IEEE Transactions on Dielectrics and Electrical Insulation Vol. 19, No.

1; February 2012 107

Anomalous Behaviour of the Dielectric Spectroscopy

Response of Nanocomposites
Roman Kochetov, Thomas Andritsch, Peter H.F. Morshuis and Johan J. Smit
Delft University of Technology, High Voltage and Management Group
Mekelweg 4, 2628CD
Delft, the Netherlands

ABSTRACT
A study on the dielectric spectroscopy of epoxy-based nanocomposites filled with
different types of particles, such as Al2O3, AlN, MgO, SiO2 and BN, is presented. The
surface of the nanoparticles was modified with a silane coupling agent, in order to make
them compatible with the organic host and create a system with homogeneously
dispersed filler material. Morphological characterizations of individual particles and
fabricated composites were performed by means of transmission and scanning electron
microscopy. The present research addresses an analysis of the complex permittivity.
The relative permittivity of nanocomposites shows an unusual behaviour. Introduction
of a low percentage of high permittivity filler results in a decrease of the permittivity of
the bulk polymer material. We propose a qualitative explanation for the reduction of
the relative permittivity, compared to the reference samples. The interface layer of
surface modified particles plays a more important role than the nature of the particles
themselves. The immobilization caused by the surface treatment of the nanoparticles
seems to be the main factor determining the relative permittivity of the composites with
fillgrade below 5 wt.%. The imaginary part of the complex permittivity, which
represents the dielectric losses in the system, does not change significantly with addition
of nanofiller up to 5 wt.%.
Index Terms - Nanocomposite, surface treatment, dielectric spectroscopy, interface
phenomena, relative permittivity.

1 INTRODUCTION influence of nanofillers is unique in such a way that the

DIELECTRIC Spectroscopy (DS) is a powerful tool for
investigating a variety of dielectric processes for both
electrical and non-electrical applications. The dielectric
properties of materials play a fundamental role in the
description of physical phenomena in many branches of
modern science and engineering. Dielectric spectroscopy
provides a link between the dynamics of molecular motion of
the individual constituents of a complex material and the
characterization of its bulk properties. We used DS to analyze
the dielectric behaviour of a new class of polymer-based
nanodielectrics, which have attracted a lot of attention from
industry and the scientific community since the terms
“nanocomposite” and “nanodielectric” were mentioned in
literature the first time [1, 2] and were implemented in
practice by Toyota [3, 4]. Polymer composites are interesting
because of their unique behaviour especially at low
concentrations of nanofiller. It has been reported that the
presence of a small amount of inorganic filler can affect the
real and imaginary parts of the complex permittivity. The
Manuscript received on 1 July 2011, in final form 5 October 2011.
incorporation of clay [5], alumina, titania or zinc oxide [6-9]
into a polymer can reduce the relative permittivity compared
to the value of the neat polymer material. The dielectric
spectroscopy results show a behaviour that cannot be
explained by rules of mixture and therefore a qualitative
model has been proposed for the tentative explanation of the
low values of the relative permittivity. Further, it is explained
what is the role of the nature (type, shape) of the
nanoparticles.
2 MATERIALS USED FOR EXPERIMENTS
2.1 SAMPLE PREPARATION AND CHEMICAL
CHARACTERIZATION
The base material chosen for this study was a commercially
available epoxy resin (ER) system, which consists of a
bisphenol-A type epoxy and anhydride hardener. The
chemical structure of ER and hardener is shown in Figures 1
and 2, respectively.
The fillers used were Nanopox®, aluminum oxide (Al2O3),
aluminum nitride (AlN) and magnesium oxide (MgO).
Nanopox® is a colloidal silicon dioxide (SiO2) sol in an epoxy
resin matrix.

1070-9878/12/$25.00 © 2012 IEEE

108 R. Kochetov et al.: Anomalous Behaviour of the Dielectric Spectroscopy Response of Nanocomposites
о with the hydroxyl groups on the surface of nanoparticles or with
о о
о о о о
Figure 1. Chemical structure of a bisphenol-A type epoxy resin.
о
о of free OH-groups on the nanoparticle surface is necessary for
о groups of the GPS. In case of oxides like Al2O3 and MgO the
Figure 2. Chemical structure of a curing agent (hardener).
Besides Al2O3, AlN, SiO2 and MgO nanoparticles, boron
nitride (BN) nano- and microparticles have been used for our
study.
The composites containing 0.5, 2, 5 and 10 wt.% of Al2O3,
AlN and MgO and 10 wt.% of BN were fabricated using an
ex-situ procedure. The as-received particles were dispersed in
ethanol using an ultrasonic bath.
It is well known that very small particles tend to
agglomerate due to their large specific surface area and
intermolecular forces [10]. The dispersion of the nanofiller
is of vital importance for the overall dielectric behaviour of
nanocomposites. If the particles are not well dispersed, the
agglomerations of nanoparticles may lead to worse dielectric
properties than microparticles of similar size. A
homogeneous dispersion of the filler in the final composite
is necessary, since agglomerates would act as weak spots
which promote electrical, mechanical or thermal failure. To
achieve an even distribution of the nanoparticles inside a
polymer host it is necessary to modify the particle surfaces
with a coupling agent (CA) [11]. A CA improves the
compatibility between the hydrophilic nanoparticles and the
hydrophobic polymer, helps in achieving an even dispersion
of the nanoparticles and provides a good adhesion by
bonding the dissimilar host and filler material. For our study
we used a silane coupling agent (SCA) in form of -
glycidoxypropyltrimethoxysilane (GPS) for the surface
functionalization of nanoparticles. The chemical structure of
GPS is shown in Figure 3.
OCH3
O
CH2 CH2 CH2
H2 C CH O CH2 Si
OCH3
Figure 3. Chemical structure of -glycidoxypropyltrimethoxysilane (silane
coupling agent).
The reaction between inorganic particles and GPS can be
illustrated in two steps. The first step is the hydrolyzation of the
alkoxy groups of the silane with water to form a silanol group.
This reaction produces silanol and an alcohol as shown in
Figure 4. The second step is the condensation of silanol groups
neighboring hydrolyzed silane molecules [12].
о OX о OH
о о  
Si OX+3H2O Si OH+3XOH
OX OH
Figure 4. Hydrolysis of the GPS.
As it is evident from the hydrolysis reaction, the availability
successful surface modification. The particles we use have a
hydroxide layer on their surface, which reacts with the OH-
presence of hydroxyl groups on the surface is easy to predict.
Oxidation causes the formation of an oxide hydroxide layer
around the oxide core. The reactivity of AlN powder with
water has been reported by Bowen et al. [13]. A thin
aluminum hydroxide shell forms on the surface of the AlN
core at room temperature. When AlN nanoparticles are
hydrolyzed, initially an amorphous layer composed of AlOOH
is formed on the surface of the AlN particles, which then
transforms into Al(OH)3. The surface treatment of BN
nanoparticles was reported in [14].
It is important to calculate the exact amount of the GPS
needed for particle treatment. An excess of GPS can lead to a
homoreaction between the molecules of the SCA. This can
subsequently lead to a formation of a gel substance in
quantities large enough, so that their influence results in the
deterioration of the properties of the nanocomposite [15]. It
was calculated that the amount of GPS needed for
functionalizing for example nanosized alumina equals 3% of
the weight of the filler used.
The surface chemistry of as-received and GPS-treated
nanoparticles was characterized by a Fourier transformed
infrared (FTIR) spectroscope and thermogravimetric analysis
(TGA). After surface treatment, the peak on the FTIR
spectrum corresponding to OH groups was reduced, which
indicates the reaction of SCA with the functional OH groups
on the particle surfaces. The FTIR spectrum of the
functionalized particles indicates peaks corresponding to
asymmetrical and symmetrical stretching of CH3 and CH2.
The GPS contains both groups in its chemical structure [16].
TGA confirmed the FTIR analysis, thus it was assumed that
the SCA was successfully grafted on the nanoparticles surface
[17]. The SCA attached to the particle surface has an organic
chain, which degrades at elevated temperatures. This leads to
OCH3 a weight loss, which can be observed at treated particles. TGA
indicated that modified alumina particles have a higher weight
loss than untreated ones.
Subsequently the ER was added to the mixture and mixed,
using a high shear force stirrer, in order to disperse the
functionalized particles in the resin. The solution of epoxy
resin, functionalized particles and ethanol was heated to
remove the solvent. A magnetic stirrer was used during the
evaporation process, in order to ensure a good dispersion of
the nanoparticles inside the epoxy. Subsequently, the hardener
was added and the mixture was stirred with a high shear
mixer. Afterwards the resin was degassed and subjected to
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 19, No. 1; February 2012 109

ultrasonication, in order to remove air voids that were trapped
during the mixing process. The final composite was cast into
preheated Al-molds and cured at 140 ºC for 4 hours.
ER-SiO2 compounds were not created using this ex-situ
polymerization process. The silica nanoparticles were
synthesized inside the epoxy resin. The binary composite
system of 31.8 wt.% of silica and the ER was diluted to
prepare appropriate concentrations of SiO2. The rest of the
procedure was the same as for the other nanocomposite
samples, i.e. mixing with hardener, degassing and curing.
The samples prepared for the study have a dimension of 42
mm diameter with 0.5 mm thickness.
The size distribution appeared to be broad for AlN and
Al2O3. However, the fraction of particles with diameters larger
than 200 nm was negligible. MgO and SiO2 particles had a
narrow size distribution.
The micrographs of Al2O3 and SiO2 show the shape of the
particles being spherical. MgO particles have spherical, egg
and truncated cubic shapes. AlN particles have spherical,
hexagonal and cubic structures.
SEM allows getting an impression of the shape and size
distribution of microparticles. The shape of BN particles with
70 nm, 1.5 μm and 5 μm is spherical. BN particles with 0.5
μm have a platelet-like structure (see Figure 7).

2.2 MORPHOLOGICAL CHARACTERIZATION

X-ray diffraction measurements of as-received particles were
performed on a Bruker-AXS D8 Advance diffractometer using
a CuK wavelength as radiation source. The 2 ranges of the
data were taken from 10° to 90° with an increment of 0.02°.
Figure 5 shows the XRD chart of BN fillers with
different average particle sizes. The BN particles have a
hexagonal crystal structure according to these spectra. The
diffraction patterns on nano-, submicron- and micro-
crystalline BN indicate that the particles have exactly the
same structure.
Figure 6. TEM micrographs of Al2O3 (left) and AlN (right) nanoparticles.
XRD analysis indicated that all nanoparticles have a
crystalline structure. AlN particles are present in cubic and
hexagonal crystalline structures, while Al2O3 particles have
orthorhombic and cubic crystalline structures [16].

4000 70 nm
0.5 m
3500
1.5 m
5 m
3000
Intensity (a.u.)

2500

2000 Figure 7. TEM micrograph of MgO nanoparticles (left) and SEM micrograph
of BN submicron particles with APS 500 nm (right).
1500

1000 Ultrathin slices of a sample for TEM analysis were obtained

500 using a diamond knife on an ultramicrotome. The quality of
the dispersion for some nanofilled specimens is shown in
0
10 20 30 40 50 60 70 80 Figure 8. For example Al2O3 and SiO2 particles are fairly well
2 ( o ) dispersed inside the epoxy matrix. Beside well-dispersed
nanoparticles, we can also observe the formation of small
Figure 5. X-ray diffraction patterns of BN particles. The spectra are hard to aggregates in some other test specimens.
distinguish from each other because of the similarities of the BN particles.

Morphological observations of as-received particles and

thin layers of created composites were performed by means of
transmission electron microscopy (TEM) and scanning
electron microscopy (SEM), using a Philips CM30T and a
Philips XL20, respectively.
TEM allows viewing individual particles in an ethanol
suspension. Particle size distribution analysis revealed that the
average particle size (APS) was approximately 20 nm for
SiO2, 22 nm for MgO, 30 nm for Al2O3 and 60 nm for AlN Figure 8. TEM images of thin slices of ER-Al2O3-2 (left) and ER-SiO2-2
(see Figures 6 and 7). (right) nanocomposites.
110 R. Kochetov et al.: Anomalous Behaviour of the Dielectric Spectroscopy Response of Nanocomposites

3 PERMITTIVITY MEASUREMENT AND

EXPERIMENTAL RESULTS 9
120ºC
100ºC
3.1 DIELECTRIC SPECTROSCOPY MEASUREMENT 8
80ºC
METHOD 60ºC
40ºC
We use a fully automated Alpha-A dielectric analyzer by 7
20ºC
Novocontrol, which allows measuring the dielectric 0ºC

'
-20ºC
properties (i.e. conductivity, permittivity, impedance and 6

tan of a dielectric as a function of frequency and

5
temperature. A schematic representation of the DS setup
can be seen in Figure 9. The system consists of a ZGS
4
Alpha active cell [18] and includes an automatic
temperature control unit (Quatro cryosystem) with a
3
precision 0.1 °C [19]. The diameter of the measuring 10
-2
10
-1 0
10 10
1 2
10
3
10 10
4
10
5
10
6 7
10
electrodes is 40 mm. Frequency (Hz)

Figure 10.   of the reference neat ER samples for different temperatures as a

function of frequency.
Cryogenic cell Sample
I
10
1 … 3V AC 120ºC
Measuring 100ºC
Voltage with phase and
variable ≈ V U A 80ºC
magnitude 60ºC
frequency from
3 µHz up to 20 1 40ºC
MHz 20ºC
T 0ºC
''

-20ºC

Figure 9. A schematic of the DS setup. The sample lies within a cryogenic 0.1
cell, which can be heated up to 400 ºC and cooled down to -160 ºC with 0.01
ºC temperature stability.

Usually we are interested in the complex permittivity  0.01

(see Eq. 1) of the material, which relates to a material’s ability -2 -1 0 1 2 3 4 5 6

10 10 10 10 10 10 10 10 10
to respond to the electric field by its polarization.
Frequency (Hz)
*
     j  (1)
Figure 11.   of the reference neat ER samples for different temperatures as a
The behaviour of the permittivity as a function of frequency function of frequency.
and temperature is sensitive to material changes, such as
changes caused by e.g. aging processes, absorbed moisture or The broad peak on   (f), which can be seen at 1 kHz for
by the addition of fillers. The complex permittivity and tan -20 °C, moves to higher frequencies with increasing
were measured in the frequency range from 10 mHz to 10 temperature and is attributed to a β-relaxation process. The
MHz. β-relaxation is linked to the mobility of side chains in the
The dielectric properties were measured at different polymer. The α-relaxation peak is not visible at high
temperatures from -20 to 120 °C. Prior to dielectric temperatures because it is obscured by the DC conductivity,
characterization, aluminum electrodes were deposited onto which increases greatly above the Tg [22].
both sides of the specimens by sputtering, in order to achieve
a good contact between the samples and measuring electrodes
3.3 INFLUENCE OF FILLER TYPE AND
[20]. In order to minimize the amount of absorbed water in
CONCENTRATION
the specimens, which may affect the measurement data [21],
The modified nanoparticles with coupling agents on the
all samples were dried at 140 °C for 2 days before the
surfaces have different microscopic properties compared to
measurements.
the bulk polymer matrix, which is relatively far away from
the particle surface. A CA might affect the amorphous
3.2 DIELECTRIC RESPONSE OF NEAT EPOXY structure of epoxy and lead to more or less aligned polymer
The real and imaginary parts of the complex permittivity of chains, which are perpendicular to the particle surface (see
the neat ER samples are shown in Figures 10 and 11. Figure 12).
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 19, No. 1; February 2012 111

3.6
 
  3.4

3.2

'
3.0 Neat ER
MgO-0.5
MgO-2
2.8 MgO-5
Figure 12. A particle without surface modification (left) and a particle with MgO-10
surface modification (right). Grey area around surface treated particle is a
layer of aligned polymer chains.
-2 -1 0 1 2 3 4 5 6 7
10 10 10 10 10 10 10 10 10 10

Frequency (Hz)

Figures 13-15 show the real part of the complex

Figure 15.   of ER-MgO samples as a function of frequency at 20 °C.
permittivity of Al2O3 -, AlN-, MgO- and SiO2 - epoxy
composites as a function of frequency, for different filler
contents at 20 °C. With an increase in frequency,   3.8

decreases monotonically in all samples. As the frequency

increases, inert dipolar groups, e.g. larger molecules, can not
keep pace with the applied field. Therefore the contributions 3.6

of these inert dipolar groups to the permittivity disappear,

which leads to a decrease of the amount of polarization in the '
system [23]. 3.4

Neat ER
3.8 SiO2-0.5
3.2 SiO2-2
SiO2-5
3.6
-2 -1 0 1 2 3 4 5 6 7
10 10 10 10 10 10 10 10 10 10
Frequency (Hz)
3.4
Figure 16.   of ER-SiO2 samples as a function of frequency at 20 °C.
'

Neat ER
3.2 Al2O3-0.5
Al2O3-2
The higher values of   for Nanopox-based composites,
Al2O3-5 compared to those of neat polymer and composites filled with other
3.0
Al2O3-10 particles are assumed due to the different specimen preparation
-2 -1 0 1 2 3 4 5 6 7
process. Nanopox composites were produced using an in-situ
10 10 10 10 10 10 10 10 10 10
polymerization process. The surface modification is not part of this
Frequency (Hz)
synthesis method, therefore we assume that there is no interfacial
Figure 13.   of ER-Al2O3 samples as a function of frequency at 20 °C. layer of aligned polymer chains in the proximity of the SiO2
nanoparticles. Thus, immobilization of epoxy chains does not take
place, which would explain why a reduction of   could not be
4.0
observed. Another reason why Nanopox-based composites have
higher relative permittivity values can be due to byproducts, which
3.8 are present inside the polymer after the in-situ precipitation of
nanosilica [24]. These byproducts are assumed to be very mobile
3.6 and lead to a higher   of the composite material overall.
The   spectra of composite materials containing surface
'

3.4 modified particles (Al2O3, AlN and MgO) show a similar

Neat ER trend. In contrast to the work of Singha and Thomas [6-8],
AlN-0.5
3.2
AlN-2
who observed the lowest relative permittivity for composites
AlN-5 filled with 0.1 wt.% of alumina, titania (TiO2) and zinc oxide
3.0
AlN-10 (ZnO) nanoparticles, our composites with 2 wt.% of modified
10
-2
10
-1
10
0
10
1
10
2
10
3
10
4
10
5
10
6
10
7
particles have the lowest relative permittivity values. The
Frequency (Hz) value of   at 0.5 wt.% is the second lowest. An
Figure 14.   of ER-AlN samples as a function of frequency at 20 °C. incorporation of 5 wt.% and further 10 wt.% gives a rise in
112 R. Kochetov et al.: Anomalous Behaviour of the Dielectric Spectroscopy Response of Nanocomposites

  . Singha and Thomas investigated the systems based on ER 100.0

and filled with 0.1, 0.5 and 1 % of TiO2, ZnO and Al2O3. The as-received Al2O3 particles
higher filler concentration reflected in the higher relative 99.5
surface treated Al2O3 particles
permittivity values of the composites. We explain the
difference between our results and the results of Singha and

Weight loss (%)

99.0
Thomas by various process of surface modification, which
might influence the dielectric behaviour of the final 98.5
nanocomposite. It can also be explained by variations in the
synthesis process, for example, different time of 98.0
ultrasonication and curing. The epoxy system the samples
were based on is also not the same. Variations in the chemical 97.5
structure of the ER can lead to a different thickness of the
interfacial layer, thus a different response of the polymer to an 0 100 200 300 400 500 600 700 800

external electric field. Temperature (ºC)

We assume two effects, which affect the polarization of the

Figure 17. TGA spectra of untreated and GPS treated Al2O3 nanoparticles.
ER. The first effect is the restriction of the chain mobility in
the vicinity of the nanoparticles, within a few nanometers of
100
the filler surface. It can be found in literature [25-27] that the as-received MgO particles
estimated thickness of the interfacial layer is between 5 and 20 98
surface treated MgO particles

nm, depending on the sample composition. This effect of

chain restriction is caused by the interface polymer layer with

Weight loss (%)

96
modified molecular structure and chain dynamics, compared
to neat polymer matrix. The dipolar groups are immobilized in 94

the proximity of nanoparticles because of GPS, which

connects these groups to a nanoparticle. 92

ER samples filled with MgO nanoparticles have the lowest

90
relative permittivity values tested. The interesting feature is
that MgO (APS is 22 nm) nanoparticles have the smallest size 88
compared to Al2O3 and AlN (APS is 30 nm and 60 nm, 0 100 200 300 400 500 600 700 800

respectively). All   (f) curves of ER-MgO have lower values Temperature (ºC)

than those of neat epoxy samples. Since the   of the bulk
material of MgO, Al2O3 and AlN is approximately the same,
namely 8.5-10 depending on the structure [28-30], we can
conclude that interaction between MgO nanoparticles and host
is the strongest. Since MgO nanoparticles are the smallest,
they have the highest surface-to-volume ratio, which leads to
larger amounts of OH groups on the surface. The evidence of
larger amounts of hydroxyl groups on the surface of MgO
nanoparticles, compared to e.g. Al2O3 nanoparticles, can be
found in TGA spectra of as-received and GPS-treated
particles of these types (see Figures 17 and 18). The higher
weight loss for MgO in Figure 18 indicates a higher amount of
GPS groups that reacted with the OH groups on the surface of
the nanoparticles. The reaction of hydroxyl groups with GPS
creates stronger bonds between particles and host. Therefore
more dipolar groups will be immobilized in the vicinity of the
MgO nanoparticles. The importance of the size of the particles
and therefore the interface region is underlined by results for
BN-ER composites shown in section 3.4.
AlN particles have the largest APS and therefore smallest
interface volume. The size of the filler plays an important role.
Since the volume of interfaces decreases for increasing filler
size at constant weight percentage, the move to larger filler
results in smaller zones of immobilization. The larger the
particles are, the larger the influence of the filler properties
become compared to the changes of the polymer structure and
the influence of the immobilization layers diminishes.
Figure 18. TGA spectra of untreated and GPS treated MgO nanoparticles.
The relative permittivity values of ER-AlN-5 are higher
than that of neat ER, while values of ER-Al2O3-5 are still
lower. The amount of hydroxyl groups on the surface of
alumina is larger than those on the surface of AlN. More GPS
groups connect to the surface of Al2O3 particles and in turn
more dipole groups of epoxy will be immobilized in the
interface polymer layer.
The second effect, which is responsible for the increase of
  , originates from the relative permittivity of the filler itself.
As the volume of filler material increases, the influence of the
material properties of the particles on the   of the matrix
becomes more pronounced. The relative permittivity of all
nanoparticles is higher than the one of ER. The properties of
the filler become predominant at 10 wt.% of Al2O3 or AlN and
therefore   (f) has higher values than neat ER. The surface of
silica nanoparticles was not modified. That explains why the
relative permittivity of ER-SiO2 systems has higher values at
all filler concentrations. The higher volume fraction of the
filler with higher   results in a higher relative permittivity of
the system itself.
The   values of all specimens at 1 kHz and the difference
compared to neat ER are shown in Table 1. Specimens with
MgO nanoparticles have values between 3.06 and 3.25, which
is a reduction by 8.5% to 15%, compared to the neat ER. The
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 19, No. 1; February 2012 113

addition of alumina nanofiller leads to a 7.4% reduction as
lowest value of   compared to unfilled ER, which has an  
equal to 3.52. The lowest value for ER-AlN composites is 3.4,
which is a 3.6% reduction. ER-SiO2 samples have a higher  
compared to reference sample.
nanoparticles can be used for electrical insulation applications
and the losses are much lower than those of microcomposite
systems filled with 60 wt.% of conventional sized alumina or
silica filler.
3.8

Table 1. The relative permittivity of the specimens at 1 kHz and the difference + 2.5%
compared to neat ER. 3.6

Composite APS, Relative   of a Difference

nm permittivity of compo- to neat ER 3.4 - 3.6%
particles   site (%)
p

'
- 7.4%
Neat ER - - 3.52 - 3.2
- 15%
ER-Al2O3-0.5 30 (8.5 – 10.0)* 3.41 -3.3 SiO2
ER-Al2O3-2 30 (8.5 – 10.0)* 3.28 -7.4 Neat ER
3.0 AlN-2
ER-Al2O3-5 30 (8.5 – 10.0)* 3.45 -2.0 Al2O3-2
ER-Al2O3-10 30 (8.5 – 10.0)* 3.69 +4.7 MgO-2
2.8
ER-AlN-0.5 60 (8.7 – 8.9)* 3.44 -2.2 10
-2
10
-1
10
0 1
10 10
2 3
10 10
4
10
5
10
6
10
7

ER-AlN-2 60 (8.7 – 8.9)* 3.40 -3.6 Frequency (Hz)

ER-AlN-5 60 (8.7 – 8.9)* 3.59 +1.9
ER-AlN-10 60 (8.7 – 8.9)* 3.78 +7.4 Figure 19. The effect of the filler type on the   of nanocomposites with 2
ER-MgO-0.5 22 (9.6 - 9.8)* 3.14 -12.1 wt.% for 20 °C as a function of frequency.
ER-MgO-2 22 (9.6 - 9.8)* 3.06 -15.0
ER-MgO-5 22 (9.6 - 9.8)* 3.18 -10.9 0.1

ER-MgO-10 22 (9.6 - 9.8)* 3.25 -8.5 Neat ER

ER-Al2O3-0.5
ER-SiO2-0.5 20 (3.8 – 7.0)* 3.58 +1.6
ER-SiO2-2 20 (3.8 – 7.0)* 3.61 +2.5 ER-Al2O3-2

ER-SiO2-5 20 (3.8 – 7.0)* 3.68 +4.5 ER-Al2O3-5

''

ER-BN-10 70 nm 70 (4.0 – 4.5)* 3.47 -1.4

ER-BN-10 500 nm 500 (4.0 – 4.5)* 3.90 +10.8
ER-BN-10 1500 nm 1500 (4.0 – 4.5)* 4.09 +16.0
ER-BN-10 5000 nm 5000 (4.0 – 4.5)* 4.33 +22.9
0.01

* Values for bulk material, depending on the literature source [28-37]. They
might be different for particles of nanometric size.
-2 -1 0 1 2 3 4 5 6
10 10 10 10 10 10 10 10 10

Frequency(Hz)
Figure 19 shows the effect of nanoparticles with a fillgrade
Figure 20.   of ER-Al2O3 samples as a function of frequency at 20 °C.
of 2 wt.% on the   of the nanocomposites as a function of
frequency. Nanocomposites with 2 wt.% are shown, because
the reduction of   is the most pronounced at this fillgrade for
0.1
Neat ER

all investigated nanocomposite types. The   of neat epoxy

ER-SiO2-2
ER-SiO2-5
has a higher value than composites filled with AlN, A2O3 and
MgO, but lower than that of ER-SiO2-2. The surface of AlN,
''

A2O3 and MgO was treated with GPS, which creates better
adhesion and aforementioned interfacial layer.
The dielectric losses of the composites containing a little
amount of nanoparticles (up to 5 wt.%) do not differ much 0.01

compared with the neat polymer (see Figures 20 and 21).

The main difference in the dielectric spectroscopy spectra
of composites with nanoparticles is observed at frequencies -2
10
-1
10
0
10
1
10
2
10
3
10
4
10
5
10
6
10

higher than 1 kHz. The polarization of epoxy dipolar groups is
more pronounced in the high frequency region, which causes
higher losses in this region. The lower losses of
nanocomposites can be indication for the presence of the
interfacial layer. The side chains, which are responsible for
the β-relaxation peak, are restricted in the vicinity of
nanoparticles, while all side chains of unfilled polymer can
move in respect to applied field and bring losses to the
dielectric system. However,   at 50 Hz is almost the same
as for ER. That means that ER filled with small amount of
Frequency (Hz)
Figure 21.   of ER-SiO2 samples as a function of frequency at 20 °C.
3.4 EFFECT OF FILLER SIZE
The relative permittivity of epoxy-based compounds with
10 wt.% of modified micro- and submicron-BN particles is
higher than of the neat ER, while   of ER filled with 70 nm
particles is lower compared to the reference sample (see
Figure 22). XRD analysis showed that the structure of BN is
114 R. Kochetov et al.: Anomalous Behaviour of the Dielectric Spectroscopy Response of Nanocomposites

the same (see Figure 5). Thus, the difference between them illustrated in Figure 24. Smaller particles with the same
could be the size of particles, their aspect ratio and surface weight percentage lead to a larger interfacial volume. In our
modification. The   increases if the size of particles case we have a directly proportional relationship between the
increases, while the concentration of them remains constant. overall volume of the interfacial layer and the amount of
polarization. The movement of the dipoles in the layer close to
the particle is restricted. Since the movement of dipoles is
limited, the   of the system is lower for smaller particles.
4.4

4.2

4.0
Particle
3.8
'

Interface
3.6

Neat ER Host
3.4
BN-10 5 m
BN-10 1.5 m
3.2
BN-10 0.5 m
BN-10 70 nm Particle/Interface Zone:
3.0
-2 -1 0 1 2 3 4 5 6 7
10 10 10 10 10 10 10 10 10 10
Frequency (Hz)
Figure 24. Illustration of the change of the ratio between particle and interface
volume as a function of particle size [38].
Figure 22. The effect of the filler size on the   of composites with 10 wt.%
BN particles for 20 °C as a function of frequency.
The dielectric losses of the samples with 1.5 m and 5 m
The imaginary part of the complex permittivity   , which particles have higher values compared to the reference sample.
represents the dielectric losses, is shown in Figure 23. The The   of the specimens filled with 70 nm and 0.5 m
dielectric losses of composites filled with BN are higher particles do not differ much compared to neat epoxy.
compared to ER in the low frequency region (<1 kHz) and
lower for frequencies higher than 1 kHz, with exception of 4 DISCUSSION
BN with 5 m APS. The broad dielectric loss peak in the high
frequency region (10 kHz) is attributed to a β relaxation It is generally accepted that the surface-to-volume ratio of
process of the epoxy network. It moves to a higher frequency the nanoscale fillers and the interface-polymer layer in the
with increasing temperature. vicinity of the filler plays a significant role for the decreased
  that has been observed [39, 40].
Tanaka et al. [41] proposed the multi-core model, where he
0.1 divided an interfacial layer into 3 layers (bonded, bond and
Neat ER
loose layers). Andritsch [22] proposed the polymer chain
ER-BN-10 (5 m)
ER-BN-10 (1.5 m) alignment model and introduced 2 layers (aligned and
ER-BN-10 (0.5 m) affected). Rätzke & Kindersberger [42] and Todd & Shi [43]
ER-BN-10 (70 nm)
proposed a single interface polymer layer or interphase. All
''

models have in common the assumption that the molecular

structure in the interface layer is reorganized, which results in
a change of the morphology and of the chain dynamics around
the particles.
0.01
In our model, we assume the existence of an interface layer
between the filler and matrix, which has different properties.
-2 -1 0 1 2 3 4 5 6
This interfacial layer is discussed to illustrate the effects of
10 10 10 10
Frequency (Hz)
10 10 10 10 10
surface modified nanoparticles on the dielectric properties of
nanocomposites. The interfacial layer was investigated and
Figure 23. The effect of the filler size on the   of composites with 10 wt.% estimated its size in prior works of Miwa et al. [25], Chen et
BN particles for 20 °C as a function of frequency. al. [26] and Hu et al. [27]. We use their work as a basis to
describe a nanodielectric system filled with different types of
The   of BN filled composites was shown to be dependent surface modified nanoparticles. Our model has its limits
on the APS. Composites with larger particles show a higher however, since current measurement techniques do not offer
permittivity. An explanation for this behaviour could be found insight in the exact nature of the interface, e.g. the value for
in the structure of the interface between the filler and polymer the permittivity of the interfacial zone.
matrix. Every modified particle has an interfacial layer of Due to the large surface area of nanoparticles, the interface
constant thickness. For larger particles this interfacial layer is layer can occupy a significant volume of the polymer in a
relatively thin compared to the particle dimensions, as nanocomposite [44, 45]. We assume that the interface layer
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 19, No. 1; February 2012
plays a role for the properties of a composite, which is more
important than the nature of the particles themselves. The total
interface volume would occupy a larger space if the size of the
particles becomes smaller for the same fillgrade. The treated
nanoparticles can change the degree of cross-linking, since
polymer groups react with SCA molecules, which are grafted
on the surfaces of the nanoparticles.
The rules of mixture, which are used to predict the relative
permittivity values of composites filled with conventional
microsized particles, cannot be applied to the systems
containing nanoparticles. Andritsch et al [22, 46] tried to use
the Looyenga formula [47] and semi-empiric models [48, 49]
to predict the relative permittivity of nanocomposites, but did
not succeed in fitting experimental data. The interface layer
which is created by SCA and resulted in the aligned polymer
chains has to be taken into account with an entirely different
permittivity [44, 50]. The immobilization caused by the
surface treatment of the nanoparticles seems to be the main
factor determining the   of composites with fillgrade below
5 wt.%.
In our model we assume that the interfacial polymer layer,
which has different properties compared to the bulk polymer,
act as a third phase. A third phase, i.e. interfacial polymer
layer, immobilizes the epoxy chains around the particles. The
dipolar groups in these immobilized chains are less able to
follow the external field, which leads to lower polarization of
the interfacial volume and as a result to lower polarization of
the whole composite. The lower polarization of a material
reflects on the macroscopic properties of a material, such as
relative permittivity. The highest degree of epoxy chain
immobilization and as a result, lowest relative permittivity
value is observed for the systems filled with 2 wt.% of the
filler. The smaller the particles are, the larger the relative
interfacial layer is and the larger the total interface volume.
We notice that the smallest MgO nanoparticles (APS is 22
nm) create the system with the lowest relative permittivity
values. Even the system containing 10 wt.% of MgO filler has
lower   than that of the neat ER. Conversely, the addition of
5 wt.% of the largest particles (AlN with an APS of 60 nm)
results in a higher   than for the neat ER.
The relative permittivity of a nanocomposite system is the
measure of the polarization of the polymer host, filler and
interface layer. If the   of the layer is much lower than that
of epoxy or filler, the contribution would lower the value of
  for the entire system. The relative volume of this layer in
the composite with nano-BN is larger than in one of
composites with bigger particles. Thus, the impact of the low
value of   in the interface layer is more pronounced in case
of the system filled with 70 nm particles, therefore the   (f)
values are lower than those of ER and composites filled with
larger particles.
Typically, the interface thickness is in the order of 5 to 20
nm [25, 26, 43]. If we assume the interface polymer layer
equal to 5 nm, which is constant for all BN-filled systems,
then the interface region volume fraction will occupy 12% of
the whole volume for 10 wt.% of the BN particles having an
115
APS of 70 nm. For the same fillgrade but particles with an
APS 1.5 m, the interface volume will be 0.5%, while for the
BN-10 with 5 m particles, this value is reduced to 0.1%. If
we make the same calculations, assuming the interface layer
of 20 nm, then the overall interface volume will occupy 70%,
2% and 0.6% for ER-BN-10 (70 nm), ER-BN-10 (1.5 m)
and ER-BN-10 5 (m), respectively (see Figure 25).
10
2
Log (interface volume fraction, %)
10
1
70 nm
1.5 m
10
0
5 m
-1
10
20 nm interface layer
-2
5 nm interface layer
10
1
10 10
2
10
3 4
10
Log (filler size, nm)
Figure 25. Log-log dependence of the filler size on interfacial volume
fraction, assuming different interface layer between matrix and filler.
We can conclude that the size of nanoparticles and the
thickness of the interfacial layer that is created by surface
modification of nanoparticles, act as a main characteristic
determining the dielectric behaviour of composites at low
filler concentrations.
5 CONCLUSIONS
Measurements of the complex permittivity revealed that the
relative permittivity of epoxy filled with 0.5, 2, 5 wt.%
alumina, 0.5, 2, 5 and 10 wt.% of magnesium oxide and 0.5
and 2 wt.% of aluminum nitride is lower than that of the
unfilled epoxy, despite the fact that the relative permittivity of
Al2O3, MgO and AlN is higher than   of the epoxy. On the
other hand, epoxy-silica composites showed higher values of
  compared to neat ER. Two effects play a role for the   of
a nanocomposite. The immobilization of the interfacial layer
and the   of the filler material. Due to the relatively small
amount of the filler, the main contribution to the altered   of
the nanocomposite comes from a restructuring of the polymer
due to the presence of the surface modified particles.
The dielectric behavior of nanocomposites was tentatively
explained by the existence of a strong interface layer between
polymer matrix and inorganic filler, and its influence on the
macroscopic properties of the composite. The dielectric losses
of the epoxy resin did not change significantly with addition
of nanofiller up to 5 wt.%.
The reduction of the relative permittivity compared to neat
epoxy resin at low filler concentration was explained by
mentioned immobilization of epoxy chains in the proximity of
nanoparticles. This immobilization is the result of the interface
polymer layer that was created by surface modification of the
116 R. Kochetov et al.: Anomalous Behaviour of the Dielectric Spectroscopy Response of Nanocomposites
nanoparticles. The minimum value of the relative permittivity
was noticed at 2 percent by weight. At this fillgrade modified
nanoparticles are far away from each other. If the distance
between particles is lower than the interfacial thickness of
approximately 20 nm, their interface layers do not overlap. Thus
the lower relative permittivity can be attributed to the intrinsic
properties of the rigid interfacial volume around the
nanoparticles.
If the volume fraction increases, the amount of the material
with high relative permittivity is increasing along with the
overlapping of the interfacial layers, which leads to a higher
relative permittivity of the overall nanocomposite. The
overlapping is possible at high filler loadings (≥ 10 vol. %) or
for very small particles.
The relative permittivity of epoxy-based compounds with
10 wt.% of modified BN particles is in all cases higher
compared to neat ER, when the particles have submicron and
micron-size dimensions. However, the   of the composite
with nano-BN is lower than that of the reference polymer.
This is attributed to the influence of interfacial volume, which
is a function of the particle size.
The loss spectrum of epoxy-based composites with nano-
and submicron BN filler has minor differences compared to
neat ER. The losses are lower in high frequency region
because the dipolar polarization of epoxy groups is
predominant. Losses are higher for composites filled with
micron-sized particles. This is explained by the influence of
interfacial polarization.
The β-relaxation peak that is observed for all
nanocomposites is reduced compared to the neat ER. It gives
an indication for the presence of the interfacial layer.
REFERENCES
[1] R. Roy and S. Komarneni, “Nanophase and nanocomposite materials”,
Mater. Res. Soc., 241, 1984.
[2] T.J. Lewis, “Nanometric dielectrics”, IEEE Trans. Dielectr. Electr. Insul.,
Vol. 1, pp. 821-825, 1994.
[3] A. Okada, Y. Fukushima, M. Kawasumi, S. Inagaki, A. Usuki, S.
Sugiyama, T. Kurauchi, and O. Kamigaito, “Composite material and process
for manufacturing same”, United States Patent 4739007, 1988.
[4] R.K. Gupta, E. Kennel, and K.-J. Kim, Polymer Nanocomposites
Handbook, CRC Press, 2010.
[5] R. Sarathi, R.K. Sahu, and P. Rajeshkumar, “Understanding the thermal,
mechanical and electrical properties of epoxy nanocomposites”, Mater. Sci.
Eng. A, Vol. 445-446, pp. 567-578, 2007.
[6] S. Singha and M.J. Thomas, “Permittivity and tan delta characteristics of
epoxy nanocomposites in the frequency range of 1 MHz-1GHz”, IEEE Trans.
Dielectr. Electr. Insul., Vol. 15, pp. 2-11, 2008.
[7] S. Singha and M.J. Thomas, “Dielectric properties of epoxy
nanocomposites”, IEEE Trans. Dielectr. Electr. Insul., Vol. 15, pp. 12-23, 2008.
[8] S. Singha, M.J. Thomas and A. Kulkarni, “Complex permittivity
characteristics of epoxy nanocomposites at low frequencies”, IEEE Trans.
Dielectr. Electr. Insul., Vol. 17, pp. 1249-1258, 2010.
[9] F. Ciuprina, I. Plesa, P.V. Notingher, T. Tudorache, and D. Panaitescu,
“Dielectric properties of nanodielectrics with inorganic fillers”, IEEE Conf.
Electr. Insul. Dielect. Phenom., Québec City, Canada, pp. 682-685, 2008.
[10] Y-W. Mai and Z-Z. Yu, Polymer Nanocomposites, Woodhead publishing
limited, 2006.
[11] E.P. Plueddemann, Silane Coupling Agents, Plenum press, 1982.
[12] E.M. Petrie, Handbook of Adhesives and Sealants, 2nd ed., McGraw-Hill,
New York, 2006.
[13] P. Bowen, J.G. Highfield, A. Mocellin, and T.A. Ring, “Degradation of
aluminum nitride powder in aqueous environment”, J. Amer. Ceram. Soc.,
Vol. 73, pp. 724-728, 1990.
[14] Y. Xu and D.D.L. Chung, “Increasing the thermal conductivity of boron
nitride and aluminum nitride particle epoxy-matrix composites by particle
surface treatments”, Compos. Interfaces, Vol. 7, pp. 243-256, 2000.
[15] S. Shokoohi, A. Arezafar, and R. Khosrokhavar, “Silane coupling agents
in polymer-based reinforced composites: a review”, J. Reinf. Plast. Compos.,
Vol. 27, pp. 473-485, 2008.
[16] R. Kochetov, T. Andritsch, U. Lafont, P.H.F. Morshuis, and J.J. Smit,
“Effects of inorganic nanofillers and combinations of them on the complex
permittivity of epoxy-based composites”, IEEE Int. Sympos. Electr. Insul.,
San Diego, CA, USA, pp. 340-344, 2010.
[17] R. Kochetov, T. Andritsch, U. Lafont, P.H.F. Morshuis, and J.J. Smit,
“The thermal conductivity in epoxy–aluminum nitride and epoxy – aluminum
oxide nanocomposite systems”, Nordic Insul. Symp., Gothenburg, Sweden,
pp. 27-30, 2009.
[18] G. Schaumburg, “New integrated dielectric analyzer extends accuracy
and impedance range for material measurement”, Dielect. Newsletter, Vol. 11,
pp. 4-6, 1999.
[19] G. Schaumburg, “Novocontrol cryo system for dielectric applications
improved by the new quatro 4.0 controllers”, Dielect. Newsletter, Vol. 4, pp.
7-10, 1995.
[20] G. Schaumburg, “On the accuracy of dielectric measurements”, Dielect.
Newsletter, Vol. 8, pp. 5-10, 1997.
[21] C. Zou, J.C. Fothergill, and S.W. Rowe, “The effect of water absorption
on the dielectric properties of epoxy nanocomposites”, IEEE Trans. Dielectr.
Electr. Insul., Vol. 15, pp. 106-117, 2008.
[22] T. Andritsch, Epoxy Based Nanocomposites for High Voltage DC
Applications. Synthesis, Dielectric Properties and Space Charge Dynamics,
PhD thesis, TU Delft, 2010.
[23] J.P. Eloundou, “Dipolar relaxations in an epoxy-amine system”, Eur.
Polym. J., Vol. 38, pp. 431-438, 2002.
[24] A. Vioux “Nonhydrolytic sol-gel routes to oxides”, Chem. Mater., Vol. 9,
pp. 2292-2299, 1997.
[25] Y. Miwa, A.R. Drews, and S. Schlick, “Detection of the direct effect of
clay on polymer dynamics: the case of spin-labeled poly(methyl acrylate)/clay
nanocomposites studied by ESR, XRD, and DSC”, Macromolecules, Vol. 39,
No.9, pp. 3304-3311, 2006.
[26] L. Chen, K. Zheng, X. Tian, K. Hu, R. Wang, C. Liu, Y. Li, and P. Cui,
“Double glass transitions and interfacial immobilized layer in in-situ-
synthesized poly(vinyl alcohol)/silica nanocomposites”, Macromolecules, Vol.
43, No.2, pp. 1076-1082, 2010.
[27] M. Hu, S. Shenogin, and P. Keblinski, “Molecular dynamics simulation
of interfacial thermal conductance between silicon and amorphous
polyethylene”, Appl. Phys. Lett., Vol. 91, 241910, 2007.
[28] N. McN. Alford, J. Breeze, and X. Aupi, “Dielectric loss of oxide single
crystals and polycrystalline analogous from 10 to 320 K”, J. Europ. Ceram.
Soc., Vol. 21, pp. 2605-2611, 2001.
[29] W. Kim, J-W. Bae, I-D. Choi, and Y-S. Kim, “Thermally conductive
EMC (Epoxy Molding Compound) for microelectronic encapsulation”,
Polym. Eng. Sci., Vol. 39, pp. 756-766, 1999.
[30] R.S. Carmichael, Handbook of Physical Properties of Rocks, CRC, Boca
Raton, FL., USA, 1982.
[31] C-Y. Hsieh and S-L. Chung, “High thermal conductivity epoxy molding
compound filled with a combustion synthesized AlN powder”, J. Appl. Polym.
Sci., Vol. 102, pp. 4734-4740, 2006.
[32] A.F. Junior and D.J. Shanafield, “Thermal conductivity of polycrystalline
aluminum nitride”, Ceramica, Vol. 50, pp. 247-253, 2004.
[33] K. Ishimoto, T. Tanaka, Y. Ohki, Y. Sekiguchi, Y. Murata, and M.
Gosyowaki, “Comparison of dielectric properties of low-density
polyethylene/MgO composites with different size fillers”, IEEE Conf. Electr.
Insul. Dielect. Phenom., Québec City, Canada, pp. 208-211, 2008.
[34] J. Kriegesmann, Technische Keramische Werkstoffe, Deutsche
Keramische Gesellschaft, 1992.
[35] J. Zeng, R. Fu, Y. Shen, H. He, and X. Song, “High thermal conductivity
epoxy molding compound with thermal conductivity pathway”, J. Appl.
Polym. Sci., Vol. 113, pp. 2117-2125, 2009.
[36] K.C. Yung and H. Liem, “Enhanced thermal conductivity of boron
nitride epoxy-matrix composite through multi-modal particle size mixing”, J.
Appl. Polym. Sci., Vol. 106, pp. 3587-3591, 2007.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 19, No. 1; February 2012
[37] Z-L. Hou, M-S. Cao, J. Yuan, X-Y. Fang and X-L. Shi, “High-
temperature conductance loss dominated defect level in h-BN:
experiments and first principles calculations”, J. Appl. Phys., Vol. 105,
076103, 2009.
[38] T. Andritsch, R. Kochetov, Y.T. Gebrekiros, P.H.F. Morshuis, and
J.J. Smit, “Short term DC breakdown strength in epoxy based BN nano-
and microcomposites”, IEEE 10th Int’l. Conf. Solid Dielectrics,
Potsdam, Germany, pp. 179-182, 2010.
[39] M. Roy, J.K. Nelson, R.K. MacCrone, L.S. Schadler, C.W. Reed,
R. Keefe, and W. Zenger, “Polymer nanocomposite dielectrics – the
role of the interface”, IEEE Trans. Dielectr. Electr. Insul., Vol. 12, pp.
629-643, 2005.
[40] T.J. Lewis, “Interfaces: nanometric dielectrics”, J. Phys. D: Appl.
Phys., Vol. 38, pp. 202-212, 2005.
[41] T. Tanaka, M. Kozako, N. Fuse, and Y. Ohki, “Proposal of a multi-
core model for polymer nanocomposite dielectrics”, IEEE Trans.
Dielectr. Electr. Insul., Vol. 12, pp. 669-681, 2005.
[42] S. Rätzke and J. Kindersberger, “Role of interphase on the
resistance to high-voltage arcing, on tracking and erosion of
silicone/SiO 2 nanocomposites”, IEEE Trans. Dielectr. Electr. Insul.,
Vol. 17, pp. 607-614, 2010.
[43] M.G. Todd and F.G. Shi, “Complex permittivity of composite
systems: a comprehensive interphase approach”, IEEE Trans. Dielectr.
Electr. Insul., Vol. 12, pp. 601-611, 2005.
[44] R. Kotsilkova, Thermoset Nanocomposites for Engineering
Applications, Smithers Rapra Technology, 2007.
[45] J.K. Nelson, Dielectric Polymer Nanocomposites, Springer, 2010.
[46] T. Andritsch, R. Kochetov, P.H.F. Morshuis, and J.J. Smit,
“Comparison of the dielectric behavior of epoxy based Al2 O 3 - and SiO 2
- nanocomposites”, Proc. Nordic Insul. Sym., Gothenburg, Sweden, pp.
137-140, 2009.
[47] H. Looyenga, “Dielectric constants of heterogeneous mixtures”,
Physica 31, pp. 401-406, 1965.
[48] J.I. Hong, P. Winberg, L.S. Schadler, and R.W. Siegel, “Dielectric
properties of zinc oxide/low density polyethylene nanocomposites”,
Mater. Lett., Vol. 59, pp. 474-476, 2005.
[49] S-H. Xie, B-K. Zhu, J-B. Lo, X-Z. Wei, and Z-K. Xu, “Preparation
and properties of polyimide/aluminum nitride composites”, Polym.
Test., Vol. 23, pp. 797-804, 2004.
[50] A.S. Vaughan, “On the dielectric response of nanofilled
materials”, unpublished.
Roman Kochetov was born in 1984 in Kazakhstan
(USSR). He received the B.Sc. and M.Sc. degrees in
technique and technology (microelectronics and
semiconductors) from Saint-Petersburg Electrotechnical
University with distinction in 2005 and 2007,
respectively. He participated in exchange two-diploma
program and received M.Sc. degree in electrical
engineering (technical physics) from Lappeenranta
University of Technology in 2007. His current Ph.D.
degree research project nanoPOWER has
the aim of designing and assessment of new, environmental friendly dielectric
materials, using nanotechnology, to realize insulation components with
improved thermal conductivity and electrical breakdown strength. His main
research interests are focused on the nanostructured systems, thermal
conductivity of polymer composites and dielectric spectroscopy of electrical
insulation materials.
117
Thomas Andritsch was born on the 17 August 1980 in
Innsbruck, Austria. He received his diploma in
electrical engineering from the Graz University of
Technology in 2006, specializing in energy
management technology. He started his Ph.D. degree
research project on nanodielectrics at the TU Delft in
2006. The goal of this project was the improvement of
insulation material for DC applications outside the
energy sector, e.g. medical X-ray systems, radar or
satellite parts. Nanoparticles are used to increase the dc
breakdown strength, thermal conductivity or reduce space charge density. He
finished his Ph.D. thesis with the title "Epoxy Based Nanocomposites for High
Voltage DC Applications - Synthesis, Dielectric Properties and Space Charge
Dynamics" in 2010 and is currently as a post-doctoral researcher at the HTM
group.
Peter H.F. Morshuis (M’95) was born in The Hague, The
Netherlands, on 23 December 1959. In 1986, he received
the Master degree in electrical engineering from Delft
University of Technology.
Between 1986 and 1988, he was involved in studies for
NKF Kabel on the effects of defects on the lifetime of high
voltage cables and in studies on new cable concepts. Since
1988, he has been a staff member of the High Voltage
Group at Delft University of Technology where he was
awarded the Ph.D. degree in electrical engineering in 1993
on the topic of ultrawideband electrical and optical studies of partial discharges in
solid dielectrics. In 1998, he was a visiting Professor at the University of Bologna.
Since 1999, he is an Associate Professor in High Voltage Engineering at Delft
University of Technology and he is involved in teaching first year students in the field
of electricity and magnetism and M.Sc. students in the field of high voltage dc.
His most important fields of interest are HVDC (materials and systems), space charge,
partial discharge, and aging of electrical insulation. He is involved in a number of
CIGRÉ activities and is an Associate Editor of the IEEE Transactions on Dielectrics
and Electrical Insulation.
Johan J. Smit (M’97), graduated in 1974 in experimental
physics and in 1979 he received the Ph.D. degree at the State
University of Leiden for his research in electro-magnetism
on behalf of the National Science Foundation. He was
employed in different research/management functions
during two decades at KEMA’s testing, consultancy and
engineering company, Arnhem NL. Until 2003 he has been
supervisory board member of the power grid company of
South Holland. Since 1996 he is a head of the High Voltage
Technology and Management group at the Delft University
of technology. In 2004 he has become Director of Education in Electrical Engineering
of the faculty of Electrical Engineering, Mathematics and Computer Science. In 2003
he has been the general chairman of the XIII Conference ISH, “International
Symposium on High Voltage Engineering” in Delft and he is currently chairing the
ISH steering committee. He is chairman of the executive board of the KSANDR,
organization for Knowledge Sharing and Research in the field of energy
infrastructures. He represents the Netherlands in the Technical Committee of IEC 98
on Electrical Insulation Systems. He performs and has performed several functions in
CIGRÉ, in particular the chairmanship of Study Committee D1, Materials and
Emerging Technologies. In September 2004 he became honorary member of CIGRÉ,
Paris.