PAPER 2009-156

Simulation of Trapping Processes for

CO2 Storage in Saline Aquifers
L. NGHIEM, V. SHRIVASTAVA, B. KOHSE, M. HASSAM, C. YANG
Computer Modelling Group Ltd.

This paper is accepted for the Proceedings of the Canadian International Petroleum Conference (CIPC) 2009, Calgary,
Alberta, Canada, 16‐18 June 2009. This paper will be considered for publication in Petroleum Society journals.
Publication rights are reserved. This is a pre‐print and subject to correction.

Abstract
Important modelling techniques for CO2 storage in saline
aquifers are discussed, in particular solubility trapping,
residual gas trapping and geochemistry for mineral trapping.
These modelling techniques are applied to the simulation of
several important aspects of CO2 storage, which include
optimizing total trapping through water injection, assessing the
security of residual trapping with regards to leakage through
the cap rocks and evaluating the conversion of CO2 into
minerals.
Introduction
Saline aquifers represent the most important venue for CO2
storage since they have the largest capacity among all other
venues (coal seams, depleted gas and oil fields). Table 1 (DOE
and NETL, Carbon and Sequestration Atlas for the United
States and Canada, 2008) shows the dominance of saline
aquifers for large scale deployment of CO2 storage.
There are several CO2 trapping mechanisms in saline
aquifers: (1) structural trapping; (2) residual gas trapping; (3)
solubility trapping and (4) mineral trapping. Structural trapping
involves the storage of CO2 in a geological structure as a free
gas or super-critical fluid. CO2 can flow and escape through the
cap rock or sealing faults if the integrity of the latter is
compromised. Residual gas trapping consists of storing CO2 as
an immobile gas in the porous media. This process has been
identified as one of the most important processes for safe CO2
storage as the immobile gas can be kept away from the cap
rock. CO2 is highly soluble in brine and solubility trapping is
essentially the impetus for CO2 storage in saline aquifers. As
+ −
CO2 dissolves in brine, it decomposes into H and HCO3 .
These ions in turn react with the minerals in place. Depending
on the mineralogy of the formation, these reactions could
induce precipitation of carbonate minerals such as Calcite,
Dolomite and Siderite, which corresponds essentially to the
conversion of CO2 into minerals. This paper discusses the
modelling of the physics associated with CO2 storage in saline
aquifers and illustrates through examples the important storage
processes.
Table 1: Storage Capacity for CO2 Storage in
North America
Formation Type 109 metric tons %
Saline Aquifers 3297 – 12,618 91.8 – 97.5
Unmineable Coal Seams 157 – 178 4.4 – 1.4
Mature Oil & Gas Reservoirs 138 3.8 – 1.1
Total Capacity 3,592 – 12,934 100.0
CO2 has a critical pressure of 7,376 kPa and a critical
temperature of 304.2 K (31ºC). Thus, CO2 is normally a
supercritical fluid at aquifer conditions. For convenience, CO2
is referred to as a “gas” in this paper. To be general, the gas
phase is treated as a multicomponent mixture.
The simulation and modelling techniques described in this
paper are based on the geochemical equation-of-state
compositional simulator, GEM™ with the greenhouse gas

1
option (Nghiem et al., 2004). This simulator has been used
extensively for modelling CO2 storage in saline aquifers in the
past five years. Pruess et al. (2003) and Zhang et al. (2007)
describe a simulator for CO2 storage in aquifers with similar
features.
Solubility Trapping
The accurate modeling of gas solubility in the aqueous phase
(brine) is important as brine can dissolve a large amount of
CO2. The gas solubility in brine is modelled as a phase
equilibrium process, which is governed by the equality of
fugacities in the gas and aqueous phase, i.e.
f ig = f ia , i = 1,..., n c ............................................................ (1)
The fugacity fig of component i in the gas phase is calculated
with an equation of state (EOS) (e.g. Peng-Robinson, 1976).
The fugacity of component i in the aqueous phase is modeled,
either with Henry’s law (Li and Nghiem, 1986), i.e.
al. (2005) provide a comprehensive summary of several residual
gas trapping models.
In this paper, the classical Land’s model (Land, 1968) is
used. Figure 2 shows a typical gas relative permeability curve
which exhibits hysteresis. When the gas (CO2) saturation
increases, the gas relative permeability follows the drainage
curve k drg (black curve in Figure 2). If at the saturation S*gi on
the drainage curve, the gas saturation reverses its course and
decreases, the gas relative permeability follows the imbibition
curve k irg (red curve). The Land’s coefficient C is expressed as
1 1
C = − ............................................................(4)
Sgt , max Sg , max
where Sg,max is the maximum gas saturation that could be
attained and Sgt,max is the maximum trapped gas saturation. The
residual gas saturation for a given S*gi is:
( )
S*gt S*gi =
S*gi
1 + C S*gi
..............................................................(5)

f ia = H i ⋅ y ia ......................................................................... (2) To promote residual gas trapping, CO2 is injected near the
bottom of the aquifer and as the CO2 rises to the top of the
or with the Søreide-Whitson-Peng-Robinson (SWPR) EOS
aquifer, imbibition occurs and traps CO2.
(Søreide and Whitson, 1992).
Henry’s constants are functions of pressure, temperature and
salinity and are normally expressed as follows:
1 p
RT ∫ p
ln H i = ln H *i + * v i dp ............................................. (3)

where
Hi = Henry’s constant at p and T
Hi * = Henry’s constant at p* and T
R = gas constant
vi = partial molar volume of component i in solution

Harvey (1996) gives correlations for Henry’s constants for

many gases including CO2 at the saturation pressure of H2O and
temperature T. Bakker (2003) provides correlations for effect of
salinity on Henry’s constant. Garcia (2001) describes a
correlation for the molar volume for CO2 in water. Gas
solubility increases with increasing pressure and decreases with
increasing temperature and salinity.
In the SWPR EOS, Søreide and Whitson use different
interaction coefficients for the gas and aqueous phase (Peng and
Robinson, 1980). In addition, they give correlations for
interaction coefficients of gaseous components in the water
phase that are functions of temperature and salinity.
An accurate prediction of the aqueous phase density is
another important aspect for modelling CO2 storage in saline
aquifers. The brine density increases with increasing CO2
solubility and creates convection currents that facilitate the
mixing of CO2 with unsaturated brine and increase solubility
trapping (Nghiem et. al, 2004; Ennis-King and Patterson, 2005).
The density of brine without gas solubility is calculated with the
Rowe and Chou (1970) correlation. The density of brine with
soluble gas is calculated as using the partial molar volume of
gaseous components in solution as described in Garcia (2001).
Residual Gas Trapping
Residual gas trapping was first suggested by Kumar et al.
(2005) as an important mechanism for trapping CO2. The idea is
to take advantage of the capillary effect and wettability effect to
trap CO2 as an immobile phase in the porous media. Spiteri et
Figure 2: Land’s residual gas trapping model
Mineral Trapping
When injected CO2 dissolves in an aqueous phase, it
dissociates into H+ and HCO3− ions through the reactions:
CO 2 (aq ) + H 2O ↔ H + + HCO3− ............................................(6)
In Equation (6), CO2(aq) is used to denote the CO2 that is
dissolved in the aqueous phase. The ions in turn react with other
ions in solution and with the mineral in place to give rise to the
precipitation and dissolution of various minerals. The reaction
rate is very slow and it takes hundreds of years to obtain a
reasonable amount of precipitation. A typical reaction is the
precipitation and dissolution of Calcite, i.e.,
Calcite + H + ↔ Ca + + + HCO3− ..............................................(7)
The above reactions are called geochemical reactions and they
occur between components in the aqueous phase, and between
minerals and aqueous components. These reactions are
reversible. The intra-aqueous reactions such as the one in
Equation (6) are very fast while the mineral dissolution and
precipitation reactions (e.g. reaction in Equation (7)) are very
slow.

2
Chemical Equilibrium Reactions
Since the intra-aqueous reactions are very fast (almost
instantaneous from a practical view point), they are modelled as
chemical equilibrium reactions. Chemical equilibrium reactions
are governed by chemical equilibrium constants (Bethke, 1996)
Q α − K eq,α = 0, α = 1,..., R aq ..........................................(8)
with
n aq
ν a (Ca + + ) ⋅ a ( HCO3− )
Qα = ∏ a k kα ......................................................................(9)
k =1
where naq is the number of aqueous components, Raq is the
number of intra-aqueous chemical equilibrium reactions, Keq,α
is the chemical equilibrium constant for the aqueous reaction α,
ak is the activity of component k, νk,α are the stoichiometry
coefficients of the chemical equilibrium reactions and Qα is the
activity product. Tables of values of Keq,α as a function of
temperature for many aqueous reactions are given in Kharaka et
al. (1989) and Delany et al. (1990). The stoichiometry
coefficients are negative for reactant species and positive for
product species. For example, the activity product for the
reaction in Equation (6) is
a (H + ) ⋅ a (HCO3− )
Q = ..................................................(10)
a (CO 2 (aq)) ⋅ a (H 2O)
The activities ak are related to the molality mi (moles per kg of
H2O) as follows:
a i = γ i m i , i = 1,.., n aq ......................................................(11)
where γi is the activity coefficient. For ideal solutions, γi = 1,
and the activity is equal to the molality. However, for most
cases, the solution is non-ideal and one of the preferred models
for ionic activity coefficients is the B-dot model (Bethke, 1996):
A γ z i2 I where E aβ is the activation energy for reaction β [J/mol] and
log γ i = −  I ..........................................(12)
+ B
1 + å i Bγ I s)].
where A γ , B γ and B  are temperature dependent parameters and
åi is the ion size parameter. I is the ionic strength, is given by:
n aq
1
I =
2
∑ m k z 2k .................................................................(13)
k =1
with zk being the charge of the k-th ion. The activity of CO2(aq)
and H2O can be assumed to be equal to 1 and the activity of
H2O is close to 1.
At high salinity, the Pitzer (1987) model is more accurate
than the B-dot model, but the computation is more demanding.
Mineral Dissolution and Precipitation Reactions
The dissolution or precipitation of minerals follows the
reaction (Bethke, 1996):
⎛ Qβ ⎞ bottom of the aquifer and the CO2 bubble rises to the top of the
rβ = Â β k β ⎜1 − ⎟, β = 1,..., R ...............................(14)
mn
⎜ K eq,β ⎟
⎝ ⎠
where Rmn is the number of mineral reactions, rβ is the rate,
 β is the reactive surface area for mineral β, kβ is the rate
constant of the mineral reaction β, Keq,β is the chemical
equilibrium constant of the mineral reaction β and Qβ is the
3
activity product of the mineral reaction β. The activity product
Qβ is analogous to the activity product for aqueous chemical
equilibrium reactions (Equation 9), i.e.
n aq νk
Qβ = ∏ak β .................................................................... (15)
k =1
For example, the activity product of the mineral reaction in
Equation (7) is:
Q = .................................................... (16)
a (H + )
The product does not involve the activities of the minerals as
they are equal to unity. The chemical equilibrium constants
Keq,β for many minerals are given in Kharaka et al. (1989) and
Delany et al. (1990). The ratio (Qβ/K eq,β) is called the
saturation index of the reaction. If (Qβ/K eq,β) < 1, mineral
dissolution occurs and if (Qβ/K eq,β) > 1, mineral precipitation
occurs. In this paper, the convention adopted is that the rate rβ is
negative for dissolution and positive for precipitation. The rate
is zero when (Qβ/K eq,β) = 1.
The rate of dissolution/precipitation in Equation (13) applies
to the mineral. The rate of formation/consumption of the
different aqueous species is obtained by multiplying rβ by the
respective stoichiometry coefficient, i.e.
rkβ = ν kβ ⋅ rβ ....................................................................... (17)
The reaction rate constants, kβ, are normally reported in the
literature at a reference temperature T0 (usually 298.15 K or
25°C). The following equation is used to calculate the rate
constant at a given temperature T:
⎡ E aβ ⎛1 1 ⎞⎤
k β = k 0β exp ⎢ − ⎜ − ⎟
⎜ T T ⎟ ⎥ ........................................ (18)
⎢⎣ R ⎝ 0 ⎠ ⎥⎦
k 0β is the reaction rate constant for reaction β at T0 [mol/(m2
Assessment of CO2 Storage Scenarios
An important goal for CO2 storage in saline aquifers is to
increase the storage security by reducing the risk of leakage
through the cap rock over the long storage period. The risk in
structural trapping is the highest as the CO2 is mobile and will
migrate through the cap rock if there are micro fractures or if
the rock fails because of geomechanical and/or geochemical
effects. Solubility trapping is safe as CO2 is stored as a soluble
component in brine and will only come out of solution if there
is a substantial drop in pressure. This is very unlikely for large
aquifers. Nevertheless, there is the potential migration of CO2
saturated brine over large distances if the aquifer is active.
Residual gas trapping is also a safe trapping mechanism,
especially if the immobile CO2 is kept away from the cap rock.
Residual gas trapping is induced when CO2 is injected near the
aquifer due to buoyancy. Residual gas trapping has been
recognized as the most rapid method to trap CO2 with time
scales in the order of years to decades (Ennis-King and
Paterson, 2002; Kumar et. al., 2005; Obi and Blunt, 2006;
Juanes et al., 2007; Qi et al., 2007). The conversion of CO2 into
carbonate minerals (Calcite, Dolomite, Siderite), also referred to
as mineral trapping, relies on the dissolution of other minerals
in the aquifer that provide Ca++, Mg++ and Fe++ for the
conversion (Thibeau et al., 2007). CO2 mineralization is
probably the safest mode for CO2 storage. However,
mineralization takes hundreds or thousands of years to yield a
reasonable quantity of minerals. Figure 3 from the IPCC report
(2005) shows qualitatively the increase in storage security with
time.
Figure 3: Increase in CO2 storage security with time
(IPCC, 2005)
In the remainder of this paper, it is shown how simulation
could be used to evaluate and optimize the storage process. The
first case involves the optimization of the solubility trapping
and residual gas trapping. A scenario of leakage through the cap
rock is also simulated to assess the storage security. The second
case involves the evaluation of the mineral trapping in an
aquifer. These cases are used to highlight the different trapping
processes. Similar simulation studies should be carried out for
specific applications.
Optimization of Residual Gas and
Solubility Trapping
An important endeavor in designing a CO2 storage process is
to speed up the storage amount by accelerating solubility and
residual gas trapping. Water (brine) injection is normally used
to achieve that goal. Leonenko and Keith (2008) suggested
brine injection to accelerate the dissolution of CO2. Qi et al.
(2007) investigated a WAG injection procedure to increase
residual gas trapping. Nghiem et al. (2008) proposed injecting
brine above the CO2 injector to accelerate and increase the
amount of residual gas trapping. They applied an optimization
procedure to determine the rate, duration and location of the
brine injector to maximize the amount of residual gas trapping
and solubility trapping for a low-permeability and a high-
permeability aquifer. Residual gas trapping and solubility
trapping are competing trapping mechanisms that occur
simultaneously during the injection and post-injection period.
This paper examines in detail the interaction of the two
mechanisms and shows how the process could be optimized.
The aquifer considered in the study consists of a quarter of
an element of symmetry shown in Figure 4. The average initial
conditions are: pressure = 21,472 kPa; temperature = 79.5ºC
and salinity = 32,000 ppm. The grid system is 71 × 71 × 100 =
504,100 with a uniform gridblock size of 15 m × 15 m × 3 m.
The aquifer thickness is 300 m and the lateral extent of the
element of symmetry is 1065 m. Large volume modifiers (107)
are applied to the outer boundary gridblocks to represent a
4
constant pressure boundary. CO2 is injected at a depth of 2276
m (near the bottom of the aquifer) on 2000-01-01 at a rate of
106 std m3/day (1869 ton/day) for a period of 25 years. The
cumulative CO2 injected is 9.1313 × 109 std m3/day or 1.7068 ×
107 tons. Figure 5 shows the gas and water relative permeability
curves used in the model. The value of Sg,max is 0.8 and the
value of Sgt,max for the calculation of residual trapped gas
saturation is 0.4. The porosity is 0.2. A low-permeability
aquifer with kh = 100 md and kv = 10 md and a high-
permeability aquifer with kh = 500 md and kv = 50 md are
considered in this study.
Figure 4: Aquifer quarter of element of symmetry
(The displayed value is the depth in m)
1.2
1 Gas
Water
Relative Permeability
0.8
0.6
0.4
0.2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Gas Saturation (Sg)
Figure 5: Relative permeability curves
The following trapping indices are defined:
Residual Gas Trapping Index RTI(t) =
Total mass of CO 2 trapped as residual gas at time t
.......... (18a)
Total mass of CO 2 injected at time t
Solubility Trapping Index STI(t) =
Total mass of CO 2 so lub le in brine at time t
............(18b)
Total mass of CO 2 injected at time t
Trapping Efficiency Index TEI(t) = RTI(t) + STI(t) .......... (18c)
Residual gas trapping and solubility trapping are competing
storage processes, i.e. an increase in residual gas trapping will
decrease the solubility trapping and vice-versa. The goal is to
maximize the TEI by adjusting the water injection rate, location
and duration. Possible values for the different water injection
parameters are shown in Table 2.
Since the enumeration of the parameter space is very small
(4×3×4 = 48), all 48 runs were performed. If the parameter
space is large, it is normally not possible to perform all the runs
corresponding to the enumeration of the parameter space.
Therefore, other optimization methods such as the one
described in Yang et al. (2007a, 2007b) or the Particle Swarm The results show that for this high permeability aquifer,
Optimization (PSO) technique (Kennedy and Eberhart, 1995) or water injection does not improve the TEI. Indeed, for high-
Genetic Algorithms can be used. permeability aquifers, there is substantial mixing between the
injected CO2 and the brine, which makes solubility trapping
significant.
Table 2: Water injector parameters
Description Values
Water rate at reservoir 1480, 2960, 4440, 5920
conditions (m3/day)
Depth of water injector (m) 2096, 2156, 2216
Water injection period (years) 10, 15, 20, 25

Low-Permeability Aquifer
The solutions for the low-permeability aquifer (kh = 100 md; kv
= 10 md) at 50 years are plotted in Figure 6. The case without
water injection is also shown. The RTI and STI for the case
without water injection is 0.456 and 0.345 respectively, which
corresponds to a TEI of 0.801. The line corresponding to TEI =
0.801 is plotted in red in Figure 6. Figure 6 clearly demonstrates
that the injection of water improves substantially the TEI
through an increase in RTI although the STI is decreased. In
Figure 6, the Pareto optimal solutions are optimal solutions that
correspond to different weightings of RTI and STI in the
objective function, i.e. J = ω1(RTI) + ω2(STI). The TEI
corresponds to ω1 = ω2 = 1. The maximum TEI corresponds to
injecting water at the largest rate (5920 res. m3/day), at the
largest depth (2216 m) for a period of 20 years. There is a
substantial increase in TEI from a value of 0.801 (80% of the
injected CO2) for the case without water injection to a value of
0.971 (97% of the injected CO2) with water injection. Note that
the optimal operating conditions found are among the discrete
values shown in Table 2. In fact, there are possibly other
operating conditions besides the values listed that would give a
better TEI. If more discrete values for each parameter are used, Figure 7: Evolution of RTI, STI and TEI with time
the parameter space is larger and the optimal operating kh = 100 md; kv = 10md
conditions found would likely be better than those currently 0.80
obtained with the small parameter space. Non Optimal Solutions
0.35 Pareto Optimal Solutions
0.75 No Water Injection
Non Optimal Solutions TEI = 0.926
Pareto Optimal Solutions
STI (50 years)

0.30 No Water Injection 0.70

TEI = 0.801
STI (50 years)

0.65
0.25

0.60

0.20

0.55
0.15 0.20 0.25 0.30 0.35
0.15 RTI (50 years)
0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80
RTI (50 years)
Figure 8: Pareto optimal solutions at 50 years
kh = 500 md; kv = 50md
Figure 6: Pareto optimal solutions at 50 years
kh = 100 md; kv = 10md
Leakage Scenario
Figure 7 shows the evolution of RTI, STI and TEI with time
for the cases with and without water injection. The cumulative It is important to assess the security of a CO2 storage process
amount of injected CO2 is also shown. regarding the potential leakage through the cap rock. Reservoir
simulation provides a convenient means to evaluate different
High-Permeability Aquifer leakage scenarios by creating “holes” in the cap rock.
For the CO2 storage in the low-permeability aquifer
For the high-permeability aquifer case with kh = 500 md and
discussed above, a “leakage” producer was opened at the top of
kv = 50 md, the RTI and STI for all the runs corresponding to
the aquifer (see Figure 9) at 50 years with a pressure drawdown
the water injection parameters in Table 2 are shown in Figure 8.
of 1000 kPa. The well relative permeabilities are designed so
The RTI and STI for the case without water injection is 0.171
that only gaseous CO2 would escape through this well. The
and 0.755 respectively, which corresponds a TEI of 0.926.
reasoning is that the brine hydro static head will prevent the
brine from escaping through the leakage well. Furthermore, the

5
brine with soluble CO2 is heavier that the brine without CO2,
which will reduce further the possibility of brine leakage.
Figure 10 shows the amount of leaked CO2 for the case with
and without water injection. The results clearly show the
significant increase in CO2 storage security when water
injection was used to promote residual gas trapping below the
cap rock.
Figure 9: Aquifer with leakage well
Figure 10: Leaked CO2 for the aquifer with
kh = 100 md; kv = 10 md
Mineral Trapping
The case study discussed in this section illustrates the effect
of the aquifer mineralogy on the mineral trapping process. The
grid used for the 3D simulation was extracted from the Sintef
3D model of the Utsira formation (Lindeberg et al., 2002). The
reduced model has 41 × 45 × 26 (= 47,970) gridblocks with a 6
km × 7 km areal extension and an average of 200 m of
thickness. Figure 11 shows a 3D view of the extracted grid. The
initial pressures at the top and bottom of the aquifer are 8,110
kPa and 10,831kPa respectively, while the temperatures at these
locations are 28.5°C and 37.6°C respectively. CO2 is injected in
gridblock (28,25,5) starting on 2000−01−01 at a rate of 106 std
m3/day for 25 years. The total injection is 3.63 × 1011 moles (16
millions tons) of CO2. The intra aqueous and mineral reactions
are shown in Table 3. The reactions are not typical of the
Sleiper/Utsira formation but are used here to illustrate the
mineralization of CO2. Further details can be found in Thibeau
et al. (2007).
The intra aqueous reactions are very fast and are modeled as
chemical equilibrium reactions. The mineral
dissolution/precipitation reactions are very slow and are
modeled as rate-dependent reactions. Reaction (3) in Table 3 is
an important reaction in CO2 storage processes. The generated
H+ ions react with many minerals (dissolution reactions) as
6
shown in Table 3. All reactions except Reaction 7 involve the
dissolution of minerals if there is a surplus of H+ in the system.
The bicarbonate ion HCO3– is crucial in the mineralization of
CO2. The HCO3– can combine with Ca++, Mg++ and Fe++ to
precipitate Calcite, Dolomite and Siderite according to
Reactions 5, 6 and 10. Those ions are supplied by the
dissolution of Anorthite which provides Ca++ according to
Reaction 12, of Illite which gives Mg++ according to Reaction 9
and of Annite which supplies Fe++ according to Reaction 12.
Without the presence of those minerals (or similar minerals) in
the aquifer), mineralization of CO2 would not occur.
Figure 11: Grid structure and depth of Utsira formation
Table 3: Geochemical reactions
Intra-aqueous chemical equilibrium reactions
1. H 2 O ↔ H + + OH −
2. CO 2 (aq ) + H 2 O ↔ H + + HCO 3−
3. HCO 3− ↔ H + + CO 3− −
4. AlOH + + + H + ↔ Al+ + + + H 2 O
Mineral dissolution/precipitation reactions
5. Calcite + H + ↔ Ca + + + HCO 3−
6. Dolomite + 2 H + ↔ Ca + + + Mg + + + 2 HCO 3−
7. Chalcedony ↔ SiO 2 (aq )
8. Kaolinite + 6 H + ↔ 5 H 2 O + 2 SiO 2 + 2 Al+ + +
9. Illite + 8 H + ↔ 5 H 2O + 0.6 K + + 0.25 Mg + + +
2.3 Al+ + + + 3.5 SiO 2 (aq)
10. Siderite + H ↔ Fe + + + HCO 3−
+
11. Annite + 10 H + ↔ 3 Fe + + + K + +
Al+ + + + 3 SiO 2 (aq) + 6 H 2 O
12. Anorthite + 8 H + ↔ Ca + + + 2 Al+ + + +
2 SiO 2 (aq ) + 4 H 2O
Figure 12 shows the evolution of the total moles of CO2(g),
CO2(aq) and HCO3– in the aquifer. After the 25-year injection
period, the amount of CO2(aq) increases as more and more
CO2(g) goes into solution. As CO2 dissolves in the aqueous
phase (brine), the density of the CO2-saturated brine increases
and migrates to the bottom of the aquifer as plumes. This
creates convection currents that bring unsaturated brine in
contact with CO2(g) and induces additional solubility. The ions
H+ and HCO3– that result from the dissociation of CO2(aq)
according to Reaction 2 reacted with the mineral in place. The
ions Ca++, Mg++ and Fe++ released combine with HCO3– to
precipitate Calcite, Dolomite and Siderite. Figure 13 shows the
total amounts of precipitation/dissolution of those minerals.
Figure 13 shows a net precipitation of Calcite and Siderite and a
net dissolution of Dolomite. At beginning of 2300, the
dissolution curve for Dolomite levels off and shows a positive
slope afterwards. This indicates a precipitation of Dolomite in
the aquifer. Figure 14 shows the dissolution of Anorthite,
Annite and Illite that provides the necessary ions for the
mineralization of CO2.
Table 4 shows the fate of the injected CO2 after 600 years
and 1000 years based on 100 moles of injected CO2. The
amount of CO2 in the form of CO3– – is negligible. Note that
there is 1 mole of CO2 in 1 mole of Calcite and Siderite while
there is 2 moles of CO2 in 1 mole of Dolomite. The
mineralization is quite effective for this aquifer with the
presence of Anorthite, Annite and Illite.
• gas solubility in brine as a function of pressure, temperature
and salinity;
• relative permeability with hysteresis for residual gas
trapping;
• geochemical reactions involving intra-aqueous equilibrium
reactions and mineral dissolution/precipitation kinetics for
mineral trapping.

Several important aspects of CO2 storage in saline aquifers

are illustrated with simulation:
• Solubility trapping and residual gas trapping are important
competing storage mechanisms;
• Water injection could be used to increase residual gas
trapping and consequently total trapping in low-permeability
aquifers; the total trapping could be maximized by adjusting
the water injector location, rate and duration.
• For high-permeability aquifers, water injection does not
increase the total trapping.
• “Leakage” producers could be used to assess the security of
Figure 12: Evolution of CO2(g), CO2(aq) and HCO3– the storage process with regards to leakage through the cap
rock. It was shown that residual gas trapping is very effective
in increasing the security of CO2 storage.
• Mineral trapping or conversion of CO2 into carbonate
minerals (Calcite, Dolomite, Siderite) occurs in aquifers that
contains providers of Ca++, Mg++, Fe++. The mineralization
process takes a long time, hundreds to thousands of years.

Acknowledgements
The authors thank Colin Card and Dingjun Chen for
discussions regarding the optimization of the storage process
with water injection.
Figure 13: Evolution of Calcite, Dolomite and Siderite
NOMENCLATURE
 k = reactive surface area
C = Land’s constant
E = activation energy [J/mol]
fij = fugacity of component i in phase j, j = a, g [kPa]
Hi = Henry’s constant for component i [kPa]
I = ionic strength
kβ = rate constant for reaction β
Kα,eq = equilibrium constant for reaction α
mi = molality of component i
naq = number of aqueous components
nc = number of components in the gas phase
p = pressure [kPa]
Figure 14: Evolution of Annite, Anorthite, Chalcedony,
Qα = activity product of reaction α
Kaolinite and Illite
R = gas constant
Raq = number of intra-aqueous reactions
Table 4: Fate of injected CO2 in moles Rmn = number of mineral reactions
based on 100 moles of injected CO2 Sg = gas saturation
Components 600 years 1000 years T = absolute temperature [K]
CO2(g) 59.5 43.5 vi = partial molar volume of component in solution
CO2(aq) 26.9 34.1 zk = charge of ion k
HCO3– 9.6 11.3 γ = activity coefficient
Calcite 6.1 8.1 ν = stoichiometry coefficient
Dolomite –5.8 –4.3
Siderite 3.7 7.3
Total Minerals 4.0 11.1 REFERENCES
1. BAKKER, R.J., “Package FLUIDS 1. Computer
Conclusion Programs for Analysis of Fluid Inclusion Data and
for Modelling Bulk Fluid Properties,” Chemical
This paper describes the modelling of the important physics Geology, Vol. 194, pp 3-23, 2003.
for CO2 storage in saline aquifers, namely:

7
2. BETHKE, C.M.,: Geochemical Reaction Modeling,
Oxford University Press, New York, 1996
3. DELANY, J.M. and LUNDEEN, S.R.: “The LLNL
Thermo-chemical Database,” Lawrence Livermore
National Laboratory Report UCRL-21658.
4. ENNIS-KING, J. and PATTERSON, L., Role of
Convective Mixing in the Long-Term Storage of Carbon
Dioxide in Deep Saline Formation, SPEJ, Vol. 10, No. 3,
pp. 349-356, September 2005.
5. HARVEY, A.H., “Semiempirical Correlation for Henry’s
Constants over Large Temperature Ranges,” AIChE J,
Vol. 42, No. 5, pp. (May 1996) 1491-1494, May 1996.
6. IPCC, Special Report on Carbon Dioxide Capture
and Storage, Metz, B., Davidson, O., de Coninck,
H., Loos, M., Meyer, L.A. (Eds.), Cambridge
University Press, New York, N.Y., 2005.
7. GARCIA, J.E.: “Density of Aqueous Solutions of
CO2,” Lawrence Berkeley National Laboratory
Report, October 2001.
8. JUANES, R., SPITERI, E.J., ORR, F.M. Jr. and
BLUNT, M.J.: “Impact of Relative Permeability
Hysteresis on Geological CO2 Storage,” Water
Resources Research, Vol. 42 (2006) W12418.
9. KENNEDY, J. and EBERHART, R.: “Particle
Swarm Optimization,” Proceedings IEEE Int.
Conf. on Neural Network, Piscataway, New
Jersey, pp. 1942-1948, 1995.
10. KHARAKA, Y.K., GUNTER, W.D.,
AGGARWAL, P.K., PERKINS, E. and
DEBRAAL, J.D., “SOLMINEQ.88: A Computer
Program for Geochemical Modeling of Water-
Rock Reactions,” U.S. Geological Survey Water-
Resources Investigations Report 88-4227, Menlo
Park, California, 1989.
11. KUMAR, A., OZAH, R., NOH, M., POPE, G.A.,
BRYANT, S., SEPEHRNOORI, K. and LAKE,
L.: “Reservoir Simulation of CO2 Storage in Deep
Saline Aquifers,” SPEJ, Vol. 10, pp. 336-348,
September 2005.
12. LAND, C.S., “Calculation of Imbibition Relative
Permeability for Two- and Three-Phase Flow from
Rock Properties,” Soc. Pet. Eng. J., Vol. 8, No. 2,
pp. 149-156, June 1968.
13. LEONENKO, Y. and KEITH, D., “Reservoir
Engineering to Accelerate the Dissolution of CO2
Stored in Aquifers,” Environ. Sci. Technol., Vol.
42, pp. 2742-2747, 2008.
14. LI, Y.-K. and NGHIEM, L.X.: “Phase Equilibria of Oil,
Gas and Water/Brine Mixtures from a Cubic Equation of
State and Henry's Law,” Can. J. Chem. Eng., Vol. 64,
pp. 486-496, June 1986.
15. LINDEBERG, E., BERGMO, P., and MOEN, A.,
“The Long-term Fate of CO2 Injected into an
Aquifer”, presented of the 6th International
conference on Greenhouse Gas Control
Technology (GHGT-6), Kyoto, Japan, 1-4 October
2002.
16. NGHIEM, L., SAMMON, P., GRABENSTETTER, J. and
OHKUMA, H., “Modeling CO2 Storage in Aquifers
with a Fully-Coupled Geochemical EOS Compositional
8
Simulator,” paper SPE 89474, presented at the 2004
SPE/DOE Fourteenth Symposium of Improved Oil
Recovery, Tulsa, Oklahoma, U.S.A., 17–21 April 2004.
17. NGHIEM, L., YANG, C., SHRIVASTAVA, V., KOHSE,
B., HASSAM, M. and COLIN, C., “Risk Mitigation
Through the Optimization of Residual Gas and
Solubility Trapping for CO2 Storage in Saline Aquifers,”
presented at the 9th International Conference on
Greenhouse Gas Control Technologies (GHGT-9),
Washington DC, 16 – 20 November 2008.
18. NGHIEM, L., YANG, C., SHRIVASTAVA, V., KOHSE,
B., HASSAM, M., CHEN, D and COLIN, C.,
“Optimization of Residual Gas and Solubility Trapping
for CO2 Storage in Saline Aquifers,” paper SPE 119080,
presented at the 2009 SPE Reservoir Simulation
Symposium, The Woodlands, Texas, U.S.A., 2–4
February 2009.
19. OBI, E-O.J. and BLUNT, M.J., “Streamline-Based
Simulation of Carbon Dioxide Storage in a North
Sea Aquifer,” Water Resources Research, Vol. 42
(2006) W03414.
20. PENG, D.Y. and ROBINSON, D.B., “A New Two-
Constant Equation of State,” Ind. Eng. Chem. Fundam.,
Vol. 15, pp. 59-64, 1976.
21. PENG, D.Y. and ROBINSON, D.B, “Two- and Three-
Phase Equilibrium Calculations for Coal Gasification
and Related Processes,” Thermodynamics of Aqueous
Systems with Industrial Applications, ACS Symposium
Series, Vol. 133, pp. 393-414, 1980..
22. PITZER, K.S., A Thermodynamic Model for Aqueous
Solutions of Liquid-Like Density, in Thermodynamic
Modeling of Geological Materials: Minerals, Fluids and
Melts, I.S.E Carmichael and H.P. Eugster (eds). Reviews
in Mineralogy, Vol. 17, pp. 97-142, 1987.
23. PRUESS, K., XU, T., APPS, J., GARCIA, J,
“Numerical Modeling of Aquifer Disposal of
CO2”, SPEJ, Vol. 8, Number 1, pp. 49-60, March
2003.
24. QI, R., BERALDO, V., LAFORCE, T. and
BLUNT, M.J.: “Design of Carbon Dioxide Storage
in a North Sea Aquifer Using Streamline-Based
Simulation,” paper SPE 109905, presented at the
2007 Annual Technical Conference and
Exhibition, Anaheim, California, U.S.A., 11–14
November 2007.
25. ROWE, A.M. and CHOU, J.C.S.: “Pressure-
Volume-Temperature-Concentration Relation of
Aqueous NaCl Solutions”, J. Chem. Eng. Data,
Vol. 15, pp. 61-66, 1970.
26. SØREIDE, I. and WHITSON, C.H., “Peng-Robinson
Predictions for Hydrocarbons, CO2, N2 and H2S With
Pure Water and NaCl-Brines,” Fluid Phase Equilibria,
Vol. 77, pp. 217-240, 1992.
27. SPITERI, E.J., JUANES, R., BLUNT, M.J. and
ORR, F.M. Jr.: “Relative Permeability Hysteresis:
Trapping Models and Application to Geological
CO2 Sequestration,” paper SPE 96448, presented
at the 2005 Annual Technical Conference and
Exhibition, Dallas, Texas, U.S.A. 9-12 October
2005.
28. THIBEAU, S., NGHIEM, L.X. and OHKUMA, H.: “A
Modeling Study of the Role of Selected Mineral in
Enhancing CO2 Mineralization During CO2 Aquifer
Storage, paper SPE 109739, presented at the 2007
Annual Technical Conference and Exhibition, Anaheim,
California, U.S.A., 11–14 November 2007.

29. YANG, C., CARD, C.C. and NGHIEM, L.:

“Economic Optimization and Uncertainty
Assessment of Commercial SAGD Operations,”
paper 2007-029, presented at the Canadian
International Petroleum Conference, Calgary,
Alberta, Canada, June 11-14, 2007a.

30. YANG, C., NGHIEM, L., CARD, C. and

BREMEIER, M.: “Reservoir Model Uncertainty
Quantification through Computer-Assisted History
Matching,” paper SPE 109825, presented at the
2007 SPE Annual Technical Conference and
Exhibition, Anaheim, California, U.S.A., 11-14
November 2007b.
31. ZHANG, K., DOUGHTY, C., WU, Y.-S. and
PRUESS,K., “Efficient Parallel Simulation of CO2
Geologic Sequestration in Saline Aquifers,” paper SPE
106206, presented at the 2007 SPE Reservoir
Simulation Symposium, Houston, Texas, U.S.A., 26-28
February 2007.