Thickening Carbon Dioxide With the
Fluoroacrylate-Styrene Copolymer
Jianhang Xu, Aaron Wlaschin, and Robert M. Enick, SPE, U. of Pittsburgh
Summary
The fluoroacrylate-styrene copolymer is the first associative thick-
ener that has been identified for carbon dioxide. Fluoroacrylate is
highly carbon dioxide-philic, enhancing the solubility of the co-
polymer in carbon dioxide. Styrene is relatively carbon dioxide-
phobic, but promotes viscosity-enhancing, intermolecular associa-
tions. The copolymer used in this study had a composition of 29
mol% styrene—71 mol% fluoroacrylate, a number-average mo-
lecular weight of 540,000, and a polydispersity index of 1.63. The
copolymer was sufficiently soluble in CO2 at typical reservoir
flooding conditions to induce a significant increase in viscosity.
Falling cylinder viscometry measurements at 298 K demonstrated
that the viscosity enhancement of CO2 increased with increasing
copolymer concentration, decreasing shear rate, and decreasing
temperature. Mobility measurements of the fluoroacrylate-styrene
copolymer-CO2 solutions flowing through 80 to 200 md Berea
sandstone cores at superficial velocities of 0.00035 to 0.028 cm/s
(1–80 ft/D) at 298 K also indicated that the copolymer was an
effective thickener. At a superficial velocity of 0.00035 m/s (1ft/
D), the addition of 1.5 wt% copolymer increased the CO2 viscosity
by a factor of 19 relative to neat CO2. Smaller increases in vis-
cosity occurred at lower copolymer concentrations and higher ve-
locities. These results demonstrate that it is possible to design a
compound that can enhance the viscosity of dense carbon dioxide.
Fluoropolymers are characterized by environmental persistence,
high cost, and unavailability in bulk quantities, however. Therefore
nonfluorous, inexpensive thickeners are currently being designed.
Introduction
The mobility ratio of carbon dioxide floods is unfavorable because
the viscosity of CO2 is low relative to the oil being displaced. A
more favorable mobility ratio would result in an increased amount
of oil recovery prior to CO2 breakthrough and increased rate of oil
recovery thereafter. Water-alternating-gas injection is typically
used to improve the mobility ratio by decreasing the relative per-
meability of the carbon dioxide. Alternatively, the mobility ratio
could be improved if the viscosity of carbon dioxide was in-
creased. The use of “thickened” CO2 would also eliminate the need
for the co-injection of water.
The identification of a compound capable of dramatically in-
creasing the viscosity of carbon dioxide when present in dilute
solution has been an elusive goal. The primary obstacle that has
hindered the efforts of numerous investigators has been the ex-
tremely low CO2-solubility of candidate thickeners.1 Most of the
pioneering efforts in this area were conducted by Heller and co-
workers1–6 in their evaluation of tri-alkyl tin fluorides, polymers,
telechelic ionomers, and 12-hydroxystearic acid. Significant con-
tributions were also made by Irani and coworkers, who success-
fully thickened CO2 using silicone-based polymers that required
the addition of an organic solvent.7 Irani and coworkers later pro-
posed that polymers that were designed to exhibit low solubility
parameters may be effective CO2 thickeners in the presence of
small amounts of an organic liquid cosolvent.8,9
Copyright © 2003 Society of Petroleum Engineers
This paper (SPE 84949) was revised for publication from paper SPE 71497, first presented
at the 2001 SPE Annual Technical Conference and Exhibition, New Orleans, 30 Septem-
ber–3 October. Original manuscript received for review 25 January 2002. Revised manu-
script received 22 January 2003. Manuscript peer approved 28 January 2003.
June 2003 SPE Journal
DeSimone and coworkers were the first to demonstrate that a
high molecular weight polymer, poly(1,1-dihydro-perfluorooctyl-
acrylate), PFOA, could exhibit remarkably high CO2 solubility in
the absence of a cosolvent.1,10 Falling cylinder viscometry was
used to demonstrate that PFOA could also increase the viscosity of
CO2. For example, at 323 K and 34 MPa, the addition of 6.7wt%
PFOA to CO2 resulted in an increase of viscosity from 0.2 mPas
to 0.6 mPas. Although the level of viscosity enhancement was
promising, the shear rate of the measurement was not provided,
and the concentration of PFOA was high.1,10
Enick and coworkers designed, synthesized, and evaluated sev-
eral CO2-soluble thickeners.11–13 Many of these thickeners were
similar in structure to oil thickeners, but were modified to incor-
porate fluorocarbon chains (rather than hydrocarbon chains) in
their structure. The fluoroalkyl, fluoroether, and fluoracrylate
functional groups in these compounds imparted CO2-philicity to
the candidate thickening agent, thereby facilitating its dissolution
in dense CO2. The polar, ionic, hydrogen-bonding, or carbon di-
oxide-insoluble groups in the thickener enhanced intermolecular
associations with similar groups in neighboring molecules. These
intermolecular associations led to the formation of macromolecu-
lar structures in solution that were capable of inducing viscosity
increases or gelation when present in dilute concentration. Thick-
ening candidates included semifluorinated trialkyltin fluorides,11
telechelic ionomers,11 small fluorinated hydrogen-bonding com-
pounds, 12 fluoroacrylate homopolymers, 13 fluoroacrylate-
hydrocarbon copolymers,13 and partially sulfonated fluoroacry-
late-styrene copolymers.13 Although each of these compounds ex-
hibited CO2 solubility of several weight percent, most induced
modest viscosity increases. Modest increases in carbon dioxide
viscosity at 298 K were detected by falling cylinder viscometry for
most of these compounds. The poly(fluoroacrylate-styrene) co-
polymer (Fig. 1) and its slightly sulfonated analogs induced tre-
mendous increases in viscosity when present in dilute concentra-
tion, however.13 It was postulated that the fluoroacrylate compo-
nent induced high CO2 solubility while the pendent phenyl groups
associated with the styrene interacted with the phenyl group of a
neighboring copolymer molecule via ␲-␲ stacking.14 Stacking of
aromatic rings is attributed to delocalization of electrons on the
carbon atoms of the ring and slight residual positive charge on the
hydrogen atoms. The hydrogen atoms are attracted to the more
electron-rich carbon atoms to give a T-shaped arrangement (the
aromatic rings stack normally to one another, forming a “T” shape
when viewed on edge).14 This hypothesis was supported by the
observation that many fluoroacrylate-based copolymers dissolved
in CO2, but only the copolymer with the pendent aromatic ring
thickened the CO2.13 Therefore, the thickening of the CO2 was
achieved by the intermolecular associations of the dissolved co-
polymer molecules, which established an associative, viscosity-
enhancing, macromolecular network to form in dense carbon di-
oxide. The optimal composition of the copolymer was 71 mol%
fluoroacrylate and 29 mol% styrene. Compositions with greater
amounts of styrene were more difficult to dissolve and yielded
smaller viscosity increases. This was probably caused by an in-
creased number of intramolecular interactions of the aromatic
groups that prevented the polymer coil from expanding and asso-
ciating with other copolymer molecules.13 Copolymer composi-
tions with less than 29% styrene are also effective because they are
even more soluble in CO2, but they are somewhat less effective as
thickeners and require higher concentrations in CO2 to attain the
desired level of viscosity enhancement.13
85
 Fig. 1—Synthesis of the fluoroacrylate-styrene random copolymer.
Objective
The use of an associative thickener to reduce the mobility of car-
bon dioxide flowing through porous media has not previously been
determined. Therefore, the objective of the current work was to
determine whether the fluoroacrylate-styrene copolymer would be
sufficiently soluble in CO2 in the 298 to 373 K temperature range
to decrease the mobility of carbon dioxide flowing through a sand-
stone core. Viscosity was initially estimated with a falling cylinder
viscometer, and aluminum cylinders of varying diameter were
used to define the non-Newtonian characteristics of the thickened
CO2 over a wide range of shear rates. The mobility of thickened
CO2 flowing through Berea sandstone cores at superficial veloci-
ties of 0.00035 to 0.028 cm/s (1–80 ft/D) was then determined.
Experimental
Copolymer Synthesis. Styrene [Aldrich, 99+%] was purified un-
der vacuum distillation. 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-
heptadecafluorodecyl acrylate [Aldrich, 97%] was used after the
inhibitor, 100 ppm monomethyl ether hydroquinone, was removed
by adding the monomer dropwise to an inhibitor removal column.
Bulk, free radical polymerization, with AIBN as initiator, was used
to obtain a random copolymer product (Fig. 1).
In a typical synthesis of the 71% fluoroacrylate, 29% styrene
copolymer, 0.82 g (0.007873 mol) styrene, 10.00 g (0.01929 mol)
fluoroacrylate, and 8.93×10−3 g (5.44×10−5 mol) of AIBN were
added to a 50-ml glass ampule under an inert N2 atmosphere. The
ampule was sealed and placed in an oil bath at 338 K for 12 hours
to promote polymerization. The reaction products were then dis-
solved in 100 g 1,1,2-trichlorotrifluoroethane. The copolymer was
then precipitated by washing with 300 g methanol, while the un-
reacted monomer remained in solution. The copolymer product, a
white solid, was recovered by filtration. The copolymer was
washed two additional times in a similar manner and dried under
vacuum overnight. The structure of the copolymer (Fig. 1) was
confirmed using standard IR and NMR techniques.13 The yield
was 85%. This copolymer was a random copolymer of the two
monomers with an overall proportion of styrene in the copolymer
of 29%. This copolymer was neither a block copolymer (a polymer
that consists of a polyfluoroacrylate section joined to a polystyrene
section) nor a physical mixture of polystyrene resin and polyfluo-
roacrylate resin.
Solubility Apparatus. The phase behavior studies of CO2-
copolymer mixtures at 298 to 373 K were performed using the
apparatus illustrated in Fig. 2. Phase behavior measurements were
conducted in a high-pressure, variable-volume, windowed cell
(D.B. Robinson Cell), rated to 69 MPa and 478 K. A thick-walled
hollow quartz cylinder was inserted in the high-pressure windowed
cell. A floating piston in the hollow cylinder isolated the sample
volume in the cell from the overburden fluid below the piston and
surrounding the tube (Fig. 3). Standard, nonsampling techniques
86
were employed to determine copolymer solubility.11–13 Specified
amounts of the copolymer and CO2 were introduced to the sample
volume. The mixture was then heated to the specified temperature
and compressed to 45 MPa. Mixing was achieved by rocking the
cell until a single, transparent phase was obtained. The sample
volume was then slowly expanded by withdrawing the overburden
fluid at a constant rate until the copolymer began to precipitate.
This pressure corresponded to the cloud-point pressure of the mix-
ture of known overall composition. This procedure was repeated
for copolymer concentrations up to 5 wt%.
Falling Cylinder Viscometry. The simplest test for estimating
high-pressure fluid viscosity is falling object viscometry.2,11,13,15
Single-phase solutions of the copolymer in CO2 were prepared at
298 K and 34 MPa. An aluminum cylinder was placed within the
sample volume of hollow quartz tube (Fig. 3). Rapid inversion of
the high-pressure cell enabled the fall of the aluminum cylinder
through the thickened CO2 to be observed (Fig. 3). Cylinders with
an outer diameter slightly less than the inside diameter of the
quartz tube were used (Table 1), thereby reducing the terminal
velocity of the falling cylinder and minimizing the acceleration
length required for the cylinder to attain its terminal velocity. The
pressure equilibration across the walls of the quartz tube elimi-
nated the possibility of any change in the dimensions of the tube
under high-pressure conditions.
The viscosity and shear rate can be determined for Newtonian
fluids given the geometry of the system, the density of the fluid
and falling cylinder, and the terminal velocity.15 Multiple cylin-
ders must be used to characterize non-Newtonian fluids.16 The
value of the calculated viscometer constant differed from the value
obtained by calibration with neat carbon dioxide by 0 to 50%. The
magnitude of the error increased as the diameter of the cylinder de-
creased and the terminal velocity of the falling cylinder increased.
The aluminum cylinder was introduced to the cell prior to the
addition of the copolymer and the carbon dioxide. The pressure
was then increased above the cloud point, yielding a single, trans-
parent phase within the sample volume. The entire high-pressure
cell was then rapidly inverted (Fig. 3), causing the aluminum cyl-
inder to fall through the neat or thickened carbon dioxide. The time
required for the cylinder to fall a specified distance, 2 to 10 cm,
was then recorded. Each measurement was repeated five times.
These measurements indicated that the cylinder attained its termi-
nal velocity within several millimeters.
The falling cylinder results provided a convenient measure of
the change in the fluid viscosity. During the initial screening of
thickeners, the solution was assumed to be Newtonian.
dVz
␶⳱−␮ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
dr
Consider the aluminum cylinder falling through a dilute poly-
mer solution that is considered to be a Newtonian fluid. The gov-
June 2003 SPE Journal
 Fig. 2—Apparatus for phase behavior, falling cylinder viscometry, and mobility measurements.

erning expression for a falling cylinder viscometer under these
conditions is provided in Eq. 2.
共␳c − ␳f兲
␮=C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)
Vc
The viscometer constant, C, can be theoretically determined
from the viscometer geometry15,16 or experimentally determined
by calibrating the viscometer with neat CO2. The disparity be-
tween these values increased with decreasing cylinder diameter
and increasing gap width (Table 1). The narrower cylinders were
also prone to wobbling and rotation during their fall. Therefore it
should be emphasized that the primary utility of falling cylinder
when comparing neat CO2 to thickened CO2. Under these constraints
for a Newtonian fluid, the ratio of the viscosity of thickened CO2 to
the viscosity of neat CO2 was equal to the ratio of the terminal
velocity in neat CO2 to the terminal velocity in thickened CO2.
␮copolymer solution Vc,CO2
= . . . . . . . . . . . . . . . . . . . . . . . . . (3)
␮CO2 Vc,copolymer solution
The viscosity of the CO2-copolymer solutions was also mod-
eled using a power law expression, providing a more accurate
representation of the non-Newtonian nature of the solution. (Note
that if non-Newtonian fluid approaches the Newtonian behavior, n
approaches unity as m approaches ␮.)

冏冏
viscometry is the rapid detection of significant increases in vis- n−1
cosity. The density of CO2 was estimated with a precise equation dVz dVz dVz
of state derived solely for carbon dioxide.17 The density of the ␶ = −m = −␩ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (4)
dr dr dr
thickened carbon dioxide was assumed to be comparable to the
density of neat carbon dioxide because the concentration of the The relative increase in the viscosity of carbon dioxide exhib-
copolymer was relatively low, and the same cylinder was used ited by the non-Newtonian copolymer solution was defined as the
following ratio18:
␩copolymer solution
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
␮CO2

Flow Through Porous Media. The most accurate and relevant

viscosity measurements were obtained by displacing the high-
pressure fluid through sandstone. A core holder rated to 100 MPa

Fig. 3—Falling cylinder viscometry.

June 2003 SPE Journal 87

(Temco, Tulsa) was used to retain the 15.24-cm long, 2.54-cm
diameter, clean, homogeneous Berea sandstone core [80 to 200
md, American Stone, ∼13% porosity]. The core sample was held
within a rubber sleeve, which had two pressure taps that permitted
pressure measurements to be made 5.08 and 10.16 cm from the
inlet face of the core. Neat CO2 and thickened CO2 were prepared
in the Robinson cell at 298 K and 20 MPa to ensure that a single
phase of copolymer-CO2 solution at 298 K was attained. The radial
overburden pressure of silicone oil was maintained at 23.5 MPa
with a positive displacement pump (Ruska, Houston). A dual posi-
tive-displacement proportioning pump (Ruska) was used to ensure
that the high-pressure fluid (CO2 or thickened CO2) was displaced
into the core at the same rate that the volume for the effluent fluid
from the core was expanded. The use of a dual-proportioning
pump eliminated the need for a backpressure regulator. The entire
system was charged to the desired pressure, and the high-pressure
fluid was then displaced from the Robinson cell into the sandstone
face at the same volumetric rate that the effluent solution was
being withdrawn from the porous media. Therefore the experi-
ments were easy to control, monitor, and reproduce, despite the
very small pressure gradients across the high-pressure core. High-
total-pressure, low-pressure-drop differential pressure transducers
(Validyne Engineering Co., DP303 transduce, Northridge, Califor-
nia) were used to monitor the pressure drop along each third of the
core. Equivalent pressure drops along each section indicated that
the copolymer was flowing uniformly through the core rather than
being retained at the inlet face.
Each core was initially saturated with neat carbon dioxide.
Carbon dioxide was then displaced through the core, followed by
80 to 90 cm3 of thickened carbon dioxide. Finally, neat CO2 was
again introduced to the core to determine if any permeability re-
duction had occurred. Fluid viscosity was determined by rearrang-
ing Darcy’s Law to solve for viscosity.
⌬P K
␮= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (6)
L v
The pressure drop measurement associated with the flow of
neat carbon dioxide was used to determine the core permeability.
Using this permeability value, the pressure drop associated with
the flow of the thickened carbon dioxide was used to determine its
viscosity. Because the same core was used for the experiments
with neat CO2 and thickened CO2, the ratio of the viscosity of
thickened CO2 to neat CO2 was estimated as the ratio of the
respective pressure drops along the length of the core.
␮copolymer solution 共⌬P Ⲑ L兲copolymer solution
= . . . . . . . . . . . . . . . . . . . . . (7)
␮CO2 共⌬P Ⲑ L兲CO2
Results and Discussion
Polymer Characterization. The number-, weight-, and Z-average
molecular weights, Mn, Mw, and Mz, respectively, were determined
using gel permeation chromatography (American Polymer Stan-
dards Corporation, Mentor, Ohio). These results are presented in
Table 2, along with the polydispersity index (PDI), composition,
and melting point of copolymer. The PDI is the ratio of Mw to Mn.
88
Solubility Results. The copolymer is soluble in dense carbon di-
oxide at concentrations up to 5 wt% at 298, 323, 348, and 373 K
(Fig. 4). The cloud-point curve shifted toward higher pressure with
increasing concentration and temperature. The cloud-point pres-
sure at any given temperature must be contrasted with the mini-
mum miscibility pressure (MMP) at the same temperature to es-
tablish the solubility of the copolymer at reservoir conditions. For
example, consider a carbon dioxide-miscible project with a mini-
mum miscibility pressure of 18 MPa (2616 psia). Fig. 4 indicates
that at a temperature of 298 K and a pressure of 18 MPa, roughly
3 wt% of the copolymer could be dissolved in CO2, more than
enough to significantly thicken the CO2. At a reservoir temperature
of 323 K and a pressure of 18 MPa, about 0.5 wt% of the copoly-
mer could be dissolved in CO2—roughly the lowest amount of
copolymer needed to make a significant (e.g., doubling) increase in
CO2 viscosity. At a pressure of 18 MPa and 348 or 373 K, the
solubility of the copolymer in CO2 would be much less than 0.25
wt%, insufficient to induce an appreciable viscosity change. If
greater copolymer solubility is required at the reservoir conditions
than is possible with the 79%fluoroacrylate-21%styrene copoly-
mer with a number average molecular weight of 540,000, then a
copolymer of comparable molecular weight with a slightly greater
fluoroacrylate composition and/or a lower molecular weight would
be required,13 thereby making the copolymer more CO2-soluble.
Yet this modification would reduce the thickening capability of the
copolymer, and a slightly greater concentration of the copolymer
would be required to attain a desired level of viscosity. Future
studies are directed at the design of inexpensive, environmentally
benign, nonfluorous thickeners, however, rather than further opti-
mization of the fluorinated copolymer composition and molecu-
lar weight.
Falling Cylinder Viscometer Results. The falling cylinder results
for CO2-rich solutions at 298 K and 34 MPa containing 0.2 to 5.0
wt% copolymer are presented in Fig. 5. Experimental data are
provided along with curves that correspond to the power law cor-
relation obtained using the parameters listed in Table 3. The plots
of ␩solution/␮CO2 as a function of shear rate illustrate that the poly-
mer solutions are shear-thinning in nature and approach Newto-
nian behavior with diminishing copolymer concentration. No dis-
cernible increase in viscosity was detected at copolymer concen-
trations less than 0.2 wt%. This falling cylinder viscometer was not
suited for the study of dilute copolymer solutions at the very low
shear rates associated with creeping flow through porous media
(1-10s−1). Still, the results clearly indicate that substantial in-
creases in viscosity can be realized at extremely low shear rates
with the fluoroacrylate-styrene copolymer present at concentra-
tions of at least 0.2 wt%.
Fig. 4—Effect of temperature on the solubility of the copolymer
in carbon dioxide.
June 2003 SPE Journal

Fig. 5—Relative viscosity increase vs. shear rate at 298 K, 34
MPa, power law model, falling cylinder viscometry.
Effect of Temperature on Copolymer Thickening Capability.
The copolymer is capable of increasing the viscosity because of its
high molecular weight and the intermolecular associations of the
aromatic groups.13,14 Although Fig. 5 demonstrated effective
thickening at 298 K, reservoir temperatures can be much higher.
An increase in temperature may influence the strength and geom-
etry of the stacking of the phenyl groups, which may result in the
change of local molecular environment, thereby effecting the ␲-␲
stacking of the aromatic groups and the thickening capability of
the copolymer. Therefore the aluminum cylinder with a diameter
of 3.157 cm was used to estimate viscosity through falling cylinder
viscometry. Fig. 6 demonstrates that the effectiveness of the thick-
ener at concentrations of 1.0 and 0.5 wt% decreased slightly with
increasing temperature in the 298 to 373 K range at 34 MPa.
Measurement of Thickened CO2 Mobility in Sandstone. The
pressure drops along the core were on the order of 100 to 100,000
Pa, depending on the superficial velocity and the copolymer con-
centration. This pressure drop was small relative to the pressure at
the inlet face of the sandstone, which was maintained at 20 MPa.
The pressure drops along each third of the core varied by no more
than 5% for every mobility measurement. These variations were
random, with no section of the core (particularly the inlet section)
exhibiting a consistently greater pressure drop than the other sec-
tions. Therefore, no evidence of copolymer retention at the inlet
face of the core was detected. Pressure drops of the neat and
thickened CO2 flowing at the same superficial velocity were used
to estimate the increase in CO2 viscosity. This viscosity increase at
298 K and 20 MPa is presented as a function of copolymer con-
Fig. 6—Temperature effect on copolymer-CO2 solution relative
viscosity increase, 34 MPa, falling cylinder viscometry.
June 2003 SPE Journal
centration for a constant superficial velocity (Fig. 7). Copolymer
concentrations as low as 0.25 wt% were capable of thickening
CO2. Higher copolymer concentrations at a constant superficial
velocity led to remarkable increases in viscosity at low superficial
velocities. The 1.0 and 1.5 wt% copolymer solutions were 8 and 19
times more viscous than neat CO2, respectively, at a velocity of
0.00035 cm/s (1 ft/D). Because of the shear-thinning nature of the
copolymer, increases in superficial velocity led to diminished
thickening. For example, the addition of 1.5 wt% copolymer in-
creased the CO2 viscosity by a factor of only 2 at a superficial
velocity of 0.028 cm/s (80 ft/D). The permeability of the core to
neat CO2 that was injected after the copolymer solution agreed
with the original measurement of permeability to within 1%.
Therefore, the flow of the copolymer solution through the sand-
stone core did not result in any detectable permeability reduction.
Copolymer adsorption and changes in the wettability of the porous
media were not determined.
Next-Generation Thickeners. Silicone polymers have been pre-
viously used to thicken carbon dioxide flowing through porous
media, as evidenced by decreased mobility and improved recov-
ery.7 Substantial amounts of an organic cosolvent were required,
however, to attain a single-phase solution.7 The fluoroacrylate-
styrene copolymer is a direct thickener in that it does not require
a cosolvent. Nonetheless, this copolymer is a highly fluorinated
copolymer that is expensive, unavailable in large quantities, and
biologically and environmentally persistent. Rather than continu-
ing to modify the structure, composition, or molecular weight of
this fluoroacrylate-styrene copolymer, a new generation of CO2
thickeners is being designed. Future studies will focus on inex-
pensive, biodegradable, nonfluorous direct thickeners that can be
produced in large amounts and do not require an organic cosolvent
to achieve dissolution. The same strategy used in the development
of the fluorinated thickener will be employed in the development
of the nonfluorous thickener.
First, novel copolymers composed of carbon, hydrogen, oxy-
gen, and nitrogen will be designed or selected that exhibit remark-
Fig. 7—Relative viscosity increase vs. concentration at 298 K,
20 MPa, flow through ∼100 md Berea sandstone.
89
ably high CO2 solubility. Polymers containing carbonate and ether
groups, 19 sugar acetates, 20,21 acetate, 22 carbonyls, 19,22 and
ethers23 are particularly attractive candidates. For example, we
have recently demonstrated that poly(vinyl acetate) is the most
CO2 soluble, nonfluorous, high-molecular-weight, commodity
polymer that can readily dissolve in CO2 over a wide temperature
range.22 The pressure required for dissolution of the polymer is
much greater than even the highest MMP values at reservoir tem-
perature, however. Therefore nonfluorous polymers with an even
greater affinity for CO2 than poly(vinyl acetate) must be designed
and synthesized.
Second, CO2-phobic, associating groups will be incorporated
into the highly CO2-philic polymer to impart thickening capabili-
ties. The associating group content of the copolymer thickener
composition will be adjusted to attain a reasonable balance of CO2
solubility and thickening capability.
Conclusions
This study has established that it is possible to design a polymer
that exhibits both high carbon dioxide solubility and the ability to
enhance carbon dioxide through intermolecular associations using
a fluorinated copolymer. A 71 mol% fluoroacrylate-29 mol% sty-
rene random copolymer was synthesized through bulk, free-radical
polymerization. This copolymer, which had a number-average mo-
lecular weight of 540,000 and a polydispersity index of 1.63, ex-
hibited remarkably high solubility in dense carbon dioxide because
of the high fluoroacrylate content of the copolymer. The thicken-
ing capability of the copolymer over the 298 to 373 K temperature
range was attributed to intermolecular stacking of the aromatic
rings associated with the styrene monomer. Falling cylinder vis-
cometry results were used to develop a power law model of the
pseudoplastic solutions of the copolymer in CO2. Pressure-drop
measurements of high-pressure solutions flowing through Berea
sandstone demonstrated that the copolymer could effectively re-
duce the mobility of carbon dioxide. For example, the apparent
viscosity of CO2 increased by a factor of 8-6 at superficial veloc-
ities of 0.00035 to 0.0035 cm/s (1 to 10 ft/D) for a 1 wt% solution
of the copolymer in CO2. Higher velocities and lower copolymer
concentrations resulted in smaller mobility reductions. Although
this copolymer successfully thickened CO2, fluorinated polymers
are expensive, unavailable in bulk quantities, and environmentally
persistent. Therefore, we are currently investigating nonfluorous
thickeners, which are composed solely of carbon, hydrogen, oxy-
gen, and nitrogen. Acetate-rich polymers appear to be particularly
promising candidates.
Nomenclature
C ⳱ falling cylinder viscometer constant
K ⳱ permeability, L2, md
L ⳱ length, L, m
M ⳱ power law model constant
MMP ⳱ minimum miscibility pressure, m/Lt2, Pa
Mn ⳱ number-average molecular weight, m/mole, g/mol
Mw ⳱ weight-average molecular weight, m/mole, g/mol
Mz ⳱ Z-average molecular weight, m/mole, g/mol
n ⳱ power law model constant
P ⳱ pressure, m/Lt2, Pa
PDI ⳱ polydispersity index
r ⳱ r coordinate, L, m
Tm ⳱ melting point, T, K
v ⳱ superficial velocity, L/t, m/s
Vc ⳱ falling cylinder terminal velocity, L/t, m/s
Vz ⳱ velocity fraction in z direction, L/t, m/s
x ⳱ mole fraction styrene in copolymer
y ⳱ mole fraction fluoracrylate in copolymer
z ⳱ z coordinate, L, m
␩ ⳱ non-Newtonian fluid viscosity, m/Lt, Pa-sec
␮ ⳱ Newtonian fluid viscosity, m/Lt, Pa-sec
␳c ⳱ cylinder density, m/L3, g/cm3
90
␳f ⳱ fluid density, m/L3, g/cm3
␶ ⳱ shear stress, m/Lt2, N/m2
Acknowledgments
The authors would like to express their appreciation to the U.S.
D.O.E. Natl. Energy Technology Center for their support of this
program through Natl. Petroleum Technology Office contracts
DOE-RA26-98BC15108 and DE-FC26-01BC15315. Aaron
Wlaschin was supported by an NSF grant for Research Experience
for Undergraduates, EEC 0097664. We would also like to thank
Bayer, Cabot Oil, and Gas and Air Products for their financial
support. Dr. Mark Thies of Clemson U. assisted us in the charac-
terization of the copolymer.
References
1. Enick, R.M.: “A Literature Review of Attempts to Increase the Vis-
cosity of Dense Carbon Dioxide,” NETL report DE-AP26-97FT25356,
U.S. D.O.E., Washington, DC (October 1998) www.netl.doe.gov.
2. Heller, J. et al.: “Direct Thickeners for Mobility Control of CO2
Floods,” SPEJ (October 1985) 679.
3. Dandge, D.K., Taylor, C., and Heller, J.P.: “Associative Organotin
Polymers: 1. Symmetric Tributyltin Fluorides, Synthesis and Proper-
ties,” J. of Polymer Science—Part A Poly. Chem. (February 1989)
1053.
4. Dandge, D.K. et al.: “Associative Organotin Polymers,” J. of Macro-
molecular Science, Chem A 26 (1989) 18, 1451.
5. Dandge, D.K., Singh, P.K., and Heller, J.P.: “Associative Organotin
Polymers in Miscible Gas Enhanced Oil Recovery,” in Oilfield Chem-
istry—Enhanced Recovery and Production Stimulation, J.K. Borchardt
and T.F. Yen (eds.), ACS Symposium Series No. 396, American
Chemical Society, Washington, DC (1989) 529–546.
6. Gullapalli, P., Tsau, J.-S., and Heller, J.P.: “Gelling Behavior of 12-
Hydroxystearic Acid in Organic Fluids and Dense CO2,” paper SPE
28979 presented at the 1995 SPE International Symposium on Oilfield
Chemistry, San Antonio, Texas, 14–17 February.
7. Bae, J.H. and Irani, C.A.: “A Laboratory Investigation of Viscosified
CO2 Process,” SPE Advanced Technology Series (April 1993) 166.
8. Irani, C.A., Harris, T.V., and Pretzer, W.R.: “Polymer Containing Pen-
dant Vinyl Ether Groups Useful in EOR Using Carbon Dioxide Flood-
ing,” U.S. Patent No. 4,945,990 (1990).
9. Irani, C.A., Harris, T.V., and Pretzer, W.R.: “Polymer Containing Pen-
dant Tertiary Alkyl Amine Groups Useful in EOR Using Carbon Di-
oxide Flooding,” U.S. Patent No. 4,945,989 (1990).
10. McClain, J.B. et al.: “Characterization of Polymers and Amphiphiles in
Supercritical CO2 using Small Angle Neutron Scattering and Viscom-
etry,” 1996 Spring Meeting of the ACS, Division of Polymeric Mate-
rials Science and Engineering, New Orleans (1996).
11. Shi, C. et al.: “Semi-Fluorinated Trialkyltin Fluorides and Fluorinated
Telechelic Ionomers as Viscosity-Enhancing Agents for Carbon Diox-
ide,” Industrial & Engineering Chemistry Research (2001) 40, 908.
12. Shi, C. et al.: “The Gelation of CO2: A Sustainable Route to the
Formation of Microcellular Materials,” Science (19 November 1999)
1540.
13. Huang, Z. et al.: “Enhancement of the Viscosity of Carbon Dioxide
Using Styrene/Fluoroacrylate Copolymers,” Macromolecules (2000)
33, 5437.
14. Weber, E.: Topics in Current Chemistry: Design of Organic Solids,
Springer-Verlag, Hamburg, Germany (1998).
15. Iezzi, A.: MS thesis, U. of Pittsburgh (1989).
16. Ashare, E. and Bird, R.B.: “Falling Cylinder Viscometer for non-
Newtonian Fluids,” AIChEJ (1965) 11, 910.
17. Span, R. and Wagner, W.: “A New Equation of State for Carbon
Dioxide Covering the Fluid Region from the Triple-Point Temperature
to 1100 K at Pressures up to 800 MPa,” J. Phys. Chem. Ref. Data
(1996) 25, 1509.
18. Enick, R., Beckman, E., and Hamilton, A.: “Novel CO2-Thickeners for
Improved Mobility Control,” Quarterly Report, April 1–June 30, 2000,
DOE-RA26-98BC15018, Washington, DC (July 2000).
19. Sarbu, T., Styranec, T., and Beckman E.J.: “Non-fluorous Polymers
with Very High Solubility in Supercritical CO2 Down to Low Pres-
sures,” Nature (May 2000) 165.
June 2003 SPE Journal
20. Raveendran, P. and Wallen, S.L.: “Sugar Acetates as Novel, Renew- direct carbon dioxide thickeners for enhanced oil recovery,
able CO2-Philes,” J. Am. Chem. Soc. (2002) 124, 7274. funded by the U.S. D.O.E. Xu holds BS and MS degrees in
21. Potluri, V. et al.: “Per-acetylated Sugar Derivatives Show High Solu- chemical engineering from the East China U. of Science and
Technology. Aaron Wlaschin is a PhD candidate at the U. of
bility in Liquid and Supercritical Carbon Dioxide,” Organic Letters
Minnesota in the Chemical Engineering and Materials Science
(2002) 4, 2333. Dept., where he is advised by Dr. Friedrich Srienc. As an under-
22. Shen, Z. et al.: “CO2-Solubility of Oligomers and Polymers that Con- graduate he participated in an NSF-sponsored Research Ex-
tain the Carbonyl Group,” Polymer (2003) 44, 1491. periences for Undergraduates program at the U. of Pittsburgh
23. O’Neill, M.L. et al.: “Solubility of Homopolymers and Copolymers in in Bob Enick’s lab. In 2002 he received an NIH Biotechnology
Carbon Dioxide,” Ind. Eng. Chem. Res. (1998) 37, 3067. training fellowship. His thesis research focuses on using anaero-
bic fermentation and elementary mode modeling to optimize
the synthesis of biodegradable polymers. Wlaschin holds a BS
degree in chemical engineering from the U. of Nebraska. Bob
Enick is a professor in the Dept. of Chemical and Petroleum
SI Metric Conversion Factors
Engineering at the U. of Pittsburgh. e-mail: enick@engrng.
cp × 1.0* E–03 ⳱ Pa⭈s pitt.edu. He joined the university immediately after his gradu-
ft × 3.048* E–01 ⳱ m ation in 1985, and he currently holds the James T. MacLeod
Professorship in the School of Engineering. He is a long-standing
mL × 1.0* E+00 ⳱ cm3
member of SPE, and his current research focuses on the de-
psi × 6.894 757 E+00 ⳱ kPa sign, synthesis, and evaluation of hydrocarbon-based, carbon
*Conversion factor is exact. dioxide-soluble thickeners and surfactants for improved mobil-
ity control. Enick is also an Oak Ridge Fellow at the U.S. D.O.E.
Natl. Energy Technology Center, where he works on the de-
velopment of hydrogen and carbon dioxide membranes. En-
Jianhang Xu is a PhD student in the Dept. of Chemical and ick holds BS, MS, and PhD degrees in chemical engineering, as
Petroleum Engineering at the U. of Pittsburgh. e-mail: well as an MS degree in petroleum engineering, from the U.
jianhang+@pitt.edu. He is currently working on the study of of Pittsburgh.

June 2003 SPE Journal 91