close

Вход

Забыли?

вход по аккаунту

?

Magnetic nanocomposites: A new synthesis method and microwave absorption properties

код для вставкиСкачать
UNIVERSITY OF CALIFORNIA
Los Angeles
Magnetic Nanocomposites:
A New Synthesis Method and Microwave Absorption Properties
A dissertation submitted in partial satisfaction o f the
requirements for the degree Doctor o f Philosophy
in Mechanical Engineering
by
Sung Sik Park
2006
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
UMI Number: 3251446
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignm ent can adversely affect reproduction.
In the unlikely event that the author did not send a com plete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3251446
Copyright 2007 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
© Copyright by
Sung Sik Park
2006
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
The dissertation of Sung Sik Park is approved.
Nasr Ghoniem
H. Thomas Hahn, Committee Chair
University o f California, Los Angeles
2006
ii
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
This dissertation is dedicated to my beloved Lord, and ever supporting
my wife Anna.
iii
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
TABLE OF CONTENTS
1. Introduction
1
1.1 Fundamentals o f Electromagnetic Waves
3
1.2 Reflection Coefficients
8
1.3 Loss Mechanism o f Microwave Absorbers
10
1.4 The Effect o f Particle Characteristics on Microwave Properties
15
1.5 Characteristics o f Magnetic Nanoparticles
19
1.6 Motivation for Magnetic Nanocomposites
as Microwave Absorbers
27
1.7 Reference
29
2. Processing of Iron Oxide/Epoxy Vinyl Ester Nanocomposites
36
2.1 Materials
36
2.2 Particle Dispersion and Stability
37
2.3 Curing Behavior
51
2.4 Conclusion
54
2.5 Reference
56
iv
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
3. Flexible Magnetic Nanocomposites Fabrication through
Particle Surface Initiated Polymerization
57
3.1 Motivation
57
3.2 Material
58
3.3 Nanocomposite Fabrication
59
3.4 Characterization
61
3.5 Result and Discussion
62
3.6 Conclusion
78
3.7 Reference
79
4. Microwave Properties of Fe2 C>3 , Fe and FeCo Based Magnetic
Nanocomposite
80
4.1 Motivation
80
4.2 Material
84
4.3 Sample Preparation
87
4.4 Characterization
90
4.5 Result and Discussion
93
4.6 Conclusion
118
4.7 Reference
120
5. Summary and Future Work
123
v
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
LIST OF FIGURES
Figure 1.1 Electromagnetic Wave.................................................................................... 7
Figure 1.2 Effect o f thickness o f microwave absorber on reflection coefficient.......... 9
Figure 1.3 Effect of //” on RL and thickness o f microwave absorber..........................14
Figure 1.4 Effect damping coefficient and anisotropy ratio on susceptibility............ 27
Figure 2.1 Influence o f ultrasonication time on dispersion quality o f
Fe 3 0 4 particles: particle volume fraction o f 2 %, ultrasonic power o f 120 W,
and sedimentation time o f 24 h r............................................................................ 38
Figure 2.2 Influence o f ultrasonication time and resin temperature
on dispersion quality o f Fe 3 0
4
particles: particle volume fraction o f 2 %,
ultrasonic power o f 400 W, and sedimentation time o f 72 h r............................39
Figure 2.3 Influence o f particle volume fraction on dispersion
quality o f Fe 3 C>4 particles: ultrasonic power o f 400 W,
cooling, and sedimentation time o f 72 h r............................................................. 40
Figure 2.4 SEM photomicrographs o f fracture surfaces o f cured resin
specimens: 4 vol. % Fe 3
0 4
particles dispersed ultrasonically
for (a) 5min,(b)l lm in, and (c) 20 m in ................................................................41
Figure 2.5 SEM photomicrographs o f fracture surfaces o f cured
resin specimens: 4 vol. % Fe 3 0 4 particles dispersed
ultrasonically for 11 minutes................................................................................ 41
vi
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F urth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 2.6 Influence o f ultrasonication time on dispersion
quality o f Fe 2 C>3 nanoparticles in the resin..........................................................43
Figure 2.7 Experimental Setup o f Fe203 Dispersion Study........................................ 44
Figure 2.8 Resin temperature vs. sonication tim e..........................................................45
Figure 2.9 a Micro photos showing particle dispersion at power level 1..................45
Figure 2.9 b Micro photos showing particle dispersion at power level 2 ..................46
Figure 2.9 c Micro photos showing particle dispersion at power level 3 ..................46
Figure 2.10 Sonication power output Vs sonication duration....................................... 49
Figure 2.11 Reaggregation o f particles after sonication................................................ 49
Figure 2.12 Manually Stirred after 48 hours past sonication....................................... 50
Figure 2.13 Measured heat flow from different particle loading.................................. 51
Figure 3.1 Chemical formulats o f the two-part polyurethane monomers used.......... 59
Figure 3.2 Surface Initiated Polymerization....................................................................60
Figure 3.3 (a) TGA o f as-received nanoparticles, catalyst-accelerator
mixture, and treated nanoparticles; (b) FT-IR spectra o f catalyst-accelerator
mixture and treated nanoparticles......................................................................... 63
Figure 3.4 TGA o f the PU-NPs after reaction for 6 hours and washed with excess
THF before p o u rin g ............................................................................................... 64
Figure 3.5 Scheme o f the NPs filled polyurethane com posites................................... 65
vii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 3.6 TGA of neat polyurethane, composites made by
DM and SIP methods measured at different sample locations......................... 66
Figure 3.7 DGT o f Poylurethane, DM and SIP nanocomposites
........................66
Figure 3.8 FT-IR spectra o f catalyst-accelerator mixture and treated nanoparticles..68
Figure 3.0 FT-IR spectra o f neat polyurethane, as-received nanoparticles and
nanocom posite........................................................................................................69
Figure 3.10 X-ray Photoelectron Spectroscopy (XPS) analysis on as-received
70
Figure 3.11. DM composite (left), neat polyurethane (middle) and SIP
composite(right).....................................................................................................72
Figure 3.12 AFM tapping mode phase images o f the 65 %wt.
composites by (a) DM method and (b) SIP method........................................... 73
Figure 3.13 Scanning electron micrographs o f composites:
SIP method (a) and (c); and DM method (b) and (d).......................................74
Figure3.14 Chemical Structure o f PET.............................................................................75
Figure 3.15 As-received nanoparticles remaining on the PET
vial wall (left), and composite nanoparticles almost completely
removed from the PET vial w all.......................................................................... 76
Figure 3.16 Tensile tress-strain curves o f 65 %wt. Composites
fabricated by DM method and SIP methods.....................................................77
viii
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 4.1. Synthesis o f carbonyl iron particle (Cl particle).......................................... 84
Figure 4. 2. Carbonyl iron particles................................................................................... 85
Figure 4.3. Processing nanoparticles through the gas condensation m ethod................86
Figure 4.4. TEM images o f a) Fe and b) FeCo nanoparticles.........................................89
Figure 4.5. 7mm set-up configuration...............................................................................91
Figure 4.6. Scattering coefficients o f a specimen on a coaxial transmission line
92
Figure 4.7 SEM image o f Fe/PU nanocomposite (65% wt. Loading)........................ 94
Figure 4.8 Normalized Ms for different particle loading.............................................95
Figure 4.9 Magnetization vs Magnetic Field for CIP/PU composite
film with different particle loading...................................................................... 96
Figure 4.10 Magnetization vs magnetic field for CIP/PU composite
films with different particle loadings..................................................................97
Figure 4.11 Magnetization vs magnetic field for FeCo/PU..........................................98
Figure 4.12 Magnetization vs magnetic field for Fe/PU ............................................... 99
Figure 4.13 The relationship between coercive force and particle
diameter.................................................................................................................. 101
ix
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 4.14 Real permittivities of CI/PU, Fe/PU and FeCo/PU
nanocomposite films............................................................................................ 104
Figure 4.15 Imaginary permittivities o f composite film s........................................... 105
Figure 4.16 Real permeabilities o f composite films ....................................................106
Figure 4.17 Imaginary permeabilities o f composite films..........................................106
Figure 4.18. Comparison between the SIP and DM on permittivity..........................108
Figure 4.19. Comparison between the SIP and DM on permeability.........................108
Figure 4.20 Metal backed reflection loss for CI/PU composite
films with thickness = 1.02 m m .........................................................................109
Figure 4.21 Metal Backed Reflection Loss for CIP/PU composite
film with thickness = 2.04 m m ............................................................................110
Figure 4.22 Metal Backed Reflection Loss for Fe/PU and
FeCo/PU nanocomposite film with thickness = 2.04 m m ............................... I l l
Figure 4.23 Metal backed reflection loss comparison between
CI/PU composite film and Fe/PU nanocomposite
for a target frequency o f 8GHz............................................................................113
Figure 4.24 Typical single absorber weight and thickness
for the different target frequencies...................................................................... 114
x
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 4.25 Reflection spectra for three model composites.
Particle size=20 nm (Co, ferrite), 10 nm (Fe);
particle volume fraction=0.2 (Co, Fe), 0.5 (ferrite).
d is the composite thickness................................................................................ 116
Figure 4.26 Multi-band absorber from RF Product,
A Laird Technology Company............................................................................117
xi
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
LIST OF TABLES
Table 1.1. Magnetic properties o f various magnetic nanoparticles............................25
Table 2.1. Influence o f iron oxide particles on curing................................................. 51
Table 2.2. Curing Behavior o f Fe203 for different particle loading..........................54
Table 4.1. Density o f CIP/PU, Fe/PU and FeCo/PU....................................................85
xii
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
ACKNOWLEDGMENTS
I would like to express m y deepest gratitude to my advisor Professor H. Thomas
Hahn, for providing the valuable chance for me to finish my Ph.D study in the
Mechanical and Aerospace Engineering Department. It is his insightful guidance,
continuous inspiration and patience throughout the duration o f the project, which
made this work possible. Especially, his advice o f job opportunity at Northrop
Grumman Corporation launched m y career and inspired the search o f better material
systems for the microwave absorber application which became the foundation o f this
thesis. For that I am sincerely grateful.
I would like to thank Prof. Nasr Ghoniem, Prof. Vijay Gupta, from Mechanical &
Aerospace Engineering Department and Prof. Jenn-Ming Yang from Material Science
Engineering for serving on my examination committee. I also thank Prof.D. Young
from Department o f Physics at Louisiana State University in the magnetic and
magnetoresistacne measurement.
xiii
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
My special thanks go to Dr. Zhanhu Guo for his appetite for writing a good scientific
journal paper. W ith the collaboration with Dr. Gou, I have learnt a lot such as how to
systematically design research, how to collect data and utilize them for interpretation,
and how to organize the data for peer-reviewed paper preparation.
Continuous
discussion o f the result and sharing o f new ideas with him were a very satisfying
experience.
Also I would like to thank Northrop Grumman Corporation for the financial support
for this project. M y special thanks to Dr. John W illis for his guidance and Terry Hall
for the proof reading o f the thesis.
And the last, but not the least, I thank m y wife, Anna.
It was her continuous
encouragement and sacrifice that made this work possible. I truly owe everything to
her for all my advances in academia as well as in my professional career.
xiv
R e p r o d u c e d w ith p e r m is s io n o f th e c o p y r ig h t o w n e r . F u rth er r ep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
VITA
1994
B.Sci., Aerospace Engineering, University o f Maryland at College
Park
1997
M.S., Aerospace Engineering, University o f California, Los Angeles
1998
Engineer, Advanced Materials and Processes
Present
Development, Northrop Grumman Corporation, Integrated Systems
Western Region
PUBLICATIONS AND PRESENTATIONS
Park, S.B., Park, S.S., Carman, G.P, and Hahn, H.T. “Measuring Strain Distribution
During Mesoscopic Domain Reorientation In A Ferroelectric Material,” Journal o f
Engineering Materials and Technology, Vol. 120, 1998, pp. 1-6.
Park, S.S, Bemet, N., de la Roche, S., and Hahn, H.T. “Processing o f Iron OxideVinyl Ester Nanocomposites,” Journal o f Composite Materials, Vol. 37, No. 5, 2003,
pp. 465-476.
Guo, Z., Park, S., Young, D., and Hahn, H.T. “Flexible High-loading Particle
Reinforced Polyurethane Magnetic Nanocomposites Fabrication Through Particle
Surface Initiated Polymerization,” In preparation
xv
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
ABSTRACT OF THE DISSERTATION
Magnetic Nanocomposites:
A New Synthesis Method and Microwave Absorption Properties
by
Sung Sik Park
Doctor o f Philosophy in Mechanical Engineering
University o f California, Los Angeles, 2006
Professor H. Thomas Hahn, Chair
Electromagnetic absorbers are a critical part o f defense systems for their contribution
to survivability in air vehicles and for use as commercial products for electromagnetic
interference (EMI) shielding.
The emergence o f nanomaterials technology opened
the door for new opportunities to further improve the functionality o f electromagnetic
absorbers. However, the challenge o f incorporating nanoparticles into a polymer is to
overcome the difficulty o f dispersing large volume fractions o f nanoparticles into the
polymer without sacrificing the mechanical properties o f the resulting composite, a
xvi
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
problem caused by the increase o f the particle surface area as the particle size
decreases to the nanometer range.
In this study, Surface Initialized Polymerization
(SIP) is used to provide physicochemical adsorption o f the catalyst/promoter onto the
magnetic nanoparticle (Fe2 C>3 ) surface in a tetrahydrofuran (THF) solution.
This
method
The
results
in
a
highly
flexible
magnetic
composite
material.
physicochemical attachment o f catalyst/promoter on the nanoparticle surface was
proven by thermal gravimetric analysis (TGA) and Fourier-transform infrared (FTIR) analysis, surface image analysis (Atomic Force Microscopy, AFM and Scanning
Electron Microscopy, SEM), and mechanical testing. The saturation magnetization,
complex permeability, and complex permittivity o f SIP magnetic nanocomposites
with nano-sized Fe and FeCo magnetic particles (< 20nm diameter) were
investigated.
The magnetic properties were measured with a 9-Telsa Physical
Property Measurement System (PPMS).
The electromagnetic properties were
obtained through transmission line measurements using an HP Network analyzer
851 OB over the frequency range from 2 to 18 GHz. The results were then compared
with the commercial electromagnetic wave absorbers that use the micron-size
carbonyl iron particles (CIP, 2~5 pm diameter).
A direct comparison was made by
analyzing the reflection loss performance between the magnetic nanocomposite films
and the CIP added composite film. This was done to show that a weight saving can
be realized by using the magnetic nanocomposite films instead o f the CIP
microcomposite.
xvii
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Chapter 1
Introduction
The use o f nanoparticles in polymeric systems has become a subject o f interest in
engineering applications due to potential dramatic changes in physical properties
o f composites.
These changes in properties come from two aspects o f
nanoparticles: increased surface area and quantum effects associated with nano­
dimensional particle structures [1].
These factors can change or enhance
properties such as reactivity, strength, and electromagnetic properties o f
composites.
The most extensively studied nanoparticles are montromillonite silicate clay
having a mean size o f a few microns [2-4]. The clay particle consists o f silicate
layers about 1 nm thick that are separated by van der Waals interlayers containing
charge-compensating cations (M+). When these silicate layers are fully dispersed
at the molecular level in a polymer through intercalation and exfoliation, the
resulting strengthening far exceeds what is possible from solid clay particles. The
improvements are believed to be due to the combined effect o f the restricted
mobility o f the polymer chains close to the silicate layers and the high aspect ratio
o f the separated layers (greater than 1000).
1
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
The discovery o f carbon nanotubes has received much attention due to its high
modulus and electrical/thermal conductivity [5]. Significant effort has been made
to apply carbon nanotubes in polymeric composite applications to improve
strength and to manipulate electrical properties [6-8]. Many potential applications
have been proposed for carbon nanotubes, including conductive and high strength
composites, field emission displays, and nanometer-sized semiconductor devices
[9].
The presence o f nanoparticles provides improvements in other properties as well,
such as erosion resistance, wear resistance, fire resistance, hardness, and
environmental resistance [10].
A judicious selection o f nanoparticles can
provide additional functionalities such as electrical, magnetic, optical, etc. The
major advantage o f using functional particles is that the resulting composite can
be made multifunctional. For example, magnetic nanocomposites can be made
with magnetic nanoparticles [11],
One application that can benefit from magnetic nanocomposites is a microwave
absorber. A typical microwave absorber consists o f micron size magnetic filler
materials in a polymer matrix. To get an optimum absorption, the particles are
loaded into the polymer up to 40% to 60% in volume. This high volume loading
2
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F urth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
o f magnetic particles poses two major problems: weight because o f high density
o f magnetic materials, and durability since high loading o f the particles negate the
high strain characteristic o f the polymer. This is especially true when it is applied
in military applications such as fighter jets where every ounce o f weight is critical
to its performance and any frequent maintenance or repair on materials is very
costly.
A thinner thus lighter and more durable absorber is needed to overcome
this disadvantage.
Also it is beneficial to have a broadband characteristic to
eliminate extra absorbers to cover different frequencies. However, the design o f
microwave absorbers is limited by the inherent material properties [12].
Recent studies have shown some promises for using magnetic nanocomposites as
a microwave absorber [13, 14].
Improved physical properties o f magnetic
nanoparticles may enable a thinner absorber to achieve the same performance as a
thicker absorber made o f traditional filler materials.
To understand how this is
possible, a review is provided on how an electromagnetic wave propagates in
magnetic materials.
1.1 Fundamentals of Electromagnetic Waves [15]
When an electromagnetic wave propagates and hits a target that is conductive, the
surface electrons become excited, creating surface currents. The surface currents
3
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
carry electromagnetic energy that is momentarily trapped on the target’s surface.
The energy will eventually be either absorbed by the target or re-radiated off.
This mechanism is exactly how radar operates. It sends out an electromagnetic
wave beam from a transmitting antenna and detects the re-radiated wave that is
bounced off the target.
Electromagnetic waves in a free space or a dielectric medium can be analyzed by
solving Maxwell’s equations. M axwell’s equations relate four fundamental field
quantities and four fundamental response quantities. They are governed by a set
o f four equations
V xE = -
V x H = J + ——,
at
V -D = p
with
E = Electric field intensity (V/m)
D = Displacement flux (C/m2)
B = Magnetic induction flux (T or Wb/m2)
H = magnetic field intensity (A/m)
p = electric charge density (C/m3)
4
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
(i)
J = electric current density (A/m2)
Solution o f any electromagnetic problems require the following electromagnetic
constitutive equations:
D = £re0E ,
B = ^ r^ 0H ,
J = <rE
(2)
where £o = 8.854 x 10'12 F/m, //0= 0.4 n x 10'6 H/m are permittivity and
permeability o f air, £, and
are complex relative permittivity and permeability,
and a is the conductivity o f the material with the unit o f S/m. Note that the
material is assumed isotropic. The relative complex pemittivity and permeability
may be written as
e t = — = e T'+ ie "
(3)
Mr= — = Mr'+iMr"
(4)
Mo
The relative values are commonly used in electromagnetic analyses as these are
what we usually measure.
Equations (1) can be rearranged to yield the following electromagnetic wave
equations
5
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
^
„
-
1 d2E
VxVxEi —
—
-
= - p t u
VxVxH + 4 - ^
c dt
= V xJ
c2 at2
dJ
—
0 at
(5)
(6)
where c is the speed o f light
c=—
SJU
(7)
For a time harmonic field expressed in complex notation as E(r,t) = E(r)e'" , the
wave equation for E reduces to
V 2E = jcoco/M - co2fis E
(8)
where cj is radian frequency. In free space or in dielectric medium where 0=0, the
plance wave solution to Equation 8 becomes
E (r ,t) = E oe_j(£ft_‘at)
(9)
/V
where k is the direction vector o f the wave propagation with wavelength
X = 2 tic / o) . The corresponding solution for the magnetic field is obtained as
6
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
H (r)= — - k x E (r)
(10)
COfJ.
with k = (w2fis)1/2.
The magnitude ratio o f E to H is the wave impedance Z and
is given by
E _ coju
=Z
-
H
k
( 11)
This suggests that the impedance depends on the permeability and permittivity o f
the medium that the wave propagates in. For the normalized impedance where the
wave impedance is divided by the impedance o f air, Z0, Equation 11 turns into
(12)
In summary, the electromagnetic wave equation is derived from M axwell’s
equations and shows that an electromagnetic wave has both electric and magnetic
field components that are perpendicular to each other oscillating in time and space
in phase.
The ratio o f E to H is the wave impedance determined by the
permeability and permittivity o f the medium. Figure 1.1 illustrates characteristics
o f an electromagnetic plane wave.
7
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 1.1. Electromagnetic Wave [16]
1.2 Reflection Coefficient [15]
When a wave impinges on the surface o f an object which is semi-infinite, part o f
the wave is reflected and part is transmitted. The reflection coefficient from the
part depends not only on the material properties but also on the propagation
frequency. The reflection coefficient, R, for normal incidence can be calculated
by using the normalized impedance Z/Zo,
Z /Z o -1
R = - V - 2-----
(13)
Z/Z„ + l
R is a complex number since the normalized impedance is a function of relative
permittivity and permeability. When discussing the reflection coefficient, it is
8
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
customary to ignore the phase angle and refer only to the amplitude o f R in
decibels. Thus the reflection coefficient can be written as
|R|(dB) = 201ogw|R|
(14)
Wave impedance provides us insight into how material properties, jur and £r, can
affect the reflection coefficient.
If the material is a very good conductor, Z
approaches 0 due to high £” (= o/a>£o). In this case, Equation 13 tells us that R =
-1, meaning the wave is entirely reflected with a phase change o f 180°.
Now if
the material has an electromagnetic property o f fiT =£r, Equation 12 suggests Z =
Zo, and there will be no reflection from the material (R =0).
For a flat metallic surface coated with a layer o f dielectric material, the
normalized impedance is modified as
z / Z o
=
t a n h ( - i k 0d ^ ///r • s r )
(15)
where d is the thickness o f the layer and k0 = 27tAo. Now substituting Equation 15
into Equation (10) we get the reflection coefficient R,
9
R e p r o d u c e d w ith p e r m is s io n o f th e c o p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
r/g, tanh(-jk0d^jjur ■sr) - 1
p _
(16)
V /v X tanh(-ik0d^jUr ■sT) +1
It is seen that the reflection coefficient depends on two material properties,
relative permeability and permittivity, and a design parameter, the layer thickness.
As fir and £r change with frequencies, the design o f an absorber involves
optimizing microwave absorption at a target frequency using its electromagnetic
properties and thickness. Figure 1.2 shows reflection coefficients with different
thicknesses for an artificial material with £’ = 12, £” = 1, fi’ = 2 and ju” = 1.
0
2
4
6
8
10
12
14
16
18
-5 -
-J
-1 5 -
—
t —1.0 mm
• — t = 1.5 mm
- - t = 2.0 mm
-25s = 1 2 s ' + 1s'
-3 0 -
n= 2u'+ V
Frequency (GHz)
Figure 1.2. Effect o f thickness o f microwave absorber on Reflection Coefficient
10
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
This illustrates a general rule o f thumb in microwave absorbers: a thicker absorber
is needed to absorb effectively at low frequencies than at high frequencies.
1.3 Loss Mechanisms of Microwave Absorbers
Typical microwave absorbers are produced by adding fillers into a polymer,
altering dielectric and magnetic properties o f the resulting composite. Common
fillers used to alter magnetic properties are ferrites, iron, and cobalt-nickel alloys.
Carbon, graphite, and metal flakes are used to alter conductive properties.
Microwave absorbing materials utilize the loss mechanisms within them to absorb
energy from the electromagnetic wave passing through them.
For magnetic materials, the energy loss is associated with the relaxation
mechanisms which control the magnetization motion during flux reversal [17].
Whenever there is a change o f magnetization motion due to a dynamic external
field applied, it is accompanied with energy loss. The loss o f energy comes from
two major sources: a domain wall movement and ferromagnetic resonance.
A ferromagnetic material consists o f many regions, called domains, within which
the magnetization is constant in magnitude and direction. However, the magnitude
11
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
and direction o f magnetization varies from domain to domain in such a way that
the net magnetization vanishes. When an external magnetic field is applied, these
domains change their volume through the displacement o f the boundary walls
between adjacent domains [18]. The transition o f walls depends on the magnetic
anisotropy energy, which aligns the magnetic moments to preferred directions,
and the exchange energy, which is the interaction energy between atomic
moments.
The competition between these two energies provides the finite
thickness o f the wall.
The transition o f a domain wall is associated with an
energy loss as its movement dissipates energy to the crystal lattice o f the material.
Another form o f energy loss comes from ferromagnetic resonance which involves
the precession o f magnetic moments. When a magnetic material is disturbed from
the equilibrium state, the magnetic moments undergo precession as a result o f the
torque exerted on the magnetization by the external magnetic field.
When a
sinusoidal magnetic field is applied, it excites precessional motion. At a frequency
near precessional frequency o f the magnetic material, the energy coupled from the
excitation field to the precessing magnetization will be large and the energy is
absorbed from the magnetic field to the crystal lattice.
For ferrites or
ferromagnetic materials, the precession frequency falls under the regime o f
12
R e p r o d u c e d w ith p e r m issio n o f th e c o p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
microwave frequencies. This makes these materials an excellent choice for
microwave absorbers.
The resonance losses are characterized by a absorption curve, a plot o f radio
frequency (rf) energy absorption vs. applied static field. The difference between
the field values at which the energy absorption is one-half the maximum
absorption is defined as the linewidth, A H [17]. The linewidth is the measure o f
the rate at which energy is transferred from the system o f precessing magnetic
moments to the crystal lattice.
In conjunction with the linewidth, the
dimensionless Gilbert damping constant, or, which is a function o f the linewidth as
shown in Equation (17), characterizes the loss:
/ • AH
a =—
r
,
(17)
2(0
where y = 2.8 MHz/Oe (gyromagnetic ratio). This phenomenological constant
governs the rate o f return o f the excited precession system to the equilibrium
state. A strong damping o f the precessional motion o f the magnetic moment leads
to a broad spectral line. The linewidth depends on volume and surface divergence
o f the magnetization, impurities, imperfections, internal strains, grain boundaries,
and particle interactions [19].
13
R e p r o d u c e d w ith p e r m is s io n o f th e c o p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
In macroscopic sense, the magnetic loss o f the absorber is measured by complex
term o f the relative permeability, /f”. This stems from the fact that the energy
absorption curve is proportional to the imaginary part o f transverse susceptibility
o f magnetic materials [17]. The susceptibility and the permeability are related to
each other by
/U = % + 1
(18)
Thus, the higher the /f”, the better the performance o f magnetic absorbing
composites can be achieved. Figure 1.3 shows the effect o f the imaginary part o f
the permeability, /f”, on the reflection loss. Again two artificial materials with £’
= 12, £” = 1, /f’ = 3, /f” =2 and £’ = 12, £” = 1 ,/i’ = 3 and/i” =5 are compared. As
/i” is increased the same absorbing performance is matched with the reduced
thickness.
14
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
0
2
4
6
8
10
12
14
16
18
— t = 1.5 mm, b'=12, e"=1, h'=3, h"=2
1 = 1.0 mm, e'=12, s"=1, h'=3, h"=5
-5 -
T
-10-
C -1 5 -
-
20-
-2 5 -
-30
Frequency (GHz)
Figure 1.3. Effect o f / f ’ on RL and thickness o f microwave absorber
As shown in Equation 12, the material with similar £r and fiT will minimize the
impedance mismatch.
Magnetic materials such as ferrites or ferromagnetic
particles are used as filler since these materials offer significantly better intrinsic
fir and £r in this regards. However, fir varies with frequencies, and at the
microwave frequencies, fiT
impedance mismatch.
is substantially lower than £r, increasing the
Recognizing that permeability plays a key role in
absorption, there have been numerous studies conducted on the effect o f different
characteristics o f magnetic particles on microwave properties.
15
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
1.4 The Effect of Particle Characteristics on Microwave Properties
The complex permittivity and permeability may be calculated from measured
reflection coefficients using Nicholson - Ross algorithm [20].
The calculated
properties give the effective permittivity (£eff) and the permeability (jueff) o f the
composites.
The Maxwell Garnet theory predicts £cff and fies o f the composite in terms o f the
known intrinsic electromagnetic properties o f the metallic filler (£f and fif), the
matrix material (£m and fim), and the volume fraction o f the fillers [21].
Two
assumptions were made when the model was derived; the filler has the shape o f
sphere and no contact is made between the fillers. This model predicts the
properties relatively well when the volume fraction is low but starts to deviate
from the experimental value as the volume fraction increases. This is the result o f
the metallic fillers starting to touch one another as the volume fraction increases,
thus violating the assumptions made in the model. A Bruggeman approach, also
called effective medium theory, addresses the above problem by assuming that
both £f and £m are embedded in the effective medium and imposing self
consistency, a condition o f zero average polarizability [22]. This model is further
refined by including the effect o f eddy current on the magnetic dipole term to the
16
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F urth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
model [23].
One can use this model to calculate the intrinsic permeability Hi by
inversely using the model with measured jUeff- This is helpful since measurement
o f intrinsic properties o f ferromagnetic fillers is extremely difficult. W ith known
intrinsic properties, an engineer can optimize the design o f absorbers utilizing the
model.
However, this holds true only if the permeability is a true intrinsic
property, independent o f particle characteristic.
Numerous studies were conducted to calculate the intrinsic permeability o f
magnetic materials. Berthault et al calculated Hf o f Permalloy (NigoFe2 o) using
modified Bruggeman’s model [24], The sample was prepared by varying particle
size from 4 jUm to 16 /urn and volume fraction from 33% to 57%. The result
showed nearly constant intrinsic permeability over different particle size and
volume. This study suggests that the permeability o f the particle is an intrinsic
property that is independent o f its size and concentration.
Intrinsic permeability jUf o f carbonyl iron spheres (d = 3 -5 /mi), cobalt spheres (d
= 2 j«m), permalloy spheres (d = 9 jum), and non-spherical atom ized iron (d = 7.5
//m) were also calculated using this model [25], The effective permeability for
each particle was measured from 130 MHz to 20 GHz with different volume
17
R e p r o d u c e d w ith p e r m is s io n o f th e c o p y rig h t o w n e r . F urth er r ep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
loadings. The derived [if showed that carbonyl iron performed the best in terms o f
having high fi” and the frequency dependence, over cobalt and permalloy, two
characteristic that are the elements o f good microwave absorption material. When
the derived intrinsic permeabilities were compared between spherical carbonyl
iron and non-spherical atomized iron, the broadening o f gf” over the frequency
was revealed for non-spherical atomized iron.
Authors could not explain this
behavior, but suggested that the particle shape may have contributed to this effect.
The broadening o f [ i f is plausible since the model is based on the assumption o f
fillers being spherical. Any deviation from the spherical shape would change the
prediction to some degree.
Over the past decade or so, synthesis o f fine magnetic particles started to emerge
from research laboratories and through private enterprises. Availability o f fine
particles led many researchers to investigate the effect o f the size o f fillers on
microwave properties.
Microwave characterization o f spherical submicron Co 2 oNigo particles revealed
very interesting phenomena that had not been reported previously [26].
The
particles are synthesized through the polyol process which the liquid polyol acts
both as a solvent o f the metallic precursors and as a mild reducing medium.
18
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F urth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Again, a modified Bruggeman model was used to calculate the intrinsic
permeability o f the particle with two different sizes; 0.3 pm and 1.5 pm. Particles
were dispersed in epoxy with volume loading o f 30 % to 50 %.
The
predicted intrinsic permeability o f submicron-size particle exhibits a unique
behavior with three resonance bands while micron-sized particle had a single
resonance band.
Previously intrinsic permeability was thought as a material
property that is independent o f its size.
This study suggests for the first time
that there is a finite size effect on the permeability o f magnetic materials when
subjected to microwaves.
Another investigation on the electromagnetic properties o f nano-sized magnetic
particles revealed an extraordinary behavior [27]. In this study nano-sized (-25
nm) and micron-sized (5 pm) Fe2 0 3 were used along with non-magnetic particle
ZnO and epoxy as a binder. The volume loading o f Fe20 3 was varied from 0 to
58%. The effective permeability and permittivity were measured over a range o f
30 MHz ~ 14 GHz. The result showed the real part o f permittivity, £eff’, o f nano­
sized particles slowly decreases while micron-sized particles maintain constant
£cff’ over the frequency range.
The larger £eff” was reported for nano-sized
particles and the author attributed this to the increase o f the density o f the
interface as surface to volume ratio increases in magnetic nanostructures. For the
19
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
effective permeability, nano-sized particle added composites showed significantly
different behavior compared to micron-sized particle added composites.
g eff
increased as volume loading o f particles increased. At a maximum volume
loading o f
Fe 2 C>3 (~50% by vol.), / W is increased by a factor o f 3 and /teff” is
increased by a factor o f 20.
These findings suggest that the continuum apporach used to explain the
electromagnetic behavior o f absorber composites may not be adequate when the
size o f the magnetic filler reaches nano-scale.
As the magnetic particle size
reduces to a nano-scale, the characteristics o f a metallic particle such as
crystalline structures, magnetic anisotropy, interaction among particles, and
surface chemistry may lead to different properties than the bulk material, thus,
influencing microwave properties o f magnetic nanocomposites.
Therefore, to
utilize magnetic nanocomposites as microwave absorber, it is essential to
understand the size effect on magnetic properties as the particle size decreases to
nanometer.
1.5 Characteristic of Magnetic Nanoparticles
As mentioned in the previous section, a magnetic material consists o f grains with
many domains. As the grain size decreases, a critical size will be reached where
20
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
the grain can no longer accommodate domain walls and hence only consists o f a
single domain [18]. This critical size depends on the saturation magnetization,
anisotropy energy, and exchange energy interaction between individual magnetic
moments.
For a spherical particle, the dimension can be anywhere between 10
nm and 800 ran, depending on the material composition [28]. As the particle
reduces nearly to a single domain or below, a fundamental question arises: do the
properties o f nanosize magnetic particles differ from the bulk properties? Many
researchers predicted and experimentally confirmed that there are indeed size
effects o f magnetic particles on properties, mainly due to particle interactions and
surface effects.
The
most
studied
finite-size
effect
in
small
particle
systems
is
superparamagnetism which occurs when the size o f particle falls below a critical
size [28]. Below this size, the thermal energy (kBT) overcomes the anisotropic
energy, and the thermal excitation induces spontaneous fluctuation o f the
magnetic moment. For superparamagnetic particles, the net magnetic moment in
zero field and T > OK will average to zero. This is analogous to paramagnetism
except that the nanoparticle has a net magnetic moment o f a single domain
containing about 105 atoms instead o f a net magnetic moment by a single atom,
and thus has much higher susceptibility than paramagnetism. The dynamics o f
21
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
magnetization o f superparamagnetic particles are governed by the thermal
activation over individual anisotropy energy barrier o f each particle when the
magnetic interaction is negligible. In the case where the magnetic interaction
cannot be ignored, a new dynamic model needs to be proposed. When a high
volume loading o f particles is required, strong exchange interactions, will be
present through the surfaces o f the particles. To study this phenomenon much
theoretical and experimental work has been conducted [29-31].
These studies reveal that one o f the properties affected by particle interactions,
which is relevant to microwave absorber applications, is the Gilbert damping
constant a .
The damping constant is a phenomenological constant deduced
from fitting the model for the relaxation time o f the magnetic moment o f particles.
Various studies show that a varies from 0.035 to 1 for different magnetic particle
embedded composites [29-31].
Compared to these values, a for a bulk Fe is
about 0.01 [32], The observed increase o f a is hypothesized to be caused by the
rearrangement o f spin direction at the surface during the rotation o f process, and
the introduction o f additional irregularities in the volume, owing to their variation
from particle interactions [30]. This increase is particularly useful in microwave
absorber design since this enhances the broadband performance if the magnitude
o f the reflection loss is still acceptable.
22
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Another hypothesis for the increase o f or is due to the surface effect,
or is a
phenomenological constant in the Gilbert equation for the entire particle assembly
including all defects; in particular, the surface defects.
Thus, the smaller the
particle, the effect o f the surface defects will be more pronounced and will
increase the resulting or [31].
Magnetization and magnetic anisotropy are other properties affected by the
surface effects o f nanoparticles.
The theoretical and experimental work by
various researchers showed that the saturation magnetization and the anisotropy
constant o f nanoparticles are different from those o f the corresponding bulk
materials. A theory has been suggested that the decreasing dimensionality can
enhance magnetic moments [33].
The increased
saturation magnetization
observed is explained by the increase in the ratio o f the surface atoms over the
volume atoms [34]. For instance, about 50% o f all atoms in a particle with 2 nm
diameter are on the surface. The magnetic moments o f Fe, Co, and Ni atoms in
clusters containing a few tens o f atoms were deduced to be 10% ~ 50% larger
than the atomic moments found in the corresponding bulk materials [35, 36].
However, when an oxide shell is present, the saturation magnetization is observed
to decrease [37].
The theory o f spin canting at the surface caused by
23
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
antiferromagnetic interaction between the shell and the core is used to explain this
phenomenon [38]. This explanation is well supported by the fact that no spin
canting is observed in pure ferromagnetic particles [28].
Another key contributor to the surface effects is surface anisotropy. As the size
gets smaller, the surface anisotropy contributes significantly to the overall
anisotropy. For a spherical particle, the following phenomenological model was
developed to deduce the surface anisotropy from the experimental data [39]:
K <r= K B+ ^ K ,
(19,
where d is the diameter o f the particle, KB is a bulk surface anisotropy energy per
unit volume and Ks is a surface density o f anisotropy energy.
The surface
anisotropy should be averaged to zero based on symmetry for a spherical shape,
but this is not true for a particle with a few atomic layers.
For a 1.8-nm Co
particle, the total anisotropy is one order o f magnitude higher than the bulk
anisotropy [40]. However, the contribution o f surface anisotropy to the effective
anisotropy becomes less significant as the particle size increases above 10 nm.
24
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
The surface treatment o f nanoparticles may have some effect on their magnetic
properties [41].
Ferrite particles (Fe3 0 4 , NiFe 2 C>4 , and CoFe 2 0 4 ) 10 nm in
diameter were coated with a surfactant (oleric acid and high molecular weight
modified polybutene).
The coating produced extremely high local anisotropy
field which prevented saturation o f magnetization even at a high magnetic field,
thus significantly reducing the saturation magnetization.
Surface impurities also affect the properties as the size decreases. Impurities may
arise from the synthesis o f nanoparticles. To investigate these influences, a metal
particle with well-defined morphology and texture is studied [42], Fine particles
with different composition o f Fe, Co, and Ni are produced through the polyol
process. A thin magnesium oxide layer with the thickness o f a few nanometers
was coated on the metallic particle through a chemical process.
Density, impurity contents (C, O, and H) and saturation magnetization were
measured with to respect to the size o f particles for different composition o f Fe,
Co, and Ni. The data revealed the linear relationship between the particle size and
the above three measurement. Based on this relationship with an assumption that
the oxide layer is much smaller than the radius o f the particle, and the
25
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
magnetization o f the oxide layer is zero, a simple model is developed to predict
the core properties.
The analysis showed that the value o f saturation magnetization o f core particle
with Co and Ni composition is less than 2% lower than the bulk value. However,
a particle composition with Fe, more than 15% knockdown on the saturation
magnetization is observed. This is mainly due to the polyol process which may
contribute the impurities o f non-metallic phase within the core. In fact when an
iron nanoparticle was synthesized by an evaporation-deposition technique, the
saturation magnetization was identical to the bulk one [43].
In summary, the fmite-size effects dominate the magnetic properties o f the
magnetic nanoparticles. Table 1.1 recaps the experimental values from previous
researches conducted on magnetic nanoparticles.
26
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Material
Composition
S ize
(nm)
Bulk
2
S h ap e
Matrix
Particle
Loading
(wt%)
Synthesis
Method
NA
None
NA
CVD
7.6
55
0.5
1
10
Chemical
Reduction
Microwave
Plasm a/ Melt
Blend
NA
Reaction Milling
NA
Microwave
Plasm a/
Polystyrene
Coated
80 [45]
PVD/Mg Coated
2 0 0 [43]
Ms
(emu/g)
I 220[30]
.
5
24
Sphere
PMMA
20
Sphere
PMMA
8
NA
ZnO
22 6 [34]
184 [441
8 4 [441
0.091 [471
0.411 [471
11.5 [471
K (J/m3)
a
0 .5x10® [39]
0.01 [30]
3x10® [391
1 [29]
Fe
1 5 -2 0
Sphere
NA
2 .5 -4 0
42/58
27.5
28/72
21.4
F e/F e203 30/70
None
1 0 .3 3 -0 .4 5 1
[191
I
150 [43]
135 [43]
PV D/Passivated
with Oxygen
NA
11.3
120 [43]
(%)
17/83
10
7/93
8
0/100
7 -1 0
91 [43]
Sphere,
Chain
L 65 r43’
NA
NA
I 1 6 6 -1 7 5
I
I401
Bulk
Co
F eC o
1 .8 4.4
Sphere
20
20
None
None
NA
NA
NA
NA
Hexaferrite
75
Microemulsion
/cationic
surfactants
Arc-discharqe
Chemical
Reduction
20/80
Co/Ni
50/50
2.7x10® [40]
3 x 1 07[40]
I 0.5(311
169.7 [46]
76.7 [42]
42
Sphere
None
NA
(%)
Polyol/Mg coated 109.2 [42]
80/20
Ni
'.1-0 .6
[30]
CVD
138.1 [42]
Bulk
9
85
Sphere/
Chain
None
NA
PVD/Passivated
with Oxvaen
54 [48]
22.5 [481
43.5 [481
Table 1.1. Magnetic properties o f various magnetic nanoparticles
27
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
1
What impacts from these changes in properties is the susceptibility o f the
magnetic materials. M any studies have been conducted on the susceptibility o f
magnetic nanoparticles due to longitudinal and transversal relaxation o f
magnetization [30, 31, 49].
Since microwave absorbers operate without an
external magnetic field, and only transverse susceptibility is relevant in the
microwave, a simplified model can be used to express the susceptibility o f
magnetic nanoparticles [50]:
where vp = particle volume fraction; ju0 = permeability in vacuum; M s =
saturation magnetization; K = effective anisotropy constant; a> = frequency o f
incident wave; a>0 = y 2 K / M s = resonance frequency; y = gyromagnetic ratio;
R1, R2, R}, R 4 = functions o f damping parameter and a - K V p /(kBT ) ; V =
particle volume; k B = Boltzman constant; and T = temperature. As seen in this
model, saturation magnetization, anisotropy constant, and the damping constant
play a major role on permeability as it is related to the susceptibility, shown in
Equation (15). Figure 1.4 plots the effect o f damping coefficient and anisotropy
ratio on susceptibility assuming Ms and K are constant. As seen in the plot, the
28
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
magnitude, shift o f peak, and the linewidth are quite sensitive to these parameters.
The changes in properties as shown in Table 1 thus affect susceptibility o f
magnetic particles, and this has a major impact on microwave absorber design
since the reflection loss is a function o f the permeability. This suggests that the
properties o f magnetic nanoparticles can be utilized to enhance the performance
o f microwave absorbers. Indeed, the simulated result o f reflection loss, using the
above model with nanomagnetic properties, showed enhanced performance o f the
microwave absorbers [13]. Using the properties o f nano Fe, a good absorbing
performance is achieved compared to the absorber using traditional ferrites even
with the low loading, 10% by volume.
29
R e p r o d u c e d w ith p e r m is s io n o f th e c o p y r ig h t o w n e r . F urth er r ep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
H /M 2 x
%\
0= 10, a=0.1
x": a=10, a=0.1
x': ct—10, a=0.3
x": o=10, a=0.3
-2 -
-3 -
H /M 2x
x’: o=5, a=0.2
x": o=5, a=0.2
X- 0=50, a=0.2
x": o=50, a=0.2
co/co
Figure 1.4. Effect damping coefficient and anisotropy ratio on susceptibility
1.6 Motivation For Magnetic Nanocomposites As Microwave Absorbers
The quest for a broadband, light weight and durable microwave absorber is
already mentioned in previous section. However, limitations o f current
microwave absorbers are well documented [12], To overcome this, a material
with superior magnetic properties compared to the existing material must be used.
The previous research on magnetic properties o f nanoparticles offers a hope to
design a better microwave absorber in this regard. Besides taking advantage o f
30
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
the intrinsic properties o f magnetic nanoparticles, there is another side benefit o f
using magnetic nanoparticles as fillers that cannot be ignored. Since the metallic
magnetic materials are conductive, the permeability decreases at high frequency
due to eddy current loss induced by electromagnetic waves [51].
However, the
eddy current loss can be suppressed if the particle size is below the skin depth. In
microwave frequency, the skin depth is around 1 gm [52].
This enables the
electromagnetic wave to fully penetrate the particle and allows the contribution o f
a whole particle to the magnetic properties o f the microwave absorber.
From the results discussed on previous chapters and the inherent benefit o f using
magnetic nanoparticles, a magnetic nanocomposite can advance the engineering
o f microwave absorption with the right fabrication processes and judicious
selection o f magnetic nanoparticles.
This dissertation focuses on those two
aspects to attain an enhanced microwave absorber.
Processing o f magnetic
nanocomposites, and microwave characterizations o f magnetic nanocomposites
are studied.
31
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Reference
1. Nanoscience and Nanotechnologies, July 2004, Royals Society & the Royal
Academy o f Engineering
2. R.A. Vaia and E.P. Giannelis,
“Polymer Nanocomposites: Status and
Opportunities, ” MRS Bulletin, Vol. 26, 2001, pp. 392-401.
3. J. Gilman., “Flammability and Thermal Stability Studies o f Polymer LayeredSilicate Nanocomposites, "A pplied Clay Science, Vol. 15, 1999, pp.31-49.
4. X. Komman, L.A. Berglund., and H. Lindberg, “Stiffness Improvements and
Molecular Mobility in Epoxy-Clay Nanocomposites, ” Mat. Res. Soc. Proc.,
Vol. 628, 2000, pp. CC11.8.1-CC11.8.7.
5. S. Iijima, “Helical microtubles o f graphite carbon,” Nature, Vol. 354, 1991,
pp. 56
6. T.W. Ebbesen, “Electrical conductivity o f individual carbon nanotubes,”
Nature, Vol. 382, 1996, pp.54-56.
7. C.
Bower,
“Deformation
o f carbon
nanotubes
in
nanotube-polymer
composites,” Applied Physics Letters, Vol. 74, 1999, pp. 3317-3319.
8. K. Lozano and E.V. Barrera, “Nanofiber-reinforced thermoplastic composites.
Thermo analytical and mechanical analysis,” Journal o f Applied Polymer
Science, Vol. 79, 2001, pp.125-133.
32
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
9. KT, Lau and D. Hui, “The revolutionary creation o f new advanced materialscarbon nanotube composites,” Composites Part B, Vol. 33, 2002, pp.263.
10. J. Gilman., “Flammability and Thermal Stability Studies o f Polymer LayeredSilicate Nanocomposites,” Applied Clay Science, Vol. 15, 1999, pp.31-49.
11. L. Chen, W.-J. Yang, and C.-Z. Yang, “Preparation o f Nanoscale Iron and
Fes 0 4 Powders in a Polymer M atrix,” Journal o f Materials Science, Vol. 32,
1997, pp. 3571-3575.
12. J.L. Wallace, “Broadband magnetic microwave absorbers: Fundamental
limitations,” IEEE Transactions on Magnetics, Vol. 29, 1993, pp.4209-4214.
13. V. Bregar, “Advantages o f ferromagnetic nanoparticle composites in
mircrowave absorbers,” IEEE Transactions on Magnetics, Vol. 40, 2004,
pp. 1679-1684.
14. J. R. Liu, M. Itoh, T. Elorikawa, E. Taguchi, H. Mori, and K. Machida, “Iron
based carbon Nanocomposites fo r electromagnetic wave absorber with wide
bandwidth in GHz ra n g e f Applied Physics A, Vol. 82, 2006, pp. 509-513.
15. E. F. Knott, J. F. Shaffer, M. T. Tully, Radar Cross Section, 2nd ed., SciTech
Publishing, New York, 2004.
16. From website o f ARC Technologies, Inc.
17. C. E. Patton, “Dynamic processes in magnetic thin film s domain wall motion
and ferromagnetic resonance,” Ph.D Thesis, 1967, Cal Tech.
33
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
18. A. Aharoni, Introduction to the theory o f ferromagnetism, 2nd ed., Oxford
University Press, New York, 2000.
19. M. Pardavi-Horvath and L. J. Swatzendruber, “Ferromagnetic resonance
relaxation in reaction milled and annealed Fe-Zn Nano composites,” IEEE
Transactions on Magnetics, Vol 35, 1999, pp.3502-3504.
20. A. M. Nicholson and G. F. Ross, “Measurement o f the intrinsic properties o f
materials by time domain techniques,” IEEE Transaction on Instrumentation
and Measurement, Vol. IM-19, 1970, pp.377-382.
21. J.C. Maxwell-Garnet, Philosophical Transactions o f Royal Society London,
Vol. 203, 1904, pp. 203
22. D. A. G. Bruggeman, Ann. Physics (Leipzig), Vol. 24, 1935, pp.636
23. G. D. Mahan, uLong-wavelength absorption o f cermets,” Physical Review B,
Solid State, Vol. 38, 1988, pp.9500-9502.
24. Berthault, A., Rousselle, D., and Zerah, G., “Magnetic properties o f
Permalloy microparticles,” Journal o f Magnetism and Magnetic Materials,
Vol. 112, 1992, pp. 477-480.
25. Olmedo, L., Chateau,
G., Deleuze, C., Forvielle, J.L., '‘‘M icrowave
characterization and modelization o f magnetic granular materials,” Journal
o f Applied Physics Vol.73, 1993, pp. 6998-6994.
34
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
26. Viau,
G., Ravel, F.,
and Acher,
O., “Preparation and microwave
characterization o f spherical and monodisperse Co2oNiso particles,” Journal
o f Applied Physics, Vol.76, 1994, pp. 6570-6572.
27. Talbot, P., Konn, A.M., and Brosseau, C., “Electromagnetic characterization
o f fine-scale particulate composite materials,” Journal o f Magnetism and
Magnetic Materials, Vol. 249, 2002, pp. 481-485.
28. X. Batle and A. Labarta, uFinite-size effects in fin e particles: magnetic and
transport properties,” Journal o f Physics D: Applied Physics, Vol. 35, 2002,
pp.R15-R39
29. J. L. Dormann, L. Bessias, and D. Fiorani, “A dynamic study o f small
interacting particles: superparamagnetic model and spin-glass laws,” Journal
o f Physics C: Solid State Physics, Vol. 21, 1988, pp.2015-2034.
30. J. L. Dormann, F. D ’Orazio, F. Lucari, E. Troc, P. Prene, J. P. Jolivet, D
Fiorani, R. Cherkaoui, and M. Nogues, “Thermal variation o f the relaxation
time o f the magnetic moment o f y-Fe 2 0 3 nanoparticles with interparticle
interactions o f various strength,” Physical Review B, Vol. 53, 1996, pp.
14291-14297.
31. W. T. Coffey, D. S. F. Crothers, J. L. Dormann, Y. P. Kalmykov, and W.
Wemsdorfer, “ Thermally activated relaxation time o f a single domain
ferromagnetic particle subjected to a uniform fie ld at an oblique angle to the
35
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
easy axis: Comparison with experimental observation,” Physical Review
Letters, Vol. 80, 1998, pp. 5655-5658.
32. L. BeSsias, L. B. Jeffel, and J. L. Dormann, ‘'‘'Relaxation time o f fine magnetic
particles in uniaxial symmetry,” Physical Review B, Vol. 45, 1992, pp. 78057815.
33. F. Liu, M. R. Press, S. N. Khahna, and P. Jena, “Magnetism and local order:
A b initio tight-binding theory,” Physical Review B, Vol.39, 1989, pp. 69146924.
34. T. Furubayashi, I. Nakatani, and N. Saegusa, “Magnetic moment and
hyperfine fie ld in colloidal fin e particles o f iron,” Journal o f the Physical
Society o f Japan, Vol. 56, 1987, pp. 1855-1858.
35. S. N. Khanna, and S. Linderoth, “Magnetic behavior o f clusters o f
ferromagnetic transition metals,” Physical Review Letters, Vol. 67, 1991, pp.
742-745.
36. I. Billas, J. Becker, A. Chatelain, and W. A. de Heer, “Magnetic moments o f
iron clusters with 25 to 700 atoms and their dependence on temperature,”
Physical Review Letters, Vol. 71, 1993, pp. 4067-4070.
37. S. Gangopadhay, G. C. Hadjipanayis, C. M. Sorensen, and K. J. Klabunde,
“Effect of particle size and suface chemistry on the interactions among fine
,35a
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
metallicparicles,” IEEE Transactions on Magnetics, Vol. 29, 1993, pp. 26192621.
38. J. M. D Coey, “Noncollinear spin arrangement in ultrafine ferrimagnetic
crystallites,” Physics Review Letter, Vol. 27, 1971, pp.l 140-1142.
39. F. Bodker, S. Morup, and S. Linderoth, “Surface effects in metallic iron
nanoparticles,” Physical Review Letters, Vol. 72, 1994, pp. 282-285.
40. J. P. Chen, C. M. Sorensen, K. J. Klabunde, and G. C. Hadjipanayis,
“Enhanced magnetization o f nanoscale colloidal cobalt particles,” Physical
Review B, Vol. 51, 1995, pp. 11527-11532.
41. A. E. Berkowitz, J. A. Lahut, and C. E. Vanburen, “Properties o f magnetic
flu id particles,” IEEE Transactions on Magnetics, Vol. 16, 1980, pp. 184-190.
42. P. Toneguzzo,, G. V iau ,, O. A rch er,, F. G uillet,, E.B runeton,, and F. FivetVincent, “CoNi and FeCoNi fin e particles prepared by the polyol process:
Physico-chemical characterization
and dynamic
magnetic properties,”
Journal o f Material Science, Vol. 35, 2000, pp. 3767-3784
43. S. Gangopadhay, G. C. Hadjipanayis, C. M. Sorensen, and K. J. Klabunde,
“Magnetism in ultrafine Fe and Co particles,” IEEE Transactions on
Magnetics, Vol. 29, 1993, pp. 2602-2607.
35b
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
44. Z. Gou, L. Henry, V. Palshin, and E. Podlaha, “Synthesis o f poly(methyl
methacrylate) stabilized colloidal zero-valence metallic nanoparticles,”
Journal o f Material Chemistry, To be published, 2006.
45. H. Srikanth, R. Hajndl, C. Chirinos, J. Sanders, A. Sampath, and T. S.
Sudarshan, “Magnetic studies o f polymer-coated Fe Nanoparticles synthesized
by microwave plasma polymerization”, Applied Physics Letters, Vol. 79,
2001, pp. 3503-3505.
46. C. Sudakar, G. N. Subbanna, and T. R. N. Kutty, “Hexaferrite-FeCo
nanocomposite particle and their electrical and magnetic properties at high
frequencies,” Journal o f Applied Physics, Vol. 94, 2003, pp. 6030-6033
47. J. L. Wilson, P. Poddar, N. A. Frey, H. Srikanth, K. Mohomed, P. Harmon, S.
Kotha, and J. Wachsmuth, “Synthesis and magnetic properties o f polym er
Nanocomposites with embedded iron nanoparticles,” Journal o f Applied
Physics, Vol. 95, 2004, pp. 1439-1443.
48. Y. Du, M. Xu, J. Wu, Y. Shi, and H. Lu, “M agnetic properties o f ultrafine
nickel particles,” Journal o f Applied Physics, Vol. 70, 1991, pp. 5903-5905.
49. V. Shilove, J. Bacri, F. Gazeau, F. Gendron, R. Perzynski, and Y. Raiker,
“Ferromagnetic resonance in ferrite nanoparticles with uniaxial surface
anisotropy,” Journal o f Applied Physics, Vol. 85, 1999, pp. 6642-6647.
35c
R e p r o d u c e d w ith p e r m is s io n o f th e c o p y rig h t o w n e r . F urth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
50. Y. Raiker and M. Shliomis, “Theory o f dispersion o f the magnetic
susceptibility o f fine ferromagnetic particles,” Soviet Physics. JETP, Vol. 40,
1975, pp. 526-532.
51. D. Rousselle, A. Berthault, O. Archer,
J. P. Bouchaud, and P. G. Zerah,
“Effective medium at finite frequency: Theory and e x p e r im e n tJournal o f
Applied Physics, Vol. 74,1993, pp. 475-479.
52. S. Sugimoto, T. Maeda, D. Book, T. Kagotani, K. Inomata, M. Homma, H.
Ota, Y. Houjou and R. Sato, “GHz microwave absorption o f a fin e a-Fe
structure produced by the disproportionation o f SmzFen in hydrogen,”
Journal o f Alloys and Compounds, Vol. 330-332, 2002, pp. 301-3.06.
35d
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Chapter 2
Processing of Iron Oxide/Epoxy Vinyl Ester Nanocomposites
The chapter describes ultrasonic dispersion o f iron oxide particles in an epoxy
vinyl ester resin. The effects o f these particles on rheology and cure behavior o f
the resulting colloidal suspensions are then discussed. The power and duration o f
sonication and the temperature o f the matrix resin were varied to obtain the
optimum dispersion.
The dispersion quality was quantified using the
sedimentation ratio and the density variation supplemented by the SEM
photomicrographs. In addition to the dispersion quality, the effect o f particle size
and particle loading on the resin curing was studied as well.
2.1 Materials
The polymer was an epoxy vinyl ester resin manufactured by The Dow Chemical
Company, with trade name DERAKANE 441-400. The liquid resin has a density
o f 1.02 g/cm3, and curing is done at room temperature by addition o f 0.2 wt% o f
cobalt naphthenate (CoNap) promoter and 2.5 wt. % o f Trigonox 239A catalyst.
Two types o f iron oxide particles were used: FesCU (magnemite, Bayer) particles
with an average diameter o f 200 nm and a density o f 4.6 g/cm3, and / -Fe 2 0 3
36
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
(maghemite, Nanophase Technologies) particles with a diameter o f 19-38 nm and
a density o f 5.2 g/cm3.
2.2 Particle Dispersion and Stability
Fe^Oa Particles
A controlled amount o f particles was mixed by hand in 80 ml o f resin.
The
mixture was then sonicated by a Branson ultrasonic horn with its tip immersed in
the mixture. The horn was operated at 20 kHz with variable power output and
variable duration. Although the dispersion quality looked good immediately after
sonication, a clear phase separation with particle sedimentation was observed 24
hours thereafter.
To quantify the dispersion stability, the height o f the supernatant layer at the top
and the total height of the suspension in the beaker were measured and used to
calculate the sedimentation ratio, which is equal to the upematant layer height
divided by the total height. Thus, the lower the sedimentation ratio, the better the
particle dispersion stability.
The change o f the 24-hr sedimentation ratio with
sonication time is shown in Figure 2.1 for a 2% particle volume content and an
ultrasonic power o f 600 W. It is seen that the best dispersion stability is achieved
at a sonication period between 8 and 15 min.
37
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
0.7
o
a
°-5
0.0
0
5
15
10
20
25
30
Time (min)
Figure 2.1. Influence o f ultrasonication time on dispersion quality o f Fe 3 C>4
particles: particle volume traction o f 2 %, ultrasonic power o f 120 W, and
sedimentation time o f 24 hr.
The resin temperature increased significantly during sonication.
This rise in
temperature induces a decrease in the resin viscosity, a decrease in the solubility of
gases present in the resin, a decrease in the surface tension, and an increase in its
saturation vapor pressure, which collectively modify the rate and intensity o f
cavitation [1], Figure 2.2 shows the effect o f cooling by ice water during sonication.
The sedimentation ratio is taken 72 hours after the sonication. It shows that the
cooling o f the mixture during sonication improves the particle dispersion quality
slightly. The optimum sonication time was around 11 minutes, which is consistent
with the previous result.
38
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
- t? —
uncooled resin
—A— cooled resin
0.0 9
0 .08
•2
^
a
|
|
0.07
n (\fy
0.04
°-°3 ■
0.02
0.01
0.00
-
-
8
9
10
11
Time (min)
12
13
Figure 2.2. Influence o f ultrasonication time and resin temperature on dispersion
quality o f Fe 3 <I) 4 particles: particle volume fraction o f 2 %, ultrasonic power o f
400 W, and sedimentation time o f 72 hr.
Figure 2.3 compares the changes o f sedimentation ratio with sonication time for
two different volume fractions o f Fe 3 C>4 particles.
When the particle volume
fraction is increased from 2 % to 4 %, the quality o f particle dispersion is slightly
reduced. However, the optimum sonication time appears to remain unchanged.
Scannining Electron Microscopy (SEM) photomigraphs o f cross-sections o f cured
resin specimens containing 4 vol% Fe 3 C>4 particles are shown in Figures 2.4a-2.4c
for sonication times o f 5, 11 and 20 min, respectively. The number and size o f
particle agglomerations observed in the figures indicate that the best particle
39
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
dispersion quality is obtained at a sonication time o f 11 min in agreement with the
sedimentation results o f Figure 2.1.
A larger magnification o f the 11-min
specimen shows a fairly good dispersion o f particles (Figure 2.5). The seemingly
high concentration o f particles revealed on the micrographs is not surprising since
the average space between the adjacent particles is only about twice their
diameter.
♦
4% Fe304
- 2% Fe304
■Series 3
0.14
0.12
0.10
2
co
0.08
a<0
0.06
T3
<D
C/5
0.04
0.02
0.00
5
15
10
20
Time (min)
Figure 2.3. Influence o f particle volume fraction on dispersion quality o f Fe 3
particles: ultrasonic power o f 400 W, cooling, and sedimentation time o f 72 hr.
40
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
0 4
Figure 2.4. SEM photomicrographs o f fracture surfaces o f cured resin specimens:
4 vol. % Fe 3 C>4 particles dispersed ultrasonically for (a) 5m in,(b)llm in, and (c)
20 min
41
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 2.5. SEM photomicrographs o f fracture surfaces o f cured resin specimens:
4 vol. % Fe 3 C>4 particles dispersed ultrasonically for 11 minutes
Fe?.Q-< Particles
Similar dispersion experiments were carried out with the smaller Fe 2 C>3 particles.
Unlike Fe3( \ sedimentation was not observed ever after 48 h after sonication.
Therefore, a density ratio was used instead o f the sedimentation ratio to quantify
the dispersion quality.
To this end, the upper 40 ml o f the suspension was
removed from the beaker and the density o f the solution remaining in the beaker
was measured. The density o f the remaining 40 ml specimens was then divided
by the density o f the initial 80 ml mixture to yield a density ratio.
The results are given in Figure 2.6 for the suspensions including lvol% and 2
vol% o f Fe2 C>3 nanoparticles 48 h after sonication with the horn delivering a
maximum power output o f 600W. For both particle volume fractions, none o f the
ultrasonic conditions is effective enough to
went sedimentation fully since the
density ratio values are always greater than unity. In both cases, the density ratio
initially decreases with increasing exposure time, suggesting that sonication
42
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
improves dispersion. However, no improvement is noticed after 2-3 minutes o f
sonication.
1% Fe203
2% Fe203
1.08
1.07
•
o
1.06
*3
5-H 105
.1?
1-04
c/5
S
Q
1.03
1.02
1.01
1.00
0
5
10
15
20
Time (min)
Figure 2.6. Influence o f ultrasonication time on dispersion quality o f Fe 2 C>3
nanoparticles in the resin.
SEM photomicrographs were taken to further understand the particle dispersion.
Since cooling during sonication seemed to improve the dispersion quality in the
earlier preliminary tests, all subsequent sonication was done in an ice bath
surrounding the beaker. The resin temperature during sonication was measured
with a thermocouple placed at the bottom o f the beaker, as shown in Figure 4.7.
Three different power levels and three different durations were then used to find
out the optimum dispersion condition. A Vibra-Cell manufactured by the Sonic
43
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
and Materials was used instead o f the Branson unltrasonic horn because its ability
to control the power level more accurately.
Sonic Horn-
T.C at the Center Bottom
Figure 2.7. Experimental Setup o f Fe203 Dispersion Study
Power levels were kept low since a high-power output increases the resin
temperature very rapidly and beyond the flash point o f the resin.
This is
illustrated in Figure 2.8 that shows the measured temperature during sonication.
The thermocouple was located one-half inch below the tip o f the horn where the
temperature should be the highest in the resin bath.
The resin temperature
44
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
increased slightly at the beginning but decreased to a lower-than-room
temperature at power level 1.
However, when the power level was increased
slightly, the resin temperature increased rapidly. The maximum temperature
reached close to 90 °C. Such a high temperature is not expected to induce curing
for the present resin as the initiator and promoter are added after sonication.
100
90
80
70
2 60
¥
3
2
4>
a
E
Q)
H
50
40
^
Power Level 1
Power Level 2
Power Level 3
I
^
30
20
10
0
0
5
10
15
20
25
30
Time (min)
Figure 2.8. Resin temperature vs. sonication time
45
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
35
Figure 2.9 a. Micro photos showing particle dispersion at power level 1.
Figure 2.9 b. Micro photos showing particle dispersion at power level 2.
46
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 2.9 c. Micro photos showing particle dispersion at power level 3.
During sonication, a small amount o f suspension was taken from the beaker after
5, 15, and 30 minutes, respectively, o f sonication. The sample was then placed on
a micro-slide and covered with a plastic cover. An optical microscope was used
to
examine
the
microphotographs.
particle
distribution.
Figures
2.9a-c
are
the
400X
Black dots in the picture are the aggregates that were not
broken during sonication. At power level 1, aggregates were still large after 5
minutes o f sonication.
At power level 2, there were fewer aggregates after 5
minutes o f sonication. Power level 3 yielded the fewest aggregates present after 5
minutes o f sonication. All three power levels showed a fairly uniform distribution
47
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
after 15 minutes o f sonication. These pictures suggest that most aggregates are
broken in the early stage o f sonication, and after 5 minutes or so dispersion
quality does not improve as rapidly. This is consistent with the density ratio study
shown in Figure 2.6. Regardless o f the power level, the dispersion quality after
30-min sonication is about the same as that after 15-min sonication. This suggests
that a long period o f sonication is needed to produce a uniform particle dispersion.
Figure 2.10 shows the actual power output at each power level.
show
All three levels
a peak output at an early stage o f sonication and leveling-off after 5
minutes. This suggests that the energy required to separate the particles is the
highest during the initial 5 minutes o f sonication. After 5 minutes the output is
uniform, suggesting no further particle dispersion is a steady process. The density
ratio, the SEM micrographs and the power output all suggest that the particle
dispersion, not surprisingly, is most effective in the early stage o f sonication.
A
higher power level generally give a better dispersion. However, a high power
output can lead to a higher temperature. Thus, it is better to use a high power
level for the initial 2-3 minutes to break up the particles and then reduce the
power for incremental dispersion at a lower temperature.
W hen a sonicated sample was left on the micro slide over a period o f time, the
particles started to re-aggregate. Figure 2.11 shows re-aggregated particles that
48
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
were sonicated 2 hours earlier.
However, the agglomerates can be broken up
easily by manual stirring as Figure 2.12 shows no agglomerates in a sample after
manual stirring. The sample did show agglomerates 48 hours after sonication.
2
0#
0
15
20
25
Time (Min)
Figure 2.10. Sonication power output Vs sonication duration
49
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
30
Figure 2.11. Reaggregation of particles after sonication
msmmmm
Figure 2.12. Manually Stirred after 48 hours past sonication
50
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Initially, nanoparticles form agglomerates because o f the van der Waals forces
between them. During sonication, agglomerates are broken up and particles are
surrounded by the polymer molecules. Although these separated particles try to
re-agglomerate because o f the van der Waals forces, they cannot form direct
contact with one another because o f the intervening polymer molecules, i.e., one
has a steric equilibrium. It is thus easy to re-disperse the particles even without
sonication.
2.3 Curing Behavior
The degree o f cure was measured for the neat resin and the composites on a Seiko
Instruments DSC 220C calorimeter.
All specimens were cured using CoNap
promoter and Trigonoxn catalyst. Samples o f approximately 15 mg were heated
in air at a rate o f 10°C/min. For comparison purposes, measurements were also
taken on specimens that were postcured for 1 hr at 85°C.
The degree o f cure a
was calculated as:
a =
( 1)
1----------— ------( l - ^ ) A H uc
51
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
where AH is the residual enthalpy o f reaction o f the specimen (in J/g), h H uc the
enthalpy o f reaction o f the initially uncured liquid resin (J/g, as determined
experimentally), and Wp the weight fraction o f particles in the resin. Results in
Table 2.1 shows that the larger Fe 3 0 4 particles have no effect on curing. However,
the smaller Fe 2 C>3 nanoparticles retards curing.
In fact, additional amounts o f
CoNap and Trigonoxa had to be used to induce cure. Even then, the attained
degree o f cure was lower. However, a postcure can still increase the degree o f
cure.
To examine the effect o f Fe2 C>3 on curing behavior in more detail, DSC
(differential scanning calorimetry) was run on resin suspensions. Particle loading
was varied from 0.5 vol% to 4 vol%. Each sample was prepared from 100 ml o f
suspension sonicated at power level 2 for 30 miniutes in an ice bath.
Trigonox
catalyst (1.25 wt%) and CoNap Promotor (0.2 wt%) were added just before the
DSC run.
As soon as the catalyst and promoter were added, the sample was
placed on the DSC calorimeter. Nitrogen environment was used to prevent any
oxidation and the temperature was ramped from 27°C to 200°C at a 5°C/min
heating rate.
52
R e p r o d u c e d w ith p e r m issio n o f th e c o p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Particle
' type
%vol. o f
particles
%wt. o f
CoNap/Trigonox
Time for
completion
o f cure (hr)
Post­
cure
Degree
o f cure
(%)
-
0.0
0.2/2.5
0.5
No
92.5
-
0.0
0.2/2.5
0.5
Yes
99.7
Fe3C>4
4.0
0.2/2.5
0.5
No
92.1
Fe 3 0 4
4.0
0.2/2.5
0.5
Yes
99.8
Fe203
4.0
0.2/2.5
>24.0
No
0.0
Fe203
4.0
0.4/5.0
>24.0
No
0.0
Fe2C>3
4.0
0.8/10.0
0.5
No
88.9
Fe203
4.0
0.8/10.0
0.5
Yes
99.8
TABLE 2.1. Influence o f iron oxide particles on curing
Figure 2.13 shows the heat flow rate versus temperature.
Table 2.2 lists the
observed peak temperatures, i.e., the temperatures where the exotherm reaches a
peak, and the total heats o f reaction. Both the peak temperature and the heat o f
reaction increase with increasing particle loading, clearly showing the delay o f
cure caused by the nanoparticles.
act as a prohibitor.
Thus the nano-size Fe203 particles appear to
It is noted again that the larger Fe 3 C>4 , whose diameter is
about 10 times larger than that o f Fe203, had no delay in curing. Preliminary
results with other materials indicate that the delayed curing is caused by the small
size o f the particles.
53
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
1.2
neat
0.50%
A
V
%
>
w
lh
2 . 00 %
z
4.00%
W
4
it
- 0.2
-0.4
Tem perat ure ( C)
Figure 2.13. Measured heat flow from different particle loading
Particle
Loading (%)
Curing Initiating
Temperature (°C)
Curing Peak
Temperature (°C)
Complete
Heat o f
Reaction
0
0.5
1
2
4
61.29
73.96
85.62
100.4
94.1
78.62
91.54
96.9
111.08
108.0
173.7
158.8
124.1
112.1
112.3
m
54
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Table 2.2. Curing Behavior o f Fe203 for different particle loading
2.4 Conclusions
Particulate composites were prepared by mixing an epoxy vinyl ester resin and
two types o f iron oxide particles: Fe 2 C>3 with an average diameter o f
approximately 26 nm and Fe 3 C>4 with an average diameter o f about 200 nm.
Although sonication was effective, it was rather difficult to achieve a completely
uniform
dispersion
o f particles.
For the
larger Fe 3 C>4 , sedimentation
measurements were used to characterize the dispersion quality. For the nano-sacle
Fe 2 C>3 , there was no visible sedimentation within a reasonable period o f time.
Therefore, the density ratio was used instead. The power output indicated that the
dispersion was most effective at the start o f sonication, and further dispersion
took much longer time.
Nevertheless, SEM micrographs still showed a few
agglomerates. Once the particles were separated by sonication, steric equilibrium
was achieved.
Fe 3 0
4
The nano-size Fe 2 0 3 particles delayed curing while the larger
did not. Preliminary results with other materials indicate that the delayed
curing is caused by the small size o f the particles. Further study is needed to
elucidate the interaction mechanisms between nanoparticles and the matrix resin.
55
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Reference
1. Dooher, R. and R. Lippaman, T. Marrone, and H. Poble. 1977. “Ultrasonic
Disintegration o f Particles,” Ultrasonic Symposium Proceedings, Phoenix, pp. 1116.
56
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Chapter 3
Flexible Magnetic Nanocomposites Fabrication through Particle Surface
Initiated Polymerization
3.1
Motivation
Magnetic nanoparticles (NPs) have attracted much interest due to their special
physicochemical properties such as enhanced magnetic moment
[1] and
coercivity [2] different from the bulk and atomic counterparts. Incorporation o f
the inorganic NPs into a polymer matrix has extended the particle applications
such as in high-sensitivity chemical gas sensors [3] due to the advantages o f
polymeric nanocomposites possessing high homogeneity, flexible processability
and tunable physicochemical properties such as improved mechanical, magnetic
and conductive properties. [4-6] High particle loading and flexibility are required
o f nanocomposite for certain applications such as photovoltaic (solar) cells,[7]
photodetectors and shape-memory devices.[8, 9]
A microwave absorber is another application that can benefit from using magnetic
nanoparticles [10-12],
A typical microwave absorber consists o f micron size
magnetic fillers in a polymer matrix. To get an optimum absorption, the particle
loading in the polymer matrix is about 40% to 60% by volume. However, the
57
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
challenge o f incorporating nanoparticles into polymer is amplified by the
difficulty o f dispersing a high loading o f nanoparticles without sacrificing the
mechanical properties o f the resulting composite.
Current high-particle-loading nanocomposites are normally fabricated by directly
mixing the particles with the polymer monomers followed by curing at a certain
temperature, or alternatively, using a solvent to disperse the particles followed by
casting to produce a thin film. Both methods result in inhomogenous dispersion
and unacceptable mechanical properties.
The major challenge is to process a
nanocomposite with uniform particle distribution and good properties.
In this
chapter a technique known as Surface Initialized Polymerization (SIP) is proposed
to overcome the above mentioned processing difficulties.
3.2
Material
The polymeric matrix used was a commercial, clear polyurethane coating
(CAAPCOAT FP-002-55X, manufactured by the CAAP Co., Inc.), which
contains two-part polyurethane monomers, i.e., 80 wt. % diisocyanate and 20
wt.% diol. The liquid resin has a density o f 0.83 g/cm3. Iron oxide (Fe2 C>3 ,
Nanophase Technologies) nanoparticles with an average diameter o f 23 nm and a
<y
specific surface area o f 32 m /g were used as nanofillers for the nanocomposite
fabrication.
Polyurethane catalyst (a liquid containing -20-65 wt.% aliphatic
58
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
amine, 1-50 wt.% parachlorobenzotrifluoride and 10-35 wt.% methyl propyl
ketone) and accelerator (polyurethane STD-102, containing 1 wt.% organotitanate and 99 wt.% acetone) were purchased from CAAP Co. Inc.
All the
chemicals were used as-received without further treatment.
O
O
II
II
C = N - R —N = C
H O -R -O H
.
diol
diisocyanate
Figure 3.1. Chemical formulats o f the two-part polyurethane monomers used.
3.3
Nanocomposite Fabrication
The nanocomposites were fabricated by two different methods. One is the direct
mixing (DM) and the other is based on the surface initiated polymerization (SIP).
The fabrication procedures are as follows. In the DM method, 7.7 g monomers,
1.03 g catalyst and 1.42 g accelarator are added into 30 ml tetrahyrofuran (THF)
solvent with ultrasonic stirring for 10 minutes. Then, 6.41 g nanoparticles are
added into the above solution and ultrasonically stirred for 10 minutes. The
suspended solution is poured out into a container for curing to obtain a nominal
particle loading o f 65%.
59
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
The SIP method starts with physicochemical adsorption o f the initiator onto the
iron-oxide (Fe2 0 3 ) nanoparticle surface in a THF solution. This will initiate the
polymerization at the surface o f particles. The surface initiated polymerization
should serve dual advantages: separation o f particles and improved mechanical
strength.
Since polyurethane polymer is chemically bonded directly onto the
particle, an improved bond strength is expected compared to the one that the
polymer just surrounding the particles. Also once polymerization is initiated from
the surface, the particles w on’t be able to agglomerate due to the polymer chain
surrounding the particles, thus ensuring a good disperstion. Figure 3.2 depicts
surface initiated polymerization.
HO
O
HO—
HO
+
H
O
II
C = N —R - N = C + H O - R - O H
diisocyanate
diol
► HO— I
polyurethane
Figure 3.2. Surface Initiated Polymerization
To achieve this, the catalyst and accelarator are added into the THF solution
containing nanoparticles and ultrasonically stirred for half an hour. The sonication
offers evaporation o f moisture that was physically adsorbed onto the particle. The
60
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
monomers are then added into the above solution drop by drop within half an hour
and the polymerization is left to continue for 6 hours. Finally, the solution is
poured into a mold to evaporate the solvent. The resulting composite is pressed
for about half an hour on a hot press at a temperature o f 266 °F and a pressure o f
10 psi.
3.4
Characterization
To validate the existence o f imitators on the surface o f particles, the following
analyses were used. Fourier transform infrared (FT-IR) spectra were recorded in
the FT-IR spectrometer (Jasco, FT-IR 420) in transmission mode under dried
nitrogen flow (10 cubic centimeters per minute, ccpm). The liquid catalyst and
accelerator were mixed with powder KBr, ground and compressed into a pellet.
Its spectrum was recorded as a reference to be compared with that o f the
nanoparticles treated with the catalyst and accelerator. Thermo-gravimetric
analysis (TGA, PerkinElmer) was done from 25 °C to 600 °C with an argon flow
rate o f 50 ccpm at a heating rate o f 10 °C/min to study the thermal degradation o f
nanocomposites.
61
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
To compare the quality o f parts that were fabricated from the direct method and
the surface initiated polymerization method, the following image and mechanical
analyses were used.
Atomic force microscope (AFM) (multimode, Digital
Instruments, Veeco Company) was utilized to characterize the dispersion quality
of
nanoparticles in the polyurethane matrix. The sample was prepared by
embedding the nanocomposite in a cured vinyl-ester-resin and polishing the
nanocomposite cross-section. Particle dispersion in PU was further characterized
with scanning electron microscope (SEM, JEOL field emission scaning electron
microscope, JSM-6700F). The SEM sample was the same as the AFM sample
except that a thin gold coating was sputtered to improve the electrical
conductivity needed for high-quality images.
The mechanical properties o f the fabricated nanocomposites were evaluated by
tensile tests following an American Society for Testing and Materials standard
(ASTM, 2002, standard D 412-98a). The samples were prepared according to the
standard. A crosshead speed o f 15 mm/min was used and strain (mm/mm) was
calculated by dividing the crosshead displacement by the gage length.
3.5
Result and Discussion
62
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
FR-IR and TGA Analysis
The physiochemical attachment o f catalyst-promoter (CP) mixture onto the
particle surface was verified by TGA and FT-IR analyses.
The sample was
prepared by dispersing the NPs into the THF solvent, adding the exact amounts
o f catalyst and accelerator, stirring for half an hour, and washing the CP-treated
nanoparticles with THF. TGA and FT-IR analyses were carried out as described
in the section 3.4 (Figure 3.3). The CP mixture is completely decomposed at
temperatures around 200° C. The as-received nanoparticles show continuous
weight loss possibly arising from the dehydration o f the physicochemically
adsorbed moisture; however the treated nanoparticles show a similar weight loss
as the catalyst/accelerator mixture by itself. This also indicates that the
physicochemically adsorbed moisture reacts with the CP mixture during
sonication. The hot spots around the nanoparticle surface during sonication favor
the loss o f physically adsorbed moisture and promote reaction between the
moisture and the CP mixture.
63
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
100
100
80-
NPs treated with catalyst+promoter
O)
8)
a 60Q.
Q.
£ 40-
O)
As-received NPs
20pure catalyst+promoter
100
200
300
400
500
600
T (°C )
Figure 3.3 TGA o f as-received nanoparticles, catalyst-accelerator mixture, and
treated nanoparticles
In order to test the amount o f polyurethane chemically bound onto the
nanoparticle surface, the SIP particle suspension in THF solution was thoroughly
washed with excess THF several times, dried under vacuum oven at room
temperature, and tested on TGA, Figure 3.4. It is clear to see that about 8 wt.%
PU is still present compared to 3.5 % for the treated NPs and 5% for as-received
NPs in Figure 3.3. This is evidence indicating the chemical bond between the
nanoparticle and the polyurethane matrix. Once polymerization is initiated from
64
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
the surface o f the particle, the extra polymer is linked by polymerization o f the
chemically bound polymer with the monomers. The chemical structure o f the
cured nanocomposite is shown in Figure 3.5.
The curved lines indicate the
polymer molecules binding a nanoparticle to the bulk PU, which is represented
by the straight lines.
100
98-
£ 96I
THF w a sh e d
90-
88
100
200
300
400
500
600
T (°C)
Figure 3.4. TGA o f the PU-NPs after reaction for 6 hours and washed with excess
THF before pouring
65
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
FG j O
Figure 3.5. Scheme o f the NPs filled polyurethane composites
The particle weight percentages remaining at 500° C in the TGA analyses in
Figure 3.6 indicate differences in particle distribution resulting from the two
processing methods. Here a sample taken from an edge o f a composite specimen
is compared with another taken from the center.
No difference in remaining
weight is seen in the SIP samples, indicating uniform particle dispersion.
However, a large difference is observed in the DM samples: the particle weight
percentage shifts from 45% to 70% as the edge is replaced by the center.
66
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
100
*
Ui
3
70
- ------ Pure PU
■ ------ Fe20 3-PU
- ------ Fe20 3-PU
------ Fe20 3-PU
------ Fe20 3-PU
\V \
DM
\Y \\
DM-edge^ \
SIP
ASIP-edge X
c
oo
0
a.
£
O)
1
4-1
20
10
0
| i l l
| 1—1-- .--- 1—i---1---1
50 100 150 200 250 300 350 400 450 500 550 600
T(°C)
Figure 3.6. TGA o f neat polyurethane, composites made by DM and SIP methods
measured at different sample locations
Figure 3.7 shows the derivative thermal gravimetric (DTG) behaviors taken from
Figure 3.6. Only one peak is observed in PU arising from thermal decomposition
o f free chain in pure PU. For the DM method, one peak at a lower temperature is
observed. The existence o f NPs promotes the decomposition o f PU with a lower
temperature due to the particle being heated. However, SIP composites show
three different peaks. The first two peaks are from the bulk matrix and the
intermediate phase between the particle and PU. The latter one is PU chemically
67
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
attached to the nanoparticle surface. This further indicates the SIP method favors
good bonding between nanoparticle and matrix.
700
600500I
400-
9
300-
o
3
IQ
PU
200-
DM
SIP
100-
100
200
300
400
500
600
Temperature (°C)
Figure 3.7. DGT o f Poylurethane, DM and SIP nanocomposites
68
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Further evidence o f the CP mixture adsorbed onto the nanoparticle surface was
shown in FT-IR.
FT-IR detects the vibrational characteristics o f chemical
functional group in a sample. Similar spectra were observed between the CA
mixture and CA-treated nanoparticles.
1 00 -
~ 90-
NPs treated with catalyst+promoter
Pure catalyst+promoter
80-
70-
60-
504000
3500
3000
2500
2000
1500
1000
Wavenumber (cm'1)
Figure 3.8
FT-IR
spectra o f catalyst-accelerator mixture and treated
nanoparticles.
The FT-IR spectra shown in Figure 3.9 reveal dramatic differences in the
transmittance properties before and after polymerization. The diisocyanate
monomer is believed to react with the physicochemically adsorbed moisture (peak
69
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
at 3433 cm '1) on the nanoparticle surface with the aid o f the CP mixture and
further copolymerize with the diols to form a composite as shown in Figures 3.1
and 3.2.
114 -
112
-
2 - 110
-
Pure PU
3
■
o>
c
m
2364
o
"
Surface Treated NP
1083433
Nanocomposite
106104-
102
4000
3500
3000
2500
2000
1500
1000
wavenumber (cm'1)
Figure 3.9. FT-IR spectra o f neat polyurethane, as-received nanoparticles and
nanocomposite
XPS as a sensitive tool to characterize the surface o f the materials were used to
distinguish the origin o f the oxygen, i.e., the effect o f CP treatment on the oxygen
in the nanoparticles, and the composite as compared to the as-received
nanoparticles. XPS spectra at O ls region are shown in Figure 3.10. There
70
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
different peaks, i.e., Fe-O-Fe at low binding energy (covalent bonding), Fe-OH
(chemically adsorbed moisture) and F e -0 -0 H 2 (physically adsorbed moisture) at
high binding energy, corresponding
3
m
c
JB
c
As-received NPs
Accelerato r&Prometer-treated NPs
PU+NPs
525
526
527
528
529
530
531
532
533
534
535
536
Binding energy (eV)
Figure 3.10. X-ray Photoelectron Spectroscopy (XPS) analysis on as-received
to three different oxygen origins were observed in all three samples. However,
the relative ratios o f the peaks varied completely among the CP-treated NPs, as-
71
R e p r o d u c e d w ith p e r m issio n o f th e c o p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
received nanoparticles and the PU/Fe 2 0 3 samples. Consistent with the TGA
analysis, as-received nanoparticles are coved with moisture as strong two peaks
at higher binding energy were present. Compared to this, CP treatment favors the
removal o f the moisture, thus showing low peak at higher binding energy for the
CP treated particles. However, some moisture residues in the form o f chemically
or physically adsorbed were present. Since as received particles are coated with
the moisture and the lower peak for Fe-O-Fe is expected and vise versa for the
CP treated particles. This was illustrated in the first peak in the figure. After the
polyurethane formed surrounding the CP-treated nanoparticles, spectra similar to
the as-received nanoparticles were observed due to the present o f the diol
molecules around the nanoparticles. Figure also showed the presence o f the
strong bonding between nanoparticle and polymer as indicated by the strong
middle peaks as compared to the AC-treated NPs. The weak peak o f Fe-O-Fe in
the PU+Fe 2 0 3 at lower is attributed to the coating o f the polyurethane.
Image Analyses
Optical micrographs are shown in Figure 3.11 for the DM composite, neat PU
and SIP composite. Large cracks are seen in the DM composite whereas no sign
72
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
o f cracking is found on the neat PU and SIP composite. The latter two specimens
were much more flexible than the first.
Figure 3.11. DM composite (left), neat polyurethane (middle) and SIP composite
(right)
Figure 3.12 shows the atomic force microscope (AFM) phase images o f a
composite with 65 % by weight o f particle loading. A clear sign o f agglomeration
is observed in the DM composite. However, discrete particles without obvious
agglomeration are observed in the SIP composite, indicating that a good
dispersion nanoparticles can be obtained by the SIP process.
73
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 3.12. AFM tapping mode phase images o f the 65 %wt. composites by (a)
DM method and (b) SIP method.
The SEM micrographs also indicate a homogeneous particle distribution within
the polymer matrix when the SIP synthesis is used. As shown in Figures 3.13 (b)
and (d), nanoparticles agglomerate during polymerization in the DM synthesis.
However, the polymer layer chemically bound to the nanoparticle prevents
agglomeration via steric forces.
74
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 3.13. Scanning electron micrographs of composites: SIP method (a) and
(c); and DM method (b) and (d).
The effect of particle surface treatment on interaction with PET vial walls
W hen the as-received nanoparticle suspension in THF solution is removed from a
PET vial, the vial wall changes from translucent white to brown as a result o f the
nanoparticles remaining on the wall. However, the 6-hour cured nanocomposite
particle suspension leaves no residue when the vial is emptied (Figure 3.13).
75
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure3.14. Chemical Structure of PET
The hydrogen bonding between the moisture on the as-received nanoparticle
surface (NP-OH) and the PET wall (C =0) favors stable attachment o f
nanoparticles onto the wall. On the contrary, the PU coating on the
nanocomposite particle has a very weak hydrogen bond and favors affinity to the
THF solvent rather than to the PET wall. The same reasoning can also explain by
the quick sedimentation o f as-received NPs in THF and the stable suspension o f
the nanocomposite particles.
76
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 3.14. As-received nanoparticles remaining on the PET vial wall (left), and
composite nanoparticles almost completely removed from the PET vial wall.
Mechanical Testing
*
The flexibility o f highly loaded composites was further evaluated by tensile test.
Figure 3.15 shows the tensile stress-strain curves o f the composites fabricated by
77
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
the DM and SIP methods. The Y oung’s moduli and tensile strengths are almost
the same for both composites. However, the elongation in the SIP composite is
about four times greater than that o f the DM composite. The strong chemical
bonding between the nanoparticles and the polyurethane polymer and uniform
particle distribution within the polymer matrix are believed to contribute to the
observed flexible behavior o f the SIP composites.
DM
SIP
4-
0
200
400
600
800
te n s ile strain {%)
78
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
1000
Figure 3.15. Tensile tress-strain curves of 65 %wt. composites fabricated by DM
method and SIP methods.
3.6
Conclusion
It has been shown that the surface-initiated-polymerization (SIP) method can
produce flexible iron oxide/polyurethane nanocomposites with a high particle
loading. Chemical, thermal, mechanical, and image analyses were conducted to
confirm that the SIP method yields better chemical bonding between the
nanoparticles and the polymer matrix with uniform particle distribution. The SIP
method thus opens a way to the fabrication o f nanocomposites with high particle
loading.
79
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Reference
1.
Billas, I.M.L., A. Chatelain, and W.A. de Heer, Magnetism from the atom
to the hulk in iron, cobalt, and nickel clusters. Science (Washington, DC,
United States), 1994. 265(5179): p. 1682-4.
2.
Gangopadhyay, S., et al., Effect o f oxide layer on the hysteresis behavior
o f fin e iron particles. Journal o f Applied Physics, 1991. 70(10, Pt. 2): p.
5888-90.
3.
Tang, H., et al., Gas sensing behavior o f polyvinylpyrrolidone-modified
ZnO nanoparticles fo r trimethylamine.
Sensors and Actuators, B:
Chemical, 2006. B113(l): p. 324-328.
4.
Zhanhu Guo, et al., Synthesis o f Poly(methyl methacrylate) Stabilized
Colloidal Zero-valence Metallic Nanoparticles. Journal o f Materials
Chemistry, 2006. 16: p. 1772-1777.
5.
Stankovich, S., et al., Graphene-based composite materials. Nature, 2006.
442: p. 282-286.
6.
Zhanhu Guo, et al., Surface functionalized alumina nanoparticle filled
polymeric nanocomposites with enhanced mechanical properties. Journal
o f Materials Chemistry, 2006. 16: p. 2800-2808.
80
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
7.
Waldo J. E. Beek, Martijin M. Wienk, and r.A.J. Jassen, Efficient hybrid
solar cells from zinc oxide nanoparticles and a conjugated polymer.
Advanced Materials, 2004. 16(12): p. 1009.
8.
Gall, K., et al., Internal stress storage in shape memory polymer
nanocomposites. Applied Physics Letters, 2004. 85(2): p. 290-292.
9.
Mohr, R., et al., Initiation o f shape-memory effect by inductive heating o f
magnetic nanoparticles in thermoplastic polymers. PNAS, 2006. 103(10):
p. 3540-3545.
10.
Talbot, P., A.M. Konn, and C. Brosseau, Electromagnetic characterization
o f fine-scale particulate composite materials. Journal o f Magnetism and
Magnetic Materials, 2002. 249(481-485).
11.
Brosseau, C., et al., Electromagnetic and magnetic properties o f
multicomponent
metal
oxides
heterostructures:
Nanometer
versus
micrometer-sized particles. Journal o f Applied Physics, 2003. 93(11): p.
9243-9255.
12.
Brosseau, C. and P. Talbot, Effective magnetic permeability o f N i and Co
Micro- and nanoparticles embedded in a ZnO matrix. Journal o f Applied
Physics, 2005. 97: p. 104325-1/11.
81
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Chapter 4
Microwave Properties of Fe2 C>3 , Fe and FeCo Based Magnetic
N anocomposites
4.1
Motivation
Electromagnetic absorbers are a critical part o f defense systems for their
contribution to survivability in air vehicles and are used in commercial products
for electromagnetic (EM) shielding.
However, existing electromagnetic
absorption materials and material forms have several drawbacks: they are heavy,
less durable, and effective only over fixed wavelength bands. The emergence o f
nanomaterials technology opened the door for new opportunities to further
improve the functionality o f electromagnetic absorbers. One way to improve the
current electromagnetic absorber is to exploit the properties o f magnetic
nanocomposites.
Recent studies performed on magnetic nanocomposites show
improvements in electromagnetic wave absorption properties with the use o f
nano-sized magnetic particles [1-3]. Magnetic nanoparticles offer novel material
properties from the core/shell structure o f the particle and the particle interaction
that bulk magnetic materials do not exhibit; thus, exploiting specific material
properties o f nanoparticles can improve the overall microwave properties o f the
system [4-6].
81
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
For an EM absorbing layer on a metallic substrate, the reflection
coefficient R o f a plane wave at normal incidence is given by
(1)
where Z the relative impedance (the input impedance over the impedance o f the
air) o f the material.
Z for a flat metallic surface coated with a layer o f dielectric
materials is given by
(2)
where ju and s are the relative permeability and permittivity, respectively, o f the
absorbing layer, d is the layer thickness, and X is the wave length in free space
[7].
The quest for a broadband, lightweight and durable microwave absorber was
already mentioned in the Introduction. The limitations o f current microwave
absorbers are well documented [8].
The study showed that the homogeneous
absorber material closely obeys Kramers-Kronig conditions. This condition states
that the material in which large values o f p ” are invariantly accompanied by
rapidly changing values o f p ’. This restriction limits the broadband absorption
characteristics o f a single layer (homogeneous material) microwave absorber.
82
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Author claims that to achieve a broadband performance, one needs a multilayer
approach in designing the absorber.
To overcome some o f these limitations, a material with superior magnetic
properties compared to the existing material must be used. The previous research
on magnetic properties o f nanoparticles offers the potential to design a better
microwave absorber. Ota et al [9] showed that by manipulating metallic content in
M-type hexagonal ferrite, one can change the resonant frequency and the
broadband absorption is still possible using a homogenous single layer
configuration in the form o f a particle mixture.
This study suggests that the
core/shell structures o f magnetic nanoparticles can be very useful in the design o f
microwave absorber.
In addition, there is another side benefit o f using magnetic nanoparticles as fillers
that cannot be ignored. Since the metallic magnetic materials are conductive, the
permeability decreases at high frequencies due to eddy current losses induced by
electromagnetic waves [10].
However, the eddy current loss can be suppressed
if the particle size is below the skin depth. At microwave frequencies o f interest,
the skin depth is typically around 1 jum [11]. This enables the electromagnetic
83
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
wave to fully penetrate the entire particle allowing the whole particle to be
effective in the operation o f the microwave absorber.
A magnetic nanocomposite can advance the technology o f microwave absorption
when combined with the right fabrication process and judicious selection o f
magnetic nanoparticles.
magnetic
This chapter discusses the microwave properties o f
nanocomposites
fabricated
by
the
highly
particle-loaded
nanocomposites fabricated by the SIP method discussed in the preceding chapter.
4.2 Material
Polymer
The same polymeric matrix as mentioned in the Chapter 3 was used in this study.
Magnetic Particles
The carbonyl iron (Cl) particle (BASF Group) is chosen as a baseline material for
comparison with magnetic nanocomposites. The Cl particle is widely used as a
magnetic filler for microwave absorbers because o f its excellent electromagnetic
properties [12]. The Cl particle is produced by thermal decomposition o f iron
pentacarbonyl (Fe(CO)s). In the course o f the decomposition process, spherical
84
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
particles form on nuclei (Figure 4.1). The particle size ranges from 2 to 5 /tm and
it has an iron content up to 99.5%. Figure 4.2 shows a SEM image o f Cl paticles.
From Iron to CIP
granules
co
I
+
Fe(CO)*-
synthesis
i
CO
FelCOU-
therm al
purification
decomposition
raw CIP
iron oxide .
red
effect
pigments
CVDooating
» sitting -
finished
CIP
►Insulating.
■ “ *fc* '
Figure 4.1. Synthesis o f carbonyl iron particle (Cl particle) [12]
Figure 4. 2. Carbonyl iron particles [12].
Fe and FeCo nanoparticles (Quantum Sphere, Inc.) and Fe203 particles, the same
particles used in Chapter 3, are chosen as nanoparticles.
Fe and FeCo
85
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
nanoparticles are produced through the gas phase condensation method as shown
in Figure 4.3 [13]. Metal is vaporized using resistance heating to a temperature
beyond the boiling point o f the material until a sufficient rate o f vaporization is
achieved. By computer control o f the metal flux, chamber pressure, temperature,
and gas flow, nanopowders are produced having the desired size and the size
distribution.
86
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
m
V lc u im i C l im b e r
m
m
'
H lb C o m p i i t e r i i e d
r
J
P io te u
'
b SS mm I
C o n tro l
T o P a c k a g in g
m
m
i
*. U i f
CEks*
■ •
r...-p, - ,
"
t
:*
■ 1 1 : j„‘ :
(^Gas
■
i
••
. . .
'
V a c u u m
W
ire
F e e d
■
p
1
I
1
r
e
f
e
r
*
—
A
&
■N H H M BN N W M H
j—
P p w e i S u p p ly
1
‘
■' » '
UmarnM Quenchant Gas
Figure 4.3. Processing nanoparticles through the gas condensation method [13]
87
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
4.3 Sample Preparation
For nanoparticles, the surface initialized polymerization (SIP) methods described
in Chapter 3 was used to provide the nanocomposites.
For Cl particles, a
conventional spraying method was used to fabricate a 1.5 mm thick composite
film. Particles were premixed with diluted polyurethane with Ketone. Then the
correct amount o f the catalyst and promoter was added to the solution. It is then
sprayed onto the aluminum substrate. The sample was then dried for 7 days for
complete cure o f the polyurethane.
Since the density o f Cl particle is known to be 7.6 g/cc, o f the resulting particle
volume fraction can be calculated. However, it is difficult to predict the particle
volume fraction for nanocomposites since the density o f nanoparticles are not
known.
Also, the limited supply o f nanoparticles made it more difficult to
prepare samples with different particle loadings. Thus two particle loadings, 35%
and 65%, using weight percentange were used to prepare nanocomposites.
The absorbing performance comparison was made with respect to the final
density o f the composite. The density was measured following the ASTM D792
standard. For the CI/PU composite, four (4) different particle volume loadings
were used: 15% (27% by wt.), 15% (55% by wt.), 25% (70% by wt.) and 35%
(79% by wt.). The density o f Cl composites was also measured. Having a broad
88
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
range o f particle loading for Cl composites ensured the density o f nanocomposite
specimens to fall within the range for Cl composites.
Material
CI/PU: 5% vol
CI/PU: 15% vol
CI/PU: 25% vol
CI/PU: 35% vol
Fe/PU: 35% wt
Fe/PU: 65% wt
FeCo/PU: 35% wt
FeCo/PU: 65% wt
Density (g/cc)
1.234
1.942
2.551
3.153
1.472
1.804
1.429
1.818
Table 4.1. Density o f CIP/PU, Fe/PU and FeCo/PU
The densities o f the fabricated composites are shown in Table 4.1.
The back
calculated densities o f Fe and FeCo nanoparticles were 3.15 and 3.5 g/cc,
respectively, which are much lower than expected. If Fe nanoparticles in reality
have a mix o f Fe and Fe 2 C>3 , the nanoparticle density should fall between 7.6 g/cc
and 5.2g/cc.
The lower-than-expected densites o f the nanoparticles can be explained by the
size effect associated with their microstructure [14], Here, we need to examine
the methods to fabricate the Fe and FeCo nanoparticles. The Fe and FeCo
nanoparticles are fabricated by using the gas condensation method. This method
89
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
produces nanoparticles with different shapes which may have some porous
structure (Figure 4.4).
The density is directly related to the atomic arrangement in the unit cell and any
misarrangement will change the density dramatically. The different shape and
possible porous structure give the nanoparticles lower weight, thus generating the
less density o f the nanocomposite. In comparison, the spherical shape o f the Cl
particles and their large size more readily allows a crystalline structure and hence
a higher density, thus closely resembling the bulk iron.
Figure 4.4. TEM images o f a) Fe and b) FeCo nanoparticles
90
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
The TEM pictures in Figure 4 also show the core-shell nature o f the particles.
The darker images are the metallic core while the lighter images are either an
oxide shell or completely oxidized particles. The particle sizes can be determined
from the TEM image. The average size was found to be around 1 0 - 2 0 nm. The
ideal particle for microwave absorption applications is to have the thinnest oxide
shell possible that will prevent the particle from oxidation in an ambient
environment while maintaining the maximum pure iron property at the core.
However, the TEM images suggest that a large percentage o f oxide is present in
the particles. Some particles are missing metallic core, suggesting complete
oxidization. The presence o f a significant amounts o f oxide within a particle will
certainly affect its magnetic properties and hence the resulting composite.
4.4 Characterization
Transmission electron microscopy (TEM) measurements were performed on a
JEOL 2010 microscope at an accelerating voltage o f 200kV.
For TEM
observation, the nanoparticles were dispersed in acetone and deposited on an
amorphous holey carbon coated gold grid.
The magnetic studies were carried out using a 9-Telsa Physical Properties
Measurement System (PPMS) by Quantum Design.
The field dependent
91
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
magnetization was measured at room temperature in the field range o f -2 Tesla to
2 Tesla.
The relative complex permeability and permittivity were measured using a
transm ission line technique. A washer-shaped specimen was cut from a thin (how
thin?) sheet o f magnetic composite. The outer diameter o f the specimen was 7.00
mm nominal and the inner diameter was 3.04 mm nominal. The specimen was
faced by abrading with 320 grit silicone carbide abrasive paper on a granite flat
until a smooth and uniform surface was achieved. The specimen-was then placed
in a sample holder that was placed in between the rigid beaded airline (APC-7)
and the flexible coaxial airline (APC-7) that were connected to the network
analyzer (HP Model 851 OB). The frequency generator was used to generate the
electromagnetic wave from 2 to 18 GHz. Figure 4.5 describes the experimental
set-up.
HP 851 OB Network Analyz
Port
Rigid
B eaded■
Airline
Flexible
Coaxial
Airline
Sample
Holder
92
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 4.5. 7mm set-up configuration
Two different measurements were obtained to characterize the electromagnetic
properties o f each specimen.
The scattering coefficients, air-backed reflection
measurement (SI 1) and the insertion loss measurement (S21), were used to
calculate the permittivity and permeability (Figure 4.6).
Figure 4.6. Scattering coefficients o f a specimen on a coaxial transmission line.
The scattering coefficients S 11 (w) and S21(&i) are related to the reflection and
transmission coefficients by
.
( 1 - R 2)z
T h iv "
O il.
x
(4)
(1 - Z2) R
(<a)=T h i v '
93
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
(5)
where R = the reflection coefficient and z = the transmission coefficient o f the
material. The permeability and permittivity are then deduced from the scattering
coefficients using a Nicholson-Ross algorithm [15]. A total o f three specimens
for each composite were measured.
Reflection Loss
Once the complex permeability and permittivity are calculated from the scattering
coefficients, these values along with the associated frequency are used as an input
for impedance Z (Equation 2) for the metal backed reflection loss. The reflection
loss coefficient R in dB is given by (see Eq. (1))
Metal Backed Reflection Loss [dB]=201og 10
Z -l
Z+l
(6)
For the present study, two different thicknesses, 1.02 mm and 2.04 mm, were used
to compare the microwave absorption performance o f magnetic nanocomposites.
The metal backed reflection loss (MBRL) coefficient is a good indicator for the
overall performance o f a microwave absorber since the absorbers are typically
applied on the surface o f a metal or a graphite/epoxy composite that is conductive.
4.5 Result and Discussion
94
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 4.7 shows the dispersion quality o f the Fe/PU nanocomposites. As seen in
the figure, no agglomeration was shown. The DM was tried but because o f high
agglomeration from the particles during the curing, the cracks appeared
ubiquitously.
The problem o f agglomeration was more profound in this case
compared to Fe 2 C>3 /PU samples due to its smaller size. The average size for Fe
and FeCo were 10 ~ 20 nm with tighter size distribution. The average size o f the
Fe 2 C>3 particle was 30nm but the size distribution was much larger, ranging from
20 to lOOnm.
This proved the SIP method was essential in fabricating high
particle loaded nanocomposites.
Figure 4.7 SEM image o f Fe/PU nanocomposite (65% wt. Loading)
95
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
4.5.1 Magnetic Properties
Figure 4.8 shows the magnetization curves for the Cl, Fe and FeCo particles. For
the Cl particle, the particle magnetization data is inferred from the CI/PU
composite data by dividing the composite magnetization with the particle weight
fraction. All four data sets o f CI/PU composite films produced almost the same
results as shown in Figure 4.7.
0,5
-
(0
2 o.o
— CI P/ PU 5 % Vol
-
0 .5
-
- CI P/ P U 1 5 % Vol
- - CI P/ P U 2 5 % Vol
- CI P/ Pu 3 5 % Vol
-
1. 0
-
-20000
0
-10000
10000
20000
Magnetic Field (O e)
Figure 4.8. Normalized Ms for different particle loading.
The particles tested exhibit the characteristics o f typical soft magnetic materials
in Figure 4.9; very low coercivity (below 125 Oe) and the lack o f a significant
96
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
hysterisis loop. The saturation magnetizations, Ms, for the Cl, Fe, and FeCo
particles are found to be 202, 98 and 104 emu/g, respectively. For the Cl, it is
close to the reported Ms o f a bulk Fe (216 emu/g) [16]. This is consistent with
the reported composition o f the Cl particle which is over 99% Fe.
200-
160120-
CIP
Fe
FeCo
40-40c
-80-
O)
05
^
-
120 -
-160-200
-20000
0
-10000
10000
20000
Magnetic Field (Oe)
Figure 4.9. Magnetization vs Magnetic Field for CIP/PU composite film with
different particle loading.
The Fe and FeCo particles show very similar magnetization behavior. The
significant reduction in Ms is the result o f the large presence o f oxides in the
97
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
particles as shown in the TEM micrographs. FeCo nanoparticles are suspected to
be a mixture o f Fe and Co nanoparticles rather than FeCo alloys. The Ms for
FeCo alloys is reported to be higher than the bulk Fe.
If the nanoparticles are
FeCo alloys, the Ms should be higher than for the Fe nanoparticles.
200
150
100
O)
Z3
E
(D
c
CIP/PU:
CIP/PU:
CIP/PU:
CIP/PU:
1.253 g/cc
1.939 g/cc
2.551 g/cc
3.153 g/cc
o
-t—«
ro
N
■4—4
0)
c
CD
(0
- 50
-100
-150
-200
-20000
-15000
-10000
0
-5000
5000
10000
15000
20000
M a g ne tic Field (O e)
Figure 4.10. Magnetization vs magnetic field for CIP/PU composite films with
different particle loadings.
Figure 4.10 shows the magnetization curves o f CI/PU composite films with
different particle loadings. As expected, the saturation magnetization increases
98
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
with the increase o f particle loading. The Fe/PU and FeCo nanocomposite film
shows a hysteresis loop within the coercivity range from 900 to 685 Oe (Figures
4.11 and 4.12).
The observed phenomenon on nanocomposite is that the hysterisis loop only
occurs in nanocomposites.
Neither the nanoparticles nor the Cl particles and
CIP/PU composite exhibited the significant hysterisis loop.
One possible
explanation for this occurrence is the effect o f domain o f the particle and the
particle interaction within the composite.
120
1008060-
f
§
— Fe Co / PU: 1 . 4 2 9 g / c c
F e Co / PU: 1 . 8 1 8 g / c c
-■ F e C o N P S
t'
40:
20-
c
o
•*—
*
co
20-
N
-
CD
C
-40-
CD
-60-
o>
-80-
100-
-120
-20000
0
-10000
10000
20000
Magnetic Field (Oe)
Figure 4.11. Magnetization vs magnetic field for FeCo/PU
99
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
120
100
-
80-
Fe/PU: 1.472 g/cc
Fe/Pu: 1.804 g/cc
FeNPs
20-
-
D)
20-
-40-60-80-
-100
-120
- 20000
- 15000
- 1 00 0 0
0
- 5000
5 00 0
10000
15000
2 0 0 00
Magnetic Field (Oe)
Figure 4.12. Magnetization vs magnetic field for Fe/PU
A ferromagnetic material consists o f many regions, called domains, within which
magnetization is constant in magnitude and direction. However, the magnitude
and direction o f magnetization varies from domain to domain in such a way that
the net magnetization vanishes. W hen an external magnetic field is applied, these
domains change their volume through the displacement o f the boundary walls
between adjacent domains [17]. The transition o f walls depends on the magnetic
anisotropy energy, which aligns the magnetic moments to preferred directions,
and the exchange energy, which is the interaction energy between atomic
100
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
moments.
The competition between these two energies provides the finite
thickness o f the wall.
When the size o f a particle decreases, it reaches a point where the particle can no
longer accommodate more than one domain.
That is, the particle becomes a
single domain (SD) particle. A single domain takes more energy to rotate the
magnetization than multiple domains (MD).
In an MD particle, only the
translation o f the domain walls is needed to change the magnetization.
A
collection o f SD particles however, requires the rotation o f all magnetization
directions, a more energetic process than translating domain walls. This, in turn,
increases the coercivity. Figure 4.12 shows the effect on coercivity o f the particle
size.
The largest coercivity occurs at a particle diameter that is close to the SD
size. Away from the SD region, the coercivity decreases either by the existence
o f an MD within a large particle or due to the randomizing effects o f thermal
energy for particles smaller than an SD (superparamagnetic).
For a pure Fe
particle, the size o f an SD is around 20 nm at 77K [18].
101
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
fikjnomagneMet x -04,
JVOX
lOOO
100
ooi
ipm
Particle diameter, d
Figure 4.13. The relationship between coercive force and particle diameter [18].
The size o f SD is dependent on the content o f the particle. Figure 4.13 suggests
that the size o f SD increases for the magnetite and hematite compared to the pure
Iron. For example, the SD for a hematite is as big as 10 pm at 300K according to
the figure. The Fe and FeCo particles are more close to the hematite (y Fe2C>3 )
than a pure iron because o f larger presence o f oxides so the SD size should fall
between the iron and the hematite. The TEM showed the particles were around
20nm, a well below the suspected SD o f our particle. This allures that the particle
is a superparamagnetic.
At room temperature, a superparamagnetic particle
shows no hysterisis loop due to the randomizing effect from the thermal agitation.
102
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
In conjunction with this, particles were packed together when the M vs H was
measured. This means that the volume o f the nanoparticles passed beyond the
percolation threshold.
Under such condition, the particles prefer a magnetic
closure-domain structure due to dipolar interaction. This system then behaves like
multi-domain structures even though the particles itself remain a SD [19]. This
may be the reason for the lack o f hysterisis loop shown in Figure 4.9 when the
nanoparticle is measured alone.
However, not knowing the exact percolation
threshold and the size o f SD for these nanoparticles makes difficult validate the
explanation and needs further study.
The coercivity (or hysterisis loop) exhibited in the nanocomposite can be
explained by dipolar interaction [20]. The dipolar interactions are function o f the
metal volume fraction.
The Fe and FeCo particles were dispersed in a non­
magnetic matrix which means the dipolar interactions were the dominant ones
among other particle interactions.
volume fraction o f metal increase.
In this case, the coercivity increases as the
The maximum occurs at the percolation
threshold and start to decrease beyond the threshold as the magnetic closuredomains begin dominating the magnetic properties.
The nanocomposites with
the particle loading below the threshold thus shows the coercivity compared to the
packed nanoparticles.
Again, not knowing detail information such as metal
103
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
content in the particle and SD size for the Fe and FeCo nanoparticles makes this a
proposition. Further investigation is needed to fully answer the existence o f the
hysterisis loop in nanocomposites.
The CIP/PU composites didn’t reveal any hysterisis loop.
This is because Cl
particle is already a multi-domain structure due to its size. So when the magnetic
field is removed, the magnetization is re-oriented to randomized direction through
the domain wall translation. This is typical o f a soft magnetic material.
4.5.2 Permittivity (£) and Permeability ( p ) Analysis
Figures 4.14 and 4.15 show the permittivities (real and imaginary) o f the CI/PU,
Fe/PU and FeCo/PU nanocomposite films. The relative permittivities are higher
for the nanocomposites than for the CIP/PU composites.
Also, the Fe/PU
nanocomposite films exhibit higher dielectric constants, s ' , (real permittivities)
compare to the FeCo/PU nanocomposite films. Since an iron particle is more
conductive than a cobalt particle, it is reasonable to see higher relative
permittivities for the Fe/PU nanocomposite films, Figure 4.14. For a conductive
particle filler o f conductivity cr in a dielectric medium, the imaginary permittivity
s" is given by [21]
(T
£
=
104
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
( 8)
where /
is the frequency.
This trend resembled the composite system with
metallic chain or flakes in a dielectric material which results in a high relative
permittivity [22]. It is possible that the Fe particles may be contacting each other
for the high loading composites, thus creating a long chain o f metallic filler.
22
20
1
8
'
16
14
12
10
8
6
4
2
2
4
6
8
10
12
14
16
18
Frequency (GHz)
Figure 4.14. Real permittivities o f CI/PU, Fe/PU and FeCo/PU nanocomposite
films.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
3.5
- - CIP/PU-.1.253 g/cc
CIP/PU: 1.939 g/cc
CIP/PU:2.551 g/cc
CIP/PU:3.153 g/cc
3.0'X
, Fe/PU: 1.804 g/cc
2.5-
2.0
-
w
Fe/PU: 1.472 g/cc
FeCo/PU: 1.818 g/cc
FeCo/PU: 1.429 g/cc
0. 5 -
0.0
2
4
6
8
10
12
14
16
18
Frequency (GHz)
Figure 4.15. Imaginary permittivities o f composite films.
Figures 4.16 and 4.17 show the real and imaginary permeabilities o f the
composites. The magnetic properties discussed in the previous chapter suggest
that the Cl particles have superior magnetic properties compared to the Fe and
FeCo nanoparticles. Because o f this, the real and complex permeabilities o f the
Fe/PU and the FeCo/PU nanocomposite films are lower than those o f the CI/PU
composite films.
106
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Fe/Pu: 1.472 g/cc
#•••• Fe/Pu: 1.804 g/cc
—0 “ FeCo/PU: 1.818 g/cc
FeCo/PU: 1.429 g/cc
CIP/PU:1.253 g/cc
CIP/PU:1.939 g/cc
CIP/PU:2.551 g/cc
CIP/PU:3.153 g/cc
4
6
8
10
12
Frequency (GHz)
Figure 4.16. Real permeabilities o f composite films.
107
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
1.05 0.900.75-
CIP/PU:1.253 g/cc
CIP/PU:1.939 g/cc
CIP/PU:2.551 g/cc
CIP/PU:3.153 g/cc
—■ — Fe/Pu: 1.472 g/cc
- $■—Fe/Pu: 1.804 g/cc
" O - FeCo/PU: 1.818 g/cc
- D -F e C o /P U : 1.429 g/cc
0 .6 0 0 .4 5 0 .3 0 0.150.00
Frequency (GHz)
Figure 4.17. Imaginary permeabilities o f composite films.
The implication o f high relative permittivity and low permeability for the Fe/PU
and FeCo/Pu nanocomposites as a mircrowave absorber will be a reduced
bandwidth in reflection loss. The characteristics o f this absorber should resemble
the dielectric absorber more than the magnetic absorber.
4.5.4 Effect of Dispersion Quality on Permittivity and Permeability
Fe203/PU nanocomposites were fabricated by both the DM and SIP methods.
The permittivity and permeability were then compared to see if any changes in
properties occur.
As seen in Chapter 3, the DM method caused many
108
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
agglomerations within the nanocomposites while the SIP method provided the
well dispersed particles. The mechanical properties showed dramatic difference
in two nanocomposites. Figure4.18 and 4.19 illustrates the effect o f dispersion
quality on electromagnetic properties.
The result showed the existence o f
agglomerations didn’t effect the electromagnetic properties o f the final
nanocomposites.
.
>
B
£
s': 65% SIP
s": 65% SIP
4-
<5
,
Ol
J
s': 65% DM
s": 65% DM
2
4
6
8
10
12
14
16
18
Frequency (GHz)
Figure 4.18. Comparison between the SIP and DM on permittivity
109
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
1.5
1. 0
-
&
0
£
0
Q- 0 .5 -
o.o
2
8
10
12
14
16
18
F r e q u e n c y (G Hz)
Figure 4.19. Comparison between the SIP and DM on permeability
4.5.3 Reflection Loss Analysis
Figure 4.20 illustrates the metal backed reflection loss (MBRL) o f the CI/PU
composite film with a thickness o f 1.02 mm. As the particle loading increases,
the MBRL gradually increases and the difference in reflection loss between
different particle loadings increases as the frequency increases. The power loss
o f 10 dB or more, which translates to a minimum 90% reduction o f the original
signature wave, is achieved around 13 GHz for the CI/PU composite with a
density o f 3.153 g/cc.
110
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
0
-5 -
00
-
10-
T3
CL
CO -15-
-
-CIP/PU: 3.153g/cc
- CIP/PU: 2.551 g/cc
- CIP/PU: 1.939 g/cc
CIP/PU: 1.234 g/cc
20-
-25 8
10
12
—I—
- 1—
14
16
18
Frequency (GHz)
Figure 4.20. Metal backed reflection loss for CI/PU composite films with
thickness = 1.02 mm.
Figure 4.21 illustrates the MBRL o f the CI/PU composite films when the film
thickness is doubled to 2.04 mm.
As the result suggests, the thickness o f the
absorber needs to be increased to get the lower frequency absorption.
In this
figure, if the target frequency o f absorption is 7 -8 GHz, and a goal o f the power
loss is 10 dB or more, then a film thickness o f 2.04mm is needed; doubling the
thickness o f film that was needed for 13 GHz as shown in Figure 4.20. As the
thickness o f the composite film increases the weight impact becomes more
profound.
Ill
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
-
ct
10-
-2 0 -
CIP/PU:
CIP/PU:
CIP/PU:
CIP/PU:
-30-
2
4
6
8
10
12
14
3.153
2.551
1.939
1.234
g/cc
g/cc
g/cc
g/cc
16
18
Frequency (G Hz)
Figure 4.21. Metal Backed Reflection Loss for CIP/PU composite film with
thickness = 2.04 mm.
Figure 4.22 shows the MBRL o f the Fe/PU and FeCo/PU nanocomposite films
with a thickness o f 2.02 mm. The MBRL characteristics are similar for these two
films. This is not unexpected since these two magnetic nanocomposites have
almost the same permeability and permittivity. It shows peak absorption around 8
GHz for the film with higher density and around 1 2 - 1 3 GHz for the fdm with
lower density. The reflection loss spectra exhibit narrow bandwidth at the peak
for the higher density nanocomposite.
This can be explained by the high
112
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
dielectric constant and the low permeability o f the nanocomposites. To get more
broadband spectra, the permeability must be increased. This means more iron
content is needed for the Fe nanoparticles.
o..
Ll ° . -
-
10-
"U
—• — Fe/PU: 1.472 g/cc
O FeCo/PU: 1.429 g/cc
- W - Fe/PU: 1.804 g/cc
□ ■ FeCo/PU: 1.818 g/cc
-30-
2
4
6
8
10
12
14
16
18
Frequency (G Hz)
Figure 4.22 Metal Backed Reflection Loss for Fe/PU and FeCo/PU
nanocomposite film with thickness = 2.04 mm.
Figure 4.23 shows a comparison between the CI/PU composite film and the
Fe/PU nanocomposite film.
For this comparison, it was assumed that the
requirement for the absorber was to have the MBRL minimum o f 20 dB at the
target frequency o f 10 GFIz. Also included in the figure is the commercially
113
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
available microwave absorber from RF Products, A Laird Technologies Company
[23]. A product called Q-Zorb Single Band Absorber (Product Number 1091) is
resonantly tuned to develop a 20 dB loss at 10 GHz. The product is a composite
film made out o f nitrile rubber with the Cl particles. The thickness o f this product
is 1.524 mm. The plot is redrawn so it can be clearly compared to the CI/PU
composite film and the Fe/PU nanocomposite film.
m
T3^
</>
in
o
_i
c
o
o
<D
5=
CD
a:
■Fe/PU Nanocomposite Film:
p = 1,804 g/cc, t=1.70 mm
Q-Zm s Single Bend Absorber
P>« 1031
12
-CIP/PU Composite Film:
p = 3.153 g/cc, t=1.575 mm
sg
- Q-Zorb Single Band Absorber(P/N 1091*):
Nitrile Rubber + CIP, t=1.524 mm
-10
\
16
15
I**"’’
/
\
1 15
a
14
\\J /
•25
RF Products, A Laird Technology Company
-40-
1
'
1
1
1
8
1
'
10
1
1------
12
i
I
14
16
18
F re q u e n cy (G H z)
Figure 4.23 Metal backed reflection loss comparison between CI/PU composite
film and Fe/PU nanocomposite for a target frequency o f 8GHz.
114
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
The optimum thickness o f the CI/PU composite film and the Fe/PU composite
film was calculated to meet this requirement. The Fe/PU nanocomposite film
(density = 1.804 g/cc) with the thickness o f 1.702 mm was able to achieve this
while the CI/PU composite film (3.153 g/cc) with a thickness o f 1.524 mm was
able to satisfy the requirement. The close resemblance o f the CIP/PU composite
film and Q-Zorb Single Absorber suggests that a similar amount o f particles must
have been added to the nitril rubber. Since the density o f nitril rubber is around
1.15 g/cc which is comparable to the density o f polyurethane, the density o f QZorb Single Absorber should be very close to the CI/PU composite film. Figure
4.24 is the weight chart versus frequency and the thickness o f the absorber,
provided by the Liard Technology Company.
According to this chart,
approximately 1.25 lbs/ft2 o f composite film is required for the target frequency
o f 10 GHz. When the thickness o f this film (0.06 inches) is converted to SI units,
the density o f this film was found to be 4.06 g/cc. This is comparable to the
CI/PU composite film.
115
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
RFSB Weight and Thickness vs. Frequency
3 .5
10
.15
.
0.14
.
0.12
0.14
O.08
0.06 1
0.04
0.02
0,5
2
4
6
8
10
12
14
16
■Weight (PSF)
•Thickness (in)
18
Frequency (GHz)
Figure 4.24 Typical single absorber weight and thickness for the different target
frequencies [23].
If the width and length o f the absorber are fixed to unity, a direct weight
comparison can be made by multiplying the density with the thickness o f the film.
For this comparison, the Fe/PU nanocomposite film was able to meet the
performance requirement with 37 and 50% reduction in weight when compare to
the CI/PU composite film and Q-Zorb Single Absorber. This suggests that the
magnetic nanocomposite can be an excellent choice for the discrete frequency
absorbing applications.
However, if the bandwidths are compared, the CI/PU composite film shows far
superior performance than the Fe/PU nanocomposite film. The CI/PU composite
film shows a bandwidth o f 7 GHz ( 7 - 1 4 GHz) while the Fe/PU nanocomposite
116
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
film shows a bandwidth o f 2 GHz (9 — 11 GHz) for a reflection loss o f 10 dB.
One way to overcome this problem is by having multi-layer magnetic
nanocomposites.
The permeability and permittivity can be varied by using different nanoparticles.
As an example, Figure 4.25 shows the predicted reflection spectra o f polymeric
nanocomposites that incorporate Fe, ferrites and Co nanoparticles [1].
These
predictions are based on the following model for the magnetic susceptibility %
[1,24]. Figure 4.25 clearly illustrates that the absorption band can be changed
substantially using a mixture o f different particles. For example, a nanocomposite
incorporating Co nanoparticles exhibits the absorption peak around 1 GHz while a
nanocomposite incorporating ferrite nanoparticles moves the absorption peak to 3
GHz.
117
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
-10
i
ICo
&
I
2
{ # > 2 .4
m m )
-3 0
Figure 4.25 Reflection spectra for three model composites. Particle size=20 nm
(Co, ferrite), 10 nm (Fe); particle volume ffaction=0.2 (Co, Fe), 0.5 (ferrite) [1].
d is the composite thickness. This is not an abnormal practice as multi-layer
absorbers are already commercially available. Figure 4.26 shows a dual resonant
absorber from a Laird Technology Company.
118
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
RFDB Multiband A bsorber
I
_j
Figure 4.26 Multi-band absorber from RF Product, A Laird Technology Company
[22].
One drawback, however, is the weight penalty associated with the design. The
weight increase can be substantial if the goal is to cover the frequency region o f a
quasi-microband (1 ~ 3 GHz). For 3 GHz, the thickness o f the layer needs to be
2.7 mm, a substantial increase from a 10 GHz absorber [23].
As the results in Figure 4.25 indicate, the use o f magnetic nanocomposites may
offer a solution to this problem. By closely controlling the particle composition,
the particle loading, and the thickness o f each layer, a broadband absorber without
substantial w eight penalty is possible. A nother possibility is the use o f the rule o f
119
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
mixture. By mixing different nanoparticles such as Co and Fe nanoparticles in a
single layer, a broad bandwidth may by realized.
4.6 Conclusion
The Fe and FeCo nanoparticles and their composite with PU matrix were
characterized for their magnetic and microwave properties. Their properties were
compared with those o f the CI/PU composite films. The results suggest that the
Fe and FeCo nanoparticles, produced by the gas condensation method, contain a
large amount o f oxides. This effectively reduces the magnetic properties o f the
nanocomposites when compared to the CIP/PU composite. Also the magnetic
properties o f the Fe and FeCo nanoparticles suggest that the gas condensation
method did not exactly produce the FeCo alloys. Rather a physical mixture o f
nano-Fe and nano-Co particles with oxides was suspected. Also studied was the
effect o f dispersion o f particles in nanocomposites on the permittivity and
permeability from FeaCVPU nanocomposites with the DM and SIP samples. The
result illustrated that the existence o f agglomerations didn’t change the properties.
The reflection loss o f the Fe/PU and FeCo/PU nanocomposite films were
calculated using the measured permeability and permittivity. The performance
was compared with those o f the CI/PU composite film and a commercially
120
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
available absorber.
The result indicates a substantial weight savings if the
magnetic nanocomposite film is used for a discrete frequency absorption
application.
This finding offers the feasibility of developing lighter-weight
microwave absorbers through the use o f magnetic nanocomposites.
A better
synthesis o f nanoparticles with higher iron content will further improve the
performance o f nanocomposites by increasing the bandwidth.
121
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Reference:
1. V. Bregar, “Advantages o f ferromagnetic nanoparticle composites in
mircrowave absorbers,” IEEE Transactions on Magnetics, Vol. 40, 2004,
pp. 1679-1684.
2.
J. R. Liu, M. Itoh, T. Horikawa, E. Taguchi, H. Mori, and K. Machida, “Iron
based carbon Nanocomposites fo r electromagnetic wave absorber with wide
bandwidth in GHz r a n g e f Applied Physics A, Vol. 82, 2006, pp. 509-513.
3. C. Sudakar, G. N. Subbanna, and T. R. N. Kutty, “Hexaferrite-FeCo
nanocomposite particle and their electrical and magnetic properties at high
frequencies,” Journal o f Applied Physics, Vol. 94, 2003, pp. 6030-6033
4. T. Furubayashi, I. Nakatani, and N. Saegusa, “Magnetic moment and
hyperfine fie ld in colloidal fin e particles o f iron,” Journal o f the Physical
Society o f Japan, Vol. 56, 1987, pp. 1855-1858.
5.
S. N. Khanna, and S. Linderoth, “Magnetic behavior o f clusters o f
ferromagnetic transition metals,” Physical Review Letters, Vol. 67, 1991, pp.
742-745.
6. F. Bodker, S. Morup, and S. Linderoth, '’‘Surface effects in metallic iron
nanoparticles,” Physical Review Letters, Vol. 72, 1994, pp. 282-285.
122
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
7. E. F. Knott, J. F. Shaffer, M. T. Tully, Radar Cross Section, 2nd ed., SciTech
Publishing, New York, 2004.
8. J.L. Wallace, “Broadband magnetic microwave absorbers: Fundamental
limitations,” IEEE Transactions on Magnetics, Vol. 29, 1993, pp.4209-4214.
9. H. Ota, M. Kimura, R. Sato, K. Okayama, S. Kondo, and M. Homma,
“Broadband microwave absorber using M-type hexagonal ferrite,” IEEE
Symposium on Electromagnetic Com patibility, Vol. 2, 1999, pp. 590-593.
10. M. Matsumoto and Y. Miyata, uThin electromagnetic wave absorber fo r
quasi-microwave band containing aligned thin magnetic metal particle, ”
IEEE Transactions on Magnetics, Vol. 33, 1997, pp.4459-4464
l l . S . Sugimoto, T. Maeda, D. Book, T. Kagotani, K. Inomata, M. Homma, H.
Ota, Y. Houjou and R. Sato, “GHz microwave absorption o f a fin e a-Fe
structure produced by the disproportionation o f Sm 2F e n in hydrogen,”
Journal o f Alloys and Compounds, Vol. 330-332, 2002, pp. 301-306.
12. From BSAF, The Chemical Company, www.inorganics.basf.com
13. From QuantumSphere Incorporation, www.qsinano.com
14. Private communication
with Dr. Z. Guo, Mechanical & Aerospace
Engineering, University o f California, Los Angeles
122a
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
15. A. M. Nicholson and G. F. Ross, “Measurement o f the intrinsic properties o f
materials by time domain techniques,” IEEE Transaction on Instrumentation
and Measurement, Vol. IM-19, 1970, pp.377-382.
16. S. Gangopadhay, G. C. Hadjipanayis, C. M. Sorensen, and K. J. Klabunde,
“Magnetism in ultrafine Fe and Co particles,” IEEE Transactions on
Magnetics, Vol. 29, 1993, pp. 2602-2607.
17. I. A. Privorotski, “Thermodynamics theory o f domain structures,” Reports on
Progress in Physics, vol 35, 1972, pp. 115-155.
18. D.
Dunlop,
“Magnetic
properties
o f fine-particle hematite,” Annual
Geophysics, Vol. 27, 1071, pp 269-293.
19. G. Xiao and C. Chien, “Giant magnetic coercivity and percolation effects in
granular Fe-(Si02) solids,” Applied Physics Letter, Vol. 51, 1987, pp 12801282.
20. Kechrakos and Trohidou, “Magnetic properties o f dipolar interacting single­
domain particles,” Physical Review B, Vol 58, No. 18, 1998, pp 1216912177.
21. S. Sindhu, M. Anantharaman, B. Thampi, K. Malini, and P. Kurian,
“Evaluation o f a.c. conductivity o f rubber ferrite composites from dielectric
measurements,” Bulletin o f Material Science, Vol. 25, No 7, 2002, pp599607.
122b
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
22. Private communication with Dr. J. Willis, Northrop Grumman Corporation
ISWR, One Hornet Way, El Segundo CA.
23. RF Product, The Laird Technology Company, www.lairdtech.com
24. Y. Raiker and M. Shliomis, “Theory o f dispersion o f the magnetic
susceptibility o f fin e ferromagnetic particles,” Soviet Physics. JETP, Vol. 40,
1975, pp. 526-532.
122 c
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Chapter 5
Summary and Future Work
5.1 Summary
A magnetic nanocomposite can advance the engineering o f microwave absorption
with the right fabrication processes and judicious selection o f magnetic
nanoparticles.
To attain an enhanced microwave absorber through a magnetic
nanocomposite, a new synthesis method o f processing magnetic nanocomposites
was developed.
Surface Initialized Polymerization (SIP) is used to provide physicochemical
adsorption o f the initiator onto the iron-oxide (FeiC^) nanoparticle surface in a
tetrahydrofuran (THF) solution. The SIP method involved the mixing of; the
catalyst, promoter and the NPs suspended in the THF solution using ultrasonic
agitation. The monomers were added into the above solution drop wise within 30
minutes the polymerization was allowed to occur within the solution over 6 hours.
Following polymerization the solution was poured into a mold to allow the
solvent to evaporate. The resulting nanocomposite was then pressed for about 30
minutes using a hot presser at a temperature o f 266 °F and a pressure o f 10 psi.
Physico-chemical analyses were used to validate the chemical adsorption o f the
catalyst and promoter onto the nanoparticle. Image analysis illustrated that the
123
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
dispersion o f nanoparticles was superior compared to the direct method.
Ultimately, the concept o f the SIP method was validated through mechanical
testing. The development o f this new synthesis is a significant progress towards
advancing magnetic nanocomposite microwave absorber technology.
processing provides
The SIP
a general method to prepare high particle loaded
nanocomposites for microwave characterization.
Utilizing the SIP method, magnetic properties and microwave properties o f the
Fe/PU and FeCo/PU nanocomposite films were studied. A comparison analysis
was performed with the CIP/PU composite film. The particle characterization
suggested that the Fe and FeCo nanoparticles, produced by the gas condensation
method, contained large amount o f oxides. This effectively reduced the magnetic
properties o f the nanocomposites compared to the CIP/PU composite. Also the
magnetic properties o f the Fe and FeCo nano particles suggested that the gas
condensation method didn’t exactly produce the FeCo alloys. Rather a physical
mixture o f the nano-Fe and nano-Co particles with oxides was suspected.
Also studied was the effect o f dispersion quality on the permittivity and
permeability. The result illustrated that the dispersion quality didn’t have any
impact on the property.
This is plausible since the agglomerations, if kept in
small amount within the sample, shouldn’t affect the electromagnetic properties.
124
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
However, a minute existence o f agglomeration can impact the mechanical
properties o f the nanocomposite since it can be the single point o f a failure.
The reflection loss o f the Fe/PU and FeCo/PU nanocomposite films was
calculated using the measured permeability and permittivity values.
The
performance was compared with the CIP/PU composite film and a commercially
available absorber.
The characteristic o f the magnetic nanocomposites was a
narrow bandwidth, making them suitable for only a discrete frequency absorber.
Due to the lower density o f the nanoparticles, a weight savings using the
nanocomposites compared to the conventional iron particles is possible.
5.2 Future Work
One o f the ultimate goals o f any R&D engineer in a commercial field is
transitioning a lab environment finding to production. This involves dealing with
many other issues that aren’t necessarily considered in the lab experiment. A
scaling-up issue is one o f them. The SIP method involves the sonication during
mixing o f the catalyst, promoter and NPs.
Sonication o f large quantity o f
polymers and polymeric solution must be addressed to realize the transition to
production.
125
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Another issue is the use o f a tetrahydrofuran (THF).
A THF is considered a
hazardous material and is prohibited from the production line [1], Use o f a
hazardous material is becoming more critical to places like California where the
environmental issues are sensitive. To make a transition to production, a possible
replacement for a THF needs to be investigated. One o f the possible replacements
is a Ketone. A Ketone is an approved solvent material that can be used as a
polymeric solvent.
Also worthwhile to investigate is the effect o f the SIP method on conventional
particle filled polymers. Typically the CIP is treated with the silane, used as a
coupling agent between the CIP and the polymer, to improve the mechanical
properties o f the absorber [2], The SIP should also work for the CIP (2-5 pm)
since the chemical adsorption from the particle to the polyurethane should be the
same as the nanoparticles.
If this is the case, then the SIP method is better
dispersion method regardless the size o f particles. It will be interesting to compare
the mechanical properties between these two methods.
As far as nanoparticle synthesis, the technology o f producing consistent and high
quality metallic particles that are stable in an ambient environment must be
further progressed.
One reason that the particle synthesis through the gas
126
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
condensation method was picked for the evaluation was that large amounts o f
particles can be synthesized in a short period o f time, thus overcoming the scaling
up issue associated with the particle synthesis [3].
However, as the result
suggested, a gas condensation method didn’t provide the particle quality needed
for the microwave absorber application.
A similar quality o f metallic
nanoparticles reported by Guo [4] is needed to truly evaluate the performance o f a
magnetic nanocomposite microwave absorber.
127
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Reference:
1. Private communication with M. Low, Northrop Grumman Corporation,
ISWR, August, 2006.
2. M. Matsumoto and Y. Miyata, “Thin electromagnetic wave absorber fo r
quasi-microwave band containing aligned thin magnetic metal particle
IEEE Transactions on Magnetics, Vol. 33, 1997, pp.4459-4464
3. D. Pesiri, C. Aumann, L. Bilger, D. Booth, R. Carpenter, R. Dye, E.
O ’Neill, D. Shelton, and K. Walter, “Industrial scale nano-aluminum
pow der m a n u fa c tu r in g Journal o f Pyrotechniques, Issue 19, 2004, pp.
19-30.
4. Z. Guo, Fabrication o f Core-Shell Nanoparticle, Ph.D Thesis, Department
o f Chemical Engineering, Louisiana State University, 2005.
128
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Документ
Категория
Без категории
Просмотров
0
Размер файла
3 587 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа