High-resolution 3D fabric and porosity model in a tight gas sandstone reservoir:A new approach to investigate microstructures from mm- to nm-scale combiningargon beam cross-sectioning and SEM imaging

Guillaume Desbois a,!, Janos L. Urai a, Peter A. Kukla b, Jan Konstanty c, Claudia Baerle d

a Structural geology, Tectonics and Geomechanics, RWTH Aachen University, Lochnerstrasse 4–20, D-52056 Aachen, Germanyb Geological Institute, RWTH Aachen University, Wüllnerstr. 2, D-52056 Aachen, Germanyc Wintershall Holding AG, Friedrich-Ebert-Straße 160, 34119 Kassel, Germanyd Wintershall Holding AG, Erdölwerke Barnstorf, Rechterner Straße 2, 49406 Barnstorf, Germany

a b s t r a c ta r t i c l e i n f o

Article history:Received 6 January 2011Accepted 6 June 2011Available online xxxx

Keywords:argon beamcross-sectioningSEMtight gas reservoirporosityrotliegend sandstone

The development of new technologies to enhance tight gas reservoir productivity could strongly bene!t froma better resolution and imaging of the porosity. Numerous methods are available to characterize sandstoneporosity. However, imaging of pore space at scales below 1 !m in tight gas sands remains dif!cult due to limitsin resolution and sample preparation. We explored the use of high resolution SEM in combination with argonion beam cross sectioning (BIB, Broad Ion Beam) to prepare smooth, and damage-free, true-2D surfaces oftight gas sandstone core samples from the Permian Rotliegend in Germany, to image porosity down to 10 nm.The quality of cross-sections allows measuring porosity at pore scale, and describing the bulk porosity byde!ning different regions with characteristic pore morphology and pore size distribution. Serial crosssectioning of samples produces a 3D model of the porous network. We present a model of fabric and porosityat 2 different scales: the scale of sand grains and the scale of the clay grains in the intergranular volume.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Producing gas economically from unconventional sources is agreat challenge. Tight gas reservoirs, with their low permeability andlow porosity, have recently receivedmuch interest because of the verylarge reserves, which could be produced with suitable technology.Tight gas reservoirs are found throughout the world and occur in allcommon types of reservoir rocks (Law and Curtis, 2002; Haines, 2005;Holditch, 2006; Littke et al., 2008).

For improved recovery of these tight gas reservoirs, it is essentialto quantify porosity at the pore scale. However, because pores intight gas reservoirs are usually below 1 !m, this is dif!cult because oftechnical and preparation limits. Thus, the fundamental informationabout pore morphologies in tight gas reservoirs is missing, althoughpore morphology and connectivity is the key to correlate porosity andpermeability data.

Traditional techniques for characterization of different aspects ofporosity in sandstone are X-ray computer tomography (micro-CT)scanning (Honarpour et al., 2003; Tono and Ingrain, 2008), magneticresonance imaging (MRI) (Pape et al., 2005), particle size analysis, pointcounting based on petrographic thin sections, environmental scanning

microscopy (ESEM) (Tiab and Donaldson, 1996), X-ray diffraction(XRD) (Ward et al., 2005), X-ray "uorescence (XRF) (McCann, 1998),confocal scanning lasermicroscopy (CSLM) (Menendez et al., 2001) andmercury porosimetry (Favvas et al., 2009). The combination of thesecomplementary methods is considered to give a robust characteri-zation of reservoir properties of sandstones (e.g. Baraka-Lokmane et al.2009) and in sandstones this has produced a growing body of litera-ture of 3D pore models (with a resolution of a few micrometers) andbased on these, numerical models of "uid "ow through porosity,which accurately predict bulk properties such as Darcy Permeability(Sholokhova et al., 2009; Tölke et al., 2010).

There is a considerable body of literature on the evolution ofporosity by compaction and diagenetic processes (Gaupp et al., 1993;Schöner et al., 2008; Zwingmann, 2008) and it appears that toinvestigate porosity at nm scale in tight gas reservoirs, SEM imaging isthe most direct approach (Nadeau and Hurst, 1991; Ziegler, 2006).This method, however, is limited by the poor quality of the inves-tigated surfaces (mainly broken or mechanically polished surfacesincluding the decoration of porosity by colored resin embedment),which make observations and interpretation dif!cult (Lanson et al.,2002; Schöner and Gaupp, 2005).

The recent development of Argon ion source milling tools whichproduce polished cross-sections of exceptional high quality offers anew alternative for high resolution SEM imaging of porosity at nano-scale (Desbois et al., 2009; Loucks et al., 2009; Holzer and Cantoni,

Journal of Petroleum Science and Engineering 78 (2011) 243–257

! Corresponding author. Tel.: +49 241 80 94352; fax: +49 241 92358.E-mail address: g.desbois@ged.rwth-aachen.de (G. Desbois).

0920-4105/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.petrol.2011.06.004

Contents lists available at ScienceDirect

Journal of Petroleum Science and Engineering

j ourna l homepage: www.e lsev ie r.com/ locate /pet ro l

2011). Until now, the Ar-BIB preparation technique was mainly usedfor metallic materials but not for polyphase geomaterials. On the onehand, because it does not induce mechanical polishing, the Ar-BIBtechnique is an alternative to the resin embedment technique forpreparation of delicate samples, which are dif!cult to polish becauseof phases with different hardness. In addition, the large area whichcan be polished (up to 2 mm2, large in comparison to the micro-porosity investigated) by the Ar-BIB allows investigating numerousfeatures in one cross-section and also allows quantitative stereolog-ical analysis (Underwood, 1970; Russ and Dehoff, 2000), which is verydif!cult when broken samples surfaces are investigated. Ar-BIB crosssectioning does not require water or epoxy impregnation, whichcould damage the clay minerals by re-wetting. Thus, the Ar-beampolishing technique provides rapid, damage-free, reproducible sur-faces in comparison to the conventional preparation techniques(Desbois et al., 2009; Loucks et al., 2009).

Focused Ion Beam (FIB) cutting was also demonstrated to be ableto prepare high quality polished cross-sections in clay-rich geomater-ials (Holzer et al., 2007, 2010; Desbois et al., 2009; Dvorkin, 2009).However, this technique induces potentially more damage of thesurface since it uses mainly Gallium source (Erdmann et al., 2006) andtypical areas produced by FIB (about !m2) are much smaller thanthose prepared by BIB (about mm2). Nevertheless, because FIB gunsprovide “focused” beam and are now fully implemented in commer-cial SEM machines, FIB instruments allows precise targeting of theregion of interest and productive serial-cross-sectioning (Holzer et al.;2010), respectively.

BIB and FIB are complementary approaches to prepare highquality surfaces suitable for high resolution SEM imaging (Holzerand Cantoni, 2011).

This contribution describes the use of the BIB high-resolutiontechnique for a better understanding of pore space geometries in 2Dand, for the !rst time, in 3D on tight gas reservoir sandstone samplesfrom the Rotliegend in the north-west Germany.

2. Methods

Four samples were prepared by slowly cutting with a miniaturediamond saw using air as cooling medium, from drill-cores slowlydried during storage in air; and faces ground dry using carbide paperfrom 800 down to 1200 grade. A "at surface of these samples ofabout 1 cm3 was cross-sectioned perpendicular to the bedding, usinga stand-alone argon beam machine (cross-section polisher JEOL SM-09010, Erdmann et al., 2006) to produce high quality polished cross-sections of about 2 mm2, removing a 100 !m thick layer which waspossibly damaged. We used 6 kV acceleration, achieving currents ofabout 150–200 nA for 8 h of milling. The principle of this technique isresumed in Fig. 1. The polished cross-sections were then coated withcarbon, suitable for SEM imaging and EDX chemical analysis. The SEM

microscopes used are a JEOL JSM-7000 F with an EDX system (EDAX-Pegasus work station) and a Zeiss Supra 55 equipped also with an EDXdetector (EDAX, Apollo 10). Typically, SEM (in SE and BSE modes)imaging and EDX studies were performed at 15 kV at a workingdistance of 10 mm. Porosity and fabrics were statistically evaluatedusing the PolyLXMatlab toolbox (Lexa et al., 2005; http://petrol.natur.

Fig. 1. The principle of BIB cross-sectioning. (a) the ion beam irradiates the edge of sample un-masked by the shielding plate to create high quality polished cross-sections suitable forSEM imaging. (b) Overview of a typical cross section performed by BIB cross-sectioning.

Fig. 2. BIB serial cross-sectioning experiment. (a) Overview of the original mechanicallypre-polished sample imaged in BSE mode; the dashed white line indicates the regionof interest selected for the serial cross-sectioning experiment. (b) Serie of BSEmicrographs from top view after the !ve successive cross-sections. These micrographsare used to measure the thickness of each cross-section: the 2nd, 3rd, 4th and 5th sliceshave a thickness of 57 !m, 55 !m, 51 !m and 51 !m respectively. In total, the selectedregion of interest was serial cross-sectioned over a depth of 210 !m.

244 G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

cuni.cz/~ondro/polylx:home) from microstructures digitized in OpenJump (open source GIS software) from ion-beam-polished-cross-section micrographs.

For one of the samples, the stand-alone argon beam machine wasalso used by following a serial cross sectioning procedure in order toestimate the evolution of microstructures in 3D. For this particularexperiment, we produced !ve successive slices of about 50 !mthick (Figs. 2 and 3). After each cut, each surface was imaged at lowmagni!cation (as a mosaic made of more than 40 single images takenat !500 magni!cation) both using SE2 and BSE detectors to producegeneral overview of fabrics and phase density maps over the entireBIB cross-section. An additional high magni!cation picture (as amosaic made of 600–1000 single images taken at !3000 magni!ca-tion) was also produced with an SE2 detector to map the porosity atpore scale within the entire BIB cross-section. This procedure wasrepeated for all of the !ve successive cross-sections. Single imageswere assembled using Autopano software (Giga 2; Kolor) whilesuccessive pictures of cross-sections were aligned and recti!ed using

ArcMap 9.3 software (ESRI) enabling reconstruction of fabric andporosity in 3D. For this serial cross-sectioning experiment, theconditions of imaging are similar to those used for single BIB cross-sectioning experiments.

In order to compare our approach with conventional techniques ofpreparation (Fig. 4), we also prepared mechanically polished thin-sections suitable for optical transmitted light microscopy and electronmicroscopy (gold coated) as well as broken samples for electronmicroscopy (gold coated).

3. Geological setting of the samples analyzed

The studied core samples originate from awell drilled in the ArstenGraben, south of Bremen. The area is situated at the southern marginof the North German Basin, which forms part of the Southern PermianBasin (SPB) (Glennie, 1986, Glennie, 1990, George and Berry, 1993,Plein, 1995, Strömbäck and Howell, 2002, Legler, 2005). Both areintegral part of the Central European Basin System (Littke et al., 2008).

Fig. 3. SE and BSE micrograph mappings of the !ve successive cross-sections produced for BIB serial cross-sectioning experiment. At the scale-view of SE micrographs, the largestintergranular volumes !lled with hairy illite (black arrows) and fractures, which !t roughly the sand-grain contours (white arrows), are clearly visible. BSE micrographs show theevolution of mineral phases over the investigated volume. Darker gray value indicates quartz grains while lighter gray values indicate mostly calcite, feldspars and calcite at the scaleof micrographs. The black square from the SE cross-section-2d cut is a missing picture in the image set.

245G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

The SPB stretches with a width of 300–600 km over ~1.700 kmfrom the eastern United Kingdom to central Poland and the CzechRepublic. The Basin has an asymmetric shape with its deepest part inthe North (McCann, 1998). It is composed of several NW–SE-trendingen echelon subbasins like the Silverpit/Dutch, the North German andthe Polish Basin, which are separated by N–S and NNW–SSE Variscanbasement highs (Börmann et al., 2006). In the Upper Rotliegend I,tectonic quiescence with subsidence being mainly driven by compac-tion and thermal relaxation prevailed, following a phase of intenseLower Rotliegend bimodal magmatism (van Wees et al., 2000).This changed with the onset of the Upper Rotliegend II, which saw anew phase of extensional faulting and associated sedimentation(Bachmann and Hoffmann, 1997).

The investigated reservoir rocks are from the Upper Rotliegend IIBahnsen Member of the Hannover Formation. In this sequence,sediments of preliminary "uvio-aeolian origin, including barchanoiddune, wet to dry interdune, braided "uvial and alluvial fan sedimentswere deposited with thicknesses ranging between 180 m and 450 m.The analyzed sandstone sample from approximately 4800 m depth isof aeolian dune origin and has a white/gray color, a porosity of 8.7%and a permeability of 0.2 mD. Porosity was measured on core plugsusing Mercury injection porosimetry. The measured porosity is thenrelated to the effective (connected) porosity, of a dried sample underlab conditions. The porosity and permeability measurements reportedre"ect the average values for the studied stratigraphic member (nocut off has been applied). The studied samples were selected frompart of the core, which are considered as representative in regards toporosity and permeability as well as mineralogy. Interesting locationswere !rst selected based on log-data; then homogeneity was checkedin 2D in large thin sections in the optical microscope and in volumeusing !CT imaging. It has been observed during testing of the well,that the more favorable reservoir quality is below the gas–watercontact, while in the gas-bearing zone above, the porosity/perme-ability is drastically reduced by authigenic minerals. This led touneconomic "ow rates and the well was subsequently plugged andabandoned.

4. Results

4.1. BIB cross-sectioning

All single BIB cross-sections (Figs. 1 and 3), produced at similarconditions (6 kV for 8 h. of milling), have comparable polished-surface sizes of about 2 mm wide and 1 mm long resulting in area upto 2 mm2. Each single BIB cross-sections have a pseudo Gaussianshape re"ecting the distribution of ion density in the beam. It wasslightly convoluted by the rocking motion of the sample duringmilling used to minimize the curtaining of milled surfaces, whichoccurs typically for heterogeneous materials. Investigations ofmicrostructures at high magni!cations (Figs. 5 to 11) give evidencefor extremely well-polished surfaces down to the resolution of SEM(!3 nm) without damage or artifacts which may disturb the imagingof sub-micrometric structures. Therefore, the high quality of surfacesallows investigating microstructures down to the resolution of SEM intrue 2D cross-sections (Figs. 5 to 11). The high precision of BIB cuttingis particularly well illustrated by Fig. 7.h, which shows that BIB is ableto cut even the hairy illite particles of 50 nm in diameter and 1–2 !mlong along the direction of length.

Serial cross-sectioning (Figs. 3 and 11) is a major improvement ofthe use of stand-alone BIB machines since it allows us to investigatemicrostructural evolutions in 3D. As far as we know, stand-alone BIBmachines were never used for serial-cross-sectioning until now.

The process of serial sectioning does not decrease the quality ofthe BIB cutting: surfaces have similar and reproducible quality andno subsequent waste redeposition was detected at scales of ourobservations. During the serial cross-sectioning experiment the slice

Fig. 4. Typical images used in conventional investigations of tight gas reservoir sandstones.(a) Optical thin sections of investigated Bahnsen sampleunder polarized transmitted light.The sandstones are mainly aggregates of quartz, calcite, feldspars and kaolinite. Only thelargest intergranular volumes at triple junction are visible (decorated in blue by resinimpregnation). The mean grain size is quite homogeneous of about 200 !m in diameter.Grain fabrics are dominated by sutured and concavo/convex contacts possibly evolved bydissolution processes and grain re-arrangement during compaction. (b) SEmicrograph ofbroken sample surface showing a grain of quartz embedded in illite clay. Two types of illiteclay are distinguishable: tangential illite with platymorphology coats the quartz grain andhairy/!brous illite grown quite perpendicular to the tangential illite substratum. (Forinterpretation of the references to color in this !gure legend, the reader is referred to theweb version of this article.)

246 G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

thicknesswas targeted to be 50 !m. In practice, slice thicknesses are inthe range of 51–57 !m (Fig. 2) resulting in an inaccuracy of 6% fromthe target point. Thickness of the slices is accurately measured byimaging the top of the cross-section (Fig. 2). The inter-slice thickness(here about 50 !m thick) is too thick to give the exact interpolation ofmicrostructures along successive slices. Using our BIB cross-sectioner,the slice thickness can be reproducibly reduced down to 20 !m.

4.2. Grain fabrics and mineralogy

The samples show the typical features of tight, diageneticallyaltered aeolian sandstone from this part of the Central European Basin(Gaupp et al., 1993). Optical thin sections (Fig. 4.a) and ion beampolished cross-sections (Fig. 5.a) show that the diameter of the quartzgrains is about 0.2 mm with sub-angular to round grains. Angulargrains are rare and were only observed as “"oating” fragments inpores cemented by clay matrix (Figs. 7.e and 9.c). Grain fabrics aredominated by sutured and concavo/convex contacts possibly evolvedby dissolution processes and grain re-arrangement during compac-tion. EDX mappings made both on mechanically-polished thin-sections and BIB cross-sections reveal similar mineralogy (Fig. 6).Quartz grains dominate largely the mineralogical composition. Other

minerals such as calcite, halite and feldspars are also present but inmuch less amount.

Optical transmitted light microscopy (Fig. 4.a), EDXmeasurementsand electron microscopy performed on broken surfaces (Fig. 4.b andc) and BIB cross-sections corroborate the presence of diagenetic clayminerals between the sand grains (Bashari, 1998). In the vicinity ofnon-clay minerals, these are predominantly illite while kaolinite ismainly present in the region surrounding the feldspars probably dueto the replacement of original feldspar grains. Clay cementations arenot post-dated by any other cement. In optical thin sections (Fig. 4.a),only the biggest primary pores at triple junctions (N10 !m in size),decorated in blue by resin impregnation, are identi!ed. However,SEM observations on broken surfaces (Fig. 4.b) and mechanicallypolished thin-sections (Fig. 4.c) show that these primary pores arenot simply voids but complex pore microstructures based on thephyllo-mineral arrangement !lling the inter-granular pore space.Fig. 4.b shows that the illite displays two different morphologies:!brous/hairy or platy. The poor quality of mechanically preparedsurfaces (Fig. 4) does not allow the detailed characterization of clay-pore microstructures. In contrast, BIB-prepared cross-sections showdetails of the morphology of pores in clay down to 10 nm in size(Figs. 5 to 11).

Fig. 5. BSE micrographs of sand fabric overview. (a) The general fabric and mineralogy of the Bahnsen sample: calcite (Cc.), quartz (Qtz.), feldspar (Felds.), halite, illite and kaolinite(Kaol.). (b) and (c) Fractures features. Illite cement is distributed along each grain contacts forming interconnected network. The two kinds of fractures were identi!ed: F1-fracturesoccur at the vicinity of grain boundaries, propagated within the clay cementation and have “tear like” morphology. F2-fractures are systematically located within the grains withstraight morphology and angular connections at fractures nodes. Both set of fractures are opened and never cemented suggesting late formation.

247G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

Fig. 6. EDX mapping for chemical composition analysis. Grains forming the matrix are mainly quartz (Si+O) but calcite (Ca+C), feldspar (Si+K+Al) and halite (Na+Cl) are alsosigni!cantly present. Large amount of clay materials, as illite (K+Al+Fe+Mg) and kaolinite (Al+Si), cements the sandstone. Yellow color gives the highest relative specie contentand black the lowest. (For interpretation of the references to color in this !gure legend, the reader is referred to the web version of this article.)

Fig. 7. SE and BSE micrographs of pore microstructures at pore scale. (a) Pore located at junction formed by four detrital grains (three quartz grains in deep black and one feldspargrain in gray). This pore is early cemented by halite (white color) and postdated by illite cementation arranged tangential at the direct vicinity of the detrital grains edge and as“hairy/!brous” toward the pore center. (b) Same pore presented in (a) showing the apparent surface porosity. (c) and (d) detail from (b). In hairy/!brous illite, three types ofporosity are visible in (d) (see the text for details). (e) Pore located at junction formed by three detrital quartz grains. This pore is locally early cemented by halite (white color) andpostdated by illite cementation arranged tangential at the direct vicinity of the sand-grains edge and as “hairy/!brous” toward the pore center. We see also euhedral quartz grain(Qtz*) overgrown from sand-grain and angular quartz grain as “"oating” fragment (Qtz+). (f) Same region presented in (e) showing the apparent surface porosity. (g) Tangentialillite coating a calcite grain serving of substratum for hairy/!brous illite. (h) Pore space as appearing in hairy/!brous material showing in detail the meshwork formed by illite !bersaggregates close to the center of the pore shown in (f). (i) Detail of a single grain–grain contact !lled with illite: tangential is coating edges of the two adjacent sand-grains whilethe median part of the contact is !lled with hairy/!brous illite. A F1 fracture is passing through the illite !lling at contact between tangential and hairy/!brous illite. (j) Same regionpresented in (i) showing the apparent surface porosity. (k) Detail of apparent porosity as appearing in a large multiple grains junction. Edges of adjacent grains are coatedby tangential illite while hairy/!brous illite !lled the center of the intergranular volume. Euhedral quartz overgrowth (Qtz*) are embedded into the illite-rich region.

248 G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

249G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

By studying the relationships and microstructures between thedifferent mineralogical phases, it appears that calcite, halite, feldsparand minor quartz species are involved in the early cementation of

the sandstones. Euhedral quartz grains occur as overgrowth on quartzgrains (Figs. 7.e, 7.k and 11.c), halite !lls locally the intergranularvolume as “pervasive-intrusions” (Figs. 5.a, 7.a, 7.c, and 7.e) andcalcite and feldspars seem to !ll simply the intergranular volume.Feldspar grains (and much less commonly calcite and quartz) exhibitclear evidence of diagenetic dissolution and alteration (Figs. 8 and 9).Hematite grains (see Fig. 8.a) are sometimes found in the clay matrixbut rare; this is in good agreement with the white/gray color ofthe sample since hematite is responsible for the “red bed” color ofRotliegend sandstones (Torrent and Schwertmann, 1987).

4.3. Pore microstructures

In the following sub-sections, observations are only based on BIBcross-sections, which allow accurate and detailed investigation ofpore microstructures at pore scale.

4.3.1. Fracture porosityAs shown in Figs. 3, 5 and 11, both samples are signi!cantly

fractured. Two kinds of fractures are identi!ed. Both are open andnever cemented. The !rst set of fractures (F1) occurs at the boundariesof clastic grains, in the clay cement (Figs. 5.b and c) or cutting theclay–clastic grain contact. F1-fractures have jagged-walls. Segmenta-tion of the F1-fracture network (on single BIB cross-sections) andanalysis using Image J 1.38! (Abramoff et al., 2004) show that thesefractures represent 1–2% of the total polished cross-sections. Thesecond set of fractures (F2) is located within the clastic grains, withstraight walls and abutting against other fractures. F2-fracturescomprise 1–2% of the total polished areas. Thus, the total porosityby fracturing (F1+F2) is estimated equal to 2–4% of the total surfacerevealed by BIB cross-sectioning.

It is important to note here that in the case of the true 2D sectionsstudied here, the area fraction in the image is a good measure ofthe volume fraction (Underwood 1970; Russ and Dehoff, 2000). Insections with topography or !nite thickness this relationship rapidlygets lost (Krabbendam and Urai, 2003).

Serial cross-sectioning experiment (Figs. 3 and 11) shows thatfracturing affects the entire mass of samples and that F1 fractures areparticularly connected and promoted according to the distributionof illite cementation.

4.3.2. Non clay mineralsMinerals such as quartz, feldspars and calcite (Fig. 8) show

evidence of intra-crystalline porosity. Feldspar commonly exhibitsintra-crystalline porosity with the pores having a pseudo-cubicmorphology, which appears to be controlled by the crystal lattice.They are homogeneously distributed and have a range size of 200–500 nm. In calcite and quartz pores are rare, and occur as smallrounded holes of around 500 nm in diameter, interpreted to be "uidinclusions cut by the ion beam. Some bigger pores (up to few !m) alsooccur with more jagged morphology and !lled with visible hair-y/!brous illite (Fig. 8.d).

4.3.3. Clay mineralsFig. 7 presents images focused on illite-rich regions. SE images are

used for topography investigations while BSE images give informationabout the nature of phases and details of phases' arrangement. Allprepared samples exhibit similar microstructures and pore morphol-ogies in illite cementation, which occurs typically at grain contactsand !lling the intergranular volume.

Non-clay minerals are systematically coated with illite minerals,which are always tangentially organized to the surface of the sandgrain, with a compacted fabric and pores close to the limit of SEMresolution. When they are visible, these pores are about 500 nm longand less than 50 nmwide, elongated parallel to the adjacent surface ofnon-clay mineral.

Fig. 8. SE and BSE micrographs of intra-crystalline porosity observed in non-clayminerals. (a) and (b) In quartz, pores occur as small rounded hole of around 500 nm indiameter. Hematite grains are sometimes founded in the “hairy/!brous” illite but rare.(c) and (d) In calcite pore occurs also as small rounded hole of around 500 nm indiameter. Moreover, some of the largest pores are !lled with “hairy/!brous” illite (seedetail in d.). (e) In feldspar grain, pores have preferentially pseudo-cubic morphology,homogeneously distributed (200–500 nm in size) and free of illite.

250 G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

Fig. 9. Distribution and segmentation of porosity from typical and representative porous regions. Tangential illite (I), hairy/!brous illite (II), kaolinite (III), feldspar (IV) and bendedtangential illite (V). different regions are outlined by red dashed lines. (a) Different adjacent porous regions (I+II+III+IV). The mineral at the down right corner is calcite.(b) Segmented porosity as appears in micrograph (a) used to quantify and describe statistically the porosity. Each region is encoded by a different color (see the legend on !gure).(c) Typical evolution of porosity into a multiple grains junction region partially cemented by illite and "oating angular quartz fragment. Illite becomes much more porous towardsthe center of pore until it developed a pronounced hairy fabric. (I+II+III+V) regions are distinguished. (d) Segmented porosity as appears in micrograph (c) used to quantifyand describe statistically the porosity. Each typical region is encoded by a different color (see the legend directly on the !gure).

251G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

Toward the center of illite-rich regions, illite becomes signi!cantlymore porous. These clay particles have hairy/!brous morphologyforming a muddle meshwork with a slight tendency to be elongatedperpendicular to the tangential illite. Here, we distinguished 3 maintypes of pore morphology: (1) Type I — elongated pores betweensimilarly oriented clay sheets, (2) Type II — crescent-shaped pores insaddle reefs of folded sheet of clay and (3) Type III — large jaggedpores where the density of hairy/!brous illite is less. Type III pores aretypically N1 !m, Type II between 1 !m and 150 nm and Type Ib150 nm.

In our samples, illite-rich regions are mainly distributed alongsand–sand contacts as an interconnected network (Figs. 5.a, 6 and 10).The study of the serial cross-sectioning experiment (Figs. 3 and 11)shows the spatial variability of illite cement distribution in 3D,corroborating that illite cement is matching the outer part of grainsforming the sand by !lling all the intergranular volume. In detail,primary multiple grain junctions are mostly !lled by hairy illite whilesingle primary grain–grain contacts are typically fully !lled bytangential illite.

Another type of clay was identi!ed to be kaolinite, which is locallyfound in the region surrounding the weathered feldspar grains.Kaolinite-rich regions have porosity characteristics within the rangeof tangential illite and hairy/!brous illite. Pore morphologies ofkaolinite-rich regions are in the range of the Type I and II as describedabove for hairy/!brous illite. However, here the typical pore sizes aresigni!cantly smaller: Type II between 400 nm–100 nm and Type Ib50 nm. Fig. 9.a and b shows an example of the coexistence of illiteand kaolinite clays.

4.4. Quanti!cation of porosity in signi!cant porous regions

Based on observations of more than 50 positions in the sample,all the different types of pores related to identi!ed speci!c mineralsand regions described in Section 4.3 are very similar in all studiedpositions. Therefore, a detailed study of few selected positions (typicalexamples in Fig. 9) is taken to be representative for porosity contentof the speci!c regions (Table 1). Further quanti!cation of this is inprogress.

Figs. 9.a and b show porosity in adjacent areas formed by thesuccession of tangential illite, hairy/!brous illite, feldspar andkaolinite. The tangential illite has low porosity at the resolution ofSEM (2%). Kaolinite has a typical porosity of 5% and pore size of0. 8 !m; feldspars have a porosity of 16% and a pore size of 1.3 !m. Thehairy/!brous illite is the most porous region (24% of porosity inregions exclusively !lled with hairy/!brous illite) and pores around1.3 !m. Generally pores have elongated shape (axial ratio greaterthan 2) but less pronounced for feldspars and kaolinite. The shape

of pores in illite is individually controlled by the orientation of claysheets at contact forming pores of Type I, II or III as described in theSection 4.3.3 (Desbois et al., 2009).

Fig. 9.c and d show a second typical porous region for primarymultiple grain junctions formed by three grains of quartz partiallycemented by illite and a "oating angular quartz fragment. Averagedtotal porosity in multiple grain junctions is about 33%. Here, illite isorganized tangentially at edges developing more and more hairytowards the center of the triple junction. Between these two end-members, illite trends to become hairy by bending its tangentialorganization according to the curvature of the quartz edges. Thisgradual development of illite microstructures is observed andquanti!ed by the systematic increasing of the mean pore size andthe total porosity contribution towards the center (Table 1).

5. Discussion

Microstructural observations presented in this contribution are ingood agreement with the previous studies about illite characteriza-tion in reservoir sandstones. New insight is gained of porosity in clay-rich diagenetically altered samples by the preparation of extreme highquality, true 2D surface, revealing unprecedented detail of pore spaceand allowing quantitative analysis of the microporosity. Moreover,the use of the BIB instrument to produce serial cross-sections allowscharacterizing the fabric in 3D.

5.1. Microstructures and diagenesis

Though the aim of this contribution is not the study of diageneticprocesses occurring in Rotliegend tight gas reservoir sandstone, werecognized many of the twelve diagenetic types distinguished inGaupp (1996) and Schöner (2006). Following the interpretations inGaupp (1996) and Schöner (2006), studied samples give evidence forthe following phases in chronological order: Sebkha type (SB), Illitecoating type (IC), Hematite type (H), Feldspar leaching type (FL),Kaolinite type (K), Illite meshwork type (IM) and possibly the latequartz type (Q) for only the Bahnsen sample. Thus, our observationsare in good agreement with the general model of Rotliegendsandstone diagenesis (Macchi, 1987; McCann, 1998; Lanson et al.,2002; Ziegler, 2006; Abraham, 2007), and are interpreted to representcommon rock and diagenesis types from the Rotliegend.

5.2. Fracturing

The morphology of F1-fractures indicates that these fractures arelate and were formed by relaxation of stress (from over 200 MPato atmospheric) after drilling by core damage (Holt et al., 1994).

Although intra-crystalline fracturing in sandstones due to com-paction is a common process (Chester et al., 2004), the F2-fracturesare not compatible with the burial history because they are free ofcementation or pressure solution. This indicates that F2-fractures alsoformed late by core damage (Holt et al., 1994; Haimson, 2007).

Core damage is a common problem when evaluating reservoirporosity and correction for core damage is an as yet unsolvedquestion. Using the technique presented here, the porosity due to coredamage can be quanti!ed by image analysis.

5.3. Intragranular volume !lled by Illite

In sedimentary basins, illite is present as detrital matrix and asdiagenetic cement. Diagenetic illite forms after burial to signi!cant

Fig. 10. BSE micrograph of the entire !rst cross-section from the serial cross-sectioningexperiment showing the position of two selected regions of interest presented inFig. 11. Darker gray value indicates quartz grains while lighter gray values indicatemostly calcite, feldspars and calcite at the scale of micrographs.

Fig. 11. Serial SE highmagni!cationmicrographs of two selected regions of interest (Position 1 and position 2) cropped from the !ve successive cross-sections produced by the serialcross-sectioning experiment. These series shown the development of pores !lled by hairy/!brous illite and fracture (F1) feature with increasing depth in sample. Because of SEmicrograph, tangential illite is not visible in this !gure but !t the F1-fracture network. Some quartz overgrowths (Qtz*) are embedded in hairy/!brous illite-rich regions. The twoblack squares appearing in 2nd and 5th cuts from position 1 are missing pictures in the image set.

252 G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

253G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

depth (Macchi, 1987; Meunier and Velde, 2004; Schöner et al., 2008).Fibrous illite growth in the Central European Basin sandstones isepisodic as determined by K–Ar measurement (Wilkinson andHasseldine, 2002), rapid as suggested by their elongated morphology(Mullin, 1961) and their nucleation is almost exclusively uponpreexisting illite grain coating (Pollastro, 1985; Whitney and Velde,1993). Illite growth is a major factor in reducing the porosity andpermeability of reservoir rocks (Stadler, 1973; Seeman, 1979;Kantorowicz, 1990) and cause major problems during enhancedrecovery (Kantorowicz et al., 1986). Until now, direct quantitativeanalysis of the morphology of illite cement and its effect onpermeability (Pallat et al., 1984; Ziegler, 2006) has been dif!cult,because it was not possible to prepare 2D cross sections and 3Dmodels of porosity (Underwood, 1970; Desbois et al., 2009) andbecause it was not clear to what extent the fabric of the illite is alteredby the methods used to clean and prepare the samples for analysis(Rahman et al., 1995), or by the analysis itself (Hildenbrand and Urai,2003).

After drilling, the core samples studied by us were only subject toslow drying in air. Critical point drying (Huggett, 1982; DeWaal et al.,1986; Nadeau, 1998;) to eliminate the capillary forces during dryingwhich are known to possibly change the fabric of the thin illite !bers,was not possible.

In the studied samples, illite is the major clay mineral cementingthe grain fabric and !lling the intergranular volume (also primaryporosity). Illite is de!ned either by !brous or platy morphology(Seeman, 1979; Huggett, 1982; Nadeau et al., 1985). Illite cementa-tion reduces the intragranular volume by more than 60% in multiplegrain junctions and by more than 95% at single grain–grain contacts.These differences are attributed by different illite particle organiza-tion: (1) in multiple grains junctions, adjacent grains are coated byvery compact and poorly porous tangential illite (about 5 !m thick)while much more porous and connected hairy/!brous illite !lls theremnant space of the intergranular volume contributing to 70% ofporosity in this particular regions; (2) at single grain–grain contacts,intergranular volume is exclusively !lled by poorly porous tangentialillite.

From a 3D point of view given by the serial cross-sectioningexperiment (Fig. 11), intragranular volume appears then fully !lled byillite. When the intragranular volume is located at multiple grainjunctions, hairy illite is able to develop toward the center of theprimary pore from tangential illite substrate and results in relativehighly porous regions. When intragranular volume is located at only

two grain interfaces, hairy illite is rare and the intragranular volume isthen entirely !lled by tangential illite and results in very low porousregions. Therefore, the intragranular volume results in a network oflarge intragranular volumes mainly !lled with relative high poroushairy illite connected by poorly porous “intragranular volume throats”!lled with tangential illite.

According to Bushell (1986), illite arranged tangential to the grainsurfaces has less effect on permeability than perpendicular clayminerals and hairy/!brous illite has a lesser effect in reducing porositythan tangential illite.

5.4. Estimation of total porosity from micro-investigations — geometrichomogenization

Because our approach combining Ar-beam cross-sectioning andhigh-resolution SEM imaging enables the detection of pore micro-structure in a high quality true 2D cross-section, porosity at pore scalecan be segmented reliably and then quanti!ed and statisticallyinterpreted. This has the potential to model porosity by a combinationof characteristics at the scale of sand grains (intergranular volume)with information at much !ner scale to predict with macro-scalepore-permeability properties.

Microstructures were linked to bulk porosity by using a simpleporosity homogenization since porosity is only based on geometricconsiderations: the total porosity was estimated by the sum ofindividual porosity content of typical representative porous regions(examples in Fig. 9) convoluted by the percentage amount of each ofthese regions (Table 1).

Based on point counting performed on thin sections, our sampleconsists mainly of quartz (52%), calcite (10%), feldspars (6%),kaolinite (4%), illite cementation at narrow grain contacts (10%),big pores at triple junction (blue epoxy in Fig. 4.a) cemented by illite(15%) and fractures (average of 3%, see section 3.3.1). The porosity ofthese different typical representative porous regions was inferredfrom segmentation of porosity from our microstructural investiga-tions at pore scale (Table 1). Quartz and Calcite are assumed non-porous and we neglected other minor regions known to occur inthese samples. Table 1 summarizes data used for the estimation ofthe total porosity.

The total porosity (9.27%) estimated for the studied sample is ingood agreement with the bulk porosity (8.72%) of a similar samplemeasured on a core plug of about 10 cm3. At !rst, our study showsthat illite !lling the intra granular volume and fractures both

Table 1Quantitative analysis of porosity and estimation of total porosity from microstructural investigations.

Region Pore size (!m) Axial ratio Amounta

(%)Porosityb

(%)Contribution to totalporosity (%)

Bulk porosityc

(%)Estimated totalporosity (%)

Geomean Std. Geomean Std.

Quartz N.P N.P N.P N.P 52 0.00 0.00Calcite N.P N.P N.P N.P 10 0.00 0.00Feldspar 1.31 1.63 1.88 1.29 6 16.22 0.97Kaolinite 0.81 0.63 2.35 1.62 4 4.93 0.20Illite at single grain–grain contacts 10 2.00 0.20

Tangential illite 0.76 0.34 3.64 1.67 8.72 9.27Illite at multiple grains junctions 15 32.65 4.90

Tangential illite 0.61 0.53 3.65 1.95Bended tangential illite 0.96 1.38 4.17 1.84Hairy/!brous illite 1.37 3.07 4.51 2.02

Fractures 3 100.00 3.00

Geomean: geometric mean — most representative value for lognormal distribution.Std: standard deviation.N.P.: considered as non porous.

a Average inferred from optical thin-sections by point counting method.b Inferred by porosity segmentation of typical regions identi!ed from microstructural investigations at pore scale.c Measured porosity from drill-core, no overburden correction.

254 G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

contribute mainly to the porosity (table 1). In second, our micro-structural investigations at pore scale show also that illite cementsand fractures form an interconnected network. Therefore, this isproposed to explain why the total porosity estimated by ourmicrostructural investigations (9.27%) matches very well with theeffective (connected) porosity measured by mercury porosimetry(8.72%). The slight discrepancy between measured bulk porosity(8.72%, which is indeed the connected porosity) and our estimatedporosity (9.27%, i.e. apparent porosity in 2D cross-sections) may bedue to local heterogeneities in the sample which we missed in ourmicrostructural observations, errors in pore counting, or/and due toporosity in feldspar grains (contribution to total porosity estimatedaround 1%, Table 1) since our estimation includes it implicitly asconnected to clay minerals (Tangential and hairy/!brous illite+kaolinite) which are assumed to be connected and contributing inmajority to the overall permeability. Anyway, this discrepancy islower than 5% and could be considered as acceptable. Because inour estimation quartz and calcite grains were assumed non-porousthough few intra-crystalline pores were detected (Fig. 8) and becausemeasured and estimated porosity are in the same range, thesesuggests that pores in quartz and calcite are poorly connected to theoverall fabric and thus do not contribute signi!cantly to the effectiveporosity.

5.5. Fabric and pore 3D-model

This contribution gives a !rst impression of the complex micro-porosity at pore scale in tight gas sandstones. Fig. 12 presents a3D model of fabric and porosity occurring in the studied sample(Rotliegend sandstone, Bahnsen member, Arsten graben, Bremen,Germany) compiling all microstructural information detailed above.

The model (Fig. 12) presents a Bahnsen sample with grain matrixcomprising quartz, calcite and feldspar partially weathered intokaolinite. The intergranular volume, corresponding to initial free-space between the grains forming thematrix, is cemented by euhedralquartz overgrowth, halite and calcite; the remnant free space is fully

!lled by illite. Tangential illite coats systematically grains forming thematrix over a thickness of 1–5 !m; at multiple grains junctions, thereis enough free-space left for hairy/!brous illite development.Therefore in 3D, the intragranular volume appears then fully !lledby illite with highly porous regions (corresponding to hairy illite atmultiple grains junctions) connected by poorly porous “intragranularvolume throats” (corresponding to tangential illite at single grain–grain contacts).

SEMmicrographs were added in this model to re"ect the diversityof typical pore morphologies observed in our samples. Feldspar grainscommonly exhibit intra-crystalline pores with pseudo-cubic mor-phology. In calcite and quartz, pores originate from "uid inclusionsappearing as small rounded holes. Some bigger pores occur also incalcite grains with more jagged morphology and !lled with visiblehairy/!brous illite. Pores in tangential illite are elongated and layingbetween two clay sheet aggregates, while hairy/!brous illite displays3 types of pores whose shapes are controlled by the organization ofclay sheet aggregates (Type I, II and III). Kaolinite-rich regions bearpores with similar morphologies as found in hairy/!brous illiteregions (Type I and II) but with smaller pore size.

Fractures identi!ed in our investigations were not included inthis model since they are not primary but induced by core samplecollection.

This 3D view of porosity fabric in a tight gas reservoir presentsthe concept of porous !ll of the Intergranular Volume with relevantimplications for "uid "ow simulation in sandstone reservoir. Thelattice-Boltzmann method was applied successfully on clean sand-stone (Sholokhova et al., 2009) with intragranular volume free of claycements. For tight gas reservoir sandstone, this simulation approach ismore challenging since the intergranular volume is cemented withpermeable clay (Tölke et al., 2010). The model presented here formsthe basis for incorporating the transport properties of the pore !ll into"uid "ow simulations in tight gas reservoirs, by assigning represen-tative permeabilities to the different diagenetic pore !lls of theIntergranular Volume, and incorporating this information in networkmodels of the intergranular volume as determined by micro-CT.

Fig. 12. Fabric and porosity 3D-model in an idealized tight gas sandstone reservoir from Rotliegend formation (Bahnsen member, Arsten Graben, south of Bremen, Germany). Thismodel compiles all microstructural information detailed in this contribution. See text for details.

255G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

6. Conclusion

The use of the Argon beam for polished cross-sections preparationproduces very high quality sections without damage and artifact toinvestigate quantitatively the pore network in tight gas sandstones atthe-state-of-the-art SEM resolution. This method has the potential tobring out a new concept for porosity investigation in tight gassandstones by bridging the information based on microstructures atpore scale with macro-scale properties. It offers also the basis for afundamental understanding of pore geometry, "uid "ow, and, incombination with cryogenic techniques (Holzer et al., 2007, Desboiset al., 2008; Desbois et al., 2009) will offer insight in the geometry andwetting characteristics of in-situ "uid phases in sandstone reservoirs.

Acknowledgments

This paper reports work done in the Wintershall Tight GasConsortium at RWTH Aachen University. We thank WintershallHolding AG for help with obtaining the Rotliegend sandstone coresample, and for funding the project. We are grateful to AlexanderSchwedt (GFE at RWTH Aachen University, Aachen, Germany) fortechnical support for the SEM imaging, Werner Krauss and ChristianDiebel (†) for preparation of the samples; Katharina Albert, JochenHürtgen and Prokop Zavada for image processing and Simon Virgoand Max Arndt for help with Autopano and ArcMap softwares.

References

Abraham, P., 2007. Relations between seismic signals and reservoirs properties of deepgas reservoirs in Northwest-Germany–Wustrow member, Rotliegend. Friedrich-Schiller University, Jena, Germany. 153p.

Abramoff, M.D., Magelhaes, P.J., Ram, S.J., 2004. Image processing with ImageJ.Biophotonics International 11 (7), 36–42.

Bachmann, G.H., Hoffmann, N., 1997. Development of the Rotliegend Basin in northernGermany. Reihe D: Mineralogie, Petrographie, Geochemie, Lagerstaettenkunde:Geologisches Jahrbuch., 103, pp. 9–31.

Baraka-Lokmane, S., Main, I.G., Ngwenya, B.T., Elphick, S.C., 2009. Application ofcomplementary methods for more robust characterization of sandstone cores. Mar.Pet. Geol. 26, 39–56.

Bashari, A., 1998. Diagenesis and reservoir development of sandstones in the Triassicrewan group, Bowen basin, Australia. J. Pet. Geol. 21 (4), 445–465.

Börmann, C., Gast, R., Görisch, F., 2006. Structural and sedimentological analysis of anearly Late Rotliegendes graben based on 3D seismic and well log data, GermanNorth Sea. Pet. Geosci. 12, 195–204.

Bushell, T.P., 1986. Reservoir geology of the Morecambe Field. In: Brooks, J., Goff, J.C.,van Hoorn, B. (Eds.), Habitat of Palaeozoic Gas in N.W. Europe: Special Publicationof the Geological Society, London, 23, pp. 189–208.

Chester, J.S., Lenz, S.C., Chester, F.M., Lang, R.A., 2004. Mechanisms of compaction ofquartz sand at diagenetic conditions. Earth Planet. Sci. Lett. 220, 435–451.

De Waal, J.A., Kantorowicz, J.D., Dicker, A.I.M., 1986. Petrophysical core analysis ofsandstones containing delicate illite. Transactions of the l0th European FormationEvaluation Symposium, April 1986. Paper Z. 21 pp.

Desbois, G., Urai, J.L., Burkhardt, C., Drury, M.R., Hayles, M., Humbel, B., 2008. Cryogenicvitri!cation and 3D serial sectioning using high resolution cryo-FIB SEM technologyfor brine-!lled grain boundaries in halite: !rst results. Geo"uids 8 (1), 60–72.

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-Earth 4,15–22.

Dvorkin, J., 2009. Digital rock physics bridges scales of measurement. Rock Physics. E&P.Sept. 2009.

Erdmann, N., Campbell, R., Asahina, S., 2006. Precise SEM cross-section polishing viaArgon beam milling. Microsc. today 2006, 22–25 May 2006.

Favvas, E.P, Sapalidis, A.A., Stefanopoulos, K.L., Romanos, G.E., Kanellopoulos, N.K.,Kargiotis, E.K., Mitropoulos, A.Ch., 2009. Characterization of carbonate rocks bycombination of scattering, porosimetry and permeability techniques. Microporousand Mesoporous Materials 120, 109–114.

Gaupp, R., 1996. Diagenesis types and their application in diagenesis mapping.Zentralbl. Geol. Palaeontol. Teil 1 1994 (11/12), 1183–1199.

Gaupp, R., Matter, A., Platt, J., Ramseyer, K., Walzebuck, J.P., 1993. Diagenesis and "uidsevolution of deeply buried Permian (Rotliegende) gas reservoir, NorthwestGermany. AAPG Bull. 77 (7), 1111–1128.

George, G., Berry, J.K., 1993. A new palaeogeographic and depositional model for theUpper Rotliegend of the UK Sector of the Southern North Sea. In: North, C.P.,Prosser, D.J. (Eds.), Characterization of Fluvial and Aeolian Reservoirs: Geol SocLondon, Spec Publ, 73, pp. 291–319.

Glennie, K.W., 1986. Development of N.W. Europe's Southern Permian Gas Basin. In:Brooks, J., Goff, J.C., Van Horn, B. (Eds.), Habitat of Paleozoic Gas in N.W. Europe:Geological Society Special Publication, 23, pp. 3–22.

Glennie, K.W., 1990. Introduction to the Petroleum Geology of the North Sea, 3rd ed.Blackwell Scienti!c, Oxford.

Haimson, B., 2007. Micromechanisms of borehole instability leading to breakouts inrocks. Int. J. Rock Mech. Min. Sci. 44, 157–173.

Haines, L. (Ed.), 2005. Tight Gas. Supplement to Oil and Gas Investor, March.Hildenbrand, A., Urai, J.L., 2003. Investigation of the morphology of pore space in

mudstones — !rst results. Mar. Pet. Geol. 20 (10), 1185–1200.Holditch, S.A., 2006. Tight gas sands. JPT 58 (6), 86–93.Holt, R.M., Brignoli, M., Fjaer, E., Unander, T.E., Kenter, C.J., 1994. Core damage effects on

compaction behaviour. Proc. SPE/ISRM Rock Mechanics in Petroleum Engineering.SPE 28027. Balkena, Rotterdam, Delft, The Netherlands, pp. 55–62.

Holzer L., Cantoni M., (2011). Review of FIB-tomography. In: Nanofabrication with FEBand FIB. P. Russell, I. Utke and S. Moshkalev, Oxford University Press: in press.

Holzer, L., Gasser, Ph., Kaech, A., Wegmann, M., Zingg, A., Wepf, R., Münch, B., 2007.Cryo-FIB-nanotomography for quantitative analysis of particle structures in cementsuspensions. J. Microsc. 227, 216–228.

Holzer, L., Münch, B., Rizzi, M., Wepf, R., Marschall, P., 2010. 3D-microstructure analysisof hydrated bentonite with cryo-stabilized pore water. Appl. Clay Sci. 47, 330–342.

Hornapour, M.M., Djabbarah, N.F., Sampath, K., 2003. Whole core analysis-experienceand challenges. Presented for presentation at the SPE 13th Middle East Oil Show &Conference, Barhain 5–8 April 2003, SPE 81575, pp. 1–16.

Huggett, J.M., 1982. On the nature of !brous illite as observed by electron microscopy.Clay Miner. 17, 433–441.

Kantorowicz, J.D., 1990. The in"uence of variation in illite morphology on thepermeability of Middle Jurassic Brent Group sandstones, Cormorant Field, UKNorth Sea. Mar. Pet. Geol. 7, 66–74.

Kantorowicz, J.D., Lievaart, L., Eylander, J.G.R., Eigner, M.R.P., 1986. The role ofdiagenetic studies in production operations. Clay Miner. 21, 769–780.

Krabbendam, M., Urai, J.L., van Vliet, L.J., 2003. Grain size stabilisation by dispersedgraphite in a high-grade quartz mylonite: an example from Naxos (Greece).J. Struct. Geol. 25 (6), 855–866.

Lanson, B., Beaufort, D., Berger, G., Bauer, A., Cassagnabère, A., Meunier, A., 2002.Authigenic kaolin and illitic minerals during burial diagenesis of sandstones: areview. Clay Miner. 37 (1), 1–22.

Law, B.E., Curtis, J.B., 2002. Introduction to unconventional petroleum systems. AAPGBull. 86 (11), 1851–1852.

Legler, B., 2005. Die Entwicklung des Südlichen Permbeckens in Abhängigkeit vonTektonik, eustatischen Meeresspiegelschwankungen des Proto-Atlantiks undKlimavariabilität (Ober-Rotliegend, NW Europa). Schriftenreihe Dt Ges Geowiss.47, 103.

Lexa, O., !típská, P., Schulmann, K., Baratoux, L., Kröner, A., 2005. Contrasting texturalrecord of two distinct metamorphic events of similar P–T conditions and differentdurations. J. Metamorph. Geol. 23 (8), 649–666.

Littke, R., Bayer, U., Gajewski, D., Nelskamp, S., 2008. Dynamics of ComplexIntracontinental Basins — The Central European Basin System. Springer-Verlag,Berlin-Heidelberg978-3-540-85084-7.

Loucks, R.G., Reed, R.M., Ruppel, S.C., Jarvie, D.M., 2009. Morphology, genesis, anddistribution of nanometer-scale pores in siliceous mudstones of the MississippianBarnett Shale. J. Sediment. Res. 79, 848–861.

Macchi, L., 1987. A review of sandstone illite cements and aspects of their signi!canceto hydrocarbon exploration and development. Geol. J. 22, 333–345.

McCann, T., 1998a. Sandstone composition and provenance of the Rotliegend of the NEGerman Basin. Sediment. Geol. 116 (3–4), 177–198 March 1998.

McCann, T., 1998b. The Rotliegend of the NE German Basin: background andprospectivity. Pet. Geosci. 4, 17–27.

Menendez, B., David, C.,Martinez,NistalA., 2001. Confocal scanning lasermicroscopy appliedto the study of pore and cracks networks in rocks. Comput. Geosci. 27, 1101–1109.

Meunier, A., Velde, B. (Eds.), 2004. Illite: Origins, Evolution, and Metamorphism.Springer. ISBN: 9783540204862. 281 pp.

Mullin, J.W., 1961. Cristallisation. Butterworths, England.Nadeau, P.H., 1998. Fundamental particles and the advancement of geoscience:

response to “implications of TEM data for the concept of fundamental particles”.Can. Mineral. 36, 1409–1414.

Nadeau, P.H., Hurst, A., 1991. Application of back-scattered electron microscopy to thequanti!cation of clay mineral microporosity in sandstones. J. Sediment. Res. 61 (6),921–925.

Nadeau, P.H., Wilson, M.J., McHardy, W.J., Tait, J., 1985. The conversion of smectite toillite during diagenesis: evidence from some illitic clays from bentonites andsandstones. Mineral. Mag. 49, 393–400.

Pallat, N., Wilson, J., McHardy, B., 1984. The relationship between permeability and themorphology of diagenetic illite in reservoirs rocks. J. Pet. Technol. 36, 2225–2227.

Pape, H., Tillich, J.E., Holz, M., 2005. Pore geometry of sandstone derived from pulsed!eld gradient NMR. J. Appl. Geophys. 58, 232–252.

Plein, E., 1995. Norddeutsches Rotliegendbecken, 183 (pp. 1±193). Courier Forsch.-lnst. Senkenberg.

Pollastro, R.M., 1985. Mineralogical and morphological evidence for the formation ofillite at the expense of illite/smectite. Clays Clay Miner. 33, 265–274.

Rahman, S.S., Rahman, M.M., Khan, F.A., 1995. Response of low-permeability, illiticsandstone to drilling and completion "uids. J. Pet. Sci. Eng. 12 (4), 309–322.

Russ, J.C., Dehoff, R.T., 2000. Practical Stereology, 2 ed. Kluwer Academic/Plenum, NewYork.

Schöner, R., 2006. Comparison of Rotliegend Sandstone Diagenesis from the Northernand Southern Margin of the North German Basin, and Implications for the

256 G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257

Importance of Organic Maturation and Migration. Friedrich-Schiller University,Jena, Germany. 160p.

Schöner, R., Gaupp, R., 2005. Contrasting red bed diagenesis: the southern and northernmargin of the Central European Basin. International Journal of Earth Sciences, 94,(5–6). Springer, Berlin/Heidelberg, pp. 897–916. doi:10.1007/s00531-005-0004-3.

Schöner, R., Lüders, V., Ondrak, R., Gaupp, R., Möller, P., 2008. Fluid-rock interactions. In:Littke, R., Bayer, U., Gajewski, D., Nelskamp, S. (Eds.), Dynamics of ComplexIntracontinental Basins — The Central European Basin System, 1–12. Springer-Verlag, Berlin-Heidelberg. ISBN: 978-3-540-85084-7.

Seeman, U., 1979. Diagenetically formed interstitial clay minerals as a factor inRotliegend sandstone reservoir quality in the North Sea. J. Pet. Geol. 1, 55–62.

Sholokhova, Y., Kim, D., Brent Lindquist, W., 2009. Network "ow modeling via lattice-Boltzmann based channel conductance. Adv. Water Res. 32, 205–212.

Stadler, P.J., 1973. In"uence of crystallographic habit and aggregate structure ofauthigenic clay minerals on sandstone permeability. Geol. Mijnb. 53, 217–220.

Strömbäch, A.C., Howell, J.A., 2002. Predicting distribution of remobilized aeolian faciesusing sub-surface data: theWeissliegend of the UK Southern North Sea. Pet. Geosci.8, 237–249.

Tiab, D., Donaldson, E.C., 1996. Petrophysics Theory and Practice of Measuring ReservoirRock and Fluid Transport Properties. Gulf Publishing Company, Houston, TX.706 pp.

Tölke, J., Baldwin, C., Mu, Y., Derzhi, N., Fang, Q., Grader, A., 2010. Computer simulationsof "uid "ow in sediment: from images to permeability. Lead. Edge 29, 68–74.

Tono, H., Ingrain, 2008. Computing Properties from 3D Imaging. E&P Magazine,November 2008. Hart Energy Publishing, L.P.

Torrent, J., Schwertmann, U., 1987. In"uence of hematite on the color of the red beds.J. Sediment. Petrol. 57 (4), 682–686.

Underwood, E.E., 1970. Quantitative Stereology. Addison-Wesley, Reading, Massachu-setts, USA978-0201076509.

VanWess, J.D., Stephenson, R.A., Ziegler, P.A., Bayer, U., McCann, T., Dadlez, R., Gaupp, R.,Narkiewicz, M., Bitzer, F., Scheck, M., 2000. On the origin of the Southern PermianBasin, Central Europe. Marine Petrol. Geol. 17, 43–59.

Ward, C.R., Nunt-jarunwong, S., Swanson, J., 2005. Use of mineralogical analysis ingeotechnical assessment of rock strata for coal mining. Int. J. Coal Geol. 64,156–171.

Whitney, G., Velde, B., 1993. Changes in particle morphology during illitization – anexperimental study. Clays Clay Miner. 41, 209–218.

Wilkinson, M., Haszeldine, R.S., 2002. Fibrous illite in oil!eld sandstones— a nucleationkinetic theory of growth. Terra nova 14, 56–60.

Ziegler, K., 2006. Clay minerals of the permian Rotliegend group in the North Sea andadjacent areas. Clay Miner. 41, 355–393.

Zwingmann, H., Gaupp, R., Higgs, K., Zhang, Y., 2008. Diagenesis evaluation insedimentary basins for hydrocarbon exploration. Geochim. Cosmochim. Acta 72(12), A1109–A.

257G. Desbois et al. / Journal of Petroleum Science and Engineering 78 (2011) 243–257