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.