Ind. Eng. Chem. Res.

2008, 47, 1261-1267 1261

Carbon Dioxide Capture Using a CO2-Selective Facilitated Transport Membrane

Jin Huang, Jian Zou, and W. S. Winston Ho*
Department of Chemical and Biomolecular Engineering, Department of Materials Science and Engineering,
The Ohio State UniVersity, 140 W. 19th AVenue, Columbus, Ohio 43210-1180

A novel CO2-selective membrane with the facilitated transport mechanism has been synthesized to capture
CO2 from the industrial gas mixtures, including flue gas. Both mobile and fixed amine carriers were incorporated
into the cross-linked poly(vinyl alcohol) (PVA) during the membrane synthesis. The membrane showed
desirable CO2 permeability (with a suitable effective thickness) and CO2/N2 selectivity up to 170 C. In the
CO2 capture experiments from a gas mixture with N2 and H2, a permeate CO2 dry concentration of >98%
was obtained, using steam as the sweep gas. The effects of the feed flow rate and the sweep:feed molar ratio
on the membrane separation performance were investigated. A one-dimensional isothermal model was
established to examine the performance of a hollow-fiber membrane module composed of the described CO2-
selective membrane. The modeling results show that a CO2 recovery of >95% and a permeate CO2 dry
concentration of >98% are achievable from a 1000 standard cubic feet per minute (SCFM) (or 21.06 mol/s)
flue gas stream with a 2 ft (0.61 m) hollow-fiber module that contained 980 000 fibers.

1. Introduction In this type of membrane, the incorporated carrier agents can

The increasing public concern over global warming has been
concentrated on the anthropogenic (man-made) emissions of
greenhouse gases, especially carbon dioxide (CO2). CO2 is
primarily emitted as the result of the combustion of fossil fuels,
and also from nature gas sweetening, the production of synthesis
gas, and certain chemical plants.1 Currently, the principal CO2
capture technologies include chemical and/or physical absorp-
tion, physical adsorption, membrane separation, and cryogenic
distillation.2 Compared to other technologies, membrane-based
CO2 capture has the advantages of low-energy consumption,
simplicity of operation, and an absence of moving parts. It has
shown great potential in natural gas purification, CO2 capture
from emissions of coal-fired power plants, and metabolic CO2
removal from enclosed breathing atmospheres (for example, in
a space shuttle or space station).3-7 Moreover, it may also be
used in the CO2-selective water-gas-shift membrane reactor for
the purification of H2 in fuel-cell fuel processing, which is a
key challenge for the proton-exchange membrane fuel cells
(PEMFCs). We have proposed and verified this idea in earlier
publications.8,9
However, todays commercial polymeric membranes, such
as cellulose acetate and polyimide, suffer from low CO2
permeability and selectivity, because they are based on a
solution-diffusion mechanism and rely primarily on the subtle
size differences of the penetrants to achieve separation.3,10,11
Also, these membranes are not suitable for high-temperature
and high-humidity operation, which is required in the direct CO2
capture from the flue gas in power plants.11,12 For the specific
application of CO2 capture from the flue gas, it has been reported
that a CO2/N2 selectivity of >70 and a minimum CO2 perme-
ability of 100 Barrers for a membrane thickness of 0.1 m (a
permeance of 1000 GPU (gas permeation unit), where 1 GPU
) 10-6 cm3 (STP)/(cm2 s cm Hg)) are required for the economic
operation.13,14 As an alternative to conventional polymeric
membranes, facilitated transport membranes have shown better
promise to satisfy these goals.4,15,16
* To whom correspondence should be addressed: E-mail address:
ho@chbmeng.ohio-state.edu.
10.1021/ie070794r CCC: $40.75 2008 American Chemical Society
Published on Web 01/16/2008
react reversibly with the target gas component. Therefore, the
reaction in the membrane creates another transport mechanism,
in addition to the simple solution-diffusion mechanism. The
target gas component dissolves in the membrane first, and it
can diffuse down its own concentration gradient or diffuse down
a concentration gradient of a carrier-gas complex. Meanwhile,
the gas components that do not react with the carrier agents
will permeate exclusively via the solution-diffusion mecha-
nism.16 The facilitated transport mechanism for CO2 removal
is shown schematically in Figure 1.
The reaction between CO2 and the amine carrier can be
described by the zwitterions mechanism, which was originally
proposed by Caplow17 and reintroduced by Danckwerts.18 First,
CO2 reacts with primary or secondary amines (RRNH, where
R is a functional group and R is a functional group or hydrogen)
to form zwitterions as an intermediate.
CO2 + RRNH h RRNH+COO- (1)
The zwitterion then is deprotonated by bases such as the amine
itself and H2O to form the carbamate ion.
RRNH+ COO- + RRNH h RRNCOO- + RRNH2+ (2)
RRNH+ COO- + H2O h RRNCOO- + H3O+ (3)
If the carbamate ion of the amine carrier is not stable, it will
react with H2O to form bicarbonate, HCO3-.
RRNCOO- + H2O h RRNH + HCO3- (4)
In addition, a water-swollen condition may provide a better
facilitation effect than a dry condition. This is due to the fact
that the CO2 hydration reaction in a water-swollen membrane
would be enhanced in the presence of amino groups, which work
as weak base catalysts. Therefore, CO2 transport is facilitated
in the forms of carbamate and bicarbonate.19 Under the partial
pressure difference in the gas separation process, these carrier-
gas reaction products (i.e., carbamate and bicarbonate) will
diffuse down their concentration gradient or pass off to the next
carrier agent.
1262 Ind. Eng. Chem. Res., Vol. 47, No. 4, 2008
Figure 1. Schematic of the facilitated transport mechanism.
Extensive studies have been done on the CO2 facilitated
transport membranes since the 1960s.8,9,15,16,19-36 Based on the
carrier mobility, facilitated transport membranes can be divided
into two categories: mobile-carrier membranes and fixed-carrier
membranes. In mobile-carrier membranes, the carrier can diffuse
in the membrane. Generally, they were prepared by immersing
the microporous supports in the carrier solutions or a reactive
solvent, which was known as the immobilized liquid membrane
(ILM).22-24 However, this configuration is relatively unstable
(e.g., the membrane can be dried out at high temperatures and
the carrier agents can be lost). To improve the membrane
stability, the ion-exchange membrane was first proposed as the
support of ILM by LeBlanc et al.25 When the ionic mobile
carriers such as monoprotonated ethylenediamine were im-
mobilized in an ion exchange membrane by the electrostatic
force, the carrier washout would be reduced, to some extent.25-28
In fixed-carrier membranes,19,29-36 in which the reactive carrier
agents are covalently bonded to the polymer backbone, CO2
reacts at one carrier site and then hops to the next carrier site
along the direction of the concentration driving force via the
hopping mechanism.20 Even though mobile-carrier membranes
have shown higher CO2 transport rates than fixed-carrier
membranes, because of the higher reactive diffusivity, fixed-
carrier membranes were considered to be the most stable
configuration.16
In the present study, a novel CO2-selective polymeric
membrane that contained both mobile and fixed amine carriers
was prepared, and the permeation properties were measured at
g100 C. The behaviors of CO2 capture from the gas mixtures
were studied with a simulated feed gas while using steam as
the sweep gas. Based on the measured CO2 permeability and
CO2/N2 selectivity, a mathematical model was developed to
evaluate the separation performance of a hollow-fiber module
that was composed of the described membrane.
2. Experimental Section
2.1. Chemicals and Materials. Poly(vinyl alcohol) (PVA)
with 99+% hydrolysis [from poly(vinyl acetate)] and a molec-
ular weight of 89 000-98 000, potassium hydroxide (KOH)
(85+ wt %), formaldehyde (37 wt % solution in water), and
2-aminoisobutyric acid (AIBA) were purchased from Aldrich
Chemicals (St. Louis, MO). Poly(allylamine hydrochloride),
with a molecular weight of 60 000 was obtained from
Polysciences, Inc. (Warrington, PA). Deionized (DI) water was
used as the solvent during the membrane preparation. All
these chemicals were used as received, without further purifica-
tion.
Two certificated gas mixtures were used as the feed gases.
The gas compositions were (1) 20% CO2, 40% H2, and 40%
N2, and (2) 1% CO, 17% CO2, 45% H2, and 37% N2. The
second gas mixture simulated the composition of synthesis gas
from the autothermal reforming of gasoline with air. Prepurified
argon (99.998%) was used as the sweep gas in the gas
permeation measurements, for ease of gas chromatography (GC)
analysis. All of these gases were obtained from Praxair, Inc.
(Danbury, CT). The microporous poly(tetrafluoroethylene)
(PTFE) support (thickness of 60 m, average pore size of 0.2
m, and 80% porosity) was donated by BHA Technologies
(Kansas City, MO).
2.2. Synthesis of CO2-Selective Membrane. The membrane
of PVA that contained amino functional groups in the thin-
film composite structure was prepared using the solution casting
method. PVA was first dissolved in DI water at 80 C under
stirring. A stoichiometric amount of formaldehyde and a certain
amount of KOH were added into the PVA aqueous solution to
achieve a 60 mol % degree of crosslinking, based on the
hydroxyl groups of PVA. The PVA/formaldehyde/KOH solution
was heated at 80 C for 20 h under stirring. The crosslinking
of PVA with formaldehyde is a condensation reaction, which
is shown schematically in Figure 2.
Free polyallyamine was prepared by mixing poly(allylamine
hydrochloride) with a stoichiometric amount of KOH in
methanol overnight. 2-Aminoisobutyric acid (AIBA) potassium
salt (AIBA-K) was synthesized by mixing AIBA with a
stoichiometric amount of KOH in DI water for 30 min. The
chemical structures of these compounds are illustrated in Figure
3. The designated amounts of these two chemicals were then
added into the cross-linked PVA solution as the fixed and mobile
carriers, respectively. The resulting solution was centrifuged to
remove air bubbles and cast to a uniform thickness with a
GARDCO adjustable micrometer film applicator (Paul N.
Gardner Company, Pompano Beach, FL) on the microporous
PTFE support. The solution was viscous enough to avoid
penetrating the pores of the support, which was verified by SEM
cross-sectional pictures. The gap setting of the film applicator
was adjusted to control the membrane thickness. The membrane
was dried first in a fume hood under ambient conditions
overnight, to remove most of the water, and then was further
cured in a 120 C furnace for 6 h to complete the cross-linking
reaction. The final composition of the dense, active layer was
50 wt % PVA (60 mol % crosslinking), 10 wt % AIBA-K,
20 wt % KOH, and 20 wt % polyallylamine. The thickness of
the active layer of the membranes ranged from 5 m to
80 m after heat treatment; this was referred to as the thick-
ness of the membrane. Details about the preparation pro-
cedure and characterization of this membrane can be found
elsewhere.35
2.3. Gas Permeation Measurements. The transport proper-
ties of the membrane were measured using a gas-permeation
unit, such as that shown in Figure 4. In the experiments, a
membrane with a thickness of 30 m was cut into a circular
piece and mounted into a circular permeation cell with an
effective area of 45.60 cm2. The circular permeation cell that
was used had flow channels on the feed and permeate sides
that were designed in a way that both the retentate and the
permeate were well-mixed, respectively, inside the cell, so that
we could neglect gas-phase boundary layer resistances. A ternary
gas mixture that contained 20% CO2, 40% H2, and 40% N2 at
a flow rate of 60 mL/min (under ambient conditions) was used
as the feed gas. We included H2 in the gas mixture to study the
potential of CO2-selective membranes in H2 purification for fuel
cells. The total pressure on the feed side was measured using a
pressure gauge and set at 2 atm via a back-pressure regulator
(Model 44-2362-24, Tescom Co., Elk River, MN). The total
pressure of argon sweep gas was set at 1.1 atm, and the flow
rate was 30 mL/min (under ambient conditions).

Figure 2. Synthesis of the cross-linked PVA with formaldehyde.
Figure 3. Chemical structures of (a) free polyallylamine and (b) AIBA-K.
On each of the feed and sweep sides, DI water at a flow rate
of 0.03 mL/min was pumped into the gas permeation unit by a
solvent delivery pump (Prostar 210, Varian, Inc., Palo Alto, CA).
Water vapor and each of the dry gases were mixed well in a
humidifier. The testing temperature was adjusted with an oven
temperature controller (Model FTU4.6+50350C, Bemco Inc.,
Simi Valley, CA). After leaving the gas permeation cell, the
water vapors in both the retentate and the permeate were trapped
in respective water knockout vessels. The dry gas compositions
of both gas streams were analyzed using a gas chromatograph
that was equipped with a thermal conductivity detector (Model
6890N, Agilent Technologies, Palo Alto, CA) and a stainless
steel micropacked column (80/100 mesh Carboxen 1004, Sigma-
Aldrich, St. Louis, MO). A counter-current flow configuration
was applied to offer the maximum driving force across the
membrane.
2.4. CO2 Removal Capacity. To illustrate the CO2 removal
capacity of the flat-sheet membrane, we constructed a rectan-
gular gas permeation cell, which had a larger effective mem-
brane area of 342.7 cm2 (19.10 cm 17.95 cm) and well-
defined gas-flow channels on both the feed and sweep sides.
The feed gas consisted of 1% CO, 17% CO2, 45% H2, and
37% N2 (on a dry basis). The water contents on the feed and
sweep sides were 35% and 95%, respectively. Membranes
with the same composition as the previous one but with
thicknesses of 20-80 m were used. The retentate CO2 dry
concentrations, from various feed flow rates (ranging from 10
mL/min to 130 mL/min, under ambient conditions), were
measured.
2.5. CO2 Capture Experiments. The experiments on CO2
capture from a gas mixture were performed in the circular
permeation unit previously described. Compared to other gases,
steam is the more desirable sweep gas in the CO2 capture
process, because water will condense under ambient conditions
and a high-purity CO2 product can be readily obtained. The
ternary gas mixture that contained 20% CO2 was used as the
feed gas. Various gas flow rates were applied to study the effects
of process parameters, such as the feed flow rate and the sweep:
feed molar ratio, on the CO2 recovery and the permeate CO2
purity. The experimental conditions are listed in Table 1. All
of the CO2 capture experiments were performed at 110 C for
the optimal membrane transport properties. Feed-side and
sweep-side pressures were set at 2 and 1.1 atm, respectively.
The membrane thickness was 28 m. The CO2 recovery was
defined as the ratio of the permeate CO2 molar flow rate to the
feed inlet CO2 molar flow rate.
Ind. Eng. Chem. Res., Vol. 47, No. 4, 2008 1263
3. Modeling Work
It is interesting to evaluate the separation performance of the
CO2-selective membrane module through the modeling based
on the transport data obtained experimentally. In the modeling
work, a one-dimensional isothermal model was configured to
simulate the CO2 capture from the flue gas with a hollow-fiber
membrane module that was composed of the CO2-selective
membrane previously described. The following assumptions
were made in the model:8
(1) The module is isothermal and operating at a steady state;
(2) The module is operated with a counter-current configu-
ration;
(3) There is no axial mixing;
(4) The pressure drops on both the lumen and shell sides are
negligible.
Based on the differential volume element on a single hollow
fiber along the length direction, as shown in the lower right-
hand side of Figure 5, the molar balances for gas species i on
the feed (lumen) side and the sweep (shell) side can be expressed
as eqs 5 and 6, respectively:
dnfi
) -dinJi (5)
dz
dnsi
) -dinJi (6)
dz
where Ji is expressed as
J i ) Pi( )pi
l
(7)
The feed gas was assumed to be a simulated flue gas mixture
that contained 9% CO2, 11% H2O, and 80% N2 at a flow rate
of 1000 standard cubic feet per minute (SCFM) (or 21.06 mol/
s). Pure steam was used as the sweep gas. The operating
temperature was assumed to be 110 C, and both the feed-side
and sweep-side pressures were set at 1 atm, considering the real
case of the flue gas from power plants. The inlet sweep:feed
molar flow rate ratio was set at 3.
The bVp4c solver in Matlab was used to solve the aforemen-
tioned differential equations of the boundary value problem.
During the calculations, the hollow fiber number was adjusted
to satisfy a constraint of CO2 recovery (i.e., > 95%). In addition,
a similar flat-sheet model was applied to simulate CO2 removal
experiments with the rectangular cell described earlier. The same
geometric dimensions and process parameters as those in the
experiment were used in the calculations.
4. Results and Discussion
4.1. Transport Properties of the CO2-Selective Membrane.
The transport properties of the gas separation membrane are
1264 Ind. Eng. Chem. Res., Vol. 47, No. 4, 2008

Figure 4. Schematic of the gas permeation unit.

Table 1. Operating Parameters for the CO2 Capture Experiments

Feed Side
sweep-side
dry gas water total water
flow rate flow rate flow rate flow rate sweep:feed
experiment (mL/min)a (mL/min) (10-5 mol/s) (mL/min) molar ratio
1 40 0.02 4.58 0.050 1.0
2 60 0.03 6.86 0.075 1.0
3 80 0.04 9.15 0.100 1.0
4 100 0.05 11.44 0.125 1.0
5 80 0.04 9.15 0.050 0.5
6 80 0.04 9.15 0.100 1.0
7 80 0.04 9.15 0.150 1.5
8 80 0.04 9.15 0.200 2.0
a Under ambient conditions.

characterized primarily by permeability and selectivity. The

permeability of a gas species i is defined in terms of the steady-
state flux and the thickness-normalized pressure driving force.
It can be calculated by rearranging eq 7:
Ji
Pi )
pi/l
The common unit of permeability is Barrer (1 Barrer ) 10-10
cm3 (STP) cm/(cm2 s cmHg)). The selectivity between two
components i and j in a gas mixture may be expressed as
yi/yj
Rij )
xi/xj
A schematic of the CO2 separation in the permeation cell is
shown in Figure 6. Based on the measured gas compositions of
both the retentate and the permeate, the mass balance was
performed with the assumption of a constant argon molar flow
rate on the sweep side. The CO2 permeability and CO2/N2
selectivity were then calculated from eqs 8 and 9.
Figure 7 shows the membrane CO2 permeability and CO2/
N2 selectivity, as a function of temperature. As the temperature
increased from 100 C to 110 C, both the permeability and
selectivity increased, which could be explained by the signifi-
cantly higher CO2/amine reaction rate but relatively stable N2
permeability in the membrane at the elevated temperature. A
CO2 permeability of 6196 Barrers and a CO2/N2 selectivity of
493 were obtained at 110 C. As the temperature increased to
Figure 5. Schematic of the hollow-fiber membrane module.
g120 C, both permeability and selectivity gradually decreased,
because the water content in the membrane decreased with
increasing temperature. However, a CO2 permeability of >2000
(8)
Barrers and a CO2/N2 selectivity of >40 were still observed at
170 C, which demonstrated that PVA, as a hydrophilic polymer,
was a desirable polymeric matrix for the CO2-selective facilitated
transport membrane. Crosslinking was necessary for PVA to
hold the membrane structure while maintaining enough water
content at the high temperatures. Compared to the traditional
facilitated transport membranes immobilized in microporous
(9) supports, such as ILMs, embedding carriers in cross-linked PVA
matrix may significantly improve the long-term membrane
stability, because the matrix has strong interactions with the
nonvolatile mobile carrier, including hydrogen bonding. AIBA-
K, poly(allylamine), and KHCO3-K2CO3 (converted from
KOH) were the carriers to facilitate CO2 transport, as discussed
previously. In principle, the mobile carrier and the fixed carrier
may have advantages on carrier mobility/membrane perme-
ability33 and membrane stability, respectively. The optimal
combination of both types of carriers is of great importance to
the overall membrane performance, particularly for membrane
permeability.33
Almost all of the CO2 facilitated transport studies reported
in the literature were conducted at approximately ambient
temperature. The excellent transport properties of this CO2-
selective membrane at the high temperatures would be more

Figure 6. Schematic of gas separation with the CO2-selective membrane.
Figure 7. CO2 permeability and CO2/N2 selectivity, as a function of
temperature. Feed gas: 20% CO2, 40% H2, and 40% N2 (on a dry basis);
water content in feed ) 41%; pf ) 2 atm, ps ) 1.1 atm; membrane thickness
) 30 m.
Figure 8. Exit dry CO2 concentration in the retentate versus the feed flow
rate. Feed gas: 1% CO, 17% CO2, 45% H2, and 37% N2 (on a dry basis);
water content in feed ) 35%; pf ) 2 atm, ps ) 1.1 atm; membrane thickness
) 20-80 m.
desirable to CO2 capture from the flue gas, because steam could
be used as the sweep gas and operating at higher temperatures
would reduce the energy loss from cooling the flue gas.
4.2. CO2 Removal Capacity. A rectangular gas permeation
cell was used to study the CO2 removal capacity. The well-
defined gas-flow channels of this cell allowed us to compare
the experimental data with the results from the one-dimensional
model, which assumed no axial mixing. CO2 removal experi-
ments were conducted using the rectangular cell at a temperature
of 120 C with different feed flow rates.
As shown in Figure 8, the membrane removed CO2 in the
retentate, from 17% to 100 ppm, at a feed gas flow rate of 60
mL/min (under ambient conditions) and to 1000 ppm at a flow
rate of 120 mL/min (under ambient conditions). Such effective-
ness in CO2 removal could be attributed to the high CO2
permeability demonstrated in Figure 7. The variation of data in
Figure 8 was presumably due to the membranes with different
thicknesses that were used in the experiments, because the
thicker membrane, with the higher mass-transfer resistance,
resulted in a higher exit dry CO2 concentration on the feed side.
The modeling results, based on the CO2 transport properties
Ind. Eng. Chem. Res., Vol. 47, No. 4, 2008 1265
Figure 9. Permeate CO2 dry concentration and CO2 recovery versus the
feed flow rate. Feed gas: 20% CO2, 40% H2, and 40% N2 (on a dry basis);
water content in feed ) 41%; pf ) 2 atm, ps ) 1.1 atm; membrane thickness
) 28 m.
obtained experimentally, are also shown in Figure 8. The
experimental data have agreed well with the modeling results.
This suggests that the one-dimensional model can accurately
describe the gas permeation process in the rectangular cell.
4.3. CO2 Capture Performance. Although flue gases from
different sources might have various compositions, N2, CO2,
and water vapor are generally the major components, accounting
for >90% of the gas stream. After the dry feed gas was mixed
with the water vapor in the vessel in the experimental setup,
the total feed gas composition used was 12% CO2, 24%
H2, 24% N2, and 40% H2O, which would provide a close
estimate of the CO2 capture from the real flue gas. Using the
circular permeation cell, we have demonstrated the CO2 capture
performance with different feed inlet flow rates. The enhancing
effect of the sweep gas on the separation performance was also
investigated experimentally.
4.3.1. Effect of Feed Inlet Flow Rate. To study the effect
of the feed molar flow rate on the membrane separation
performance, the CO2 capture experiments were performed with
four different flow rates while the feed gas composition and
the sweep:feed ratio were kept constant (refer to experiments
1-4 in Table 1). As shown in Figure 9, a permeate CO2 dry
concentration of >98% was achieved for feed flow rates as high
as 11.44 10-5 mol/s (a 100 mL/min dry feed gas flow rate,
as shown in Table 1). Furthermore, if we deducted the H2 portion
from the permeate, the CO2 dry concentration would be >99%.
Obviously, the high CO2/N2 selectivity accounted for this high
permeate CO2 purity. A slight increase in the permeate CO2
dry concentration was observed with the increases in the feed
flow rate. This was due to the fact that the higher water flow
rate that was associated with the higher feed rate (to keep the
same water content in the feed gas) was more favorable for
keeping a higher water content in the membrane, which was
also more favorable for the facilitated transport mechanism. The
CO2 recovery decreased as the feed inlet flow rate increased,
1266 Ind. Eng. Chem. Res., Vol. 47, No. 4, 2008
Figure 10. Permeate CO2 dry concentration and CO2 recovery versus the
sweep-to-feed molar ratio. Feed gas: 20% CO2, 40% H2, and 40% N2 (on
a dry basis); water content in feed ) 41%; pf ) 2 atm, ps ) 1.1 atm;
membrane thickness ) 28 m.
which was understandable, considering that the residence time
of the gas mixture in the permeation cell was reduced as the
feed inlet flow rate increased.
4.3.2. Effect of the Sweep:Feed Ratio. The impact of the
sweep gas on the CO2 capture was investigated using four sweep
water rates while the gas and water flow rates on the feed side
were kept constant. As shown in Table 1 (from experiments
5-8), the inlet sweep:feed molar flow rate ratios were 0.5, 1,
1.5, and 2, respectively. Figure 10 illustrates the effect of the
sweep:feed ratio on the permeate CO2 dry concentration and
the CO2 recovery. The permeate CO2 dry concentrations were
all >95% and did not change significantly as the sweep:feed
ratio increased. The slightly declining trend was presumably
due to the fact that a slightly larger amount of CO2 was absorbed
in the water knockout when a higher water flow rate on the
sweep side was used. The CO2 recovery increased significantly
as the sweep:feed ratio increased from 0.5 to 1.5. A higher
sweep:feed ratio resulted in a lower CO2 concentration on the
sweep side and then a higher CO2 permeation driving force.
Beyond a sweep:feed ratio of 1.5, higher sweep:feed ratios
did not show a higher CO2 recovery, which suggested that
the excessively higher steam rate on the sweep side would
not significantly increase the permeation driving force,
because the sweep-side CO2 concentration was already suf-
ficiently low.
4.4. Modeling Study. As a common type of commercialized
membrane modules, the hollow-fiber membrane module has
shown excellent mass-transfer performance, because of its large
surface area per unit volume.4 In the modeling calculations,
hollow fibers were assumed to have a length of 61 cm (2 ft),
an inner diameter of 0.1 cm, and a porous support with a
porosity of 50% and a thickness of 30 m. The membrane was
assumed to have an effective thickness of 5 m, a CO2/N2
selectivity of 500, and a CO2 permeability of 6000 Barrers.
These values of the CO2/N2 selectivity and the CO2 permeability
were based on the experimental data shown previously. As
reported in an earlier paper from our group,35 a constant CO2
permeability of 6000 Barrers for a CO2 partial pressure of
<0.4 atm was observed experimentally. Generally, the typical
CO2 partial pressure in a flue gas is <0.2 atm. Thus, the constant
CO2 permeability of 6000 Barrers was used for a CO2 partial
pressure of <0.2 atm in the flue gas for CO2 capture. Note that
the CO2 permeance was 1200 GPU for the CO2 permeability
of 6000 Barrers and an effective thickness of 5 m.
Figure 11. CO2 concentration profiles along the length of the membrane
module.
Based on the calculations, 980 000 hollow fibers with the
given dimensions were needed to recover >95% of the CO2
from the 1000 SCFM (or 21.06 mol/s) of flue gas, which
contained 9% CO2. Figure 11 shows the profiles of the CO2
concentration on both feed and sweep sides along the length of
the counter-current hollow-fiber membrane module. The model-
ing results demonstrated that this membrane module could
reduce the feed-side CO2 concentration (on the wet basis) from
9% to 0.487%. The calculated permeate CO2 concentration (on
the dry basis) and CO2 recovery were 98.25% and 95.06%,
respectively. Because it is well-known that there is no significant
scale-up effect for the membrane module,4 simply adding the
hollow fiber number or membrane area would result in a
predictable capacity increase, if a larger amount of the flue
gas must be processed. Therefore, CO2 capture with such
hollow-fiber membrane modules could have great potential to
provide a viable approach for reducing CO2 emissions from flue
gas.
5. Conclusions
The membrane-based CO2 capture, using a new facilitated
transport CO2-selective membrane, was studied. The membrane
that was synthesized containing poly(allylamine) and AIBA-K
as the amine carriers in the cross-linked PVA network showed
good CO2 permeability and CO2/N2 selectivity up to 170 C.
The highest CO2/N2 selectivity of 493 and the highest CO2
permeability of 6196 Barrers were observed at 110 C. With a
larger effective membrane area of 342.7 cm2, we have removed
CO2 in the retentate, from a concentration of 17% to 100 ppm
at a feed gas flow rate of 60 mL/min and to 1000 ppm at a
flow rate of 120 mL/min. The experimental data have agreed
well with the modeling results. In the CO2 capture experiments,
a permeate CO2 dry concentration of >98% was achieved using
steam as the sweep gas. The effects of inlet feed flow rate and
sweep:feed ratio on the separation performance also were
studied. As the feed flow rate increased, the permeate CO2 dry
concentration slightly increased, and the CO2 recovery de-
creased, because of the reduced residence time. Increasing the
sweep:feed ratio enhanced the permeation driving force and
resulted in higher CO2 recovery, while the permeate CO2 dry
concentration did not change significantly.
We have developed a one-dimensional isothermal model to
evaluate the separation performance of a hollow-fiber module
that is composed of the described CO2-selective membrane.
From the modeling results, a 2-ft hollow-fiber module that
contains 980 000 fibers was sufficient to recover >95% of the
CO2 and obtain a permeate CO2 concentration of >98% (on a
dry basis) from a 1000 SCFM (or 21.06 mol/s) flue gas stream
that contained 9% CO2.

Nomenclature
d ) hollow fiber diameter (cm)
J ) permeation flux (mol/(cm2 s))
l ) membrane thickness (cm)
n ) molar flow rate (mol/s)
p ) pressure (atm)
P ) permeability (Barrer)
x ) retentate molar fraction
y ) permeate molar fraction
z ) axial position along the length of reactor (cm)
Greek Letters
R ) selectivity
Subscripts
f ) feed side
i, j ) species
in ) inside of the hollow fiber
s ) sweep side
Acknowledgment
We would like to thank Chris Plotz and BHA Technologies,
Inc., for giving us the BHA microporous Teflon support used
in this work. We would also like to thank the National Science
Foundation (NSF), the Ohio Department of Development
(through Wright Center of Innovation Grant No. 342-0561), and
The Ohio State University for the financial support. Part of this
material is based on work supported by the NSF, under Grant
No. 0625758.
Literature Cited
(1) U.S. Department of Energy. Emissions of greenhouse gases in the
United States 2004. Available via the Internet at http://www.eia.doe.gov/
oiaf/1605/ggrpt/pdf/057304.pdf, 2006.
(2) Aaron, D.; Tsouris, C. Separation of CO2 from flue gas: a review.
Sep. Sci. Technol. 2005, 40, 321-348.
(3) Zolandz, R. R.; Fleming, G. K. Gas permeation. In Membrane
Handbook; Ho, W. S. W., Sirkar, K. K., Eds.; Chapman and Hall: New
York, 1992.
(4) Ho, W. S. W.; Sirkar, K. K., Eds. Membrane Handbook; Chapman
and Hall: New York, 1992.
(5) Kesting, R. E.; Fritzche, A. K. Polymeric Gas Separation Mem-
branes; Wiley: New York, 1993.
(6) Koros, W. J.; Fleming, G. K. Membrane based gas separation. J.
Membr. Sci. 1993, 83, 1-80.
(7) Xomeritakis, G.; Tsai, C.-Y.; Brinker, C. J. Microporous sol-gel
derived aminosilicate membrane for enhanced carbon dioxide separation.
Sep. Purif. Technol. 2005, 42, 249-257.
(8) Huang, J.; El-Azzami, L.; Ho, W. S. W. Modeling of CO2-selective
water gas shift membrane reactor for fuel cell. J. Membr. Sci. 2005, 261,
67-75.
(9) Zou, J.; Huang, J.; Ho, W. S. W. CO2-selective water gas shift
membrane reactor for fuel cell hydrogen processing. Ind. Eng. Chem. Res.
2007, 46, 2272-2279.
(10) Meisen, A.; Shuai, X. Research and development issues in CO2
capture. Energy ConVers. Manage. 1997, 38 (Suppl.), S37-S42.
(11) Koros, W. J.; Mahajan, R. Pushing the limits on possibilities for
large scale gas separation: which strategies. J. Membr. Sci. 2000, 175, 181-
196.
(12) Bredesen, R.; Jordal, K.; Bolland, O. High-temperature membranes
in power generation with CO2 capture. Chem. Eng. Process. 2004, 43,
1129-1158.
(13) Hirayama, Y.; Kase, Y.; Tanihara, N.; Sumiyama, Y.; Kusuki, Y.;
Haraya, K. Permeation properties to CO2 and N2 of poly(ethylene oxide)-
Ind. Eng. Chem. Res., Vol. 47, No. 4, 2008 1267
containing and crosslinked polymer films. J. Membr. Sci. 1999, 160, 87-
99.
(14) Bounaceur, R.; Lape, N.; Roizard, D.; Vallieres, C.; Favre, E.
Membrane processes for post-combustion carbon dioxide capture: A
parametric study. Energy 2006, 31, 2220-2234.
(15) Ward, W. J.; Robb, W. L. Carbon dioxide-oxygen separation:
facilitated transport of carbon dioxide across a liquid film. Science 1967,
156, 1481-1484.
(16) Way, J. D.; Noble, R. D. Facilitated transport. In Membrane
Handbook; Ho, W. S. W., Sirkar, K. K., Eds.; Chapman and Hall: New
York, 1992.
(17) Caplow, M. Kinetics of carbamate formation and breakdown. J.
Am. Chem. Soc. 1968, 90, 6795-6803.
(18) Danckwerts, P. V. The reaction of CO2 with ethanolamines. Chem.
Eng. Sci. 1979, 34, 443-446.
(19) Matsuyama, H.; Teramoto, M.; Sakakura, H. Selective permeation
of CO2 through poly(2-N,N-dimethylaminoethyl methacrylate) membrane
prepared by plasma-graft polymerization technique. J. Membr. Sci. 1996,
114, 193-200.
(20) Cussler, E. L.; Aris, R.; Bhown, A. On the limits of facilitated
diffusion. J. Membr. Sci. 1989, 43, 149-164.
(21) Nobel, R. D. Analysis of facilitated transport with fixed site carrier
membranes. J. Membr. Sci. 1990, 50, 207-214.
(22) Teramoto, M.; Nakai, K.; Ohnishi, N.; Huang, Q.; Watari, T.;
Matsuyama, H. Facilitated transport of carbon dioxide through supported
liquid membranes of aqueous amine solutions. Ind. Eng. Chem. Res. 1996,
35, 538-545.
(23) Kovvali, A. S.; Sirkar, K. K. Carbon dioxide separation with novel
solvents as liquid membranes. Ind. Eng. Chem. Res. 2002, 41, 2287-2295.
(24) Marzouqi, M. H. A.; Abdulkarim, M. A.; Marzouk, S. A.; El-Naas,
M. H.; Hasanain, H. M. Facilitated transport of CO2 through immobilized
liquid membrane. Ind. Eng. Chem. Res. 2005, 44, 9273-9278.
(25) LeBlanc, O. H.; Ward, W. J., Jr.; Matson, S. L.; Kimura, S. G.
Facilitated transport in ion-exchange membranes. J. Membr. Sci. 1980, 6,
339-343.
(26) Way, J. D.; Nobel, R. D. Competitive facilitated transport of acid
gases in perfluorosulfonic acid membrane. J. Membr. Sci. 1989, 46, 309-
324.
(27) Matsuyama, H.; Teramoto, M.; Iwai, K. Development of a new
functional cation-exchange membrane and its application to facilitated
transport of CO2. J. Membr. Sci. 1994, 93, 237-244.
(28) Yamaguchi, T.; Boetje, L. M.; Koval, C. A.; Noble, R. D.; Bowman,
C. N. Transport properties of carbon dioxide through amine functionalized
carrier membranes. Ind. Eng. Chem. Res. 1995, 34, 4071-4077.
(29) Matsuyama, H.; Teramoto, M.; Sakakura, H.; Iwai, K. Facilitated
transport of CO2 through various ion exchange membranes prepared by
plasma graft polymerization. J. Membr. Sci. 1996, 117, 251-260.
(30) Quinn, R.; Laciak, D. V. Polyelectrolyte membranes for acid gas
separations. J. Membr. Sci. 1997, 131, 49-60.
(31) Matsuyama, H.; Terada, A.; Nakagawara, T.; Kitamura, Y.;
Teramoto, M. Facilitated transport of CO2 through polyethylenimine/poly-
(vinyl alcohol) blend membrane. J. Membr. Sci. 1999, 163, 221-227.
(32) Kim, T.-J.; Li, B.; Hagg, M.-B. Novel fixed-site-carrier polyviny-
lamine membrane for carbon dioxide capture. J. Polym. Sci., Part B: Polym.
Phys. 2004, 42, 4326-4336.
(33) Tee, Y. H.; Zou, J.; Ho, W. S. W. CO2-selective membranes
containing dimethylglycine mobile carriers and polyethylenimine fixed
carrier. J. Chin. Inst. Chem. Eng. 2006, 37, 37-47.
(34) Yi, C.; Wang, Z.; Li, M.; Wang, J.; Wang, S. Facilitated transport
of CO2 through polyvinylamine/polyethylene glycol blend membranes.
Desalination 2006, 193, 90-96.
(35) Zou, J.; Ho, W. S. W. CO2-selective polymeric membranes
containing amines in crosslinked poly(vinyl alcohol). J. Membr. Sci. 2006,
286, 310-321.
(36) Shil, G.; Ho, W. S. W. Synthesis and characterization of interfacially
polymerized membranes for CO2 separation. J. EnViron. Eng. Manage. 2006,
16, 233-241.
ReceiVed for reView June 8, 2007
ReVised manuscript receiVed November 12, 2007
Accepted November 15, 2007
IE070794R