© Zahir Razzaz, 2019
Continuous Production of Porous Hollow Fiber Mixed Matrix Membranes for Gas Separation
Thèse
Zahir Razzaz
Doctorat en génie chimique
Philosophiæ doctor (Ph. D.)
Québec, Canada
ii
Continuous Production of Porous Hollow Fiber
Mixed Matrix Membranes for Gas Separation
Thèse
Zahir Razzaz
Sous la direction de :
Denis Rodrigue, directeur de recherche
iii
Résumé
Ce travail présente une nouvelle méthode sans solvant pour la production de membranes à fibres
creuses pour la séparation des gaz. La technologie repose sur une extrusion continue suivie de
l’étirage de polyéthylène expansé présentant une densité cellulaire élevée et une distribution
uniforme de la taille des cellules. Pour atteindre cet objectif, une optimisation expérimentale et
systématique a été appliquée afin de produire une morphologie de mousse riche et uniforme pour
développer une structure adaptée aux performances de la membrane pour la séparation des gaz.
À partir des échantillons obtenus, un ensemble complet de caractérisations comprenant les
propriétés morphologique, mécanique, physique et de transport gazeux a été réalisé. En
particulier, les performances de séparation ont été étudiées pour différents gaz (CO2, CH4, N2, O2
et H2). La première étape a consisté à combiner du polyéthylène linéaire de basse densité
(LLDPE) avec un agent d'expansion chimique (azodicarbonamide, CBA) afin d'optimiser le
procédé en termes de la concentration en CBA et du profil de température, ainsi que la vitesse
d'étirage. Les résultats ont confirmé que des échantillons avec une densité cellulaire plus élevée
peuvent améliorer les propriétés de perméation des gaz des membranes. La deuxième partie a
examiné l’ajout de polyéthylène de basse densité (LPDE) afin d’améliorer la structure cellulaire
grâce à une densité cellulaire plus importante et à une vitesse d’étirement plus élevée. Il a été
constaté qu'un mélange LLDPE/LDPE (70/30) augmentait de 10 fois la densité cellulaire et
réduisait également l'épaisseur de la mousse de 50% par rapport aux mousses de LLDPE seul.
Dans la troisième partie, l’addition de nanoparticules a été étudiée et s’est révélée être une
stratégie très efficace pour améliorer encore plus la structure cellulaire via un effet de nucléation
hétérogène. Les résultats ont montré que l'introduction de zéolithe poreuse (5A) comme agent de
nucléation cellulaire/modificateur de perméation des gaz améliorait considérablement la densité
cellulaire de la mousse (1,2×109 cellules/cm3) tout en réduisant les tailles moyennes de cellules
(30 µm). Les propriétés membranaires de cette membrane moussée à matrice mixte optimisée
(MMFM) ont également été considérablement améliorées, en particulier avec l’ajout de 15% en
poids de zéolithe car la perméance de l’hydrogène ainsi que la sélectivité H2/CH4 et H2/N2 ont été
augmentées d’un facteur 6,9, 3,8 et 5,9 respectivement, par rapport à la matrice seule (sans
zéolithe) et non moussée. Par conséquent, une combinaison de l'addition de particules (structure
cellulaire), d'étirement (surface interne) et de moussage (porosité) a conduit à la production d'une
iv
structure multi-poreuse à l'intérieur des membranes afin d'améliorer les propriétés de transport
des gaz. On s'attend à ce que ces MMFM puissent être efficaces et rentables en termes de vitesse
de production (méthode continue), en particulier pour l'industrie pétrolière où la séparation
H2/CH4 et H2/N2 est essentielle pour la purification de l’hydrogène.
v
Abstract
This work presents a novel solvent-free method to produce hollow fiber membranes for gas
separation. The technology is based on continuous extrusion followed by stretching of foamed
polyethylene having a high cell density and uniform cell size distribution. To achieve this
objective, a systematic experimental optimization was applied to produce a rich and uniform foam
morphology and to develop a suitable structure for gas separation membrane performance. From
the samples obtained, a complete set of characterizations including morphological, mechanical,
physical and gas transport was performed. In particular, the separation performances were
investigated for different gases (CO2, CH4, N2, O2 and H2). The first step was to combine linear
low density polyethylene (LLDPE) with a chemical blowing agent (azodicarbonamide, CBA) to
optimize the processing in terms of CBA content and temperature profile along with stretching
velocity. The results confirmed that samples with a higher cell density can improve the membrane
gas permeation properties. The second part investigated the addition of low density polyethylene
(LPDE) to improve the cellular structure by having a higher cell density and using higher
stretching speed. It was found that a LLDPE/LDPE (70/30) blend increased the cell density by a
factor of 10 times and also decreased the foam thickness by 50% compared to neat LLDPE foams.
In the third part, nanoparticle addition was investigated and found to be a very effective strategy
to further improve the cellular structure via a heterogeneous nucleation effect. The results showed
that the introduction of porous zeolite (5A) as a cell nucleation agent/gas permeation modifier,
substantially improved the foam cell density (1.2×109 cells/cm3) while decreasing the average
cell size (30 µm). The membrane properties for this optimized mixed matrix foam membrane
(MMFM) were also significantly improved, especially at 15 wt.% zeolite as the H2 permeance,
as well as H2/CH4 and H2/N2 selectivity were increased by 6.9, 3.8 and 5.9 times respectively,
compared to the unfoamed neat (unfilled) matrix. Hence, a combination of particle addition (cell
structure), stretching (internal surface area) and foaming (porosity) led to the production of a
multi-porous structure inside the membranes to improve the gas transport properties. It is
expected that these MMFM can be efficient and cost-effective in terms of processing rate
(continuous method), especially for the petroleum industry where H2/CH4 and H2/N2 separation
are essential for H2 purification.
vi
Table of contents
Résumé .................................................................................................................................. iii
Abstract ................................................................................................................................... v
Table of contents ....................................................................................................................vi
List of Tables .........................................................................................................................xi
List of Figures ...................................................................................................................... xii
Abbreviations .................................................................................................................... xviii
Symbols................................................................................................................................xix
Dedication ............................................................................................................................xxi
Acknowledgments.............................................................................................................. xxii
Forewords ......................................................................................................................... xxiii
General introduction ............................................................................................................... 1
Membrane technology ..................................................................................................... 1
Porous structure polymer membrane ............................................................................... 5
Cellular foam structure .................................................................................................... 6
Stretching method .......................................................................................................... 10
Thesis objectives and research contributions .................................................................... 11
1 Chapter 1........................................................................................................................ 15
1.1 Introduction .............................................................................................................. 15
1.2 Nanocomposite (mixed matrix) membranes ............................................................ 16
1.3 Principles of gas transport in membranes ................................................................. 17
1.3.1 Free volume .................................................................................................... 20
1.3.2 Tunable transport ............................................................................................ 21
1.3.3 Nano-gap formation hypothesis ...................................................................... 22
1.4 Porous structure membrane fabrication techniques .................................................. 23
1.4.1 Phase separation (inversion) membranes ........................................................ 24
vii
1.4.2 Hollow fiber membrane preparation ............................................................... 25
1.4.3 Melt spinning .................................................................................................. 26
1.4.4 Stretching ........................................................................................................ 27
1.4.5 Solution spinning ............................................................................................ 29
1.4.6 Particle leaching and solvent casting .............................................................. 31
1.4.7 TIPS and NIPS methods ................................................................................. 32
1.4.8 Other techniques ............................................................................................. 34
1.5 Composite foams structure ....................................................................................... 35
1.5.1 Foam processing ............................................................................................. 36
1.5.2 Chemical blowing agent ................................................................................. 36
1.5.3 Effect of CBA content on the foaming expansion and morphology ............... 37
1.5.4 Physical blowing agent ................................................................................... 38
1.5.5 Influence of nucleating agent .......................................................................... 39
1.5.6 Influence of processing conditions ................................................................. 41
1.6 Material selection for membrane .............................................................................. 42
1.6.1 Polymer materials for membrane .................................................................... 43
1.6.2 Gas transport properties in polyolefins ........................................................... 43
1.6.3 Polyethylene foam and challenges in the foam processing ............................ 45
1.6.4 Zeolites ............................................................................................................ 46
1.7 Conclusion ................................................................................................................ 49
2 Chapter 2........................................................................................................................ 50
Effect of processing conditions on the cellular morphology of polyethylene hollow fiber foams
for membrane ........................................................................................................................ 50
Résumé .................................................................................................................................. 51
Abstract ................................................................................................................................. 52
2.1 Introduction .............................................................................................................. 53
2.2 Experimental investigation ....................................................................................... 55
2.2.1 Materials used ................................................................................................. 55
2.2.2 Sample preparation and foaming process ....................................................... 55
2.2.3 Foam characterization ..................................................................................... 57
2.2.4 Morphology..................................................................................................... 57
viii
2.2.5 Pure gas permeation ........................................................................................ 58
2.2.6 Mechanical properties ..................................................................................... 59
2.2.7 Thermal properties .......................................................................................... 59
2.3 Results and Discussion ............................................................................................. 60
2.3.1 Optimization of the foaming process .............................................................. 60
2.3.2 Effect of CBA content .................................................................................... 60
2.3.3 Effect of the die temperature........................................................................... 63
2.3.4 Stretching speed .............................................................................................. 67
2.3.5 Gas permeation through hollow fiber foamed LLDPE membrane ................. 70
2.4 Conclusion ................................................................................................................ 72
Acknowledgments ............................................................................................................. 73
3 Chapter 3........................................................................................................................ 74
Gas transport properties of cellular hollow fiber membranes based on LLDPE/LDPE blends74
Résumé .................................................................................................................................. 75
Abstract ................................................................................................................................. 76
3.1 Introduction .............................................................................................................. 77
3.2 Experimental ............................................................................................................ 80
3.2.1 Materials ......................................................................................................... 80
3.2.2 Sample preparation ......................................................................................... 80
3.2.3 Characterization .............................................................................................. 82
3.2.4 Thermal properties .......................................................................................... 82
3.2.5 Morphological analysis ................................................................................... 82
3.2.6 Gas permeation ............................................................................................... 83
3.2.7 Mechanical properties ..................................................................................... 84
3.3 Results and discussion .............................................................................................. 84
3.3.1 Optimization of the foaming process .............................................................. 84
3.3.2 Effect of stretching speed................................................................................ 85
3.3.3 Thermal properties .......................................................................................... 89
3.3.4 Gas transport properties .................................................................................. 92
3.3.5 Mechanical properties of hollow fiber foamed LLDPE membrane and its blends
……………………………………………………………………………… 97
ix
3.4 Conclusion ................................................................................................................ 99
3.5 Acknowledgements ................................................................................................ 100
4 Chapter 4...................................................................................................................... 101
Hollow fiber porous nanocomposite membranes produced via continuous extrusion: Morphology
and gas transport properties ................................................................................................ 101
Résumé ................................................................................................................................ 102
Abstract ............................................................................................................................... 103
4.1 Introduction ............................................................................................................ 104
4.2 Materials and methods ............................................................................................ 107
4.2.1 Materials ....................................................................................................... 107
4.2.2 Sample preparation ....................................................................................... 107
4.2.3 Characterization ............................................................................................ 108
4.2.4 Gas permeation ............................................................................................. 110
4.3 Results and discussion ............................................................................................ 111
4.3.1 Preparation and optimization of MMFM ...................................................... 111
4.3.2 Morphology................................................................................................... 112
4.3.3 Effect of blending method on the cellular structure...................................... 114
4.3.4 Effect of the stretching rate on the cellular structure .................................... 117
4.3.5 Gas permeation performances....................................................................... 119
4.3.6 Influence of stretching speed on the membrane’s gas transport ................... 124
4.3.7 Mechanical and thermal properties ............................................................... 125
4.4 Conclusion .............................................................................................................. 129
Acknowledgement .............................................................................................................. 130
Conclusions and Recommendations for future work .......................................................... 131
General conclusions ........................................................................................................ 132
Recommendations for future work .................................................................................. 134
References ........................................................................................................................... 136
Appendix A: Abstract ......................................................................................................... 158
x
Appendix B: Equipment Conditions .................................................................................. 159
B1 Extruder and Die ................................................................................................... 159
B2 Calendaring equipment ........................................................................................... 160
B3 Air flow controller system ....................................................................................... 160
xi
List of Tables
Table 0.1. Most important suppliers and industrial applications of gas separation membranes .. 3
Table 1.1. Comparison between polymeric and inorganic membranes with MMM [58]. ……..16
Table 1.2. Molecular weight, kinetic diameter and diffusivities inside a typical polymer
membrane: low density polyethylene at 25oC [68]. ..................................................................... 19
Table 1.3. Classification of foams by cell size and cell density [129]......................................... 35
Table 1.4. Gas permeation results for different polyethylene grades. ......................................... 45
Table 2.1. Conditions tested for the foamed LLDPE hollow fibers. …………………………..56
Table 2.2. DSC results for the effect of stretching speed at a constant die temperature of 172C.
...................................................................................................................................................... 69
Table 3.1. Specifications of the hollow fibers produced ……………………………………. 81
Table 3.2. Heat of fusion and melting temperatures for the neat polymers and their blends. ..... 91
Table 4.1. Specifications of the hollow fibers produced . …………………………….......... 108
xii
List of Figures
Figure 0.1. Schematic representative of gas mixture via a hollow fiber membrane reproduced
from ref [4]. .................................................................................................................................... 2
Figure 0.2. U.S. energy consumption distribution (*Quads: a unit of energy, ~0.3045 kWh) [7].2
Figure 0.3. Membrane classification based on different characteristics [15]. ............................... 4
Figure 0.4. Polymer foam market by applications and different applications for polymer foams
[32]. ................................................................................................................................................ 7
Figure 0.5. Twin-screw extruder at Université Laval. ................................................................... 8
Figure 0.6. The screw configuration used for the experimental work. .......................................... 9
Figure 0.7. Schematic representation of: a) tubular/annular die and b) flat die (pslc.ws/macrog).
........................................................................................................................................................ 9
Figure 0.8. Typical SEM images of PP/HDPE blend films produced via melt extrusion: a) without
air cooling, b) with air cooling including cold stretching (draw ratio = 35%)) and followed by hot
stretching (draw ratio = 55%)) [42]. SEM images of a LLDPE film filled with zeolite (40 wt.%)
at different draw ratios: c) 50% and d) 200% [43]. ..................................................................... 11
Figure 0.9. Schematic diagram of the first part of objectives. ..................................................... 12
Figure 0.10. Overview of MMFM membrane preparation and optimization. ............................. 13
Figure 0.11. Overview of the different subjects covered in this thesis. ....................................... 14
Figure 1.1. Schematic representation of the different membrane systems. Reproduced from [59].
…………………………………………………………………………………………..17
Figure 1.2. Illustration of three types of gas diffusion through a membrane............................... 18
Figure 1.3. Schematic representation of the free volume for the transport of small penetrant
molecules in polymers [70]. ......................................................................................................... 20
Figure 1.4. Schematic representation of macroscale interconnecting pathways inside a nano-
colander network to provide enhanced transport characteristics stability [83]. ........................... 22
Figure 1.5. Schematic representation of the nano-gap formation in MMM or composite
membranes. Reproduced from [86]. ............................................................................................ 23
Figure 1.6. Different processes available to produce porosity in polymers................................. 24
Figure 1.7. Schematic diagram of membrane production via phase inversion (PI) [88]. ............ 25
xiii
Figure 1.8. Common hollow fiber membrane preparation methods used for gas separation [89].
...................................................................................................................................................... 26
Figure 1.9. Schematic representation of the cross-section of a spinning die: a) air and b) molten
polymer [90]................................................................................................................................. 27
Figure 1.10. Typical scanning electron micrograph of the surface of microporous membranes
produced using different hot stretching ratio at 120oC: a) 30%, b) 60% and c) 100%. The arrows
show the stretching direction [94]. d) Schematic representation of the production technique of
porous membranes [98]................................................................................................................ 28
Figure 1.11. General procedure for the production of porous hollow fiber membranes by melt-
spinning coupled with stretching. ................................................................................................ 28
Figure 1.12. Typical SEM micrographs of hollow fiber MMM with HKUST-1: (a),(b) cross-
section of hollow fiber MMM, (c) outer surface, (d) surface of finger-like porous (void) and (e)
sponge-like structure [105]. ......................................................................................................... 30
Figure 1.13. SEM micrographs reporting the cross-section of asymmetric hollow fiber
membranes: a) neat polysulfone, b) polysulfone/Cloisite 15A MMM and c) close-up view of b)
[103]. ............................................................................................................................................ 31
Figure 1.14. Schematic diagram of TIPS extrusion for hollow fiber membranes preparation.
Reproduced from [125]. ............................................................................................................... 33
Figure 1.15. Cross-section images of porous PSF membranes: a) by immersion into propanol and
b) by immersion in butanol, and c) porous PSF zeolite-filled MMM with 50 wt.% zeolite [127].
...................................................................................................................................................... 34
Figure 1.16. Typical image presenting the macrovoid structure of 6FDA-based MMM with
different SiO2 loading: (a) 5, (b) 10, (c) 15 and (d) 20 wt.% [26]. .............................................. 35
Figure 1.17. SEM micrographs of foamed polypropylene and CaCO3 (1% wt.) in the flow (left)
and transverse (right) direction [152]. ......................................................................................... 38
Figure 1.18. Typical SEM micrographs of LDPE foams containing: a) 0% and b) 5% of CaCO3
as a nucleation agent [165]. ......................................................................................................... 40
Figure 1.19. Schematic diagram of gas transport across MMM (A: ideal pathway) [53]. .......... 42
Figure 1.20. Schematic structure of some widespread zeolites [54]. ........................................... 47
Figure 1.21. Trend of the number of investigations during the last decade using zeolites for
separation application (Web of Science, Nov. 2018). ................................................................. 47
xiv
Figure 1.22. Effect of particle size on the permeability of different gases for PDMS-silicate
composite membranes [211]. ....................................................................................................... 48
Figure 2.1. Effect of particle size on the permeability of different gases for PDMS-silicate
composite membranes [211]. 57
Figure 2.2. Schematic representation of the hollow fiber membrane module set-up with an actual
image of a mounted sample. ........................................................................................................ 59
Figure 2.3. Micrographs of the samples produced with different chemical blowing agent contents
for images taken in the transverse (left column) and flow (right column) directions. Sample codes
are defined in Table 2.1. .............................................................................................................. 61
Figure 2.4. Cell size, cell densities, expansion ratio, and aspect ratio of the samples produced with
different chemical blowing agent contents. ................................................................................. 62
Figure 2.5. Young’s modulus, tensile strength and elongation at break of the samples produced
with different chemical blowing agent contents. ......................................................................... 63
Figure 2.6. Micrographs of the samples produced with different die temperatures (sample codes
are defined in Table 2.1). ............................................................................................................. 64
Figure 2.7. Effect of die temperature on cell density, average cell size and expansion ratio. Sample
codes are defined in Table 2.1. .................................................................................................... 65
Figure 2.8. Optimum foaming window for the cellular LLDPE hollow fiber membrane under the
conditions studied. ....................................................................................................................... 66
Figure 2.9. Effect of the die temperature on Young’s modulus, tensile strength and elongation at
break of the samples. Sample codes are defined in Table 2.1. .................................................... 67
Figure 2.10. Micrographs of the HFM produced using different stretching speeds. T direction (top
row) and F direction (bottom row). Sample codes are defined in Table 2.1. .............................. 68
Figure 2.11. Effect of stretching speed on cell density, expansion ratio and aspect ratio with 1.75%
CBA. Sample codes are defined in Table 2.1. ............................................................................. 69
Figure 2.12. Effect of stretching speed on Young’s modulus, tensile strength and elongation at
break of the HFM (sample codes are defined in Table 2.1). ........................................................ 70
Figure 2.13. Results for gas (CO2, N2, O2, H2, CH4) permeance (a) and selectivity (b) at 30C and
30 psia for unfoamed and foamed LLDPE hollow fiber membranes produced at different
stretching speed (sample codes are defined in Table 2.1). .......................................................... 71
xv
Figure 3.1. The most important specifications related to membrane preparation methods. ........ 78
Figure 3.2. Schematic representation of the experimental set-up for the continuous production of
hollow fiber foamed membranes. ................................................................................................ 81
Figure 3.3. Micrographs of the foamed LLDPE/LDPE blends of 100/0 (HFM), 70/30 (HFMB7-
5), and 60/40 (HFMB6-5) with a stretching speed of 5 m/min (see Table 3.1 for sample details).
...................................................................................................................................................... 85
Figure 3.4. Cell density, expansion ratio, and average cell sizes in the flow (F) and transverse (T)
directions of the samples with different LDPE contents in LLDPE/LDPE blends at a stretching
speed of 5 m/min (see Table 3.1 for sample details). .................................................................. 85
Figure 3.5. SEM micrographs in the T (left column) and F (right column) direction for different
stretching speeds. Sample codes are defined in Table 3.1. .......................................................... 87
Figure 3.6. Effect of stretching speed on: a) cell density, b) cell aspect ratio, c) expansion ratio
and d) sample thickness for samples produced with 1.75% CBA. Sample codes are defined in
Table 3.1. ..................................................................................................................................... 88
Figure 3.7. Standard deviation (SD) of the cell size distribution in the F and T direction as a
function of stretching speed for 5, 7, 9, 11, and 13 m/min. Sample codes are defined in Table 3.1.
...................................................................................................................................................... 89
Figure 3.8. DSC curves of LLDPE, LDPE and their blends. The heating rate is 10C/min. ...... 90
Figure 3.9. Degree of crystallinity as a function of stretching speed for the foamed hollow fiber
membranes based on LLDPE and its blend with LDPE. Sample codes are defined in Table 3.1.
...................................................................................................................................................... 91
Figure 3.10. Schematic presentation of the formation mechanism of the crystalline and amorphous
structure in LLDPE and its blends with LDPE during stretching. ............................................... 92
Figure 3.11. Permeance of different gases (CH4, CO2, N2, O2 and H2) at 30C and 30 psia for
unfoamed and foamed hollow fiber LLDPE membranes and the blends produced at a stretching
speed of 5 m/min (sample codes are provided in Table 3.1). ...................................................... 93
Figure 3.12. Results for: (a) H2 gas permeance and (b) H2/CH4 ideal selectivity at 30C and 30
psia for foamed hollow fiber LLDPE membranes and its blends with LDPE produced at different
stretching speed (sample codes are presented in Table 3.1). ....................................................... 94
xvi
Figure 3.13. Results for H2 gas permeance, H2/CH4 selectivity (at 30C and 30 psia) and
crystallinity for hollow fibers based on unfoamed LLDPE and foamed 70/30 LLDPE/LDPE
(HFMB7-11). ............................................................................................................................... 95
Figure 3.14. Schematic presentation of H2 and CH4 molecular diffusion into the cellular structure
of a foamed membrane................................................................................................................. 95
Figure 3.15. Permeance of CH4, CO2, N2, O2 and H2 through the HFMB-11 membrane (optimum
sample) as a function of testing temperature. .............................................................................. 97
Figure 3.16. Young’s modulus (a), tensile strength (b) and elongation at break (c) of the foamed
and unfoamed hollow fibers based on LLDPE and its blends with LDPE. ................................. 98
Figure 3.17. Effect of stretching speed on the Young’s modulus (a), tensile strength (b) and
elongation at break (c) of selected HFMB7 and LL70 (70/30) blends. ....................................... 99
Figure 4.1. Schematic representation of the extrusion set-up to produce hollow fiber mixed matrix
foamed membranes. …………………………………………………………………………108
Figure 4.2. Schematic representation of the (right) permeation set-up with the (left) hollow fiber
membrane module set-up. .......................................................................................................... 111
Figure 4.3. High-resolution images showing zeolite nanoparticle dispersion in the matrix. ..... 112
Figure 4.4. Micrographs of the hollow fiber foam membranes with different zeolite contents (0,
10, 15, and 20 wt. %) at a stretching speed of 5 m/min. ............................................................ 113
Figure 4.5. (a) Cell density and expansion ratio, (b) average cell sizes in the transverse (T) and
flow (F) directions of mixed matrix foam membrane (MMFMs) with different zeolite contents (0,
10, 15, and 20 wt %) at a stretching speed of 5 m/min. ............................................................. 114
Figure 4.6. Effect of the production method on the MMFM morphology with 15 wt.% zeolite:
(left) melt compounding and (right) dry blending. .................................................................... 115
Figure 4.7. Effect of blending method (melt compounding and dry blending) on the cell size in
both T and F directions, as well as cell density for MMFM0 and MMFM15. .......................... 116
Figure 4.8. Schematic representation of the porous and nonporous nucleating agent mechanism
of zeolite during foaming. .......................................................................................................... 116
Figure 4.9. Micrographs of MMFM15 produced at different stretching speeds (m/min) for the
(top row) transverse (T) and (bottom row) flow (F) directions. ................................................ 118
Figure 4.10. Effect of the stretching speed on cell density and cell aspect ratio (AR) in the flow
direction for MMFM15 (2.5% CBA)......................................................................................... 119
xvii
Figure 4.11. Effect of (a) zeolite loading and (b) stretching speed on the crystallinity of MMFMs.
.................................................................................................................................................... 120
Figure 4.12. Permeance of various gases (CH4, CO2, N2, O2, and H2) at 30 C and 30 psia for
foamed and unfoamed hollow fiber MMFMs produced with different zeolite loading (0, 10, 15,
and 20 wt %) at a stretching speed of 5 m/min. ......................................................................... 121
Figure 4.13. Permeation results for (a) H2 gas permeance and (b) H2/CH4 and H2/N2 ideal
selectivity at 30 C and 30 psia for different membranes as a function of zeolite loading. ...... 122
Figure 4.14. Effect of melt compounding and dry blending on H2/CH4 and H2/N2 ideal selectivity
and H2 permeance of the selected MMFM15 membranes with 15 wt % zeolite. ...................... 123
Figure 4.15. (a) Scanning electron microscopy (SEM) of the selected MMFM15 membranes and
(b) schematic representation of the gas molecule diffusion pathways. ..................................... 123
Figure 4.16. Results for H2 gas permeance and H2/CH4 and H2/N2 ideal selectivity at 30 C and
30 psia for MMFM15 at different stretching speeds. ................................................................ 124
Figure 4.17. (a) Young’s modulus, (b) tensile strength, and (c) elongation at break of the foamed
and unfoamed hollow fibers at different zeolite loading. .......................................................... 126
Figure 4.18. Effect of stretching speed on the (a) Young’s modulus, (b) tensile strength, and (c)
elongation at break of selected MMFM15 and L-15. ................................................................ 127
Figure 4.19. Derivative thermogravimetric analysis (DTGA) curves of MMFM samples
compared to unfoamed samples. ................................................................................................ 128
Figure 4.20. Thermogravimetric analysis (TGA) curves of MMFM samples compared to
unfoamed samples. ..................................................................................................................... 129
xviii
Abbreviations
ADC azodicarbonamide
ALPO aluminophosphate
Azo azodicarbonamide
CBA Chemical blowing agent
DC Direct current
DMA Dynamic mechanical analysis
DNA DeoxyriboNucleic Acid
DSC Differential scanning calorimetry
DTGA Derivative of the thermogravimetric analysis
GED Gas diffusion expansion
HDPE High density polyethylene
HMS High-melt-strength
L-PP Linear PP
NaCl Sodium chloride
PE Polyethylene
PEN Poly(ethylene naphthalate)
PEO Poly(ethylene-co-octene)
PET Poly(ethylene terephthalate)
PMP Poly(4-methyl-2-pentyne)
PP Polypropylene
PP-g-MA Maleic anhydride grafted PP
PS Polystyrene
PVDF Poly(vinylidene fluoride)
SC Supercritical
SCF Supercritical fluid
SD Standard deviation
SEM Scanning electron microscope
TGA Thermogravimetric analysis
xix
Symbols
A Area (cm2)
a Length of unit-cell (m)
AR Cell aspect ratio (-)
b Width of unit-cell (m)
D Average cell diameter (m)
E Tensile storage moduli (MPa)
F Formability factor (-)
L Thickness of the sample (m)
Mp Pressure sensitivity of microphone (V/Pa)
n Number of cells (-)
Nf Cell density (cells/cm3)
Nnucl. Cell nucleation rate (cells/cm3.s)
N0 Nucleation density (cells/cm3)
p Pressure (Pa)
P Permeation coefficient (Barrer)
PC Critical pressure (Pa)
T Temperature (C)
TC Critical temperature (C)
Tg Glass transition temperature (C)
Tm Melting temperature (C)
Tmax. dec Maximum decomposition temperature (C)
T10 Temperature of 10% mass loss (C)
wf Weight fraction of additives (-)
Y Young’s modulus (MPa)
Hexp Heat of fusion (J)
H* Heat of fusion of pure sample (J)
𝛥𝑃 Pressure drop of the gas/polymer solution (Pa)
ρ Density (g/cm3)
Expansion ratio (-)
xx
𝜒𝐶 Degree of crystallinity (-)
xxi
Dedication
Dedicated to my family
for their endless love, support and encouragement
When something is important enough
you do it even if the odds are not
in your favor!
Elon Musk
xxii
Acknowledgments
I would like to highly specify my eternal gratitude to my father and my mother for their constant
support and love for my whole life. I am also in deep appreciation to my sisters and their husbands
for their continuous support and encouragement.
I wish to offer my deepest appreciation to my supervisor, Professor Denis Rodrigue, for his great
support, advices and for its unique way of leadership and unfailing ideas. I am grateful for all his
considerations, help and patience to correct and return within hours all my manuscripts in the
course of my Ph.D. That was outstanding.
The help of the chemical engineering department technical staff, Mr. Yann Giroux, our group
technician, for his training and support on different equipment and also Jérôme Noël, Marc Lavoie
and Jean-Nicolas Ouellet, as well as André Ferland for their technical support during this research
project is also appreciated.
I would like also to thank Dr. Abolfazl Mohebbi and Dr. Xiao Yuan Chen for all their kind and
scientific help, all my friends and my colleagues in the chemical engineering department at
Université Laval for their friendship, discussion and guidance. I also would like to thank the thesis
jury and evaluation committee members for their time and valued input in reviewing my work.
Finally, I acknowledge the technical and financial support from the Center for Applied Research
on Polymers and Composites (CREPEC), Natural Sciences and Engineering Research Council of
Canada (NSERC), Research Center on Advanced Materials (CERMA), and Fonds Québécois de
la Recherche sur la Nature et les Technologies (FRQNT).
xxiii
Forewords
This thesis is composed of six main chapters and presented as papers insertion. This project was
supervised by Prof. Denis Rodrigue and the chapters are the following:
The first section presents a general introduction on the subject, some experimental strategies and
the objectives are briefly outlined. In chapter 1, a general literature review is presented on both
aspects of the study: foaming and membranes properties. A comprehensive study on the
fabrication of porous structure for membrane application is also given. Then, chapters 2-4
present the main experimental findings in the form of published or submitted journal papers. An
overview of these work is:
Chapter 2
Z. Razzaz, A. Mohebbi, D. Rodrigue, Effect of processing conditions on the cellular morphology
of polyethylene hollow fiber foams for membrane applications, Cellular Polymers, 37, 4-6, 169-
188 (2018).
As a first step, Chapter 2 presents a continuous method without any solvent to produce porous
hollow fibers for membrane applications. In this study, linear low density polyethylene (LLDPE)
was combined with azodicarbonamide (chemical blowing agent) to produce samples via
extrusion. A systematic investigation on the effect of processing conditions (chemical blowing
agent content and temperature profile) and post-processing (stretching rate) on the cellular
structure and gas permeation properties is presented based on a complete set of characterizations
xxiv
(morphological, mechanical, physical and gas transport) was performed. As for permeability and
selectivity of the hollow fiber foamed membranes (HFM), those were found to depend on the
cellular structure and also their crystallinity. A membrane with high cell density and uniform cell
size distribution has increased gas separation performance.
Chapter 3
Z. Razzaz, A. Mohebbi, D. Rodrigue, Gas transport properties of cellular hollow fiber
membranes based on LLDPE/LDPE blends, Cellular Polymers, submitted (2018).
Chapter 3 is devoted to further optimize the cellular structure of the hollow fiber foamed
membranes by adding LDPE. The addition of this polymer helped to produce a cellular structure
with higher cell density and higher stretching speed leading to thinner films (about 300 µm).
Finally, based on the morphological, gas transport, thermal and mechanical characterization
made, the optimum cellular structures for gas separation membrane application was determined.
The results showed that the permeance properties of semi-crystalline porous polymers the result
of a complex combination of the foam morphology (cell sizes and cell density) and the matrix
structure (composition and crystallinity).
xxv
Chapter 4
Z. Razzaz, D. Rodrigue, Hollow Fiber Porous Nanocomposite Membranes Produced via
Continuous Extrusion: Morphology and Gas Transport Properties, Materials, 11(11), 2311
(2018).
In Chapter 4, foamed hollow fiber mixed matrix membranes are produced using zeolites to
determine the effect of their concentration. The combination of particle addition, stretching and
foaming is shown to create a multi-porous structure to facilitate gas transport properties. The
results show that significant permeance and selectivity improvements were obtained compared to
the neat matrix. These results show that mixed matrix foamed membranes can be cost-effective,
easy to process and efficient in terms of processing rate, especially for the petroleum industry
where H2/CH4 and H2/N2 separation are important for H2 recovery.
Conclusion
Concludes the study and presents the main results obtained in this investigation. Possible future
directions for more investigation on hollow fiber mixed matrix membranes are also discussed.
xxvi
Further findings obtained from this study were presented in other forums:
X.Y. Chen, Z. Razzaz, S. Kaliaguine, D. Rodrigue, Mixed matrix membranes based on silica
nanoparticles and microcellular polymers for CO2/CH4 separation. Journal of Cellular Plastics,
54(2), 309-331 (2018).
Z. Razzaz, D. Rodrigue, Morphology and gas transport properties of foamed hollow fiber
membranes based on LLDPE/LDPE blends.16th International Conference on Advances on Foams
Materials & Technology, Montreal, Canada, (poster presentation), (SPE Foam 2018).
Z Razzaz, D Rodrigue, Production and characterization of hollow fiber foam LLDPE membranes
for gas separation via melt processing. 10th Annual CREPEC Colloquium, Montreal, Canada
(poster presentation) (2017).
Z. Razzaz, D. Rodrigue, Effect of chemical blowing agent concentration on the morphological
and mechanical properties of linear low-density extruded polyethylene foams, (oral and
presentation), 66th Canadian Chemical Engineering Conference, Quebec, Canada (2016).
1
General introduction
Membrane technology
Recently, the population and economy of the world are rapidly developing, leading to an increase
in energy demand and emission of greenhouse gases. Current environmental issues related to
industrial gas purification and manufacturing have driven scientists to look for more sustainable
and green technologies to solve these problems. During the last few decades, membrane
separation technology has appeared as an alternative and modern technique to design new
industrial applications since no chemical additives are used and no phase change is needed.
Membranes technology is seen as a promising environmentally friendly method for gas separation
such as hydrogen purification and recovery, nitrogen separation from the air [1], landfill gas
treatment, gas sweetening [2], carbon dioxide removal [3], biogas upgrading [4], and several
more. Modern membrane systems with innovative and effective designs are aimed to develop
processes to reduce production costs, energy use, waste generation and equipment size [5]. Hence,
membrane technology and separation science are well-known tools to overcome and solve some
important global issues (see Figure 0.1) [6]. Overall, traditional separation systems such as
evaporation, distillation and absorption are involved in 10 to 15% of the world’s energy use, for
which distillation alone accounts for about 50% of the total commercial energy consumption. But
a rough estimation concludes that membrane-based operations use about 90% less energy
compared to conventional distillation systems [7]. Figure 0.2 shows the US distribution and
supply of energy information in different area [7].
2
Figure 0.1. Schematic representative of gas mixture via a hollow fiber membrane reproduced
from ref [4].
Figure 0.2. U.S. energy consumption distribution (*Quads: a unit of energy, ~0.3045 kWh) [7].
3
In 1979, the first commercial membrane gas separation operation was used by Monsanto based
on their Prism® membrane for the separation of H2 from N2 during the manufacturing of ammonia
[8],[9]. Subsequently, membranes got more attention and have been used in several facilities
intended for a variety of applications [10]. Currently, Air Liquide (France), Aquillo (The
Netherlands), Air Products (USA), Ube (Japan), MTR (USA), UOP (USA) and Cameron (USA)
are the main players in gas separation having a significant proportion in the membrane gas
separation companies [11],[12]. The form and the size of this industry keeps on changing [13].
Membrane applications can be classified into three general categories: gas separation, gas
generation and gas purification. Table 0.1 presents a summary of the most important industrial
membrane-based gas separation applications [14]. Other classification have been reported for
membrane separation processes as presented in Figure 0.3.
Table 0.1. Most important suppliers and industrial applications of gas separation membranes [14].
Gas separation Application Supplier
O2-N2 Nitrogen generation
Oxygen enrichment
Permea (Air Products), Generon
(IGS), IMS (Praxair), Medal (Air
Liquid), Parker Gas Separation, Ube
H2-hydrocarbons Refinery hydrogen recovery Air Products, Air Liquide, Praxair,
Ube
H2-N2 Ammonia purge gas Air Products, Air Liquide, Praxair,
Ube
Hydrocarbons-air Pollution control hydrocarbon
recovery
Borsig, MTR, GMT, NKK
H2O-hydrocarbon Natural gas dehydration Kvaerner, Air Products
Hydrocarbons from
process streams
Organic solvent recovery
Monomer recovery
Borsig, MTR, GMT, SIHI
4
Figure 0.3. Membrane classification based on different characteristics [15].
Advantages and disadvantages of gas separation by membrane
There are significant advantages and features of using membranes for commercial processes
compared to conventional techniques. The most important ones are:
- Being efficient for consistent manufacturing with extremely high selectivity.
- Very simple to process and design.
- Simple automation and remote-control system.
- No need for phase change and chemical additives.
- Low cost and low energy consumption.
- High mechanical property and environmentally friendly.
- Possibility of recycling the byproducts and high productivity for raw materials use.
But the membranes also have some drawbacks such as: lower product purity than adsorption, low
contaminant resistance and low level of technological maturity (more research needed for
industrial scale use) [10],[16],[17].
5
The selection of a membrane material for gas separation is based on specific chemical and
physical properties because a material must be designed to meet higher properties to separate
specific gas mixtures (design safety). The membrane material must be strong (mechanical), stable
(long-term properties) and resistant (temperature and chemicals) for the intended gas separation
to be performed [6].
In general, the gas separation properties of a membranes depend on [6]:
- The materials (solubility, diffusivity, selectivity, separation factors),
- The membranes structure and thickness (permeability and permeance),
- The membrane configuration (flat, hollow fiber, porosity),
- The module and system design.
Porous structure polymer membrane
Several strategies have been developed to produce symmetric and asymmetric porous polymer
membranes, However, most of them are based on solvent casting followed by phase
separation/inversion [8],[18]. In these methods, solvents toxicity and costs
(recycling/elimination/treatment) are the main disadvantages of these non-environmentally
approaches. On the other hand, a few techniques have been proposed to manufacture a
microporous membrane without any solvents. Using leachable particles (salts) was shown to
produce an open-cell structure. Other non-solvent approaches are stretching with or without
particles [12] and direct foaming of semi-crystalline thermoplastics [13],[19]. But several stages
of thermal post-treatment are needed to stabilize the newly produced microporous and crystalline
structure to avoid shrinkage and warpage which are the main drawbacks of these techniques.
Asymmetric hollow fibers have become increasingly used for large gas volume separation
applications. The development of an asymmetric structure with a very thin selective layer (defect-
free) and a porous support layer with strong mechanical properties without any additional coating
is one of the most significant challenges in membrane production. However, the cellular structure
of a polymer can also be created through different foaming processes resulting in waste and
weight reduction, high strength to weight ratio, heat resistance, multi-functionality, low cost,
recyclability, as well as easier processing. The processing of porous cellular polymers consists of
6
various steps, each one needing to be controlled and optimized through various variables for a
specific technique. These issues will be further discussed in Chapter 2 based on a recent literature
review.
Cellular foam structure
Thermoplastic foams are polymers containing two main phases: a continuous solid polymer
matrix filled with a distribution of gas bubbles as the dispersed phase. Polymer foams are made
by introducing a gas within a polymer matrix via a foaming process. The two main foam structures
are closed-cell and open-cell. While the closed-cell structure has more mechanical (rigidity)
properties, open-cell foams are more flexible. Moreover, several applications have been reported
for open-cell foams such as tissue engineering, battery separators, filtration membrane and sound
insulators [20],[21]. Open-cell foams were also used as a porous support layer for gas separation
and liquid filtration. The cells are generally created by introducing a gas or a gas mixture using
either a physical or a chemical blowing agent. The gases are typically generated by the thermal
decomposition of a particle or a chemical reaction for chemical blowing agents (CBA). On the
other hand, inert gases or volatile liquids under high pressure/temperature are used as physical
blowing agents (PBA) to dissolve in the polymer matrix via a saturation process [22]. A key
concern with PBA is dosing and delivery which is done by specially designed equipment. By
using a PBA, a uniform mixture inside the polymer is done by saturating the polymer at high
pressure and/or temperature [23]. Then, cell nucleation and cell growth proceed after a
thermodynamic instability is induced (pressure and/or temperature drop). On the other hand,
chemical foaming can be more directly performed as the CBA (usually a powder) is releasing the
gas via thermal decomposition of a solid. As decomposition proceeds, the gas generated is
dissolved into the polymer and foaming can occur as for PBA [24]. Most of the chemical blowing
agents are in a powder or masterbatch form and are easier to handle with simple equipment
compared with physical blowing agents [25]. The morphological properties, such as cell size
uniformity as well as cell density, can control the physical, mechanical and gas transport
properties of hollow fiber polymer membranes [19],[26],[27].
Investigations on the cellular structure of polymer foams is a well-known and important topic in
material science because of their unique relation between morphology and multi-functionality for
a wide range of applications [28]. Thermoplastic foams have various benefits compared to
7
unfoamed/solid polymers including high impact strength [29], lightweight, good sound and
thermal insulation, high strength to weight ratio [18],[21], high fatigue resistance [30], low
materials and processing costs, as well as simple processing methods. These materials were
developed to answer commercial needs to decrease material consumption and costs for different
polymer applications [31]. As a result of these unique properties, cellular polymers are used in
numerous commercial applications such as packaging, automotive, biotechnology, filtration
process, acoustic absorbents and sporting equipment [20],[21]. The global polymer foam market
and common applications are presented in Figure 0.4. This wide range of applications is important
considering the great benefits these microcellular structures can offer and their low cost
processing. Nevertheless, ongoing investigations on cellular nanocomposite polymers are of
interest to extend their potential. A recent application is for gas separation membrane.
Figure 0.4. Polymer foam market by applications and different applications for polymer foams
[32].
8
Extrusion is the single best-known melt processing techniques to manufacture continuous
products, which is a high-volume manufacturing process shaping martials like films, sheet and
tubular forms. Figure 0.5 shows the extruder used in this investigation. Extrusion is generally
classified as single or twin-screw rotating inside a cylindrical barrel. This kind of device,
equipped with heaters, includes three distinct zones from the solid conveying or feeding zones,
to melting and pumping zones. Feeding the polymer powder or pellet into this equipment leads
to a molten polymer exiting from the die (shaping element). Besides mixing or blending with
other polymers, addition of particles and blowing agents are other operations performed by this
equipment. Direct extrusion is applied in a number of manufacturing applications such as
packaging, pipe/tubing and thermoplastic coating. In these cases, a twin-screw extruder is more
suitable because of better mixing performance and material control. A typical twin-screw
configuration is presented in Figure 0.6.
Figure 0.5. Twin-screw extruder at Université Laval.
9
Figure 0.6. The screw configuration used for the experimental work.
Furthermore, the dies, controlling the extruder production (flow rate), also control the final shape
of the polymer product. Two main kinds of dies are generally encountered in extrusion:
annular/tubular and flat. Polymer tubes (hollow fibers) can be produced by using a tubular/annular
die, while flat dies may be used to generate films (flat membranes). A schematic representation
and some examples of standard dies are shown in Figure 0.7.
Figure 0.7. Schematic representation of: a) tubular/annular die and b) flat die (pslc.ws/macrog).
10
Stretching method
This method is mainly applied to draw polymer films containing filler particles [33], melt
spinning follow by stretching (hot and cold) (Figure 0.8a and 0.8b) [34], droplets of an immiscible
polymer blend [35], hexagonal, beta unit cell crystals [36], a mixed solvent [37], or a stacked
lamellar structure [38],[39]. Since the purpose of this thesis is to produce porous hollow fiber
nanocomposite membranes by a foaming process followed by stretching, a presentation of the
other stretching strategies might help in understanding the importance of the procedure selected.
In stretching of a filled polymer, the polymer is initially blended with some particles and then
extruded to prepare a composite hollow fiber [40] or film [39]. By applying stretching, voids are
created at the polymer/particle interface due to poor adhesion and discontinuities in mechanical
properties between both phases. Consequently, the size and shape of these voids are controlled
by the level of stretching imposed (Figure 0.8c and 0.8d). The pore size and its distribution also
depend on the particles size and geometry, as well as particle concentration, processing conditions
and limited by the sample thickness. To eliminate the particle from the matrix following drawing,
the particles need to have weak adhesion with the polymer phase [41]. At the end, annealing is
performed to stabilize the morphology created.
11
Figure 0.8. Typical SEM images of PP/HDPE blend films produced via melt extrusion: a)
without air cooling, b) with air cooling including cold stretching (draw ratio = 35%)) and
followed by hot stretching (draw ratio = 55%)) [42]. SEM images of a LLDPE film filled with
zeolite (40 wt.%) at different draw ratios: c) 50% and d) 200% [43].
Thesis objectives and research contributions
The objective of this thesis is to develop a continuous and solvent-free manufacturing technique
to produce low cost hollow fiber porous nanocomposite (mixed matrix) membranes. The matrix
is made by cellular foam extrusion followed by stretching, which is using available common and
inexpensive polymers and fillers. The hollow fiber mixed matrix foamed membranes (MMFM)
to be produced should have a high cell density, uniform cell size distribution and low thickness.
For this purpose, this work is composed of two main parts. For the first part, the parameters
controlling the cell morphology of foamed LLDPE with and without zeolite as gas permeation
modifier/nucleating agent in the foam structure and using a chemical blowing agent are
investigated. For the first time, the zeolite nanoparticles are added to serve as a nucleation agent
for foaming and to optimize the cellular structure.
Thus, the specific objectives are to:
12
optimize the various effective processing parameters to control the foam structure quality
through continuous extrusion followed by stretching (calendaring speed, temperature
profile and die temperature) to produce hollow fiber PE foams having low thickness.
develop various strategies using polymer and nanoparticle blending to enhance the
foamability of LLDPE (type and concentration).
compare the cellular structure of foamed neat LLDPE and LLDPE/LDPE blend to
optimize the LDPE content based on the final application.
compare foamed LLDPE/LDPE blends with and without zeolite (nucleating agent) to
obtain an optimum nanoparticle content.
perform a complete set of characterization of the hollow fiber foams from a morphological
(cell density, cell size, cell aspect ratio, foam density) and to relate with the mechanical
properties (tension properties).
These objectives can be summarized in Figure 0.9.
Figure 0.9. Schematic diagram of the first part of objectives.
The aim of the second part is to establish and understand the relations between the porous
structure produced and the gas permeation property of the PE foams. Moreover, the effect of cell
sizes and shape on the permeation property of the hollow fiber MMFM will be studied and
13
optimized. Furthermore, the work will investigate the presence of zeolite as an active
filler/nucleating agent to determine its effects on the membrane cellular structure and gas
separation performances (permeance and selectivity). The specific objectives of the second part
can be described as to:
characterize the membranes and investigate the effect of morphology (cell density, cell
size distribution) on the mechanical (tension) and thermal (crystallinity) properties with
gas transport.
develop a new hollow fiber MMM based on a solvent free method and to study the effect
of filler content on the membranes gas separation.
investigate the effect of stretching (speed) on the morphological (cell density, cell size
distribution, cell aspect ratio), gas transport, mechanical (tension) and thermal
(crystallinity) properties.
Figure 0.10 presents an overview of the work for this second part.
Figure 0.10. Overview of MMFM membrane preparation and optimization.
14
Finally, improved gas transport properties of the foamed membranes through foam extrusion
followed by stretching, which combines the benefit of both methods to obtain a new multi-porous
polymer structure, is made. Figure 0.11 presents an overview of the whole project.
Figure 0.11. Overview of the different subjects covered in this thesis.
15
1 Chapter 1
This chapter presents a review on porous membranes for gas separation. It begins with an
introduction on the mechanisms related to gas transport in mixed matrix membranes (MMM).
Then a short overview of the fabrication of porous membranes is made with a focus on foaming
processes.
1.1 Introduction
Background and literature review
In general, membrane producers are looking for materials having high permeability and
selectivity leading to lower membrane production costs and energy savings. However, the
inherent disadvantage of conventional polymer membranes is a trade-off between selectivity and
permeability as reported by Robeson [44],[45]. This means that highly permeable membranes
have relatively low selectivity and vice-versa. But recent development showed that some
inorganic membranes have good permselectivity performances above the upper-bound trade-off
reported by Robeson. Nevertheless, these membranes are still only at laboratory scale. For
industrial and large-scale applications, they are limited because of poor processability as well as
high production costs. To solve these issues, investigations focussed on polymer membranes. It
had been reported that the cost of polymer membranes is estimated to be 1-3 times lower than for
inorganic membranes [6].
Over the last decade, significant efforts were devoted to improve the performance of polymer
membranes for gas separation to overcome the intrinsic trade-off. One approach to improve these
membrane performances was to use microporous adsorbents (zeolites, silica, etc.) as inorganic
fillers dispersed inside the polymer matrix leading to the concept of mixed matrix membranes
(MMM). It is believed that MMM have the potential to solve, at least partially, the trade-off limit
of neat polymer membranes by combining the excellent gas separation properties of inorganic
membranes with the good processability of polymers [46],[47]. If the MMM perm-selectivity is
enhanced by 10% or higher compared to the corresponding neat polymeric membrane, the MMM
is described as showing a "mixed matrix effect" [48].
16
1.2 Nanocomposite (mixed matrix) membranes
Good processability of gas separation membranes is a key requirement for commercial
applications. Kemp and Paul [49] were the first to report on the concept of MMM by improving
the selectivity of silicon rubber using zeolite 5A. Among the numerous types of membranes
available, mixed matrix membranes (MMM) based on polymer matrices filled with inorganic
particles have been extensively investigated [50],[51],[27]. Mixed matrix membranes (MMM)
are currently a very active research area where the objective is to improve the properties and
performances by careful selection of the polymer matrix and solid particles [52].
MMM are produced by the introduction of inorganic fillers into a continuous and bulk polymer
phase, actually producing composites or hybrid materials. The selection must take into account
processability and cost to get good performances [53],[54]. But these systems have different
limitations associated to the intrinsic properties of the polymer (resistance to contaminants,
thermal and chemical stability, mechanical resistance) [55]. Similar limitations are related to the
particles (high cost, lack of technology for continuous production, inherent brittleness and defect-
free membranes) [56],[57]. The general characteristics of the different membranes are compared
in Table 1.1.
Table 1.1. Comparison between polymeric and inorganic membranes with MMM [58].
Properties Polymeric
membrane
Inorganic
membrane
Mixed matrix
membrane
Cost Economical to
fabricate
High fabrication
cost Moderate
Chemical and thermal
stability Moderate High High
Mechanical strength Good Poor Excellent
Compatibility with
solvent Limited Wide range Limited
Swelling Frequently occurs Free of swelling Free of swelling
Separation performance Moderate Moderate Exceed Robeson upper
bounds
Handling Robust Brittle Robust
17
Figure 1.1. Schematic representation of the different membrane systems. Reproduced from [59].
A schematic configuration of the three types of membranes is presented in Figure 1.1. The gas
transport through polymer, inorganic and mixed matrix membranes is assumed to follow the
solution-diffusion model which is driven by preferential sorption or molecular sieving, as well as
the inorganic particles and polymer phase properties [59]. In particular, several efforts (post-
treatment) have been made to avoid the creation of defective pores in MMM. For instance, the
dual layer co-extrusion method was used to fabricate such membranes [60],[61].
1.3 Principles of gas transport in membranes
Gas separation via membranes is based on the splitting a gas mixture through the use of dense
(nonporous) polymer films having a selective permeability to each gas. The transport of each
molecule across the polymer is based on the solution-diffusion mechanism. The separation is
governed by a pressure difference across the membrane. The membrane can be in two forms:
hollow fiber and flat sheet. Generally, hollow fibers are preferred since they have higher effective
surface area.
Three fundamentals gas transport properties are commonly used in membrane technology:
Diffusion is the process where molecules move from regions of high concentration to regions of
lower concentration. The theories distinguish three main types of diffusion: Knudsen diffusion,
molecular diffusion or molecular sieving, and solution-diffusion (Figure 1.2) [62] [62].
Solution-diffusion: This mechanism relates the gas molecules transport through a membrane
based on permeability (P) or permeance (Q) [63]. A typical hollow fiber involves two parts: a
thin nonporous film (skin) along with a thick porous asymmetric tubular support (core). The thick
18
support layer makes it possible for the gas to pass freely and give mechanical strength, while the
nonporous layer does the separation function. This means that merely a thin layer controls the
separation performance. Then, the hollow fiber skin layer thickness (l) could be measured through
the ratio of gas permeability, as calculated from the flat membrane, as well as permeance within
the similar testing conditions while using identical material. However, the flat membrane
permeability is presumed to become nominally equal to the hollow fiber thin dense skin
permeability. The permeance is the gas flow rate normalized by the surface area and the pressure
difference across the membrane, it can be used for asymmetric membranes and (multilayer)
composite films (especially when its thickness is unknown). On the other hand, permeability is
used for dense films (flat sheet) with a specific thickness (δ). The permeance is normalized by the
thickness (P = Q δ). The unit are mol s-1 m-2 Pa-1, generally reported as Gas Permeation Units
(GPU) (1 GPU = 10-6 cm3 (STP) s-1 cm-2 cmHg-1). On the other hand, permeability is normally
reported in Barrer where 1 Barrer = 10-10 cm3 (STP) s-1 cm cm-2 cmHg-1 [64],[65].
Figure 1.2. Illustration of three types of gas diffusion through a membrane.
Generally, the permeability for any given gas is controlled by the membrane material properties
such as chemical and physical structures, but also by the permeant molecules (shape, polarity and
size) as well as the interactions between the permeant and the membrane. Typically, the shape
and size (kinetic diameter) of gas determine the diffusion characteristics across a membrane, but
the membrane-permeant interactions, related to gas solubility in the polymer membrane, is a
thermodynamic parameter [66]. Therefore, these kinds of membranes are able to carry out
separations requiring higher selectivity compared to separations solely based on size [11]. The
solution-diffusion mechanism has been mathematically presented as Equation (2.1) for the
19
measurement of the permeability coefficient (P) where D and S are the diffusion coefficient
(kinetics parameter) and solubility coefficient (thermodynamics parameter), respectively:
𝑃 = 𝐷 × 𝑆 (2.1)
In membrane gas separation, the permeant molecules are initially sorbed at the higher pressure
side of the membrane. Then, it diffuses inside the membrane due to a concentration gradient and
finally desorbs on the low pressure side. The permselectivity or ideal selectivity represents the
ability of a membrane to separate one gas (A) from another (B). It is presented as the permeability
ratio between two gases as:
𝛼𝐴/𝐵 =𝑃𝐴
𝑃𝐵=
𝑆𝐴
𝑆𝐵
𝐷𝐴
𝐷𝐵 (2.2)
Equation (2.2) shows that the permselectivity can be divided into two ratios: the solubility
selectivity and the diffusivity selectivity [10].
It is known that gas permeation is affected by [53],[67]:
1. Diffusivity and solubility of the small molecule in the polymer.
2. Crystallinity, chain packing, side group complexity, polarity, orientation, fillers, humidity and
plasticization.
Small gas molecules diffuse more quickly in dense materials compared to large molecules. For
instance, the diffusivities of various gases in LDPE at 25oC are presented in Table 1.2. It is clear
that the diffusivity decreases with increasing molecular size [14],[68].
Table 1.2. Molecular weight, kinetic diameter and diffusivities inside a typical polymer
membrane: low density polyethylene at 25oC [68].
Gas Molecular weight
(g/mole)
Kinetic diameter
(Å)
Diffusivity
(10-6 cm2/s)
CO2 44 3.30 -
O2 32 3.46 -
N2 28 3.64 0.320
H2O 18 2.65 -
CH4 16 3.80 -
He 4 2.60 6.8
H2 2 2.89 4.74
20
1.3.1 Free volume
One of the theories for elucidating the gas permeability inside polymer is related to chain mobility
and free volume. “The free volume in polymers is defined as the volume not directly occupied by
the atoms constituting the polymer chains” [11]. It can be expected that good polymer candidates
have a potential for several applications by trapping the molecules in a large amount of
interconnected free volume like micro-porous materials and potentially have applications in
membrane gas separation [69], storage of electric energy [70] and filtration pharmaceuticals [71].
A schematic representation of the free volume is presented Figure 1.3.
Figure 1.3. Schematic representation of the free volume for the transport of small penetrant
molecules in polymers [70].
The main factors influencing the permeability and selectivity of the polymer membranes are: (i)
polymer chains mobility, (ii) free volume between the chains and (iii) penetrant-polymer
interactions [72]. In general, glassy polymers have more selectivity while having lower
permeability than rubbery polymers. On the other hand, rubbery polymers have higher free
volume leading to higher permeabilities [73]. Moreover, since the glassy polymers have lower Tg
and low chain mobility, diffusivity selectivity is generally more effective for gas transport. So
small molecules like H2 are moving faster than larger ones such as CO2. On the other hand, the
free volume will increase by adding nanoparticles in a polymeric matrix. Hence, the resulting
MMM may show higher solubility selectivity [74].
For example, He et al. investigated the effect of particle size in MMM based on poly(4-methyl-
2-pentyne) (PMP) [75]. They found that the permeability of n-butane and the n-butane/methane
selectivity of silica-filled PMP significantly increased as the filler particle size decreased. An
important result of this investigation was that the n-butane/methane selectivity of PMP with 30
wt.% of nano-silica increased from 22 to 26 when the size decreased from 30 to 7 nm. Although
ho
le size
Jumplength ℷ
21
PMP is a glassy polymer, its high free volume might decrease the importance of solubility-
selectivity and make diffusivity-selectivity dominant for gas transport. This leads to condensable
gases like CO2 being more permeable than non-condensable ones like H2. This phenomenon is
called reverse selectivity.
A relationship between nitrogen permeability with free volume was observed for MMM based on
silica nanoparticle membranes as investigated by Winberg et al. [76] and Merekl et al. [77]. The
authors reported that the permeability coefficient increased with increasing particle loading. They
also showed that the particles can affect the polymer chains and increase the free volume between
polymer chains leading to higher gas diffusion enhancing gas permeability. This effect was also
reported in various investigations [78],[79].
Based on the relationships between the polymer free volume and molecular transport inside
polymers, a great deal of effort was devoted to find a way for creating/controlling the free volume
via the synthesis of various polymers and nanoparticles addition [69],[80]. Moreover, increasing
the free volume into a polymer matrix leads to higher solubility selectivity [52]. So several
investigations were devoted to artificially introduce free volume in polymer structures to improve
permeability and selectivity [81]. This was actually successful as reported in some experimental
works [78].
1.3.2 Tunable transport
The development of tunable transport properties for a membrane can be a way to improve the
performance of membranes by simultaneously increase both the surface area activity and gas
transport [82]. Porous interconnected network structures with tunable transport properties are an
interesting design for a variety of applications like energy devices and separation, which can
systematically control the selectivity and permeability of porous structure [69],[83],[84]. Kim et
al. [83] reported a multiscale porous network structure with tunable transport properties for
membrane separation using Au nano-filler in solution (Figure 1.4).
22
Figure 1.4. Schematic representation of macroscale interconnecting pathways inside a nano-
colander network to provide enhanced transport characteristics stability [83].
The effect of tunable gas transport properties on membrane performance has been reported for
different polymers and was shown to increase the membrane separation performance. Zhuang et
al. [85] reported the synthesis of copolyimides TB (Tröger’s base) membranes showing excellent
performance in molecular separation, especially for hydrogen, by increasing the tunability
properties. They presented promising gas membrane design for large-scale gas separation
applications in adverse environments. Moreover, due to their microporous structural range
through dianhydride monomers, the polymers produced may be effectively tunable by optimizing
the functional monomers. A microcellular polymer based on Matrimid 5218 and
tetraethoxysilane, tetramethoxysilane and tetrapropoxysilane with silica nanoparticles using a
sol-gel approach was also investigated by Chen et al. [27]. The performances of this microcellular
structure provided much higher CO2 permeability, but ideal selectivity was not improved even if
the results were close to the Robeson upper bound. Nevertheless, the high permeability of these
membranes makes them interesting for commercial application and further optimization.
1.3.3 Nano-gap formation hypothesis
Cong et al. [86] observed that the dispersion of unmodified silica in polymer matrices can enhance
the diffusivities and permeabilities of CH4 and CO2 without having significant change on the
CO2/CH4 selectivity compared with the neat polymer. Because of weak compatibility between
the polymer and the nanoparticles, the polymer chains are not in contact with the particles creating
some voids surrounding the particles as presented in Figure 1.5. In this case, the gas diffusion
pathway is reduced leading to increased gas diffusion and permeability. This explains why
nanoparticles addition improves gas permeability without affecting the gas selectivity. If the
nanoparticle have good compatibility with the matrix, the nano-gaps do not form eliminating this
defect [86].
23
Figure 1.5. Schematic representation of the nano-gap formation in MMM or composite
membranes. Reproduced from [86].
1.4 Porous structure membrane fabrication techniques Membranes today provide superior separations enabling them to be manufactured from several
materials and processes. The introduction of voids or microscopic pores inside polymer matrices
has been studied for several years leading to the development of pores creating processes. A few
of them have been applied for commercial application. So far, most of these methods are
performed as academic investigations and Figure 1.6 presents different techniques depending on
the materials and the equipment selected.
24
Figure 1.6. Different processes available to produce porosity in polymers.
*NIPS (Non solvent induced phase separation)
*TIPS (Thermally induced phase separation)
The processing of porous cellular polymer consists of several steps. The following section
provides a summary of the methods to generate porosity inside polymers with a focus on hollow
fiber membranes.
1.4.1 Phase separation (inversion) membranes
Loeb and Sourirajan [87] were the first to prepare cellulose acetate membranes by using phase
inversion in the 1960s. They proposed a novel method to prepare a wide range of membranes.
Various strategies including solution casting and interfacial polymerization were developed to
produce permeable and selective membranes. All of them are still used today. However, the phase
inversion process continues to be the most critical and conventional fabrication technique,
particularly for industrial membrane manufacturing. In this technique, a polymer solution can be
cast on an appropriate support with a casting knife, after which it is immersed in a coagulation
bath. This method leads to an asymmetric membrane having a nonporous top layer along with a
porous sublayer. The generation of this two layer structure is governed through several parameters
25
in the matrix dope solution including coagulant temperature, composition and additives. But this
process can be used to make both porous and non-porous structures. To obtain a specific
membrane morphology and separation performance, the phase inversion process needs to be very
well controlled. The preparation of a hollow fiber or flat sheet membrane is a complex procedure
since it involves several steps such as: the dope preparation, immersion-precipitation, rheology
of the casting solution, air gap and coagulant bath. Figure 1.7 presents a schematic representation
of the set-up.
Figure 1.7. Schematic diagram of membrane production via phase inversion (PI) [88].
1.4.2 Hollow fiber membrane preparation
Hollow fiber membranes are generally asymmetric and nonporous (dense) structures. These
structures vary merely within the process used for gel filament. The asymmetric structure is
generally created via solution spinning (phase separation) where a nonporous membrane is
produced by melt spinning. Asymmetric hollow fibers progressively became desirable and
intended for a large volume of gas separation applications. The development of asymmetric
structure with a very thin selective layer (defect-free) as well as a porous support layer with strong
26
mechanical properties without coating a thin layer is among the significant challenges. Hollow
fiber (spinning) is a complex physical procedure that typically consists of the following stages:
solution formulation, extrusion, coagulation and treatment after coagulation (Figure 1.7). The
main preparation methods of hollow fiber membranes used for gas separation are provided in
Figure 1.8.
Figure 1.8. Common hollow fiber membrane preparation methods used for gas separation [89].
1.4.3 Melt spinning
In this process, a molten polymer flow is made inside an extruder coupled with a spinning die. A
typical cross-section of this type of die (spinneret) is presented in Figure 1.9. Hollow fibers of
thermoplastic polymers can be manufactured by this technique. Polyolefins have been shown to
be a good option for hollow fiber membranes production [89]. PMP (poly(4-methyl-1-pentene))
is an example which was reported to have good gas transport properties [89]. Melt spinning is
economical and a non-solvent technique for hollow fiber membrane preparation since no
wastewater and hazardous materials are included. It also results in the creation of homogenous
nonporous structures. To create a microporous structure, spinning needs to be coupled with the
stretching [90].
27
Figure 1.9. Schematic representation of the cross-section of a spinning die: a) air and b) molten
polymer [90].
1.4.4 Stretching
A commercial size pore structure manufacturing method is via the controlled stretching of a
polymer film. The process consists of the melt-spinning or melt processing of a polymer and
followed by a drawing or stretching stage (Figure 1.10). This method enables to stretch in one
(uniaxial) or two (biaxial) directions [91]. The pores are generally formed as a result of
mechanical deformation, in which the morphology is determined based on the polymer properties
and stretching speed as well as temperature [91]. Melt spinning followed by stretching is a general
procedure for hollow fiber membrane preparation (Figure 1.11). This process is a very cost-
effective, environmentally friendly and easy method since membrane production by this process
does not use any solvent and any phase inversion step. Various papers are available on the
production of microporous structures such as PP [92],[93],[34],[94] PP/HDPE [42],[95] and PE
[96],[97],[40]. As presented in Figure 1.10d, industrial production was extended to
polytetrafluoroethylene (PTFE) which was commercialized by W. L. Gore and associates in their
products along with Celgard [98]. The limitation of this methods is that thermal post-treatments
are needed to stabilize the newly created microporous and crystalline structure to avoid
shrinkage/warpage which are the main drawbacks of such methods. The fabrication of hollow
fiber membranes is summarized in Figure 1.11.
28
Figure 1.10. Typical scanning electron micrograph of the surface of microporous membranes
produced using different hot stretching ratio at 120oC: a) 30%, b) 60% and c) 100%. The arrows
show the stretching direction [94]. d) Schematic representation of the production technique of
porous membranes [98].
Figure 1.11. General procedure for the production of porous hollow fiber membranes by melt-
spinning coupled with stretching.
29
1.4.5 Solution spinning
This technique is derived from the phase separation process. Based on the conditions applied in
solution spinning, this method can produce microporous or dense, asymmetric structure of hollow
fiber membranes. Generally, solution spinning is categorized into three approaches: wet spinning
[99],[100], dry spinning [101] and dry/wet spinning [102],[103]. Presently, there is no reported
industrial gas separation membrane manufactured by dry spinning. But recently, thin skin coating
layer hollow fiber membranes were shown to have excellent selectivity for CO2 and CH4
separation and these membranes were produced by a dry/wet forced evaporation technique [104].
1.4.5.1 Hollow fiber mixed matrix membrane
Hu et al. [105] used HKUST-1 into a PI (PMDA-ODA) polymer to produce hollow fiber MMM
applying the dry-wet spinning technique. Figure 1.12 presents the morphology of hollow fiber
MMM showing suitable interaction between the polymer and the particles since no interfacial
defect was found. Consequently, the separation performance of these MMM was improved even
at low particle content. For instance, 6 wt.% HKUST-1 led to 4% H2 permeability improvement
while the H2/CO2, H2/CH4, H2/O2 and H2/N2, ideal selectivities were improved by three fold
compared to the neat matrix.
30
Figure 1.12. Typical SEM micrographs of hollow fiber MMM with HKUST-1: (a),(b) cross-
section of hollow fiber MMM, (c) outer surface, (d) surface of finger-like porous (void) and (e)
sponge-like structure [105].
Zulhairun et al. [103] studied the separation performance (CO2/CH4) of hollow fibers based on
polysulfone and Cloisite 15A MMM membranes produced by dry/wet spinning. They reported
that the CO2 permeance increased by about four times by increasing the dope extrusion rate
modifying the MMM porous structure as showed in Figure 1.13.
31
Figure 1.13. SEM micrographs reporting the cross-section of asymmetric hollow fiber
membranes: a) neat polysulfone, b) polysulfone/Cloisite 15A MMM and c) close-up view of b)
[103].
Several spinning parameters are available to produce defect-free good separation performance
membranes. The spinning conditions including spinneret dimensions [106], force convection gas
flow rate [107], dry gap length [106],[108],[109], bore fluid flow rate and composition
[110],[109], temperature of the spinneret and dope solution [111], dope extrusion rate
[110],[108], coagulation bath [109], and take-up speed [106],[108],[112],[113],[114] have all
been reported to play an important role in achieving good morphology and separation
performances.
1.4.6 Particle leaching and solvent casting
Particle (particulate, salt, porogen) leaching can be used in combination with a variety of different
methods to obtained open-cell pores including compression molding, solvent casting, template
leaching. Template leaching as a non-solvent and continuous process is applied by combining
melt-extrusion followed by template leaching to form a porous structure membrane. In the particle
leaching technique, particles such as sugar, salt and especially created spheres are embedded
inside a polymer matrix. Then, the dissolved particles are washed out producing (additional)
porosity in the scaffold [115]. Using this method makes sure that membranes along with highly
controlled porosity as well as pore sizes are generally created [116]. On the other hand, this
method may not be suitable for all materials (soluble protein scaffolds). Also, post-treatment such
as washing out is usually time and cost consuming, as well as a high risk of residues remaining
(unleached particles) in the matrix. There is also some restrictions of the final thickness of about
1-2 mm [115]. This process is mainly intended for polymers which are not soluble in common
32
organic solvents [115],[116]. Porous open-cell structure membranes fabricated through this
technique consist of PE [117],[118] membranes using tapioca starch as the leachable component
[119] and also NaCl (as the leachable part) in polystyrene [120] and polypropylene [121]. Porous
membranes based on this method are mostly open-cell without a dense layer, So they need a dense
coating layer to improve the gas separation performance [122].
1.4.7 TIPS and NIPS methods
Polymer traditionally used as porous membranes were produced using the thermally induced
phase separation (TIPS), non-solvent-induced phase separation (NIPS), or a combination of both
[123],[124]. A polymer solution along with a solvent are put in contact with the polymer. Due to
primary interfusion between the solvent along with the non-solvent, the polymer solution
concentration changes up to a point where phase inversion takes place (the pores in the membrane
after solvent extraction). Thermally induced phase separation (TPIS) is an alternative approach
used for manufacturing porous structures based on the quality of solvent decreasing with lower
temperature. A schematic diagram of TIPS extrusion to prepare hollow fiber membranes is
presented in Figure 1.14. In comparison to NIPS, the main advantage of TIPS is that it is easier
to control the membrane structure due to only a few factors having an influence on creating the
porous structure [125].
33
Figure 1.14. Schematic diagram of TIPS extrusion for hollow fiber membranes preparation.
Reproduced from [125].
A few studies are available combining NIPS and zeolite-filled poly(vinyl acetate) [126] and
polysulfone [127] porous MMM for gas separation applications. Hollow fiber and flat sheet
module zeolite-filled porous polysulfone (PSF) MMM were prepared using NIPS [127]. Oxygen
permeance increased from 0.07 to 0.13 GPU compared to porous PSF, while the selectivity
slightly decreased from 5.6 to 5.1. The gas transport properties of this porous structure were found
to be highly dependent on the zeolite (4A) loading. The permeance increased up to 3.71 GPU
with a substantial reduction in selectivity (1.3). This was related to undesirable voids in PSF and
particle interface defects, especially with increasing particle content (50 wt.%) as presented
Figure 1.15.
34
Figure 1.15. Cross-section images of porous PSF membranes: a) by immersion into propanol
and b) by immersion in butanol, and c) porous PSF zeolite-filled MMM with 50 wt.% zeolite
[127].
1.4.8 Other techniques
Composite membranes can also be produced by other methods. One possibility is a sol-gel
procedure where nanoparticles could be synthesized in situ at the same time as the polymerization
of the monomers or organic polymer. Chen et al. [26] reported that macrovoid structured MMM
consisting of silica nanoparticles and co-polyimide were produced from 6FDA-based polymers
through a sol-gel method. Typical SEM micrographs of these MMM for different SiO2 content
are presented in Figure 1.16. The results showed that a series of silica–6FDA-based MMM with
different SiO2 loadings could improve the CO2/CH4 separation performance. Chen et al. [27] also
successfully produced Matrimid (BTDA-DAPI polyimide) MMM incorporating nano-silica
synthesized via a sol-gel method using various precursors. The nano-silica was produced in situ
through membrane casting with a homogeneous distribution within a cellular polyimide structure.
The CO2 permeability improved by about five times (9.9 to 46.3 Barrer) compared to the neat
polymer matrix, where a reduction in CO2/CH4 selectivity from 35.3 to 28.3 was observed.
35
Figure 1.16. Typical image presenting the macrovoid structure of 6FDA-based MMM with
different SiO2 loading: (a) 5, (b) 10, (c) 15 and (d) 20 wt.% [26].
The main issues regarding MMM are poor spatial distribution associated to particle
agglomeration [27], as well as possible interfacial rigidification and pore blockage [128].
1.5 Composite foams structure
Cell density and cell size are two parameters used to classify foams [129]. Table 1.3 presents the
four main types of foams including conventional, fine celled, micro-cellular and nano-cellular
foams.
Table 1.3. Classification of foams by cell size and cell density [129].
Foam type Cell size
(μm)
Cell density
(cells/cm3)
Conventional > 300 < 106
Fine-celled 10 - 300 106 - 109
Micro-cellular 0.1 - 10 109 - 1015
Nano-cellular < 0.1 > 1015
Cellular foam processing is generally divided in different cell density between 106 cells/cm3 [129]
and 1015 cells/cm3 [130]. Suh and Martini proposed this definition for the first time in the 1980s
by producing a material with high toughness and low materials use [131]. Colton and Suh
obtained a microcellular foam with semi-crystalline polymers for the first time [132],[133]. Then,
36
the batch foaming process of semi-crystalline polymers was proposed by Doroudiani et al. [134].
They reported that the crystalline region influenced the blowing agent diffusivity and solubility
to modify the cellular structure. The foams were developed by continuous extrusion by Dow
[135].
1.5.1 Foam processing
The foaming process is composed of four main steps: I) saturation (foaming agent addition), II)
nucleation (bubble formation), III) bubble growth, and IV) stabilization (morphology). The most
typical category based on the process and mechanism which gas can be liberated as a result of
these compounds are called blowing agents including chemical and physical blowing agents
[136],[137].
These four steps have specific objectives:
The polymer is saturated by dissolving the high-pressure inert gas or the gas is liberated
from CBA in the saturation step.
Nucleation, several nuclei are produced by creating an unstable thermodynamic state via
a temperature and/or pressure drop.
Growth, various bubbles gradually increase in size by the diffusion of the dissolved gas.
Stabilization, based on the foaming process, cell growth is restricted by cooling. Thus,
the cell structure is fixed.
Two different nucleation mechanisms are known: homogenous and heterogeneous [138].
Homogeneous nucleation takes place because of the presence of a critical amount of gas
molecules dissolved inside the polymer matrix. Heterogeneous nucleation occurs at an interface
with a particle acting as a nucleating agent inside the polymer [139].
1.5.2 Chemical blowing agent
Typically, chemical blowing agents (CBA) are categorized based on whether the decomposition
reaction is exothermic or endothermic. Endothermic CBA are also called inorganic CBA since a
carbonate species, sometimes sodium bicarbonate, is used in the generation of gases (mostly CO2
with a minor content of H2O). The active ingredients of an endothermic CBA can be introduced
during polymer processing as either a powder or within a polymer pellet comprised of a low
melting temperature carrier resin (masterbatch). Materials like sodium bicarbonate, citric acid and
37
their salts are used as endothermic chemical foaming agents. The most widely used exothermic
chemical foaming agent is azodicarbonamide. Nowadays, a mixture of exothermic and
endothermic chemical foaming agent is also used for a better temperature control [140].
Two main benefits of using CBA exist: it is easy to handle and add to the composition, and they
can be simply processed by using typical equipment (no modification required). Other CBA
advantages include self-nucleation, wider operating window, as well as finer cell size. CBA could
be added in almost any thermoplastic system to create cellular foams including extrusion,
rotational molding, injection molding, calendering, casting and coating [141].
Some requirements of an ideal CBA are [137]: The decomposition temperature must be close to
(but above) the melting point of the matrix, non-toxic, non-flammable and non-corrosive. The
gases released are another requirement for selecting a chemical blowing agent [137]. The gas
released must be in high amount and must easily dissolve/distribute within the molten polymer.
The decomposition reaction must be fast but controlled and the residues must not cause
discoloration [142].
1.5.3 Effect of CBA content on the foaming expansion and morphology
The foaming expansion increased with blowing agent content if all the gas generated by the CBA
decomposition stays inside a polymer. The amount of gas liberated available for foaming is
directly related to the CBA content [24]. Rodrigue et al. [143] reported that blowing agent content
on foamed PE, prepared by compression molding, significant by affects the foam density, but
there is a limitation. If the polymer viscosity does not support the high gas pressure, the surface
starts to be rough and cell collapse occurs [144]. The viscosity can be increased by using a higher
molecular weight polymer by decreasing the temperature. But at higher viscosity, the force related
to gas expansion and bubble growth might not be effective due to high resistance [24], [145].
Considering that solubility associated with a gas in the molten polymer at high temperature is
limited, an excess amount of CBA is not able to create more foaming due to diffusion (gas loss)
leading to reduced performance of the blowing agent. Competition between the amount of
resistance (viscosity) and gas diffusion exists [144]. For example, Sahagun et al. [146]
investigated the effect of blending and foaming high-density polyethylene/polypropylene blends
through extrusion using azodicarbonamide as a chemical blowing agent. They found that the foam
morphology can be improved by adding a compatibility agent (Kraton D 1102).
38
1.5.4 Physical blowing agent
To produce microcellular foams, the polymer can be saturated by using an inert gas in a
supercritical state. A high number of nuclei can be generated by this type of blowing agent [147].
A good physical blowing agent must satisfy the following requirements: be inert in the polymer
matrix and should not modify the chemical and physical properties of the other components. It
should be nontoxic, odorless, nonflammable, environmentally safe and economical. A
supercritical fluid (SCF) can be defined as a substance for which its temperature and pressure are
higher than its critical values. So, the properties of the fluid are between those of liquids and
gases. In this condition, the material is able to dissolve like a liquid and diffuse into solids like a
gas [148]. Supercritical CO2 and N2 are well established as PBA. The pressure and temperature
in the critical state for CO2 are Pc = 7.38 MPa and Tc = 304 K [149]. In this condition, CO2 can
reduce the viscosity of polymers by means of plasticization [150]. Similarly, the conditions for
N2 are Pc = 3.4 MPa and Tc = 126 K in the supercritical state [151].
In a comparison between N2 and CO2, it should be mentioned that the supercritical temperature
of CO2 is above room temperature and also much higher than N2. Thus, it is only required to
increase the pressure above the critical pressure to get supercritical N2. In a previous work in our
team, Mohebbi et al. [152] produced closed-cell polypropylene microcellular foams by twin-
screw extrusion using N2 as a physical blowing agent and calcium carbonate as a nucleating agent.
The extrusion conditions were optimized to create ellipsoidal cells via stretching to improve the
internal specific surface area for piezoelectric applications (Figure 1.17).
Figure 1.17. SEM micrographs of foamed polypropylene and CaCO3 (1% wt.) in the flow (left)
and transverse (right) direction [152].
39
1.5.5 Influence of nucleating agent
Some additives, like micro or nano-particles, can be used as nucleating agents to control the
foaming process leading to more uniform foams. Furthermore, improving the functional and
mechanical properties of polymeric foams can be done by adding fillers having specific
characteristics [153]. Moreover, solid particles can improve the morphology and macroscopic
properties at the same time [154]. Therefore, adding particles leads to higher cell density and
expansion ratio, but also to cell walls thinning in the final cellular structures [155]. Generally
speaking, dissolved gases like N2 could merely diffuse inside the amorphous structure while the
crystalline structure can also play the role of nucleation sites [134]. This leads to higher melt
strength/viscosity to prevent cell coalescence and provide better cell growth resistance [156]. The
morphology of homogeneous polyolefin foams is generally not uniform (cell structure) due to a
competition between heterogeneous nucleation of the foam and the crystals. This is especially the
case when nanoparticles (nanoclays) are added [157],[158]. So the crystallinity of semi-
crystalline polymers like HDPE and PP may decrease the role of nano-fillers for improving the
cellular structure uniformity [158].
Based on the classical nucleation theory, filler particles can improve the foaming process by
decreasing the critical cell radius [159]. So any changing in the critical radius can lead to an
effective modification of the final cell structure [160]. Moreover, by increasing the matrix melt
strength it is possible to create better cell structures through well-distributed filler particles in the
foam processing. McClurg reported that to produce desirable foams (small cell size and high cell
density), uniformly distributed particles is necessary to get uniform distributions [161]. In some
cases, blending the polymer matrix with another immiscible polymer can be used to create
nucleation sites for microcellular foaming. Tejeda et al. [162] studied the effect of polyethylene
and polypropylene blending on the mechanical and morphological properties of the foams. It was
shown that the interface between immiscible polymers led to improved nucleation site to lower
cell sizes. Several other studies about nucleating agents like clay [163],[164], CaCO3 [165],[152],
silica [25],[166],[27], and Fe3O4 (OA-Fe3O4) [167] are available.
A chemical blowing agent (CBA) thermally decomposes in a temperature range to generate the
gases needed for cell nucleation and growth in the polymer melt. For thermoplastic polyolefins
and their composites, CBA have been more popular because of their easy use when high density
40
(100-450 kg/m3) foams are needed [146]. However, CBA leave residues in the polymer that
sometimes are not desired.
For example, Saiz-Arroyo et al. [25] produced LDPE foams by using silica as a nucleating agent
along with azodicarbonamide (ADC) as a chemical blowing agent via compression molding.
They reported that ADC increased the mechanical properties by modifying the cell structure as a
result of the incorporation of silica particles [25]. Abbasi et al. [165] improved the properties of
foamed low-density polyethylene by using nano-CaCO3 as a nucleating agent in comparison with
neat LDPE microcellular foams. They created a microcellular foam having a uniform and fine
cell structure. Also, they reported that a non-uniform cell structure was formed when the
nanoparticles agglomerated (Figure 1.18).
Figure 1.18. Typical SEM micrographs of LDPE foams containing: a) 0% and b) 5% of CaCO3
as a nucleation agent [165].
Based on the literature, it is believed that similar behaviors are obtained with CBA and PBA,
especially when particles are added. Saiz-Arroyo et al. [25] reported that low density polyethylene
and silica were foamed via CBA and PBA in a batch process. It was also possible to improve the
process using compression molding. They also reported a similar behavior for both blowing
agents: adding silica up to an optimum content led to cell size reduction. But the optimum silica
content for both methods was the main difference: 6 wt.% for CBA and 1 wt.% for PBA. These
differences can be described by the classical nucleation theory [139]. However, the nucleation
ratio, a parameter providing the effect of a particle as a nucleating agent, was lower for the PBA
foams than for CBA.
41
1.5.6 Influence of processing conditions
Foaming temperature has a significant effect on the final foam structure. Firstly, the cell growth
is facilitated by increasing temperature. On the other hand, decreasing temperature leads to wider
cell size distributions and small cell sizes. This is due to the increase melt-strength and viscosity
of the polymer limiting cell growth [156],[168]. Secondly, among the thermoplastic polymers,
foam processing of a semi-crystalline polymer is more challenging compared with amorphous
polymers, owing to the narrow foaming temperature window and the crystalline structure [169].
The foaming process of large parts or continuous processes is more challenging with narrower
foaming temperature [170]. Some studies reported that the presence of nanoparticles enables to
carry out foaming at a lower temperature compared to neat polymers [156]. Moreover, some
researchers showed that composites are less sensitive to temperature variation compared to neat
polymers [171],[170]. It was also reported that blending HDPE with PP reduced the temperature
dependence of HDPE [171]. Lee et al. [172] reported that in a batch system, the temperature
influences the cell density of foamed LLDPE in two ways: melt strength and skin creation. Xu
Y., et al. [173] showed that in a batch system, the control of both the saturation pressure and
temperature has an important effect on the foaming behavior. So the pressure, pressure drop and
pressure drop rate all have an important effect in cell nucleation [174]. The classical nucleation
theory can be used to determine the effect of pressure drop on cell nucleation rate (Nnucl.)
[139],[154]:
𝑁𝑛𝑢𝑐𝑙. = 𝑓0𝐶0exp(−∆𝐺𝑛𝑢𝑐𝑙.
𝑘𝑇) (2.3)
where ∆𝐺𝑛𝑢𝑐𝑙. is the Gibbs free energy determined by:
∆𝐺𝑛𝑢𝑐𝑙. =16𝜋𝛾𝑏𝑝
3
3∆𝑃2 (2.4)
where ∆P is the pressure drop of the gas/polymer solution, C0 is the concentration of gas
molecules, f0 is the frequency factor of gas molecules joining the nucleus, k is the Boltzmann’s
constant, γp is the surface energy of the polymer-bubble interface and T is the temperature.
Consequently, the cell density should be constant at a constant pressure drop by having an abrupt
pressure drop rate. Since the pressure drop occurs during a limited period, the pressure drop rate
can have an influence on the nucleation time as well as the nucleation rate [174]. Moreover,
changing the mass flow rate and the die geometry in extrusion can modify the pressure drop and
pressure drop rate [175]. Reducing the diameter/length of the die can cause a higher pressure drop
42
rate, leading to increased nucleation rate [150]. The influence of pressure on the foaming behavior
of polypropylene and carbon fiber by sc-CO2 as a physical blowing agent was investigated by
Wang et al. [176]. At the minimum pressure (10 MPa), large cell size distribution was obtained:
small cells were created around the crystalline areas while larger cells were generated in the
amorphous zones. But more uniform cell size distribution was obtained when the saturation
pressure was increased to 15 and 20 MPa.
1.6 Material selection for membrane
Several variables have a direct effect on membrane separation performance including membrane
structure and material, membrane thickness, system design and module configuration like flat
sheet or hollow fiber [177]. The main purpose of MMM fabrication is to use the excellent gas
transport and permeation properties of inorganic fillers while keeping the flexibility of polymer
materials. In an ideal MMM, good adhesion between the polymer and fillers is needed to limit
interfacial defects (see Figure 1.19) [53]. Numerous polymeric materials have been used for
membranes fabrication such as cellulose acetate, polyamides, polysulfone, etc. But there is still a
limitation in the commercial use of these materials despite all the efforts made for their
development for gas separation applications [178],[179].
Figure 1.19. Schematic diagram of gas transport across MMM (A: ideal pathway) [53].
43
1.6.1 Polymer materials for membrane
Polymeric membranes for commercial separations (e.g. gas separation, pervaporation and
filtration) are attractive owing to low manufacturing cost and investment, energy savings, easy
processability and operation, as well as mechanical stability compared with other traditional
methods for gas separations [180]. However, there is a limitation in polymer application due to
low selectivity, weak stability at high temperature and low solubility [181],[182].
Basically, commercial gas separation membranes are divided in two groups: rubbery and glassy
polymers. Glassy polymers are glass-like, rigid, and used below their glass transition temperature.
They have long relaxation times and low chain segmental mobility. On the other hand, rubbery
polymers show reverse properties. They are flexible and operate above their glass transition
temperature. They have short relaxation times and higher chain segmental mobility. Based on the
literature, glassy polymers have high selectivity and low permeability while rubbery ones have
low selectivity and high permeability [183]. Examples of glassy polymers are polyacetylenes,
polyamides and cellulose acetate, while rubbery polymers includes polydimethylsiloxane
(PDMS), polypropylene and polyethylene. To ensure a suitable interaction between the polymer
and particles, which is the most critical problems in the selection of polymer material, in
comparison to glassy polymers, rubbery ones have high chain mobility getting defect-free
interfacial membranes, leading to strong particle-polymer adhesion.
1.6.2 Gas transport properties in polyolefins
Common petroleum-based polyolefin polymers are largely used in industrial applications such as
polyethylene (PE), low-density polyethylene (LDPE), linear low-density polyethylene (LLDPE)
and polypropylene (PP). Polyolefins can provide good properties to produce composite
membranes. Polyolefins are the most widely thermoplastic materials because of their good
properties such as: processability, thermomechanical properties, recyclability and operational
flexibility. Polyolefins are inexpensive and commonly used in microporous membranes due to
their good resistance of solvent and thermal [184]. Moreover, the production cost of polyethylene
hollow fiber membranes is much less than that of polyvinylidene fluoride (PVDF) as well as other
fluorinated polymers. The properties and chemical stability of polyethylene are generally better
than polypropylene due to fewer side groups on the molecular backbone [40]. For all these
44
reasons, polyethylene was selected in this investigation as a suitable porous membrane material
[81].
Extensive reports have been done over the 1960-1989 period upon gas transport properties of
polyolefins (low and high density polyethylene [185], polypropylene [186], poly(ethylene-co-
vinyl acetate) (EVA) [101] and poly(4-methyl pentene)) [187]. However, these polymers are
semi-crystalline, thus their gas permeation properties are mainly depending on crystallinity
degree as well as morphology [188]. They have been commonly used in microfiltration [189] and
gas separation as support layers [190]. The use of dense polyolefin films for gas separation as a
membrane are limited. Because of that, they show low permeation behavior as a result of their
semi-crystalline properties [191]. Microporous or micro-sized pores polyolefin membranes have
been commonly produced by cold-stretching [39],[42], melt-spinning [34] and thermally induced
phase separation (TIPS) [184]. Covarrubias et al. [192] reported on the gas permeation across a
PE/aluminosilicate film membrane having higher carbon monoxide (CO) and H2 permeability
compared to the neat polyethylene film because of the intercalated structure of the
aluminosilicate. The aluminosilicate filling led to increased gas transport across their microporous
network and reduced the crystal size of polyethylene.
Moreover, disturbing the chain packing in the polyethylene is the main reason for higher gas
permeability. It offers a molecular sieve effect by creating functional pathways for small permeant
gas molecules. In another report, Covarrubias et al. [193] used ethylene polymerization reaction
of metallocenes filled by zeolite nanocrystals. They reported that polyethylene has the ability to
improve from internal and external sites around the sieving zeolite crystal and incorporating the
zeolite filler inside the matrix. The permeation behavior of N2-plasma and NH3-plasma treatment
with rubbery PE film membranes was investigated for O2, N2 and CO2 [194]. The permeability
of both N2-plasma and NH3-plasma increased for all the gases permeability as well as CO2/N2
and O2/N2 selectivity. Gholizadeh et al. [101] used LDPE and poly(ethylene-co-vinyl acetate)
(EVA) as a polymeric matrix produced by film blowing as well as thermal dry phase inversion
method. They measured permeability, solubility and diffusivity of CO2 and O2 to investigate the
influence of temperature and pressure on this polymeric system. They found that the permeability
and solubility of the polyethylene as a rubbery polymer and EVA decreased, but diffusivity
increased with increasing film thickness. This is related to the crystalline content. For O2, the
effect of pressure on permeability was low but permeability increased with increasing
45
temperature. They also reported that the polymer chains, by having polar groups, will increase
the CO2 permeability as well. However, polyolefin nanocomposites preparation for gas separation
membrane is barely studied. Checchetto et al. [195] prepared LDPE/graphite nano-platelets
composite membranes by film blowing. The nanoparticle addition led to lower gas permeability
due to barrier effects, but without significant change in selectivity. The separation behavior of
some polyethylene membranes is reported in Table 1.4. The gas transport properties of PE
(LLDPE and LDPE) are mainly determined by crystallinity; i.e. permeability is inversely
proportional to crystallinity. This reduction of permeability is relatively higher for larger gas
molecules because the mobility of the amorphous polymer chains is restrained by the physical
crosslinking effect of the crystallites. As the gas molecular size increase, the effect is more
important.
Table 1.4. Gas permeation results for different polyethylene grades.
Polymer Permeability
(Barrer) Selectivity
Crystallinity
(%) Ref.
N2 CO2 O2 CH4 He He/N2 CO2/ N2
LDPE 9.5 154 - - 15.8 - 16.3 0 [185],[196]
LDPE 4.0 48.4 - - 4.9 3.9 12.1 0.29 [185],[196]
LDPE 0.97 12.6 - - 1.1 5.1 13 0.43 [185],[196]
LDPE 0.14 0.36 - - 8.0 2.5 0.77 [185],[196]
LDPE 1 12.8 2.9 5 4.9 - - - [197]
PE - 14.6 4.8 - - - - - [101]
PE 0.6 - 2.0 - 2.9 - - 60 [192]
1.6.3 Polyethylene foam and challenges in the foam processing
Polyethylene foam is one of the most common polymeric foams with the largest volume
production among the thermoplastic foams. Its commercial production started in the 1940s [24].
Since then, several works were published on the structure-properties relationships of different PE
46
grades such as high density polyethylene (HDPE) [24],[145], medium density polyethylene
(MDPE) [198] and low density polyethylene (LDPE) [25],[166],[199]. PE has a significant
worldwide use because of its excellent overall performance such as chemical resistance,
mechanical behavior, high ductility and impact strength, good processability and low cost [200].
However, foaming polyethylene, which is a semi-crystalline polymer, is limited due to its narrow
foaming temperature window. This is especially true for “linear” molecules like HDPE and linear
low-density polyethylene (LLDPE) having very few branches and limited molecular weight
leading to poor rheological behavior in the melt state (low melt strength). This is why very limited
works have been published on LLDPE foaming. However, some literature on rotomolding [201]
and batch foaming system [202] can be found. The cell walls of polyethylene foam may have cell
rupture during foaming because of poor melt strength [203],[204]. Methods to improve the
nucleation behavior and melt strength of linear polyethylene were proposed: blending with high
melt strength polymers [205],[206], adding particles such as nucleating agents [156],[207],
crosslinking [208] and introducing long chain branching [131],[175].
1.6.4 Zeolites
The most commonly used filler in mixed matrix membrane fabrication are zeolites. The most
common family of microporous solids are crystalline aluminosilicates also called molecular
sieves. Their remarkably specific pore structure and size selective inherently lead to consider
them as excellent candidates for MMM (Figure 1.20). The number of investigations during the
last decade on zeolites for separation application are presented in Figure 1.21. In membrane gas
separation, some polymers with low packing density are less attractive, so a combination of
zeolites particle inside polymer matrices provide a possibility to overcome this issue [54]. Zeolites
with a highly polar surface inside the pores is the most important factor for adsorption. This
differentiate them with various commercially available adsorbents and made them applicable for
water purification as well as polar components removal such as CO2. Zeolites and molecular
sieves absorbents have different absorption rate based on molecular and physical properties such
as: shape, size, polarity and counter-ions [209].
47
Figure 1.20. Schematic structure of some widespread zeolites [54].
Figure 1.21. Trend of the number of investigations during the last decade using zeolites for
separation application (Web of Science, Nov. 2018).
Biswas et al. [43] used LLDPE as the matrix and zeolite as an inorganic filler for producing a
microporous film by extrusion. A morphological analysis by scanning electron microscopy
(SEM) showed that good adhesion and distribution between the polymer matrix and filler was
obtained. Also, the mechanical properties increased with increasing zeolite concentration (up to
40 wt.%). Another study reported that the CO2, N2 and O2 gas permeability inside LDPE with
zeolite improved [210]. Chen et al. [65] investigated the synergistic effect of zeolite and
polyimide in MMM for CH4/CO2 mixture. They reported two different pathways for the diffusion
48
of CH4 and CO2, in which CO2 mostly diffused through the zeolites while CH4 diffused through
the polymeric phase.
For gas transport properties, firstly gas molecules transport through zeolites by adsorbing into the
pores, secondly diffusing across the pore and lastly desorb. Inside a polymer matrix, the effect of
the zeolite on the MMM gas separation uses the selective adsorption or size/shape discrimination,
according to the selective adsorption properties and the pore sizes. As an example, according to
the molecular sieving influence, zeolite 4A and 5A having a pore size of 3.8 Å and 4.3 Å, can be
more suitable for the separation of O2 from N2 and CO2 from CH4.
1.6.4.1 Effect of zeolite particle size
The effect of the zeolite particle size on the performance of silicalite-PDMS mixed matrix
membranes was investigated at two different zeolite loadings (20 and 40 wt.%). In this study,
zeolites with different sizes (0.1, 0.4, 0.7, 0.8 and 8.0 micron) were used for the separation of O2,
N2 and CO2 as shown in Figure 1.22a. The permeabilities of these MMM increased with
increasing particle size, but the particle size effect was more evident at higher zeolite loading.
Figure 1.22b compares the ideal selectivity between different gases and particle size on. The
selectivity is mainly controlled by the particle (zeolite) loading, while particle size has less effect
[211].
Figure 1.22. Effect of particle size on the permeability of different gases for PDMS-silicate
composite membranes [211].
Most investigations on MMM are based on fillers having a large particle size; i.e. in the micron
range. But sometimes, smaller particles have more surface area and therefore better membrane
performance can be obtained. He et al. [75] investigated the effect of particle size in MMM based
a) b)
49
on PMP. They found that the n-butane permeability and the n-butane/methane selectivity of silica-
filled PMP significantly increased as the filler size decreased. An important result of this
investigation was that the PMP n-butane/methane selectivity with 30 wt.% of nano-silica
increased from 22 to 26 when the particle size decreased from 30 to 7 nm.
1.7 Conclusion
A variety of techniques have been developed to produce porous structure membranes. But mixed
matrix membranes used for gas separation are still prepared using high cost and toxic solvents as
well as diluents, making their preparation non-environmentally friendly and non-economical.
Using the cellular structure in polymers, could be beneficial in several areas. However, not all
polymers can be easily made porous by different methods. This is why optimization must be made
based on several properties.
The porous structure can be produced via chemical reactions. Examples are phase separation,
solvent casting and polymer mixing. Other methods involve mechanical stresses (stretching) or
physical (foaming and leaching) methods.
The next chapters will present a continuous and non-solvent technique combining the benefits of
both stretching and particle addition foaming to generate a multi-porous structure inside a
polymer matrix to continuously produce membranes with improved gas transport properties.
50
2 Chapter 2
Effect of processing conditions on the cellular
morphology of polyethylene hollow fiber foams for
membrane
Zahir Razzaz, Abolfazl Mohebbi, Denis Rodrigue, Cellular Polymers, 37, 169-188 (2018).
51
Résumé
Une méthode continue sans aucun solvant est proposée pour produire des fibres creuses poreuses
destinées aux applications à membrane (HFM). Dans ce cas, du polyéthylène linéaire de basse
densité (LLDPE) a été combiné avec de l’azodicarbonamide pour produire des échantillons par
extrusion. En particulier, les conditions de traitement (teneur en agent gonflant chimique et profil
de température) et de post-traitement (vitesse d'étirage) ont été optimisées pour obtenir une
structure cellulaire présentant une densité cellulaire élevée et une distribution uniforme de la taille
des cellules. À partir des échantillons obtenus, un ensemble complet de caractérisations a été
réalisé (propriétés morphologique, mécanique, physique et de transport de gaz). Les résultats
montrent que les HFM ayant une densité cellulaire plus élevée peuvent augmenter la perméabilité
aux gaz, en particulier pour l'hydrogène. Globalement, il a été montré que des polyoléfines à
faible coût ayant une structure cellulaire appropriée peuvent être utilisées pour les membranes de
séparation de gaz.
52
Abstract
A continuous method without any solvent is proposed to produce porous hollow fibers for
membrane (HFM) applications. In this case, linear low density polyethylene (LLDPE) was
combined with azodicarbonamide to produce samples via extrusion. In particular, the processing
(chemical blowing agent content and temperature profile) and post-processing (stretching
velocity) conditions were optimized to obtain a cellular structure having a high cell density and
uniform cell size distribution. From the samples obtained, a complete set of characterizations was
performed (morphological, mechanical, physical, and gas transport). The results show that HFM
having a higher cell density can increase gas permeability, especially for hydrogen. Overall, it is
shown that low cost polyolefins having a suitable cellular structure can be used for gas separation
membranes.
Keywords: Polyethylene, foams, membranes, extrusion, optimization, gas transport.
53
2.1 Introduction
Membranes separation based on polymers are commercially interesting because of their low costs
and operating temperature combined with easy processability. Several attempts have been made
to create a membrane improving the gas transport properties based on material selection,
especially for large-scale applications. However, very few investigations focused on the structure
control (morphology) of the membrane [27],[69],[83].
Several methods are available today for the production of symmetric and asymmetric porous
polymeric membranes depending on their final applications. But most of the polymer membranes
are produced by the well-known phase inversion technique [212] or solvent casting using polymer
solutions [27]. Unfortunately, the presence of solvents in these methods is a serious drawback
because they are costly, non-environmentally friendly and need to be removed/recycled/treated.
To the best of our knowledge, only a few solvent-free methods to produce porous polymer
membranes having microcellular morphology were developed [40],[213]. The most important
one is melt-spinning and stretching based on the melt-processing of homogeneous semi-
crystalline polymers such as polypropylene [34]. However, this method needs several thermal
post-treatment and mechanical stretching to stabilize the crystalline structure while avoiding
membrane shrinkage [34]. Other developments for the production of open-cell polymers and
composites are using leachable salts with polypropylene [120], low-density polyethylene [122],
and polystyrene [121].
To control the porous structure, polymer foams were developed which can be easily produced
using several matrices (mostly thermoplastics) and methods (melt-processing). These polymers
contain two main phases: a continuous solid polymer matrix filled with a distribution of gas
bubbles as the discontinuous phase [22]. In most cases, the production of polymer foams is
performed by the introduction of gases inside the polymer matrix. The gaseous phase can be
obtained from a physical blowing agent (PBA) or a chemical blowing agent (CBA). A key
concern with PBA is dosing and delivery which is done by specially designed equipment. By
using a PBA, a uniform mixture inside the polymer is done by saturating the polymer at high
pressure and/or temperature [23]. Then cell nucleation and cell growth proceeds after a
thermodynamic instability is induced (pressure and/or temperature drop). On the other hand, the
chemical foaming approach can be more directly performed as the CBA is releasing the gas via
thermal decomposition of a solid [214,215]. As decomposition proceeds, the gas generated is
54
dissolved into the polymer and foaming can proceeds as for PBA [24]. Generally, most chemical
blowing agents are in a powder or masterbatch form and are easier to handle with simple
equipment compared with physical blowing agents [25].
Microcellular foams are known for their excellent properties such as high impact strength [29],
lightweight, good sound and thermal insulation, high strength to weight ratio [21],[227], high
fatigue resistance [30], low materials and processing costs, as well as simple processing methods.
These materials were developed to answer commercial needs to decrease material consumption
and costs for different polymer applications [31]. Today, microcellular foams are commonly used
for insulation, automotive, renewable energies, biotechnology and packaging applications
[20,21]. This wide range of applications is important considering the great benefits these
microcellular structures can offer and their economical processing. Nevertheless, ongoing
investigations on microcellular polymers is of interest to extend their potential. A recent
application is for gas separation membrane applications.
Polyethylene (PE) foam is one of the most common thermoplastic foams with the largest volume
production. Its commercial production started in the 1940s [24]. Since then, several works were
published on the structure-property relationships of different PE grades such as high density
polyethylene (HDPE) [24,145], medium density polyethylene (MDPE) [198], and low density
polyethylene (LDPE) [25,166,199]. PE has a significant worldwide use because of its excellent
overall performance such as chemical resistance, high ductility and impact strength, good
processability, and low cost [200]. However, foaming polyethylene, which is a semi-crystalline
polymer, is limited due to its narrow foaming temperature window. This is especially true for
“linear” molecules like HDPE and LLDPE because they have very few branches and limited
molecular weight leading to poor rheological behavior in the melt state (mostly no melt strength).
This is why limited works have been published on LLDPE foaming. However, some literature on
rotomolding [201], batch foaming [202,206] and extrusion [218] can be found. Hence,
azodicarbonamide (Azo) is used to foam linear low density polyethylene (LLDPE) here. It is
expected that polyethylene crystallites can act as local gas barriers and increase the tortuous
diffusion path for the penetrant gas molecules and membranes produced with linear low density
polyethylene with relatively low crystallinity has been reported [219].
Polymer foams can be obtained through various processing methods such as extrusion [220],
compression molding [221], injection molding [222], phase separation [223], and in-situ
55
polymerization [224]. It should be mentioned that among these methods, extrusion is the only
continuous process. So extrusion is known to be economic, fast and effective to produce excellent
low density foams [137],[225]. As expected, the foaming conditions have a substantial influence
on the final cellular structure (cell size, density, geometry, orientation, etc.) [144], and therefore
on the thermal, physical and mechanical properties. But barrier properties have been less
investigated.
So it is expected that a cellular structure having a high cell density and uniform cell size
distribution can have highly tunable transport properties for membrane applications [27],[26].
The performance of a membrane can be improved by simultaneously increasing both surface area
and gas transport properties [82]. Hence, the main objective of this study is to develop a novel
continuous and non-solvent method to produce microcellular foams by extrusion and to optimize
the processing conditions to achieve a membrane having tunable transport properties and good
membrane performance in terms of permeability and selectivity [129].
2.2 Experimental investigation
2.2.1 Materials used
For this work, linear low density polyethylene powder LL 8460 from ExxonMobil (Canada) with
a melting temperature 126C, a melt index of 3.3 g/10 min (190C at 2.16 kg) and a density of
0.938 g/cm3 was used as the matrix. Activated azodicarbonamide (Celogen 754A, ChemPoint,
USA) was used as a chemical foaming agent. Based on the information from the provider, this
grade of azodicarbonamide decomposes between 165 and 180C, leading to a gas generation of
200 cm3/g of a mixture of N2, CO2, CO, and NH3.
2.2.2 Sample preparation and foaming process
LLDPE and azodicarbonamide were firstly dry-blended at different CBA concentrations (1, 1.5,
1.75 and 2.5% wt.). Then, chemical foaming was carried out via continuous extrusion on a
Leistritz ZSE-27 co-rotating twin-screw extruder (diameter = 27 mm and length/diameter = 40)
with a tubular die (outside and inside diameter of 5 mm and 3.5 mm, respectively) and 12
individually heated/cooled zones followed by a two-roll calendaring system (roll diameter and
width of 7 cm and 25 cm, respectively) to obtain the hollow fiber foamed LLDPE membrane. The
extruder was operated with different die temperature (165, 172, and 179C) with a screw
56
rotational speed of 30 rpm. A gravimetric feeder with a feeding rate of 17 g/min was used to feed
the materials. The extrusion system is presented in Figure 2.1 and the conditions tested in Table
2.1. For comparison, unfoamed samples under the same processing conditions as the foamed
samples were also prepared.
Table 2.1. Conditions tested for the foamed LLDPE hollow fibers.
Sample code
CBA
concentration
(wt.%)
Uniaxial
stretching speed
(m/min)
Die temperature
(C)
LL 0 - -
F1-5 1 5 172
F2-5 1.5 5 172
F3-5 1.75 5 172
F4-5 2.5 5 172
FC-7 1.75 7 172
FC-9 1.75 9 172
FD65 1.75 5 165
FD79 1.75 5 179
57
Figure 2.1. Effect of particle size on the permeability of different gases for PDMS-silicate
composite membranes [211].
2.2.3 Foam characterization
2.2.4 Morphology
Scanning electron microscopy (SEM) was performed on a JEOL JSM 840A to study the
morphology of the foamed samples. Firstly, the samples were cut using a sharp knife in two
perpendicular directions (transverse (T) and flow (F)) to get a full 3D analysis of the cells due to
deformation in the stretching direction. Secondly, the exposed surfaces were coated with a thin
palladium/gold layer to take images at different magnification. The images were analyzed using
the ImageJ software (National Institute of Health, USA) to report on cell density and cell size.
The reported values are the averages of at least three images in each direction. Because of cell
deformation, the cell density (N) was calculated as [222]:
𝑁 = 𝑁1(𝑁2)12⁄ (3.1)
where N1 and N2 are defined as the surface cell densities in the flow and transverse directions
respectively, and calculated as:
𝑁𝑖 =𝑛
𝐴 (3.2)
58
where A is the analyzed area and n is the number of cells.
Finally, foam density (f) was determined by a gas (nitrogen) pycnometer model Ultrapyc 1200e
(Quantachrome Instruments, USA). The reported results are based on a minimum of five
measurements. From the density, the volume expansion ratio (∅) is defined as:
∅ =𝜌
𝜌𝑓 (3.3)
where is the density of the neat (unfoamed) polymer.
2.2.5 Pure gas permeation
The gas permeation coefficient of all the samples was measured with the variable pressure and
constant volume method. As shown in Figure 2.2, the HFM was set securely in a permeation
hollow fiber module. Then, both sides of the feed and permeate module were evacuated
(vacuumed) for at least 6 h. A gas was loaded to the feed side at a pressure of 30 psi, while the
permeate gas on the other side of the membrane was stored in a closed chamber. The pressure
variation was recorded using a pressure transducer and the permeate pressure as a function of
time was used to measure permeability by the solution-diffusion model. The gas permeation
measurement at 30C with commercially important gases of different kinetic diameters (CO2 (3.3
Å), N2 (3.64 Å), O2 (3.46 Å), H2 (2.89 Å), and CH4 (3.8 Å)) through both the hollow fiber
unfoamed and foamed LLDPE membranes were performed. The reported values are based on the
average of five measurements and the permeance coefficient was determined as:
𝑄 = (𝑃
𝑙) =
22,414
𝐴×
𝑉
𝑅𝑇∆𝑝×
𝑑𝑝
𝑑𝑡 (3.4)
where Q is the permeance (GPU), P is permeability, l is membrane thickness, V is the constant
downstream volume (cm3), 𝑑𝑝
𝑑𝑡is the rate of permeation in the chamber under steady state (psi/s),
A is the membrane area (cm2), ∆𝑝 is the difference between the feed pressure and permeate
pressure (psi), T is the absolute temperature (K) and R is the universal gas constant. Based on the
individual gas permeance, their ratio is used to get the ideal selectivity as:
𝛼𝐴/𝐵 =𝑄𝐴
𝑄𝐵 (3.5)
59
Figure 2.2. Schematic representation of the hollow fiber membrane module set-up with an
actual image of a mounted sample.
2.2.6 Mechanical properties
The tensile property were measured on an Instron universal testing machine (model 5565, Instron,
USA) with a 500 N load cell. The tensile tests were carried out with a crosshead speed of 10
mm/min at ambient temperature (23C). The standard deviation and the average data of Young’s
modulus, tensile strength, and elongation at break are based on a set of 5 to 10 samples.
2.2.7 Thermal properties
The thermal properties (crystallinity and melting temperature) were determined by differential
scanning calorimetry (DSC) on a DSC 7 from Perkin Elmer (USA). Around 3 mg were
encapsulated in aluminum pans and a heating/cooling/heating cycle from 50 to 200C at a rate of
10C/min in a nitrogen atmosphere was applied. The degree of crystallinity (XC) as calculated by
dividing the experimental heat of fusion (∆Hexp) by the value for 100% crystalline LLDPE (H*
= 288 J/g, [155]) to give [226]:
𝑋𝐶 =∆𝐻𝑒𝑥𝑝
∆𝐻∗ (3.6)
60
2.3 Results and Discussion
2.3.1 Optimization of the foaming process
In general, foaming polyolefins is more difficult than most polymers such as polystyrene because
semi-crystalline polymers have a very narrow foaming temperature window [137]. Based on the
literature, a cellular structure having a high cell density about 107 to 109 cells/cm3, good
distribution of cell and small number of cells is considered as a good cellular structure [22]. The
influence of the processing parameters on the final HFM properties is presented here. Typical
morphology and results are presented next for the effect of the different parameters.
2.3.2 Effect of CBA content
Typical images in the flow (F) and transverse (T) directions of hollow fiber foamed membranes
for different CBA content are presented in Figure 3. It can be seen that the number and size of
cells vary with increasing CBA content. It is expected that more gas is released with increasing
CBA decomposition which is used for cell nucleation and growth [199],[227].
The effect of CBA content on the average cell size in both directions is presented in Figure 2.4.
The average cell size increases with increasing CBA content reaching a maximum value of 105
and 165 µm at 1.75% CBA in the F and T direction, respectively. This condition also produced
the most uniform cell morphology by having the lowest standard deviation variation. At higher
CBA content, gas loss may occur leading to lower cell sizes (no further increase).
The cell density and expansion ratio are also reported as a function of CBA content in Figure 2.4.
It was found that with increasing CBA concentration, the cell density increases up to 1.75%, then
slowly decrease (2.5% CBA). The cell density increases significantly (33.3%) from 7x106 to
10.5x106 cells/cm3 for F1-5 to F3-5 by increasing the CBA content from 1% (F1-5) to 1.75% (F3-
5). A similar trend is observed for the expansion ratio: from 25% to 48%. Similar results were
reported for wood composite foams and high-density polyethylene [199],[228]. An optimum
CBA value has been reported to be related to gas loss and cell coalescence when the CBA content
is too high as seen at 2.5% CBA (F4-5) having a non-uniform cellular structure (Figure 2.3).
61
Figure 2.3. Micrographs of the samples produced with different chemical blowing agent
contents for images taken in the transverse (left column) and flow (right column) directions.
Sample codes are defined in Table 2.1.
62
Figure 2.4. Cell size, cell densities, expansion ratio, and aspect ratio of the samples produced
with different chemical blowing agent contents.
Mechanical analyses was performed on both the unfoamed and foamed hollow fibers. Figure 2.5
shows that, as expected, the mechanical properties decreases with increasing CBA content
[222],[229]. This result is related to the weight reduction as less material is available to sustain
the applied stresses. Overall, les lowest properties were obtained for F3-5 which has the optimum
CBA content (1.75%) as reported in Figure 2.4 (highest expansion ratio). Above this critical
concentration, the tensile properties increased while the standard deviations increased due to a
degraded morphology; i.e. non-uniform cellular structure of the sample F4-5. It is clear that a
close relationship exists between the tensile properties and the cellular structure.
63
Figure 2.5. Young’s modulus, tensile strength and elongation at break of the samples produced
with different chemical blowing agent contents.
2.3.3 Effect of the die temperature
Effect of foaming temperature (die temperature) on the foam structure can be seen in Figure 2.6
showing several micrographs. As expected, melt elasticity (strength) and viscosity are highly
related to the melt temperature, and both parameters are controlling the final foam structure in
terms of cell density and cell size, as well as cell geometry and deformation due to post-extrusion
stretching. But other properties of interest for foaming are also modified with temperature: surface
tension and gas solubility/diffusivity. These parameters also come into play when optimizing the
foaming process [230].
64
Figure 2.6. Micrographs of the samples produced with different die temperatures (sample codes
are defined in Table 2.1).
From the image analyses performed, cell density as a function of die temperature is presented in
Figure 2.7a. It can be seen that the cell density at 1 and 1.5% CBA slightly increased from 165 to
172C and then decreased when the temperature increased to 179C. This optimum in temperature
represents a balance low temperature where cell growth is very difficult (high matrix viscosity
165C 172C 179C
F1-5
F2-5
F3-5
F4-5
FD65 FD79
65
and elasticity) and high temperature (gas loss and cell coalescence). In our case, the highest cell
density (10.5x106 cells/cm3) was obtained for an intermediate value of CBA content (1.75%) and
die temperature (172C) indicating that both parameters must be optimized for a specific
gas/matrix system [231]. Figure 2.7b presents the expansion ratio as a function of die temperature.
Again, the optimum CBA content is clearly 1.75% for the range of conditions studied. Figure
2.7c shows that the cell size might slightly decreased with increasing die temperature, but the
variation might not be significant due to large distributions on this parameter. In the literature
[24], it has been reported that cell sizes may decrease due to the limited gas solubility and high
diffusivity in the polymer at high temperature (gas loss) and the low melt strength/viscosity (cell
coalescence). This is in agreement here as increasing the die temperature from 165 to 179C led
to an extremely non-homogeneous open cell foamed structure along with cell
coalescence/collapse. This effect is mainly associated to the difficulty of stabilizing the cells after
their growth stage [152]. Similar trends have also been reported for foamed crosslink low-density
polyethylene [156], [166].
Figure 2.7. Effect of die temperature on cell density, average cell size and expansion ratio.
Sample codes are defined in Table 2.1.
66
Figure 2.8. Optimum foaming window for the cellular LLDPE hollow fiber membrane under
the conditions studied.
As mentioned before, linear polyolefins have a very limited foaming condition window
[137],[232],[233]. Generally, foam processing is much easier and stable when a wide range of
conditions are possible, while a narrow window needs good process control to be stable [233].
Based on our results, Figure 2.8 presents an approximation for the processing window. This range
of conditions will lead to low density, high cell densities and uniform cell distributions. So it is
estimated that the die temperature must be 1724C. Consequently, a broader range of 7C was
selected to include FD65 at 165C and FD79 at 179C for optimization.
Figure 2.9 shows the effect of die temperature on the tensile properties in terms of Young 's
modulus, tensile strength, and elongation at break. In general, the properties are following the
same trends as the expansion ratio. The lowest modulus and strength were obtained for F3-5
(172C and 1.75% CBA) which can be attributed to a reduced amount of material able to sustain
the applied stresses (Figure 2.7). But elongation at break is more related to the foam structure; i.e.
the defects inside the foams as open cells due to coalescence and this is why SD79 gives le lowest
value. As observed in (Figure 2.6), cell wall thickness must also be optimized for this property.
67
Figure 2.9. Effect of the die temperature on Young’s modulus, tensile strength and elongation at
break of the samples. Sample codes are defined in Table 2.1.
2.3.4 Stretching speed
Typical SEM micrographs in the transverse (T) and flow (F) directions for the samples produced
with 1.75% CBA at a constant die temperature of 172C are presented in Figure 2.9 for different
stretching speeds (5, 7 and 9 m/min). As expected, increasing the stretching speed from FC-5 to
FC-7 resulted in more elongated cells in the F-direction. Nevertheless, further increase in
stretching speed (9 min/m for FC-9) led to some cell wall ruptures and a non-uniform cellular
structure.
Based on the image analysis, general trends can be observed in Figure 2.10a. Firstly, increasing
the stretching speed resulted in lower cell density. For example, at the lowest stretching speed,
F3-5 produced the highest cell density (10.5x106 cells/cm3) which decreased (5.8x106 cells/cm3)
with increasing stretching speed (9 m/min for FC-9). This is mostly related to the fact that
68
stretching the samples gives the same amount of cells, but they have a larger area (volume) due
to their deformation as reported elsewhere [152]. Secondly, the expansion ratio slightly increased
with stretching speed as reported in Figure 2.10a. On average, the values are about 5% higher.
On the other hand, the cell aspect ratios did not changed much with increasing stretching speed
from F3-5 to FC-9 as shown in Figure 2.10b.
Figure 2.10. Micrographs of the HFM produced using different stretching speeds. T direction
(top row) and F direction (bottom row). Sample codes are defined in Table 2.1.
F3-5_T
F3-5_F
FC-7_T FC-9_T
FC-7_F FC-9_F
69
Figure 2.11. Effect of stretching speed on cell density, expansion ratio and aspect ratio with
1.75% CBA. Sample codes are defined in Table 2.1.
The effect of stretching speed on the mechanical properties are presented in Figure 2.11. It can
be seen that increasing the stretching speed led to higher Young’s modulus and tensile strength.
On the other hand, the elongation at break decreased from F3-5 to FC-9. There trends are typical
for linear polymers, especially related to chain orientation [137]. As a result, this also influences
the sample crystallinity as reported in Table 2.2. It can be seen that crystallinity increased from
37.2% for F3-5 to 39.6% for FC-9. Usually, lower crystallinity results in higher elongation at
break [234],[235]. This is the case here as the lowest stretching rate produced the lowest
crystallinity (Table 2.2) but the highest elongation at break (Figure 2.12).
Table 2.2. DSC results for the effect of stretching speed at a constant die temperature of 172C.
Sample
CBA
concentration
(wt.%)
Uniaxial stretching
speed (m/min) ∆H (J/g) XC (%)
F3-5 1.75 5 107.4 37.2
FC-7 1.75 7 110.3 38.8
FC-9 1.75 9 114.0 39.6
70
Figure 2.12. Effect of stretching speed on Young’s modulus, tensile strength and elongation at
break of the HFM (sample codes are defined in Table 2.1).
2.3.5 Gas permeation through hollow fiber foamed LLDPE membrane
The gas transport properties of PE (LLDPE, LDPE) as a semi-crystalline polymer are mainly
related to its crystallinity. Polyethylene crystals can act as gas barriers thus increasing the
tortuosity (diffusion path) of the penetrant gas molecules. Increasing crystallinity should lower
permeability, especially for larger molecules. This is related to polymer chains mobility
restriction by the crystallites [185], but a similar effect is possible for physical/chemical
crosslinks.
The effect of stretching speed on the permeance of the optimized HFM (172C and 1.75% CBA)
is presented in Figure 2.13a. The higher permeance of CO2 comes from its high condensability
increasing its solubility in LLDPE. It can also be seen that the gas permeance was higher than for
unfoamed LLDPE for all the gases. H2 with the smallest molecule size (2.98 Å) has the highest
permeance compared to the other gases (larger molecules). The H2 permeance increases by 70%
71
for sample F3-5 compared to neat LLDPE. This behavior is associated to the large amount of
voids inside the foams. But when the stretching speed increased, the permeance of FC-5 and FC-
9 slightly decreased probably due to their higher crystallinity (Table 2.2).
The ideal selectivity for different gas pairs (CO2/CH4, O2/N2, H2/N2, and H2/CH4) are presented
in Figure 2.13b. In most cases, the selectivity was slightly improved. However, H2/CH4 was only
significant increase (47%) between the three samples at different stretching speed (F3-5, FC-7,
and FC-9). This selectivity improvement is caused by the relatively higher increase of H2
permeance compared to CH4.
So, the development of tunable transport properties of a polymer membrane can lead to improved
membranes performance. This can be done by simultaneously increasing both surface area
activity and gas transport properties [82]. To achieve the best foam structure, process optimization
must be done to maximize the cell density and minimize cell size (more surface area per unit
volume). This optimization must include blowing agent content, die temperature, and stretching
speed, but other properties can be important even for a fixed system (polymer, blowing agent and
processing set-up). Based on the results obtained, sample F3-5 was determined as the optimum
which was processed using 1.75% CBA with a stretching speed 5 m/min and a die temperature
of 172C. Nevertheless, work is needed to improve on the gas transport properties of these
materials.
Figure 2.13. Results for gas (CO2, N2, O2, H2, CH4) permeance (a) and selectivity (b) at 30C
and 30 psia for unfoamed and foamed LLDPE hollow fiber membranes produced at different
stretching speed (sample codes are defined in Table 2.1).
72
2.4 Conclusion
In this study, a continuous processing method was developed to prepare a porous HFM using melt
extrusion. The system was shown to work using Azo (CBA) and LLDPE (as the matrix). The
method is believed to be cost-effective, fast, simple, and no solvent is used. However, since the
system is complex, the process and formulation must be optimized.
Based on the investigation performed, a foamed hollow fiber was optimized having a cell density
of 10.5 106 cells/cm3, and an average cell size of 105 and 165 mm in the F and T directions,
respectively, with an expansion ratio of 1.8 was produced. As reported several times in the
literature, the grade of LLDPE used has a very narrow foaming temperature window of 172+4C.
Increasing the foaming temperature above this range led to cell opening and nonuniform
distribution, while below this range limited cell nucleation and growth occurred. For the blowing
agent content, the optimum was 1.75 wt.% as lower amount did not achieve full blowing potential
(not enough gas generated), while higher content led to oversaturation and gas loss. For the
stretching speed, 5 m/min was found to be the optimum since lower speed did not stretch much
of the samples, while higher velocity led to sample breakup (cell wall rupture). The tensile
mechanical properties (Young’s modulus, tensile strength, and elongation at break) were
reported. The results indicated a clear relation between the foam structure and the macroscopic
behavior of the materials because the properties were direct functions of the cell size, cell density,
and matrix crystallinity.
The gas transport properties of the hollow fibers were determined. The results showed again a
clear relation between the cellular structure and the permeance/selectivity values. Because
LLDPE is a semi crystalline polyolefin, crystallinity increase was found to decrease permeance
but increase selectivity, especially for H2 and H2/CH4.
Finally, it can be concluded that foamed hollow fibers based on LLDPE can be considered as a
low-cost and an easily processable membrane for gas separation applications, especially for the
H2 purification industry. However, more work is needed to further improve these results and
scale-up the process for industrial production.
73
Acknowledgments
Financial support from the Natural Sciences and Engineering Research Council of Canada
(NSERC) was received for this work. Samples were kindly provided by Exxon Mobil (LLDPE).
The technical assistance of Mr. Yann Giroux is highly appreciated.
74
3 Chapter 3
Gas transport properties of cellular hollow fiber
membranes based on LLDPE/LDPE blends
Zahir Razzaz, Abolfazl Mohebbi, Denis Rodrigue, Cellular Polymers, submitted (2018).
75
Résumé
La production de membranes en forme de fibres creuses expansées (HFM) est présentée à base
des mélanges de polymères utilisant différentes concentrations de polyéthylène linéaire de basse
densité (LLDPE) et de polyéthylène basse densité (LPDE) associé à de l'azodicarbonamide (agent
d'expansion chimique) pour la préparation d'échantillons par extrusion à double vis. En
particulier, la concentration en agent gonflant ainsi que la vitesse d’étirage sont les paramètres les
plus importants pour obtenir une bonne structure cellulaire pour l’application de membranes. À
partir des échantillons obtenus, un ensemble complet de caractérisations morphologiques,
thermiques et de transport de gaz a été réalisé. Les résultats montrent que les mélanges
LLDPE/LDPE comparés au LLDPE seul conduisent à une densité cellulaire plus élevée, surtout
à une vitesse d'étirement élevée, ce qui convient aux membranes ayant une perméabilité aux gaz
et une sélectivité plus élevée en raison de l'épaisseur plus faible de la paroi cellulaire. Les résultats
montrent également que la structure cellulaire développée offre un potentiel élevé pour la
production continue de membranes à fibres creuses pour différentes séparations de gaz, en
particulier pour la récupération d'hydrogène.
76
Abstract
The production of foamed hollow fiber membranes (HFM) is presented based on polymer blends
using various concentrations of linear low-density polyethylene (LLDPE) and low density
polyethylene (LPDE) combined with azodicarbonamide (chemical blowing agent) to prepare
samples via twin-screw extrusion. In particular, the blowing agent concentration as well as the
stretching speed were found to be the most important parameters to achieve a good cellular
structure for membrane application. From the samples obtained, a complete set of morphological,
thermal and gas transport characterization was performed. The results show that LLDPE/LDPE
blends compared to neat LLDPE lead to higher cell density at high stretching speed, which is
appropriate for membranes having higher gas permeability and selectivity due to lower cell wall
thickness. The results also show that the developed cellular structure has high potential for the
continuous production of hollow fiber membranes for different gas separation, especially for
hydrogen recovery.
Keywords: polyethylene, blends, extrusion, cellular structure, gas separation.
77
3.1 Introduction Investigations on the cellular structure of polymer foams is a well-known and important topic in
material science because of their unique relation between morphology and multi-functionality for
a wide range of applications [236]. Thermoplastic foams are cellular materials containing gaseous
bubbles as dispersed phases surrounded by a polymer matrix as a continuous phase [22],[144].
The cellular structure can be obtained using chemical or physical foaming agents. Nevertheless,
chemical blowing agents, compared to physical blowing agents, are easier to handle [82],[237].
Microcellular foams have widespread uses for automotive, insulation, biotechnology, food
packaging, renewable energies, and separation [134],[164]. This wide range of applications is
related to their high impact strength, high stiffness to weight ratio, lightweight, high fatigue
resistance, high thermal insulation, low processing and material costs, and easy processing. In our
pervious study, polyethylene foams have also been proposed as membranes for gas separation
[19].
Polymeric membranes for gas separation are commercially attractive due to their low operation
and material costs, easy processability, and low energy consumption [19]. But very few studies
were devoted to control the morphological structure of these membranes in terms of material
selection and cell morphology to optimize their gas transport properties [27],[213]. Nowadays,
according to the final membrane applications, different structures are available like symmetric
and asymmetric porous polymer membranes, but most of these methods, such as solvent casting
[27] or phase inversion [212], are using toxic and expensive solvents which are non-
environmentally friendly and must eventually be removed/recycled/treated. Nevertheless, a few
non-solvent techniques were proposed to prepare porous polymer membranes having a
microcellular structure [213],[238]. Melt-processing can be done by stretching and melt-spinning
of semi-crystalline thermoplastics such as polypropylene [239]. But this method requires several
mechanical stretching and thermal post-treatment to stabilize the crystalline and porous structure
to prevent membrane shrinkage [240]. Another non-solvent technique is to use leachable particles
(salts) to prepare a composite leading to an open-cell structure. This method was found to work
well with low density polyethylene [122], polystyrene [121], and polypropylene [120]. In our
previous work, a melt processing method was developed to produce hollow fiber membranes
(HFM) based on microcellular linear low-density polyethylene (LLDPE). In this work,
continuous extrusion is used to optimize the cellular structure and the overall properties of these
78
membranes by adding low density polyethylene (LDPE) in the formulation [19]. Each method
has its own benefits and drawbacks. As presented in Figure 3.1, several important specifications
must be considered to determine the suitability of membrane preparation.
Figure 3.1. The most important specifications related to membrane preparation methods.
In our previous works, it was shown that the morphological properties, such as cell density and
cell size uniformity, are controlling the mechanical, physical, and gas transport properties of
polymer membranes [19],[27],[26]. Since these properties are tunable, this led to membrane
performance improvement via simultaneous increase in both the gas transport properties and
surface area [19],[82]. But foam processing of semi-crystalline polymers, such as polyethylene
and polypropylene (PP), is more challenging compared to amorphous resins because they have a
very narrow foaming temperature window [22],[134]. On the other hand, different grades of
polyethylene including high density polyethylene (HDPE), low density polyethylene and linear
low density polyethylene [237] are highly produced and consumed due to their excellent overall
performance, high impact strength and ductility, chemical resistance, low cost, and good
processability. However, the crystalline regions can act as nucleation sites for foaming but behave
as local gas barriers [164], which is important for membrane application [241]. Therefore
79
membranes based on LLDPE with relatively low crystallinity have been introduced [219]. In our
previous work, it was shown that foamed hollow fiber membranes with good gas permeability
were possible to produce. Hence, the main objective of this study is to further improve on the
cellular structure of these membranes to increase their performance. Although several methods
are available to improve the foamability of a polymer [22],[236], it was found that polymer
blending is a very easy and effective strategy to enhance the cellular structure of polymer foams
in terms of cell density and cell size uniformity. This approach has been successfully applied for
polyolefin blends [146],[242],[167], especially for PE and PP blends [243],[236]. For example,
Sahagun et al. [146] used a three-dimensional analysis to study the interactions between PP and
HDPE over the complete concentration range (0:100 to 100:0). Foaming was performed via flat
sheet extrusion using a chemical blowing agent (azodicarbonamide). They showed that with the
addition of a suitable amount (10%) of compatibilizer (Kraton D1102), the cell morphology of
PP/HDPE foamed blends was improved due to better interfacial adhesion and smaller sizes of the
dispersed phase.
Therefore, to achieve our objective, a design of experiments was performed based on
LLDPE/LDPE blends. The addition of LDPE to LLDPE is believed to improve processability
and foamability of the LLDPE matrix used for its better mechanical properties without having
incompatibility problems between both polymers [244]. It is known that LDPE, due to its highly
branched structure, has high elongational viscosity improving bubble stability and controlling
bubble growth with respect to LLDPE foaming [245]. Hence, the main purpose of this study is to
achieve a hollow fiber membrane based on foamed LLDPE/LDPE having a good cellular
structure for gas separation applications. The samples are produced via foam extrusion (a
nonsolvent and continuous method) and a focus is made on the interaction between the extrusion
and post-extrusion (stretching ratio) conditions with respect to the amount of LDPE added. The
membrane performances are finally compared in terms of gas transport properties (permeability
and selectivity) [129].
80
3.2 Experimental
3.2.1 Materials
Linear low-density polyethylene powder LLDPE 8460 with a MFI of 3.3 g/10 min (190C at 2.16
kg), a melting temperature of 126C and a density of 0.938 g/cm3 was supplied by ExxonMobil
(Canada). The low density polyethylene (LDPE) used was LA-0219-A with a density of 0.919
g/cm3 and a MFI of 2.3 g/10 min (190C at 2.16 kg) provided by Nova Chemicals (Canada).
Activated azodicarbonamide, Celogen 754A from ChemPoint (USA), was used as a chemical
blowing agent (CBA) with a decomposition temperature of 165-180C and a gas generation of
200 cm3/g (mixture of NH3, N2, CO2 and CO).
3.2.2 Sample preparation
LLDPE and LDPE were first dry-blended with different LDPE contents (30 and 40% wt.) and an
optimum CBA concentration (1.75% wt.) was used. More detail for the optimum CBA content
can be found in our previous study [19], while the LDPE range selected was based on preliminary
runs. Sample foaming was performed on a co-rotating twin-screw extruder (Leistritz ZSE-27)
(length/diameter (L/D) = 40 and diameter = 27 mm) with 10 individually heated/cooled zones. A
tubular die with an outside diameter of 5 mm and inside diameter of 3.5 mm was installed at the
extruder’s end and coupled with a two-roll calendaring system (roll width = 25 cm and roll
diameter = 7 cm) to control stretching (post-extrusion). The extruder was run with a screw speed
of 30 rpm. The experimental set-up is presented in Figure 3.2 with the processing conditions
reported in Table 3.1. Unfoamed (solid) samples were also prepared for reference using the same
processing conditions as for their foamed counterparts.
81
Table 3.1. Specifications of the hollow fibers produced*.
Sample code
LLDPE
(wt.%)
LDPE
(wt.%) CBA
concentration
(wt.%)
Uniaxial
stretching
speed (m/min)
LLDPE 100 0 0 -
LL70 70 30 0 -
LL60 60 40 0 -
HFM 100 0 1.75 5
HFMB7-5 70 30 1.75 5
HFMB7-7 70 30 1.75 7
HFMB7-9 70 30 1.75 9
HFMB7-11 70 30 1.75 11
HFMB7-13 70 30 1.75 13
HFMB6-5 60 40 1.75 5
HFMB6-7 60 40 1.75 7
HFMB6-9 60 40 1.75 9
HFMB6-11 60 40 1.75 11
HFMB6-13 60 40 1.75 13
* HFMB7 and HFMB6 stand for the foamed samples with a composition of LLDPE/LDPE
70/30 and 60/30, respectively.
Figure 3.2. Schematic representation of the experimental set-up for the continuous production
of hollow fiber foamed membranes.
82
3.2.3 Characterization
3.2.4 Thermal properties
The thermal properties were determined in terms of crystallinity and melting temperature via
differential scanning calorimetry (DSC) on a DSC 7 (Perkin Elmer, USA). The experiments were
performed with 5 mg of sample under a nitrogen atmosphere with a rate of 10C/min of
heating/cooling/heating cycle from 50 to 200C. The crystallinity (XC) was calculated by
comparing the experimental heat of fusion (∆Hexp) with that of fully crystalline polyethylene
( H*) (288 J/g [155]) as [226]:
𝑋𝐶 =∆𝐻𝑒𝑥𝑝
∆𝐻∗ (4.1)
3.2.5 Morphological analysis
The morphological characterization was performed using scanning electron microscopy (SEM)
(JEOL JSM 840A). To obtain a complete 3D analysis of the deformed cellular structure resulting
from uniaxial stretching (post-extrusion), the samples were cut by using a sharp knife in both
flow (F) and transverse (T) directions. Then, a thin gold/palladium coating was applied on the
exposed surfaces. Finally, the images were characterized using the ImageJ software (National
Institute of Health, USA) to measure cell density and cell sizes [19]. The values reported are the
averages of at least three measurements for each direction.
Foam density (f) was measured using a gas pycnometer (Ultrapyc 1200e, Quantachrome
Instruments, USA). The reported values are the average of at least five measurements for each
sample. Then, the volume expansion ratio (∅) was determined as:
∅ =𝜌
𝜌𝑓 (4.2)
where is the density of the unfoamed (solid) samples.
3.2.5.1 Cell size distribution
In addition to the cell density, cell size distribution is another important parameter affecting the
macroscopic properties and membrane performance of porous materials. A good membrane
performance can be achieved by a more homogenous cellular structure with a narrower cell size
83
distributions [246]. Moreover, the standard deviation of the cell size distribution (SD) for each
samples was reported to better understanding the cellular structure homogeneity for different
samples [246],[240], which is calculated as:
𝑆𝐷 = √∑(𝜑𝑖−𝜑)
2
𝑛𝑛𝑖=1 (4.3)
where 𝜑 is the average diameter of the cells, 𝜑𝑖 is the individual cell diameter and n is the total
number of cells analyzed.
3.2.6 Gas permeation
The gas permeation coefficient was measured for all the membranes using the constant-volume
and variable-pressure technique. The samples were securely fixed inside a permeation hollow
fiber module. Then, the module was evacuated from both sides (permeate and feed) of the module
for at least 6 h. Then, a gas was loaded from the feed-side of the system at a pressure of 2 bar,
while on the other side of the membrane the permeate gas was kept in a constant volume/closed
chamber. An electronic pressure transducer recorded the pressure variation as a function of time.
The results were used to determine the permeability coefficient using the solution-diffusion
model. The measurement were performed using commercially important gases having different
kinetic diameters including H2 (2.89 Å), CO2 (3.30 Å), O2 (3.46 Å), N2 (3.64 Å) and CH4 (3.80
Å). All the experiments were performed at 30C. The results are based on the average of at least
five measurements for the permeance coefficient determined as:
𝑄 = (𝑃
𝑙) =
22,414
𝐴×
𝑉
𝑅𝑇∆𝑝×
𝑑𝑝
𝑑𝑡 (4.4)
where Q is the permeance (GPU), P and l are the permeability and membrane thickness (cm),
respectively. A is the membrane area (cm2), 𝑑𝑝
𝑑𝑡is the rate of permeation in the closed chamber
under steady state condition (psi/s), V is the constant downstream volume (cm3), ∆𝑝 is the
difference between the upstream pressure and downstream pressure (psi), R is the universal gas
constant (6236.56 cm3 cmHg/mol K) and T is the absolute temperature (K). The ideal selectivity
(𝛼𝐴/𝐵) for gases A and B can be calculated from their individual gas permeance as:
𝛼𝐴/𝐵 =𝑄𝐴
𝑄𝐵 (4.5)
84
3.2.7 Mechanical properties
The tensile properties were characterized using an Instron universal testing machine (model 5565,
Instron, USA) with a crosshead speed of 10 mm/min. The tensile tests were performed with a 500
N load cell at ambient temperature. The average results for the Young’s modulus, tensile strength
and elongation at break were determined with standard deviations based on several (5 to 10)
samples.
3.3 Results and discussion
3.3.1 Optimization of the foaming process
Foaming of polyolefins, such as LLDPE, is more challenging compared to other polymers (such
as polystyrene) because a very narrow foaming temperature window for these semi-crystalline
polymers exists [137],[241]. In our previous study, it was shown that a cellular structure having
a high cell density (1x107 cells/cm3) and a uniform cell size distribution can enhance membrane
performance. In the current work, blending LLDPE with LDPE was performed to improve these
properties. Multiphase polymer materials, compared to neat polymers, can have different foaming
behavior because of the presence of interfacial regions. Moreover, melt strength is one of the most
important parameters in foaming polymers. It is known that foaming of a polymer having a low
melt strength leads to a poor cell morphology and high cell coalescence/rupture due to weak cell
walls elasticity under biaxial deformation around the cells [155]. Hence, different LDPE contents,
having a high melt strength and a close structural/chemical structure to LLDPE [247], was added
to sustain part of these high extensional deformation and deformation rate [248], to improve the
LLDPE cellular structure for better membrane performance.
Typical SEM pictures of the foamed LLDPE/LDPE blends of 100/0 (HFM), 70/30 (HFMB7),
and 60/40 (HFMB6) with a stretching speed of 5 m/min are presented in Figure 3.3. It can be seen
that the cell uniformity and cell size were slightly improved by increasing LDPE content from 0
to 40% wt. Figure 3.4a presents the cell density as a function of LDPE content for a stretching
speed of 5 m/min and a die temperature of 172C. The results show that the cell density and
expansion ratio were both increased from HFM to HFMB6, while limited improvement was
observed between HFMB7 and HFMB6. The effect of LDPE content on the average cell size in
both F and T directions is reported in Figure 3.4b. It was found that increasing the LDPE content
85
led to negligible decrease in cell size and similar results have been reported for
LLDPE/polystyrene (PS) blends [206],[238].
Figure 3.3. Micrographs of the foamed LLDPE/LDPE blends of 100/0 (HFM), 70/30 (HFMB7-
5), and 60/40 (HFMB6-5) with a stretching speed of 5 m/min (see Table 3.1 for sample details).
Figure 3.4. Cell density, expansion ratio, and average cell sizes in the flow (F) and transverse
(T) directions of the samples with different LDPE contents in LLDPE/LDPE blends at a
stretching speed of 5 m/min (see Table 3.1 for sample details).
3.3.2 Effect of stretching speed
The foaming behavior of LLDPE as well as two LLDPE/LDPE blends (70/30 and 60/40) with
1.75% CBA and different stretching speeds (5, 7, 9, 11 and 13 m/min) are presented in Figure
3.5. As expected, increasing the stretching speed led to more elongated cells in the F direction for
HFM, HFMB7 and HFMB6. Accordingly, the cell aspect ratio (longer axis divided by the shorter
one) increased at higher stretching speed as presented in Figure 3.6a. As shown in Figure 3.6b,
increasing the stretching speed for neat LLDPE (HFM) led to lower cell density from 1.05x107
86
to 0.58 x107 cells/cm3. Moreover, increasing the stretching speed above 9 m/min resulted in cell
wall rupture and cell collapse. But the results show that adding LDPE led to higher cell density
by 62% and 66% for HFMB7-11 and HFMB6-11, respectively. Moreover, as shown in Figure
3.6c, a similar trend was found for the expansion ratio. This uniaxial stretching also decreased the
samples thickness by about 50%, from 560 µm (HFM) to 265 µm (HFMB7-11) and 160 µm
(HFMB7-13), but the latter presented a highly non-uniform morphology (see Figure 3.5d) due to
the high stretching speed (13 m/min).
Furthermore, blending LLDPE with LDPE enabled to produce foamed hollow fibers at high
stretching velocity with more uniform and higher cell density leading to lower film thickness.
This improvement can be related to the fact that stretching the blends provided about the same
number of cells but in a much smaller area (volume) by also preserving the structure uniformity.
For instance, at the lowest stretching speed, HFMB7-5 had the lowest cell density (1.2 x107
cells/cm3) which increased for HFMB7-11 (1.7 x107 cells/cm3) and then slightly decreased for
HFMB7-13. Nevertheless, there is a limit as further increase (HFMB6-13) for both LLDPE/LDPE
composition led to cell wall rupture and break-up losing the foam structure uniformity.
87
Figure 3.5. SEM micrographs in the T (left column) and F (right column) direction for different
stretching speeds. Sample codes are defined in Table 3.1.
88
Figure 3.6. Effect of stretching speed on: a) cell density, b) cell aspect ratio, c) expansion ratio
and d) sample thickness for samples produced with 1.75% CBA. Sample codes are defined in
Table 3.1.
The size distribution uniformity at different stretching speed is presented in Figure 3.7 in terms
of the cell size distribution standard deviation (SD). It is interesting that a more homogeneous
cellular structure (lower SD) was achieved for the blends, but a very similar cell size distribution
was obtained for HFMB7 and HFMB6. The SD values for HFMB7 and HFMB6 decreased with
increasing stretching speed, while a reverse trend was observed for HFM. Once again, the
LLDPE/LDPE blends are better than neat LLDPE leading to a better cellular structure with
increasing stretching speed up to an optimum (13 m/min) limited by the matrix melt strength.
89
Figure 3.7. Standard deviation (SD) of the cell size distribution in the F and T direction as a
function of stretching speed for 5, 7, 9, 11, and 13 m/min. Sample codes are defined in Table
3.1.
3.3.3 Thermal properties
Figure 3.8 shows the DSC results for the first heating of neat LLDPE, LDPE and their blends.
For the latter, a single peak is obtained due to good compatibility between both polymers which
is shifted to higher temperature with increasing LDPE content and similar results have been
reported elsewhere [247],[248],[249],[250]. Form these curves, the main parameters are reported
in Table 4.2 [251],[234].
90
Figure 3.8. DSC curves of LLDPE, LDPE and their blends. The heating rate is 10C/min.
Table 4.2 also presents the difference between the melting onset and endpoint temperature in DSC
which is indicative of the melting peak width. This difference for LLDPE, LL70, LL60 and LDPE
was 7, 4, 5 and 8 oC, respectively. Hence, the melting peak is narrower for the blends (LL70 and
LL60) which may be associated to a narrower crystal size distribution compared to the neat
polymers (LLDPE and LDPE) [252]. Furthermore, for the LL60 sample, having a higher LDPE
content compared to LL70, a higher heat of fusion and crystallinity were obtained. Finally, the
crystallization temperature of LLDPE is lower than that of the LDPE. So the LDPE crystalline
phase can act as nucleating sites for LLDPE crystallization and cell nucleation [253]. A summary
of the results is presented in Table 3.2.
91
Table 3.2. Heat of fusion and melting temperatures for the neat polymers and their blends.
Material
Melting onset
temperature
(C)
Melting end
temperature
(C)
Melting
temperature range
(C)
Melting peak
temperature
(C)
Heat of
fusion
(J/g)
LLDPE 104 111 7 107 106
LL70 108 112 4 111 110
LL60 109 114 5 112 112
LDPE 112 120 8 116 125
The effect of the stretching speed on the sample crystallinity is presented in Figure 3.9. It can be
seen that increasing the stretching speed led to lower crystallinity for HFMB7 and HFMB6
blends. This is mostly related to the fact that the long chain branches of LDPE prevented
recrystallization during stretching of the blends [253],[254]. This phenomenon is schematically
presented in Figure 3.10. Since this is not the case for LLDPE (HFM), the crystallinity increased
with increasing stretching because the molecules are more aligned in the machine direction
leading to higher chain mobility and organization to create and grow crystals.
Figure 3.9. Degree of crystallinity as a function of stretching speed for the foamed hollow fiber
membranes based on LLDPE and its blend with LDPE. Sample codes are defined in Table 3.1.
92
Figure 3.10. Schematic presentation of the formation mechanism of the crystalline and
amorphous structure in LLDPE and its blends with LDPE during stretching.
3.3.4 Gas transport properties
The crystalline regions of LLDPE and LDPE, as semi-crystalline polymers, can also play an
important role in the gas transport properties or polymer membranes as they are generally
accepted as being impermeable,[241], [185]. So, these crystalline regions can be considered as
gas barrier phases (negligible diffusion coefficient and solubility) compared to the amorphous
regions. Hence, their presence increases the tortuosity (diffusion path) of the gas molecules
(permeates). As a result, the overall permeability should decrease with increasing crystallinity,
especially for larger molecules [255],[256]. This can be described by the limitation of polymer
chains mobility in the crystalline regions. However, the same effect is expected when
chemical/physical crosslinks are present [19],[185].
The separation permeance of the foamed hollow fiber LLDPE membranes and the blends are
presented in Figure 3.11. The results showed that the permeance of the foamed samples slightly
decreased with increasing LDPE content (100/0 to 60/40) for all the gases. This is due to increased
crystallinity with increasing LDPE content (Figure 3.9), while there was a negligible difference
in cell density for HFM, HFMB7 and HFMB6 (Figure 3.6b). But the gas permeance of the foamed
samples was much higher than for the unfoamed samples for all the gases studied (CO2, N2, O2,
H2 and CH4). The higher CO2 permeance compared to N2, O2, and CH4 was related to its high
solubility in PE (higher condensability), while the higher permeance for H2 is related to its small
molecular size (2.98 Å) compared to the other gas molecules: CO2 (3.30 Å), O2 (3.46 Å), N2 (3.64
Å) and CH4 (3.80 Å). The hydrogen permeance increased by about 70% for HFM membrane
93
compared to the unfoamed LLDPE membrane due to the large amount of existing voids of the
cellular structure.
Figure 3.11. Permeance of different gases (CH4, CO2, N2, O2 and H2) at 30C and 30 psia for
unfoamed and foamed hollow fiber LLDPE membranes and the blends produced at a stretching
speed of 5 m/min (sample codes are provided in Table 3.1).
The influence of the stretching speed on H2 permeance is shown in Figure 3.12a. Increasing the
stretching speed led to higher H2 permeance for HFMB7 and HFMB6, but lower H2 permeance
for HFM. This trend is directly related to the morphology of these samples since increasing the
stretching speed led to increased cell density and decreased crystallinity for HFMB7 and HFMB6,
while a reverse effect on cell density and crystallinity for HFM membranes was observed as
presented in Figure 3.6b and Figure 3.9. In the literature [257], the same trend was observed for
neat LLDPE (unfoamed) films as permeability decreased with increasing stretching rate. It was
found that the H2 permeance of HFMB7 increased with increasing the stretching rate and the
highest permeance was achieved for HFMB7-11 (stretching speed of 11 m/min). This permeance
improvement for HFMB7-11 was about twice the value for unfoamed LLDPE, while HFMB6
94
showed a similar trend. However, the H2 permeances for HFMB6 were lower than that of HFMB7
for all stretching speed. The reason is related to the higher crystallinity degree of the former with
negligible cell density change between both membranes as obtained for a wide range of stretching
speed (5, 7, 9, 11 and 13 m/min in Figure 3.6b). For HFMB7-13 and HFMB6-13, the standard
deviation was higher due to a degraded morphology at high stretching speed (13 m/min) leading
to a non-uniform cellular structure (see Figure 3.5 and Figure 3.6b).
Figure 3.12. Results for: (a) H2 gas permeance and (b) H2/CH4 ideal selectivity at 30C and 30
psia for foamed hollow fiber LLDPE membranes and its blends with LDPE produced at
different stretching speed (sample codes are presented in Table 3.1).
The H2/CH4 ideal selectivity of HFM, HMFB7 and HFMB6 membranes with different stretching
speed are presented in Figure 3.12b. Negligible variation was observed with increasing LDPE
content (HFM, HFMB7-5 and HFMB6-5) at constant stretching speed (5 m/min). However, the
selectivity of HFMB7 and HFMB6 was slightly increased with increasing stretching speed up to
the optimum value of 11 m/min. The highest selectivity of 5.8 was obtained for HFMB7-11 which
is about 75% higher than the value for unfoamed LLDPE. This selectivity improvement is mainly
due to higher H2 permeance compared to CH4 permeance.
As shown in Figure 3.13, the crystallinity of HFMB-11 and unfoamed LLDPE are similar, while
the selectivity and permeance were increased from the unfoamed to the foamed (LLDPE to
HFMB7-11) membranes. It can be inferred that a direct relationship between the cell density and
selectivity exists. Therefore, CH4 molecules compared with H2 molecules, were mostly trapped
95
inside the cells and moved more slowly than smaller molecules (Figure 3.14). It was also found
that the permeance is a function of cell density and crystallinity.
Figure 3.13. Results for H2 gas permeance, H2/CH4 selectivity (at 30C and 30 psia) and
crystallinity for hollow fibers based on unfoamed LLDPE and foamed 70/30 LLDPE/LDPE
(HFMB7-11).
Figure 3.14. Schematic presentation of H2 and CH4 molecular diffusion into the cellular
structure of a foamed membrane.
96
Therefore, improvement of the tunable transport properties of a polymer membrane results in
better membrane performance. This can be done by simultaneously improving the gas transport
properties as well as the surface area activity [82]. To achieve a good cellular structure, process
optimization should be carried out to maximize the cell density. This optimization consists of
stretching speed and optimum blend composition, although other parameters are essential to set
a system (die temperature, flow rate, etc.). According to the results obtained, HFMB7-11 was
identified as the optimum sample which is based on a 70/30 blend of LLDPE/LDPE with
1.75%wt. of CBA (chemical blowing agent) and a 11 m/min stretching speed. Nevertheless, more
study would be required to improve the gas transport properties of these materials. Therefore, cell
density and crystallinity are the main parameters to investigate the gas transport properties in
foamed semi-crystalline polymers.
Figure 3.15 presents the effect of testing temperature on the gas transport properties of the
optimum membrane (HFMB7-11). The results showed that the permeance increased with
increasing testing temperature for all gases. Increasing the testing temperature led to higher
diffusion coefficient for all gases [27]. Then, this diffusivity increase led to higher permeance
[257]. The gas permeance increased with increasing temperature in the following order: CO2 >
H2 > O2 > N2 > CH4. As shown in Figure 3.15, increasing rate of CO2 permeance with increasing
temperature was more important compered to H2, which is related to more important increase of
CO2 solubility at higher temperature [257],[197].
97
Figure 3.15. Permeance of CH4, CO2, N2, O2 and H2 through the HFMB-11 membrane
(optimum sample) as a function of testing temperature.
3.3.5 Mechanical properties of hollow fiber foamed LLDPE membrane and
its blends
The mechanical properties, including Young's modulus, tensile strength, and elongation at break
are plotted in Figure 3.16. In general, the foamed samples showed lower tensile modulus and
tensile strength compared to the unfoamed sample. This is related to less material (polymer
matrix) available to sustain the applied stresses. It can be observed that the modulus of HFMB6-
5 is slightly higher than HFMB7-5, and HFMB7-5 is slightly higher than HFM. This indicates
that the modulus increased with decreasing LLDPE content. This can be due to higher cell density
with decreasing LLDPE content, as presented in Figure 3.6b. A similar trend was observed for
the tensile strength. However, the elongation at break is more associated with the cellular structure
and crystallinity [234],[254],[235]. The elongation at break of HFMB7 and HFMB6 was higher
than HFM. This results from the presence of LDPE which is more elastic. The elongation at break
of HFMB6-5 (LLDPE/LDPE = 60/40) was slightly lower than HFMB7-5 (70/30), which is due
to a higher cell density of the former. For the unfoamed samples, the modulus, tensile strength
98
and elongation at break increased with LDPE addition. This is in agreement with previous reports
[236],[253].
Figure 3.16. Young’s modulus (a), tensile strength (b) and elongation at break (c) of the foamed
and unfoamed hollow fibers based on LLDPE and its blends with LDPE.
Figure 3.17 presents the effect of the stretching speed on the tensile properties of the optimum
hollow fiber membrane (HFMB7-11). It was found that the Young's modulus and tensile strength
of HFMB7-11 decreased by increasing the stretching speed. This is due to increase in cell density
and degree of crystallinity by increasing the stretching speed. Above a critical stretching speed
(11 m/min), the tensile properties were slightly increased while experimental error was increased
because of the degraded cell morphology. However, the elongation at break increased with
increasing stretching rate (Figure 3.17c). This is mostly related to the fact that the long chain
branches of LDPE prevents recrystallization during stretching of the blends [253],[254].
99
Figure 3.17. Effect of stretching speed on the Young’s modulus (a), tensile strength (b) and
elongation at break (c) of selected HFMB7 and LL70 (70/30) blends.
3.4 Conclusion
In this work, foamed hollow fiber membranes were produced and optimized based on a simple,
precise, cost-effective and rapid continuous process (extrusion) using a blend of LLDPE/LDPE
as the matrix and azodicarbonamide as a chemical blowing agent. The effect of blend content and
stretching speed were mainly investigated to optimize the foam structure.
The results showed that a stretching speed of 11 m/min was optimum for both LLDPE/LDPE
blends (70/30 and 60/40) leading to a cell density of around 1.7x107 cells/cm3. The results also
showed that the thermal (DSC) and mechanical (tension) properties were directly related to the
cellular structure (cell size and cell density), as well as the crystallinity of the matrix blend.
Finally, the gas permeation results showed that the transport properties inside semi-crystalline
polymers are a complex combination of the foam cellular morphology (cell sizes and cell density)
and the matrix structure (composition and crystallinity). The permeance of different gases (CH4,
100
CO2, N2, O2 and H2) was found to increase with decreasing foam density. In particular, hydrogen
permeance and H2/CH4 selectivity of the unfoamed LLDPE membranes and the optimum foamed
blend 70/30 (HFMB7-11) increased from 39 to 78 GPU (100%) and 3.3 to 5.8 (76%) respectively,
when a cellular structure was formed inside the PE membrane. However, further investigation in
under way to improve on these results and for potential industrial applications.
3.5 Acknowledgements
This work was financially supported by the Natural Sciences and Engineering Research Council
of Canada (NSERC) and the Centre de Recherche sur les matériaux avancés (CERMA). PE
samples were kindly provided by Exxon Mobil (LLDPE) and Nova Chemicals (LDPE). The
technical support of Mr. Yann Giroux is gratefully acknowledged.
101
4 Chapter 4
Hollow fiber porous nanocomposite membranes
produced via continuous extrusion: Morphology
and gas transport properties
Zahir Razzaz, Denis Rodrigue, Materials, 11(11), 2311 (2018).
DOI: 1 doi.org/10.3390/ma11112311 (2018).
102
Résumé
Dans ce travail, des membranes nanocomposites poreuses en forme de fibres creuses ont été
préparées avec succès par l’incorporation d’une nanoparticule poreuse (zéolite 5A) dans un
mélange de polyéthylène linéaire de basse densité (LLDPE)/polyéthylène basse densité (LDPE)
associé à de l’azodicarbonamide comme agent gonflant chimique (CBA). La mise en œuvre a été
effectuée par extrusion continue en utilisant une extrudeuse à double vis couplée à un système de
calandrage. Le procédé a d'abord été optimisé en termes de conditions d'extrusion et de post-
extrusion, ainsi que de formulation pour obtenir une bonne structure cellulaire (distribution
uniforme de la taille des cellules et densité cellulaire élevée). La microscopie électronique à
balayage (SEM) a été utilisée pour déterminer la structure cellulaire ainsi que la dispersion des
nanoparticules. Ensuite, les échantillons ont été caractérisés en termes de stabilité mécanique et
thermique via des tests de traction et une analyse thermogravimétrique (TGA), ainsi que par
calorimétrie différentielle à balayage (DSC). Les résultats ont montré que les nanoparticules de
zéolithe pouvaient jouer le rôle d'agents de nucléation efficaces au cours du procédé de moussage.
Cependant, la teneur optimale en nanoparticules était fortement liée aux conditions de moussage.
Enfin, les performances de séparation membranaire ont été étudiées pour différents gaz (CO2,
CH4, N2, O2 et H2), montrant que l’incorporation de zéolithe poreuse améliorait de manière
significative les propriétés de transport de gaz des membranes de polyoléfines semi-cristallines
en raison de l’épaisseur réduite de la paroi cellulaire (contrôle de perméabilité) et de meilleures
propriétés de séparation (contrôle de la sélectivité). Ces résultats montrent que les membranes à
matrice mixte (MMM) peuvent être rentables, faciles à fabriquer et efficaces en termes de vitesse
de traitement, en particulier pour l'industrie pétrolière où la séparation/purification de H2/CH4 et
H2/N2 est importante pour la récupération de l'hydrogène.
103
Abstract
In this work, hollow fiber porous nanocomposite membranes were successfully prepared by the
incorporation of a porous nanoparticle (zeolite 5A) into a blend of linear low-density polyethylene
(LLDPE)/low-density polyethylene (LDPE) combined with azodicarbonamide as a chemical
blowing agent (CBA). Processing was performed via continuous extrusion using a twin-screw
extruder coupled with a calendaring system. The process was firstly optimized in terms of
extrusion and post-extrusion conditions, as well as formulation to obtain a good cellular structure
(uniform cell size distribution and high cell density). Scanning electron microscopy (SEM) was
used to determine the cellular structure as well as nanoparticle dispersion. Then, the samples were
characterized in terms of mechanical and thermal stability via tensile tests and thermogravimetric
analysis (TGA), as well as differential scanning calorimetry (DSC). The results showed that the
zeolite nanoparticles were able to act as effective nucleating agents during the foaming process.
However, the optimum nanoparticle content was strongly related to the foaming conditions.
Finally, the membrane separation performances were investigated for different gases (CO2, CH4,
N2, O2, and H2) showing that the incorporation of porous zeolite significantly improved the gas
transport properties of semi-crystalline polyolefin membranes due to lower cell wall thickness
(controlling permeability) and improved separation properties (controlling selectivity). These
results show that mixed matrix membranes (MMMs) can be cost-effective, easy to process, and
efficient in terms of processing rate, especially for the petroleum industry where H2/CH4 and
H2/N2 separation/purification are important for hydrogen recovery.
Keywords: polyethylene; hollow fiber; mixed matrix membranes; gas separation; extrusion;
cellular structure
104
4.1 Introduction
Microporous polymer membranes are currently considered as commercially attractive due to their
low operating temperature and manufacturing costs, as well as good processability. However, due
to limited polymer thermal stability and the plasticization effect, their applications have been
restricted to separation processes where severe conditions are not encountered [258],[259]. One
effective approach to improve the performance of polymer membranes is to incorporate inorganic
nanoparticles such as zeolites and carbon-based molecular sieves since they have higher thermal
resistance and chemical stability, combined with molecular sieving property. This led to the
concept of mixed matrix membranes (MMMs) which were shown to have increased selectivity
and permeability compared to neat polymer membranes [53][260]. Nevertheless, the addition of
these inorganic particles makes the membranes more fragile.
A large number of investigations focused on membrane production to improve the gas transport
properties based on material selection, especially for commercial-scale applications. However,
very few studies focused on the control of the morphological structure in relation with the
membranes’ performances [19],[27],[26].
Several methods are available to prepare asymmetric and symmetric porous polymer membranes,
but most of them are based on solvent casting followed by phase separation [223], where solvent
toxicity and costs (recycling/elimination/treatment) are the main drawbacks of these non-
environmentally friendly methods. On the other hand, a few methods have been proposed to
create a microporous membrane without any solvents. The use of leachable particles such as salts
was shown to produce an open-cell structure based on polystyrene [120], polypropylene [121],
and polyethylene [117]. Other solvent-free methods are stretching with or without particles [261]
and direct foaming of semi-crystalline thermoplastics such as high-density polyethylene (HDPE)
and polypropylene (PP) [120],[122]. For stretching methods, the porosity is created by
delamination between the matrix and local stress concentration points such as solid particles or
crystalline zones, but thermal post-treatment is needed to stabilize the newly created microporous
and crystalline structure to avoid shrinkage and warpage, which are the main drawbacks of such
methods.
However, the cellular structure of a polymer can be produced via direct foaming processes leading
to weight and waste reduction, heat resistance, high strength to weight ratio, low cost,
recyclability, multi-functionally, as well as easier processing. Today, polymer nanocomposite
105
foams can be produced via extrusion [152], injection molding [262], compression molding [221],
in situ polymerization [224], and phase separation [223]. Extrusion is however the only one being
continuous. Thus, extrusion is very effective, fast, and economic to produce low-density foams
[137],[225]. The cellular structure can be produced using either chemical or physical blowing
agents (PBAs), but chemical blowing agents (CBAs) are generally easier to handle with standard
equipment compared to physical blowing agents [19], [25].
The authors’ previous works showed how the morphological properties, such as cell size
uniformity as well as cell density, can control the physical, mechanical and gas transport
properties of hollow fiber polymer membranes [19],[27],[263]. Since these properties are tunable,
this led to membrane performance enhancement via simultaneous improvement of both the
surface area as well as gas transport properties [19],[82]. Then, a melt processing method was
optimized to produce hollow fiber foamed membranes (HFM) based on a microcellular blend of
linear low-density polyethylene (LLDPE) and low-density polyethylene (LDPE) to get a balance
between foamability and mechanical strength. Furthermore, the blowing agent content,
processing (temperature profile, flow rate, screw speed, etc.), and post-processing (stretching)
conditions were the main parameters to control the porous membrane structure. Direct foaming
is interesting because it is solvent free, continuous, and fast to produce a porous structure in
hollow fibers [19],[26],[263]. This configuration is the most interesting because of its higher
active surface area per unit volume compared with flat membranes.
Among a wide range of resins, polyolefins such as polyethylene (PE) are today one of the most
widely used thermoplastic resins due to their excellent overall performance and
thermomechanical properties, chemical resistance, and low cost, as well as easy processability
and recyclability. Polyethylene membranes have been commonly used in microfiltration and
blood oxygenation applications [93],[117],[189],[264] , but their application as gas separation
membranes has been limited because of their low gas transport properties owing to their semi-
crystalline nature. However, this can be modified. For example, Covarrubias and Quijada [192]
investigated the effect of aluminophosphate (ALPO) addition into PE membranes through melt
compounding. The results showed that a substantial improvement in gas transport properties was
obtained by the creation of a gas transport pathway provided via the swollen ALPO. Hydrogen
permeability was increased from 1.4 to 11.9 Barrers and H2/CO selectivity increased from 1.2 to
17. Nevertheless, polyolefin-based composite membranes for gas separation have barely been
106
investigated, and this lack of information limits their use since the relation between membrane
performances and morphology is not well understood. This is why one of the objectives of this
work is to investigate the effect of foaming conditions on cell density and cell size distribution
by the addition of nanoparticles to control not only the mechanical, but also the gas permeation
properties of these materials. However, the foaming of semi-crystalline polymers, such as
polyethylene, is more challenging compared to amorphous resins, since particle addition has a
direct effect on foam cell nucleation and crystal nucleation. These resins are also known to have
a very narrow foaming temperature window [22],[134],[219].
Nevertheless, the authors’ previous results showed that hollow fiber membranes based on foamed
LLDPE/LDPE blends can be a good starting point due to their interesting gas permselectivity.
Although several approaches are available to improve the foamability of a polymer [22],[28], it
was found that nanoparticle addition as a nucleating agent can be a simple but very effective
strategy to improve the cellular structure of polymer foams in terms of cell density and cell size
uniformity [146],[265]. Two different types of nucleation mechanisms are known (homogeneous
and heterogeneous) [82], but heterogeneous nucleation becomes predominant when particles with
a high surface area are used [129]. This approach has been successfully applied for polyolefins
using Ca2CO3 [152], clay [164], silica [27],[25], [166], and Fe3O4 (OA-Fe3O4) [167].
To improve membrane properties, different particles have been used. The focus for this study was
the zeolite nanoparticle [19],[263]. However, the main issues in matrix membrane production are
particle spatial distribution (agglomeration) [22], as well as possible interfacial rigidification and
pore blockage [128]. For this reason, zeolite nanoparticles were added through a well-controlled
extrusion process to limit these issues.
This work is also a second step to improving upon the results obtained from the authors’ previous
studies [27],[19],[263]. In fact, the idea was to propose a sustainable alternative production
method for hollow fiber mixed matrix membranes. The matrix was made from foamed
LLDPE/LDPE (70/30) and the addition of zeolite as nucleation agent/gas permeation modifier
was performed to investigate its effects on the cellular structure and gas separation performance
of the resulting nanocomposite foams referred to as mixed matrix foam membranes (MMFMs).
The membranes were prepared through a continuous non-solvent method (foam extrusion) with
a focus on zeolite content. For the separation performances, permeability and ideal selectivity
were determined.
107
4.2 Materials and methods
4.2.1 Materials
The material used in this study was linear low-density polyethylene (LLDPE) powder LL 8460
with a melting temperature of 126 C, a density of 0.938 g/cm3, and a melt index of 3.3 g/10 min
(190 C at 2.16 kg) provided by ExxonMobil (Calgary, AB, Canada). The low-density
polyethylene (LDPE) used was LA-0219-A supplied by Nova Chemicals (Calgary, AB, Canada).
It has a melt flow index (MFI) of 2.3 g/10 min (190 C at 2.16 kg) and a density of 0.919 g/cm3.
As a chemical blowing agent (CBA), activated azodicarbonamide (Celogen 754A) was purchased
from ChemPoint (Midland, MI, USA). According to the information from the supplier, this CBA
has a gas released of 200 cm3/g of a mixture of N2, CO2, CO, and NH3. Its decomposition takes
place between 165–180 C. Zeolite 5A nanoparticles were purchased from Sigma-Aldrich
(Oakville, ON, Canada) and used as received. These nanoparticles have a size less than 3 µm.
The permeation performance of each sample was performed by using commercially important
gases having various kinetic diameters, including H2 (2.89 Å), CO2 (3.30 Å), O2 (3.46 Å), N2 (3.64
Å), and CH4 (3.80 Å).
4.2.2 Sample preparation
Two blending techniques were used to produce a cellular structure with a LLDPE/LDPE (70/30)
blend with different zeolite nanoparticles contents (10, 15, and 20 wt.%) and an optimum
chemical blowing agent content (2.5 wt.%). Before blending, both polymers, CBA, and zeolite
were dried at 70 C for 24 h. In the first approach, LLDPE/LDPE/zeolite compounds were
prepared by melt compounding (extrusion) at 140 C and 50 rpm. The process was done on a co-
rotating twin-screw extruder (ZSE-27, Leistritz, Nürnberg, Germany) with a length/diameter
(L/D) of 40 and a diameter of 27 mm with 10 individually heated/cooled zones. Further details
for the optimum base conditions can be obtained in the authors’ previous study [263]. In the
second approach, since all the materials were in a powder form (both polymers, zeolite, and
CBA), they were simply dry-blended for 20 min. For foaming, the same extruder as for melt
compounding was used, but a tubular die (inside diameter of 3.5 mm and outside diameter of 5
mm) was used instead of a circular die and the screw speed was decreased to 25 rpm. At the
extruder’s exit, the material was introduced in a two-roll calendaring system to impose different
108
drawing speeds and have a control on the stretching (post-extrusion) extrusion. The complete
experimental set-up is shown in Figure 4.1. The foamed membranes prepared by using direct
foaming are referred to as mixed matrix foamed membrane (MMFM) followed by a number
showing the zeolite content (wt.%). For comparison, unfoamed (compact) samples have also been
produced with the same zeolite contents using the same processing conditions. These samples (L)
are named as L-0, L-10, L-15, and L-20 where the number refers to the zeolite content. More
details can be found in Table 4.1.
Figure 4.1. Schematic representation of the extrusion set-up to produce hollow fiber mixed
matrix foamed membranes.
Table 4.1. Specifications of the hollow fibers produced *. Unfoamed (Solid) Foamed
Sample
Zeolite
Concentration
(wt %)
Uniaxial
Stretching
Speed
(m/min)
Sample CBA
(wt %)
Uniaxial
Stretching
Speed (m/min)
Zeolite
Concentration
(wt %)
L-0 0 5, 7, 9, 11, 13 MMFM0 1.75 5, 7 ,9, 11, 13 0
L-10 10 5, 7, 9, 11, 13 MMFM10 2.5 5, 7 ,9 10
L-15 15 5, 7, 9, 11, 13 MMFM15 2.5 5, 7 ,9 15
L-20 20 5, 7, 9, 11, 13 MMFM20 2.5 5, 7 ,9 20
* All unfoamed and foamed samples have a composition of LLDPE/LDPE 70/30 (wt %).
4.2.3 Characterization
Scanning electron microscopy (SEM) was performed on a JSM 840A (JEOL, Tokyo, Japan) to
analyze the zeolite dispersion and foam structure. For each sample, images were taken in two
109
perpendicular directions, namely, the flow (F) and transverse (T) directions, to get a complete 3D
evaluation of the structure related to deformation in the stretching direction. The samples were
cut with a doctor blade and the exposed surfaces were made conductive by deposition of a thin
gold/palladium coating. Cell density and cell size were determined by using the Image J software
(National Institutes of Health, Bethesda, MD, USA). The values are reported as an average of at
least three images in each direction. Due to cell deformation, the cell density (N) was determined
as [222]:
𝑁 = 𝑁1(𝑁2)12⁄ (5.1)
where N1 and N2 are defined as the surface cell densities in the F and T directions, respectively,
and determined as:
𝑁𝑖 =𝑛
𝐴 (5.2)
where n is the number of cells in a defined area A (cm2).
Density was obtained by using a gas (nitrogen) pycnometer Ultrapyc 1200e (Quantachrome,
Boynton Beach, FL, USA). Thermogravimetric analysis (TGA) was used to determine the weight
loss curves and evaluate the samples’ thermal stability as well as to confirm the particle contents.
This was done on a Q5000IR (TA Instruments, New Castle, DE, USA) from 50 to 800 °C at a
heating rate of 10 °C/min under nitrogen. The matrix crystallinity was determined via differential
scanning calorimetry (DSC) on a DSC 7 from Perkin Elmer (Waltham, MA, USA). The
experiments were conducted using approximately 5 mg of samples in aluminum pans under N2
at a rate of 10 C/min. A heating/cooling/heating between 50 to 200 °C was used to run the
experiments. The results extracted from the first heating cycle were noted to show the influence
of foaming history on the crystallization degree of the polymer blend/nanofiller foamed
composites. The mechanical properties were obtained with a 500 N load cell and a 10 mm/min
crosshead speed at ambient temperature (23 °C) using a universal testing machine model 5565
(Instron, Norwood, MA, USA). To report on the Young’s modulus, tensile strength, and
elongation at break, a minimum of five replicates was used.
110
4.2.4 Gas permeation
The separation performance of the membranes was measured by gas permeation analysis. As
presented in Figure 4.2, a sample was fixed in a permeation hollow fiber module. The module
was evacuated (feed and permeate sides) for a minimum of 6 h under vacuum. To start the
experiment, a gas was introduced on the feed side at a pressure 30 psi, while the permeate side
had a constant volume. Then, the pressure variation on the permeate side was measured as a
function of time until a steady state was achieved, which was used to determine the permeance
by the solution-diffusion model. All the tests were carried out at 30 °C. The values reported are
the average of at least five measurements and the permeance coefficient (Q in gas permeation
unit (GPU)) was determined as follows:
𝑄 = (𝑃
𝑙) =
22414
𝐴×
𝑉
𝑅𝑇∆𝑝×
𝑑𝑝
𝑑𝑡, (5.3)
where V is the constant volume of the permeate side (cm3) and L and A are membrane thickness
(cm) and the membrane area (cm2), respectively. 𝑑𝑝
𝑑𝑡is the rate of permeation in the constant
volume under steady state condition (psi/s), ∆p is the variation between the upstream pressure
and permeate pressure, T is the absolute temperature (K), and R is the universal gas constant
(6236.56 cm3 cmHg/mol K). The ideal selectivity between two gases A and B is defined as the
ratio between the more permeable gas (A) and the less permeable one (B), as follows:
𝛼𝐴/𝐵 =𝑄𝐴
𝑄𝐵 (5.4)
111
Figure 4.2. Schematic representation of the (right) permeation set-up with the (left) hollow
fiber membrane module set-up.
4.3 Results and discussion
4.3.1 Preparation and optimization of MMFM
The foaming of polyolefins, such as polyethylene, is generally more challenging in comparison
with most polymers (such as polystyrene) due to a quite narrow foaming temperature window for
semi-crystalline polymers [137],[265]. In the authors’ previous study, it was shown that a cellular
structure of LLDPE/LDPE (70/30) having a high cell density (1.7 × 107 cells/cm3) and a uniform
cell size distribution could enhance the membrane permeance and selectivity by approximately
100% and 75%, respectively, in comparison with a compact (unfoamed) LLDPE/LDPE
membrane [137]. The current work confirms their previous findings and provides additional
evidence suggesting that zeolite addition improved these properties. Multiphase polymer
materials, in comparison with neat polymers, can have a different foaming behavior because of
low-energy interfacial regions. It is known that using a polymer with a low melt strength for
foaming leads to poor cell morphology due to cell rupture/coalescence resulting from poor cell
wall elasticity associated with biaxial elongational flow around growing cells. Moreover, to
improve the matrix (LLDPE/LDPE) melt strength and its nucleation behavior, particle addition
(nucleating agent) has been reported [25],[153] [265].
112
4.3.2 Morphology
The dispersion degree of nanoparticles in the polymer matrix has a considerable effect on the
final results [28]. In this study, the zeolite dispersion degree was even more critical to control the
gas permeation. An excellent dispersion has a positive nucleating effect, whereas particle
agglomeration decreases the effectiveness of these nanoparticles as bubble/crystal nucleators
during the foaming process. They can also create non-selective channels for the gas molecules
leading to poor separation performance [126],[266],[267].
Firstly, Figure 4.3 presents typical images for the unfoamed nanocomposites. The bright white
spots are zeolite nanoparticles where well-dispersed aggregates of approximately 1 to 3 µm are
seen. It is also clear that the aggregate number increases with zeolite content. These results are in
agreement with other studies performed on polyethylene as the matrix and zeolite nanoparticles
with larger diameters [43],[268].
Figure 4.3. High-resolution images showing zeolite nanoparticle dispersion in the matrix.
Figure 4.4 presents typical cross-section images for MMFMs with different zeolite particles
loading (0, 10, 15, and 20 wt %) using a stretching speed of 5 m/min. From the micrographs, it
can be observed that the addition of the porous zeolite particles produced a significant increase
in the number of cells and a reduction in cell size. Based on the images taken, Figure 4.5a presents
the cell density as a function of zeolite content for a die temperature of 168 C and a stretching
speed of 5 m/min. These results show that with zeolite addition, the cell density significantly
increased (two orders of magnitude) from 1.2 × 107 cells/cm3 for MMFM0 to 120 × 107 cells/cm3
for MMFM15. A negligible variation was observed at higher zeolite content (20 wt.%), probably
due to severe particle agglomeration.
113
Figure 4.4. Micrographs of the hollow fiber foam membranes with different zeolite contents (0,
10, 15, and 20 wt. %) at a stretching speed of 5 m/min.
On the other hand, the expansion ratio did not change much (1.4–1.9) between all the samples
studied. Nonetheless, the effect of zeolite on the average cell size (F and T directions) is clearly
seen in Figure 4.5b. The results reveal a significant cell size reduction in the transverse direction
compared to the neat foam (0% zeolite). For example, the average cell size in the T direction
decreased from 105 to 30 µm at 10% zeolite. Then, a slight increasing is observed with increasing
zeolite content. These results are consistent with those reported for a LDPE/silicon system [25],
[166]. Similarly, filler addition decreased the average cell size in the flow direction from 165 to
110 µm. The optimum zeolite content may be related to gas loss and cell coalescence when the
zeolite content is too high (Figure 4.4). Finally, based on the above findings, the reason for the
enhanced foaming potential of LLDPE/LDPE/zeolite blends is the presence of zeolite, leading to
increased heterogeneous nucleation. Also, it should be mentioned that samples with less than 10%
of zeolite did not show a significant increase in those parameters, the reason for which the findings
are not presented.
114
Figure 4.5. (a) Cell density and expansion ratio, (b) average cell sizes in the transverse (T) and
flow (F) directions of mixed matrix foam membrane (MMFMs) with different zeolite contents
(0, 10, 15, and 20 wt %) at a stretching speed of 5 m/min.
4.3.3 Effect of blending method on the cellular structure
Before the foaming process, two methods were used to introduce the zeolite in the polymer
matrix, namely, melt compounding and dry blending. Figure 4.6 presents typical results for the
sample MMFM15. Interestingly, samples based on dry blending led to lower cell sizes compared
to melt compounding. Figure 4.7 reports on the results after image analysis. The cell size
reduction is assumed to be related to the possible absorption of gases which are released from the
CBA by the porous zeolite nanoparticles. As melt compounding may lead to possible pore
blockage and/or particle collapse/shrinkage, more gas is available and released for cell growth
compared with dry blending. This effect is more important as zeolite content increases since a
higher cell density is obtained at a higher zeolite concentration (10–20 wt %) with 2.5% CBA.
Based on this observation, it is expected that the addition of porous zeolite nanoparticles can lead
to lower cell sizes, especially using a dry blending method before foaming.
115
Figure 4.6. Effect of the production method on the MMFM morphology with 15 wt.% zeolite:
(left) melt compounding and (right) dry blending.
The effect of zeolite on PE foam nucleation is thus affected by the blending method, and a
schematic representation in presented in Figure 4.8 to explain the mechanisms. The energy barrier
associated with homogeneous nucleation is usually much higher compared to heterogeneous
nucleation [269]. Consequently, the surface of a nucleating agent has lower surface free energy
(energy barrier) and the particles act as heterogeneous nucleation points leading to higher cell
density and smaller cell sizes. Interestingly, there were negligible differences in the foam density
with zeolite addition. However, zeolite addition via dry blending prior to foaming seems to
provide a higher surface area because of more available open pores due to the gas cavities.
Therefore, the foams have higher cell density and smaller cell size. This finding is in agreement
with recent results on poly(methylmethacrylate) (PMMA) foams using nonporous (compact)
(solid silica) and porous particles (spherical ordered mesoporous silica) as nucleating agents
[270]. This effect will be further discussed in the permeation section.
116
Figure 4.7. Effect of blending method (melt compounding and dry blending) on the cell size in
both T and F directions, as well as cell density for MMFM0 and MMFM15.
Figure 4.8. Schematic representation of the porous and nonporous nucleating agent mechanism
of zeolite during foaming.
117
4.3.4 Effect of the stretching rate on the cellular structure
Figure 4.9 presents typical SEM micrographs in the flow (F) and transverse (T) directions of
MMFM15 produced with 2.5% CBA at a constant die temperature (168 C) and different
stretching speeds (5, 7, and 9 m/min). As expected, based on the authors’ previous investigations,
by increasing the stretching speed from 5 to 9 m/min more elongated cells in the flow direction
are produced. Consequently, the cell aspect ratio increased with the stretching rate, as presented
in Figure 4.10. Based on the image analysis, it can be observed that the cell density might slightly
decrease with increasing stretching speed, but the variation may not be significant due to large
distributions. However, the further increase in stretching speed (9 m/min) resulted in several cell
wall ruptures and voids were created around the nanoparticle cellular structure. This is different
than what was reported in the authors’ previous investigation where stretching speeds up to 13
m/min were used. This was possible because an unfilled PE matrix was used (no nanoparticles).
In this study, interfacial voids around the nanoparticles were created because of delamination due
to high stretching speed [261],[43]. Therefore, an optimum stretching speed must be selected to
get the best foam structure in terms of high cell density and narrow cell size distribution without
any defects in the polymer structure for good membrane performances (i.e., no short-circuit).
118
Figure 4.9. Micrographs of MMFM15 produced at different stretching speeds (m/min) for the
(top row) transverse (T) and (bottom row) flow (F) directions.
119
Figure 4.10. Effect of the stretching speed on cell density and cell aspect ratio (AR) in the flow
direction for MMFM15 (2.5% CBA).
4.3.5 Gas permeation performances
The gas transport properties of polyolefins, as well-known semi-crystalline polymers, are a result
of their crystallinity. Their crystalline regions can behave as gas barriers (regions of negligible
diffusion and solubility), hence the penetrant gas molecules have a more tortuous path (diffusion
path) [185],[192],[241]. Therefore, the permeability should overall decrease with an increasing
degree of crystallinity, especially for larger gas molecules [195],[271]. This can also be explained
through the limitation of polymer chain mobility in the crystalline areas [19],[185]. Overall, the
size of the gas molecules can influence their mobility inside the polymer structure. Polyethylene,
with a rubbery character, can also favor the condensation of gases in its free volume developed
by the mobile and flexible chain molecules [192],[272]. The permeability in porous dense
polymer membranes is defined through the solution-diffusion model [27]. Thus, the effect of
zeolite content on the film crystallinity is presented in Figure 4.11a. It can be seen that the
crystallinity of the authors’ PE blend/zeolite foamed composites decreased as zeolite content
increased, and similar results have been reported for PE above 10 wt % zeolite [43],[268].
Generally, the introduction of inorganic particles not only replaces polymer crystals and occupies
120
the crystal structure sites, but also creates more amorphous areas, especially at higher particle
concentration [22],[164],[273].
Figure 4.11. Effect of (a) zeolite loading and (b) stretching speed on the crystallinity of
MMFMs.
The permeance (Equation (5.3)) of the MMFMs with different zeolite contents is presented in
Figure 4.12. The results show that the permeance significantly increases with zeolite addition
compared to hollow fibers based on foamed and unfoamed MMFM0 membranes for all the gases
studied. It was also found that the permeance increases with increasing zeolite concentration up
to MMFM15 and then slightly decreases for MMFM20. This optimum content represents again
a balance between the cell morphology created (cell size and cell density), the PE crystallinity
level and the zeolite dispersion (see Figure 5.5a).
121
Figure 4.12. Permeance of various gases (CH4, CO2, N2, O2, and H2) at 30 C and 30 psia for
foamed and unfoamed hollow fiber MMFMs produced with different zeolite loading (0, 10, 15,
and 20 wt %) at a stretching speed of 5 m/min.
However, the gas permeance of the foamed hollow fiber MMFM0 membranes is higher than the
unfoamed (L-0) membranes for all the gases analyzed (H2, O2, CH4, CO2, and N2). The
incorporation of zeolite nanoparticles within the polymer matrix produced a significant increase
in the gas transport properties due to their porous nature. As expected, H2 permeance is the highest
because of its smaller molecular size (2.98 Å) compared with other gases (O2: 3.46 Å, CO2: 3.30
Å, N2: 3.64 Å, and CH4: 3.80 Å). This also explains the higher permeance of CO2 compared to
CH4, O2, and N2, especially due to its high condensability leading to higher solubility in
polyethylene. In particular, Figure 4.13a shows that the H2 permeance increased by seven and
three times for MMFM15 compared to the unfoamed (L-0) and foamed (MMFM0) membranes,
respectively. This result can be related to more gas diffusion pathways provided by the porous
nature of zeolites [126] and a larger number of voids in the foam structure [19],[26],[263].
The ideal selectivity (Equation (5.4)) results for H2/CH4 and H2/N2 are presented in Figure 4.13b.
A negligible variation was observed in the unfoamed L-0 and L-15 samples for both values, but
slight improvements were observed for the foamed membranes (MMFM0). However, all
MMFMs exhibited superior H2/CH4 and H2/N2 ideal selectivity compared to samples without
zeolite (MMFM0). At the optimum zeolite loading (15 wt %), the highest H2/CH4 ideal selectivity
122
was 12.5, whereas the ideal selectivity for H2/N2 increased from 3.9 to 22.9. As mentioned above,
since the small molecular size of H2 leads to a higher permeance than for the other gases, any
membrane permeance improvement mostly favors H2 permeation. However, the main idea in
porous zeolite addition is to improve the separation performances by developing some free
volume in the polymer, as well as to control the porosity (foam structure) and crystallinity (see
Figure 4.11) by facilitating H2 transport and reducing the other gases by adsorption/molecular
sieving effect, especially for N2 [274].
Figure 4.13. Permeation results for (a) H2 gas permeance and (b) H2/CH4 and H2/N2 ideal
selectivity at 30 C and 30 psia for different membranes as a function of zeolite loading.
According to Figure 4.6, it can be seen that the effect of the dry blending and melt compounding
techniques on the MMFM15 structure will have a direct effect on the membrane separation
performance. Figure 4.14 clearly shows this effect in terms of gas permeation results. Partial pore
blockage of the porous zeolite might be the main reason for the decreasing permeance and
selectivity of melt blended membranes [53]. As discussed above, apart from pore blockage, the
gas released by the CBA is distributed differently as new gas molecule diffusion pathways are
provided by dry blending; this is schematically presented in Figure 4.15.
123
Figure 4.14. Effect of melt compounding and dry blending on H2/CH4 and H2/N2 ideal
selectivity and H2 permeance of the selected MMFM15 membranes with 15 wt % zeolite.
Figure 4.15. (a) Scanning electron microscopy (SEM) of the selected MMFM15 membranes
and (b) schematic representation of the gas molecule diffusion pathways.
124
4.3.6 Influence of stretching speed on the membrane’s gas transport
The effect of the stretching speed on the performance of MMFM15 was investigated by preparing
membranes at different stretching speed (5, 7, and 9 m/min). The H2 permeance and H2/CH4 and
H2/N2 ideal selectivity of these MMFMs are presented in Figure 4.16 . Increasing the stretching
speed significantly increased the H2 permeance above 500 GPU, whereas the H2/CH4 and H2/N2
ideal selectivity substantially decreased. This implies that some undesirable pores were created
between the PE matrix and zeolite particle interface, where the gas molecules might go through
(short-circuit). Consequently, the H2/CH4 and H2/N2 ideal selectivity significantly decreased
below 4. As seen in Figure 4.15, these interfacial voids shaped mostly in the island area, that is,
the area between cells or containing particles, whereas good interfacial adhesion for MMFM15
at 5 m/min was obtained. Based on these observations, the structure produced combined the
benefits of both particle addition and stretching [261], as well as foaming to produce multiporous
structure membranes to facilitate gas transport properties [19],[263].
Figure 4.16. Results for H2 gas permeance and H2/CH4 and H2/N2 ideal selectivity at 30 C and
30 psia for MMFM15 at different stretching speeds.
Zeolite addition to the unfoamed hollow fiber (L-15) membrane led to a negligible variation in
separation performance compared to the unfilled and unfoamed hollow fiber (L-0) samples
(Figure 4.13). For H2, diffusion more than solubility was responsible for the improved permeance
of polyolefins. On the other hand, both CH4 and N2 solubility and diffusivity were decreased by
125
zeolite addition [192]. Since CH4 and N2 are larger molecules than H2, they move more slowly
and/or are trapped inside the cells [263]. Moreover, the incorporation of porous zeolites increased
gas adsorption, especially for N2. Therefore, the permeance is a function of cell density, zeolite
loading, interfacial state, and PE crystallinity.
As a conclusion based on the results obtained, a 70/30/15 blend of LLDPE/LDPE/zeolite with 2.5
wt % of CBA and stretched at 5 m/min seems to be the optimum. To complete the authors’
characterization, mechanical and thermal properties are discussed next.
4.3.7 Mechanical and thermal properties
The tensile behavior of unfoamed and MMFMs at different zeolite content is shown in Figure
4.17. Generally, cellular materials have lower modulus and strength when compared to their
unfoamed counterpart. This is associated with the fact that less polymer (matrix) was available to
sustain the applied stresses. As expected, the tensile modulus and tensile strength of both L-0
(280 MPa, 16 MPa) and L-15 (350 MPa, 18 MPa) (unfoamed samples) are higher than the foamed
samples. For the foams, the moduli of MMFM0 to MMFM20 were found to increase with zeolite
concentration. Figure 4.17b shows that the tensile strength trend is similar to the modulus. This
is related to the inherent rigid properties of inorganic particles, at improving the mechanical
properties of MMFMs for the concentration range studied. On the other hand, the elongation at
break of unfoamed samples decreased with zeolite addition (L-0 and L-15), whereas the values
for MMFMs decreased less since they are more brittle. These findings are consistent with
previous studies [268],[275].
126
Figure 4.17. (a) Young’s modulus, (b) tensile strength, and (c) elongation at break of the
foamed and unfoamed hollow fibers at different zeolite loading.
The effect of the stretching rate on the tensile behavior of the selected MMFM15 membranes is
shown in Figure 4.18 a,b. It can be seen that the modulus and strength decreased with increasing
stretching rate for both MMFM0 and MMFM15 membranes. However, the elongation at break
of MMFM0 increased with increasing stretching speed as seen in Figure 4.18c, whereas the
elongation at break for MMFM15 decreased by increasing the stretching speed due to the
presence of zeolite particles.
127
Figure 4.18. Effect of stretching speed on the (a) Young’s modulus, (b) tensile strength, and (c)
elongation at break of selected MMFM15 and L-15.
Characteristic temperatures from TGA curves are presented in Figure 4.19. This analysis suggests
that the presence of zeolite not only improved the cell size distribution and cell density, but also
the membrane thermal stability. Inorganic particles typically have inherently excellent thermal
properties compared to polymers. Therefore, it can be seen in the derivative thermogravimetric
analysis (DTGA) curves (Figure 4.19) that the main decomposition temperature (Tmax)
significantly increased from 456 to 462 and 463 C for MMFM10, MMFM15, and MMFM20
membranes, respectively, compared to MMFM0 (451 C).
128
Figure 4.19. Derivative thermogravimetric analysis (DTGA) curves of MMFM samples
compared to unfoamed samples.
The effect of zeolite content of the MMFMs has also been studied by TGA (Figure 4.20). It was
found that these membranes are more stable compared to MMFM0. It can be observed that above
475 C, MMFM0 is totally degraded, whereas the zeolite content (associated with the residues
above 500 C) in MMFM10, MMFM15, and MMFM20 membranes was confirmed to be close
to 10%, 15%, and 20%, respectively. Moreover, the values of T10% (the temperature for 10%
mass loss) of MMFM0, MMFM10, MMFM15, and MMFM20 are also increasing: 401, 412, 421,
and 425 C, respectively.
129
Figure 4.20. Thermogravimetric analysis (TGA) curves of MMFM samples compared to
unfoamed samples.
4.4 Conclusion
In this work, hollow fiber cellular composite membranes were prepared and optimized from a PE
blend (LLDPE/LDPE 70/30) with different zeolite loadings (0, 10, 15, and 20 wt.%) combined
with azodicarbonamide as a chemical blowing agent. The effects of zeolite content as well as
stretching rate on the cellular structure and gas separation performance of these mixed matrix
foamed membranes (MMFMs) were reported. The membranes were prepared via a fast, cost-
effective, and easy continuous extrusion process. SEM images confirmed good zeolite dispersion
during manufacturing with this non-solvent technique.
The findings showed that the addition of a porous nanoparticle (zeolite 5A) significantly
increased the cell density for the optimum membrane (MMFM15) with a value of 1.2 × 109
cells/cm3 and a substantial reduction of cell sizes by approximately 72%. Overall, from a
microcellular foaming point of view, the results obtained in this investigation indicate that the
zeolites, with their porous structures, have a significant potential to act as nucleating agents for
polymer foaming.
130
The gas permeation results showed that the H2 permeance and H2/CH4 and H2/N2 selectivity of
the unfoamed (L-0) membranes compared to the optimum membrane (MMFM15) significantly
increased from 39 to 270 GPU and 3.3 (H2/CH4) and 3.9 (H2/N2) to 12.5 and 22.9, respectively,
by introducing zeolite leading to a better cellular structure inside the polyethylene membranes.
Furthermore, combining the benefits of both particle addition and stretching, as well as foaming,
helped to create multiporous structure membranes to facilitate gas transport properties. Finally,
significant permeance and selectivity improvements were obtained compared to the neat matrix.
Acknowledgement
The authors would like to thank the Centre de recherche sur les matériaux avancés (CERMA) and
Mr. Yann Giroux for technical support. Materials were kindly provided by ExxonMobil (LLDPE)
and Nova Chemicals (LDPE). This research was funded by the Natural Science and Engineering
Research Council of Canada (NSERC), grant number RGPIN-2016-05958.
131
Conclusions and Recommendations for future work
132
General conclusions
Academic investigation and industrial development led to the development of a variety of dense
and porous polymeric membranes. The production of porous polymer membranes was the focus
of this thesis. It was expected that gas separation through controlled porosity (cell density and
cell size) would control the permeability and diffusivity of a polymer which are two fundamental
requirements for a good membrane. To improve on permeability and selectivity, the porous
membrane were “doped” with a porous particle to create mixed matrix membranes.
Based on the experimental work performed, several conclusions were obtained via an
experimental step-step approach. The main information obtained is:
1. A continuous processing technique was developed to produce porous hollow fiber
membranes using melt foam extrusion followed by stretching. The system was shown to
perform well using linear low density polyethylene (LLDPE) as the matrix and
azodicarbonamide (ADC) as a chemical blowing agent. The technique is considered to be
cost-effective, simple, fast and solvent-free. Since the process is complex, the processing
conditions must be optimized to obtain a suitable hollow fiber foam having high cell
density and cell size uniformity leading to high separation performance. As reported
several times in the literature, the selected LLDPE has a very narrow foaming temperature
window of 1724C. Increasing the foaming temperature above this range resulted in cell
opening (collapse) along with non-uniform cell distribution, while below this optimum
range limited cell nucleation and cell growth occurred. For foaming, the optimum blowing
agent concentration and the stretching speed were 1.75% wt. and 5 m/min, respectively.
The morphological results showed that a hollow fiber foamed LLDPE was optimized by
having an average cell size of 165 and 105 µm in the transversal (T) and extrusion (F)
direction respectively, leading to a cell density of 10.5×106 cells/cm3 with an expansion
ratio of 1.8. The gas permeation results showed a clear relation between the cellular
structure and the permeance/selectivity, especially for H2 and H2/CH4. Since LLDPE is a
semi-crystalline polyolefin, crystallinity was also found to have key role in the membrane
performances with an inverse relationship between crystallinity and permeance. Finally,
133
the membrane having the highest cell density with better cell size uniformity was shown
to have the lowest Young’s modulus and tensile strength as well.
2. Based on the results of the first part, it was possible to improve the polymer matrix
foamability by adding different LDPE contents, which is known to have a high melt
strength, while having a chemical structure close to LLDPE. The effect of using this
polymer blend led to higher possible stretching speed and better cell structure. It was
found that a stretching speed of 11 m/min was the optimum for both LLDPE/LDPE blends
(70/30 and 60/40) leading to a cell density of about 1.7×107 cells/cm3. The gas permeation
results showed that the permeance of all the gases studied (CH4, CO2, N2, O2 and H2)
increased with lower foam density. In particular, the H2 permeance and H2/CH4 selectivity
of the unfoamed LLDPE membranes and the optimum foamed blend (70/30) were
increased by 100% and 76%, respectively.
3. Another strategy to further improve the foamability of a polymer is by using nanoparticles
which will act as nucleating agents. Multiphase polymer materials, compared with neat
polymers, can have different foaming behaviors due to the presence of interfacial regions.
The influence of zeolite (5A) content and stretching rate on the porous structure and gas
separation performance of these mixed matrix foamed membranes (MMFM) were
investigated. The results showed that the introducing of this porous nanoparticle (zeolite
5A) substantially improved the foam cell density (1.2×109 cells/cm3) with a significant
decrease in cell sizes (30 µm). The membrane properties for this optimized MMFM were
found to substantially increase, especially at 15 wt.% zeolite as the H2 permeance, as well
as H2/CH4 and H2/N2 selectivity, were improved by 6.9, 3.8 and 5.9 times respectively,
compared to the unfoamed neat (unfilled) matrix.
To conclude, the findings from this thesis show that the next generation of hollow fiber mixed
matrix membranes might be achieved from further optimization of this new preparation method.
Moreover, combining the benefits of both particle addition and stretching with foaming helped to
generate a multi-porous structure in the membranes to better control the gas transport properties.
Finally, the mixed matrix foam membranes were shown to have higher permeance and selectivity
compared to the unfoamed neat polymer matrix.
134
To our knowledge, no investigation has been performed before on the development of hollow
fiber porous composite polyethylene membranes produced by extrusion foaming followed by
stretching which is a continuous process. Therefore, the original contributions of this work can
be described as:
The production of very thin hollow fiber foamed polymer by adding zeolite as a nucleation
agent.
The development of a continuous, fast and low-cost process for hollow fiber mixed matrix
membrane production.
The steady production of MMM with controlled morphology.
The development of a solvent-free method to produce MMM hollow fiber membranes
with a compact (dense) surface and a multi-porous structure for better gas separation
applications.
Recommendations for future work
Some investigations were performed in this thesis with respect to the production of MMM, but
more investigations are still possible in the future. For further development of these materials,
more work can be made on the following subjects:
In the first part of the study, the CBA content was optimized for a specific grade of ADC.
But other chemical blowing agents could be used to determine their effect on the cellular
structure improvement. In addition, using physical blowing agents (PBA), such as N2 and
supercritical CO2, can be done. Other important process parameters optimization should
include the die (size and configuration) and the cooling time (die-calendar distance and
calendar roll temperature) to further control the cellular morphology. There is also the
possibility to develop a local open-cell structure which would be interesting to get at the
same time as a compact selective layer (foam skin) and a porous support (foam core).
The production can also be performed using other matrices like polyethylene oxide (PEO),
polypropylene (PP), polysulfone (PSF), (PVDF) alone or blended together.
The cellular structure of the MMFM was improved using zeolite as a nucleation agent/gas
permeation modifier. It would be interesting to examine different types of zeolites with
135
different sizes and geometries such as zeolite 4A and 13X. Other porous nanoparticles
(MOF) or non-porous (nanoclays and silica) can be added to get a better control of the
foam permeance.
The work was mainly done for H2/CH4 and H2/N2 separation. It would be interesting to
investigate other gases such as C3H6/C3H8 and C2H4/C2H6.
Finally, modeling would be interesting for design and optimization purposes from both a
foam structure and a membrane separation point of view. Developing some relationships
between all the parameters involved would be of great help for future optimization and
industrial application/scale-up. In particular, more information on thickness (active layer
and cell wall) effect would be important to complement the work.
136
References
[1] T.E. Rufford, S. Smart, G.C.Y. Watson, B.F. Graham, J. Boxall, J.C. Diniz da Costa, E.F.
May, The removal of CO2 and N2 from natural gas: A review of conventional and
emerging process technologies, Journal of Petroleum Science and Engineering. 94–95
(2012) 123–154. doi:10.1016/j.petrol.2012.06.016.
[2] C.A. Scholes, G.W. Stevens, S.E. Kentish, Membrane gas separation applications in
natural gas processing, Fuel. 96 (2012) 15–28. doi:10.1016/j.fuel.2011.12.074.
[3] H. Yang, S. Fan, X. Lang, Y. Wang, J. Nie, Economic Comparison of Three Gas Separation
Technologies for CO2 Capture from Power Plant Flue Gas, Chinese Journal of Chemical
Engineering. 19 (2011) 615–620. doi:10.1016/S1004-9541(11)60031-1.
[4] X.Y. Chen, H. Vinh-Thang, A.A. Ramirez, D. Rodrigue, S. Kaliaguine, Membrane gas
separation technologies for biogas upgrading, RSC Advances. 5 (2015) 24399–24448.
doi:10.1039/C5RA00666J.
[5] F.M. Dautzenberg, M. Mukherjee, Process intensification using multifunctional reactors,
Chemical Engineering Science. 56 (2001) 251–267. doi:10.1016/S0009-2509(00)00228-
1.
[6] P. Bernardo, E. Drioli, G. Golemme, Membrane Gas Separation: A Review / State of the
Art, Industrial & Engineering Chemistry Research. 48 (2009) 4638–4663.
doi:10.1021/ie8019032.
[7] D.S. Sholl, R.P. Lively, Seven chemical separations to change the world, Nature News.
532 (2016) 435.
[8] J.M.S. Henis, M.K. Tripodi, A novel approach to gas separations using composite hollow
fiber membranes, Separation Science and Technology. 15 (1980) 1059–1068.
[9] J.M.S. Henis, M.K. Tripodi, The developing technology of gas separating membranes,
Science. 220 (1983) 11–17.
[10] R.W. Baker, B.T. Low, Gas Separation Membrane Materials: A Perspective,
Macromolecules. 47 (2014) 6999–7013. doi:10.1021/ma501488s.
[11] Y. Yampolskii, Polymeric gas separation membranes, Macromolecules. 45 (2012) 3298–
3311. doi:10.1021/ma300213b.
[12] W.J. Koros, G.K. Fleming, Membrane-based gas separation, Journal of Membrane
Science. 83 (1993) 1–80.
[13] D.R. Paul, Polymeric Gas Separation Membranes, CRC Press (2018).
[14] A.F. Ismail, K.C. Khulbe, T. Matsuura, Gas separation membrane materials and structures,
in: Gas Separation Membranes, Springer, 2015: pp. 37–192.
137
[15] Z. Dai, L. Ansaloni, L. Deng, Recent advances in multi-layer composite polymeric
membranes for CO2 separation: A review, Green Energy & Environment. 1(2) (2016) 102–
128. doi:http://dx.doi.org/10.1016/j.gee.2016.08.001.
[16] J. Mulder, Basic principles of membrane technology, Springer Science & Business Media,
2012.
[17] W.W. Ho, K.K. Sirkar, Membrane handbook part III, Chapman/Hall, New York/London.
(1992).
[18] S. Husain, W.J. Koros, Mixed matrix hollow fiber membranes made with modified HSSZ-
13 zeolite in polyetherimide polymer matrix for gas separation, Journal of Membrane
Science. 288 (2007) 195–207. doi:10.1016/j.memsci.2006.11.016.
[19] Z. Razzaz, A. Mohebbi, D. Rodrigue, Effect of processing conditions on the cellular
morphology of polyethylene hollow fiber foams for membrane applications, Cellular
Polymers. 37 (2018) 169–188.
[20] K.W. Suh, C.P. Park, M.J. Maurer, M.H. Tusim, R. De Genova, R. Broos, D.P. Sophiea,
Lightweight cellular plastics, Advanced Materials. 12 (2000) 1779–1789.
[21] C. Okolieocha, T. Köppl, S. Kerling, F.J. Tölle, A. Fathi, R. Mülhaupt, V. Altstädt,
Influence of graphene on the cell morphology and mechanical properties of extruded
polystyrene foam, Journal of Cellular Plastics. 51 (2015) 413–426.
doi:10.1177/0021955X14566084.
[22] A. Mohebbi, F. Mighri, A. Ajji, D. Rodrigue, Current issues and challenges in
polypropylene foaming: a review, Cellular Polymers. 34 (2015) 299–337.
[23] H. Guo, A. Nicolae, V. Kumar, Solid-state poly (methyl methacrylate)(PMMA)
nanofoams. Part II: Low-temperature solid-state process space using CO2 and the resulting
morphologies, Polymer. 70 (2015) 231–241.
[24] Y. Zhang, D. Rodrigue, A. Ait-Kadi, High-Density Polyethylene Foams. I. Polymer and
Foam Characterization, Journal of Applied Polymer Science. 90 (2003) 2111–2119.
doi:10.1002/app.12821.
[25] C. Saiz-Arroyo, M.Á. Rodríguez-Pérez, J.I. Velasco, J.A. De Saja, Influence of foaming
process on the structure-properties relationship of foamed LDPE/silica nanocomposites,
Composites Part B: Engineering. 48 (2013) 40–50.
doi:10.1016/j.compositesb.2012.10.045.
[26] X.Y. Chen, H. Vinh-Thang, D. Rodrigue, S. Kaliaguine, Effect of macrovoids in nano-
silica/polyimide mixed matrix membranes for high flux CO2/CH4 gas separation, RSC
Advances. 4 (2014) 12235–12244. doi:10.1039/C3RA47208F.
[27] X.Y. Chen, Z. Razzaz, S. Kaliaguine, D. Rodrigue, Mixed matrix membranes based on
silica nanoparticles and microcellular polymers for CO2/CH4separation, Journal of
138
Cellular Plastics. 54 (2018). doi:10.1177/0021955X16681453.
[28] R. Chen, L., Rende, D., Schadler, L.S. and Ozisik, Polymer nanocomposite foams, Journal
of Materials Chemistry A. 1 (2013) 3837–3850 |. doi:10.1016/j.compscitech.2005.06.016.
[29] D.I. Collais, D.G. Baird, Tensile toughness of microcellular foams of polystyrene, styrene-
acrylonitrile copolymer, and polycarbonate, and the effect of dissolved gas on the tensile
toughness of the same polymer matrices and microcellular foams, Polymer Engineering &
Science. 35 (1995) 1167–1177.
[30] M. Shimbo, I. Higashitani, Y. Miyano, Mechanism of strength improvement of foamed
plastics having fine cell, Journal of Cellular Plastics. 43 (2007) 157–167.
[31] J.B. Bao, A. Nyantakyi Junior, G.S. Weng, J. Wang, Y.W. Fang, G.H. Hu, Tensile and
impact properties of microcellular isotactic polypropylene (PP) foams obtained by
supercritical carbon dioxide, Journal of Supercritical Fluids. 111 (2016) 63–73.
doi:10.1016/j.supflu.2016.01.016.
[32] Technavio, Polymer Foam Market Will Have a Steady Growth of 8% Until 2021, Says
Technavio, Sample-Global Polymer Foam Market. (2016).
[33] S. Nagō, S. Nakamura, Y. Mizutani, Structure of microporous polypropylene sheets
containing CaCO3 filler, Journal of Applied Polymer Science. 45 (1992) 1527–1535.
[34] J.-J. Kim, T.-S. Jang, Y.-D. Kwon, U.Y. Kim, S.S. Kim, Structural study of microporous
polypropylene hollow fiber membranes made by the melt-spinning and cold-stretching
method, Journal of Membrane Science. 93 (1994) 209–215.
[35] M. Xanthos, C. Chandavasu, K.K. Sirkar, C.G. Gogos, Melt processed microporous films
from compatibilized immiscible blends with potential as membranes, Polymer Engineering
& Science. 42 (2002) 810–825.
[36] G. Shi, F. Chu, G. Zhou, Z. Han, Plastic deformation and solid‐ phase transformation in
β‐ phase polypropylene, Die Makromolekulare Chemie: Macromolecular Chemistry and
Physics. 190 (1989) 907–913.
[37] A. Peterlin, J. Williams, H. Olf, Process for the preparation of opencelled microporous
films, (1974) U.S. Patent No. 3,839,516. Washington, DC: U.S. Pa.
[38] S.H. Tabatabaei, P.J. Carreau, A. Ajji, Microporous membranes obtained from
polypropylene blend films by stretching, Journal of Membrane Science. 325 (2008) 772–
782. doi:10.1016/j.memsci.2008.09.001.
[39] F. Sadeghi, A. Ajji, P.J. Carreau, Analysis of microporous membranes obtained from
polypropylene films by stretching, Journal of Membrane Science. 292 (2007) 62–71.
doi:10.1016/j.memsci.2007.01.023.
[40] Z. Xi, Y. Xu, L. Zhu, C. Du, B. Zhu, Effect of stretching on structure and properties of
139
polyethylene hollow fiber membranes made by melt-spinning and stretching process,
Polymers for Advanced Technologies. 19 (2008) 1616–1622. doi:10.1002/pat.
[41] T.-H. Yu, Processing and structure-property behavior of microporous polyethylene: from
resin to final film, (1996) (Doctoral dissertation, Virginia Tech).
[42] S.H. Tabatabaei, P.J. Carreau, A. Ajji, Microporous membranes obtained from PP/HDPE
multilayer films by stretching, Journal of Membrane Science. 345 (2009) 148–159.
doi:10.1016/j.memsci.2009.08.038.
[43] J. Biswas, H. Kim, S. Choe, P.P. Kundu, Y.-H. Park, D.S. Lee, Linear low density
polyethylene (LLDPE)/zeolite microporous composite film, Macromolecular Research. 11
(2003) 357–367. doi:10.1007/BF03218377.
[44] L.M. Robeson, Correlation of separation factor versus permeability for polymeric
membranes, Journal of Membrane Science. 62 (1991) 165–185. doi:10.1016/0376-
7388(91)80060-J.
[45] L.M. Robeson, The upper bound revisited, Journal of Membrane Science. 320 (2008) 390–
400. doi:10.1016/j.memsci.2008.04.030.
[46] F. Peng, L. Lu, H. Sun, Y. Wang, J. Liu, Z. Jiang, Hybrid Organic−Inorganic
Membrane: Solving the Tradeoff between Permeability and Selectivity, Chemistry of
Materials. 17 (2005) 6790–6796. doi:10.1021/cm051890q.
[47] F. Peng, L. Lu, H. Sun, Y. Wang, H. Wu, Z. Jiang, Correlations between free volume
characteristics and pervaporation permeability of novel PVA-GPTMS hybrid membranes,
Journal of Membrane Science. 275 (2006) 97–104. doi:10.1016/j.memsci.2005.09.008.
[48] S.J. Miller, A. Kuperman, D.Q. Vu, Mixed matrix membranes with small pore molecular
sieves and methods for making and using the membranes, (2007) U.S. Patent No.
7,166,146. Washington, DC: U.S. Pa.
[49] D.R. Kemp, D.R. Paul, Gas sorption in polymer membranes containing adsorptive fillers,
Journal of Polymer Science: Polymer Physics Edition. 12 (1974) 485–500.
[50] X.Y. Chen, H. Vinh-Thang, D. Rodrigue, S. Kaliaguine, Amine-Functionalized MIL-53
Metal − Organic Framework in Polyimide Mixed Matrix Membranes for CO2/CH4
Separation, Industrial & Engineering Chemistry Research, 51 (2012) 6895-6906.
[51] C.Y. Feng, K.C. Khulbe, T. Matsuura, A.F. Ismail, Recent progresses in polymeric hollow
fiber membrane preparation, characterization and applications, Separation and Purification
Technology. 111 (2013) 43–71. doi:10.1016/j.seppur.2013.03.017.
[52] T.C. Merkel, B.D. Freeman, R.J. Spontak, Z. He, I. Pinnau, P. Meakin, J. Hill,
Ultrapermeable, reverse-selective nanocomposite membranes., Science (New York). 296
(2002) 519–522. doi:10.1126/science.1069580.
140
[53] M. Rezakazemi, A.E. Amooghin, M.M. Montazer-Rahmati, A.F. Ismail, T. Matsuura,
State-of-the-art membrane based CO2 separation using mixed matrix membranes
(MMMs): an overview on current status and future directions, Progress in Polymer
Science. 39 (2014) 817–861. doi:10.1016/j.polymer.2017.09.053.
[54] G. Dong, H. Li, V. Chen, Challenges and opportunities for mixed-matrix membranes for
gas separation, Journal of Materials Chemistry A. 1 (2013) 4610–4630.
doi:10.1039/C3TA00927K.
[55] S. Kulprathipanja, R.W. Neuzil, N.N. Li, Separation of gases by means of mixed matrix
membranes, (1992) U.S. Patent 5,127,925.
[56] N. Du, G.P. Robertson, J. Song, I. Pinnau, M.D. Guiver, High-performance carboxylated
polymers of intrinsic microporosity (PIMs) with tunable gas transport properties,
Macromolecules. 42 (2009) 6038–6043.
[57] N. Du, H.B. Park, G.P. Robertson, M.M. Dal-Cin, T. Visser, L. Scoles, M.D. Guiver,
Polymer nanosieve membranes for CO2-capture applications, Nature Materials. 10 (2011)
372.
[58] A.F. Ismail, P.S. Goh, S.M. Sanip, M. Aziz, Transport and separation properties of carbon
nanotube-mixed matrix membrane, Separation and Purification Technology. 70 (2009) 12–
26. doi:10.1016/j.seppur.2009.09.002.
[59] A.W. Thornton, D. Dubbeldam, M.S. Liu, B.P. Ladewig, A.J. Hill, M.R. Hill, Feasibility
of zeolitic imidazolate framework membranes for clean energy applications, Energy &
Environmental Science. 5 (2012) 7637–7646.
[60] L.Y. Jiang, T.S. Chung, S. Kulprathipanja, Fabrication of mixed matrix hollow fibers with
intimate polymer–zeolite interface for gas separation, AIChE Journal. 52 (2006) 2898–
2908.
[61] 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) 28–37.
[62] A. Javaid, Membranes for solubility-based gas separation applications, Chemical
Engineering Journal. 112 (2005) 219–226.
[63] X.Y. Chen, S. Kaliaguine, Mixed gas and pure gas transport properties of copolyimide
membranes, Journal of Applied Polymer Science. 128 (2013) 380–389.
doi:10.1002/app.37728.
[64] K. Ghosal, B. D. Freeman, Gas separation using polymer membranes: an overview,
Polymer for Advanced Technologies. 5 (1994) 673–697.
[65] X.Y. Chen, O.G. Nik, D. Rodrigue, S. Kaliaguine, Mixed matrix membranes of
141
aminosilanes grafted FAU/EMT zeolite and cross-linked polyimide for CO2/CH4
separation, Polymer (United Kingdom). 53 (2012) 3269–3280.
doi:10.1016/j.polymer.2012.03.017.
[66] W.J. Koros, G.K. Fleming, S.M. Jordan, T.H. Kim, H.H. Hoehn, Polymeric membrane
materials for solution-diffusion based permeation separations, Progress in Polymer
Science. 13 (1988) 339–401. doi:10.1016/0079-6700(88)90002-0.
[67] B. Freeman,Y. Yampolskii, I. Pinnau, Materials Science of Membranes for Gas and Vapor
Separation, John Wiley & Sons. (2006) doi:10.1002/047002903X.
[68] G. V Sengbusch, Future of membranes, technological and economical aspects, Vortrag:
Synthetic Membranes in Science and Industry, Tüebingen. (1994).
[69] H.B. Park, C.H. Jung, Y.M. Lee, A.J. Hill, S.J. Pas, Polymers with cavities tuned for fast
selective transport of small molecules and ions, Science. 318 (2007) 254–258.
[70] M. Hedenqvist, A. Angelstok, L. Edsberg, P.T. Larssont, U.W. Gedde, Diffusion of small-
molecule penetrants in polyethylene: Free volume and morphology, Polymer. 37 (1996)
2887–2902. doi:10.1016/0032-3861(96)89384-0.
[71] S.B. Lee, D.T. Mitchell, L. Trofin, T.K. Nevanen, H. Söderlund, C.R. Martin, Antibody-
based bio-nanotube membranes for enantiomeric drug separations, Science. 296 (2002)
2198–2200.
[72] H. Yamamoto, Y. Mi, S.A. Stern, A.K. St Clair, Structure/permeability relationships of
polyimide membranes. II, Journal of Polymer Science Part B: Polymer Physics. 28 (1990)
2291–2304.
[73] S.S. Ozdemir, M.G. Buonomenna, E. Drioli, Catalytic polymeric membranes: preparation
and application, Applied Catalysis A: General. 307 (2006) 167–183.
[74] B. Freeman, I. Pinnau, Separation of gases using solubility-selective polymers, Trends in
Polymer Science. 5 (1997) 167–173.
[75] Z. He, I. Pinnau, A. Morisato, Nanostructured poly (4-methyl-2-pentyne)/silica hybrid
membranes for gas separation, Desalination. 146 (2002) 11–15.
[76] P. Winberg, K. DeSitter, C. Dotremont, S. Mullens, I.F.J. Vankelecom, F.H.J. Maurer,
Free volume and interstitial mesopores in silica filled poly (1-trimethylsilyl-1-propyne)
nanocomposites, Macromolecules. 38 (2005) 3776–3782.
[77] T.C. Merkel, Z. He, I. Pinnau, B.D. Freeman, P. Meakin, A.J. Hill, Sorption and transport
in poly(2,2-bis(trifluoromethyl)-4,5-difluoro-1,3-dioxole-co-tetrafluoroethylene)
containing nanoscale fumed silica, Macromolecules. 36 (2003) 8406–8414.
[78] W.F. Yong, K.H.A. Kwek, K.-S. Liao, T.-S. Chung, Suppression of aging and
plasticization in highly permeable polymers, Polymer. 77 (2015) 377–386.
142
[79] J. Ahn, W.-J. Chung, I. Pinnau, J. Song, N. Du, G.P. Robertson, M.D. Guiver, Gas transport
behavior of mixed-matrix membranes composed of silica nanoparticles in a polymer of
intrinsic microporosity (PIM-1), Journal of Membrane Science. 346 (2010) 280–287.
[80] P.M. Budd, N.B. McKeown, D. Fritsch, Free volume and intrinsic microporosity in
polymers, Journal of Materials Chemistry. 15 (2005) 1977–1986. doi:10.1039/b417402j.
[81] N. Petzetakis, C.M. Doherty, A.W. Thornton, X.C. Chen, P. Cotanda, A.J. Hill, N.P.
Balsara, Membranes with artificial free-volume for biofuel production, Nature
Communications Hierarchical Nanofabrication of Microporous Crystals. 6 (2015) 7529.
doi:10.1038/ncomms8529.
[82] W. Fan, M. Snyder, S. Kumar, P.-S. Lee, W.C. Yoo, A.V. McCormick, R. Lee Penn, A.
Stein, M. Tsapatsis, Hierarchical nanofabrication of microporous crystals with ordered
mesoporosity, Nature Materials. 7 (2008) 984–991. doi:10.1038/nmat2302.
[83] Y.H. Kim, H. Kang, S. Park, A.R. Park, Y.M. Lee, D.K. Rhee, S. Han, H. Chang, D.Y.
Ryu, P.J. Yoo, Multiscale porous interconnected Nanocolander network with tunable
transport properties, Advanced Materials. 26 (2014) 7998–8003.
doi:10.1002/adma.201402436.
[84] Y.M. Lee, B. Jung, Y.H. Kim, A.R. Park, S. Han, W. Choe, P.J. Yoo, Nanomesh-
Structured Ultrathin Membranes Harnessing the Unidirectional Alignment of Viruses on a
Graphene-Oxide Film, Advanced Materials. 26 (2014) 3899–3904.
[85] Y. Zhuang, J.G. Seong, Y.S. Do, W.H. Lee, M.J. Lee, Z. Cui, A.E. Lozano, M.D. Guiver,
Y.M. Lee, Soluble, microporous, Troger’s Base copolyimides with tunable membrane
performance for gas separation, Chemical Communications. 52 (2016) 3817–3820.
doi:10.1039/C5CC09783E.
[86] H. Cong, X. Hu, M. Radosz, Y. Shen, Brominated poly (2,6-diphenyl-1,4-phenylene
oxide) and its silica nanocomposite membranes for gas separation, Industrial &
Engineering Chemistry Research. 46 (2007) 2567–2575.
[87] S. Loeb, S. Sourirajan, Sea water demineralization by means of an osmotic membrane,
University of California, Department of Engineering, 1960.
[88] A. Figoli, T. Marino, S. Simone, E. Di Nicolò, X.-M. Li, T. He, S. Tornaghi, E. Drioli,
Towards non-toxic solvents for membrane preparation: a review, Green Chemistry. 16
(2014) 4034. doi:10.1039/C4GC00613E.
[89] K. Twarowska-Schmidt, A. Włochowicz, Melt-spun asymmetric poly (4-methyl-1-
pentene) hollow fibre membranes, Journal of Membrane Science. 137 (1997) 55–61.
[90] P.S. Puri, Fabrication of hollow fibre gas separation membranes, Gas Separation &
Purification. 4 (1990) 29–36.
[91] B.S. Lalia, V. Kochkodan, R. Hashaikeh, N. Hilal, A review on membrane fabrication:
143
Structure, properties and performance relationship, Desalination. 326 (2013) 77–95.
[92] S.H. Tabatabaei, P.J. Carreau, A. Ajji, Effect of processing on the crystalline orientation,
morphology, and mechanical properties of polypropylene cast films and microporous
membrane formation, Polymer. 50 (2009) 4228–4240.
doi:10.1016/j.polymer.2009.06.071.
[93] F.J. Wei, H.J. Shao, B. Wu, K.Z. Zhang, D.J. Luo, S.H. Qin, Z. Hao, Effect of spin-draw
rate and stretching ratio on polypropylene hollow fiber membrane made by melt-spinning
and stretching method, International Polymer Processing. 33 (2018) 13–19.
doi:10.3139/217.3303.
[94] A. Saffar, P.J. Carreau, A. Ajji, M.R. Kamal, Influence of stretching on the performance
of polypropylene-based microporous membranes, Industrial and Engineering Chemistry
Research. 53 (2014) 14014–14021. doi:10.1021/ie502300j.
[95] C. Du, Y. Xu, B. Zhu, Structure formation and characterization of PVDF hollow fiber
membranes by melt‐ spinning and stretching method, Journal of Applied Polymer Science.
106 (2007) 1793–1799.
[96] J. Kim, S. Soo, M. Park, M. Jang, Effects of precursor properties on the preparation of
polyethylene hollow fiber membranes by stretching, Journal of Membrane Science. 318
(2008) 201–209. doi:10.1016/j.memsci.2008.02.050.
[97] L. Shen, Z. Xu, Y. Xu, Preparation and characterization of microporous polyethylene
hollow fiber membranes, Journal of Applied Polymer Science. 84 (2002) 203–210.
[98] R.W. Gore, Process for producing porous products, (1976) U.S. Patent No. 3,953,566.
Washington, DC: U.S. Pa.
[99] G.C. Kapantaidakis, G.H. Koops, M. Wessling, Effect of spinning conditions on the
structure and the gas permeation properties of high flux polyethersulfone—polyimide
blend hollow fibers, Desalination. 144 (2002) 121–125.
[100] X. Liu, B. Cao, P. Li, Effects of Spinning Temperature on the Morphology and
Performance of Poly (ether sulfone) Gas Separation Hollow Fiber Membranes, Industrial
& Engineering Chemistry Research. 57 (2017) 329–338.
[101] M. Gholizadeh, J. Razavi, S.A. Mousavi, Gas permeability measurement in polyethylene
and its copolymer films, Materials & Design. 28 (2007) 2528–2532.
doi:10.1016/j.matdes.2006.09.018.
[102] Z. Ye, Y. Chen, H. Li, G. He, M. Deng, Preparation of a novel polysulfone/polyethylene
oxide/silicone rubber multilayer composite membrane for hydrogen–nitrogen separation,
Materials Chemistry and Physics. 94 (2005) 288–291.
doi:10.1016/j.matchemphys.2005.05.001.
[103] A.K. Zulhairun, B.C. Ng, A.F. Ismail, R. Surya Murali, M.S. Abdullah, Production of
144
mixed matrix hollow fiber membrane for CO2/CH4 separation, Separation and Purification
Technology. 137 (2014) 1–12. doi:10.1016/j.seppur.2014.09.014.
[104] H. Julian, I.G. Wenten, Polysulfone membranes for CO2/CH4 separation: State of the art,
Journal of Engineering. 2 (2012) 484–495.
[105] S. Zhao, X. Cao, Z. Ma, Z. Wang, Z. Qiao, J. Wang, S. Wang, Mixed-matrix membranes
for CO2/N2 separation comprising a poly (vinylamine) matrix and metal–organic
frameworks, Industrial & Engineering Chemistry Research. 54 (2015) 5139–5148.
[106] N. Peng, T.S. Chung, The effects of spinneret dimension and hollow fiber dimension on
gas separation performance of ultra-thin defect-free Torlon® hollow fiber membranes,
Journal of Membrane Science. 310 (2008) 455–465.
[107] S.J. Shilton, Forced convection spinning of hollow fibre membranes: Modelling of mass
transfer in the dry gap, and prediction of active layer thickness and depth of orientation,
Separation and Purification Technology. 118 (2013) 620–626.
[108] D.W. Wallace, C. Staudt-Bickel, W.J. Koros, Efficient development of effective hollow
fiber membranes for gas separations from novel polymers, Journal of Membrane Science.
278 (2006) 92–104.
[109] S.-H. Choi, F. Tasselli, J.C. Jansen, G. Barbieri, E. Drioli, Effect of the preparation
conditions on the formation of asymmetric poly (vinylidene fluoride) hollow fibre
membranes with a dense skin, European Polymer Journal. 46 (2010) 1713–1725.
[110] A.F. Ismail, S.J. Shilton, I.R. Dunkin, S.L. Gallivan, Direct measurement of rheologically
induced molecular orientation in gas separation hollow fibre membranes and effects on
selectivity, Journal of Membrane Science. 126 (1997) 133–137.
[111] O.M. Ekiner, G. Vassilatos, Polymeric membranes, (1992) U.S. Patent 5,085,774.
[112] W.F. Yong, F.Y. Li, Y.C. Xiao, T.S. Chung, Y.W. Tong, High performance PIM-
1/Matrimid hollow fiber membranes for CO2/CH4, O2/N2 and CO2/N2 separation,
Journal of Membrane Science. 443 (2013) 156–169.
[113] Y.E. Santoso, T.S. Chung, K.Y. Wang, M. Weber, The investigation of irregular inner skin
morphology of hollow fiber membranes at high-speed spinning and the solutions to
overcome it, Journal of Membrane Science. 282 (2006) 383–392.
[114] O.M. Ekiner, G. Vassilatos, Polyaramide hollow fibers for H2/CH4 separation: II.
Spinning and properties, Journal of Membrane Science. 186 (2001) 71–84.
[115] J. Sa-nguanruksa, R. Rujiravanit, P. Supaphol, S. Tokura, Porous polyethylene membranes
by template-leaching technique: preparation and characterization, Polymer Testing. 23
(2004) 91–99.
[116] P. Roy Chowdhury, V. Kumar, Fabrication and evaluation of porous 2,3‐
145
dialdehydecellulose membrane as a potential biodegradable tissue-engineering scaffold,
Journal of Biomedical Materials Research Part A: An Official Journal of The Society for
Biomaterials, The Japanese Society for Biomaterials, and The Australian Society for
Biomaterials and the Korean Society for Biomaterials. 76 (2006) 300–309.
[117] S. Mosadegh-Sedghi, J. Brisson, D. Rodrigue, M.C. Iliuta, Morphological, chemical and
thermal stability of microporous LDPE hollow fiber membranes in contact with single and
mixed amine based CO2 absorbents, Separation and Purification Technology. 96 (2012)
117–123. doi:10.1016/j.seppur.2012.05.025.
[118] G. Decher, Fuzzy nanoassemblies: toward layered polymeric multicomposites, Science.
277 (1997) 1232–1237.
[119] U. Sampath, Y. Ching, C. Chuah, J. Sabariah, P.-C. Lin, Fabrication of porous materials
from natural/synthetic biopolymers and their composites, Materials. 9 (2016) 991.
[120] M. Narkis, E. Joseph, Tensile Properties of Rigid Polymeric Foams Produced by Salt
Extraction, Journal of Cellular Plastics. 14 (1978) 45–49.
doi:10.1177/0021955X7801400107.
[121] L. Verdolotti, S. Colini, G. Porta, S. Iannace, Effects of the addition of LiCl, LiClO4, and
LiCF3SO3 salts on the chemical structure, density, electrical, and mechanical properties
of rigid polyurethane foam composite, Polymer Engineering & Science. 51 (2011) 1137–
1144. doi:10.1002/pen.21846.
[122] K. Shahidi, D. Rodrigue, Production of Composite Membranes by Coupling Coating and
Melt Extrusion/Salt Leaching, Industrial and Engineering Chemistry Research. 56 (2017)
1306–1315. doi:10.1021/acs.iecr.6b04362.
[123] S. Berghmans, H. Berghmans, H.E.H. Meijer, Spinning of hollow porous fibres via the
TIPS mechanism, Journal of Membrane Science. 116 (1996) 171–189.
[124] J.R. Werber, C.O. Osuji, M. Elimelech, Materials for next-generation desalination and
water purification membranes, Nature Reviews Materials. 1 (2016) 16018.
[125] J. Lee, B. Park, J. Kim, S. Bin Park, Effect of PVP, lithium chloride, and glycerol additives
on PVDF dual-layer hollow fiber membranes fabricated using simultaneous spinning of
TIPS and NIPS, Macromolecular Research. 23 (2015) 291–299.
[126] C. V. Funk, D.R. Lloyd, Zeolite-filled microporous mixed matrix (ZeoTIPS) membranes:
Prediction of gas separation performance, Journal of Membrane Science. 313 (2008) 224–
231. doi:10.1016/j.memsci.2008.01.002.
[127] J.T. Chen, C.C. Shih, Y.J. Fu, S.H. Huang, C.C. Hu, K.R. Lee, J.Y. Lai, Zeolite-filled
porous mixed matrix membranes for air separation, Industrial and Engineering Chemistry
Research. 53 (2014) 2781–2789. doi:10.1021/ie403833u.
[128] Y. Li, T. Chung, S. Kulprathipanja, Novel Ag+‐ zeolite/polymer mixed matrix membranes
146
with a high CO2/CH4 selectivity, AIChE Journal. 53 (2007) 610–616.
[129] S.-T. Lee and Chul B. Park, Foam extrusion: principles and practice, CRC press. 2014.
[130] J.E. Martini, The production and analysis of microcellular foam. Diss. Massachusetts
Institute of Technology, 1981.
[131] C.B. Park, L.K. Cheung, A study of cell nucleation in the extrusion of polypropylene
foams, Polymer Engineering & Science. 37 (1997) 1–10.
[132] J.S. Colton, The nucleation of microcellular foams in semi crystalline thermoplastics,
Material and Manufacturing Process. 4 (1989) 253–262.
[133] J.S. Colton, N.P. Suh, Microcellular foams of semi-crystaline polymeric materials, (1992)
U.S. Patent 5,160,674.
[134] S. Doroudiani, C.B. Park, M.T. Kortschot, Effect of the crystallinity and morphology on
the microcellular foam structure of semicrystalline polymers, Polymer Engineering &
Science. 36 (1996) 2645–2662.
[135] D.D. Imeokparia, C.D. Shmidt, K.W. Suh, Extruded, open-cell alkenyl aromatic polymer
foam and process for making, (1995) U.S. Patent 5,411,687.
[136] D. Feldman, Integral/structural polymer foams. Technology, properties and applications,
Journal of Polymer Science Part C: Polymer Letters. 24 (1986) 657.
doi:10.1080/03003930500157457.
[137] D. Klempner, V. Sendijareviʹc, R.M. Aseeva, Handbook of Polymeric Foams and Foam
Technology, Hanser Publishers, 2004. https://books.google.ca/books?id=hlIfwH8cDzQC.
[138] D.A. Porter, K.E. Easterling, M. Sherif, Phase Transformations in Metals and Alloys,
(Revised Reprint), CRC press, 2009.
[139] J.S. Colton, N.P. Suh, Nucleation of Microcellular Thermoplastic Foam With Additives:
Part I: Theoretical Considerations, Polymer Engineering and Science. 27 (1987) 485–492.
doi:10.1002/pen.760270702.
[140] M.R. Barzegari, J. Yao, D. Rodrigue, Mechanical properties of density graded foams:
Tensile properties, Cellular Polymers. 32 (2013) 323–342.
[141] D. Eaves, Handbook of polymer foams, Polimeri. 25 (2004) 1–2.
[142] A.K. Błędzki, O. Faruk, H. Kirschling, J. Kühn, A. Jaszkiewicz, Microcellular polymers
and composites, Polimery. 51 (2006).
[143] J. Yao, D. Rodrigue, Density graded polyethylene foams produced by compression
moulding using a chemical blowing agent, Cellular Polymers. 31 (2012) 189–206.
[144] C.H. Lee, K.J. Lee, H.G. Jeong, S.W. Kim, Growth of gas bubbles in the foam extrusion
147
process, Advances in Polymer Technology. 19 (2000) 97–112. doi:10.1002/(SICI)1098-
2329(200022)19:2<97::AID-ADV3>3.0.CO;2-B.
[145] Y. Zhang, D. Rodrigue, A. Ait-Kadi, High Density Polyethylene Foams. II. Elastic
Modulus, Journal of Applied Polymer Science. 90 (2003) 2120–2129.
doi:10.1002/app.12822.
[146] C.Z. Sahagun, R. Gonzalez-Nunez, D. Rodrigue, Morphology of Extruded PP/HDPE
Foam Blends, Journal of Cellular Plastics. 42 (2006) 469–485.
doi:10.1177/0021955X06063521.
[147] A. Mohebbi, A. Mehrabani-Zeinabad, M. Navid-Famili, Dynamic behavior of nucleation
in supercritical N2 foaming of polystyrene-aluminum oxide nanocomposite, Polymer
Science Series A. 53 (2011) 1076–1085. doi:10.1134/S0965545X11110071.
[148] C.A. Eckert, B.L. Knutson, P.G. Debenedetti, Supercritical fluids as solvents for chemical
and materials processing, Nature. 383 (1996) 313–318. doi:10.1038/383313a0.
[149] S.P. Nalawade, F. Picchioni, L.P.B.M. Janssen, Supercritical carbon dioxide as a green
solvent for processing polymer melts: Processing aspects and applications, Progress in
Polymer Science (Oxford). 31 (2006) 19–43. doi:10.1016/j.progpolymsci.2005.08.002.
[150] M. Sauceau, J. Fages, A. Common, C. Nikitine, E. Rodier, New challenges in polymer
foaming: A review of extrusion processes assisted by supercritical carbon dioxide,
Progress in Polymer Science (Oxford). 36 (2011) 749–766.
doi:10.1016/j.progpolymsci.2010.12.004.
[151] N. Zong, V. Yang, Cryogenic fluid jets and mixing layers in transcritical and supercritical
environments, Combustion Science and Technology. 178 (2006) 193–227.
[152] A. Mohebbi, F. Mighri, A. Ajji, D. Rodrigue, Effect of Processing Conditions on the
Cellular Morphology of Polypropylene Foamed Films for Piezoelectric Applications,
Cellular Polymers. 36 (2017) 13. doi:10.1177/026248931703600102.
[153] M. Antunes, G. Gedler, J.I. Velasco, Multifunctional nanocomposite foams based on
polypropylene with carbon nanofillers, Journal of Cellular Plastics. 49 (2013) 259–279.
doi:10.1177/0021955X13477433.
[154] L.J. Lee, C. Zeng, X. Cao, X. Han, J. Shen, G. Xu, Polymer nanocomposite foams,
Composites Science and Technology. 65 (2005) 2344–2363.
doi:10.1016/j.compscitech.2005.06.016.
[155] O. Almanza, M.A. Rodríguez-Pérez, B. Chernev, J.A. De Saja, P. Zipper, Comparative
study on the lamellar structure of polyethylene foams, European Polymer Journal. 41
(2005) 599–609. doi:10.1016/j.eurpolymj.2004.10.028.
[156] H.-X. Huang, J.-K. Wang, X.-H. Sun, Improving of Cell Structure of Microcellular Foams
Based on Polypropylene/High-density Polyethylene Blends, Journal of Cellular Plastics.
148
44 (2008) 69–85. doi:10.1177/0021955X07086082.
[157] X.L. Jiang, T. Liu, Z.M. Xu, L. Zhao, G.H. Hu, W.K. Yuan, Effects of crystal structure on
the foaming of isotactic polypropylene using supercritical carbon dioxide as a foaming
agent, Journal of Supercritical Fluids. 48 (2009) 167–175.
doi:10.1016/j.supflu.2008.10.006.
[158] Y.H. Lee, C.B. Park, K.H. Wang, M.H. Lee, HDPE-clay nanocomposite foams blown with
supercritical CO2, Journal of Cellular Plastics. 41 (2005) 487–502. doi:Doi
10.1177/0021955x05056964.
[159] J. Gibbs, The scientific papers of JW Gibbs, (1961), Dover, New York.
[160] S.N. Leung, A. Wong, Q. Guo, C.B. Park, J.H. Zong, Change in the critical nucleation
radius and its impact on cell stability during polymeric foaming processes, Chemical
Engineering Science. 64 (2009) 4899–4907. doi:10.1016/j.ces.2009.07.025.
[161] R.B. McClurg, Design criteria for ideal foam nucleating agents, Chemical Engineering
Science. 59 (2004) 5779–5786. doi:10.1016/j.ces.2004.06.025.
[162] D. Tejeda, E. H., Sahagún, C. Z., González-Núñez, R., & Rodrigue, Morphology and
Mechanical Properties of Foamed Polyethylene-Polypropylene Blends, Journal of Cellular
Plastics. 41 (2005) 417–435. doi:10.1177/0021955X05056959.
[163] S.M. Seraji, M.K. Razavi Aghjeh, M. Davari, M. Salami Hosseini, Sh. Khelgati, Effect of
Clay Dispersion on the Cell Structure of LDPE/Clay Nanocomposite Foams, Polymer
Composites. 32 (2011) 1095-1105. doi:10.1002/pc.
[164] W. Zhai, T. Kuboki, L. Wang, C.B. Park, E.K. Lee, H.E. Naguib, Cell structure evolution
and the crystallization behavior of polypropylene/clay nanocomposites foams blown in
continuous extrusion, Industrial & Factors Controlling Successful Formation of Mixed-
Matrix Gas Engineering Chemistry Research. 49 (2010) 9834–9845.
doi:10.1021/ie101225f.
[165] J.M.E. Mohammadhadi Abbasi, S. Nouri Khorasani, R. Bagheri, J. Esfahani, Microcellular
foaming of low-density polyethylene using nano-CaCO3 as a nucleating agent, Polymer
Composites. 32 (2011) 1718–1725. doi:10.1002/pc.
[166] E. Laguna-Gutierrez, C. Saiz-Arroyo, J.I. Velasco, M.A. Rodriguez-Perez, Low density
polyethylene/silica nanocomposite foams. Relationship between chemical composition,
particle dispersion, cellular structure and physical properties, European Polymer Journal.
81 (2016) 173–185. doi:10.1016/j.eurpolymj.2016.06.001.
[167] G. Zhang, S. Zhang, J. Qiu, Z. Jiang, H. Xing, M. Li, T. Tang, Insight into the influence of
OA-Fe3O4 nanoparticles on the morphology and scCO2 batch-foaming behavior of
cocontinuous LLDPE/PS immiscible blends at semi-solid state, Polymer (United
Kingdom). 129 (2017) 169–178. doi:10.1016/j.polymer.2017.09.053.
149
[168] P. Zhang, X. Wang, Y. Yang, N. Zhou, Effect of dynamic shear on the microcellular
foaming of polypropylene/high‐ density polyethylene blends, Journal of Applied Polymer
Science. 114 (2009) 1320–1328.
[169] P. Shahi, A.H. Behravesh, A. Haghtalab, G. Rizvi, F. Goharpei, An experimental study on
foaming of linear low-density polyethylene/high-density polyethylene blends, Journal of
Cellular Plastics. (2016) 1–23. doi:10.1177/0021955X16639033.
[170] J. Ding, J. Shangguan, W. Ma, Q. Zhong, Foaming behavior of microcellular foam
polypropylene/modified nano calcium carbonate composites, Journal of Applied Polymer
Science. 128 (2013) 3639–3651.
[171] W.G. Zheng, Y.H. Lee, C.B. Park, Use of nanoparticles for improving the foaming
behaviors of linear PP, Journal of Applied Polymer Science. 117 (2010) 2972–2979.
[172] S.T. Lee, K. Lee, Surrounding temperature effects on extruded polyethylene foam
structure, Advances in Polymer Technology. 19 (2000) 87–96. doi:10.1002/(SICI)1098-
2329(200022)19:2<87::AID-ADV2>3.0.CO;2-F.
[173] J. Ding, W. Ma, F. Song, Q. Zhong, Foaming of polypropylene with supercritical carbon
dioxide: An experimental and theoretical study on a new process, Journal of Applied
Polymer Science. 130 (2013) 2877–2885.
[174] C.B. Park, D.F. Baldwin, N.P. Suh, Effect of the pressure drop rate on cell nucleation in
continuous processing of microcellular polymers, Polymer Engineering & Science. 35
(1995) 432–440.
[175] P. Spitael, C.W. Macosko, Strain hardening in polypropylenes and its role in extrusion
foaming, Polymer Engineering & Science. 44 (2004) 2090–2100.
[176] C. Wang, S. Ying, Z. Xiao, Preparation of short carbon fiber/polypropylene fine-celled
foams in supercritical CO2, Journal of Cellular Plastics. 49 (2013) 65–82.
[177] P. Luis, T. Van Gerven, B. Van Der Bruggen, Recent developments in membrane-based
technologies for CO2 capture, Progress in Energy and Combustion Science. 38 (2012)
419–448. doi:10.1016/j.pecs.2012.01.004.
[178] C.M. Zimmerman, A. Singh, W.J. Koros, Tailoring mixed matrix composite membranes
for gas separations, Journal of Membrane Science. 137 (1997) 145–154.
doi:10.1016/S0376-7388(97)00194-4.
[179] R. Mahajan, W.J. Koros, Factors controlling successful formation of mixed-matrix gas
separation materials, Industrial and Engineering Chemistry Research. 39 (2000) 2692–
2696.
[180] S.A. Stern, Polymers for gas separations: the next decade, Journal of Membrane Science.
94 (1994) 1–65.
150
[181] 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. 32 (2007) 483–507. doi:10.1016/j.progpolymsci.2007.01.008.
[182] M. Freemantle, Membranes for gas separation, Chemical Engineering News. 83 (2005) 3.
[183] 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, Journal of Membrane Science. 252 (2005) 89–
100. doi:10.1016/j.memsci.2004.12.004.
[184] H. Matsuyama, K. Hayashi, T. Maki, M. Teramoto, N. Kubota, Effect of polymer density
on polyethylene hollow fiber membrane formation via thermally induced phase separation,
Journal of Applied Polymer Science. 93 (2004) 471–474. doi:10.1002/app.20461.
[185] A.S. Michaels, H.J. Bixler, Flow of gases through polyethylene, Journal of Polymer
Science. 50 (1961) 413–439. doi:10.1002/pol.1961.1205015412.
[186] H. Alter, U. Carbide, P. Company, B. Brook, A Critical Investigation of Polyethylene Gas
Permeability, Journal of Polymer Science. 57 (1962) 925–935.
[187] A.C. Puleo, D.R. Paul, P.K. Wong, Gas sorption and transport in semicrystalline poly (4-
methyl-1-pentene), Polymer. 30 (1989) 1357–1366.
[188] S. Kanehashi, A. Kusakabe, S. Sato, K. Nagai, Analysis of permeability; solubility and
diffusivity of carbon dioxide; oxygen; and nitrogen in crystalline and liquid crystalline
polymers, Journal of Membrane Science. 365 (2010) 40–51.
[189] J.-H. Cao, B.-K. Zhu, H. Lu, Y.-Y. Xu, Study on polypropylene hollow fiber based
recirculated membrane bioreactor for treatment of municipal wastewater, Desalination.
183 (2005) 431–438. doi:10.1016/j.desal.2005.02.056.
[190] S.R. Armstrong, G.T. Offord, D.R. Paul, B.D. Freeman, A. Hiltner, E. Baer, Co-extruded
polymeric films for gas separation membranes, Journal of Applied Polymer Science. 131
(2014) 1–11. doi:10.1002/app.39765.
[191] S. Sinha Ray, M. Okamoto, Polymer/layered silicate nanocomposites: A review from
preparation to processing, Progress in Polymer Science (Oxford). 28 (2003) 1539–1641.
doi:10.1016/j.progpolymsci.2003.08.002.
[192] C. Covarrubias, R. Quijada, Preparation of aluminophosphate/polyethylene
nanocomposite membranes and their gas permeation properties, Journal of Membrane
Science. 358 (2010) 33–42. doi:10.1016/j.memsci.2010.04.026.
[193] C. Covarrubias, R. Quijada, R. Rojas, Ethylene polymerization using nanosized ZSM-2
zeolite particles as support for metallocene catalyst, Applied Catalysis A: General. 347(2)
(2008) 223–233.
151
[194] M. Nakata, H. Kumazawa, Gas permeability and permselectivity of plasma-treated
polyethylene membranes, Journal of Applied Polymer Science. 101 (2006) 383–387.
doi:10.1002/app.23850.
[195] R. Checchetto, A. Miotello, L. Nicolais, G. Carotenuto, Gas transport through
nanocomposite membrane composed by polyethylene with dispersed graphite
nanoplatelets, Journal of Membrane Science. 463 (2014) 196–204.
doi:10.1016/j.memsci.2014.03.065.
[196] A.S. Michaels, H.J. Bixler, Solubility of gases in polyethylene, Journal of Polymer
Science. 50 (1961) 393–412.
[197] J.P.G. Villaluenga, B. Seoane, Experimental estimation of gas-transport properties of
linear low-density polyethylene membranes by an integral permeation method, Journal of
Applied Polymer Science. 82 (2001) 3013–3021. doi:10.1002/app.2156.
[198] J. Yao, M.R. Barzegari, D. Rodrigue, Density graded linear medium density polyethylene
foams produced under a temperature gradient with Expancel microbeads, Cellular
Polymer, 30 (2011) 157-185.
[199] A. Raymond, D. Rodrigue, Foams and wood composite foams produced by rotomolding,
Cellular Polymers. 32 (2013) 199–212.
[200] S.J. Moore, S.E. Wanke, Solubility of ethylene, 1-butene and 1-hexene in polyethylenes,
Chemical Engineering Science. 56 (2001) 4121–4129.
[201] G. Liu, C.B. Park, J. Lefas, Production of low-density LLDPE foams in rotational molding,
Polymer Engineering and Science. 38 (1998) 1997–2009. doi:10.1002/pen.10369.
[202] P. Guo, Y. Xu, M. Lu, S. Zhang, Expanded Linear Low-Density Polyethylene Beads:
Fabrication, Melt Strength, and Foam Morphology, Industrial and Engineering Chemistry
Research. 55 (2016) 8104–8113. doi:10.1021/acs.iecr.6b01545.
[203] H.E. Naguib, C.B. Park, N. Reichelt, Fundamental foaming mechanisms governing the
volume expansion of extruded polypropylene foams, Journal of Applied Polymer Science.
91 (2004) 2661–2668.
[204] J. Wang, L. Zhang, J.B. Bao, Supercritical CO2 Assisted Preparation of Open-cell Foams
of Linear Low-density Polyethylene and Linear Low-density Polyethylene, Chinese
Journal of Polymer Science. 34 (2016) 889–900. doi:10.1007/s10118-016-1806-4.
[205] P. Shahi, A.H. Behravesh, A. Haghtalab, G. Rizvi, R. Pop-Iliev, F. Goharpei, Effect of
Mixing Intensity on Foaming Behavior of LLDPE/HDPE Blends in Thermal Induced
Batch Process, Polymer-Plastics Technology and Engineering. 55 (2016) 949–964.
doi:10.1080/03602559.2015.1132444.
[206] G. Zhang, Y. Wang, H. Xing, J. Qiu, J. Gong, K. Yao, H. Tan, Z. Jiang, T. Tang, Interplay
between the composition of LLDPE/PS blends and their compatibilization with
152
polyethylene-graft-polystyrene in the foaming behaviour, RSC Advances. 5 (2015) 27181–
27189. doi:10.1039/C4RA16084C.
[207] A. Durmus, A. Kasgöz, C. Macosko, Mechanical Properties of Linear Low-density
Polyethylene (LLDPE)/clay Nanocomposites: Estimation of Aspect Ratio and Interfacial
Strength by Composite Models, Journal of Macromolecular Science, Part B. 47 (2008)
608–619. doi:10.1080/00222340801957780.
[208] S. Abe, M. Yamaguchi, Study on the foaming of crosslinked polyethylene, Journal of
Applied Polymer Science. 79 (2001) 2146–2155. doi:10.1002/1097-
4628(20010321)79:12<2146::AID-APP1022>3.0.CO;2-Q.
[209] O. Talu, S.Y. Zhang, D.T. Hayhurst, Effect of cations on methane adsorption by NaY,
MgY, CaY, SrY, and BaY zeolites, The Journal of Physical Chemistry. 97 (1993) 12894–
12898.
[210] S.W. Hwang, Y. Chung, B.C. Chun, S.J. Lee, Gas Permeability of Polyethylene Films
Containing Zeolite Powder, Polymer (Korea). 28 (2004) 374–381.
[211] Ş.B. Tantekin-Ersolmaz, Ç. Atalay-Oral, M. Tatlıer, A. Erdem-Şenatalar, B. Schoeman, J.
Sterte, Effect of zeolite particle size on the performance of polymer–zeolite mixed matrix
membranes, Journal of Membrane Science. 175 (2000) 285–288.
[212] I. Pinnau, W.J. Koros, A qualitative skin layer formation mechanism for membranes made
by dry/wet phase inversion, Journal of Polymer Science Part B: Polymer Physics. 31
(1993) 419–427.
[213] S.-H. Lin, K.-L. Tung, W.-J. Chen, H.-W. Chang, Absorption of carbon dioxide by mixed
piperazine–alkanolamine absorbent in a plasma-modified polypropylene hollow fiber
contactor, Journal of Membrane Science. 333 (2009) 30–37.
[214] C.D. Han, Y.W. Kim, K.D. Malhotra, A Study of Foam Extrusion Using a Chemical
Blowing Agent, Journal of Applied Polymer Science. 20 (1976) 1583–1595.
[215] J.A. Reglero Ruiz, M. Vincent, J.F. Agassant, T. Sadik, C. Pillon, C. Carrot, Polymer
foaming with chemical blowing agents: Experiment and modeling, Polymer Engineering
& Science. 55 (2015) 2018–2029. doi:10.1002/pen.
[216] D. Klempner, K.C. Frisch, Handbook of polymeric foams and foam technology, Hanser
Munich, 1991.
[217] C. Okolieocha, D. Raps, K. Subramaniam, V. Altstaedt, Microcellular to nanocellular
polymer foams: Progress (2004-2015) and future directions - A review, European Polymer
Journal. 73 (2015) 500–519. doi:10.1016/j.eurpolymj.2015.11.001.
[218] X.M. Zhang, S. Elkoun, A. Ajji, M.A. Huneault, Oriented structure and anisotropy
properties of polymer blown films : HDPE, LLDPE and LDPE, 45 (2004) 217–229.
doi:10.1016/j.polymer.2003.10.057.
153
[219] V. Compañ, L.F. Del Castillo, S.I. Hernández, M.M. López-González, E. Riande, On the
crystallinity effect on the gas sorption in semicrystalline linear low density polyethylene
(LLDPE), Journal of Polymer Science Part B: Polymer Physics. 45 (2007) 1798–1807.
doi:10.1002/polb.
[220] J.A. Reyes-Labarta, M.M. Olaya, A. Marcilla, DSC and TGA study of the transitions
involved in the thermal treatment of binary mixtures of PE and EVA copolymer with a
crosslinking agent, Polymer. 47 (2006) 8194–8202. doi:10.1016/j.polymer.2006.09.054.
[221] R. Guan, B. Wang, D. Lu, Q. Fang, B. Xiang, Microcellular thin PET sheet foam
preparation by compression molding, Journal of Applied Polymer Science. 93 (2004)
1698–1704. doi:10.1002/app.20614.
[222] R. Gosselin, D. Rodrigue, Cell morphology analysis of high density polymer foams,
Polymer Testing. 24 (2005) 1027–1035. doi:10.1016/j.polymertesting.2005.07.005.
[223] Y.S. Nam, T.G. Park, Biodegradable polymeric microcellular foams by modified thermally
induced phase separation method, Biomaterials. 20 (1999) 1783–1790.
doi:10.1016/S0142-9612(99)00073-3.
[224] N. Sarier, E. Onder, Thermal characteristics of polyurethane foams incorporated with
phase change materials, Thermochimica Acta. 454 (2007) 90–98.
doi:10.1016/j.tca.2006.12.024.
[225] S.T. Lee, Foam Extrusion Overview, Foam Extrusion, Technomic Publishing Co. (2000).
[226] H. Teymoorzadeh, D. Rodrigue, Morphological, mechanical, and thermal properties of
injection molded polylactic acid foams/composites based on wood flour, Journal of
Cellular Plastics. 54 (2018) 179–197. doi:10.1177/0021955X16671304.
[227] K.A. Neavor, A.J. Lesser, T.J. McCarthy, Preparation and characterization of microcellular
polystyrene foams processed in supercritical CO2, Macromolecules. 32 (1998) 4614–
4620. doi:10.1021/ma971811z.
[228] C.B.P. R., Pop-Iliev, N. Dong, D., Xu, Visualization of the Foaming Mechanism of
Polyethylene Blown by Chemical Blowing Agents under Ambient Pressure, Journal of the
Polymer Processing Institute. 26(4) (2007) 213–222. doi:10.1002/adv.
[229] F.J. Moscoso-Sanchez, E. Mendizabal, C.F. Jasso-Gastinel, P. Ortega-Gudino, J.R.
Robledo-Ortiz, R. Gonzalez-Nunez, D. Rodrigue, Morphological and mechanical
characterization of foamed polyethylene via biaxial rotational molding, Journal of Cellular
Plastics. 51 (2015) 489–503. doi:10.1177/0021955X14566207.
[230] K.A. Arora, A.J. Lesser, T.J. Mccarthy, Preparation and Characterization of Microcellular
Polystyrene Foams Processed in Supercritical Carbon Dioxide, Macromolecules. 9297
(1998) 4614–4620.
[231] J.W.S. Lee, C.B. Park, Use of nitrogen as a blowing agent for the production of fine-celled
154
high-density polyethylene foams, Macromolecular Materials and Engineering. 291 (2006)
1233–1244. doi:10.1002/mame.200600203.
[232] S.-T. Lee, C.B. Park, Foam extrusion: principles and practice, CRC press, 2014.
[233] J.G. Burt, The Elements of Expansion of Thermoplastics Part II, Journal of Cellular
Plastics. 14 (1978) 341–345. doi:10.1177/0021955X7801400605.
[234] A. Mohebbi, F. Mighri, A. Ajji, D. Rodrigue, Polymer ferroelectret based on
polypropylene foam : Piezoelectric properties improvement using post-processing
thermomechanical treatment, Journal of Applied Polymer Scienc. 134 (2017) 1–10.
doi:10.1002/app.44577.
[235] C.M.O. Müller, J. Borges, F. Yamashita, Effect of cellulose fibers on the crystallinity and
mechanical properties of starch-based films at different relative humidity values,
Carbohydrate Polymers. 77 (2009) 293–299. doi:10.1016/j.carbpol.2008.12.030.
[236] Y.-G. Zhou, B. Su, L.-S. Turng, Fabrication of super-ductile PP/LDPE blended parts with
a chemical blowing agent, Journal of Applied Polymer Science. 133 (2016) 1–16.
doi:10.1002/app.44101.
[237] H. Ruckdäschel, P. Gutmann, V. Altstädt, H. Schmalz, A.H.E. Müller, Foaming of
microstructured and nanostructured polymer blends, in: Complex Macromolecular
Systems I, Springer, 2009: pp. 199–252.
[238] M. Lee, C. Tzoganakis, C.B. Park, Extrusion of PE/PS blends with supercritical carbon
dioxide, Polymer Engineering and Science. 38 (1998) 1112–1120. doi:10.1002/pen.10278.
[239] A. Mohebbi, D. Rodrigue, Energy Absorption Capacity of Ferroelectrets Based on Porous
Polypropylene, Polymer Engineering and Science. 58(3) (2017) 2–11. doi:10.1002/pen.
[240] W. Gong, J. Gao, M. Jiang, L. He, J. Yu, J. Zhu, Influence of cell structure parameters on
the mechanical properties of microcellular polypropylene materials, Journal of Applied
Polymer Science. 122 (2011) 2907–2914.
[241] E. Soles, J.M. Smith, W.R. Parrish, Gas transport through polyethylene membranes,
AIChE Journal. 28 (1982) 474–479. doi:10.1002/aic.690280316.
[242] V. Altstädt, Complex macromolecular system I, Springer Science & Business Media
(2010).
[243] A. Maani, H.E. Naguib, M.-C. Heuzey, P.J. Carreau, Foaming behavior of microcellular
thermoplastic olefin blends, Journal of Cellular Plastics. 49 (2013) 223–244.
doi:10.1177/0021955X13477435.
[244] D. Abraham, K.E. George, D.J. Francis, Rheological characterization of blends of low
density with linear low density polyethylene using a torque rheometer, European Polymer
Journal. 26 (1990) 197–200.
155
[245] A. Ajji, P. Sammut, M.A. Huneault, Elongational rheology of LLDPE/LDPE blends,
Journal of Applied Polymer Science. 88 (2003) 3070–3077.
[246] A. De Saja, C.S. Jose, Moulded polypropylene foams produced using chemical or physical
blowing agents : structure – properties relationship, Journal of Materials Science. 47(15)
(2012) 5680–5692. doi:10.1007/s10853-012-6357-7.
[247] G.J. Field, P. Micic, S.N. Bhattacharya, Melt strength and film bubble instability of
LLDPE / LDPE blends, Polymer International. 48(6) (1999) 461–466.
[248] A.H. Behravesh, C.B. Park, E.K. Lee, Formation and Characterization of Polyethylene
Blends for Autoclave-Based Expanded-Bead Foams, Polymer Engineering & Science.
50(6) (2010) 1161–1167. doi:10.1002/pen.
[249] T. Joo, B. Lee, B. Shuk, Y. Soo, C. Lee, H. Hoon, R.S. Stein, M. Ree, Composition-
dependent phase segregation and cocrystallization behaviors of blends of metallocene-
catalyzed octene-LLDPE (D) and LDPE (H), Polymer. 51 (2010) 5799–5806.
doi:10.1016/j.polymer.2010.09.075.
[250] C. Liu, J. Wang, J. He, Rheological and thermal properties of m-LLDPE blends with m-
HDPE and LDPE, Polymer. 43 (2002) 3811–3818.
[251] C.S. Brazel, S.L. Rosen, Fundamental principles of polymeric materials, John Wiley &
Sons, 2012.
[252] E. Piorkowska, Crystallization in polymer composites and nanocomposites, Handbook of
Polymer Crystallization. Wiley, Hoboken. (2013) 379–397.
[253] M. Yamaguchi, S. Abe, LLDPE/LDPE Blends. I. Rheological, Thermal, and Mechanical
Properties, Journal of Applied Polymer Science. 74(13) (1999) 3153–3159.
[254] P.P. Kundu, J. Biswas, H. Kim, S. Choe, Influence of film preparation procedures on the
crystallinity, morphology and mechanical properties of LLDPE films, European Polymer
Journal. 39 (2003) 1585–1593. doi:10.1016/S0014-3057(03)00056-9.
[255] S. Amirabadi, D. Rodrigue, C. Duchesne, Characterization of PLA-talc films using NIR
chemical imaging and Multivariate Image Analysis techniques, Polymer Testing. 68
(2018) 61–69. doi:10.1016/j.polymertesting.2018.03.047.
[256] A. Ghassemi, S. Moghaddamzadeh, C. Duchesne, D. Rodrigue, Effect of annealing on gas
permeability and mechanical properties of polylactic acid/talc composite films, Journal of
Plastic Film & Sheeting. 33 (2017) 361–383.
[257] J.P.G. Villaluenga, B. Seoane, D. De Fisica, A.I. Facultad, Influence of drawing on gas
transport mechanism in LLDPE films, Polymer. 39 (1998) 3955–3965.
[258] P. Pandey, R.S. Chauhan, Membranes for gas separation, Progress in Polymer Science. 26
(2001) 853–893. doi:10.1016/S0079-6700(01)00009-0.
156
[259] S.R. Reijerkerk, K. Nijmeijer, C.P. Ribeiro Jr, B.D. Freeman, M. Wessling, On the effects
of plasticization in CO2/light gas separation using polymeric solubility selective
membranes, Journal of Membrane Science. 367 (2011) 33–44.
doi:10.1016/j.memsci.2010.10.035.
[260] B. Ge, Y. Xu, H. Zhao, H. Sun, Y. Guo, W. Wang, High Performance Gas Separation
Mixed Matrix Membrane Fabricated by Incorporation of Functionalized Submicrometer-
Sized Metal-Organic Framework, Materials. 11 (2018) 1421. doi:10.3390/ma11081421.
[261] X. Wei, C. Haire, Biaxially oriented microporous membrane, (2014) WO Patent
2007098339.
[262] H. Yousefian, D. Rodrigue, Effect of nanocrystalline cellulose on morphological, thermal,
and mechanical properties of nylon 6 composites, Polymer Composites. 37 (2016) 1473–
1479. doi:10.1002/pc.23316.
[263] Z. Razzaz, A. Mohebbi, D. Rodrigue, Gas transport properties of cellular hollow fiber
membranes based on LLDPE/LDPE blends, Cellular Polymers. Submited (2018).
[264] Z. Razzaz, A.W. Mohammad, E. Mahmoudi, Foaming Prevention in Absorption Columns
through Removal of Contaminants from Amine-Based Solutions Using a Solvent Resistant
Nanofiltration (SRNF) Membrane, Industrial & Engineering Chemistry Research. 54
(2015) 12135–12142. doi:10.1021/acs.iecr.5b02642.
[265] A. Mohebbi, Optimization of polypropylene cellular films for piezoelectric applications,
Ph.D. Thesis, Université Laval, Quebec, Canada (2016).
[266] R.T. Adams, J.S. Lee, T.H. Bae, J.K. Ward, J.R. Johnson, C.W. Jones, S. Nair, W.J. Koros,
CO2-CH4 permeation in high zeolite 4A loading mixed matrix membranes, Journal of
Membrane Science. 367 (2011) 197–203. doi:10.1016/j.memsci.2010.10.059.
[267] M.J.C. Ordoñez, K.J. Balkus Jr, J.P. Ferraris, I.H. Musselman, Molecular sieving realized
with ZIF-8/Matrimid® mixed-matrix membranes, Journal of Membrane Science. 361
(2010) 28–37.
[268] H. Kim, J. Biswas, S. Choe, Effects of stearic acid coating on zeolite in LDPE, LLDPE,
and HDPE composites, Polymer. 47 (2006) 3981–3992.
doi:10.1016/j.polymer.2006.03.068.
[269] L. Chen, H. Sheth, R. Kim, Gas absorption with filled polymer systems, Polymer
Engineering and Science. 41 (2001) 990–997. doi:10.1002/pen.10800.
[270] J. Yang, L. Huang, Y. Zhang, F. Chen, P. Fan, M. Zhong, S. Yeh, A new promising
nucleating agent for polymer foaming: Applications of ordered mesoporous silica particles
in polymethyl methacrylate supercritical carbon dioxide microcellular foaming, Industrial
and Engineering Chemistry Research. 52 (2013) 14169–14178. doi:10.1021/ie4018447.
[271] G. Bounos, K.S. Andrikopoulos, H. Moschopoulou, G.C. Lainioti, D. Roilo, R.
157
Checchetto, T. Ioannides, J.K. Kallitsis, G.A. Voyiatzis, Enhancing water vapor
permeability in mixed matrix polypropylene membranes through carbon nanotubes
dispersion, Journal of Membrane Science. 524 (2017) 576–584.
doi:10.1016/j.memsci.2016.11.076.
[272] T. Graunke, K. Schmitt, J. Wöllenstein, Organic membranes for selectivity enhancement
of metal oxide gas sensors, Journal of Sensors. 2016 (2016) 2435945.
doi:10.1155/2016/2435945.
[273] P. Barber, S. Balasubramanian, Y. Anguchamy, S. Gong, A. Wibowo, H. Gao, H.J. Ploehn,
H.C. Zur Loye, Polymer composite and nanocomposite dielectric materials for pulse power
energy storage, Materials. 2 (2009) 1697-1733. doi:10.3390/ma2041697.
[274] R.C. Dutta, S.K. Bhatia, Transport diffusion of light gases in polyethylene using atomistic
simulations, Langmuir. 33 (2017) 936–946. doi:10.1021/acs.langmuir.6b04037.
[275] H. Yousefian, D. Rodrigue, Nano-crystalline cellulose, chemical blowing agent, and mold
temperature effect on morphological, physical/mechanical properties of polypropylene,
Journal of Applied Polymer Science. 132 (2015) 2–10. doi:10.1002/app.42845.
158
Appendix A: Abstract
Mixed matrix membranes based on silica nanoparticles and microcellular polymers for
CO2/CH4 separation
X.Y. Chen, Z. Razzaz, S. Kaliaguine, D. Rodrigue
Published in: Journal of Cellular Plastics, 54(2), 309-331 (2018).
Abstract
Mixed matrix membranes made from silica nanoparticles and microcellular polymers were prepared from
Matrimid® 5218 combined with tetramethoxysilane, tetraethoxysilane, and tetrapropoxysilane via the sol–
gel method. The nanoparticles were prepared in situ during membrane casting yielding a homogeneous
distribution inside a foamed polyimide structure. Mixed matrix membranes with SiO2 contents up to 16%
wt. were treated at 60℃, 100℃, 150℃, and 200℃. Thermal gravimetric analysis and Fourier transform
infrared spectroscopy analyses were performed providing information on chemical composition and
thermal stability, while the porous structure (average cell diameter and cell density) was studied by
scanning electron micrograph. Also, dynamic mechanical analysis was used to determine the glass
transition temperature (Tg) and elastic modulus. Finally, the gas transport properties were studied in terms
of treatment temperature, feed pressure, SiO2 loading, and testing temperature. CO2 permeability was
found to increase by a factor of 3–4 at 3% SiO2 content using tetraethoxysilane in Matrimid, while ideal
selectivity for CO2/CH4 separation was constant. Finally, the plasticization effect was practically
eliminated by the introduction of SiO2 nanoparticles.
In this work, my contribution was on the morphological analysis of the samples.
159
Appendix B: Equipment Conditions
B1 Extruder and Die
The extruder was equipped with a tubular die as a continuous technique to produce hollow fiber
foamed LLDPE. The extruder screw configuration design is shown as follow. The numbers above
the elements indicate the screw profile orders.
160
A tubular-die was installed on the extruder as follow:
Outside diameter (mm) 5
Inside diameter (mm) 3.5
B2 Calendaring equipment
A calendaring system including two-roll was used to apply uniaxial stretching on the produced
hollow fiber foamed. The specifications related to the calendaring system are as follow:
Roll diameter (cm) 7
Roll width (cm) 25
Roll temperature (C) 25 (ambient)
B3 Air flow controller system
Air was used to form the tubular geometry in the production process, but an accurate control
system was needed. The specifications of the mass flow meter/controller are as follow:
Gas temperature (C) 25 (ambient)
Gas pressure (psi) 0.01 - 0.05
Gas injection point Die of the extruder
