1526 P. Preetha and M. J.

Thomas: AC Breakdown Characteristics of Epoxy Nanocomposites

AC Breakdown Characteristics of Epoxy Nanocomposites

P. Preetha and M. Joy Thomas
Nanodielectrics Laboratory
Department of Electrical Engineering
Indian Institute of Science
Bangalore, 560 012, India

ABSTRACT
Experiments were conducted to measure the ac breakdown strength of epoxy alumina
nanocomposites with different filler loadings of 0.1, 1 and 5 wt%. The experiments
were performed as per the ASTM D 149 standard on samples of thickness 0.5 mm, 1
mm and 3 mm in order to study the effect of thickness on the ac breakdown strength of
epoxy nanocomposites. In the case of epoxy alumina nanocomposites it was observed
that the ac breakdown strength was marginally lower for 0.1 wt% and 1 wt% filler
loadings and then increased at 5 wt% filler loading as compared to the unfilled epoxy.
The Weibull shape parameter (β) increased with the addition of nanoparticles to epoxy
as well as with the increasing sample thickness for all the filler loadings considered.
DSC analysis was done to study the material properties at the filler resin interface in
order to understand the effect of the filler loading and thereby the influence of the
interface on the ac breakdown strength of epoxy nanocomposites. It was also observed
that the decrease in ac electric breakdown strength with an increase in sample
thickness follows an inverse power-law dependence. In addition, the ac breakdown
strength of epoxy silica nanocomposites have also been studied in order to understand
the influence of the filler type on the breakdown strength.
Index Terms — Nanocomposites, breakdown strength, epoxy, alumina, silica,
interface, inverse power law, thermal conductivity.

1 INTRODUCTION are being conducted to study the characteristics of nanofilled

polymer composites for different high voltage applications.
FOR safe and reliable operation of any electrical equipment, After the introductory paper published by Lewis [2] on
the insulating materials used in the equipment should have polymer nanodielectrics, researchers are actively involved in
excellent dielectric and thermo mechanical properties. studying the various electrical characteristics of several new
Electrical insulation in many high voltage power equipments polymer nanocomposites [3] formed by the addition of
is epoxy based and it could be subjected to different types of different fillers at different weight percentages. The studies on
electrical discharges when the equipment is in operation. electrical breakdown in epoxy nanocomposites are few and
These discharges produce gradual deterioration of the epoxy they report conflicting results. Zhe Li et al [4] conducted
insulation by the bombardment of electrons and ions generated experiments on epoxy alumina nanocomposites (5 wt %),
by the discharges. The products of ionization may also lead to microcomposites (60 wt %) and nano micro composite (2
chemical reaction with the insulation. Also the high wt%+ 60 wt%) using sphere sphere electrodes on 0.2 mm
temperature in the region of the discharges may lead to thick flat samples. It was observed that the breakdown
degradation of the insulation [1]. The combined effect of the strength of epoxy alumina nanocomposites is higher than
discharges along with the degradation induced by the unfilled epoxy as well as all the other epoxy composites
discharges will increase the losses and can finally result in the considered. In another study, Nelson and Hu [5] show that the
electrical failure of the insulation system. Thus electrical 60 Hz ac electric strength in epoxy titania nanocomposites is
breakdown strength in such polymers is an important limiting marginally higher than base epoxy and significantly higher
phenomenon in most electrical devices and components. In than microcomposites. In addition, the Weibull shape
order to meet certain functional requirements such as higher parameter β also shows an increase for the nanocomposites as
mechanical strength, better heat handling capability etc., most compared to unfilled samples. On the other hand, it was
of the polymeric insulations used in high voltage applications reported by Imai et al [6] that epoxy nanocomposites with 5
use micron sized fillers. With the advent of nanotechnology wt% filler loading of titania and layered silicate had
and the easy availability of nanosized fillers, various studies marginally lower breakdown strength than unfilled epoxy. The
same study also reports a higher value of breakdown strength
Manuscript received on 26 November 2010, in final form 27 April 2011. for epoxy silica nanocomposites.

1070-9878/11/$25.00 © 2011 IEEE

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 18, No. 5; October 2011 1527

The above discussions clearly show that there are conflicting 40 oC for one hour to remove trapped air bubbles and
results published by different authors on the breakdown moisture. Similarly, the alumina nanoparticles were vacuum
strength of epoxy nanocomposites with few of the groups dried at 150 oC for at least 24 hours before using it for
reporting lower breakdown strengths as compared to the processing. During the sample preparation, care was taken
unfilled epoxy. Thus the breakdown phenomena in polymer such that the nanoparticles are exposed to the atmosphere only
nanocomposites is not yet clearly understood and since this is for a very short duration so that they do not absorb moisture.
a nascent area, an attempt has been made in this paper to study The epoxy nanocomposite samples were prepared by two
the breakdown behavior of epoxy nanocomposites. different types of processing techniques – mechanical mixing
The area and thickness dependence of the insulation on the in a completely closed high shear mechanical mixer at a speed
dc dielectric breakdown field have been carried out by Diaham of 700 rpm for 2 minutes which is followed by ultrasonic
et al [7] using the Weibull statistical model of failure for agitation at a frequency of 24 kHz for 1 hour. Then the
polyimide thin films in the temperature range from 25 to 400 hardener is added to the particle-resin mixture and mixed
o
C. It is observed that the Weibull scale parameter α showed a thoroughly and then evacuated to remove the air bubbles
decrease with increasing sample thickness and the shape trapped inside. The mixed epoxy particle mixture was casted
parameter β increased with an increase in the sample in a mould at 60 oC for 4 hours, to prepare samples of 75 mm
thickness. Yilamz and Kalenderli [8] have reported the effect diameter and 0.5 mm, 1 mm and 3 mm thicknesses. The
of sample thickness on the ac breakdown strength in thin samples prepared are then kept in a desiccator under vacuum
polyester films of thickness in the range 0.012 to 0.2 mm. It for at least 24 hours to release residual mechanical stresses if
was observed that the electric strength increased with reducing any before carrying out the ac breakdown studies.
thickness. The effects of specimen size and filler content on
2.3 SAMPLES PREPARED FOR THE PRESENT WORK
the electrical-impulse strength of epoxy polymer have also
been reported by Richard Lovell [9]. It was observed that the The alumina and silica filler loadings for the samples to be
electrical impulse strength decreases with increasing thickness prepared for the present study were arrived at by calculating
as well as with increasing filler content. The field the interparticle distance and the surface area per unit volume
enhancement is reported to be due to the shape and orientation using the equation given by Tanaka et al [11], and are shown
of the particles, the mismatch in the permittivity between the in Table 1 and Table 2 respectively. It is observed that the
filler and the polymer and the voids present near the fillers. interparticle distance decreases and the surface area increases
This field enhancement led to reduced electric field strength in with increase in the filler loading. It is observed that with
the case of filled polymers. increase in alumina filler loading to 5 wt%, the interparticle
Even though there are some studies reporting the thickness distance becomes less than 100 nm. So in order to study the
dependence of the breakdown strength of unfilled epoxy and effect of the interface on the ac breakdown strength, epoxy
epoxy microcomposites, no such data is available for epoxy alumina nanocomposites with filler loadings of 0.1, 1, 5, 10
nanocomposites. So in this study, the short time ac breakdown and 15 wt% and epoxy silica nanocomposites with filler
strength of unfilled epoxy and epoxy alumina nanocomposites loadings of 0.1, 1 and 5 wt% were prepared.
were measured to understand the effect of thickness on the Table 1. Interparticle distance and surface area of fillers in epoxy alumina
breakdown strength by performing experiments for three different nanocomposites
sample thicknesses of 0.5, 1 and 3 mm. The effect of the Interparticle Surface area (km2
Sl. No. wt%
nanoparticle-epoxy interface was studied by measuring the glass distance (nm) /m3)
transition temperature of the epoxy alumina nanocomposites. 1 0.1 447 0.044
2 1 244 0.438
2 PROCEDURE FOR SAMPLE PREPARATION 3 5 90 2.255
4 10 62 4.683
2.1 MATERIALS USED
5 15 48 7.303
The samples were prepared using Bisphenol-A epoxy resin
(CY 1300, density 1.16 g/cm3) along with hardener (HY 956, Table 2. Interparticle distance and surface area of fillers in epoxy silica
density 1.02 g/cm3) supplied by Huntsman. This was used as nanocomposites
the base polymer material. The fillers used in the present study Interparticle Surface area (km2
Sl. No. wt% distance (nm) /m3)
are commercially available Al2O3 and SiO2 fillers. The
average particle size (APS) of the alumina nanofillers is 40 nm 1 0.1 190 0.134
and density is 4 g/cm3. For silicon dioxide nanopowder APS is 2 1 78 1.346
20 nm and density is 2.6 g/cm3. Aluminium oxide particles of
3 5 37 6.882
average particle size 5 µm were used as microfillers. All the
fillers were supplied by Sigma-Aldrich, USA. 2.4 SEM STUDIES
2.2 NANOCOMPOSITE PREPARATION The dispersion of the fillers in the polymer matrix was
The epoxy nanocomposites were prepared in the laboratory analyzed using a scanning electron microscope. A good
by direct dispersion method following a protocol developed by dispersion of the fillers in the polymer matrix is obtained for
one of the authors to get the best possible dispersion [10]. the nanocomposites. The SEM image showing the dispersion
Initially, the epoxy resin and the hardener are both degassed at of the alumina fillers in the epoxy matrix is shown in Figure 1.

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
1528 P. Preetha and M. J. Thomas: AC Breakdown Characteristics of Epoxy Nanocomposites

the end curvature region of the electrode. The thickness of the

sample around the breakdown region was measured and the ac
breakdown stress was evaluated using the equation E=V/d,
where ‘V’ is the puncture voltage and ‘d’ is the thickness of
the sample in the region close to the puncture point.
3.2 DSC STUDIES
The glass-transition behaviour of epoxy alumina
nanocomposites has been investigated with differential
scanning calorimetry (DSC) technique using Mettler USA
make DSC equipment. A sample of approximately 15 mg was
sealed in a hermetic aluminium pan. Experiments were
conducted at a ramp rate of 10 °C/min, from the ambient
temperature to 150 °C, to obtain the curing heat-flow diagram
and thereby the glass transition temperature of the composite.
The glass transition temperature for epoxy alumina
Figure 1. SEM image of the epoxy alumina nanocomposite (5 wt%) nanocomposites is shown in Figure 3. In the present study the
Tg values obtained are less than the unfilled epoxy for all the
3 EXPERIMENTAL DETAILS filler loadings considered. The lowest value of Tg is at 0.1
wt% filler loading. With further increase in the filler loading
3.1 AC BREAKDOWN STUDIES the Tg value increases with an exception at 10 wt% filler
Experiments were performed on unfilled epoxy samples and loading.
epoxy alumina nanocomposite samples of 0.1, 1 and 5 wt% It has been reported that the interaction between the
filler loadings as per ASTM D 149 [12]. A set of 10 samples nanoparticle and the polymer chains leads to the formation of
for each filler loading and thickness were tested to evaluate an interfacial nanolayer around the nanoparticle and it is this
the breakdown voltage. The experiment was conducted for interfacial region that determines the glass transition of the
three different thicknesses (0.5 mm, 1 mm, and 3 mm). polymer nanocomposites [13, 14]. The possible explanation
Experiments at higher filler loadings of 10 wt% and 15 wt% for the Tg behavior obtained in the present study is based on
were conducted for samples having a thickness of 0.5 mm the dual nanolayer theory suggested by Tsagarapoulos et al
only. All tests were carried out at room temperature. The [15]. A representative model for the theory considered is the
electrodes used for the experiments were 25 mm in diameter dual layer model and the same is also explained by Santanu
with an edge radius of 3.2 mm. The experimental set up for and Joy Thomas [16]. According to this model the layer
the ac breakdown studies is as shown in Figure 2. The entire around the nanoparticle comprises of two regions, one tightly
bound region and the other, the loose polymer region. The
tightly bound region is the region around the nanoparticle
 where the nanoparticles are bonded by weak hydrogen bonds
to the polymer chain. The tightly bound region of the

Figure 2. Experimental setup for the ac breakdown studies

electrode assembly was immersed in a vessel containing

Figure 3. Glass transition temperature of epoxy alumina nanocomposites
transformer oil to avoid surface flashover. The experiment was
performed by applying an ac voltage to the high voltage nanocomposite reduces the mobility of the polymer chains. At
electrode at the rate of 0.5kV/s until the sample got punctured. low filler loadings the number of nanoparticles are smaller. So
With increase in voltage, bubbling was observed in the oil the tightly bound layer is less, leading to an increase in the
near the end of the electrodes at the curved region. Most of the mobile regions in the nanocomposites leading to a reduction in
punctures in the samples occurred through the region under Tg. With increase in the filler loading, the number of

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 18, No. 5; October 2011 1529

nanofillers increases. So the tightly bound region around the It is observed that the thermal conductivity of unfilled epoxy
nanoparticle increases. An increase in the immobile regions in is 0.137 W/m/K. With the addition of 0.1 wt% of alumina
the interface may lead to an increase in the Tg. nanofiller to epoxy, the thermal conductivity increases to
0.169 W/m/K. With further increase in the filler loading, the
3.3 THERMAL CONDUCTIVITY MEASUREMENTS
thermal conductivity shows an increasing trend and reaches
The thermal conductivity was measured using the Unitherm 0.206 W/m/K at a filler loading of 5 wt%. A similar trend in
model 2022 guarded heat flow thermal conductivity the thermal conductivity of unfilled epoxy by the addition of
measurement system of Anter Corporation, USA. The tests are alumina nanofiller was observed by Kochetov et al [19] and
in accordance with ASTM E 1530 standard [17]. A sample of Okasaki et al [20].
the material of 3 mm thickness and 50 mm diameter is held
under a uniform compressive load between two polished 4 RESULTS AND DISCUSSIONS
surfaces, each controlled at a different temperature. The lower
surface is part of a calibrated heat flow transducer. The heat Breakdown in a polymeric insulator results in narrow
flows from the upper surface, through the sample, to the lower conducting breakdown channel between the electrodes and is
surface, establishing an axial temperature gradient in the stack. always catastrophic in the sense that it is irreversible and
After reaching thermal equilibrium, the temperature difference destructive. All catastrophic breakdowns in solids are
across the sample is measured using the output from the heat electrically power driven and ultimately thermal in the sense
flow transducer. These values and the sample thickness are that the discharge track involves at least the melting and
then used to calculate the thermal conductivity. The probably the carbonization or vaporization of the dielectric.
temperature drop through the sample is measured with So the breakdown processes are subdivided in to
temperature sensors in the highly conductive metal surface electromechanical breakdown, thermal breakdown,
layers on either side of the sample. electronic breakdown, partial discharge and free volume
breakdown [21].
Electromechanical breakdown occurs due to the electrostatic
attraction of the electrodes which decreases the width of the
insulation by an amount depending on the Young’s modulus.
So if the applied voltage is the same, the field increases due to
the decrease in thickness. Mechanical crack propagation will
be initiated if the strain energy released due to electric stress is
higher than that required for the material deformation leading
to the breakdown. In thermal breakdown, electrical power
dissipation causes heating of the insulation to a temperature
above a critical value which results directly or indirectly in the
failure of the insulation. In electronic breakdown, the field
causes either the number or the energy of the electrons to
reach unstable magnitudes such that they rise catastrophically.
Ultimately this causes local destruction of the lattice. In partial
discharge breakdown, sparks occur within voids in the
Figure 4. Thermal conductivity of epoxy alumina nanocomposites
insulation causing degradation of the void walls. The
progressive deterioration of the dielectric occurs by the
The experimental and the calculated values of thermal bombardment of energetic electrons and ions generated by
conductivity based on the Agari and Uno model [18] for the discharges and the chemical action on the dielectric by the
epoxy alumina nanocomposites with different filler loadings products of ionization of the gas and high temperature in the
are plotted in Figure 4. The Agari and Uno model for the region of discharges. In free volume breakdown carriers are
prediction of the thermal conductivity is expressed as follows accelerated through spaces within low-density amorphous
log λc =V2C2 log λ2+ (1-V2) log(C1λ1) (1) regions, the energy thereby gained is lost through collision.

where λc is the thermal conductivity of the epoxy alumina 4.1 AC BREAKDOWN STUDIES
nanocomposites,V2 is the volume fraction of the filler (1.45% for The ac breakdown strength of unfilled epoxy and epoxy
5wt% filler loading, 0.029% for 0.1 wt% filler loading and 0.29 alumina nanocomposites with filler loadings of 0.1, 1 and 5 wt%
wt% for 1 wt% filler loading), λ2 is the thermal conductivity of and thicknesses of 0.5 mm, 1 mm and 3 mm were measured and
the alumina filler (30 W/m/K), λ1 is the thermal conductivity of are shown in Figure 5. It is observed that the ac breakdown
the polymer matrix (0.137 W/m/K) and C2 is the factor strength decreases for 0.1 wt% filler loading for all the
introduced to define the ability of forming continuous network of thicknesses considered. With an increase in the filler loading to 5
filler in the matrix. It is reported by Agari that the value of C2 lies wt%, the ac breakdown strength increases. It can be seen from
between 0 and 1. Considering that the preparation procedure of Figure 5 that the increase in the breakdown strength with increase
the composites can affect the crystallinity and thereby the thermal in the filler loading is more prominent at lower thickness of 0.5
conductivity, a constant C1 is introduced. The values of C1 and C2 mm. For 1 mm and 3 mm thick samples, the ac breakdown
are calculated from the experimental data. The values of C1 and strength increases with filler loading but not as much as in the
C2 are 1.2 and 0.99 respectively. case of 0.5 mm thick samples.

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
1530 P. Preetha and M. J. Thomas: AC Breakdown Characteristics of Epoxy Nanocomposites

90 0.5 mm
80 1 mm
3 mm
AC breakdown strength

70
60
(kV/mm)

50
40
30
20
10
0
0 0.1 1 5
Filler loading(wt%)
Figure 5. AC breakdown strength of epoxy alumina nanocomposites at
different filler loadings Figure 8. Weibull probability plot of epoxy alumina nanocomposites of 3 mm
Two- parameter Weibull distribution has been used to assess the thickness
breakdown behavior of the epoxy nanocomposites. The relevant The shape parameter and scale parameter are shown in Table 3
parameters of the distribution are the shape parameter β that and Table 4 respectively. It is observed from Table 3 that the
represents the inverse of data scatter and the scale parameter α shape parameter β increases with the addition of nanofillers to
that represents the breakdown strength at the cumulative failure epoxy. The β–parameter showed an increase with increasing
probability of 63.2% and is called the Weibull breakdown thickness for unfilled epoxy as well as epoxy alumina
strength. The Weibull probability plots for the samples with nanocomposites for all the filler loadings considered. With the
different thicknesses are shown in Figures 6 to 8. increase in the thickness of the samples, the breakdown strength
(α) decreases for both unfilled epoxy as well as epoxy
nanocomposites as shown in Table 4. It is also observed that the
decrease in the ac electric field strength with an increase in
sample thickness follows the inverse power-law dependence as
shown in Figure 9. The values of the power law coefficient are
given in Table 5. A similar trend for the Weibull shape and scale
parameters has also been observed by Diaham et al [7].

Table 3. Shape parameter (β) for different sample thickness of unfilled epoxy
and epoxy alumina nanocomposites

Sample Unfilled Epoxy alumina nanocomposites

thickness Epoxy
(mm) 0.1 wt% 1 wt% 5 wt%

0.5 12.23 12.65 16.07 15.28

Figure 6. Weibull probability plot of epoxy alumina nanocomposites of 1 15.6 16.95 16.2 16.21
0.5 mm thickness
3 18.08 18.48 18.16 18.78

Table 4. Scale parameter (α) for different sample thickness of unfilled epoxy
and epoxy alumina nanocomposites
Sample Unfilled Epoxy alumina nanocomposites
thickness Epoxy
(mm) 0.1 wt% 1 wt% 5 wt%
0.5 63.47 58.74 59.86 70.72
1 40.16 36.37 38.61 42.77
3 19.67 18.13 19.07 20.15

Table 5. Inverse power law coefficient (n) for epoxy alumina nanocomposites
Epoxy alumina nanocomposites
Unfilled
Epoxy 0.1 wt% 1 wt% 5 wt%
Figure 7. Weibull probability plot of epoxy alumina nanocomposites of 0.65 0.65 0.64 0.70
1 mm thickness

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 18, No. 5; October 2011 1531

loading to 5 wt% showed a further increase in the breakdown

strength above that of the unfilled epoxy. In this case the number of
nanoparticles is much more and it has been seen from Table 1 that
the interparticle distance is also less than 100 nm. So there is a
possibility of overlapping of the loose polymer regions in the
nanocomposites leading to a reduction in the loose polymer
regions. This along with a large number of fillers can obstruct the
discharge path as shown in Figure10.

Figure 9. Log-log plot of the ac breakdown strength of epoxy alumina

nanocomposites vs. thickness (a) At low nanofiller loading (b) At high nanofiller loading
4.2 EFFECT OF THE FILLER LOADING ON THE Figure 10. Possible nature of charge transfer mechanism in nanocomposites
BREAKDOWN STRENGTH at (a) low nanofiller loading and (b) at high nanofiller loading
When nanofillers are introduced in to the polymer matrix there is Another possibility for the increase in the ac breakdown
a change in the morphology of the polymer introduced by the strength could be the increase in the thermal conductivity of
interaction of the polymer with the nanofiller. It has been reported the nanocomposite at higher filler loading. This higher thermal
that the interaction between the nanoparticle and the polymer conductivity can cause better dissipation of the heat generated
chains leads to the formation of an interfacial nanolayer around the resulting in higher breakdown strength.
nanoparticle. A dual layer model has been proposed by Santanu In order to understand the effect of higher filler loading on
and Joy Thomas [16] for the nanocomposites and is discussed in the ac breakdown strength of epoxy alumina nanocomposites,
section 3.2. According to this model the layer around the experiments were conducted with filler loadings of 10 wt%
nanoparticle comprises of two regions, one tightly bound region and 15 wt% as well as epoxy alumina microcomposite of 10
and the other the loose polymer region. The tightly bound region is wt%. The Weibull probability plot for the breakdown strength
the region around the nanoparticle where the nanoparticles are is shown in Figure 11 for 0.5 mm thick samples.
bonded by hydrogen bonds to the polymer chain. So the mobility in It was observed that with increase in the filler loading above
this region is restricted. With 0.1 wt% filler loading, it was 5 wt%, the breakdown strength is less than the unfilled epoxy.
observed that the ac breakdown strength decreases as compared to Also it has been observed that the breakdown strength is
unfilled epoxy. In this case the number of particles is less and the minimum for epoxy alumina microcomposite of 10 wt% filler
interparticle distance is more. So the volume fraction of the loose loading. With increase in the nanofiller loading to 10 and 15
polymer layer is large. When high electric stress is applied, the wt%, the reduction in the breakdown strength may be due to
transfer of charge carriers between the electrodes becomes easy the overlapping of the tightly bound polymer regions in the
through this loose polymer nanolayer [22]. Also the filler particles interface as the interparticle distances are comparable to the
have a permittivity higher than epoxy. So there will be field filler diameter as shown in Table 1. The interphase region
enhancement at these particles leading to partial discharges. It has close to the nanoparticle is reported to be conductive by Zhang
been reported earlier that the field enhancement by filler particles et al [23]. As the conductive interphase region tends to
does not give a complete explanation of the reduced breakdown overlap, the neat epoxy region is reduced leading to an easier
strength observed. It is probable that, despite the care in preparation conducting path for the charge transfer and thereby a reduction
of the filled materials, there may be some voids present near the in the breakdown strength. The variations in the ac breakdown
filler particles which lead to further field enhancement [9]. The strength at different filler loadings for 0.5 mm thick epoxy
combined effect of the surface discharges, the localized partial alumina nanocomposite samples are shown in Figure 12. It is
discharges at the particle sites and the easy transfer of charge observed that the trend in the ac breakdown strength is similar
carriers through the loose polymer layer can lead to the final to the one observed by Le Wang et al [24].
breakdown of the polymer. In the case of epoxy alumina microcomposite, the number of
With increase in filler loading to 1 wt% it has been observed that fillers will be less than that of the nanocomposites for the
the ac breakdown strength is less than the unfilled epoxy, but is same filler loading and so the interparticle distance will be
slightly greater than that for the 0.1 wt % epoxy alumina larger in this case. This causes the discharge to penetrate
nanocomposite. Even though the particles act as defects, in this through a larger bulk of the polymer. Once the bulk epoxy
case the effect of the interface also comes into picture. The number material is degraded, the microparticles get exposed to
of particles is large as compared to 0.1 wt%. This large number of discharges. So there is a possibility of field enhancement at
fillers can act as barriers for the penetration of the charge carriers these particle extremes which can lead to enhanced discharges
through the bulk of the sample. A further increase in the filler resulting in reduced breakdown strength.

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
1532 P. Preetha and M. J. Thomas: AC Breakdown Characteristics of Epoxy Nanocomposites

aromatic polymers have also been reported by Bjellheim and

Helgee [25,26].
It has been reported by Peek [27] that the gradient at the
edges of the electrodes used for the experiment, even when
these are rounded off is generally higher than the average
gradient V/d, and this edge effect is different for different
thicknesses of the insulation. This edge effect becomes
relatively greater as the thickness of the insulation is
increased. This may lead to a reduction in the ac breakdown
strength of the samples with increased thickness. The increase
in the breakdown strength of thin samples may also be
attributed to the better heat dissipation in these thin samples.
With increase in the thickness, the heat generated becomes
greater than the heat dissipated leading to a thermal
breakdown of the insulation.
4.4 EFFECT OF THE FILLER TYPE ON THE
Figure 11. AC breakdown strength of epoxy alumina nano/microcomposites
BREAKDOWN STRENGTH
of 0.5 mm sample thickness In order to understand the effect of the fillers on the ac
breakdown strength, experiments were also conducted with
90 epoxy silica nanocomposites of 0.5 mm thickness. The Weibull
AC breakdown strength (kV/mm)

80 probability plot for epoxy silica nanocomposites is shown in

70 Figure 13. In the case of epoxy silica nanocomposites it was
observed that the ac breakdown strength increases with 0.1 and 1
60
wt% filler loadings and decreases slightly for a higher filler
50 loading of 5 wt%. The increase in the ac breakdown strength at
40 0.1 wt% and 1 wt% when compared to the epoxy alumina
30
nanocomposites of the same filler loading could be attributed to
the difference in permittivity between the alumina and silica
20 nanoparticles. The permittivity of silica nanoparticles is lower as
10 compared to alumina nanoparticles. So the enhancement in the
0 field at the nanoparticles will be less in the case of epoxy silica
0 5 10 15 nanocomposites leading to higher breakdown strength. Also the
Filler loading (wt%) number of nanoparticles for the same filler loading is more for
silica as the nanoparticle size is around 20 nm as compared to
Figure 12. AC breakdown strength of epoxy alumina nanocomposites of 0.5 around 40 nm for alumina nanoparticles. The presence of large
mm thick samples at different filler loadings
number of nanoparticles can obstruct the penetration of the
4.3 EFFECT OF SAMPLE THICKNESS ON THE AC charged carriers through the bulk of the epoxy silica
BREAKDOWN STRENGTH nanocomposite sample leading to an increase in the breakdown
strength. But at higher filler loading of 5 wt%, the ac electric
With increase in the thickness, the ac breakdown strength breakdown strength decreased slightly as compared to unfilled
decreases for unfilled epoxy and epoxy alumina nanocomposites. epoxy. It has been reported by Qi Wang et al [28] that the
By plotting the ac breakdown strength (α) values vs. the presence of nanoparticles enhances the ionization process inside
sample thickness in a log -log plot, good linear correlation was
observed for unfilled epoxy as well as epoxy alumina
nanocomposites. From this, one can conclude that there is an
inverse power-law relationship between the ac breakdown
field and the sample thickness for unfilled epoxy and epoxy
alumina nanocomposites. That is the ac breakdown strength
obeys the inverse power law given by
E α kd (1)
where ‘d’ is the sample thickness in mm and ‘k’ is the
arbitrary thickness parameter. In logarithmic form, equation
(1) can be expressed as
log E α log k log d (2)
The value of n obtained lies in the range 0.64 to 0.70 as
shown in Table 4.
A similar exponential relationship between the short time ac
breakdown strength and the thickness ‘d’ of the films of five Figure 13. AC breakdown strength of epoxy silica nanocomposites.

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 18, No. 5; October 2011 1533

the sample at higher filler loadings and that the type of the filler IEEE Conf. Electr. Insul. Dielectr. Phenomena (CEIDP), Kansas City,
Missouri, USA, pp. 306-309, 2006.
has a significant influence on the space charge accumulations. It [7] S. Diaham, S. Zelmat, M. L. Locatelli, S. Dinculescu, M. Decup and T.
is also reported that a higher content of nano sized silica filler (5 Lebey, “Dielectric Breakdown of Polyimide Films: Area, Thickness and
wt%) in epoxy will lead to more complicated charge Temperature Dependence”, IEEE Trans. Dielectr. Electr. Insul., Vol. 17,
accumulation. So this may be one of the reasons for the decrease pp. 18-27, 2010.
[8] G. Yilamz and O. Kalenderli, “The Effect of Thickness and Area on the
in the breakdown strength of epoxy silica nanocomposites at Electric Strength of Thin Dielectric Films”, IEEE Int’l. Sympos. Electr.
higher filler loading. Insul., Montreal, Quebec, Canada, pp. 478-481, 1996.
[9] Richard Lovell, “The Effect of Specimen Size on the Electric
5 CONCLUSIONS Breakdown of Unfilled and Filled Epoxy Polymers”, IEEE Trans. Electr.
Insul., Vol. 11, pp. 110-114, 1976.
AC dielectric strength of epoxy alumina nanocomposites [10] Santanu Singha and M. Joy Thomas, “Polymer Composite/
shows a slight reduction with filler loadings of 0.1 wt% and 1 Nanocomposite Processing and its Effect on the Electrical Properties”
IEEE Conf. Electr. Insul. Dielectr. Phenomena (CEIDP), Kansas city,
wt%. However the AC breakdown strength is higher for 5wt% Missouri, USA, pp. 557-560, 2006.
alumina filler loading as compared to the unfilled epoxy. The [11] T. Tanaka, M. Kozako, N. Fuse, Y. Ohki, “Proposal of a Multi-core
breakdown strength of epoxy alumina microcomposite was the Model for Polymer Nanocomposite Dielectrics”, IEEE Trans. Dielectr.
least as compared to the unfilled epoxy and epoxy alumina Electr. Insul., Vol.12, pp. 669-681, 2005.
[12] ASTM-D149-97a (reapproved 2004), Standard Test Method for
nanocomposites for all the filler loadings studied. It is also Dielectric Breakdown Voltage and Dielectric Strength of Solid
observed that the ac breakdown strength of nanocomposites is Electrical Insulating Materials at Commercial Power Frequencies.
lower with further increase in the filler loadings to 10 wt% and [13] A. M. Mayes, “Softer at the boundary”, Nature Materials, Vol. 4, pp.
15 wt%. Further studies are required to clarify the reason for 651-652, 2005.
[14] F. W. Starr, T. B. Schroder and S. C. Glotzer, “Effects of a nanoscopic
the reduction in the ac breakdown strength with increase in the filler on the structure and dynamics of a simulated polymer melt and the
filler loadings to 10 wt% and 15 wt%. The ac breakdown relationship to ultra thin films”, Phys. Rev. E, Vol. 64, pp. 021802 (1) –
strength is also found to be dependent on the thickness of the 021802 (5), 2001.
sample considered. With increase in the thickness, the [15] G. Tsagaropoulos and A. Eisenberg, “Dynamic Mechanical Study of the
Factors Affecting the Two Glass Transition Behavior of Filled
breakdown strength decreased for both unfilled epoxy and Polymers. Similarities and Differences with Random Ionomers”,
epoxy alumina nanocomposites. It was observed that the Macromolecules, Vol. 28, pp. 6067-6077, 1995.
breakdown strength of epoxy silica nanocomposites increased [16] S. Singha and M. Joy Thomas, “Dielectric Properties of Epoxy
at filler loadings of 0.1 wt% and 1 wt% and decreased slightly Nanocomposites”, IEEE Trans. Dielectr. Electr. Insul., Vol.15, pp. 12-
23, 2008.
at a filler loading of 5 wt%. [17] ASTM E 1530 (2004), “Standard Test Method for Evaluating the
The short term dielectric data presented here for the Resistance to Thermal Transmission of Materials by the Guarded Heat
nanodielectric materials can be considered as a parameter to Flow Meter Technique”.
check the quality of the material. In reality, the dielectric [18] Y. Agari and T. Uno, “Estimation on Thermal Conductivities of Filled
Polymers”, J. Appl. Polymer Sci., Vol. 32, pp. 5705-5712, 1986.
behaviour of the insulation system under long term aging tests [19] R. Kochetov, T. Andritsch, U. Lafont, P.H.F. Morshuis, and J. J. Smit,
simulating real operating conditions are of greater practical “Thermal Conductivity of Nano-filled Epoxy Systems”, IEEE Conf.
significance and needs to be carried out. Electr. Insul. Dielectr. Phenomena (CEIDP), Virginia Beach, USA, pp.
658-661, 2009.
[20] Y. Okasaki, M. Kozako, M. Hikita and T. Tanaka, “Effects of Addition
ACKNOWLEDGMENT of Nano-scale Alumina and Silica Fillers on Thermal Conductivity and
Dielectric Strength of Epoxy/Alumina Microcomposites”, Int’l. Conf.
The authors would like to thank Mr. Riaz Ahmed for the Solid Dielectr. (ICSD), Potsdam, Germany, July 4-9, pp. 1-4, 2010.
help in conducting the experimental work. [21] L. A. Dissado and J. C. Fothergill, Electrical Degradation and
Breakdown in Polymers, IEE Materials and Devices Series 9, Peter
Peregrinus Ltd, London, UK, 1992.
REFERENCES [22] Toshikatsu Tanaka, “Dielectric Nanocomposites with Insulating
[1] L. L. Alston, High Voltage Technology, Oxford University Press, Properties”, IEEE Trans. Dielectr. Electr. Insul., Vol. 12, pp. 914-928,
London, 1968. 2005.
[2] T. J. Lewis, “Nanometric Dielectrics”, IEEE Trans. Dielectr. Electr. [23] C. Zhang, M. Ralf and G. Stevens, “Preparation, Characterization and
Insul., Vol.1, pp. 812-825, 1994. Dielectric Properties of Epoxy and Polyethylene Nanocomposites”, IEEJ
[3] T. Tanaka, G. C. Montanari and R. Mulhaupt, “Polymer Trans. Fundamentals of Materials (Japan), Vol.126, pp. 1105-1111,
Nanocomposites as Dielectrics and Electrical Insulation–Prospective for 2006.
Processing Technologies, Material Characterisation and Future [24] Le Wang, Man Xu, Junqiang Feng, Xiaolong Cao, “Study on AC
Applications”, IEEE Trans. Dielectr. Electr. Insul., Vol.11, pp. 763-784, Breakdown Property of Nano-Ag/Epoxy Resin Composite”, IEEE Int’l.
2004. Conf. Properties and Applications of Dielectric Materials (ICPADM),
[4] Z. Li, K. Okamoto, Y. Ohki and T. Tanaka, “Role of Nano-Filler on Bali, Indonesia, pp. 163-166, 2006.
Partial Discharge Resistance and Dielectric Breakdown Strength of [25] P. Bjellheim and B. Helgee, “Aromatic Polyimides: Synthesis,
Micro-Al2O3/Epoxy Composites”, IEEE Int’l. Conf. Properties and Characterization, and Evaluation of Electric Strength”, J. Appl. Polymer
Applications of Dielectric Materials (ICPADM), China, pp. 753-756, Sci., Vol. 48, Issue 9, pp. 1587-1596, 2003.
2009. [26] B. Helgee and P. Bjellheim, “Electric Breakdown Strength of Aromatic
[5] J. K. Nelson and Y. Hu, “Electrical Properties of TiO2 Polymers- Dependence on Film Thickness and Chemical Structure”,
Nanocomposites”, IEEE Conf. Electr. Insul. Dielectr. Phenomena IEEE Trans. Electr. Insul., Vol. 26, pp. 1147-1152, 1991.
(CEIDP), Albuquerque, New Mexico, USA, October 19-22, pp. 719- [27] F. W. Peek, Dielectric phenomena in High Voltage Engineering,
722, 2003. McGraw Hill, New York, Third edition, 1929.
[6] T. Imai, F. Sawa, T. Ozaki, Y. Inoue, T. Shimizu and T. Tanaka, [28] Qi Wang, G. Chen and A. S. Alghamdi, “Influence of Nanofillers on
“Comparison of Insulation Breakdown Properties of Epoxy Electrical Characteristics of Epoxy Resin Insulation”, Int’l. Conf. Solid
Nanocomposites under Homogeneous and Divergent Electric Fields”, Dielectrics, Potsdam, Germany, July 4-9, pp. 1-4, 2010.

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.
1534 P. Preetha and M. J. Thomas: AC Breakdown Characteristics of Epoxy Nanocomposites

P. Preetha was born in Kerala, India in 1973. She M. Joy Thomas (S’85-M’95) was born in Kerala,
received the B.Tech. degree in electrical engineering India in 1961. He received the B.Tech. degree in
from N. S. S. College of Engineering, Palakkad, electrical engineering from the Institute of
India in 1995, PG diploma in thermal power plant Technology, BHU, Varanasi, India, the M.S. and the
engineering from RPTI Neyveli in 1997 and the Ph.D. degrees in Electrical Engineering from the
M.E. degree in high voltage engineering from Indian Indian Institute of Science, Bangalore, India.
Institute of Science, Bangalore, India in 2003. Presently, he is working as an Assistant Professor at
Presently she is doing her Ph.D. degree at the the Department of Electrical Engineering, Indian
Department of Electrical Engineering, Indian Institute of Science, Bangalore, India. He is also a
Institute of Science, Bangalore, India. Her areas of member of CIGRE. His areas of interest are
interest are aging and discharge resistant characteristics of nanocomposites, EHV/UHV power transmission, high voltage engineering, dielectrics and
and GIS. electrical insulation, condition monitoring and asset management of high
voltage power apparatus, nanodielectrics, biodielectrics, pulsed power
engineering, plasma science and technology, high power electromagnetics,
lightning, EMC and engineering pedagogy.

Authorized licensed use limited to: University of Manchester. Downloaded on December 15,2021 at 16:15:49 UTC from IEEE Xplore. Restrictions apply.