Environ Geol (2008) 54:1707–1722

DOI 10.1007/s00254-007-0946-9

ORIGINAL ARTICLE

Effects of in-situ conditions on relative permeability

characteristics of CO2-brine systems
Stefan Bachu Æ Brant Bennion

Received: 2 June 2006 / Accepted: 29 November 2006 / Published online: 24 July 2007

Springer-Verlag 2007

Abstract Carbon dioxide capture and geological storage Keywords CO2-brine systems
Relative permeability

(CCGS) is an emerging technology that is increasingly Interfacial tension
Capillary pressure
Pore size
being considered for reducing greenhouse gas emissions to distribution
Alberta basin
the atmosphere. Deep saline aquifers provide a very large
capacity for CO2 storage and, unlike hydrocarbon reser-
voirs and coal beds, are immediately accessible and are Introduction
found in all sedimentary basins. Proper understanding of
the displacement character of CO2-brine systems at in-situ It is generally accepted that the main cause of global
conditions is essential in ascertaining CO2 injectivity, warming is the increase in atmospheric concentrations of
migration and trapping in the pore space as a residual gas greenhouse gases, such as carbon dioxide (CO2) and
or supercritical fluid, and in assessing the suitability and methane (CH4). To meet mid- to long-term targets in
safety of prospective CO2 storage sites. Because of lack of reducing either CO2 emissions or their intensity, various
published data, the authors conducted a program of mea- mitigation approaches need to be considered, foremost
suring the relative permeability and other displacement among them being CO2 capture and geological storage
characteristics of CO2-brine systems for sandstone, car- (CCGS), which will play an important role at least in the
bonate and shale formations in central Alberta in western first half of this century if reduction targets are to be met
Canada. The tested formations are representative of the in- (IEA 2004; IPCC 2005). Carbon dioxide can be seques-
situ characteristics of deep saline aquifers in compacted on- tered in geological media through a variety of mechanisms:
shore North American sedimentary basins. The results static or hydrodynamic trapping below caprock in oil and
show that the capillary pressure, interfacial tension, relative gas reservoirs and deep saline aquifers; dissolution in res-
permeability and other displacements characteristics of ervoir oil or formation water; in-situ mineral carbonation;
CO2-brine systems depend on the in-situ conditions of and/or adsorption onto organic matter in coal and shale
pressure, temperature and water salinity, and on the pore (Gunter et al. 2004). More recently, CO2 retention as an
size distribution of the sedimentary rock. This paper pre- immobile phase trapped in the pore space of deep saline
sents a synthesis and interpretation of the results. aquifers has also been identified as an important CO2
storage mechanism (Kumar et al. 2005). Deep saline
aquifers offer the advantage that they are ubiquitous in
S. Bachu (&) sedimentary basins, have the largest storage capacity of all
Alberta Energy and Utilities Board,
geological media and can be used immediately, unlike oil
4999-98th Avenue, Edmonton, AB,
Canada T6B 2X3 and gas reservoirs that have to be produced first, or coal
e-mail: Stefan.Bachu@gov.ab.ca beds for which the technology is still in the development
stage (IPCC 2005).
B. Bennion
For temperatures T and pressures P above the critical
Hycal Energy Laboratories Ltd,
1338 A-36th Avenue NE, Calgary, point (Tc = 31.1C and Pc = 7.38 MPa), CO2 is in super-
AB, Canada T2E 6T6 critical phase, behaving like a gas but having the density of

123
1708 Environ Geol (2008) 54:1707–1722

a liquid. These conditions are found at depths that vary in situ conditions became apparent, such as the dependency
between ~600 and ~1,000 m, depending on pressure and of interfacial tension (IFT) on in-situ conditions, and of
geothermal regimes (Bachu 2003). For conditions below relative permeability and irreducible saturation on IFT. To
the critical point (i.e., at depths less than that corresponding verify these trends, an additional series of controlled tests
to the critical point), CO2 is either a gas or a liquid, were conducted that indeed confirmed the observed
depending respectively on the position in the P–T space dependencies. Building on the work presented previously,
below or above the vaporization curve (see, e.g., Bachu this paper presents the complete set of results and a syn-
2003). It is generally accepted that, in order to maximize thesis of this work, including the additional unpublished
the use of the pore space, CO2 should be injected at depths data, that will likely be of use for applications where
where the pressure and temperature are such that CO2 similar conditions prevail, particularly in consolidated on-
would be a dense fluid (i.e., in liquid or supercritical phase) shore sedimentary basins.
(IPCC 2005), although there might be instances when CO2
could be in gas phase. Consequently, depending on the
injection depth because pressure and temperature in the Geological setting and rock characteristics
subsurface increase with depth, CO2 injected into deep
saline aquifers will initially be in free gas, liquid or dense Alberta, a major North American energy producer, is the
supercritical fluid phase, and will form a plume around the province with the largest CO2 emissions in Canada
injection well that will rise to the top of the formation and (~30%). Approximately two thirds of the annual CO2
expand below the caprock. Due to diffusion-limited mass emissions in Alberta, estimated currently to be in the
transfer it will take centuries to millennia for CO2 to dis- order of 250 Mt CO2/year, originate in large stationary
solve in formation water and react with rock minerals sources (>100 kt CO2/year each), such as oil sands plants,
(Ennis-King and Patterson 2003; Xu et al. 2003), during fossil–fuel power plants, refineries, upgraders, chemical
which time a plume of injected CO2 may migrate updip, and petrochemical plants, compressor stations, cement
driven by buoyancy if the CO2 is injected in a deep aquifer plants, paper mills and gas plants. The province is
and is not trapped in stratigraphic or structural traps. The underlain by the Alberta basin which, except for its
processes of injection, migration and storage of CO2 as an shallow northeastern corner, is suitable for CO2 geologi-
immobile phase in the pore space through irreducible sat- cal storage (Bachu 2003) in oil and gas reservoirs, deep
uration in the wake of a migrating plume all depend on the saline aquifers and in coal beds if the latter technology
relative permeability of CO2 and formation water systems proves successful. This combination of large stationary
and on the character of CO2 - brine capillary pressure. sources and potential CO2 geological sinks makes CCGS
Furthermore, numerical models for predicting the fate of particularly attractive and promising. The largest con-
the injected CO2 also need information about these two centration of large CO2 sources is in central Alberta in a
important parameters. However, because CO2 storage in region spanning from the coal-fired power plants near
geological media, and particularly in deep saline aquifers, Wabamun Lake southwest of Edmonton to the petro-
is a recently emerging field, to date no relevant data have chemical plants, refineries and bitumen upgraders, and
been published regarding the relative permeability and chemical industry in Fort Saskatchewan northeast of Ed-
capillary pressure of CO2-brine systems at in-situ condi- monton (Fig. 1). The Pembina oil field, the largest oil
tions, and only limited sets of data were published field in Canada and one of the largest in North America,
regarding the interfacial tension (IFT) between CO2 and with original oil in place of 1,237 million m3 (7.78 billion
water or brine, which will be reviewed subsequently. To barrels), is also located southwest of Edmonton and
date, interest has been focused on CO2-oil systems because Wabamun Lake (Fig. 1). This oil field has undergone
of the potential use of CO2 for enhanced oil recovery. extensive water flooding (secondary recovery) and cur-
To cover the gap in data, a program of laboratory rently the oil industry is planning to build a pipeline that
measurements of relative permeability and displacement will bring CO2 from sources in the Fort Saskatchewan
characteristics of CO2-brine systems was commenced by area in the northeast for use in CO2 enhanced oil recovery
the authors in 2004, with partial results being reported at operations (tertiary recovery). However, this giant oil field
technical conferences (Bennion and Bachu 2005, 2006a, b, and other oil and gas fields in the region, as well as local
c). The laboratory measurements were carried out at in-situ uneconomic coal seams, do not have sufficient capacity
conditions on core samples from deep sandstone and car- and/or cannot accommodate the potential annual rate of
bonate aquifers in the Wabamun Lake area southwest of captured CO2 and/or are not yet available for CO2 stor-
Edmonton in Alberta, western Canada. As results were age, such that CO2 storage in deep saline aquifers that
obtained and analyzed, the dependence of certain relative underlie the whole region is likely to occur either before
permeability and other displacement characteristics on or simultaneously with storage in hydrocarbon reservoirs.

123
Environ Geol (2008) 54:1707–1722 1709

Fig. 1 Location of large

stationary CO2 sources in
central Alberta, Canada, and of
wells in the Wabamun Lake
area from which core samples
were used in experiments for
determining the displacement
characteristics of CO2-brine
systems. The location of the
well reaching the Precambrian
basement, whose downhole
model is presented in Fig. 2, is
also shown. Names by wells
indicate the geological unit from
which the rock was sampled and
tested

The Alberta basin, which, together with the Canadian the basement, with low heat flow and geothermal gradients
part of the Williston basin, is known as the Western Can- in the south, and higher heat flow and geothermal gradients
ada Sedimentary Basin, is underlain by a stable crystalline in the north (Bachu 1993).
Precambrian platform and comprises a platform-margin In the Wabamun Lake area southwest of Edmonton, the
sedimentary succession that is dominated by carbonates Upper Cretaceous strata at the top of the sedimentary
and evaporites with intervening shales, and an overlying succession (Fig. 2) are slightly sub-hydrostatic, the Car-
foreland succession that consists of regional-scale thin dium Formation, which is host of the giant Pembina Car-
sandstones and thick shales (Porter et al. 1982). The basin dium oil field, is overpressured, the underlying Lower
dips from a zero-edge in the northeast to the southwest, Cretaceous Viking and Mannville strata are severely un-
reaching more than 6 km depth at the thrust and fold belt of derpressured, and the underlying Devonian and Cambrian
the Rocky Mountains. The Paleozoic strata of the Alberta aquifers are underpressured. Geothermal gradients are less
basin are generally underpressured, being slightly sub- than 25C/km, such that temperatures at the base of the
hydrostatic. The Cretaceous strata are overpressured in succession are <80C. Oil and gas is produced in the area
places, due to active hydrocarbon generation, but also se- from reservoirs in the Cretaceous Belly River, Cardium,
verely underpressured (sub-hydrostatic) over large areas Viking, Upper Mannville and Ellerslie strata (Fig. 2), with
due to post-erosional and post-glacial rebound in thick no production from deeper Paleozoic strata. The Upper
shales, and normally pressured in shallow strata, particu- Cretaceous strata of the Edmonton and Belly River groups
larly along the edge of the basin (Bachu 1999). The salinity are not suitable for CO2 storage because they are too
of formation waters in the basin increases with depth from shallow and no regionally-extensive shale barriers are
freshwater in shallow groundwater systems to more than present to prevent upward migration of injected CO2. The
350,000 mg/l in deep Paleozoic strata, particularly in strata water-flooded Pembina oil field, still in production, is
adjacent to extensive halite beds. The geothermal regime in currently being considered for CO2-EOR; hence CO2 will
the basin is controlled mainly by conduction of heat from be stored in the Cardium Formation as a result. Carbon

123
1710 Environ Geol (2008) 54:1707–1722

Fig. 2 Stratigraphic and

hydrostratigraphic downhole
model for strata in the
Wabamun Lake area Alberta,
Canada (well 00/10–05–052–
02W5). The location of this well
that reaches the basement is
shown in Fig. 1. The side
arrows indicate the geological
units from which core samples
have been analyzed and tested

123
Environ Geol (2008) 54:1707–1722 1711

dioxide storage is possible in all the underlying Cambrian Colorado Gp. Shale and Cardium Fm. sandstone, both from
to Lower Cretaceous strata since all meet the conditions of the same well and very close to each other, were taken in a
pressure, temperature and confinement required for CO2 well-located downdip of all other wells such that the
storage. sample depth is greater than that of most other samples. In
Rock samples were collected for the sandstone Cardium, addition, unlike all the other formations, the Cardium Fm.
Viking, Ellerslie and Basal Cambrian formations, for the is overpressured, such that the in-situ pressure and
carbonate Wabamun Group and Nisku and Cooking Lake temperature for these samples are greater than the corre-
formations, and for shales from the Colorado Group and sponding values for most other samples from stratigraphi-
Calmar Formation. The core samples (3.81 cm OD—outer cally underlying strata. Water salinity in the basin, as
diameter—plug samples, approx. 5–6 cm in length) were measured also in analyses of formation water in nearby
obtained from core stored at the Core Research Centre wells, increases generally with stratigraphic depth rather
facility of the Alberta Energy and Utilities Board in Cal- than true depth (Table 1). Additional core samples from
gary. The location of the wells with core from which rock different wells were collected for the Viking and Nisku
samples were taken is shown in Fig. 1, and the strati- formations and the Wabamun Group, and a second sample
graphic position of the samples is shown in Fig. 2. Each was collected from the same well for the Cardium Fm., and
sample was subjected to standard laboratory measure- were tested for capillary pressure and relative permeability
ments, described in the next section, for pore size distri- at the same in-situ conditions of pressure, temperature and
bution, porosity and permeability, and subsequently tested water salinity as the corresponding first set, for studying the
for capillary pressure, relative permeability and displace- effect of variability in rock properties on the displacement
ment characteristics of CO2-brine systems at the respective characteristics of CO2 and brine. The characteristics of all
in situ conditions of pressure, temperature and salinity of 13-rock samples are provided in Table 2.
formation water. The in-situ pressure, temperature and The 13 rock samples exhibit various pore size distribu-
water salinity conditions are given in Table 1, where the tions, of which the most representative are: uni-modal
samples are listed stratigraphically in descending order. ‘‘spike’’ for the very homogeneous siliciclastics (Fig. 3a,
Note that for all the core samples the in-situ pressure and b), wide uni-modal and ‘‘jagged’’ bi-modal for carbonates
temperature correspond to CO2 supercritical conditions (Fig. 3c) and tri-modal for heterogeneous siliciclastics
(P > 7.38 MPa and T > 31.1C). Pressure and water (Fig. 3d). Generally, the carbonate rocks display more
salinity were obtained, respectively, through interpolation complex pore-size distributions than the siliciclastic rocks.
of pressures measured in drillstem tests and of formation The pore distribution of each rock sample is given in Ta-
water analyses taken in nearby wells. Temperatures were ble 2 in terms of fraction of micropores (<1 lm diameter
calculated on the basis of geothermal gradients and depth. pore throats), mesopores (1–3 lm diameter pore throats)
The samples from the Colorado Gp. and Cardium Fm. are and macropores (>3 lm pore throat diameters), as well as
only 8 m apart in the same well, and they were considered the median (average) pore throat diameter and threshold
as having the same in-situ conditions. Note that the rock intrusion pressure representing the pressure required to first
sample from the Cooking Lake Fm., although stratigraph- push mercury into the largest pores present in the sample.
ically below the Nisku Fm., is at a shallower depth, hence Two samples record a threshold intrusion pressure of
at lower temperature and pressure than the sample from the 3.5 kPa (Table 2), which is the minimum initial pressure
latter (Table 1), because it is from a relatively shallower level for mercury intrusion of the apparatus. Samples that
updip well (Fig. 1). Conversely, the samples from the exhibit this low threshold pressure typically contain large

Table 1 In-situ conditions for

Unit Lithology Depth Pressure Temperature Salinity IFT
rock samples from the
(m) (MPa) (C) (mg/l) (mN/m)
Wabamun Lake area, Alberta,
Canada, used in the analysis of Colorado Gp. Shale 1,618 20.00 43 27,100 19.8
relative permeability and
displacement characteristics of Cardium Fm. #1 Sandstone 1,627 20.00 43 27,100 19.8
CO2-brine systems Viking Fm. #1 Sandstone 1,240 8.60 35 28,300 32.1
Ellerslie Fm. Sandstone 1,463 10.90 40 97,200 32.5
Wabamun Gp. #1 Carbonate 1,603 11.90 41 144,300 29.5
Samples for which a second set
Calmar Fm. Shale 1,566 12.25 43 129,700 27.6
was subsequently collected are
identified by a numeral. Core Nisku Fm. #1 Carbonate 2,050 17.40 56 136,800 34.6
depth is rounded to the nearest Cooking Lake Fm. Carbonate 1,889 15.50 55 233,400 35.7
meter, and water salinity is Basal Cambrian Fm. Sandstone 2,734 27.00 75 248,000 27.0
rounded to the nearest 100 mg/l

123
1712 Environ Geol (2008) 54:1707–1722

Table 2 Pore and permeability characteristics of rock samples from the Wabamun Lake area, Alberta, Canada, used in the analysis of relative
permeability and displacement characteristics of CO2-brine systems
Unit Micro Meso Macro Median pore Porosity Threshold capillary Permeability
porosity (%) porosity (%) porosity (%) size (lm) (%) pressure (kPa) to brine (mD)

Colorado 95.5 1.9 2.6 0.011 4.4 172.0 0.0000788

Cardium #1 67.0 32.1 0.9 0.690 15.3 31.8 0.356
Cardium #2 30.1 9.0 60.9 4.887 16.1 44.3 21.17
Viking #1 33.8 13.5 52.7 3.760 12.5 28.8 2.700
Viking #2 27.6 7.6 64.8 10.763 19.8 3.5 21.72
Ellerslie 59.6 20.9 19.5 0.277 12.6 56.3 0.376
Wabamun #1 81.0 18.9 0.1 0.645 7.9 493.6 0.018
Wabamun #2 11.1 11.8 77.1 8.760 14.8 57.2 66.980
Calmar 100. 0 0 0.006 3.9 72,827.0 0.00000294
Nisku #1 54.0 3.1 42.9 0.066 9.7 58.0 45.920
Nisku #2 33.2 18.8 48.0 4.327 10.4 3.5 21.020
Cooking Lake 11.9 5.1 83.0 15.860 9.9 11.1 65.300
Basal Cambrian 65.1 29.5 5.4 0.919 11.7 31.9 0.081
Multiple samples from the same unit are differentiated by a roman numeral. Micro pores are pore connected by pore throats of less the 1 lm in
diameter, meso pores by throats of 1–3 lm in diameter and macro pores by pore throats of over 3 lm in diameter

exposed surface dissolution pores and/or exposed vugular Fig. 3c). If this data point is discarded, the relationship
porosity. between absolute permeability and pore size is significantly
No correlation was found between porosity and pore improved (see dashed line in Fig. 4). A similar correlation
size characteristics (fraction of micro-, meso- and to that shown in Fig. 4 was found between permeability
macroporosity, or median pore size), but a reasonable link and the fraction of large pores, but basically the two cor-
was obtained between the latter and permeability to brine relations are an expression of the same dependency and
(see solid line in Fig. 4). This is intuitively an expected flow process.
conclusion, because porosity is a measure of the volume
(amount) of pores, not of their size and distribution, while
permeability expresses the ability of fluids to pass through Laboratory procedures
the pore system, hence it is related to pore characteristics.
It is worth noting that the data point in the upper left corner Knowledge of the interfacial tension (IFT) between CO2
of Fig. 4 corresponds to the Nisku #1 core sample that and brine at in-situ conditions is needed for precise
has an unusual bimodal, ‘‘seesaw’’ pore distribution (see measurements of capillary pressure, which in turn

Fig. 3 Examples of various

pore size distributions for
different rock samples from the
Wabamun Lake area, Alberta,
Canada: a uni-modal
homogeneous shale (Colorado
Group), b ‘‘widespread’’ and
‘‘spike’’ uni-modal
homogeneous sandstones
(Cardium and Viking
formations), c ‘‘widespread’’
uni-modal and bi-modal, ‘‘on-
off’’ carbonate (Nisku Fm.), and
d heterogeneous tri-modal
sandstone (Ellerslie Fm)

123
Environ Geol (2008) 54:1707–1722 1713

the IFT and contact angle of mercury (485 dyn/cm and

approximately 130, respectively), the specific pore throat
size distribution of the sample being tested and the porosity
fraction of the sample that is accessed by these specific
pore throats were derived. The mercury injection data were
converted to appropriate reservoir conditions for the CO2-
brine systems being evaluated using the measured IFT data
and assuming strongly water wet reservoir behaviour (a
valid assumption as all the systems evaluated were clean
Fig. 4 Relationship between permeability to brine and median pore sandstones or carbonates containing no hydrocarbon liquid
size for siliciclastic and carbonate rock samples from the Wabamun saturation) at the appropriate reservoir conditions. For
Lake area, Alberta, Canada. The solid line corresponds to the more details see Bennion and Bachu (2006a).
relationship obtained by including all the data points, the dashed line
corresponds to the relationship obtained if the Nisku #1 core sample
Coreflood experiments were used to measure the relative
(upper left corner) is excluded permeability for CO2 displacing water (drainage) and water
displacing CO2 (imbibition) in core samples. Each core
sample was placed in a 3.81 cm ID (internal diameter)
impacts relative permeability. To obtain a stable IFT the flexible confining sleeve. Lead was used as a sleeve
two fluids must be in thermodynamic equilibrium. For material due to the propensity of CO2 to diffuse and pen-
this reason the brine and CO2 were precontacted with etrate through most flexible conventional core sleeves. The
each other at in-situ temperature and pressure conditions, ductility of the sleeve allows the confining external over-
and shaken until an equilibrium condition of saturation burden pressure to be transferred to the core in a radial and
had been established between the two phases. The phases axial mode to simulate in-situ overburden pressure. The
were then separated maintaining constant temperature core, mounted within the sleeve, was placed inside a
and pressure, and charged into a high pressure and 7.5 cm ID steel core holder that simulated reservoir con-
temperature drop-pendant IFT cell. A droplet of equi- fining pressures of up to 68.9 MPa. The core-holder ends
librium brine, surrounded by the equilibrium gas, is each contain two ports to facilitate fluid displacement and
suspended at bottomhole temperature and pressure con- pressure measurements at each end of the core. The por-
ditions from the tip of the small bore capillary tube tions of the core holder directly adjacent to the inflow and
which extends into the optically visible portion of the outflow ends of the test cores are equipped with radial
sight cell, and a digital high magnification image of the distribution plates to ensure that fluid flow is uniformly
configuration of this drop is obtained. When the drop distributed into and out of the core sample. Fluid was
matures, the profile is digitally captured and extracted in displaced through the core samples using pulsation-free
a file. An optimization software package is used to match positive displacement pumps that can operate at constant
the surface curvature of the drop with the solution of the rate or pressure. All injection fluids were pressure-injected
Laplace equation for the solution of the IFT. More de- through a 0.5 sintered stainless steel filter element prior to
tailed descriptions of the experimental method and of the entry into the core sample to avoid the entrainment of any
experiments are given elsewhere (del Rio and Newman potentially plugging particulate materials. The pressure
1997; Bennion and Bachu 2006a). difference across the sample was monitored using quartz
Capillary pressure measurements were conducted using strain gauge pressure transducers mounted directly across
a computerized 400 MPa Micromuretics air–mercury cap- the core. The core holder and associated injection fluids
illary pressure apparatus. This technique has been well were contained in a temperature controlled air bath to
established in the industry and is discussed in numerous simulate reservoir temperature. More details about the
technical papers (e.g., Pickell and Swanson 1966). Sample experimental procedure are found in Bennion and Bachu
rock chips of approx 2.5 cm diameter were taken from the (2005).
endcap sections of the plugs that were tested for relative
permeability, to provide samples as representative as pos-
sible of the actual porous medium being tested. Liquid Interfacial tension
mercury was used to represent the non wetting phase, while
the mercury vapor phase represented the wetting phase. The surface tension of liquids displays a dependence on
Highly accurate capacitance measurements were used to pressure and temperature, as shown theoretically by
evaluate the exact volume of mercury injected into the Navascués (1979). Furthermore, the surface tension of
sample as a function of applied pressure, to generate aqueous electrolytes increases with increasing salinity as a
the mercury injection capillary pressure curve. Knowing result of hydration of ions (Levin and Flores-Mena 2001).

123
1714 Environ Geol (2008) 54:1707–1722

Thus, a dependence of the interfacial tension (IFT) of CO2-

brine systems on pressure, temperature and water salinity
should be expected. As mentioned previously, there are
reported studies of interfacial tension for CO2-oil systems
(e.g., Rosman 1977; Zuo and Stenby 1998) where
decreasing trends in IFT as a function of increasing pres-
sure and oil API gravity were observed) and also for
hydrocarbon-brine systems. For the latter, Cai et al. (1996) Fig. 5 Variation of interfacial tension (IFT) between CO2 and brine
report a weak dependence of IFT on pressure (actually an for the in-situ conditions of rock samples from the Wabamun Lake
increase of IFT with increasing pressure), but a significant area, Alberta, Canada (see Table 1 for values)
dependence of IFT on temperature and salinity (IFT in-
creases with increasing temperature and/or NaCl concen- constant for a large range of molal concentrations and that
tration). Data have been published recently also for CO2- depends on temperature before the constant IFT value is
water systems (e.g., Hebach et al. 2002), but only very reached.
recently the possibility of using CO2 storage in oil reser- The IFT measurements reported here for the specific in-
voirs and deep saline aquifers as a climate-change miti- situ conditions of the rock samples from the Wabamun
gation strategy raised the interest in CO2-brine systems. In Lake area are given in Table 1. They display a broad
a study of CO2-crude oil systems and CO2-brine systems, dependence on pressure (Fig. 5). The trend is defined
Yang et al. (2005) report for both cases a decrease in IFT as mostly at low pressures, while the middle region shows the
pressure increases and as temperature decreases. Their absence of a clear trend, suggesting the plateau identified
study covers the pressure range of 0.1–31.4 MPa for tem- by Chalbaud et al. (2006). However, a detailed analysis of
peratures of 27C and 58C. A relatively low-salinity oil- the data revealed that the scatter reflects the variations in
field brine (4,270 mg/l total dissolved solids) was used in temperature and formation water salinity of the in-situ
the experiments. conditions (Table 1; Bennion and Bachu 2006a).
While the current paper was in review, Chalbaud et al. To verify the detected dependence of IFT on tempera-
(2006) presented IFT measurements for CO2-brine systems ture and water salinity, two series of controlled tests were
in the pressure range of 4.5–25.5 MPa, for temperatures of devised and run. In the first series, water salinity was
27C, 71C and 100C that correspond to liquid and maintained constant at 144,304 mg/l, characteristic of the
supercritical CO2, and NaCl concentrations between 0.085 Wabamun Gp. water (Table 1), and pressure was varied
and 2.56 molal. They conclude that the IFT for CO2-brine between 2,000 and 27,000 kPa for the following tempera-
systems decreases with increasing pressure, and increases tures: 41C (the in-situ temperature of the Wabamun Gp.
with increasing temperature and salinity, but that above a sample), 60C, 75C, 100C and 125C. For the given
certain pressure IFT reaches a plateau that is independent salinity, the results show indeed a strong dependence of
of temperature and pressure. For pressures below the pla- IFT on temperature (Fig. 6a). For a constant temperature,
teau limit, Chalbaud et al. (2006) conclude that the increase the IFT decreases exponentially with increasing pressure,
in IFT with increasing salinity is linear, with a slope that is the decrease being more pronounced for pressures less than

Fig. 6 Variation with pressure

of interfacial tension (IFT)
between CO2 and brine for:
a different temperatures and
constant water salinity of
144,304 mg/l; and b different
salinity at constant temperature
of 41C. Note that the curves
are not mathematical fits, but
are manually drawn to evidence
the patterns

123
Environ Geol (2008) 54:1707–1722 1715

10 MPa and becoming asymptotic towards a constant value between gaseous CO2 and brine is higher than that between
for pressures greater than 15 MPa (the ‘‘plateau’’ suggested supercritical CO2 and brine (Chun and Wilkinson 1995;
by Chalbaud et al. 2006). This dependency can be ex- Hebach et al. 2002). The derivation of a general functional
pressed functionally by IFT = A
PB, where P is pressure. form to express the dependence of IFT of CO2-brine sys-
As temperature increases, IFT increases, this increase tems on pressure, temperature and pressure requires a
being less pronounced for low pressures (less than 5 MPa systematic program of laboratory tests that is currently
in this case), but significant for higher pressures (Fig. 6a). being conducted by the authors and whose results will be
Fitting an exponential function through all the data pro- reported in a future publication.
duced the following expression for IFT for CO2 and brine
of 144,304 mg/l salinity:
Capillary pressure
IFT ¼ ð0:19255
T 2  54
7246
T þ 4; 287:29Þ

Pð0:0026
T0:5712428Þ ; R2 ¼ 0:09658: ð1Þ Capillary pressure measurements using a primary drain-
age method from a condition of 100% water saturation
where IFT (mN/m) is interfacial tension, P (kPa) is pres- were conducted for all 13 rock samples. In the case of the
sure and T (C) is temperature. Cardium #1 sample, the availability of IFT data at dif-
In the second series of tests, temperature was main- ferent pressures allowed for the conversion of the data to
tained constant at 41C (the in-situ temperature of the reservoir conditions that corresponded to different IFT
Wabamun Gp. sample) and water salinity was varied as levels. Most of the results have been reported previously
follows: freshwater (zero salinity), 75,780, 225,462 and (Bennion and Bachu 2006a, c), with the examples shown
282,768 mg/l, the last three values corresponding to here being selected to illustrate the dependence of capil-
analyses of formation water from the same Wabamun lary pressure on rock properties, which is expected, but
Gp. at other locations in the Alberta basin. The results also on the in-situ conditions that affect IFT. Figure 7
show a strong dependency of IFT on salinity (Fig. 6b) of presents the capillary pressure for the same core samples
the same exponential type as the dependence on pressure as those whose pore size distribution is shown in Fig. 3.
and temperature at constant salinity. For constant pres- Figure 7a and b show the increase in capillary pressure
sure and temperature, IFT increases as salinity increases. with the decrease in pore size, illustrated by both median
In a similar manner with the previous series of tests, the pore size and ‘‘spike’’ shape of the pore size distribution
IFT dependence on pressure and salinity at constant (Table 2, Fig. 3a and b). Bi- or tri-modal pore size dis-
temperature of 41C can be expressed by the following tributions (Fig. 3c, d) are reflected in the shape of the
function: capillary pressure curve (Fig. 7c, d) that looks, respec-
tively, like the superposition of two or three S-shaped
IFT ¼ ð0:001328
S þ 608:025Þ
P0:338
S ; R2 ¼ 0:9437: curves characteristic of uni-modal pore size distributions.
ð2Þ The ‘‘on-off’’ pore size distribution observed between the
peaks of the bi-modal distribution of the Nisku #1 sample
where S (mg/l) is salinity. (Fig. 3c) manifests itself in the jagged, step-wise capillary
These results show clearly a strong dependence of the pressure observed in the range of 0.5–0.65 water satura-
IFT for CO2-water systems on pressure, temperature and tion fraction for the same sample (Fig. 7c). The capillary
water salinity. Interfacial tension decreases with increasing pressure curves do not lend themselves to Leverett scaling
pressure with an asymptotic trend for high pressures, and (normalization by porosity) because, as the results and
increases with increasing temperature and salinity, in a illustrative examples in Fig. 7 indicate, their shape
functional form of the type: strongly depends on pore size distribution and on rock
type, and there is a wide range of variation of facies types
IFT ¼ AðT; SÞ
PBðT;SÞ ð3Þ that were evaluated in the test program of radically dif-
ferent pore geometry and type.
However, although for constant temperature and water More important, it was found that IFT affects capillary
salinity it seems that IFT tends toward a ‘‘plateau’’ value, pressure to an extent comparable to that of pore size
definitely this value depends also on temperature, not only characteristics (Bennion and Bachu 2006c). Figure 8
on salinity as concluded by Chalbaud et al. (2006). shows the effect of decrease in capillary pressure with
The effects of pressure, temperature and salinity on IFT decreasing IFT, measured on the Cardium #1 core sample.
are likely due to CO2 solubility and phase effects, i.e., CO2 The measurements were conducted on the same core plug
solubility increases with increasing pressure and decreases to eliminate possible effects of different pore distributions
with increasing temperature and salinity, and the IFT and other rock characteristics. Given the IFT dependence

123
1716 Environ Geol (2008) 54:1707–1722

Fig. 7 Dependence of capillary

pressure on rock characteristics:
a extremely high capillary
pressure for Colorado Gp. shale
(95% microporosity), b
decrease in capillary pressure
with increasing pore size for
Viking #1 and Viking #2
sandstone samples; c effect of
bi-modal pore distribution and
of ‘‘on-off’’ pore sizes: Nisku #1
and Nisku #2 carbonate
samples, and d effect of tri-
modal heterogeneous pore
distribution: Ellerslie Fm.
sandstone

on pressure, temperature and salinity as shown previously, usually tested in the oil industry (water–oil and oil–water),
the variation in IFT was obtained by maintaining temper- but few data are available for water–gas (CO2) systems.
ature and salinity constant at the specific in-situ conditions The results show that capillary pressure, hence displace-
(Table 1) and varying only the pressure to obtain different ment characteristics and seal integrity, are affected not only
interfacial tensions, thus allowing maintaining constant the by the pore structure and characteristics of the injection
majority of the parameters used in the test, such as tem- aquifer and/or overlying aquitard, but also by the in-situ
perature, flow rate and fluid composition. The effect of IFT pressure, temperature and water salinity.
on capillary pressure and subsequently on relative perme-
ability has been observed previously for other fluid pairs
Relative permeability

Initially core samples were collected for the most repre-

sentative sandstones in the Wabamun Lake area: Viking
(#1), Ellerslie and Basal Cambrian, and carbonates:
Wabamun (#1), Nisku and Cooking Lake. Because the
Wabamun sample proved to have uncharacteristically low
permeability, a second sample (Wabamun #2) was col-
lected. These samples were tested and relative permeability
measurements were conducted on these samples only for
the drainage cycle that corresponds to CO2 injection. The
Wabamun #2 sample was tested also for freshwater rather
than brine, to check the effect of water salinity, and also for
low pressure and temperature (but original water salinity)
to test the effect of these two parameters on displacement
characteristics. Detailed results for the drainage cycle were
presented previously in both tabular and graphic form
(Bennion and Bachu 2005). This paper also detailed the
experimental equipment and procedures used in the test
programs. The tests were conducted in a horizontal dis-
placement mode, to simulate the primary direction of flow
that would occur in most of the reservoir applications. The
small diameter (3.81 cm) of the samples and relatively
high capillary pressure values present in the relatively low
Fig. 8 Dependence of capillary pressure on the interfacial tension
permeability media tested are thought to have minimized
(IFT) of the CO2-brine system, as measured for the Cardium #1 the potential for any significant gravity override effects in
sandstone sample the tested samples. To study the relative permeability and

123
Environ Geol (2008) 54:1707–1722 1717

displacement characteristics during the imbibition cycle (Calmar Fm. and Colorado Gp., respectively), and these
that corresponds to the post CO2-injection stage when the additional samples were tested for both drainage and
plume of CO2 migrates and water/brine invades back imbibition, with results being very recent and not reported
the porous medium, additional samples were collected for previously. Because previous results were reported in a
the most representative sandstone and carbonate (Viking piece-meal fashion as the research program advanced
#2 and Nisku #2, respectively) and tested for both drainage (Bennion and Bachu 2005, 2006a, b, c), no interpretation
and imbibition CO2-brine displacement, with detailed re- and synthesis of the finally emerging multi-facetted pattern
sults being also reported (Bennion and Bachu 2006b). To were presented, which is the purpose of this paper. Table 3
study the effect of IFT on relative permeability, a sand- presents a summary of the relevant displacement charac-
stone sample characterized by a wide uni-modal distribu- teristics during drainage for all 13 core samples and for all
tion was collected (Cardium #1) and tested for both 17 different test conditions for the drainage cycle of CO2-
drainage and imbibition for three different IFT conditions, brine systems. Table 4 presents the results of displacement
with the detailed results also reported previously (Bennion characteristics during the imbibition cycle for the six rock
and Bachu 2006c). The Cardium sample has also unchar- samples and eight test conditions for which these tests were
acteristically low permeability, but was chosen based on its run. Samples are listed in stratigraphically descending or-
pore distribution (Fig. 3b) because it was thought that a der in both tables.
porous medium of this type is a better candidate for illus- As a general observation, it seems that, on average, the
trating the effect of interfacial tension on the displacement permeability to CO2 at irreducible water saturation at the
characteristics of two-phase systems. The slight variation original in-situ conditions is approximately one seventh of
in initial permeability to brine at 100% water saturation for that measured for brine at conditions of 100% brine satu-
the three tests conducted on the low-permeability Cardium ration (solid line in Fig. 9), at least for rocks with perme-
#1 sample (Table 3) are within the realm of experimental ability greater than 1 mD. It should be noted that in all
accuracy and it can be concluded that the sample character these cases CO2 is in supercritical state. The case of very
remains unchanged throughout the three different IFT level low relative permeability to CO2 at irreducible water sat-
tests that were conducted on this sample. A second, more uration corresponds to the heterogeneous Cooking Lake
representative Cardium Fm. sample (Cardium #2) was Fm. carbonate sample which contains a significant amount
selected, together with Devonian and Cretaceous shales of large vugular porosity that likely has contributed to

Table 3 Relative permeability and displacement characteristics for the drainage cycle in CO2-brine systems for the rock samples from the
Wabamun Lake area, Alberta, Canada, tested for various conditions of pressure, temperature and water salinity
Rock sample CO2 State IFT Viscosity Kbrine at 100% Kr CO2 at irreducible Sb-irr
(mN/m) ratio lb =lCO2 saturation (mD) brine saturation

Colorado S 19.8 8.68 0.0000788 0.0148 0.605

Cardium #1 in-situ S 19.8 8.68 0.356 0.2980 0.379
Cardium #1 at P = 6,890 kPa S 33.5 18.57 0.373 0.4560 0.271
Cardium #1 at P = 1,378 kPa G 56.2 46.36 0.359 0.5260 0.197
Cardium #2 S 19.8 8.68 21.170 0.1290 0.425
Viking #1 S 32.12 13.98 2.700 0.3319 0.558
Viking #2 S 32.12 13.98 21.720 0.2638 0.423
Ellerslie S 32.45 14.52 0.376 0.1156 0.659
Wabamun #1 S 29.53 15.41 0.018 0.5289 0.595
Wabamun #2 S 29,53 15.41 66.980 0.1883 0.569
Wabamun #2 at low T&P G 49.24 75.00 61.160 0.0860 0.682
Wabamun #2 at S = 0 S 28.89 11.32 60.640 0.1993 0.552
Calmar S 27.60 19.14 0.00000294 0.1875 0.638
Nisku #1 S 34.56 11.80 45.920 0.1768 0.330
Nisku #2 S 34.56 11.80 21.020 0.0999 0.492
Cooking Lake S 35.74 15.43 65.300 0.0685 0.476
Basal Cambrian S 27.01 11.82 0.081 0.5446 0.294
The in-situ conditions are given in Table 1, and rock characteristics are given in Table 2. The CO2 phase (S supercritical or G gaseous) is
indicated in the second column

123
1718 Environ Geol (2008) 54:1707–1722

Table 4 Relative permeability and displacement characteristics for the imbibition cycle in CO2-brine systems for selected rock samples from the
Wabamun Lake area, Alberta, Canada, tested for various conditions of pressure, temperature and water salinity
Rock sample CO2 state IFT Viscosity Kbrine at 100% Kr brine at irreducible SCO2 irr
(mN/m) Ratio lb =lCO2 saturation (mD) gas saturation (trapped gas)

Colorado S 19.8 8.68 0.0000788 0.0024 0.349

Cardium #1 in-situ S 19.8 8.68 0.356 0.905 0.102
Cardium #1 at P = 6,890 kPa S 33.5 18.57 0.373 0.860 0.107
Cardium #1 at P = 1,378 kPa G 56.2 46.36 0.359 0.404 0.224
Cardium #2 S 19.8 8.68 21.170 0.267 0.253
Viking #2 S 32.12 13.98 21.720 0.365 0.297
Calmar S 27.60 19.14 0.00000294 0.282 0.256
Nisku #2 S 34.56 11.80 21.020 0.550 0.218
The in-situ conditions, rock characteristics and drainage cycle characteristics are given in Tables 1, 2 and 3, respectively. The CO2 phase (S
supercritical or G gaseous) is indicated in the second column

macroscopic channeling effects. This probably led to presence of some small dissolution-enhanced porosity, a
bypassing of a portion of the pore system, resulting in more small fraction of macro and meso porosity is present in the
extensive multiphase flow interference and a lower result- Colorado shale, which is not observed in the Calmar shale
ing endpoint relative permeability to CO2. This appears to (Table 2). This results in a much lower initial threshold
be a singular characteristic of the particular geometry of entry pressure for the Colorado shale than for the Calmar
the pore system in this sample and is not apparent in the shale. Initial brine permeability was slightly higher for the
other more uniform and intergranular-dominated pore Colorado shale than for the Calmar shale, likely due to the
systems tested in the other nine samples. If this sample is presence of this small fraction of larger pores. Upon con-
removed, then the permeability to supercritical CO2 at tact with CO2, however, it can be observed that the Colo-
irreducible water saturation is about one fifth of that rado shale exhibited much more adverse relative
measured for brine (dashed line in Fig. 9). permeability character, with much lower permeability to
No other relationship was found between rock (pore) both CO2 and imbibing water being observed in contrast to
and displacement characteristics (i.e., between pore size the Calmar shale. The Colorado shale IFT was lower than
and/or distribution on one side, and irreducible saturations that of the Calmar shale (which should have in theory
for both drainage and imbibition cycles, and trapped gas, improved the relative permeability character), so the data
on the other side). The irreducible water saturation on the suggest that a significant variation in pore scale geometry
drainage cycle varies between ~0.2 and 0.68 (Table 3), or surface spreading (wetting) character must be present,
with an average of 0.47, which is surprisingly high. It is resulting in increased multiphase relative permeability
interesting to note that both the lowest and highest irre- interference effects. The distribution of micro porosity (if
ducible water saturations occur for cases of CO2 being in the macro and meso pores are excluded from the Colorado
gaseous phase (Cardium #1 at 1,378 kPa and Wabamun #2 shale) are similar between the two samples, which suggests
at low T&P, Table 3). that specific pore geometry, surface roughness and/or pore
The Colorado shale sample exhibits significantly dif- geometry shape may play a role in the observed differences
ferent permeability, relative permeability and capillary between the tests.
pressure character than the Calmar shale. Due to the The displacement of formation water by the injected
CO2 (drainage cycle), and of CO2 by water in the wake of a
migrating CO2 plume (imbibition cycle) depends, intui-
tively, on the interfacial tension between the two fluids. For
lower IFT it should be easier for the non-wetting phase to
pass through pore throats, leading to higher relative per-
meability. This conclusion was suggested by relative per-
meability measurements on the same core samples (Viking
#2 and Nisku #2) for both CO2 and H2S (Bennion and
Bachu 2006b) which show that, for the same in-situ con-
Fig. 9 Relation between maximum permeability to CO2 and perme-
ditions of pressure, temperature and water salinity, H2S has
ability to brine at 100% saturation for various rock samples at in-situ
conditions from the Wabamun Lake area, Alberta, Canada (solid line lower IFT and higher relative permeability than CO2. To
all 13 rock samples; dashed line Cooking Lake Fm. sample removed) verify this hypothesis, a series of controlled tests was

123
Environ Geol (2008) 54:1707–1722 1719

conducted on the Cardium #1 sample, where two more characteristics (end-point saturation and shape) on IFT.
drainage and imbibition measurements for different IFT The relative permeability curves correspond to the Car-
were conducted besides the measurement at in-situ condi- dium #1 sample tested for different pressures and constant
tions (Bennion and Bachu 2006c). The variation in IFT was temperature and water salinity (see Table 3).
obtained by conducting the measurements for the same Previously it was generally thought that actually it is
temperature and water salinity but for lower pressures viscosity that has an effect on the displacement charac-
(Tables 3 and 4). Previous to these measurements, a series teristics of fluid pair systems (e.g., Odeh 1959; Du Prey and
of measurements were conducted on the Wabamun #2 Lefevre 1973; Bennion and Bachu 2005). The argument
sample (drainage cycle only) to test the effect of pressure, would be that, as the viscosity contrast between the two
temperature and salinity on displacement characteristics fluids increases, it is easier for the more mobile (less vis-
(Bennion and Bachu 2005), by which one test was run for cous) fluid to flow through the pore space. Indeed, plots of
the same pressure and temperature as the in-situ conditions CO2 and water relative permeability during drainage and
but with freshwater, and a second one for the same water imbibition, respectively, and of irreducible water saturation
salinity but for CO2 in gaseous rather than supercritical and trapped gas against the viscosity ratio between for-
phase (low pressure and temperature) (Table 3). The IFT mation water and CO2 (Fig. 11) show a striking similitude
values for the Wabamun #2 series of tests were retrieved with the plots showing the variation of the same variables
after the Cardium #1 measurements, and used together to as a function of IFT (Fig. 10). Unfortunately it is not
examine the effect of IFT on relative permeability. All possible with the current data set to separate these two
these results clearly show that indeed, for the same rock dependent variables, IFT and viscosity ratio, that both de-
properties, IFT has a significant effect on the displacement pend on the primary in-situ independent variables pressure,
characteristics of CO2-brine systems during both CO2 temperature and water salinity. As seen previously, the IFT
injection (drainage) and water invasion (imbibition) cycles. for CO2-brine systems decreases with increasing pressure
Endpoint relative permeability to CO2 during drainage and and increases with increasing temperature and water
to brine during imbibition both decrease with increasing salinity. Water viscosity increases with pressure and
IFT (Fig. 10a, b), confirming the initial hypothesis. salinity, and decreases with increasing temperature (Adams
Increasing IFT has the opposite effect on irreducible water and Bachu 2002), while CO2 viscosity also increases with
saturation during drainage and on trapped gas (CO2) during pressure and decreases with increasing temperature
imbibition, which both increase with increasing IFT (Fenghour et al. 1998). Because the variation of CO2 vis-
(Fig. 10c, d). Examination of results for the drainage cycle cosity is stronger than that of brine viscosity, the net effect
would suggest that the effect is linear (Fig. 10a, c), but the is that the ratio of brine to CO2 viscosities varies direc-
imbibition cycle clearly shows that this is not the case tionally in the same manner as IFT (in other words, the
(Fig. 10b, d). Figure 11 shows comparative plots of rela- mobility of CO2 increases at the same time as IFT de-
tive permeability curves as a function of IFT for brine creases). It is not possible to devise any experiment in
displacement by CO2 (primary drainage) and CO2 dis- which IFT could be varied without varying viscosity, hence
placement by brine (imbibition) that clearly illustrate the the viscosity ratio at the same time because both depend on
dependence of the relative permeability displacement the pressure and temperature of the two fluids and on water

Fig. 10 Dependence of
displacement characteristics of
CO2-brine systems on IFT:
a endpoint relative permeability
to CO2 during drainage,
b endpoint relative permeability
to water during imbibition;
c irreducible water saturation
during drainage, and d gas
(CO2) trapped during imbibition
(dots Cardium #1., triangles
Wabamun #2)

123
1720 Environ Geol (2008) 54:1707–1722

Fig. 11 Dependence on IFT

of the relative permeability
during: a drainage, and
b imbibition, for the Cardium
#1 core sample

salinity. However, based on the displacement characteris- saline aquifers is likely to be implemented sooner rather
tics of both drainage and imbibition processes, it seems that than later because oil and gas reservoirs either do not exist
IFT dominates and is the determining factor in changes in within economic distance from the CO2 source, or do not
relative permeability and saturations, although viscosity have the necessary capacity, or are still in production. The
may also play a role (Bennion and Bachu 2006c) (Fig. 12). injection and spread of CO2 during the operational phase,
If pressure, temperature and water salinity are primary, the migration of the CO2 after cessation of injection, and
independent variables, then one may consider IFT and the trapping of CO2 at irreducible saturation in the wake of
viscosity ratio (or mobility ratio) as secondary variables a CO2 plume depend on the in-situ characteristics of the
depending on the first ones, and relative permeability and injection aquifer: rock pore system and water pressure,
irreducible water and CO2 saturations as tertiary variables temperature and salinity.
that are dependent on the secondary variables, hence ulti- Laboratory measurements at in-situ pressure and tem-
mately on the primary variables. perature conditions on rock samples of various lithologies
from the Wabamun Lake area in Alberta, Canada, and with
water with respective aquifer salinity produced a wide
Conclusions range of results regarding the displacement characteristics
of CO2-brine systems. In situ conditions vary between 8.6
Geological storage of CO2 is a means of reducing signifi- and 27 MPa for pressure, 35C and 75C for temperature,
cantly atmospheric CO2 emissions from large stationary and 27 and 248 g/l for salinity. The interfacial tension
sources such as those in the energy and power generating between CO2 and brine was found to depend strongly on
industries. Although not associated with any enhanced oil pressure for relatively low pressures (<10 MPa), decreas-
or gas recovery to offset some of the costs, storage in deep ing as pressure increases, and displays an asymptotic

Fig. 12 Dependence of
displacement characteristics of
CO2-brine systems on the
viscosity ratio between the two
fluids: a endpoint relative
permeability to CO2 during
drainage, b endpoint relative
permeability to water during
imbibition; c irreducible water
saturation during drainage, and
d gas (CO2) trapped during
imbibition (dots Cardium Fm.,
triangles Wabamun Gp)

123
Environ Geol (2008) 54:1707–1722 1721

behavior towards a constant value for higher pressures. References

Besides pressure, IFT depends strongly on both tempera-
ture and water salinity, increasing significantly as either Adams JJ, Bachu S (2002) Equations of state for basin geofluids:
algorithm review and intercomparison for brines. Geofluids
one of these increases. This dependency is attributed to 2:257–271 (Erratum (2004) Geofluids 4(3):250)
solubility and phase effects. A laboratory program of sys- Bachu S (1993) Basement heat flow in the Western Canada
tematic measurements is needed to develop a generalized Sedimentary Basin. Tectonophysics 222:119–133
relationship for the dependency of IFT on these three in- Bachu S (1999) Flow systems in the Alberta basin: patterns types and
driving mechanisms. Bull Can Petrol Geol 47:455–474
situ primary variables. Bachu S (2003) Screening and ranking of sedimentary basins for
Capillary pressure depends, as expected, on rock prop- sequestration of CO2 in geological media in response to climate
erties, mainly on the pore size distribution (size and shape), change. Environ Geol 44:277–289 doi:10.1007/s00254-003-762-
but also on IFT, decreasing as IFT decreases. In the case of 9doi:10.1007/s00254-003-762-9
Bennion B, Bachu S (2005) Relative permeability characteristics for
the 13 tested rock samples, permeability to brine shows an CO2 displacing water in a variety of potential sequestration
exponential dependence on pore size, with values in the zones in the Western Canada Sedimentary Basin. Paper SPE
range of millidarcies to tens of millidarcies (10–15 to 95547, 15 p., presented at the 2005 SPE Technical Conference
10–14 m2) for sandstones and carbonates characterized by a and Exhibition, Dallas, TX, 9–12 October 2005
Bennion B, Bachu S (2006a) The impact of interfacial tension and
large fraction of macro porosity and median pore size, and pore size distribution/capillary pressure character on CO2
dropping to picodarcies (10–21 to 10–20 m2) for shales. relative permeability at reservoir conditions in CO2-brine
More interestingly, the permeability to CO2 at irreducible systems. Paper SPE 99325, 10 p., presented at the SPE/DOE
brine saturation displays a broad linear relationship with fifteenth symposium on improved oil recovery, Tulsa, OK, USA,
22–26 April 2006
permeability to brine, generally being about one seventh to Bennion B, Bachu S (2006b) Supercritical CO2 and H2S—brine
one fifth of the latter for the sandstone and carbonate rocks drainage and imbibition relative permeability relationships for
with permeability greater than one millidarcy (10–15 m2). intercrystalline sandstone and carbonate formations. Paper SPE
No identifiable dependency was found between irreducible 99326, 13 p., presented at the SPE Europec/EAGE annual
conference and exhibition, Vienna, Austria, 12–15 June 2006
brine and CO2 saturations during both drainage and imbi- Bennion B, Bachu S (2006c) Dependence on temperature, pressure
bition cycles and pore characteristics. However, a clear and salinity of the IFT and relative permeability displacement
dependency of displacement characteristics (end point characteristics of CO2 injected in deep saline aquifers. Paper SPE
relative permeability and shape of relative permeability 102138, 10 p., presented at the 2006 SPE technical conference
and exhibition, San Antonio, TX, 24–27 September 2006
curve and irreducible saturations) on IFT was observed in Cai B-Y, Yang J-T, Guo T-M (1996) Interfacial tension of
two series of measurements conducted on the same rock hydrocarbon + water/brine systems under high pressure. J Chem
sample (a sandstone and a carbonate) to avoid effects of Eng Data 41:493–496
rock characteristics. Endpoint relative permeability to CO2 Chalbaud C, Robin M, Egermann P (2006) Interfacial tension of CO2/
brine systems at reservoirs conditions. In: Gale J, Rokke N,
decreases with increasing IFT for both drainage and Zweigel P, Swenson H (eds) Proceedings of the 8th international
imbibition cycles, while irreducible water saturation for the conference on greenhouse gas control techn. Elsevier, Amster-
drainage cycle, and irreducible CO2 saturation for the dam CD ROM
imbibition cycle increase with increasing IFT. Although a Chun BS, Wilkinson GT (1995) Interfacial tension in high pressure
carbon-dioxide mixtures. Ind Eng Chem Res 34:4371–4377
similar dependency between these displacement charac- Ennis-King JP, Paterson L (2003) Role of convective mixing in the
teristics and the viscosity ratio between brine and CO2 is long-term storage of carbon dioxide in deep saline formations.
observed, the shape of the relative permeability curves Paper SPE 84344, presented at the SPE annual technical
suggests that the dependency is rather on IFT than on conference and exhibition, Denver, CO, USA, 5–8 October 2003
Fenghour A, Wakeham WA, Vesovic V (1998) The viscosity of
viscosity ratio. carbon dioxide. J Phys Chem Ref Data 27(1):31–44
The synthesis presented here of the results of an exper- Gunter WD, Bachu S, Benson SM (2004) The role of hydrogeological
imental program of laboratory measurements regarding the and geochemical trapping in sedimentary basins for secure
displacement characteristics of CO2-brine systems clearly geological storage for carbon dioxide. In: Baines SJ, Worden RH
(eds) Geological storage of carbon dioxide, vol 233. Geological
indicates that relative permeability depends on pore size Society, London, pp 129–145 (special publication)
characteristics and on the interfacial tension between the Hebach A, Oberhof A, Dahmen N, Kögel A, Ederer H, Dinjus E
two fluids, which in turn depends on the in-situ conditions (2002) Interfacial tension at elevated pressure: measurements
of pressure, temperature and water salinity. Neglect of these and correlations in the water + carbon dioxide system. J Chem
Eng Data 47:1540–1546
dependencies when assessing through numerical modeling IEA (International Energy Agency) (2004) Prospects for CO2 Capture
the short and long term fate of CO2 injected in deep saline and Storage. IEA/OECD, Paris, 249 p
aquifers will inevitably lead to erroneous results which may IPCC (Intergovernmental Panel on Climate Change) (2005) IPCC
affect decisions regarding siting, operating parameters and special report on carbon dioxide capture and storage. In: Metz B,
Davidson O, de Coninck HC, Loos M, Mayer LA (eds)
monitoring procedures of operations for CO2 storage in Cambridge University Press, Cambridge 442 p
deep saline aquifers.

123
1722 Environ Geol (2008) 54:1707–1722

Lefebvre du Prey EJ (1973) Factors affecting liquid relative perme- del Rio OI, Neumann AW (1997) Axisymmetric drop shape analysis:
ability of a consolidated porous medium. SPE J 13(1):39–47 computational methods for the measurement of interfacial
Levin Y, Flores-Mena JE (2001) Surface tension of strong electro- properties from the shape and dimensions of pendant and sessile
lytes. Europhys Lett 56(2):187–192 drops. J Coll Int Sci 196:136–147
Kumar A, Noh MH, Pope GA, Sepehrnoori K, Bryant SL, Lake LW Rosman A (1977) Experimental studies of low IFT displacement by
(2005) Simulating CO2 storage in deep saline aquifers. In: CO2 injection. Paper SPE 6723, 10 pp, presented at the SPE
Thomas D, Benson SM (eds) Carbon dioxide capture for storage annual fall technical conference and exhibition, Denver, CO,
in deep geologic formations—results from the CO2 Capture 9–12 October 1977
Project, vol 2. Elsevier, Amsterdam, pp 877–896 Xu T, Apps JA, Preuss K (2003) Reactive geochemical transport
Navascués G (1979) Liquid surfaces: theory of surface tension. Rep simulation to study mineral trapping for CO2 disposal in deep
Prog Phys 42(7):1131–1186 arenaceous formations. J Geophys Res 108(B2):2071
Odeh AS (1959) Effect of viscosity ratio on relative permeability. Zuo Y-X, Stenby EH (1998) Prediction of interfacial tensions of
Petrol Trans AIME 216:346–353 reservoir crude oil and gas condensate systems. Paper SPE
Pickell JJ, Swanson BF (1966) Application of air–mercury and oil–air 38434, SPE Journal 3(2):134–145
capillary pressure in the study of pore structure and fluid Yang D, Tontiwachwuthikul P, Gu Y (2005) Interfacial tensions of
distribution. SPE J 6(1):55–61 the crude oil + reservoir brine + CO2 systems at pressures up to
Porter IW, Price RA, McCrossan RG (1982) The Western Canada 31 MPA and temperatures of 27C and 58C. J Chem Eng Data
Sedimentary Basin. Phil Trans R Soc London, Series A 50:1242–1249
305:169–182

123