Clay and clay Minerals, Vol. 47 No. 3, 269-285, 1999
 

                    Lhoussain Hassouta, Martine D. Buatier, Jean-Luc Potdevin and Nicole liewig *
 

Université Lille 1, URA 719, Laboratoire de Sédimentologie et Géodynamique, 59655 Villeneuve d’Ascq, France

* Centre de Géochimie de la Surface, CNRS, 1, rue Blessig , 67084 Strasbourg Cedex, France

Corresponding author : Lhoussain Hassouta, Université de Lille 1, URA 719, Laboratoire de Sédimentologie et Géodynamique, 59655 Villeneuve d’Ascq, FRANCE

Email: Lhoussain.Hassouta@univ-lille1.fr

Abstract.

The nature, composition, and relative abundance of clay minerals in the sandstones of the Brent Group reservoir have been studied between 3200 and 3300 m in a well of the Ellon Field (Alwyn area, North Sea). They have a heterogeneous calcite cement which occurred during early-diagenesis. Clay diagenesis of the cemented and uncemented sandstones has been investigated using optical microscopy, scanning electron microscopy (SEM), X-ray diffraction analyses (XRD) and infrared spectroscopy. The influence of cementation on clay neoformation is demonstrated in this study. Detrital illite and authigenic kaolinite are present in both the calcite cemented and uncemented sandstones suggesting that kaolinite precipitated before calcite cementation. In the uncemented sandstones, blocky dickite replaces vermiform kaolinite with increasing depth. At 3205 m, authigenic illite begins to replace kaolinite and shows progressive morphological changes (fibrous to lath-shape transition). At 3260 m, all sandstones are not cemented by calcite. and Iillite which is the only clay mineral in these samples and showsdisplays a platelet morphology .

In the cemented samples, vermiform kaolinite is preserved at all depths suggesting that dickite transformation has been inhibited by the presence of the calcite cement. This observation suggests that calcite cement would prevent fluid circulation and dissolution-precipitation reactions.

Key-Words. - North Sea, diagenesis, sandstone, illite, dickite, kaolinite, calcite, pressure-solution.
 

Introduction

Clay diagenesis can greatly change the physical properties of sandstones in petroleum reservoirs. Precipitation of clay minerals in the pore spaces and the evolution of clays through burial and temperature have been described by many authors; for instance, in the Brent Group reservoirs of the (Middle Jurassic) North Sea, the evolution of illite and kaolin morphologies with depth and temperature is well documented (Kantorowicz 1984; Thomas 1986; Glasmann et al. 1989a; Giles et al. 1992; Haszeldine et al. 1992). Ehrenberg et al. (1993) and Lanson et al. (1996) demonstrated the relationship between kaolin morphology (i.e., vermicular or blocky kaolin) and the nature of the kaolin polytype (i.e., kaolinite or dickite). In the Brent Group reservoirs (Middle Jurassic), illite replaces kaolinite and muscovite and shows morphological changes with burial depth. The illitization of kaolinite was described by Bjørlykke (1983) by the reaction: Kaolinite + K-feldspar ---> Illite + Quartz. According to Ehrenberg et al. (1993) and Lanson et al. (1996), the kaolin polytype evolution and the illitization reaction are temperature dependant. The purpose of this paper is to describe clay diagenesis in a sandstone reservoir of the Brent Group in the North Sea which is heterogeneously cemented by calcite. We compare clay diagenesis in early calcite cemented sandstones with sandstones in which the calcite cement was absent during burial diagenesis. These comparisons allow discussion of the mechanisms of illite precipitation and the kaolinite to dickite transition with reference to depth, lithology and cementation.

Geological setting

The studied sandstones were collected from a well of the Ellon Field (North Sea), located in the Great Alwyn area ( Fig1.GIF ). The Alwyn area is a intermediate faulted terrace between the East Shetland Platform and the Viking Graben. The structure of the Ellon Field is mainly due to rifting of the Viking Graben in two stages (Faure 1990). The first stage occurred from Permian to Triassic. The second stage is Jurassic and resulted in tilted block faulting. Then an important erosion and a meteoric fluid flow episode occurred during the Cimmerian phase (Sommer 1978). A new stage of tectonic extension occurred during the late Cretaceous with a high sedimentation rate which sealed the basin structure. The studied sandstones belong to the Brent Group which contains the most prolific petroleum reservoirs in the northern North Sea Basin (Blanche and Whitaker 1978, Hancock and Taylor 1978, Giles et al. 1992). The Brent Group is composed of five Jurassic formations (late Bajocian to early Bathonian): Broom, Rannoch, Etive, Ness and Tarbert (Bowen 1975, Deegan and Scull 1977). From a sedimentological point of view, the Brent sequence records the progression from the coastal shallow-marine environments of the Lower Brent Group, to the alluvial or deltaic interbedded sands, shales and coals of the Middle Brent, and ends with the transitional to marine sands of the Upper Brent Group (Jourdan et al. 1987). In the studied well, only the Ness B and Tarbert Formations have been cored (Figure 2). The Ness B is divided into a lower tide-dominated section and an upper section of three shoreface-tidal complexes. These two parts are separated by thin lagoonal or coaly marsh facies. The lower Tarbert Formation comprises two upper shorefaces separated by a thin transgressive lag. The upper Tarbert Formation consists of a thick transgressive lag (Figure 2). The sandstones from the Ness and Tarbert formations are heterogeneously cemented by calcite from 3200 to 3260 m. Below 3260 m, the calcite cement is absent.

Sampling approach

Sampling has focused on comparison of calcite cemented sandstones and sandstones where the calcite cement is absent ("uncemented" sandstones). Both sandstones were sampled at different depths in the Ness and Tarbert Formations of the Brent Group (Figure 2 and Table 1). Calculated and measured porosities of the studied sandstones are about 20% in the uncemented sandstones and less than 3% in the cemented sandstones (Potdevin and Hassouta 1997). Considering the calcite content of the cemented sandstones, the initial porosity before cementation was about 40%. This high initial porosity and the sparitic texture of the cement suggest that calcite cementation was an early diagenetic process occurring at shallow depth as suggested by Zieglar and Spotts (1978). The calcite cementation is heterogeneous. The contact between cemented and uncemented sandstones is always very sharp (Figure 3) and does not seem to correspond to an initial heterogeneity but may be a dissolution or a precipitation front. The early calcite cementation and the very low porosity of the cemented sandstones preserved the sandstones grains from water-rock interactions and fluid flow during burial. Studying the cemented and uncemented sandstones allow to compare clay burial diagenesis in water present and water absent sandstones of the same initial composition. Twenty four pairs of adjacent cemented and uncemented sandstones have been sampled to compare clay diagenesis as a function of depth, lithology and cementation (Table 1).
 
 

Methods and analytical techniques

Thin sections were examined using petrographic and cathodoluminescence microscopes. A Technosyn Cold Cathodoluminescence Model 8200 MkII operating at 15-20 kv and 350 µA was used. Bulk samples were also observed by scanning electron microscopy (SEM) at 15 kv with a Cambridge Stereoscan 240. Samples located in the oil zone were washed at the Total-Beauplan Centre to remove organic matter (hydrocarbons) from the pore spaces. The clay size fractions (<2 µm and <5 µm) were separated from bulk samples by settling in a water column. Prior to separation, each sample was dispersed in deionized water, disaggregated and decarbonated with a 0.2 M HClsolution and washed several times. Clay mineral assemblages were determined by X-ray diffraction (XRD) on oriented and powdered samples with a Philips PW 1729 diffractometer using Cu Ka radiation and a Ni filter with a scan speed of 2° 2q /min. The modelling program of Lanson and Besson (1992) was used to analyze XRD patterns of clay minerals in the range 5 to 12.5° 2q. This program allows the determination of the composition of the clay mineral fraction by fitting the XRD pattern to the sum of the theoretical individual patterns of different clay minerals. Infrared spectra were recorded in the 4800-200 cm^-1 range on a Nicollet 510 infrared spectrophotometer using pressed disks prepared by mixing 4 mg of sample with 300 mg of KBr. Chemical analyses of selected minerals were performed using the electron microprobe Camebax, SX 50 model, of the Camparis Center of Paris VII University. Standards were natural silicates and oxides, the accelerating voltage was 15 kv, the counting time 10-20 s, the current intensity 15 nA and the spot size 1 µm.

Results

Petrography

The bulk mineralogy is very similar in all sandstones (Table 2). The detrital grains are quartz, feldspars (albite and microcline), mica and various types of lithic fragments. Mineralogic variations concern principally the diagenetic phases which are calcite cement, quartz overgrowth, pyrite, kaolin and illite. Minor ankerite was also detected by XRD. Calcite and ankerite are only present in cemented sandstones and are absent below 3260 meters. Kaolin minerals occur in cemented and uncemented sandstones above 3260 m. Quartz overgrowths and illite are very rare in cemented sandstones but abundant in uncemented sandstones located below 3205 m.

QUARTZ. Quartz represents 50-70% of the sandstones. It is mostly detrital but quartz overgrowths are common in the uncemented sandstones. The detrital core is rimmed by one or two stages of diagenetic quartz reducing the pore space between grains. Overgrowths give an euhedral shape to quartz grains as shown on optical microscope images (Figure 4A). Indented and stylolitized contacts between quartz grains indicate that pressure-solution was an efficient mechanism decreasing rock porosity. Quartz overgrowths are absent in the calcite cemented sandstones.

FELDSPAR. K-feldspar is the most abundant feldspar but its concentration depends on the sedimentary facies. According to modal analyses, K-feldspar represents 4-5% of the rock volume in upper shoreface but 1 to 7% in lower shoreface facies. The amount of K-feldspar also varies with increasing depth and cementation. When depth increases, most feldspars display dissolution marks in uncemented sandstones (Figure 4B). In calcite cemented sandstones, K-feldspars are well preserved (Figure 4C) and can display an euhedral shape which results in thin K-feldspar overgrowths. The microprobe analyses show a very pure K-feldspar composition of this overgrowths. They are black (nonluminescent) in cathodoluminescence indicating a very low trace elements content which results of a low temperature of crystallization (i.e., <100°C, Marshall 1988) that confirms their authigenic origin. The important dissolution of K-feldspar explain that K-feldspar overgrowths are not observed in uncemented sandstones. The lack of overgrowths could result in important dissolution of the K-feldspar grains in these sandstone..

DETRITAL MICA. Detrital muscovite is present in all samples. Its amount is not related to cementation, but dependant on the sedimentary facies. Undulatory extinction and flexed mica crystals in the uncemented sandstones under the optical microscope their deformation during compaction. In calcite cemented sandstones, mica grains are well preserved with minor kaolinitization developed only at the grain boundaries (Figure 4D), whereas in uncemented sandstones the micas display open layers filled by kaolin and illite. In some crystals, these two minerals entirely replace the mica.

CALCITE. In cemented sandstones, a calcite cement fills the primary porosity. It occurs as large crystals in which detrital grains are enclosed. Despite evidence of intense recrystallization, cathodoluminescence reveals growth zones which show the sparitic nature of the early cement. The microprobe analyses show that the calcite crystals havethe structural formulae: Ca_0.948Mg _0.047 Mn_0.004 Sr_0.001Ba _0,001CO₃. Some relics of the calcite cement have been found inside quartz overgrowths suggesting a dissolution episode of the calcite cement prior to quartz cementation.

KAOLIN MINERALS. Three morphologies of kaolin minerals have been observed by SEM.

- Vermiform kaolin which fill the primary porosity (Figure 5A).

- Cluster kaolin (Figure 5B) which form adjacent to, or between expanded layers of detrital mica. This morphology is common in the Ness Formation, particularly in the shoreface facies, and is probably due to the higher concentration of muscovite in these sandstones.

Vermiform and cluster morphologies are present both in cemented and uncemented sandstones suggesting that kaolinite precipitation preceded calcite cementation.

- Blocky kaolin (Figure 5C) are present in the deeper uncemented samples (3 3232 meters) but absent in the corresponding cemented sandstones and very rare in impermeable coaly marsh facies. These crystals are pseudo-hexagonal in shape, 5 to 10 µm in thickness with a diameter of 10 to 20 µm. This morphology is absent in all calcite cemented sandstones.

SEM observations of the kaolin morphologies in the uncemented sandstones show a progressive change with depth. In the most shallow samples (3200 m), aggregates of vermiform kaolin crystals were observed, about 80-100 µm, and the diameter of the vermicules is smaller than 10 µm. Deeper in the core (about 3204 m), the vermiform crystals are progressively replaced by larger crystals (diameter > 10 µm) (Figure 5D). Between 3254 and 3260 m, the blocky morphology is dominant. In the cemented sandstones, only vermiform and cluster kaolinite crystals are observed. Below 3260 m, no kaolin was observed by SEM.

DIAGENETIC ILLITE. Diagenetic illite is only present in the uncemented sandstones. It occurs as aggregates of small crystals filling the pore spaces. Illite can replace the kaolin minerals and detrital muscovite. In SEM images , it can be seen that kaolinite aggregates are replaced by fibrous illite. The SEM images show that illite morphology varies with depth. Filamentous illite (Figure 6A) is observed in the shallowest samples (3204 m), it is followed by lath morphology (3238 m) (Figure 6B). Both morphologies occur with kaolin minerals until 3260 meters depth. Deeper, illite is the only clay mineral identified. It appears as platelet crystals (Figure 6C) which fill the pores and replace detrital mica.

Clay mineralogy (XRD)
 

The clay fraction (< 5 µm) consists mainly of illite and kaolin (Figure 7). Chlorite is rare and is absent in the upper part of the Tarbert Formation. Kaolins are characterized by the (001) and (002) reflections respectively at 7.14 Å or ; 7.22 Å and 3.58 Å. Illite is characterized by the presence of three peaks at 10 Å, 5 Å and 3.3 Å corresponding respectively to the reflection along the (001), (002) and (003) planes. In all of the cemented samples, kaolinite is the major phase in the clay fraction, whereas in the uncemented samples, the intensity of the illite peaks increases with depth. Below 3260 meters, illite is the only clay present in the clay fraction (Table 2).

Kaolin polytype characterization

X-ray diffraction DATA. The random mounts of the less than 5 µm fraction of the samples were analyzed to characterize the kaolin polytype present in the sandstones. The polytype identification can be difficult because of the superposition of the quartz and feldspar reflections with those of the kaolin minerals (Ehrenberg et al. 1993). However, some of the characteristic reflections could be used to distinguish the various polytypes. In the cemented samples, kaolinite is the only kaolin polytype and is characterized by reflections at 7.21, 3.58, 3.136, 3.32 and 2.34 Å (Figure 8A). In the uncemented samples, the XRD patterns display mineralogical changes with depth (Figure 9 and Table 2). Samples located between 3200 and 3223 m contain kaolinite. Below 3223 m, the occurrence of a reflection at 4.13 Å indicates the presence of dickite. In samples where dickite is the major kaolin polytype, reflections at 2.32 Å and 4.13 Å are clearly visible (Figure 8B and 9).

IR SPECTROSCOPY. More accurate distinction between kaolinite and dickite can be performed by assessing the position and relative intensity of the OH-stretching bands in the 3600-3700 cm-1 region of infra-red patterns (IR). Well crystallized kaolinite shows a strong absorption at 3697 cm-1, a band of medium-strong intensity at 3620 cm-1 and two bands with relatively weak intensity at 3669 and 3652 cm-1. On the other hand, dickite shows a strong absorption band at 3621 cm-1 and two medium-strong bands at 3704 and 3654 cm-1 (Ehrenberg et al. 1993). IR spectra on the <5 µm fraction of samples from different depths are shown in Figure 10. They confirm the XRD data, i.e. the presence of kaolinite (well crystallized) in the cemented samples and the presence of dickite in the uncemented samples located at 3254 m (Figure 10A and B). The uncemented samples between 3230 and 3244 m contain both kaolin polytypes. The increase of dickite proportion with depth is clearly shown by the IR spectra (Figure 10B). In the sample at 3257 m, a shoulder at 3668 cm-1 characterizing the presence of kaolinite is still visible in the IR spectra. The higher proportion of kaolinite in this clay fraction could be related to the relatively low porosity of this sample (about 13%) compared with the other samples (porosity of about 18%) where kaolinite is in smaller proportion according the IR spectra. In the cemented samples, kaolinite is the only kaolin polytype detected at all depths (<3260 m) (Figure 10A).
 
 

Illite characterization
 

XRD DATA. The diffraction patterns of the oriented clay fraction of both cemented and uncemented samples display a large and asymmetric peak between 5 and 11°2q (Figure 11). In the air dried samples located above 3260 m, the most intense peak is at 10 Å; below 3260 m, the illite reflections are the most intense, and a double reflection is visible with a peak at 10.5 Å and an other one at 10 Å. After glycolation, the 10.5 Å peak disappears, whereas the peak at 10 Å becomes more intense, but the reflection is asymmetric in the small angles of 2q. A comparison of the XRD data with data generated by the Newmod program (Reynolds 1985) indicates the presence of illite rich I/S mixed layers (about 90% of illite layer). With the decomposition method of Lanson and Besson (1992), the simulated XRD pattern reproduces the air dried experimental pattern in the range 5-11°2q when they are generated by three gaussian curves respectively at about 11 (8.03°2q), 10.5 (8.5°2q) and 10 Å (8.8°2q) (Figure 11 and 12). Experimental patterns of the glycolated samples are reproduced correctly with four gaussian curves respectively at about (7.3°2q), (8.3°2q), (8.7°2q) and (8.8°2q). These decompositions are very similar to the decompositions obtained by Lanson et al. (1996) for diagenetic illite from the sandstones of the Rotliegend Formation. According to these authors, these reflections indicate that the illite particles are composed of well crystallized illite (8.8°2q in air-dried and glycolated samples), poorly crystallized illite (8.6°2q in air dried samples ) and illite-smectite mixed layers with about 80% of illite layers (about 8 °2q in air dried samples). All the samples can be decomposed with similar gaussian fit curves (e.g. Lanson et al 1996, Lanson and Besson 1992). However, comparison of samples from different depths, show variations of the relative intensity of each gaussian curve (Figure 11 and figure12). PCI (poorly crystallized illite) is more intense in the deepest samples whereas WCI (well crystallized illite) decreases and IS presents the same intensity (Table 3).

RELATION TO ILLITE MORPHOLOGY. XRD patterns of the clay fraction of samples with filamentous and lath illite particles (3204 to 3260 meters depth) have asymmetrical 10 Å reflections . XRD patterns of samples below 3260 m with illite crystals having a platelet morphology display a double peak reflection at 10 and 10.5 Å. The 10.5 Å peak is less intense (Figure 11) or absent on the XRD pattern of the samples showing illite with filamentous and lath morphologies. However this peak is more intense in the deeper samples where only platelet illite crystals are present (Figure 12).

Discussion

Chronology of diagenetic processes
 
 

According to the petrographical data the chronology of the main diagenetic stages is the following (Figure 13):

1. K-feldspar overgrowths and kaolinite precipitation. Both minerals are enclosed in calcite and therefore formed prior to calcite cementation.

2. Heterogeneous calcite cementation. The calcite cementation is an early diagenetic process which occurred at depths shallower than 1000 m (Potdevin and Hassouta 1997).

3. Kaolinite-dickite transition. This transition is only visible in uncemented sandstones and therefore occurred after cementation.

4. Illite precipitation, quartz overgrowth and feldspar dissolution. Illite replaces kaolin in samples below 3204 m.

In the deepest sandstones, kaolin are completely absent, and the pore spaces are filled with diagenetic illite. Quartz overgrowths, illite authigenesis and feldspar dissolution are very rare in the cemented sandstones and therefore occurred after calcite cementation.
 

Kaolinite precipitation
 
 

Petrographic investigations suggest that kaolinite precipitated early in the burial history of these sandstones. Kaolinite precipitation is a major diagenetic event in North Sea sandstones and has been discussed by several authors. Sommer (1978) and Blanche and Whitaker (1978) suggested that kaolinite precipitated during the later stages of burial diagenesis, whereas an early origin from the precipitation by meteoritic fluids is suggested by other authors (Glasmann 1992; Bjørlykke and Aagaard 1992; Haszeldine et al. 1992). The present study confirms the latter hypothesis. Indeed, in the studied sandstones, kaolinite is embedded in a calcitic cement which precipitated at less than 1000 meters depth (Potdevin and Hassouta 1997).
 

Calcite cementation

Calcite cementation is an important factor limiting the diagenetic reactions (Kantorowicz et al. 1987; Saigal and Bjørlykke, 1987; Walderhaug et al. 1989a). According to Bjørlykke et al. 1992, the carbonate cement intervals in the Brent Group are generally thin, usually less than 1 meter and usually more abundant in the marine part of the Brent Group (i.e. Rannoch, Etive) than in the deltaïc environment of the Ness Formation. This is well demonstrated in studies of the Heather Field (Glasmann et al. 1989a). Scotchman et al. (1989) show that in the NW Hutton Field the sub-littoral sheet sand and the wave-dominated delta front sand contain more carbonate cement than distributary mouth bar and crevasse splay lobe facies. In the Ellon Field, no relationships are observed between the amount of carbonate cement and the depositional facies. According to Potdevin and Hassouta (1997), the cement fills 34-42% of the sandstone volume, suggesting a porosity before cementation of 37-44% (present porosity in cemented sandstones is about 2 to 3%, see Table 1). These high porosity values indicates that calcite cementation is an early diagenetic episode occurring at shallow depth (<1000 m) (Zieglar and Spotts 1978).

The absence of carbonate fossils in the studied sandstones and the presence of pyrite inclusions inside calcite crystals indicate that the supply of CO₂ could result from organic matter degradation. A similar explanation is given by Walderhaug and Bjørkum (1992) for the calcite origin of the Oseberg Formation. Calcium might be derived from plagioclase, potassium feldspar and/or heavy minerals dissolution.
 
 

Kaolinite-dickite transformation
 
 

Changes in the kaolin polytype morphology in sedimentary basins were initially described for a coal basin in Russia (Kossovskaya and Shutov 1963, Shutov et al. 1970), and more recently in many reservoir sandstones from the North Sea. Ehrenberg et al. (1993) described a transition from vermicular to blocky kaolin crystals in Triassic and Jurassic reservoirs of the Norway platform. The same morphological changes have been described by McAulay et al. (1993) in Brent sandstones from the Hutton and NW Hutton Field (North Sea) and by Lanson et al. (1996) in the sandstone reservoirs from the Rotliegend Formation (North Sea). The transition from vermiform to blocky crystals of kaolin is also observed in the present study, but only in uncemented sandstones. IR and XRD data confirm the structural modifications, i.e. the polytype is dependent on the morphological changes. kaolinite-dickite transition is observed in uncemented sandstone at 3240 m and is characterized by the progressive replacement of vermiform kaolinite by blocky dickite crystals (Figure 14). The depth of the transition is similar to the depth interval (2.8-3.6 km) proposed by Ehrenberg et al. (1993) and Beaufort et al. (1998) in the sandstone reservoirs from the continental platform of Norway and from the Rotliegend Formation respectively. Ehrenberg et al. (1993) suggested that the kaolinite-dickite transition is temperature dependant and that it occurs at about 120 °C.

The morphological changes observed suggest that kaolinite-dickite transition is a dissolution-precipitation mechanism. However, Lanson et al. (1996) suggested that this could be a two step process, consisting of a solid-state transformation without any morphological changes followed by a dissolution-precipitation reaction involving replacement of vermiform kaolin by blocky crystals. Further work is needed to confirm this hypothesis.
 
 

Diagenetic illite
 
 

SEM observations of the sandstones above 3260 m show the presence of filamentous and lath particles of illite. These particles seem to grow from kaolinite layers suggesting that an alteration of kaolinite into illite occurred in these sandstones. Illite authigenesis coinciding with kaolinite and K-feldspar dissolution has been described previously in the Brent sandstones from the North Sea by several authors (Bjørlykke 1983; Bjørkum et al. 1993; Bjørlykke and Aagaard 1992). The reaction which results the formation of illite is expressed as (Bjørlykke 1983):

Al₂Si₂O₅(OH)₄ + 0.75KAlSi₃O₈ Æ K_0,75Al₂(Si_3.25Al_0.75O₁₀)(OH)₂ + SiO₂ + H₂O

Kaolinite + K-Feldspar Æ Illite + Quartz

A similar reaction can be proposed in the uncemented sandstones. For instance, in thin sections, K-feldspar dissolution coincides with appearance of filamentous illite. In order to characterize and to quantify the mobility of elements during illite precipitation in the studied sandstones mass balance calculations by the Gresens method (Potdevin 1993) were performed comparing the chemical composition of cemented and uncemented sandstones (Potdevin and Hassouta 1997). Gluyas and Coleman (1992) performed a similar comparison in about ten reservoir sandstones, of permian to tertiary ages, from oilfields world-wide. They show that some early carbonate concretions preserve both the sedimentary fabric and the composition of the sandstones before calcite cementation from later diagenetic processes. In our case, the cemented sandstone represents the sandstones before illite authigenesis and the uncemented sandstone containing illite represent the final stage of the rock after illite precipitation. Gresens composition-volume diagrams show that silicon, aluminium and potassium among other elements are immobile during diagenesis (Potdevin and Hassouta 1997, Figure 15). Therefore, these elements are not brought by fluid flow and illite precipitation occurred in a closed system to silicon, aluminium and potassium. The potassium of illite is provided by dissolution of K-feldspar and the silica released by kaolinite and feldspar dissolution results in quartz overgrowths. In the studied uncemented sandstones, both mass balance computations and observations agree with the reaction proposed by Bjørlykke (1983) to explain illite authigenesis.

Below 3260 m, illite with a platelet morphology is the only clay mineral present in the pore-spaces of the uncemented sandstones. The morphological changes of illite particles with depth observed in the studied well could be related to an increase in the crystallinity with burial. However, below 3260 m, clays are in smaller proportion according modal analyses and kaolinite is absent; platelet illite crystals seem to have different origin, they could have precipitated directly from fluids without kaolin precursor. Greenwood et al. (1994) described similar diagenetic phases in the sandstones sampled in the Hingin formation from the Brae in the North Sea, these authors suggested that diagenetic kaolin was probably inhibited by temperature and depth of burial. Furthermore, the decomposition of the XRD patterns between 5 and 11°2q suggests a decrease of the illite crystallinity, i.e., an increase of the PCI (10.5 Å) curve versus the WCI (10 Å) with depth. The increase of the 10.5 Å reflection is correlated with the presence of platelet illite, whereas above 3260 m where illite particles present a filamentous or lath morphology, the 10 Å reflection is more intense. Filamentous and lath morphologies are found at depth lower than 3260 m in the shoreface facies where detrital micas are very abundant. Therefore the contribution of the diagenetic illite (10.5 Å peak) could be masked by the stronger contribution of the detrital mica (10 Å peak) within the clay fraction. In the deeper samples (> 3260 m) the sedimentary facies changes (tidal complex), detrital micas are less abundant and the diagenetic illite reflection becomes relatively more intense. The second explanation for the increase of the 10.5 Å reflection in the deeper samples is that platelet particles could correspond to poorly crystallized illite that directly precipitated from fluids.
 
 

Factors promoting clay diagenesis in sandstone reservoirs
 

The clay diagenetic evolution described in this study is clearly different in cemented sandstones and in sandstones where the calcite cement is absent (Figure 13). Mineralogical data (IR, XRD and SEM) suggest that kaolinite is the unique authigenic clay mineral in all the studied cemented sandstones whereas a kaolinite-dickite transition is observed in the uncemented sandstones with increasing depth.

These data suggests that the kaolinite-dickite transition can be inhibited by the calcite cement and that temperature and burial are not the only parameters controlling this transition. Zemmerle et al. (1991) suggested that dickite occurrence could be restricted to sedimentary rocks of high porosity and permeability. This is also confirmed in the present study: in cemented sandstones (0 to 3% of porosity) kaolinite is the only kaolin phase formed before calcite cementation and in uncemented sandstones (13 to 20% of porosity) both kaolinite and dickite were identified. By comparison, Ehrenberg et al. (1993), McAulay et al. (1993), McAulay et al. (1994) indicate that temperature is the major factor controlling the kaolinite-dickite transition. A recent study of Buatier et al. (1997) of the clay minerals evolution in thrust faults from the south-Pyrenean basin demonstrated that deformation could be an important factor promoting dickite crystallization. The present study demonstrates that porosity and permeability are also important parameters. The calcite cement would prevent fluid flow in cemented sandstones. By filling the pores, calciteosity it will prevent fluid-rock interactions and chemical reactions by dissolution-precipitation like the kaolinite-dickite transition.
 
 

Conclusions

The sandstones of the Ellon Field reservoir (North Sea) show a complex diagenetic evolution, the first authigenic clay mineral is kaolinite which precipitated during early diagenesis. Then, the sandstones have been heterogeneously cemented by calcite. This calcite cement preserved kaolinite whereas in uncemented sandstones kaolinite has been progressively replaced by dickite with increasing depth. Later, illite precipitated in uncemented sandstones. Its morphology changes with depth: filamentous, lath and platelet illites are successively found at 3204 m, 3244 m and 3275 m. This study shows that temperature and depth are not the only parameters which control the clay reactions (kaolinite-dickite transition, illite precipitation). Pore space availability, fluid composition or fluid flow could be also important factors. Our data suggest that the kaolinite-dickite transition cannot be used as a simple geothermometer.
 
 

Acknowledgements

We are indebted to S.N. Erhenberg and M. Batchelder for their constructive review of the manuscript and to TOTAL-CST France for permission to release this study. C. Demars et F. Sommer (TOTAL-CST) are acknowledged for providing the studied samples and S. Petit (University of Poitiers) for helping us to in IR spectra study. The comments and discussions with E. Brosse, C. Durand (IFP) and B. Lanson (University of Grenoble) were also greatly appreciated.
 
 

Figure captions

Figure 1. Location map of the studied well.

Figure 2. Schematic log of the sedimentary facies within the studied well. Symbols include: IT = Transgressive lag, SS = Upper shoreface, M = Marsh, S = Shoreface, CT = Tidal complex, BE = Mouth bar. Coal and organic shale are in black. The cemented zones represent shown in grey. The filled circles are cemented samples. The open circles represent uncemented samples.

Figure 3. Section of sample (3244) showing the sharp transition between (A) a calcite cemented sandstone and (B) uncemented sandstone.

Figure 4. Optical microscope image of studied sandstones. (A) Quartz grains showing one or two phases of overgrowth (O1 and O2), indented contacts between detrital grains of quartz are shown by arrows (B) dissolved K-feldspar in uncemented sandstone (C) preserved microcline (Kfs) in cemented sandstone and (D) alteration of muscovite to kaolinite in a cemented sandstone. The kaolinite crystals appears at the edge of the muscovite crystal. Plane polarized light. Symbols include: Qtz = Quartz, Kfs = K-Feldspar, Mus = Muscovite, Kln = Kaolinite, iIll = iIllite, Cal = Calcite.

Figure 5. SEM micrographs of the kaolin polytype morphologies. (A) Vermiform kaolin crystals (3204 m), (B) Cluster kaolin crystals (3224 m), (C) Blocky kaolin crystals (3254 m). (D) Replacement of vermiform kaolin by blocky crystals (3204 m). (A), (C) and (D) are secondary electron images. (B) is backscattered electron image. Symbols include: Mus = Muscovite, Kln = Kaolinite.

Figure 6. SEM micrographs of illite morphology changes with depth. (A) Filamentous illite (3204 m), (B) lath illite (3244 m), (C) platelet illite (>3260 m).

Figure 7. XRD patterns of a pair of uncemented (a) and cemented (b) sandstones from 3238 m. The kaolinite peaks are more intense than the illite in the cemented sandstones. (c) XRD patterns of a deep uncemented sandstone (3275 m). The kaolinite peaks are absent.

Figure 8. XRD pattern of a powder clay fraction from a calcite cemented sandstone (A) and an uncemented sandstone (B). Peak labels include: I = Illite, Q = Quartz, F = Feldspar, D = Dickite, K = Kaolinite.

Figure 9. XRD pattern of a random oriented powder clay fraction (< 5 mm) of uncemented sandstones showing the evolution of the crystallographic structure of kaolin (polytype) with depth. Peak labels include: I = Illite, F = Feldspar, Q = Quartz.

Figure 10. Infrared spectra of clay fractions (< 5 mm) from a calcite cemented sandstone (A) and an uncemented sandstone (B). The four characteristic bands of kaolinite are present in the calcite cemented sandstone. Whereas only three bands are present in uncemented sandstone from the same depth, characterizing the dickite polytype. The evolution of the Infrared reflectance spectra as a function of the depth in the uncemented sandstone is showed in B whereas in calcite cemented sandstones (A) there is no change with depth.

Figure 11. Decomposition of both (A) air dried and (B) glycolated XRD profile of a sample 3244 m. Symbols include: WCI = Well crystallized illite, PCI = Poorly crystallized illite, IS = Ordered mixed layers.

Figure 12. Decomposition of both (A) air dried and (B) glycolated XRD profile of a sample 3275 m.

Symbols include: WCI = Well crystallized illite, PCI = Poorly crystallized illite, IS = Ordered mixed layers.

Figure 13. Relationships between the clay diagenetic evolution of these sandstones and the major diagenetic events occurring in the North Sea.

Figure 14. Schematic diagram of changes to Kaolin morphologies with depth in the uncemented sandstones.

Figure 15. An example of mass balance calculations of illite authigenesis and quartz overgrowths using the Gresens method (1967). A. The chemical analysis of the calcite cemented sandstone (sample 3238 c) gives the rock composition before illite authigenesis and quartz overgrowths. The chemical analysis of an uncemented sandstone sampled from the same lithology (3238 nc) gives the rock composition after illite authigenesis and quartz overgrowths. B. Composition-volume diagram of absolute mobility of major oxides Dm/M_o versus Fv. Dm/M_o, the gain or loss of an oxide (in mass percentage of initial rock) is given by the general relation derived by Gresens (1967). In this diagram, the absolute mobilities of the oxides are graphically shown versus the possible volume changes accompanying the transformation from a cemented sandstones to an uncemented one. m_o and m are the mass of the oxide in the initial and final rock (i.e., the cemented and uncemented sandstones), c_o and c are the weight percentages in this oxide, d_o and d the densities and Fv the volume factor or ratio of final volume to initial volume. In the composition-volume diagram, the three major elements Si, Al and K are simultaneously immobile for a Fv value of 0,77 (i.e., a volume change of 23%). It cannot be a coincidence (Gresens, 1967) and these elements are really immobile during illite authigenesis and quartz overgrowths. CaO, MgO, MnO, FeO mobilities are related to carbonate cementation (see Potdevin and Hassouta, 1997 for a more comprehensive study of mass balance calculations).
 
 

Table captions

Table 1. Summary of major physical properties of the studied samples. Symbols include: X = present, SS = Upper shoerface, S = Shoreface, BE = Mouth bar, CT = Tidal complex, PM = poor to moderate, w = well, nc = uncemented sandstone, c = cemented sandstone. Symbole include: nc = uncemented sandstone, c = cemented sandstone.

Table 2. Mineralogical data on the studied samples. Relative abondances of each mineral were calculated from modal analyses calculated from point counting on thin section. XRD data are also presented in this table. Mineral labels include: Cal = Calcite, Rf = Rock fragments, Om = Organic matter, Py = Pyrite, Q = Quartz, K-F = K-Feldspar, Ab = Albite, M = Mica, Qo = Quartz Overgrowths, ill = illite, Kln = Kaolin, K^? = Kaolinite; D^? = Dickite, ill^? = illite, nc = uncemented sandstone, c = cemented sandstone. ^?Data from XRD, X= present, XX = dominant Symbole include: nc = uncemented sandstone, c = cemented sandstone

Table 3. Relative intensity of each gaussian curve obtained by decomposition of air dried XRD pattern between 5 and 11° 2q, PCI is more intense in the deepest samples whereas WCI decreases and IS presents the same intensity. Symbols include: WCI = Well crystallized illite, PCI = Poorly crystallized illite, IS = Ordered mixed layers, nc = uncemented sandstone, c = cemented sandstone.
 
 

References