Embed Size (px)
Citation preview
HETEROGENEOUS POLYELECTROLYTE GEL MEMBR4NES: EFFECT OF MORPHOLOGY ON
STIMULI-RESPONSIVE PERMEATION CONTsROL
Josephine Turner
A thesis submitted in conformity with the requirements For the degree of Doctor of Philosophy,
Graduate Department of Chemical Engineering and Applied Chernistry, io the University of Toronto
O Copyright by Josephine Turner, 2001
National Library I*I of Canada Bibliothèque nationale du Canada
Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395, rue Wellingîon Ottawa ON K1A ON4 Onawa ON K I A ON4 Canada Canada
Your ive Votre réfdrence
Our IWe Nom tdidrence
The author has granted a non- L'auteur a accordé une licence non exclusive licence dowing the exclusive permettant a la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of this thesis in microfom, vendre des copies de cette thèse sous papcr or electronic formats. ia forme de microfiche/nIm, de
reproduction sur papier ou sur fomat électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thése. thesis nor substantial extracts fiom it Ni la thése ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent ê e imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.
Heterogeneous Polyelectrolyte Gel Membranes: Effect of Morphology on
Stimuli-Responsive Permeation Control
Doctor of Philosophy, 2001
Josephine Turner Department of Chemical Engineering and Applied Chemistry
University of Toronto
ABSTRACT
Poly(methacry1ic acid)-poly(dimethy1 siloxane) (PMAA-PDMS) composites and
interpenetrating polymer networks (IPNs) were prepared in order to examine the effect of a
heterogeneous hydrogel-elastomer morphology on the variable transport properties of
stimuli-responsive membranes. The composites consisted of PMAA particles ( d o pn
diameter) dispersed within a PDMS network and IPNs were composed of PMAA channels
(a. 1 p diameter) which spanned the thickness of the PDMS network.
The mechanism of permeation control for the composite was based on the
percolation or comectivity of the dispersed PMAA particles. The fast swelling-deswelling
rates of the surface-resident PMAA particles resulted in a dynamic permeation response to
pH change in the order of minutes and an ON/OFF (pH 7/pH 3) permeability ratio of 160
for vit- BI*. This compared favourably to the ON/OFF pemeability ratio of 7 estimated
for homogeneous PMAA membranes.
The mechanism of pemeation control for IPNs was based on size exclusion of the
permeant fiom the PMAA channel. Although the ON/OFF penneability ratio (1400)
improved over both PMAA and composite membranes, the response time for the dynamic
24 h permeation cycle was very slow. The slow swelling kinetics of the membrane-
spanning PMAA channel fiom pH 3 to 7 resulted in a lag t h e of 15 h before permeation
resumed at pH 7. A combination of the composite and IPN morphologies is expected to
improve upon the deficiencies of the two membrane types. For example, a hydrogel-PDMS
IPN can be prepared where the gel channel in the surface region is stimuli-responsive and
that within the bulk remains hydrated and non-responsive.
A novei monomer immersion method was deveIoped to prepare stimuli-responsive,
bicontiuuous, PMAA-PDMS IPN membranes. The method involved immersing the pre-
IPN in rnethacrylic acid (MAA) during IPN formation in order to obtain a unifonn MAA
concentration profile within the pre-IPN.
A unique appmach was used to examine PMAA-PDMS IPN morphology with the
laser scanning confocal microscope and fluorescent probes of varying diameters. The
results revealed cornplex, superimposed structures of PMAA domains of varying sizes and
spatial distributions. These observations had not been reported previously and present a new
understanding of morphology development in IPNs.
iii
1 would like to express much gratitude to my supervisor, Professor Yu-Ling Cheng, for her constant support, patience and unyielding drive to bring forth the best fiom her graduate students. This thesis would not have been possible without her guidance and support.
I'd like to thank the members of the thesis cornmittee, Professors Sefton, Woodhouse, Santerre and Baike for their valuable suggestions and insights. A special acknowledgement to Professor Sefton, my undergraduate and masters supervisor, who fmt introduced me to the fascinating world of biomaterials and h g delivery.
I'd like to thank the fiiendly, generous and intelligent labmates who have graced Rm 366 with their presence. That is where the real learning took place and 1 will look back on those sessions of collaborative discovery with much fondness. A special acknowledgernent to Jennifer Smith who has always been generous with her tirne and knowledge towards myself and whoever else happens to walk through the door of lab 366.
I'd like to thank my parents for their constant and unquestionhg love, support and enthusiasm regardless of which path I decide to pursue.
To my husband, Steven, Much love and gratitude for your great generosity, Your kind and gentle spirit, Your loving and knowing ways.
This thesis is dedicated to my children. May your fiiture work be full of wonder and joyfùl discovery; Meditat ive labour, challenges and triumphs. Much as this work has been for me.
Knowledge is not a copy of reality. To b o w an object, to know an event, is not simply to
look at it and muke a mental copy, or image of it. To know an object is to act on it. To
know is to modzjj, to tramform the object, und to understand the process of this
transformation, and as a consequence to understand the way the object is conshwcted.
Piaget
TABLE OF CONTENTS
TABLE OF CONTENTS mmm.......mm.m..............................m.......m.......................................... VI
LIST OF TABLES ......................................................................................................... XII
CHAPTER 2 :THESIS OBJECTIVE AND HYPOTHESIS ...mm...m.........e..m.........m.m.... ,..6
2.1 P~PARATION OF HETEROGENEOUS PMAA-PDMS COMPOSITE AND IPN
MEMBRANES. ................................................................................................................... .6
2.2 MEcHANIsM OF PERMEATION CONTROL FOR PMAA-PDMS COMPOSITE AND IPN
MEMBRANES.. .................................................................................................................. .7
2.3 VARIABLE TRANSPORT PROPERTES OF PMAA-PDMS COMPOSITE AND IPN
MEMJ~RANES. ................................................................................................................... .9
2.4 THESIS HYPOTHESIS ...................................... ,. 1 2.5 THESIS~UTLINE ..................................................................................................... 11
................................................................. 3.1 STIMULI-RESPONSWE DRUG DELNERY 14
....................................................................................... 3.2 RESP~NS~VE HYDROGELS 17
32.1 Thennodynamic Basis Of Hydration Change For Responsive Hydrogels .... 17
......................................................................................... 3.2.1.2 Osmotic Pressure of Elastic Retraction 19
........................................................................................................ 3.2.1.3 lonic ûsmotic Pressure 2 0
3.2.1.4 Total ûsmotic Pressure ............................................................................................................. 23
3.2.2 Responsive Hydrogels used for Drug Delivery: Current State of Art .......... 24
3.2.2.1 Slow response time ....................................................................................................................... 25
................................................................................................................... 3.2.2.2 Mechanical Strength 2 6
......................... 3.2.2.3 Low ONIOFF ratios ... .................................................................................. 26
.................................................................... 3.3 HETEROGENEOUS POLYMER SYSTEMS 27
...................................................................................................... 3.3.1 Elastomers 28
.................................................................................................... 3.3.2 Composites 30
............................................................... 3.3.3 Interpenetrating Polymer Networks 32
............................................................................................................ 3.3.3.1 Morphology Development 33
.............. 3.4 PERMEABILITY CONCEPTS FOR RESPONS~VE HETEROGENEOUS SYSTEMS 38
3.4.1 Solute Difision Through Hydrogels: Free Volume Theory ........................ 38
3.4.2 Solute diffision in PDMS Elastomers ........................................................... 40
3.4.3 Mass Transfer in Heterogeneous Media: Percolation Theory ...................... 41
3.4.4 Variable Penneability Membranes: Mechanisms of Penneation Control .... 42
CHAPTER 4 : EXPERIMENTAL METHODS ............m.. .........e..e..........m........m....... 45
vii
.............................................................................. 4.2 FREPARATION OF MEMBRANES -47
4.2.1 Preparation of P m gel .............................................................................. 47
4.2.1.1 Materials ................................................-................................................................................. 47
4.2.1.2 Synthesis ................................................................................................................................... 47
....................................................................... 4.2.2 Preparation of PDMS network 48
..................................,.. .......................................................................................... 4.2.2.1 Materials ... 48
................................................................................................................................... 4.2.2.2 Synthesis 48
..................................................... 4.2.3 Preparation of P M - P D M S composite -49
4.2.3.1 Materials ............................................................................................................................... 49
4.2.3.2 Synthenis ...................................................................................................................................... 49
.............................................................. 4.2.4 Preparation of PMAA-PDMS IPN 3 0
4.2.4.1 Materials ....................................................................................................................................... 50
.................................................................................. 4.2.4.2 Synthesis: Monomer-Immersion Method 50
........................................................................................... 4.2.4.3 Synthssis: Air-IPN interface Method 51
...................................................................................... 4.2.4.4 Synthesis: Glass-IPN Interface Method. 51
........................................................................ 4.3 METHODS OF CHARACTERIZATION -52
................................................... 4.3.1 Determination of M, for PDMS Networks .52
........................................................ 4.3.2 Detemination of Membrane Hydration 53
............................................................................................... 4.3.2.1 Preparation of pH B a r Solutions 53
.......................................................................................................................... 4.3.2.2 Swelling Studies 54
....................................................................................... 4.3.2.3 Equations of Hydration for Membranes 54
...................................... 4.3.3 Pre-Equilibrated and Dynamic Penneation Studies 55
............................................................................................. 4.3.4 LSCM Studies S6
viii
.............................................................................................................. 7.3 RESULTS 105
......................................................................................................... 7.4 Drscuss~o~ 107
.................................. 7.4.1 PMAA-PDMS IPN Morphology at a Specific Depth 108
....... 7.4.2 PMAA-PDMS IPN Gel Domain Morphology as a Function of Depth 112
....................................................................................................... 7.5 CONCLUSION 1 3
................................................................................................... 8.1 INTRODUCTION 1 1 8
............................................................................................................ 8.2 METHODS -121
8.3 RESULTS ............................................................................................................... 122
........................................................................................................ 8.4 DISCUSSION 1 2 4
8.4.1 Mechanism of Permeation Change .............................................................. 124
........................... 8.4.2 EfEect of IPN Morphology on Hydration and Permeation 126
8.4.2.1 (a) Equilibrium Hydration and Pemeation Properties ............................................................... 126
8.4.2.2 (b) Dynarnic Hydration and Pemeation Properties ................................................................... 127
....................................................................................................... 8.5 CONCLUSIONS 129
....................... CHAPTER 9 : CONCLUSIONS AND RECOMMENIBATIONS ..... 138
9.1 PREPARATION OF HETEROGENEOUS PMAA-PDMS COMPOSITE AND IPN
M E ~ R A N E S ................................................................................................................. 139
9.2 MECHANISM OF PERMEATION CONTROL FOR PMAA-PDMS COMPOSITE AND IPN
................................................................................................................ W ~ R A N E S 141
9.3 VAIUABLE TRANSPORT PROPERTIES OF PMAA-PDMS COMPOSITE AND IPN
MEMBRANES ................................................................................................................. 142
9.4 FUTURE RESEARCH WORK ................................................................................... 144
APPENDIX A: Surface Analysis of IPN Membranes .......................................... 157
APPENDIX B: Detailed description of the preparation of PMAA-PDMS IPN
membranes using the rnonomer immersion method ............................................... 162
APPENDIX C: Mass Balance of MAA monomer during formation of PMAA-PDMS
IPN with Air-IPN Intedace ....................................................................... -165
LIST OF TABLES
Table 2.1 Stimuli which are utilized for the regulation of hydrogel membrane.
adapted fiom [Brondsted 199 11 ............................................... 16
Table 8.1 Hydration of PMAA gel. Composite and IPN membranes and of gel
components in Composite and IPN membranes at pH 7. 5 and 3 ........ 132
Table 8.2 Permeability of PMPLA gel. Composite and IPN membranes
equilibrated at pH 7. 5 and 3 ........................................ 132
xii
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 5.1
Figure 5.2
Figure 5.3
Schematic illustration of a h g delivery system connected to
a sensor that monitors a relevant body fùnction continuously and
produces a feedback signal to the computerized delivery module
[Heilmann1984].. ............................................................... 15
Chernical equilibrium of ionized poiy(methacry1ic acid). ................ .20
Chernical structure of poly(dimethy1 siloxane). ........................... 29
Miscibility curves corresponding to miscibility (line a), phase
separation(1ine b) and immiscibility (line c) [Kiefferlggg].. ............ 34
Phase diagrarn of a polyrner mixture with a lower critical solution
temperature. .................................................................... . 3 5
Schematics of the unit ce1 for a CO-continuous structure via spinodal
decomposition and the concentration fluctuation along a straight line
passing hough the unit ce11 [Inouel 9953.. ................................ - 3 8
Percolation curves and schematics of composite membranes which
demonstrate hypothesized synergistic effects of hydration and
percolationon effective diffisivity of membranes.. ........................ 73
Swelling ratio vs. pH for PMAA gels (0.25% TEGDMA) and
composite membranes (28% dry gel loading). Ermr bars indicate
standard deviations (n=3). ..................................................... 74
Hydration vs. pH for PMAA gels (0.24% TEGDMA) and
... Xll l
composite membranes (28% dry PMAA gel loadings). Error bars
indicate standard deviations (n=3). .......................................... 75
Figure 5.4 Permeability of caffeine through PMAA gels (0.25% TEGDMA)
and composite membranes (17%, 22%, 28%, 33% dry PMAA gel
loading) as a function of pH. Emor bars indicate standard
deviations ( ~ 3 ) . ............................................................... ,76
Figure 5.5 Permeability of vitamin BI 2 h u g h PMAA gels (0.25%
TEGDMA) and composite membranes (1 7%, 22%, 28%,3 3%
dry PMAA gel loading) as a fùnction of pH. Error bars indicate
standard deviations (n = 3). .................................................. 77
Figure 5.6 Seni-log plot of permeability of cafEeine and vitamin B12 through
PMAA gel membranes vs. the inverse of hydration of gel
membrane. ....................................................................... 78
Figure 5.7 Serni-log plot of permeability of caffeine îhrough composite
membranes vs. the inverse of gel hydration in membrane.. ............. .79
Figure 5.8 Semi-log plot of permeability of vitamin Biz through composite
membranes vs. inverse of hydration of gel in composite: ( )
detectable permeation indicated the existence of percolating
clusters; ( ------- ) non-detectable permeation indicated no
percolating c1usters.. .......................................................... .80
Figure 5.9 Pemeability of cafTeine and vitamin B12 VS. the volume fiaction of
hydrated PMAA gel in composite membranes.. ........................... 81
Figure 5.10 Dynamic permeation profile for caffeine through composite
xiv
membranes with 28% dry PMAA gel loading.. ........................... .82
Figure 5.1 1 Dynamic permeation profile for vitamin B 12 through composite
.............................. membrane with 28% dry PMAA gel loading.. 83
Figure 5.12 Composite membrane concentration profiles in a dynamic
permeation study: (A) membrane in pH 7 solution, geI particles
swell and connect, steady state concentration profile; (B)
membrane switched to pH 3 solution, surface gel particles deswell,
and disconnect, concentration equilibration between interior,
swollen comected particles; (C) membrane switched to pH 7
solution, surface particles swell and connect, sharp concentration
......................... gradient at membranelrelease medium interface.. 84
Figure 6.1 LSCM images of depth-profile of IPN prepared with a glas
substrate (glass-IPN): ( a ) swface; ( b ) 2 p; ( c ) 14 p;
( d ) 18 pm; ( e ) 24 p; ( f ) 30 p .......................................... 94
Figure 6.2 LSCM images of depth profile of IPN prepared with a fiee
surface (air-IPN): ( a ) surface, ( b ) 5 pm; ( c ) 10 p; ( d ) 40 pm;
( e ) 50 pm; (f) 60 pn ......................................................... 95
Figure 6.3 LSCM images of depth-profile of IPN prepared using monomer
immersion method: ( a ) surface; ( b ) 10 pm; ( c ) 20 pn;
( d ) 3 0 pm; ( e ) 40 pm; ( f )50pm .......................................... 96
Figure 6.4 Percent FIuorescent Area as a function of Depth fiom IPN Surface.. ... 97
Figure 7.1 Phase diagram of a polymer blend depicting the lower critical
solution temperature and quench depth of the system.. .................... 99
Figure 7.2
Figure 7.3
Figure 7.4
Figure 7.5
Figure 7.6
Figure 7.7
Figure 8.1
Figure 8.2
Figure 8.3
Figure 8.4
Figure 8.5
Phase diagram of a polymer blend system undergohg PIPS. The
LCST @oint A) decreases with t h e as the molecular weight
of the polymers increase due to polymerization and crosslinking
reactions. TEmson of thesystem remains constant. ......................... 100 Phase Diagram for a polymer blend system. The kinetic
mechanisms of phase separation are illustrated for each region
[Kieferl999].. ................................................................... 1 O 1
Change in morphology with tirne during thermally-induced spinodal
decornposition pnoue 1 9951.. ................................................. 102
LSCM images of depth profile of PMAA-PDMS IPN immersed in
fluorescein solution.. .......................................................... ,115
LSCM images of depth-profile of IPN immersed in FITC-dextran
(4,400 Da) solution.. ............................................................ 1 16
LSCM images of depth-profile of PMAA-PDMS IPN immersed in
FITC-dextran (70,000 Da) solution.. ......................................... 1 17
LSCM images of PMAA-PDMS IPN equilibrated at
pH7(a)-(b),pH5(c)-(d),andpH3 ( e ) - ( f ) i n
FITC-dextran (4,400 Da) solution at 10 and 20 p.. ..................... .13 1
2 h hydration cycle for PMAA-PDMS IPN membrane. ................. .13 3
24 h hydration cycle for PMAA-PDMS IPN membrane. ................. 134 4 h permeation cycle of vitamin B12 through PMAA-PDMS IPN
membrane ...................................................................... 135
24 h permeation cycle of vitamin BI2 thr~ugh PM.kM-PDMS IPN
xvi
membrane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 3 6
Figure 8.6 Depiction of hydration response of hydrogel domain channel of
IPN at different pH conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 1 3 7
xvii
d diameter
h.d. hydrodynamic diameter
H hydration
M mass
molecular weigbt between crosslinks
MW molecular weight
M, the d q mass of the PDMS network in grarns
Mf the swollen mass of the PDMS network in grams
P permeability
Il osmotic pressure
elastic osmotic pressure
osmotic pressure of electrostatic repulsion
Hi, ionic osmotic pressure
Il,, osmotic pressure of mixing
Ppoiyma density of the polymer network before swelling
VI molar volume of the swelling soivent at room temperature
V2rn volume fiaction of the PDMS network in the swollen state
X polymer-solvent interaction parameter
m s density of the PDMS network in g/L
p density of cyclohexane in g/L
xviii
G Gibbs Free Energy
volume hction
p chemical potential
A wavelength of concentration fluctuation
Ts spinodal temperature
Vf fkee volume
Pm permeability through membrane
Do solute diffusion coefficient
Deff effective diffusivity
xix
GLOSSARY OF TERMS & ACRONYMS
air-IPN -term used b denote an IPN which has been prepared with IPN surface interfixing
a fke surface such as air or nitrogen during polymerization and crosslinking reactions
bicontinuous morphology -description of a two phase morphology in which both phases
span the bulk of the material beginning at one surface to the opposite surface
dynamic permeation response -a material responds to changes in stimuli via changes in
the permeability of solutes through the material
glass-IPN -tenn used to denote an IPN which has been prepared with IPN surface
interfacing a glass substrate during polymerization and crosslinking reactions
Heterogeneous Poiyelectro~e Geis -a material prepared h m a polyelectrolyte gel and a
second polymer which is not compatible with the polyelectrolyte gel so that the material
demonstrates a micro- or macro phase separated structure
HPG -heterogeneous polyelectrolyte gel
IPlY 4nterpenetrating polymer network
interpenetrating poïymer network -a multicomponent material comprised of two or more
crosslinked networks that are at least partially interlaced on a molecular scale, but are not
covalently bonded to each other
LSCM -Laser Scanning Confocal Microscope
LCST -Lower Critical Solution Temperaiure
MAA -methacrylic acid
ON/OFF drug permeability ratio -the ratio of pemeabiiity of solute in the ON state
(maximum permeability) to permeability in the OFF state (minimum permeability). This is
a measure of the range of permeability values attainable by a system for a particular solute.
PFPS -Polyrnerization Induced Phase Separation
pre-IPN -polymer network which has been swdlen with monomer, crosslinking agent and
intitiator prior to the polymerization and crosslinking reactions and the formation of the IPN
PMAA -poly(methacrylic acid)
PDMS -poly(dimethyl siloxane)
ses-island morphology -description of a two phase morphology in which the first phase is
dispersed (forms islands) within the second phase (the sea)
TEGDMA -triethylene glycol dimethacrylate
UCST -upper critical solution temperature
CHAPTER 1 : INTRODUCTION
A pulsatile h g delivery regimen for the administration of peptide or protein
drugs improves therapeutic effectiveness, minimizes side effects and prevents dmg
tolerance by optimizing dmg release profiles to mimic cyclical, physiologic release
patterns [Li 1 987, Breimer 19931. Variable-perrneability , stimuli-responsive membranes
are currently being exarnined as potential control elements for pulsatile drug delivery
systems. These membranes may act as ON-OFF switches or permeability valves to
produce pulsatile drug release profiles, where the period and rate of drug delivery can be
controlled by extemal triggers (i.e. pH or electric field) or physiologic parameters (blood
sugar levels). Ideally, response of the membrane system to stimuli should be fast,
accurate and reproducible. In order for the system to be applicable to different
therapeutic regimens, it is desirable that the membrane be able to produce a wide range of
permeation values or a hi@ ON/OFF cimg permeability ratio (the ratio of maximum
pemeability to minimum permeability for a particular membrane-permeant system). The
OFF state should result in zero flux of dmg.
Past research efforts have focussed on the use of homogeneous polyelectro~yte
hydrogels as potential responsive membranes for pulsatile h g delivery applications
rGehrkel990, Gehrke20001. Polyelectrolyte gels are permeable to large peptides and
proteins, and their hydration can be altered via stimuli such as pH, ionic strength, electric
field, electric current, temperature and chemical agents voshidal9931. In most cases,
the mechanism of permeation control through hydrogels is based on their hydration
[Yasuda1969], where an hcrease in hydraaion leads to increases in the diffusivity of
solute through the hydrogel. Accordingly, the development of variable permeability
hydrogel membranes has focused on producing changes in hydration to effect changes in
membrane permeability [Weiss1 986, Sawahatal990, Hofian1 986, Gehrke1989,
Bae1989, Braze119961. Unfortunately, polyelectrolyte gels are unable to provide a wide
range of permeation values due to their relatively high hydration levels in the OFF state.
For example, Gehrke et al. [Gehrke 19891 found that acrylarnide-co-MAA copolymers
had an ON/OFF (pH 9.2/pH 4) vitamin Bt2 penneability ratio of 3 when hydration of the
gel changed fiom 0.98 at pH 9.2 to 0.85 at pH 4. More recently, polyw-
isoproplyacrylamide) (PNiPAAmI-PMAA IPNs were investigated as variable
permeability membranes and produced an ON/OFF (pH 7.4lpH4) oxprenolol
permeability ratio of 2.7 [ZhangZOOOa].
Another disadvantage of the homogeneous hydrogel membrane is the slow
hydration response to change in stimuli [Gehrkel990]. A 1 mm thick gel slab with a
difision coefficient of 10" cm2/s requires more than 6 h to reach 90% of equilibrium
hydration in response to stimuli. Since difision times scale with the square of the
dimension, decreasing the characteristic dimension will significantly decrease the
response tirne. Thus for sub-micron sized gels, millisecond response times are possible.
However, gels of such dimensions are not practical as variable permeability membranes.
Another approach has been the preparation of porous hydrogel microstructures that swell
and shrink in response to stimuli thousands of times faster than homogeneous, nonporous
gels. The interco~ected porous structure allows water to be expelled h m pores by a
convective process rather than a diffisive one [Chen2000]. In this case the gel's
mechanical strength is compromised by the porous microstructure making these materials
unsuitable for membrane applications.
Heterogeneous hydrogels, defmed by Bae as a mixture or blend of polymers which
demonstrate micro- or macrophase separated structures [Bae1993], have also been examined
as variable permeability membranes. It was expected that heterogenous hydrogels would
provide new mechanisms of permeation control, irnprove upon the mechanical strength and
provide additional functionality when compared to the homogeneous hydrogel.
Heterogeneous hydrogels prepared by the copolymerization of hydrophobic and
hydrophilic monomers have been used primarily as dnig delivery matrices [Siegel1 988,
Dong1990, Inoue1997J The main advantages provided by these systems have been
impmved mechanical strength and the ability of the matrix to act as a reservoir for
hydrophobic drugs. Hydrophobic polyelectrolyte gels were prepared by Siegel and
Firestone [Siegel19881 h m copolymerization of hydrophobic methyl methacrylate (MMA)
and ionhbIe N,N-dimethylaminoethyl methacrylate @MA). The 70MO MMA/I)MA
copolymer gel demonstrated a sharp transition in hydration fkom 0.1 to 0.7 at pH 6.6. This
change in hydration was a significant improvement over hydrophilic polyelectrolyte gels
that are known to undergo relatively smaller changes in hydration as a fùnction of pH.
Unfortunately the slow response of hydration to changes in pH made these materials
impractical for use as variable permeability membranes. The materials have demonstrated
potential as matrices for oral h g delivery applications [Firestonel988].
Heterogeneous hydrogel-elastomer composites have been prepared for use as
permeability membranes [Schwendemanl992, Cifkoval990, Hron19971. However, these
materials have not been examhed as stimuli-responsive membranes. Lopour and
Janatova [Lopour1995] prepared silicone rubber-hydrogel composite materials (silicone
elastomers filled with very fine particles of hydrogels) that were permeable to low
molecular weight, water-soluble compounds. They found that, in spite of the
hydrophobic rubber matrix, the materials were highly permeable to low molecular weight
solutes and the composites behaved as homogeneous water-swollen hydrogels where the
mechanism of penneation control was a fûnction of membrane hydration.
Heterogeneous hydrogels have also been prepared as înterpenetrating polyrner
networks. Conventional methods of preparing hydrogel-elastomer IPNs were unable to
produce bicontinuous membranes permeable to water-soluble compou11ds when the
elastomer formed the continuous phase wurayamal 993, He 1995, Ingenito1998J For
heterogeneous IPN hydrogels which have been used for delivery of water-soluble
compounds, the hydrogel has formed the continuous phase and the hydrophobic domains
were dispersed within the continuous phase (sea-island morphology) [Bae1991,
Lim1997J. In these situations the hydrophobic domains acted as barriers to diffision but
pemeability was still largely dependent upon hydration of the hydrogel component. A
thenno-responsive IPN composed of poly(acry1amide-CO-butyl methacrylate) and
poly(acry1ic acid) was used as a dnig delivery matrix for the pulsatile release of
ketoprofen when the system temperature was cycled between 10°C and 30°C [Lim1997].
Dmg release at 30°C was 20 times larger than release at 10°C. The mechanism of
permeation control for this systern was attributed to the hydrogel-fiee volume changes
that occurred during IPN hydration.
Although heterogeneous hydrogel systems are expected to show great promise as
variable pemeability membranes Pae19931, systerns which have been prepared to date
have been used primarily as h g delivery matrices where the mechanism of permeation
control continues to rely on hydration of îhe hydrogel component. Not surprisingly, the
range of penneation values or the ON/OFF drug permeability ratio has not improved
greatly over that of homogeneous hydrogels. As well, minimal effort has been made to
reduce the dimensions of the hydrogel component in order to improve upon the slow
response associated with homogeneous hydrogel systems. Heterogeneous hydrogel-
elastomer materials have shown significant improvements over homogeneous hydrogels
with respect to their mechanical strength, but there is much which remains to be
understood and exploited with regards to the combination of such highly incompatible
and intrinsically diverse polymeric components.
This thesis has focussed on improving the transport properties of homogeneous
polyelectrolyte hydrogels by using heterogeneous polyelectrolyte gels (HPG) coinposed
of permeable, pH-responsive poly(methacry1ic acid) (PMAA) gel domains dispersed
within a water-impermeable, hy drophobic poly(dimethy1 siloxane) (PDMS) network.
Emphasis was placed on exploiting the heterogeneous morphology of these materials to
produce new mechanisrns of permeation control in order to improve upon the
performance of membranes whose permeability was controlled by hydration only.
Efforts were also made to minimize the domain size of the PMAA gel component in
order to improve upon the slow response associated with hydrogel membranes. A novel
method was also developed to produce a permeable, bicontinuous hydrogel-elastomer
IPN, a unique material with promising applications as both a biomaterial and a
membrane. This type of material has not been investigated to date due to the inability to
form such membranes using conventional methods of IPN preparation.
CHAPTER 2 :THESIS OBJEC TIVE AND HYPOTHESIS
The main objective of this thesis has been to examine the effect of a hydrogel-
elastomer morphology on the variable transport properties of stimuli-responsive
membranes.
This objective has been subdivided fiirther into three prirnary goals for this thesis.
1. Preparation of two bicontinuous, stimuli-responsive membranes fiom the
same polymer components but having very different heterogeneous
morphologies.
2. Investigation of the rnechanism of permeation control for each membrane
system.
3. Evaluation and cornparison of the variable transport properties for each
membrane system.
2.1 Preparation of Heterogeneous PMAA-PDMS Composite and IPN
Membranes
Composites and intepenetrating polymer networks (IPNs) were prepared from
PMAA and PDMS to form heterogeneous hydrogel membrane systerns. In both cases
PMAA was the minor component (approximately 30% wt fiaction) and, therefore, it was
important that both membrane types be bicontinuous (both polymer components form a
continuous pathway from one membrane surface to the other) to ensure pemeability to
water-soluble species. As noted in the introduction, for the preparation of hydrogel-
elastomer IPN membranes conventional rnethods of preparation are unable to produce
biconiinuous, permeable membranes. Thus it was necessary to develop a novel method for
the preparation of bicontinuous, stimuli-responsive PMAA-PDMS IPN membranes. This
method is well-docwented in chapter 6 which is based on a published paper [Turner2000].
A U.S. patent has been allowed regarding the process and materials fonned fiom the
process.
Composite and IPN membrane types were specifically chosen in this work
because their morphology was very different, thus allowing significant conclusions to be
drawn regarding the effect of a heterogeneous membrane morphology on variable
transport properties. The main differences in morphology were expected to be based on
the size and connectivity of the PMAA domains. Composite membranes were prepared
using micron-sized P m gel particles and IPNs were expected to form PMAA gel
domains in the nanometer range, resulting in a 100-fold difference in PMAA domain size
between the two membrane types. Furthemore, it was expected that the gel particles in
the composite membrane would be homogeneously dispersed throughout the PDMS
network but not connected to each other. Gel particle connectivity was expected to occur
as a result of the increase in hydration, and therefore volume of the gel particle. In
contrast, it was expected that IPNs wouId be formed having PMAA gel channels which
spanned the thickness of the membrane, forming permanent connected pathways through
the IPN membrane.
2.2 Mechanism of Permeation Control for PMAA-PDMS Composite
and IPN Membranes
Both composite and IPN membranes consist of pH-responsive PMAA gel
domains dispersed within a rubbery PDMS network. The composite and IPN membranes
differ only in their morphology or the arrangement of the PMAA gel domains within the
elastomeric PDMS network. Because hydrophilic solutes can permeate through the gel
domains only, this work is based on the premise that for HPG materials transport
properties can be controlled not only by hydration of the gel domains, but also by their
connectivity and/or size.
It is expected that the presence of hydrophobie, impermeable areas will lead to
percolation (composite) and size exclusion (IPN) effects which will dominate the
rnechanisms of permeation control and improve upon the shortcomings of membranes in
which permeation is controlled by hydration only. These new mechanisms of permeation
controI are expected to increase the ON/OFF dmg permeability ratios and produce zero
flux in the OFF state. The small size of the hydrogel domain is expected to improve the
membrane response time to stimulus.
In order to determine the primaxy mechanism of permeation control in these
membranes it was important to understand the morphology of the membranes. This was a
straightforward task for the composite membrane since it was prepared fkom a blend of gel
particles and PDMS resin. However for the IPN membrane, the morphology developed
dwing IPN formation was dependent upon a variefy of parameters not yet clearly
understood by the research community.
The literature has alluded to the formation of domains of multiple length scales in
P N morphology [Tm-Congl999a, Seul1995l. However, there has been no direct evidence
of more than two length scales in IPNs [Chou1994, Widrnaierl9951 and ample evidence has
been provided over the past decade that PNs are composed of a constant domain size
[Chen1998, Bufordl 989, Yeo1983, Donatellil 9761. In order to understand the morphology
and domain sizes fonned within the PMAA-PDMS IPN, a unique approach was used to
visuaiize the IPN morphology. Fluorescent probes having diameters larger, comparable and
smalier than the PMAA domains were used to differentiate between the difFerent sized
domains of the IPN. The method and results of this work are documented in chapter 7
which is based on a paper submitted to the journal Macromolecules for publication.
2.3 Variable Transport Properties of PMAA-PDMS Composite and
IPN Membranes
HPG membranes should improve upon the pemeation control properties of the
homogeneous hydrogel system by pmviding a wider range of permeation values, exhibiting
faster response times to similar changes in pH and zero flux in the OFF state.
The range of permeation values achievable by a membrane system was
characterized by the ON/OFF dmg pemeability ratio. This was the ratio of permeability at
pH 7, when the PMAA hydrogel was at maximum hydration (pH 7), divided by the
pemeability at pH 3, when the PMAA hydrogel was at minimum hydration (pH 3).
Therefore, the larger the ON/OFF drug delivery ratio, the larger the range of pemeation
values that a membrane system exhibited.
The flux in the OFF state was also used in this thesis as a measure of pemeation
control which a membrane system was able to exhibit. Ideally, a membrane system should
be able to produce zero flux in the OFF state.
The response t h e of permeation and hydration to stimuli was another parameter
used to compare the performance of variable pemeability membranes. Since the hydrogel
domains of HPG me~nbranes are quite small (pm - nm range) it was hypothesized that
hydration and volume changes (and change in permeability) would occur much faster for
HPG membranes than they would for homogeneous gel membranes.
For h g delivery applications, the ideal specifications for these parameters would
depend highly on the disease state being treated and the drug being adrninistered. Because
there was no particular clmg deiivery application for which our membrane systems have
b e n designed, the membrane systems were evduated based on the widest range of
permeability values achievable @&$est ONIOFF drug permeability ratios), the smallest flux
achievable in the OFF state and the fastest response time of permeation to stimulus.
It was expected that the mass flux through the heterogeneous membrane would be
dependent upon the number of permeable pathways formed within the membrane. For
PMAA-PDMS membranes, these permeable pathways consisted of connected PMAA gel
particles or chameh of a diameter larger than that of the solute. By controllhg the
connectivity and domain size of these channels, the number of permeable pathways and
corresponding mass flux could be controlled. More importantly, these new mechanisms of
permeation control should ampli@ the small changes in hydration brought about by pH
change to produce much larger changes in mass flux than was possible in hydrogel systems.
With respect to response times, the ability to completely stop mass flux through the
membrane by disconnecting the gel particles which make up the permeable path or by
making the path too small should lead to a wider range of permeation values, as well as zero
flux in the OFF state.
2.4 Thesis Hypothesis
The thesis objective and goals outlined above have k e n guided by the followhg
hypothesis:
Permeation control in stimuli-responsive bydrogel membranes, whicb
occurs via hydration only, can be enhanced using heterogeneous
systems where hydration changes may be coupled with changes in gel
domain connectivity (percolation) andlor gel domain size (size
exclusion).
Validation of this hypothesis and realization of the thesis objective and goals
outlined above will represent a significant contribution to the area of variable permeability
membranes and the field of controlled drug delivery. Based on the current literature, no
concerted effort has k e n made to improve upon the performance of variable perrneability
membranes by devising new mechanisrns of permeation control.
2.5 Thesis Outhe
The body of the thesis has been divided into chapters that are based on published and
submitted papers. Thus chapters 5,6 ,7 and 8 each focuses on a significant contribution
that the thesis has made to the current body of literature. At the same t h e these chapters
demonstrate that the thesis has successfully realized its objective and goals, and that the
main hypothesis has been validated.
In chapter 5 [Turner1 9981 responsive h g delivery using PMAA-PDMS composite
membranes and the mechanism of permeation control for these membranes are presented,
discussed and compared to the stimuli-responsive transport properties of homogeneous
PMAA membranes.
The preparation, characterizattion and use of heterogeneous IPNs are presented in
Chapters 6,7 and 8. Chapter 6 [Turner2000] deals with the preparation of PMAA-PDMS
IPNs wing the novel monomer-immersion method which allows for the preparation of
permeable, bicontuiuous, stimuli-responsive hydrogel-elastomer membranes, characteristics
conventional methods of IPN preparation were unable to achieve. Chapter 7 is concerned
with characterization of the morphology of such PMAA-PDMS IPNs. Using fluorescent
markers and the laser scanning codwal microscope, direct evidence of the formation of
domains composed of dierent characteristic length scales and the morphology formed
during progressive stages of polymerization-induced phase separation was provided via high
resohtion, high magnification images. Chapter 8 examines the permeation and hydration
properties of PMAA-PDMS IPNs and provides results and discussion supporthg the
hypothesis that the dominant mechanism of permeation control for PMAA-PDMS IPN
membranes was due to a size exclusion effect. The benefits of an IPN morphology to the
variable transport properties of IPN membranes are discussed with reference to hydrogel
and composite membranes.
The literature review is presented in Chapter 3, although relevant sections that
pertain to a particular discussion have also been included in chapters 5 - 8. The review first
focuses on stimuli-responsive hydrogels, the various mechanisms for hydration change in
responsive hydrogels, and the current state of the art regarding their use in dnig delivery
systems. The review then centers on heterogeneous systems, the important properties of
elastomers used in such systems, the preparation, morphology and permeability concepts of
composites and IPN membranes and their current status in variable pemeability membrane
appIications.
The preparation of composites and IPNs as stimuli-responsive membranes presented
a unique challenge due to the large incompatibility of the elastomer and hydrogel
components and the requisites of bicontinuity and pH-responsiveness. Chapter 4 contains a
description of the experimental materials and novel methods used to prepare PMAA-PDMS
composites and IPNs. The methods used to characterize the morphology, hydration and
permeation properties of these materials are also described.
This work successfùlly combined the responsive, permeable nature of
polyelectrolyte hydrogels with the mechanical strength and unique transport properties of
hetemgeneous materials to fom stimuli-responsive, variable permeability membranes fiom
composite and IPN rnaterials. A summary of the most significant conclusions in relation to
the thesis objective and goals, and recommendations for fiitwe work are presented in
Chapter 9.
C W E R 3 : LITERATURE REVIEW
The focus of this thesis has been the formulation of novel heterogeneous,
membrane materials and their application as control elements for pulsatile dnig delivery
systems. These materials were prepared by combining hydrophobic
poly(dimethylsi1oxane) (PDMS) rubber and stimuli-responsive poly(methacry1ic acid)
(PMAA) to fom composites and interpenetrating polymer networks (IPNs). It was
hypothesized that the heterogeneous morphology of these materials would produce new
mechanisms of permeation control and irnprove upon the performance of stimuli-
responsive hydrogels currently examined as potential membranes for drug delivery
applications.
The literature review fint examines stimuli-responsive hydrogels, the various
mechanisms for hydration change in responsive hydrogels, and the current state of the art
regarding their use in drug delivery systems. The review then concentrates on
heterogeneous polymeric systems, the Unportant properties of the elastomer used in such
systems, the preparation, morphology and permeability concepts of composite and IPN
membranes, and their current status in variable permeability applications. Emphasis is
placed on a fiuidamental understanding of morphology development in IPNs since novel
methods were developed in this thesis to prepare bicontinuous hydrogel-elastomer PNs and
to characterize their complex morphology.
3.1 Stimuli-Responsive Drug Delivery
The goal of stimuli-responsive h g delivery devices is to adrninister drugs at rates
that Vary according to the therapeutic needs of the patient by responding in a pre-determined
fàshion to signals or triggers provided by the surroundhgs. Potential advantages of such
devices include (i) maximizing therapeutic effectiveness, (ii) minimizing side effects and
(iii) avoiding h g tolerance [Li1 987, Breirnerl9931. These devices are especially suited for
the delivery of protein dmgs such as hormones that require moddated delivery in much the
same way as the physiological secretion of endogenous hormones Fernmer 199 11.
An ideal stimuli-responsive h g delivery device would be able to (1) monitor
related pharmacokinetic parameters of the patient (2) produce a continuous feedback signal
to the device and (3) administer a pre-determined dose as dictated by the signal (Figure 3.1).
WeIler 1 9931.
Therapeutic System Biosystem
Negative Feedback Signal
0 ~ m m m , m m m m m m ~ m m m a ~ m m m m m m m m m m m m m m m w m a
............... Therapeutic effect
........................ Desired dmg concentration at target side
Figure 3.1 Schematic illustration of a drug delivery system connected to a sensor that monitors a relevant body fùnction continuously and produces a feedback signal to the computerized delivery module peilmannl984].
Responsive polymers are o h n investigated as the controI element of stimuli-
responsive dnig delivery devices. Drug diffusion and mass transfer in responsive polymers
are dependent on the polymers physical properties, which may change in response to and
cm be controlled by certain stimuli. Table 3.1 lists the triggering stimuli that cari be used to
control the mass transfer properties of responsive polymers. Although a variety of stimuli
have been examined, the majority of responsive polymers used to modulate protein drug
delivery have been hydrogels, and change in hydration has been the prirnary mechanism of
rnass transfer control.
1 Ionic strength
Chemical species
Enzyme- substrate
Table 3.1. Stimul membranes, adapte4
Hyd rogel
Acidic or basic hydrogel
Ionic hydrogel
Hydrogel containing electron-accepting groups Hydrogel containing immobilized enzymes Magnetic particles dispersed in polyrners
Themoresponsive hy drogels
Polyelectrolyte hy drogels
Mechanism
Change in hydration caused by change in ionization. Change in hydration caused by change in concentration of ions inside gel.
Change in hydration cuased by electron- donating compounds and the formation of charge/transfer cornplex.
Change in hydration caused by the enzymatic conversion of substrate and subsequent change in concentration inside the gel.
Change in percolating volume fraction caused by change in pore conformation.
Change in hydration caused by the change in polymer-polymer and water- polymer interactions
Change in hydration caused by membrane charging. Transport of solute caused by electrophoresis of charged dmg.
that are utilized for the regulation of mass trarnsfer in hydrogel fiom @3rondstedl99 11.
3.2 Stimuli-Responsive Hydrogels
"Intelligent" polymers, also describai as smart, stimuli-responsive or
envkonmentally sensitive, exhibit relatively large physical or chemical changes in response
to stimuli. Although glass and melting transitions of solid polymers can also fit within this
definition, most of the interest in intelligent polymers for h g delivery focuses on hyhgels
[Hofnnan1995].
Hydrogels are crosslinked polymers that can absorb more than 20 % of their weight
in water while maintainhg a distinct three-dimensional structure. The key property that
makes hydrogels valuable for drug delivery applications is gel hydration which, in turn,
affects 0 t h important h g delivery properties, such as permeability to drugs, mechanical
strength and bioçompatability [Gehrke2000].
3.2.1 Thermodynamic Basis Of Hydrntion Change For Responsive Hydrogels
Hydration (H) is defined as the mass hction of swollen hydrogel which is water:
The chemical potential of water in a hydrogel solution (gel water) is less than the
chemical potential of pure water, due to the lowering of the vapour pressure of gel water by
the presence of the polymer. Immersion of a dry hydrogel in water leads to the diffusion of
water into the hydrogel, until the chemical potentials of gel water and smounding water are
equal. Equilibrium can also be established by applying pressure to the hydrogel to raise the
chemical potential of gel water to equal that of the surrounding water. The additional
pressure on the hydrogel that is required to establish this equiiibrium condition is called the
osmotic pressure. Ali factors that affect the chemical potential of gel water cm be expressed
in terrns of an osmotic sweliing pressure 0, so that at equilibrium, when the chemical
potentiais of gel water and surromding water are equaI, the total osmotic pressure of water
(Ilfod) in the gel is also dehed as zero. When ntoEal is negative the gel will swell, when it is
positive the gel will deswell.
3.2.1.1 Osmotic Pressure of MIXing
The osmotic sweiling pressure due to mixing of the polymer molecules with the
solvent (TI-), or polymer dissolution, depends primarily on the hydrophilicity of the gel's
polymer molecules. The higher the concentration of solvated (or dissolveci) polymer
molecules in the solvent, the lower the chemical potential of gel water and, therefore, the
gceater the amount of water required to difise into the gel to raise the chemical potentiai of
gel water and establish equilibrium hydration conditions.
Stimuli, such as temperature, may be used to change the solubility of a polymer in
water, leading to changes in the &, term and resulting in new equilibrium hydration
values. A classic exarnple in the field of dmg delivery is poly(N-isopropylacrylamide)
(PNiPAAm). III aqueous solution, PNiPAAm has a lower critical solution temperature
(LCST) of about 3 1 OC. At this temperature a macromolecular transition fkom a hydrophilic
to a hydrophobie structure takes place so that the polymer solubility in water or the osmotic
pressure of k i n g term decreases and the hydrogel deswells [Cole1987]. PoIy (NiPAArn-
CO-BMA) hydrogel matrix demonstrated reversible swelling in response to a stepwise
temperature change between 20 and 30°C in phosphate buffered saline (pH 7.4) resulting in
the pulsatile release of indomethacin Fael99 11.
Photosensitive compounds such as azobenzenes undergo a conformational change
upon photoirradiation. Photoresponsive polymers have been prepared by incorporation of
these compounds to a polymeric backbone where change in confonnation of the
photosensitive compound led to changes in the & term and related changes in hydration.
These polymers were used for the photochemicai control of penneation of various solutes,
such as metal salts, proteins and amino acids Dshiharal9861.
3.2.1.2 Osmotic Pressure of Elastic Refraction
An elastic retractive force exists within hydrogels that may be described as a positive
osmotic pressure acting on the solution within the gel. al, increases the chernical
potential of the gel water enabling equilibrium conditions to be established between two
otherwise very different solutions.
The crosslink density of the hydrogel affects nk, where an increase in crosslink
density results in an increase in al, and the IIm. A variety of stimuli may affect the
number of physical crosslinks (complexations) contained in hydrogels, resulting in changes
to the elastic osrnotic pressure term. For example, for the IPN structure of poly(acry1amide-
CO-butylmethacrylate) and poly(acrylamide), the temperature-dependent hydrogen bonding
interaction between the two polyrner networks becomes weaker with increasing
temperature, leading to a decrease in the number of functional crosslinks and an increase in
dnig release due to gel swelling (Katanol9911. Complexation may also be sensitive to pH
conditions pe1119941.
Antigen responsive hydrogels have been prepared by grafting the antigen and
corresponding antiboày to the polymer network. Binding between the two molecules
introducecl crosslinks into the network. Cornpetitive binding of the fkee antigen triggered a
change in gel volume owing to the break-up of these noncovalent crosslinks, producing
hydrogels whose swelling response was dictated by a specific protein [Miyata1999]
3.2.1.3 Ionic Osmotic Pressure
Polyelectrolyte gels contain pendant ionizable groups along their polymeric
backbone. When a polyelectrolyte gel is placed in water, ionic groups will dissociate to a
certain degree dependent upon the hydrogen ion concentration and ionic strength of the
aqueous solution. For example, poly(methacry1ic acid) (PMAA), the responsive
polyelectrolyte gel used in this work (Figure 3.2)
Figure 3.2 Chernical equilibrium of ionized poly(methacry1ic acid).
contains carboxylic acid goups which may dissociate in water to release positively charged
counter-ions (hydrogen atoms) into the gel water. In order to maintain electroneutrality
within the gel, these comterions do not difise out f?om the gel and, therefore, d u c e the
chernical potential of gel water even M e r . This phenornenon can be quantitatively
described by the ionic osmotic pressure (Ilion) term. The dissociated ions dong the gel
backbone may M e r cause gel expansion due to the electrostatic repulsion of like charges
bound to the gel, decreasing the osmotic swelling pressure by an amount b,. (osmotic
pressure of electrostatic repulsion).
The greatest contribution, however, to polyelectroIyte gel swelling cornes fkom the
ionic osmotic pressure term [Gehrke2000]. In cornparison, IT- and mi, contribute to the
s w e b g to a much smaller degree and ni, inhibits swelling by acting as a retractive
pressure on the gel solution. Attempts to mathematically mode1 the polyelectrolyte gel-
swelling phenornenon via caiculation of the various osmotic pressure ternis pasal 975,
Vasheghani-Farahanil990, Hariharan 1993, Doi 1992 J have been qualitatively successfil but,
as of yet, quantitatively inaccurate when correlated to experimental resdts.
Stimuli such as pH, ionic strength, electric field, electric current and photoirradiation
can al1 affect the number of ionized groups dong the polyelectrolyte gel, and, therefore,
niim. For example, weakly acidic or basic groups of polyelectrolyte gels will dissociate
depending upon the hydrogen ion concentration or pH of the smounding solution. The
extent of dissociation as a function of pH is ofien expressed as the dissociation constant or
pK, for that particular ionic group. Providing the ionic strength of the pH solution remains
constant, increases in solution pH result in decreases in hydrogen ion concentration. This
decrease causes the pendant carboxflic acid groups of PMAA to give up a hydrogen ion
(proton) and become ionized in order to maintain a certain equilibrium or ratio of ionized to
unionized species in solution (figure 3.2). This results in an increase in counterions in the
gel water and a corresponding decrease in the ionic osmotic pressure terni. The exact
degree of dissociation may be calcuiated fiom the dissociation constant. Likewise, as pH is
decreased, the number of ionized groups within the polymer decreases and the hydrogel
deswells. Thus, changing the pH of the surrounding environment may control sweIling or
hydration of a polyelectrolyte gel [Weissl986].
Immobilking enzymes such as glucose oxidase onto the polyelectrolyte gel has
effected changes in pH within a hydrogel. When the enzyme cornes in contact with glucose,
it converts glucose to gluconic acid, thereby altering the local pH within the gel, resulting in
membrane hydration and permeabiiity changes. This mechanism has been developed for
insulin delivery to diabetics [Ishihara1986, PoduaI2000]. Changes in pH may also occur at
the eiectrode surface due to generation of hydrogen ions by hydrolysis, when an electric
current is applied through the polyelectrolyte gel. Weiss reported pemeability changes for a
neutral solute across a poly(MAAc) membrane due to the increase in pH of the difision ce11
by electric current application, resulting in ionization of the membrane Weiss1 9861.
Buffer composition and ionic strength also affect the swelling of polyelectrolyte
hydrogels. As ionic strength increases, the swelling decreases due to shielding of charges on
the polymer chah and, more importantly, increased counterion concentration which
substantially reduces the concentration difference of ions inside and outside the gel, thereby
decreasing ioniç osmotic pressure and gel hydration.
Electnc field is another stimuli known to affect the ionic osmotic pressure terni via
changes in the local counterion concentration within the gel. Ionized groups may be created
via the electrodiffiision of counterions or by the protonation of ionized polyelectrolyte
networks [Tomer 1995, Kwon1990J. For example, application of an electrical field to
crosslinked hyaluronic acid caused rapid deswelling of the hydrogels, due to the partial
protonation of the ionized polyelecîrolyte gel network -1 9991.
Responsive hydrogels with more than one mechanism of hydration change have also
been prepared by combining different stimuli-responsive poIpers, resulting in new
materials with interesting release properties. For example, when temperature sensitive
polymers were combined with pH-sensitive polymers, the LCST of the copolymer was
especially sensitive to pH, due to the strong hydrophilic character of the ionized state of the
pH-sensitive component. In some cases the LCST phenornenon of the copolymer was
eliminated when the gel was in the ionized state r]HofEnan 1 9951.
Siegel and Firestone [Siegel1 9881 prepared hydrophobic polyelectrolyte gels fkom
hydrophobic methyl methacrylate (MMA) and ionizable N,N-diethylaminomethacrylate
@MA). The 70/30 MMA/DMA copolymer gel dernonstrated a sharp transition in
hydration fiom 0.1 to 0.7 at pH 6.6. This change in hydration was a significsult
improvement over hydrophilic polyelectrolyte gels that are known to undergo relatively
srnaller changes in hydration as a fiinction of pH due to their large osmotic pressure of
mixing tem. In the unionized state the DM-MMA copolymer is essentially hydrophobic
resulting in a very small osmotic pressure of m m when immersed in aqueous solutions.
As the pH fdls below 6.6, the copolymer becomes ionized significantly increasing the
osmotic pressure of the gel via the osmotic pressure tenn leading to a large increase in gel
hydration. Udortunately the slow response time of hydration to changes in pH which
bracket the swelling phase transition made these materials irnpractical for use as variable
permeability membranes. They have demonstrated potential as matrices for oral dnig
delivery applicaîions~irestone 19881.
In sumrnary, the swelling process of a gel can be descnbed in t e m of several
independent contributions to the osmotic swelling pressure (IItotal), which is equal at
equilibrium, to the externally applied pressure. For fiee swelling conditions, the externally
applied pressure equals zero so that
IItotai = IImix + Glas + Hion + alec = O at equilibrium.. . . . . . . . ..... (3 -2)
When any of these parameters is altered due to exposure to stimuIus, gel hydration
and its dependent gel properties, (i.e. solute permeability) will be affectecl. Accordingly
research in the area of responsive polymeric h g delivery has focussed on effecting changes
in gel hydration via changes to one or more of these osmotic pressure ternis, in order to
produce self-regulated dtug delivery profiles.
3.2.2 Responsive Hydrogels as Drug Delivery Membranes: Current State of Art
Since most peptides or proteins are impermeable ùirough dense polymers, the type
of polymer cornmonly developed for peptide h g delivery applications are hydrogels. In
addition to king permeable to larger, water-soluble species, many hydrogels are
biocompatible and their structure and physical properties can also be altered via a variety of
stimuli so that responsive dnig delivery is a possibility [Yoshida1993].
As with any burgeoning technoiogy, there are some disadvantages that must be
overcome in order for stimuli-responsive hydrogels to be effectively used in the control
element of dnig delivery systems. They include: 1. Slow response time 2. Low mechanical
strength and 3. Low ON/OFF permeability ratios.
3.2.2.1 Slow response time
The slow response t h e of hydrogels is mainly a function of swelling kinetics and
gel configuration IGehrkel990J. Gel hydration modulated by chernical stimuli is difision
limited. Since the time scale of diffusion is inversely proportionai to the square of the length
scale of the gel, gel dimension plays an important role in determining the tirne required for
these drug delivery systems to respond to input. Using thinner membranes may shorten
response times, but mechanical strength may be compromised as a result. Another approach
has been to prepare hydrogel microspheres. In these situations the rnicrospheres would be
used as matrices rather than membranes ll\Jakazawal996].
More recently, superporous hydrogels were prepared with pore sizes in the range of
100 pm using l4IPAA.m and acrylarnide [Chen1999]. The hydrogels were polymerized and
crosslinked in the presence of gas bubbles. The connected pores fonned open capiliary
channels that provided a thousand-fold irnprovement for gel swellingdeswelling response
times. This approach compromised the mechanical strength of the hydrogel and so
hydrophilic particulate materials were added to the hydrogel to form composites
[Chen2000]. Mechanical strength of the hydrogel was irnproved due to the increased
physical crosslinks between gel and particulates.
Another approach has been to synthesize a poly(NIPAAm) hydrogel by using a
water/acetone mixed solvent. Because NIPAAm consists of both hydrophilic and
hydrophobic groups, when polymerized in a mixed solvent the polymer c h a h are more
soluble and expanded, producing a polymer system with a greater tendency to deswell
[Zhang2000].
Poly(]?3iPAAm) gels have also been prepared with poly(ethy1ene glycol) gr& chains
having fieely mobile ends. The graft-type gels demonstrated rapid deswelling changes in
response to temperature inmeases. The graft chahs foxmed water releasing channe1s during
the gel desweiling changes facilitating water outflow fiom inside the gels. In contrast, gels
without graft chains demonstrated a slow deswelling change due to formation of a dense
skin layer at the surfàce which retarded water outfiow ftom the gels meko1998,
Yoshidal995, Kaneko 19961.
Using membrane support structures where the responsive gel is incorporated into the
pore or the sutface of the membrane [Pengl998] has increased the mechanical strength of
responsive hydrogels. Stronger responsive hydrogels have also been prepared by
copolymerization with a more hydrophobic cornonomer wuratore2000], or by formation of
composites with a stronger hydrophobic material [Chen2000]. Interpenetrating polymer
networks (IPNs) of PNIPAArn with polyurethane domains were synthesized. The presence
of the urethane network improved the mechanical strength, but reduced swelling and dmg
release rates due to its hydrophobic characteristics birn 19971.
3.2.2.3 Lo w ON/OFF drug permeabifity ratios
Gehrke et al [Gehrke1989] were among the fmt to note that although diffusion of
various solutes through gels of different hydration were consistent with trends predicted by
f k volume tbeory, the diffusion coefficients of the solutes did not change very much as a
fùnction of hydration. They concluded that for vitamin B12, a dramatic increase in the rate
of solute diffusion into the gel during swelling was not likely. Swelling ratios varied fiom
6.8 to 47 and the diffusion coefficient varied fiom 1.2 to 3.7 x 106 cm2/s, with an ON/OFF
ratio of 3.
Permeation control was greatly improved when responsive hydrogels were gdted
on surfaces [Ito2000] of porous membranes or within their pores [Peng2001j as polymer
brushes. The grafted polymers acted as gate valves where pore size was regulated by the
extent of hydration
3.3 Heterogeneous Polymer Systems
A heterogeneous hydrogel, in the context of drug delivery, was defmed by Bae
[Bae 19931 as a material derived from a mixture of polymers with opposing characteristics
demonstrating micro or macrophase separated structures tbat was expected to lead to
improved mechanical properties as well as novel release mechanisms. For the hydrophilic-
hydrophobie systems used in controlled release applications, the microdomains were
expected to provide improved drug selectivity, spatially dependent drug transport, and a
greater degree of solvation mland19931.
Heterogeneous hydrogels have been developed in the form of graR and block
copolymers, composites and interpenetrating polymer networks. Bae et al @3ad 9911
prepared heterogeneous interpenetrating polymer networks composed of a polyurethane
network and hyàrophobichydrophilic balanced vinyl network. The materiai was used as a
h g delivery matnx for both hydrophilic and hydroghobic solutes. The resuitant
morphology was not examined. Ulman et al ~lmanl989J prepared PDMS-PEO gr& and
block copolymers to be used as drug delivery matrices.
Merriil's group [Sung19901 also prepared copolymers of PEO and PDMS. Dmg
release of hydrophilic solute fiom a series of networks of varying FE0 content suggested
that for networks containing less than 35% PEO, the domains had fonned a discontinuous
phase remking in markedly reduced pemeability.
Hofhan's group [Dong1990] prepared the f k t stimuli-responsive heterogeneous
system consisting of a thermally reversible hydrogel with hydrophobic domains for delivery
of hydrophobic drugs. NiPAAm and bis-vinyl teminated PDMS were gamma irradiated,
DSC was used to confinn the existence of microdomain structure. The material was used as
a matnx to provide zero order release of progesterone.
Hofhan more recently prepared a hydrophobicdly-modified polyelectroIyte
hydrogel by gra£üng oligomers of methyl methacrylate to the backbone of poly(acry1ic acid)
hydrogel for use as a drug delivery matrix of hydrophilic and hydrophobic drugs
Doue1997J. The addition of the methymethacrylate oligomer graA enhanced the h g
release of hydrophilic solutes due to the enlargement of the aqueous pore size of the
hydrogel. However, the release rate of hydrophobic solutes decreased due to the adsorption
of the dmg onto the M ' A domains and the lack of intercomection between the
hydrophobic domains.
3.3.1 Elastomers
Elastomeric materials consist of relatively long polymeric c h a h that exhibit a high
degree of flexibility and mobility and that are joined into a network structure. Upon
application of an extemal stress, the long chains alter their configurations rapidly because of
the high chah mobility. As the chains are stretched and become more ordered, entropy
decreases. When the force is released, the elastomer retums to its original configuration
with concomitant entropy increase. In orslinary solids at high defonnations, the atoms slide
past each other and either flow takes place or the material fractures. The response of
e l a s tom is entirely intramolecular. Extemally applied forces are transmitted to the long
chahs where each chah acts like a spring in response to the extemal stress Wk19941.
The elastomer used in the heterogeneous systems of this thesis is crosslinked
poly(dimethy1siloxane) (PDMS). Rather than a carbohydrate backbone, as is found in most
polymers, PDMS consists of silanol functional groups, a silicon -oxygen backbone, with
methyl groups attached to the silicon atom (figure 3.3).
Figure 3.3 Chemical Structure of Poly(dimethylsi1oxane).
This unique structure leads to four fundamental characteristics that set the polymeric
properties of PDMS apart &orn all other polymers [Clarsonl993].
They are 1. the low intemolecular forces between methyl groups, 2. the unique
flexibility of the siloxane backbone, 3. the hi& bond energy of the siloxane bond and 4. the
partial ionic nature of the siloxane bond.
These characteristics of PDMS lead to many unique properties that include:
unusually low bulk viscosity, unusually high pemeabiüty, low surface tension (16-2 1 d m
at 20 OC), moderaâe interfacial tension against water, hi& water repellency, but high
permeability to water vapour, large fiee volume, low glass transition temp (-1 25 OC), and
inert material with good biocompatability.
In summary, PDMS is particularly suitable for biomedicai applications due to its
good biocompatability and inert nature. However, it is one of the more difficult polymers to
blend, particularly with hydrophilic polymers. Due to its very low surface tension, PDMS
has a tendency to segregate and form a surface layer upon &hg with other polymers.
Thus, even though the bulk may consist of a blend of two polymers, the surface morphology
is that of a homogeneous polymer [Clarson1993]. This is an important consideration for
materials where bicontinuity is necessary, such as membranes and in biomedical
applications where s h c e chemistry plays an important role in biocompatability issues.
3.3.2 Composites
Polymer composites consist of two or more physically distinct and mechanicaily
separable materials. The t em iypically refers to polymers that have been reinforced with
filler, such as a fibre or a particulate, in order to improve properties such as mechanical
strength. They are prepared by mixing separate materials in such a way that the dispersion
of one material in the other improves some desired property. In most cases, the k a 1
properties are superior to the properties of their individual components and are dictated by
the morphology of the blended system.
One of the first composites prepared in the field of controlled h g delivery was the
combinatim ~f hydrophobic silicone elastomers with low molecular weight compounds
[Follcman1964]. Langer used sirnilar composite systems of ethylene-vinyl acetate matrix
and dispersed polypeptide powder particles for the sustained release of macromolecules
banger1976, Hsu1985, Siegel1 9841. Difiion was found to occw through the
interconnecting pores created by the dissolution of the polypeptide particles.
Composites can also be a convenient means of blending two othemise incompatible
polymers, as has k e n carried out in this thesis work where dry crosslinked PMAA gel
particles (40 pm diameter) were rnixed with PDMS resin with subsequent crosslinking of
the PDMS matrix. Similar methods have been used to prepare elastomer-hydrogel
composites for biomedical applications.
Lopour et al [Lopour1995] prepared silicon rubber-hydrogel composite materials
(silicone elastomers filled with very fine particles of hydrogels) that were permeable to low
molecular weight, water-soluble compounds. They fomd that in spite of the hydrophobie
rubber mat& the materiais were highly permeable to low molecular weight solutes and the
composites behaved as hornogeneous water-swollen hydrogels (30% water content) based
on concepts developed fkom fite volume theory.
DiColo et al [Carelli1995] also prepared a silicone-based matrix containing
crosslinked polyethylene glycol (PEG) granules loaded with various solutes to be used as a
drug delivery maîrix for oral applications. The PEG granules (354-425 pm size range) were
loaded to a 35% weight fiaction in order to irnprove upon the fiaction of drug released with
other osmotic agents such as sodium chloride. The irnproved release was due to the ability
of the crosslinked PEG to form swollen hydrogels in the matrix upon hydration, without
dissolving and diffùsing into water.
Solute flux for a composite membrane of PNiPAAm dispersed in a gelatin matrix
was found to increase 3-4 fold when the PNiPAAm particles deswelled with increasing
temperature. It was hypothesized that the P'NiPAAm particles in the composite acted as gate
valves, where permeation hcreased when the particles shnink and decreased when the
particles swelled [Chunl996].
Using a similar approach, composite microcapsules (d = 100 p) were prepared that
consistecl of a core of drug particles and a responsive coating of an ethylcellulose mahrix
containing nanosized thermosensitive hydrogel particles (1 5% w/w) [Ichikawa2000]. The
shrinkage of hydrogel particles as the temperature increased created voids in the coating,
imparting higher water penneability to the coating. An O W F F drug delivery ratio of 15
was achieved with permeation response times of less than one minute. For hydrogel particle
loading of less than IO%, the voids did not contribute to the release of solute fiom the
ethylcellulose matrix. At loadings of 20% or more, h g release was no longer
thennosensitive due to the formation of connected hydrogel networks that led to rapid solute
release.
3.3.3 Interpenetrating Polymer Networks
Interpenetrating polyrner networks (IPNs) are multicomponent materials comprised
of two or more crosslinked networks that are at le& partially interlaced on a molecular scale
but not covalently bonded to each other, and cannot be separated unless chemical bonds are
broken [Sperling1997]. IPNs are distinguishable h m blends, block copolymers, and graR
copolymers by (1) their ability to swell but not dissolve in solvents, and (2) suppression of
their creep and flow. The interlocked structure of the crosslinked components ensures the
stability of the bulk and surface morphologies, regardess of the miscibility of the two
polymeric component. [Sutharl997]. IPNs are also known to fom frnely divided phases of
only tens of nanometers in size and to exhibit dual phase continuity, where constituent
polymers form phases that are continuous on a macroscopic scale.
lPN rnorphology is largely determined by the phase sepmtion of the polymer
components during IPN formation. In this thesis, hydrogel-elastomer IPNs were prepared
using the sequential method. Briefly, the elastomer network was first synthesized, then
swollen with the monomer of the hydrogel (the guest polymer). Polymerization and
crosslinking of the monomer to form a hydrogel network within the elastomer was then
effected. In such systems, phase separation occurs as the guest monomer begins to form its
own independent network within the host polymer network due to increases in molecular
weight caused by both polymerization and crosslinking reactions. This process is termed
polymerization induced phase separation (PIPS).
Morphology development in PIPS is intimately related to the competing and
interdependent rates of polymerization, crosslinking and phase separation. It is the increase
in molecular weight caused by the polymerization reaction that provides the therrnodynamic
driving force for phase separation. At the same tirne, polymerization and crosslinking
reactions increase the viscosity of the system and freeze the morphology developed during
the phase separation process due to vibrification and network formation. Thus phase
separation in IPNs is not allowed to reach îhermodynamic equilibrium, but rather the
morphology is fiozen at an intermediate stage of the phase separation process.
From a thennodynamic viewpoint, phase separation is the result of a change in the
fiee energy of the system. According to the Flory-Huggins equation, the fiee energy of
mixing is a fiinction of the temperahue, volume fiaction and molecular weight of the
polymer components. Figure 3.4 shows a fixe energy curve of rnixing for two polymers as a
function of polymer volume fiaction at a particular temperature and polymer molecular
weight. The two polymer components are completely miscible when AG is negative over
the entire composition range (Iine a). The two polymer components are completely
immiscible when AG is positive (Iine c). Line b reflects a phase separation where a
transition occurs fiom the miscibIe to the immiscible state. Line b is characterized by a
curve with two minima, one maximum and two inflection points. The inflection points in
the curve, defïned as ô AG / @2 = O, represents the thennodynarnic condition
I linc c
I line a
Palper volume Etaction Solvcnt
Figure 3.4. Miscibility curves corresponding to rniscibility (line a), phase separation (line b) and immiscibility (line c) [Kieffer 19991.
for spinodal decomposition @iefer1999]. For any composition between the two inflection
points, the mixture will phase separate by spinodal decomposition. The two minima of the
fiee energy curve correspond to points that satis@ the thermodynamic conditions for
equilibrium, where both points have the same chemical potential ( = ) because they
have a common tangent 3 AG / i3) = p . A phase separation diagram for a PIPS process, as
shown in figure 3.5, illustrates the phase separation behaviour as a function of the increase
in molecular weight (or % conversion) of a polymer system with a lower critical solution
temperature (LCST). The schematic phase diagram contains two lines and several regions.
The b e r line is called the spinodal line and the outer line is the binodal line. The binodal
Iine results fiom the fkee energy curves of the system as a function of % monomer
conversion, by interconnecting al1 the points having a cornmon tangent (similar chemical
potentiais). The spinodal iine results fiom the summation of inflection points of fiee energy
curves as a function of percent conversion.
Figure 3.5: Phase diagram of a polymer mixture with a lower critical solution temperature.
The transition fiom one area of a phase diagram into another is called a quench,
where quench depth is defmed as the difference between the LCST and the reaction
temperature. PIPS undergoes a chemical quench due to the continuous increase in
molecular weight causeci by polymerization and crosslinking reactions as opposed to the
more widely studied thermally-induced phase separation (TIPS) resulting fiom a single
temperature quench. The phase separation mechanism and the fmal morphology depend on
the region that is entered during the quench [Kiefer1999]. If the meta-stable region is
entered, phase sepmation occurs via the nucleation and growth mechanism (NG), in which
isolated concentration fluctuations with an equilibrium composition initially appea. and then
grow to yield an irregular two-phase structure Fouel 9951. If the unstable or spinodal
region is entered the concentration fluctuations begin smail and grow in amplitude and
wavelength as phase separation proceeds, to fom interconnected guest polymer domain
structures. Depending upon the viscosity changes occurring during the increase in
molecular weight, the bicontinuous structure may coarsen with tirne and form large, discrete
or intercomected globules
In the past decade it has been found that morphology formed via PIPS is very
different from that fonned by TES. This has been attributed to the concentration
fluctuations formed during the phase separation process. A blend of two polymers contains
fluctuations in concentration due to random thermal motion of the blend components. In the
single phase regime the fiee energy darnpens these thermal fluctuations and a one phase
system is maintained. As one approaches the critical point the Eee energy curve flatte- out,
especially in the vicinity of the critical composition so this damping effect is less
pronounced allowing concentration fluctuations to grow at the critical point.
Beyond the critical point, within the spinodal regime, phase separation is
spontaneous for ail composition fluctuations. This means that any concentration fluctuation
leads to a l o w e ~ g of the free energy. Usually on a micron scale a light scattering peak is
observed in the scattering for polyrner systems undergoing spinodal decomposition. The
peak reflects a "wave-like" fluctuation in the composition with a preferred wavelength. The
system contains some mtural size-constant which may be viewed as being similar to a time-
constant for a temporally oscillating system such as a pendulum, where oscillations are in
concentration rather than in tirne. The system selects certain "fiequencies" in spatial size
which are preferred for growth. At the early stage of demixing, the concentration
fluctuation becomes monochromie; i.e. sine waves of concentration with wavelength Am
predominate [Inoue1995]. The sine waves overlap with each other to render a replarly
phase separated structure s h o w schematically in Figure 3.6.
The periodic distance in this structure is nearly equal to Am. The wavelength, Am, is a
function of the quench depth ( Ts - T2 1 where:
1 is the interaction length, (typicaily 30 nm for a polymer-polymer mixture) and Ts is the
spinodal temperature. Equation 3.3 suggests that a deeper quench depth yields a shorter A,.
Figure 3.6: Schematics of the unit ce11 for a CO-continuous structure via spinodal decomposition and the concentration fluctuation along a straight line passing through the unit ce11 [Inouel 9951.
3.4 Permeabiiity Concepts For Responsive Heterogeneous Systems
nie main objective of this thesis is to control the delivery of solutes using a
responsive heterogeneous polymeric membrane composed of a hydrophobie
poly(dimethy1siloxane) component and a hydrophilic poly(methacry1ic acid) component.
Therefore, it is usehl to fmt understand the pemeation of solutes through each of these
very different polymers.
3.4.1 SoIute Diffusion Through Hydrogeis: Free Volume Theory
Two basic mechanisms, the partition and the pore mechanisrn are used to describe
penneation through gel membranes. These mechanisms are dependent on the
hydrophilicity/hydrophobicity of the polymer, it's pore size, degree of swelling, .sis well as
the size and hydrophilici@ of the solute [Gehrke2000, Kou2000J.
The pore mechanisrn of transport is associated with the gel's bulk water, where
solutes permeate the membrane by diffusion through the solvent filled pores. The polymer
network serves only to physically obstruct motion of the solute. The relative size of the
solute and pore are important for determining transport rates. Hydrophilic, low molecular
weight h g s are assumed to diffuse via water filled pores in the hydrogel by the pore
mechanism.
Many models have been proposed for the pore mechanisrn of diffusion through
polymer networks. The most sucçessfùl bas been the Free Volume Theory proposed by
Yasuda and CO-workers. This theory was based on the work of Cohen and Turnbull
[Cohenl1959, Turnbulll96 11 concerning the diffusion of hard spheres in a liquid. The
migration rate of a large sphere was proportional to the probability of fhding a hole of
volume V* or larger adjacent to the sphere and the volume in a liquid was composed of the
volume occupied by the liquid molecules and the fkee volurne surroundhg these molecules.
Using the pore mechanisrn of pemeation, fiee voIume theory arguments, and the
assumption that the polymer contribution to the fke volume was negligible, Yasuda et al
Tyasuda1969, Yasudal971] derived an expression for solute penneability through a
hydrogeI membrane, Pm in terms of the fkee volume, Vfof pure water only:
where Do is the solute diffusion coefficient through pure water, B is a constant, q2 is the size
of the solute, and <pz(q2) is a sieving factor that accounts for exclusion of solutes when the
solute size is Iarger than hydrogel mesh size. aH is the partition coefficient between the gel
and surrounding water, and cm often be approximated as a constant. H is the hydration
defined as the volume fiaction of water in the hydrogel, and q is the swelling ratio defmed as
the ratio of swollen gel volume to dry gel volume. This expression clearly indicated that for
a particular solute, changes in gel hydration or swelling ratio would lead to changes in the
permeability of the solute through the gel. Accordingly, the development of variable
permeability membranes has focused on producing changes in the hydration of hydrogels to
effect changes in permeability @ofban1 986, Gehrke 1 989, Bael9891.
3.4.2 Solute Diffusion in PDMS Elastomers
Liquid water and water-soluble solutes cannot diffuse through pure PDMS,
however, transport of water-soluble solutes through PDMS has occurred via aqueous
channels created in the material due to the presence of hydrophilic contaminants or the
intentional inclusion of salts or hydrophilic drugs. Water vapour diffuses rapidly in
polydimethylsiloxane (PDMS) [Schirrer 19921. Since PDMS is not permeable to dissolved
salts, the materiai acts as a semi-permeable membrane. The water vapour entering PDMS is
entrapped by hydrophilic inclusions and the osmotic pressure created in these pockets
becomes large enough to create microcracks around the pockets. The microcracks grow
creating small channels between the salt pockets and the sumounding water. The osmotic
pressure differential between the pockets and the extemal water creates a salt flow fiom the
pocket towards the water tbrough micro-channels. The salt or h g flow is constant, due to
the constant crack growth speed. Healing at the interface of these microcracks may occur
by interdiffusion. For PDMS this has been described in terms of the concentration of minor
chahs, where minor chains are known to play an important role in the healing process of
h t u r e d polymers [Kiml996].
In some cases, depending upon the modulus of the PDMS material, microcracks do
not fom. Water vapour diffuses hto the &ber, as before, and forms pools around particles
of water-soluble impurities. The intemal solution becomes diluted, and the swounding
rubber exerts a retractive pressure because of the local deformation. When a point is
reached that the retractive pressure of the rubber becomes equal to the osmotic pressure of
the interna1 solution, a state of equilibrium is established and water content reaches a
constant value. In these cases, diffusion of the aqueous solutes through the material does not
mur [.dwards1985].
Amsden and Cheng [Arnsden 1996, Amsdenl995, Amsden 19943 have used osmotic
excipient. to facilitate delivery of protein h g s fiom ethylene vinyl acetate polymers. The
osmotically induced membrane rupture subsequent to water imbibition aided the release of
protein dmg. The main parameters affècting release were osmotic activity, saturation
concentration and density of the incorporated agent, as weU as the tensile strength, elastic
modulus and hydraulic pemeability of the polymer.
3.4.3 Mass Transfer in Heterogeneous Media: Percolation Theory
Transport through porous systems, such as bicontinuous hydrogel-elastomer
materials, can be descnbed by percolation theory [Mohanty1982, SiegeI19891. If the
pemeating solute diffuses through only the hydrogel regions, then the effective d ih iv i t y
through the heterogeneous medium depends on the volume fiaction of the swollen gel in the
membrane (qsl). At low < P B ~ I , gel regions are isolated h m each other, so that the
accessible volume fiaction of gel regions, (PA, as well as the normalized effective diffiisivity
in the heterogeneous medium (relative to diflhivity in the homogeneous gel), D,fP' a, are
both zero. As <p,l increases, the percolation threshold, q@l,, is reached when the k t matrix
spanning connected gel cluster is formed. As <p,l increases even M e r , more isolated gel
regions are recruîted into connected clustas so that both c p ~ and De& increase, and both
reach a value of 1 at qgel = 1 corresponding to a homogenous hydrogel. The exact value of
4&,, and the shapes of the De& and < p ~ curves depend on the geometry of the gel regions.
3.4.4 Variable Permeability Membranes: Mechanisrns of Permeation Control
Mechanisms of diffusion through homopolymer or heteropolymeric systems having
constant properties are fhkly well understood. However, once properties, such as hydration
or for heterogeneous systems, connectivity and domain size, become dynarnic under the
influence of an extemal stimulus then transport mechanisms become more cornplex.
For variable perrneability hydrogel membranes, literature reports have included
examples where permeation is controlled by hydration changes Fei1 199 11 and also by
formation of a surface skin ~ o f h a . 1 9 8 6 , Mukael 990, Bae 199 11 or bulk squeezing of the
membrane [Hofltmanl986, Sawahatal990]. Crosslinked poly(NIPAAm-co-
butylmethacrylate) was used to separate out three solutes of different sizes, uranine, FITC-
dextran 4.4K and l5OK. As hydration decreased, permeation of larger solutes was impeded,
according to the fiee-volume theory, due to the effects of sohte size and water volume
hction in the membrane peil1991]. In some cases, immediate shrinking of the outer layer
restricts M e r bu& water outflow fiom the interior of the gel. This response of the gel
surface to temperature changes may be used as an on-off switch for h g release @3ael989].
In work carried out by Hofian et al ~ o f i a n 1 9 8 6 ] , poly(NiPAArn) and its copolymers
were shown to exhibit two pattern of temperature-modulated h g release; bulk squeezing
and swface regulation. When the swollen gels in drug solution at low temperature were
iransferred to release media at 50°C, the initial rapid release was followed by a slow release
rate. This phenomenon was interpreted as the squeezing effect accompanying gel
deswelling? which caused out flm of dissolved h g with water flow in addition to
diffusional flux. On the other hand, the copolymers of NiPAAm with more hydrophobie
cornonomers such as n-butylmethacrylate showed dense skin formation durhg the
deswelling process when the temperature increased past LCST values. The dense skin was
able to block the release of drugs fkom the matrix, limiting diffusion through the membrane
and resulting in the ON-OFF control of solute release. The repeated cycle of constant
release of loaded drugs during the ON stage was explained by the redistribution of drug
concentration inside the matrix during the OFF stage. Similar results were obtained f?om
IPNs composed of NiPAAm and PTMO Pae19911 or NiPAAm and PEO-DMS-EO
mukael 9901.
Polymeric complexes in gaft copolymers~elll994] and IPNsmishil986 ] have
been prepared fiom PMAA, PEG and PAA, PEO, respectively. The reversible formation of
polymer complex due to pH or ionic strengîh resulted in abrupt increase in hydration and a
decrease in obstructions due to the dissolution of ionic crosslinks or complexation between
the polymers. Both aictors allowed for a rapid and wide range of permeation control.
A novel approach to pulsatile h g delivery has been devised by Siegel et al.
[Siegel1 9951 where the hydrogel membrane acts as an oscillator that modulates dnig release
in a pulse-oscillatory manner. The system couples m a s transfer and enzymatic reactions to
gel swelling. The permeability of poly(NiPAArn-co-MAA) gel membrane to glucose was
inhibited due to gel dehydration caused by protons produced fiom the reaction of glucose
with the enzyme glucose oxidase in the receptor chamber of a permeation cell. The
decrease in glucose flux from the donor chamber decrezsed the production of protons in the
recqtor chamber, leading to an increase in gel hydration and glucose permeation. The
hystensis in the characteristic curve relating gel permeabiiity to pH indicated ptential for
the membrane as a repetitively pulsing drug delivery device [Bakerl996J.
Responsive hydrogels have been grafted onto the surface and within the pores of
porous, solid polymeric substrates parbucci 1991, Peng19981 and prous hydrogel supports
[Chun19961 for use as variable pemeability membranes. Pulsatile dnig delivery was
achieved due to the sweiling and shrinking of the responsive gel to externa1 stimuli. As the
gels swelled, the pores became fded with gel and pemeabiiity was obstructed/decreased.
As the gels shnink, pores opened alIowing permeation through. Peppas et al [Zhang2000]
synthesized a PMAA-PNiPAArn IPN, where permeation decreased above the LCST of
PNiPAAm (shrunken state). This was attributed to the deswelling of PNiPAAm, which
created more spaces or open gates in the swollen PMAA gel, in order for the solute to
permeate througb.
A heterogeneous hydrogel-elastomer membrane has recently been prepared fiom
p(NiPAAm) and PMMA [Lu2000]. The hydrogel formed the continuous phase and the
hydrophobie component, present at lower volume fiactions of S%, 10% and 15%, formed
the disconnected domains. This new material was stronger than the hydrogel, but swelling
kinetics and release rates were much slower than the hydrogel only. The primary
mechanism of permeation control in this membrane was hydration.
CHAPTER 4 : EXPERIMENT AL METHODS
4.1 Overview
Stimuli-responsive, heterogeneous materials that are to be used as membranes for
variable mass taansport applications must be: 1. responsive to stimuli, 2. have a continuous
permeable pathway and 3. be able to undergo reversible property changes. A good deal of
experimentation was required to acfiieve these properties for P m - P D M S composites and
TPNs. For example, the PMAA gel required good swelling response to changes in pH, but
also mechanical integrity to withstand handling. The crosslink density of the PDMS
network was chosen to provide optimum retractive pressure, so that the network remained
cohesive during handling, but was also able to deform and allow the incorporated gel
domains to swell and deswell. Furthemore, the morphology needed to be such as to allow
water and water-soluble solutes to difhse through the material. Thus bicontinuity in the
swollen state was also an important requisite.
The PMAA hydrogel and PDMS networks were prepared individually and then in
combination with each other as composites and TPNs. The fabrication of composites
followed relatively well hown procedures, although the formation of very small (micron-
sized) dry PMAA gel pariicles proved to be diffrcult. Various grinding apparatus were
tested and the best results were achieved by fkeeze drying the hydrogel in the swollen state
to produce a fiagile porous structure which was then crushed with a laboratory blender.
Another challenge was related to formulating a PDMS network that allowed the PMAA gel
to swefl and deform. In order to minimize the retractive pressure of the PDMS network, the
molecular weight between crosslinks (MJ was rnaximized. Commercially available vinyl-
terminated PDMS had a molecular weight in the order of 100,000 Da. Hydride terrninated
PDMS was used as a chah extender to iaiease the molecular weight of vinyl termjnated
PDMS. The terrninated hydride and vinyl groups on the PDMS molecules reacted via the
hydrodilyation reaction to form a longer PDMS rnolecule. Addition of this chah extender
allowed the M, of the PDMS network (1 66,000 Da) to increase to values larger îhan the
M.W. of the vinyl terminated PDMS (1 16,000 Da). The final PDMS formulation was
chosen based on a compromise between maximiWng M& the ability of the material to form a
cohesive structure, and stimuli-responsive hydration properties. Another important variable
in the composite fabrication was the amount of PMAA gel particle loading required in order
to fonn a percolating system. Again, formulations with various loadings were made using
hydration and perineation studies to determine the optimum loading.
The preparation of PMAA-PDMS IPN membranes posed various challenges as well.
Although IPNs could be easily fonned within the bulk of the PDMS network, the surface of
the IPN was not bicontinuous when the pre-IPN (PDMS network swollen with methacrylic
(MM) monomer) interfaced with either air or glas during IPN formation. Thus the
"moonomer-immersion" method was developed where the pre-IPN was immersed in MAA
in order to maintain a uniforrn concentration of monomer witbin the pre-ZPN during IPN
formation. This method produced a responsive, bicontinuous IPN membrane.
The membranes were characterized in t e m of their stimuli-responsive hydration
and permeation properties. Both swelling and permeation studies were carried out on the
homogeneous and heterogeneous membranes as a fùnction of pH. The morphology of the
membranes was also characterized in order to determine how morphology aflected the
membranes stimuli-responsive properties. The Laser Scanning Confocal Microscope
(LSCM) proved to be an excellent tool to examine the morphology of both the bulk and
surface of these membranes. Transmission electron microscopy (TEM), Field Emmision - Scanning Electron Microscopy (FE-SEM), Atomic force microscopy (AFM) and Electron
Spectroscopy for Chemical Analysis (ESCA) were aiso used in order to gather more
information regarding the surface of these membranes. However, these techniques were not
successfùl and will not be mentioned M e r in this section. Appendix A contains a brief
description of each method and a summary of the problems encountered when using the
method to analyze the surface of PMAA-PDMS IPN membranes.
4.2 Preparation of Membranes
4.2.1 Preparation of PMAA gel.
The materiah used to prepare poly(methacry1ic acid) gel particles and films include
methacrylic acid and triethylene glycol dimethacrylate polysciences Inc., Warrington, PA);
ethylene glycol (Fisher Scientific, Fairlawn, NJ); ammonium persulfate and sodium
metabisulfite (Sigma Chemical Co).
Poly(methacry1ic acid) gel films were prepared by mkhg 20 mL of rnethacrylic
acid with 8 rnL of distilled, deionized water and 3 mL of ethylene glycol in a round bottom
flask. The solution was dowed to mix for 15 minutes, at which point 1 mL of 20%
ammonium persulphate solution and 1 rd, of 7.5% sodium metabisuIfite solution along wiîh
a specified amount of crosslinking agent, triethylene glycol dimethacrylate (TEGDMA),
were added. The mass ratio of TEGDMA with respect to methacrylic acid monomer was
varied fkom 0.25% to 1% and 4%. The polymer solution was mked for an additional 15
minutes and then pipetted between two glass plates separateci by a 0.75 mm gasket.
Polymerization and crosslinking were carried out at 55 OC for 2 h.
The resultant PMAA f i h s were washed using distilled water for two weeks, with
the wash water being changed twice M y . 33 mm diameter sections of the P M film
were die cut for swelling and permeation studies.
4,2.2 Preparation of PDMS network
4.2.2.1 Materials
The materials used to prepare the PDMS network portion of HPG membranes were
vinyl terrninated PDMS (65,000 cst, 1 16,OO Da), hydride terminated PDMS (1 000 cst),
platinum-divinyltetramethyldisiloxane (United Chernical Technologies, Bristol, PA) and a
cyclic hydride-containing crosslinking agent (MDX4-42 1 0, Dow Corning, Midland
Michigan).
4,2.2,2 Synthesis
Polydimethylsiloxane (PDMS) resin (1 16,000 Da, 65,000 cst) containhg 60 ppm of
platinum diviny ltetra-methyldisiloxane complex and 4% MDX4-42 1 0, a cyclic,
multibctional silicone hydride crosslinker, was spread to a thicimess of approximately 0.5
pm on a mylar sheet using Polyethylene (PE) spacers. The resin was then placed under a
vacuum of 25 mm Hg for approximately 6 hours in order to remove any entcapped air
bubbles within the bulk of the fh. This was followed by a hydrosilyation addition reaction
at 55 OC for 24 h. The cured PDMS resin was die cut into circular sections ( d = 20 mm).
Each circular PDMS section was placed in cyclohexane with mixing for 24 h in
order to wash away any unreacted PDMS components. The degree of swelling in
cycloheiane was also used to estimate the apparent molecular weight between crosslinks of
the PDMS polymer network using the procedure in section 4.3.1.
4.2.3 Preparation of PMAA-PDMS composite membrane.
4.2.3.1 Materials
The materials for the preparation of PMAA-PDMS composites were the same as
those used in sections 4.2.1 and 4.2.2.
4.2.3.2 Synthesis
The swollen PMAA hydrogel film prepared as in section 4.2.1 was ground using a
laboratory mixer and îhen lyophilized. The lyophilized product was further ground ushg
the laboratory mixer. The dry PMAA gel particles were sieved to isolate particles in the 30-
45 pm size range.
HPG membranes composed of PDMS and varying amounts of PMAA gel particles
(1 7%, 22% or 28% on a dry mass basis) were prepared using the following procedure.
PMAA gel particles (30-45 pn in size) were prepared fkom PPvZQA hydrogel films
containing 0.25% TEGDMA. A specified amount of PMAA gel particles was added to a
mixture of 1.5 g of vinyl-teminated PDMS (65,000 cst) with 60 ppm platinum
divinyltetramethyldisiloxane catalyst, and stirred to ensure d o m dispersion. Then 0.2 g of
hydride terminated PDMS (1000 cst) and 0.02 g of MDX4-4210 crosslinking agent were
added and mixing was carried out using a homogenizer. The PMAA-PDMS gel particle
resin was cast in a teflon mold and degassed overnight in a vacuum oven. The polymer
mixture was then placed in an oven at 55 OC for 24 h to accelerate the hydrosilyation
reaction. The resdtant composite membranes were approximately 1 mm in thickness and
wwe die cut to a diameter of 33 mm. They were subsequently washed in distilled deionized
water for a 2 week pend to remove any umeacted components.
4.2.4 Preparation of P U - P D M S IPN membrane.
The materials used to prepare the PDMS host polyrner network were vinyl
termhated PDMS (65,000 cst), platinumdivinyltetramethyldisiloxane (United Chernical
Technologies, Bristol, PA) and a cyciic hydride-containing crosslinking agent (MDX4-
42 10, Dow Coming, Midland Michigan). The PMAA guest polymer network was prepared
fiom methaczylic acid, triethyIene glycol dimethacrylate (TEGDMA, Polysciences Inc.,
Warrington, PA) and 2,2 - dimethoxyacetophenone (Irgacure 65 1, Ciba Geigy), the UV
sensitive fiee radical initiator.
PDMS f i s , 1 mm in thickness and 20 mm in diameter, were prepared according to
the method in Section 4.2.3. The washed PDMS sections were immersed in methacrylic
acid monomer solution containhg 1 % w/v of 2,2 - dimethoxyacetophenone, and 1 % v/v of
TEGDMA crosslinker for approximately 18 h. The PDMS network absorbed
approximately 100 % of its rnass in MAA monomer solution. The monomer swoiien
PDMS network (pre-IPN film) was purged with N2, immersed into a MAA monomer
solution only (this solution âid not contain crosslinking agent or photoinitiator) and placed
under UV lights having an intensity of 32 W and a wavelength of 3SO nm for 1 h. The
resultant IPN, a hard, tough opalescent material was washed extensively in distille4
deionized water to remove any unreacted components and to reach equilibrium swelling.
Once completely washed, the IPN contained approxirnately 30% PMAA gel on a dry
membrane basis. More detailed description and figures of the apparatus used to prepare
PMAA-PDMS IPN membranes using the monomer immersion method is found in
Appendix B.
4.2.4.3 Synthesis: Air-IPN Interface Method
The circular PDMS sections, 1 mm thickness and 20 mm diameter, were placed in
methacrylic acid monorner solution as outlined in section 4.2.4.3. Using a glove box, the
pre-IPN film was purged with N2 and transferred to an empty scintilfation vial. The film
was placed agakt the wall of the glas vial so that one surface contacted the gIass wall and
the other forrned a fiee surface. The vial was capped and placed under W irradiation and
subsequently washed as detailed above. It should be noted that in the first experiments
using this method the IPN interfaced air. Later experiments the IPN was made to interface
N2 only. The results of both experiments were the same and the IPN is denoted as having an
air-IPN interface.
4.2.4.4 Synthesilr: Glass-IPN interface Method
Within a glove box, the pre-IPN film prepared according to the method in section
4.2.4.2 was transferred to a square cover glass slide. The film was smunded by a PE
spacer which was held in place using silicone grease. A second cover glass slide was placed
on top of the film. The pre-IPN film was essentially sandwiched between the two glass
slides and then placed under W irradiation for 1 h and subsequently washed as detailed
above.
4.3 Methods of Characterization
4.3.1 Determination of Mc for PDMS Networks
Molecular Weight between Crosslinks @&) was determined for PDMS networks
using the Flory-Rehner Equation [Treloar1975]:
where p is the density of the PDMS network before swelling
V1 is the molar volume of the solvent at room temperature
VZm is the volume fiaction of the PDMS network in the swollen state
x is the polyrner-solvent interaction parameter
The volume fiaction V2rn was determined by sweiling the fDMS network in
cyclohexane to equilibrim and measuring the swollen mass and dry mass of the network
using the following equation:
where & is the dry mass of the PDMS network in grams
Mf is the swollen mass of the PDMS network in grams
p p ~ ~ s is the density of the PDMS network in g/L
PcyC~dexane is the density of cyclohexane in g/L
The polymer solvent interaction parameter, X, which accounts for the specific
interactions between polymer segments and solvent molecules in a particular polyrner-
solvent system, was estimated by the bulk-cured mode1 [Brotzman1982].
Circular PDMS polymer network samples with diameters of 20 mm were placed in
20 mL of cyclohexane at r o m temperature for 72 h in sealed cuvettes. The equilibriurn
swollen mass of the PDMS network was measured and the solvent was evaporated £tom the
PDMS film mtil a constant mass was reached (approxirnately 24 h), at which point the mass
of the dry PDMS network was deterrnined. Corrections for the specific gravity of the
solvent, the amount of soluble material extracted, and the prevention of evaporative loss of
solvent fiom the specimen during weighing were îaken into consideration.
4.3.2 Determination of Mernb~qe Hydration
4.3.2. I Preparation of pH Buffer Solutions
The citrate buffer solution (pH 3,I = 0.015 M) was prepared by adding 1 L of
distilled, deionized water to 9.765 g of citric acid and 1 .O3 g sodium citrate. Citrate buffer
solution (pH 5, I = 0.012 M) was prepared by addhg 1 L of water to 4.307 g citric acid and
8.68 g of sodium citrate. Phosphate bufTer solution (pH 7 , I = 0.01 M) was prepared by
adding 16.352 g sodium phosphate dibasic and 5.166 g sodium phosphate, monobasic to 1 L
of water.
Swelling studies were carried out for PMAA gel films, composite membranes and
PDMS-PMAA IPN membranes in buffers of pH 3,5, and 7. Equilibrium mass, thickness
and diameter were recorded for PMAA films of different crosslinking density, composite
membranes containhg varying amounts of PMAA gel particles and IPN membranes
prepared using the "monomer immersion rnethod". The membranes were placed in a buffer
solution of particular pH and allowed to corne to swelling equilibnum, at which point the
above-noted measurements were taken. Three samples were evaluated for each
experimental condition.
4.3.2*3 Equations of Hydration for Membranes
Modified definitions of hydration and swelling ratio based on mas, rather than
volume, of water and PMAA gel were used since mass could be more accurately measured.
Hydration was defined as either
based on PMAA gel only, or
- mass of water H ~ p ~ - total mars of W O Z ~ HPG
based on the entire HPG membrane,.or
- mass of water HP* , m m - mass of water+mass of dry PUAAgel particles (4.7)
Likewise, swelling ratio based on PMAA gel only was defmed as
mass of swoZZen P W gel 'B'= rnass of d,y P M gel
while swelling ratio of the entire HPG membrane was defmed as
total mass of swollen HPG q*pG'mass of d y PDM + mass of dry PMAA gel particles
and the swelling ratio of the gel particles in the HPG membrane was def~ned as
- mass of water + mass of diy PMAA gel particles in HPG - (4.1 0) mass of d y P M M gel particles
4.3.3 Pre-Equilibrated and Dynamic Permeation Studies
CaEeine (MW 198, Aldrich) and vitarnin B 1 ~ (MW 1,355 Sigma Chemical Co.)
were used as the mode1 permeants for equilibrium and dynamic permeation studies for
composite membranes and vitamin BI1 only was used in permeation studies for IPN
membranes. Al1 permeation studies were camied out at 37OC using standard two
cornpartment &ion cells. The donor chamber was filled with the penneant (cafTeine or
vitamin Bl2) dissolved in a bufEer solution of specified pH. The receptor chamber was filled
with the same buffer but did not contain permeant. The receptor chamber was sampled at
specified tirne intervals. The volume removed was replaced with fiesh buRer. Approximate
infihite sink condition was maintained in ali experiments. The sarnple solutions were
monitored at 272 nm for caffeine and 362 nm for vitarnin B12 ushg a Hewlett Packard 8450
UV-Vis Spectrophotometer.
For pre-equilibrated shidies, permeation was carried out using homogeneous PMAA
gel films, composite membranes or IPN membranes pre-equilibrated with buffer at pH
conditions correspondhg to permeation experirnents. Solute pemeability through the
membrane was calculated fiom the average steady state flux based on the swollen
membrane thickness.
For dynamic pemeation studies, the membranes were pre-swollen and equilibrated
in pH 3 buffer. Permeation was monitored for a specified time period, then both donor and
receptor solutions were replaced by solutions of pH 7 while the membrane remained
clamped to the diffusion cell. Permeation was monitored for 24 h starting immediately after
the pH Increase, then both the donor and receptor media were replaced with pH 3 solutions
while monitoring of the receptor solution continued. Permeability was calculated fkom the
average h g flux during each 24 h period. EquiIibrium composite and IPN membrane
thkkness at the appropriate pH was used for the pemeability calculations.
4.3.4 LSCM Studies
Fluorescein (sodium salt, 332 Da, hydrodynamic diameter (h.d.)=0.8 nm) and
fluorescein isothiocyanate dextran ((FITC-de-) 4,000 Da, h.d.=9.4 nm and 70,000 Da,
h.d.=37.4 m) were used as the hydmphilic fluorescent probes for LSCM studies. Although
LSCM has been used to examine the morphology of blends and composites, there are no
examples in the literature where it has been used for IPNs, possibly because IPN domain
sizes are usuaily much smaller than LSCM resolution (1 00 - 200 nm). In this work, we
demonstrated that by exarnining PMk4-PDMS IPNs in their swollen state (pH 7), the
hydrogel domains were sufficiently large that LSCM visualization of the morphology was
possible.
LSCM studies were caxrîed out using an Înverted microscope, the Car1 Zeiss LSCM
5 10 using a 488 nm Argon laser. Two diensional optical sections at diHmnt depths dong
the optical axis of the microscope (z direction) were taken with a water-immersible x63
objective lem (NA. = 1.2). The size of the images obtained in the x-y plane was 29.2 p x
29.2 p. The image was captured with a resolution of 5 12 pixels x 5 12 pixels and 8-bit
grayscale depth that was later converteci to an 8-bit duotone scale with red filter. Resolution
along the z-axis was 0.65 p. The spatial resolution was approximately 244 nm. Images
were taken at the surface and at correspondhg planes below the surface at 1 jm intervals.
Use of a fluorescent marker (fluorescein sodium salt) to distinguish the hydrogel
domains fiom the rubbery network allowed for sharp image contrast between the rubber and
hydrogel regions. The red regions in the LSCM images represented the PMAA hydrogel
domains of the IPN membrane that were accessible to the fluorescein probes. The black
regions represented fluorescein-free portions of the IPNs, either PDMS or isolated PMAA
hydrogel domains that were inaccessible to fluorescein. PDMS networks immersed in
fluorescein solution for a one month p&od produced images which were black. This was
attributed to the irnpermeability of the PDMS network to the fluorescein solute which was
visually confhned by the absence of yellow colour in the PDMS network. LSCM images
of the PMAA gel irnmersed in the fluorescein solution for a 24 hr pend resulted in dark red
images throughout the depth of the PMAA gel. These results were expected since the
hydrophilic fluorescein solute could not permeate through PDMS but could easily pemeate
through the PMAA gel. Images of different IPN samples were taken at three different spots
on the sample and&ee different sections were analyzed for each different PN. The
microscopy results were found to be very reproducible and two samples prepared in the
same way showed similar tomorgraphic results. The images were fiuther analyzed using
Cord Photopaint to obtain percent fluorescent area as a fùnction of depth for each IPN-
interface system. Al1 pixels in a bitrnap image above a certain red tonal value (threshold
value = 41) were defined as being fluorescent. Since fluorescent areas represent the
accessible gel regions in the membrane, these figures allowed us to assess the relative
accessible gel regions as a fùnction of depth in a semiquantitative fashion. The results were
not used to compare the difTerent IPNs since fluorescent intemify was highly dependent on
LSCM parameters and membrane conditions at the t h e of image capture.
CWAPTER 5 :PRZAA-PDMS C OMPOSITE MEMBRANES
In chapter 5 the preparation of PMAA-PDMS composite membranes, their
mechanism of penneation control and their performance as stimuli-responsive, variable
permeability membranes are discussed. The mechanism of permeation coneol was
dependent primarily on the percolating volume fraction of the PMAA gel particles. The
critical percolation threshold for diffusion of vitamin BI:, in PMAA-PDMS composites with
28% gel loading occurred at a hydrated gel volume hction of 0.38. The hydrated gel
volume fiaction of the composite membrane could be rnanipulated with pH to fa11 above
(pH 7, ON state) or below (pH 3, OFF state) the percolation threshold in order to Vary the
permeability through these membranes. Although the OFF state did not produce zero flw
(P at pH 3 = 1.7 x 1 0 ~ cm&, the permeation response of the membrane to changes in pH
occurred in the order of minutes and the ONIOFF vitamin BI2 permeability ratio was 160 (
equilibrated ONIOFF vitamin B12 pexmeability ratio for PMAA was 7). Thus, there was a
significant improvement in the penneation properties of composite membranes compared to
the hornogeneeous PMAA hydrogel.
5.1 Introduction
One of the fmt composites prepared in the field of controlled dmg delivery was the
combination of hydrophobie silicone elastomers with low molecular weight compounds
~oikman1964], Langer et al. used similar composite systems of ethylene-vinyl acetate
mat* and dispersed polypeptide powder particles for the sustained release of
macromolecules Langer1 976, Hsu1985, Siegel 1 9841. Difision was found to occur
through the interconnecteci pores created by the dissolution of the polypeptide particles.
59
Composite systems can also provide a convenient means of blending two othenvise
incompatible polymers, as has been carried out in this thesis work, where dry crosslinked
P m gel particles ( a 0 pn diameter) were mixed with PDMS resin followed by
crosslinking of the PDMS matrix. Similar methods have been used to prepare hydrogel-
elastomer composites for biomedical applications.
Lopour and Janatova [Lopour1995] prepared silicone rubber-hydrogel composite
materials (silicone elastomers filled with very fine particles of hydrogels) that were
permeable to low molecular weight, water-soluble compounds. They found that in spite of
the hydrophobic rubber matrix, the materials were highly permeable to Iow molecular
weight solutes and the composites behaved as homogeneous water-swollen hydrogels (30%
water content) based on concepts developed fiorn fke volume theory.
Carelli et al. [Carelli1995] also prepared a silicone-based matrix containing
crosslinked polyethylene glycol (PEG) granules loaded with various solutes to be used as a
drug delivery matrix for oral applications. The PEG granules (354-425 pn size range) were
loaded to a 35% weight hction in order to improve upon the fiaction of drug released with
other osmotic agents such as sodium chloride. The improved release was due to the ability
of the crosslinked PEG to fom swollen hydrogek in the matrix upon hydration, without
dissolving and diffushg into water.
Solute flux for a composite membrane of PNiPAAm dispersed in a gelatin matrix
was fond to increase 3-4 fold when the PNiPAAm particles deswelled with increasing
temperature. It was hypothesized that the PNiPAAm particles in the composite acted as gate
valves, where pameation increased when the particles shrunk and decreased when the
particles swelled [ChunlW6].
Using a similar approach, composite microcapsules (d = 100 p) were prepared that
consisted of a core of drug particles and a responsive coating of an ethylcellulose matnx
containhg nanosized thennosensitive hydrogel particles (1 5% w/w) ~chikawa2000]. The
shrinkage of hydrogel particles as the temperature increased created voids in the coating,
imparling higher water permeability to the coating. An ON/OFF drug delivery ratio of 15
was achieved with permeation response times of less than one minute. For hydrogel particle
loadings of less than IO%, the voids did not contribute to the release of solute fiom the
ethylce1ldose matrix. At loadings of 20% or more, drug release was no longer
thennosensitive due to the formation of connective hydrogel networks that led to rapid
solute release.
In this work, we attempted to combine the responsive, permeable nature of
polyelectrolyte gels with the mechanical shrength and unique transport properties of
composite materials. In this chapter we focused our work on composite membranes that
consisted of polyelectrolyte gel particles dispersed within a mechanically stronger
hydrophobic elastic network. We hypothesized that (1) the polyelectrolyte gel particles
would remai. responsive to extemal stimuli when dispersed within the hydrophobic matrix,
aiiowing for extemal control of gel particle hydration, and that (2) extemally triggered
changes in hydration would result in changes in the polyelectrolyte gel volume fiaction
(<p,,) and, therefore, effective dihivity through the composite. Ttius hydration and
percolation would act synergistically to enhance the permeability response with respect to
changes in hydration. Furthemore, extemal triggers could dynarnically control transport
through composite membranes. Our hypotheses are shown schematically in Figure 5.1.
5.2 Methods
PDMS films, PMAA gel films and particles were prepared according to the methods
described in sections 4.2.1,4.2.2 and 4.2.3, respectively. Pre-equilibrated and dynamic
hydration and permeation studies were used to characterize the PMAA gel film, and the
PMAA-PDMS composite. These results were also used to determine the mechanism of
permeation change of PMAA-PDMS composite membranes.
5.3 Results and Discussion
5.3.1 Equiübrium SwelIing Studies
Figures 5.2 and 5.3 show the swelling ratio (q) and hydration (H), respectively, as a
function of pH for (a) PMAA gels prepared with 0.25 % TEGDMA (%I or H@), @)
composite membranes containing 28% dry gel pariide loading (qmpi& or and (c)
the PMAA gel (0.25% TEGDMA) particle in the composite membrane (%i, -mit, or Hgei
For PMAA gels a large change in swelling ratio, h m 8 to 16, was observed
between pH 5 and 7 (Figure 5.2), in agreement with literature reports that carboxylic acid
groups in PMAA have a pKa of 5.5 [Shatayeval979]. Hydration of the PMAA gel as a
function of pH (figure 5.3) increased fiom 0.75 at pH 3 to 0.93 at pH 7 and 9. No distinct
jump in hydration was apparent near the pl&. Because the hydration of the un-ionized gel at
pH 3 was already quite hi&, the increase in swelling brought about by ionkation did not
result in a drarnatic increase in hydration in the region of the p&. It should be noted that
larger changes in hydration occurred for gels prepared with 1 .O% and 4.0% TEGDMA.
However, it was found that these gels were dficult to work with because they were quite
brittle at the lower levels of hydration and &cause they often broke apart when clamped or
when the extemal pH was changed fiom 3 to 7.
The hydration and swelling ratio of composite membranes, brnpik and qmpik,
were lower than the homogeneous gel membrane and gel particles. PDMS absorbs only
0.02 to 0.2% water [Fedorsl980], the presence of PDMS in composite membranes,
therefore, ~ i ~ c a n t l y reduced the swelling ratio and hydration exhibited by composite
membranes relative to the homogeneous PMAA gel membranes.
From Figures 5.2 and 5.3 it was also obvious that the dispersed PMAA gel particles
in composite films hydrated to the same extent as the homogeneous PMAA gels. This
confinned the fïrst hypothesis of this work that stated that polyelectrolyte gel particles
would remain responsive to extemal stimuli when dispersed within a hydrophobie elastic
network. The PDMS network developed in this work had a high M, that reduced the
retractive pressure it exerted on the PMAA gel particles, dlowing them to hydrate to the t
same extent as the PMAA gel films.
5.3.2 Permeation through Pre-equiiibrated Membranes
The pemeability of caffeine and vitamin B 12 (VB 12) through homogeneous PMAA
gel membranes (0.25% TEGDMA) and composite membranes (17%, 22%, 28% and 33%
dry gel particle loading) as a hction of pH are shown in Figures 5.4 and 5.5. As expected,
the higher molecular weight VI312 (1 200 Da) had a lower membrane pemeability than
caffeine (192 Da) and s o u e pemeability through composites was lower than through the
PMAA gels.
This work, however, was specificdy concemed with the change in permeability as a
fiinction of pH or the ONIOFF drug permeability ratio which was defhed as the ratio of
permeability at pH 7(ON) and pH 3(OFF). It should be noted that the ratio of flues at pH 7
and pH 3 is different than the ratio of pemeability shce the membrane thickness is not the
same at pH 7 and 3. Although the nwi ratio was more relevant practically, we chose to
define ONIOFF ratio based on penneability in order to examine the mechanism that controls
transport in these membranes.
As pH increased fiom 3 to 7, the permeability of c&eine through the PMAA gel
membranes increased fiom 2.4 x 10 cm2/s to 4.6 x lo4 cm2/s and the ON/OFF dmg
permeability ratio was 1.9. The ONIOFF drug deIivery ratio for VB 12 through PMAA gel
membranes was 7. This trend was in agreement with Yasuda's predictions that the rate of
change in permeability as a fiinction of hydration increases with increasing solute size
~asuda1969, Yasudal9711.
Grodzinsky et al [Grimshawl99O,Gr&inskyl990] have also examined the
permeability of solutes through PMAA gel membranes as a hct ion of pH. They found, as
we did, that the swelling ratio (called hydration in their papers) increased significantly in the
region of the gel's pK, ranging fkom a value of 4 at pH 3 to 19 at pH 6. For a permeant of
300 Da, the ON/OFF flux ratio through the membrane was found to be less than 2, similar to
our results with cafZeine. As the molecular weight of the penneant increased to 10,000 Da
the ONIOFF flux ratio increased to 26, again in agreement with our observation that the
ON/OFF ratio increased with increasing penneant size.
Composite membranes with 17% and 22% PMAA gel particle loading showed little
change in permeability in response to pH changes, even though the hydration of the gel
particles in these membranes increased with increasing pH. This was attributed to the low
gel volume M o n of these membranes resulting in little or no increase in gel particfe
connectivity despite increases in gel particle hydration. Composite membranes with 28%
and 33% PMAA gel particle loading showed a 10-fold and 40-fold increase in permeability
for caEeine and VI312 respectively, when the pH was changed h m 3 to 7. These O W F F
perrneability ratios of 10 and 40 represented an increase by a factor of approxirnately 5 for
both cafEeine and VB 12 when compared to perrneation through corresponding PMAA gel
membranes.
Lopour et al. Fopour19953 also demonstrated the dependence of permeation on the
dry gel content of composites prepared with HEMA-MAA copolymers dispersed in silicone
rubber. They found that permeability increased as the hydrogel content increased, with
large changes in permeability o c c ~ g in composites containing between 22% and 35% gel
particles. They atûibuted this observation to the formation of mutuai contacts between gel
particles. They demonstrated that composite membranes contahing kss than 17% dry gel
phase (30% water content) did not conduct electricity, whereas composites containing larger
amounts of gel particles had a marked increase in conductivity. These results indicated a
percolation threshold for their membranes corresponding to 17% dry gel particle loading.
They were concerned with the development of composite systems with percolating hctions
well above their percolation threshold, which behaved like homogeneous hydrogels in tems
of their transport properties, but mechanically were much stronger. No attempt was made to
manipulate the percolating h t i o n of their materials via extemal stimuli to develop variable
permeability membranes.
A semi-logarithmic plot of permeability versus the inverse of PMAA gel membrane
hydration was found to be linear for both cafFeine and VI312 (Figure 5.6), showing that the
mechanism of solute permeation tbrough these gel membranes was consistent with the kee
volume mechanism described by Yasuda [Yasudal969,Yasudal971]. No sieving effect was
observed for either solute.
A semi-log plot of P versus 1/H,l for permeation of caffeine and VB12 through
composite membranes (Figures 5.7 and 5.8) showed that permeability varied with the
hydration of the gel particles in the composite membrane. As discussed previously, it was
expected that gel hydration and percolation wouId act synergisticaily to control membrane
pemeability, giving rise to an enhanced response in permeability with respect to changes in
hydration. The larger slopes observed in Figures 5.7 and 5.8 for the composite membranes
relative to the P W gel results in Figure 5.6 were consistent with our hypothesis.
The caffeine-composite membrane permeation data (Figure 5.7) showed two distinct
clusters corresponding to composite membranes prepared wîth lower (1 7% and 22%) and
higher (28% and 33%) gel particle loading. The slope corresponding to the lower gel
particle loading (-2.42 +/- 0.62,95% CL.) was statistically equal to the slope of the
caffeinePMAA gel permeation data shown in figure 5 (-1.97 +/-0.76,95% C.L.). This was
interpreted to signirjl that hydration was the dominant mechanism of permeability response
in the lower loading region. In this region, a limited number of suface-connected pathways
existed due to the finite nature of the composite membrane. In addition, caffeine had a finite,
albeit low pemeability through PDMS (1.6 x 10- cm2/s) so that isolated PMAA gel
particles and PDMS phases existing in series also formed permeation pathways. Although
inc~ases in hydration increased the permeability of these two pathways, isolated gel
particles were sufficiently fa. apart that increases in hydration did not result in the
recruibnnt of additional gel particles into the membrane-spanning clusters. Thus
connectivity of the gel particles (percolation) did not play a role in the permeability response
of these membranes.
Zn Figure 5.7, data for caffeine permeation through the higher particle loading
composite membranes showed a larger slope (-3.37 +/- 0.53,95% CL.) than the
cafI'einePMAA gel membrane permeation results (- 1.97 +/-0.76,95% C.L.). These results
were consistent with the hypothesis that comectivity and hydration act synergistically to
give an enhanced response of permeability to pH or hydration changes. At these higher
loadings, the gel particles were sufficiently close together that the swelling of particles
which were initially isolated fiom each other led to the formation of new connections,
thereby increasing the number of gel particles within connected, membrane-spanning
permeable clusters.
Similar trends were observed for VB12 permeation through composite membranes
(Figure 5.8). The cluster of data points giving a zero slope corresponded to membranes 4th
low gel particle loading (17% and 22%) and pH 3 gel particle hydration. For these
membranes no VB 12 permeation was detected. Based on the detection limits of the
experirnents, an upper bound on pemeability of 1.6 x 1 u13 cm2 /s was calculated for these
conditions; the data points shown represent th is upper bound. The second cluster of data
gave a slope (-13.81 +/- 2.03,95% CL.) approximattely four times larger than VB12
permeation through PMAA gel membranes (-3.3 7 +/-0.53,95% C.L., figure 5), indicating
the synergistic pemreability enhancement of hydration and percolation effects. The more
pronounced enhancement in permeability response for VB12 relative to cafEeine may be
attributed to the higher rnolecular weight of VB 12.
The data points in Figure 5.8 represent varying combinations of PMAA loading and
pH; this gives rise to a large number of HN, mmposik values. As seen in the figure, at some
values of HSl, ~ ~ , i < r , very different pmeabilites were observed. The same value of Hsl
i,, rnpik only indicates the sarne local diffusion coefficient through the PMAA gel phase
within ihe composite; whether those gel particles are connected greatly influence the overall
permeability through the composite membrane. At the point of incipient percolation, slight
differences in hydration may yield very large differences in permeability. The fact that
composite membrane data in this figure c m be grouped into two clusters - one
representative of "percolating membranes" and the second representing ''uripercolated"
membranes supports this point.
When the sarne data were plotted as permeability versus the hydrated gel volume
fiaction, 4gel in hm, (Figure 5.9), it was evident that the shapes of the curves are characteristic
of percolation behaviour. It should be noted that these plots represent both percolation and
hydration effects, and since percolation curves at constant hydration do not exist, no
cornparison can be made to demonstrate enhanced permeation response due to synergy
using this type of plot. From this plot it is evident that the critical percolation threshold for
VB 12 in PMAA-PDMS composites occurs when the hydrated gel volume hction equals
0.38.
53.3 Dynamic Permeation Studies
Dynamic permeation studies were unsuccessfidly carried out for PMAA gel
membrane due to breakage of the gel membranes during experiments. As the pH of the
solution surrounding the clamped gel membrane was chged , stresses within the gel due to
swelling/deswelling caused the gel to break apart. The weak mechanical properties of the
gel membranes, regardless of crosslinking density used, prevented these gels £tom being
examined as variable permeability membranes. Furthemore, although the PMAA gels
prepared in these saidies experienced large changes in swelling ratio, the hydration changes
as a fiinction of pH produced ON/OFF permeability ratios not larger than 4 for caffeine or
22 for vitamin Bl2. Thus it was concluded that the use of hydration change as a mechanism
for penneability modulation was sub-optimal for the PMAA gel membranes and solutes
examined in this study.
Having demonstrated that at 28% dry gel particle loading, composite membrane
permeability could be modulated via the pH-dependence of the percolating ftaction,
permeability studies were carried out to characterize the responsiveness and reversibility of
composite membranes in a dynamic situation.
5.3.3.1 Caffeine
For dynamic permeability studies of cafTeine through composite membranes
containing 28% PMAA gel particles, extemal pH conditions were changed every 24 h or 48
h in the sequence 3 + 5 + 3 + 7+ 3+ 9 and back to 3. Figure 5.10 shows that
permeability changes arising fiom these pH triggers were reversible, and observable
changes in flux usually occurred within two hours after a new pH was imposed. As the pH
increased h m 3 to 7, composite membrane permeability increased due to an increase in
Hel m mmposac and consequently i,, çanpasiw It is plausible that as the gel particles increased
in volume the surroundiug PDMS network stretched and the gel particles were able to
connect either through the mesh of the PDMS network which had an M, of 166,000 Da or
due to small ruptures in the PDMS network which were reversible. As the pH decreased
back down to 3, the gel particles decreased in volume, becoming isolated and reducing the
-ber of conducting gel clusters. The PDMS retracted back to its original conformation,
the small ruptures were sealed and the PDMS network surrounded the gel particles
providing a relatively impermeable barrier.
An average permeability during each constant pH time interval was calculated based
on the cumulative mass of solute released into the receptor cornpartment during that tirne
fiame. Large changes in permeability were seen when pH was cycled between 3 and 7, and
between 3 and 9, while only small changes were seen between pH 3 and pH 5.
The dynamic ON/OFF permeability ratio (pH 7/pH 3) was found to be 4.5 for
caffeine with a pemeability of 2.5 x IO* cm2/s at pH 7 and 5.6 x 1 O-' cm2/s at pH 3. This
ON/OFF ratio was lower than the value of 1 1.5 found in pre-equilibrated permeability
experiments, most likely because the composite membranes did not reach equilibrium
swelling in 24 h and, therefore had a lower 4sl. ,- than the pre-equilibrated
membranes. The ON permeability, however, was larger in the dynamic pemeation study
than that obtained in the pre-equilibrated membrane study by a factor of 2.
Figure 5.1 1 shows that VI3 12 permeation through composite membranes was also
responsive to dynamic changes in pH. The results indicated reversible and rapid changes in
permeability upon changes in pH, in agreement with earlier observations for caeine. VB 12
permeability h u g h the composite membrane was found to be 1.7 x 1 w8 cm2/s at pH 3 and
2.7 x 104 crn2/s at pH 7, resulthg in an ON/OFF ratio of 160. The OFF (pH 3) permeation
value was comparable to that obtained in pre-equilibrated membrane pemeation studies.
However the ON (pH 7) permeah value was approximately 5 times larger than that
obtained in the pre-equilibrated membrane studies and approxirnately 1.7 times larger than
the permeability of VB 12 through homogeneous PMAA gel membranes (1.6 x 1 O& cm2/s).
Higher permeation values in the ON state during dynamic studies also occurred
when caffeine was used as the penneant. These unexpected results may be explained as
follows. It has been obsewed through the use of the laser scanning confocal microscope that
surface resident gel particles swell and shrink in response to extemal pH changes within a
matter of minutes, even though composite membranes of approximately 1 mm thickness
require h o m to exhibit significant swelling changes and up to two weeks to reach swelling
equilibrium. Considering the situation when the composite membrane had been exposed to
pH 7 for some the , and a steady state linear concentration profile of VI3 12 existed across
the membrane (Figure 5.12 (A)); changing the pH fkom 7 to 3, caused the surface gel
particles to shrink, and closed off permeation channels to VI3 12. This resulted in a rapid
decrease in VB12 flux. In the pH 3 state, VB12 molecules trapped within the composite
membrane had equilibrated throughout the thickness of the membrane (Figure 5.12 (B)), so
that when the pH was changed back to 7 again (]Figure 5.12 (C)), a sharp concentration
gradient of VI31 2 fomed at the membrane/release medium interface, giving rise to a hi&
release flux. Thus the unexpectedly high obsewed permeability at pH 7 may be the result of
a concentration gradient which was higher than that which would exist under steady state
conditions.
Further work is needed to c o d m this hypothesis. However, similar explanations
have been offered in the literature for similar behavior observed in other systems Yoshida
rYoshida19931 estabfished ON-OFF regdation of h g permeation and release using
thermoresponsive copolymer gels of N-isopropylacrylamide and allq4 methacrylates @oly
(NiPAAm-CO-RMA) where the gel surface acted as an 'ON-OFF' switch. The dense skin
layer formed on the surface immediately after increasing the temperature stopped dmg
permeation through the gel. They also found that permeabilify through the membrane
increased in the second cycle, suggesting that the drug concentration profile in the
membrane changed during the "OFF" penod, similar to our hypothesis for the behaviour of
concentration profiles within composite membranes.
5.4 CONCLUSIONS
A novel concept for responsive permeation based on both hydration and percolation
mechanisms in composite membranes has been demonstrated. It was found that PMAA
gel particles dispersed in PDMS networks swelled and deswelled in response to external
pH changes in a marner similar to hornogeneous PMAA gels. At high enough Ioadings
of PMAA gel particles in PDMS, pH-induced swelling resulted in the formation of
comected, penneable pathways, and resulted in ON/OFF permeability ratios through pre-
equilibrated composite membranes that were higher than the ON/OFF ratio achievable
with gel hydration alone. Lastly, unusually high permeabiiity at pH 7 occurred for
vitamin Blz and caffeine in the dynamic state. It was hypothesized that the rapid
deswelling of surface gel particles at pH 3 trapped solute within the membrane,
concentration equilibration across the membrane resulted, giving rise to high release
fluxes at pH 7 when the membranes gel pathways were "re-openedy'.
Figure 5.2 Swelling ratio vs. pH for PMAA gels (0.25% TEGDMA) and composite membranes (28% dry gel ioading). Error bars indicate standard deviations (n=3).
PMAA
I HPG-28% PMAA
A PMAAin HPG
Figure 5.3 Hydration vs. pH for PMAA gels (0.24% TEGDMA) and composite membranes (28% dry PMAA gel loadings). Error bars indicate standard deviations (n=3).
--
Membrane Type * - HPG-17% PMAA
A HPG-22% PMAA -6- HPG-28% PMAA --+- - HPG-33% PMAA 4 100% PMAA
Figure 5.4 Permeability of caffeine through PMAA gels (0.25% TEGDMA) and composite membranes (17%, 22%, 28%, 33% dry PMAA gel loading) as a function of pH. Emor bars indicate standard deviations (n=3).
Membrane Type
* - HPG-17% PMAA
A HPG22% PMAA -Tm HPG-28% PMAA
-6- HPG-33% PMAA 100% PMAA
Figure 5.5 Penneability of vitamin BIZ through PMAA gels (0.25% TEGDMA) and composite membranes (1 7%, 22%, 28%, 33% dry PMAA gel loading) as a function of pH. Error bars indicate standard deviations ( n = 3 ).
S l o ~ e with 95% Confidence Interval
Figure 5.6 Semi-log plot of permeability of caffeine and vitamin Btz through PMAA gel membranes vs. the inverse of hydration of gel membrane.
Dry PMAA Gel loadin
17%
22%
A 28%
v 33%
Slope with 95% Confidence Interval
U -2.42 +/- 0.62
-Ir-f- -3.37 +/- 0.53
Figure 5.7 Semi-log plot of permeability of caffeine through composite membranes vs. the inverse of gel hydration in membrane.
Dry PMAA Gel Loadiig
Slo~e with 95% Confidence Interval
-13.81 +/- 2.03
Figure 5.8 Semi-log plot of permeability of vitamin Bit through composite membranes vs. inverse of hydration of gel in HPG: ( ) detectable permeation indicated the existence of percolating clusters; ( ------- ) non-detectable penneation indicated no percolating clusters.
a
Caffeine 15 - a
0 a
I O - a
a * l
5 - e -
a m & :*a
0 - 1 I I
15 -
10 -
I I
I 5 - 5
Vitamin B,,
Figure 5.9 Penneability of caffeine and vitamin B12 VS. the volume fraction of hydrated PMAA gel in composite membranes.
I I I I 1 I
O 50 IO0 150 200 250
Time (h)
Figure 5.10 Dynamic permeation profile for caffeine through composite membranes with 28% dry PMAA gel Ioading.
100 150 200
Time (h)
Figure 5.1 1 Dynamic pemeation profile for vitamin through composite membrane with 28% dry PMAA gel Ioading.
OFF
Figure 5.12 Composite membrane concentration profiles in a dynamic permeation study : (A) membrane in pH 7 solution, gel particles swollen and connected, steady state concentration profile; (B) membrane switched to pH 3 solution, surface gel particles deswell, and disconnect, concentration equilibration between interior, swollen connected particles; (C) membrane switched to pH 7 solution, surface particles swell and connect, sharp concentration gradient at membranelrelease medium interface.
CHAPTER 6 : PMAA-PDMS 1 PN USING THE MONOMER
XMIIMERSION lMETHOD
The morphology at and adjacent to the IPN surface affects the mass transfer
properties of an IPN, and is therefore an important aspect for its use in dnig delivery
applications. In this chapter we examined the effect of the IPN-substrate interface and IPN
preparation method on the surface morphology of hydrogel-elastomer IPNs. Specificaily,
IPNs of polydimethylsiloxane and polymethacrylic acid were prepared using glass, air or
pure methacryiic acid monomer as the contacting surface during IPN formation. The
morphology at the surface region (100 pm fiom the surface) of the resulting IPN was
characterized using laser scanning confocal microscopy (LSCM), and the existence of a
bicontinuous morphology throughout the membrane bulk was M e r corroborated via
permeation studies.
Conventional methods of preparing IPNs using glas or air as the contacting surface
did not produce bicontinuous morphologies at the surface region and were not permeable to
water-soluble solutes. The use of pure methacrylic acid monomer as the contacting medium
during PN preparation - the "monomer immersion method" - is a novel method of IPN
synthesis developed in this work to produce permeable IPN membranes with bicontinuous
morphologies due to the homogeneous distribution of monorner in the pre-IPN film.
6.1 Introduction
Gradient morphology formation and surfàce enrichment have been examined to a
greater extent for blends [Chen 1999, Zhang19981 and copolymers [Senshu1997] than IPNs.
For these systems, it has long been known that the surface composition cm be different
h m the bulk and that surface segregation is significantly affected by the interfacial energy
at the polyrner-substrate interface. The polymer component that minimizes the interfacial
energy will segregate to the surface. The fonnation of a gradient morphology in the surface
region is a direct consequence of the sudiace segregation of the one polymer.
Conventional methock of hydrogel-elastomer IPN preparation has involved placing
the IPN against a surface of either Teflon wurayamal 993, McGareyl989,
Murayamal 993al Mylar Pae19903 or glass Figenit01 998, He 19951 without explicitly
controlling monomer disûibution in the pre-IPN films. Surface segregation of the rubber
component was fond in some studies. He et al. w 9 9 5 ] placed a polysiloxane/acrylic
acid monomer mixture onto a glass sheet to fom an IPN. They found that the surface
consisted of a 5 nm layer of plysiloxane and that the swface wettability was very Iow.
Murayama et al [Murayamal 9931 prepared IPNs fiom poly(2-hydroxyethyl methacrylate)
OpHEMA) and polystyrene (PS) using the sequentiaf method by placing the pre-IPN films
ont0 Teflon. XPS studies found that the PS component was enriched in the swfàce of the
IPN and that a gradient composition of the two polymers existed in the fmt 100 p of the
surface region. The gradient composition at the surface was attributed to the inhibition of
styrene polymerization due to diffbsing into the IPN, as well as the effects of surface
thermodynamics at the IPN-substrate interface during polymerization.
Lipatov and Semenovich [Lipatov 19993 found that the structure and composition of
the surface layers formed during IPN fonnation near the interface with solid were dependent
on the surface energy of the solid. Using attenuated total reflectance infked spectroscopy,
the composition of the system near the interface was monitored. When a pre-IPN
composition was applied to the surface, an excess interfacial energy arose between the
substrate and polymer mixture as a result of the diEmence in surface tension between the
substrate and the pre-IPN. Those components that compensated this dflerence segregated
to the surface.
In this chapter PMAA-PDMS IPNs were prepared using the sequential method
where the monomer-swollen pre-IPN film was pIaced ( i )against a glas substrate (glass-
IPN ), ( ii )interfaced with air (or nitrogen) (air-IPN) or ( iii )immersed in the pure monomer
of the IPNs guest polymer (monomer-IPN), during guest monomer polymerization and cure.
At pH 7, al1 three materials had an equilib- hydration value of approximately 0.84.
Despite the high hydration, no permeation of vitamin B12 h u g h the air-IPN and glass-IPN
membranes could be detected over a one month permeation experiment. This corresponded
to a value of perrneability of less than 1.9 x 1 cm2/s. In contrast, the permeability of
VB12 through the monomer-IPN membrane was found to be 6.5 x lod crn2/s at pH 7. The
morphology of these materials was M e r examined ushg LSCM in order to understand
why some materials of significant hydration were not penneable to water soluble
compounds and to determine the effect of substrate and preparation methods on the
bicontinuous morphology of the IPN - particularly in the surface and sub-surface regions.
The hypothesis was that the IPN-substrate interface had a signifîcant effect on the guest
monomer concentration profile at the surface and sub-surface regions, which in turn affected
the morphology in each IPN system.
6.2 Methods
IPNs were prepared using a glas substrate, against air (or nitrogen) and immersed in
a MAA monomer solution as described in section 4.2.4.
LSCM of IPNs pre-immersed in a concentrated fluorescein (332 Da) solution was
canied out ushg the procedure outlined in section 4.3.4. These images were subsequently
andyzed in terms of percent fluorescent area as a function of depth (section 4.3.4).
6.3 Results and Discussion
6.3.1 Glass-IPN Interface.
Figures 6.1 (a) - (f) are optical LSCM sections of an IPN prepared against a glas
surface, a substrate known to thermodynamically favour contact with hydrophilic polymers
and ionic monomers. The IPN was pre-immersed in a concentrated fluorescein sodium salt
solution in order that fluorescein could permeate through the hydrophilic, connected
domains and channels of the IPN which appear red, and the impermeable PDMS ma& or
isolated hydrogel domains which appear black.
The image of the PMAA-PDMS IPN surface (Figure 6.1 (a)) shows large (0.5 - 2 p
m), irregular PMAA hydrogel domains. Images taken between 1 p and 12 pn below the
surface were black (Figure 6.1 (b)) implying that either no hydrogel domains were present,
or that hydrogel domains which may have been present were encapsulated in PDMS and
inaccessible to the fluorescent probe. The latter scenario is typical of a dispersed sea-island
morphology fomed by the nucleation and growth (NG) phase separation mechanism, and is
indicative of low monomer concentrations in the pre-IPN film.
Kim et al. park1997, Kim19991 were able to visualize such a morphology using
SEM because their guest polymer domains were sigiiGr;àiiiiy iargei' in the dry state (1 pm
diameter) than those in this study. Based on the results of Kim and images taken of layers
above and below the 14 pm fluorescent-fiee area, a plausible argument cm be made that
dispersed hydrogel domains existed in this area due to the low monomer concentration
created during the migration of monomer to the substrate surface. A dispersed, discrete
hydrogel domain morphology would be impermeable to water-soluble compounds such as
fluorescein and vitamin B12.
At a depth of 14 pm, spherical connected hydrogel do& approximately 3-5 pm
in diameter were visible (Figure 6.1 (c)). From 14 to 18 pm in depth (Figure 6.1 (d)) the
spherical hydrogel domains increased in concentration. At 24 p (Figure 6.1 (e)), spherical
domains were still visible, however, the presence of small nodula. domains within the
spheres became evident. This region of dual phase morphology, where large dispersed
spheres have begun to cmect to form a bicontinuous, penneable structure due to the phase
separation taking place by both spinodal decomposition (SD) and NG, is clearly evident in
the LSCM images. This mgion represents a transition in mechanisms of phase separation,
and is consistent with a higher monomer concentration than that present at the surface layer.
At 30 pm (Figure 6.1 (0) only a small, nodular gel domain morphology was present, and
this morphoiogy contùlued to a depth of about 60 pm. Monomer concentration had
increased at this depth leading to SD mechanism of phase separation.
It should be noted that other than the first five microns near the surface, the
morphology spectrum just described was seen in al1 glass-IPN films exarnined. The
morphology at the surface varied with lateral position, and ranged fkom isoiated hydrogel
domains to unifom hydrogel layers up to 5 p in thickness. The enrichment of hydrogel at
the s d a c e is consistent with the hydrophilic nature of the glas surface. The non-
uniformity of the surface layer can be attributed to uneven contact between the pre-IPN f h
and glass and evaporation of the MAA monomer prior to placing the pre-IPN film q a h s t
the glass substrate.
Figure 6.4 (a) quantitatively shows the percent fluorescent area versus depth for the
glass-IPN based on the series of images of Figure 6.1. The most striking feature of Figures
6.1 and 6.4 (a) was the 14 pm fluorescent-fiee layer adjacent to the surface of the glass-IPN
that was attnauted to a region of discrete, impermeable hydrogel domains. The pattern of
morphology s h o w in these images was sirnilar to the morpholo~ spectnun described by
Kim and was attributed to the monomer concentration gradient created due to the migration
of monomer to the glas sirbstrate surface.
6.3.2 Air-IPN Interface
Figures 6.2 (a-f) are optical sections of an IPN prepared with an air4PN interface
(air-IPN). At the surface of the air-IPN (Figure 6.2 (a)), there is a PDMS-e~ched layer
with some nodular hydrogel domains (0.7 % fluorescent area, Figure 6.4 @)). Similar
nodular hydrogel domains were also visible at 5 pn below the surface (Figure 6.2 (b)), but
at a higher concentration (98% fluorescent area, Figure 6.4 (b)). From 10 pn to 30 pm
(Figure 6.2 (c)) the LSCM images were black, indicating the absence of accessible PMAA
hydmgel regions. The existence of a black sub-surface layer can be attributed to the
presence of discrete hydrogel domains that were impermeable to the fluorescent probe. The
above tomographie results were reproducible at three different spots for each of three
different Ah-IPN samples anaiyzed.
It was expected that at the air-IPN interface a surfaçe layer of PDMS would exist
due its low surface energy, as well as due to the volatilization of the monomer and
crosslinker, therefore, the large increase in hydrogel domains at 5 pn below the surface was
unexpected. The enclosed nitmgen enviromnent in which the polymerization and
crosslinking of the IPN took place was most likely saturated with monomer vapour fiom the
pre-IPN film and this vapour may have polymerized at the polymer swrface regions affecting
the final hydrogel composition of the surface layer. A mass balance calculation based on the
vapour pressure of MAA at room temperature and the composition of gel at the air-IPN
surface supports this explanation and can be fond in Appendix C.
Proceeding inward fiom the black layer, the morphology at 40 pn (Figure 6.2 (d))
changed to large spheres of PDMS-rich domains surrounded by PMAA-rich boundaries to
give a honeycomb-like appearance. A similar morphology was seen at 50 p (Figure 6.2-
(e)); however, the P M - r i c h boundaries appeared to be more diffuse, and within each
honeycomb, small, nodular PMAA-rich domains were visible. This honeycomb appearance
for the hydrogel component was different than the large hydrogel spheres observed in the
region of dual phase morphology for the giass-IPN membrane. In this membrane, it was the
PDMS-rich regions that have formed spheres. This structure may possibly be due to a phase
inversion that has taken place during IPN formation. At 60 pn (Figure 6.2 (f) no evidence
of spheres or honeycomb-like structures could be seen; the separate phases had a nodular
shape typical of SD. This morphology continued up to 100 p below the surface, at which
point LSCM images were no longer discernible. The gIass and air substrates affected the
morphology of the IIPN to depths of 30 pn and 60 pn fkom the surface, respectively. The
substrate affected a fairly large region of the surface because of the phase separation process
which took place simultaneously with the surface segregation occurring at the substrate-IPN
interface.
6.3.3 Monomer-IPN Interface.
Figures 6.3 (a) - (f) show the LSCM images of a rnonomer-IPN starhg at the
surfàce of the IPN, and progressing into the bulk of the IPN to a depth of 50 p. There was
a slightly lower concentration of PMAA hydrogel regions at the surface of the IPN (Figures
6.3 (a) and 6.4 (c). This was attributed to the diffusional release of crosslinking agent and
initiator h m the pre-IPN film when it was placed in pure MAA monomer prior to UV
irradiation. The difisional release would slow considerably as polymerization and
crossilliksig progressed, thus isolating MAA depletion to oniy the near-surface region.
The qualitative morphological features of the PMAA-PDMS IPN prepared by the
monomer immersion method were similar throughout the IPN. The IPN consisted of
uniforiniy dispersed hydrogel domains (approximately 100-200 nm in dimension) within the
host PDMS network. The morphology was characteristic of bicontinuous structures fomed
by SD. Figure 6.4 ( c ) illustrates the uniform concentration of accessible hydrogel domains
as a hc t ion of gel depth produced by the monomer-mimersion method.
The monomer-immersion rnethod of IPN preparation was developed to overcome
the problem of non-uniform morphologies created due to the formation of a monomer
concentration gradient were used. It was hypothesized that the gradient morphology
spectnun formed in membranes prepared using conventional modes (formation of the IPN
against a substrate (glass-IPN) or as a ikee surface (air-IPN)) was due to a poorly controlled
monomer concentration gradient in the pre-IPN ~. Monomer concentration profiles were
most h l y a resdt of surface segregation at the IPN-substrate interface. By surrounding the
pre-IPN film with guest monomer duririg IPN formation, a uniform distribution of MAA in
the pre-IPN film was maintained while polymerization and crosslinking of MAA took place
resulting in a uniform, bicontinuous morphology throughout the membrane thickness.
In this work, the monomer immersion method was developed for an IPN system in
which the guest monorner could di- into the host polymer without requiring a co-
solvent. In less compatible systems that require a CO-solvent, the same method can be
applied simply by carrying out guest monomer pc-ilymerization and crosslinking while
immersing the pre-IPN in a solution of the monomer and CO-solvent.
6.4 Conclusions
The effect that different substrates have on the surface morphology of the rubber-
hydrogel IPN material was examined. IPNs prepared against a glass substrate or as a £ke
surface produced a morphology spectnun ranging h m dispersed hydrogel domains near the
surface to a dual phase morphology, followed by a bicontinuous morphology indicative of
spinodal decomposition. This morphology spectnui3 created an Ilripermeable layer near the
surface of the membrane, rendering even highly, hydrated membranes inpermeable to
water-soluble compounds. The formation of the impermeable fayer was attributed to the
monomer concentration gradient formed due to the re-distribution of monomer at the IPN-
substrate interface. The monomer concentration gradient was dependent on the particulzr
substrate used during synthesis.
The monomer-immersion method of IPN preparation produced bicontinuous
hydrogel-elastomer IPN membranes with phase morphology indicative of spinodal
decomposition. This allowed for high hydration values and permeability to water-soluble
solutes at pH 7 conditions. The monomer-immersion method minirnized the monomer
concentration gradient in the pre-IPN fïim caused by surface thennodynamics and produced
a uniform, bicontinuous morphology throughout the membrane thickness.
Figure 6.1 LSCM images of depth-profile of IPN prepared with a glass substrate (glass- IPN): ( a ) surface; ( b ) 2 Pm; ( c ) 14 Pm; ( d ) 18 prn; ( e ) 24 pm; ( f ) 30 pm.
Figure 6.2 LSCM images of depth profile of IPN prepared with a free surface (air-IPN):
( a ) surface, (b) 5 p; ( c ) 10 pn; ( d ) 40 pm; ( e ) 50 p; ( f ) 60 W.
Figure 6.3 LSCM images of depth-profile of IPN prepared using monomer immersion method: (a)surfàce;(b) 10pm;(c)20p;(d)30pm;(e)40p;(f)50pm.
O 10 20 30 4 0 50 60 70
Depth from IPN Surface (pm)
Figure 6.4 Percent Fluorescent Area as a function of Depth fiom IPN Surface
for LSCM images of glas-IPN ( a ), air-IPN ( b ), and monomer-IPN ( c ).
CHAPTER 7 : MORPHOLOGY OF PlMAA-PDMS IPN
MEMBRANES
In chapter 7, direct visualization of IPN morphology fonned during polymerization
induced phase separation (PIPS) of PMAA-PDMS IPNs is reported. Using laser scanning
confocal microscopy (LSCM) and hydrophilic fluorescent probes of varying molecular
weights, the hydrophilic domains in PMAA-PDMS IPNs were observed. The results reveal
cornplex, superimposed structures of hydrophilic domains of varying sizes and spatial
distributions. This morphology is athibuted to the phase-separated structures formed and
partially arrested at each successive quench depth during the PIPS process. These
observations have not been reported previously and represent a new understanding of
morphology development in IPNs.
7.1 Introduction
IPN morphology is largely determined by phase separation of the polymer
components during IPN formation. For sequential IPNs, phase separation occurs as a result
of the increase in molecular weight during polymerization and network formation of guest
monomers within the host polyrner network. This process is termed polymerization induced
phase separation (PIPS).
The morphology fonned during the PIPS process is highly dependent u p n the
relative rates of the polymerization and crosslinking reactions versus the rate of phase
separation. At t = O, the monorner and polymer of the pre-IPN are miscible and form a
single phase. Once polymerization begins, the hcreasing molecular weight of the
polymerizing component(s) reduces the entropy of mixing of thc polyrner system,
98
CHAPTER 7 : MORPHOLOGY OF PMAA-PDMS IPN
MEMBRANES
In chapter 7, direct visualization of IPN morphology formed during polymerization
induced phase separation (PPS) of PMAA-PDMS IPNs is mported. Using laser scanning
confocal microscopy (LSCM) and hydrophilic fluorescent probes of varying molecular
weights, the hydrophilic domains in PMAA-PDMS IPNs were observed. The results reveal
complex, superimposed structures of hydrophilic domains of varying sizes and spatial
distributions. This morphology is attributed to the phase-separated structures formed and
partially arrested at each successive quench depth during the PIPS process. These
observations have not been reported previously and represent a new understanding of
morphology development in PNs.
7.1 Introduction
IPN morphology is largely determined by phase separation of the polymer
components during IPN formation. For sequential IPNs, phase separation occurs as a result
of the increase in molecular weight during polymerization and network formation of guest
monomers within the host polymer network. This process is temed polymerization induced
phase separation (PIPS).
The morphology formed during the PLPS process is highly dependent upon the
relative rates of the polymerization and crosslinking reactions versus the rate of phase
separation. At t = 0, the monomer and polymer of the pre-IPN are miscible and form a
single phase. Once polymerization begins, the increasing molecular weight of the
polymerizing component(s) reduces the entropy of mWng of the polymer system,
eventually reaching a point where the Gibbs fkee energy of mixing becomes positive and
phase separation occurs. The accompanying vitrification and increase in viscosity caused by
the polyrnerization and crosslinking reactions kinetically hinder and eventually arrest phase
separation. These two opposing interactions dictate the thermodynamic quench depths of
the PIPS process, as weîl as the kinetic mechanism of phase separation.
The transition fiom one ara of the phase diagram to another is called a quench.
Quench depth is quantitatively defined as the temperature ciifference between the lower or
upper critical solution ternperature (LCST or UCST) of the polymer system and the reaction
temperature. Figure 7.1 illustrates the defrnition of quench depth for a LCST undergohg
phase separation due to a temperature jump or 'Yherma1 quench".
Quench
Figure 7.1 Phase diagram of a polymer blend depicting the lower critical solution ternperature and quench depth of the system.
During PIPS, as the polymerization and crosslinking reactions increase the
rnolecular weight of the polymerizing component, the LCST decreases shifting the phase
diagram to lower temperatures which resdts in a "chernical quench". Although the reaction
temperature remains constant, the quench depth continuously increases during the PIPS
process due to the continuous change in the LCST of the system caused by the increase in
X X
(a) t = (b) t =
the molecular weight of the polyrners in the system (Figure 7.2).
Figure 7.2 Phase diagram of a polymer blend system undergohg PIPS. The LCST (point A) decreases with time as the molecular weight of the polyrners increase due to polymerization and crosslinking reactions. of the system remains constant.
In Figure 7.2, (a) at t = O the polyrner system forms a single phase at TdOn denoted
by point B. The LCST of the system is higher than TMOn and is denoted by point A. As
polymerization and crosslinking proceed the LCST of the system decreases. At t = tl
(Figure 7.2 (b)), Treaction is equal to the LCST and the polyrner system is located at the
critical point of the binodal line of the phase diagram, where phase separation begins to
occur. In this region the chemical quench is just beginning to occur and the quench depth is
very small or "shallow". At t = t ( Figure 7.2 (c)), continued polymerization and
crosslinking lowers the LCST even M e r so that point B is located within the two phase
regime of the phase diagram. Phase separation has occumed within the system and the
quench depth has increased significantly. At greater quenches the driving force for phase
separation is larger resulting in smaller phase separated domains than those created at
shallow quench depths. nerefore, the size of the phase separated domain is inversely
proportional to the quench depth [vanAartsenl970, Binder1973J. Theoretically, the
continuous change in quench depth during PIPS should result in a variety of gel domain
sizes in the final IPN morphology [Tran- Cong 1 999% Seul 1 995 1.
volUmm fraction Polyma B
Figure 7.3 Phase Diagram for a polymer blend system which illustrates the kinetic mechanisms of phase separation for each region [Kieferl999].
The kinetic mechanism of phase separation is determined by the location of the
polymer system in the binary phase diagram during PIPS (Figure 7.3). Providing
temperature and composition rernain constant, this is dependent upon the extent of
polymerization d o r crosslinking. Since these conditions are continuously changing with
time, the region of the phase diagram describing the reaction system also continuously
changes. The system may first exist in a metastable region of the phase diagram, where
phase separation occurs via a nucleation and growth (N&G) mechanism, but as the phase
diagram changes the condition of the reacting system may become part of an the unstable
region where phase separation occurs via spinodal decomposition.
Spinodal decomposition is the predominant mechanism during the formation of
IPNs ESperlingl994, Lipatov1997J It is a spontaneous process that may be initiated by a
change in temperature (thermally induced phase separation (TIPS)), where the quench depth
does not Vary or by PIPS, where quench depth continuously changes with tirne. SD initiated
by TIPS has been studied extensively and a 4-stage process has been identified.
Figure 7.4 Change in morphology with thne during themally-induced spinodal decomposition [Inouel 9951.
During the early stages of SD, a highly interconnected, two-phase morphology with
a unique periodicity or a domain structure with relatively narrow size distributions may be
obtained (Figure 7.4 (a) and (b)). As phase separation progresses, both the concentration
and domain size of the sepamted domains increase and the interconnected structure yields
ihgmented domains (Figure 7.4 (c)) followed by spherical domains which may grow in size
and coalesce (Figure 7.4 (d)). The domain sizes, even in the late stages of SD are of a
uniform size and dispersed quite regularly Foue19951. Although similar stages of SD were
found to occur in PIPS there has been some indication in the literature [Tm-Congl999a,
Seul 19951 that SD due to PIPS in IPNs does not produce uniform domain sizes, most likely
due to the the-variant nature of quench depth in this process. However, different shed
domains in IFW morphology, created by a continuously changing quench depth, have not
been observed.
Chou et. al. [Chou1994] examined morphology development in a polyurethane and
polystyrene IPN using a phase contrast optical microscope and a transmission electron
microscope (TEM). An interconnected phase developed which coalesced to form a periodic
droplet and matrix type morphology. A second level of phase separation producing srnaller
domains was found to occur within the dropiet and m a t - phases produced by SD.
Widmaier Widmaierl99SJ examined semi-IPNs of crossluiked polyurethane (PU) and
linear polystyrene by light transmission studies, optical microscopy and scanning electron
microscopy (SEM) and found that when the reaction medium phase separated before
gelation of PU, the final morphology was a superposition of two levels of phase separation:
(i) a fine dispersion of the components and (ii) a gross phase separation of polystyrene
noduli surrounded by a PU-rich shell. Although these limited examples allude to a more
coniplex morphology formed in IPNs due to PIPS, they only directly demonstrate the
occurrence of two phase separated domah structures with widely different length scales.
Theoretically, due to the continuously changing quench depth, one may expect a
continuum of domain sizes and macrostructures formed in the IPN. However, either a
constant domain size or evidence of only two different domain sizes has been reported. We
suspect that part of the reason for these limited observations have been that the analfical
methods used to examine IPN morphology were not able to distinguish between the various
levels of phase separation and domain structures present. The andytical techniques that
have been used to investigate IPN morphology include TEM, SEM, phase contrast optical
microscopy and to a lesser extent, small angle neutron scattering and SAXS [Hourston
19981. In this thesis, PN morphology is examined using the laser scanning confocal
microscope (LSCM), a device which uses a laser to visualize flourescent markers within a
material. The unique advantage of the LSCM over conventional light microscopy is that it
incorporates a pin-hole in fiont of a photo-detector to only admit light that is in a single,
tightly defined focal plane. The rejection of out-of focus light results in images with good
contrast and clarity, under atmospheric conditions. The specimen stage is stepped up or
down to collect a series of 2-D irnages (or slices) at each focal plane, thus specimen
preparation is minimal requiring only standard optical microscope techniques.
PMAA-PDMS IPNs prepared using the monomer immersion method (chapter 6)
consist of hydrophilic PMAA permeation pathways within a rubbery, water impermeable
PDMS matrix. Therefore, when hydrophilic fluorescent probes permeated through the
membrane, they only diffised through the intercomected PMAA hydrogel domain
channels. When LSCM was used to image the LPN these interconnected hydrogel domains
which contained the fluorescent probe were visible and distinguishable fi-om the rest of the
IPN. In this chapter we present work in which different sized fluorescent probes were used
to distinguish between the different sized hydrogel domains which may have formed in the
IPN during the PIPS process.
The LSCM images reveaied a complex overlap of hydrogel domains and
macrostructures that had formed and were arrested by crosslinking reactions during the
different quench depths and stages of SD taking place during the PIPS process. Direct
evidence of such IPN morphology development has not been demonstrated previously.
7.2 Methods
The morphology of the PMAA-PDMS IPN prepared according to the monomer
immersion method of section 4.2.4 was examined as a fùnction of depth using the laser
scanning confocal microscope (LSCM) (section 4.3.4 ). Sections of the PMAA-PDMS IPN
membrane were placed in solutions of different molecular weight fluorescent probes;
fluorescein (sodium salt, M.W. 332 Da), and fluorescein isothiocyanate dextran @TTC-
dextran) of M. W. 4,000 Da (FDX4.4KD) and 70,000 Da (FDX70KD). These hydrophilic
markers penneated the interconnected hydrogel domains of the IPN and were used to
distinguish the hydrogel domains fiom the PDMS regions of the IPN.
Each IPN section was imaged at depths of 2 pn intervals from the surface of the
IPN to approximately 100 pm below the surface. Selected images are show in Figure 7.5
(fluorescein marker) Figure 7.6 VDX4.4KD ) and Figure 7.7 (FDX70KD). The IPN was
not irnaged at depths greater than 100 p because the images became too dark and
docussed. In LSCM such images are commonly observed as the depth of the focal plane
increases due to rehction, reflection, and absorption of the excitation beam and emitted
light fiom the region above the focal plane.
7.3 Results
Figure 7.5 contains LSCM images of the PMAA-PDMS IPN that had been pre-
equilibrated in a concentrated solution of fluorescein (hydrodynamic diameter (h.d.) 0.8
nm). Images show at the surface and at selected depths up to 70 pn fiom the surface
consisted of small red cylhdrical, homogeneously dispersed regions. The red regions
represent PMAA hydrogel domains that were permeable or accessible to the fluorescein
solute. The black regions represent the PDMS network or hydrogel domains that the marker
could not permeate through due to size exclusion (ie. the diameter of the hydrogel domain
or the mesh size of the hydrogel was smaller than the diameter of the marker). In these
images the PMAA gel domains appeared uniformly dispersed and of approxirnately the
same small diameter.
Figure 7.6 shows LSCM images of the PDMS-PMAA IPN pre-equilibrated in the
concentrated solution of a much larger probe, FDX 4.4KD (h.d.. 9.4 nm). Images of the fmt
50 pn of the membrane were similar to images of the fluoresceinequilibrated IPN in Figure
7.5. The domains appeared small, of uniform size and homogeneously dispersed. Beyond
50 pm, LSCM images of the IPN cross-section abruptly changed. Large, black regions,
indicating an absence of fluorescent solute, and red regions, which formed a rnacrostructure
containing smaller, red cylindrical gel domains, were observed.
In Figure 7.6 (c) at a depth of 52 pn, a macro-structure of gel domains fonned
which was similar to that of intemediate-stage spinodal decomposition, where the
interconnected cylindrical domains had grown in size and become globular in order to
minimize the interfacial area (see Figure 7.4 (c)). At 54 ptn the macrostructure formed
spherical domains (Figure 7.6 (d)) similar to late stage SD (Figure 7.4 (d)). Figures 7.6 (e)
and (0 show that these spherical domains grew and coalesced into a connected globular
stnickire. It was also evident that the larger spherical domains contained smaller cylindrical
gel domains.
Figures 7.7 (a)-(f) contain LSCM images of the PMAA-PDMS IPN soaked in
FDX70KD (h.d. 37.4 nm). Zn conbast to the smaller molecular weight probes of 4,400 or
332 Da, the non-random distribution of inaccessible gel domains was evident irnmediately
at 5 p m fiom the surface (Figure 7.7 (b), where clusters of red regions and black regions had
formed.
At a depth of 10 pm (Figure 7.7 (c)), a large decrease in the volume fiaction of gel
dornains accessible to FDX70KD occurred. The upper left-hand corner contained an area of
very small, spherical, gel regions sirnilar to morphology created fiom the nucleation and
growth mechanism of phase separation. At 15 pm (Figure 7.7 (d)) larger globules of a
similar diameter containhg srnaller cylindncal gel domains were evident. These globules
increased in size and began to coalesce as the image depth increased to 20 pm (Figure 7.7
(e)) and 25 pn (Figure 7.7 (f)). This morphology was similar to that created in the late
stages of the spinodai decomposition mechanism.
7.4 Discussion
LSCM images of PMAA-PDMS IPNs immersed in a fluorescein solution (Figures
7.5 (a) - (f)) show PMAA hydrogel domains of a uniform size, homogeneously dispersed
throughout the 100 pm surface region. Past studies [Chen1998, Burfordl989,
Donatellil 976, Yeo19831 which have used transmission electron microscopy (TEM)
(Burfordl989, Donatelli 19761 or scanning electron microscopy (SEM) to examine the
morphology of IPN materials have produced simila images. This has led these researchers
to conclude that the process of PIPS in IPNs results in a morpholojg of uniform domain size
sirnilar to the morphology produced in polyrner blends undergohg themally induced phase
separation (TIPS). M e r researchers [Tran-Congl999, Tran-Congl999al have questioned
these conclusions because the process of PIPS and TIPS are very different. In particular, it
is hown that the quench depth continuously increases during the PIPS process and that the
domain size during phase separation is determined by the quench depth of the process.
Theoretically this continuous change in quench depth during PIPS shouid result in a variety
of gel domain sizes in the final nm\J morphology.
7.4.1 PMAA-PDMS LPN Morphology at a Specific Depth
Figure 7.6 shows LSCM images of an IPN pre-equilibrated in FDX4.4KD (h.d. 9.4
nrn). Images of the fmt 50 pn of the membrane were similar to images of the IPN soaked
in fluorescein (hd. 0.8 nm, Figure 7.5). The hydrogel domains appeared small, of uniform
size and homogeneously dispersed.
Beyond 50 pm (J?igure 7.6 (e)-(O), large, black regions, indicating an absence of the
fluorescent probe, and red regions, which formed a macros~cture containing smaller, red
cylindrical gel domains, were observed. Based on images (Figure 7.5 (a)+')) obtained h m
IPNs immersed in fluorescein, it is known that hydrogel domains were uniforrniy present at
this depth. It was concluded, therefore, that hydrogel domains that were accessible to
fluorescein and visible in Figures 7.5 (a)-(f), were not accessible to the larger probe and
were imaged as black due to the absence of the larger fluorescent probe in these areas.
These figures provided direct evidence that domains of different length scales were
produced during polymerization-induced phase separation (PIPS). Most likely, the larger
domains were created at a shallow quench and were arrested during crosslinking and
polymerization reactions which occurred at that quench depth. Subsequently, smaller
domains formed and set at deeper quench depths. The smaller domains were superimposed
upon the larger ones but had their own permeation pathways.
Tran-Cong et al [Tran-Congl9991 indirectly demonstrated the existence of domains
of two different length scales during photo-crosslinking of two components of a temary
polymer blend By imaging the morphology using a phase-contrast optical microscope and
analyzing the images using two dimensional fast Fourier transfom (2D-FFT) they were able
to distinguish between the different sized domains formed in the crosslinked polymer blend.
This morphology was attributed to the inhomogeneous fkeezing kinetics of the crosslinking
process.
Upon M e r examination of Figures 7.6 (dHf), the macrostructure was found to
contain smaller, cylindrical gel domains within the larger phase separated globules,
providing another example of the different levels of phase separation taking place. In this
case "secondary" phase separation by spinodal decomposition occurred at later times
(deeper quench) to produce smaller domains within the macrostructure. These domains
were fiozen within the macrostmcture during the initial stages of spinodal decomposition
before any coarsening was allowed to take place and were, therefore, cylindrical in shape
and highly interconnected. Their small size was attributed to the deep quench caused by the
high molecular weight of the polymer components during phase separation.
Yang et al Wang19981 fmt coined the tem, "secondary spinodal decomposition" in
1998 when semi-IPNs prepared fkom polyphenylether and crosslinked poly (diallyl
phthalate) were visualized using TEM. They observed systems of fine polyDAP domains
on the order of 10 nm present in the larger, micrometer-size dispersed particles. They
concluded that the fine domains were formed by successive spinodal decomposition, under
very deep quench after the micrometer scale particle/matrk morphology was arrested by
partial cure.
Figures 7.7 (a)-(f) contain LSCM images of IPN pre-equilibrated in FDX70KD.
Compared to the smaller molecular weight probes of 4,000 or 322 Da, the non-random
distribution of inaccessible gel domains was evident immediately at 5 pm, where clusters of
red regions and black regions had fonned. The hydrodynamic radius of the solute was 37.4
m. At the 5 p depth there were domains that were larger than 37.4 nm and smaller than
37.4 nm, again confirming the existence of a variety of length scales within the final PN
morphology .
At 10 pm (Figure 7.7 (c)), a large decrease in the volume fiaction of gel domains
accessible to solutes of 37.4 nm in size occwed. The upper lefi-hand corner contained an
area with spherical gel regions similar to nucleation and growth type morphology. Since
domain sizes decreased as PIPS proceeded, it was fogical to assume that domains which
allowed the passage of larger diameter solutes such as FDX70KD, were formed durhg the
very early stages of phase separation. This corresponded with the evidence of a nucleation
and growth mechanism that would have occurred as the system entered the metastable
region during initial poIymerization and crosslinking reactiom. The hydrogel domains that
comprised this structure coexisted with smaller hydrogel domains observed in Figures 7.5
and 7.6 at this depth. mese images provided direct evidence that decomposition by
nucleation and growth occurred in this system and had an impact on the fmal morphology
and mass transport properties of the IPN.
Harada [Haradal 9971 observed similar nucleation and growth type structures phase
contrast optical microscopy for a polymer blend crosslinked by irradiation at different t h e
intervals. The appearance of nuclei was h t noted, followed by the appearance of
interconnecting structures around the nuclei at later times. The secondary interconnected
structures grew with tirne until the phase separation process was fiozen by the crosslinking
reaction. This process was termed nucleation assisted spinodal decomposition.
By using different molecular weight fluorescent probes, different domain sizes
created during the phase separation process were distinguished. This also allowed for
examination of morphology at different quench depths and different time intervals. The
phase-separated structures fomed during earlier quenches were '%ozen" to some extent by
vitrification and increased viscosity. By using fluorescent probes that could not access
domains created at later stages, the morphology development and phase separation that
occurred at earlier, more shallow quenches were observed. Interestingiy, the larger sized
domains created during the f'irst quenches CO-existed with the smaller domains created at
deeper quenches. The analytical method developed in this chapter along with the
hydrophobic-hydrophilic IPN prepared by the monomer immersion method in chapter 6,
Iend themselves particularly well to the study of morphology development in IPN systems.
Although isolated examples of different domain sizes have been provided in the literature,
this is the fkst example of the cornplex, multi-layered morphology with multiple length
scales developed in IPNs. This morphology is attributed to the continuously changing
quench depths and associated crosslinking reactions that serve to partially arrest the
morphology formed at the different quench depths. It is expected that an increase in the
reaction temperature of the system would serve to increase the quench depth of the system.
Thus the resulting domains would be much smaller, but a variety of domain sizes would still
be formed due to the continuously decreasing LCST of the polymer system.
7.4.2 PMU-PDMS IPN Gel Domain Morphology as a Function of Depth
For the larger sized fluorescent probes (FDX4.4KQ FDX70K.D) the morphology of
the IPN varied as a function of depth (Figures 7.6 and 7.7). In Figure 7.6 (c) at 52 pm, a
macro-structure of gel domains formed which was sirnilar to that of intermediate-stage
spinodal decomposition. At this depth fiom the surface and at this particular hydrogel
domain size (domain diameter > 9.4 nm ), phase separation was fMer than the
polymenization and crosslinking reactions and the PMAA-rich phase began to separate out
into spherical domains which became more distinct (Figure 7.6 (d) ), grew in size and
coalesced in order to minirnize their interfacial area (Figure 7.6 (e) and (0). As the depth
increased the macrostnicture was "fiozen" at a later stage of SD. A similar pattern of
morphology as a hc t ion of depth is seen in Figures 7.7 (e-f).
This variation in morphology with depth was attributed to the diffierent extents of
polymerization and crosslinking that occurred as a fùnction of depth due to the non-uniform
distributions of crosslinker, initiator and UV intensity. The UV intensity was expected to
decrease with depth and î h e due to interference fiom polymer formation. Crosslinker and
initiator concentrations decreased towards the surface of the IPN due to diffusion out into
the surrounding monomer during IPN formation.
PIPS began at the surface of the IPN at shallow quench and progressed to a deeper
quench at a much faster rate than the interior of the IPN. Thus smaller domains were
expected to be present near the surface. At 50 pm depths fiom the surface (Figure 7.6 (c))
there was evidence of late stage S.D., since phase separation was faster than polymerization
and crosslinking due to the decrease in the UV light intensity.
In most systems used to study morphology development in polymeric systems,
attempts were made to keep parameters such as temperature, crosslinking agent, initiator
concentrations and UV light intensity constant. The samples were made thin (50 p), but
not so thin that the wetting effects of the polymer-substrate intexface dominated the bulk
morphology of the system [Trrm-Congl9991. Studies which were intentionally canied out
using a temperature gradient [Tran-Congl999bl or gradient in the guest monomer
concentration 19991 resulted in a gradient of morphological features that ranged from
interco~ected structures of constant domain size during early stages of SD to uniformly
dispersed domains fiom late stage SD.
7.5 Conclusion
LSCM examination of IPN networks pre-equilibrated with different molecular
weight fluorescent probes is a novel analytical method that allows for the visualization of a
complex series of phase separations produced at different quench depths during PIPS and
arrested by partial cure to fonn the fuial structure of the IPN.
At any one particular depth a variety of domain sizes and macrostnictms co-
existed. Morphology also changed as a function of depth due to gradients created in the
initiator, crosslinking agent and W light intensity during IPN formation. Exarnples of IPN
morphology that contained different domain sizes or phase separated structures have been
observed in the past. However, this is the first tirne that these structures have been s h o w to
co-exist in the same polymeric system as part of a series of phase separation processes
which have taken place at different quench depths, and at different rates of polymerkttion,
crosslinking and phase separation. These results imply that IPN morphology (i.e. domain
size and comectivity) is continuously changing as a hc t ion of quench depth. Furthemore,
morphology formed during earlier stages of phase separation cm become partiaily arrested,
due to increases in viscosity and vitrification, to co-exist with morphology created in the
iater stages of phase separation. This work brings a new understanding of morphology
development that takes place durhg the PIPS process of IPNs.
Figure 7.5 LSCM images of depth profile of PDMS-PMAA IPN immersed in fluorescein solution.
Figure 7.7 LSCM images o f depth-profile of PDMS-PMAA IPN immersed in FITC-dextran (70,000 Da)
CHAPTER 8 : pH DEPENDEN CE OF PDMS-PMAA IPN
MORPHQLGY AND TRANSPORT PROPERTIES
Stimuli-responsive PMAA-PDMS IPNs prepared by the monomer immersion
method of chapter 6 were examined for their use as variable pemeability membranes. The
mechanism of permeation control was dependent on a size exclusion effect. Laser scadng
confocal microscope &SCM) images of the IPN surface region (û-100 pm fiom the
sdace), showed that fluorescein isothiocyanate dextran, a hydrophilic fluorescent
penneant, was able to access the gel domain channels at pH 7, but not at pH 3. This
observation correlated to a significant decrease in pemeability at pH 3 and was ateibuted to
the decrease in hydrogel domain size during hydrogel dehydration at pH 3.
The TPN morphology of small, interconnected hydrogel domain channels did not
shorten the permeation response times of IPN membranes relative to composite membranes
(chapter 5) for 24 h penneation cycles, as was expected. The nanometer-scale gel channel
morphology did result in faster response of membrane hydration to changes in pH.
However, complete dehydration of the membrane-spanning gel domah channel at pH 3
during the 24 h pemeation cycle required complete hydration (approximately 25 h) at pH 7,
before pH 7 permeability could be re-established.
8.1 Introduction
Mechanisms of diffusion through homopolyrner or heteropolymeric systems having
time-independent properties are well understood. However, once the properties of polyrner
systems become dynarnic under the influence of external stimuli, the transport properties of
these materials become more cornplex. In this chapter we examine the eflect of IPN
morphology on the stimuli-responsive transport properties of PMAA-PDMS IPN
membranes.
For variable permeability hydrogel membranes, literature reports have included
examples where penneation is controlled by hydration changes Fei11 99 11 and also by
formation of a s&e skin ~ofEnan1986, Bae1991J or bulk squeezing of the membrane
~ofEmnl986 , Sawahatal9901. Heterogeneous polymeric systems have also been prepared
where permeation change was accomplished by formation of a surface skin. For example,
copolymers of NiPAAm with more hydrophobie comonomers, such as n-butyhethacrylate,
showed dense &.in formation during the deswelling process when the temperature increased
pst LCST values. The dense skin blocked the release of drugs fiom the matrix and Iimited
diffusion through the membrane resulting in the on-off control of solute release. Similar
results were obtained fiom IPNs composed ofNiPAAm and PTMO @3ae1991].
In another example, responsive hydrogels were grafted onto the surface and within
the pores of porous, solid polymerric substrates @3arbuccil99 1, Peng200 11 and porous
hydrogel supports [Chun19961 for use as variable pemeability membranes. Pulsatile drug
delivery was achieved due to the swelling and shrinking of the responsive gel to external
stimuli. As the gels swelled, the pores became filled with gel and permeability was
obstnicted/decreased. As the gels shrunk, pores opened allowing penneation.
Stimuli-responsive IPNs have been prepared by other researchers [Gudemanl995,
Lee1996, Byun1996, Ruckensteinl9961 primarily to improve the wet strength of the
responsive hydrogel component. In most cases, both the host and guest polymer networks
have been hydrogels. Complexation between the two IPN polymer components due to
hydrogen bonding mishi1 985, Yao1993, Aoki 19941 has also been manipulated via pH and
thermal stimuli to produce variable permeability membranes. Ln al1 cases, the phase
morphology of the resultmt IPN was not exarnined, presumably because changes in the
morphology of the R N would not significantly alter the permeation properties in these
systems.
A heterogeneous hydrogel-elastomer IPN membrane was recently prepared f?om
p(NiPAAm) and PMMA [Lu2000]. The hydrogel formed the continuous phase and the
hydrophobic component present at lower volume fiactions of 5%, 10% and 15%, f o n d the
discomected domains. This new material was stronger than the hydrogel but swelling
kinetics and release rates were much slower than the hydrogel only due to the presence of
the hydrophobic domains. Permeation control for this membrane was based on hydration.
The stimuli-responsive PMAA-FDMS IPN membranes w d in this chapter were
prepared using the monomer immersion method described in Chapter 6. The IPN
membrane has been shown to have morphology very different fiom the composite
membrane of Chapter 5 due to the mechanism of phase separation that occurred during IPN
formation (Chapter 7). Unlike the discrete domains of the composite membrane, the
hydrogel domains of the IPN were connected throughout the membrane thickness. The
diameter of the hydrogel domains (ranging fiom less than 10 nm to p a t e r than 40 nm)
were much smaller than those of composite membranes (90 pm at pH 7) and comparable in
size to the permeants of interest in this work (0.8 nrn to 37.4 nm).
For the PMAA-PDMS IPNs prepred in this work, it was hypothesized that the
polyelectrolyte gel component in the IPN would swell and deswell as a function of pH,
making the hydration of and pemeation tbrough the IPN responsive to pH stimuli. More
specXcally, it was hypothesized that the mechanism of permeation would be based on a size
exclusion effect, where the size of the hydrogel domain channel would be dependent upon
pH and penneating solutes would be excluded h m regions of hydrogel channels as the
diameter of the hydrogel channel became smaller than the diameter of îhe solute. It was also
hypothesized that the size-exclusion rnechanism of permeability change dong with the
small diameter hydrogel domain channels would increase the ON/OFF dnig delivery ratios
and decrease the hyhtion and permeation response times relative to both homogeneous gel
membranes and composite membranes. This proposed mechanism for membrane
permeation control was very diEerent h m the mechanisms which control permeability
change in hydrogels (hydration, bulk squeezing, surface skin), composites (Chapter 5,
percolation) and the heterogeneous polymer systems mentioned above.
8.2 Methods
Sections of the PMAA-PDMS IPN membrane were placed in pH 7,s and 3 buffers
and allowed to corne to equilibrim hydration in a similar rnethod to that described in
Chapter 4.3.2. The sections were then placed in fluorescein isothiocyanate dextran
(FDX4.4K.D, M.W. 4,400 Da, h.d. 9.4 nrn) solution of comesponding pH for a 2 week
period. The sections were subsequently examined using the LSCM as a fùnction of depth in
a similar rnethod to that described in Chapter 4.3.4.
Equilibnum and dynarnic hydration and permeation studies were carried out in a
similar method to that described in section 4.3.3. The permeant used in these studies was
vitamin B12 (1,355 Da). The hydration cycles were conducted for 2h and 24 h cycles.
Permeation cycles were conducted for 4h and 24 h cycles.
8 3 Results
Figure 8.1 is comprised of LSCM images of PMAA-PDMS IPNs that had been
equilibrated to pH 7, pH 5 and pH 3 hydration and then immersed in concenbated solutions
of FDX4.4KD. Figure 8.1 (a) and (b) are images of PDMS-PMAA IPN at pH 7 and depths
of 10 and 20 p respectively. Figure 8.1 (c) and (d) are images of PDMS-PMAA IPN at
pH 5 and depths of 10 and 20 jm, respectively. Figure 8.1 (e) and (f) are images of PDMS-
PMAA IPN at pH 3 and depths of 10 and 20 p., respectively. The images show that the
fluorescent regions of the IPN decreased as pH decreased, so that at pH 3 and 20 p there
were very few fluorescent regions. The fluorescent regions represent the hydrophilic
domains that are accessible to the permeating solute, FDX4.4KD. This implied that as pH
decreased the accessible hydrophilic domains decreased.
Table 8.1 lists the hydration of the P U hydrogel, the PMAA-PDMS composite
(prepared in Chapter 5), the PMAA hydrogel component of the composite, the PMAA-
PDMS IPN (prepared in Chapter 6) and the hydrogel component of the IPN at pH 7 , s and
3. The composite and IPN membranes have sirnilar hydration at pH 7 and 5. The hydrogel
component of the composite and IPN membranes has similar hybtion to the PMAA
hydrogel at pH 7 and pH 5. At pH 3, the hydrogel component of the IPN has siwlar
hydration to the PMAA hydrogel, but that of the composite membrane is much higher.
Thus, the hydrogel component of the composite membrane did not reach complete
dehydration at pH 3, but the gel component of the IPN did.
Table 8.2 iists the permeability at pH 7,s and 3 and the ON/OFF ratios for vitamin
B12 through the PMAA hydrogel, the PMAA-PDMS composite and the PMAA-PDMS IPN.
Permeability decreased with decreasing pH. The IPN membrane had the lowest
permeability and the largest ON/OFF permeability ratio.
Figure 8.2 is a graph of hydraion of the PMAA-PDMS IPN membrane as a function
of time and pH. The membrane was initially equilibrated to pH 7 hydration and then placed
in pH 3 for a 2 h interval where hydrogel dehydration was observed. After 2 h the IPN was
placed in pH 7 where dehydration continued to occur for the following 2 h period, at which
point the IPN began to hydrate.
Figure 8.3 is a graph of the hydration of the PDMS-PMAA IPN membrane as a
fùnction of thne and pH. The membrane was initially equilibrated to pH 7 and then placed
in pH 3 for 24 h when gel dehydration was observed and equilibrium hydration at pH 3 was
reached. After 24 h the IPN was placed in pH 7 where hydration began immediately and
approximately 25 h was required for the membrane to reach equilibrium hydration at pH 7.
Figure 8.4 is a graph of the mas permeated through PDMS PMAA IPN as a
fùnction of time and pH. Initially the permeability at pH 7 was constant at 6.53 x 1 O* cm2/s.
As the membrane was placed in pH 3 for a 4 h cycle, the mass transfmed and
corresponding permeability was zero. When the pH was changed to 7, permeability
immediately returned to a constant value of 6.74 x 1 od cm2/s, comparable to the
permeability before the pH change.
Figure 8.5 is a graph of the mass permeated through PDMS PMAA IPN as a
function of tirne and pH for a 24 h pemeability cycle. Initidly the permeability at pH 7 was
constant at 2.8 x IO-' c d s . As the membrane was placed in pH 3 for a 24 h cycle, mass
transfer and corresponding permeabiiity was zero. When the pH was changed to 7, zero
flux continued for appmximately 20 h, at which point permeation through the membrane
began again but at a substantially reduced permeability of 1.4 x 10-~ cm2 /S.
8.4 Discussion
8.4.1 Mechanism of Permeation Change
In Chapter 7 the accessible or permeable gel domain channels were visualized using
LSCM and different sized fluorescent solutes. The gel domain size ranged fiom less than 10
nm to greater than 40 nm. The domains were comparable in size to the solutes used in the
permeation studies. It was hypothesized that during dehydration at pH 3, the gel domains
becarne even smaller than the diameter of the solute, effectively impeding the diffusion of
solutes through the gel channels, resulting in a pH dependent variable permeability
membrane. In order to provide evidence that the dominant mechanism of perrneation
control in IPNs occmed via a size exclusion effect, images were taken of the accessible
regions of the IPN membrane pre-equilibrated in FDX4.4KD at different pH (7,s and 3)
and at depths of 10 and 20 pn fkom the surface (Figures 8.1 (a) - (0).
At pH 7 (Figures 8.1 (a) and (b)), it was evident that FDX4.4KD was permeable
through all the domains at depths of 10 and 20 pm from the surface. Both figures consisted
of a homogeneous dispersion of very smali red domains that represented hydrogel domains
that were accessible to the fluorescent solute.
At pH 5 (Figures 8.1 (c) and (d)), hydration of the IPN had decreased to 0.63 (fjrom
0.8 at pH 7). Black regions appeared in both the images taken at 10 p and 20 p. The
black regions represented areas that did not fluoresce because the fluorescent solute had not
been able to permeate through these regions. Since we know fkom Figures 8.1 ((a) and (b))
that hydrogel domains were present at these depths, it was concluded that the fhorescent
solute was unable ta penneate into these regions because the hydrogel domains had
dehydrated and their diameter was less than 9.4 nm. This indicated that the diameter of the
gel domain was actually smaller than the mesh size of the gel since we know that this solute
permeated the gel at pH 3. Although crosslinks may not have been present almg the cross
section of the domain due to its small size, it is conceivable that the gel domain is
crosslinked along its length and that one of the factors which allows the nm-scale domah to
remain stable is the presence of these cross^ along the length of the gel domain charnel.
In Chapter 7, we have shown that although a large range of domain sizes coexisted at any
given depth, the general trend was for hydrogel domains to become smaller as the depth
from the membrane surface decreased. This observation correlated with the larger
proportion of inaccessible black regions as the depth from the surface increased.
At pH 3 (Figures 8.1 (e) and (f)), hydration of the IPN had decreased to 0.47. At 10
pm it appeared that the permeable gel domains had aggregated together to form larger
spheres with distinct black areas surrounding these spheres. At this depth there were still a
significant number of domains that were permeable to the solute. However, at 20 pn there
were only a few clusters of accessible gel domains present. The majority of hydrogel
domains had become impermeable to the solute due to dehydration and decrease in gel
domain diameter. Since permeation did not occur in this membrane, it was expected that as
the depth increased, there would be no permeable, fluorescent domains present. This was
corroborated by LSCM images taken at greater depths, which appeared black.
Thus direct evidence of the size exclusion effect in IPN membranes was provided
using the LSCM, where hydrogel domain charnels were no longer accessible to the
permeating solute as the pH decreased to 3 and the hydrogel domains dehydrated and
became smaller in diameter.
8.4.2 Effect of IPN Morphology on Hydration and Permeation.
8.4.2.1 (a) Equilibrium Hydration and Permeation Ploperties
TabIe 8.1 shows equilibrated hydration data for PMAA hydrogel membranes,
PDMS-PMAA composites, PDMS-PMAA IPNs, and the hyhge l component in both the
composite and IPN membranes at pH 7,5 and 3. Hydration values decreased as pH was
Iowered due to the stimuli-responsive properties of the PMAA hydrogel component.
Hydration of the IPN hydrogel component was found to be the same as that of the
PMAA hydrogeI, based on a 30% dry PMAA hydrogel content. For exarnple, at pH 7 the
hydration of the PMAA hydrogel and that of the hydrogel in the IPN were both 0.94. At pH
3, the hydroge1 domain chaunels completely dehydrated to 0.74, the same hydration value as
PMAA hydrogels at pH 3. This was not the case for the hydrogel component found in the
composite membrane. Hydration of the hydrogel component at equilibrated pH 3 conditions
(0.83) was simcantly larger than that of the PMAA hydrogel at pH 3 (0.74). This was
attributed to the morphology of the composite membrane. Because the hydrogel domains
were discrete, as deswelüng occurred in the outer regions of the membrane, îhe hydrogel
domains deswelled, decreased in size and the charnels became discomected. Complete
dehydration did not occur for the hydrogel domains in the interior regions because water
was unable to diffuse out of the membrane. For IPN membranes, the connected hydrogel
domain channels allowed for complete dehydration to equilibrium pH 3 values, since water
had a channel or pathway to d i h e out fiom. Thus, the difTerent hydrogel hydration results
at pH 3 for the composite and IPN membranes were atûibuted to their different morphology.
The equilibrated permeation values (Table 8.2) for vitamin B12 through the IPN
membrane were much smaller than for the hydrogel and composite membranes even though
the hydration of al1 membranes were simila- at pH 7. This was attributed to the smaller
diameter, more tortuous channel path located within the IPN membrane. Although the
general trend of decreased penneation as the pH decreased applied to both systems, the
equilibrated ON/OFF ratios for the l[PN were much larger due to the negiigible perrneability
at pH 3. Most likely, at pH 3, there existed regions of very small diameter gel domain
channels within the bulk of the IPN membrane that did not allow the solute to pemieate
through the membrane.
8.4.2.2 fi) Dynamic Hydration and Permeation Roperties
One of the advantages that the unique morphology of the IPN was expected to
produce was a faster response of hydration and penneation to changes in surrounding pH. It
was hypothesized that the relatively small, connected hydrogel domain channels would
respond very quickly to changes in pH, due to much fwter d i f i i on of buffer salts and ions
through the smalI diameter hydrogel channels than through the much larger, discrete
hydrogel domains of the composite membrane. However, experirnental data provided in
Figures 8.2 - 8.5 do not support this hypothesis.
In Figure 8.2, a membrane was placed in pH 3 bufEer solution for a 2 h period. The
response was a reduction in hydration from 0.87 to 0.78. This rate of reduction in
membrane hydration was sirnilar to that for composite membranes. However, when the
membrane was retumed to pH 7 conditions, hydration of the membrane continued to
decrease for approximately another 2 h. Re-hydration to former levels required another 46
h. When the IPN membrane was placed in pH 3 conditions for 24 h, the membrane came to
equilibrium pH 3 hydration values within the 24 h period (Figure 8.3). Again this rate of
dehydration was similar to that of composite membranes. When the dehydrated membrane
was placed in pH 7, re-hydration immediately began to take place. However, 24 h was
required in order for the IPN membrane to rehydrate to former pH 7 equilibrium hydration
values.
The continued desweiling at shorter intervals was unexpected and was atüibuted to
the channel morphology of the IPN. In Figure 8.6 (a) the gel channel is depicted as a series
of blocks, the larger blocks represent hydrogel domains, the smaller blocks the smaller
diameter comecting hydrogel channels. When the IPN was placed in pH 3 fiom pH 7, the
pH 3 buffer difhed into the channels and the portion of the channe1 at the surface
dehydrated rapidly due to the srna11 diameter. Dehydration of the channels at the surface
substantially reduced the bulk flow of water out of the gel channel for dehydration purposes.
Thus at small dehydration cycle times, bulk water was trapped within the mid-section of the
gel channel. When the membrane was placed in pH 7 conditions, the channels at the surface
quickly rehydrated, bulk water in the mid-section was allowed to flow out of the gel c h d
and produced a deswelling effect as pH 3 buffer diffised into the midsection of the hydrogel
channel, until the entire channel reestablished pH 7 hydration conditions.
Interestingly, the continued deswelling which took place at short cycle times did not
adversely affect the permeation kinetics that took place at short cycle times. In Figure 8.4
the permeability of the IPN membrane at pH 3 conditions was abruptly reduced to zero.
After 4 h, when pH 7 conditions resumed, the membrane permeability quickly returned to its
former value. This was attributed to the short cycle times, where only the surface gel
domain channels were able to dehydrate and the rnid-section of the channel remained
hydrated at essentially pH 7 conditions.
The main disadvantage to variable permeation control of the IPN membrane was the
complete dehydration of the gel channel when placed at pH 3 conditions for 24 h (Figure
8.5). Zero permeability continued for approximately 20 h after the membrane was retumed
to pH 7. When permeability did resume it was less than half the equilibrium permeability at
pH 7 conditions for the following 10 h period. Thus, the small diameter of the gel domain
channel allowed for complete dehydration at pH 3 during a 24 h cycle. When the membrane
was retumed to pH 7, permeation did not resume until the hydrogel domain channel was
larger than the penneating solute along the entire length of the channel. For the above
permeation conditions, this required 20 h. Furthemore, equilibrium pH 7 permeability was
not re-established after 15 h.
8.5 Conclusions
The mechanism of permeation conirol for PMAA-PDMS IPN membranes was
detennined to be a size exclusion effect. With dehydration at pH 3 the diameter of the
hydrogel domain channels decreased to sizes smaller than the diameter of the permeating
solute, thereby stopping diffiion of îbe solute through the membrane. The large time lag
for 24 h stimuli-responsive pemeation cycles for the PMAA-PDMS IPN membrane
provided conclusive evidence that the membrane-spanning gel domain channel of the IPN
was not the most appropriate morphology for stimuli-responsive mass transfer applications.
Depending upon the particular application, the composite morpho1ogy was superior to the
IPN morphology for variable mass transfer. For composites, partial dehydration (Le. surface
particles only) within a 24 h period was adequate to substantially reduce the penneability.
When the composite was placed in pH 7 (the ON state) permeability was quickly
reestablished because only the surface particles needed to undergo hydration in order to
restore full membrane pemeation. The main advantage of the IPN was the ability to
produce small gel domains, the fact that they formed channels which spanned the thickness
of the membrane did not provide any advantaes for variable rnass transport applications.
Most likely, the preparation of an IPN with small interconnected hydrogel domain channels,
where the swcfàce portion of the gel charme1 is responsive but the interior portion of the
hydrogel channel remains hydrated and non-responsive wouId be the best compromise
between the composite and PN morphologies studied in this thesis.
1 1 Hydra tion 1 Membrane Type
PMAA gel
1 Gel in Composite ( 0.94 1 0.86 1 0.83 1
PH 7
Composite
0.94
Table 8.1 : Hydration of PMAA gel, Composite and IPN membranes and of gel components in Composite and IPN membranes at pH 7,5 and 3.
PH 5
0.82
Gel in IPN
Permeability (x IO-' cm2/s)
PH 3
0.86
Membrane 1 PH 7 1 pH5 1 PH 3 1 ON/OFF 1
0.74
0.63
0.93
0.58
0.88
Type
Table 8.2 Permeability of PMAA gel, Composite and IPN membranes equilibrated at pH 7,5 and 3. O W F F ratio is ratio of penneability at pH 7 over permeability at pH 3.
- - - -
0.74
Ratio
PMAA gel
Composite
IPN
18
5.2
1.7
19
3.7
0.024
3
0.14
~ 1 . 7 x lo6
6
40
>1.7 x 106
Figure 8.2 2 h Hydration cycle for PDMS-PMAA IPN membrane.
Tirne (h)
Figure 8.3 24 h Hydration cycle for PDMS-PMAA IPN membrane.
Time (h) t
Figure 8.4 4 h Permeation cycle of vitamin B12 b u & PDMS-PMAA IPN membrane.
Time (h) I
Figure 8.5 24 h perrneation cycle of vitamin Br2 though PDMS-PMAA IPN membrane.
IPN MEMBICANE PLACED IN
(a) pH 7 equilibrium hydration of hydrogel channel in IPN
IPN MEMBRANE PLACED IN pH 3
(b) pH 3 begins to difise into surface section of hydrogel channel. Difision ofbulk water fiom mid-section considerably reduced due to dehydration of surface channels. Midsection remains hydrated.
IPN MEMBRANE PLACED IN pH 7
(c) IPN placed in pH 7. Surface domains hydrate to pH 7, allowing bulk water form interior to diffise out of membrane resulting in further dehydration.
Figure 8.6 Depiction of hydration response of hydrogel domain channel of IPN placed at different pH conditions.
CHAPTER 9 : CONCLUSIONS AND RECOMMENDATIONS
The main objective of this thesis has been to examine the effect of a hydrogel-
elastomer morphology on the variable transport properties of stimuli-responsive
membranes. This objective has been subdivided M e r into three primary goals:
1. Preparation of heterogeneous PMAA-PDMS composite and IPN membranes.
2. Investigation of the mechanism of permeation control for each membrane system.
3. Evaluation and comparison of the variable transport properties for each membrane
system.
The thesis objective and goals were guided by the following hypothesis:
Permeation control in stimuli-responsive hydrogel membranes, which
occurs via hydration only, can be enhanced using heterogeneous
systems where bydration changes may be coupled with changes in gel
domain connectivity (percolation) and/os gel domain sue (size
exclusion).
The following sections summarize the conclusions of this thesis with reference to the
goals outlined above. It is shown that the conclusions validate the thesis hypothesis and the
work as a whoIe contributes to the current body of knowledge in the area of variable
permeability membranes and the field of controlled drug delivery. The last section of this
chapter briefly outlines areas in which fiirther research work can be carried out as a natural
progression of the contributions made in this thesis.
9.1 Preparation of Heterogeneous PMAA-PDMS Composite and IPN
Membranes
Bicontinuos, permeable, pH-responsive PMAA-PDMS composite and IPN
membranes were prepared in this work. Although the dry PMAA gel content for both
membranes was approximately 30% both membranes hydrated relatively quickly in water,
and were penneable to water-soluble compounds. Membrane hydration was a reversible
fhction of pH. Composite membranes required weeks in order to come to swelling
equilibrium whereas IPN membranes achieved swelling equilibrium within a 24 h period.
This difference in swelling kinetics was attributed to the small nm-scale PMAA channel
murphology of the IPN.
For the PMAA-PDMS IPN, the monomer immersion method was developed for the
preparation of bicontinuous hydrogel-elastomer IPN membranes. This represented an
improvement over conventional methods of IPN preparation that produced hydrogel-
elastomer membranes with a sea-island morphology that was imperneable to water-soluble
compounds. Immersion of pre-IPN films in the guest monomer throughout synthesis
emured an even monomer concentration profile, and produced a uniform bicontinuous
morphology throughout the 1PN. In cornparison, PMAA-PDMS sequential IPNs
synthesized while in contact with glass or air (conventional methods) resulted in
impermeable, spatiaiiy varying morphology which ranged fiom dispersed hydrogel domains
near the surface to a bicontinuous morpblogy some distance below the surface. A paper
has been published [TumerZOOO] and a U.S. patent has been allowed regarding the process
and the materials fonned fkom the process.
Composite and IPN membrane types were specifically chosen in this work because
their morphology was very different, thth allowing significant conclusions to be drawn
regardhg the effect of a heterogenous membrane morphology on variable transport
properties. Composites were fonned fiom a blend of E 40 j.m PMAA gel particles
homogeneously dispersed within PDMS resin. The PDMS resin was c d after dispersion
of the gel particles. LSCM images of fluorescently labelled composite membranes confirm
that the composite morphology consisted of a uniform dispersion of PMAA gel particles and
that gel connectivity increased as pH and hydration of the gel particles increased
[Chan1 9951.
The morphology (size and distribution of PMAA gel domains) fonned in IPN
membranes was more difficult to ascertain. LSCM images of IPNs equilibrated in
fluorescein solution show that the PMAA-PDMS IPN morphology consists of a
homogeneous dispersion of nodular-shaped PMAA gel domains of approximately uniform
size. The current literature has different opinions and evidence as to whether the domain
size produced in IPN systems is d o m i or of multiple length scales. Theory, however,
suggests that the final IPN morphology is most likely composed of a variew of different
domain sizes based on the constantly changing quench depth which takes placed during IPN
formation.
A different approach was developed in this work to examine the IPN morphology.
The LSCM was used to examine IPN networks preswollen with different rnolecular weight
fluorescent probes. The resultant images reveal cornplex, superimposed structures of
hydrophilic domains of varying sizes (z 10 - 100 nm ) and spatial distributions. This
morphology was attributed to the phase-separated structures formed and parîidly arrested at
each successive quench depth during Polymerization hduced Phase Separation. These
observations have not been reported previously and present a new understanding of
morphology developrnent in IPNs.
9.2 Mechanism of Permeation Control for PMAA-PDMS Composite
and IPN Membranes
The mechanism of permeation control for composite membranes was found to be a
synergistic function of PMAA gel domain connectivity and hydration. A semi-logarithmic
plot of permeability of vitamin B12 through the composite versus the inverse of hydration of
the PMAA gel domain yielded a straight Iine that was consistent with the hydration
mechanism described by Yasuda. The slope of this line was much larger than the slope of
the h e of a similar plot based on PMAA gel membrane permeation studies. This indicated
that for similar changes in hydration composite membranes experienced much larger
changes in permeabiiity. This enhancement in penneation response was attnbuted to the
increase in gel domain connectivity that took place as hydration of the PMAA gel domains
in the composite increased.
The mechanism of permeation control for IPN membranes was shown to occur via a
size exclusion effect by using LSCM to visualke the accessible gel domain channels of the
IPN as a function of decreasing pH. As pH decreased fiom 7 to 3, the number of gel
domains that were accessible to the fluorescent probe decreased dramatically. This was
attnbuted to the decrease in the domain size of the probe due to dehydration of the PMAA
gel channel caused by the Iowered pH. The PMAA gel c h e l decreased to sizes smaller
than the diameter for the permeating solute, thereby preventuig diffusion of the solute
through the membrane.
The current state of the art for variable penneability membranes focuses on
hydration as the enabling mechanism for pemeation control for both homogeneous and
heterogeneous hydrogels. The above conclusions demonstrate that a heterogeneous
morphology may be used to produce new mechanisms of permeation control in variable
permeability membrane applications.
9.3 Variable Transport Properties of PMAA-PDMS Composite and
IPN Membranes
This thesis demonstrated that heterogeneous composite and IPN membranes
consisting of PMAA gel domains dispersed within a PDMS network had larger ON/OFF
drug permeability ratios, much smaller flux in the OFF state and faster response times to pH
changes than PMAA gel membranes. These improvements in h g delivery properties were
attributed to percolation and size exciusion rnechanisms, which controlled permeation
change for the composite and IPN membranes, respectively. These mechanisms of
pemeation control were a fiinction of the heterogeneous morphology of the composite and
PN systems. The above conclusions validate the thesis hypothesis.
Composites and PNs each had advantages (and disadvantages) as variable
permeability membranes that could be attributed to their different morphology. The IPN is
the only type of polymer blend system that is able to produce stable domains with
nanometer-sale diameters fiom highly incompatible polyrner components. The nanometer-
scale, interconnected gel domain morphology of the IPN membrane was able to produce
near-zero aux h the "OFF" state due to a size exclusion effect. The very high ON/OFF
ratios of the IPN system was also a reflection of the wide range of permeability values
which could be achieved by the system. The main disadvantage of the IPN morphology was
the large lag time encountered during 24 h ONIOFF permeation cycles. The slow swelhg
kinetics of the membrane-spanning PMAA channel h m pH 3 to 7 resulted in a lag time of
15 h before permeation resumed at pH 7.
In conhast the discrete gel particles of the composite membrane allowed for very
fast pemeation response to changes in pH d u ~ g both short and long 24 h ON/OFF
pemeation cycles. This highly favourabe result was attributed to the surface hydrogel
pariicles that controlled permeability. In order for permeability to be reduced to "OFF"
values, only the surfàce gel particles were required tu disconnect. Another factor was that
the hydrogel particles within the buik of the composite membrane did not completely
dehydrate at pH 3. This was an advantage in dynamic permeation experiments because it
allowed for pH 7 conditions within the membrane to be quickly re-estabIished. The main
disadvantages of the composite morphology was the relatively high flux in the "OFF" state
and the relatively low (1 60) ON/OFF ratio (compared to IPN membranes). These were
attributed to the number of permeation pathways that remained connected at pH 3.
In summry, the heterogeneous morphology was superior to homogeneous
hydrogels for stimuli-responsive mass transfer applications because of the different
mechanisms of permeation control inherent in such morphologies. The heterogeneous
morphology most likely to produce optimum, variable permeation control would consist of a
combination of both the IPN and composite morphology. For example, PDMS-hydrogel
IPNs with small interconnected hydrogel domain channels could be prepared, where the
surface portion of the gel channel was stimuli-responsive but the interior portion of the
hydrogel channel remained hydrated and non-responsive to external stimuli. In these cases
the hydrogel channel would consist of a responsive hydrogel near the surface regions and a
non-responsive hydrogel within the bulk of the membrane.
9.4 Future Research Work
Future research work may focus on one of three broad areas: (i) the improvement
of the PDMS-PMAA IPN membrane for use in controlied drug delivery applications. (ii)
the M e r understanding of morphology development during PIPS of IPNs and (iii)
preparation of hydrogel-elastomer materials using different intefices to produce unique
morphological structures.
The first area, the irnprovernent of the hydrogelelastorner IPN membrane for use in
controlled h g delivery applications, may involve the formation of an IPN where only the
surface gel domains are responsive to stimuli and the interior rernains hydrated throughout
the application. Swelhg the PDMS network initially with a hydrophilic, unresponsive
monomer such as vinyl pyrrolidone and then sweiling the surface regions only with a
responsive, hydrophilic monomer such as methacrylic acid may be used to prepare such
membranes. The eventual polymerization and crosslinking of this pre-IPN material using
UV light should result in a bicontinuous rubber hydrogel IPN that is stimuli-responsive at
the surface regions only. Such a morphology would allow the IPN membrane to achieve the
k t response times observed in composite membranes. Maintainhg a nm-scale gel channel
morpholom wilI allow the membrane to produce high ON/OFF permeability ratios and very
low flux in the OFF state.
In another application, the P U - P D M S IPN may be used as a h g delivery
mtrix where h g is preloaded into the pre-IPN prior to IPN formation. The material would
then exist as three phases; rubbery PDMS, hydrophilic PMAA gel, and h g particles. The
kinetics of dmg reIease when placed in water and as a fiuiction of pH would be very
interesting. It is expected that drug release would be pH-responsive.
The second area of research may focus on a fùrther understanding of morphology
development during PIPS in IPNs. Having devised a method of preparing bicontinuous
hydrogel-elastomer membrane which are not disturbed by surface effects and a method to
examine the complex morphology of IPN membmnes, both of these methods can be used to
gain fùndamental laiowledge regarding morphology development in IPNs. Changes in the
parameters which affect the rates of polymerization, crosslinking and phase separation
during IPN formation wiil lead to IPN materials of varying morphology and serve to gather
insight into the morphology development of IPN materials. Parameters that rnay be changed
include the M, of PDMS, the crosslinker concentration of PMAA, UV Uradiation intensiiy,
length of time irradiated, and the reaction temperature during IPN formation.
A third area of research rnay concentrate on the preparation of hydrogel-elastomer
PN materials having different morphologies based on contact with different surface types.
For example, preparation of an IPN materid which is bicontinuous only at one surface or
bicontinuous throughout the bulk but containhg a thin hydrogel layer at the surface wouId
be useful morphologies in applications such as wound dressings or contact lenses,
respectively. Their preparation could be easily carried out by selecting the proper substrate
that encourages the migration of one of the components to the surface-substrate interface.
Amsden BG, Cheng Y. Enhanced Fraction Releasable above Percolation Threshold fiom Monoliths containing Osmotic Excipients. Journal of Controlled Release 1 994;3 1 (1):2 1 -32.
Amsden B, Cheng Y. A Generic Protein Delivery System Based on Osmotically Rupturable Monoliths. Journal of Controlled Release l995;33(1):99- 105.
Amsden BG, Cheng YL. Enhanced hction releasable above percolation threshold fkom monoliths containhg osmotic excipients. Joumal of Controlled Release l996;39(l): 1 O9
Aoki T, Kawashima M, Katono H, Sanui K, Ogata N, Okano T, Sakurai Y. Temperature-Responsive Interpenetrating Polymer Networks Constructed with Poly(acry1ic acid) and Poly(N,N dimethy lacry larnide). Macromolecules 1 994;27:94742.
Bae YH, Okano T, Kim SW. Insulin Permeation through Thermo-sensitive Hydrogels. Journal of Controlled Release l989;9:271-9.
Bae, Y.H., Okano, T., and Kim, S.W. inventors. University of Utah, assignee. Heterogeneous Interpenetrating Polymer Networks for the Contrulled Release of Dnigs. United States. 493 1287. 1990; p. 1 206,593.
Bae YH, Okano T, Ebbert C, Heiber S, Dave S, Kim SW. Heterogeneous interpentrating polymer networks for drug delivery. Journal of Controlled Release 1991;16:189-96.
Bae YH, Kim SW. HydrogeI delivery systems based on polymer blends, block co- polymers or interpenetrating networks. Advanced Drug Delivery Reviews 1993;11(1,2):109-36.
Barbucci R, Casolaro M, Magnani A. The role of polyelectrolytes in the permeability controI of insulin: behavïour of poly(N-acryloyl-glycine) graftedon porous cellulose membrane. Journal of Controlled Release 199 1 ; 17:79-88.
Bell CL, Peppas NA. ; Mikos AG, Murphy RM, Bernstein H, Peppas NA, editors.Biomaterials for Drug and Ce11 Delivery. Pittsburgh: Materials Research Society, 1994;Poly(methacrylic acid-g-ethylene glycol) Hydrogels as pH Responsive Biomedical Materials. p. 199-204.
Binder K, Stauffer D. Physical Review Letter 1973;33 : 1 O06
Brazel CS, Peppas NA. Pulsatile local delivery of thrombolytic and antithrombotic agents using poly(N-ispropylacrylamide-CO-methacrylic acid) hydrogels. Joumal of Controlled Release l996;3 9:57-64.
Breirner DD. ; Gurny R, Junginger HE, Peppas NA, editorsPulsatile h g Delivery: Current Applications and Future Trends. Stuttgart: Wissenschafiliche Verlagsgesellschafi mbH, 1993; 2, The Need for Rate and Time Prograrnrning in Future Drug Delivery. p. 25-40.
Brondsted, H., Ph.D. Thesis, University of Utah, Salt Lake City; 1991;
Brotzman RW, Eichinger BE. Volume dependence of the elastic equation of state. Solution-cured Poly(dhnethy1 siloxane). Macromolecules 1 98 1; 14: 1445-8.
Burford R, Chaplin R, Mai YM. ; Culbertson My editors.Multiphase Macromolecular Systems. New York: Plenum, 1989;
Byun J, Lee YM, and Cho C-S. Swelling of Thennosensitive Interpenetrating Polymer Networks Composed of Poly(vinyI alcohol) and Poly(acry1ic acid). Journal of Applied Polymer Science l996;61697-702.
Carelli V, DiColo G, Guerrini C, Namipieri E. Dnig Release from Silicone Elastomer through Controlled Polymer Cracking: an extension to macromolecular drugs. Inernational Joumal of Phannaceutics l995;5O: 1 8 1-8.
Chen WJ, Li XL, Dong T, Jiang M. Spinodal decomposition induced by cross-linking reaction in a binary polyrner mixture. Macromolecular Chemistry and Physics 1998; 199(3):327-33.
Chen Z, Ward R, Tian Y, Eppler AS, Shen YR, and Sornojai GA. Surface Composition of Biopolyrner Blends Biospan-SP/Phenoxy and Biospan-FPhenoxy Observed with SFG, WS, and Contact Angle Goniometry. Journal of Physical Chemistry B 1999;103(15):2935-42.
Chen J, Park K. Synthesis and characterization of superporous hydrogel composites. Journal of Controlled Release 2000;65(1-2):73-82.
Chou YC, Lee W. Reaction-Induced Phase Separation during the Formation of a Polyurethane-Unsaturated Polyester Interpenetrating Polymer Network. Polymer Engineering and Science l994;34(16): 1239-49.
Chun S and Kim J. A novel hydrogel-dispersed composite membrane of poly(N- isopropylacrylarnide) in a gelatin matrix and its thennally actuated permeation of 4- aceatmidophen. Journal of Conîrolled Release 1996;3839-47.
Cifkova 1, Lopour P, Vondracek P, Jelinek F. Silicone rubber-hydrogel composites as polymenc biomaterials. Biomaterials 1 990; 1 1 :393-6.
Clarson SJ. ; Clarson SJ, Semylen JA, editors.Siloxane Polyrners. New Jersey: Prentice Hall, 1993;
Cohen MH, Tumbull D. Molecular Transport in Liquids and Gases. Journal of Chemical Physics 1 959;3 1 : 1 1 64-9.
Cole CA, Schreiner SM, Priest JH, Monji N, H o f i a n AS. Lower critical solution temperatures of aqueous copolymers of N-isopropylacrylamide and other N-substituted acrylamides. ACS Symposium Series 1 987;3 50:255-84.
Doi M, Matsumoto My Hirose Y. Deformation of Ionic Polymer Gels by Electric Fields. Macromolecules l992;Z :5 504-1 1.
Donatelli AA, Sperling LH, Thomas DA. Macromolecules 1976;9:67 1-6.
Dong L, Hofban AS. Synthesis and Application of Thermally Reversible Heterogels for Drug Delivery. Journal of Controlled Release 1990; 1 3 :2 1-3 1.
Edwards DC. Water Absorption Phenornena in Elastomers. Elastomerics l985;25-9.
Fedors RF. Osmotic Effects in Water Absorption by Polymers. Polymer l98O;2 1 :207- 12.
Fei1 H, Bae YH, Feijen J, Kim SW. Molecular Separation by Thennosensitive Hydrogel Membranes. Journal of Membrane Science 199 1 ;64:283-94.
Folkman J, Long DM. Journal of Surgical Research 1964;4:139-42.
Gehrke S. ; Amidon GL, Lee PI, Topp EM, editors.Transport Processes in Pharmaceutical Systems. New York: Marcel Dekker, 2000;Synthesis and Properties of Hydrogels.
Gehrke SH, Lee PI. ; Tyle P, editors.Specia1ized h g Delivery Systems. New York: M.Dekker, 1990; 8, Hydrogels for Drug Delivery Systems. p. 333-89.
Gehrke SH, Cussler EL. Mass Transfer in pH-Sensitive Hydrogels. Chemical Engineering Science 1 989;44(3):5 5 9-66.
Grimshaw PE, Grodzinsky AJ, Yarmush ML, Yamiush DM. Selective Augmentation of Macromolecular Transport in Gels by Electrodifision and Electrokinetics. Chemical Engineering Science lWO;45(9):2917-29.
Grodzinslq AJ, Gnlmshaw PE. ; Kost J, editors.Pulsed and Self-Regulated Drug Delivery. New York: Plenum Press, 1990;Electrically and Chemically Controlled Hydrogels for h g Delivery. p. 47-64.
Gudeman LF and Peppas NA. pH-Sensitive membranes fiom poly(viny1 a1cohol)l poly(acry1ic acid) interpenetrating networks. Journal of Membrane Science 1995; 107239- 48.
Harada A, Tran-Cong Q. Modulated Phases Observed in Reacting Polymer Mixtures with Competing Interactions. Macromolecules 1 997;6: 1643-50.
Hourston DJ, Song My Schafer FU, Pollock HM, Hammiche A. Modulated differential scanning calorimetry : 13. Analysis of morphology of poly(ethy1 rnethacrylate) polyurethane interpenetrating polymer networks. Thermochimica Acta l998;324(1-2): 109-2 1.
Hariharan D, Peppas NA. Swelling of Ionic and Neutra1 Polymer Networks in Ionic Solutions. Journal of Membrane Science 1 993;78: 1 - 1 2.
Harland RS, Peppas NA. Hydrophilichydrophobic, block and graft copolymeric hydrogels: synthesis, characterization, and solute partition and peneîration. Journal of Controlled Release 1993;26: 157-74.
Hasa J, Ilavsky M, Dusek K. Deformational, Swelling and Potentiometric Behaviour of Ionized Poly(methacry1ic Acid) Gels. 1. Theory. Journal of Polymer Science: Polymer Physics Edition 1975; 13 :253 -62.
He, X., Herz, J., Meyer, G. et al., inventors. Norsolor, assignee. Process For The Manufacture Of A P o l p e n c Material Containhg Interpenetrated Polysiloxane- Polyacry lic Networks. United States. 5424375. 1 995; p. 1 34,620.
Heilmann K. ; Heilmann K, editors.Therapeutic Systems. Rate Controlled Drug Delivery : Concept and Development. Stuttgart: Georg Thiem Verlag, 1984;
Heller J. Polymers for Controlled Parenteral Delivery of Peptides and Proteins. Advanced Drug Delivery Reviews 1993; 10: 1 63-204.
H o h a n AS, Afrassiabi A, Dong LC. Thermally reversible hydrogels. II Delivery and selective removal of substances fkom aqueous solutions. Journal of Controlled Release 1986;4:213-22.
H o f i a n AS. Intelligent Polymers in Medicine and Biotechnology. Macromolecular Symposium 1 995;98 :645-64.
Hron P, Vymazalova 2, and Lopour P. Water-swellable rubbers containing powdery poly(acry1amide) hydrogel. Biornaterials 1 997; 1 8 1 O69
Hsu T, Langer R. Polymers for the controllcd release of macromolecules: Effect of molecular weight of ethylene-vinyl acetate copolymer. Journal of Biomedical Materials Research 1985;19:445-60.
Ichikawa H, Fukumori Y. A novel positive thennosensitive controlled release microcapsule with membrane of nano-sized poly(N-isopropylacrylamide) gel dispersed in ethylcellulose matrix. Journal of Controlled Release 2000;63: 1 07- 19.
Ingenito, DR., Rugge, H.F., Soane, D.S. et al., inventors. Rasor Associates Inc. assignee. Ocular Lens Composition And Method of Formation. United States. 5 789483. 1998; p. 1 706,280.
Inoue T. Reaction-Induced Phase Decomposition in Polymer Blends. Progress in Polymer Science 1995;SO: 1 19-53.
Inoue T, Chen G, Nakame K, Hofban AS. A hydrophobically rnodified bioadhesive polyelectrolyte hydrogel for dmg delivery. Journal of Controlled Release 1997;49: 167- 76.
Ishihara, K. Synthesis of stimuli responsive polymers and their biomedical applications. Waseda University, Japan; 1 986;
Ito Y, Park YS. Signal-responsive gating of porous membranes by polymer brushes. Polymer for Advanced Technologies 2000; 1 l(3): 136-44.
Kaneko Y, Nakamura S, Sakai K, Kikuchi A, Aoyagi T, Sakurai Y, Okano T. Pubatile h g release regdation ushg fast swelling-deswelling changes of novel thermo- responsive polymer gel containing poly(ethy1ene glycol) graft chahs. Kagaku Kogaku Ronbunshu 1 998;24(2):200-4.
Kaneko Y, Sakai K, Kikuchi A, Sakurai Y, Okano T. Fast swelling/deswelling khetics of comb-type grafted poly(N-isopropylacrylmide) hydrogels. Macrornolecular Symposium 1996; lO9:4 1-53.
Katono H, Maniyarna A, Sanui K, Ogata N, Okano T, Sakurai Y. Thermo-responsive swelling and drug release switching of interpnentrating polymer networks composed of poly(acry1amide-CO-butyl methacrylate) and poly (acrylic acid). Journal of Controlled Release 1991;16:215-28.
Kiefer J, Hedrick JL, Hilborn JG. Macroporous Thermosets by Chemically Induced Phase Separation. Advanced in Polymer Science 1999; 147: 161 -247.
Kim HJ, Lee K, Lee HH. Healing of hctured polymers by interdiffision. Polymer 1996;37(20):4593-7.
Kim Y and Kim S. Properties of Polyetherimide/Dicyanate Semi-Interpenetrating Polymer Network having the Morphology Spectrum. Macromolecules 1999;32(7):2334- 41.
Kim SY, Lee YM. Drug Release Behaviour of Electrical Responsive Poly(viny1 alcohol)/Poly(acrylic acid) IPN Hydrogels under an Electric Stimulus. Journal of Applied Polymer Science 1999;74: 1752-6 1.
Kou 3H. ; Amidon GL, Lee PI, Topp EM, editors.Transport Processes in Pharmaceutical Systems. New York: Marcel Dekker, 2000;Transport in Polymer Systems.
Kwon IC, Bae YH, Okano T, Berner B, Kim S W. Stimuli Sensitive Polymers for Drug Delivery Systerns. Makromol Chem Macromol Symp 1990;33:265-77.
Langer R, F o h a n J. Polymers for the sustained release of proteins and other macromolecules. Nature 1 976;263 :797-9.
Lee YM, Kim SH, and Cho CS. Synthesis and swelling characteristics of pH and thenno-responsive Interpenetrating Polymer Network Hydrogel composed of Poly(viny1 alcohol) and Poly(acrylic acid). Journal of Applied Polymer Science l996;623 0 1 - 1 1.
Lemmer B. Circadian Rhythms and Drug Delivery. Journal of Controlled Release f 991;16:63-74.
Li VHK, Robinson JR, Lee VIX. ; Robinson JR, Lee VHL, editors.Controlled h g Delivery. 2nd ed. New York: Marcel Dekker, Inc. 1987; 1, Influence of Drug Properties and Routes of Drug Administration on the Design of Sustained and Controlled Rdease Systems. p. 3-94.
Lim YH, Kim D, Lee DS. Drug Releasing Characteristics of Thermo- and pH- Sensitive Interpenetrating Polymer Networks Based on Poly (N-isopropylacrylamide). Journal of Applied Polymer Science 1997;64(13):2647-55.
Lipatov YS, Nesterov AE. ; Lipatov YS, Nesterov AE, editors.Thermodynamics of Polymer BIends. Lancaster: Technomic Publishing, 1997; 5, Mechanism and Kinetics of Phase Separation in Polymer Solutions and BIends. p. 275-362.
Lipatov YS, Sernenovich GM. The interrelation between the kinetics of the IPN formation at the interface with solid and surface segregation. Polymer 1999;40:6485-92.
Lopour P and Janatova V. Silicone rubber-hydrogel composites as polymeric biomaterials. Biomaterials 1995; 16633-40.
Lopour P, Vondracek P, Janatova V, Sulc J, Vacik J. Silicone rubber-hydrogel composites as polyrneric biomaterials. II. Hydrophilicity and penneability to water- soluble low-molecular-weight cornpounds. Biomaterials 1990; 1 1 :3 97-402.
Lu X, Zhai M, and Li J. Radiation preparation and thermo-responsive swelling of interpenetrating polymer network hydrogel composed of PNTPAAm and PMMA. Radiation Physics and Chemistry 2000;57(3-6):477-80.
Mark JE. ; Mark JE, Erman B, Eirich FR, editors.Science and Technology of Rubber. 2nd ed. San Diego: Academic Press, 1994; 4, The Molecular Basis of Rubberlike Elasticity .
McGarey B, McLenaghan DW, Richards RW. Guest Network Co-continuity in Interpenetrating Polymer Networks: A Diffision Study. British Polymer Journal l989;2 1 :227-32.
Miyata T, Asami N, Uragami T. A reversibly antigen-responsive hydrogel. Nature lW9;399(673 8):766-9.
Mohanty KK, Ottino JM, Davis HT. Reaction and Transport in Disordered Composite Media: Introduction of Percolation Concepts. Chemical Engineering Science 1982;37(6):905-24.
Mukae K, Bae YW, Okano T, Kim SW. Polymer Journal 1990;22:206-l7.
Muratore LM, Davis TP. Self-reinforcing hydrogels comprised of hydrophobic methyl methacrylate macromers copolymerized with N,N-dirnethylacrylarnide. Journal of Polymer Science Part A - Polyrner Chernistry 2OOO;38(5):8 10-7.
Murayama S, Kuroda S, and Osawa Z. Hydrophobic and hydrophilic interpenetrating polymer networks composed of polystyrene and poly(2-hydroxyethyl methacrylate): 1. PS-PHEMA sequential IPNs synthesized in the presence of a common solvent. Polymer 1993;34(l3):3845-52.
Murayama S, Kuroda S, and Osawa 2. Hydrophobic and hydrophilic interpenetrating polymer networks (IPNs) composed of polystyrene and poly(2-hydroxyethyl methacrylate): 2. Gradient composition in the IPN's synthesized by photopolymerization. Polymer 1993;34(18):3893-8.
Nakazawa Y, Kamijo Y, Fujimotto K, Kawaguchi H, Yuguchi Y, Urakawa H, Kajiwara K. Preparation and Structural Characteristics of Stimuli-Responsive Hydrogel Microspheres. Angewandte Makrornolekulare Chemie 1996;240: 1 87-96.
Nishi S, Kotaka T, Cornplex-formiag Poly(oxyethylene):Poly(acrylic acid) Interpenetrating Polyrner Networks. 1. Preparation, Structure, and Viscoelastic Properties. MacromolecuIes 1985; 18(8): 15 19-25.
Nishi S, Kotaka T. Complex-forming Poly(oxyethylene)/Poly(acrylic acid) Interpenetrating Polymer Networks. 2. Function as a Chemical Valve. Macromolecules 1986; 19:978-84.
Park JW, Kim SC. ; Kim SC, Sperling LH, editors.IPNs Around the World. John Wiley & Sons Ltd. 1997; 2, Phase Separation During Synthesis of PolyetherimideBEpoxy Semi-PNs. p. 27-46.
Peng T, Cheng Y. Temperature responsive permeability of porous PNIPAAm-g-PE membranes. Journal of Applied Polymer Science 1998;70(11):2 133-42.
Peng T and Cheng Y. PNIPAAril and PMAA CO-grafied porous PE membranes: Living radical CO-grafting mechanism and multi-stimuli responsive permeability. Polymer 200 1 ;42(5):2091- 100.
Podual K, Doyle FJ, Peppas NA. Preparation and dynamic response of cationic copolymer hydrogels containhg glucose oxidase. Polymer 2OOO;4 l(1 l):3975-83.
Ruckenstein E and Liang L. Poly(acry1ic acid)-Poly(viny1 alcohol) Semi- and Interpenetrating Polymer Network Pervaporation Membranes. Journal of Applied Polymer Science 1996;62973-87.
Sawahata K, Hara My Yasunaga H, Osada Y. Electrically Controlled Drug Delivery System Using Polyelectrolyte Gels. Journal of Controlled Release 1990; 14:253-62.
Schirrer R, Thepin P, Torres G. Water absorption, swelling, rupture and salt release in salt-silicone rubber compounds. Journal of Materials Science 1992;27:3424-34.
Schwendeman SP, Amidon GL, Meyerhoff ME, Levy RJ. Modulated Drug Release Using Iontophoresis through Heterogeneous Cation-Exchange Membranes: Membrane Preparation and Influence of Resin Cross-Linkage. Macromolecules l992;25:253 1-40.
Senshu K, Funizono T, Koshizaki N, Yamashita S, Matsumoto T, Kishicla A, and Akashi M. Novel Functional Polymers: Poly(dimethylsi1oxane)-Polyamide Multiblock Copolymer. 8.Surface Studies of Aramid-Silicone Resin by Means of XPS, Static SIMS, and TEM. Macromolecules 1997;30(15):442 1-8.
Seul M, Andelman D. Domain Shapes and Patterns: The Phenomenology of Modulated Phases. Science 1 995;267:476-83.
Shatayeva LK, Samsonov GV, Vacik J, Kopecek J, Kalal J. Penneability of heterogeneous membranes based on methacrylic acid. Journal of Applied Polymer Science l979;23 :2245-5 1.
Siegel RA, Langer R. Controlled Release of Polypeptides and other Macromolecules. Phannaceutical Research 1984; 1 :2-10.
Siegel RA. ; Rosoff M, editors.Controlled Release of h g s : Polymers and Aggregate - - -
Systems. New York: VCH, 1989; 1, Modeling of Drug Release fiom Porous Polymers. p. 1-51.
Sperling LH. ; Klempner Dy Sperling LH, Utracki LA, editors.Interpenetrating Polymer Networks. Advances in Chemistry Series, 1994; 1, Interpenetrating Polymer Networks: An Overview. p. 1-43.
Sperling LH, Mishra V. ; Kim SC, Sperling LH, editors.IPNs Around the World. John Wiley & Sons Ltd. 1 997; 1, The Current Status Of Interpenetrating Polymer Networks. p. 1-25.
Sung C, Sobarzo MR, Merrill EW. Synthesis and Characterization of Polymer Networks made fiom poly(ethy1ene oxide) and polysiloxane. Polymer 1990;3 1556-63.
Suthar B, Xiao HX, Klempner D, et al.; Kim SC, Sperling LH, editors.IPN1s Around the World. John Wiley & Sons Ltd. 1997; 3, A Review of Kinetic Studies on the Formation of Interpenetrating Polymer Networks. p. 49-73.
Tran-Cong Q, Kawai J, Endoh K. Modes selection in polymer mixtures undergoing phase separation by photochernical reaction. Chaos 1999;9(2):298-307.
Tran-Cong Q, Kawai J, Nishikawa Y, Jinnai H. Phase separation with multiple length scales in polymer mixtures induced by autocatalytic reactions. Physical Review E l999;60(2):Rll5O-rll53.
Tran-Cong Q, Okinaka J. Polymer blends with spatially graded structures prepared by phase separation under a temperature gradient. Polymer Engineering and Science 1999;39(2):365-70.
Treloar LRG. The Physics of Rubber Elasticity. London: Oxford University Press, 1975;
Tomer R, Dimitrijevic Dy Florence AT. Electrically controlled release of macromolecules fiom cross-linked hyaluronic acid hydrogels. Jomal of Controlled Release l995;33:405-13.
Turnbull Dy Cohen MH. Free volume mode1 of the amorphous phase: Glass Transition. Journal of Chemical Physics 196 1 ;34: 120-5.
Turner JS and Cheng Y. Heterogeneous polyelectrolyte gels as stimuli-responsive membranes. Journal of Membrane Science 1 998; 148207022.
Turner JS and Cheng Y. Preparation of P D M S - P W IPNs using the Monomer- Immersion Method. Macromolecules 2000;33(10):3714-8.
Ulman KL, Larson KR, Lee CL, Tojo K. Dmg Permeability of Modified Silicone Polymers. II. Silicone-ûrganic Gr& Copolymers. Jomal of Controlled Release 1989; 1 O:261-72.
Vasheghani-Farahani E, Vera JH, Cooper DG, Weber ME. Swellhg of Ionic Gels in Electrolyte Solutions. Industrial Chernical Engineering Research 1 990;29:554-60.
van Aarsten JJ. European Polymer Journal l970;6:919
Weiss AM, Grodzinsky AJ, Yarmush ML. Chemically and Electrically Controlled Membranes: Size Specific Transport of Fluorescent Solutes through PMAA Membranes. AIChE Symposium Series 1986;82(250):85-98.
Widmaier JM. Microphase Separation during the Concurrent Formation of 2 Polymer Networks. Macromolecular Symposia 1995;93 : 179-86.
Yang Y, Fujiwara H, Chiba T, Inoue T. Morphology development in a thermoset/thermoplastic blend: DAP/PPE systern via apparent two-step spinodal decomposition. Polymer 1 W8;3 9(13):274S-SO.
Yasuda H, Peterlin A, Colton CK, Smith KA, Merrill EW. Permeability of Solutes through Hydrated Polymer Membranes. Die Makromolekulare Chemie 1969; 126: 177-86.
Yasuda H, Lamaze CE. Penneselectivity of Solutes in Homogeneous Water-Swollen Polymer Membranes. Journal of Macromolecular Science --B Physics 1 971;5(l): 1 1 1-34.
Yao KD, Sun S. pH Stimulus Response of Complex-foming Poly[(ethylene glycol- CO-propylene glycol)-g-acrylamide]:cr-Poly(acry1ic acid). Polymer International l993;32: 19-22
Yeo JK, Sperlhg LW, Thomas DA. Polymer 1983;24:307
Yoshida R, Sakai K, Okano T, Sakurai Y. Pulsatile h g delivery systems using hydrogels. Advanced h g Delivery Reviews 1993; 1 1 (l-2):85-108.
Yoshida R, Uchida K, Kaneko Y, Sakai K, Kikuchi A, Sakurai Y, Okano T. Comb- type grafted hydrogels with rapid CO-swelling response to temperature changes. Nature l995;347:240-2.
Bang J, Peppas NA. Synthesis and Characterization of pH- and Temperature- Sensitive Poly(methacry1ic acid)/Poly(N-isopropylacrylamide) Interpenetrating Polymeric Networks. MacromoIecules 2000;33(1): 102-7.
Zhang XZ, Zhuo RX. Novel synthesis of temperature-sensitive poly(N- isopropylacrylamide) hydrogel with fast deswelling rate. European Polymer Journal 2000;36(3):643-5.
Zhang D, Gracias DH, Ward R., Gauckler M, Tian Y, Shen YR, and Somorjai GA. Surface Studies of Polymer Blends by Sum Frequency Vibrational Spectroscopy, Atomic
Force Microscopy, and Contact Angle Goniometry. Joufllal of Physical Chemistry B 1 W8;l O2(32):6225-30.
APPENDM A: Surface Analysis of IPN Membranes
Transmission Electron Microscopy (TEM), Field Emission - Scanning Electron
Microscopy (FE-SEM), Tapping Mode - Atomic Force Microscopy (TM-AFM) and
Electron Spectroscopy for Chernical Analysis (ESCA) were also used in order to gather
more information regard'mg the surface morphology of IPN membranes prepared using
different substrates (N2, glass or monomer). The methods were not successful in
determinhg the composition of the IPN surface and were not used M e r . This appendix
details the problems encountered in using these methods to analyze the surface of the
PMAA-PDMS IPN membranes.
The main disadvantages of the surface analysis methods listed above included:
1. Sample materials analyzed must be dry. ln' this work the main focus was the
morphology of the membrane in the hydrated state at different pH conditions in order
to understand the mechanisms of permeation control of the different membrane
morphology. It would be expected that once the materials are dehydrated for analysis
the surface morphology would change significantly and that the ratio of hydrophobic
PDMS to hydrophilic PMAA would increase due to the change in interfacial energy
between water-IPN and air-IPN. With LSCM, IPN membranes were analyzed at
different pH conditions in the hydrated state and in contact with water .
2. Only the swface of the membrane (FE-SEM, AFM), near surface (ESCA) or a very
thin Iayer of the membrane ( ~ 5 0 nm, TEM) can be analyzed at one tirne. For LSCM,
images of 1 um slices up to 100 um into the interior of the sample could be made with
minimal smple preparation or sample destruction, so that an overall picture of
morphology as a hc t ion of depth could be established.
3. Sample preparation for TEM, FE-SEM, and ESCA was a very important aspect of
analysis. The required experience and specialized equipment necessary for
elastomeric polymer samples were not always available at laboratories used in this
work. Because the resolution of these methods was very hi& artifacts introduced
into the sample during preparation greatly affected the results of the surface analysis.
Transmission Electron Microscopy
TEM was carried out using the Hitachi 7000 TEM in Room 1239 of the Medical
Sciences Building at the University of Toronto by Mr. Battista Calvalieri. TEM is the
most commonly used method to investigate IPN morphology. TEM examines a materials
structure by passing a beam of electrons through the sample. In order for the electrons to
pass through the sample it must be very thin, approximately 30 - 70 nrn. The thin
sections of sample are stained with electron absorbing heavy metal salts (in our case
osmium tetroxide) to provide contrast and reveal details of the materials structure. The
Mages produced by the transmitting elecbrons are displayed on a phosphorescent screen
which is then photographed. TEMs are capable of resolving structures only 0.2 nm apart.
Thin sections of the sarnple are usually prepared by first embedding the sample in
resin and then using an ultrarnicrotome to slice the thin sections. This particular step is
critical in sample preparation because artifacts cm be easily introduced during sectioning
of the sample. Since PMAA-PDMS IPNs are primarily elastomcric, it was very difficult
to sectioo thin slices ffom the sample. Either the sample pulled away or sections which
were cut did not produce a srnooth surface. Ideally elastomeric samples such as PMAA-
PDMS IPNs should be sectioned using a cryostat ultramicrotome, where the polymer
sample is fiozen and sliced below its Tg. This intstrument (approx. cost $45,000) was
not available at the laboratory used. This method of surface analysis was abandoned.
Field Emission - Scanning Electron Microscopy
FE-SEM was carried out usbg the Hitachi S4500 FE-SEM in the Metallurgy and
Materials Science Department of the University of Toronto by Mr. Fred Neub. FE-SEM
has the ability to resolve structures in the nm length scale. A special field ernission
electron source is projected onto the sample surface. Secondary electrons emitted fiom
the sample are detected using a cathode ray tube which displays an image of the variation
of secondary electron intensity as a fùnction of x-y position on the sample resulting in a
topographical image of the sample. Insulating samples such as polymer samples must be
either coated with a conductive film such as gold, or imaged at very low keV.
The IPN material was h t analyzed using low voltages in the 2-3 keV range and
the material ww destroyed by the electron field. Attempts to coat the sample with gold
was not carried out due to the inexperience of the laboratory in working and preparing
such samples. When imaging nanometer sized features on coated surfaces, great care
must be taken to distinguish surface topography fiom those caused or accentuated by the
conductive coating since the coating microstructure is easily resolvable using FE-SEM.
Analysis of swface morphology using this method was not explored M e r .
Tapping Mode - Atomic Force Microscopy
TM-AFM was canied out using the Nanoscope III A Multimode SPM in room
402 C-D, Roseburgh Building, under the supervision of Professor Yip. The AFM
~owalewskil998] directly measures ?he XYZ profiles of surfaces by monitoring the
vertical motion of a microcantilever with an ultra-sharp tip which interacts with the
sample through intermolecular and surface forces. In tapping mode the cantilever is
oscillated at high fiequency and its oscillation amplitude is monitored as a function of
distance fkom the surface. For deformable surfaces the cantilever amplitude decreases
more slowly than with rigid surfaces. Thus cantilever motion in tapping mode can give a
quantitative characterization of sample rigidity, and should be able to distinguish between
the hard PMAA domains of the IPN and the sofi elastomeric PDMS regions. Sample
preparation for this method was minimal and involved dehydrating the sample using a
laboratory dessicant.
However, inconsistent results were obtained. Depending upon the oscillation
amplitude of the cantilever the same sarnple either had no PMAA domains at the d a c e or
some domains were present. This reswlt could be atûibuted to the high deformation of the
PDMS component which may allow the subsurface morphology to be detected by thk
method. However, it was not understood whether a layer of PDMS existed on the surface or
whether the oscillation amplitude needed to be increased to detect PMAA which existed on
the surface because dehydrated PMtt4 domains had formed pits in the surface of the IPN.
These results were similar to those achieved by Kowalewski ~owalewski 1 9981 where
PDMS films filled with calcium carbonate particulate were examined. At certain
amplitudes the surface appeared featureless, at larger amplitudes particulate features
appeared. This method of anaiysis was not investigated m e r .
Etectroli Spectroscopy for Chernical Analysis
For ESCA samples are irradiated with x-rays which cause the emission of
photoelectrons fkom the surface of the sample. The electron bindng energies of the
emitted photoelectrons are measured by a hi& resolution electron spectrorneter and are
used to identie the elements and valence states or chemical bonding environment of the
elements present on the surface. The depth of analysis is typically the first 3 nm of the
sample. It is dependent on the escape depth of the photoelectrons and the angle of the
sample plane relative to the spectrophotometer.
Sample preparation for ESCA requires that the sample be completely dehydrated
since the samples are investigated under high vacuum and moisture could damage the
equipment. The samples were lyophilized fiom the hydrated state in order that the
analysis reflect the morphology present in the hydrated membranes. Lyophilization fiom
the hydrated state resulted in samples with burnpy, uneven surfaces, in some cases the
surface was peeling away in layers. Not suprisingly, results in terms of PMAA content
at the surface did not follow suiy psirticular trend with regards to the method of
preparation used for each IPN membrane. Most iikely, the buk of the IPN had become
exposed and contributed to the PMAA content detected by the instrument. This method
was not used m e r .
163
APPENDIX B: Detadeci description of the preparation of PMAA-
PDMS IPN membranes using the monomer immersion method.
This appendix describes in greater detail the procedure used to prepare PMAA-PDMS
IPN membranes using the monomer immersion method. The procedure involves four
steps:
Swelling the PDMS film with a solution of monomer, crosslinking agent and
photoinitiator to form the pre-IPN film.
Placing the pre-IPN film in a glove box and purging the film with N2.
Imrnersing the pre-IPN film into MAA monomer only.
Placing the pre-IPN film swounded by MAA monomer in a W reactor where
polymerization and crosslinking reactions take place.
1. A circular 20 mm diarneter PDMS network is suspended within a 20 ml; glass via1
containing 18 mL of methacrylic acid monomer along with 0.18 g (or 1 % by weight)
of 2,2 - dimethoxyacetophenone, the UV sensitive fiee radical initiator and 72 pL -
720 pL (or 0.04% to 4% v/w) of triethylene glycol dimethacrylate, the crosslinking
agent, for approximately 18 h or until the diffusion of monomer into the PDMS
network comes to equilibrium. At equilibrium the PDMS network has increased it's
mass by approximately 100% due to the imbibed monomer solution.
2. A rubber gasket is used to replace the screw cap in the glas vial opening. Two
syringe needles are used to pierce the gasket. One of the needles is Mmersed into the
swounding monomer solution to act as the N2 inlet and the other needle is placed in
the air space just above the monomer solution, and acts as a vent or outlet. The
monomer solution with the swollen PDMS network is purged with Nz for 20 minutes,
at which point the vial with swollen PDMS is placed in an inert glove bag, filled with
N2.
3. A clean glas scintillation vial containing a section of upright glass in the middle of
the vial attached to the inside, bottom surface of the glass vial using an adhesive such
as Crazy Glue is filled with 18 mL of methacrylic acid monorner only. The glass
insert is used to maintain the pre-IPN film in a vertical position. The monomer Iiquid
is purged with N2 using the same method as described above for a total of 20 minutes.
During the purge, the monomer is heated to a temperature of 50 C. This vial is also
placed in the inert glove bag and the pre-IPN film is quickly transferred into the
monomer only containing vial.
4. This system is imrnediately capped, and placed under UV lights having an intensity of
32 W and a wavelength of 350 nrn and an operating temperature of 50 C. The film is
placed in this UV reactor for a one hour period. Afier an hour, the resultant IPN is a
very hard, tough crystalline material and is surrounded by a waxy solid of the
polyrnerized monomer which s~~rounds the IPN. The vial is broken open, the
surrounding waxy polymer is removed and the IPN is washed extensively in distilled
water to remove any unreacted components and to swell the resultant IPN.
APPENDIX C: Mass Balance of MAA monomer during formation of
PM1AA-PDMS PPN with Air-PPN Interface.
The air-IPN preparation procedure involved placing a film of PDMS that has been
pre-equilibrated with MMi, D M . and TEGDMA in a scultialltion vial, then irradiating
the film with UV. Evaporation of would occur within the vial, possible saturathg
the gaseous space. Upon UV irradiation, and thus depletion of monomer within the
PDMS disk due to polymerization, a concentration gradient would resuIt that would drive
the re-absorption of MAA fiom the gaseous phase back into the solid film. This
sequence of events could result in a PMAA rich region near the surface. To support this
speculation, the mass of M h i in the saturated vapor phase was calculated and correlated
to the amount of MAA seen in the intensely fluorescent layer.
Mass Balance of MAA monomer
Vapour Pressure of MiU monomer at 25 O C = 1 mm Hg or 1/760 atm T = 298 OK v = 0.02 L -
R = 0.082 1 L-atm/mol-K
M.W. of MA4 = 86.09 g/mol Mass = 9.3 x IO-' g
The swollen gel film formed in the Brst 5 p layer has a diameter o f 30 mm. Volume of swollen gel = x (15 mm)2 x O.OO5mm = 3.53 mm3.
We will assume that the swol1en gel formed fkom the pre-IPN with 50% MAA monomer to produce IPN with 30% PMAA on dry basis and that hydration = 0.84.
Also, that the mass is mainly water and density of gel is 1 &m3. Mass of gel = 3.53 x lV3 g. Mass of water in gel = 2.97 x 10') g Mass of PMAA-PDMS polyrner = 5.6 x 104 g Mass of PMAA = 1.6 x 104 P. Therefore, it is reasonable to speculate that the MAA
vapour could be a source for the formation of gel layer at the air-IPN surface.
