20
Clay Minerals (1998) 33, 15-34 Clay mineral diagenesis in sedimentary basins a key to the prediction of rock properties. Examples from the North Sea Basin K. BJORLYKKE Department of Geology, Box 1047 Blindern, University of Oslo, N-0316 Oslo, Norway (Received 23 September 1996," revised 9 June 1997) A B S T RA C T : Dissolution of feldspar and mica and precipitation of kaolinite require a through flow of meteoric water to remove cations such as Na + and K § and silica. Compaction driven pore-water flow is in most cases too slow to be significant in terms of transport of solids. The very low solubility of A1 suggests that precipitation of new authigenic clay minerals requires unstable Al-bearing precursor minerals. Chlorite may form diagenetically from smectite and from kaolinite when a source of Fe and Mg is present. In the North Sea Basin, the main phase of illite precipitation reducing the quality of Jurassic reservoirs occurs at depths close to 4 km (130-140~ but the amount of illite depends on the presence of both kaolinite and K-feldspar. Clay mineral reactions in shales and sandstones are very important factors determining mechanical and chemical compaction and are thus critical for realistic basin modelling. The presence of clay minerals and clastic sheet silicates strongly influence the physical and chemical properties of both sandstones and shales. The primary sediment composition and the early diagenetic reactions determine the burial diagenetic reactions and rock properties as a function of depth. Clay minerals will also, in most cases, reduce their shear strength and increase the surface area of the sediments and change chemical properties such as ion exchange capacity. The primary clastic compo- sition of sedimentary rocks is related to source rocks, weathering and erosion in the source area, transport processes and to the depositional environ- ment. Each basin has a different basin subsidence and depositional history and clay diagenesis is influenced by many different factors. Diagenetic reactions are driven towards higher thermodynamic stability at a rate which is controlled by the kinetics of the mineral reactions. The main principles for clay mineral diagenesis should therefore be the same for all basins even if they have very different initial mineralogy and thermal history. If these principles can be agreed upon, the main problem is making the right assumptions about variables such as provenance, facies, sedimentation rates and geothermal gradi- ents. The same diagenetic reactions that we can study in sandstones probably also apply to mudstones, even if the texture and mineralogy may be different. The North Sea Basin and Haltenbanken (Mid-Norway) Basin are particularly good 'laboratories' for studying clay mineral diagenesis. Both basins are extensively cored and large amounts of geochemical and mineralogical data are available on the composition of the sediments, as well as the pore-water (Egeberg & Aagaard, 1989; Aagaard et al., 1992; Warren & Smalley, 1994; Bjorlykke et al., 1995). With the exception of the marginal parts of the basins, there has been almost continuous subsidence and sedimentation through the Cenozoic. An overview of the regional geology and stratigraphy of the North Sea Basin is provided by Glennie (1990) and of Haltenbanken by Koch & Heum (1995). 1998 The Mineralogical Society

Clay mineral diagenesis in sedimentary basins a key to the

  • Upload
    dotuyen

  • View
    223

  • Download
    2

Embed Size (px)

Citation preview

Page 1: Clay mineral diagenesis in sedimentary basins a key to the

Clay Minerals (1998) 33, 15-34

Clay mineral diagenesis in sedimentary basins a key to the prediction of rock properties. Examples from the North Sea

Basin

K. B J O R L Y K K E

Department of Geology, Box 1047 Blindern, University of Oslo, N-0316 Oslo, Norway

(Received 23 September 1996," revised 9 June 1997)

A B S T RA C T : Dissolution of feldspar and mica and precipitation of kaolinite require a through flow of meteoric water to remove cations such as Na + and K § and silica. Compaction driven pore-water flow is in most cases too slow to be significant in terms of transport of solids. The very low solubility of A1 suggests that precipitation of new authigenic clay minerals requires unstable Al-bearing precursor minerals. Chlorite may form diagenetically from smectite and from kaolinite when a source of Fe and Mg is present. In the North Sea Basin, the main phase of illite precipitation reducing the quality of Jurassic reservoirs occurs at depths close to 4 km (130-140~ but the amount of illite depends on the presence of both kaolinite and K-feldspar. Clay mineral reactions in shales and sandstones are very important factors determining mechanical and chemical compaction and are thus critical for realistic basin modelling.

The presence of clay minerals and clastic sheet silicates strongly influence the physical and chemical properties of both sandstones and shales. The primary sediment composition and the early diagenetic reactions determine the burial diagenetic reactions and rock properties as a function of depth. Clay minerals will also, in most cases, reduce their shear strength and increase the surface area of the sediments and change chemical properties such as ion exchange capacity. The primary clastic compo- sition of sedimentary rocks is related to source rocks, weathering and erosion in the source area, transport processes and to the depositional environ- ment. Each basin has a different basin subsidence and depositional history and clay diagenesis is influenced by many different factors.

Diagenetic reactions are driven towards higher thermodynamic stabili ty at a rate which is controlled by the kinetics of the mineral reactions. The main principles for clay mineral diagenesis should therefore be the same for all basins even if they have very different initial mineralogy and

thermal history. If these principles can be agreed upon, the main problem is making the right assumptions about variables such as provenance, facies, sedimentation rates and geothermal gradi- ents. The same diagenetic reactions that we can study in sandstones probably also apply to mudstones, even if the texture and mineralogy may be different. The North Sea Basin and Haltenbanken (Mid-Norway) Basin are particularly good ' laboratories ' for studying clay mineral diagenesis. Both basins are extensively cored and large amounts of geochemical and mineralogical data are available on the composition of the sediments, as well as the pore-water (Egeberg & Aagaard, 1989; Aagaard et al., 1992; Warren & Smalley, 1994; Bjorlykke et al., 1995). With the exception of the marginal parts of the basins, there has been almost continuous subsidence and sedimentation through the Cenozoic. An overview of the regional geology and stratigraphy of the North Sea Basin is provided by Glennie (1990) and of Haltenbanken by Koch & Heum (1995).

�9 1998 The Mineralogical Society

Page 2: Clay mineral diagenesis in sedimentary basins a key to the

16 K. Bj~rlykke

The present-day geothermal gradients in the North Sea vary mostly between 30-40~ (Glennie, 1990; Hermanrud et al., 1991) and there is no evidence to suggest that the geothermal gradients were very much higher earlier in Cenozoic times. Even if the geothermal gradients should have been considerably higher during the Mesozoic or Lower Tertiary, it is not likely that the temperature of a particular rock should have had a higher absolute temperature because of the Pliocene and Pleistocene subsidence. This is also supported by the fact that fluid inclusions in quartz from the North Sea and Haltenbanken record temperatures up to the present-day bottom-hole temperature but not higher (Walderhaug, 1990, 1994; Saigal et al., 1992). Where there has been no hydrothermal activity or uplift, the present burial depths and temperatures can, in most cases, be taken as maximum values because of the rapid late Cenozoic subsidence. It is reasonable to assume that the North Sea Basin has experienced recent progressive burial diagenetic processes with increasing temperatures, except in the uplifted marginal parts of the basins.

The depth ranges of authigenic minerals such as kaolinite, illite, chlorite and quartz provide very important constraints on the interpretations derived from petrographic analyses. This paper is an attempt to present a summary of the main clay mineral reactions typical of the North Sea and Hattenbanken basins and to discuss the principles of diagenetic processes involving clay minerals. However, a detailed discussion of the regional variations in clay mineralogy within these basins is beyond the scope of this paper. For recent reviews of elastic diagenesis see Wilson (1994).

C L A Y M I N E R A L O G Y A N D S A N D S T O N E D I A G E N E S I S

Diagenetic reactions must have a thermodynamic drive so that the minerals precipitated are more stable than the minerals which are dissolving. At shallow depths and low temperatures, hydrous minerals such as gibsite, kaolinite and smectite form as a result of weathering or early diagenetic processes during meteoric water flow. Such early diagenetic processes may be considered a continua- tion of the weathering process even if the pore- water is reducing. The overall reaction:

rock (feldspar, mica) + water = clay + cations.

These minerals become unstable at greater burial depth and higher temperatures and this reaction is often referred to as reversed weathering:

clay (kaolinite, smectite) + cations (K +) = aluminosilicate (illite) + quartz + water.

The above reactions are modified from Velde (1995).

In the North Sea and Haltenbanken basins, diagenetic studies have focused mostly on the reservoir sandstones of Jurassic age. The main primary minerals such as feldspar and mica are unstable when exposed to meteoric water of low ionic strength near the surface (weathering), but comprise a stable mineral assemblage during burial diagenesis at higher temperatures and lower flow rates. It is well known that arkoses have their feldspars well preserved after exposure to greens- chist facies or higher grades of metamorphism. Only if kaolin, smectite or other potentially unstable clay minerals form at shallow depth will clay mineral reactions such as precipitation of illite take place at greater burial due to higher temperatures. In such well-sorted reservoir sand- stones, nearly all the clay minerals are authigenic and the distribution of clay minerals then depends on the diagenetic processes.

Early (shallow) diagenesis

The term early diagenesis is used here to include processes near the surface where the diagenesis may be strongly influenced by meteoric water and sea-water. Early marine diagenesis is strongly influenced by the accumulation of biogenic carbonate and silica on the sea floor and by interaction with sea-water by diffusion near the red/ox boundary. These changes in the primary elastic sediment composition are related to marine facies and may strongly influence diagenetic reactions at greater burial. Meteoric water may flow deep into sedimentary basins and many of the reservoirs in the northem North Sea have salinities which are about 50% of that of normal sea-water (Warren & Smally, 1994 ). However, the flux of meteoric water is highest in the marginal and shallow parts of the basin. It also depends on the climate, topography, water table and on aquifers and aquitards in the basin. Fluvial and shallow marine sediments will be flushed by meteoric water after deposition while more distal shelf facies and turbidites normally will be subjected

Page 3: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins 17

FIG. 1. Leached feldspar and authigenic kaolinite from the Brent Group (Ness Formation, Huldra Field, depth, 3722 m (Nedkvitne & Bjorlykke, 1992). Porosity is gained by feldspar dissolution (secondary porosity) but lost

through precipitation of kaolinite.

to much less meteoric water flushing. The dissolution of feldspar and mica and precipitation of kaolinite (Fig. 1) is a weathering reaction and this type of early diagenesis may be referred to as subsurface weathering:

2K(Na)A1Si3Os + 2 W + 9H~O = A12SiOs(OH)4 + 2HaSiO4 + 2K(Na +)

It is clear that this reaction cannot take place in a closed system because it requires the supply of protons and the removal of cations such as Na + and K + and silica by fluid flow. Most ground-waters are in the stability field of kaolinite (Garrels & Christ, 1965). Calculations show that a total flow of 103--104 m3/m 2 through sandstones is required to dissolve significant quantities of feldspar and mica and precipitate a few percentages of kaolinite (Bjorlykke, 1994). This rate of flow is obtained in

fluvial and shallow marine environments if the climate is humid. Silica will normal ly not precipitate as quartz at such low temperatures and must be removed along with the alkali ions in order for the pore-water to remain in the stability field for kaolinite. Meteoric water may penetrate deeply into sedimentary basins in some cases, but meteoric water fluxes, significant in relation to dissolution of feldspar and mica, probably occur mostly at depths shallower than I00 m - - in many cases shallower than 10 m (Bjorkum et al., 1990). The model indicating mixing of meteoric water and compac- tion driven flow from opposite direction into the Brent Gr. (Osborne et al., 1994), is hydrodynami- cally very problematic. Meteoric water flowing into sedimentary basins will gradually approach equili- brium with the mineral phases present, including also K-feldspar and mica as the distance from the

Page 4: Clay mineral diagenesis in sedimentary basins a key to the

18 K. Bjor lykke

area of recharge increases. High fluxes are required for the pore-water to remain in the stability field of kaolinite (low K+/H+), but only a few exchanges of pore-water are required before the pore-water composition will be dominated by the isotopic signature of meteoric water. Osborne et al. (1994) suggested that authigenic kaolinite formed in the Brent sandstones because of meteoric water flow into the basin in the Late Cretaceous to Early Eocene. At that time, however, the Middle Jurassic Brent Group was in most places covered by a 1 -2 km thick sequence of mudstones, which probably had low permeability, thereby reducing the potential for large fluxes of meteoric water to flow into the Jurassic sediments and up to the surface again. In addition, the land areas adjacent to the northern North Sea such as the Shetland platform and western Norway were transgressed by the sea during Late Cretaceous and Early Tertiary times (Hancock, 1984; Jordt et al. , 1995) providing little head for meteoric water flow into the basin. Late Eocene and Oligocene smectitic mudrocks with very low permeability probably reduced the potential for fluid into the underlying Mesozoic sequence as well as back up to the surface.

In the North Sea Basin it has been shown that the distribution of kaolinite in sandstones can be related to facies and climate (Bjorlykke & Aagaard, 1992). The Permian and Triassic sandstones in the North Sea Region, which were deposited in a dry climate, generally contain little kaolinite compared to the Jurassic fluvial and shallow marine sandstones deposited in a more humid climate. Practically all samples of sandstones from the Brent Group which are not carbonate cemented contain authigenic kaolin and show evidence of feldspar dissolution (Morton e t a l . , 1992). Below ~4 km depth, however, much of the kaolinite has been dissolved and replaced by illite (Fig. 2). The Brent Group represents a deltaic facies where both the fluvial and shallow marine sediments would have been flushed by meteoric water shortly after deposition. The most effective leaching of feldspar and mica occurs at shallow depths (<10-20 m), even though meteoric water may extend much deeper into the basin. The degree of dissolution of feldspar and mica varies through the Brent Group as a function of facies (Nedkvitne & Bjorlykke, 1992).

The 8180 values of authigenic kaolinite indicate rather low formation temperatures. The exact temperatures cannot be calculated because of the

uncertainty of the isotopic composition of the pore- water but it is highly unlikely that it could be formed at high (>100~ temperatures (Glasmann e t

al. , 1989a,b). Authigenic kaolinite is common in sandstones of

the Brent Group in the shallowest reservoirs at <1700 m in the Emerald Field (Osborne et al., 1994) and at c. 1.8 km depth in the Gullfaks Field (Bjorlykke e t al. , 1992; Giles e t al. , 1992) and in Upper Jurassic sandstones of the Troll Field at 1.5-1.6 km indicating that kaolinite has precipi- tated at shallower depth. Osborne et al. (1994) suggested that kaolinite precipitated from meteoric water at temperatures between 25-80~ at a burial depth between 571-2143 m at a time interval between 47-86 Ma. They assumed that kaolinite was precipitated from a mixture of meteoric water and compactional water with 81So values between -6.5 to -3.5~ SMOW. The composition of Jurassic meteoric water is, however, poorly constrained and may vary locally depending on various factors such as wind directions during rainfall. Assuming that the kaolinite precipitated from meteoric ground water with 8180 values between - 7 to -9%0 SMOW, most of the kaolinite could have precipitated at low temperatures (20-30~ which is also suggested by McAulay e t al. (1990). Seventy six analyses of diagenetic kaolinite from the Brent Group (Osborne e t al. , 1994) showed that 5180 values decrease with increasing present-day burial depth from an average of 17.2%o at 1600-1700 m, 16.4%o between 2-3000 m and 14.2%o at >3300 m. This could indicate that some degree of re-equilibration had occurred during burial, but then the pore-water must have continued to have low 5180 at greater depth. Possibly, some of the kaolinite now observed in reservoir sandstones could have been recrystal- lized from amorphous aluminium phases or from less crystalline kaolinite thus explaining somewhat elevated temperatures. Some of the kaolinite in the deeper reservoirs may be dickite and this may also change the oxygen isotopic composition. Although kaolinite is a pore-filling mineral, it is often partly enclosed in authigenic quartz at greater burial depth (Fig. 3).

Upper Jurassic sandstones from the North Sea Basin representing more distal shelf facies and turbidites contain very little or no authigenic kaolinite and much less evidence of feldspar leaching compared to the underlying Brent Group of delta facies where these features are ubiquitous.

Page 5: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins

Biogenic carbonate and s i I i c a M e t . . w a t e r

/ I Basin f i l l ~

Carbonate cement ~I BURIAL Verdine(Fe) DEPTH Opal A-CT - quartz Facies

Little aut. kaolinite o-3.s(4) km

19

Detrital supply

N Dissolution of feldspar

and mica, precipitation of authigenic kaol in i te

2KAISi308+2H++ 9H20 =

AI2Si205(OH) 4 +4H4SiO 4 +2K +

BURIAL Extensive Chlorite Quartz cement, DEPTH > quartz cementation coatings? Illitization if 3.5(4) Little illite if Little kaolinite and Km kaolinite and quartz K-feldspar are present

smect i te are absent cement KAISi308+AI2Si2Os(OH)4 =

KAI3Si3010(0H)2+2Si02 +2H20

FI~. 2. Model for relationships between provenance, facies-related early diagenesis and diagenesis at greater burial depth. Provenance and early diagenesis in meteoric or marine environments strongly influence the

diagenesis at deeper burial.

In the Claymore Field, a thin turbiditic sandstone (Ten Foot Sand) within the Kimmeridge Clay Formation shows no significant authigenic kaolinite and feldspar dissolution (Spark & Trewin, 1986), while the Piper Formation, which is a paralic deposit, contains authigenic kaolinite and inten- sively leached feldspar and mica. The Fulmar Formation is an example of a distal shelf and turbidite facies which contain little diagentic kaolinite (Stewart, 1986). This may be because it has received too little meteoric water flushing to cause leaching of feldspar and mica (Saigal et al., 1992). If the leaching was related to the generation of CO2 or organic acids as suggested by several authors (Schmidt & McDonald, 1979; Surdam et al., 1984, 1989; Burley et al., 1985; Burley, 1986), then the sandstones which were close to the source rock would be expected to show the most leaching.

This is clearly not the case. In Upper Jurassic sandstones, representing more proximal shallow marine facies such as in the Piper and Tartan formations, authigenic kaolinite is observed, prob- ably because they have been more extensively flushed by meteoric water (Burley, 1986). Near the top of rotated fault blocks, when they were submerged as islands, the Brent Group has been exposed to meteoric water flushing (Bjorlykke & Brendsdal, 1986). Because of uplift and erosion, relatively little of the section affected by meteoric flushing below the unconformity may be preserved (Bjorkum et al., 1990).

Dissolution and precipitation of minerals due to fluid flow during deeper burial cannot be expected to be facies selective as is the case with early diagenetic reactions. Shallower sandstones are rarely cored but authigenic kaolinite has been

Page 6: Clay mineral diagenesis in sedimentary basins a key to the

20 K. Bjorlykke

FIG. 3. Authigenic kaolinite enclosed by quartz overgrowth. Rarmock Formation (Brent Group, Statfjord Field). The scale bar represents 0.001 mm.

observed in cuttings from shallower sandstones, i.e. from the Pliocene Nordland Group (36/1-2) at 500 m present burial depth (Singh, 1996). Most Lower Tertiary sandstones in the North Sea contain little authigenic kaolinite, probably because they represent distal marine and turbidite facies which have not been subjected to extensive flushing by meteoric water (Bjorlykke & Aagaard, 1992). Minor amounts of authigenic kaolinite may, however, be found in Paleocene sandstones (Stewart et al., 1990). The Permian and Triassic sandstones, which were deposited in a rather dry climate, contain little authigenic kaolinite and mostly smectite and illite, which is typical of

desert environments today (Weaver, 1989). In Upper Triassic sediments, i.e. the Lunde Formation at the Snorre Field, the kaolin content is higher, probably due to a slightly less arid climate and it is clearly linked to increased ground waterflow through fluvial sandstone.

The depth and temperature of formation of kaolinite in the basins like the North Sea has been the subject of considerable controversy. It was a widely held view that the dissolution of feldspar and precipitation of kaolinite occurred in connec- tion with release of acids from generation of oil from kerogen (Burley, 1986; Surdam et al., 1984, 1989). As shown above, observations from cores

Page 7: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis

show that abundant kaolinite has already precipi- tated prior to deep (>1.5-2.0 km) burial. This is also supported by isotopic data suggesting a relatively low temperature (Glasmann et al., 1989a). Precipitation from meteoric water at 1 -2 km depth as suggested by Osborne et aL (1994) cannot be disproved by data since there are no cores from depths <1.5 km. The interpretation that kaolinite is formed by leaching at very shallow depth is based on indirect reasoning about meteoric water fluxes required to produce significant dissolution (weathering).

Authigenic clays change the pore-size distribu- tion and therefore also oil saturation and the production capability of reservoir sandstones (Pittman, 1978). Prediction of such rock properties depends very much on the diagenetic model for minerals like kaolinite which is important in shallow reservoirs and even more important as precursor for illite at greater depth.

B u r i a l d iagenes i s

Meteoric water flow may reach deep into sedimentary basins driven by the head of ground water from nearby land areas. The rate of flow is highest near the surface and decreases with depth and distance from land areas, depending on the distribution of aquifers. At a certain depth, however, the compaction process in subsiding basins will build up sufficient over-pressure and prevent penetration of meteoric water. This depth is very difficult to estimate but it follows from what has been stated above, that meteoric water can flow more deeply into uplifted sediments which are not undergoing compaction.

Weathering reactions like dissolution of feldspar and precipitation of kaolinite require that K § and silica are removed so that the pore-water can remain in the stability field of kaolinite (Fig. 4). The pore-water need not be acidic but must have a low K+/H + ratio. In sandstones of the North Sea Basin, there is nearly always some carbonate present and the pore-water is in equilibrium with calcite and the pH is to a large extent determined by the COz. At greater depth (>3-4 km), clay mineral reactions have the highest pH buffering capacity. Organic acids have much lower buffering capacities than both the silicate and the carbonate system and therefore do not influence the pH very much (Hutcheon et al., 1992). In the presence of K- feldspar, the K+/H + ratio of the pore-water will be

in sedimentary basins 21

too high to be in the stability field of kaolinite. If some K-feldspar or mica should dissolve, the K concentration in the pore-water will increase until the reaction stops, since there is normally no other mechanism for removing K.

At burial depths >2-3 km, the kaolin mineral present is commonly not kaolinite but dickite (Ehrenberg et al., 1993). Much of what has previously been described as kaolinite from North Sea reservoir sandstones should mineralogically be classified as dickite. The transition of kaolinite to dickite is still poorly understood. It is not known if there are factors other than temperature influencing this transition and to what extent it involves total dissolution of kaolinite and precipitation of dickite so that the oxygen isotopic ratios are reset.

Kaolinite in shales is probably mostly clastic although this is difficult to prove because the primary textures are usually difficult to observe because they have been destroyed by compaction. Mudstones are normally not subjected to much meteoric water flow due to their low permeability. At least in the presence of K-feldspar, the pore- water in the shales should be expected to be in the stability field of illite. Authigenic kaolinite, however, may be observed in coarse-grained mudstones. It is often not quite clear to what extent kaolinite in these cases has precipitated due to leaching of feldspar and mica or has precipitated from clastic gibbsite minerals or amorphous aluminous gels (Foscolos, 1984) reacting with biogenic silica. The stability of smectite is reduced with higher temperatures and, as the rate of quartz precipitation increases, the pore-water will be less supersaturated with respect to quartz.

Illite. Authigenic illite often occurs as a fibrous pore-filling mineral which strongly reduces the permeability in reservoir sandstones (Fig. 5). The percentage of authigenic illite is difficult to quantify by XRD because of interference from clastic illite and mica. A strong increase in the amount of illite relative to kaolinite in the clay fraction is observed below 3.7-4.0 km, both in the northern North Sea (Giles et al., 1992; Bjorlykke et al., 1992) and Haltenbanken (Bjorlykke et al., 1986; Ehrenberg & Nadeau, 1989). North Sea Jurassic reservoir sandstones which have been buried to depths <3.5 km generally show little pore-filling illite in thin-section or by SEM. Even if the pore-water composition in the North Sea Basin in most cases falls in the stability of illite, little precipitation occurs due to an extremely low kinetic

Page 8: Clay mineral diagenesis in sedimentary basins a key to the

22 K. Bjorlykke

ol ""... 60 "" .......... K-Fe ldspar

"%

, %~

80 ~E]~ o ~ ""''""" ................................

o•'• lOO

g , t l l " \ I l l i te D Viking Graben, < 50,000 ppm CI

1 ~ . o % o �9 Viking Graben, > 50,000 ppm CI

�9 �9 ,, Central Trough, < 50,000 ppm CI

140t ~-111 �9 Central Trough, > 50,000 ppm CI I 1 ~ ~60 1 I.A~ o Haltenbanken, < 50,000 ppm CI 7 ~ T

�9 Haltenbanken, > 50,000 ppm CI K a o l i n i t e ~ \ ~ l

~8o, ,~ , ~ , . , . . . . . . . . , , . . . . , . . . . , . . . . , . . . . I 0 500 1000 1500 2000 2500 3000 3500 4000

K ( p p m )

Equilibrium between kaolinite and illite at 0% salinity

Equilibrium between kaolinite and illite at 10% salinity

........................ Equilibrium between illite and K-feldspar at 0% salinity

. . . . . . Metastable equilibrium between kaolinite and K-feldspar at 0% salinity

FIG. 4. Geochemical composition of formation water from North Sea and reservoirs in relation to the stability field of kaolinite, illite and K-feldspar. All the pore-water analyses fall in the stability field of illite but precipitation of illite depends on the available A1 from precursor minerals and on the kinetics which is very slow below 120-140~ At higher temperatures, the pore-water composition falls close to the boundary between the stability field of kaolinite and illite as should be expected when kaolinite is replaced by illite. From Bj~rlykke et

al. (1995).

precipitation rate at low temperature (< 120-140~ (Bjgrlykke et al., 1995).

High concentrations of authigenic illite are nearly always associated with dissolution of an unstable aluminous mineral phase which in North Sea Jurassic reservoirs is mostly kaolin. Another precursor mineral for illite is smectite but analyses of shallow samples (<2 km) suggest that the Jurassic sandstones had a low primary smectite content. Tertiary sandstones, however, may have had a high smectite content (Bjorlykke et al., 1995).

Authigenic illite may form by different reactions (Bjorlykke et al., 1995):

smectite + K § = illite + silica (via mixed-layer minerals) (1)

A12SiOs(OH)4 + KAISi308 = Kaolinite K-feldspar

KA13Si3Olo(OH)2 + 2SIO2 + H20 (2a) Illite Quartz

3A12Si2Os(OH)4 + 2KA1Si308 + 2Na + = Kaolinite K-feldspar

2KA13Si301o(OH)2 + 2NaA1Si308 + 2H + + 3H20 (2b)

Illite Albite

Page 9: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins 23

FIG. 5. Pore-filling authigenic illite from a Jurassic reservoir, Haltenbanken (4.2 km depth).

The distribution of smectite in the North Sea sediments suggest that smectite dissolves at temperatures of ~65-75~ and at 80-100~ the mixed-layer minerals contain >70% illitic layers (Dypvik, 1983). This reaction depends, however, on several other factors including the supply of K and the time factor (Boles & Franks, 1979).

Reaction 2a is isochemical and does not require the supply or removal of ions by pore-water flow. It does require, however, that K +, A13+ and silica are transported from the surface of the dissolving K-feldspars and kaolinite to the site of illite growth. The SEM pictures frequently show that the authigenic illite growth is closely associated with or replacing dissolved kaolinite. The rate- limiting process for illite growth will then be the kinetics of illite precipitation and the transport of K + from dissolving K-feldspar. At low temperature and slow precipitation rate, the pore-water will be highly supersaturated with respect to illite and less under-saturated with respect to K-feldspar, thus reducing the dissolution rate of K-feldspar and the diffusive transport of K. In the second reaction,

illitization is combined with albitization and there is no excess silica which can be precipitated as quartz. There are several pieces of evidence suggesting that illite requires relatively high temperatures to form. (1) Geochemical analyses of pore-water from the North Sea show that the pore-waters are mostly supersaturated with respect to illite (Fig. 4). (2) A strong increase in the amount of diagenetic illite is observed in reservoir sandstones at depths close to 3 . 8 - 4 . 0 km c o r r e s p o n d i n g to 1 2 0 - 1 4 0 ~ (Bjorlykke et al., 1986; Ehrenberg, 1990). (3) Illite may also form from dissolving smectite at somewhat lower temperatures (80-100~

The stability of smectite is reduced as quartz starts to precipitate, reducing the supersaturation with respect to quartz (Aagaard & Helgeson, 1983; Sass et al., 1987). The rate of illitization of smectite also depends on the supply of K (Hower et al., 1976) and in sediments without zeolites and K-bearing evaporites, this will mainly be K-feldspar. It is still not known how far K can be transported by diffusion within sandstones and between sandstones and shales. Detailed petrographic and

Page 10: Clay mineral diagenesis in sedimentary basins a key to the

24 K. Bjorlykke

XRD analyses from the Garn Formation at Haltenbanken show, however, that samples with relatively high kaolinite content at 4 km depth have little K-feldspar (Ehrenberg, 1991). Potassium appears not to have been supplied by diffusion from K-feldspar 1 0 - 2 0 m away. Mudstones containing smectite may represent a sink for K from adjacent sandstones during illitization but if the mudstones contain K-feldspar there is no concentration gradient to drive such transport.

Illite datings. A large number of K-Ar dates of illite from the North Sea and Haltenbanken have been published. The ages obtained from Jurassic reservoir sandstones range from 100-30 Ma, often with a concentration of ages between 40-60 Ma (Thomas, 1986; Liewig et al., 1987; Jourdan et al., 1987; Glasmann et al., 1989a,b). Analyses of Jurassic sandstones from Haltenbanken gave ages from 55-31 Ma, but Ehrenberg & Nadeau (1989) interpreted these ages to be much too old due to contamination of old feldspar and illite and interpreted the illite to have formed in the last few Ma at temperatures close to 140~ Also from other basins, like the Paris Basin, the possible detrital contamination of illite and the validity of K/Ar datings have been debated (SpiStl et aL, 1996; Clauer et al., 1996). If these illite dates represent the time of the main phase of illite growth, this poses several problems. The problems of such datings have recently been discussed in detail by Clauer & Chaudhuri (1995). If the Early Tertiary Jurassic reservoir sandstones, presently at -4 km depth, were buried to only about 2 km or less, the illite must then have formed at rather low temperatures (50-80~ or the geothermal gradients must have been about twice that of the present day for a long time in the Late Cretaceous and Early Tertiary. However, the distribution of authigenic illite in reservoir rocks seems to be strongly controlled by the present burial depth. In the North Sea and the Haltenbanken basins, there is a strong increase in the amount of illite below 3.7-4.0 km depth. Below this depth authi- genic kaolinite can commonly be observed to have been replaced by illite (Bjorlykke et aL, 1986,, 1992; Ehrenberg & Nadeau, 1989; Ehrenberg, 1990).

In the Brent Group, a marked decrease in the K-feldspar content is commonly observed suggesting that it has been dissolved in the process of illitization of kaolinite and possibly also smectite (Bjorlykke et al., 1992). In the Brent Group, the illite content in sandstones commonly increases at 3 . 7 - 4 k m (11000-12000 It) depth

(Giles et al., 1992; Scotchman et aL, 1989). This is the same depth as Haltenbanken despite lower Late Cenozoic subsidence suggesting that the illitization is controlled by the present depth.

A lower degree of illitization and more unaltered kaolinite have been observed in reservoir sand- stones like those of the Hild Field where the K- feldspar content is very low, suggesting that the supply of K is the limiting factor for illitization (Lonoy et al., 1986; Bjorlykke et al., 1992; Thyberg, 1993).

The amount of kaolinite formed at shallow depth can be the limiting factor for the formation of illite during deeper burial diagenesis, particularly in the distal shelf and ~n'bidite facies where meteoric water flushing is not usually very effective and contains little authigenic kaolinite, thus reducing the potential for illitization (Fig. 2).

The illite datings suggest that authigenic illite formed at 75~ at -2 km or less (Hamilton et al., 1992). As documented above, a pronounced increase in the illite content is observed at -3.7-4.0 km depth, both in the North Sea, although the thickness of the Pliocene/Pleistocene is much greater at Haltenbanken (1 km) than in the North Sea, suggesting that the distribution of illite is controlled by the present burial depth. Observations both from the North Sea and Haltenbanken (Ehrenberg, 1991; Lonoy et al., 1986; Thyberg, 1983) show that kaolinite remains stable to higher temperatures than 140~ (4 km) when K-feldspar is not locally available to supply the K. The reduced illitization and improved reservoir quality can then be related to the amount of clastic K-feldspar.

Chlorite. Chlorite is common as a clastic mineral in the Pliocene-Pleistocene sequences of the North Sea Basin because of limited weathering in this partly glacially-influenced cold climate. In the warmer climate of the Lower Tertiary and Mesozoic, the clastic sediments supplied to the North Sea probably contained little chlorite. Most of the chlorite minerals in the Lower Tertiary sequence were probably formed diagenetically during early diagenesis on the sea floor or by burial diagenesis from smectite or volcanic detritus. Similar smectite-rich mudstones of volcanic origin have been found in Upper Mesozoic and Lower Tertiary sequences along the Atlantic margin (Chamley, 1992). Authigenic chlorite occurs as grain-coating cement in some sandstones and is particularly common in the Jurassic Tilje Formation at Haltenbanken (Ehrenberg, 1993). Grain-coating

Page 11: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins 25

FIG. 6. Authigenic chlorite coating quartz grains retarding the growth of authigenic quartz thus preserving abnormally high porosity. From Jurassic reservoirs buried to -5 km at Haltenbanken.

chlorite cement (Fig. 6) inh"bits quartz overgrowth, thus preserving higher porosity than normal at depths of 4.0-5.5 km (Ehrenberg, 1993). Chlorite crystals become more coarse grained with increasing burial depth as a result of grain coarsening (Jahren & Aagaard, 1989).

Precipitation of chlorite requires a source of Fe and Mg and possible sources are clastic biotite, basic rock fragments and volcanic rock fragments or early diagenetic Fe minerals (verdine and glaucony) formed in deltaic or estuarine environments by the supply of Fe from rivers as demonstrated from the Niger Delta (Odin et al., 1988; Ehrenberg, 1993). It is possible that such early diagenetic Fe minerals formed in tidal or estuarine environments could be important precursors for chlorite coatings forming at greater depth. In sandstones of the Brent Group, the North Sea chlorite cement is in most cases rare or absent (Giles et al., 1992; Bjorlykke et al., 1992), possibly because it represents mostly a proximal marine and fluvial facies.

Lower Jurassic Statfjord Formation and Intra Dunlin sand in the Veslefrikk Field of the North

Sea, however, contain chlorite coatings (Ehrenberg, 1993; Hillier, 1994).

D I A G E N E S I S OF M U D S T O N E S A N D S H A L E S

Mechanical compaction is the reduction in volume due to reorientation and breakage of grains, a function of grain strength and effective stress. Chemical compaction involves mineral dissolution and precipitation and is a function of mineral stability and kinetics of precipitation of cements, processes that are strongly influenced by tempera- ture. These two types of compaction have very different driving forces and must be treated separately. In basin modelling, however, compac- tion of mudstones is assumed to be a function of effective stress (Hermanrud et al., 1991; Illiffe & Dawson, 1996). Hermanrud pointed out the great variations in shale porosity in the published data, i.e. from Rieke & Chillingarian (1974). Similar variations are found in North Sea mudstones but these are functions of the initial grain size and

Page 12: Clay mineral diagenesis in sedimentary basins a key to the

26 K. Bjgrlykke

mineralogy and can be predicted. As a general rule, it can be stated that with increasing burial depth, the rate of compaction becomes more chemical and more a function of temperature and less of the effective stress.

Mechanical compaction o f mudstones

The North Sea and Haltenbanken basins are characterized by three main types of mudstones and mixtures of them: (1) Glacial marine mudstones of Pliocene and Pleistocene age. These are miner- alogically immature mudstones with a high feldspar content, dominantly clastic chlorite and also, frequently, unstable rock fragments such as pyroxenes and amphiboles (Thyberg, personal communication; Rundberg, 1989; Karlsson et al., 1979). (2) Smectitic mudstones, mostly of Lower Tertiary age. In particular, Eocene and Oligocene mudstones representing a distal facies may have a very high smectite content (>50%) and almost no quartz or feldspar (Huggett, 1992; Thyberg, personal communication). These mudstones are derived from volcanic ash resulting from subaerial volcanicity during the opening of the Norwegian- Greenland Sea. It is also present onshore in Denmark (Nielsen & Heilman-Clausen, 1988). Mudstones of more proximal facies contain more kaolinite and quartz (Rundberg, 1989; Thyberg et aL, 1998). (3) Mesozoic mudstones and shales consisting mostly of illite and variable amounts of kaolinite, smectite and mixed-layer I-S and chlorite. Most of these sediments have been buried to >2.0-3.0 km and the smectite and the mixed- layer content may have been higher at shallower depths of burial. These types of mudstones have very different properties, particularly during mechanical compaction (Rieke & Chillingarian, 1974). Mudstones and shales should, therefore, not be treated as one category during basin modelling. The difference between these types, with respect to compaction, is clearly seen on velocity and density logs from the North Sea (i.e. 34/7-1) (Thyberg, personal communication).

The Pliocene and Pleistocene glacial marine mudstones are characterized by high velocities (2.5-2.7 kin/s) and densities (low porosity), producing a strong velocity and density inversion compared to the underlying smectitic mudstones, which have typical velocities o f - 2 . 0 km/s. The driving force for mechanical compaction is the effective stress (~e) transmitted at the grain

contacts. Compaction causes reduction of pore- space and can only occur if the fluid (water) is able to escape. The coarse-grained glacial marine sediments are relatively permeable, allowing rapid dewatering, and thus avoiding overpressure. The smectitic mudstunes have very high surface area and low permeability and in such fine-grained sediments the rate of fluid expulsion may be the rate-limiting process in compaction (Thyberg, personal communication). Fractures in mudstones can only remain open if the pore-pressure is equal to the least horizontal stress (~h), which in such sediments is 80-90% of the vertical stress (Crv) (Garenstroom et al., 1993). The effective stress (~e) to drive compaction would then be very small and there would be low pressure gradients for water to flow into the fracture. In addition, the fracture must connect with high permeability pathways all the way to the surface. According to Gaarenstrom (1993) the degree of overpressure in the Central graben may have been increasing through the Tertiary reaching fracture pressures in the Late Tertiary. The effective stress would then have been reduced and the mechanical compaction must have stopped, but the chemical compaction which is not so sensistive to stress has continued.

Chemical compaction o f mudstones and shales

The onset of chemical compaction depends on the stability of the mineral phases. As discussed above, smectite is replaced by mixed-layer minerals and illite in the temperature range of 70-100~ which, in basins like the North Sea, corresponds to burial depths of 2 -3 km and this causes compaction which could not have been obtained mechanically. Experimental data by Chillingarian & Knight (1960) suggest that even at effective stresses corresponding to 15000 ft, the porosity of the montmorillonite clay is only reduced to ~45%, while kaolinite and illite are compacted much more efficiently at the same overburden stress (Fig. 7). This suggests that smectitic mudstones can only compact mechanically to -30-40%, depending on the smectite content. Both porosities derived from density logs (Fig. 7) and from analyses of core samples (Tyridal, 1994; Tyridal, personal communication) suggest that Lower Tertiary smectitic mudstones have porosities in this range down to ~2 km. Further compaction probably depends on the transformation of smectite to illite via mixed-layer minerals, which increases the particle

Page 13: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins 27

1 I t ~

2

3 �84

4

o 0 .5 o

6

7

.o 8 L

n g

0

--" 10 n

r~ 11

t - 12 o

13 o E

1 4 " 0

1 5 0 2 0 4 0

P o r o s i t y ( ' / . )

E x p e r i m e n t a l

t i 6 0 8 0

Gul f Coast T e r t i a r y

- V e n e z u e l a T e r t i a r y

-- - -- "- - C o m l S o s i t e - R e c e n t M i o

Oklahoma - Palaeozoic

. . . . . Gulf C o a s t - T e r t i a r y

. . . . . . . JQpo~- Tertiary

Experimental

F~G. 7. Experimental and observed porosity depth curves from mudstones. Porosity depth data (points) for Tertiary mudstone from the Central Graben based on density logs and estimates of mineral matrix composition from six wells (Lauvrak, 1996). The mudstones are fine grained and smectite-rich and show little reduction in porosity down to nearly 2 km depth. This may be due partly to moderate overpressure reducing the effective stress. Experimental compaction curves (Chillingarian & Knight, 1960) show that pure clay minerals do not compact readily even at stresses equivalent to 5 km. The much greater rates of compaction between 2 - 3 km depth may reflect chemical compaction, mainly dissolution of smectite and precipitation of mixed-layered

minerals and illite.

size and therefore the permeability and the rate of compaction.

Also in shales, K-feldspar reacts with kaolinite to form illite, but the reaction rate may be lower due to slower diffusion of K from K-feldspar in the shale mat r ix which has low poros i ty and permeability.

Kaolinite in shales, as in sandstones, is subject to dissolution and precipitation of illite at depths close to 4 km (130-140~ if K-feldspar is present to

supply K. Authigenic illite is more fine grained than the dissolving kaolinite and this process may contribute to the reduction in permeability.

Clastic kaolinite may be a part of the clastic framework carrying effective stress. When dissolved, illite may precipitate in the available pore-space and allow for more efficient compaction. Both the reaction from smectite to illite and kaolinite to illite release water, which may contribute to the build-up of pore-pressure. Dehydration of minerals

Page 14: Clay mineral diagenesis in sedimentary basins a key to the

28 K. Bjor lykke

involves a partial phase change from solid to fluid thus increasing the porosity and fluid/solid. Dehydration of clay minerals can generate a significant percentage of the total compaction- driven flux. Shales containing 20% kaolinite may generate water corresponding to ~4% of the rock volume which could contribute to the build-up of overpressure (Bjorlykke, 1996). Most North Sea mudstones, however, will probably have had a lower initial kaolinite content. Compaction and generation of petroleum are in most cases important contribu- tors to overpressure (Buhrig, 1989). If a constant porosity is assumed and the permeability/depth curve is constant, the modelling of overpressure will necessarily be a function of the sedimentation rate. However, since both mechanical and chemical sediment compaction of mudstones are a function of time, high sedimentation rates will imply that the porosity and permeability at a certain depth is also higher. The higher permeability may partly or totally compensate for the increased flux due to the higher sedimentation rate. The observed porosity/depth trends of mudstones vary greatly as a function of grain size and mineralogy (Fig. 7). As shown above, there is a limit to how much mudstones will compact mechanically and dissolution and precipitation of minerals make it possible to form a mineral fabric with a much lower porosity (Fig. 8). Mudstones often have silt- or sand-sized quartz grains and dissolution and reprecipitation of quartz may be an important factor in producing closer packing (Fig. 8). This process has been shown in sandstones to be mainly a function of temperature and textural relationships and relatively insensitive to variations in stress (Bjorkum, 1995; Walderhaug, 1996). The same processes probably also apply to mudrocks and shales but it is more difficult to study textural relationships between quartz, mica and clay minerals in such f ine -g ra ined sediments (Fig. 9). Overpressures built up at depths where chemical compaction is dominant (>2-3 km depth, 70-100~ should not be expected to have significantly higher porosity than normally pressured rocks. Chemical compaction cannot be modelled based on effective stress because the temperature is the most important factor (Fig. 9). The transition from mechanical to dominantly chemical compaction is not fixed and will depend on the mineralogy and the burial history. Smectite-rich mudstones will compact chemically at shallower depth than kaolinite-rich mudstones. Both mechanical and chemical compaction provide a ductile response to

tectonic stress (low strain rates) and will therefore reduce the potential to transmit plate tectonic stress in sedimentary basins during subsidence (Bjorlykke & Hoeg, 1997).

D I S C U S S I O N

Predictions of burial diagenetic reactions depend on whether or not the chemical composition of the sediments can be assumed to be constant during burial. Changes in the bulk composition during burial must be due to transport in pore-water by diffusion or by fluid flow (advection). With increasing temperature, the pore-water approaches equilibrium with the constituent minerals and the concentration gradients for driving diffusion will be very small. Where there are important differences in the primary lithology and mineral assemblage, there may locally be higher concentration gradients between pore-waters buffered by different minerals. The presence or absence of minerals like calcite, K- feldspar and kaolinite may be critical for such buffering producing concentration gradients near lithological boundaries (Thyne et al., 1996). It is difficult to estimate how effective diffusion driven by mineral buffered pore-water is, but it is significant probably to the order of a few metres. At temperatures >80-100~ the pore-water is probably close to quartz saturation and the K concentration will normally be in the stability field of illite and in some cases also chlorite. The K § and Mg ++ may be transported by diffusion locally but the North Sea formation water does not have high concentrations of these elements except near evaporites (Warren & Smalley, 1994). During early marine diagenesis near the sea floor there may be effective diffusion from sea water into the sediments. Fluid flow rates are orders of magnitude greater during meteoric water flushing than during burial diagenesis (Giles, 1987). The potential for mass transport is also greater during early diagenesis because at low temperatures the pore- water may be highly over saturated or under saturated with respect to mineral phases. During burial diagenesis, when the temperature is greater and the flow rates are much smaller, the pore-water is closer to equilibrium with the minerals present. The volume of minerals dissolved or precipitated due to advective flow can then be calculated (Wood & Hewett, 1984; Bjorlykke, 1994):

Vc = F t sin[3(0T/~Z) ~v/p (3)

Page 15: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins

IO'v (overburden , . . . . . . . lit _ stre_ ss) "

I i t I

, I

I

O C - 'Gv - Pp

_ _ J

Mechanical compaction-

a function of effect ive stress oe

X ) f" . . . . . . . . . . 4

I I'+ Effective stress at grain contacts G e

Chemical compaction involving dissolution and precipi tat ion of minerals mainly as a function of temperature

Smectitic mudstones (30-60% porosity) (partly bound water)

70~ / Via mixed-

I layer minerals to

lo0oc , Illite

29

M ~ : : ~ - , . ~ = Clastic kaolinite

l L { { 7 ~ t7 : , : ; , # , ' : : 7 : " I b__~_:• _ ~ _ - ~ _ ~ l Dissolution of clastic kaolinite and precipitation of authigenic illite in the avilable pores will increase compaction.

FQrther compaction depends on the dissolution and precipitation of quartz

FIG. 8. Schematic representation of mechanical and chemical compaction of mudstones.

Here, Vc is the volume of cement precipitated, F is the total flux of pore-water (cm3/cm 2 s - - l ) , t =

time (s), ~ is the angle between the direction of flow and the isotherm, ~T/~Z is the geothermal

gradient, Ctr is the solubility-temperature function (transfer coefficient) and p is the density of the mineral. In the case of clay minerals, the mobility of A1 is particularly critical. Calculations using

Page 16: Clay mineral diagenesis in sedimentary basins a key to the

30 K. Bj#rlykke

2-3 km

D E P T H

4kin

Porosity IP

m e compaction .

/ Effective stress: / Compact,on / ~v=.(prgh-P) and /

compaction modulus t 70-1000C

~ " Chemical i compaction

j dominant. / " Mainly a function

of temperature and mineralogy

FIG. 9. Relationships between compaction, effective stress and temperature in mudstones. The depth at which chemical compaction becomes dominant over mechanical compaction depends on the mineralogy and textural

relationships.

SOLMINEQ 88 show that the solubility of A1 is <1 ppm at temperatures <140~ and that the solubility is not increased in the presence of o r g a n i c a c i d s ( B j o r l y k k e et a l . , 1995). Significantly higher A1 concentrations have not been reported from North Sea reservoirs. Because of the low mobility of AI, growth of diagenetic clay minerals like illite and chlorite require a local precursor aluminous mineral like smectite and kaolinite and the supply of K and Mg. The distribution of illite and chlorite at depth must be linked to provenance and facies and the distribution of early diagenetic kaolinite due to meteoric water flushing (Fig. 5). Modelling of compaction driven flow indicates that flow rates are very low so that the system is characterized by low Peclet numbers and thus both heat transport and mass transport are dominated by diffusion (Bethke, 1985; Ungerer et al., 1990; Ludviksen et al., 1993; Bjorlykke, 1994). Diagenesis is still, to a large extent, based on what was traditionally called 'sedimentary petrology' and careful mineralogical petrographic data and obser- vations are always valuable. Diagenetic theories should be tested and calibrated against observations,

but the processes cannot be inferred from petro- graphic data alone. Changes in sediment composi- tion and mineralogy with burial depth are often inferred to be due to diagenetic processes. In a single well there are frequently very distinctive stratigraphical variations in the lithology. To observe changes in one stratigraphic interval with depth, several wells where this unit is cored must be studied and it must be considered that the observed trend with depth may also be due to other factors such as provenance and facies. Because of lateral variations in facies and provenance, it is impossible to study the same rock at different burial depths. Sandstones may vary laterally with respect to facies and provenance but this is also the case with mudstones. An increase in illite and K content with depth may be due to primary enrichment of illite or smectite in the distal fine-grained facies. Because there may always be lateral changes in the primary composition of sandstones and shales, geochemical changes cannot be inferred from sampling at different burial depths.

I f the sediments undergo diagenesis in a geochemically closed system, the diagenetic reac-

Page 17: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins 31

tions and the rock properties of the sediments are direct functions of the primary sediment composi- tion and the burial history. Predictions of burial diagenetic reactions must therefore be linked to facies and provenance. Rock properties like porosity, permeability, specific surface and thermal conductivity all depend very much on the primary clay content and on the burial diagenetic reactions.

C O N C L U S I O N S

Clay minerals play a crucial role in both mechanical and chemical compaction of sandstones and shales. The amount and type of clay mineral is a function of the provenance of clastic minerals and of diagenetic reactions at shallow and greater depth. Early diagenetic reactions may be relatively open (from a geochemical point of view), due to diffusion from seawater or due to meteoric water flow. Dissolution of feldspar and mica and precipitation of authigenic kaolinite require that K and silica are removed by fluid flow as is the case during weathering. During burial diagenesis, the pore-water flow is very much smaller and calculations show that advective transport of dissolved ions is in most cases relatively insignif- icant. Diffusion, however, may be significant over shorter distances depending on differences in mineral composition controlling the concentration gradients of the dissolved ions. Clay mineral reactions are therefore close to being isochemicaI within the range of effective diffusion transport (1-10 m?). Sandstones without clay minerals such as kaolinite and smectite and unstable rock fragments will remain stable and porosity will then be reduced mainly by mechanical compaction and dissolution/precipitation of quartz and carbo- nate minerals.

Prediction of diagenetic reactions at greater burial must be linked to provenance facies and early diagenetic reactions. North Sea formation water is nearly always in the stability field of illite, but precipitation of illite requires an aluminous precursor mineral like smectite or kaolinite and is mainly a function of temperature and supply of K from K-feldspar. Mudstones and shales from the North Sea and Haltenbanken vary greatly in mineralogical composit ion and grain size. Immature, partly glacial Pliocene and Pleistocene mudstones in which illite and chlorite are the main clay minerals, compact much faster and have

greater velocities than the underlying Eocene and Oligocene smectitic mudstone and commonly develop overpressures. Dissolution of smectite and kaolinite and precipitation of illite releases crystal- bound water which adds to the fluid flux and may potentially contribute to the build-up of over- pressure. At burial depths >2.0-3.0 km, porosity loss by compaction is mainly chemical, involving dissolution and precipitation of minerals and this compaction can not be calculated as a function of effective stress as is assumed in most basin modelling programmes. Realistic prediction of rock properties can only be made if the primary sediment composition and burial diagenetic processes are known. Mineralogy, temperature and time are the main factors controlling compaction both of reservoir sandstones and shales, and the degree of overpressure plays a minor role at depths >2-3 km (70-100~ During progressive burial, sediments and mudstones in particular compact both mechanically and chemically and show a ductile response to stress at slow strain rates. This reduces the transfer of tectonic stress in sedimentary basins.

ACKNOWLEDGMENTS

Financial support from the Norwegian Research Council (NFR) and from Norwegian oil companies and useful discussions with Per Aagaard are gratefully acknowledged. Two anonymous reviewers provided useful comments to help improve the manuscript.

REFERENCES

Aagaard P. & Helgeson H.C. (1983) Activity/composi- tion relations among silicates and aqueous solutions: II. Chemical and thermodynamic consequences of ideal mixing of atoms of homological sites in montmorillonites, illites, and mixed-layer clays. Clays Clay Miner. 31,207-217.

Aagaard P., Jahren J.S. & Egeberg P.K. (1992) North Sea clastic diagenesis and formation water con- straints. Pp. 1147-1152 in: Water-Rock Interaction Proceedings (Y.K. Kharaka & A.S. Maest, editors). Park City, Utah, USA.

Bethke C.M. (1985) A numerical model of compaction driven flow and heat transfer and its application to the paleohydrology of intercratonic sedimentary basins, d. Geophys. Res. 90, 6817-6828.

Bjorkum P.A. (1995) How important is pressure in causing dissolution of quartz in sandstones? J. Sed. Res. 66, 147-154.

Bjorkum P.A., Mjos A., Walderhaug O. & Hurst A. (1990) The role of the late Cimmerian unconformity

Page 18: Clay mineral diagenesis in sedimentary basins a key to the

32 K. Bjorlykke

for the distribution of kaolinite in the Gullfaks Field, Northern North Sea. Sedimentology, 37, 395-406.

Bjorlykke K. (1983) Diagenetie reactions in sandstones. Pp. 169-213 in: Sediment Diagenesis. NATO Advanced Study lnst. (A. Parker & B.W. Sellwood, editors). Reidel Publ. Co., Reading, U.K.

Bjorlykke K. (1994) Fluid flow and diagenesis in sedimentary basins. Pp. 127-140 in: Geofluid: Origin, Migration and Evolution of Fluids in Sedimentary Basins (J. Parnell, editor). Geol. Soc. London Spec. Publ. 78.

Bjorlykke K. (1996) Lithological control on fluid flow in sedimentary basins. Pp. 15-34 in: Fluid Flow and Transport in Rocks -- Mechanisms and Effect (B. Jamtveit & B.W.D. Yardley, editors). Chapman & Hall, London.

Bj~rlykke K. & Aagaard P. (1992) Clay Minerals in North Sea Sandstones. Pp. 65-80 in: Origin, Diagenesis, and Petrophysics of Clay Minerals in Sandstones (D.W. Houseknecht & E.D. Pittrnan, editors). SEPM Special Publication 47.

Bjorlykke K. & Brendsdal A. (1986) Diagenesis of the Brent sandstone in the Statfjord Field, North Sea. Pp. 57-166 in: Roles of Organic Matter in Sediment Diagenesis (D.L. Gautier, editor). SEPM Spec. Publ. No. 38.

Bjorlykke K. & Hoeg K. (1997) Effects of burial diagenesis on stress, compaction and fluid flow in sedimentary basins. Mar. Petr. Geol. 14, 267-276.

Bjorlykke K., Aagaard P., Dypvik H., Hastings D.S. & Harper A.S. (1986) Diagenesis and reservoir proper- ties of Jurassic sandstones from the Haltenbanken area, offshore mid-Norway. Pp. 275-286 in: Habitat of Hydrocarbons on the Norwegian Continental Shelf (A.M. Spencer et al., editors). Nor. Petr. Soc. (Graham & Trotman).

Bjorlykke K., Aagaard P., Egeberg P.K. & Simmons S.P. (1995) Geochemical constraints from formation water analyses from the North Sea and Gulf Coast Basin on quartz, feldspar and illite precipitation in reservoir rocks. Pp. 33-50 in: The Geochemistry of Reservoir (J.M. Cubitt & W.A. England, editors). J. Geol. Soc. London Spec. Publ. 86.

Bjorlykke K., Nedkvitne T., Ramm M. & Saigal G. (1992) Diagenetic processes in the Brent Group (Middle Jurassic) reservoirs of the North Sea - - an overview. Pp. 263-287 in: Geology of the Brent Group (A.C. Morton, R.C. Haszeldine, M.R. Giles & S. Brown, editors). Geol. Soc. London Spec. Publ. 61.

Boles J.R. & Franks, S.G. (1979) Clay diagenesis in Wilcox sandstones. J. Sed. Pet. 49, 55-70.

Buhrig C. (1989) Geopressured Jurassic reservoirs in the Viking Graben: modelling and geological signifi- cance. Mar. Petr. Geol. 6, 31-48.

Burley S.D. (1986) The development and destruction of porosity within upper Jurassic reservoir sandstones

of the Piper and Tartan fields, outer Moray Firth, North Sea. Clay Miner. 21, 649-694.

Burley S.D., Kantoriwicz J.D. & Waugh B. (1985) Clastic diagenesis. Pp. 189-226 in: Sedimentology: Recent and Applied Aspects, (P. Brenchley & B.P.J. Williams, editors). Geol. Soc. London Spec. Publ. 18.

Chamley H. (1992) Clay Mineral Diagenesis. Pp. 161-188 in: Quantitative Diagenesis: Recent Developments and Applications to Reservoir Geology (A. Parker & B.W. Sellwood, editors). Kluwer.

Chillingarian G.V. & Knight L. (1960) Relationship between pressure and moisture content of kaolinite, illite and montmorillonite clays. A.A.P.G. Bull. 44, 101-106.

Clauer N. & Chaudhuri S. (1995) Clays in Crustal Environments. Isotope Dating and Tracing. Springer Verlag.

Clauer N., O'Neil J.R., Furlan S. & Mossmann J.R. (1996) Clay minerals as recorders of temperature conditions and duration of thermal anomalies in the Paris Basin, France: A reply to the discussion by Sp6tl et al. Clay Miner. 31, 209-215.

Dypvik H. (1983) Clay mineral transformations in Tertiary and Mesozoic Sediments from the North Sea. A.A.P.G. Bull. 67, 160-165.

Egeberg P.K. & Aagaard P. (1989) Origin and evolution of formation waters from oil fields on the Norwegian Shelf. Appl. Geochem. 4, 131-142.

Ehrenberg S.N. (1990) Relationship between diagenesis and reservoir quality in sandstone. A.A.P.G. Bull. 74, 1538-1559.

Ehrenberg S.N. (1991) Kaolinized, potassium-leached zones at the contacts of the Garn Formation, Haltenbanken, mid-Norwegian Continental Shelf. Mar. Petrol Geol. 8, 250-269.

Ehrenberg S.N. (1993) Preservation of anomalously high porosity in deeply buried sandstones by grain- coating chlorite: Examples from the Norwegian Continental Shelf. A.A.P.G. Bull. 77, 1260-1286.

Ehrenberg S.N. & Nadeau P.H. (1989) Formation of diagenetic illite in sandstones of the Garn Formation, Haltenbanken area, mid-Norwegian Continental Shelf. Clay' Miner. 24, 233-253.

Ehrenberg S.M., Aagaard P., Wilson M.J., Fraser A.R. & Duthie D.M.L. (1993) Depth-dependent transfor- mation of kaolinite to dickite in sandstones of the Norwegian Continental Shelf. Clay Miner. 28, 325-352.

Foscolos A.E. (1984) Diagenesis 7, Catagenesis of Argillaceous Sedimentary Rocks. Geosci. Canada, 11, 67-74.

Gaarenstroom L., Tromp R.A.J., Jong M.C. de & Brandenburg A.M. (1993) Overpressures in the Central North sea: implication for trap integrety and drilling safety. Pp 1305-1313 in: Petroleum

Page 19: Clay mineral diagenesis in sedimentary basins a key to the

Clay mineral diagenesis in sedimentary basins 33

Geology of Northwest Europe. Proc. 4th Conf. (J.R. Parker, editor).

Garrels R.M. & Christ C.L. (1965) Solutions, Minerals and Equilibira. Harper & Row.

Giles M.R. (1987) Mass transfer and problems of secondary porosity creation in deeply buried hydro- carbon reservoirs. Mar. Petr. Geol. 4, 188-201.

Giles M.R., Stevenson S., Martin S.V. & Cannon S.J.C. (1992) The reservoir properties and diagenesis of the Brent Group: a regional perspective. Pp. 289-327 in: Geology of the Brent Group (A.C. Morton, R.S. Haszeldine, M.R. Giles & S. Brown, editors). Geol. Soc. London Spec. Paper 61.

Glasmann J.R., Larter S., Briedis N.A. & Lundegard D. (1989b) Shale diagenesis in the Bergen High Area, North Sea. Clays Clay Miner. 37, 97-112.

Glasmann J.R., Lundegard P.D., Clark R.A., Penny B.K. & Collins I.D. (1989a) Geochemical evidence for the history of diagenesis and fluid migration: Brent sandstone, Hether Field, North Sea. Clay Miner. 24, 255-284.

Glennie K.W. (1990) Introduction tO the Petroleum Geology of the North Sea. Blackwelt, Oxford.

Hamilton P.J., Giles M.R. & Ainsworth P. (1992) KJAr datings of illite in Brent reservoir: a regional perspective. Pp. 377-400 in: Geology of the Brent Group (A.C. Morton, R.S. Haszeldine, M.R. Giles & S. Brown, editors). Geol. Soc. London Spec. Publ. 61.

Hancock J.M. (1984) Cretaceous. Pp~ 133-150 in: Introduction to the North Sea (K.W. Glennie, editor). Btackwell, pp 133- t50.

Hermanrud C., Eggen S. & Larsen R.M. (1991) Investigation of the thermal regime of the H0rda platform by basin modelling: implications for the hydrocarbon potential of the Stord Basin, northern North Sea. Spec. Publ. Europ, Assoc. Petrol. Geosci. 1, 65-73.

Hillier S. (1994) Pore-lining chlorites in siliciclastic reservoir sandstones; electron microprobe, SEM and XRD data, and implications for their origin. Clay Miner. 29, 665-679.

Hower J., Eslinger E.V. & Perry E.A. (1976) Mechanism of burial metamorphism of argillaceous sediments: mineralogical and geochemical evidence. Geol. Soc. Amer. 87, 725-737.

Huggett J.M. (1992) Petrography, mineralogy and diagenesis of overpressured Tertiary and Late Cretaceous mudrocks from the East Shetland Basin. Clay Miner. 27, 487-506.

Hutcheon I., Shevalier M. & Abercrobie H.J. (1992) pH buffering by metastable mineral fluid equilibria and evolution of carbon dioxide fugacity during burial diagenesis. Geochim. Cosmochim. Acta, 57, 1017-1027.

Illiffe J.E. & Dawson M.R. (1996) Basin modelling history and predictions. Pp. 83- t05 in: AD, 1995: NW Europe's Hydrocarbon Industry (K. Glenn]e &

H. Hurst, editors). Geol. Soc. London. Jahren J.S. & Aagaard P. (1989) Compositional varia-

tions in diagenetic chlorites and illites, and relation- ships with formation water chemistry. Clay Miner. 24, 157-170.

Jordt H., Faleide J.I., Bjorlykke K. & Ibrahim M.T. (1995) Cenozoic sequence stratigraphy in the Central and Northern North Sea: tectonic development, sediment distribution and provenance areas. Mar. Petrol. Geol. 12, 845-879.

Jourdan A., Thomas M., Brevart O., Robson P., Sommer F. & Sullivan M. (1987) Diagenesis as the control of the Brent sandstone reservoir properties in the Greater Alwyn area (East Shetland Basin). Pp. 951-961 in: Petroleum Geology of North West Europe (J. Brooks & K. Glennie, editors). Graham & Trotman.

Karlsson W., Vollset J., Bjorlykke K. & Jorgensen P. (1979) Changes in mineralogical composition of Tertiary Sediments from North Sea wells. Proc. Int. Clay Conf. Oxford, 281-289.

Koch LO. & Heum O.R. (1995) Exploration trends of the Halten Terasse. Pp. 36-42 in: Petroleum Exploration and Exploration in Norway (S. Hanslien, editor). Norw. Petr. Soc. Spec. Publ. 4.

Lauvrak O. (1996) Overtrykk i sentralgraben. En studie basert pd HPHT-bronner. (Over-pressure in the Central Graben. A study based on HPHT wellS). Cand. Scient. thesis, Univ. Oslo, Norway.

Liewig N., Clauer N. & Sommer F. (1987) Rb-Sr and K/ Ar Dating of Clay Diagenesis in Jurassic Sandstone Oil Reservoir, North Sea. A.A.P.G. Bull. 71, 1467-1474.

Ludvigsen A., Gran K., Palm E. & Bjorlykke K. (1993) Effects of thermal convection-currents on heat transfer in sedimentary basins. Pp. 353-359 in: Basin Modelling. Advances and Applications (A.G. Dor6 et al., editors). Norw. Petr. Soc, Spec. Pubi. 3.

Lonoy A., Akselsen J. & Ronning K. (1986) Diagenesis of deeply buried sandstones from a deeply buried sandstone reservoir: Hild Field, Northern Nortk Sea. Clay Miner. 21,497-511.

McAulay G.E., Burley S.D., Fallick A.E. & Kusznir N.J. (1990) Palaeohydrodynamic fluid flow regimes during diagenesis of the Brent Group in the Hutton-NW Hutton reservoirs; Constraints from oxygen isotope studies of authigenic kaolin and reverse flexural modelling. Clay Miner. 29, 609-626.

Morton A.C., Haszeldine R.C., Giles M.R. & Brown S. (1992) Geology of the Brent Group. GeoL Soc. London Spec. Publ. 61.

Nedkvitne T. & Bj~rlykke K. (1992,) Secondary porosity in the Brent Group (Middle Jurassic) Huldra Field, North Sea; implication for predicting lateral con- tinuity of sandstones. J. rSed. Pet. 62, 23-34~

Nielsen O.B. & Heilmann-Ctausen C. (1988) Paleogene

Page 20: Clay mineral diagenesis in sedimentary basins a key to the

34 K. Bjorlykke

volcanism: the sedimentary record in Denmark. Pp. 395-405 in: Early Tertiary Volcanism and the Opening of the NE Atlantic (A.C. Morton & L.M. Parson, editors). Geol. Soc. London Spec. Publ. 39.

Odin G.S., Debeney J.B. & Masse J.M. (1988) The verdine facies identified in 1988. Pp 131-148 in: Green Marine Clays (G.S. Odin, editor). Devs. in Sedimentology 45.

Osborne M., Haszeldine R.S. & Fallick A.E. (1994) Variations in kaolinite morphology with temperature in isotopically mixed pore-fluids, Brent Group, UK, North Sea. Clay Miner. 29, 59l -608.

Pittman E.D. (1978) Porosity, diagenesis and productive capability of sandstone reservoirs. Pp 2159-2173 in: Aspects o f Diagenesis (P.A. Scholle & P.R. Schluger, editors). SEPM Spec. Publ. 26.

Rieke H.H. & Chillingarian G.V. (1974) Compaction of Argillaceous Sediments. Devel. Sedimentol. 16. Elsevier.

Rundberg Y. (1989) Tertiary sedimentary history and basin evolution between 60-62~ an integrated approach. Dr. Eng. thesis, Univ. Trondheim, Norway.

Saigal G.C., Bjorlykke K. & Larter S.R. (1992) The Effects of oil emplacements on diagenetic processes

- - examples from the Fulmar Reservoir sandstones, Central North Sea. A.A.P.G. Bull. 76, 1024-1033.

Sass B.M., Rosenberg P.E. & Kittfick A. (1987) The stability of illite/smectite during diagenesis: An experimental study. Geochim. Cosmochim. Acta, 51, 2103-2115.

Schmidt V. & McDonald D.A. (1979) The role of secondary porosity in the course of sandstone diagenesis. Pp 209-225 in: Aspects of Diagenesis (P.A. Scholle & P.R. Schluger, editors). SEPM Spec. Publ. 26.

Scotchman I.C., Jones L.H. & Miller R.S. (1989) Clay mineral diagenesis and oil migration in the Brent Group, NW Hutton Field, UK, North Sea. Clay Miner. 24, 339-374.

Singh B. (1996) Study of diagenesis and provenance of the sedimentary sequences in the Northern North Sea. Cand. Scient. Thesis, Univ. Oslo, Norway.

Spark I.S.C. & Trewin N.H. (1986) Facies related diagenesis in the main Claymore oilfield sandstones. Clay Miner. 21,479-496.

Spotl C., Worden R.H. & Walgenwitz F. (1996) Clay minerals as recorders of temperature conditions and duration of thermal anomalies in the Paris Basin, France: discussion. Clay Miner. 31,203-208.

Stewart D.J. (1986) Diagenesis of the shallow marine Fulmar Formation in the Central North Sea. Clay Miner. 21, 537-564.

Stewart R.N.T., Fallick A.E. & Haszeldine R.S. (1994)

Kaolinite growth during pore-water mixing: isotopic data from Paleocene sands, North Sea, UK. Clay Miner. 29, 627-636.

Surdam R.C., Boese S.W. & Crossey L.J. (1984) The chemistry of secondary porosity. Pp. 127-149 in: Clastic Diagenesis (D.A. McDonald & R.C. Surdam, editors). A.A.P.G. Memoir 37.

Surdam R.C., Crossley L.J., Hagen E.S. & Heasler P. (1989) Organic-inorganic interaction and sandstone diagenesis. A.A.P.G. Bull. 73, 1-23.

Thomas M. (1986) Diagenetic sequences and K/Ar dating in Jurassic sandstones, central Viking Graben effects on reservoir properties. Clay Miner. 21, 695-710.

Thyberg B. (1993) En studie av avsetningsmiljo i jurassiske og terticere sedimenter, Hild og Sleipnerfeltet. Cand. Scient. thesis, Univ. Oslo, Norway.

Thyne G., Bjorlykke K. & Harrison W. (1996) Chemical reaction and solute transport rates as contraints on diagenetic models. Ann. Meet. Abstracts Am. Assoc. Petrol. SEPM 5, 140.

Tyridal D.S. (1994) Litologisk og mineralogisk sam- mensetning av slamsteiner i relasjon til loggrespons. Cand. Scient. thesis, Univ. Oslo, Norway.

Ungerer P., Burrus J., Doliger B., Doliger P.Y., Chenet P.Y. & Bessis F. (1990) Basin evaluation by integrated two-dimensional flow, hydrocarbon gen- eration and migration. Am. Assoc. Petrol. Geol. 74, 309-335.

Velde B. (1995) Origin and Mineralogy of Clays. Springer Verlag.

Walderhaug O. (1990) A fluid inclusion study of quartz cemented sandstones from off-shore mid Norway, possible evidence for continued quartz cementation during oil emplacements. J. Sed. Pet. 60, 203-210.

Walderhaug O. (1994) Temperatures of quartz cementa- tion in Jurassic sandstones from the Norwegian Continental She l f - - evidence from fluid inclusions. J. Sed. Res. 64, 224-333.

Walderhaug O. (1996) Kinetic modelling of quartz cementation and porosity loss in deeply buried sandstone reservoirs. A.A.P.G. Bull. 80, 731-745.

Warren E.A. & Smalley P.C. (1994) North Sea Formation Water Atlas. Geol. Soc. London Memoir 15.

Weaver C.E. (1989) Clays, Mud and Shales. Devel. Sedimentol. 44. Elsevier.

Wilson M.D. (1994) Reservoir quality assessments and prediction in clastic rocks. SDEPM Short Course no. 30.

Wood J.R. & Hewett T.A. (1982) Fluid convection and mass transfer in porous sandstones - a theoretical model. Geochim. Cosmochim. Acta, 46, 1707-1713.