DC Current in Nanosilica-based Polyethylene
Nanocomposites
Yan Wang,
Zhiqiang Xu, George Chen and Alun Vaughan
The Tony Davies High Voltage Laboratory, University of Southampton
Southampton, United Kingdom
yw14g13@soton.ac.uk
Abstract – In the present paper, DC current of polyethylene
nanocomposites containing different nanosilica loading ratios,
either the untreated or the surface treated using the
trimethoxy(propyl)silane coupling agent has been investigated.
TGA was used to identify the true loading concentration in the
samples and the nanofillers dispersion was studied using SEM. A
range of electric field from 10 kV/mm to 50 kV/mm were applied
It has been found that two dynamic processes are involved in the
current observed over a period of 3 hours. The initial process was
dominated by trap filling where a decreasing current versus time
was observed. The second process was related to the transient
space charge limited current in the unfilled sample but taken
over by the effect of nanofillers that shows a continuous current
increase versus time rather than peaking. The exact mechanism
responsible for the increasing current is not known yet. The
influence of surface treatment of the nanofillers on the current is
significant, resulting in a lower current comparing with the
untreated samples. A current dip has been observed for samples
with lower loading concentration, supporting the deep trap
concept.
Keywords—nanodielectrics; nanosilica; conductivity;
I. I
NTRODUCTION
For meeting the demand of requirements for future
insulation materials, the research has been focused on finding a
high performance and multifunction material by loading the
nanometer-sized particles, which is currently termed as
nanodielectrics or nanocomposites. During past 20 years, many
experimental results have shown that some electrical properties
of the nanocomposites such as discharge resistance and space
charge suppression can be improved [1]. Despite this, the
current achievement is still far away from the practical
application because it is hard to find materials that can
effectively operate under high voltage but at the same time
offer excellent thermal and mechanical properties. The space
charge reduction in nanodielectrics has frequently been
reported but the related mechanisms are not fully understood.
In addition, the measurement of the DC conductivity in
nanodielectrics also does not obtain consistent results, which
leads to difficulty in mechanism explanation [1, 2].
Nevertheless, connecting the result of space charge behavior
associated with DC conductivity will be a very informative
approach to further investigate the interaction between the
nanoparticles and a host polymer [3, 4]. This is because the
results of DC conductivity can directly embody the space
charge accumulation and movement in nanodielectrics [4]. It
has also been reported that after corona charged, the surface
potential decay of nanocomposites when in higher loading
ratios is faster than the pure polymer [5].
In this paper, the current which can be used to compare the
conductivity of different insulation materials is measured after
applying constant electric fields from 10kV/mm to 50kV/mm
at an interval of 10kV/mm. The current measurement is
sensitive to the nanofillers loading ratio [4], possibly to the
dispersion and the aggregation of nanofillers in the composites.
However, the latter is difficult to attain and evaluate [6-8]. It is
feasible that some of the nanofillers can be lost during the
sample preparation, resulting in lower loading level. One way
that can be used to obtain the useful information about real
loading is to utilize the thermogravimetric analysis (TGA) [6, 9
- 10]. Furthermore, it has been widely reported that the
electrical properties of nanocomposites is sensitive to the
absorbed water [6-11]. The possible effect on DC conductivity
of the nanocomposites has been discussed.
II. E
XPERIMENTAL
D
ETAILS
R
EFERENCE
A. Specimen Preparation
The nanosilica powder was obtained from Sigma-Aldrich,
and the range of its size is from 10nm to 20nm. Four weight
percentages including 0.5wt%, 2wt%, 5wt% and 10wt% were
applied in this investigation. The functionalization used was
the trimethoxy(propyl)silane for each loading ratio via an
anhydrous route [12]. The nanocomposites without surface
treated referred as untreated while those after treatment
referred as C3-treated. The host polymer utilized was blend
20% high-density polyethylene (HDPE) grade Rigidex
HD5813A, obtained from BP Chemical with 80% low-density
polyethylene (LDPE) grade LD100BW, obtained from
ExxonMobil Chemicals. The control group without the
addition of nanofillers is referred as unfilled. The required
amount of nanofillers was dispersed in xylene with probe
sonication by applying Hielsher UP200S probe sonicator for 1
hour. Each sample was dried by using a vacuum oven at 60
℃
for at least 72 hours [4]. The required specimen thickness for
conductivity test was ~120μm and this was achieved by
utilizing a heated hydraulic large press at 150
℃
. Once
removed from the press, the specimen was placed into an oil
bath going through isothermal crystallization at 115
℃
for 1
hour. After that, all samples were stored in a desiccator. The
DC conductivity measurements were performed at 19±3
℃
and
45-65% RH.
515
2015 Annual Report Conference on Electrical Insulation and Dielectric Phenomena
978-1-4673-7498-9/15/$31.00 ©2015 IEEE
B. Specimen Characterisation
To observe the dispersion/distribution of nanofillers in
polyethylene blends, SEM was carried out. The emphasis in the
present study was placed on the influence of surface treatment
on the nanofillers dispersion. The specimens from each type
were etched before imaged using scanning electron microscopy
(SEM), the etching procedure was performed based on a
standard permanganic reagent [6, 7]. After etching processing,
all sample was put on the aluminium stub and sputter-coated
with gold. A Jeol JSM6500F high resolution FEG-SEM was
used to measure all sample at 15kV.
Perkin Elmer Pryis 1 TGA system was employed and 5mg
sample tested in the ambient environment. The heating rate is
20
℃
/min from 100
℃
to 670
℃
. The conductivity
measurements were carried on a Keithley 6487 picoammeter
and the electrode with a diameter of 20 mm was applied. The
sample for conductivity test was gold sputtered to achieve
better electrical contact. A high voltage dc supply was used to
provide the required electric field and the current flowing
through the sample was recorded every 5 seconds for 3 hours.
All the current measurements were carried out at room
temperature.
III. R
ESULTS AND
A
NALYSIS
A. SEM images
Fig. 1 shows that the selected images display the
influence of C3-treatment on the nanofillers dispersion in
polyethylene blends.
(a) (b)
(c) (d)
Fig. 1. SEM micrograph of (a) 2% untreated; (b) 2% C3-
treated; (c) 10wt% untreated; (d) 10wt% C3-treatd
Nanofillers dispersion in the untreated sample is generally
poor with significant nanofillers aggregation as shown in Fig.
1 (a) and (c). It can also be seen that although one still
observes aggregation in the C3-treated sample, the surface
treatment does improve the nanofillers dispersion especially
for samples with higher loading ratios. Shells around the
nanofillers have been noticed in some of the images in both
C3-treated and untreated samples. It is most likely caused by
the artifacts introduced by the etching process.
B. TGA
(a) (b)
(c) (d)
Fig. 2. TGA results of nanocomposites (a) 0.5wt% (b) 2wt% (c)
5wt% (d) 10wt%.
TABLE I. COMPARISON OF THE MEASURED CONCENTRATION AND
NOMINATED LOADING CONCENTRATION IN NANOCOMPOSITE
SAMPLES
Sample
Real concentration
（
%
）
Unfilled
-0.07 ±0.02
0.5wt% untreated
0.57 ± 0.04
0.5wt% C3-treated
0.21±0.05
2wt% untreated
1.74±0.18
2wt% C3-treated
2.36±0.21
5wt% untreated
4.96±0.12
5wt% C3-treated
4.56 ±0.23
10wt% untreated
7.59±0.7
10wt% C3-treated
6.8 ±0.8
The TGA results for all the samples are shown in Fig. 2.
Basing on the weight loss at high temperature, the true amount
of nanofillers in each sample can be estimated as shown in
Table I. Objectively, considering the imperfection of specimen
preparation, the error between the true and believed value of
nanofillers is expected. However, the water absorbed in the
nanopowder should be taken into account, which may impact
on the weight measured [6].
As mentioned earlier, the specimen preparation for
treated nanofillers was proceeded via an anhydrous route.
Hence, the nanosilica may already absorb some moisture in
the storage. The moisture effect can clearly be seen for 10wt%
nanocomposites sample and weight loss at low temperature
range as indicated by the arrow in Fig. 2 (d). However, the
onsite temperature of other samples to lose weight is around
250
℃
. The weight reduction in 10wt% untreated
nanocomposites is 1.8wt% below 250
℃
while for 10wt% C3
treated sample is 0.8wt%, which is relative large comparing
with that in [6]. Although the weight loss in other loading ratio
nanocomposites from 100
℃
to 250
℃
is below 0.02wt%, it
can not be certain that these nanocomposites are free of
516
moisture. This is because moisture bound on the surface of
nanosilica needs at least 400
℃
to remove [9].
C. DC Current
A typical current versus time at an applied electric field of
30 kV/mm is shown in Fig. 3 for various nanocomposite
samples, including unfilled one for comparison. It is expected
that the current versus time for a typical dielectric material
show a decrease initially (called absorption current) followed
by a steady current (called conduction current). The duration
for the former process varies depending on many factors such
as the material, the applied electric field and temperature etc.
It has been widely reported [13 -14] that at high electric fields
the latter process may experience a dynamic period before
becoming stable. One of the candidates for the dynamic
process is the transient space charge limited current (SCLC)
that has been observed by many researchers. It has been
demonstrated that the transient space charge peak is related to
the charge packet observed in polyethylene [15]. From Fig. 3,
it is clear to identify two processes in operation. As expected,
the absorption current that decreases with time is observed for
all samples. For untreated nanofillers, it is noticed that the
magnitude of the absorption current increases with the
nanofillers loading ratio in the nanocomposite samples. The
magnitude of the current is higher when compared with the
unfilled sample except the one with a lowest loading ratio of
0.5% as shown in Fig. 3 (a).
(a)
(b)
Fig. 3. The currents versus time for all nanocomposites at an
applied electric field of 30kV/mm (a) untreated (b) C3-treated.
On the other hand, the dynamic second process only occurs
in some of the samples where high loading nanofillers are
involved such as 2%, 5% and 10% samples for untreated
nanocomposites. Rather than a steady conduction current, the
current that increases with time is observed. Furthermore, it is
also obvious that surface treatment on nanofillers plays a role
in the observed dynamic process, the increase in current is not
as big as the one observed with untreated.
Fig. 4 shows the effect of the applied electric field on the
current versus time in 5% nanofillers composites. For
comparison, the current versus time for unfilled samples are
also shown in Fig. 4 (a). The flat shoulder in the middle
section can be attributed to transient SCLC and shifts towards
shorter time with the increasing applied field.
(a)
(b)
(c)
Fig. 4. The current versus time for 5wt% nanocomposites (a)
unfilled, (b) untreated, (c) C3-treated.
When comparing the results in Fig. 3 (a) and (b), the
influence of the surface treatment on the current behaviours
can be readily demonstrated. A lower current has been
observed for all the samples after C3-treated. In semi-
crystalline polymer, there are many defect sites that are treated
as traps. The absorption current observed can be considered as
trap filling process. In addition, it can be seen the absorption
current for 0.5% and 2% C3-treated nanocomposites is lower
than that from the unfilled sample. The inclusion of nanofillers
into the polymer has been recognised as the introduction of
deep traps. Recent work by Chen [16] indicates that charge
dynamics in the nanocomposites can be seriously affected by
the presence of deep traps. The capture of the injected charge
by the deep traps adjacent to the electrode will suppress further
charge injection, which will lead to a low conductivity. And
there are many other factors that can influence the trap depth
[16], including the surface treatment on the nanofillers. Deeper
traps may be formed with C3-treatment on the surface of the
nanofillers and this may explain why the absorption current of
517
2% C3-treated sample still shows a lower value than the
unfilled sample. Charge dynamics is also affected by the
concentration of deep traps [16]. If it is too high, the other
process may be initiated, such as tunnelling. Consequently, the
high current can be expected as shown in nanocomposites with
high loading ratio. At a lower applied electric field, it has
been noticed that the dynamic second process observed in
nanocomposites shows a similarity to the transient SCLC. For
the traditional transient SCLC, the peak shifts towards a shorter
time when the applied field increases. The flat shoulder in 2%
C3-treated sample may be related to the transient SCLC.
However, care must be excised for those observed in higher
loading concentrations. It can be seen the current keeps
increasing without any sign of peaking in the 3 hours testing
duration. It is possible that a new process that is a direct
consequence of the introduction of nanofillers in the polymer
may take over. A plan is in hand to perform a longer
observation for both the current and space charge dynamics.
In general, the higher the applied electric field, the higher the
current. When comparing the effect of surface treatment, the
current from the untreated nanocomposites is higher than the
current from the C3-treated nanocomposites. The reason for the
difference can be attributed to the deep traps resulting from the
C3-treatment as outlined in the previous section.
Due to dynamic processes involved in the nanocomposites
the current never settles down in the time duration used in the
present study, which causes some trouble in identifying the
effect of the applied electric field and loading concentration.
However, there is a period from 5 to 15 minutes where the
current is relatively stable. Fig. 5 illustrated the effect of
nanofillers concentration on the current measured at 10
minutes. It can be seen that the current experiences a dip at a
loading concentration of 0.5% for untreated nanocomposites
then show an increase in general. However, the current
becomes too high for 10% untreated and C3-treated samples.
From the general increasing in the current point of view for
high loading concentration, the moisture observed in TGA test
for 10% nanofillers samples may also contribute to this
extremely high current. For C3-treated nanocomposites, the
current dip extends to 2% loading concentration before it starts
to increase at higher loading concentration. It is worth noting
that the rate of current increment is higher for those samples
with high loading concentration.
(a) (b)
Fig. 5. The current versus nanofillers loading concentration
where the current taken at 10min (a) untreated, (b) C3-treated.
IV
C
ONCLUSION
The current of nanosilica filled polyethylene blends has been
studied under a range of the applied dc electric fields. The
following conclusions may be drawn:
The current versus time in nanocomposites has two dynamic
processes. Initial trap filling process leads to a decreasing
current and this is followed by the effect of nanofillers,
resulting in an increasing current versus time. This current
increases more severe in the samples with higher loading ratio
than the lower one. The effect of surface treatment of
nanofillers on the current is significant and a lower current has
been observed under the same condition when compared with
the untreated sample. A current dip occurs when nanofillers
loading in the sample is lower, supporting the concept of deep
trap intruded by nanofillers. Further research has been planned
to measure the current over a longer period at different
temperatures to identify the mechanism for the current increase.
R
EFERENCE
[1] T. Tanaka and T. Imai, "Advances in nanodielectric materials over the
past 50 years," Electrical Insulation Magazine, IEEE, vol. 29, pp. 10-23,
2013.
[2] T. Tanaka, "Dielectric nanocomposites with insulating properties,"
Dielectrics and Electrical Insulation, IEEE Transactions on, vol. 12, pp.
914-928, 2005.
[3] G. Chen, C. Zhang, and G. Stevens, "Space charge in LLDPE loaded
with nanoparticles," in Electrical Insulation and Dielectric Phenomena,
2007. CEIDP 2007. Annual Report - Conference on, 2007, pp. 275-278.
[4] K. Lau, A. Vaughan, G. Chen, I. Hosier, and A. Holt, "Absorption
current behaviour of polyethylene/silica nanocomposites," in Journal of
Physics: Conference Series, 2013, p. 012003.
[5] J. Zha, G. Chen, Z. Dang, and Y. Yin, "The influence of TiO 2
nanoparticle incorporation on surface potential decay of corona-resistant
polyimide nanocomposite films," Journal of Electrostatics, vol. 69, pp.
255-260, 2011.
[6] I. Hosier, M. Praeger, A. Vaughan, and S. Swingler, "Electrical
Properties Of Polymer Nano-composites Based On Oxide And Nitride
Fillers," 2015.
[7] I. Hosier, M. Praeger, A. Holt, A. Vaughan, and S. Swingler, "Effect of
water absorption on dielectric properties of nano-silica/polyethylene
composites," in 2014 IEEE Conference on Electrical Insulation and
Dielectric Phenomena (CEIDP), 2014.
[8] Y. Wang, G. Chen, and A. Vaughan, "Space charge dynamics in silica-
based polyethylene nanocomposites," in Electrical Insulation and
Dielectric Phenomena (CEIDP), 2014 IEEE Conference on, 2014, pp.
727-730.
[9] M. Praeger, I. Hosier, A. Vaughan, and S. Swingler, "The Effects Of
Surface Hydroxyl Groups In Polyethylene-Silica Nanocomposites,"
2015.
[10] T. M. Andritsch, Epoxy Based Nanodielectrics for High Voltage DC
Applications: Synthesis, Dielectric Properties and Space Charge
Dynamics: TU Delft, Delft University of Technology, 2010.
[11] H. Le, L. S. Schadler, and J. K. Nelson, "The influence of moisture on
the electrical properties of crosslinked polyethylene/silica
nanocomposites," Dielectrics and Electrical Insulation, IEEE
Transactions on, vol. 20, pp. 641-653, 2013.
[12] K. Lau, "Structure and electrical properties of silica-based polyethylene
nanocomposites," PhD thesis, University of Southampton, 2013.
[13] G. Chen, H.M. Banford, R.A. Fouracre and D.J. Tedford, “Electrical
conduction in low density polyethylene”, 3rd Int. Conf. Conduction and
Breakdown in Solid Dielectrics, pp. 277 -281, 1989.
[14] R. Patsch, “Space charge phenomena in polyethylene at high electric
fields”, J. Phys. D: Appl. Phys. Vol. 23, pp.1497-1505, 1990
[15] W. S. Lau and G. Chen, “Simultaneous space charge and conduction
current measurements in solid dielectrics under high dc electric field”,
CMD, Changwon Korea, 2006.
[16] G. Chen, S. T. Li and L. S. Zhong, “Space charge in nanodielectrics and
its impact on electrical perfroemance” ICPADM, Sydney, Austrialia,
2015.
518
