Subscriber access provided by University of Newcastle, Australia

Article
Membrane gas separation processes for carbon capture and use (CCU) from
wet post combustion flue gases: physical vs chemically reactive membranes
Marc Pfister, Bouchra Belaissaoui, and Eric Favre
Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.6b03969 • Publication Date (Web): 19 Dec 2016
Downloaded from http://pubs.acs.org on December 19, 2016

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7 Membrane gas separation processes from wet
8
9 post combustion flue gases for carbon capture
10
11
12
and use (CCU): a critical reassessment
13
14
15
16
17
18
19
20
21 Marc PFISTER, Bouchra BELAISSAOUI, Eric FAVRE*
22
23
24
25
26
27 Laboratoire Reactions & Genie des Procedes (LRGP), (UMR 7274) ENSIC,
28
29 Université de Lorraine, 1 rue Grandville-BP 20451. 54001 Nancy FRANCE
30
31
32
33
34
35
36 *: Corresponding author eric.favre@univ-lorraine.fr
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 2 of 30

1
2
3 Abstract:
4
5
6
7 Membrane processes have been shown to offer interesting separation performances for post
8 combustion carbon capture and storage (CCS) from coal power plant flue gases. A 90% CO2 purity and
9 recovery is typically required for that application. This set of constraints does not apply within a
10 Carbon Capture and Use (CCU) scenario, because the objective is to ensure carbon dioxide
11 transformation at minimal cost. In this study a dense polymeric glassy membrane (Matrimid) and a
12
chemically reacting Fixed Site Carrier Membrane (FSCM) are investigated in order to achieve a partial
13
14 carbon dioxide capture and concentration, which could be of interest in order to intensify the carbon
15 dioxide transformation step. The importance of humidity in the carbon capture performances with
16 solution-diffusion membranes is highlighted: flue gases compression significantly changes the inlet
17 water content of the membrane module, water decreases the membrane separation performances,
18 increases the specific energy requirement and decreases the membrane surface area due to an
19
internal sweep effect. Based on a case study, the performances of a FSCM are shown to be close to
20
21 the solution-diffusion membrane performances in terms of surface area and energy requirement.
22 The FSCM offers however two major advantages compared to the solution-diffusion membrane:
23 compatibility to wet flue gases and high CO2 purity. A FSCM single stage module, operated with a
24 50% CO2 capture ratio, is suggested to offer the best trade-off for CCU applications with a high
25 carbon dioxide purity (95%) and a low specific surface area and energy requirement.
26
27
28
29
30
31
32
33 Keywords:
34
Membrane, process, simulation, water, carbon capture, energy
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 3 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5 1. INTRODUCTION
6
7
8
Post-combustion carbon capture from large emitting sources (such as coal power plants), in order to
9
10 achieve a significant mitigation of greenhouse gases (GHG) emissions, has received considerable
11 attention for more than 20 years. The development of a capture process with a low energy
12 requirement (typically less than 2 GJ, thermal basis, per ton of recovered CO2) is considered as a
13 priority for Carbon Capture and Storage (CCS) to be economically acceptable. Consequently, the
14 search for an energy and cost efficient carbon capture process has been abundantly discussed and a
15
great number of separation concepts and process designs have been investigated through simulation
16
17 studies, laboratory tests and, in some cases, pilot units. Basically, a CO2 purity (y) and carbon dioxide
18 recovery ratio (R) larger than 90% are classically taken as process specifications of the capture step
19 for CCS projects.
20
21 Besides CCS, Carbon Capture and Use (CCU) has been more recently considered and receives major
22 attention these last years. The overall framework of CCU is however very different from CCS:
23
24 i) First, the target carbon dioxide flux to be recovered for CCU is much lower than for CCS.
25 While CCS typically aims a Gt per year GHG reduction on a world basis, with a Mt
26 captured flux per unit and [per year (i.e. a 500 MW power plant), CCU targets are
27 typically in the Mt range on a world basis (typical of the market for carbon based
28 molecules) and the carbon capture flux per unit usually scales down kt per year.
29
30 ii) Because of the previous point, only part of the carbon dioxide flux can be transformed
31 from large emitting sources when CCU is aimed. Consequently, carbon recovery ratio R is
32 not imposed to be larger than 90% and smaller recovery ratios can be explored.
33 iii) Carbon dioxide purity specifications for CCU is not very documented, given the large
34 portfolio of transformation solutions (thermochemical, catalytic, electrochemical,
35 biological).
36
37 iv) Finally, while CCS mostly corresponds to a separation process challenge, CCU requires a
38 hybrid separation / reaction solution to be designed, with minimal overall cost.
39
40
Figure 1 summarizes the CCS and CCU specificities, with an emphasis to the carbon capture unit.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56 Figure 1: CCS vs CCU overall framework.
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 4 of 30

1
2
3 *: CO2 transformation processes include chemical (thermochemical, catalytic), biochemical
4 (fermentation, algae), electrochemical and mineralization.
5
6 CO2 absorption in chemical solvents is often presented as the most relevant technology in order to
7 achieve post combustion carbon capture. Membrane separation processes have also more recently
8 attracted attention and could offer interesting potentialities compared to the gas-liquid absorption
9 baseline technology. It should be stressed that, given the purity and recovery constraints for CCS,
10
multistage membrane processes or hybrid processes (e.g. membrane + cryogeny) are needed in
11
12 order to attain the specifications. Despite a significant number of studies, several questions remain
13 however open for a better understanding of membrane processes relevancy for carbon capture
14 purposes:
15
16 i) A large majority of studies focuses on dry CO2/N2 mixtures, with high purity (>90%) and
17 recovery targets 1,2. A very limited number of investigations addressed the impact of
18 water (i.e. humidity), which is a non negligible component of flue gases 3–5.
19 ii) In terms of membrane materials, a physical separation process through dense polymers
20
is most often investigated (solution-diffusion mechanism), but chemically reacting
21
22 membranes have also been proposed in order to improve selectivity performances 6–8.
23 This situation is similar to the gas-liquid absorption process, where chemical solvents are
24 favoured when a larger selectivity is needed. Few process comparisons between physical
25 and chemical separation membranes have been however reported 9 10.
26 iii) Chemically reactive membranes require water (humidity) in the gas mixture for the
27
reaction in the membrane material to be effective 11–13. Water content in flue gas being
28
29 impacted by compression (partial condensation effect), it is necessary to combine
30 compression and separation units in order to carefully evaluate the humidity content in
31 the membrane unit inlet flowrate. The interplay between compression conditions and
32 the associated feed water content of a FSCM carbon capture unit is largely unexplored.
33
34 Based on the state of the art detailed above, this study intends to provide a critical analysis of the
35 potentialities and limitations of a membrane separation / concentration step within a post
36 combustion carbon capture and use (CCU) framework. First, a multicomponent flue gas mixture,
37
including water, will be taken into account in order to compare results to the dry mixture
38
39 approximation. The impact of a compression step on water content will be assessed through a
40 rigorous Process Systems Engineering analysis (including equation of state computations). The
41 separation process analysis is achieved on two different membrane materials: a glassy material for
42 physically based separations (i.e. solution diffusion process) and a Fixed Site Carrier Membrane
43 (FSCM) for chemically based membrane processes. The range of carbon dioxide purity and recovery
44
values which is attainable for different set of operating conditions is determined for the two different
45
46 membrane materials and the wet vs dry feed mixture performances are compared. Special attention
47 is focused on the key impact of the driving force choice, namely feed compression vs vacuum
48 pumping. Finally, the specific energy requirement and associated membrane surface area are
49 evaluated for a series of membrane, feed mixture and operating conditions.
50
51 More generally, it is expected that this study provides key data for the best place and role of a
52 physical or chemically reactive membrane unit to be identified in a CCU scenario. Membrane
53 processes have been effectively shown to offer attracting energy efficiency potentialities for
54
moderate carbon dioxide purity and/or recovery targets. Given the completely different context of
55
56 CCU compared to CCS, it is likely that a moderate purity and/ or a limited capture ratio fit the
57 requirement of a hybrid separation / chemical transformation unit. In that event, the specific energy
58 requirement of the capture/concentration step is needed for operating expenses (OPEX) evaluation;
59
60
ACS Paragon Plus Environment
Page 5 of 30 Industrial & Engineering Chemistry Research

1
2
3 similarly, compression conditions associated to membrane surface area enable capital expenses
4 (CAPEX) to be assessed. This information set is a necessary pre-requisite for technico-economical
5 studies of CCU scenarios, including a concentration and transformation step, to be performed (Figure
6 2).
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 Figure 2: Overall framework of the study: A wet or dry flue gas mixture is compressed and separated
22
23 in a single stage membrane unit. For each set of operating conditions, the separation performances
24 of the membrane unit (carbon dioxide purity y and recovery R) are computed through PSE (Aspen
25 software) and the corresponding specific energy requirement (E) and membrane surface area (A) are
26 calculated.
27
28 *: Fixed Site Carrier Membrane (chemically reactive membrane)
29
30
31
32 2. BACKGROUND: MEMBRANE MATERIALS SELECTION AND PROCESS SIMULATION
33 METHODOLOGY
34
35 2.1. Membrane materials selection for simulation purposes
36
37 One first objective of this study is to explore the impact of water on the carbon capture
38 performances of a membrane unit. Generally speaking, water shows a very large permeability in
39
40 dense polymeric membranes and is usually considered as the fastest component compared to gases
41 or organic vapours. A series of water permeability data into different dense polymers is reported in
42 Table 1.
43
44 More importantly, water is known to affect membrane permeability and selectivity of gaseous
45 components in different ways. Water impact interpretations are not systemic and difficult to predict
46 because water interaction with the different species present in the membrane material (polymeric
47 chains, dissolved water and gases) is complex. Consequently, experimental data with wet gas
48 mixtures are necessary for simulation purposes. For flue gas separation, several studies specifically
49
50 investigated the role of humidity when a rubbery polymer membrane is used 14,15. In that case, water
51 molecules can swell the rubbery matrix and increase the free volume. In consequence, an increased
52 diffusivity of the other components can be observed. The situation is very different for glassy
53 polymers 16–22. In that case, water permeation can cause competitive sorption in the Langmuir voids,
54 plasticization and cluster formation. For instance, Lasseuguette 23 observed a loss of CO2 permeability
55 under humid conditions in a hydrophobic PIM-1 membrane. The CO2 mass transfer was supposed to
56
57 be hindered by cluster formation inside the free volumes of the membrane. With hydrophilic
58 membranes however, humidity tends to increase components permeability. Liu 24 compared CO2
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 6 of 30

1
2
3 permeability in hydrophilic PVA membrane under dry and humid conditions. The results showed that
4 CO2 permeability in a PVA membrane is much higher with humidity compared to dry conditions. The
5 explanation is expected to come from co-solubility effects. CO2 permeability is speeded up because
6 CO2 solubility in water is higher than in the PVA membrane material.
7
8 Figure 3 shows that CO2 and N2 permeability in different membranes are not identical under dry and
9 humid conditions. As discussed before, it can be seen that humidity is responsible of a CO2
10
permeability increase or decrease in dense, non-reacting, polymers (solution diffusion mechanism).
11
12 Separation performances logically never exceed the Robeson upper bound limit. FTMs are able
13 however to overtake that limit due to the additional selective flux provided by the chemical reaction.
14
15 In order to achieve a rigorous comparison of dry vs wet flue gas separation and physical vs chemical
16 gas separation membranes, representative permeability data of the different components of post
17 combustion flue gases are absolutely necessary. Based on a detailed permeance data review, two
18 different membrane materials have been selected for the simulations: a glassy membrane material,
19 classically used for gas separation applications (Matrimid, a polyimide type material) 16,18,25 and a
20
Fixed Site Carrier Membrane (FSCM), BAM-1 (PolyVinylAlcohol with a blend of amine) 26. The Bam-1
21
22 composition is the following: 41.66% wt PVA, 8.33% wt PVP, 10% wt KOH, 25% wt PAA and 15% wt
23 AHPD. The experimental effective permeabilities of the two membranes as well as the experimental
24 conditions (T, RH, P) and membrane thicknesses are summarized in Table 2. Two set of permeability
25 are offered with the matrimid. The first set corresponds to the effective permeabilities in a dry
26 system. The second set corresponds to effective permeabilities with a saturated feed flow. Because
27
these data result from experiments performed with real gas mixtures (dry and wet flue gases), it is
28
29 expected that the process simulation will be representative of the incidence of water on membrane
30 module separation performances.
31
32 It is important to stress that, for the FSCM simulation, a large relative humidity is required for the
33 facilitated transport reaction into the polymer to operate. Deng 11 observed that best separation
34 performance are reached when relative humidity is larger than 90%. The facilitated transport effect
35 was reported to be efficient only under these high relative humidity conditions. It is thus of major
36 importance to carefully evaluate the flue gas humidity content into the gas separation module.
37
Moreover, the impact of flue gas compression on water content has to be rigorously evaluated so
38
39 that the right inlet gas composition is used for module simulation purposes. This point is detailed
40 hereafter.
41
42
43
44
2.2. Impact of flue gas compression on membrane unit inlet water concentration
45 A dry gas mixture is most often postulated for carbon capture studies with post combustion flue
46
47
gases. Flue gases are however saturated in water vapour under real conditions.
48 Water content in a gas mixture is usually expressed through relative humidity (
.  ).
.  is linked to
49
the total pressure ( ), the water molar fraction ( 
 ) and the saturated pressure (  ) which
50
51 depends on the operating temperature only (Equation 4).
52  ∗
53
.  = 
(1)
54
55 During gas compression, compressor work induces an increase of total pressure and a temperature
56 rise 27. An increase of total feed pressure with saturated flue gas at constant temperature involves a
57 partial water condensation until water vapour partial pressure becomes equal to the saturated
58
59
60
ACS Paragon Plus Environment
Page 7 of 30 Industrial & Engineering Chemistry Research

1
2
3 pressure. In practice however, no condensation takes place at the outlet stream of the compressor
4 due to the increase of the temperature. Nevertheless, a heat cooler exchanger, before the
5 membrane module is necessary in order to cool the temperature down to the operating temperature
6 compatible to membrane use (typically 20 to 60 C). At this cooling step, water condensation is
7
8
observed.
9 In summary, two variables control water condensation: membrane module operating temperature
10
and compression rate. The interplay of these two key variables is usually not taken into account. A
11
12 rigorous computation, through Process System Engineering simulation, including thermodynamic
13 properties data (Peng Robinson equation of state) has been performed. Results are shown in Figure
14 4. Compression involves a decrease of water vapour fraction in the feed flow. An increase of the
15 temperature entails an increase of the saturated pressure. It can be seen that the compression /
16 cooler pack significantly impacts the flue gas water content. This has two major effects on the
17
membrane module inlet conditions: a decreased total gas flowrate (water condensation) and a
18
19 decreased water content (humidity change). This preliminary study shows the importance to
20 systematically include the compressor and heat exchanger cooler units when flue gas post-
21 combustion carbon capture is investigated. This point is of particular significance for membrane
22 carbon capture studies because the choice of the driving force for the membrane separation results
23 in different compression conditions. As a consequence, each set of feed compression and membrane
24
25
module operating temperature will correspond to a specific feed mixture composition. This strategy
26 has been applied afterwards.
27
28
29
2.3. Membrane transport simulation
30
31 Gas separation membranes are most often based on a dense polymeric active layer which has
32 selective properties. The local mass transfer in such a gas permeation membrane does not involve a
33
34 chemical reaction (physical separation process) and is based on a so called solution-diffusion
35 mechanism. Component molecules are dissolved in the high pressure side of the membrane, cross
36 through the membrane by diffusion and desorb on the permeate side. The driving force of this
37 separation process is a differential of partial pressure (∆ ) between the both sides of a membrane
38 with a thickness () 28.  the flux per unit area of the component i can be computed through
39
(Equation 1):
40
41  =  ∆ / (2)
42
43 Permeability () defines the membrane capacity to be crossed by one component and can be
44 described as the product of a kinetic factor (diffusion coefficient  in m²/s) and a thermodynamic
45 factor (solubility coefficient  in cm3 (STP).cm-3.cmHg-1) (Equation 2). The selectivity ( ) for a binary
46
47 stream is the ratio between the fastest permeability over the slowest permeability (Equation 3). The
48 key material performance are gas permeability ( in cm3 (STP) cm. cm-2.s-1.cmHg-1)) and selectivity for
49 a gas pair ( ).
50
51  =  (3)
52
53  =  / (4)
54
Water permeability is often high in polymeric membranes (Table 1). It generally means that sorption
55
56 coefficient and diffusion coefficient are in favour of water compared to the others components.
57 Carbon dioxide permeability is always larger than nitrogen into dense polymers. As a consequence,
58 membrane gas separation of flue gases produces a wet CO2 enriched stream on the permeate side.
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 8 of 30

1
2
3 The evolution of the gas composition in a membrane module for a multicomponent gas mixture can
4 be calculated from the permeability of the different species and membrane thickness data. Matrimid
5 permeability values for dry and wet mixtures is detailed in Table 2.
6
7 A different mass transfer mechanism takes place into facilitated transport membranes such as FSCM.
8 In that case, a reversible chemical reaction occurs within the membrane material and enables to
9 enhance CO2 permeability only, providing that relative humidity is high enough. In the case of carbon
10
capture, site carriers containing amine group are used and can be fixed or mobile. Mobile carrier in
11
12 liquid membranes have been intensively investigated during the 90’s 29–32. Despite their impressive
13 performances, liquid membranes have unfortunately shown instability issues. FSCM, which overtake
14 mobile carrier membrane limits, have been developed more recently in order to overcome this
15 limitation. In presence of water, CO2 molecules are hydrolyzed on an amine carrier (AmH) and a free
16 bicarbonate anion is formed. The chemical reaction is the following one:
17
18 %&'(
19 !"# + # " )***+ !",- + %
20 Only CO2 molecules take the benefit of facilitated transport. Thereby CO2 permeability and selectivity
21
increase. The global chemical reaction can be described in two steps:
22 011
23 !"# + ./ )*+ ./ % !"" -
24
0112
25 ./ % !""- + # " )**+ ./#% + !",-
26
27 The amine carrier (./) role is similar to a catalyst. The chemical reaction has to be reversible in
28 order to release CO2 on the permeate side. Steric hindrance effects help the hydration reaction
29 because the intermediate ./% !""- is instable inside the membrane. The stoichiometry of the
30
31 hydration reaction is one CO2 molecule and one amine carrier. This characteristic shows a real
32 interest compared to zwitterion mechanism where the stoichiometry is one CO2 molecule and two
33 amine carriers. The mass transfer performances of FSCM can be expressed through an effective
34 permeability concept. Nevertheless, it should be stressed that a given experimental effective
35 permeability will depend on humidity and membrane thickness. Thus the FSCM simulations detailed
36 afterwards, based on experimental data, only apply for the membrane thickness and humidity
37
38 indicated in Table 2.
39
40
41 2.4. Membrane process simulations
42
43 Membrane module separation performances can be simulated as soon as a mass transfer expression
44 is used, in combination to a series of assumptions. A cross flow model (i.e. plug flow on the retentate
45
side and perfectly mixed conditions on the permeate side) is most often proposed for that purpose. A
46
47 set of assumptions, classically proposed for membrane module simulations, has been used:
48
- Steady state
49
50 - Isothermal conditions,
51 - Constant permeability (Table 2)
52 - Negligible pressure drop along the surface
53
54 The membrane module program is a tailor made computer code 33. Gas permeation module is
55 simulated thanks to this computer code, which is inserted into Aspen Plus software. Additional
56 operations such as compressor, heat exchanger, expander can be used from the Aspen Plus library.
57
58
59
60
ACS Paragon Plus Environment
Page 9 of 30 Industrial & Engineering Chemistry Research

1
2
3 Additionally, thermodynamic properties can be calculated thanks to in use packages. In this study,
4 the Peng-Robinson equation was used as thermodynamic model.
5
6 A standard coal power plant post combustion flue gas composition has been considered: 12% CO2,
7 73% N2, 10% H2O and 5% O2. NOx and SOx are supposed to be removed from flue gas in a previous
8 step. The corresponding composition of the dry gas (after water removal through a drying step) is
9 13.3% CO2, 81.1% N2 and 5.6% O2. The overall flue gas feed flow rate is fixed at 5 ton/h, typical of a
10
post combustion coal power plant. This flow rate value logically will translate into large membrane
11
12 surface area requirement. The simulation conditions are detailed in the Table 3 and 4.
13
After feed conditions have been fixed, membrane module simulation logically requires the driving
14
15 force to be defined. Three different possibilities, shown in Figure 5, can be proposed: feed
16 compression, permeate vacuum or a combination of compression and vacuum. The key separation
17 performances which are obtained are the CO2 purity on the permeate side (y) and the carbon dioxide
18 recovery ratio (R). Additionally, membrane surface area and specific energy requirement can be
19 obtained (Figure 2).
20
21 It is important to note that a single stage membrane process can by no means reach the capture ratio
22 and CO2 purity (RCO2>90% and yCO2>90%) required for CCS 1,10,33,34. Multistaged membrane designs,
23
including cryogenic condensation are needed and have been investigated through parametric studies
24 9
25 , [16]-[21]. The objective of the simulations described hereafter is to evaluate the interest of
26 solution diffusion or FSCMs to achieve a carbon dioxide partial recovery and concentration from a
27 real (wet) flue gas within a CCU perspective. Thus, no target on purity and capture ratio for CO2 has
28 been fixed.
29
30 A thin active layer involves a porous support in order to withstand the transmembrane pressure
31 difference. Generally, the membrane support is chosen porous or highly permeable and its
32 corresponding mass transfer resistance may be negligible. But the support nature may be more or
33
less hydrophobic and influence the water transport. However, water permeability stays high even in
34
35 a dense hydrophobic support in the most cases. Indeed water diffusion coefficient is higher than
36 other components diffusion coefficient.
37
38 In a first step, a comparison of dry vs wet flue gas will be performed for a solution-diffusion
39 polymeric membrane (Matrimid). The important effect of water when different compression
40 strategies are applied, for similar pressure ratio values, will be shown. Real behaviour with mixed gas
41 properties and the decrease of components’ permeability for humid condition (extracted from
42 literature data) enable a detailed analysis of the influence of water on permeation properties to be
43
achieved. The interplay between CO2 purity, CO2 recovery, energy and surface requirement can be
44
45 discussed.
46
In a second step, a comparison between the solution diffusion membrane performances and the
47
48 FSCM performances under similar operating conditions (wet flue gas only) will be presented. The
49 benefit of the facilitated transport is made explicit in terms of CO2 purity, CO2 capture ratio, surface
50 area and energy requirement.
51
52
53
54 3. RESULTS AND DISCUSSION
55
56 3.1 Polymeric membrane (Matrimid): Dry vs humid flue gas
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 10 of 30

1
2
3 A large majority of membrane carbon capture studies makes use of a dry binary CO2/N2 mixture
4 assumption for sake of simplicity. The trade-off between purity and capture ratio (y vs R) is then a
5 key issue and it has been largely discussed, especially for the single stage membrane unit 35. More
6 specifically, a single purity / recovery performance data set is obtained for a dry mixture when
7
8
membrane selectivity and pressure ratio (i.e. downstream pressure over upstream pressure) are
9 fixed. This peculiarity is of interest in order to achieve generic parametric sensitivity studies, through
10 adimensional numbers. Nevertheless, humidity in the flue gas feed complicates the situation and the
11 analysis.
12
13 The comparison to results obtained with a humid flue gas feed and dry flue gas feed is shown on
14 Figure 6a and 6b. The major influence of humidity is noticeable (Figure 6a). First, inlet flow
15 composition depends on the feed pressure, as explained in the previous section 9. For example, a 10
16 Bar flue gas compression leads to an important water removal effect through condensation, with a
17
2% water vapour fraction in the inlet stream. The different feed pressure conditions involve that
18
19 water vapour molar fraction in the inlet stream varies between 2% and 10%. 10 Bar feed
20 compression with flue gas saturated in water vapour logically shows separation performance close to
21 the performance obtained with dry flue gas (solid line) because a large part of water vapour has
22 condensed during the compression step. Correspondingly, the vacuum strategy leads to the lowest
23 CO2 purity in the permeate stream, due to the important water content and water dilution effect
24
25
exerted on the permeate side. Indeed the high water permeability involves that water is the main
26 component of the permeate flow when the carbon capture ratio stays small. This effect translates
27 into an increasing CO2 purity when low capture ratio and vacuum are combined. Figure 6a clearly
28 shows that the dilution effect is determined by the compression strategy chosen.
29
30 Figure 6b presents the same simulation results than in Figure 6a but on dry basis. Water dilution
31 effect in the permeate stream tends to increase the gradient of partial pressure which is the driving
32 force of the system. In consequence a higher CO2 purity is reached than with the dry flue gas feed at
33 same CO2 capture ratio (R) whereas the CO2/N2 selectivity with a dry flue gas is superior to the
34
selectivity with a humid flue gas (Table 2). However the influence of the water dilution is observable
35
36 with a low CO2 capture ratio (R).
37
Thus, the necessity to systematically include compression and membrane separation units for (wet)
38
39 flue gas treatment is highlighted. A permeate vacuum is interesting in that it has the maximal dilution
40 effect. But the decrease of the component’s permeability by the water interaction are not benefit in
41 terms of separation performance. The compression feed limits the impact of humidity but also the
42 associated dilution effect. This options have also an incidence on energy and surface requirements.
43 This point will be detailed hereafter.
44
45
46
47 3.2 Polymeric membrane (Matrimid): analysis of the mechanisms of water influence
48
49
50 The strong effect of humidity on carbon capture performances call for a detailed analysis and
51 interpretation of the phenomenon. Based on the rigorous modelling of combined water, CO2 and N2
52
previously reported for Matrimid membrane [23], the impact of water can be analysed in detail.
53
54 An estimation of the surface area for different situations with a Matrimid membrane is shown in
55
Figure 7. Vacuum pumping only has been taken because the sweep effect and water interaction are
56
57 the strongest (Figure 6a). Humid gas is generally speaking responsible of two opposite phenomena
58 which impact surface area estimation. On one hand, water is responsible for a loss of CO2
59
60
ACS Paragon Plus Environment
Page 11 of 30 Industrial & Engineering Chemistry Research

1
2
3 permeability in Matrimid membrane due to a competitive sorption and free volume reduction effect
16
4 . This phenomenon logically leads to an increased membrane surface area compared to the dry
5 mixture case. On the other hand, water permeation also generates an internal sweep effect which is
6 beneficial; the dilution effect exerted by water on the permeate side increases the CO2 driving force.
7
8
The consequence is a decrease of the surface area for the same CO2 capture ratio.
9 These two opposite effects have been separately evaluated on Figure 7 Simulation results in terms of
10
surface area estimation show that the difference between a dry and a humid gas is small because the
11
12 dilution effect counterbalances water interaction effect. For a high CO2 capture ratio however, CO2
13 concentration gradient decreases whereas the permeability loss with humidity is independent of the
14 CO2 capture ratio. The water negative effects due to competitive sorption and free volume reduction
15 tend to exceed the impact of the water sweep effect. As result, the surface area with humid gas is
16 above the dry gas case.
17
18 This water influence analysis suggests to use membranes for which sorption competition effects do
19 not exist. Rubbery membranes can be expected to offer this characteristic. There is no sorption site
20
limitation in a rubbery matrix, which is subject to swelling effects when gaseous compounds dissolve
21
22 in the polymer. In that case, the only beneficial internal sweep effect of water will occur. It is
23 interesting to notice that rubbery membranes have been mostly proposed for carbon capture
24 applications, even though their interest because of non competitive sorption, such as suggested
25 above, have not been pointed out up to now.
26
27
28
29 3.3 Polymeric membrane (Matrimid): Surface and energy requirement
30
31 Besides carbon dioxide purity and recovery, membrane surface area and specific energy requirement
32 are of major importance in order to evaluate the potentialities of membrane processes for CCS and
33 CCU (Figure 2).
34
35 The membrane surface area and the associated energy requirement of a Matrimid membrane
36 module with dry and wet inlet flue gas is shown on Figures 8a and 8b. Because compression rate
37 impacts the water vapour fraction in the inlet flow rate, it can be expected that the difference
38 between the dry and wet case are not always equal. When the inlet water fraction decreases (i.e.
39
40
larger p’ value), water influence logically lessens too. Figure 8.a shows the simulation results for the
41 different cases of compression corresponding to a pressure ratio of 0.1. The difference of surface
42 area estimation between a dry gas and a humid one is low even if surface area estimation with
43 humid flue gas tends to stay superior. The smallest surface area is clearly and logically obtained with
44 compression.
45
46 The energy requirement corresponding to the different cases discussed in Figure 8.a are shown on
47 Figure 8.b. The net energy requirement including compressor, expander and/or vacuum pump is
48 indicated as a function of the capture ratio. It can be seen that the trend is right the opposite to the
49
membrane surface area. The minimal energy requirement (i.e. corresponding to the maximal energy
50
51 efficiency) is obtained for large capture ratios. The trade-off between membrane surface area
52 (leading to CAPEX) and energy requirement (OPEX) is self speaking. The best combination requires a
53 technico-economical analysis to be carried out (Figure 2). A wet inlet flow generates a larger molar
54 flowrate than the dry gas case. The difference of flow rate is responsible of the difference of the
55 energy requirement between a humid flue gas and dry flue gas. That difference in energy
56
requirement is maximal for the vacuum pump only case because there is no water condensation
57
58 before the membrane module. And finally driving force increase by the water dilution effect is too
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 12 of 30

1
2
3 limited on the energy requirement and has not real positive effect on the energy requirement. The
4 energy requirement with humidity is always over 2GJ/ton CO2 (the CCS energy requirement target
5 taken here for comparison purpose).
6
7 This simulation study shows the limitation of a single stage gas permeation process in terms of CO2
8 purity and recovery. More generally, a high purity and high capture ratio cannot be reached with a
9 single stage fed by a dry mixture when CO2 composition in the feed flow is under 20% 28 36. The
10
strong, mostly negative impact of water on separation performances has been shown here. In order
11
12 to reach high purity and high capture ratio, a membrane with better permeability and selectivity is
13 required. The possibility to achieve separation under wet conditions, without needing a drying step
14 would also be of interest. FSCM combine these two advantages and are compared to the solution
15 diffusion case in the next section.
16
17
18
19 3.4 FSC membrane: the benefit of the facilitated transport
20
21 With FSCM, humid gas conditions are always taken for simulations, because the chemical reaction is
22 only effective under high relative humidity conditions. Thus, there is no need to remove water before
23 membrane module. The combined use of compressor and vacuum pump is often suggested for an
24 optimal driving force with FSCM. High feed pressure indeed induces amine site saturation whereas
25
high vacuum tends to dry the membrane in the permeate side. Both situations are not desired
26
27 because FSCM performance will largely decrease.
28 Facilitated transport is chemically specific to CO2 molecules. CO2 permeability and CO2/N2 selectivity
29
30 expected with a FSCM are high. FSCM has the capacity to overtake the Robeson plot limit. Mondal 26
31 observed under specific operating condition that CO2 permeability was 1850 Barrer with a
32 permselectivity around 430 with the BAM-1 membrane. Simulation results show that at same
33 operating pressure and same CO2 capture ratio (RCO2=80%), CO2 purity with a Matrimid membrane
34 cannot exceed 50% whereas a CO2 purity close to 78% can be reached with FSCM as shown in Figure
35 9. Under these operating conditions, FSCM enables a 95% CO2 purity together with a CO2 capture
36
37 ratio of 50%. The facilitated transport not only improves the CO2 permeability but also the selectivity
38 of the membrane (Figure 3). In order to better settle the interest of FSCM, a simulation has been
39 performed in order to identify a virtual physical (solution diffusion) membrane which would show
40 similar separation performances under the constraints of Robeson trade-off 37. Taking the CO2 / N2
41 upper bound proposed for dense polymeric membranes, a fictional membrane with a selectivity
42 around 1000 is necessary. According to Robeson trade-off curve, this level of selectivity corresponds
43
44 to a very low CO2 permeability, around 0.1 Barrer. This hypothetic membrane could obviously not be
45 used for practical applications because of a too low permeability (a huge surface area would be
46 needed).
47
48 On the one hand, the analysis about the surface area estimation in Figure 10.a shows the high
49 potential of the FSCM but also its limits. The consequence of high CO2 permeability and selectivity
50 are an important decrease of the surface area requirement compared to a Matrimid membrane for a
51 CO2 capture ratio below 90%. The membrane surface area of FSCM is however larger than the
52 Matrimid membrane for a larger capture ratio. It is important to note that the large FSCM thickness
53
54 (45µm) compared to Matrimid (1 µm) does not result into large membrane surface area differences.
55 The facilitated transport effectively improves the membrane mass transfer performances enough so
56 as to compensate the rather large thickness. At the difference to solution-diffusion membranes, an
57 extrapolation for a thinner FSCM is unfortunately not possible. For a FSCM indeed, each membrane
58
59
60
ACS Paragon Plus Environment
Page 13 of 30 Industrial & Engineering Chemistry Research

1
2
3 thickness value corresponds to a given reaction time and the separation performances cannot be
4 simply related to the membrane thickness 13. Moreover, for each FSCM system, there is a minimal
5 optimal thickness [26]. If the active layer is too small, equilibrium of chemical reaction inside the
6 membrane cannot be established. The facilitated transport is limited in that case and separation
7
8
performance decreases. A rigorous modelling and parametric study of the optimal FSCM thickness in
9 order to minimize the carbon capture cost, which is beyond the scope of this study, would obviously
10 be of interest.
11
12 The energy requirement of Matrimid and FSCMs are compared on Figure 10.b. The difference
13 between the two membranes is globally low. The pressure operating conditions are similar and
14 compressor energy requirement represents the major contribution of the system. For low CO2
15 capture ratio, the energy for compression is dominant (low permeate flow generating a low vaccum
16 energy requirement). For a larger CO2 capture ratio, the difference in permeate and retentate flow
17
rate for the two membranes stays close. In summary, the benefit of FSCM over Matrimid in terms of
18
19 energy requirement for similar carbon capture ratio is limited. Nevertheless, the much larger CO2
20 purity obtained by a FSCM under similar operating conditions (Figure 9) remains valid and shows the
21 key interest of this type of material.
22
23
24
25
26
27 4. CONCLUSION
28 This study intended to provide a detailed comparative analysis of a solution-diffusion vs a chemically
29
30 reactive membrane carbon capture process within a perspective of CCU from coal power plant post
31 combustion wet flue gases. Based on a series of previously reported experimental data obtained on
32 wet gas mixtures with a glassy dense polymeric membrane (Matrimid) and a Fixed Site Carrier
33 Membrane (FSCM), the possibilities and limitations of these two types of membrane materials have
34 been investigated. The main conclusions of this process simulation study can be summarized as
35 follows:
36
37
38
39 i) The flue gas compression step induces a significant change in water content due to
40 partial condensation effects. It is thus recommended to systematically combine
41 compression, heat exchanger cooler and membrane separation unit for Process System
42
Engineering (PSE) studies.
43
44
45 ii) The impact of humidity on solution-diffusion membrane separation performances is very
46 important. Water systematically decreases the separation performances when compared
47 to dry gas mixtures.
48
49
iii) A humid flue gas requires a larger energy compared to a dry flue gas. This effect simply
50
51 results from the larger flowrate to be treated.
52
53 iv) Humidity surprisingly decreases the required surface area, due to an internal sweep
54 effect (dilution in the permeate improves the CO2 driving force compared to a dry
55 mixture separation). In glassy polymers however, an additional competitive sorption
56
effect of water towards CO2 takes place and globally annihilates the benefits of the
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 14 of 30

1
2
3 internal sweep effect. It is thus recommended to use rubbery polymers in order to take
4 advantage of the internal sweep effect only.
5
6 v) FSCMs offer a very large improvement of CO2 purity compared to solution- diffusion
7
8
membranes and do not require a drying step because humidity is necessary for the
9 reaction to take place, both on the retentate and permeate side.
10
11 vi) Based on a comparative study, the most interesting separation performances in a CCU
12 perspective correspond to a single stage FSCM module with a 50% CO2 capture ratio. In
13 that case a 95% CO2 purity stream can be produced, with low membrane surface area (22
14
15 000 m2.kg CO2.s-1) and energy requirement (2.6 GJ/ton).
16
17
18 The interest of this set of performances for a complete Carbon Capture and Use process remains to
19 be evaluated through a technico-economical analysis. Depending on the reaction used for CO2
20
transformation (thermochemical, catalytic, electrochemical, biotechnological), the potential interest
21
22 of a pre-concentration step for a lower overall cost can be investigated based on the surface area,
23 compressor, vacuum pump and the associated energy requirement reported in this study. This work
24 is currently under progress for a chemical transformation of CO2 to methanol.
25
26
27
28
29
30
31 REFERENCES
32
33 (1) Ho, M. T.; Allinson, G. W.; Wiley, D. E. Reducing the Cost of CO 2 Capture from Flue Gases
34 Using Membrane Technology. Ind. Eng. Chem. Res. 2008, 47 (5), 1562–1568.
35 (2) Zhao, L.; Riensche, E.; Blum, L.; Stolten, D. Multi-Stage Gas Separation Membrane Processes
36
Used in Post-Combustion Capture: Energetic and Economic Analyses. J. Membr. Sci. 2010, 359
37
(1–2), 160–172.
38
39 (3) Merkel, T. C.; Lin, H.; Wei, X.; Baker, R. Power Plant Post-Combustion Carbon Dioxide Capture:
40 An Opportunity for Membranes. J. Membr. Sci. 2010, 359 (1–2), 126–139.
41 (4) Scholes, C. A.; Ho, M. T.; Wiley, D. E.; Stevens, G. W.; Kentish, S. E. Cost Competitive
42 Membrane—cryogenic Post-Combustion Carbon Capture. Int. J. Greenh. Gas Control 2013, 17,
43 341–348.
44 (5) Scholes, C. A.; Ho, M. T.; Aguiar, A. A.; Wiley, D. E.; Stevens, G. W.; Kentish, S. E. Membrane
45 Gas Separation Processes for CO2 Capture from Cement Kiln Flue Gas. Int. J. Greenh. Gas
46 Control 2014, 24, 78–86.
47 (6) He, X.; Hägg, M.-B. Energy Efficient Process for CO2 Capture from Flue Gas with Novel Fixed-
48 Site-Carrier Membranes. Energy Procedia 2014, 63, 174–185.
49
(7) He, X.; Fu, C.; Hägg, M.-B. Membrane System Design and Process Feasibility Analysis for CO2
50
Capture from Flue Gas with a Fixed-Site-Carrier Membrane. Chem. Eng. J. 2015, 268, 1–9.
51
52 (8) Vakharia, V.; Ramasubramanian, K.; Winston Ho, W. S. An Experimental and Modeling Study
53 of CO2-Selective Membranes for IGCC Syngas Purification. J. Membr. Sci. 2015, 488, 56–66.
54 (9) Low, B. T.; Zhao, L.; Merkel, T. C.; Weber, M.; Stolten, D. A Parametric Study of the Impact of
55 Membrane Materials and Process Operating Conditions on Carbon Capture from Humidified
56 Flue Gas. J. Membr. Sci. 2013, 431, 139–155.
57 (10) Gottschlich, D. . A Theoretical Comparison of Facilitated Transport and Solution-Diffusion
58 Membrane Modules for Gas Separation. Chem. Eng. Lab. 1988.
59
60
ACS Paragon Plus Environment
Page 15 of 30 Industrial & Engineering Chemistry Research

1
2
3 (11) Deng, L.; Hägg, M.-B. Swelling Behavior and Gas Permeation Performance of PVAm/PVA Blend
4 FSC Membrane. J. Membr. Sci. 2010, 363 (1–2), 295–301.
5 (12) Sandru, M.; Haukebø, S. H.; Hägg, M.-B. Composite Hollow Fiber Membranes for CO2 Capture.
6 J. Membr. Sci. 2010, 346 (1), 172–186.
7 (13) Ansaloni, L.; Zhao, Y.; Jung, B. T.; Ramasubramanian, K.; Baschetti, M. G.; Ho, W. S. W.
8 Facilitated Transport Membranes Containing Amino-Functionalized Multi-Walled Carbon
9 Nanotubes for High-Pressure CO2 Separations. J. Membr. Sci. 2015, 490, 18–28.
10
(14) Potreck, J.; Nijmeijer, K.; Kosinski, T.; Wessling, M. Mixed Water Vapor/gas Transport through
11
12
the Rubbery Polymer PEBAX® 1074. J. Membr. Sci. 2009, 338 (1–2), 11–16.
13 (15) Jan Roman, P.; Detlev, F.; Thomas, K.; Klaus-Viktor, P. Gas Permeation Measurement under
14 Defined Humidity via Constant Volume/variable Pressure Method. J. Membr. Sci. 2012, 389,
15 343–348.
16 (16) Chen, G. Q.; Scholes, C. A.; Doherty, C. M.; Hill, A. J.; Qiao, G. G.; Kentish, S. E. The Thickness
17 Dependence of Matrimid Films in Water Vapor Permeation. Chem. Eng. J. 2012, 209, 301–312.
18 (17) Chen, G. Q.; Scholes, C. A.; Doherty, C. M.; Hill, A. J.; Qiao, G. G.; Kentish, S. E. Modeling of the
19 Sorption and Transport Properties of Water Vapor in Polyimide Membranes. J. Membr. Sci.
20 2012, 409–410, 96–104.
21 (18) Chen, G. Q.; Scholes, C. A.; Qiao, G. G.; Kentish, S. E. Water Vapor Permeation in Polyimide
22 Membranes. J. Membr. Sci. 2011, 379 (1–2), 479–487.
23
(19) Chen, G. Q.; Kanehashi, S.; Doherty, C. M.; Hill, A. J.; Kentish, S. E. Water Vapor Permeation
24
through Cellulose Acetate Membranes and Its Impact upon Membrane Separation
25
26 Performance for Natural Gas Purification. J. Membr. Sci. 2015, 487, 249–255.
27 (20) Tsvigu, C.; Pavesi, E.; De Angelis, M. G.; Giacinti Baschetti, M. Effect of Relative Humidity and
28 Temperature on the Gas Transport Properties of 6FDA–6FpDA Polyimide: Experimental Study
29 and Modelling. J. Membr. Sci. 2015, 485, 60–68.
30 (21) Scholes, C. A.; Freeman, B. D.; Kentish, S. E. Water Vapor Permeability and Competitive
31 Sorption in Thermally Rearranged (TR) Membranes. J. Membr. Sci. 2014, 470, 132–137.
32 (22) Scholes, C. A.; Kanehashi, S.; Stevens, G. W.; Kentish, S. E. Water Permeability and Competitive
33 Permeation with CO2 and CH4 in Perfluorinated Polymeric Membranes. Sep. Purif. Technol.
34 2015, 147, 203–209.
35 (23) Lasseuguette, E.; Ferrari, M.-C.; Brandani, S. Humidity Impact on the Gas Permeability of PIM-
36 1 Membrane for Post-Combustion Application. Energy Procedia 2014, 63, 194–201.
37
(24) Liu, L.; Chakma, A.; Feng, X. Gas Permeation through Water-Swollen Hydrogel Membranes. J.
38
39 Membr. Sci. 2008, 310 (1–2), 66–75.
40 (25) Moghadam, F.; Omidkhah, M. R.; Vasheghani-Farahani, E.; Pedram, M. Z.; Dorosti, F. The
41 Effect of TiO2 Nanoparticles on Gas Transport Properties of Matrimid5218-Based Mixed
42 Matrix Membranes. Sep. Purif. Technol. 2011, 77 (1), 128–136.
43 (26) Mondal, A.; Barooah, M.; Mandal, B. Effect of Single and Blended Amine Carriers on CO2
44 Separation from CO2/N2 Mixtures Using Crosslinked Thin-Film Poly(vinyl Alcohol) Composite
45 Membrane. Int. J. Greenh. Gas Control 2015, 39, 27–38.
46 (27) Moore, J. CO2 Capture Tech Update 2013 Advanced Compression. U.S. Department of Energy
47 2013.
48 (28) Kentish, S. E.; Scholes, C. A.; Stevens, G. W. Carbon Dioxide Separation through Polymeric
49 Membrane Systems for Flue Gas Applications. Recent Pat. Chem. Eng. 2008, 1 (1), 52–66.
50
(29) Matsuyama, H.; Teramoto, M.; Kiyoshi, I. Development of a New Functional Cation-Exchange
51
52 Membrane and Its Application to Facilitated Transport of CO2. J. Membr. Sci. 1994, 93, 237–
53 244.
54 (30) Teramoto, M.; Nakai, K.; Ohnishi, N.; Huang, Q.; Watari, T.; Matsuyama, H. Facilitated
55 Transport of Carbon Dioxide through Supported Liquid Membranes of Aqueous Amine
56 Solutions. Ind. Eng. Chem. Res. 1996, 35, 538–545.
57 (31) Saha, A. K.; Bandyopadhyay, S. S. KINETICS OF ABSORPTION OF C02 INTO AQUEOUS
58 SOLUTIONS OF 2-AMINO-2-METHYL-1-PROPANOL. Chem. Eng. Sci. 1995, 50 (22), 3587–3598.
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 16 of 30

1
2
3 (32) Littel, R. J.; Versteeg, G. F.; Van Swaaij, W. Kinetics of CO2 with Primary and Secondary Amines
4 in Aqueous solutions—I. Zwitterion Deprotonation Kinetics for DEA and DIPA in Aqueous
5 Blends of Alkanolamines. Chem. Eng. Sci. 1992, 47 (8), 2027–2035.
6 (33) Bounaceur, R.; Lape, N.; Roizard, D.; Vallieres, C.; Favre, E. Membrane Processes for Post-
7 Combustion Carbon Dioxide Capture: A Parametric Study. Energy 2006, 31 (14), 2556–2570.
8 (34) Yuan, M.; Narakornpijit, K.; Haghpanah, R.; Wilcox, J. Consideration of a Nitrogen-Selective
9 Membrane for Postcombustion Carbon Capture through Process Modeling and Optimization.
10
J. Membr. Sci. 2014, 465, 177–184.
11
12
(35) Belaissaoui, B.; Willson, D.; Favre, E. Membrane Gas Separations and Post-Combustion Carbon
13 Dioxide Capture: Parametric Sensitivity and Process Integration Strategies. Chem. Eng. J. 2012,
14 211–212, 122–132.
15 (36) Mondal, M. K.; Balsora, H. K.; Varshney, P. Progress and Trends in CO2 Capture/separation
16 Technologies: A Review. Energy 2012, 46 (1), 431–441.
17 (37) Robeson, L. M. The Upper Bound Revisited. J. Membr. Sci. 2008, 320 (1–2), 390–400.
18 (38) Giacinti Baschetti, M.; Minelli, M.; Catalano, J.; Sarti, G. C. Gas Permeation in
19 Perflurosulfonated Membranes: Influence of Temperature and Relative Humidity. Int. J.
20 Hydrog. Energy 2013, 38 (27), 11973–11982.
21 (39) Scholes, C. A.; Stevens, G. W.; Kentish, S. E. The Effect of Hydrogen Sulfide, Carbon Monoxide
22 and Water on the Performance of a PDMS Membrane in Carbon Dioxide/nitrogen Separation.
23
J. Membr. Sci. 2010, 350 (1–2), 189–199.
24
(40) Scholes, C. A.; Chen, G. Q.; Tao, W. X.; Bacus, J.; Anderson, C.; Stevens, G. W.; Kentish, S. E.
25
26 The Effects of Minor Components on the Gas Separation Performance of Membranes for
27 Carbon Capture. Energy Procedia 2011, 4, 681–687.
28 (41) Cersosimo, M.; Brunetti, A.; Drioli, E.; Fiorino, F.; Dong, G.; Woo, K. T.; Lee, J.; Lee, Y. M.;
29 Barbieri, G. Separation of CO2 from Humidified Ternary Gas Mixtures Using Thermally
30 Rearranged Polymeric Membranes. J. Membr. Sci. 2015, 492, 257–262.
31 (42) Zhao, Y.; Winston Ho, W. S. Steric Hindrance Effect on Amine Demonstrated in Solid Polymer
32 Membranes for CO2 Transport. J. Membr. Sci. 2012, 415–416, 132–138.
33 (43) Francisco, G. J.; Chakma, A.; Feng, X. Membranes Comprising of Alkanolamines Incorporated
34 into Poly(vinyl Alcohol) Matrix for CO2/N2 Separation. J. Membr. Sci. 2007, 303 (1–2), 54–63.
35 (44) Matsuyama, H.; Terada, A.; Nakagawara, T.; Kitamura, Y.; Teramoto, M. Facilitated Transport
36 of CO 2 through Polyethylenimine/poly (Vinyl Alcohol) Blend Membrane. J. Membr. Sci. 1999,
37
163 (2), 221–227.
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 17 of 30 Industrial & Engineering Chemistry Research

1
2
3 Polymer Water permeability References
4 (barrer)
5 18
Matrimid 3200-3800
6 6FDA-MPDS 20000-27000 18
7 Cellulose Acetate CDA 22500 19
8 19
Cellulose Acetate CTA 14000
9 22
10 Teflon AF 1600 430
22
11 Hyflon AD60 810
12 Table.1 Water permeability in different polymeric membranes
13
14
15 Experimental CO2 N2 O2 αCO2/N2 Reference
16
conditions (T Permeability Permeability Permeability
17
18 / R.H) (barrer) (barrer) (barrer)
19
Matrimid 35°C / 0% 8,3 0,2 1,9 41,5 [16], [18],
20
performance [25]
21
22 (z=12µm) 35°C / >90% 5,66 0,17 1,46 33 [16], [18]
23
24 BAM-1 95°C / >90% 1850 4,3 60 430 [26]
25 (FSCM)
26 (z=45µm)
27
28
29 Table 2. (Modified) Component permeability in the two membranes used in simulation (Matrimid
30 and FSCM)
31
32
33
34
35
36 Matrimid with saturated flue Matrimid with dry flue gas
37 gas
38 Pressure ratio ψ 0,1 0,1
39 Feed flue gas (ton/h) 5 4,69
40 Temperature (°C) 35 35
41 Relative Humidity (RH) 100% 0%
42
Inlet steam composition
43
44
CO2 12% 13,3%
45 N2 73% 81,1%
46 O2 5% 5,6%
47 H2 O 10% 0%
48
49
50 Table. 3 Operating conditions used for the comparison between FSCM and dense polymeric
51 membrane
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 18 of 30

1
2
3 Matrimid with saturated flue Facilitated transport
4 gas membrane (BAM-1) with
5 saturated flue gas
6 Upstream pressure (bar) 2,8 2,8
7 Downstream pressure (bar) 0,25 0,25
8
Feed flue gas (ton/h) 5 5
9
10 Temperature (°C) 35 50
11 Relative Humidity (RH) 100% 100%
12 Thickness (µm) 1 45
13 Inlet steam composition
14 CO2 12% 12%
15 N2 73% 73%
16
O2 5% 5%
17
H2 O 10% 10%
18
19 Table. 4 Simulation conditions used for the comparison between FSCM and dense polymeric
20 membrane
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 19 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5 Robeson upper bound
6 1000
Nafion 117
7
8 Nafion 117 (RH:70%)
9 Aquivion
10 Aquivion (RH:70%)
11 6FDA-6FpDA
12 100 6FDA-6FpDA (RH:90%)
13
Selectivity αCO2/N2

14 PIM-1
15 PIM-1 (RH:40%)
16 PDMS
17 PDMS (RH:41%)
18
10 PEBAX 1657
19
20 PEBAX 1657 (RH:80%)
21 Polysulfone
22 PPO
23 PBO
24
1 Polyactive
25
26 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 FTM
27 Permeability CO2 (barrer) FSC membrane
28
29 Figure 3: (Modified) Comparison of different membrane CO2 / N2 separation performances for dry
30
and humid gas mixtures 38 20 23 28 39 40 9 15 41 42 43 26 44.
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 20 of 30

1
2
3
4
5 12%
6 T=10°C
7 x water vapour (%) 10%
T=30°C
8 8%
9
T=50°C
10 6%
11 4%
12
13 2%
14
0%
15
1 3 5 7 9
16
17 compressor pressure (bars)
18
19
20 Figure 4: Water vapour volume fraction from a compressor / heat exchanger cooler outlet stream. A
21 saturated flue gas is taken at the temperature indicated on the curves for inlet compressor
22 conditions. The inlet saturated feed flow composition in dry basis is 13.3% CO2, 81.1% N2 and 5.6%
23 O2.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 21 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7
8
9
10
11 (a)
12
13
14
15
16
17
18
19
20
21
22
(b)
23
24
25
26
27
28
29
30
31
32 (c)
33
34
Figure 5: Driving force strategies ((a) compression/vacuum, (b) compression and (c) vacuum)
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 22 of 30

1
2
3 80.00%
4
70.00% p'=1/p''=0.1 bar
5
6 60.00%
p'=2/p''=0.2 bars
7 p'=5/p''=0.5bars
8 50.00%
y co2 (%)

9 p'=10/p''=1 bars
40.00%
10 Dry gas
11 30.00%
12
13 20.00%
14
10.00%
15
16 0.00%
17 10.00% 30.00% 50.00% 70.00% 90.00%
18 Rco2 (%)
19
20
21 Figure 6a: CO2 purity in a humid permeate flow in function of the compression system at same
22 pressure ratio (inlet pressure (p’)/permeate pressure (p’’) =0.1) with a Matrimid membrane
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 23 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5 90%
6
80%
7
8 70%
9
10 60% p'=1/p''=0.1 bar
yco2(%)

11 p'=2/p''=0.2 bars
50%
12
13 40% p'=5/p''=0.5 bars
14 p'=10/p''=1 bars
15 30%
dry gas
16 20%
17
18 10%
19 10.00% 30.00% 50.00% 70.00% 90.00%
20 Rco2 (%)
21
22
23 Figure 6b: Influence of the driving force with a saturated feed flow on the CO2 purity in the permeate
24 stream in dry basis. Simulation results are in function of the compression / vacuum conditions under
25 similar pressure ratio (inlet pressure (p’)/permeate pressure (p’’) =0.1) with a Matrimid membrane.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 24 of 30

1
2
3
4
5 7.E+06
6
6.E+06
Surcae area (m²/(Kg of co2/s))
7 Dry gas
8
9 5.E+06
10 Humid gas without FV
4.E+06
11 reduction and
12 3.E+06 competitive sorption
13
Humid gas
14 2.E+06
15
16 1.E+06
Humid gas without
17
water sweep effect
18 0.E+00
19 10% 30% 50% 70% 90%
20 Rco2 (%)
21
22
23 Figure 7: Membrane surface area for dry and humid flue gas and impact of the competitive sorption
24 and the Free Volume (FV) reduction and the water sweep effect. Matrimid membrane, vacuum
25 operation (inlet pressure (p’)=1bar/permeate pressure (p’’)=0.1bar).
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 25 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7
8 3.60E+06
9
10
Surface area (m²/(kg of co2/s))

3.10E+06
11
12
2.60E+06 Dry gas p'=10/p''=1 bar
13
14 Humid gas p'=10/p''=1 bar
2.10E+06
15
Dry gas p'=5/p''=0.5 bar
16
17 1.60E+06 Humid gas p'=5/p''=0.5 bar
18 Dry gas p'=2/p''=0.2 bar
19 1.10E+06
Humid gas p'=2/p''=0.2 bar
20
21 6.00E+05
22
23 1.00E+05
24 10.00% 30.00% 50.00% 70.00% 90.00%
25 Rco2 (%)
26
27
28 Figure 8.a: Impact of driving force strategies (compression and/or vacuum) for the same pressure
29 ratio value (inlet pressure (p’)/permeate pressure (p’’) =0.1) on the membrane surface area of a
30 Matrimid membrane.
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 26 of 30

1
2
3
4
5 1.00E+01
6
7 9.00E+00
8 Dry gas p'=10/p''=1 bar
Energy requirement (GJ/ton of co2)

8.00E+00
9
10 7.00E+00 Humid gas p'=10/p''=1 bar
11
12 6.00E+00
Dry gas p'=5/p''=0.5 bar
13
5.00E+00
14
Humid gas p'=5/p''=0.5 bar
15 4.00E+00
16
17 3.00E+00 Dry gas p'=1/p''=0.1 bar
18
19 2.00E+00 Humid gas p'=1/p''=0.1 bar
20
1.00E+00
21 Amine absorpber energy
22 0.00E+00 requirement
23 10.00% 30.00% 50.00% 70.00% 90.00%
24 Rco2 (%)
25
26
27 Figure 8.b: Impact of driving force strategies (compression and/or vacuum) for the same pressure
28 ratio value (inlet pressure (p’)/permeate pressure (p’’) =0.1) on the energy requirement of a
29
Matrimid membrane.
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 27 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5 100%
6 90%
7 80%
FSCM
8
70%
9
60%
yco2 (%)

10
11 50% Matrmid
12 40%
13 30%
14
20%
15 Hypothetique
16 10% membrane
17 0% PCO2=0.1 and
15.00% 35.00% 55.00% 75.00% 95.00% CO2/N2 s=996
18
19 Rco2 (%)
20
21
22 Figure 9: Comparison of CO2 purity in the permeate flow between a Matrimid membrane, BAM-1 and
23 a hypothetical polymeric membrane with performance on the Robeson upper bound (CO2
24 permeability of 0.1 Barrer and CO2/N2 selectivity of 996). Same operating conditions: p’=2.8 Bar and
25 p’’= 0.25 Bar.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 28 of 30

1
2
3
4
5
6 Surface area (m²/(kg of CO2/s)) 4.E+06
7
8 3.E+06
9 3.E+06
10
2.E+06
11 Matrimid
12 2.E+06
13 FSCM
1.E+06
14
15 5.E+05
16 2.E+04
17 15% 35% 55% 75% 95%
18
RCO2 (%)
19
20
21 Figure 10.a: Surface area comparison between Matrimid and BAM-1 membrane. Same operating
22
conditions: p’=2.8 Bar and p’’= 0.25 Bar.
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Page 29 of 30 Industrial & Engineering Chemistry Research

1
2
3
4
5
6
7 1.E+01
Energy requirement (GJ/ton of co2)

8 9.E+00
9 8.E+00
10
11 7.E+00
12 6.E+00
13 5.E+00
14 FSCM
4.E+00
15 Matrimid
16 3.E+00
17 2.E+00
18 1.E+00
19
20 0.E+00
21 15% 35% 55% 75% 95%
22 Rco2 (%)
23
24
25 Figure 10.b: Energy requirement comparison between Matrimid and BAM-1 membrane. Same
26 operating conditions: p’=2.8 Bar and p’’= 0.25 Bar.
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
 Industrial & Engineering Chemistry Research Page 30 of 30

1
2
3 TOC Graphic
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment