 

 

 

 

 

 

 

 

 

 

 

 

 

.. _'

....... .ruv .

mm. mm
.....,

I
4-

m"

.41.
. «u

,1
1'
r

,1‘
5.
5
r

u} ~ 3;! 13!;
" "1.55431; » s
£1,123!“ . .‘ yl ‘
is J .1?“ ‘ A
1%?“
, D

1? .55

 

10°C

 

LIBRARY
Mrcnigan State
University

 

 

 

This is to certify that the
dissertation entitled

MICROWAVE PROCESSING OF EPOXY RESINS AND
SYNTHESIS OF CARBON NANOTUBES BY MICROWAVE
PLASMA CHEMICAL VAPOR DEPOSITION

presented by

LIMING ZONG

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemical Engineering

 

 

7V2 w": 50%“67

 

Major Professor’s Signature
”7%; / p/ a2 &0 (-

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RE1URN BOX to remove this checkout from your record.
1‘0 AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 a/CIW.m—p'.1s"

 

MICROWAVE PROCESSING OF EPOXY RESINS AND SYNTHESIS OF CARBON
NAN OTUBES BY MICROWAVE PLASMA CHEMICAL VAPOR DEPOSITION

By

Liming long

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemical Engineering and Materials Science

2005

ABSTRACT

MICROWAVE PROCESSING OF EPOXY RESINS AND SYNTHESIS OF CARBON
NANOTUBES BY MICROWAVE PLASMA CHEMICAL VAPOR DEPOSITION

By
Liming Zong

Microwave processing of advanced materials has been studied as an attractive
alternative to conventional thermal processing. In this dissertation, work was preformed
in four sections. The ﬁrst section is a review on research status of microwave processing
of polymer materials.

The second section is investigation of the microwave curing kinetics of epoxy
resins. The curing of diglycidyl ether of bisphenol A (DGEBA) and 3, 3'-
diaminodiphenyl sulfone (DDS) system under microwave radiation at 145 °C was
governed by an autocatalyzed reaction mechanism. A kinetic model was used to describe
the curing progress.

The third section is a study on dielectric properties of four reacting epoxy resins
over a temperature range at 2.45 GHz. The epoxy resin was DGEBA. The four curing
agents were DDS, Jeffamine D-230, m-phenylenediamine, and diethyltoluenediamine.
The mixtures of DGEBA and the four curing agents were stoichiometric. The four
reacting systems were heated under microwave irradiation to certain cure temperatures.
Measurements of temperature and dielectric properties were made during free convective
cooling of the samples. The cooled samples were analyzed with a Differential Scanning
Calorimeter to determine the extents of cure. The Davidson-Cole model can be used to

describe the dielectric data. A simpliﬁed Davidson-Cole expression was proposed to

calculate the parameters in the Davidson-Cole model and describe the dielectric
properties of the DGEBA/DDS system and part of the dielectric data of the other three
systems. A single relaxation model was used with the Arrhenius expression for
temperature dependence to model the results. The evolution of all parameters in the
models during cure was related to the decreasing number of the epoxy and amine groups

in the reactants and the increasing viscosity of the reacting systems.

The last section is synthesis of carbon nanotubes (CNTs) on silicon substrate by
microwave plasma chemical vapor deposition of a gas mixture of methane and hydrogen.
The catalyst was nickel, which was not directly deposited on the substrate but migrated
from catalyst supplier during microwave plasma pretreatment. The Si wafer was coated
with amorphous carbon before synthesis. Additional heating sources and DC bias on
graphite substrate were not employed. Scanning Electron Microscopy and Transmission
Electron Microscopy were used to characterize the morphologies and microstructures of
the synthesized CNTs. The lengths and diameters of the CNTs changed with gas
composition and growth temperature. Vertically-aligned CNTs with a length range of
350-500 pm were synthesized. The diameter of CNTs is around 30-60 nm. The plasma
gases included 20 sccm methane and 80 sccm hydrogen. The growth temperature was
800-810°C and the growth time was 20 minutes. The CNTs exhibit bamboo-like structure

and appear to grow via a root-grth mechanism.

Copyright by
LIMING ZONG

2005

ACKNOWLEDGEMENTS

The author would like to thank his academic advisor, Dr. Martin C. Hawley, for
his invaluable guidance during this research. Appreciation is also given to the other Ph.D.
committee members, Dr. Krishnamurthy J ayaraman, Dr. Michael E. Mackay, Dr. Leo C.

Kempel, and Dr. Gregory L. Baker, for their time and insightful suggestions.

The author is grateﬁrl to Dr. Shuangjie Zhou and Rensheng Sun for their countless
help during the research. Acknowledgement is extended to Michael Rich for help in DSC
operation and Gregory Charvat for building the dielectric measurement and heating
switch system. Furthermore, Dr. Jes Asmussen, Stanley S. Zuo, Nikki A. Sgriccia, and
Susan Farhat helped the author in experiments on carbon nanotubes. Ewa Danielewicz
and Dr. Xudong Fan helped the author in SEM and TEM imaging at the Center for
Advanced Microscopy at Michigan State University. Yijun Zhang and Dr. David

Grummon made the nickel coated silicon substrates for the CNT synthesis.

Last but not the least, special thanks are expressed to the author’s family members

for their endless love and support that words cannot describe.

This work was supported by the National Science Foundation under contract

number DMI-0200346.

TABLE OF CONTENTS

LIST OF FIGURES ................................................................................... x
LIST OF TABLES ................................................................................... xv
CHAPTER 1 INTRODUCTION ...................................................................................... 1
CHAPTER 2 BACKGROUND ON MICROWAVE PROCESSING .............................. 4
2.1 ELECTROMAGNETIC FIELDS IN A MICROWAVE ENCLOSURE ........... 4
2.1.1 Maxwell’s Equations .................................................................................. 4
2.1.2 Resonant Modes in a Cylindrical Single Mode Cavity ............................... 6
2.1.3 Electromagnetic Fields in a Cylindrical Single Mode Cavity .................... 9

2.2 MICROWAVE/MATERIALS INTERACTIONS ........................................... 10
2.2.1 Mechanisms of Microwave/Materials Interactions ................................... 10
2.2.2 Dielectric Properties .................................................................................. 13

2.3 LITERATURE SURVEY ................................................................................. 16
2.3.1 Polymer Dielectric Properties ................................................................... 17
2.3.2 Microwave-Assisted Polymer Processing ................................................. 19
2.3.2.1 Thermosets ................................................................................................. 19
2.3.2.2 Reaction Rate and Kinetics ........................................................................ 19
2.3.2.3 Properties ................................................................................................... 21
2.3.2.4 Monitor and Heat Transfer Model ............................................................. 22

2.3.2.5 Thermoplastics ........................................................................................... 24
2.3.2.6 Composites Bonding .................................................................................. 24

2.3.3 Microwave-Assisted Polymer Synthesis ................................................... 25
2.3.4 Microwave Plasma Modiﬁed Polymer Surface ........................................ 28
2.3.5 Microwave Plasma Polymerization .......................................................... 34
2.3.6 Polymer Degradation ................................................................................ 35

vi

2.3.7 Nanomaterials ........................................................................................... 36

2.3.8 Applications .............................................................................................. 36
CHAPTER 3 MICROWAVE CURING MECHANISM OF EPOXY RESIN S ............ 38
3. 1 INTRODUCTION ............................................................................................ 3 8
3.2 MECHANISMS OF EPOXY RESIN CURING REACTIONS ....................... 39
3.3 EPOXY RESIN CURING KINETICS ............................................................. 42
3.3.1 Thermal Curing ......................................................................................... 42
3.3.1.1 Mechanistic Models ................................................................................... 42
3.3.1.2 Phenomenological Models ......................................................................... 45
3.3.1.3 Microwave Curing Kinetic Model ............................................................. 46
3.3.2 Kinetic Model Used in This Study ............................................................ 46
3.4 EXPERIMENTAL ............................................................................................ 47
3 .4.1 Experimental Equipment .......................................................................... 47
3.4.1.1 Experimental Circuit .................................................................................. 47
3.4.1.2 Temperature Sensing Systems ................................................................... 48
3.4.1.3 Microwave Applicators .............................................................................. 49
3.4.1.4 Cavity Characterization and Process Control ............................................ 51
3.4.2 Experimental Materials and Procedure ..................................................... 53
3.5 RESULTS AND DISCUSSION ....................................................................... 55
3.5.1 Temperature and Power Deposition Proﬁles ............................................ 55
3.5.2 Kinetics ..................................................................................................... 58
3.6 CONCLUSIONS ............................................................................................... 60
CHAPTER 4 DIELECTRIC ANALYSIS OF CURING EPOXY RESINS ................... 62
4. 1 INTRODUCTION ............................................................................................ 62
4.2 BACKGROUND ON DIELECTRIC ANALYSIS .......................................... 64
4.2.1 Fundamental Theories for Dielectric Relaxation ...................................... 64
4.2.1.1 Complex Dielectric Constant ..................................................................... 64
4.2.1.2 Typical Graphs ........................................................................................... 69
4.2.1.3 Dielectric Relaxation Time ........................................................................ 7O

vii

4.2.1.4 Typical Dielectric Relaxation Processes .................................................... 71

4.2.2 Literature Review ...................................................................................... 72
4.3 EXPERIMENTAL ............................................................................................ 74
4.3.1 Experimental Systems ............................................................................... 74
4.3.2 Experimental Materials ............................................................................. 75
4.3.3 Sample Preparation ................................................................................... 76
4.3.4 Measurements ........................................................................................... 77
4.4 RESULTS AND DISCUSSION ....................................................................... 78
4.4.1 DGEBA/DDS System ............................................................................... 79
4.4.2 DGEBA/Jeffamine D-230 System ............................................................ 95
4.4.3 DGEBA/mPDA System .......................................................................... 105
4.4.4 DGEBA/Epikure W System ................................................................... 114
4.4.5 Parameters in the Models for the Four Systems ..................................... 121
4.5 CONCLUSIONS ............................................................................................. 127
CHAPTER 5 SYNTHESIS OF CARBON NAN OTUBES BY MPCVD .................... 130
5.1 INTRODUCTION .......................................................................................... 130
5.2 LITERATURE REVIEW ............................................................................... 134
5.3 EXPERIMENTAL .......................................................................................... 138
5.3.1 Experimental System .............................................................................. 138
5.3.2 Experimental Materials ........................................................................... 140
5.3.3 Experimental Procedure .......................................................................... 140
5.4 RESULTS AND DISCUSSION ..................................................................... 142
5.4.1 CNT Grth Conditions ......................................................................... 142
5.4.2 CNT Growth Results ............................................................................... 145
5.4.3 Morphology of CNTs by SEM ............................................................... 146
5.4.4 TEM Results ........................................................................................... 149
5.4.5 CNT Growth Mechanism ........................................................................ 151
5.5 CONCLUSIONS ............................................................................................. 1 54
CHAPTER 6 CONCLUSIONS ..................................................................................... 155

viii

CHAPTER 7 FUTURE WORK .................................................................................... 159

REFERENCES ............................................................................................................... 160

LIST OF FIGURES

Figure 2.1 TE modes in an empty cavity with a diameter of 17.78 cm .............................. 8
Figure 2.2 TM modes in an empty cavity with a diameter of 17.78 cm ............................. 8
Figure 3.1 Three-step epoxy curing mechanism ............................................................... 40
Figure 3.2 Uncatalyzed and autocatalyzed epoxy curing reaction mechanisms ............... 41
Figure 3.3 Circuit of the microwave curing system .......................................................... 48
Figure 3.4 A cylindrical single-mode resonant cavity ...................................................... 50
Figure 3.5 Mode spectrum of the loaded cavity for microwave curing ............................ 52
Figure 3.6 Electrical ﬁeld of TM 020 mode ..................................................................... 52
Figure 3.7 Electrical ﬁeld of TM 021 mode ..................................................................... 53
Figure 3.8 Temperature proﬁles during microwave cure at 145 °C ................................. 55
Figure 3.9 Power proﬁles during microwave cure at 145°C ............................................. 56
Figure 3.10 Extent of cure vs. curing time ........................................................................ 59
Figure 4.1 Schematic diagram of the Debye model .......................................................... 65
Figure 4.2 Schematic diagram of the Davidson-Cole model ............................................ 67
Figure 4.3 Schematic diagram of Cole-Cole plot of the Debye model ............................. 69
Figure 4.4 Schematic diagram of Cole-Cole plot of the Davidson-Cole model ............... 69

Figure 4.5 Schematic illustration of the microwave processing and diagnostic system... 75

Figure 4.6 Dielectric properties vs. temperature for DGEBA, DDS, and uncured
DGEBA/DDS mixture ................................................................................... 79

Figure 4.7 Temperature dependence of dielectric properties for the DGEBA/DDS epoxy

resins at different extents of cure (%) ............................................................ 81

Figure 4.8 DGEBA resin curing mechanism .................................................................... 82
Figure 4.9 e" vs. 8' for the DGEBA/DDS epoxy resins at different extents of cure (%).. 87
Figure 4.10 In (6-8...) vs. lOOO/T for the curing DGEBA/DDS system ............................ 88

Figure 4.11 In (8") vs. 1000/1‘ of the curing DGEBA/DDS system ................................. 88

Figure 4.12 Comparison between the experimental and calculated dielectric properties of

the curing DGEBA/DDS system ................................................................... 89
Figure 4.13 Cole-Cole plots for the curing DGEBA/DDS system ................................... 90
Figure 4.14 n and (so-s...) vs. extent of cure for the curing DGEBA/DDS system ........... 92
Figure 4.15 Ea vs. extent of cure for the curing DGEBA/DDS system ............................ 93
Figure 4.16 I vs. 1000/T for the curing DGEBA/DDS system ........................................ 94

Figure 4.17 Dielectric properties vs. temperature for DGEBA, J effarnine D-230, and
uncured DGEBA/Jeffamine D-230 mixture .................................................. 95

Figure 4.18 Temperature dependence of dielectric properties for the curing

DGEBA/Jeffamine system ............................................................................ 97
Figure 4.19 8" vs. 8' for the curing DGEBA/Jeffamine system ........................................ 98
Figure 4.20 In (858...) vs. lOOO/T for the curing DGEBA/Jeffamine system .................... 99
Figure 4.21 1n (8") vs. lOOO/T for the curing DGEBA/Jeffamine system ........................ 99
Figure 4.22 Cole-Cole plots for the curing DGEBA/Jeffamine system ......................... 101

Figure 4.23 n and (so-em) vs. extent of cure for the curing DGEBA/Jeffamine system . 102

Figure 4.24 E21 vs. extent of cure for the curing DGEBA/Jeffamine system .................. 103

Figure 4.25 I vs. 1000/T for the curing DGEBA/Jeffamine system ............................... 104

xi

Figure 4.26 Dielectric properties vs. temperature for DGEBA, mPDA, and uncured

DGEBA/mPDA mixture .............................................................................. 105

Figure 4.27 Temperature dependence of dielectric properties for the curing

DGEBA/mPDA system ............................................................................... 107
Figure 4.28 e" vs. 2' for the curing DGEBA/mPDA system ........................................... 108
Figure 4.29 In (858...) vs. 1000/T for the curing DGEBA/mPDA system ....................... 109
Figure 4.30 In (8") vs. lOOO/T for the curing DGEBA/mPDA system ........................... 109
Figure 4.31 Cole-Cole plots for the curing DGEBA/mPDA system .............................. 110

Figure 4.32 n and (so-s...) vs. extent of cure for the curing DGEBA/mPDA system ...... 112

Figure 4.33 E, vs. extent of cure for the curing DGEBA/mPDA system ....................... 112
Figure 4.34 ’C vs. 1000/T for the curing DGEBA/mPDA system ................................... 113
Figure 4.35 Dielectric properties vs. temperature for DGEBA, W, and uncured
DGEBA/W mixture ..................................................................................... 114
Figure 4.36 Temperature dependence of dielectric properties for the curing DGEBA/W
system .......................................................................................................... 115

Figure 4.37 Comparison between the experimental and calculated data of the curing

DGEBA/W system. ..................................................................................... l 18
Figure 4.38 n and (so-s...) vs. extent of cure for the curing DGEBA/W system ............. 119
Figure 4.39 E, vs. extent of cure for the curing DGEBA/W system .............................. 120
Figure 4.40 I vs. 1000/T for the curing DGEBA/W system ........................................... 120
Figure 4.41 (so-so) vs. extent of cure for the four curing systems .................................. 123
Figure 4.42 n vs. extent of cure for the four curing systems .......................................... 124

xii

Figure 4.43 E, vs. extent of cure for the four curing systems ......................................... 125

Figure 4.44 1: vs. extent of cure for the four curing systems at 80°C .............................. 126
Figure 4.45 Calculated e' and e" vs. extent of cure for the four curing systems ............ 126
Figure 5.1 Schematic illustrations of four carbon forms ................................................ 130
Figure 5.2 Schematic illustrations of relation between graphite and CNTs ................... 131
Figure 5.3 Schematic illustrations of CNTs: (a) SWNT, (b) MWNT. ........................... 132

Figure 5.4 Schematic illustrations of three SWN'I's of different chiralities: (a) armchair,

(b) zigzag, (c) chiral. ................................................................................... 132
Figure 5.5 TEM picture of a bamboo-like carbon tube .................................................. 134
Figure 5.6 Schematic diagram of the MPCVD apparatus at MSU ................................. 138
Figure 5.7 Photo of the microwave plasma reactor ........................................................ 139
Figure 5.8 Photo of the control panel of the microwave plasma reactor ........................ 139
Figure 5.9 Photo of Si wafer on a graphite substrate ...................................................... 141

Figure 5.10 Temperature proﬁle during CNT growth in experiment No. 1-4: solid points
represent Si wafer; hollow points represent Ni catalysts. ........................... 144
Figure 5.11 Optical images of the Si wafers: (a) before CNT growth, after CNT growth in
(b) experiment 1, (0) experiment 2, ((1) experiment 3, (e) experiment 4, (f)
experiment 5. ............................................................................................... 145
Figure 5.12 Optical images of the graphite substrate with Si wafer and Ni catalyst in
experiment 3: (a) before CNT growth, (b) after CNT growth ..................... 146
Figure 5.13 SEM images (45° tilted) at different magniﬁcations of CNTs in sample 1. 147

Figure 5.14 SEM images (45° tilted) at different magniﬁcations of CNTs in sample 2. 147

xiii

Figure 5.15 SEM images (45° tilted) at different magniﬁcations of CNTs in (a) sample 3,

(b) curled CNT part and (c) aligned CNT part. ........................................... 148
Figure 5.16 TEM images of CNTs in sample 1: (a) body, (b) tip, (c) root ..................... 150
Figure 5.17 TEM images of CNTs in sample 2: (a) body, (b) tip, (c) root ..................... 150

Figure 5.18 TEM images of CNTs in sample 3: (a) body, (b) root, (0) root with a catalyst
particle. ........................................................................................................ 1 5 1

Figure 5.19 Schematic illustrations of root-growth mechanism of a bamboo-like CNT 153

xiv

LIST OF TABLES
Table 1.1 Comparison between microwave and thermal heating methods ........................ 2
Table 3.1 Properties of the reactants at 25°C .................................................................... 53
Table 3.2 Time and average power required in the two modes ........................................ 57
Table 3.3 Resonant frequencies of the two microwave modes ......................................... 58
Table 3.4 DSC results of cured epoxy resins .................................................................... 59
Table 3.5 Values of the kinetic parameters ....................................................................... 60
Table 4.1 Properties of the epoxy resin and curing agents ............................................... 76
Table 4.2 DSC results of the curing DGEBA/DDS system .............................................. 81
Table 4.3 Values of the parameters for the curing DGEBA/DDS system ........................ 91
Table 4.4 DSC results of the curing DGEBA/Jeffamine system ...................................... 96
Table 4.5 Values of the parameters for the curing DGEBA/Jeffamine system .............. 102
Table 4.6 DSC results of the curing DGEBA/mPDA system ......................................... 106
Table 4.7 Values of the parameters for the curing DGEBA/mPDA system ................... 111
Table 4.8 DSC results of the curing DGEBA/W system ................................................ 116
Table 4.9 Values of the parameters for the curing DGEBA/W system .......................... 117
Table 5.1 Literature summary on synthesis of CNTs by MPCVD ................................. 136
Table 5.2 Experimental conditions of the CNT growth .................................................. 142
Table 5.3 CNT growth pressure (Torr) ........................................................................... 143
Table 5.4 CNT growth and catalyst temperatures (°C) ................................................... 143
Table 5.5 Summary of morphology, length, and diameter of CNTs .............................. 146

XV

CHAPTER 1 INTRODUCTION

Microwaves are one form of electromagnetic radiation. It is a wave motion
associated with electric and magnetic forces. Microwave refers to electromagnetic waves
in a frequency range from 300MHz to 300GHz or a characteristic wavelength range from
1m to 1mm. Heating is one of the major non-communication applications of microwaves.
The fundamental electromagnetic property of nonmagnetic materials for microwave
heating and diagnosis is complex dielectric constant (8* = s' - js"). The real part of the
complex dielectric constant is dielectric constant, which is related to the microwave
energy stored in the materials. The imaginary part is dielectric loss factor, which is
related to microwave energy dissipated as heat in materials. The dielectric loss factor of
materials is generally due to contributions from the motion of dipoles and charges,
conductivity, etc. Polymers have polar groups to interact with electromagnetic ﬁelds and
exhibit dielectric relaxation. These polar groups can absorb microwave energy directly
and the localized heating on the reactive polar sites can initiate or promote
polyrnerizations that require heat.

Microwave processing of materials has been studied as an attractive alternative to
conventional thermal processing. Thermal heating is a surface—driven, non-selective
process. The heating efﬁciency is controlled by the heat transfer coefﬁcient at the
material surface and the thermal conductivity of the material. During thermal heating,
heat ﬂows from the surface to the interior of the material. This tends to cause remarkable
temperature gradients in thick materials. Residual thermal stresses resulting from large

temperature gradients will reduce the physical and mechanical properties of the materials.

In addition, the production cycle is long because of the difﬁculty in heating poor thermal
conductors like polymers.

Microwave heating offers a number of advantages over thermal heating in a wide
range of heating applications. A comparison between microwave and thermal heating is
summarized in Table 1.]. Microwave heating is selective, instantaneous, and volumetric
with heat loss at the boundaries while thermal heating is nonselective and depends on
temperature gradient. Microwave heating can be easily controlled by fast changes in the
applied electric ﬁeld whereas thermal heating is characterized with long lag times and
difﬁculty for composite cure control. The heat source of microwave heating can be
readily removed to prevent thermal excursion. Microwave processing has potential for

rapid processing of thick-section and complex-shaped composites.

Table 1.1 Comparison between microwave and thermal heating methods

 

 

Thermal heating Microwave Heating
Heat conduction/convection Energy coupling/transport
Surface heating Molecular level coupling
Slow Fast
Surface Volumetric
Non-selective Selective
Less property dependent Material property dependent
Surface temperature control Intelligent control
Established technology Emerging technology

 

This research is directed towards investigation of microwave processing of
polymer and composites. Four speciﬁc topics are studied and discussed in the following
chapters. The ﬁrst research topic in Chapter 2 is survey of the research status of

microwave processing of polymers and composites. The second research topic in Chapter

3 is study on kinetics of epoxy/amine curing at 2.45 GHz microwave. The third topic in
Chapter 4 is modeling the dielectric properties of curing epoxy/amine systems at 2.45
GHz. The fourth topic in Chapter 5 is synthesis of carbon nanotubes using microwave
plasma chemical vapor deposition method. The research ﬁndings and achievements are

summarized in Chapter 6, and suggestions for future work are proposed in Chapter 7.

CHAPTER 2 BACKGROUND ON MICROWAVE PROCESSING

2.1 Electromagnetic Fields in a Microwave Enclosure

Electromagnetic ﬁeld strength and distribution patterns are essential factors that
influence microwave heating efﬁciency and uniformity. They are determined by
microwave operating conditions, applicator dimensions, and material properties. To
understand microwave heating characteristics, the fundamentals in microwave processing

are reviewed.

2.1.1 Maxwell’s Equations

The basic laws governing electromagnetic wave propagation are Maxwell's
Equations [1], which describe the relations and variations of the electric and magnetic
ﬁelds, charges, and currents associated with electromagnetic waves. Maxwell's Equations
can be written in either differential or integral form. The differential form, shown as

follows, is most widely used to solve electromagnetic boundary-value problems.

V x E = -951} (Faraday's law) (2.1)
V x H = J + 96% (Ampere's law) (2.2)
V - D = p (Gauss law) (2.3)

V - B = 0 (Gauss law - magnetic) (2.4)

where E is the electric ﬁeld intensity, H is the magnetic ﬁeld intensity, D is the electric
displacement density or electric ﬂux density, B is the magnetic ﬂux density, J is the

electric current density, and p is the charge density. D is deﬁned as:

D = 50E + P (2.5)

where so is the dielectric constant of free space, P is the volume density of polarization,
the measure of the density of electric dipoles. B can be expressed as:

B = ,uO(H + M) (2.6)
where no is the permeability of free space, H is the magnetic ﬁeld intensity, and M is the
volume density of magnetization, the measure of the density of magnetic dipoles in the
material. In a simple isotropic medium, the ﬁeld quantities are related as follows:

D = 5E (2.7)
B = yH (2.8)
where sis the dielectric constant, and p is the permeability.
In addition to the Maxwell's Equations, the Equation of Continuity holds due to
the conservation of electric charge:

a
V-J+— =0 2.9
atp ( )

In the Maxwell's Equations, only two are independent. Usually Equations 2.1 and
2.2 are used with Equation 2.9 to solve for electromagnetic ﬁelds.

Maxwell’s Equations are ﬁrst-order differential equations with E and B coupled.
They can be converted into uncoupled second-order wave equations through

mathematical manipulations:

 

6J5
a 52 E V£+
[VZ-ﬂaa_ﬂg-at—2{B}= (a) ”a: (2-10)
-,quJs

where o is the conductivity, and J5 is the source current term. In a source free region,

Equations 2.10 become:

a 62 E
V2 - “0-_ _ ug— _— 0 2.11
[ at 6t21IB} ( )

Equations 2.1-2.4 are the time-domain representation of Maxwell's Equations. If
the source functions, J(r, t) and p(r, t), oscillate with a constant angular frequency 0) in
the system, all the ﬁelds will oscillate at the same frequency. The Maxwell’s equations
can be written in time-harmonic form:

rV x E(r) = —in(r)

< V x H(r) = J(r) + iwD(r)
V - 1)(r) = p(r)

(V - B(r) = 0

(2.12)

 

In time-harmonic case, Equation 2.10 becomes Helmholtz Equations and Equation 2.11

becomes Helmholtz equations in source-free region:

[v2 + wzps*]{:} = o (2.13)

5* = 5(1-1'1) (2.14)
(08

where 8* is the complex dielectric constant.

2.1.2 Resonant Modes in a Cylindrical Single Mode Cavity
In a cylindrical single mode cavity, there are two types of resonant modes,
transverse electric (TE) and transverse magnetic (TM). In TE modes, the electric ﬁeld

components are transverse, and the magnetic ﬁeld components are parallel to the

direction of wave propagation, which is the axial direction. In TM modes, the electric
ﬁeld components are parallel, and the magnetic ﬁeld components are transverse to the
direction of wave propagation. Three subscripts, n, p, and q, are used to represent the
physical appearance of the corresponding mode in an empty cavity, i.e. TEnpq and TMan.
The subscript n denotes the number of the periodicity in the circumferential direction,
n=0,1,2...; p denotes the number of ﬁeld zeroes in the radial direction, p=1,2,3...; q
denotes the number of half wavelengths of the equivalent circular waveguide, q=0,l,2. ..
for TM modes and q=1,2,3 . .. for TB modes.

Theoretically the relationship between the frequency and cavity diameter and
length for a given resonant mode can be calculated. The equations for TB and TM modes

in an empty cylindrical single mode cavity [2] are:

 

2
(f 155., = ﬁ‘lx'np2+[%j (2.15)

 

2
(f )522 = 573217;:anp2 {1251] (2.16)

where f is frequency, a is cavity diameter, h is cavity height, xnp and x' are tabulated

np
zeros of the Bessel’s function and the derivative of the Bessel’s function, respectively.
Images in this dissertation are presented in color. Figure 2.1 and Figure 2.2 show
the previous research results at Michigan State University (MSU) about the relation
between resonant frequency of some TE and TM modes and cavity length in an empty

cavity with a diameter of 17.78 cm.

Frequency (GHz)

- o 45':::3 93x:

 

 

CI 2;:
T .0 +A I:“A:Q; ‘g‘xxx**XXxxxx
x“ o +A " 66::2‘3AAAAAAAA
.0x0_ 0 + A l'. o
. I 9008
n— ’9 + A I . 6o
X . A I I - 080
. D- .‘o I I ' .0
X n- I
3D- . x D-- ++ A '..:°°'f§00930
I
X a -— + A .-
- A I
P x n -__+ A II
0 x D ‘-+_ AA
‘1 x no + ------ AA
' xx “RUDD
0 xx 00000
o xxxx ++
ZD- . Xxx +4-
0. xxXXxx
'0
.0
. o
o
0......
1.0 T I T I Y I T I I Y I r# I r r r r r f r r r 1
4 6 8 10 12 14 16 18

Cavity length (cm)

aTE011
ITE012
oTE013
xTE021
o TE111
+ TE112
-TE113
oTE121
oTE122
xTE211
ATE212
ATE221
-TE311
-TE312

Figure 2.1 TB modes in an empty cavity with a diameter of 17.78 cm

4.5 - :1 M12

'0 :1 x A ITM013

ix» psi-"20K °TM020
.0 +A

..,,.1.....1:1."£,.,......‘c ............ 0TM021

x .3 u D‘s: ++13AAAAAAA TM022
3‘5 TIAAA AAA, A,.D aa‘ AAAAAA“+ ‘AQPAA

a o . too. 1 _++++ + A ATM110

x «”9” ‘3’»...tgg”: x TM111

g
v x
g l)oowo¢5woooua$800ﬁgmg “8° 1: TM112
g oooooooooobqooooooooaagooooood Sago! -TM120
xx
11. 2.5 ~ "XXXXXX: - TM121
xxxxxx Ralﬁéxx O'HMQ10

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAE O'HM211

 

 

+TM212
‘TM310
4681012141618
Lengthofmvitywm)

Figure 2.2 TM modes in an empty cavity with a diameter of 17.78 cm

2.1.3 Electromagnetic Fields in a Cylindrical Single Mode Cavity

In a homogeneous, source-free cylindrical single mode cavity with perfectly
conducting walls, the electromagnetic ﬁelds inside the cavity can be derived from
Maxwell's equations and boundary conditions.

In TE modes, the electromagnetic ﬁeld components inside an empty cavity are

[2]:

2
4% E =__1_£°_Vﬁaq_
” A apaz P p 641
Z
2
¢ ziia ”’an 13¢ ._.___°"’"P‘I (2.17)
219 042162 6p
2
1 0 w
Hz=:(_'3_gﬂ+k2V/npq) Ez=0
Z
Z

/\
where (p, (I), z) are the cylindrical coordinates, z = j wy,k2 = (172/108, and ‘Vnpq is the

wave potential for TEnpq modes:

_ x'Lp sinn¢ , ﬂ
V’npq ‘Jn( a p){cosn¢}51n( h 2) (2-18)

where a is the diameter, and h is the height of the cavity.

In TM modes, the ﬁeld components inside an empty cavity are [2]:

2
15 mm H zlaV’npq

E =
p A G/ﬁZ p p a¢
y
2
._._1-_1____a ”’an H =———°"’"Pq (2.19)
¢ Ap awz ¢ 6/)
y
2
1 6 11!
E2 :7(_Ez_rzzp_q+k2'/’"Pq) Hz =0
y

A
where y = java, and wnpq is the wave potential for Tanq modes:

mm = J. (531’- p){::::} cox-"hie (2.20)

When the cavity is loaded with materials, Equations 2.17-2.20 are no longer
applicable. For simple materials, analytic methods are useful for calculating the
electromagnetic ﬁeld inside the materials and the cavity. For complex situations
(inhomogeneous or anisotropic materials, irregular shapes, etc.), numerical techniques are
usually used to solve for the electromagnetic ﬁeld. The most widely used numerical

techniques include the method of moments, the ﬁnite-element method, and the ﬁnite

difference method.

2.2 Microwave/Materials Interactions

2.2.1 Mechanisms of Microwave/Materials Interactions
Materials are classiﬁed into conductors, semiconductors and dielectrics according

to their electric conductivity. Conductors contain free charges, which are conducted

10

inside the material under alternating electric ﬁelds so that a conductive current is induced.
Electromagnetic energy is dissipated into the materials while the conduction current is in
phase with the electric ﬁeld inside the materials. Dissipated energy is proportional to
conductivity and the square of the electric ﬁeld strength. Conduction requires long-range
transport of charges.

In dielectric materials, electric dipoles, which are created when an external
electric ﬁeld is applied, will rotate until they are aligned in the direction of the ﬁeld.
Therefore, the normal random orientation of the dipoles becomes ordered. These ordered
polar segments tend to relax and oscillate with the ﬁeld. The energy used to hold the
dipoles in place is dissipated as heat into the material while the relaxation motion of
dipoles is out of phase with the oscillation of the electric ﬁeld. Both the conduction and
the electric dipole movement cause losses and are responsible for heat generation during
microwave processing. The contribution of each loss mechanism largely depends on the
types of materials and Operating frequencies. Generally, conduction loss is dominant at
low frequencies while polarization loss is important at high frequencies. Most dielectric
materials can generate heat via both loss mechanisms.

There are mainly four different kinds of dielectric polarization:

1. Electron or optical polarization occurs at high frequencies, close to
ultraviolet range of electromagnetic spectrum [3, 4]. It refers to the displacement of the
electron cloud center of an atom relative to the center of the nucleus, caused by an
external electric ﬁeld. When no electric ﬁeld is applied, the center of positive charges

(nucleus) coincides with the center of negative charges (electron cloud). When an

11

external electric ﬁeld is present, the electrons are pushed away from their original orbits
and electric dipoles are created.

2. Atomic polarization is also referred to as ionic polarization. It occurs in
the infrared region of the electromagnetic spectrum. This type of polarization is usually
observed in molecules consisting of two different kinds of atoms. When an external
electric ﬁeld is applied, the positive charges move in the direction of the ﬁeld and the
negative ones move in the opposite direction. This mainly causes the bending and
twisting motion of molecules. Atomic polarization can occur in both non-ionic and ionic
materials. The magnitudes of atomic polarization in non-ionic materials are much less
than that in ionic or partially ionic materials.

3. Orientation or dipole alignment polarization occurs in the microwave
range of the electromagnetic spectrum. It is the dominant polarization mechanism in
microwave processing of dielectrics. Orientation polarization is usually observed when
dipolar or polar molecules are placed in an electric ﬁeld. At the presence of external
electric ﬁeld, the dipoles will rotate until they are aligned in the direction of the ﬁeld. The
dipolar rotation of molecules is accompanied by intermolecular ﬁiction, which is
responsible for heat generation. Orientation polarization is fundamentally different from
electronic and atomic polarization. The latter is due to the fact that the external ﬁeld
induces dipole moments and exerts displacing force on the electrons and atoms, while the
orientation polarization is because of the torque action of the ﬁeld on the pre-existing
permanent dipole moments of the molecules.

4. Interfacial or space charge polarization occurs at low frequencies, e.g.

radio frequency (RF). It is a ﬁmdamental polarization mode in semiconductors. This type

12

of polarization is caused by the migration of charges inside and at the interface of

dielectrics under a large scale ﬁeld.

2.2.2 Dielectric Properties

Most polymers and composites are non-magnetic materials. The electromagnetic
energy loss is only dependent on the electric ﬁeld. Incident electromagnetic ﬁelds can
interact with conductive and nonconductive materials. The fundamental electromagnetic
property of nonmagnetic materials for microwave heating and diagnosis is the complex

dielectric constant:

5 = a — jg (2.21)
The real part of the complex dielectric constant is dielectric constant. The higher
polarizability of a molecule, the larger its dielectric constant. The imaginary part is
dielectric loss factor, which is related to energy dissipated as heat in the materials.

Usually, the relative values with respect to the dielectric constant of free space are used:

a = acts} — 162)) (2.22)
where so is the dielectric constant of free space, s,’ is the relative dielectric constant, and
Seff” is the effective relative dielectric loss factor. The loss factor of materials consists of
both polarization and conduction loss. The polarization loss is further contributed by all

four polarization mechanisms mentioned earlier. The effective relative loss factor is

expressed as [4]:

seﬂ (a2) = a; (w) + 82(0)) + a; + a: + :32)— (2.23)

13

where the subscripts d, e, a, and 5 refer to dipolar, electronic, atomic and space charge
polarization, respectively. The loss factor is a function of material structures,
compositions, angular frequency, temperature, pressure, etc.

The ratio of the effective loss factor to the dielectric constant is deﬁned as the loss

tangent, which is also commonly used to describe dielectric losses:

0'

at:

gr

tan aeﬂ = (2.24)

When introduced into a microwave ﬁeld, materials will interact with the
oscillating electromagnetic ﬁeld at the molecular level. Different materials will have
different responses to the microwave irradiation. Microwave heating of conductive
materials, such as carbon ﬁbers and acid solutions, is mainly due to the interaction of the
motion of ions or electrons with the electric ﬁeld. However, conductors with high
conductivity will reﬂect the incident microwaves and can not be effectively heated.

The ﬁelds attenuate towards the interior of the material due to skin effect, which
involves the magnetic properties of the material. The conducting electrons are limited in
the skin area to some extent, which is called the skin depth, d5. Deﬁned as the distance
into the sample, at which the electric ﬁeld strength is reduced to l/e, the skin depth is
given by [2]:

d, = 1 1 (2.25)

1 . E
Ear/101w

where a) is the frequency of the electromagnetic waves in rad/sec, p0 (=41t10‘7 H/m) is the

 

permeability of the free space, u' is the relative permeability, and o is the conductivity of

14

the conductor in mhos/m. For example, 0 = 7x104 mhos/m and d5 = 38.4 p m for graphite

at 2.45 GHz in a free space. The skin depth decreases as frequency increases. For a
perfect conductor, the electric ﬁeld is reﬂected and no electric ﬁeld is induced inside a
perfect conductor at any frequency. Therefore, no electromagnetic energy will be
dissipated even though the conductivity of the perfect conductor is inﬁnite.

Microwave heating of nonconductive materials, such as polymers, glass ﬁbers,
and Kevlar ﬁbers, is mainly due to the interaction of the motion of dipoles with the
alternating electric ﬁeld. Microwave processing of therrnosets is self-adjusting. As the
crosslinking occurs, the mobility of dipoles decreases because of the “trapping” or
reaction and the dielectric loss factor decreases. Energy absorbed by crosslinking
molecules decreases and microwaves are concentrated in unreacted molecules. During
microwave processing of thermoplastics, the dielectric loss factor usually increases with
temperature and thermal runaway is likely to occur. Thermal runaway can be prevented
by decreasing or even turning off power at a temperature close to thermal excursion.
Microwave heating selectivity of polymer composites depends on the magnitude of
dielectric loss factor of polymers and ﬁbers. When non-conducting ﬁbers, such as glass,
are used, microwaves will selectively heat the polymer matrix. When conducting ﬁbers
like graphite are used, energy is preferably absorbed by the conductive ﬁbers and heat is
conducted ﬁ'om the ﬁber to the matrix. In this case, loss factor is mainly due to the ﬁber
conductivity and can not be used to diagnose the curing process of the low loss matrix
materials.

Dielectric measurement of epoxy curing systems has shown that generally both

the dielectric constant and dielectric loss factor increase with temperature and decrease

15

with extent of reaction [5]. This dependence on temperature and extent of reaction is non-
linear. During microwave processing, the dielectric properties of materials change as a
result of heating and reaction. This affects the electrical ﬁeld strength and power
absorption in the materials. The change in electric ﬁeld and power absorption in turn
affects the temperature and extent of reaction inside the materials. Thus, the modeling of
microwave heating is a coupled non-linear problem, which involves Maxwell’s equations
for solving the electric ﬁeld strength, a heat transfer equation for obtaining the
temperature distribution inside the material, and a reaction kinetic equation for

calculating extent of reaction.

2.3 Literature Survey

Within the portion of the electromagnetic spectrum, frequencies are used for
communication and non—communication applications. Major non-communication
applications exist in medicine and heating. Two frequencies, 0.915 and 2.45 GHz, are
most widely used for microwave heating. However, other frequencies including 5.8,
24.125, 61.25, 122.5, and 245.0 GHz are also reserved by the Federal Communications
Commission (FCC) for industrial, scientiﬁc, and medical (ISM) applications [6].

Microwave processing is well established in the food, rubber, textile, and wood
products industries. Studies of microwave processing of polymeric materials in the early
1960s led to a successful industrial application in the rubber industry. Since the mid-
1980’s, there has been a great deal of interest in microwave processing of polymeric
materials worldwide. The discipline can be categorized in two major ﬁelds: microwave-

assisted polymer physics (MAPP) and microwave-assisted polymer chemistry (MAPC).

l6

In the ﬁeld of MAPP, microwave heating is used to assist the dissolution of polymers in
such solvents as water and nitric acid. The polymer dissolution is a sample preparation
step prior to analysis such as multi-element determination of major elements in polymer
additives and polymers, and molecular weight [7-11]. Microwave heating is also used to

extract additives from polymers and dry polymers [12-18].

2.3.1 Polymer Dielectric Properties

Dielectric properties of polymers are important for microwave processing,
especially the polymers used in electronic components. The dielectric properties change
during processing. As therrnosets are curing, their dielectric loss factors decrease
signiﬁcantly because of the formation of crosslinking structures. Thus, therrnosets absorb
less microwave energy so that the reaction is self-quenching. During microwave heating
of semicrystalline thermoplastics, heating can be difﬁcult until a critical temperature is
reached, where the loss factor increases signiﬁcantly [19, 20]

Polymers are usually not used as neat materials in commercial applications.
Additives are generally used in the resin formulation and the resulting composites may
have improved thermal, mechanical, and dielectric properties. The dielectric loss factor of
neat polymers is usually small. For example, 8" of the diglycidyl ether of bisphenol A
(DGEBA) epoxy at 2.45 GHz at room temperature microwave is 0.21 [21]. a" of Nylon
66 at 3 GHz at room temperature is 0.039 and that of polystyrene is 0.00085 [2].
Conducting species have much larger dielectric loss factors than polymers. A small
amount of these species can be added into polymers to improve the dielectric loss. In

iodine doped polyblends of polystyrene (PS) and polymethyhnetha-acrylate (PMMA), s'

17

and a" increase as iodine percentage increases. This can be attributed to the complex
formation in polymer due to iodine doping [22]. Carbon black is an important conducting
material that can increase dielectric properties of polymers [23-27]. Filler materials, e. g.
aluminum, copper, and silver, can be also blended with polymers. The dielectric
properties of the composite materials have been investigated as a function of volume
fraction and frequency. Normally, dielectric properties do not increase readily when a
trace amount of ﬁllers are used, but increase signiﬁcantly as the ﬁllers are further added
until a saturation point is reached. Dielectric properties also vary with changing
ﬁ‘equency nonlinearly. Furthermore, they were found to depend on inclusion dispersion
microstructure as well as constitutive properties [28-33].

Negi et al. prepared low glass-transition-temperature (T8) polymer composites
with relatively high dielectric constants, which could be modulated reversibly by voltage
variation. Polymers are glassy (hard and brittle) below T8, and rubbery (soﬁ and ﬂexible)
above T8. These polymers could be used in electronic devices [34-36]. The dielectric
constant and loss factor of poly (vinylidene ﬂuoride) ﬁlms decrease as frequency
increases from 4 to 13 GHz [37]. The dielectric properties of conductive polyaniline and
its composites have been studied [3 8-40]. Sengwa et al. studied microwave dielectric
relaxations in binary mixtures of poly (ethylene glycols) in solution [41-43].

Measuring dielectric properties of polymer composites can be used as a standard
nondestructive testing technique to determine the stratiﬁed structure of composites [44],
molecular orientation and dielectric anisotropy [45], the quality of materials such as

porosity and defect dimensions [46-49], and process control [50].

18

2.3.2 Microwave-Assisted Polymer Processing

2.3.2.1 Thermosets

Most of research on microwave processing of polymer composites focuses on
therrnosets. The commonly observed advantages of microwave processing are shortened
processing time, and improved properties [51, 52]. Although some studies suggested that
microwavesdid not change the reaction rate [53], many reports concluded that the cure
speed is faster in microwave cure than in thermal cure. Thermosets cured under
microwave irradiation include epoxies [54-81], polyesters [82, 83], polyimides [84-87],

polyurethanes [88], and others [19, 89-91].

2.3.2.2 Reaction Rate and Kinetics

Kinetic models about microwave curing of therrnosets are either mechanistic or
phenomenological. Mechanistic models are obtained based on reaction mechanisms. The
key point is the assumption of free radical polymerization. Several researchers tried to
use the concept to model the cure process of therrnosetting resins [81]. However,
derivation of mechanistic models can be difﬁcult or even impossible because of the
complexity of cure reactions. In most cases, phenomenological models are preferred in
the studies of curing because they are simpler [81]. Among phenomenological models
[74], a semi-empirical has been widely used to represent the cure kinetics of epoxy with
unsaturated polyesters:

da m _ ,,
—dt-—(k,+k2a )(1 a) (2.26)

19

where a is the degree of cure, t is time, k1 and k; are rate constants with Arrhenius
temperature dependency, and m and n are constants independent of temperature.

Wei et al. at MSU [78, 79] reported that stoichiometric mixtures of DGEBA
(diglycidyl ether of bisphenol A) with DDS (diaminodiphenyl sulfone) and mPDA (meta
phenylene diamine) were isotherrnally cured by microwave and conventional heating.
Microwave heating enhanced the reaction rates of both systems and a phenomenological
model was used to ﬁt the experimental data. Hedreul et al., who studied the reaction
kinetics of DGEBA/DDS and rubber-modiﬁed epoxy, reached the same conclusion that
the phenomenological models ﬁtted the experimental results well [59]. Fang et al.
reported the reaction kinetics of a phenylethynyl-terminated imide model compound and
an oligomer, and carbon ﬁber reinforced polyimide composites. Microwave heating gave
a much higher reaction rate for both systems than thermal heating [92, 93]. The kinetic
studies of the crosslinking of a nadic end-capped imide model compound in microwave
heating and thermal heating were investigated. At the same temperature, the reaction rate
is about 10 times faster in the microwave heating than in the thermal heating [86]. The
kinetics of simultaneous polymerization and degradation of PMMA under microwave
radiation were studied. A model was developed to predict the change of molecular weight
distribution [94]. Other studies made the same conclusion that the microwave heating can
shorten cure time and enhance reaction rate compared to thermal heating [76, 82, 83].
However, Mijovic et al. claimed that there is no difference between microwave and
thermal heating methods for processing polymers including epoxies, polyimides, and

bismaleirnides [53].

20

Some high dielectric loss materials were added into polymer systems to increase
heat absorption during microwave processing. Liu et al. studied glass-graphite-polyimide
composites and found that a small quantity of absorber, chopped carbon ﬁber, can
accelerate the cure dramatically. Furthermore, soapstone mold material was found to be
an efﬁcient absorber to accelerate the cure process [87]. Thermoplastics that contain even
modest polar groups can also be used as additives to accelerate the cure rate of epoxy
under microwave heating [67]. In contrast, the effects of carbon black concentration on
microwave curing of DGEBA/DDS were studied [95]. The magnitude of the dielectric
properties increased drastically and the reaction rate constants decreased as concentration
of carbon black increased.

Pultrusion is a continuous manufacturing process and important to industrialize
microwave processing of polymer composites. Microwave-assisted pultrusion of a
number of glass reinforced epoxy composites was studied and the pulling force was about
an order of magnitude smaller than for conventional pultrusion. It is stated that the
pulling force reﬂects a stick-slip mechanism for the crosslinked composites within the

MAP die and a slip mechanism for the uncrosslinked composites [70, 71].

2.3.2.3 Properties

Fang et al. reported that higher Tg, ﬂexible strength, moduli, and shear strength
were observed in the microwave-cured composites than in the therrnal-cured composites
[92]. MSU researchers [58, 78, 79] reported that stoichiometric mixtures of
DGEBA/DDS and DGEBA/mPDA were isothermally cured by microwave radiation and

conventional heating using thin ﬁlm sample conﬁgurations. While similar Tg for both

21

heating methods was obtained at low conversion, higher T8 was observed in microwave
cured samples at an extent of cure larger than 0.6. However, Tg was similar for
microwave and thermal cured poly (methyl methacrylate) (PMMA) [72]. The interfacial
properties of Kevlar ﬁber reinforced epoxy composites post-cured by both conventional
and microwave heating were examined [55]. The interfacial shear strengths and critical
lengths of the microwave post-cured composites are comparable with those for thermally
post-cured ones. Similarly, regarding the impact and ﬂexural strengths, the new
microwave-heated polyurethane-based polymer offered no advantage over the existing
thermal-heated and microwave-heated PMMA-based denture base polymers. But, it has
rigidity comparable to that of the microwave-polymerized PMMA [88]. The average
particle size of microwave-heated PMMA was much larger and the particle size
distribution was narrow and nearly symmetrical. Morphology of DGEBA/DDS epoxy
composites versus microwave heating rate was studied [60]. The heating rate did not have
a strong inﬂuence on morphology. But morphology of thermoplastic toughened

DGEBA/DDS epoxy can be controlled by varying the microwave power [67].

2.3.2.4 Monitor and Heat Transfer Model

To monitor the cure of polymer composites, a Time-Domain-Reﬂectometry
system was studied [96]. During microwave processing of polymer composites, high
temperatures due to exothermic cure reaction can degrade the mechanical properties of
the composite. To control the reaction and ensure uniformity of polymer composite
materials, temperature was obtained. Using the temperature information, the occurrence

of material degradation due to resin over-temperature can be reduced. In addition, a

22

theoretical model is presented that helps elucidate the inﬂuence of the microwave
parameters on the temperature proﬁle [97]. The improper control of microwave
processing of polymers, especially thick polymer laminates, can lead to quality control
problems, such as the formation of voids, non-uniform heating, and over curing. The
control system for microwave curing of polymer composites has been studied. In
particular, some quality and production issues, and the control of the process parameters,
e.g. pressure and temperature, were discussed [98].

Pichon et a1. presented a practical electromagnetic-thermal simulation of
microwave heating during the curing process of polymer resin. The model is based on
ﬁnite elements developed for the case of asymmetric geometries and ﬁelds [99]. Joly et
al. also used the ﬁnite element method to model a heat transport phenomenon in a
polymer sample heated by microwaves [100, 101]. To understand microwave cure
reaction kinetics properly, the results for both microwave and thermal curing polymer
were related by obtaining a temperature equivalent value using a phenomenological
logarithmic approach [102]. A ﬁnite difference numerical simulation was developed to
predict the one-dimensional transient temperature proﬁle of the composite laminate
during both microwave and thermal heating. Numerical and experimental results were
presented for a glass/epoxy laminate with the thickness of 25 mm. It is possible to cure
thick laminate composites unifonnly and eliminate temperature excursions caused by
exothermic reaction [77]. To simulate microwave heating of Nylon-6 inside a ridge
waveguide, Maxwell’s electromagnetic equations were coupled to the heat transfer
equation and solved numerically [86]. A program to control the temperature for

microwave curing of an epoxy has also been developed [103].

23

2.3.2.5 Thermoplastics

The heating characteristics during microwave processing of thermoplastics are
different from those of therrnosets due to different dielectric behavior during heating.
During microwave heating of semicrystalline thermoplastics, heating can be difﬁcult
until a critical temperature is reached, at which point the loss factor increases
signiﬁcantly [20]. The critical temperature is related to increased molecular mobility but
is not necessarily the same as T8 of the polymer. If the critical temperature is above Tg,
rapid heating rates can be obtained until the melting temperature of the polymer is
reached. Usually, amorphous polymer can be heated more effectively than
semicrystalline polymers [104]. The reason might be that the molecules in amorphous
polymers are not restricted by the crystal lattice and thus are more mobile. O’Brien et al.
used dual beam microwaves to heat glass mat thermoplastic sheets, reducing cycle times

up to 60% [105]. Microwaves are also used to process foamable thermoplastics and

therrnosets [106].

2.3.2.6 Composites Bonding

To bond composite by microwave, a conductive polymer is placed between the
parts being joined to serve as a preferential site for electromagnetic energy absorption.
Sufﬁcient heat can be generated to weld the joint without heating the entire part,
therefore shortening process time and limiting the part distortion [104, 107-109].
Staicovici et al. studied welding and disassembly of high density polyethylene (HDPE)
bars, placing an electromagnetic absorbent material polyaniline at the interface. By

controlling the amount of remaining polyaniline at the interface, the welded samples can

24

be placed in the microwave, reheated, and disassembled for recycling and reuse [110-
112]

Zhou et al. at MSU studied microwave adhesive bonding, using a glass reinforced
ethylene/methacrylic acid copolymer, and a nylon 6 and ethylene/methacrylic acid
copolymer as the substrates and an epoxy based material as the adhesive [21, 113].
Signiﬁcantly shorter bonding times and stronger bonds were obtained. Furthermore,
unlike single mode microwaves, variable frequency microwaves (V FM) can obtain
uniform heating in microwave adhesive bonding of large-size materials. Similar results
were obtained for bonding urethane-based glass ﬁber composite panels and ﬁberglass
reinforced polyester panels using VFM [114]. VFM can also produce strong bonds for
polystyrene and low-density polyethylene [115]. Shanker and Lakhtakia used extended
Maxwell Gamett formalism to predict the dielectric constant of a metal-dope composite

adhesive for joining polymers [1 16].

2.3.3 Microwave-Assisted Polymer Synthesis

Traditionally, polymer synthesis can be divided into polycondensation and
polyaddition. For polycondensation, the repeating unit of a polymer lacks certain atoms
which are present in the monomers. For polyaddition, however, the repeating unit of a
polymer has same macromolecular structure as the monomers forming the polymer.

Microwave-assisted polycondensation of benzoguanamine and pyromellitic
dianhydride has been studied. It is found that compared with thermal heating method,
microwave-assisted polymerization not only had faster heating and complete imidization,

but saved time and resources as well [117]. Mallalcpour et al. used microwave radiation to

25

synthesize a number of novel optically active and thermally stable poly (amide-imide)s
and poly(ester-imide)s. The microwave-assisted polycondensation proceeded rapidly and
resulted in a high yield of products [85, 118-124]. Poly (arylene ether sulfone)
functionalized with either hydroxyl or t—butylphenyl end group was synthesized.
Microwave processing of the unmodiﬁed polymers resulted in fast reaction rates but
incompletely cured products. However, in the therrnoplastic-modiﬁed networks, the
addition of the thermoplastic led to vastly improved control over system temperature and
therefore fully cured products with high reaction rates. Furthermore, networks generated
with a faster cure had much ﬁner morphologies [125]. Phase transfer catalysis (PTC) is
frequently used in the synthesis of polymers. The polycondensation of polyethers by
microwave-assisted PTC had shortened reaction time, did not require stirring, and
resulted in larger molecular-weight products [126]. Microwave radiation was also used in
the solid state polyether-ester polyamic acid imidization, resulting in decreased reaction
temperature and reaction time [127]. Others achieved a similar result that microwave
radiation can reduce the reaction time of the irnidization of polyamic acids [128]. A new
rapid synthesis of aliphatic polyarnides was presented by the microwave-assisted
polycondensation of (ii—amino amides and nylon salts in the presence of a small amount
of a polar organic medium as a primary microwave absorber. The reaction proceeded
faster than the conventional method [129-131].

Ring-opening polyaddition of s-caprolactone was investigated with a microwave
furnace at 2.45 GHz. The polymerization was accelerated and improved dramatically by
microwave heating [132]. A similar open-ring synthesis of Poly (e-caprolactam-co-s-

caprolactone) was carried out with a VFM oven. Compared with the thermal heating

26

products, microwave-assisted copolymers had equivalent molecular weight, but higher
yield, amide composition, and Tg [93]. Jacob, Chia, and Boey studied microwave-assisted
polymerization of poly (methyl acrylate) (PMA), polystyrene (PS), and
polyrnethylmethacrylate (PMMA). Microwave heating can accelerate reaction rates and
the “microwave effect” increased as the power increased [133].

Cationic polymerization of epoxies under microwave irradiation was studied by
Stoffer et al. Reaction selectivities, and reaction temperature shifts at different microwave
powers were observed [134, 135]. Solid state polymerization of poly (ethylene
terephthalate) (PET) and nylon 66 was studied. Theoretical analysis and experimental
data show that the increase in the reaction rate under microwave radiation was not caused
by an increase in the bulk temperature, but by enhanced diffusion rates due to direct
heating of condensate [68]. Microwave can reduce the reaction time and help obtain good
yields of products used to synthesize conjugated polymers such as polyphenylacetylene,
which has interesting optical and electrical properties [136]. Microwave is also used in
the ﬁeld of solution polymerization [137, 138]. A study of emulsion polymerization of PS
in a polar solvent concluded that the reaction could be carried out rapidly using
microwave radiation [139].

Copolymerization mechanism of dibutyltin maleate and allyl thiourea was
studied. The copolymer may be used as soluble polymer agents for metal ions. Effect of
composition of the feed, the power and time of microwave radiation on conversion and
intrinsic viscosity was investigated [140-142]. Graft copolyrnerization of hydroxyethyl
methacrylate onto wool fabrics was studied. Microwave heating could improve the

reactivity of the monomer. The inﬂuence of various parameters of reaction including

27

time, microwave intensity, catalyst, and monomer concentration on reaction were
investigated [143]. Microwave was used to initiate the copolymerization of methyl
methacrylate and 2-hydroxyethyl methacrylate monomers. The product was obtained in a
very short time and the molecular weight was almost double the values obtained by
thermal heating reactions [144]. Microwave can be used to obtain porous biodegradable
polymer composites with adequate micro- and macro-porosity and promising mechanical

properties, which may be used in the biomaterials ﬁeld [145].

2.3.4 Microwave Plasma Modified Polymer Surface

Polymer materials are inexpensive and normally easy to process. They have
excellent physical and chemical properties and can be used in such industries as plastics,
rubber, ﬁber, adhesive, medicine, etc. But, polymers may not have surface properties
needed for some special applications, e. g. stability, adhesion, special surface energy,
wettability, and biocompatibility. Therefore, research on technologies to modify polymer
surface is important for polymer applications. The ultimate goal of plasma modiﬁcation
is to produce polymer materials with chosen bulk properties and with particular surface
properties. Microwave plasma at low temperatures has become an interesting and
effective method to modify polymer surface. Different types of gases are used to produce
plasma. The depth of modiﬁed polymer surface is usually several hundred angstroms and
the bulk properties of the polymers are not changed [146-148].

The microwave plasma is used to modify the surface properties of polymers, such
as durable surface, adhesion, surface tension, wettability, wear resistance, and

biocompatibility.

28

In order to obtain a durable, functional surface of poly (tetraﬂuoroethylene)
(PTF E), water microwave plasma has been used to modify PTFE ﬁlms. The water plasma
introduces functional groups and radicals, serving as reactive species for the gas phase
graft polymerization of acrylic acid. A homogeneous and stable poly (acrylic acid)
(PAAC) layer with a thickness of about 70 nm was generated on the surface of PTFE
foils [149]. A study of vacuum plasma and atmospheric pressure plasma modifying
polymer surface showed that VFM-assisted nonequilibrium plasmas (APNEPs) at
atmospheric pressure could be effective in modifying polymer surfaces. The source gases
were air, nitrogen, argon, helium and gas mixtures. Polymers were LDPE, HDPE,
PMMA, polypropylene (PP), and PET. VFM APNEP can clean polymer surface, form
durable surface, and, thus, enhance surface energy [150]. Oxidation of polymer surface is
an effective method to protect polymers from oxidative degrading. A facile method of the
surface oxidation of PE and PP in the solid phase was developed, using potassium
permanganate as an oxidizing agent. The oxidation did not affect thermal properties of
the polymer [120, 122]. Polydimethylsiloxane (PDMS) is usually used as outdoor high-
voltage composite insulator. Exposure of PDMS to oxygen can cause a loss of
hydrophobicity and thus accelerate aging. To simulate the aging mechanism, oxygen
microwave plasma was used to treat PDMS materials. The oxidized surface layer with a
thickness of 130-160 nm was thinner alter longer plasma exposure [151]. Hollander et al.
studied the mechanism of oxidation of PP and PE by oxygen plasma and concluded that
short wavelength radiation contributed appreciably to the surface modiﬁcation [152].

Lianos et al. also studied the mechanisms of modiﬁcation by remote oxygen microwave

29

plasma on LDPE, PS, and PMMA. The relative density of ground state atomic oxygen in
the plasma initiated the oxidation [153].

A coating of polymers is often required to provide additional functional
properties, such as a barrier against permeation of gases, controlled optical properties,
and abrasion resistance. The permeation of water and other gases leads to aging of plastic
packaging materials. To enhance the stability of PEC and polycarbonate (PC), a new
multilayer coating system using microwave plasma was investigated and proved effective
to prevent oxygen and water from diffusing into the polymers [154]. A dual
microwave/radio frequency (MW/RF) reactor was designed to deposit an Optical thin ﬁlm
on polymer substrates [155]. The deposition of silicon alloys for protective and optical
coatings on polymers is an interesting topic. The inﬂuence of N2, H2, and 02 plasma
modiﬁcation on pure and commercial PC was investigated to enhance the adhesion of
plasma deposited silica ﬁlms. The major inﬂuence of plasma on pure PC was chain
scission and that on commercial PC was crosslinking, in which additives in commercial
PC played a major role [156]. Adhesion of amorphous hydrogenated silicon nitride
(SiNu) and oxide (SiOz) ﬁlms on PC and on silicon substrates, by using a dual-mode
MW/RF plasma system, has also been studied. The goal was to get a controllable
adhesion and optical properties of the interface. The adhesion strength was a function of
the substrate material and the energy of bombarding ions, and related to the mechanical
properties of the ﬁlms [157]. The plasma-treated PC contains a crosslinked surface layer
with a depth of 50 nm and a less dense transition region between the polymer and the

ﬁlm. The thickness of this interface layer was about 100 nm [158].

30

The study of adhesion at metal-polyrner interfaces is of great interest. In general,
modiﬁcation of the polymer surface with inert gas plasma prior to metal deposition can
improve the adhesion properties of the metal-polymer interfaces. A study of surface
modiﬁcation of ﬂuoropolymers including PTFE by the remote hydrogen plasma showed
that the treatment can improve adhesion of copper metal and the polymers [159].

Surface wettability and adhesive properties of polyamides were improved by
ammonia and nitrogen/oxygen microwave plasma [160]. To improve adhesion of PET
and metals, such as Ag, nitrogen and argon plasmas were deposited on a PET substrate
[161]. Low pressure plasmas were used to improve adhesion of a ﬂuoropolyrner and Cu.
Nitrogen was most efﬁcient among all the gases used, c. g. N2, 02, N2/H2, 02/H2, and H2.
The reasons of the improvement may be surface cleaning, increased wettability, and the
formation of chemical linkages at the interface [162]. Argon plasma treatment on PC
could improve adhesion of PC with metals, possibly caused by crosslinking in the plasma
treated surface [163].

Mechanical properties and Tg of composites of cellulose and polymers such as PS,
PP, and chlorinated polyethylene (CPE) were improved if cellulose ﬁbers were treated by
ammonia, nitrogen, and methacrylic acid (MMA) plasmas [164]. Microwave plasma can
treat polymer ﬁllers, such as, CaCO3, TiOz, and carbon black, to improve adhesion of
ﬁllers with polymer matrix, and therefore improve mechanical properties of the
composites [165]. The wettability of PE can be modiﬁed by oxygen plasma and the radio
frequency can achieve much faster treatment than microwave plasma [166]. The

wettability of poly (ether ether ketone) (PEEK), PC, PMMA [167], and poly

31

(phthalazinone ether sulfone ketone) (PPESK) [168] were improved by microwave
plasma treatment.

The wear resistance of PMMA was improved by microwave plasma treatment of
the surface. The agent was plasma of CH4 diluted in Ar gas, which foamed a transparent
polymer-like carbon ﬁlm on the PMMA surface [169]. Scratch-resistant PC ﬁlms were
achieved by remote argon plasma modiﬁcation on the PC surface. A good sticking
interface polymer layer grew with argon discharging [170]. MW/RF plasma technology
can also be used to remove polymers effectively from a material surface in the
semiconductor industry [171].

Polymers can be used as biomaterials. In the ﬁeld of tissue engineering, poly (3-
hydroxybutyrate) (PHB), produced by many types of microorganisms, has become
commercially available. But, it is hydrophobic while biomaterials should be hydrophilic.
Ammonia plasma can be used to modify the surface properties of PHB. A durable
conversion of the hydrophobic material into hydrophilic was obtained and no signiﬁcant
changes in the morphology of the surface was observed. Furthermore, there were amides
and amino groups on the surface, which will be useful in the necessary biochemical
reaction [172]. PTFE is one of the common polymers applied in medical devices and
long-term blood-contacting implants. However, the hydrophobic property of PTF E is a
drawback because PTFE can adsorb protein strongly. In order to change PTFE into a
hydrophilic material, poly (ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide)
(PEO-PPO-PEO) triblock copolymers were immobilized on the surface of ﬂuoropolymer
using argon plasma method successfully [173]. MW/RF plasmas have been used to

modify biomedical polyurethane and silicone. Feed gases were CO2, H20, and NH3. The

32

plasmas can improve the stability of the materials but have no signiﬁcant improvement
regarding antibiotic properties. Furthermore, MW-plasma system was better than RF-
plasma system in terms of deposition rate, barrier properties, and introduction of amine
functional groups onto the surface [174]. Schroder et al. concluded that ammonia plasma
treatment is a fairly universal method to introduce amino bioftmctional groups onto the
surfaces of polymeric biomaterials including PE, PS, PC, PEEK, PET, PMMA,
polyethylenenaphthalate (PEN) [175]. To improve biocompatibility of PE, microwave
plasma polymerization of allyl alcohol can be used to introduce hydroxyl groups into the
polymer surface [176]. In order to be applied in biomedical ﬁelds, the surface
modiﬁcation of polyimides by H20 and D20 plasma was studied [177].

Microwave plasma modiﬁcation mechanism is an interesting topic for many
researchers. Kobayashi et al. studied the mechanism of polyacrylonitrile (PAN)
modiﬁcation by oxygen plasma [178]. A Langmuir probe can be used to detect the end
point of polymer etching by plasma [179]. The reaction of imidazole molecules plasma
with poly (vinyl chloride) (PVC) was investigated and a mechanism was proposed [180].
Bichler et al. investigated the adhesion mechanisms of aluminum, aluminum oxide, and
silicon oxide on biaxially oriented polypropylene (BOPP), poly (ethyleneterephthalate)
(PET), and poly(vinyl chloride) (PVC) [181]. The microwave plasma method had a
drawback of surface degradation and the degradation rate was high when the polymer
materials were directly immersed in the plasma because of ion and electron
bombardments. Therefore, modifying polymers in the ﬂowing afterglow of the discharge
may decrease the degradation. The mechanisms of surface modiﬁcations of polymers,

such as PP, PE, and PC, by the ﬂowing afterglow of oxygen plasma were studied [182].

33

The mechanism of A1 and Cu metallization of untreated and oxygen plasma treated PE
and PET was investigated [183]. Spin-coated specimens of crosslinked

polydimethylsiloxane modiﬁed by oxygen plasma were investigated [151].

2.3.5 Microwave Plasma Polymerization

Microwave-activated monomer plasma can be used to synthesize polymers, which
have unique features. Transparent and ﬂuorescent polymer ﬁlms were synthesized and
deposited on to glass slides at 10 Pa by microwave plasma using three types of volatile
aromatics: benzene, toluene, and styrene [184]. The microwave plasma assisted
deposition of hexarnethyldisiloxane [HMDSO-(CH3)3SiOSi(CH3)3] ﬁlms for corrosion
protection of Al metal sheet surfaces was synthesized [185]. A plasma polymer from
acrylic acid deposited on PE by a pulsed plasma method was studied and a mechanism
was proposed to explain experimental data [186]. Ultrathin (<5 nm) ﬂuorinated polymer
ﬁlms of homogeneous thickness were synthesized using CF4/H2 plasma. The surface
energy of the ﬁlms was more than four times that of PTFE. The deposition can be
separated into two phases, a growth phase and a treatment phase. The depth of the ﬁlms
was limited by plasma parameters. The thickness limiting behavior can be explained by
the dualism of etching and polymerization occurring in ﬂuorocarbon discharges. The
ﬁlms showed the excellent anti-adhesion to PC and good adhesion to the Ni substrate
[187]. The properties of polymers from n-hexane and n-heptane plasma were studied

[188]. A book about plasma polymerization has been published [189].

34

2.3.6 Polymer Degradation

To protect the environment, petroleum-based polymer wastes should be recycled.
A study of microwave pyrolysis of HDPE and aluminum/polymer laminates showed that
microwave method had the same features as conventional pyrolytic method and can also
treat laminates. Clean aluminum can be recycled after the microwave treatment [190].
Devulcanization is one new method of recycling waste rubber products, meaning that the
cleavage of cross-linking sulfur bonds without destroying the polymer chain bonds.
Microwave can be used in the rubber devulcanization process and was proved effective
[191]. Microwave has also proven to be an effective energy source in the solvolysis of
polyamide-6 [192], and PET [193]. The depolymerizations were ﬁnished in 4-20 minutes.

Some polymers containing silicon can convert to ceramic when heated at
temperatures higher than 800°C. Six preceramic polymers were pyrolyzed into ceramics
using microwave and thermal heating methods. The heating method could affect the
amount and size of the B—SiC nanocrystals and the graphitization of residual carbon
[194]. A polymeric method based on the Pechini process was investigated in synthesizing
alpha-alumina. Different polymers were prepared using the microwave and thermal
heating methods in the polyesteriﬁcation reaction. Microwave heating at 2.45 GHz can
reduce the polyesteriﬁcation time dramatically and the mechanism of the reaction did not
change [195]. The electrochemical properties of superﬁne spine] LiMn2O4 ceramics

synthesized by microwave polymer method were studied [196].

35

2.3.7 Nanomaterials

Nanomaterials have been a great interesting interdisciplinary ﬁeld of material
science in the past ten years. Microwave can be used in production of nanomaterials. A
concentrated uniform PS nanoparticle in solution was obtained under microwave
radiation and a structural model was proposed to predict the resultant particle size [197].
Microwave heating method helped obtain monodisperse TiO2 nanoparticles from inverse
polystyrene-poly(ethylene oxide) diblock copolymer micelles in toluene by hydroxylation
of titanium alkoxide [198]. PS-coated Fe nanoparticles were obtained by microwave
plasma polymerization and their magnetic properties were studied. PS coating plays an
important role in the underlying magnetic response of the particles [199]. Uniform
polymer-stabilized metal nanoparticles can be produced continuously under microwave
irradiation [200]. Microwave plasma method can also be used to produce

polymer/ceramics nanoparticles [ZOO-204].

2.3.8 Applications

The use of microwave energy can be classiﬁed as either a communication or non-
communication application. Since the discovery of electromagnetic waves, it has been
widely used in communications. The earliest reported non-communication use of
microwaves in polymer processing was an attempt to cure plywood cement in 1940
[205]. In the I960S, microwave processing was successfully applied in the vulcanization
of the rubber in the tire industry [206]. By now, the vulcanization of extruded rubber
weather-stripping for the automotive and construction industries has been one of the most

successful applications of microwave heating in industry [207]. Microwave heating has

36

also been used in forest industry and food industry [206, 208-213]. Examples of
industrial applications of microwave processing include curing of Flip-Chip underﬁll and
pultrusion of wood products [214]. In Flip-Chip process, the silicon chip is attached to
the Printing Circuit Board (PCB) via solder bumps to reduce the assembly size. Recently
organic PCBs are used to replace the expensive ceramic boards. Underﬁlls are used to
reduce the mismatch of coefﬁcient of thermal expansion between the silicon chips and
the substrates. Microwave techniques are applied in this process to selectively heat the
underﬁlls without heating up the PCBs and to reduce the cure cycles.

Microwave has also been used in pultrusion of wood products. The pultrusion
process is a continuous manufacturing method that can be used to produce composites.
The shape of the product is determined by continuously pulling the composite material
through a die to produce uniform proﬁle parts. The key step in a pultrusion process is to

control the solidiﬁcation process within the die.

37

CHAPTER 3 MICROWAVE CURING MECHANISM OF EPOXY RESINS

3.1 Introduction

The previous research results at Michigan State University have shown that
microwaves reduced the curing times of epoxy resins [215-220]. These results motivate
further investigation on microwave heating mechanisms to provide explanations for
reaction rate enhancement by microwaves. The enhancement of polymer curing rate has
been demonstrated in a number of studies [78, 89, 221-223]. Some investigators
suggested that the reaction rate enhancement was because of microwave thermal effect,
which is localized superheating [78, 221]. Some other investigators attributed it to
speciﬁc microwave non-thermal effects, such as accelerated reaction of the secondary
amine group [222], and improved difﬁision rate of reactive species [223].

Fu et al. studied the microwave thermal or non-thermal effect by comparing
continuous-power and pulsed-power microwave curing of epoxy resins [56].
Experimental results showed that continuous-power microwave curing had only slightly
higher reaction rates and ultimate extents of cure than pulsed-power curing. The results
seemed to support the theory of thermal effect. But non-thermal effect could not be
disproved because the power level in pulsed-power curing was much higher than that in
continuous-power curing. Microwave power has large inﬂuence on both microwave
thermal and non-thermal effects. Further, it has been pointed out that microwave heating
of materials depends largely on dielectric properties [4]. Microwaves can be more
efﬁciently coupled into components with larger dielectric loss factor. Fillers with high

dielectric properties can be added into resins to modify microwave thermal effect without

38

signiﬁcantly affecting the non-thermal alignment of polar groups in the electromagnetic
ﬁeld.
In this chapter, microwave heating mechanism is investigated via studying

microwave curing kinetics of epoxy resins.

3.2 Mechanisms of Epoxy Resin Curing Reactions

The conversion of thermosetting resins to rigid solid is brought about by small
chain polymer molecules reacting with curing agents or each others to form a crosslinked
molecular network. Thermosets, such as phenolics, amino resins, polyesters,
polyurethanes, poly-isocyanurates, silicones, and polyimides, react in a similar fashion as
epoxy resins. However, epoxy resins have great versatility, low shrinkage, good chemical
resistance, high mechanical properties, outstanding adhesion, and reaction without the
evolution of byproducts. Epoxy resin curing may accomplish at room temperature or
require the addition of external heat, depending on the type of the curing agent. For
example, the epoxy/amine resins used in this study required heat to initiate.

Epoxy groups react with amine via a ring—opening mechanism. Functionality of
epoxy resin or curing agent is determined by the number of reactive groups per molecule.
The three-step epoxy reaction is shown in Figure 3.1 [224]. In the ﬁrst step, an epoxy
group reacts with a primary amine to form a secondary amine. In the second step, another
epoxy group reacts with the secondary amine to form a tertiary amine. The third step is
etheriﬁcation, which is reaction of a formed hydroxyl group and an epoxy group to form

an ether crosslinking epoxy. However, etheriﬁcation is insigniﬁcant for stoichiometric

39

mixtures. The progress of the reaction is deﬁned in terms of extent of cure or percentage

of available epoxy groups reacted.

 

 

Step 1
R1 R2
R1-NH2 + I>__R2 \NH/W/
O
OH
Step 2
OH
R1 R2 R
\NH 2
+ R3 —. R1 N R3
O
OH
OH
Step 3
R6
R5
R4 R5
Y + R6 —-> O OH
0 R4

Figure 3.1 Three-step epoxy curing mechanism

The hydroxide groups, formed during the reaction, can act as catalysts so that the
reaction is autocatalytic, which is shown in Figure 3.2 [215]. The electron pair of the

amine group bonds to the chain terminating carbon in the epoxy group, causing a bond

40

breakage of the carbon-oxygen bond. The hydrogen atom detaches itself from the amine
group and reattaches to the oxygen atom to form a hydroxide. The hydroxide group

catalyzes further epoxy/amine addition by providing a hydrogen bond to the epoxy group.

 

Uncatalyzed
R, R2
R,—NH2 + l>—R2 ———> \NH
O
OH
Autocatalyzed
R1\Ii'/H* + I>__'R2 + HO-R3
H O
R,\ ,H*
——>
'f -------- CH2——CH R2
l-I \ /
\
\\
\
O
I
i
I
I
HO—R3
R1\ R2
———> NH + HID—R3

OH

Figure 3.2 Uncatalyzed and autocatalyzed epoxy curing reaction mechanisms

41

3.3 Epoxy Resin Curing Kinetics

Epoxy resins are the most widely used matrix materials for advanced composites.
A large amount of work has been performed in the curing of the general class of epoxy
resins. A variety of models, proposed for curing of neat epoxy resins, have been ﬁirther
applied to thermal curing of doped resins and microwave curing of neat resins. Research
efforts on the curing kinetics of epoxy and other commonly used resins, such as

vinylester and polyester, are reviewed.

3.3.1 Thermal Curing

There are mainly two categories of kinetic models for the curing process [81].
Mechanistic model is obtained based on reaction mechanisms while phenomenological
model is developed without considering the details of cure reactions. Although
mechanistic models offer the advantages of better prediction and interpretation without
conducting cure experiments for each new variable in the cure system, the
phenomenological models usually have simpler forms with less kinetic parameters. In
addition, the complexity of cure reactions sometimes makes the derivation of mechanistic
models very difﬁcult or even impossible. Therefore, phenomenological models have been
used in most studies of cure kinetics. A summary of mechanistic and phenomenological

models for cure reactions is presented as follows.

3.3.1.1 Mechanistic Models
The proposed reaction kinetic mechanism for epoxy-aromatic diamine system is

shown in Equation 3.1 [225],

42

K1

0] +e__)az +OH
Ki
K2

az+e_)a3+0H (3.1)
K2

0H + e :3) et + 0H
K 3
where a1, a2, a3, e, and et are primary amine, secondary amine, tertiary amine, epoxy, and
ether group, respectively; K, and K,‘, i=1, 2, 3, are speciﬁc reaction rate constants for the
catalytic and non-catalytic reactions, respectively. From the kinetic mechanism,
mechanistic models for the curing process can be derived.
For the simpliﬁed case of no etheriﬁcation, steric hindrance or OH impurity, a

cure kinetic expression for epoxy has been derived as follows [226]:

{—3 = (k, + kza)(1—— a)(B — a) (3-2)

where B is the ratio of the initial hardener equivalents to epoxy equivalents, B=l for a
stoichiometric mixture; or is the extent of cure; k1 and k2 are the catalytic and non-
catalytic polymerization reaction rate constants, respectively.

The above equation holds well up to the gelation point. To model the whole

curing reaction, the following kinetic model has been proposed [227]:

do:

(3.3)
da -k 1 h
E— 3( _a), W ena>agel

43

where k3 is the ﬁrst order reaction rate constant with Arrhenius temperature dependency,
and age] is the extent of cure at the gelation point.

The etheriﬁcation of a stoichiometric mixture of epoxy and amine can be
neglected at low curing temperatures [79, 228-231]. However, it can no longer be ignored
at high curing temperatures or with excess epoxy [79, 215]. In addition, the reaction rate
constant for primary and secondary amine is not always the same. For the generalized
case of epoxy curing with etheriﬁcation, steric hindrance and an OH impurity, the curing
kinetics has been derived for a stoichiometric mixture of epoxy and amine [79]. The

kinetic models are shown in the following equations:

g3 2(1-n)¢+n¢"/ 2

 

 

d =1 +LF(¢)](1—a)[k1 +k2F(¢)1 (3.4)
t 2—n
n/2
e0 2—n
(l—¢)(l—n)(2—L)+2(1—¢"/2)(1—£)—(2—n)L(l+L0—Ii]—°—)ln¢
a = " 30 (3.6)

 

2(2 - n)

where n is the reaction rate constant ratio between the secondary amine-epoxy reaction
and the primary amine-epoxy reaction, n = K2/K1 = K2'/K1'; L is the reaction rate constant
ratio between the etheriﬁcation and the primary amine-epoxy reaction, L= K3/K1 =
K3'/K1'; [OH]0 is the initial concentration of OH impurity; (Pal/co; e0 is the initial epoxy
concentration; k1=eoK1'; and k2=eozK1.

If L=0 (i.e. no etheriﬁcation), n=1 (i.e. no steric hindrance) and [OH]0=O (i.e. no
OH impurity), the above reaction kinetics simpliﬁes into the following equation for a

stoichiometric epoxy-amine mixture:

44

71—? = (k1 + kza)(l - (1)2 (3.7)

This kinetic equation is consistent with Equation 3.2 because of B=1 for a stoichiometric

mixture in Equation 3.2.

3.3.1.2 Phenomenological Models
The simplest phenomenological model is the nth order reaction kinetic model [74,
232], which assumes that the kinetics can be expressed as:

do:

Et- = k(T)f(a) (3-8)

where or is the extent of cure, t is the time, the function f(o.) is expressed as (l-or)", and

k(T) is the overall reaction rate constant which obeys the Arrhenius relation:
k(T) - Aex (——E—) (3 9)
p RT '

The nth order reaction kinetics is computationally simple. According to this
model, the maximum reaction rate should occur at the beginning of the reaction.
However, in real cases, or=0.3 ~ 0.4 at maximum reaction rate, which is better explained
by the autocatalyzed reaction mechanism [233, 234]. The reactions between amines and
epoxy are autocatalyzed by the hydroxide groups formed in the reactions. The initial rate
should be slow due to lack of catalytic hydroxide groups. The cure kinetic expression of
autocatalyzed reaction for a stoichiometric reactant mixture is given by:

d m n
—d:£=(k1+k2a )(1—a) (3.10)

45

where k1 is the non-catalytic polymerization reaction rate constant, k2 is the autocatalytic
polymerization reaction rate constant, in is the autocatalyzed polymerization reaction
order, and n is the non-catalyzed polymerization reaction order. This model has been
widely used to represent adequately the cure kinetics of epoxy and unsaturated polyester

cure systems [233-241].

3.3.1.3 Microwave Curing Kinetic Model

Thermal cure kinetic models have been used in modeling the reaction kinetics of
microwave cured epoxy resins [56, 79, 218]. It was demonstrated that the microwave
cure kinetics of epoxy resin systems could be described by the autocatalytic kinetic
model up to vitriﬁcation [79, 218]. In the study of continuous-power and pulsed-power
microwave curing of epoxy resins [56], a semi-empirical kinetic model was used:

%=(k1+k2am)(au -a)" (3.11)

where or is the extent of cure, R1 and k2 are rate constants, m and n are constants, and 01.,
is the ultimate extent of cure. This model is similar to Equation 3.10 except that the
ultimate extent of cure oru is included in the equation. This is because, at certain stage of
the reaction, gelation and vitriﬁcation take place and, thus, the reaction rates are

controlled by physical deposition. The ultimate extent of cure is usually less than 100%.
3.3.2 Kinetic Model Used in This Study

The phenomenological kinetic model in Equation 3.10 is used in this study. The

reaction rate constants k1 and k2 obey the Arrhenius relation:

46

W) = A. exp<—%.) (3.12)

where i=1 for non-catalytic polymerization reaction or 2 for autocatalytic polymerization
reaction, A, is the Arrhenius frequency factor, and E, is the activation energy. E, and A,

can be obtained from the reaction rate constants at different temperatures.

3.4 Experimental
3.4.] Experimental Equipment

3.4.1.1 Experimental Circuit

The experimental circuit was assembled for microwave processing. The circuit
directs microwaves into the applicator, allows the measurement of temperature, incident
and reﬂected powers, and reduces the power reﬂected back to the power source to
prevent damage to the power source. The microwave circuit is illustrated in Figure 3.3.

Microwave signal generator is a sweep oscillator (HP8350B) connected with a RF
plug-in (I-[P86235A). A variable frequency ampliﬁer (Lambda LT-1000) is used to
amplify the signal. The ampliﬁed power signal is in the range from 0 to 200 Watts.
Microwave frequency can be adjusted from 2 GHz to 4 GHz either manually or
automatically. A 3-port circulator is used to prevent the reﬂected power from damaging
the power source. The input and reﬂected microwave powers are decoupled with 20db
directional couplers (Narda 3043-20) and measured with power meters (HP435B). A
dummy load is used to absorb most of the reﬂected power. A multi-channel LUXTRON

ﬂuoroptic thermometer and a multi-channel Nortech NoEMI-TS ﬁberoptic thermometer

47

are used for sample temperature measurement. The probes are electrically nonconductive

so that they do not perturb or be perturbed by the microwave ﬁelds.

 

Microwave Circulator Directional Microwave
Power Source Coupler Applicator
AF—__l

 

 

 

 

”/l 'l I

 

. . C: Thermometry
Directional P p.
Coupler ’ '

 

 

 

 

 

 

 

 

 

 

 

 

 

'F"""""""""""

 

 

1 Reﬂected Incident (:1: ND
‘— Power Meter Power Meter I
Dummy Load r——1
I—J 1:] Computer
Controller

 

 

 

 

Figure 3.3 Circuit of the microwave curing system

A cylindrical single mode cavity with a diameter of 17.78 cm is used. The
coupling probe is side mounted 3 cm above the bottom of the cavity. The cavity length
(L) and the probe depth (Lp) are adjusted to be 13.2 cm and 2.0 cm, respectively. A

sample is loaded at the center of the cavity.

3.4.1.2 Temperature Sensing Systems
Two types of thermometers are used in this study. One is multi-channel
LUXTRON ﬂuoroptic thermometer. The Luxtron thermometer uses Decay Time

Technology to measure the temperature of the sensor [242]. Luxtron sensors contain a

48

small amount of magnesium ﬂuorogerrnate. The sensors are attached at the tip of the

optic ﬁber. The optical system excites the sensor with blue light. In turn the sensor
ﬂuoresces a red light, the intensity of which decays exponentially with time. The time
constant of the decay is inversely proportional to the temperature. Therefore, the
temperature can be obtained by measuring the decay time. The temperature probes are
electrically nonconductive, which will not perturb or be perturbed by the microwave
ﬁelds. The other one is multi-channel NoEMI-TS ﬁberoptic thermometer. The working
principle is based on the absorption of light by a semi-conducting crystal bonded to the
end of an optical ﬁber [243]. The crystal is in well contact with the materials to be
processed with microwaves. As the crystal temperature increases, more low-energy
photons are captured and absorbed by the band. The absorption edge is moved towards
the longer wavelengths. Therefore, measuring the position of the absorption shift gives a
measurement of the crystal's temperature and, thus, the temperature of the materials. The

sensor is immune to and does not perturb the electromagnetic ﬁeld.

3.4.1 .3 Microwave Applicators
The most commonly used microwave applicators include waveguide, commercial
multimode microwave ovens, and single mode applicator. Waveguides are hollow metal
tubes, the high-reﬂectivity walls of which allow microwaves to propagate. Commercial
multimode microwave ovens have large dimensions compared to the operating
wavelength, allowing the establishment of multi modes at the same time. The EM ﬁeld
inside multimode ovens is not uniform and shows many peaks and valleys. Turntables are

usually used to rotate the materials to be processed for more uniform heating. Single

49

 

 

 

mode cavity supports one mode at one time and has well-deﬁned electric ﬁeld pattern.
The single mode cavity system has higher energy efﬁciency to transfer microwave power
into the processed materials. A cylindrical single mode cavity is used in this study. Since
one mode heating is not uniform with high ﬁeld intensity conﬁned to small regions,
variable frequency techniques can be used to excite several modes with complementary

heating patterns sequentially to obtain more uniform heating.

A sketch of the cylindrical single-mode cavity used in this study is shown in

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.4.
1 _
Shorting Plate ‘‘‘‘‘ I J
--~~~$
Coupling Probe ‘, 1
‘x I
~. 1——-~ :L,
L
Microwave —> I :1 I
' I
1;___.. ,
H—I-i-p— +1 I

 

 

 

Bottom Plate """"" *I

Figure 3.4 A cylindrical single-mode resonant cavity

The cavity is made out of a length of metal circular waveguide with both ends
shorted by brass. The cavity has an inner diameter of 17.78 cm with cavity length
adjustable from 10 cm to 30 cm. Microwave energy is introduced into the cavity by a

coaxial coupling probe. The coupling probe is side mounted 3 cm above the base of the

50

 

 

 

cavity. The probe is adjustable in the radial direction so that the coupling probe depth Lp
can be changed for locating critical coupling conditions. The range of the probe depth is
from 0 mm to 50 mm. The top shorting plate is adjustable so that the cavity length Lc can
be changed. The bottom plate is removable for sample loading. Both the top and the

bottom plates are shorted with the cavity wall by metallic ﬁnger stocks.

3.4.1.4 Cavity Characterization and Process Control

Before microwave curing, the loaded cavity was characterized to locate the
heating modes. The mode spectrum, as shown in Figure 3.5, was obtained with
measuring the incident power (P,) and the reﬂected power (P,) as a function of frequency.
The frequency with minimum reﬂectance (Pr/Pi) was the resonant frequency of a mode.
Among many available electromagnetic modes, two center heating modes TM020 and
TM021 were selected because the material sample was loaded at the center of the cavity.
For the experimental setup in this study, the resonant frequency of TM020 was around
2.89 GHz and the resonant frequency of TM021 was around 3.17 GHz. Figures 3.6 and
3.7 describe the theoretical electric ﬁeld of TM020 and TM021 in a empty cavity with

FEMLAB, respectively.
Traditional proportional-integral-differential (PID) method was used to control
the curing temperature by adjusting the incident power level. The PID controller was
programmed with LabView as a subroutine. The three parameters K, T, and Td were
obtained with Ziegler-Nichols frequency response method. The details of the method are

described in literature [244].

51

 

 

 

1

E:

in. 0.5
0 I I TI 1’
2.4 2.9 34 3.9

Frequency (GHz) .

Figure 3.5 Mode spectrum of the loaded cavity for microwave curing

High

Power

 

 

 

Figure 3.6 Electrical ﬁeld of TM 020 mode

52

 

1‘
High

Power

Low
Power

 

 

 

Figure 3.7 Electrical ﬁeld of TM 021 mode

3.4.2 Experimental Materials and Procedure
The epoxy resin was diglycidyl ether of bisphenol A (DGEBA), and the curing
agent was 3, 3'-diaminodiphenyl sulfone (DDS). DGEBA was DER332 ﬁ'om Dow
Chemical with an epoxy equivalent weight of 173 and a molecule weight of 346 g/mol.
The curing agent DDS was from TCI America with an amine equivalent weight of 62 and

a molecule weight of 248 g/mol. The properties of the reactants are shown in Table 3.1.

Table 3.1 Properties of the reactants at 25°C

 

 

Materials Chemical Structure Density (g/cm3) e' s"
CH ,__<|
3 O
DGEBA O 0H0 1.16 4.31 0.383
[>—/ on,
H2N NH2

 

DDS 80> <O§ 1.33 4.04 0.116

 

 

 

53

 

All of the materials were used as received without further puriﬁcation. In
preparing the neat epoxy resin, stoichiometric DGEBA/DDS (2.79: 1 by weight) were
mixed at 130°C. The mixture was well stirred by hand in a 130°C oil bath until DDS was
completely dissolved (in approximately 5-10 minutes). The resins were degassed at 0.02
bar and 100°C for 5 minutes. Fresh samples were kept in a -20°C freezer and used within
2 weeks.

For each microwave curing experiment, 0.1 i 0.003 grams of resin was loaded
into a cylindrical Teﬂon holder with 2 cm high, the inner diameter of which was 1.25 cm.
The thickness of the resin was around 0.6 mm. Because of the small dimensions of the
sample, the temperature within the resin was assumed as uniform. The cylindrical Teﬂon
holder with fresh DGEBA/DDS mixture was loaded on the bottom of the cavity and
heated under microwave radiation from room temperature to the cure temperature 145°C

for a certain time, ranging from 5 to 100 min.
After the completion of each experiment, extents of cure were determined by
Differential Scanning Calorimetry (DSC). The samples were scanned in the DSC pan at a
heating rate of 5°C/min from 20 to 320°C. The extents of cure were then calculated with

the following equation:

(3.13)

where a is the extent of cure, H, is the total heat of reaction per gram of the uncured

sample, Hlr is the residual heat of reaction per gram of the cured sample.

54

3.5 Results and Discussion

3.5.1 Temperature and Power Deposition Profiles
During microwave curing, data acquisition of temperature, incident and reﬂected
powers was performed every second. With the PID controller used, the curing
temperature was controlled within 1°C of the set point temperature. The typical
temperature proﬁles during microwave heating and curing of epoxy resins are shown in
Figure 3.8. The power deposition was calculated as the difference between the incident
and reﬂected powers. The reﬂected power was close to zero for the experimental setup in
this study. The typical power deposition proﬁle during microwave curing at 145°C is
shown in Figure 3.9. The temperatures in Figure 3.8 and deposition powers in Figure 3.9

are the average numbers in every 15 seconds.

200 7

p...

(Jr

C
l

  
 

+ TM 020 mode
—9-— TM 021 mode

U1
O
1

 

Temperature (°C)
5'
O

O

 

I I I l

0 2 4 6 8 10

Time (minute)

Figure 3 .8 Temperature proﬁles during microwave cure at 145 °C

55

O
'. .r
‘
ll
.‘
3
:‘I
ll
ll
'6
.s
'1

 

 

 

Al " r
E .
l3
3
0
m5.
+TM 020 mode
+TM021mode
0 TX I l I l
0 2 4 6 8 10

Time (minute)

Figure 3.9 Power proﬁles during microwave cure at 145°C

The temperature proﬁles of the two modes in Figure 3.8 are similar. The

temperature increased with heating time until 145°C and then ﬂuctuated around 145°C.

The difference is that it took longer time to heat DGEBA/DDS mixture to the curing
temperature in TM 021 mode than TM 020 mode.

The power deposition curves for two modes in Figure 3.9 are similar. At the

beginning, initial high power levels dropped to set power level 10 W. Then, high power

levels were required to heat up the materials from room temperature to the isothermal

curing temperature. After the isothermal curing temperature was reached, the power level

was stable for a while and then dropped. This phenomenon can be explained by that the

:xothennal heat generated and dielectric property changed during the curing process. The

:uring of DGEBA/DDS was governed by an autocatalyzed reaction mechanism and the

naximurn curing rate usually occurred at the extent of cure of 0.3-0.4. Therefore, at the

56

 

 

 

beginning of curing, the curing rate kept increasing until the maximum rate was reached.
During this period, high power level was needed. Correspondingly the rate of heat
generation from the exotherrnal curing reaction increased. This tended to lower the power
requirement. Then, the increase of curing extent in DGEBA/DDS mixtures tended to
decrease the power requirement and the decrease in dielectric properties in cross-linking
process tended to increase the power requirement. Therefore, power Was still required to
maintain the curing reaction but not as much as at the begimring of curing. However, the
overall power requirements in two modes are not same, more power is required in TM
021 mode than in TM020. This is consistent with the phenomenon of temperature proﬁle
in Figure 3.8.

Table 3.2 shows times used to heat DEGBA/DDS reacting system from room
temperature to 145°C and average power depositions in the two modes. Less time and

power depositions were required in TM 020 mode than TM 021 mode.

Table 3.2 Time and average power required in the two modes

 

 

 

 

 

Curing Time Heating time from room Average power
(minutes) temperature to 145°C (min) deposition (W)
TM 021 TM 020 TM 021 TM 020
5 3.9 1.2 10.5 7.8
10 6.9 2.4 12.7 5.4
20 3.2 2.5 10.7 8.5
30 6.0 0.9 17.3 3.4
40 1.6 1.6 7.7 6.4
60 2.6 0.9 11.7 6.1
80 3.0 2.8 20.6 11.0
100 2.4 2.0 13.9 11.0
Average 3.7 1.8 13.1 7.4

 

57

Based on the above experimental results and analysis, it is concluded that TM 020
mode is more effective and efﬁcient than TM 021 mode. This can by explained by the
theoretical electric ﬁeld distribution of the two modes in Figures 3.6 and 3.7. The highest
power concentration of TM 020 mode is located in the center of the cylindrical cavity
while that of TM 021 mode is located in the centers of the top and bottom circles of the
cylindrical cavity. Since the sample was put in the cylindrical center a little above the
bottom circle, the power, that the sample can be absorbed, is not as much in the

electromagnetic ﬁeld of TM 021 mode as in that of TM 020 mode.

0 3.5.2 Kinetics

Microwave curing of the epoxy resins were performed at 145°C. The average
resonant frequencies and standard deviations of microwave modes used in the
experiments are shown in Table 3.3. The extent of cure of the resins was tested with DSC
as a function of curing time. At least three samples were measured for each time point.
The experiment data are shown in Table 3.4. The extent of cure as a function of curing

time is shown in Figure 3.10.

Table 3.3 Resonant frequencies of the two microwave modes

 

 

Mode TM 020 TM 021
Resonant Frequency (GHz) 3.123 2.887
Standard Deviation 0.002 0.006

 

58

Table 3.4 DSC results of cured epoxy resins

 

 

 

 

Time TM 020 Time TM 021
(min) Absorption Extent Standard (min) Absorption Extent Standard
Heat (I / g) of Cure Error Heat (J / g) of Cure Error
0 407 0% 5.1% 0 406 0% 3.6%
5 382 6% 4.4% 5 346 15% 2.4%
20 203 50% 1.0% 10 260 36% 1.4%
30 156 62% 3.7% 20 209 49% 1.3%
40 117 71% 6.6% 30 156 62% 2.6%
60 22 95% 0.4% 60 56 86% 0.4%
80 13 97% 0.3% 80 35 91% 2.1%
100 14 96% 0.3% 100 15 96% 0.2%

 

From Table 3.4 and Figure 3.10, it can be seen that curing curves under both
modes, TM 020 and TM 021, have the typical shape of autocatalytic reaction. The initial
curing rate was slow due to lack of catalytic hydroxide groups. As the reaction

proceeded, hydroxide groups were generated and the maximum curing rate occurred at

the extent of cure of around 30 to 40%.

 

 

 

 

100% 2 g,
80% ‘ // 1/
//‘
60% r / ‘
e ,4
40% “ . TM 021
20% a i 4 TM 020
0% . T 7
0 3O 60 90 120
Time (Min)

Figure 3.10 Extent of cure vs. curing time

59

POLYMATH was used in data regression to determine the kinetic parameters.
The Levenberg—Marquardt algorithm was used to ﬁnd the parameter values, which
minimizes the sum of squares of the errors. The kinetic parameters in Equation 3.10 were
ﬁrst assigned initial values.

d m n
T? =(k1+k2a )(1—a) (3.10)

where k1 is the non-catalytic polymerization reaction rate constant, k2 is the autocatalytic
polymerization reaction rate constant, m is the autocatalytic polymerization reaction
order, and n is the non-catalytic polymerization reaction order. The calculated values of
parameters are shown in Table 3.5. k2 is higher than k1. This veriﬁes that the epoxy

curing is mainly governed by autocatalytic reaction mechanism.

Table 3.5 Values of the kinetic parameters

 

m 0.67

n 1.30

k, 0.026
kl/k2 0.7

 

3.6 Conclusions

The curing of DGEBA/DDS epoxy resin system under microwave radiation at
145 °C was governed by an autocatalyzed reaction mechanism. A kinetic model can be
used to describe the curing progress.

da
—=k+kaml—a"
dt (1 2 )( )

60

where k1 is the non-catalytic polymerization reaction rate constant, k2 is the autocatalytic
polymerization reaction rate constant, m is the autocatalyzed polymerization reaction

order, and n is the non-catalyzed polymerization reaction order.

61

CHAPTER 4 DIELECTRIC ANALYSIS OF CURING EPOXY RESINS

4.1 Introduction

The study of the dielectric properties during polymerizations of epoxy resins
under microwave irradiation has both fundamental and practical interests. It provides
useful information about details of the curing reactions and, thus, has been developed as a
nondestructive testing technique to monitor the curing processes.

A microwave cavity technique is used here to obtain dielectric data and to
diagnose the microwave (MW) curing of epoxy/amine resins at 2.45 GHz. Dielectric
properties of epoxy/amine resins are not from the ionic contribution, but due to the
presence of polar groups in the molecules. Microwave energy can be directly absorbed by
these polar groups to cause localized heating, which initiates the curing processing. The
absorption occurs primarily at epoxy and amine groups during the early stages of heating.
These polar groups are consumed during cure while hydroxyl groups are formed, which
also absorb microwave radiation and, subsequently, react with epoxy groups. The
mobility of polar chains decreased as the crosslinking reaction proceeds. The effect of
changing mobility and population of these polar groups on the dielectric properties of the
system can be determined by dielectric measurement. The rate of absorption of
microwave energy is determined by dielectric loss factor and electric ﬁeld strength. The
crosslinked molecules, normally with low dielectric properties, do not absorb microwave
energy as readily as the monomers with high dielectric properties. Therefore, dielectric
properties can be an index of energy dissipation in the resins and an index of extents of

reaction. The advantages of using microwave heating and dielectric diagnostics to initiate

62

and monitor the curing process are: (1) selective and controlled direct heating due to
absorption of microwave energy by polar groups; (2) increased control of materials
temperature-time proﬁle and input power to decrease thermal degradation and optimize
the cure cycle. These advantages may cause microwave cured materials to have superior
mechanical characteristics by chemically modifying the polar groups or by intelligently
selecting the microwave power cycle conditions.

Epoxy resin is one of the most versatile materials used in such areas as general
purpose, electrical, and aerospace. In the general-purpose area, the diglycidyl ether of
bisphenol A (DGEBA) epoxy resin is the preferred material and has been used as pipers,
adhesives, protective coatings, and electrical insulations. Until now, the studies on
DGEBA epoxy resins primarily focus on measuring dielectric properties as a function of
frequency at a constant temperature. However, microwave heating apparatus usually
Operates at a constant frequency, e. g. 2.45 GHz for all domestic microwave ovens. The
Federal Communications Commission (FCC) allocated a number of microwave
frequencies for Industrial, Scientiﬁc and Medical applications (ISM), among which 2.45
GHz is the major operating frequency worldwide. In order to improve industrial
applications of microwave processing of epoxy resins, understanding the dielectric
properties and relaxation during cure at an ISM microwave frequency is essential. The
objective of this chapter is to do the dielectric analysis of the curing systems of DGEBA
and different curing agents at 2.45 GHz over a temperature range. The dielectric
properties changing with the reaction were ﬁt by models and the evolutions of dielectric
parameters, e. g. dielectric strength, shape parameter, and relaxation activation energy,

were analyzed.

63

4.2 Background on Dielectric Analysis

4.2.1 Fundamental Theories for Dielectric Relaxation

In linear systems a response and a stimulus are proportional to one another in
equilibrium. Relaxation is a delayed response to a changing stimulus in such a system.
Dielectric relaxation is relaxation occurring in dielectrics. The stimulus is always an
electrical ﬁeld, and the response is a polarization [245]. If a material is under a dynamic
electric ﬁeld, the linear response to this ﬁeld is deﬁned as complex dielectric constant.
Complex dielectric constant of the material has dipolar contribution and ionic
contribution. For non-conducting materials, complex dielectric constant is mainly

contributed by presence of the dipoles [218].

4.2.1.1 Complex Dielectric Constant

The complex dielectric constant (or permittivity) is expressed as follows:

is" = £‘—s" j = (a; — 1::th (4.1)
where the unit is F/m. s' is the dielectric constant, 8" is the dielectric loss factor, so

(=8.85 x 10'12 F/m) is the permittivity of free space, s'r is the relative dielectric constant,

and s"r is the relative dielectric loss factor. Although s'r is referred to as the dielectric

constant in many papers, it is a function of temperature, pressure, humidity, and other
conditions. In this paper, 8' refers to s'r while 8" refers to 8”,.

The ﬁrst effort to explain absorption of electromagnetic power by polar materials

Was that of Peter Debye [246]. Absorption occurs when dipoles, ordered with respect to

an electric ﬁeld, relax to a random orientation. The energy required to hold the dipoles in

an ordered orientation is released as thermal energy when the dipoles relax. The
relaxation time is an indication of the amount of time required for a collection of dipoles

in an electric ﬁeld to revert to a random orientation once the ﬁeld is removed. Following

is the Debye model:
8* = s +——(£° —£°°)
l+wy
8r = 800 + lie-$0,} (4.2)
1+ (0 r

n (80 — £00 )0”
° = "T

1 + a) r
where s' is the dielectric constant, a" is the dielectric loss factor, j is the imaginary unit, to

(=21tf, f is the oscillator frequency in Hz) is the radial frequency of the electric ﬁeld in s",

t is the relaxation time in s, so is the low frequency dielectric constant, and sq, is the high
frequency dielectric constant. The functions represented by Equation 4.2 are illustrated in

Figure 4.1, for the speciﬁc values so =8 and s,o =2.

 

 

 

 

 

0.01 0.1 1 10 100

Figure 4.1 Schematic diagram of the Debye model

65

Experimentally, one usually measures the quantities s' and the loss tangent, which

is deﬁned in Equation 4.3. Loss tangent can also be expressed as Equation 4.4, according

 

to the Debye equations.
tan a = % (4.3)
tan 5 = (80 — 89;”: (4.4)
80 + 80060 1’

Although Debye model works well for systems containing small molecules in gas
media or media with low viscosity, it does not work for complicated systems including
polymers. Therefore, Cole and Cole proposed an empirical equation for a distribution of
the relaxation time process which considered that each dipole had its own relaxation
[247]. This distribution of dielectric relaxation time process can be represented and
derived using a generalized Maxwell relaxation model. Davidson and Cole proposed
another empirical equation from the experimental observation of the asymmetric
dielectric dispersion curve due to interaction of neighboring dipoles [248]. The Cole-Cole

model and Davidson-Cole model are shown in Equations 4.5 and 4.6, respectively.

 

£*_goo: (SO—80(1))
[1+(J'wT) '0]

(80 — 8.011 + (wow sin(g—)]
s' = 800 + 2 (4.5)
1 + 2(m)"a sun?) + (m)2(1’“)

 

(80 — 6.. km)”z cost?)

 

s" =
l + 2(wr)l"a QM?) + (021)2(1’a)

66

g*'£w : (60 ’800)
cum)”

8, = a. + (so-6m)COS(/g6’)

(Heinz)?

 

(4.6)

 

n (80 " £00)S1n(ﬂ0)
° = a
(new)?

19 = arc tan(a)r)
where or and B are the shape parameters with a range from 0 to 1. The functions
represented by the Davidson-Cole model in Equation 4.6 are illustrated in Figure 4.2, for

the speciﬁc values, so =8,s.,O =2, and B=0.2.

 

 

 

 

O I T l

0.01 l (or 100 10000 1000000

Figure 4.2 Schematic diagram of the Davidson-Cole model

67

Havriliak and Negami combined the Davidson-Cole and Cole-Cole models, and
proposed a general empirical model, which is known as Havriliak-Negami (H-N) model

[249,250)

(60 —500) 7

[1+(jwr)1_a}0

where a is dispersion width parameter with the range 0 S a S 1, and B is dispersion

* _
8 —8w _

 

(4.7)

skewness parameter with the range 0 S B S 1. If B equals to 1, the H-N model is Cole-
Cole model; If (1 equals to O, the H-N model is Davidson-Cole model. Using the complex
identity in Equation 4.8 and DeMoivere’s theorem, the H-N model in Equation 4.7 can be
separated into the real part and the imaginary part, described in Equation 4.9.

jPrr

jP =6 2 (4.8)

. (So-8w)008(,39)

s =800+

[1 + 2(wz)1‘a MIX?) + (021)2(1—a) ]7

8.: (So—866)Sin(,59) ,3 (4.9)

[1 + 2am)” swag) + (m)2(1‘“)]3

(mum) cos(°—”)
t9 = arc tan 2

 

1+ (coma—a) sin(92£)

68

4.2.1.2 Typical Graphs

Complex dielectric constant is a function of frequency, temperature, pressure, and
material structure and composition. Two most important factors in the dielectric spectrum
are temperature and frequency. Four types of graphs based on these two variables are
ususlly used to present dielectric data:

(1) Cole-Cole plot (Arc diagram). The Cole-Cole plot is obtained by plotting
the dielectrc loss factor versus dielectric constant. The Debye model is a semicircle in the
Cole-Cole plot. This plot is usually used to verify the above dielectric models. The
following ﬁgures are typical Cole-Cole plots for the Debye model and the Davidson-Cole

model, for the speciﬁc values, so =8,s.,o =2, and B=0.2.

 

 

 

 

 

 

 

 

Figure 4.4 Schematic diagram of Cole-Cole plot of the Davidson-Cole model

69

(2) 1 Complex dielectric constant versus frequency. Plotting complex dielectric
constant versus frequency at constant temperature is to illustrate the time-temperature
superposition.

(3) The relaxation time at maximum loss versus the reciprocal of temperature.
The purpose of this graph is to identify the type of the molecular relaxation motions and
to calculate activation energy for the relaxation transition. In practice, frequency is
usually used as a parameter and temperature as a variable.

(4) Complex dielectric constant versus temperature. Complex dielectric
constant versus temperature is plotted at a ﬁxed frequency. Thus, the dielectric spectra
can be easily compared with mechanical relaxation spectra, and thermomechanical,
dilatometric, and differential scanning calorimetric curves which are also measured as a

function of temperature.

4.2.1.3 Dielectric Relaxation Time
Debye gave the relaxation time for a spherical polar molecule as

T=3V77

H (4.10)

where V is the molecular volume, n is the viscosity, k is Boltzmann’s constant, and T is
the temperature [246].

Kauzmann and Erying [251, 252] expressed the relaxation process of local
molecular transitions as a ﬁrst-order reaction with an activation energy, which ﬁtted to

the Arrhenius expression
( a I
r=Axe RT (4.11)

70

where E, is the activation energy in J/mol, R is the gas constant in J/mol-K, T is the
temperature in K, and A is the relaxation time in the high temperature limit in 3.

Normally, the dipoles will have a short relaxation time at high temperatures, in
low viscous medium (low intermolecular attraction), and with small molecular sizes (or
small dipole sizes). Large molecules in highly viscous (or solid or networked) media can
be expected to have large relaxation times, and the relaxation time would decrease as the
temperature increases.

For structural transitions, it is generally recognized that the temperature
dependence of the relaxation time above glass transition temperature is governed by WLF

equation [253]

_A(T-Tg )]

B+T—Tg

r = 2'(Tg )e[ (4.12)

where T is temperature, Tg is glass transition temperature, A and B are constants.

4.2.1.4 Typical Dielectric Relaxation Processes

Three major types of molecular relaxation processes are classiﬁed in the structural
transition region of materials [218]:

(1) or" process is the motion of larger crystallite on the melting phase.

(2) 01' process is the motion of macromolecules in crystalline region at the
beginning of the melting process.

(3) or process. This process randomizes the dipole moments through random
Brownian motion of large segments of a polymer chain. These relaxations take place at

low ﬁequencies and, therefore, with long relaxation times. The characteristic relaxation

71

time increased over that of the liquid state due to reduced molecular motion in the solid
state. a process is a principal relaxation process.

Several secondary types of molecular motions below the principal relaxation
process are deﬁned in the local molecular transition regions:

(4) B process. This process occurs at higher ﬁequency. It randomizes the
dipole moments by reorientation of individual dipoles in the molecule chain, involving
segmented motion of a small section of the polymer chain than a process. A plot of the
loss tangent versus temperature would yield a broad, ﬂat distribution due to the fact that
there are a number of different dipoles involved in a B process, each with its own
characteristic absorption. B process occurs at higher ﬁ'equencies than a process.

(5) y process is motion of individual groups of atoms in or attahced to the
backbone chains at higher frequencies in the microwave region.

(6) 7' process is motion of groups of atoms in branches or at the ends.

4.2.2 Literature Review

Dielectric techniques under microwave irradiation have been applied extensively
to monitor thermal curing of thermosetting resins and composites by several investigators
[254-262]. These dielectric measurements are usually performed by measuring an
admittance of the materials placed between two conducting electrodes under microwave
irradiation over frequencies less than 10 MHz. Signiﬁcant data variations in dielectric
measurement due to the interference of electrode polarization at low frequency or glass

transition at high cure temperatures have been reported in the literatures [261, 262].

72

Studies on the dielectric properties of epoxy resins have received considerable
attention [5, 263-281]. One main relaxation, a, and two secondary relaxations, B and y,
were found in DGEBA [263]. The structural a relaxation, which takes place at low
frequencies and high temperatures, was due to the Brownian motions of the polymer
chain [264-269]. The secondary B process, occuning at higher frequencies and lower
temperatures, was assigned to a smaller section of the polymer chain than the a process,
e.g. the hydroxyether group —CH2-CHOH-CH2-O- [270]. The secondary y relaxation,
which is found located at higher frequencies than the B process, was the motion of
individual groups of atoms, such as, epoxide groups, amine groups, and hydroxyl groups
[5, 266-269]. The evolution of the dielectric constant and the dielectric loss factor during
cure is related to the disappearance of dielectric dipoles in the reactants [273, 276-278]
while the changes in the relaxation time depend on the viscosity of the reacting mixture
[258, 259, 279-281].

Koike et al. studied the dielectric properties of a series of epoxy prepolymers
including DGEBA [264, 265]. a relaxation was found in all studied prepolymers and ﬁts
to the empirical H-N equations. The WLF equation can be used to describe the
relationship between the relaxation time and the temperature higher than the glass
transition temperature.

Researchers at University of Pisa, Italy studied the dielectric properties of
DGEBA and CGE (a monoepoxide, cresyl-glycidyl-ether) at different temperatures and
frequencies [266-269, 274-277]. Two relaxation processes were found [275]. One is the
structural a relaxation process, freezing at the glass transition. The glass transition

temperature of DGEBA is -16°C. The other one is a secondary y relaxation. When the

73

temperature increases, a relaxation shifts toward higher frequencies and, eventually,
merged with the y relaxation. The spectra of the two relaxations show a non-Debye and
asymmetric behavior. The experimental data ﬁt a superposition of two H-N equations for
the two relaxation processes [266]. The relaxation time of y relaxation process were ﬁtted
by the Arrhenius equation, while that of or relaxation process was non-Arrhenius and can

be described by the empirical Vogel-Fulcher-Tammen equation [266].

4.3 Experimental

4.3.] Experimental Systems

The microwave diagnostic and processing system at MSU was developed by Dr.
Jow. The details of the system and the single mode perturbation method can be found in
his dissertation [218]. A switch between the heating and diagnostic systems was
developed by Mr. Charvat [282]. A software program, called Diane which was developed
by Dr. Jow at MSU, is used to calculate the dielectric constant and dielectric loss factor
of materials. Here is a brief description of the system.

The microwave diagnostic and processing system consists of a microwave
external circuit (an energy source, transmission lines, and the coupling probe), diagnostic
elements (e.g. an X-Y oscilloscope, power meters, the E-ﬁeld diagnostic probe, and a
temperature sensor device), and the loaded cavity (e. g. the cavity and the loaded
material).The system is shown in Figure 4.5. A cylindrical TM 012 mode cavity is used
to process epoxy resins at a frequency of 2.45 GHz. A ﬂuoroptic temperature sensing

device (Luxtron 750). which does not absorb electromagnetic energy or interfere with

74

electromagnetic ﬁelds . is used to on-line monitor the material temperature. The

ﬂuoroptic probe is protected by a 3 mm O.D. pyrex capillary tube. A cylindrical Teﬂon
sample holder is also required to contain the liquid samples. The cylindrical geometry of
both the cavity and loaded materials is selected in order to facilitate diagnosis, modeling,

and theoretical analysis.

Fluoroptic Temperature
Sensing System

 

 

 

 

 

 

 

 

 

 

Power Meter
Pio
Microwave
Source Pi = P“ _ P Loaded
TWT . 1 r , , Material
Ampliﬁer Clrculator Coaxral Coupllng
. Cable Probe

 

 

 

 

 

 

 

   
  

 

 

 

' < Cavity
Pb

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Directional
Coupler Pr E Field
Diagnostic
Reference X Y 20 dB Attenuator Probe
Signal
Crystal Detector
X-Y Oscillosco e
p Pro Pbo
Power Meter Power Meter

Figure 4.5 Schematic illustration of the microwave processing and diagnostic system

4.3.2 Experimental Materials
The epoxy resin used in this study was diglycidyl ether of bisphenol A (DGEBA,
DER 332 by Dow Chemical). The curing agents were 3,3-diaminodipheny1 sulfone (DDS

by TCI America), a difunctional primary amine (J effamine D-230 by Huntsman), m-

75

phenylenediamine (mPDA by Sigma-Aldrich), and diethyltoluenediamine (Epikure W by
Resolution Performance Product). The chemical structures and properties of the reactants

are shown in Table 4.1.

Table 4.1 Properties of the epoxy resin and curing agents

 

Name Chemical Structure Epoxy/amine Density
equivalent at 25°C
weight (g/ml)

 

CH,
DGEBA OHO/ <O| 171-175 1.16
O
[>—/ CH,

DDS sz‘ﬁ‘OH 62 1.33
S
I I
HZN /20H NH 2
Jeffamine \CH/ OCH/ 2&0er 60 0 948
[>230 | | x=2-6 '
CH 3 CH3
HZN NH2
mPDA U 27 1.14
Epikure
w 43-46 1.02

 

 

4.3.3 Sample Preparation
All of the materials were used as received without further puriﬁcation. In
preparing neat DGEBA/DDS epoxy resins, stoichiometric DGEBA and DDS (2.79:1 by

weight) were mixed in a glass beaker. The mixtures were well stirred by hand in a 130°C

76

oil bath until the DDS was completely dissolved (in approximately ﬁve minutes). Finally,
the resins were degassed at 0.02 bar at 100°C for ﬁve minutes.

In order to prepare neat DGEBA/Jeffamine epoxy resins, DGEBA was ﬁrst
preheated in a glass beaker at 50°C to melt any crystals and then a stoichiometric
J effamine (weight ratio of DGEBAzJeffamine 2.88: 1) was added. The mixture was stirred
for ﬁve minutes with a magnetic bar at room temperature and degassed at 0.02 bar at
room temperature for ﬁve minutes.

To prepare neat DGEBA/mPDA epoxy resins, stoichiometric DGEBA and mPDA
(6.4:1 by weight) were mixed in a glass beaker. The mixtures were well stirred with a
magnetic bar at 65°C for ﬁve minutes and then degassed at 0.02 bar at room temperature
for ﬁve minutes.

To prepare neat DGEBA/W epoxy resins, stoichiometric DGEBA and W (3.89:1
by weight) were mixed in a glass beaker. The mixtures were well stirred with a magnetic
bar at 50°C for ﬁve minutes and then degassed at 0.02 bar at room temperature for ﬁve

minutes.

4.3.4 Measurements

An empty Teﬂon holder with a ﬂuoroptic probe was located at the position of the
strongest electric ﬁeld in the TM 012 mode cavity by a cotton thread. This cavity was
critically coupled with a microwave external circuit and initial measurements were made
by single and swept ﬁ'equency methods at 2.45 GHz. The Teﬂon holder was removed

from the cavity.

77

The degassed liquid epoxy resins were poured into the Teﬂon holder. The sample
volumes (about 2.00 cm3) of epoxy resins were experimentally determined so that the
resonant frequency shift was much less than the resonant frequency throughout the entire
process. The Teﬂon holder with a ﬂuoroptic probe and epoxy resin sample was relocated
at the position of the highest electric ﬁeld for the TM012 cavity mode at 2.45 GHz.

The single ﬁ'equency microwave curing and diagnostic system was used to heat
the liquid samples in the Teﬂon holder. The fresh DGEBA/DDS samples were heated to
react at 145°C for speciﬁed reaction time periods, e.g. 1, 5, and 20 minutes, with the
exception of those for the 0% cured epoxy resin, which were heated to 100°C. The curing
temperatures for the DGEBA/Jeffamine, DGEBA/mPDA, DGEBA/W systems were 90,
110, and 160 °C, respectively. And the peak temperatures for the unreacted epoxy resins
were 80, 90, and 100°C. Thereafter, the single frequency microwave curing system was
switched to a low-power swept frequency diagnostic system by changing the switch
position. Measurements of temperature and dielectric properties using the swept
frequency method were made during free convective cooling of the samples. The cooled
samples were analyzed with a Differential Scanning Calorimeter (DSC) to determine the
residual heat of reaction per gram and thus the extents of cure. The calculation of extent
of cure of epoxy resins is illustrated in Chapter 3.4.1.The reported extent of cure data is

an average value at least three samples.

4.4 Results and Discussion

78

4.4.1 DGEBA/DDS System
Dielectric properties of DGEBA, DDS, and uncured DGEBA/DDS mixture at

2.45 GHz and different temperatures are shown in Figure 4.6.

 

 

 

 

 

 

 

 

6.5
.1
l o 0 ” O .
d o . 1:
5.5 — . 9 a
- 4 0 °
to a o u
_ 0. AD
4-5: o Mm ° oDGEBA
_ it" n DGEBA/DDS
* ADDS
3.5 l l I I 1 T T I I I I I I I T I I I I
20 40 6O 80 100 120
Terrperature (°C)
1.0
0.8 : 0’ ° 0
: o . Go
" o
- 0.6j o a u D O .
-00 : Q U
0.4 r U
j 0 o ODGEBA
02 :1 D D D DGEBA/DDS
. l A ° “ ‘ ‘ ADDS
0.0 d T T T I I I I I r I I I T I I I 7 I I
20 40 6O 80 100 120

Tenperature (0 C )

Figure 4.6 Dielectric properties vs. temperature for DGEBA, DDS, and uncured
DGEBA/DDS mixture

79

The dielectric constant of DGEBA increases as the temperature increases from 20
to 80°C, remains stable around 80 to 100°C, and then decreases. The dielectric loss factor
increases ﬁrst and then decreases with a peak value around 70 to 80°C. The dielectric
constant and dielectric loss factor of DDS and uncured DGEBA/DDS mixture increase as
the temperature increases. The dielectric constant of DGEBA/DDS is smaller than that of
DGEBA but similar to that of DDS while the dielectric loss factor of DGEBA/DDS is
between that of DGEBA and DDS. The added DDS increases the viscosity of DGEBA
matrix and hinders the relaxation time of DGEBA. Furthermore, the dielectric properties
of DDS are smaller than those of DGEBA. Therefore, the dielectric properties of
DGEBA/DDS are less than those of DGEBA.

The dielectric constant and dielectric loss factor of reacting DGEBA/DDS epoxy
resins at different temperatures are shown in Figure 4.7. The extents of cure were
calculated ﬁom DSC data, which are shown in Table 4.2. Due to the ununiformity of
microwave heating, a distribution of extents of cure exists in the samples. From Figure
4.7, it is found that both dielectric constant and dielectric loss factor increase as the
temperature increases and decrease as the curing reaction proceeds. According to the
Debye model in Equation 4.2, the dielectric properties increase as the relaxation time
decreases. Generally, the relaxation time decreases as the temperature increases. Hence,

the dielectric properties should increase as the temperature increases.

80

 

 

 

 

4
i A
6: D A o
..;u ‘00 +
7.05: ODD; % o +
00 A0 +
1 0° to +
j “00’ + + I . X
44‘“ + . I x
— + + 'x x x A A A
I ' *A A A A A
3 I IIrITITI I III FIITfr I I
20 40 6O 80 100 120
Terrperature(°C)
0.8
: . a .
0.6: . A
q o a ‘
J 0 an 0 °
$0.44 o a ‘ 0 ° +
: o a ‘ 4 o + +
_ .%: ‘0 0 + .- I
0.2~ ’39 o + +. I
4 ° “’1 ' x X
?0 .+. x X X A
:IAngxA X A A A A
0.0 l l 1'1 III I111I IIII
20 40 60 80 100 120
Ten'peratm'e(°C)

 

 

 

 

o 0%

a 9%

A 14%
o 22%
+ 37%
I 42%
x 64%
A 79%

o 0%

D 9%

A 14%
o 22%
+ 37%
I 42%
x 64%
A 79%

Figure 4.7 Temperature dependence of dielectric properties for the DGEBA/DDS epoxy

resins at different extents of cure (%)

Table 4.2 DSC results of the curing DGEBA/DDS system

 

 

Reaction Heat (J / g) 432 387 395 324 249 168 203 72
448 436 415 350 253 299 153 109
419 403 360 362 269 262 58 63
439 370 330 370 294 222 57 64
429 365 363 290 308 297 319 157
Average 433 392 372 339 275 250 158 93
Extent of Cure 0% 9% 14% 22% 37% 42% 64% 79%
Standard Error 1% 3% 3% 3% 3% 6% l 1% 4%

 

81

The reaction mechanism of the DGEBA/DDS epoxy resin system is shown in
Figure 4.8 [224]. DGEBA reacts with amines via a ring-opening mechanism. Two-step
reactions occur during cure. In the ﬁrst step, an epoxy group reacts with a primary amine
to form a hydroxyl group and a secondary amine. In the second step, the formed
secondary amine reacts with another epoxy group to produce a hydroxyl group and a
tertiary amine [281]. The progress of the reaction is deﬁned in terms of extent of cure.
The viscosity of the reacting systems rise as the polymerization goes on [281] and two

distinguishable transitions, i.e. the gelation and the vitriﬁcation, are crossed.

R, R2
R1—NH2 + l> R2 > \NH
0
OH

 

 

OH
R,\ R2 R2
NH R —> R N
+ 3 1 R3
0
OH
OH

Figure 4.8 DGEBA resin curing mechanism

In general, three relaxation processes, or, B, and 7, may occur [263-281]. The a
relaxation, which takes place at low ﬁequencies, randomizes the dipole moments through
the Brownian motion of whole molecules [266]. Due to the high frequency used in the

experiments and crosslinking between DGEBA and DDS, the whole molecules of the

82

system may not relax. Therefore, the or relaxation does not dominate in the reacting
system. The B relaxation occurs at a higher frequency, being attributed to the hydroxyl
groups attached to the backbone of polymers [263]. Considering the low concentration of
the hydroxyl groups in the reacting system, the B relaxation gives a negligible
contribution to the relaxation processes. Therefore, the relaxations that occurred in the
subject material are mainly y relaxation.

To interpret the evolution of the dielectric properties during the curing reaction,
one should know the dielectric behavior of all involved dipolar groups, which are,
however, too complicated to differentiate. Since the dipolar groups within the reacting
system and their dynamics are substantially similar, the dielectric properties of the
reacting system reﬂect the combination of all the dipolar groups involved in the reaction.

The y relaxation is the motion of dipolar groups of atoms, which should include:
epoxy and amine groups with the unreacted DGEBA and DDS; dipolar groups with the
intermediate products and ﬁnal polymers, e.g., -NH- and —OH. As the reaction goes on,
epoxy dipoles disappear, amine dipoles change to -NH- and =N-, and new dipoles, such
as hydroxyl groups, appear. Overall, the total number of the dipolar groups within the
reacting system is stable. Taking into account decreasing dielectric properties during
cure, it is reasonable to suppose that the contribution of different dipolar groups to the
relaxation is different. The unreacted epoxy groups, -O-, and amine groups, -NH2-, within
the reaction system are the main driving forces for the apparent combined 7 relaxation
observed in this study. The disappearance of the epoxy and amine groups during the
reaction is one reason accounting for the changes of the dielectric properties. Other

researchers reported similar results [273, 276-278]. However, it may not be the only

83

reason. According to the classical Debye theory, the dielectric dipoles are regarded as
spheres in a continuous medium having a macroscopic viscosity [246]. Schonhals and
Schlosser studied the dielectric relaxation in polymeric solids and argued that the
environment with high viscosity hindered the diffusion process of dipolar groups, causing
the dielectric behavior of polymers far from ideal Debye materials [283]. In this study the
transitions from liquid to gel and then to solid was observed during the cure reactions for
all the four reacting systems. The increasing viscosity of the reacting systems should
hinder the mobility of the dipolar groups associated with the relaxation and cause the
relaxation time to increase [258, 259, 279-281]. The two reasons accounting for the
evolution of the dielectric constant and the dielectric loss factor are a decrease in the
number of the dipolar groups in the reactants and an increase in the viscosity during the
reaction.

The Davidson-Cole model [248] can be used to describe the dielectric behavior of

DGEBA epoxy resins:

: (50 - £00)
(1+J'm)"

, + (£0 — £00 )cos(n 6?)

 

 

e = 800 n
<1+<wr>2>5
(4.13)
8,, ___ (so — 800 )sin(n 6?)
(1 + (we?
6 z are tan(a)r)

where n is the shape parameter with a range of 0 S n5 1.

84

Consideration of the experimental dielectric data leads to a simplifying
assumption. The observation that the dielectric constant and loss factor increased as the
temperature increases happens only when the product (an) is greater than unity, as shown
in Figures 4.1 and 4.2. Therefore, it is logical to assume that the value of ((0102 is much
greater than one. In that case, the expressions of 8' and 8"in Equation 4.13 reduce to:

(£0 — £00 )cos(n %)

n

 

a = 8m +
(M)

(4.14)

(80 - £00 )Sin(n 325)

(601)"

 

Combining the calculation rules of complex numbers in Equations 4.15 and 4.16,

the complex dielectric constant can be derived in Equation 4.17.

 

1 . . _ 1
:[cos(—t) + _] srn(—t)] .. 40080) + mum] (4.15)
[cos(t) + j sin(t)]" = cos(nt) + j sin(nt) (4.16)
8* = £'——£" j
= 600 + Mk0“): 5) — j sin(n 35)]
(607)" 2 2
zngEO’gw) 1 (4.17)

 

(01)" it. ' ' f.
cos(n 2)+ Jsrn(n 2)

 

= £00 + (£0 _ £00) 1
W)" 1cos<§>+ jsin(12[-)]"
: £00 + (£0 - £00)

UM)"

85

The proposed simpliﬁed Davidson-Cole expression to describe the dielectric

properties of DGEBA/DDS system is shown in Equation 4.18.

(80 - Eco )COSOI %)

(607)"

 

(4.18)

(80 7' £00 )Sin(n %)

(607)"

 

where 8* is the complex dielectric constant, 8' is the dielectric constant, e" is the

dielectric loss factor, j is the imaginary unit, or (=21tf, f is the oscillator frequency in Hz)
is the radial ﬁ'equency of the electric ﬁeld in s", t is the relaxation time in s, so is the low
frequency dielectric constant, an is the high frequency dielectric constant, and n is the
shape parameter with a range of 0 S n3 1.

It is observed that Equation 4.19 can be derived from Equation 4.18.

 

tan(n£) = g
2 SL500

(4.19)
The proposed simpliﬁed Davidson-Cole expression in Equation 4.18 is similar to,

but more speciﬁc than, a model proposed by Schonhals and Schlosser (S-S) [283, 284]:

£"((:)) ~ 02'" ((0 << 020) (4.20)

n

£"(a)) ~ (0’ (a) >> (00)

where m and n are shape parameters, and (no is the angular ﬁequency at maximum 3". If

Equation 4.18 is correct, a plot of 3" versus 8' should yield a straight line. Figure 4.9

86

shows 8” versus 8’ at different extents. Straight lines represent the calculated data from

the proposed simple model and points represent experimental data.

0.7 -

. /
5%

0.5— /‘/X//:/'.

 

 

 

0.4 -
= A 0%
60
X x 9%
0.3 - , /
/ x 14%
// . 22%
0.2 - I 37%
n 42%
0.1 _ o 64%
A 79%
0.0 ' l I I I I 1 l
3 4 5 6 7

Figure 4.9 2" vs. 8' for the DGEBA/DDS epoxy resins at different extents of cure (%)

The 7 relaxation time should ﬁt the Arrhenius expression, shown in Equation
4.11. Combined with Equation 4.11, Equation 4.18 can be rewrite in Equation 4.21. If the
proposed simple model is correct and the Arrhenius expression is applicable to the 'y
relaxation, plots of In (sham) and ln(s") versus I/T should yield straight lines. Figures

4.10 and 4.11 demonstrate that this is indeed the case.

87

(£0 — goo )cos(n 1;) Ea

 

 

 

 

ln(8 £00) — ln (0)14)" n RT
. 7r
1n(8')=1n (£0 ”500)51n(n'§) —n-1-;i
(604)" RT
1.5 ..
A 0%
1.0 r x 9%
A x 14%
fi’ 05 d o 22%
30.0 _ I 37%
a 42%
-0.5 — o 64%
d A 79%
-l.0 1 I I I 1

 

2.4 2.6 2.8 3.0 3.2 3.4
lOOO/I‘

Figure 4.10 In (858,0) vs. lOOO/T for the curing DGEBA/DDS system

 

 

A 0%
-0.5 _ x 9%
x 14%
E -l.5 N . 22%
I 37%
'25“ % ° 42%
° 64%
’3'5 ‘ ‘ r ’ A 79%
2.4 2.6 2.8 3.0 3.2 3.4

1 000/T

Figure 4.11 In (8") vs. 1000/T of the curing DGEBA/DDS system

88

(4.21)

2.1,, n, E3, and a relation between A and so in Equations 4.11 and 4.18 can be
ﬁgured out from the slopes and intercepts of straight lines in Figures 4.9-4.11. The value
of so was estimated from the references [266, 269] and then A was calculated. Figure 4.12
shows the comparison of the experimental and calculated data of the reacting
DGEBA/DDS system, where curves represent the calculated data from the simpliﬁed

Davidson-Cole expression and points represent experimental data.

 

 

 

 

 

 

 

 

 

7 —<
: o 0%
6 E /‘A a 9%
i /’ A 14%
’w 5 : /% o 22%
“p32?” + 37%
4 M - 42%
Z M x 64%
3‘1,.,,.,,,,,,,,,,,,m”9%
20 40 60 80 100 120
Temperature (°C)
0.8 W . 0%
j u 9%
0'6 : A 14%
= : o 22%
to 0.4 E + 37%
— l 42%
0.2 r
: W x 64%
0_0°HHHHHHHWYHﬂ A79%
20 40 60 80 100 120
Temperature (°C)

Figure 4.12 Comparison between the experimental and calculated dielectric properties of
the curing DGEBA/DDS system

89

The proposed simple model ﬁts experimental results exactly. The difference

between the experimental and the calculated data is caused by the calculation error based
on the approximation functions and the experimental error due to ﬂuctuation of measured
data including temperature, resonant frequency, and half-power frequency bandwidth.

To verify the assumption of the proposed expression, comparison of the
experimental data and Cole-Cole plots of the calculated data is shown in Figure 4.13,
where curves represent the calculated data by the Davidson-Cole model and points
represent experimental data. The experimental data are located in the left side of the

spectra, where tangent lines can be used to represent curves approximately. The tangent

lines are related to the simpliﬁed Davidson-Cole expression.

 

 

 

10 4 ‘ 0%
o / \x
, \
-1 . 9% Ir, \\
/ ,—\
I / \
0.8 — ‘ 14% I I ’1’ ’_,\ \\ \\\
/ ,l l” \\ \ \
r . 22% ,’ , ,1 ’__ \ X \\
[IA/"l ’x \‘\ \‘ ‘\ \
__, z ’ \
= 0.6 + 37% ’, [Ag/2" _\ \\\ \\ \‘ 1‘
w _ o IV,*’¢I;L-\ ‘\ I 1 \‘ 1‘
\ \
.. 42%» r "'2’“ \ x 1 x x
0.4 — . , \ .
t \ ‘ ‘ \
—< -\ \ \‘ \l ‘ l l
‘\ ‘\ , \ \ '| l.
l l I
0.2 - \ ‘1 ‘. x : '. 1
1 \ l 1 '
A | 1
l I I . : I I
00 1 l I l 1 I1 1 l j1 1
2 4 6 8 10

Figure 4.13 Cole-Cole plots for the curing DGEBA/DDS system

90

Table 4.3 shows the calculated values of all parameters the simpliﬁed Davidson-
Cole expression, and the Arrhenius rate law. The parameter n describes the skewness of
the dispersion of the relaxation times, which increases as n ranges from unity to zero. The
value of n for an ideal Debye material is unity while that for polymeric solutions is
around 0.5 [285]. As the reaction goes on, the parameter n decreases linearly, from 0.17
down to 0.07 (see Figure 4.14). The parameter 11, according to Schonhals and Schlosser,
is related to the intrarnolecular movement for the main a relaxation of polymers [283,
284]. However, in this experiment it is shown that n is mainly connected with motion of
dipolar groups for the secondary y relaxation. During the reaction the motion of dipolar
groups is hindered by the medium with increasing viscosity caused by curing epoxy
resins. Therefore, 11 decreases. The rationale is similar to the argument that n is connected
with the local chain dynamics of a polymer and deceases in the range of 0-0.5 with an
increase of hindrance of orientational diffusion in the polymer [283]. Another similar
result is that n for DGEBA prepolymers decreases as the molecule weight increases

[265].

Table 4.3 Values of the parameters for the curing DGEBA/DDS system

 

Extent of Cure n so[266,269] ego (so-coo) Ea (kJ/Mol) A (s)

 

0% 0.17 9.10 3.29 5.81 110 4.2E-25
9% 0.15 8.67 3.33 5.34 110 1.7E-24
14% 0.14 8.47 3.30 5.16 112 2.1E-24
22% 0.13 8.12 3.43 4.69 107 5.2E—23
37% 0.11 7.45 3.11 4.34 119 8.6E-24
42% 0.10 7.19 2.89 4.30 133 7.1E-25
64% 0.08 6.24 2.93 3.31 129 5.4E-23
79% 0.07 5.57 2.73 2.83 121 7.2E-20

 

91

The parameter so represents the equilibrium behavior while s.no represents the

instantaneous behavior. Therefore, (so-soc) is the effective moment of the orienting

dipoles [249]. The 7 relaxation strength (so-s00) was found to decrease during the

polymerization (see Fig 4.14), which is consistent with the decreasing number of the

epoxy and amine groups of the reactants. Researchers at University of Pisa reached same

results for similar reacting systems [267, 269] while Sheppard and Senturia reported that

the relaxed dielectric constant 800 deceased as the reaction progressed and could be

linearly related to the extent of cure for DGEBA/DDS reacting system [273]. In addition,

the relaxation strength was found to diminish with increasing molecular weight of

 

 

 

 

DGEBA prepolymers [265].
6 o 0.3
l “‘.\ :
: \I‘\ ‘— (So-8m)
1.! _
4 - e- 0.2
Itxt “u :
AxAx ‘ ‘u t
2~ 3““ —~ 0.1
a ‘A. _
n—> “A .
O T I I I I I I I I I I I I I I T T I I 0.0
0% 20% 40% 60% 80% 100%
Extent ofCure

Figure 4.14 n and (so-soc) vs. extent of cure for the curing DGEBA/DDS system

The activation energy of the y relaxation ﬁrst increases, and then decreases during

cure in Figure 4.15. The activation energy is the mean value of a distribution of activation

92

energies [251] and changes with the polymerization [259]. The phenomenon of
increasing activation energy is consistent with the fact that the viscosity increases as the
polymerization progresses. However, after the extent of cure reaches around 50%-60%
the activation energy start to decrease. It may be explained by that the hindrance ability
of the existing polymer chains may be weaker than that of dipoles in the reactants and
thus less energy is needed for dipolar groups to relax after the peak point. Inasmuch as
the gel point for the DGEBA/DDS system is 58% [281], the peak around 50-60% extent

may be related to the gel point of the curing system.

 

 

140 -
A
A A
E .
a 120 e ‘
”3 A
A A
A
100 1 1 r 1 I
0% 20% 40% 60% 80% 100%
Extent ofCure

Figure 4.15 Ea vs. extent of cure for the curing DGEBA/DDS system

Figure 4.16 shows the calculated 7 relaxation time of DGEBA/DDS epoxy resin at
different temperatures and extents. The relaxation time increases as the temperature
decreases and the curing reaction goes on. The main reason for the remarkable increase

of the relaxation time during the reaction is the rise of the medium viscosity.

93

 

 

 

 

0
1.E+02 g 79/0
3 64%
l.E-Ol ; 42%
i 37%
s : 43.24
P 1.1304 g 9023
1.12.07
lcE-IO I I I I I
2.5 2.7 2.9 3.1 3.3 3.5 3.7
rooorr

Figure 4.16 I vs. lOOO/T for the curing DGEBA/DDS system

Conclusion: The dielectric properties of a crosslinking DGEBA/DDS system as a
function of temperature in a range of 20-120°C at 2.45 GHz have been investigated. The
dielectric properties of DGEBA/DDS mixture are less than those of DGEBA. The
dielectric constant and dielectric loss factor of the DGEBA/DDS system increase as the
temperature increases while they decrease during the reaction. The experimental data
ﬁtted the proposed simpliﬁed Davidson-Cole expression well. The 1 relaxation has been
identiﬁed. The Arrhenius expression is applicable to the y relaxation. The evolution of all
parameters as the reaction proceeds was related to the facts that the number of the dipolar

groups involved in the reaction decreases and medium viscosity increases.

94

4.4.2 DGEBA/Jeffamine D-230 System
Dielectric properties of DGEBA, Jeffamine D-230 (Jeffamine hereafter), and

uncured DGEBA/Jeffamine mixture at 2.45 GHz versus temperature is shown in Figure

4.17.

 

 

 

 

 

 

 

 

7
on Q Q . O
o D 9. .
6 ‘ n . ‘
_w 1“ ADAD’A.AA“
00.3.
5 j 0. o DGEBA
- ’ . Jeﬂirrnine 13-230
0 DGEBA/Jeﬂ‘arnine D-230
4 T I T r I I I I T T T I j I I I I I f
20 40 60 80 100 120
Temperature (°C)
I Duo
0.8 e .0
r u a .9 . .
j A D ’ o
r A D O
- 0.6 ~ 0" :2
- , D ‘
w _ A‘ ‘ ‘ o
_ o ‘ ‘ o
0-4 :0 o DGEBA
- A Jeﬂarnine D230
0 DGEBA/Jeffamine D-230
0.2 I I I I I T I I I I I I I I I I I T T
20 40 60 80 100 120
Temperature (°C)

Figure 4.17 Dielectric properties vs. temperature for DGEBA, Jeffamine D-230, and
uncured DGEBA/Jeffamine D-23O mixture

95

The dielectric constant of DGEBA increases with an increase in temperature from
20°C to 80°C, remains stable around 80°C to 100°C, and then decreases. The dielectric
loss factor increases ﬁrst, and decreases with a peak value around 70°C. The dielectric
constant of Jeffamine is stable while its loss factor decreases as the temperature increases.
Changes of the dielectric properties of uncured DGEBA/Jeffamine mixture are similar to
those of DGEBA over a temperature range of 20 to 80°C. The dielectric constant and loss
factor of the mixture are similar to those of DGEBA. Compared with DDS (see Table
4.1), Jeffamine has longer molecular chains. The long molecular chains, which act as
lubricant, may decrease the viscosity of DGEBA matrix. Since the mixture is easier to
relax than DGEBA, the dielectric properties of the mixture are larger than those of
DGEBA.

The extents of cure were calculated from DSC data, which are shown in Table 4.4.
Dielectric constant and dielectric loss factor of reacting DGEBA/Jeffamine epoxy resins
at a temperature range of 20-90°C are shown in Figure 4.18, where points represent the

experimental data and lines represent calculated data.

Table 4.4 DSC results of the curing DGEBA/Jeffamine system

 

Absorbed Heat (J/ g) 260 296 298 222 177 160 l 19 62
313 253 229 185 208 143 125 44
299 238 192 205 156 161 82 50

 

 

Average 291 262 240 204 180 154 109 52
Extent of Cure 0% 10% 18% 30% 38% 47% 63% 82%
Standard Error 5% 6% l 1% 4% 5% 2% 5% 2%
Physical State Liquid Liquid Liquid Liquid Gel Solid Solid Solid

 

96

 

 

 

 

 

7: 90%
63 010%
i A 18%
5: 030%
3 +38%
43 I 47%
3 : x63%
d A82%

2 rfrrmrr...,,..,,,,,,,

20 40 60 80 100 120

Temperature(°C)

1.0? .O%
0.83 D o10%
A 18%
0'6: °° 030%
0.43 // +38%
3 -47%
0'2; W x63%
‘ W A82%
0.0 urtmlwmlmwﬂwm

20 40 60 80 100 120

Temperature(°C)

DGEBA/Jeffamine system

97

Figure 4.18 Temperature dependence of dielectric properties for the curing

The dielectric properties of DGEBA/Jeffamine epoxy resins decrease as the extent
of cure increases and increase with temperature except s" at 0 and 10% extents at high
temperatures. The explanation of the phenomena is similar to that of DGEBA/DDS
system. The reaction mechanism of the DGEBA epoxy resin system is shown in Figure

4.8. The two reasons accounting for the decrease in the dielectric constant and loss factor

during the reaction are a decrease in the number of the dipolar groups in the reactants and
an increase in the viscosity.

The Davidson-Cole model was used to describe the dielectric behavior of
DGEBA/Jeffamine epoxy resins while the simpliﬁed Davidson-Cole expression is used
to calculate the parameters. Furthermore, the expression is applicable to dielectric data at
the extent of cure over 30% and low temperature parts at the extent of cure from 0 to
30%. The Davidson-Cole model is given by Equation 4.13 while the simpliﬁed
Davidson-Cole expression is shown in Equation 4.18.

Figure 4.19 shows plots of s" versus 8', which yield straight lines. The values of
s" and s' of DGEBA/Jeffamine at 0%, 10%, 18%, and 30% extents are low temperature

values in Figure 4.18, since the values at high temperatures did not yield straight lines.

0.7 —< ° 0%

 

 

 

Figure 4.19 3" vs. 8' for the curing DGEBA/Jeffamine system

98

Similar to the DGEBA/DDS system, the predominant relaxation in the
DGEBA/Jeffamine system is the y relaxation [5], which should ﬁt the Arrhenius
expression. According to Equation 4.21, plots of In (s'-s.o) versus UT and ln(s") versus
III“ of experimental data in Figure 4.19 yield straight lines, shown in Figures 4.20 and

4.21. Curves represent ﬁt of the proposed expression to the experimental data points.

 

 

 

 

1.5
o 0%
\‘ﬁr’: o 10%
A 0.5 4 A 18%
fo 0 30%
:59 + 38%
-415 r I.4796
x 63%
A 82%
~15 7— 1 T
2.6 2.8 3.0 3.2 3.4
lOOO/T

Figure 4.20 In (s'-s,o) vs. 1000/T for the curing DGEBA/Jeffamine system

 

 

 

 

0
ﬂ 0 0%
a 10%
_ -1 4 \ A 18%
E K o 30%
_2 _ + 38%
A A I 47%
\ x 63%
-3 r , T A 82%
2.6 2.8 3.0 3.2 3.4
1000/T

Figure 4.21 ln (8") vs. 1000/T for the curing DGEBA/Jeffamine system

99

The parameters 11 and s,o of the curing DGEBA/Jeffamine epoxy resins were
calculated from slopes and intercept of Figure 4.19. Ba and a relation between so and A
can be calculated based on Figures 4.20 and 4.21. The values of so at 0%, 10%, 18%,
30% extents were modiﬁed until the calculated data from the Davidson-Cole model ﬁtted
the experimental data well. so at 38%, 47%, 63%, and 82% extents were estimated based
on the modiﬁcation and the rule that (so-soc) decreases during curing.

Comparison of the experimental data with the calculated data is shown in Figure
4.18. The calculated data at 0%, 10%, 18%, and 30% extents were calculated from the
Davidson—Cole model while the others were calculated using the simpliﬁed Davidson-
Cole expression.

Figure 4.22 shows the comparison of the experimental data with the calculated
Cole-Cole arcs from the Davidson—Cole model. According to Figures 4.18 and 4.22, the
calculated values ﬁtted the experimental data well. The difference between the
experimental data and the calculated data is caused by the calculation error based on the
approximation functions and the experimental error due to ﬂuctuation of measured data
including temperature, resonant frequency, and half-power frequency bandwidth.

The simpliﬁed Davidson-Cole expression is applicable to part of the experimental
data of the DGEBA/Jeffamine system while it is applicable to all measured data for the
DGEBA/DDS system. Compared with Jeffamine, DDS has a rigid chain which hinders
the relaxation of dipolar groups. Therefore, the dielectric behavior of DGEBA/DDS
epoxy resins for low extents of cure under 2.45 GHz is similar to that of polymers, which

can be described by the simpliﬁed Davidson-Cole expression, while that of

100

DGEBA/Jeffamine is similar to that of low molecular weight chemicals, e.g. glycerol,

which can be described by the Davidson-Cole model.

0.6

‘° 0.4

0.2

0.0

1 1 I 1

 

o 0% ‘
A 10%
x 18%
o 30%
+ 38%
- 47%
x 63%
A 82%

 

Figure 4.22 Cole-Cole plots for the curing DGEBA/Jeffamine system

Table 4.5 shows the calculated values of all parameters in the models. The gel

point of the reacting DGEBA/Jeffamine system is between 30% and 40% extents. Before

the extent of cure reaches 30%, the reacting system is not viscous and the relaxation time

is small. Therefore, the simpliﬁed Davidson-Cole expression is not suitable to describe its

dielectric behavior.

As the reaction goes on, the parameter 11 decreases from 0.172 down to 0.112 (see

Figure 4.23). The parameter n is mainly connected with the motion of dipolar groups for

the secondary y relaxation. During the reaction the motion of dipolar groups is hindered

lOl

by crosslinking epoxy resins. Therefore, the parameter r1 decreases. The parameter so
,represents the equilibrium behavior while s,o represents the instantaneous behavior.
Therefore, (so-s00) is the effective moment of the orienting unit. The 7 relaxation strength

(so-s00) was found to decrease during the polymerization, as shown in Figure 4.23. This is

consistent with the decreasing number of the dipolar groups in the reactants.

Table 4.5 Values of the parameters for the curing DGEBA/Jeffamine system

 

 

Extent of n so s,o (so-s00) E, A (s)
cure (kJ/mol)
0% 0.172 7.60 2.92 4.68 69 8.6E-21
10% 0.155 7.55 2.74 4.81 80 2.6E-22
18% 0.156 7.30 2.88 4.42 99 8.9E-25

30% 0.158 7.00 2.96 4.04 105 8.7E-25
38% 0.134 6.85 2.83 4.02 131 3.0E-28
47% 0.150 6.70 2.74 3.96 123 2.0E-26
63% 0.114 6.20 2.45 3.75 135 9.8E-27
82% 0.112 5.90 2.56 3.34 110 2.2E-21

 

 

 

 

6 _ 0.3
7 H . ‘— (80-800) *
1 ----- - ....... .

4 .. I "I n. ...... '.. T 02
iL-"K'm ----- é A H“.

5 .............. _

2 '1 n—>A “A —: 01
4 I

O r I I I f T T I I V I I I I I I ' ' I 0'0
0% 20% 40% 60% 80% 100%

Extent ofCure

Figure 4.23 n and (so-soc) vs. extent of cure for the curing DGEBA/Jeffamine system

102

The activation energy of the y relaxation ﬁrst increases and then decreases during
the reaction in Figure 4.24. The increasing activation energy during cure is consistent
with the fact that the viscosity of the reacting system increases as the polymerization
progresses. However, after the extent of cure reaches around 60% the activation energy
start to decrease. It may be explained by that the hindrance ability of the existing polymer
chains may be weaker than that of dipoles in the reactants and thus less energy is needed

for dipolar groups to relax after the peak point.

 

 

150 —
‘ A
g 100 ~ 4
uﬂ .
IL
50 I I I I I
0% 20% 40% 60% 80% 100%

Extent of Cure

Figure 4.24 Ea vs. extent of cure for the curing DGEBA/Jeffamine system

Figure 4.25 shows the calculated 7 relaxation time of DGEBA/Jeffamine epoxy
resin at different temperatures and extents. The relaxation time increases as the
temperature decreases and the curing reaction goes on. The main reason for the

remarkable increase of the relaxation time during the reaction is the rise of the medium

viscosity.

103

1 .E+00

 

 

a
5 82%
- 63%
1.13-03 47%
3 38%
’5? 7 30%
7: 113-06; 18%
: 10%
1.309 a 0%
1.E-12 1 r r r I
2.5 2.7 2.9 3.1 3.3 3.5

1 000/T

Figure 4.25 1: vs. 1000/T for the curing DGEBA/Jeffamine system

Conclusion: The dielectric properties of a crosslinking DGEBA/Jeffamine D-230
epoxy resin system as a function of temperature in a range of 20-90°C at 2.45 GHz have
been investigated. The dielectric constant and dielectric loss factor of the uncured
DGEBA/Jeffamine D-230 mixture are larger than those of DGEBA. Normally, the real
and the imaginary part of the complex dielectric constant increased with temperature
while they decreased as the extent of cure increased. The Davidson-Cole model can be
used to describe the experimental data. The simpliﬁed Davidson-Cole expression is used
to calculate the parameters in the Davidson-Cole model. It is also applicable to the
experimental data of extent of cure larger than 30%. The Arrhenius-like dependent y
relaxation has been identiﬁed. The evolution of the parameters in the models, e.g. the
shape parameter n, the relaxation strength (so-sac), the activation energy Ea, and the
relaxation time 1:, was related to the facts that the dipolar groups in the reactants decrease

in number and medium viscosity increases during the polymerization.

104

4.4.3 DGEBA/mPDA System
Dielectric properties of DGEBA, mPDA, and uncured DGEBA/mPDA mixture at

2.45 GHz versus temperature are shown in Figure 4.26.

 

 

 

 

 

 

 

 

7
‘ A ‘ ° 5 o '
‘ o . o
6 - ‘ ° 0
_ 5 A a O
- o O
0)
_ .3
5 _ D
q J3 o DGEBA
q 0 A mPDA
_ D DGEBA/mPDA
4 r I I I f T I I m I I I I T I
20 40 60 80 100
Temperature (0 C)
J o DGEBA
0.9 i A mPDA
: D DGEBA/mPDA D
‘ no? 9 0
- 0.7 T r .
-w ‘ . O
_ .0
_ ‘ ‘
0.5 : ‘ ‘0 Cl ‘ ‘
_ z D A
0.3 T I ‘ I I I I I I I I I I I j
20 40 60 80 100
Temperature (0 C)

Figure 4.26 Dielectric properties vs. temperature for DGEBA, mPDA, and uncured
DGEBA/mPDA mixture

105

The dielectric constant of DGEBA increases as the temperature increases from
20°C to 80°C, remains stable around 80°C to 100°C, and then decreases. The dielectric
loss factor increases ﬁrst, and then decreases with a peak value around 70°C. The
dielectric constant of mPDA increases while its loss factor has a peak around 40°C as the
temperature increases. The dielectric properties of uncured DGEBA/mPDA mixture are
ahnost same as those of DGEBA over temperatures ranging from 20 to 80°C. Compared
with DDS (see Table 4.1), mPDA has small molecules and low amine equivalent weight.
Added mPDA did not change the dielectric properties of DGEBA too much due to its
small molecules, which may not change the viscosity of DGEBA.

Dielectric constant and dielectric loss factor of reacting DGEBA/mPDA epoxy
resins over a temperature range of 20-100°C are shown in Figure 4.27, where points
represent the experimental data and lines represent calculated data. The extents of cure

were calculated from DSC data, which are shown in Table 4.6.

Table 4.6 DSC results of the curing DGEBA/mPDA system

 

Absorbed Heat (J/ g) 467 488 319 31 1 21 l 89 72
492 370 297 215 192 93 60
477 403 302 223 198 1 12 81

 

 

 

Average 479 420 306 249 200 98 71
Extent of Cure 0% 11% 35% 47% 58% 79% 85%
Standard Error 2% 7% 1% 6% 1% 1% 1%
Physical state Liquid Liquid Liquid Solid Solid Solid Solid

 

The dielectric properties of DGEBA/mPDA epoxy resins decrease as the extent of
cure increases and increase with temperature except s" at 11% extent of cure at high

temperatures. The explanation of the phenomena is similar to that of DGEBA/DDS and

106

DGEBA/Jeffamine systems. The reaction mechanism of the DGEBA epoxy resin system
is shown in Figure 4.8. The predominant relaxation in the DGEBA epoxy resins at 2.45
GHz is the y relaxation [5]. The two reasons accounting for the decrease in the dielectric
constant and loss factor during the reaction are a decrease in the number of the dipolar

groups in the reactants and an increase in the viscosity.

 

 

 

 

 

 

 

 

 

7 _.
i o u .0%
6E A a11%
. D 435%
1.75; +47%
Z I58%
42 Mow-4’ 079%
I W x85%
344r
10 30 50 70 90 110
Temperature(°C)
1.0
. 00%
03: n n11%
0.62 2 A 135%
=w : +47%
0.4; A I58%
: g 079%
0.2; W x85%
0.0‘1.111.....41m...,.
10 30 50 70 90 110
Temperature(°C)

Figure 4.27 Temperature dependence of dielectric properties for the curing
DGEBA/mPDA system

107

The Davidson-Cole model was used to describe the dielectric behavior of
DGEBA/mPDA epoxy resins while the simpliﬁed Davidson-Cole expression can be used
to calculate the parameters. The simpliﬁed Davidson—Cole expression is also applicable
to dielectric properties of DGEBA/mPDA epoxy resins over 47% extents and low-
temperature dielectric parts of those from 0% to 47% extents. The Davidson-Cole model
is given by Equation 4.13 while the simpliﬁed Davidson-Cole expression is shown in
Equation 4.18. The dominant y relaxation in this system should ﬁt the Arrhenius
expression in Equation 4.1 1.

Figure 4.28 shows plots of 3" versus 8', which yield straight lines. The values of

s" and s' at 11, 35, and 47% extents are low-temperature data in Figure 4.27.

 

 

 

0.8 i
- ‘ 0
0.6 ~ 0/°
7 x 11%
= f . 35%
(.0
0.4 1 . 47%
~ 1: 58%
- a o
0.2 ~ 794’
‘ M + 85%
0.0 I I I I I I I I I I I I I I f I
3 4 5 6 7
8'

Figure 4.28 s" vs. 8' for the curing DGEBA/mPDA system

108

According to Equation 4.21, plots of In (853.0) versus UT and ln(s") versus UT of
experimental data in Figure 4.28 yield straight lines, shown in Figures 4.29 and 4.30.

Curves represent ﬁt of the proposed expression to the experimental data points.

 

 

 

 

1.5
1.0 - \ o 0%
’2? \ n 11%
3 0.5 ~ ‘ 35%
E + 47%
o 79%
O 5 x 85%
2.6 2.8 3.0 3.2 3.4
lOOO/l"

Figure 4.29 In (8'-800) vs. lOOO/T for the curing DGEBA/mPDA system

 

 

 

 

O
A
\\ o 0%
'1 ‘ a 11%
p ;
~51 A 35%
5 + 47%
-2 ~ M n 58%
o 79%
_3 x 85%
2 6 2.8 3.0 3.2 3.4
lOOO/T

Figure 4.30 In (8") vs. lOOO/T for the curing DGEBA/mPDA system

109

The parameters n and 8.0 of the curing DGEBA/mPDA epoxy resins were
calculated ﬁom slopes and intercept of Figure 4.28. Ea and a relation between so and A
can be calculated based on Figures 4.29 and 4.30. The values of so at 0%, 11%, 35%,
47% extents were modiﬁed until the calculated data from the Davidson-Cole model ﬁtted
the experimental data well. so at 58%, 79%, and 85% extents were estimated based on the
modiﬁcation and the rule that (so-sac) decreases during curing.

Comparison of the experimental data with the calculated data is shown in Figure
4.27. The calculated data of 0% to 47% extents were calculated from the Davidson-Cole
model while the others were calculated using the simpliﬁed Davidson-Cole expression.

Figure 4.31 shows the comparison of the experimental data with the calculated Cole-Cole

plots from the Davidson-Cole model.

 

 

 

 

1.0
l
0.8 1 O
‘ ." O i 9 0%
- /" 0
0.6 ~ .r - 11/..
= j ,./‘«\ . 35%
w ./ \
I / ,/ \ + 47°/
0.4 “ {i I’L’l’\\\ ‘\ 0
“ ,”/lp’,’ \ I ' 58%
-I /}’/l /_‘\ “ I|
0 2 d I’ll /// \\ “ ‘I o 79%
° . X I 2 x85%
. __A.’“" I i i
_ , / , i I
0.0 T I I / I l I I I i 1I4 I I i I
2 4 6 8
8'

Figure 4.31 Cole-Cole plots for the curing DGEBA/mPDA system

110

Table 4.7 shows the calculated values of the parameters of the Davidson-Cole
model and the Arrhenius expression. As the reaction goes on the parameter n decreases
from 0.153 down to 0.073 (see Figure 4.32). The parameter n is mainly connected with
the motion of dipolar groups for the secondary y relaxation. During the reaction the
motion of dipolar groups is hindered by curing epoxy resins. Therefore, the parameter 11
decreases. so represents the equilibrium behavior while so, represents the instantaneous
behavior. Therefore, (so-sw) is the effective moment of the orienting unit. The y
relaxation strength (so-s00) was found to decrease during the polymerization in Figure
4.32. This is consistent with the decreasing number of the dipolar groups in the reactants.

The activation energy of the y relaxation ﬁrst increases and then decreases during
the reaction, shown in Figure 4.33. The increasing activation energy during cure is
consistent with increasing viscosity of the reacting system during curing. However, after
the extent of cure reaches around 40%, gel point shown in Table 4.6, the activation
energy start to decrease. It may be explained by that the hindrance ability of the existing
polymer chains may be weaker than that of dipoles in the reactants and thus less energy is

needed for dipolar groups to relax after the peak point.

Table 4.7 Values of the parameters for the curing DGEBA/mPDA system

 

Extent of cure n so sq, (so-s00) Ea (kJ/mol) A (s)

 

0% 0.153 8.20 3.07 5.13 93 4.6E-24
11% 0.152 7.50 2.94 4.56 98 7.1E-25
35% 0.153 7.10 2.93 4.17 101 6.2E-25
47% 0.137 7.00 2.85 4.15 113 6.3E-26
58% 0.139 6.70 3.18 3.52 109 8.4E-25
79% 0.108 6.32 2.92 3.40 53 4.3E-14

85% 0.073 6.00 2.73 3.27 79 9.2E-16

111

 

 

 

 

4’ I-
." ’
Fﬁ“.~“‘ (80-800) '-
“‘~~‘ . >-

44 “x-“ 4 0.2
1 ' ““4« i
‘ _
r ‘ F

2 4 n—’ \ 4 0.1
d ‘ ‘-

0 l l T I T l T I T T l I l f T I T l l 0.0

0% 20% 40% 60% 80% 100%

Extent of Cure

Figure 4.32 n and (so-s00) vs. extent of cure for the curing DGEBA/mPDA system

 

 

120 -
A 100 _ 5 ‘
'5 A
E
a 80 ‘ A
I13
60 4
40 I I I I I
0% 20% 40% 60% 80% 100%

Extent of Cure

Figure 4.33 Ea vs. extent of cure for the curing DGEBA/mPDA system

Figure 4.34 shows the calculated 7 relaxation time of DGEBA/mPDA epoxy resin
at different temperatures and extents. The relaxation time increases as the temperature
decreases and the curing reaction goes on. The main reason for the remarkable increase

of the relaxation time during the reaction is the rise of the medium viscosity.

112

 

1.E+00 :
i 8596
1.503 / 79%
58%
6‘ 4 / 47%
:j lJEk06 g 3596
E 11%
i 0%
1.E-O9 :4
1.E‘12 j T l ﬂ T

 

 

 

2.5 2.7 2.9 3.1 3.3 3.5 3.7
1000/T

Figure 4.34 I vs. 1000/T for the curing DGEBA/mPDA system

Conclusion: The dielectric properties of a crosslinking DGEBA/mPDA epoxy
resin system as a function of temperature in a range of 20-100°C at 2.45 GHz have been
investigated. The dielectric constant and dielectric loss factor of the uncured
DGEBA/mPDA mixture are similar to those of DGEBA. Normally, the real and the
imaginary part of the complex dielectric constant increased with temperature while they
decreased as the extent of cure increased. The Davidson-Cole model can be used to
describe the experimental data. The simpliﬁed Davidson-Cole expression is used to
calculate the parameters and is also applicable to the experimental data of extent of cure
larger than 47%. The Arrhenius-like dependent y relaxation has been identiﬁed. The
evolution of the parameters in the models, e. g. the shape parameter n, the relaxation
strength (so-soc), the activation energy E, and the relaxation time I, can be related to facts
that the dipolar groups in the reactants decrease in number and medium viscosity

increases during the polymerization.

113

4.4.4 DGEBA/Epikure W System

Dielectric properties of DGEBA, Epikure W (W hereafter), and uncured

DGEBA/W mixture in a temperatures range of 20 to 120°C are shown in Figure 4.35.

 

 

 

 

 

 

 

 

7.
Z .000 .0 9
6E .0 '0..ch Dunn
D
51.3000
-0045
33 AA A A A“ “ “
22
3 ODGEBAAWDDGEBA/w
1 rTlITTIIIITIIYIIlTlIT
20 40 60 80 100 120
Tenperature(°C)
0.8
ﬂ O..U°.t'.°n
a O
: 9 D O
0.6 ‘30
: n O .
- i 9 o
-0304“.
: ‘ A AA AA AA
- “ A
0.24
~ ODGEBA AW UDGEBA/W
0.0 IIIIIIIIIﬁIrTIIIrITTI
20 40 60 80 100 120
Temperature(°C)

Figure 4.35 Dielectric properties vs. temperature for DGEBA, W, and uncured

DGEBA/W mixture

The dielectric constant of DGEBA increases as the temperature increases from

20°C to 80°C, remains stable around 80°C to 100°C, and then decreases. The dielectric

114

loss factor increases ﬁrst and then decreases with a peak value around 70°C. The
dielectric constant and loss factor of W increases as the temperature increases. The
dielectric properties of uncured DGEBA/W mixture are similar to those of DGEBA.

The dielectric constant and dielectric loss factor of reacting DGEBA/W epoxy

resins over a temperature range of 20-130°C are shown in Figure 4.36. The extents of

cure were calculated from DSC data, which are shown in Table 4.8.

 

 

 

 

 

 

 

 

_ ..A AA
_ o .4.1; x x 80%
o A x
~ °AA xxx ' A8%
5 ‘ 03 x I ' .
_ or“, x .- + x21%
- - .Axx +++ o
w J 4 ..I' ++ u D 340A)
4 _ I +‘*‘ a D a ‘ +50%
A
++++u 0 1A “ “ 066%
~ CA 2‘ 3‘ .840/0
3 IIII IIIII Ij I I l I I III I I II
20 40 60 80 100 120 140
Terrperature(°C)
0.8
: ’0’...
06: . AAA 00%
. _ ...A AA A AA A8%
a A
=..04: * I ”1%
' — 40°
4 A xxx: .1 l:+:+ ' f)
02: XXX-I's’+ at: noon +50/o
- A84%
0.0 I I 7 I I I I I I l I I I TI I I I I I l FT
20 40 60 80 100 120 140
Terrperature(°C)

Figure 4.36 Temperature dependence of dielectric properties for the curing DGEBA/W
system

115

Table 4.8 DSC results of the curing DGEBA/W system

 

Absorbed Heat (J / g) 309 257 245 157 167 75 35
261 250 217 179 120 138 49
282 276 213 175 142 80 52

 

 

Average 284 261 225 170 143 97 45
Extent of Cure 0% 8% 21% 40% 50% 66% 84%
Standard Error 5% 3% 4% 2% 5% 7% 2%
Physical state Liquid Liquid Gel Solid Solid Solid Solid

 

The dielectric constant of the DGEBA/W epoxy resins increases as the
temperature increases and decreases as the extent of cure increases. The dielectric loss
factor of the DGEBA/W system ﬁrst increases and then decreases as the temperature
increases, except at 66% and 84% extents. The dielectric loss factor at 66% and 84%
extents increases with temperature. Generally, the changes of the dielectric properties

with (art) are illustrated in Figure 4.1 and 4.2. An increase in temperature leads to a

decrease in the relaxation time and (our). The changing pattern of the dielectric properties
of DGEBA/W system is same as shown in Figure 4.1 and 4.2. It is known from Figure
4.36 that one relaxation exists in the DGEBA/W system. The reaction mechanism of the
DGEBA epoxy resin system is shown in Figure 4.8. The predominant relaxation in the
DGEBA epoxy resins at 2.45 GHz is the y relaxation.

The Davidson-Cole model was used to describe the dielectric behavior of
DGEBA/W epoxy resins while the simpliﬁed Davidson-Cole expression can be used to
calculate the parameters. The Davidson-Cole model is given by Equation 4.13 while the
simpliﬁed Davidson-Cole expression is shown in Equation 4.18. The calculation

procedure is same as that for DGEBA/mPDA system in chapter 4.4.3.

116

Figure 4.37 shows the comparison of the experimental data with the calculated
data from the Davidson-Cole model. The calculated values ﬁtted the experimental data
well for data at low temperatures and high extents of cure. The difference between the
experimental data and the calculated data is, normally, caused by the calculation error
based on the approximation functions and the experimental error due to ﬂuctuation of
measured data including temperature, resonant frequency, and half-power frequency
bandwidth. The difference at low extents and high temperatures in Figure 4.37 is mainly
caused by rapid reaction at low extents and high temperatures under microwave
irradiation and the extent of cure changed during measuring the dielectric properties.
Therefore, the dielectric properties deviated from the models, e. g. the Cole-Cole model,
the Davidson-Cole model, the H—N model, and the simpliﬁed Davidson-Cole expression.
The ﬁtting results in Figure 4.37 are best from the models.

Table 4.9 shows the calculated parameters of the Davidson-Cole model and the
Arrhenius expression. The gel point of DGEBA/mPDA epoxy resin is between 21% and
40% extents, shown in Table 4.8. Before the extent of cure reaches 21%, the reacting

system is low in viscosity and the relaxation time is small.

Table 4.9 Values of the parameters for the curing DGEBA/W system

 

Extent of cure n so s,o (so-s00) Ea (kJ/mol) A (s)

 

0% 0.161 7.10 2.70 4.40 1.57 64
8% 0.145 7.00 2.70 4.30 1.57 79
21% 0.123 6.40 3.04 3.36 1.57 94
40% 0.121 6.00 3.24 2.76 1.57 101
50% 0.123 5.30 2.92 2.38 1.57 110
66% 0.092 5.20 2.60 2.60 1.57 107

84% 0.084 4.90 2.56 2.34 1.57 94

117

o 0%

A 8%

x 21%
I 40%
+ 50%
o 66%
A 84%

111444

 

 

 

3 I—IIIri—rIfIIIIIIIIIIIIII]

20 40 60 80 100 120 140

 

 

Terrperature(°C)
0.8—
f. 00%
0.6: AAAA A8%
i /“I x21%
190.43 ﬂ u40%
j +50%
0.2} $3: 066%
2 A84%
0.0 nmmurmnnlurnn,
20 40 60 80 100 120 140
Tenperature(°C)

Figure 4.37 Comparison between the experimental and calculated data of the curing
DGEBA/W system.

As the reaction goes on the parameter n decreases ﬁ'om 0.161 down to 0.084 (see
Figure 4.38). The parameter n is mainly connected with the motion of dipolar groups for
the secondary y relaxation. During the reaction the motion of dipolar groups is hindered
by curing epoxy resins. Therefore, the parameter n decreases. The parameter so represents
the equilibrium behavior while so, represents the instantaneous behavior. Therefore, (so-

ego) is the effective moment of the orienting unit. The 7 relaxation strength (so-s00) was

118

found to decrease during the polymerization, as shown in Figure 4.38. This is consistent

with the decreasing number of the dipolar groups in the reactants.

 

 

 

 

5i 0.4
N I :
4 - «— (So-Soc) ~
‘~.‘ 44 0.3
- I x“ C
3 _ “~-\ I
. ‘~.‘ .—
4~,I
_, . \4\\ . '7: 0.2
.1 \‘4 b
2 i“‘A-~--‘_ r
4 A ‘‘‘‘‘ A~—-é__ r
1_ ‘“A~~_-NA ‘: 0.1
n—> ~
0 I I I I I T I I I I I I WI I I I I I 0.0
0% 20% 40% 60% 80% l 00%
ExtentofCure

Figure 4.38 n and (so-soc) vs. extent of cure for the curing DGEBA/W system

The activation energy of the y relaxation of the DGEBA/W system ﬁrst increases
and then decreases during the reaction, shown in Figure 4.39. The increasing activation
energy during cure is consistent with the fact that the viscosity of the reacting system
increases as the polymerization progresses. However, after the extent of cure reaches
around 40%, gel point shown in Table 4.9, the activation energy start to decrease. It may
be explained by that the hindrance ability of the existing polymer chains may be weaker
than that of dipoles in the reactants and thus less energy is needed for dipolar groups to

relax after the peak point.

119

120 4

 

 

A
A
A 100 - A ‘
'-‘ A
E
a 80 4 A
US
60 —“
40 I I I I ﬂ
0% 20% 40% 60% 80% 100%
Extent of Cure

Figure 4.39 Ea vs. extent of cure for the curing DGEBA/W system

Figure 4.40 shows the calculated 7 relaxation time of the DGEBA/W epoxy resins
at different temperatures. The relaxation time increases as the temperature decreases or
the curing reaction goes on. The main reason for the remarkable increase of the relaxation

time during the reaction is the rise of the medium viscosity.

1.E-02 3 84%
1 66%
1.E-04 jg 50%
2 40%
A 1.E—06 E“ 21%
3; E 8%
1.E—08 g 0%
1.15.10 '9
1.1312 I T I I I .

 

 

2.5 2.7 2.9 3.1 3.3 3.5 3.7
1000/T

Figure 4.40 I vs. 1000/T for the curing DGEBA/W system

120

Conclusion: The dielectric properties of a curing DGEBA/W epoxy resin system
as a function of temperature in a range of 20-130°C at 2.45 GHz have been investigated.
The dielectric constant and dielectric loss factor of the uncured DGEBA/W mixture are
similar to those of DGEBA. The Davidson-Cole model can be used to describe the
experimental data. The simpliﬁed Davidson-Cole expression was used to calculate the
parameters and is also applicable to the experimental data at low temperatures or high
extents of cure. The Arrhenius-like dependent y relaxation has been identiﬁed. The
evolution of the parameters in the models, e. g. the shape parameter n, the relaxation
strength (so-soc), the activation energy Ba, and the relaxation time r, can be related to facts
that the dipolar groups in the reactants decrease in number and medium viscosity

increases during the polymerization.

4.4.5 Parameters in the Models for the Four Systems

The dielectric properties of the reacting systems of DGEBA epoxy resin and the
four curing agents have been investigated at 2.45 GHz over a temperature range. The 7
relaxation was identiﬁed in the four systems. The Davidson-Cole model in Equation 4.13
can describe the dielectric behaviors while the Arrhenius expression in Equation 4.11 was
used to describe the y relaxation time. Although the proposed simpliﬁed Davidson-Cole
model in Equation 4.18 can only represent the dielectric properties of DGEBA/DDS
system and part of those of the other three systems, it can be used to calculate the

parameters in the Davidson-Cole model.

r=Axe[%T—] (4.11)

121

_ (80 —800)

 

 

 

s * *500 —
(1+ I'M)"
s' = 800 + (so — sq, )cos(: (9)
(1 + (my)?
(4.13)
8' = (80 _ goo )Sin(n 0)
(1 + (web?
6 z are tan(a)r)
8* _ 8m + (80 —800)
UM)"
(so — soo )cos(n 15)
s' = 8,, + 2 (4.18)

(607)"

(80 4' 800 )Sin(n %)

 

(601)"

(so-soc) is the effective moment of the orienting dipoles [249]. Figure 4.41 shows

the y relaxation strength as a function of extent of cure for the four reacting systems. (so-

soo) was found to decrease during the polymerization of the four systems, which is

consistent with the decreasing number of the epoxy and amine groups in the reactants.

Researchers at University of Pisa reached same results for similar reacting systems [267,

269] while Sheppard and Senturia reported that the relaxed dielectric constant s,o

deceased as the reaction progressed and could be linearly related to the extent of cure for

122

DGEBA/DDS reacting system [273]. In addition, the relaxation strength was found to

diminish with increasing molecular weight of DGEBA prepolymers [265].

 

 

 

 

6
A DDS
5 A A I Jeﬂimrine D-230
I A x mPDA
I» .x - ,9 A . Epikure W
A 4 ~ I I g
go "I I
c': j 9 X A x'x
$3 3 ~
A
_ Q .
Z 0 o
2 :
1 d I I I I I I If r I I I I I I I T I I I
0% 20% 40% 60% 80% 100%
Extent of Cure

Figure 4.41 (so-s00) vs. extent of cure for the four curing systems

Figure 4.42 shows the shape parameter'n as a function of extent of cure for the
four systems. As the reaction goes on, n decreases from 0.18 down to 0.06. The four
systems have same trend. The parameter describes the skewness of the dispersion of the
relaxation times, which increases as n ranges from unity to zero. The value of n for an
ideal Debye material is unity while that for polymeric solutions is around 0.5 [285]. The
relaxation time is mainly connected with the mobility of the dipolar groups. During the
reaction, the motion of dipolar groups is hindered by the medium with increasing
viscosity. Therefore, the parameter 11 decreases. The rationale is similar to the argument

that n is connected with the local chain dynamics of a polymer and deceases in the range

123

0.5-0 with an increase of hindrance of orientational diffusion in the polymer [283].
Another similar result is that the parameter n for DGEBA prepolymers decreases as the

molecule weight increases [265].

 

 

 

 

0.20 _
0.15 i I ' ' x I
I A A I X x
: . A. . I X.
‘1 0.10: A O
a A O
: sons 1. X
0.05 ‘I I Jeffamine D-230
I meDA
j 0 Epikure W
0.00 IIIIIIITIIIrTIIIIII
0% 20% 40% 60% 80% 100%
Extent ofCure

Figure 4.42 n vs. extent of cure for the four curing systems

The activation energy of the y relaxation for all four systems ﬁrst increases, and
then decreases during cure (see Figure 4.43). The activation energy is the mean value of a
distribution of activation energies [251] and changes with the polymerization [259]. The
phenomenon of increasing activation energy is consistent with the fact that the viscosity
increases as the polymerization progresses. However, after the extent of cure reaches
around 50%, the activation energy start to decrease. It may be explained by that the
hindrance ability of the existing polymer chains may be weaker than that of dipoles in the

reactants and thus less energy is needed for dipolar groups to relax after the peak point.

124

Inasmuch as the gel point for the DGEBA/DDS system is 58% [281] and, in this study,

the changing temperatures of the other three systems from liquid to solid are around 40-

50%, the peak around 50% extent may be related to the gel point of the curing system.

Ea (kJ/mol)

 

 

 

 

160
. I A .A
120 - A ' A
11 A A . x. x , '
x o
2: x . ’
I
80 II A DDS x
I Jeffamine D-230
x mPDA x
o Epikure W
40 I I I T I I I
0% 20% 40% 60% 80% 100%
Extent of Cure

Figure 4.43 Ea vs. extent of cure for the four curing systems

The relaxation time increases as the reaction proceeds. For instance, the

calculated 7 relaxation time at 80°C has been reported as a function of the extent of cure

in Figure 4.44. This is consistent with Kauzman’s study on relaxation on polymers [251].

The main reason for the remarkable increase of the relaxation time during the reaction is

the rise of the medium viscosity. Furthermore, it is shown in Figure 4.44 that the

relaxation time of DGEBA/DDS system is about two decades larger than that of

DGEBA/Jeffamine, DGEBA/mPDA, and DGEBA/W systems.

125

 

 

 

1.E+00 :
5 A DDS A
; I Jeﬂilmine D—230
I x mPDA A
1.13-04 = ° Rpm W x
E A I
g A X o
P - A A .9
A I
1.308 . - Q X. x
0‘ - x
1.E‘12 1 I I I I I I I I I

 

0% 20% 40% 60% 80% 100%
Extent of Cure

Figure 4.44 I vs. extent of cure for the four curing systems at 80°C

Figure 4.45 shows the calculated dielectric constant as a ﬁrnction of extent of cure
for the four isothermal curing systems. The dielectric constant exhibits a linear relation to
the extent of cure and may be used to in-situ monitor the polymerization. Other

researchers reached similar result [276].

 

  
 
     

 

 

 

10
A DDS
. I Jeﬂirmine D-230
8 ‘ x mPDA
o Epikure W
’0.) 6 ‘
4 Z
2 If r I I I I I I I T T T T I I f I I T
0% 20% 40% 60% 80% 100%

Extent of Cure

Figure 4.45 Calculated s' and 8" vs. extent of cure for the four curing systems

126

 

4.5 Conclusions

A single-frequency microwave heating and diagnostic system, developed at
Michigan State University, was used to heat and measure the dielectric constant and
dielectric loss factor of curing epoxy resins. A cylindrical TM 012 mode cavity is used to
process epoxy resins at 2.45 GHz. The epoxy resin was diglycidyl ether of bisphenol A
(DGEBA). The four curing agents were 3, 3-diaminodiphenyl sulfone (DDS), a
diﬁmctional primary amine (J effamine D-230), m-phenylenediamine (mPDA), and
diethyltoluenediamine (Epikure W). The mixtures of DGEBA and the four curing agents
were stoichiometric. The four reacting systems were heated under microwave irradiation
to certain cure temperatures, i.e. 145°C for DGEBA/DDS, 90°C for DGEBA/Jeffamine
D-230, 110°C for DGEBA/mPDA, and 160°C for DGEBA/Epikure W. Measurements of
temperature and dielectric properties using the swept frequency method were made
during free convective cooling of the samples. The cooled samples were analyzed with 3
Differential Scanning Calorimeter to determine the extents of cure.

The major conclusions of this chapter are:

0 The Davidson-Cole model can be used to describe the dielectric properties
of the four curing systems.

0 The simpliﬁed Davidson-Cole expression is proposed to describe the
dielectric properties of polymeric materials. It works well for the DGEBA/DDS system
and part of the data of the other three systems. The model is more speciﬁc than the
Schonhals-Schlosser model. It can be used to calculate the parameters of the Davidson-

Cole model as well. The simpliﬁed Davidson-Cole expression is as follows:

127

8* = £00 + (£0 ‘50:)
(1'60?)

(80 — £00 )cos(n %)

n

I

a =£00+

 

(M)

(£0 — goo )Sin(n :25)

 

(607)"
where 3* is the complex dielectric constant, 8' is the dielectric constant, 8" is the
dielectric loss factor, j is the imaginary unit, 0) (=21Ef, f is the oscillator frequency in Hz)
is the radial frequency of the electric ﬁeld in s", t is the relaxation time in s, so is the low
frequency dielectric constant, a... is the high ﬁequency dielectric constant, and n is the
shape parameter with a range of 0 S n3 1.

0 The secondary 'y relaxation has been identiﬁed in the four reacting systems.
The Arrhenius expression is applicable to the y relaxation.

- The evolution of all parameters in the models, e. g. the shape parameter n,
the relaxation strength (150-800), the activation energy 13,, and the relaxation time 1:, during
cure was related to the facts that the number of the dipolar groups in the reactants
decreases and medium viscosity increases during the polymerization.

o The relaxation strength (so-aw) was found to decrease during the
polymerization of the four systems, which is consistent with the decreasing number of the
epoxy and amine groups in the reactants.

o The shape parameter n decreases as the reaction proceeds for four systems.

11 is related to the local chain dynamics. During the reaction the motion of dipolar groups

128

is hindered by the medium with increasing viscosity. Therefore, the parameter n
decreases.

o The activation energy of the four systems increases ﬁrst, and then decrease
around 50% extent of cure. The increase of activation energy before the peak is
consistent with increasing viscosity of the systems during cure. The decrease after the
peak point may be explained by that the hindrance ability of the reacted polymer chains
may be weaker than that of dipoles in the reactants and, thus, less energy is needed for
dipolar groups to relax. The peak points are related to the gel points of the four systems.

0 The relaxation time increases as the reaction proceeds due to increasing
medium viscosity for the four systems.

0 The dielectric constants of the four reacting systems exhibit a linear

relation to the extent of cure and may be used to in-situ monitor the curing reaction.

129

l

 

CHAPTER 5 SYNTHESIS OF CARBON NANOTUBES BY MPCVD

5.1 Introduction
Carbon is found in many different compounds in the world while carbon alone
forms graphite, diamond, fullerence and nanotubes. The illustrations of four forms in

which the element carbon exists are shown in Figure 5.1 [286].

   
    
      
       

 
 
  

.‘f‘T.
IQ! .
U
i it.)
“:2.

  

'.. WM
0,03?
! U ,
O...)
"5".
9.9

J‘ na‘tw '
‘.
M!’ ‘r
' 4.11., O
C 10.
‘
\
79

 

graphite (10, 10) tube

Figure 5.1 Schematic illustrations of four carbon forms

In diamond, the carbon atoms are connected to each other in all three dimensions,
making it a very hard material. Graphite consists of layers of graphene sheets, layers of
hexagonally patterned carbon atoms, which form a two-dimensional structure. The two-
dimensionality of graphite makes it a soﬁer material. Fullerenes consist of a caged
structure similar to the shape of a soccer ball. Fullerenes were discovered in 1985 by

Robert F. Curl Jr., Sir H. W. Kroto, and Richard E. Smalley, who were awarded the 1996

130

Nobel Prize in Chemistry. Fullerenes can be found in nature, whereas nanotubes are only
man—made.

Carbon nanotubes (CNTs) were ﬁrst discovered in 1991 by the Japanese electron
microscopist Sumio Iijima of NEC Corporation, who was studying the material deposited
on the cathode during the arc-evaporation synthesis of fullerenes [287]. A carbon
nanotube is a tube-shaped material with a diameter measuring on the nanometer scale,
which is made of carbon. CNTs are large macromolecules that are unique for their size,
shape, and remarkable physical properties [288]. They are formed from hexagonal arrays
of carbon atoms and can be thought of as a sheet of graphite rolled into a cylinder shown

in Figure 5.2 [289].

 

Graphene sheet SWNT

Figure 5.2 Schematic illustrations of relation between graphite and CNTs

There are two main types of CNTs that can have structural perfections, which are
shown in Figure 5.3 [289]. Multi-wall nanotubes (MWNTs) comprise an array of such
nanotubes that are concentrically nested like rings of a tree trunk. The CNTs that Sumio
Iijima found in 1991 were MWNTs. Single-wall nanotubes (SWNTs), found in 1993
[290, 291] consist of a single graphite sheet seamlessly wrapped into a cylindrical tube.
SWNTs can be deﬁned by their diameter, length, and chirality. The SWNTs have a

tubular form with a diameter as small as 0.4 nm [292] and a length of a few nanometers

131

to micrometers. The SWNTs with three different chiralities are shown in Figure 5.4

[289].

(a) Cb)

 

Figure 5.4 Schematic illustrations of three SWNTs of different chiralities: (a) armchair,

(b) zigzag, (c) chiral.

132

CNTs possess many unique and remarkable chemical, physical, and electronic
properties, which make them desirable for many applications.

SWNTs are incredibly stiff and tough mechanically. They may have a high
Young’s modulus (up to 1 TPa) [293] and high tensile strength (around 30 GPa) at an
elongation of almost 6% [294, 295]. The density-normalized modulus and strength of
SWNTs are around 20 and 50 times that of steel wire, respectively. Therefore, SWNTs
are considered to be excellent reinforcement material for composites. The addition of a
small weight percent of SWNT can result in signiﬁcant improvement in mechanical
properties of the composites. This demonstrates the potential of using SWNT for ﬁbers at
the microscale level.

Nanotubes conduct heat as well as diamond at room temperature. They are very
sharp, and thus can be used as probe tips for scanning-probe microscopes, and ﬁeld-
emission electron sources for lamps and displays.

Although SWNTs are similar to a single sheet of graphite in structure, which is
semiconductor with zero band gap, they may be either metallic or semiconducting,
depending on the structures of SWNTs in Figure 5.4. The electronic properties of
MWNTs are similar to those of SWNTs due to the weak coupling between the cylinders
in MWNTs. Since both metals and semiconductors can be made from the same all-carbon
system, CNTs are ideal candidates for molecular electronics technologies.

The ﬁrst realized major commercial application of MWNTs is their use as
electrically conducting components in polymer composites [296]. Other applications of
CNTs include electrochemical devices, hydrogen storage, ﬁeld emission devices,

nanometer-sized electronic devices, sensors, probes [297].

133

The objective of this chapter is to synthesize carbon nanotubes on silicon
substrates by microwave plasma chemical vapor deposition (MPCVD) of a mixture of
methane and hydrogen. The catalyst was nickel, migrated from a small piece of catalyst
supplier to the substrate surface during microwave plasma pretreatment. Scanning
Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) were used to

characterize the morphologies of the CNTs.

5.2 Literature review

CNTs are usually made by arc discharge, laser ablation, and chemical vapor
deposition (CVD) methods. The arc discharge method produces high quality SWNTs
with few structural defects and does not require a catalyst for synthesis of MWNT. But
the purity of the nanotubes is usually low. For instance, Shi et al. synthesized SWNTs by
do. are discharge method [298]. A Y-Ni alloy composite graphite rod was used as anode
for do are discharge. A cloth-like soot, containing about 40% SWNTs, was produced.
The diameter of SWNTs is 1.3nm. Saito observed a bamboo-shaped carbon tube
produced by the arc evaporation of nickel-loaded graphite [299]. The tube, shown in
Figure 5.5, consists of a linear chain of hollow compartments that are spaced at nearly

equal separation ﬁ'om 50-100 pm. The outer diameter of the bamboo tubes is about 40

nm, and the length is typically several urn. A growth model of the bamboo tubes was

proposed.

 

Figure 5.5 TEM picture of a bamboo-like carbon tube

134

The laser ablation method produces SWNTs with high quality and high purity, but
the process is very costly. For example, the yield of SWNTs is about 80% in the
literatures [300, 301]. The diameter of SWNTs is about 1.4 nm while the length is in the
order of 1-10 micrometer.

The arc discharge and laser ablation methods involve high temperatures, e.g.
5,000—20,000°C for arc discharge, and 4,000-5,000°C for laser ablation. However, the
CVD method is used for rapid synthesis of MWNTs with high purity at lower
temperatures and is easy to scale up for commercial production. The nanotube alignment
is easy to control with this method. However, the nanotubes synthesized with CVD
usually have more structural defects compared with the other two methods. MWNTs with
diameters of 10-100 nm were produced by the ﬂoating catalyst method [302]. Aligned
nitrogenerated amorphous carbon nano-rods, with a diameter of approximately 100-250
nm and a length of approximately 50-80 pm, were synthesized on a porous alumina
template, using an electron cyclontron resonance CVD system and a microwave-excited
plasma of QR; and N2 as precursors [303]. Mauron et al. synthesized oriented nanotube
ﬁlms (20-35 pm thick) on ﬂat silicon substrates by CVD of a gas mixture of acetylene
and nitrogen. The diameter of the nanotubes is 20-25 nm. An iron nitrate ethanol solution
was coated onto a silicon substrate before heating [304]. Long SWNTs, with lengths of
10 and 20 cm, were synthesized by an optimized catalytic CVD technique with a ﬂoating
catalyst method [3 05].

Microwave plasma chemical vapor deposition (MPCVD) has gained increasing
popularity as a controllable and deterministic method for growing vertically aligned

CNTs while conventional thermal CVD has been successfully employed for self-oriented

135

growth of CNTs. A summary of literatures about synthesis of carbon nanotubes by
MPCVD [306-330] is listed in Table 5.1. The well-aligned CNTs or carbon nanowires
[306-308] synthesized by MPCVD are MWNTs. Silicon wafer are widely used as
substrates while metals, e. g. Ni, Fe, are used as catalyst. Microwave plasma gas includes
CH4, H2, C2H2, NH3, N2, and C02. The key factors of synthesis of CNTs by MPCVD are
gas composition, CNT growth temperature, pressure, growth time. The microwave power
can affect the grth temperature. The grth pressure is usually far less than
atmospheric pressure (736 Torr), approximately 1-80 Torr. However, one literature
synthesized CNTs at atmospheric pressure [309]. Most CNT growth temperature ranges
from around 600 - 800°C, except two references synthesizing CNTs at 300°C [310, 311].
The typical growth time is about 10-30 minutes. The CNTs synthesized by MPCVD are
usually 1-100 urn long and 20-100 nm in diameter. The bamboo-like structure was found

within MWNTs by many researchers.

Table 5.1 Literature summary on synthesis of CNTs by MPCVD

 

 

 

 

Ref. Topic Catalyst on Gas composition; MW Power; Temp. ; CNT Length;
substrate Total ﬂow rate (sccm) Pressure; Time Diameter

306 Carbon nanowires Ni/ Si N/A;
having the sea urchin 46nm
structure

307 Synthesis of Ni/ Si CH4:H2=2.5:57.5 sccm 600 W; 80 Torr; N/A;
interconnecting 900°C; 5 min 10-100 nm
MWNT island

308 Synthesis of high- Ni, Fe / Ti / CH4:H2:N2=6:90210 N/A; N/A; 1-50 pm;
density MWNT coils Si sccm 650°C; 15 min 20-400 nm

 

309 Producing MWNTs at Iron carbonyl Ar:CO:Fe(CO)5= 500- <1 kW; Atmospheric 5 mm;
atmospheric pressure gas / Al tube 600:900:10-30 sccm Pres; >3000 K; 4 h 50 um

 

 

310 Low temp. growth of Fe / Si CH4:CO2 = 30:30 sccm 300 W; 15 Torr; N/A;
vertically-aligned <33O °C; 20 min 15-20nm
MWNTs

311 A high yield of Fe, Ti / Si CH4:CO2=30:30 sccm 300 W; 15 Torr; N/A;
aligned MWNTs <330 °C; 20 min 15-300 nm

 

136

Table 5.] Literature summary on synthesis of CNTs by MPCVD (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ref. Topic Catalyst on Gas composition; MW Power; Temp. ; CNT Length;
substrate Total ﬂow rate (sccm) Pressure; Time Diameter
312 Well-aligned CNTs Co (~2nm)/ C2H2zNH3=10-30: 100; 1 kW; 20 Torr; 12 pm;
perpendicular to Si Si Total 200 sccm 825 °C; 2 min 30 nm
substrate
313 Simple and Fe, Ni, Co/ CIL:H2:N2=20:80:80 1.5 kW; 6.6 kPa; 5 11m;
straightforward Si sccm 680 °C; 20 min 10—20 nm
synthesis of MWNTS
314 Growth of MWNTS Ni / TiN / Si CH4/(CH4+H2)=10-20% 400 W; 10 Torr; N/A;
at low temp. 520-700°C; 10-50 10-15 nrn
min
315 Synthesis of verticallyNi / Si CH4/(CH4+H2)=20% 400 W; 10 Torr; 30 um;
aligned MWNTS 700°C; 5 min 10-35 nm
316 Well aligned MWNTsNi / Si CH4zNH3=150:150, 2.2 kW; 21 Torr; 100 um;
with high aspect ratio 200: 100; 240:60 sccm 800°C; 40 min 20-50 nm
317 Plasma breaking of Fe / Si CH4:(CH4+N2)=2-3O%; 700-900 W; 15 Torr; 15 um;
catalyst ﬁlms for 100 sccm 850°C; 15 min 20-120 nm
MWNT growth
318 Plasma breaking of Fe / Si 200 um;
catalyst ﬁlms for 5-30 nm
MWNT growth
319 The effect of catalysis Ni/Cu alloy / CH4zH2=O.5: 100 sccm 1.1 kW; 30 Torr; N/A;
on MWNT growth Si 600°C; 50-200 nm
320 Role of N2 in CNT Ni, Co / Si CH4:H2 or N2 = 10:100 960 W; 16 Torr; 10-20 pm;
growth sccm 650°C; 10 min 50 nm
321 In-situ growth of Ni / Al/ Si CIL:H2 = 0.5:100 sccm 1.1 kW; 4 kPa; N/A;
Spiral shape CNTs 600°C; 30 min 100 nm
322 Bias-enhanced growtth / Si CH4:H2 = 04:80 sccm 1.1 kW; 30 Torr; 2.3 pm;
of aligned CNTs N/A; 2-20 min 40-90 nm
323 Synthesis of multi- Pd / Si CH42H2 = 0.5: 100 sccm 1.1 kW; 4 kPa; N/A;
branched CNTs N/A; 2-20 min 60-300 nm
324 Synthesis of large Fe / Si C2H2:H2=15:60 sccm 100 W; 1200 mTorr; 5-20 pm;
area aligned CNTs 700°C; 5-20 min 40-90 nm
325 Synthesis of MWNTS Ni / Si CH42H2 1 kW; 9.33 kPa; 10-20 “In;
with narrow diameter 850°C; N/A 2-30 nm
326 The inﬂuence of N2 Fe / Si CH4:N2=20:80 700 W; N/A; 30 p.111;
on CNT growth 600°C; N/A 20-100 mm
327 In situ synthesis of Cu electrodes CH4:H2=1 :1 800 W; 2.3 kPa; N/A;
branched Cu—ﬁlled / Si 1150°C; 30 min 40-80 nrn
CNTs
328 Analysis of diameter Cu/ Si CH4:N2=20:80 N/A; 20 Torr; 1 pm;
distribution of CNTs 700-800°C; 5-10 min 20-120 nm
329 Metal analysis of tip Ni, Fe-Ni-Cr CH4:H2=10-20:80-9O 500 W; 250-300 Pa; 1 pm;
of MWNT by EDX substrates sccm 650°C; 30-60 min 60-80 nm
330 Large arrays of well- Ni / Glass C2H2:NH3:N2 N/A; N/A 20 um;
aligned CNTs <666°C; 3-25 min 180-350 mn

 

137

5.3 Experimental

5.3.1 Experimental System
The MPCVD system used in this study is illustrated in Figure 5.6. The system
was built for diamond coating at ﬁrst by the Fraunhofer Center for Coatings and Laser

Applications at MSU. The synthesis process of CNTs by the system is ﬁled to apply

 

 

 

 

 

 

    
 

 

 

 

patent [331].
Microwave -- __ Cavity Side
Excitation Probe Wall
/ Pyrometer

Air Blower Inlet '* "I,

' ' ' Plasma
Substrate """""" "

- - - Quartz Dome
Inlet Gas

 

Outlet Gas

Figure 5.6 Schematic diagram of the MPCVD apparatus at MSU
The photos of the microwave plasma reactor and its control panel are shown in
Figures 5.7 and 5.8, respectively. The microwave power was controlled by hand while
the gas ﬂows were controlled by the computer. The growth temperature of CNTs was

measured by a pyrometer via the screen side window.

138

Microwave ‘ ' Screen Side Windo
Plasma Reactor ‘ .‘ .. .
.’ ; J , :‘

" . 7
ii-
1’

.——*

 

k 0' l I .1 I
1‘... r “f“ ~
y _' . Microwave Power

1 _ __. Control Panel

air -I. .._ _. .-_. __, ____,_,_.1~l4:‘~

 

Figure 5.8 Photo of the control panel of the microwave plasma reactor

139

5.3.2 Experimental Materials

The silicon substrate used to grow CNTs is a boron-doped-p-type Si <100>
substrate with an electric resistivity of 10 Qcm from Silicon Sense Inc. The diameter of
the Si wafer is 2 inch and the dopant is Boron. Its orientation is <100> and resistivity is
1-10 ohm—cm. Its thickness is 254-304 um and grade is test.

The unpuriﬁed SWNTs, which were coated on the Si wafers before CNT growth,
are from the labs of Professor James M. Tour at Center of Nanoscale Science and
Technology at Rice University. The unpuriﬁed SWNTs contain amorphous carbon, Fe,
and SWNTs. The iron content of the SWNTs is on the order of 7%. The manufacturing
method can be found in the literature [332].

The catalyst for CNT growth is Ni on a Si wafer. A thin ﬁlm of Ni with a
thickness of 30 nm is deposited using dc magnetron sputtering on a Si wafer. A 3 inch
pure Ni (99.999%) target from K. J. Lesker Company was used. The deposition of Ni
ﬁlms was carried out under an Ar pressure of 4.0 mTorr at a substrate temperature of
400°C. The working distance from the Ni target to the Si wafer was 2.5 inch while

deposition power was 200W.

5.3.3 Experimental Procedure

1 mg unpuriﬁed SWNT was mixed with 40mL HPLC grade acetone in a beaker.
The mixture was stirred two hours under ultrasonic agitation. The beaker was covered
with aluminum foil to prevent evaporation loss and spill. Then, the Si wafer was coated

with 0.15-0.2 ml SWNT mixture. The Si wafer was dried in atmosphere.

140

Three small pieces (around 5 x 5 m2) of Ni catalyst supplier were placed on one
quarter of one SWNT-coated Si wafer. The Si wafer was placed on a graphite plate (see
Figure 5.9) and the plate was put into the microwave plasma reactor of the MPCVD

system in Figure 5.7.

Ni catalyst

  

Si wafer

Graphite
substrate

Figure 5.9 Photo of Si wafer on a graphite substrate

The system was vacuumed pumped to a base pressure of less than 5 mTorr and
purged with Argon at 365 sccm (Standard Cubic Centimeters per Minute) for 20 minutes.
After turning off Ar gas, the system was pumped to around 0 torr for 5 minutes.

Hydrogen with a set ﬂow rate started and microwave plasma was ignited at 2kW
when the pressure of the reactor reached 5 Torr. The Ni catalysts were heated by H2
plasma to about 700°C with a microwave power of 1.7 kW (total power 2 kW). Then,
methane with a set ﬂow rate was introduced into the plasma chamber to start CNT
growth. Microwave power was increased to 2.2 kW with a total power of 3 kW to keep
the CNT growth temperature is about 750-800°C. If the temperature was higher than

800°C, the microwave p0wer should be lowered. The growth time was 20 minutes.

141

After synthesis of CNTs, a SEM (JEOL 6300F) was used to examine the
morphology of the CNTs. A small piece of CNTs on Si wafer was cut and mounted on a
SEM holder (45° tilted). The sample was coated gold before SEM study. A TEM (J EM-

2200FS) was used to investigate the microstructure of CNTs.

5.4 Results and discussion

5.4.1 CNT Growth Conditions

The grth conditions of CNTs for ﬁve samples are shown in Table 5.2. CH4 was
the source of carbon and the percentage of CH4 in the gas mixture of CH4 and H2
changed from 10% to 100%. The MW power in ﬁve experiments was 2.2 kW except
experiment 2, the power in which was 1.7 kW. The CNT growth pressure in experiment
1-4 were around 34 Torr while that in experiment 5 was lower than 30 Torr, shown in

Table 5.3. Experiment 5 stopped at 12 minute due to disappearance of MW plasma.

Table 5.2 Experimental conditions of the CNT grth

 

No Grown CNTs CH4: H2 Total flow rate (sccm) Power (kW)

 

1 Some 10:90 100 2.2
2 Many 20:80 100 1.7
3 Some 30:70 100 2.2
4 None 50:50 60 2.2
5 None 100:0 36.8 2.2

 

142

 

Table 5.3 CNT growth pressure (Torr)

 

Time (min) Sample 1 Sample 2 Sample 3 Sample 4 Sample 5

 

0 37.5 27.0 28.2 28.7 13.5

2 32.0 30.6 32.4 32.2 16.2

4 36.7 34.6 33.4 36.2 18.8

6 33.9 34.1 34.3 34.3 21.4

8 34.3 34.4 34.2 34.0 23.9

10 34.2 34.2 34.2 34.1 26.6

12 34.2 34.2 34.2 34.1 29.1 If,
14 34.2 34.1 34.2 34.1 '
16 34.2 34.2 34.1 34.2

18 34.1 34.2 34.1 34.2

20 34.2 34.2 34.1 34.1

 

 

 

The temperature data in ﬁve experiments are listed in Table 5.4 while the
temperature proﬁles in experiments 1-4 are shown in Figure 5.10. The temperatures of Si

and Ni are the highest temperatures on Si wafer and Ni catalyst suppliers, respectively.

Table 5.4 CNT growth and catalyst temperatures (°C)

 

Time No.1 No.2 No. 3 No.4 No. 5
(min) Si Ni Si Ni Si Ni Si Ni Si Ni
679 737 676 706 640 680 660 696 662 692
722 737 706 748 732 748 703 731 666 714
761 750 800 N/A 739 747 709 714 650 686
770 765 810 727 760 780 720 743 679 699
770 790 810 741 769 760 717 789 639 671
765 798 810 752 799 761 721 756 619 624
762 790 810 768 797 780 726 737 590 737
14 769 764 804 778 740 779 728 764

16 760 787 793 756 753 776 734 802

18 760 792 787 737 753 776 N/A N/A

20 759 793 808 740 799 767 745 819

 

 

Eamox-ANO

 

The CNT grth temperature in experiment 1 was about 760-770°C and the

catalyst temperature was around 790°C. The CNT growth temperature in experiment 2

143

was stable around 800-810°C, which was higher than the catalyst temperature ranging
from 740 to 770°C. The CNT growth temperature in experiment 3 changed from 740 to
800°C and the catalyst temperature was about 760-780°C. The Si wafer temperature in
experiment 4 was lower than 730°C while that in experiment 5 was lower than 700°C.
The catalyst temperature in experiment 4 was higher than the Si wafer temperature,

ranging ﬁ'om 730 to 800°C but that in experiment 5 was lower than 700°C.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

850 850
No.1 No.2
" _ 6 -
a 800 P” ..... R 5/3/0 I, 9:800
5 3° k'“*’ ““I——I l 5
a 750 — ,n a 750 -
o I "CI" 0
CL
5‘ /" E
o 700 ~ [2 700 -‘
1“ ‘ l
650 I I I I I I I I I 650 I I I I I I I I I
0 2 4 6 8101214161820 0 2 4 6 8101214161820
Tirm (min) Time (min)
850 850
No.3 No.4 0
’ ".y-O~-o 0-" Q) , .‘ ,-
\ Q - / I‘; 0 , \ ,0
§750 ~ 8.3/ ° ° ,0 ‘ €750 ~ 0 ° . . ..
a) J/Q‘” ‘9 0 , 2 o .
3‘ / 3‘ ‘s‘ ’ " '
g 700 <> ,2 700 za‘__,,°'
650 I I I I I f I I I 650 II I I I I T I I I I
0 2 4 6 8101214161820 0 2 4 6 8101214161820
Time (min) Time (min)

Figure 5.10 Temperature proﬁle during CNT grth in experiment No. 1-4: solid points
represent Si wafer; hollow points represent Ni catalysts.

144

5.4.2 CNT Growth Results

Figure 5.11 shows photos of SWNT-coated Si wafer before CNT growth and aﬁer
synthesis in experiment 1-5. Unpuriﬁed SWNTs on a Si wafer were visible in Figure 5.11
(a). After CNT synthesis, a layer of dark ﬁlm of CNTs was visible to the eyes. Visual
observation of the Si wafers suggests that CNTs grew on Si wafers in experiment 1—3

while the thickest CNTs happened in experiment 2. No CNT grew in experiments 4 and

 

5. Further, unpuriﬁed SWNTs were blown off by microwave plasma gas.
0))

(a) (C)
(d) ( ) (0

Figure 5.11 Optical images of the Si wafers: (a) before CNT growth, after CNT growth in
(b) experiment 1, (c) experiment 2, (d) experiment 3, (e) experiment 4, (1) experiment 5.

   

e

Pictures of a graphite substrate with Si wafer and Ni catalyst before and aﬁer
CNT growth in experiment 3 are shown in Figure 5.12. CNTs grew not only on the Si

wafer, but on the graphite substrate and Ni catalyst as well.

145

 

Figure 5.12 Optical images of the graphite substrate with Si wafer and Ni catalyst in
experiment 3: (a) before CNT growth, (b) after CNT growth.

5.4.3 Morphology of CNTs by SEM

The morphology of grown CNTs of samples 1-3 was investigated by SEM.
Figures 5.13-5.15 show SEM images (45° tilted) of CNTs of samples 1-3 while the
summary of information about morphology, length, and diameter of samples is shown in
Table 5.5. Vertically aligned CNTs can be found in three samples while the diameters of
three samples are similar, around 30-60 nm. Nevertheless, the lengths of CNTs and

morphologies of samples are different.

Table 5.5 Summary of morphology, length, and diameter of CNTs

 

 

Sample No. Morphology CNT length (um) CNT diameter(nm)
1 Not well aligned CNTs 40-60 30-60
2 Well aligned CNTs 350—500 30-50
3 Two parts: curled CNTs 0.9-1.1 20-50 (curled CNTs)
and aligned CNTs 40-60 (aligned CNTs)

 

146

 

Figure 5.13 SEM images (45° tilted) at different magniﬁcations of CNTs in sample 1

g ”I. ”5“} \«Nxﬁt‘ f]; ‘3 -" II“ 1

 

Figure 5.14 SEM images (45° tilted) at different magniﬁcations of CNTs in sample 2

147

 

Figure 5.15 SEM images (45° tilted) at different magniﬁcations of CNTs in (a) sample 3,
(b) curled CNT part and (c) aligned CNT part.

Sample 2 has longest vertically well-aligned CNTs with a lengﬂr of approximately
350-500 pm, which may be the longest CNTs synthesized by MPCVD, compared with a
CNT length range of 1-100 urn according to the available literatures [306-330]. The
aspect ratio (Length/Diameter) of sample 2 is around 10,000, which is comparable to
calculated L/D ratio in the literature [318] and larger than that in other surveyed
literatures by MPCVD. The CNTs of sample 2 connected to each other when taken at
larger magniﬁcations, showing in Figure 5.14 (b)-(d). Figures 5.14 (c) and (d) shows

different parts of CNTs of sample 2.

148

According to Figures 5.11 and 5.13-5.15, the inlet gas composition in sample 2
was 20% methane. This is same as synthesis condition in the literatures [313-315, 329].
The CNT grth pressure of samples 1-3 was around 34 Torr (see Table 5.3). The
grth temperature of sample 2 is about 800—810°C while the grth temperature of
samples 1 and 3 was changing from 740-800 °C (see Table 5.4 and Figure 5.10). Other
researchers reached similar results about growth temperature with a range of 800-850°C
[312, 316, 317, 325]. Since the graphite substrate was heated directly by plasma without
other heating sources, the temperature difference among samples 1-5 was caused by the
CNT growth. Taking into account of lower microwave power in experiment 2 than in
other experiments (see Figure 5.2), 1.7 kW vs. 2.2 kW, it is proposed that more growing
CNTs, higher growth temperature.

The CNTs in sample 1, shown in Figure 5.13 (a) and (b), have a shorter length but
maybe a wider distribution than those of sample 2. The length of CNTs in sample 3 (<1
mm) is far less than that in sample 1 (20-40 um) and 2 (250-350 urn). Furthermore, the
CNTs in sample 3 have different structure. The cross-sectional structure of sample 3 in
Figure 5.15 (a) consists of two parts, one aligned upper section and one curly or random
lower section. The length of the curled part is about 150 nm while that of the aligned part
is around 600 nm. The aligned CNT density, shown in Figure 5.15 (c), is higher than the

curly CNT density in Figure 5.15 (b).

5.4.4 TEM Results
The TEM images of CNTs in samples 1-3 are shown in Figures 5.16-5.18.

Bamboo-like CNTs exist in three samples. Other researchers had similar observation

149

[299, 308, 313, 316, 317, 320]. The width of a hollow region is around 20—30 mm, with a
wall thickness in the range of 10-15 nm. The heights of the hollow compartments are not
equal. The inside diameters of the CNTs are around 20—30 nm and the external diameters
are about 40-60 nm. The tip of the arrowhead is in the range of 20-30 nm and no catalyst
particles found inside the tip. The root of CNTs is open-ended while one catalyst particle

is found with the root in sample 3. These phenomena suggest a root-grth mechanism.

 

Figure 5.17 TEM images of CNTs in sample 2: (a) body, (b) tip, (c) root.

150

 

Figure 5.18 TEM images of CNTs in sample 3: (a) body, (b) root, (c) root with a catalyst
particle.

5.4.5 CNT Growth Mechanism
The CH4 source gas was an atomic carbon source for CNTs while Ni was catalyst.
The question is why unpuriﬁed SWNTS were coated on Si wafers before experiments.
This study was based on previous research results by Shuan.ie Zhou within the research
group. The unpuriﬁed SWNTS, coated on Si wafer before experiments, contain
amorphous carbon, Fe, and SWNTS. In order to investigate if the unpuriﬁed SWNTS was
essential for CNT growth, the comparable experiments were carried out. The result that
no CNT grew without the unpuriﬁed SWNTs showed that the unpuriﬁed SWNTS was not
catalyst for CNT growth. The comparable experiments to investigate if Fe was catalyst
for CNT growth were carried out with and without Ni catalyst supplier on the Si wafer.
The result that no CNT grew without Ni catalyst supplier suggested that Fe in the
unpuriﬁed SWNTS was not catalyst for CNT growth. The comparable experiments to
investigate if amorphous carbon in the unpuriﬁed SWNTS was important for CNT grth
were carried out using puriﬁed SWNTS without amorphous carbon. The result that
nothing grew suggested that the amorphous carbon coated on Si wafer is essential for

CNT growth.

151

Growth mechanisms of bamboo-shape MWNTS have been discussed [299, 311,
316, 333]. The root-grth mechanism of bamboo-shaped MWNTS in this study is
illustrated in Figure 5.19. The Ni catalyst supplier was etched by H2 microwave plasma
before the temperature of catalyst reached 700°C, shown in Figure 5.19 (a). The nano-
sized Ni grains formed under the hydrogen plasma heating migrated from the catalyst
supplier to the Si wafer, attaching to the amorphous carbons from unpuriﬁed SWNTS
coated on the Si wafer (see Figure 5.19 (b)). The role of amorphous carbons was to hold
Ni nanoclusters on the Si wafer. The Ni particles in this process remained solid because
the temperatures of growing CNTs and catalysts were lower than 810°C, which is far
below the melting temperature of nickel (1455°C). After CH4 was introduced into the
microwave plasma reactor, the hydrocarbon species decomposed on the surfaces of the
Ni particle, shown in Figure 5.19 (c). Carbon atoms, disassociated from methane,
deposited on the Ni surfaces to form a saturated carbon ﬁhn, shown in Figure 5.19 ((1).
After the Ni and Si substrate surfaces were saturated with carbon layers, the graphitic
sheath was pushed upward while carbon atoms were depositing into graphite layers,
illustrated in Figure 5.19 (e). The graphitic sheath left the Ni particle and another
graphitic sheath deposited on the Ni particle. Therefore, compartments were formed
while carbon atoms were continuous supplied and diffused onto the vertically growing
MWNTS, shown in Figure 5.19 (f). DC bias imposed on the substrate surface was used to
align MWNTS in some literatures [310, 311, 319, 322, 323], but it is not required in this

process.

152

 

 

 

(a) x/ (b)

c8199. 3 $8“

 

 

 

[ L _ , , Sisnbstrate 1 I Si siilisﬂtgte 3]

 

(c) (d)

 

   

 

 

 

l’ _ 51511135811!th

 

 

 

 

 

 

 

(6)

Figure 5.19 Schematic illustrations of root-growth mechanism of a bamboo-like CNT

153

5.5 Conclusions

Carbon Nanotubes were synthesized on silicon substrate by microwave plasma
chemical vapor deposition (MPCVD) of a gas mixture of methane and hydrogen. The
catalyst was nickel, which was not deposited on the substrates directly but migrated from
a small piece of catalyst supplier to the substrate surface during microwave plasma
pretreatment. The Si wafer was coated with amorphous carbon before synthesis.
Additional heating sources and DC bias on graphite substrate are not required. SEM and
TEM were used to characterize the morphologies and microstructures of the CNTs. The
lengths and diameters of CNTs changed with gas composition and grth temperature.
Long vertically-aligned CNTs with a length range of 350-500 tun were synthesized. The
plasma gases included 20 sccm methane and 80 sccm hydrogen. The CNT grth
temperature was 800-810°C. The growth time was 20 minutes. The diameter of CNTs is
around 30-60 nm. The CNTs exhibit bamboo-like structure and may grow via a root-

growth mechanism.

154

CHAPTER 6 CONCLUSIONS

In this study, experimental research was conducted in three sections. The ﬁrst
section was investigation of the curing kinetics of epoxy resins. The second section was
study on dielectric properties of several epoxy resin reacting systems over a temperature
range at 2.45 GHz. The last section was synthesis of carbon nanotubes by microwave
plasma chemical vapor deposition.

The curing of diglycidyl ether of bisphenol A (DGEBA) and 3, 3'-
diarninodiphenyl sulfone (DDS) system under microwave radiation at 145 °C was
governed by an autocatalyzed reaction mechanism. A kinetic model was used to describe
the curing progress.

515:3 = (k1 +k2a'")(1—a)"

where k1 is the non-catalytic polymerization reaction rate constant, k2 is the autocatalytic
polymerization reaction rate constant, m is the autocatalyzed polymerization reaction
order, and n is the non-catalyzed polymerization reaction order.

A single frequency microwave heating and diagnostic system, developed at
Michigan State University, was used to heat and measure the dielectric constant and
dielectric loss factor of curing epoxy resins. A cylindrical TM 012 mode cavity is used to
process epoxy resins at 2.45 GHz. The epoxy resin was DGEBA. The four curing agents
were DDS, a diﬁmctional primary amine (Jeffamine D-230), m-phenylenediamine
(mPDA), and diethyltoluenediamine (Epikure W). The mixtures of DGEBA and the four

curing agents were stoichiometric. The four reacting systems were heated under

155

microwave irradiation to certain cure temperatures, i.e. 145°C for DGEBA/DDS, 90°C
for DGEBA/Jeffamine D-230, 110°C for DGEBA/mPDA, and 160°C for
DGEBA/Epikure W. Measurements of temperature and dielectric properties using the
swept ﬁequency method were made during free convective cooling of the samples. The
cooled samples were analyzed with a Differential Scanning Calorimeter to determine the
extents of cure. The major conclusions of this section are:

0 The Davidson-Cole model can be used to describe the dielectric properties
of the four curing systems.

0 A simpliﬁed Davidson-Cole model is proposed to describe the dielectric
properties of polymeric materials. It works well for the DGEBA/DDS system and low-
temperature or high-extent-of-cure data of the other three systems. It can be used to
calculate the parameters of the Davidson-Cole model as well. The simpliﬁed Davidson-

Cole expression is as follows:

+ (£0 —£oo)

(1601)"

*—

(ao - ego )cos(n [25)

 

£=ew+ ( )n
(or

,, (so — £00 )sin(n %)

8:

 

Wt)"
where 8* is the complex dielectric constant, a' is the dielectric constant, a" is the

dielectric loss factor, j is the imaginary unit, 0) (=21tf, f is the oscillator ﬁ'equency in Hz)

is the radial frequency of the electric ﬁeld in s", t is the relaxation time in s, so is the low

156

 

 

frequency dielectric constant, a... is the high frequency dielectric constant, and n is the

shape parameter with a range of 0 S n3 1.

o The secondary y relaxation has been identiﬁed in the four reacting systems.

The Arrhenius expression is applicable to the y relaxation.

o The evolution of all parameters in the models, e. g. the shape parameter n,
the relaxation strength (so-8.0), the activation energy Ba, and the relaxation time 1, during
cure was related to the facts that the number of the dipolar groups in the reactants
decreases and medium viscosity increases during the polymerization.

o The relaxation strength (ED-80°) was found to decrease during the
polymerization of the four systems, which is consistent with the decreasing number of the
epoxy and amine groups in the reactants.

o The shape parameter 11 decreases as the reaction proceeds for four systems.
n is related to the local chain dynamics. During the reaction the motion of dipolar groups
is hindered by the medium with increasing viscosity. Therefore, the parameter 11
decreases.

- The activation energy of the four systems increases ﬁrst, and then decrease
around 50% extent of cure. The increase of activation energy before the peak is
consistent with increasing viscosity of the systems during cure. The decrease after the
peak point may be explained by that the hindrance ability of the reacted polymer chains
may be weaker than that of dipoles in the reactants and, thus, less energy is needed for
dipolar groups to relax. The peak temperatures of the four systems may be related to the

gel points.

157

l 200W.?!

 

“'5:- ’ 'l .

 

 

o The relaxation time increases as the reaction proceeds due to increasing
medium viscosity for the four systems.

0 The dielectric constants of the four reacting systems exhibit a linear
relation to the extent of cure and may be used to in-situ monitor the curing reaction.

Carbon Nanotubes (CNTs) were synthesized on silicon substrate by microwave
plasma chemical vapor deposition of a gas mixture of methane and hydrogen. The
catalyst was nickel, which was not directly deposited on the substrates but migrated from
a small piece of catalyst supplier to the substrate surface during microwave plasma
pretreatment. The Si wafer was coated with amorphous carbon before synthesis.
Additional heating sources and DC bias on graphite substrate were not employed.
Scanning Electron Microscopy and Transmission Electron Microscopy were used to
characterize the morphologies and microstructures of the CNTs. The lengths and
diameters of CNTs changed with gas composition and grth temperature. Long
vertically-aligned CNTs with a length range of 350-500 mm were synthesized. The
plasma gases included 20 sccm methane and 80 sccm hydrogen. The CNT growth
temperature was 800-810°C. The grth time was 20 minutes. The diameter of CNTs is
around 3060 nm. The CNTs exhibit bamboo-like structure and may grow via a root-

growth mechanism.

158

CHAPTER 7 FUTURE WORK

Development of mathematical models for microwave heating is beneﬁcial for the
process in a number of ways. In addition to setting criteria for material selection and
realizing smooth and precise control, the occurrence of hot spots and thermal runaway
can be predicted with the model and procedures can be taken to avoid these undesired
phenomena. Therefore, modeling of electromagnetic ﬁelds and curing reaction is an
essential to scale-up and apply microwave processing of polymer and composites.

Regarding synthesis of carbon nanotubes, many interesting topics should be
investigated: effect of grth time on CNT growth, effect of catalysts other than Ni,
effect of adding other gas (e. g. N2), the growth mechanism (root-growth or tip-growth),

the possible applications of CNTs.

159

 

10.

ll.

12.

13.

14.

15.

16.

l7.

18.

19.

REFERENCES

Maxwell J C. A Treatise on Electricity and Magnetism. Dover, New York, 1954.

Harrington RF. Time Harmonic Electromagnetic Fields. McGraw-Hill, New York
1961.

Ku CC, Liepins R. Electrical Properties of Polymers: Chemical Principles. Hanser,
Munich, 1987.

Saltiel C, Datta AK. Advances in Heat Transfer 1999, 33, 1.

long L, Zhou 8, Sun L, Kempel LC, Hawley MC. J Polym Sci Part B: Polym Phys
2004, 42, 2871.

long L, Zhou S, Sgriccia N, Hawley MC, Kempel LC. J Microwave Power E E
2003, 38, 49.

Besecker KD, Rhoades CB Jr., Jones BT, Barnes KW. Atom Spectrosc 1998, 19, 55.
J ankowski K, etc. Analytica Chimica Acta 2001 , 440, 215.

LeBlanc GN. Am Lab 2000, 32, 32.

Matusiewicz H. Anal Chem 1994, 66, 751.

Wang Q, Wood PJ, Cui W. Carbohydr Polym 2002, 47, 35.

Bakass M, Mokhlisse A, Lallemant M. Thermochim Acta 2000, 356, 159.
Contat-Rodrigo L, Haider N, Ribes-Greus A et al. J Appl Polym Sci 2001, 76, 1101.
Costley CT, Dean JR, Newton I et al. Anal Commun 1997, 34, 89.

Latham RJ, Linford RG, Pynenburg RAJ. Solid State Ionics 1993, 60, 105.

Nielson RC. J Liq Chromatogr 1993, 16, 1625.

Richter RC, Shah S. Am Lab 2000, 32, 14.

Vandenburg HJ, Clifford AA, Bartle KD, et al. Analyst 1997, 122, 101R.

Chen M, Siochi EJ, Ward TC, Mcgrath JE. Polym Eng Sci 1993, 33, 1092.

160

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

Chen M, Siochi EJ, Ward TC, Mcgrath JE. Polym Eng Sci 1993, 33, 1110.
Zhou S, Yang H, Zong L, Hawley MC. Proc 2001 SPE Auto Compos Conf 2001.
Meghal P, Ubale S, Adgaonkar CS. Indian J Eng Mater S 2001, 8, 55.

Achour ME, Malhi ME, Miane J L, et al. J Appl Polym Sci 1996, 61, 2009.
Achour ME, Malhi ME, Miane IL, et al. J Appl Polym Sci 1999, 73, 969.
Chang JT, Fanning MW, Meaney PM, etc. IEEE T Electromagn C 2000 42, 76.
Lucchese L, Liauw CM, Allen NS, et al. Polym Bull 2000, 44, 187.

Kanynak A, Polat A, Yilmazer U. Mater Res Bull 1996, 31, 1195.

Barnes A, Despotakis A, Wright PV, et al. Electrochimica Acta 1998, 43, 1629.
Biemholtz H, Kenig S, Dodiuk H. Polym Advan Technol 1992, 3, 125.

Gan YX, Chen CQ, Li CG. Sci Eng Compos Mater 1994, 3, 23..

Golosovsky M, Tsindlekht M, Davidov D. Phys Rev 1992, 46, 439.

Schultz JW, Colton J S, Berkowitz CK. Intl SAMPE Symp 2000, 45, 1863.
Yoshida S, Sato M, Sugawara E, et al. J Appl Phys 1999, 85, 4636.

Khan SM, Negi S, Khan IM. Polym News 1997, 22, 414.

Negi S, Gordon K, Khan SM, Khan IM. ACS Sym Ser 1999, 726, 145.

Negi S, Gordon K, Khan SM, Khan IM. Polym Prepr 1998, 39, 721.

Osaki S. J Polym Sci Part B: Polym Phys 1995, 33, 685.

Faez R, Martin IM, DePaoli M, Rezende MC. J Appl Polym Sci 2002, 83, 1568.
Jousse F, Delnaud L, Olmedo O. SAMPE Quart 1995, 360.

Wan M, Li J, Li S. Polym Advan Techn012001, 12, 651.

Sengwa RJ. Polym Int 1998, 45, 202.

Sengwa RJ. Polym Int 1998,45, 43.

161

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

Sengwa RJ, Chaudhary R. Polm Int 2001, 50, 432.

Zoughi R, Lai J, Munoz K. Mater Eva12002, 60, 171.

Osaki S, Tashiro K. Macromolecules 1998, 31, 1661.

Gray S, Ganchev S, Qaddoumi N, et al. Mater Eval 1995, 53, 404.

Kim Y], De Flaviis F, Jofre L, Feng MQ. Proc 46th Intl SAMPE Symp 2001, 439.
Qaddoumi N, Carriveau G, Ganchev S, Zoughi R. Mater Eval 1995, 53, 926.
Parneix JP, Miane J L. European Association for Composite Materials 1992, 821.
Prevors L, Pabiot J. Materiaux et Techniques 1994, 82, 17.

Mijovic J, Wijaya J. Polym Composite 1990, 11, 184.

Wang TB, Liu J. J Electron Manuf 2000, 10, 181.

Mijovic J, Corso WV, Nicolais L, et al. Polym Advan Technol 1998, 9, 231.
Chen M, Hellgeth JW, Ward TC, Mcgrath JE. Polym Eng Sci 1995, 35, 144.

Day RJ, Yau SHC, Hewson KD. Plast Rub Compos Pro 1998, 27, 213.

Pu B, Hawley MC. Polym Eng Sci 2000, 40, 2133.

Gourdenne, A. Polym Mater Sci Eng 1992, 66, 430.

Hawley MC, Wei J, Adegbite V. Mater Res Soc Symp Proc 1994, 347, 669.
Hedreul C, Galy J, Dupuy J. J Appl Polym Sci 1998, 68, 543.

Hedreul C, Galy J, Dupuy J, Delrnotte M, More C. J Appl Polym Sci 2001, 82, 1118.
Jullien H, Petit A. Polym Mater Sci Eng 1992, 66, 378.

Jullien H, Petit A, Merienne C. Polymer, 1996, 37, 3319.

King RJ, Werner MJ, Mayorga GD. J Reinf Plast Comp 1993, 12, 173..
King RJ, Werner M, Mayorga G. SAMPE J 1993, 28, 35.

Liu F, Turner 1, Siores E, Groombridge P. J Microwave Power E E 1996, 31, 71.

162

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

Lefeuvre S, Azzouz F, Salrnoria GV, et al. Ceramic Transactions 2001, 11 l, 153.
Liptak SC, Wilkinson SP, Hedrick JC, et al. Polym Prepr 1990, 31, 320.

Mallon FK, Ray WH. J Appl Polrn Sci 1998, 69, 1203.

McMillan, A; Lauf, RJ; Paulauskas, FL. Proc. Am Chem Soc 1995, 72..
Methven J. Mater Technol 1999, 14, 183.

Methven JM, Ghaffariyan SR, Abidin AZ. Polym Composite 2000, 21, 586.
Muhtarogullari IY, Dogan A, et al. J Appl Polym Sci 1999, 74, 2971.

Outifa L, Jullien H, More C, Delrnotte M. Ind Eng Chem Res 1995, 34, 688.
Prime B. Thermal Characterization of Polymeric Materials. Academic Press, 1981.
Qiu Y, Adegbite v, Hawley MC. Proc 27th Int SAMPE Technol Conf 1995, 173.
Tanikella RV, Allen SAB, Kohl PA. J Appl Polym Sci 2001, 83, 3055.
Thostenson ET, Chou TW. Polym Composite 2001, 22, 197.

Wei J, Hawley MC, Delong JD. Polym Eng Sci 1993, 33, 1132.

Wei J, Hawley MC, Demeuse MT. Polym Eng Sci 1995, 35, 461.

Youngs 1, Power T, Robinson D, et al. Ceramic Transactions 2001, 111, 93.
Youseﬁ A, Laﬂeur PG, Gauvin R. Polym Composite 1997, 18, 157.

Chia LHL, Jacob J, Boey FYC. J Mater Process Tech 1995, 48, 445.

Ng LT, Chia LHL. Polym Int 1999, 48, 952.

Fang X, Hutcheon R, Scola DA. J Polym Sci Part A: Polym Chem 2000, 38, 2526.
Mallakpour SE, Hajipour AR, Khoee S. J Appl Polym Sci 2000, 77, 3003.
Liu Y, Sun XD, Xie XQ. J Polym Sci Part A: Polym Chem 1998, 36, 2653.
Liu Y, Xiao Y, Sun X. J Appl Polym Sci 1999, 73, 2391.

Memon, MS; Yunus, N; Razak, AAA. Int J Prosthodont 2001, 14, 214.

163

89. Boey FYC, Lee WL. J Mater Sci Lett 1990, 9, 1172.

90. Rodrigues DE, Wilkes GL. Polym Mater Sci Eng 1992, 384.

91. Ward TC, Chen M. Polym Mater Sci Eng 1992, 335.

92. Fang X, Scola DA. J Polym Sci Part A: Polym Chem 1999 37, 4616.

93. Fang X, Hutcheon R, Scola DA. J Polym Sci Part A: Polym Chem 2000, 38, 1379.
94. Madras G, Karmore V. Polym Int 2001, 50, 1324.

95. Zhou S, Hawley MC. Proc 46th Int SAMPE Symp 2001, 2243.

96. Hager N, Domszy R. Time-Domain-Reﬂectometry Cure Monitoring. A Materials
and Processes Odyssey; Book 2 of 2 Books (USA), 2001, 2252.

97. Shull PJ, Hurley DH, Spicer JWM, Spicer JB. Polym Eng Sci 2000, 40, 1157.
98. Lye SW, Boey FYC. Mater Manuf Process 1994, 9, 851.

99. Pichon L, Meyer 0. IEEE T Magn 2002, 38, 977.

100. Joly A, Joly JP, Karmazsin E, Veau JL. J Therm Anal 1993, 39, 1285.
101.Joly A, Joly JP, Veau JL, Karmazsin, E. J Therm Anal 1995, 45, 375.

102. Boey FYC, Rath SK. Adv Polym Tech 2000, 19, 194.

103.Karmazsin E, Bechu S. J Therm Anal 1994, 41, 1639.

104. Wei J, Hawley MC. Am Chem Soc 1995, 10.

105. O'Brien KT, Kasturi S, Mekkaoui AM. Polym Composite 1994, 15, 231.

106. Pharn TA, Lancaster GM. Auszuge aus den Europaischen Patentanmeldungen, Teil I
1991, 7, 4189.

107.Kathirgamanathan, P. Polymer 1993, 34, 3105.
108. Margolis JM. Rev Plast Mod 1990, 59, 377.
109. Reitz WE, Oman RM. Adv Mater Process 2000, 158, 49.

110. Staicovici S, Wu CY, Benatar A. J Reinf Plast Comp 1999, 18, 35.

164

 

111.Wu CY, Benatar A. Technomic 1992, 1771.

112. Wu CY, Benatar A. Polym Eng Sci 1997, 37, 738.

113. Zhou S, Hawley MC. Compos Struct 2003, 61, 303.

114. Wei JB, Ngo K, Tucker D, et al. J Microwave Power E E 1998, 33, 10.

115. Ku HS, Ball JAR, MacRobert M, Siores E. Plast Rubber Compos 2000, 29, 278.
116. Shanker B, Lakhtakia A. J Compos Mater 1993, 27, 1203.

117. Lu J, Jiang Q, Zhu X, Wang F. J Appl Polym Sci 2001, 79, 312.

118. Mallakpour SE, et al. J Polym Sci Part A: Polym Chem 2000, 38, 1154.

119. Mallakpour SE, Hajipour AR, Faghihi K. Polym Int 2000, 49, 1383.

120. Mallakpour SE, Hajipour AR, Zadhoush A, et al. J Appl Polym Sci 2001, 79, 1317.
121.Ma11akpour SE, Hajipour AR, Habibi S. Eur Polym J 2001, 37, 2435.

122. Mallakpour SE, Hajipour AR, Mahdavian AR, et al. Eur Polym J 2001, 37, 1199.
123. Mallakpour SE, Hajipour AR, F aghihi K, et al. J Appl Polym Sci 2001, 80, 2416.
124. Mallakpour SE, et al. J Appl Polym Sci Part A: Polym Chem 2001, 39, 177.

125. Srinivasan SA, Joardar SS, Kranbeuhl D, et al. J Appl Polym Sci 1997, 64, 179.
126. Hurduc N, Abdelylah D, Buisine JM, et al. Eur Polym J 1997, 33, 187.

127. Chen J, Chen Q, Yu X. J Appl Polym Sci 1996, 62, 2135.8 (1996).

128. Kishanprasad VS, Gedam PH. J Appl Polym Sci 1993, 50, 419.

129.1mai Y, Nemoto H, Watanabe S, Kakimoto MA. Polym J 1996, 28, 256.
130.1mai Y, Nemoto H, Kakimoto MA. J Polym Sci Part A: Polym Chem 1996, 34, 701.
131.Park KH, Watanabe S, Kakimoto M. Irnai Y. Polym J 1993, 25, 209.

132. Liao LO, et al. J Polym Sci Part A: Polym Chem 2002, 40, 1749.

133.Jacob J, Chia LHL, Boey FY C. J Appl Polym Sci 1997, 63, 787.

165

134. Stoffer JO, Zhang D, Crivello JV. Polym Mater Sci Eng 1999, 81, 118.

135. Stoffer J 0, Zhang D, Crivello JV. Am Chem Soc Div Polym Chem 1999, 40, 669.
136.Dhanalakshmi K, Sundararajan G. Polym Bull 1997, 39, 333.

137. Bogdal D, et al. Polimery 2000, 45, 363.

138ng KS, et al. Polym-Korea 2001, 21, 609.

139. Correa R, Gonzalez G, Dougar V. Polymer 1998, 39, 1471.

140. Lu J, Zhu X, Zhu J, Yu J. J Appl Polym Sci 1997, 66, 129.

14l.Lu J, Zhu X, Ji S, Zhu J, Chen Z. J Appl Polym Sci 1998, 68, 1563.

142. Lu J, Chen N, Ji S, Sun S, Ding M, Zhu X. J Appl Polym Sci 2001, 82, 1356.
143.Xu W, Bao J, Zhang J, Shi M. J Appl Polym Sci 1998, 70, 2343.

144. Palacios J, Sierra J, Rodiguez MP. New Polym Mat 1992, 3, 273.

145. Reis RL, Ribeiro AS, Malafaya PB. 31St Int Biomater Symp 1999, 555.

146. Chan CM, Ko TM, Hiraoka H. Surf Sci Rep 1996, 24, 3.

147.Coen MC, Groening P, Schlapbach L. Le Vide: Science Technique et Applications
1999, 54, 162.

148. Li RZ, Ye L, Mai YW. Compos Part A-Appl 1997, 28, 73.

149. Konig U, et al. Colloid Surface B 2002, 24, 63.

150. Shenton MJ, et al. J Polym Sci Part A: Polym Chem 2002, 40, 95.
151.Hillborg H, et a1. Polymer 2000, 41, 6851.

152.Hollander A, et al. J Polym Sci Part A: Polym Chem 1995, 33, 2013.

153. Lianos L, Parrat D, Hoc TQ, Duc TM. J Vac Sci Technol A 1994, 12, 2491.
154. Walther M, Spallels M. Surf Coat Tech 1996, 80, 200.

155. Etemadi R, et al. J Vac Sci Technol A 1994, 12, 320.

156. Hoﬁichter A, Bulkin P, Drevillon B. J Vac Sci Technol A 2002, 20, 245.

166

157. Rats D, Hajek V, Martinu L. Thin Solid Films 1999, 340, 33.

158.Bergeron A, et al. J Vac Sci Technol A 1998, 16, 3227.

159.1nagaki N. Macromol Symp 2000, 159, 151.

160. Lennon P, et al. Int J Adhes Adhes 1999, 19, 273.

161. Shi MK, et al. J Adhes Sci Technol 2000, 14, 1485.

162. Klemberg—Sapieha J E, Shi MK, et a1. Society Plastics Engineers 1995, 2862.
163.Keil M, et al. Appl Surf Sci 1998, 125, 273..

164. Felix J, Gatenholrn P, Schreiber HP. J Appl Polym Sci 1994, 51, 285.

165. Schreiber HP, St Germain F. J Adhes Sci Eng 1990, 4, 319.

166. Aumann T, Theirich D, Engemann J. Surf Coat Tech 2001, 142, 169.

167. Tomcik B, Popovic DR, Jovanovic IV, Prtrovic ZL. J Polym Res 2001, 8, 259.
168. Peng J, Jian XG, Li Q, Deng XL, Gu B. ACTA Polymerica SINICA 2000, 3, 354.
169. Zhou XT, et al. Surf Coat Tech 2000, 123, 273.

170. Korzec D, Traub K, Werner F, Engemann J. Thin solid ﬁlms 1996, 281-282, 143.
171 . Graff W, Matson M, Kellner T, et al. Solid State Technol 2001, 44, 37.
172.Nitschke M, Schmack G, J anke A, et al. J Biomed Mater Res 2002, 59, 632.
173.Nitschke M, Menning A, Werner C. J Biomed Mater Res 2000, 50, 340.

174. Sodhi RNS Sahi VP, Mittelrnan MW. J Electron Spectrosc 2001, 121, 249.

175. Schroder K, Meyer-Plath A, Keller D, et al. Contrib Plasm Phys 2001, 41 , 562.
176. Wickson BM, Brash JL. 31St Int Biomater Symp 1999, 57.

177. Chou NJ, Marwick AD, Goldblatt RD, et al. J Vac Sci Technol 1992, 10, 248.
178. Kobayashi T, Sasama T, Wada H, Fujii N. J Vac Sci Techn012001, 19, 2155.

179. Castro RM, Verdonck P, Pisani MB, et al. IEEE T Plasms Sci 2000, 28, 1043.

167

 

180. Schmitt BR, Kim H, Urban MW. J Appl Polym Sci 1999, 71, 1.

181.Bich1er CH, Langowski HC, Moosheimer U, et al. J Adhes Sci Eng 1997, 11, 233.
182. Normand F, Granier A, Leprince P, et a1. Plasma Chem Plasma P 1995, 15, 173.
183. Ringengbach A, Jugnet Y, Duc TM. J Adhes Sci Technol 1995, 9, 1209.

184. Kaiser M, Walker M, Baumgartner KM, et a1. Surf Coat Tech 1998, 105, 165.
185. Schwarz J, Schmidt M, Ohl A. Surf Coat Tech 1998, 98, 859.

186. Behnisch J, Mehdom F, Hollander A, et al. Surf Coat Tech 1998, 98, 875.

187. Groning P, Schneuwly A, Schlapbach L, et al. J Vac Sci Technol A 1996, 14, 3043.
188.Hlavac M, Ohlidal M, Slavinska D, Int J Electron 1992, 73, 1067.

189. Biederrnan H, Osada Y. Plasma Polymerization Processes. Elsevier, 1992.

190. Ludlow-Palafox C, Chase HA. Ind Eng Chem Res 2001, 40, 4749.

191.Kleps T, Piaskiewicz M, Parasiewicz W. J Therm Anal Calorim 2000, 60, 271.
l92.Klun U, Krzan A. Polymer 2000, 41, 4361.

193.Krzan A. J Appl Polym Sci 1998, 69, 1115.

194.Danko GA, Silberglitt R. J Am Ceram Soc 2000, 83, 1617.

195. Hernandez MT, Gonzalez M. Euro Ceramics VII (1-3) 2002, 206-2, 71.
196.Yang ST, Zhang YF, Lu QZ, et al. J Inorg Mater 2000, 15, 309.

197. Wu C, Gao J, Li M, Zhang WM, Jiang M. Macromol Symp 2000, 150, 219.
198. Spatz J, Mossmer S, Moller M, et al. Adv Mater 1998, 10, 473.

199. Srikanth H, Hajndl R, Chirinos C, Appl Phys Lett 2001, 79, 3503.

200. Tu WX, Lin HF. Chem Mater 2000, 12, 564.

201. Szabo DV, Vollath D, Arnold W. Ceramics Transactions 2001, 111, 217.

202. Szabo DV, Vollath D. Adv Mater 1999, 11, 1313.

168

 

203.Vollath D. Mater Technol 1997, 12, 89.

204. Willert-Porada MA. MRS Bull 1993, 18, 51.

205. White WC. Proc Inst Radio Electron Eng 1962, 50, 1129.

206. Chabinsky IJ. Mater Res Soc Symp Proc 1988, 17, 124.

207. Krieger B. Proc Am Chem Soc Div Polym Mater Sci Eng 1992, 66, 339.
208. Smith RD. EPRI Report EM-3645 (1984).

209. Nelson SO. J Microwave Power E E 1985, 30, 65.

210. Vetsuypens JP, Loock WV. J Microwave Power E E 1986, 31, 110.
211.Lightsey G, et al. J Microwave Power E E 1988, 33, 11.

212. Thiebaut JM, Akye C, Roussy G. IEEE T Instrurn Meas 1988, 37, 114.
213. Lightsey G, George C, Russell LD. J Microwave Power E E, 1986, 31, 86.

214. Lee, SM. International Encyclopedia of composites. VCH publishers, New York,
1990, 6, 339.

215. Finzel M. PhD Dissertation. Michigan State University, 1991.

216. Wei J. PhD Dissertation. Michigan State University 1992.

217. Adegbite V. PhD Dissertation, Michigan State University 1995.

218. Jow J. PhD Dissertation. Michigan State University, 1988.

219. Qiu Y. PhD Dissertation. Michigan State University 2000.

220. Zhou S. PhD Dissertation, Michigan State University, 2002.

221.Lewis DA, Summers JD, et al. J Polym Sci Part A: Polym Chem 1992, 30, 1647.
222. Marand E, Baker KR, Graybeal JD. Macromolecules 1992, 25, 2243.

223. Hedrick JC, Lewis DA, et al. Proc ACS Div Polym Mater Sci Eng 1989, 60, 438.

224. Lee H, Neville K. Epoxy Resins: Their Applications and Technology. McGraw-Hill,
New York, 1957, 41.

169

 

225. Schechter L, Wynstra J, Kurkjy RP. Ind Eng Chem 1956, 48, 94.

226. Sourour S, Kamal MR. Thermochim Acta 1976, 14, 41.

227. Lee WI, Loos AC, Springer GS. J Compos Mater 1982, 16, 510.
228.Dannenberg H. SPE Trans 1963, 3, 78.

229. Horie K, Hiura H, Sawada M, Mita I, Kambe H. J Polym Sci A1 1970, 8, 1337.
230.Dusek K, Ilavsky M, Lunak S. J Polym Sci Pol Sym 1976, 53, 29.
231.Riccardi CC, Adabbo HE, Williams RJ. J Appl Polym Sci 1984, 29, 2481.

232. Acitally MA, Prime RB, Sacher E. Polymer 1971, 12, 335.

233.Kamal MR. Polym Eng Sci 1974, 14, 231.

234. Kamal MR, Sourour S. Polym Eng Sci 1973, 13, 59.

235. Dutta A, Ryan ME. J Appl Polym Sci 1979, 24 635.

236.Mijovic J. J Appl Polym Sci 1986, 31, 1177.

237. Moroni A, Mijovic J, Pearce EM, Foun CC. J Appl Polym Sci 1986, 32, 3761.
238. Progelhof RC, Throne JL. Polym Eng Sci 1975, 15, 690.

239. Chan AW, Hwang ST. Polym Eng Sci 1991, 31, 1149.

240. Lin R, Lee LJ, Liou M. Intern Polym Process 1991, VI, 356.

241.Gebart BR. J Appl Polym Sci 1994, 51, 153.

242. The LUXTRON model 750 Fluoroptic Thermometry TM System Service and

Technical Reference Manual. LUXTRON F luoroptic TM Temperature Sensing,
California, 1986.

243. The NoEMI-TSTM Family Fiberoptic Thermometer Systems User’s Guide. Nortech
Fibronic Inc., Quebec Canada. 1996.

244. Levine WS. The control handbook. Boca Raton Fla CRC Press, 1996.
245. Daniel W. Dielectric Relaxation. London Academic Press, 1967.

246. Debye P. Polar Molecules. Dover, New York, 1945.

170

247. Cole KS, Cole RHJ. Chem Phys 1941, 9, 341.

248.Davidson DW, Cole RHJ. Chem Phys 1951, 19, 1484.

249.Havriliak S, Negami S. J Polym Sci Part C 1966, 14, 99.

250. Havriliak S, Negami S. Polymer 1967, 8, 161.

251.Kauzmann W. Rev Mod Phys 1942, 14, 12.

252. Erying H. J Chem Phys 1936, 4, 283.

253.Williams ML, Landel RF, Ferry JD. J Am Chem Soc 1955, 77, 3701.
254.Delmonte J. J Appl Polym Sci 1959, 2, 108.

255.Haran EN, Gringras H, Katz D. J Appl Polym Sci 1965, 9, 3505.

256. Kranbuehl D, Delos S, Yi E, et a1. Polym Eng Sci 1986, 26, 338.

257. Day DR. Polym Eng Sci 1986, 26, 362.

258. Lane JW, Seferis JC, Bachmann MA. J Appl Polym Sci 1986, 31, 1155.
259. Lane JW, Seferis JC, Bachmann MA. Polym Eng Sci 1986, 26, 354.

260. Sanjana ZN. Polym Eng Sci 1986, 26, 373.

261. Senturia SD, Seppard NF. Adv Polym Sci 1986, 80, 1.

262.Zukas WX, Wentworth SE. Polym Compos 1987, 8, 232.

263. Pochan JM, Gruber RJ, Pochan DF. J Polym Sci Polym Phys 1981, 19, 143.
264.Koike T. Adv Polym Sci 1999, 148, 139.

265. Koike T, Tanaka R. J Appl Polym Sci 1991, 42, 1333.

266. Casalini R, Fioretto D, Livi A, et al. Phys Rev B 1997, 56, 3016.

267 .Butta E, Livi A, Levita G, et al. J Polym Sci Part B: Polym Phys 1995, 33, 2253.
268. Casalini R, Livi A, Rolla PA, Levita G, Fioretto D. Phys Rev B 1996, 53, 564.

269. Gallone G, Capaccioli S, Levita G, et al. Polym Int 2001, 50, 545.

171

 

270. Domenici C, Levita G, Marchetti A, Frosini V. J Appl Polym Sci 1987, 34, 2285.
271 . Simpson JO, Bidstrup SA. J Polym Sci Part B: Polym Phys 1993, 31, 609.

272. Sheppard NF, Senturia SD. J Polym Sci Part B: Polym Phys 1989,27, 753.

273. Sheppard NF, Senturia SD. Polym Eng Sci 1986, 26, 354.

274. Corezzi S, Capaccioli S, Gollone G, et al. J Phys-Condens Mat 1997, 9, 6199.
275. Capaccioli S, Corezzi S, Gallone G, et al. J Non-Cryst Solids 1998, 235-237, 576.
276. Livi A, Levita G, Rolla PA. J Appl Polym Sci 1993, 50, 1583.

277. Carrozzino S, Levita G, Rolla P, Tombari E. Polym Eng Sci 1990, 30, 366.

278. Eloundou JP, Gerard JF, Pascault JP, et al. Macromol Chem Phys 2002, 203, 1974.
279. Kortaberria G, Arruti P, Mondragon I. Polym Int 2001, 50, 957.

280. Lairez D, Emery JR, Durand D, Pethrick RA. Macromolecules 1992, 25, 7208.
281.Simpson JO, Bidstrup SA. J Polym Sci, Part B: Polym Phys 1995, 33, 55.

282.Zong L, Charvat GL, Kempel LC, Hawley MC. 26th AMTA Meeting & Symp,
Stone Mountain, GA, 2004.

283. Schonhals A, Schlosser E. Colloid Polym Sci 1989, 267, 125.

284. Schlosser E, Schonhals A. Colloid Polym Sci 1989, 267, 133.

285. Viovy J L, Monnerie L, Brochon JC. Macromolecules 1983, 16, 1845.
286. http://cohesion.rice.edu/naturalsciences/smalley/emplibrary/allotropes.jpg
287.1ijima, S. Nature 1991, 354, 56.

288. http://www.labs.nec.co.jp/Eng/innovative/E1/top.html

289. http://www.nanotech-now.com/nanotube-buckyball-sites.htm

290. Iijima S, Ichihashi T. Nature 1993, 363, 603.

291. Bethune DS, Kiang CH, DeVries MS, et al. Nature 1993, 363, 605.

292. Qin L, Zhao X, Hirahara K, Miyamoto Y, Ando Y, Iijima S. Nature 2000, 408, 50.

172

 

293. Gao G, Cagin T, Goddard WA. Nanotechnology 1998, 9, 184.

294. Yu M, Files BS, Arepalli S, Ruoff RS. Phys Rev Lett 2000, 84, 5552.

295. Wong EW, Sheehan PE, Lieber CM. Science 1997, 277, 1971.

296. http://www. ﬁbrils.com

297. Baughman RH, Zakhidov AA, DeHeer WA. Science 2002, 297, 787.

298. Shi Z, Lian Y, Zhou X, et a1. Carbon 1999, 37, 1449.

299. Saito Y. Carbon 1995, 33, 979.

300. Maser WK, Munoz E, Benito AM, et a1. Chem Phys Lett 1998, 292, 587.
301.Puretzky AA, Geohegon DB, Fan X, Pennycook SJ. Appl Phys Lett 2000, 7 6, 182.
302. Ci L, Li Y, Wei B, Liang J, Xu C, Wu D. Carbon 2000, 38, 1933.

303. Liu XW, Lin JH, Hsieh WJ, Shih HC. Diam Relat Mater 2002, 11, 1193.

304. Mauron Ph, Emmenegger Ch, Zuttel A, et al. Carbon 2002, 40, 1339.

305. Zhu HW, Xu CL, Wu DH, Wei BQ, Vajtai R, Ajayan PM. Science 2002, 296, 884.
306. Ting J, Liu R. Carbon 2003, 41, 579.

307. Kim S, Park J, Kim B, et al. Mater Sci Eng C 2004, 24, 297.

308. Kim U, Pcionek R, Aslam DM, Tomanek D. Diam Relat Mater 2001, 10, 1947.
309. Chen C, Perry WL, Xu H, Jiang Y, Phillips J. Carbon 2003, 41, 2555.

310. Chen M, Chen C, Chen C. J Mater Sci 2002, 37, 3561.

31 1. Chen M, Chen C, Koo H, Chen C. Diam Relat Mater 2002, 37, 3561

312. Bower C, Zhu W, Jin S, Zhou, 0. Appl Phys Lett 2000, 77, 830.

313. Chen L, Wen C, Liang C, et al. Adv Funct Mater 2002, 12, 687.

314. Choi YC, Bae DJ, Lee YH, et al. J Vac Sci Technol A 2000, 18, 1864.

315. Choi YC, Shin YM, Lee YH, et al. Appl Phys Lett 2000, 76, 2367.

173

316. Cui H, Zhou O, Stoner BR. J Appl Phys 2000, 88, 6072.

317. Gao, J S, Umeda K, Uchino K, et al. Mater Sci Eng A 2003, 352, 308.

318. Gao, J S, Umeda K, Uchino K, et a1. Mater Sci Eng B 2004, 107, 113.

319.Jong WJ, Lai SH, Hong KH, Lin HN, Shih HC. Diam Relat Mater 2002, 11, 1019.
320. Lin CH, Chang HL, Hsu CM, Lo AY, Kuo CT. Diam Relat Mater 2003, 12, 1851.
321.Tsai SH, Shiu CT, Lai SH, et al. J Mater Sci Lett 2002, 21, 1709.

322. Tsai SH, Chao CW, Lee CL, Shih HC. Appl Phys Lett 1999, 74, 3462.

323. Tsai SH, Shiu CT, Lai SH, Shih HC. Carbon 2002, 40, 1597.

324. Wang YH, Lin J, Huan CHA, Chen GS. Appl Phys Lett 2001, 79, 680.

325. Wang SG, Zhang Q, Yoon SF, Ahn J. Scripta Mater 2003, 48, 409.

326. Wang EG, Guo ZG, Ma J, et al. Carbon 2003, 41, 1827.

327.Zhang GY, Wang EG. Appl Phys Lett 2003, 82, 1926.

328. Zhu YB, Wang WL, Liao KJ, Ma Y. Diam Relat Mater 2003, 12, 1862.

329. Okai M, Muneyoshi T, Yaguchi T, Sasaki S. Appl Phys Lett 2000, 77 , 3468.

330. Ren ZF, Huang ZP, Xu JW, et al. Science 1998, 282, 1105.

331. Hawley MC, Asmussen J, Zhou S, Zuo S. An integrated catalyst deposition and

nanotube growth process with microwave plasma chemical vapor deposition. Patent
application in process.

332.Nikolaev P, Bronikowski MJ, Bradley RK, et al. Chem Phys Lett 1999, 313, 91.

333.Har1is PJF. Carbon Nanotubes and Related Structures. Cambridge University Press.
1999, 30—33.

174