used to examine the meso- and macropore size distributions (Gan et al., 1972; Gensterblum et al., 2009; Prinz et al., 2004), this paper uses the pore size distribution from individual macerals and compares the results to nitrogen adsorption. Coal is known to display anisotropic characteristics. Massarotto et al. (2003) studied the effects of anisotropy on coal permeabilities and

0166-5162/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.coal.2013.02.009

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

2

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

found that the investigated coals displayed a greater permeability in the face cleat direction. Coal permeability is more strongly influenced by flow through the cleats than through the pore systems. Nevertheless, cleats and pores are still affected by the same paleo-tectonic settings. In a recent work quantifying pore size distributions using SAXS and SANS, Radlinski et al. (2004) showed that pores perpendicular to bedding are smaller than those in the bedding plane. In this study, we will quantify pore sizes in the bedding plane and perpendicular to it using visual techniques. The characterization of pore morphology, e.g. size distributions or shapes, has recently become the focus of various studies. SEM imaging can provide important information on pore structures. However, pores located in a rough surface, due to mechanical polishing or even with resin-enhancement, are often difficult to interpret (Desbois et al., 2011). Better results have recently been obtained combining either focused ion beam (FIB) or broad ion beam (BIB) sample preparation with high resolution scanning electron microscopy (SEM) to obtain detailed images of clear pore structures (Chalmers et al., 2012; Curtis et al., 2010, 2012; Desbois et al., 2009; Houben et al., 2013; Loucks et al., 2009). These studies investigated the porosity of gas shales or claystones. As of yet, little attention has been given to the application of the two methods to coals. This paper utilizes BIB sample preparation with SEM imaging on macerals in coal to investigate the pore size distribution of meso- and macropores, which are a part of the fluid pathway when determining reservoir permeability, as well as the pore shape, which plays a role in many volumetric reservoir calculations. 2. Methods

and references within for depositional conditions). The burial history of the coal-bearing Carboniferous succession has been summarized by Karg et al. (2005). Characteristics of the studied coal samples are given in Table 1. To ensure that the areas selected for the pore analysis of maceral type were actually of the desired maceral group, i.e. vitrinite or inertinite, sample areas of a core were first polished and then examined under the optical microscope (Fig. 1A, step 1). An image of the core slice was acquired and used as a guide to maceral selection (Fig. 1A, heavy black box). The maceral was BIB polished and SEM images were acquired of the maceral, as viewed parallel to the bedding plane (step 2). Then the sample was rotated 90°. This surface was then BIB polished and examined under the SEM; thus, allowing us to investigate the same maceral perpendicular to bedding (step 3). Based on core slice images (e.g. Fig. 1A, step 1), the specific maceral type of each maceral group can be identified. For the bright lithotype chosen from Well 1, a fusinite particle was chosen to represent the inertinite maceral group and a telocollinite was chosen to represent the vitrinite maceral group. For the dull lithotype from this well, macrinite was the inertinite maceral chosen and telocollinite was the vitrinite maceral chosen. All BIB-milled and examined sections are listed in Table 2. Sample labeling is based first by site location (Ibbenbüren, Beustfeld, Prosper Haniel, and Well 1). Second, to distinguish between the different sample sections from Well 1, the samples are labeled by maceral group (inertinite or vitrinite). In the parentheses, the lithotype in which the sample was selected is listed (D = dull lithotype, B = bright lithotype), followed by the image's relationship to bedding (II = parallel to bedding; ⊥ = perpendicular to bedding) and subsequently the magnification at which the image was examined (2kx, 5kx, etc.).

2.1. Sample background and selection 2.2. Sample preparation, image acquisition and processing Samples with varying degrees of coalification from within the Ruhr and Munsterland basins were first selected to test applicability of using BIB-milling in combination with SEM imaging to study pore morphology in coals. The original Ibbenbüren and Prosper Haniel coal samples were derived from mining activity, while the original Beustfeld sample was from a core. The samples, as used in this study, were selected from particles, and thus the original bedding relationship is ambiguous. A second set of coal samples was taken from Well 1 in the Munsterland basin, Germany, to evaluate the role of lithotype on pore size and shape and the role of pore size and shape relative to bedding. The term lithotype is purely a macroscopic description of a coal section 3 to 10 mm thick (Taylor et al., 1998); however, this term will be used throughout the paper to reference the source of the individual samples. The two lithotypes (bright coal, dull coal) are closely related in depth (Table 1), thus ruling out any potential influences due to a different maturity or compaction history. All of the samples represent Carboniferous coals of tropical origin (see Böcker et al. (2013), Jasper et al. (2009), and Littke and ten Haven (1989) Table 1 List of samples that were BIB-polished and viewed under the SEM. Vitrinite reflectance and maceral analysis are bulk parameters for the entire bed from which the samples were selected. VRr = random vitrinite reflectance. Sample label symbols: D = dull lithotype; B = bright lithotype; II = parallel to bedding; ⊥ = perpendicular to bedding. Site location

Ibbenbüren Beustfeld Prosper Haniel Well 1 Well 1

Sample label

Depth [m]

VR, [%]

Maceral analysis Vitrinite [%]

Inertinite [%]

Liptinite [%]

Ibbenbüren Beustfeld Prosper Haniel

1500 545 915

3.29 2.44 0.86

99 83 57

1 17 31

0 0 12

891 891

1.02 1.02

63 63

25 25

12 12

Inertinite Vitrinite

A detailed description of the maceral sample preparation for BIB-milling/SEM imaging is as follows: the coal sample was dried in an oven (air drying under b1 bar to prevent oxidation at 65 °C) for 12 h. Subsequently, the vitrinite and inertinite samples were cut out of the coal core with a mini rotary saw blade. The samples were not embedded in epoxy. These samples were then ground and polished with silicon carbide sandpaper, which also aids in preparing parallel surfaces for mounting the sample to a sample holder. After this preliminary preparation, the samples were stored in an exicator with silica gel to prevent hydration. The sample size is dictated by the capacity of the BIB machine (JEOL SM-09010), which polishes a cross-section of about 2 mm 2, using an argon beam for about eight hours of milling under a vacuum (10 −3–10 −4 Pa, 6 kV and 150 μA). The milling process removes about 100 μm of material and creates a cross-section with a topography of about ±5 nm (Klaver et al., 2012). The BIB-polished samples were gold sputtered to prevent electrostatic charging and subsequently imaged with a Zeiss Supra 55, utilizing a high-resolution secondary electron (SE) and back-scatter electron (BSE) detectors at RWTH Aachen University. Typical porous areas to study the pores within the BIB-polished cross sections were selected (Table 2). The final image of these areas is a mosaic, created by taking hundreds of pictures at magnifications of either 2000, 5000, 10,000, 20,000, or 30,000× using 10–20% overlap. The BSE detector was additionally used to create several mosaics to gain information about the relative material density and mineralogy. An energy dispersive X-ray spectroscope (EDX) was attached to the SEM for elemental analysis. When using FIB–SEM or BIB–SEM, artifacts like local composition or phase changes as well as redeposition of the sputtered material can occur. No action was taken to prevent such artifacts, but for example, redeposition of the sputter material was also not observed in the studied samples. Bassim et al. (2012) studied the effect of FIB milling on lignite coal and found that only minor chemical changes

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

3

Fig. 1. (A) Schematic of sample selection and orientation. Dashed lines indicate bedding planes (not to scale). Dotted lines signify BIB polished sections. Heavy black boxes are parallel to bedding; double-dashed boxes are perpendicular to bedding. Solid grey areas of the cubes represent non-imaged sections. (B) Visualization of BIB-milled surface (SEM image; (a)) on sample vitrinite (B/II) and vitrinite (B/⊥/5) with its comparable micrograph (oil immersion, reflected light (b)). The pores in sample vitrinite (B/⊥/5) and vitrinite (B/⊥/20) are most likely not located in the telocollinte but instead in an underlying, different maceral type. (The micrograph is analogous to the SEM image. Dotted lines contain the BIB-milled surface. Solid grey areas of the cubes represent non-imaged sections.)

Table 2 Statistics of the mosaics with detectible pores, i.e. those mosaics used in the pore analysis. The samples vitrinite (D/II) and vitrinite (D/⊥) as well as vitrinite (B/II) are not listed in the table, since they contained practically no detectible pores. In addition to the sample label symbols (Table 1), a number after the bedding orientation indicates the level of magnification times 1000. A letter after the magnification (i.e. 20a or 20b) indicates two different mosaics taken at the same magnification. Sample label

Magnification

Total no. of pores

No. of pores > p.p.r.a

No. of images in mosaic

Mosaic area [nm2]

Pixel size [nm]

Practical pore resolution (eq. diameter)b [nm]

Ibbenbüren (−/−/10) Beustfeld (−/−/30) Prosper Haniel (−/−/20) Inertinite (D/II/20) Inertinite (D/⊥/20a) Inertinite (D/⊥/20b) Inertinite (B/II/2) Inertinite (B/II/10) Inertinite (B/⊥/5) Inertinite (B/⊥/20) Vitrinite (D/II/2)c Vitrinite (D/⊥/2)c Vitrinite (B/II/2)c Vitrinite (B/⊥/5) Vitrinite (B/⊥/20)

10 kx 30 kx 20 kx 20 kx 20 kx 20 kx 2 kx 10 kx 5 kx 2 0kx 2kx 2 kx 2 kx 5 kx 20 kx

182 3898 73 5121 2431 1851 1403 281 1213 56 b35 b20 b20 5183 333

109 2383 23 3427 2123 1449 940 242 1139 56 – – – 1504 245

132 168 104 224 120 156 197 180 169 88 136 77 126 196 56

5.12 · 1010 7.5 · 109 6.51 · l09 2.53 · 1010 1.24 · 1010 1.38 · 1010 1.20 · 1012 8.70 · 1010 3.18 · 1011 1.15 · 1010 1.87 · 1014 l.33 · 1012 1.44 · 1012 5.06 · 1011 9.22 · 109

29.3 9.8 14.7 14.7 14.7 14.7 146.7 29.3 58.6 14.7 146.7 146.7 146.7 58.6 14.7

113.5 38.0 56.9 56.9 56.9 56.9 568.2 113.5 227.0 56.9 – – – 227.0 56.9

– Not determined. a p.p.r.: practical pore resolution. b Pore resolution assumes a minimum of 15 pixels. c Due to low number of visible pores, no high resolution mosaics were created to determine pore size distributions.

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

4

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

were induced. Yao et al. (2011) used atomic force microscopy to study the nanostructure of coals. They used anhydrous ethanol to remove impurities from the surface before observation; this step may have been of interest for our study, since a few dust particles were observed on our samples' BIB-milled surface. Loucks et al. (2009) used argon-ion beam milling to eliminate artifacts, such as pits or holes that could be mistaken for pores, created by conventional polishing techniques. During milling, surface damage caused by incident ions is reduced by a shield covering most of the sample (Drobne et al., 2007). The individual images for the mosaics of the large areas were stitched together using a bicubic interpolation in Autopano 2. The pores were segmented manually in ArcGIS 9.3. Pores are recognizable down to a few pixels in size. A practical pore resolution of 15 pixels is taken as the minimum pore resolution, which is the minimum size in which all pores are detected. 2.3. Data acquisition The data format of the segmented pores is an ArcGIS shape file. From this shape file, the geometrical properties of each pore can be obtained, including area, length, width and orientation. Area is calculated from the shape file itself, while length, width and orientation were derived using a minimum bounding geometry. In the results section of this paper, frequency diagrams of pore area are used to compare the individual samples. The pore size distribution is given as a normalized discrete function (Klaver et al., 2012): Ni −D ¼ CSpore bi Smosaic

ð1Þ

By taking the log of each side, we get:  log

Ni bi Smosaic



  ¼ −D log Spore þ logðC Þ

ð2Þ

where Ni is the number of pores with a characteristic pore area Spore within the bin width bi. Bin width bi doubles with each subsequent bin (1, 2, 4, 8, 16, etc.). Smosaic is the surface area of the respective mosaic. C is a constant and D is the power-law exponent. The pore size distribution is plotted on a log–log graph and the range between the largest pore and the smallest (practical pore size) defines the limits of the power-law distribution, as presented in Klaver et al. (2012). In addition, the frequency of pore areas of samples from Well 1 is compared with their bulk mesopore distribution. The mesopore analysis was determined by low pressure nitrogen adsorption after the method discussed in Prinz et al. (2004) and evaluated using the quenched solid density functional theory (QSDFT) presented in Neimark et al. (2009). 3. Results 3.1. Qualitative description of pore morphology 3.1.1. Ibbenbüren Much of the BIB-milled surface of the Ibbenbüren sample contained little visible porosity. The apparent pores are rounded without any sharp corners. The contact between the pore wall and the BIB-milled surface is smooth. The pores have a gentle curvature to their shape. A BSE detector combined with an EDX detector shows that an inorganic component – kaolinite - is present in the coal matrix. Pores appear to be concentrated within the organic material adjacent to the kaolinite. No pores were observed within the kaolinite. Fig. 2A shows the relationship between the coal organic porosity and adjacent kaolinite.

3.1.2. Beustfeld Microcracks are prominent in the BIB-milled surface area of the Beustfeld sample. The presence of minerals either completely or partially filling the cracks indicates that they originated in-situ, and thus not due to damage or drying effects. Framboidal pyrite occurs in layers along these cracks, which may be an indication of bedding. Towards the base of the milled surface area, a layer is perforated with many pores. These pores are rather angular with visible pore walls dipping shallowly from the surface into the sample. Opposite pore walls sometimes constrict, either closing off a singular large pore into two smaller ones or leaving a small slit to connect the larger pore cavities. These pores are located primarily in the coal matrix or as a gap between the framboidal pyrite and organic matrix (Fig. 2B). The spatial relationship of the pores to one another is shown in Fig. 3(1) and Fig. 3(A). 3.1.3. Prosper Haniel The pores in the Prosper Haniel sample can be divided into two types: 1) pores in the matrix, and 2) pores associated with the framboidal pyrite grains located along partially open microcracks. The majority of the BIB-milled surface area of the Prosper Haniel matrix showed no visible pores; however, there was one cluster of small pores. The shape of the pores surrounding the pyrite grains outlines the cubic nature of pyrite. The Prosper Haniel sample also contained frambodial pyrites (Fig. 2D). In contrast to sample Beustfeld, the pores in the Prosper Haniel sample are found around each pyrite mineral in the frambodial pyrite group. A gap has formed between the organic matrix and the individual framboidal pyrites, resulting in a rather porous structure on the whole. This gap may be the result of matrix shrinkage caused by desorbed methane or by drying (Day et al., 2008). However, the opposite can also occur: pyrite can fill the pore space, thus effectively reducing total porosity and creating more torturous pathways for gas flow (Fig. 2C). 3.1.4. Well 1 The investigated surface area of the inertinite sample series from Well 1 revealed a variety of pore shapes and sizes. The pores in the inertinite in the dull lithotype, both parallel and perpendicular to bedding, are characterized as angular, small, – compared to the sample in the inertinite in the bright lithotype – and multitudinous (Fig. 3(2)). The BIB-milled surface area perpendicular to bedding appeared to be cross-cut by microcracks, which potentially mark bedding planes. Pores in the inertinite sample in the bright lithotype, in contrast, were generally elliptical. These pores also appear to be slightly oriented, stretching from the lower left-hand corner to the upper right-hand corner. Even at the magnification of the whole surface area (2000×), the outline of the pore walls was well defined. Often these pores were partially filled; albeit the original pore outline was still visible. However, the mineral(s) filling the pores did not seal them completely. Fusinite from the inertinite (B/II) sample is shown in Fig. 4. Fig. 4A and B indicates that a significant proportion of the pores within the fusinite are filled with inorganic material. This agrees with observations using the reflected light microscope (Fig. 4C). Higher magnification reveals that the entire pore is not completely filled (Fig. 4D). Fig. 4E shows an exemplary pore from the sample inertinite (B/II) along with the corresponding EDX images and interpretation of the mineralogy (Fig. 4F). Based on the EDX measurements, the data suggests that the inorganic substances are kaolinite, dolomite, and ankerite. The mineralogy found in the pores of this study's samples is similar to that in the Ruhr Basin coals studies by Dawson et al. (2012). The vitrinite sample series showed very little visible porosity. Only a few very sparsely scattered, small round pores were found. These pores were so small in number that the samples are characterized as having no meso- and macroporosity. The one exception to this in

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

5

Fig. 2. Images of samples Ibbenbüren, Beustfeld, and Prosper Haniel. The relationship between pores in the organic matrix and their proximity to inorganic components is shown. (A) BSE image showing pores (red) located between clays (light grey areas) in sample Ibbenbüren (−/−/10). (B) BSE image of sample Beustfeld with segmented pores. Majority of the pores are in the organic material between the frambodial pyrites and not in between the pyrites (compare to (D)). (C) Euhedral pyrite crystal in pore in sample Prosper Haniel. (D) is a BSE image of frambodial pyrite grains, sample Prosper Haniel.

the vitrinite series was the sample viewed perpendicular to bedding from the bright lithotype (Fig. 1B). In this sampled surface area, numerous pores were visible. However, these pores probably do not belong to a vitrinite maceral. The lower section of the image, which includes the porous parts, is within an inertinte maceral. Only the top part of the BIB-milled section belongs to a vitrinite, even though the whole sample section is labled as a vitrinite.

3.2. Quantitative description of pore shape 3.2.1. Pore length, width, and aspect ratio Pore dimensions of length and width for each sample are compared in Fig. 5. In the figure real pore images are shown next an ellipse of mean pore shape. The ellipse of mean pore shape is a measure of the average length and width of the minimum bounding box for all segmented pores. The average pore length, width, and aspect ratio are based on all the pores greater than the practical pore resolution (equivalent to 15 pixels in size). With respect to the length and width, all pores compared to one another are small (lengths and widths less than one micrometer) and slightly elliptical. The exception is the pores found in the fusinite sample. These pores have lengths greater than or equal to approximately one micrometer. Aspect ratio is not so strongly dependent on magnification as the individual dimensions and therefore, makes a good parameter to compare results from the same sample that were determined at different magnifications. Aspect ratio is an indicator for the pore roundness. It may also be a way to eliminate the effect of magnification on pore dimensions. Ibbenbüren, Beustfeld, and Prosper Haniel have aspect ratios of 2.6,

2.7, and 1.6, respectively. The aspect ratio for the inertinite in the dull lithotype does not change much, regardless of the orientation to bedding. The inertinite in the bright lithotype has a high aspect ratio parallel to bedding, and a comparatively lower aspect ratio perpendicular to bedding. The aspect ratio for the vitrinite in the bright lithotype is similar to that of inertinite in the dull lithotype.

3.2.2. Pore area The distribution of pore area for each mosaic is shown on a log–log diagram of normalized frequency versus pore area (Fig. 6). The valid range for each linear regression is defined by the filled-in markers, the variables D and log C, and the dashed lines crossing the x-axis guide the eye to end-member equivalent pore diameters. The open markers represent segmented pores with a size of the practical pore resolution. The parameters D and log C range for the samples Ibbenbüren, Beustfeld, and Prosper Haniel from 1.47 to 1.81 and −3.90 to −6.98, respectively. The parameters D and log C range for the vitrinite and inertinite sample sections from 1.13 to 1.79 and −4.31 to −8.63, respectively. For most of the samples, the frequency of pores less than approximately 15 pixels in size decreases with decreasing pore size. The exceptions are the samples inertinite (D/⊥/20a), inertinite (D/⊥/20b) and inertinite (D/II/20). These samples showed the greatest frequency of pores at a size of 90 nm in diameter. This peak indicates the lower boundary of the self-similar behavior of the pore size distribution. For each sample analyzed with different magnifications, the data align over the middle range of pore areas, indicating the overlap in magnification (Krohn and Thompson, 1986). The resulting linear regression of the pore size distribution for each magnification or section per sample also aligns for the most part.

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

6

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

Fig. 3. Comparison of spatial relationship and morphology of pores from the samples Beustfeld, inertinite and vitrinite.

Fig. 7 shows the percent distribution of pore volume over pore diameter for the nitrogen low-pressure isotherm data (solid line). The data is derived from the bulk coal sample from Well 1. As a comparison, the percent of pore area for the sample inertinite (B/II/5) is also plotted over pore diameter. The graph shows that there is a minimal overlap in pore diameter between the two methods. Based on the nitrogen adsorption results, the majority of the pores are about 2 nm in diameter. However, these are the throats of the accessible pores (Prinz and Littke, 2005), leaving the larger pore cavities, which store and transport gas, uncharacterized. And in turn, BIB–SEM determined pore sizes do

not capture pore throat diameters. This could explain the lack in overlap between the two methods shown in Fig. 7. 4. Discussion 4.1. Porosity and pore morphology 4.1.1. Maceral type: inertinite versus vitrinite In all BIB-milled sections but one, essentially no visible pores were detected in the vitrinite samples, which agrees with the findings of

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

7

Fig. 4. Fusinite maceral as seen with the BSE (A) and SEM (B) and reflected light microscope under oil immersion (C). (D) A pore (outlined with dashed line) and the remaining porosity after mineral fill or even secondary porosity after partial dissolution of mineral fill. Color coding in image A: dark grey = organic matrix, light grey = inorganic components; image C: white = organic fragments; light grey = pore-filling mineral(s). (E) BSE image of pore-filling minerals in sample inertinite (B/II) with an interpretation (F). Color coding in image (E): dark grey = organic matrix; light grey = inorganic components; black = open porosity. Color coding in (F): dotted = dolomite, stripped = kaolinite, checked = ankerite.

Bustin and Clarkson (1998), Clarkson and Bustin (1997), and Mastalerz et al. (2008a). The low-pressure nitrogen sorption experiment shows that the majority of coal porosity is located in the micropore range (Fig. 7), which is below the pixel resolution. The few pores that were found in the vitrinites were associated with either a fracture network or interspersed between mineral phases. Less than 2% of all the pores studied were found in vitrinite. The amount of porosity in the inertinites varied based on maceral type. The samples Ibbenbüren, Beustfeld, and Prosper Haniel are not initially referenced by a maceral group. However, based on the findings of the image analysis and a subsequent polishing of the sections to

examine them under the microscope, conclusions can be drawn about the investigated macerals. The matrix sections of the Ibbenbüren and Prosper Haniel samples contained very few pores, which would fit with the vitrinite group. Additionally the polished sections for the optical microscope also showed vitrinites. The Beustfeld sample is most likely of an internite maceral; it contained a multitude of pores, with a similar pore morphology and size distribution to the studied macrinite sample (inertinite (D/II and ⊥)). It is known that the distribution of micropore volume with rank for dry coals decreases down to 1.2–1.4 VRr and then increases (Prinz and Littke, 2005). The majority of the micropore volume is located in

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

8

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

Fig. 5. Comparison of pore dimensions for each analyzed sample. The ellipses of mean pore shape, based on the minimum bounding box for all pores, are scaled to one another and are based on the given mean lengths and widths for all pores greater than the practical pore resolution. Aspect ratio is given by length divided by width and is based on all data points greater than the practical pore resolution.

vitrinites (Harris and Yust, 1976; Unsworth et al., 1989), and a similar trend for the amount of meso- and macropores in vitrinites was found in this study. Mastalerz et al. (2008b) found that collotelinite positively influenced the meso- and micropore volume. Maceral composition is important to determine the adsorbed methane and thus the potential of the coalbed reservoir (Lamberson and Bustin, 1993). Methane adsorption can be used as an indicator for gas storage capacity. Due to the low amount of internal surface area of inertinites, they have the lowest methane adsorption capacity (Lamberson and Bustin, 1993). However, the inertinites contained the largest pores (Fig. 5), which may provide the best transport pathways, considering that smaller pores may be inaccessible to fluids like water (Han et al., 2010; Prinz and Littke, 2005). A combination between a good storage capacity, e.g. microporous vitrinites, and large, open transport pathways, e.g. cleats and connected meso- and macroporous inertinites, represents the most promising reservoirs for coalbed methane. Given that the high volitale bituminous coal samples of Well 1 and Prosper Haniel as well as the semianthracite Beustfeld sample contain about 55–85% vitrinite and 15–30% inertinite (Table 1), these samples have the best potential for gas storage and transport. 4.1.2. Lithotype: dull versus bright Bright and dull lithotypes stem from a difference in depositional environment and/or accumulation rate or degradation of organic material, where a change from bright to dull lithotype may represent a transition from woody to herbaceous milieus (Lamberson and Bustin, 1993). The change in depositional input, demonstrated by lithotype, has an effect on the total maceral composition, which in turn, affects the dominating type of porosity. Based on Fig. 6 it would appear that the greatest difference in pore area distributions occurs between the lithotypes (i.e. individual macerals) and not between pores with respect to the bedding orientation. For example, the pores in sample vitrinite (B/⊥) (only labeled as a vitrinite, even though this section probably belongs to an inertinite

maceral) that were located in a bright lithotype matched the pore size dimensions of those in sample inertinite (D/⊥), which were located in a dull lithotype. In a future study, more samples of the same maceral type in different lithotypes should be investigated. Porosity as a function of lithotype for whole coals has been described in the literature. Many of these studies have used sorption techniques to determine porosity (Clarkson and Bustin, 1997; Crosdale et al., 1998). They found that dull coal lithotypes desorb more rapidly due to a higher inertinite content, and thus macroporosity than their bright lithotype counterparts. However, the bright lithotypes (vitrinite-rich coals) contain more micropores and thus have a greater gas storage capacity due to larger surface areas. The visual images of coal pore structures presented in this paper support the previously mentioned authors' conclusions: i.e. there is a higher meso- and macropore volume in dull lithotypes and inertinites. 4.1.3. Orientation to bedding: perpendicular versus parallel Pore shape reflects the paleostrain at the time of coalification, where pore sizes perpendicular to bedding are expected to be smaller than those in the bedding plane (Radlinski et al., 2004). A trend between the pore dimensions – given by length, width, and axial ratio (Fig. 5) – and the relationship to bedding was not discernible in this study. Nevertheless, pore roundness described by the axial ratio may lead to a better understanding of pore shape models. For example, the shape of a low-pressure isotherm provides clues as to the shape of the pores (Sing et al., 1985). 4.2. Pore size distribution in coal A practical pore resolution of 15 pixels was chosen to model the pore size distribution in coal. However, this boundary should be the subject of modification in future works. Pores less than 15 pixels in size were deemed beyond the resolution of the system, and thus, not all of the visible pores below this size can be potentially captured. Therefore, the pore size distribution begins to decrease in frequency towards smaller

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

9

Fig. 6. Pore size distribution for all studied samples.

pore sizes due to resolution limits; an effect known as truncation (Ortega et al., 2006). Another possible explanation for the change towards smaller pore sizes may be that there are no SEM visible pores below a certain pore diameter. For example, the inertinite samples examined perpendicular to bedding at 20,000× magnification, i.e. the samples inertinite (B/⊥/20), inertinite (D/⊥/20a), and inertinite (D/⊥/ 20b), did not contain pores smaller than approximately 60 nm in diameter. The smallest pore in the sample inertinite (B/⊥/20) is about

100 nm (Fig. 6). The smallest detected pores in these samples are still above the practical pore resolution, which was used as a visibility cut-off. The power law exponent, i.e. parameter D, gives the slope of the regression line for the log-log diagrams of pore frequency versus pore size. The relatively low power law exponent (b2) indicates that the small pores within the visible range of the SEM – from 10 nm to >1 μm – contribute less to the total porosity than the larger pores.

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

10

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

Fig. 7. Pore size distribution for bulk coal sample from Well 1 (solid line) using nitrogen adsorption at 77 K and QSDFT (cf. Neimark et al., 2009) and for the sample inertinite (B/II/2) using image quantification techniques (dashed line).

Nevertheless, the pore size distribution, derived from low-pressure adsorption data (Fig. 7), shows that most of the pore volume is due to micropores. In order for the micropores to contribute the most to the total porosity, the slope of the regression line in this range should be greater than the slope of the regression line for the meso-/macropores (determined from the SEM images). Therefore, we interpret that pores in the meso-/macropore range have a different pore size distribution – and ensuing power law exponent and coefficient – than pores in the micropore range. The power law exponent can be further used to assess relationships between samples. By comparing the D values for all sections of the inertinite samples, the biggest difference is between the inertinite in the bright lithotype, i.e. fusinite; D ≈ 1.25, and the inertinite in the dull lithotype, i.e. macrinite (D ≈ 1.61). The sample labeled vitrinite (B/⊥) has a an average D of 1.57, which is very similar to the macrinite, which is further evidence that only the top part of the sample is vitrinite while the pore-rich areas are located in inertinite. The pore area distributions for the Posidonia shale (Klaver et al., 2012) using the exact same method as in this study are different from those for the coals. The power law exponent for the rock is approximately 2, while the power law regression for the nanofossil porosity showed a dual power-law distribution, with an exponent varying between 1 and 3. The power law exponents for the inertinite samples averaged to about 1.5. There is the potential that the inertinites are also governed by a dual power-law; however, the change in the slope of the regression line occurs at pore diameters smaller than the resolution capabilities of the SEM. A dual power law – indicating a bi-modal system – for the studied inertinites agrees the bulk coal data sets in the literature. Gan et al. (1972), who used mercury porosimetry and nitrogen adsorption to determine their results, found a bi-modal system. Clarkson and Bustin (1999) determined a multi-modal pore system for four coals from the Cretaceous Gates Formation.

allows for the investigation of pores within an individual maceral. However, the areas examined may not be representative of the whole, and thus cannot be easily interpreted from a coal seam standpoint. Nevertheless, these semi-quantitative images do provide insight into the macropore diversity of individual macerals and thus the heterogeneity of coals. This study is the first to use BIB-milling to examine a smooth coal surface, which allows for a definitive characterization of pore morphology and their location within individual macerals. The studied vitrinite samples showed mostly no meso- and macropores, while the intertinite samples contained the meso- and macropores. The application of BIB–SEM imaging to coals as well as the nature of meso-/ macropore morphology is summarized below.

5. Conclusions

This work is funded by ExxonMobil Production Germany GmbH within the framework of the project “RWTH-EMPG Investigation of Coal Seam Gas: Gas storage and transport processes in coal seams”. The constructive comments of two anonymous reviewers greatly improved the original manuscript. The authors would like to thank J. Yeakel and Y. Gensterblum for the fruitful discussions. U. Wollenberg

The state-of-the-art method combining broad-ion beam milling with a high resolution SEM provides clear images of the meso- and macroporosity in coals, especially for pore diameters between 10 nm and 1 μm. The small area of the milled surface (ca. 2 mm by 2 mm)

• BIB–SEM imaging provides clear images of pores in individual coal macerals at high resolutions, i.e. lower micrometer to upper nanometer. • Pores were often found in close proximity to mineral matter or the mineral matter was pore-filling, like in the fusinite voids. The inorganic components that occurred most often in the samples studied were kaolinite and framboidal pyrite. • The most elongated pores, which also happened to be the largest pores, were found in the fusinite. • The sample sections that were imaged at both high and low magnifications, e.g. 20,000× and 10,000×, yield pore area distributions that are in good agreement; i.e. the pore area distributions from the highand low-magnification mosaics have overlapping sections on the log-log plots. The pore area distributions shown in the log–log plots follow a power law distribution in the BIB–SEM range. • More samples of the same maceral type from different settings should be investigated at several different resolutions in the future. • Pore systems in coals differ from those in organic-rich shales by their power law regression parameters. Acknowledgments

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009

S. Giffin et al. / International Journal of Coal Geology xxx (2013) xxx–xxx

is acknowledged for his support with the BIB-milling. We would also like to acknowledge S. Fellmin, M. Sartorius, K.A. Wilkins for their help with pore segmentation.

References Bassim, N.D., deGregorio, B.T., Kylcoyne, A.L.D., Scott, K., Chou, T., Wirick, S., Cody, G.D., Stroud, R.M., 2012. Minimizing damage during FIB sample preparation of soft materials. Journal of Microscopy 245, 288–301. Böcker, J., Littke, R., Hartkopf-Fröder, C., Jasper, K., Schwarzbauer, J., 2013. Organic geochemistry of Duckmantian (Pennsylvanian) coals from the Ruhr Basin, western Germany. International Journal of Coal Geology 107, 112–126. Bustin, R.M., Clarkson, C.R., 1998. Geological controls on coalbed methane reservoir capacity and gas content. International Journal of Coal Geology 38, 3–26. Chalmers, G.R., Bustin, R.M., Power, I.M., 2012. Characterization of gas shale pore systems by porosimetry, pycnometry, surface area, and field emission scanning electron microscopy/transmission electron microscopy mage analyses: examples from the Barnett, Woodford, Haynesville, Marcellus, and Doig units. AAPG Bulletin 96, 1099–1119. Clarkson, C.R., Bustin, R.M., 1996. Variation in micropore capacity and size distribution with composition in bituminous coal of the Western Canadian Sedimentary Basin: implications for coalbed methane potential. Fuel 75, 1483–1498. Clarkson, C.R., Bustin, R.M., 1997. Variation in permeability with lithotype and maceral composition of Cretaceous coals of the Canadian Cordillera. International Journal of Coal Geology 33, 135–151. Clarkson, C.R., Bustin, R.M., 1999. The effect of pore structure and gas pressure upon the transport properties of coal: a laboratory and modelling study. 1. Isotherms and pore volume distributions. Fuel 78, 1333–1344. Crosdale, P.J., Beamish, B.B., Valix, M., 1998. Coalbed methane sorption related to coal composition. International Journal of Coal Geology 35, 147–158. Curtis, M.E., Ambrose, R.J., Sondergeld, C.H., Rai, C.S., 2010. Structural characterization of gas shales on the micro- and nano-scales. Canadian Unconventional Resources and International Petroleum Conference. Society of Petroleum Engineers, Calgary, Alberta, Canada, p. 15 (SPE 137693). Curtis, M.E., Sondergeld, C.H., Ambrose, R.J., Rai, C.S., 2012. Microstructural investigation of gas shales in two and three dimensions using nanometer scale resolution imaging. AAPG Bulletin 96, 665–677. Dawson, G.K.W., Golding, S.D., Esterele, J.S., Massarotto, P., 2012. Occurrence of minerals within fractures and matrix of selected Bowen and Ruhr Basin coals. International Journal of Coal Geology 94, 150–166. Day, S., Sakurovs, R., Weir, S., 2008. Supercritical gas sorption on moist coals. International Journal of Coal Geology 74, 203–214. Desbois, G., Urai, J.L., Kukla, P.A., 2009. Morphology of the pore space in claystones — evidence from BIB/FIB ion beam sectioning and cryo-SEM observations. eEarth 4, 15–22. Desbois, G., Urai, J.L., Kukla, P.A., Konstanty, J., Baerle, C., 2011. High-resolution 3D fabric and porosity model in a tight gas sandstone reservoir: a new approach to investigate microstructures from mm- to nm-scale combining argon beam crosssectioning and SEM imaging. Journal of Petroleum Science and Engineering 78, 243–257. Drobne, D., Milani, M., Leser, V., Tatti, F., 2007. Surface damage induced by FIB milling and imaging of biological samples is controllable. Microscopy Research and Technique 70, 895–903. Gamson, P.D., Beamish, B.B., Johnson, D.P., 1993. Coal microstructure and micropermeability and their effects on natural gas recovery. Fuel 72, 87–99. Gan, H., Nandi, S.P., Walker Jr., P.L., 1972. Nature of the porosity in American coals. Fuel 51, 272–277. Gensterblum, Y., van Hemert, P., Billemont, P., Busch, A., Charriére, D., Li, D., Krooss, B.M., de Weireld, G., Prinz, D., Wolf, K.-H.A.A., 2009. European inter-laboratory comparison of high pressure CO2 sorption isotherms. I: activated carbon. Carbon 47, 2958–2969. Han, F., Busch, A., van Wageningen, N., Yang, J., Liu, Z., Krooss, B.M., 2010. Experimental study of gas and water transport processes in the inter-cleat (matrix) system of coal: anthracite from Qinshui Basin, China. International Journal of Coal Geology 81, 128–138. Harris, L.A., Yust, C.S., 1976. Transmission electron microscopy observations of porosity in coal. Fuel 55, 233–236. Houben, M.E., Desbois, G., Urai, J.L., 2013. Pore morphology and distribution in the shaly facies of Opalinus Clay (Mont Terri, Switzerland): insights from representa-

11

tive 2D BIB–SEM investigations on mm to nm scale. Applied Clay Science 71, 82–97. Jasper, K., Krooss, B.M., Flajs, G., Hartkopf-Fröder, Chr, Littke, R., 2009. Characteristics of type III kerogen in coal-bearing strata from the Pennsylvanian (Upper Carboniferous) in the Ruhr Basin, Western Germany: comparison of coals, dispersed organic matter, kerogen concentrates and coal mixtures. International Journal of Coal Geology 80, 1–19. Karacan, C.Ö., Okandan, E., 2000. Fracture/cleat analysis of coals from Zonguldak Basin (northwestern Turkey) relative to the potential of coalbed methane production. International Journal of Coal Geology 44, 109–125. Karg, H., Carter, A., Brix, M., Littke, R., 2005. Late- and post-Variscan cooling and exhumation history the northern Rhenish massif and the southern Ruhr Basin: new constraints from fission track analysis. International Journal of Earth Sciences 94, 180–192. Klaver, J., Desbois, G., Urai, J.L., Littke, R., 2012. BIB–SEM study of the pore space morphology in early mature Posidonia Shale from the Hils area, Germany. International Journal of Coal Geology 103, 12–25. Krohn, C.E., Thompson, A.H., 1986. Fractal sandstone pores: automated measurements using scanning-electron-microscope images. Physical Review B 33, 6366–6374. Lamberson, M., Bustin, R.M., 1993. Coalbed methane characteristics of Gates Formation coals, northern British Columbia: effect of maceral composition. AAPG Bulletin 77, 2062–2076. Littke, R., ten Haven, H.L., 1989. Palaeoecologic trends and petroleum potential of Upper Carboniferous coal seams of Western Germany as revealed by their petrographic and organic geochemical characteristics. International Journal of Coal Geology 13, 529–574. Loucks, R.G., Reed, R.M., Ruppel, S.C., Jarvie, D.M., 2009. Morphology, genesis, and distribution of nanometer-scale pores in siliceous mudstones of the Mississippian Barnett shale. Journal of Sedimentary Research 79, 848–861. Massarotto, P., Rudolph, V., Golding, S., 2003. Anisotropic permeability characterisation of Permian coals. Proceedings of the International CBM Symposium. University of Alabama, Tuscaloosa, USA (CD ROM Paper 0359, 11 pp.). Mastalerz, M., Drobniak, A., Rupp, J., 2008a. Meso- and micropore characteristics of coal lithotypes: implications for CO2 adsorption. Energy & Fuels 22, 4049–4061. Mastalerz, M., Drobniak, A., Strapoc, D., Solano-Acosta, W., Rupp, J., 2008b. Variations in pore characteristics in high volatile bituminous coals: implications for coal bed gas content. International Journal of Coal Geology 76, 205–216. Neimark, A.V., Lin, Y., Ravikovitch, P.I., Thommes, M., 2009. Quenched solid density functional theory and pore size analysis of micro-mesoporous carbons. Carbon 47, 1617–1628. Ortega, O.J., Marrett, R.A., Laubach, S.E., 2006. A scale-independent approach to fracture intensity and average spacing measurement. AAPG Bulletin 90, 193–208. Prinz, D., Littke, R., 2005. Development of the micro- and ultramicroporous structure of coals with rank as deduced from accessibility to water. Fuel 84, 1645–1652. Prinz, D., Pyckhout-Hintzen, W., Littke, R., 2004. Development of the meso- and macroporous structure of coals with rank as analysed with small angle neutron scattering and adsorption experiments. Fuel 83, 547–556. Radlinski, A.P., Mastalerz, M., Hinde, A.L., Hainbuchner, M., Rauch, H., Baron, M., Lin, J.S., Fan, L., Thiyagarajan, P., 2004. Application of SAXS and SANS in evaluation of porosity, pore size distribution and surface area of coal. International Journal of Coal Geology 59, 245–271. Rouquerol, J., Avnir, D., Fairbridge, C.W., Everett, D.H., Haynes, J.H., Pernicone, N., Ramsay, J.D.F., Sing, K.S.W., Unger, K.K., 1994. Recommendations for the characterization of porous solids. Pure and Applied Chemistry 68, 1739–1758. Sing, K.S.W., Everett, D.H., Haul, R.A.W., Moscou, L., Pierotti, R.A., Rouquerol, J., Siemieniewska, T., 1985. Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity. Pure and Applied Chemistry 57, 603–619. Taylor, G.H., Teichmüller, M., Davis, A., Diessel, C.F.K., Littke, R., Robert, P., 1998. Organic Petrology. Gebrüder Borntraeger, Berlin-Stuttgart. Unsworth, J.F., Fowler, C.S., Jones, L.F., 1989. Moisture in coal, 2. Maceral effects on pore structure. Fuel 68, 18–26. Yao, Y., Liu, D., 2009. Microscopic characteristics of microfactures in coals: an investigation into permeability of coal. Procedia Earth and Planetary Science 1, 903–910. Yao, Y., Liu, D., Che, Y., Tang, D., Tang, S., Huang, W., 2010a. Petrophysical characterization of coals by low-field nuclear magnetic resonance (NMR). Fuel 89, 1371–1380. Yao, Y.B., Lui, D.M., Cai, Y.D., JunQian, L.I., 2010b. Advanced characterization of pores and fractures in coals by nuclear magnetic resonance and X-ray computed tomography. Science China Earth Sciences 53, 854–862. Yao, S., Jiao, K., Zhang, K., Hu, W.X., Ding, H., Li, M.C., Pei, W.M., 2011. An atomic force microscopy study of coal nanopore structure. Chinese Science Bulletin 56, 2706–2712.

Please cite this article as: Giffin, S., et al., Application of BIB–SEM technology to characterize macropore morphology in coal, International Journal of Coal Geology (2013), http://dx.doi.org/10.1016/j.coal.2013.02.009