marine and petroleum geology - rwth aachen university · 2013-11-18 · a comparative study of...

lable at ScienceDirect

Marine and Petroleum Geology 49 (2014) 143e161

Contents lists avai

Marine and Petroleum Geology

journal homepage: www.elsevier .com/locate/marpetgeo

A comparative study of representative 2D microstructures in Shalyand Sandy facies of Opalinus Clay (Mont Terri, Switzerland) inferredform BIB-SEM and MIP methods

M.E. Houben*,1, G. Desbois, J.L. UraiStructural Geology, Tectonics and Geomechanics, RWTH Aachen University, Lochnerstrasse 4-20, 52056 Aachen, Germany

a r t i c l e i n f o

Article history:Received 20 February 2013Received in revised form18 October 2013Accepted 22 October 2013Available online 30 October 2013

Keywords:Opalinus ClayBIB millingSEM-imagingPore morphologyPore size distributionMercury intrusion porosimetry

* Corresponding author. Utrecht University, DepBudapestlaan 4, Postbus 80021, 3508 TA Utrecht, The2531409.

E-mail address: [email protected] (M.E. Houben)1 Now at: Structural Geology and Tectonics/HPT-La

pestlaan 4, Postbus 80021, 3508 TA Utrecht, The Neth

0264-8172/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.marpetgeo.2013.10.009

a b s t r a c t

A combination of Broad-Ion-Beam (BIB) polishing and Scanning Electron Microscopy (SEM) has beenused to study qualitatively and quantitatively the microstructure of Opalinus Clay in 2D. High quality 2Dcross-sections (ca. 1 mm2), belonging to the Shaly and Sandy facies of Opalinus Clay, were investigateddown to the nanometre scale. In addition Mercury Intrusion Porosimetry (MIP) and X-Ray powderDiffraction experiments have been used to extend characterization of the microstructure to the mmecmscale on bulk volume sample material. Interestingly, both end-member samples of the Opalinus Clayshow qualitatively similar mineralogy and pore characteristics as well as a comparable pore size dis-tribution and pore morphology within the different mineral phases and mineral aggregates. Differencesbetween the facies are mainly due to variations in mineral size and mineral amount present in thealternating layers of the different facies. Six different porous mineral phases have been identified and thepores have been subdivided into ten different pore types. Pores visible in the SEM images are mostabundant in the clay matrix and these seem to follow a power law distribution with a power lawexponent of ca. 2.25 independent of the sample location. Furthermore, all common mineral grains showcharacteristic porosity, pore shape and pore size distribution in 2D and are proposed to be considered aselementary building blocks for Opalinus Clay. Combined these homogeneous elementary building blocksmake up the heterogeneous fabric of the different facies of Opalinus Clay. Based on extrapolation of thepower law size distribution in the clay matrix below SEM resolution results in a porosity of 10e25% forclay rich layers (60e90% of clay matrix), whereas sand and carbonate layers show an extrapolatedporosity of 6e14%. These extrapolated porosities are in agreement with water-loss and physical porositymeasurements performed on bulk material of comparable samples.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Shale formations are known for their low permeability andexcellent sealing properties (Bossart and Thury, 2007). This is whyin the Mont Terri Rock Laboratory (Switzerland) the Opalinus Clay(a Mesozoic shale formation) is investigated thoroughly as apossible host rock for long-term geological disposal of radioactivewaste (Bossart and Thury, 2007). Opalinus Clay is a fine-grainedsedimentary rock, which comprises a sequence of dark grey,calcareous, mica rich, silty mudstones with up to 20 cm thick

artment of Earth Sciences,Netherlands. Tel.: þ31 (0)30

.b, Utrecht University, Buda-erlands.

All rights reserved.

interbeds of calcareous cemented fine-grained sandstone/siltstonelenses and laminae as well as sandy marls, biodetritic limestones,pyrite-, siderite- and chalk concretions, and some ammonites(NAGRA, 2002).

Characterization of Opalinus Clay, in a lab or in-situ at the MontTerri rock laboratory, is related to the investigated scale. On theregional scale the lateral variability of facies and lithology is low(NAGRA, 2002), with three layered lithological facies (Pearson et al.,2003): (1) Shaly facies - a dark grey silty calcerous shale andargillaceous marl in the lower half of the sequence, (2) Sandy-Carbonate rich facies e a grey sandy and argillaceous limestonein the middle of the sequence which is almost absent in north-eastern Switzerland (Nussbaum et al., 2011) and (3) Sandy faciese silty to sandy marls with sandstone lenses cemented with car-bonate in the upper part. On the millimetre to decimetre scale,there is visible bedding due to sedimentation of the clay minerals;the bedding has become more pronounced by burial and

Delta:1_given name

Delta:1_surname

Delta:1_given name

mailto:[email protected]

http://crossmark.crossref.org/dialog/?doi=10.1016/j.marpetgeo.2013.10.009&domain=pdf

www.sciencedirect.com/science/journal/02648172

http://www.elsevier.com/locate/marpetgeo

http://dx.doi.org/10.1016/j.marpetgeo.2013.10.009



Table 1Investigated drill cores.

Facies Sample Mont TERRINiche

Corediameter(cm)

Distance fromthe gallery wall(m)

No. of BIB-polishedcross-sections

Sandy BWS-H TT-Niche 15 0.4 9Shaly BDR1_OC DR-Niche 30 5.4 5Shaly BCS-2 Gallery 98 10 2.7 2

M.E. Houben et al. / Marine and Petroleum Geology 49 (2014) 143e161144

compaction (Wenk et al., 2008). The bedding causes anisotropy inthe hydrological, transport related and rock mechanical propertiesof the rock (NAGRA, 2002; van Loon et al., 2004; Wenk et al., 2008)and from the anisotropy of the shale’s microstructure it is notsurprising that pores in the Shaly facies of Opalinus Clay are alsoanisotropic in 3D and show a preferred orientation along thebedding (Keller et al., 2011). Mineral clasts and biogenetic mineralsare mainly present in the size range from 10 mm to 1 mm (NAGRA,2002). Qualitatively themineralogical composition of Opalinus Clayis similar in the different facies present at the Mont Terri rocklaboratory, common minerals encountered are: calcite, dolomite/ankerite, siderite, quartz, albite/plagioclase, K-feldspar, illite, illite/smectite, chlorite, kaolinite, pyrite and organic carbon (Thury andBossart, 1999).

Due to the fine-grained nature of claystones direct investigationof microstructures below 1 mm is still challenging. Distribution ofpores in claystones has up to now mainly been illustrated asschematic diagrams/drawings (Yven et al., 2007; Ortega et al.,2009). These schematic pore models show that a percolating porenetwork exists in the clay matrix interconnected with minerals andbioclasts present as single grains or clusters. Recent use of ion beammilling tools enables the production of smooth damage free sur-faces even for fine-grained heterogeneous geological materials(e.g.: Holzer et al., 2007, 2010; Desbois et al., 2008, 2009, 2010a,2010b, 2011b; de Winter et al., 2009; Loucks et al., 2009; Holzerand Cantoni, 2011; Heath et al., 2011; Keller et al., 2011;Schneider et al., 2011; Klaver et al., 2012), suitable for investiga-tion of the microstructure below the mm scale. Typical Gaþ FocusedIon Beam (FIB) and Broad Ion Beam (BIB) polishing in combinationwith Scanning Electron Microscopy (SEM) allows imaging withresolutions that are about two orders of magnitude higher than X-ray microtomography, where FIB-/BIB-SEM achieve pixel sizesdown to a few nm’s and X-ray microtomography normally achievespixel sizes of a few mm’s (Holzer and Cantoni, 2011). Gaþ FIB-SEM ismainly used to image 3Dmicrostructures via serial cross-sectioningof volumes which are ca. 10 � 10 � 10 mm in size (e.g.: Holzer et al.,2006, 2010; Desbois et al., 2009; Curtis et al., 2010; Keller et al.,2011; Bera et al., 2012) giving information about the 3D porenetwork present in claystones and hence information about thepercolating network present (Keller et al., 2013b). The BIB-SEMmethod allows investigation of a polished area up to 1 mm2

(Desbois et al., 2009; Loucks et al., 2009; Schneider et al., 2011;Klaver et al., 2012; Houben et al., 2013), which is typically largerthan the representative elementary area (REA) for quantification ofthe mineralogy and porosity in claystones and other fine-grainedmaterials, where previous studies show that the REA for fine-grained claystones is smaller than 150 � 150 mm (140 � 140 mmfor Posidonia Shale, Klaver et al., 2012; 100 � 100 mm for the Shalyfacies of Opalinus Clay, Houben et al., 2013). Limitation of bothtechniques is the maximum area/volume of the polished sections,the number of sequential slices made and the resolution used toimage the polished sections (Holzer and Cantoni, 2011). Moreover,investigation of one sample (typically< 1mm3) cannot be upscaledto reservoir size due to the small scale stratigraphic and lateralheterogeneity of shale strata (Chalmers et al., 2012; Houben et al.,2013).For a full microstructural clay model from the nm-scaleideally a combination of m-CT, BIB-SEM, FIB-SEM and TEM shouldbe used. Keller et al. (2011) investigated the microstructure of threeOpalinus Clay (Shaly facies) samples using Gaþ FIB-serial cross-section, determining the spatial orientation of pore space, porepath orientation, pore path tortuosity and pore path length.Furthermore, Keller et al. (2013a) used a STEM-, FIB- and X-raytomography approach to characterize the pore structure in the fine-grained clay matrix of the Shaly facies of Opalinus Clay. Based onmicrostructural observations, using the BIB-SEM images, Houben

et al. (2013) concluded that the microstructure of the Shaly faciesof Opalinus can be modelled based on predictable homogeneousporous and non-porous regions with comparable porecharacteristics.

In order to research differences and consistencies in themicrostructure of the Opalinus Clay, in this contribution pore sizedistribution, pore shape and mineralogy are investigated for thetwo end-member facies of Opalinus Clay (Shaly and Sandy facies)using the BIB-SEM method. This is validated with optical thin-sections investigated with transmitted light, Mercury IntrusionPorosimetry (MIP) and X-Ray powder Diffraction (XRD) measure-ments. The present 2D BIB-SEM study can in the future be com-bined with local 3D FIB-SEM data of similar samples in order tolearn something about pore connectivity in 3D (Keller et al., 2013a,2013b).

2. Material and methods

2.1. Geological setting and samples

The underground Mont Terri Rock Laboratory is situated ingalleries and niches beside the security gallery of a motorwaytunnel (A16) in north-western Switzerland near the town of St-Ursanne, and is cross-cutting a ca. 240 m section of Opalinus Claywhich dips towards the south-east (Bossart and Thury, 2008). Rocksin the Jura mountains are of Triassic to early Cretaceous age andwere deposited mostly in marine environments with generallyuniform subsidence (Pearson et al., 2003). They comprise lime-stones, evaporitic rocks, and marls/shales. The Mont Terri tunnel ispresently located ca. 300 m below the surface (Thury and Bossart,1999) and passes through all lithological facies of Opalinus Clay.The samples studied in this contribution originate from differentboreholes along galleries and niches in the Rock Laboratory and arelocated in both the Shaly and Sandy facies (Table 1). The cores weredrilled with a tungsten carbide drill bit (diameters ranging from10 cm to 30 cm for the Shaly facies samples and a diameter of 15 cmfor the Sandy facies samples) cooled with dried air. Sandy faciessamples (BWS-H) investigated originate from two different bore-holes, both situated in the TT-niche (Mont Terri Rock Laboratory)and both are cored parallel to the bedding about 40 cm into thegallery wall. One Shaly facies sample (BCS-2) originates from gal-lery 98, is cored perpendicular to the bedding (Mont Terri RockLaboratory) about 2.70 m into the wall. The second Shaly faciessample (BDR1) was drilled in the DR-niche perpendicular to thebedding (Mont Terri Rock Laboratory) at a distance of ca. 5.4 m intothe gallery wall. After drilling, the cores were stored immediatelyunder vacuum in aluminium barrier foil to prevent them fromdehydration.

2.2. Optical thin sections

Samples were unpacked from the aluminium barrier foil anddried at room temperature. Ultra-thin sections were produced (W.Tschudin, Geoprep, Basel, Switzerland; e.g.: Stipp et al., 2006; Milkeet al., 2009) by impregnating the samples with epoxy, and after

Figure 1. A. Cross-section of the BWS-H drill core, Sandy facies Opalinus Clay. In the drill core one can distinguish four sub-areas, a dark grey fine-grained clay layer, a light colouredcoarse grained sand layer, a pinkish/yellowish coarse grained carbonate lens and a mixture of clay, sand and carbonate. B. Cross-section of the BCS-2 drill core, Shaly facies OpalinusClay. On the drill core scale there are no visible heterogeneities. C. Thin section of the Sandy facies showing alternating clay rich and quartz rich layers. D. Thin section of the Shalyfacies showing some bigger minerals (as calcite, quartz and pyrite) floating in the clay matrix. E. Clay layer present in the Sandy facies (thin section). F. Sandy facies sand layer (thinsection). G. Shaly facies clay layer (thin section) wherein quartz, calcite, pyrite and fossils are visible. H. BSE image of a BIB polished cross-section of a clay layer in the Sandy facies. I.BSE image of a BIB polished cross-section of a sand layer in the Sandy facies. J. BSE image of a BIB polished cross-section of the Shaly facies showing a similar calcite fossil as in thethin section (bottom right). K. SE-image of the clay matrix in the clay layer of the Sandy facies. L. Clay present in the Sand layer of the Sandy facies. M. Clay matrix of the Shaly facies.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

M.E. Houben et al. / Marine and Petroleum Geology 49 (2014) 143e161 145

hardening, the samples were polished on both sides down to athickness<10 mm. The ultra-thin sections were investigated using aZeiss Axioscope optical microscope under transmitted light using amagnification of 10� (pixel size of 0.7 mm), using both plain-polarized and cross-polarized light.

2.3. SEM imaging of broad-ion-beam milled surfaces

2.3.1. Sample preparationThe samples were unpacked from the aluminium barrier foil,

where after they were cut dry into subsamples (1 � 0.5 � 0.5 cm)suitable for BIB polishing, using a 0.3 mm diamond saw at low

speed (3 rev./sec.). Due to the heterogeneity of the Sandy faciesvisible at the mmecm scale various subsamples were selected to beBIB polished based on mineralogical composition. Regions of in-terest were based on colour and texture visible at the drill core scaleand details visible in the optical thin sections (Figure 1AeG.). TheShaly facies is homogeneous on the mmecm scale, and subsampleswere sampled from the middle of the drill core as far away from thedrill-core wall as possible. Samples were carefully dried at roomtemperature in a closed 60 mL container with a small opening inthe cover so that the relative humidity and the temperature in thecontainer was the same as in the lab (respectively 35% and 22 �C). Inorder to optimize the use of the BIB cross-section polisher, the dried


samples were pre-polished dry using carbide paper (down to a gritsize of 1200). Pre-polishing of the samples causes the surfaceroughness to be reduced to ca. 20 mm, which results in a betterbroad-ion-beam polished cross-section by minimizing thecurtaining effect (Desbois et al., 2011a; Klaver et al., 2012; Houbenet al., 2013). The cross-section polisher removes about 100 mm fromthe outermost edge of the sample and removes pre-polishingdamage. Cross-sections were all cut perpendicular to the beddingusing Ar-ion beam milling with the JEOL (SM 09010) cross-sectionpolisher. The cross-section polisher (JEOL, SM 09010) producesdamage-free polished surfaces up to ca. 1 mm2, in 6.5 h (6 kV,150 mA). In this contribution 16 BIB polished cross-sections areinvestigated. The polished cross-sections were produced perpen-dicular to the bedding, and nine of the sections originate from theSandy facies and seven were located in the Shaly facies of OpalinusClay (Table 1).

2.3.2. BIB-SEM imaging of microstructureCross-sections produced with the BIB cross-section polisher

were imaged using a Scanning Electron Microscope (SEM, ZEISSsupra 55). This method allows direct imaging of the clay fabric andporosity down to the nm scale on about 1mm2 areas (Desbois et al.,2009, 2011b; Loucks et al., 2009, 2012; Schneider et al., 2011; Klaveret al., 2012; Houben et al., 2013). Resolution of a SEM is defined asthe minimum spacing at which two features can be recognized asseparated objects (Goldstein et al., 2007). For this study that meansthat using the secondary electron (SE) detector objects smaller than45 nm (3 pixels) in diameter could not be detected when an ac-celeration voltage of 5 kV and a magnification of 20,000 � wasused. The Back Scattered Electron (BSE) detector was used at amagnification of 2000 � (equivalent pixel size is 147 nm) and anacceleration voltage of 20 kV to map mineral phase contrast;mineralogy was checked by Energy-dispersive X-ray spectroscopy(EDX) also using an acceleration voltage of 20 kV. Before imaging,the samples were sputter coated with gold (Goldstein et al., 2007).Theoretically a REA is an area large enough to be statisticallyrepresentative of the composite and should include all micro-structural heterogeneities present (Kanit et al., 2003). For a naturalmaterial like a claystone this area should then be infinitesimal largeto incorporate all microstructural heterogeneities present (Kelleret al., 2013b). Practically we defined the BIB-SEM REA as an arealarge enough to represent the mineralogy and porosity in one1mm2 BIB polished cross-sectionwithin a 10% error range. The BIB-SEM REA for mineralogy was defined by using the point countingmethod (Kameda et al., 2006; Klaver et al., 2012; Houben et al.,2013) based on segmented BSE-images. A grid consisting ofdifferent square box sizes was placed over a segmented BSE imageand for every box size themineralogy wasmeasured. Theminimum(box sized) area of an image that displays a mineralogical contentfluctuating less than 10% on average from the mineralogy in theprevious box is then called the REA. For this we assume that theporosity in the clay is linked to mineralogy and thus the mineraldistribution (see also Houben et al., 2013), and the REA estimatedby the box counting method is an approximation of the REA forporosity. Secondly, the REA for porosity was verified using a sta-tistically more rigorous method based on the variance ofsegmented porosity, which is also able to define the error of thechosen REA (Kanit et al., 2003; Keller et al., 2013b). In order toimage the REA with sufficient resolution to detect pores, SE- mi-crographs were combined into one high resolution image (>100million pixels) using Kolor Autopano giga 2.0 software (Kolor, 2008;Houben et al., 2013). Imaging of the porosity was done using the SE-detector at a low acceleration voltage of 5 kV and a working dis-tance of 6e8 mm. Magnifications used were between 10,000� and50,000 � .

2.3.3. SEM image processingThe only visible topography after using the BIB cross-section

polisher was caused by the pores forming openings in the pol-ished surface. Visible porosity in the micrographs was segmentedusing a MATLAB routine (2010b, The MathWorks, 2010) based on acombination of thresholding and edge detection (Gonzalez et al.,2009; Houben et al., 2013). Thresholding (Gonzalez et al., 2009)was based on the image histogram. SEM image background shows anormal grey scale distribution around one value which depends onthe settings in the microscope and the threshold should be set justbelow this (Houben et al., 2013). Performing edge detection on thethresholded pictures provided a better representation of the poresin the original SEM image. The segmented images were loaded intoARCMAP 10 (ESRI, 2008) together with the original SEM image sothat the segmented image could be compared to the original imageand manually corrected when necessary. Minerals were segmentedmanually, based on their characteristic grey values in the BSE im-ages using ARCMAP 10 software (ESRI, 2008). Mineral compositionsencountered were measured with EDX point measurements andEDX maps to assign a mineral name to each BSE grey-value. Aftersegmentation of the minerals, all pores were classified according tothe mineral phase in which they occured.

2.4. Bulk chemical analysis by X-ray powder diffraction

To characterize the mineral content, X-ray powder diffraction(XRD) analysis was performed on three samples of the Sandy andone sample of the Shaly facies. The samples were unpacked fromthe aluminium barrier foil, where after they were dried using thesame method as was performed on the BIB-SEM samples. Threedifferent samples were selected from the Sandy facies based ondifferent colour and textures visible in the drill core (Figure 1A).These included a clay-rich, a quartz-rich and a calcite-rich sample.The Shaly facies drill core did not showmineralogical differences atthe drill core scale (Figure 1B), hence XRD-measurements wereperformed on one randomly chosen sample. Twenty grams ofsample material was first turned into powder, where after themineral content was measured using the Bruker D5000 at theGeological Institute of the RWTH Aachen University (GIA). Rietveldanalysis (TOPAS; Bruker AXS) was performed for mineral quanti-fication (Toraya, 2000; Kahle et al., 2002; Scrivener et al., 2004).Rietveld analysis is based on iteratively comparing X-ray patternswith a pattern simulated based on the presumed amounts of crystalparameters and equipment parameters of a mixture of knownphases. This means that when not correctly used, the large amountof fitting parameters can cause a good but mineralogical unrealisticfit (Scrivener et al., 2004). Because SEM measurements were per-formed on similar samples the expected mineral composition ofOpalinus Clay was known from BSE and EDX measurements. Thiswas used as input for the Rietveld analysis.

2.5. Mercury intrusion porosimetry

To characterize the porosity and pore size distribution on bulksample material mercury intrusion porosimetry (MIP) was con-ducted (Abell et al., 1998; Galle, 2001; Hildenbrand and Urai, 2003;Klaver et al., 2012; Houben et al., 2013), where mercury acts as anon-wetting liquid and intrudes progressively smaller pores withincreasing pressure. Assuming that the porous network can berepresented by a bundle of non-intersecting tubes (Howard,1991) apore size distribution curve can be obtained from the intrudedvolume of mercury at each pressure step. Pressure differences canbe converted to pore throat diameters using the Washburn equa-tion (Washburn, 1921).


Shaly facies samples were broken off from the drill core,whereas for the Sandy facies specific locations were identified withdifferent mineralogies and samples were drilled out of the drill coreusing a 0.3 mm diamond saw at low speed (3 rev./sec.). All sampleswere first dried at 60 �C, where after the samples were placed in asample cup filled with mercury in a pressure vessel. The pressurewas stepwise increased up to 4.13$108 Pa, which corresponds to apore diameter of 3 nm. To check whether the mercury penetratedthe sample, samples were weighted before and after the MIP ex-periments. The pore throat (Sth) size distribution was extractedfrom the MIP data using the Washburn equation to calculate thepore throat diameters. These were then turned into disk shapedareas and the matching infiltrated mercury volumes were con-verted to number of pores (Ni).

3. Results

3.1. Mineralogy and mineral fabric of Opalinus Clay

Differences inmineral fabric on themmecm scale are illustratedin Figure 1AeG, which shows a Shaly and a Sandy facies drill coreand the corresponding thin sections. Figure 1AeG illustrates themmecm scale heterogeneity of the Sandy facies, whereas the Shalyfacies is homogeneous at this scale. The mineralogy of the Sandyfacies drill-core on the mmecm scale is mainly based on colour andgrain size, where a white coarse-grained layer corresponds to aquartz rich layer, the light yellow/pinkish coarse grained lensescorrespond to carbonate rich lenses and the black areas are finegrained clay rich regions (Figure 1A). This is supported by the thinsections, where the thin sections are imaged using cross-polarizedlight (XPL) (Figure 1CeG). Minerals like quartz (Light grey to blackmineral grains), calcite (High order coloured mineral grains) andpyrite (Black round grains) in the different layers of Sandy faciescan be identified. In the Shaly facies thin section (mm-mm scale,Figure 1D, G), larger calcite fossils (bivalve, Bossart and Thury,2008), calcite grains, quartz grains and pyrite grains can be differ-entiated from the clay matrix.

The overall size of themineral grains present in the SEMmosaicsin the Carbonate lens, Sand layer and ‘Clay þ Sand’ layer is larger(average mineral size of 250 mm2; Figure 2A,B) than the size of theminerals present in the Clay-rich layers (average mineral size of50 mm2; Figure 2C, D). The microfabrics of the different lithologicalfacies vary due to different percentages of certain mineral grainsand their size. Figure 2 shows that the mineral grains are all‘floating’ within a clay matrix, with a grain size < 2 mm. Themicrofabric of all the clay layers (Sandy and Shaly facies) appearssimilar (Figure 2DeF), although XRD measurements show that thequartz content is higher in clay layers of the Sandy facies whereasthe calcite content is higher in the clay layers of the Shaly facies.

Minerals identified in all the BIB polished cross-sections of theOpalinus Clay facies are qualitatively similar (Figure 1HeM, Fig. 2).The claystone mainly comprises: quartz, calcite, siderite, mica, py-rite and a clay matrix. Furthermore, some feldspar and organicmatter was encountered in the BIB polished cross-sections as wellas insignificant amounts of minerals like apatite and zircon.Quantitatively, on the other hand, the percentage of a certainmineral per SEM mosaic differs; the mineralogy of the differentpolished cross-sections is summarized in Table 2. XRD measure-ments aimed to measure the mineralogy in a sand, clay and car-bonate layer in the Sandy facies and a clay layer in the Shaly faciesshow comparable mineralogical differences as found in the SEMmosaics (Table 3). Results presented in Tables 2 and 3 show thatclay layers, Sandy and Shaly facies, appear to have ca. 56e57% clayminerals and a clay matrix making up between 65 and 85% of thesample. The sand layer (Sandy facies) only features 14.4% clay

minerals and 32% of the SEM images is classified as clay matrix.Results for the carbonate layer show 8.4% clay minerals and 34%clay matrix in the SEM images. The investigated samples from theSand layer (Sandy facies) have a quartz content of around 50% fromboth XRD and BIB-SEM measurements, and in the carbonate layerthe calcite content is >60% of the sample (XRD and BIB-SEMmeasurements).

3.2. Representative elementary area

The box counting method (Kameda et al., 2006; Klaver et al.,2012; Houben et al., 2013) was used to determine a first estima-tion of the representative elementary area (REA; Figure 3A) formineralogy present in the different cross-sections. The mineralphases taken into account for the box counting method are: pyrite,mica, siderite, calcite, quartz, feldspar, fossil shell and clay matrix.We started with a box size of 100 � 100 pixels and increased thiswith 100 pixels in x and y for every following box. For every box weidentified the percentage of themineral phases present as shown inFigure 3A, B. For this particular sample, Opalinus Clay Shaly faciesBDR, the REA for mineralogy is on average 148 � 148 mm (Table 4),where results do not depend on the chosen starting point for thebox counting method (Houben et al., 2013). Furthermore, to illus-trate the range of REAs found in different BIB-SEM samples of thesame drill core, minimum and maximum values are shown inTable 4. The mineralogical REAs for the other investigated samplesare also presented in Table 4. The REA for porosity was also iden-tified (Figure 3C, D), using an expression for the smallest area (Ar)with a given relative error (er), for the truemeanmeasured porosityvalue Mp and N realizations adapted from Kanit et al. (2003) andKeller et al. (2013b):

ArðN; erÞ ¼ �4�1�Mp

�A3

��hN Mpe2r

i(1)

where A3 (i.e. integral range) for the porosity was approximated bycomputing the variance for measured porosities in both x and ydirections (see also Keller et al., 2013b, Fig. 3). The difference be-tween the variance in x and y direction is expressed as theanisotropy factor which is really close to one for all samples(Table 4). For a relative error of 10% and N ¼ 1 the REA for porosityin sample BDR is on average 131 � 131 mm (Table 4, Fig. 3). Theaverage REA for porosity for the other investigated samples can befound in Table 4. The REAs of all the cross-sections show that itdepends on the facies, and more specifically the layer within thatfacies, what the extent of the REA is. Because in the Sandy facies,and mostly in the sand and carbonate layers, the mineral grains arelarger, making the REA for these layers also larger than the REA forthe clay rich layers (Table 4). A BIB-polished cross-section is in theorder of 1 mm2, meaning that the area is large enough to be calledrepresentative for mineralogy and porosity for all samples inves-tigated. Based on the box counting method, the minimum areainvestigated (accepting a variation in mineralogy of 10%) should bein the order of 250 � 250 mm for the Sandy facies samples, whereinthe mosaics are imaged at a magnification of 10,000 � . For theShaly facies samples, the minimum area of investigation should bein the order of 180 � 180 mm where the mosaics are imaged at amagnification of 20,000 � . Comparison of REAs for porosity basedon the covariance analysis with corresponding REAs based on boxcounting method indicates that REAs based on the box countingmethod are a good estimation of the REA for porosity accepting arelative error of 10% (Table 4). A lower magnification of10,000 � was used to image the Sandy facies samples due to thefact that otherwise >700 single images had to be made to enable

Figure 2. BSE micrographs of the Shaly and Sandy facies of Opalinus Clay, illustrating the differences in microstructure in the different layers per facies. A. Sand layer of the Sandyfacies drill core, where the image is mainly made up of quartz grains and some calcite, siderite and feldspar minerals, with the clay matrix present in between the minerals. B. A sandand clay mixture from the Sandy facies, where the mineral size of the calcite and quartz minerals is similar as found in the sand layer, but ca. 50% of the segmented image consists ofclay matrix. C. Sandy facies carbonate lens, mainly consists of calcite with a few siderite and quartz grains and the clay matrix in between the mineral grains. D. Clay layer as presentin the Sandy facies, where more than 75% of the micrograph exists of the clay matrix with some quartz, calcite, siderite, pyrite and mica minerals ‘floating’ in the clay matrix. E. Shalyfacies BDR drill core showing that quartz, calcite, siderite, pyrite and mica grains are surrounded by the clay matrix. F. Image of the microstructure present in the BCS drill-core(Shaly facies), where quartz, calcite, siderite, pyrite and mica minerals and fossils are embedded within the clay matrix.


that one mosaic covered the REA (realistic limit is around 500images; Houben et al., 2013).

3.3. Minerals featuring characteristic porosity

Visible porosity in the SEM mosaics (pixel size between 15 and30 nm) in all studied polished cross-sections ranges from 0.7% to3.1% (Table 5). The majority of the over 700,000 pores segmentedare present in the clay matrix. Six mineral phases were identifiedthat also display characteristic porosity (Table 3, Fig. 4). Based on

characteristic pore morphologies visible in SEM images Desboiset al. (2009) and Heath et al. (2011) classified 7 different poretypes, which are either intergranular pores or intragranular pores(Loucks et al., 2012).

Intergranular pores present in the clay matrix (Pore Types Ie III,Desbois et al., 2009) account for 50% or more of the visible porosityin all investigated samples (Table 5, Figure 4H), independent of thelayer or facies investigated. Average equivalent pore radius (req) forthe pores present in the clay matrix is 52 nm. Although the pores inthe clay matrix are most abundant, the average visible porosity in

Table 2Mineral content of the Shaly and Sandy facies per investigated sample in the segmented SEM images. Mineralogy is based on EDXmeasurements and Back-Scattered-Electronimages.

Sandy facies Shaly facies

BWS BDR BCS2

Clay Clay þ Sand Sand Clay þ Sand þCarbonate

Carbonate Clay þ Carbonate Clay Clay

Mosaic area (mmxmm) 250 320 275 300 320 320 190 180Clay matrix 85 45 32 44 34 33 67 85Quartz 9 35 42 28 2 8 15 4Calcite 3 12 13 20 62 52 12 3Mica 1 1 1 3 e e 1 4Pyrite 0.5 1 1 1 e 1 1 1Siderite 0.5 1 2 1 2 1 1 2Feldspar e 1 8 2 e e 1 e

Other 1 4 1 1 e 5 2 1


the clay matrix of the clay-rich layers is about 1% and 2.5e9% in theclay matrix of the sand and carbonate layers/lenses (Table 5). Microcracks (intergranular, req ¼ 88 nm) were identified in almost allpolished cross-sections (Type V, Heath et al., 2011, Figure 4G). Theyhave a high aspect ratio, rough pore edges and sharp crack tips atboth ends. Microcracks contribute 12e28% to the visible porosity inboth facies with the exception of the carbonate and sand-richlayers, where only 0e2.5% of the pores are characterized ascracks (Table 5). Intergranular pores in framboidal pyrite(req ¼ 66 nm, Figure 4B) were present, where the single non-porous, polygonal pyrite grains are loosely packed forming aframboidal pyrite grain (Pore Type VII, Heath et al., 2011). Due touneven packing of the irregular grains pyrite aggregate porosityvaries between 1.5 and 15% (Table 5). Pores in mica are mainlypresent between mica lamellae, making these pores very elongatedwith smooth pore walls, and regularly triangular (req ¼ 59 nm,Figure 4E). The pores are mostly situated where the lamellae arebent (also observed by White and Johnston, 1981). The porositywithin mica grains ranges from 0.4 to 1.8% (Table 5). Pores in quartz(Figure 4A) are disk shaped intragranular pores within singlecrystals. Quartz minerals have a visible porosity<<1% (req¼ 52 nm,Table 5). Calcite pores are intragranular (req ¼ 52 nm, Figure 4C),circular to elongated and the minerals show on average a porosityof about 0.2% (Table 5). Siderite minerals exhibit a characteristicintragranular porosity of 5e18% (req¼ 62 nm, Table 5), the pores areelongated to roughly circular with jagged edges and lobate featureson the inner pore wall (Figure 4D). In the Shaly facies fragments ofcalcareous half-moon shaped fossils (req ¼ 63 nm; Schizospaerella,Pittet and Mattiolo, 2001; Mattiolo and Pittet, 2002) with a distinctporosity are present (Figure 4F). The pores have a rugged shapewhere they intersect with the polished cross-section and smoothpore walls. The fossils display a visible porosity in the range of 7e15% (Table 5).

Table 3Mineralogy per investigated sample inferred from X-ray powder Diffractionmeasurements.

Sandy facies Shaly facies

BWS BCS

Clay Sand Carbonate Clay

Calcite 5.1 27.8 69.2 18.6Quartz 31.4 53.2 11.7 17.6Ankerite/Dolomite 0.4 2.4 2.0 0.5Siderite 0.9 0.0 7.3 2.2Pyrite 3.0 0.2 0.7 1.4Feldspar 3.2 2.0 0.6 2.5Clay minerals 56.0 14.4 8.4 57.2

3.4. Pore shape factors

Pore shape can be described by a number of different mathe-matical shape factors (Connell et al., 1999). In this contribution, allBIB cross-sections were cut perpendicular to the bedding and asshown by Houben et al. (2013) pores show similar size distribution/shape independent of the orientation of the 2D section perpen-dicular to the bedding. Due to the imaging method used theseshape factors could only be determined for 2D sections and 3D FIBserial cross-sectioning is needed when one would like to investi-gate the pore shape in 3D (Keller et al., 2011, 2013a, 2013b). For allpores segmented the inverse of the Axial ratio [W/L] and squaredcircularity [4pA/P2] were calculated, where; W ¼ Pore width,L ¼ Pore length, A ¼ Pore area, P ¼ Pore perimeter. Axial ratio andsquared circularity are chosen to describe pore shape because of thefact that the axial ratio describes pore elongation whereas circu-larity is a measure of compactness of a shape incorporating surfaceroughness.

To verify the dependency of the shape factors on the magnifi-cation ca. 30 pores were traced manually in pictures taken at amagnification of 10,000 � , 20,000 � , 25,000 � , 30,000 � and50,000 � (Fig. 5). Table 6 summarizes the dependency of themagnification on the pore area, perimeter, length and width, wherethe magnification of 10,000 � is used as a reference (for perimeterof magnification X e.g.: Px/P10000). The table shows that there is nocritical difference (based on the standard error) regarding the de-pendency of perimeter, area, length andwidth on themagnificationused in this calibration. On average the pore shape/area does notchange significantly using anymagnification between 10,000� and50,000� . Results on the pore shape per mineral/mineral aggregateare plotted in Figure 6. The data density points (the darker thecolour the more data points are represented) of this bivariate his-togram are plotted in 2D using 20 � 20 bins. The pore shape cal-culations are based on more than 700,000 segmented pores (mostof them present in the clay matrix) and showsW/L ratio versus thecircularity for all segmented pores per facies and per mineral/mineral aggregate. Cracks and mica have clearly a more elongatedshape than pores associated with other minerals/mineral aggre-gates. Pores in the clay matrix show the largest range from elon-gated to rounded pores, whereas disk shaped pores are mostabundant in calcite, quartz and siderite minerals and pyrite mineralaggregates.

3.5. Clay matrix porosity

Seven BIB polished cross-sections of the Shaly facies and nineBIB-polished cross-sections from two drill cores of the Sandy facieswere investigated (Table 1). The smallest pores visible were present

Figure 3. A. Showing a segmented mineralogy SEM-image of one of the subsamples made from the BDR drill core (Shaly facies) with the grid used for the box counting method.From the solid boxes onwards the imaged area is called representative due to the fact that the change in mineralogy from here onwards is less than 10%. B. Table showing themeasure mineralogy per box. The mineralogy does not vary more than 10% anymore after passing box 10, hence this is interpreted to be the REA for this clay layer. C. Graph showingthe variance along x and y in a segmented porosity SEM-image from the BDR drill core, where the variance along x and y is similar with the exception of some edge effect due to thefact that the image finite (180 mm in length to be exact). D. Showing the relative error vs. the length of the REA (LREA; when the REA is squared) based on the variance of thesegmented porosity map.


in the clay matrix. In terms of frequency, the pores in the claymatrix were most abundant in number, accounting in all mosaicsfor ca. 50e90% of the visible porosity (Table 5). Clay rich layers inboth the Sandy and Shaly facies show a pronounced pore orienta-tion along the bedding (Figure 7A, B; also found by Keller et al.,2011; Houben et al., 2013 for the Shaly facies of Opalinus Clay).This preferred orientation along the bedding is poor for the pores inthe clay matrix of sand layers and orientation of pores in the claymatrix of the carbonate lens is random (Figure 7C, D).

Table 4Length of the REA (Squared box) for mineralogy (Box Countin Method; BCM) and pororiginating from the different drill cores. For all samples investigated from the same drillmaximum values are given to show the spread in the data. Furthermore, the anisotropy

Sample Subsample REA BCM (10% error) (mm)

Mean Max. value Min. Valu

BDR Clay layer 148 184 113BCS2 Clay layer 111 115 107BWS-H Carbonate layer 210 259 188

Clay layer 215 258 102Clay þ Sand layer 245 256 235Sand layer 250 265 191

BCM ¼ Box Counting Method.CovM ¼ Covariance Method.

To describe size distributions of grains, pores and soil pore in-terfaces concepts of fractal geometry have beenwidely used (Krohnand Thompson, 1985; Hansen and Skjeltorp, 1987; Mandelbrot,1987; Korvin, 1992; Lipiec et al., 1998; Martin and Tarquas, 1998;Li et al., 2000; Dathe et al., 2001; Menendez et al., 2005; Desboiset al., 2009; Yu et al., 2009; Houben et al., 2013; Hemes et al., inpress).The 2D size distribution of pores in the BIB-SEM imageswithin the clay matrix can be described using a continuous fractalset equation, from which it follows that the data displays a power

osity (Covariance Method; CovM) for the different layers investigated for samplescore the mean REA values are given in the table, where also both the minimum andfactor show the difference in variance in the segmented porosity graph in x and y.

REA CovM (10% error) (mm) Anisotropy factor

e Mean Max. value Min. Value

131 184 60 0.990131 135 126 1.005203 241 166 1.04082 121 43 0.995

157 201 112 0.984181 181 180 0.989

Table

5Th

etablesu

mmarizes

thetotalv

isible

porosityper

inve

stigated

ofthedifferentSE

Mmosaics,andtheco

ntribution

totheporosityforthedifferentiated

mineralsan

dmineral

aggreg

ates.Furthermore,2D

porosityper

segm

ented

mineral/m

ineral

aggreg

ateis

also

specified

.

Sandyfacies

Shalyfacies

BW

SBDR

BCS2

Clay

Clayþ

Sand

Sand

Clayþ

Sandþ

Carbo

nate

Carbo

nate

Clayþ

Carbo

nate

Clay

Clay

Totalvisible

4(%)

0.67

1.72

3.62

1.46

3.08

1.46

0.90

2.35

Con

t.4

4Con

t.4

4Con

t.4

4Con

t.4

4Con

t.4

4Con

t.4

4Con

t.4

4Con

t.4

4

Claymatrix

66.3

0.5

67.0

3.5

68.5

5.3

53.0

1.7

95.0

9.0

57.0

2.5

62.2

0.8

49.0

1.4

Quartz

<1.0

0.05

<1.0

0.05

<1.0

0.06

1.5

0.08

1.0

0.7

1.0

0.09

<1

0.03

0.0

0.0

Calcite

1.3

0.2

1.5

0.3

3.0

0.4

3.0

0.2

2.0

0.1

3.0

0.1

1.6

0.1

0.0

0.0

Mica

<1.0

0.4

3.5

1.1

<1.0

0.8

2.5

0.6

ee

ee

<1

1.8

ee

Pyrite

4.3

2.8

<1.0

3.1

1.5

7.3

11.5

14.6

ee

1.0

2.7

<1

0.6

ee

Siderite

7.3

12.1

7.0

11.9

6.5

8.9

2.0

7.6

2.0

4.3

10.0

18.0

10.0

10.9

2.5

4.6

Fossil

ee

ee

ee

ee

ee

ee

4.0

7.1

27.5

14.8

Other

<1.0

1.4

5.5

3.4

16.5

4.9

3.0

4.1

ee

ee

2.2

2.0

ee

Crack

s19

.0e

15.0

e2.5

e23

.5e

ee

28.0

e18

.0e

21.0

e

Totalvisible4¼

Visible

porosityin

theen

tire

SEM

micrograp

h.

Con

t.4¼

Con

tribution

tothetotalvisibleporosityin

percent.

4¼

Visible

porosityper

mineral

aggreg

atein

percent.


law area distribution, and for different mosaics the pore size dis-tributions can be compared when the pore size distribution isnormalized to the total investigated area, expressed as (Houbenet al., 2013):

Log½Ni=ðbi$SmosaicÞ� ¼ �D$Log½Si� þ LoghC*

i(2)

where Ni is the frequency of the pores with a characteristic surfacesmaller than Si within the bin-size bi (b1 ¼1 and bi ¼ 2. bi-1), Smosaicis the area of the mosaic, D is the power law exponent and C* is aconstant of proportionality. At the low end of the pore size distri-bution the data is restricted by the magnification used (Smin), poreareas smaller than 15 pixels could not be consistently and accu-rately detected and therefore these are not included in the sizedistribution results. At the high end of the pore sizes (Smax) the datais limited by the fact that the investigated area is not infinite and/orbigger pores are not present in this rock, meaning that the powerlaw distribution of the pore size is only valid in a certain rangedefined by Smin and Smax. Figure 8 shows the frequency of poresbound to the clay matrix as a function of the binned-area meanpore area (see also Klaver et al., 2012; Houben et al., 2013; Hemeset al., in press). The power law exponent and constant of propor-tionality are inferred from the best power-law fit to the data be-tween Smin and Smax, where the outermost liner fits through thedata are used as a measure of uncertainty for the best fit (Klaveret al., 2012).

In Figure 8 the normalized pore size distribution of the porespresent in the clay matrix in all mosaics is plotted. The pores in theclay matrix follow a power law distribution with a power lawexponent between 2.6 < D < 2.1, with an average power law dis-tribution for pore areas of D ¼ 2.25 � 0.079 (Fig. 8) for pore areasbetween 104 and 108 nm2. The log C* is �1.06 � 0.433 for the samerange of pore areas on average.

3.6. Mercury intrusion porosimetry

Sandy facies samples from a carbonate lens, clay layer and sandlayer and two Shaly facies samples were intruded with mercury(Table 7). The connected MIP porosity measured is between 9% inthe carbonate lens of the Sandy facies and 13% in a clay layer of theShaly facies (Table 7, Figure 9A and Figure 9B), where the highestpressure step used measured pore throat diameters down to 3 nm.MIP data shows that most abundant pores have a pore throatdiameter smaller than 200 nm in all the investigated samples(Figure 9C), where clay layers show more abundant pores withsmaller pore throats (average < 20 nm in diameter) than the car-bonate lens and the sand layer (average 50e100 nm in diameter).All data with pore diameters larger than 100 mm are interpreted asporosity measured due to the surface roughness of the samples and‘drying/stress relaxation’ cracks present in the samples (see alsoKlaver et al., 2012; Houben et al., 2013), and hence the data pre-sented in Table 7 and Figure 9 is corrected data. Samples wereweighed before and after the experiment (Table 7); a weight gainwas measured in four out of five samples and one sample showed aweight loss. Extrusion measurements and weight gain of thesample show comparable results for the Shaly facies samplesinvestigated (Table 7), inferring that the mercury infiltrated thesamples during these experiments. On the other hand according tothe extrusion measurements there should still be 53e59% of mer-cury in all the Sandy facies samples, but after weighing only 9% ofHg was present in the sand layer and 34% of Hg was present in theclay layer, where the carbonate lens even lost weight.

The difference between intruded and extruded mercury can beconverted to pore-body to pore-throat ratios (Table 7;Webb, 2001).

Figure 4. Typical minerals and their corresponding pores found in the Opalinus Clay. A. A quartz grain with intragranular circular pores. B. Framboidal pyrite with intergranularpores, where shape and size is depending on the pyrite grain packing. C. Calcite grain with intragranular porosity. D. Siderite mineral with characteristic toothed edged pores withlobate features on the inner pore wall. E. Mica with very elongated mainly triangular pores in between lamellae. F. Half-moon shaped fossil with rugged shaped pores that exhibitsmooth pore walls. G. A crack, these are mainly present in the clay matrix. The cracks are very elongated, show rough edges and sharp thin tips at both ends. H. Pores in the claymatrix can be subdivided into three pore types (i) elongated pores between similar oriented clay sheets, (ii) crescent shaped pores in saddle reefs of folded sheets of clay and (iii)large jagged pores surrounding mineral grains.


Figure 5. Pore A e E. are pictures of the same pore imaged at different magnifications, left to right magnifications used are; 10,000 � , 20,000 � , 25,000 � , 30,000 � and 50,000 � .This pore has been traced on the different images generating information of dependency of the pore area, perimeter, length and width on the magnification.


The pore-body to pore-throat ratios are: 1:41, 1:30 and 1:33 for theclay layers, 1:22 for the sand layer and 1:09 for the carbonate lens.Pore throat areas (Sth) follow a power law size distribution (10e106 nm2) with a (best fit) power law exponent of 2.25� 0.059 and alog C* of �1.01 � 0.205 (Figure 9D), where again the outermostpower law fits through the data are used as a measure of uncer-tainty for the best fit.

4. Discussion

4.1. Microstructure of Opalinus Clay

In all layers investigated of both the Shaly and Sandy facies, thesame abundant mineral phases were found in the BIB-SEM mi-crographs: quartz, calcite, mica, pyrite, siderite, feldspar and a claymatrix. Variance in quantity of a certain mineral phase definesdifferent sedimentary layers within the different facies. The min-eral content measured in the BIB-SEM mosaics (Table 2) differsfrom the mineral content measured using XRD measurements(Table 3). This is due to a difference in measuring method, whereinminerals <2 mm in size are not segmented in the BIB-SEM mosaicsand hence part of the clay matrix (Guggenheim and Martin, 1995),whereas using XRD, these minerals smaller than 2 mm will beclassified as the specific mineral. This accounts for a systematicallyhigher quartz and calcite content in the XRD measurements thanthe quartz and calcite content derived from the BIB-SEM data.

The REA for mineralogy and porosity for BIB-SEM polishedsections of Opalinus Clay are comparable for a 10% error range(Table 4). Areas investigated should at least be 250� 250 mm for theSandy facies and 180 � 180 mm for the Shaly facies samples, forsmaller areas the representativeness of the mosaic becomes less

Table 6This table illustrates the average shape change with respect to SEM magnification used to10,000 � is used as a reference (Xx/X10,000).

Perimeter Area

Average Standard error Average Standard erro

50000 1.076 0.0423 1.003 0.120530000 1.063 0.0429 1.063 0.118625000 1.067 0.0387 1.091 0.113320000 1.081 0.0533 1.170 0.1784All 1.072 0.0224 1.082 0.0682

(and the error becomes larger), meaning that the information aboutporosity and mineralogy can only be interpreted on a local scale.This also implies that FIB-SEM measurements (usually sample sizeis around 1000 mm3) should be accurately located within the tar-getedmineral phase (Keller et al., 2011, 2013a, 2013b). Furthermore,an anisotropy factor of around one, when calculating the variancein x and y, implies that perpendicular to the bedding the porestructure is isotropic (see also Houben et al. (2013)), whereas in 3Dan anisotropic pore structure has been found by Keller et al. (2011).

Pores in the clay matrix are most abundant; more than 700,000pores were segmented in all mosaics and about 90% of these arelocated in the clay matrix. Pore morphology of the different min-erals and mineral aggregates is shown qualitatively in Figure 4.Figure 6 illustrates the quantitative pore shape, illustrating thatpores in different minerals show different circularity and axial ra-tios. Specifically, pores in mica and cracks could be easily differ-entiated based on their pore shape in 2D. Pores in the differentminerals show qualitatively as well as quantitatively similar shapeswithin the Shaly and Sandy facies.

Pore shape is thought to reflect the rock formation history(burial, diagenesis, etc.), where pores in the clay matrix are mostlikely a result of the burial and compaction history (Heath et al.,2011; Loucks et al., 2012). Formation of the cracks in the claymatrix is more obscure, but due to their high aspect ratio,orientation along the bedding, rough pore edges and sharp cracktips cracks are interpreted as not being part of the in-situmicrostructure. Cracks can originate from either drying of thesamples (Soe et al., 2009), drilling and stress relaxation (Corkumand Martin, 2007; Klinkenberg et al., 2009) or a combination ofboth. Different drying techniques performed on Shaly faciessamples did not show noticeable differences to the

image the pore, together with a standard error. Bear in mind that a magnification of

Length Width

r Average Standard error Average Standard error

1.080 0.0446 0.965 0.04471.089 0.0458 0.988 0.04661.096 0.0448 1.003 0.04151.103 0.0542 1.035 0.05551.092 0.0239 0.998 0.0239

Figure 6. Data density graphs of the bivariate histograms of the pore shape factors. Showing 1/Axial ratio vs. circularity of the pores per mineral/mineral aggregate they areoccurring in. The graphs with the grey background show the pore shape factors encountered in the Shaly facies and the graphs with the white background are the pore shape factorsas calculated for the samples originating from the Sandy facies.


microstructure with respect to the microcracks (Houben et al.,2013), hence more research is needed to investigate the originof the cracks. Mica pores are caused by interlayer delaminationwhere the interlayer bond is weak due to stacking mistakes (Viti,2011). The pores in framboidal pyrites are due to uneven packingof the single pyrite grains, leaving irregularly shaped pore spacesbetween grains. Distinct pores in the fossils are part of the shell(Pittet and Mattiolo, 2001; Mattiolo and Pittet, 2002) and couldbe slightly altered by the burial and compaction history of therock. Intragranular pores in quartz, calcite and siderite couldhave formed due to the presence of impurities incorporated intothe crystal lattice or defects, such as screw dislocations, edgedislocations and subgrain boundaries (Audetat and Gunther,1999; Fernandez-Diaz et al., 1996). In addition, temperatureand pressure changes can cause high enough stresses to generateintragranular pores due to recrystallization and/or crystal growth(e.g.: Mancktelow et al., 1998; Schmatz and Urai, 2011). Processeslike formation of fluid inclusions, chemical alteration of mineralsand partial dissolution of minerals can cause formation ofintergranular pores.

4.2. Mercury intrusion porosimetry measurements

Although all samples show a lower weight gain after the MIPexperiments than expected by the MIP extrusion curves it is clearthat the mercury infiltrated both Shaly facies samples (see alsoHouben et al., 2013) and the clay layer of the Sandy facies. Themeasured weight of the sand and carbonate samples (Sandy facies)after MIP measurements tells us that (almost) all the mercury leftthese samples, while the extrusion curves show that ca. 50% of themercury stayed behind in the samples. This discrepancy impliesthat for these particular samples it is not clear whether the sampleswere: 1) infiltrated with mercury at all and this extruded thesample after the measurement, 2) a piece of sample material brokeoff resulting in a lower weight after the MIP measurements, or 3)that the sample was never infiltrated with mercury and was onlymechanically compacted (see also Hildenbrand and Urai, 2003). Notfinding clear evidence for mercury infiltrating the samples makesinterpretation of the MIP data ambiguous. The measured porositywithin these two samples could be interpreted as being mechanicalcompaction, whereof resp. 5.22 and 5.13% was irreversible

Figure 7. Pore orientation in the clay matrix in different layers of the Sandy (Saf) and Shaly (Shf) facies. A. Sandy facies clay matrix shows a pronounced pore orientation along thebedding. B. Clay matrix of the Shaly facies showing a distinct pore orientation along the bedding. C. Pores in the clay matrix of the carbonate lens showing a random poreorientation. D. Pores in the clay matrix in a sand layer in the Sandy facies. The pore orientation makes an angle to the bedding.

Figure 8. Pore size distribution of the pores in the clay matrix. In this graph pore areas are plotted versus no. of pores below a certain size i, which is than normalize by the area ofthe mosaic and the bin size. The graph show that the pores in the clay matrix in all the mosaics (depending on the magnification: 104 nm2 < Si < 108 nm2) show a similar power lawpore size distribution with a mean power law dimension of 2.25 � 0.079 and a constant of proportionality of �1.06 � 0.433.


Table 7Table summarizing the MIP data for the five different samples investigated.

Shaly facies Sandy facies

Clay layer e I Clay layer e II Carbonate lens Clay layer Sand layer

Volume (cm3) 0.05 0.27 0.41 0.38 0.33Density (g/cm3) 2.42 2.24 2.54 2.44 2.43Porosity (%) 11.56 13.01 8.84 9.90 9.68Weight before experiment (g) 0.16 0.61 1.13 1.02 0.89Weight after experiment (g) 0.25 1.00 1.10 1.21 0.93Pore body to pore throat ratio 1:41 1:30 1:09 1:33 1:22Hg in sample after extrusion (vol.%) 9.94 8.72 5.22 5.74 5.13Hg in sample by weight (vol.%) 9.83 7.94 �0.53 3.37 0.87


compaction due to for instance pore collapse, resulting in an uppervalue for porosity measured with MIP in these two layers. The restof the measured ‘porosity’ can then be interpreted as elasticcompaction. In addition, pore throat size distributions inferredfrom MIP measurements do show a similar size distribution as thepore body size distributions inferred from BIB-SEM. However, usingthe Washburn equation to convert pressures to throat sizes for thecarbonate and sand layer is not a valid method since it is unclearwhether the mercury actually penetrated the samples. Recalculat-ing the throat size distributions for only the samples that weresurely intruded with mercury (shale layers) results in a power lawdistribution exponent of 2.22 � 0.053 and log C* ¼ �0.85 � 0.236.

Figure 9. A. MIP intrusion and extrusion for the 5 different samples measured. B. MIP porosthe carbonate lens to ca. 13% in the clay layers of the Shaly facies. C. MIP pore size distribucarbonate and clay layers in both the Shaly (Shf) and Sandy (Saf) facies. D. Pore throat size

Porosity values inferred from MIP measurements are: 12.3% claylayer Shaly facies, 9.9% clay layer Sandy facies, 5.2% carbonate layerSandy facies and 5.1% sand layer Sandy facies.

4.3. Effect of the minerals on the microstructure of Opalinus Clay

The size and shape of particles reflects the formation history ofthe rock, e.g. sedimentation, burial and chemical processes (Choet al., 2006). Packing density of clay condenses porosity andmineralogical information about the Opalinus Clay into oneparameter and can be calculated when the total area is partitionedinto three categories, porosity, clay volume fraction and non-clay

ity vs. the pore diameter (nm). The total porosity measured with MIP ranges from 9% intion data showing differences in abundant pore throat diameter for the pores in sand,distribution inferred from the MIP measurements.

Figure 10. Different microstructures encountered in the Opalinus Clay, from left to right; a quartz rich layer in the Sandy facies, a clay and quartz mixture in the Sandy facies and aclay rich layer in the Shaly facies. The figure illustrates the difference in amount of clay matrix, clay packing density, mineral content, average mineral size and clay matrix porosityin these two layers.


volume fraction that sum to one (Bobko and Ulm, 2008). The claypacking density h is then calculated by (Bobko and Ulm, 2008)

h ¼ 1� 4=½1� finc� (3)

where 4 is the porosity and finc is the non-clay volume fraction. Theclay packing density depends on the used method for calculation ofporosity andmineral volume fraction (Bobko and Ulm, 2008). In thework presented here, the clay packing density can be calculated bycombining the XRD and MIP data or from the BIB-SEM data. Theclay packing density calculated for the XRD and MIP data is higherin the clay layers in both the Shaly and Sandy facies (respectively0.78 and 0.82) than in the sand layer (0.63) and carbonate lens(0.33). Using the same approach for the 2D BIB-SEM data set gives acomparable result; clay-packing densities in 2D for clay layers ofthe samples BWS, BDR and BCS2 are higher (0.98) than for the sandlayer (0.88) or the carbonate lens (0.91). One has to bear in mindthat both the porosity values and mineral amounts as calculatedfrom the BIB-SEM method are different than these from the XRD/MIP data. This explains the different values calculated for the claypacking densities for the two methods. However, they follow acomparable trend showing that the packing density is higher forclay-rich layers. The difference in clay packing density in thedifferent layers is also reflected by the visible porosity observed inthe clay matrix, which is between 0.5 and 1.4% for the clay layersand 5.3 and 9.0% for the sand layer and carbonate lens, respectively(Table 5). Furthermore, average pore size increases with highermineral content: in the sand layer and carbonate lens (Sandy facies)the average equivalent pore radius is respectively 74 and 71 nm,whereas in the clay layer of the Sandy facies the average equivalentpore radius is 52 nm.

In all the investigated mosaics mineral grains are embeddedwithin the clay matrix, indicating that in both the Shaly and Sandyfacies the microstructure is matrix supported. Although the poresize distribution in the clay matrix shows a comparable power-lawarea-size distribution in all investigated samples, the porosity (in %)in the claymatrix differs. A higher value for the clay packing densityimplies a lower visible porosity in the clay matrix. Pores in the claymatrix are also more randomly orientated with increasing quartzand calcite content (Fig. 7). Furthermore, clay matrix porosity(Table 5) and average pore size increases with increasing quartz andcalcite minerals, which is interpreted to reflect a less intense me-chanical compaction in the latter two (Fig. 10), where an increasingmineral size (related to the increasing amount of rigid phases, i.e.non-clay minerals) may prevent the clay matrix from compaction.

4.4. Porosity of Opalinus Clay

Dimensions of the BIB-SEM mosaics investigated and corre-sponding number of visible pores (>680,000 in all mosaics) areextensive enough to contain statistically significant information tobe representative for the investigated polished cross-sections.Conversion of 2D data (BIB-SEM) to the 3rd dimension (MIP) isnot straightforward, especially when the 2D pores are not disk-shaped (Higgins, 2000). However, by using the basic principles ofstereology being in a representative area gives a good estimate ofthe 3D bulk porosity (Underwood, 1970).

The BIB-SEM data shows a power-law distribution for pore area-sizes in the clay matrix between 104 and 108 nm2 with a power-lawexponent D ¼ 2.25 � 0.079. This value is comparable to the porearea distribution power-law exponent D ¼ 2.44 � 0.157 asdescribed in Houben et al. (2013). Assuming we can use theWashburn equation to convert pressures to pore throats, the visibleporosity in the BIB-SEM mosaics is similar to that measured withMIP at the resolution of the BIB-SEM method (Fig. 11A and B). Thisconfirms that the 2D data inferred from the BIB-SEM images isrepresentative of the bulk porosity. In addition, the similar valuesfor the power-law exponents for throat (MIP) and area (BIB-SEM)size distributions suggests the presence of a connected tube-likepore network in the clay matrix of Opalinus Clay. This is not inaccordancewith the pore body to pore throat ratios calculated fromthe MIP data, nor the fact that according to the MIP measurementsmore than 50% of the mercury stayed inside the samples after theexperiments.

Porosity measured with MIP (ca. 5e12%; Table 7, Fig. 9) is higherthan the visible porosity in the BIB-SEM mosaics (0.7e3.1%;Table 5). Differences in the porosity values are mainly due to dif-ferences in resolution between the twomethods:MIPmeasures theconnected porosity down to a pore throat size of 3 nm, whereasBIB-SEM measures the visible porosity down to a certain pixel sizedepending on the SEM magnification used (30 nm at a magnifica-tion 10,000 � and 15 nm at a magnification of 20,000 � ; Klaveret al., 2012; Houben et al., 2013). In addition, water-loss porosityand physical porosity (samples dried at a temperature of 105 �C)measured on different but comparable samples of the same Sandyfacies drill cores by Peters et al. (2011) show porosities of 17.2% and17.7% (clay layer), 6.2% and 6.1% (bioclastic limestone) and 10.0%and 9.9% (sandy clays) (Fig. 11). Water-loss porosity and physicalporosity measured for the Shaly facies are respectively: 12.6e21.1%and 14.0e24.7% (Bossart and Thury, 2008, Fig. 11). Furthermore, N2BET-analysis show a BET surface of 19e49 m2/g, with a mean value

Figure 11. Porosity in the Sandy and Shaly facies of Opalinus Clay, where the graph illustrates the difference in porosity values measured MIP porosity, water loss porosity, totalphysical porosity, visible BIB-SEM porosity and extrapolated BIB-SEM porosity. A. Measured porosity values for the Shaly facies of Opalinus Clay, where the range of water-lossporosities and total physical porosities are published by Bossart and Thury (2008). B. Measured porosity values for a shale layer of the Sandy facies of Opalinus Clay, water lossand physical porosity from Peters et al. (2011). C. Measured porosity values for a sand layer of the Sandy facies of Opalinus Clay, water loss and physical porosity from Peters et al.(2011). D. Measured porosity values for a carbonate layer of the Sandy facies of Opalinus Clay, water loss and physical porosity from Peters et al. (2011).


of 28 m2/g (NAGRA, 2002; Keller et al., 2011). Correspondingporosity, for BET surface areas of 20e21 m2/g calculated by Kelleret al. (2011), is 10e12%. N2 BET surface measured for the Sandyfacies, clay layer is 35.1 m2/g and for the Sandy facies (bioclasticlimestone) the N2 BET surface measured is 6.37 m2/g (Peters et al.,2011). Unfortunately only Keller et al. (2011) calculated corre-sponding porosity values using the ‘modelless’ method imple-mented in the Beckman Coulter Sorption Analysis software. Resultsdo show that the BET surface in the carbonate lens of the Sandyfacies is significantly lower than the BET surface measured for theclay layer (Sandy facies) and the Shaly facies. This implies a lowerporosity and/or larger pores in the carbonate layer of the Sandyfacies, hence the overall trend in the gas adsorption data withrespect to the porosity is similar to that found with the BIB-SEMand MIP methods. Differences in MIP and water-loss or physicalporosities, observed for clay layers of the Sandy and Shaly facies,can be explained first by the fact that MIP measures only porositythat is connected to the samples outside down to a pore throatdiameter of 3 nm, whereas it is assumed that water-loss andphysical porosity methods measure the porosity down to a porethroat diameter <3 nm. Corresponding pore throats for water-lossand physical porosity depend on the temperature used for sampledrying (105 �C), an exact corresponding pore throat value is notknown, but it is assumed to be in the range of 1e3 nm becausebound (adsorbed and interlayer) water has not been or has onlypartly been removed by drying at 105 �C (Pearson et al., 2003;Marschall et al., 2005). In claystones interlayer adsorbed water isremoved up to temperatures of 250 �C (Plötze et al., 2007) and in

addition water adsorption and desorption data (NAGRA, 2002)shows a most abundant pore diameter of 3 nm. Secondly, differ-ences between MIP and water-loss or physical porosities may alsobe due to the fact that these porosity measurements are done oncomparable but not on the same samples.

Differences in porosity values between the BIB-SEM measure-ments and MIP, water-loss, physical porosity measurements areexplained by the fact that smaller pores are expected to be presentin the clay matrix although they are not visible in the SEM mosaicsinvestigated here. In trying to evaluate the total porosity, it isassumed that the distribution of pore-areas below the resolution ofBIB-SEM method follow a similar power-law area-size distributionmeasured for pore-areas detected in the SEMmosaics. This enablesthe estimation of the total porosity by extrapolation of the power-law down to the resolution of MIP; i.e. the pore areas are extrap-olated down to pore areas of 7.07 nm2 corresponding to a porediameter of 3 nm. Extrapolation of the pore areas, usingD ¼ 2.25 � 0.079 and log C* ¼ �1.06 � 0.433, results in a totalextrapolated clay matrix porosity between 20.5 and 25.3% whenextrapolating both to larger and to smaller pore areas. Assumingthat each non-clay porous mineral/mineral aggregate has a char-acteristic porosity in the range of their visible porosities in the SEMimages (Table 5), the combination of porosity born by non-clayminerals and the extrapolated clay matrix porosity enables calcu-lating the total porosity in the SEM mosaics. Total porosity of theBCS drill-core sample is for the best fit 15.3% (between 9.4 and27.4% using the two outermost fits), which is comparable to themeasured physical/water-loss porosity range for the Shaly facies

Figure 12. The ‘elementary building blocks’ for the Opalinus Clay microstructure are shown in the top row. These ‘blocks’ can be used to schematically draw the microstructurewhen the mineralogy of sample is known. When combined in an appropriate way the homogeneous elementary building blocks can be put together to illustrate the differences inthe different layers encountered in the heterogeneous Opalinus Clay. Shaly facies drill-cores (mmecm scale) are relatively homogeneous and this is reflected in the microstructureof this layer. The Sandy facies is more heterogeneous on mmecm scale, hence microstructurally this facies is subdivided in clay layers, sand layers and carbonate lenses, for themicrostructure it is important than to know the mineral composition and distribution per sub-facies.


(12.6e24.7%; Bossart and Thury, 2008, Fig. 11 A) and slightly higherthan the MIP porosity. The total extrapolated porosity in the claylayer of the Sandy facies (BWS drill core) is 14.1% (8.63e25.2%),which is again in agreement with thewater-loss/physical porosities(17.2e17.7%) and higher than theMIP porosity of 9.9% (Fig.11 B). Forthe sand layer, the total extrapolated porosities are calculated to bebetween 6.6 and 13.0% and on average ca. 9.3%, which is higherthan the MIP porosity (5.1) and comparable to the water-loss/physical porosity of 10% (Fig. 11C). The carbonate layer shows anextrapolated porosity of 10.5% which is higher than both the MIPand water loss/physical porosities (Fig. 11 D). Overall sand andcarbonate rich layers display a lower total porosity (ca. 6e10%;Fig. 11) than the clay-rich layers (ca. 12e25%; Fig. 11). This is alsoobserved by Crawford et al. (2008) and Koroleva et al. (2011) whostated that as soon as the clay content in a rock becomes highenough, so that the mineral grains are ‘floating’ in the matrix, theporosity of the rock will increase with increasing clay content.

Although the Sand and Carbonate layers in the Sandy faciesdisplay an overall lower porosity than the clay layers in eitherSandy or Shaly facies, the porosity in the clay matrix is actuallyhigher (see also previous paragraph). Since all samples are matrixsupported this can actually mean that the clay matrix of sand layersand carbonate layers are preferred pathways for possible flowthrough in Opalinus Clay due to the fact that the pores are actuallylarger here. That this is also true in 3D for carbonate layers is shownby Keller et al. (2013b). Whether this also applies to the sand layersand whether the pores in the clay matrix of the sand layer actuallyform a connected pore network has to be investigated using theFIB-SEM and/or m-CT methods.

4.5. Opalinus Clay ‘elementary building blocks’

The Opalinus Clay comprises the following minerals: quartz,calcite, mica, siderite, pyrite, feldspar, organic matter, ankerite/


dolomite and clay minerals. Qualitatively both Shaly and Sandyfacies, and all differentiated layers within, comprise a number of allthese minerals, quantitatively the amount of a certain mineralpresent shows a large range (Tables 2 and 3) and depends on thespecific layer and facies investigated. Based on 16 BIB polishedcross-sections and more than 700,000 segmented pores, it isconcluded that all identified porous minerals/mineral aggregatesdisplay each a characteristic porosity range (Table 5), characteristicpore geometry based on the images (Fig. 4) and a characteristicpore shape based on axial ratio and circularity (Fig. 6). Therefore,we propose that each porous mineral can be considered as anelementary building block for Opalinus Clay (Desbois et al., 2011b).

The two end-member facies (Sandy and Shaly) of Opalinus Clayshow the following common homogeneous elementary buildingblocks (Fig. 12): (1) minerals featuring intragranular porosity notconnected to the clay matrix at SEM magnifications used (e.g.:quartz, calcite, siderite), (2) intragranular pores that are connectedto the clay matrix (e.g.: fossil, mica), (3) intergranular pores be-tween minerals (e.g. framboidal pyrite), (4) intergranular cracksand (5) intergranular pores in the clay matrix. Different facies ofOpalinus Clay are composed of different amounts of theseelementary building blocks (mineral composition; see also Houbenet al., 2013). Combining the right amounts (mineral composition) ofthe different homogeneous building blocks and by distributingthem in a suitable way (mineral distribution) the different homo-geneous elementary building blocks together can be used to outlinethe microstructure of the heterogeneous Opalinus Clay. Hence, assoon as the mineralogy and the mineral distribution are known,one can visualize the microstructure schematically (Fig. 12)together with a first approximation of the porosity present in thatparticular sample. In addition, identifying common porous min-erals in the BIB-SEM cross-sections helps guide FIB experiments todifferent targets for investigation of the 3D (connected) porosity inthe different mineral phases.

5. Conclusions

BIB-SEM measurements are made on representative elementaryareas and extrapolated porosity values give similar results to bulksample porosity measurements, like; Water-loss, physical and MIPporosimetry. Increasing mineral size may prevent the clay matrixfrom compaction resulting in a higher clay packing density in theclay-rich layers and a lower visible BIB-SEM porosity. Sand and car-bonate rich layers display a total porosity (BIB-SEM, water-loss andphysical) of 6e10% and the clay-rich layers show a total porosity of12e25%. Although the overall porosity is lower in the sand and car-bonate rich layers (Sandy facies), the actual clay matrix pore size isslightly larger than in the clay-rich layers. This could mean thatthrough flow rates are slightly higher in these particular layers.

Common minerals are considered together with the cracksand a clay matrix as elementary building blocks of Opalinus Clay.Combining the right amounts of the homogeneous buildingblocks together, as present in the different layers, illustratesschematically the microstructure of Opalinus Clay. Furthermore,knowing the mineralogy can give a first approximation of thetotal porosity present in the different layers of the heterogeneousOpalinus Clay.

Acknowledgements

U. Wollenberg from the Geological Institute RWTH-Aachenshould be thanked for his help with sample preparation, SEM mi-croscopy and performing the XRD measurements. We would like toshow our gratitude to NAGRA for funding and supporting thisproject and in particular to Dr. D. Traber. The Swisstopo team at the

Mont Terri rock laboratory is thanked for their hospitality duringdrilling campaigns and providing the drill cores. Furthermore, Prof.M.Mazurek (University of Bern) andDr. D. Jaeggi (Swisstopo) shouldbe thanked for providing technical reports on the investigatedsamples. Furthermore, the Hiwi’s are thanked for their help with themanual interpretation of the SEM mosaics, in particular: J. Hürtgen,J. Schneider, Jörg Schoel, C. Diebely, K.R. Nokar, T. Thiel, F. Halpaap.

References

Abell, A.B., Willis, K.L., Lange, D.A., 1998. Mercury intrusion porosimetry and imageanalysis of cement-base materials. J. Colloid Interf. Sci. 211, 39e44.

Audetat, A., Gunther, D., 1999. Mobility and H2O loss from fluid inclusions in naturalquartz crystals. Contrib. Mineral. Petrol 137, 1e14.

Bera, B., Siva Kumar Gunda, N., Mitra, S.K., Vick, D., 2012. Characterization ofnanometer-scale porosity in reservoir carbonate rock by focused ion beam-scanning electron microscopy. Microsc. Microanal. 18, 171e178.

Bobko, C., Ulm, F.J., 2008. The nano-mechanical morphology of shale. Mech. Mater.40, 318e337.

Bossart, P., Thury, M., 2007. Research in the Mont Terri rock laboratory: Quo vadis?Phys. Chem. Earth 32, 19e31.

Bossart, P., Thury, M., 2008. Mont Terri Rock Laboratory e Project, Programme 1996to 2007 and Results, vol. 3. Swiss Geological Survey, Wabern.

Cho, G.C., Dodds, J., Santamarina, J.C., 2006. Particle shape effects on packing den-sity, stiffness, and strength: natural and crushed sands. J. Geotechnical Geo-environ. Eng., 591e602.

Chalmers, G.R., Bustin, R.M., Power, I.M., 2012. Characterization of gas shale poresystems by porosimetry, pycnometry, surface, area, and field emission scanningelectron microscopy/transmission electron microscopy image analysis: exam-ples from the Barnett, Woodford, Haynesville, Marcellus, and Doig units. AAPGBulletin 96, 1099e1119.

Connell, H., Zhu, J., Bassi, A., 1999. Effect of particle shape on crossflow filtrationflux. J. Membr. Sci. 153, 121e139.

Corkum, A.G., Martin, C.D., 2007. The mechanical behaviour of weakmudstone (Opalinus Clay) at low stresses. Int. J. Rock Mech. Mining Sci.44, 196e209.

Crawford, B.R., Faulkner, D.R., Rutter, E.H., 2008. Strength, porosity, and perme-ability development during hydrostatic and shear loading of synthetic quartz-clay fault gouge. J. Geophys. Res. 113 (B03207), 1e14.

Curtis, M.E., Ambrose, R.J., Sondergeld, C.H., Rai, C.S., 2010. Structural Character-ization of Gas Shales on the Micro- and Nano-scales. Canadian society for un-conventional gas, SPE 137693, p. 15.

Dathe, A., Eins, S., Niemeyer, J., Gerold, G., 2001. The surface fractal dimensionof the soil-pore interface as measured by image analysis. Geoderma 103,203e229.

Desbois, G., Urai, J.L., Burkhardt, C., Drury, M.R., Hayles, M., Humbel, B., 2008.Cryogenic vitrification and 3D serial sectioning using high resolution cryo-FIBSEM technology for brine-filled grain boundaries in halite: first results. Geo-fluids 8, 60e72.

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. E-Earth4, 15e22.

Desbois, G., Urai, J.L., De Craen, M., 2010a. In-situ and Direct Characterization ofPorosity in Boom Clay (Mol Site, Belgium) by Using Novel Combination of IonBeam Cross-sectioning, SEM and Cryogenic Methods. In: Motivations, FirstResults and Perspectives. SCK�CEN. External report 10/MDC/P-26.

Desbois, G., Urai, J.L., Houben, M., Sholokhova, Y., 2010b. Typology, morphology andconnectivity of pore space in claystones from reference site for research usingBIB, FIB and cryo-SEM methods. In: EPJ Web of Conferences (European PhysicalJournal), 6 (ICEM 14 e 14th International Conference on Experimental Me-chanics, 22005.

Desbois, G., Urai, J.L., Kukla, P.A., Konstanty, J., Baerle, C., 2011a. High-resolution 3Dfabric and porosity model in a tight gas sandstone reservoir: a new approach toinvestigate microstructures from mm- to nm-scale combining argon beamcross-sectioning and SEM imaging. J. Pet. Sci. Eng. 78, 243e257.

Desbois, G., Urai, J.L., Houben, M., Hemes, S., Klaver, J., 2011b. BIB-SEM of repre-sentative area clay structures: insights and challenges. In: NEA ClayClubWorkshop Proceedings e Clays under Nano- to Microscopic Resolution, Karls-ruhe (Germany), 6e8 September 2011.

ESRI, 2008. ArcMap 9.3. In: Arcinfo Desktop 9.3. ESRI, Redlands, California, U.S.Li, J.F., Li, L., Stott, F.H., 2000. Fractal characteristics of apparent pores present on

polished cross sections of alumina coatings prepared by laser-assisted flamehybrid spraying. Thin Solid Films 453e454, 229e233.

Fernandez-Diaz, L., Putnis, A., Prieto, M., Putnis, C.V., 2006. The role of magnesiumin the crystallization of calcite and aragonite in a porous medium. J. Sediment.Res. 66 (3), 482e491.

Galle, C., 2001. Effect of drying on cement-based materials pore structure as iden-tified by mercury intrusion porosimetry, a comparative study between oven-,vacuum-, and freeze-drying. Cement Concrete Res. 31, 1467e1477.

Goldstein, J., Newbury, D., Joy, D., Lyman, C., Echlin, P., Lifshin, E., Sawyer, L.,Michael, J., 2007. Scanning Electron Microscopy and X-ray Microanalysis, thirded. Springer Science þ Business Media, LLC, ISBN 978-0-306-47292-3.

http://refhub.elsevier.com/S0264-8172(13)00261-4/sref1
























http://refhub.elsevier.com/S0264-8172(13)00261-4/sref82a










































































Gonzalez, R.C., Woods, R.E., Eddins, S.L., 2009. Digital Image Processing. Gatesmarkpublishing LLC, USA, p. 826.

Guggenheim, S., Martin, R.T., 1995. Definition of clay and clay mineral: joint reportof the AIPEA nomenclature and CMS nomenclature committees. Clays and ClayMinerals 43 (2), 255e256.

Hansen, J.P., Skjeltorp, A.T., 1987. Fractal pore space and rock permeability impli-cations. Phys. Rev. B. 38 (4), 2635e2638.

Heath, J.E., Dewers, T.A., McPherson, B.J.O.L., Petrusak, R., Chidsey Jr., T.C.,Rinehart, A.J., Mozley, P.S., 2011. Pore networks in continental and marinemudstones: characteristics and controls on sealing behavior. Geosphere 7 (2),429e454. http://dx.doi.org/10.1130/GES00619.1.

Hemes, S., Desbois, G., Urai, J., Craen de, M., Honty, M., 2013. Variations in themorphology of porosity in the Boom Clay Formation: insights from 2D highresolution BIB-SEM imaging and Mercury injection Porosimetry. Neth. J. Geosci.(in press).

Higgins, M.D., 2000. Measurement of crystal size distribution. Am. Mineral. 85,1105e1116.

Hildenbrand, A., Urai, J.L., 2003. Investigation of the morphology of pore space inmudstones e first results. Mar. Petroleum Geol. 20, 1185e1200.

Holzer, L., Cantoni, M., 2011. Review of FIB-tomography. In: Utke, I., Moshkalev, S.A.,Russell, Ph. (Eds.), Nanofabrication Using Focused Ion and Electron Beams:Principles and Applications. Oxford University Press, NY, USA, ISBN9780199734214.

Holzer, L., Muench, B., Wegmann, M., Gasser, Ph, 2006. FIB-nanotomography ofparticulate systems e part I: particle shape and topology of interfaces. J. Am.Ceram. Soc. 89, 2577e2585.

Holzer, L., Gasser, Ph., Kaech, A., Wegmann, M., Zingg, A., Wepf, R., Muench, B., 2007.Cryo-FIB-nanotomography for quantitative analysis of particle structures incement suspensions. J. Microsc. 227, 216e228.

Holzer, L., Münch, B., Rizzi, M., Wepf, R., Marschall, P., Graule, T., 2010. 3D-micro-structure analysis of hydrated bentonite with cryo-stabilzed pore water. Appl.Clay Sci. 47, 330e342.

Houben, M.E., Desbois, G., Urai, J.L., 2013. Pore morphology and distribution in theshaly facies of Opalinus Clay (Mont Terri, Switzerland): insights from repre-sentative 2D BIB-SEM investigations on mm to nm scale. Appl. Clay Sci. 71, 82e97.

Howard, J.J., 1991. Porosimetry measurement of shale fabric and its relationship toillite/smectite diagenesis. Clays and Clay Minerals 39, 355e361.

Kahle, M., Kleber, M., Jahn, R., 2002. Review of XRD-based quantitative analyses ofclay minerals in soils: the suitability of mineral intensity factors. Geoderma 109,191e205.

Kameda, A., Dvorkin, J., Keehm, Y., Nur, A., Bosl, W., 2006. Permeability-porositytransforms from small sandstone fragments. Geophysics 71 (1), N11eN19.

Kanit, T., Forest, S., Galliet, I., Mounoury, V., Jeulin, D., 2003. Determonation of thesize of the representative volume element for random composites: statisticaland numerical approach. Int. J. Solids Struct 40, 3647e3679.

Keller, L.M., Holzer, L., Wepf, R., Gasser, P., 2011. 3D geometry and topology of porepathways in Opalinus clay: implications for mass transport. Appl. Clay Sci. 52,85e95.

Keller, L.M., Schuetz, P., Erni, R., Rossell, M.D., Lucas, F., Gasser, P., Holzer, L., 2013a.Characterization of multi-scale microstructural features in Opalinus Clay.Micropor. Mesopor. Mat. 170, 83e94.

Keller, L.M., Holzer, L., Schuetz, P., Gasser, P., 2013b. Pore space relevant for gaspermeability in Opalinus clay: statistical analysis of homogeneity, percolationand representative volume element. J. Geophys. Res. Solid Earth 118, 2799e2812. http://dx.doi.org/10.1002/jgrb.50228.

Klaver, J., Desbois, G., Urai, J.L., Littke, R., 2012. BIB-SEM study of the pore spacemorphology in early mature Posidonia Shale from the Hils area, Germany. Int. J.Coal Geol.. http://dx.doi.org/10.1016/j.coal.2012.06.012.

Klinkenberg, M., Kaufhold, S., Dohrmann, R., Siegesmund, S., 2009. Influence of car-bonate microfabrics on the failure strength of claystones. Eng. Geol. 107, 42e54.

Kolor, 2008. Autopano Giga 2.0. Challes-les Eaux, France.Koroleva, M., Lerouge, C., Mader, U., Claret, F., Gaucher, E., 2011. Biogeochemical

processes in a clay formation in situ experiment: Part B e results from over-coring and evidence of strong buffering by the rock formation. Appl. Geochem.26, 954e966.

Korvin, G., 1992. Fractal Models in the Earth Science. Elsevier, Amsterdam, TheNetherlands, p. 396.

Krohn, C.E., Thompson, A.H., 1985. Fractal sandstone pores: automated measure-ments using scanning-electron-microscope images. Phys. Rev. B 33 (9), 6366e6374.

Lipiec, J., Hatano, R., Slowinska-Juriewicz, A., 1998. The fractal dimension of poredistribution patterns in variously-compacted soil. Soil Tillage Res. 47, 61e66.

van Loon, L.R., Soler, J.M., Muller, W., Bradbury, M.H., 2004. Anisotropic diffusion inlayered argillaceous rocks: a case study with Opalinus Clay. Environ. Sci.Technology 38, 5721e5728.

Loucks, R.G., Reed, R.M., Rupper, S.C., Jarvie, D.M., 2009. Morphology, genesis, anddistribution of nanometer-scale pores in siliceous mudstones of the Missis-sippian Barnett shale. J. Sediment. Res. 79, 848e861.

Loucks, R.G., Reed, R.M., Ruppel, S.C., Hammes, U., 2012. Spectrum of pore types andnetworks in mudrocks and a descriptive classification for matrix-relatedmudrock pores. AAPG Bull. 96 (6), 1071e1098.

Mancktelow, N.S., Grujic, D., Johnson, E.L., 1998. An SEM study of porosity and grainboundary microstructure in quartz mylonites, Simplon fault zone, Central Alps.Contrib. Mineral. Petrol 131, 71e85.

Mandelbrot, B.B., 1987. Die fraktal geometrie der natur. Birkhauser verlag, Base-l$Boston, p. 491.

Marschall, P., Horsman, S., Gimmi, T., 2005. Characterisation of gas transportproperties of the Opalinus Clay, a potential host rock formation for radioactivewaste disposal. Oil Gas Sci. Technol. 60 (1), 121e139.

Martin, M.A., Tarquas, F.J., 1998. Fractal modeling, characterization and simulationof particle-size distributions in soil, Proceedings: mathematical. Phys. Eng. Sci.454 (1973), 1457e1468.

Mattiolo, E., Pittet, B., 2002. Contribution of calcareous nannoplankton to carbonatedeposition: a new approach applied to the Lower Jurassic of central Italy. Mar.Micropaleontol. 45, 175e190.

Menendez, I., Caniego, J., Gallardo, J.F., Olechko, K., 2005. Use of fractal scaling todiscriminate between and macro- and meso-pore sizes in forest soils. Ecol.Model. 182, 323e335.

Milke, R., Abart, T., Kunze, K., Koch-Müller, M., Schmid, D., Ulmer, P., 2009. Matrixrheology effects on reaction rim growth I: evidence from orthopyroxene rimgrowth experiments. J. Metamorphic Geol. 27 (1), 71e82.

NAGRA, Dezember 2002. Technischer Bericht 02e03, Projekt Opalinuston: Syntheseder geowissenschaftlichen Untersuchungsergebnisse.

Nussbaum, C., Bossart, P., Amann, F., Aubourg, C., 2011. Analysis of tectonic struc-tures and excavation induced fractures in the Opalinus Clay, Mont Terri un-derground rock laboratory (Switzerland). Swiss J. Geosci. 104, 187e210.

Ortega, J.A., Ulm, F.-J., Abousleiman, Y., 2009. The nanogranular acoustic signature ofshale. Geophysics 74 (3), D65eD84.

Pearson, F.J., Arcos, D., Bath, A., Boisson, J.-Y., Fernandez, A.M., Gäbler, H.-E.,Gaucher, E., Griffault, L., Hernan, P., Waber, H.N., 2003. Mont Terri Project eGeochemistry of Water in the Opalinus Clay Formation at the Mont Terri RockLaboratory. Reports of the FOWG, no. 5, Bern.

Peters, M., Mazurek, M., Jaeggi, D., Müller, H., Nov. 2011. WS-H Experiment:Heterogeneities in the Sandy Facies of Opalinus Clay on a Scale on Milli-meters to Centimeters. Mont Terri Project, Technical Note TN 2010-76,pp. 1e64.

Pittet, B., Mattiolo, E., 2001. The carbonate signal and calcareous nannofossil dis-tribution in an Upper Jurassic section (Balingen-Tieringen, Late Oxfordian,southern Germany). Palaeogeogr. Palaeoclimatol. Palaeoecol. 179, 71e96.

Plötze, M., Kahr, G., Dohrmann, R., Weber, H., 2007. Hydro-mechanical, geochemicaland mineralogical characteristics of the bentonite buffer in a heater experi-ment: the HE-B project at the Mont Terri Rock Laboratory. Phys. Chem. Earth 32,730e740.

Schmatz, J., Urai, J.L., 2011. The interaction of migrating grain boundaries and fluidinclusions in naturally deformed quartz: a case study of a folded and partlyrecrystallized quartz vein from the Hunsrück Slate, Germany. J. Struct. Geol. 33,468e480.

Schneider, J., Flemings, P.B., Day-Stirrat, R.J., Germaine, J.T., 2011. Insights into pore-scale controls on mudstone permeability through resedimentation experi-ments. Geology 39, 1011e1014.

Scrivener, K.L., Füllmann, T., Gallucci, E., Walenta, G., Bermejo, E., 2004. Quantitativestudy of Portland cement hydration by X-ray diffraction/Rietveld analysis andindependent methods. Cement Concrete Res. 34, 1541e1547.

Soe, A.K.K., Osada, M., Takahashi, M., Sasaki, T., 2009. Characterization of drying-induced deformation behaviour of Opalinus Clay and tuff in no-stress regime.Environ. Geol. 58, 1215e1225.

Stipp, M., Tullis, J., Behrens, H., 2006. Effect of water on the dislocation creepmicrostructure and flow stress of quartz and implications for the recrystallizedgrain size piezometer. J. Geophys. Res. 111 (B04201), 19.

The Math Works, 2010. MATLAB 2010b. The MathWorks, Natick, MA, USA.Thury, M., Bossart, P., 1999. Mont Terri Rock Laboratory: Results of the Hydro-

geological, Geochemical and Geotechnical Experiments Performed in 1996 and1997. Geologisch Bericht e Rapports geologiques e Rapporti geologici Nr. 23,Bern.

Toraya, H., 2000. Estimation of statistical uncertainties in quantitative phase anal-ysis using the Rietveld method and the whole-powder-pattern decompositionmethod. J. Appl. Crystallogr. 33 (6), 1324e1328.

Underwood, E.E., 1970. Quantitative Stereology. Addison Wesley, Reading,Massachusetts.

Viti, C., 2011. Exploring fault rocks at the nanoscale. J. Struct. Geol. 33l, 1715e1727.Washburn, E.W., 1921. Note on a method of determining the distribution of pore

sizes in a porous material. Physics 7, 115e116.Webb, P.A., 2001. An Introduction to the Physical Characterization of Materials by

Mercury Intrusion Porosimetry with Emphasis on Reduction and Presentationof Experimental Data. Micromeritics Intrument Corp., Norcross, Georgia (USA).

Wenk, H.-R., Voltolini, M., Mazurek, M., van Loon, L.R., Vinsot, A., 2008. Preferredorientation and anisotropy in shales: Callovo-Oxfordian shale (France) andOpalinus Clay (Switzerland). Clays Clay Miner 56, 285e306.

White, S.H., Johnston, D.C., 1981. A microstructural and microchemical study ofcleavage lamellae in a slate. J. Struct. Geol. 3 (3), 279e290.

de Winter, D.A.M., Schneijdenberg, C.T.W.M., Lebbink, M.N., Licht, B., Verkleij, A.J.,Drury, M.D., Humbel, B.M., 2009. Tomography of insulating biological andgeological materials using focused ion beam (FIB) sectioning and low-kV BSEimaging. J. Microsc. 233, 372e383.

Yu, B., Cai, J., Zou, M., 2009. On the physical properties of apparent two-phase fractalporous media. Soil Sci. America 1 (8), 177e196.

Yven, B., Sammartino, S., Geraud, Y., Homand, F., Villieras, F., 2007. Mineralogy,texture and porosity of Callovo-Oxfordian argillites of the Meuse/Haute-Marneregion (eastern Paris Basin). Mem. Soc. Geol. France 178, 73e90.










http://dx.doi.org/10.1130/GES00619.1























































http://dx.doi.org/10.1002/jgrb.50228

http://dx.doi.org/10.1016/j.coal.2012.06.012


















































































































































marine and petroleum geology - rwth aachen university · 2013-11-18 · a comparative study of...

Documents