See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228676182

Carbon Dioxide Separation through Polymeric Membrane Systems for Flue Gas
Applications

Article · January 2010

DOI: 10.2174/1874478810801010052

CITATIONS READS
140 13,547

3 authors:

Colin Scholes Sandra Kentish

University of Melbourne University of Melbourne
99 PUBLICATIONS   1,878 CITATIONS    351 PUBLICATIONS   9,960 CITATIONS   

SEE PROFILE SEE PROFILE

Geoff Stevens
University of Melbourne
538 PUBLICATIONS   10,380 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Gas separation by membrane technology View project

Solvent extraction - Pulsed extraction column View project

All content following this page was uploaded by Sandra Kentish on 01 June 2014.

The user has requested enhancement of the downloaded file.

52 Recent Patents on Chemical Engineering, 2008, 1, 52-66

Carbon Dioxide Separation through Polymeric Membrane Systems for

Flue Gas Applications
Colin A. Scholes, Sandra E. Kentish* and Geoff W. Stevens
Cooperative Research Centre for Greenhouse Gas Technologies, Department of Chemical and Biomolecular
Engineering, The University of Melbourne, VIC, 3010, Australia

Received: October 27, 2007; Accepted: November 6, 2007; Revised: November 15, 2007
Abstract: The capture and storage of carbon dioxide has been identified as one potential solution to greenhouse gas
driven climate change. Efficient separation technologies are required for removal of carbon dioxide from flue gas streams
to allow this solution to be widely implemented. A developing technology is membrane gas separation, which is more
compact, energy efficient and possibly more economical than mature technologies, such as solvent absorption. This
review examines the recent patented developments in polymeric based membranes designed for carbon dioxide separation
from mixed-gas systems. Initially, the background to polymeric membrane separation is provided, with an overview of
past polymeric designs. This is followed by a discussion on the current state of the art; in particular developments in
mixed matrix polymeric membranes and facilitated transport polymeric membranes for improved carbon dioxide
permeation and selectivity. Recent developments in other membrane types, carbon and inorganic, are reviewed for
comparison purposes with polymeric developments. Finally, a brief comment on the future directions of polymeric
membrane gas separation technologies is provided.
Keywords: Polymeric membranes, gas separation, carbon dioxide, mixed matrix membranes, facilitated transport membranes.

1. INTRODUCTION Initially, a brief introduction into membranes and perfor-

mance characterization is provided, followed by an overview
The control of anthropogenic carbon dioxide emissions is
of prior patented polymeric membrane designs. The review
one of the most challenging environmental issues facing
then focuses on the most recent polymeric membrane
industrialized countries, because of the implications to designs, mainly mixed matrix membranes and facilitated
atmospheric carbon dioxide levels and climate change [1].
transport membranes, which provide improved carbon
Burning of fossil-fuels is responsible for the majority of
dioxide separation over previous polymeric designs. Both
these carbon dioxide emissions [1], and therefore there is
membrane gas absorption and non-polymeric membranes are
significant interest in developing technologies that will
covered briefly in regards to their performance and recent
reduce carbon dioxide emissions. In particular, the capture of
advances for comparison purposes with polymeric designs.
carbon dioxide from large point sources allows storage
options, such as geo-sequestration, to reduce emission levels. The concept of membrane separation was originally
The conventional process for carbon dioxide capture is by proposed by Graham in 1866 [13] with Loeb and Sourirajan
reversible solvent absorption [2, 3]. In general, this process [14] developing the first anisotropic membrane in 1961.
has high energy consumption [4], for example, regenerating Initially, the majority of membrane separation research was
the solvent requires a high heating utility in the stripper directed towards reverse osmosis, which is extensively
reboiler. The associated cost and environmental impact covered in the patent literature [15, 16]. Gas separation
means that there is need for other more efficient separation membranes were first commercialized in 1977 when
processes to be applied to carbon dioxide capture. Monsanto/Perma released their hydrogen recovery system
[17]. The success of this and other gas membrane systems by
The energy efficiency and simplicity of membrane gas
Cynara, Separex and Generon [17-19] led to substantial
separation makes it extremely attractive for carbon dioxide
innovation during the 1980s and 1990s into membrane mate-
capture [5-7]. The ability to selectively pass one component rials. These innovations have improved the gas separation
in a mixture while rejecting others describes the perfect
efficiency and membrane durability, making membrane gas
separation device. While no membrane system truly behaves
separation commercially competitive with existing sepa-
this way, membrane gas separation does have a number of
ration technologies. Gas separation membranes are now
advantages over conventional processes and a number of
applied to a wide range of industrial processes. In particular,
reviews examining their benefits exist [8-12]. This review
many of the early patents were awarded for the sweetening
covers the recent patented advances in membrane design for of natural gas by the removal of carbon dioxide [20, 21] and
carbon dioxide capture, and more broadly for the separation
this is the foundation on which current carbon dioxide
of polar gases from multiple component gas streams. In
capture research is based.
particular, this review will focus on advances in polymeric
membrane design for improved carbon dioxide separation. Membranes act as filters to separate one or more gases
from a feed mixture and generate a specific gas rich
*Address correspondence to this author at the Department of Chemical and permeate, as shown in Fig. (1). Two characteristics dictate
Biomolecular Engineering, The University of Melbourne, VIC, 3010, membrane performance, permeability; that is the flux of a
Australia; Tel: +61 3 83446682; Fax: +61 3 83444153; specific gas through the membrane, and selectivity; the
E-mail: sandraek@unimelb.edu.au

1874-4788/08 $100.00+.00 © 2008 Bentham Science Publishers Ltd.

Carbon Dioxide Separation through Polymeric Membrane Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 53

Fig. (1). Schematic of membrane gas separation.

membrane’s preference to pass one gas species and not P = DS (1)

another. There are five possible mechanisms for membrane where P is the permeability coefficient, a measure of the flux
separation [18, 22]; Knudson diffusion, molecular sieving,
of the membrane (cm3(STP) cm-2 s-1 cmHg-1). The common
solution-diffusion separation, surface diffusion and capillary
unit of permeability is the barrer (10-10 cm3(STP) cm-2 s-1
condensation, of which the first three are schematically
cmHg-1). D is the diffusivity coefficient (cm2 s-1), the
represented in Fig. (2). Molecular sieving and solution-
mobility of molecules within the membrane and S the
diffusion are the main mechanisms for nearly all gas
solubility coefficient (cm3(STP) cmHg-1), which measures
separating membranes. Knudson separation is based on gas the solubility of gas molecules within the membrane.
molecules passing through membrane pores small enough to
prevent bulk diffusion. Separation is based on the difference For ideal gases, the permeability is related to the gas
in the mean path of the gas molecules due to collisions with permeation rate through the membrane (Q), the surface area
the pore walls, which is related to the molecular weight of the membrane (A), the thickness of the membrane (l) and
(Table 1). Specifically, the selectivity for any gas pair is the driving force for separation, the pressure difference
determined by the inverse ratio of the square root of their across the membrane (p):
molecular weight. For CO2 /N2 and CO2/H2 separation, P Q
Knudsen diffusion predicts a selectivity of less than unity. = (2)
Molecular sieving relies on size exclusion to separate gas l A
p
mixtures. Pores within the membrane are of a carefully The ideal selectivity (
) of one gas, A, over another gas,
controlled size relative to the kinetic (sieving) diameter of B, is defined as:
the gas molecule. This allows diffusion of smaller gases at a
much faster rate than larger gas molecules. In this case, the PA (3)

=
CO2/N2, selectivity is greater than unity, as CO2 has a PB
smaller kinetic diameter than N2. Surface diffusion is the
migration of adsorbed gases along the pore walls of porous Polymeric membranes are further classified as rubbery or
membranes [23, 24]. The rate of surface diffusion is glassy, dependent on operating temperature relative to the
determined by the level of interaction between the adsorbed glass transition temperature of the polymer [27]. Rubbery
gases and pore surface. Thus, molecules diffuse along the membranes, operating above the glass transition temperature,
pore walls relative to the strength of this interaction, and are able to rearrange on a meaningful time scale and are
separation is mainly achieved by the difference in the degree usually in thermodynamic equilibrium. Therefore, gas
of this interaction for the individual gases. An extension of solubility within the polymer matrix follows Henry’s Law
surface diffusion is when the vapour pressure becomes low, and is linearly proportional to the partial pressure, or
adsorbed gas can undergo partial condensation within the fugacity, ƒ:
pores. This condensed component diffuses more rapidly
through the pore than gases, causing separation of the
CD = K D ƒ (4)
condensable gas. This is known as capillary condensation Where CD is the concentration of gas in the polymer matrix
[25, 26]. and is proportional through the Henry’s Law constant (KD).
Polymeric membranes are generally non-porous, and Conversely, glassy membranes operate below the glass
therefore gas permeation through them is described by the transition temperature and therefore polymer rearrangement
solution-diffusion mechanism [18]. This is based on the is on an extraordinarily long time scale meaning the mem-
solubility of specific gases within the membrane and their brane never reaches thermodynamic equilibrium. Hence, the
diffusion through the dense membrane matrix. Hence, polymer chains are packed imperfectly, leading to excess
separation is not just diffusion dependent but also reliant on free volume in the form of microscopic voids in the
the physical-chemical interaction between the various gas polymeric matrix. Within these voids Langmuir adsorption
species and the polymer, which determines the amount of of gases occurs that increases the solubility. Therefore, the
gas that can accumulate in the membrane polymeric matrix. total concentration of absorbed gas within glassy membranes
The relationship between permeability, diffusivity and (C) can be described by [18]:
solubility can be described by:
54 Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 Kentish et al.

Fig. (2). Schematic representation of three of the different possible mechanisms for membrane gas separation, Knudsen diffusion, molecular
sieving and solution-diffusion.

Table 1. Molecular Weight (Da) and Kinetic Diameter (Å) of

Gases Encountered in Membrane Gas Separation

Molecule Molecular Weight Kinetic Diameter (Å)

CO2 44 3.3

O2 32 3.46

N2 28 3.64

H2O 18 2.65

CH4 16 3.8

H2 2 2.89

C = CD + CH (5)
Where CH is the standard Langmuir relationship
Fig. (3). Schematic representation of the relationship between the
C' H b ƒ
CH = (6) polymer specific volume and temperature in an amorphous polymer
(1 + b ƒ ) (reprinted with permission from [28]).

C’H is the maximum adsorption capacity, b is the ratio of

rate coefficients of adsorption and desorption, defined as: absorption separation [29]. The majority of this cost is
associated with creating the pressure difference across the
CH (7) membrane to drive separation, usually achieved through feed
b= gas compression [30]. To make gas membrane separation
(C' H
C H )ƒ
cost competitive for this CO2 content, it has been suggested
Hence, the dual-mode sorption for glassy membranes is that CO2/N2 selectivity needs to exceed 120 [29]. When the
written: CO2 content of the gas exceeds 20%, experienced in a range
of industrial processes such as cement production, the selec-
C' H b ƒ tivity needed reduces to >60. While these degrees of sepa-
C = KD ƒ + (8)
(1 + b ƒ ) ration appear not to be achievable with current commercial
membrane material [31], many of the recently patented
The relationship between the specific volumes and polymeric materials discussed here overcome these
temperature of a polymer from the glassy to rubbery state is benchmarks. Similarly, membrane permeability is inversely
shown in Fig. (3). proportional to the membrane area required for separation.
Currently, for a flue gas of 10% CO2 content, the cost of Thus high permeability leads to lower capital cost. However,
membrane separation post-combustion is greater than solvent for most membranes, there is a trade-off between selectivity
and permeability. A highly permeable membrane tends to
have low selectivity, and visa versa. Robeson [32] has
suggested that this trade-off may be represented as an upper
Carbon Dioxide Separation through Polymeric Membrane Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 55

bound to membrane performance. This upper bound can carbon dioxide flux of cellulose acetate based membranes
clearly be seen in Fig. (4), for a range of membranes invol- decreases substantially with time, due to the material being
ved in CO2/N2 separations. Overcoming this upper bound is susceptible to plasticisation and compaction under feed
the focus of many recently awarded patents in polymeric stream conditions. Therefore, later polymeric membrane
membranes, because achieving both high carbon dioxide patents focus on more robust polymers that achieve greater
permeability and selectivity is desirable. selectivity and/or permeability. Improving the performance
of the CO2-selective polymeric membrane is achieved by
two approaches; increasing the solubility of carbon dioxide
in the membrane through changes in polymeric composition,
and increasing the diffusion of carbon dioxide by altering the
polymer packing within the membrane. Diffusion is
generally enhanced by increasing the volume of free space
within the membrane and this can sometimes be achieved
through the addition of bulk substituent groups [9]. The
polymer packing in glassy membranes and thus the free
volume is also influenced by the casting method and
annealing conditions. Therefore, along with the polymeric
materials, casting methods have also been patented [35].
The combination of these approaches has produced a
wide range of polymeric membranes with reasonable
permeability and selectivity to provide good carbon dioxide
separation. Some patented polymeric membranes are based
on polyamides [36-44], polysemicarbazides [45], polycarbo-
nates [46-50], polyarylates [51, 52], poly(phenylene oxide)
[53-57], polyaniline [58-60] and polypyrrolones [61]. These
all have reasonable permeability and selectivity, with some
achieving performance around Robeson’s upper bound. The
difference between individual patents for each polymeric
system is through the addition of bulky substituent and
Fig. (4). Comparison of Robeson’s curve for CO2/N2 separation by functional groups to the polymer, as well as cast history.
carbon membranes (o) and polymeric membranes (•) [9, 33].
One of the most widely patented polymeric materials are
polysulfones [62-72]. They are regarded as among the most
Improving permeability and selectivity are not the only
membrane properties that are important. For membrane chemically and thermally durable thermoplastic polymers
available; and polysulfones have been extensively applied to
materials to be viable, they need to be thermally and chemi-
gas separation. More recently, polyimides based membranes
cally robust, resistant to plasticisation and aging affects to
have out performed polysulfones, displaying some of the
ensure continual performance over long time periods, and be
best permeability and selectivity properties for purely
cost effective to manufacture as standard membrane
polymeric membranes [73-95]. This coupled with their
modules.
thermal, chemical and plasticisation resilience, as well as
Patents have been awarded for a wide range of polymeric considerable mechanical strength makes them an attractive
based membranes that claim to meet these aims. Here, material for gas separation membranes. The performance of
original polymeric and inorganic membrane patents are these two polymers has resulted in a large number of patents.
reviewed briefly to inform the reader of past developments Differences between patents deal with substituent groups to
in the art, since the majority of these patents were awarded in change the carbon dioxide solubility, membrane packing
the 1980s and early 1990s. This review will then focus on density and free volume, as well as improving the membrane
recent novel approaches in polymeric membranes that resistance to harsh environments.
achieve separation performance above Robeson’s upper
More recent patents on purely polymeric membranes
bound and therefore are possibly more commercially com-
have focused on combining different polymers to produce
petitive than present membrane gas separation technologies.
composite polymeric membranes. The copolymers used
2. MEMBRANE TECHNOLOGY - STATE OF THE generally have a glassy (hard) polymer segment and a
ART rubbery (soft) polymer segment [9]. The hard segment forms
the structural frame and provides the mechanical support.
2.1. Polymeric Membranes
The rubbery segment generally forms continuous microdo-
The first carbon dioxide separating membranes were mains within the membrane and the flexible nature of the
based on cellulose acetate and derivatives thereof, originally structure allows the transportation of gas, hence greater
designed for reverse osmosis [14, 34]. These polymeric permeability. The idea is to combine the selectivity of one
membranes are characterized by a thin, dense selective polymer with the permeability of the other to provide a better
surface ‘skin’ on a less dense porous support that is non- performance membrane. Again, a considerable number of
selective. This is the basis of all asymmetric membrane patents have been awarded for copolymer membranes, with
materials that are readily used in industry. However, the the best performance based on blends of polyimide [77, 96-
56 Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 Kentish et al.

100]. These copolymer blends have improved separation molecules (Table 1) dictates which molecular sieves are
performance over polyimides, while retaining high mecha- useful and provides an indication of selectivity [17]. For
nical, thermal and chemical stability. example, Zeolite 3A separates hydrogen effectively from
hydrocarbon feeds because the pore diameters are ~3 Å
2.2. Inorganic Membranes [109]. Zeolite 4A has pore sizes of ~ 4 Å, which will
Inorganic membranes represent an alternative gas separate carbon dioxide from nitrogen and methane.
separation technology; however it is not the purpose of this Molecular sieving based inorganic membranes incorpo-
review to cover these membranes in detail. There are two rate such materials into the pores of a porous support. This
major categories of inorganic membranes, porous and non- integrates the size exclusion properties of the molecular
porous. Non-porous membranes are generally used in highly sieve within the pores, providing selective gas separation
selective separation of hydrogen, where transportation is [110]. A major difficulty with this approach is to achieve an
through alloys of palladium [101] or oxygen through perovs- efficient trap of the sieve particles within the pores. Both the
kite systems [102-104]. Porous inorganic membranes are pore and sieve sizes must be well-known and sufficiently
generally cheaper but less selective. The attraction of inor- similar to ensure strong packing of the pores. Otherwise, the
ganic systems is their ability to operate at high temperatures. gaps between the sieves and the pore walls will allow bulk
A particular high temperature operation of interest is the diffusion of gas molecules, diminishing selectivity
separation of carbon dioxide from hydrogen in syngas performance [111].
processes.
3. RECENT DEVELOPMENTS IN POLYMERIC
The large size of carbon dioxide relative to hydrogen
MEMBRANE SYSTEMS
means that achieving a carbon dioxide rich permeate stream
by simple molecular sieving is not possible Table 1. The current research into polymeric membranes is
Inorganic membranes overcome this by functionalizing the focused on increasing the polymeric performance above
pores of the membrane to increase the carbon dioxide Robeson’s upper bound to become cost competitive with
loading. Permeability then becomes a function of surface solvent absorption, as well as improving the resilience of the
diffusion. Recent patents describe inorganic membranes that membrane material. McKeown et al. [112] disclose the
consist of a ceramic support, such as Al2O3, onto which a formation of microporous polymeric structures with very
porous separating layer is added, often silica, alumina or high free volumes. These polymers of intrinsic microporosity
zirconia. To this a functional layer is added, BaTiO3 by Ku et (PIMs) are rod-like randomly contorted structures which
al. [105] or MgO by Gobina [106]. These functional groups generate intrinsic cavities within the membrane. They exhibit
have a high chemical affinity for carbon dioxide and behaviour analogous to conventional molecular sieves, but
therefore the pore walls become saturated, which increases have greater solubility.
the permeability, as shown in Fig. (5). For BaTiO3 on
Young et al. [113] detail the cross-linking of polybenzi-
alumina, for 5 nm pores a CO2/H2 selectivity of 3.1 is
midazole for improved mechanical properties. Unmodified
achieved at 500oC and for 1 nm pores a selectivity of 18.4 is
polybenzimidazole has been covered in past patents and
achieved. Similarly for MgO, a CO2/N2 selectivity of 120 is provides reasonable permeability and selectivity for carbon
obtained with a carbon dioxide permeability of > 0.02 barrer
dioxide separation from methane [114, 115]. Cross-linking
at 350oC. These membrane performances approach Robe-
of the polymer improves the mechanical properties, by
son’s upper bound for carbon dioxide separation and are
increasing the yield stress of the membrane, which also
therefore compatible with polymeric membranes.
results in increased separation performance. The patented
example has a carbon dioxide permeation of 7.9 barrer with a
selectivity over nitrogen of 27, when the polymer is cross-
linked, compared to a carbon dioxide permeability of 0.3
barrer with a selectivity of 18 for unmodified linear poly-
benzimidazole, at 23oC tested by single gases. The effect of
cross-linking on performance is dependent on the linking
agent.
Wang and Yeager [116] present solvent resistant
polymeric membranes that can reduce the plasticisation
effects of hydrocarbons in the feed stream. The patent covers
Fig. (5). Schematic of inorganic membrane operation through polyketone, polyether ketone, polyarylene ethere ketone,
Knudsen and surface diffusion (adapted from [107]). polyimide, polyetherimide and polyphenylene sulphide,
which have intrinsic solvent inertness and can therefore
Molecular sieves exist in both natural and synthetic withstand organic rich operation conditions. Ekiner and
forms, and can be classed as zeolites (aluminosilicate com- Simmons [117] describe the controlled annealing of hollow
positions), or non-zeolites, such as aluminophosphates, fiber polyimide membranes. The annealing conditions
silico-aluminophosphates and silica [108]. The molecular patented are for the polyimide based membrane to exist
sieve framework forms a well-defined repeating structure of under a vacuum less than 15 inches of mercury at between
regular channels and cages. Gas separation is dependent on 100-250oC for 6-30 hours. This controlled annealing
the size of these channels and cages relative to the kinetic improves chemical resistance and ensures that the resulting
diameter of the gas. The difference in kinetic diameter of gas hollow fibers have the necessary mechanical strength for
Carbon Dioxide Separation through Polymeric Membrane Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 57

high pressure and temperature applications. However, results in a doubling of selectivity, with a loss in
annealing conditions for polymeric membranes are relatively permeability. This arises because the compatibilizer prevents
well known and therefore the novelty of some of the patent’s phase separation within the membrane and hence reduces the
claims is questionable. Weinberg [83] deals with aromatic- free volume. Too much PSMA in the membrane leads to
polyimide based membranes that have improved perfor- itself phase separating, which affects both selectivity and
mance at low temperature. This allows the membrane to be permeability to the detriment of performance Table 3.
utilized in cold temperature applications without the need for
heating of the feed gas. The improved performance relies on 3.1. Carbon Membranes
the increased solubility of carbon dioxide at these Carbon membranes are a variant of polymeric memb-
temperatures being greater than the loss in diffusivity. The ranes that operate as molecular sieves [119]. These
patented example of asymmetric hollow fibers of this membranes generally have higher permeability and selec-
aromatic polyimide can be seen in Table 2. However, none tivity for carbon dioxide than polymerics, Fig. (4). However,
of these recent patents are able to achieve a selectivity and the carbon structures are not easily formed into thin layers
permeability combination that exists above Robeson’s upper and so while the permeability may be high, the actual flux
bound. through the thick membranes may not. The membranes are
Table 2. The Temperature Dependent Permeability (Barrer) fabricated by pyrolyzing organic materials, typically
and Selectivity of an Aromatic-Polyimide Based Mem- polymers, in a non-oxidizing atmosphere upon a structural
brane, for a CO2/N2 Gas Mixture of 72:28 Mole Ratio support that can withstand the high temperatures. The
at 448 kPa [81]. resulting membrane is comprised almost entirely of carbon
and manifests as an active surface with controlled pore
structure. These pore structures allow for effective separation
Temperature performance of novel aromatic-polyimide membranes
of composite gas mixtures through both a molecular sieve
-22oC 10oC 50oC mechanism, as well as by surface diffusion. The choice of
the organic precursor material is important, because this will
CO2 Permeability (barrer) 0.232 0.147 0.091 influence pore size and structure, and hence determine the
N2 Permeability (barrer) 0.0035 0.0045 0.0053 performance of the membrane. Suitable precursors include
resins, surfactants, graphite and polymers [120]. Their
CO2/N2 66.3 32.7 17.1 advantage over polymeric membranes is that they can with-
stand high temperatures and are more durable. Therefore,
carbon membranes are often employed in chemical reactors
Two recent patents have been awarded for polymeric and catalytic systems. Different production processes for
blends. The patent awarded to Ekiner and Simmons [100] carbon membranes have been covered [121-124]. These
covers polyamide, polyimide and polyamide-imide blends, patents generally involve heating of the organic material
which have increased mechanical strength and chemical above its decomposition temperature in an argon or nitrogen
resistance over the pure polymeric membranes, with only rich environment.
slight changes in permeability and selectivity. Seo [118]
describes semi-crystalline amorphous polymeric blends with Carbon membranes can be relatively brittle and this can
a compatibilizer agent within the membrane matrix to ensure lead to cracking, greatly diminishing their performance
miscibility of the two polymers. The patent examples of [119], as well as making fabrication of carbon based
poly(phenylene oxide) with Nylon 6 are provided in Table 3. membrane modules difficult. Furthermore, differences in the
Poly (phenylene oxide) has been patented previously as a gas thermal expansion coefficient with the support can
separation membrane material, with reasonable selectivity exasperate cracking at high temperature or if the membrane
and permeability. The addition of the semi-crystalline Nylon experiences thermal cycling. This increases the associated
6, results in a decrease in permeability for only a slight cost of the membrane modules compared with polymeric
increase in selectivity. However, addition of the membranes. The recent patent of Foley, Rajagopalan and
compatibilizer poly(styrene-co-maleic anhydride) (PSMA) Merritt [125] strengthens carbon membranes by incor-

Table 3. Permeability (Barrer) and Selectivity of Poly (Phenylene Oxide) (PPO) Membranes with Nylon 6, with the Addition of a
Compatibilizer, Poly(Styrene-co-Maleic Anhydride) (PSMA), for Single Gases at 35oC and 1 atm [118].

Separation performance of poly (phenyl oxide) nylon 6 blends

Compatibilizer Permeability CO2 Permeability N2 Selectivity CO2/N 2

PPO 61 4.1 14.4

PPO - Nylon(20%) 24.8 1.53 16.15

PPO - Nylon (20%) PSMA (2wt%) 12.0 0.309 38.7

PPO - Nylon(20%) PSMA (4wt%) 17.0 0.474 35.87

PPO - Nylon (20%) PSMA (6wt%) 12.9 0.469 27.43

58 Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 Kentish et al.

porating particulate matter into the membrane. Having

particulate matter inside relieves stress on the carbon
structure, hence improving the mechanical strength and
reducing the possibility of cracking. Furthermore, there is
evidence that if the particulate matter is selected with care,
enhanced permeability can also be achieved. The patent
covers a wide range of suggested materials, ranging from
oxides, nitrides, carbides, oxynitrides, oxycarbides, as well
as nanoparticles of Si, Al, Ti, Zr, Fe and the different forms
of carbon. These are doped in the membrane at the polymeric
precursor preparation stage. High selectivity for H2/CO (21)
and for H2/CO2 (5) makes the membrane applicable for
syngas separation and hydrogen production through the Fig. (6). Schematic of mixed matrix membranes hollow fiber, with
water-shift reaction of syngas [126]. the zeolite particulate matter within the polymeric phase (adapted
from [141]).
Carbon membranes are also known to suffer from aging
effects, with a corresponding loss in permeability and selec- fones, polyetherimides and polyimides. The particle size of
tivity. These aging affects differ from polymeric membranes,
the molecular sieve within the membrane influences perfor-
and involve pore filling, cracking, and chemical degradation
mance, with average dimensions of 200-900 Angstroms
of the carbon matrix [118]. In particular, exposure to oxygen
suggested [130] to prevent the membrane thickness beco-
causes the surface to become hydrophilic, and after this any
ming too great to operate effectively. This is in view of some
exposure to water results in a loss of performance. Similar to
of the early mixed matrix patents quoting thickness of the
polymeric membranes, various techniques have been order of 100-1000 micrometers [131], which are too thick for
proposed to regenerate the membrane, including thermal,
effective membrane modules.
chemical, electrothermal, ultrasonic or microwave processes.
Hagg and Lie [127] deal with methods of increasing the An example of molecular sieves in mixed matrix
performance of the membrane at the same time as improving membranes is covered by Hasse et al. [132] (Table 4). The
the regenerative properties. The patent focuses on memb- inclusion of Zeolite 4A into the membrane results in a loss of
ranes formed from mixtures of cellulose and hemicellulose, permeability and selectivity compared to the polymer-only
with an acid used to degrade the cellulosic structure before membrane. The cause of this is poor association between the
pyrolysis to oligo- and monosaccharides. Acid pre-treatment zeolite and the polymer leading to an increased in free
for 6 days is ineffective, but after 2 weeks of acid pre- volume within the membrane. For mixed matrix membranes
treatment a critical threshold in the hydrolysis process is to perform effectively, there needs to be a high affinity
overcome, and separation performance improves. The between the sieve and the polymeric phase. Otherwise,
examples quoted for permeability and selectivity are above during casting the polymer will dissociate from the sieve’s
Robeson’s bound. Furthermore, this patented membrane surface leaving micro-cavities throughout the membrane.
design has improved regeneration properties, taking This will enhance the permeability of gas and limit the
advantage of the conductive nature of carbon membranes for sieving mechanism. Therefore, many of the patents in this
electrothermal regeneration. field deal with functionalizing molecular sieves to have high
affinity with the polymeric phase.
3.2. Mixed Matrix Membranes
KFI is an aluminosilicate molecular sieve (pore size 3.9
A relatively recent advance that takes advantage of Å) that has a greater affinity for poly vinyl acetate. This
inorganic and polymeric membrane approaches is mixed reduces the micro-cavities in the mixed membrane and
matrix membranes. These are characterized by a hetero- enhances the selectivity of the membrane through size
geneous gas separation layer comprising a dispersed phase of exclusion. Chabazite molecular sieves of differing forms (H-
discrete inorganic particles in a continuous polymeric phase ZK-5, Na-SSZ-13, SAPO-34 and SAPO-44) also associate
[128, 129]. The inclusion of dispersed particles can have well with the poly vinyl acetate, controlling micro-cavity
three possible effects on the permeability of gases; the formation within the membrane and therefore the permea-
discrete particles can act as molecular sieves, altering bility of nitrogen.
permeability in relation to molecular size, the particles can
disrupt the polymeric matrix resulting in increased micro- To improve the affinity of the particles for the continuous
cavities and hence increase permeability, or they can act as a phase, other chemically grafted zeolites have been patented
barrier to the gas transport and reduce permeability. A [133]. Common binding agents are the monofunctional
possible arrangement within a hollow fiber is shown in Fig. organosilicon compounds [130, 134]. These bond to the
(6). The mixed matrix membranes provide the opportunity to zeolite surface with the silicon segment miscible with the
overcome the individual deficiencies of molecular sieves and continuous polymer phase. This is covered by Kulkarni et al.
polymers, and achieve carbon dioxide separation perfor- [135], with carbon dioxide selectivities provided in Table 5.
mances above Robeson’s upper bound. The treated Na-SSZ-13 in polyvinylacetate has a 22%
increase in selectivity over the non-treated Na-SSZ-13
The continuous phase of the mixed matrix membrane can zeolite because the aminopropyldimethylethoxysilane (APD-
be almost any polymeric material; examples are polysul- MS) binding agent increases association with the polymeric
matrix and hence reduces micro-cavity formation.
Carbon Dioxide Separation through Polymeric Membrane Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 59

Table 4. Carbon Dioxide Permeability (Barrer) and Selectivity in Mixed Matrix Membranes Based on Polyvinylacetate Within the
Inclusion of Various Zeolites, Measured for Single Gases at 35oC and 410 kPa [132].

Separation performance of mixed matrix membranes based on poly vinyl acetate

Continuous Phase Zeolite Conc. (pph) CO2 Permeability Selectivity (CO2/N 2)

Polyvinyl acetate 3.1 34.7

Polyvinyl acetate 4A 15 2.4 30.7

Polyvinyl acetate KFI 20 4.9 53.6

Polyvinyl acetate H-ZK-5 15 4.9 41.0

Polyvinyl acetate Na-SSZ-13 15 4.5 41.7

Polyvinyl acetate SAPO-34 15 4.4 44.4

Polyvinyl acetate SAPO-44 15 4.9 51.8

Silicalite, is a common non-zeolite hydrophobic crystal- Table 6. The Carbon Dioxide Permeability (Barrer) and Selec-
line silica. Grose and Flanigen [136] show that treating such tivity of Liquid-Phased Based Mixed Matrix Memb-
a non-zeolite with organosilicon compounds also improves ranes Made of Silicone Rubber on a Porous Poly-
the selectivity of a polyvinyl acetate based mixed matrix sulfone Support, with the Inclusion of Poly Ethylene
membrane through the reduction of micro-cavities, in this Glycol (PEG), Activated Carbon or Carbonate to
case by 33% (Table 5). Improve Separation Performance, Measured for
Single Gases at Ambient Temperature
Table 5. Gas Selectivity of Mixed Matrix Membranes Based on
Polyvinylacetate with Various Molecular Sieves
Included, Measured for Single Gases at 35oC and 410 Phase Particle CO2 Selectivity
kPa [135]. Permeability (CO2/N2)

Silicone 14.3 11
Continuous Molecular Sieve Selectivity Rubber
Phase (CO2/N2)
Silicone Polyethylene Glycol 4.89 42
Poly vinyl acetate Na-SSZ-13 34.7 Rubber

Poly vinyl acetate APDMS treated Na-SSZ-13 51.2 Silicone Activated Carbon 29.4 15
Rubber
Poly vinyl acetate Silicalite 32.3
Silicone Polyethylene Glycol + 24.2 47
Poly vinyl acetate APDIPS-treated Silicalite 36.9
Rubber Activated Carbon
Poly vinyl acetate APDMS-treated silicalite 43.1
Silicone Activated Carbon + 16.2 20.1
Rubber carbonate

Additional treatments of molecular sieves can be under- Silicone PEG + Activated 14.3 40.5
taken to improve performance, with washed zeolites covered Rubber Carbon + carbonate
in the patent awarded to Kulkarni et al. [137]. Washing
zeolites in aqueous ionic solutions alters their operating pH
and surface functionality, in this case making a more basic increase in permeability of 70% and selectivity by a factor of
structure that will therefore increase the solubility of carbon 4 is achieved, compared to previous silicone rubber memb-
dioxide within the mixed matrix. ranes on porous and non-porous supports [140].
Mixed matrix membranes based on polymers with Marand and Kim [141] cover the inclusion of meso-
intrinsic porosity (PIMs) as well as rubbery continuous porous silica into mixed matrix membranes. The patent
phases have also been patented [138, 139]. Kulprathipanja example is based on polysulfone for the polymeric phase,
and Charoenphol [138] add activated carbon to increase the with obtained permeabilities and selectivities provided in
selectivity of silicone rubber by surface diffusion. The Table 7. The inclusion of mesoporous silica MCM41 and
activated carbon has a particle size between 0.1 and 5 MCM48 into polysulfone increases the permeability of
microns. Improvements in both the selectivity and permea- carbon dioxide with slight changes to the selectivity.
bility are shown in Table 6. The inclusion of polyethylene Increasing the weight of the silica substantially increases the
glycol acts as a plasticizer on the silicone rubber and permeability of carbon dioxide because of the increase in
therefore assists the emulsion with activated carbon. An micro-cavities within the membrane, resulting in a small loss
in selectivity. Functionalizing the silica with amine groups
increases the association with the polysulfone phase and
60 Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 Kentish et al.

Table 7. Permeabilities (Barrer) and Selectivity of Polysulfone Based Mixed Matrix Membranes with Mesoporous Silica [141]

Membrane Mesoporous silica wt% CO2 Permeability N2 Permeability Selectivity CO2/N 2

Poly sulfone 0 4.46 0.18 24.8

MCM41/ poly sulfone 20 7.59 0.30 25.3

MCM41/ poly sulfone 30 22.93 0.98 23.4

Amine-MCM41/ poly sulfone 20 7.25 0.25 29

MCM48/ PSF 10 8.45 0.32 26.4

MCM48/PSF 20 18.21 0.77 23.6

therefore reduces the number of micro-cavities present, this

produces an increase in selectivity as nitrogen permeability
is reduced because of size exclusion. The relative
performance of some of these mixed matrix membranes can
be seen in Fig. (7). In addition, Freeman et al. [142] incor-
porates various nanoparticles into a range of polymeric
membrane to increase the performance, an example is shown
in Fig. (8). Increasing the weight percentage of nanoparticle
corresponds with an increase in permeability, at the expense
of selectivity, again due to increased micro-cavity formation.
However, the patent examples show this is not true for all
mixed matrix membranes.

Fig. (8). Variation in mixed matrix membrane performance with

change in the amount of MgO nanoparticles present within a poly
(1-trimethylsilyl-1-proyne) (PTMSP) based membrane (reproduced
from [142]).

[144]. The resulting membranes showed mixed performance,

Matrimid based mixed matrix hollow fibers have no
enhancement in selectivity over just Matrimid hollow fibers
[145]. In contrast, P84 polyimide has improved CO2/N 2
selectivity with SAPO-34 and carbon molecular sieves, 20.2
and 25.5 respectively, in hollow fiber modules.
Mixed matrix membranes containing molecular sieves
can foul rapidly due to the accumulation of condensable
impurities, most importantly water [138, 146, 147]. This
hinders the molecular sieve mechanism and greatly reduces
the selectivity of the membrane. Much more research in
Fig. (7). Carbon Dioxide permeability against selectivity relative to
mixed matrix membranes will continue in the future, with
nitrogen of mixed matrix membrane examples presented here, grey,
the type of the dispersed particles of most interest. Recent
and polymeric membranes, black [9], with Robeson’s upper bound
publications in the literature are focused on the inclusion of
on performance.
nanoparticles and it would be expected that future patents
will be awarded in this area.
The conversion of mixed matrix membrane designs to 3.3. Facilitated Transport Membranes
asymmetric membrane systems, the preferred form for
industrial use, is covered by Kulkarni and Hasse [143], based Facilitated transport membranes rely on a chemical
on the asymmetric method proposed by Ekiner and Fleming reaction occurring between the gas of interest and a
component of the membrane (carrier). The reacted species is
Carbon Dioxide Separation through Polymeric Membrane Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 61

readily carried across the membrane, whereas diffusion of carbon dioxide, of which arsenite salts are suggested to be
non-reactive gases is inhibited, see Fig. (9). The active the most practical. With the inclusion of carbonate in these
carrier is generally basic in nature, given that carbon dioxide membrane systems, as well as the catalysts, the permeability
is acidic. The driving force for gas transportation remains the and selectivity are dramatically increased. The difference in
partial pressure difference across the membrane; however the carbon dioxide permeability and selectivity can be compared
facilitator carrier increases both the permeability and in Table 8.
selectivity of the membrane through the increased loading.
The immobilized liquid nature of the facilitator
The facilitator carrier can be either fixed-sited within the
membrane poses practical problems, such as leakage and
polymeric matrix or mobile. An illustrated schematic of a
evaporation of water, as well as loss of the facilitator through
fixed-sited carrier, polyvinylamine, in operation is provided
degradation. Hence, their performance over long periods of
in Fig. (9). time presents problems for large scale applications. Much of
the recently patented facilitated transport membranes have
focused on reducing these aging effects. In this regard,
Sirkar et al. [151] disclose the use of dendrimers as carriers.
The use of an amine-rich dentritic structure such as the
Starburst polyamidoamine (PAMAM) [152] leads to high
CO2 solubiltiy. These structures can be immobilised in a
polymeric matrix or alternatively, dissolved in a CO2-philic
liquid solvent which is placed within a mesoporous subs-
trate.
Recent patents have been awarded for the incorporation
of fixed carriers into existing porous polymeric supports,
generating a composite structure [153, 148]. This is then
swollen by the addition of water vapour, facilitating the
transfer of carbon dioxide along the polymer chain through
the membrane. Methods differ for the casting and cross-
linking of the fixed carrier layer, as well as attachment to the
porous polymeric support, with the desired result improving
mechanical and thermal properties of the composite layer to
increase stability of the membrane [153, 148]. Use of a
Fig. (9). Schematic of fixed carrier facilitator transport membrane polyvinylamine selective layer on a porous poly sulfone
for transport of carbon dioxide (reprinted from [148]). support, when polymer cross linking occurs by glutaral-
dehyde, results in a carbon dioxide permeability of 4100
The first facilitated transport membranes patent was barrer and a CO2/CH4 selectivity of 70, at room temperature
awarded to General Electric in 1967 [149, 150]. This related for a CO2/CH4 gas mixture. In contrast, fluorine based cross-
to cellulose acetate films swollen by the inclusion of an linkers, such as NH4F, provide fixed facilitator layers with
aqueous carbonate solution. Carbon dioxide readily dissolves greater permeability (3095 Barrer) and CO2/CH4 selectivity
and reacts with water to form the bicarbonate anion. (1143) [148]. A possible reason for this is the fluorine anion,
being considerably basic, allows increased carbon dioxide
H2O + CO2 (aq)
H2 CO3 (aq)
H + (aq) + HCO-3 (aq) loading within the membrane. The molecular weight of the
Carbonate acts as a carrier by increasing the amount of fixed facilitator polymer influences the selectivity but not the
carbon dioxide absorbed: permeability of carbon dioxide in the membrane, as shown in
Fig. (10).
CO32- (aq) + CO2 (aq) + H2O
2HCO-3 (aq) (+ Heat)
More recent patents have focused on hydrogel films from
Hence, this reaction occurs on the feed side of the cross-linked hydrophilic polymers, such as polyvinyl alc-
membrane and the bicarbonate anion transports through to ohol, polyvinylacetate, polyvinyl pyrrolidone, polyethylene
the permeate side, where the reverse reaction occurs and oxide, polyacrylamide, bends and copolymers, cast on
carbon dioxide is released. This patent also allows for the permeable supports [154, 155]. These hydrogels have a high
inclusion of catalysts into the facilitator membrane to water absorbing power and therefore can accommodate
increase the rate of reaction between the carbonate and

Table 8. Wet Cellulose Acetate Facilitated Membranes for Carbon Dioxide Separation, at Ambient Temperature and 4 kPa [149]

Facilitator CO2 Permeability (barrer) Selectivity (CO2/O 2)

Cellulose Acetate Pure water 400 22

Cellulose Acetate 2 N KHCO3 500 78

Cellulose Acetate 2 N KHCO3 + 0.5 N NaAsO2 (catalyst) 2000 600-800

62 Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 Kentish et al.

carbonate combination provides permeability and selectivity

well above Robeson’s upper bound, Fig. (11), making this
membrane an attractive option.
Patented examples of fixed and mobile carrier
combinations in polyvinylalcohol (PVA) are provided in
Table 10. The selectivity and permeability of the membrane
is unaffected by cross-linking with formaldehyde, and
therefore the mechanical strength can be increased using this
approach without penalty. The hydroxide anion present
ensures the immobilized liquid is basic and therefore
increases the carbon dioxide loading. This is most obvious
for the dimethylgylcine - polyethylene system, where the
selectivity for carbon dioxide against hydrogen is tripled
when hydroxide is present. The higher the weight percentage
of fixed carrier within the membrane, the greater selectivity
is achieved with the permeation of carbon dioxide remaining
Fig. (10). Influence of the molecular weight of the fixed carrier on
the same [154], similar to that observed for simple fixed
carbon dioxide selectivity over methane (reproduced from [148]).
carrier membranes. The type of mobile carrier present infl-
uences the permeability of carbon dioxide, with amino-
considerable loading of the carrier species. Copolymer isobutyric acid better than dimethylglycine and glycine.
composite membranes take advantage of the different However, the relationship between mobile carrier concen-
swelling potential of each polymer component, whereas tration and carbon dioxide permeability is not clear in these
cross-linking agents are used to generate a strong polymer mixed systems.
matrix. As a consequence, even though the membranes are
thin, the high water content hydrogel film can retain its As previously mentioned, dehydration of facilitated
shape upon being subjected to pressure. Hence, they can membranes is a major concern, and the cross linking of the
function as a separation membrane with a long service life polymer is believed to minimise water loss. For the fixed
by exhibiting good water- retention and weatherability [156]. carrier polyvinylamine facilitated membrane, exposure to dry
feed gas for 5 days decreased the carbon dioxide
Fixed carriers are generally polyamines, examples are permeability by 20% whereas exposure for 35 days resulted
polyallylamine, polyethylenimine and polyvinylamine. in a 90% permeability decline, as shown in Fig. (12) [153].
Mobile carriers are basic compounds, and are often a combi- The flux was restored when the membrane was re-wetted
nation of hydroxide salts, organic ammonium salts, amino- implying that permeability loss is due to evaporation, rather
acids, carbonates, alkanolamines and polydentate ligands, than carrier degradation. Similarly, with a simple mobile
such as EDTA [154]. carbonate carrier, the permeability decreased by 35% after
Carbon dioxide selectivity and permeability through 30 days with a similar loss in selectivity [153]. In industrial
various mobile carrier facilitator membranes are provided in applications, the performance loss would be lower, since real
[157,158] Table 9, based on patented examples. Correlation flue gases are saturated with water and this would keep the
between mobile carriers is difficult because of the different facilitated membranes hydrated to some degree.
chemistry and kinetics involved. However, carbonate The highly selective nature of both facilitated and mixed
appears to provide very high carbon dioxide permeabilities matrix membranes means that a combination of the two may
through the membranes. Interestingly, with the addition of be advantageous. Kulprathipanja and Charoenphol [156]
other mobile carriers the permeability is reduced. However, describe such a combination, with a mixed matrix membrane
as a trade off, having mixed mobile carriers with carbonate of polyethylene glycol in silicone rubber, within which
improved the selectivity against nitrogen. The EDTA -

Table 9. Carbon Dioxide Permeability (Barrer) and Selectivity in a Range of Mobile Carrier Facilitated Transport Membranes.
Test Gases are Mixtures of 90% N2 10% CO2 at 98 kPa or 75% H2 25% CO2 at 300 kPa, at Ambient Temperature

Facilitator CO2 Permeability Selectivity Selectivity

(CO2/N 2) (CO2/H2)

Polyvinylalcohol [157] Ethylene diamine (50 %wt) 161 26.1

Polyvinylalcohol [157] Aminoisobutyric acid – ethylene 67 14.3

diamine (50 %wt)

Vinyl alcohol - acrylate copolymer [158] 2 M K2 CO3 6100 317

Vinyl alcohol / acrylate copolymer [158] 2 M K2 CO3 and 0.05 M 18-crown-6 740 670

Vinyl alcohol / acrylate copolymer [158] 2 M K2 CO3 and 0.05 M EDTA 2400 1417
Carbon Dioxide Separation through Polymeric Membrane Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 63

Table 10. Carbon Dioxide Permeability (Barrer) and Selectivity in a Range of Mobile, Fixed and Combined Carrier Facilitated
Transport Membranes, for Gas Mixtures of 40/20/40 or 75/25/0 of H2, CO2 with N2, at 200-300 kPa and Ambient
Temperature

Facilitator CO2 Selectivity

Permeability (CO2/H2)

Polyvinyl alcohol [155] Glycine (50 %wt) - polyethylenimine (25 %wt) 194 28

Polyvinyl alcohol (cross linked Glycine (50 %wt) - polyethylenimine (25 %wt) 186 31
formaldehyde) [155]

Polyvinylalcohol (cross linked) [155] Dimethylglycine (23.6 %wt) - polyethylenimie (23.6 %wt) 602

Polyvinylalcohol (cross linked (23.6 % wt) Dimethylglycine - polyethylenimine (23.6 %wt) - KOH (6.4 %wt) 338 1782
formaldehyde) [155]

Polyvinylalcohol (cross link) [155] Aminoisobutyric acid (27.2 %wt) - polyallylamine (10.1 %wt) - KOH (16.8 %) 6196 262

Polyvinylalcohol (cross link) [155] Aminoisobutyric acid (19.6 %wt) - polyallylamine (9.8 %wt) - KOH (18.1 8278 170
%wt)

Fig. (12). Changes in facilitated transport membrane flux over time,

at the 35 day mark the feed is altered to be wet, which rehydrates
the membrane (reproduced from [148]).

on the feed side with a molten carrier, such as potassium

carbonate, diffuses through the molten carbonate channels
within the solid oxide structure, and is released on the
permeate side. The corresponding oxide anion is transferred
back through the oxide support, schematically shown in Fig.
(13). This novel design allows the facilitator mechanism to
Fig. (11). Carbon Dioxide permeability against selectivity relative operate at high temperatures, above the melting point of the
to nitrogen of facilitated transport membrane examples presented carrier, for carbonates 600-900oC. Thus, facilitator memb-
here, grey, and polymeric membranes, black [9], with Robeson’s rane separation can be applied to processes such as partial
upper bound on performance. combustion recycle, or syngas production, which are not
available to polymeric membranes. While, the quoted
inorganic particles (0.1 - 200 um) are dispersed. Examples selectivity of 5, with values up to 500 possible, are small
are activated carbon, zeolites, alumina and silica. The mobile compared to previously mentioned facilitated membrane
facilitator is carbonate and improves the selectivity of the patents, the advantage of operating at high temperature is
membrane. attractive and more patents in this area will be seen over the
coming years.
One of the restrictions of facilitated transport membranes
is their inability to operate at high feed temperatures due to 3.4. Membrane Gas Absorption
evaporation problems. Indeed in the mixed carrier memb-
Membrane gas absorption (MGA) is a hybrid of
ranes performance drops off substantially above 100oC due membrane and solvent separation that seeks to exploit the
to evaporation losses which shuts down the transport process
advantages of both processes [160-162]. MGA involves the
within the membrane. Lackner et al. [159] overcomes this by
transfer of carbon dioxide through a non-selective membrane
using a molten carrier phase within a porous solid oxide
before chemically absorption into a solvent, Fig. (14). The
structure. In this non-polymeric system carbon dioxide reacts
polymeric membrane facilitates a controlled flow of gas into
64 Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 Kentish et al.

Fig. (14). Schematic of Membrane Gas Absorption for carbon

dioxide separation.

sions of the pore are dependent on the operating character-

istics, such as pressure difference across the membrane, and
typically range from 0.01 to 0.1 micron [168, 169]. Some
examples of patented polymeric membranes for MGA are
Fig. (13). Schematic of molten carbonate facilitated transport porous polysulfone, polypropylene, Teflon, polyethylene,
membrane for high temperature [159]. polyethersulfone, polyethene and polypropene [170, 171].
4. CURRENT & FUTURE DEVELOPMENTS
the solvent and provides high gas-liquid contact surface area.
This physical separation of the liquid and gas flows elimi- As has been shown, the current trend in polymeric
nates foaming and channelling problems that can occur in membranes is to incorporate an additional agent into the
classical solvent absorption processes. polymeric phase to improve separation performance. This
can be another polymer (polymeric blends), particulate
Early patents in MGA studied the removal of acid gases matter (mixed matrix membranes) or a carrier molecule
[163, 164], and this has been developed in recent years for (facilitated transport). Alternatively, a solvent can be used in
specific carbon dioxide separation. The focus of the majority conjunction with a polymeric membrane for the same effect
of research has been on solvent development. For a solvent (membrane gas absorption). This research will continue into
to be acceptable, it must be able to absorb carbon dioxide at the future, shifting Robeson’s bound further to the upper
relatively low partial pressures, achieve high carbon dioxide right corner of a Permeability/Selectivity plot. However,
loading, as well as have low volatility and viscosity. membranes that succeed commercially need to be cheap and
Importantly in membrane absorption applications, it must not readily processed into hollow fibre format. Therefore, there
wet through the membrane pores, as this leads to inefficient will also be future patents awarded into manufacturing
mass transfer. Further, it must be inert towards the memb- techniques, not covered in this review, that make possible
rane material. Amine based solvents [165], such as mono- the transformation of many of these novel polymeric memb-
ethanolamine (MEA), diethanolamine and methyldiethanol- rane materials into acceptable industrial module designs.
amine, are the industry standard for column-based solvent Future polymeric membrane research will also be focused on
absorption but are known to cause pore wetting and improving the membrane durability under adverse condi-
membrane degradation in MGA applications. Other examp- tions. Future patents will be awarded for methods that imp-
les are potassium carbonates and hydroxide solutions [166]. rove the mechanical strength, chemical and thermal resis-
Potassium carbonate is more resilient to degradation and less tance, as well as reducing aging affects, without compro-
toxic compared to MEA. The major drawback of this solvent mising performance. This can already be seen with patents
is the slow reaction time. This can be overcome with the awarded for a range of polymer cross-linking agents and
addition of promoters such as piperazine which improve the methodologies.
performance. Aqueous ammonia has also been suggested as
an alternative solvent because the byproduct, ammonium Importantly, in all gas separation membrane applications
bicarbonate, can potentially be used as a fertilizer [167]. The minor components exist that will influence performance.
patent literature also covers the possibility of solvent Condensable vapour, such as hexane in natural gas and water
mixtures, which alter the physical characteristics of the in flue gas, can accumulate within the membrane matrix
liquid phase to enhance absorption. Some examples are altering both the permeability and selectivity. Their existence
solvent compatible alcohols, esters and glycols [168, 169]. within the polymer can also lead to plasticization, altering
the mechanical properties of the membrane and increase the
The importance of the polymeric membrane material is chance of membrane failure [172, 173]. Minor gas compo-
different for MGA compared to gas separation processes, nents, such as CO, NH3, H2S, SOx and NOx, found in both
because it has a non-selective role. Generally, the major natural and flue gases, can also degrade the polymeric mate-
criteria are the hydrophobicity of the membrane, pore size rial, reducing performance and causing premature aging.
and chemical resistance to the solvent. The hydrophobicity However, very little research on the effect of these compo-
of the membrane ensures wetting of the pores is minimal and nents has been reported in the patent literature. Therefore,
therefore that mass transfer rates remain high. The dimen- given the interest in maintaining membrane performance and
Carbon Dioxide Separation through Polymeric Membrane Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 65

the efforts to prevent membrane degradation, the study of [27] Plate N, Yampol'skii YP. Relationship between structure and
minor component effects will be an area of interest in the transport properties for high free volume polymeric materials, in:
Polymeric Gas Separation Membranes. Baton Rouge: CRC Press
near future. 1994; 115-208.
To conclude, with the rapidly increasing interest in [28] Kanehashi S, Nagai K. Analysis of dual-mode model parameters
for gas sorption in glassy polymers. J Membrane Sci 2005; 253:
carbon dioxide capture to mitigate global warming, gas 117-138.
separation polymeric membranes will remain an active [29] Favre E. Carbon dioxide recovery from post-combustion processes:
research area and the patented literature will continue to Can gas permeation membranes compete with absorption? J
grow as new discoveries are made in the art. Membrane Sci 2007; 294: 50-59.
[30] Hendriks C. Carbon dioxide removal from coal-fired power plants.
REFERENCES Dortecht: Kluwer Academic Publishers; 1994.
[31] Roberts C., Gibbins J.R., Panesar R., Kelsall G. Improvement in
[1] Carapellucci R, Milazzo A. Membrane systems for CO2 capture power generation with post-combustion capture of CO2.
and their integration with gas turbine plants. Proc Inst Mech Eng Cheltenham: IEA Greenhouse Gas R & D Programme; 2004.
Part A. J Power Energy 2003; 217: 505-517. [32] Robeson L. Correlatoin of separation factor versus permeability for
[2] Thambimuthu K, Davidson J, Gupta M. CO2 capture and resuse. In: polymeric membranes. J Membr Sci 1991; 62: 165-185.
Proc. Canda; IPCC Workshop on carbon capture and storage; [33] Saufi S, Ismail A. Fabrication of carbon membranes for gas
2002. separation - a review. Carbon 2004; 42: 241-259.
[3] Aaron D, Tsouris C. Separation of CO2 from flue gas: a review. [34] Watson, E.R., Rowley, G.V., Wunderlich, C.R.: US3432585
Separ Sci. Technol 2005; 40: 321-348. (1969).
[4] Rao A, Rubin E. A technical, economic and environmental [35] Lee, M.-S., Choi, S.-H., Shin, Y.-C.: US20030134550 (2003).
assessment of amine-based CO2 capture technology for power plant [36] Hoehn, H.H., Richter, W.J.K.: US3899309 (1975).
greenhouse gas control. Environ Sci Technol 2002; 36: 4467-4475. [37] Hoehn, H.H.: US3822202 (1974).
[5] Feron P. CO2 capture: The characterisation of gas separation/ [38] Bentz, F., Elfert, K., Kunzel, H.E., Wolf, G.D.: US4217227 (1980).
removal membrane systems applied to the treatment of flue gases [39] Richter, W.J.K., Hoehn, H.H.: DE1941022 (1970).
arising from power generation using fossil fuel. Cheltenham: IEA [40] Richter, W.J.K., Hoehn, H.H.: DE1941932 (1970).
Greenhouse Gas R & D Programme; 1992. [41] Steadly, H., Laccetti, A.J.: US4770777 (1988).
[6] Abertz V, Brinkmann T, Dijkstra M, et al. Developments in [42] Hayes, R.A.: US5076817 (1991).
membrane research: from material via process design to industrial [43] Harris, J.E., Berger, A., Chopdekar, V.M., Matzner, M., Spanswick
application. Adv Eng Mater 2006; 8: 328-358. J.: US4713438 (1987).
[7] Basu A, Akhtar J, Rahman M, Islam M. A Review of separation of [44] Manos, P.: EP0219878 (1987).
gases using membrane systems. Petrol Sci Technol 2004; 22: 1343- [45] Richter, W.J.K., Hoehn, H.H.: US3567632 (1971).
1368. [46] Sanders, E.S.J., Wan, H.S., Beck, H.N.: US4975228 (1990).
[8] Stern S. Polymers for gas separation: the next decade. J Membrane [47] Sanders, E.S.J., Overman, D.C.: EP0500974 (1992).
Sci 1994; 94: 1. [48] Beck, H.N., Sanders, E.S.J., Lipscomb, G.G.: US4962131 (1990).
[9] Powell C, Qiao G. Polymeric CO 2/N2 gas separation membranes [49] Anada, J.N., Feay, D.C., Bales, S.E., Jeanes, T.O.: EP0242147
for the capture of carbon dioxide from power plant flue gases. J (1987).
Membrane Sci 2006; 279: 1-49. [50] Jeanes, T.O.: EP0316960 (1989).
[10] Maiser G. Gas separation with polymer membranes. Angew Chem [51] Chen, N., Tien, C.-F., Patton, S.M.: US5232471 (1993).
Int Ed. 1998; 37: 2960. [52] Tien, C.-F., Surnamer, A.D.: EP0455216 (1991).
[11] Koros W. Gas separation membranes: needs for combined [53] Pedretti, U., Gandini, A., Roggero, A., Sisto, R., Valentini, C.,
materials science and processing approaches. Macromol Symp Assogna, A., Stopponi, A.: US5169416 (1992).
2002; 188:13. [54] Laverty, B.W., Vujosevic, R., Dang S., Yao B., Matsuura T.,
[12] Baker R. Future directions of membrane gas separation technology. Chowdhury G.: GB2334526 (1999).
Ind Eng. Chem Res 2002; 41: 1393. [55] Percec, E.S., Li, G.S.: US4596860 (1986).
[13] Graham TP. Mag 1866; 32: 402. [56] Li, G.S.: US4586939 (1986).
[14] Loeb, S., Sourirajan, S.: US3133132 (1964). [57] Farias, O., Gandini, A., Monga, R., Roggero, A., Sisto, R.,
[15] Jawad M. Future for desalination by reverse osmosis. Desalination Valentini, C.: EP0360318 (1990).
1989; 72: 23-28. [58] Hachisuga, H.: JP11342322 (1999).
[16] Podall H. Recent advances in reverse osmosis membranes for [59] Hachisuga, H., Matsumoto, K., Obara T.: JP6238138 (1994).
desalination. Chem Eng Prog 1971; 67: 260-266. [60] Illing, G.: DE19936044 (2001).
[17] Koros W. Gas Separation, in: Membrane Separation Systems - [61] Koros, W.J., Walker, D.R.B.: US5262056 (1993).
Recent Developments and Future Directions Baker RW. William [62] Chiao, C.C.: US4717395 (1988).
Andrew Publishing; 1991. [63] Bikson, B., Coplan, M.J., Goetz, G.: US4508852 (1985).
[18] Paul D, Yampol'skii Y. Polymeric gas separation membranes. [64] Coplan, M.J., Park, C.H., Williams, S.C.: US4414368 (1983).
baton rouge: CRC Press 1994. [65] Rose, J.B.: US4268650 (1981).
[19] Spillman R. Economics of gas separation by membranes. Chem [66] Rose, J.B.: US4273903 (1981).
Eng Prog 1989; 85: 41. [67] Quentin, J.: US4054707 (1977).
[20] Mazur W, Chan M. Membranes for natural gas sweetening and [68] Quentin, J.: US3709841 (1973).
CO2 enrichment. Chem Eng Prog 1982; 78: 38-43. [69] Chiao, C.C.: US4828585 (1989).
[21] Coady A, Davis J. CO2 recovery by gas permeation. Chem Eng [70] Bourganel, J.: US4026977 (1977).
Prog 1982; 78: 43-49. [71] Graefe, A.F., Saltonstall, C.W.J., Schell, W.J.: US3875096 (1975).
[22] Fritzsche A, Kurz J. The separation of gases by membranes, in: [72] Kawakami, J.H., Bikson B., Gotz, G., Ozcayir, Y.: EP0426118
Handbook of industrial membrane technology porter MC, editor. (1991).
William Andrew Publishing 1990; 559-593. [73] Hayes, R.A.: US4880442 (1989).
[23] Hill T. Surface diffusion and thermal transpiration in fine tubes and [74] Hayes, R.A.: US4717393 (1988).
pores. J Chem. Phys. 1956; 25; 730-745. [75] Hayes, R.A.: US4705540 (1987).
[24] Hwang S-T, Kammermeyer K. Membrane Separations. New York: [76] Kohn, R.S., Coleman, M.R., Chung, T.-S.: US5055116 (1991).
Wiley-Interscience 1975. [77] Macheras, J.T.: US5635067 (1997).
[25] Rhim H, Hwang S-T. Transport of capillary condensate. J Colloid [78] Black, L.E., Boucher, H.A.: US4571444 (1986).
Interf. Sci. 1975; 52: 174-181. [79] Wan, W.-K.: US4836927 (1989).
[26] Lee K-H, Hwang S-T. The transport of condensible vapors through [80] White, L.S.: US20016180008 (2001).
a microporous Vycor glass membrane. J Colloid Interf. Sci. 1986; [81] Carlsen, D.B., Andrus, R.G., Hall, R.T.: US5605627 (1997).
110: 544-555. [82] Ekiner, O.M., Simmons, J.W.: WO9405404 (1994).
[83] Weinberg, M.G.: EP0554862 (1993).
 66 Recent Patents on Chemical Engineering, 2008, Vol. 1, No. 1 Kentish et al.

[84] Iwama, A., Iwahori, H., Kazuse, Y.: US4358378 (1982). [130] Rojey, A., Deschamps, A., Grehier, A., Robert, E.: US4925459
[85] Makino, H., Nakatani, M.: US4690873 (1987). (1990).
[86] Alegranti, C.W.: US4113628 (1978). [131] Rhone, P.: FR2079460 (1971).
[87] Pfeifer, J., Ciba, G.A.: EP0141781 (1985). [132] Hasse, D.J., Kulkarni, S.S., Corbin, D.R.: US2003089227 (2003).
[88] Makino, H., Yoshihiro, H.T., Shimazaki, H.: US4370290 (1983). [133] Sterzel, H.-J., Sanner, A.: EP0154248 (1985).
[89] Makino, H., Kusuki, Y., Harada, T., Shimazaki, H., Isida, T.: [134] Kulprathipanja, S., Neuzil, R.W., Li, N.N.: US5127925 (1992).
US4528004 (1985). [135] Kulkarni, S., Hasse, D.J., Corbin, D.R., Patel, A.N.:
[90] Makino, H., Kusuki, Y., Harada, T., Shimazaki, H., Isida, T.: US20036508860 (2003).
US4474858 (1984). [136] Grose, R.W., Flanigen, E.M.: US4061724 (1977).
[91] Shimatani, S., Yamamoto, M., Shimazu, A., Iwama, A.: [137] Kulkarni, S., Ekiner, O.M., Hasse, D.J.: US2005230305 (2005).
US4964887 (1990). [138] Kulprathipanja, S., Charoenphol, J.: US2003089228 (2003).
[92] Blinka, T.A., Itatani, H., Wang, I.-F.: US5042992 (1991). [139] Te, H.H.J.C., Mulder, M.H.V., Smolders, C.A., Bargeman, D.,
[93] Jeanes, T.O., Summers, J.D., Sanders, E.S.J.: US4988371 (1991). Schroder, G.A., Setec, B.V.: EP0254759 (1988).
[94] Chung, T.-S., Chng, M.L., Shao, L.: US20040177753 (2004). [140] Henis, J.M.S., Tripodi, M.K.: US4230463 (1980).
[95] Lee, Y.-M., Park, H.B., Lee, C.-H.: US20040236038 (2004). [141] Marand, E., Kim, S.: US2007022877 (2007).
[96] Maeda, M.: US5969087 (1999). [142] Freeman, B.D., Matteucci, S., Lin, H.: US2007137477 (2007).
[97] Simmons, J.W.: WO0249747 (2002). [143] Kulkarni, S., Hasse, D.J.: US2007199445 (2007).
[98] Langsam, M.: US5939520 (1999). [144] Ekiner, O.M., Fleming, G.K.: US5468430 (1995).
[99] Sakellaropoulos, G.P., Kaldis, S.P., Kapantaidakis, G.C., Dabou, [145] Vu D, Koros W, Miller S. Mixed matrix membranes using carbon
X.S.: EP0778077 (1997). molecular sieves. J Membrane Sci 2003; 211: 335.
[100] Ekiner, O.M., Simmons, J.W.: US20060156920 (2006). [146] Miller, S.J., Yuen, L.-T.: US2005043167 (2005).
[101] Hsieh H. Inorganic membranes. Membr Mater Proc 1990; 84: 91. [147] Miller, S.J., Kuperman, A., Vu, D.Q.: US2006107830 (2006).
[102] Iwahara, H., Yamaji, T., Azuma, S.: JP56041804 (1981). [148] Hagg, M.-B., Kim, T.-J., Li, B.: WO05089907 (2005).
[103] Oshima, H., Seki, Y.: JP8071385 (1996). [149] Gen, E.: GB1076438 (1967).
[104] Shreiber, E.H., Eardley, E.P., Srinvasan, V., van Hassel, B.A., [150] Gen, E.: GB1167797 (1967).
Shah, M.M.: US20026623714 (2002). [151] Sirkar, K.K., Kovvali, S., Chen, H.: WO03008070 (2003).
[105] Ku, A.Y.-C., Ruud, J.A., Molaison, J.L., Schick, L.A., [152] Kovvali A, Chen H, Sirkar KK. Dendrimer Membranes: A CO2-
Ramaswamy, V.: WO07037933 (2007). Selective Molecular Gate. J Am Chem Soc 2000; 122: 7594-7595.
[106] Gobina, E.: US2006112822 (2006). [153] Wang, Z., Wang, S., Lu, Q.: CN1363414 (2002).
[107] Ku, A.Y.-C., Ruud, J.A., Molaison, J.L., Schick, L.A., [154] Ho, W.S.W.: US20006099621 (2000).
Ramaswamy, V.: WO07037933 (2007). [155] Ho, W.S.W.: WO06050531 (2006).
[108] Dyer A. An Introduction to Zeolite Molecular Sieves. New York: J [156] Kulprathipanja, S., Charoenphol, J.: WO03039728 (2003).
Wiley & Sons 1988. [157] Ho, W.S.W.: US5611843 (1997).
[109] Guiver, M.D., Robertson, G.P., Thi, H.N.L.: CA2421875 (2004). [158] Nakabayashi, M., Okabe, K., Mishima, T., Mano, H., Haraya, K.:
[110] Suzuki, H.: EP0180200 (1986). EP0638353 (1995).
[111] Berger, C.: US3567666 (1971). [159] Lackner, K., West, A.C., Wade, J.L.: WO06113674 (2006).
[112] McKeown, N.B., Budd, P.M., Msayib, K., Ghanem, B.: [160] Qi Z, Cussler EL. Microporous hollow fibres for gas absorption II:
WO05012397 (2005). Mass transfer across the membrane. J Membrane Sci 1985; 23:
[113] Young, J.S., Long, G.S., Espinoza, B.F.: WO06028594 (2006). 333-345.
[114] Davis,H.J., Thomas, N.W.: US4020142 (1977). [161] Feron P, Jansen A, Klaassen R. Membrane technology in Carbon
[115] Dye, R.C., Jorgensen, B., Pesiri, D.R.: US20046681648 (2004). Dioxide removal. Energy Conv Man 1992; 33: 421-428.
[116] Wang, H., Yeager, G.W.: US2007056901 (2007). [162] Yeon S-H, Lee K-S, Sea B, ParkY-I, Lee K-H. Application of
[117] Ekiner, O.M., Simmons, J.W.: US2006196355 (2006). pilot-scale membrane contractor hybrid system for removal of
[118] Seo, Y.: US2007185264 (2007). carbon dioxide from flue gas. J Membrane Sci 2005; 257: 156-160.
[119] Ismail A., David L. A review on the latest development of carbon [163] Thakore, Y.B., Stoy, V.: US4954145 (1990).
membranes for gas separation. J Membrane Sci 2001; 193: 1-18. [164] Ward, W.J.: US4147754 (1979).
[120] Soffer, A., Koresh, J.E., Saggy, S.: US4685940 (1987). [165] Lee, G.S., Lee, G.H., Park, Y.I., Seo, B.G., Yoen, S.H.:
[121] Rao, M.B., Sircar, S., Golden, T.C.: US5104425 (1992). KR0042656 (2004).
[122] Rao, M.B., Sircar, S., Abrado, J.M., Baade, W.F.: US5354547 [166] Baba, T., Fukazama, T., Nishino, Y.: JP3218912 (1991).
(1994). [167] Yuan H.: US2007163432 (2007).
[123] Rao, M.B., Sircar, S., Golden, T.C.: US5431864 (1995). [168] Birbara, P.J., Nalette, T.A.: US5281254 (1994).
[124] Lee, Y.M., Park, H.B., Seo, I.Y.: KR008782 (2002). [169] Jansen, A.E., Feron, P.H.M.: US5749941 (1998).
[125] Foley, H.C., Rajagopalan, R., Merritt, A.R.: US2007017861 [170] Feron, P.H.M., Jansen, A.E.: NL9400483 (1995).
(2007). [171] Hesse, H.J.F.A., Smit, M.J., du Toit, F.J.: US20016312655 (2001).
[126] Lee S. Handbook of alternative fuel technologies. New York: [172] Al-Juaied M., Koros W.J. Performance of natural gas membranes
Taylor & Francis 2007. in the presence of heavy hydrocarbons. J Membrane Sci 2006; 274:
[127] Hagg, M.-B., Lie, J.A.: WO07017650 (2007). 227-243.
[128] Zimmerman C, Singh A, Koros W. Tailoring mixed matrix [173] Vu DQ, Koros WJ, Miller SJ. Effect of condensable impurities in
composite membranes for gas separation. J Membrane Sci 1997; CO2/CH 4 gas feeds on carbon molecular sieve hollow-fiber
137: 145. membranes. Ind Eng Chem Res 2003; 42: 1064-1075.
[129] Mahajan R, Zimmerman C, Koros W. Fundamental, practical
aspects of mixed matrix gas separation membranes. ACS Symp
Series 1999; 733: 277.

View publication stats