Khalilpour 2015

Journal of Cleaner Production xxx (2014) 1e15
Contents lists available at ScienceDirect
Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro
Review
Membrane-based carbon capture from flue gas: a review

Rajab Khalilpour a, *, Kathryn Mumford b, Haibo Zhai c, Ali Abbas a, Geoff Stevens b,
Edward S. Rubin c
a
Department of Chemical and Biomolecular Engineering, University of Sydney, NSW 2006, Australia
b
Department of Chemical and Biomolecular Engineering, University of Melbourne, VIC 3010, Australia
c
Department of Engineering and Public Policy, Carnegie Mellon University, PA 15213, USA
a r t i c l e i n f o a b s t r a c t
Article history: There has been an increasing interest in the application of membranes to flue gas separation, primarily
Received 2 February 2014 driven by the need of carbon capture for significantly reducing greenhouse gas emissions. Historically,
Received in revised form there has not been general consensus about the advantage of membranes against other methods such as
2 September 2014
liquid solvents for carbon capture. However, recent research indicates that advances in materials and
Accepted 16 October 2014
process designs could significantly improve the separation performance of membrane capture systems,
Available online xxx
which make membrane technology competitive with other technologies for carbon capture. This paper
mainly reviews membrane separation for the application to post-combustion CO2 capture with a focus on
Keywords:
Post-combustion carbon capture
the developments and breakthroughs in membrane material design, process engineering, and engi-
Flue gas neering economics.
Membrane © 2014 Elsevier Ltd. All rights reserved.
Process systems
1. Introduction Tsouris, 2005; Desideri and Paolucci, 1999; Rao and Rubin, 2002),
while other technologies, including membranes, lie at different
In the United Nation's Earth Summit held June 1992 in Brazil, a developmental stages from concept to pilot-scale testing. Solvent-
treaty known as Nations Framework Convention on Climate Change based PCC, in spite of being the best available technology (BAT)
(UN-FCCC) was made in order to stabilize greenhouse gas con- for dealing with various flue gas streams given in Table 1, may not
centrations in the atmosphere. This treaty came into force gradually be a long-term desired technology for PCC due to its high energy
as more countries signed it, which accelerated research and penalty associated with solvent regeneration (Feron, 2009).
development on different carbon dioxide (CO2) capture technolo- Different from solvent technologies for post-combustion CO2 cap-
gies. Since that time, a few technologies have been developed for ture, membrane technology is a physical separation process shown
post-combustion carbon capture (PCC), including liquid solvent in Fig. 2 where gas mixtures consisting of two or more components
absorption, membranes, solid-sorbent adsorption (Krishnamurthy are separated by a semipermeable barrier into a retentate stream
et al., 2014), membranes (Ho et al., 2008b), and mineralization and a permeate stream. The main advantages of membrane sepa-
(Zevenhoven and Fagerlund, 2009). ration against other technologies include compactness, modularity,
The first commercial interest in the separation of CO2 from flue ease of installation by skid-mounting, ability to be applied in
gas was its application for enhanced oil recovery (EOR) (Herzog remote areas (such as offshore), flexibility in operation and main-
et al., 1997). Application of CO2 for EOR started in the 1970s and tenance, and, in most cases, lower capital cost as well as lower
has proved to be an effective economic approach over the years energy consumption (Seader and Henley, 2006). They also require
(Long et al., 2006). When the amount of CO2 product from natural very little chemicals compared to conventional separation pro-
gas purification is insufficient, the CO2 from flue gases emitted from cesses (He and Ha €gg, 2012).
nearby power plants can be utilized using various carbon capture Aaron and Tsouris (2005) reviewed various methods for CO2
technologies for the EOR application. Currently, amine-based car- removal from flue gas including solvent absorption, pressure- and
bon capture technologies (shown Fig. 1), such as Flour Econamine temperature-swing adsorption using various solid sorbents, cryo-
Plus, are commercially available for implementation (Aaron and genic distillation, and membranes. They concluded that the most
promising method for CO2 separation is liquid absorption using
monoethanolamine (MEA); however, the development of ceramic
* Corresponding author. and metallic membranes for membrane diffusion will produce
E-mail address: [email protected] (R. Khalilpour). membranes that are significantly more efficient at separation than
http://dx.doi.org/10.1016/j.jclepro.2014.10.050
0959-6526/© 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Khalilpour, R., et al., Membrane-based carbon capture from flue gas: a review, Journal of Cleaner Production
(2014), http://dx.doi.org/10.1016/j.jclepro.2014.10.050
2 R. Khalilpour et al. / Journal of Cleaner Production xxx (2014) 1e15
Fig. 1. Typical solvent-based process for carbon capture from flue gas streams.
Table 1 hybrid system combining membranes with an absorption process is

Typical conditions of a flue gas stream from various sources (Metz et al., 2005). used to separate the bulk of CO2 via membrane and the remainder,
Stream sources CO2 concentration Pressure CO2 partial down to pipeline specification, with absorption system; (2) Unlike
%vol (dry) range pressure (bar) well gas streams with high pressures, flue gas streams are typically
Gas turbines 3e4 Atmospheric 0.03e0.04 at atmospheric pressure and require either costly compression or
Fired boilers of oil refinery 8 “ 0.08 vacuum pumping to generate driving force for membrane gas
and petrochemical plant separation; and (3) Different molecular characteristics of CH4/CO2
Natural gas fired boiler 7e10 “ 0.07e0.10 separation and CO2/N2 separation involve in membrane gas sepa-
Oil fired boilers 11e13 “ 0.11e0.13
Coal fired boilers 12e14 “ 0.12e0.14
ration processes.
IGCC after combustion 12e14 “ 0.12e0.14 Membrane research and development for CO2 capture could be
Blast furnace (after 27 “ 0.27 divided into two categories: membrane material design (MMD)
combustion) and membrane systems engineering (MSE). The fundamental goal
Cement process 14e33 “ 0.14e0.33
of MMD is to synthesize membranes of desirable permeance and
selectivity while having chemically and physically stable struc-
tures. On the other hand, the major goal of MSE is to develop
the liquid absorption method. The special report by the Intergov- membrane capture processes with optimal configurations to ach-
ernmental Panel on Climate Change (IPCC) (Metz et al., 2005) ieve the separation targets (e.g., CO2 removal rate and product
mentioned that membranes for flue gas separation had not yet purity) at minimum capital and operational expenditures (CAPEX
been trialed in large scale systems, and their reliability and costs and OPEX).
were not fully evaluated at that time. But the report was optimistic This paper mainly reviews membrane separation for the appli-
about the future of this application via worldwide R&D efforts that cation to post-combustion CO2 capture with a focus on the de-
aimed at the manufacturing of more suitable membrane materials velopments and breakthroughs in membrane material design,
for CO2 capture in large-scale applications. process engineering, and process economics. In the following sec-
Membranes for carbon dioxide separation have been applied to tions, we first review and analyze the studies on MMD for PCC
natural gas sweetening where the concentrations of CO2 and H2S applications. This is followed by an MSE literature survey. The next
contained in high-pressure natural gas should be lowered to the focus is on the techno-economics of MPCC in comparison with
levels of meeting the gas pipeline specifications (CO2 < 2% and other capture technologies. Last, we discuss the limitations of
H2S < 4 ppm) (Baker and Lokhandwala, 2008; Ettouneya et al., current MPCC research and potential improvement opportunities.
1995; Peters et al., 2011). Numerous commercial membrane tech-
nologies have been available for natural gas sweetening. However, 2. Membrane material design (MMD)
this does not necessarily ensure the feasibility of membranes for
flue gas treatment because of several key differences between the 2.1. Non-facilitated membranes
two applications: (1) Membranes are good for bulk separation, but
generally are not an economical option for cases with either a low Non-facilitated membranes rely on the “solution-diffusion”
feed concentration or a high-purity product requirement. Thus, transport process whereby the permeate first dissolves into the
membranes are used for natural gas sweetening only when the CO2 membrane and then diffuses through it. The solubility of the
concentration of a well gas stream is sufficiently high. Otherwise, a permeate follows the Henry's law. For a given temperature, the
Fig. 2. Schematic of membrane-based CO2 separation process from flue gas streams.
R. Khalilpour et al. / Journal of Cleaner Production xxx (2014) 1e15 3
amount of a gas dissolved per unit volume is proportional to the compared separation efficiency with the high-pressure feed on the
partial pressure of the gas in equilibrium with it, described as outside and the inside of the fibers and reported permeances of
Eq. (1): about 20, 3, and 0.8 GPUs1 for CO2, N2, and O2, respectively. They
also claimed that the inherent membrane separation capability
Ci ¼ Pi =KHi ¼ Pi Ki (1) appeared the same whether the feed was inside or outside of the
fibers. Sada et al. (1992) examined separation characteristics of
where Ci is the dissolved concentration, Pi is the partial pressure, CO2/N2/O2 mixtures using hollow fiber modules of asymmetric
KHi the Henry's Law constant, and Ki is the solubility of component i. cellulose triacetate membrane with a countercurrent plug flow
Diffusion is governed by Fick's first law of diffusion as presented model. They reported ideal selectivity of CO2 relative to N2 in the
in Eq. (2): range of 21e24. Polyethersulfones (PES) have a higher glass-
transition temperature than cellulose triacetate, thus making
Ji ¼ Di DCi =L (2) them more resistant to plasticization. Kumazawa et al. (1993) used
PES membranes and reported CO2/N2 ideal selectivities of 35 and
where Ji is the flux per unit area, Di is the diffusion co-efficient in 40 for asymmetric and homogenous membranes respectively.
the membrane, and DCi is the difference in concentration across the Cellulose acetate membranes were later discarded due to their
membrane width, L. problems with plasticization (Bernardo et al., 2009; Scholes et al.,
Substituting Eq. (1) into Eq. (2) gives: 2008) though some studies showed that blending CA with other
polymers such as PEG could improve the membrane performance
Ji ¼ Ki Di DPi =L (3)
(Li et al., 1995).
Tokuda et al. (1997), with knowledge that Cardo polyimides
where DPi is the difference in partial pressure across the mem-
have high CO2/N2 permselectivity, studied the impact of various
brane. The membrane selectivity is dependent upon the relative
functional groups on selectivity of Cardo polyimides. They noticed
rates of these processes.
that Cardo polyimide (PI-BT-COOMe), having a CO2 affinitive
. methylcarboxyl functional group, shows outstanding CO2/N2
ai=j ¼ Ji Jj (4) selectivity behavior. More specifically, Cardo polyimides which had
3,30 ,4,40 -Benzophenonetetracarboxylic dianhydride as a monomer
A schematic representation of this mechanism is presented in
(PI-BT) obtained the highest separation with CO2/N2 selectivity of
Fig. 3.
52 at 25 C. Later, it was reported that a bis(phenyl)fluorene-based
The major advantage of non-facilitated membranes is that there
Cardo polyimide could result in membranes with notably high
is a high tunable degree of controlling the membrane permeability
permeance of 1000 GPU with good CO2/N2 selectivity of 40
and selectivity via the manipulation of polymer preparation and
(Kazama et al., 2004). Yoshino et al. (2000) studied the effects of
chemical composition. Counteracting this is the characteristic
hard-segment polymers (namely polyurethanes (PUs), polyamides
swelling and plasticization of the materials as a result of CO2 ab-
(PAs), and polyamides (PIs)) on CO2/N2 separation properties of
sorptions. Additionally, this material is generally only suitable for
poly(ethylene oxide) (PEO). The study showed that compared with
low temperature applications.
PEO-PIs, PEO-PUs and PEO-PA had much smaller CO2 permeability
The pioneering work on CO2/N2 separation using non-facilitated
values and somewhat lower or similar CO2/N2 selectivity. This was
membranes may be traced back to the study by Kawakami et al.
due to small diffusivity and solubility values for CO2 as a result of
(1982) who blended a low permeable glassy polymer (cellulose
the incomplete phase separation. They also noticed that in worst
nitrate) with a plasticizer membrane (polyethylene glycol (PEG)).
phase separation conditions, PEO-PIs still had CO2/N2 selectivity
They noticed that with the increase of PEG percentage and/or
above 50. They reported the best CO2/N2 selectivity value of 58 for a
molecular weight, the permeability of CO2 as well as CO2/N2
PEO-PI. In a similar study, Bondar et al. (2000) studied properties of
selectivity increased. However, above a threshold concentration
poly(ether-b-amide) segmented block copolymers. They noticed
(30% PEG) plasticization was observed. The interesting point of this
that CO2/N2 selectivities were greater in polymers with higher
research was that the authors did not have a clear idea on industrial
concentrations of polar groups. They obtained CO2/N2 selectivities
applications for this process and only projected that the separation
as high as 56 with CO2 permeability of approximately 220 Barrers2
of CO2 from N2 might be used “in order to recover carbon resources
at 35 C.
or to control CO2 concentration in an artificial atmosphere.”
Kim et al. (2001) used a membrane composed of two polymeric
Sidhoum et al. (1988) studied CO2/N2 separation using cellulose
materials: a porous substrate and a filling polymer that fills the
acetate (CA) hollow fibers with dense skin on the outside. They
pore of the substrate. By using this method, they obtained CO2/N2
selectivity of 32.4 which is higher than their non-filling mem-
branes. The results justified their proposal that the high selectivity
of the pore-filling membrane is attributed to the high solubility
selectivity due to the affinity of CO2 to PEO segment. Nakagawa
et al. (2002) synthesized three copolyimide membranes by the
condensation polymerization of alicyclic 2,3,5-tricarboxy cyclo-
pentyl acetic dianhydride (TCDA) with diamino diphenyl ether
(DADE) and bis(aminopropyl)polydimethylsiloxane (BAS). They
observed CO2/N2 selectivity of 60 for TCDAeDADE. Lin and Freeman
(2004) studied the impact of pressure and temperature on the
solubility, diffusivity, and permeability of various gases including
CO2 and N2 using PEO. PEO exhibited CO2 permeability coefficient
1
1 GPU ¼ 106 cm3 (STP)/cm2 s cmHg.1 GPU ~ 3.4 1010 mol/m2 s Pa.
Fig. 3. Non-facilitated transport mechanism. 2
1 Barrer ¼ 1010 cm3 (STP) cm/(cm2 s cmHg).
of 12 Barrers, and CO2/N2 pure gas selectivity of 48 at 35 C. PE CO2 permeance and selectivity with increasing CO2 feed partial
under similar conditions showed CO2 permeability of 13 Barrers pressure. A schematic representation of the mechanisms for a fixed
with the cost of very low CO2/N2 selectivity of 13. Charmette et al. carrier amine based facilitated transport membrane is provided in
(2004) studied properties of cross-linked membranes prepared Fig. 4. As shown, the CO2 is absorbed and reacts to form bicar-
from PEO and poly(epichlorohydrin) (PEP) at different composition bonate, the form in which it tranverses the membrane. Upon
ratios. At PEO/PEP mixtures with high PEO concentrations such as reaching the other side it back reacts to form CO2 and is then
93/7 and 96/4, a high CO2/N2 selectivity above 62 was observed. released on the permeate side.
In a thorough study, Powell and Qiao (2006) reviewed available Support liquid membranes (SLMs) were the first facilitated
membrane materials for flue gas separation, possible design stra- transport membranes developed. These systems consisted of a
tegies, synthesis, fabrication, and the role of novel materials. Their liquid phase containing a carrier species in the pores of a polymer
survey included a number of different classes of polymers such as support. The mobile carriers are able to move freely across the
polyacetylenes, poly(arylene ether), polyarylates, polycarbonates, membrane and perform well for CO2 separation processes. How-
poly(ethylene oxides), polyimides, 6FDA-based polyimides, poly(- ever, these systems are relatively unstable as the carriers can be
phenylene oxide)s, and polysulfones as well as carbon and mixed washed out from the system as they are not bonded to the polymer
matrix membranes. Their survey, which covered some 190 reports, matrix. SLM design was improved by Leblanc et al. (1980) through
concluded that copolymers and polymer blends had higher po- the development of ion exchange membranes. Electrostatic forces
tential for further research e of which two materials were the most between ionic carriers and the membrane better held the carriers
interesting. The first types are polyimides that incorporate 6FDA within the membrane. In recent years, ionic liquids have also been
due to their gas transport properties, good physical properties, used in these systems. Ionic liquids are favorable as they have high
potential structural variations and ease of membrane formation. temperature durability and negligible volatility. Another variety of
The second are the PEO segmented copolymers with polyimides facilitated transport membranes are fixed carriers. Fixed carriers
having both high selectivities and a high CO2 permeability possibly are where the reactive group is bound to the polymer backbone of
due to high solubility of CO2 in PEO. the membrane material. Rather than being transported through the
Scholes et al. (2008) reviewed the state-of-the-art development membrane with the mobile carrier, the species ‘hops’ from one
in mixed matrix polymeric membranes, facilitated transport poly- fixed carrier to the next through the membrane. Different to gas
meric membranes, and other membrane types, e.g., carbon and absorption membranes, facilitated transport membranes rely on
inorganic. They showed that the trend of research on polymeric the membrane being saturated to enable the required reaction to
membranes aims at improving their performance by incorporating occur.
an additional agent into the polymeric phase that could be another Kasahara et al. (2012) prepared amino acid ionic liquid based
polymer (polymeric blends), particulate matter (mixed matrix facilitated transport membranes with tetrabutylphosphonium
membranes), or a carrier molecule (facilitated transport). They amino acid ionic liquids and with glycine, alanine, proline, and
predicted that such studies will shift Robeson's bound to the upper serine as the anion. Tetrabutylphosphonium was selected due its
right corner of a Permeability/Selectivity plot having moderate superior thermal stability. The proline based membrane showed
selectivities and high permeabilities. the highest permeability and best CO2/N2 selectivity most likely
Toy et al. (2012), in a project for U.S. Department of Energy, due to its high higher water holding ability.
synthesized a vinylidene fluoride (VDF)-based copolymer mem- Zhao and Ho (2012) developed a fixed-site-carrier facilitated
brane (VDF-co-B). They observed that the bulky, high-dipole transport membrane incorporating sterically hindered polyamines
comonomer B helped to increase gas permeability in the polymer with cross-linked polyvinylalcohol. The incorporation of amine
matrix, and its greater polarity enhanced the CO2 affinity (solubil- steric hindrance in a solid phase exhibited significantly improved
ity) of the matrix. They reported 2.5e3 times higher CO2/N2 enhancement of CO2 permeability and CO2/H2 and CO2/N2 separa-
selectivity and six times higher CO2 permeability than those of the tion performances at 110 C and a free pressure of 2 atm. Zhao and
PVDF homopolymer. Ho prepared a new membrane incorporating moderately hindered
Some studies have also addressed metallic membranes for PCC poly-N-isopropylallylamine as a fixed site carrier alongside mobile
applications. For instance, Carapellucci and Milazzo (2003) evalu- carriers creating a cross-linked polyvinylalcoholepoly(siloxane)
ated the influence of different operating conditions on the behavior network (Zhao and Ho, 2013). This new membrane exhibited high
of pre- and post-combustion separation units, based on metallic or
polymeric membranes. They reported encouraging results for CO2
capture with integration of a metallic membrane system into a
chemically recuperated gas turbine (CRGT) power plant. There has
also been an increasing interest in carbon membranes. Although
these membranes have shown potential for CO2 separation, they
require further improvements, specially by reducing fiber wall
thickness (He and H€ agg, 2011).
2.2. Facilitated transport membranes
Facilitated transport membranes, in addition to exhibiting “so-

lution-diffusion” characteristics, also contain an active transport
mechanism that increases the permeability and selectivity of the
membrane material. The target species reversibly reacts with either
a fixed or mobile carrier present in the membrane that then dif-
fuses across the membrane driven by a concentration gradient in
the complex rather than a gradient in the permeate.
Unlike non-facilitated transport membranes, the main charac- Fig. 4. Schematic of the mechanisms for a fixed carrier amine based facilitated
teristic of these membranes is represented by the decrease of both transport membrane.
CO2 permeability and CO2/N2 and CO2/H2 selectivity. Bai and Ho the membrane provides. However, when the system is operated in
(2011) fabricated CO2 selective facilitated transport membranes the “non-wetted” mode, i.e., pores filled with gas, not solvent, the
based on cross-linked poly(vinyl alcohol) (PVA) and sulfonated resistance is several orders of magnitude smaller than that of the
polybenzimidazole (SPBI) matrixes with 2-aminoisobutyric acid liquid film resistance which results in a negligible overall effect on
(IBA-K) and potassium carbonateebicarbonate as mobile carriers the overall mass transfer coefficient. A schematic is presented in
and poly(allylamine) (PAA) or 2-bromobutane functionalized PAA Fig. 5.
(PAA-C4H9) as fixed carriers for CO2 transport. The cross-linked PVA Prediction of whether an MGC system will operate in the
based membranes showed the best separation performance, likely ‘wetted’ or “non-wetted” mode is primarily controlled by the pa-
due to its great hydrophilicity. rameters of the YoungeLaplace equation (DPc ¼ 4l cos q/d). DPc, is
Deng et al. (2009) developed an ultra-thin PVAm/PVA blend the critical breakthrough pressure, i.e., the differential pressure at
facilitated transport membrane cast on a porous polysulfone (PSf) which pores will become “wetted,” l is the surface tension of the
support for CO2eN2 separation. A separation factor of up to 174 and solvent, q is the contact angle in degrees for the gasesolvent-
a CO2 permeance up to 0.58 m3 (STP)/(m2 h bar) were documented. membrane system, and d is the effective diameter of the membrane
The fixed amino groups in the PVAm matrix function as CO2 carriers pore (assuming that the pores are circular in shape).
to facilitate the transport whereas the PVA adds mechanical The membranes themselves are unlike gas separation mem-
strength. branes as they exhibit no intrinsic selectivity; rather, selectivity of
the system is provided by the solvent used. In CO2 capture systems
2.3. Mixed matrix membranes the most common solvents used are alkanolamines (i.e., mono-
ethanolamine (MEA), Diethanolamine (DEA), etc.) due to their high
Mixed-matrix membranes (MMMs), filled with inorganic par- CO2 loading capacities, high absorption rates, and low regeneration
ticles, provide a means to improve the gas separation performance rates. However, alkanolamines also have low surface tensions, and
of polymeric membranes. The key challenge with this type of so investigation into performance of other solvents is warranted.
membrane is to ensure that the inorganic particles are well bonded As high solvent to membrane contact angles are favorable,
with the polymer thus preventing voids that may result in a loss of membranes are typically synthesized from hydrophobic materials
selectivity. Pera-Titus (2014) presented a review of inorganic ma- such as polypropylene (PP), polyethylene (PE), polytetrafluoro-
terials that may be suitable for use in polymer membranes for CO2 ethylene (PTFE), poly(tetrafluoroethylene-co-perfluorovinylether)
capture. Inorganic fillers trialed have included zeolites (Bastani (PFA), and polyvinylidene fluoride (PVDF). A wide variety of in-
et al., 2013; Junaidi et al., 2014; Nik et al., 2011; Sublet et al., vestigations involving these membranes in association with alter-
2012), carbon nanotubes (Ahmad et al., 2014; Aroon et al., 2013; nate solvents are presented in Table 2. PP membranes are
Rajabi et al., 2013), and metal organic frameworks (MOFs) (Nafisi commercially available and inexpensive, and therefore are most
and Hagg, 2014; Perez et al., 2009). The inorganic fillers have commonly used; however, they generally exhibit a poorer perfor-
been found to improve both the permeability and selectivity of mance. PP is not as hydrophobic as the fluorine containing poly-
membranes over the base polymer and also effectively restrict the meric membranes (PTFE and PVDF) and so are more susceptible to
mobility of polymer changes at high CO2 partial pressures, thus “pore wetting” inhibiting performance.
suppressing CO2 induced plasticization (Shahid and Nijmeijer, Yeon et al. (2003) investigated the mass transfer behavior of
2014). carbon dioxide through a membrane contactor system comprising
Jiang (2012) has provided a comprehensive review on MOF PTFE and PVDF hollow fiber membranes with MEA. They found that
membranes and their application for carbon capture. According to the PVDF was more resistant to “pore wetting” and so exhibited
the author, unlike MOF sorbents, very little work has been carried superior carbon dioxide mass transfer performance. As suggested
out on MOF membranes and they are still at their infancy, though at by the investigations presented in Table 2, the base membranes
very progressive stage. Even amongst the studies on the application used for membrane contactors have been well characterized. Cur-
of MOF membranes for carbon capture, there has been little rent work is focused on modifications of these materials to improve
attention paid to CO2/N2 separation (Krishna and van Baten, 2011). performance.
Therefore, given the very high costs of MOFs synthesis and their A number of methods based on the modification of the base
lab-scale research status, they may not find a place amongst the membrane materials are being investigated to improve the per-
potential commercial membranes for MPCC in the medium term. formance of gas absorption membranes. These include: plasma
2.4. Gas membrane contactors
Gas membrane contactors make use of membranes to provide

an interface between the gas and liquid as opposed to the using
column arrangements, i.e., packed beds, spray towers, etc.
Feron et al. (1992) proposed using gas absorption membranes,
with solvents such as MEA on the permeate side. They reported that
such configuration was more promising as it combines the advan-
tages of absorption technology (high selectivity) and membrane
technology (compactness of equipment). These types of equipment
are now generally referred to as membrane gas contactors (MGC).
The major advantage of these systems over traditional packed bed
based solvent systems is the elimination of the need for phase
dispersion. In membrane contactors, the mass transfer interface is
set in the membrane pores, where the pore volume is occupied by
one of the two phases.
The most common and widely reported negative characteristic Fig. 5. Gaseliquid membrane contactor for CO2 removal from flue gas showing
of MGC systems is the additional resistance to mass transfer that “wetted” and “non-wetted” pores.
Table 2
Gas absorption membranes.
Traditional membrane material
Membrane material Reference Gases separated Solvent(s) trialed Comments

(abbreviation)
Polypropylene (PP) (Wang et al., 2013) N2/CO2 MEA, DEA, MDEA, MEA/MDEA, DEA/2-amino-
2-methyl-1-propanol (AMP), and MDEA/PZ
(Scholes et al., 2012) CO2, CO, H2, N2, CH4 Monoethanolamine, potassium carbonate Pore wetting observed with MEA,
but not potassium carbonate
(Ozturk and Hughes, Water, 10 wt% MEA, 10 wt% DEA
2012)
(Khaisri et al., 2009) CO2 MEA
(Zhang et al., 2008) N2/CO2 Water Rate controlled by absorbent flow rate
(Bottino et al., 2008) N2/CO2 Monoethanolamine (MEA), [3 M]
(Zhang et al., 2008) N2/CO2 Diethanolamine (DEA) [2 M]
(Yan et al., 2007) N2/CO2/O2 Potassium Glycinate (PG) Limited wetting observed
(Yan et al., 2007) N2/CO2/O2 Monoethanolamine (MEA) [0.5e3 M]
(Yan et al., 2007) N2/CO2/O2 Methyldiethanolamine (MDEA) [0.5e3 M]
Polyvinylidene fluoride (Zhang et al., 2008) N2/CO2 Water Rate controlled by absorbent flow rate
(PVDF) (Zhang et al., 2008) N2/CO2 Diethanolamine (DEA) [2 M] Membrane wetting
(Mansourizadeh and N2/CO2 Diethanolamine (DEA)
Ismail, 2009)
(Yeon et al., 2003) 25% CO2, 75% N2 Monoethanolamine, MEA, 5 wt%
Polytetrafuloroethylene (Scholes et al., 2012) CO2, CO, H2, N2, CH4 Monoethanolamine [30 wt%], potassium Pore wetting observed with MEA,
(PTFE) carbonate [30 wt%] but not potassium carbonate
(Yeon et al., 2003) 25% CO2, 75% N2 Monoethanolamine, MEA, 5 wt%
(Kim and Yang, 2000) CO2, N2 2-Amino-2-methyl-1-propanol (AMP)
treatment, applying a thin hydrophobic layer to base membrane Ahmad et al. (2013) deposited a superhydrophobic layer of low-
materials, incorporation of salts into the membranes during density polyethylene (LDPE) onto poly(vinylidene fluoride) (PVDF).
manufacture, and changing manufacturing parameters to alter The material developed was porous with a 152 ± 3.2 water contact
membrane characteristics such as pore diameter. angle and maintained a higher CO2 flux than the unamended PVDF
Bae et al. (2001) compared the performance of Freon-116, membrane. Mosadegh-Sedghi et al. (2013) developed highly hy-
allylamine, water, and acrylic acid for gas plasma treatment of drophobic microporous low density polyethylene membranes
polypropylene membranes. They found that the oxygen containing through incorporation of sodium chloride particles into the struc-
chemical species, i.e., water and acrylic acid, deteriorates the ture. This treatment resulted in an increase of the water contact
membrane pores and allylamine grafted to the surface thereby angle from 98 to 130 .
reducing the membrane effectiveness. Freon-116 gas treatment Ghasem et al. (2012) investigated extrusion temperature effects
made no change to the surface morphology and no deterioration of (from 150 to 170 C) on the pore size, water contact angle, strength,
the pores was observed, yet fluoro-carbon molecules were grafted and porosity for poly(vinylidene fluoride) (PVDF) hollow fiber
to the surface. Additionally, only Freon-116 resulted in an increase membranes. With increased temperatures the pore radius
in solvent membrane contact angle while the others were reduced. decreased and the porosity and water contact angle increased,
Lin et al. (2009b) examined plasma treated PP membranes and resulting in improved CO2 fluxes across the membrane.
reported significant improvements including: an increase in the Rajabzadeh et al. (2013) fabricated PVDF hollow fiber mem-
water contact angle of the membrane from 117 to 143 , increased branes with different membrane porosities on the inner surface.
durability, an 8% increase in the CO2 absorption flux when using They then applied the membranes for CO2 absorption using two
1.0 mol/dm3 AMP as the absorbent, and a decrease in the wetting concentrations of MEA: 2 M and 4 M. No difference in CO2 ab-
ratio with mixed AMPePZ absorbent from 0.0674% to 0.027%. sorption rates was observed for any membrane porosity when used
Franco et al. (2012) undertook polytetrafluoroethylene (PTFE) in associated with a 2 M MEA solvent, while the capture rate for the
plasma sputtering of polypropylene (PP) membranes. This process larger porosity membrane increased when 4 M MEA was used. The
resulted in the formation of an ultrathin fluorinated hydrophobic long-term studies reported wetting for all materials when used
surface that retained the microporous surface of the PP membrane. with 4 M MEA, but only for the larger porosity membranes for the
The novel material outperformed the unamended PP membrane for 2 M concentration. Lin et al. (2013) coated highly porous SiO2
CO2 mass transfer rate with MEA solvent flowing through the fiber aerogels, modified with hydrophobic fluorocarbon functional
lumen by approximately 30%. Upon exposure to MEA, the plasma groups (eCF3), onto a macroporous Al2O3 membrane. This hydro-
treated membranes were found less susceptible to degradation and phobic coating enabled long-term continuous CO2 capture with
wetting as compared to the untreated material. Lin et al. (2009a) large absorption flux enhancements. The authors also examined
plasma treated (CH4) PVDF and used an aqueous mixture of alka- the use of methyltrimethoxysilane (MTMS) precursors for coating.
nolamines (PZ, MEDA and AMP) as absorbents. The water, AMP and The MTMS-derived hydrophobic membrane exhibited CO2 ab-
AMP þ PZ mixture contact angles ranged 145e155 for the modi- sorption with fluxes at least 500% higher than the uncoated MTMS-
fied PVDF, whilst PVDF and PTFE ranged 132e136 with the contact based aerogel membranes (Lin et al., 2014a).
angle sequence mirrored for CO2 absorption rates. Another study
reported that PVDF/siloxane nanofibrous membranes, compared 2.5. Discussion on membrane material design
with PVDF, exhibit higher solvent resistance and mechanical
strength (Lin et al., 2014b). These properties make them better al- Seader and Henley (2006) have identified six attributes for a
ternatives for carbon capture applications. desirable membrane: (1) good permeability, (2) high selectivity, (3)
chemical and mechanical compatibility with the process environ- or even under the condition of real flue gas compositions, while
ment, (4) stability, freedom from fouling, and reasonable useful considering oxygen and water vapor at least.
lifetime, (5) amenability to fabrication and packaging, and (6) Therefore, it can be concluded that research in the field of
resistance to high pressures. Suitable membranes for carbon cap- membrane material design (MMD) is more complex and has a
ture should have all these attributes. However, most of the existing wider spectrum than just focusing on improving the permse-
studies on CO2 capture membranes are focused on improving lectivity of CO2/N2 over a narrow operating condition. Other
selectivity and lack sufficient attention to other important re- membrane attributes, i.e., stability and compatibility, while con-
quirements. Furthermore, the impacts of other flue gas compo- siderations of the real process environment and operational pa-
nents, such as water vapor, O2, SOx, NOx, NH3, etc., have been rameters are yet to be well explored. Commercialization of
ignored widely in literature. With a few exceptions (Hussain and membranes might not be possible without addressing these critical
Ha €gg, 2010; Merkel et al., 2010; Reijerkerk et al., 2011; Sada et al., technical and operational concerns in various scales from lab to
1992; Scholes et al., 2010), the existing studies mainly focus on a pilot plant.
binary mixture of N2 and CO2 and most of the available selectivity
values are reported only for this binary mixture. 3. Process systems design for membrane-based carbon
The existing experimental studies mainly analyze the perme- capture (MPCC)
ability of pure gases and then estimate the CO2/N2 selectivity.
However, permeability of a gas at pure versus mixture conditions The objective of membrane system engineering (MSE) is to
may vary remarkably due to different molecules' competition in design the entire membrane process systems in optimal configu-
diffusion and sorption (Koros and Fleming, 1993). Understanding rations to achieve required purities at minimum capital and oper-
the important role of minor components in the design and opera- ational expenditures (CAPEX and OPEX). Modeling and optimization
tion of membranes for carbon capture, Scholes et al. (2009) are essential in MSE work which are nowadays more facilitated by
reviewed those polymeric membranes considering the impacts of faster computers and enhanced algorithms. Membranes require
minor gases on membrane permeability, plasticization, and aging high pressures in order to generate driving force and facilitate mass
effects. They concluded that “while many minor components can transfer. With notable increase in energy prices within the last
affect performance both through competitive sorption and plasti- decade, the importance of optimal system design has remarkably
cization, much remains unknown. This limits the selection process increased in order to minimize the operational costs and thus
for membranes in this application.” One good example for illus- enhance the overall performance of membranes, especially in the
tration is regarding the effect of water vapor. There has been conditions that there are competing alternatives for membranes.
traditional thought that water, due to its higher permeability than Bernardo et al. (2009) in a thorough review of state of the art
CO2 plus solubility of CO2 in water, will cause a plasticization effect. membranes for gas separation point out the fact that the concept of
It will moreover result in part of the CO2 to be lost due to disso- application of membranes for PCC is not fully explored and “sig-
lution in water. However, Merkel et al. (2010) found that the nificant design optimization would be required in order to identify
presence of water has a beneficial effect of providing an internal efficient, feasible, and environmentally sound technical solutions.”
sweep due to diluting the permeate and thus increasing the CO2 Similar conclusions are reported elsewhere (He et al., 2013). Here,
permeation driving force through the membrane. The water con- we review membrane processes in unit and network levels and
tent of permeate could be removed through the intercoolers of the discuss the complementary nature of MMD and MSE.
multistage compression of CO2 before transport to a storage site.
Similar results on the positive impact of water have been reported 3.1. Membrane unit modeling
for PVAm facilitated transport membranes (Hussain and Ha €gg,
2010) and poly(ethylene oxide) based block copolymers There have been numerous efforts in membrane modeling over
(Reijerkerk et al., 2011). In a contradicting study, Scholes et al. the last six decades. Weller and Steiner (1950a,b) first proposed a
(2010) associated water with negative impacts. The work of method to calculate the permeate composition of a binary gas
Scholes et al. (2010) involved the experimental determination of mixture in a cross-flow process while assuming negligible pressure
the effect of minor gas components such as carbon monoxide, drop. Brubaker and Kammermeyer (1954) were the first to tackle
hydrogen sulfide, and water on the performance of a polydimethyl the topic of multicomponent membrane separation. They intro-
siloxane (PDMS) rubbery membrane in CO2/N2 separation. Exam- duced a method to solve a perfect-mixed membrane. However, the
ining the syngas of integrated gasification combined cycle pro- model was limited to three or four components, exposing the
cesses, Scholes et al. (2010) showed that the permeability of both complex nature of membrane modeling. Stern et al. (1965) pre-
CO2 and N2 over PDMS decreased with the addition of these com- sented a general algorithm for multicomponent gas mixtures. It
ponents due to competitive sorption of these gases, with the was a trial and error method in perfect-mixing flow condition.
exception of positive impact of H2S on N2 permeability. The nega- Though there were other efforts (Pan and Habgood, 1978a, b;
tive effect of water vapor on the permeability of N2 and CO2 is due Sengupta and Sirkar, 1984) that tackled this problem, it was
to the hydrophobic nature of PDMS resulting in very low water Shindo et al. (1985) who presented a complete calculation method
sorption. Using the FloryeHuggins theory, they justified that water for multicomponent gas mixtures at five different flow patterns
occupies the free volume within PDMS and then results in lower (one-side-mixing, perfect-mixing, cross flow, countercurrent, and
CO2 and N2 permeability. Low et al. (2013) have also highlighted cocurrent). The study was based on flat membranes with the
that though water vapor in flue gas can have a slight positive sweep assumption of negligible pressure drop on both sides of the
effect on process performance, its effect on membranes can be membrane. It concluded that countercurrent flow could result in
complex depending on possible interactions of sorbed water with higher permeate purity. A similar study was carried out by Li et al.
the membrane. An interesting discussion also is provided by Jiang (1990) reporting the outperformance of countercurrent flow.
(2012) on the notably complex impact of water on CO2 capture Recently, Alshehri et al. (2013) improved the Shindo et al. (1985)
over MOFs. For instance, water impact can be positive on CO2 model by adding HagenePoiseuille approximation formula for
separation at a certain pressure range and negative at another describing the permeate pressure change in the fiber lumen.
pressure range. Future studies, therefore, may need to measure the Pan (1986), expanding on his previous models (Pan, 1983; Pan
permeability and selectivity of CO2/N2 under the mixture condition, and Habgood, 1974, 1978a, b), presented a model of
multicomponent permeation with high-flux, asymmetric hollow- 1 When the concentration of the target component in the inlet
fiber membranes. The model considered the permeate pressure feed gas is high, membranes can perform very well compared to
variation inside the fiber. A thorough review of efforts done for gas other kinds of separation systems, resulting in high purity
membrane modeling till 1992 could be found in Kovvali et al. permeate. However, they fail in achieving a retentate stream
(1992). with a low concentration of target component. One example is
In a different type of modeling, Coker et al. (1998) introduced an the case of a gas sweetening process for which membrane is an
algebraic stage-wise method by dividing the length of hollow fiber economical option to separate bulk of sour gas, whilst it is not
into a series of S stages in the axial direction and mass balances capable of achieving a retentate (processed natural gas) with
were enforced in each section. They developed a model for cocur- required pipeline specifications, i.e., sour gas concentration less
rent, countercurrent, and cross flow contacting patterns consid- than 2%. Therefore, a hybrid of membrane and amine system is
ering permeate sweep, pressure-dependent permeability introduced, in which the membrane takes the responsibility of
coefficients, and bore side pressure gradients. The existence of a the bulk of separation and the amine solvent further reduces the
boundary layer on either side of the membrane results in the for- sour concentration of retentate.
mation of pressure and concentration gradients near the mem- 2 When the concentration of the target component in the inlet
brane surface (addressed as concentration polarization) which feed gas is comparatively low, the membrane can neither satisfy
reduces the driving force. high purity permeate nor low-concentration retentate.
Scholz et al. (2012) have correctly highlighted the fact that
conventional studies have generally ignored the process non- In most of the flue gas cases, the concentration of CO2 is notably
idealities and there exist very few studies which have studied the low (Table 1). At such conditions, a single-stage membrane cannot
non-idealities such as concentration polarization (Ohlrogge et al., produce high quality permeate or retentate even at very high inlet
2010), JouleeThomson effect (Gorissen, 1987), real gas behaviors pressures and/or over large membrane areas. The solution is a
(Soni et al., 2009), and pressure drop on both sides of the mem- combination of a few-stage membrane, in parallel or in series to
brane. Scholz et al. (2012) developed and validated a model reach higher qualities of permeate and retentate. Such arrange-
considering concentration polarization, the JouleeThomson effect, ments result in higher CAPEX (due to high membrane area) and
pressure losses, and real gas behavior. Their case-study on CO2/ OPEX (due to high compression costs) making the appropriacy of
propane and CO2/methane showed that the non-ideality in the case membrane debatable for such conditions of low concentration
of CO2/methane was notable. This fact might be extendable to CO2/ feeds. In such scenarios, membranes may not be the best available
N2 mixtures. We expect that future research on gas membranes will technology (BAT) and other separation technologies might be
focus more on addressing the process non-idealities. competitive. Selection of the right option will therefore require
With this background on the modeling of membrane unit, it detailed techno-economical investigations. The success of mem-
would be interesting to know that except for very limited studies brane systems will be very much dependent on process synthesis,
(e.g., Alshehri et al., 2013; Hasan et al., 2012; Khalilpour et al., 2012; configuration, and design. Here, we review the works on mem-
Merkel et al., 2010), the mainstream PCC membrane modeling tasks brane process network (in general and for PCC processes) and we
have been carried out with assumption of binary CO2/N2 mixture will discuss techno-economics in the next section.
for flue gas. We will address the binary studies in later sections As seen in many industrial applications, single-stage membrane
when we review membrane systems and the techno-economic separation has limitations in achieving high quality permeate or
studies. Merkel et al. (2010) used multicomponent flue gas in retentate while typically the objective of separation is either a high
their study, but did not present their model. Khalilpour et al. (2012), quality permeate or a high quality retentate. The reason is due to
using the multicomponent gas model of Pan (1986), investigated the permeation requiring a driving force between two membrane
the impact of numerous parameters such as area, length, pressure, sides, i.e., xjPf yjPp > 0, resulting in:
permeance, and selectivity on the separation of CO2 from flue gas
over co/countercurrent hollow fiber membranes. From a merely xj Pf xj
technical point of view, they showed that when the concentration yj < ¼
Pp b
of CO2 in the inlet feed is comparatively low (less than 15%),
membrane separation, neither in pressurized nor in vacuum con- Therefore, when the concentration of target component j in the
figurations, can satisfy high purity permeate or high recovery. Their feed gas (xj) is low, even high pressure ratios (b) may not achieve
study showed selectivity and permeability have a combined impact high permeate purity (yj). As such, more stages are required in or-
on PCC membrane performance and a high selectivity (e.g., >50) is der to achieve the desired quality and high recovery. The first so-
not merely advantageous. A similar proposal was given earlier by lution was a patent by Pfefferle (1960) who presented a two-stage
Kaldis et al. (1998) who studied flue gas with binary CO2/N2 system with permeate recycled in order to reach high purity
composition. Khalilpour et al. (2012) also identified optimal regions permeate. Following the two-stage presentation, a cascade of
of flue gas pressures and membrane area within which a techno- membrane systems was introduced for a binary gas mixture sep-
economical process system design could be carried out. Alshehri aration (Kakuta et al., 1978; Ozaki et al., 1978). Gruzdev et al. (1984)
et al. (2013) introduce a model by accommodating the proposed a method for calculation of a cascade system when there
HagenePoiseuille approximation formula in Shindo et al. (1985) in is a multicomponent gas mixture. Moreover, some studies have
order to describe the permeate pressure change in the fiber lumen. proposed using different membrane materials in the cascade sys-
With this model, they investigated the performance of a flue gas tem (Stern et al., 1984).
with four components (CO2, N2, O2, and H2O) over co- and counter- The cascade model (shown in Fig. 6) is the most general
current membrane configurations. multistage membrane design. The design idea is borrowed from
conventional mass transfer operation designs such as multi-stage
3.2. Multi-stage membrane network synthesis and design distillation or extraction systems (Treybal, 1955). The down-
stream membranes enrich permeate to higher purities of the target
There is general consensus that membranes are good for bulk components while upstream membranes strip remaining traces of
separation. We can categorize membrane processes into two major target components to desired values. In many processes, either of
groups: the enriching or stripping section is required. For instance, in the
Fig. 6. Schematic of cascade membrane system with recycle.
case of oxygen separation from air via air separation units (ASU), permeate product, design downstream section (enriching) is the
the downstream enriching section is not required. In the case of best choice. Datta and Sen (2006) techno-economically reviewed
natural gas sweetening, the objective is to minimize the concen- ten different one-, two-, and three-stage configurations and re-
tration of sour gas (CO2 and H2S) in retentate. Therefore, an up- ported that the selection of the best configuration is highly related
stream stripping section is required. Moreover, the concentration of to feed quality, separation objectives, and market parameters.
permeate, i.e., sour gas, is also important in order to minimize It can be concluded that the selection of proper multi-stage
natural gas loss in the permeate stream. Therefore, a downstream systems is highly dependent on the separation strategy which is
stripping section is also needed. mostly affected by economical inputs. When the retentate is the
It is, however, discussed that, except very low feed concentra- target product and its partial loss with permeation is not
tions (Li et al., 1990) or low-efficient membranes (Avgidou et al., economically important, the strategy will be mere stripping. For
2004), a series of two (Ozaki et al., 1978) or three stage (Pettersen mere stripping objectives, a two-stage (Fig. 7b) or three-stage
and Lien, 1995) membrane system is usually the most techno- (Fig. 9a) stripper would be selected. However, when the objective
economically optimal configuration. It is proven that the intro- is high purity permeate, the strategy will be to enrich and a two/
duction of more stages results in slightly less membrane area and three-stage enriching system (Fig. 9b) will then be chosen.
compression energy, and the increase in the number of compres- There might be a condition where high quality retentate is
sors ultimately neutralizes this advantage (Pettersen and Lien, required, but minimum loss in permeate is also desired. Alterna-
1995). tively, high quality permeate might be required with minimum loss
There have been numerous studies for optimal configurations of in the retentate stream. In such conditions, when both high quality
two/three-stage membrane systems. Studying the two-stage sys- permeate and retentate are desired, a combination approach could
tems, Kao et al. (1989) compared the continuous column mem- be Fig. 7a (Pfefferle, 1960), Fig. 8 (Bhide and Stern, 1993), or Fig. 10
brane, or CMC (Fig. 7a) (Pfefferle, 1960), to the two strippers in one-stage enriching with two-stage stripping (Pettersen and Lien,
series permeator system, TSSP (Fig. 7b). They reported sweet-spot 1995). According to Seader and Henley (2006), the best method
operating conditions for each configuration. According to them, in designing cascade systems is to choose parameters in a way that
TSSP is superior to CMC configuration unless the objective is to the composition of a permeate recycled to any stage i is identical to
minimize membrane area or to have high purity permeate. A that of feed entering the same stage.
similar result was reported by Qiu et al. (1989), in terms of eco- Attention to multistage membrane network has taken off in
nomic evaluation, stating that TSSP is the best configuration when recent years. There have been numerous efforts to develop so-
the objective is high quality retentate, while in the case of high called “superstructures” for membrane process synthesis
quality permeate, CMC is more efficient. (Agrawal, 1996; Qi and Henson, 2000; Uppaluri et al., 2004; Saif
Bhide and Stern (1993) studied seven different one-, two-, and et al., 2009; Alshehri et al., 2013; Gassner and Mare chal, 2010).
three-stage configurations with respect to minimum costs. They The earliest study, was that by Vandersluijs et al. (1992) who
found that a three-stage system with a single permeation stage in a investigated the merits of single-stage and two-stage cascade
series with a two-stage permeation cascade with recycle (Fig. 8) membrane systems for PCC application. They used the Shindo et al.
was the best. (1985) model for membrane unit modeling with the assumption of
Pettersen and Lien (1995) studied the intrinsic behavior of binary CO2/N2 flue gas. They reported that for high purity CO2
several single-stage and multi-stage permeator systems. They also products (>80%), the two-stage system can notably outperform the
divided the multi-stage system into enriching and stripping cas- single-stage process. Similar results were reported by Carapellucci
cades which could be two- or three-stages. According to them, if and Milazzo (2003) who highlighted that the two-stage design is
the objective is to have retentate with minimum concentration of a the best option for enriching CO2 stream while the addition of the
high permeable component, then the upstream section of the third stage does not notably improve the CO2 purity while increases
cascade (stripping) will be chosen. In the case of a high purity the process complexity markedly.
Fig. 7. Schematic of a) CMC and b) TSSP.
with permeate recycle. In the second step, the CO2-enriched air is

then cooled to room temperature to remove the water content. The
dry gas is then subcooled to around 20 C and sent to fraction-
ation unit. The fractionation bottom is liquid CO2 ready to be
pumped and sent for sequestration. The overhead gas is mainly
incondensable nitrogen and some CO2. This gas stream goes into
the third membrane unit where the CO2 is further removed and
recycled to the front of the membrane unit while the retentate is
vented.
The idea of combustion air sweep was also utilized by Zhai and
Fig. 8. Schematic of a single permeation stage in series with two-stage permeation
cascade. Rubin (2013b) who investigated a two-stage, two-step membrane
system with air sweep for capturing CO2 at coal-fired power plants.
In the second step, the retentate passes through a membrane with
sweep air flow to further dilute the flue gas before being vented.
Alshehri et al. (2013) introduced an N-stage membrane network
superstructure considering possible flowsheet configurations. An
optimization formulation was then developed and solved using an
objective function that minimizes the costs associated with oper-
ating and capital expenses. The membrane unit model was devel-
oped using the Shindo et al. (1985) multicomponent gas model
with the addition of pressure drop formulations inside the fiber.
The model was able to synthesize the network configuration and
identify the optimal design and operation variables. Their case-
study for a 300 MW coal-fired power plant resulted in a two-
stage membrane system for CO2/N2 selectivity of 100 and a three-
stage for selectivity of 50.
3.3. Discussion on membrane systems
A single-stage membrane is not a feasible solution for MPCC due

Fig. 9. Schematic of a) a three-stage stripping and b) a two/three-stage enriching to low CO2 content of flue gas (Table 1), though it has been widely
system. used in the modeling works for MPCC. The other issue is the sig-
nificant impacts of other flue gas components on process system
design while the mainstream studies are based just on the binary
Ho et al. (2008a) investigated three processes, i.e., a single-stage
CO2/N2 assumption. Furthermore, some studies assume a mixed
and two-stage cascades with and without retentate recycle. They
flow pattern that may be subject to error, as in practice cocurrent
studied both vacuum and pressurized scenarios for single-stage
and countercurrent flows are used. There seems a general
while for two-stage processes only the vacuum scenario was
consensus in the literature on the positive impact of vacuum
studied. They employed the Shindo et al. (1985) model for devel-
configuration and the application of sweep gas for MPCC.
oping a binary cross-flow gas separation model of the membrane
unit. According to their results, the vacuum two-stage with reten-
4. Techno-economics of MPCC systems
tate recycle could achieve the highest CO2 purity, while in terms of
cost and without considering purity, vacuum single-stage was the
A number of techno-economic studies have been carried out to
best.
evaluate the feasibility of membrane systems for removing CO2
Hussain and H€ agg (2010) simulated two process configurations
from flue gases and improve the viability of membrane technology
with facilitated transport membranes in an Aspen Hysys environ-
for carbon capture. An early study by Vandersluijs et al. (1992)
ment. The first configuration was a two-stage process with sweep
investigated the technical feasibility and mitigation costs of poly-
gases of water vapor at the first stage (vacuum mode) and permeate
mer membranes for the recovery of CO2 from flue gases of a power
at the second stage. The second configuration used recycled
plant. They employed a cross-flow permeation model in conjunc-
permeate as sweep gas for both stages. The result showed that the
tion with optimization analysis for the membrane process to
former configuration had a better performance. Merkel et al. (2010)
determine CO2 abatement costs that depend on the separation
also introduced a two-stage and two-step membrane system
targets. For the 50% product purity and 75% recovery for CO2, the
considering countercurrent flow at vacuum mode and using
minimum achievable cost was estimated to be US$48 per tonne of
incoming combustion air as a sweep gas to generate driving force.
CO2 avoided. Improving the product purity and the degree of CO2
In their process, the flue gas first goes under a two-stage membrane
recovery to the higher levels of 95% and 90% respectively would
significantly increase the cost of CO2 avoided by nearly 50% (US$71
per tonne of CO2 avoided). They illustrated that for the membranes
available at the time of their study, membrane capture technology
could not compete with solvent-based technologies; and in order
for membranes to become economically attractive for carbon cap-
ture, membranes with CO2/N2 selectivity higher than 200 along
with high permeability would be required. This threshold value
regarding the selectivity has thereafter been cited by different
membrane studies (Aresta, 2003; Favre, 2007; Feron et al., 1992;
Fig. 10. Schematic of one-stage enriching with two-stage stripping. Wolsky et al., 1994).
Kazama et al. (2004) carried out an economic analysis to eval- evaluated three membranes: Poly (phenylene oxide) (PPO), Poly-
uate their developed Cardo polyimide hollow fiber membrane with imide (PI), and PEO. The three membranes have CO2/N2 selectivities
high CO2 permeance of 1000 GPUs and CO2/N2 selectivity of 40. of 20, 30, and 50, respectively, resulting in a total sequestration cost
They illustrated that the success of membranes as an alternative to of US$55 to 61/tonne CO2 avoided. Their techno-economic assess-
existing amine-based capture systems depends on the CO2 con- ment confirmed the general consensus that such membrane se-
centration of source gases. When the inlet CO2 concentration is lectivities require a high feed gas compression rate, which makes
around 25% or more, membrane systems become economically membranes not economically competitive against MEA systems.
advantageous over amine-based capture systems. In addition, the They suggested several approaches to improve the viability of
energy penalty of vacuum pumping for producing the driving force membrane technology, including improvements in CO2 perme-
for gas separation would contribute 50% or more of the total sep- ability and selectivity and changes to the process configurations
aration cost, which implies that reducing the power use for the and operating pressures as well as reductions in the cost of
capture process would help lower the cost for removing CO2 from membranes.
flue gases. Membrane capture systems, therefore, become more Favre (2007) offered a critical comparison of dense polymeric
viable with pressurized gas streams via reducing the energy membrane processes with amine absorption for post-combustion
requirement for gas separation. carbon capture and addressed that the potential of dense poly-
Matsumiya et al. (2005) estimated energy consumption of a meric membranes to solve the flue gas treatment problem may
novel ultrafiltration hollow fiber module with facilitated transport have been underestimated. Although membranes in conjunction
membrane for the separation of CO2 from the flue gas and evalu- with a single-stage module, even with CO2/N2 selectivity as high as
ated the effects of a range of major parameters including the 120, cannot compete with amine absorption for low CO2 content,
permeate side pressure, temperature, gas and liquid flow rates, and flue gases such as those from power plants, may compete with
the inner diameter of hollow fiber membranes. In the membrane amines for application to flue gas sources with high CO2 concen-
module, both the flue gas and carrier solution (DEA) are supplied to trations above 20% (such as flue gas from cement or steel produc-
the lumen side (slightly above atmospheric pressure) of the hollow tion). That paper also comparatively presented the Robeson plots
fiber membrane, and CO2 is dissolved in the carrier solution and revealing the tradeoff between permeability and selectivity for
permeates. As soon as the solution enters the shell side, being rubbery and glassy polymers and showed that rubbery polymers
under vacuum pressure, CO2 is liberated and the solution is recy- may have relatively higher potential for flue gas treatment. How-
cled. Among the key parameters, increasing the inner diameter of ever, both of the upper bound lines appear somewhat more opti-
the hollow fiber would remarkably decrease the energy penalty. mistic than the similar ones presented by others (Bernardo et al.,
The energy consumption was reported to be in the range from 2009; Powell and Qiao, 2006; Scholes et al., 2008).
0.072 kWh/kg-CO2 (0.259 GJ/tonne-CO2) to 0.211 kWh/kg-CO2 Ho et al. (2008a) also compared the two major strategies of
(0.796 GJ/tonne-CO2) when the hollow fiber inner diameter was producing the driving force for membrane gas separation at coal-
changed from 1.4 mm to 0.8 mm. In addition, they evaluated two fired power plants: compressing feed gas stream versus vacuum
different designs for achieving the required driving force between pumping in the permeate side, which is similar to other studies
feed and permeate sides. In one design the flue gas entering the (Bounaceur et al., 2006; Matsumiya et al., 2005). They showed that
feed side was compressed while the permeate side was kept under while the vacuum strategy required relatively high membrane area,
the atmospheric pressure. In contrast, the other design kept the it could achieve 35% less capture cost per tonne of CO2 avoided,
feed flue gas under the atmospheric pressure but made a vacuum compared to the compression design. But, neither of the driving
condition in the permeate side. They found that the energy con- force designs could result in a capture cost competitive with amine
sumption required for achieving the driving force using the vacuum solvents.
strategy was significantly less than the one with compressing the Zhao et al. (2008) investigated the effects on membrane per-
feed flue gas. A similar conclusion was made by Zhai and Rubin formance (CO2 separation degree and CO2 purity) of operating
(2013b) based on the analysis for two-stage membrane systems. conditions (e.g., CO2 concentration in the flue gas, pressure and
Bounaceur et al. (2006) systematically conducted parametric temperature, etc.) as well as membrane properties (permeability,
analyses to identify the potential as well as the limitations of single- selectivity) and membrane area. They modeled a single-stage
stage membrane separation processes for CO2/N2 separation in membrane process and also found that the single-stage mem-
comparison with amine-based solvent processes. Amine absorption brane process could not achieve high purity CO2 and separation
is shown to be the best option when a high CO2 purity is desired as degree at the same time.
the high purity induces a large energy penalty for membrane sys- Merkel et al. (2010) reported development of a new membrane
tems. However, when the CO2 recovery ratio and permeate with CO2 permeances of greater than 1000 GPUs and a CO2/N2
composition do not exceed 0.8 and the CO2 concentration of flue selectivity of 50 at 30 C, which has permeance 10 times higher
gas is more than 20%, the membrane separation has significant than those of commercial CO2 membranes, but has a selectivity
advantage over amine absorption because of its lower energy within the range reported for non-facilitated transport materials.
requirement of about 0.5e1 GJ/tonne-CO2 recovered. This study They developed a novel two-stage membrane system using com-
concluded that for the membranes with CO2/N2 selectivities less bustion air as sweep gas that carries a portion of permeated CO2
than 50, the membrane process is not feasible and higher selec- back to the boiler and finally increases the CO2 concentration of flue
tivities more than 100 are required to facilitate the feasibility for gas into the capture unit. Their study claimed that the membrane
carbon capture. Similarly, another study also claimed that MEA- process with combustion air sweep for 90% CO2 capture would
based capture technology is less expensive than membrane cap- account for 16% of the power output of a coal-fired power plant and
ture technology even for CO2 purity requirements as low as 60% (Ho result in a capture cost as low as $23/tonne CO2. Their study also
et al., 2006b). suggested that improving the CO2 permeance may become more
Ho et al. (2006a) compared the feasibility of a single-stage important than increasing the CO2/N2 selectivity to reduce the
membrane process for PCC against an amine-based system and overall cost.
examined the effects of membrane characteristics, operating pa- Hussain and Ha €gg (2010) used a novel CO2-selective facilitated
rameters, and system design on sequestration costs. They used the transport membrane (PVAm) with selectivity of 200 in their
Shindo et al. (1985) model considering binary flue gas and techno-economic study. They compared two-stage membrane
processes with amine absorption in terms of energy requirement. versus amine-based CCS systems are basically limited to the 90%
They reported that membrane process using the facilitated trans- CO2 capture efficiency. However, the costs of CO2 avoided or
port membrane is feasible, even for low CO2 concentration (10%) in captured by two types of CCS systems appear different trends as a
flue gas while achieving more than 90% CO2 purity and recovery. function of CO2 capture efficiency. For a typical amine-based cap-
Zhai and Rubin (2013b) evaluated the performance and costs of ture system, the cost-effective CO2 capture often happens to the
multiple membrane capture systems including single- and multi- 90% removal efficiency. In contrast, the minimum cost of CO2
stage configurations on a common framework. Multi-stage mem- captured by a two-step/sweep membrane process is at about 70% of
brane systems were identified to be able to simultaneously achieve removal efficiency (Merkel et al., 2010). Therefore, more careful
the 90% removal efficiency and 95% or more product purity for CO2 process development and economic assessments are needed for
at pulverized coal power plants. The comparative assessment re- membrane systems designed for partial CO2 capture.
sults indicate that the multi-stage membrane system with com-
bustion air sweep is capable of achieving the separation targets at 5. Conclusion
an avoidance cost that is 15% less than that of an amine-based
capture process. Consistent with the finding by Merkel et al. There has been an increasing interest in the application of
(2010), enhancing the CO2 permeability of membrane materials membranes as an alternative technology in post-combustion car-
appear to be more economically attractive than increasing the CO2/ bon capture (PCC). Historically, there has not been a general
N2 selectivity. However, they also highlighted a few key caveats agreement on the advantage of membranes against other methods
such as the influences of minor gas components and humidity on such as amine-based absorptionedesorption processes for CO2
membrane performance. capture. In this paper, we reviewed the progress of membrane
based PCC (MPCC) from both material and process design per-
4.1. Discussion on MPCC economics spectives plus engineering economics.
The research on polymeric membranes for carbon capture likely
The utilization of diverse membrane process models (simplified move towards improving their performance by incorporating
or rigorous) might be one of the main sources of discrepancies additional agents into the polymeric phase that could be another
amongst the techno-economic studies, whilst efficient general polymer (polymeric blends), particulate matter (mixed matrix
models already exist in the literature. Therefore, it appears neces- membranes), or a carrier molecule (facilitated transport). Facili-
sary that future studies employ more rigorous models for techno- tated transport membranes (e.g., PVA based) seem to be promising
economic assessments. for carbon capture. Membrane gas contactors have been suffering
More importantly, though technical and cost studies along with from the additional mass transfer resistance. However, a number of
optimization analyses have been conducted actively for membrane approaches to modify membrane materials are being investigated
capture systems, the costing scope and methodology appear to improve the performance of gas absorption membranes. These
inconsistent across existing studies (Zhai and Rubin, 2013b), which include: plasma treatment, applying a thin hydrophobic layer to
makes it hard for cost comparisons among different membrane base membrane materials, incorporation of salts into the mem-
materials and process designs. There are big differences in the branes during manufacture, and changing manufacturing param-
capital costing scope across various studies. Many cost studies eters to alter membrane characteristics such as pore diameter.
include direct cost components such as membrane module, sup- Permeability and selectivity are two critical attributes of a good
port framework, and power equipment, but largely ignore such membrane. An industrially desirable membrane, however, requires
indirect capital costs as process and project contingencies and (generally and specifically for carbon capture) other key features
owner's costs, which can account for more than 40% of total capital such as chemical and mechanical compatibility with the process
cost. As one of the key cost components of a CCS (carbon capture environment, stability, freedom from fouling, reasonable useful
and sequestration) system, CO2 product compression is not lifetime, amenability to fabrication and packaging, and resistance to
included in the scope of some techno-economic studies that eval- high pressures. Nevertheless, most of the existing studies on CO2
uate the feasibility of membrane capture systems. The cost metric capture membrane materials are focused on improving permse-
of “cost of CO2 captured” has been adopted by many membrane lectivity and they lack sufficient attention to other important re-
studies, but this measure typically employs arbitrary electricity quirements. The impacts of minor gas components such as water
price assumptions that vary by more than a factor of two across vapor, O2, SOx, NOx, NH3, etc. have been ignored widely in the
existing studies, which is inappropriate for reporting and literature and the existing studies, with few exceptions, mainly
comparing CCS costs because of the big energy requirement of CCS focus on a binary mixture of N2 and CO2. As such, commercializa-
systems (Zhai and Rubin, 2013b). Rubin et al. (2013) proposed a tion of membranes might not be possible without addressing these
common costing methodology and provided detailed guidelines to critical technical and operational concerns in various perspectives.
improve the clarity and consistency of CCS cost estimates. The widely used single-stage membrane configuration is not a
Furthermore, the deployment of CCS in fossil fuel-fired power feasible solution for MPCC due to the low CO2 content of flue gas,
plants likely will start with partial CO2 capture. In September 2013, even for membranes with high permselectivities. In the literature,
the U.S. Environmental Protection Agency proposed emission per- there is an agreement with the positive impact of vacuum config-
formance standards to limit CO2 emissions from new fossil fuel- uration in reducing the system energy penalty. Multi-stage mem-
fired power plants. The compliance with the proposed emission brane systems are feasible to simultaneously achieve the separation
performance standards would require roughly 40e50% CO2 capture targets of high product purity and high removal efficiency for CO2.
for new pulverized coal power plants, depending on plant designs To improve the membrane technology's viability, optimizing the
(Zhai and Rubin, 2013a). However, in the existing studies, the design of driving force for membrane gas separation appears
process designs and economics evaluation of membrane capture important to reduce the energy penalty and the capture cost for
systems have been focused heavily on the 90% CO2 capture effi- membrane capture systems. Although recycling a portion of
ciency. Although some studies have evaluated partial CO2 capture, captured CO2 via the gas sweep design may increase the CO2 partial
the CO2 product of membrane capture processes has a lower purity pressure of feed flue gas, and in turn, remarkably reduce the system
less than 90e95%, which is not suitable for transport and geological power use, the effects of CO2-rich air combustion on traditional
storage. In addition, the existing cost comparisons for membrane- boiler performance and other environmental control systems
remain unclear, which needs more careful analyses. Many existing Agrawal, R., 1996. Membrane cascade schemes for multicomponent gas separation.
Ind. Eng. Chem. Res. 35, 3607e3617.
studies only evaluate the performance and costs of membrane
Ahmad, A.L., Jawad, Z.A., Low, S.C., Zein, S.H.S., 2014. A cellulose acetate/multi-
capture systems alone, but do not explore the relations of power walled carbon nanotube mixed matrix membrane for CO2/N2 separation.
generation system designs to the process parameters and material J. Memb. Sci. 451, 55e66.
properties that influence the overall performance and cost of Ahmad, A.L., Mohammed, H.N., Ooi, B.S., Leo, C.P., 2013. Deposition of a polymeric
porous superhydrophobic thin layer on the surface of poly(vinylidene fluoride)
membrane capture technologies. A plant-level analysis is needed to hollow fiber membrane. Pol. J. Chem. Technol. 15, 1e6.
help explore tradeoffs in meeting performance and cost objectives Alshehri, A., Khalilpour, R., Abbas, A., Lai, Z., 2013. Membrane systems engineering
and identify the most promising system designs and R&D targets of for post-combustion carbon capture. Energy Proced. 37, 976e985.
Aresta, M., 2003. Carbon Dioxide Recovery and Utilization. Kluwer Academic Pub-
material properties for advancing membrane capture technologies. lishers, Dordrecht, Boston.
The differences in gas separation and process models (multi- Aroon, M.A., Ismail, A.F., Matsuura, T., 2013. Beta-cyclodextrin functionalized
component versus binary, multi-stage versus single-stage, etc.) MWCNT: a potential nano-membrane material for mixed matrix gas separation
membranes development. Sep. Purif. Technol. 115, 39e50.
employed by existing membrane studies are considered to be one Avgidou, M.S., Kaldis, S.P., Sakellaropoulos, G.P., 2004. Membrane cascade schemes
of the key sources of discrepancies amongst the techno-economic for the separation of LPG olefins and paraffins. J. Memb. Sci. 233, 21e37.
studies. Furthermore, the costing scope and methodology appear Bae, B., Chun, B.H., Kim, D., 2001. Surface characterization of microporous
polypropylene membranes modified by plasma treatment. Polymer 42,
inconsistent across existing studies and application of a common 7879e7885.
costing methodology could improve the clarity and consistency of Bai, H., Ho, W.S.W., 2011. Carbon dioxide-selective membranes for high-pressure
MPCC cost estimates. synthesis gas purification. Ind. Eng. Chem. Res. 50, 12152e12161.
Baker, R.W., Lokhandwala, K., 2008. Natural gas processing with membranes: an
overview. Ind. Eng. Chem. Res. 47, 2109e2121.
List of abbreviations Bastani, D., Esmaeili, N., Asadollahi, M., 2013. Polymeric mixed matrix membranes
containing zeolites as a filler for gas separation applications: a review. J. Ind.
Eng. Chem. 19, 375e393.
ASU air separation units Bernardo, P., Drioli, E., Golemme, G., 2009. Membrane gas separation: a review/state
BAT best available technology of the art. Ind. Eng. Chem. Res. 48, 4638e4663.
BAS bis(aminopropyl)polydimethylsiloxane Bhide, B.D., Stern, S.A., 1993. Membrane processes for the removal of acid gases
from natural gas. 1. Process configurations and optimization of operating-con-
CA cellulose acetate ditions. J. Memb. Sci. 81, 209e237.
CAPEX capital and operational expenditure Bondar, V.I., Freeman, B.D., Pinnau, I., 2000. Gas transport properties of poly(ether-
CCS carbon capture and sequestration b-amide) segmented block copolymers. J. Polym. Sci. B Polym. Phys. 38,
2051e2062.
CMC continuous membrane column Bottino, A., Capannelli, G., Comite, A., Di Felice, R., Firpo, R., 2008. CO2 removal from
CRGT chemically recuperated gas turbine a gas stream by membrane contactor. Sep. Purif. Technol. 59, 85e90.
DADE diamino diphenyl ether Bounaceur, R., Lape, N., Roizard, D., Vallieres, C., Favre, E., 2006. Membrane pro-
cesses for post-combustion carbon dioxide capture: a parametric study. Energy
EOR enhanced oil recovery 31, 2556e2570.
GPU gas permeation unit Brubaker, D.W., Kammermeyer, K., 1954. Separation of gases by plastic membranes
IPCC Intergovernmental Panel on Climate Change e permeation rates and extent of separation. Ind. Eng. Chem. 46, 733e739.
Carapellucci, R., Milazzo, A., 2003. Membrane systems for CO2 capture and their
LDPE low-density polyethylene
integration with gas turbine plants. Proc. Inst. Mech. Eng. A J. Power Energy 217,
MEA monoethanolamine 505e517.
MGC membrane gas contactors Charmette, C., Sanchez, J., Gramain, P., Rudatsikira, A., 2004. Gas transport prop-
erties of poly(ethylene oxide-co-epichlorohydrin) membranes. J. Memb. Sci.
MMD membrane material design
230, 161e169.
MPCC membrane-based carbon capture Coker, D.T., Freeman, B.D., Fleming, G.K., 1998. Modeling multicomponent gas
MSE membrane systems engineering separation using hollow-fiber membrane contactors. AIChE J. 44, 1289e1302.
MTMS methyltrimethoxysilane Datta, A.K., Sen, P.K., 2006. Optimization of membrane unit for removing carbon
dioxide from natural gas. J. Memb. Sci. 283, 291e300.
OPEX operational expenditure Deng, L.Y., Kim, T.J., Hagg, M.B., 2009. Facilitated transport of CO2 in novel PVAm/
PA polyamides PVA blend membrane. J. Memb. Sci. 340, 154e163.
PAA poly(allylamine) Desideri, U., Paolucci, A., 1999. Performance modelling of a carbon dioxide removal
system for power plants. Energy Convers. Manag. 40, 1899e1915.
PCC post-combustion carbon capture Ettouneya, H.M., Al-Enezia, G., Hughesa, R., 1995. Modelling of enrichment of nat-
PE polyethylene ural gas wells by membranes. Gas Sep. Purif. 9, 3e11.
PEO poly(ethylene oxide) Favre, E., 2007. Carbon dioxide recovery from post-combustion processes: can gas
permeation membranes compete with absorption? J. Memb. Sci. 294, 50e59.
PEP poly(epichlorohydrin) Feron, P.H.M., 2009. The potential for improvement of the energy performance of
PES polyethersulfones pulverized coal fired power stations with post-combustion capture of carbon
PI polyamides dioxide. Energy Proced. 1, 1067e1074.
Feron, P.H.M., Jansen, A.E., Klaassen, R., 1992. Membrane technology in carbon-
PP polypropylene
dioxide removal. Energy Convers. Manag. 33, 421e428.
PTFE polytetrafluoroethylene Franco, J.A., deMontigny, D.D., Kentish, S.E., Perera, J.M., Stevens, G.W., 2012. Poly-
PU polyurethanes tetrafluoroethylene (PTFE)-sputtered polypropylene membranes for carbon
dioxide separation in membrane gas absorption: hollow fiber configuration.
PVA poly(vinyl alcohol)
Ind. Eng. Chem. Res. 51, 1376e1382.
PVDF poly(vinylidene fluoride) Gassner, M., Mare chal, F., 2010. Combined mass and energy integration in process
SLM support liquid membranes design at the example of membrane-based gas separation systems. Comput.
SPBI sulfonated polybenzimidazole Chem. Eng. 34, 2033e2042.
Ghasem, N., Al-Marzouqi, M., Rahim, N.A., 2012. Effect of polymer extrusion tem-
TCDA alicyclic 2,3,5-tricarboxy cyclopentyl acetic dianhydride perature on poly(vinylidene fluoride) hollow fiber membranes: properties and
TSSP two stripper in series permeator performance used as gaseliquid membrane contactor for CO2 absorption. Sep.
UN-FCCC United Nations Framework Convention on Climate Purif. Technol. 99, 91e103.
Gorissen, H., 1987. Temperature changes involved in membrane gas separations.
Change Chem. Eng. Process. Process Intensif. 22, 63e67.
VDF vinylidene fluoride Gruzdev, E.B., Laguntsov, N.I., Nikolaev, B.I., Todosiev, A.P., Sulaberidze, G.A., 1984.
A method of calculating membrane-element cascades for separating multi-
component mixtures. Sov. Atom. Energy 57, 553e557.
References Hasan, M.M.F., Baliban, R.C., Elia, J.A., Floudas, C.A., 2012. Modeling, simulation, and
optimization of postcombustion CO2 capture for variable feed concentration
Aaron, D., Tsouris, C., 2005. Separation of CO2 from flue gas: a review. Sep. Sci. and flow rate. 1. Chemical absorption and membrane processes. Ind. Eng. Chem.
Technol. 40, 321e348. Res. 51, 15642e15664.
He, X., H€ agg, M.-B., 2011. Hollow fiber carbon membranes: investigations for CO2 Lin, Y.-F., Chen, C.-H., Tung, K.-L., Wei, T.-Y., Lu, S.-Y., Chang, K.-S., 2013. Mesoporous
capture. J. Memb. Sci. 378, 1e9. fluorocarbon-modified silica aerogel membranes enabling long-term contin-
He, X., H€ agg, M.-B., 2012. Membranes for environmentally friendly energy pro- uous CO2 capture with large absorption flux enhancements. ChemSusChem 6,
cesses. Membranes 2, 706e726. 437e442.
He, X., Yu, Q., Ha €gg, M.-B., Hoek, E.M.V., Tarabara, V.V., 2013. CO2 Capture, Ency- Lin, Y.-F., Ko, C.-C., Chen, C.-H., Tung, K.-L., Chang, K.-S., 2014a. Reusable
clopedia of Membrane Science and Technology. John Wiley & Sons, Inc. methyltrimethoxysilane-based mesoporous water-repellent silica aerogel
Herzog, H., Drake, E., Adams, E., 1997. CO2 Capture, Reuse, and Storage Technologies membranes for CO2 capture. RSC Adv. 4, 1456e1459.
for Mitigating Global Climate Change. Lin, Y.-F., Wang, C.-S., Ko, C.-C., Chen, C.-H., Chang, K.-S., Tung, K.-L., Lee, K.-R.,
Ho, M.T., Allinson, G., Wiley, D.E., 2006a. Comparison of CO2 separation options for 2014b. Polyvinylidene fluoride/siloxane nanofibrous membranes for long-term
geo-sequestration: are membranes competitive? Desalination 192, 288e295. continuous CO2-capture with large absorption-flux enhancement. Chem-
Ho, M.T., Allinson, G.W., Wiley, D.E., 2008a. Reducing the cost of CO2 capture SusChem 7, 604e609.
from flue gases using membrane technology. Ind. Eng. Chem. Res. 47, Long, R., Felber, B., Enick, R., 2006. CO2 EOR Technology e Technologies for To-
1562e1568. morrow's E&P Paradigms. U. S. Department of Energy.
Ho, M.T., Allinson, G.W., Wiley, D.E., 2008b. Reducing the cost of CO2 capture from Low, B.T., Zhao, L., Merkel, T.C., Weber, M., Stolten, D., 2013. A parametric study of
flue gases using pressure swing adsorption. Ind. Eng. Chem. Res. 47, the impact of membrane materials and process operating conditions on carbon
4883e4890. capture from humidified flue gas. J. Memb. Sci. 431, 139e155.
Ho, M.T., Leamon, G., Allinson, G.W., Wiley, D.E., 2006b. Economics of CO2 and Mansourizadeh, A., Ismail, A.F., 2009. Hollow fiber gaseliquid membrane contactors
mixed gas geosequestration of flue gas using gas separation membranes. Ind. for acid gas capture: a review. J. Hazard. Mater. 171, 38e53.
Eng. Chem. Res. 45, 2546e2552. Matsumiya, N., Teramoto, M., Kitada, S., Matsuyama, H., 2005. Evaluation of energy
Hussain, A., H€ agg, M.-B., 2010. A feasibility study of CO2 capture from flue gas by a consumption for separation of CO2 in flue gas by hollow fiber facilitated
facilitated transport membrane. J. Memb. Sci. 359, 140e148. transport membrane module with permeation of amine solution. Sep. Purif.
Jiang, J., 2012. Metal-organic frameworks for CO2 capture: what are learned from Technol. 46, 26e32.
molecular simulations. In: Ortiz, O.L., Ramirez, L.D. (Eds.), Coordination Poly- Merkel, T.C., Lin, H., Wei, X., Baker, R., 2010. Power plant post-combustion carbon
mers and Metal Organic Frameworks. Nova Science Publishers, Inc., dioxide capture: an opportunity for membranes. J. Memb. Sci. 359, 126e139.
pp. 225e247. Metz, B., Davidson, O., Coninck, H.D., Loos, M., Meyer, L., 2005. IPCC Special Report
Junaidi, M.U.M., Leo, C.P., Ahmad, A.L., Kamal, S.N.M., Chew, T.L., 2014. Carbon di- on Carbon Dioxide Capture and Storage. Cambridge University Press, for the
oxide separation using asymmetric polysulfone mixed matrix membranes Intergovernmental Panel on Climate Change, Cambridge.
incorporated with SAPO-34 zeolite. Fuel Process. Technol. 118, 125e132. Mosadegh-Sedghi, S., Rodrigue, D., Brisson, J., Iliuta, M.C., 2013. Highly hydrophobic
Kakuta, A., Ozaki, O., Ohno, M., 1978. Gas permeation in freeze-dried cellulose-ac- microporous low-density polyethylene hollow fiber membranes by melt-extrusion
etate membrane. J. Polym. Sci. A Polym. Chem. 16, 3249e3257. coupled with salt-leaching technique. Polym. Adv. Technol. 24, 584e592.
Kaldis, S.P., Kapantaidakis, G.C., Papadopoulos, T.I., Sakellaropoulos, G.P., 1998. Nafisi, V., Hagg, M.B., 2014. Development of dual layer of ZIF-8/PEBAX-2533 mixed
Simulation of binary gas separation in hollow fiber asymmetric membranes by matrix membrane for CO2 capture. J. Memb. Sci. 459, 244e255.
orthogonal collocation. J. Memb. Sci. 142, 43e59. Nakagawa, T., Nishimura, T., Higuchi, A., 2002. Morphology and gas permeability in
Kao, Y.K., Qiu, M.M., Hwang, S.T., 1989. Critical evaluations of 2 membrane gas copolyimides containing polydimethylsiloxane block. J. Memb. Sci. 206,
permeator designs e continuous membrane column and 2 strippers in series. 149e163.
Ind. Eng. Chem. Res. 28, 1514e1520. Nik, O.G., Chen, X.Y., Kaliaguine, S., 2011. Amine-functionalized zeolite FAU/EMT-
Kasahara, S., Kamio, E., Ishigami, T., Matsuyama, H., 2012. Effect of water in ionic polyimide mixed matrix membranes for CO2/CH4 separation. J. Memb. Sci.
liquids on CO2 permeability in amino acid ionic liquid-based facilitated trans- 379, 468e478.
port membranes. J. Memb. Sci. 415, 168e175. Ohlrogge, K., Wind, J., Brinkmann, T., 2010. Membranes for recovery of volatile
Kawakami, M., Iwanaga, H., Hara, Y., Iwamoto, M., Kagawa, S., 1982. Gas perme- organic compounds. In: Enrico, D., Lidietta, G. (Eds.), Comprehensive Membrane
abilities of cellulose nitrate poly(ethylene glycol) blend membranes. J. Appl. Science and Engineering. Elsevier, Oxford, pp. 213e242.
Polym. Sci. 27, 2387e2393. Ozaki, O., Heki, H., Ohno, M., 1978. Characteristics of separation cell with 2 kinds of
Kazama, S., Morimoto, S., Tanaka, S., Mano, H., Yashima, T., Yamada, K., Haraya, K., membrane differing in gas permeability tendency. J. Atom. Energy Soc. Jpn. 20,
2004. Cardo polyimide membranes for CO2 capture from flue gases. In: Seventh 723e725.
International Conference on Greenhouse Gas Control Technologies. Greenhouse Ozturk, B., Hughes, R., 2012. Evaluation of mass transfer characteristics of non-
Gas Control Technologies, vol. I, pp. 75e82. Vancouver. porous and microporous membrane contactors for the removal of CO2. Chem.
Khaisri, S., Demontigny, D., Tontiwachwuthikul, P., Jiraratananon, R., 2009. Eng. J. 195, 122e131.
Comparing membrane resistance and absorption performance of three different Pan, C.Y., 1983. Gas separation by permeators with high-flux asymmetric mem-
membranes in a gas absorption membrane contactor. Sep. Purif. Technol. 65, branes. AIChE J. 29, 545e555.
290e297. Pan, C.Y., 1986. Gas separation by high-flux, asymmetric hollow-fiber membrane.
Khalilpour, R., Abbas, A., Lai, Z.P., Pinnau, I., 2012. Modeling and parametric analysis AIChE J. 32, 2020e2027.
of hollow fiber membrane system for carbon capture from multicomponent flue Pan, C.Y., Habgood, H.W., 1974. Analysis of single-stage gaseous permeation process.
gas. AIChE J. 58, 1550e1561. Ind. Eng. Chem. Fundam. 13, 323e331.
Kim, J.H., Ha, S.Y., Nam, S.Y., Rhim, J.W., Baek, K.H., Lee, Y.M., 2001. Selective Pan, C.Y., Habgood, H.W., 1978a. Gas separation by permeation .1. Calculation
permeation of CO2 through pore-filled polyacrylonitrile membrane with methods and parametric analysis. Can. J. Chem. Eng. 56, 197e209.
poly(ethylene glycol). J. Memb. Sci. 186, 97e107. Pan, C.Y., Habgood, H.W., 1978b. Gas separation by permeation .2. Effect of permeate
Kim, Y.-S., Yang, S.-M., 2000. Absorption of carbon dioxide through hollow fiber pressure-drop and choice of permeate pressure. Can. J. Chem. Eng. 56, 210e217.
membranes using various aqueous absorbents. Sep. Purif. Technol. 21. Pera-Titus, M., 2014. Porous inorganic membranes for CO2 capture: present and
Koros, W.J., Fleming, G.K., 1993. Membrane-based gas separation. J. Memb. Sci. 83, prospects. Chem. Rev. 114, 1413e1492.
1e80. Perez, E.V., Balkus, K.J., Ferraris, J.P., Musselman, I.H., 2009. Mixed-matrix mem-
Kovvali, A.S., Vemury, S., Krovvidi, K.R., Khan, A.A., 1992. Models and analyses of branes containing MOF-5 for gas separations. J. Memb. Sci. 328, 165e173.
membrane gas permeators. J. Memb. Sci. 73, 1e23. Peters, L., Hussain, A., Follmann, M., Melin, T., Ha €gg, M.B., 2011. CO2 removal from
Krishna, R., van Baten, J.M., 2011. In silico screening of metal-organic frameworks in natural gas by employing amine absorption and membrane technologyda
separation applications. Phys. Chem. Chem. Phys. 13, 10593e10616. technical and economical analysis. Chem. Eng. J. 172, 952e960.
Krishnamurthy, S., Rao, V.R., Guntuka, S., Sharratt, P., Haghpanah, R., Rajendran, A., Pettersen, T., Lien, K.M., 1995. Design studies of membrane permeator processes for
Amanullah, M., Karimi, I.A., Farooq, S., 2014. CO2 capture from dry flue gas by gas separation. Gas Sep. Purif. 9, 151e169.
vacuum swing adsorption: a pilot plant study. AIChE J. 60, 1830e1842. Pfefferle, W.C., 1960. Diffusion Purification of Gases. USA.
Kumazawa, H., Wang, J.S., Sada, E., 1993. Gas-transport through homogeneous and Powell, C.E., Qiao, G.G., 2006. Polymeric CO2/N2 gas separation membranes for the
asymmetric polyethersulfone membranes. J. Polym. Sci. B Polym. Phys. 31, capture of carbon dioxide from power plant flue gases. J. Memb. Sci. 279, 1e49.
881e886. Qi, R.H., Henson, M.A., 2000. Membrane system design for multicomponent gas
Leblanc, O.H., Ward, W.J., Matson, S.L., Kimura, S.G., 1980. Facilitated transport in mixtures via mixed-integer nonlinear programming. Comput. Chem. Eng. 24,
ion-exchange membranes. J. Memb. Sci. 6, 339e343. 2719e2737.
Li, J.T., Nagai, K., Nakagawa, T., Wang, S., 1995. Preparation of polyethyleneglycol Qiu, M.M., Hwang, S.T., Kao, Y.K., 1989. Economic-evaluation of gas membrane
(PEG) and cellulose-acetate (CA) blend membranes and their gas permeabil- separator designs. Ind. Eng. Chem. Res. 28, 1670e1677.
ities. J. Appl. Polym. Sci. 58, 1455e1463. Rajabi, Z., Moghadassi, A.R., Hosseini, S.M., Mohammadi, M., 2013. Preparation and
Li, K., Acharya, D.R., Hughes, R., 1990. Mathematical-modeling of multicomponent characterization of polyvinylchloride based mixed matrix membrane filled with
membrane permeators. J. Memb. Sci. 52, 205e219. multi walled carbon nano tubes for carbon dioxide separation. J. Ind. Eng. Chem.
Lin, H., Freeman, B.D., 2004. Gas solubility, diffusivity and permeability in poly(- 19, 347e352.
ethylene oxide). J. Memb. Sci. 239, 105e117. Rajabzadeh, S., Yoshimoto, S., Teramoto, M., Al-Marzouqi, M., Ohmukai, Y.,
Lin, S.H., Tung, K.L., Chang, H.W., Lee, K.R., 2009a. Influence of fluorocarbon flat- Maruyama, T., Matsuyama, H., 2013. Effect of membrane structure on gas ab-
membrane hydrophobicity on carbon dioxide recovery. Chemosphere 75, sorption performance and long-term stability of membrane contactors. Sep.
1410e1416. Purif. Technol. 108, 65e73.
Lin, S.H., Tung, K.L., Chen, W.J., Chang, H.W., 2009b. Absorption of carbon dioxide by Rao, A.B., Rubin, E.S., 2002. A technical, economic, and environmental assessment of
mixed piperazine-alkanolamine absorbent in a plasma-modified polypropylene amine-based CO2 capture technology for power plant greenhouse gas control.
hollow fiber contactor. J. Memb. Sci. 333, 30e37. Environ. Sci. Technol. 36, 4467e4475.
Reijerkerk, S.R., Jordana, R., Nijmeijer, K., Wessling, M., 2011. Highly hydrophilic, Tokuda, Y., Fujisawa, E., Okabayashi, N., Matsumiya, N., Takagi, K., Mano, H.,
rubbery membranes for CO2 capture and dehydration of flue gas. Int. J. Greenh. Haraya, K., Sato, M., 1997. Development of hollow fiber membranes for CO2
Gas Control 5, 26e36. separation. Energy Convers. Manag. 38, S111eS116.
Rubin, E.S., Short, C., Booras, G., Davison, J., Ekstrom, C., Matuszewski, M., Mccoy, S., Toy, L., Kataria, A., Gupta, R.P., 2012. CO2 Capture Membrane Process for Power Plant
2013. A proposed methodology for CO2 capture and storage cost estimates. Int. Flue Gas. RTI International, North Carolina.
J. Greenh. Gas Control 17, 488e503. Treybal, R.E., 1955. Mass-transfer Operations. McGraw-Hill, New York.
Sada, E., Kumazawa, H., Wang, J.S., Koizumi, M., 1992. Separation of carbon-dioxide Uppaluri, R.V.S., Linke, P., Kokossis, A.C., 2004. Synthesis and optimization of gas
by asymmetric hollow fiber membrane of cellulose triacetate. J. Appl. Polym. permeation membrane networks. Ind. Eng. Chem. Res. 43, 4305e4322.
Sci. 45, 2181e2186. Vandersluijs, J.P., Hendriks, C.A., Blok, K., 1992. Feasibility of polymer membranes
Saif, Y., Elkamel, A., Pritzker, M., 2009. Superstructure optimization for the synthesis for carbon-dioxide recovery from flue-gases. Energy Convers. Manag. 33,
of chemical process flowsheets: application to optimal hybrid membrane sys- 429e436.
tems. Eng. Optim. 41, 327e350. Wang, Z., Fang, M.X., Yan, S.P., Yu, H., Wei, C.C., Luo, Z.Y., 2013. Optimization of
Scholes, C., Kentish, S., Stevens, G., 2009. Effects of minor components in carbon blended amines for CO2 absorption in a hollow-fiber membrane contactor. Ind.
dioxide capture using polymeric gas separation membranes. Sep. Purif. Rev. 38, Eng. Chem. Res. 52, 12170e12182.
1e44. Weller, S., Steiner, W.A., 1950a. Engineering aspects of separation of gases e frac-
Scholes, C.A., Kentish, S.E., Stevens, G.W., 2008. Carbon dioxide separation through tional permeation through membranes. Chem. Eng. Prog. 46, 585e590.
polymeric membrane systems for flue gas applications. Recent Patents Chem. Weller, S., Steiner, W.A., 1950b. Separation of gases by fractional permeation
Eng. 1, 52e66. through membranes. J. Appl. Phys. 21, 279e283.
Scholes, C.A., Simioni, M., Qader, A., Stevens, G.W., Kentish, S.E., 2012. Membrane Wolsky, A.M., Daniels, E.J., Jody, B.J., 1994. CO2 capture from the flue-gas of con-
gasesolvent contactor trials of CO2 absorption from syngas. Chem. Eng. J. ventional fossil-fuel-fired power-plants. Environ. Prog. 13, 214e219.
195e196, 188e197. Yan, S.-P., Fang, M.-X., Zhang, W.-F., Wang, S.-Y., Xu, Z.-K., Luo, Z.-Y., Cen, K.-F., 2007.
Scholes, C.A., Stevens, G.W., Kentish, S.E., 2010. The effect of hydrogen sulfide, Experimental study on the separation of CO2 from flue gas using hollow fiber
carbon monoxide and water on the performance of a PDMS membrane in membrane contactors without wetting. Fuel Process. Technol. 88, 501e511.
carbon dioxide/nitrogen separation. J. Memb. Sci. 350, 189e199. Yeon, S.H., Sea, B., Park, Y.I., Lee, K.H., 2003. Determination of mass transfer rates in
Scholz, M., Harlacher, T., Melin, T., Wessling, M., 2012. Modeling gas permeation by PVDF and PTFE hollow fiber membranes for CO2 absorption. Sep. Sci. Technol.
linking nonideal effects. Ind. Eng. Chem. Res. 52, 1079e1088. 38, 271e293.
Seader, J.D., Henley, E.J., 2006. Separation Process Principles, second ed. Wiley, Yoshino, M., Ito, K., Kita, H., Okamoto, K.I., 2000. Effects of hard-segment polymers
Hoboken, NJ. on CO2/N2 gas-separation properties of poly(ethylene oxide)-segmented co-
Sengupta, A., Sirkar, K.K., 1984. Multicomponent gas separation by an asymmetric polymers. J. Polym. Sci. B Polym. Phys. 38, 1707e1715.
permeator containing 2 different membranes. J. Memb. Sci. 21, 73e109. Zevenhoven, R., Fagerlund, J.I.C.S.E.E.S., 2009. CO2 fixation by mineral matter; the
Shahid, S., Nijmeijer, K., 2014. High pressure gas separation performance of mixed- potential of different mineralization routes. IOP Conf. Ser. Earth Environ. Sci. 6.
matrix polymer membranes containing mesoporous Fe(BTC). J. Memb. Sci. 459, Zhai, H., Rubin, E.S., 2013a. Comparative performance and cost assessments of coal-
33e44. and natural-gas-fired power plants under a CO2 emission performance standard
Shindo, Y., Hakuta, T., Yoshitome, H., Inoue, H., 1985. Calculation methods for regulation. Energy Fuels 27, 4290e4301.
multicomponent gas separation by permeation. Sep. Sci. Technol. 20, 445e459. Zhai, H., Rubin, E.S., 2013b. Techno-economic assessment of polymer membrane
Sidhoum, M., Sengupta, A., Sirkar, K.K., 1988. Asymmetric cellulose-acetate hollow systems for postcombustion carbon capture at coal-fired power plants. Environ.
fibers e studies in gas permeation. AIChE J. 34, 417e425. Sci. Technol. 47, 3006e3014.
Soni, V., Abildskov, J., Jonsson, G., Gani, R., 2009. A general model for membrane- Zhang, H.-Y., Wang, R., Liang, D.T., Tay, J.H., 2008. Theoretical and experimental
based separation processes. Comput. Chem. Eng. 33, 644e659. studies of membrane wetting in the membrane gaseliquid contacting process
Stern, S.A., Perrin, J.E., Naimon, E.J., 1984. Recycle and multimembrane permeators for CO2 absorption. J. Memb. Sci. 308, 162e170.
for gas separations. J. Memb. Sci. 20, 25e43. Zhao, L., Riensche, E., Menzer, R., Blum, L., Stolten, D., 2008. A parametric study of
Stern, S.A., Sinclair, T.F., Gareis, P.J., Vahldiec, N.P., Mohr, P.H., 1965. Helium recovery CO2/N2 gas separation membrane processes for post-combustion capture.
by permeation. Ind. Eng. Chem. 57, 49. J. Memb. Sci. 325, 284e294.
Sublet, J., Pera-Titus, M., Guilhaume, N., Farrusseng, D., Schrive, L., Chanaud, P., Zhao, Y.A., Ho, W.S.W., 2013. CO2-selective membranes containing sterically hin-
Siret, B., Durecu, S., 2012. Technico-economical assessment of MFI-type zeolite dered amines for CO2/H2 separation. Ind. Eng. Chem. Res. 52, 8774e8782.
membranes for CO2 capture from postcombustion flue gases. AIChE J. 58, Zhao, Y.N., Ho, W.S.W., 2012. Steric hindrance effect on amine demonstrated in solid
3183e3194. polymer membranes for CO2 transport. J. Memb. Sci. 415, 132e138.

Khalilpour 2015

Uploaded by

Copyright:

Available Formats

Khalilpour 2015

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Khalilpour 2015

Uploaded by

Copyright:

Available Formats

Journal of Cleaner Production xxx (2014) 1e15

Contents lists available at ScienceDirect

Journal of Cleaner Production

Membrane-based carbon capture from ﬂue gas: a review

Table 1 hybrid system combining membranes with an absorption process is

2.2. Facilitated transport membranes

Facilitated transport membranes, in addition to exhibiting “so-

2.4. Gas membrane contactors

Gas membrane contactors make use of membranes to provide

Traditional membrane material

Membrane material Reference Gases separated Solvent(s) trialed Comments

Fig. 6. Schematic of cascade membrane system with recycle.

Fig. 7. Schematic of a) CMC and b) TSSP.

with permeate recycle. In the second step, the CO2-enriched air is

3.3. Discussion on membrane systems

A single-stage membrane is not a feasible solution for MPCC due

You might also like