Membranes and Post Combustion CarbonDioxide Capture: Challenges and Prospects
Eric FAVRE
ENSIC Nancy UniversitéLaboratoire Réactions & Génie des Procédés CNRS
NANCY FRANCE
i) Introduction: Membranes & Post combustion capture
ii) Membrane contactors
iii) Gas separation membranes
iv) Chemically reactive membranes and alternative approaches
vi) Conclusion
Outline
0 2 4 60
2
4
6
c2 (g
/L)
c1 (g/L)
SMB Classique M3C Continu M3C Moyen
Introduction
Separation processes in industry at a glance
I. Thermal energy
II. Auxiliary phase
III. Membrane
Separation agent:
(Humphrey & Keller, Separation Process Technology, 2005)
Van der Sluis, J.P. et al. C. Feasability of polymer membranes for carbon dioxide recovery from flue gases, Energy Conversion & Management, 33, 5-8 (1992) 429.
O. Davidson, B., Metz, Special Report on Carbon Dioxide Capture and Storage, International Panel on Climate Change , Geneva, Switzerland, 2005
Gas separation membranes and post combustion carbon capture
Source: Figueirao J. et al. DOE (2007) Int. J. Greenhouse Gas Control
Carbon capture processes: A tentative roadmap 1/2
Carbon capture processes: A tentative roadmap 2/2
Carbon capture strategy
Target mixture Conditions First generation separation
process
Possible breakthrough
membrane process
Oxycombustion O2/N2 P atmosphericT ambient
Cryogeny Ion Transfer Membranes (ITM)
Precombustion CO2/H2 P up to 80 BarT 300 – 500 C
Gas-liquid absorption in
physical solvent
Membrane reactor
Postcombustion CO2/N2 P atmosphericT 100 – 250 C
Gas-liquid absorption in
chemical solvent (MEA)
Membrane gas separation
Membranes & carbon capture processes
Membranes and post-combustion carbon capture: Synopsis
• Intensified gas liquid absorption: Membrane contactors
• Gas permeation membranes: physical
• Gas permeation membranes: chemical
• Alternative approaches & miscellaneous
Procédés de capture du CO2
i) Absorption gaz liquide conventionnelle
ii) Absorption par contacteurs à membranes
iii) Adsorption
iv) Membranes0 2 4 6
0
2
4
6
c2 (g
/L)
c1 (g/L)
SMB Classique M3C Continu M3C Moyen
Membranes contactors for intensified
CO2 absorption processes
Membranes and post-combustion carbon capture: Synopsis
• Intensified gas liquid absorption: Membrane contactors
• Gas permeation membranes: physical
• Gas permeation membranes: chemical
• Alternative approaches, miscellaneous
CAPEX
Membrane contactors for intensified gas-liquid absorption: Principle
dP �� cos4max��
• Key issue: design a porous materialand use it between the liquid solvent and the gas phase under non wettingconditions
• Advantages:- increased interfacial area (a) - no flooding or weeping limitations- improved liquid distribution, no sensitivity to orientation- limited solvent losses- scale-up modular and easy
• Drawbacks:-additional mass transfer resistance- no economy of scale (numbering up)
Membrane contactors: a novel intensification process for fluid mass tranfer processes
Hollow fiber moduleInternal fiber diameter ~ 0.2-5 mmSpecific interfacial area (a): 1000-5000 m-1
� � � �Lmlm
e
eG
eINTG
kEmrkr
rkrK
..1
..
1,
�
�
�
Membrane contactors in industry at a glance
Membrane phase contactor unit for stripping of oxygen from water. (Provided by Hoechst Celanese Corporation).
• Numerous industrialapplications (carbonatedbeverages, aroma recovery, oxygen stripping…
• Dozens of publications each month on CO2 capture at lab scale…
• Very limited feedback and limited number of actors atpilot and industrial scale(Statoil, TNO)
• 2 main membrane materials: PP and PTFE
r
r e
MEA
solution
CO2mixture
Gas -liquid
interfacer
Liquid
boundary
layer
Gas
boundary
layer
Microporous
membrane
Liquidbulk Gasbulk
z
��
..m
mD
k �
Membrane contactors for CO2 capture: Materials challenge
• D: diffusion coefficient (bulk – Knudsen)
• �: Porosity
• : Membrane thickness
• �: Tortuosity
Trade-off between permeability and breakthrough pressure
Polypropylene (PP)
Polyvinylfluoridilène (PVDF)
Polytetrafluoroethylene (PTFE)
Nylon
Porous hydrophobic membrane materials: Examples of different structures
Membrane contactors and CO2 absorption in 30% MEA:A parametric study
Reference case: Packed column
Membrane contactor
Inlet solvent (MEA)
loading ( IN)
0.24
(Abu-Zahra et al., 2007)
Variable (usually 0)
Outlet solvent (MEA)
loading ( IN)
0.48
(Abu-Zahra et al., 2007)
Variable
Gas phase pressure drop
(�PG)
5000 – 7000 Pa
(NETL, 2007)
Usually not mentioned
Gas-liquid interfacial area
(a)
200-500 m-1
(Tobiesen et al., 2007)
1000 – 5000 m-1
(Gabelman & Hwang,
1999)
Overall mass transfer
coefficient (K)
4x10-3 – 10-2 m.s-1
(Tobiesen et al., 2007)
3x10-4 – 10-3 m.s-1
(Feron et al., 2002)
Overall volumetric CO2
absorption capacity (C)
1 mol CO2 .m-3.s-1
(Tobiesen et al., 2007;
NETL, 2007; Mangalapally
et al., 2010)
0.7 - 10 mol CO2 .m-3.s-1
(Nishikawa et al., 1995;
Feron and Jansen, 2002;
Yeon et al., 2003; Hoff et
al., 2004; deMontigny et
al., 2005)
Parametric study: Targets and framework
1. Compute, through a number of simulations, the volumetric absorption capacity of differentmembrane contactors (isothermal conditions, plug flow on gas side, liquid in, countercurrent mode, 30 % MEA, 15%CO2 in feedmixture, capture ratio 0.9)
2. Select the set of data with a maximal gaspressure drop of 50 mBar and liquid pressure drop of 1 Bar
3. Identify the requirements in terms of membrane permeance, fiber geometry and module packing fraction which enableintensification to be achieved
0
20
40
60
80
100
0 0,01 0,02 0,03u
G
Lab scale tests and simulations with membrane contactors
A typical lab scale set-up for membrane contactors testing
Prediction performances of simulations with 30 % MEA
Variable
Variation range
Unit
re
hollow fiber external
radius
10-4 - 3.10-3
[m]
�
membrane thickness
5.10-5 - 10-4
[m]
�
module packing
fraction
0.1 - 0.6
[-]
uG
interstitial gas velocity
0.1 - 5
[m.s-1]
km
membrane effective
mass transfer coefficient
10-2 - 10-5
[m.s-1]
Parametric sensitivity: variables and range of variation
0
1
2
3
4
5
6
7
8
9
0 2000 4000 6000 8000 10000 12000
Inte
nsifi
catio
n fa
ctor
(I)
Parametric study: summary of results
Approximately 11 000 simulations performed
An intensification factor larger than 1 isobtained within the �P limits for ca 300 cases, with module length between 0.1 and 2m
A maximal intensification factor of 8.5 seems to be attainable for a very limitednumber of cases
An intensification larger than 1 requires: km >7x10-4 m/s < 80 µm and re < 400 µm
A packing fraction of 0.5-0.6 is requiredfor an intensification factor larger than 8
Ranking of the simulation results in termsof intensification factor
Membrane contactors and CO2 capture: one step forward….
Long time scale results are required for material selection
Experimental results at a large scale, on real flue gases are of primiraly importance
Non isothermal conditions, water evaporation, gas and liquid dispersion effectsshould be added in simulation packages
Desorption thanks to temperature resistantMC should be more systematically explored
Dense skin membrane contactors in order to prevent liquid wetting problems
Procédés de capture du CO2
i) Absorption gaz liquide conventionnelle
ii) Absorption par contacteurs à membranes
iii) Adsorption
iv) Membranes0 2 4 6
0
2
4
6
c2 (g
/L)
c1 (g/L)
SMB Classique M3C Continu M3C Moyen
Post combustion capture by gas
separation membranes
Membranes and post-combustion carbon capture: Synopsis
• Intensified gas liquid absorption: Membrane contactors
• Gas permeation membranes: physical
• Gas permeation membranes: chemical
• Alternative approaches, miscellaneous
CAPEX + OPEX
Dense membrane permeation
JA
Membrane thickness
Upstream pressurepfeed
Downstream pressurepperm
�
Component AComponent B
pfeed > pperm
(1) Sorption on upstream side(2) Diffusion down partial pressure gradient(3) Desorption on downstream side
Permeability of A � PA = DA SA
where DA = diffusion coefficient
SA = solubility coefficient
Selectivity
/A A A
A BB B B
P S DP S D
� � � � � �
� �� � � � � �� � � � � �
Membrane Gas Separation: Applications & Market
0.1
1
10
100
1000
0.1 1 10 100 1000 104
*
Permeability (Barrers)
O2/N
2
CO2/N
2
CO2/CH
4
H2/N
2
Trade-off curves predictions for physical process
Step 1: Selectivity challenge
Feed composition xCapture ratio RPurity y
Selectivity challenge
Capture Cost
in
p
��
P’ : Upstream pressure
P’’ : Downstream pressureQp: Flowratey: CO2 mole fraction
Stage cut
PIn: Feed pressureQin: Flowratexin: CO2 mole fraction
RetentateFeed
Permeate To compression
Qout: Flowratexout: CO2 mole fraction
Key variables
Pressure ratio
2
2*N
CO
��
�
'''PP
��
Selectivity
Gas permeation module design: single stage case
Separation performances:
R: Recovery= Capture ratio
y: Permeate content CO2
Q in Q out x in x out
y Q p
dQ
Q Q - dQ x x - dx
P ’
P ’’
FEED RETENTATE
PERMEATE
z dA
Permeation module simulation: Cross plug flow case
ODE system resolution : DASSL (M3Pro© software)
���
��� ��
dsdQxyx
dsdxQ p
*...* �
� ����
��� �� )1.(11.*
pp yxyxds
dQ �
�
���
����
�
�
)1.(1.
.1 p
p
p
p
yxyx
yy
��
Recovery (target > 80%)Pe
rmea
te c
onte
nt y
Bounaceur, R., Lape, N., Roizard, D., Vallières, C., Favre, E. (2005) Membrane processes for post-combustion carbon dioxide capture : a parametric study. Energy
Gas permeation module master / curves
10% CO2 in flue gas:
Impossible to achieve
R> 0.8 and y >0.8
unless a > 200
20% CO2 in flue gas:
Possible to achieve
R> 0.8 and y >0.8
if > 50
Gas permeation module master curves
Step 2: Energy challenge
Feed composition xCapture ratio RPurity y
Selectivity challenge
Operating Conditions�����
Energy Challenge
E < 2 GJth/tonOPEX
Energy cost
Capture Cost
Energy requirement computation: Feed compression
���
�
�
���
�
��
�
����
�
�
1PP
.1
RT..QE
1
min
maxIn1
��
��
!
P’ : Upstream pressure
P’’ : Downstream pressureQp: Flowratey: CO2 mole fraction
PIn: 1 atmQin: Flowratexin: CO2 mole fraction
RetentateFeed
Permeate
Qout: Flowratexout: CO2 mole fraction
0
2
4
6
8
10
12
0 50 100 150 200
Ener
gy E
(GJ /
ton
CO2 re
cove
red)
xin = 0.3
xin = 0.2
xin = 0.1 R = 0.8 y = 0.8
1. High parametric sensitivity: E strongly influenced by xin2. For a 10% CO2, content the energy requirement and the minimal
membrane selectivity are too high
Parametric study results: Feed compression
Energy requirement computation: Vacuum pumping
���
�
�
���
�
��
�
����
�
�
1PP
.1
RT..'
QE
1
min
maxP1
��
��
!
P’ : Upstream pressure
P’’ : Downstream pressureQp: Flowratey: CO2 mole fraction
PIn: 1 atmQin: Flowratexin: CO2 mole fraction
RetentateFeed
Permeate Pipe & injection
Qout: Flowratexout: CO2 mole fraction
0
0.5
1
1.5
2
2.5
3
0 50 100 150 200
Ene
rgy
E (
GJ
/ ton
CO
2 re
cove
red)
xin = 0.1
xin = 0.2
xin = 0.3
R = 0.8 y = 0.8
Parametric study results: Vacuum pumping
1. E below 0.5 GJ per ton can be achieved 2. For a 20% CO2, content the required membrane selectivity
for E = 0.5 GJ per ton is around 60, i.e. a realistic figure
The ultimate (and largely unexplored) challenge: energy, productivity, cost
Feed composition xCapture ratio RPurity y
Selectivity challenge
Operating Conditions�����
Energy Challenge
E < 2 GJth/ton
Feed flow ratePermeabilityThickness
ProductivityChallenge
Maximal flux
OPEX
CAPEX
Energy cost
Membrane cost(€/m2)
Capture Cost
Tackling the energy (pressure ratio) / capacity trade off
Pressure ratio
Capture ratio
Surface area
15% CO2 in feed mixture 90% purity at permeate
Ind. Eng. Chem. Res. (2008) 47, 5, 1562
Main conclusion:
Best solution : single stage withvacuum
Selectivity: > 40 required (but y
will be limited to 0.6)
Ind. Eng. Chem. Res. (2008) 47, 5, 1562
Main conclusion (ctd):
High permeability required (500 Barrer with z 0.125 µm)
Ind. Eng. Chem. Res. (2008) 47, 5, 1562
Main conclusion (ctd):
Low membrane cost required (< 50 $/sqm)
* CO2/N2maximum ~ 60
� CO2maximum ~ 500 Barrer
Selectivity challenge: "Physical" separation membranes
Favre, E. (2007) Carbon dioxide recovery from post combustion processes: Can gas permeation membranes compete with absorption? Journal of Membrane Science, 294, 50-59
MTRPolaris membrane
~ 501000 GPU
The only way to overcome the energy challenge seems to apply vacuum, nevertheless, this option may show major limitations:
- energy efficiency can be low compared to compressors (! ~ 0.85)- vacuum pump foot print is large and leaks can be problematic
Option 1: Single stage membrane system
Herzog et al., Environ. Prog. (1991) 10, 64-74.
CO2 recovery 80%, CO2 purity 90%Energy requirement 50-75 % of combustion energy of coal
(MEA 47-79 %)
One of the earliest membrane flowsheet for post-combustion CCS application
Option 2: Multi-stage membrane systems
Challenges & unsolved issues: Water
1
10
100
1000
10000
0,1 1 10 100 1000 10000
P CO2
(Barrer)
CO2 / N
2
CO2 / Ar
CO2 / O
2
CO2 / SO
2
CO2 / NO
CO2 / CO
0,1
1
10
100
0,1 1 10 102 103 104 105
CO
2/O
2
CO2 permeability (Barrer)
Challenges & unsolved issues: Oxygen and “minor species”
Rejection increases
Procédés de capture du CO2
i) Absorption gaz liquide conventionnelle
ii) Absorption par contacteurs à membranes
iii) Adsorption
iv) Membranes0 2 4 6
0
2
4
6
c2 (g
/L)
c1 (g/L)
SMB Classique M3C Continu M3C Moyen
MMembranes and post
combustion carbon capture:
Chemically reactive membranes
Membranes and post-combustion carbon capture: Synopsis
• Intensified gas liquid absorption: Membrane contactors
• Gas permeation membranes: physical
• Gas permeation membranes: chemical
• Alternative approaches, miscellaneous
CAPEX + OPEX
Looking for improved performances : Chemically reactive membranes with fixed sites
• Polyelectrolyte membranes (PVBTAF) with water saturatedmixture permeation: = 900 (Quinn,1977)
• Plasma grafted polyacrylic acidmembrane with amine carrier (Matsumya, 1994) = 4700, P = 106 Barrer!!
• Polyvinylamine (Hagg, 2005) = 150-250
• PAMAM (Duan, 2006) , = 260
Looking for improved performances : Liquid membranes
• Numerous studies for a long time!
• WWard & Robb (1966) > 4600
• SSirkar, Ho, Noble…
Membrane type Material and/or carrier CO2/N2selectivity
CO2 permeability (Barrer) or permeance
(GPU)
Gas separation membrane (dense
polymers)
PEO-PBTPEG/Pebax©
PEG-DME/ Pebax©
PEGDA/PEGMEAPolaris™
7047434150
120 Barrer151 Barrer600 Barrer570 Barrer1000 GPU
Fixed Site Carrier
Membrane(FSCM)
PAAM-PVA / PSPVAm/PVAPEI / PVAPDMA/PSPDMAMA
801452305380
24 GPU212 GPU
1 GPU30 GPU5 GPU
Liquid Membrane
(LM)
PVAm-PVA/PSPVAm/PVAAmines/PVA
Carbonic anhydraseAmines / PVA
9090500250493
22 GPU15 GPU250 GPU80 GPU
693 Barrer
Chemically reactive membranes: Performances
Assymetric PVTMS membrane selective layer thickness 0.2 µm
gas-penetrant
gas-carrier
liquid carrier260 µm
membranes
Carrier K2CO3 aqueous solution
Liquid membrane processes
Shalygin, M.G., Roizard, D. Favre, E., Teplyakov V.V. (2008). Journal of Membrane Science, 318, 317-326.
PVTMS
PTMSP
PE PDMSEthyl
cellulose
PP
Water
Ethanol Methanol
Propylenecarbonate
1,0E+00
1,0E+01
1,0E+02
1,0E+03
1,0E+04
1,0E+01 1,0E+02 1,0E+03 1,0E+04 1,0E+05CO2 permeability coefficient, Barrer
CO
2/N
2 s
elec
tivi
ty
Selectivity challenge: "Chemical" separation membranes
Fixed site carrier membranes:Selectivity 80-250Water required
Liquid membranes:Selectivity up to 4500Water required
Key issues:Productivity (l ~5-7 µm)Driving force (sweep)Stability
Carbonic anhydrase LM: Carbozyme concept
Procédés de capture du CO2
i) Absorption gaz liquide conventionnelle
ii) Absorption par contacteurs à membranes
iii) Adsorption
iv) Membranes0 2 4 6
0
2
4
6
c2 (g
/L)
c1 (g/L)
SMB Classique M3C Continu M3C Moyen
Membranes processes for CO2 capture:
Alternative approaches
Looking for alternative emission sources…
��
���
��"�
yyyyy
xxxxxTW )-1ln().-1()ln(. -)-1ln().-1()ln(..-min
0
5000
10000
15000
0 0,2 0,4 0,6 0,8 1
W
(J.mol -1)
x (% vol)
y = 1
y = 0.9
y = 0.7
y = 0.5
AirGas Coal Steel,
cementFermentation
NH3, EO
Wmin
x
y
Towards breakthrough membrane materials?
Zeolite membranes performances. Predictions based on MD simulations
(J. Memb. Sci., in press)
Unconventional approach : Reverse selective membranes
CO2 selective membrane N2 selective membrane
y
RFavre, E., Roizard, D., Koros, W.J. (2009) Ind. Eng. Chem. Res., 48 (7) 3700-3701.
Hybrid processes
+=
x)x-1ln().x-1()xln(.x
T.R-W 'min
��
���
�"���
��� �
"�y
)y-1ln().y-1()yln(.y .Tx
)x-1ln().x-1()xln(.xT.-W ''min
2
1
Air Power
3
CompressioTransport Storage
4
y x’ x
E1 E2
N2 �� O2
O2 N2
CO2 H2O N2
Fuel N2 �� CO2
Flue gas
CO2
H2O
CO2 N2
Hybrid process: Rationale
Cryogeny Gas permeation membrane module
Combustion Flue gas drying
44'.x.R10.1
"P'P.
'T.
1n21n3.
RE
2E1EE3
1
2O
��
�
�
��
�
�
��
���"
�
����
��
!��
Hybrid process:Principle
Natural gas (n =1)Most challenging!
Single module
Feed compression
R = 0.9
y = 0.9
Hybrid process: Simulation results
�50-60% oxygen in feed air seems to be the best compromise
Favre, E., Bounaceur, R., Roizard, D. (2009). Separation & Purification Technology, 68, 30-36.
0,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
0 0,2 0,4 0,6 0,8 1
y
E [
GJ/
ton
CO
2]
XCH4 = 0.95
XCH4 = 0.90
XCH4 = 0.80
XCH4 = 0.70
XCH4 = 0.50R = 0.9y = 0.9 = 50
Carbon dioxyde capture by a membrane unit from biogas combustion unit : simulation results
Oxygen mole fractionin enriched air
Methane mole fractionin biogas
Favre, E., Bounaceur, R., Roizard, D. (2009) Biogas, membranes and carbon dioxide capture. Journal of Membrane Science, 328, 11-14.
Miscellaneous….
• Flue gas drying
•Compressor purge
•Alernate driving forces: temperature, wator vapor sweep
• Pressurised combustion (e.g.Combicap concept)
• Integrated approach (S, N, C removal)
• High temperature permeation
….
Procédés de capture du CO2
i) Absorption gaz liquide conventionnelle
ii) Absorption par contacteurs à membranes
iii) Adsorption
iv) Membranes0 2 4 6
0
2
4
6
c2 (g
/L)
c1 (g/L)
SMB Classique M3C Continu M3C Moyen
Conclusion
• Membranes processes offer a large variety of potential applications in a CCS framework (intensification, separation, concentration)
• For the capture step, the purity and recovery specifications play a key role(strong parametric sensitivity, strong flexibility)
•Material science & process engineering collaboration is crucial!
• Pilot studies & technico-economical analyses are needed!
•Post combustion CCS is a complex, multivariable landscape: no silver bullet!
•Gas separation membranes missed (as usual) the first generation
Gas permeation membranes and post combustion CCS: utopy or opportunity?
Acknowledgements:
N. Boucif, E. Chabanon, P.T. Nguyen
R. Bounaceur, C. Castel, S. Rode, D. Roizard
Support & funding:
CNRS
ANR
EU / FP7 (CESAR)
Target Membrane process Main characteristics
Intensified gas liquid absorption
Membrane contactor(hydrophobic porous membrane)
Membrane stability (non wetting conditions) has to be ensured for long time operationRegeneration step non achievableEffective intensification factor remains to be clearly evaluated
Carbon dioxide capture
1. Gas separation membranes: physical (GS) (usually a dense polymer)
Numerous materials investigated, few technico-economical studies, almost no pilot scale process on real flue gasWater in flue gas could be a problemRequire multistage processes in order to achieve purity target
2. Chemically reactive membranes (FSCM)
Require water, on both sides of the membrane,to be effectiveMay be applicable through a single stage process if a high selectivity is attained Industrial feedback on real flue gas still not achieved
3. Liquid membranes (LM) Very high selectivity can be achievedRequire water, on both sides of the membraneto be effectiveSolvent losses due to volatility can be a problem, possibly solved by ionic liquidsEffective drivig force, stability and effective permeability in a real psot-combustion situation remain to be
Liquid nitrogen
Adsorption
Cryogeny
Membrane
N2 purity
Flowrate (Nm3.h-1)10 250 5000
90
95
99.9
Source: Greenhouse Issues (2006), 84, 12
Source: Greenhouse Issues (2006), 84, 12