Downloaded from geosphere.gsapubs.

org on July 3, 2012

Geosphere

3D characterization of sandstone by means of X-ray computed tomography

V. Cnudde, M. Boone, J. Dewanckele, M. Dierick, L. Van Hoorebeke and P. Jacobs

Geosphere 2011;7;54-61
doi: 10.1130/GES00563.1

Email alerting services click www.gsapubs.org/cgi/alerts to receive free e-mail alerts when new
articles cite this article
Subscribe click www.gsapubs.org/subscriptions/ to subscribe to Geosphere
Permission request click http://www.geosociety.org/pubs/copyrt.htm#gsa to contact GSA
Copyright not claimed on content prepared wholly by U.S. government employees within scope of
their employment. Individual scientists are hereby granted permission, without fees or further
requests to GSA, to use a single figure, a single table, and/or a brief paragraph of text in subsequent
works and to make unlimited copies of items in GSA's journals for noncommercial use in classrooms
to further education and science. This file may not be posted to any Web site, but authors may post
the abstracts only of their articles on their own or their organization's Web site providing the posting
includes a reference to the article's full citation. GSA provides this and other forums for the
presentation of diverse opinions and positions by scientists worldwide, regardless of their race,
citizenship, gender, religion, or political viewpoint. Opinions presented in this publication do not reflect
official positions of the Society.

Notes

© 2011 Geological Society of America

 Downloaded from geosphere.gsapubs.org on July 3, 2012

3D characterization of sandstone by means of X-ray computed tomography

V. Cnudde1, M. Boone2, J. Dewanckele1, M. Dierick2, L. Van Hoorebeke2, and P. Jacobs1

1
Department of Geology and Soil Science–UGCT, Ghent University, Krijgslaan 281–S8, 9000 Gent, Belgium
2
Department of Physics and Astronomy–UGCT, Ghent University, Proeftuinstraat 86, 9000 Gent, Belgium

ABSTRACT (Russ, 2002). In order to overcome this problem, Although statistical and process-based models
several orthogonal sections or serial sectioning have their advantages, X-ray microtomography
Due to technological development, state- coupled with digital image analysis to construct (micro-CT) can also be applied in the study of
of-the-art high-resolution X-ray computed a three-dimensional image of the microstructure internal structures of geomaterials. Micro-CT
tomography (CT) systems can be developed, can be necessary to reveal 3D mineral structures has improved enormously in recent years thanks
enabling the internal visualization of geo- (Lin and Cohen, 1982; Holt et al., 1996). to rapid technological developments and is more
materials in three dimensions. However, in Many research topics in geology concern widely available now, with a constantly improv-
order to obtain structural information one the study of internal structures of geomaterials ing resolution and image quality. This technique
also needs proper three-dimensional (3D) on a pore-scale level in order to estimate enables rapid, nondestructive, 2D and 3D exam-
analysis software. In this paper, the potential their macroscopic behavior (Bakke and Øren, ination and analysis of almost any kind of mate-
for petrographic purposes of high-resolution 1997). In order to compute properties such as rial, including rock and soil (Carlson, 2006).
X-ray CT in combination with the 3D analy- permeability and electrical conductivity, there Tomography-based methods rebuild the inter-
sis software Morpho+ is explored for a Bel- is a need for an adequate description of the nal structural information within an object by
gian sandstone sample. The advantage of the complex internal microstructure. Bakke and mathematically reconstructing it from a series
CT technique is the fact that it is an ideal tool Øren (1997) developed a process-based recon- of projections (Russ, 2002). All transmission
to characterize the internal structure of a struction procedure that incorporates grain- tomography devices are based on the same prin-
rock in three dimensions in a nondestructive size distribution and other petrographic data ciple: the object is positioned between a source
way while a limitation of this CT technique obtained from 2D thin sections to reconstruct and a detector and rotated while transmission
is that only small samples can be combined 3D sandstones. The essence of this approach images are recorded. Different kinds of probes
with a high spatial resolution and therefore was to build sandstone models that were ana- can be used for the visualization: X-rays, neu-
often many samples will need to be scanned logs of actual sandstones by stochastically trons, gamma rays, etc.
in order to obtain representative volumes. modeling the results of the main sandstone In this paper X-ray computed tomography
The relationship between sample size and forming processes: sedimentation, compac- (CT) is used. A set of 2D radiographs of the
obtained spatial resolution are discussed as tion, and diagenesis. While statistical models sample are collected at a sequence of transmis-
well as the influence on the spatial resolution based on correlation-function reconstruction sion angles, providing information on X-ray
exerted by some important technical aspects retain a similar degree of isotropy and homo- attenuation, which depends on the density and
like the used X-ray source and detector. This geneity as the experimental microstructure, atomic number of the components. By means
paper focuses in detail on the structures that they tend to significantly underestimate the of reconstruction software the radiographs are
can be determined by means of micro- and connectivity properties. Although the recon- converted into a stack of 2D cross sections that
high-resolution X-ray CT in combination struction method proposed by Bakke and Øren allow 3D renderings of the sample. Although in
with 3D algorithms. was promising, it was unclear how accurately the beginning X-ray computed tomography was
it can reproduce more heterogeneous and dia- mainly used in medicine, it soon became clear
INTRODUCTION genetically complex sandstones such as those that X-ray CT had a large potential for other
encountered in nature. Øren and Bakke (2002) applications, including palaeontology (Fourie,
Conventional microstructure analysis of rock stated that although X-ray microtomography 1974; Conroy and Vannier, 1984; Haubitz et al.,
samples is usually carried out by studying two- was extremely useful, it was still not readily 1988; Chapman et al., 2003; Hlusko et al.,
dimensional (2D) polished thin sections with available in practice and therefore information 2004), sedimentology (Kenter, 1989; Peyton
optical microscopy or scanning electron micros- about the pore structure of porous materials et al., 1992; Zeng et al., 1996; Boespflug et al.,
copy (SEM). When looking at 2D images, it is was often limited to 2D thin-section images. 1995), petrology (Van Geet et al., 2001), soil
necessary to keep in mind that the shape of min- Consequently, the attractive approach was to science (Allan et al., 2002), and fluid-flow
erals and their spatial relationships can be much reconstruct a 3D description of the pore struc- research (Wellington and Vinegar, 1987; Géraud
more complex in three dimensions. Although ture from readily available 2D images, by et al., 2003; Dunsmuir et al., 1991; Coles et al.,
there are powerful quantitative tools to interpret means of statistical models for reconstructing 1994; Spanne et al., 1994; Hazlett, 1995; Coles
data derived from 2D images for the character- 3D porous media from 2D thin-section images et al., 1996; Coker et al., 1996).
ization of three-dimensional (3D) structures, (Joshi, 1974; Adler et al., 1990; Adler et al., Industrial X-ray CT scanners proved to be
many quantitative and qualitative aspects of 1992; Adler, 1992; Hazlett, 1997; Yeong and successful for research on geomaterials but with
structures remain inaccessible from 2D images Torquato, 1998a, 1998b). a resolution of a few hundreds of μm only. To

Geosphere; February 2011; v. 7; no. 1; p. 54–61; doi: 10.1130/GES00563.1; 11 figures; 1 table.

For permission to copy, contact [email protected]

54
© 2011 Geological Society of America
 Downloaded from geosphere.gsapubs.org on July 3, 2012

3D characterization of sandstone

obtain (sub)micron-scale resolution, currently if the entire sample width is not imaged. Since in a heterogeneous material can be determined
two options exist: synchrotron sources and X-ray detectors have a limited number of pixels, either by a few measurements on large volumes,
laboratory high-resolution X-ray CT (HRXCT) image resolution is limited to the discretization or by measurements of many small volumes of
scanners. Synchrotron sources allow submicron of the sample width, the latter forming the main material. They conclude that the representative
resolution with a monochromatic beam of well- practical limit for computed tomography. volume element must be considered as a func-
defined energy, but prove to be inaccessible for As mentioned earlier, a small sample size tion of several parameters including the physical
daily use for most research groups. More acces- is crucial for a high spatial resolution. How- property, the contrast of properties, the volume
sible laboratory equipment includes medical ever, working with small rock samples raises fractions of components, the desired relative
and micro-CT scanners (resolutions varying the question of the representative volume. In precision for the estimation of the effective
from roughly 100 to 1 μm, respectively) using order to obtain an acceptable estimation of the property, and the number of samples associated
X-ray tubes instead of synchrotron sources. general rock properties, the samples to be ana- with computations that one is ready to carry
HRXCT systems are currently being developed lyzed should be representative of the rock body out. The primary question, however, remains:
with submicron resolution. The CT scanners itself, including its normal spatial variations. what characteristic or property does one want
with micron or submicron resolution provide an Theoretically, a sample is only representative to see in 3D by means of X-ray? Is it to deter-
alternative to synchrotron-based research. One when its analysis results are both accurate and mine overall parameters like total porosity to be
of the advantages of CT systems is that they can reproducible (Gy, 1994). Sampling accuracy used as the overall porosity value for the stone
nondestructively provide stacks of more than is achieved when a few qualitative rules are in general? Or does one want to look in detail
1000 2D cross sections of the sample under respected. Correct sampling will provide all con- at a certain volume of interest in order to solve
investigation much faster than by performing stitutive elements to be evaluated with an equal some remaining questions? In this paper we
serial sectioning. Since no system is perfect, it probability of being selected and of belonging will focus on the detail that can be detected
is important to understand the advantages and to the sample (Gy, 1994). Three major factors with high-resolution X-ray CT rather than on
limitations of this technique in order to properly influence the accuracy of sampling: the sample the porosity values or the grain-size distribution
use it to address particular geological questions. grain size, the sample volume, and the amount results as the overall properties. As an example,
An important issue when working with CT is of samples taken (Smith, 1999). All these fac- we present an analysis conducted on the Bray
the relationship between sample size and spa- tors are in relation with each other. If a rock sandstone. The Bray sandstone has already
tial resolution. In HRXCT, the best achievable type has a high internal heterogeneity, a larger been characterized in different ways (Cnudde
resolution is related to the spot size of the X-ray amount of samples should be taken. Many stud- and Jacobs, 2004; Cnudde, 2005; Cnudde et al.,
source and the magnification of the system. ies were performed on sampling by Gy (1967, 2008), so the results of these experiments can be
The magnification is limited by the diameter 1971, 1979), Goodsall and Mathews (1970), and compared with the results of the high-resolution
of the sample under investigation. The resolv- Sedman and Stanley (1990). General rules like scan presented here.
ing power of an imaging system is the smallest the one formulated by Gy (1979) predict that Several parameters play a role in the classifi-
distance between two features (point, lines) at a good estimate of the required sample size is cation of different minerals in 3D derived from
which these features can be distinguished from obtained if X-ray CT scanning: the contrast in X-ray attenua-
one another. tion between those minerals, which is dependent
The resolving power that can be achieved Ms ≥ K· d 3, (3) on the density and average atomic number of
within the image is these minerals, and the size of the minerals in
where Ms is the sample weight and d is the diam- relation to the spatial resolution. Comprehen-
d ⎛ 1⎞ eter of the largest particle (Gy, 1979). Depend- sion of the image acquisition technique based on
R= + 1 − ⎟ s, (1)
M ⎜⎝ M⎠ ing on the degree of variation in grain size, the X-ray transmission and of the methods used for
parameter K is given a certain value (Gy, 1979). extraction of 3D data from a data set is crucial
with R the achievable resolution (Equation 1) in In case of a high degree of variation in grain if one is to obtain reliable quantitative measures
the object, s the spot size of the X-ray source, d size, the sample size might have to be doubled of features of interest. 3D image analysis starts
the pixel size of the detector, and M the magni- compared to samples with a low degree of varia- with high-quality images, since this is essential
fication, which is related to the position of the tion. However, in the case of sampling rocks to to retrieve valuable and useful information from
object and the separation between the source be used as building stone, the size of the blocks the original images. The use of image analysis
and detector: produced has to be taken into account as well generally involves the implementation of com-
as an estimation of the variation between the plex image processing (Starkey and Samantaray,
M = (total source-detector distance)/ blocks themselves. Jefferson (1993) suggests 1994) and of advanced calculation algorithms
(2)
(total source-object distance). that, as the properties of a block of stone such as (Pirard, 1994). Classical image analysis consists
its durability are determined by the properties of of noise reduction, segmentation, and binary
For low magnifications, the detector pixel size its constituent parts, the size of these constituent image editing. But before performing image
is the limiting factor for the image resolution parts can be used as a guide to determining the analysis, the original images often need to be
(Equation 1). For high magnifications, the X-ray sample size. Another approach, suggested by processed. This image processing is primarily
source spot size becomes significant, resulting Kanit et al. (2003) states that the representative performed to improve the visual appearance of
in a fuzzy image. The spot size thus limits the volume element size can be associated with a the images and to prepare for measurements of
resolution of the whole setup. Furthermore, to given precision of the estimation of the wanted features and structures present. A wide range
achieve high magnification the targeted part of overall property and the number of realizations of different techniques exists for each of these
the sample must be positioned very close to of a given volume of microstructure one is able tasks. Several software packages are available
the X-ray source, which is impossible for large to consider. For example, Kanit et al. (2003) for analysis of 2D images. Software for 3D
samples. Moreover, images can contain artifacts argue that the overall volume fraction of a phase analysis, however, is less frequently available or

Geosphere, February 2011 55

 Downloaded from geosphere.gsapubs.org on July 3, 2012

Cnudde et al.

sometimes not well adapted to specific needs. the roundness in thin sections, the relation of the at 300 ms exposure time in order to obtain a
Therefore the new 3D software Morpho+ various particle diameters (length, width, and high signal-to-noise ratio. The total scan time
(Vlassenbroeck et al., 2007) was developed in thickness) is calculated; specifically the degree was ~90 min. A source-to-detector distance
order to overcome the existing limitations to to which the shape of a grain approaches that of of 890 mm and a source-to-object distance of
analyze (pore) structures inside reconstructed a sphere (Bates and Jackson, 1980) is of inter- 51.8 mm resulted in a magnification of 17.2,
micro-CT images. est. In general, there is no real preferred orienta- equivalent to a voxel pitch of 7.4 μm (detector
In this paper, the potential of high-resolution tion of the grains; they are also closely packed pixel pitch of 127 μm). For the small sample,
X-ray CT in combination with the Morpho+ 3D (Fig. 1). The average open porosity of the Bray the spot size was set to its minimal size of below
analysis software for petrographic purposes is sandstone determined with water absorption 1 μm, limiting the available target current to
explored on a Belgian Bray sandstone sample. under vacuum is 14%, with a minimum of 9 μA, or ~1 W target power. A voxel pitch of
The advantages and limitations of these tech- 4% and a maximum of 24% (Cnudde, 2005; 700 nm was achieved by a source-to-detector
niques are discussed. The data obtained with Cnudde et al., 2008). The average pore diameter distance of 890 mm and a source-to-object dis-
X-ray CT is compared with results from the determined with mercury intrusion porosimetry tance of 4.9 mm, equivalent to a magnification
more conventional techniques of optical micros- (MIP) ranges from 4.7 to 20.1 μm, with an aver- of 181.6. Due to the low X-ray flux, no beam-
copy and scanning electron microscopy. age of 15.7 μm. The average threshold pressure hardening filter was applied. Four frames at an
derived from MIP was between 11 μm and exposure time of 1000 ms were averaged for each
MATERIALS 16 μm depending on the degree of cementation. of the 1500 projections, extending the total scan
The density of the Bray sandstone is on average time to nearly 3 h. The Varian PaxScan 2520V
The Bray Sandstone 2281 kg/m3 (Cnudde, 2005). (a-Si flat panel, CsI screen, 1880 horizontal by
Two different samples of the Bray sandstone 1496 vertical pixels, 127 μm pixel size) recorded
The Bray sandstone is a quartz arenite from were analyzed for this study. A large rectangular images for both samples, although its orienta-
the Thanetian (Paleocene, Paleogene). This sample with a cross section of 6.6 × 7.4 mm2 tion was changed for the different purposes. This
natural building stone has mainly been used for was scanned. Because of the size of the sample, resulted in an image width of 1496 pixels for the
monuments in Binche, Mons, and Bray (Bel- only part of the sample height could be imaged low resolution scan, and 1880 pixels to achieve
gium) and in houses (Cnudde, 2005). The Bray in order to obtain the desired resolution. The maximum resolution for the small sample. Both
sandstone is a heterogeneous accumulation of second sample is a drilling core with only 1 mm samples were reconstructed using the soft-
mostly continental deposits consisting of quartz diameter and a height of ~500 μm. Its small size ware package Octopus (http://www.inct.be).
grains with a siliceous cement. Its color varies allowed scanning at very high resolution. The software Morpho+ (Vlassenbroeck et al.,
from gray to yellowish brown, depending on the 2007) was used for the 3D analysis of the
amount of iron oxide and/or hydroxides, which METHODS pores and the grains of the Bray sandstone. In
occur as a very thin coating around the grains. 3D analysis, a volume composed of voxels is
In addition to monocrystalline quartz grains and High-Resolution X-ray CT (HRXCT) processed to extract quantitative data about the
sometimes polycrystalline quartz, feldspars, Combined with 3D Analysis by Morpho+ structural composition of the sample. The vol-
rutile, zircon, some mica, and clays can be iden- ume can be represented as a stack of images,
tified in thin sections. Some quartz grains show A flexible multifunctional high-resolution although the operations need to be performed
a typical overgrowth; in many cases the shape X-ray CT (HRXCT) scanner allowing 2D in 3D. The objective of the analysis is to obtain
of the original grain is delineated by a thin iron radiography and 3D visualization and quanti- parameters such as size, shape, and orientation
oxide and/or hydroxides or clay coating between fication on a (sub)micron scale was used. The for each object (for example, a pore or a grain)
the overgrowth and the grain itself. The presence scanner was developed at the “Centre for X-ray inside the sample. In Morpho+, the analysis
of the clay minerals in the sandstone can be very Tomography” of the Ghent University, Belgium
significant as it strongly affects rock permeabil- (Masschaele et al., 2007). The transmission tube
ity and porosity. The Bray sandstone is a fine- head was operated at a high voltage of 100 kV
grained, poorly sorted sandstone (Cnudde et al., for both samples. The large sample was scanned
2004). Grains are angular to subrounded, with at a target current of 80 μA, resulting in 8 W
an average grain size of 0.156 mm and low in target power. In this configuration, the X-ray
sphericity (Cnudde and Jacobs, 2004). Accord- spot size is ~5 μm in diameter. A 550-μm-thick
ing to the formulas of Folk and Ward (1957) aluminum plate was used as a beam-hardening
(Table 1), the grains could be defined as poorly filter. In total 1000 projections were recorded,
sorted and positively fine-skewed. To determine each projection being an average of three frames

TABLE 1. FORMULAS FOR THE CALCULATION OF GRAIN-SIZE PARAMETERS FROM A

GRAPHIC PRESENTATION OF THE DATA IN A CUMULATIVE FREQUENCY PLOT
Parameter Graphic formula
Mean M = (φ16 + φ50 + φ84)/3
Median Md = φ50 200 μm
Sorting σφ = [(φ84 − φ16 )/4] + [(φ95 − φ5)/6.6]
Skewness Sk = [(φ16 + φ84 − 2φ50 )/2(φ84 − φ16 )] + [(φ5 + φ95 − 2φ50 )/2(95 − φ5)]
Kurtosis KG = (φ95 − φ5)/2.44(φ75 − φ25) Figure 1. Scanning electron microscopy
Notes: Formulas after Folk and Ward (1957). φx represents the phi size values to the x (SEM) image of the Bray sandstone (mag-
percentiles read from the cumulative frequency curve. nification 70×).

56 Geosphere, February 2011

 Downloaded from geosphere.gsapubs.org on July 3, 2012

3D characterization of sandstone

process is composed of several steps, includ- als present in the Bray sandstone, zircon has the equal to 7.4 μm are not included in this volume
ing filtering, phase segmentation or threshold- highest attenuation coefficient, while quartz has analysis since they are below the resolution.
ing, labeling, and object separation. By using an the lowest (Cnudde, 2005; Gualda and Rivers, In addition to determining porosity and over-
algorithm-based dual threshold, the original data 2006). The quartz grains can be clearly detected, all pore shapes and spatial relationships, each
was transformed into a binary data set with the as well as some dense inclusions of rutile and individual grain can be analyzed as well. There-
aim to extract structural features of the selected zircon, based on a microscopic examination fore grains are first segmented and automati-
pore volume. Dual thresholding uses two inter- of the samples prior to CT scanning. Besides cally identified by a watershed separation algo-
vals; voxels with a gray value in the first interval cross sections, 3D renderings (Fig. 4) can be rithm based on the Euclidean distance transform
are classified as foreground voxels, while voxels produced for visual interpretation of the mineral (Russ, 2002). Based on analysis of the grain
in the second interval are only defined as fore- shape and their spatial relationships. However, size in the 343 mm3 volume sample, an aver-
ground voxels if they are connected to voxels for quantitative results a 3D analysis is required. age equivalent diameter for the Bray sandstone
from the first interval. This approach reduces the Analyzing the pore-volume of the scanned sam- grains of 194 (±53) μm was found together with
sensitivity to residual image noise. One of the ple with Morpho+ revealed an average porosity an average maximum opening of 117 (±34) μm.
analyzed parameters, which is necessary for our of 18 vol% over an analyzed partial volume of In this volume, a total of 53,103 grains were
shape analysis, is the maximum opening, which 343 mm3. However, all pores smaller than or analyzed.
can be extracted from the distance transform.
The maximum opening is defined as the diam-
eter of the maximum inscribed sphere which fits
inside the object (Fig. 2). For each grain its total
volume can be determined as well as its surface. Figure 2. Two-dimensional
If a sphere with the same total volume of a grain (2D) illustration of the maxi-
is constructed, its corresponding diameter is mum opening of an object.
defined as the equivalent diameter of the grain.
Morpho+ calculates the sphericity S as the ratio
of the maximum opening over the equivalent
diameter. This parameter gives a rough approxi-
mation of the object shape since it expresses
how much an object resembles a sphere (S = 1).
S ≈ 0 typically corresponds to a large network
composed of narrow channels.
Additionally, each object can be modeled
as an equivalent ellipsoid which has the same
moments of inertia. The three-dimensional
moments of the inertia tensor are then calcu-
Figure 3. Reconstructed cross
lated and diagonalized. The corresponding rota-
section of the Bray sandstone
tion of the eigenvectors defines the orientation
scanned with a resolution of
of the object, while the eigenvalues correspond
7.4 μm.
to the lengths of the principal axes of the equiva-
lent ellipsoid. Since Morpho+ can determine the
orientation of the objects of interest such as the
grains constituting the sandstone, it is possible
to produce a stereoplot derived from the 3D
analysis of the grains. The processed volume
1000 μm
can be visualized after each step in the analysis
process. Since all analysis algorithms operate in
three dimensions, cross sections of the volume
according to the different principal planes can
be visualized.

RESULTS

Micro-CT Scan (Resolution 7.4 μm) Figure 4. Three-dimensional

(3D) rendered volume of the
Scanning of the Bray sandstone with a cross Bray sandstone scanned with
section of 6.6 × 7.4 mm2 resulted in a resolution a resolution of 7.4 μm.
of the micro-CT data of 7.4 μm. Figure 3 rep-
resents a reconstructed cross section. Based on
the different X-ray attenuation of the composing
minerals, different gray values between the dif- 1000 μm
ferent minerals can be detected. Of all miner-

Geosphere, February 2011 57

 Downloaded from geosphere.gsapubs.org on July 3, 2012

Cnudde et al.

Figure 5 shows details of the 3D renderings the intergranular structure, and to identify The grain contact varies from point (grains
of the results of Morpho+, where the grains in details inside the grains (Fig. 7). Monocrystal- touch each other) to sutured contacts (mutual
the selected volume have been relabeled (color line quartz grains can easily be detected on the interpenetration of grains).
coded) based on a selected parameter, such as, reconstructed cross sections as well as mineral Volume rendering of the 2D reconstructions
e.g., their equivalent diameter. From the grain inclusions like rutile and zircon, and voids. One reveals the internal structure of the stone with
analysis performed on the micro-CT scan, the has to be very careful when identifying minerals high 3D detail (Fig. 8). Besides looking at the
largest equivalent diameter found was 375 μm on X-ray CT reconstructed images since X-ray structures themselves, it is also possible to virtu-
and the largest maximum opening 231 μm. The attenuation coefficients are a function of both ally cut the sample at any desired location and/or
average sphericity of the analyzed grains based atomic number and material density. However, angle and to make certain components transpar-
on the micro-CT scan was 0.6. This value indi- based on experience with the material most ent while rendering other features opaque.
cates that on average the grains are not perfectly probably clays and local thin iron oxide coat- When investigating the images of the high-
spherical. Figure 6 illustrates the frequency dis- ings can be detected in the reconstructions. resolution scan with Morpho+, a total volume
tribution of the grain sphericity. The frequency Since feldspars and quartz have a more similar of 0.57 mm3 was analyzed, corresponding to a
distribution shows a well-sorted curve, which attenuation coefficient, they are more difficult mass of 0.013 g. Although much more detail
corresponds with the estimation made on the to distinguish from one another. A small varia- can be detected in this high-resolution CT scan,
basis of the 2D reconstructed images. tion in attenuation coefficient toward the edge many similar samples should be examined in
of the sample can be seen on Figure 7, caused order to obtain a representative element vol-
High-Resolution CT Scan by phase-contrast. However, this feature is not ume. As the subsample volume is so small, only
(Resolution 0.7 μm) problematic for the segmentation purpose since a small amount of grains, a total of 710, could
by using dual thresholding the correct features be analyzed on their equivalent diameter, maxi-
In order to achieve a resolution on the order can still be segmented. mum opening, and sphericity. Figure 9 shows a
of 700 nm, a subsample of the Bray sandstone Based on the reconstructed 2D cross sections, 3D rendering of the results of Morpho+, where
was selected. This resolution enables one to the grains of the Bray sandstone were character- each grain has been relabeled (color coded)
clearly distinguish the different grains, to assess ized as subangular over subrounded to rounded. based on its equivalent diameter. The average
grain sphericity calculated based on the high-
resolution scan was 0.57. This value indicates
that on average the grains are not perfectly
spherical, although some grains definitely are
(sphericity value = 1).

Comparison of High-Resolution and

Micro-CT Scan

The results of the equivalent grain diameter

analysis of the high-resolution (0.7 μm) and
micro-resolution (7.4 μm) scans are shown in
Figure 10, where the ratio of the volume in μm3
of the grains over the total analyzed volume is
plotted against their corresponding equivalent
500 μm 500 μm diameter results. The 3D analysis demonstrated
that in the high-resolution scan on average, more
Figure 5. Three-dimensional (3D) rendered volume of the color-coded Bray sandstone
scanned with a resolution of 7.4 μm analyzed with Morpho+: the blue objects have a small
equivalent diameter, while the red ones are the largest.

Frequency of sphericity
25,000

20,000
Frequency

15,000
Figure 6. Sphericity distribution
of the analyzed grains obtained 10,000
by the micro-CT scan.
5000 200 μm

Figure 7. Reconstructed cross section of the

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Bray sandstone sample scanned with a reso-
Sphericity lution of 700 nm.

58 Geosphere, February 2011

 Downloaded from geosphere.gsapubs.org on July 3, 2012

3D characterization of sandstone

Figure 8. Three-dimensional
(3D) rendered volume of the
Bray sandstone sample scanned
with a resolution of 700 nm.
The dense inclusions are visual-
ized in red.

200 μm 200 μm 200 μm

smaller objects were found, which could not A B

be distinguished in the micro-scan. Also more
grains with a lower equivalent diameter were
detected than on the micro-CT scan. The aver-
age equivalent diameter of the grains scanned at
high resolution was 39 (±56) μm and the aver-
age maximum opening 24 (±38) μm, which
was much lower than the respective values
calculated from the micro-CT scan. Based on
Equation (3), it is likely that the amount of ana- C D
lyzed volume in the high-resolution scan is too
small to correctly characterize the average grain
sizes. Additionally, the amount of small grains
is much higher than in the micro-scan, where
small grains were not detected due to resolution
limitations. Based on the frequency distribution
curve calculated for the high-resolution scan
(Fig. 11), it appears that the amount of exam- 100 μm
ined grains is too low to provide a good estimate
of sorting and skewness. Figure 9. (A) Three-dimensional (3D) rendering of the grains after labeling with Morpho+
based on their equivalent diameter (blue objects have the smallest equivalent diameter,
CONCLUSION AND DISCUSSION while red objects have the largest); (B) selection of the largest grains; (C) selection of the
middle-sized grains; and (D) selection of the small grains.
In this paper, the potential of high-resolution
X-ray CT combined with 3D analysis software
for petrographic purposes was explored for a
Belgian Bray sandstone sample. The advantages the scanned volume, the higher the detail, but An average grain size between 0.125 and
of this technique reside in the fact that X-ray this fact also influences the representative ele- 0.25 mm was determined by means of point
CT is an ideal tool to characterize the internal ment volume. counting on thin sections. An average equivalent
structure of a rock in three dimensions in a non- The total porosity of the Bray sandstone was diameter of 194 (±53) μm and an average maxi-
destructive way. Besides looking at the struc- determined in two different ways. The average mum opening of 117 (±34) μm were found for the
tures themselves, CT also allows us to virtually open porosity determined with water absorption grains scanned with the micro-CT scan while an
cut through the sample at any given location and under vacuum was 14%, with a minimum of average equivalent diameter of 39 (±56) μm and
to make certain components transparent while 4% and a maximum of 24%. The average pore an average maximum opening of 24 (±38) μm
rendering other features opaque. Due to rapid diameter determined with MIP ranged from were detected from the high-resolution scan.
technological developments, laboratory equip- 4.7 to 20.1 μm, with an average of 15.7 μm. Point counting overlooked the smallest grains,
ment at the moment reaches resolutions at sub- When analyzing the pore-volume of the micro- which increases the average grain size, bringing
micron scale. The flexible high-resolution X-ray CT scanned sample, its average porosity was it closer to the results of the micro-CT scan. On
CT scanner at the Centre for X-ray Tomography 18 vol% over an analyzed volume of 343 mm3. the other hand, it should be kept in mind that the
(Ghent University, Belgium) is capable of imag- However, all pores smaller than and equal to analyzed volume of the high-resolution scan is
ing at a resolution of 700 nm for rock samples. 7.4 μm are not included in this volume analysis too low to provide representative data.
At this magnification it is possible to clearly dis- due to the limited resolution. For shape determi- Based on optical microscopy and SEM
tinguish different grains with diameters larger nation a minimum amount of voxels need to be images, the grains are visually qualified as
than the resolution, to determine and appreci- present. In this study objects with a radius equal angular to subrounded, with a low sphericity
ate their intergranular structure, and to identify to or larger than 3 voxels were included in the and poorly sorted. In general, no real preferred
details inside the grains. However, the smaller analysis. orientation of the grains is observed; they are

Geosphere, February 2011 59

 Downloaded from geosphere.gsapubs.org on July 3, 2012

Cnudde et al.

Equivalent diameter grains REFERENCES CITED

Volume (μm3) grains/total analyzed volume (μm3) (× 106)

160 Adler, P.M., 1992, Porous media: Geometry and transports:

Stoneham, Butterworth/Heineman.
Adler, P.M., Jacquin, C.G., and Quiblier, J.A., 1990, Flow
140 in simulated porous media: International Journal of
Multiphase Flow, v. 16, p. 691–712, doi: 10.1016/
120
0301-9322(90)90025-E.
Adler, P.M., Jacquin, C.G., and Thovert, J.F., 1992, The
formation factor of reconstructed porous media: Water
100 Resources Research, v. 28, p. 1571–1576, doi: 10.1029/
92WR00059.
Allan, C.C., Tsapin, A.I., Kuebler, K., Haskin, L., and Wang,
80 A., 2002, Analysis inside the box- studying rock and
micro-scan soil in biological quarantine: Lunar Planet. Science
XXXIII, abstract 1222.
60 high-resolution (per 3)
Bakke, S., and Øren, P.E., 1997, 3-D pore-scale modelling of
sandstones and flow simulations in the pore networks:
40 SPE Journal, v. 2, no. 2, p. 136–149.
Bates, R.L., and Jackson, J.A., 1980, Glossary of geology
(second edition): Falls Church, Virginia, American
20 Geological Institute, 751 p.
Boespflug, X., Long, B.F.N., and Occhietti, S., 1995, Cat-
scan in marine stratigraphy—A quantitative approach:
0 Marine Geology, v. 122, no. 4, p. 281–301, doi: 10.1016/
0 50 100 150 200 250 300 350 400 0025-3227(94)00129-9.
Carlson, W., 2006, Three-dimensional imaging of earth
μm)
Equivalent diameter (μ and planetary materials: Earth and Planetary Science
Letters, v. 249, p. 133–147, doi: 10.1016/j.epsl.2006
Figure 10. Equivalent grain diameter data for both scans. .06.020.
Chapman, R.E., Andersen, A., and Wilcox, B., 2003, Study-
ing extinct animals using three-dimensional visualiza-
tion, scanning, animation, and prototyping, in Woods,
A.J., Bolas, M.T., Merritt, J.O., and Benton, S.A., eds.,
Frequency of sphericity Stereoscopic displays and virtual reality systems x,
Proceedings of the Society of Photo-Optical by Instru-
250 mentation Engineers (SPIE) no. 5006, p. 469–477.
Cnudde, V., and Jacobs, P., 2004, Monitoring of weathering
200 and conservation of building materials through non-
destructive X-ray computed microtomography: Envi-
Frequency

Figure 11. Sphericity distri- 150 ronmental Geology, v. 46, p. 477–485, doi: 10.1007/
s00254-004-1049-5, doi: 10.1007/s00254-004-1049-5.
bution of the analyzed grains Cnudde, V., 2005, Exploring the potential of X-ray tomog-
obtained by the high-resolution 100 raphy as a new non-destructive research tool in conser-
CT scan. vation studies of natural building stones [Ph.D. thesis]:
50 Ghent University, p. 192.
Cnudde, V., Dierick, M., Vlassenbroeck, J., Masschaele,
0 B., Lehmann, E., Jacobs, P., and Van Hoorebeke, L.,
2008, High-speed neutron radiography for monitoring
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 the water absorption by capillarity in porous materi-
Sphericity als: Nuclear Instruments and Methods in Physics
Research, Section B: Beam Interactions with Materials
and Atoms, v. 266, no. 1, p. 155–163, doi: 10.1016/
j.nimb.2007.10.030.
Coker, D.A., Torquato, S., and Dunsmoir, J.H., 1996, Mor-
close packed with more concave-convex con- data to obtain statistically correct grain-size phology and physical properties of Fontainebleau
tacts. The average sphericity of the analyzed distributions. sandstone via tomographic analysis: Journal of Geo-
physical Research, v. 101, p. 17,497–17,506, doi:
grains based on the micro-CT scan was 0.6. The A convincing advantage of CT data is formed 10.1029/96JB00811.
frequency distribution shows a more well-sorted by the fact that measurement results can be Coles, M.E., Hazlett, R.D., Muegge, E.L., Jones, K.W.,
Andrews, B., Siddons, P., Peskin, A., and Soll, W.E.,
curve, which corresponds with the estimation visualized in 3D, where grains in a selected vol- 1996, Developments in synchrotron X-ray micro-
based on the 2D reconstructed images. On the ume can be relabeled (color coded) based on a tomography with applications to flow in porous media:
reconstructed 2D cross sections of the high- chosen parameter. Unfortunately CT still has its Proceedings 1996 SPE Annual Technical Conference
and Exhibition, Denver, Oct. 6–9, Society of Petroleum
resolution scan, the Bray sandstone grains can limitations and should be used for the right pur- Engineers, paper SPE 36531.
visually be classified as angular to subrounded. poses in the correct conditions. Problems like a Coles, M.E., Spanne, P., Muegge, E.L., and Jones, K.W.,
The grain contact varies from point contacts to large sample size combined with a high spatial 1994, Computer microtomography of reservoir core
samples: Proceedings 1994 Annual SCA Meeting,
sutured contacts. The average sphericity of the resolution will always be encountered when Stavanger, Norway, Sept. 12–14, Society of Core
grains analyzed on the high-resolution scan was using X-ray CT and often many samples will Analysts.
Conroy, G.C., and Vannier, M.W., 1984, Noninvasive
0.57, corresponding to the value determined need to be scanned in order to obtain represen- three-dimensional computer imaging of matrix-filled
from the micro-CT scan; on average the grains tative volumes. fossil skulls by high-resolution computed tomog-
are not perfectly spherical. raphy: Science, v. 226, p. 456–458, doi: 10.1126/
ACKNOWLEDGMENTS science.226.4673.456.
High-resolution CT scans provide more Dunsmuir, J.H., Ferguson, S.R., D’Amico, K.L., and Stokes,
structural detail due to their higher resolution. The Fund for Scientific Research—Flanders J.P., 1991, X-ray microtomography: A new tool for
(FWO) is acknowledged for the post-doctoral grant to the characterization of porous media: Proceedings
However, since the volume examined is smaller V. Cnudde. The Institute for the Promotion of Innova- 1991 SPE Annual Technical Conference and Exhibi-
than for micro-CT scans, the amount of analyzed tion by Science and Technology in Flanders, Belgium, tion, Dallas, Oct. 6–9: Society of Petroleum Enginers,
grains is much smaller, introducing insufficient is acknowledged for the Ph.D. grant to J. Dewanckele. paper SPE 22860.

60 Geosphere, February 2011

 Downloaded from geosphere.gsapubs.org on July 3, 2012

3D characterization of sandstone

Folk, R.L., and Ward, W., 1957, Brazos River bar: A study thickness: A baboon model: American Journal of Smith, M.R., 1999, Stone: Building stone, rock fill and
in the significance of grain size parameters: Journal of Physical Anthropology, v. 124, no. 3, p. 223–233, doi: armourstone in construction: Geological Society
Sedimentary Petrology, v. 27, p. 3–26. 10.1002/ajpa.10353. [London], Engineering Geology, Special Publication
Fourie, S., 1974, The cranial morphology of Thrinaxondon Holt, R.M., Fjoer, E., Torsoeter, O., and Bakke, S., 1996, No. 16, 478 p.
liohinus Seeley: Annals of the South African Museum, Petrophysical laboratory measurements for basin and Spanne, P., Thovert, J.F., Jacquin, C.J., Lindquist, W.B.,
v. 65, p. 337–400. reservoir evaluation: Marine and Petroleum Geology, Jones, K.W., and Adler, P.M., 1994, Synchrotron
Géraud, Y., Surma, F., and Mazerolle, F., 2003, Porosity v. 13, p. 383–391, doi: 10.1016/0264-8172(95)00091-7. computed microtomography of porous media: Topol-
and fluid flow characterization of granite by capillary Jefferson, D.P., 1993, Building stone: The geological ogy and transports: Physical Review Letters, v. 73,
wetting using X-ray computed tomography, in Mees, dimension: Quarterly Journal of Engineering Geol- p. 2001–2004, doi: 10.1103/PhysRevLett.73.2001.
F., Swennen, R., Van Geet, M., and Jacobs, P., eds., ogy, v. 26, p. 305–319, doi: 10.1144/GSL.QJEGH Starkey, J., and Samantaray, A.K., 1994, A microcomputer-
Applications of X-ray computed tomography in the .1993.026.004.06. based system for quantitative petrographic analysis:
Geosciences: Geological Society [London] Special Joshi, M., 1974, A class of stochastic models for porous Computers & Geosciences, v. 20, no. 9, p. 1285–1296,
Publication 215, p. 95–106. media [Ph.D. thesis]: Lawrence, University of Kansas. doi: 10.1016/0098-3004(94)90055-8.
Goodsall, G.D., and Mathews, D.H., 1970, Sampling of road Kanit, T., Forest, S., Galliet, I., Mounoury, V., and Jeulin, D., Van Geet, M., Swennen, R., and Wevers, M., 2001, Towards
surfacing materials: Journal of Applied Chemistry, 2003, Determination of the size of the representative 3D petrography: Application of microfocus computed
v. 20, p. 361–366. volume element for random composites: Statistical and tomography in geological science: Computers & Geo-
Gualda, G., and Rivers, M., 2006, Quantitative 3D petrog- numerical approach: International Journal of Solids sciences, v. 27, p. 1091–1099, doi: 10.1016/S0098-3004
raphy using x-ray tomography: Application to Bishop and Structures, v. 40, p. 3647–3679, doi: 10.1016/ (00)00154-0.
Tuff pumice clasts: Journal of Volcanology and Geo- S0020-7683(03)00143-4. Vlassenbroeck, J., Dierick, M., Masschaele, B., Cnudde,
thermal Research, v. 154, p. 48–62, doi: 10.1016/ Kenter, J.A.M., 1989, Applications of computerized tomog- V., Van Hoorebeke, L., and Jacobs, P., 2007, Software
j.jvolgeores.2005.09.019. raphy in sedimentology: Marine Geotechnology, v. 8, tools for quantification of X-ray microtomography
Gy, P., 1967, L’échantillonnage des Minerais en Vrac. p. 201–211, doi: 10.1080/10641198909379868. at the UGCT: Nuclear Instruments and Methods in
Tome 1: Théorie Générale. Special issue of Revue de Lin, C., and Cohen, M.H., 1982, Quantitative methods of Physics Research, Section A: Accelerators, Spectrom-
L’industrie Minérale, January 15 1967. Saint-Etienne: microgeometric modeling: Journal of Applied Physics, eters, Detectors and Associated Equipment, v. 580,
Société de l’Industrie Minérale. v. 53, p. 4152, doi: 10.1063/1.331238. p. 442–445, doi: 10.1016/j.nima.2007.05.073.
Gy, P., 1971, L’échantillonnage des Minerais en Vrac. Masschaele, B., Cnudde, V., Dierick, M., Jacobs, P., Van Wellington, S.L., and Vinegar, H.J., 1987, X-ray computer-
Tome 2: Théorie Générale, Erreurs Opératoires, Com- Hoorebeke, L., and Vlassenbroeck, J., 2007, UGCT: ized tomography: Journal of Petroleum Technology,
pléments. Mémoires du B.R.G.M., No. 67. Orléans: New X-ray radiography and tomography facility: v. 39, p. 885–898, doi: 10.2118/16983-PA.
Bureau de Recherches Géologiques et Minières. Nuclear Instruments and Methods in Physics Research, Yeong, C.L.Y., and Torquato, S., 1998a, Reconstructing
Gy, P., 1979, Sampling of particulate materials, theory Section A: Accelerators, Spectrometers, Detectors random media: Physical Review E: Statistical Physics,
and practice: Developments in Geomathematics v. 4: and Associated Equipment, v. 580, p. 266–269, doi: Plasmas, Fluids, and Related Interdisciplinary Topics,
Amsterdam, Elsevier Scientific Publishing Company. 10.1016/j.nima.2007.05.099. v. 57, p. 495–506.
Gy, P.M., 1994, Sampling or gambling: Process Control and Øren, P.-E., and Bakke, S., 2002, Process based reconstruc- Yeong, C.L.Y., and Torquato, S., 1998b, Reconstructing
Quality, v. 6, no. 2–3, p. 97–102. tion of sandstones and prediction of transport proper- random media. II. Three-dimensional media from
Haubitz, B., Prokop, M., Dohring, W., Ostrom, J.H., and ties: Transport in Porous Media, v. 46, p. 311–343, doi: two-dimensional cuts: Physical Review E: Statistical
Wellnhofer, P., 1988, Computed tomography of 10.1023/A:1015031122338. Physics, Plasmas, Fluids, and Related Interdisciplinary
Archaeopteryx: Paleobiology, v. 14, p. 206–213. Peyton, R.L., Haeffner, B.A., Anderson, S.H., and Gantzer, Topics, v. 58, p. 224–233.
Hazlett, R.D., 1995, Simulation of capillary dominated dis- C.J., 1992, Applying X-ray CT to measure macropore Zeng, Y., Gantzer, C.J., Peyton, R.L., and Anderson, S.H.,
placements in microtomographic images of reservoir diameters in undisturbed soil cores: Geoderma, v. 53, 1996, Fractal dimension and lacunarity of bulk density
rocks: Transport in Porous Media, v. 20, p. 21–35, doi: p. 329–340. determined with X-ray computed tomography: Soil
10.1007/BF00616924. Pirard, E., 1994, Shape processing and analysis using the Science Society of America Journal, v. 60, p. 1718–
Hazlett, R.D., 1997, Statistical characterization and stochas- calypter: Journal of Microscopy, v. 175, p. 214–221. 1724, doi: 10.2136/sssaj1996.03615995006000060016x.
tic modelling of pore networks in relation to fluid flow: Russ, J.C., 2002, The image processing handbook (fourth
Journal of Mathematical Geology, v. 29, p. 801–822, edition): Boca Raton, Florida, CRC press, 732 p.
doi: 10.1007/BF02768903. Sedman, J.H.F., and Stanley, L., 1990, Variations in the MANUSCRIPT RECEIVED 9 NOVEMBER 2009
Hlusko, L.J., Suwa, G., Kono, R.T., and Mahaney, M.C., physical properties of porous building limestones: REVISED MANUSCRIPT RECEIVED 26 SEPTEMBER 2010
2004, Genetics and the evolution of primate enamel Stone Industries, v. 25, no. 6, p. 22–24. MANUSCRIPT ACCEPTED 2 OCTOBER 2010

Geosphere, February 2011 61