Thermo

Download as pdf or txt
Download as pdf or txt
You are on page 1of 219

Thermomechanics of Nano-Filled Elastomers

by
Shawna M. Liff
Master of Science in Mechanical Engineering
Massachusetts Institute of Technology, 2005
Bachelor of Science in Mechanical Engineering
Northeastern University, 2003
SUBMITTED TO THE DEPARTMENT OF MECHANICAL ENGINEERING IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY IN MECHANICAL ENGINEERING
AT THE
A

A CC A OTU

TIVQTTQ ITTITT

ET OF TF

H1FnT

OLGV
MASSACHUSETS INSTITTE
OF TECHNOLOGY

AUGUST 2008

DEC 07 2008
2008 Shawna M. Liff. All rights reserved.

LIBRARIES

The author hereby grants to MIT permission to reproduce and to distribute publicly paper and
electronic copies of this thesis document in whole or in part in any medium now known or
hereafter created.

Signature of Author
Department of Mechanical Enginee /g
August 13, 008

Certified by
Garet5lH. Mc
ing
Professor of Mechanical Engi
Thesis Supervisor
Accepted by
Lallit Anand
Professor of Mechanical Engineering
Chair, Committee on Graduate Students

ARCHIVES

Thermomechanics of Nano-Filled Elastomers


by
Shawna M. Liff
Submitted to the Department of Mechanical Engineering
on August 13, 2008 in partial fulfillment of the
requirements for the Degree of
Doctor of Philosophy in Mechanical Engineering

Abstract
The incorporation of nanoparticles into engineering thermoplastic elastomers affords engineers
an opportunity to formulate flexible, tough and multifunctional polymer nanocomposites that
potentially rival the most advanced materials in nature. Development of these materials is
difficult since thermodynamic and kinetic barriers inhibit the dispersal of inorganic, hydrophilic
nanoparticles into inherently hydrophobic polymer matrices. Thermoplastic polyurethanes
(TPUs) are particularly attractive nanocomposite matrix materials due to their vast range of
potential applications (e.g. in artificial organs, coatings, foams, and active wear), their
mechanical versatility, and tunable block-polymeric structure. In this thesis we explore methods
for systematically nanoreinforcing such materials by exploiting the microphase structure,
differential polarities and multiple thermomechanical phase transitions of the macromolecular
blocks that constitute the elastomeric matrix.
Using a solvent exchange technique we show that it is possible to preferentially nanoreinforce
the hard micro-domains of thermoplastic elastomers with smectic clay nanofillers that have
characteristic dimensions similar to the hard segment. The adhesion between the clay and the
hard micro-domains coupled with the formation of a percolative network not only stiffens and
toughens, but increases the heat distortion temperature (HDT) of the material. The discotic clay
platelets induce morphological ordering over a range of length scales that results in significant
thermomechanical enhancement and expands high temperature applications. This thesis seeks to
further enhance the understanding and utility of thermoplastic polyurethane nanocomposites by
answering two questions: (1) what thermo-physical interactions between nano-clay and
elastomeric thermoplastic polyurethane are taking place? and (2) how can these thermo-physical
interactions be exploited?
To answer these questions the nano-reinforced-hard micro-domain morphology was monitored
during deformation using in-situ wide angle x-ray scattering and combined with the results of
extensive quasi-static mechanical testing which enabled the identification two characteristic
relaxation times. A one-dimensional constitutive model to account for such morphological
changes augmenting the previous model for unfilled polyurethanes developed by Qi and Boyce
(2005) is discussed. Finally, the thermo-mechanical influence of nano-clay fillers on the shape
memory effects exhibited by polyurethane nanocomposites is examined and multi-responsive
shape memory polyurethane fibrous mats are developed via electrospinning.

Quantifying and controlling the thermo-physical interactions between a block-copolymer with


polar segments (e.g. thermoplastic polyurethane) and inorganic nanoparticles (e.g. nano-clay) is
important for future nanocomposite processing strategies: the efficacy of nanoreinforcement
hinges upon the close matching of characteristic length scale and the adhesion of the
nanoparticles to the targeted polymer phase morphology. Exploiting the different polarity of the
blocks in conjunction with solvent exchange approach developed in this thesis and solution
processing techniques such as electro-spinning, offers an avenue toward the development of high
performance, hierarchically-ordered materials that rival natural materials.
Thesis Advisor: Gareth H. McKinley
Title: Professor of Mechanical Engineering

Acknowledgements
This research was supported by the US Army through the Institute for Soldier Nanotechnologies,
under Contract DAAD-19-02-D-0002 with the US Army Research Office. The content does not
necessarily reflect the position of the Government, and no official endorsement should be
inferred. I also want to thank the National Science Foundation for the financial support it
provided in the form of a graduate research fellowship.
I would like to thank Professor Gareth H. McKinley, my advisor, and the rest of my committee,
Professors Mary C. Boyce, Robert E. Cohen, and Paula T. Hammond for sharing their expertise,
for asking thought-provoking questions and for offering support and guidance. Gareth, thank you
for broadening my knowledge of polymer science and non-Newtonian fluids, teaching me to
prepare for and think about the tough questions I think you or someone else will ask, and then
always asking a question I did not consider. Professor Boyce thank you for allowing me to work
with you as a teaching assistant for the undergraduate Mechanics and Materials II course (2.002),
for adopting me as a group member, and continuing to tutor me in polymer mechanics. I am
grateful to have worked with you both. I am also grateful for the friendship and mentorship Dr.
Nitin Kumar and Dr. LaShanda James-Korley afforded me my first two years at MIT. Their
support and knowledge was invaluable. I also want to thank my lab-mates at the Institute for
Soldier Nanotechnologies and at the Non-Newtonian Fluids Laboratory for sharing their
expertise. I am particularly thankful for the help Merideth Silberstein provided with the ELF, the
useful wide angle x-ray diffraction analysis advice Ryan Waletzko offered, the TEM preparation
Dr. Steve Kooi completed with the FIB, the TEM imaging completed by Dr. Mark Johnson, Dr.
Yong Zhang, and Daniel Alcazar, and the Mercury Porosimetry tests completed by Joseph
Lowery. I also want to thank Wonjae Choi, Shreerang Chattre, and Dr. Anish Tuteja for their
assistance with the goniometer and preparation and supply of the oleophobic coating used on the
electrospun mats, as well as Dr. Brian Pate, Emily Chen and Cate Morgan for their data
collection at CHESS, and Patrick Boisvert for his ESEM and SQUID assistance.
Finally, and most importantly I would also like to wholeheartedly acknowledge and thank my
family for their continuing support, patience, and love. Mom, Dad, and Harmony, thank you for
always answering the phone and talking with me no matter if I were ecstatic or a just a stressed
out mess, for supporting me always, and loving me no matter what. Mom for five years I always
looked forward to our phone conversations, especially on Saturday morning when I was in the
lab and even when I wanted to crush the DMA with a sledgehammer. Shannon, you are my
lighthouse. You stand resolute no matter what the conditions and I rely heavily on you to keep
me grounded and to guide me home. I may have been able to do this without you, but I certainly
would not have smiled as big as when I came home at the end of each day. Thanks for your
patience when I set the alarm every two hours through the night so that I could tend to
experiments, for altering our plans last-minute so that I could meet deadlines, and for always
being supportive and positive. Most importantly, thank you for sharing your love and life with
me.

Contents
List of Figures ................. .

..

L ist o f T ab les ...............................

.. .........................

... .....................
............. .......... ....................

...

...... ............................... 15
17

1. Introduction..............................................

.................... 22

1.1 References........................................... ................................. ........ ...

24

2. Background ..........................................................
2.1 Inspiration .............................

.... .... . . ........

... ..... ...... ........................

24

2.2 Introduction to Polyurethanes .........................................

27

2.3 Importance of Polymer-Clay Nanocomposites ...............................

29

2.4 Brief Synopsis of Previous Polyurethane Nanocomposite Research............................ 33


2.5 Summ ary.................

.... .........

49

.....................

.................... 50

2.6 R eferences ................................... .. .......... ..... ......................... ..... ............

3. Nanoparticle Dispersal Via Solvent Exchange .......................................... 55


. . ... 55

3.1 Introduction ...................................................................................................


3.2 Previous Dispersal Techniques Utilized ......................................

.................55

3.2.1 Dispersion Via Melt and Mixing ........................................................ 56


........... 57
...................
3.2.2 In Situ Polymerization ...............................
. ............... 57
....
3.2.3 Common Solvent Process ....................
3.3 Solvent Exchange Approach................................

..................... 58

3.2.1 Method to Disperse Laponite in Thermoplastic Polyurethane ....................... 60


3.4 Proof of Exfoliation ...................................................................

............

.....

..
.. ...................
....
3.4.1 Materials .....................
................
3.4.2 Dispersal Characterization ....................................................
............
..
3.4.2.1 Transmission Electron Microscopy .......................
.................
3.4.2.2 Atomic Force Microscopy ..........................
......................................................
3.4.2.3 Wide Angle X-ray Diffraction
............................
3.4.3 Results and Dicussion .............................
3.5.1 Influence of Segment Polarity.................................................................
. .... ... ... ....................
3.5.1.1 Materials.................
...................
3.5.1.2 Evidence of Exfoliation ...................
3.5.1.3 Evidence of Preferential Reinforcement ..........................................
.................. ...................
3.5.2 Influence of Processing Time ...................
............

...........................

63
64
64
64
64
64
67

3.5 Influence of Processing Parameters ............................................

3.6 Conclusion

63

67
67
68
71
74

.................... 75

3.7 References .............................................................

.....

............ ............................ 76

4. High Performance Elastomeric Nanocomposites via Solvent Exchange Processing ............. 79


4.1 Introduction.................................

.....

....... ....

.....................

4.2 Materials and Characterization Methods ......................


4.2.1
4.2.2
4.2.3
4.2.4

79

.................. 80

M aterials .................................... .
..
... . ....... .................... 80
................... 81
...
..
Atomic Force Microscopy ..................
Mechanical Characterization ..........................
.....................
81
Thermal Characterization .............................
...
......................... 81

4.3 Results and Discussion .........................................


4.4 C onclusion ..........................................
4.5 References .........................................................

.....................

81

. ........................

..................... 95

.......................... ....... .......................... 95

5. Influence of Nano-Clay on Morphology and Mechanics of High Performace Elastomeric


N anocom posites ..................................................

........................ . ............

5.1 Introduction....................................

................ 99
.....................

5.2 Mechanical and Morphological Characterization Techniques ..................................

99
101

101
5.2.1 Mechanical Test Descriptions ........................................
102
........................................
5.2.1.1 Hysteresis .....
102
5.2 .1.2 S oftenin g .......................................................................................................
...........
..................... 102
............. ....
5.2.1.3 Rate D ependence .....................
102
5.2.1.4 Time Dependence ........................................
.......... .... 102
5.2.1.5 Relaxed Response and Equilibrium Paths ............
5.2.2 Morphological Characterization Techniques ............... .........................103
5.2.2.1 Wide Angle X-Ray Scattering .................................................... 103
5.2.2.2 Attenuated Total Reflectence Fourier Transform Infrared Spectroscopy.... 104
5.2.2.3 Cross-Polarized Microscopy ..................................................... 105
5.3 R esults and D iscussion ................................................................................
5.3.1
5.3.2
5.3.3
5.3.4
5.3.5
5.3.6
5.3.7
5.3.8

........... 105

105
. .....................
Hysteresis................. .. . ..... ..
................ 107
Softening................. ......................
.................. 109
Rate Dependence ................ ..............
..... .............
.... ..................... 111
Time Dependence ....................
Relaxed Response and Equilibrium Paths..................................120
122
......... ..................
X-Ray Scattering ......................................
Attenuated Total Reflectence Fourier Transform Infrared Spectroscopy ......... 125
................... 128
Cross-Polarized Microscopy .................. .................

5.4 Conclusions

........ 129

...........................................

5.5 References ..................... .. ................

...

.............

...................... 130

6. Influence of Laponite on the Shape Memory Behavior of Polyurethane .......................... 133


6.1 Introduction

...................... 133

............................

136

6.2 Materials and Experimental Methods ........................................


.. ........ ... ........... ........................................
6.2.1 M aterials .....................
6.2.2 Composite Preparation................................
6.2.3 Characterization Methods ........................................
6.2.3.1 Laponite Dispersal ........................................
6.2.3.2 Thermal Transitions ........................................
6.2.3.3 Thermal Expansion Coefficient ........................................
6.2.3.4 Stress-Strain Behavior Above and Below the Transition Temperature.......
................................
6.2.3.5 Free Strain Recovery......
6.2.3.6 Constrained Strain, Stress Recovery................................

142

6.3 Results and Discussion ...........................................


6.3.1 Laponite Dispersal ...........................................
6.3.2 Thermal Transitions ..........................................
6.3.3 Thermal Expansion Coefficient ..........................................................
6.3.4 Stress-Strain Behavior Above and Below the Transition Temperature..............
........
6.3.5 Free Strain Recovery..........................
6.3.6 Constrained Strain, Stress Recovery................................
6.4 C onclusion .........................................................................

136
136
136
136
137
137
137
140
142

...... ....................

142
143
147
148
154
160
163

6.5 References ..................................... ...... ..... ..................................... 164


7. Development of Multi-Responsive Electrospun Fabric....................
7.1 Introduction............... . ...........

............

................. 166
................

166
169

7.2 Materials and Experimental Methods ........................................


.... ............................. .....................
7.2.1 M aterials .........................
7.2.2 Solution Preparation..............................
7.2.3 Electrospinning Parameters and Mat Preparation ......................................
7.2.4 Characterization Methods ........................................
7.2.4.1 Interrogation of Solution Conductivity ........................................
7.2.4.2 Interrogation of Fiber Morphology ........................................
7.2.4.3 Interrogation of Non-Woven Mat..............................
7.2.4.4 Interrogation of Pore Size ....................................................... .....................

169
169
170
171
171
171
171
172
173

7.3 Results and Discussion ...........................................

........... 173
7.3.1 Fibrous M orphology .......................................
7.3.2 Results of Thermal and Aqueous Actuation ........................................ 176
182
7.3.4 Evidence of Reversible Surface Energy ............................................................
7.4 Conclusions

........ 185

...................
......................

7.5 References ........................................

...........

............... 186

8. Concluding Remarks.........................
8.1 Summary of Accomplishments....
8.2 Future Work ............................

.. ............................................
............. .................

................................

.. ........

190

................................ 190
...... ........ ................ 192

8.2.1 Scale-Up of Time Efficient, Environmentally-Friendly, Solvent Exchange


Approach ........................ ............................................... 192
8.2.2 Deformation Induced Orientation and Damage by Nano-Clay .................. 194
8.2.3 Utility of Magnetite in Shape Memory Polyurethane Composites ................ 196
8.3 References ........................ .. ............. .. ... .......
A ppendices..........................

.....................................

....................
....
.

......

....................................

200
. 201

A-i Movie Demonstrating the Mechanical Enhancement of Elasthane at High Temperature


by the Addition of Laponite ...........................................
201
A-2 Elasthane/Laponite Composite Moduli Measured with a Video Extensometer.......... 202
A-3 Discussion of Appropriate Constitutive Model ........................................................ 206
A-4 2-D WAXS Spectra Collected at Each Specimen Stretch for Pure Elasthane and that
Reinforced with 8 wt% Laponite ........................................
213
A-5 Movies of Electrospun Mat Lateral Shrink .............................

219

List of Figures
Figure 1-1: The aspect ratio of various shaped nano-filer is the ratio of the longest dimension, 1,
to the shortest dimension, d.............................................................. 17
Figure 1-2: Schematic visually describing how the surface area to volume ratio of a discotic
platelet increases when thinner, higher aspect ratio platelets are utilized ........ 18
Figure 1-3: Prediction of nanoparticle, nanocomposite formulation and nanocomposite
component market size by the chemical company BASF for the year 2015 taken
.. 19
from [4] ...................................................................
Figure 2-1: Engineering stress-strain curves of dragline and viscid silk as displayed [2].........25
.................... 27
Figure 2-2: A Urethane Linkage ..................................... .......
Figure 2-3: Thermomechanical behavior of various polyurethanes .................................. 28
Figure 2-4: Schematic illustrations of various polymer-clay nanocomposites with different
extents of clay dispersal .............................................................. 32
Figure 2-5: Comparison of the Montmorillonite and Laponite crystal structure as shown in
.. 32
[27] .......................................
Figure 2-6: Tensile properties of the polyurethane-clay nanocomposites prepared with the
longest onium ion Montmorillonite-modification at various clay loadings as
reported in [28]: (A) tensile strength, (B) tensile modulus, and (C) strain-at...
.......................... 33
break.................................. ................................ .... .. ...
Figure 2-7: Dispersal of 1 wt% Montmorillonite modified with most reactive swelling agent in
the PU containing 39 wt% hard segment as shown in [34] ............................. 35
Figure 2-8: Representative force-extension curves of the polyurethane urea nanocomposites in
.. 36
[36] ..................................... ..............................
Figure 2-9: Storage moduli (left) and stress-strain plots (right) of the polyurethane containing 0
....................37
(a), 4 (b), 20 (c), and 40 (d) wt% clay from [38] ...............
Figure 2-10: Transmission Electron Micrographs of the solvent cast (left) and melt compounded
(right) soft polyurethane nanocomposites containing 3 wt% Montmorillonite
............ 39
.....
depicted in [42]..................................................
Figure 2-11: Stress-strain response of a polyurethane filled with the same concentration of
synthetic fluoromica (3 wt% ) but of different aspect ratio [43] ........................ 40
Figure 2-12: Hysteresis exhibited by a polyurethane filled with the same concentration of
synthetic fluoromica (3 wt% ) but of different aspect ratio [43] ........................ 40
Figure 2-13: Storage modulus versus strain amplitude curves of the polyethylene-polyurethane
(PEPU) nanocomposites containing 0, 2.5, and 5 wt% clay depicted in [46].........41
Figure 2-14: The recovery stress of the shape memory polyurethane with 0, 1, 3 and 5 wt% clay
at a fixed strain of 100% while a 40 C min' heating rate is applied [47].........42
Figure 2-15: Evolution of d-spacing with respect to a nanocomposite loaded with 5 wt% clay
44
.................
during stretching [53].............................
wt%
hard
36
containing
polyurethane
(a)
pure
images
of
phase
Figure 2-16: Tapping mode AFM
segment and (b) the same polyurethane loaded with 1 wt% clay in [14]. ........ 45
Figure 2-17: Mechanical properties of the waterborne PU filled with various weight
concentrations of Saponite as shown in [68] ....................................... 48
Figure 2-18: Transmission electron micrographs of sepiolite (a) and 3 wt% sepiolite in
................ 49
polyurethane (b) [64] ...................................... ..........

Figure 3-1:

Vapor-Liquid equilibrium curve for H20 and for DMAc (left) and vapor-liquid
equilibria of H2 0-DMAc determined via experiment in [33] ......................... 61
Figure 3-2: Structure of Elasthane 80A based on available non-proprietary information .......... 63
Figure 3-3: a, b, & c, Transmission electron micrographs (TEM) of nanocomposites containing
4, 10, and 20 wt% Laponite show fully exfoliated, uniformly dispersed and
randomly oriented, Laponite platelets................................... ...........
65
Figure 3-4: Laponite dispersal within the polyurethane matrix via WAXD, using a different
diffraction instrument and different nanocomposite samples (Dispersal
Characterization). ........................
.. . ....
. ........................66
Figure 3-5: Structure of PEO-based polyurethane [42] .............................
67
Figure 3-6: Structure of PTMO-based polyurethane [44]........................................ 68
Figure 3-7: (a) TEM image of 10 wt % Laponite dispersed in PTMO:HDI-BDO PU; 50 nm
scale bar. (b)TEM image of 10 wt % Laponite dispersed in PEO-PPO-PEO:HDIBDO PU; 50 nm scale bar. (c) TEM image of 10 wt % Laponite dispersed in PEOPPO-PEO:HDI-BDOPU; 20 nm scale bar...........
......
................ 69
Figure 3-8: WAXD spectra of HDI:BDO hard segment, of pure Laponite and of the pure and 10
wt % Laponite-filled polyurethane nanocomposites ..........................
.......70
Figure 3-9: Flexural storage modulus, E' (a), and loss tangent, tan 6 (b), both determined via
DMA and DSC thermograms and (c) during initial heating and cooling cycles of the
pure (black) and 10 wt % Laponite-filled PUs (red) .................................... 72
Figure 3-10: Representative engineering stress-strain tensile curves of the pure and 10 wt %
Laponite filled polyurethanes are presented, and the corresponding cross-polarized
im age............................................................................................
...73
Figure 4-1: Structure of Elasthane 80A and Laponite dispersal within the polyurethane
matrix................... .. ... ...............
..................... 82
Figure 4-2: Hard and soft micro-domain distribution in nanocomposites containing 0, 4, 10, and
20 wt % Laponite as observed via AFM phase imaging. ................................84
Figure 4-3: Impact of Laponite concentration on the mechanical properties of Elasthane.
.................................................................................................. 86
Figure 4-4: Impact of Laponite concentration on the thermomechanical properties of Elasthane..
................................................................................................... 89
Figure 4-5: DSC of the nanocomposites containing 0, 4, 10, and 20 wt% Laponite.
................................................................................................... 91
Figure 4-6: Laponite concentration and impact on thermal degradation. ................................ 92
Figure 4-7: Evidence of long range order or crystallization in the nanocomposites is apparent in
images obtained with a cross-polarized microscope. .................................... 93
Figure 4-8: Polarized optical microscopy images of the ordered, fractal-like domains in various
nanocomposite films exhibit birefringence. ...........................
94
Figure 5-1: Hysteresis of pure Elasthane and composites filled with 4, 10, and 15 wt% Laponite
when thin-film specimens are stretched 80% and 160%................................ 106
Figure 5-2: Comparison of the hysteresis of pure Elasthane and a composite filled with 15 wt%
Laponite when thin-film specimens are stretched 80% and unloaded and then to
160% ...............
............................................
.. 106
Figure 5-3: Cyclic softening behavior of pure Elasthane and nanocomposites filled with 4, 10,
and 15 st% Laponite when the specimens are stretched to nominal strains equal to
0.1, 0.2, 0.4, 0.8, 1.2, 1.6, and 2.0. ......... ......
.....................
107

Figure 5-4: Evolution of the maximum stress and plastic strain as 0, 2, 4, 10, and 15 wt%
Laponite-filled nanocomposites are cycled 5 times at applied nominal strains equal
to 0.1, 0.2, 0.4, 0.8, 1.2, 1.6, and 2.0 .................................................. 108
Figure 5-5: Rate dependence of 0, 4, 10, and 15 wt% Laponite-filled Elasthane when stretched
to a nominal strain of 1.0 at three different nominal strain rates .................. 109
Figure 5-6: The flow stress, or engineering stress as e=0.3, at the various strain rates is plotted
against the Laponite concentration ...................................................... 110
Figure 5-7: Comparison of the engineering stress-strain response of 0, 4, 10, and 15 wt%
Laponite-filled Elasthane when stretched to a a nominal strain of 1.0 at three
different nominal strain rates .......................................................... 111
Figure 5-8: Applied strain history for stress relaxation tests as well as the engineering stress
response vs. time and applied strain for nanocomposites containing 0, 4, 10 and 15
wt% Laponite .............................................................. 112
Figure 5-9: One-dimensional description of the rheological model of the experimental stress
relaxation and corresponding fir to the pure Elasthane stress relaxation after the
113
............
specimen is stretched 10%......................................
the
of
behavior
relaxation
Figure 5-10: Two characteristic time constants observed in the stress
114
........................
polyurethane nanocomposites ...................................
[18]..............114
in
depicted
as
network
Figure 5-11: Schematic of the stretched eight-chain
Figure 5-12: Relationship between the orientation parameter as wel as the normalized orientation
parameter and the principal stretch of an eight-chain RVE for an incompressible
material ............................................................. ... 116
Figure 5-13: Evolution of the two characteristic time constants observed in the stress relaxation
behavior of the polyurethane nanocomposites with orientation ...................... 116
Figure 5-14: Evolution with tensile stretch of the three moduli that capture the stress relaxation
................... 117
.
behavior of the polyurethane nanocomposites .................
Figure 5-15: Power-law decay of the polyurethane nanocomosite moduli with
.............................. 117
orientation ........................................
Figure 5-16: Comparison of the polyurethane nanocmposite moduli evolution with increased
orientation determined from experimental data and that prediced via empirical
relations .............................................................. ... 119
Figure 5-17: Stress relaxation behavior of pure polyurethane and that filled with 4, 10, and 15
wt% Laponite through five load-unload cycles ........................................ 120
Figure 5-18: Equilibrium paths of the 0, 4, 10, and 15 wt% Laponite-filled nanocomposites
determined from the stress relaxation behavior during the 1 st and 5 th
121
..... ..........
cycle............................................
is
material
the
as
Elasthane
Laponite-filled
wt%
8
and
pure
of
Figure 5-19: WAXS spectra
deform ed ....................................... .......................................... 123
Figure 5-20: Azimuthal scan of the pure and 8 wt% Laponite-filled polyurethane composite as
the deformation increases at d-spacing between 0.447 and 0.468 nm............... 124
Figure 5-21: Azimuthal scan of the pure and 8 wt% Laponite-filled polyurethane composite as
the deformation increases at d-spacing between 0.409 and 0.427 nm............. 125
Figure 5-22: ATR-FTIR spectrum of pure Elasthane 80A and that filled with 6 wt%, 10 wt%,
and 15 wt% Laponite ................................................................... .. 127
Figure 5-23: Comparison of the long range order exhibited through cross-polarized microscopy
of pure Elasthane prior to deformation and after deformation to yield.............. 128

Figure 5-24: Comparison of the long range order exhibited through cross-polarized microscopy
of 10 wt% Laponite-filled Elasthane prior to deformation and after significant
stretch and soft-segment reorientation and alignment .............................. 129
Figure 6-1: Images of a pure MM5510 polyurethane film inside the ELF 3200 as seen with the
video extensom eter. ..........................
........................ ................... 139
Figure 6-2: Visual description of the free strain recovery test and the locations with respect to
engineering stress, strain and temperature .......... ................ ........................ 141
Figure 6-3: Transmission electron micrographs of MM5 510 PU nanocomposites containing 4,
10, and 15 wt% Laponite ...............................
......... 143
Figure 6-4: First and second heat flow cycles of pure MM5510 polyurethane and composites
filled with 4, 10, and 15 wt% Laponite using differential scanning
calorimetry..........................................................144
Figure 6-5: Dynamic mechanical response of pure MM5510 polyurethane and composites filled
with 4, 10, and 15 wt% Laponite soon after preparation (left) and nine months later
146
(right) ..................................... ..................................
Figure 6-6: a) Thermal expansion of pure MM5510 polyurethane films and composites filled
with 4, 10, and 15 wt% Laponite from -150 0 C to 125 0 C while under a 10 kPa
tensile load. b) Enlarged view of the thermal expansion behavior within the shape
memory operational temperature range interrogated. c & d) The thermal expansion
coefficients below and above the transition temperature (55'C), respectively, are
elucidated ................................................................... 147
Figure 6-7: True stress-strain response of the MM5510 polyurethane composites at 25'C, a
temperature below the transition temperature (55'C), and 70'C, a temperature
above the transition temperature when the films are stretched 2.5 times their
original length ...................................................
149
Figure 6-8: Uniaxial tensile response of the various nanocomposites at 70C, a temperature
above the transitition temperature........................
........ .................... 150
nanocomposites
at
70'C and the
Figure 6-9: Uniaxial tensile response of the various
corresponding response at 25'C after five cycles of deformation at the elevated
temperature ...................................................
151
0
Figure 6-10: Uniaxial tensile response of the various nanocomposites at 25 C, a temperature
153
below the transition temperature ....................................
Figure 6-11: Stretch recovery of the nanocomposites after stretching the specimen to 1.1, 1.2,
1.4, 1.7, or 2.0 times its original length at 700 C using the TA Instruments Q800
DMA .....................................................
.................................
154
Figure 6-12: Comparison of the stretch recovery of the nanocomposites after stretching the
specimens to 1.1, 1.2, 1.4, 1.7, or 2.0 times their original length at 70'C using the
TA Instruments Q800 DMA .......... ..........
.................................... 155
Figure 6-13: Comparison of the extent of strain recovery, total strain recovery and strain fixity
for the polyurethane nanocomposites determined after deformation at elevated
temperature .....................................
.
.........................
.. 156
Figure 6-14: Comparison of the strain recovery of the nanocomposites after stretching the
specimens to 1.4, 1.7, or 2.5 times their original length at 25'C using the ELF
3200.....................................158

Figure 6-15: The strain recovery of the polyurethane nanocomposites containing 0, 4, 10, and 15
wt% nano-clay after stretching the speciments to 1.4, 1.7, or 2.5 times their original
length at 25 0 C using the ELF 3200 ..................................................... 159
Figure 6-16: Comparison of the stress recovered when the strain is constrained after stretching
the polyurethane nanocomposite specimens to 1.4, 1.7, or 2.5 times their original
length at 25'C using the ELF 3200 .................................... .................. 160
Figure 6-17: The stress recovered when the strain is constrained after stretching each
polyurethane nanocomposite specimen to 1.4, 1.7, or 2.0 times their original length
at 25C using the ELF 3200.......................................... .......... ........ 162
Figure 7-1: Schematic representation of the parallel-plate fiber spinner .......................... 170
Figure 7-2: Scanning electron micrographs demonstrate the change in fibrous morphology as
theconcentration of polyurethane in solution increases ............................... 173
Figure 7-3: Complex plane impedance diagram of the 17 wt% polyurethane solution with (b)
and without Laponite (a) from which the solution conductivity is calculated....... 175
Figure 7-4: Microscopic images of the as spun, pure, non-woven mat prior to and after
actuation and of a 4 wt% Laponite-filled mat .................. ...................... 175
Figure 7-5: Differential scanning calorimetry of the as spun non-woven mat, the thermally
actuated mat, and the water soaked mat prior to drying and after drying......... 177
Figure 7-6: Evolution of water uptake in the electrospun mat as it is soaked in a DI H20
...178
.
bath ............................................................
Figure 7-7: Influence of polyurethane concentration in solution on fiber diameter (a) and the
extent of strain recovered after actuation (b & c) ...................................... 179
Figure 7-8: Flexural storage moduli of the BOAS electrospun mat at various stages of
.... 180
.................................
actuation ................
Figure 7-9: Resultant change in pore size distribution within the pure and Laponite composite,
'beads on a string' morphology, non-woven mats prior to and after thermal and
aqueous actuation ................................................................ 181
Figure 7-10: Scanning electron micrographs of the pure, non-woven, electrospun mat coated
with a 50/50 blend of fluorinated POSS & PMMA in the relaxed, free-standing, as
spun configuration, in the thermally actuated and shrunk state, the re-stretched state,
and the again shrunk or recovered state and the corresponding static contact angle
of a drop of hexadecane ...................................... .. ....................... 183
Figure 7-11: Comparison of the top and bottom side of an electrospun polyurethane mat filled
w ith 10 w t% magnetite....................................................................185
Figure 8-1: Stress-strain curves of poly(etherurethane urea) (PEUU-A) and that filled with 5
wt% vitamin E (PEUU-VE) [2] ................................... .................. .193
Figure 8-2: Comparison of the viscosity of DI-a-tocopherol (Vitamin E) as a function of
temperature when constant shear stress or constant shear rate is applied......... 194
Figure 8-3: Transmission electron micrographs of magnetite nanoparticles dispersed in a
commercial thermoplastic polyurethane (MM 5510) after precipitation and
compression molding: scale bar 500nm in a and 100 nm in b, c, and d.............198
Figure 8-4: Magnetization of commercial thermoplastic polyurethane DiaplexTM MM 5510
filled with 5 wt% magnetite (Fe 30 4) as a function of applied magnetic field (left)
and temperature when no magnetic field is applied (right) ............................ 199
Figure A-2-1: Images of 2 wt% Laponite-filled polyurethane film inside the Zwick ZO 10 as seen
203
...........................
with the video extensometer ...................................

Figure A-2-2: Selection of the sequential images taken with the video extensometer as the
Elasthane filled with 2 wt% nano-clay is deformed using the Zwick ZO10
mechanical tester at a nominal strain rate of 0.0167 s 1 .......... .. . . . . . . . . . . . . . . . . 204
Figure A-2-3: Comparison of the engineering stress-strain response of 2 wt% nano-clay filled
Elasthane calculated from the change in length reported by the Zwick mechanical
tester crosshead and that measured via the video extensometer at various sections
along the length of the specimen gauge length ......................
............ 205
Figure A-3-1: One dimensional rheological model and material parameters summary of the Qi
and Boyce thermoplastic polyurethane constitutive model as shown in [1].........207
Figure A-3-2: Cartoon of the soft segment evolution with matrix stretch ........................ 208
Figure A-3-3: Eight chain rubber elasticity model developed by Arruda and Boyce.............209
Figure A-3-4: The equilibrium or hyperelastic response of the Elasthane/Laponite
nanocomposites is the same despite Laponite concentration and can be captured by
the constitutive model of Qi and Boyce [1, 5] ................... .................. 210
Figure A-3-5: Comparison of the equivalent shear stress versus shear strain for pure Elasthane in
tension to that of the polyurethane evaluated by Qi and Boyce in compression
[1]............................................
.....................
211
Figure A-4-1: The specimen stretches equal to 1.1, 1.6, 3.0 and 9.0 correspond to the locations
shown above in the stress-strain curve due to specimen slip during testing.........213
Figure A-4-2: Enlarged scatter patterns at specimen stretches equal to 1.1, 1.6, 3.0 and 9.0 for
pure Elasthane ...................
...............
......................................... 214
Figure A-4-3: Enlarged scatter patterns at specimen stretches equal to 1.1, 1.6, 3.0 and 9.0 for
the 8 wt% Laponite-filled Elasthane nanocomposite.............
...................... 215
Figure A-4-4: 1-D SAXS spectra of pure Elasthane and that filled with 8 wt% Laponite.......218

List of Tables
List and comparision of the various clays investigated for polyurethane
reinforcem ent [7-11] ..................... .............. ..... ...................... 19
Table 2-1: Comparison of the mechanical properties of silk to other high-performance
. .............................. 25
materials [2]..............................
and water absorption of the pure
elongation,
tensile
strength,
weight,
molecular
2-2:
The
Table
polyurethane and polyurethane-Montmorillonite nanocomposites prepared in
..................... 34
[32]... ................ ................. ............................... ...........
Table 2-3: List and comparison of the various clays investigated for polyurethane
47
reinforcement [19, 61-64] .......................................
parameter
the
solubility
of
Table 3-1: Polyurethane soft segment polarity expressed in terms
which was calculated using the group contribution method and values given in
71
...................
..........................
[45]........
Table 5-1: Parameters describing the power-law decay of the polyurethane nanocomposite
moduli with increased orientation ........................................ ................ 118
Table 5-2: ATR-FTIR spectral assignments for Elasthane determined by Christenson et al in
................. 126
....... . ...............
[28] ........................
Table 5-3: Ratio of area under the absorbance peaks of the hydrogen-bonded carbonyl bonds
and the free carbonyl bonds as well as the ratio of these areas to the total
absorbance area constituting carbonyl bonds described by [30] .................. 127
Table 6-1: Analysis of the glass transition temperatures observed between O'C and 100 0 C via
dynamic mechanical analysis and differential scanning calorimetry upon heating
freshly prepared and aged samples ...................................... .................. 146
Table 6-2: Predicted thermal stress of the various polyurethane nanocomposites when the
................... .160
length is constrained ...........................................
as seen via the
nanocomposite
each
in
achieved
stretch
true
Table 6-3: Comparison of the
video extensometer during the constrained strain, stress recovery test when the
specimen was stretched to 1.4, 1.7, or 2.5 its original length ......................... 161
Table 7-1: Electrospinning parameters for each fibrous morphology discussed ............. 170
Table A-3- 1:Five parameters needed to describe the hyperelastic or equilibrium response of the
Elasthane/Laponite nanocomposites ..................................................... 209

Table 1-1:

1. Introduction
Stronger, stiffer polymeric materials are attractive to numerous industries due to their light
weight, easy manufacturing, low-cost, and recyclability [1]. Reinforcement of polymer matrices
with inorganic nanoparticles produces composite materials that exhibit enhanced mechanical,
thermomechanical and barrier properties without significantly increasing the weight of the
material. Excellent stiffness and strength are achieved while utilizing far less high-density
inorganic material than is utilized in conventional composites filled with micron or millimeter
size particles. Whether the nano-filler is cylindrical (carbon nanotubes), spherical (carbon black),
or a platelet (clay) the efficiency of the inorganic nano-reinforcement depends on three
parameters: filler mechanical properties, filler aspect ratio, and adhesion between the matrix and
the filler [2]. The filler aspect ratio, a, is given by the ratio of the longest dimension, 1, to the
shortest dimension, d. The aspect ratio of various shaped nano-filler is visually compared in
Figure 1-1.

Fiber

Platelet

Figure 1-1: The aspect ratio of various shaped nano-filer is the ratio of the longest dimension, 1, to the
shortest dimension, d. The aspect ratio of spherical particles like carbon black is equal to 1 while that of the
clay platelets can range from 25 to over 200.

The aspect ratio of the nanoparticles is important, because as shown for the discotic platelets of
constant diameter I in Figure 1-2, as the aspect ratio increases the surface-area (As) to volume (V)
ratio increases. Consequently, when the surface area to volume ratio is high, the nanoparticles
occupy a small volume and do not significantly increase the weight of the material, but afford the
polymer a large, more rigid, area on to which stress can be transferred.

!I

J
d
d

=d

A_

Figure 1-2: Schematic visually describing how the surface area to volume ratio of a discotic platelet increases
when thinner, higher aspect ratio platelets are utilized.

The advantages provided by the widespread use of polymer nanocomposites can be significant.
For example, if automobile weight was reduced by the judicious use of polymer nanocomposites
vehicle manufactures could reduce their gasoline need by 1.5 billion liters and carbon dioxide
emissions by 10 billion pounds in just one year's time [3]. The elimination of multi-polymer,
multi-layer

materials

in food and beverage

packing and replacement

with polymer

nanocomposites would significantly increase the recyclability of these materials and further ease
the burden placed on the environment. Consequently, the chemical company BASF predicts that
nanocomposite formulations and bulk materials will account for 50 billion Euros of the
nanotechnology market in 2015 as shown in Figure 1-3. Meanwhile, resulting nanocomposite
components will account for an even greater share-over 500 billion Euros of that same
market [4].

Market potential 2015


in billion Euro
500-

50-

0.5-

NanoparWtices
orgamic and
mnrgamic

00
n
UU" -

I . _ ll:lij

Source: Griw I. ecast Ejropean ComnrIsior, DG 5anr~ 5CC


Consufre Marke Data fho- Sal Oppenelmr, In Reals and Evouticn Capia

Figure 1-3: Prediction of nanoparticle, nanocomposite formulation and nanocomposite component market
size by the chemical company BASF for the year 2015 taken from [41.

The most often utilized inorganic nanoparticle used for polymer matrix reinforcement is
montmorillonite-a natural silicate [3, 5]. Clay platelets like montmorillonite that are 1-2 nm
thick are attractive inorganic nanoparticles because they are naturally-occurring, inexpensive,
and their aspect ratio ranges from 25 to 600 depending on the type utilized. Table 1-1 lists and
compares a series of clays that can be utilized as inorganic nanoreinforcement. Since the
development of Nylon-6/Montmorillonite composites in the late 1980s by Toyota Research Labs
[6], polymer-clay nanocomposites have attracted much attention. By adding less than 5 wt% clay
to nylon, researchers were able to significantly enhance the mechanical properties, heat
resistance, and dimensional stability of the nylon and utilize the material in an under-the-hood
automotive application, i.e. as a timing belt cover [1].
Clay

Empirical Formula

Hectorite
Laponite
Montmorillonite

Nao.3Mg 2 .7 Li0.3 Si 4Oio(OH) 2


0 7
Na+ 7 [(Si 8Mg. 5Li. 3 )O 20(OH) 4 10 H 2 0
Na 0 .2Cao.lA12 Si 401 0(OH) 2 O

Rectorite

Nao.6Ca

O. 3

K 0 1 A1 6 Si 6

"

0
20

(OH) 4 2H 2 0

Lateral dimension,
D [nm]
600
25
70-218

Thickness, t
[nm
1
1
1

d-spacing
[nm]
1.53
1.28
1.50

200-300

2.22

Saponite

2
Nao. 1Cao.Mg2 .25Fe 0.75Si 3A1Oo(OH) 2 -4H20

165

1.23

Sepiolite

(OH )4(OH28H
(Sil 2Mg 8030)(OH)
20
2 4

25

20-1500

1.03

Table 1-1: List and comparison of the various clays investigated for polyurethane reinforcement [7-11].

Because well-dispersed clay can significantly increase the strength, heat distortion temperature,
barrier properties, flame retardance, and dimensional stability of a polymer at small volume
fractions, polymer-clay nanocomposites are desired in numerous applications. Endeavors by the
Army to design light-weight vehicles, contamination-protecting

and freshness-enhancing

packaging for food, and uniforms that better protect soldiers from shrapnel each rely upon
polymer nanocomposites [12]. Work by electronics companies like Intel to design low-cost,
flexible substrates and packaging for semiconductors, and by energy and biomedical companies
to design 'smart,' actuating devices indicate that polymer-clay nanocomposites will comprise an
even greater share in future product markets [13-15].

Tough, high energy absorbing materials which retain sufficient elasticity are desired for the
closefitting fabric of soldier's uniforms.

Fabric materials which can be worn comfortably, yet

fit closely for improved chemical and biological protection, are currently of interest to the Army.
Traditional polyurethane materials like Lycra/Spandex do not provide satisfactory strength and
toughness when compared to Nylon/cotton fiber blends.

Consequently, development of higher

modulus, light-weight, elastomeric materials that can exhibit enhanced cut, tear and puncture
resistance, resistance to fabric wear, in conjunction with breathability are needed. These superior,
lightweight, energy absorbing materials would also be useful in helmets and in other gear that
protects the soldier's body, especially his or her extremities, from low to moderate strain rate
impacts. These impacts could include those received during general field maneuvers, moving
across sharp objects, or from falling projectiles. The current, baggie BDU uniform is heavy and
suffocating. Through utilization of a more closefitting, lightweight garment with enhanced
mechanical properties the soldier gains mobility through enhanced comfort and protection.
Adding nanoreinforcement to polyurethane is one way in which to strengthen and toughen
polyurethane.

Polyurethanes are attractive nanocomposite matrix materials due to their numerous product
markets, mechanical versatility, and tunable block-polymeric structure. Polyurethanes are wellstudied multipurpose polymers that comprise about 5% of the polymer market today and are used
in a variety of applications, including in chair cushions, bra padding, artificial organs and other
biomedical applications, in active wear (spandex/lycra), footwear, roller ski and skate wheels, ski

boots, wake boards, and in paints and coatings [16, 17]. The various soft and hard blocks of
polyurethane can interact with nanoparticles and organic modifiers depending on the polarity of
these segments. In the past eight years numerous research groups [5] have examined various
large aspect-ratio (a >100) nano-clays, organic modifiers, and dispersal techniques for clay in
polyurethane, most of which have relied upon in situ polymerization. These nanocomposite
systems often display thermomechanical enhancement and/or better barrier properties; however,
clay intercalation persists and limits the extent of these improvements.
Two challenges exist in the development of polymer/clay nanocomposites. The first challenge is
fully dispersing and exfoliating inorganic, hydrophilic clay within organic and often hydrophobic
polymer matrices. The second major challenge, as Krishnamoorti and Vaia assert, "is developing
nanocomposites

for

systems

ranging

from

high

performance

to

commodity

polymers...[without]...even simple structure-property models" [3]. By using a novel solvent


exchange approach and a true nano-clay filler-meaning the longest length scale is on the order
of tens of nanometers-elastomeric thermoplastic polyurethane-clay nanocomposites with
superior thermomechanical properties have been developed without reducing the extensibility of
the matrix elastomer. Consequently this nanocomposite system and other polyurethane
nanocomposites containing discotic, Laponite RD, platelets which measure 25 nm in diameter
and less than 1 nm in thickness are excellent systems in which to investigate the structureproperty relations because the size of the inorganic filler is similar to that of polyurethane hard
domains.
This thesis details the development of elastomeric polyurethane/clay nanocomposites that are
preferentially reinforced and describes the resulting structure-property relations. Chapter 2
introduces the reader to polyurethanes, clay reinforcement, specifically the nano-clay Laponite,
and previous polyurethane-clay nanocomposite research findings. In Chapter 3 our method for
the nanoparticle preparation is described and its efficacy compared to other dispersal techniques.
Then in Chapter 4 the resultant characterization and thermomechanical properties of the model
Elasthane-Laponite nanocomposite are emphasized. The complete thermomechanical response is
detailed and morphological evolution examined in Chapter 5. Furthermore, a constitutive model
that captures the room-temperature, uniaxial tensile mechanical response of this polyurethane-

clay nanocomposite is discussed. The knowledge gained from the thorough analysis of the model
nanocomposite system is utilized in conjunction with experimental results to evaluate the
influence of Laponite reinforcement on the strain recovery and the resulting actuating stress of a
commercial shape memory polyurethane/clay nanocomposite in Chapter 6. Furthermore,
discussion of the extension of the solvent exchange approach for polyurethane nanocomposites
to an alternate, "fast," processing technique-electrospinning-in order to develop multiresponsive, porous, non-woven, shape-memory fabrics is presented in Chapter 7. Finally, in
Chapter 8 a summary of the results is listed with recommendations for future work.

1.1 References
[1]

[2]

[3]
[4]

[5]
[6]

J. DeGaspari, "Prospecting Paydirt,"


http://www.memagazine.org/backissues/april01 /features/prospect/prospect.html, April
2001: Accessed February 17, 2005.
N. Sheng, M. C. Boyce, D. M. Parks, G. C. Rutledge, J. I. Abes, and R. E. Cohen,
"Multiscale micromechanical modeling of polymer/clay nanocomposites and the
effective clay particle," Polymer, vol. 45, pp. 487-506, 2004.
R. Krishnamoorti and R. A. Vaia, "Polymer Nanocomposites: Synthesis,
Characterization, and Modeling," in ACS Symposium Series, 1st ed. Washington, D.C.:
American Chamical Society, 2002.
BASF, "Nanotechnology: We innovate for growth,"
http://www.corporate.basf.com/en/sustainability/dialog/pohitik/nanotechnologie/service/p
raesentationen.htm?getasset=fileabs 1&name=BASF Equitvy Story Nanotechnology Jul
y07.pdf&MTITEL=Charts&suffix=.pdf&id=2mpN9Circbcp-2W, 2008: Accessed August
1, 2008.
S. S. Ray and M. Okamoto, "Polymer/layered silicate nanocomposites: a review from
preparation to processing," Progressin Polymer Science, vol. 28, pp. 1539-1641, 2003.
T. K. Chen, Y. I. Tien, and K. H. Wei, "Synthesis and characterization of novel
segmented polyurethane clay nanocomposite via poly(epsilon-caprolactone)/clay,"
JournalofPolymer Science Parta-Polymer Chemistry, vol. 37, pp. 2225-2233, 1999.

[7]
[8]

[9]

D. Barthelmy, "Mineralogy Database," 2004.


X. Y. Ma, H. J. Lu, G. Z. Liang, and H. X. Yan, "Rectorite/thermoplastic polyurethane
nanocomposites: Preparation, characterization, and properties," Journal Of Applied
Polymer Science, vol. 93, pp. 608-614, 2004.
K. Yano, A. Usuki, and A. Okada, "Synthesis and Properties of Polyimide-Clay Hybrid
Films," Journalof Polymer Science PartA-Polymer Chemistry, vol. 35, pp. 2289-2294,

[10]

[11]

1997.
C. Zilg, F. Dietsche, B. Hoffmann, C. Dietrich, and R. Mulhaupt, "Nanofillers based
upon organophilic layered silicates," MacromolecularSymposia, vol. 169, pp. 65-77,
2001.
H. X. Chen, M. S. Zheng, H. Y. Sun, and Q. M. Jia, "Characterization and properties of

sepiolite/polyurethane nanocomposites," MaterialsScience And EngineeringA-Structural


MaterialsPropertiesMicrostructureAnd Processing,vol. 445, pp. 725-730, 2007.

[12]
[13]
[14]

[15]

[16]
[17]

MIT Institute for Soldier Nanotechnologies, "ISN Day: The ISN-Army-Industry


Collaboration Conference." Cambridge, MA, May 28th, 2008.
D. Ratna and J. Karger-Kocsis, "Recent advances in shape memory polymers and
composites: a review," JournalOf Materials Science, vol. 43, pp. 254-269, 2008.
P. P. Kundu, V. Sharma, and Y. G. Shul, "Composites of proton-conducting polymer
electrolyte membrane in direct methanol fuel cells," CriticalReviews In Solid State And
Materials Sciences, vol. 32, pp. 51-66, 2007.
Intel, "Technology Innovation,"
http://www.intel.com/technology/index.htm?iid=gg work+home technology, 2008:
Accessed August 1, 2008.
D. Randall and S. Lee, "The Polyurethanes Book," 2nd ed. New York: Huntsman
International, 2002, pp. 477.
T. Thomson, Polyurethanesas Specialty Chemicals: Principles andApplications. Boca
Raton, Florida: CRC Press, 2005.

2. Background
The following sections provide more detailed information regarding the findings of previous
researchers, which inspired the development of the thermoplastic polyurethane-Laponite
nanocomposites as synthetic, high-energy-absorbency, multi-functional materials. The structural,
mechanical, and thermal properties of spider silk and thermoplastic polyurethanes are examined
and compared. Furthermore, the efficacy of clay nanoreinforcement in polymers is evidenced
and the major challenge facing the development of polymer-clay nanocomposites is discussed as
well as the polyurethane-clay research over the past ten years summarized.

2.1 Inspiration
A strong, elastomeric material with excellent toughness would be ideal in numerous products
ranging from clothing and food packaging for soldiers, to compression stockings for flight
attendants. The material would have a high tolerance to puncture and tear yet would be very
flexible and easy to wear. Spider silk is nature's high-performance polymer that best exhibits
these qualities.

The orb-web-weaving araneid spider can produce a family of high-performance silk fibers that
exhibit superior energy absorbency capabilities in conjunction with various combinations of
extensibility, stiffness, and strength. Each fiber in the family has a different function, depending
on the mechanical properties of the fiber. These mechanical properties in turn depend upon the
protein combinations excreted by the spider. Consequently, a spider is a technologically
advanced fiber manufacturer, producing fibers for various functions by manipulating amino acid
sequences, water content, and spinning speeds [1].
The orb-web-weaving araneid spiders are most studied because these spiders have seven glands
which produce different proteins that each has a specific function [1]. For example, the major
ampullate gland produces the dragline or web, frame silk and the flagelliform glands produce the
core of the spiraling, fly-catching, viscid silk [1]. These fibers have completely different
mechanical functions and behaviors, yet have similar energy absorbency capabilities as shown in
Figure 2-1 and Table 2-1.

i ll

Aramms dadedatus MA sik


dagline and web frame

iI\

1.21-0

.os
lrmns diadh

s viscid sil

catching spiral of aib-web

060-4-

Eim-=1O GP
o0-

~ooo
"z-J

n.

15

0.5

2.5

I
3

Strain

Figure 2-1: Engineering stress-strain curves of dragline and viscid silk as displayed in 121.

The dragline silk, which must hold the weight of the spider and form the frame of the web,
exhibits an initial modulus of 10 GPa, an ultimate strength of 1.1 GPa, a 27% extensibility and a
toughness of 160 MJ m-3 [2]. The viscid silk, which must bend extensively to slow and trap flies
-3
that collide with the web, exhibits a similar toughness of 150 MJ m , but instead an initial
modulus of 3 MPa, a 270% extensibility, and an ultimate strength of 500 MPa [2]. Other natural
and synthetic materials may exhibit similar combinations of stiffness, extensibility and strength
to silk, but the toughness of native-spider silk is unmatched as is shown in Table 2-1.

Viscid Silk
Synthetic Rubber
Nylon Fiber
Silkworm Silk
Wool
Kevlar 49 Fiber
Elasthane
Carbon Fiber
Tendon Collagen
High-Tensile Steel
Bone

0.003
0.001

5
7
0.5
130
0.038

U.5

Z.1

13U

0.05
0.95
0.6
0.2
3.6
0.032

8.5
0.18
0.18
0.5
0.027
3.9
0.013
0.12
0.008
0.03

100
80

300

1.5
200
20

0.15
1.5
I

0.16

70
60

50
48
25
7.5
I

6
4

Table 2-1: Comparison of the mechanical properties of silk to other high-performance materials 121.

Synthetic rubber comes closest to matching the toughness of the weaker, viscid silk, but the

ultimate strength exhibited by the viscid silk is unachievable in synthetic or natural rubbers.
Consequently, knowledge of the amino acid sequence and morphology of spider silk, and their
relation to the mechanical properties is critical for the development of new high-performance
materials.

Investigation of the amino acid sequence and structural motif of the various spider silks has
shown that dragline spider silk, in particular, is an optimized block copolymer. The silk is based
on two proteins, spidroin I and spidroin II, sequenced by Xu and Lewis in 1990 [3]. Like multiblock polymeric systems, the resultant silk consists of hydrophilic, soft glycine-rich blocks and
hydrophobic, alanine-rich hard blocks. The alanine-rich blocks hydrogen bond with each other
and form small crystalline 3-sheets that comprise 12% of the fiber by volume and measure
approximately 2 x 5 x 7 nm 3 [4]. These highly oriented P-sheets are coupled to the amorphous
glycine-rich soft blocks via the less oriented P-sheets. Consequently during deformation these
crystalline sheets are moved and aligned by the soft glycine-rich matrix [5]. The soft blocks give
the fiber its elasticity and their structural motif is not easily identified and may consists of a
combination of 31-helices, P-spirals, and random coils, which are probably highly ordered and
hydrogen bonded [6-8]. Furthermore, the soft blocks may change conformation depending on
extension and water content, providing a means for the silk to obtain the superior ultimate
strength it exhibits prior to failure [7].

Dragline silk also exhibits remarkable thermomechanical behavior. At -200'C, dynamic


mechanical analysis shows that the flexural storage modulus of dragline silk drops from -10 GPa
to -2 GPa when the intermolecular hydrogen bonds of the non-crystalline domains break down
[8]. This behavior is analogous to the soft segment glass transition in block-co-polymers. This
material exhibits amazing heat resistance, only undergoing mechanical failure at 371C the point
when the 0-crystallites break down [8]. This observed behavior is analogous to block-polymeric
hard segment melting. Not only, is spider silk, dragline silk specifically, remarkable at room
temperature, but the static mechanical properties between -60'C and 150'C are superior when
compared to other properties [8]. Although the mechanical properties of spider silk, like
polymers, are time and temperature dependent, spider silk is functional over a wider temperature
range.

Spider dragline silk is a highly evolved and optimized protein polymer that has a combination of
thermomechanical properties that are unmatched by synthetic materials. The strength, toughness,
and heat resistance exhibited by dragline silk fibers are desirable. If these fibers could be
replicated on a mass scale the potential applications are vast. At this point synthetic analogs of
spider silk, like nylons, rubbers, and polyurethanes are still thermomechanically inferior to
native-spider silk.

2.2 Introduction to Polyurethanes


Polyurethanes are well-studied versatile polymers that have been used in a variety of
applications, including in chair cushions, bra padding, artificial organs and other biomedical
applications, in active wear (spandex/lycra), footwear, roller ski and skate wheels, ski boots,
wake boards, and in paints and coatings [9, 10]. Any polymer that contains urethane linkages,
depicted in Figure 2-2, is considered a polyurethane.
O

C
N

Figure 2-2: A urethane linkage


Polyurethanes or polyurethane ureas are made by reacting diisocyanates with dialcohols or
diamines, respectively [11]. The urethane linkages like to hydrogen bond and can form rigid,
crystalline aggregates. In numerous applications segmented polyurethanes are desired and
polyols, polymeric compounds containing alcoholic hydroxyl groups, of various lengths
(Mw -600-3000 g mo'-1) are polymerized with the urethane reactants in either a one-step or twostep reaction [12]. The material properties of the segmented polyurethane can be tuned between
that of a hard plastic and a flexible elastomer depending on the polyol added and the resultant
chemical or physical crosslinks formed [12].
Thermoplastic polyurethanes are those segmented, elastomeric polyurethanes that contain soft

polymeric segments that have a glass transition temperature (Tg) below room temperature [13].
Consequently, at temperatures below the soft segment Tg a thermoplastic polyurethane behaves
as a glass, while at temperatures above the Tg the polyurethane behaves as an elastomer. The soft
segment imparts good extensibility and the rigid hydrogen bonded crystal aggregates, identified
as hard domains, act as physical crosslinks imparting moderate stiffness and strength [9].

The

stiffness,

strength,

extensibility,

and

thermomechanical

behavior

of segmented

polyurethanes are dependent upon numerous factors. These include the extent of phase
separation between the hard and soft segment, the concentration of hard segment to soft segment,
the type of polyol added, the molecular weight of the polymer, and the extent of the polymer
crystallinity [12].

(A

3
0
cc
"o

o
0)
X
4)
0
C,
x
ol
L.
0)

TT,

Temperature

T
Tm

Figure 2-3: Thermomechanical behavior of various polyurethanes.

For example, soft segment crystallinity or hard segment content can be increased in segmented
polyurethanes to increase the magnitude of the rubbery plateau modulus, as shown in Figure 2-3.
Furthermore, an increase in polyurethane molecular weight can increase the hard segment
melting temperature (T,,) or point of liquid flow without influencing the magnitude of the
rubbery plateau modulus [12]. Consequently, there are numerous ways in which the
thermomechanical properties of polyurethanes can be tailored.

The semi-crystalline, block-polymeric structure of elastomeric polyurethane is similar to spider


silk. Each contains crystalline aggregates that act as physical crosslinks. The aggregates or hard
domains in polyurethane, which have a lateral dimension measuring 3-32 nm in length
depending on the chemistry, are bigger than the 3-sheets in spider silk [14-16] and are easily
tailored. Unfortunately, the properties of thermoplastic polyurethanes are inferior to spider silk.
Because the onset of the urethane dissociation is near 200 0 C, polyurethanes undergo a threeorders-of-magnitude

change in flexural storage modulus between -100C and 2500 C.

Comparatively, spider dragline silk undergoes less than an order-of-magnitude change within
0
this temperature range because the dissociation of the P-sheets is well above 200 C [8, 12].

Polyurethane-based materials have the potential to be spider silk analogs if the mechanical
properties and thermal stability of the polyurethane can be significantly enhanced.

2.3 Importance of Polymer-Clay Nanocomposites


Nano-clay has proven to be effective polymer filler, enhancing mechanical properties, thermal
stability, and decreasing permeability of numerous polymers at small volume fractions. For
example, in the late 1980s Toyota Research Labs developed a Nylon 6-Monmorillonite
nanocomposite that at only a 4.2 wt% clay loading exhibited a 100% increase in initial modulus,
a 50% increase in ultimate strength, and a 80'C increase in heat distortion temperature [17]. This
thermomechanical enhancement was greater than that produced by glass or talc at over 30 wt%
loading

[18].

This finding

initiated numerous

investigations

of other polymer-clay

nanocomposites.
Nylon was not the only polymer that exhibited thermomechanical enhancements and improved
barrier properties when filled with clay. The inclusion of well-dispersed clay has exhibited
enhancements in a handful of other polymer systems as is shown in the list below.
*

In 1993, Yano and colleagues from Toyota Research Labs in Japan found that when
polyimide was filled with 2 wt% synthetic mica the water vapor permeability coefficient
was less than 10% that of the pure polyimide permeability coefficient. Furthermore, the
thermal expansion coefficient of the composite was 60% of the pure polyimide
coefficient [19].

In 1994, Lan and Pinnavai found that when rubbery epoxy was filled with exfoliated
Montmorillonite at 15 wt%, the composite exhibited more than a 10-fold increase in both
tensile strength and modulus when compared to the properties of the pure epoxy [20].

In 2001, Nam and colleagues reported that a 6 wt% loading of modified Montmorillonite
in polypropylene foam resulted in a 44'C increase in heat distortion temperature when
compared to the properties of the pure polypropylene [21].

In 2002, Lepoittevin and colleagues reported that poly(c-caprolactone) exhibited a higher


thermal stability when filled with modified Montmorillonite. Whether the clay was
exfoliated or intercalated the degradation temperature at 50% mass loss was at least 250 C
higher than the degradation temperature of the pure poly(E-caprolactone) at the same
mass loss percentage [21].

Currently, as the price of petroleum increases and concern for the environment becomes
more prevalent, there has been a push to develop polymers from renewable bio-based
materials. While some researchers have developed soybean-oil- or fatty-acid-derivativebased polyurethane thermosets and castor-oil and rapeseed-oil based waterborne
polyurethane dispersions, Lu and Larock have reinforced a corn oil based blockcopolymer with modified Montmorillonite [22]. As was exhibited in the petroleum based
polymer

composites, these

composites

display increased thermal

stability and

improvements in mechanical properties when only 2-3 wt% modified Montmorillonite is


added.

In the past twenty years, numerous researchers have shown that clay can effectively reinforce
polymer systems, whether petroleum or bio-based, and significantly improve the mechanical,
thermal, and barrier properties of the pure material. The reinforcement of polyurethanes with
clay appears to be a feasible method to enhance the thermomechanical properties of polyurethane
and develop synthetic spider silk analogs-high performance, flexible, strong, energy absorbing,
multi-functional composites.

The major challenge in developing polymer/clay nanocomposites is fully dispersing and


exfoliating inorganic, hydrophilic clay platelets within organic and often hydrophobic polymer
matrices. Numerous clays, including Montmorillonite, Hectorite, and Saponite among others,

have been investigated as polymer filler. Whether natural or synthetic, these silicates contain
alkali and alkaline earth cations absorbed on the negatively charged silicate faces which act as
glue binding the silicates in stacks [21]. If the silicates and polymer matrix do not interact and
the silicates remain aggregated in stacks, the mechanical properties of the material are poor [23].
However, if the silicate and polymer interact favorably the silicate stacks can be delaminated and
the small-volume, high aspect-ratio, stiff platelets can offer a large surface on to which the soft
polymer can transfer stress and can provide a torturous path in which permeants must
traverse [17]. Consequently, it is important that when clay is blended with polymer that the
silicates are delaminated and well-dispersed, so that the maximal enhancements are obtained.

Three different techniques are employed to disperse the clay within the polymer: melt dispersion,
in situ polymerization, and use of a common solvent. Each technique as well as a new, solvent
exchange technique is discussed in Chapter 3. The effectiveness of these dispersal methods
depends on numerous factors, including the polarity of the clay and/or organic modification, the
reactivity of the clay and/or organic modification, as well as the mobility of the polymer chains.
The extent of the clay dispersal in the polymer is defined by the gallery spacing (distance
between silicate faces) and the orientation of the silicate platelets. When the silicate stacks are
not delaminated at all and, instead, remain in tight aggregated tactoids, the clay acts as macroscale filler and the resultant material can be considered a conventional composite [17]. These
aggregates may increase the stiffness of the polymer at the expense of elongation, strength, and
toughness and are unfavorable. When the gallery spacing of the clay increases because a few
molecular layers of polymer have been inserted between the silicates, the clay is intercalated. In
some instances the edges of the intercalated clay interact and the clay flocculates and is
considered intercalated-and-flocculated [21]. When the clay is delaminated and the individual
layers are separated throughout the polymer matrix and there are few to no steric interactions
between platelets, the clay is considered exfoliated [24]. However, as the clay content increases,
the average spacing between platelets necessarily decrease. In some instances, an ordered
exfoliated structure, which is detected by wide angle X-ray diffraction (WAXD), can result [24].
Various possible clay dispersals are depicted in Figure 2-4.

Aggregated Tactoids

Ordered Exfoliated

00

Disordered Exfoliated

Figure 2-4: Schematic illustrations of various polymer-clay nanocomposites


with different extents of clay dispersal.

Typically, research groups use very large aspect-ratio (a = L / t ) clays, like Montmorillonite
(a -70 to 218), that are approximately 1 nm thick to fill their polymers [19, 25]. Because the
lateral dimensions of these silicates are large, often when exfoliation is achieved at small weight
fractions, intercalation persists at higher concentrations. These large thin platelets are
substantially bigger than the hard domains within polyurethanes. Consequently, I chose to
reinforce polyurethane with the smallest aspect ratio clay available, Laponite (D = 25 nm, t = 1
nm), so that the lateral dimension of the silicate was similar to the lateral dimension of the
polyurethane hard domain. Therefore, just as spiders reinforce silk with P-sheets, and tire-makers
reinforce rubber with carbon black, we reinforce polyurethanes with Laponite [26].
Al ul Mg a
cryrwea Lc

Montmorillonite

Mg nd U In

arbsl stek

Laponite

Figure 2-5: Comparison of the Montmorillonite and Laponite


crystal structure as shown in 127].

2.4 Brief Synopsis of Previous Polyurethane Nanocomposite Research


The following section will briefly review a random selection of approximately 25% of the
polyurethane/clay nanocomposite literature published and listed in the ISI Web of Science
database. These articles do not constitute all polyurethane/clay nanocomposite literature, but are
a representative subset. The versatility of polyurethanes and the previously exhibited potential of
clay nanocomposites have made polyurethane/clay nanocomposites very desirable.

After the dramatic thermomechanical enhancement exhibited by Montmorillonite in Nylon-6,


numerous research groups began investigating the affects of clay reinforcement in other polymer
matrices [18]. It was not until 1998 that Wang and Pinnavaia reported the first instance of
successful reinforcement of elastomeric polyurethane (PU) with smectic clay [28]. In this work,
Montmorillonite, organically-modified with alkylammonium ions containing either 12 or 18
carbons, were solvated in polyols-polymeric compounds containing alcoholic hydroxyl groups,
of various molecular weight. The researchers found that the extent of clay intercalation was
determined by the length of the onium ion, the complex cation modifying the clay, not by the
molecular weight of the polyol. At a 10 wt% loading of Montmorillonite modified with the
longest onium ion a more than 100% increase in modulus, strength, and extensibility was
exhibited as shown in Figure 2-6 [28]. This result suggested that onium ions acted as a
plasticizer, improving elasticity, an observation noted by other researchers later [28, 29].
1.u

1 0

160
a.

t140

S3

I.

120

I-

40

10

Organo Clay Loading (wt %)

Organo Clay Loading (wl %)

1n

15

Organo Clay Loading (wt %)

Figure 2-6: Tensile properties of the polyurethane-clay nanocomposites prepared with the longest onium ion
Montmorillonite-modification at various clay loadings as reported in [28]: (A) tensile strength, (B) tensile
modulus, and (C) strain-at-break.

Less than a year later Chen, Tien, and Wei published the first of a set of publications examining
polyurethane/Montmorillonite nanocomposites [30]. In their first study, inspired by the Nylon-

6/Montmorillonite work nearly a decade earlier, polycaprolactone/Montmorillonite prepolymer


was prepared and substituted in as chain extender during PU nanocomposite synthesis.
Characterization showed that the clay was well dispersed within the polycaprolactone. However,
each platelet behaved as a nucleation seed inducing crystallinity that enhanced the tensile
strength of the PU nanocomposite while significantly hindering the extensibility at loadings
greater than 0.74 wt% clay [30]. In a completely different PU system, five years later, Moon and
colleagues found that the clay behaved in the same way, nucleating crystals in the polybutylene
succinate soft segments [31].
Not wanting the clay to nucleate crystallization at the expense of the soft domain, in the
following study Chen, Tien, and Wei investigated the affect of clay swelling agents, 12aminolauric acid (12COOH) and benzidine (BZD), on a 4,4'-diphenyl methane diisocyanate
(MDI), polytetramethylene glycol (PTMG, MW=1000 g mol-'), and 1,4-butanediol (BDO) based
PU nanocomposite [32]. Although x-ray diffraction suggested that the 12COOH better exfoliated
the clay, the composites made using the BZD exhibited greater thermal stability, more than a 3fold increase in elongation, and more than 2-fold increase in strength when compared to the pure
PU [32]. However, similar to the previous study, at loadings greater than 1 wt%
Montmorillonite, the mechanical enhancements deteriorated as shown in Table 2-2.

Pamre 1
1% 12COCBmm*aPU
3% 12CO0EI-f*lmPU
5% 12COElU
e
WU
1%B2D-maVU
3% B2D.-nmPU
5% BD-mamPU

A1

Tm1se SibOacgfC)

Ekinafim (%

Waw abs(piim(%)

10660
11330'
10720
106W
114604
1143W0
11040"

66.8
76-2
76.1
533
144
130
129

220
230
360
180
680
590
600

17
1.3
15
1.6
1.4
1.6
1.7

Table 2-2: The molecular weight, tensile strength, elongation, and water absorption of the pure polyurethane
and polyurethane-Montmorillonite nanocomposites prepared in [32].

Tien and Wei extended this work, again utilizing the BZD swelling agent, and investigated the
nanocomposite behavior when the hard segment content was increased from 39 wt% to 50 wt%
and then to 63 wt% [33]. They found that when the hard segment fraction increased, FTIR
showed that the carbonyl hydrogen bonding increased; however, as the clay content increased the
hydrogen bonding decreased and the extent of intercalation diminished. Ultimately, the
reinforcing ability of the clay is compromised when the hard segment hydrogen bonds are

disrupted and phase mixing results, and when the more rigid PU molecules are unable to diffuse
between and separate clay platelets. Consequently, the reduction in hydrogen bonding (37%) was
greatest for the PU nanocomposite containing 63 wt% hard segment and 5 wt% clay [33].

Attempting

to better enhance the clay-polymer

interaction and further improve

the

thermomechanical properties, in the next investigation Tien and Wei used the same PU system
with 39 wt% hard segment and evaluated four different swelling agents [34]. Three of the four
swelling agents contained at least one hydroxyl group which could react with the isocyanates of
the PU and tether the clay to the polymer matrix. Tien and Wei found that the extent of clay
dispersion increased with swelling agent reactivity and that the mechanical enhancement was
best when the clay platelets were tethered to the PU rather than trapped within it [34]. In the
following publication Tien and Wei evaluated the thermomechanical and thermal degradation
properties of the nanocomposites formed using the three reactive swelling agents [35]. They
found that the PU soft segment glass transition temperature (Tg) remained unaffected while the
hard segment Tg increased with increasing clay content [35].

Figure 2-7: Dispersal of 1 wt% Montmorillonite modified


with most reactive swelling agent in the PU containing 39
wt% hard segment as shown in [34].

The greatest improvement in degradation temperature, a 40'C increase above the degradation
temperature of the pure PU, was displayed in the nanocomposite that utilized the most reactive
swelling agent (containing 3 hydroxyl groups) with a loading of 1 wt% clay. This nanocomposite
also exhibited the best clay dispersal as shown in Figure 2-7. Ultimately, the maximum
thermomechanical enhancements are exhibited in the nanocomposites with the best clay

dispersal [35].

In 2001 Xu et al, a group from Pennsylvania State University, investigated a biomedical PU


system similar to that Tien and Wei studied in [30, 32-35], except the soft segment in this PU
had a higher molecular weight (MW=2000 g mol-), the hard segments (22 wt%) were chain
extended with 1,4-diaminocyclohexane

and ethylenediamine

(25:75 by mole) forming

polyurethane urea segments, and the Montmorillonite (Cloisite 15A) sodium ions were
exchanged with dimethyl ditallow ammonium (65%

C18 , 30% C16 , and 5% C14 ) [36].

Representative tensile behavior of the various nanocomposites is shown in Figure 2-8.

vod% akaet

.9%

Extensie (%)

Figure 2-8: Representative force-extension


polyurethane urea nanocomposites in 1361.

curves

of the

Xu et al found that at 20 wt% clay loading the nanocomposite exhibited over a 300% increase in
initial modulus, over a 30% increase in strength, over a 50% increase in extensibility, and over a

5-fold reduction in water vapor permeability when compared to the pure polyurethane urea [36].
The silicates reduce the permeability of the PU by forming torturous paths through which the
water vapor molecules must diffuse. Consequently, the reduction in permeability is dependent
upon the amount of clay added as well as the extent of its exfoliation. In a later publication, the

authors reported that Fourier transform infrared spectroscopy (FTIR) showed that the clay
platelets were kinetically trapped in the PU matrix and differential scanning calorimetry (DSC)
showed that the soft segment Tg was not influenced by clay loading [37].

In 2002 Tortora et al, a group from Italy, published a study similar to the first study published by

Chen, Tien, and Wei, in which they investigated the reinforcement effects of PU composites
containing poly(c-caprolactone)/Montmorillonite prepolymer as a chain extender [38]. In this
new study the prepolymer replaced di(ethylene glycol) in the hard-segment and induced the same
crystallinity in the poly(e-caprolactone) as that seen by Chen, Tien, and Wei with x-ray
diffraction peaks at 21.30 and 23.7' [38]. Similarly, the nanocomposite became more brittle with
increased clay content. The storage modulus increased with clay loading, particularly at
temperatures above the soft segment Tg, and resultantly the material flow temperature increased
by 500 C at 20 wt% clay loading as shown in Figure 2-9. Furthermore, although the elastic
modulus increased by over 5-fold at 20 wt% loading, there was a significant reduction in the
strain hardening slope which indicated that the clay disrupted the soft segment mobility and
alignment as shown in Figure 2-9, reducing the strength and toughness of the material [38].
'4-7

'0

K.

t'

46
hii

-IZ5

-10

-75

-25

054"

T(

I 'll

4g

75

1.~_

IlK)

*t

'Im

4u,

6I"

wo

4iM

Strain (%)

Figure 2-9: Storage moduli (left) and stress-strain plots (right) of the polyurethane
containing 0 (a), 4 (b), 20 (c), and 40 (d) wt% clay from [38].

In 2003 Osman et al evaluated the gas permeability of a PU adhesive filled with Montmorillonite
modified by three different ammonium ions [39]. The Montmorillonite was exchanged with
organic cations that had either an aliphatic hydrocarbon nature, an aromatic moiety, or had a
polarity that matched the PU polarity. They found that the water vapor permeability decreased
by 40-50% at a 3 vol% clay loading for all three Montmorillonite modifications. However, the
oxygen permeability decreased by 30% at 3 vol% loading only when the aromatic and polar
modifications were made [39]. The nanocomposite composed of clay modified with the
hydrocarbon-based ammonium ion instead increased oxygen permeability. The authors speculate
that this was because the non-polar modification phase separated at the PU-silicate interface and
increased the free volume at these interfaces enough that small oxygen molecules could diffuse

through, but large clusters of water vapor molecules could not [39].

Also in 2003, Wang et al, who were interested in developing novel electrolytes for solid-state
rechargeable batteries, investigated the influence of Montmorillonite addition on the ionic
conductivity of a polyurethane electrolyte [40]. The electrolyte was filled with between 0 and 6
wt% Montmorillonite and exhibited the best ionic conductivity at 2.5 wt% loading [40]. The
enhanced conductivity was due to the ability of clay to reduce the PU crystallinity, to provide
more ionic transport sites, and to maintain the PU mechanical stability.

In 2004, Chen-Yang et al published work in which they examined the thermal and anticorrosive
affects that Montmorillonite conferred on a PU, containing polypropylene glycol (MW=3000 g
mol-1), toluene diisocyanate (TDI) and 1,4-butanediol (BDO) [41]. They found that the thermal
stability, char formation, and flame retardance all increased with increasing clay content.
Furthermore,

Montmorillonite

enhanced the anticorrosive properties

of the PU. The

nanocomposite that exhibited the best clay dispersion, 2 wt% Montmorillonite modified with
4,4'-diaminodiphenylmethane, exhibited an order of magnitude slower corrosion rate (1.09 x 10-5
mm yr - ) than that exhibited by the pure PU (2.62 x 10-4 mm yr - ') [41].

Finnigan et al, a group from Australia evaluated the reinforcement effects of two different clay
dispersal techniques (1. solution intercalation and 2. melt intercalation via a twin-screw extruder)
in both hard (55 wt% hard segment) and soft (35 wt% hard segment) PUs composed of PTMO
(MW=1000 g mol-'), MDI, and BDO [42]. Finnigan et al found that clay dispersion via extrusion
was slightly better as shown in Figure 2-10 because during the slow process of solvent
evaporation the clay platelets had time to re-aggregate.

However, the high temperatures needed for the melt in twin-screw extrusion caused the
Montmorillonite modifying surfactant and polyurethane to react and reduce the molecular weight
of the nanocomposite which in turn degraded the material properties of the polyurethane.
Interestingly, both dispersal methods increased the material stiffness and decreased the ultimate
strength and extensibility of the composite with increasing clay content. The increase in stiffness
was greater in the solvent cast films (3.2-fold increase at 7 wt% loading) [42]. The authors

believe the Montmorillonite modification surfactant is responsible for the degradation of


composite extensibility and strength and are currently investigating this behavior.

Figure 2-10: Transmission Electron Micrographs of the solvent cast (left) and melt
compounded (right) soft polyurethane nanocomposites containing 3 wt% Montmorillonite
depicted in [42].

In 2005, Finnigan et al utilized the same TPU with to evaluate the influence of silicate lengthscale on the microphase morphology and mechanical. The TPU was filled with a single weight
fraction (3 wt%) of clay (Somasif MEE, a synthetic fluoromica), but of different characteristic
length: 30, 75, 200 and 650 nm. The authors show that the aspect ratio of the silicates modifies
the mechanical behavior, as duplicated in Figure 2-11, but does not change the phase
morphology [43]. In 2006, Finnigan et al evaluated the influence of this same filler at its smallest
and largest aspect ratio on the mechanical properties and morphology of polyurethanes
containing poly(hexamethylene oxide) soft segments of increasing molecular weight [44]. They
found that the particle size did not affect the microphase morphology in the system either.
However, the enhancement in modulus provided by the filler was greatest in the least phaseseparated system. In contrast, when the smaller nano-clay was added to the polyurethane
containing poly(tetramethylene oxide) soft segments, a dramatic upturn in the stress-strain curve
combined with increased tensile strength resulted. This indicated that the soft segment formed
secondary bonds with the filler and improved stress transfer by reducing molecular slippage [43,
44].

1.2

o.s

0.4
0.0
0
IV

(b)

4
6
Strain (%

10

TPU-30

60 -

TPU-75
TPU

. 50C.
240-

TPU-200

30-

TPU-650

2010 0-

400

800

1200

1600

Strain (%)
Figure 2-11: Stress-strain response of a polyurethane filled with the same
concentration of synthetic fluoromica (3 wt% ) but of different aspect ratio

1431.
Extending their previous work, which evaluated the influence of length scale of filler on the
microphase behavior and the mechanical response to failure in tension, Finnigan et al completed
a more thorough analysis of the tensile mechanical behavior of these composites in the PTMO
based polyurethane [45]. At strains less than 100%, as platelet size increased the relaxation rate,
hysteresis, and permanent set of the composite increased.
50

40

-30

to

0
0

200

400

600

800

Strain (%)

1000

1200

Figure 2-12: Hysteresis exhibited by a polyurethane filled with the same


concentration of synthetic fluoromica (3 wt% ) but of different aspect ratio [43].

However, at large strains the hysteresis became less sensitive to platelet size, while set became
more sensitive as shown if Figure 2-12. This is likely due to chain slippage at the filler interface

and void formation. In contrast, the smallest filler (-30 nm) reduced the permanent set,
indicating that favorable interfacial interactions limited slip [45]. Also, at large strains, the
slowest relaxation time of the composite (which is affiliated with hard segment rich structures)
decreased when large filler particles were included.

Mishra, Kim, and Ha from Korea have used the same modified Montmorillonite (Cloisite 30B)
that Finnigan et al used in their initial study in melt intercalated and compression molded
nanocomposites containing low-density polyethylene and millable polyurethane [46]. In contrast
to the work of Finnigan et al, they found that with increasing clay content both the stiffness and
strength of the composite increased, albeit minimally, while the extensibility decreased slightly.

lamX)

Figure 2-13: Storage modulus versus strain amplitude curves


of the polyethylene-polyurethane (PEPU) nanocomposites
containing 0, 2.5, and 5 wt% clay depicted in [461.

The heat shrinkability-the

ability of a previously stretched, but currently unstrained,

nanocomposite to reduce its length when heated--decreased with increasing filler content. Melt
rheology showed that this reduction is because the Montmorillonite disturbs and deforms the
entanglement network, decreasing the extent of the linear viscoelastic deformation region of the
nanocomposite material as shown in Figure 2-13 [46].
Cao and Jana evaluated the influence of 1, 3 and 5 wt% of this same clay, Montmorillonite
(Cloisite 30B), on the shape memory behavior of a custom synthesized polyurethane, containing
polycaprolactone [47]. The melting transition of crystalline polycaprolactone was utilized as the
shape memory transition trigger. At a 1 wt% loading the clay is exfoliated, but at higher

concentrations the clay is intercalated. Consequently, at the I wt% loading, the increase in
recovery stress (when strain is fixed) is 20% greater than that of the pure polyurethane as
displayed in Figure 2-14. This increase in recovery stress is also greater than that exhibited by
the other composites because the greater the clay loading, the greater the disruption of the hard
domains and soft segment crystallinity. This disruption increases the rate of stress relaxation that
the resultant composite undergoes, minimizing the achievable recovery stress.

co
I:

20

30

40

50

60

70

Temperature ('C I

Figure 2-14: The recovery stress of the shape memory polyurethane with 0,
1, 3 and 5 wt% clay at a fixed strain of 100% while a 4 0 C min- 1 heating rate
is applied 1471.

A majority of the polyurethane/clay nanocomposite studies published in the past ten years used
solution or melt, reactive or non-reactive in situ polymerization to prepare the nanocomposites.
Recently, Dan and co-workers evaluated the influence of two solvents, dimethylacetamide and
tetrahydrofuran, on the resultant dispersal of 5 wt% Montmorillonite with different organic
modifications (Cloisite 30B, 25A, 15A, and pristine) in polyether and polyester-based
thermoplastic polyurethanes [48]. Cloisite 30B was best dispersed with dimethylacetamide,
while tetrahydrofuran was the better solvent for Cloisite 15A. However, while the solvent-clay
interactions are important, they are secondary when there are strong interactions between the
polymer and clay. Furthermore, FTIR analysis indicates that the type of solvent as well as the
polymer-clay and solvent-clay interactions play a role in the resultant phase-behavior and
morphology [48].

Meng et al evaluated the roles of molecular diffusion and shear in the exfoliation process when
polyurethane and organically modified Montmorillonite (Cloisite 30B) are melt-mixed [49].

While molecular diffusion plays a role initially, during the first minute of mixing, the slippage of
plates due to applied shear accounts for particle exfoliation. At least three minutes of mixing is
necessary to maximize separation minimize the number average length of the platelets dispersed
within the composite. Prolonged mixing exposes the polyurethane matrix to the organic modifier
at high temperature and leads to degradation and deterioration of the mechanical properties [49].

Tan and Nie evaluated the dispersion of montmorillonite in polyurethane urea when three
different photo-polymerizable ammonium surfactants containing reactive methacrylate groups
were utilized [50]. The spacing between platelet faces (d-spacing) increased with the length of
the polymerizable surfactant. While the presence of well-dispersed silicates can form a tortuous
path decreasing water permeation, the composites are still susceptible to water sorption. Tan and
Nie found that the addition of the long alkyl, hydrophobic surfactant acted as a barrier to water
sorption, decreasing the content uptake from 3.4% for the unmodified montmorillonite to 0.77%
for the longest surfactant [50].

Rhoney et al evaluated the dispersion and reinforcement effects of nanocomposites when the
various fillers, i.e. reactive Montmorillonite, non-reactive Montmorillonite, and large silicon
nitride particles, were first mixed in the polyol via different agitation methods prior to in situ
polymerization [51]. The mixing methods included (i) sonication with stirring at 300 rpm, (ii)
high-speed mixing at 5,000 rpm, and (iii) high-shear mixing at 24,000 rpm. Steady shear
rheology showed that only the Montmorillonite/polyol solutions mixed via sonication evidenced
a yield stress at low shear rates and shear thinning at high shear rates, behavior that is indicative
of well dispersed clay platelets [52]. From wide angle x-ray diffraction (WAXD) spectra the
authors determined that the Montmorillonite with the reactive modification was well exfoliated
while the non-reactive clay was aggregated. Interestingly, at 3 wt% clay loading, the reactive
clay exhibited a 30% increase in storage modulus below the soft segment Tg and a 55% increase
in storage modulus above the Tg when compared to the unfilled PU. Conversely, the non-reactive
clay exhibited a 50% increase in storage modulus below the soft segment Tg and a 30% increase
in storage modulus above the Tg [51]. The high-shear mixed nanocomposite exhibited dynamic
mechanical behavior similar to the sonicated samples below the Tg but reduced moduli above the
Tg, indicating that the filler was not interacting with the PU.

Song and Yao published an interesting article in 2004 in which they examined the evolution of
the basal plane spacing of the clay or d-spacing in elastomeric polyurethane (containing
polypropylene glycol, methylene biscyclohexyl isocyanate, and BDO) nanocomposites via
WAXD during increasing cyclic mechanical deformations [53]. The authors found that the dspacing, the distance between the basal planes of clay platelets, increased during deformation,
but recovered to the original spacing upon release of the stretching force when the extent of
deformation was within the elastic range (up to 100% strain). From the original state to -200%
strain the d-spacing of a 5 wt% filled nanocomposite increased from -3.7 nm to -4.4 nm. Yet, at
strains greater than 200% the clay spacing exhibited less than a 0.2 nm increase as shown in
Figure 2-15 [53].
4.0-

4.4- I
E

4.0-

U
I.

3.6-

3.20

YT~T~

150

300
450
600
Elongation (%)

750

900

Figure 2-15: Evolution of d-spacing with respect to a


nanocomposite loaded with 5 wt% clay during stretching

1531.

The authors also evaluated the fatigue life of the nanocomposites. At 3 wt% clay loading the
fatigue life increased from 55,160 cycles for the pure PU to 821,352 cycles [53]. The authors
speculate that the clay intercalated between silicates forms a nano-spring and believes this type
of material could extend the useful life of tires.

In a later publication pertaining to the same nanocomposite system, Song et al analyzed the
morphology of the nanocomposites when the hard segment content was varied using small angle
x-ray scattering and atomic force microscopy (AFM) [14]. The nanocomposites evaluated
contained 18, 26, and 36 wt% hard segment. They found that the inter-domain repeat distance
increased with hard segment content but decreased with increased clay content. Furthermore, the

hard domain size increased with hard segment content and decreased minimally with increasing
clay content. The change in hard segment size was minimal because in this PU system the clay
interacts primarily with the soft, polyol segments and has little affect on hard segment formation.
AFM phase images showed that the hard domains formed spherical aggregates in the PU and the
aggregate size decreased from -800 nm to -500 nm upon the addition of clay as shown in Figure
2-16 [14].

Figure 2-16: Tapping mode AFM phase images of (a) pure polyurethane
containing 36 wt% hard segment and (b) the same polyurethane loaded
with 1 wt% clay in [141. The small spherical aggregates are hard domains
and the large masses shown in (b) are clay tactoids.

Xia et al also evaluated the relaxation behavior of pure polyurethane of various hard segment
content (16, 26, 36, and 46 wt%) and that of the pure polyurethane with 26 wt% hard segment
and 1, 3 and 5 wt% modified Montmorillonite (Cloisite 20A) [54]. In contrast to the results of
Finnigan et al, they found that the addition of clay encouraged phase separation and the hard
domains decreased in size. As the hard segment content increased the polyurethane characteristic
relaxation time decreased. However, the addition of clay, like Finnigan et al reported, too
increased the relaxation time. As the applied strain increased to values above 100%, multiple
relaxation mechanisms became apparent. Xia et al correlate these peaks in the relaxation
spectrum to interconnected hard domain break-up, soft segment chain pullout from hard
domains, and disentanglement of the chains within the soft segments [54].
Rahman et al prepared numerous waterborne polyurethane/montmorillonite (Cloisite 15A)
nanocomposites with three different concentrations of 2,2-dimethylol propionic acid (DMPA):
3.75, 5.41, and 6.17 wt% [55]. They found that as the concentration of DMPA increased, and the

hydrophilic nature of the pre-polymer increased, the mean particle size decreased. However, for
each DMPA concentration there was a corresponding optimum clay concentration which
provided the largest enhancement in tensile strength, elastic modulus, and adhesive strength: 0.5,
1, and 2 wt% respectively [55]. Surprisingly, these clay concentrations did not correlate to the
smallest mean particle size.

Because polyurethane foams are the most demanded of all polymeric foams [56], Cao and
colleagues investigated the affects of adding Montmorillonite to two polyurethane foams, one
more rigid than the other [57]. When clay was added, it acted as nucleation sites for cell
formation, creating more small cells than that formed in pure PU foams. The inclusion of clay
increased the soft segment Tg of the flexible foam and decreased the soft segment Tg of the rigid
foam [57]. Similarly, the clay increased the reduced (normalized with respect to density)
compressive strength (780%) and reduced modulus (650%) of the flexible foam and decreased
the reduced compressive strength (760%) and reduced modulus (610%) of the rigid foam.
Ultimately, there appears to be a competition between the clay induced reinforcement and
enhanced foam morphology and the disruption of the PU hydrogen bonding and network
structure [57].

In 2005, Song et al investigated how the addition of Montmorillonite in conjunction with


melamine polyphosphate powder affects the flame retardance and mechanical properties of the
PU [58]. Although the clay was intercalated the tensile strength increased from 1.53 MPa to 3.85
MPa when 5 wt% clay was added with 6 wt% melamine polyphosphate [58]. Thermogravimetric
analysis on this same nanocomposite showed that the mass loss rate, the CO and CO 2 release
amount, specific extinction area, and the heat release rate all decreased significantly, increasing
the material flame retardance. The authors speculate that the flammability is retarded because
the carbonaceous-silicious-phosphorated char layers coat the materials protecting the interior
polymer.

Recently, Rajendran and co-workers evaluated the release of Bifenthrin I, a termiticide,


encapsulated in a polyurethane containing 0, 0.5, and 5 wt% commercial Bentonite (an impure
clay composed primarily of Montmorillonite) [59]. Unless the polyurethane is subjected to a

solvent that reacts with the matrix polymer and significantly swells the polyurethane, the
Bifenthrin persists. Although, Bifenthrin has a high affinity for silicates and should bind to the
platelets, the diffusion coefficients in the clay filled systems were faster than that of the pure
polyurethane, indicating either the presence of cavities or organic carbon that results in
inadequate binding [59].

In most of the work published in the past ten years, Montmorillonite was the filler of choice;
however, clays other than Montmorillonite have been investigated as polyurethane fillers. Table
3.2 lists and compares the empirical formulae, the size, and the d-spacings of clays investigated
for polyurethane reinforcement. Large aspect-ratio clays are often used and considered desirable
because they can efficiently enhance the stiffness at small volume fractions due to their high
surface area-to-volume ratios [60]. However, the probability that the platelets shield the matrix
from straining and negatively affect load transfer is greater when the aspect ratio of the clays is
large [60]. Ultimately, the aspect ratio of the clay, the interactions between the polyurethane and
clay, and the resultant dispersal all influence the mechanical enhancement.
Clay

Empirical Formula

Hectorite

Nao.3Mg 2.7Li0. 3Si 4010(OH) 2

Lateral
dimension, D
[nm]
600

1.53

Laponite
Montmorillonite
Rectorite

Na+.7[(Si 8Mg.5Lio.3)O 20(OH) 4f-0 7


Na 0 .2CaO. 1A12 Si 4010(OH) 2 10H 20
Nao.6Cao.3Ko.A16 Si 6020(OH)4-2H 20

25
70-218
200-300

1
1
1

1.28
1.50
2.22

Saponite

Nao.ICao. 1Mg2. 25Fe2 +0.75Si 3AlO 10(OH) 2 -4H20

165

1.23

Sepiolite

(Si 12Mg8030)(OH) 4(OH 2)4 "8H 20

25

20-1500

1.03

Thickness, t
[nm]

d-spacing
[nm]

Table 2-3: List and comparison of the various clays investigated for polyurethane reinforcement [19, 61-64].

In 1999, Zilg et al evaluated the reinforcement capabilities of synthetic fluoromica (Hectorite)


[65]. They found that the particles modified with bis(hydroxyethyl)-methyl-dodecylammonium
chloride negatively influenced the Young's modulus of the nanocomposites while the
unmodified fillers increased the modulus. From these results the authors inferred that the
delaminated platelets were soft when compared to the original stacked platelets and did not
enhance the nanocomposite stiffness [65]. However, in 2004 Varghese et al effectively
reinforced polyurethane rubber (PUR) latex and PUR/natural rubber latex blends with the same
unmodified Hectorite [66]. In the PUR, specifically, the inclusion of the Hectorite increased the
storage modulus by 1200-1500 MPa at temperatures below the soft segment glass transition

7-fold
temperature (Tg = 100 C). Furthermore, the modulus of the PUR at 100% elongation had a
with 10
increase and at 300% elongation over a 9-fold increase when the material was filled
wt% Hectorite [66].
and
In 2004, Ma and colleagues evaluated the reinforcement properties of Rectorite
Montmorillonite when they were ion-exchanged with three different quaternary ammonium salts
as
and then melt intercalated in a different PUR [62]. Ma et al found that the Rectorite was
effective a filler as Montmorillonite, but that the key to successful reinforcement is choosing the
right quaternary ammonium salt for ion exchange [62].
loaded with
Saponite was used to reinforce a waterborne PU. When the nanocomposite was
4 wt% Saponite, the Young's modulus increased from 55.7 MPa to 126.1 MPa, the strength
increased from 3.9 MPa to 7.6 MPa, the extension increased from 27.7% to 57.2%, and the wear
loss was one-third that of the pure PU [67].
13

, dX
SOW60

Ik

i00

4S3..

27.7

Si
2D 3..9 4.

4.9

,1

10

62
2

Cao.tacf clay (pit

,1
3

0
0

......
3
2
IpWb
h
day .afutg

Figure 2-17: Mechanical properties of the waterborne PU filled with various weight
concentrations of Saponite as shown in 168].

More recently, modified sepiolite, a fibrous or tunnel-like silicate, was dispersed within
the
polyurethane at concentrations of 1, 3 and 5 wt% as shown in Figure 2-17 [64]. As
concentration of sepiolite increased, the tensile strength, modulus and extensibility increased,
indicating strong interactions and stress transfer between the matrix and filler. The addition of
sepiolite also increases the thermal stability. This increases the onset temperature of thermal
degradation but also acts as an insulator and retards volatile product generation [64]. Ultimately,
any clay can be as effective a reinforcing agent as Montmorillonite, if the appropriate swelling
agent or ammonium salt is utilized for ion exchange and the silicates are well-dispersed.

Figure 2-18: Transmission electron micrographs of sepiolite (a) and 3 wt% sepiolite in polyurethane (b) [641.

2.5 Summary
Over the past ten years researchers have investigated numerous polyurethane systems, clay
fillers, chemical modifiers of clay, dispersal techniques, and nanocomposite synthesis
procedures. In general, the previous research suggests that the addition of clay most effectively
stiffens soft polyurethanes rather than hard polyurethanes. This is because the hard domains and
soft segment crystallinity are responsible for polyurethane stiffness. When clay is added to a soft
polyurethane, the rigidity of the polymer is enhanced, but when added to a hard polyurethane, the
previously established hard domain network is disrupted and the rigidity compromised [12]. The
resultant strength and extensibility of a polyurethane nanocomposite increases when the clay is
well dispersed and the clay, the chemical modification, and the polyurethane all interact
favorably. However, if the clay induces crystallinity in the soft segment, the strength and
extensibility are degraded [31, 38, 69]. Furthermore, the water vapor permeability of the
nanocomposites decreases with increasing clay concentration and better clay dispersal [36, 39].
The oxygen permeability, however, decreases only if the polyurethane adheres to the clay
platelets [39]. These investigations show that typically the soft segment glass transition
temperature is unaffected by the addition of clay while the thermomechanical and thermal
degradation behavior is inconsistent and depends heavily upon the interactions between the clay,
chemical modification, and polyurethane. Furthermore, the greatest thermomechanical, thermal,
and barrier enhancements are exhibited in the nanocomposites with the best clay dispersal [60].
These publications also suggest that the reinforcement capability of the clay depends primarily
upon the dispersal of the clay and that the type of clay is of lesser importance [65-67]. The clay

dispersal tends to be best in soft polyurethanes, those containing less than 50 wt% hard segment,
because it is more difficult for the rigid hard segments to diffuse between silicates [33]. While
the melt processing-based dispersal technique may be more environmentally friendly than using
a common solvent, the high temperatures needed for the melt can cause the chemical
modification of the clay and polyurethane to interact unfavorably and deteriorate the mechanical
properties of the polyurethane [42]. If the common solvent technique is used, the clay is best
dispersed when the solution is sonicated [51]. Unfortunately, in all of these publications the
intercalation of the large aspect-ratio clays investigated has persisted and the maximum
thermomechanical, thermal, and barrier enhancements have been exhibited at small clay weight
concentrations.

Ultimately, in order to develop a polyurethane/clay nanocomposite that exhibits a large increase


in stiffness and strength, good dispersal of clay is necessary. To avoid possible deleterious
interactions between clay chemical-modification and polyurethane, a dispersal technique that
does not rely upon the chemical modification of the clay is needed. Furthermore, the
reinforcement capabilities of smaller aspect-ratio clay need to be evaluated. Small aspect-ratio
clays may be more easily dispersed and the thermomechanical enhancements could prove
interesting since smaller aspect-ratio clay is less likely to shield the polyurethane from straining
[60].

2.6 References
[1]
[2]

F. Vollrath and D. P. Knight, "Liquid crystalline spinning of spider silk," Nature, vol.
410, pp. 541-548, 2001.
J. M. Gosline, P. A. Guerette, C. S. Ortlepp, and K. N. Savage, "The mechanical design
of spider silks: From fibroin sequence to mechanical function," JournalofExperimental
Biology, vol. 202, pp. 3295-3303, 1999.

[3]

M. Xu and R. V. Lewis, "Structure of a Protein Superfiber - Spider Dragline Silk,"


Proceedings of the NationalAcademy of Sciences of the United States ofAmerica, vol.

[4]
[5]

87, pp. 7120-7124, 1990.


D. T. Grubb and L. W. Jelinski, "Fiber morphology of spider silk: The effects of tensile
deformation," Macromolecules,vol. 30, pp. 2860-2867, 1997.
L. W. Jelinski, "Establishing the relationship between structure and mechanical function
in silks," Current Opinion in Solid State & Materials Science, vol. 3, pp. 237-245, 1998.

[6]

C. Y. Hayashi, N. H. Shipley, and R. V. Lewis, "Hypotheses that correlate the sequence,


structure, and mechanical properties of spider silk proteins," InternationalJournalof
BiologicalMacromolecules, vol. 24, pp. 271-275, 1999.

[7]
[8]

[9]
[10]
[11]
[12]
[13]
[14]

[15]

[16]

[17]
[18]
[19]

[20]
[21]
[22]

[23]
[24]

[25]
[26]
[27]

J. P. O'Brien, S. R. Fahnestock, Y. Termonia, and K. C. H. Gardner, "Nylons from


nature: Synthetic analogs to spider silk," Advanced Materials, vol. 10, pp. 1185-+, 1998.
Y. Yang, X. Chen, Z. Z. Shao, P. Zhou, D. Porter, D. P. Knight, and F. Vollrath,
"Toughness of spider silk at high and low temperatures," Advanced Materials,vol. 17,
pp. 84-88, 2005.
D. Randall and S. Lee, "The Polyurethanes Book," 2nd ed. New York: Huntsman
International, 2002, pp. 477.
T. Thomson, Polyurethanesas Specialty Chemicals: PrinciplesandApplications. Boca
Raton, Florida: CRC Press, 2005.
M. Michalovic, "The Macrogalleria: A Cyberwonderland of Polymer Fun," 1995.
M. Szycher, Szycher's Handbook ofPolyurethanes.Boca Raton, Florida: CRC Press,
1999.
R. J. Young and P. A. Lovell, Introduction to Polymers. Cheltenham, U.K.: Nelson
Thornes Ltd., 1991.
M. Song, H. S. Xia, K. J. Yao, and D. J. Hourston, "A study on phase morphology and
surface properties of polyurethane/organoclay nanocomposite," EuropeanPolymer
Journal,vol. 41, pp. 259-266, 2005.
C. D. Eisenbach, A. Ribbe, and C. Gunter, "Morphological-Studies Of Model
Polyurethane Elastomers By Element-Specific Electron-Microscopy," Macromolecular
Rapid Communications, vol. 15, pp. 395-403, 1994.
D. J. Martin, G. F. Meijs, P. A. Gunatillake, S. P.Yozghatlian, and G. M. Renwick, "The
influence of composition ratio on the morphology of biomedical polyurethanes," Journal
ofApplied Polymer Science, vol. 71, pp. 937-952, 1999.
P. C. LeBaron, Z. Wang, and T. J. Pinnavaia, "Polymer-layered silicate nanocomposites:
an overview," Applied Clay Science, vol. 15, pp. 11-29, 1999.
J. DeGaspari, "Prospecting Paydirt," in MechanicalEngineeringMagazine, vol. April,
April 2001.
K. Yano, A. Usuki, and A. Okada, "Synthesis and Properties of Polyimide-Clay Hybrid
Films," JournalofPolymer Science PartA-Polymer Chemistry, vol. 35, pp. 2289-2294,
1997.
T. Lan and T. J. Pinnavaia, "Clay-Reinforced Epoxy Nanocomposites," Chemistry Of
Materials, vol. 6, pp. 2216-2219, 1994.
S.S. Ray and M. Okamoto, "Polymer/layered silicate nanocomposites: a review from
preparation to processing," Progress in Polymer Science, vol. 28, pp. 1539-1641, 2003.
Y. S. Lu and R. C. Larock, "Bio-based nanocomposites from corn oil and functionalized
organoclay prepared by cationic polymerization," MacromolecularMaterialsAnd
Engineering,vol. 292, pp. 863-872, 2007.
E. P. Giannelis, "Polymer layered silicate nanocomposites," Advanced Materials,vol. 8,
pp. 29-35, 1996.
C. M. Koo, H. T. Ham, S.O.Kim, K. H. Wang, I. J. Chung, D. C. Kim, and W. C. Zin,
"Morphology evolution and anisotropic phase formation of the maleated polyethylenelayered silicate nanocomposites," Macromolecules, vol. 35, pp. 5116-5122, 2002.
Rockwood Specialties Inc., "Nanoclay.com."
Y. Termonia, "Molecular Modeling of Spider Silk Elasticity," Macromolecules, vol. 27,
pp. 7378-7381, 1994.
T. B. Murphy and A. F. Sawyer, "Advancement of natural and synthetic clays in personal

[28]
[29]

[30]

[31]

[32]
[33]

[34]

[35]

[36]
[37]

[38]

[39]

[40]

[41]

[42]

[43]

care applications," 2004.


Z. Wang and T. Pinnavaia, "Nanolayer Reinforcement of Elastomeric Polyurethane,"
Chemistry of Materials,vol. 10, pp. 3769-3771, 1998.
K. J. Yao, M. Song, D. J. Hourston, and D. Z. Luo, "Polymer/layered clay
nanocomposites: 2 polyurethane nanocomposites," Polymer, vol. 43, pp. 1017-1020,
2002.
T. K. Chen, Y. I. Tien, and K. H. Wei, "Synthesis and characterization of novel
segmented polyurethane clay nanocomposite via poly(epsilon-caprolactone)/clay,"
JournalofPolymer Science Parta-Polymer Chemistry, vol. 37, pp. 2225-2233, 1999.
S. Y. Moon, J. K. Kim, C. Nah, and Y. S. Lee, "Polyurethane/montmorillonite
nanocomposites prepared from crystalline polyols, using 1,4-butanediol and organoclay
hybrid as chain extenders," European Polymer Journal,vol. 40, pp. 1615-1621, 2004.
T. K. Chen, Y. I. Tien, and K. H. Wei, "Synthesis and characterization of novel
segmented polyurethane/clay nanocomposites," Polymer, vol. 41, pp. 1345-1353, 2000.
Y. I. Tien and K. H. Wei, "Hydrogen bonding and mechanical properties in segmented
montmorillonite/polyurethane nanocomposites of different hard segment ratios,"
Polymer, vol. 42, pp. 3213-3221, 2001.
Y. I. Tien and K. H. Wei, "High-Tensile-Property Layered Silicates/Polyurethane
Nanocomposites by Using Reactive Silicates as Pseudo Chain Extenders,"
Macromolecules, vol. 34, pp. 9045-9052, 2001.
Y. I. Tien and K. H. Wei, "The Effect of Nano-Sized Silicate Layers from
Montmorillonite on Glass Transition, Dynamic Mechanical and Thermal Degradation
Properties of Segmented Polyurethane," JournalofApplied Polymer Science, vol. 86, pp.
1741-1748, 2002.
R. Xu, M. Evangelos, A. J. Snyder, and J. Runt, "New Biomedical Poly(urethane urea)-Layered Silicate Nanocomposites," Macromolecules, vol. 34, pp. 337-339, 2001.
R. J. Xu, E. Manias, A. J. Snyder, and J. Runt, "Low penneability biomedical
polyurethane nanocomposites," Journal Of Biomedical MaterialsResearch PartA, vol.
64A, pp. 114-119, 2003.
M. Tortora, G. Gorrasi, V. Vittoria, G. Galli, S. Ritrovati, and E. Chiellini, "Structural
characterization and transport properties of organically modified
montmorillonite/polyurethane nanocomposites," Polymer, vol. 43, pp. 6147-6157, 2002.
M. A. Osman, V. Mittal, M. Morbidelli, and U. W. Suter, "Polyurethane adhesive
nanocomposites as gas permeation barrier," Macromolecules, vol. 36, pp. 9851-9858,
2003.
X. J. Wang, J. J. Kang, Y. P. Wu, and S. B. Fang, "Novel composite polymer electrolytes
based on poly(ether-urethane) network polymer and modified montmorillonite,"
Electrochemistry Communications, vol. 5, pp. 1025-1029, 2003.
Y. W. Chen-Yang, H. C. Yang, G. J. Li, and Y. K. Li, "Thermal and anticorrosive
properties of polyurethane/clay nanocomposites," JournalOf Polymer Research, vol. 11,
pp. 275-283, 2004.
B. Finnigan, D. Martin, P. Halley, R. Truss, and K. Campbell, "Morphology and
properties of thermoplastic polyurethane nanocomposites incorporating hydrophilic
layered silicates," Polymer, vol. 45, pp. 2249-2260, 2004.
B. Finnigan, K. Jack, K. Campbell, P. Halley, R. Truss, P. Casey, D. Cookson, S. King,
and D. Martin, "Segmented polyurethane nanocomposites: Impact of controlled particle

[44]

[45]

[46]

[47]
[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

[56]
[57]
[58]

size nanofillers on the morphological response to uniaxial deformation,"


Macromolecules, vol. 38, pp. 7386-7396, 2005.
B. Finnigan, P. Halley, K. Jack, A. McDowell, R. Truss, P. Casey, R. Knott, and D.
Martin, "Effect of the average soft-segment length on the morphology and properties of
segmented polyurethane nanocomposites," JournalofApplied Polymer Science, vol. 102,
pp. 128-139, 2006.
B. Finnigan, P. Casey, D. Cookson, P. Halley, K. Jack, R. Truss, and D. Martin, "Impact
of controlled particle size nanofillers on the mechanical properties of segmented
polyurethane nanocomposites," InternationalJournalOfNanotechnology, vol. 4, pp.
496-515, 2007.
J. K. Mishra, I. Kim, and C. S. Ha, "Heat shrinkable behavior and mechanical response of
a low-density polyethylene/millable polyurethane/organoclay ternary nanocomposite,"
MacromolecularRapid Communications,vol. 25, pp. 1851-1855, 2004.
F. Cao and S. C. Jana, "Nanoclay-tethered shape memory polyurethane nanocomposites,"
Polymer, vol. 48, pp. 3790-3800, 2007.
C. H. Dan, Y. D. Kim, M. H. Lee, B. H. Min, and J. H. Kim, "Effect of solvent on the
properties of thermoplastic polyurethane/clay nanocomposites prepared by solution
mixing," JournalOfApplied Polymer Science, vol. 108, pp. 2128-2138, 2008.
X. Y. Meng, X. H. Du, Z. Wang, W. G. Bi, and T. Tang, "The investigation of exfoliation
process of organic modified montmorillonite in thermoplastic polyurethane with different
molecular weights," Composites Science And Technology, vol. 68, pp. 1815-1821, 2008.
H. L. Tan and J. Nie, "Photopolymerization and characteristics of
polyurethane/organoclay nanocomposites," MacromolecularReaction Engineering,vol.
1, pp. 384-390, 2007.
I. Rhoney, S. Brown, N. E.Hudson, and R. A. Pethrick, "Influence of processing method
on the exfoliation process for organically modified clay systems. I. Polyurethanes,"
JournalofApplied Polymer Science, vol. 91, pp. 1335-1343, 2004.
R. Krishnamoorti and R. A. Vaia, "Polymer Nanocomposites: Synthesis,
Characterization, and Modeling," 1st ed. Washington, D.C.: Oxford University Press,
2002,pp. 242.
M. Song and K. J. Yao, "X-ray diffraction detection of compliance in polyurethaneorganoclay nanocomposites," MaterialsScience And Technology, vol. 20, pp. 989-992,
2004.
H. S. Xia and M. Song, "Intercalation and exfoliation behaviour of clay layers in
branched polyol and polyurethane/clay nanocomposites," Polymer International,vol. 55,
pp. 229-235, 2006.
M. M. Rahman, J. H. Kim, and H. D. Kim, "Characterization of waterborne
polyurethane/clay nanocomposite adhesives containing different amounts of ionic
groups," JournalOfAdhesion Science And Technology, vol. 21, pp. 1575-1588, 2007.
K. C. Khemani, "Polymeric Foams: Science and Technology," 1st ed. Washington, D.C.:
American Chemical Society, 1997, pp. 236.
X. Cao, L. J. Lee, T. Widya, and C. Macosko, "Polyurethane/clay nanocomposites foams:
processing, structure and properties," Polymer, vol. 46, pp. 775-783, 2005.
L. Song, Y. Hu, Y. Tang, R. Zhang, Z. Y. Chen, and W. C. Fan, "Study on the properties
of flame retardant polyurethane/organoclay nanocomposite," Polymer DegradationAnd
Stability, vol. 87, pp. 111-116, 2005.

[59]

[60]

[61]
[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

S. Rajendran, R. G. Gilbert, B. Cribb, R. Truss, B. Noller, A. Stewart, and M. Zalucki,


"Effect of inert fillers on diffusion and controlled release of penetrant in polymer
matrices," Journal Of Applied Polymer Science, vol. 108, pp. 3593-3600, 2008.
N. Sheng, M. C. Boyce, D. M. Parks, G. C. Rutledge, J. I. Abes, and R. E. Cohen,
"Multiscale micromechanical modeling of polymer/clay nanocomposites and the
effective clay particle," Polymer, vol. 45, pp. 487-506, 2004.
D. Barthelmy, "Mineralogy Database," 2004.
X. Y. Ma, H. J. Lu, G. Z. Liang, and H. X. Yan, "Rectorite/thermoplastic polyurethane
nanocomposites: Preparation, characterization, and properties," JournalOfApplied
Polymer Science, vol. 93, pp. 608-614, 2004.
C. Zilg, F. Dietsche, B. Hoffmann, C. Dietrich, and R. Mulhaupt, "Nanofillers based
upon organophilic layered silicates," MacromolecularSymposia, vol. 169, pp. 65-77,
2001.
H. X. Chen, M. S. Zheng, H. Y. Sun, and Q. M. Jia, "Characterization and properties of
sepiolite/polyurethane nanocomposites," MaterialsScience And EngineeringA-Structural
MaterialsPropertiesMicrostructureAnd Processing,vol. 445, pp. 725-730, 2007.
C. Zilg, R. Thomann, R. Mulhaupt, and J. Finter, "Polyurethane nanocomposites
containing laminated anisotropic nanoparticles derived from organophilic layered
silicates," Advanced Materials,vol. 11, pp. 49-52, 1999.
S. Varghese, K. G. Gatos, A. A. Apostolov, and J. Karger-Kocsis, "Morphology and
mechanical properties of layered silicate reinforced natural and polyurethane rubber
blends produced by latex compounding," JournalOf Applied Polymer Science, vol. 92,
pp. 543-551, 2004.
H. C. Kuan, W. P. Chuang, and C. C. M. Ma, "Synthesis and characterization of a
clay/waterborne polyurethane nanocomposite," Journal Of MaterialsScience, vol. 40, pp.
179-185, 2005.
H. C. Kuan, C. C. M. Ma, W. P. Chuang, and H. Y. Su, "Hydrogen bonding, mechanical
properties, and surface morphology of clay/waterborne polyurethane nanocomposites,"
JournalOfPolymer Science PartB-Polymer Physics, vol. 43, pp. 1-12, 2005.
A. M. Chen, Y. Tian, B. Han, G. D. Ji, S. S. Wu, and J. Shen, "Synthesis and
characterization of polyurethane/montmorillonite nanocomposites," Acta Polymerica
Sinica, pp. 591-594, 2003.

3. Nanoparticle Dispersal via Solvent Exchange Processing


[Part of this work has been published previously, in slightly different form, in "High Performance Elastomeric
Nanocomposites via Solvent Exchange Processing" by S.M. Liff, N. Kumar, and G.H. McKinley, Nature Materials,
2007, 6 (1): 76-83, "Preferential Association of Segment Blocks in Polyurethane Nanocomposites" by L.T.J. Korley,
S.M. Liff, P.T. Hammond, and G.H. McKinley, Macromolecules, 2006, 39 (20): 7030-7036, and in "Method to
disperse and exfoliate nanoparticles" by N. Kumar, S. M. Liff & G.H. McKinley, U.S. Patent Application
11/253,219 (Filed October 18, 2005). ]

3.1. Introduction
High performance biomaterials, such as silk [1-5] and bone [6, 7] exhibit unparalleled
combinations of stiffness, strength, extensibility, and toughness by exploiting hierarchical
structures in which stiff nanometer-size crystallites are embedded and dispersed in softer protein
matrices [1-3, 8]. Polymer nanocomposites which utilize engineering thermoplastics as the
matrix material and stiff, inorganic, anisotropic nanoparticles to tailor thermomechanical
properties have been the subject of many recent studies [9, 10]. Optimal mechanical
enhancement of polymer matrices via nanoreinforcement is expected only when the
nanoparticles are fully exfoliated in the polymer matrix. In the present work we outline a novel
solvent exchange approach to efficiently exfoliate nanoparticles, particularly synthetic smectic
clays, such as Laponite RD (Southern Clay Products, Inc.) in polymers like commercial
thermoplastic

polyurethane

(TPU) elastomers,

specifically

Elasthane TM

80A (Polymer

Technology Group, Inc.)

3.2 Previous Dispersal Techniques


Polymer nanocomposites have the potential to fulfill many high performance applications which
require enhanced mechanical, thermal, electrical, barrier and/or optical properties, as is explained
in the two previous Chapters. However, the potential for maximum enhacement occurs when the
nanoparticles are fully exfoliated within the polymer matrix. For numerous nanoparticle systems,
specifically clay particles, dispersion is a two part process--1) exfoliation of stacked clay
particles, and 2) dispersion of the separated particles within the matrix. Incomplete exfoliation
and dispersion results in nanocomposites with little or no improvement in the mechanical,
thermal, electrical, barrier, and/or optical properties [11-15].

Full exfoliation of nanoparticles, specifically smectic clay particles in polymers is a major


challenge. In the dry form, smectic clays exist as agglomerates of stacked platelets tightly bound
to each other via strong ionic and van der Waals forces [16]. The current methods to exfoliate
nanoparticles in engineering thermoplastics have been only partially successful. Exfoliation in
high molecular weight and/or strongly hydrophobic polymers that are insoluble in water has
proven to be particularly difficult [17, 18]. Typical methods [19] include (i) compounding and
melt extrusion of the polymer melt with unmodified and organically modified clay and (ii)
dispersion of organically modified clay in organic solvent followed by in-situ polymerization
[20] or dissolution of the matrix polymer followed by solvent casting or precipitation. Organic
modification of the particle surface has been found to be necessary to increase the affinity of clay
with hydrophobic polymers or organic solvents [21]; however, disadvantages such as premature
degradation of the polymer and/or decreased thermomechanical properties have been reported
[21]. Even with organic modification, good exfoliation remains a challenge [21-23] as previously

reported in Chapter Two and further discussed here.


3.2.1 Dispersion Via Melt and Mixing
Dispersion via melt and mixing describes generally those nanoparticle dispersion techniques that
rely upon application of high temperatures and in most cases additional mechanical work input
(such as extruding or compounding). The nanoparticles may be chemically modified to promote
favorable interactions with the polymer matrix; however, selection of a modifier that does not
degrade at the high processing temperatures is critical. This technique is advantageous because it
is environmentally friendly; hazardous solvents (which are used in the other methods below) are
not needed. Consequently, numerous thermoplastic polymers may be processed in this way with
a wide variety of nanoparticles. Significant drawbacks still remain. When the matrix polymer
and reinforcing particle do not interact favorably e.g. a polarity mismatch, large quantities of
mechanical work are needed to disperse the particles. Not only is this extra work costly, but too
much deformation of the polymer may cause cleavage of the polymer chains and an irreversible
reduction in the molecular weight, resulting in inferior mechanical properties. Another drawback
is that the concentration of nanoparticles is limited to concentrations below or near the
percolation threshold. Otherwise, not only does the jammed particle network prevent further
particle migration within the matrix, but jamming can cause instruments used to apply

deformation, such as a twin-screw extruder, to fail. These jamming concentrations may be quite
small (-2 vol%) for large aspect ratio particles.
3.2.2 In-Situ Polymerization
In-situ polymerization as a method to disperse nanoparticles is necessary in many thermoset
polymers, but may also be used with thermoplastics. Again, chemical modification of the
nanoparticle is not always necessary with this technique, but often helps target the particle
placement within the polymer. The nanoparticles with or without modification can be swollen
within the monomer. The particles can then be used either as the catalyst, as cross-linking sites or
chain extenders, or as reinforcement embedded within the non-cross-linked phase [24, 25].
Meanwhile, polymerization can be induced via heat, radiation, diffusion of a suitable initiator, or
catalysis. In-situ polymerization was first successful and demonstrated when Toyota developed
Nylon-6 and Montmorillonite composites with enhanced thermal and mechanical properties [26].
While much of the chemistry needed to develop such systems is published, reproducible
synthesis of large batches of materials for characterization can be difficult. This is due to
inconsistencies that arise during the scale-up of polymerization processes that include nano-filler,
e.g. impurities, nucleation sites, and kinetic limitations.
3.2.3 Common Solvent Process
A common solvent that disperses the nanoparticles and in which the polymer or pre-polymer is
soluble is the third method often utilized to exfoliate nanoparticles. Chemical modification of the
nanoparticles is often necessary in order for the same solvent that solubilizes the matrix polymer
to also disperse the particles. The resulting nanocomposites can then be prepared by solvent
casting or precipitating out the molecularly-mixed solids and then compression molding or melt
compounding the precipitate. Obviously, the limiting factor is determination of a good solvent or
chemical modifier that does not have a deleterious impact on the morphology or thermo-physical
properties of the matrix polymer. Mackay et al have also shown that it is considerably easier to
disperse nanoparticles within linear polymer solutions when the radius of gyration of the
polymer is larger than the radius of the nanoparticle because there is an enthalpy gain when more
molecular contacts become available [27].
Each of the methods discussed above is not consistently reliable for multiple polymernanoparticle pairs, nor are the resulting modifications inexpensive.

For example, chemical

modification of nanoparticles may result in the early onset of thermal degradation of the

polymer.

Furthermore, the modifying agent may be incompatible with and alter the matrix

polymer morphology. The extra processing steps needed for dispersal and exfoliation, like
monomer modification, adds to production costs.

Some of these methods are of limited

usefulness because of the limited number of solvents that can both dissolve polymers and
disperse nanoparticles.

As a result, a new method to effectively and uniformly disperse

nanoparticles in polymers, and consistently improve the mechanical, thermal, electrical, barrier,
and/or optical properties of composites is desirable.

3.3 Solvent Exchange Approach


This new dispersal technique does not rely upon chemical modification, but exploits the affinity
of unmodified polar nanoparticles, like Laponite for the polar segments of block copolymers, i.e.
the hard micro-domains of polyurethanes, like Elasthane. Laponite powder can be dispersed in
many organic, polar solvents due to its hydrophilic character which promotes the separation of
clay platelets via osmotic pressure produced by the hydration of intercalated ions [20]. In nonpolar solvents, the large energy barrier created by the reduced wettability and the absence of
osmotically driven platelet separation prevents direct dispersion of Laponite. A novel solvent
exchange method [28] was developed to disperse charged discotic nanoparticles, such as
Laponite or Montmorillonite, in organic solvents. This procedure requires the use of two
solvents, denoted A and B, that meet four criteria; (i) solvent A must fully disperse the
hydrophilic clay, (ii) solvent B must dissolve the polymer matrix and not cause re-aggregation of
the nano-clay, (iii) solvents A and B must be fully miscible, and (iv) solvent B must have a higher
boiling point than solvent A. Solvents with high dipole moments and dielectric constants were
found to be the most suitable B solvents. Once the nanoparticle-solvent pair is prepared, solvent
A may be removed via evaporation and the desired polymer subsequently dissolved in the
nanoparticle-solvent B solution. Once the nanoparticle-solvent B-polymer solution is prepared
the polymer nanocomposites can either be prepared via slow solvent casting or fast precipitation
and compression molding.
Solvent A may include aqueous or polar solvents or solvent mixtures. A particularly suitable
solvent A is water, but alcohols and other volatile polar solvents in which the particles are
dispersible may be used as well.

Other solvent A possibilities include methanol, ethanol,


chloroform, and ethylene glycol among others [28]. Salts, surfactants, and/or other additives

may be utilized to increase the solubility of the particles or optimize another property of the
dispersion, e.g. ionic strength, pH, or viscosity. Various salts, including sodium chloride and
sodium citrate, may be added to the dispersion to change its ionic strength. Likewise, ionic or
non-ionic surfactants may be employed. Alternatively or in addition, acids or bases, such as
potassium hydroxide, sodium hydroxide, or sulfuric acid, may be added to the dispersion to
change its pH [28]. Low molecular weight polyethylene glycol or polypropylene glycol may
also be added and used to modify the viscosity of the nanodispersion.

The concentration of nanoparticles added to form the nanoparticle-solvent A solution varies from
one system to the next. The particle shape and dispersity are significant factors governing the
concentration limit.

The theoretical maximum volume fraction achievable for the random

packing of monodisperse spherical particles is 0.56 [29]. However, higher concentrations may be
achieved when bimodal regularly shaped particles are utilized [30, 31]. More irregularly shaped
particles will jam, or form a percolated network at low concentrations, depending on their aspect
ratio. The dispersion of the nanoparticles and the ability of the nanoparticles to remain in a
stable and non-aggregated state is a result of the surface charge, surface chemistry, and
wettability of the nanoparticles, and the ion solvation, dipole moment and dielectric constant of
solvent A [28]. Chemical modification is not necessary before dispersion of the nanoparticles in
solvent A but may be utilized, if necessary.
Solvent B is a second solvent or solvent mixture that when added to the nanoparticle-solvent A
solution does not disrupt the nanoparticle dispersion or stability. Solvent B may be, but is not
limited to, n-methyl pyrrolidinone (NMP), dimethylacetamide (DMAC), dimethylformamide
(DMF), dimethylsulfoxide (DMSO), dichlorobenzene, or ethylene glycol.

For polyurethane-

Laponite composites, dimethylacetamide and dimethylformamide are excellent B solvents. In


some instances, solvent B could be a polymerizable reagent. Consequently, it may be desirable to
agitate or heat the dispersion as solvent B is added. Solvent B may be added to solvent A in
practically any ratio, as long as when solvent A is removed the resultant concentration of
nanoparticles in solvent B does not result in sample gelation-as a consequence the matrix
polymer will not dissolve without thermal energy or mechanical work.

3.3.1 Method to Disperse Laponite in Thermoplastic Polyurethanes


In advance of preparing the polyurethane-Laponite thin-film nanocomposites, the Laponite had
to be completely dispersed and the agglomerated platelet stacks delaminated without chemical
modification. This is done by hydrating the Laponite with deionized water at concentrations less
than 1 wt% Laponite: at larger concentrations the solution forms a gel. However, the matrix
polyurethanes investigated were not soluble in water, so a process was needed to disperse the
Laponite in a polyurethane solution.

For the polyurethane-Laponite systems discussed in this thesis, the optimal solvent pair
determined was deionized water (DI; A) and dimethylacetamide or dimethylformamide (DMAc
or DMF; B), depending on the ease of solubility of the polyurethane in solvent B.

When

Elasthane or custom synthesized polyurethanes were utilized solvent B was DMAc; the shape
memory polyurethane discussed in Chapters Six and Seven employed DMF as solvent B. The
following method details the procedure for Laponite dispersal when Elasthane is utilized, but the
procedure remains nearly the same when a different polyurethane and DMF is utilized.
Initially, Laponite is mixed with deionized water at a concentration of -0.70 wt% and the
solution is stirred with a Teflon-coated, magnetic stirbar on a hot plate stirrer at a moderate stir
speed for at least two days. Then, solvent B was added to the solution so that if all the water were
removed, the weight fraction of Laponite would be between 0.50% and 0.65%. Again the
solution was stirred at a moderate speed for two to three days. Then the liquid-vapor phase
equilibrium of the binary solvent system was utilized in order to evaporate off all the water from
the Laponite-H 2 0-DMAc solution.
From examination of the liquid-vapor phase equilibrium curves of DMAc and H2 0 separately, it
was clear that at any temperature, the vapor pressure of DMAc was an order of magnitude less
than H20. For example, at 75'C, the vapor pressure of DMAc is 4.2 kPa while the vapor pressure
of H20 is 38.6 kPa as shown in Figure 3-1 [32]. This difference could be exploited to remove all
the water from the solution if the liquid-vapor phase equilibrium curve of the binary system did
not exhibit significant hydrolysis or the formation of an azeotrope. In 1972, Carli and colleagues
investigated two binary solvents pairs-acetic acid-DMAc and H20-DMAc at two different

pressures and varied the temperature. They found that the H20-DMAc system displayed only a
slight negative deviation from Raoult's Law and unlike the other system, no azeotrope formation
0
[33]. Consequently, we assumed that at a constant temperature of 75 C the vapor pressure would
change linearly from 38.6 kPa with 100% H20 to 4.2 kPa at 100% DMAc and used a rotary

evaporator (Rotovap) to remove the water.


1U

.4

10
160

140

1 2
101

120
0

IL

100

100

80

101

rn

A-2

-100

100

200

300

Temperature rC]

400

0.0

0.2

0.6
0.4
mole fraction H2 0

0.8

1.0

Figure 3-1: Vapor-Liquid equilibrium curve for H20 and for DMAc (left) and vapor-liquid equilibria of H 20DMAc determined via experiment in [33].

The flask containing the Laponite-H 20-DMAc solution with the spin bar was lowered into a hot
water bath at 75 0 C and connected to the Rotovap at the attachment point. The flask was rotated
at 85 rpm. This rotation and the inclusion of the spin bar, which acted as a bubble nucleation
point, ensured that the solution did not char or thermally decompose. A vacuum was applied and
the pressure immediately set at 25.0 kPa. At this point, in a solution containing 40% DMAc,
boiling begins and water vapor begins to rise up toward the condenser, where the condensate is
then caught in a second flask. The vacuum was increased very slowly until the pressure
decreased to less than 4.0 kPa. The small incremental changes in pressure ensured that no
splashing occurred and no DMAc or Laponite was lost in the Rotovap attachment. When all
water appeared to be removed the mass of the flask was measured. The evaporation step was not
considered to be complete until the mass removed from the flask through evaporation was
greater than or equal to the mass of the H20 added initially. Typically the pressure was further
reduced to 3.0 kPa as a precautionary measure to ensure all H20 was removed. When

dimethylformamide was utilized the pressure was drawn down to 5.5 kPa because the vapor
pressure of DMF at 75C is 7.9 kPa [32].

To ensure complete removal of the water using the Rotovap it was necessary to remove some
DMAc as well. However, it was important not to remove so much DMAc that the concentration
of Laponite in the DMAc solution was greater than 0.70 wt%. At concentration higher than this it
was found that the resultant solution does not reliably dissolve the polyurethane. When
concentrations greater than 0.70 wt% were prepared, heat was necessary to dissolve the
polyurethane in solution. Occasionally, even after a few hours at an elevated temperature, even
as high as 140 0 C, the solution appeared to be emulsion-like and cloudy rather than completely
dissolved and the resultant films were lumpy and discontinuous. This emulsion-like state may
indicate that at Laponite weight fractions greater than 0.007 in DMAc, the system is more of a
gel rather than a solvent dispersion. To avoid such situations, concentrations below 0.70 wt%
were utilized.

This DMAc-Laponite solution was then mixed with various quantities of pure DMAc and
approximately 0.5 g of polyurethane in 60 ml, thick-walled, glass vials. The total amount of
solution necessary was determined by the mass of the polyurethane in the vial and the desired
polyurethane concentration in solution, which was approximately 1 wt%. The final concentration
of nanoparticle has a dramatic affect on the resulting composite material properties and therefore
was kept constant from one nanocomposite to the next. Nanocomposites containing, 0, 1, 2, 4, 6,
8, 10, 15, and 20 wt% Laponite in Elasthane were made. The amount of DMAc-Laponite
mixture was determined by the desired Laponite concentration in polyurethane. The amount
needed could be calculated by solving the following equation
m

CA
CA + mpu

(3.1)

where p,m specifies the weight fraction of Laponite in the PU desired, C is the resultant weight
fraction of Laponite in the DMAc-Laponite mixture, mpu is the mass of the polyurethane, and A
is the mass of the DMAc-Laponite mixture needed. The mass of pure DMAc needed is
determined by taking the difference between the total mass of solvent needed less the mass of the
of the DMAc-Laponite mixture needed. The resultant PU-DMAc-Laponite mixture was then

rolled on a Wheaton mixer for 24 hours or until the PU was completely dissolved. In a few
instances when the concentration of Laponite in DMAc was excessively high (> 0.70 wt%) or
when custom synthesized polyurethanes were used it was necessary to heat the solutions to
140 0 C for 3-6 hours. After complete dissolution the solution was sonicated at room temperature
for 60 minutes. At this point, each vial contained a transparent PU-DMAc-Laponite mixture in
which the Laponite was fully dispersed.

3.4 Proof of Exfoliation


3.4.1 Materials
Biomedical polyetherurethane, Elasthane TM 80A, was obtained from the Polymer Technology
Group, Inc. (Berkeley, CA). Elasthane contains poly(tetramethylene oxide) (PTMO) soft
segments and 4,4'-methylene bisphenyl diisocyanate (MDI) hard segments chain extended with
1,4-butanediol (BDO), as depicted in Fig 3-2. The hard-to-soft segment ratio of Elasthane 80A is
-3
40:60 wt%, its molecular weight is 290,000 g mol' , and its density is 1.12 g cm [34].

o
cH2N-C-N

NH"C(C4 H8)"

Hard segment

Figure 3-2: Structure of Elasthane 80A based on available non-proprietary information. Elasthane 80A is an
elastomeric block-copolymer composed of MDI-BDO hard segments and PTMO soft segments. Hard
segments (length 3-11 nm[35, 361) phase segregate to form hard crystalline domains via hydrogen
bonding[211.
-3
Laponite RD, a synthetic smectic clay with a density of 2.65 g cm [37] was obtained from
0
Southern Clay Products, Inc. (Gonzales, TX). Dimethylacetamide (DMAc; B.P. 165 C) was

obtained from Sigma Aldrich Co. (St. Louis, MO) and used as received. Laponite was dispersed
in DMAc via the solvent exchange procedure detailed above. The Elasthane concentration in the
solution was maintained at 1.4-1.6 wt%. Elasthane/Laponite nanocomposite films with 120-150
pim thicknesses were prepared in Teflon containers (6 cm x 4 cm) by controlled evaporation of
DMAc. The final Laponite concentration in the dry film was controlled by varying the Laponite
concentration in the Laponite-Elasthane-DMAc solution according to eq 3.1. Note, for a discotic
clay such as Laponite, the percolation threshold,

p, can be approximated by dividing the

random-close-packed hard sphere percolation volume fraction (-0.64) by the aspect-ratio, a, of

the clay. The evaporation rate of DMAc was maintained by keeping the films in a closed 60'C
oven with a 0.02 m 3 hr- N2 purge flow.
3.4.2 Dispersal Characterization
3.4.2.1 Transmission Electron Microscopy
The 10 wt% Elasthane-Laponite nanocomposite transmission electron microscope (TEM) sample
was cut on a RMC MT-X Cryo-capable Ultramicrotome with a diamond knife at -170'C and
collected on 400 mesh copper grids. The other polyurethane-Laponite nanocomposite TEM
samples were prepared in a JEOL JEM 9310 focused ion beam (FIB) instrument. These samples
were first sputter-coated with -200 nm of gold, and then a localized -1 p.m thick carbon
protective film was deposited atop the selected area for lamella preparation. The samples were
milled and polished in the FIB to give lamella measuring 10 jim by 10 pm by 80 nm. These
samples were transferred to the copper grids using a micromanipulation system which consisted
of a position-controlled polished glass rod that allowed for electrostatic pick-up of the lamella.
The Elasthane lamellae filled with 4 wt% and 20 wt% Laponite as well as the custom
synthesized polyurethane composites described in 3.5.2 were observed in a JEOL 2010 TEM
containing a LaB6 filament and imaged using a Gatan digital camera. The TEM image of the
nanocomposite containing 10 wt% Laponite was obtained using a JEOL 200CX operated at 200
kV.
3.4.2.2 Atomic Force Microscopy
On a separate occasion, atomic force microscope (AFM) phase images of the nanocomposite was
taken in tapping mode on the 10 wt% Laponite cryo-microtomed TEM sample at a different time
with a unique NanoProbe (amplitude set point = 0.85 Volts, scale = 20').
3.4.2.3 Wide Angle X-ray Diffraction
Wide angle x-ray diffraction (WAXD) spectra from 100 to 80' were obtained using a Rigaku
RU300 with an 18 kW rotating CuKa anode x-ray generator, a 185 mm diffractometer, and a
scintillation counter, in the Bragg-Brettano mode. WAXD spectra from 2' to 380 were obtained
using a Molecular Metrology small-angle X-ray scattering system with a custom-machined
attachment so that WAXD d-spacings could be obtained.
3.4.3 Results and Discussion
The efficacy of the solvent exchange approach for exfoliating the Laponite is demonstrated in the
TEM micrographs, AFM phase images, and WAXD data of the resultant Elasthane/Laponite
composites presented in Figures 3-3 and 3-4. The TEM images show exfoliated and randomly
oriented single Laponite platelets.

--

nm

7 12
S8
0 4

0
o

10

2A

20
J3
fvrianmzo1

Figure 3-3: a, b, & c, Transmission electron micrographs (TEM) of nanocomposites containing 4, 10, and 20
wt% Laponite show fully exfoliated, uniformly dispersed and randomly oriented, Laponite platelets. Above
the percolation concentration (5p = 5.9 wt%) jamming of Laponite platelets is observed. In the 95= 20 wt%
micrograph some instances of intercalation are evident with a 2 nm spacing between platelets. d, Atomic
Force Microscope (AFM) phase image of the surface of the same 9 = 10 wt% TEM sample shows stiff, single
Laponite particles (bright fields) cut at different orientations. Image width is 383 nm. e, Wide angle x-ray
diffraction (WAXD) spectra of nanocomposites containing 0, 4, 10 and 20 wt% Laponite and of Laponite
powder[38] (from bottom to top).

As the Laponite

concentration,

0, exceeds the percolation threshold,

p = 5.9 wt%

(or p z 2.5 vol%), particle-particle interactions become increasingly important during the slow
solvent evaporation and the instances of intercalated platelets with a spacing of 2 nm, however
infrequent, increase. The percolation threshold, Op, represents the concentration at which the
average inter-particle spacing is less than one particle diameter. The d001 spacing of Laponite is
1.28 nm (20 = 6.92') according to Southern Clay Products, Inc. the provider. This diffraction
shoulder is observed in the 0 = 10 and 20 wt% samples with increased intensity at a spacing of

2.02 nm (20 = 4.36'), indicating the instances of Laponite intercalation increase as surpasses

4.

The incomplete exfoliation at 0 > Op is a consequence of the increased Laponite volume


fraction and the slow solvent evaporation which results in weak particle-particle interactions [17,
21] (See Figure 3-3e and 3-4). The peak at 20 = 20' corresponds to hard segment crystallites
with lattice spacing of 4.43 A [21, 38].
4
r n

.i12

S8

4
L

ZO

0
10

20

30

40

50

60

70

80

20B [degrees]
Figure 3-4: Laponite dispersal within the polyurethane matrix via WAXD, using a different diffraction
instrument and different nanocomposite samples (Dispersal Characterization). Spectra from 20 = 100 to 800
of nanocomposites containing 0, 4, 10 and 20 wt% Laponite and of Laponite powder [391 (from bottom to
top).

Because Laponite has only weak reflections due to imperfect platelet registration and its small
size, the broad diffraction halo of the Elasthane would mask the peak at 20 = 19.5'. Although no
large angle peaks are observed in the 0 < 20 wt% samples, the Laponite peak at 27.40 is observed
in the 0 = 20 wt% spectra shifted to 26.40. At this high Laponite volume fraction, particle-

particle interactions developed over the long time allowed for slow solvent removal cause the

platelets to re-aggregate in an intercalated state. Note, the WAXD spectra of numerous


nanocomposite samples at various Laponite concentration have been examined using numerous
WAXD instruments, including the GI beamline at the Cornell high-energy synchrotron source
(CHESS) [40], and are broadly consistent. This will be discussed in Chapter Five. In numerous
instances at Laponite concentrations above the percolation threshold, but below 20 wt%
(gp < 0 < 20 wt%), nanocomposite spectra did not exhibit any peaks or shoulders corresponding
to the pure Laponite spectra, indicating that the platelets were well dispersed and no longer in
agglomerated stacks.

3.5 Influence of Processing Parameters


The solvent exchange method is an effective way to disperse nanoparticles in solution and
exfoliate nanoparticles in a solid polymer matrix, as is detailed in the previous section. The
exfoliation and resultant property enhancement can be further tailored by exploiting polymer
polarity and processing time.
3.5.1 Influence of Segment Polarity
To verify that segment polarity and hydrophilicity determine the location of preferential
reinforcement in either the soft or hard domain, thermoplastic polyurethane/Laponite
nanocomposites were prepared using two custom-synthesized thermoplastic polyurethanes
(TPUs) that differed in soft segment (SS) polarity and hydrophilicity[41].
3.5.1.1 Materials
Two custom synthesized polyurethanes were synthesized and differ from Elasthane in
composition. One is poly(ethylene oxide) based while the other is poly(tetramethylene oxide)
based. They were each filled with 10 wt% Laponite as discussed previously using our solventexchange process. The PEO-based polyurethane contains a 1,6-hexamethylene diisocyanate1,4-butanediol (HDI-BDO) hard segment (33 wt%) and a poly(ethylene oxide)-poly(propylene
oxide)-poly(ethylene oxide) (PEO-PPO-PEO) soft segment (1900 g mol - ) [42].

4
v

hard segment

Figure 3-5: Structure of PEO-based polyurethane [42].

The total weight-average

molecular weight of this crystalline, aliphatic hard segment

polyurethane was determined to be 110,000 g mol -' using gel permeation chromatography. This
polyurethane displays a soft segment glass transition at -61 0 C, a hard segment melting transition
at 146 0 C, and a hard segment crystallization transition at 108 0 C [42]. Like Elasthane, the novel
PTMO-based polyurethane contained poly(tetramethylene oxide) soft segments but the soft
segment was of higher molecular weight (2000 g mo1-) and was more crystalline at room
temperature. Note soft segment crystallinity is dependent upon segment molecular weight [43].
The hard segment (37 wt%) is composed of HDI chain extended with BDO, as in the PEO-based
polyurethane [44].

HO-

28

soft segment

N
H

HN

4
hard segment

Figure 3-6: Structure of PTMO-based polyurethane [44].

The total weight-average molecular weight of this polyurethane is 233,000 g mo'-1. The HDIPTMO PU displays a soft segment glass transition at -65 0 C, a hard segment melting transition at
174 0 C, and a hard segment crystallization transition at 1460 C [44].

The following section details the characterization results of these two custom-synthesized
polyurethane-Laponite nanocomposites. These composites were characterized in the same
manner as the Elasthane-Laponite nanocomposites described in 3.4.2.
3.5.1.2 Evidence ofExfoliation
Both wide angle x-ray diffraction (WAXD) and transmission electron microscopy (TEM)
provide evidence that the Laponite is well dispersed throughout the custom-synthesized
polyurethanes, further supporting the efficacy of the novel solvent exchange procedure.

7 1ntorclato

IEx'oo]

adRcutIo.n'

io

Figure 3-7: (a) TEM image of 10 wt % Laponite dispersed in PTMO:HDI-BDO PU; 50 nm scale bar. (b)TEM
image of 10 wt % Laponite dispersed in PEO-PPO-PEO:HDI-BDO PU; 50 nm scale bar. (c) TEM image of
10 wt % Laponite dispersed in PEO-PPO-PEO:HDI-BDOPU; 20 nm scale bar.

However, the Laponite appears to be fully exfoliated in the HDI-PTMO PU system, while in the
PEO-based system there is evidence of some platelet flocculation. Although the clay is randomly
dispersed throughout the PEO-based matrix, regions of more closely packed Laponite platelets
are observed as well as regions of more even, random dispersal. The HDI-PTMO PU
nanocomposite does not display phase separated regions of clay dispersal, but instead even,
random dispersal throughout the image. The differences in the Laponite dispersal can be seen in
the TEM micrographs displayed in Figure 3-7.
WAXD spectra between a scattering angle of 5' and 80' further confirm the extent of Laponite
dispersion observed in the TEM images. The WAXD spectra of the pure HDI:BDO hard
segments of these polyurethanes, the pure polyurethanes, and the nanocomposites filled with 10
wt% Laponite, as well as the pure Laponite are shown below in Figure 3-8.

d-spacing [nm]

12

d00

11---

cic d

dd

Pure PU

--- 10 wt% Laponite


-

.10

100 wt% Laponite

9-

Lap

N
E

7u

HDI-PEO-PPO PU-

HDI-PTMO PU-

3-

HDI:BDO

10

20

30 40 50 60
20B [degrees]

last ane

70

80

10 wt %
Figure 3-8: WAXD spectra of HDI:BDO hard segment, of pure Laponite and of the pure and
Laponite-filled polyurethane nanocomposites.

There is virtually no difference in the pure HDI- PTMO PU and corresponding nanocomposite
6.92o) and d020
spectra above 50. The absence of the Laponite basal plane spacing peak (20 =

reflection (20 = 19.50) in the HDI-PTMO PU nanocomposite spectrum indicates that the platelets
have been separated and the gallery spacing increased to a random undetectable distance with
this apparatus, indicating full exfoliation. However, there are differences between the spectra of
pure PEO-based polyurethane and its corresponding nanocomposite. In the PEO-based
If
polyurethane-Laponite nanocomposite spectrum there is a small peak at approximately 100.
the clay was intercalated and the gallery spacing increased, the peak in the scattering angle at
to
6.90, indicating the Laponite basal plane spacing when the platelets are stacked, should shift
lower angles. Consequently, the presence of a peak at a higher scattering angle implies either that
the clay platelets are stacking more tightly than exhibited when Laponite is in a dry powder state
.
or that the peak is a higher-order reflection of a peak present below a scattering angle of 5
Based on the TEM image, it is reasonable to assume that the peak is a secondary reflection of an

intercalation peak or ordered exfoliation peak. However, to determine with any certainty the
cause of the peak, an investigation utilizing small angle x-ray scattering is necessary, as will be
described in Chapter Five.
3.5.1.3 Evidence of PreferentialReinforcement
The results of characterization of these polyurethane/Laponite nanocomposites suggest that by
increasing the polarity and/or hydrophilicity of the polyurethane soft segment the Laponite can
be directed away from the hard domain and attracted into the polar soft domains. When
polyurethanes containing highly polar, hydrophilic soft segment constituents, such as
polyethylene oxide (PEO), are employed, the Laponite is directed away from the hard domains
and embedded within the soft domain. The resulting material is less extensible than the pure
polyurethane and very brittle. The other polyurethane investigated utilized the same hard
segment at similar weight concentrations, but the soft segments, although of similar molecular
weight,

had

different

polarities

and

hydrophilicities.

One

polyurethane

contained

poly(tetramethylene oxide) (PTMO) soft segments while the other polyurethane contained
polyethylene oxide-polypropylene oxide-polyethylene oxide (PEO-PPO-PEO) soft segments. As
can be seen from Table 3-1 PEO is more polar and has a higher solubility parameter than the
other soft segment constituents and attracts the Laponite platelets.
Polyurethane

Soft Segment
Constituent

Solubility Parameter
[J1 /2cm -3/2]

Elasthane
HDI-PTMO PU

PTMO

18.039

HDI-PEO-PPO-PEO PU

PEO
PPO
PEO-PPO-PEO

19.625
16.873
18.493

Table 3.1: Polyurethane soft segment polarity expressed in terms of the solubility
parameter which was calculated using the group contribution method and values
given in [451.

Consequently, the novel polyurethane/Laponite nanocomposite that has a PTMO soft segment
but hexamethylene diisocyanate (HDI) hard segment behaves similar to Elasthane/Laponite
nanocomposites-exhibiting an increase in strength, toughness, and stiffness without a loss in
extensibility when Laponite is added. This will be described in the next Chapter. Furthermore,
like the Elasthane nanocomposites, the HDI-PTMO PU nanocomposite exhibits a disruption in
the hard segment crystallinity seen in DSC and a decrease in lossiness (tan 6) at the glass
transition in DMA. There is no change in the value of the soft segment glass transition

temperature when Laponite is added (See Figures 3-9 and 3-10). Therefore, as in the
Elasthane/Laponite nanocomposites, the Laponite is preferentially embedded within the hard
domains of the HDI-PTMO PU and effectively strengthening and toughening the polyurethane.

1000
O

D -IPU
HDI-PTMO

0.1

100

10
0_01
001

1000
HDI-PEO-PPO PU

100

-AAAA

10000UU

SI

DI '

'

U.

c 0.1

- 100

1
zi

Lu 10
1I

0.01
00.

-I UUU

1000

C0.1

n. 100

10

t
-100

100
0
Temperature [C]

200

-100

100
0
Temperature rC]

200

-100

100 200
0
Temperature [OC]

300

Figure 3-9: Flexural storage modulus, E' (a), and loss tangent, tan 6 (b), both determined via DMA and DSC
thermograms and (c) during initial heating and cooling cycles of the pure (black) and 10 wt % Laponite-filled
PUs (red).

The novel polyurethane/Laponite nanocomposite that has the same HDI-based hard segments but
a more polar and hydrophilic, PEO-based soft segment exhibits contrasting behavior. The
extensibility of this polyurethane decreases dramatically when Laponite is added while the
stiffness of the material exhibits no distinct change. The dramatic reduction in extensibility
decreases the strength and toughness of the material significantly, and the nanocomposite
behaves like a brittle plastic rather than as a stronger, tougher elastomer as is displayed in Figure
3-10. Furthermore, DSC shows that the hard segment is unaffected by the addition of Laponite.
DMA shows that the value of the soft segment glass transition temperature is unchanged,
although the transition is broadened significantly, indicating a restriction in soft segment
mobility as shown in Figure 3-9b. Characterization of this PEO-based Laponite nanocomposite
suggests that the Laponite is embedded within the soft domain of polyurethane, where it restricts
soft segment mobility and negatively affects the extensibility of the material.

50
40
a
=E 30
20
10
0
E [mm/mm]

HDI-PTMO PU
Initial Modulus (MPa)
Ultimate Strength (MPa)

Failure Strain
Toughness at 30% Strain (MJ m 3)
Total Toughness, (MJ m')

260 18
40 + 5

6.28 + 0.35
2.57 0.23
149 21

HDI-PTMO PU & 10
wt/o Laponite

HDI-PEOPPO PU

HDI-PEO-PPO PU &
10 wt% Laponite

474 26
48 2
6.17 0.26

181 26
16 3

157 3
5.0 + 0-4
0.62 0.21

3.51 0.12
170 13

3.85 + 0.35
1.60 + 0.26
43 9

1.16 0.04
2.7 1.1

Figure 3-10: Representative engineering stress-strain tensile curves of the pure and 10 wt % Laponite filled
polyurethanes are presented, and the correspondingly labeled cross-polarized image (height) 2.2 mm) of four
of the six samples is shown (a-d), where (a) is unfilled PTMO:HDIBDO PU, (b) is 10 wt % Laponite filled
PTMO:HDI-BDO PU, (c) is unfilled PEO-PPO-PEO:HDI-BDO PU, and (d) is 10 wt % Laponite filled PEOPPO-PEO:HDI-BDO PU. Note that the stress-strain curves are offset by A = 2.0 for ease of viewing. The
cross-polarized images of PTMO:MDI-BDO and PTMO:HDI-BDO PU thin films after deformation are
similar. A white arrow runs tangent along the grip locale, separating the undeformed portion of the
polyurethane (right) from the deformed portion (left) which was stretched in a direction perpendicular to the
white arrow.
In summary, this novel solvent exchange approach can produce fully-exfoliated dispersions of
nanoparticles that enhance the thermomechanical properties of polyurethanes which contain nonpolar soft segments without hindering material extensibility. If the soft segment attracts the
nanoparticles (in this case the charged clay discs) the Laponite preferentially segregates into the
soft micro-domains. This significantly hinders soft segment mobility and therefore dramatically
reduces the TPU extensibility [41]. Ultimately, the solvent exchange approach combined with
the thermodynamic attraction of the unmodified Laponite platelets to polar and/or hydrophilic
constituents

afford material engineers

the ability to select and control the location of

nanoreinforcement. Thus, the thermomechanical properties of polar block-copolymers such as


polyurethanes and polyureas can be tailored.
3.5.2 Influence of Processing Time
When this novel solvent exchange approach is utilized, the solvent is removed slowly over three
to four weeks at elevated temperature. Ultimately, the polyurethane and nano-filter has an
extended period of time in which to preferentially self-assemble. Consequently, when the matrix
contains non-polar soft segments, the Laponite is preferentially embedded within the
polyurethane hard domains, where it significantly enhances the thermomechanical properties of
the material while maintaining material extensibility. A small fraction of Laponite may be
embedded within the soft domain, but this is not enough to negatively influence the material
behavior. However, when the matrix polyurethane contains polar constituents, such as
polyethylene oxide, the Laponite is preferentially embedded within the soft domain, restricting
soft segment mobility and decreasing the material extensibility. Again, a small fraction of the
Laponite may be embedded within the hard domain, but the Laponite within the soft domain
dominates the material behavior.

This same extent of preferential reinforcement can not be achieved if a quicker processing
method is utilized. The physical properties of the solvent cast films were observed to depend on
the evaporation rate. As noted in section 3.4.1 a constant evaporation rate was maintained for all
nanocomposite samples by keeping the films in a closed 60'C oven with a 0.02 m3 hr -1 N2 purge
flow. This rate was maintained because when the study was initiated the evaporation rate was not
strictly monitored and the mechanical properties from one batch of composites to the next
displayed significant differences. Those composites that were prepared under more rapid
evaporative conditions exhibited inferior properties. Consequently, when comparatively fast
processes like electrospinning and precipitation with compression molding are utilized, there is
less time for microphase segregation and preferential reinforcement. Inferior mechanical
properties are then inevitable.

These solvent-based nanodispersion methods are further complicated when Laponite or other
high aspect ratio particles are used, because the nanoparticles induce gelation at low
concentrations. When nanocomposite solutions for precipitation in solvents like methanol are

prepared the concentration of polyurethane in solution is constrained by the solubility limit. This
limit is further constrained when nanoparticles are added. For example, 5 grams of polyurethane
will not fully dissolve in a 100 gram solution that contains greater than 0.3 grams of Laponite (or
a concentration of Laponite that is greater than or equal to 6 wt% of the resulting solid). The
same solution without the Laponite will easily dissolve the polyurethane. Similar solutions that
contain concentrations of Laponite in the resulting solid that are greater than the percolation limit
are not able to be successfully electrospun because addition of such high concentration of
Laponite particles leads to jamming. This limits the extent of the solution viscoelasticity and
promotes capillary break-up. So while the solvent exchange approach successfully exfoliates
nanoparticles, the resultant property enhancement can only be achieved but at a cost-long
processing time and utilization of a lot of environmentally unfriendly solvents, such as dimethyl
acetamide and dimethylformamide. Should an environmentally-friendly

solvent, like a-

tocopherol or supercritical carbon dioxide be able to be utilized instead to maintain nanoparticle


dispersion, the resultant solution could be used as an additive to numerous polymeric products.
For example, nano-reinforced polyurethanes and polyureas could increase the strength and wear
resistance of prosthetics or protective eye-wear. Polyolefins too could be reinforced and the
barrier resistance and flame retardance of cable sheathing improved

3.6 Conclusions
The new solvent exchange method outlined here allows block-copolymers with polar
constituents, like thermoplastic polyurethanes, to be preferentially nanoreinforced with inorganic
nanoparticles, like smectic clays. This method utilizes a combination of solvents; one solvent
separates and disperses the nanoparticles and the second solvent maintains the dispersion
stability but also dissolves the matrix polymer. While this technique may be applied to numerous
nanoparticle-polymer systems, the extent of property enhancement depends on the time allowed
for phase-segregation and preferential reinforcement and the concentration limits of solution
feasibility relies heavily on the aspect ratio of the nanoparticles employed. Ultimately, the use of
block polarity in conjunction with this solvent exchange approach offers an avenue toward the
development of high performance materials that rival natural materials.

3.7 References
[1]

[2]
[3]
[4]

[5]

[6]

[7]

[8]

[9]

[10]
[11]

[12]

[13]

[14]

[15]

D. P. Knight and F. Vollrath, "Biological liquid crystal elastomers," Philosophical


Transactionsof the Royal Society ofLondon Series B-Biological Sciences, vol. 357, pp.
155-163, 2002.
F. Vollrath and D. P. Knight, "Liquid crystalline spinning of spider silk," Nature, vol.
410, pp.541-548, 2001.
Y. Termonia, "Molecular Modeling of Spider Silk Elasticity," Macromolecules, vol. 27,
pp. 7378-7381, 1994.
A. H. Simmons, C. A. Michal, and L. W. Jelinski, "Molecular orientation and twocomponent nature of the crystalline fraction of spider dragline silk," Science, vol. 271,
pp. 84-87, 1996.
J. D. van Beek, L. Beaulieu, H. Schafer, M. Demura, T. Asakura, and B. H. Meier,
"Solid-state NMR determination of the secondary structure of Samia cynthia ricini silk,"
Nature, vol. 405, pp. 1077-1079, 2000.
H. J. Gao, B. H. Ji, I. L. Jager, E. Arzt, and P. Fratzl, "Materials become insensitive to
flaws at nanoscale: Lessons from nature," Proceedings of the NationalAcademy of
Sciences of the United States ofAmerica, vol. 100, pp. 5597-5600, 2003.
B. L. Smith, T. E. Schaffer, M. Viani, J. B. Thompson, N. A. Frederick, J. Kindt, A.
Belcher, G. D. Stucky, D. E. Morse, and P. K. Hansma, "Molecular mechanistic origin of
the toughness of natural adhesives, fibres and composites," Nature, vol. 399, pp. 761-763,
1999.
J. Gosline, P. Guerette, C. Ortlepp, and K. Savage, "The mechanical design of spider
silks: from fibroin sequence to mechanical function," JExp Biol, vol. 202, pp. 32953303, 1999.
H. Koerner, G. Price, N. A. Pearce, M. Alexander, and R. A. Vaia, "Remotely actuated
polymer nanocomposites - stress-recovery of carbon-nanotube-filled thermoplastic
elastomers," Nature Materials,vol. 3, pp. 115-120, 2004.
R. A. Vaia and H. D. Wagner, "Framework for nanocomposites," Materials Today, vol.
7, pp. 32-37, 2004.
A. M. Chen, Y. Tian, B. Han, G. D. Ji, S. S. Wu, and J. Shen, "Synthesis and
characterization of polyurethane/montmorillonite nanocomposites," Acta Polymerica
Sinica, pp. 591-594, 2003.
S. Y. Moon, J. K. Kim, C. Nah, and Y. S. Lee, "Polyurethane/montmorillonite
nanocomposites prepared from crystalline polyols, using 1,4-butanediol and organoclay
hybrid as chain extenders," European Polymer Journal,vol. 40, pp. 1615-1621, 2004.
M. Song, D. J. Hourston, K. J. Yao, J. K. H. Tay, and M. A. Ansarifar, "High
performance nanocomposites of polyurethane elastomer and organically modified layered
silicate," Journal ofApplied Polymer Science, vol. 90, pp. 3239-3243, 2003.
Z. Z. Yu, C. Yan, M. S. Yang, and Y. W. Mai, "Mechanical and dynamic mechanical
properties of nylon 66/montmorillonite nanocomposites fabricated by melt
compounding," PolymerInternational,vol. 53, pp. 1093-1098, 2004.
X. M. Zhang, R. J. Xu, Z. G. Wu, and C. X. Zhou, "The synthesis and characterization of
polyurethane/clay nanocomposites," Polymer International,vol. 52, pp. 790-794, 2003.

[16]

[17]
[18]

[19]
[20]

[21]

[22]

[23]

[24]
[25]
[26]

[27]

[28]

[29]
[30]

[31]
[32]

S. S. Ray, K. Okamoto, and M. Okamoto, "Structure-property relationship in


biodegradable poly(butylene succinate)/layered silicate nanocomposites,"
Macromolecules, vol. 36, pp. 2355-2367, 2003.
S. Sinha Ray and M. Okamoto, "Polymer/layered silicate nanocomposites: a review from
preparation to processing," Progress in Polymer Science, vol. 28, pp. 1539-1641, 2003.
M. Alexandre and P. Dubois, "Polymer-layered silicate nanocomposites: preparation,
properties and uses of a new class of materials," MaterialsScience and Engineering:R.:
Reports, vol. 28, pp. 1-63, 2000.
J. Njuguna and K. Pielichowski, "Polymer nanocomposites for aerospace applications:
Fabrication," Advanced EngineeringMaterials,vol. 6, pp. 193-203, 2004.
D. Burgentzle, J. Duchet, J. F. Gerard, A. Jupin, and B. Fillon, "Solvent-based
nanocomposite coatings: I. Dispersion of organophilic montmorillonite in organic
solvents," Journalof ColloidandInterface Science, vol. 278, pp. 26-39, 2004.
B. Finnigan, D. Martin, P. Halley, R. Truss, and K. Campbell, "Morphology and
properties of thermoplastic polyurethane nanocomposites incorporating hydrophilic

layered silicates," Polymer, vol. 45, pp. 2249-2260, 2004.


X. H. Dai, J. Xu, X. L. Guo, Y. L. Lu, D. Y. Shen, N. Zhao, X. D. Luo, and X. L. Zhang,
"Study on structure and orientation action of polyurethane nanocomposites,"
Macromolecules, vol. 37, pp. 5615-5623, 2004.
I. Rhoney, S. Brown, N. E. Hudson, and R. A. Pethrick, "Influence of processing method
on the exfoliation process for organically modified clay systems. I. Polyurethanes,"
JournalofApplied Polymer Science, vol. 91, pp. 1335-1343, 2004.
P. C. LeBaron, Z. Wang, and T. J. Pinnavaia, "Polymer-layered silicate nanocomposites:
an overview," Applied Clay Science, vol. 15, pp. 11-29, 1999.
S. S. Ray and M. Okamoto, "Polymer/layered silicate nanocomposites: a review from
preparation to processing," Progress in Polymer Science, vol. 28, pp. 1539-1641, 2003.
Y. Kojima, A. Usuki, M. Kawasumi, A. Okada, Y. Fukushima, T. Kurauchi, and 0.
Kamigaito, "Mechanical-Properties Of Nylon 6-Clay Hybrid," JournalOf Materials
Research, vol. 8, pp. 1185-1189, 1993.
M. E. Mackay, A. Tuteja, P. M. Duxbury, C. J. Hawler, B. Van Horn, Z. Guan, G. Chen,
and R. S. Krishnan, "General strategies for nanoparticle dispersion," Science, vol. 311,
pp. 1740, 2006.
N. Kumar, S. M. Liff, and G. H. McKinley, " Method to disperse and exfoliate
nanoparticles," U. S. P. Office, Ed. United States: Massachusetts Institute of Technology,
Filed October 18, 2005.
R. G. Larson, The Structure and Rheology of Complex Fluids. New York, NY: Oxford
University Press, 1998.
C. S. O'Hern, L. E. Silbert, A. J. Liu, and S. R. Nagel, "Jamming at zero temperature and
zero applied stress: The epitome of disorder," PhysicalReview E, vol. 68, pp. 011306,
2003.
V. Trappe, V. Prasad, L. Cipelletti, P. N. Segre, and D. A. Weitz, "Jamming phase
diagram for attractive particles," Nature, vol. 411, pp. 771-772, 2001.
C. L. Yaws, X. Lin, L. Bu, D. R. Balundgi, and S. Tripathi, "Vapor Pressure," in
ChemicalPropertiesHandbook, C. L. Yaws, Ed. New York: McGraw-Hill, 1999, pp.
784.

[33]

[34]

[35]

[36]

[37]

[38]

[39]
[40]

[41]

[42]

[43]
[44]

[45]

A. Carli, S. Di Cave, and E. Sebastiani, "Thermodynamic characterization of vapour-liquid equilibria of mixtures acetic acid--dimethylacetamide and water-dimethylacetamide," Chemical EngineeringScience, vol. 27, pp. 993-1001, 1972.
M. J. Wiggins, M. MacEwan, J. M. Anderson, and A. Hiltner, "Effect of soft-segment
chemistry on polyurethane biostability during in vitro fatigue loading," JournalOf
Biomedical MaterialsResearch PartA, vol. 68A, pp. 668-683, 2004.
D. J. Martin, G. F. Meijs, P. A. Gunatillake, S. P. Yozghatlian, and G. M. Renwick, "The
influence of composition ratio on the morphology of biomedical polyurethanes," Journal
ofApplied Polymer Science, vol. 71, pp. 937-952, 1999.
C. D. Eisenbach, A. Ribbe, and C. Gunter, "Morphological-Studies Of Model
Polyurethane Elastomers By Element-Specific Electron-Microscopy," Macromolecular
Rapid Communications,vol. 15, pp. 395-403, 1994.
I. Grillo, P. Levitz, and T. Zemb, "Insertion of small anisotropic clay particles in swollen
lamellar or sponge phases of nonionic surfactant," EuropeanPhysicalJournalE, vol. 5,
pp. 377-386, 2001.
M. M. Malwitz, S. Lin-Gibson, E. K. Hobbie, P. D. Butler, and G. Schmidt, "Orientation
of platelets in multilayered nanocomposite polymer films," JournalofPolymer Science
PartB-Polymer Physics, vol. 41, pp. 3237-3248, 2003.
Cornell University, "Cornell High Energy Synchrotron Source - GLine,"http://www.chess.cornell.edu/gline/index.htm, 2007: Accessed August 20, 2005.
A. Aneja and G. L. Wilkes, "A systematic series of'model' PTMO based segmented
polyurethanes reinvestigated using atomic force microscopy," Polymer, vol. 44, pp. 72217228, 2003.
L. T. J. Korley, S. M. Liff, N. Kumar, G. H. McKinley, and P. T. Hammond,
"Preferential association of segment blocks in polyurethane nanocomposites,"
Macromolecules, vol. 39, pp. 7030-7036, 2006.
L. J. Korley, PEO-containingCopolymers as PolyurethaneSoft Segments in the
Development ofHigh PerformanceMaterials,Ph.D. in Program in Polymer Science and
Technology, Cambridge, MA:Massachusetts Institute of Technology, 2005.
M. Szycher, Szycher's Handbook ofPolyurethanes.Boca Raton, Florida: CRC Press,
1999.
G. Pollock, Synthesis and Characterizationof Mechanically Enhanced,Nanostructured
ThermoplasticPolyurethaneElastomers,Ph.D. in Program in Polymer Science and
Technology, Cambridge, MA:Massachusetts Institute of Technology, 2005.
D. W. van Krevelen, PropertiesofPolymers: Their Correlationwith Chemical Structure;
Their Numerical Estimation andPredictionFrom Additive Group Contributions,3rd ed.
Amsterdam, NL: Elsevier Science, 1997.

4. High Performance Elastomeric Nanocomposites via


Solvent Exchange Processing
[Part of this work has been published previously, in slightly different form, in "High Performance Elastomeric
Nanocomposites via Solvent Exchange Processing" by S.M. Liff, N. Kumar, and G.H. McKinley, Nature Materials,
2007, 6 (1): 76-83 and in "Preferential Association of Segment Blocks in Polyurethane Nanocomposites" by L.T.J.
Korley , S.M. Liff, P.T. Hammond, and G.H. McKinley, Macromolecules, 2006, 39 (20): 7030-7036. ]

4.1 Introduction
The incorporation of nanoparticles into engineering thermoplastics affords engineers an
opportunity to synthesize polymer nanocomposites that potentially rival the most advanced
materials in nature. Development of these materials is difficult since thermodynamic and kinetic
barriers inhibit the dispersal of inorganic, often hydrophilic nanoparticles in hydrophobic
polymer matrices. Using the new solvent exchange approach, detailed thoroughly in Chapter 3,
we preferentially nanoreinforce the hard micro-domains of thermoplastic elastomers with a
smectic clay of similar characteristic dimensions. The strong adhesion between the clay and the
hard micro-domains of the block copolymer coupled with the formation of a percolative network
not only stiffens and toughens, but increases the heat distortion temperature of the material and
induces reversible thermotropic liquid-crystalline transitions. The discotic clay platelets induce
morphological ordering over a range of length scales that results in significant thermomechanical
enhancement and expands high temperature applications. Merging block-copolymer processing
techniques with this method for nanoparticle preferential-ordering facilitates the development of
new, hierarchically-ordered materials or "'molecular-composite hybrids'"' [1].

High performance biomaterials, such as silk [2-6] and bone [7, 8] exhibit unparalleled
combinations of stiffness, strength, extensibility, and toughness by exploiting hierarchical
structures in which stiff nanometer-size crystallites are embedded and dispersed in softer protein
matrices [2-4, 9]. Polymer nanocomposites which utilize engineering thermoplastics as the
matrix material and stiff, inorganic, anisotropic nanoparticles to tailor thermomechanical
properties have been the subject of many recent studies as discussed in Chapter 2 [10, 11].
Successful mechanical enhancement of polymer matrices via nanoreinforcement is expected only
when the nanoparticles are fully exfoliated in the polymer matrix. This is difficult since the

organic, hydrophobic polymer matrices thermodynamically and kinetically inhibit the dispersal
of the

inorganic,

hydrophilic

clay.

Despite

numerous

advances

synthetic

polymer

nanocomposites have yet to display the superior toughness of natural biomaterials, such as silk.
Like a majority of engineering materials the nanocomposites are either stiff but not extensible,
like steel, or extensible but not stiff, like rubber.

In this Chapter, we present the material

characterization of a high performance nanocomposite prepared via the novel solvent exchange
approach of Chapter Three composed of synthetic smectic clay, Laponite RD, and a
commercial thermoplastic polyurethane (TPU) elastomer, ElasthaneTM 80A.

4.2 Materials and Characterization Methods


4.2.1 Materials
Biomedical polyetherurethane, ElasthaneTM 80A was obtained from the Polymer Technology
Group, Inc. (Berkeley, CA). Elasthane contains poly(tetramethylene oxide) (PTMO) soft
segments and 4,4'-methylene bisphenyl diisocyanate (MDI) hard segments chain extended with
1,4-butanediol (BDO). The hard-to-soft segment ratio of Elasthane 80A is 40:60 wt%, its
molecular weight is 290,000 g/mol, and its density is 1.12 g.cm -3 [12]. Laponite RD, a
synthetic smectic clay with a density of 2.65 g.cm -3 [13] was obtained from Southern Clay
Products, Inc. (Gonzales, TX). Dimethylacetamide (DMAc; B.P. 165 0 C) was obtained from
Sigma Aldrich Co. (St. Louis, MO) and used as received. Laponite was dispersed in DMAc via
the new solvent exchange procedure [14] explained in the previous Chapter. Solutions of
Elasthane, Laponite, and DMAc were prepared from the resulting Laponite-DMAc dispersion.
The Elasthane concentration in the solution was maintained at 1.4-1.6 wt%. Elasthane/Laponite
nanocomposite films with 120-150 pm thicknesses were prepared in Teflon containers (6 cm x
4 cm) by controlled evaporation of DMAc. The final Laponite concentration in the dry film was
controlled by varying the Laponite concentration in the Laponite-Elasthane-DMAc solution.
Note, for a discotic clay such as Laponite, the percolation threshold, 0p, can be approximated by
dividing the random-close-packed hard sphere percolation volume fraction (-0.64) by the aspectratio, a, of the clay. Because the diameter and thickness of Laponite platelets are 25 nm and
Inm, respectively, the aspect ratio of Laponite is 25 and the percolation threshold is 2.5 vol% or
6 wt%. The evaporation rate of DMAc was maintained by keeping the films in a closed 60 0 C
oven with a 0.02 m 3.hr - 1 N2 purge flow. A constant evaporation rate was maintained for all
samples because the physical properties of the dry films were observed to depend on the

evaporation rate. The mechanical properties of films prepared at higher evaporation rates proved
repeatedly to be inferior (see section 3.5.2.).
4.2.2 Atomic Force Microscopy
Atomic force microscope (AFM) phase images of the nanocomposite samples (which were
sandwiched between epoxy and cryo-microtomed) were taken in tapping mode using moderate
force (Amplitude Set Point = 0.7 Volts, scale = 60') and high force (0.3 Volts, 15') on a Veeco
Instruments Dimension 3100 AFM with the same Veeco NanoProbe (130 [tm, 280-361 Hz).
4.2.3 Mechanical Characterization
Tensile tests were performed on thin film samples approximately 45 x 5 x 0.1 mm 3 and were
carried out on an Zwick/Roell Z010 mechanical tester with a 500 N load cell at a constant
crosshead speed of 100% initial length per minute. Dynamic mechanical analysis (DMA)
measurements were made at 1 Hz and at a heating rate of 3'C.min -' using the TA Instruments
Q800 DMA. To demonstrate the thermomechanical integrity of these nanocomposites two films
containing 0 and 20 wt% Laponite were held in improvised grips under a constant load equal to
1.08 N (initial stress of 1.75 MPa) and heated from 45'C to 125'C at -1C.min -'. Photographs
were captured as the experimental set-up underwent heating in an oven with a transparent
window.
4.2.4 Thermal Characterization
Differential scanning calorimetry (DSC) measurements were performed on a TA Instruments
Q1000 DSC at heating rate of 10oC.min -1. The theoretical Laponite concentration in the films
was verified by thermogravimetric analysis (TGA). TGA was performed under a 20 ml.min -1 N2
purge at a heating rate of 10'C.min -' with a Perkin Elmer TGA7.

Multiple samples at each

Laponite concentration were prepared and all DMA, DSC, and tensile tests were repeated a
minimum of three times. A Carl Zeiss Axioskop 2MAT Polarizing Microscope was used to
obtain birefringent images of the nanocomposites.
4.3 Results and Discussion
Elasthane, an elastomeric block-copolymer, is composed of hard segments (HS) and soft
segments (SS) which combine to give the TPU moderate stiffness, high rubber-like extensibility,
and easy processibility [12, 15] (See Figure 4-1). Thermodynamic incompatibility of the HS and
SS drives the polymer system into a two-phase morphology in which hydrogen-bonded,
crystalline hard micro-domains form amid the rubbery soft domains. Depending on the specific
segmental composition and the interaction between soft and hard segments, the hard micro-

domains can form fibrillar, globular, or lamellar structures within a continuous soft matrix or
form an interconnected hard domain network [16, 17]. The hard segments (length 3-11 nm [18,
19]) are observed to phase segregate in polyurethanes similar to Elasthane and form hard
cylindrical crystalline domains via hydrogen bonding [20]. Because the polyurethane HS has a
high melting temperature (Tn,HS > Troom) and the SS has a low glass transition temperature (Tg,ss

< Troom), at room temperatures the hard and soft domains are in crystalline and rubbery states,
respectively. The stiffness (or elastic modulus, E) and strength (tensile stress at break, Omx) of
the TPU increases with HS content while the extensibility (strain at failure,

max)

increases with

SS content [18, 21-24]. The flexibility of urethane chemistry enables the chemical structure of
the HS and SS domains to be varied to control the thermal transition temperatures and
thermomechanical properties of the polyurethane [25, 26]. Synthetically it is difficult to increase
the elastic modulus of a polyurethane and maintain its high extensibility, or vice-versa-an
increase in one often results in a decrease in the other [15].

'.4
'.4

Hard Segment

Soft Segment

Figure 4-1: Structure of Elasthane 80A and Laponite dispersal within the polyurethane matrix. Elasthane
80A is an elastomeric block-copolymer composed of MDI-BDO hard segments and PTMO soft segments.
Hard segments (length 3-11 nm [18, 191) phase segregate to form hard crystalline domains via hydrogen
bonding [201. The charged Laponite platelets (in orange) preferentially reinforce the polar hard domains.

Recent studies of high performance biomaterials such as spider silk show that araneid spiders
modify the stiffness and strength of silk by regulating ordering of polypeptide chains and/or
through formation of nano-crystallites during spinning [27, 28]. Following previous polymer
nanocomposite work [10, 11, 29, 30], we thus chose to stiffen and strengthen a commercial
synthetic TPU with low aspect-ratio clay platelets. Very recently, Finnigan et al. have studied a
similar TPU (but with higher soft segment weight fraction and with a spherical morphology)
filled with a single weight fraction of clay (3 wt%) [16]. They show that the aspect-ratio of the

silicates modifies the mechanical behavior but does not change the phase morphology. To
increase E and

ma,, of

Elasthane while maintaining its high Em,

it is necessary to preferentially

nanoreinforce the polyurethane hard micro-domains with unmodified Laponite nanoparticles.


The stronger affinity of the hydrophilic, non-surface-modified Laponite with the more polar HS
rather than with the SS allows the Laponite to segregate into the hard micro-domains during the
processing of the TPU. By preferentially restricting the Laponite particles to the hard microdomains the undesired stiffening of the soft domain is avoided and the large

&,a

of the

polyurethane composite remains uncompromised. To optimize this micro-segregation it is key


that the dimension of the discotic clay platelets is similar in magnitude to the natural dimensions
of the HS domains that microphase separate (HS length

= 3-11 nm [18, 19], Figure 4-1).

Consequently, Laponite (diameter, d= 25 nm, thickness l = 1 nm, aspect-ratio a = d / = 25) is


more suitable for the preferential nanoreinforcement of the hard micro-domains than other high
aspect-ratio clays (e.g. Montmorillonite and Hectorite with d=100-300 nm [29]) that are
commonly used in polymer nanocomposites.
AFM phase imaging is often used to distinguish hard and soft micro-domains in polyurethanes
because there is often insufficient contrast in electron density between phases and effective
staining procedures prove too time-consuming to make TEM effective [31, 32]. Extensive
analysis of AFM images showed repeatedly that as more of the high modulus Laponite is added
the bright areas brighten and grow in size, indicating that the Laponite imposes a change in hard
domain texture. As the Laponite concentration increases the discrete hard micro-domains
become interconnected, and in conjunction with the Laponite this accounts for the increased
initial modulus and distinct yield point. The bright white regions on the left hand column, which
correspond to regions occupied by the very highest modulus moieties on the right, are believed to
be a combination of Laponite and hard polyurethane micro-domains at the surface. With AFM it
is difficult to distinguish the Laponite platelets from the hard polyurethane micro-domains due to
the preferential association of the Laponite and the irregular shape formation of the hard microdomains. The hard micro-domains in the bulk Elasthane are unlike the fibrillar hard microdomain structures observed in AFM phase images of piperazine-butanediol-poly(tetramethylene
oxide) and hexamethylene diisocyanate-butanediol-poly(ethylene oxide)-poly(propylene oxide)

20 wt %
Figure 4-2: Hard and soft micro-domain distribution in nanocomposites containing 0, 4, 10, and
nanocomposite
Laponite as observed via AFM phase imaging. The same areal extent (1 gm x 1 gm) on each
with high
was imaged with moderate tapping force (amplitude set point = 0.7 Volts, scale = 600, left) and
the
within
modulus
high
of
material
tapping force (0.3 Volts, 150, right). The bright regions correspond to
softer matrix material.

84

poly(ethylene oxide) polyurethanes [33] or the large, spherical hard micro-domain structures
observed in methylene-bis(cyclohexyl isocyanate)-butanediol-poly(propylene glycol) [34]. The
soft domains observed in the present study do not form the irregular spherical soft microdomains which are observed in AFM phase images of bisphenyl methylene diisocyanatedendrimer-poly(tetramethylene oxide) polyurethanes [19, 35].

Instead the hard micro-domain

morphology observed in these images is similar to the morphology observed by Eisenbach et al.
via element-specific electron-microscopy and by McLean and Sauer using tapping-mode AFM in
polyurethanes of similar composition [19, 35, 36].
Because Laponite preferentially reinforces

the hard micro-domains of Elasthane and

concentrations greater than 5.9 wt% induce jamming, the influence on the mechanical properties
of Elasthane is profound. The tensile curve of pure Elasthane (0 = 0 wt%) shown in Figure 4-3a
is similar to that of other TPUs [37, 38] in which three distinguishable deformation regions are
readily identified. The first region where the stress, o, increases linearly with strain, , is a region
of elastic deformation and is governed by TPU crystallinity, hard segment content and ordering
[38].
Yield corresponds to break-up of an interconnected hard domain network and is followed by two
distinct regions of plastic deformation [17, 38]. The region of moderate slope is indicative of soft
domain deformation plus rotation and alignment of the smaller hard micro-domains. The last
region characterized by the steep upturn in the c-e curve corresponds to the stretching and straininduced crystallization of the soft-segments as well as further break-up of the hard microdomains [17], later discussed in Chapter Five. The addition of Laponite significantly increases
the initial modulus of the material and the magnitude of the plateau and ultimate stress, but does
not affect the slope of the curve due to soft segment stretching after yield or the slope of the
curve due to soft-segment strain-induced crystallization or hardening. However, the onset of
plastic deformation and strain-hardening both occur at smaller deformations with increasing
Laponite concentration (see Chapter Five).

M1

W
CL

-oo
/I
800

4'600

400
200 i

,
J

200
.

0
0

,
15

[wtO/o]

I I
25
20

I
0

, I
5

10

15

I
20

[wto]

Figure 4-3: Impact of Laponite concentration on the mechanical properties of Elasthane. a, Representative
engineering stress-strain (C-s) curves of thin film nanocomposites containing between 0 and 20 wt%
Laponite. b, Initial elastic modulus, E, increases with increasing Laponite concentration, 0. For 0 < Op the
prediction of Guth's model (Equation 1; red line) is in excellent agreement with experimental values. The
agreement of Guth's model using a = 25 further demonstrates that the Laponite is completely exfoliated. For
> Op the sharp increase in E follows the typical scaling (Equation 2; blue dashed-line) of jammed
systems[18, 20]. c, Toughness, Uo.3, (defined as the area under the 0-c curve up to E = 0.3) increases
quadratically with 0 (Equation 3; blue line), so Uo.3 of a nanocomposite containing 20 wt% Laponite is 4.25
times greater than that of pure Elasthane. The error bars in b and c are equal to one standard deviation
(minimum of five samples) plus the contribution based on instrument imprecision.

The dramatic impact of Laponite on the elastic modulus of the polyurethane is shown in Figure
4-3. The modulus, E, increases with increasing Laponite concentration from 38 MPa at b= 0
wt% to 70.5 MPa at 0 = 4 wt% to 856 MPa at 0 = 20 wt%. This 23-fold increase in modulus, and
specifically the sharp increase in E at 0 > 6 wt%, is a consequence of the percolative Laponite
network. Below Op the platelets are randomly dispersed with an average inter-particle spacing
greater than one particle diameter. Above Op, the Laponite forms a sample-spanning jammed
structure which significantly hinders the mobility of hard and soft micro-domains during
imposed deformation. Note the nanocomposites containing 15 and 20 wt% Laponite exhibit a

sudden decrease in cr after yield, a behavior typical of glassy polymers [39]. This result
corresponds to the sudden collapse of the jammed structure at which point the yield stress
exceeds the stress that can be supported by the soft domains without significant reorientation and
alignment.
The increase in E below the percolation limit can be modeled by Guth's equation [40] which was
extended for anisotropic particles:
E = Eo(1+ 0.67a

+ 1.62a 2

), for

<~

(4.1)

Above the percolation limit, a power law scaling typical of jammed systems, such as gels [36]
and glasses [41], is observed:
E = E, + A(

where

)"n for

(4.2)

>

is the volume fraction of Laponite,

Pis the percolation threshold volume fraction of

Laponite, Eo is the elastic modulus of pure Elasthane, and A and n are model parameters. The
predicted values of E using Eq. 4.1 with Eo = 38 MPa and a = 25 agree with experimental data.
The aspect ratio, a = 25, of a single Laponite platelet further indicates that the Laponite platelets
are fully exfoliated. Above the percolation threshold Eq. 4.2 is used with {E , A, n} = {110
MPa, 1.54x10 5 MPa, 2}. This power-law scaling is typical of critical-like behavior systems.
While this behavior is similar to equilibrium critical phenomena, jamming is clearly a nonequilibrium transition [36].

The scaling factor, n, depends on numerous system properties

including particle size, shape, polydispersity, and the particle-particle interaction potential [36,
41]. The magnitude of the scaling factor increases with the extent of particle-particle
interactions [42]. As shown in Figure 4-3c the toughness at C = 0.3 increases quadratically with
0 in a manner similar to the Guth prediction for modulus. Thus
Uo.3 = Uo. (+ 0.415a

+ 0.345a2

2)

(4.3)

-3
where U.3 = 0.916 MJ.m -3 and Uo. 3 increases from 0.86 MJ.m -3 at 0 = 0 wt% to 3.7 MJ.m at

S= 20 wt%.

The impact of Laponite concentration on the extensibility (a),

strength

(oma),

and overall

toughness (Umax) is not as dramatic. There is no significant change in the extensibility (max) of
Elasthane with Laponite--all nanocomposites exhibit max = 4.23 + 0.4. This suggests that the
ultimate extensibility of the elastic network of soft domains is not compromised by reinforcing
the hard domains. Both

max

and Umax increase when Laponite is added to Elasthane. In

comparison, other elastomeric nanocomposites reinforced with either smectic clays [20, 43-45],
carbon nanotubes [46], or fullerene [47] have only exhibited up to a six-fold increase in E and
these enhancements are often accompanied by reductions in either max,
Elasthane/Laponite nanocomposite, the ultimate tensile strength
exhibiting a 50% increase from 32 MPa at 0 = 0 wt% to 50 MPa at
plateaus at

max
i

max,

or both. In the

plateaus at I = 10 wt%,

= 10 wt%. Similarly, Umax

= 15 wt% and exhibits a 100% increase, raising Uma from 48 MJ m -3 at 0 = 0 wt%

to 95 MJ m -3 at b = 15 wt%. The Laponite toughens the Elasthane matrix in a manner similar to


that in which hard, crystalline, mineral platelets toughen bone and nacre [7, 48]. The successful
toughening of the Elasthane with Laponite is a result of the strong affinity, and consequent
strong interfacial adhesion, of the Laponite with the polyurethane hard domain.

The 23-fold increase in E and increase in max, U0.3 and

Umax

(without a concomitant reduction in

cmax, or any change in the slope of the tensile curve during strain-hardening) arises because the
Laponite is preferentially embedded within the polyurethane hard micro-domains rather than in
the soft domain and drives interconnections between the hard micro-domains as observed with
AFM (Figure 4-2). As shown in Figure 4-4a and Figure 4-5 the insignificant change in the soft
segment glass transition temperature, Tg,ss, combined with the large increase in heat distortion
temperature (HDT) and large increase in the magnitude of the flexural storage modulus, E' (0 ,
T), for temperatures T > Tg,ss, confirms that Laponite platelets are preferentially embedded in the
hard micro-domains. Only at

2 15 wt% does Laponite begin to observably enter the soft

domains of the Elasthane and induce soft segment crystallinity.


In addition to identifying the location of preferential nanoreinforcement, the dynamic mechanical
analysis (DMA) data for the various nanocomposites depicted in Figure 4-4b elucidates the
influence of Laponite concentration on the flexural storage modulus, E' (0, T).

10

10

a. S10
C 10"

102
0

In
I0

in

-100

-50

100
50
0
Temperature [C]

150

200

-100

-50

100
50
0
Temperature [C]

150

200

125 C

Gislp,t l no] deformati

a, The ratio
Figure 4-4: Impact of Laponite concentration on the thermomechanical properties of Elasthane.
curve, for
tan(b)
or
of the flexural storage modulus to the loss modulus at various temperatures,
Analysis
Mechanical
Dynamic
by
nanocomposites containing between 0 and 20 wt% Laponite measured
0
curve,
the
in
C
-25
at
peak
by the
(DMA). The soft segment glass transition temperature (Tg,ss), indicated
0
0 2 15
with
nanocomposites
in
observed
C
29
does not change significantly with increasing q. A new peak at
data
DSC
the
in
appear
also
transitions
Equivalent
wt% suggests the onset of crystallization in soft domains.
significantly
not
does
nanocomposite
the
in
Laponite
of
shown in Figure 4-5. b, Increasing the concentration
glass
change the magnitude of the flexural storage modulus (E) prior to the onset of the soft segment
fall
steep
0
The
0.
with
significantly
increases
E'
of
magnitude
the
temperature
transition at -50 C. Above this
0
the
of
melting
the
by
caused
Elasthane
the
of
softening
the
to
due
is
C)
130
(T>
in E' at high temperatures
which
at
hard domains. The heat distortion temperatures (HDT), which may be defined as the temperature
134,
E ' = 8 MPa (see ISO 75-method C) for nanocomposites with 0= 0, 2, 4, 8, 10, and 20 wt% are 101,111,
could not
161, >160, >200 0 C, respectively. The exact values of the HDT for nanocomposites with 0 > 10 wt%
at
Elasthane
of
be measured due to instrument limitations. c, Demonstration of the mechanical enhancement
little
exhibits
Elasthane
pure
The
high temperature by the addition of Laponite (See Appendix A-1).
increase in temperature, reaching the maximum
resistance to heat distortion, deforming significantly with
0
0
The nanocomposite containing 20 wt% Laponite
C.
120
at
allowable extension at 95 C, and finally breaking
0
upon heating; however, at 125 C the grips
deformation
any
visually
show
resists heat distortion and does not
slipped from the sample. Length of scale bar: 46 mm.

The magnitude of the rubbery plateau modulus at 25'C scales like the elastic modulus, E, and
exhibits a 25-fold enhancement from E'= 24 MPa at 0 = 0 wt% to E'= 600 MPa at

= 20 wt%.

Again, this enhancement is due to the percolative network established by the exfoliated Laponite
platelets which are preferentially embedded within the TPU hard micro-domains. Consequently,
the addition of Laponite to shape memory polyurethanes may significantly improve performance,
offering a means to enhance the rubbery modulus and thus the restoring force [10] without
changing the glass transition temperature.

The addition of 20 wt% Laponite to Elasthane induces an increase in the HDT of more than
100C. This allows the nanocomposite to be used potentially in higher temperature applications
without risk of permanent sagging or deformation. Figure 4-4c and the movie in Appendix A-1
illustrate that under a constant load Elasthane films elongate by more than 100% and break at
120 0 C. In contrast, thin films containing 20 wt% Laponite exhibit no deformation at
temperatures as high as 125'C. Furthermore, because the Laponite platelets are not organically
modified, TGA shows that there is no risk of thermal degradation or decomposition at
temperatures below 250'C (Figure 4-6).

Further investigation is necessary to determine what thermo-physical mechanism is responsible


for the dramatic increase of the HDT when Laponite is added to the polyurethane-presumably
both the percolative Laponite network and the enhancement in magnitude and intensity of the
hard segment melting temperature which is observed in DSC data (Figure 4-5) play a role. In
general, the melting temperature of a material increases if the intermolecular cohesive bonds
holding the material together are strengthened, and the intensity of a hard segment melting
endotherm increases when the crystallite is more perfect [39]. Consequently, Laponite platelets
must amplify a crystalline morphology with the hard micro-domains which accounts for the HS
strengthening and the enhancement in melting endotherm intensity.

06

06la

0.4

0.4

0.2

0.2

0.0

S0.0
-1-

-0.2

-0.2

-0.4

-0.4

,.,N

-N A

-100

-50

50 100 150
0
Temperature [OC]

200

250

-100

-50

50 100 150
0
Temperature [OC]

200

250

Figure 4-5: DSC of the nanocomposites containing 0, 4, 10, and 20 wt% Laponite. a, DSC curves during the
first heating and cooling cycle show both reversible and irreversible transitions. b, Subsequent heating and
cooling cycles show reversible transitions. Pure Elasthane exhibits a reversible hard segment melting
0
0
endotherm and crystallization exotherm at 150 C and 77 C, respectively, which agree with the values
reported in literature. With increasing Laponite concentration, , this peak reduces in intensity, and finally
disappears in nanocomposites with q > 6 wt%. This behavior is attributed to the restricted motion of the hard
segments due to the disruption by and jamming of Laponite above the percolation concentration 0 ($p = 5.9
wt%). Furthermore, an irreversible melting peak observed in the first heating cycle near 227 C is not
observed in the samples during the subsequent heating cycles. The increased intensity of this peak is
attributed to interactions between the hard segment and the Laponite since the peak intensity increases with
0
0. The origin of the broad dip in the first heating cycle near 150 C observed in the 10 wt% and 20 wt%
nanocomposite thermograms cannot be resolved but may result from soft segment and Laponite interactions.
0
The glass transition temperature of the soft segment, Tg,ss at -41 C does not change, but nanocomposites with
0
0
_ 15 wt% exhibit an interphase reversible melting/crystallization peak at 10 C and -25 C during the
heating and cooling cycles, respectively. Similar results are observed in DMA measurements (Figure 4-4a).

II
1.0

20

0.9
o

0.8

a,
.=

0.7
0.6
0o

0.5

I-

S10

u. 0.4
S0.3
0.2

"0

0.1
0.0
00
Temperature

rC]

0
0 [wt%]

Figure 4-6: Laponite concentration and impact on thermal degradation. a, Thermogravimetric analysis
(TGA) of nanocomposites containing 0 to 20 wt% Laponite. The rapid loss in weight at T = 350 0C is due to
the degradation of the Elasthane. b, Residue weight fraction at 1000 0C of the various nanocomposites is
nearly equal to the weight fraction of Laponite blended with Elasthane in each case. This is demonstrated by
the agreement of the experimental data with a line of slope one. Increasing Laponite fraction decreases the
thermal degradation temperature which is defined as the temperature at which the sample weight decreases
by 5%.

Microscopic images of the nanocomposite films taken under cross-polarized light like those
depicted in Figures 4-7 and 4-8 help to elucidate the influence of Laponite on the morphology of
the polyurethane nanocomposite. The local mechanical properties and morphology of the ordered
domains and the amorphous matrix are very different from one another. The birefringence and
fractal-like shape of the ordered domains depicted in Figure 4-7b and Figure 4-8 suggest a
quenched phase transition process [49]. The spatial scale of these ordered domains (20-100 pm)
is much larger than the size of the hard micro-domains (20-175 nm) [20, 31]. The fraction of the
image area occupied by these ordered-domains increases slightly with ; however, this increase
was close to the sample-to-sample variation with the same 0. As shown in Figure 4-7a, there is a
rapid reduction in the fractional area of bright regions upon heating between 100'C and 120'C.
This transition is similar to the local melting transition in thermotropic liquid-crystalline
polymers. This long-range ordering is the result of an isotropic-nematic (I-N) transition of the
hard micro-domains with, and without, the Laponite discs. Upon annealing the samples at 60 0 C,
new, ordered, crystalline domains appeared after numerous days of annealing as shown in Figure

4-7c. This re-crystallization behavior was not observed in the pure Elasthane sample even after
annealing for many days.

N
Grip location Stretch direction

L
(U
S

30

M Annealing at 60*C
U.

60
90
Temperature [ C]

120

150

Unstretched

7
"7

Stretched/

Figure 4-7: Evidence of long range order or crystallization in the nanocomposites is apparent in images
obtained with a cross-polarized microscope. a, Bright crystalline domains in a nanocomposite containing 20
wt% Laponite are due to the long-range ordering of the Elasthane hard domains. b, The fractal-like shape of
the crystalline domains is typical of a quenched phase transformation. Upon heating, the bright domains
undergo a nematic-isotropic transition between 100 and 120 0C. c, Crystalline domains reappeared upon
annealing the sample for a few days at 60 0C. d, Birefringent image of a nanocomposite containing 10 wt%
Laponite after tensile testing. The unstretched region on the left side of the grip line exhibits these same
bright domains. Upon stretching the amorphous matrix becomes birefringent as a result of strain induced
alignment and crystallization of the polymer chains during plastic deformation[501. e, The ordered domains
become disordered upon stretching as evidenced by the darkening of these previously bright domains.
Similar birefringence results are also obtained in pure Elasthane films. Image widths are as follow: (a) 2700
gm, (b) 337 jim, (c) 270 gm, (d) 2700 gm, and (e) 337 gm.

The re-crystallization behavior is again similar to that observed in thermotropic liquid-crystalline


polymers in which mesogens are chemically attached to the polymer chains [51]. In our
nanocomposites it appears that the Laponite platelets act as mesogens that are physically
embedded in the hard micro-domains rather than chemically attached to the polymer chain.

M
I/

l!

FIGURE 4-8: Polarized optical microscopy images of the ordered, fractal-like domains in various
nanocomposite films exhibit birefringence. Image width is as follows: (a) 270 gm, (b) 135 gm, (c) 675 gm, (d)
270 gm, (e) 675 jim, (f) 270 gm, (g) 135 jm, and (h) 337 pm.

These ordered domains did not exhibit the Maltese cross pattern indicative of spherulites
observed in some polyurethane systems [31]. Spherulites which consist of numerous crystallites
that grow radially outward form spheres that measure approximately 10 Vpm in diameter [39].
Instead these ordered domains maintain non-descript, fractal-like shapes that measure
approximately 200 pm in size until they are disrupted by stretching. After stretching the
amorphous matrix becomes birefringent via strain induced alignment and crystallization of the
polymer chains during plastic deformation [50]. The ordered domains become disordered during
stretching as evidenced by the darkening of the previously bright fields (Figures 4-7d-e).

4.4 Conclusion
The new solvent exchange method outlined in Chapter Three allows block-copolymers with
polar constituents, like thermoplastic polyurethanes, to be preferentially nanoreinforced with
organic

nanoparticles,

like

smectic

clays.

The

efficacy

of

nanoreinforcement

on

thermomechanical properties hinges upon the close matching of characteristic length scale and
the adhesion of the nanoparticles to the targeted polymer phase morphology. The hierarchical
order induced by the addition of discotic Laponite platelets creates a percolated network that
significantly stiffens and toughens the elastomer without reducing its extensibility. The use of
block polarity in conjunction with this solvent exchange approach and processing techniques,
such as electrospinning (later discussed in Chapter Seven), offers an avenue toward the
development of high performance materials that rival natural materials. Furthermore, the
addition of Laponite to shape memory polyurethanes or other actuating polymers may
significantly improve performance, offering a means to enhance the restoring force without
changing the glass transition temperature.

4.5 References
[1]
[2]

[3]
[4]
[5]

[6]

[7]

[8]

E. Manias, "Stiffer by design," Nature Materials,vol. 6, pp. 9-11, 2007.


D. P. Knight and F. Vollrath, "Biological liquid crystal elastomers," Philosophical
Transactionsof the Royal Society ofLondon Series B-BiologicalSciences, vol. 357, pp.
155-163, 2002.
F. Vollrath and D. P. Knight, "Liquid crystalline spinning of spider silk," Nature, vol.
410, pp. 541-548, 2001.
Y. Termonia, "Molecular Modeling of Spider Silk Elasticity," Macromolecules, vol. 27,
pp. 7378-7381, 1994.
A. H. Simmons, C. A. Michal, and L. W. Jelinski, "Molecular orientation and twocomponent nature of the crystalline fraction of spider dragline silk," Science, vol. 271,
pp. 84-87, 1996.
J. D. van Beek, L. Beaulieu, H. Schafer, M. Demura, T. Asakura, and B. H. Meier,
"Solid-state NMR determination of the secondary structure of Samia cynthia ricini silk,"
Nature, vol. 405, pp. 1077-1079, 2000.
H. J. Gao, B. H. Ji, I. L. Jager, E. Arzt, and P. Fratzl, "Materials become insensitive to
flaws at nanoscale: Lessons from nature," Proceedingsof the NationalAcademy of
Sciences of the UnitedStates ofAmerica, vol. 100, pp. 5597-5600, 2003.
B. L. Smith, T. E. Schaffer, M. Viani, J. B. Thompson, N. A. Frederick, J. Kindt, A.
Belcher, G. D. Stucky, D. E. Morse, and P. K. Hansma, "Molecular mechanistic origin of
the toughness of natural adhesives, fibres and composites," Nature, vol. 399, pp. 761-763,
1999.

[9]

[10]

[11]
[12]

[13]

[14]

[15]
[16]

[17]

[18]

[19]

[20]

[21]

[22]
[23]

[24]

J. Gosline, P. Guerette, C. Ortlepp, and K. Savage, "The mechanical design of spider


silks: from fibroin sequence to mechanical function," JExp Biol, vol. 202, pp. 32953303, 1999.
H. Koerner, G. Price, N. A. Pearce, M. Alexander, and R. A. Vaia, "Remotely actuated
polymer nanocomposites - stress-recovery of carbon-nanotube-filled thermoplastic
elastomers," Nature Materials,vol. 3, pp. 115-120, 2004.
R. A. Vaia and H. D. Wagner, "Framework for nanocomposites," Materials Today, vol.
7, pp. 32-37, 2004.
M. J. Wiggins, M. MacEwan, J. M. Anderson, and A. Hiltner, "Effect of soft-segment
chemistry on polyurethane biostability during in vitro fatigue loading," Journal Of
Biomedical MaterialsResearch PartA, vol. 68A, pp. 668-683, 2004.
I. Grillo, P. Levitz, and T. Zemb, "Insertion of small anisotropic clay particles in swollen
lamellar or sponge phases of nonionic surfactant," European PhysicalJournalE, vol. 5,
pp. 377-386, 2001.
N. Kumar, S. M. Liff, and G. H. McKinley, "Method to disperse and exfoliate
nanoparticles," U. S. P. Office, Ed. United States: Massachusetts Institute of Technology,
Filed October 18, 2005.
M. Szycher, Szycher's handbook ofpolyurethanes. Boca Raton: CRC Press, 1999.
B. Finnigan, K. Jack, K. Campbell, P. Halley, R. Truss, P. Casey, D. Cookson, S. King,
and D. Martin, "Segmented polyurethane nanocomposites: Impact of controlled particle
size nanofillers on the morphological response to uniaxial deformation,"
Macromolecules, vol. 38, pp. 7386-7396, 2005.
F. Yeh, B. S. Hsiao, B. B. Sauer, S. Michel, and H. W. Siesler, "In-situ studies of
structure development during deformation of a segmented poly(urethane-urea)
elastomer," Macromolecules, vol. 36, pp. 1940-1954, 2003.
D. J. Martin, G. F. Meijs, P. A. Gunatillake, S. P. Yozghatlian, and G. M. Renwick, "The
influence of composition ratio on the morphology of biomedical polyurethanes," Journal
ofApplied Polymer Science, vol. 71, pp. 937-952, 1999.
C. D. Eisenbach, A. Ribbe, and C. Gunter, "Morphological-Studies Of Model
Polyurethane Elastomers By Element-Specific Electron-Microscopy," Macromolecular
Rapid Communications,vol. 15, pp. 395-403, 1994.
B. Finnigan, D. Martin, P. Halley, R. Truss, and K. Campbell, "Morphology and
properties of thermoplastic polyurethane nanocomposites incorporating hydrophilic
layered silicates," Polymer, vol. 45, pp. 2249-2260, 2004.
H. D. Kim, T. J. Lee, J. H. Huh, and D. J. Lee, "Preparation and properties of segmented
thermoplastic polyurethane elastomers with two different soft segments," Journalof
Applied Polymer Science, vol. 73, pp. 345-352, 1999.
B. R. Nair, V. G. Gregoriou, and P. T. Hammond, "FT-IR studies of side chain liquid
crystalline thermoplastic elastomers," Polymer, vol. 41, pp. 2961-2970, 2000.
H. Yeganeh and M. A. Shamekhi, "Poly(urethane-imide-imide), a new generation of
thermoplastic polyurethane elastomers with enhanced thermal stability," Polymer, vol.
45, pp. 359-365, 2004.
J. L. Stanford, R. H. Still, and A. N. Wilkinson, "Effects of soft-segment prepolymer
functionality on structure - property relations in RIM copolyurethanes," Polymer, vol. 44,
pp. 3985-3994, 2003.

[25]

[26]
[27]

[28]

[29]
[30]
[31]

[32]

[33]

[34]

[35]

[36]
[37]

[38]
[39]
[40]
[41]

A. J. Ryan, J. L. Stanford, and A. J. Birch, "Copolyureas formed by reaction injection


moulding: correlations between chemical structure, thermal properties and microphase
separation," Polymer, vol. 34, pp. 4874-4881, 1993.
D. K. Lee and H. B. Tsai, "Properties of segmented polyurethanes derived from different
diisocyanates," JournalofApplied Polymer Science, vol. 75, pp. 167-174, 2000.
D. P. Knight, M. M. Knight, and F. Vollrath, "Beta transition and stress-induced phase
separation in the spinning of spider dragline silk," InternationalJournalOf Biological
Macromolecules, vol. 27, pp. 205-210, 2000.
D. Porter, F. Vollrath, and Z. Shao, "Predicting the mechanical properties of spider silk as
a model nanostructured polymer," EuropeanPhysicalJournalE, vol. 16, pp. 199-206,
2005.
S. Sinha Ray and M. Okamoto, "Polymer/layered silicate nanocomposites: a review from
preparation to processing," Progress in Polymer Science, vol. 28, pp. 1539-1641, 2003.
E. T. Thostenson, C. Li, and T.-W. Chou, "Nanocomposites in context," Composites
Science and Technology, vol. 65, pp. 491-516, 2005.
A. Aneja and G. L. Wilkes, "A systematic series of'model' PTMO based segmented
polyurethanes reinvestigated using atomic force microscopy," Polymer, vol. 44, pp. 72217228, 2003.
L. T. J. Korley, B. D. Pate, E. L. Thomas, and P. T. Hammond, "Effect of the degree of
soft and hard segment ordering on the morphology and mechanical behavior of
semicrystalline segmented polyurethanes," Polymer, vol. 47, pp. 3073-3082, 2006.
M. Song, H. S. Xia, K. J. Yao, and D. J. Hourston, "A study on phase morphology and
surface properties of polyurethane/organoclay nanocomposite," EuropeanPolymer
Journal,vol. 41, pp. 259-266, 2005.
C. P. Chen, S. A. Dai, H. L. Chang, W. C. Su, T. M. Wu, and R. J. Jeng, "Polyurethane
elastomers through multi-hydrogen-bonded association of dendritic structures," Polymer,
vol. 46, pp. 11849-11857, 2005.
R. S. McLean and B. B. Sauer, "Tapping-mode AFM studies using phase detection for
resolution of nanophases in segmented polyurethanes and other block copolymers,"
Macromolecules, vol. 30, pp. 8314-8317, 1997.
V. Trappe, V. Prasad, L. Cipelletti, P. N. Segre, and D. A. Weitz, "Jamming phase
diagram for attractive particles," Nature, vol. 411, pp. 772-775, 2001.
H. S. Lee, S. R. Yoo, and S. W. Seo, "Domain and segmental deformation behavior of
thermoplastic elastomers using synchrotron SAXS and FTIR methods," Journalof
Polymer Science PartB-Polymer Physics, vol. 37, pp. 3233-3245, 1999.
C. B. Wang and S. L. Cooper, "Morphology and properties of segmented polyether
polyurethaneureas," Macromolecules, vol. 16, pp. 775-786, 1983.
R. J. Young and P. A. Lovell, Introduction to Polymers, 2nd ed. London: Chapman and
Hall, 1991.
E. Guth, "Theory of filler reinforcement," JournalofApplied Physics, vol. 16, pp. 20-25,
1945.
C. S. O'Hern, L. E. Silbert, A. J. Liu, and S. R. Nagel, "Jamming at zero temperature and
zero applied stress: The epitome of disorder," PhysicalReview E (Statistical,Nonlinear,
and Soft Matter Physics), vol. 68, pp. 011306, 2003.

[42]

[43]

[44]

[45]

[46]

[47]

[48]
[49]
[50]

[51]

B. N. H. Wang, Rheological and morphologicalcharacterizationof hierarchically


nanostructuredmaterials, Ph.D. in Chemical Engineering, Cambride, MA:Massachusetts
Institute of Technology, 2007.
X. H. Dai, J. Xu, X. L. Guo, Y. L. Lu, D. Y. Shen, N. Zhao, X. D. Luo, and X. L. Zhang,
"Study on structure and orientation action of polyurethane nanocomposites,"
Macromolecules, vol. 37, pp. 5615-5623, 2004.
Z. S. Petrovic, I. Javni, A. Waddon, and G. Banhegyi, "Structure and properties of
polyurethane-silica nanocomposites," JournalofApplied Polymer Science, vol. 76, pp.
133-151, 2000.
I. Rhoney, S. Brown, N. E. Hudson, and R. A. Pethrick, "Influence of processing method
on the exfoliation process for organically modified clay systems. I. Polyurethanes,"
JournalofApplied Polymer Science, vol. 91, pp. 1335-1343, 2004.
M. D. Frogley, D. Ravich, and H. D. Wagner, "Mechanical properties of carbon
nanoparticle-reinforced elastomers," Composites Science and Technology, vol. 63, pp.
1647-1654, 2003.
T. Zhang, K. Xi, X. Yu, M. Gu, S. Guo, B. Gu, and H. Wang, "Synthesis, properties of
fullerene-containing polyurethane-urea and its optical limiting absorption," Polymer, vol.
44, pp. 2647-2654, 2003.
Z. Y. Tang, N. A. Kotov, S. Magonov, and B. Ozturk, "Nanostructured artificial nacre,"
Nature Materials,vol. 2, pp. 413-U8, 2003.
I. Dierking, "Liquid crystalline fractals: dilatation invariant growth structures in the phase
ordering process of 'banana-phases'," Liquid Crystals Today, vol. 12, pp. 1-10, 2003.
K. Kojio, S. Nakamura, and M. Furukawa, "Effect of side methyl groups of polymer
glycol on elongation-induced crystallization behavior of polyurethane elastomers,"
Polymer, vol. 45, pp. 8147-8152, 2004.
D. Acierno, E. Amendola, C. Carfagna, S. Concilio, P. Iannelli, L. Incamato, and P.
Scarfato, "Synthesis and characterisation of a nematic homo-polyurethane," Polymer, vol.
44, pp. 4949-4958, 2003.

5. Influence of Nano-Clay on Morphology and Mechanics of


High Performance Elastomeric Nanocomposites
5.1 Introduction
Hierarchical order induced by the addition of discotic platelets creates a percolated network that
significantly stiffens and toughens an elastomer without reducing its extensibility when its hard
micro-domains are preferentially reinforced. As is described in Chapter 4, the addition of 20 wt%
Laponite not only increases the initial modulus of Elasthane 23-fold, but strengthens and
toughens the commercial polyurethane and induces an increase in the heat distortion temperature
of more than 100 0 C. The remarkable thermomechanical enhancement and hierarchical order
exhibited in this model segmented, block-copolymer polyurethane nanocomposite prompts two
questions:

(1) what

thermo-physical

interactions

between nano-clay

and elastomeric

thermoplastic polyurethane are taking place? and (2) can these thermo-physical interactions be
exploited further?
To answer these questions, the Laponite-hard micro-domain morphology during deformation is
interrogated and evaluated and extensive mechanical testing and characterization completed.
Quantifying the influence of nano-clay on the hysteresis, time-dependence and stress-softening
behavior typical of elastomers is imperative

for these materials to be utilized in

thermomechanical applications. Furthermore, by utilizing various characterization techniques,


i.e. x-ray scattering, fourier transform infrared spectroscopy, and cross-polarized microscopy in
conjunction with various mechanical tests, the local orientational changes of Laponite within the
thin film and the influence of the nano-clay on the soft and hard micro-domains can be inferred.
Small and wide angle x-ray scattering allows the spacing between hard micro-domains and the
spacing between Laponite platelets and between hard segments to be determined as well as the
extent of soft segment orientation [1-4]. Using x-ray scattering with in situ stretching allows the
spacing between the Laponite platelets, hard segments, and hard micro-domains to be
interrogated as the thin film is progressively deformed. Song et al have completed a small angle
x-ray scattering (SAXS) study in which they found that within the elastic regime of the matrix

material the d-spacing of the clay platelets varied in a reversible manner, increasing upon
deformation but recovering to the original spacing upon removal of the imposed load [4].
Ultimately, the pairs of platelets with interspersed polymers acted as tiny nano-springs.
Quantifying the variation in the resultant platelet d-spacing, soft segment orientation, and hard
segment spacing with respect to macro-scale deformation when the nano-clay is preferentially
embedded with the polyurethane hard micro-domains allows the physical influence of the nanoclay to be determined.

Not only is x-ray scattering with in situ stretching useful, but FTIR and cross-polarized light
microscopy can be used to correlate morphological changes in the jammed Laponite network to
the thermomechanical behavior. By monitoring the intensity of the IR signals associated with the
pure and nanocomposite bond signature, such as the free and hydrogen-bonded carbonyl bonds
of the hard segment at 1733 cm'-1 and 1712 cm -1, respectively, the association of Laponite with
the polyurethane hard micro-domains can be determined [5, 6]. Subsequent observations
monitoring how the bond signature changes with respect to deformation can provide clues
regarding clay-segment association and matrix damage.
Ordered, fractal-like domains are observed within the Elasthane nanocomposites under crosspolarized light. Monitoring these birefringent regions in the thin-film nanocomposite during a
tensile test visually ties the various regions of tensile deformation, i.e. Laponite jamming, soft
segment alignment, strain hardening or hard micro-domain break-up within the nanocomposite,
to the disruption of the crystallite structures observed under cross-polarized light (shown in
Figure 5-23). These domains are comprised of polyurethane hard micro-domains, Laponite, and
conformationally-confined

soft domains. These domains, which measure approximately

20-100 jim in size, are disrupted upon large deformations as shown in the previous chapter. The
scale of these ordered domains is much larger than the size of the hard micro-domains
(20-175 nm) [7, 8], indicating that the local mechanical properties and morphology of the
ordered domains and the amorphous matrix are very different from one another.

The above morphological evaluation provides evidence to help explain how the nano-clay
interacts with the elastomer to modify the mechanical response. Physical attachment between the

100

elastomer and nano-clay can be verified with attenuated reflectance Fourier transform
spectroscopy (ATR-FTIR). X-ray scattering and cross-polarized microscopy help tie the
morphological evolution upon deformation to the mechanical behavior of an elastomer when
nano-clay preferentially reinforces the hard micro-domains. It is important for the continued
advancement of engineering thermoplastics that these thermo-physical interactions between the
elastomer and nano-clay are not only understood but able to be modeled. A one-dimensional
constitutive model that augments the previous polyurethane model developed by Qi and
Boyce [9] is discussed in Appendix A-3. This knowledge coupled with the appropriate model
offers a means for the development of new devices that exploit the thermo-physical interactions
between elastomers and nano-clay.

5.2 Mechanical and Morphological Characterization Techniques


Thermoplastic polyurethanes demonstrate strong hysteresis, time dependence, and cyclic
softening [9]. In section 5.2 a series of uniaxial tensile tests as well as morphological
characterization techniques are described. In the following section, 5.3, the results which
quantify these features in the model nano-clay filled elastomer are conveyed and discussed.
5.2.1 Mechanical Test Descriptions
Uniaxial tensile tests were conducted using a computer controlled Zwick mechanical tester,
model Z010, with screw tightened grips (Model No. 8133). Each grip had one flat polyurethane
jaw face and one convex aluminum face. The jaw faces ensured a line of contact across the width
of the thin film specimen and eliminated slip. Specimens were subjected to constant nominal
strain rate loading and unloading cycles. Test specimens were rectangular and approximately
5 mm in width and measured 120-180 pm in thickness. The specimen thickness did vary through
the length by +20 pm. Measurements made with a video extensometer proved that the modulus
calculated via engineering stress-strain results utilizing the average specimen thickness and the
Zwick crosshead as the means to measure the change in specimen length was valid (See
Appendix A-2). Consequently, the results reported are in terms of engineering stress-strain and
when true stress-strain is reported, the material was assumed incompressible. True strain was
defined as the logarithm of the current length over the initial length, where the current length of
the sample was taken to be the crosshead distance. True stress was calculated by multiplying
nominal stress by the stretch ratio. Elastomers are generally incompressible. Consequently it is
reasonable to assume the Poisson ratio ranges from 0.48 to 0.50. The nominal strain rate

101

maintained for each test was 0.0167s - unless otherwise noted.


5.2.1.1 Hysteresis
The polyurethane nanocomposites were stretched to 1.8 and 2.6 times the original specimen
lengths and then unloaded. Additional strain recovery occurred with time after unloading, so the
specimens were unloaded to a slightly positive force of 15 mN so that specimen buckling was
avoided. This 'no-load condition' was held for 180 seconds before subsequent loading. Each
load-unload cycle was completed twice at a specified specimen stretch before the larger
subsequent stretch was applied. The response after the first and second cycle (N= 1 and N=2) are
reported and the residual strains were measured prior to re-loading.
5.2.1.2 Softening
The polyurethane nanocomposites were stretched to 1.1, 1.2, 1.4, 1.8, 2.2, 2.6, and 3.0 times the
original specimen length and then unloaded. Additional strain recovery occurred with time after
unloading, so the specimens were unloaded to a slightly positive force of 15 mN so that
specimen buckling was avoided. This 'no-load condition' was held for 180 seconds before
subsequent loading. Each load-unload cycle was completed five times at a specified specimen
stretch before the larger subsequent stretch was applied. The response after the first and fifth
cycle (N= 1 and N=5) are reported and the residual strains were measured prior to re-loading.
5.2.1.3 Rate Dependence
The rate dependence of the material at low strain rates was evaluated. Each nanocomposite
specimen was stretched 2.0 times the original specimen length and unloaded five times at three
different nominal tensile strain rates: 0.0017 s-1, 0.0167 s-', and 0.1667 s- 1.
5.2.1.4 Time Dependence
During loading or unloading if the test is suspended, time dependence in the stress response (if
the strain is held constant) or in the, creep behavior (when the stress is held constant) ensues.
Stress relaxation tests were conducted during loading and unloading up to a maximum stretch of
2.0 in one instance and 3.0 in another instance. These load-unload cycles with intermittent holds
were repeated four times and stretched at a nominal strain rate of 0.0167 s- 1 with intermittent 600
second holds at stretches of 1.1, 1.2, 1.4, 1.6, 1.8, and 2.0 in one instance and 1.1, 1.2, 1.4, 1.8,
2.2, 2.6 and 3.0 in another instance.
5.2.1.5 Relaxed Response andEquilibrium Paths
In stress relaxation tests, the stress relaxes towards an equilibrium state during the intermittent
holds. The stress relaxation behavior during the 1st cycle and the subsequent cycles of the test
differ after cycling between stretches of 1.0 and 2.0. Consequently, the stress relaxes towards

102

two distinct softened paths. The softened stress at each intermittent hold was taken as the stress
value prior to loading or unloading to the subsequent stretch. The equilibrium paths then are
defined to fall halfway between the softened response of the first softened load-unload path at
each applied strain and the subsequent softened load-unload path.
5.2.2 Morphological Characterization Techniques
5.2.2.1 Wide Angle X-ray Scattering
WAXS data was acquired at the GI beamline at the Cornell High-Energy Synchrotron Source
(CHESS). At the CHESS GI beamline, k = 1.23 A, the beam was collimated to a horizontal
divergence of 1 mrad via the wiggler 'K' factor and horizontal-focusing optics. This natural
collimation was supplemented by use of two 1 mm slits and a third, which functioned as a guard
slit. The relative X-ray intensity was measured at positions (0) just before the sample, using a
helium-filled proportional counter, and (2) upstream of the collimation slits, using a nitrogenfilled proportional counter.

A MedOptics CCD detector was used to measure the scattered

radiation, with pixel size 47.19 ptm pixel-' and size 1024 x 1024 pixels.
A home-built stretcher was used to elongate the 0.1-0.2 mm thick samples to static positions that
subsequently increased the specimen length to 1.1, 1.2, 1.4, 1.6, 1.8, 2.0, 2.5, 3.0, 4.0, 5.0, 6.0,
7.0, 8.0, and 9.0 times the original specimen length. Obviously, the samples slipped because
tensile failure data presented in the previous Chapter indicates that the films fail consistently at
nominal stretches of approximately 5.0. Consequently, the resulting scatter patterns can only be
qualitatively associated with various tensile deformation regions, rather than correlated to
specific strains. Scattering patterns were collected for 20 seconds in WAXS (time-averaged). A
background pattern was collected for each sample environment, for a collection time equal to the
sample collection time. A dark (blocked beam) pattern was measured for each collection time
employed.

The appropriate dark pattern was subtracted from each sample and background

scattering pattern. Each sample pattern was then corrected using the following formula:
Correlated Scattering Pattern= Sample Pattern - T - Background Pattern.

(5.1)

The transmission ratio, T, was taken to be the following ratio


,sample
OT
IO, background

sample
2

(5.2)

0 background

103

where Io and 12 is the measured beam intensity at the positions denoted by numerical subscript 0
and 2 mentioned previously and lo' is the intensity measured at position 0 during collection.
Data was reduced from 2D (intensity vs. 20, X) to 1D (intensity vs. 20) format by integrating
over all values ofX , for each value of 20, where 20 is the scattering angle and ; is the azimuthal
angle. The data was also reduced to ID format by integrating over values of 20

between

15.09' and 15.830 (d-spacing = 0.447 nm to 0.468 nm) and 16.56' and 17.280 (d-spacing = 0.409
nm to 0.427 nm) for each value of ;.

Small angle x-ray scattering data with in-situ stretching was also collected, but the distinct SAXS
peaks observed in the pure polyurethane were not identifiable in the Laponite filled sammple
(See Appendix A-4). This is most likely the result of electron-density contrast matching between
the soft and hard blocks due to the disrupted morphology encouraged by the addition of the rigid
clay platelets as seen in atomic force microscopy phase images [10]. If a fraction of the soft
segment chains are constrained within the Laponite-hard micro-domain substructures, as is
suggested via cross-polarized microscopy, the pseudo electron density of the continuous domain
is raised, limiting the perceived contrast between the soft and hard domains. Consequently, the
SAXS data is not reported.
5.2.2.2 Attenuated Total Reflectance FourierTransform InfraredSpectroscopy
The nanocomposite films were stretched to a static position using a home-built, manual stretcher
and characterized with attenuated total reflectance Fourier transform infrared spectroscopy
(ATR-FTIR). A Nicolet Nexus 870 spectrometer with a DuraSampl!R II Diamond ATR from
SensIR Technologies was used to collect data between wave numbers of 4 cm -' and 4000 cm - .
The home-built stretcher was comprised of a digital caliper with cylindrical rods used as grips
attached to the jaws of the caliper. The grip on each jaw consisted of a small cylindrical rod with
two threaded holes at either end set parallel to the measuring flat of the jaw. A specimen,
measuring 25 x 5 x -0.15 mm 3 was placed between the threaded holes and as the screws through
the rod were tightened the rod made a line of contact along the width of the specimen. Then the
thin-film specimen was deformed by manually opening the jaws of the caliper. Each position
was held by tightening the screw on the caliper that stops the lower jaw from moving and the
thin film characterized. The area in which the film was characterized was circled with a black
permanent marker. Upon deformation, the marker smeared, making it difficult to interrogate the
same area at each subsequent increasing stretch. As a consequence, the bond signatures did not
104

change systematically, but seemed to hover around the static signature despite the magnitude of
stretch. Therefore, only the static signature is reported.
5.2.2.3 Cross-PolarizedMicroscopy
The long range order, deformation, and crystalline morphologies of the nanocomposites during
deformation were examined using a Carl Zeiss Axioskop 2MAT Polarizing Microscope with
cross-polarized light and the same home-built, manual stretcher used with ATR-FTIR. The
nanocomposite films were stretched to a static position and then the mid-section was imaged.
Although the grips did induce local deformation at the line of contact, specimen slip was
eliminated. Because the middle of the specimen was imaged, the resultant stretches reported
describe the overall specimen stretch and the value is greater than that actually achieved by the
area interrogated.

5.3 Results and Discussion


5.3.1 Hysteresis
Figure 5-1 shows

the

tensile

engineering

stress-engineering

strain

behavior

of

Elasthane/Laponite nanocomposite films at engineering strains equal to 0.8 and 1.6. The loading
curves exhibit an initial stiff response that increases its magnitude with Laponite concentration
and rolls over to a more compliant behavior. This rollover in the 15 wt% nanocomposite appears
as a yield point typical of amorphous polymers due to the jamming of nano-clay platelets and
initiation of localized deformation. As the Laponite concentration increases, the rollover occurs
at smaller and smaller strains. The pure polyurethane rolls over at a strain of -0. 15 while the 15
wt% Laponite filled rolls over at -0.03. The unloading paths exhibit large hysteresis and
significant residual strain that both increase with increasing Laponite concentration. This
behavior is obvious in Figure 5-2 when the behavior of the pure and 15 wt% Laponite-filled
behavior is compared directly.
When the Laponite concentration is less than the percolation limit, the increase in hysteresis, or
energy absorbed, compared to that of the pure polyurethane increases significantly without
considerably increasing the residual strain. In contrast, when the Laponite concentration is
greater than the percolation limit, there is both a marked increase in energy absorbed and residual
strain. The addition of nano-clay reinforces the stress response of Elasthane, but also locally
damages the polyurethane matrix, especially at concentrations above the percolation threshold.

105

0.0

0.2

0.6

0.4

0.0

0.8

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Engineering Strain

Engineering Strain

Figure 5-1: Hysteresis of pure Elasthane to composites filled with 4, 10, and 15 wt% Laponite when thin-film
specimens are stretched 80% and 160%.

CU
0.
o)

C_

O(D
1

0
U.
I

C
Wm

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Engineering Strain
Figure 5-2: Comparison of the hysteresis of pure Elasthane and a composite filled with 15 wt% Laponite
when thin-film specimens are stretched 80% and unloaded and then to 160%.

106

5.3.2 Softening

0.0

0.5

1.0

1.5

Engineering Strain

Engineering Strain

.a
c

C
IU

CI
C
LU

0.0

0.5

1.0

1.5

0.0

2.0

0.5

1.0

1.5

2.0

Engineering Strain

Engineering Strain

Figure 5-3: Cyclic softening behavior of pure Elasthane and nanocomposites filled with 4, 10, and 15 wt%
Laponite when the specimens are stretched to nominal or engineering strains equal to 0.1, 0.2, 0.4, 0.8, 1.2, 1.6
and 2.0. The response shown in black is that of an independent sample stretched to failure. The stress-strain
response shown in red is that upon the first cycle of loading and unloading at each stretch (N=1) and that in
blue corresponds to the fifth cycle (N=5).

Figure 5-3 exhibits the engineering stress-strain response of Elasthane filled with 0, 4, 10 and 15
wt% Laponite during the first and fifth cycle of loading-unloading tests to engineering strains of
107

0.1, 0.2, 0.4, 0.8, 1.2, 1.6, and 2.0. Typical of most elastomers, the response of the subsequent
cycle is more compliant than that observed in the first deformation cycle [11, 12]. The extent of
softening that occurs between the first and second deformation cycle is greatest, but after a few
cycles the softened response stabilizes, as shown in Figure 5-4. The residual or plastic strain too
exhibits the largest increase between the first and second deformation cycle and then stabilizes
during subsequent cycling. There is slight softening in the material response upon unloading, but
when compared to the softening upon loading it can be considered negligible. When the previous
maximum strain achieved during previous cycles is surpassed the stress response approaches that
typical of a non-softened specimen and the extent of softening increases with applied stretch.
Sly

12

O wt%
S2wt%/
4 wt%
M1

14

0.6

10 wt%/

15 wt%

0.5

-j

10
CL
S8

'

2wt%

4 wt%
10 wt%

15wt%
/

I I

'

*
--

.00.4

0.2

4
2

-,

0.1

0.0
0
I

0.0

0.5

1.0

_n i

I a

1.5

I I

2.0

0.0

Applied Strain

0.5

1.0

1.5

2.0

Applied Strain

Figure 5-4: Evolution of the maximum stress and plastic strain as 0, 2, 4, 10, and 15 wt% Laponite-filled
nanocomposites are cycled 5 times at applied nominal strains equal to 0.1, 0.2, 0.4, 0.8, 1.2, 1.6, and 2.0. The
plastic strain is that obtained after the specimen is held under a no-load condition for 180 s. For ease of
viewing the resulting trends of maximum stress and plastic strain are plotted against modified applied strain
(modified applied strain = applied strain + 0.01).

When nano-clay is added to the polyurethane the softening behavior typical of elastomers is still
apparent. However, increasing the nano-clay concentration increases the magnitude of the
softened, stress-strain response of the nanocomposites from that exhibited by the pure
polyurethane. The extent of softening between the first and fifth cycle too increases with
Laponite concentration, particularly when concentrations exceed the percolation limit. The

108

addition of Laponite similarly increases the plastic strain. Obviously the material response of
these elastomers with and without nano-clay depends strongly upon the material strain history.
5.3.3 Rate Dependence

Co

,-

Co
C

U)
0

._
C)
0)

Nominal Strain

Nominal Strain

CL

0
4)
C

,_

0)
a.
~fo

C
0
a)
UJ

0.0

0.2

0.4

0.6

0.8

1.0

Nominal Strain

Nominal Strain

Figure 5-5: Rate dependence of 0, 4, 10, and 15 wt% Laponite-filled Elasthane whenq stretched to a nominal
-1,
-1
strain of 1.0 at three different nominal strain rates: 0.0017 s , 0.0167 s and 0.1667 s-

The stress-strain behavior of the Elasthane/Laponite nanocomposites at three low strain rates is
shown in Figures 5-5, 5-6 and 5-7. Upon loading the stress-response increases with increased

109

strain rate, a response typical of polymers [13]. The observed increase in stress with rate is
greatest upon loading, enhancing the flow stress by 1-2 MPa with each order of magnitude
increase in rate. Meanwhile the increase is so slight upon unloading that it can be assumed
negligible. Consequently, loading is more rate dependent than unloading.
The flow stress or engineering stress at an engineering strain equal to 0.3 at each strain rate
increases with the Laponite volume fraction as shown in Figure 5-6. This increase in flow stress
is similar to the increase in viscosity exhibited in filled solutions. Einstein showed that the
viscosity of a dilute solution filled with spheres increases linearly with the filler volume fraction
[14]. Similarly, the flow stress increases linearly, indicating the higher concentration solids are
more rate sensitive.
12

0.
e
*10

CO
U

to

S6

In 4
C
M
44..

2
0
-1

0.0017s 0.0167 s "'

0.1667 s

$,(vol%)

Figure 5-6: The flow stress, or engineering stress at 6 = 0.3, at the various strain rates is plotted against the
Laponite concentration. The flow stress appears to increase linearly with volume fraction.

At the slowest strain rate the stress-strain response deviates from that reported in Chapter Four
(nominal strain rate = 0.0167 s-1) where the modulus and flow stress increases with Laponite
concentration and the slope of soft segment stretch and re-orientation region remains constant
despite the addition of nano-clay. As shown in Figure 5-7, at the slowest rate the 0 and 4 wt%
filled nanocomposite exhibits a larger hardening slope than that of the more concentrated
nanocomposites. Deformation at this slow strain rate is like testing the specimen at a slightly

110

elevated temperature (T > Tamb); the Laponite and hard micro-domains are able to re-orient with
the soft segment and retard the hardening.

a
I
(0

,,=
a
a

0_0

02

0.4

0.6

Nominal Strain

0.8

1.0

Nominal Strain

Nominal Strain

Figure 5-7: Comparison of the engineering stress-strain response of 0, 4, 10, and 15 wt% Laponite-filled
-1
Elasthane when stretched to a nominal strain of 1.0 at three different nominal strain rates: 0.0017 s ,
.
0.0167 s-1, and 0.1667 s-

For a more complete analysis of the rate dependence of these nanocomposites evaluation at
moderate and high strain rates are needed [15].
5.3.4 Time Dependence
Tensile stress relaxation tests were conducted at sequentially increasing nominal strains and then
unloaded at those same strains. Figure 5-8 show the corresponding strain profile, stress response
and stress-strain response. A majority of the stress relaxation occurs within the first 20 seconds.
During loading, the stress decreases while the strain is held constant and upon unloading, the
stress increases. Note that there was no intermittent hold at a nominal strain of 2.0 prior to
unloading. Consequently, the stress-recovery during unloading is masked by the material
relaxation that continues to evolve. The extent of stress-relaxation is greater than that of stressrecovery due to evolving chain orientation and matrix damage.

111

time, s

Engineering Strain

01

=)
0)
C
0

0
uJ
.5,
w,

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

5500

6000

6500

time, s
Figure 5-8: Applied strain history for stress relaxation tests as well as the engineering stress response vs. time
and applied strain for nanocomposites containing 0, 4, 10 and 15 wt% Laponite.

As the magnitude of the applied strain increases, the relaxed stresses of the Laponite-filled
nanocomposites appear to converge to a value greater than the pure polyurethane. There also
appears to be two relaxation mechanisms and consequently two relaxation time constants that
depend on both Laponite concentration and the magnitude of stretch applied. It is difficult to
clearly distinguish the influence of Laponite on the stress relaxation behavior without further
analysis. The experimental relaxation data at each nominal or engineering strain was fit to the
112

following equation that describes the rheological model in Figure 5-9:


=0 Eo+ , ex p

(5.3)

+E2 exp -t

where the specimen stretch, A,is equal to 1+e, where e is the applied engineering strain.
30'

1'
I~

'

' I'

'

'

'

'

'

'

l'

'

I '

it

2.8F

I '

- EY'I

2.6-

2.4-

2.2

2.0
-50

I
0

50

100

150

200

350
250 300
ds)e
* son

400

I
450

,i l
500

I
550

I ,
600 650

stress relaxation and


Figure 5-9: One-dimensional description of the rheological model of the experimental
corresponding fit to the pure Elasthane stress relaxation after the specimen is stretched 10%.

After fitting the relaxation data for composites containing 0, 4, 10 & 15 wt% Laponite upon
loading to stretches equal to 1.1, 1.2, 1.4, 1.8, 2.2, and 2.6 the resulting moduli and time
constants were plotted against the applied stretch. Below in Figure 5-10, find the time constants,
of the
r, and r2 . The short relaxation time constant (12-16 s) appears to depend on the magnitude
stretch applied. The long relaxation time constant displays significant scatter, but appears to
remain constant at -200 s despite the magnitude of stretch applied. The magnitude of these
relaxation times are within the same range that Xia et al observe for a similar polyurethane [16].
The long relaxation time corresponds to the hard micro-domains and the short relaxation time
corresponds to the soft segment. As the magnitude of the applied stretch increases but does not
reach the strain hardening regime, the soft segments stretch, align and orient while the hard
segments rotate but do not break-up. Consequently, the long relaxation time exhibits scatter and
remains constant while the short relaxation time increases with stretch. It is important to note that
the initial data point taken at a nominal stretch of 1.1 falls within the region of yield for the nanoclay filled materials. This flow, in conjunction with the stress relaxation, combines to exhibit
effectively smaller relaxation time constants. In future analyses these three data points are
excluded.

113

L4U

II
SO

220

wt/o

4wt%

10 wt%/

15wt%

-A

,I

f ......
I
I

,
I

-------

Owt%
S4wtO/
A 10wtO/
v l5wt%/o
5
-

Ae 0

vn

1.0

1.2 1.4

1.6 1.8 2.0

2.2 2.4 2.6

2.8 3.0

1.0 1.2 1.4

1.6 1.8 2.0 2.2

2.4 2.6 2.8 3.0

Figure 5-10: Two characteristic time constants observed in the stress relaxation behavior of the polyurethane
nanocomposites. The first relaxation time constant corresponds to the hard micro-domain relaxation and the
second faster relaxation constant corresponds to the soft segment that becomes less mobile upon stretch and
orientation.

In order to extend the analysis of the stress relaxation behavior exhibited by these
nanocomposites in tension to other modes of deformation, the relaxation behavior was examined
as a function of orientation [17]. Arruda and Boyce have shown that the eight-chain network
model depicted in Figure 5-11 captures the finite strain behavior of polymers subject to complex
deformations using only two model parameters [18].

Figure 5-11: Schematic of the stretched eight-chain network as depicted in 1181.

The three chain angles that correspond to the angles between the three principal directions (1, 2

114

and 3) and a vector pointing from the center of the representative volume element (RVE) along
one of the eight chains to its vertex can be described by

A = acos

(5.4)

A2

where fli is the chain angle with respect to a principal direction, i, and 2i is the stretch describing
the RVE deformation with respect to a principal direction [19]. Dupaix and Boyce found it useful
to utilize a physically based orientation parameter, a, to describe the molecular relaxation
observed in amorphous polymers like poly(ethylene terephthalate)-glycol [17]. They define the
orientation parameter as the compliment of the maximum chain angle:
a =

(5.5)

-m-ax

The orientation parameter as a function of principal stretch for an incompressible material


deformed in tension, compression and equibiaxial tension and compression is shown in
Figure 5-12. Note that the orientation parameter when a polymer solid is not deformed is equal to
0.615 rad and decreases upon deformation. It is not convenient to plot the stress relaxation time
constants and moduli against a non-linear, decreasing function that starts at a maximum value
and approaches zero. For ease, the orientation parameter was normalized by the undeformed
orientation parameter (ao-a), so that the undeformed normalized orientation parameter starts at
0 rad and monotonically increases to 0.615 rad.
When the two, characteristic relaxation time constants are plotted against the normalized
orientation parameter in Figure 5-13 it is clear that zr = 200 s and r2 increases as the soft
segment stretches and aligns. Initially the soft segment exhibits a relaxation time constant of
-12 s but upon deformation the value increases according to the following relation:
T2

(5.6)

= 12.3 + 9(a -a).

This trend is observed in the pure Elasthane as well as nanocomposites filled with 4, 10, and 15
wt% Laponite.

115

0.5

0.5

0.4

0.4

tS 0.3

0.3

0.2

0.2

0.1

0.1

0.0
0

0.0
1

2
it

normalized orientation
Figure 5-12: Relationship between the orientation parameter, a, as well as the
material. Note that
incompressible
an
for
RVE
eight-chain
an
of
stretch
parameter, ao-a, and the principal
less than unity.
equibiaxial compression is equivalent to tension but the principal stretch is given by a value

This increase in relaxation time with soft segment stretch and alignment is either due to strain
induced crystallization (the specimen film transitions from transparent to opaque upon
stretching) or interrogation of the longer soft segment chains upon increased stretch [20-22].
I0

I
'

240-

17

230

16

220-

15

S20C-

210-

190-

13

12
-

11

S0 wtO/%
wt%0
4wtY

10

180
*
*

170 -

160

0.0

0.1

10 wt%

rv 15 wt%/
0.2

14

0.3

0.4

*
A

9
F
0.0

a.-a

A
v

I
0.1

I I
0.2

I I
0.3

00 wt%
4 wt%/o
lOwt%_
15wt%
I
0.4

a.-a

behavior of
Figure 5-13: Evolution of the two characteristic time constants observed in the stress relaxation
the polyurethane nanocomposites with orientation.

116

The three relaxation moduli given in equation 5.3 are plotted against the tensile stretch, as
depicted in Figure 5-14.
r
1L

",

*
e

0 wt%
4 wt%

Owt%
4wt%
4
4wt%

S Owt/.
*

A 10 wt%
S15 wt%-

4wt%
wt%-

A10

S10 Wt%

-0

15 wt%

A
A
-3
CI
0.

-0

Ui
2-

S V
A
1.0.

,.o I-M

A
VO

U
U

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.8

2.6

3.0

1.2

1.0

1.4

1.6

1
1.8

2.0

2.2

2.4

2.6

2.8

3.0

1.2

1.0

1.4

1.8

1.6

2.0

2.2

2.4

2.6

2.8

3.0

Figure 5-14: Evolution with tensile stretch of the three moduli that capture the stress relaxation behavior of
the polyurethane nanocomposites.

The resultant moduli (given by Eo, El, and E2) decay with increasing stretch. The addition of
nano-clay increases the moduli significantly at small stretches and minimally at large stretches.
When the natural logarithm of the moduli is plotted against the natural logarithm of the
normalized orientation parameter the trends fall linearly, indicating that the moduli scale with
orientation following a power-law. The addition of nano-clay increases the sensitivity of the
material to orientation induced softening. This behavior is exhibited by the increasing slope of
the linear fits shown in Figure 5-15. The parameters, A (the y-intercept) and m (the slope) of
these fits are given in Table 5-1. The parameters were plotted against the Laponite volume
fraction and equations capturing their dependence on concentration and aspect ratio determined.
3
0
A

S4wt%.

10wt%

15wtI%

4wt%
10wt/.
S15wt%

[
I-

-4

-3

-2

In(a 0o-a)

-1

-4

-3

.2

In(a0o-a)

-1

-4

-3

-2

-1

In(a 0-a)

Figure 5-15: Power-law decay of the polyurethane nanocomposite moduli with orientation. The magnitude of
the modulus and the extent of softening with increased orientation depend upon the Laponite concentration.

117

Laponite Fraction, q9
wt

Fit Parameters
E

vol

0.00
0.04
0.10
0.15

0.000
0.017
0.045
0.069

A
0.37
0.51
0.34
0.29

Laponite Fraction, P
wt
vol
0.00
0.04
0.10
0.15

Ei

m
-0.55
-0.58
-0.70
-0.74

0.000
0.017
0.045
0.069

A
0.44
0.44
0.44
0.44

m
-0.54
-0.59
-0.67
-0.74

A
-0.85
-0.64
-0.78
-0.71

m
-0.23
-0.30
-0.56
-0.64

Parameters Based on q)and a


E
A
m
-0.78
-0.22
-0.78
-0.33
-0.51
-0.78
-0.78
-0.66

A
-0.84
-0.73
-0.83
-0.77

m
-0.20
-0.38
-0.63
-0.74

E,
A
-0.80
-0.80
-0.80
-0.80

m
-0.23
-0.37
-0.59
-0.78

Table 5-1: Parameters describing the power-law decay of the polyurethane nanocomposite moduli with
increased orientation. The parameters in the upper table are determined via fits to experimental data. The
parameters in the lower table are approximations to the experimental fit values given by empirical trends
that quantify the influence of Laponite concentration, (p,and aspect ratio, a.

The linear decay of the log-log plot of polyurethane moduli with orientation provides the
following relation:
InE = Ai + m, In(ao - a),

(5.7)

where A, is the y-intercept and mi is the slope. The resulting relationship between each of the
three moduli, which capture the nanocomposite stress relaxation behavior, and the normalized
orientation parameter is given by
E = C (ao - a)" ,

(5.8)

where the constant, Ci = eA and the slope, mi prescribes the extent of decay. It is assumed that
the contribution to each modulus, Eo the softened polyurethane response, E1 the hard microdomain contribution, and E2 the soft segment contribution is independent of nano-filler
concentration or aspect ratio. The parameters Ai can be assumed constant: AEo = 0.44,
AE,

=-0.78, and AE = -0.80.

The validity of this assumption relies upon evidence that the

addition of nano-filler does not significantly increase or disrupt polyurethane crystallinity or


microphase segregation (discussed further in 5.3.9). The extent of modulus decay with
orientation is likely dependent upon nano-filler concentration, aspect ratio and polydispersity.
Consequently, the slopes obtained for composites containing 0, 4, 10, and 15 wt% Laponite (with

118

aspect ratio: a = 25) for each distinct modulus was plotted against the nano-filler volume
, and the equation below fit:

fraction,

(5.9)

mi = c,i (1+ a c 2,i),

where ci and c2,i are constants empirically determined and a is the nano-filler aspect ratio [23].
The empirical relations for the parameter mi are described in the following three equations:
= -0.54(1+0.21a)

(5.10)

mE = -0.22(1+1.14a )

(5.11)

mE = -0.22(1+1.40a ).

(5.12)

The resultant parameter values determined from these relations are listed in the lower portion of
Table 5-1. There is good agreement between the moduli calculated from these empirical relations
and the experimental data as shown in Figure 5-16.

a.
UJ

0.0

0.1

0.2

a--a

0.3

0.4

0.0

0.1

0.2

ao-a

0.3

0.4

0.0

0.1

0.2

0.3

0.4

ao--a

Figure 5-16: Comparison of the polyurethane nanocomposite moduli evolution with increased orientation
determined from experimental data (solid symbols) and that predicted via empirical relations (solid thin line).

The stress relaxation behavior of the 0, 4, 10, and 15 wt% Laponite-filled nanocomposites
indicates that the polyurethane behavior can be captured well using two characteristic relaxation
mechanisms. The first corresponds to the hard micro-domain relaxation (rl - 200s) and remains
constant despite Laponite concentration when the extent of matrix deformation and orientation
has not induced hard segment break-up. The second mechanism corresponds to the uncoiling and
relaxation of the soft segments (Z2 -12 s). This characteristic relaxation time constant increases
with deformation and soft segment orientation. The moduli governing the material stress
response exhibits a power-law decay with orientation. Ultimately, a constitutive model that

119

accounts for these two distinct relaxation mechanisms is necessary.


5.3.5 Relaxed Response and Equilibrium Paths

CL

C
4)
U)
0

C
0I
0,

,
iJ

Nominal Strain

Nominal Strain

Nominal Strain

Nominal Strain

Figure 5-17: Stress relaxation behavior of pure polyurethane and that filled with 4, 10, and 15 wt% Laponite
through five load-unload cycles. The points used to distinguish the softened response after the 1st and 5 th
cycle.

Figure 5-17 depicts the stress relaxation behavior of the 0, 4, 10, and 15 wt% Laponite filled
composites during the first and fifth cycle of testing. The relaxed value of stress at a particular

120

strain depends upon the deformation history of the material. Similar to the softening results
discussed previously, the recovered response upon unloading in each cycle is virtually the same.
do
Similarly, the subsequent relaxed stresses achieved upon loading for all loading cycles that
not surpass the maximum stretch achieved upon initial loading are nearly identical. The softened
or relaxed responses for each of the Elasthane nanocomposites are compared in Figure 5-18.

0.

W
(0
)

w)

0.0

0.2

0.4

0.6

0.8

0.0

1.0

0.2

0.4

0.6

0.8

1.0

Nominal Strain

Nominal Strain

determined from
Figure 5-18: Softened response of the 0, 4, 10, and 15 wt% Laponite-filled nanocomposites
between the
halfway
falls
path
equilibrium
The
cycle.
th
and
5
the stress relaxation behavior during the 1 st
response.
softened loading and unloading

by
Increasing the Laponite concentration increases the initial stiffness and flow stress exhibited
the relaxed or softened path upon initial loading. The addition of Laponite at concentrations
the
above the percolation limit decreases the hardening slope of the relaxed response as well as
stress achieved upon un-loading. This is likely due to increased localized plastic deformation
a
caused by the presence of the rigid filler. As a consequence the softened responses converge at
nominal strain equal to unity upon initial loading and the second path converges to the same
stress. The softened path upon subsequent loading for each composite is quite similar. Despite
these differences upon initial loading, Laponite does not significantly influence the equilibrium
path as shown in Appendix A-3.

121

5.3.6 X-ray Scattering


Morphological studies allow for further elucidation of the structural characteristics of these
semicrystalline polyurethane nano-clay composites while undergoing tensile deformation. Martin
et al have reported that Elasthane 80A exhibits a broad, weak hard segment diffraction pattern
when the material is compression molded and annealed at 135'C [24]. The broad peak suggests
minimal long range order and the peak which corresponds to a hard segment lattice d-spacing of
0.439 nm falls within the expected range exhibited by 4,4'-methylene diphenyl diisocyanate hard
segments chain extended with 1,4-butanediol (MDI-BDO) [3, 25]. WAXS spectra of a pure and
8 wt% Laponite-filled composite are shown below in Figure 5-19 as the materials are stretched.
The diffraction patterns are analyzed at four positions that correlate to the undeformed, yield,
soft segment orientation, and strain hardening regimes of the stress-strain profile. The diffraction
patterns are displayed in full in Appendix A-4. The pure polyurethane exhibits a broad peak at a
d-spacing of 0.447 nm (20 = 15.8') and shoulder at 0.421 nm (20 = 16.80). The presence of two
peaks is a result of slow solvent casting and microphase segregation. When the material is
stretched the intensity of this peak decreases, indicating hard domain disruption and reorientation as well as specimen thinning. When the stretch achieved corresponds to that within
the strain hardening regime, a broad peak that spans 100 and peaks at a d-spacing of 1.21 nm (20
= 5.8') results. It is not clear what structure feature this peak corresponds to because the interdomain spacing between hard domains upon significant stretch and re-orientation is typically
10-20 nm and between crystallized soft segments is -0.4 nm. However, this peak it is not
observed in the Laponite-filled samples.
Reflections indicative of Laponite powder are not observable in the 8 wt% Laponite-filled
nanocomposite diffraction pattern; specifically the absence of the d001 diffraction peak suggests
the nano-clay is fully exfoliated. It is important to note that the dOO1 and d020/1 10 peak reported
by the manufacturer as 1.28 nm and 0.452 nm are instead observed at 1.38 nm and 0.390 nm
respectively. This discrepancy is unexpected and may result due to a difference in specimen
preparation. Typically powders are ground to a fine dust before x-ray interrogation, but because a
sharp pattern was captured from the powder as constituted this step was avoided. However, the
addition of nano-clay does increase the intensity of the hard segment reflections when compared

122

to those of the pure specimen. Similarly, the hard segment melting endotherms observed with
differential scanning calorimetry exhibit increased intensity when nano-clay is added.

8 wt%

0 wt%

12

15

18

21

24

27

30

28, degree

12

15

18

21

24

27

30

20, degree

Figure 5-19: WAXS spectra of pure and 8 wt% Laponite-filled Elasthane as the material is deformed. The
spectra shown corresponds to four positions along the stress-strain curve: 1. undeformed (black circle), 2.
yield (red), 3. soft segment re-orientation (green), and 4. strain hardening (blue). The diffraction pattern of
Laponite powder is included with the 8 wt% Laponite-filled composite spectra.

The intensity of a hard segment melting endotherm increases when the crystallite is more pure
and/or perfect [26]. The preferential association of the Laponite platelets to the hard microdomains at a concentration greater than the percolation limit likely restricts mobility and
prevents crystallite adulteration. The addition of nano-clay increases the hard segment lattice
spacing from 0.447 nm (20 = 15.80) to 0.456 nm (20 = 15.50) and maintains the secondary
spacing at 0.421 nm (20 = 16.8'). Upon increased deformation the intensity of the broad peak
123

indicative of hard segment lattice spacing too decreases in the 8 wt% sample. If a broad peak of
low intensity does occur near 6', like that in the pure polyurethane at the greatest extent of
deformation, it is masked by the presence of Laponite. Small angle x-ray scattering would help
to further identify the hard micro-domain evolution, but as mentioned in the previous section this
data in the Laponite filled samples was indistinguishable (See Appendix A-4). Consequently,
inference from azimuthal analysis of the WAXS spectra must be relied upon.

Azimuthal scans within d-spacings of 0.447 nm and 0.468 nm (15.09' to 15.830) and 0.409 nm
and 0.427 nm (16.560 and 17.28 ) help to elucidate the how the hard segments evolve upon
deformation, shown in Figures 5-20 and 5-21. The peaks at 0.447 nm for pure Elasthane and
0.456 nm for the 8 wt% filled nanocomposite are initially randomly oriented indicating the
matrix is isotropic prior to deformation. Upon deformation, they orient with the equator. The
shoulder at 0.421 nm in both materials also begins randomly distributed but upon deformation
orients with the equator. A six-fold symmetry begins to develop in the composite, typical of soft
segment alignment, at the largest stretch. This result indicates that the nano-clay does not restrict
soft segment alignment, but rather its presence amplifies the chain alignment as well as the hard
segment intensity.
n

to/. n

ARRnm

Rlr u%
nU
A8nm
'7W
#U
. AA7-_4
ml
.BI
I
I

4
(U

cU

(U
C5

a)
4.

_A
C

0,
4.

Ci

-60

--120

180

240

300

360

X, degree

60

120

180

300

360

X, degree

Figure 5-20: Azimuthal scan of the pure and 8 wt% Laponite-filled polyurethane composite as the
deformation increases at d-spacing between 0.447 and 0.468 nm (20 = 15.090 to 15.830).

124

8 wt%: 0.409-0.427nm

0 wt%: 0.409-0.427nm

=.
(U
U)i

(I)
u,

(u

R
-q

60

120

180

240

300

360

60

120

180

240

300

360

X, degree

x, degree

as the
Figure 5-21: Azimuthal scan of the pure and 8 wt% Laponite-filled polyurethane composite
17.280).
to
deformation increases at d-spacing between 0.409 and 0.427 nm (20 = 16.560

The preferential association of Laponite to the hard micro-domains imposes spatial restrictions
and/or constraints on the hard segments. Upon deformation re-orientation of the hard microdomain/Laponite structures encourage local soft segment alignment. The hard microdomain/Laponite structures act like rigid fillers around which the soft chains must stretch. As
observed in the previous section, the consequence of this amplified orientation is increased
plastic strain and matrix damage.
5.3.7 Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy
From the dynamic mechanical analysis results detailed in Chapter Four, it is clear that the heat
distortion temperature of the Elasthane increases with increasing Laponite concentration.
However, it is not clear whether this behavior is a consequence of Laponite jamming or
enhancement of hard segment melting temperature. In general the melting temperature of a
material is increased if the intermolecular cohesive bonds holding the material together are
stronger [26]. Furthermore, the intensity of a hard segment melting endotherm increases when
the crystallite is more pure and/or perfect [26]. Both of these characteristics are exhibited in the
first heating cycle of DSC analysis when the Laponite concentration increases as shown in
Chapter Four. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR)
can be used to measure how the intensity of the free and hydrogen-bonded NH and C=O of the
125

Elasthane varies with increasing Laponite concentration [5, 6]. If as the clay concentration
increases, the ratio of the areas under the peaks at 1703 cm -' and 1730 cm ~', corresponding to the
hydrogen-bonded and non-bonded carbonyl group of the urethane, increases, it can be concluded
that the degree of hydrogen-bonding in Elasthane is increasing [5, 27]. By evaluating the ratio of
the intensities of the bonded and free peaks of the ATR-FTIR spectrum the origin of the
enhancement in heat distortion temperature can be determined and the preferential association of
the nano-clay with the hard micro-domains of Elasthane can be further confirmed [5].

The ATR-FTIR spectrum of the Elasthane/Laponite composites are shown in Figure 5-22. The
spectral signature of this polyurethane has been characterized previously by Christenson et al
and is displayed in Table 5-2 [28]. The peak at 1000 cm ~', absent in the pure polyurethane
spectrum, corresponds to the presence of stretched Si-O bonds in the Laponite platelets [29].
Consequently as the Laponite concentration increases the intensity of the peak at this wavelength
increases. The ratio of the areas under the peaks corresponding to the hydrogen and non-bonded
(or free) carbonyl bonds (HCO and FCO subscripts respectively) decreases initially upon the
addition of 6 wt% Laponite as shown in Table 5-3. As the Laponite concentration increases the
ratio approaches that of pure Elasthane. However, if the extent of the free or hydrogen-bonded
carbonyl bond absorbance to the total carbonyl bond absorbance is instead compared, as
explained in [30], the addition of Laponite does not appear to influence inter-urethane bonding.
This trend is the same as that when the -NH bond of the urethane linkages (3325 cm-': not
shown) is evaluated against the standard -CH stretching signature of the polyether (2860-2940
cm -': not shown) [27].
Wavenumber
1730
1703
1501
1536
1414
1310
1251 (shi
1223
1174"
1110
107~

ElasthaneTM 80A (PEU)


C==O urethane nonbonded
C-O urethane H-bonded
C==C aromatic ring
C-N+ N-H amide
C--C aromatic ring
C-N and N-H
C---0
C-N
Branched ether
C----C aliphatic ether
C----C of C---C=C'urethane

"Degradation products.

Table 5-2: ATR-FTIR spectral assignments for Elasthane determined by Christenson et al in 1281.

126

no

0.8
0.7

L0.6
0.5
I-

S0.4
0.3
0 0.2

<0.1
0.0

- _1

1800

1600

1000
1400
1200
Wavenumbers (cm 1 )

800

600

Figure 5-22: ATR-FTIR spectrum of pure Elasthane 80A and that filled with 6 wt%, 10 wt% and 15 wt%
Laponite.
Laponite

AHCO/AFHo

AHco/Aco

AFCO/ACO

0
4
10
15

2.72
2.66
2.67
2.69

0.73
0.73
0.73
0.73

0.27
0.27
0.27
0.27

Table 5-3: Ratio of area under the absorbance peaks of the hydrogen-bonded carbonyl bonds and the free
carbonyl bonds (AHco/AFco) as well as the ratio of these areas to the total absorbance area constituting
carbonyl bonds described by [30].

These results indicate that the addition of clay does not reduce the hard segment inter-urethane
hydrogen bonding and order, nor does it reduce the hydrogen bonding between the hard segment
and soft segment. Whether the Laponite platelets are physically linked to the matrix or
kinetically trapped, which can not be determined conclusively from these results, Laponite does
not influence microphase separation. Ultimately, the long processing time utilized via slow
solvent casting allows for self-assembly and complete microphase separation irregardless of
Laponite concentration. Furthermore, the preferential association of the nano-clay with the hard
micro-domains adds further order via spatial constraint.

127

5.3.8 Cross-Polarized Microscopy


Ordered, fractal-like domains are observed under cross-polarized light which are comprised of
polyurethane hard micro-domains, Laponite, and conformationally-confined soft domains. These
domains which measure approximately 20-100 ptm in size are disrupted upon yield and continue
to break up with increasing stretch as shown in Figure 5-23 and 5-24. Micrographs taken under
cross-polarized light of a pure Elasthane film and a nanocomposite film containing 10 wt%
Laponite elucidate this behavior. The bright ordered domains seen in the undeformed specimens
become dark and amorphous upon deformation. At yield, these ordered domains or substructures are stressed and stretched. Consequently, the bright domains intensify and begin to
break up, expanding in size as shown in Figure 5-23. It proved difficult to capture this point of
initial domain break-up in nanocomposites containing a percolated Laponite network because the
onset of yield begins almost immediately upon deformation. Upon greater tensile deformation
these bright domains continue to stretch parallel to the stretch direction and break-up further. As
a result, the bright, ordered domains become dark and amorphous. Complete disruption of the
ordered domains is observed prior to strain hardening, but after significant soft segment
orientation.

Figure 5-23: Comparison of the long range order exhibited through cross-polarized microscopy of pure
Elasthane prior to deformation (left) and after deformation to yield (image width is 2.14 mm). The bright
ordered domains increase intensity and begin to crack and separate upon yield.

128

Figure 5-24: Comparison of the long range order exhibited through cross-polarized microscopy of 10 wt%
Laponite-filled Elasthane prior to deformation (left) and after significant stretch and soft-segment reorientation and alignment. The bright ordered domains become dark and amorphous indicating that these
sub-structures have been completely disrupted at large stretch.

Ultimately, upon deformation these ordered domains are irreparably damaged--after inelastic
stretch and mechanical unloading, the original ordered domains remain disrupted and
unrecoverable. These cross-polarized images provide visual evidence that indeed the matrix does
undergo irreparable damage upon deformation, especially in the localized areas where Laponite
and hard micro-domains constrain the soft segment.

5.4 Conclusions
Thermoplastic polyurethane filled with nano-clay demonstrates strong hysteresis, time
dependence, and cyclic softening. The rate sensitivity of the polyurethane increases with
Laponite concentration as does the extent of softening between the first and fifth cycle of
loading. Similarly, the addition of Laponite increases the residual or plastic strain, particularly
when concentrations exceed the percolation limit. The pure and Laponite filled polyurethanes
exhibit two characteristic relaxation times. The long relaxation time corresponds to the hard
micro-domains and the short relaxation time corresponds to the soft segment. As the magnitude
of the applied stretch increases the soft segments stretch, align and orient while the hard
segments rotate but do not break-up. Consequently, the long relaxation time exhibits scatter and
remains constant while the short relaxation time increases with stretch. The addition of nano-clay
too increases the sensitivity of the material to orientation induced softening. The extent of
modulus decay with orientation is dependent upon nano-filler concentration, aspect ratio and
polydispersity. Ultimately, nano-clay reinforces the stress response of Elasthane, but also locally

129

damages the polyurethane matrix, especially when concentrations above the percolation
threshold are utilized. Laponite does not influence the hyperelastic or equilibrium response of the
material-that is matrix dependent. Instead, the nano-clay influences the non-equilibrium
response-the rate sensitivity and time dependence-much like rigid particles influence the
viscosity of dilute solutions.
ATR-FTIR results indicate that the addition of clay does not reduce the hard segment interurethane hydrogen bonding and order, nor does it reduce the hydrogen bonding between the hard
segment and soft segment. Laponite does not influence microphase separation. This confirms
that the equilibrium matrix polyurethane response is unaffected by the addition of nano-clay.
Ultimately, the long processing time utilized via slow solvent casting allows for self-assembly
and complete microphase separation irregardless of Laponite concentration. The preferential
association of Laponite to the hard micro-domains imposes spatial restrictions and/or constraints
on the hard segments. Wide angle x-ray scattering indicates that upon deformation the hard
micro-domain/Laponite structures re-orient and encourage local soft segment alignment. The
hard micro-domain/Laponite structures act like rigid fillers around which the soft chains must
stretch. As is observed via mechanical characterization, the consequence of this amplified
orientation is increased plastic strain and matrix damage. Cross-polarized images provide visual
evidence that indeed the matrix does undergo irreparable damage upon deformation, especially
in the localized areas where Laponite and hard micro-domains conformationally constrain the
soft segment initially. Ultimately, understanding the morphological and mechanical influence of
nano-clay on the elastomeric behavior is important for the continued advancement of engineering
thermoplastics.

5.5 References
[1]

[2]

[3]

L. J. Korley, B. D. Pate, E. L. Thomas, and P. T. Hammond, "Effect of the degree of soft


and hard segment ordering on the morphology and mechanical behavior of
semicrystalline segmented polyurethanes," Polymer, vol. 47, pp. 3073-3082, 2006.
D. J. Martin, G. F. Meijs, P. A. Gunatillake, S. J. McCarthy, and G. M. Renwick, "The
effect of average soft segment length on morphology and properties of a series of
polyurethane elastomers.2. SAXS-DSC annealing study," Journalof Applied Polymer
Science, vol. 64, pp. 803-817, 1997.
J. T. Koberstein, A. F. Galambos, and L. M. Leung, "Compression-molded polyurethane
block copolymers. 1. Microdomain morphology and thermomechanical properties,"

130

[4]

[5]

[6]

[7]

[8]

[9]
[10]

[11]

[12]

[13]
[14]
[15]
[16]

[17]

[18]
[19]

Macromolecules, vol. 25, pp. 6195-6204, 1992.


M. Song and K. J. Yao, "X-ray diffraction detection of compliance in polyurethaneorganoclay nanocomposites," MaterialsScience And Technology, vol. 20, pp. 989-992,
2004.
X. H. Dai, J. Xu, X. L. Guo, Y. L. Lu, D. Y. Shen, N. Zhao, X. D. Luo, and X. L. Zhang,
"Study on structure and orientation action of polyurethane nanocomposites,"
Macromolecules, vol. 37, pp. 5615-5623, 2004.
Y. I. Tien and K. H. Wei, "Hydrogen bonding and mechanical properties in segmented
montmorillonite/polyurethane nanocomposites of different hard segment ratios,"
Polymer, vol. 42, pp. 3213-3221, 2001.
A. Aneja and G. L. Wilkes, "A systematic series of'model' PTMO based segmented
polyurethanes reinvestigated using atomic force microscopy," Polymer, vol. 44, pp. 72217228, 2003.
B. Finnigan, D. Martin, P. Halley, R. Truss, and K. Campbell, "Morphology and
properties of thermoplastic polyurethane nanocomposites incorporating hydrophilic
layered silicates," Polymer, vol. 45, pp. 2249-2260, 2004.
H. J. Qi and M. C. Boyce, "Stress-strain behavior of thermoplastic polyurethanes,"
Mechanics of Materials,vol. 37, pp. 817-839, 2005.
L. J. Korley, PEO-containingCopolymers as PolyurethaneSoft Segments in the
Development of High PerformanceMaterials,Ph.D. in Program in Polymer Science and
Technology, Cambridge, MA:Massachusetts Institute of Technology, 2005.
J. Diani, M. Brieu, and J. M. Vacherand, "A damage directional constitutive model for
Mullins effect with permanent set and induced anisotropy," EuropeanJournalof
Mechanics A: Solids, vol. 25, pp. 483-496, 2006.
H. J. Qi and M. C. Boyce, "Constitutive model for stretch-induced softening of the stressstretch behavior of elastomeric materials," Journalof the Mechanics and Physics of
Solids, vol. 52, pp. 2187-2205, 2004.
J. D. Ferry, Viscoelastic PropertiesofPolymers, 3rd ed. New York, NY: John Wiley &
Sons, Inc., 1980.
A. Einstein, "Investigations on the Theory of Brownian Movement," Annalen der Physik,
vol. 4, pp. 371-381, 1906.
J. Yi, M. C. Boyce, G. F. Lee, and E. Balizer, "Large deformation rate-dependent stressstrain behavior of polyurea and polyurethanes," Polymer, vol. 47, pp. 319-329, 2006.
B. Finnigan, P. Casey, D. Cookson, P. Halley, K. Jack, R. Truss, and D. Martin, "Impact
of controlled particle size nanofillers on the mechanical properties of segmented
polyurethane nanocomposites," InternationalJournalOfNanotechnology, vol. 4, pp.
496-515, 2007.
R. B. Dupaix and M. C. Boyce, "Constitutive modeling of the finite strain behavior of
amorphous polymers in and above the glass transition," Mechanics of Materials,vol. 39,
pp. 39-52, 2007.
E. M. Arruda and M. C. Boyce, "A three-dimensional constitutive model for the large
stretch behavior of rubber elastic materials," Journalof the Mechanics andPhysics of
Solids, vol. 41, pp. 389-412, 1993.
J. S. BergstrOm and M. C. Boyce, "Deformation of elastomeric networks: relation
between molecular level deformation and classical statistical mechanics models for
rubber elasticity," Macromolecules, vol. 34, pp. 614-626, 2001.

131

[20]

[21]

[22]
[23]
[24]

[25]

[26]
[27]

[28]

[29]

[30]

M. Niesten and R. J. Gaymans, "Tensile and elastic properties of segmented


copolyetheresteramides with uniform aramid units," Polymer, vol. 42, pp. 6199-6207,
2001.
M. Niesten, J. W. ten Brinke, and R. J. Gaymans, "Segmented copolyetheresteraramids
with extended poly(tetramethyleneoxide) segments," Polymer, vol. 42, pp. 1461-1469,
2001.
J. P. Flory, Principlesof Polymer Chemistry. Ithaca, NY: Cornell University Press, 1953.
E. Guth, "Theory of Filler Reinforcement," Journalof Applied Physics, vol. 16, pp. 2025, 1945.
D. J. Martin, G. F. Meijs, P. A. Gunatillake, S. P. Yozghatlian, and G. M. Renwick, "The
influence of composition ratio on the morphology of biomedical polyurethanes," Journal
ofApplied Polymer Science, vol. 71, pp. 937-952, 1999.
B. Finnigan, D. Martin, P. Halley, R. Truss, and K. Campbell, "Morphology and
properties of thermoplastic polyurethane nanocomposites incorporating hydrophilic
layered silicates," Polymer, vol. 45, pp. 2249-2260, 2004.
R. J. Young and P. A. Lovell, Introduction to Polymers. Cheltenham, U.K.: Nelson
Thornes Ltd., 1991.
C. H. Dan, Y. D. Kim, M. H. Lee, B. H. Min, and J. H. Kim, "Effect of solvent on the
properties of thermoplastic polyurethane/clay nanocomposites prepared by solution
mixing," JournalOfApplied Polymer Science, vol. 108, pp. 2128-2138, 2008.
E. M. Christenson, M. Dadsetan, M. Wiggins, J. M. Anderson, and A. Hiltner,
"Poly(carbonate urethane) and poly(ether urethane) biodegradation: In vivo studies,"
Journalof Biomedical MaterialsResearch, vol. 69A, pp. 407-416, 2004.
J. A. Caulfield, T. A. Wells, and K. E. Miller, "Novel method for transmission infrared
analysis of clay minerals using silicon wafer substrates," Clays and Clay Minerals,vol.
55, pp. 213-219, 2007.
I. Yilgor, E. Yilgor, I. G. Guler, T. C. Ward, and G. L. Wilkes, "FTIR investigation of the
influence of diisocyanate symmetry on the morphology development in model segmented
polyurethanes," Polymer, vol. 47, pp. 4105-4114, 2006.

132

6. Influence of Laponite on the Shape Memory Behavior of


Polyurethane
6.1 Introduction
Materials that exhibit shape change, whether metal alloy, ceramic, polymer or gel, have
generated great interest in the past decade as actuators in the aerospace, textile and biomedical
industries. For instance shape memory polymer foams have attracted attention as space
deployable support structures, shelters for space habitation, and rover components, as well as
embolic sponges due to the low density, increased surface area, and compressibility of the
foam [1]. Should actuating polymers be able to respond as quickly and with as much strength as
natural skeletal muscle, numerous devices that were previously unattainable could be realizedsuch as artificial muscle, haptic devices, electronic Braille screens, rehabilitation gloves, tremor
suppressors, and micro-wings [2, 3]. Shape memory is the ability of a material to maintain a
deformed shape and upon external stimulus recover the original, pre-deformed shape. The
actuating external stimulus is most often thermal inducing entropic recovery, but for polymeric
materials and gels light may also be employed. Ionic strength, pH, and solvent quality may also
be used to induce shape transitions in gels [4]. Shape memory polymers which include
amorphous polymers as well as chemically or physically cross-linked multi-block copolymers
with either a melting transition or glass transition above the storage temperature (often ambient
temperature, Tamb or 370 C in the human body) are attractive actuated materials because they
exhibit moderate strength and significant strain recovery. Actuation stresses span the range from
kilopascals to megapascals and recovery strains extend from 10% to 100% depending on the
polymer employed.
Shape memory capability in polymers is attributed to both its chemical components and
morphology. Linear multi-block polymers are often utilized in shape memory applications
because they allow the actuation temperature, or transition temperature (Tans) to be tuned via
synthetic methods. These block copolymers contain at least two distinct, immiscible
morphological phases. One phase is characterized by thermal transitions above the programming
temperature of the shape memory polymer (SMP) and acts as a permanent physical link to the
133

other phase, maintaining memory of the original shape prior to deformation. The second phase is
characterized by either a melting or glass transition temperature (Trans) above the storage
temperature. Typical operation of a SMP includes production of the material in some specified
and desired permanent configuration, e.g. as a cylindrical fiber of length lo and radius ro. Then
the SMP is programmed or deformed to meet specifications of the desired temporary shape, e.g.
a cylindrical fiber of longer length I ( 1 > lo ) and radius r ( r < ro). Often the deformation

temperature is set above the inherent transition temperature (Trans=Tg or Tm) of the polymer-the
temperature which triggers entropic recovery-so that less work is needed to deform the
material. If deformed at an elevated temperature, the polymer is then cooled to the storage
temperature, vitrifying or crystallizing the second phase and holding the deformed, temporary
shape. This temporary shape is maintained while the temperature is held below Trans; however,
upon an increase in temperature to a value greater than or equal to Ttrans the mobile chains of the
second phase entropically return to their previous configuration and the permanent shape is
recovered-the fiber returns to its original length lo and radius ro. In order to obtain the
temporary shape specified a second time, or a different temporary shape, the programming cycle
must be repeated and more work applied to deform the material.

Polyurethanes are attractive shape memory polymers because they are well-studied linear multiblock copolymers consisting of soft segments with melting transition greater than ambient
temperature (Tm, ss> Tamb) or hard segments with glass transition temperature greater than
ambient temperature (Tg, Hs > Tamb) that are either covalently cross-linked to the soft segment in
thermosets or linked to the soft segment through hydrogen or ionic bonding in thermoplastics.
This linkage between the soft and hard blocks of the polyurethane maintains the permanent shape
unless the programming temperature approaches the hard segment melting temperature (Tm,

HS)

or the programmed deformation is to such an extent that the cross-links are permanently
damaged. Versatile polyurethane chemistry allows the soft segment glass transition and melting
temperature to be tailored for specific mechanical responses and surface character such as
biocompatibility, hydrophilicity, or blood compatibility [5]. For instance, polyurethanes
containing poly(c-caprolactone)

diol (PCL) soft segments are desirable for biomedical

applications, i.e. as sutures or in drug delivery applications due to its biocompatibility and its

134

melting transition temperature (Tm,PCL = 44-55 0 C depending on molecular weight) that is just
greater than the temperature of the human body, Tbody=37'C [4].
The capability of nano-clay to significantly enhance the thermomechanical properties of
polyurethane is discussed in Chapter 4. The fact that the addition of Laponite increases the
flexural storage modulus of Elasthane above and below its glass transition indicates that it may
be possible to reinforce shape memory polyurethane with nano-clay and enhance the recovery
stress when the material is subjected to a fixed strain and actuated. However, it has also been
shown that Laponite forms a percolated network at concentrations above 2.5 vol% or 6 wt% and,
upon deformation this network is disrupted, as is the polyurethane network which undergoes
irreparable damage. Likely, there will be an optimum loading in which the free strain recovery
and stress recovery at constrained strain are maximized.
Rezanejad and Kokabi showed in [6] that modified-Montmorillonite nano-clay (Closite 15A)
enhanced the recovery stress of cross-linked, low-density polyethylene (LDPE) by -200%.
However, as a consequence the added nano-filler hinders chain motion and disrupts chain
recovery so that in the 8 wt% filled LDPE nanocomposite there is a 12% decrease in recoverable
strain. Similarly, Cao and Jana evaluated the influence of 1, 3 and 5 wt% of modifiedMontmorillonite (Cloisite 30B), on the shape memory behavior of a custom synthesized
polyurethane,

containing

polycaprolactone

[7].

The

melting transition

of crystalline

polycaprolactone was utilized as the shape memory transition trigger. At a 1 wt% loading, the
clay is exfoliated, but at higher concentrations the clay is intercalated. At the 1 wt% loading, the
increase in recovery stress (when strain is fixed) is 20% greater than that of the pure
polyurethane. Cao and Jana believe this increase in recovery stress is greater than that exhibited
by the other composites because the larger clay loadings further disrupt the hard micro-domains
and soft segment crystallinity. This disruption in turn increases the rate of stress relaxation the
resultant composite undergoes, minimizing the achievable recovery stress. In this study we
utilized commercial, biocompatible, thermoplastic polyurethane, MM 5510 from the DiAPLEX
Series provided by Mitsubishi Heavy Industries, Ltd. and Laponite to evaluate the influence of
nano-clay on the shape memory properties of polyurethane.

135

6.2 Materials and Experimental Methods


6.2.1 Materials
The shape memory polyurethane MM 5510 (Mitsubishi Heavy Industries, Ltd.) was supplied in
pellet form and used as received. This polyurethane is composed of diphenylmethane-4,4diisocynate, adipic acid, ethylene oxide, polypropylene oxide, 1,4-butanediol, and bisphenol A
[8]. The glass transition temperature, Ttrans, of the soft segment is specified by the provider as
55 0 C. Dimethyl formamide (DMF) from Sigma Aldrich was used as received to dissolve the
polyurethane pellets. Nano-clay platelets, Laponite RD (Southern Clay Products, Inc.), with
diameter of -25 nm and thickness of 1 nm were used as received in powder form.
6.2.2 Composite Preparation
Polyurethane-Laponite-DMF solutions were prepared via the solvent exchange method discussed
in Chapter 3 so that the final concentration of polyurethane in solution was 1 wt% and the final
Laponite concentration in polyurethane was 0, 4, 10 and 15 wt%. The solutions were then rollmixed at a moderate speed for two days prior to use to ensure that the polyurethane was
completely dissolved. MM5510/Laponite nanocomposite films with 120-180 pm thicknesses
were prepared in Teflon containers (6 cm x 4 cm) by controlled evaporation of DMF. Solvent
casting took place immediately after the solutions were sonicated for 60 minutes to ensure
complete dispersal of the Laponite in solution. The evaporation rate of DMF was maintained by
keeping the films in a closed 50'C oven with a 0.02 m 3 h- 1 N2 purge flow. The resultant films
were characterized using the following procedures.
6.2.3 Characterization Methods
6.2.3.1 Laponite Dispersal

The 4 wt%, 10 wt% and 15 wt% Laponite nanocomposite TEM samples were prepared in a
JEOL JEM 9310 focused ion beam (FIB) instrument. These samples were first sputter-coated
with -200nm of gold, and then a localized -1 pm-thick carbon protective film was deposited atop
the selected area for lamella preparation. The samples were milled and polished in the FIB to
give lamellae measuring 10 pm x 10pm x 80 nm. These samples were transferred to the copper
grids using a micromanipulation system that consisted of a position-controlled polished glass rod
that allowed for electrostatic pick-up of the lamellae. These lamellae were observed in a JEOL
200CX TEM operated at 200 kV.

136

6.2.3.2 Thermal Transitions


To determine the influence of Laponite on the glass transition and melting transitions of the
polyurethane differential scanning calorimetry measurements were carried out on a TA
Instruments Q1000 DSC at a heating rate of 100 C min - ' between -90' and 250'C, using
aluminum pans. Dynamic mechanical measurements were also made at a frequency of 1 Hz and
at a heating rate of 3C min-' from -150'C to 2000 C using the TA Instruments Q800 DMA.
6.2.3.3 Thermal Expansion Coefficient
The thermal expansion coefficient of the nanocomposites was determined using TA Instruments
0
Q800 DMA. A controlled force test was run from -150 0 C to 150'C at 2 C min ' and a tensile
force was applied so that the nominal stress in the sample would be 10 kPa based on the initial

width and thickness of the rectangular specimen. The length displacement data was used to
calculate the linear axial coefficient of thermal expansion,
a = 1 dl
dl
10 dT'

(6.1)

where 1/10 is the axial thermal stretch and 10 the original length prior to testing. Each test was repeated
on a minimum of two, unique samples and only at temperatures above 80'C were there
discrepancies in behavior. This can be attributed to instrument difficulty in holding the 10 kPa
load while the sample was undergoing significant thermal softening.
6.2.3.4 Stress-StrainBehavior Above and Below the Transition Temperature
Various tensile tests were completed on each of the nanocomposites using the EnduraTEC ELF
3200 with heat chamber. One test procedure required that each specimen be stretched 150% and
unloaded to its initial length five times. Each nanocomposite was tested at room temperature
(-25'C) and at 700 C, a temperature above the glass transition or trigger temperature of the
polyurethane. A second procedure required that each specimen be stretched to 1.2, 1.4, 1.7, 2.0
and 2.5 times its original length and unloaded five times at each stretch before proceeding to the
larger deformation. The second procedure was also completed at room temperature and at the
elevated temperature. The third procedure required that each nanocomposite be stretched at an
elevated temperature to 1.4, 1.7, or 2.5 times its original length and unloaded five times. Upon
completion each specimen was cooled to room temperature (approximately 10 minutes of forced
air flow was needed to cool the sample) and then stretched to a value equal to or greater than the

137

previous stretch and again unloaded. For all tests the gauge length of the specimens was 8 mm
and the crosshead speed was 0.13 mm s 1.

During testing, using the ELF 3200, the thin-film sample was lit with a fiber optic light and
videoed using a Qimaging Retiga 1300 CCD camera equipped with a 200 mm Canon f14.0
Canon lens and a Nikon 200 mm extension. Dots using a fine tip black permanent marker were
placed in alternating pattern along the length of the sample as shown below in Figure 6-1 and the
specimen width and thickness at each location were measured with a micrometer. The camera
was placed at a distance of approximately 1 m from the specimen and images captured every
0.5 s. The images were analyzed with a digital image correlation algorithm developed by
Correlated Solutions Incorporated (CSI). The deformed image was correlated to the reference
image via two distinct points of interest, i.e. two black dots, connected by a straight line. The
software then output the x and y pixel position of these two distinct points. The change in
longitudinal distance was used to calculate the true strain while the change in horizontal distance
was used to calculate the true stress as shown below:
In

true

(6.2)
(6.2)

2 - Y1

, = ]nrY2,0 - Y1,0

and
tue

wo

(wt)

(woto)

(woto)

x2,0

x2 -

1,(6.3)
X1

when changes in thickness are assumed negligible, t

t o . This video extensometer was used

because strain localization was common within these samples since the thickness of the
specimen varied 20 tm through the length due to processing.

It is important to note that the grips used in ELF 3200 were not best suited for this material;
however, time did not allow for procurement of another pair. The grips were screw tightened and
the jaw faces were serrated. If the films were placed in these grips by themselves the serrated
faces cut and damaged the sample, causing the data that was collected on that particular sample

138

to be invalid. So instead a soft, slightly tacky elastomer was placed on either side of the
specimen as a means to prevent specimen damage.

X2.(k

X1.f)

Figure 6-1: Images of a pure MM5510 polyurethane film inside the ELF 3200 as seen with the video
extensometer. On the left is the un-deformed film with an 8 mm gauge length. On the right is the same film
stretched to the instruments upper limit; the upper grip moves 12 mm upward, however, there are localized
areas of deformation so the true strain in the sample does not reach 0.92 as is evidenced by the highlighted
markers.

While the inclusion of this elastomer prevented puncture and tear of the specimen, it encouraged
slippage at the grip locale or localized deformation there due to over-tightening of the grip
screws. Often, the contact line or surface where the grip held the specimen was not even with the
grip ends and could not be visualized from start to finish, as is the case in Figure 6-1. The extent
of deformation that takes place within these regions cannot be accounted for. These problems
manifested in the later tests which required thermal cycling. Consequently, when the instrument
was programmed to stretch the specimen a certain amount, often the local deformations
visualized from start to end within the specimen were not realized. Instead there was some slip or
localized deformation outside the initial gauge length of the specimen and a lesser amount of
strain was accounted for using the video extensometer. Therefore, throughout this chapter the
specimen stretch programmed is reported as Aspecimen and the local, true strain measured using the
video extensometer is reported as true strain or A,,e.

139

6.2.3.5 Free Strain Recovery


The free strain recovered in these nanocomposites under virtually no load was examined two
ways. The first method utilized the ELF 3200 with heat chamber. Each specimen was stretched
to 1.4, 1.7, or 2.5 times it original length using displacement control at room temperature. Upon
reaching the desired stretch the instrument feedback was changed to force control and the load
was manually reduced over approximately five minutes time to virtually no load without
buckling the specimen (in actuality a small positive tensile force in the specimen exists). Then
while this load was maintained the temperature in the oven was increased from room
temperature, -25 0 C to 70'C over approximately 30 seconds and then held constant until the
specimen stopped recovering strain and began lengthening. The second method utilized the TA
Instruments Q800 DMA and did not allow for visualization of the specimen deformation. A
custom program was made which equilibrated the temperature in the oven at 70C for one
minute and then the nominal strain was ramped at 0.0167 s-1 to 10%, 20%, 40%, 70% or 100%
nominal or engineering strain. After holding this length for one minute the temperature was
ramped to 25C at -10'C min - . The force was then reduced to and held at 0.001 N and the
temperature ramped back up to 70C at 100 C min - and the strain recovery monitored. This
procedure was repeated five times on each sample. This allowed the extent of strain recovery,
total strain recovery and strain fixity defined below by Landlein and Kelch in [4] to be
determined.

The extent of strain recovery, R, (N), compares the strain recovered from one cycle to the
previous cycle:
Rm -P(N)

R, (N)

(N

(6.4)

cm - Eg (N -1)'

where em is the maximum nominal strain applied, e, is the nominal irrecoverable strain during the
specified cycle denoted by the integer N.

The extent of total strain recovery, Rr, tota (N), compares the strain recovered during one cycle to

the deformed state:

140

r, total

(N)

(6.5)

(N)

where Em is the maximum nominal strain applied, ep is the nominal irrecoverable strain during
cycle number N.

The extent of strain fixity, Rf (N), quantifies the ability of the material to hold its temporary
shape after consecutive deformation cycles:
R (N) Eu

(6.6)

(N)
Cm

where em is the maximum nominal strain applied, eu is the nominal strain realized under the no
load condition prior to actuation and recovery during cycle number N. The positions of em, cu, ep
are described visually in the three dimensional deformation plot shown in Figure 6-2.

10
9
8
C

4.

Figure 6-2: Visual description of the free strain recovery test and the locations of c,, ,u, ,, with respect to
engineering stress, strain and temperature. The first step demarks deformation at an elevated temperature
above Tans. The second step demarks cooling to a temperature below Tran-s. The third step demarks unloading
to a no load condition. The fourth and final step demarks entropic recovery during the temperature ramp to
a temperature above Ttas.

141

6.2.3.6 ConstrainedStrain, Stress Recovery


The constrained strain, stress recovery test also utilized the ELF 3200 with heat chamber. Each
specimen is stretched to 1.4, 1.7, or 2.5 times its original length at room temperature and held at
that stretch and allowed to relax for five minutes. Then, the length of the specimen is manually
reduced until there is virtually no load without buckling the sample (in actuality a small positive
tensile force in the specimen exists). This manual reduction in length is a slow process that takes
approximately five minutes in order to allow for the time-dependent stress-recovery (opposite to
stress relaxation) at room temperature to plateau. After the load is removed, the displacement is
held constant and the temperature is increased from 25C to 70 0 C over approximately thirty
seconds and then held constant for ten minutes. Meanwhile, the load, which increases when the
specimen attempts to shrink because the specimen is restricted, is monitored. The engineering
stress is reported.

6.3 Results and Discussion


6.3.1 Laponite Dispersal
The Laponite platelets, which measure approximately 25 nm in diameter and 1 nm in thickness,
are well dispersed within the polyurethane matrix even at concentrations above the percolation
threshold (bp z 2.5 vol% or 6 wt%). Transmission electron micrographs of the nanocomposites
containing 4, 10 and 15 wt% (shown in Figure 6-3) nano-clay further elucidate the efficacy of
the solvent exchange technique to fully exfoliate the clay particles. In the 4 wt% sample the
platelets are randomly oriented and separated by more than one particle diameter. At 10 wt%
nano-clay necklaces span the sample, exemplifying a percolated network. It proved difficult to
solvent cast the 10 wt% sample in particular because the percolated clay network encouraged
crack growth upon solvent removal and often rather than a smooth continuous film, numerous
film pieces resulted. In contrast, the 15 wt% sample starts to exhibit an aligned network of clay
platelets due to volume restrictions. Just as it is impossible to re-fill a matchbox with matches
once they have been removed without aligning them, the clay platelets must align and orient
within the polyurethane matrix at this high concentration. This effect, coupled with the slow
solvent removal and the presence of polar constituents such as ethylene oxide, promotes
intercalation and flocculation as is observed in the lower left corner of the 10 wt% micrograph
and the upper right corner of the 15 wt% micrograph. This clay dispersal is similar to that seen in
the custom synthesized polyurethane containing poly(ethylene oxide) and 10 wt% Laponite

142

discussed in Chapter 3. These micrographs visually suggest that the Laponite is not only
reinforcing the hard micro-domains, but embedded within the soft domain as well.

Figure 6-3: Transmission electron micrographs of MM5510 PU nanocomposites containing 4, 10, and 15 wt%
Laponite (a, b and c, respectively). Note the Laponite platelets in the 15 wt% specimen are more aligned and
there are instances of intercalation and flocculation in the specimens containing Laponite concentrations
greater than the percolation threshold.

6.3.2 Thermal Transitions


Differential scanning calorimetry was used to verify that the transition temperature of the
polyurethane was in fact 55C as the manufacture reported. However, instead the complex nature
of this multi-block polyurethane was revealed after solvent casting. Three glass transitions were

143

0
0
observed upon the first heating cycle between 0 C and 100 C in addition to the hard segment
melting endotherm near 200'C as shown in Figure 6-4.

1st Cycle

"'

W.p

'

S
MM5510

2nd Cycle

" '

'

MM5510
wf/o
wt%
-10
wt%
-15
-

-4

0.4

0.4

0.3

0.3

S 0.2

0.2

o0.1

0.1
U.

LL

-r

So.o0

'~

0.0

-0.1

-0.1

-0.2

-0.2

-0.3

-0.3
-Exo

Down

Exo Down ,

-0.4

-100

i
I

-0.

100

Exo Down
I

-100

200

100

200

Temperature rC]

Temperature [C]

filled with 4, 10,


Figure 6-4: First and second heat flow cycles of pure MM5510 polyurethane and composites
calorimetry.
and 15 wt% Laponite using differential scanning

Upon cooling and then reheating one soft segment glass transition dominates at -35C and then
-45 0C, respectively. Aware of the results of Yang et al, who found that similar Mitsubishi shape
memory polyurethane was plasticized upon soaking in a water, great care was taken to avoid
water sorption [9]. All specimens were kept in a polycarbonate case and then within a desiccator
when not being tested or zip-locked within two bags between tests. These results, in conjunction
0
with dynamic mechanical analysis, suggest that the slow solvent removal and annealing at 50 C
is responsible for the presence of the multiple glass transition temperatures at approximately
20 0 C, 500 C, and 75C upon initial heating rather than water sorption.
Like in the Elasthane-Laponite composites upon initial heating, the hard segment melting
temperature of the nanocomposite increases as does the intensity of the melting endotherm as the
concentration of nano-clay increases. But upon cooling, a corresponding exotherm is
unnoticeable. Upon reheating, the endotherm is broadened due to incomplete and imperfect recrystallization due to the fast 100 C min ' temperature ramp. However, it is not clear whether this
behavior is a consequence of Laponite jamming and constrained motion or a change in

144

morphology and hard micro-domain texture. In general the melting temperature of a material is
increased if the intermolecular cohesive bonds holding the material together are stronger [10].
Furthermore, the intensity of a hard segment melting endotherm increases when the crystallite is
more pure and/or perfect [10].
The thermal transitions observed with dynamic mechanical analysis, while within the same
temperature range as the differential scanning calorimetry results, only illuminate one glass
transition in freshly prepared samples and two transitions in aged samples. After numerous tests
on thin-film nanocomposite samples prepared following the same procedure but at different dates
there were discrepancies in the glass transition behavior. In some instances, particularly when
samples were tested soon after preparation, only one soft segment glass transition was observed
between 44'C and 540 C, depending on Laponite concentration. When numerous days passed
these same nanocomposite samples exhibited different transition behavior, often exhibiting two
transitions separated by 100C in the pure sample and 30-40 0 C in the highly-filled samples
although the nanocomposites were carefully stored in a dessicator and sealed from laboratory
humidity (which ranged from 17% to 42% relative humidity). Despite these discrepancies, the
magnitude of the flexural storage and loss moduli were consistent from one sample to the next.
The addition of nano-clay enhances the flexural storage modulus most significantly at
temperatures above the transition temperature. While water sorption has been shown by Yang et
al to decrease the glass transition temperature in similar shape memory polyurethanes, only one
transition was observed, not two or three transitions as was observed in these nanocomposites [9,
11]. Due to the care taken to store these samples in a desiccator, it is believed the multiple
transitions are a result of processing and aging in the pure sample as well as due to the addition
of nano-clay to the composites.
The glass transition temperature can be defined as either the temperature at which the flexural
loss modulus reaches a local maximum or the temperature at which the ratio of the flexural
storage modulus to the flexural loss modulus (tan 6) reaches a local maximum. The value
determined via the flexural loss modulus is typically less than that determined via tan 8. In this
case there was as much as an 180 C discrepancy as is shown in Table 6-1. Consequently, while
the tan 8-based soft segment glass transition is observed at 45-55'C, the peak in E" is closer to

145

30'C (as seen in Figure 6-5). This indicaties that these polyurethanes could in fact be used in
biomedical applications since the body temperature is of similar magnitude (37oC). The addition
of nano-clay too decreases the soft segment glass transition; however, the influence is not
systematic. Likely, preferential reinforcement of the more polar soft segment creates a second
sub-structure or morphology that develops further while the sample is stored at room
temperature.
----10000

1000
2 100
10
10
-1 00

-50

100
50
0
Temperature (OC)

200

150

-100

-50

100
50
Temperature (OC)

150

200

-50

100
50
0
Temperature (OC)

150

200

-50

150

200

1000

100

100

10(

010
LU1

n1
-100

0.1
Temperature (C)
0.8

0.8

0.6

0.6

0.4

r 0.4

0.2

0.2
nn
v.v
-100

0n
,.
-100

-50

100
50
0
Temperature (OC)

200

150

100
50
Temperature (OC)

Figure 6-5: Dynamic mechanical response of pure MM5510 polyurethane and composites filled with 4, 10,
and 15 wt% Laponite soon after preparation (left) and nine months later (right).
Glass Transition(s)

Aag)

DMA (new)
analysis
mechanical
and differential

M.5510 PU
4wt%
wt%
10
15 wt%

W52-C
29.30C
31,4"C
32.iC

scanning calorimetry

533T
44.0"C
45,44'13.1-C,

5a14.5-C,
4
1

DSC
upon heating

Cprepared
on frest

13.3-C
51.9-C,
C ,
22.0- 52.4 74.6-C
18.6-C, 54.7-C,
73 -C
51.9-C, 77.5-C
C,

andCaged

42.9'C
400C
38.7-C
45 9 C

Table 6-1: Analysis of the glass transition temperatures observed between 0C and 1001C via dynamic
mechanical analysis and differential scanning calorimetry upon heating on freshly prepared and aged
samples.

146

6.3.3 Thermal Expansion Coefficient


The addition of nano-clay to these polyurethane films not only alters the transition temperature,

Temperature (oC)
IV

Temperature (OC)
.

0 wt%
-4 wt%/
----- 10 wt
15wt%
1.008
-

1.006
6.14e6.21 e-5
1.004

4.76e-5
3.970-5

1.002

I.UUU

I
I

Temperature (oC)
Temperature (OC)
Figure 6-6: a) Thermal expansion of pure MM5510 polyurethane films and composites filled with 4, 10,
and
15 wt% Laponite from -150 0 C to 125 0 C while under a 10 kPa tensile load. b) Enlarged view of the thermal
expansion behavior within the shape memory operational temperature range interrogated. c & d)
The
thermal expansion coefficients below and above the transition temperature (55 0 C), respectively,
are
elucidated.

but decreases the in-plane thermal expansion coefficient at temperatures above and below the
transition temperature, as seen in Figure 6-6. The transition temperature is demarked by the

147

change in slope of the axial thermal stretch of the specimens at temperatures between 35C and
45 0 C. The influence of nano-clay observed via the axial thermal stretch behavior is opposite to
that seen with DSC and DMA. The addition of 15 wt% nano-clay does not appear to
significantly change the transition temperature, but the addition of 4 wt% and 10 wt%
0
0
systematically increases the transition temperature from 33 C to 350 and 42 C, respectively. The
expansion behavior of the sample filled with 4 wt% Laponite, a concentration less than the

percolation threshold, is similar to that of the pure polyurethane. At concentrations above the
percolation threshold the influence of the nano-clay is more dramatic. The addition of 10 wt%
-1
5
and 15 wt% nano-clay decreases the coefficient of thermal expansion from 6.2x10 oC to
4.8x10-5 C-1 and 4.0x10 -5 oC -1, respectively, at temperatures below the transition temperature. At

temperatures above the transition temperature, despite the Laponite concentration, the addition of
-3
-1
-4
nano-clay reduces the coefficient of thermal expansion 28% to 8.0x10 C from 1.1x10 0C-1
It is important to note that the influence of Laponite is most significant in the in-plane directions,
through the width and length, rather than in the transverse direction through the thickness [12].
0
0
Based on these results, it is clear that within the operating range of 25 C and 70 C thermal
expansion accounts for only 4% or less of the specimen lengthening and recovery.
6.3.4 Stress-Strain Behavior Above and Below the Transition Temperature
The preferential insertion of nano-clay within the soft domains of this shape memory
polyurethane is elucidated in the hysteretic tensile results above and below the transition
temperature. As the concentration of nano-clay increases, so does the modulus of the
polyurethane composite. The 15 wt% composite, due to the incomplete exfoliation and tendency
for intercalation and flocculation of the nano-clay, exhibits an increase in modulus when
compared to the pure polyurethane and a decrease when compared to the 10 wt% composite. In
the results both above and below the transition temperature (Figure 6-7), it is obvious that the
addition of nano-clay suppresses strain-hardening, further indicating that the nano-clay is
preferentially embedded within the soft domain of the polyurethane. Strain-hardening is
demarked by an incline in slope of the stress-strain behavior after yield. The pure polyurethane
appears to strain harden immediately after yield, while the addition of nano-clay systematically
increases the onset of strain hardening and decreases the resulting slope. Greater deformation is
necessary in order to rotate, re-orient and order the Laponite filled soft segment in order to
release soft segments constrained by the clay platelets and allow those chains to fully uncoil and
148

stretch.

T=70 0C

T=25 0C

LU

a
L

0
0

I0
U)
0

I-u)

I-

..

I-

I-

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1

0.7

True Strain

True Strain

0
Figure 6-7: True stress-strain response of the MM5510 polyurethane composites at 25 C, a temperature
0
0
when the
temperature
transition
the
below the transition temperature (55 C), and 70 C, a temperature above
length.
films are stretched 2.5 times their original

Interestingly, at temperatures below the transition temperature the addition of nano-clay


increases the plastic strain contribution and energy absorbed upon the first loading-unloading
cycle, but decreases the absorbable energy upon re-loading and unloading. In contrast, at
temperatures above the transition temperature the plastic strain too increases as does the
absorbable energy upon re-loading and unloading.
In the hysteresis results collected within an oven at 70'C, depicted in Figure 6-8, there is quite a
discrepancy in flow stress achieved when the specimens are stretched to 1.4 and 1.7 times their
original length. This discrepancy is likely due to a change in the way in which the oven heated.
The previous data reported was collected prior to an event that changed the oven configuration.
While in all instances the heating rate prescribed was a step function to 70'C from room
temperature, these initial tests took approximately ten minutes to equilibrate at 70C and there
was no thermal overshoot. For these initial tests the thermocouple inside the oven which
monitored the temperature and controlled the feedback loop was surrounded by an insulator.

149

Consequently, the monitored thermal response was more gradual and smooth. After a change in
configuration and the thermocouple insulator irreversibly eliminated, as was the case for the later
T=7NrC

True Strain

T=7n

True Strain
T=70 0C

T=7nor.

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1

True Strain

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1

True Strain

Figure 6-8: Uniaxial tensile response of the various nanocomposites at 700 C, a temperature above
the
transition temperature. The red curve depicts the response of the I st and th cycle stretching the specimen
1.4
5
times its original length and unloading. The green curve depicts the response of the st and th cycle
stretching
1
5
the specimen 1.7 times its original length. The black curve depicts the response of the St cycle stretching
the
specimen 2.5 times its original length. The blue curve depicts the response of the 1st and
1
5 th cycle as the
specimen is stretched to 1.2, 1.4, 1.7, 2.0, and 2.5 times its original length, consecutively.

tests, the thermal response of the oven was more choppy and coarse. As a result, the feedback

150

loop to the oven forced faster heating initially and the prescribed temperature of 700 C was
consistently overshot. While the thermocouple measured an equilibrated temperature ranging
between 720 C and 75'C an independent thermometer placed within the oven read 800 C and
85C. This discrepancy leads to greater thermal softening and accounts for the variation in
response. While the thermocouple output a temperature value within 5'C of the prescribed value
the actual temperature could have been 100 C warmer and fully accounted for the drop in flow
stress in these later tests.

Despite the discrepancy in flow stress the nanocomposites at

temperatures above the glass transition temperature exhibit the stretch-induced softening typical
of elastomers like rubber [13].
^

=2.5. T=700C

0
a.

S
0-

a.

*1

6,

It

True Strain
Post .f

40

-0
-

True Strain

n=1.4 @ T , T=25C

wt%
4 wt/.
10 wt%

-15

wt%

T
a.
U

a.

30

True Strain

020

Si

a,

I.-

10

0.0

0.1

0.2

0.3

0.4

True Strain

0.5

0.6

0.7

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

True Strain

True Strain

Figure 6-9: Uniaxial tensile response of the various nanocomposites at 70 C and the corresponding response
at 25 0 C after five cycles of deformation at the elevated temperature.

Because shape memory polymers are often programmed for initial deformation at an elevated
temperature and then cooled for storage in order to reduce the amount of work needed to deform
the material, the response of the polyurethane nanocomposites at the colder, storage temperature

151

after deformation at the elevated temperature is of interest. Like many elastomers, we observed
that as the applied specimen stretch increased cyclically at elevated temperature, the initial
modulus subsequently softened and the amount of plastic strain incurred increased.
After undergoing a small amount of plastic deformation at 70C, the nanocomposites exhibit
increasing initial moduli with nano-clay concentration at 25'C, as is depicted in the bottom left
image in Figure 6-9. The locations of these responses with respect to the mechanical behavior of
the polyurethane composites at room temperature are visualized in Figure 6-10. The plastic
strain, given by the location of increasing stress in the bottom images of Figure 6-9, also
increases with nano-clay concentration, except in the 15 wt% specimen. This discrepancy is
likely due to the occurrence of localized deformation that takes place outside the observable
range of the video extensometer and within the specimen grips. A similar discrepancy is
observed when the specimens are stretched to 1.7 times the original length at an elevated
temperature in the 4 wt% filled sample. Again, at this stretch and the larger stretch, the plastic
deformation incurred increases with nano-clay concentration, especially when the concentration
is greater than the percolation limit. However, after deformation at these larger applied strains,
the moduli of the nanocomposites are essentially constant at room temperature no matter the
nano-clay concentration.

The response of these materials from one sample to the next at room temperature is generally
compatible.

However, there were a few discrepancies, especially in the magnitudes of the flow

stress of the 10 wt% and 15 wt% nano-clay filled specimens after deformation at elevated
temperatures. These discrepancies can be attributed to natural differences that occurred during
sample preparation, post-processing, and aging. As mentioned in section 6.3.2, multiple softsegment glass transitions between 10"C and 65'C were observed via dynamic mechanical
analysis when the samples were allowed to age within desiccators. This behavior was easily
observed in the highly filled composites because the transitions present in no two specimens of a
particular concentration coincided exactly. So while the samples were protected from
atmospheric humidity, due to limited accessibility to the ELF 3200, two months time passed
between the first and last tests, and likely manifested the discrepancies in stress-strain behavior.

152

T=25 0C

T=25C

.
(U
()

(,

u,

I-

I-

True Strain

True Strain

T=25 0C

T=250C
-

0I.

aCL

:E

()
U)

(0

7,

I-

I-

True Strain

True Strain

0
Figure 6-10: Uniaxial tensile response of the various nanocomposites at 25 C, a temperature below the
transition temperature. The red curve depicts the response after cyclic stretching of the specimen 1.4 times
its original length at an elevated temperature. The green curve depicts the response after cyclic stretching of
the specimen 1.7 times its original length at an elevated temperature. The cyan curve depicts the response
after cyclic stretching of the specimen 2.5 times its original length at an elevated temperature. The black
curve depicts the response of the Ist cycle stretching the specimen 2.5 times its original length. The blue curve
its
depicts the response of the 1 st and 5th cycle as the specimen is stretched to 1.2, 1.4, 1.7, 2.0, and 2.5 times
original length, consecutively.

153

6.3.5 Free Strain Recovery


The addition of nano-clay, particularly at concentrations above the percolation threshold, to
shape memory polyurethane significantly decreases the strain recovered during thermal actuation
when no load is applied. Figures 6-11 and 6-12 depict the nominal stretch recovery of the
polyurethane nanocomposites when a maximum stretch of 1.1, 1.2, 1.4, 1.7, or 2.0 is applied at
70 0 C.
Owt %

r
2.0-

'

'I

2.0-

-1.1-

'

4 wt %

1-

1.9

I-

1.

-1.2-

1.9-

1.7

1.6

1.7-

1.61.4-

1.4

1.3-

1.3

1.1

1~

1.1

1.2040 30

1.1

1.(I

220

II

30

40

50

60

60

20

30

40

&

50

II I

1.7

70

Temperature (C)

1.9
1.8

|,

60

10 wt %
I

1.-

Temperature (C)

2.0

1.7 2.0

I-

-1.1 -1.2
-1.4-1.7
- 2.0-

15 wt%
2.01.2.1.4

1. 1.8

--1

1.7 -

1.
3
I
1

-<1.5
1.4

1.3 -

1.3

1.2-

1.2

1. -

1.1

1.1
1.

~
1.0
2(

I
30

I
40

I
50

i I

20

Temperature (C)

30

I I

40

50

60

70

Temperature (C)

Figure 6-11: Stretch recovery of the nanocomposites after stretching the specimen to 1.1, 1.2, 1.4, 1.7, or 2.0
times its original length at 70C using the TA Instruments Q800 DMA. The stretch reported is the nominal
stretch or specimen stretch.

The initial steep drop in specimen stretch at 250 C correlates to removal of a tensile load. Then

154

slow recovery ensues until the transition temperature of the nanocomposite is approached. At
that point, the strain is recovered at a faster rate. The rate of recovery increases as the applied
strain increases. This is easily observed in the pure polyurethane sample in Figure 6-12 when
comparing the recovery within the 40'C-60'C temperature range at increasing applied stretch.
Furthermore, the addition of nano-clay, particularly at concentrations above the percolation limit,
decreases the rate at which strain is recovered during this thermal region. The addition of nanoclay at concentrations below the percolation limit decreases the temperature at which the strain
recovery begins by as much as 80 C. At concentrations above the percolation limit, the nano-clay
suppresses the onset of strain recovery by as much as 10'C for the 15 wt% filled sample.
2.0
1.9
1.8
1.7
1.6
,< 1.5

1.4
1.3
1.2
1.1-1.0c
I

20

30

40

50

60

70

Temperature (C)
Figure 6-12: Comparison of the stretch recovery of the nanocomposites after stretching the specimens to 1.1,
1.2, 1.4, 1.7, or 2.0 times their original length at 70 0C using the TA Instruments Q800 DMA. The stretch
reported is the nominal stretch or specimen stretch.

At 580 C the strain recovery rate slows until 65'C when the recovery rate again increases. The
temperatures at which these rates in strain recovery change correspond to the variation in
transition temperature behavior observed with the DMA. The flexural storage modulus
undergoes a similar behavior of slow softening, followed by a large quick drop and then gradual
decay with increasing temperature through the same temperature range. While the DMA
155

interrogates the visco-elastic response of the material, it is obvious that those same soft segment
components and Laponite are responsible for the resultant recovery.

95%4%3

0.4

0.

a
a
a

4 wt%

15 wt%

0 wt%
-

10 wt%
I

0.
0. 0

0.2

0.4
0.6
Applied Strain

0.8

1.0
Applied Strain
S I

'

'

95%2%
0.8

0.6

S0 wt%

0.0

13

4 wt%

10 wt%

15 wt%

0.0

0.2

0.4
0.6
Applied Strain

0.8

1.0

Figure 6-13: Comparison of the extent of strain recovery, total strain recovery and strain fixity for the
polyurethane nanocomposites determined after deformation at elevated temperature.

Despite the concentration of Laponite added to the polyurethane the extent of strain recovery
from one cycle to the next is approximately 95%. Just as when an elastomer is cycled to one
stretch and unloaded the plastic strain increases with each consecutive stretch, the extent of strain

156

recovery decreases. However, like with the increase in plastic strain, the extent of strain recovery
lost eventually reaches an equilibrium value. Although, the presence of nano-clay does not
0
influence the extent of strain fixity (-95%) when the polyurethane is quickly stretched at 70 C
and then quickly cooled at 100 C min -'. The influence of the nano-clay with regard to strain fixity
will be more important in deformation programs where the cooling rates are slower or the
specimen is left at a temperature near or above the transition temperature for longer periods of
time. The greatest influence of the nano-clay as shown in Figure 6-13 is on the extent of total
strain recovery, particularly at concentrations above the percolation threshold. The 4 wt% filled
nanocomposite exhibits only a slight decrease in the extent of total strain recovery from 64% for
the pure sample to 61%. When 10 wt% and 15 wt% nano-clay is dispersed within the
polyurethane matrix, the extent of total strain recovery drops to 47% and 32%, respectively. As
is observed in the previous Chapter, the addition of nano-clay, particularly at concentrations
above the percolation threshold, causes irreparable damage. Irreparable damage upon
deformation is significant in these composites where the nano-clay is preferentially embedded
within the soft domain.
In contrast to the response of nanocomposites deformed at an elevated temperature of 70'C, the
nanocomposites deformed at room temperature exhibit a large drop in strain when the load is
removed, indicating poor strain fixity. This behavior is apparent when comparing the point of
maximum applied strain in Figure 6-14, denoted by the open circle, to the initial data point of the
solid line that describes the strain recovery upon thermal actuation. The difference is accounted
for by greater inelastic strain recovery typically observed in glassy polymers [14]. As the
concentration of Laponite increases so too does the permanent plastic strain and, consequently,
the strain fixity of the most concentrated samples appears to be better than the pure polyurethane
because the inelastic strain recovery is hindered. Similar to the previous results, the extent of
total strain recovered increases with applied stretch, but decreases with nano-clay concentration.
Furthermore, the temperature demarking the onset of strain recovery increases with nano-clay
concentration.

157

0.25

n0

. . .. ..

S-4

-wt%
wt%.
-10
-t5wt%.

-4wt%
-10 wt%

0.45

0.6

-15wt%

0.40

0.7

0.5

.1.

0.35

SO
0

0.20

I
n /

0.30

0.15
S0.25

20.4

0.20

I-

0.10

0.3
0.15S
0.2
0.10

0.05

0.1
0.05
0.00

ntm
.00

20

30

40

so

60

70

Temperature (C)

20

so0

40

30

Temperature ( C)

60so

70

0.0
0.0
20

30

40

so

so0

70

Temperature, OC

Figure 6-14: Comparison of the strain recovery of the nanocomposites after stretching the specimens to 1.4,
0
was
1.7, or 2.5 times their original length at 25 C using the ELF 3200 (from left to right). The deformation
line
continuous
the
while
circle
empty
an
with
visualized and the maximum true strain achieved was reported
demarks the strain recovery during the no load condition.

In contrast to the previous results, the specimens, deformed at room temperature with the ELF
0
0
3200, exhibit strain recovery onset at temperatures as small as 30 C. At 50 C, the pure sample
begins to elongate. By including a percolated network of nano-clay, the temperature at which the
0
nanocomposites begin to elongate can be suppressed by as much as 20 C depending on the
extent of strain applied. Again, it is important to note that the qualitative behavior described in
Figures 6-14 and 6-15 is accurate. However, the temperatures at which the transitions take place
are accurate to within -5 0 C. A digital thermometer displayed as much as a 5oC discrepancy
0
between the output results from the ELF 3200 during the thermal ramp to 70 C.
Ultimately, the greater the stretch applied to the specimen the greater the orientation and
ordering of the soft chains and consequently the greater the available strain for entropic recovery.
As is shown in Figure 6-15, when the strain applied does not significantly stretch the sample past
yield, the extent of strain recovered is small. At stretches approximately three times greater,
when the specimen has been stretched well past yield, the strain recovery is significantly
increased. However, the presence of rigid nano-clay within the soft segments retards orientation
and alignment. The soft segment then incurs irreparable damage in order to achieve orientation
and chain alignment when the chains break, stretch or slip around the platelets, break away from
the platelets, or rotate the platelets [15].

Consequently, the addition of nano-clay, at

concentrations above the percolation threshold, significantly limits the extent of total strain
recovery upon thermal actuation.

158

0 wt%/o
I

'

'

-x

4wtc

n
-

'

'

specimen

=
1.4
S
specimen=

=,cimen
i =1.7
S--s-- " specimen=2.

=1.40.8

0.8

Specimen

Sspecimen=2.5-

0.7

0.6

0.6
0.5

specimn

0.7

0O

0.5

0.4

0.4
O
0.3

0.3
0

0.2

0.2

III

0.1

0.1

0.0

0.0
30

20

50

40

20

70

60

30

0.9

10 wt%
I

'

'

'

40

60

50

70

Temperature (OC)

Temperature (oC)
I

'

15 wt%

09

'
l

Sspecimen
0.8

0.8

Sspecimen =1.7
--- "

0.7

specimen

~=2.5

0.7

=1.4

2specimenl.

=1.7specimen

0.6

0.6

-eX

U-

0.5

0.

S0.4

0.4
C

0.3

0.3

0.2

0.2

2.5I

S-I
1-

0.1
I

30

40

50

60

20

70

40

30

50

60

70

Temperature (OC)

Temperature (OC)

Figure 6-15: The strain recovery of the polyurethane nanocomposites containing 0, 4, 10 and 15 wt% nanoclay after stretching the specimens to 1.4, 1.7, or 2.0 times their original length at 250C using the ELF 3200.
The deformation was videoed and the maximum true strain achieved was reported with an empty circle while
the continuous line demarks the strain recovery during the no load condition.

159
-,

6.3.6 Constrained Strain, Stress Recovery


The attraction of researchers to shape memory polymers for numerous applications
is often based
on utilization of an actuation stress, or recovery stress when the polymer is constrained
from
recovering strain [2, 16-20]. However, it is difficult to isolate the stress due to entropic
recovery
under a constrained strain condition from thermal expansion/contraction stresses,
thermal
softening and stress relaxation. Thermal expansion, thermal softening and stress relaxation
under
a constrained strain conditions causes a drop in the tensile stress.
Initial Modulus (MPa)
T = 25*C
T = 70C
142
29
290
34
461
81
451
137

0 wt%
4 wt%
10 wt%
15 wt%

CTE (117C)
T = 25C
T = 70*C
6.14E-05
1.12E-03
6.21E-05
8.12E-04
4.76E-05
8.04E-04
3.97E-05
7.98E-04

Predicted Thermal Stress (MPa)


min
max
0.39
1.46
0.81
1.24
0.99
2.93
0.81
4.92

Table 6-2: Predicted thermal stress of the various polyurethane nanocomposites


when the length is
constrained. The initial modulus at room temperature multiplied by the axial coefficient
of thermal expansion
at room temperature and the operating temperature difference (70 0C-25 0 C=450 C)
gives a lower estimate of
the predicted thermal stress. The upper estimate is given by the initial modulus at
the elevated temperature
multiplied by the axial coefficient of thermal expansion at the same elevated
temperature and operating
temperature difference.

0t.

'

(A5

(5
034

_=,

C
w

C 5

03
.,
C

CD

0'C
LU

C1
C
W

0
20

30

40

50

Temperature (OC)

60

70

0
20

30

40

50

Temperature (OC)

60

70

20

30

40

50

60

70

Temperature (OC)

Figure 6.16: Comparison of the stress recovered when the strain is constrained
after stretching the
polyurethane nanocomposite specimens to 1.4, 1.7, or 2.5 times their original
length at 25 0 C using the ELF
3200. The engineering stress is reported because the sample exhibited localized
deformation that was not
entirely captured by the video extensometer.

The magnitude of the drop due to thermal expansion can be estimated by multiplying
the
modulus of the material, E, by the axial coefficient of thermal expansion, a, and the
change in
temperature. Table 6-2 predicts the minimum and maximum thermal stress
reduction. The
minimum thermal stress reduction predicted for pure polyurethane is -0.4 MPa.
This drop in
stress can be observed in Figure 6-16 over the 25 0 C to 70'C temperature
range when the

160

specimen was constrained after stretching the specimen 1.4 times its original length. The nanoclay reinforced samples exhibit drops greater in magnitude as predicted. This drop is augmented
by thermal softening of the material and negated by time-dependent stress recovery.
Figure 6.16 depicts the recovery stress of each nanocomposite after stretching the specimen to
1.4, 1.7, or 2.5 times its original length. In actuality though, due to the localized deformation
outside the extensometer viewing area discussed previously, the actual stretches achieved are
listed in Table 6-3. Consequently, the actual stretches achieved by the 10 wt% and 15 wt% are
less than that achieved by the 4 wt% sample. The resulting actuation stresses achieved are twothirds of that achieved by the 4 wt% sample. While increasing the stretch applied increases the
available energy for recovery, the addition of nano-clay also increases the available energy and,
consequently, the resulting actuation stress achieved. This is confirmed when the pure
polyurethane and 4 wt% filled composite stretched 1.7 times their original length are compared
or when the pure and 10 wt% filled polyurethanes are stretched 2.5 times their original length. In
both instances the nano-clay filled composite exhibits the greater actuation or recovery stress.

A specimen

0 wt%

4 wt%

10 wt%

15 wt%

1.4
1.7
2.5

1.22
1.36
1.73

1.30
1.35
1.53

1.17'
1.29
1.73

1.11
1.49
1.68

Table 6-3: Comparison of the true stretch achieved in each nanocomposite as seen via the video extensometer
during the constrained strain, stress recovery test when the specimen was stretched to 1.4, 1.7, or 2.5 its
original length.

As the nano-clay concentration increases the actuation stress too increases, except when 15 wt%
nano-clay is added. This drop in actuation stress is due to the poorer Laponite dispersal and
increased plastic deformation due to the increased presence of rigid, flocculated, intercalated
platelets. The peak in actuation stress of the pure polyurethane occurs almost immediately at
-30 0 C and the stress plateaus at nearly this same value until the temperature is held at 70'C for
an extended period of time, as shown in Figure 6-17. In contrast, the nanocomposite actuation
stress peaks at -40'C and then drops as the material thermally softens, and expands until 70'C,
when the soft segments easily reptate and flow and the rate of softening significantly increases.

161

0 wtO/%

8
0

0
4 wt/o
wt /o

0wV'!
00

pecimen=1.4
.
specimen
Xspecimen

=1.7=2 5

0.

06

5
U)
CO4

0)4
I.. 3

03
OI

1
n

v
30

40

50

60

20

70

30

40

50

60

70

60

70

Temperature (OC)

Temperature (oC)
8

M 6

M6

0.

O4

04
1..

L 2
01
Cio
UI

n S-20

30

40

50

60

20

70

30

40

50

Temperature (OC)

Temperature (OC)

each polyurethane
Figure 6.17: The stress recovered when the strain is constrained after stretching
0
ELF 3200. The
the
using
C
25
at
length
original
their
times
2.5
or
nanocomposite specimen to 1.4, 1.7,
not entirely
was
that
deformation
localized
exhibited
sample
the
because
reported
engineering stress is
extensometer.
video
the
by
captured

This behavior contrasts the behavior reported by Miaudet et al [21] when the actuation stress of a
carbon nanotube-polyvinyl alcohol peaks at a temperature roughly equal to the deformation
temperature. In this polyurethane-nano-clay system, the peak in actuation stress occurs nearly
162

150 C above the deformation temperature or room temperature when clay is added. These peaks
instead correspond to the transition behavior observed in the dynamic mechanical results.

6.4 Conclusions
Nano-clay is preferentially embedded within the soft domains of commercial shape memory
polyurethane using the solvent exchange approach previously described. The nano-clay is fully
exfoliated; however, at concentrations above the percolation limit, intercalation persists and
flocculation amid the most polar soft segment constituents is observable in the most concentrated
composite. The complex nature of this multi-block copolymer is further manifested when nanoclay is added. The manufacturer reports the soft segment glass transition as 55'C, but differential
scanning calorimetry, as well as dynamic mechanic analysis, indicates that two and three
transitions within 100 C and 65'C are present. At temperatures above the transition temperature,
the nanocomposites behave like typical elastomers-exhibiting stretch-induced softening and
hysteresis. At temperatures above and below the transition temperature, the addition of nano-clay
decreases the axial coefficients of thermal expansion as well as increases the plastic strain. At
temperatures below the transition temperature, the nanocomposites act like glassy polymersexhibiting significant inelastic recovery. However, the addition of nano-clay decreases the extent
of inelastic recovery. Consequently, the strain fixity of the material when stretched at
temperatures below the transition temperature is poor compared to that when the nanocomposites
are stretched at an elevated temperature and cooled quickly. No matter the deformation
temperature,

the extent of total strain recovered

decreases with increasing Laponite

concentration, particularly at concentrations above the percolation threshold. The addition of


nano-clay also delays the onset of strain recovery as well as stress recovery. However, as the
stretch applied to each specimen increases so too does the actuation stress. The addition of well
dispersed nano-clay to the polyurethane increases the actuation stress; however, as the
temperature increases, or is held at an elevated value, the recovery stress decays significantly
while the pure polyurethane holds constant. Ultimately, the optimum nano-clay concentration
where the greatest strain recovery is balanced by the largest actuation stress is equal to the
percolation limit. At larger concentrations, the amount of strain able to be recovered decreases
significantly while the actuation stress increases. The actuation stress will increase with filler
loading until the nano-clay dispersal becomes too intercalated that plastic deformation
dominates. At smaller concentrations, the strain able to be recovered is maximized, but the

163

obtainable recovery stress is minimized. Ultimately, the addition of nano-clay to shape memory
polyurethane allows another tool for the design of actuating polymers with prescribed stress and
strain recovery.

6.5 References
[1]

[2]

[3]

[4]
[5]
[6]

[7]
[8]
[9]

[10]
[11]

[12]

[13]

[14]
[15]

M. A. Di Prima, M. Lesniewski, K. Gall, D. L. McDowell, T. Sanderson, and D.


Campbell, "Thermo-mechanical behavior of epoxy shape memory polymer foams,"
Smart Materials& Structures, vol. 16, pp. 2330-2340, 2007.
G. M. Spinks, V. Mottaghitalab, M. Bahrami-Samani, P. G. Whitten, and G. G. Wallace,
"Carbon-Nanotube-Reinforced Polyaniline Fibers for High-Strength Artificial Muscles,"
Advanced Materials, vol. 18, pp. 637-640, 2006.
K. J. Pawlowski, H. L. Belvin, D. L. Raney, J. Su, J. S. Harrison, and E. J. Siochi,
"Electrospinning of a micro-air vehicle wing skin," Polymer, vol. 44, pp. 1309-1314,
2003.
A. Lendlein and S. Kelch, "Shape-memory polymers," Angewandte Chemie-International
Edition, vol. 41, pp. 2034-2057, 2002.
M. Szycher, Szycher's Handbook ofPolyurethanes.Boca Raton, Florida: CRC Press,
1999.
S. Rezanejad and M. Kokabi, "Shape memory and mechanical properties of cross-linked
polyethylene/clay nanocomposites," European Polymer Journal,vol. 43, pp. 2856-2865,
2007.
F. Cao and S. C. Jana, "Nanoclay-tethered shape memory polyurethane nanocomposites,"
Polymer, vol. 48, pp. 3790-3800, 2007.
M. Y. Razzaq and L. Frormann, "Thermomechanical studies of aluminum nitride filled
shape memory polymer composites," Polymer Composites, vol. 28, pp. 287-293, 2007.
B. Yang, W. M. Huang, C. Li, and J. H. Chor, "Effects of moisture on the glass transition
temperature of polyurethane shape memory polymer filled with nano-carbon powder,"
European Polymer Journal,vol. 41, pp. 1123-1128, 2005.
R. J. Young and P. A. Lovell, Introduction to Polymers. Cheltenham, U.K.: Nelson
Thornes Ltd., 1991.
B. Yang, W. M. Huang, C. Li, C. M. Lee, and L. Li, "On the effects of moisture in a
polyurethane shape memory polymer," Smart Materials & Structures, vol. 13, pp. 191195, 2004.
N. Sheng, Multiscale MicromechanicalModeling of the Thermal/MechanicalProperties
ofPolymer/Clay Nanocomposites, Ph.D. in Dept. of Mechanical Engineering,
Cambridge, MA:Massachusetts Institute of Technology, 2006.
H. J. Qi and M. C. Boyce, "Constitutive model for stretch-induced softening of the stressstretch behavior of elastomeric materials," Journalof the Mechanics and Physics of
Solids, vol. 52, pp. 2187-2205, 2004.
0. A. Hasan and M. C. Boyce, "A constitutive model for the nonlinear viscoelastic
viscoplastic behavior of glassy polymers," Polymer EngineeringAnd Science, vol. 35, pp.
331-344, 1995.
J. Diani, M. Brieu, and J. M. Vacherand, "A damage directional constitutive model for
Mullins effect with permanent set and induced anisotropy," EuropeanJournalof
Mechanics A: Solids, vol. 25, pp. 483-496, 2006.

164

[16]

[17]

[18]

[19]
[20]
[21]

P. R. Buckley, G. H. McKinley, T. S. Wilson, W. Small, W. J. Benett, J. P. Bearinger, M.


W. McElfresh, and D. J. Maitland, "Inductively heated shape memory polymer for the
magnetic actuation of medical devices," Ieee Transactions On Biomedical Engineering,
vol. 53, pp. 2075-2083, 2006.
K. Gall, M. L. Dunn, Y. P. Liu, G. Stefanic, and D. Balzar, "Internal stress storage in
shape memory polymer nanocomposites," Applied Physics Letters, vol. 85, pp. 290-292,
2004.
K. Gall, P. Kreiner, D. Turner, and M. Hulse, "Shape-Memory Polymers for
Microelectromechanical Systems," Journalof MicroelectromechanicalSystems, vol. 13,
pp. 472-483, 2004.
A. Lendlein and R. Langer, "Biodegradable, Elastic Shape-Memory Polymers for
Potential Biomedical Applications," Science, vol. 296, pp. 1673-1676, 2002.
D. Ratna and J. Karger-Kocsis, "Recent advances in shape memory polymers and
composites: a review," JournalOf Materials Science, vol. 43, pp. 254-269, 2008.
P. Miaudet, A. Derre, M. Maugey, C. Zakri, P. M. Piccione, R. Inoubli, and P. Poulin,
"Shape and temperature memory of nanocomposites with broadened glass transition,"
Science, vol. 318, pp. 1294-1296, 2007.

165

7. Development of Multi-Responsive Electrospun Fabric


[This work will be submitted for publication in a slightly different form, as "Multi-Responsive Non-Woven
Electrospun Polyurethane Mats Recover Intrinsic Stretch" by S.M. Liff and G.H. McKinley to Polymer in
September 2008.]

7.1 Introduction
Electrospun, non-woven, polymer nano-fibers exhibit interesting properties due to the large
surface area and large draw-ratio induced by the electric field and three-dimensional whipping
instability [1-3]. The present work evaluates the entropic strain recovery of electrospun, nonwoven, thermoplastic polyurethane mats composed of either smooth continuous fibers or
composed of 'beads on a string' (BOAS) morphology that respond to both thermal and aqueous
stimuli. The BOAS mats exhibit the greatest strain recovery and annealing the mats at a
temperature near the glass transition temperature releases the intrinsic stretch 'frozen' into the
material via electro-spinning. In response to either stimuli the porous, non-woven mats shrink
macroscopically, maintaining mechanical integrity and simultaneously exhibiting fiber diameter
growth coupled with pore diameter shrinkage that nearly eliminates all pores greater than three
pm in size. Furthermore, this paper demonstrates that addition of nanoparticles to the responsive,
biocompatible, thermoplastic polyurethane mats using the solvent exchange technique detailed in
Chapter 3 allows the fiber morphology to be tuned and treatment of the surface with an
oleophobic coating allows development of fabrics with reversible wettability.
Materials that exhibit shape change, whether metal alloy, ceramic, polymer or gel, have
generated great interest in the past decade as actuators in the aerospace, textile and biomedical
industries. Shape memory, as reviewed in Chapter 6, is the ability of a material to maintain a
deformed shape and upon external stimulus recover the original, pre-deformed shape. The
actuating external stimulus induces entropic recovery and is most often thermal but for
polymeric materials and gels ultraviolet stimulation may also be employed [4, 5]. Ionic strength,
pH, and solvent quality may also be used to induce shape transitions in gels [5]. Actuation
stresses span the range from 103 Pa to 106 Pa and recovery strains may vary from 10% to 100%
depending on the polymer employed.

166

Typical preparation of a SMP includes production of the material in some specified and desired
permanent configuration, e.g. as a cylindrical fiber of length lo and radius ro. The SMP is then
programmed or deformed to meet specifications of the desired temporary shape, e.g. a longer
thinner cylindrical fiber of length I ( I > lo ) and radius r ( r < ro). Often the deformation
temperature is selected to be above the inherent transition temperature (Tra,,s=Tg or T,n) of the
polymer-the temperature which triggers entropic recovery-so that less work is needed to
deform the material. If deformed at an elevated temperature the polymer is then cooled to the
storage temperature, vitrifying or crystallizing the softer phase and holding or 'setting' the
deformed, temporary shape. This temporary shape is maintained while the temperature is held
below T,an,,s; however, upon an increase in temperature to a value greater than or equal to Trans
the mobile chains of the second phase entropically return to their previous configuration and the
permanent shape is recovered-the fiber returns to its original length lo and radius ro. It would be
advantageous to utilize a polymeric processing method that deforms the material to the desired
temporary shape as it is processed so that no extra mechanical work input is needed. One such
processing method is electrospinning. Furthermore, utilization of shape recovery is not a new
idea-Shrinky Dinks, popular toys since 1973 that contract to one-third their original size and
grow nine times through their thickness upon oven heating, exclusively utilize shape memory
[6]. However, neither the development of porous, shape memory materials, the characterization
of their respective recovery, nor devices that exploit shape change and porosity have been
addressed previous to this work. Consequently, the development of electrospun, shape memory
materials, their characterization and their exploitation discussed in this chapter is the first of its
kind and may transform future filter and drug delivery device design.
Electrospinning is a processing method that accelerates a polymer solution from a nozzle under
action of an electric field to make fibers with diameters that range from 10 nm to 10 tm. The
process exploits surface charge-repulsion, polymer viscoelasticity, surface tension, extensional
viscosity, and solvent evaporation to deform and solidify material elements without capillary
breakup [1, 3].

The large draw-ratio induced by the electric field and three-dimensional

whipping instability significantly stretches the polymeric chains. Upon removal of solvent the
intrinsic stretch is 'frozen' or locked in the fiber if the polymer utilized has a Tg or Tm which is
greater than Tamb. The glass transition temperature of a polymer in solution depends strongly on

167

the concentration of polymer utilized at concentrations within the dilute solution regime [7-9].
As the polymer concentration increases so too does the glass transition temperature until the
maximum or bulk value is obtained. The glass transition temperature of concentrated solutions,
however, is not dependent upon polymer concentration, but instead remains constant and equal
to the bulk glass transition temperature. If Tg or Tm of the polymer is below Tamb the chains
within the fiber are able to flow and relax and the strain applied during processing is lost upon
fiber collection. Likewise, if the polymeric chains of the material with Tg or Tm greater than Tamb
are allowed time to relax, the intrinsic stretch applied during processing is lost. An example of
such a process is if the fibers are annealed at a temperature below but near T,,ans or the solvent is
not completely removed. Consequently, electrospinning offers a method to make non-woven
mats randomly composed of stretched polymeric fibers with Tg or Tm above Tamb that recovers
entropy and shrinks upon thermal stimulus. However, the extent of shrinkage of the non-woven
mats and fibers from which the mat is composed is a priori unknown due to an incomplete
understanding of the electrospinning process and an inability to control or monitor the draw
down of the polymeric solution during whipping.
Development of new, responsive, polymeric electrospun fibers and systems offer potential for
development of high-performance engineering thermoplastics that may be used as smart,
actuating textiles [10], selective filters [11] and coatings [12-14], as artificial muscles [4, 15-17],
shape-changing scaffolds, wound dressing or other biomedical devices [18-23]. Barhate and coworkers reviewed the status of nano-fibrous filteration in 2007 and indicate that prototypes have
been successfully prepared for antimicrobial air filtering, catalytic filtering, ion exchange
filtering among other applications; however, the key for further advancement relies upon
advances in surface treatment [11]. Yoshida et al reviewed the current trend in advanced biocoatings of utilizing biodegradable, shape memory materials to develop multi-functionalized
biomaterials that can be used to control cell growth of tissues or in drug delivery systems [14].
Pawlowski and co-workers utilized electroactive polymer fibers composed of poly(vinylidene
fluoride) and trifluoroethylene electrospun onto the wing frame of a micro air vehicle (mass less
than 50 g) to induce wing vibration at 6.7 Hz with a 2.0 kV amplitude [16]. Samatham and coworkers used electrospun polyacrylonitrile as a type of artificial muscle. By changing the pH of
the solution in which the non-woven mat was submerged, the fibers were observed to expand in

168

basic solutions and contract in acidic solutions more quickly than commercially available larger
diameter fibers [17]. There are only a handful of publications investigating electrospun, shape
memory polyurethanes and the emphasis of this previous work was often to identify spinning
parameters that enable production of a uniform fibrous morphology. Minimal mechanical
analysis or shape memory characterization was presented, and only after applying stretch after
processing and formation of the fibrous mat [10, 24-26]. The following chapter investigates the
entropic recovery of electrospun, shape memory polyurethane (SMPU) mats after processing and
discusses further applications for this behavior.

7.2 Materials and Experimental Methods


7.2.1 Materials
In this study we utilized a commercial, biocompatible, thermoplastic polyurethane, MM 5510
from the DiAPLEX Series provided by Mitsubishi Heavy Industries, Ltd. This polyurethane is
composed of diphenylmethane-4,4-diisocynate, adipic acid, ethylene oxide, polypropylene oxide,
1,4-butanediol, and bisphenol A [27] and the glass transition temperature of the soft segment is
specified by the provider as 55oC. The shape memory polyurethane (SMPU) was supplied in
pellet form and used as received. Dimethyl formamide (DMF) from Sigma Aldrich was used as
received to dissolve the polyurethane pellets. Nano-clay platelets, Laponite RD (Southern Clay),
with diameter of -25 nm and thickness of 1 nm were used as received in powder form.
7.2.2 Solution Preparation
Polyurethane-DMF solutions were prepared so that the final concentration of polyurethane in
solution was 13, 17, 20, 24 and 27 wt%. The solutions were then roll-mixed at a moderate speed
for two days prior to use to ensure that the polyurethane was completely dissolved. To disperse
Laponite in the 17 wt% polyurethane-DMF mixture, Laponite was first dispersed in pure DMF
at -0.6 wt% using the solvent exchange approach [28] detailed in Chapter 3 and then added to
the polyurethane-DMF mixture. Consequently, the resultant polyurethane concentration in
solution was less than that desired. After thoroughly mixing the solution using a magnetic stir bar
(for approximately two days) and placing the solution on a Cimarec digital hot plate stirrer, the
cap was removed and some of the excess DMF was then evaporated off the mixture at 50C
while the stir bar rotated slowly at setting of 2. The mass of the remaining solution was measured
intermittently until the desired polyurethane concentration of 17 wt% was obtained. Once
thoroughly mixed the temperature was reduced to ambient and the solution continually stirred

169

until the sample was removed for electrospinning.


7.2.3 Electrospinning Parameters and Mat Preparation
The solution was electrospun using the parallel plate setup described and depicted schematically
in [29] is also shown below in Figure 7-1.
Syringe pump

Flow-rate

v]

Cf

Figure 7-1: Schematic representation of the parallel-plate fiber spinner.

The spinning parameters for each of the solutions are given in Table 7-1. The protrusion of the
spinneret from the upper plate was 2 cm and distance, d, between the capillary tip and the
collector plate is given in Table 7-1. The fiber mats were collected on a grounded non-stick
aluminum foil.
PU Concentration, c

Voltage, V

Flow Rate, Q

Distance, d

Temperature, T

wt% in solution

kV

mL.s-1

Relative Humidity
%

13
17
17
17 (4 wt% Laponite)
20
24
27

20.0
24.0
32.7
33.8
28.0
37.1
34.3

1.67E-04
1.67E-04
1.82E-04
4.25E-05
1.67E-04
1.67E-04
1.67E-04

0.37
0.37
0.48
0.43
0.44
0.44
0.46

21
21
24
22-24
22
22
21-22

27-31
24-28
24
11-16
32-37
25-38
25-26

Table 7-1: Electrospinning parameters for each fibrous morphology discussed.

The voltage and flow rate applied to the solution as well as the laboratory temperature and
relative humidity during spinning were all recorded; however, note the electrical current, I,
during spinning was not monitored. Immediately after collection the sample and aluminum foil
were placed in an oven for forty-eight hours at 20'C with a N2 purge flow of 0.03 m3 h' and then
stored in zip-lock plastic bags at ambient temperature until samples were needed for
170

characterization. Each sample mat used for characterization was a small film-like section cut
from the original electrospun mat.
7.2.4 Characterization Methods
7.2.4.1 InterrogationofSolution Conductivity
The Solartron 1287A Potentiostat and 1260A Impedance/Gain-Phase Analyzer were used in
combination as an impedance analyzer to measure the conductivity of the pure 17 wt%
polyurethane solution and 4 wt% Laponite-filled solution. The solution was placed in an
electroporation cuvette with a 0.2 cm gap (dgap), surface area equal to 2.3 cm2 (A) and was
connected to the instrument via double-sided copper tape and alligator clamps. A frequency
sweep was run from 10-1 to 107 Hz with an AC current of 0.1 Amps and the complex plane
impedance diagram recorded. The conductivity, or, was then calculated following the procedure
detailed in [30], as shown below:
d

r7
= dgp

(7.1)

A-Z

where IZ is the closest absolute complex impedance value to the real axis when the semicircular electrical response transitions to the monotonically increasing tail, as seen in Fig 7-3.
7.2.4.2 InterrogationofFiberMorphology
The general morphology of the resultant mats was characterized using a JEOL JSM-6060
scanning electron microscope (SEM) operating at 10-15 kV with a spot size of 45. Prior to
imaging the electro-spun mats were sputter coated with a 9-12 nm coating of gold (i.e. 30
seconds of sputter time). The average fiber diameter and standard deviation was calculated from
twenty different fiber locations in one representative SEM image taken at a magnification of
2000 using the ruler tool in Adobe Photoshop CS. In situ thermal-actuated shrinkage of mats
without the gold sputter coating was captured in the FEI/Philips XL30 FEG Environmental SEM
by adhering two edges of a long rectangular mat specimen to a thermoelectric cooler using
double-sided copper tape and increasing the temperature in an instantaneous step from O'C to
62 0 C.
7.2.5.2 Interrogationof Non- Woven Mat
To actuate the material the mats were either placed on a Cimarec hot plate held at 70'C for 120
seconds or placed in a beaker of hot deionized water for twenty-four hours unless otherwise
denoted and then dried on a chem-wipe for five minutes. In either case the mats were allowed to
recover the intrinsic strain (or shrink) freely without any constraining load. The dimensions of
171

the mat samples prior and after actuation were measured using calipers. The stretch recovered in
each of the lateral directions was calculated as follows
L
&ecovered

(7.2)

where Lo represents the initial lateral dimension and L represents that same dimension after
actuation. The average value and standard deviation reported are based on two measurements
taken from four different mat samples.

The thermal transition temperatures of the fiber mats were evaluated and the presence of water
absorbed within the fibers was verified using a TA Instruments Q1000 differential scanning
calorimeter (DSC) with a 10C min - ramp rate from -90'C to 250'C. The thermomechanical
properties of the mats were evaluated using a TA Instruments Q800 dynamic mechanical
analyzer (DMA) with a 30 C min -' ramp rate from -150'C to 2000 C.
The amount of water absorbed into the electrospun mats was monitored with a laboratory mass
balance with a 0.1 mg resolution. At various recorded times the mat was removed from the
beaker of deionized water, wrapped in a chem-wipe for 120 seconds to wick away any excess
water trapped amid pores or pinned to the surface and then the mass measured. Obviously, this
technique is filled with numerous sources of error; however, the results allow an order of
magnitude estimate of the mass diffusivity to be determined.
7.2.5.3 Interrogationof Pore Size
The pore size distribution of the electrospun mats was measured using Micromeritics AutoPore
IV 9500 mercury porosimeter. As the pressure increased incrementally from 3.65 kPa to
68.8 MPa the mercury invaded smaller and smaller specimen pores. Pore size was determined
based on evaluation of the capillary pressure given by the following equation
P=,

4y cos 0
D

(7.3)

where P is the pressure, y = 0.48 N m -' is the surface tension of mercury, 0 = 140' is the
assumed contact angle between mercury and the pore wall, and D is the calculated cylindrical
pore diameter following the arguments of Washburn [31, 32].

172

7.3 Results and Discussion


7.3.1 Fibrous Morphology
While many polymer solutions have a small envelope of spinnable concentrations [1] this
commercial, shape memory polyurethane (SMPU) was found to be spinnable in DMF over a
wide range of concentrations from 13 wt% to 27 wt% in solution.

Figure 7-2: Scanning electron micrographs demonstrate the change in fibrous morphology as the
concentration of polyurethane in solution increases: a) 13 wt%, b) 17 wt%; low Q; low E, c) 17 wt%; high Q;
high E, d) 20 wt%, e) 24 wt%, and f) 27 wt% (scale bar is 10 pm in all images).

The resulting fibrous morphologies that correspond to the spinning parameters of Table 7-1 are
displayed in Figure 7-2. As the concentration increases above 20 wt% polyurethane in solution,
the fiber morphology transitions from 'beads on a string' (BOAS) to smooth continuous fibers. It
is important to note that the resulting fiber diameters as well as bead diameters are polydisperse.
The beads observed in the 13 and 17 wt% based mats are frequent and quite spherical in contrast
to the fewer, more elongated beads observed in the mat obtained from 20 wt% solution. The
resulting shape of the bead is determined by the relaxation time afforded the polymeric material
0
comprising the beads. By utilizing a more volatile solvent, such as tetrahydrofuran (B.P. = 66 C)
instead of dimethylformamide (B.P. = 153oC) or by increasing the characteristic relaxation of the
polymer with nano-filler the beads become more elongated and less spherical [33]. While the

173

beads resulting from the polyurethane-dimthylformamide solution appear smooth and spherical,
the beads Tuteja et al observe in [13] when poly(methyl methacrylate) and tetrahydrofuran are
electrospun appear to be oblong and wrinkled. These oddly shaped beads result when high
concentrations of polymer coupled with volatile solvents are utilized in the electrospinning
solution because a solid skin forms along the fiber due to quick evaporation. Meanwhile, solvent
still remains within the fiber unless the material is further post-processed.
It is important to note that scanning electron micrographs shown in Figure 7-2b and 7-2c are of
two separate electrospun mats both utilizing 17 wt% polyurethane in solution. However, as noted
in Table 7-1, the flow rate, Q, and the magnitude of the electric field, E, given by the voltage, V,
divided by the distance, d, between capillary tip and collection plate differ. The morphology
exhibited in Figure 7-2b which is characterized by a polydisperse assortment of large beads,
most of which measure 8-10 pm in diameter, connected by strings with an average diameter of
170 + 70 nm, are produced when a flow rate of Q = 1.67x10-4 mL s- 1 and an electric field of E =
6.52x104 kV m

are utilized. In contrast, the resultant morphology when a higher flow rate

(Q= 1.82x10-4mL s-1) and electric field magnitude (E=6.81x10 4 kV min)

are utilized is

characterized by numerous beads measuring 2-4 jim in diameter connected by strings with an
average diameter of 130 30 nm.

By adding Laponite platelets at a small to moderate concentration, 0 < 0, which does not
exceed the percolation limit (Op = 6 wt% [28]),

to the 17 wt% polyurethane solution it is

possible to further tailor the resulting fibrous morphology and extend the spinnable solution
envelope. If the Laponite concentration exceeds the percolation limit the microstructure in the
Laponite/polymer solution jams, hindering the development of solution viscoelasticity and
promoting capillary break-up. Even when a small concentration of Laponite is added to the
solution (4 wt% in the solid or -0.7 wt% in solution) the conductivity increases to -60 times that
of the pure solution. Figure 7-3 depicts the complex plane impedance diagram of the two
solutions. Note that the absolute magnitude of the complex impedance of the polyurethane
solution is 9.8x104 Q while that of the Laponite filled solution is 1.9x103 Q; therefore, the
resultant conductivity is 8.9x10 -7 S cm -1 and 5.5x10 -5 S cm-1, respectively.

174

ann
3WVV

'u, ,J'J

150000

2000

100000

1000
50000

1500

2500

2000

3000

Zr (n)

Z (a)

solution with (b) and without


Figure 7-3: Complex plane impedance diagram of the 17 wt% polyurethane
alculated.
Lap,

17

and of a 4
Figure 7-4: Microscopic images of the as spun, pure, non-woven mat prior to and after actuation
(c) pure
shrink,
and
actuation
thermal
after
mat
pure
(b)
mat,
spun
as
wt% Laponite-filled mat: (a) pure,
smaller
Note
spun.
as
mat
Laponite-filled
wt%
4
(d)
and
shrink,
and
actuation
mat after aqueous
the
actuation
upon
magnification in image c allows greater viewing area. Despite the addition of Laponite
strings relax and re-coil and grow in diameter while the beads remain unchanged.

175

The charged Laponite discs act effectively like a salt increasing the conductivity of the solution
but simultaneously increasing the solution viscosity and relaxation time [34, 35]. Consequently,
as shown in Figure 7-4d, the addition of enough Laponite so the final concentration in the solid is
4 wt% allows a slower flow rate and greater electric field magnitude to be applied
(Q = 4.25x10-5 mL s; E= 7.83x10 4 kV m'-1). This produces smaller, elongated beads that
measure less than 2 tm in diameter and are connected by finer fibers with average diameter of
120 20 nm.

7.3.2 Results of Thermal and Aqueous Actuation


The ambient temperature (Tamb) at which the material is spun and stored is between 20-25'C. The
glass transition of the soft phase or transitioning phase (Tg = 55'C) of the SMPU is greater than
Tamb. Despite the fiber morphology, the stretch applied during electrospinning is 'frozen' in the
strings or fibers. Consequently, upon thermally stimulating the non-woven mat (e.g. by placing it
atop a hot plate set at 700 C) the mat shrinks in each lateral direction and the thickness of the
fibers increases as shown in the movies in Appendices A-5. The extent of the change in fibrous
morphology upon heating, namely the increase in fiber diameter and noticeable relaxation and
coiling of the strings, is clearly observed in the micrographs presented in Figures 7-4a and 7-4b.
Upon heating, the taut polymer chains relax and re-coil, attempting to recover the entropy lost
during electrospinning. The elongated chains are driven to return to the coiled configuration in
solution. Consequently, the average diameter of strings after thermal stimulation measure 260 +
80 nm and are two times thicker than they were prior to stimulation.
These fibrous materials respond not only to thermal stimulus but also to aqueous stimulus [36].
Because the polyurethane utilized contains ethylene oxide, a very hydrophilic component, within
its transition segment or softer phase it is possible to plasticize the segments with water and
allow the stretched polymeric chains to relax and recovery entropy without increasing the
temperature. Although the water is absorbed and plasticizes the ethylene oxide component of the
linear block-copolymer, the Tg measured via differential scanning calorimetry does not
significantly change as is shown in Figure 7-5.

176

4,

f%-la

I
i0

..2

U.

-50

100
50
Temperature ("C)

150

-40

200

100
50
Temperature ("C)

150

200

Figure 7-5: Differential scanning calorimetry of the as spun non-woven mat, the thermally actuated mat, and
the water soaked mat prior to drying and after drying. Note the image on the right is a zoomed view of that
on the left.

The rate of absorption of water into the fibers is highly dependent on the surface area to volume
ratio of the available fibers [37]. Consequently, fibers of high aspect ratio (a = 1 / d where 1 is
fiber length and d is fiber diameter) absorb water more quickly than mats composed of shorter or
larger diameter fibers. Mats consisting of a 'beads on a ' morphology may absorb less water than
a mat consisting of thicker continuous fibers if the overall surface area to volume ratio of the
strings and beads is less than that of the other mat composed larger diameter fibers [37]. This
actuatable system is more complex than that of the thermal stimuli system because the water is
absorbed first. The 'BOAS' electrospun mat prepared at C= 17 wt% with a high electric field and
high flow rate increases its mass almost fourfold in five minutes when exposed to water.
However, as the mat continues to be soaked in the water bath the polymer chains relax and the
mat contracts laterally and water is squeezed from the mat so that an equilibrium mass of only
1.8 times the original mass is achieved (see Figure 7-6 below). In forty-eight hours the
equilibrium mass is reached and in twenty-four hours the mass is within 10% of the equilibrium
value. Despite utilization of high aspect ratio nanofibers, response to aqueous stimulation is
observed to be -102 times slower than the response to thermal stimulation.

177

at

45

yI

r-~

40

S1035
(a

30
25

L0

20

IA

00:00

00:10

00:20

00:30

00:4

00:50

01:001

Elapsed Time (HH:mm)

24:00

48:00

72:00

96:00

120:00

144:00

168:00

192:00

Elapsed Time (HH:mm)

Figure 7-6: Evolution of water uptake in the electrospun mat as it is soaked in a DI H 0 bath.
Notice the
2
elapsed time scale changes after one hour.

The Lewis number (Le),


Le = a
D

(7.4)

a dimensionless number which compares the thermal diffusivity (a) of a system to the mass
diffusivity (Dc), typically used to characterize fluid flows, can be approximated in the following
manner. The thermal diffusivity can be estimated to be:
a

pCp

0m

k - 5 - 7 x 10

(7.5)

where k is the thermal conductivity (and varies from 0.12 to 0.2 W m-lK -' for this polyurethane
depending on temperature [27, 38]), p is the density ( 1200 kg m-3 [27]), and c, is the specific
heat which is 2.0 kJ kg-'K ]- at 300'C and 2.7 kJ kg-'K - 1 above the transition temperature (based
on results from DSC). The mass diffusivity is not easily calculated for this fibrous polyurethane
system due to the inherent heterogeneities in microphase-separated polyurethanes coupled with
the fact that the diffusivity changes as the concentration of water within the fiber
increases [39,
40]. However, based on the results of Mrotek et al. [40] and Dolmaire et al. [39] the mass
diffusivity can be approximated as 10-12-10 - 13 m2 s- 1. As a result the Lewis number is much
greater than unity (Le -104-10 5 ) and corroborates our experimental observations that the thermal
response is quicker than the aqueous response. The discrepancy in the ratio of experimentally
observed response times and the ratio of thermal diffusivity to mass diffusivity exists because the

178

diffusion of water is not Fickian. There are strong interactions between the polyurethane and the
water, i.e. swelling and/or plasticization, and the most noticeable polymer chain relaxation takes
place after the maximum water uptake is achieved [40].
The morphological change of the beads on string morphology after aqueous stimulation and
drying is dramatic as is depicted in Figure 7-4c. The average diameter of the strings increases
almost sixfold to 810 + 330 nm, at which point it becomes difficult to distinguish the location of
beads. The scanning electron micrographs in Figure 7-4 indicate that the stretch applied during
electrospinning is 'frozen' in the fibrous ligaments and not in the beads, because there is no
observable change in bead dimension after thermal or aqueous stimulation. The polymer chains
within the beads are completely relaxed.
It is apparent when comparing the morphologies depicted in the series of micrographs that as the
concentration of polyurethane in solution increases the magnitude and variation of diameter of
the strings or fibers also increases as. This trend is emphasized in Figure 7-7a. The extent of
stretch recovered laterally within the mat correlates inversely with diameter as shown in Figure
7-7b. The greater the average diameters of the fibers that compose the mat, the smaller the
average lateral shrink, irrespective of the mode of stimulation.

4,1

M
I

The"
t I, I '

0.6

3.6

Aqueous Actuation

3.2
Low Q; Low

2
E 1

0.2
0.5

xC) 0.

'U
S2.4

2,
.,

P1B

tow Q; LwE-

01,

S0.2
I

0.1

0.I

oe

_T1

Low Q; Low E

12

I, I,
Is

I
20

, ,_
24

Polyurethane Concentration, wt%

0.0
28

1A

16

10

&-

4,

i;
8

Polyurethane Concentration, wt%

Figure 7-7: Influence of polyurethane concentration in solution on fiber diameter (a) and the extent of strain
recovered after actuation (b & c). A photo of the as spun, non-woven mat prior to (c: top) and after aqueous
actuation (bottom: C = 17 wt%; high Q; high E) visually highlights the extent of lateral shrink achieved.

179

The beads on a string mats produced from 17 wt% polyurethane in solution exhibit the greatest
lateral shrink of -50% when thermally actuated and -30% when aqueously actuated. The lateral
shrink recovered via aqueous actuation was always less than that recovered via thermal
actuation. This difference occurs because the swelling upon exposure to water increases the fiber
diameter to such an extent that the swollen fibers interfere with neighboring ones and inhibit
further shrinkage. However, upon removal of all the water absorbed in the aqueously actuated
samples via heating or drying (the samples are left in a case with desiccant for 48 hours) the mats
shrink further and approach the shrink values obtained after thermal actuation. It is likely that the
stretched ligaments in the 13 wt% based mat are stretched further; however, its recovery is not
measurable because the mat is so fragile that a continuous film-like sample could not be peeled
away from the collector plate and foil. All of the other mats maintain their mechanical integrity
prior to, during and after shape memory actuation. The flexural storage modulus determined via
dynamic mechanical analysis of one mat (C=17 wt%; high E; high Q) at these various stages of
actuation is depicted in Figure 7-8. Note that each mat has a modulus on the order of E-10 MPa
without accounting for the high porosity at temperatures below 30'C.
1U

103

O.

2 101
LU
101

in

-100

-50

50

100

150

Temperature (OC)
Figure 7-8: Flexural storage moduli of the BOAS electrospun mat (C=17 wt%; high Q; high E) at various
stages of actuation. Note that the modulus values are based on the initial dimensions of the film-like samples
and that it does not account for the porosity of the non-woven mat. Furthermore, the instrument does not
record values during the transition region when the mat is continuing to shrink and recover entropy because
an equilibrium has not been achieved.

When the temperature is close to the transition temperature the DMA, a load controlled
instrument, has difficulty tracking the length of the specimen and determining the load needed to
180

keep the specimen from buckling. This difficulty results because the instrument is trying to
oscillate the specimen in tension while the mat is thermally softening and attempting to recover
more entropy and shrink. Finally, the temperature becomes high enough that the material just
flows as a viscoelastic liquid.
Thin fibers recover more intrinsic strain than larger diameter fibers as is confirmed by the trends
portrayed in Figure 7-7. Although the 27 wt% based sample appears to contain an equal ratio of
large and small diameter fibers the mat exhibits the smallest lateral shrink of-10% after thermal
stimulation. The overall mat shrink is diminished because the thicker fibers are effectively stiffer
than the thin fibers and restrict the extent of strain the thin fibers may recover. From the inverse
correlation of lateral shrink with fiber diameter we can infer that the strings of the BOAS fiber
morphology are stretched further than the larger diameter, smooth fibers. Yet, the extent of
stretch applied via electrospinning can not be quantified without a more controlled study and
collection of monodisperse fibers.
6

.ia

3ii

0
2

0
I

0.01

0.1

10

0.01

100

0.1

Actuation
Aqueous ACtUaton

10

100

Pore Size Diameter (pm)

Pore Size Diameter (pm)

Average Diameter
130 nm
260 nm

Diameter
do
20 do

ulu nm

6.2 do

Length
Lo
0,32 Lo

A 13I o
v.v

Estimated Pore Size


Do
0 32 Do
An A
. vo

Figure 7-9: Resultant change in pore size distribution within the pure (a; C = 17 wt%; high Q; high E) and
Laponite composite (b), 'beads on a string' morphology, non-woven mats prior to and after thermal and
aqueous actuation. c) Cartoon exhibiting expected pore size change, assuming affine deformation and that
there are no beads.

181

The dramatic macroscopic change in mat shape shown in Figure 7-7c and microscopic change in
fiber morphology depicted in Figure 7-4 should be accompanied by a significant reduction in
pore size as shown in the cartoon in Figure 7-9c. As the results of mercury porosimetry indicate
in Figures 7-8a and 7-8b, there is significant reduction in the pore size distribution after aqueous
and thermal stimulation in the 17 wt% based solutions (with high Q and high E or with 4 wt%
Laponite). The estimated shrink, assuming affine deformation, are not achieved. This is likely
due to the presence of the beads which interfere and hinder further strain recovery. However,
virtually all pores greater than the bead diameter prior to actuation disappear after stimulus. If the
beads could be made smaller or the bead frequency decreased by utilization of nanoparticles,
conducting salts or alteration in spinning parameters, it is possible to tailor a non-woven mat to
have pores of a specific size. This technology could be useful in filtration and protective textile
applications [11].

7.3.4 Evidence of Reversible Surface Energy


It is possible to exploit the change in macroscopic mat shrink and microscopic fiber topology and
spacing to develop a fabric with reversible wettability [13, 41, 42]. Tuteja and coworkers[13]
show that it is possible to design superoleophobic surfaces using rough, electrospun surfaces
created from PMMA/FluoroPOSS blends. FluoroPOSS or 1H,1H,2H,2H-heptadecafluorodecyl
polyhedral oligomeric silsesquioxane (fluorodecyl POSS) is a very low surface energy
nanoparticle due to the presence and arrangement of -CF 2 and -CF

groups. By exploiting the

ability of these mats to recover strain and shrink it is possible to develop a fabric that changes its
wettability. On its own, each of these mats is wetted by water (7, = 72.1 mN m -1') and
hexadecane (Tvy,
= 27.5 mN m -') irregardless the fiber morphology. This result is not surprising
since similar polyurethanes have been reported to have a solid-vapor surface energy, yV, of
approximately 40 mN m- 1 [43] and as mentioned previously, this polyurethane absorbs water
(due to the presence of ethylene oxide within the polyurethane). Consequently, a low surface
energy coating was needed. A solution of a 50/50 blend of poly(methyl methacrylate) and
fluorodecyl POSS in Asahiklin AK 255 was provided by A. Tuteja, W. Choi, and S. Chattre. The
mats described above could be dip-coated in this solution. Note the maximum advancing contact
angle on a spin-coated surface consisting of a fluorodecyl POSS/poly(methyl methacrylate)

182

blend (44 wt% fluorodecyl POSS) is 800 [13]. After coating, the mats still were wetted by water,
although not instantaneously as was observed on the untreated mats. This wetting was likely due
to incomplete coverage of the fibers with the FluoroPOSS coating and the presence of large gaps
between fibers. The affinity of ethylene oxide for water caused the droplet to be locally pinned to
the mat and the water was absorbed by the polyurethane. However, the lower energy hexadecane
was non-wetting as shown in Figure 7-10a. The mat (C=20 wt%) is rough enough and the
spacing between fibers great enough so that the solid area fraction is small and a Cassie state
exists [41]. Upon thermally actuating the mat the fibers grow in diameter and the pores shrink.

U.

8 'U

Mat Treatment

As Spun

Heat Shrunk

Heat Shrunk & Stretched

Heat Shrunk Again

Average Diameter (jtm)

1.4 + 0.4

1.5 0. 4

0.97+ 0.3

1.4 + 0.4

verage Spacing (tm)

12 + 4

9.6 3

5.6 + 2

14 + 6

32
Fully Wetted
42
45
0 (degrees)
Figure 7-10: Scanning electron micrographs of the pure, non-woven, electrospun mat (C = 20 wt%) coated
with a 50/50 blend of fluorinated POSS & PMMA in the relaxed, free-standing, as spun configuration (a: L =
L0 ), in the thermally actuated and shrunk state (b: L = 0.84LO), the re-stretched state (c: L = 2.30L 0 ) and the
again shrunk or recovered state (d: L = 1.05LO) and the corresponding static contact angle of a drop of
hexadecane. Note the inset in c depicts the anisotropic or elliptical wetting pattern of the hexadecane on the
stretched mat. The average fiber diameter and spacing as well as the roll-off-angle, AO, for each surface is
reported.

Consequently, the solid area fraction grows and the spacing between fibers decreases and the
hexadecane wets the surface as shown in Figure 7-10b. Upon manually stretching the mat in
uniaxial tension 2.3 times its original length, the fibers become aligned in the stretch direction
and closely spaced. The hexadecane then wets the mat anisotropically, leaving an elliptic residue
with the long axis aligned with the direction of stretch shown in Figure 7-10c. Upon thermal
actuation a second time, the mat shrinks once again, recovering 95% of its original shape.
183

Consequently, Figure 7-10d provides evidence that the spacing between fibers is somewhat
larger than that of the relaxed, as spun configuration and the hexadecane becomes slightly more
non-wetting. The contact angle hysteresis of these materials can be measured using a goniometer
with a tilt stage. The roll-off-angle, AO, or tilt-angle when the droplet begins to roll along the
surface is equal to the difference between the advancing and receding contact angles,
AO = 0, - 0 and quantifies the contact angle hysteresis of the surface [44]. The resultant contact
angle hysteresis or roll-off-angles are reported in Figure 7-10. These surfaces exhibit significant
hysteresis and the stretched specimen wets fully in less than 10 s making it difficult to quantify
the anisotropic wetting. Ultimately, these coated, electrospun, shape memory mats demonstrate
that fabrics with reversible surface energies can be achieved. With more careful control of the
spinning conditions and fiber morphology and monitoring of the influence of uniaxial or biaxial
stretch it may be possible to develop highly tunable, porous fabrics that could be used as
actuatable filters and bandages or even as packaging, allowing the substance inside to wet when
desired.

Switchable surfaces, specifically surfaces with switchable hydrophobicity and anti-fouling


character are of particular interest in marine and biomedical industries: e.g. as coatings for ship
hulls and organ implants or as bio-sensors for drug delivery [45-47]. Genzer has shown that by
adding perfluoroalkanes at specific loacations along a stretched polydimethylsiloxane film the
contact angle of water will increase from 1000 to 130' when the film is released and allowed to
elastically recover its original shape [47]. This is an example of mechanical switching. Switching
can also be induced via an electrochemical stimulus, a change in environment i.e. thermal
stimulus, or photo-stimulus as reviewed by Gras et al in [46]. This is a new research area and
much work is still needed to develop switchable surfaces; specifically optimization of stimuli
magnitude with resulting response is needed. Surfaces must also be tuned and the extent of their
reversibility and stability quantified before devices that utilize this capability can be designed
[46].

It should be noted that by adding magnetite nanoparticles to the electrospun mat it is possible to
stimulate shrinkage without contact, via inductive heating. Numerous researchers have shown
that by adding magnetite nanoparticles to polymers or solutions and placing the resulting

184

composite or solution in an alternating magnetic field it is possible to heat the particles and the
surrounding polymer by exploiting N6el relaxation or Brownian motion depending on the size of
the particles and the viscosity of the matrix background [18, 48-50]. N6el relaxation arises from
the rotation in direction of the magnetic vector of a magnetic particle with alternating external
field and is the main contribution toward heating in solid polymeric systems [51]. It is possible to
disperse magnetite nanoparticles and this shape memory polyurethane in a solution containing
tetrahydrofuran with the aid of agitation; however, electrospinning the solution proved
unsuccessful. The slow flow rate in the syringe allowed enough time for the dense magnetite
particles (Pmag = 4.8-5.1 g cm-3 [49]) to precipitate to the bottom of the syringe prior to ejection.
As a consequence, the resultant electrospun mat, spun in high humidity conditions, had large
globules packed with magnetite particles atop a thick section of thin fibers with few magnetite
particles (see Figure 7-11 below). If a rotating magnet was utilized around the syringe containing
solution it may be possible to obtain electrospun, shape memory-magnetite composite mats that
shrink when exposed to a third stimulus-an alternating electric field; however, time did not
allow for additional experiments in such a system.

Figure 7-11: Comparison of the top and bottom side of an electrospun polyurethane mat filled with 10 wt%
magnetite. The top side appears to be brown due to the presence of magnetite particles while the bottom side
appears white as is typical of a pure polyurethane mat. Scanning electron micrographs show that the top side
is composed of large, magnetite-filled globules while the bottom side is composed of pure polyurethane beads
and thin strings.

7.4 Conclusions
By electrospinning a shape memory polymer, in this case polyurethane, it is possible to develop
a porous, non-woven mat that will shrink and recover via thermal actuation the intrinsic stretch
'frozen' into the fibers during the spinning process. By adjusting the spinning parameters, the

185

fiber morphology in the non-woven mat may be tuned. Specifically, by decreasing the polymer
concentration in solution it is possible to progressively transition from mats comprised of smooth
continuous fibers to fibers that resemble beads on a string. The diameter of the strings in this
morphology is significantly smaller than the diameter of the continuous fibers. The intrinsic
stretch 'frozen'

in the fibers is inversely proportional to the string or fiber diameter.

Consequently, a non-woven mat composed of beads on a string morphology exhibits greater


shrinkage than a mat composed primarily of larger diameter, smooth fibers.

Furthermore, we have demonstrated that when the shape memory polymer is a multi-block
copolymer containing a very hydrophilic component such as ethylene oxide within its
transitioning segment (or softer phase) then aqueous exposure can also be used as another
stimulus inducing entropic recovery. However, the response to aqueous stimulation is
approximately one hundred times slower than the response to thermal actuation. We have also
demonstrated that the addition of nanoparticles to the shape memory polymer allows for further
tuning of fiber morphology. Irregardless of the method of actuation, we have shown that the
significant entropic contraction, coupled with the beads on a string fiber morphology, provides a
method to tailor the pore size distribution within the electrospun mat and offers an avenue for the
production of selective filters or fabrics with reversible wettability. Ultimately, an electrospun,
non-woven, fibrous mat such as the one presented here, that exhibits one-way response to
multiple stimuli has significant potential in numerous applications, including drug delivery, food
packaging, and filtration.

7.5 References
[1]
[2]
[3]
[4]
[5]
[6]

S. V. Fridrikh, J. H. Yu, M. P. Brenner, and G. C. Rutledge, "Controlling the fiber


diameter during electrospinning," PhysicalReview Letters, vol. 90, 2003.
G. C. Rutledge and S. V. Fridrikh, "Formation of fibers by electrospinning," Advanced
Drug Delivery Reviews, vol. 59, pp. 1384-1391, 2007.
J. H. Yu, S. V. Fridrikh, and G. C. Rutledge, "The role of elasticity in the formation of
electrospun fibers," Polymer, vol. 47, pp. 4789-4797, 2006.
C. Liu, H. Qin, and P. T. Mather, "Review of progress in shape-memory polymers,"
JournalOfMaterials Chemistry, vol. 17, pp. 1543-1558, 2007.
A. Lendlein and S. Kelch, "Shape-memory polymers," Angewandte Chemie-International
Edition, vol. 41, pp. 2034-2057, 2002.
K & B Innovations, "The Magical World of Shrinky
Dinks,"http://www.shrinkydinks.com, 2007: Accessed July 5th, 2008.

186

[7]
[8]
[9]

[10]

[11]
[12]

[13]

[14]
[15]

[16]

[17]

[18]

[19]
[20]

[21]
[22]

[23]

[24]

J. D. Ferry, Viscoelastic PropertiesofPolymers, 3rd ed. New York, NY: John Wiley &
Sons, Inc., 1980.
J. P. Flory, Principles ofPolymer Chemistry. Ithaca, NY: Cornell University Press, 1953.
D. Huang, Y. Yang, G. Zhuang, and B. Li, "Influence of entanglements on the glass
transition and structural relaxation behaviors of macromolecules: 1. Polycarbonate,"
Macromolecules, vol. 32, pp. 6675-6678, 1999.
S. J. Hong, W. Yu, J. H. Youk, and Y. R. Cho, "Polyurethane smart fiber and shape
memory function: experimental characterization and constitutive modelling," FibersAnd
Polymers, vol. 8, pp. 377-385, 2007.
R. S. Barhate and S. Ramakrishna, "Nanofibrous filtering media: Filtration problems and
solutions from tiny materials," Journalof Membrane Science, vol. 296, pp. 1-8, 2007.
M. L. Ma, Y. Mao, M. Gupta, K. K. Gleason, and G. C. Rutledge, "Superhydrophobic
fabrics produced by electrospinning and chemical vapor deposition," Macromolecules,
vol. 38, pp. 9742-9748, 2005.
A. Tuteja, W. Choi, M. L. Ma, J. M. Mabry, S. A. Mazzella, G. C. Rutledge, G. H.
McKinley, and R. E. Cohen, "Designing superoleophobic surfaces," Science, vol. 318,
pp. 1618-1622, 2007.
M. Yoshida, R. Langer, A. Lendlein, and J. Lahann, "From advanced biomedical coatings
to multi-functionalized biomaterials," Polymer Reviews, vol. 46, pp. 347-375, 2006.
X. Jin and Y. L. Hsieh, "pH-responsive swelling behavior of poly(vinyl
alcohol)/poly(acrylic acid) bi-component fibrous hydrogel membranes," Polymer, vol.
46, pp. 5149-5160, 2005.
K. J. Pawlowski, H. L. Belvin, D. L. Raney, J. Su, J. S. Harrison, and E.J. Siochi,
"Electrospinning of a micro-air vehicle wing skin," Polymer, vol. 44, pp. 1309-1314,
2003.
R. Samatham, I. Park, K. J. Kim, J. Nam, N. Whisman, and J. Adams, "Electrospun
nanoscale polyacrylonitrile artificial muscle," Smart Materials & Structures, vol. 15, pp.
N152-N156, 2006.
P. R. Buckley, G. H. McKinley, T. S. Wilson, W. Small, W. J. Benett, J. P. Bearinger, M.
W. McElfresh, and D. J. Maitland, "Inductively heated shape memory polymer for the
magnetic actuation of medical devices," Ieee TransactionsOn BiomedicalEngineering,
vol. 53, pp. 2075-2083, 2006.
S.Chaterji, I. K. Kwon, and K. Park, "Smart polymeric gels: Redefining the limits of
biomedical devices," ProgressIn Polymer Science, vol. 32, pp. 1083-1122, 2007.
C. M. Yakacki, R. Shandas, C. Lanning, B. Rech, A. Eckstein, and K. Gall,
"Unconstrained recovery characterization of shape-memory polymer networks for
cardiovascular applications," Biomaterials,vol. 28, pp. 2255-2263, 2007.
M. Martina and D. W. Hutmacher, "Biodegradable polymers applied in tissue
engineering research: a review," Polymer International,vol. 56, pp. 145-157, 2007.
K. Gall, C. M. Yakacki, Y. P. Liu, R. Shandas, N. Willett, and K. S. Anseth,
"Thermomechanics of the shape memory effect in polymers for biomedical applications,"
JournalOfBiomedical MaterialsResearch PartA, vol. 73A, pp. 339-348, 2005.
G. Baer, T. S. Wilson, D. L. Matthews, and D. J. Maitland, "Shape-memory behavior of
thermally stimulated polyurethane for medical applications," Journal OfApplied Polymer
Science, vol. 103, pp. 3882-3892, 2007.
H. T. Zhuo, J. H. Hu, S. J. Chen, and L. P. Yeung, "Preparation of polyurethane

187

[25]
[26]

[27]
[28]

[29]

[30]
[31]
[32]

[33]
[34]

[35]
[36]

[37]

[38]
[39]

[40]
[41]
[42]

nanofibers by electrospinning," Journal OfApplied Polymer Science, vol. 109, pp. 406411, 2008.
H. T. Zhuo, J. L. Hu,and S. J. Chen, "Electrospun polyurethane nanofibres having shape
memory effect," Materials Letters, vol. 62, pp. 2078-2080, 2008.
D. I. Cha, H. Y. Kim, K. H. Lee, Y. C. Jung, J. W. Cho, and B. C. Chun, "Electrospun
nonwovens of shape-memory polyurethane block copolymers," Journal Of Applied
Polymer Science, vol. 96, pp. 460-465, 2005.
M. Y. Razzaq and L. Frormann, "Thermomechanical studies of aluminum nitride filled
shape memory polymer composites," Polymer Composites, vol. 28, pp. 287-293, 2007.
S.M. Liff, N. Kumar, and G. H. McKinley, "High-performance elastomeric
nanocomposites via solvent-exchange processing," Nature Materials, vol. 6, pp. 76-83,
2007.
Y. M. Shin, M. M. Hohman, M. P. Brenner, and G. C. Rutledge, "Experimental
characterization of electrospinning: the electrically forced jet and instabilities," Polymer,
vol. 42, pp. 9955-9967, 2001.
J. R. MacCallum and C. A. Vincent, "Polymer Electrolyte Reviews," 1st ed. London:
Elsevier Applied Science, 1987, pp. 368.
E. W. Washburn, "The dynamics of capillary flow," The PhysicalReview, vol. 17, pp.
273-283, 1921.
E. W. Washburn, "Note on a method of determining the distribution of pore sizes in a
porous materials," Proceedingsof the NationalAcademy of Sciences of the United States
ofAmerica, vol. 7, pp. 115-116, 1921.
Knovel, "Knovel Critical Tables: Basic Physical Properties of Chemical Compounds,"
Knovel Corporation, 2008.
V. K. Daga and N. J. Wagner, "Linear viscoelastic master curves of neat and laponitefilled poly(ethylene oxide)-water solutions," RheologicaActa, vol. 45, pp. 813-824,
2006.
E. Guth, "Theory of Filler Reinforcement," Journalof Applied Physics, vol. 16, pp. 2025, 1945.
B. Yang, W. M. Huang, C. Li, and J. H. Chor, "Effects of moisture on the glass transition
temperature of polyurethane shape memory polymer filled with nano-carbon powder,"
European Polymer Journal, vol. 41, pp. 1123-1128, 2005.
A. Laforgue, L. Robitaille, A. Mokrini, and A. Ajji, "Fabrication and characterization of
ionic conducting nanofibers," MacromolecularMaterialsAnd Engineering,vol. 292, pp.
1229-1236, 2007.
G. R. Lomax, "Breathable polyurethane membranes for textile and related industries,"
JournalOfMaterials Chemistry, vol. 17, pp. 2775-2784, 2007.
N. Dolmaire, E. Espuche, F. Mechin, and J.-P. Pascault, "Water transport properties of
thermoplastic polyurethane films," JournalofPolymer Science: PartB: Polymer Physics,
vol. 42, pp. 473-492, 2004.
J. L. Mrotek, M. J. Matthewson, and C. R. Kurkjian, "Diffusion of moisture through
optical fiber coatings," Journal OfLightwave Technology, vol. 19, pp. 988-993, 2001.
A. B. D. Cassie and S. Baxter, "Wettability of porous surfaces," Transactionsof the
Faraday Society, vol. 40, pp. 546-551, 1944.
R. N. Wenzel, "Resistance of solid surfaces to wetting by water," Industrialand
EngineeringChemistry, vol. 28, pp. 988-994, 1936.

188

[43]

[44]
[45]
[46]
[47]
[48]

[49]

[50]

[51]

M. Lupu, D. Macocinschi, G. Ioanid, M. Butnaru, and S. Ioan, "Surface tension of


poly(ester urethane)s and poly(ether urethane)s," Polymer International,vol. 56, pp. 389398, 2007.
G. McHale, N. J. Shirtcliffe, and M. I. Newton, "Contact-angle hysteresis on superhydrophobic surfaces," Langmuir, vol. 20, pp. 10146-10149, 2004.
S. Krishnan, C. J. Weinman, and C. K. Ober, "Advances in polymers for anti-biofouling
surfaces," Journal Of Materials Chemistry, vol. 18, pp. 3405-3413, 2008.
S. L. Gras, T. Mahmud, G. Rosengarten, A. Mitchell, and K. Kalantar-Zadeh, "Intelligent
control of surface hydrophobicity," Chemphyschem, vol. 8, pp. 2036-2050, 2007.
J. Genzer and K. Efimenko, "Recent developments in superhydrophobic surfaces and
their relevance to marine fouling: a review," Biofouling, vol. 22, pp. 339-360, 2006.
V. S. Kalambur, B. Han, B. E. Hammer, T. W. Shield, and J. C. Bischof, "In vitro
characterization of movement, heating and visualization of magnetic nanoparticles for
biomedical applications," Nanotechnology, vol. 16, pp. 1221-1233, 2005.
R. Mohr, K. Kratz, T. Weigel, M. Lucka-Gabor, M. Moneke, and A. Lendlein, "Initiation
of shape-memory effect by inductive heating of magnetic nanoparticles in thermoplastic
polymers," Proceedings Of The NationalAcademy Of Sciences Of The United States Of
America, vol. 103, pp. 3540-3545, 2006.
K. Okawa, M. Sekine, M. Maeda, M. Tada, and M. Abe, "Heating ability of magnetite
nanobeads with various sizes for manetic hyperthermia at 120kHz, a noninvasive
frequency," JournalofApplied Physics, vol. 99, pp. H102, 2006.
R. E. Rosensweig, Ferrohydro-dynamics.Mineola, NY: Dover Publications, Inc., 1985.

189

8. Concluding Remarks
8.1 Summary of Accomplishments
This thesis demonstrates that the use of block polarity in conjunction with the solvent exchange
approach described here offers an avenue toward the development of high performance materials
that rival natural materials. While this solvent exchange dispersal technique may be applied to
numerous nanoparticle-polymer systems, the extent of property enhancement depends on the
time allowed for phase-segregation and preferential reinforcement. Unfortunately, the aspect
ratio of the nanoparticles employed significantly limits the concentration envelope that may be
utilized so that matrix polymer dissolution is achieved. Ultimately, the success of this approach
hinges upon utilization of dilute solutions in conjunction with slow processing which encourages
optimal property enhancement.

Using the solvent exchange technique described, it is possible to preferentially nanoreinforce the
hard micro-domains of thermoplastic elastomers with smectic clay that has characteristic
dimensions similar to the hard segment length. The adhesion between the clay and the hard
micro-domains coupled with the formation of a percolative network not only stiffens and
toughens, but increases the heat distortion temperature (HDT) of a commercial poly(ether
urethane). The discotic clay platelets induce morphological ordering over a range of length scales
that results in significant thermomechanical enhancement and expands high temperature
applications.

The role of the nano-clay reinforcement in polyurethane, however, evolves with deformation,
specifically with chain stretch and orientation. ATR-FTIR results indicate that the addition of
clay does not reduce the hard segment inter-urethane hydrogen bonding and order, nor does it
reduce the hydrogen bonding between the hard segment and soft segment. The long processing
time utilized via slow solvent casting allows for self-assembly and complete microphase
separation irregardless of Laponite concentration. Because Laponite does not influence
microphase separation the equilibrium or hyperelastic response of the polyurethane matrix is
unaffected by the addition of nano-clay. The preferential association of Laponite to the hard

190

micro-domains imposes spatial restrictions and/or constraints on the hard segments. The hard
micro-domain/Laponite structures act like rigid fillers around which the soft chains must stretch.
Consequently, the composite develops anisotropy and the matrix undergoes irreparable damage.
These thermoplastic polyurethane nano-clay composites demonstrate strong hysteresis, time
dependence, and cyclic softening. The pure and Laponite filled polyurethanes exhibit two
characteristic relaxation times. The long relaxation time corresponds to the hard micro-domains
and remains constant with deformation. The short relaxation time corresponds to the soft
segment. As the magnitude of the applied stretch increases the soft segments stretch, align and
orient and the relaxation time increases. Ultimately, nano-clay reinforces the stress response of
Elasthane, but also locally damages the polyurethane matrix, especially when concentrations
above the percolation threshold are utilized. Laponite does not influence the hyperelastic or
equilibrium response of the material-that is matrix dependent. Instead, the nano-clay influences
the non-equilibrium response-the rate sensitivity and time dependence-much like rigid
particles influence the viscosity of dilute solutions.

To exploit these thermo-physical interactions, nano-clay was dispersed within commerciallyavailable shape memory polyurethane in order to enhance the restoring or actuation stress. While
adding nano-reinforcement, specifically at concentrations above the percolation threshold, can
significantly enhance the modulus of the material at temperatures above the glass transition there
is trade-offs. The rigid reinforcement constrains the soft, entropy recovering segments restricting
the chain mobility. A small change in glass transition temperature is observed with nano-clay
reinforcement, while the actuated strain recovery is delayed as the filler concentration increases.
Furthermore, as the nano-clay concentration increases, irreparable damage to the polyurethane
matrix upon deformation is unavoidable. Consequently, the extent of strain recovered
deteriorates, particularly at concentrations above the percolation limit. So while the actuation or
recovery stress can be enhanced with the addition of exfoliated nano-clay, the percolation
threshold denotes the critical point, beyond which the strain recovered significantly decreases.
By electrospinning a shape memory polyurethane, it is possible to develop a porous, non-woven
mat that will shrink and recover the intrinsic stretch 'frozen' in the fibers during the spinning

191

process via thermal actuation. Aqueous actuation may also be utilized if the multi-block
copolymer contains a very hydrophilic component such as ethylene oxide within its transitioning
segment; however, the response time is approximately a hundred times slower. The resulting
string (of a 'beads on a string' fiber morphology) or fiber diameter is inversely proportional to
the intrinsic stretch 'frozen' in the fibers. Consequently, a 'beads on a string' fiber morphology
exhibits greater shrinkage than a mat composed primarily of large diameter, smooth fibers. The
addition of nano-clay to the shape memory polyurethane solution allows for further tuning of
fiber morphology by increasing solution conductivity and relaxation time, while fluorinated
coatings offer a means to tailor the resultant surface energy of the fibrous mat. This control
coupled with significant entropic shrink of the porous material provides a method to tailor the
pore size distribution within the electrospun mat and offers an avenue for the production of
selective filters or fabrics with reversible wettability. Ultimately, an electrospun, non-woven,
fibrous mat that exhibits one-way response to multiple stimuli has significant potential in
numerous applications, including drug delivery, food packaging, and filtration.

8.2 Future Work


Quantifying and controlling the thermo-physical interactions between a block-copolymer with
polar segments (e.g. thermoplastic polyurethane) and inorganic nanoparticles (e.g. nano-clay) are
important for future nanocomposite processing strategies: the efficacy of nanoreinforcement
hinges upon the close matching of characteristic length scale and the adhesion of the
nanoparticles to the targeted polymer phase morphology. Exploiting block polarity in
conjunction with the solvent exchange approach developed in this thesis and processing
techniques, such as electro-spinning, offers an avenue toward the development of high
performance, hierarchically-ordered materials that rival natural materials. Further interrogation
of the thermo-physical interactions of nano-clay and polyurethane, particularly at varying
temperature and high strain rate, is of interest to the various military branches since polyurethane
coatings have proven to mitigate damage due to blast [1]. Characterization of other inorganic
nanoparticle interactions with polyurethane is of interest and necessary for continued
development of high performance, multifunctional composites.
8.2.1 Scale-Up of Time Efficient, Environmentally-Friendly, Solvent Exchange Approach
While this solvent exchange dispersal technique discussed in this thesis proved successful, and
may be applied to numerous nanoparticle-polymer systems, the extent of property enhancement

192

depends upon utilization of dilute solutions in conjunction with slow processing for optimal
enhancement. Both requirements, dilute solutions and time, are expensive and prevent this
solvent exchange dispersal technique from being commercially viable, despite the interest of
companies that produce cable sheathing, luggage, and women's active wear and intimate apparel
for preferentially nano-reinforced polymer composites. Furthermore, unless the solvent utilized
is water, or another

naturally

occurring liquid,

rather than

dimethylformamide

or

dimethylacetamide which were utilized in this thesis, the environmental impact of the solvent
must be considered-particularly when it is needed in such large quantities. Consequently, scaleup of this dispersal technique to a more time-efficient, environmentally conscious method could
prove beneficial.
One such method for scale-up, includes identification and evaluation of a naturally occurring
"solvent" that would disperse nanoparticles but could be mixed and remain within the polymer
matrix as an "additive" after processing. DI-a-tocopherol, otherwise known as vitamin E (a
naturally occurring antioxidant), could prove to be a viable option to act as both solvent and
additive. Schubert et al have shown that the addition of 5 wt% vitamin E improves the
biocompatibility and biostability of poly(ether urethane urea) elastomers in vivo with only a
small amount of plasticization as shown in Figure 8-1 [2].
40
35
30
25
20
15
10
5
n
0

50

100 150 200 250 300 350 400 450 500 550 600 650 70
%Strain

Figure 8-1: Stress-strain curves of poly(etherurethane urea) (PEUU-A) and that

filled with 5 wt% vitamin E (PEUU-VE) [2].

193

It is possible that the slight plasticization caused by the addition of vitamin E could be negated
be utilized
by the inclusion of nanoreinforcement. Another difficulty that arises should vitamin E
is that it is quite viscous at room temperature. Consequently, dispersal of nanoparticles like
Laponite would require mixing the materials at elevated temperatures. Constant shear stress tests,
-1
when the shear stress equals 15 Pa, and constant shear rate tests, when the shear rate equals 1 s
ramped from 25'C to 100 0 C at 2.5 0 C min -' using TA Instruments AR 2000,
indicate that viable mixing temperatures would be greater than 70'C, shown in Figure 8-2.
at temperatures

2.5

m 2.0

> 1.0

0.5

nl
20

30

40

50

60

70

80

90

100

Temperature (oC)

Figure 8-2: Comparison of the viscosity of DI-a-tocopherol (Vitamin E) as a function of temperature


when constant shear stress or constant shear rate is applied.

Ultimately, more research is necessary to determine whether vitamin E is a viable


environmentally friendly solvent and polymeric additive and to identify other such materials.
8.2.2 Deformation Induced Orientation and Damage by Nano-Clay

This thesis makes clear that the role of the nano-clay reinforcement changes as the polyurethane
composite is deformed and the polymer chains are aligned. However, it is not clear whether the
platelets resist rotation and orientation or exactly how, or to what extent, the platelets induce
irreparable damage. Nor are the interactions between the platelets and hard micro-domains

194

within the strain hardening deformation region completely understood. While initial attempts at
small angle x-ray scattering (SAXS) with in situ stretching proved unsuccessful because
electron-density matching blurred any distinguishing features in the Laponite-filled samples,
future attempts utilizing thicker specimens may allow the spacing between the Laponite platelets
and hard micro-domains to be interrogated. Song et al have completed a similar study in which
they found that within the elastic regime of the matrix material the d-spacing of the clay platelets
varied in a reversible manner, increasing upon deformation but recovering to the original spacing
upon removal of the imposed load [3]. Ultimately, the pairs of platelets with interspersed
polymer acted as tiny nano-springs. It would be interesting to evaluate the nano-clay d-spacing in
preferentially reinforced elastomeric polyurethanes. Quantifying the variation in the resultant
platelet d-spacing with respect to macro-scale deformation would allow the movement,
breakdown, and orientation of the jammed nano-clay network structure, as well as matrix
damage, to be connected to physical mechanisms within the various deformation regimes, i.e.
plastic yield, soft segment alignment, and strain-hardening. TEM images of the films stretched to
these various deformation regimes would also offer visual evidence supporting the structural
changes inferred through x-ray scattering and pin-point the mechanism behind irreparable matrix
damage.
Interrogation of these preferentially reinforced polyurethane composites at various temperatures
and high strain rates is also of particular interest to the United States military. Polyurethane
coatings have proven to mitigate damage due to blast or ballistic event time and again, yet the
mechanism(s) responsible for increased survivability are not entirely understood [1, 4]. In
different temperature or frequency regimes, the rate sensitivities of polymers change as various
molecular mobility mechanisms are stimulated. Nano-clay alters the morphology of the
polyurethane and as a consequence may alter the rate-dependent mechanical deformation and
failure behavior of the polyurethane.

Ultimately, understanding the correspondence between deformation and structural changes,


specifically the orientation of Laponite within the nanocomposites, may help engineers design
materials for specialty applications, such as ballistic protection or tear resistance in stockings.
The structural changes coupled with strain rate and temperature sensitivity data can then be used

195

to modify the constitutive model developed by Dr. Renaud Rinaldi (see Appendix A-3) to
account for the addition of nano-clay and other modes of deformation interrogated and
accounted. Progress in the utility of polymer composites, particularly polyurethane composites,
depends upon reliable material models that are easily applied by design engineers.
8.2.3 Utility of Magnetite in Shape Memory Polyurethane Composites
Characterization of other inorganic nanoparticle interactions with polyurethane is of interestspecifically that of magnetite nanoparticles with shape memory polyurethane. Actuating
biomedical devices, filters and protective coatings with switchable surface hydrophobicity are
currently of great interest [5-11]. Often, however, these shape memory polymers need to be
thermally stimulated in order to induce actuation. By adding magnetite nanoparticles it is
possible to induce actuation without contact via inductive heating. A few researchers have shown
that by adding magnetite nanoparticles to polymers or solutions and placing the resulting
composite or solution in an alternating magnetic field it is possible to heat the particles and the
surrounding matrix by exploiting Neel relaxation or Brownian motion (depending on the size of
the particles and the viscosity of thematrix background) [7, 12-14]. While Buckley et al and
Mohr et al have investigated the feasibility of actuating polyurethane via inductive heating, a
systematic study evaluating the thermo-physical interaction between true, non-coated,
nanoparticles and the polyurethane have not yet been evaluated and development of an efficient
dispersal technique is needed [7, 13].
It is possible to disperse commercially available magnetite nanoparticles (Ocean NanoTech),
measuring -15 nm in diameter within commercial shape memory polyurethane (Mitsubishi
Diaplex TM Series MM 5510 PU). Ocean NanoTech provides these nanoparticles with oleic acid
surface ligands dispersed in chloroform at a concentration of 5 mg m'-1 [15]. But the
polyurethane

is not soluble in chloroform. The chloroform can be exchanged with

tetrahydrofuran by first adding acetone to the magnetite-chloroform solution in a ratio of 3:1.


The acetone-magnetite-chloroform solution is then centrifuged at 3,000 rpm for fifteen minutes.
The supernatant is then discarded and the magnetite precipitate dispersed in tetrahydrofuran at a
concentration of 5 mg ml-1. Then polyurethane and pure tetrahydrofuran are added to the
magnetite-tetrahydrofuran solution so that 5 wt% polyurethane in solution resulted as well as a
concentration of 5 wt% magnetite in solid when the polyurethane-magnetite composite was

196

precipitated out using excess methanol. The precipitate is dried and then compression molded at
150'C. The dispersal of the magnetite nanoparticles within the compression molded films were
then characterized using transmission electron microscopy after lamellae were prepared using a
focused ion beam (discussed in Chapter 4). While the resultant film appears to be a
homogeneous rust brown color, instead of transparent as it typically is without magnetite filler,
the nanoparticle dispersal is not homogeneous. As is observed in Figure 8-3, while many of the
nanoparticles are well dispersed, there are numerous small flocculates of 20-30 nanoparticles and
a few huge agglomerates measuring 100-500 nm in diameter. Obviously, this dispersal is not
ideal. The method for dispersal may require improvement, but at the least provides a means to
prepare composites in which the thermo-physical interactions between magnetite particles and
polyurethane matrix may be evaluated.

197

P~ ix4
a

5--""

i~a. "
;-1
U:;

- .g

;;
i~i: 4

d~ :D

-iaf

?".

"r-sic,

yC
;x
~

,.i "s~

az

y ,
/:

I-

\,

Figure 8-3: Transmission electron micrographs of magnetite nanoparticles dispersed in a commercial


thermoplastic polyurethane (MM 5510) after precipitation and compression molding: scale bar 500nm in a
and 100 nm in b, c, and d.

Despite the inhomogeneous magnetite dispersal, the composite films still display paramagnetic
behavior. For instance, when a strong magnet is placed a couple inches above the composite
film, the film jumps off the countertop and magnetically bonds to the magnet. The magnetization
of the composite can be evaluated using a superconducting quantum interference device, the

198

results of which are shown in Figure 8-4. As the applied magnetic field approaches 1500 Oe the
magnetization of the composite saturates.
However, when no magnetic field is applied there is small remnant magnetization which decays
with increased temperature. This can be explained by the N'el relaxation which dominates and
varies exponentially with inverse temperature[ 12]. Consequently, when the applied field is zero,
the dipole moments of some particles remain polarized and result in a small remnant
magnetization, however, as the temperature increases the remnant magnetization decreases.
......

1.0x10

H=0 Oe

0.6 -

E 0.4 -E

9.0x10'

\000

O
E

0.2 --

S8.0x10

"

0.0O% \

0.0

r 7.0x10

-0.4

6.0x10

-I-0\

-0.6

\o0

-0.8

.4i

-1500 -1000

I
-500

I
0

I
500

i
1000

a ni4n

20

1500

I
40

I
60

I
80

I
100

Temperature (C)

Applied Magnetic Field, H (Oe)

Figure 8-4: Magnetization of commercial thermoplastic polyurethane DiaplexTM MM 5510 filled with 5 wt%
magnetite (Fe 3O4) as a function of applied magnetic field (left) and temperature when no magnetic field is
applied (right).

Currently nanoparticle dispersal, magnetic characterization and proof of actuation via inductive
heating are the extent to which most polymer-magnetite systems have been evaluated. It would
be useful to understand how the inclusion of magnetite in shape memory polymers, like the
polyurethane shown here, influences the response time, actuation stress, and strain recovery of
the composite so that these materials can be utilized in real problems, i.e. in drug delivery or as
stroke or aneurysm treatment [7].

199

8.3 References
[1]

[2]

[3]

[4]
[5]

[6]
[7]

[8]
[9]
[10]
[11]
[12]

[13]

[14]

[15]

J. Porter, R. Dinan, M. Hammons, and K. Knox, "Polymer coatings increase blast


resistance of existing and temporary structures," AMPTIAC Quarterly, vol. 6, pp. 47,
2003.
M. A. Schubert, M. J. Wiggins, K. M. DeFife, A. Hiltner, and J. M. Anderson, "Vitamin
E as an antioxidant for poly(etherurethane urea): In vivo studies," Journalof Biomedical
MaterialsResearch, vol. 32, pp. 493-504, 1996.
M. Song and K. J. Yao, "X-ray diffraction detection of compliance in polyurethaneorganoclay nanocomposites," MaterialsScience And Technology, vol. 20, pp. 989-992,
2004.
J. Yi, M. C. Boyce, G. F. Lee, and E. Balizer, "Large deformation rate-dependent stressstrain behavior of polyurea and polyurethanes," Polymer, vol. 47, pp. 319-329, 2006.
G. Baer, T. S. Wilson, D. L. Matthews, and D. J. Maitland, "Shape-memory behavior of
thermally stimulated polyurethane for medical applications," JournalOfApplied Polymer
Science, vol. 103, pp. 3882-3892, 2007.
R. S. Barhate and S. Ramakrishna, "Nanofibrous filtering media: Filtration problems and
solutions from tiny materials," Journalof Membrane Science, vol. 296, pp. 1-8, 2007.
P. R. Buckley, G. H. McKinley, T. S. Wilson, W. Small, W. J. Benett, J. P. Bearinger, M.
W. McElfresh, and D. J. Maitland, "Inductively heated shape memory polymer for the
magnetic actuation of medical devices," leee Transactions On Biomedical Engineering,
vol. 53, pp. 2075-2083, 2006.
M. Yoshida, R. Langer, A. Lendlein, and J. Lahann, "From advanced biomedical coatings
to multi-functionalized biomaterials," Polymer Reviews, vol. 46, pp. 347-375, 2006.
S. Krishnan, C. J. Weinman, and C. K. Ober, "Advances in polymers for anti-biofouling
surfaces," JournalOf Materials Chemistry, vol. 18, pp. 3405-3413, 2008.
S. L. Gras, T. Mahmud, G. Rosengarten, A. Mitchell, and K. Kalantar-Zadeh, "Intelligent
control of surface hydrophobicity," Chemphyschem, vol. 8, pp. 2036-2050, 2007.
J. Genzer and K. Efimenko, "Recent developments in superhydrophobic surfaces and
their relevance to marine fouling: a review," Biofouling, vol. 22, pp. 339-360, 2006.
V. S. Kalambur, B. Han, B. E. Hammer, T. W. Shield, and J. C. Bischof, "In vitro
characterization of movement, heating and visualization of magnetic nanoparticles for
biomedical applications," Nanotechnology, vol. 16, pp. 1221-1233, 2005.
R. Mohr, K. Kratz, T. Weigel, M. Lucka-Gabor, M. Moneke, and A. Lendlein, "Initiation
of shape-memory effect by inductive heating of magnetic nanoparticles in thermoplastic
polymers," ProceedingsOf The NationalAcademy Of Sciences Of The United States Of
America, vol. 103, pp. 3540-3545, 2006.
K. Okawa, M. Sekine, M. Maeda, M. Tada, and M. Abe, "Heating ability of magnetite
nanobeads with various sizes for manetic hyperthermia at 120kHz, a noninvasive
frequency," JournalofApplied Physics, vol. 99, pp. H102, 2006.
W. W. Yu, J. C. Falkner, C. T. Yavuz, and V. L. Colvin, "Synthesis of monodisperse iron
oxide nanocrystals by thermal decomposition of iron carboxylate salts," Chemical
Communications,pp. 2306-2307, 2004.

200

Appendix A-1-Movie Demonstrating the Mechanical


Enhancement of Elasthane at High Temperature by the
Addition of Laponite.

Movie titled Appendix A-1 displays thermal creep of nanocomposite films containing 0 and
20 wt% Laponite when a constant load equal to 1.08 N is applied. The pure Elasthane deforms
significantly, increasing its length by 100% at 95 0 C, and breaks at 120'C. Meanwhile the
nanocomposite containing 20 wt% Laponite does not measurably deform at all. However, at
125oC the improvised grip holding the constant load slips off the sample.

201

Appendix A-2-Elasthane/Laponite Composite Moduli


Measured with a Video Extensometer
The initial modulus reported in chapter 4 from the engineering stress-strain curves utilizing the
Zwick Z010 mechanical tester were verified using a video extensometer. The thin-film sample
was lit with a fiber optic light and videoed using a Qimaging Retiga 1300 CCD camera equipped
with a 200 mm Canon f14.0 Canon lens and a Nikon 200 mm extension. A series of X's were
placed along the length of the sample using a fine tip black permanent marker as shown below in
Figure A-2-1 and the specimen width and thickness were measured at the vertex of each X and
measured with a micrometer. The camera was placed at a distance of approximately 1 m from
the specimen and images were captured every 0.5 s. The images were analyzed with a digital
image correlation algorithm developed by Correlated Solutions Incorporated (CSI). The
deformed image was correlated to the reference image via two distinct points of interest, i.e. two
black dots, connected by a straight line. The software then output the x and y pixel position of
these two distinct points. The change in longitudinal distance was used to calculate the true strain
while the change in horizontal distance was used to calculate the true stress as shown below:

YUC=Ln Y22,0 -

Y1,0

(A-2.1)

and

"true

F(wt)

(WotF )

(wot)

x2 - XI
x

-x

(A-2.2)

when changes in thickness are assumed negligible, t

to. This video extensometer was used

because strain localization was common within these samples since the thickness of the
specimen varied +20 [pm through the length due to solvent casting.

202

Y2
X2

exi

0
0
.Yi

AND

Figure A-2-1: Images of a 2 wt% Laponite-filled polyurethane film inside the Zwick Z010 as seen with the
video extensometer. On the left is the un-deformed film with 45 mm gauge length. On the right is the same
film after the gauge length has been extended 30%. It is important to note that 40% of the un-deformed
sample is masked by the grip: recall the grip consists of a convex and flat jaw face so the upper and lower
gauge length extends into the grip and is unobservable.

It is important to note that the grips used in the Zwick consisted of one flat, polyurethane jaw
face placed flush against a convex, aluminum jaw face. Consequently, a portion at either end of
the specimen extends between the grips and up to the line of contact that holds the specimen
fixed. These grips were screw tightened and as a consequence localized deformation near the

203

grip contact lines was common. So the deformation in the middle of the various nanocomposites
was monitored.

Figure A-2-2: Selection of the sequential images taken with the video extensometer as
Elasthane filled with 2
wt% nano-clay is deformed using the Zwick Z010 mechanical tester at a nominal strain rate
of 0.0167s'.

Figure A-2-2 depicts the deformation captured by the video extensometer typical
of the
nanocomposites. The results of the localized deformation captured via the video extensometer
are plotted against the engineering results determined by the Zwick cross-head in Figure A-2-3.
The engineering stress-strain response within the elastic regime falls within the scattered
response of that calculated with the video extensometer. This behavior was true of the
more
highly concentrated samples as well. Consequently, the modulus measurements reported
can be
considered valid.

204

2.5

0" 2.0
()

1.5

1.0
C

0.5
NN

0.00

0.02

0.04

0.06

0.08

Nominal Strain
Figure A-2-3: Comparison of the engineering stress-strain response of 2 wt% nano-clay filled Elasthane
calculated from the change in length reported by the Zwick mechanical tester crosshead and that measured
via the video extensometer at various sections along the length of the specimen gauge length.

205

Appendix

A-3-Discussion

of Appropriate Constitutive

Model
This appendix reviews the constitutive model developed by Qi and Boyce to capture the stressstrain behavior of thermoplastic polyurethanes [1]. Within this appendix the ability of the model
to capture the hyperelastic equilibrium behavior of the high performance, Elasthane/Laponite
nanocomposites and its inability to accurately capture the non-equilibrium response is discussed.

A-3.1 Review of Qi & Boyce Thermoplastic Polyurethane Model


Qi and Boyce developed a constitutive model that predicts the stress-strain behavior of
thermoplastic polyurethanes well by coupling the mechanical behavior to the evolution of the
hard and soft domain morphology of the polyurethane matrix as strain increases [1]. The model
captures the nonlinear hyperelastic behavior, time-dependence, stress-softening, and hysteretic
behavior by virtue of a hyperelastic spring and visco-plastic Maxwell element. This hyperelastic
spring acts in parallel to the Maxwell element which consists of an elastic spring and viscoplastic dashpot that are placed in series. The one-dimensional rheological model, constitutive
equations, and parameters are shown below in Figure A-3-1. The definitions of the ten model
parameters are listed below.
p : product of Boltzmann's constant, the absolute temperature, and the chain density
N: number of "rigid links" between two crosslinks and/or strong physical entanglements
vo: effective initial soft domain volume fraction
vss: effective saturation soft domain volume fraction
A:
Ev:

parameter characterizing the evolution of the soft domain volume fraction


elastic modulus of the linear, elastic spring in the viscoelastic-plastic Maxwell
element
v: Poisson's ratio
ro : factor proportional to the visco-plastic shear strain rate

AG: activation energy in the limit of zero applied stress


s,: initial athermal shear strength

206

Vi

kic
cscoNlticp

nsyumyqirug
mrbbaY swingg
00)

Summary of cosituatve model and material parmeters


T=

'+eT

Equilitaimn

Hyperedastc

strs-strain
spons T'

rubbery spri
element

Non-equihim
rate-dependent

Linear spring
element

Hyperlastic "filera

PL,
N

Linear Elic
E,,

arestin
response Ty

TV - -rL[InV1
a. =I - V.

Time dependWMdncM

Viscoplastic

dashpot

ct

irm= *(ffe-'(Ay)
x - 1 + 3.5( - ) + 1S(1 - S)2

t, AG

element

so

Figure A-3-1: One dimensional rheological model and material parameters summary of the Qi and Boyce
thermoplastic polyurethane constitutive model as shown in [1].

Note that the soft domain volume fraction, vs, is not constant but evolves with deformation. This
is because as the polyurethane stretches, the hard domain network breaks down and the effective
soft domain volume fraction increases (see Figure A-3-2). By accurately capturing the evolution
of the soft domain volume fraction, the hysteresis, softening, and time-dependence of the
material can be predicted. However, the effective soft segment volume fraction does not
correlate to the chemically-determined composition because prior to deformation some of the
soft segments are constrained by hard micro-domains and effectively increases the hard domain
volume fraction. Similarly upon extensive deformation not only is the soft segment unconstrained, but the hard micro-domains are disrupted and effectively increases the soft volume
fraction.

207

__

sU0

" ss

Figure A-3-2: Cartoon of soft segment evolution with matrix stretch. Red circles denote regions of soft
segment constrained by hard micro-domains (black) in the polyurethane microstructure cartoon borrowed
from [2]. Initially the effective soft segment volume fraction is vso is less than the chemically-determined, soft
segment volume fraction. Upon deformation the hard domains break up (at high hard segment
concentration) and/or rotate and align releasing constrained soft segment. The effective soft segment volume
fraction, v,, increases with deformation until a saturated value is achieved, v,,. This saturated value
corresponds to a value larger than the chemically-determined, soft segment volume fraction due to the
inclusion of damaged hard-micro-domains.

Qi and Boyce utilize this evolution to amplify the hyperelastic response and drive the evolution
in the non-equilibrium response. The effective hard segment volume fraction, Vh, is defined as
(A-3.1)

vh = 1- v s

and the effective soft volume fraction evolves non-linearly with the maximum applied stretch
according to the following equation:
,Zlock -

K :A(v"

-Vs)

chain

where A

(A-3.2)

(,lock chainA max)2 A chain',


chain

is an amplification of the equivalent maximum stretch achieved defined in [1].

A-3.1.1 Hyperelastic Rubbery Network Behavior


The hyperelastic response of the polyurethane was based on the response devised by Arruda and
Boyce when an eight-chain network representative volume element (shown in Figure A-3-3) was
utilized [3].

208

(a)

(c)

(b)

FX74--'.
10
Figure A-3-3: Eight chain rubber elasticity model developed by Arruda and Boyce. The undeformed (a),
uniaxially extended (b), and biaxially extended (c) configurations are displayed as shown in [3].

However, the equivalent chain stretch Achain was redefined as A chain and the previously defined
value amplified to account for the presence of obstacles, such as rigid particles, around which
polymer chains must stretch. The amplified equivalent chain stretch is defined as follows
2
(A-3.3)
+1
Achain = X
c2han -)
where the amplification factor, X, is defined as

X = 1+3.5(1- v, ) + 18(1- v)2.

(A-3.4)

It is clear that the amplification of the hyperelastic stress-response, X, is an extension of the


Guth-Gold rigid particle filler model in which the hard domain volume fraction (1-v,) is utilized
as the rigid particle filler volume fraction [1, 4].
The resulting hyperelastic response can fit the equilibrium behavior (described in section 5.3.5)
exhibited by pure Elasthane and the Laponite-filled nanocomposites. When the five parameters
are defined as follows in Table A-3-1, the resulting response shown in Figure A-3-4 captures the
equilibrium behavior of the Elasthane/Laponite nanocomposites. The addition of Laponite does
not influence the equilibrium or hyperelastic response-this response is matrix driven.

N
3.7

r
1.6 MPa

vs

3.6

0.45

vSS
0.85

Table A-3-1: Five parameters needed to describe the hyperelastic or equilibrium response of the
Elasthane/Laponite nanocomposites.

209

l
0

0.0

0.2

0.4

0.6

0.8

Nominal Strain

Nominal Strain

1.0

Nominal Strain

Figure A-3-4: The equilibrium or hyperelastic response of the Elasthane/Laponite nanocomposites is the
same despite Laponite concentration and can be captured by the constitutive model of Qi and Boyce [1, 5].
The equilibrium paths of the 0, 4, 10 and 15 wt% Laponite filled Elasthane determined experimentally from
the relaxed response are described by the solid thick lines in the left-most figure. The response is the same
despite the Laponite concentration. Using the constitutive model described by Qi & Boyce this equilibrium or
hyperelastic response is captured. The hyperelastic model (grey solid line) falls within the relaxed response of
the pure (center) and 15 wt% Laponite-filled (right) Elasthane.

A-3.1.2 Viscoelastic-Plastic Behavior


The non-equilibrium behavior or viscoelastic-plastic behavior of this constitutive model too
depends on the evolution of the soft volume fraction. The complex Maxwell element with linear
spring and non-linear dashpot accounts for the rate dependence and stress relaxation of the
material. The equivalent shear stress was defined as follows:
, = -T" *

(A-3.5)

"

where
(A-3.6)

TV = T- TN.

The equivalent shear stress coupled with a softening parameter, s, defined by


S=

(A-3.7)

-vso

with constant model parameters defines the visco-plastic shear strain rate, j:
S= fo exp

(A-3.8)

While this model exhibits two characteristic relaxation times upon relaxation, both the short and
long time constant evolve with stretch or orientation. It is impossible to separate the nonequilibrium contribution of the hard segment volume fraction from the soft segment volume

210

fraction contribution as currently prescribed. It is also not realistic to assume that the spring in
the Maxwell element is linear and does not evolve too with stretch or orientation. Furthermore,
the method to identify visco-plastic parameters identified by Qi and Boyce proved inadequate
because the equivalent shear stress deviated from the behavior the authors observed as is shown
in Figure A-3-5.

I I I I I I I

. II I II I

0 wt% Laponite

I II I
0.0017 s "i
0.0167 s'
0.1667 s'

cc

a.
e
v1

a-)

1
I
0.0

.2I
I I I
0.2 0.4 0.6

I I I I 1.4
It. 1 1 1
0.8 1.0 1.2 1.4 1.6 1.8
vy

(b)

YV

Figure A-3-5: Comparison of the equivalent shear stress versus shear strain for pure Elasthane in tension to
that of the polyurethane evaluated by Qi and Boyce in compression [1].

The equivalent shear stress observed by the authors reached a plateau while the equivalent shear
stress of pure Elasthane decayed with increased stretch or equivalent shear strain. This behavior,
or decay, was also observed in the Laponite-filled composites. Consequently augmentation of the

Qi and Boyce constitutive model is needed.


A-3.2 Necessary Augmentation of the Constitutive Model
To more accurately capture the behavior of the Elasthane/Laponite nanocomposites a
constitutive model that explicitly separates the non-equilibrium stress contribution from the soft
volume fraction and the hard volume fraction is needed. A more appropriate model would
employ the same similar hyperelastic spring placed parallel to two Maxwell elements (see Figure
5.9). As described in Chapter 5.3.4 one Maxwell element would correspond to the soft segment,
non-equilibrium stress contribution while the other element would correspond to the hard
segment, non-equilibrium stress contribution. This would allow the soft segment relaxation time

211

constant to evolve with orientation. Meanwhile, the hard micro-domain relaxation behavior
could act independently and its characteristic relaxation time maintained. In both instances the
stress contribution from the spring comprising the Maxwell element must decay non-linearly,
following a power-law scaling that depends on Laponite concentration, aspect ratio and
polydispersity. Simultaneously, the dashpots of these Maxwell elements must accurately predict
the strain rate sensitivity exhibited by these nanocomposites.

Unfortunately, I was unable to develop a model to capture this behavior in my last four months at
MIT. It has recently come to my attention that Dr. Renaud Rinaldi, a post-doc working for
Professor Mary C. Boyce, has developed a constitutive model for polyurethanes that separates
the hard and soft segment non-equilibrium contribution. It is likely that this model could easily
be modified to capture the behavior exhibited by the high performance, Elasthane/Laponite
nanocomposites.

A-3.3 References
[1]
[2]

[3]

[4]
[5]

H. J. Qi and M. C. Boyce, "Stress-strain behavior of thermoplastic polyurethanes,"


Mechanics of Materials,vol. 37, pp. 817-839, 2005.
D. J. Martin, G. F. Meijs, P. A. Gunatillake, S. P. Yozghatlian, and G. M. Renwick, "The
influence of composition ratio on the morphology of biomedical polyurethanes," Journal
ofApplied Polymer Science, vol. 71, pp. 937-952, 1999.
E. M. Arruda and M. C. Boyce, "A three-dimensional constitutive model for the large
stretch behavior of rubber elastic materials," Journalof the Mechanics andPhysics of
Solids, vol. 41, pp. 389-412, 1993.
E. Guth, "Theory of Filler Reinforcement," JournalofApplied Physics, vol. 16, pp. 2025, 1945.
H. J. Qi and M. C. Boyce, "Constitutive model for stretch-induced softening of the stressstretch behavior of elastomeric materials," Journalof the Mechanics and Physics of
Solids, vol. 52, pp. 2187-2205, 2004.

212

Appendix A-4-2-D WAXS Spectra Collected at Each


Specimen Stretch for Pure Elasthane and that Reinforced
with 8 wt% Laponite
The spectra at specimen stretches equal to 1.1, 1.6, 3.0 and 9.0 are used in the results displayed
in Chapter Five. These stretches correspond to the approximate locations in the stress-strain
curve given by the black, red, green and blue empty circles shown below.

t:

6
above in
Figure A-4-1: The specimen stretches equal to 1.1, 1.6, 3.0 and 9.0 correspond to the locations shown
testing.
during
slip
specimen
to
the stress-strain curve due

All of the scatter patterns collected are shown on the fourth and fifth page of this appendix.
Below find enlarged patterns for the four positions of interest for the pure and the 8 wt%
Laponite-filled Elasthane composite.

213

Figure A-4-2: Enlarged scatter patterns at specimen stretches equal to 1.1, 1.6, 3.0, and 9.0 for pure
Elasthane.

214

Figure A-4-3: Enlarged scatter patterns at specimen stretches equal to 1.1, 1.6, 3.0, and 9.0 for the 8 wt%
Laponite-filled Elasthane nanocomposite.
=1.6

-=11

215

1.4

1.6

1.8

2.0

2.5

3.0

4.0

5.0

..

6.0

7.0

8.0

9.0

Owt%

%IM 8
0"6

0"8

91

V"!,

O'L

0"9

0*9

8L

p p

--

The 1-D SAXS pattern collected on a static film of pure Elasthane and filled with 8 wt%
Laponite is shown below in Figure A-4-4. The behavior of the 8 wt% filled sample is
representative of the spectra exhibited by nanocomposites filled with various concentrations of
Laponite. SAXS data has been expressed in terms of q, where q = -- sin(). The peaks
identifiable in the pure polyurethane that correspond to the hard micro-domain size and spacing
are indistinguishable in the Laponite-filled samples. Consequently, the SAXS data is not
reported.

.0

0,,

a,

*8

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

q, nm"1
Figure A-4-4: I-D SAXS spectra of pure Elasthane and that filled with 8 wt% Laponite. The broad peaks at q
- 0.18 and -0.50 are masked by the presence of Laponite due to electron-density contrast matching.

218

Appendix A-5-Movies of Electrospun Mat Lateral Shrink

The movie titled Appendix A-5-1 demonstrates the quick shrink of the as spun non-woven, mat
(17wt% Polyurethane in solution) when it is thermally actuated by placing it on a hot plate ( T=
75'C; each grid edge measures 1 cm).

The movie titled Appendix A-5-2 demonstrates the micron-scale change in mat morphology
observed in an environmental scanning electron microscope. The two ends of the film are held in
position while the middle of the film is allowed to shrink freely-consequently, the film appears
to narrow. The slow strain recovery occurs because the peltier plate can only achieve a
temperature of 62 0 C. Notice that the BOAS morphology re-coils along the film edges where the
morphology is least constrained.

219

You might also like