PIEZORESISTIVE POLYVINYLIDENE FLUORIDE/CARBON FILLED
NANOCOMPOSITES
Shailesh Vidhate, B. E.
Thesis Prepared for the Degree of
MASTER OF SCIENCE
UNIVERSITY OF NORTH TEXAS
May 2011
APPROVED: Nandika D’Souza, Major Professor Vijay Vaidyanathan, Committee Member Witold Brostow, Committee Member Nigel Shepherd, Committee Member Jaycee Chung, Committee Member Narendra Dahotre, Chair of the
Department of Materials Science and Engineering
Costas Tsatsoulis, Dean of the College of Engineering
James D. Meernik, Acting Dean of the Toulouse Graduate School
Vidhate, Shailesh. Piezoresistive Polyvinylidene Fluoride/Carbon Filled
Nanocomposites
This thesis examines the value of using dispersed conductive fillers as a
stress/strain sensing material. The effect of the intrinsic conductivity of the filler on the
ability to be effective and the influence of filler concentration on the conductivity are also
examined.
. Master of Science (Materials Science and Engineering), May 2011,
74 pp., 25 figures, 4 tables, and chapter references.
To meet these objectives, nanocomposites of polyvinylidene fluoride (PVDF) with
carbon nanofibers (CNFs) and carbon nanotubes (CNTs) were prepared by melt-
blending using a twin screw extruder. Since PVDF has a potential to be piezoresistive
based on the type of crystalline phase, the effect of CNFs on PVDF crystallinity,
crystalline phase, quasi static and dynamic mechanical property was studied
concurrently with piezoresponse. Three time dependencies were examined for
PVDF/CNTs nanocomposites: quasi-static, transient and cyclic fatigue. The transient
response of the strain with time showed viscoelastic behavior and was modeled by the
4-element Burger model. Under quasi-static loading the resistance showed negative
pressure coefficient below yield but changed to a positive pressure coefficient after
yield. Under cyclic load, the stress–time and resistance–time were synchronous but the
resistance peak value decreased with increasing cycles, which was attributed to charge
storage in the nanocomposite.
The outcomes of this thesis indicate that a new piezoresponsive system based
on filled polymers is a viable technology for structural health monitoring.
ii
Copyright 2011
by
Shailesh Vidhate
iii
ACKNOWLEDGEMENTS
I would like to express my heartfelt gratitude to my major advisor Professor
Nandika D’Souza for accepting me to conduct research in her lab and for his continual
guidance, support and encouragement throughout my master’s program without which
this thesis would not have been possible.
I would like to thank Professor Vijay Vaidyanathan for his guidance as a
committee member. Advice from my committee members Professor Whitold Brostow
and Professor Nigel Shepherd is highly appreciated.
I am grateful to all the help rendered by Sandeep Manandhar whenever I
needed. A special thanks to Ali Shaito and Koffi Dagnon for training me on all the PMRL
lab equipments. I would like to thank the members of MTSE departmental staff: Joan
Jolly, Wendy Agnes, April Porter, John Sawyer and David Garrett. Thanks to Dr. Dave
Diercks and Dr. Nancy Bunce of UNT-CART for giving me equipments training and
generous assistance in the use of CART equipments.
I am grateful to all the help and support rendered by Mohammad Maneshian, Tea
Datashvili, Mangesh Nar, Emmanuel Ogunsona, Dong Le and all the students and
staffs of Materials Science and Engineering for the good times together.
I am forever indebted to my parents Pandit Vidhate and Padma Vidhate; and my
fiancé Gauri Khandekar for their care and support in every aspect of my life.
iv
TABLE OF CONTENTS
Page ACKNOWLEDGEMENTS ................................................................................................... iii LIST OF TABLES ............................................................................................................... vii LIST OF ILLUSTRATIONS ............................................................................................... viii Chapters
1. POLYMER/CARBON NANOPARTICLE NANOCOMPOSITES .................. 1
1.1 Introduction ......................................................................................... 1
1.2 How Nanocomposites Work? ............................................................ 2
1.3 How to Produce Nanocomposites? ................................................... 2
1.3.1 Melt Mixing .............................................................................. 3
1.3.2 In Situ Polymerization ............................................................. 3
1.3.3 Solution Technique ................................................................. 4
1.4 Polymer Carbon Nanoparticle Composites ....................................... 4
1.5 Carbon in Various Forms ................................................................... 5
1.5.1 Carbon Nanofibers .................................................................. 5
1.5.2 Carbon Nanotubes .................................................................. 7
1.6 Dispersion of Nanotubes in Polymer Matrix ...................................... 9
1.7 Percolation Theory ........................................................................... 12
1.8 Piezoresitivity in CNT Filled Nanocomposites ................................ 14
1.9 Reference List .................................................................................. 15 2. CRYSTALLIZATION, MECHANICAL AND RHEOLOGICAL BEHAVIOR
OF POLYVINYLIDENE FLUORIDE/CARBON NANOFIBER COMPOSITES ............................................................................................ 18
2.1 Introduction ....................................................................................... 18
2.2 Experimental .................................................................................... 20
2.2.1 Materials ................................................................................ 20
2.2.2 Preparation of PVDF Fibers ................................................. 20
2.2.3 Differential Scanning Calorimetry (DSC) ............................. 21
2.2.4 X-Ray Diffraction ................................................................... 21
v
2.2.5 Mechanical Testing ............................................................... 21
2.2.6 Dynamic Mechanical Measurements (DMA) ....................... 21
2.2.7 Scanning Electron Microscopy (SEM).................................. 22
2.2.8 Melt Rheology ....................................................................... 22
2.3 Results and Discussion.................................................................... 22
2.3.1 Crystallization Behavior ........................................................ 22
2.3.2 Dynamic Mechanical Behavior ............................................. 25
2.3.3 X-Ray Diffractometry............................................................. 28
2.3.4 Mechanical Properties .......................................................... 29
2.3.5 Rheological Measurements .................................................. 32
2.4 Conclusion ........................................................................................ 35
2.5 Reference List .................................................................................. 36 3. TIME DEPENDENT PIEZORESISTIVE BEHAVIOR OF
POLYVINYLIDENE FLUORIDE/CARBON NANOTUBE CONDUCTIVE . 38
3.1 Introduction ....................................................................................... 38
3.2 Burgers Model .................................................................................. 39
3.3 Experimental .................................................................................... 40
3.3.1 Materials ................................................................................ 40
3.3.2 Sample Preparation .............................................................. 41
3.3.3 Measurements ...................................................................... 41
3.4 Results and Discussion.................................................................... 41
3.4.1 Compressive Stress and Resistance Response under Quasi-Static Loading ....................................................................... 41
3.4.2 Compressive Creep and Resistance under Transient Creep
............................................................................................... 43
3.4.3 Cyclic Loading and Electric Resistance Response of Sample
............................................................................................... 45
3.5 Conclusions ...................................................................................... 45
3.6 Reference List .................................................................................. 47 4. RESISTIVE-CONDUCTIVE TRANSITIONS IN TIME DEPENDENT
PIEZORESPONSE OF PVDF-MWCNT COMPOSITES ........................... 48
4.1 Introduction ....................................................................................... 48
4.1.1 Burgers Model ....................................................................... 51
vi
4.2 Experimental .................................................................................... 53
4.2.1 Materials ................................................................................ 53
4.2.2 Sample Preparation .............................................................. 53
4.2.3 Measurements ...................................................................... 53
4.2.4 Raman Spectroscopy ........................................................... 56
4.3 Results and Discussion.................................................................... 56
4.3.1 Compression Test ................................................................. 56
4.3.2 Resistance Response for Creep and Relaxation of PVDF/MWCNT Composites ................................................. 60
4.4 Conclusions ...................................................................................... 69
4.5 Reference List .................................................................................. 70 5. SUMMARY .................................................................................................. 73
5.1 Effect of Carbon Nanofibers on Thermo-Mechanical Properties ... 73
5.2 PVDF/CNTs Nanocomposites’ Time Dependent Piezoresistive Effect................................................................................................. 74
5.3 Resistive to Conductive Transition in PVDF/CNTs Nanocomposites .......................................................................................................... 74
vii
LIST OF TABLES
Page 2.1 Comparison of various materials’ properties ........................................................... 8
2.2 DSC results from second-heating and second-cooling thermograms .................. 25
4.1 Results of the Burger model .................................................................................. 43
4.2 Results for electric fit .............................................................................................. 61
viii
LIST OF ILLUSTRATIONS
Page
1.1 Schematic structure of carbon naotubes (a) graphene layer, (b) stacked cone sherringboned nanofiber, and (c) nanotube ............................................................ 6
1.2 Schematic illustrations of the structures of (A) armchair, (B) zigzag, and (C) chiral SWNTs ................................................................................................................... 8
1.3 (a) Overall statistics of the journal papers reviewed in this article which addresses the influence of various pre-treatment in polymer/CNT composites and compares with respect to pristine CNT; (b) Statistics showing the strengths (+) and weaknesses (-) of covalent and (c) non-covalent types of pre-treatment on the composite properties (statistics also includes paper which report simultaneous improvement in both structural and electrical properties). ............. 11
1.4 Percolation theory: (a) Well dispersed conductive nanofillers: Non conductive composite (b) High concentration of fillers, well distributed but not forming conductive path (c) Filler concentration at percolation threshold forming conductive path ...................................................................................................... 13
2.1 DSC second-heating curves showing an increase in melting point with increasing CNF content ........................................................................................................... 23
2.2 DSC second-cooling curves showing an increase in melt recrystallization temperatures with increasing CNF content ........................................................... 24
2.3 DMA results showing a decrease in tan δ peak height and temperature with increasing CNF content (a) for the β transition, (b) for the α transition ................ 27
2.4 X-ray diffraction spectrographs for PVDF and PVDF/CNF fibers ......................... 29
2.5 Stress strain curves for PVDF and its composites ................................................ 30
2.6 SEM images of PVDF composites showing fiber pull out mechanisms dominating failure (a) PVDF1 (b) PVDF2 (c) PVDF4 ............................................................... 31
2.7 Dynamic strain sweep tests showing storage modulus as a function of strain .... 32
2.8 (a) Storage modulus G’ versus frequency at 180 ºC temperature, (b) Loss modulus, G versus frequency at 180 °C temperature, (c) Complex viscosities η* at 180 °C temperature ............................................................................................ 34
3.1 Schematic diagram of Burgers model ................................................................... 40
3.2 PPC and NPC phenomenon in PVDF MWCNT conductive composite ............... 42
ix
3.3 (a) Creep compliance versus time in compressive creep test (b) Change in fractional resistance in creep test .......................................................................... 44
3.4 Resistance response under cyclic loading ............................................................ 45
4.1 (a) A typical creep relaxation curve of a viscoelastic material. (b) Schematic diagram of Burgers model and equivalent electric model..................................... 52
4.2 Experimental procedure showing sample preparation and electrical, mechanical response measurement technique. ....................................................................... 55
4.3 (a) Compressive stress strain curves. (b) Yield stress and compressive modulus values comparison for PVDF/MWCNT composites. ............................................. 57
4.4 Resistance change as a function of the applied pressure in piezoresistive composites. Presence of PPC and NPC phenomenon in (a) PVDFCNT1 and (b) PVDFCNT2, (c) PVDFCNT4 and (d) PVDFCNT10 showing only NPC behavior.59
4.5 Experimental and predicted creep compliance versus time curves ..................... 61
4.6 (a) Creep compliance versus time curves and (b) simultaneously recorded change in resistance versus time curves of PVDF/MWCNT composites. ........... 63
4.7 Schematic showing the effect of MWCNT-MWCNT contact leading to time dependent resistive response at low concentrations and conductive response at high concentrations. ............................................................................................... 64
4.8 Experimental and predicted resistance creep versus time curves ....................... 66
4.9 Raman spectra of PVDF/CNT composites using (a) line mapping to examine large area of sample. (b) The peaks arising from C-MWNTs (D, G, and G’ bands) are indicated in normalized spectra. (c) Raman line mapping spectra acquired from positions along the line for PVDFCNT1, (d) PVDFCNT2, (e) PVDFCNT4, and (f) PVDFCNT10. .............................................................................................. 68
1
CHAPTER 1
3BPOLYMER/CARBON NANOPARTICLE NANOCOMPOSITES
1.1 7BIntroduction
Nanotechnology is now recognized as one of the most promising technologies of
the 21st century. Among various materials research, polymer nanocomposites is
emerging as a multidisciplinary research activity. Results obtained through the research
of polymer nanocomposites can broaden the applications of polymers to a great extent.
Multifunctional advanced polymeric nanocomposites can be used for wide variety of
applications in various different fields.
Polymer nanocomposites, a multiphase solid material where one of the phases is
less than 100 nm size, are becoming popular and being manufactured commercially for
various diverse applications. In the last twenty five years there has been intensive
research on polymer nanocomposites. Simultaneously, growth in the computer
simulation techniques, scanning electron, and transmission electron microscopy has
made the characterization and prediction of the polymer nanocomposites’ properties
easier. In addition, nanocomposites can be processed using conventional processing
techniques and does not need any special or costly processing techniques. Today
various types of nanomaterials with various shapes and sizes are being used to prepare
polymer nanocomposites. The nanofillers can be in the form of nanoparticles (e.g.
carbon, metal powder), nanoplateletes (e.g. silicates), nanowires (e.g. carbon
nanotubes, ceramic nanowires), fullerences (e.g. C60), etc. To fulfill the objective of this
research carbon nanofibers and carbon nanotubes were used as nanofillers to make
2
nanocomposites with polyvinylidene fluoride (PVDF) polymer. The details about carbon
nanofibers and carbon nanotubes are discussed in this chapter.
1.2 How Nanocomposites Work?
Transition of fillers from macro size to nano size drastically increases the surface
area per unit volume of the particles and also causes change in their physical properties
[ 3F1]. Small size of particles provides large interfacial area between particles and host
material. Nanofillers provide higher reinforcing efficiency due to their high aspect ratio.
In general, shape and size of the particles have direct effect on properties of the
prepared nanocomposites. Along with the individual properties of the host material and
fillers, interfacial region shared by both the components affect nanocomposite’s
properties. Other factors like aspect ratio of the nanoparticles, filler dispersion in the
matrix, physical or chemical interaction of nanoparticles with host material affect
properties of nanocomposits. In the early 1990s, Toyota Central R&D Laboratories in
Japan pioneered the work on nanocomposites showing considerable improvement in
thermal and mechanical properties of Nylon-6 nanocomposite made by addition of
small amount of nano fillers [ 4F2]. Since then polymer nanocomposites research became
commercially and scientifically attractive topic all over the world.
1.3 9BHow to Produce Nanocomposites?
Polymer nanocomposites can be produced using various techniques. The goal of
any processing technique to produce nanocomposite is achieving maximum possible
dispersion of nanofillers in polymer matrix. Techniques to produce nanocomposites are
discussed below.
3
1.3.1 32BMelt Mixing
Since the invention of nanocomposites the melt mixing technique is highly
attractive owing to its simplicity [ 5F3, 6F4]. This technique has been widely studied and well
explored with polymer clay system and generated knowledge with this system can be
used easily with other nanofillers and polymer nanocomposites. In this technique a
thermoplastic polymer and nano fillers mixed using conventional melt mixing methods
such as extrusion, batch mixing, or injection molding. No solvent is required in this
technique and fillers mixed in the molten matrix using high shear forces. Viscosity of the
melt plays important role in shear mixing of the nanofillers and polymer melt during
processing [ 7F5- 8F9F7]. Compatibilizers can also be used to improve the interfacial adhesion
between polymer and fillers. However, organic surface modifications are prone to
thermal damage and optimum processing conditions need to be selected. Increased
mixing time can improve the dispersion of nanofillers in polymer but long processing
time can degrade the heat sensitive polymers. Elongational flow and orientation of
extrudate during extrusion leads to orientation of fillers in the direction of extrusion.
Large amount of polymer composites can be processed compared to other techniques.
Achieving complete exfoliation or dispersion of nanofillers in polymer matrix with this
technique is difficult.
1.3.2 33BIn Situ Polymerization
In this technique monomer is dissolved or suspended in solvent [ 10F811F12F13F- 14F12 ]. The nano
particles are dispersed or swelled in liquid monomer by ultra sonication or vigorous
mechanical stirring. Low viscosity of the monomer improves the dispersion of the
nanofillers. Combined solution of monomer and fillers is then polymerized using initiator
4
at elevated temperature or using radiation. Subsequent polymerization of the monomer
leads to uniforn intercalation or dispersion of fillers in polymer matrix. This method is
common for thermoset resins as non reacted liquid resin can be crosslinked in between
the dispersed nanofilres [ 15F13]. This approach has also been successfully used for
thermoplastics like nylon [ 16F14], polystyrene (PS) [17F15], poly metyl methacrylate (PMMA)
[ 18F16], etc. Nandika et al. [19F17] polymerized layered double hydroxide (LDH) dispersed
styrene monomer by using free radical polymerization method and benzoyl peroxide as
a initiator to make PS/LDH nanocomposite.
1.3.3 34BSolution Technique
In this technique polymer is dissolved in suitable solvent [ 20F18, 21F19]. Nanofillers
which tend to form stacked structure or agglomerate can also be dispersed in the same
solvent using sonication or mechanical stirring. The polymer chains get adsorbed on the
filler surface or get intervened between nanofillers. When the solvent is evaporated,
remaining polymer and nanofillers form a nanocomposite. Removal of the solvent is
critical issue in this technique.
1.4 Polymer Carbon Nanoparticle Composites
Today, carbon nanofillers are ideal fillers for polymers owing to their high
mechanical and electrical properties. By the late 1980s, carbon filaments were able to
manufacture in several gram size quantities. But at the same time the ability of those
nano size fillers to enhance the mechanical properties as well as electrical properties of
the polymers was not realized by the researchers at that time. Also, the cost of
manufacturing carbon nanotubes was barrier to commercialize and use them into
composite systems.
5
After Iijima discovered carbon nanotubes in 1991 [ 22F20 23F21], significant efforts have
been made to incorporate carbon nanotubes in various materials. Carbon nanotubes
have high aspect ratio, high modulus, and strength and therefore they are promising
fillers in polymer composites. Significant enhancement in the tensile modulus and
tensile strength has been reported. The estimated mechanical properties of carbon
nanotubes are higher than the real values but still represent high potential filler
materials for polymer nanocomposites.
The glass transition temperature and thermal degradation temperature is also
observed to be significantly increased by addition of carbon nanofillers in polymer. Also,
addition of carbon nanotubes in polymeric system can impart improved electrical and
thermal conductivity [ 24F22- 25F26F27F28F29F30F31F32F30]. Also, many potential applications have been proposed for
carbon nanotubes and carbon nanotubes based systems, including high strength
conductive composite, sensors, thermal conductors, energy storage, semiconductor
devices and probes, etc.
1.5 Carbon in Various Forms
The structure and properties of various forms of carbon are completely different.
Carbon forms various structures like graphite, diamond, carbon fibers, carbon black,
bucky balls (C60), carbon nanofibers, carbon nanotubes, etc. In the experimental work
of this thesis I have used carbon nanofibers and carbon nanotubes as a conductive
fillers so these forms of carbon has been elaborately discussed below.
1.5.1 35BCarbon Nanofibers
Carbon nanofibers and carbon nanotubes looks similar under electron microscope
but Melechko et al. [33F31] showed that the planar structures and arrangement of graphen
in both
graphen
graphen
propertie
150 nm
were us
Figure 1sherring
structure
ne planes a
ne planes
es of carbo
with sever
ed and deta
.1 Schemagboned nan
is differen
are wrappe
is presen
on nanofibe
al micron le
ails of these
atic structurofiber, and
t (Figure
ed into trun
nt in carbo
ers is diffic
engths. In t
e carbon fib
re of carbon(c) nanotu
6
1.1). Unlike
ncated con
on nanofib
cult. Carbon
this work ca
bers are giv
n naotubes be [31]
e in nano
nic sections
bers so de
n fiber diam
arbon fiber
ven in chap
(a) graphe
tubes, car
s. Anisotro
efining intr
meter rang
rs supplied
pter 1.
ene layer, (b
bon nanof
pic structu
rinsic elec
es from 10
by Pyrogra
b) stacked
fibers
re of
ctrical
00 to
aph®
cone
7
1.5.2 36BCarbon Nanotubes
Carbon nanotubes are the allotropes of carbon, cylindrical in nature and made by
rolled graphene sheets. CNTs show incredible thermal, mechanical, and electrical
properties. Chemical composition of the carbon nanotubes is simple but shows diverse
structure property relationship. Each carbon atom in the nanotube is attached to three
neighboring carbon atoms by sp2 bonding in x-y plane. The presence of sp2 bonding in
carbon nanotubes which forms hexagonal lattice is stronger than sp3 bonding which
forms cubic structure in diamond. Theoretical tensile strength of carbon nanotube is
130-150 GPa and modulus is over 1000 GPa [ 34F32]. In laboratories actual measured
value of up to 63 GPa has been reported which is higher than diamond, Kevlar, or
spider’s silk. Comparison of various materials properties are shown in Table 1.1.
The delocalized pi-electron cloud in the z-direction of graphene plane gives unique
electrical properties to carbon nanotubes [ 35F33]. Thermal and electrical conductivity of the
carbon nanotubes depends on the chirality or twist of the carbon nanotubes (chirality is
the chiral angle between hexagons of carbon nanotubes and the tube axis). Fig. 1.2
shows different structures formed by CNTs due to arrangement of hexagonal lattice
structure. Single wall nanotubes can be either conductive or semiconductive according
to structure. However, measuring electrical and thermal properties of single nanotubes
is challenging.
Table 1
M
Single w
Multi wa
Di
K
W
Figure 1chiral SWaxis are
Compariso
Material
wall nanotub
all nanotub
amond
Kevlar
Steel
Wood
1.2 SchemaWNTs. Projon the top
on of variou
Young
(
be 1
be 1
atic illustratjections noand bottom
s materials
's Modulus
GPa)
1054
1200
600
186
208
16
tions of thermal to the
m, respectiv
8
s’ properties
Tensile S
(GP
15
15
13
3.
1
0.0
structurestube axis a
vely [34].
s
Strength
Pa)
50
50
30
6
1
008
s of (A) armand perspe
Density
(g/cm3)
1.4
2.6
3.5
7.8
7.8
0.6
mchair, (B) zective views
zigzag, ands along the
d (C) tube
9
There are two types of carbon nanotubes: 1) Single wall carbon nanotubes
(SWNTs) and 2) Multi wall carbon nanotubes (MWNTs). MWNTs consists of one atom
thick single graphene sheet rolled around itself. Diameter of nanometer and length upto
1000 nm has been achieved by many researchers depending on the synthesis process
variables. SWNTs show better transport properties than MWNTs. MWNTs consist of
many concentrically nested SWNTs [ 36F34]. Space between concentric nanotubes is
slightly larger than the single graphene sheet because of the presence of geometrical
strains occurred during formation of nanotubes. MWNTs having outer diameter of 4-50
nm, inner nanotube diameter of 2.2 nm, and length up to several microns has been
reported by researchers. In this work, I have used MWNTs supplied by Baytubes® and
the detail information has been given in chapter 3 and 4.
1.6 Dispersion of Nanotubes in Polymer Matrix
Due to high polarizability of the pi-electrons there is a strong interaction between
adjacent nanotubes through van der Walls’ forces and weak interaction between
graphene sheets. Attraction between nanotubes to nanotubes leads to formation of
agglomerates of nanotubes and that hinders well dispersion of nanotubes in polymer
matrix. The dispersion of nanotubes in polymer matrix can be improved by covalent
treatment or non-covalent treatment of carbon nanotubes (CNTs). In covalent treatment
attachment of the chemical structure on the surface of the nanotubes or at the end of
nanotubes is possible. Covalent treatment causes change in possible hydrogen bonding
and results in separation of nanotubes and can be dispersed in common solvent of
polymer [37F35]. This treatment of solvent also aids to improve the interfacial adhesion
between CNT and polymer matrix. Functionalized CNTs ensure the reactive coupling of
10
CNT to polymer matrix. This manifest stronger interface and efficient load transfer
between CNTs and polymer matrix. Covalent treatment of CNT can ensure the better
mechanical properties of composite due to better dispersion of CNT in polymer matrix
but at the time covalent treatment adversely affect the intrinsic properties of CNTs. This
route causes the change in aspect ratio of the CNTs which generally affect the
percolation threshold of the composite. Electrical properties of CNTs can also be
hampered due to adverse effect of covalent treatment on Fermi level and perturbs pi-
conjugation of CNTs [ 38F36]. Covalent treatment also decreases the phonon-scattering
length which affects the thermal conductivity of the CNTs [ 39F37]. The changes in intrinsic
properties of CNTs directly affect the overall properties of polymer/CNT composite.
Non-covalent treatment is effective to disperse the CNTs without disturbing pi-
conjugation of the CNTs so that the intrinsic properties of the CNTs can be preserved.
S. Bose et al. [40F38] have shown that the electrical properties of the non-covalently treated
CNTs and polymer composites are better than the covalently treated CNTs and polymer
composites.
Figure 1addresscompareweaknescomposimprove
Go
interacti
dispersio
advanta
polymer
polymer
transfer
stress w
would tr
basically
1.3 (a) Oves the infes with ressses (-) oite properment in bot
ood mecha
on and be
on in polym
ges of CN
r and filler g
r to filler. Al
depends u
would transf
ransfer the
y occurs by
erall statistfluence of spect to prof covalentties (statisth structura
nical and
etter disper
mer matrix
NT polyme
gives better
so load sho
pon the sh
fer load from
e load over
y three me
tics of the various pr
ristine CNTt and (c) stics also al and elect
electrical p
rsion of C
x is one of
r Nanocom
r mechanic
ould get dis
ear stress
m matrix to
r long dista
chanisms;
11
journal pare-treatmen
T; (b) Statisnon-covaleincludes
rical proper
properties c
NT in poly
f the major
mposites. B
cal propertie
stributed ov
between m
o filler at sh
ance. Load
chemical b
apers revient in polymstics showent types paper wh
rties) [38].
can be ac
ymer matri
r challenge
Better inter
es due to e
ver the supp
matrix and fi
horter distan
d transfer
bonding, we
ewed in thimer/CNT cing the strof pre-tre
hich report
chieved if g
ix is achie
es before g
rfacial adh
effective loa
porting fille
ller interfac
nce while lo
between m
eak van de
is article wcomposites rengths (+) atment ont simultan
good interf
eved [41F39].
getting the
hesion betw
ad transfer
r. Effective
ce. A high s
ow shear st
matrix and
er Walls fo
which and and
n the eous
facial
CNT
real
ween
from
load
shear
tress
filler
rces,
12
and physical micro interlocking [ 42F40]. Physical micro interlocking is difficult in CNT due
atomically smooth surface.
When CNTs are dispersed randomly in polymer matrix, thermal, mechanical, or
electrical properties are expected to be isotropic. As the filler content in the polymer
matrix increases gradual transition from predominant polymer matrix properties to filler
properties occurs. Above certain concentration of fillers in polymer, overall properties of
the composite are dominated by filler properties and that transition point is called as
percolation threshold. At percolation threshold there is a formation of percolating
network of filler in polymer matrix which helps to transfer electrical, thermal, or
mechanical forces from one end to another. The required concentration for percolation
threshold reduces with increased dispersion of fillers in matrix [43F41].
1.7 Percolation Theory
The bulk electrical conductivity of the single atom thick graphene is highly isotropic.
Along the graphene planes the electrical conductivity takes place without any scattering
while conductivity perpendicular to graphene plane involves considerable scattering
losses. Graphene’s in plane electrical conductivity is 2.1 × 106 S/m and perpendicular to
graphene plane is 5 × 102 S/m [ 44F42]. Carbon nanotubes are nothing but rolled graphene
sheets so at least from electrical point of view the intrinsic electrical conductivity of
nanotubes is similar to graphene sheets. Bulk electrical conductivity can be different
depending on the factors like orientation, entanglement, agglomeration, number of walls
in nanotubes, etc. In MWNTs Ballistic conductivity like superconductivity (electron
transfer without any collision losses) is observed by Poncharal et al. [45F43].
Add
make p
Figure 1
when di
or highly
of condu
whole c
conduct
paths th
the diele
a critica
suddenly
Figure 1compospath (c)
(a)
dition of su
olymers co
1.4) is base
spersed in
y dispersed
uctive partic
composite
ive particle
hrough diele
ectric prope
al concentr
y conductiv
.4 Percolatite (b) HighFiller conce
ch highly c
onductive o
ed on the
polymer m
d particles i
cles comes
can becom
es in dielect
ectric mater
erties of the
ration of c
ve.
tion theory:h concentraentration at
onductive p
or semi-co
way those
matrix. Addi
in polymer
s in contact
me conduc
tric materia
rial. Also t
e polymer m
conductive
: (a) Well dation of fillet percolatio
(b)
13
particles in
onductive. P
nanopartic
tion of very
matrix can
t with each
ctive. With
al increases
he conduct
matrix. Thu
particles a
ispersed coers, well din threshold
dielectric m
Percolation
cles forms
y low amou
n’t impart co
other they
h increase
s the chanc
tivity of who
us percolati
at which i
onductive nstributed b
d forming co
materials lik
n theory (a
link to con
unt of cond
onductivity.
form cond
in the co
ces of form
ole compos
on thresho
insulating
nanofillers: but not formonductive p
(c)
ke polymers
as illustrate
nduct elect
ducting part
. When add
uctive path
oncentratio
ming condu
site depend
old is define
matrix bec
Non conduming condupath
s can
ed in
tricity
ticles
dition
h and
on of
cting
ds on
ed as
come
uctive uctive
14
1.8 Piezoresitivity in CNT Filled Nanocomposites
Piezoresistivity is phenomenon in which electrical resistance of the material
changes with change in applied stress or strain. By addition of carbon nanofillers
dielectric polymer matrix can be made electrically conductive. In electrically conductive
material change in the electrical resistance can be measured easily. So when applied
stress or strain on the material is changed there is change in measured electrical
resistance of the material. Thus change in the electrical resistance of the material is
directly related to the change in the stress or strain of the material. In this way, strain
sensing is possible with piezoresistive materials. Polymeric nanocomposites materials
are the best suited materials for this application as they can be designed for required
stress or strain. To achieve piezoresistivity in polymeric nanocomposites there are
important issues need to be studied and investigated in detail.
In this thesis poly vinylidene fluoride (PVDF) was used as a host material and
CNFs and CNTs were used as a conductive nanofillers. PVDF/CNF nanocomposites
and PVDF/MWCNT nanocomposites were prepared by non covalent melt blending
technique. Various issues like mechanical properties, thermal properties, morphological
structure, piezoresistivity under dynamic and transient stress loading conditions, non-
linear piezoresistive response, etc were studied and discussed elaborately in this thesis.
15
1.9 Reference List
[1] Luo JJ, and Daniel IM. Compos. Sci. Technol. 2003; 63 (11): 1607–1616.
[2] Usuki A, Kawasumi M, Kojima Y, Okada A, Kurauchi T and Kamigaito O J. Mater.
Res.1993, 8 (5): 1174.
[3] Dennis HR, Hunter D, Chang D, Kim S and Paul DR. Polymer 2001; 42: 9513–9522.
[4] Vaia RA, Jant KD, Kramer EJ and Giannelis EP. Chem. Mater. 1996; 8: 2628–2635.
[5] Vaia RA, Ishii H and Giannelis EP. Chem. Mater. 1993; 5: 1694–1696.
[6] Rehab A and Salahuddin N. Materials Science and Engineering A 2005; 399: 368–
376.
[7] Burnside SD and Giannelis EP. Chem. Mater. 1995; 7: 1597-1600.
[8] Usuki A, Kawasumi M, Kojima Y, Okada A, Kurauchi T and Kamigaito OJ. Mater.
Res. 1993; 8(5): 1174.
[9] Alexandre M and Dubois P. Mater. Sci. Eng. Rep. 2000; 28: 1–63.
[10] Beron L, Wang Z and Pinnavia TJ. Applied Clay Science 1999; 15: 11–29.
[11] Halvatty V and Oya A. Appl. Clay Sci. 1994; 9: 199–210.
[12] Hussain F, Dean D and Haque A. Structures and Characterization of Organoclay-
Epoxy-Vinyl ester Nanocomposite, ASME International Mechanical Engineering
Congress and Exposition, LA, USA, IMECE 2002; 33552.
[13] Lan T, Kaviratna PD and Pinnavaia TJ. 1995; 7(11): 2144–2150.
[14] Kawasumi M, Hasegawan M, Usuki A and Okada A. Macromolecules 1997; 30:
6333–6338.
[15] Zhang F, Zhang H, Su Z. Polymer Bulletin 2008; 60: 251–257.
[16] Maurizio A, Maria EE, and Ezio M. Nano letters 2001; 1(4): 213-217.
16
[17] Ogboma SM, Richardson MC, Braterman PS, Xu ZP and D’Souza NA.
“Polystyrene Nanocomposite Materials By In Situ Polymerization Into Zn Al Layer
Double Hydroxide Stearate Hexadecane Styrene Monomer Interlayer“, Annual
Technical Conference Proceedings of the Society of Plastics Engineers Milwaukee,
2008.
[18] Kornmann X, Linderberg H and Bergund L A. Polymer 2001; 42: 4493–4499.
[19] Lee DC and Jang LW. J. Applied Polymer Sci. 1998; 68(12): 1997–2005.
[20] Iijima S. Nature 1991; 354 (6348): 56.
[21] Iijima S and Ichihashi T. 1993; 363(6430): 603–605.
[22] Bower C, Rosen R, Jin L, Han J, Zhou O. Applied Physics Letters 1999; 74(22):
3317–3319.
[23] Cooper CA, Ravich D, Lips D, Mayer J and Wagner HD. Composites Science and
Technology, 2002; 62(7–8): 1105–1112.
[24] Haggenmueller R, Gommans HH, Rinzler AG. Chemical Physics Letters 2000;
330(3–4): 219–225.
[25] Jin L, Bower C and Zhou O. Applied Physics Letters 1998; 73(9): 1197–1199.
[26] Jin Z, Pramoda KP, Xu G and Goh SH. Chemical Physics Letters 2001; 337(1–3):
43–47.
[27] Kearns JC and Shambaugh RL. Journal of Applied Polymer Science 2002; 86(8):
2079–2084.
[28] Lozano K and Barrera EV. Journal of Applied Polymer Science 2001; 79(1): 125–
133.
[29] Potschke P, Fornes TD and Paul DR. Polymer 2002; 43(11): 3247–3255.
17
[30] Safadi B, Andrews R and Grulke EA. Journal of Applied Polymer Science 2002;
84(14): 2660– 669.
[31] Melechko A. Journal of Applied Physics 2005; 97: 041301.
[32] Yu M, Lourie O, Dyer M J, Moloni K, Kelly TF, Ruoff RS. Science 2000; 287(5453):
640.
[33] Terrones M. Annu Rev Mater Res 2003; 33: 419–501.
[34] Baughman RH, Zakhidov AA and Walt A. de Heer. Science 2002; 297 (5582): 787
– 792.
[35] Dyke CA, Tour JM. J Phys Chem A 2004; 108(51): 11151–9.
[36] Kamaras K, Itkis ME, Hu H, Zhao B, Haddon RC. Science 2003; 301(5639): 1501.
[37] Padgett CW, Brenner DW. Nano Lett. 2004; 4(6): 1051–3.
[38] Bose S, Khare RA, Moldenaers P. Polymer 2010; 51: 975–993.
[39] Wagner HD, Vaia RA. Nanocomposites: issues at the interface. Mater. Today 2004;
7: 38–42.
[40] Schadler LS, Giannaris S C and Ajayan PM. Applied Physics Letters 1998; 73(26):
3842–3844.
[41] Strano MS, Nanocomposites: Polymer-wrapped nanotubes. Nat. Mater. 2006; 5:
433–434.
[42] Matsubara K, Sukihara K, and Tusuku T. Phys. Rew. B 1990; 41: 969.
[43] Poncharal P, Berger C, Yi Y, Wang ZL, and de Heer WA. Journal of Phys. Chem. B
2002; 106: 12104.
18
CHAPTER 2
4BCRYSTALLIZATION, MECHANICAL AND RHEOLOGICAL BEHAVIOR OF
POLYVINYLIDENE FLUORIDE/CARBON NANOFIBER COMPOSITES0F
*
2.1 16BIntroduction
Reinforcement of polymers by carbon-based nanofillers has been of increasing
interest because of the multifunctional properties that they result in. Polyvinylidene
fluoride (PVDF) has attracted interest because it is a piezoelectric, pyroelectric, and
ferroelectric material [ 46F1- 47F48F49F4]. PVDF is a semicrystalline polymer with a high molecular
weight and typically has around 50% amorphous content. PVDF shows various
interesting properties like ease of processability, good mechanical properties, thermal
stability, and chemical resistance [ 50F5]. Five crystal structures are present in PVDF. The
electrical properties have been correlated to the β phase, which has been found to
induce polarity in the crystal structure. When PVDF is uniaxially oriented, it results in
longitudinal deformation of polymer chains in the crystals and increased β-phase
formation [51F6]. Serrado Nunes et al. [52F7] have shown that the α phase can be converted
into β phase by mechanical stretching below 100 ºC using a stretching ratio of about 3
to 5, or directly fromsolution at a given temperature. Some researchers have processed
PVDF into porous and nonporous films that have had a 100% β phase [ 53F8- 54F55F10].
PVDF with nanofillers like carbon nanotubes (CNTs), carbon black, and calcium
carbonate has been widely studied. The concept of nanoreinforcement is based on the
fact that a low percentage (3% to 5%) of loading can result in a major change in the
* This entire chapter is reproduced from Shailesh Vidhate, Ali Shaito, Jaycee Chung, Nandika Anne
D’Souza, “Crystallization, Mechanical and Rheological Behavior of Polyvinylidene Fluoride/Carbon Nanofiber Composites”, Journal of Applied Polymer Science, Vol. 112, 254–260 (2009), with permission from Wiley Periodicals, Inc.
19
properties of polymers. Mechanical properties, thermal conductivity, electrical
conductivity, flame retardance, and wear resistance have all shown benefits from
nanofillers [ 56F11]. Single-walled carbon nanotube (SWCNT) and PVDF composites have
demonstrated an increase in mechanical, conducting, and ferromagnetic properties [ 57F12].
Yu et al. [58F13] showed that montmorillonite clays act as nucleating agent and cause the
formation of a γ phase. For clay content greater than 1 weight percent, α and β phases
coexist. When multiwalled carbon nanotubes (MWCNTs) are incorporated, the
crystallites are transformed from the nonpolar α form to polar β form. A percolation
threshold for electrical and thermal conductivity was observed at 2 to 2.5 weight
percentage of MWCNT [ 59F14, 60F15]. MWCNTs also offer ease of processing, flexibility, and
good dielectric behavior of PVDF film [ 61F16, 62F17]. CNT-filled PVDF thin films indicated an
excellent acoustic response, acting as a transducer over a broadband frequency range.
In addition the films were transparent (invisible sound monitors for military applications),
flexible, and lightweight [63F18]. Among the various nanofillers, an increase in electrical
conductivity was also observed with the addition of carbon black [64F19]. Vapor-grown
carbon fibers (VGCFs) have been attracting much research interest as fillers in
composites because of their good electrical, thermal conductivity, and mechanical
properties [ 65F20]. There is to date limited information on carbon nanofiber reinforcement
of PVDF, to our knowledge. However, cost comparisons of multiwalled carbon
nanotubes, single-walled carbon nanotubes, and carbon nanofibers indicate that carbon
nanofibers remain very cost-competitive. The purpose of this study is to investigate the
effects of CNFs on the thermal, mechanical, and rheological properties of PVDF at
different weight percentage loadings.
20
2.2 17BExperimental
2.2.1 Materials
The PVDF used was supplied by Arkema (Kynar® 721, powder form) and had the
following properties:
• Density: 1.78 g/cc
• Melt flow index (MFI): 10 g/10 min
• Tensile strength: 54 MPa
• Melting temperature: 168 ºC
CNFs were obtained from obtained from Pyrograf® Products, Inc. (PR-24-XT-LHT),
with the following material properties:
• Bulk density: 1.95 g/cc (ASTM D1513-86)
• Average diameter: 107 nm (JEOL 5300 SEM)
CNFs were used as received without further purification. Prior to melt-mixing, both
the materials were vacuum-dried at 80 ºC for 6 hours. PVDF and CNFs were dry-mixed
via tumbling in a bottle. The contents of CNFs in PVDF powder were 0, 1, 2, and 4
weight percentage; and the compositions were coded as PVDF, PVDF1, PVDF2, and
PVDF4, respectively.
2.2.2 Preparation of PVDF Fibers
Melt-blending of PVDF and CNFs was performed in a twin-screw co-rotating
extruder. The extruder temperatures were set from 170 ºC at the feed zone to 210 ºC at
metering zone. Screw rpm was 200. Substantial shear forces are necessary during the
composite processing step in order to disperse nanofibers in the polymer and to achieve
good mechanical and electrical properties [ 66F21]. Fiber pulling roll speed was set to 230
21
rpm to ensure mechanical stretching of fibers, which is anticipated to lead to orientation
of carbon nanofibers and polymer crystallites in the direction of the pulling [ 67F22].
Extruded fibers with an average diameter of 0.5 mm were obtained in product form.
2.2.3 39BDifferential Scanning Calorimetry (DSC)
The crystallization and melting behavior of PVDF/CNF compositions were
investigated by using the Perkin Elmer DSC 6 in a nitrogen atmosphere. Approximately
4 to 6 mg of sample was sealed in an aluminum pan. Heating and cooling scans was
performed at 10 ºC/min between 30 and 220 ºC. Samples were held at 220 ºC in the
molten state for 5 min to eliminate previous thermal history prior to cooling scan.
2.2.4 40BX-Ray Diffraction
The crystal structure of PVDF and composites were studied by wide-angle X-ray
diffraction (WAXD). The diffraction patterns were obtained with a Rigaku Ultima III using
CuKα radiation with a wavelength of 0.154 nm at 40 kV and 100 mA. Measurements
were made between 2θ values of 2º and 40º with a scan speed of 2 º /min.
2.2.5 41BMechanical Testing
Tensile tests were carried out on the extruded fiber samples with a TA Instruments
RSA III DMA in the tensile mode. The shapes of the samples were cylindrical with 50
mm gauge length and 0.5 mm diameter. The crosshead speed was set at 5 mm/min.
For each data point, three samples were tested, and the average value was taken.
2.2.6 42BDynamic Mechanical Measurements (DMA)
DMA was conducted on a TA Instruments RSA III under nitrogen using a heating
rate of 3 ºC/min. and a frequency of 1 Hz between -100 and 120 oC. A fiber sample with
a 0.5 mm diameter and 40 mm length was used.
22
2.2.7 43BScanning Electron Microscopy (SEM)
A high resolution SEM (FEI Nova 200 Dual Beam FIB/ FEGSEM) was used to
observe the dispersion of CNF in the PVDF matrix. The samples were dipped in liquid
nitrogen for 3 minutes and fractured. Gold coating was done on the fractured surface to
avoid overcharging of polymeric samples during SEM imaging. The gold coated surface
was imaged using beam of 1.7 nA at 5 kV of accelerating voltage.
2.2.8 44BMelt Rheology
Rheological measurements were carried out on a TA Instrument’s ARES strain-
controlled rheometer. For the rheological study, a 25 mm parallel plate setup was used.
Extruded fibers were used to prepare rheological disc samples having a diameter of 25
mm and a thickness of 2.5 mm in a compression press at 180 oC. Dynamic strain sweep
measurements were carried out at a frequency of 1 Hz, a temperature of 180 ºC, and a
strain of 0.1% to 100% to determine the linear viscoelastic region. The gap between the
two parallel plates was 0.051 mm.
2.3 18BResults and Discussion
2.3.1 45BCrystallization Behavior
DSC results of the pure PVDF and composites are summarized in Table 1. Melting
temperature (Tm), melting enthalpy (ΔHm), crystallization temperature (Tmc), melt
crystallization temperature (Tmc), and melt crystallization enthalpy (ΔHmc) were obtained
from the second-heating and second-cooling thermograms.
Figure 2increasin
Fig
a meltin
compos
observe
loading
remaine
2.1 DSC sng CNF con
gure 2.1 sho
ng tempera
itions (PV
d; but mel
of 4% res
ed lower tha
second-heantent
ows the DS
ature of 16
DF1 and
ting enthal
sulted in a
an the pure
ating curve
SC second-
8 ºC and
PVDF2),
pies (ΔHm)
significant
PVDF but
23
es showing
-heating cu
a melting
no signific
) dropped t
t increase
a little high
g an increa
urves for al
enthalpy o
cant chang
to 41.29 an
in melting
her than PV
ase in me
l compositi
of 58.08 J/g
ge in melt
nd 31.56 J
g to 170 ºC
VDF3 (42.03
lting point
ons. PVDF
g. For low
ting point
J/g. A high
C, but enth
3 J/g).
with
F has
-filler
was
-filler
halpy
Figure 2tempera
Fig
the com
crystalliz
crystalliz
145.25 º
for PVD
X
where, ∆
weight f
60%, 50
complem
2.2 DSC satures with i
gure 2.2 sh
mpositions
zation tem
zation temp
ºC, respect
F. The corr
Xc = [∆Hc/∆H
∆H0 is the
fraction of P
0%, 38%, a
ment the re
econd-coolincreasing
hows the D
increased
mperature
peratures (
tively. Thes
responding
H0] × Xm × 1
e enthalpy o
PVDF in th
and 47% for
esults of the
ling curvesCNF conte
DSC second
for all com
increased
Tmc) of PV
se tempera
degree of c
100%,
of 100% c
he composi
r PVDF, PV
e heating s
24
s showing nt
d-cooling cu
mposites r
with the
VDF and co
tures show
crystallinity
rystalline P
ite. The de
VDF1, PVD
scan and th
an increas
urves. The
relative to
increase
omposition
w that CNFs
y (Xc) was d
PVDF (105
egree of cry
DF2, PVDF4
he tendenc
se in melt
crystallizat
that of pu
in CNF
PVDF4 are
s act as nu
determined
J/ g) [ 68F23],
ystallinity w
4, respectiv
cy for the e
recrystalliz
tion point fo
ure PVDF.
content.
e 138.4 ºC
ucleating ag
by
, and Xm is
was found t
vely. The tr
enthalpy to
ation
or all
The
The
C and
gents
s the
to be
ends
drop
25
with concentration for PVDF1 and PVDF2 but to increase for PVDF4 is once again
observed. While the CNF enhanced the nucleation efficiency of the PVDF, the
crystallinity decreased with increasing concentration. With an unchanged melting point
but increased recrystallization temperature, the difference between melting and
recrystallization temperatures decreased. This indicates a reduced degree of
supercooling in the composites with the presence of CNF. The glass transition of PVDF
was undetected by DSC.
Table 2 DSC results from second-heating and second-cooling thermograms
Sample
ID
Tm (ºC) ∆Hm (J/g of
PVDF) Tmc
∆Hmc
(J/g)
Xc (%)
PVDF 168 58.081 138 -62.625 55.32
PVDF1 168 41.298 141 -52.068 39.33
PVDF2 168 31.568 144 -40.455 30.06
PVDF4 170 42.0395 145 -50.288 40.04
2.3.2 46BDynamic Mechanical Behavior
DMA was used to determine the dynamic mechanical properties of the samples in
which the sample is subjected to repeated small-amplitude strains in a cyclic manner.
The DMA Tg was found by examining the peak temperature of the tan δ (E/E) curve. E’
(storage modulus) is a measure of the energy stored elastically, whereas E (loss
modulus) is a measure of the energy lost. Tan δ, also called damping, indicates how
efficiently material loses energy to molecular rearrangements and internal friction.
26
Figure 2.3 (a) shows the β transition region of the PVDF and its composites. PVDF
shows a broad β relaxation related to side chain relaxation. For PVDF1 and PVDF2,
the curves overlap and indicate a slight decrease in the damping factor. A significant
decrease in peak height for the PVDF4 is obtained. In addition the β transition is shifted
toward a lower temperature indicated inhibited mobility. The α relaxation region is
depicted in Figure 2.3 (b). The Tg of PVDF is 40.16 ºC. As can be seen the glass
transition of the composites is shifted to higher temperatures (around 80 ºC) I note
however that PVDF4 does not exhibit a glass transition temperature within the range
investigated and the fiber compliance prevented additional data collection. The DSC
results on the decreased fractional crystallinity coupled to the increased glass transition,
indicates that the improved mechanical performance in the composite is best attributed
to the CNF presence.
Figure 2increasin
2.3 DMA reng CNF con
esults showntent (a) fo
wing a decrer the β tran
27
ease in tannsition, (b) f
δ peak hefor the α tra
eight and teansition
emperature with
28
2.3.3 47BX-Ray Diffractometry
WAXD was used to observe the effect of CNF content on the microstructure of
PVDF. Figure 2.4 shows the X-ray diffraction of PVDF and its composites. PVDF
reflections are located at 2θ = 17.8o (100), 18.6o (110), 19.8o (020), 26.62o (021), and
38.2o (002). These correspond to assignments for the α-phase crystal which has non-
polar trans-gauche-trans-gauche (TGTG) conformation. I note that the composite fibers
show retention of the α-phase crystal. No conversion to a β-phase is observed as
indicated by an absence of a peak at an angle of about 20.6º to 20.8º. The alpha phase
however does undergo a change with CNF presence. Two intense peaks at 17.8 and
18.6 observed in the PVDF merge into a single broad peak. I also note that the peak
intensity of the (020) reflection ratioed to the (110) reflection is approximately 2 for the
PVDF but drops to 1.5 in all composites. This ratio is retained when ratioing the (002)
peak intensity to the (020) reflection. I therefore conclude that the transformation of
crystal structure does not take place from α to β but the nature of the α phase is
affected by the presence of CNF. I predict that the crystal phase transformation did not
occur since the extruded fibers were cold stretched and quenched in water on exit from
the die.
Figure 2
2.3.4 48BM
The
PVDF c
those fo
microstr
pure PV
PVDF1,
increase
addition
2.4 X-ray dif
Mechanical
e tensile pr
composites
or most of
ructure’s co
VDF, PVDF
PVDF2, a
e in yield s
of 4 weigh
ffraction sp
Properties
roperties fo
offer impr
the polym
onversion i
F1 and PVD
and PVDF4
strength an
ht percenta
pectrograph
or extruded
roved tens
ers. At str
nto fibrillar
DF2. Upper
4 were 25,
nd an 88%
age. It is, th
29
s for PVDF
d fibers are
sile propert
ains past t
r morpholog
r yield stre
40, 47.5 a
increase
herefore, no
F and PVDF
shown in
ties. Tensil
the yield, a
gy extensiv
ngth and m
and 55.5 M
in modulus
ot surprisin
F/CNF fiber
Figure 2.5
le curves w
a broad pl
ve plastic
modulus va
MPa, respec
s were obs
g that the
rs
. It is clear
were simila
ateau indic
deformatio
alues for PV
ctively. A 1
served with
CNFs at h
r that
ar to
cates
n for
VDF,
22%
h the
igher
weight-p
break w
attribute
the amo
role in th
causing
less pro
toughne
Figure 2
percentage
was decrea
ed in part to
orphous ph
his behavio
a decrease
onounced
ess.
2.5 Stress s
loading le
ased for 4
o modificat
ase impos
or. The rest
e in ductilit
in both P
strain curve
eads to a h
4 weight-pe
ions in the
ed by the
riction sites
y and an in
PVDF1 an
s for PVDF
30
higher yield
ercentage
crystalline
unidirection
s prevent th
ncrease in s
nd PVDF2
F and its com
d stress. H
CNF filler
e fraction in
nally aligne
he polymer
stiffness. T
2 composit
mposites
owever, th
r content,
n the matrix
ed nanofibe
r from defor
This decrea
tion, indica
e elongatio
which can
x. Restrictio
ers plays a
rming, there
se in ductil
ating a hi
on at
n be
on of
a key
efore
lity is
igher
Figure 2dominat
2.6 SEM iting failure (
images of (a) PVDF1
PVDF co(b) PVDF2
31
mposites s (c) PVDF4
showing fib4
ber pull ouut mechannisms
Fig
that with
of matrix
interfaci
2.3.5 49BR
Figure 2
Fig
frequenc
80% str
region d
based o
complex
and (c)
frequenc
gure 2.6 sho
h increased
x on the na
al adhesion
Rheological
2.7 Dynamic
gure 2.7 sh
cy of 1 Hz.
rain; but w
decreases
on a const
x viscosities
respective
cy. At low
ows the SE
d concentra
anofibers co
n between t
Measurem
c strain swe
hows the d
The linear
with an incr
rapidly. Su
ant strain
s (*) of PV
ely. The sig
frequency,
EM images
ation of CN
oupled to po
the fibers a
ments
eep tests sh
dependence
r viscoelast
rease in th
ubsequent
of 0.1%. T
VDF and co
gnificant eff
PVDF/CN
32
of the frac
F, increase
ores indicat
and the PVD
howing stor
e of elastic
tic region fo
e percenta
frequency
The elastic
ompositions
fect of CN
F composi
ctured surfa
ed fiber pu
ting comple
DF.
rage modu
c modulus
or pure PV
age of CNF
sweeps w
moduli (G
s are compa
Fs can be
te melts ha
ace of the f
llout occurs
ete pullout,
lus as a fun
on strain
DF is very
Fs, the line
were there
G), loss m
ared in Fig
e seen, par
ave higher
fibers. It is
s. The abs
indicates l
nction of str
at 180 ºC
wide and u
ear viscoel
fore condu
oduli (G),
ures 2.8 (a
rticularly at
r elastic mo
clear
ence
ower
rain
at a
up to
lastic
ucted
and
), (b)
t low
oduli,
33
loss moduli, and complex viscosities compared with pure PVDF and show monotonic
increase with CNF content. It is conjectured that CNFs and PVDF interact and that
these structures become stronger with an increase in the percentage of CNF
concentration. At high frequency, however, the elastic moduli, loss moduli, and complex
viscosities of PVDF and its composites are similar, indicating matrix dominance. At low
frequency, a temporal structure is formed between CNFs and PVDF chains which is
strong enough to withstand the flow, resulting in the higher values of * at a low-
frequency region. At high frequency, flow destroys some of the structure, leading to a
decrease of viscosity.
Figure 2modulusºC temp
2.8 (a) Stors, G” versusperature
rage modus frequency
lus G’ versy at 180 ºC
34
sus frequenC temperatu
ncy at 180 ure, (c) Com
ºC tempermplex visco
rature, (b) osities η* at
Loss t 180
35
2.4 19BConclusion
I prepared PVDF/CNF composites by melt-blending and explored the potential of
CNFs as mechanical reinforcements in PVDF composite fibers. DSC showed that CNFs
decrease the fractional crystallinity in the composite. The increase in crystallization
temperature with relatively no change in melting point indicates decreased supercooling
in the composite. X-ray diffraction analysis indicated some change in α-phase
crystallites, but β-phase transformation did not occur. A decrease in peak tan δ for both
the α and β relaxation was observed. The transition temperature of the α relaxation
underwent a significant increase with the presence of CNF. The increased amorphous
fraction coupled to the absence of a β phase transformation is attributed to the use of a
quick quenching of the fibers in the cooling bath on exit from the extruder die. CNFs
however were found effective in improving mechanical properties. The addition of CNFs
results in an increase in ultimate tensile stress and modulus values of PVDF,
suggesting that nanofibers play an important role in enhancing the mechanical
properties of a polymer matrix. An increase in storage moduli, loss moduli, and melt
viscosities was observed with increased CNF concentration and was significantly
dependent on test frequency. I note that when 4% CNF were added to PVDF a
transition in stress-strain curves is observed together with slight increases in
crystallinity. I note that higher concentrations of CNF were not processable in the
extruder fiber die. This is reflected in the viscosity measurements which show values
>104 Pa-s for the PVDF4 concentration.
36
2.5 20BReference List
[1] Vivek B, Nath RJ. Phys D: Appl Phys 2001; 34: 667.
[2] Yu X, Rajamani R, Stelson KA, Cui T. Sensors and Actuators A 2006; 132: 626.
[3] Pantelis P. Phys Technol 1984; 15: 239.
[4] Ray HB, Changxing C, Anvar A, Zafar I, Josef NB, Geoff MS, Gordon GW, Alberto
M, Danilo DR, Andrew GR, Oliver J, Siegmar R, Miklos K. Science New Series 1999;
284: 1340.
[5] Gregorio RJ. J Appl Polym Sci 2006; 100: 3272.
[6] Mohammadi B, Yousefi AA, Bellah SM. Polymer Testing 2007; 26: 42.
[7] Nunes JS, Sencadas VW, Kholkin AL, Vilarinho PM, Lanceros-Méndez S. Mater
Res Soc Symp Proc 2007; 949: 0949-C03-02.
[8] Lovinger AJ. Poly(vinylidene fluoride), D. C. Bassett (Ed.), Developments in
Crystalline Polymers, Applied Science Publishers, London, 1982, Chap. 5.
[9] Lovinger J. Ferroelectric Polymers Science 1983; 220: 1115.
[10] Schaffner F, Jungnickel BJ. The Electric Moment Contribution to the Piezoelectricity
of PVDF IEEE Transactions on Dielectrics and Electrical Insulation, 1994; 553; 1-4.
[11] D’Souza NA, Ranade A, Strauss W, Hernandez-Luna A, Sahu L. Polymer
nanocomposite processing, Chap. 13, Handbook of Polymer Processing 2006,
Harper, C. Ed., Wiley).
[12] Owens FJ, Jayakody JRP, Greenbaum SG, Composites Science and Technology
2006; 66: 1280.
[13] Yu W, Zhao Z, Zheng W, Song Y, Li B, Long B, Jiang Q. Materials Letters 2008; 62:
747.
37
[14] Nam YW, Kim WN, Cho YH, Chae DW, Kim GH, Hong SP, Hwang SS, Hong SM.
Macromol Symp 2007; 478: 249-250.
[15] Wang M, Shi J-H, Pramoda K P, Goh SH. Nanotechnology 2007; 18: 235701.
[16] Dang Z M, Fan L Z, Shen Y, Nan CW. Mater. Sci and Eng 2003; B103: 140.
[17] Wang L, Dang Z-M. Appl Phy Letters 2005; 87: 042903.
[18] Yu X, Rajamani R, Stelson KA, Cui T. Sensors and Actuators A 2006; 132: 626.
[19] Feng J, Chan C-M. Carbon black filled immiscible blend of pol(vinylidene fluoride)
and high density polyethylene: Electrical properties and morphology. Annual
Technical Conference of the Society of Plastics Engineers, 1998.
[20] Tibbetts G, Lake ML, Karla L, Strong KL, Rice BP. Composites Science and
Technology 2007; 67(7-8): 1709-1718.
[21] Chen G-X, Li Y, Simizu H. Carbon 2007; 45 (12): 2334-2340.
[22] Mhetre SK, Kim YK, Warner SB, Patra PK, Katangur P, Dhanote A. Mat Res Soc
Symp Proc 2004 ; 788 : L11.17.1.
[23] Nakagawa K, Ishida Y. Polym. Sci. Phys 1973; 11: 2153.
38
CHAPTER 3
5BTIME DEPENDENT PIEZORESISTIVE BEHAVIOR OF POLYVINYLIDENE
FLUORIDE/CARBON NANOTUBE CONDUCTIVE COMPOSITES*
3.1 21BIntroduction
Over the last decade, polymer composite materials containing nanofiller
reinforcements have become a popular material for structural applications [ 69F1, 70F2].
Nanofillers such as carbon nanotubes offer multifunctional benefits of concomitant
strength and thermal/electrical conductivity [ 71F3- 72F73F74F75F7] enhancement leading to novel
multifunctional materials. The change in resistance due to change in strain have been a
reliable means of developing strain based sensors [76F8-77F78F10]. Piezoresistive ceramics
based on barium strontium titanate and lead zirconia titanate have been used effectively
but their reliability over time is limited by poor adhesion to the surface, their brittleness
and cost of manufacture. In contrast polymer carbon nanotube composites are easy to
make by melt blending. This leads to reduced cost, good mechanical strength and ease
of stress or strain monitoring.
The application of a stress to a nanotube based composite can be expected to give
resistance changes based on the extent of interchain contact throughout the matrix.
When a mechanical force is applied on such a composite, a morphological change in
network structure of the filler and polymeric matrix would take place leading to a change
in resistivity. In this paper I will focus on how the resistivity response is dependent on
stress and time.
* This entire chapter is reproduced from Shailesh Vidhate, Jaycee Chung, Vijay Vaidyanathan, Nandika
Anne D’Souza, “Time dependent piezoresistive behavior of polyvinylidene fluoride/carbon nanotube conductive composite”, Materials Letters 63 (2009) 1771–177 with permission from Elsevier.
3.2 22BBurg
Am
model [ 79F1
Fig. 3.1
constitut
respons
3.1. The
the sprin
Where,
element
dashpot
element
as:
gers Model
mong the n
11-80F12] is w
, the mode
tive equatio
e under co
e total strai
ng and dash
the subsc
ts respectiv
t, and Kelv
ts and the
numerous v
widely used
el consists
on for a B
onstant stre
n at time
hpot in the
cripts B, M
vely; M1, M
vin unit, re
initial cond
viscoelastic
to analyze
of a Maxw
Burgers mo
ess of each
e t is a sum
Maxwell m
M, and K
M2 and K a
espectively.
itions, the
39
c creep m
the viscoe
well and a K
odel can be
h coupled e
m of the str
odel are co
indicate B
are the stra
Consideri
total strain
odels, the
elasticity of
Kelvin unit
e derived
element in
rains in the
onsidered a
Burgers mo
ains of the
ing the co
n for Burge
Burgers o
materials.
connected
by conside
series as d
ese three e
as two elem
odel, Maxw
e Maxwell s
onstitutive r
rs model c
or four-elem
As illustrate
d in series.
ering the s
depicted in
elements, w
ments, thus:
well and K
spring, Max
relations of
can be obta
ment
ed in
The
strain
n Fig.
where
Kelvin
xwell
f the
ained
Figure 3
where E
respectiv
respectiv
3.3 23BExpe
3.3.1 50BM
The
as follow
tempera
Material
4 nm, le
without f
3.1 Schema
EM and M
vely; EK an
vely; 0 is t
erimental
Materials
PVDF used
ws: Density
ature 168
Science, w
ength 1-10 µ
further puri
atic diagram
are the m
nd are th
he initially a
d was supp
y: 1.78 g/cc
0C. MWCN
with 3-15 wa
µm and bu
fication. Pr
m of Burgers
odulus and
he modulus
applied stre
plied by Ark
c, MFI: 10
NTs (Baytu
alls, outer m
lk density 1
rior to melt
40
s model
d viscosity
s and viscos
ess; = K /
kema (Kyna
g/ 10 min
ubes® C15
mean diam
140-160 kg/
mixing both
of the Max
sity of the K
/ K is the re
ar® 721, pow
, Tensile S
50 P) wer
meter 13-16
/m3. MWCN
h the mater
xwell sprin
Kelvin sprin
etardation t
wder form)
Strength: 54
re obtained
nm, inner
NT were us
rials were v
g and dash
ng and dash
time.
with prope
4 MPa, Me
d from Ba
mean diam
sed as rece
vacuum drie
hpot,
hpot,
erties
elting
ayer®
meter
eived
ed at
41
150 0C for one hour. PVDF and MWCNT were dry mixed via tumbling in a bottle. The
content of MWCNT in PVDF powder was 10 wt %.
3.3.2 51BSample Preparation
MWCNT and PVDF were melt blended in twin screw extruder at 230°C and 200 rpm,
followed by a compression molding at 220 °C under 10 MPa for 10 min. to form a sheet
with smooth surfaces. After natural cooling to room temperature, the sheet was cut into
samples with a size of 25 × 25 × 3 mm3. Silver paste and copper mesh was mounted on
the both surfaces to make better electric contact.
3.3.3 52BMeasurements
The compression tests were performed on MTS 810 Material Test System, a
universal testing machine, in which the upper platen was fixed and the bottom platen
was mobile. A two-probe method was used to measure the resistance, as the resistance
of the highly conductive metal wires and contacts can be ignored. The compression test
was done at the speed of 0.5 mm/min. The axial compressive force and the
displacement data were automatically recorded in a computer. For creep testing under
compression, the specimen was compressed with a certain axial stress which was
maintained during the creep period. Compressive creep tests on composite samples
were performed under axial stresses of 20, 30, and 40 MPa. Fatigue tests were
conducted between 0 and 48 MPa in a triangle wave at 1 cycle per minute.
3.4 24BResults and Discussion
3.4.1 53BCompressive Stress and Resistance Response under Quasi-Static Loading
Depending on the % loading of the conductive filler in composite and the stress
level, positive pressure coefficient (PPC) or negative pressure coefficient (NPC) can be
observe
pressure
called N
compres
the resis
compos
modulus
compos
distance
in the re
exceeds
deforma
MWCNT
orientati
MWCNT
increase
Figure 3
d in piezo
e is called
NPC. As sh
ssive stress
stance befo
ite system
s is very hig
ite, the co
e of the MW
esistance o
s the yield
ation, the m
T and con
on of the
T and the d
e in the com
3.2 PPC and
oresistive m
PPC and
hown in Fi
s both PPC
ore loading
, MWCNT
gh (0.9 to 5
mpressibilit
WCNT. This
of the comp
stress, a P
matrix flow
sequently
MWCNT
destruction
mposite res
d NPC phe
materials. A
a decrease
ig. 3.2 at
C and NPC
and R is th
can be tr
5.5 TPa) [ 81F13
ty of the m
s forms clos
posite i.e. N
PPC effect
leads to a
a PPC ef
in the tra
of the cond
istance.
nomenon i
42
An increase
e in resista
10 wt. % o
C phenome
he resistanc
reated as
3]. When a
matrix leads
se conducti
NPC effect
can be se
an increase
ffect. Stres
ansverse d
ducting pat
n PVDF MW
e in resista
ance with a
of MWCNT
non was o
ce under th
incompres
a compress
s to a dec
ing paths, w
t. When the
een. As the
e in the inte
sses above
irection, b
th formed b
WCNT cond
ance with
an increase
T loading i
bserved. In
he loading c
sible since
ive stress i
crease in th
which resu
e magnitud
e matrix un
er-particle
e the yield
uckling or
by MWCNT
ductive com
an increas
e in pressu
in PVDF u
n Fig. 3.2,
condition. In
e their You
s applied to
he inter-pa
lt in a decr
de of the st
dergoes pl
distance o
d stress c
breakdow
T resulting i
mposite
se in
ure is
under
R0 is
n the
ung’s
o the
rticle
ease
tress
lastic
f the
ause
wn of
in an
43
3.4.2 54BCompressive Creep and Resistance under Transient Creep
For the creep test, the applied stress selected was in the NPC region since the
range of stresses showing PPC behavior was small. The plots of the creep strain versus
time at the different axial stress are shown in Fig. 3.3a together with the results of the fit
to the Burgers model. The fit parameters are provided in Table 3.1. With increasing
magnitude of constant stress, the elastic factor decreases for both the time independent
(EM) and the time dependent (EK) factors. The relaxation time also decreases with
increasing stress.
Table 3 Results of the Burger model.
Sample 0
(MPa)
E Compressive
(Mpa)
EM
(Mpa)
EK
(Mpa)
ηM
(Gpa h)
ηK
(Gpa h)
(h)
PVDF
10
20
3200
4264 20876 1394 532 25.5
30 4045 18647 1330 413 22.1
40 3954 15634 1273 302 19.3
The corresponding fractional resistance ∆R/ R0= R/R0 – 1, is shown in Figure 3.3b.
The resistance sharply decreased under the instantaneous application of the
compressive stress. Under constant load there was a negligible change in resistance
over time. This correlates to the increase in conductivity on load application and the
consequent decrease in resistance. The marginal change in resistance during the
constant stress shows that the material has a potential for sensing constant load with no
time dependent resistive response.
44
0 500 1000 1500 2000 2500 3000 3500 40000.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0.13
C
reep
Co
mp
lian
ce (
MP
a-1)
Time (Sec)
20 MPa Burgers fit at 20 MPa 30 MPa Burgers fit at 30 MPa 40 MPa Burgers fit at 30 MPa
0 400 800 1200 1600 2000 2400 2800 3200 3600-0.8
-0.6
-0.4
-0.2
0.0-0.8
-0.6
-0.4
-0.2
0.0-0.8
-0.6
-0.4
-0.2
0.0
0 400 800 1200 1600 2000 2400 2800 3200 3600
Time(Sec.)
20 MPa
R
/ R0
30 MPa
40 MPa
Figure 3.3 (a) Creep compliance versus time in compressive creep test (b) Change in
fractional resistance in creep test
(b)
(a)
45
3.4.3 55BCyclic Loading and Electric Resistance Response of Sample
Cyclic loading was applied within the elastic limit of the sample. During cyclic
loading, as shown in Fig. 3.4, the specimen resistance undergoes an increase and
decrease with stress on the sample. The decrease in resistance was observed with
increasing time at unloading condition due to time dependency of piezoresistance or
building up of some permanent residual stain after every cycle of loading.
Figure 3.4 Resistance response under cyclic loading
3.5 25BConclusions
Under quasi-static loading the PVDF-10% MWNT showed a PPC effect before the
yield stress and NPC behavior after the yield. This was related to the elastic matrix
response before yield and the plastic flow after yield. Under transient creep, the
resistance response during the instantaneous loading mimicked the response of the
‐0.8
‐0.7
‐0.6
‐0.5
‐0.4
‐0.3
‐0.2
‐0.1
0
‐30000
‐25000
‐20000
‐15000
‐10000
‐5000
0
0 100 200 300 400 500 600 700
Load
(N)
Time (Sec.)
∆R/R
0
46
tensile test but there was no time dependent resistance response under the constant
stress application. For cyclic fatigue, the stress-time response was synchronous with
the resistance but the peak resistivity decreased over time. This was related to residual
conductance during one cycle that built up with increasing cycles.
47
3.6 Reference List
[1] Veedu V P, Cao A, Li X, Ma K, Soldano C, Kar S, Ajayan P M, Ghasemi-Nejhad M
N. Nat. Mater. 2006; 5: 457.
[2] Suhr J, Koratkar N, Keblinski P, Ajayan P M. Nat. Mater. 2005; 4: 134.
[3] Baughman RH, Sakhidov AA, De Heer WA. 2002; 297: 787–92.
[4] Baughman RH. Science 1999; 284: 1340–4.
[5] Tahhan M, Truong VT, Spinks GM, Wallace GG. Smart Mater. Struct. 2003; 12:
626–32.
[6] Wood JR, Zhao Q, Frogley MD, Neurs ER, Prins AD, Peijs T, Dunstan DJ, Wagner
HD. Phys. Rev. B 2000; 62: 7571–5.
[7] Ajayan PM, Zhou O. Applications of carbon nanotubes Carbon Nanotubes
Synthesis, Structure, Properties and Applications, M S Dresselhaus, G Dresselhaus
and P Avouris 2001 (Berlin: Springer).
[8] Ponomarenko AT, Shevchenko VG, Klason C, Pristupa AI. Smart Mater Struct.1994;
3: 409.
[9] Xia HS, Wang Q. Chem. Mater. 2002; 14: 2158.
[10] Mei Z, Guerrero VH, Kowalik DP, Chung DDL. Polym. Compos. 2002; 23: 425.
[11] Findley WN, Lai JS, Onaran K, Creep and relaxation of nonlinear viscoelastic
materials: with an introduction to linear viscoelasticity. New York: Dover
Publications; Inc. (1989).
[12] Ranade A, Kasinath N, Debora F, D’Souza N. Polymer 2005; 46: 7323-33.
[13] Liu TT, Wang X, Phys. Lett. A, 2007; 365: 144-148.
48
CHAPTER 4
6BRESISTIVE-CONDUCTIVE TRANSITIONS IN TIME DEPENDENT
PIEZORESPONSE OF PVDF-MWCNT COMPOSITES*
4.1 27BIntroduction
Piezoresisivity is a phenomenon in which electrical resistance of a material
changes with applied stress or strain. This phenomenon can be employed to make
sensors which can monitor the change in the stress or strain of the material by
analyzing the electrical response of the material. Many researchers are applying this
phenomenon using various types of material systems like thermoplastics [3-6,8,12-18]
thermoset resins [7,21], cement [82F1] etc. with a range of fillers. Since the last decade,
polymeric composite materials containing nanofiller reinforcements have become
popular materials for structural application. Among the various types of nanofillers,
carbon nanotubes are dominant when conductivity is needed, as they provide high
strength and modulus [ 83F2,84F3] at low concentration. If polymer-carbon nanotube
composites provide strain sensing, then the conventional expensive electronic sensors
are not necessary. Polymer carbon nanotube composites are also easy to make by melt
blending based techniques such as extrusion and injection molding. This means
reduced cost, good mechanical strength and ease of strain monitoring can be realized.
The piezoresistive effect can be used to develop various strain sensors or self sensing
This entire chapter is reproduced from Shailesh Vidhate, Jaycee Chung, Vijay Vaidyanathan, Nandika
Anne D’Souza, “Resistive–conductive transitions in the time-dependent piezoresponse of PVDF-MWCNT nanocomposites”, Polymer Journal 42, 567–574 (2010). Reprinted with permission from Macmillan Publishers, Ltd.
49
composite structures and some mechanical damage based self-monitoring materials [ 85F4-
86F87F88F89F90F91F10].
By incorporating carbon nanotubes, multifunctional mechanical and electrical
response is facilitated. Increases in the mechanical strength and electrical conductivity
are simultaneously obtained [ 92F11- 93F94F95F96F97F16]. These materials have attracted a great deal of
scientific and commercial interest because they exhibit unique electrical and mechanical
properties in addition to some exclusive properties pertaining to polymeric materials
such as light weight, low cost, ease of processing, and corrosion resistance. When
mechanical force is applied to MWCNT composites, morphological changes in the
network structure of the filler in the polymeric matrix take place which leads to change in
resistivity measurements. Change in the resistance of the conductive composites is
mainly because of the change in inter-particle separation distance. Any process which
can change the particle to particle distance can change the resistivity response. For
example, by application of stress on a filled system can change particle to particle
distances. Also, depending upon the filler concentration loading level, time and stress
dependent changes can be observed. The increase in resistance with increase in
pressure is called as a positive pressure coefficient (PPC) phenomenon and the
opposite is a negative pressure coefficient (NPC) phenomenon.
Much work has been published on the study of various aspects related to
piezoresitivity of polymer and conducting filler composites. The main reason for the
piezoresistance is due to differences in compressibility of matrix and filler constituents,
material composition, load and filler content [ 98F17]. From the literature it can be inferred,
that with increase in concentration of the filler content, resistance of the material
50
decreases slowly up to the percolation threshold and decreases rapidly afterwards till
the conducting particles come in close contact with each other and after that remains
constant at very high filler concentration [ 99F18-100F101F102F21]. Mechanical strain due to tensile or
compressive stress also causes a remarkable change in resistance [103F22-104F105F106F25].
In this study, experimental results on PVDF/MWCNT conductive composites
have been demonstrated. Previous work on PVDF/MWCNT composites showed various
outstanding properties like, low percolation threshold for electrical conductivity [ 107F26],
improved piezoresitivity [108F27], improved strain sensing ability [ 109F28], good interfacial
adhesion between nanotubes and PVDF [ 110F29], improved crystallinity [17], etc. There are
very few reports on time dependent piezoresistive behavior of conducting composites.
For HDPE/ short carbon fiber conductive composite Q. Zheng et al. [111F30] suggested that
the molecular motion of the matrix due to creep brings about the local rearrangement of
the percolation network leading to the resistance creep and resistance relaxation. In
another paper Chen et al. [112F31] worked on HDPE/ graphite composites demonstrating
that there exists a critical stress above which resistivity increases with time and below
which resistivity decreases with time.
Although several research attempts have been made to understand the
piezoresistive behavior of the conducting material, questions still remain regarding a)
how the concentration of the conducting filler affects the time dependent behavior of the
material and b) how this time dependent behavior can be predicted based on a
mathematical model. Herein, I study the piezoresistive behavior of the conductive
composites using various concentrations of fillers using compression and creep tests.
An analogy between electrical and mechanical laws [ 113F32] is used to generate a model to
predict t
resistan
4.1.1 56BB
A
[ 114F33] is w
with a M
Burgers
of each
sum of t
model a
Where,
element
dashpot
element
A simila
laws of m
Based o
Based o
In whic
equiviva
time depen
ce respons
Burgers Mod
Among the
widely used
Maxwell an
model can
coupled ele
the strains i
re consider
the subsc
ts, respecti
t, and Kelv
ts strain in B
ar equation
mechanical
on Viscoela
on an Elect
h the load
alent to the
ndent piezo
se.
del
numerous v
to analyze
d a Kelvin
n be derived
ement in se
in these thr
red as two
cripts B, M
ively; M1,
vin unit, re
Burgers mo
can be w
l deformatio
stic continu
tric network
d stress
e electric p
oresistivity
viscoelastic
the viscoe
unit conne
d by consid
eries as dep
ree elemen
elements, t
M, and K
M2 andk
espectively.
odel can be
ritten by co
on and elec
uum:
k: I=Y.V; I=
is equiva
potential V;
51
with regard
c creep mo
lasticity of
ected in se
dering the s
picted in Fig
ts, where th
thus:
indicate B
k the strain
Consideri
e finally obta
onsidering
ctric circuits
C.V’
alent to th
; the mate
ds to the d
odels, Burg
materials, a
eries. The c
strain respo
g. 4.1. The
he spring a
Burgers mo
ns of the
ing the co
ained as fo
the analog
s [32]
he electric
rial modulu
deformation
ers or four-
as illustrate
constitutive
onse under
total strain
and dashpo
odel, Maxw
Maxwell s
onstitutive r
llows:
gy between
current I;
us E is eq
n and fract
-element m
ed in Fig. 4.
e equation
constant s
n at time
t in the Max
well and K
spring, Max
relations of
n the gover
; the strai
quivalent to
tional
model
.1 (b)
for a
tress
t is a
xwell
Kelvin
xwell
f the
rning
nis
o the
conduct
characte
Figure 4diagram
In
dashpot
dashpot
K / K.
ance Y; a
eristic C.
4.1 (a) A typm of Burgers
Fig. 4.1, E
t, respective
t, respective
and the v
pical creep s model and
M and M a
ely; EK and
ely; 0 is th
viscosity p
relaxation cd equivalen
are the mo
K are the
e initially a
52
parameter
curve of a vnt electric m
odulus and
e modulus a
pplied stres
is equ
viscoelasticmodel.
viscosity o
and viscosit
ss; Retarda
uivalent to
c material. (
of the Maxw
ty of the Ke
ation time is
the capa
(b) Schema
well spring
elvin spring
s defined a
acitor
atic
and
g and
s =
53
4.2 28BExperimental
4.2.1 57BMaterials
PVDF used was supplied by Arkema (Kynar® 721, powder form) with properties
as follows; Density: 1.78 g/cc, MFI: 10 g/ 10 min, Tensile Strength: 54 MPa, Melting
temperature 168 0C. MWCNTs (Baytubes® C150 P) were obtained from obtained from
Bayer® MaterialScience, with outer number of walls 3-15, outer mean diameter 13-16
nm, inner mean diameter 4 nm, length 1-10 mm and bulk density 140-160 kg/m3.
MWCNT were used as received without further purification. Prior to melt mixing both the
materials were vacuum dried at 150 oC for one hour. PVDF and MWCNT were dry
mixed via tumbling in a bottle. The contents of MWCNT in PVDF powder were 0, 1, 2, 4
and 10 weight percentage; and the compositions were coded as PVDF, PVDFCNT1,
PVDFCNT2, PVDFCNT4, and PVDFCNT10 respectively.
4.2.2 58BSample Preparation
MWCNT and PVDF were melt blended in twin screw extruder at 230 °C and 200
rpm, followed by a compression molding at 220 °C under 10 MPa for 10 min. to form a
sheet with smooth surfaces. After natural cooling to room temperature, the sheet was
cut into samples with a size of 25 × 25 × 3 mm3. Silver paste and copper mesh was
mounted on the both surfaces to make better electric contact. All experiments were
conducted in triplicate.
4.2.3 59BMeasurements
The compression tests were performed on MTS 810 Material Test System,
universal testing machine, in which the bottom platen was fixed and the upper platen
was mobile only along the uniaxial direction. The two-probe method was used to
54
measure the volume resistance by using Agilent 34410A multimeter. The two-probe
method is based on Ohm’s law, i.e., V = IR with V, I and R being voltage, current and
resistance between the two electrodes respectively. Since the copper electrode-material
resistivity difference was significant, a two point over four point measurement
configuration was found to be equitable. When connecting electrical multimeter to the
two ends of the conductive wires, a circuit is formed through conductive composite
sample, in which a direct current (DC) is produced by the power of the meter. Based on
the current and voltage and using Ohm’s law, a resistance was measured.
Figure mechan
T
volume
test, the
maintain
unloade
done at
4.2 Experical respon
The experim
resistance
e specimen
ned during
d to zero fo
the speed
rimental pnse measur
mental setu
measurem
n was com
the creep
orce and m
of 0.5 mm
rocedure rement tech
up for simu
ents is sch
mpressed w
period. Sim
aintained th
m/min. The
55
showing shnique.
ultaneous m
ematically
with a certa
milarly, dur
hrough rela
axial comp
sample pre
measureme
illustrated i
ain axial co
ring relaxat
axation peri
pressive fo
eparation
ents of stre
in Fig. 4.2.
onstant str
tion stresse
iod. Compr
orce and th
and elect
ess, strain,
For each c
ress which
ed sample
ression test
e displacem
trical,
and
creep
was
was
t was
ment
56
data were automatically recorded in a computer. The engineering stress was
determined as the ratio of the axial force to the cross-section area of the specimen, and
the compression stain was defined as X100%. Where, l is the deformed length
at time t. For creep tests, l0 is the axial length of the composite at the beginning of the
creep process, and ε represents the creep strain. Compressive creep test for all the
samples was performed under the constant stress below the maximum yield stress of
the sample.
4.2.4 60BRaman Spectroscopy
For recording Raman spectra, film samples of all compositions were used. All
Raman spectra were recorded on a Thermo Nicolet Almega XR Dispersive Raman
equipped with a microscope, through a 20 fold magnification objective, by co-adding
four spectra with collection times of 10 seconds each. Argon-ion laser of 514 nm
wavelength was used. The multiple grating that provides a resolution starting from 1000
cm-1 to 2800 cm-1 for the argon-ion laser was used. The abscissa was calibrated with
the 520.7 cm-1 peak of a silicon standard, and the sharp Raman shifts are accurate
within the limits of the resolution. To eliminate the influence of experimental parameters
all the compositions were measured on the same day.
4.3 29BResults and Discussion
4.3.1 61BCompression Test
Compression tests were performed to first determine the maximum yield stress
value of the sample. The compression test results for all compositions are shown in Fig.
4.3 (a) and (b). It is clear that with an increase in MWCNT content in PVDF,
compressive yield stress and modulus values increase. Yield stress values for PVDF,
PVDFCN
MPa res
Figure 4modulus
As
promine
stress,
resistan
NT1, PVDF
spectively.
4.3 (a) Cos values co
s shown in
ently in 1 a
both PPC
ce before
FCNT2, PV
ompressive mparison fo
n Fig. 4.4
and 2% MW
and NPC
loading an
VDFCNT4,
stress straor PVDF/M
4 at variou
WCNT com
C phenome
nd R is th
57
and PVDF
ain curvesWCNT com
us wt. %
mposites, F
enon was
he resistan
FCNT10 ar
. (b) Yieldmposites.
of MWCN
ig. 4.4 a a
observed.
nce under
re 28, 45,
d stress an
NT concen
and b) unde
In the fig
the loadin
50, 58, an
nd compres
tration (tho
er compres
gure, R0 is
g condition
d 80
ssive
ough
ssive
s the
n. In
58
composites, MWCNT can be considered as incompressible as their Young’s modulus is
very high (0.9 to 5.5 TPa) [ 115F34]. When the composite is compressed, the compressibility
of the matrix leads to a decrease in inter-particle distance of the MWCNT and close
conducting paths can result in decreasing the resistance of the composite i.e. NPC
effect, but upon further compression above yield stress results in plastic deformation
which may cause MWCNT slippage and an increase in interparticle distance. In
addition, at compressive stresses above yield, orientation of the MWCNT in transverse
direction, buckling or breakdown of MWCNT and the destruction of the conducting path
formed by MWCNT result in increasing the resistance of the sample at strains greater
than yield. This results in PPC behavior. When a higher concentration (4 and 10%) of
MWCNT, increased concentration leads to increased particle contact. Thus the slippage
or realignment of MWCNT does not influence the particle to particle distance and little to
no PPC behavior is evident (Fig. 4 c and d).
Figure 4composPVDFCN
4.4 Resistaites. PreseNT2, (c) PV
ance changence of PPVDFCNT4 a
ge as a funPC and Nand (d) PVD
59
nction of thPC pheno
DFCNT10 s
he applied omenon in showing on
pressure in(a) PVDF
nly NPC beh
n piezoresiFCNT1 andhavior.
istive d (b)
60
4.3.2 62BResistance Response for Creep and Relaxation of PVDF/MWCNT Composites
Transient tests were performed using the method of Fotheringham and Cherry [ 116F35]
which involves first stressing a sample and then immediately removing the applied
stress and allowing the sample to relax at zero stress. The specimens were loaded
using a mechanical test system using a stress ramp up rate of 0.5 MPa/minute. When
the sample reached a predetermined value of stress, 20 MPa, the stress was kept
constant for 1 hour. The strain continued to be monitored for 1 hour following creep.
Creep compliance was calculated by dividing obtained strain values by constant stress.
All the tests were performed at ambient temperature. As shown in Fig. 4.5, the effect of
sample composition on creep compliance can be seen. With increase in filler
concentration the creep compliance decreased. It can be seen that at higher % MWCNT
concentration, higher elasticity in the material causes relaxation to occur faster than the
lower % MWCNT filled samples. The fitting parameters corresponding to the
mechanical response are shown in Table 4.1. From the mechanical fits, I deduce that
the Maxwell initial elastic response trends are similar to the quasi-static elastic modulus
(Figure 4.3 b). The short term time dependence corresponding to the Kelvin element
shows similar trends for both elastic and viscous response with concentration. Long
term constant rate viscosities also increased with concentration.
Figure 5
Table 4
Sam
PVD
PVDF C
PVDF C
PVDF C
PVDF C
Sim
concent
concent
5.5 Experim
Results for
ple E
M
(MP
DF 350
CNT1 347
CNT2 364
CNT4 382
CNT10 426
milar to PP
ration of fi
ration, the
mental and p
r electric fit
M
a)
EK
(MPa) (G
09 8320
70 12320
43 13276
21 16982
64 20876
PC and N
ller wt. %
e observed
predicted cr
ηM
Gpa h)
ηK
(Gpa
746 180
945 324
875 360
1284 456
1394 532
PC, two n
in the PV
d resistan
61
reep compl
K
a h)
=
ηK/E
K
0 21.6
4 26.3
0 27.1 1
6 26.9
2 25.5 44
new pheno
VDF matrix.
ce increas
liance versu
Ym
(S)
Yk
(S)
9.8 3.4
19.5 3.2
3.4 3.3
429.5 1.1 -
omena wer
. At low w
sed with
us time cur
Cm
(µF)
Ck
(µF
11700 -88
59500 -13
-88200 221
-232000 4.4
re observe
wt. (1 & 2)
time and
rves
k
F)
R=Ck/Ym
1 -261.7
5 -41.8
1 66.6
4 4.1
d by chan
% of MW
the opp
nging
WCNT
posite
62
phenomenon was observed at high wt. % (4 & 10) concentration of MWCNT. The
corresponding resistance curves are shown in Fig. 4.6 b. I note that initial resistance
response is always conductive but that time dependent behavior led to concentration
dependent resistive/conductive response dependent on MWNCT concentration. At low
concentrations I would expect less particle to particle contact. With the applied
compressive stress, the particles are pushed and result in decreased interparticle
distance. Thus the resistivity initially decreases but the time dependent phenomenon is
more influenced by the polymer constrained between the conductive MWCNTs. The
corresponding fitting parameters show conductivity of the instantaneous response. The
transition to the long term response fits (Table 1) indicates negative capacitance.
Negative capacitances have been reported due to charge injection in mixtures of
materials with differences in resistivity of the constituents or when charges trapped at
the interface are released [ 117F36]. I note that the electrical resistance of the MWCNT is far
less than of PVDF matrix. At high concentration the gap between the MWCNT particles
is small enough for tunneling to occur, leading to formation of local conductive paths.
This has been associated with negative capacitances in semiconductor-conductor
mixtures. The time dependent polymer relaxation resulted in increased interparticle
distance and a long term capacitive/resistive response.
Figure 4change
At
material
injection
therefore
4.6 (a) Crein resistanc
higher conc
and theref
n due to th
e negative
ep compliace versus ti
centrations
fore a resis
he current b
capacitanc
ance versusime curves
, the increa
stive/capaci
being appl
ces towards
63
s time curvof PVDF/M
ased interpa
tive initial r
ied leads t
s the end of
ves and (bMWCNT co
article cont
response. O
to the form
f the test.
) simultanemposites.
tact results
Over long te
mation of a
eously reco
in a condu
erm, the ch
n electrets
orded
uctive
harge
and
Figure 4dependeconcent
Du
increase
correspo
4.7 Schemaent resistivrations.
ring recove
e in resistan
onds to the
atic showine response
ery followin
nce was ob
instantane
ng the effece at low co
ng creep, w
bserved at
eous respon
64
ct of MWConcentration
when the s
both high a
nse of the p
CNT-MWCNns and con
stress is re
and low wt
polymer. At
NT contact nductive re
moved to z
t. % of con
t low conce
leading to esponse at
zero, a sud
centration.
entrations w
time high
dden
This
where
65
resistivity is dominant, the remnant resistivity remains relatively time independent with
low time dependence. This indicates that the lack of particle to particle contact is
responsible. The resistive response persists when the load is removed and the particle
to particle distances are relatively unchanged as the material recovers. At higher
concentrations, the interparticle distances that were shortened during the application of
compressive loads coupled to increased filler-filler contact caused increased
conductivity. Dimensional recovery results in an increase in junction distances causing a
resistivity increase. With time however, the polymer recovery leads to a gradual
relaxation of the polymer matrix causing a decrease in resistivity as the material
approaches its architecture prior to the application of stress. The difference between the
low and high concentrations is shown in the schematic in Fig. 4.7. At low
concentrations, under the applied stress, the matrix between the nanotubes is the active
piezoresponsive element. There are fewer MWCNT-MWCNT contacts and the response
is largely resistive. At higher concentrations, more MWCNT-MWCNT contact leads to a
conductive response. The fitted results for the electrical analogy are shown in Fig. 4.8.
The schematic of the transition from resistive to conductive response is replicated in the
model through the annihilation of the capacitive response for the high concentrations.
The model fits show the potential for this applied functionality in sensor deployment.
Figure 4
To
Fig. 4.9
the inten
several
used to
2470 cm
intensitie
in the in
increase
et. al. [118F3
concent
4.8 Experim
understan
b, the MWC
nse peaks
tangential C
correlate d
m-1 is assoc
es in Rama
ntensity and
e in the con
37] Similarl
ration. In R
mental and p
d the role
CNT peaks
is the G-ba
C-C stretch
dispersion
ciated with
an spectros
d area of th
ncentration
ly the PVD
Raman line
predicted re
of dispersi
s are found
and of the
hing transiti
of nanotub
the PVDF
scopy are p
he G band
of the MW
DF band pe
mapping (
66
esistance c
on I utilize
at 2700 cm
MWCNT at
ions of the
bes in the p
F band whic
proportiona
peak, G’, a
WCNT which
eak area re
(Fig. 4.9 c,
reep versus
e Raman sp
m-1, 1590 cm
t 1590 cm-
MWCNT c
polymer ma
ch arises f
l to MWCN
and D band
h is also re
educed with
, d, e, and
s time curv
pectroscop
m-1 and 13
-1 which is
carbon atom
atrix. The i
from C-H s
NT concent
d peaks ca
ported by C
h an increa
f) a multi
ves
y. As show
24 cm-1. On
associated
ms. This ca
intense pea
stretching. P
ration. Incr
an be seen
C. G. Salzm
ase in MW
spectrum f
wn in
ne of
with
an be
ak at
Peak
rease
with
mann
WCNT
file is
67
acquired. Each spectrum represents the Raman response of the composition present at
the point where laser is focused. The image was obtained by integrating over all Raman
lines. Using both images and the spectral data at different locations can be clearly
linked to distribution of MWCNT in the polymer matrix. Examining the G band as a
function of distance it can be seen that low dispersion exists at 1% corresponding to the
least particle to particle contact and highest resistance. For 2 and 10% well dispersed
particles are evident with periodic peaks and valleys showing spatial uniformity (albeit
with increased concentration the period is smaller). For 4% I see agglomeration as the
concentration transitions from no particle to particle concentration to a continuous
network.
Figure 4large arindicatedpositionsPVDFCN
4.9 Raman ea of sampd in norms along thNT10.
spectra of ple. (b) The
malized spehe line for
PVDF/CNTe peaks ariectra. (c)
PVDFCNT
68
T compositeising from CRaman linT1, (d) PV
es using (aC-MWNTs
ne mappingVDFCNT2,
a) line mapp(D, G, and
g spectra (e) PVDF
ping to exad G’ bands
acquired FCNT4, an
mine ) are from d (f)
69
4.4 30BConclusions
A transition from NPC to PPC behavior for materials in compression was
determined. The extent of the PPC was related to the degree of particle to particle
contact and was thus a function of the material response. A novel transition from
electrically resistive to conductive time dependent response was determined in MWCNT
modified PVDF with concentration. The magnitude of the resistance change could be
similar for high and low concentrations indicating that resistive based piezoresponse
can also be considered without requiring particle-particle contact in a non
piezoresponsive polymer. An electrical and mechanical analog for time dependent
viscoelastic response was developed to describe the resistive response. As shown, the
result from the model agrees well with the experimental data and the creep analysis of a
viscoelastic continuum can be easily carried out by means of the electric analogy.
Raman spectroscopy was ideally suited for the investigation of MWCNT dispersion
since the laser probe can interrogate large surface area. Of longer range ramifications,
the results in this investigation show that even a resistive piezoresponse is of value
because the magnitude of the change in resistance whether conductive or resistive is a
valuable parameter in correlating stress effects.
70
4.5 31BReference List
[1] Vipulanandan C, Garas V. J. of Materials in Civil Engg 2008; 20: 93-101.
[2] Veedu V, Cao A, Li X, Ma K, Soldano C, Kar S, Ajayan P. Ghasemi-Nejhad M. Nat.
Mater. 2006; 5: 457-62.
[3] Suhr J, Koratkar N, Keblinski P, Ajayan P. Nat. Mater. 2005; 4: 134-37.
[4] Ponomarenko A, Shevchenko V, Klason C, Pristupa A. Smart Mater Struct. 1994; 3:
409-15.
[5] Xia H, Wang Q. Chem. Mater. 2002; 14: 2158-65.
[6] Mei Z, Guerrero V, Kowalik D, Chung D. Polym. Compos. 2002; 23: 425-32.
[7] Song Y, Zheng Q, Yi X. J. Polym. Sci., B: Polym. Phys. 2004; 42: 1212-17.
[8] Yamaguchi K, Busfield J., Thomas AJ. Polym. Sci., B: Polym. Phys. 2003; 41: 2079-
89.
[9] Sau K, Khastgir D, Chaki, T. Die Angewandte Makromolekulare Chemie 1989; 258:
11-17.
[10] Pramanik P, Khastgir D, De S and Saha T. J. Mater. Sci. 1990; 25: 3848-53.
[11] Baughman R, Zakhidov A, Heer W. Science 2002; 297:787-92.
[12] Baughman R, Cui C, Zakhidov A, Iqbal Z, Barisci J, Spinks G, Wallace G, Mazzoldi,
A, De Rossi D, Rinzler A, Jaschinski O, Roth S, Kertesz M. Carbon Nanotube
Actuators. Sci. 1990; 284: 1340-44.
[13] Tahhan M, Truong V, Spinks G, Wallace G. Smart Mater. Struct. 2003; 12: 626-32.
[14] Wood J, Zhao Q, Frogley M, Meurs E, Prins A, Peijs T, Dunstan D, Wagner H.
Carbon Nanotubes: From Molecular to Macroscopic Sensors. Phys Rev. B. 2000;
62: 7571-5.
71
[15] Peng S, O’Keeffe J, Wei C, Cho K, Kong J, Chen R. Carbon Nanotube Chemical
and Mechanical Sensors. Prceeding, 3rd Int. Workshop on Structural Health
Monitoring, Stanford, CA, pp. 1-8 (Sep. 2001).
[16] Ajayan P, Zhou O. in Carbon Nanotubes Synthesis, Structure, Properties and
Applications (eds. Dresselhaus, M., Dresselhausand, G., Avouris, Ph.) Applications
of Carbon Nanotubes, 391-425 (Springer, 2001).
[17] Kang I, Schulz M, Kim J, Shanov V, Shi D. Smart Mater. Struct. 2006; 15: 737-48.
[18] Nam Y, Kim W, Cho Y, Chae D, Kim G, Hong S, Hwang S, Hong S. Macromol.
Symp. 2007; 249: 478-484.
[19] Qu S, Wong S. Composites Sci. and Tech. 2007; 67: 231-7.
[20] Wu M, Shaw L. J. of Power Sources 2004; 136: 37-44.
[21] Xiao M, Sun L, Liu J, Li Y, Gong K. Polymer 2001; 43: 2245-8.
[22] Taya M, Kim W, Ono K. Mechanics of Mat. 1998; 28: 53-9.
[23] Mattmann C, Clemens F, Troster G. Sensors 2008; 8: 3719-32.
[24] Wang X, Chung D. J. of Material Research 1999; 14: 4224-29.
[25] Zhang X, Pan Y, Zheng Q, Yi X. Journal of Polymer Science. B. 2000; 38: 2739-49.
[26] Dang Z, Wu J, Xu H, Yao S, Jiang M, Bai J. Applied Physics Lett. 2007; 91: 072912
1-3.
[27] Kim J, Loh K, Lynch J. Piezoelectric Polymeric Thin Films Tuned by Carbon
Nanotube Fillers. Proceedings of SPIE –15th Annual International Symposium on
Smart Structures and Materials, 693232-1 (2008).
[28] Laxminarayan K, Jalili N. Text. Res. J. 2005; 75: 670-80.
72
[29] Owens F, Jayakody J, Greenbaum S. Composites Sci. and Tech. 2006; 66: 1280-
84.
[30] Zheng Q, Zhou J, Song Y. J. Mater. Res. 2004; 19: 2625-34.
[31] Chen G, Lu J, Lu W, Wu D, Wu C. Polym. Int. 2005; 54: 1689-93.
[32] Scotta R, Vitaliani R. Journal of Structural Engineering. 1996; 122: 1118-21.
[33] Findley W, Lai J, Onaran K. Creep and Relaxation of Nonlinear Viscoelastic
Materials: With an Introduction to Linear Viscoelasticity, Ch. 5, 50-70 (Dover
Publications, Inc., New York, 1989).
[34] Liu T, Wang X. Phys. Lett. A. 2007; 365: 144-8.
[35] Fotheringham D, Cherry B. Mater. Sci. Lett. 1976; 11: 1368–1371.
[36] Matsumura M, Hirose Y. Jpn. J. of Appl. Phys. 2000; 39: L123-L125.
[37] Salzmann C, Chu B, Tobias G, Llewellyn S, Green M. Carbon. 2007; 45: 907.
73
CHAPTER 5
SUMMARY
The objective of this thesis was to estabilish the viability of using particle dispersed
polymers as piezoresponsive materials using the electrical properties of the filler.
Carbon is a conductive filler and two particles were investigated: CNFs with a dimension
of 107 nm average diameter and 50-100 µm length and electrical resitivity of 1000 micro
ohm-cm and carbon nanotubes with 3-15 walls 3-15, outer mean diameter 13-16 nm,
inner mean diameter 4 nm, length 1-10 µm and 2-4 micro ohm-cm resistivity. To
summarize we can state:
5.1 Effect of Carbon Nanofibers on Thermo-Mechanical Properties of PVDF
The DSC results showed the decrease in fractional crystallinity of PVDF/CNFs
composites. The increased crystallization temperature with relatively no change in
melting point indicated decreased supercooling in the composite. X-ray diffraction
analysis indicated some change in α-phase crystallites, but β-phase transformation did
not occur due to weak interface formation between CNFs and PVDF matrix and lack of
sufficient chain orientation. DMA results indicated decrease in tan δ values for both the
α and β relaxation. However, CNFs were found effective in improving mechanical
properties. Improved ultimate tensile stress and modulus values of PVDF were
observed by addition of CNFs. In rheological investigation an increase in storage
moduli, loss moduli, and melt viscosities were observed with increased CNF
concentration.
74
5.2 PVDF/CNTs Nanocomposites’ Time Dependent Piezoresistive Effect
NPC and PPC behavior was observed under quasi static loading conditions for
PVDFCNT10 nanocomposite. Elastic matrix response caused the NPC effect and the
plastic deformation of composites resulted into PPC effect. Under transient creep, the
resistance response during the instantaneous loading mimicked the response of the
tensile test but there was no time dependent resistance response under the constant
stress application. For cyclic fatigue, the stress-time response was synchronous with
the resistance but the peak resistivity decreased over time.
5.3 Resistive To Conductive Transition in PVDF/CNTs Nanocomposites
A novel transition from electrically resistive to conductive time dependent response
was determined in MWCNTs modified PVDF. The magnitude of the resistance change
could be similar for high and low concentrations indicating that resistive based
piezoresponse can also be considered without requiring particle-particle contact in a
non piezoresponsive polymer. An electrical and mechanical analog for time dependent
viscoelastic response was developed to describe the resistive response. Also, the
developed electrical model based on Burgers model agreed well with the experimental
data. Raman spectroscopy was successfully employed to observe the dispersion of
MWCNTs in PVDF nanocomposites andto correlate the results with electrical resistivity
response.
In conclusion, the electrically conductive PVDF/CNTs nanocomposites have
shown the prospective to be used as a stress or strain sensor for real time structural
health monitoring.