Identification and Three-Dimensional Characterization of Micropore Networks Developed in Granite using Micro-Focus X-ray CT

Abstract

We analyzed the three-dimensional distribution of micropores and internal structures in both fresh and weathered granite using micro-focus X-ray computed tomography (micro-CT). Results show that the pore radius in fresh granite is mostly in the range of $17-50{\mu}m$, the throat radius is in the range of $5-25{\mu}m$, and the coordination number (CN) of pores is less than 10. In contrast, the pore radius in weathered granite is mostly in the range of $20-80{\mu}m$, the throat radius is in the range of $8-30{\mu}m$, and the CN is less than 12. In general, a positive linear relationship exists between pore radius and CN. In addition, both the size and the density of pores increase with an increasing degree of rock weathering. The size of the throats that connect the pores also increases with an increasing degree of weathering, which induces fracture propagation in rocks. Micro-CT is a powerful and versatile approach for investigating the three-dimensional distributions of pores and fracture structures in rocks, and for quantitatively assessing the degree of pore connectivity.

Introduction

In recent decades, micro-focus X-ray computed tomography (micro-CT) has been increasingly used as a diagnostic tool in medicine and in other fields, as the technique is non invasive and provides high-resolution, three-dimensional images of internal structures. The technique of micro-CT has also been increasingly applied in geological investigations, to study internal microstructures in a variety of geological materials such as rocks, fossils, and minerals (Nakashima et al., 2004; Cnudde et al., 2006; Jeong and Takahashi, 2010; Kim et al., 2011, Cnudde and Boune, 2013). Micro-CT is especially effective for evaluating the strength, permeability, and failure properties of rock materials, and for characterizing and analyzing the structure and distribution of internal pores and their complex networks and interconnections in porous media (Nemoto et al., 2009; Watanabe et al., 2009, 2011; Wildenschild and Sheppard, 2013). Using micro-CT, it is possible to image the heterogeneity of the rock matrix and to determine the grain size and porosity distributions of rock samples. Therefore, micro-CT has proven to be a promising technique in diverse domains of geological and materials research (Ketcham, 2005; Cnudde et al., 2006).

Although the impact of subsurface geochemical reactions on porosity is relatively well understood, the permeability characteristics of geological formations are difficult to estimate (Beckingham et al., 2013). Many uncertainties are encountered in the analysis of rock properties owing to inadequate information about the characteristics of underground sites and the variable nature of geological and geotechnical parameters, which are both inherent to the materials and a result of measurement errors. Starting in the late 1970s, with the advent of high-performance computing, the CT technique rapidly evolved when analog to digital converters were interfaced with monitoring systems, and computed tomography was increasingly applied in many industrial fields (Cnudde et al., 2006). Many CT applications were developed for medical purposes, using scanners with typically mm-scale resolutions and X-ray intensities that matched the density of the human body. The evolution of CT technology has resulted in the development of industrial scanners specifically designed to image geological samples and other highdensity materials. The application of X-ray CT imaging in the geosciences was introduced in the petroleum industry in the 1980s, allowing the nondestructive, three-dimensional visualization of internal microstructures in rock formations to be performed.

Although two-dimensional imaging is widely accessible and provides information quickly about both subgranular and intragranular mineralogical variations, two-dimensional images cannot provide direct information about network connectivity (Beckingham et al., 2013). Three-dimensional micro-CT images are particularly useful for understanding the relationship between fracture propagation/damage intensity and the nature of structures and microfabric orientations in rocks (Yun et al., 2013). The use of threedimensional methods to characterize pore networks has increased in recent years. Therefore, methodological approaches that allow systematic and quantitative analyses of rock characteristics to be conducted have become a subject of increasing interest in rock engineering, especially for the quantitative evaluation of the internal characteristics of rocks in three dimensions.

An understanding of fracture and porous network properties in rocks is crucial in many fields of engineering and environmental geology, because such physical properties significantly influence the strength of rocks and affect their permeability, especially rocks exposed to weathering, stress, or changes in loading due to excavation. In addition, such information is important for precisely predicting the storage capacities of rocks, as well as the mobility of fluids along fracture zones in a rock mass, which are commonly encountered in applications in fields such as engineering, groundwater, CO2 sequestration, and seismology. However, so far, the characteristics of pore net works have been difficult to identify. Thus, the objective of this research was to demonstrate the capability of a micro-CT approach for reliably calculating pore sized distributions and characterizing fracture patterns in granites, and thereby to improve our understanding of rock properties from the perspective of rock microstructure. In this study, we investigated two granite samples recovered from cores, especifically, samples of fresh granite and of highly weathered granite.

 

Experimental analysis and samples

Micro-CT

Micro-CT is a powerful and nondestructive technique that enables the internal structures of rock samples to be visualized (Takahashi et al., 2002, 2004). The images, which reflect variations in the atomic composition and density of materials, are obtained by reconstructing twodimensional sections perpendicular to an axis of rotation. The X-ray images record differences in the degree of attenuation of the X-rays, which is both material and energy dependent. The interactions responsible for attenuation are related mostly to Compton scattering and photoelectrical absorption. The contribution of the photoelectric effect depends on the atomic numbers of the constituent elements. As the X-ray attenuation is primarily a function of the X-ray energy and of the density and atomic composition of the material being imaged, it is possible that information on the chemical composition of the examined object can also be obtained (Cnudde et al., 2006).

The micro-CT system is composed of an X-ray source, a subject, and an image magnifier (Fig. 1). After X-ray transmission, the data are manipulated to reconstruct a magnified three-dimensional CT image (Fig. 2). In this study, we used a micro-CT system Micro-Focus X-Ray CT Scanner, model HMX 225-ACTIS+3. The analytical conditions were as follows: voltage of 225 kV; focal spot size of 5 μm; resolution, 12-bit charge-couple device (CCD); input window for image magnifier with diameter of Φ= 150 mm; scan type, BIR ACTIS+3 with off sets of 180°/360°/3 slices; and reconstruction time of 30 s.

Fig. 1.Sample holder and configuration of the micro-focus X-ray CT apparatus.

Fig. 2.Schematic of micro-focus X-ray CT apparatus.

Micro-CT resolution is defined in terms of voxels (a voxel is a volumetric pixel) and the z number, which is the spacing thickness) of a planar measurements lice. The x–y resolution, which is generally defined as the crosssectional diameter divided by the number of image pixels (512 or 1024), is largely affected by the number of detectors and by channel types. Fig. 2 illustrates the micro-CT apparatus, showing a collimated planar X-ray fan beam directed at the sample, with readings being taken on a linear detector array at constant-angle intervals through a full rotation. The resulting data are used to reconstruct a sample cross-section along the slice plane. A collimated planar X-ray fan beam is directed at the sample, and readings are taken on a linear detector array at constantangle intervals through a full rotation. The resulting data are used to reconstruct a sample cross-section along the slice plane.

Morphology and connectivity of micropores

Porosity, which is a measure of the void space present in a material, is represented as a ratio of the volume of void space to total volume (i.e., the void fraction), also expressed as the percentage of void space (in the range of 0%-100%). A variety of methods are available to analyze the porosity of a material, including industrial CT scanning.

The coordination number (CN) of a pore is the number of other pores to which a port directly connects; thus, CN is equivalent to the number of throats through which a pore connects to other pores. Reconstructions of a threedimensional pore network are possible using micro-CT, in contrast to two-dimensional SEM imaging, which can analyze both subgranular and intragranular mineralogical variations, but not the network geometry or connectivity. However, pore network models informed by analyses of both two-dimensional and three-dimensional images at comparable resolutions produce permeability estimates that are in relatively good agreement with observed values (Beckingham et al., 2013).

Fig. 4 presents a conceptual model of the microstructural features of pores and throats. A pore is connected via throats to other pores. The model allows an analysis to be made of the connectivity of the pores, of their aspect ratios, and of their tortuosity. The effective porosity is controlled by percolation clusters, which connect sample surfaces via enhanced flow pathways, and through which water readily penetrates. Fig. 4 also illustrates pore–throat networks in a two-dimensional sketch that shows the medial axis of a pore network, in which the axis of rotational symmetry of a throat (medial axis, MA) is a straight line passing through the center points of adjacent pores.

Fig. 3.Core samples for the micro-CT analysis (Φ = 20 mm), obtained by NX coring (Φ = 54 mm) of samples that show different degrees of weathering.

Fig. 4.Two-dimensional sketch of a pore–throat network showing medial axes (MAs) (red lines), which are axes of rotational symmetry passing through the center points of pores and throats. The lower diagram is adapted from Lindquist (2002).

Samples

We investigated two granite samples, a fresh granite (F-1) and a highly weathered granite (HW-2). For the micro-CT measurements, the granite samples (Φ= 20 mm) were recovered from NX coring (Φ= 54 mm) (Fig. 3).

 

Results and discussion

Internal structures of rocks

Micro-CT reveals the intrinsic patterns of pores and microstructures developed in rocks. When an X-ray bombards a sample, electrons in the elements or molecules absorb its energy and are subsequently scattered, giving rise to various interference effects. Thus, the characteristics of micro-CT images depend on the chemical and mineralogical onstituents of the sample.

As the numbers of electrons in the elements that constitute minerals are large, the interference phenomena associated with mineral samples are stronger than in samples of other types. That is, minerals composed of elements of high atomic number (i.e., with many electrons) show brighter tones than do those of elements of low atomic number, which display darker tones. In particular, the sections of samples that contain pores and fractures, that is, where no matter is present, are rendered as dark areas, as no elements are present in the voids to react with the X-rays. Thus, micro-CT essentially measures the density distribution of the material, which is especially useful for analyzing strain localization in rocks, as well as the nature of interactions between fracture orientations and microstructural fabrics (Yun et al., 2013).

In granite, it is expected that some minerals, such as biotite, hematite, and zircon, will be displayed as bright and high-contrast areas, because these minerals are composed of heavy elements, such as Fe and Zr. In this study, small samples, 3 mm in diameter, were analyzed by micro-CT using continuous imaging from the core sample surface to the interior.

Fig. 5 shows micro-CT images of fresh granite (F-1) obtained at depths of 0.15, 1.05, 1.95, and 2.85 mm from the surface. The granite is composed mostly of quartz and feldspar, whose grain boundaries correspond to fractures and voids. It is evident that small fractures are present even in fresh granite, according to X-ray CT observations. Imaging in the internal parts of the samples shows only small variations in fracture patterns, pores, and textures, except at a depth of ~3 mm, which indicates that the material is nearly homogeneous in terms of its physical properties. No conspicuous fractures or pores are present in the fresh granite, with the exception of small microfractures developed within grains. In a strict sense, no pores are present in the fresh granite.

Fig. 5.Image slices of fresh granite (sample F-1) obtained over an area of 3 mm × 3 mm, at depths of 0.15 mm (a), 1.05 mm (b), 1.95 mm (c), and 2.85 mm (d).

In contrast, micro-CT images of the highly weathered granite (HW-2) at depths of 0.15, 1.05, 1.95, and 2.85 mm, show characteristic distribution patterns of fractures and pores (Fig. 6). Grains of quartz and feldspar are randomly arranged in the rock, and numerous fractures and pores of various sizes are present between adjacent grains. In the internal parts of the sample, the number and distribution of pores and fractures is similar to that observed at the surface, but their concentrations are slightly lower at a depth of ~3 mm. Overall, it is reasonable to conclude that the extent of fracturing and the distribution of pores is fairly constant in the samples, although some variations in mineral assemblages are observed at this scale. Fracturing of the samples is insignificant, as fractures are not developed within the internal parts of grains, nor do they extend through grains; also, the fracture patterns vary with depth. Fractures and pores that are present in the highly weathered granite are commonly developed in feldspar and along the boundaries between grains, as shown in the dark-gray area of Fig. 8. In comparison, fractures within quartz grains are mostly absent, as shown in the brightgray area of Fig. 6. Hence, in granite, the total porosity of the rock gradually increases with an increasing degree of fracturing within feldspar.

Fig. 6.Image slices of highly weathered granite (sample HW-2) obtained over an area of 3 mm × 3 mm, at depths of 0.15 mm (a), 1.05 mm (b), 1.95 mm (c), and 2.85 mm (d). The dark-gray tones represent fractured feldspar and the brightgray tones represent resistant quartz.

Fig. 7.Micro-CT images showing the distributions of pores and fractures in fresh granite (sample F-1), as well as the threedimensional distribution of medial axes, volume data, and pore volume data. (a) Original rock; (b) Mineral grains; (c) Pore spaces; (d) Lines; symbols: grains, pores, isolated pores, respectively.

Fig. 8.Micro-CT images showing the distribution of pores and fractures in highly weathered granite (sample HW-2), as well as the three-dimensional distribution of medial axes, volume data, and pore volume data. (a) Original rock sample; (b) Mineral grains; (c) Pore spaces; (d) Lines; symbols: grains, pores, isolated pores, respectively.

The brightest areas are characteristic of irregular grain boundaries or anhedral grains of iron oxides, which form by precipitation of heavy elements leaching out of mafic minerals during weathering. Subhedral or angular grains displaying the brightest contrasts are resistant accessory minerals, such as zircon, which are commonly present in granite and which retain their own form with little or only slight corrosion. From the micro-CT observations of microtextures, it is evident that alteration of feldspar, which is particularly susceptible to weathering, is the primary cause of the formation of pores in granite, with pore space associated with quartz and heavy minerals being less important. After immobile elements such as Fe, Ti, Zr have been leached out of the heavy minerals, zircon precipitates along fractures or within pores, subsequently affecting the total porosity and fracture patterns; the heavy elements tend to fill up fractures, reducing the total porosity.

Three-dimensional pore patterns

Three-dimensional images of the distributions of pores, throats, and microfractures in the internal parts of the granite samples were reconstructed from micro-CT data. According to the three-dimensional image analysis, the spatial distributions of pore size and frequency are heterogeneous and concentrated locally, and the sizes and spatial frequencies of the throats that connect the pores are variable. Connections between large pores are especially abundant, and the connections (throats) tend to expand and propagate with time.

In fresh granite (F-1), although pores are largely absent, the pores that are present are concentrated locally (Fig. 7). Since minor pores are concentrated in certain regions, it is assumed that they might have resulted from the weathering of minerals susceptible to alteration, or by fracturing associated with failure. In general, fractures in rocks tend to be preferentially developed in certain directions, and, when formed by failure, are systematically arranged. Therefore, it seems clear that most pores are the result of fracturing, in association with slight alterations of minerals susceptible to weathering, as small and minor amounts of pores are distributed locally.

Similarly, highly weathered granite (sample HW-2) contains pores concentrated locally, but the pore density is much higher than that in the fresh granite (Fig. 8). In weathered rocks, mineral assemblages and textures serve as important influences on the formation and development of fractures and pores. It is highly probable that because feldspar and mica are the main constituents of granite, and because these minerals are highly sensitive to weathering (as compared with quartz), the development of pores is noticeable predominantly in areas where these minerals are vulnerable to alteration by weathering. Based on micro-CT images obtained from granite samples in which quartz is dominant, and resistance to weathering is high, the development of pores is negligible, not only at the surface but also in internal regions.

Effective pore size and developmental patterns of throats

Micro-CT provides data on the three-dimensional distribution of pore size and connectivity. Fig. 8 shows the size distribution of pores and fractures within the rock, as well as quantitative information on the connectivity of pores (which allows permeability values to be calculated).

The sizes of effective pores in the granite cores vary in the range of 12-100 μm. The connectivity of pores in the two cores is remarkably different, and is dependent on the degree of weathering. The size frequency distributions of the pores and of the throats that connect the pores in the fresh granite also differ from those in the weathered granite. The radii of effective pores in the fresh granite (F-1) are in the range of 17-75 μm, with most in the range of 17-50 μm (Fig. 9). The radii of throats in the fresh granite are in the range of 3-33 μm, with the highest frequency of occurrence in the range of 5-25 μm. The relationship between the effective radii of pores and CN is slightly linear, despite the low correlation; most CNs are less than 10.

Fig. 9.Analysis of fresh granite (F-1), showing the distribution of the number of pores as a function of pore size (upper), the distribution of the number of throats as a function of throat radius (middle), and the distribution of the coordination number of pores as a function of the effective pore radius (lower).

In the highly weathered granite (HW-2), the radii of effective pores are in the range of 17-105 μm, and are most commonly in the range of 20-80 μm, which is considerably greater than that observed in fresh granite (Fig. 10). The radii of throats are in the range of 5-43 μm, with the highest frequency in the range of 8-30 μm. The relationship between the effective radii of pores and the CN is slightly linear, despite the very low correlation, with most CNs being less than 12. One point shows an exceptionally high CN of ~40 and is considered to be an unusual case.

Fig. 10.Analysis of highly weathered granite (HW-2) showing the distribution of the pore number as a function of pore size (upper), the distribution of the number of throats as a function of throat radius (middle), and the distribution of the coordination number of pores as a function of the effective pore radius (lower).

These observations of the size distributions of effective pores and throats indicate that with an increasing degree of weathering, even in the same rock type, both the size and frequency of effective pores increase. We therefore conclude that the size of throats connecting pores increases with an increasing degree of weathering, which induces fracture propagation in rocks.

Application of X-ray CT to rock engineering

Alteration processes that commonly accompany weathering include dissolution, cementation, and precipitation within rock interiors, which causes changes in pore properties. Rock properties such as porosity and permeability are largely dependent on the interior structures of the rock, especially the characteristics of rock pores. A research tool that can quantify changes in permeability, and thus account for changes in flow network topology, is of particular use in pore-network models that characterize pore size, pore connectivity, and pore-throat size (Beckingham et al., 2013; Mehmani and Prodanovic, 2014). Importantly, underground fluids tend to migrate along irregularly developed fractures in rocks, and flow may occur along permeability gradients.

Information on the connectivity of pores and the data from the three-dimensional imaging of internal structures in rocks using micro-CT is useful for the prediction of the strength of rocks and the development of pores by weathering. For example, a micro-CT analysis of weathered and porous rocks enables valuable information to be obtained on the three-dimensional distributions of pores, fracture properties, and pore connectivity, as well as imaging to be performed of the geometry of internal structures (Ikeda et al., 2000; Ketcham and Carlson, 2001; Wildenschild et al., 2002; Jeong and Takahashi, 2010). More recently, high-resolution micro-CT has demonstrated the ability to resolve details with resolutions as small as a few microns in size, even in high-density materials (Nakashima, et al., 2004; Cnudde and Boone, 2013). High-resolution micro-CT is a non-invasive and nondestructive technique for visualizing the three-dimensional geometries and properties of the internal structures of rocks, as well as for providing digital information useful for elucidating rock characteristics from a rock-engineering perspective. Three-dimensional imaging is essential for informing pore-network models, because such imaging is the only method that can characterize pore-network structures and pore connectivity. Nevertheless, to improve interpretations of three-dimensional scans, the resolution capabilities of three-dimensional imaging should be improved, so as to provide more accurate descriptions of pore-network topologies (Beckingham et al., 2013). While a minimum resolution is crucial for the visualization of pores in rock materials (Cnudde et al., 2006), techniques such as micro-CT can be applied to measure damage to rocks at nano to micron scales, as well as predict the flow paths of groundwater, fluids, and contaminants transported through fractured or porous media. Thus, micro-CT is regarded as sufficiently powerful to potentially capture the characteristic changes in pore arrangement developed in weathered rocks.

 

Concluding remarks

This study analyzed the three-dimensional distribution of micropores and internal structures in fresh and weathered granite using micro-focus X-ray CT. In fresh granite, the pore radius distribution is mostly in the range of 17-50 μm, the throat radius mostly in the range of 5-25 μm, and the CN of pores less than 10, whereas in weathered granite, the pore radius is mostly in the range of 20-80 μm, the throat radius mostly in the range of 8-30 μm, and the CN less than 12. In general, the relationship between pore radius and CN is linear and positive, despite the low correlation. These observations regarding the size distributions of effective pores and throats indicate that with an increasing degree of weathering, even in the same rock type, both the size and frequency of effective pores increase. The size of the throats connecting pores increases with an increasing degree of weathering, which induces fracture propagation in rocks. Micro-CT provides threedimensional image information on the internal pores and fracture structures of rocks, which assists the quantitative assessment of pore connectivity.