Download Novel Superdielectric Materials: Aqueous Salt Solution Saturated

materials
Article
Novel Superdielectric Materials: Aqueous Salt
Solution Saturated Fabric
Jonathan Phillips
Energy Academic Group, Naval Postgraduate School, Monterey, CA 93943, USA; [email protected]
Academic Editor: Christof Schneider
Received: 6 September 2016; Accepted: 3 November 2016; Published: 11 November 2016
Abstract: The dielectric constants of nylon fabrics saturated with aqueous NaCl solutions,
Fabric-Superdielectric Materials (F-SDM), were measured to be >105 even at the shortest discharge
times (>0.001 s) for which reliable data could be obtained using the constant current method, thus
demonstrating the existence of a third class of SDM. Hence, the present results support the general
theoretical SDM hypothesis, which is also supported by earlier experimental work with powder and
anodized foil matrices: Any material composed of liquid containing dissolved, mobile ions, confined
in an electrically insulating matrix, will have a very high dielectric constant. Five capacitors, each
composed of a different number of layers of salt solution saturated nylon fabric, were studied, using
a galvanostat operated in constant current mode. Capacitance, dielectric constant, energy density and
power density as a function of discharge time, for discharge times from ~100 s to nearly 0.001 s were
recorded. The roll-off rate of the first three parameters was found to be nearly identical for all five
capacitors tested. The power density increased in all cases with decreasing discharge time, but again
the observed frequency response was nearly identical for all five capacitors. Operational limitations
found for F-SDM are the same as those for other aqueous solution SDM, particularly a low maximum
operating voltage (~2.3 V), and dielectric “constants” that are a function of voltage, decreasing for
voltages higher than ~0.8 V. Extrapolations of the present data set suggest F-SDM could be the key to
inexpensive, high energy density (>75 J/cm3 ) capacitors.
Keywords: dielectric; capacitance; energy storage
1. Introduction
Two parallel, but technically distinct, efforts are underway for improving the energy density of
capacitors: (i) employing electrically conductive materials with the highest possible surface areas as
electrodes (EDLC, also known as supercapacitors); and (ii) finding/inventing electrically insulating
materials with higher dielectric constants. In contrast to EDLC, the electrodes area in the high dielectric
constant capacitors are quite low, approximately equal the macroscopic surface area. Increasing capacitance
via the former route may have reached a limit with the deployment of graphene, “single layer”
electrically conductive graphitic carbon, with a theoretical surface area limit of ~2600 m2 /gm.
Until recently, efforts via the latter route focused primarily on improving barium titanate. Arguably this
approach was not successful and significant improvements in energy storage density were not achieved.
However, the recent invention of super dielectric materials, which employ dispersed liquids containing
dissolved salts, not solids, as the dielectric (SDM) has dramatically improved energy density in
capacitors based on this second approach. Indeed, the published energy density of prototypes of
the two approaches are similar, with graphene based EDLC approaching 450 J/cm3 , and SDM based
Novel Paradigm Supercapacitors (NPS) using aqueous solutions of NaCl in a anodized titania matrix
as the dielectric with energy density of nearly 400 J/cm3 [1,2].
Previously, two types of materials were demonstrated to have dielectric constants greater than
5
10 , and hence qualify as super dielectric materials (SDM): (i) porous oxides (e.g., alumina) filled
Materials 2016, 9, 918; doi:10.3390/ma9110918
www.mdpi.com/journal/materials
Materials 2016, 9, 918
2 of 12
with aqueous salt solutions [3–5], so called Powder SDM (P-SDM); and (ii) anodized titania films,
“Tube” SDM (T-SDM), also filled with aqueous salt solutions [1,2]. The measured dielectric constants of
materials described in those studies was >1010 in many cases, hence orders of magnitude higher than
required to meet the required minimum, greater than 105 , for a material to be classified as an SDM.
In the present study a prediction of the general SDM hypothesis was tested: Any material composed
of liquid containing dissolved, mobile ions, confined in an electrically insulating matrix, will have
a very high dielectric constant. Specifically, a novel type of material that fits the above description
was studied as a dielectric; nylon fabric saturated with aqueous NaCl salt solutions. Five capacitors
composed of layers of this material were studied and each performed as superdielectric over the entire
range of discharge rates studied. Once again, for slow discharge (ca. 100 s), dielectric constants as high
as 1011 were measured, and, even at the shortest discharge rates, 0.001 s, no dielectric constant <107
was measured. Thus, the present work supports the general SDM hypothesis.
Potentially, F-SDM are attractive engineering materials: inexpensive (fabric and salt water), easy
to fabricate, and with a potential for high energy density. However; in order to better assess the
applicability of F-SDM for fast discharge processes, the “roll-off” of capacitance with discharge time
was studied.
The very theory of SDM suggests they may be relatively slow to release energy. SDM are based
on diffusion of ions inside the liquid filling the pores of the electrically insulating matrix material to
form “giant” (ca. microns in length) dipoles upon exposure to an electric field. It takes time for the
ions to move and form the multi-micron length dipoles required to explain super dielectric behavior.
That is, ions initially randomly dispersed in a liquid must travel the order of the thickness of the
dielectric layer (many microns in all cases studied to date) to ultimately create giant dipoles. Similarly,
once formed, it takes time for the dipoles to “dissolve” allowing electrons on the electrodes to release.
This model of multi-micron ionic travel under the influence of an applied field suggests there should
be a time dependence to the capacitance of SDM. Qualitatively, it is anticipated that if the applied
field is switched too quickly from one polarity to the opposite, the dipoles will not achieve their
maximum possible magnitude. That is, the ions that form the dipole will simply not have time to
travel to their equilibrium, fully polarized, positions. This in turn implies smaller effective dipoles at
higher frequency, hence a lower effective dielectric constant. This should lead to a smaller measured
capacitance with increasing frequency. One major objective of the present work is to quantify the effect
of frequency on capacitance.
The present data, based on constant current analysis between 2.3 and 0.1 volts, clearly show there
is a strong roll-off of capacitance for F-SDM. The capacitance, from ~1 volt, at ~10 Hz is of the order
of 1 percent that at ~0.01 Hz or to put it another way, the capacitance for a discharge time of 100 s is
about 100 times greater than the capacitance for a discharge of 0.1 s. The dielectric constant shows
nearly an identical rate of roll off, and the energy density measured over the full voltage range shows
a slightly higher rate of roll off. This suggests the forming process for full strength dipoles in F-SDM
take of the order of tens of seconds. This type of information clearly will help determine the suitability
of F-SDM for particular applications.
2. Results
The capacitance, measured dielectric constant, energy density and power delivery characteristics
of five capacitors were studied as a function of discharge time (DT) defined as the time the voltage takes,
during the discharge half-cycle, to drop from 2.3 to 0.1 volts. These capacitors were designed to be
nearly identical in all respects except for the thickness, corresponding to the number of layers of nylon
employed to create the dielectric layer (Table 1). Typical observed behavior for the “constant current”
tests is shown in Figure 1 for several DT. A material with a voltage independent dielectric constant
will produce a straight line, and it was consistently found that for the F-SDM studied herein, that is
only approximately the case for DT > 50 s. As with previous studies of SDM materials, the dielectric
constant is found to vary with voltage for DT less than about 50 s. Indeed, there are roughly three
Materials 2016, 9, 918
3 of 12
<0.8 volts (Region III). For the present study, the dielectric constant values presented are those of
regions III.
of near constant dielectric constant: 2.3–1.6 V (Region I), 1.6–0.8 volts (Region II) and <0.8 volts
Region
(Region III). For the present study, the dielectric constant values presented are those of Region III.
Table 1. Dielectric and Energy Density as a function of number of layers for 50 s DT.
Table 1. Dielectric and Energy Density as a function of number of layers for 50 s DT.
Approx. Dielectric Constant *
Approx. Energy Density (J/cm3) ^
Layers
Thickness (cm)
DT **
50 s (J/cm3 ) ˆ
** 50 s Constant *
Approx. DT
Dielectric
Approx. Energy
Density
Thickness
(cm)
9
1Layers
0.036
5
×
10
0.7
DT ** 50 s
DT ** 50 s
2
0.074
2 × 1010 9
0.6
1
0.036
0.7
5 × 10
3
0.114
3 × 101010
0.2
2
0.074
0.6
2 × 10
10
0.15
5 3
0.191
5
×
10
0.114
0.2
3 × 1010
11
0.08
10 5
0.38
1.2
× 10
0.191
0.15
5×
1010
* Dielectric
constants
voltage
voltage range, the
10
0.38 listed are for the 1.2
× 1011range 0.8–0.1 volts. Over this 0.08
dielectric
constant
was
in
fact
constant.
As
noted
in
the
text,
the
value
of
the
constant
* Dielectric constants listed are for the voltage range 0.8–0.1 volts. Over this voltage range, thedielectric
dielectric constant
was in factabove
constant.
in energy
the text, density
the valueis
ofnot
the dielectric
decreases
above
~0.8 V. ˆ The
energy
decreases
~0.8AsV.noted
^ The
constantconstant
but decreases
with
increasing
thickness.
density is not constant but decreases with increasing thickness. ** DT—Discharge time. Defined in text.
** DT—Discharge time. Defined in text.
Figure 1. Cont.
Materials 2016, 9, 918
Materials 2016, 9, 918
Materials 2016, 9, 918
4 of 12
4 of 12
4 of 12
Figure 1.
1. Five
54 ss showing
showing one
one cycle,
cycle, and
and illustrating
illustrating three
three
Figure
Five Layers
Layers of
of Nylon
Nylon Cycles:
Cycles: (A)
(A) 12
12 mA,
mA, DT
DT ~~54
Figure
1.regions;
Five Layers
ofmA,
Nylon
Cycles:
(A)
12 mA,
DT(C)
~ 54
s mA,
showing
one
cycle,
and illustrating
three
voltage
(B)
12
DT
~
54
s,
7
cycles;
and
90
DT
~
0.25
s,
7
cycles.
As
shown,
the
voltage regions; (B) 12 mA, DT ~54 s, 7 cycles; and (C) 90 mA, DT ~0.25 s, 7 cycles. As shown, the first
voltage
regions;
(B) 12 at
mA,
DT
~ 54 s,value
7 cycles;
andtypical”
(C) 90 mA,
DT
~used
0.25ins,computation.
7 cycles. As shown,
the
first
~3
cycles
collected
any
current
are
“not
and
not
Note:
Data
~3 cycles collected at any current value are “not typical” and not used in computation. Note: Data were
first
cycles collected
at anyofcurrent
value arethese
“not two
typical” and0.01
not used in
computation.
Note: Data
were~3collected
theof
basis
the smaller
s, in all
cases.
(The
collected
on theon
basis
the smaller
of theseoftwo
values, values,
0.01 volts or volts
0.01 s,orin0.01
all cases.
(The
plotting
were
collected
on
the
basis
of
the
smaller
of
these
two
values,
0.01
volts
or
0.01
s,
in
all
cases.
(The
plottingfor
routine
for this
figure data
lumped
data to approximately
value
routine
this figure
lumped
to approximately
one valueone
every
0.5every
s.) 0.5 s.)
plotting routine for this figure lumped data to approximately one value every 0.5 s.)
One general aspect of SDM behavior not described in earlier papers is the symmetry of
One general
general aspect
aspect of
of SDM
SDM behavior
not described
described in
in earlier papers
papers is
is the
symmetry
of
One
behavior
not
theSupercapacitors
symmetry of
charge/discharge
behavior with
respect
to the polarity
of the voltage.earlier
As Novel Paradigm
charge/discharge
behavior
with
respect
to
the
polarity
of
the
voltage.
As
Novel
Paradigm
charge/discharge
behaviorcreated
with respect
polarityof
of two
the voltage.
As Novel
Supercapacitors
(NPS), that is capacitors
fromtoa the
sandwich
low surface
areaParadigm
electrodes
with an SDM
Supercapacitors
(NPS), that
is capacitors
created from
a sandwich
of two
lowelectrodes
surface area
electrodes
(NPS),
that
is
capacitors
created
from
a
sandwich
of
two
low
surface
area
an SDM
dielectric between, are completely symmetric, the expectation is that the charge/dischargewith
behavior
will
with
an
SDM
dielectric
between,
are
completely
symmetric,
the
expectation
is
that
the
charge/discharge
dielectric
between,
completely
symmetric,
the expectation
is thattothe
behavior
be symmetric
with are
respect
to the sign
of the voltage.
This is shown
be charge/discharge
true in Figure 2. As
shown,will
the
behavior
will be
symmetric
with
respect
to the
sign of
theisvoltage.
This
is shown
to be2.true
in Figure
2.
be
symmetric
with
respect
to
the
sign
of
the
voltage.
This
shown
to
be
true
in
Figure
As
shown,
the
voltage limits were set to be from −2.3 to +2.3 volts, whereas in all prior reports of SDM behavior the
As
shown, the were
voltage
set
from
−2.3whereas
to +2.3 volts,
in all of
prior
reports
of SDM
voltage
setlimits
to be were
from
−2.3to
tobe+2.3
volts,
in all whereas
prior reports
SDM
behavior
the
voltage limits
was never permitted
to drop
below
zero
volts.
behavior
the
voltage
was
never
permitted
to
drop
below
zero
volts.
voltage was never permitted to drop below zero volts.
2.0
2.0
VOLTS
VOLTS
1.0
1.0
0
0
-1.0
-1.0
-2.0
-2.0
500
1000
500
1000
TIME, s
TIME, s
1500
2000
1500
2000
Figure 2. Symmetric Capacitance. NPS capacitors are symmetric, hence can be charged and discharged
Symmetric
Capacitance.
NPS
capacitors
are symmetric,
symmetric,
hence
can be
be
Figure
2. Symmetric
Capacitance.
NPS
are
can
and discharged
from −2.3
to +2.3 Volts.
Data shown
arecapacitors
from a five-layer
sample,hence
operated
at charged
12 mA constant
current
−
2.3
to
+2.3
Volts.
Data
shown
are
from
a
five-layer
sample,
operated
at
12
mA constant
from
−2.3
to
+2.3
Volts.
Data
shown
are
from
a
five-layer
sample,
operated
at
12
mA
constant
current
mode.
current mode.
mode.
2.1. Capacitance
2.1. Capacitance
Capacitance for the voltage Region III is shown in Figure 3 for all five capacitors tested. In all
forthe
the
voltage
Region
III
is shown
inpower
Figure
3 all
forwith
all capacitors
five
capacitors
tested.
In
all
Capacitance
for
voltage
Region
is shown
in aFigure
3 for
five
tested.
In all cases,
cases,Capacitance
the capacitance
“rolls
off”
very III
smoothly
as
law
decreasing
discharge
time.
cases,
the
capacitance
“rolls
off”
very
smoothly
as
a
power
law
with
decreasing
discharge
time.
the
capacitance
“rolls
off”
very
smoothly
as
a
power
law
with
decreasing
discharge
time.
Except
for
Except for the three-layer case, the same power law describes the roll off of all the test capacitors very
Except
for the three-layer
case,power
the same
lawthe
describes
rollthe
offtest
of all
the test capacitors
the
three-layer
case, the same
lawpower
describes
roll offthe
of all
capacitors
very well:very
well:
well:
−0.55
(1)
Capacitance
(1)
100×
× (100/DT)
(100/DT)−0.55
Capacitance= =CC
100
−0.55
(1)
Capacitance = C100 × (100/DT)
Materials 2016, 9, 918
Materials 2016, 9, 918
5 of 12
5 of 12
where C100
is the
the capacitance
capacitance at
at 100
100 s,
s, and
and DT
DT is the discharge time.
time. Roughly,
Roughly, this is equivalent to a
100 is
roll off of 0.55 dB capacitance for 1.0 dB decrease in “period”. The “curve
“curve fit” line of Equation (1) is
the dashed curve in Figure 3. The same relationship also fits the dielectric data, substituting dielectric
Inaddition,
addition, nearly
nearly the
the same
same relationship
relationship fits virtually all
constant for capacitance and “D 100”” for C100
100.. In
the
energy
data
(“E
”
for
C
),
except
the
exponent
is
changed
from
−
to indicating
−0.60, indicating
” for C100100
), except the exponent is changed from −0.55 to0.55
−0.60,
a slightlya
the energy data (“E100100
slightly
fasterroll
energy
faster
energy
off. roll off.
Figure
Time.
ForFor
long
discharge
timetime
(>100
s) capacitance
increases
with
Figure 3.
3. Capacitance
Capacitancevs.
vs.Discharge
Discharge
Time.
long
discharge
(>100
s) capacitance
increases
each
additional
layer,
but
for
all
capacitors,
except
three-layer,
approximately
the
same
simple
power
with each additional layer, but for all capacitors, except three-layer, approximately the same simple
relationship
(slope) (slope)
between
discharge
time and
describes
the capacitive
roll-off
with
power relationship
between
discharge
timecapacitance
and capacitance
describes
the capacitive
roll-off
decreasing
discharge
time.
The
dashed
curve
(above
others)
is
the
fit
to
Equation
(1).
with decreasing discharge time. The dashed curve (above others) is the fit to Equation (1).
2.2.
2.2. Dielectric
Dielectric Constant
Constant
The
The variation
variation of
of the
the dielectric
dielectric constant
constant in
in Region
Region III
III is
is an
an important
important parameter,
parameter, but
but does
does not
not tell
tell
the
entire
story
(Figure
4).
The
dielectric
constant
decreases
with
increasing
voltage,
and
the
ratio
the entire story (Figure 4). The dielectric constant decreases with increasing voltage, and the ratio of
of
the
the dielectric
dielectric constants
constants in
in the
the regions
regions is
is aa function
function of
of DT.
DT. For
For aa DT
DT of
of approximately
approximately 11 s,
s, the
the dielectric
dielectric
constant
for the
the Region
Region III
III is
is twice
twicethat
thatfor
forRegion
RegionIIIIand
andfour
fourtimes
timesthat
thatfor
forRegion
RegionI.I.For
Fora aDT
DTofof0.01
0.01s,
constant for
s,
the
dielectric
constant
for
Region
III
is
three
time
that
for
Region
II
and
10
times
that
of
Region
the dielectric constant for Region III is three time that for Region II and 10 times that of Region I. I.
There
There is
is aa clear
clear trend:
trend: as
as the
the DT
DT gets
gets shorter,
shorter, the
the voltage
voltage drops
drops more
more quickly
quickly in
in Regions
Regions II and
and II
II
relative
to
Region
III.
More
detail
on
this
behavior
is
found
in
earlier
articles
[1–5].
Rather
than
focus
relative to Region III. More detail on this behavior is found in earlier articles [1–5]. Rather than focus
on
complex behavior,
behavior, in
inthis
thispaper,
paper,the
thetrends
trendsinintotal
totalenergy
energydensity,
density,
and
power
a function
on this
this complex
and
power
as as
a function
of
of
DT,
which
includes
contributions
from
all
voltage
regions,
is
reported.
The
total
energy
and
power
DT, which includes contributions from all voltage regions, is reported. The total energy and power
are
voltage
regions.
ThisThis
lumping
is necessary
to more
readily
describe
“nonare “lumped”
“lumped”data
datafrom
fromallall
voltage
regions.
lumping
is necessary
to more
readily
describe
textbook”,
that
is
voltage
dependent
capacitance,
behavior.
Moreover,
the
lumped
values
reflect
net
“non-textbook”, that is voltage dependent capacitance, behavior. Moreover, the lumped values reflect
energy
and and
power
behavior
as a as
function
of the
rate, rate,
and concomitantly
frequency.
The
net energy
power
behavior
a function
ofdischarge
the discharge
and concomitantly
frequency.
lumped
values
are
significant
indicators
of
performance
in
the
likely
applications:
energy
storage
The lumped values are significant indicators of performance in the likely applications: energy storage
and/or
and/orpower
powerdelivery.
delivery.
One
significant
One significant outcome
outcome from
from this
this work,
work, evident
evident from
from Figure
Figure 44 and
and Table
Table 1,
1, are
are the
the remarkable
remarkable
11
dielectric
constants
for
Region
III
at
long
DT
(>50
s)
with
the
highest
directly
observed
value3.5
3.5×× 10
1011
dielectric constants for Region III at long DT (>50 s) with the highest directly observed value
(350
(350 billion)
billion) for
for aa discharge
discharge time
time of
of the
the 10
10 layer
layer F-SDM
F-SDM of
of >200
>200s.s.Clearly,
Clearly,F-SDM
F-SDMare
aresuperdielectrics.
superdielectrics.
One
particular
feature
of
the
dielectric
constants
is
that
they
are
not
constant,
but
rather
increase
with
One particular feature of the dielectric constants is that they are not constant, but rather
increase
increasing
thickness.
It
is
revealing
to
compare
this
behavior
with
that
of
P-SDM
for
which
with increasing thickness. It is revealing to compare this behavior with that of P-SDM for which the
the
“dielectric constants” are constant with thickness [3–5], and with T-SDM [1,2], for which the dielectric
constants were found to increase approximately as 1/t2, where t is thickness of the dielectric layer.
Materials 2016, 9, 918
6 of 12
Materials 2016,
9, 918
6 of 12
“dielectric
constants”
are constant with thickness [3–5], and with T-SDM [1,2], for which the dielectric
constants were found to increase approximately as 1/t2 , where t is thickness of the dielectric layer.
The
increasing values
valuesof
of“dielectric
“dielectricconstant”
constant”for
forthe
the
F-SDM
illustrate
a behavior
“between”
of
The increasing
F-SDM
illustrate
a behavior
“between”
thatthat
of the
the
P-SDM
and
the
T-SDM.
That
is,
any
increase
is
inconsistent
with
observed
P-SDM
behavior,
but
P-SDM and the T-SDM. That is, any increase is inconsistent with observed P-SDM behavior, but the
the
measured
of increase
much
found
for T-SDM.
measured
raterate
of increase
waswas
much
lessless
thanthan
thatthat
found
for T-SDM.
DIELECTRIC CONSTANT
1.10
1.0x10 11
1.00E+11
1.0x1010
1.00E+10
1.0x109
1.00E+09
1 layer
2 layers
1.0x108
1.00E+08
3 layers
5 layers
1.00E+07
0.001
0.01
0.1
1
DISCHARGE TIME, S
10
100
1000
Figure
Figure 4.
4. Dielectric
DielectricConstant
Constantvs.
vs.Discharge
DischargeTime.
Time.In
Inall
allcases,
cases,even
evenfor
fordischarge
dischargetimes
timesof
oforder
order0.001
0.001s,s,
the
>1077.. At
the measured
measured dielectric constant was >10
At relatively
relatively long
long discharge
discharge times,
times, >100
>100 s,
s, the
the dielectric
dielectric
10 in
constants
constants were, remarkably, >1010
inall
allcases.
cases.
2.3.
2.3. Energy
Energy Density
Density
Energy
Energy density
density data
data were
were obtained
obtained in
in two
two steps.
steps. First,
First, the
the area
area under
under the
the entire
entire constant
constant current
current
curve,
2.3–0.1
V,
is
integrated
to
yield
“volt-seconds”,
this
is
multiplied
by
the
constant
current
to
curve, 2.3–0.1 V, is integrated to yield “volt-seconds”, this is multiplied by the constant current to yield
yield
energy
(integration
of
“power”,
where
power
is
volts
×
current).
This
procedure
is
required
energy (integration of “power”, where power is volts × current). This procedure is required because
because
energybecannot
be determined
from simple
the usual
simple relationship
between
and
energy cannot
determined
from the usual
relationship
between energy
and energy
capacitance
capacitance
because
capacitance
changes
with
voltage.
Nonetheless,
there
is
a
smooth
relationship
because capacitance changes with voltage. Nonetheless, there is a smooth relationship between energy
between
energy
andEnergy,
discharge
time. Energy,
like
off smoothly
for
and discharge
time.
like capacitance,
rolls
offcapacitance,
smoothly forrolls
all capacitors
tested
asall
thecapacitors
discharge
tested
as the
discharge
times
reduced
(Figure
5). Once
again,
the roll-off
rates are
nearly “parallel”
times are
reduced
(Figure
5).are
Once
again,
the roll-off
rates
are nearly
“parallel”
independent
of the
independent
of
the
number
of
layers.
An
interesting
feature
of
the
data
is
that,
even
for
the
relatively
number of layers. An interesting feature of the data is that, even for the relatively thick 1 layer capacitor
thick
1 layer
capacitor
(360greater
μ), an energy
greater
thanmeasured
1 J/cm3 was
measured
fortimes.
very
(360 µ),
an energy
density
than 1 density
J/cm3 was
directly
fordirectly
very long
discharge
long
discharge
times.
Although
not
even
close
to
that
recorded
for
T-SDM
[1,2],
relative
to
the
best
Although not even close to that recorded for T-SDM [1,2], relative to the best barium titanate based
barium
titanate
based
electrostatic
capacitors
this
is
a
remarkable
value
[6–8].
electrostatic capacitors this is a remarkable value [6–8].
The
The increasing
increasing energy
energy density
density with
with decreasing
decreasing thickness
thickness suggests
suggests that
that aa thin
thin fabric
fabric type
type SDM
SDM
could
rival
the
TSDM
energy
density
observed
in
earlier
studies.
Indeed,
a
conservative
extrapolation,
could rival the TSDM energy density observed in earlier studies. Indeed, a conservative extrapolation,
based
on aa the
the empirical
empiricalfinding
findingofofone
oneorder
order
magnitude
increase
in net
energy
density
for each
based on
ofof
magnitude
increase
in net
energy
density
for each
one
one
order
of magnitude
decrease
in thickness
(see Table
1),anand
an energy
density
100 sthat
twice
order
of magnitude
decrease
in thickness
(see Table
1), and
energy
density
at 100 at
s twice
at that
50 s,
3 for a 5 micron thick nylon layer and discharge time
3 for
at
50 s, indicates
an density
energy density
of >75
J/cm
indicates
an energy
of >75 J/cm
a 5 micron
thick nylon layer and discharge time of 100 s.
of 100 s.
Materials 2016, 9, 918
7 of 12
Materials 2016, 9, 918
Materials 2016, 9, 918
7 of 12
7 of 12
1
1
ENERGY DENSITY, J/CM3
ENERGY DENSITY, J/CM3
0.1
0.1
0.01
1 layer
0.01
2 layers
1 layer
3 layers
2 layers
5 layers
3 layers
10 layers
5 layers
0.001
0.001
0.0001
0.001
0.0001
0.001
10 layers
0.01
0.1
1
10
100
DISCHARGE TIME, S
0.01
0.1
1
10
100
S
Figure 5. Energy Density vs. Discharge Time. DISCHARGE
The initialTIME,
energy
density at long discharge times is
Figure 5. Energy Density vs. Discharge Time. The initial energy density at long discharge times is
highest for one layer and decreases with each added layer, but the power law relationship (slope)
Figurefor
5. Energy
Density
vs. Discharge
initial
energy
at long
times
is
highest
one layer
and decreases
withTime.
each The
added
layer,
but density
the power
law discharge
relationship
(slope)
between energy density and discharge time is very similar for all capacitors studied.
highestenergy
for onedensity
layer and
added
layer,
power law
relationship (slope)
between
anddecreases
dischargewith
timeeach
is very
similar
forbut
allthe
capacitors
studied.
between energy density and discharge time is very similar for all capacitors studied.
2.4. Power Density
2.4. Power Density
The Density
power delivery as a function of frequency clearly increases as the discharge times are
2.4. Power
The power
delivery
as a function
of frequency
increases as
discharge
are reduced
reduced
(Figure
6). Power
was computed
simplyclearly
by determining
thethe
total
energy,times
computed
as
The power delivery as a function of frequency clearly increases as the discharge times are
described
above,
and
dividing itsimply
by the by
discharge
time. The
in power
delivered
a discharge
(Figure
6). Power
was
computed
determining
theincrease
total energy,
computed
asat
described
above,
reduced (Figure 6). Power was computed simply by determining the total energy, computed as
of 0.01its,by
a typical
high power
application
time,in
is power
more than
an orderatofa magnitude
greater,
andtime
dividing
the discharge
time.
The increase
delivered
discharge time
of in
0.01 s,
described
above,
and
dividing
it by the
discharge
time.
The increase
in powerincrease
delivered
at
a discharge
all
cases,
than
the
power
delivered
during
a
100
s
discharge.
The
measured
in
power
with
a typical high power application time, is more than an order of magnitude greater, in all cases, than
time
of 0.01 s,
typical
high power
application
time, the
is more
than anroll
order
magnitude
greater,
decreasing
DTa is
anticipated
system for which
capacitance
offpower
isofless
than
1 dB
per 1 dBinDT
theallpower
delivered
during
a for
100any
s discharge.
The
measured
increase
in
with
decreasing
cases,
than
the
power
delivered
during
a
100
s
discharge.
The
measured
increase
in
power
with
decrease in discharge time. That is, as long as the exponent in Equation (1) is greater than negative
is anticipated
for
any
system
for
which
the
capacitance
roll
off
is
less
than
1
dB
per
1
dB
decrease
decreasing
DTwill
is anticipated
fordecreasing
any systemDT
forvalue.
which the capacitance roll off is less than 1 dB per 1 dB in
one, power
increase with
discharge
That is,time.
as long
asis,
the
inexponent
Equationin(1)
is greater
negative
power
decreasetime.
in discharge
That
asexponent
long as the
Equation
(1) than
is greater
thanone,
negative
0.5
will
increase
with
decreasing
DT
value.
one, power will increase with decreasing DT value.
0.5
POWER, W/CM3
POWER, W/CM3
0.05
0.05
1 layer
0.005
2 layers
1 layer
3 layers
0.005
25layers
layers
310
layers
layers
5 layers
0.0005
0.001
10
layers
0.01
0.1
1
DISCHARGE TIME, S
10
100
0.0005
Figure 6. Power vs.
Discharge Time. The power observed decreases monotonically with the number
0.001
0.01
0.1
1
10
100
of layers. Notably, the power increases with DISCHARGE
decreasing
times. This is because the roll-off
TIME,discharge
S
of energy is slower than the “roll-off” of discharge time.
Figure 6. Power vs. Discharge Time. The power observed decreases monotonically with the number
Figure
6. Power vs. Discharge Time. The power observed decreases monotonically with the number of
of layers. Notably, the power increases with decreasing discharge times. This is because the roll-off
layers. Notably, the power increases with decreasing discharge times. This is because the roll-off of
of energy is slower than the “roll-off” of discharge time.
energy is slower than the “roll-off” of discharge time.
Materials 2016, 9, 918
Materials 2016, 9, 918
8 of 12
8 of 12
It should
should be noted
noted that
that the
the typical
typical “Ragone
“Ragone chart” (Figure 7) presentation
presentation of data reflects the
the
same
aspects
of
behavior
seen
in
the
above
capacitance,
energy
and
power
vs.
discharge
time
plots
[9].
same
Specifically,
Specifically, as
as frequency
frequency increases
increases (nearly
(nearly equivalent
equivalent to
to decreasing
decreasing discharge
discharge time),
time), the
the capacitance
capacitance
rolls
linear
relationship
between
energy
and capacitance,
as frequency
increases,
energy
rolls off.
off.Given
Giventhe
the
linear
relationship
between
energy
and capacitance,
as frequency
increases,
density
decreases.
As
shown
in
Figure
7,
the
Energy
vs.
Power
curve
can
be
readily
fit
with
a
singlea
energy density decreases. As shown in Figure 7, the Energy vs. Power curve can be readily fit with
power
law forlaw
F-SDM.
In contrast,
for many
apparently,
there is athere
change
“power
single power
for F-SDM.
In contrast,
forsupercapacitors,
many supercapacitors,
apparently,
is a in
change
in
law”
such
that
power
output
reaches
a
maximum
even
as
energy
continues
to
drop.
“power law” such that power output reaches a maximum even as energy continues to drop.
1
ENERGY J/CM3
0.1
0.01
0.001
1 Layer
10 Layers
0.0001
0.0001
0.001
0.01
POWER J/CM3 S
0.1
Figure 7.
7. Ragone
Ragone Plot
Plot Presentation.
Presentation. The
The data
data obtained
obtained for
for two
two of
of the
the F-SDM
F-SDM based
based capacitors
capacitors studied
studied
Figure
are
shown.
As
with
other
types
of
capacitors,
as
the
energy
density
decreases,
the
power
output
are shown. As with other types of capacitors, as the energy density decreases, the power output
increases.
The
basis
for
this
relationship
is
that
as
the
discharge
time
decreases,
the
energy
density
increases. The basis for this relationship is that as the discharge time decreases, the energy density
decreases, but
but more
more slowly.
slowly. Thus,
Thus, the
the energy
energy delivered
delivered per
per unit
unit time
time (power)
(power) increases
increases with
with decreasing
decreasing
decreases,
discharge time.
time.
discharge
3. Discussion
3.
Discussion
The basis
basis for
for the
the SDM
SDM hypothesis:
hypothesis: Dielectrics
capacitance by
reducing the
the field
field
The
Dielectrics increase
increase capacitance
by reducing
everywhere
relative
to
the
no
dielectric
case.
Specifically,
any
dielectric
exposed
to
a
field
created
by
everywhere relative to the no dielectric case. Specifically, any dielectric exposed to a field created by
charge on
on electrodes
electrodes forms
forms dipoles
dipoles that
that create
create fields
fields “opposite”
“opposite” the
the field
field generated
generated by
by the
the charge
charge on
on
charge
the
electrodes.
Relative
to
the
“dielectric
free”
situation,
this
reduces
the
net
electric
field
strength
the electrodes. Relative to the “dielectric free” situation, this reduces the net electric field strength
everywhere. InInturn,
this
reduces
the the
work,
that is
the is
integral
of electric
field over
anyover
path any
(voltage),
everywhere.
turn,
this
reduces
work,
that
the integral
of electric
field
path
necessary to
bring another
charge
to thecharge
electrode
surface.
Thus,surface.
when a dielectric
is present
between
(voltage),
necessary
to bring
another
to the
electrode
Thus, when
a dielectric
is
parallel plates,
more
charge
is necessary
to create
same voltage
reached
in thevoltage
absencereached
of a dielectric.
present
between
parallel
plates,
more charge
is the
necessary
to create
the same
in the
This is of
equivalent
to This
increasing
the charge/voltage
ratio,
which is the
definition
of increasing
absence
a dielectric.
is equivalent
to increasing the
charge/voltage
ratio,
which is the
definition
capacitance.
any dielectric
produce
dipoles
larger than
those
in solids
of
increasingHence,
capacitance.
Hence, that
any can
dielectric
that
can produce
dipoles
larger
than barium
those intitanate
solids
should have
a dielectric
constant
higherconstant
than thathigher
foundthan
for that
understanding
barium
titanate
should have
a dielectric
that material.
found forThe
thatabove
material.
The above
motivated the design
of the
wit: Awork.
nylonTofabric
saturated
with saturated
“salt water”
should
understanding
motivated
thepresent
designwork.
of theTo
present
wit: A
nylon fabric
with
“salt
have
dipoles
as
long
as
the
thickness
of
the
nylon,
at
least
thousands
of
times
larger
than
that
of any
water” should have dipoles as long as the thickness of the nylon, at least thousands of times larger
solid, including BaTiO3. Thus the dielectric constant of an SDM is predicted to be much higher than
than that of any solid, including BaTiO3 . Thus the dielectric constant of an SDM is predicted to be
that of any solid.
much higher than that of any solid.
The first major empirical finding supports the above predicted high dielectric: the measured
The first major empirical finding supports the above predicted high dielectric: the measured
dielectric constants are above 1077 (vs. <1044 for barium titanate) even at very short discharge times,
dielectric constants are above 10 (vs. <10 for barium titanate) even at very short discharge times,
establishing F-SDM as a new class of SDM. Second, the capacitance and dielectric constant changes
establishing F-SDM as a new class of SDM. Second, the capacitance and dielectric constant changes
with voltage in a very similar pattern to that for aqueous based P-SDM and T-SDM. That is, the
with voltage in a very similar pattern to that for aqueous based P-SDM and T-SDM. That is, the
capacitance changes with voltage, but can be well modeled as constant in three voltage zones for
capacitance changes with voltage, but can be well modeled as constant in three voltage zones for
discharge times less than ca. 25 s. Third, the relatively slow increase in dielectric constant with
increasing layer thickness demonstrates this class of materials is not like P-SDM, and, the measured
Materials 2016, 9, 918
9 of 12
discharge
Materials
2016,times
9, 918 less
than ca. 25 s. Third, the relatively slow increase in dielectric constant 9with
of 12
increasing layer thickness demonstrates this class of materials is not like P-SDM, and, the measured
with
dielectric
layer
thickness
indicates
behavior
that is
distinct
from
and clear
clear drop
dropininenergy
energydensity
density
with
dielectric
layer
thickness
indicates
behavior
that
is distinct
T-SDM.
The behavior
of F-SDM
is unique
(Figure
8). Fourth,
there
is aistypical
pattern
decrease
in
from
T-SDM.
The behavior
of F-SDM
is unique
(Figure
8). Fourth,
there
a typical
pattern
decrease
capacitance,
dielectric
constant
and
energy
density
with
decreasing
discharge
time.
Fifth,
extrapolating
in capacitance, dielectric constant and energy density with decreasing discharge time. Fifth,
the energy density
values
for thevalues
relatively
thick
samples
studied
suggests
that
capacitors
with
extrapolating
the energy
density
for the
relatively
thick
samples
studied
suggests
thatmade
capacitors
3 ).
3
thin
fabric
layers
(ca.
5
µ)
might
have
exceptional
energy
densities
(ca.
>75
J/cm
made with thin fabric layers (ca. 5 μ) might have exceptional energy densities (ca. >75 J/cm ).
NORMALINZED ENERGY DENSITY
1
0.1
Measured
Standard Theory
SDM Theory
0.01
1
NUMBER OF LAYERS
10
Figure
toto
SDM
theory
(solid
line),
thethe
energy
density
is
Figure 8.
8. Energy
Energy Density
Densityvs.
vs.Thickness.
Thickness.According
According
SDM
theory
(solid
line),
energy
density
independent
of
thickness.
According
to
standard
theory
(dashed
line)
the
energy
density
falls
as
is independent of thickness. According to standard theory (dashed line) the energy density falls as
inverse
inverse thickness
thickness squared.
squared. The
The actual
actual data
data (points/dotted
(points/dotted line)
line) for
for energy
energy density
density as
as aa function
function of
of
dielectric
thickness
fall
between
these
two
theories.
dielectric thickness fall between these two theories.
A low surface area nylon mesh containing a salt solution is predicted to have a high dielectric
A low surface area nylon mesh containing a salt solution is predicted to have a high dielectric
constant by the SDM theory, but not by any other theory. A brief review shows no viable alternative
constant by the SDM theory, but not by any other theory. A brief review shows no viable alternative
models in the literature. Alternative theories of high dielectric constants include high dielectric
models in the literature. Alternative theories of high dielectric constants include high dielectric
constants at the percolation threshold [10–12], a nano metal particle model [13–15], and a quantum
constants at the percolation threshold [10–12], a nano metal particle model [13–15], and a quantum
“surface state” model for colossal dielectric behavior [16–20]. As discussed elsewhere, none of these
“surface state” model for colossal dielectric behavior [16–20]. As discussed elsewhere, none of these
models appears to provide a reasonable framework for understanding the current system [1–5].
models appears to provide a reasonable framework for understanding the current system [1–5].
Specifically, there is no conceivable percolation process [21], there are no metal particles, and there
Specifically, there is no conceivable percolation process [21], there are no metal particles, and there
are no surface states associated with the liquid phase. Finally, NPS are clearly not a variation on
are no surface states associated with the liquid phase. Finally, NPS are clearly not a variation on
electric double layer capacitors (EDLC), as the NPS capacitors are missing two elements of standard
electric double layer capacitors (EDLC), as the NPS capacitors are missing two elements of standard
supercapacitors: a high surface area conductive electrode (e.g., graphene), and a thin separator to
supercapacitors: a high surface area conductive electrode (e.g., graphene), and a thin separator to
allow ion transport, but to keep the two high surface area electrodes from touching/shorting [22,23].
allow ion transport, but to keep the two high surface area electrodes from touching/shorting [22,23].
Another issue requiring discussion is the “in-between” PSDM and TSDM behavior observed. In
Another issue requiring discussion is the “in-between” PSDM and TSDM behavior observed.
P-SDM dipole length, critical in determining dielectric constant, is established by the pore structure
In P-SDM dipole length, critical in determining dielectric constant, is established by the pore structure
of the powder employed. Pore structure does not change with thickness, hence the dielectric constant
of the powder employed. Pore structure does not change with thickness, hence the dielectric constant
is unchanging, as observed [3–5]. Energy density drops, consequently, as expected (thickness−−22) for
is unchanging, as observed [3–5]. Energy density drops, consequently, as expected (thickness ) for
any standard dielectric material. In T-SDM, the dipole length is determined by the length of the open
any standard dielectric material. In T-SDM, the dipole length is determined by the length of the
pores in the titania structure. These pores run the width of the oxide layer. Thus, as the dipole length
open pores in the titania structure. These pores run the width of the oxide layer. Thus, as the dipole
increases exactly at the same rate as the dielectric layer thickness, the amount of salt also increases
length increases exactly at the same rate as the dielectric layer thickness, the amount of salt also
proportional to the pore length, and concomitantly so does the dielectric constant. This is not
increases proportional to the pore length, and concomitantly so does the dielectric constant. This is
“standard” behavior for dielectric materials, but is anticipated for TSDM as explained in earlier
not “standard” behavior for dielectric materials, but is anticipated for TSDM as explained in earlier
publications [1,2]. Moreover, the theory of energy density for a system of ever increasing pore length
is that energy density should be largely independent of dielectric thickness, as observed.
It is argued that pore length change with increasing dielectric layer thickness for an F-SDM is
between those observed for PSDM and TSDM. The pore length is anticipated to increase with
dielectric layer thickness, but at a slower rate than that of the dielectric layer itself. That is, it is
Materials 2016, 9, 918
10 of 12
publications [1,2]. Moreover, the theory of energy density for a system of ever increasing pore length is
that energy density should be largely independent of dielectric thickness, as observed.
It is argued that pore length change with increasing dielectric layer thickness for an F-SDM is
between those observed for PSDM and TSDM. The pore length is anticipated to increase with dielectric
layer thickness, but at a slower rate than that of the dielectric layer itself. That is, it is anticipated
that there is not perfect alignment between the layers of nylon fabric. Hence, liquid containing holes
(~50% by area) in the mesh only partially “line up”, limiting the average pore length increase with
the addition of each layer. This limits the average increase in dipole length with increasing number of
layers. The average dipole length does increase with each layer added, but not as quickly as the overall
thickness. This qualitative explanation is consistent with the observed change in energy density with
number of layers.
The final point of interest is the decrease in dielectric constant and concomitantly energy density
with decreases in discharge time. A reasoned extrapolation of the SDM theory suggests this is to be
expected. Indeed, each time the applied field is reversed the original dielectric constant will not be
recovered until all the ions in the solution within a pore physically move by diffusion, or field assisted
convection, to the “other side” of the pore. This takes time. If the discharge time is too short, complete
reversal of the ions will not take place, leading to a smaller net dipole strength within the pore, and
hence a lower net dielectric constant. As the charge time gets shorter, the net motion is less, and
hence the value of the dielectric constant should fall as the charge time decreases. This model leads
to several questions for future work. First, will other dielectrics for which physical motion of ions is
required show a similar fall off of capacitance as discharge time is reduced? In particular, will a similar
discharge time dependence of dielectric constant be observed in EDLC? In addition, if the charging
time is increased at any given current, will the energy density at any given discharge current increase?
4. Materials and Methods
Capacitor Fabrication: All capacitors, so-called Novel Paradigm Supercapacitors (NPS), were
created from three components: nylon fabric, aqueous solutions of NaCl (30 wt %), and GTA grade
Grafoil [24,25] electrodes (0.4 mm thick × 5 cm diameter), a commercially available, paper-like
(~0.4 mm thick), moderate surface area (~20 m2 /gm) material composed of compressed graphite
(>99%). Nylon fabric squares, ~5.1 cm on a side, nominal thickness 0.36 mm, 50% open space, were
dipped into the salt solution for approximately one hour, and then smoothed onto a Grafoil electrode.
For the multi-layer samples, additional salt solution saturated fabric layers were added one at a time.
The second Grafoil electrode was then placed on top and the thickness of the capacitor determined
from an average of four measurements with a micrometer. Once the capacitor “sandwich” was created,
in order to retard drying, it was placed on a small plastic block, and the block placed in a plastic bag
containing water saturated cloth (3).
Measurement: The capacitive behavior was determined using a BioLogic Model SP 300
Galvanostat (Bio-Logic Science Instruments SAS, Claix, France) in constant current charge/discharge
mode. Data were collected from the smaller of these intervals: 0.01 s or 0.01 V Other methods were
judged to be inappropriate for determination of characteristics as prior SDM studies clearly show
that the capacitance of NPS is a strong function of voltage. Hence, methods such as impedance
spectroscopy that measure the dielectric constant at a single voltage, generally 0 ± 15 mV [8,9], do
not yield the full story regarding energy storage characteristics. For a reliable understanding of the
behavior expected for energy storage devices, the behavior over the entire voltage operating range
must be studied [6,26,27].
At each selected constant current at least ten complete cycles were recorded, and generally
twenty. This method does not permit the selection of the discharge rate; however, the discharge rate
is a function of the discharge current. Thus, the trend in dielectric constants and energy density was
determined by variation of the controllable parameter, discharge current. In all cases the charging
current was the same magnitude as the discharge current, but of opposite sign.
Materials 2016, 9, 918
11 of 12
There are several potential sources of error in determination of capacitance and dielectric constant.
First, is the measured current. Observation clearly showed this to be within 3% of the nominal current
in all cases. A second source is in measurement of the capacitor thickness. In all cases, that was
determined to be no greater than ±0.03 mm, which, for all capacitors studied, was less than 10%.
Variation in the slope of the discharge (dI/dV), determined by measurement of multiple discharges,
was found to be no more than 5%, indicating precision. Consideration of all these errors suggests the
values presented in the results section are ±15% of true values. It is notable in this regard that a 220 µF
commercial superdielectric capacitor (Maxwell) was studied as a “control” in the same circuit, over the
same voltage range. The measured capacitance at 0.03 Hz was 215 µF, and was constant over nearly
the entire voltage range.
The voltage range in most cases was selected to be between 2.3 and 0.1 volts (Figure 1) as earlier
studies showed this range to be compatible with aqueous solution based SDM [1–5]. For illustration
purposes, one example of a charge/discharge cycle operating between −2.3 and +2.3 volts is shown
(Figure 2).
Acknowledgments: The author is grateful that this study was fully supported by the Office of Naval Research
and the Naval Research Program administered by the Naval Postgraduate School, Monterey, CA, USA.
Conflicts of Interest: The author declares no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Gandy, J.; Quintero, F.; Phillips, J. Testing the Tube Super Dielectric Material Hypothesis: Increased Energy
Density Using NaCl. J. Electron. Mater. 2016. [CrossRef]
Quintero, F.; Phillips, J. Tube-Super Dielectric Materials: Electrostatic Capacitors with Energy Density Greater
than 200 J·cm−3 . Materials 2015, 8, 6208–6227.
Fromille, S.; Phillips, J. Super Dielectric Materials. Materials 2014, 7, 8197–8212. [CrossRef]
Quintero, F.; Phillips, J. Super Dielectrics Composed of NaCl and H2 O and Porous Alumina. J. Electron. Mater.
2015, 44, 1367.
Jenkins, N.; Petty, C.; Phillips, J. Investigation of Fumed Aqueous NaCl Superdielectric Material. Materials
2016, 9, 118. [CrossRef]
Reynolds, G.J.; Krutzer, M.; Dubs, M.; Felzer, H.; Mamazza, R. Electrical properties of thin-film capacitors
fabricated using high temperature sputtered modified barium titanate. Materials 2012, 5, 644–660. [CrossRef]
Kinoshita, K.; Yamaji, A. Grain-size effects on dielectric properties in barium titanate ceramics. J. Appl. Phys.
1976, 47, 371–373. [CrossRef]
Arlt, G.; Hennings, D. Dielectric properties of fine-grained barium titanate ceramics. J. Appl. Phys. 1985, 58,
1619–1625. [CrossRef]
Christen, T.; Carlen, M.W.J. Theory of Ragone Plots. J. Power Sources 2000, 91, 210–216. [CrossRef]
Bergman, D.J.; Imry, Y. Critical behavior of the complex dielectric constant near the percolation threshold of
a heterogeneous material. Phys. Rev. Lett. 1977, 39, 1222. [CrossRef]
Efros, A.L.; Shklovskii, B.I. Critical Behaviour of Conductivity and Dielectric Constant near the
Metal-Non-Metal Transition Threshold. Phys. Status Solidi B 1976, 76, 475–485. [CrossRef]
Efros, A.L. High volumetric capacitance near the insulator-metal percolation transition. Phys. Rev. B 2011, 84,
155134. [CrossRef]
Pecharroman, C.; Esteban-Betegon, F.; Bartolome, F.; Lopez-Esteban, J.F.; Moya, J. New Percolative BaTiO3 –Ni
Composites with a High and Frequency-Independent Dielectric Constant (r ≈ 80000). Adv. Mater. 2001, 13,
1541–1544. [CrossRef]
Pecharroman, C.; Esteban-Betegon, F.; Jimenez, R. Electric field enhancement and conduction mechanisms in
Ni/BaTiO3 percolative composites. Ferroelectrics 2010, 400, 81–88. [CrossRef]
Saha, S.K. Observation of giant dielectric constant in an assembly of ultrafine Ag particles. Phys. Rev. B 2004,
69, 125416. [CrossRef]
Lunkenheimer, P.; Krohns, S.; Riegg, S.; Ebbinghaus, S.G.; Reller, A.; Loidl, A. Colossal dielectric constants in
transition-metal oxides. Eur. Phys. J. 2009, 180, 61–89. [CrossRef]
Materials 2016, 9, 918
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
12 of 12
Lunkenheimer, P.; Bobnar, V.; Pronin, A.V.; Ritus, A.I.; Volkov, A.A.; Lodl, A. Origin of apparent colossal
dielectric constants. Phys. Rev. B 2002, 66, 052105. [CrossRef]
Hu, W.; Lau, K.; Liu, Y.; Withers, R.L.; Chen, H.; Fu, I.; Gong, B.; Hutchison, W. Colossal Dielectric Permittivity
in (Nb + Al) Codoped Rutile TiO2 Ceramics: Compositional Gradient and Local Structure. Chem. Mater.
2015, 27, 4934–4942. [CrossRef]
Krohns, S.; Lu, J.; Lunkenheimer, P.V.; Brize, V.; Autret-Lambert, C.; Gervais, M.; Gervais, F.; Bouree, F.;
Porcher, F.; Loidl, A. Correlations of structural, magnetic, and dielectric properties of undoped and doped
CaCu3 Ti4 O12 . Eur. Phys. J. B 2009, 72, 173–182. [CrossRef]
Kant, C.; Rudolf, T.; Mayr, F.; Krohns, S.; Lunkenheimer, P.; Ebbinghaus, S.G.; Loidl, A. Broadband dielectric
response of CaCu3 Ti4 O12 : From dc to the electronic transition regime. Phys. Rev. B 2008, 77, 045131.
[CrossRef]
Gerenrot, D.; Berlyand, L.; Phillips, J. Random network model for heat transfer in high contrast
composite materials. IEEE Trans. Adv. Packag. 2003, 26, 410–416. [CrossRef]
Conway, B.E.E. Electrochemical Supercapacitors. In Scientific Fundamentals and Technological Applications;
Spinger Science and Business Media: Berlin/Heidelberg, Germany, 2013.
Gualous, H.; Bouquain, D.; Berthan, A.; Kauffmann, J.M. Experimental study of supercapacitor serial
resistance and capacitance variations with temperature. J. Power Sources 2003, 123, 86–93. [CrossRef]
Phillips, J.; Clausen, B.; Dumesic, J.A. Iron pentacarbonyl decomposition over Grafoil. Production of small
metallic iron particles. J. Phys. Chem. 1980, 84, 1814–1822. [CrossRef]
Phillips, J.; Dumesic, J.A. Iron pentacarbonyl decomposition over Grafoil: II. Effect of sample outgassing on
decomposition kinetics. Appl. Surf. Sci. 1981, 7, 215–230. [CrossRef]
Kim, Y.; Kathaperumal, M.; Chen, V.W.; Park, Y.; Fuentes-Hernandez, C.; Pan, M.J.; Kippelen, B.;
Perry, J.W. Bilayer Structure with Ultrahigh Energy/Power Density Using Hybrid Sol–Gel Dielectric and
Charge-Blocking Monolayer. Adv. Energy Mater. 2015. [CrossRef]
Gogotski, Y.; Simon, P. True performance metrics in electrochemical energy storage. Science 2011, 334, 917–918.
[CrossRef] [PubMed]
© 2016 by the author; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).