97
-
OR
-
07
INDONESIAN PETROLEUM ASSOCIATION Proceedings of the Petroleum Systems of
SE
Asia and Australasia Conference, May 1997
HOW
OVERPRESSURE AND DIAGENESIS
INTERACT
IN
SEDIMENTARY BASINS
-
CONSEQUENCES FOR POROSITY PRESERVATION IN HPHT RESERVOIR SANDSTONES
Ma
,J.
Osbome*
Richard
E.
Swahrick
ABSTRACT
This paper critically examines the link between diagenesis <and overpressure, how diagenetic reactions may prodrice overpressure, and conversely, how overpressure may influence diagenetic reactions. These questions are important, because they may hold
a
key
LO
predicting overpressure and reservoir porosity in sedimentary basins. Our results are summarised below:
Effect of diagenetic reactions on overpressum
Smectite dehydration
is
unlikely to be a primary cause
of
overpressuring in sedimentary basins, because the volume of fluid released
is
small, and dehydration is dibited by the build up
of
pressure. In addition, the exact reaction involved In the smectite to illite trapsition
is
not presently known. Thus it is not certain that there
is
an actual volume increase during the transformation. Quartz cementation is also unlikely to be
a
direct cause of overpressuring or underpressuring, because extensive cementation/dissolution requires an open systeim in which fluid is free to move and dissipate abnormal pressures. By contrast, the transformation of gypsum to anhydrite can potentially generate a fluid pressure significantly in excess of overburden pressure, depending on the rock permeability. The existence
of
basin-wide diagenetic pressure seals remains unsupported by direct evidence.
Effect
of
overpressure on diagenetic reactions
Overpressuire inhibits pressure solution, and retards the development of late diagenetic quartz overgrowths, helping to preserve high porosities in reservoir sandstones during deep burial. The rate of quartz
University
of
Durham
growth is fastest under conditions of high effective stress.
INTRODUCTION
Considerable technical problems are encountered when drilling in HPHT (High Pressure High Temperature) areas, for example, unexpectedly high overpressure and highly variable reservoir quality. The main causes of overpressure in such areas ar~likely
to
be disequilibrium compaction and gas generation (Osborne
&
Swarbrick, 1997; Swarbrick
&
Osborne 1997). Diagenetic mineral reactions have
also
been proposed as a source of overpressure, plus growth of diagenetic minerals will act to change rock permeability, influencing the rate of overpressure build up. Overpressure may exert a control on diagenetic mineral reactions and sediment compaction, with important consequences for porosity evolution in reservoir sandstones. This paper critically examines rhe link between diagenesis and overpressure; (1) how diagenetic reactions may produce overpressure, and
(2)
conversely, how overpressure may influence diagenetic reactions. These questions are important, because we need to know both the srcin of the overpressure and how overpressure may control reservoir quality in HPHT sandstones. Answering such questions may hold a key to predicting overpressure and reservoir porosity in sedimentary basins.
EFFECT OF DIAGENESIS ON OVERPRESSURE Smectite dehydmtion
Smectite is a common detrital mineral in shales, and contains abundant interlayer water in its crystal structure. The water released during simple dehydration is thought to result in overpressuring,
947
 © IPA, 2006 - Proceedings of an International Conference on Petroleum Systems ofSE Asia and Australasia, 1997
 
948
because some of the interlayer water molecules are arranged in
a
denser packing than those of ordinary water. Thus when the interlayer water is expelled to become pore water, there is an expansion in volume, and abnormal pressures will result from the density change. Authors have proposed that the dehydration is staged in two (Powers, 1967) or three pulses (Burst, 1969) From 'thermodynamic considerations, Colton-Bradley, (1987) calculated that in highly overpressured rocks (effective stress approaching zero), smectite
ss
stable
as
two or three water layer complexes at temperatures of <200"C. Under conditions of high effective stress, one water layer is expelled at temperatures of
<6OoC,
loss of a second water layer occurs at 67-8loC, and expulsion of the last layer requires much higher temperatures of 172-1 92°C (Colton-Bradley, 1987). Hence, for typical temperatures in sedimentary basins, smectite dehydration is only likely to occur under conditions
of
high effective stress. Development of overpressure in low permeability shale would actually inhibit smectite dehydration, casting doubt
on
its validity as an overpressuring mechanism. Release of water cannot cause pore pressure increase unless there is an expansion in volume associated with the process
of
dehydration. Smectite dehydration could possibly lead to overpressure because there is an expansion of the interlayer water on expulsion, because interlayer water density is greater than that of the bulk pdrewater (Colton-Bradley, 1987). Bruce (1984) calculated that dehydration could theoretically produce an increase in the amount
of
pore water present by
up
to
6.6
weight
.
The maximum volume expansion associated with smectite dehydration can be calculated if the volume and density of the water layers are known. Assuming that the solid proportion of the rock is initially
100
smectite, it
is
possible to calculate that at each dehydration stage about
10
water is released
as
a proportion
of
the total rock volume (Osborne
&
Swarbrick, 1997). However, although
-10
volume
YQ
water is released at each dehydration stage, this will not lead to overpressure if the crystal structure of smectite compacts by
a
similar amount, and there is no overall expansion in volume. Calculations indicate that the overall increase in volume associated with simple dehydration is only a maximum
of
4.0 volume%, with the water release occuring in three pulses of -1.3 volume% (Osborne
&
Swarbrick, 1997). Decreasing the geothermal gradient will increase the depth at which dehydration occurs, but will not greatly affect the volume of fluid released. Unless the rock is perfectly sealed, such small volume expansion will not generate significant overpressure, however the transformation of solid into liquid will modify the rheology of the mudrock, and this, in turn, may generate overpressure due to disequilibrium compaction.
Smectite
to
illite transition
Release of structurally bound water from smectite could also occur during its transformation
to
illite, by the addition' of A1 and
K
ions, and the release of
Na,
Ca, Mg, Fe, Si ions plus water. The general form of the reaction is: K-feldspar
+
Smectite
===>
Illite
+
Quartz+ Water. In the Gulf Coast of the
U.S.A.,
there is a broad correlation between the onset
of
overpressuring and the smectite to illite transformation (Bruce, 1984), but this does not prove that the reaction produces the abnormal pressures. In the low geothermal gradient, highly overpressured environment of the Caspian Sea, for example, there is no change in the smectite-illite ratio down to
6.0
km (96"C, Bredehoeft,
t
al.
1988). hence the transition is not necessary for overpressuring to occur. The amount of water released is difficult to ascertain because the exact nature of the numerous reactions involved are not yet clear (Ahn
&
Peacor, 1986). The overall volume change and the amount
of
water released by the reaction varies widely depending on which hypothetical reaction is used
in
the calculation, the molar volume of smectite, and whether ions are assumed to be exported from the system or not.
If
all the reactions occur in
a
closed system (i.2. no ions are exported or imported), there is an overall volume
decrease
of -25% for the reaction of Hower
et
alls,
and a volume
increase
of -14% for Boles
&
Frank's. Ten possible smectite to illite reactions reported in Osborne
&
Swarbrick (1997), and volumetric calculations using those reactions indicate either volume increases of 0.1 to 4.1%, or volume decreases of 0.7 to
8.4%,
during the transformation. Because of the uncertainty in the exact reaction volumetrics, and our lack of knowledge about the permeability of mudrocks in the subsurface, it
is
not possible to entirely discount the smectite to illite reaction as a cause of overpressure in smectite nch lithologies. However, Audet (1995) has shown that the amounts
 
949
of overpressure from clay diagenesis remain small relative to disequilibrium compaction even under the most favourable conditions. Another potential consequence of the mineral transformation from smectite to illite is the sealing effect produced by the release of Si, Ca, Fe and Mg ions. Boles
&
Franks (1979) suggest that the ions released froim the shales can migrate into adjacent sandstones imd precipitate quartz, chlorite, ankerite and calcite lcements. This cementation at the shale- sand contact would help to retain pore waters within the shales. Freed
&
Peacor (1989), however, have argued that the coincidence of overpressure near the depth of the smectite to illite transformation results from a reduciion in the permeability of the shales, by the formation of illite packets within the smectite. The transformation of srnectite to illite clay may also be accomp,anied by changes in the physical characteristics. of the sediments. First, the collapse of the smectite clay framework and the release of bound water influences the compressibility of the sediment. If the rock is made more compressible, and the overburden induces more compaction, the newly released water must be expelled from the rock.
If
the permeability acts to retain the fluid, overpressure results, i.e. disequilibrium compaction induced by mineral dehydration.
Gypsum
to
anhydrite
transition The reaction of gypsum transforming to anhydrite is thought to be an important mechanism to generate changes in pore pressure in evaporite sections. The reaction
is
CaS0,.2H20
---->
CaSO,
+
2H,O, and will liberate 39% water by volume. The primary controls
on
the reaction are the activity
of
water in the pore- fluid, temperature and pressure (Jowett
t
a1.
1993). In hydrostatic pressure conditions, where
fluid
IS
free to escape from the rock, the reaction will occur at
40-
50 C,
the water released will be dissipated by fluid movement, and
no
overpressure will result (Jowett
et
a/.
1993). However, in impermeable conditions, pore pressures
maiy
increase, and further conversion of gypsum to anhydrite will be inhibited until slightly higher temperatures are reached (60°C; Jowett
t
al.
1993). The amount
of
overpressure will depend
on
the permeability of the surrounding rocks, Hanshaw
&
Bredehoeft 1968) has calculated that hydraulic conductivities must be as low as
lo-''
cmls for the mechanism
to
be effective. Only very tight shales and evaporites have hydraulic conductivities of this magnitude. The reaction will typically occur during shallow burial (low temperatures, 40-60°C), and is unlikely to be responsible for overpressures that occur at great depth, unless the basin has an anomalously low geothermal gradient. Diagenetic
seals
Cementation to
form
pressure seals which now mark the top
of
the overpressure zone was suggested by Hunt (1990). However many instances of diagenetic seals reported in the literature remain speculative, as no evidence from rock core or cuttings is available to confirm their existence. Many "pressure seals" described in the literature are clearly lithological rather than diagenetic in srcin, although diagenesis can clearly modify the sediment permeability (Swarbrick
&
Osborne, 1995). One instance where petrographic evidence does support the idea of diagenetic seals is discussed by Weedman
t
a/.
(1992).
The seal
is
a highly compacted sandstone (up to 60m thick) showing extensive pressure solution and fracturing of detrital grains. The seal occurs at considerable depth
(-5
600m TVD). Weedman
et
a/.
suggest that highly focussed corrosive fluids created a zone of secondary porosity which subsequently compacted to form a seal below which overpressure developed. Sandstones below the seal are presently overpressured and less compacted than sandstones above the seal: Drzewiecki et
al.
(1994) describe diagenetically banded dolomite and quartz cemented "seals" occurring in sandstones at the top of overpressured compartments in the Michigan basin,
U.S.A.
The srcin of such banded seals remains uncertain, but pressure solution is an important process in their formation (Shepherd et al. 1994). Silica budgets for these seals indicate that more silica was released by pressure solution than is present as cement, hence silica has been exported by fluid movement (Shepherd et al. 1994). Silica export implies open system conditions, and hence hydrostatic pore pressures, at the time
of
seal formation. Pressure solution will proceed most rapidly under conditions
of
high effective stress (low overpressure). Seal formation is likely to pre-date overpressure build
up
Banded diagenetic seals have only been reported from very old (Ordovician) basins, suggesting they may take considerable periods of geological time to form.
 
95
0
In summary, banded diagenetic seals are unknown in young, rapidly subsiding basins, but may occasionally occur in very old basins. Most seals described in the literature are clearly lithological in srcin, with diagenesis playing only a secondary role in permeability redultion. The limited evidence currently available suggests that diagenetic seals pre-date overpressure build up.
Quartz
cementation
s
a
cause
of
ovelpmsure
Bjorkum
&
Nadeau, (1996) have proposed that quartz cementation in reservoir sandstones encased within low permeability lithologies can be a direct cause of overpressure within the sandstone. The model requires chemical compaction at stylolites within the sandstones, resulting in a net thickness reduction,' at the same time silica released by dissolution reprecipitates in the adjacent pore space as quartz cement, also resulting in a net decrease in porosity. In a perfectly sealed and incompressible rock, this decrease in porosity could cause a massive increase in fluid pressure; precipitation of 2.0 volume% quartz cement could increase pressures by -8OOOpsi, sufficient to blow the rock apart at depths <16OOOft (4900m) The problem with this model is that surrounding seals would require great thickness and extremely low permeability to prevent fluid movement and pressure dissipation from the sandstones. Hence this mechanism is considered an unfeasible source of overpressure unless impermeable seals (e.g. evaporites) encase the sandstones.
EFFECT
OF
OVERPRESSURE
ON
DIAGENESIS Chemical compaction
Overpressuring is likely to inhibit quartz cementation where the major source of silica is from pressure solution, because when pore pressure increases effective stress decreases, reducing the stress at grain contacts. Fluid inclusion studies
of
quartz cements in North Sea oilfields (Swarbrick, 1994) supports quartz precipitation at near hydrostatic conditions. Alternatively, quartz cementation may be temperature related and increase with depth of burial, regardless of overpressure (Walderhaug,
1994).
To
test this hypothesis, petrographic, petrophysical and fluid inclusion data were compiled
from
Fulmar Formation reservoir sandstones in 6 oillcondensate fields in the
UK
North Sea with highly variable pressure states and depths of burial. In these deeply buried reservoirs small volumes (average
3.0
~01%)
f
macroquartz cement are found as overgrowths on detital quartz grains. Below 3000m, macroquartz cement abundance does not increase with present day depth of burial, suggesting that quartz cementation is not strongly temperature controlled. By contrast, there does appear to be a relationship between quartz abundance and effective stress in reservoirs buried to depths
of
4-6km. This relationship between high effective stress and high quartz overgrowth abundance can be explained
if
pressure solution at grain contacts is the source of much
of
the cement. Pressure solution could be inhibited by build-up of overpressure (low effective stress), hence explaining why there is relatively little quartz cement in
HPHT
reservoirs compared to reservoirs which are only slightly overpressure In order to test this hypothesis, the source of the quartz cement was investigated
*
using cathodoluminesence on a scanning electron microscope. The amount of quartz cement and the amount of pressure solution accurately quantified for sandstones in two wells,. one of which was highly overpressured, the other slightly overpressured. From Figure
1
it
can be seen that pressure solution can supply much of the quartz in the sandstones
from
these
two
wells. Note that more quartz cement occurs in the slightly overpressured well compared to highly overpressured well. However, three samples have more quartz cement than can be supplied by pressure solution alone. These samples are sandstones which are immediately adjacent to siltstones. Petrographic examination
of
these siltstones indicates they contain highly sutured quartz grains surrounded by clays and micas, suggesting quartz grains have undergone extensive pressure solution. Further information about the srcin of the quartz overgrowths can be obtained using fluid inclusion microthermometq. Typically, the homogenisation temperatures in each sample span a temperature range of about 40°C, and often have values approaching present day reservoir temperature, based on palaeotemperature gradients of 35-40°C/km, suggesting that the quartz must have formed during deep burial, at relatively high temperatures.. If the fluid inclusion temperature data are combined with
of 8