SPE 113370

A Novel Foam Concept with CO

2
Dissolved Surfactants
Viet Q. Le, SPE, Quoc P. Nguyen, SPE, The University of Texas at Austin, and Aaron W. Sanders, The Dow
Chemical Company
Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE/DOE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, U.S.A., 1923April2008.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.


Abstract
CO
2
injection has become the most attractive solution for enhanced oil recovery (EOR). However, this process frequently
suffers from viscous fingering, gravity override, and channeling of CO
2
in heterogeneous formations and the inefficient
displacement of oil in below-miscibility-pressure reservoirs. These challenging issues are closely related to the very limited
ability to control CO
2
mobility. CO
2
foam stabilized with CO
2
soluble surfactants has exhibited more economical and
technical advantages in effective control of CO
2
mobility in porous media than the existing direct methods of CO
2

viscosification. An integrated research framework was developed to strengthen the role of molecular design for these features
as wells as provide a better understanding of foam behavior with complex formation of viscous water/oil or oil/water
emulsions. A simple block model was developed in CMG/STARS to demonstrate the advantages of the novel foam concept
over the conventional foam processes in the field.
We proposed a novel injection strategy which involves dissolving the surfactant in the CO
2
. This method drastically
lowers the injection costs, reduces the loss of surfactant onto the rock surface due to adsorption, and improves in-situ foam
generation to significantly increase oil recovery. Two different novel methods, continuous CO
2
-dissolved-surfactant injection
and water-alternating-gas with CO
2
-dissolved-surfactant injection, have been studied in this paper.
Foam corefloods performed with carbonate core show that CO
2
-dissolved surfactants greatly reduce the mobility of the
injected gas compared to conventional injection strategies. This is consistently observed in the numerically simulated foam
process at the field level.

Introduction
CO
2
flood have been applied worldwide as a routine EOR technology, particularly for reservoir candidates whose pressures
are above minimum miscibility pressure (MMP). However, reservoir heterogeneity and high gas mobility reduce sweep
efficiency and drastically decrease oil recovery (Renkema and Rossen, 2007). For formations with relatively high vertical
permeability and vertical well pattern, the injected CO
2
tends to rise to the top of the reservoir due to its low density and
overrides the bottom oil-rich zone, leading to early gas breakthrough. CO
2
displacement could be improved by water-
alternating gas (WAG) injection. The efficiency of this process relies on the ability to reduce gas relative permeability in the
presence of the aqueous phase and to promote gas trapping in depleted zones that diverts both the injected gas and water into
the oil-rich zones. However, WAG injection has frequently encountered some important issues such as the reduction of CO
2
-
oil contact in the presence of water and low injectivity commonly observed in carbonate reservoirs. Gravity segregation tends
to impair further the advantages of this injection strategy. This has inspired the application of surfactant stabilized CO
2
foam
to improve CO
2
conformance as foam could generate not only a large amount of stable trapped gas but also a significantly
high local pressure gradients induced by aqueous foam films that disperse the gas phase.
CO
2
foam in porous media with surfactant injected in an aqueous phase has been extensively studied (Wang, 1984;
Lescure and Claridge, 1966; Di Julio and Emanuel, 1989; Lee and Heller, 1990; Kuhlman et al., 1992; Alkan et al., 1991;
Dixit et al., 1994; Grigg et al., 2002; Dong and Rossen, 2004; Du and Zitha, 2007). These experimental and theoretical
studies have contributed to the success of several foam field applications. One of the largest full-scale field demonstrations of
foam for gas mobility control is the Foam Assisted Water Alternating Gas (FAWAG) project in the Snorre field on the
Norwegian Continental Shelf from 1997 to 2000. It was estimated that the FAWAG treatment contributed approximately
250000 Sm
3
of oil for every $1 million cost of treatment. More than twenty five successful steam-foam field tests and small
commercial applications have also been reported (Patzek, 1996). Unfortunately, field experiences have shown that
conventional foam with surfactant injected in water has some important limitations. Even though surfactant solution-
alternating-gas injection (SAG) has a strong advantage over co-injection with respect to the strong foam reduced injectivity
2 SPE 113370
(Patzek, 1996), the injected surfactant slugs do not improve CO
2
-oil contact. Gravity override and macroscopic heterogeneity
also challenge the success of surfactant placement into theft zones where the presence of foam is desired. Severe surfactant
adsorption can aggravate this potential issue. This is particularly true in carbonate reservoirs where commonly used anionic
surfactants are known to exhibit strongest adsorption (Lawson, 1978).
A novel foam concept has been developed to mitigate the above shortcomings from the conventional SAG injection. In
this concept, foam could be generated in-situ by injection of surfactant in the CO
2
phase. The injection water may not be
required if the thief zones are sufficiently saturated with water that is available for the foaming process. Carbon dioxide
exhibits very low solubilities for polar and ionic substances and for many (non-polar) compounds with high molecular
weight. This behavior arises from its very low dielectric constant values (Moriyoshi, 1993) and polarisability (3.24 10
4

0
C
m2 V
1
)(Lide, 1995). Consequently, non-polar CO
2
has by far weaker intermolecular van der Waals interactions than
lipophilic hydrocarbon solvents. This means that substances such as many conventional surfactants with appreciable
solubility in latter solvents show very poor solubilities in carbon dioxide. Hoefling et al. (Hoefling et al., 1991) developed the
first effective fluorinated surfactants for application in CO
2
. Subsequently, numerous research efforts have been made in the
last decade to find CO
2
-philic surfactants (Eastoe and Gold, 2005; Dupont et al., 2004; Sagisaka et al., 2004; Lee et al., 2003)
of which non-fluorinated AOT (sodium bis(2-ethylhexyl)-sulfosuccinate) and non-ionic surfactants are the most interesting
(Ihara et al., 1995; Hutton et al., 1999; J. Eastoe et al., 2001; Liu et al., 2001; McFann et al., 1994).
In this work coreflood experiments were conducted with different injection strategies including conventional SAG
(surfactant alternated CO
2
with surfactant injected in water), novel WAGS (water-alternating-gas with surfactant injected in
CO
2
), and novel CO
2
(continuous CO
2
injection with dissolved-surfactant). These foam injection strategies were then scaled
to field where the effects of macroscopic heterogeneity and buoyancy could be better addressed. For this purpose, a simple
double-layer geological model was built in the CMG/STARS foam simulator to demonstrate the soundness of the novel foam
concept.

Experimental description
Materials
Rock. Core plugs were taken from two different sets of carbonate reservoir cores. The plugs were heterogeneous in
porosity some of which contained a large amount of gypsum. CT-Xray scan were performed in conjunction with air
permeability measurement to choose plugs with relatively homogeneous permeability. The properties of the core plugs for
foam flooding are presented in Table 1.
Crude oil. Midland Farm (West Texas) crude oil was used in all experiments. The crude oil is light (36.2 API), low
viscosity (5.2 cp).
Carbon dioxide. Industrial-grade carbon dioxide was purchased from Matheson Gas. CO
2
was delivered in cylinders
with about 800 psig initial pressure. Since all corefloods were conducted with 1500 psig backpressure, injected CO
2
must be
compressed to 1800 psig using an accumulator.
Surfactant and synthetic brine. Surfactant provided by Dow was used for coreflooding with 0.1 wt% and 0.15 wt% as
injected in CO
2
and in a synthetic brine solution, respectively. These surfactant concentrations were chosen to ensure that the
same total amount of surfactant was used in each foam coreflood. The composition of the synthetic brine was 5.02 wt%
sodium chloride (NaCl), 0.58 wt% calcium chloride (CaCl
2
), and 0.23 wt% magnesium chloride in deionized water. This
brine was also used to waterflood the cores before foam flooding.

Apparatuses
The foam coreflood apparatuses are schematically shown in Figure 1. It includes three main components: (a) CO
2
and
aqueous solution injection system, (b) coreholder and pressure transducers, and (c) backpressure and effluent collection
system.
Fluid injection system. A Digital HPCL Dual Pump (model 1500) with maximum working pressure of 5000 psig was
used to directly inject brine or surfactant solution. CO
2
was injected at a constant rate by another Digital HPCL Dual Pump
(model 1500) through a floating-piston accumulator. Two accumulators of Model CFR-50-200 manufactured by the Temco
were used, each of which has a floating piston driven by a displacing fluid (Dow Corning fluid). The pistons made from
PEEK (Polyether-Ether-Ketone) are mechanically strong enough to work under high pressure and resistant to CO
2
diffusion.
Each piston had three grooves with o-rings to separate the Dow Corning fluid and CO
2
. One accumulator was used for pure
CO
2
injection while the other contained CO
2
with dissolved surfactant. The surfactant and CO
2
were mixed together in the
latter accumulator by a mechanical mixer. A Model MM-D06 mixer manufactured by PPI was used.
Core holder and pressure transducers. A Phoenix Hassler-type core holder was mounted in the vertical direction and
fluids were injected from the top to the bottom. Hydraulic oil was used as an overburden fluid, which compressed and sealed
the -inch-thick Viton rubber sleeve to assure the axial flow of the injected fluids. The core holder has two end caps one of
which has an adjustable end plug length to accommodate different core lengths. There were three pressure taps along the core
holder connected to two pressure transducers as shown in Figure 1. The pressure transducers measured the pressure drops
over two sections of the same length (called Section 1 and 2) that constituted the core.
Backpressure regulator (BPR) and effluent collector. A Mity Mite backpressure regulator (BPR) from Grove Valve and
Regulation Company (Oakland, CA) was used to maintain a constant backpressure during coreflood experiments. Two BPRs
SPE 113370 3
were used for the experiment. One was installed at the CO
2
injection line operated at 1800 psi to provide a constant injection
rate to the core. The other one installed at the outlet of the apparatus was operated at 1500 psi. The effluent was collected for
quantifying oil recovery using a fractional collector.

Experimental Procedure
Core preparation. Since high-pressure CO
2
could diffuse through the rubber sleeve into the overburden fluid, the core
was wrapped in aluminum foil and a thin Teflon tube. A hand pump, Enerpac model P-391 (Rex Supply Corporation, Austin,
TX), was used to exert confining pressure around the rubber sleeve. The wrapped core was then placed in the core holder
whose outlet was connected to the vacuum pump, and the inlet was closed. Both axial and radial confining pressures were
maintained to force the Teflon tubing, aluminum foil, and the sleeve into good contact. The core was vacuumed for 10 hours
to make sure there was no air trapped in the core before measuring core porosity and saturating the core with formation brine.
Porosity measurement. The core outlet was then switched from the vacuum pump and connected to a burette that was
filled with a brine solution through a 3-way valve. The brine level in the burette was recorded. The total volume of brine
imbibed into the core could be calculated by the differences in brine levels. To determine the pore volume of the core, the
dead volume from tubing lines must be subtracted from the total volume.
Core Saturation Procedure. After performing porosity measurement, the inlet was connected to the HPLC pump and
brine solution was injected at 2 ml/min into the core while the outlet remained closed. Once the core pressure increased to the
atmospheric pressure, the outlet was opened to the BPR set at 1500 psi throughout all experiments. The brine injection
stopped after 10 injected PV.
Permeability measurement. The core permeability was measured with brine solution after the porosity measurement.
Permeability measurement is based on Darcys law, which can be rearranged into the following form:

q P
k
A L

= (1)
where q is the flow rate; the viscosity of fluid; A the cross-sectional area of the core; k the permeability; P the pressure
drop along the core; and L the length of the core. Normally pressure drops at different flow rates are measured. Then q/A is
plotted versus P/L. A straight line, forced to go through the origin, can be fitted to the data; the slope of the line represents
the permeability of the core. If the data deviate significantly or systematically from the linear trend, there may have been an
experimental artifact in the data.
Oil Saturation. Oil was injected into the core while the outlet was connected to the fractional retriever. This process was
stopped when the effluent fluid contained almost 100% oil. The initial oil volume in the core was calculated by subtracting
the amount of produced oil and the dead volume from the amount of injected oil. A simple mass balance was performed to
determining residual water saturation and oil relative permeability.
Waterflood. The next step was to brine flood the core until an almost 100% water cut. Residual oil saturation and water
relative permeability were calculated using the same procedure as that for the oil saturation step.
Foam injection. Foam corefloods were then performed with three different injection strategies (see detailed injection
strategies in Table 2):
Conventional SAG injection included 2 SAG cycles with a total 1.2 injected PV chased with 0.8 PV of brine
injection. Each SAG cycle included 0.2 PV of surfactant injected in the aqueous solution at 2 ft/day, followed with
0.4 PV of pure CO
2
at 10 ft/day.
Novel WAGS injection used the same procedure as that for the conventional SAG injection, except the surfactant was
injected in the CO
2
phase.
Novel CO
2
injection included only one single fluid stage in which CO
2
with dissolved surfactant was continuously
injected at 10 ft/day for 2 PV.

Simulation description
Reservoir model. A simple geological model developed in STARS represents a rectangular reservoir of 400 ft long, 400
ft wide and 200 ft high. The grid size was refined towards the wellbore region. All sides of the reservoir were bounded by no-
flow boundaries. One vertical injection well and one vertical production well were diagonally placed at two corner blocks of
the reservoir. The wells were perforated over the entire reservoir thickness. Macroscopic heterogeneity of the reservoir was
represented by two layers. The horizontal permeabilities of the top and bottom layers were 300 md and 100 md, respectively.
The same vertical-to-horizontal permeability ratio of 0.25 was used for both layers. Table 3 lists all reservoir properties.
Fluid properties. Flow of four-components (i.e. water, surfactant, dead oil, and CO
2
) and three phases (i.e. aqueous,
oleic, and gaseous) were simulated. Fluid properties are listed in Table 3. The partitioning of surfactant between the free CO
2

and water was assumed to be instantaneous and determined by a specified partitioning coefficient. The value of this
coefficient was based on the work of Lui (Liu et al., 2002.)
4 SPE 113370
Relative permeability curves. The relative permeability curves for gas-liquid and water-oil systems are shown in
Figures 2 and 3. These curves were available in a field-scale foam simulation developed by CMG/STARS. Note that the end
points of the relative permeability curves were experimentally determined from the corefloods.
Foam model. A simple empirical foam model in STARS was used. This model simply rescales the gas relative
permeability for non-foamed gas-liquid flow, using a dimensionless scaling factor. This scaling factor represents foam
strength whose maximum value of one indicates no foam present. It is a function of several parameters that allows describing
different effects on foam strength such as limiting foam strength (fmmob), surfactant concentration (fmsurf and epsurf),
capillary number (fmcap and epcap), oil (fmoil and epoil), etc. For the sake of demonstration, only the first two effects were
taken into account. The values of fmmob, fmsurf, and epsurf were arbitrarily chosen within the typical ranges suggested by
STARS.
Injection strategies. To be consistent with the experiments, the reservoir was water flooded before foam injection. Five
different foam injection strategies were simulated (see details of injection strategies in Table 4):
Water-alternating-gas injection included 2 cycles, each of which started with injection of a brine slug of 0.2 PV at
1000 bbl/day, followed by a CO
2
slug of 0.4 PV injected at 20,000 ft
3
/day under standard conditions. After the 2
cycles, the reservoir was flooded with 0.8 PV of brine at 1000 bbl/day.
Conventional SAG injection was similar to the WAG injection, except that a surfactant was added to the water slug.
Novel WAGS started with a slug of CO
2
with dissolved surfactant. No surfactant was thus injected in the water slug.
All slug sizes and injection rates were the same as those in the previous injection strategies.
Conventional and novel CO
2
injections were a single fluid stage process with a total of 3 injected PV of CO
2
. The
injected CO
2
in the latter case carried the dissolved surfactant.

Results and discussion
Foam corefloods
Conventional SAG injection
Figure 4 shows the oil recovery and the respective pressure drop profiles over Sections 1 and 2 during the conventional SAG
injection. The pressure drops over the two sections were low and steady during placement of the first surfactant slug. They
were almost the same as the pressure drops induced by the waterflood at residual oil saturation. As a result, there was no oil
recovered during this fluid stage. However, once the injection of the first CO
2
slug commenced, the sectional pressure drops
sharply increased first in phase for a short time and then out of phase afterwards. The former characterized by two pressure
peaks (i.e. about 3.5 and 5.8 psi for Section 1 and 2, respectively) at around 0.23 IPV is due to the displacement of a small
water volume in the injection line upstream of the core entrance at a rate five times higher than that of the surfactant slug
injection (see conventional SAG in Table 2). The pressure drops decreased when CO
2
invaded the core. Strong foam was
generated in-situ and propagated towards the core outlet as indicated by the pressure peaks that subsequently occurred at
about 0.3 and 0.4 IPV in Section 1 and 2, respectively (see Figure 4).
A higher pressure drop in Section 2 than in Section 1 also indicated that foam was much stronger as it propagated
downstream. This is consistent with the previous experimental and theoretical studies on transient foam behavior in porous
media (Rossen, 1996; Kovscek et al., 1994; Ettinger and Radke, 1992). Therefore, a sharp increase in oil recovery shown in
Figure 4 could be ascribed to an increase in local capillary numbers induced by the viscous strong foam during CO
2
injection.
It is important to note that the displacing CO
2
is not always completely dispersed into the leading surfactant slug to form a
foam bank. Part of the surfactant slug will be displaced together with oil as a consequence of the increased local capillary
numbers. As a result, foam becomes weaker in a region that has been swept by a strong foam bank due to lack of surfactant
solution. This is responsible for the subsequent pressure drops in Sections 1 and 2 at late time of CO
2
injection (see Figure 4).
It is important to note that the pressure drops leveled off at values above those observed during the first slug of surfactant.
Examining the pressure profile in Section 1 shows that, since the ordinary supercritical CO
2
viscosity is significantly lower
than water viscosity under the experimental conditions, weak foam did not completely die out in this section where most of
surfactant solution had been displaced at the end of the first CO
2
slug. Otherwise, there must have been a sufficiently large
fraction of trapped gas present due to strong foam. Gas trapping during strong foam flow has been intensively studied (Nahid,
1971; Radke and Gillis, 1990; Friedmann et al., 1991; Tang and Kovscek, 2004; Nguyen et al., 2007, Nguyen et al., 2008).
Nguyen et al. developed a novel gas tracer technique that allows visualizing in-situ local foam mobility, and found that
trapped gas saturation could be higher than 60% (Nguyen, 2004; Nguyen et al., 2007, Nguyen et al., 2008).
CO
2
injection was stopped at 0.6 IPV. The following second slug of surfactant led to a uniform sharp decrease in
pressure drop over the entire core as shown in Figure 4. This is due primarily to the injection rate that was reduced by a factor
of 5 for the surfactant slug (see Table 2). However, the sectional pressure drops approach values that are higher than those
observed from the first surfactant slug. This implies a significant reduction of liquid relative permeability in the presence of
foam, particularly in Section 2. Nguyen (Nguyen, 2004; Nguyen et al., 2005, Zitha et al., 2006, Nguyen et al., 2008)
visualized in-situ fluid partitioning in sandstone cores using a CT-Xray scan technique and reported that the surfactant
solution injected after foam flood did not displace foam in a piston-like manner. Rather, it fingered through the foam bank,
resulting in a large fraction of bypassed foam. The liquid finger continuously grew at a rate that was dependent on the
SPE 113370 5
injection rate and foam strength. During this process, more foam was displaced downstream that may help explain the higher
pressure drop in Section 2 than that in Section 1 as shown in Figure 4. It is important to note that foam could be generated
during surfactant solution injection if the ensuing decrease in pressure drop causes a significant gas expansion (Nguyen,
2004, Nguyen et al., 2005). Newly generated foam helps maintain low water relative permeability, and thus improves sweep
efficiency.
Moreover, the local distribution of the injected surfactant solution has a strong influence on foam behavior in the second
SAG cycle. Indeed, the second CO
2
slug injected at the end of the surfactant slug (i.e. after 0.8 IPV) did not result in strong
pressure peaks as observed in the first cycle. This is because the dispersion of the injected CO
2
into the second surfactant slug
is significantly reduced due to the liquid fingering phenomenon discussed above. However, CO
2
mobility in the second cycle
was clearly reduced as indicated by an increase of the pressure drop in Section 2 by a factor of 4 during CO
2
injection in the
second cycle (i.e. see Section 2 pressure profile between 0.8-1.2 IPV). The pressure drop in Section 1 modestly increased due
primarily to the reduced dispersion of CO
2
and the surfactant solution and the increasing foam strength during its
propagation. Even though the overall pressure drop was significantly lower than that in the first cycle, the oil recovery
continuously increased during the second cycle, accounting for almost 50% of the total recovery after the two cycles. The oil
recovery slightly increased and leveled off after CO
2
injection was switched to brine injection (i.e. at 1.2 IPV). The pressure
drop in Section 1 approached that observed during the first surfactant stage while it remained higher in Section 2. This could
be ascribed to foam was weakened and mobilized by the injected brine in Section 1, and that the destabilization of foam by
the brine in Section 2 was hindered by a combined effect of brine fingering and strong foam. Note that only 0.8 PV of brine
was injected.

Novel WAGS injection
Figure 5 shows the pressure drop profiles and oil recovery for the novel WAGS injection. The profiles are very similar in
trend and magnitude to those for the conventional SAG (Figure 4) indicating almost the same foam performance in both
injection strategies. Such similarity can be expected for laboratory scale 1D foam flow where the gas-liquid dispersion is
essentially the same in either strategy. However, there are several factors that may make a significant difference between the
two strategies in the field. In fact, surfactant adsorption may be more severe when it is injected in the aqueous phase than in
CO
2
because water injection rate is normally much lower than gas injection rate, thus apparently increasing surfactant
retention time. Additionally, diffusion of surfactant from the CO
2
phase towards the rock surface is significantly resisted by
the CO
2
-water interfaces in the presence of foam (Nguyen et al., 2005). Gravity segregation may impair the success of
surfactant placement in the conventional SAG. This could be overcome by the novel foam that could be generated wherever
CO
2
with dissolved surfactant mixes with formation water.

Novel CO
2
injection
To maximize the benefit of CO
2
miscible displacement and improve the injectivity, it is desired to eliminate the injected
water in the novel WAG injection. This particular novel CO
2
concept was tested under the same experimental conditions as
in the novel WAGS, except only CO
2
was continuously injected. The striking results are shown in Figure 6. The pressure
drop shown in this figure increased immediately at the onset of CO
2
injection, indicating strong foam propagation in the core
(note that the initial pressure drop was due to waterflood before CO
2
injection). The peak quickly reached a maximum and
gradually decreased, while forming the leading oil bank. The foam became weaker as more resident water was desaturated.
Despite the continuous fall of the pressure drop, oil recovery increased and approached a surprising plateau value of almost
60% OOIP, compared to 51% by the novel WAG injection. This successfully demonstrates the soundness of the novel CO
2

injection concept.

Foam simulation
Comparison of WAG, conventional SAG, and novel WAGS injections
Figure 7 shows the flowing bottom pressure of the injector as a function of time. For WAG injection, the injection pressure
exhibits an alternating wave in harmony with the alternation of gas and water slugs. This could well be explained on basis of
a typically high gas-water mobility ratio. However, such an oscillatory pressure is expected to be less obvious in the presence
of foam during either SAG or WAGS injection because of the reduced gas-water mobility ratio. Examining the pressure
profiles for SAG and WAGS injections in Figure 7 indicates an excellent consistency between these two strategies as well as
qualitatively with the experimental results (see Figures 4 and 5). This is confirmed with the cumulative oil and gas production
as shown in Figure 8. Such an indiscernible difference between the injection strategies on the field scale is likely due to
surfactant adsorption, which was not taken into account in our simulation work. Further, the specified permeability contrast
is prevailing over the gravity effect. In fact, the simulation results (not shown) could illustrate that both surfactant solution
and CO
2
preferentially invade the top high permeability layer where they mix to generate foam and thus self-divert into the
low permeability layer. The capture of practical reservoir characteristics that promote a significant gravity segregation of the
injected gas and water would better demonstrate the advantages of the novel WAGS over the conventional SAG. This
together with surfactant adsorption will be presented in our future work.
6 SPE 113370
An important point to be made in the present work is that foam causes a significant delay of CO
2
breakthrough for
almost 0.5 IPV (see Figure 8), thus improving not only CO
2
utilization but also geological CO
2
storage. This is illustrated in
Figures 10 and 12. During WAG injection a large fraction of residual oil saturation after waterflood is bypassed as the
injected fluids preferentially flow into the top high permeability layer (Figure 9). Particularly, high vertical communication at
the layer boundary promotes gas gravity override. As a consequence, most of the injected gas is displaced by the injected
water as indicated by very low gas saturation after 0.8 IPV of brine (i.e. following two WAG cycles) in Figure 10. Foam
helps correct this undesired injection profile by redistributing CO
2
and water more evenly in both layers as shown in Figures
11-12.

Comparison of the conventional and novel CO
2
injection
Figure 13 shows the injection pressure profiles during CO
2
injection with and without dissolved surfactant. The initial
pressure is the same for both injection strategies and equal to that during waterflood. In the absence of surfactant, one would
expect a decrease in injection pressure due to high gas mobility as discussed above. This is not observed in the novel CO
2

injection where the injection pressure first increases and then gradually decreases due to the in-situ foam propagation. It is
interesting to observe a good agreement between the simulated (Figure 13) and the experimental (Figure 6) pressure trends.
Furthermore, the benefits of the injection of surfactant in CO
2
are well demonstrated in Figure 14. Compared to the multi-
fluid stage process above, Figures 7 and 8 show that the novel CO
2
injection results in a much longer delay in CO
2

breakthrough and a much more pronounced difference in oil recovery between foam and no foam (Figure 14). This could be
explained based on the local fluid distribution as shown in Figures 15-18. The severe gas gravity override reduced vertical
sweep efficiency in the absence of foam, while CO
2
conformance is strongly improved with uniform foam propagation. The
above benefits of the novel CO
2
injection is highly appreciated where the foam improved CO
2
miscible flood is performed
under constraints of limited water source, high water treatment cost, and low injectivity.

Conclusions
Coreflood experiments with oil recovery by the novel foam concept have been reported for the first time in this work. WAGS
and the conventional SAG injection resulted in almost the same ultimate oil recovery, while the CO
2
injection with dissolved
surfactant gave higher recovery without injected water. This novel injection strategy required relatively lower injection
pressure which is important in the improvement of well injectivity. The pressure gradients due to foam were relatively high
as normally expected on the laboratory scale. This leads to our future experimental work including the determination of
critical conditions for in-situ foam generation with respect to surfactant adsorption and partitioning, scaling behavior of the
novel foam, and optimization of injection strategies.
The simulation results are qualitatively in good agreement with the experiments. The combined effects of high
permeability contrast and high vertical communication between the layers reduce the distinction between WAGS and the
conventional SAG injection, but strengthen the advantages of the CO
2
injection with dissolved surfactant. The foam model
developed in STARS is just a heuristic model for demonstration of foam values. Its capability to scale up foam process to
field has not been proven. A more mechanistic model for foam is needed, taking into account the kinetics of surfactant
partitioning between CO
2
and aqueous phases.

Acknowledgement
The authors would like to thank Jeff Simmons from Occidental Petroleum Cooperation and Dow Chemical Company for
supporting this work. We would also like to thank Dr. Gary A. Pope (The University of Texas at Austin) for his valuable
advice on the field-scale foam process design, and Dr. Keith Johnston (The University of Texas at Austin) for his useful
discussions on surfactants.


References
Alkan, H., Goktekin, A. and Satman, A.: A Laboratory Study of CO
2
-Foam Process for Bati Raman Field, Turkey, paper SPE 21409
presented at the SPE Middle East Oil Show, Bahrain, November 16-19, 1991.
Dixit, A., Tsau J-S. and Heller, J.P.: Laboratory Study on Surfactant-Based Selective Mobility Control, paper SPE 27729 presented at the
1994 SPE Permian Basin Oil & Gas Recovery Conference, Midland, TX, March 16-18, 1994.
Di Julio, S.S. and Emanuel, A.S.: Laboratory Study of Foaming Surfactant for CO
2
Mobility Control, SPE Reservoir Engineering, Vol.
390, 136-142, May 1989.
Du, D. and Zitha, P.L.J.: Carbon Dioxide Foam Rheology in Porous Media: A CT Scan Study, SPE Journal, 245-252, June 2007.
Dupont, A., Eastoe, J., Murray, M., Martin, L., Guittard, F., Givenchy, E.T. and Heenan, R.K.: Hybrid Fluorocarbon-Hydrocarbon CO
2
-
philic Surfactants. 1. Synthesis and Properties of Aqueous Solutions, Langmuir 20, 9953-9959, 2004.
Eastoe, J. and Gold, S.: Self-Assembly in Green Solvents, Phys. Chem. Chem. Phys. 7, 1352-1362, 2005.
Eastoe, J., Paul, A., Nave, S., Steytler, D.C., Robinson B.H., Rumsey, E., Thorpe, M. and Heenan, R.K.: Micellization of Hydrocarbon
Surfactants in Supercritical Carbon Dioxide, J. Am. Chem. Soc. Vol. 123, 988-989, 2001.
Ettinger, R.A. and Radke, C.J.: Influence of Texture on Steady Foam Flow in Berea Sandstone, SPE Reservoir Engineering, Vol. 83,
1992.
SPE 113370 7
Friedmann, F., Chen W.H. and Gauglitz, P.A.: Experimental and Simulation Study of High-Temperature Foam Displacement in porous
Media, SPE Reservoir Engineering, Vol. 37, 1991.
Grigg, R.B., Tsau, J-S and Martin, D.F.: Cost Reduction and Injectivity Improvements for CO
2
Foams for Mobility Control, paper SPE
75178 presented at the SPE/DOE Improved Oil Recovery Symposium, Tulsa, OK, April 13-17, 2002.
Hoefling, T.A., Enick, R.M. and Beckman, E.J.: Microemulsions in Near-Critical and Supercritical CO
2
, J. Phys. Chem., Vol. 95, 7127-
7129, 1991.
Hutton, B.H., Perera, J.M., Grieser, F. and Stevens, G.W.: Investigation of AOT Reverse Microemulsions in Supercritical Carbon
Dioxide, Colloids Surfactant, Vol. 146, 227-241, 1999.
Ihara, T., Suzuki, N., Maeda, T., Sagaram, K. and Hobo, T.: Extraction of Water-Soluble Vitamins from Pharmaceutical Preparations
Using AOT (Sodium Di-2-ethylhexyl Sulfosuccinate)/Pentane Reversed Micelles," Chem. Pharm. Bull., Vol. 43, 626-630, 1995.
Kim, J. S., Dong, Y and Rossen, W.R.: Steady-State Flow Behavior of CO
2
Foam, paper SPE 89351 presented at the 2004 SPE/DOE
Fourteenth Symposium on Improved Oil Recovery, Tulsa, OK, April 17-21, 2004.
Kovscek, A.R., Patzet, T.W. and Radke, C.J.: Mechanistic Prediction of Foam Displacement in Multidimensions: A Population-Balance
Approach, SPE Journal, Vol. 71, 1994.
Kuhlman, M.I., Falls, A.H., Hara, S.K., Monger-McClure, T.G. and Borchardt, J.: CO
2
Foam with Surfactants Used Below Their Critical
Micelle Concentrations, SPE Reservoir Engineering, 445-452, November 1992.
Lawson, J.B.: The Adsorption of Non-Ionic and Anionic Surfactants on Sandstone and Carbonate, 5
th
Symp. SPE IOR, Tulsa, OK, 1978.
Lee, C.T., Ryoo, W., Smith, P.G., Arellano, J., Mitchell, D.R., Lagow, R.J., Webber, S.E. and Johnston, K.P.: Carbon Dioxide-in-Water
Microemulsions, J. Am. Chem. Soc. Vol. 125, 3181-3189, 2003.
Lee, H.O. and Heller, J.P.: Laboratory Measurement of CO
2
-Foam Mobility, SPE Reservoir Engineering, 193-197, May 1990.
Lescure, B. M. and Claridge, E. L.: CO
2
Foam Flooding Performance vs. Rock Wettability, paper SPE 15445 presented at the 61st
Annual Technical Conference and Exhibition of the Society of Petroleum Engineers, New Orleans, LA, October 5-8, 1966.
Lide D.R. (Editor), CRC Handbook of Chemistry and Physics 76th Edition 1995-1996, CRC Press, Boca Raton, 1995.
Liu, J., Han, B., Wang, Z., Zhang, J., Li, G. and Yang,G.: Solubility of Ls-36 and Ls-45 Surfactants in Supercritical CO
2
and Loading
Water in the CO
2
/Water/Surfactant Systems Langmuir, Vol. 18, 3086-3089, 2002.
Liu, J., Han, B., Li, G., Zhang, X., He, J. and Liu, Z.: Investigation of Nonionic Surfactant Dynol-604 based Reverse Microemulsions
Formed in Supercritical Carbon Dioxide, Langmuir, Vol. 17, 8040-8043, 2001.
Liu, J., Han, B., Zhang, J., Li, G., Zhang, X., Wang, J. and Dong, B.: Formation of Water-in-CO
2
Microemulsions with Non-Fluorous
Surfactant Ls-54 and Solubilization of Biomacromolecules, Chem. Eur. J., Vol. 8, 1356-1360, 2002.
Liu, J., Han, B., Wang, Z., Zhang, J., Li, G. and Yang, G.: Solubility of Ls-36 and Ls-45 Surfactants in Supercritical CO
2
and Loading
Water in the CO
2
/Water/Surfactant Systems, Langmuir, Vol. 18, 3086-3089, 2002.
McFann, G.J., Johnston, K.P. and Howdle, S.M.: Solubilization in Nonionic Reverse Micelles in Carbon Dioxide, AIChE J. Vol. 40,
543-555, 1994.
Moriyoshi, T., Kita, T. and Uosaki, Y.: Static Relative Permittivity of Carbon Dioxide and Nitrous Oxide up to 30MPa, Ber. Bunsenges.
Phys. Chem., Vol. 97, 589-596, 1993.
Nahid, B.H.: Non-Darcy Flow of Gas through Porous Media in the Present of Surface Active Agents, PhD Dissertation 1971, University
of Southern California.
Nguyen, Q.P., Zitha, P.L.J., Currie, P.K. and Rossen, W.: Determination of Gas Trapping With Foam Using X-Ray CT and Effluent
Analysis, SPE Journal, in press.
Nguyen, Q.P., Currie, P.K., Buijse, M. and Zitha, P.L.J.: Mapping of Foam Mobility in Porous Media, Journal of Petroleum Science and
Engineering, Vol. 58, 119132, 2007.
Nguyen, Q.P., PhD dissertation, Delft University of Technology, 2003.
Nguyen, Q.P., Zitha, P.L.J. and Currie, P.K.: Effect of Capillary Cross-Flow on Foam-Induced Diversion in Layered Formations, SPE
Journal 2005.
Nguyen, Q.P., Zitha, P.L.J., Currie, P.K. and Rossen, W.: CT Study of Liquid Diversion with Foam, SPE Production & Operations, in
press.
Patzek, T.W.: Field Applications of Foam for Mobility Improvement and Profile Control, SPE Reservoir Eng. J., Vol. 79, 1996.
Radke, C.J. and Gillis, J.V.: A Dual Tracer Technique for Determining Trapped Gas Saturation during Steady Foam Flow in Porous
Media, paper SPE 20519 presented at the 65th Annual Technical Conference and Exhibition, New Orleans, 1990.
Renkema, W.J. and Rossen, W.R.: Success of SAG Foam Processes in Heterogeneous Reservoirs, paper SPE 110408 presented at the
2007 SPE Annual Technical Conference and Exhibition, Anaheim, CA, November 11-12, 2007.
Rossen, W.R.: Foams in Enhanced Oil Recovery, in Foams: Theory, Measurements, and Application, R.K. Prudhomme and S.A. Khan
(eds.), Marcel Dekker, Vol. 413, 1996.
Sagisaka, M., Fujii, T., Ozaki, Y., Yoda, S., Takebayashi, Y., Kondo, Y., Yoshino, N., Sakai, H., Abe, M. and Otake, K.: Interfacial
Properties of Branch-Tailed Fluorinated Surfactants Yielding a Water/Supercritical CO2 Microemulsion, Langmuir, Vol. 20, 2560-
2566, 2004.
Tang, G.Q. and Kovscek, A.R.: Measurement and Theory of Gas Trapping in Porous Media During Steady-State Foam Flow, paper SPE
91042 presented at the SPE Annual Technical Conference and Exhibition, Houston, TX, 2004.
Wang, G.C.: A Laboratory Study of CO
2
Foam Properties and Displacement Mechanism, paper SPE 12645 presented at the SPE/DOE
Fourth Symposium on Enhanced Oil Recovery, Tulsa, OK, April 15-18, 1984.
Zitha, P.L.J., Nguyen, Q.P. and Currie, P.K.: Coupling of Foam Drainage and Viscous Fingering in Porous Media Vevealed by X-ray
Computed Tomography, Transport in Porous Media, Vol. 64, 301313, 2006.




8 SPE 113370

Table 1. Core properties
Rock Type
Core Dimension
(length - diameter, cm)
Average Permeability
to Brine
(md)
Porosity
(%)
C1 (SAG)
C2 (CO
2
)
25.4 - 4.6
24.8 4.6
152
141
24
23



Table 2. Experimental foam injection strategies
1
st
and 2
nd
slug of liquid 1
st
and 2
nd
slug of gas Chase brine
Foam Injection after
Waterflood
Fluid Size Rate Fluid Size Rate Size Rate
Conventional SAG
Aqueous surfactant
solution
0.2 PV
2
ft/day
Pure CO
2
0.4 PV
10
ft/day
0.8 PV 2 ft/day
Novel WAGS Formation brine 0.2 PV
2
ft/day
CO
2
with dissolved
surfactant
0.4 PV
10
ft/day
0.8 PV 2 ft/day



Table 3. Fluid and Reservoir properties used in foam simulation
Fluid Properties
Properties Water Surfactant Dead Oil CO
2

CMM (lb/lbmole) 18.015 230 280 44
P
c
(Psi) 3198 20.885 489.36 1069.3
T
C
(F) 705.47 980.6 385.61 87.89
Reservoir Properties
Size of reservoir 448x408x108 [ft]
Number of grid blocks 34*34*30 [-]
Absolute permeability of top layer
Absolute permeability of bottom layer
300
100
[mD]
[mD]
Porosity 0.24 [-]
Reservoir temperature 80 [F]
Initial reservoir pressure 2500 [psi]
Constant pressure at production well 1500 [psi]
Maximum allowable injection pressure 3500 [psi]



Table 4. Simulation injection strategies
1
st
and 2
nd
slug of liquid 1
st
and 2
nd
slug of gas
Brine Injection
after 2 cycles
Foam Injection after
Waterflood
Solution Size Rate Solution Size Rate Size Rate
WAG Formation brine 0.2 PV
1000
bbl/day
Pure CO
2
0.4 PV
20000
ft3/day
0.8 PV
1000
bbl/day
Conventional SAG
Aqueous surfactant
solution
0.2 PV
1000
bbl/day
Pure CO
2
0.4 PV
20000
ft3/day
0.8 PV
1000
bbl/day
Novel WAGS Formation brine 0.2 PV
1000
bbl/day
CO
2
with dissolved
surfactant
0.4 PV
20000
ft3/day
0.8 PV
1000
bbl/day





SPE 113370 9









0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 0.2 0.4 0.6 0.8 1
Water Saturation (Sw)
R
e
l
a
t
i
v
e

p
e
r
m
e
a
b
i
l
i
t
y
Krw Krow


0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 0.2 0.4 0.6 0.8 1
Liquid Saturation (Sl)
R
e
l
a
t
i
v
e

P
e
r
m
e
a
b
i
l
i
t
y
Krg Krog
Fig. 2Water-oil relative permeability curves Fig. 3Liquid-gas relative permeability curves
Fig. 1Schematic of the experimental foam coreflood setup
10 SPE 113370



0
4
8
12
16
20
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Injected PV
P
r
e
s
s
u
r
e

D
r
o
p

(
p
s
i
)
0%
12%
24%
36%
48%
60%
%
O
O
I
P
Sec 1 Sec 2 Cumulative Oil







0
4
8
12
16
20
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Injected PV
P
r
e
s
s
u
r
e

D
r
o
p

(
p
s
i
)
0%
12%
24%
36%
48%
60%
%
O
O
I
P
Sec 1 Sec 2 Cumulative Oil




Fig. 5Pressure drop and oil recovery profiles from the novel WAGS experiment
Fig. 4Pressure drop and oil recovery profiles from the conventional SAG experiment
SPE 113370 11



0
4
8
12
16
20
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Injected PV
P
r
e
s
s
u
r
e

D
r
o
p

(
p
s
i
)
0%
12%
24%
36%
48%
60%
%
O
O
I
P
Sec1 Sec 2 Cumulative Oil





1500
1650
1800
1950
2100
2250
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Injected PV
B
o
t
t
o
m

H
o
l
e

P
r
e
s
s
u
r
e

(
p
s
i
)
WAG Conventional SAG Novel WAGS



Fig. 7Bottomhole pressure in the injection well during WAG, conventional SAG, and
novel WAGS injections
Fig. 6Pressure drop and oil recovery profiles from the novel CO
2
experiment
12 SPE 113370



0%
5%
10%
15%
20%
25%
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Injected PV
C
u
m
u
l
a
t
i
v
e

O
i
l

(
%
O
O
I
P
)
0.E+00
1.E+06
2.E+06
4.E+06
5.E+06
6.E+06
C
u
m
u
l
a
t
i
v
e

G
a
s

(
f
t
3
)
WAG_CumOil
CONSAG_CumOil
NovelWAGS_CumOil
WAG_CumGas
ConSAG_CumGas
NovelWAGS_CumGas















Fig.8Cumulative oil and gas production during WAG, conventional SAG, and novel WAGS injections
Fig. 9Oil saturation in the reservoir after total 2 IPV
including 2 cycles of WAG injection followed with brine
injection. Only half of the reservoir is shown, and both
injector and producer are on the vertical cross section
with the latter on the right.
Fig. 10Gas saturation in the reservoir after total 2
IPV including 2 cycles of WAG injection followed with
brine injection. Only half of the reservoir is shown,
and both injector and producer are on the vertical
cross section with the latter on the right.
SPE 113370 13












1500
1650
1800
1950
2100
2250
0.00 0.50 1.00 1.50 2.00 2.50 3.00
Injected PV
B
o
t
t
o
m

H
o
l
e

P
r
e
s
s
u
r
e

(
p
s
i
)
Conventional CO2 Novel CO2







Fig. 13Bottom hole pressure in the injection well during the conventional CO
2

and novel CO
2
injection
Fig. 11Oil saturation in the reservoir after total 2 IPV
including 2 cycles of novel WAGS injection followed with
brine injection. Only half of the reservoir is shown, and
both injector and producer are on the vertical cross
section with the latter on the right.
Fig. 12Gas saturation in the reservoir after total 2 IPV
including 2 cycles of novel WAGS injection followed with
brine injection. Only half of the reservoir is shown, and
both injector and producer are on the vertical cross
section with the latter on the right.
14 SPE 113370

0%
4%
8%
12%
16%
20%
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Injected PV
C
u
m
u
l
a
t
i
v
e

O
i
l

(
%
O
O
I
P
)
0.0E+00
4.0E+06
8.0E+06
1.2E+07
1.6E+07
2.0E+07
C
u
m
u
l
a
t
i
v
e

G
a
s

(
f
t
3
)
ConCO2_CumOil NovelCO2_CumOil
ConCO2_CumGas NovelCO2_CumGas













Fig. 14Cumulative oil and gas production during conventional CO
2
and novel CO
2
injection
Fig. 15Oil saturation in the reservoir after total 3 IPV
of CO
2
without surfactant. Only half of the reservoir is
shown, and both injector and producer are on the
vertical cross section with the latter on the right.
Fig. 16Oil saturation in the reservoir after total 3 IPV of
CO
2
without surfactant. Only half of the reservoir is
shown, and both injector and producer are on the vertical
cross section with the latter on the right.
SPE 113370 15














Fig. 17Oil saturation in the reservoir after total 3 IPV of
CO
2
with dissolved surfactant. Only half of the reservoir
is shown, and both injector and producer are on the
vertical cross section with the latter on the right.
Fig. 18Oil saturation in the reservoir after total 3 IPV of
CO
2
with dissolved surfactant. Only half of the reservoir
is shown, and both injector and producer are on the
vertical cross section with the latter on the right.