View
0
Download
0
Category
Preview:
Citation preview
DEVELOPMENT OF MIXED MATRIX MEMBRANES
FOR GAS SEPARATION APPLICATION
LI YI
(M. Eng., Tsinghua University, P. R. China)
A THESIS SUBMITTED
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
DEPARTMENT OF CHEMICAL AND BIOMOLECULAR
ENGINEERING
NATIONAL UNIVERSITY OF SINGAPORE
2006
i
ACKNOWLEDGEMENTS
I wish to take this opportunity to express my sincere appreciation to all the
contributors whose cooperation and assistance were essential in helping me to
gradually acquire sharper tools toward the completion of my PhD study reported
herein:
First of all, I especially wish to record my deepest appreciation and thanks to Prof.
Chung. He has given me every opportunity to learn about membrane science and
provide the essential facilities to carry out my research. His enthusiasm, positive
outlook and belief in my abilities kept me going through the most difficult phase of
research. I am also indebted to my co-supervisors, Dr. Pramoda Kumari Pallathadka
and Dr. Liu Ye for their keen efforts and consistent consultation throughout my
candidature.
I may also like to express my appreciation to my PhD thesis committee members, Prof.
Hong Liang and Dr. Chen Jia Ping. Their suggestions on my PhD proposal were
constructive throughout my candidature in NUS.
Special thanks are due to all the team members in Prof. Chung’s research group. Dr.
Cao Chun and Ms. Jiang Lan Ying are especially recognized for their guidance and
help in my initial study step. Special thanks go to Dr. Huang Zhen for providing the
ii
zeolite Beta and Ms. Guan Huai Min for providing the silane modification method of
zeolite surface that are indispensable to my research. The suggestions on permeation
cell set-up and modification from Ms. Chng Mei Lin and Dr. Tin Pei Shi were
precious. All the members in Prof. Chung’s group are kind and helpful to me, which
have made my study in NUS enjoyable and memorable.
I would like to gratefully acknowledge the research scholarship offered to me by the
National University of Singapore (NUS), which provided me a positive, conducive
and professional atmosphere for researching. Appreciation also goes to the staff in the
Department of Chemical and Biomolecular Engineering that have helped me in
various characterization techniques and given me professional suggestions.
I would also like to convey my thanks to Dr. S. Kulprathipanja from UOP LLC for his
valuable advices in my work on mixed matrix membranes and zeolites. Thanks also go
to NUS and UOP LLC for the financial support with the grant numbers of R-279-000-
108-112, R-279-000-140-592, R-279-000-140-112 and R-279-000-184-112.
Last but not least, I must express my special thanks to my wife, Wu Qiong, for her
unwavering and unconditional love and support. My family members also deserve the
special recognition for their love and endless encouragement and support.
iii
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENT…………………………………………….………………..i
TABLE OF CONTENTS……………………………………………………………..iii
SUMMARY…………………………………………………………………..………xii
NOMENCLATURE………………………………………………………….………xv
LIST OF TABLES…………………………………………………………………..xxii
LIST OF FIGURES…………………………………………………………………xxv
CHAPTER 1 INTRODUCTION OF GAS SEPARATION MEMBRANE……….1
1.1 Scientific Milestones……………...…………………………………………….3
1.2 Importance of Gas Separations Using Membranes………………….………….7
1.2.1 Separation of O2 and N2…………….…………………………………. 8
1.2.2 Separation of H2 and Hydrocarbon Gases……….…………………….10
1.2.3 Separation of H2 and CO……………………….……………………...10
1.2.4 Separation of H2 and N2……………………………………….………11
1.2.5 Acid Gas Removal from Natural Gas……………………………....…12
1.3 Basic Concept of Membrane Separation………………………………………14
1.4 Types of Membrane Structures………………………………………………..18
1.4.1 Dense vs. Porous Membranes……………………………………..…..18
1.4.2 Symmetric vs. Asymmetric Membranes…………….………………...19
1.4.3 Dynamic in-situ Membranes………………………………….……….22
iv
1.4.4 Liquid Membranes…………………………………………….……....23
1.5 Mechanisms of Membrane Separation………………………………………...24
1.5.1 Poiseuille Flow………………………………………………………...25
1.5.2 Knudsen Diffusion…………………………………………………….25
1.5.3 Molecular Sieving……………………………………………………..26
1.5.4 Solution-diffusion……………………………………………………..28
1.5.4.1 Diffusion………..…………..…………………………………29
1.5.4.2 Sorption………………………………………………………..30
1.5.4.3 Selectivity……………………………………………………...31
1.6 Membrane Modules and Design Instructions……………….………………..33
1.6.1 Plate-and-Frame Modules……….…………………………………….33
1.6.2 Spiral-Wound Modules…………………….………………………….34
1.6.3 Hollow-Fiber Modules…………………….…………….…………….35
1.7 Research Objectives and Organization of Dissertation………………………..37
1.8 References……………………………………………………………………..41
CHAPTER 2 MIXED MATRIX MEMBRANES FOR GAS SEPARATION…..48
2.1 Emergence of Mixed Matrix Membranes……………………………………..48
2.1.1 Polymeric (Organic) Membrane Materials……………………………50
2.1.2 Inorganic Membrane Materials………………………………………..53
2.1.3 Mixed Matrix Membranes (MMMs)…………………………………..57
2.2 Development of Mixed Matrix Membrane…………………………………....58
v
2.2.1 Flat Dense Mixed Matrix Membranes………………..….……………58
2.2.2 Hollow Fiber Mixed Matrix Membranes……………………….……..64
2.3 Prediction of Gas Separation Performance of MMMs………………….……..66
2.3.1 Prediction for MMMs with an Ideal Interface………………………...66
2.3.2 Prediction for MMMs with an Nonideal Interface………….…………67
2.4 References……………………………………………………….…………….73
CHAPTER 3 MATERIALS AND EXPERIMENTAL PROCEDURES….…..…81
3.1 Materials……………………………………………………………………….82
3.1.1 Polymers………………………………………………………….……82
3.1.2 Molecular Sieves………………………………………………………83
3.1.3 Silane Coupling Agents…………………………………...…………..84
3.1.4 Others……………………………………………………………..…...85
3.2 Fabrication of Dual-layer PES Hollow Fiber Membranes with a Neat Polymeric
Outer Layer…………………………………………………………….……..85
3.2.1 Spinning Line………………………………………………………….85
3.2.2 Preparation of Spinning Dope…………………………………………86
3.2.3 Spinning Process and Solvent Exchange……………………………...87
3.2.4 Post-treatment Protocols………………………………………………88
3.3 Fabrication of Flat Dense PES-Zeolite A Mixed Matrix Membranes with Silane
Modified Zeolite or Unmodified zeolite…………………………………..….90
3.3.1 Chemical Modification Method of Zeolite Surface…..……….………90
vi
3.3.2 Preparation Procedures of Flat Dense Mixed Matrix Membranes…….91
3.4 Fabrication of Dual-layer PES/P84 Hollow Fiber Membranes with a PES-
Zeolite Beta Mixed Matrix Outer Layer…………………………………..….92
3.4.1 Preparation of Spinning Dope…………………………………………92
3.4.2 Spinning Process and Solvent Exchange……………………………...93
3.4.3 Post-treatment Methods……………………………………………….94
3.5 Characterization of Physical Properties…………………………………….…95
3.5.1 Brunauer-Emmett-Teller (BET)…………………………………….…95
3.5.2 Dynamic Light Scattering (DLS)……………….…………………..…95
3.5.3 Differential Scanning Calorimetry (DSC)…………………………….95
3.5.4 Elemental Analysis…………………………………………………....96
3.5.5 Scanning Electron Microscope (SEM)…………………………..……97
3.5.6 Energy Dispersion of X-ray (EDX)………………………….……..…97
3.5.7 X-ray Photoelectron Spectroscopy (XPS)……………………………..98
3.5.8 X-ray Diffraction (XRD)……………………………………..….……98
3.6 Characterization of Gas Transport Properties……………..………….……….98
3.6.1 Pure Gas Permeation Test……………………………………………..99
3.6.1.1 Neat Polymeric hollow fibers……………..…………….…….99
3.6.1.2 Flat Dense Neat Polymeric or Mixed Matrix Membranes…...101
3.6.1.3 Mixed Matrix Hollow Fibers……………..………….………105
3.6.2 Mixed Gas Permeation Test………………………………….………107
3.6.2.1 Neat Polymeric hollow fibers………………………….…….107
vii
3.6.2.2 Flat Dense Neat Polymeric or Mixed Matrix Membranes..….108
3.6.2.3 Mixed Matrix Hollow Fibers………………….………..……112
3.7 References……………………………………………………………………112
CHAPTER 4 FABRICATION OF DUAL-LAYER POLYETHERSULFONE
(PES) HOLLOW FIBER MEMBRANES WITH AN ULTRATHIN
DENSE SELECTIVE LAYER FOR GAS SEPARATION….......115
4.1 Introduction……………………………………………………………..……115
4.2 Results and Discussion……………………………………………………….119
4.2.1 The Effect of Different Post-treatment Protocols on Gas Separation
Performance………………………………………..……………......119
4.2.2 Membrane Morphology…………………………………………..….122
4.3 Conclusions……………………………………………..……………………128
4.4 References……………………………………………………………………129
CHAPTER 5 THE EFFECTS OF POLYMER CHAIN RIGIDIFICATION,
ZEOLITE PORE SIZE AND PORE BLOCKAGE ON
POLYETHERSULFONE (PES)-ZEOLITE A MIXED MATRIX
MEMBRANES…..............................................................................135
5.1 Introduction…………………………………………………………………..135
5.2 Results and Discussion……………………………………………………….138
viii
5.2.1 Effect of Membrane Preparation Methodology on Gas Separation
Performance…………………………………………………………138
5.2.2 Effect of Zeolite Loadings on Gas Separation Performance……..….142
5.2.3 New Modified Maxwell Model to Predict Gas Separation
Performance…………………………………………………………149
5.2.4 Effect of Pore Sizes of the Zeolite on Gas Separation Performances..154
5.3 Conclusions………….…………..……..…………………………….………156
5.4 References……………………………………………………………………158
CHAPTER 6 EFFECTS OF NOVEL SILANE MODIFICATION OF ZEOLITE
SURFACE ON POLYMER CHAIN RIGIDIFICATION AND
PARTIAL PORE BLOCKAGE IN POLYETHERSULFONE
(PES)-ZEOLITE A MIXED MATRIX MEMBRANES.…...........162
6.1 Introduction……………………................…………………………………..162
6.2 Results and Discussion……………………………………………………….165
6.2.1 Characterization and Comparison of Unmodified and Modified
Zeolites…………………………….……………………………..….165
6.2.2 Effect of Chemical Modification of Zeolite Surface on Gas Separation
Performance…………………………………………………………168
6.2.3 Effect of Zeolite Loadings on Gas Permeability……………….…….173
6.2.4 Effect of Zeolite Loadings on Gas Permselectivity….………..……..177
ix
6.2.5 Applicability of the Modified Maxwell Model to Predict Gas Separation
Performance…………………………………………………………178
6.3 Conclusions…………………………………………………………………..186
6.4 References…………………………………………………...……………….188
CHAPTER 7 DUAL-LAYER POLYETHERSULFONE (PES)/BTDA-TDI/MDI
CO-POLYIMIDE (P84) HOLLOW FIBER MEMBRANES WITH
A SUBMICRON PES-ZEOLITE BETA MIXED MATRIX
DENSE-SELECTIVE LAYER FOR GAS SEPARATION…......192
7.1 Introduction…………………………………………………..………………192
7.2 Results and Discussion……………………………………………………….197
7.2.1 Effects of Heat-treatment Temperature on Gas Separation
Performance……………………………….……………………..….197
7.2.2 Effect of Mixed Matrix Outer-layer Thickness on Gas Separation
Performance…………………………...…………………………….201
7.2.3 Effect of Air Gap on Gas Separation Performance…………..………206
7.2.4 Mixed Gas Separation Performance….…………..………………….208
7.3 Conclusions………………………………………………………….……….209
7.4 References……………………………………………………………………210
CHAPTER 8 CONCLUSIONS AND RECOMMENDATIONS…......................217
8.1 Conclusions………………………………………………..…………………217
x
8.1.1 Fabrication of Dual-layer Polyethersulfone (PES) Hollow Fiber
Membranes with an Ultrathin Dense Selective Layer for Gas
Separation………………………………..…………………………..218
8.1.2 The Effects of Polymer Chain Rigidification, Zeolite Pore Size and Pore
Blockage on Polyethersulfone (PES)-Zeolite A Mixed Matrix
Membranes……………………………..……………………………219
8.1.3 Effects of Novel Silane Modification of Zeolite Surface on Polymer
Chain Rigidification and Partial Pore Blockage in Polyethersulfone
(PES)-Zeolite A Mixed Matrix Membranes…………………..…….221
8.1.4 Dual-layer Polyethersulfone (PES)/BTDA-TDI/MDI Co-polyimide
(P84) Hollow Fiber Membranes with a Submicron PES-Zeolite Beta
Mixed Matrix Dense-Selective Layer for Gas Separation…………..222
8.2 Recommendations for Future Work………………………………………….223
8.2.1 Flat Dense Polymer-Zeolite Mixed Matrix Membranes……………..223
8.2.2 Dual-layer Hollow Fibers with a Mixed Matrix Outer Layer………..225
8.2.3 Other gas separation applications of MMMs…………...……………227
APPENDIX A SYNTHESIS AND CHARACTERIZATION OF ZEOLITE
BETA (CHAPTER 3)…………………………………………….228
A.1 Synthesis of Zeolite Beta…………………………….…………..………….228
A.1.1 Synthesis of Zeolite Beta Particles…………….…………………….228
A.1.2 Template Removal from Freshly Prepared Zeolites……..….………229
xi
A.2 Characterization of Zeolite Beta…………………………………………….231
A.3 References………………………………………….………………………..233
APPENDIX B A NEW TESTING SYSTEM TO DETERMINE THE O2/N2
MIXED GAS PERMEATION THROUGH HOLLOW FIBER
MEMBRANES WITH AN OXYGEN ANALYZER (CHAPTER
3)………………………………………………………………..…234
B.1 Introduction………………………………………….…………..…………..234
B.2 System Design and Measurement Procedure………………………………..236
B.3 Results and Discussion…………………………..………….……………….244
B.4 Conclusions………………………………………………………………….246
B.5 References……………………………………………..…………………….246
PUBLICATIONS………………………….……………………………………….250
xii
SUMMARY
Organic polymer-inorganic molecular sieve composites have received world-wide
attention during last two decades. This is due to the fact that the resultant materials
may potentially offer superior performance in terms of the permeability and
permselectivity for gas/liquid separation. The purpose of this work is to evaluate the
combined use of commercially available polyethersulfone (PES) as a matrix and
various zeolites as a dispersive phase, and to prepare the high-performance mixed
matrix membranes (MMMs) for gas separation. A comprehensive research study,
which covers the fabrication and characterization of three types of membranes,
particularly dual-layer hollow fiber neat polymeric membranes, flat dense MMMs and
dual-layer hollow fiber MMMs, is presented. Various instruments were employed to
screen the physical properties and gas separation performance of these membranes.
Emphases were put on the separation of He/N2, H2/N2, O2/N2 and CO2/CH4 gas pairs
because of their high market impact.
Firstly, the dual-layer PES hollow fiber membranes with an ultrathin dense-selective
layer of 40nm were successfully fabricated by using co-extrusion and dry-jet wet-
spinning phase inversion techniques with the aid of heat treatment at 75oC. To our best
knowledge, this is the thinnest thickness that has ever been reported for dual-layer
hollow fiber membranes. The newly developed dual-layer hollow fibers had an O2
permeance of 10.8 GPU and O2/N2 selectivity of 6.0 at 25oC after heat-treated at 75oC.
xiii
However, the heat-treatment at 150oC resulted in a significant reduction in both
permeance and selectivity because the resistance of gas transport in the non-selective
substructure was enhanced significantly. The effects of different post heat-treatments
on the membrane morphology were also studied by SEM pictures. (This work was
published in the J. Membr. Sci., 245 (2004) 53).
Secondly, the effects of membrane preparation methodology, zeolite loading and pore
size of zeolite on the gas separation performance of PES-zeolite MMMs were studied.
SEM and DSC were performed to characterize the morphology of MMMs and the Tg
change of MMMs with zeolite loading, respectively. The experimental data indicated
that a higher zeolite loading resulted in a decrease in gas permeability and an increase
in gas pair selectivity. The unmodified Maxwell model failed to correctly predict the
permeability decrease induced by polymer chain rigidification and the partial pore
blockage of zeolites. A new modified Maxwell model was therefore proposed. This
new model showed much consistent performance predication with experimental data.
(This work was published in the J. Membr. Sci., 260 (2005) 45).
Thirdly, a novel silane coupling agent, (3-aminopropyl)-diethoxymethyl silane
(APDEMS) was used to modify zeolite surface. Elementary analysis, XPS spectra and
BET measurement were applied to characterize the silane chemical modification. Both
permeability and selectivity of MMMs made from APDEMS modified zeolite were
higher than those of MMMs made from unmodified zeolite because of a decrease in
xiv
the degree of partial pore blockage of zeolites. The permeability with increasing
zeolite content showed the different change trend for MMMs made from zeolite 4A
and 5A. The permeability and selectivity predictions by the modified Maxwell model
showed very good agreement with experimental data, indicating that the modified
Maxwell model is fully capable of predicting the gas separation performance of
MMMs made from both unmodified and modified zeolite. (This work was published
in the J. Membr. Sci., in press. A US Provisional Patent Application has been filed on
29 Jun 2005).
Lastly, the dual-layer PES/P84 hollow fibers with a PES-zeolite Beta mixed matrix
dense-selective layer of 0.55µm have been successfully fabricated by adjusting the
ratio of outer layer flow rate to inner layer flow rate during the spinning with the aid of
heat-treatment at 235oC and two-step coating. To our best knowledge, this is the
thinnest thickness that has ever been reported for dual-layer hollow fibers with the
mixed matrix outer layer. SEM and DSC were used to characterize the morphology
and Tg of dual-layer hollow fiber MMMs, respectively. These newly developed dual-
layer hollow fibers exhibited an enhanced O2/N2 and CO2/CH4 selectivity of around
10~20% compared with that of neat PES dense films. Their performance has also been
confirmed in mixed gas tests, and showed comparable permeance and selectivity of
O2/N2 and CO2/CH4 in both pure and mixed gas tests. (This work was published in the
(1) J. Membr. Sci., in press; and (2) Ind. Eng. Chem. Res., 45 (2006) 871).
xv
NOMENCLATURE
A Effective area of the membrane available for gas transport (cm2)
b Langmuir affinity constant (1/atm)
C Local penetrant concentration in the membrane (cm3 (STP)/cm3 (polymer))
CD Penetrant concentration in Henry’s sites (cm3 (STP)/cm3 (polymer))
CH Penetrant concentration in Langmuir sites (cm3 (STP)/cm3 (polymer))
CH’ Langmuir capacity constant (cm3 (STP)/cm3 (polymer))
D Outer diameter of the testing fibers (cm),
D Diffusion coefficient (cm2/s)
Davg Average diffusion coefficient (cm2/s)
DAK Diffusion coefficient in the interface voids of MMMs (cm2/s)
DD Henry’s diffusion coefficient (cm2/s)
DH Langmuir diffusion coefficient (cm2/s)
Dk Knudsen diffusivity (m2/s)
dp/dt Change of pressure with time in the downstream chamber of the
permeation cell (mmHg/s)
d Diameter of gas molecules (Å)
d Average d-space (Å)
dg Diameter of gas molecules (Å)
EP Activation energy of permeation (KJ/mol)
F DH/DD
xvi
∆GM Gibbs free energy of mixing
J Permeation flux (cm3/cm2-s)
K c’H b/kD
K Constant of proportionality of power law
L Thickness of a membrane selective layer (cm)
l Thickness of a membrane selective layer (cm)
l Effective length of the modules (cm),
lI Thickness of the interface voids of mixed matrix membranes (Å)
lφ Thickness of rigidified polymer region in mixed matrix membranes (Å)
lφ’ Thickness of zeolite skin with the reduced permeability in mixed matrix
membranes (Å)
M Molecular weight (g/mole)
MA Gas molecular weight (g/mole)
Mw Gas molecular weight (g/mole)
NA Steady state flux of the permeating gas at standard temperature and
pressure (cm3 (STP)/s)
NA Avogadro’s constant
n Number of moles
n Molecular concentration
n Number of fibers in one testing module
n Shape factor of the dispersed (sieve) phase
n Power law exponent
xvii
P Permeability coefficient of a membrane to gas (1 Barrer = 1 x 10-10
cm3(STP)-cm/s-cm2-cmHg)
P0 Pre-exponential factor for the activation energies of permeation (1 x 10-10
cm3(STP)-cm/s-cm2-cmHg)
Peff Effective permeability of a gas penetrant in a mixed matrix membrane (1
Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
Pc Gas penetrant permeability in continuous phase in mixed matrix
membranes (1 Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
Pd Gas penetrant permeability in dispersed (sieve) phase in mixed matrix
membranes (1 Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
PI Gas penetrant permeability in interface voids of mixed matrix membranes
(1 Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
P3rd Composite permeability of the third phase in mixed matrix membranes (1
Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
P2nd Composite permeability of the second phase in mixed matrix membranes
(1 Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
PMMM Composite permeability of a three-phase mixed matrix membrane (1
Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
Pblo Permeability of the zeolite skin affected by the partial pore blockage in
mixed matrix membranes (1 Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
Prig Permeability of the rigidified polymer region in mixed matrix membranes
(1 Barrer = 1 x 10-10 cm3(STP)-cm/s-cm2-cmHg)
xviii
p Atmospheric pressure (atm)
p0 Feed pressure of the penetrant (psi)
∆p Pressure different between the upstream and the downstream of a
membrane (psi)
∆pA Partial pressure different between the upstream and the downstream of a
membrane (psi)
P/L Permeance of a membrane to gas (1 GPU = 1 x 10-6 cm3(STP)/s-cm2-
cmHg)
Q Volumetric flow rate of pure gas (cm3/s),
R Universal gas constant
R Gas transport resistance through the membrane (=l/PA)
r Effective pore radius (Å)
r Average pore radius (Å)
r The radial distance
r Thickness of rigidified polymer region in mixed matrix membranes (Å)
r’ Thickness of zeolite skin with the reduced permeability in mixed matrix
membranes (Å)
rd Radius of the dispersed particles (µm)
T Absolute temperature (K)
Tg Glass transition temperature (oC)
t Permeation time (s)
V Volume of the downstream chamber in permeation cell (cm3)
xix
v Average molecular velocity (m/s)
v2 Volume fraction of the third phase in the second phase
v3 Volume fraction of the bulk of zeolite in the third phase
vblo Volume fraction of the zeolite skin with the reduced permeability in the
total mixed matrix membrane
vD Volume fraction of the bulk of zeolite in the total mixed matrix membrane
vrig Volume fraction of the rigidified polymer region in the total mixed matrix
membrane
vz Axial velocity
z The axial distance
αA/B Ideal selectivity of component A over B
αD i,j Ideal selectivity of a gas pair for diffusivity
α* i,j Ideal selectivity of a gas pair for permeability
αS i,j Ideal selectivity of a gas pair for solubility
β Polymer chain immobilization factor in mixed matrix membranes
β’ Permeability reduction factor induced by partial pore blockage of zeolite in
mixed matrix membranes
χi Binary interaction parameters
φ Volume fraction
dφ Volume fraction of dispersed (sieve) phase
iφ Volume fraction of interface voids
xx
sφ Volume fraction of sieve phase in the combined sieve and interface voids
phase.
φ Draw ratio
γ& Shear rate (s-1)
λ Adsorption wavelength (Å)
λ Mean free of a gas (Å)
ρ Density (g/cm3)
ρmembrane Density of a membrane (g/cm3)
ρliquid Density of a solvent (g/cm3)
∆µi Chemical potential of species “i” relative to its reference state
vi Molar volume of species “i”.
θ Time lag (s)
θ X-ray diffraction angle of the peak (o)
τ Shear stress (Nm-2)
Abbreviations
APDEMS (3-aminopropyl)-diethoxymethyl silane
ATR Attenuated total reflection
BET Brunauer-Emmett-Teller
DL Dual-layer hollow fiber
DLS Dynamic light scattering
xxi
DSC Differential scanning calorimetry
EDX Energy dispersion x-ray
FESEM Field-emission scanning electron microscopy
FTIR Fourier transformed infrared spectroscopy
ID Inner diameter of the hollow fiber
IL Inner layer of the dual-layer hollow fiber
MMMs Mixed matrix membranes
NMP N-methyl-pyrolidinone
OD Outer diameter of the hollow fiber
OL Outer layer of the dual-layer hollow fiber
P84 BTDA-TDI/MDI co-polyimide
PES Polyethersulfone
SEM Scanning electron microscopy
SL Single-layer hollow fiber
Tg Glass transition temperature
XPS X-ray photoelectron spectroscopy
XRD X-ray diffraction
xxii
LIST OF TABLES
Table 1.1 Scientific developments of membrane gas transport…...……………..5
Table 1.2 Industrial applications of gas separation membranes.………...…..…..8
Table 1.3 Size of materials retained, driving force and type of membrane used
for each separation process…...……………………………………..15
Table 1.4 Examples of membrane applications and alternative separation
processes……………………………………………………………..16
Table 1.5 Qualitative comparison of various membrane modules…………..…37
Table 2.1 Materials for gas separation membranes……………..……………...49
Table 2.2 Properties of major synthetic zeolites…………………………….....56
Table 2.3 Comparison of the modified Maxwell Model for Cases 2, 3, and 4...72
Table 3.1 Chemical structures and properties of PES and P84……………..…83
Table 3.2 Main characteristics of zeolites……………………………..………84
Table 4.1 Spinning parameters of dual-layer PES hollow fiber membranes....119
Table 4.2 Gas separation performances of dual-layer PES hollow fiber
membranes with different post-treatment protocols………..……....120
Table 5.1 Change of O2 and N2 permeability in different regions of PES-zeolite
4A mixed matrix membranes.……………………………………...144
Table 5.2 Change of glass transition temperatures of mixed matrix membranes
over pure PES dense film……………………..……………………145
Table 5.3 Calculated volume fraction data of dispersed phase in different phases
xxiii
of PES-zeolite 4A MMMS in the new modified Maxwell model which
simultaneously considers both polymer chain rigidification and partial
pore blockage of zeolites………………..…………………………150
Table 6.1 Comparison of elementary analysis data of zeolites before and after
the chemical modification………………………………………….166
Table 6.2 Comparison of total pore volume and multipoint BET surface area of
zeolites before and after the chemical modification……….………168
Table 6.3 Change of glass transition temperatures of MMMs over neat PES
dense film before and after the chemical modification of zeolite
surface…………………………………………………..……….....172
Table 6.4 Change of O2 and N2 permeability in different regions of the PES-
zeolite 4A-NH2 MMM.……………………………………..……...184
Table 6.5 Calculated volume fraction data of dispersed phase in different phases
of PES-zeolite 4A-NH2 MMMs in the modified Maxwell model which
simultaneously considers both polymer chain rigidification and partial
pore blockage of zeolites…………….……………………………..184
Table 7.1 Intrinsic permeation properties of flat dense neat PES membrane and
PES-zeolite Beta MMM.…………………………………….……..196
Table 7.2 Spinning conditions and parameters of dual-layer PES/P84 hollow
fiber membranes with a mixed matrix outer layer.…………...…....197
Table 7.3 The effects of heat-treatment temperature on the gas separation
performance of DL2A dual-layer PES/P84 hollow fiber membranes
xxiv
with a mixed matrix outer layer………………………………..…..200
Table 7.4 Gas separation performance of dual-layer PES/P84 hollow fiber
membranes with different mixed matrix outer layer thicknesses after
heat-treated at 235oC.........................................................................203
Table 7.5 Change of glass transition temperatures of dual-layer hollow fiber
membranes with the mixed matrix outer layer over neat PES dense
film…………………………………………………………………205
Table 7.6 Effects of air gap on the gas separation performance of dual-layer
PES/P84 hollow fiber membranes with a mixed matrix outer layer
after heat-treated at 235oC………………….……………………....207
Table 7.7 Comparison of pure gas and mixed gas separation performance of
dual-layer PES/P84 hollow fiber membranes with a mixed matrix
outer layer…………….…………………………………………….208
Table B.1 Module specifications and experimental conditions (24°C)……….239
Table B.2 Comparison of separation performance of hollow fiber membranes
between the O2/N2 mixed gas and pure gas measurements………...245
Table B.3 Comparison of separation performance of hollow fiber membranes
between the CO2/CH4 mixed gas and pure gas measurements…….246
xxv
LIST OF FIGURES
Figure 1.1 Schematic diagram of pressure swing adsorption.……….….….......2
Figure 1.2 Milestones on the development of membrane gas separations..…....7
Figure 1.3 Schematic diagram of gas separation process by a membrane……14
Figure 1.4 Schematic diagram of (left) dead-end filtration and (right) cross-
flow filtration.………………………………………………….......17
Figure 1.5 Typical membrane morphology………………………………...…21
Figure 1.6 Schematic presentation of main mechanisms of membrane-based gas
separation……...…………………….....……………………….…24
Figure 1.7 Schematic drawing of a plate-and frame module………..………..34
Figure 1.8 Spiral-wound elements and assembly.………………….……..…..35
Figure 1.9 Hollow fiber separator assembly…………………………………..36
Figure 2.1 Schematic diagram of the specific volume of polymer as a function
of temperature.…………………………………………………….51
Figure 2.2 Trade-off line curve of oxygen permeability and oxygen/nitrogen
selectivity………………………………………………………….53
Figure 2.3 The structure of turbostratic carbon.….….…………….……...…..57
Figure 2.4 The SEM picture and sketch map of gas diffusion pathway in
MMMs……………………………………………………………..60
Figure 2.5 Interaction model between zeolite, TAP and polyimide..…….…...61
Figure 2.6 Schematic diagram of chemical modifications of zeolite surface
xxvi
using a silane coupling agent…………………………….………..63
Figure 2.7 Schematic diagram of various nanoscale interface morphology of
MMMs……………………………………………………………..68
Figure 3.1 Chemical structure of APDEMS………………………….……….84
Figure 3.2 Schematic diagram of the lab-scale hollow fiber spinning line and
dual-layer spinneret design………………………………….……..86
Figure 3.3 Flowchart of the chemical modification of zeolite surface……......90
Figure 3.4 Flowchart of the preparation methodology of flat PES-zeolite A
mixed matrix membranes……………………….…………………92
Figure 3.5 Pure gas permeation testing apparatus for the neat polymeric hollow
fibers…………………………………………………………….....99
Figure 3.6 Schematic diagram of the flat dense membrane gas permeation
testing apparatus………………………………………………….102
Figure 3.7 Schematic diagram of the double O-ring permeation cell for the flat
dense membranes…………………………….………………..…103
Figure 3.8 Schematic diagram of gas permeation testing apparatus for the
mixed matrix hollow fibers………………………………………106
Figure 3.9 Apparatus of the mixed gas permeation test in neat polymeric
hollow fibers……………………………………………………...108
Figure 3.10 Schematic diagram of mixed gas permeation test apparatus for flat
dense membranes……………………………………….…..……111
Figure 4.1 Integrity of as-spun dual-layer hollow fiber membranes (A: Overall
xxvii
profile; B: Membrane wall; C: Outer layer’s outer edge; D: Inner
layer’s inner bulk; E: Inner layer’s inner skin; F: Outer layer’s outer
skin)…………………………………………..…………………..123
Figure 4.2 Visual estimation of the dense-selective layer thickness of dual-layer
hollow fiber membranes with different heat-treatment methods.
(left: Condition A; right: Condition B)………………….........….125
Figure 4.3 A comparison of the cross-section morphology of dual-layer hollow
fiber membranes with different post-treatment protocols (A, D: as-
spun; B, E: Condition A; C, F: Condition B)…………………….126
Figure 4.4 Pore size comparison of the inner layer’s inner skin of dual-layer
hollow fiber membranes with different post-treatment protocols
(left: as-spun; middle: Condition A; right: Condition B)………...127
Figure 5.1 Comparison of gas permeability of PES-zeolite A mixed matrix
membranes with different cooling protocols (A: H2 permeability; B:
O2 permeability; C: N2 permeability)…………………………….139
Figure 5.2 Comparison of gas pair selectivity of PES-zeolite A mixed matrix
membranes with different cooling protocols (left: H2/N2 selectivity;
right: O2/N2 selectivity)………………………………………......140
Figure 5.3 Comparison of cross-section SEM images of mixed matrix
membranes (A, B, C: PES-zeolite 3A, 4A and 5A MMMs with
immediate quenching, respectively; D, E, F: PES-zeolite 3A, 4A and
5A MMMs with natural cooling, respectively)……….………….142
xxviii
Figure 5.4 Effect of zeolite loadings on H2, O2 and N2 permeability of PES-
zeolite 4A and PES-zeolite 5A mixed matrix membranes (left: H2
permeability; right: O2 and N2 permeability)………………….....143
Figure 5.5 Comparison of O2 permeability of PES-zeolite 4A MMMs between
experimental data and predictions from the Maxwell model and
various modified Maxwell models……………………………….144
Figure 5.6 Schematic diagram for the new modified Maxwell model.….…..147
Figure 5.7 Effect of zeolite loadings on H2/N2 and O2/N2 selectivity of PES-
zeolite 4A and PES-zeolite 5A mixed matrix membranes (left: H2/N2
selectivity; right: O2/N2 selectivity)…………………....……..….151
Figure 5.8 Comparison of O2/N2 selectivity of PES-zeolite 4A MMMs between
experimental data and predictions from the Maxwell model and the
new modified Maxwell model……………………..……………..152
Figure 5.9 Effect of different pore sizes of zeolite on H2 and O2 permeability of
mixed matrix membranes (left: H2 permeability; right: O2
permeability)…………………………..…………………………154
Figure 5.10 Effect of different pore sizes of zeolite on H2/N2 and O2/N2
selectivity of mixed matrix membranes (left: H2/N2 selectivity; right:
O2/N2 selectivity)…………………………………………………155
Figure 6.1 Comparison of XPS spectra of the zeolite before and after the
chemical modification (A: zeolite 3A; B: zeolite 3A-NH2; C: zeolite
4A; D: zeolite 4A-NH2)………………………………….…...….167
xxix
Figure 6.2 Comparison of gas permeability of PES-zeolite A MMMs before
and after the chemical modification of zeolite surface (A: He
permeability; B: H2 permeability; C: O2 permeability; D: CO2
permeability)……………………………………………………..169
Figure 6.3 Comparison of gas pair selectivity of PES-zeolite A MMMs before
and after the chemical modification of zeolite surface (A: He/N2
selectivity; B: H2/N2 selectivity; C: O2/N2 selectivity; D: CO2/CH4
selectivity)……………………………………………………......170
Figure 6.4 Comparison of cross-section SEM images of MMMs before and
after the chemical modification of zeolite surface (A, B, C: PES-
zeolite 3A, 4A and 5A MMMs, respectively; D, E, F: PES-zeolite
3A-NH2, 4A-NH2 and 5A-NH2 MMMs, respectively)…..………171
Figure 6.5 Effect of zeolite loadings on gas permeability of PES-zeolite 4A-
NH2 and PES-zeolite 5A-NH2 MMMs (A: He and H2 permeability;
B: CO2 permeability; C: O2 permeability; D: N2 and CH4
permeability)……….………………………………………….....174
Figure 6.6 Effect of zeolite loadings on gas pair selectivity of PES-zeolite 4A-
NH2 and PES-zeolite 5A-NH2 MMMs (A: He/N2 selectivity; B:
H2/N2 selectivity; C: O2/N2 selectivity; D: CO2/CH4 selectivity)..175
Figure 6.7 Schematic diagram for the modified Maxwell model.….………..179
Figure 6.8 Comparison of O2 permeability of PES-zeolite 4A-NH2 MMMs
between experimental data and predictions from the Maxwell model
xxx
and the modified Maxwell model…...……………………………185
Figure 6.9 Comparison of O2/N2 selectivity of PES-zeolite 4A-NH2 MMMs
between experimental data and predictions from the Maxwell model
and the modified Maxwell model.……………………..……...….186
Figure 7.1 Comparison of SEM morphologies of dual-layer PES/P84 hollow
fiber membranes with a mixed matrix outer layer for DL2A before
and after the heat-treatment (A, B and C: overall profile, outer
layer’s outer edge and outer layer’ outer surface of as-spun hollow
fibers, respectively; D, E and F: overall profile, outer layer’s outer
edge and outer layer’ outer surface of hollow fibers heat-treated at
235oC, respectively)……………….……………………………..199
Figure 7.2 Comparison of partial cross-section SEM images of dual-layer
PES/P84 hollow fiber membranes with different mixed matrix outer
layer thicknesses after heat-treated at 235oC (A and D (enlarged):
DL3A; B and E (enlarged): DL3B; C and F (enlarged): DL3C)...202
Figure 7.3 Comparison of glass transition temperatures of dual-layer hollow
fiber membranes with a mixed matrix outer layer with neat PES
dense film (A, B, C and D: the enlarged second heating cycle of
DSC curves of neat PES dense film, DL2A-235oC, DL3A-235oC
and DL3C-235oC, respectively)………………….………………204
Figure A.1 Synthesis procedure of zeolite-Beta……………………………...229
Figure A.2 Polymer-networking procedure for the template removal of zeolite
xxxi
Beta………………….……………………………………………231
Figure A.3 XRD characterization of the self-synthesized zeolite Beta for the
particle crystalline structure confirmation…………….………….232
Figure A.4 FESEM characterization of the self-synthesized zeolite Beta for the
particle shape observation…………………………….………….232
Figure A.5 Laser light scattering (LLS) characterization of the self-synthesized
zeolite Beta for the particle size determination…………………..233
Figure B.1 Schematic diagram of apparatus using an oxygen analyzer for O2/N2
mixed gas permeation tests through a hollow fiber module….…..238
1
CHAPTER ONE
INTRODUCTION OF GAS SEPARATION MEMBRANE
The separation of one or more gases from complex multicomponent mixture of gases
is necessary in a large number of industries. Such separations currently are undertaken
commercially by processes such as cryogenic distillation, pressure swing adsorption
(PSA) and membrane separation.
The cryogenic air separation process is routinely used in large or medium scale plants
to produce nitrogen, oxygen, and argon as gases and/ or liquid products. The
cryogenic air separation is the most cost effective technology for larger plants and for
producing very high purity oxygen and nitrogen (99.999%). It is the only technology
that will produce liquid products. The energy required to operate cryogenic plants
depends on the product mix and required product purities. Gas-producing plants use
less power than those producing some or the entire product as liquid. More than twice
as much power is required to produce a unit of product in liquid form than as a gas.
The gas separation process of pressure swing adsorption is to make air pass through a
column packed with a bed of pellets or powder with a large surface area per weight.
When air is passed over this bed, air molecules will adsorb (stick to the surface) to the
pellets. Using the right pellets, all the oxygen from the air stream can be removed
2
leaving only nitrogen and traces of argon. Using different pellets, all the nitrogen
could be removed from the air stream, leaving mostly oxygen. The schematic diagram
of PSA is shown in Figure 1.1.
Inlet Flow Product Flow
Figure 1.1: Schematic diagram of pressure swing adsorption
Today, a large scale membrane gas separation system has found acceptance in many
industrial sectors. Membrane technology compares favorably with other conventional
separation techniques due to its multidisciplinary character, which is often faster, more
capital and energy efficient. The specific features and inherent advantages of
membrane separation process can be recapitulated as follows:
1. Simplicity of operation and installation;
2. Lower capital outlay and large reduction in power (electricity and fuel, etc.)
consumption. No additional utilities/additives are required for membrane
systems unless a compressor is needed;
3. Economic viability even at high system-capacity. Membrane processes are
flexible, where the modules can be simply arranged in stages to accommodate
3
higher capacity and scaled to small sizes;
4. Membrane devices and systems are always compact in size and modulus, which
generally are space and weight efficient;
5. Membrane processes can be carried out under mild conditions, for example, air
separation able to be operated at atmosphere pressure and room temperature
instead of a cryogenic condition in distillation of air;
6. Membrane separation can be carried out continuously;
7. Membranes can be “tailor-made” to a certain extent, thus their separation
properties are viable and can be adjusted to a specific separation task;
8. Membrane processes can easily combined with other separation processes for
effective hybrid processing.
1.1 SCIENTIFIC MILESTONES
The first scientific observation associated with gas separation was laid down by J.K.
Mitchell in 1831 [1, 2]. However, the most remarkable contribution was made by
Thomas Graham, a Scottish chemist, who laid down the foundation of diffusion of
gases and liquids. Graham’s law [3] states that the diffusion rate of a gas is inversely
proportional to the square root of its density. He found that certain substances (e.g.
glue, gelatin and starch) pass through a membrane more slowly than others (e.g.
inorganic salts), leading to the distinction between two types of particles: colloids for
the former, and crystalloids for the latter. In this connection he discovered dialysis. At
4
approximately the same time, A. Fick, an outstanding physiologist postulated the
concept of diffusion and formulated the well-known Fick’s first law by studying the
gas transport through nitrocellulose membranes [4].
However, many significant scientific observations about membrane separation, such as
the first quantitative measurement of the rate of gas permeation were accomplished by
Sir Thomas Graham, the discoverer of Graham’s law of gas effusion. He proposed
“solution-diffusion” mechanism for gas permeation through a membrane by repeating
Mitchell’s experiments with the films of natural rubber in 1866 [5]. Approximately 13
years later in 1879, Von Wroblewski quantified Graham’s model and defined the
permeability coefficient as the permeation flux multiplied by the membrane thickness
divided by the transmembrane pressure [6]. He also characterized the permeability
coefficient as a product of diffusivity and solubility coefficients, which soon became
an important model in membrane permeation. A decade later in 1891, H. Kayser
demonstrated the validity of Henry’s law for the absorption of carbon dioxide in
rubber [7].
The progress of membrane separation techniques was very slow in the early stage.
Nevertheless, many fundamental scientific works and contributions related to gas
separation membranes were carried out in the twentieth century, as summarized in
Table 1.1 [8]. Particularly, H.A. Daynes developed the time lag method from
5
nonsteady-state transport behavior of gases via a membrane to determine diffusion
coefficient [9].
Table 1.1: Scientific developments of membrane gas transport [8]
Scientist (Year) Event
Graham (1829) First recorded observation
Mitchell (1931) Gas permeation through natural rubbers
Fick (1855) Law of mass diffusion
von Wroblewski (1879) Permeability coefficient product of diffusion and absorption coefficient
Kayser (1891) Demonstration of validity of Henry’s Law for the absorption of carbon dioxide in rubber
Lord Rayleigh (1900) Determination of relative permeabilities of oxygen, nitrogen and argon in rubber
Knudsen (1908) Knudsen diffusion defined
Shakepear (1917-1920) Temperature dependence of gas permeability independent of partial pressure difference across membranes
Daynes (1920) Developed time lag method to determine diffusion and solubility coefficient
Barrer (1939-1943) Permeabilities and diffusivities followed Arrhenius equation
Matthes (1944) Combined Langmuir and Henry’s law sorption for water in cellulose
Meares (1954) Observed break in Arrhenius plots at glass transition temperature and speculated about two modes of solution in glassy polymers
6
Barrer, Barrie and Slater (1958)
Independently arrived at dual mode concept from sorption of hydrocarbon vapors in glassy ethyl cellulose
Michaels, Vieth and Barrie (1963)
Demonstrated and quantified dual mode sorption concept
Vieth and Sladek (1965) Model for diffusion in glassy polymers
Paul (1969) Effect of dual mode sorption on time lag and permeability
Petropoulos (1970) Proposed partial immobilization of sorption
Paul and Koros (1976) Defined effect of partial immobilizing sorption on permeability and diffusion time lag
The above fundamental works provide the foundation in membrane processes, which
conduce to the commercialization of membrane separation technology in industrial
applications. Following the first breakthrough of asymmetric phase-inverted
membranes made of cellulose acetate for reverse osmosis by Loeb and Sourirajan in
1960 [10-12], membrane gas separation appeared to be a competitive separation tool
for industry processes in the 1970’s. The first commercially viable gas separation
membrane, Prism® was produced subsequent upon the method of repairing pinhole
size defects in the thin layer of asymmetric membranes by Henis and Tripodi [13]. As
a consequence, the successful application of the first commercial gas separation
membrane has accelerated the development of novel membrane materials as it offer an
attractive alternative for specific separation applications. Figure 1.2 displays the
important milestones in the history and scientific development of membrane gas
separation technology [14].
7
Figure 1.2: Milestones on the development of membrane gas separations [14]
1.2 IMPORTANCE OF GAS SEPARATIONS USING MEMBRANES
Membrane gas separation process becomes an emerging technology on industrial scale
in the late seventies when Prism® was introduced in 1978. However, the utilization of
membrane technology in gas separation has rapidly expanded and observed the broad
8
usage/interest in industrial application. Membrane gas separation impacts the
separation business with US$250 million a year. The multitude applications of gas
separation membranes are listed in Table 1.2. The major applications for gas
separation membranes are discussed.
Table 1.2: Industrial applications of gas separation membranes [15, 16]
Gas separation Application
O2/N2 Oxygen enrichment, nitrogen (Inert gas) generation
H2/Hydrocarbons Refinery hydrogen recovery
H2/CO Syngas ratio adjustment
H2/N2 Ammonia purge gas
CO2/Hydrocarbons Acid gas treatment enhanced oil recovery, landfill gas upgrading
H2S/Hydrocarbons Sour gas treating
H2O/Hydrocarbons Natural gas dehydration
He/Hydrocarbons Helium separations
H2O/Air Air dehydration
He/N2 Helium recovery
Hydrocarbons/Air Pollution control, hydrocarbon recovery
Hydrocarbons from process streams
Organic solvent recovery, monomer recovery
1.2.1 Separation of O2 and N2
9
Separation of O2 and N2 from the essentially free air has wide applications. About
60% of separation business involves air-either producing nitrogen or oxygen [17],
where O2 and N2 are the third and fifth largest bulk chemicals produced worldwide.
Mainly N2 is produced by membranes as an inert gas blanketing for storing and
shipping of flammable liquids, fresh fruits and vegetables. Membranes are most
competitive in this application when the volume and the acceptable purity level are
both low. The market share of membrane technology in producing nitrogen is growing
and currently producing 30% of total nitrogen [18]. Furthermore, the nitrogen
produced is already pressurized. For those applications where high pressure gas is
needed or convenient, air separation by membrane has additional values.
As of now, there is limited utility of membranes for commercial-scale production of
oxygen and it is primarily for medical purposes. No large volume market exists today
to support high volume production although enhanced combustion has the potential.
Air is one of the major feedstock for coal processing. Oxygen-enriched air improves
combustion efficiency of coal and is also a desirable feedstock for gasifiers. Oxygen
currently produced by organic membranes has a purity limit of 50% [19]. Ideally, the
new membrane materials with desire permeability (~250 Barrer) and the oxygen
separation factor of 4-6 are needed to increase the practicability of membrane
technology for industrial oxygen separation [18].
10
1.2.2 Separation of H2 and Hydrocarbon Gases
Recovery of H2 by membranes is important in the refining industry [20]. Associated
with the use of hydrogen at refinery plants, light hydrocarbon gases are produced
which must be purged from the reaction system. The hydrogen value in the purge gas
can be used as fuel in the past. However, there are two major incentives to recover
hydrogen for refinery process instead. One is that hydrogen in the purge stream, if
recovered, can be as much as 20% of the total hydrogen generated at the refinery
plants. The other factor is that the cost of hydrogen manufacture is increasing for
various refining processes such as coking, hydrocracking and hydrodesulphurization.
In these applications hydrogen is separated from the purge gas containing an
appreciable amount of hydrogen as well as hydrocarbon gases. The recovered
hydrogen is returned to the plant to supplement hydrogen made by the extensive direct
production route to improve the quality of fuels or the feed quality for various
processes. Thus membrane separation in this case improves process efficiency.
Membrane processing competes with other separation processes like pressure swing
adsorption, cryogenics and oil scrubbing for these applications. Separation of
hydrogen from methane has been identified as one of the promising applications of
inorganic membrane technology for the refinery industry in the future.
1.2.3 Separation of H2 and CO
11
Produced from natural gas, oil or coal using different processes, synthesis gas contains
a mixture of mostly hydrogen and carbon monoxide plus low percentages of carbon
dioxide and nitrogen. With transition metal catalysis, it is used in a number of
reactions for making a wide variety of organic and inorganic compounds such as
alcohols, aldehydes, acrylic acids and ammonia. The stoichiometry of the feed gas for
these reactions must be adjusted according to the process requirements. The ratios of
hydrogen to carbon monoxide and other gases in the feed streams in those chemical
production processes can be varied by selectively removing hydrogen as a permeate
through a membrane. In this case carbon monoxide comes out of the membrane
separation process at high pressure which is an advantage over other separation
processes such as cryogenics and pressure swing adsorption. Another important
application is the removal of carbon monoxide from the valuable hydrogen produced
in steam reformed natural gas.
1.2.4 Separation of H2 and N2
Ammonia is typically made from synthesis gas at a temperature of about 500oC and a
pressure of about 300bar. Incomplete conversion to ammonia results in a need to
recycle unreacted gases back to the reactor. Some gases inert to the reaction such as
argon and unreacted methane build up and require a continuous gas purge to manage
these contaminants at an acceptable level. The purge gas contains valuable hydrogen
12
and nitrogen as the major components. To reuse them after purge, H2 and N2 need to
be separated from the purge gas, which can be achieved cost-effectively by organic
membranes in comparison to other traditional separation processes. In this application
hydrogen is separated as a permeate at the lower pressure side of the membrane.
Consequently it must be recompressed to the pressure of the reactor before reuse.
1.2.5 Acid Gas Removal from Natural Gas
The total worldwide market for new natural gas separation equipment is estimated at
approximately US$ 5 billion/year [14]. The natural gas treatment process is dominated
by amine absorption. In recent year, membrane process has been accepted as a
promising technology for natural gas separation [8]. The high efficiency is observed
by membrane separation in upgrading natural gas by the removal of CO2, water and
acid gases (H2S, etc.). These impurities must be removed from natural gas stream
before delivery to a pipeline aimed to reduce the toxicity and avoid the pipeline
corrosion. Again, polymeric membranes have become the promising materials for
natural gas separation [8]. However, glassy polymeric membranes suffering from
plasticization in the presence of trace quantities of condensable heavy hydrocarbons,
which cause the loss of selectivity [21]. Hence, the development of robust membrane
materials is imperative to widespread the commercialization of membrane separations
in natural gas treatment.
13
On top of above-mentioned 3 applications, the dehydration of natural gas/air and the
olefin/paraffin separation appear to be the emerging technology and future opportunity
for membrane gas separation [19]. Membrane process competes with the glycol
absorption process for natural gas dehydration. Membranes offer the advantages over
glycol process by being more environmental friendly (loss of glycol due to
contamination with aromatic hydrocarbons by glycol absorption) and low energy
consumption. On the other hand, the separation of paraffin and olefin gases such as
ethane/ethylene and propane/propylene is one of the important tasks in petroleum
refining and petrochemical industries. Unsaturated hydrocarbons, specifically ethylene
and propylene are important feedstock in the production of petrochemical products
such as polyethylene, polypropylene, copolymer ethylene/propylene, acrylonitrile and
cumene [22, 23]. Low temperature distillation is widely performed for olefin/paraffin
separations. The major drawback of distillation is extremely energy intensive. As a
consequence, the membrane separation with simple operation, low power
consumption and more economical exhibits the high potential for the large scale
application in olefin/paraffin separations.
Besides all the above, gas separation technology is mostly used in aviation. Its
primary purpose is to remove fumes from fuel tanks that could explode due to any
electrical discharge. An electrical discharge could be caused by a short circuit in a
wire, a strong electromagnetic wave from lightning, or interference produced by other
electrical devices. As liquid fuel is used, the fumes from that fuel are left in the
14
tank. These fumes are much more likely to catch fire or explode than the liquid fuel,
so they need to be removed.
1.3 BASIC CONCEPT OF MEMBRANE SEPARATION
The definition of a membrane can be stated as a semipermeable active or passive
barrier which, under a certain driving force, permits preferential passage of one or
more selected species or components (molecules, particles or polymers) in a gaseous
and/or liquid mixture solution [24, 25].
Membrane
Feed Permeate
Driving Forces: ∆C, ∆p etc.
Figure 1.3: Schematic diagram of gas separation process by a membrane
The primary species rejected by the membrane is called retentate or sometimes just
“solute”, while those species passing through the membrane is usually termed
“permeate” or sometimes “solvent” shown in Figure 1.3. The driving force can exist in
the form of pressure, concentration, or voltage difference across the membrane.
15
Depending on the driving force and the physical sizes of the separated species,
membrane processes are classified accordingly: microfiltration (MF), ultrafiltration
(UF), nanofiltration (NF), reverse osmosis (RO), dialysis, electrodialysis (ED),
pervaporation (PV), and gas separation. Table 1.3 shows the size of materials retained
for each process, the driving force behind separation and the type of membrane, while
Table 1.4 summarizes some of the applications of each process and their alternatives.
Table 1.3: Size of materials retained, driving force and type of membrane used for each separation process
Process Size of materials
retained Driving force
Type of
membrane
Microfiltration 0.1-10µm Pressure difference
(0.5 - 2 bar) Porous
Ultrafiltration 1-100nm Pressure difference
(1 - 10 bar) Microporous
Nanofiltration 0.5-5nm Pressure difference
(10 - 70 bar) Microporous
Reverse Osmosis <1nm Pressure difference
(10-100 bar) Nonporous
Dialysis <1nm
Concentration
difference
Nonporous or
microporous
Electrodialysis <1nm
Electrical potential
difference
Nonporous or
microporous
Pervaporation <1nm
Concentration
difference Nonporous
Gas Permeation <1nm
Partial pressure
difference (1-100 bar)Nonporous
16
Table 1.4: Examples of membrane applications and alternative separation processes
Membrane processes can be operated in two major modes according to the direction of
the feed stream relative to the orientation of the membrane surface: dead-end filtration
Process Applications Alternative
Processes
Microfiltration Separation of bacteria and cells from solutionsSedimentation,
Centrifugation
UltrafiltrationSeparation of proteins and virus,
concentration of oil-in-water emulsions Centrifugation
NanofiltrationSeparation of dye and sugar,
water softening
Distillation,
Evaporation
Reverse
Osmosis
Desalination of sea and brackish water,
process water purification
Distillation,
Evaporation,
Dialysis
Dialysis Purification of blood (artificial kidney) Reverse osmosis
Electrodialysis Separation of electrolytes from nonelectrolytesCrystallization,
Precipitation
Pervaporation Dehydration of ethanol and organic solvents Distillation
Gas
Permeation
Hydrogen recovery from process gas streams,
dehydration and separation of air
Absorption,
Adsorption,
Condensation
Membrane
Distillation Water purification and desalination Distillation
17
and cross-flow filtration (see Figure 1.4). The majority of the membrane separation
applications use the concept of cross flow where the feed flows parallel to and past the
membrane surface while the permeate penetrates through the membrane overall in a
direction normal to the membrane. The shear force exerted by the flowing feed stream
on the membrane surface help to remove any stagnant and accumulated rejected
species that may reduce permeation rate and increase the retentate concentration in the
permeate. Predominant in the conventional filtration process, dead-end filtration is
used in membrane separation only in a few cases such as laboratory batch separation.
In this mode, the flows of the feed stream and the permeate are both perpendicular to
the membrane surface.
Permeate InfluentInfluent
Permeate
Figure 1.4: Schematic diagram of (left) dead-end filtration and (right) cross-flow filtration
Two of the most important parameters that describe the separation performance of a
membrane are its permeability and permselectivity (or simply as selectivity).
Permeability is typically used to provide an indication of the capacity of a membrane
for processing the permeate; a high permeability means a high throughput. A high
throughput is useless, however, unless another membrane property, permselectivity
also exceeds an economically acceptable level. On the other hand, a membrane with a
high permselectivity but a low flux or permeability may require such a large
18
membrane surface area that it becomes economically unattractive. Simply put,
permselectivity is the ability of the membrane to separate the permeate from the
retentate.
As a general rule, the membrane technology is a competitive separation method for
small to medium volumetric flowrate applications and for either primary separation or
separation with a requirement of purity level of 95%~99%.
1.4 TYPES OF MEMBRANE STRUCTURES
1.4.1 Dense vs. Porous Membranes
Membranes can be divided into two categories according to their structural
characteristics which can have significant impacts on their performance as separators
and/or reactors: dense and porous membranes. The dense membranes are free of
discrete, well-defined pores or voids. The difference between the two types can be
conveniently detected by the presence of any pore structure under electron microscopy.
The effectiveness of a dense membrane strongly depends on its material, the species to
be separated and their interactions with the membrane.
The microstructure of a porous membrane can vary according to the method of
preparation. Those membranes that show essentially straight pores across the
membrane thickness are referred to as straight pore or nearly straight pore membranes.
19
The majorities of porous membranes, however, have interconnected pores with
tortuous paths and are called tortuous pore membranes.
1.4.2 Symmetric vs. Asymmetric Membranes
When the separating layer and the bulk support designed for mechanical strength are
indistinguishable and show an integral, homogeneous structure and composition in the
direction of the membrane thickness, it is called a symmetric or isotropic membrane.
Since the flow rate through a membrane is inversely proportional to the membrane
thickness, it is very desirable to make the homogeneous membrane layer as thin as
possible.
However, very thin stand-alone membranes typically do not exhibit mechanical
integrity to withstand the usual handling procedures and processing pressure gradients
found in many separation applications. A practical solution to the dilemma has been
the concept of an asymmetric or composite membrane where the thin separating layer
and the open-cell mechanical support structure are distinctly different [24, 26-28].
In this “anisotropic” arrangement, the separation of the species in the feed stream and
ideally the majority of the flow resistance (or pressure drop) also take place primarily
in the thin separating layer. The underlying support layer should be mechanically
20
strong and so porous that it does not contribute to the transport resistance of the
penetrants along the membrane to any significant extent.
If a membrane has a graded pore structure but is made in one processing step,
frequently from the same material across its thickness, it is called an asymmetric
membrane [26-28]. If, on the other hand, the membrane has two or more distinctively
different layers made at different steps, the resulting structure is called an asymmetric
composite membrane [26-28]. A predominantly thick layer in this type of membrane
provides the necessary mechanical strength to other layers and is called the support
layer or bulk support. Asymmetric composite membranes have the advantage that the
separating layer and the support layer(s) can be tailored with different materials.
Permselectivity and permeability properties of the membrane material are critically
important for the separating layer, while the material of the support layer(s) is chosen
based on its mechanical strength and other consideration such as chemical resistance.
Ideally, the interface between the two layers and the bulk structure of the support layer
must be porous in order to have a minimal gas transport resistance. In addition, the
two layers must be delamination-free at the interface.
The asymmetric composite membranes can have more than two layers in which the
separating layer is superimposed on more than one support layer. In this case the
intermediate layer serves as a major purpose of regulating the pressure drops across
the membrane by preventing any appreciable penetration of the very fine constituent
21
particles into the pores of the underlying support layer(s). Typically the intermediate
support layer(s) is also thin. Figure 1.5 illustrates the structural difference among
symmetric and asymmetric membranes [28].
1. Dense film membrane
2. Asymmetric membrane
3. Asymmetric composite membrane
4. Microporous composite membrane
Uniform defect-free dense structure
Defect-free selective layerAsymmetric structure
Asymmetric structureDefective selective layer
Caulking/property enhancing layer
Selective layer
Gutter layerMicroporous structure
5. Matrix composite membraneCAN BE ANY ONE OF THE ABOVE
Figure 1.5: Typical membrane morphology [28]
Another promising development on asymmetric membranes worthy of mentioning is
dual-layer hollow fibre membranes [29-37]. Compared with flat membranes, hollow
fibers is a more favored configuration due to the following advantages: 1) a large
membrane area per volume unit; 2) high gas flux (i.e. high productivity); and 3) good
flexibility and easy handling in the module fabrication. Although high-performance
22
polymeric materials have been synthesized recently with a development of material
chemistry, it is not economically attractive due to high price if these materials are
applied in the entire single-layer hollow fibers. Therefore, dual-layer hollow fiber
membranes made by a co-extrusion technology could be a potential solution. The two
layers not only can be made from different materials to save costs, they also can be
fabricated in one step to save time.
1.4.3 Dynamic in-situ Membranes
There is a different type of inorganic membranes that are formed in-situ in the
application environment. They are called dynamic membranes which have been
studied a great deal especially in the 1960s and 1970s. The general concept is to filter
a dispersion containing the suspended inorganic or polymeric colloids through a
microporous support to form a layer of the colloids on the surface of the support. This
layer becomes the active separating layer (membrane).
Over time, this permselective layer is eroded or dissolves and must be replenished as
they are washed away in the retentate. Commonly used materials for the support are
porous stainless steel, carbon or ceramics. Dynamic membranes have been studied for
reverse osmosis applications such as desalination of brackish water, but found to be
difficult to provide consistent performance and the added cost of the consumables
23
makes the process unattractive economically. Therefore only limited applications for
recovering polyvinyl alcohol in the field of textile dyeing is commercially practiced.
1.4.4 Liquid Membranes
There is another type of membrane called liquid membrane where a liquid complexing
or carrier agent supported or immobilized in a rigid solid porous structure function as
the separating transport medium [38-40]. The liquid carrier agent completely occupies
the pores of the support matrix and reacts with the permeating component on the feed
side. The complex formed diffuses across the membrane/support structure and then
releases the permeate on the product side and at the same time recovers the carrier
agent which diffuses back to the feed side. Thus permselectivity is accomplished
through the combination of complexing reactions and diffusion.
This is often known as facilitated transport which can be used for gas separation or
coupled transport which can separate metal compounds through ion transport. An
example of the former is some molten salts supported in porous ceramic substrate that
are selective toward oxygen and of the latter is some liquid ion exchange reagent for
selectively transporting copper ions. In this configuration, the composite of the liquid
membrane and its support can be considered to be a special case of dense membranes.
Despite their potential for very high selectivity, liquid membranes suffer from physical
instability and are not likely to be a major commercial force in the separation industry.
24
1.5 MECHANISMS OF MEMBRANE SEPARATION
Gas membrane separation is a straightforward process concept driven by the pressure
gradient imposed between upstream and downstream streams. Permeation is a rate-
controlled process and separation degree is determined by the selectivity of membrane
at the conditions of separation, including pressure, temperature, flow rate, and
membrane area [41]. A membrane will separate gases only if some components pass
through the membrane more rapidly than others. Depending on the pore sizes in
membrane matrix, there are four fundamental transport mechanisms in gas separation
membranes, namely (1) Poiseuille flow (2) Knudsen diffusion (3) Molecular sieving
and (4) Solution-diffusion, as illustrated in Figure 1.6 [42].
Figure 1.6: Schematic presentation of main mechanisms of membrane-based gas separation [42]
25
From Figure 1.6, it is obvious that there are two major types of membranes involved
in the gas separation. The first is the “porous” membrane in which the gases are
separated on the basis of their molecular size through the small pores in the membrane
matrix, by Poiseuille flow, Knudsen diffusion or Molecular sieving. Nevertheless, the
vast majority of commercial application is based on nonporous membranes (contain
no holes or pores in the conventional sense) through solution-diffusion mechanism.
1.5.1 Poiseuille Flow
The Poiseuille flow also known as viscous flow occurs when the mean pore diameter
is larger than the mean free path of the gas penetrants. The mean free path here refers
to the average distance traversed by a gas molecule between collisions and it is
pressure and temperature dependent. In this condition, membrane contains pores large
enough to allow convective flow, where gas molecules collide exclusively with each
other and no separation is obtained between the gas components. This type of
transport mechanism is observed for the membranes having the much larger pore sizes
than gas molecules, at pores size, dp > 10µm and the flux is proportional to r4.
1.5.2 Knudsen Diffusion
Convective flow will be replaced by Knudsen diffusion in a porous membrane, whose
pore sizes are less than the mean free path of the gas molecules [24]. Gas molecules
26
therefore interact with the pore walls much more frequently than with one another and
allow lighter molecules to preferentially diffuse through pores to achieve separation.
Knudsen diffusion principally takes place in the membranes with the pore size of 50-
100Å in diameter [42].
Knudsen diffusion occurs when the permeating species flow via the membrane almost
independent of one another. Hence, the Knudsen diffusion coefficient, Dk (m2/s) is
independent of pressure. For an equimolar feed, the permeation rate of Knudsen
diffusion is inversely proportional to the square root of the molecular weight of the
different compounds in the following equation [43]:
wk M
TrrvD 97667.0 == (1.1)
where the average pore radius is given by r (m), v is the average molecular velocity
(m/s) and T is the operating temperature. Whilst, the highest attainable separation
factor between two different gas molecules i and j equals to the square root of the ratio
of the two gas molecular weights, shown in equation (1.2).
iw
jwij M
M
=α (1.2)
Consequently, such membranes are not commercially attractive in general for standard
application due to their relative low selectivity.
1.5.3 Molecular Sieving
27
Molecular sieving separation is primarily based on the precise size discrimination
between gas molecules through ultramicropores (< 7Å in diameter). Molecular sieving
membranes become increasingly important in gas separation especially for inorganic
membranes due to their higher productivity and selectivity than solution-diffusion
polymeric membranes [44, 45]. Their porous nature has led to high permeability,
while the high selectivity is achieved through effective size and shape separation
between the gas species. This happens when the pore diameters are small enough to
allow the permeation of smaller molecules while obstructing the larger molecules to
diffuse through. Even both can enter the pores; the larger one would experience
stronger repulsive forces. This energetically biased selectivity is called “energetic
selectivity”. It is also believed that the more subtle contribution to selectivity is from
“entropic selectivity” [46], in which the rotational freedom of one component is
restricted to a higher degree than that of the other components.
Carbon molecular sieve membranes (CMSMs) and zeolites are the typically
membranes dominated by molecular sieving mechanism and give the high separation
performance. The ratio of the gas molecular size to the micropore diameter controls
the gas permeation rate and separation in molecular sieving materials [47]. For
example, zeolite 4A with a pore size of 3.8Å has an O2 and N2 selectivity of
approximately 37 at 35oC [48]. The CO2/CH4 ideal selectivity has reached as high as
around 200 in the carbon molecular sieve membranes [49]. The pore size of zeolite
28
may be controlled or modified by choosing suitable synthesis, dealumination, and ion-
exchange [50]; as for CMSMs, the pore size can be controlled by choosing polymer
precursor, pyrolysis conditions, pretreatment and post treatment [51].
1.5.4 Solution-diffusion
The last mechanism, solution-diffusion through the selective layer of a nonporous
membrane occurs in the absence of direct continuous pathways for the transportation
of gas penetrants across the membrane. This transport mechanism produces high
performance membranes, which are used exclusively in commercial separation
devices to conveniently separate wide spectrum of gas pairs. Solution-diffusion
mechanism is conceptually assumed that gas molecules from the upstream gas phase
first adsorb into the membrane, then diffuse across it and finally desorb into the
downstream gas phase side [25, 52]. This mechanism commonly found in the gas
transportation through polymeric membranes.
The permeation of molecules through membranes is controlled by two major
mechanisms. They are diffusivity (D) and solubility (S) [53]. Diffusivity is the
mobility of individual molecules passing through the holes in a membrane material,
and solubility is the number of molecules dissolved in a membrane material.
Permeability (P) is defined as an ability measurement of a membrane to permeate
molecules:
29
P = D × S (1.3)
The ability of a membrane to separate two molecules, A to B, is the ratio of their
permeability, called as the membrane selectivity αAB.
BB
AA
B
AAB SD
SDPP
==α (1.4)
1.5.4.1 Diffusion
The diffusion coefficients of common gases in polymers were recognized early as a
strong function of the effective diameter of the gas molecule. DA/DB is the ratio of the
diffusion coefficients of the two molecules, and can be interpreted as the mobility or
diffusivity selectivity, reflecting the different sizes of the two penetrating molecules.
Diffusion in rubbery polymeric materials involves the generation of a sufficiently
large gap in the polymer adjacent to the sorbed penetrant for the penetrant to move
into, with the subsequent collapse of the sorbed cage that was previously occupied by
the penetrant. Thermally induced motions of the polymer segments are responsible for
creation and destruction of these transient gaps or holes. The rate of diffusion depends
on the concentration of holes that are sufficiently large to accept diffusing molecules.
Diffusion in glassy polymeric materials is different from rubbery polymers primarily
because of the difference in the characteristic scales of the micromotions that occur at
30
segmental level for the two states. In glassy polymers the micromotions are much less
extensive than rubbery polymers and are believed to be torsional oscillations. Some
differences arise due to the presence of trapped excess free volume in glassy materials.
It has been well recognized that the diffusion coefficient is the primary factor in
determining the absolute value of gas permeability in polymers [54]. The diffusivity of
gases was shown to decrease rapidly as the collision diameter of the gas molecule
(determined from gas viscosity data) increases. According to the study by Berens and
Hopfenberg [55], the diffusivities of a wide variety of gases and vapors in poly(vinyl
chloride) exhibited a systematic progression. The diffusion coefficient changed ten
orders of magnitude with an order of magnitude change in diameter. Other molecular
size parameters proposed include molar volume, square root of molecular weight, and
kinetic or Lennard-Jones diameter. The interactional relationship of these quantities
may give different results. A specific example is CO2 which has a low kinetic diameter
but a larger molar volume or molecular weight square root.
1.5.4.2 Sorption
SA/SB is the ratio of the Henry’s Law sorption coefficients of the two molecules, and
can be interpreted as the sorption or solubility selectivity of the two molecules.
Sorption in rubbery polymers is similar to the dissolution of gases in liquids. However,
for several gases utilized in the gas pairs, such as CO2 and CH4, dual mode sorption is
31
commonly observed in glassy polymers. The complication arises due to the presence
of unrelaxed volume in glassy polymers. This excess volume is considered to be
present in the form of penetrant-scale, semi-permanent microcavities that can act as
adsorption sites. Thus, sorption in glassy polymers is modeled as simple dissolution in
the bulk of polymer and specific sorption in the microcavities [53].
The dual mode sorption theory comprises a sorption isotherm consisting of a Henry’s
law “dissolved” solubility and a Langmuir “hole-fitting” solubility.
bpbpCkpC H
++=
1
'
(1.5)
where k is the Henry’s law constant, p is the pressure, C’H, is the Langmuir capacity
constant and b is the Langmuir affinity constant. The value of C’H was shown to be a
linear function of the Tg with an intercept with a value of zero at a temperature equal
to the Tg.
It has been demonstrated that the solubility constant of CO2 is more sensitive to
polymer structure than O2, N2, and H2 gas molecules with a less polarity. With
increasing polarity, the solubility constant of CO2 increases slightly whereas a
decrease is observed for O2, N2, and H2.
1.5.4.3 Selectivity
Polymeric materials derive their selectivity primarily from their ability to separate
32
gases based on subtle differences in penetrant size. The size distribution of the
transient gaps that form and fade throughout the polymer due to thermally induced
motions of the polymer chain segments determines the polymer diffusion selectivity.
If the available gap is larger than the critical dimension of a penetrant, a diffusion step
can take place; if the gap is too small the jump is precluded. The fraction of diffusion
jumps of penetrant A to penetrant B is small in rubbery polymers unless the size
difference between the two penetrants is large. For glassy polymers, the range of
motion and thus the size of transiently opening gaps are much more narrowly
distributed, leading to a higher selectivity.
The balance between the solubility selectivity and the diffusivity selectivity
determines whether a membrane material is selective for molecule A or B. Membranes
with both high permeability and selectivity are desirable since a more permeable
membrane needs a smaller area to treat a given volume of gas mixtures and a more
selective membrane can produce a higher purity.
In addition to the above-mentioned correlations of two main parameters, diffusivity
and solubility, with the permeability and selectivity, several other correlations have
been mentioned in the literature. For example, Pilato et al. [56] studied the relationship
between the permeability and polymer density for a series of poly(aryl ethers) and
polycarbonates materials.
33
1.6 MEMBRANE MODULES AND DESIGN INSTRUCTIONS
Membrane gas separation system offers great benefit that enables it to be applied in
various industrial processes. Principally, the membrane module is the heart of a
membrane system to determine the yield and efficiency of a separation process.
Industrial membrane modules are regularly based on both flat and hollow fiber
configurations. The vast majority of industrial membrane modules are constructed into
five basic designs: plate-and-frame, spiral wound, hollow fiber, tubular and capillary.
Flat sheet membranes are usually contained in the plate-and-frame devices and spiral
wound elements, whereas the modules of hollow fiber, tubular and capillary involve
the hollow fiber configuration. Three most general modules employed in industrial
membranes separation processes will be discussed.
1.6.1 Plate-and-Frame Modules
Plate-and-frame design replicates conventional filtration setup. It is conceptually
simple, which consists of a package of flat sheet membranes. Plate-and-frame module
is easy to fabricate and the area of the membranes are well defined. The flat sheet
membranes, mainly used for experimental purpose to characterize the intrinsic
properties of membrane are stacked together like a multilayer sandwich in a frame.
This module consists of a cylindrical tube, spacer materials to separate the membrane
envelopes and rubber gaskets to direct the flow through the module and seal the
34
assembly [27]. The plate-and-frame package design is illustrated by a schematic
diagram shown in Figure 1.7. Lowest surface area/unit separator volume (~100-400
m2/m3) is the major drawback of plate-and-frame module. Consequently, this module
is disfavor in gas separation application and relegated to oxygen enrichment for small
scale medical application [8].
Figure 1.7: Schematic drawing of a plate-and frame module [25, 57]
1.6.2 Spiral-Wound Modules
The spiral-wound module maintains the simplicity of flat membranes fabrication, but
it is the next logical step from a flat membrane. This element increases the packing
density (membrane surface per module volume) remarkably to 300-1000 m2/m3 as
compared to plate-and-frame modules. As shown in Figure 1.8, the assembly consists
of a sandwich of flat sheet membranes to form an envelope enclosing a
separator/spacer in them to provide mechanical strength and permeate flow space. The
membranes envelope is wound around a central core of a perforated collecting tube.
35
When a spiral-wound module is in operation, the feed gas flows outside the membrane
envelope and the permeate is collected inside and removed through the central
collector.
Figure 1.8: Spiral-wound elements and assembly [27]
1.6.3 Hollow-Fiber Modules
The hollow fiber module consists of a large number of hollow fibers assembled
together into a pressure vessel. The membranes are in the shape of thin hollow tubes
with a very small diameter. The geometric arrangement of module is similar to
conventional heat-exchanger assembly, as presented in Figure 1.9. It offers a fairly
high packing density with a maximum membrane surface area per unit volume as high
as 10,000m2/m3 [25, 58]. Typically, the high pressure feed enters either into shell side
36
or bore side and permeate is collected from the other side. However, the major
disadvantage associated with hollow fibers is significant pressure drops at bore side
when large quantities of gas permeating the membrane. Therefore, for the gas
separation, feed stream must be relatively clean and shell side feeding is preferable to
avoid high pressure drop and attain a high membrane area.
Feed
Permeate
Reject (residue)
Fiber bundle
Figure 1.9: Hollow fiber separator assembly [58]
The selection and application of membrane module are principally depending on
economic consideration, separation performance and practicability (ease of cleaning,
ease of maintenance, ease of operation, compactness of the system, scale and the
possibility of membrane replacement). Table 1.5 shows the characteristic of 5 basic
membrane modules.
With referring to Table 1.5, spiral-wound and hollow fiber modules are the two major
choices for gas separation application. Spiral-wound modules tend to have lower
37
permeate pressure, whereas hollow fiber modules exhibit the highest packing density
among all the modules. To overcome the problem of pressure drop in hollow fibers,
the compression and recompression of product gas must often be considered, but they
usually result in a significant cost addition for a given process. Universally, the need
to achieve high packing density in a cost-effective manner is probably the most critical
criteria to be considered in a membrane system.
Table 1.5: Qualitative comparison of various membrane modules [25]
1.7 RESEARCH OBJECTIVES AND ORGANIZATION OF DISSERTATION
This research study comprised of three main aspects, which are the fabrication of dual-
layer hollow fiber membranes with a neat polymeric outer-layer, flat dense mixed
matrix membranes (MMMs), and dual-layer hollow fiber membranes with a mixed
matrix outer-layer. The main objective of my research study is to evaluate the
combined utilization of suitable polymer materials as a matrix and inorganic materials
38
as a dispersive phase to prepare mixed matrix membranes for gas separation in both
flat membrane and hollow fiber configurations. In this study, the variables of interests,
especially the effect of zeolite loadings and zeolite types on gas separation
performance of MMMs will be discussed. For flat dense MMMs, the interaction
between polymer matrix and molecular sieve phases was examined in order to
investigate the gas transport mechanism in MMMs. On the other hand, the
development of fabrication technique of dual-layer hollow fiber membranes with an
ultrathin dense-selective layer was also a major interest in this research study. In order
to achieve the above objective, the scopes of this study have been drawn as follows:
1. Development of dual-layer hollow fiber membranes with an ultrathin neat
polymeric dense-selective layer for gas separation;
2. Investigation of the effects of the interaction between polymer matrix and
zeolite phases on the gas separation performance of flat dense MMMs;
3. Study of the effects of novel silane modification of zeolite surface on the gas
separation performance of flat dense MMMs;
4. Fabrication and characterization of dual-layer hollow fiber membranes with a
submicron mixed matrix dense-selective layer for gas separation.
This dissertation is organized and structured into eight chapters and two appendices.
Chapter One is an introductory chapter of this dissertation. It provides the introduction
of the historical survey, important industrial applications, basic concepts, structural
characteristics and transport mechanisms of gas separation membranes as well as the
39
membrane module design. The research objective and outline of this dissertation are
also presented in this chapter.
The emergence, development and prediction of mixed matrix membranes are given in
Chapter Two. This chapter highlights the advantages of MMMs, problems and
solutions during the MMM development. The models for the prediction of the gas
separation performance of MMMs and their limitation for the nonideal interface
between polymer matrix and molecular sieve phases are also depicted in this chapter.
The general experimental approaches and methodologies, along with the materials
involved in all areas are documented in Chapter Three. The details of the membranes
preparation and membranes physical characterization techniques are addressed. The
constant volume-variable pressure method for both pure gas and mixed gas
permeation tests of flat membranes, as well as permeation tests of dual-layer hollow
fiber membranes are also reported in this chapter.
Chapter Four reports the fabrication of dual-layer polyethersulfone (PES) hollow fiber
membranes with an ultrathin dense-selective layer by using co-extrusion and dry-jet
wet-spinning phase inversion techniques with the aid of heat-treatment. SEM images
are applied to investigate the influence of heat-treatment with different temperatures
on the morphology of dual-layer hollow fibers. SEM pictures also confirm the
formation mechanism of macrovoids proposed by previous researchers.
40
Chapter Five investigates the effects of membrane preparation methodology, zeolite
loading and pore size of zeolite on the gas separation performance of PES-zeolite 3A,
4A and 5A MMMs. SEM pictures indicate the effect of a modified membrane-casting
procedure at high processing temperatures close to Tg of PES on the interface
between polymer matrix and zeolite phases. DSC is applied to characterize the
rigidification of polymer chains after the addition of zeolites. A new modified
Maxwell model is proposed in this chapter to correctly predict the gas separation
performance of MMMs through combining effects of polymer chain rigidification and
partial pore blockage of zeolites into calculation.
Chapter Six explores the effect of the chemical modification of zeolite surface using a
novel silane coupling agent, (3-aminopropyl)-diethoxymethyl silane (APDEMS) on
the gas separation performance of MMMs. Some physical characterization equipments,
such as elementary analysis, XPS, BET, SEM and DSC, are used to describe the
change of physical properties and morphology of zeolites and MMMs after the silane
modification of zeolite surface. In this chapter, the modified Maxwell model is again
applied to predict the permeability and selectivity of MMMs by slightly adjusting
some parameters due to the silane modification of zeolite surface.
Chapter Seven presents the fabrication of dual-layer PES/P84 hollow fiber membranes
with a submicron PES-zeolite Beta mixed matrix dense-selective layer by adjusting
41
the ratio of outer layer flow rate to inner layer flow rate during the spinning with the
aid of heat-treatment and two-step coating. SEM images characterize the effect of
heat-treatment on the morphology of dual-layer hollow fibers, and DSC data suggest
the rigidification of polymer chains due to the addition of zeolites in the outer layer.
These newly developed dual-layer hollow fibers are conducted in both pure gas and
mixed gas permeation tests in this chapter.
General conclusions drawn from this research study are summarized in Chapter Eight.
Inclusive in this ending chapter are some recommendations/suggestions for future
research related to this work.
Appendix A describes the synthesis and characterization of zeolite Beta, and a new
testing system with an oxygen analyzer to determine the O2/N2 mixed gas permeation
through hollow fiber membranes is provided in Appendix B.
1.8 REFERENCES
[1] J.K. Mitchell, On the penetrativeness of fluids, Am. J. Med. Sci., 7 (1830) 36.
[2] J.K. Mitchell, On the penetration of fluids, Am. J. Med. Sci., 13 (1833) 100.
[3] T. Graham, On the law of the diffusion of gases, Phil. Mag., 2 (1833) 175, 269 and
351.
[4] A. Fick, Uber diffusion, Ann. Physik, 94 (1855) 59.
42
[5] T. Graham, On the absorption and dialytic separation of gases by colloid septa. Part
I. Action of a septum of caoutchouc, Phil. Mag., 32 (1866) 401.
[6] S.V. Wroblewski, Wied. Ann. Phys, 8 (1879) 29.
[7] H. Kayser, Wied. Ann. Phys., 43 (1891) 544.
[8] R.E. Kesting, A.K. Fritzsche, Polymeric gas separation membranes, John Wiley &
Sons, New York, 1993.
[9] H.A. Daynes, The process of diffusion through rubber membrane, Proc. Roy. Soc.
(London), 97A (1920) 286.
[10] S. Loeb, S. Sourirajan, Sea water demineralization by means of a semipermeable
membrane, Report No. 6060, UCLA., 1960.
[11] S. Loeb, S. Sourirajan, Sea water demineralization by means of an osmotic
membrane, Adv. Chem. Ser., 38 (1962) 117.
[12] S. Loeb, S. Sourirajan, The preparation of high-flow semi-permeable membranes
for separation of water from saline solutions, US Patent 3,133,132 (1964).
[13] J.M.S. Henis, M.K. Tripodi, Multicomponent membranes for gas separation, US
Patent 4,230,463 (1980).
[14] R.W. Baker, Future directions of membrane gas separation technology, Ind. Eng.
Chem. Res., 41 (2002) 1393.
[15] R.W. Spillman, Economics of gas separation membranes, Chem. Eng. Prog., 85
(1989) 41.
[16] R.D. Noble, S.A. Stern, Membranes separation technology: Principles and
applications, Amsterdam, Elsevier Science, New York, 1995.
43
[17] A. Crull, Applications for membrane technology, Proc. of the 1998 Membrane
Technology/Separation Planning Conference (1998) 1.
[18] P.S. Puri, Gas separation membranes-current status, Chem. Ind. Digest, 9 (1996)
98.
[19] W.J. Koros, R. Mahajan, Pushing the limits on possibilities for large scale gas
separation: Which strategies? J. Membr. Sci., 175 (2000) 181.
[20] R.R. Zolandz, G.K. Fleming, Gas permeation applications, In Membrane
Handbook, ed. by W.S.W. Ho, K.K. Sircar, Chapman & Hall, New York, (1992)
pp. 78.
[21] G. Djoekita, Characterization and analysis of asymmetric hollow fiber
membranes for natural gas purification in the presence of hydrocarbons, M.S.
Thesis, University of Texas at Austin, 2000.
[22] K. Othmer, Encyclopedia of chemical technology, 3rd edition, Vol 19, John Wiley
& Sons, New York, 1982.
[23] L. Rosenberg, V. Chong, Propylene in the twenty-first century, Business
Intelligence Program, SRI Consulting, 1997.
[24] W.J. Koros, G.K. Fleming, Membrane-based gas separation, J. Membr. Sci., 83
(1993) 1.
[25] M. Mulder, Basic principles of membrane technology, Dordrecht: Kluwer
Academic Publishers, 1996.
[26] J.M.S. Henis, M.K. Tripodi, Composite hollow fiber membranes for gas
separation: The resistance model approach, J. Membr. Sci., 8 (1981) 233.
44
[27] D.R. Paul, Y.P. Yampol’skii (Eds.), Polymeric Gas Separation Membranes, CRC
Press, Boca Raton, 1994.
[28] T.S. Chung, A review of microporous composite polymeric membrane
technology for air-separation, Polym. and Polym. Comp., 4 (1996) 269.
[29] E. Kuzumoto, K. Nitta, Manufacture of permselective hollow membranes,
Japanese Patent 01,015,104 (1989).
[30] T. Yoshinaga, H. Shimazaki, Y. Kusuki, Polyimide bilayer hollow membranes
and their manufacture, Japanese Patent 02,169,019 (1990).
[31] S. Nago, Y. Mizutani, Microporous polypropylene hollow fibers with double
layers, J. Membr. Sci., 116 (1996) 1.
[32] H.H. Suzuki, K. Tanaka, H. Kita, K. Okamoto, H. Hoshino, T. Yoshinaga, Y.
Kusuki, Preparation of composite hollow fiber membranes of poly(ethylene
oxide)-containing polyimide and their CO2/N2 separation properties, J. Membr.
Sci., 146 (1998) 31.
[33] D.F. Li, T.S. Chung, R. Wang, Y. Liu, Fabrication of fluoropolyimide/
polyethersulfone (PES) dual-layer asymmetric hollow fiber membranes for gas
separation, J. Membr. Sci., 198 (2002) 211.
[34] Y. Liu, T.S. Chung, R. Wang, D.F. Li, M.L. Chng, Chemical cross-linking
modification of polyimide/poly(ether sulfone) dual-layer hollow-fiber
membranes for gas separation, Ind. Eng. Chem. Res., 42 (2003) 1190.
45
[35] L.Y. Jiang, T.S. Chung, D.F. Li, C. Cao, S. Kulprathipanja, Fabrication of
Matrimid/Polyethersulfone dual-layer hollow fiber membranes for gas separation,
J. Membr. Sci., 240 (2004) 91.
[36] D.F. Li, T.S. Chung, R. Wang, Morphological aspects and structure control of
dual-layer asymmetric hollow fiber membranes formed by a simultaneous co-
extrusion approach, J. Membr. Sci., 243 (2004) 155.
[37] Y. Li, C. Cao, T.S. Chung, K.P. Pramoda, Fabrication of dual-layer
polyethersulfone (PES) hollow fiber membranes with an ultrathin dense selective
layer for gas separation, J. Membr. Sci., 245 (2004) 53.
[38] M. Ashraf Chaudry, S. Ahmad, M.T. Malik, Supported liquid membrane
technique applicability for removal of chromium from tannery wastes, Waste
Manage., 17 (1997) 211.
[39] J. de Gyves, E. Rodriguez desan Miguel, Metal ion separations by supported
liquid membrane, Ind. Eng. Chem. Res., 38 (1999) 2182.
[40] F.J. Alguacil, M.A. Villegas, Liquid membranes and the treatment of metal
bearing wastewaters, Rev. Metal. Madrid, 38 (2002) 45.
[41] R.W. Spillman, M.B. Sherwin, Gas separation membranes: The first decade,
Chemtech, 20 (1990) 378.
[42] W.J. Koros, Gas separation in membrane separation systems: Recent
developments and future directions, ed by R.W. Baker, E.L. Cussler, W. Eykamp,
W.J. Koros, R.L. Riley and H. Strathmann, Noyes Data Corporation, United States
of America, (1991) pp. 189-241.
46
[43] R.J.R. Uhlhorn, K. Keizer, A.J. Burggraaf, Gas and surface diffusion in modified
gamma alumina syatems, J. Membr. Sci., 46 (1989) 225.
[44] J. Koresh, A. Soffer, The carbon molecular sieve membranes: General properties
and the permeability of methane/hydrogen mixtures, Sep. Sci. Technol., 22 (1987)
973.
[45] J.D. Way, D.L. Roberts, Hollow fiber inorganic membranes for gas separations,
Sep. Sci. Technol., 27 (1991) 29.
[46] A. Singh, W.J. Koros, Significance of entropic selectivity for advanced gas
separation membranes, Ind. Eng. Chem. Res., 35 (1996) 1231.
[47] A.B. Fuertes, D.M. Nevskaia, T.A. Centeno, Carbon composite membranes from
Matrimid and Kapton polyimides for gas separation, Microporous and Mesoporous
Mat., 33 (1999) 115.
[48] D.M. Ruthven, R.I. Derrah, Diffusion of monatomic and diatomic gases in 4A
and 5A zeolites, J. Chem. Soc., Faraday Trans., 71 (1975) 2031.
[49] K. Steel, Carbon membranes for challenging gas separations, Ph.D. Dissertation,
The University of Texas at Austin, 2000.
[50] J.C. Jansen, Advanced zeolite science and applications, Amsterdam, Elsevier,
New York, 1994.
[51] P.S. Tin, Membrane materials and fabrication for gas separation, Ph.D. Thesis,
National University of Singapore, 2004.
47
[52] W.J. Koros, G.K. Fleming, S.M. Jordan, T.H. Kim, H.H. Hoehn, Polymeric
membrane materials for solution-diffusion based permeation separations, Prog.
Polym. Sci., 13 (1988) 339.
[53] R Mahajan, Formation, characterization and modeling of mixed matrix membrane
materials, Ph.D. Thesis, The University of Texas at Austin, 2000.
[54] L.M. Robeson, Correlation of separation factor versus permeability for polymeric
membranes, J. Membr. Sci., 62 (1991) 165.
[55] A.R. Berens, H.B. Hopfenberg, Diffusion of organic vapors at low concentrations
in glassy PVC, polystyrene, and PMMA, J. Membr. Sci., 10 (1982) 283.
[56] L. Pilato, L. Litz, B. Hargitay, R. Osborne, A. Farnham, J. Kawakami, P. Fritze, J.
McGrath, Polymers for permselective membrane gas separations, ACS Polymer,
16 (1975) 42.
[57] H. Strathmann, Membrane Separation Processes, J. Membr. Sci., 9 (1981) 121.
[58] C.J. Geankoplis, Membrane separation processes: In transport processes and unit
operations, Prentice-Hall, Inc., American, (1995) 745.
48
CHAPTER TWO
MIXED MATRIX MEMBRANES FOR GAS SEPARATION
2.1 EMERGENCE OF MIXED MATRIX MEMBRANES
The selection of membrane materials is clearly the most important factor in membrane
separation technology. The membrane materials limit the preparation technique
employed, morphology obtained and the separation principles applied. Matching of
desirable materials’ performance characteristic with the separation application is very
important. Chemical interaction between a membrane material and a gas penetrant
determined the separation efficiency of a membrane separation process [1]. The choice
of materials is not arbitrary, but based on the specific properties of a given material
originating from its nature and structural factors. The key requirements of effective
separation materials include (1) high separation efficiency with reasonable high flux,
(2) good chemical resistance, (3) good mechanical stability, (4) high thermal stability,
(5) engineering feasibility, (6) satisfactory manufacturing reproducibility and (7) low
cost [2, 3]. It is of utmost importance that the material chosen must be highly
permeable to the gas of interest, which effectively minimizes the total membrane area
needed and reduces the capital cost. In practical gas separation application, a durable
membrane must be resistant to the separation environments exposed. Since pressure
difference is the driving force, membrane materials have to maintain the mechanical
49
rigidity against the stress implied. Additionally, many gas feed streams are
contaminated with chemically reactive organic vapors (aromatic compounds,
lubricating oils and solvents) that will probably destroy (swell or dissolve) the
membrane materials, therefore the membrane materials need to have strong chemical
resistance in order to prolong the lifetime of membrane. Excellent thermal stability is
also required to withstand the high operating temperatures. Conclusively, an ideal
material should be easily converted to cost-effective membrane, as well as possesses
the intrinsic characteristics of high permeability, selectivity and durability.
Over the last two decades, membrane material science is rapidly developed to produce
wide range of materials with different structures and specific functions to separation
goals. Two common synthetic membrane materials, polymeric (organic) and inorganic
membranes are the focus of attention in membranes technology, as classified in Table
2.1. Some properties of these materials will be described briefly in the later part.
Table 2.1: Materials for gas separation membranes [4]
Organic Polymers Inorganic Materials
Polysulfone, polyethersulfone Carbon molecular sieves
Cellulose acetate Nanoporous carbon
Polyimide, polyetherimide Zeolites
Polycarbonate (brominated) Ultramicroporous amorphous silica
Polyphenyleneoxide Palladium alloys
50
Polymethylpentene Mixed conducting perovskites
Polydimethylsiloxane
Polyvinyltrimethylsilane
2.1.1 Polymeric (Organic) Membrane Materials
Amorphous polymeric material, which is cost-effective with sufficient selectivity and
good processability, is the dominating material in the membrane separation
technology. It offers the greatest promise for both industrial application and fruitful
academic research. Gas transport through membrane materials and the intrinsic
permeation characteristic of polymers are significantly influenced by polymeric chain
structures. Principally, the amorphous polymers exist in either glassy or rubbery state
depending on the operating temperatures. In the glassy state, polymer is hard and rigid,
while it becomes soft and flexible in the rubbery state. Glass transition temperature, Tg
denotes the boundary between these two states, where it is the temperature at which
the thermal expansion coefficient changes in going from the rubbery to glassy state.
As illustrated in Figure 2.1, the material is rubbery, which exhibits the viscoelastic
behaviors above the Tg. While a glassy polymer is an polymeric material that is below
its Tg under the conditions of use. For most gases, rubbery membranes generally
exhibit a higher diffusivity and result in a higher productivity as compared to glassy
membranes. However, lower separation efficiency is always achieved by rubbery
51
materials as a consequence of their small diffusivity selectivity. Conversely, glassy
materials are characterized by a low intrasegmental mobility and long relaxation time
[5]. These materials offer enhanced “mobility selectivity” as compared to rubbery
polymers due to the more restricted segmental motions in glassy polymers.
Accordingly, glassy materials are inherently more size and shape selective, resulting in
a higher selectivity and mechanical stability relative to rubbery materials. Thus, the
glassy polymers receive industrial interest and are more commonly used in separation
membranes processes [6].
Figure 2.1: Schematic diagram of the specific volume of polymer as a function of temperature [7]
52
Dozens of polymers were developed for membrane gas separation over last two
decades. In spite of these efforts, only less than 10 glassy polymers have been utilized
as the membrane materials in gas separation. The primary choice of polymers is
included polycarbonates [8-10], polysulfones [11-13], polyesters [14, 15],
polypyrrolones [16, 17] and polyimides [18-21].
An interesting issue, namely “upper bound trade-off curve” was raised by Robeson,
depicting the inverse relationship between the gas permeability and gas pair selectivity
for various membrane materials [22]. In other words, an increase in the gas pair
selectivity often occurs together with a decrease in the gas permeability and vice versa.
This trend still occurs even though the synthesis of polymeric materials has been
significantly developed during the last few decades. All polymeric materials are
empirically lying on or below the straight line of upper bound, as shown in Figure 2.2
[23]. To further extend the industrial application of the membrane-based gas
separation technology, it is very necessary to synthesize new membrane materials with
both high selectivity and permeability [24, 25].
53
Figure 2.2: Trade-off line curve of oxygen permeability and oxygen/nitrogen selectivity [23]
2.1.2 Inorganic Membrane Materials
Today, polymeric membrane systems are widely used in different separation processes
and current market for inorganic membranes is extremely small. However, inorganic
membranes technology is rapidly receiving global attention owing to the superior
separation properties when compared to polymeric membranes. The market share of
inorganic membranes will definitely increase in the near future. Inorganic membranes,
54
especially molecular sieving materials such as silica, porous glass, crystalline zeolites,
microporous beryllium oxide powders and carbon [26-28] are research-intensive due
to their potential in surpassing the upper separation capability limit of glassy
polymeric membranes, as shown in Figure 2.2.
Molecular sieving materials are named by J.W. McBain in 1932 according to their
property of acting as sieves on a molecular scale [29]. These materials show extremely
attractive performance with a significantly high productivity and selectivity [30].
Besides, they can operate over higher temperatures than polymeric membranes,
therefore during gas separations whereby high temperatures are required, molecular
sieve membranes can help to avoid the trouble of ramping down the temperature to
maintain the physical integrity of an organic membrane and ramping up the
temperature again after separation is complete. Molecular sieve membranes can also
withstand organic solvents, chlorine and other chemicals better than organic
membranes. This also permits the use of more effective and yet corrosive cleaning
procedures and chemicals. Furthermore, they are not susceptible to microbial attack,
mechanical instability.
Among molecular sieve materials, zeolites and carbon molecular sieves have long
been attractive inorganic porous materials as adsorbents for adsorptive separations;
their application as a membrane material has also received great attention in recent
years [30]. Zeolites are crystalline, hydrated aluminosilicate (as formed in nature or
55
synthesized) with cations in group I and II, especially Na, K, Ca, Mg, Sr and Ba [31,
32]. They are principally synthesized by hydrothermal crystallization of reactive alkali
metal aluminosilicate gels [31-33]. Zeolites can be represented by the empirical
formula of M2/nO.Al2O3.xSiO2.yH2O. In this formula, n is the cation valence, x is
greater than or equal to two and y is a function of the porosity of the framework, the
Si/Al ratio and the cations present.
Zeolites are shown to be crystalline, microporous. The framework consists of a three
dimensional network of SiO4 and AlO4 tetrahedral, linked to each other by sharing the
oxygen atoms to form a rich variety of structures, The size of the zeolite pore opening
is determined by the number of tetrahedral required to form the pore and the nature of
the cations present in or at the pore entrance.
The cations are present to balance the negative charge introduced into the framework
by the substitution of Si4+ by Al3+. Because of the presence of the cations, aluminum-
rich zeolites all show a high affinity for water coupled with a fairly high saturation
capacity (~ 25 wt%). The stability of zeolites is roughly inversely related to the
amount of aluminum in the framework [32]. Table 2.2 lists some common zeolite
sieves and their pore apertures. For more complete overview of zeolite structures, one
could consult the reference [34].
56
Table 2.2: Properties of major synthetic zeolites
Carbon molecular sieves (CMSs), a new class of carbon, are produced by thermal
decomposition in a controlled chemical and thermal environment of non-melting
polymer materials or by carbonization of coal [35-37]. The pore size distribution in
carbon molecular sieves is narrow and the mean pore size is in the range of molecular
dimension (3-6Å) [37-39]. Aromatic microdomains or amorphous carbon is the
fundamental building block of the CMS materials. The misalignment of aromatic
microdomains during carbonization results in the formation of void spaces in carbon
matrix. The microporosity of carbon membranes is attributed to these void spaces
arising from a natural consequence of structural dicilanation [40]. The turbostratic
structure is used to represent the structure of graphite-like layers in carbon domains, as
shown in Figure 2.3 [41].
57
Figure 2.3: The structure of turbostratic carbon [41]
Nonetheless, the main obstacle of zeolite and carbon molecular sieve materials for
large-scale membrane separation process is their high production cost, fragile and
some principle difficulties during reproducible production. It is extremely difficult to
prepare large-scale defect-free molecular sieve membranes [42].
2.1.3 Mixed Matrix Membranes (MMMs)
A recent survey of the present state of gas separation membranes indicated that
current polymeric materials are inadequate to fully exploit existing opportunities [43].
Although molecular sieve materials such as zeolite and CMS lie well above the upper
bound trade-off curve of polymeric materials, these materials are expensive and
58
difficult to process as membranes. Therefore, the deficiencies in both pure molecular
sieves (fragile and high production cost) and polymers (low chemical resistance,
mechanical and thermal stability) encourage the development of alternative materials,
which are engineering feasible, economical, mechanically stable and efficient in
separation.
Organic-inorganic hybrid material is proposed to use molecular sieves as inserts in
polymer matrix, known as mixed matrix membrane [44]. Preparation of membrane
materials with advanced separation properties at competitive price is the major
purpose in developing mixed matrix membranes. Additionally, it is verified that this
hybrid material can combine the advantages of (1) excellent separation capability and
desirable stability of molecular sieves, and (2) good processability of polymeric
materials. The development of mixed matrix membranes in the past 25 years will be
elaborated in the next section.
2.2 DEVELOPMENT OF MIXED MATRIX MEMBRANES
2.2.1 Flat Dense Mixed Matrix Membranes
To explore the possibility enhancing the membrane gas separation performance by
means of the addition of zeolites, quite a few pioneer investigations on mixed matrix
membranes have been carried out and mainly focused on the flat dense membrane
configuration. Progress has been made in rubbery polymer-zeolite mixed matrix
59
membranes [45, 46], which showed a significant increase in O2/N2 selectivity,
especially at a high zeolite loading. The above-mentioned results validated the mixed
matrix composite membrane concept; regrettably, they are not practically attractive
because rubbery polymers might lack mechanical stability and desirable inherent
transport properties relative to rigid glassy polymers at high temperatures.
Therefore, some researchers selected rigid glassy polymers, which possess properties
closer to the “upper-bound”, as the base of mixed matrix membranes. Süer et al. [47]
found that PES-zeolite 4A MMMs exhibited a change from α (O2/N2) = 3.7 to 3.9, 4.2,
and 4.4 for 16.6, 33.3 and 50.0 wt% zeolite loadings, respectively. However, such gas
separation performance still remained far below the desirable selectivity of current
high-performance glassy polymers.
Matrimid®-zeolite 4A MMMs were fabricated [48] and showed a much higher
permeability and similar permselectivity than those of the neat polymeric membrane.
These strange results suggested the void existence between the polymer matrix and
zeolite phases, allowing the gases to bypass the zeolites, as shown in Figure 2.4. The
void formation during the fabrication of mixed matrix membranes has also been
proved by other researchers [49-52].
60
Zeolite
Polymer
Gas diffusion pathway
Gas bypasses the zeolite
Zeolite
Polymer
Matrimid®-zeolite 4A MMM
Figure 2.4: The SEM picture and sketch map of gas diffusion pathway in MMMs
The possible reasons of the void formation are poor contact and interaction between
the polymer matrix and sieve phases, different shrinkage generated during the solvent
removal and poor relaxation of a rigid polymeric matrix material [53, 54]. The
interaction between the polymer and sieve phases is probably a function of the
chemical nature of the two surfaces and can be attractive, repulsive or neutral. The
shrinkage of the matrix due to solvent removal can cause stress, especially for a rigid
material such as Matrimid®. Glassy polymers (high Tg and low flexibility) interact
poorly with the zeolite surface and are prone to result in a series of non-selective,
microporous voids around the molecular sieve domains. Therefore, how to overcome
the void formation is the largest challenge when the rigid glassy polymer materials are
selected as a matrix of MMMs.
61
To fabricate the void-free mixed matrix membrane and fulfill the selectivity
enhancement, some approaches were introduced based on the mechanisms of void
formation. The first method is filling voids using some small and suitable chemicals.
Yong et al. [52] introduced a compatibilizer, 2,4,6-triaminopyrimidine (TAP), to
prepare the void-free Matrimid®-zeolite MMMs by simultaneously forming hydrogen
bonding between them (see Figure 2.5). Results showed that CO2/N2 and O2/N2
selectivity of Matrimid®-zeolite 4A-TAP MMMs increased around three-fold
compared with the neat Matrimid® membrane, while CO2 and O2 permeability
decreased at least forty-fold. Therefore, these results are not attractive for industrial
applications due to low productivity.
Figure 2.5: Interaction model between zeolite, TAP and polyimide
62
The second method is the zeolite surface modification with some appropriate silane
coupling agents. Duval et al. [51] demonstrated that the adhesion between polymer
matrix and zeolite phases can be improved by modifying zeolite surface with some
silane coupling agents, as shown in Figure 2.6. Regrettably, there was no significant
improvement on permselectivity though SEM micrographs indicated good coupling
between silane and zeolite. Mahajan and Koros [55] also reported similar phenomena.
Although the improvement on interface between polymer matrix and zeolite phases
was clearly observed by SEM images, their MMMs made from modified zeolites
exhibited worse performance (i.e. a decrease in both permeability and selectivity)
compared with those made from unmodified zeolites. This may be due to the fact that
the addition of silane coupling agents may reduce the void dimension between
polymer matrix and zeolite phases, but cannot eliminate them completely. The size of
these voids may be in the range of nanometer or sub-nanometer [54-56], which is still
larger than the sizes of gas molecules. As a consequence, the addition of silane
coupling agents on the zeolite surface may help reduce the void size and result in an
increase in gas transport resistance across the membrane, but there is almost no much
improvement on permselectivity. Unless a proper silane coupling agent is chosen, it is
very difficult to completely prevent the gas penetrants from passing through these
voids.
63
zeolite surface
OH OHOH
R(CH2)nS i(OMe)3 + 3H2OHydrolysis
R(CH2)nS i(OH)3
Si OOH
OH
Si O Si OH
OH
R(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)n
OH
m
Si OOH Si O Si OH
R(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)nR
(CH 2)n
m
+
zeolite surface
O OO
Condensation
adsorption on zeolite surface and reaction
Silane
Figure 2.6: Schematic diagram of chemical modifications of zeolite surface using a silane coupling agent
The third method is an application of high processing temperatures. Mahajan et al. [48,
54] proposed to maintain the polymer flexibility during the membrane formation by
applying high processing temperatures close to Tg of polymeric materials coupled with
the use of a non-volatile solvent. For polymers with high Tg, the authors decreased Tg
by means of the incorporation of a plasticizer into the polymer matrix because it is
very difficult to find a non-volatile solvent with an enough high boiling point to match
the temperature requirement during the membrane formation. Their experimental
results showed that the addition of zeolite in polymer matrix clearly led to an
enhanced separation performance, thus demonstrating the validity of this solution.
Unfortunately, the addition of a plasticizer may lower the intrinsic gas separation
performance of polymeric materials. Therefore, choosing glassy polymers with an
64
intermediate Tg should be a more appropriate alternative for this solution. However,
few studies have been done on the application of polymeric materials with an
intermediate Tg as a matrix in MMMs by means of high processing temperature.
2.2.2 Hollow Fiber Mixed Matrix Membranes
As has already been noted, most of the studies on the advanced MMMs have been
dealing with the configuration of flat dense membrane. To further expand the
application of the promising MMMs, a more effective membranes structure, the
asymmetric membrane, especially the hollow fibers, should be explored. However,
there is little work reporting the hollow fiber membranes consisting of a mixed matrix
dense-selective layer formed by the phase inversion technique. Work on mixed matrix
hollow fiber membranes for gas and hydrocarbon separations was initially carried out
by Miller et al. [57], Ekiner and Kulkarni [58], and Koros et al. [59]; nevertheless, the
pioneering work was released in patents which emphasize on the product development.
The hollow fiber membranes are normally composed of a dense-selective layer and a
porous supporting layer, and its gas separation performance is mainly determined by
the dense-selective layer. Therefore, how to distribute most of molecular sieves in the
dense-selective layer has become an inevitable problem to make full use of the effect
of molecular sieves on the improvement of gas separation performance. Jiang et al.
[60] have demonstrated that three factors played important roles in determining the
65
distribution of molecular sieves in the resultant hollow fibers. They are (1) shear stress
within the spinneret, (2) die swell when exiting from the spinneret and (3) elongation
drawing in the air gap region. Based on their study, a majority of molecular sieves
were located in the middle part of porous supporting layer when applying MMM
materials in the formation of single-layer hollow fibers, and could not play any role in
improving the gas separation performance.
Therefore, the dual-layer hollow fiber configuration has to be used to fulfill the
application of MMM materials in hollow fibers. Jiang et al. [61] have successfully
fabricated the dual-layer hollow fibers with a mixed matrix outer layer, and
significantly enhanced the He/N2 and O2/N2 selectivity compared with neat polymeric
membranes by a combination of the heat-treatment and two-step coating methods.
Regrettably, although the heat-treatment method really narrowed voids between
polymer matrix and zeolite phases to the range that can be cured by silicon rubber
coating, it also greatly decreased the gas permeation flux due to densifying the whole
mixed matrix outer layer. In their work, the calculated thinnest mixed matrix dense-
selective layer thickness is around 2.5µm based on the intrinsic O2 permeability of
corresponding flat dense polysulfone-zeolite Beta MMMs [61]. To fully compete and
potentially replace the hollow fiber membranes made from neat polymeric materials,
the co-extrusion technology need to be further developed to fabricate dual-layer
hollow fiber membranes with an ultrathin mixed matrix dense-selective layer.
66
2.3 PREDICTION OF GAS SEPARATION PERFORMANCE OF MMMS
2.3.1 Prediction for MMMs with an Ideal Interface
Mixed matrix membranes combine the polymer matrix and the zeolite into a
homogeneous structure; therefore, the gas transport mechanism in MMMs should be
influenced by the properties of the two individual phases. Several theoretical models
have been used to predict the steady state permeation properties of MMMs as a
function of the permeability of the continuous and dispersed phases. A particularly
useful model was developed by Maxwell in 1873 to predict the steady-state dielectric
properties in a conducting suspension of identical spheres [62]. The constitutive
equations governing electrical potential and the flux through membranes are
analogues, thus permitting the applicability of Maxwell’s Model to the transport in
MMMs. The solution to calculate the effective permeability of MMMs with a dilute
concentration of the dispersed phase is [63]:
−+−+−−−−+
=)()1(
)()1()1(
dcdcd
dcdcdceff PPnPnnP
PPnPnnPPP
φφ
(2.1)
where, Peff is the effective permeability of a gas penetrant in a MMM with a volume
fraction (Φd) of dispersed phase (d) in a continuous matrix phase (c), Pc and Pd
represent the gas penetrant permeability in the continuous (polymer) and dispersed
(sieve) phases, respectively, and n is the shape factor of the dispersed (sieve) phase.
The situation of n = 0 corresponds to the transport of gaseous component through a
MMM made of side-by-side layers of the two phases (laminate) which are parallel to
67
the gas transport direction and the permeability can be rewritten as an arithmetic mean
of the permeability of the dispersed and continuous phases:
dddceff PPP φφ +−= )1( (2.2)
Another limit of n = 1 corresponds to the gas transport through the two phases (or
laminate) in series:
cddd
dceff PP
PPP
φφ +−=
)1( (2.3)
For the dilute suspension of spherical particles (n = 1/3), the equation can be rewritten
as the well-known Maxwell equation:
−++−−+
=)(2)(22
dcdcd
dcdcdceff PPPP
PPPPPP
φφ
(2.4)
More detailed explanation about shape factor which is closely associated with the
aspect ratio can be found in other reference [64]. The Maxwell model has been widely
applied to predict the steady state permeation properties of MMMs when the
dispersion and distribution of zeolite in polymer matrix are homogeneous.
2.3.2 Prediction for MMMs with a Nonideal Interface
The transport properties of organic-inorganic mixed matrix membranes are strongly
dependent on the nanoscale morphology of membranes. Although the interface
68
between polymer matrix and zeolite phases occupies an extremely small volume
fraction (i.e., less than 10-10%), it is particularly a critical determinant of the overall
transport property [48, 53-56, 65]. Figure 2.7 shows the schematic diagram of various
nanoscale interface structure of mixed matrix membranes. Case 1 represents an ideal
morphology which corresponds to the ideal Maxwell Model prediction in Equation 2.1;
Case 2 shows the detachment of polymer chains from the zeolite surface, which causes
the interface voids; Case 3 indicates the polymer chain rigidification in the vicinity of
the zeolite; Case 4 displays a situation in which the zeolite pore has been partially
blocked within the surface section of the zeolite. The factors involving in the
morphology of the interface have been investigated by many researchers. With the
comprehension of this description about the mixed matrix membranes, it is reasonable
to expect some modifications in the Maxwell Model.
Sieve
Polymer
Case 1
Polymer
Case 2
SieveSieve
Interface voids
Polymer
Case 3
SieveSieve
Rigidified polymer region
Polymer
Case 4
Sieve
Zeolite skin with the reduced permeability
Figure 2.7 Schematic diagram of various nanoscale interface morphology of MMMs
The theoretical work of the influence of interface voids (Case 2) was attributed to
several researchers including R. Mahajan el al. [48, 53-56]. It was suggested in their
researches that the gas flow through the interface voids (Case 2) follows Knudsen
diffusion with a slight modification to account for the finite size of the gas molecules:
69
×
−××××= −
rd
MTrD g
AAK 2
1107.9 5 (2.5)
where DAK is the diffusion coefficient, r is the effective pore (defects) radius in Å, T is
the absolute temperature, MA is the gas molecular weight and dg is the diameter of the
gas molecule in Å. The solution coefficient in these voids is assumed to be the same as
in the gas phase and have a partition coefficient taking into considerations the size of
the gas molecules relative to the pore radius:
2
211
−=
rd
RTS g (2.6)
The gas permeability through this interface is the product of DAK and S.
Due to the existence of the interface voids, a three-phase system which includes the
molecular sieve phase, the polymer phase and the interface between these two main
phases appears. This three-phase model can be assumed as a pseudo two-phase
structure with the polymer matrix being one separate phase and the combined
molecular sieve and the interface being the other phase. Firstly, the Maxwell model
can be used to obtain the permeability of the combined molecular sieve and interface
phase with the interface as the continuous phase and the molecular sieve phase as the
disperse phase. Thus we can obtain the permeability of the combined molecular sieve
and interface phase [63]:
−++−−+
=)(2)(22
dIsId
dIsIdIeff PPPP
PPPPPP
φφ
(2.7)
here, Peff is the permeability of the combined molecular sieve and interface phase, Pd
70
is the permeability of the dispersed or sieve phase, PI is the permeability of the
interface, Φd is the volume fraction of the molecular sieve phase, Φi is the volume
fraction of the interface, and Φs is the volume fraction of the molecular sieve phase in
the combined phase and is given by a simple expression:
id
ds φφ
φφ
+= (2.8)
The value of the combined permeability of molecular sieve and interface, Peff can then
be used along with the continuous polymer phase permeability, Pc to obtain a
predicted permeability PMMM of the three-phase mixed matrix membranes by applying
the Maxwell equation a second time:
−+++
−+−+=
))((2))((22
effcidceff
effcidceffcMMM PPPP
PPPPPP
φφφφ
(2.9)
Thus, if one can make an estimate of the volume fraction and the permeability of the
interface region, the Maxwell model can easily be applied to these more complicated
structures. When the interface diameter is enough large which is common in mixed
matrix membranes with rigid glassy polymers as the continuous phase, the predicted
selectivity by Maxwell Model taking consideration of the Knudsen diffusion along the
interface has never exceeded the selectivity of neat polymer phase, even though the
particles have extremely high selectivity. The predictions corresponded well with the
experimental observations [48, 54].
71
This approach of predicting the gas separation performance of MMMs in Case 2 can
be readily extended to Case 3 and Case 4; and such conditions as the polymer chain
rigidification and partial pore blockage of zeolite are chosen to substitute the Knudsen
diffusion. The parameters and equations for Cases 2 to 4 are detailed in Table 2.3 [66].
72
Table 2.3: Comparison of the modified Maxwell Model for Cases 2, 3, and 4 [66]
73
2.4 REFERENCES
[1] T. Aoki, Macromolecular design of permselective membranes, Prog. Polym. Sci.,
24 (1999) 951.
[2] K. Scott, Membrane separation technology, Scientific & Technical Information,
Oxford, 1990.
[3] H. Strathmann, Membrane separation processes: Current relevance and future
opportunities, AIChE Journal, 47 (2001) 1077.
[4] S.P. Nunes, K.-V. Peinemann, Membrane technology in the chemical industry,
Wiley-VCH, Weinheim, 2001.
[5] S.A. Stern, Polymers for gas separations: The next decade, J. Membr. Sci., 94
(1994) 1.
[6] R.Z. Raymond, G.K. Fleming, Gas permeation, In Membrane Handbook, ed. by
W.S.W. Ho, K.K. Sircar, Van Nostrand Reinhold, New York, (1992) pp. 19.
[7] W.J. Koros, M. Hellums, Transport properties, In Encyclopedia of Polymer
Science and Engineering, Vol. Supplement Volume, John Wiley and Sons, Inc,
New York, (1989) pp. 724.
[8] M.W. Hellums, W.J. Koros, G.R. Husk, D.R. Paul, Fluorinated polycarbonates for
gas separation applications, J. Membr. Sci., 46 (1989) 93.
[9] J.S. McHattie, W.J. Koros, D.R. Paul, Effect of isopropylidene replacement on gas
transport properties of polycarbonates, J. Poym. Sci., 29 (1991) 731.
74
[10] M. Aguilar-Vega, D.R. Paul, Gas transport properties of polycarbonates and
polysulfones with aromatic substitutions on the bisphenol connector group, J.
Polym. Sci, Part B: Polym. Phys., 31 (1993) 1599.
[11] J.S. McHattie, W.J. Koros, D.R. Paul, Gas transport properties of polysulphones: 1.
Role of symmetry of methyl group placement on bisphenol rings, Polymer, 32
(1991) 840.
[12] J.S. McHattie, W.J. Koros, D.R. Paul, Gas transport properties of polysulphones: 2.
Role effect of biphenol connector group, Polymer, 32 (1991) 2618.
[13] C.L. Aitken, W.J. Koros, D.R. Paul, Gas transport properties of bisphenol
polysulphones, Macromolecules, 25 (1992) 3651.
[14] L.A. Pessan, W.J. Koros, Isomer effects on transport properties of polyesters
based on bisphenol A, J. Polym. Sci., Part B: Polym. Phys., 31 (1993) 1245.
[15] M.R. Pixton, D.R. Paul, Gas transport properties of polyacrylates Part I:
Connector and pendant group effects, J. Polym. Sci., Part B: Polym. Phys., 33
(1995) 1135.
[16] D.R.B. Walker, W.J. Koros, Transport characterization of a polypyrrolone for gas
separation, J. Membr. Sci., 55 (1991) 99.
[17] X.S. Gao, Z.S. Tan, F.C. Lu, Gas permeation properties of some polypyrrolones, J.
Membr. Sci., 88 (1994) 37.
[18] T.H. Kim, W.J. Koros, G.R. Husk, K.C. O’Brien, Relationship between gas
separation properties and chemical structures in a series of aromatic polyimides, J.
Membr. Sci., 37 (1988) 45.
75
[19] S.A. Stern, Y. Mi, H. Yamamoto, A.K.St. Clair, Structure/permeability
relationships of polyimide membranes: Applications to the separation of gas
mixtures, J. Polym. Sci, Part B: Polym. Phys., 27 (1989) 1887.
[20] K. Tanaka, H. Kita, K. Okamoto, Permeability and permselectivity of gases in
fluorinated and non-fluorinated polyimides, Polymer, 33 (1992) 585.
[21] Y. Liu, T.S. Chung, R. Wang, D.F. Li, M.L. Chng, Chemical cross-linking
modification of polyimide/poly(ether sulfone) dual-layer hollow-fiber membranes
for gas separation, Ind. Eng. Chem. Res., 42 (2003) 1190.
[22] L.M. Robeson, Correlation of separation factor versus permeability for polymeric
membranes, J. Membr. Sci., 62 (1991) 165.
[23] A. Singh, W.J. Koros, Significance of entropic selectivity for advanced gas
separation membranes, Ind. Eng. Chem. Res., 35 (1996) 1231.
[24] W.J. Koros, G.K. Fleming, Polymers for gas separations: the next decade, J.
Membr. Sci., 83 (1993) 1.
[25] W.J. Koros, R. Mahajan, Pushing the limits on possibilities for large scale gas
separation: which strategies? J. Membr. Sci., 175 (2000) 181.
[26] D.W. Brek, Zeolite molecular sieves: Structure, chemistry, and use, Robert E.
Krieger Publishing Company, Malabar, Florida, 1973.
[27] T.A. Centeno, A.B. Fuertes, Carbon molecular sieve gas separation membranes
based on poly(vinylidene chloride-co-vinyl chloride), Carbon, 38 (2000) 1067.
[28] S.M. Saufi, A.F. Ismail, Fabrication of carbon membranes for gas separation-1
Review, Carbon, 42 (2004) 241.
76
[29] J.W. McBain, The sorption of gases and vapors by solids, Rutlrdgr & Sons,
London, 1932.
[30] S. Morooka, K. Kusakabe, Microporous inorganic membranes for gas separation,
MRS Bulletin, 24 (1999) 25.
[31] D.W. Breck, Zeolite molecular sieves, John Wiley & Sons, New York, 1974.
[32] H. van Bekkum, E.M. Flanigen, J.C. Janse, Introduction to zeolite science and
practice (eds), Elsevier Science Publishers, 1991.
[33] D.E.W. Vaughan, The synthesis and manufacture of zeolites, Chem. Eng. Prog.,
84 (1988) 25.
[34] W.M. Meier, D.H. Olson, Atlas of zeolite structure types, 3rd revised edition,
Butterworth-Heinemann, Stoneham, 1992.
[35] J.E. Koresh, A. Soffer, Study of molecular sieve carbons Part 1. Pore structure,
gradual pore opening and mechanism of molecular sieving, J. Chem. Soc., Faraday
I, 76 (1980) 2457.
[36] J.E. Koresh, A. Soffer, Study of molecular sieve carbons Part 2. Estimation of
cross-sectional diameters of non-spherical molecules, J. Chem. Soc., Faraday I, 76
(1980) 2472.
[37] J.E. Koresh, A. Soffer, The carbon molecular sieve membrane: General properties
and the permeability of methane/hydrogen mixtures, Sep. Sci. Technol., 22 (1987)
973.
77
[38] C.W. Jones, W.J. Koros, Carbon molecular sieve gas separation membranes I.
Preparation and characterization based on polyimide precursors, Carbon, 32 (1994)
1419.
[39] D.Q. Vu, W.J. Koros, S.J. Miller, High pressure CO2/CH4 separation using
carbon molecular sieve hollow fiber membranes, Ind. Eng. Chem. Res., 41 (2002)
367.
[40] G.K. Jenkins, K. Kawamura, Polymeric carbons, Cambridge University Press,
Cambridge, 1976.
[41] M. Acharya, J. Mathews, M.S. Strano, S.J.L. Billinge, S. Subramoney, H.C.
Foley, Simulation of nanoporous carbons: A chemically constrained structure,
Philos. Mag. B, 79 (1999) 1499.
[42] A.J. Bird, D.L. Trimm, Carbon molecular sieves used in gas separation
membranes, Carbon, 21 (1983) 177.
[43] C.M. Zimmerman, A. Singh, W.J. Koros, Mixed matrix composite membranes
for gas separation, J. Membr. Sci., 137 (1997) 145.
[44] S. Kulprathipanja, R.W. Neuzil, N.N. Li, Separation of fluids by means of mixed
matrix membranes in gas permeation, US Patent No. 4,740,219 (1988).
[45] M. Jia, K.V. Peinemann, R.D. Behling, Molecular sieving effect of zeolite filled
silicone rubber membranes, J. Membr. Sci., 57 (1991) 289.
[46] J.M. Duval, B. Folkers, M.H.V. Mulder, G. Desgrandchamps, C.A. Smolders,
Adsorbent-filled membranes for gas separation. Part 1. Improvement of the gas
78
separation properties of polymeric membranes by incorporation of microporous
adsorbents, J. Membr. Sci., 80 (1993) 189.
[47] M.G. Süer, N. Baç, L. Yilmaz, Gas permeation characteristics of polymer-zeolite
mixed matrix membranes, J. Membr. Sci., 91 (1994) 77.
[48] R. Mahajan, R. Burns, M. Schaeffer, W.J. Koros, Challenges in forming
successful mixed matrix membranes with rigid polymeric materials, J. App. Polym.
Sci., 86 (2002) 881.
[49] A. Erdem-Şenatalar, M. Tatlıer, Ş.B. Tantekin-Ersolmaz, Estimation of the
interphase thickness and permeability in polymer-zeolite mixed matrix membrane,
in: Proceedings of the 13th International Zeolite Conference, Montpellier, France,
2001.
[50] Ş.B. Tantekin-Ersolmaz, L. Şenorkyan, N. Kalaonra, M. Tatlıer, A. Erdem-
Şenatalar, n-Pentane/i-pentane separation by using zeolite-PDMS mixed matrix
membranes, J. Membr. Sci., 189 (2001) 59.
[51] J.M. Duval, A.J.B. Kemperman, B. Folkers, M.H.V. Mulder, G. Desgrandchamps,
C.A. Smolders, Preparation of zeolite filled glassy polymer membrane, J. Appl.
Polym. Sci., 54 (1994) 409.
[52] H.H. Yong, H.C. Park, Y.S. Kang, J. Won, W.N. Kim, Zeolite-filled polyimide
membrane containing 2,4,6-triaminopyrimidine, J. Membr. Sci., 188 (2001) 151.
[53] R. Mahajan, W.J. Koros, Factors controlling successful formation of mixed-
matrix gas separation materials, Ind. Eng. Chem. Res., 39 (2000) 2692.
79
[54] R. Mahajan, Formation, characterization and modeling of mixed matrix
membrane materials, Ph.D. Dissertation, The University of Texas at Austin, 2000.
[55] R. Mahajan, W.J. Koros, Mixed matrix membrane materials with glassy polymers.
Part 1, Polym. Eng. & Sci., 42 (7) (2002) 1420.
[56] R. Mahajan, W.J. Koros, Mixed matrix membrane materials with glassy polymers.
Part 2, Polym. Eng. & Sci., 42 (7) (2002) 1432.
[57] S.J. Miller, C.L. Munson, S.S. Kulkarni, J. Hasse, J. David, Purification of p-
xylene using composite mixed matrix membranes, US Patent 6,500,233 (2002).
[58] O.M. Ekiner, S.S. Kulkarni, Process for making hollow fiber mixed matrix
membranes, US Patent 6,663,805 (2003).
[59] W.J. Koros, D. Wallace, J.D. Wind, S.J. Miller, C.S. Bickel, D.Q. Vu,
Crosslinked and crosslinkable hollow fiber mixed matrix membrane and method of
making same, US Patent 6,755,900 (2004).
[60] L.Y. Jiang, T.S. Chung, C. Cao, Z. Huang, S. Kulprathipanja, Fundamental
understanding of nano-sized zeolite distribution in the formation of the mixed
matrix single- and dual-layer asymmetric hollow fiber membranes, J. Membr. Sci.,
252 (2005) 89.
[61] L.Y. Jiang, T.S. Chung, S. Kulprathipanja, An investigation to revitalize the
separation performance of hollow fibers with a thin mixed matrix composite skin
for gas separation, J. Membr. Sci., in press.
80
[62] D.Q. Vu, W.J. Koros, S.J. Miller, Mixed matrix membranes using carbon
molecular sieves, II. Modeling and permeation behavior, J. Membr. Sci., 211
(2003) 335.
[63] R.H.B. Bouma, A. Checchetti, G. Chidichimo, E. Drioli, Permeation through a
heterogeneous membrane: The effect of the dispersed phase, J. Membr. Sci., 128
(1997) 141.
[64] P.M. Subramanian, I.G. Plotzker, Barrier materials for blending, in polymer
blends volume 2: Performance, Ed. D.R. Paus, C.B. Bucknall, John Wiley and
Sons, New York, 2000.
[65] T.T. Moore, R. Mahajan, De Q. Vu, W.J. Koros, Hybrid membrane materials
comprising organic polymers with rigid dispersed phases, AICHE Journal, 50 (2)
(2004) 311.
[66] T.T. Moore, W.J. Koros, Non-ideal effects in organic-inorganic materials for gas
separation membranes, J. Mol. Struct., 739 (2005) 87.
81
CHAPTER THREE
MATERIALS AND EXPERIMENTAL PROCEDURES
The experimental work involved in this research study consisted of four parts. The
work scopes of this study are listed as follows:
1. Fabrication of dual-layer PES hollow fiber membranes with an ultrathin neat
polymeric dense-selective layer for gas separation.
2. Investigation of the effects of polymer chain rigidification, zeolite pore size and
pore blockage on flat PES-zeolite A mixed matrix membranes.
3. Study of the effects of novel silane modification of zeolite surface on polymer
chain rigidification and partial pore blockage in flat PES-zeolite A mixed matrix
membranes.
4. Fabrication and characterization of dual-layer polyethersulfone (PES)/BTDA-
TDI/MDI co-polyimide (P84) hollow fiber membranes with a submicron PES-
zeolite Beta mixed matrix dense-selective layer for gas separation.
Details of the above experiments will be individually reported in the following
chapters. This chapter describes the materials that involved in this study, the
membrane preparation techniques and the material characterization methods. The
motivation of above experiment work may not be clear at this stage, but will be
elaborated in latter chapters.
82
3.1 MATERIALS
3.1.1 Polymers
In chapters 4, 5 and 6, only polyethersulfone (PES) was chosen as the polymeric
material due to its appropriate Tg of 215oC, which facilitates the application of high
processing temperature method during mixed matrix membrane formation to reduce
the void formation between polymer matrix and zeolite phases. Meanwhile, PES
possesses superior mechanical and membrane forming characteristics, and various
applications in gas separation [1].
In chapter 7, PES was still chosen as the continuous polymeric matrix of the mixed
matrix outer layer of dual-layer hollow fibers. Nevertheless, a glassy BTDA-TDI/MDI
co-polyimide commercially named as P84 was selected as the polymer material of the
microporous inner layer of dual-layer hollow fibers due to its excellent thermal
resistance and mechanical stability. The enormous difference between the glass
transition temperatures (Tgs) of PES (215oC) and P84 (315oC) [2] permits us to heat-
treat the PES-zeolite mixed matrix outer layer to reduce or eliminate the voids
between polymer matrix and zeolite phases without damaging the microporous
structure of inner layer. Table 3.1 shows the chemical structures, glass transition
temperatures, and densities for these two polymeric materials.
83
Table 3.1: Chemical structures and properties of PES and P84
OO
N
O
O
N
O
CH3
C
H
H80 % 20 %
n
OO
N
O
O
N
O
CH3
C
H
H80 % 20 %
n
1.37215PES
1.31315P84
Density(g/cm3)
Tg
(oC)Chemical structurePolymeric
materials
1.37215PES
1.31315P84
Density(g/cm3)
Tg
(oC)Chemical structurePolymeric
materials
The commercial Radel A-300 PES powder was purchased from Amoco Performance
Products Inc., OH, USA. It has a weight-average molecular weight of about 15,000.
The P84 co-polyimide powder was supplied from HP Polymer GmbH, Austria. These
two polymers were dried at 120oC in a vacuum oven overnight before use.
3.1.2 Molecular Sieves
The molecular sieves involved in the preparation of mixed matrix membranes were
various zeolites with special crystal structure, pore size and polarity. The zeolites used
in this investigation were zeolite 3A, 4A, 5A with an average particle size of 3µm
purchased from Aldrich Chemical Company, Inc.(USA) and with an average particle
size of 1.5µm supplied by UOP LLC, Des Plaines, IL, USA, and nano-sized zeolite
Beta self-synthesized in our laboratory. The synthesis procedure and characterization
of nano-sized zeolite Beta will be discussed in Appendix A, so it is not necessary to
84
duplicate it here. All zeolites were dehydrated at temperatures ranging from 250oC to
350oC under vacuum for 2h before use to remove the adsorbed water vapor or other
organic vapors. The physical properties of zeolites are summarized in Table 3.2.
Table 3.2: Main characteristics of zeolites
1.83Na+167.4BEABeta
1.48Ca+14.8LTA5A
1.52Na+13.8LTA4A
1.69K+13.0LTA3A
Densitycrystal
(g/cm3)CationSi/Al ratioPore size
(Å)StructureZeolite
1.83Na+167.4BEABeta
1.48Ca+14.8LTA5A
1.52Na+13.8LTA4A
1.69K+13.0LTA3A
Densitycrystal
(g/cm3)CationSi/Al ratioPore size
(Å)StructureZeolite
3.1.3 Silane Coupling Agents
(3-Aminopropyl)diethoxymethylsilane (APDEMS) was purchased from Sigma-
Aldrich as a silane coupling agent, and used without further purification to modify the
zeolite surface. The scheme of its chemical structure is illustrated in Figure 3.1.
Figure 3.1: Chemical structure of APDEMS
|
|(CH2)3 Si
CH3
O
H 2 N
CH2CH3
OCH2CH3
85
3.1.4 Others
N-methyl-2-pyrrolidone (NMP) was purchased from Merck and toluene was from
Fisher Scientific UK Ltd. They were dried using the activated molecular sieve 4A
beads with a diameter of 3-5mm supplied by Research Chemicals Ltd., and then
filtered through a 0.2µm Teflon filter before use. Ethanol, methanol and hexane were
purchased from Merck Inc., J.T. Baker and EM Science, an associate of Merck,
respectively. All these three chemicals were used as received.
3.2 FABRICATION OF DUAL-LAYER PES HOLLOW FIBER MEMBRANES
WITH A NEAT POLYMERIC OUTER LAYER
3.2.1 Spinning Line
In this research study, the solution spinning and phase inversion processes were used
to prepare the hollow fiber membranes with neat polymeric materials. Figure 3.2
describes the schematic diagram of lab-scale hollow fiber spinning line applied in this
research study and dual-layer spinneret design. The real pilot scale spinning line was
supplied by the Motianmo Technology Ltd. Company from Tianjin, China.
86
Spinneret
Syringe pumps
Polymer dope
Bore fluid
Coagulation bath Flushing bath
Water spray
Bore fluid Inner layer
Outer layer
Design of spinneret
Figure 3.2: Schematic diagram of the lab-scale hollow fiber spinning line and dual-layer spinneret design
3.2.2 Preparation of Spinning Dope
The composition of the outer-layer dope solution was 35/50/15 (in wt%)
PES/NMP/ethanol, while the composition of the inner-layer dope solution was
25/61/14 (in wt%) PES/NMP/ethanol. The two dope solutions were prepared by the
procedure described elsewhere [3, 4]: PES polymeric powder was first dispersed in a
cold NMP/ethanol mixture (0-3oC) with a high speed stirrer. The chilled solvent
reduced the dissolving rate of PES powders and thus prevented powders from
87
agglomeration. The dope container was then agitated in an ice bath for 1.5h and at
room temperature until it was fully dissolved. After the dope solutions became
homogenous, they were degassed for several hours in the bottle and then poured into
500ml syringe pumps (ISCO Ltd., shown in Figure 3.2) followed by degassing for 1
day before spinning.
3.2.3 Spinning Process and Solvent Exchange
By applying pressure from the syringe pumps, the dope solutions were extruded
through the channels of the spinneret (NMB Ltd., shown in Figure 3.2) with the
outer/inner/bore diameters of 1.2/0.81/0.3 mm, and exited at the orifice to form the
nascent hollow fibers, while a bore fluid was extruded at the same time from the
center channel of the spinneret to form the bore or the lumen. Right before entering
the chamber of the spinneret, the fluids went through a package of 15µm sintered
metal filters (Swagelok®). The fibers in this stage are called nascent hollow fibers.
Thereafter, the fibers were phase separated and vitrified by the diffusive removal of
the solvent and/or addition of non-solvent in the coagulant bath (see Figure 3.2). Tap
water was used as the non-solvent coagulant in this research study. In some cases, the
air gap between the spinneret and the coagulant was employed and this process is
called dry-jet/wet spinning. If the nascent hollow fibers are extruded directly into the
quench bath (i.e. no air gap), the process is called wet-jet spinning. Finally, the as-
88
spun fibers were taken by a drum and then cut into segments around 1 meter long, and
rinsed in a clean water bath for at least 3 days to remove the remaining solvent.
These as-spun hollow fibers were carried out the solvent exchange without further
drying after the rinse for 3 days. The procedure of solvent exchange was as follows: 1)
immersed these fibers into methanol under stirring for 30min; 2) repeated the same
procedure using another fresh methanol for two more times; then 3) immersed these
fibers into n-hexane under stirring for 30min; 4) repeated the same procedure using
another fresh n-hexane for two more times. Finally, these fibers were dried in the air at
ambient temperature for further test and study.
The solvent exchange process was found to play an important role in controlling the
morphology and separation performance of hollow fibers. If the water-wet asymmetric
membranes are dried in air directly, the skin structure will become dense due to the
enormous capillary force during the water removal process. Therefore, the solvent
exchange to replace water with organic volatile compound with much lower surface
tension was developed and reported in the relative literature [5].
3.2.4 Post-treatment Protocols
Two post-treatment protocols were used in this part to seal the defects on the outer
surface of resultant dual-layer hollow fibers. One was to make modules directly from
89
the as-spun hollow fiber membranes and then dip-coat these modules with a 3 wt%
silicone rubber solution [3]. The other was to heat-treat the hollow fiber membranes;
then make them into modules; finally perform the same dip-coating process. Two
heat-treatment conditions were employed. Condition A directly heated these fibers in a
vacuum oven from room temperature to 75oC, held there for 3h and then cooled down
naturally. Condition B heated these fibers from room temperature to 150oC with a
heating rate of 1oC/min under vacuum; held at 150
oC for 1h and then cooled down
naturally. The use of these two annealing temperatures was based on our experience
and literature information.
The use of heat treatment to control pore size for composite membranes has been often
practiced in industry [6-7] and their effects on membrane morphology have been
summarized elsewhere [8]. Annealing reduces a substrate’s pore size and improves
selectivity. Without annealing, the inventors could not produce useful membranes.
The annealing technology was probably developed 30 years ago for PBI membranes in
RO applications [9]. Without annealing in hot ethylene glycol (140-180°C), PBI
membranes had a poor RO selectivity. This technology was extended by Cabasso and
Tamvakis [10] for the development of polyethyleneimine/polysulfone (PS) hollow
fibers for RO. They observed intrusion of a polymeric solution into pores to a depth as
great as 0.5 µm. They contracted surface pore sizes and reduced intrusion by means of
annealing microporous PS hollow fibers at 110-150°C for less than 30 minutes.
90
3.3 FABRICATION OF FLAT DENSE PES-ZEOLITE A MIXED MATRIX
MEMBRANES WITH SILANE MODIFIED ZEOLITE OR UNMODIFIED
ZEOLITE
3.3.1 Chemical Modification Method of Zeolite Surface
Figure 3.3 showed the flowchart of zeolite surface modification with a novel silane
used in this research study. This procedure was derived with some modifications from
the Plueddemann’s method [11]. It consisted of three steps: 1) stirred the mixture of
toluene (500ml) + APDEMS (20ml) + zeolite (2.5g) for 24h at room temperature
under N2 for the reaction between the zeolite and APDEMS; 2) filtered and washed
the zeolite with 500ml toluene and then 250ml methanol to remove the unreacted
silane; 3) dried the modified zeolite for 1h at 110oC under vacuum.
OH
OHSi
H3C
RO
RO
(CH2)3SiO
H3C
O
Modified zeolite (zeolite-NH2)
room temperature N2, 24 h
zeolite
(CH2)3 NH2
NH2
silane
+
Filtering Drying
Washing withtoluene and methanol
In toluene
At 110°C for 1 h under vacuum
APDEMS (R = CH3CH2)3A, 4A and 5A
OH
OHSi
H3C
RO
RO
(CH2)3SiO
H3C
O
Modified zeolite (zeolite-NH2)
room temperature N2, 24 h
zeolite
(CH2)3 NH2
NH2
silane
+
Filtering Drying
Washing withtoluene and methanol
In toluene
At 110°C for 1 h under vacuum
APDEMS (R = CH3CH2)3A, 4A and 5A
Figure 3.3: Flowchart of the chemical modification of zeolite surface
91
3.3.2 Preparation Procedures of Flat Dense Mixed Matrix Membranes
The conventional solution-casting techniques to form mixed matrix membranes tend
to create voids between polymer and zeolite phases, which is one of the largest
challenges for the mixed matrix membranes. Therefore, based on other literatures [12,
13], the method by means of applying high processing temperatures during the
membrane formation was adopted to eliminate the voids between two phases in this
part. Firstly, the unmodified zeolite or APDEMS modified zeolite was dispersed in the
solvent of NMP under a sonication for 5 minutes and stirred for 1h. Secondly, the
polymeric material, PES was added into the mixture of zeolite/NMP at twice and
stirred for long enough time to ensure the complete homogeneity of dope solution.
However, there were three exceptions taken: 1) prior to film casting, the
polymer/zeolite/solvent mixture was degassed in vacuum for 4h; 2) after the
membrane was formed and held at 200oC for 12h, the oven temperature was increased
to 250oC at 10oC/20min and then the membrane was annealed at 250oC for 12h; and 3)
after annealing, the membranes were cooled down naturally to room temperature
instead of immediate quenching. The flowchart of the preparation methodology of flat
PES-zeolite A mixed matrix membranes was shown in Figure 3.4. The thickness of
dried MMMs used for the permeability measurements varied from 60µm to 90µm. The
nomenclature of resultant flat dense MMMs is given in the form <Polymer-zeolite
mixed matrix membranes (MMMs)>.
92
Dispersion of zeolite in NMP
Stirring for 1hSonicating for 5min
Addition of 10 wt% of total PES in the NMP/zeolite mixture
Addition of remaining PES bulk in the mixture to achieve 30 wt% polymer concentration
Stirring for 4h
Casting mixture on glass surface at room temperature
Stirring overnightVacuuming for 4h
Placing cast film in oven and applying partial vacuum with N2
Preheating oven to 170oCPurging oven with N21h laterApplication of
complete vacuum in oven
Holding for 12h
Increasing to 200oC at 10oC/h
Holding for 12h and then natural cooling down to room temperature
Increasing to 250oC at 10oC/20min
Figure 3.4: Flowchart of the preparation methodology of flat PES-zeolite A mixed matrix membranes
3.4 FABRICATION OF DUAL-LAYER PES/P84 HOLLOW FIBER
MEMBRANES WITH A PES-ZEOLITE BETA MIXED MATRIX OUTER
LAYER
3.4.1 Preparation of Spinning Dope
The composition of the inner-layer dope solution was 20/67/13 (in wt%)
P84/NMP/ethanol, and the preparation procedure has been described in the section
3.2.2. The composition of the mixed matrix outer-layer dope solution was 35/50/15 (in
wt%) PES/NMP/ethanol plus 20 wt% zeolite loading in total solid. Its preparation
process was slightly different from the neat polymeric dope solution. Firstly, zeolite
93
Beta was dispersed in the solvent and stirred for around 3h at a high speed to force
particles to separate homogenously. Secondly, a desirable amount of PES was added
in the zeolite/solvent mixture and stirred to form the homogenous polymer solution.
Thirdly, the PES/zeolite/solvent mixture was vacuumed for 4~5h under the continuous
but slow stir to remove the air trapped near the zeolite surface, which may exacerbate
the formation of voids between polymer and zeolite phases. Finally, the non-solvent
(ethanol) was added in the mixture and stirred for at least 1 day. The ethanol addition
made the composition of outer-layer dope solution closer to the point of phase
separation, thus creating more pores in the outer-layer substructure. After both inner-
layer and outer-layer dope solutions were prepared, they were degassed for several
hours in the bottles and then poured into two 500ml syringe pumps (ISCO Ltd.)
followed by degassing for 1 day before spinning.
3.4.2 Spinning Process and Solvent Exchange
The dual-layer hollow fiber membranes with a mixed matrix outer layer were
fabricated by the co-extrusion technique using a dual-layer spinneret as depicted in the
sections 3.2.1 and 3.2.3. The flow rates of the bore fluid and two polymeric dope
solutions were controlled by three ISCO syringe pumps.
The as-spun dual-layer hollow fibers were rinsed in the clean water bath for at lease 3
days and then carried out the solvent exchange without further drying. The procedure
94
of solvent exchange has been outlined in the section 3.2.3. Finally, these fibers were
dried in the air at ambient temperature for further test and study.
3.4.3 Post-treatment Methods
In this part, two post-treatment methods were used to eliminate the voids on the outer
surface of dual-layer hollow fiber membranes induced by the unfavorable interaction
between polymer and zeolite Beta phases. One was the heat-treatment method and its
procedure was to heat these as-spun fibers from room temperature to a high
temperature at a heating rate of 0.6oC/min under vacuum; then hold them there for 2h
and finally cool down naturally. The high temperatures involved in this part included
200, 210, 220 and 235oC. They were determined based on Tg values of hollow fibers,
our past studies and information in the literatures. The heat-treatment processes were
all performed in a precise high-temperature programmable furnace (CenturionTM
Neytech Qex) under vacuum.
The other was a two-step coating approach which has been discussed extensively by
Jiang et al. in our group [14]. The coating solutions included 1) 0.2 wt%
diethyltoluenediamine in iso-octane and 2) a mixture of 0.2 wt% 1, 3, 5-
benzenetricarbonyl chloride and 2 wt% silicon rubber in iso-octane. Each step of
coating was conducted by dip coating fibers under vacuum for 30min. Finally, these
fibers were dried at 100oC for 2h before the gas separation characterization.
95
3.5 CHARACTERIZATION OF PHYSICAL PROPERTIES
3.5.1 Brunauer-Emmett-Teller (BET)
The surface area and total pore volume of zeolites before and after the silane
modification were measured at 77K on a high speed gas sorption analyzer of model
NOVA 3000 Series. The surface area of zeolites was determined through the standard
multipoint Brunauer-Emmett-Teller (BET) method using nitrogen as the adsorbate.
The pore volume results were derived from the amount of vapor adsorbed at a relative
pressure close to unity by assuming that pores are then filled with liquid adsorbate.
3.5.2 Dynamic Light Scattering (DLS)
Mean particle size and polydispersity index were obtained by Dynamic Light
Scattering (DLS) on a Brookhaven 90Plus submicron particle size analyzer with a
15mW solid state laser at a wavelength of 677nm. Measurements were always carried
out at a scattering angle of 90°.
3.5.3 Differential Scanning Calorimetry (DSC)
The glass transition temperature (Tg) of membranes was analyzed with differential
scanning calorimetry (DSC) on a Mettler-Toledo DSC 822e. A small piece of
96
membrane was first stored under vacuum at 110oC for 1 day to remove adsorbed water;
then weighed and placed into aluminum DSC pans. The measurements were
conducted in a dry nitrogen environment with a flow rate of 20 ml/min. The testing
samples were heated from 25oC to 300oC at a rate of 10oC/min in the first DSC cycle
to remove thermal history; held for 10min at 300oC and then cooled from 300oC to
25oC at the rate of 10oC/min; finally carried out with the same procedure in the second
cycle. Tg of the sample was determined as the midpoint temperature of the transition
region in the second heating cycle. Tg for each sample was estimated from the average
value of two specimens. Tg provided a qualitative estimation of the flexibility of
polymer chains, and was very useful to compare the polymer chain rigidity of MMMs
at different zeolite types and loadings with that of pure polymeric membranes [15, 16].
3.5.4 Elemental Analysis
The elemental analysis was carried out on a Perkin-Elmer PE 2400 Series II CHNS/O
analyzer. This technique determined the presence of elements such as C, H and N in a
substance. The results obtained are the percentage amounts of these atoms against the
total weight. The substance was combusted under an oxygen stream in a furnace at
high temperatures (950°C). The end product of combustion would be mostly the
oxides of concerned elements in the form of gases, such as CO2, H2O and N2. These
gases are then separated and carried to a detector using inert gas like helium or argon
and the composition was measured as a function of thermal conductivity.
97
3.5.5 Scanning Electron Microscope (SEM)
SEM is a widely applied technology to obtain the morphology of a membrane. In this
research study, it is used to analyze the surface and cross-section morphology of either
the flat sheet membrane or the hollow fibers, and the contact between the molecular
sieves and the polymeric matrix phases. The morphology was observed by scanning
electron microscope (SEM JEOL JSM-5600LV) or field emission scanning electron
microscope (FESEM JEOL JSM-6700F). The samples for the cross-section
characterization were fractured in liquid nitrogen. After mounting all the specimens on
the stub using a double-side conductive carbon adhesive tape, the specimens were
further dried under vacuum overnight. All samples were sputter coated with platinum
of 200-300Å in thickness using JEOL JFC-1200 Ion Sputtering Device before testing.
3.5.6 Energy Dispersion of X-ray (EDX)
The SEM samples were also used to measure their surface elemental content by
Oxford Inca energy dispersion of X-ray (EDX). The mapping mode was performed on
the surface of zeolite, flat mixed matrix membranes and dual-layer hollow fibers to
detect the existence of certain elements. The line scan spectrum was applied to the
hollow fiber cross-section to determine the elemental distribution along the cross-
section.
98
3.5.7 X-ray Photoelectron Spectroscopy (XPS)
The XPS measurements were carried out on a Shimadzu Kratos XPS System-AXIS
HSi-165 Ultra spectrometer using a non-monochromatic Mg Kα photon source (400
eV photons). All core-level spectra were obtained at a photoelectron take-off angle of
90° with respect to the sample surface.
3.5.8 X-ray Diffraction (XRD)
An X-ray diffractometer (GADDS XRD system, Bruker AXS) was performed to
quantitatively measure the ordered dimension and interchain spacing of polymer, the
structure of zeolites and carbon molecular sieve membranes at room temperature. A
small piece of the film or double-side adhesive with a thin layer of powders on one
side was mounted onto the sample holder. Ni-filtered Cu Kα with a radiation
wavelength λ = 1.54Å was used at 40kV and 40mA. The average intersegmental
distance of polymer chains were reflected by a broad peak center on each X-ray
pattern. The d-space was calculated by the Bragg's equation as follows:
nλ = 2d sinθ (3.1)
where θ is the X-ray diffraction angle of the peak.
3.6 CHARACTERIZATION OF GAS TRANSPORT PROPERTIES
99
3.6.1 Pure Gas Permeation Test
3.6.1.1 Neat Polymeric hollow fibers
Figure 3.5 depicts the schematic diagram of a hollow fiber module structure and the
pure gas permeation testing apparatus [17]. A module was normally composed of 5 or
10 pieces of hollow fibers with a length of around 10cm. One end of the module is
sealed with 5min rapid epoxy resin (Araldite®, Switzerland), while the shell side of the
other end was glued onto the metal module with regular epoxy resin (Eposet®). It took
around 8h for the regular epoxy to be fully cured. A feed pressure was applied to the
shell side and the permeate side (lumen side) was connected with the atmosphere.
Both feed side and permeate side were sealed with O-ring to prevent the leakage.
1
23
4 4 4
5 5 5
6
1 Gas Cylinder
2 Pressure Regulator
3 Pressure Gauge
4 Membrane Module
5 Purge Gas
6 Permeate Gas
Rapid epoxy
Regular epoxy
Figure 3.5: Pure gas permeation testing apparatus for the neat polymeric hollow fibers
100
Once potted into the modules, the feed chamber would be purged using the desired gas
at 50psia for 10 times. Then the fibers were permeated with pure gases He, O2, N2,
CH4 and CO2 in the shell feed as shown in Figure 3.5. For testing the permeance of
pure gases in neat polymeric hollow fibers, a preconditioning period of 0.25h for O2
and N2 or 1.5h for CH4 and CO2 was required before conducting the permeance
measurements. The feed pressures were 200psia for He, O2 and N2 gases; 100 and
200psia for CH4 gas; while for CO2 gas, the measurement were conducted by applying
the feed side with series of feed pressures including 25, 50, 100, 150, 200psia, so as to
investigate the plasticization phenomenon in asymmetric hollow fiber membranes
with an ultrathin dense selective-layer. The gas flow rate in the permeate side was
recorded using a digital bubble flow meter (Optiflow 570, error ~2%). At least 3-4
modules were tested at room temperature for each experimental condition.
The permeance, P/L, was calculated by the following equation [18]:
PDlnQ
PAQ
LP
∆=
∆=
π (3.2)
where P is the permeability of separating layer (Barrer) (1 Barrer = 1 x 10-10 cm3
(STP)-cm/cm2-s-cmHg), L is the thickness of the apparent dense selective-layer (cm),
Q is the normalized pure gas flux (cm3 (STP)/s), A is the membrane effective surface
area (cm2), n is the number of fibers in one testing module, D is the outer diameter of
the testing fibers (cm), l is the effective length of the modules (cm), and ∆P is the gas
pressure difference cross the membrane (cmHg). We use GPU as the unit of
permeance of a membrane (1 GPU = 1 x 10-6 cm3 (STP)/cm2-s-cmHg).
101
The ideal selectivity of an asymmetric hollow fiber membrane is defined as the ratio
of the permeances of the two gases [18]:
B
ABA LP
LP)/()/(α / = (3.3)
The dense selective-layer thickness is estimated by the following equation [18]:
)membranefiber (hollowO
)film (dense
2
2
)(P/LP
L O= (3.4)
3.6.1.2 Flat Dense Neat Polymeric or Mixed Matrix Membranes
For the characterization of the pure gas permeability through the flat dense neat
polymeric or mixed matrix membranes, the membrane of interest was mounted into
the permeation cell in a way as shown in Figures 3.6 and 3.7 [19], so that it formed the
only passage for the gas flow from the upstream cell to the downstream cell. The
thickness of membrane was measured using a digital film comparator (Mitutoyo
Corp.). The efficient membrane area available for gas permeation has a diameter of
20mm. The permeation cell chamber was separated from the atmosphere by sealing
with a double O-ring design in Figure 3.7.
It is assumed that the gas transport from the upstream to the downstream through the
adhesive or the O-ring is negligible compared to that through the membrane. In
addition, the leakage from the atmosphere through the O-ring and the adhesive into the
102
down stream was also shown to be trivial [13, 20]. Both the upstream and downstream
were evacuated for at least 24 hours to remove any residue gas or vapor trapped in the
membrane and the cell before testing. Permeation test were carried out by introducing
desired gas to the upstream of the membranes and closing the downstream valve to
record the gas pressure increase due to the permeation through the membranes. The
gas-of-interest was controlled at desired temperature and pressure. The measurement
involves maintaining a constant pressure of the permeating gas on one side of the film
(the upstream) and measuring the flux through the film on the other side (downstream
or permeate side). The flux is measured by monitoring the pressure rise due to the gas
permeating through the membrane in a constant known volume.
From gas supply
To vacuum pump
C3
DS
C5
Feed valve
Permeation Cell
To atmosphere
C2C1
US
C4
Gas reservoir
PG
PT
INT-DS
Release valve
US: Upstream; DS: Downstream; PT: Pressure transducer; PG: Pressure gauge; INT-DS: Internal downstream valve; C1, 2, 3, 4, 5: Valves
Figure 3.6: Schematic diagram of the flat dense membrane gas permeation testing
apparatus
103
Figure 3.7 Schematic diagram of the double O-ring permeation cell for the flat dense membranes
A feed ballast volume of approximately 1 liter and a Swagelok® valve is used on the
upstream side to prevent or reduce any pressure fluctuations of the gases during the
experiment. The pressure is measured using a 0-500psia pressure transducer (Ashcroft).
On the permeate side, the pressure is measured using a 0-10Torr pressure transducer
(MKS instrument PDR-C-1C Power supply readout & MKS Technology and
productivity Baratron® capacitance Manometer). A digital readout connected to IOtech
Data shuttle (High-resolution data acquisition) to record the pressure rise. The
temperature in the permeation cell was controlled by a Temperature controller and
heater (Eurotherm), a heating tape inside the chamber and a cooling blower (TOYO).
104
The plumbing was all stainless steel, with Swagelok® valves and VCR metal face seal
connections. The downstream accumulator volume was known and calibrated.
Pure gas permeability of flat dense membranes was determined by a constant volume
method reported previously [19]. The permeability was obtained in the sequence of He,
H2, O2, N2, CH4 and CO2 at 35°C and 10atm (except for H2 tested at 3.5atm). For the
CO2 plasticization study, the CO2 permeability was tested with variable upstream
pressures from 3.5atm to 32atm at a constant temperature of 35°C. The gas
permeability P was determined from the rate of pressure increase (dp/dt) at the steady
state using equation (3.5):
)()7.14/)76((760
1015.273
0
10
dtdp
pATVlP
××= (3.5)
where P is the permeability of a membrane to a gas in Barrer (1 Barrer = 1 x 10-10 cm3
(STP)-cm/cm2-s-cmHg), V is the volume of the down-stream chamber (cm3), A is the
effective area of the membrane (cm2), l is the membrane thickness (cm), T is the
experimental temperature (K), dp/dt is the rate of pressure increase measured by a
pressure sensor in the down-stream chamber (mmHg/s) and the pressure of the feed
gas in the up-stream is given by p0 in psia. The ideal selectivity of a flat dense
membrane for gas A to gas B is defined as follows:
αA,B = PA/PB (3.6)
The experimental error in the permeability data was estimated to be around ±2% and
this error would be smaller if the gas permeability was higher. The percentage error
105
was estimated by comparing the data with the nominal values obtained from a
reference membrane. The gas permeation test of the reference membrane was
conducted each time before and after the testing of the flat dense membranes used in
this study. The experiment on each piece of the membrane was repeated for three times.
The volumes of the downstream compartments and the measured thickness of the
membrane were likely to contribute most to the overall experimental error.
3.6.1.3 Mixed Matrix Hollow Fibers
The pure gas permeation for the mixed matrix hollow fibers was measured by a
constant volume method described in the literature [19] with some modifications in
order to be suitable for the hollow fiber testing. The modified permeation cell is
schematically shown in Figure 3.8. Hollow fiber modules were prepared containing 1
fiber with the effective length of 5cm, and at least five modules were tested at 35oC as
an order of O2, N2, CH4 and CO2 for each spinning condition. Both the upstream and
downstream were evacuated for at least 24 hours lo remove any residue gas or vapor
trapped in the membrane. Permeation tests were carried out by introducing the desired
gas to the shell side of the hollow fibers. The testing pressure ranged from 2-10atm to
prevent the CO2 plasticization.
The permeation rate was calculated from the steady-state pressure increase as a
function of time dp/dt (mmHg/s) using the equation (3.7):
106
)()7.14/)76((760
1015.273
0
10
dtdp
pATV
LP
××
= (3.7)
in which, P/L is the permeance of a membrane to a gas with the unit of GPU (1 GPU =
1 x 10-6 cm3 (STP)/cm2-s-cmHg), V is the volume of the down-stream chamber (cm3),
A is the effective area of the membrane (cm2), T is the experimental temperature (K),
dp/dt is the rate of pressure increase measured by a pressure sensor in the down-stream
chamber (mmHg/s) and the pressure of the feed gas in the up-stream is given by p0 in
psia. The ideal selectivity of a mixed matrix hollow fiber module for gas A to gas B is
defined as follows:
B
ABA LP
LP]/[]/[
/ =α (3.8)
From gas supply
To vacuum pump
C3
DS
C5
Feed valve
Modified Permeation Cell
To atmosphere
C2C1
US
C4
Gas reservoir
PG
PT
INT-DS
Release valve
EpoxyEpoxy
Hollow fibermodule in union tee
Upstream
Downstream
US: Upstream; DS: Downstream; PT: Pressure transducer; PG: Pressure gauge; INT-DS: Internal downstream valve; C1, 2, 3, 4, 5: Valves
Figure 3.8: Schematic diagram of gas permeation testing apparatus for the mixed
matrix hollow fibers
107
3.6.2 Mixed Gas Permeation Test
3.6.2.1 Neat Polymeric hollow fibers
In the mixed gas permeation measurement of the neat polymeric hollow fibers, it is
necessary to measure the composition of the streams in addition to the flux. Figure 3.9
depicts the schematic diagram of the mixed gas permeation testing apparatus for the
neat polymeric hollow fibers designed by our senior research group members [17]. The
compositions of the permeate and retentate were analyzed by a gas chromatograph
system (Hewlett Packard (HP) 6890 Series GC) using a Carboxen 1010 column and a
thermal conductivity detector (TCD), while their flow rates were measured by a digital
bubble flow meter (Opliflow 570, error ~2%). The calculation of permeance was based
on one set of differential equations developed by Wang et al. [21], in which the non-
ideal gas behavior and pressure drop inside the hollow fibers along the permeator have
been considered.
The testing procedure is a standard one in our lab. First, the feed pressure was set at 50
psia and the fibers were conditioned for 1.5 hours; for the following pressures, the
conditioning time was at least 30 minutes. In each pressure, the flow rate of both
permeate and retentate were first determined and adjusted by the bubble flow meter,
thus achieving a stage cut around 5%. Thereafter, the permeate and/or the retentate
were connected to the GC for sampling of 15 minutes with the gas mixture being
carried to the GC system by the inert gas helium. Subsequently, the composition of the
108
mixture was determined by calculating the peak area for each gas that was eluted out
from the column at different retention time. Finally, the composition of the feed gas
from gas cylinder was also tested with the GC system.
Figure 3.9: Apparatus of the mixed gas permeation test in neat polymeric hollow fibers
3.6.2.2 Flat Dense Neat Polymeric or Mixed Matrix Membranes
For binary gas permeation measurements of flat dense membranes, the pure gas
permeation cell was modified for the continuous flowing feed by connecting to a
needle valve to control the upstream pressure as shown in Figure 3.10. Besides, it was
also equipped with a retentate channel to avoid the accumulation of feed gas at the
upstream. Then the receiving volume of the permeation cell was connected to a gas
109
chromatograph (GC) system by a valve (C6). The mixed gas system, with combination
of permeation cell and gas chromatograph allows straightforward determination of gas
permeability, which is similar to the techniques used by O’Brien et al. [22]. The GC
used was a Hewlett Packard (HP) 6890 Series GC completed with a HP 5973 Mass
Selective Detector. Before permeation testing, GC was calibrated by one set of Messer
gas mixtures of CO2 and CH4 with known composition to obtain the GC peak area as a
function of gas mole fraction.
The sampling process was initiated by evacuating the line from the receiving volume
(the lower chamber) to GC by vacuum pump (P1). The compositions of the feed and
permeate were analyzed by the GC. When the permeation rate reached steady state,
and pressure of gas collected in the downstream volume approaching 100Torr, the
valve C6 was opened to allow the permeate gas to expand into the line connecting to
GC. Then, the valve C6 is closed again in order to inject the permeate gas to GC for
composition analysis. The permeability was calculated with the consideration of non-
ideal gas behaviour, described by Wang and coworkers [21]. Similar to the pure gas
permeability, the mixed gas steady state permeation rate were then determined by
following two equations:
)())(
7.1476(760
10273
0
10
2
2
2 dtdp
pxAT
VLyP
CO
COCO
×= (3.9)
)()])1)[(
7.1476(
)1(760
10273
0
10
2
2
4 dtdp
pxAT
VLyP
CO
COCH
−
−×= (3.10)
110
where PCO2 and PCH4 are the permeability of a membrane to gas CO2 and CH4,
respectively, p0 is the upstream feed gas pressure (psia), xCO2 is the CO2 molar fraction
in the feed gas and yCO2 is the CO2 molar fraction in the permeate. The mole fraction
yCO2 was determined by batch sampling from the receiving volume. Subsequently, the
separation factor of mixed gas is characterized by the ratio of downstream (y) and
upstream (x) mole fraction of two components (CO2 and CH4) by the following
equation:
)/()/(
42
42
42 /CHCO
CHCO
CHCO xxyy
=α (3.11)
In the case of negligible downstream pressure, the separation factor (αA/B) is equal to
the ideal separation factor (α*A/B ) that measured the intrinsic selectivity of membrane
material. Therefore, the selectivity of the membrane for gas CO2 to gas CH4 is thus
computed by the following equation which is equivalent to the ratio of their
permeability measured at the partial pressure:
4
2
42 /CH
CO
CHCO PP
=α (3.12)
111
Figure 3.10: Schematic diagram of mixed gas permeation test apparatus for flat dense membranes
112
3.6.2.3 Mixed Matrix Hollow Fibers
The mixed gas permeation measurements of the mixed matrix hollow fibers will be
discussed in Appendix B, so it is not necessary to duplicate it here.
3.7 REFERENCES
[1] J.S. Chiou, Y. Maeda, D.R. Paul, Gas permeation in polyethersulfone, J. App.
Polym. Sci. 33 (1987) 1823.
[2] J.N. Barsema, G.C. Kapantaidakis, N.F.A. van der Vegt, G.H. Koops, M. Wessling,
Preparation and characterization of highly selective dense and hollow fiber
asymmetric membranes based on BTDA-TDI/MDI co-polyimide, J. Membr. Sci.,
216 (2003) 195.
[3] D.L. Wang, W.K. Teo, K. Li, Preparation and characterization of high-flux
polysulfone hollow fibre gas separation membranes, J. Membr. Sci., 204 (2002)
247.
[4] T.S. Chung, E.R. Kafchinski, The effects of spinning conditions on asymmetric
6FDA/6FDAM polyimide hollow fibers for air separation, J. Appl. Polym. Sci., 65
(1997) 1555.
[5] D.T. Clausi, W.J. Koros, Formation of defect-free polyimide hollow fiber
membranes for gas separations, J. Membr. Sci., 167 (2000) 79.
113
[6] B. Bikson, J.K. Nelson, Composite membranes and their manufacture and use, US
Patent 4,826,599 (1989).
[7] A.Z. Gollan, Anisotropic membranes for gas separation, US patent: 4,681,605
(1987).
[8] T.S. Chung, A review of microporous composite polymeric membrane technology
for air-separation, Polym. and Polym. Comp., 4 (1996) 269.
[9] F.S. Model, L.A. Lee, in: H.K. Lonsdale, H.E. Podall (Eds.), Reverse Osmosis
Membrane Research, Plenum Press, New York, 1972, p. 285.
[10] I. Cabasso, A.P. Tamvakis, Composite hollow fibre membranes, J. of Appl.
Polym. Sci., 23 (1979) 1509.
[11] E.P. Plueddemann, Silane coupling agents, 2nd ed., Plenum Press, New York,
1991.
[12] R. Mahajan, R. Burns, M. Schaeffer, W.J. Koros, Challenges in forming
successful mixed matrix membranes with rigid polymeric materials, J. App. Polym.
Sci., 86 (2002) 881.
[13] R. Mahajan, Formation, characterization and modeling of mixed matrix
membrane materials, Ph.D. Dissertation, The University of Texas at Austin, 2000.
[14] L.Y. Jiang, T.S. Chung, S. Kulprathipanja, An investigation to revitalize the
separation performance of hollow fibers with a thin mixed matrix composite skin
for gas separation, J. Membr. Sci., in press.
[15] R. Sirnha, R.F. Bayer, General relation involving the glass temperature and
coefficients of expansion of polymers, J. Chem. Phys. 37 (1962) 1003.
114
[16] A. Eisenberg, B. Hird, R.B. Moore, A new multiplet-cluster model for the
morphology of random ionomers, Macromolecules 23 (1990) 4098.
[17] C. Cao, R. Wang, T.S. Chung, Y. Liu, Formation of high-performance 6FDA-2,6-
DAT asymmetric composite hollow fiber membranes for CO2/CH4 separation, J.
Membr. Sci., 209 (2002) 309.
[18] D.F. Li, T.S. Chung, R. Wang, Y. Liu, Fabrication of fluoropolyimide/
polyethersulfone (PES) dual-layer asymmetric hollow fiber membranes for gas
separation, J. Membr. Sci., 198 (2002) 211.
[19] W.H. Lin, R.H. Vora, T.S. Chung, Gas transport properties of 6FDA-durene/1,4-
phenylenediamine (pPDA) copolyimides, J. Polym. Sci. Polym. Phys., 38 (2000)
2703.
[20] T.T. Moore, W.J. Koros, Characterization of low permeability gas separation
membranes and barrier materials: design and operation considerations, J. Membr.
Sci., 245 (2004) 227.
[21] R. Wang, S.L. Liu, T.T. Lin, T.S. Chung, Characterization of hollow fiber
membranes in a permeator using binary gas mixtures, Chem. Eng. Sci., 57 (2002)
967.
[22] K.C. O'Brien, W.J. Koros, T.A. Barbari, E.S. Sanders, A new technique for the
measurement of multicomponent gas transport through polymer films, J. Membr.
Sci., 29 (1986) 229.
115
CHAPTER FOUR
FABRICATION OF DUAL-LAYER POLYETHERSULFONE (PES) HOLLOW
FIBER MEMBRANES WITH AN ULTRATHIN DENSE SELECTIVE LAYER
FOR GAS SEPARATION
4.1 INTRODUCTION
In the last 20 years, the asymmetric single-layer hollow fiber membrane is always a
favorable configuration in the membrane-based systems for gas separation [1-6]. The
silicone rubber technology is proven an effective means to seal surface defects for gas
separation membranes. To further expand membranes for industrial applications, one
must find ways to enhance the gas permeation flux (productivity) and permselectivity
by combining the synthesis of high-performance polymeric membrane materials with
the innovation of membrane fabrication technology.
Making the dense selective layer of asymmetric hollow fiber membranes thinner is
one of the most effective approaches to increase the gas permeation flux. With the
development of technology, how to fabricate single-layer hollow fiber membranes
with an ultrathin dense selective layer has been no longer a challenge [7-15]. Excellent
and informative reports to prepare single-layer hollow fiber membranes with an
ultrathin dense selective layer have been published by researchers at Permea [2, 7-9],
116
at Prof. Koros’ group at the University of Texas [10-12], and at National U of
Singapore [13-17]. They concluded that the keys to fabricate hollow fiber membranes
with an ultrathin dense selective layer are 1) to control the chemistry of the internal
coagulant and the bore-fluid flow rate, and 2) to have a dope exhibiting significant
chain entanglement.
A significant advance on polymeric materials for gas separation has also been made in
the last 20 years [18-21], many high-permeability and high permselectivity materials
have been discovered and synthesized. However, these high performance polymeric
materials are often very expensive, while some of them are brittle. As a result, the
fabrication of integrally-skinned asymmetric membranes is either no longer feasible or
economically attractive because it is too costly to prepare the entire membrane from
the same material. One of potential solutions to overcome these problems is to
fabricate dual-layer hollow fiber membranes by the co-extrusion technology. The
dual-layer hollow fiber membrane basically consists of an asymmetric separating outer
layer and a microporous supporting inner layer. The main function of the outer layer is
to provide the permselectivity; therefore it should be made of a high-permeability and
high-selectivity polymeric material. The main function of the inner layer is to provide
the necessary mechanical support for the outer layer; therefore it may be made of low-
cost polymers with good mechanical properties. Ideally, the interface between the two
layers and the bulk structure of the inner layer must be porous in order to have a
minimal gas transport resistance; otherwise the sub-structure resistance deduced from
117
the interface and the supporting layer may result in significant decreases in both
permeance and permselectivity [22]. In addition, the dual-layer hollow fiber
membranes must be delamination-free at the interface and have a D/∆h (dual-layer
fiber diameter/wall thickness) ratio less than 2 in order to withstand high-pressure feed
streams.
There have been a few pioneer investigations on dual-layer hollow fiber membranes
since late 1980’s [23-31]. Ekiner et al’s early work at DuPont is informative because
the membranes developed are delamination-free at the interface and has good gas
separation performance [25]. Suzuki et al [27] also fabricated dual-layer hollow fiber
membranes composed of a dense polyimide outer layer and a sponge-like inner layer
made of another polyimide. However, their membranes may not withstand high-
pressure environments because of high D/∆h ratios and their thinnest dense-selective
layer is about 0.95µm. Li et al. [30] conducted the first systematic study to investigate
the effects of spinning conditions on dual-layer hollow fiber membranes and the
causes of interfacial delamination between the two layers. They concluded that the
inner-layer dope concentration, bore fluid composition and subsequent solvent
exchange play important roles to produce delamination-free dual-layer membranes
because the delamination is resulted from the difference in shrinkage rate between the
two layers during the precipitation. Nevertheless, their dense-selective layer thickness
is around 8500Å which is still too high when compared with those of single-layer
asymmetric membranes. Jiang et al extended Li et al’s work and developed almost
118
defective-free Matrimid/polyethersulfone dual-layer hollow fiber membranes with a
dense-selective layer thickness of 2886Å [32] by controlling the dope composition and
phase-inversion conditions. Later, Li et al modified the dope compositions and
demonstrated the dual-layer hollow fiber membranes with a dense-selective layer
thickness of about 1500Å [33].
To fully compete and potentially replace the single-layer hollow fiber membranes, one
must develop the co-extruding technology to produce dual-layer hollow fiber
membranes with a dense-selective layer thickness of less than 1000Å. To our best
knowledge, so far there is no academic literature available on this subject. Therefore,
the objective of this study is to fabricate dual-layer hollow fiber membranes with an
ultra-thin dense selective layer. For concept demonstration and the elimination of
interfacial delamination, polyethersulfone (PES) was chosen as the material for both
outer and inner layers because of its superior mechanical and membrane forming
characteristics. However, the spinning dopes for outer and inner layers have different
polymer and solvent/non-solvent compositions. The O2/N2 selectivity of PES was
reported to be 6.1 with an O2 permeability of 0.51 Barrer at 30oC and to be 5.1 with an
O2 permeability of 0.81 Barrer at 50oC [34]. By using an Arrhenius relationship
between permeability and temperature, the O2 permeability and O2/N2 selectivity at
25oC can be calculated to be around 0.44 Barrer and 6.3 [13-15], respectively.
119
4.2 RESULTS AND DISCUSSION
4.2.1 The Effect of Different Post-treatment Protocols on Gas Separation
Performance
The dual-layer hollow fiber membranes were fabricated by the dry-jet wet-spinning
technique using a dual-layer spinneret as shown in our previous literature [30] and the
flow rates of the bore fluid and both dope solutions were controlled by three ISCO
pumps. The spinning parameters were listed in Table 4.1.
Table 4.1: Spinning parameters of dual-layer PES hollow fiber membranes
1.5
2525
Tap water 307 (keep free falling)
0.3
NMP/H2O: 95/5 in wt.%
0.6
PES/NMP/EtOH: 25/61/14 in wt.%
0.2
PES/NMP/EtOH: 35/50/15 in wt.%Values
Inner-layer dope composition
Inner-layer dope flow rate (ml/min)
Take up rate (cm/min)
Spinning temperature (oC)
Air gap (cm)
External coagulant
Coagulation bath temperature (oC)
Bore fluid flow rate (ml/min)
Bore fluid composition
Out-layer dope flow rate (ml/min)
Outer-layer dope compositionParameters
1.5
2525
Tap water 307 (keep free falling)
0.3
NMP/H2O: 95/5 in wt.%
0.6
PES/NMP/EtOH: 25/61/14 in wt.%
0.2
PES/NMP/EtOH: 35/50/15 in wt.%Values
Inner-layer dope composition
Inner-layer dope flow rate (ml/min)
Take up rate (cm/min)
Spinning temperature (oC)
Air gap (cm)
External coagulant
Coagulation bath temperature (oC)
Bore fluid flow rate (ml/min)
Bore fluid composition
Out-layer dope flow rate (ml/min)
Outer-layer dope compositionParameters
Table 4.2 summarizes the gas separation performance and the apparent dense-selective
layer thickness of dual-layer hollow fiber membranes with different post-treatment
protocols. Before the silicone rubber coating, it can be found that the permeance
120
decreases with an increase in heat-treatment temperature. This is due to the fact that
the heat-treatment induces relaxation of the stresses imposed in hollow fibers when
they were fabricated. Hollow fibers shrink inwards gradually (i.e., reduced outer
diameter), leading to higher packing density of polymer chains. Therefore, the heat-
treatment not only reduces the surface defects but also the bulk porosity. However, all
membranes with or without post-treatment exhibit an O2/N2 selectivity of 0.93-0.94,
indicating that Knudsen diffusion is still dominant and the surface pore size cannot be
not reduced to follow the sorption-diffusion mechanism solely by the heat-treatment.
Table 4.2: Gas separation performances of dual-layer PES hollow fiber membranes with different post-treatment protocols
1089
3.54
1.14
4.04
0.93
93.6
86.8
Heat-treated fibers with Condition B
(150oC for 1h)
407
6.00
1.80
10.8
0.93
422
394
Heat-treated fibers with Condition A
(75oC for 3h)
389
5.26
2.15
11.3
0.94
544
513
As-spun fibers
Skin thickness (Å)
Ideal selectivity
Permeance (N2) (GPU)
Permeance (O2) (GPU)
After coating2
Ideal selectivity
Permeance (N2) (GPU)
Permeance (O2) (GPU)
Before coating1
Dual-layer hollow fiber membrane name
1089
3.54
1.14
4.04
0.93
93.6
86.8
Heat-treated fibers with Condition B
(150oC for 1h)
407
6.00
1.80
10.8
0.93
422
394
Heat-treated fibers with Condition A
(75oC for 3h)
389
5.26
2.15
11.3
0.94
544
513
As-spun fibers
Skin thickness (Å)
Ideal selectivity
Permeance (N2) (GPU)
Permeance (O2) (GPU)
After coating2
Ideal selectivity
Permeance (N2) (GPU)
Permeance (O2) (GPU)
Before coating1
Dual-layer hollow fiber membrane name
1. Tested at 25 oC, 70 psi2. Tested at 25 oC, 200 psi
After dip-coating in the silicone rubber solution, it can be found that the as-spun
membranes have an O2/N2 selectivity of 5.26 which is 83% of the intrinsic selectivity
121
(i.e., 6.3) of PES. The apparent dense-selective layer thickness is 389Å calculated
from Equation (3). The O2/N2 selectivity increases impressively to 6.0 after heat-
treating the dual-layer hollow fiber membranes at the Condition A (i.e., 75oC for 3
hours). This selectivity is very close to the intrinsic selectivity of 6.3. In addition, the
permeance only dips slightly to 10.8GPU after heat-treatment at 75oC. This permeance
is equivalent to an apparent dense-selective layer of 407Å, which is the thinnest one
ever reported in the literature for the dual-layer hollow fiber membranes and is also
comparable to that of state-of-the-art single-layer asymmetric hollow fibers.
Further increasing the heat-treatment temperature to 150oC not only lowers the O2/N2
selectivity to 3.54, but also thickens the apparent dense-selective layer to 1089Å. This
phenomenon may be arisen from the fact that a high-temperature heat-treatment may
induce densification of both selective skin and substructure. The densification of a
selective skin may result in a reduction of permeance, while the densification of the
non-selective substructure is manifested by a lower selectivity. This is because the
Knudsen diffusion controls the gas transport in the substructure, whereas the sorption-
diffusion dominates the gas transport in the selective skin. Therefore, when the
resistance in the substructure cannot be neglected compared with the resistance in the
selective skin after the densification, the Knudsen diffusion may worsen the resultant
selectivity in the whole asymmetric membranes.
122
4.2.2 Membrane Morphology
Figure 4.1 shows a typical morphology of the as-spun dual-layer hollow fiber
membrane produced in this study. Its outer diameter is around 560µm and inner
diameter is around 290µm. Because of using the same PES and similar solvents in
both layers, it significantly enhances the interfacial diffusion as pointed out in our
previous paper [32], thus no clear delamination can be observed between the inner and
outer layers as shown in Figures 4.1B and 4.1C. Based on the mass balance calculation
using the information provided in Table 1, the thickness of the outer layer is less than
15µm. The outer layer has an asymmetric structure, while the inner layer is fully
porous. In addition, the bulk of the inner layer (Figure 4.1D) and the inner surface of
the inner layer (Figure 4.1E) are fully porous, while the outer skin of the outer layer
(Figure 4.1F) is fully dense.
123
D E F
A B CA B C
Figure 4.1: Integrity of as-spun dual-layer hollow fiber membranes (A: Overall profile; B: Membrane wall; C: Outer layer’s outer edge; D: Inner layer’s
inner bulk; E: Inner layer’s inner skin; F: Outer layer’s outer skin)
There exist multiple-layer macrovoids in the inner layer as shown in Figure 4.1B and
the macrovoid structures of these two layers are somewhat different. The macrovoids
near the outer layer have a finger-like structure and have a smooth and less porous
skin, meanwhile, most pores are created at the bottom of these macrovoids; while the
macrovoids near the inner cavity of the fiber show a structure more like tear drops
than fingers and have a fully porous skin interconnecting with surrounding spongy
cells. The macrovoid formation is a common phenomenon in phase-inversion
membranes; however, its mechanisms are very complicated and have been heavily
debated in the last 40 years [5, 32-33, 35-42]. It is generally accepted that fast
precipitation with the aid of unbalanced localized stresses and solvent diffusion favors
124
the formation of finger like macrovoids through the nucleation and growth of the
polymer lean phase as well as polymer rich phase and the spinodal separation.
The multiple-layer macrovoids can often be observed in fibers spun with low polymer
concentrations possibly due to a rapid solvent exchange. The finger-like macrovoids
near the outer layer start just beneath the outer skin of inner layer as illustrated in
Figures 4.1B and 4.1C. This phenomenon may suggest that the external strong
coagulant, that is tap water, can move into the thin outer layer dope solution during the
formation of nascent membranes and result in the formation of macrovoids in the inner
layer with the aid of unbalanced localized stresses and rapid solvent diffusion. The
tear drop-like macrovoids near the inner cavity of the fiber may be induced by the
combination of the finger-like macrovoids with a slower precipitation deduced by the
internal coagulant. The structural difference between two types of macrovoids may be
resulted from different coagulation rates and solvent exchange rates. The external
coagulant, tap water, is a strong coagulant so that there is no enough time for
macrovoids near the outer layer to propagate before the precipitation and is also such a
powerful penetrant that it can diffuse downwards through a long distance in a short
time; therefore macrovoids near the outer layer form a long and narrow finger-like
structure. In contrast, macrovoids near the inner cavity of the fiber have plenty of time
to relax outwards because the internal coagulant is composed of 95/5 (in wt%)
NMP/H2O, but the penetration rates of the coagulant greatly decrease simultaneously;
thus leading to a short and wide tear drop-like structure.
125
Figure 4.2: Visual estimation of the dense-selective layer thickness of dual-layer
hollow fiber membranes with different heat-treatment methods. (left: Condition A; right: Condition B)
Figure 4.2 displays the cross-section morphology near the outer edge of the outer layer
after various heat-treatments. Interestingly, the calculated thicknesses of the apparent
dense-selective layers are in agreement with the SEM observation. They are
approximately 40 and 110nm for membranes heat-treated at Conditions A and B,
respectively. It can also be found from this figure that the condition B significantly
reduces the size and the amount of pores immediately beneath the dense-selective
layer because of the high heat-treatment temperature.
126
A B C
D E F
Figure 4.3: A comparison of the cross-section morphology of dual-layer hollow fiber membranes with different post-treatment protocols
(A, D: as-spun; B, E: Condition A; C, F: Condition B)
Figure 4.3 exhibits a comparison of cross-section morphology of dual-layer hollow
fiber membranes with different post-treatment protocols. 3 or 4 SEM samples were
prepared for each protocol to obtain the high accuracy and consistence.
Macroscopically, there is no sharp difference in the overall cross-section morphology
(Figures 4.3A, 4.3B and 4.3C) except that membranes treated with a higher
temperature (Condition B) apparently have slightly smaller pores and less slim finger-
like macrovoids. Future studies will be focused on if a higher heat-treatment
temperature can lead to the disappearance of small-size macrovoids.
127
as-spun
Condition A
Condition B
Figure 4.4: Pore size comparison of the inner layer’s inner skin of dual-layer hollow fiber membranes with different post-treatment protocols (left: as-spun; middle: Condition A; right: Condition B)
Figures 4.3D, 4.3E and 4.3F show the substructure morphology underneath the outer
layer, while Figure 4.4 shows a comparison of porous structure in the inner surface of
the inner layer as a function of post-treatment temperature. The as-spun fibers and the
fiber heat-treated at Condition A (i.e., 75oC) apparently have similar substructure
morphology and inner surface pores, while the fiber heat-treated at Condition B (i.e.,
150oC) has denser substructure and smaller pores in the inner surface. This
observation is consistent with our previous explanation why both permeance and
selectivity decrease for membranes heat-treated at Condition B (i.e., 150oC). Based on
the Resistance model [43], an increased substructure resistance decreases the
permeance of a fast gas more than that of a slow gas. As a result, the overall selectivity
also decreases.
128
4.3 CONCLUSIONS
Dual-layer PES hollow fiber membranes with an ultrathin dense-selective layer of
407Å have successfully been fabricated by using co-extrusion and dry-jet wet-
spinning phase inversion techniques with the aid of heat-treatment. The effects of
heat-treatment on membrane morphology and gas separation performances have also
been investigated. The following conclusions can be drawn from this chapter:
1) Heat-treatment at 75oC for 3 hours with the aid of silicone rubber coating can
effectively improve gas separation performance. The as-spun dual-layer hollow
fiber membranes have an apparent dense-selective layer thickness of 389Å and
an O2/N2 selectivity of 5.26 after silicone coating. The O2/N2 selectivity
increases to 6.00 after heat treatment at 75oC, while the apparent dense-
selective layer thickness increases to 407Å.
2) Heat-treatment at 150oC for 1 hour can lower both selectivity and permeance
values due to the enhanced substructure resistance induced by the heat-
treatment. SEM observation confirms this phenomenon.
3) Macrovoids formed just beneath the outer skin of inner layer have a long and
narrow finger-like structure, while macrovoids near the inner cavity of the
fiber have a short and wide tear drop-like structure. The structural difference
between these two types of macrovoids may be resulted from different
coagulation rates and solvent exchange rates.
129
4.4 REFERENCES
[1] W.J. Koros, I. Pinnau, Membrane formation for gas separation process, in
“Polymeric Gas Separation Membranes”, Chapter 5, ed. by D.R. Paul and Y.P.
Yampol’skii, CRC Press, Boca Raton, 1994.
[2] R.E. Kesting, A.K. Fritzsche, Polymeric Gas Separation Membranes, John Wiley
and Sons, New York, NY, 1993.
[3] I.M. Wienk, R.M. Boom, M.A.M. Beerlage, A.M.W. Bulte, C.A. Smolders, H.
Strathmann, Recent advances in the formation of phase inversion membranes
made from amorphous or semi-crystalline polymers, J. Membr. Sci., 113 (1996)
361.
[4] T.S. Chung, A review of microporous composite polymeric membrane technology
for air-separation, Polym. and Polym. Comp., 4 (1996) 269.
[5] H. Strathmann, K. Kock, P. Amar, R.W. Baker, The formation mechanism of
asymmetric membranes, Desalination, 16 (1975) 179.
[6] W.S.W. Ho, K.K. Sirkar, Membrane Handbook, Ed., Van Nostrand Reinhold, New
York, 1992.
[7] R.E. Kesting, A.K. Fritzsche, M.K. Murphy, C.A. Cruse, A.C. Handermann, R.F.
Malon, M.D. Moore, The second-generation polysulfone gas-separation membrane,
I: The use of Lewis acid : base complexes as transient templates to increase free
volume, J. Appl. Polym. Sci., 40 (1990) 1557.
130
[8] R.E. Kesting, A.K. Fritzsche, M.K. Murphy, A.C. Handermann, C.A. Cruse, R.F.
Malon, Process for forming asymmetric gas separation membranes having graded
density skins, U.S. Patent 4,871,494, 1989.
[9] A.K. Fritzsche, M.K. Murphy, C.A. Cruse, R.F. Malon, R.E. Kesting,
Characterization of asymmetric hollow fiber membranes with graded-density skin,
Gas Sep. Purif., 3 (1989) 106.
[10] I. Pinnau, W.J. Koros, Defect-free ultrahigh flux asymmetric membranes, U.S. Patent
4,902,422 (1990)
[11] S.C. Pesek, W.J. Koros, Aqueous quenched asymmetric polysulfone hollow
fibers prepared by dry/wet phase separation, J. Membr. Sci., 88 (1994) 1.
[12] D.T. Clausi, W.J. Koros, Formation of defect-free polyimide hollow fiber
membranes for gas separations, J. Membr. Sci., 167 (2000) 79.
[13] T.S. Chung, E.R. Kafchinski, P. Foley, Development of asymmetric hollow
fibers from polyimide for gas separation, J. Membr. Sci., 75 (1992) 181.
[14] D.L. Wang, K. Li, W.K. Teo, Polyethersulfone hollow fiber gas separation
membranes prepared from NMP/alcohol systems, J. Membr. Sci., 115 (1996) 85.
[15] D.L. Wang, W.K. Teo, K. Li, Preparation and characterization of high-flux
polysulfone hollow fibre gas separation membranes, J. Membr. Sci., 204 (2002)
247.
[16] T.S. Chung, S.K. Teoh, X.D. Hu, Formation of ultrathin high-performance
polyethersulfone hollow-fiber membranes, J. Membr. Sci., 133 (1997) 161.
131
[17] J.J. Qin, T.S. Chung, Effects of orientation relaxation and bore fluid chemistry
on morphology and performance of polyethersulfone hollow fibers for gas
separation, J. Membr. Sci., 229 (2004) 1.
[18] W.J. Koros, G.K. Fleming, S.M. Jordan, T.H. Kim, H.H. Hoehn, Polymeric
membrane materials for solution-diffusion based permeation separations, Prog.
Polym. Sci., 13 (1988) 339.
[19] S.A. Stern, Polymers for gas separations: the next decade, J. Membr. Sci., 94
(1994) 1.
[20] L.M. Robeson, W.F. Burgoyne, M. Langsam, A.C. Savoca, C.F. Tien, High
performance polymers for membrane separation, Polymer, 35 (1994) 4970.
[21] B.D. Freeman, Basis of permeability/selectivity tradeoff relations in polymeric
gas separation membranes, Macromolecules, 32 (1999) 375.
[22] I. Pinnau, W.J. Koros, Relationship between substructure resistance and gas
separation properties of defect-free integrally-skinned asymmetric membranes,
Ind. Eng. Chem. Res., 30 (1991) 1837.
[23] E. Kuzumoto, K. Nitta, Manufacture of permselective hollow membranes,
Japanese Patent 01,015,104 (1989).
[24] T. Yoshinaga, H. Shimazaki, Y. Kusuki, Polyimide bilayer hollow membranes
and their manufacture, Japanese Patent 02,169,019 (1990).
[25] O.M. Ekiner, R.A. Hayes, P. Manos, Novel multicomponent fluid separation
membranes, US Patent 5,085,676, (1992).
132
[26] S. Nago, Y. Mizutani, Microporous polypropylene hollow fibers with double
layers, J. Membr. Sci., 116 (1996) 1.
[27] H. Suzuki, K. Tanaka, H. Kita, K. Okamoto, H. Hoshino, T. Yoshinaga, Y.
Kusuki, Preparation of composite hollow fiber membranes of poly (ethylene
oxide)-containing polyimide and their CO2/N2 separation properties, J. Membr.
Sci., 146 (1998) 31.
[28] S.H. Yang, W.K. Teo, K. Li, Formation of annular hollow fibers for
immobilization of yeast in annular passages, J. Membr. Sci., 184 (2001) 107.
[29] T. He, M.H.V. Mulder, H. Strathmann, M. Wessling, Preparation of composite
hollow fiber membranes: co-extrusion of hydrophilic coating onto porous
hydrophobic support structures, J. Membr. Sci., 207 (2002) 143.
[30] D.F. Li, T.S. Chung, R. Wang, Y. Liu, Fabrication of fluoropolyimide/
polyethersulfone (PES) dual-layer asymmetric hollow fiber membranes for gas
separation, J. Membr. Sci., 198 (2002) 211.
[31] C.C. Pereira, R. Nobrega, K. V. Peinemann, C.P. Borges, Hollow fiber
membranes obtained by simultaneous spinning of two polymer solutions: a
morphological study, J. Membr. Sci., 226 (2003) 35.
[32] L.Y. Jiang, T.S. Chung, D.F. Li, C. Cao, S. Kulprathipanja, Fabrication of
Matrimid/Polyethersulfone dual-layer hollow fiber membranes for gas separation,
J. Membr. Sci., 240 (2004) 91.
133
[33] D.F. Li, T.S. Chung, R. Wang, Morphological aspects and structure control of
dual-layer asymmetric hollow fiber membranes formed by a simultaneous co-
extrusion approach, J. Membr. Sci., 243 (2004) 155.
[34] D.L. Wang, Polyethersulfone hollow fiber gas separation membranes prepared
from solvent systems containing nonsolvent-additives, Ph.D dissertation,
National U. of Singapore, 1996.
[35] L. Broens, F.W. Altena, C.A. Smolders, D.M. Koenhen, Asymmetric membrane
structures as a result of phase separation phenomena, Desalination, 32 (1980) 33.
[36] H. Strathmann, K. Kock, The formation mechanism of phase inversion
membranes, Desalination, 21 (1977) 241.
[37] A.J. Reuvers, C.A. Smolders, Formation of membranes by means of immersion
precipitation. Part ІІ: The mechanism of formation of membranes prepared from
the system cellulose acetate-acetone-water, J. Membr. Sci., 34 (1987) 67.
[38] C.A. Smolders, A.J. Reuvers, R.M. Boom, I.M. Wienk, Microstructures in
phase-inversion membranes, 1. Formation of macrovoids, J. Membr. Sci., 73
(1992) 259.
[39] S.S. Shojaie, W.B. Krantz, A.R. Greenberg, Dense polymer film and membrane
formation via the dry-cast process, 2. Model validation and morphological study,
J. Membr. Sci., 94 (1994) 281.
[40] S.A. Mckelvey, W.J. Koros, Phase separation, vitrification, and the
manifestation of macrovoids in polymeric asymmetric membranes, J. Membr.
Sci., 112 (1996) 29.
134
[41] R. Matz, The structure of cellulose acetate membranes, 1. The development of
porous structures in anisotropic membranes, Desalination, 10 (1972) 1.
[42] M.R. Pekny, J. Zartman, W.B. Krantz, A.R. Greenberg, P. Todd, Flow-
visualization during macrovoid pore formation in dry-cast cellulose acetate
membranes, J. Membr. Sci., 211 (2003) 71.
[43] J.M.S. Henis, M.K. Tripodi, Composite hollow fiber membranes for gas
separation: The resistance model approach, J. Membr. Sci., 8 (1981) 233.
135
CHAPTER FIVE
THE EFFECTS OF POLYMER CHAIN RIGIDIFICATION, ZEOLITE PORE
SIZE AND PORE BLOCKAGE ON POLYETHERSULFONE (PES)-ZEOLITE
A MIXED MATRIX MEMBRANES
5.1 INTRODUCTION
Following the discovery by researchers at UOP on mixed matrix membranes (MMMs)
[1], the latest emerging membrane materials comprising molecular sieve entities
embedded in a polymer matrix can potentially surpass the “upper bound” limit of the
permeability-selectivity relationship [2] by means of combining the easy
processability of polymers with the superior gas separation properties of rigid
molecular sieve materials [1, 3-9]. Progress has been made in rubbery polymer-zeolite
mixed matrix membranes [4-5], which showed a significant increase in O2/N2
selectivity, especially at a high zeolite loading. Regrettably, they are not practically
attractive because rubbery polymers might lack mechanical stability and desirable
inherent transport properties relative to rigid glassy polymers at high temperatures.
Therefore, some researchers selected rigid glassy polymers, which possess properties
closer to the “upper-bound”, as the base of mixed matrix membranes. However, it
seems not easy to improve the gas separation performance when choosing glassy
136
polymers as the matrix due to the poor polymer-zeolite compatibility, wherein leading
to the formation of voids between the two phases [8-11].
To fabricate the void-free mixed matrix membranes and fulfill the selectivity
enhancement, some approaches have been proposed [9, 12, 13]. Duval et al. [12]
promoted the adhesion between zeolite sieve and polymer matrix phases by modifying
zeolite surface with silane-coupling agents. Regrettably, significant permselectivity
improvements were not observed despite indications of good coupling between silane
and zeolite with SEM micrographs. This may be due to the difficulty in selecting a
suitable silane-coupling agent to completely prevent the penetrants from going
through voids. Yong et al. [13] introduced a compatibilizer, 2,4,6-triaminopyrimidine
(TAP), to prepare the void-free Matrimid-zeolite mixed matrix membrane by
simultaneously forming hydrogen bonding between them. Results showed that CO2/N2
and O2/N2 selectivity of Matrimid-zeolite 4A-TAP membranes increased around three-
fold compared with pure Matrimid membranes, while CO2 and O2 permeability
decreased at least forty-fold. Mahajan et al. [9] proposed to maintain the polymer
flexibility during the membrane formation by applying high processing temperatures
close to Tg of polymeric materials coupled with the use of a non-volatile solvent. For
polymers with high Tg, one may consider decreasing Tg by means of the incorporation
of a plasticizer into the polymer matrix because it is very difficult to find a non-
volatile solvent with an enough high boiling point to match the temperature
requirement during the membrane formation. However, the addition of a plasticizer
137
may lower the intrinsic gas separation performance of polymeric materials. Therefore,
choosing glassy polymers with an intermediate Tg should be a more appropriate
alternative for this solution.
Besides, a proper choice of zeolites is vital to make high-performance mixed matrix
membranes. Zeolite 4A may be more favorable for the separation of O2/N2 gas pair
than other zeolites due to its appropriate pore size of 3.8 Å. Except for the intrinsic
properties of polymers and zeolites, the interaction between them is also very
important for the gas separation performance of mixed matrix membranes. Moore et al.
[14] have proved that the polymer chain rigidification induced by the addition of
zeolites led a significant decrease in the gas permeability of mixed matrix membranes.
The purpose of this work is to demonstrate that other factors resulting from the
interaction between polymer and zeolite phases may also remarkably influence the gas
separation performance of mixed matrix membranes, such as the partial pore blockage
of zeolites by the polymer chains. Structurally, Zeolite 3A and 5A are very similar to
zeolite 4A, except they have different cations in the zeolite and pore sizes. The cations
in zeolite 3A and 5A are potassium and calcium, respectively, instead of sodium in
zeolite 4A, and their pore sizes are 3 Å and 4.5 Å, respectively. Therefore, zeolite 3A,
4A and 5A were used as the dispersed phase in this work in order to study the effect of
the partial pore blockage of zeolites. Polyethersulfone (PES) was chosen as the
continuous polymer matrix due to its appropriate Tg of 215oC and various applications
in gas separation [15]. PES-zeolite A mixed matrix membranes were fabricated based
138
on a relatively straightforward approach by applying high processing temperatures
during the membrane formation [9]. The gas permeation rates of these PES-zeolite A
MMMs were measured as a function of membrane preparation conditions, zeolite
loading and pore size of zeolite. The morphology study and Tg examination of these
developed MMMs were investigated with SEM and DSC, respectively.
5.2 RESULTS AND DISCUSSION
5.2.1 Effect of Membrane Preparation Methodology on Gas Separation
Performance
Our previous work has demonstrated that vacuum degassing can effectively avoid the
formation of voids between polymer and zeolite phases resulting from the air adsorbed
on the surface of zeolite particles. With the development of membrane fabrication
technology, another important factor influencing gas separation performance has been
identified, that is the cooling protocol after annealing at 250oC (i.e., immediate
quenching or natural cooling down). A comparison of gas permeability of PES-zeolite
3A, PES-zeolite 4A and PES-zeolite 5A mixed matrix membranes with different
cooling protocols is given in Figure 5.1, where it shows that H2, O2 and N2
permeability of PES-zeolite 3A and PES-zeolite 4A membranes with natural cooling
decreases a little compared with that of corresponding membranes with immediate
quenching. A comparison of gas pair selectivity of these MMMs is listed in Figure 5.2,
where it shows that H2/N2 and O2/N2 selectivity of these membranes increases fairly
139
except for the H2/N2 selectivity of PES-zeolite 3A membrane. The zeolite loadings of
these six mixed matrix membranes all are 20 wt%.
A
0
2
4
6
PES-3A MMM PES-4A MMM PES-5A MMM
H2 p
erm
eabi
lity
immediate quenching
natural cooling down
B
0
0.25
0.5
PES-3A MMM PES-4A MMM PES-5A MMM
O2 p
erm
eabi
lity
immediate quenching
natural cooling down
C
0
0.035
0.07
PES-3A MMM PES-4A MMM PES-5A MMM
N2 p
erm
eabi
lity
immediate quenching
natural cooling down
Figure 5.1: Comparison of gas permeability of PES-zeolite A mixed matrix membranes with different cooling protocols (A: H2 permeability; B: O2 permeability;
C: N2 permeability)
140
70
80
90
100
PES-3A MMM PES-4A MMM PES-5A MMM
H2/N
2 sel
ectiv
ity
immediate quenching
natural cooling down
5
6
7
PES-3A MMM PES-4A MMM PES-5A MMM
O2/N
2 sel
ectiv
ity
immediate quenching
natural cooling down
Figure 5.2: Comparison of gas pair selectivity of PES-zeolite A mixed matrix membranes with different cooling protocols (left: H2/N2 selectivity; right: O2/N2
selectivity)
Clearly, immediate quenching and natural cooling result in MMMs with different
performance. This may arise from the fact that immediate quenching after annealing
membranes above Tg may make polymer chains being frozen quickly as polymer
chains are still in the random status. Therefore, the resultant membrane has a higher
free volume in the polymer matrix and subsequently higher gas permeability without
the loss of selectivity. This quench method has been applied widely in the membrane
fabrication process [16-18]. However, if applying this method to form the mixed
matrix membranes, polymer chains may be detached from the zeolite surface as
polymer chains are suddenly frozen, resulting in the formation of voids between
polymer and zeolite phases because of their different thermal coefficients of expansion.
The purpose of applying the natural cooling after annealing is to make polymer chains
harden and shrink much slowly and gradually with the decrease of temperature. The
zeolite is less affected by the decrease of temperature due to its inorganic properties;
141
therefore, polymer chains may better adhere on the zeolite surface. The cross-section
SEM micrographs of these six mixed matrix membranes are illustrated in Figure 5.3.
The contact between polymer and zeolite phases with natural cooling is apparently
better than that with immediate quenching. The natural cooling method may slightly
decrease the gas permeability because of the loss of a part of free volume in
membranes; however, it clearly increases the selectivity due to the improvement of
contact between polymer and zeolite phases. As shown in Figures 5.1 and 5.2, natural
cooling results in PES-zeolite 3A and PES-zeolite 4A membranes with lower H2, O2
and N2 permeability, but higher H2/N2 and O2/N2 selectivity except that the H2/N2
selectivity of PES-zeolite 3A membrane maintains the same. However, both
permeability and selectivity of PES-zeolite 5A membrane made from natural cooling
are higher than those of the corresponding membrane made from immediate
quenching. This interesting phenomenon implies that other factors may also play an
important role on the interface between polymer and zeolite and affect the gas
separation performance of MMMs. This will be discussed in the next section.
142
CA B
D E F
Figure 5.3: Comparison of cross-section SEM images of MMMs (A, B, C: PES-zeolite 3A, 4A and 5A MMMs with immediate quenching, respectively; D, E, F: PES-zeolite
3A, 4A and 5A MMMs with natural cooling, respectively)
5.2.2 Effect of Zeolite Loadings on Gas Separation Performance
PES-zeolite 4A and PES-zeolite 5A mixed matrix membranes with different zeolite
loadings were fabricated by the protocol of natural cooling. The pure PES dense film
was also fabricated with the same procedure for comparison. Effect of different zeolite
loadings on the permeability of different gases for these MMMs is listed in Figure 5.4.
This figure indicates that permeability of all studied gases generally decrease with
increasing zeolite content. Although the magnitude of permeability and their variations
with zeolite loadings are different for different gases, the general trend is the same,
implying a similar permeation mechanism. The decreasing trend of permeability
143
seems counter-intuitive at first because the pore sizes of both zeolite 4A and 5A are
larger than the molecular diameters of H2 and O2. Figure 5.5 compares the
experimental results (solid line) with the theoretical calculations (dotted line) using the
Maxwell equation written as follows [19]:
])(2)(22
[DCDCD
DCDCDCMMM PPvPP
PPvPPPP
−++−−+
= (5.1)
where PMMM is the composite permeability of MMMs, v is the volume fraction, and the
subscripts C and D refer to the continuous and dispersed phases, respectively. Both PC
and PD values for O2 and N2 gases are listed in Table 5.1. Based on the Maxwell
equation, the O2 permeability of these PES-zeolite 4A MMMs should increase with
zeolite loadings due to the relatively larger pore size of zeolite 4A. However, the
experimental results shown in Figure 5.5 exhibit an opposite trend.
3
6
9
0 10 20 30 40 50 60
Zeolite loading (%)
Perm
eabi
lity
(Bar
rer)
H2 of PES-zeolite 4A membranes
H2 of PES-zeolite 5A membranes
0
0.3
0.6
0 10 20 30 40 50 60Zeolite loading (%)
Perm
eabi
lity
(Bar
rer)
O2 of PES-zeolite 4A membranesN2 of PES-zeolite 4A membranesO2 of PES-zeolite 5A membranesN2 of PES-zeolite 5A membranes
Figure 5.4: Effect of zeolite loadings on H2, O2 and N2 permeability of PES-zeolite 4A and PES-zeolite 5A mixed matrix membranes (left: H2 permeability; right: O2 and N2
permeability)
144
0
0.3
0.6
0.9
1.2
0 10 20 30 40 50 60
Zeolite loading (%)
Perm
eabi
lity
(Bar
rer)
O2 (experimental data)
O2 (prediction by the Maxwell model)
O2 (prediction by the modified Maxwell model assuming polymer chain rigidification)
O2 (prediction by the new modified Maxwell model simultaneously assuming polymerchain rigidification and partial pore blockage of zeolites)
Figure 5.5: Comparison of O2 permeability of PES-zeolite 4A MMMs between experimental data and predictions from the Maxwell model and various modified
Maxwell models
Table 5.1: Change of O2 and N2 permeability in different regions of PES-zeolite 4A mixed matrix membranes
0.7700.003080.1600.479 O2
0.0825
Permeability in the PES matrix (PC)a
0.0206
Permeability in the rigidified PES region
(Prig=PC/βc)
0.00208
Permeability in the zeolite-4A skin with partial pore
blockage (Pblo=PD/β’d)
0.0208N2
Permeability in the bulk of zeolite-4A
(PD)b
0.7700.003080.1600.479 O2
0.0825
Permeability in the PES matrix (PC)a
0.0206
Permeability in the rigidified PES region
(Prig=PC/βc)
0.00208
Permeability in the zeolite-4A skin with partial pore
blockage (Pblo=PD/β’d)
0.0208N2
Permeability in the bulk of zeolite-4A
(PD)b
a: Gas permeability data in the PES matrix come from experiment results of the pure PES dense film.b: Gas permeability data in the bulk of zeolite-4A come from the reference 26.
c: β is 3 for O2 gas, while β is 4 for N2 gas.
d: β’ is 250 for O2 gas, while β’ is 10 for N2 gas.
145
Two possible hypotheses are considered for the decline in the permeability. One is due
to the inhibition of polymer chain mobility near the polymer-zeolite interface; in other
words, the presence of zeolite seems to rigidify polymeric chains [20]. The 1st
hypothesis can be somewhat confirmed by the Tg results given in Table 5.2. It can be
found that the magnitude of Tg increment is not very pronounced; however, its
increasing trend with zeolite loadings indicates a fair interfacial interaction between
PES and zeolite. Therefore, the rigidified polymeric chains around the zeolite surface
may be one of the causes for permeability reduction.
Table 5.2: Change of glass transition temperatures of mixed matrix membranes over pure PES dense film
219±1oC218±1oC217±1oC216±1oC215±1oCGlass transition temperature (oC)
PES-zeolite 5A (50 wt% loading)
PES-zeolite 5A (40 wt% loading)
PES-zeolite 5A (30 wt% loading)
PES-zeolite 5A (20 wt% loading)
Pure PES dense film
Membrane name
215±1oC
Pure PES dense film
218±1oC
PES-zeolite 4A (20 wt% loading)
219±1oC
PES-zeolite 4A (30 wt% loading)
221±1oC_Glass transition temperature (oC)
PES-zeolite 4A (50 wt% loading)
PES-zeolite 4A (40 wt% loading)
Membrane name
219±1oC218±1oC217±1oC216±1oC215±1oCGlass transition temperature (oC)
PES-zeolite 5A (50 wt% loading)
PES-zeolite 5A (40 wt% loading)
PES-zeolite 5A (30 wt% loading)
PES-zeolite 5A (20 wt% loading)
Pure PES dense film
Membrane name
215±1oC
Pure PES dense film
218±1oC
PES-zeolite 4A (20 wt% loading)
219±1oC
PES-zeolite 4A (30 wt% loading)
221±1oC_Glass transition temperature (oC)
PES-zeolite 4A (50 wt% loading)
PES-zeolite 4A (40 wt% loading)
Membrane name
Since Tg measurements provide qualitative confirmation of rigidified polymeric chains,
a modified Maxwell model is used to quantitatively estimate the effect of the chain
rigidification on the permeability of MMMs [14, 21]. Firstly, the whole system of
mixed matrix membranes can be imaged as a pseudo two-phase composite, as
illustrated in Figure 5.6. The first phase is the polymer matrix, and the second phase is
composed of the whole zeolite phase and rigidified polymer region near the zeolite.
146
Secondly, the Maxwell equation can be used to obtain the permeability of the second
phase as the following equation assuming the rigidified polymer region as the
continuous phase and the whole zeolite as the dispersed phase.
])(2)(22
[2DrigSrigD
DrigSrigDrignd PPvPP
PPvPPPP
−++
−−+= (5.2)
here P2nd is the composite permeability of the second phase, PD is the permeability of
the dispersed zeolite phase, Prig is the permeability of the rigidified polymer region,
and vS is the volume fraction of the zeolite phase in the second phase and can be
determined by the following simple equation:
rigD
DS vv
vv+
= (5.3)
where vD is the volume fraction of dispersed zeolite phase in the total membrane and
vrig is the volume fraction of the rigidified polymer region in the total membrane.
Finally, one can obtain the permeability of mixed matrix membranes by applying the
Maxwell equation a second time as the following equation by considering the polymer
matrix as the continuous phase (i.e., the first phase) and the second phase as the
dispersed phase.
]))((2))((22
[22
22
ndCrigDCnd
ndCrigDCndCMMM PPvvPP
PPvvPPPP
−+++
−+−+= (5.4)
147
Rigidified polymer region Prig = PC/β
Zeolite skin affected by the partial pore blockage
Pblo = PD/β’
The second phase [the third phase (dispersed phase) + the rigidified
polymer region (continuous phase)]
The first phase (polymer matrix)
r’r r’r
The third phase [the bulk of zeolite (dispersed phase) + the zeolite skin with partial pore blockage (continuous phase)]
Polymer chains
Zeolite 4A
Bulk of zeolite PD
Polymer matrix PC
3.8Ǻ
Figure 5.6: Schematic diagram for the new modified Maxwell model
Zeolite 4A Linde crystals have an O2 permeability of around 0.77 Barrer and an O2/N2
selectivity of around 37 at 35oC [22]. Therefore, only two unknowns in Equations
(5.2)-(5.4) need to be set; namely, Prig, the permeability in the rigidified polymer
region, and vrig, the volume fraction of this region. Prig can be assumed to be the
permeability of the polymer matrix divided by a chain immobilization factor β [14, 21,
23]. A value of 3 is assumed as the parameter β for O2 gas in this work based on the
previous related studies [14, 21]. The zeolite 4A involved in this work has a cubic
morphology with an average side length of 3 µm; therefore, the thickness of rigidified
polymer region near the zeolite, r, is assumed to be 0.30 µm in order to provide a
148
reasonable fit to the experimental data. In fact, these two parameters can be obtained
through minimizing the sum of the errors in gas permeability for all PES-zeolite 4A
MMMs. However, Moore et al. [14] have demonstrated that predictions are not very
sensitive to the values used for β and r. Moreover, the trend is more important than the
actual fit; therefore, only 3 and 0.30 are used as the values of β(O2) and r, respectively,
in this work. The O2 permeability of PES-zeolite 4A membranes predicted by the
modified Maxwell model assuming the chain rigidification is also given in Figure 5.5.
Even though the predictions do not exactly match the experimental data, the predicted
trend is getting closer than that obtained by the unmodified Maxwell mode. Therefore,
the gas permeability of mixed matrix membranes filled with zeolite possessing
relatively larger pore sizes may exhibit a decrease with an increase in zeolite loading
due to the rigidified polymer chains.
The other hypothesis for the decline in permeability is possibly due to the partial pore
blockage of zeolites by polymer chains [24]. Even though polymer chains can hardly
enter into the zeolite pores, they may obstruct a part of pores through the attachment
on the zeolite surface due to enough flexibility when applying high processing
temperatures during the membrane formation [14, 21]. The effect of the partial pore
blockage on permeability of mixed matrix membranes has been explored by Koros
and his coworkers using similar concepts of the modified Maxwell model derived for
the inhibition of polymer chain mobility [14, 21]. Here we propose a new concept by
combing both effects of chain rigidification and partial pore blockage into calculation.
149
5.2.3 New Modified Maxwell Model to Predict Gas Separation Performance
As illustrated in Figure 5.6, two parameters are introduced; parameter β’ characterizes
the average decline factor of gas permeability near the zeolite skin because of the
partial pore blockage and r’ refers to the average thickness of this influenced region.
Apparently, the permeability decline in the zeolite skin induced by the partial pore
blockage should be much larger than that arising from the inhibition of polymer chain
mobility. In addition, the thickness of this region should be much smaller as the
polymer can hardly go into the bulk of zeolite. Therefore, based on other related
literatures [14, 21], β’ for O2 gas and r’ are assumed to be 250 and 100 Å, respectively.
The prediction simultaneously considering the polymer chain rigidification and partial
pore blockage of zeolites can be carried out by imaging the whole system of mixed
matrix membranes as a pseudo three-phase composite illustrated in Figure 5.6 and
applying the Maxwell equation three times. The permeability of the third phase can be
calculated with the aid of the Maxwell equation by assuming the bulk of zeolite as the
dispersed phase and the affected zeolite skin with low permeability as the continuous
phase. Subsequently, the permeability of the second phase and resultant mixed matrix
membrane can also be determined with similar concepts derived for the inhibition of
polymer chain mobility. All calculated volume fraction data of dispersed phase in the
third phase, second phase and whole MMM in this new modified Maxwell model are
listed in Table 5.3. The calculation method to obtain these volume fractions has been
150
discussed extensively by Moore et al. [14], so it is not necessary to duplicate it in this
work. It can be found that the volume fraction of the second phase in the whole MMM
reaches 0.824 at 50 wt% zeolite loading; that is, the whole polymer matrix may be
rigidified and the volume fraction of the second phase approaches the maximum value
of 1 with a further increase in zeolite loadings. However, this situation doesn’t occur
in this work because there are too many defects among zeolite particles to test in
MMMs with a zeolite loading higher than 50 wt%.
Table 5.3: Calculated volume fraction data of dispersed phase in different phases of PES-zeolite 4A MMMS in the new modified Maxwell model which simultaneously
considers both polymer chain rigidification and partial pore blockage of zeolites
0.32120 wt% zeolite loading (18.6 vol% zeolite loading)
0.48530 wt% zeolite loading (28.1 vol% zeolite loading)
0.65340 wt% zeolite loading (37.8 vol% zeolite loading)
0.82450 wt% zeolite loading (47.7 vol% zeolite loading)
0.579Calculated volume fraction of the third phase (considered as the dispersed phase) in the second phase
(defined in Figure 7)
Calculated volume fraction of the second phase (considered as the dispersed phase) in the whole mixed matrix membrane
(defined in Figure 7)
Calculated volume fraction of the bulk of zeolite 4A (considered as the dispersed phase) in the third phase
(defined in Figure 7)
0.980
0.32120 wt% zeolite loading (18.6 vol% zeolite loading)
0.48530 wt% zeolite loading (28.1 vol% zeolite loading)
0.65340 wt% zeolite loading (37.8 vol% zeolite loading)
0.82450 wt% zeolite loading (47.7 vol% zeolite loading)
0.579Calculated volume fraction of the third phase (considered as the dispersed phase) in the second phase
(defined in Figure 7)
Calculated volume fraction of the second phase (considered as the dispersed phase) in the whole mixed matrix membrane
(defined in Figure 7)
Calculated volume fraction of the bulk of zeolite 4A (considered as the dispersed phase) in the third phase
(defined in Figure 7)
0.980
Figure 5.5 illustrates the ultimate prediction of O2 permeability of PES-zeolite 4A
membranes. It can be found that the prediction is in good agreement with experimental
data. Therefore, it is possible that the partial pore blockage of zeolites by polymer
151
chains plays a much important role in the permeability decrease of MMMs, not caused
only by the polymer chain rigidification.
Effects of different zeolite loadings on H2/N2 and O2/N2 selectivity of PES-zeolite 4A
and PES-zeolite 5A membranes are illustrated in Figure 5.7. Results indicate that
selectivity of both kinds of mixed matrix membranes increase with an increase in
zeolite loadings. The maximum increment occurs at the 50 wt% zeolite loading. The
maximum percentages of selectivity increments of H2/N2 and O2/N2 for PES-zeolite
5A membrane reach approximately 50% and 30%, respectively. This increasing trend
is easily understandable due to the molecular sieving mechanism; however, the
magnitude of increment seems to be still far below from what one expects. The
predicted O2/N2 selectivity using the Maxwell model will be discussed in the next
section by choosing the PES-zeolite 4A mixed matrix membrane as an example.
70
90
110
130
0 10 20 30 40 50 60
Zeolite loading (%)
H2/N
2 sel
ectiv
ity
PES-zeolite 4A membrane
PES-zeolite 5A membrane
5
6
7
8
0 20 40 60
Zeolite loading (%)
O2/N
2 sel
ectiv
ity
PES-zeolite 4A membranePES-zeolite 5A membrane
Figure 5.7: Effect of zeolite loadings on H2/N2 and O2/N2 selectivity of PES-zeolite 4A and PES-zeolite 5A mixed matrix membranes (left: H2/N2 selectivity; right: O2/N2
selectivity)
152
Figure 5.8 shows the O2/N2 selectivity of PES-zeolite 4A membranes predicted by the
Maxwell model. It can be found that the Maxwell model predicts a much higher
magnitude of O2/N2 selectivity increment than that of experimental data. The reasons
may be still due to the inhibition of polymer chain mobility near the polymer-zeolite
interface and the partial pore blockage of zeolites by the polymer chains.
5
10
15
0 10 20 30 40 50 60
Zeolite loading (%)
O2/
N2
sele
ctiv
ity
O2/N2 (experimental data)
O2/N2 (pridiction by the Maxwell model)
O2/N2 (pridiction by the new modified Maxwell model)
Figure 5.8: Comparison of O2/N2 selectivity of PES-zeolite 4A MMMs between experimental data and predictions from the Maxwell model and the new modified
Maxwell model
Therefore, the Maxwell model is modified with the same assumptions and parameters
mentioned in the previous sections in order to include these two effects (i.e., chain
rigidification and partial pore blockage). However, two very important differences
need to be noticed. The first one is that the chain immobilization factor, β, is different
153
for O2 and N2 gases because the polymer chain rigidification normally results in the
larger resistance to the transport of gas with a larger molecular diameter and higher
selectivity of O2/N2. Therefore, this parameter is 3 for O2 gas in the rigidified polymer
region, while for N2 gas, this parameter is 4.
The second one is that the average decline factor of gas permeability near the zeolite
skin due to the partial pore blockage, β’, is also different for O2 and N2 gases. Without
the partial pore blockage of zeolite 4A, the oxygen molecule can rotate along either of
its two axes since both its dimensions are slightly smaller than the 3.8 Å aperture,
while the rotational motion of oxygen molecule is significantly hindered and can only
characterize the vibrational motion with the partial pore blockage [21, 25]. For the
nitrogen molecule, it cannot rotate around either of it axes and can only characterize
the vibrational motion no matter if the partial pore blockage happens [21, 25].
Therefore, the average decline of O2 permeability in the zeolite skin with partial pore
blockage should be much larger than that of N2 permeability, thus leading that the
O2/N2 selectivity in this region should be much smaller than that in the bulk of zeolite
4A. β’(O2) has been assumed to be 250, so β’(N2) is assumed to be 10 in this work. O2
and N2 permeability data in the rigidified polymer region and the zeolite skin with
partial pore blockage are listed in Table 5.1. The ultimate predictions of O2/N2
selectivity of PES-zeolite 4A membranes by the new modified Maxwell model are
also shown in Figure 5.8. It can be found that these predictions are much closer to
experimental data than those predicted by the unmodified Maxwell model. This
154
demonstrates that both polymer chain rigidification and partial pore blockage of
zeolites actually affect the utility of zeolites.
5.2.4 Effect of Pore Sizes of the Zeolite on Gas Separation Performances
The pore size of zeolites is another important factor to influence the gas separation
performance of mixed matrix membranes as well as the zeolite loadings. Therefore, in
order to study the effect of pore sizes, PES-zeolite 3A mixed matrix membranes with
30 wt% zeolite loading were also fabricated by the same protocol of natural cooling.
The results in Figure 5.9 exhibit that H2 and O2 permeability almost increases with an
increase in zeolite pore size. This trend is easily understandable because the larger the
zeolite pore size, the easier for gases to pass through.
0.25
0.35
0.45
PES-3AMMM
PES-4AMMM
PES-5AMMM
O2 p
erm
eabi
lty (B
arre
r) 30 wt% zeolite loading
4
5
6
PES-3AMMM
PES-4AMMM
PES-5AMMM
H2 p
erm
eabi
lity
(Bar
rer) 30 wt% zeolite loading
Figure 5.9: Effect of different pore sizes of zeolite on H2 and O2 permeability of mixed matrix membranes (left: H2 permeability; right: O2 permeability)
155
However, it is very strange that H2/N2 and O2/N2 selectivity also increases with an
increase in zeolite pore size as shown in Figure 5.10. If only assuming the molecular
sieving mechanism, the gas separation performance of PES-zeolite 4A membranes
should be better than that of PES-zeolite 5A membranes because the pore size of
zeolite 4A can just exclude the pass of N2 gas. However, this work shows reversed
results, i.e. H2/N2 and O2/N2 selectivity of PES-zeolite 5A membranes is much higher
than that of PES-zeolite 4A membranes.
70
90
110
Pure PESdense film
PES-3AMMM
PES-4AMMM
PES-5AMMM
H2/N
2 sel
ectiv
ity
30 wt% zeolite loading
5.5
6.5
7.5
Pure PESdense film
PES-3AMMM
PES-4AMMM
PES-5AMMM
O2/N
2 sel
ectiv
ity
30 wt% zeolite loading
Figure 5.10: Effect of different pore sizes of zeolite on H2/N2 and O2/N2 selectivity of mixed matrix membranes (left: H2/N2 selectivity; right: O2/N2 selectivity)
In fact, this reversed phenomenon may indirectly support the previous arguments that
the partial pore blockage of zeolites by polymer chains may play a much important
role on gas separation performance of PES-zeolite mixed matrix membranes.
Interestingly, the O2/N2 selectivity of PES-zeolite 3A and PES-zeolite 4A membranes
is very similar, while O2/N2 selectivity of PES-zeolite 5A membranes suddenly
enhances remarkably. This implies the blockage may narrow a part of zeolite 5A pores
156
to approximately 4Å which can discriminate the gas pair of O2 and N2, and narrow a
part of zeolites 3A and 4A pores to smaller sizes which can result in an increase in the
resistance to O2 transport. The H2/N2 selectivity increases with the augmentation of
zeolite pore size for these three kinds of mixed matrix membranes. This may arise
from the fact that these three types of zeolites all can reject the pass of N2 due to their
partial pore blockage, plus the resistance to H2 transport becomes lower with an
increase in zeolite pore size. The intrinsic pore size of zeolite 3A is smaller than the
molecular diameter of O2 gas, and its pore size is perhaps further reduced to smaller
than the molecular diameter of H2 gas due to the partial pore blockage. Therefore, it is
somewhat strange that H2/N2 and O2/N2 selectivity of PES-zeolite 3A membranes still
increases compared with the pure PES dense film. This phenomenon probably
attributes to the polymer chain rigidification near the zeolite, which normally results in
the larger resistance to the transport of gas with a larger molecular diameter and higher
selectivity for H2/N2 and O2/N2 gas pairs. Therefore, the above-mentioned analysis
strongly indicates that the partial pore blockage of zeolites by the polymer chains has
equivalent or more influence on the separation properties of mixed matrix membranes
compared with the polymer chain rigidification.
5.3 CONCLUSIONS
The following conclusions can be drawn from this chapter:
157
(1) The gas separation performance of mixed matrix membranes with natural cooling
is superior to that of mixed matrix membranes with immediate quenching probably
because natural cooling can make polymer chains better adhere on the zeolite surface
proved by SEM micrographs.
(2) The permeability of all gases for PES-zeolite 4A and PES-zeolite 5A membranes
decreases with an increase in zeolite loading, which is opposite to the prediction by the
Maxwell model. There are two possible hypotheses. One is the polymer chain
rigidification near the zeolite; the other is the partial pore blockage of zeolites by the
polymer chains. A new modified Maxwell model which combines and compounds
these two factors into calculation has been proposed. The predicted permeability and
selectivity show good agreement with experimental data.
(3) The permeability of all gases for PES-zeolite 3A, PES-zeolite 4A and PES-zeolite
5A membranes generally increases with an increase in zeolite pore size, which is
consistent with the molecular sieving mechanism. However, it is surprising that the
gas separation performance of PES-zeolite 5A membranes is much higher than that of
PES-zeolite 3A and PES-zeolite 4A membranes, which does not completely obey the
molecular sieving mechanism. This odd phenomenon just confirms that the partial
pore blockage of zeolites by the polymer chains may significantly affect the gas
separation performance of mixed matrix membranes.
158
5.4 REFERENCES
[1] S. Kulprathipanja, R.W. Neuzil, N.N. Li, Separation of fluids by means of mixed
matrix membranes in gas permeation, US Patent No. 4,740,219 (1988).
[2] L.M. Robeson, Correlation of separation factor versus permeability for polymeric
membranes, J. Membr. Sci., 62 (1991) 165.
[3] H.J.C. Te Hennepe, Zeolite Filled Polymeric Membranes: A New Concept in
Separation Science, Thesis, University of Twente, 1988.
[4] M. Jia, K.V. Peinemann, R.D. Behling, Molecular sieving effect of zeolite filled
silicone rubber membranes, J. Membr. Sci., 57 (1991) 289.
[5] J. M. Duval, B. Folkers, M.H.V. Mulder, G. Desgrandchamps, C.A. Smolders,
Adsorbent-filled membranes for gas separation. Part 1. Improvement of the gas
separation properties of polymeric membranes by incorporation of microporous
adsorbents, J. Membr. Sci., 80 (1993) 189.
[6] D.R. Paul, D.R. Kemp, The diffusion time lag in polymer membranes containing
adsorptive fillers, J. Polym. Sci. Symp., 41 (1973) 79.
[7] R. Mahajan, W.J. Koros, Factors controlling successful formation of mixed-matrix
gas separation materials, Ind. Eng. Chem. Res., 39 (2000) 2692.
[8] M.G. Süer, N. Baç, L. Yilmaz, Gas permeation characteristics of polymer-zeolite
mixed matrix membranes, J. Membr. Sci., 91 (1994) 77.
159
[9] R. Mahajan, R. Burns, M. Schaeffer, W.J. Koros, Challenges in forming successful
mixed matrix membranes with rigid polymeric materials, J. App. Polym. Sci., 86
(2002) 881.
[10] A. Erdem-Şenatalar, M. Tatlıer, Ş.B. Tantekin-Ersolmaz, Estimation of the
interphase thickness and permeability in polymer-zeolite mixed matrix membrane,
in: Proceedings of the 13th International Zeolite Conference, Montpellier, France,
2001.
[11] Ş.B. Tantekin-Ersolmaz, L. Şenorkyan, N. Kalaonra, M. Tatlıer, A. Erdem-
Şenatalar, n-Pentane/i-pentane separation by using zeolite-PDMS mixed matrix
membranes, J. Membr. Sci., 189 (2001) 59.
[12] J.M. Duval, A.J.B. Kemperman, B. Folkers, M.H.V. Mulder, G.
Desgrandchamps, C.A. Smolders, Preparation of zeolite filled glassy polymer
membrane, J. Appl. Polym. Sci., 54 (1994) 409.
[13] H.H. Yong, H.C. Park, Y.S. Kang, J. Won, W.N. Kim, Zeolite-filled polyimide
membrane containing 2,4,6-triaminopyrimidine, J. Membr. Sci., 188 (2001) 151.
[14] T.T. Moore, R. Mahajan, De Q. Vu, W.J. Koros, Hybrid membrane materials
comprising organic polymers with rigid dispersed phases, AICHE Journal, 50 (2)
(2004) 311.
[15] J.S. Chiou, Y. Maeda, D.R. Paul, Gas permeation in polyethersulfone, J. App.
Polym. Sci. 33 (1987) 1823.
160
[16] J.Y. Dolveck, G.H. Dai, P. Moser, M. Pineri, P. Aldebert, M. Escoubes, R.
Avrillon, J.C. Mileo, Free volume measurements in polyimides by positronium
annihilation, Mater. Sci. Forum, 105-110 (1992) 1549.
[17] H. Kazuo, M. Kazutaka, O. Yoshiyuki, O. Yoshuki, Polymer membranes with
high capacity in gas separation, Japanese Patent 62,019,206 (1987).
[18] A. Takanori, O. Yoshuki, Manufacture of gas-separation membrane, Japanese
Patent 61,101,206 (1986).
[19] K. Mizoguchi, K. Terada, Y. Naito, Y. Kamiya, S. Tsuchida, S. Yano, Miscibility
blends and gas permeability of poly(ethylene-co-3,5,5-trimethylhexyl
methacrylate)-polydimethylsiloxane, Colloid. Polym. Sci. 86 (1997) 275.
[20] S.S. Chan, T.S. Chung, R. Wang, Z. Lu, C. He, Characterization of permeability
and sorption in Matrimid/C60 mixed matrix membranes, J. Membr. Sci., 218
(2002) 215.
[21] R. Mahajan, Formation, characterization and modeling of mixed matrix
membrane materials, Ph.D. Dissertation, The University of Texas at Austin,
2000.
[22] C.M. Zimmerman, A. Singh, W.J. Koros, Tailoring mixed matrix composite
membranes for gas separation, J. Membr. Sci. 137 (1997) 145.
[23] A.S. Michaels, R.B. Parker, Sorption and flow of gases in polyethylene, J. Polym.
Sci. 41 (138) (1959) 53.
161
[24] D.Q. Vu, W.J. Koros, S.J. Miller, Mixed matrix membranes using carbon
molecular sieves. Part 1. Preparation and experimental results, J. Membr. Sci.,
211 (2003) 311.
[25] P.S. Rallabandi, D.M. Ford, Entropic and energetic selectivity in air separation
with microporous materials, AICHE Journal, 46 (1) (2000) 99.
162
CHAPTER SIX
EFFECTS OF NOVEL SILANE MODIFICATION OF ZEOLITE SURFACE
ON POLYMER CHAIN RIGIDIFICATION AND PARTIAL PORE
BLOCKAGE IN POLYETHERSULFONE (PES)-ZEOLITE A MIXED
MATRIX MEMBRANES
6.1 INTRODUCTION
During the last two decades, polymer-based organic-inorganic composites have
received world-wide attention in the field of material science. This is because the
resultant materials may offer superior performance in terms of mechanical toughness
for engineering resins, permeability and selectivity for gas/liquid separation, and
photoconductivity for electronics [1-3]. This new concept has also been applied to the
membrane-based gas/liquid separation by combining the easy processability of organic
polymers with the excellent separation properties of inorganic materials [2, 4-10].
However, it has been found that there is an obstacle to the successful introduction of
inorganic molecular sieve materials into an organic polymer matrix because of the
poor compatibility between molecular sieve and polymer matrix, especially for the
case of rigid glassy polymer materials [9-13].
163
Therefore, some methods have been proposed to improve the interfacial strength in
order to enhance the separation performance [13-18]. One of them is the chemical
modification of the surface of molecular sieve with silane coupling agents such as (γ-
aminopropyl)-triethoxy silane, N-β-(aminoethyl)-γ-aminopropyltrimethoxy silane, (γ-
glycidyloxypropyl)-trimethoxy silane and (3-aminopropyl)-dimethylethoxy silane [14-
17]. By means of FTIR and 29Si MAS NMR, Matsumoto et al have identified that
there are three types of silanol groups on the surface of zeolite [19], i.e. single
((SiO)3Si-OH); hydrogen-bonded ((SiO)3Si-OH—OH-Si(SiO)3) and geminal
((SiO)2Si(OH)2). Among these silanol groups, free SiOH groups are the most reactive
groups and may provide the sites for the physical and chemical adsorption of silane
coupling agents. With appropriate silane coupling agents, one may not only modify
surface properties of zeolite from hydrophilic to hydrophobic, but also increase zeolite
affinity to the functional groups of the polymer matrix as illustrated in references [20-
22].
Duval et al. [16] demonstrated that the adhesion between zeolite and polymer matrix
phases can be improved by modifying zeolite surface with some silane coupling
agents. However, there was no significant improvement on permselectivity though
SEM micrographs indicated good coupling between silane and zeolite. Mahajan and
Koros [14, 23] also reported similar phenomena. Although the improvement on
interface was clearly observed by SEM, their mixed matrix membranes (MMMs)
made of polymer-modified zeolite exhibited worse performance (i.e. decreases in both
164
permeability and selectivity) compared with those made of polymer-unmodified
zeolite. This is because the addition of silane coupling agents may reduce the voids
between unmodified zeolite and polymer phases, but cannot eliminate them
completely. The size of these voids may be in the range of nanometer or sub-
nanometer [14, 15, 23], which is still larger than the sizes of gas molecules. As a
consequence, the addition of silane coupling agents on zeolite surface may help reduce
voids and result in an increase in gas transport resistance across the membrane, but
there is almost no improvement on permselectivity. Unless a proper silane coupling
agent is chosen, it is very difficult to completely prevent the gas penetrants from going
through these voids.
Therefore, the first purpose of this study is to introduce a novel silane coupling agent,
(3-aminopropyl)-diethoxymethyl silane (APDEMS), on zeolite surface and to
investigate its effects on the separation performance of MMMs. To our best
knowledge, so far there is no academic literature available on using APDEMS to
modify the zeolite surface for MMMs. Zeolites with different pore sizes are used in
order to identify if the effectiveness of silane modification is pore-size dependent. In
addition, although the interphase occupies an extremely small volume fraction (i.e.,
less than 10-10%), it appears to have a significant effect on the separation performance
of mixed matrix membranes [10, 13-15, 23-25]. The origins of the imperfect
interphase are complicated. Poor compatibility between molecular sieve and polymer
matrix, uneven shrinkages and stresses of these two components during the membrane
165
formation may be some of the possible causes. As a result, polymer chain
rigidification and partial pore blockage of zeolites have been hypothesized and
partially confirmed from the lower gas permeation data of mixed matrix membranes
[10, 13, 15, 23-25]. Therefore, the 2nd purpose of this paper is to investigate if and how
the novel silane coupling agent affects polymer chain rigidification and partial pore
blockage of zeolites and to study if we can reduce the nanometric interphase.
For easy comparison with our previous work [25], zeolite 3A, 4A and 5A were used as
the dispersed phase in this work, and polyethersulfone (PES) was chosen as the
continuous polymer matrix due to its appropriate Tg of 215oC and various applications
in gas separation [26]. Zeolite and its surface were characterized by elementary
analysis, XPS and Brunaer-Emmett-Teller (BET). A high processing temperature was
employed in order to maintain the flexibility of polymer chains during membrane
formation [13-15, 23, 25]. The gas permeation rates of the PES-zeolite A MMMs were
measured as a function of zeolite loading and zeolite pore sizes. SEM and DSC were
utilized to study the morphology and Tg of these developed MMMs, respectively.
6.2 RESULTS AND DISCUSSION
6.2.1 Characterization and Comparison of Unmodified and Modified Zeolites
In order to determine if the chemical modification of zeolite surface is actually
successful, Table 6.1 lists the results of elementary analysis before and after the
166
chemical modification. Zeolite 3A, 4A and 5A mean the zeolite before the chemical
modification, while zeolite 3A-NH2, 4A-NH2 and 5A-NH2 denote the zeolite after the
chemical modification. Results show that more nitrogen and carbon elements are
detected after the modification, thus confirming that the silane coupling agent has been
attached to the zeolite surface.
Table 6.1: Comparison of elementary analysis data of zeolites before and after the chemical modification
2.447.422.47zeolite 3A-NH2*
2.160.160zeolite 3A
2.307.302.42zeolite 4A-NH2*
2.120.160zeolite 4A
2.287.282.43zeolite 5A-NH2*
2.100.150zeolite 5A
H (%)C (%)N (%)Sample name
2.447.422.47zeolite 3A-NH2*
2.160.160zeolite 3A
2.307.302.42zeolite 4A-NH2*
2.120.160zeolite 4A
2.287.282.43zeolite 5A-NH2*
2.100.150zeolite 5A
H (%)C (%)N (%)Sample name
*: NH2 means the zeolite after the modification with APDEMS.
XPS spectra of zeolite 3A and 4A before and after the chemical modification are
exhibited in Figure 6.1. After modification, a peak clearly appears at the electronic
binding energy of about 397.2 eV that attributes to the nitrogen atom of the silane
coupling agent containing amino group. This indicates that some molecules of the
silane coupling agent have been grafted onto the external surface of zeolite, which is
in good agreement with the elementary analysis results.
167
AA BB
CC DD
Figure 6.1: Comparison of XPS spectra of the zeolite before and after the chemical modification (A: zeolite 3A; B: zeolite 3A-NH2; C: zeolite 4A; D: zeolite 4A-NH2)
That APDEMS is selected as the silane coupling agent because it has two ethoxy
groups (CH3CH2O) and an additional hydrophobic group (CH3). Compared to 3-
aminopropyl triethoxy silane with 3 ethoxy groups, APDEMS may lower the number
of coupling points with zeolite surface during the silanization reaction and thus avoid
blocking the micropore of zeolites. Table 6.2 lists the results of the multipoint BET
surface area and total pore volume of zeolites before and after the chemical
modification. There are no sharp changes in the surface area and total pore volume of
zeolites after the modification. These results imply that the addition of APDEMS does
not alter the micropore of zeolites.
168
Table 6.2: Comparison of total pore volume and multipoint BET surface area of zeolites before and after the chemical modification
556.7
561.6
553.5
after modifying
561.1
567.0
557.6
before modifying
multipoint BET surface area(m2/g)
after modifyingbefore modifying
0.196
0.198
0.196
0.198zeolite 5A
0.200zeolite 4A
0.197zeolite 3A
total pore volume(cm3/g)
sample name
556.7
561.6
553.5
after modifying
561.1
567.0
557.6
before modifying
multipoint BET surface area(m2/g)
after modifyingbefore modifying
0.196
0.198
0.196
0.198zeolite 5A
0.200zeolite 4A
0.197zeolite 3A
total pore volume(cm3/g)
sample name
6.2.2 Effect of Chemical Modification of Zeolite Surface on Gas Separation
Performance
PES-zeolite 3A, PES-zeolite 4A, and PES-zeolite 5A MMMs before and after the
chemical modification of zeolite surface were fabricated with a zeolite loading of 20
wt%. Figure 6.2 shows their permeability for 4 fast gases, while Figure 6.3
summarizes their selectivity for 4 gas pairs. Interestingly, different from the previous
studies [14, 16], the permeability of MMMs with modified zeolite is higher than those
MMMs with unmodified zeolite. This increasing trend in permeability seems counter-
intuitive at first because the contact between polymer and zeolite phases becomes
better as illustrated in Figure 6.4 and the leakage through the interface becomes
impossible after the modification.
169
0
4
8
PES-3A MMM PES-4A MMM PES-5A MMM
He permeability (Barrer) unmodified zeolite
modified zeoliteA
0
3.5
7
PES-3A MMM PES-4A MMM PES-5A MMM
H2 permeability (Barrer) unmodified zeolite
modified zeoliteB
0
0.2
0.4
0.6
PES-3A MMM PES-4A MMM PES-5A MMM
O2 permeability (Barrer) unmodified zeolite
modified zeolite
C
0
0.8
1.6
2.4
PES-3A MMM PES-4A MMM PES-5A MMM
CO2 permeability (Barrer) unmodified zeolite
modified zeoliteD
Figure 6.2: Comparison of gas permeability of PES-zeolite A MMMs before and after the chemical modification of zeolite surface (A: He permeability; B: H2 permeability;
C: O2 permeability; D: CO2 permeability)
170
70
95
120
PES-3A MMM PES-4A MMM PES-5A MMM
He/
N2 s
elec
tivity
unmodified zeolite
modified zeolite
A
60
85
110
PES-3A MMM PES-4A MMM PES-5A MMM
H2/
N2 s
elec
tivity
unmodified zeolite
modified zeolite
B
18
29
40
PES-3A MMM PES-4A MMM PES-5A MMM
CO2
/CH4
sele
ctiv
ity
unmodified zeolite
modified zeoliteD
5.8
6.4
7
PES-3A MMM PES-4A MMM PES-5A MMM
O2 /
N2 s
elec
tivity
unmodified zeolite
modified zeolite
C
Figure 6.3: Comparison of gas pair selectivity of PES-zeolite A MMMs before and after the chemical modification of zeolite surface (A: He/N2 selectivity; B: H2/N2
selectivity; C: O2/N2 selectivity; D: CO2/CH4 selectivity)
171
CA B
D E F
Figure 6.4: Comparison of cross-section SEM images of MMMs before and after the chemical modification of zeolite surface (A, B, C: PES-zeolite 3A, 4A and 5A MMMs,
respectively; D, E, F: PES-zeolite 3A-NH2, 4A-NH2 and 5A-NH2 MMMs, respectively)
Perhaps we can explain this increasing trend from an opposite perspective. Two
possible mechanisms have been proposed in the previous work [23-25] in order to
explain the significant decrease in gas permeability for MMMs; namely, the inhibition
of polymer chain mobility near the polymer-zeolite interface and the partial pore
blockage of zeolites induced by polymer chains. The first mechanism can be easily
detected by the Tg shift. Table 6.3 shows a comparison of Tg values of PES-zeolite A
MMMs made of modified and unmodified zeolite. Their Tg values are almost the same,
indicating the degree of polymer chain rigidification for these MMMs are almost the
same. Therefore, we may be able to rule out a decrease in polymer chain rigidification
as a possible cause for the permeability increase.
172
Table 6.3: Change of glass transition temperatures of MMMs over neat PES dense film before and after the chemical modification of zeolite surface
220±1oC220±1oCPES-zeolite 5A (40 wt% loading)218±1oC218±1oCPES-zeolite 4A
(40 wt% loading)
215±1oCneat PES dense film215±1oCneat PES dense
film
PES-zeolite 5A (50 wt% loading)
PES-zeolite 5A (30 wt% loading)
PES-zeolite 5A (20 wt% loading)
membrane name
221±1oC
218±1oC
217±1oC
after modifying
221±1oC
219±1oC
218±1oC
before modifying
Glass transition temperature (oC)
after modifying
before modifying
219±1oC
217±1oC
216±1oC
219±1oCPES-zeolite 4A (50 wt% loading)
217±1oCPES-zeolite 4A (30 wt% loading)
216±1oCPES-zeolite 4A (20 wt% loading)
Glass transition temperature (oC)
membrane name
220±1oC220±1oCPES-zeolite 5A (40 wt% loading)218±1oC218±1oCPES-zeolite 4A
(40 wt% loading)
215±1oCneat PES dense film215±1oCneat PES dense
film
PES-zeolite 5A (50 wt% loading)
PES-zeolite 5A (30 wt% loading)
PES-zeolite 5A (20 wt% loading)
membrane name
221±1oC
218±1oC
217±1oC
after modifying
221±1oC
219±1oC
218±1oC
before modifying
Glass transition temperature (oC)
after modifying
before modifying
219±1oC
217±1oC
216±1oC
219±1oCPES-zeolite 4A (50 wt% loading)
217±1oCPES-zeolite 4A (30 wt% loading)
216±1oCPES-zeolite 4A (20 wt% loading)
Glass transition temperature (oC)
membrane name
Previous researchers have measured the chain length of (3-aminopropyl)-triethoxy
silane molecules of around 5-9Å [27, 28] by means of spectroscopic ellipsometry.
Because APDEMS and (3-aminopropyl)-triethoxy silane have similar chemical
structures, APDEMS may have a molecular chain length of approximately 5-9Å. If the
attachment of the polymer matrix upon zeolite surface is the main cause of partial pore
blockage of zeolites [23-25], the addition of 5-9Å thick APDEMS on top of zeolite
surface may effectively reduce the direct partial pore blockage induced by the polymer
matrix. As a result, the gas permeability of MMMs increases if APDEMS modified
zeolite is used.
173
The same reasons may be applicable to explain the permselectivity increase if MMMs
are made of APDEMS modified zeolite as illustrated in Figure 6.3. Previous works
have demonstrated that, owing to partial pore blockage, the selectivity obtained from
the experiments is far below from the prediction calculated from the Maxwell model
[23-25]. Because of the unique structure of APDEMS, the thin APDEMS layer on
zeolite surface not only can reduce the pore blockage as discussed in the previous
section, but also can diminish the transport of gas penetrants via the interphase due to
better adhesion between APDEMS and the polymer matrix and the existence of
various functional groups in this space (i.e. CH2 and CH3). Therefore, the originally
high selectivity nature of zeolite remains almost intact and the resultant mixed matrix
membranes have a higher permselectivity.
6.2.3 Effect of Zeolite Loadings on Gas Permeability
PES-zeolite 4A-NH2 and PES-zeolite 5A-NH2 mixed matrix membranes with different
zeolite loadings were fabricated. Neat PES dense films were also prepared with the
same procedure for comparison. The permeability of all gases and selectivity of 4 gas
pairs at different zeolite loadings are shown in Figures 6.5 and 6.6, respectively. The
permeability of all gases decreases with an increase in zeolite content for PES-zeolite
4A-NH2 MMMs, while the relationship between gas permeability and zeolite loading
for PES-zeolite 5A-NH2 MMMs is not so straightforward. Their gas permeability
shows a decrease then an increase with increasing zeolite loading. The maximum
174
percentages of permeability increments of He, H2, O2 and N2 for PES-zeolite 5A-NH2
MMMs are approximately 73%, 67%, 35% and 11%, respectively, at 50 wt% zeolite
loading.
0
10
20
0 10 20 30 40 50 60Zeolite loading (%)
Perm
eabi
lity
(Bar
rer)
He of PES-modified 4A membranesHe of PES-modified 5A membranesH2 of PES-modified 4A membranesH2 of PES-modified 5A membranes
A
0
2
4
0 10 20 30 40 50 60Zeolite loading (%)
Perm
eabi
lity
(Bar
rer) CO2 of PES-modified 4A membranes
CO2 of PES-modified 5A membranesB
0
0.5
1
0 10 20 30 40 50 60Zeolite loading (%)
Perm
eabi
lity
(Bar
rer) O2 of PES-modified 4A membranes
O2 of PES-modified 5A membranesC
0
0.07
0.14
0 10 20 30 40 50 60Zeolite loading (%)
Perm
eabi
lity
(Bar
rer)
N2 of PES-modified 4A membranesN2 of PES-modified 5A membranesCH4 of PES-modified 4A membranesCH4 of PES-modified 5A membranes
D
Figure 6.5: Effect of zeolite loadings on gas permeability of PES-zeolite 4A-NH2 and PES-zeolite 5A-NH2 MMMs (A: He and H2 permeability; B: CO2 permeability; C: O2
permeability; D: N2 and CH4 permeability)
175
80
125
170
0 10 20 30 40 50 60Zeolite loading (%)
He/
N2 s
elec
tivity
PES-modified 4A membranes
PES-modified 5A membranes
A
70
105
140
0 10 20 30 40 50 60Zeolite loading (%)
H2/N
2 sel
ectiv
ity
PES-modified 4A membranes
PES-modified 5A membranes
B
5
7
9
0 10 20 30 40 50 60Zeolite loading (%)
O2/N
2 sel
ectiv
ity
PES-modified 4A membranes
PES-modified 5A membranesC
28
38
48
58
0 10 20 30 40 50 60Zeolite loading (%)
CO
2/CH
4 sel
ectiv
ity
PES-modified 4A membranes
PES-modified 5A membranes
D
Figure 6.6: Effect of zeolite loadings on gas pair selectivity of PES-zeolite 4A-NH2 and PES-zeolite 5A-NH2 MMMs (A: He/N2 selectivity; B: H2/N2 selectivity; C: O2/N2
selectivity; D: CO2/CH4 selectivity)
The decreasing trend of gas permeability of PES-zeolite 4A-NH2 MMMs with zeolite
loading may be easily understandable because it has been previously demonstrated
that both polymer chain rigidification and partial pore blockage of zeolites may lead to
a decrease in the permeability of MMMs [23-25]. However, for PES-zeolite 4A-NH2
MMMs, the polymer chain rigidification may play a much more important role than
the partial pore blockage. This is due to the fact that the surface modification by
APDEMS has lowered the degree of pore blockage in zeolite. In addition, the slight
increase in Tg with zeolite loading as shown in Table 6.3 implies greater chain
rigidification with increasing zeolite loading. As a result, permeability of PES-zeolite
176
4A-NH2 MMMs decreases with an increase in zeolite loading. A modified Maxwell
model [25] will be used to quantitatively estimate the effect of both polymer chain
rigidification and partial pore blockage of zeolites on the O2 permeability of PES-
zeolite 4A-NH2 MMMs in the next section.
However, why the permeability of PES-zeolite 5A-NH2 MMMs decreases and then
increases with increasing zeolite content as shown in Figure 6.5? This trend is
different from our previous results [25] where the permeability of PES-zeolite 5A
MMMs monotonously decreases with an increase in the zeolite loading. This
difference may arise from three factors: 1) The pore size of zeolite 5A, 4.8Å, is larger
than the molecular kinetic diameter of all tested gases; therefore, the permeability of
PES-zeolite 5A MMMs should increase with an increase in zeolite loadings, 2) The
effect of polymer chain rigidification may still play an important role when the zeolite
loading is small. The benefit of using large pore-size zeolite only magnifies when its
content is greater than approximately 30%, and 3) Even though the APDEMS surface
modification can reduce the degree of partial pore blockage, the partial pore blockage
cannot be eliminated completely. A part of zeolite 5A pores may be narrowed to
approximately 4Å, thus leading to a permeability decrease in PES-zeolite 5A MMMs.
Once the zeolite loading is further increased, the positive effect of using large pore-
size zeolite gradually offsets the negative effect of partial pore blockage on
permeability. As a consequence, permeability becomes increase with an increase in
zeolite loading.
177
The above three factors compete one another and result in different influence on the
relationship between gas permeability and zeolite loading for different gases. For He
and H2 gases with smallest kinetic diameters, their permeability starts to surpass that
of neat PES dense film from 20 wt% zeolite loading, while for O2 gas, it begins from
40 wt% loading; for N2, it starts from 50 wt% loading; and for CH4 gas, it is always
lower than that of neat PES dense film in our experimental range. The information
hints different gases exhibit different responses to polymer chain rigidification and
partial pore blockage of zeolites.
The above phenomena and arguments may not be applicable to the case of PES-zeolite
4A-NH2 MMMs. This arises from the fact that the pore size of zeolite 4A is about
3.8Å, which is very approximate with the kinetic diameter of tested gases. Therefore,
the APDEMS induced chemical modification cannot bring much positive effects on
the permeability vs. zeolite content relationship.
6.2.4 Effect of Zeolite Loadings on Gas Permselectivity
Figure 6.6 exhibits that the selectivity of both PES-zeolite 4A-NH2 and PES-zeolite
5A-NH2 MMMs increases with an increase in zeolite loadings. The maximum
increment occurs at 50 wt% zeolite loading. The maximum percentages of selectivity
increments of He/N2, H2/N2, O2/N2 and CO2/CH4 for PES-zeolite 4A-NH2 MMMs are
178
approximately 33%, 34%, 32% and 44%, respectively, and for PES-zeolite 5A-NH2
MMMs are around 56%, 51%, 22% and 15%, respectively. Amazingly, PES-zeolite
5A-NH2 MMMs not only increase the permeability of He, H2 and O2, but also enhance
the selectivity of He/N2, H2/N2 and O2/N2, especially at 50 wt% loading. This
simultaneous advance in both permeability and selectivity is very attractive to
potential industrial applications.
The increasing trend in gas pair selectivity of both PES-zeolite 4A-NH2 and PES-
zeolite 5A-NH2 MMMs with an increase in zeolite loadings is easily explainable due
to the influence of molecular sieving mechanism via zeolite. Although the magnitude
of increment is still below from the calculation of the Maxwell model, the gas
separation performance of the newly developed MMMs made from APDEMS
modified zeolite is much higher than those made from un-modified zeolite [25].
6.2.5 Applicability of the Modified Maxwell Model to Predict Gas Separation
Performance
In our previous work [25], a modified Maxwell model was proposed to predict the gas
separation performance. It takes the combined effects of the polymer chain
rigidification and partial pore blockage of zeolites into calculation. This modified
Maxwell model needs intrinsic gas permeation properties of zeolites to predict the gas
separation performance of MMMs in the calculation. So far only O2 and N2
179
permeability data of zeolite 4A can be found in the literatures and relatively consistent,
while there are no consistent CO2 and CH4 permeability data of zeolite 4A available.
For other zeolites, such as 5A, Silicalite-1, no consistent data on intrinsic gas
permeability can be found in the literatures. Therefore, the PES-zeolite 4A-NH2
MMM is chosen as a sample to predict O2 permeability and O2/N2 selectivity by using
this modified Maxwell model in this study. This exercise may provide convincing
information to quantitatively estimate the effect of chemical modification on the
polymer chain rigidification and partial pore blockage of zeolites.
Rigidified polymer region Prig = PC/β
Zeolite skin affected by the partial pore blockage
Pblo = PD/β’
The second phase [the third phase (dispersed phase) + the rigidified
polymer region (continuous phase)]
The first phase (polymer matrix)
r’r r’r
The third phase [the bulk of zeolite (dispersed phase) + the zeolite skin with
partial pore blockage (continuous phase)]
Polymer chains
Zeolite 4A
Bulk of zeolite PD
Polymer matrix PC
3.8Ǻ
Figure 6.7: Schematic diagram for the modified Maxwell model
180
Firstly, we assume the whole system of mixed matrix membranes as a pseudo three-
phase composite, as illustrated in Figure 6.7. The first phase is the polymer matrix; the
second phase is composed of the third phase and rigidified polymer region near the
zeolite, and the third phase comprises the bulk of zeolite and the zeolite skin affected
by the partial pore blockage. Secondly, the permeability of the third phase can be
calculated with the aid of the Maxwell equation [29] by assuming the bulk of zeolite
as the dispersed phase and the zeolite skin affected by the partial pore blockage as the
continuous phase.
])-(2)-(2-2
[3
33
DblobloD
DblobloDblord PPvPP
PPvPPPP
+++
= (6.1)
here P3rd is the composite permeability of the third phase, PD is the permeability of the
bulk of zeolite, Pblo is the permeability of the zeolite skin affected by the partial pore
blockage, and v3 is the volume fraction of the bulk of zeolite in the third phase and can
be determined by the following equation:
bloD
D
vvvv+
=3 (6.2)
where vD is the volume fraction of the bulk of zeolite in the total membrane and vblo is
the volume fraction of zeolite skin affected by the partial pore blockage in the total
membrane.
181
Thirdly, the permeability of the second phase can be determined with the aid of
Maxwell equation a second time by regarding the rigidified polymer region as the
continuous phase and the third phase as the dispersed phase.
])-(2)-(2-2
[323
3232
rdrigrigrd
rdrigrigrdrignd PPvPP
PPvPPPP
++
+= (6.3)
here P2nd is the composite permeability of the second phase, Prig is the permeability of
the rigidified polymer region, and v2 is the volume fraction of the third phase in the
second phase and can be estimated by the following equation:
rigbloD
bloD
vvvvv
v++
+=2 (6.4)
where vrig is the volume fraction of the rigidified polymer region in the total membrane.
Finally, one can obtain the permeability of the whole mixed matrix membrane with the
aid of Maxwell equation a third time by considering the polymer matrix as the
continuous phase (i.e., the first phase) and the second phase as the dispersed phase.
])-)((2)-)((2-2
[22
22
ndCrigbloDCnd
ndCrigbloDCndCMMM PPvvvPP
PPvvvPPPP
++++
+++= (6.5)
here PMMM is the composite permeability of mixed matrix membranes and PC is the
permeability of the polymer matrix (i.e., the first phase).
Zeolite 4A Linde crystals have an O2 permeability of around 0.77 Barrer and an O2/N2
selectivity of around 37 at 35oC [30]. The neat PES dense film has an O2 permeability
of around 0.479 Barrer and an O2/N2 selectivity of 5.81 at 35oC according to our
182
experimental results. Therefore, four unknowns in Equations (6.1)-(6.5) need to be set
to predict the resultant permeability of mixed matrix membranes. They are Pblo, v3, Prig
and v2, respectively. Prig and Pblo can be assumed to be the permeability of the polymer
matrix divided by a chain immobilization factor β and the permeability of the bulk of
zeolite divided by an average decline factor β’ induced by the partial pore blockage,
respectively [23-25, 31]. v2 and v3 can be calculated if the volume fraction of the
rigidified polymer region and affected zeolite skin in the total membrane are known.
The polymer chain rigidification near the zeolite surface is probably resulted from the
inhibited isotropic contraction of the polymer. Zeolite with a smaller particle size may
have a less effect on the inhibition of the polymer contraction, thus leading to a
smaller thickness of rigidified polymer region [24]. The thickness of rigidified
polymer region near the zeolite, r, has been assumed to be 0.66µm [24], 0.3µm [25]
and 0.066µm [24] as the average particle size of zeolite 4A or carbon molecular sieve
is 5µm [24], 3µm [25] and 1µm [24], respectively. Therefore, r is assumed to be
0.12µm in this work because the average particle size of used zeolite 4A is 1.5µm.
This inhibited isotropic contraction is assumed to affect only the thickness of rigidified
polymer region, but no effect on the extent of polymer chain rigidification [24].
Therefore, the chain immobilization parameter β is still assumed to be 3 for O2 gas and
4 for N2 in this work which are the same as our previous work [25]. The partial pore
blockage of zeolites is resulted from the attachment of polymer chains on the zeolite
surface. The thickness of this affected zeolite skin, r’, is assumed to be almost
183
unchanged with zeolite particle size [23-25]. However, the chemical modification of
zeolite surface may reduce the extent of the partial pore blockage, so in this work, r’ is
assumed to be 70Å lower than that (i.e., about 100Å or more) in other related
literatures [23-25]. To simplify the calculation, we presume neither the particle size of
zeolites nor the chemical modification has any influence on the magnitude of
permeability reduction in this zeolite skin affected by the partial pore blockage;
therefore, β’ is still assumed to 250 for O2 and 10 for N2 in this work. The reason that
β and β’ have different values for different gases has been explained in our previous
work [25]. O2 and N2 permeability data in different regions of PES-zeolite 4A-NH2
MMMs are shown in Table 6.4. All calculated volume fraction data of dispersed phase
defined in Figure 6.7 in the third phase, second phase and whole MMM in this
modified Maxwell model are listed in Table 6.5. The calculation method to obtain
these volume fractions has been discussed extensively by Moore et al. [24], so it is not
necessary to duplicate it in this work.
184
Table 6.4: Change of O2 and N2 permeability in different regions of the PES-zeolite 4A-NH2 MMM
0.7700.003080.1600.479 O2
0.0825
Permeability in the PES matrix (PC)a
0.0206
Permeability in the rigidified PES region
(Prig=PC/βc)
0.00208
Permeability in the zeolite 4A skin affected with partial pore blockage (Pblo=PD/β’d)
0.0208N2
Permeability in the bulk of zeolite 4A
(PD)b
0.7700.003080.1600.479 O2
0.0825
Permeability in the PES matrix (PC)a
0.0206
Permeability in the rigidified PES region
(Prig=PC/βc)
0.00208
Permeability in the zeolite 4A skin affected with partial pore blockage (Pblo=PD/β’d)
0.0208N2
Permeability in the bulk of zeolite 4A
(PD)b
a: Gas permeability data in the PES matrix come from experiment results of neat PES dense film.
b: Gas permeability data in the bulk of zeolite 4A come from the reference [30].
c: β is 3 for O2 gas, while β is 4 for N2 gas.
d: β’ is 250 for O2 gas, while β’ is 10 for N2 gas.
Table 6.5: Calculated volume fraction data of dispersed phase in different phases of PES-zeolite 4A-NH2 MMMs in the modified Maxwell model which simultaneously
considers both polymer chain rigidification and partial pore blockage of zeolites
0.29020 wt% zeolite loading (18.6 vol% zeolite loading)
0.43930 wt% zeolite loading (28.1 vol% zeolite loading)
0.59040 wt% zeolite loading (37.8 vol% zeolite loading)
0.74550 wt% zeolite loading (47.7 vol% zeolite loading)
0.641Calculated volume fraction of the third phase (considered as the dispersed phase) in the second phase (v2)
(defined in Figure 8)
Calculated volume fraction of the second phase (considered as the dispersed phase) in the whole mixed matrix membrane
(defined in Figure 8)
Calculated volume fraction of the bulk of zeolite 4A (considered as the dispersed phase) in the third phase (v3)
(defined in Figure 8)
0.972
0.29020 wt% zeolite loading (18.6 vol% zeolite loading)
0.43930 wt% zeolite loading (28.1 vol% zeolite loading)
0.59040 wt% zeolite loading (37.8 vol% zeolite loading)
0.74550 wt% zeolite loading (47.7 vol% zeolite loading)
0.641Calculated volume fraction of the third phase (considered as the dispersed phase) in the second phase (v2)
(defined in Figure 8)
Calculated volume fraction of the second phase (considered as the dispersed phase) in the whole mixed matrix membrane
(defined in Figure 8)
Calculated volume fraction of the bulk of zeolite 4A (considered as the dispersed phase) in the third phase (v3)
(defined in Figure 8)
0.972
Figures 6.8 and 6.9 illustrate the ultimate prediction of O2 permeability and O2/N2
selectivity of PES-zeolite 4A-NH2 MMMs, respectively. It can be found that the
185
prediction from the modified Maxwell model is in very good agreement with
experimental data. Therefore, this demonstrates our revised Maxwell model can be
utilized for MMMs made from both unmodified and modified zeolite.
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50 60
Zeolite loading (%)
Perm
eabi
lity
(Bar
rer)
O2 (experimental data)
O2 (prediction by the Maxwell model)
O2 (prediction by the modified Maxwell model assuming polymer chainrigidification and partial pore blockage of zeolites)
Figure 6.8: Comparison of O2 permeability of PES-zeolite 4A-NH2 MMMs between experimental data and predictions from the Maxwell model and the modified Maxwell
model
186
5
7
9
11
13
0 10 20 30 40 50 60
Zeolite loading (%)
O2 /
N2 s
elec
tivity
O2/N2 (experimental data)O2/N2 (pridiction by the Maxwell model)O2/N2 (pridiction by the modified Maxwell model)
Figure 6.9: Comparison of O2/N2 selectivity of PES-zeolite 4A-NH2 MMMs between experimental data and predictions from the Maxwell model and the modified Maxwell
model
6.3 CONCLUSIONS
The following conclusions can be drawn from this chapter:
(1) Results from elementary analysis and XPS demonstrate that the silane coupling
agent (APDEMS) has been successfully grafted onto the external surface of zeolite
after the chemical modification. BET results show that the surface area and total pore
volume of zeolites before and after modification are almost the same, indicating the
micropore of zeolites is almost not influenced by the addition of APDEMS. These
187
characterizations assure the zeolite modified by APDEMS as a good candidate which
can be applied in the fabrication of MMMs.
(2) Both gas permeability and gas pair selectivity of PES-zeolite A-NH2 MMMs are
higher than those of PES-zeolite A MMMs at 20 wt% zeolite loading. This may be
mainly due to the fact that the presence of APDEMS introduces a distance of around
5-9Å between polymer chains and zeolite surface, thus reduces the extent of the partial
pore blockage of zeolites induced by polymer chains.
(3) The permeability of all studied gases decreases with increasing zeolite content for
PES-zeolite 4A-NH2 MMMs, while for PES-zeolite 5A-NH2 MMMs, the gas
permeability decreases and then increases with an increase in zeolite loadings. Clearly,
a high loading of zeolite 5A-NH2 offsets the effects of partial pore blockage and
polymer chain rigidification on permeability. The selectivity of both PES-zeolite 4A-
NH2 and PES-zeolite 5A-NH2 MMMs increases with an increase in zeolite loadings
due to the influence of molecular sieving mechanism.
(4) A modified Maxwell model proposed in our previous work [25] was applied to
predict the gas separation performance of the newly developed PES-zeolite 4A-NH2
MMMs with adjusted parameters. The predicted permeability and selectivity show
very good agreement with experimental data.
188
6.4 REFERENCES
[1] G.W. Peng, F. Qiu, V.V. Ginzburg, D. Jasnow, A.C. Balazs, Forming
supramolecular networks from nanoscale rods in binary, phase-separating mixtures,
Science, 288 (2000) 1802.
[2] T.C. Merkel, B.D. Freeman, R.J. Spontak, Z. He, I. Pinnau, P. Meakin, A.J. Hill,
Ultrapermeable, reverse-selective nanocomposite membranes, Science, 296 (2002)
519.
[3] Y. Wang, N. Herron, X-ray photoconductive nanocomposites, Science, 273 (1996)
632.
[4] S. Kulprathipanja, R.W. Neuzil, N.N. Li, Separation of fluids by means of mixed
matrix membranes in gas permeation, US Patent No. 4,740,219 (1988).
[5] H.J.C. Te Hennepe, Zeolite Filled Polymeric Membranes: A New Concept in
Separation Science, Thesis, University of Twente, 1988.
[6] M. Jia, K.V. Peinemann, R.D. Behling, Molecular sieving effect of zeolite filled
silicone rubber membranes, J. Membr. Sci., 57 (1991) 289.
[7] J. M. Duval, B. Folkers, M.H.V. Mulder, G. Desgrandchamps, C.A. Smolders,
Adsorbent-filled membranes for gas separation. Part 1. Improvement of the gas
separation properties of polymeric membranes by incorporation of microporous
adsorbents, J. Membr. Sci., 80 (1993) 189.
[8] D.R. Paul, D.R. Kemp, The diffusion time lag in polymer membranes containing
adsorptive fillers, J. Polym. Sci. Symp., 41 (1973) 79.
189
[9] M.G. Süer, N. Baç, L. Yilmaz, Gas permeation characteristics of polymer-zeolite
mixed matrix membranes, J. Membr. Sci., 91 (1994) 77.
[10] R. Mahajan, W.J. Koros, Factors controlling successful formation of mixed-
matrix gas separation materials, Ind. Eng. Chem. Res., 39 (2000) 2692.
[11] A. Erdem-Şenatalar, M. Tatlıer, Ş.B. Tantekin-Ersolmaz, Estimation of the
interphase thickness and permeability in polymer-zeolite mixed matrix membrane,
in: Proceedings of the 13th International Zeolite Conference, Montpellier, France,
2001.
[12] Ş.B. Tantekin-Ersolmaz, L. Şenorkyan, N. Kalaonra, M. Tatlıer, A. Erdem-
Şenatalar, n-Pentane/i-pentane separation by using zeolite-PDMS mixed matrix
membranes, J. Membr. Sci., 189 (2001) 59.
[13] R. Mahajan, R. Burns, M. Schaeffer, W.J. Koros, Challenges in forming
successful mixed matrix membranes with rigid polymeric materials, J. App.
Polym. Sci., 86 (2002) 881.
[14] R. Mahajan, W.J. Koros, Mixed matrix membrane materials with glassy
polymers. Part 1, Polym. Eng. & Sci., 42 (7) (2002) 1420.
[15] R. Mahajan, W.J. Koros, Mixed matrix membrane materials with glassy
polymers. Part 2, Polym. Eng. & Sci., 42 (7) (2002) 1432.
[16] J.M. Duval, A.J.B. Kemperman, B. Folkers, M.H.V. Mulder, G.
Desgrandchamps, C.A. Smolders, Preparation of zeolite filled glassy polymer
membrane, J. Appl. Polym. Sci., 54 (1994) 409.
190
[17] Y.L. Liu, Y.H. Su, K.R. Lee, J.Y. Lai, Crosslinked organic-inorganic hybrid
chitosan membranes for pervaporation dehydration of isopropanol-water
mixtures with a long-term stability, J. Membr. Sci., 251 (2005) 233.
[18] H.H. Yong, H.C. Park, Y.S. Kang, J. Won, W.N. Kim, Zeolite-filled polyimide
membrane containing 2,4,6-triaminopyrimidine, J. Membr. Sci., 188 (2001) 151.
[19] A. Matsumoto, K. Tsutsumi, K. Schumacher, K.K. Unger, Surface
functionalization and stabilization of mesoporous silica spheres by silanization
and their adsorption characteristics, Langmuir, 18 (2002) 4014.
[20] E.P. Plueddemann, Silane coupling agents, 2nd ed., Plenum Press, New York,
1991.
[21] D.H. Flinn, D.A. Guzonas, R.H. Yoon, Characterization of silica surfaces
hydrophobized by octadecyltrichlorosilane, Colloids Surfaces A: Physicochem.
Eng. Aspects, 87 (1994) 163.
[22] K.C. Vrancken, K. Possemiers, P. Van Der Voort, E.F. Vansant, Surface
modification of silica gels with aminoorganosilanes, Colloids Surfaces A:
Physicochem. Eng. Aspects, 98 (1995) 235.
[23] R. Mahajan, Formation, characterization and modeling of mixed matrix
membrane materials, Ph.D. Dissertation, The University of Texas at Austin,
2000.
[24] T.T. Moore, R. Mahajan, D.Q. Vu, W.J. Koros, Hybrid membrane materials
comprising organic polymers with rigid dispersed phases, AICHE Journal, 50 (2)
(2004) 311.
191
[25] Y. Li, T.S. Chung, C. Cao, S. Kulprathipanja, The effects of polymer chain
rigidification, zeolite pore size and pore blockage on polyethersulfone (PES)-
zeolite A mixed matrix membranes, J. Membr. Sci., 260 (2005) 45.
[26] J.S. Chiou, Y. Maeda, D.R. Paul, Gas permeation in polyethersulfone, J. App.
Polym. Sci., 33 (1987) 1823.
[27] V.V. Tsukruk, V.N. Bliznyuk, Adhesive and friction forces between chemically
modified silicon and silicon nitride surfaces, Langmuir, 14 (1998) 446.
[28] D.F. Siqueira Petri, G. Wenz, P. Schunk, T. Schimmel, An improved method for
the assembly of amino-terminated monolayers on SiO2 and the vapor deposition
of gold layers, Langmuir, 15 (1999) 4520.
[29] K. Mizoguchi, K. Terada, Y. Naito, Y. Kamiya, S. Tsuchida, S. Yano, Miscibility
blends and gas permeability of poly(ethylene-co-3,5,5-trimethylhexyl
methacrylate)-polydimethylsiloxane, Colloid. Polym. Sci., 86 (1997) 275.
[30] C.M. Zimmerman, A. Singh, W.J. Koros, Tailoring mixed matrix composite
membranes for gas separation, J. Membr. Sci., 137 (1997) 145.
[31] A.S. Michaels, R.B. Parker, Sorption and flow of gases in polyethylene, J. Polym.
Sci., 41 (138) (1959) 53.
192
CHAPTER SEVEN
DUAL-LAYER POLYETHERSULFONE (PES)/BTDA-TDI/MDI CO-
POLYIMIDE (P84) HOLLOW FIBER MEMBRANES WITH A SUBMICRON
PES-ZEOLITE BETA MIXED MATRIX DENSE-SELECTIVE LAYER FOR
GAS SEPARATION
7.1 INTRODUCTION
During the last few decades, the membrane-based gas separation technology has held a
part of market share in competition with the traditional separation processes due to its
various advantages such as low capital investment, ease of operation and low energy
consumption [1]. However, the further development of polymeric membrane
separation technology has been constrained by a performance “upper bound” tradeoff
curve between the gas productivity and permselectivity [2]. To expand the industrial
application of membrane separation technology, it is very necessary to enhance the gas
permeation flux (productivity) and permselectivity by combining the synthesis of
high-performance membrane materials with the innovation of membrane fabrication
technology. Following the discovery by researchers at UOP LLC on mixed matrix
membranes (MMMs) [3], the latest emerging polymer-based organic-inorganic
composite membrane materials may potentially surpass this “upper bound” limit by
193
means of combining the easy processability of organic polymers with the excellent gas
separation properties of inorganic molecular sieve materials.
Significant progress has been made in MMMs under the endeavor of researchers in
spite of using rubbery or glassy polymers as the matrix [4-18]. There are mainly three
approaches to reduce or eliminate the voids between polymer and molecular sieve
phases, which is the largest challenge to fabricate ideal MMMs, especially for glassy
polymers. The first one is to promote the adhesion between polymer matrix and
molecular sieve phases by modifying the zeolite surface with silane coupling agents
[12, 13, 18]. The second one is to introduce a compatibilizer to fill the voids between
polymer and molecular sieve phases [14]. The third one is to apply high processing
temperatures close to Tg of polymeric materials to maintain the polymer chain
flexibility during the membrane formation [11, 15, 17, 18]. However, the above MMM
studies focused on the formation of flat dense membranes. The flat dense MMMs can
provide the intrinsic properties of this type of organic-inorganic composite materials
for the academic research; however, they are not appropriate for the industrial
application due to its much thicker dense selective layer and much lower gas
permeation flux.
In the last 20 years, the asymmetric hollow fiber membrane is a favorable
configuration in the membrane-based gas separation systems due to its various
advantages [19-24]. Therefore, it will be a great development if applying this type of
194
promising polymer-based organic-inorganic composite materials with the
configuration of hollow fibers for gas separation. Some pioneering progresses have
been made [25-30]. The asymmetric hollow fiber membrane is composed of a dense
selective layer and a porous supporting layer, and its gas separation performance is
mainly determined by the dense selective layer. Therefore, how to distribute most of
molecular sieves in the dense selective layer has become an inevitable problem to
make full use of the effect of molecular sieves on the improvement of gas separation
performance. Jiang et al. [29] have demonstrated that three factors played important
roles in determining the distribution of molecular sieves in the resultant hollow fibers.
They are (1) shear stress within the spinneret, (2) die swell when exiting from the
spinneret and (3) elongation drawing in the air gap region. Based on their study, a
majority of molecular sieves were located in the middle part of porous supporting
layer when applying mixed matrix membrane materials in the formation of single-
layer hollow fibers, and could not play any role in improving the gas separation
performance.
Therefore, the dual-layer hollow fiber has to be used to fulfill the application of
organic-inorganic composite materials in hollow fibers [29, 30]. The dual-layer hollow
fiber membranes formed by the co-extrusion technology represent a breakthrough in
the hollow fiber fabrication technology, and basically consist of an asymmetric
separating outer layer and a microporous supporting inner layer. Dual-layer hollow
fibers made from neat polymeric materials have been investigated extensively [31-37]
195
because they can provide a solution for the wide application of high-performance but
expensive polymeric materials by using these materials in the outer layer instead of the
whole fiber. Similarly, when hybrid organic-inorganic composite materials are applied
in the dual-layer hollow fiber membranes, one can efficiently lower the material cost
and easily control the distribution of molecular sieves by using composite materials
only in the thin outer layer. Jiang et al. in our research group have successfully
fabricated the dual-layer hollow fibers with a mixed matrix outer layer, and
significantly enhanced the He/N2 and O2/N2 selectivity compared with neat polymer
membranes by a combination of the heat-treatment and two-step coating methods [30].
Regrettably, although the heat-treatment method really narrowed voids between
polymer and zeolite phases to the range that can be cured by silicon rubber coating, it
also greatly decreased the gas permeation flux due to densifying the whole mixed
matrix outer layer. In their work, the calculated thinnest mixed matrix dense-selective
layer thickness is around 2.5µm based on the intrinsic O2 permeability of
corresponding flat dense polysulfone-zeolite Beta MMMs [30].
To fully compete and potentially replace the hollow fiber membranes made from neat
polymeric materials, one must develop the co-extruding technology to fabricate dual-
layer hollow fiber membranes with a mixed matrix dense selective layer thickness of
less than 1µm at the same time keeping the advanced separation performance of mixed
matrix materials. To our best knowledge, so far there is no academic literature
available on this subject. Therefore, the objective of this study is to fabricate dual-
196
layer hollow fiber membranes with a mixed matrix dense selective layer thickness of
less than 1µm. In the mixed matrix outer layer of dual-layer hollow fibers,
polyethersulfone (PES) was chosen as the continuous phase and self-synthesized
zeolite Beta in our lab was used as the dispersed phase. A glassy BTDA-TDI/MDI co-
polyimide commercially named as P84 was selected as the polymer material of the
microporous inner layer due to its excellent thermal and mechanical properties. The
enormous difference between the glass transition temperatures (Tgs) of PES (215oC)
and P84 (315oC) [38] permits us to heat-treat the PES-zeolite Beta mixed matrix outer
layer to reduce or eliminate the voids between two phases without damaging the
microporous structure of inner layer, similar to the previous work in our group [30].
The intrinsic permeation properties of flat dense neat PES membrane and PES-zeolite
Beta MMM measured in our group are listed in Table 7.1 for comparison and
calculation in the subsequent sections.
Table 7.1: Intrinsic permeation properties of flat dense neat PES membrane and PES-zeolite Beta MMM
Both flat dense membranes were tested at 35oC and 10atm.
32.66.41.630.0500.0510.33Flat dense PES-zeolite beta MMM (20 wt% zeolite loading)
5.9
P(O2)/P(N2)
29.9
P(CO2)/P(CH4)Ideal Selectivity Permeability (Barrer)
2.510.0840.0820.48Flat dense neat PES membrane
CO2CH4N2O2
32.66.41.630.0500.0510.33Flat dense PES-zeolite beta MMM (20 wt% zeolite loading)
5.9
P(O2)/P(N2)
29.9
P(CO2)/P(CH4)Ideal Selectivity Permeability (Barrer)
2.510.0840.0820.48Flat dense neat PES membrane
CO2CH4N2O2
197
7.2 RESULTS AND DISCUSSION
7.2.1 Effects of Heat-treatment Temperature on Gas Separation Performance
The dual-layer hollow fiber membranes with a mixed matrix outer layer were
fabricated by the co-extrusion technique using a dual-layer spinneret as depicted in
relative literatures [29, 30, 33]. The flow rates of the bore fluid and both dope
solutions were controlled by three ISCO pumps. The detailed spinning conditions and
parameters were listed in Table 7.2.
Table 7.2: Spinning conditions and parameters of dual-layer PES/P84 hollow fiber membranes with a mixed matrix outer layer
116.8
0
0.35
1:6.7
1.0
0.15
DL3A
170.9
0
0.55
1:10
1.5
0.15
DL3B
170.9
0
0.55
1:15
1.5
0.10
DL3C
1:5.3Ratio of out layer dope flow rate to inner layer dope flow rate
25
25
Tap water
96.9
0
0.35
NMP/H2O: 95/5 (in wt%)
0.80
P84/NMP/ethanol: 20/67/13 (in wt%)
0.15
PES/NMP/ethanol: 35/50/15 (in wt%) + 20 wt% zeolite loading of beta in total solid
DL2A
119.0
0.5
0.35
0.80
0.15
DL2B
141.1
1.5
0.35
0.80
0.15
DL2C
Air gap (cm)
Inner layer dope composition
Inner layer dope flow rate (ml/min)
Spinning temperature (oC)
Take up rate (cm/min) (keep free falling)
External coagulant
Coagulation bath temperature (oC )
Bore fluid flow rate (ml/min)
Bore fluid composition
Out layer dope flow rate (ml/min)
Outer layer dope composition
ID
116.8
0
0.35
1:6.7
1.0
0.15
DL3A
170.9
0
0.55
1:10
1.5
0.15
DL3B
170.9
0
0.55
1:15
1.5
0.10
DL3C
1:5.3Ratio of out layer dope flow rate to inner layer dope flow rate
25
25
Tap water
96.9
0
0.35
NMP/H2O: 95/5 (in wt%)
0.80
P84/NMP/ethanol: 20/67/13 (in wt%)
0.15
PES/NMP/ethanol: 35/50/15 (in wt%) + 20 wt% zeolite loading of beta in total solid
DL2A
119.0
0.5
0.35
0.80
0.15
DL2B
141.1
1.5
0.35
0.80
0.15
DL2C
Air gap (cm)
Inner layer dope composition
Inner layer dope flow rate (ml/min)
Spinning temperature (oC)
Take up rate (cm/min) (keep free falling)
External coagulant
Coagulation bath temperature (oC )
Bore fluid flow rate (ml/min)
Bore fluid composition
Out layer dope flow rate (ml/min)
Outer layer dope composition
ID
The largest challenge when applying mixed matrix membrane materials in the dual-
layer hollow fibers is to minimize the voids induced by the unfavorable interaction
198
between polymer and zeolite phases. The size of these voids is normally in the range
of tens of nanometers and cannot be cured only by silicon rubber coating; therefore, it
is obvious that such hollow fibers will present the Knudsen diffusion even after
coating with silicon rubber. Jiang et al [30] have demonstrated the heat-treatment is an
effective method to narrow voids to the range that can be cured by silicon rubber
coating. The use of heat-treatment to control pore size for composite membranes has
been often practiced in industry [39, 40] and their effects on membrane morphology
have been summarized elsewhere [22]. The heat-treatment induces the relaxation of
the stresses imposed in hollow fibers as spinning, thus leading to higher packing
density of polymer chains and reducing the outer surface defects. Figure 7.1 shows the
comparison of SEM morphologies of dual-layer PES/P84 hollow fibers with a mixed
matrix outer layer before and after the heat-treatment, where the heat-treatment
temperature is 235oC which is 20oC above the Tg (215oC) of the neat PES material. It
can be seen that the heat-treatment not only eliminates the delamination between the
outer layer and inner layer, but also remarkably reduces voids between polymer and
zeolite phases. However, it can be easily predicted that the heat-treatment will greatly
decrease the gas permeation flux and increase the dense selective layer thickness due
to densifying the whole mixed matrix outer layer, as shown in the study of Jiang et al
[30].
199
A B C
D E F
Figure 7.1: Comparison of SEM morphologies of dual-layer PES/P84 hollow fiber membranes with a mixed matrix outer layer for DL2A before and after the heat-
treatment (A, B and C: overall profile, outer layer’s outer edge and outer layer’ outer surface of as-spun hollow fibers, respectively; D, E and F: overall profile, outer layer’s
outer edge and outer layer’ outer surface of hollow fibers heat-treated at 235oC, respectively)
In order to increase the gas productivity and decrease the dense selective layer
thickness, there are two relatively straightforward methods. One is to decrease the
heat-treatment temperature; the other is to make the mixed matrix outer layer as thin
as possible. In this section, we will investigate the effect of different heat-treatment
temperatures on the gas separation performance.
200
Table 7.3: The effects of heat-treatment temperature on the gas separation performance of DL2A dual-layer PES/P84 hollow fiber membranes with a mixed
matrix outer layer
32.932.732.6Ideal selectivity(CO2/CH4)
Before coating
5.25.45.8Calculated dense selective layer thickness (μm)
0.007960.007310.00561Permeance (CH4) (GPU)
0.2620.2390.183Permeance (CO2) (GPU)
0.936.566.566.60Ideal selectivity(O2/N2)
86.30.009680.009330.00867Permeance (N2) (GPU)
80.00.06350.06120.0572Permeance (O2) (GPU)
DL2A-235oC DL2A-220oC DL2A-210oC
0.95
2.77
2.63
0
DL2A-200oC
Air gap (cm)
After coating
Permeance (N2) (GPU)
Ideal selectivity(O2/N2)
Permeance (O2) (GPU)
hollow fiber sample name
32.932.732.6Ideal selectivity(CO2/CH4)
Before coating
5.25.45.8Calculated dense selective layer thickness (μm)
0.007960.007310.00561Permeance (CH4) (GPU)
0.2620.2390.183Permeance (CO2) (GPU)
0.936.566.566.60Ideal selectivity(O2/N2)
86.30.009680.009330.00867Permeance (N2) (GPU)
80.00.06350.06120.0572Permeance (O2) (GPU)
DL2A-235oC DL2A-220oC DL2A-210oC
0.95
2.77
2.63
0
DL2A-200oC
Air gap (cm)
After coating
Permeance (N2) (GPU)
Ideal selectivity(O2/N2)
Permeance (O2) (GPU)
hollow fiber sample name
All fibers were tested at 35oC. The testing pressure was 10atm for O2 and N2, and 8atm for CH4 and CO2.All fibers were coated with the two-step coating method if they showed Knudsen diffusion before coating.
Table 7.3 summarizes the gas separation performance and the calculated dense
selective layer thickness of dual-layer hollow fiber membranes with the mixed matrix
outer layer for DL2A fibers at different heat-treatment temperatures. The apparent
dense-selective layer thickness is calculated based on the intrinsic O2 permeability
(0.33 Barrer) of flat dense PES-zeolite Beta MMMs shown in Table 7.1. There are
three points worth notice in Table 7.3. The first point is that the O2/N2 (~6.6) and
CO2/CH4 (~33) selectivity of these dual-layer hollow fibers increases around 10%
compared with 5.9 and 29.9 of the neat PES dense film shown in Table 7.1 when the
heat-treatment temperature is at or above 210oC. The enhanced selectivity may be due
to two reasons: polymer chain rigidification and partial pore blockage of zeolite Beta.
201
The effects of these two factors on the gas separation performance will be discussed in
the next section.
The second point is that the advanced selectivity is obtained only by the heat-treatment
at or above 210oC for the DL2A fibers. Normally, PES hollow fiber membranes
cannot present the good selectivity before silicon rubber coating due to the existence
of defects [37]. This indicates that the heat-treatment at or above 210oC is enough to
completely eliminate the defects on the outer surface of dual-layer hollow fibers for
the DL2A fibers, which may greatly simplify the membrane module fabrication.
The third point is that the calculated apparent dense-selective layer thickness only
decreases slightly with a reduction in heat-treatment temperature from 235oC to 210oC,
which is far from our objective “a mixed matrix dense selective layer thickness of less
than 1µm”. With a further decrease in heat-treatment temperature (i.e. 200oC), dual-
layer hollow fibers show the Knudsen diffusion even after a two-step silicon rubber
coating. The results imply that 200oC is not enough high to reduce voids to the range
that can be cured by the subsequent silicon rubber coating. Therefore, decreasing heat-
treatment temperature is not an effective method to fulfill our purpose.
7.2.2 Effect of Mixed Matrix Outer-layer Thickness on Gas Separation
Performance
202
By means of adjusting the ratio of outer-layer flow rate to inner-layer flow rate during
the spinning, one may produce the mixed matrix outer layer as thin as possible. Figure
7.2 shows partial cross-section SEM images of dual-layer PES/P84 hollow fiber
membranes with different mixed matrix outer-layer thicknesses after heat-treated at
235oC. The outer-layer thicknesses after heat-treatment vary from around 7 to 1.8µm.
A B C
D FE
Figure 7.2: Comparison of partial cross-section SEM images of dual-layer PES/P84 hollow fiber membranes with different mixed matrix outer layer thicknesses after heat-
treated at 235oC (A and D (enlarged): DL3A; B and E (enlarged): DL3B; C and F (enlarged): DL3C)
Table 7.4 summarizes the gas separation performance of dual-layer PES/P84 hollow
fiber membranes with different mixed matrix outer-layer thicknesses after heat-treated
at 235oC. It can be found that the thinner mixed matrix outer-layer thickness measured
from SEM pictures, the higher gas permeance. All of these heat-treated fibers show
203
the Knudsen diffusion before the two-step coating treatment, which indicates the
defects on the outer surface of DL3A, DL3B and DL3C fibers still exist even after
heat-treatment. This may be due to the fact that more defects probably may be formed
when the mixed matrix outer layer becomes thinner, and thus they cannot be
completely eliminated simply by means of heat-treatment.
Table 7.4: Gas separation performance of dual-layer PES/P84 hollow fiber membranes with different mixed matrix outer layer thicknesses after heat-treated at 235oC
2.081.560.486Permeance (CO2) (GPU)
7.037.117.16Ideal selectivity(O2/N2)
0.05620.04210.0133Permeance (CH4) (GPU)
37.037.136.5Ideal selectivity(CO2/CH4)
DL3C-235oCDL3B-235oCDL3A-235oChollow fiber sample name
1:151:101:6.7Ratio of out layer dope flow rate to inner layer dope flow rate
Before coating
0.920.930.92Ideal selectivity(O2/N2)
1269364112Permeance (N2) (GPU)
1173338103Permeance (O2) (GPU)
2.7 [7.35]
0.0169
0.121
0.77 [3.08]
0.0605
0.430
0.55 [1.83]
0.0861
0.605
0Air gap (cm)
After coating
Permeance (N2) (GPU)
Calculated dense selective layer thickness [measured mixed matrix outer layer thickness
from SEM pictures] (μm)
Permeance (O2) (GPU)
2.081.560.486Permeance (CO2) (GPU)
7.037.117.16Ideal selectivity(O2/N2)
0.05620.04210.0133Permeance (CH4) (GPU)
37.037.136.5Ideal selectivity(CO2/CH4)
DL3C-235oCDL3B-235oCDL3A-235oChollow fiber sample name
1:151:101:6.7Ratio of out layer dope flow rate to inner layer dope flow rate
Before coating
0.920.930.92Ideal selectivity(O2/N2)
1269364112Permeance (N2) (GPU)
1173338103Permeance (O2) (GPU)
2.7 [7.35]
0.0169
0.121
0.77 [3.08]
0.0605
0.430
0.55 [1.83]
0.0861
0.605
0Air gap (cm)
After coating
Permeance (N2) (GPU)
Calculated dense selective layer thickness [measured mixed matrix outer layer thickness
from SEM pictures] (μm)
Permeance (O2) (GPU)
All fibers were tested at 35oC. The testing pressure was 2atm for O2 and N2 before coating; 10atm for O2 and N2, and 8atm for CH4 and CO2 after coating.All fibers were coated with the two-step coating method if they showed Knudsen diffusion before coating.
After the two-step coating, the O2/N2 and CO2/CH4 selectivity of heat-treated DL3A,
DL3B and DL3C fibers increases impressively to around 7.1 and 37, respectively,
which are around 20% and 24% superior to those of neat PES dense films. The
possible reasons for the enhanced selectivity have been mentioned previously; that are
204
polymer chain rigidification and partial pore blockage of zeolite Beta. The effects of
these two factors on the gas separation performance of flat dense MMMs have been
extensively studied [15-18]. To prove the occurrence of polymer chain rigidification,
Figure 7.3 and Table 7.5 show a comparison of Tgs between a heat-treated dual-layer
hollow fiber membrane and a neat PES dense film. The Tg increase of these fibers
provides qualitative confirmation for polymeric chain rigidification. The polymer
chain rigidification usually leads to an increase in gas selectivity.
B
C D
A
Figure 7.3: Comparison of glass transition temperatures of dual-layer hollow fiber
membranes with a mixed matrix outer layer with neat PES dense film (A, B, C and D: the enlarged second heating cycle of DSC curves of neat PES dense film, DL2A-
235oC, DL3A-235oC and DL3C-235oC, respectively)
205
Table 7.5: Change of glass transition temperatures of dual-layer hollow fiber membranes with the mixed matrix outer layer over neat PES dense film
222±1oC
DL3B-235oC
222±1oC
DL3C-235oC
222±1oC
DL3A-235oC
220±1oC
DL2A-235oC
215±1oC
Neat PES dense film
Glass transition temperature (oC)
Hollow fiber sample name
222±1oC
DL3B-235oC
222±1oC
DL3C-235oC
222±1oC
DL3A-235oC
220±1oC
DL2A-235oC
215±1oC
Neat PES dense film
Glass transition temperature (oC)
Hollow fiber sample name
The pore openings in the linear channels of zeolite Beta are approximately 5.7 x 7.5Ǻ
[41]. Therefore, it may be a reasonable assumption that zeolite Beta is probably
selective for the transport of O2 over N2 and CO2 over CH4 when the pore size of
zeolite Beta is reduced due to the attachment of polymer chains on zeolite surface.
This partial blockage assumption has been confirmed by the gas separation
performance of MMMs made of zeolite 5A in our previous work [17, 18], where
zeolite 5A with a pore size of 5Ǻ exhibits good separation properties for O2/N2 and
CO2/CH4 gas pairs due to the partial pore blockage. Therefore, the combination of
polymer chain rigidification and partial pore blockage of zeolite Beta may lead to an
enhancement in the O2/N2 and CO2/CH4 selectivity of dual-layer hollow fiber
membranes with a PES-zeolite Beta mixed matrix outer layer.
A more important achievement shown in Table 7.4 is that the calculated mixed matrix
dense selective layer thickness decreases greatly; i.e. the gas permeance increases
greatly. The calculated mixed matrix dense-selective layer thickness for DL3C-235oC
fibers only is 0.55µm, which is the thinnest one ever reported in the literatures for the
dual-layer hollow fiber membranes with a mixed matrix outer layer. This will
206
significantly enhance their competitive ability with hollow fiber membranes made
from neat polymeric materials.
7.2.3 Effect of Air Gap on Gas Separation Performance
It can be seen from Table 7.6 that the permeance and selectivity of DL2B-235oC fibers
are very similar to those of DL2A-235oC fibers as the air gap increases from 0 to
0.5cm. However, when the air gap increases to be 1.5cm, the gas permeance of DL2C-
235oC fibers greatly increases and the O2/N2 selectivity shows the Knudsen diffusion
before coating, which indicates that there are defects on the outer surface of DL2C-
235oC fibers with a further increase in air gap. Although DL2C-235oC fibers exhibit
the solution-diffusion after the two-step coating, their O2 and CO2 permeance is still
larger than that of DL2A-235oC and DL2B-235oC fibers, and their O2/N2 and
CO2/CH4 selectivity is slightly lower than that of DL2A-235oC and DL2B-235oC
fibers. This may be due to relatively porous skin structure of DL2C-235oC fibers.
The elongation stress in the air gap region can enhance the polymer molecular
orientation which possibly causes the polymer molecules to pack closer to each other
and leads to a tighter skin structure. However, there exists a critical elongation stress
[42]. The entangled polymer chains may be broken with a further increase in
elongation stress and beyond this critical point, it may lead to the formation of defects.
207
The critical stress when spinning dual-layer PES/P84 hollow fiber membranes with a
mixed matrix outer layer occurs when the air gap is about 0.5 cm.
Table 7.6: Effects of air gap on the gas separation performance of dual-layer PES/P84 hollow fiber membranes with a mixed matrix outer layer after heat-treated at 235oC
0.946.606.60Ideal selectivity(O2/N2)
34.10.008550.00867Permeance (N2) (GPU)
32.00.05640.0572Permeance (O2) (GPU)
0.281Permeance (CO2) (GPU)
5.97Ideal selectivity(O2/N2)
0.00944Permeance (CH4) (GPU)
29.8Ideal selectivity(CO2/CH4)
DL2C-235oCDL2B-235oCDL2A-235oChollow fiber sample name
1.50.50Air gap (cm)
Before coating
33.032.6Ideal selectivity(CO2/CH4)
0.006120.00561Permeance (CH4) (GPU)
0.2020.183Permeance (CO2) (GPU)
0.0117
0.0699After coating
Permeance (N2) (GPU)
Permeance (O2) (GPU)
0.946.606.60Ideal selectivity(O2/N2)
34.10.008550.00867Permeance (N2) (GPU)
32.00.05640.0572Permeance (O2) (GPU)
0.281Permeance (CO2) (GPU)
5.97Ideal selectivity(O2/N2)
0.00944Permeance (CH4) (GPU)
29.8Ideal selectivity(CO2/CH4)
DL2C-235oCDL2B-235oCDL2A-235oChollow fiber sample name
1.50.50Air gap (cm)
Before coating
33.032.6Ideal selectivity(CO2/CH4)
0.006120.00561Permeance (CH4) (GPU)
0.2020.183Permeance (CO2) (GPU)
0.0117
0.0699After coating
Permeance (N2) (GPU)
Permeance (O2) (GPU)
All fibers were tested at 35oC. The testing pressure was 10atm for O2 and N2, and 8atm for CH4 and CO2.All fibers were coated with the two-step coating method if they showed Knudsen diffusion before coating.
Our previous work indicated that the critical air gap was about 1.5cm when spinning
dual-layer “neat” PES hollow fiber membranes with the same composition except the
zeolite addition [37]. A comparison of these data, it can be found that the critical air
gap of dual-layer hollow fibers made from organic-inorganic composite materials is
smaller than that of dual-layer hollow fiber made from neat polymeric materials. This
may be due to the fact that the addition of zeolite Beta in the polymer matrix increases
the weight (gravitation force) as well as rigidifies the polymer chains. The latter may
208
weaken the toughness of polymer chains, thus making polymer chains more easily
broken with an increase in elongation stress in the air gap region.
7.2.4 Mixed Gas Separation Performance
There may be some differences between pure gas and mixed gas separation results.
Possible reasons are the competition in sorption among the penetrants, the
plasticization induced by CO2 or hydrocarbon gases, the concentration polarization,
and the non-ideal gas behavior [43-45]. Therefore, the mixed gas measurement is
highly recommended for hollow fiber membranes in order to obtain the true
membrane performance in industrial applications.
Table 7.7: Comparison of pure gas and mixed gas separation performance of dual-layer PES/P84 hollow fiber membranes with a mixed matrix outer layer
Pure gas at 35oCMixed gas (50/50 mol% CO2/CH4) at 24oCMembrane name
CH4CO2CH4CO2
Ideal selectivity (CO2/CH4)
Permeance (GPU)Selectivity (CO2/CH4)
Permeance (GPU)Total feed pressure (psig)
29.20.006410.18733.80.004850.164190DL-235oC
6.50.01140.07416.90.009180.0631200DL-235oC
N2O2
Permeance (GPU)Total feed pressure (psig)
Mixed gas (21 mol%±1% O2 in air) at 24oCMembrane name
N2O2
Ideal selectivity (O2/N2)
Permeance (GPU)Selectivity (O2/N2)
Pure gas at 35oC
Pure gas at 35oCMixed gas (50/50 mol% CO2/CH4) at 24oCMembrane name
CH4CO2CH4CO2
Ideal selectivity (CO2/CH4)
Permeance (GPU)Selectivity (CO2/CH4)
Permeance (GPU)Total feed pressure (psig)
29.20.006410.18733.80.004850.164190DL-235oC
6.50.01140.07416.90.009180.0631200DL-235oC
N2O2
Permeance (GPU)Total feed pressure (psig)
Mixed gas (21 mol%±1% O2 in air) at 24oCMembrane name
N2O2
Ideal selectivity (O2/N2)
Permeance (GPU)Selectivity (O2/N2)
Pure gas at 35oC
Pure O2 and N2 gases were tested at 132psig.
Pure CO2 and CH4 gases were tested at 100psig.
Both purified air containing 21 mol%±1% O2 and CO2/CH4 (50/50 mol%) mixed
gases were selected as separation objects in this work. Table 7.7 summarizes the
209
mixed gas and pure gas separation performance of dual-layer PES/P84 hollow fiber
membranes with a mixed matrix outer layer heat-treated at 235oC. It can be seen that
the difference between mixed gas and pure gas results is almost negligible after
considering the effect of different testing temperatures, which demonstrates the
developed dual-layer hollow fiber membranes with a mixed matrix outer layer in this
study are applicable in industry for the gas separation.
7.3 CONCLUSIONS
The following conclusions can be drawn from this chapter:
(1) The developed dual-layer PES/P84 hollow fiber membranes with a PES-zeolite
Beta mixed matrix outer layer exhibit enhanced O2/N2 and CO2/CH4 selectivity (i.e.
around 10~20%) after heat-treatment and two-step silicon rubber coating compared
with that of neat PES dense films, although the selectivity still is below what might be
expected from MMMs. Polymer chain rigidification and partial pore blockage of
zeolite Beta may be the main causes to enhance the selectivity.
(2) Through adjusting the ratio of outer layer flow rate to inner layer flow rate during
the spinning, the calculated apparent mixed matrix dense-selective layer thickness of
DL3C-235oC fibers is only 0.55µm which is the thinnest one ever reported in the
literatures for the dual-layer hollow fiber membranes with a mixed matrix outer layer.
210
(3) For wet-spun dual-layer hollow fibers of DL2A, the heat-treatment at or above
210oC is enough to completely eliminate the defects on the outer-layer surface of dual-
layer hollow fibers without coating, which may greatly simplify the module
fabrication. However, for dry-jet wet-spun dual-layer hollow fibers with a reasonable
length of air gap, heat-treatment alone may be able to reduce the voids to the range
that can be cured by silicone rubber coating. There exists a critical air gap for dry-jet
wet-spun dual-layer hollow fibers with a mixed matrix outer layer.
(4) The newly developed dual-layer PES/P84 hollow fiber membranes with a PES-
zeolite Beta mixed matrix outer layer show comparable permeance and selectivity of
O2/N2 and CO2/CH4 in both pure and mixed gas tests.
7.4 REFERENCES
[1] R.W. Spillman, Economics of gas separation membranes, Chem. Eng. Prog., 85
(1989) 41.
[2] L.M. Robeson, Correlation of separation factor versus permeability for polymeric
membranes, J. Membr. Sci., 62 (1991) 165.
[3] S. Kulprathipanja, R.W. Neuzil, N.N. Li, Separation of fluids by means of mixed
matrix membranes in gas permeation, US Patent No. 4,740,219 (1988).
211
[4] C.M. Zimmerman, A. Singh, W.J. Koros, Tailoring mixed matrix composite
membranes for gas separation, J. Membr. Sci., 137 (1997) 145.
[5] M. Jia, K.V. Peinemann, R.D. Behling, Molecular sieving effect of zeolite filled
silicone rubber membranes, J. Membr. Sci., 57 (1991) 289.
[6] J. M. Duval, B. Folkers, M.H.V. Mulder, G. Desgrandchamps, C.A. Smolders,
Adsorbent-filled membranes for gas separation. Part 1. Improvement of the gas
separation properties of polymeric membranes by incorporation of microporous
adsorbents, J. Membr. Sci., 80 (1993) 189.
[7] D.R. Paul, D.R. Kemp, The diffusion time lag in polymer membranes containing
adsorptive fillers, J. Polym. Sci. Symp., 41 (1973) 79.
[8] R. Mahajan, W.J. Koros, Factors controlling successful formation of mixed-matrix
gas separation materials, Ind. Eng. Chem. Res., 39 (2000) 2692.
[9] M.G. Süer, N. Baç, L. Yilmaz, Gas permeation characteristics of polymer-zeolite
mixed matrix membranes, J. Membr. Sci., 91 (1994) 77.
[10] Ş.B. Tantekin-Ersolmaz, L. Şenorkyan, N. Kalaonra, M. Tatlıer, A. Erdem-
Şenatalar, n-Pentane/i-pentane separation by using zeolite-PDMS mixed matrix
membranes, J. Membr. Sci., 189 (2001) 59.
[11] R. Mahajan, R. Burns, M. Schaeffer, W.J. Koros, Challenges in forming
successful mixed matrix membranes with rigid polymeric materials, J. App. Polym.
Sci., 86 (2002) 881.
212
[12] J.M. Duval, A.J.B. Kemperman, B. Folkers, M.H.V. Mulder, G. Desgrandchamps,
C.A. Smolders, Preparation of zeolite filled glassy polymer membrane, J. Appl.
Polym. Sci., 54 (1994) 409.
[13] Y.L. Liu, Y.H. Su, K.R. Lee, J.Y. Lai, Crosslinked organic-inorganic hybrid
chitosan membranes for pervaporation dehydration of isopropanol-water mixtures
with a long-term stability, J. Membr. Sci., 251 (2005) 233.
[14] H.H. Yong, H.C. Park, Y.S. Kang, J. Won, W.N. Kim, Zeolite-filled polyimide
membrane containing 2,4,6-triaminopyrimidine, J. Membr. Sci., 188 (2001) 151.
[15] R. Mahajan, Formation, characterization and modeling of mixed matrix
membrane materials, Ph.D. Dissertation, The University of Texas at Austin, 2000.
[16] T.T. Moore, R. Mahajan, De Q. Vu, W.J. Koros, Hybrid membrane materials
comprising organic polymers with rigid dispersed phases, AICHE Journal, 50 (2)
(2004) 311.
[17] Y. Li, T.S. Chung, C. Cao, S. Kulprathipanja, The effects of polymer chain
rigidification, zeolite pore size and pore blockage on polyethersulfone (PES)-
zeolite A mixed matrix membranes, J. Membr. Sci., 260 (2005) 45.
[18] Y. Li, H.M. Guan, T.S. Chung, S. Kulprathipanja, Effects of novel silane
modification of zeolite surface on polymer chain rigidification and partial pore
blockage in polyethersulfone (PES)-zeolite A mixed matrix membranes, J. Membr.
Sci., in press.
213
[19] W.J. Koros, I. Pinnau, Membrane formation for gas separation process, in: D.R.
Paul, Y.P. Yampol’skii (Eds.), Polymeric Gas Separation Membranes, CRC Press,
Boca Raton, 1994, Chapter 5.
[20] R.E. Kesting, A.K. Fritzsche, Polymeric Gas Separation Membranes, Wiley, New
York, NY, 1993.
[21] I.M. Wienk, R.M. Boom, M.A.M. Beerlage, A.M.W. Bulte, C.A. Smolders, H.
Strathmann, Recent advances in the formation of phase inversion membranes
made from amorphous or semi-crystalline polymers, J. Membr. Sci., 113 (1996)
361.
[22] T.S. Chung, A review of microporous composite polymeric membrane
technology for air-separation, Polym. Polym. Comp., 4 (1996) 269.
[23] K.Y. Wang, T. Matsuura, T.S. Chung, W.F. Guo, The effects of flow angle and
shear rate within the spinneret on the separation performance of poly (ethersulfone)
(PES) ultrafiltration hollow fiber membranes, J. Membr. Sci., 240 (2004) 67.
[24] W.S.W. Ho, K.K. Sirkar, Membrane Handbook, Ed., Van Nostrand Reinhold,
New York, 1992.
[25] R. Kiyono, G.H. Koops, M. Wessling, H. Strathmann, Mixed matrix microporous
hollow fibers with ion-exchange functionality, J. Membr. Sci., 231 (2004) 109.
[26] O.M. Ekiner, S.S. Kulkarni, Process for making hollow fiber mixed matrix
membrane, US Patent No. 6,663,805 (2003).
214
[27] W.J. Koros, D. Wallace, J.D. Wind, S.J. Miller, C.S. Bickel, D.Q. Vu,
Crosslinked and crosslinkable hollow fiber mixed matrix membrane and method of
making same, US Patent Application No. 20030140789 (2003).
[28] S.J. Miller, C.L. Munson, S.S. Kulkarni, D.J. Hasse, Purification of p-xylene
using composite mixed matrix membranes, US Patent No. 6,500,233 (2002).
[29] L.Y. Jiang, T.S. Chung, C. Cao, Z. Huang, S. Kulprathipanja, Fundamental
understanding of nano-sized zeolite distribution in the formation of the mixed
matrix single- and dual-layer asymmetric hollow fiber membranes, J. Membr. Sci.,
252 (2005) 89.
[30] L.Y. Jiang, T.S. Chung, S. Kulprathipanja, An investigation to revitalize the
separation performance of hollow fibers with a thin mixed matrix composite skin
for gas separation, J. Membr. Sci., in press.
[31] E. Kuzumoto, K. Nitta, Manufacture of permselective hollow membranes,
Japanese Patent 01,015,104 (1989).
[32] T. Yoshinaga, H. Shimazaki, Y. Kusuki, Polyimide bilayer hollow membranes
and their manufacture, Japanese Patent 02,169,019 (1990).
[33] D.F. Li, T.S. Chung, R. Wang, Y. Liu, Fabrication of fluoropolyimide/
polyethersulfone (PES) dual-layer asymmetric hollow fiber membranes for gas
separation, J. Membr. Sci., 198 (2002) 211.
[34] Y. Liu, T.S. Chung, D.F. Li, R. Wang, High-selectivity, anti-plasticization cross-
linked polyimide/polyethersulfone dual-layer hollow fiber membranes, Ind. Eng.
Chem. Res., 42 (2003) 1190.
215
[35] L.Y. Jiang, T.S. Chung, D.F. Li, C. Cao, S. Kulprathipanja, Fabrication of
Matrimid/Polyethersulfone dual-layer hollow fiber membranes for gas separation,
J. Membr. Sci., 240 (2004) 91.
[36] D.F. Li, T.S. Chung, R. Wang, Morphological aspects and structure control of
dual-layer asymmetric hollow fiber membranes formed by a simultaneous co-
extrusion approach, J. Membr. Sci., 243 (2004) 155.
[37] Y. Li, C. Cao, T.S. Chung, K.P. Pramoda, Fabrication of dual-layer
polyethersulfone (PES) hollow fiber membranes with an ultrathin dense selective
layer for gas separation, J. Membr. Sci., 245 (2004) 53.
[38] J.N. Barsema, G.C. Kapantaidakis, N.F.A. van der Vegt, G.H. Koops, M.
Wessling, Preparation and characterization of highly selective dense and hollow
fiber asymmetric membranes based on BTDA-TDI/MDI co-polyimide, J. Membr.
Sci., 216 (2003) 195.
[39] B. Bikson, J.K. Nelson, Composite membranes and their manufacture and use,
US Patent 4,826,599 (1989).
[40] A.Z. Gollan, Anisotropic membranes for gas separation, US Patent 4,681,605
(1987).
[41] J.B. Higgins, R.B. LaPierre, J.L. Schlenker, A.C. Rohrman, J.D. Wood, G.T. Kerr,
W.J. Rohrbaugh, The framework topology of zeolite Beta, Zeolites, 8 (1988) 446.
[42] T.S. Chung, The limitations of using Flory-Huggins equation for the states of
solutions during asymmetric hollow fiber formation, J. Membr. Sci., 126 (1997)
19.
216
[43] P.C. Raymond, W.J. Koros, D.R. Paul, Comparison of mixed and pure gas
permeation characteristics for CO2 and CH4 in copolymers and blends containing
methyl methacrylate units, J. Membr. Sci., 77 (1993) 49.
[44] J.H. Kim, S.Y. Ha, Y.M. Lee, Gas permeation of poly(amide-6-b-ethylene oxide)
copolymer, J. Membr. Sci., 190 (2001) 179.
[45] M.R. Coleman, W.J. Koros, Conditioning of fluorine-containing polyimides. 2.
Effect of conditioning protocol at 8% volume dilation on gas-transport properties,
Macromolecules, 32 (1999) 3106.
217
CHAPTER EIGHT
CONCLUSIONS AND RECOMMENDATIONS
8.1 CONCLUSIONS
In recent years, current available membrane materials have reached Robeson’s upper
bound trade-off limit. Therefore, both the development of membrane materials with
superior separation properties for gas separation and the innovation of membrane
fabrication technology are research-intensive, aimed to produce high performance
membranes that can accomplish/conquer the separation challenges. Accordingly, an
investigation of membrane materials and fabrications based on commercially available
PES for gas separation (He/N2, H2/N2, O2/N2 and CO2/CH4) was carried out in this
research study. It consisted of four aspects:
1. Fabrication of dual-layer PES hollow fiber membranes with an ultrathin neat
polymeric dense-selective layer for gas separation.
2. Investigation of the effects of polymer chain rigidification, zeolite pore size
and pore blockage on flat PES-zeolite A mixed matrix membranes.
3. Study of the effects of novel silane modification of zeolite surface on polymer
chain rigidification and partial pore blockage in flat PES-zeolite A mixed
matrix membranes.
218
4. Fabrication and characterization of dual-layer polyethersulfone (PES)/BTDA-
TDI/MDI co-polyimide (P84) hollow fiber membranes with a submicron PES-
zeolite Beta mixed matrix dense-selective layer for gas separation.
The above three explored membranes in this study, namely dual-layer hollow fiber
neat polymeric membranes, flat dense mixed matrix membranes and dual-layer hollow
fiber mixed matrix membranes exhibited the excellent gas separation capability. The
following conclusions were derived and summarized as a result of physical
characterization and gas permeation measurement of these materials.
8.1.1 Fabrication of Dual-layer Polyethersulfone (PES) Hollow Fiber Membranes
with an Ultrathin Dense Selective Layer for Gas Separation
By using co-extrusion and dry-jet wet-spinning phase inversion techniques with the
aid of heat treatment at 75oC, we have fabricated dual-layer polyethersulfone (PES)
hollow fiber membranes with an ultrathin dense-selective layer of 407Å. To our best
knowledge, this is the thinnest thickness that has ever been reported for dual-layer
hollow fiber membranes. The newly developed dual-layer hollow fibers had an O2
permeance of 10.8 GPU and O2/N2 selectivity of 6.0 at 25oC after heat-treated at 75oC.
It was also found that the heat-treatment at 75oC improved the gas permeance and
ideal selectivity, while the heat-treatment at 150oC resulted in a significant reduction
in both permeance and selectivity due to the enhanced substructure resistance.
219
The effects of different post heat-treatments on the membrane morphology were also
studied. SEM pictures confirmed that a higher heat-treatment temperature can
significantly reduce pore sizes and the amount of pores in the substructure
immediately underneath the dense-selective layer. SEM pictures also showed that
macrovoids formed just beneath the outer skin of inner layer had a long and narrow
finger-like structure, while macrovoids near the inner cavity of the fiber had a short
and wide tear drop-like structure. The structural difference between these two types of
macrovoids might be resulted from different coagulation rates and solvent exchange
rates.
8.1.2 The Effects of Polymer Chain Rigidification, Zeolite Pore Size and Pore
Blockage on Polyethersulfone (PES)-Zeolite A Mixed Matrix Membranes
The polyethersulfone (PES)-zeolite 3A, 4A and 5A mixed matrix membranes (MMMs)
were fabricated with a modified solution-casting procedure at high temperatures close
to the glass transition temperatures (Tg) of polymer materials. The effects of
membrane preparation methodology, zeolite loading and pore size of zeolite on the gas
separation performance of these mixed matrix membranes were studied in this part.
SEM results showed the interface between polymer and zeolite in MMMs
experiencing natural cooling was better (i.e., less defective) than that in MMMs
experiencing immediate quenching. The increment of glass transition temperature (Tg)
220
of MMMs with zeolite loading confirmed the polymer chain rigidification induced by
zeolite.
The experimental results indicated that a higher zeolite loading resulted in a decrease
in gas permeability and an increase in gas pair selectivity. The unmodified Maxwell
model failed to correctly predict the permeability decrease induced by polymer chain
rigidification near the zeolite surface and the partial pore blockage of zeolites by the
polymer chains. A new modified Maxwell model was therefore proposed. It took the
combined effects of chain rigidification and partial pore blockage of zeolites into
calculation. The new model showed much consistent permeability and selectivity
predication with experimental data.
Surprisingly, an increase in zeolite pore size from 3Å to 5Å generally not only
increased gas permeability, but also gas pair selectivity. The O2/N2 selectivity of PES-
zeolite 3A and PES-zeolite 4A membranes was very similar, while the O2/N2
selectivity of PES-zeolite 5A membranes was much higher. This implied that the
blockage might narrow a part of zeolite 5A pores to approximately 4Å which can
discriminate the gas pair of O2 and N2, and narrow a part of zeolites 3A and 4A pores
to smaller sizes. It was concluded that the partial pore blockage of zeolites by the
polymer chains has equivalent or more influence on the separation properties of mixed
matrix membranes compared with that of the polymer chain rigidification.
221
8.1.3 Effects of Novel Silane Modification of Zeolite Surface on Polymer Chain
Rigidification and Partial Pore Blockage in Polyethersulfone (PES)-Zeolite
A Mixed Matrix Membranes
A novel silane coupling agent, (3-aminopropyl)-diethoxymethyl silane (APDEMS)
was used in this part to modify zeolite surface for mixed matrix membranes (MMMs).
Both elementary analysis and XPS spectra confirmed the chemical modification, while
BET measurements showed no changes in zeolite surface area and total pore volume
after the modification. Polyethersulfone (PES)-zeolite 3A, 4A and 5A MMMs were
fabricated at high processing temperatures using unmodified and chemical modified
zeolite. SEM images of these MMMs indicated that the interface between polymer and
zeolite phases becomes better if modified zeolite is used.
The effects of chemical modification of zeolite surface and zeolite loadings on the gas
separation performance of these MMMs were investigated. Both permeability and
selectivity of MMMs made from APDEMS modified zeolite were higher than those of
MMMs made from unmodified zeolite at 20 wt% zeolite loading because of a decrease
in the degree of partial pore blockage of zeolites. The permeability of all studied gases
decreased with increasing zeolite content for PES-zeolite 4A-NH2 MMMs, while for
PES-zeolite 5A-NH2 MMMs, the gas permeability decreased and then increased with
an increase in zeolite loadings. This unique phenomenon implied that using large pore-
222
size zeolite for MMMs would potentially offset the negative effects of partial pore
blockage and polymer chain rigidification on permeability.
A modified Maxwell model with adjusted parameters was applied to study the PES-
zeolite 4A-NH2 MMM system. The predicted permeability and selectivity showed
very good agreement with experimental data, indicating that the modified Maxwell
model is fully capable of predicting the gas separation performance of MMMs made
from both unmodified and modified zeolite.
8.1.4 Dual-layer Polyethersulfone (PES)/BTDA-TDI/MDI Co-polyimide (P84)
Hollow Fiber Membranes with a Submicron PES-Zeolite Beta Mixed
Matrix Dense-Selective Layer for Gas Separation
Dual-layer PES/P84 hollow fiber membranes with a PES-zeolite Beta mixed matrix
dense-selective layer of 0.55µm have been successfully fabricated in this study by
adjusting the ratio of outer layer flow rate to inner layer flow rate during the spinning
with the aid of heat-treatment at 235oC and two-step coating. To our best knowledge,
this is the thinnest thickness that has ever been reported for dual-layer hollow fiber
membranes with the mixed matrix outer layer. Scanning electron microscope (SEM)
images demonstrated that heat-treatment could effectively reduce or eliminate the
voids induced by the unfavorable interaction between polymer and zeolite phases.
Differential scanning calorimetry (DSC) results proved that heat-treated dual-layer
223
hollow fiber membranes with a mixed matrix outer layer possessed a higher glass
transition temperature (Tg) over neat PES dense film.
.
The newly developed dual-layer hollow fibers exhibited an enhanced O2/N2 and
CO2/CH4 selectivity of around 10~20% compared with that of neat PES dense films.
Both the polymer chain rigidification and partial pore blockage of zeolite Beta may be
the main causes for the selectivity increment. The outer surface of wet-spun hollow
fibers showed fewer defects than that of hollow fibers spun at 1.5cm air gap probably
because of the overlarge elongation stress with an increase in air gap. The
performance of these newly developed dual-layer hollow fiber membranes has also
been confirmed in mixed gas tests, and showed comparable permeance and selectivity
of O2/N2 and CO2/CH4 in both pure and mixed gas tests.
8.2 RECOMMENDATIONS FOR FUTURE WORK
Based on the experimental results obtained, discussions presented and conclusions
drawn from this research study, the following recommendations given may provide
further insight for future investigations related to the development of membrane
materials with potentially high-separation properties and the innovation of membrane
fabrication technology.
8.2.1 Flat Dense Polymer-Zeolite Mixed Matrix Membranes
224
The flat dense MMMs are still very useful to help us understand the mechanism about
the separation performance variation of MMMs as a function of zeolite types and
zeolite loadings, since some important physical properties, such as the dense-selective
layer thickness and the zeolite loading, can be obtained accurately. In the previous
study in this work and by other researchers, many factors were applied to explain the
change in the separation performance of the continuous phase after the addition of
zeolites. Yet, except for few works, most of these factors have not been completely
and accurately confirmed. One aspect of interest is the effect of particle size of zeolites
on the membrane performance. Only three types of zeolites, 3A, 4A and 5A, were
applied in this research study and may not be accurate enough to deduce a general
conclusion suitable for various zeolites. Therefore, in the future studies, more zeolites,
such as Silicalite-1 and 13X, need to be applied in the fabrication and characterization
of MMMs in order to obtain more detailed and accurate gas transport mechanism in
polymer-zeolite MMMs.
Moreover, in the future, more characterization methods should be conducted to prove
our assumptions in this work, such as the polymer chain rigidification and partial pore
blockage of zeolites. The dynamic mechanical analysis may be used to indicate the
modification in the chain rigidity near the zeolite surface through the peak broadening
and shifting of the tan delta.
225
The modified Maxwell model with some adjustable parameters was applied to predict
the gas separation performance of flat dense polymer-zeolite MMMs and prediction
results showed very good agreement with experimental data. Regrettably, only O2 and
N2 permeability of PES-zeolite 4A MMMs was predicted using the modified Maxwell
model in this work and may not provide enough convincing information to
quantitatively estimate the combined effects of polymer chain rigidification and partial
pore blockage of zeolites on resultant gas separation performance of MMMs. This is
mainly due to the fact that so far only O2 and N2 permeability data of zeolite 4A can
be found in the literature and these data are relatively consistent, while there are no
consistent CO2 and CH4 permeability data of zeolite 4A available. For other zeolites,
such as 5A and Silicalite-1, no consistent data on intrinsic gas permeability can be
found in the literature. The modified Maxwell model developed in this work could be
applied in the prediction of other MMMs after uniform data on the gas permeability,
diffusion and sorption coefficients of neat zeolite and polymer membranes are
obtained by means of the cooperation with other research groups. Besides these
intrinsic properties of used organic and inorganic materials, the flexible selection of
model parameters based on the nature of used materials and chemical modification
method etc. also is essential to perfect the modified Maxwell model in the prediction
of the resultant gas separation performance of MMMs.
8.2.2 Dual-layer Hollow Fibers with a Mixed Matrix Outer Layer
226
The hollow fiber membranes are more advantageous for industry applications due to
their high surface area per unit volume of membrane module. Therefore, it is
imperative to convert the advanced mixed matrix materials developed in this work into
the hollow fiber configuration. Two crucial problems concerning the hollow fibers in
this work need to be solved before the further investigation. Firstly, these dual-layer
hollow fiber mixed matrix membranes still exhibited quite low permeance due to the
relatively thicker dense-selective layer, although the thinnest mixed matrix dense-
selective layer has been successfully fabricated in this work. Secondly, the voids
between the polymer and zeolite phases during the phase inversion in hollow fiber
spinning entails more complicated post-treatment/repair procedures compared to the
flat dense membranes by casting-evaporation. The major scopes that may be helpful
in overcoming these challenges involve:
1) To choose the suitable pair of polymer-particle and silane coupling agents in order
to reduce the void dimension between the polymer and zeolite phases induced by
the incompatibility between them;
2) The micro-sized zeolite particles are too large and inappropriate for the fabrication
of defect-free asymmetric hollow fibers with an ultrathin (e.g., 100nm) dense-
selective layer. Thus, the synthesis of nano-sized zeolite, preferably, <100nm is
required to induce the high dispersion in the polymer solution;
3) To develop an optimized dual-layer hollow fiber spinning technology (e.g.
spinneret design, spinning conditions) to get an ultrathin dense-selective layer with
zeolite particles.
227
After overcoming these problems, the zeolite loading in dual-layer hollow fiber mixed
matrix membranes should be increased gradually and steadily to obtain more advanced
separation performance in the future studies, which needs a more skillful hollow fiber
spinning technique.
8.2.3 Other gas separation applications of MMMs
The separation of olefins and paraffins is of primary importance in the chemical
process industry due to the high demand for polymers and other specialty chemicals
derived from mono-olefins. The membrane-based separation is attractively adopted for
gas separation due to its low capital outlay and high energy efficient; however, the
permeability and selectivity achieved by polymeric membrane materials is very low
and has restricted the membrane application in hydrocarbon separation. The mixed
matrix membrane materials combine the easy processability of polymeric materials
with the superior gas separation properties of rigid molecular sieve materials; therefore,
this type of composite material may be helpful to enhance the separation performance
of olefin/paraffin mixture. Moreover, a combination of dual-layer hollow fibers with
nano-sized zeolite will be more favorable to investigate the transport mechanism of
hydrocarbon gases in mixed matrix composite materials due to the much higher
surface area per unit volume and much thinner dense-selective layer of hollow fibers.
228
APPENDIX A
SYNTHESIS AND CHARACTERIZATION OF ZEOLITE BETA (CHAPTER 3)
A.1 SYNTHESIS OF ZEOLITE BETA
A.1.1 Synthesis of Zeolite Beta Particles
Zeolite Beta can be synthesized using many different ways, strongly dependent on the
silica sources used. The method used here for the zeolite Beta synthesis was similar to
that reported by Borade and Clearfield [1]. In the synthesis in our lab, the batch
composition was 1.5Na2O: 1.0Al2O3: 10TEAOH: 30SiO2: 245H2O, which was slightly
different from that of [1]. Based on the above formula, 1.546g sodium aluminate
(Na2O, 33.2 wt%, Na2O/Al2O3=1.5) was measured and put into a 250ml beaker; the
sodium aluminate was dissolved in 8ml DI water under constant stirring. The clear salt
solution thus formed was then added with 19ml tetraethyl ammonium hydroxide
(TEAOH, 40 wt% aqueous solution, 1.06g/ml) and 5ml of DI water balance while
stirring. 15 minutes later, 10.0 g fumed silica was added into the resultant solution
under stirring; the mixture initially appeared to be a dry powder but turned into a very
dense and homogeneous gel after about 3-hour stirring. The viscous gel was
transferred into a 200ml Teflon container that was set in a stainless steel autoclave
sealed tightly. The crystallization of zeolite Beta was carried out at 170oC under an
autogenous pressure for 24-48 hours. The autoclave was directly quenched with tap
229
water (cooling water is better to obtain the zeolite with a smaller particle size) after
taken out from the oven. The resultant product was thoroughly washed with DI water
till a pH value of 8-9 and then was centrifuged at 10000circles/minute for 5 minutes to
obtain powders. The final product was dried at 70oC for 12 hours. A simple flowchart
about the synthesis of zeolite Beta particles is shown in Figure A.1.
Sodium aluminate
& DI H2O
TEAOH(40wt%)
Clear?
Fumed silica-380
Aging at room temperature
Crystallization
Reaction gel
autoclave
Oven
Teflon container
Reaction at 170oC for 24 ~ 48 hrs
Quenching the reactor in cold water and washing with DI water till a pH value of 8-9Powders with a
centrifuge at 10000 rpm for 5 minutes
Figure A.1: Synthesis procedure of zeolite-Beta
A.1.2 Template Removal from Freshly Prepared Zeolites
Zeolite materials are generally synthesized by using self-assembling organic
surfactants that serve as the template. After synthesis, the template must be removed to
form the sieve pore prior to any potential applications. Conventionally, the organic
230
template is removed by high temperature calcinations; however, with such method, the
organic template is directly burned off and the final particles might form agglomerates.
In this sense, the synthesized zeolites are large zeolite crystals which are not suitable
to prepare homogeneous zeolite suspension to develop practical composite membranes
with polymer matrix.
Recently, Wang et al. [2] proposed a polymer-networking method as a supplement to
the conventional procedure for template removal; by applying a certain polymer-
network in the particle suspension/around the particles, the zeolite aggregations can be
avoided. Therefore, this novel method was adopted in our method with some trivial
modifications in order to prevent the nanocrystal aggregation. Firstly, the washed
zeolites were dispersed into DI water by using an oscillator. The monomer
(acrylamide), crosslinker (N, N’-methylenebisacrylamide) and initiator ((NH4)2S2O8)
were added into the zeolite particle suspension under agitation. Typically, 1.0 g
monomer, 0.1mg crosslinker and 25mg initiator were added into a 10-g zeolite water
suspension which contains around 5 wt. % zeolite. Then, the resultant mixture was
stirred for 30 minutes before heated to 70oC for polymerization. The crosslinked
elastic hydrogel was dried in air at 90oC overnight. Thereafter, the obtained solid
polymer-zeolite composite was directly calcined at 550oC for 10 hours at a scanning
rate of 5oC/min to remove the template and polymer in air. The approach proposed by
Wang et al. [2] was first doing carbonization at 500oC under N2 followed by the
231
calcination under O2. A simple flowchart about the polymer-networking procedure for
the template removal of zeolite Beta is illustrated in Figure A.2.
Calcination at 550oC for 10 hrs in air
Individual particles
Polymer- networking @70oCDried at 80~100oC for 24 hrs
Nano-sized zeoliteDI water
Colloidal suspension (5wt% zeolite)
Addition of monomer (acrylamide)
Addition of cross-linker (N,N’-methylene-bis-acrylamide) and initiator (ammonium persulfate )
Stirring for 10min
Stirring for 20min
After polymerization
Stirring
Figure A.2: Polymer-networking procedure for the template removal of zeolite Beta
A.2 CHARACTERIZATION OF ZEOLITE BETA
Figure A.3 shows the XRD pattern of zeolite Beta, which confirms that the zeolite
self-synthesized in our group is zeolite Beta due to the same crystalline structure as
reference [1]. Figure A.4 displays the FESEM picture of self-synthesized zeolite Beta
and Figure A.5 indicates the particle size determined by the dynamic light scattering
(DLS) system. The latter two characterization methods demonstrate that the particle
size of self-synthesized zeolite Beta in our group is slightly smaller than 300nm.
232
0
2000
4000
6000
5 15 25 35 45 552 Theta(o)
Intensity
Reference (Borade and Clearfield, 1996)
Experimental in my research study
Figure A.3: XRD characterization of the self-synthesized zeolite Beta for the particle crystalline structure confirmation
257.7nm
206.2nm299.0nm
195.9nm206.2nm
195.9nm
268.0nm
206.2nm
Figure A.4: FESEM characterization of the self-synthesized zeolite Beta for the particle shape observation
233
Figure A.5: Laser light scattering (LLS) characterization of the self-synthesized zeolite Beta for the particle size determination
A.3 REFERENCES
[1] R.B. Borade, A. Clearfield, Preparaion of aluminum-rich Beta zeolite,
Microporous Mater., 5 (1996) 289.
[2] H. Wang, B.A. Holmberg, Y. Yan, Homogeneous polymer-zeolite nanocomposite
membranes by incorporating dispersible template-removed zeolite nanocrystals, J.
Mater. Chem., 12 (2002) 3640.
234
APPENDIX B
A NEW TESTING SYSTEM TO DETERMINE THE O2/N2 MIXED GAS
PERMEATION THROUGH HOLLOW FIBER MEMBRANES WITH AN
OXYGEN ANALYZER (CHAPTER 3)
B.1 INTRODUCTION
Pure gas permeation measurements have often been used to estimate the intrinsic
separation properties of polymeric membranes for mixed gas pairs. However, some
differences between pure gas and mixed gas permeation results may exist. Possible
causes are 1) the competition in sorption among the penetrants, 2) the plasticization
induced by CO2 and hydrocarbon gases, 3) the concentration polarization, and 4) the
non-ideal gas behavior [1-8]. Therefore, the conduction of mixed gas measurements is
highly recommended for polymeric membranes in order to obtain the true membrane
performance in industrial applications.
During the last two decades, the techniques based on gas chromatography (GC) to
measure the mixed gas permeation through polymer membranes have been well
developed. At first, researchers determined the actual permeability of each species in a
gas mixture for flat membranes by using a sweep gas at the downstream side of the
membrane to carry the permeate gas to a GC system for composition analysis [9, 10].
235
However, the undesirable back-diffusion of sweep gas across the membrane to the
feed side might not be negligible. O'Brien et al. [11] devised a modified manometric
(i.e., constant volume) permeation cells for the measurement of multicomponent gas
transport through polymer membranes. In combination with GC, their approach
significantly simplified the measurement process and could directly estimate the true
permeability and selectivity over a wide range of mixed gas feed pressures and
compositions. However, their technique was designed for flat membranes.
With the advance in membrane fabrication technology, asymmetric hollow fiber
membranes have become a favorable configuration in the membrane-based systems
for gas separation [12-15]. To investigate their separation performance for mixed
gases, Jones and Koros may be the pioneers on mixed gas tests for hollow fibers [16],
where they revised a flat permeation cell for hollow fiber carbon membranes with a
modified manometric permeation cell. Other approaches have also been proposed
using bubble flow meters to determine the retentate and permeate flow rates of a
hollow fiber module and then back calculated the individual permeance depending on
the flow pattern [17-19]. However, no matter which type of mixed gases was separated
and which method was applied to test the gas flow rate, the compositions of the
permeate, retentate and feed gases were all determined by a GC system in these
reports. A modern GC system is fairly expensive. In addition, it is not trivial to
identify proper operation conditions for GC and to accurately calculate the individual
fluxes when overlap peaks occur or the permeance is very small. This is especially
236
true for O2/N2 mixed gas separation. Based on our experience, it is more difficult to
measure the separation performance of O2/N2 than that of CO2/CH4 mixed gas when
using the same GC-based system due to the partial overlapping of peak areas of O2
and N2.
Therefore, the purpose of this short note is to introduce a relatively economical and
easier testing system for the characterization of hollow fiber membranes for mixed O2
and N2 separation based on an oxygen analyzer. The detailed design, testing procedure
and data validation will be elucidated.
B.2 SYSTEM DESIGN AND MEASUREMENT PROCEDURE
A schematic diagram of the apparatus using an oxygen analyzer for O2/N2 mixed gas
permeation tests through a hollow fiber module was shown in Figure B.1. Swagelok®
316 and 316L stainless steel components such as tubing, union tees and connectors
were used in the construction of testing system. A hollow fiber module was installed
in a temperature controlled chamber and tested in the shell-side feed method. This
module was made by assembling several pieces of fibers into a bundle. The open end
of the bundle was sealed with a 5min rapid epoxy resin (Araldite®, Switzerland), while
the other end was glued onto the Swagelok VCR® union tee using a regular epoxy
resin (Eposet®). Eight hours were required to fully cure the Eposet® resin. Two hollow
237
fiber modules made of different polymers were tested. Table B.1 summarizes their
specifications and experimental conditions.
Purified air containing 21 mol%±1% O2 was feed to the shell side of hollow fiber
membranes from a compressed gas cylinder shown in Figure B.1. A very small part of
the feed gas went through the wall of hollow fiber membranes and entered the
permeate side, while most of the feed gas entered the retentate side from the mid-arm
of union tee. The feed gas pressure was controlled by the needle valve 1 and displayed
by a digital pressure gauge. For O2/N2 mixed gas permeation tests, the feed gas
pressure was approximately 200psig, the permeate side was open to atmosphere, and
the system is at ambient temperature. However, it can be also conducted at different
temperatures and pressures.
238
A module consisting of hollow fibers
Feed
8-hour epoxy
Inner diameter: 10.2mm
73.7mm
VCR Union TeeRetentate
Permeate5-minute epoxy
Mass flow controller (1)
Bubble flow meterBubble flow meter
Mass flow controller (2)Mass flow controller (2) O2 analyzerO2 analyzer inletoutlet
Gas reservoirGas reservoir
Gas cylinder (air)
Gas cylinder (air)
Needle valve 1
Pressure gauge
Needle valve 3 Needle valve 2
inletoutlet
inletoutlet
Vacuum pump
Needle valve 4
Needle valve 5
Atmosphere
Figure B.1: Schematic diagram of apparatus using an oxygen analyzer for O2/N2 mixed gas permeation tests through a hollow fiber module
239
Table B.1: Module specifications and experimental conditions (24°C)
21 mol%±1% O2 in air or CO2/CH4 (50/50 mol%)
21 mol%±1% O2 in air or CO2/CH4 (50/50 mol%)
Compositions of mixed gas
0.57 (for air separation)2.70 (for CO2/CH4)
0.50 (for air separation)1.38 (for CO2/CH4)
Retentate flow rate (ml/min)
Hollow fiber 2Hollow fiber 1Membrane name
550520Inner diameter of fibers (μm)
4.55.5Length of fibers (cm)
1000950Outer diameter of fibers (μm)
200 (for air separation) or 200 (for CO2/CH4)
200 (for air separation) or 190 (for CO2/CH4)
Feed gas pressure (psig)
411Number of fibers
0.0283 (for air separation)0.135 (for CO2/CH4)
0.0219 (for air separation) 0.0819 (for CO2/CH4)
Permeate flow rate (ml/min)
~ 0~ 0Permeate gas pressure (psig)
Composite membranes based on polysulfone
Composite membranes based on polyethersulfone
Membrane material
21 mol%±1% O2 in air or CO2/CH4 (50/50 mol%)
21 mol%±1% O2 in air or CO2/CH4 (50/50 mol%)
Compositions of mixed gas
0.57 (for air separation)2.70 (for CO2/CH4)
0.50 (for air separation)1.38 (for CO2/CH4)
Retentate flow rate (ml/min)
Hollow fiber 2Hollow fiber 1Membrane name
550520Inner diameter of fibers (μm)
4.55.5Length of fibers (cm)
1000950Outer diameter of fibers (μm)
200 (for air separation) or 200 (for CO2/CH4)
200 (for air separation) or 190 (for CO2/CH4)
Feed gas pressure (psig)
411Number of fibers
0.0283 (for air separation)0.135 (for CO2/CH4)
0.0219 (for air separation) 0.0819 (for CO2/CH4)
Permeate flow rate (ml/min)
~ 0~ 0Permeate gas pressure (psig)
Composite membranes based on polysulfone
Composite membranes based on polyethersulfone
Membrane material
240
In order to obtain accurate mixed gas permeation results, the testing system comprises
the following four steps. The first step is to decide the retentate and permeate flow
rates by a mass flow controller and bubble flow meter, respectively. The model of
mass flow controller used in this work is GFC 17 (Aalborg®), which costs S$1,740
(about US$1,000) including calibration. Its measurement range is 0-10ml/min (under
standard temperature and pressure). The accuracy of mass flow controller is around
±1.5% of full scale. For different gases, the actual flow rate is the product of the set
point of the mass flow controller times an adjustable coefficient because this
equipment was calibrated with N2 as received. In our case, the adjustable coefficient is
approximately equal to 1 for air because the adjustable coefficient is 1 for N2 and
0.996 for O2.
The following briefly describes our procedure to control the stage cut approximately at
or below 5% to eliminate the effect of concentration polarization [11, 18]. Firstly, the
set point of mass flow controller (1) (see Figure B.1) was randomly set at a certain
value, for example 2ml/min. Secondly, the permeate flow rate was measured by the
bubble flow meter shown in Figure B.1. If the ratio of permeate flow rate to retentate
flow rate is around 0.05 or lower (i.e. the stage cut was at or below 0.05), the first step
(i.e. measurement of flow rates) was finished. Otherwise, the set point of mass flow
controller (1) would be readjusted to a new value until this requirement was achieved.
In this part of work, the ultimate retentate and permeate flow rates in the mixed O2 and
241
N2 separation were 0.50 and 0.0219 ml/min for hollow fiber 1, respectively; and were
0.57 and 0.0283 ml/min for hollow fiber 2, respectively, as shown in Table B.1.
The second step is to measure the composition of retentate gas using the oxygen
analyzer. An Advanced Micro Instrument (AMI) oxygen analyzer (model 201) was
applied in this study. The analyzer costs S$5500 (about US$3,300), inclusive of the
installation fee. Its measurement range is 0-100% O2 concentration. This O2 analyzer
is calibrated on air and its reading is adjusted to 20.9 in the calibration at moderate
temperature and humidity. Both sensitivity and accuracy of AMI model 201 oxygen
analyzer used in this part of work are 0.5% of full scale. Therefore, the measurement
error from oxygen analyzer is much smaller compared with that from the mass flow
controller, and the oxygen analyzer is sensitive and accurate enough to measure the
concentration changes in feed and retentate gases.
The measurement of O2 concentration is carried out by an electrochemical oxygen
sensor, the basic functioning of which is similar to a battery. Oxygen gas diffuses
through a membrane; contacts an electrode and is reduced to a negatively charged
hydroxyl ion. This ion moves through an electrolyte in the oxygen sensor to a
positively charged electrode typically made of lead. The hydroxyl ion reacts with the
lead and releases electrons. The electron flow is measured and can mathematically be
converted to an oxygen concentration. This electrochemical oxygen sensor is sealed in
the cell compartment of O2 analyzer by an O-ring. The gas sample can only enter and
242
exit from the cell compartment through the ¼ inch Swagelok® tube; therefore, there is
no internal leak within the O2 analyzer.
Because the retentate flow rate was too small in our study, it was impossible to
produce a high positive pressure from the inlet to the outlet of the oxygen analyzer to
prevent the interference of atmosphere entering into the oxygen analyzer. To
overcome this problem, a special design was made here. Another mass flow controller
(denoted as mass flow controller (2)) was connected with the O2 analyzer, as shown in
Figure B.1. As a result, the gas can only flow from the inlet to the outlet in the mass
flow controller and cannot flow inversely due to its special inner structure; therefore,
the gas in atmosphere could not diffuse into the system. This design (i.e. the use of the
second mass flow controller) is very important because it significantly extends the
application range of our testing system at various gas flow rates.
After opening the needle valves 2, 3 and 5 and closing the needle valve 4, the line
from the outlet of mass flow controller (1) to the outlet of mass flow controller (2) was
evacuated to remove the residual gas of the system (i.e., hollow fiber module and
tubing) before test. The electric power of O2 analyzer must be shut down during this
period; otherwise, this action may damage the sensor of O2 analyzer by boiling the
electrolyte. Thereafter, the needle valves 3 and 5 were closed and the electric power of
O2 analyzer was turned on. Thereupon, the retentate gas accumulated inside the O2
analyzer was indicated by a gradual increase in the reading of O2 analyzer. After an
243
appropriate period of time, needle valves 3 and 4 were opened and the superfluous gas
stored in the O2 analyzer streamed into the atmosphere through the mass flow
controller (2). The set point of mass flow controller (2) should be regulated to a value
slightly higher than that of the mass flow controller (1) in order to eliminate the
accumulation of the retentate gas in the O2 analyzer, which may mislead the reading.
With a decrease in the superfluous gas contents in the O2 analyzer, the set point of
mass flow controller (2) should be gradually adjusted until it was the same as that of
the mass flow controller (1) (i.e. 0.50ml/min for hollow fiber 1 and 0.57ml/min for
hollow fiber 2). Under these conditions, the O2 analyzer would provide a stable and
reliable reading of O2 concentration in the retentate gas. The above proposed operation
sequence is very critical to obtain a stable and accurate gas flow rate through the O2
analyzer, especially at low gas flow rates, which determines the reliability of the
reading of O2 concentration.
In order to validate the O2 composition of the feed gas to be 21 mol%, the third step is
to measure the composition of the feed gas using the O2 analyzer and then calculate
the composition of permeate gas through the mass balance. In this study, the
composition of permeate gas was not measured directly by using the O2 analyzer
because the permeate flow rate was so small that a much larger error may be
introduced during the measurement. The tubing of the feed gas was connected directly
to the inlet of mass flow controller (1) as illustrated in Figure B.1, and the composition
of the feed gas was measured similarly to test of the retentate gas.
244
The last step was that the apparent permeance of each species in the O2/N2 mixed gas
was calculated based on one set of differential equations developed by Wang et al.
[18], in which the non-ideal gas behavior and pressure drop inside the hollow fiber
have been considered. In the case of negligible permeate gas pressure, the selectivity
of hollow fiber membranes for a mixed gas was equal to the ideal selectivity; that is, it
can be calculated from the ratio of multicomponent permeances measured at the
partial pressure [1, 11, 20].
B.3 RESULTS AND DISCUSSION
The performance of this new testing system is evaluated by measuring the permeance
of two types of hollow fiber membranes in the mixed gas. Their pure gas permeance
was first measured using the similar technique described in Jones and Koros’ paper
[16]. The O2/N2 separation results of mixed gas and pure gas are listed in Table B.2. It
can be found that both permeance and selectivity of mixed gas are very close to those
of pure gas when considering the effect of different testing temperatures. The results
demonstrate that the new testing system devised in this work can effectively determine
the permeance and selectivity of hollow fiber membranes in the separation of O2/N2
mixed gas. The total price of apparatus applied in this new testing system including
one oxygen analyzer, two mass flow controllers and one bubble flow meter is no more
than S$10,000 (about US$6,000) which is much lower than that of a GC system.
245
Table B.2: Comparison of separation performance of hollow fiber membranes between the O2/N2 mixed gas and pure gas measurements
0.0423
0.00918
N2
0.271
0.0631
O2
Permeance (GPU)
200
200
Total feed pressure (psig)
Mixed gas (21 mol%±1% O2 in air) at 24oC
Hollow fiber 2
Hollow fiber 1
Membrane name
N2O2
Ideal selectivity
(O2/N2)
Permeance (GPU)
Selectivity (O2/N2)
6.20.04930.3066.4
6.50.01140.07416.9
Pure gas at 35oC
0.0423
0.00918
N2
0.271
0.0631
O2
Permeance (GPU)
200
200
Total feed pressure (psig)
Mixed gas (21 mol%±1% O2 in air) at 24oC
Hollow fiber 2
Hollow fiber 1
Membrane name
N2O2
Ideal selectivity
(O2/N2)
Permeance (GPU)
Selectivity (O2/N2)
6.20.04930.3066.4
6.50.01140.07416.9
Pure gas at 35oC
Pure O2 and N2 gases were tested at 132psig.
In order to further prove the authenticity of O2/N2 mixed gas permeation results
measured by the O2 analyzer, the same hollow fiber samples were used to separate the
CO2/CH4 (50/50 mol%) mixture in this study. The testing procedure was similar to
that applied in the mixed O2/N2 separation and the only difference is that the
composition of CO2/CH4 mixed gas was determined by GC instead of the O2 analyzer.
Both O2 analyzer-based testing system and GC-based testing system use the mass flow
controller or bubble flow meter to measure the gas flow rate, therefore, the error of
these two testing systems should be comparable and is mainly determined by the
accuracy of mass flow controller. Experimental conditions and parameters can be
found in Table B.1. CO2/CH4 separation results of mixed gas and pure gas are given in
Table B.3. It can be seen that the difference between mixed gas and pure gas is almost
negligible if considering the effect of different testing temperatures, which again
justify that this new testing system with the much lower cost can fulfill the purpose to
246
measure O2/N2 mixed gas separation performance through the hollow fiber
membranes.
Table B.3: Comparison of separation performance of hollow fiber membranes between the CO2/CH4 mixed gas and pure gas measurements
0.0249
0.00485
CH4
0.807
0.164
CO2
Permeance (GPU)
200
190
Total feed pressure (psig)
Mixed gas (50/50 mol% CO2/CH4) at 24oC
Hollow fiber 2
Hollow fiber 1
Membrane name
CH4CO2
Ideal selectivity (CO2/CH4)
Permeance (GPU)
Selectivity (CO2/CH4)
31.2 0.02670.83532.4
29.20.006410.18733.8
Pure gas at 35oC
0.0249
0.00485
CH4
0.807
0.164
CO2
Permeance (GPU)
200
190
Total feed pressure (psig)
Mixed gas (50/50 mol% CO2/CH4) at 24oC
Hollow fiber 2
Hollow fiber 1
Membrane name
CH4CO2
Ideal selectivity (CO2/CH4)
Permeance (GPU)
Selectivity (CO2/CH4)
31.2 0.02670.83532.4
29.20.006410.18733.8
Pure gas at 35oC
Pure CO2 and CH4 gases were tested at 100psig.
B.4 CONCLUSIONS
A new economical testing system using an oxygen analyzer has been designed to
determine the permeance and selectivity of hollow fiber membranes in the O2/N2
mixed gas separation. Two types of hollow fiber membranes are used to evaluate the
performance of this new testing system. Experimental results indicate the difference of
both permeance and selectivity between mixed gas and pure gas is almost neglectable,
which demonstrates this new testing system with the much lower cost can accurately
measure the permeance of each species in the O2/N2 mixed gas separation.
B.5 REFERENCES
247
[1] W.J. Koros, R.T. Chern, V.T. Stannett, H.B. Hopfenberg, A model for permeation
of mixed gases and vapors in glassy polymers, J. Polym. Sci. Polym. Phys. Ed., 19
(1981) 1513.
[2] M.R. Coleman, W.J. Koros, Conditioning of fluorine-containing polyimides. 2.
Effect of conditioning protocol at 8% volume dilation on gas-transport properties,
Macromolecules, 32 (1999) 3106.
[3] M.D. Donohue, B.A. Minhas, S.Y. Lee, Permeation behavior of carbon dioxide-
methane mixture in cellulose acetate membrane, J. Membr. Sci., 42 (1989) 197.
[4] E. Sada, H. Kumazawa, P. Xu, S.T. Wang, Permeation of binary gas mixture
through glassy polymer membranes with concentration-dependent diffusion
coefficient, J. Polym. Sci. Polym. Phys., 30 (1992) 105.
[5] R.S. Prabhakar, B.D. Freeman, I. Roman, Gas and vapor sorption and permeation
in poly(2,2,4-trifluoro-5-trifluoromethoxy-1,3-dioxole-co-tetrafluoroethylene),
Macromolecules, 37 (2004) 7688.
[6] P.C. Raymond, W.J. Koros, D.R. Paul, Comparison of mixed and pure gas
permeation characteristics for CO2 and CH4 in copolymers and blends containing
methyl methacrylate units, J. Membr. Sci., 77 (1993) 49.
[7] H. Ettouney, U. Majeed, Permeability functions for pure and mixture gases in
silicone rubber and polysulfone membranes: Dependence on pressure and
composition, J. Membr. Sci., 135 (1997) 251.
[8] J.H. Kim, S.Y. Ha, Y.M. Lee, Gas permeation of poly(amide-6-b-ethylene oxide)
copolymer, J. Membr. Sci., 190 (2001) 179.
248
[9] H. Yasuda, K.J. Rosengren, Isobaric measurement of gas permeability of polymers,
J. Appl. Polym. Sci., 14 (1970) 2839.
[10] D.G. Pye, H.H. Hoehn, M. Panar, Measurement of gas permeability of polymers.
II. Apparatus for determination of permeabilities of mixed gases and vapors, Appl.
Polym. Sci., 20 (1976) 287.
[11] K.C. O'Brien, W.J. Koros, T.A. Barbari, E.S. Sanders, A new technique for the
measurement of multicomponent gas transport through polymer films, J. Membr.
Sci., 29 (1986) 229.
[12] R.E. Kesting, A.K. Fritzsche, Polymeric Gas Separation Membranes, Wiley, New
York, NY, 1993.
[13] I.M. Wienk, R.M. Boom, M.A.M. Beerlage, A.M.W. Bulte, C.A. Smolders, H.
Strathmann, Recent advances in the formation of phase inversion membranes
made from amorphous or semi-crystalline polymers, J. Membr. Sci., 113 (1996)
361.
[14] T.S. Chung, A review of microporous composite polymeric membrane
technology for air-separation, Polym. and Polym. Comp., 4 (1996) 269.
[15] K.Y. Wang, T. Matsuura, T.S. Chung, W.F. Guo, The effects of flow angle and
shear rate within the spinneret on the separation performance of
poly(ethersulfone) (PES) ultrafiltration hollow fiber membranes, J. Membr. Sci.,
240 (2004) 67.
249
[16] C.W. Jones, W.J. Koros, Carbon molecular sieve gas separation membrane. 1.
Preparation and characterization based on polyimide precursors, Carbon, 32 (8)
(1994) 1419.
[17] M.H. Hassan, J.D. Way, P.M. Thoen, A.C. Dillon, Single component and mixed
gas transport in a silica hollow fiber membrane, J. Membr. Sci., 104 (1995) 27.
[18] R. Wang, S.L. Liu, T.T. Lin, T.S. Chung, Characterization of hollow fiber
membranes in a permeator using binary gas mixtures, Chem. Eng. Sci., 57 (2002)
967.
[19] C. Cao, R. Wang, T.S. Chung, Y. Liu, Formation of high-performance 6FDA-
2,6-DAT asymmetric composite hollow fiber membranes for CO2/CH4
separation, J. Membr. Sci., 209 (2002) 309.
[20] P.S. Tin, T.S. Chung, Y. Liu, R. Wang, S.L. Liu, K.P. Pramoda, Effects of cross-
linking modification on gas separation performance of Matrimid membranes, J.
Membr. Sci., 225 (2003) 77.
250
PUBLICATIONS
Journal Papers:
1. T.S. Chung, L.Y. Jiang, Y. Li, S. Kulprathipanja, Mixed matrix membranes
(MMMs) comprising organic polymers with dispersed inorganic fillers for gas
separation, Progress in Polymer Science, in press.
2. Y. Li, T.S. Chung, S. Kulprathipanja, Novel Ag+-zeolite/polymer mixed matrix
membranes with a high CO2/CH4 selectivity, AIChE Journal, in press.
3. Y. Li, L.Y. Jiang, T.S. Chung, A new testing system to determine the O2/N2
mixed gas permeation through hollow fiber membranes with an oxygen analyzer,
Industrial & Engineering Chemistry Research, 45 (2006) 871.
4. Y. Li, T.S. Chung, Z. Huang, S. Kulprathipanja, Dual-layer polyethersulfone
(PES)/BTDA-TDI/MDI co-polyimide (P84) hollow fiber membranes with a
submicron PES-zeolite Beta mixed matrix dense-selective layer for gas separation,
Journal of Membrane Science, 277 (2006) 28.
5. Y. Li, H.M. Guan, T.S. Chung, S. Kulprathipanja, Effects of novel silane
modification of zeolite surface on polymer chain rigidification and partial pore
blockage in polyethersulfone (PES)-zeolite A mixed matrix membranes, Journal
of Membrane Science, 275 (2006) 17.
251
6. Y. Li, T.S. Chung, C. Cao, S. Kulprathipanja, The effects of polymer chain
rigidification, zeolite pore size and pore blockage on polyethersulfone (PES)-
zeolite A mixed matrix membranes, Journal of Membrane Science, 260 (2005) 45.
7. Y. Li, C. Cao, T.S. Chung, K.P. Pramoda, Fabrication of dual-layer
polyethersulfone (PES) hollow fiber membranes with an ultrathin dense selective
layer for gas separation, Journal of Membrane Science, 245 (2004) 53.
8. Y. Li, Y.D. Wang, Y.X. Li, Y.Y. Dai, Extraction of glyoxylic acid, glycolic acid,
acrylic acid, and benzoic acid with trialkylphosphine oxide, Journal of Chemical
and Engineering Data, 48 (2003) 621.
9. Y.D. Wang, Y. Li, Y.X. Li, Y.Y. Dai, Liquid-liquid extraction of carboxylic acid
with trialkylphosphine oxide, Chemical Engineering (China), 31 (6) (2003) 8.
10. Y.D. Wang, Y.X. Li, Y. Li, J.Y. Wang, Z.Y. Li, Y.Y. Dai, Extraction equilibria of
monocarboxylic acids with trialkylphosphine oxide, Journal of Chemical and
Engineering Data, 46 (2001) 831.
Conference Papers:
1. Y. Li, P.S. Tin, T.S. Chung, S. Kulprathipanja, A novel ion exchange treatment of
zeolite for the application of mixed matrix membranes in natural and hydrocarbon
separation, presented in AIChE Annual Meeting 2006, San Francisco, California,
USA, Nov 12-17 2006.
252
2. T.S. Chung, L.Y. Jiang, Y. Li, S. Kulprathipanja, Fabrication and characterization
of zeolite/polymer mixed matrix nano-composite hollow fiber membranes,
presented in the Third Conference of Aseanian Membrane Society, Beijing, China,
August 23-25, 2006.
3. T.S. Chung, L.Y. Jiang, Y. Li, P.S. Tin, S. Kulprathipanja, Fabrication of
polymer/zeolite and carbon/zeolite mixed matrix membranes for gas separation,
presented in the 17th Annual Meeting of the North American Membrane Society
(NAMS), Chicago, Illinois, USA, May 12-17, 2006.
4. Y. Li, T.S. Chung, S. Kulprathipanja, Enhanced gas separation performance in
polyethersulfone (PES)-modified zeolite mixed matrix membranes, presented in
AIChE Annual Meeting 2005, Cincinnati, OH, USA, Oct 30-Nov 04 2005.
5. T.S. Chung, L.Y. Jiang, Y. Li, S. Kulprathipanja, Fabrication and characterization
of zeolite/polymer mixed matrix nano-composite hollow fiber membranes,
presented in AIChE Annual Meeting 2005, Cincinnati, OH, USA, Oct 30-Nov 04
2005.
6. T.S. Chung, L.Y. Jiang, Y. Li, S. Kulprathipanja, Fundamental understanding of
the science and engineering of the fabrication of zeolite/polymer nano-composite,
presented in the 4th International Membrane Conference on Membrane Science
and Technology, Chung-Li, Taiwan, Aug 18-19, 2005.
7. Y. Li, T.S. Chung, S. Kulprathipanja, Enhanced gas separation performance in
polyethersulfone (PES)-zeolite mixed matrix membranes, presented in the 16th
253
Annual Meeting of the North American Membrane Society (NAMS), Providence,
Rhode Island, USA, Jun 11-15, 2005.
8. T.S. Chung, D.F. Li, L.Y. Jiang, Y. Li, S. Kulprathipanja, Morphological and
structure control of dual-layer hollow fiber membranes formed by a phase-
inversion co-extrusion approach, presented in the 16th Annual Meeting of the
North American Membrane Society (NAMS), Providence, Rhode Island, USA,
June 11-15, 2005.
9. Y.D. Wang, Y.X. Li, Y. Li, J.Y. Wang, Extraction equilibrium of acetic acid and
its derivatives by Cyanex 923, presented in the Third Joint China/USA Chemical
Engineering Conference, Beijing, China, 2000, 07-049-053.
Patents:
1. Y. Li, P.S. Tin, T.S. Chung, S. Kulprathipanja, Ion exchange treatment of zeolites
for the application of mixed matrix membranes for natural gas and hydrocarbon
separation, US. Patent US60/792,094 (2006).
2. H.M. Guan, Y. Li, T.S. Chung, S. Kulprathipanja, Chemical modification of
zeolite surface for mixed matrix membrane, US. Patent US60/695,097 (2005).
Recommended